stereoelectronic effects in nucleosides and nucleotides and their

Stereoelectronic Effects in

Nucleosides and Nucleotides and

their Structural Implications

(including Appendix on literature up to 2005)

Christophe Thibaudeau, Parag Acharya and Jyoti Chattopadhyaya*

*To whom correspondence should be addressed.

E-mail: [email protected]

F +4618554495

T +46184714577

www.boc.uu.se

Department of Bioorganic Chemistry

Biomedical Center, Uppsala University, Sweden

Chattopadhyaya et al, "Stereoelectronic Effects in Nucleosides & Nucleotides and their Structural Implications", Dept of Bioorganic Chemistry, Box 581, Uppsala University, S-75123 Uppsala, Sweden, Ver 160205 [email protected]

2

Christophe Thibaudeau, Parag Acharya and Jyoti Chattopadhyaya*

Dept of Bioorganic Chemistry

Uppsala University

Biomedical Center, Box 581

S-751 23 Uppsala

Dr C. Thibaudeau has completed his Ph.D

in Feb, 1999 at the Dept of Bioorganic Chemistry

under the supervision of Prof J. Chattopadhyaya

Dr P. Acharya has completed his Ph.D

in Dec, 2003 at the Dept of Bioorganic Chemistry under

the supervision of Prof J. Chattopadhyaya

Dr J. Chattopadhyaya is professor of Bioorganic Chemistry

at the Uppsala University

Use the following for citation in your reference:

Uppsala University Press

First Edition: 1999; Second Edition, 2005

Copyright © May, 1999 by J. Chattopadhyaya, C. Thibaudeau and P. Acharaya

ISBN 91-506-1351-0

pp 166

Chattopadhyaya et al, "Stereoelectronic Effects in Nucleosides & Nucleotides and their Structural Implications",

Dept of Bioorganic Chemistry, Box 581, Uppsala University, S-75123 Uppsala, Sweden, Ver 160205 [email protected]

3

PREFACE

The three essential components of DNA and RNA are

the aglycone, the pentofuranose sugar in β-D stereochemistry

and the phosphodiester. The phosphates at the backbone

makes DNA/RNA to behave as polyelectrolyte, the pentose

sugar gives the intrinsic flexibility with relatively low energy

barrier for conformational interconversions (according to the

demand of the environment)compared to the hexopyranoses,

and the aglycones help in the self-assembly process throgh the

stacking and hydrogen bonding. Thus, the stability of the

folded structure of nucleic acids is usually attributed to the

interplay of various forces such as hydrogen bonding,

stacking, electrostatics, hydrophobics and hydration.

Are the physico-chemical roles of the aglycone, sugar

and phosphate correlated and interdependent? If so, what are

their thermodynamics? Energetically, how do they respond to

the local change of the environment, and affect the

recognition process that culminate into function?

It has long been qualitatively known that they are not

isolated structural elements. The electronic nature of the

aglycone dictates certain preferred sugar torsions, which are

again correlated with certain phosphate torsions – they are

interdependent. X-ray crystallographic and NMR studies have

demonstrated how the sugar and phosphate moieties can

adopt different conformations in DNA and RNA. It has been

shown that the rotations about C-O and P-O ester bonds are

restricted, and certain sugar-phosphate torsions are preferred

over the others.

Prior to the work summarized here from the Uppsala

group, very little was known about the dynamic

interdependency of the conformational changes between the

three essential components of DNA and RNA or the energetics

involved in this process. Through our solution NMR studies,

we have attempted to show that the electronic nature of the

aglycone as well as those of all other substituents of the sugar

moiety dictate the intrinsic character of the pentose-sugar

conformation, which in turn dictate the phosphate backbone

torsions. The important aspect of this dynamic

interdependency of the aglycone-sugar-phosphate orientation

is that it can be modulated by the change of the environment

with a relatively much smaller energy penalty compared to the

hexopyranose counterparts. This intrinsic flexibile character of

the pentose ring is perhaps the evolutionary basis for their

adoption in nucleic acids for storage of genetic information,

almost error-free transcription, as well as for selective gene

expression in the translation machinery. The present evidences

suggest that the mechanism of this modulation of the pentose

conformation in DNA and RNA is stereolectronic in character,

and the concerted conformational change is unidirectional,

originating from the aglycone to the sugar and further to the

phosphate. In this monograph, we have explored the nature of

the stereoelectronic forces arising from the gauche and

anomeric interactions, that are partly responsible for the self-

organization of nucleosides and nucleotides. Most importantly,

we have experimentally measured the strength and the

interplay of these interactions, for the first time, and shown the

significance of their effects in dictating the overall dynamics

and the structure of nucleos(t)ides and their analogs. This

process has enabled engineering of specific conformations in a

predictable manner in nucleos(t)ides by having appropriate

substituent(s) in the sugar moiety.

The intrinsic flexibility of pentoses in natural

nucleosides and nucleotides, owing to their lower

energy barrier for interconversions compared to the

hexopyranoses, is dictated by the energetics of

stereoelectronic effects, which simply can be tuned

and modulated by choice of substituents and their

ionization state as well as by their complexation with

potential ligands present in the medium.

Stereoelectronic effects operate by appropriate

orbital overlap between the donor and acceptor

through-bond and through-space interactions. The

strength of the stereoelectronic effects induced

stabilization is proportional to the square of the

overlap of the donor and acceptor orbitals, and is

inversely proportional to their energy difference.

Finally, a new section has been added

covering the latest in the field (Appendix: Chapter

10), which makes this monograph current till the end

of 2002.

We believe this monograph is of considerable

value to those medicinal and pharmaceutical

chemists in the academia or in the industry, who

wish to understand fundamental mechanism involved

in the design of structure of modified nucleos(t)ides,

and use this knowledge to rationally develop

antisense, RNAi or triplexing agents or specific

enzyme inhibitors.

We thank Swedish Natural Science Research

Council (Vetenskapsrådet), Swedish Organization

for Strategic Research (SSF) and Uppsala University

for generous financial support in our research

projects


4

Table of Contents

page

1. Introduction. Stereoelectronic effects in hexopyranoses 8

1.1 Equatorial monosubstituted cyclohexane 9

1.2 The Edward-Lemieux effect 9

1.3 ∆∆G° versus ∆∆H° estimates for the anomeric effect 10

1.4 The generalized anomeric effect 11

1.5 Influence of the nature and configuration of substituents on the anomeric effect 11

1.6 Effect of the polarity of the solvent on the anomeric effect 13

1.7 The reverse anomeric effect 13

1.8 Hyperconjugation as the origin of the anomeric effect 43

1.9 The nature of the electron lonepairs 16

1.10 Dipole-dipole (electrostatic) interactions as origin of the anomeric effect 17

1.11 Alternative explanations for the origin of the anomeric effect 18

1.12 The gauche effect 19

1.13 The energetics of the gauche effect 20

1.14 Possible origins of the gauche effect 20

2. Stereoelectronic effects in nucleosides and nucleotides 22

2.1 Structure of nucleic acids 22

2.2 The pseudorotation concept 22

2.3 The two-state N � S equilibrium in β-nucleosides 25

2.4 The two-state N � S equilibrium in α-nucleosides 27

2.5 The two-state N S equilibrium in carbocyclic nucleosides 27

2.6 Energy barriers of the pseudorotation cycle of β-D-nucleosides 34

2.7 Steric effect of the nucleobase on the sugar conformation 35

2.8 O4'-C1'-N1/9 stereoelectronic effect (the anomeric effect) 36

2.9 Effect of the aglycone on the conformation of nucleos(t)ides and oligos 38

2.9.1 Configuration-dependent sugar conformation in furanosides 39

2.9.2 Effect of electron-withdrawing nucleobases on pseudorotamer populations 39

2.9.3 Effect of the protonated nucleobase on pseudorotamer populations 40

2.9.4 Effect of base-modifications on the stability of nucleic acids 42

2.10 The gauche effects in α- and β-nucleosides 43

2.10.1 2'-Deoxynucleosides 43

2.10.2 Ribonucleosides and nucleotides 44

2.11 The gauche effects of sugar substituents and the self-organization of DNA/RNA 44

2.11.1 Studies on nucleosides 44

2.11.2 Studies on oligonucleotides 46

3. Methods to quantitate stereoelectronic effects in nucleos(t)ides 46

3.1 Thermodynamics of the two-state N � S equilibria 46

3.2 1H-NMR spectra [temperature, pD, ligand-dependent spectra of nucleos(t)ides] 47

3.3 Pseudorotational analyses of 3JHH with PSEUROT and some practical hints 48

3.3.1 Incorporation of coupling constant errors in PSEUROT calculations 48

3.3.2 Parameters to be fixed or optimized during PSEUROT calculations 48

3.3.3 Electronegativity of the substituents on each HCCH fragment 49

3.3.4 Priority rule to number the substituents on the HAC1C2HB fragment 49



5

3.3.5 Translation of HCCH torsion angles into endocyclic torsion angles 50

3.3.6 General operational conditions for PSEUROT 50

3.4 Generalised Karplus-type equation 53

3.4.1 EOS Karplus-Altona equation for 3JHH 53

3.4.2 Reparametrized EOS equation for carbocyclic nucleosides 54

3.4.3 Karplus equation for interpretation of 3JHF 55

3.4.4 Refined Karplus equation for 3JHH based on Fourier formalism 55

3.5 Translation of experimental 3JHH and 3JHF into pseudorotational parameters 56

3.6 Principle of iterations with PSEUROT 66

3.7 Estimation of ∆Hº, ∆Sº and ∆Gº of the N� S equilibrium 61

3.7.1 Methodology 61

3.7.2 Accuracy of thermodynamics 62

3.7.3 Influence of λN1/9 on the thermodynamics 62

3.7.4 Influence of the nature of the aglycone dictates the thermodynamics 66

3.8 New Karplus equation to interpret 3JHF coupling constants 66

3.8.1 Dataset of (3JHF,ΦHF) pairs for monofluoronucleosides 67

3.8.2 Parametrization of the Karplus equation 68

3.8.3 Pseudorotational analyses of 3JHF in fluoronucleosides to validate the Karplus equation 69

4. The quantitation of the stereoelectronic effects in nucleos(t)ides 71

4.1 Quantitation of the anomeric and gauche effects by regression analyses 71

4.1.1 Stereoelectronic effects in neutral β-D-Ns 71

4.1.2 Effect of the 5'CH2OH versus 5'CH2OMe 72

4.1.3 Effect of the nucleobase 73

4.1.4 Gauche effects 73

4.1.5 Energetic equivalence of mirror-image β-D-dNs and β-L-dNs 74

4.1.6 3'-phosphate has stronger gauche effect than 3'-hydroxy 74

4.1.7 Stereoelectronic effects in neutral α-D-N 75

4.1.8 Weakening of the effects of 5'CH2OH and 5'CH2OMe in α-nucleosides 75

4.1.9 Weakening of the effect of the nucleobase in α-nucleosides 75

4.1.10 Weaker 3'-hydroxy gauche effect in α- compared with β-D-Ns 76

4.1.11 Limitations of regression analysis to quantitate stereoelectronic effects 76

4.2 Quantitation of the anomeric and gauche effects by pairwise comparisons 78

4.3 Strengths of ∆H° and -T∆S° to ∆G° of pseudorotation of neutral nucleosides 78

4.4 Nucleobase-dependent anomeric effects in neutral nucleosides 78

4.4.1 Effect of the 5'CH2OH versus 5'CH2OMe 82

4.4.2 The effect of the nucleobase is sugar-dependent 82

4.5 The gauche effect of the 3'-substituent in neutral dNs 85

4.5.1 The influence of the C1'-substituent 85

4.5.2 3'-substituent electronegativity dictates 3'-gauche effect 86

4.5.3 Stronger gauche effect in nucleosides than in 1,2-difluoroethane 90

4.6 The 2'-OH effect in ribonucleos(t)ides is nucleobase-dependent 90

4.7 3'-gauche effect modulation by 2'-OH in ribonucleos(t)ides 92

4.8 Drive of pseudorotation in β-nucleosides by the nature of the nucleobase 93


6

4.8.1 The two-state N � S equilibrium is evidenced by pKa values from ∆G° 93

4.8.2 Identical pKas of the nucleobases in 2',3'-dideoxy, 2'-deoxy and ribo series 93

4.8.3 Anomeric effect in β-D-ddNs is modulated by the nature of the nucleobase 95

4.8.4 The orbital mixing as the origin of the O4'-C1'-N1/9 anomeric effect 95

4.8.5 No reverse anomeric effect in pentofuranosyl nucleosides 97

4.8.6 Variable tunablity of anomeric effect in β-D-ddNs, β-D-dNs and β-D-rNs 97

4.8.7 Correlation of the electronic nature of aglycone with pseudorotational state 99

5. Comparison of stereoelectronic effects in α- and β-D-nucleosides 102

5.1 The relative magnitude of the anomeric and gauche effects in Neutral state 102

5.1.1 Anti orientation of the nucleobase in α/β-D-ddN and α/β-D-dN 103

5.1.2 The balance of ∆H° and -T∆S° in neutral α- and β-D-N 104

5.1.3 Weakening of 5'-substituent effect in α- compared with β-nucleosides 104

5.1.4 3'-gauche effect weakens anomeric effect in α-D-dN compared to α-D-ddN 105

5.1.5 Weakening of the 3'-gauche effect in α-D/L-dN compared with β-D/L-dN 105

5.2 The relative magnitude of the anomeric and gauche effects in the ionic states 105

5.2.1 Virtually identical pKa values of the nucleobase in α- and β-nucleosides 105

5.2.2 Predominant enthalpy over entropy in the ionic states of α-D-ddN and -dN 105

5.2.3 Weaker anomeric effect in α-D-ddN gives poorer flexibility 105

5.2.4 The interplay of pD-independent ∆H° and ∆S° in α-D-dN 106

5.2.5 Poor correlation of the nature of aglycone with pseudorotation in α-D-N 106

6. Quantitation of the anomeric effects in C- and N-nucleosides 107

6.1 The anomeric effect in C-nucleosides 107

6.1.1 Effect of the C1'-pyrimidine aglycone on the conformation of the sugar 108

6.1.2 Effect of the C1'-purine aglycone on the drive of the sugar conformation 110

6.1.3 pD-tunable anomeric effect in C-nucleosides 110

6.1.4 Transmission of the nature of the C-aglycone drives the N� S equilibrium in C-nucleosides 111

6.1.5 Estimates for the thermodynamics of the N � S equilibrium 111

6.1.6 Enhanced anomeric effect upon protonation of the aglycone 111

6.1.7 Weaker anomeric effect upon deprotonation of the aglycone 112

6.1.8 Quantitation of the anomeric effect in C-nucleosides 112

6.1.9 Comparison of pD-induced flexibility in C- and N-nucleosides 112

6.1.10 Correlation of the effect of the aglycone and its electronic nature 113

6.2 Quantitation of anomeric effect in N-nucleosides using C-nucleoside as reference 114

7. The interdependency of the sugar and phosphate conformation 118

7.1 Methods to assess the preferred conformation across the phosphate backbone 119

7.2 No correlation of sugar and phosphate conformation in 2'-dN-3'-ethylphosphates 120

7.3 Interaction of 2'-OH with vicinal 3'-phosphate in ribonucleotides 121

8. Application of stereoelectronic effects in oligonucleotides 124

8.1 Design of antisense oligonucleotides via the gauche engineering 124

8.2 Fused nucleosides to engineer preorganized DNA structures 127

8.3 Enzyme recognition by fused carbocyclic nucleosides of fixed conformation 127

8.4 The conformational transmission in the self-cleaving Lariat-RNA 127



7

8.5 The conformational transmission by dihydrouridine in RNA 129

8.6 Preferential recognition of 3'-anthraniloyladenosine by Elongation Factor Tu 131

8.7 Conformational changes in nucleotides induced by interaction with metal ions 133

8.8 RNA as molecular wire 136

8.9 The importance of O4' in the self-organization of oligo-DNA 146

9. References

148

10. Appendix (New Additions of references during 2003-4) 160

10.1 The nature of Gauche Effect between 2'-OH and Glycosyl-Nitrogen.

10.2 The pseudorotational barrier of the pentose moiety of nucleosides and nucleotides.

10.3 The hydrogen bonding, hydration and pKa of 2'-OH in adenosine and adenosine 3'-

ethylphosphate.

10.4 Latest articles (1998 – 2002) on the aspects of Anomeric (AE) and Gauche (GE) Effects and

discussions/comments on these works

10.5 The sensitivity of the RNase H discriminates the local structure changes owing to

conformational transmission induced by 3'-endo sugar constrained nucleotides in the

antisense strand of the antisense-RNA hybrid duplex.

10.6 The influence of flouro substitution at the sugar moiety in modulating the furanose

onformation of flourinated nucleosides.


8

1. Introduction. Stereoelectronic effects in hexopyranoses

Stereoelectronic effects1-7 can be defined as phenomenological effects, which allow to explain

"seemingly abnormal" conformational preferences. Their influence upon the structure of transition

states and the course of chemical reactions is known. The extensive application by Deslongchamps2

and Kirby1 of the principles underlying anomeric hyperconjugation to rationalize and predict reaction

pathways has given rise to the theories of stereoelectronic control, kinetic anomeric effect and to the

antiperiplanar lone-pair hypothesis. The preponderant role of steroelectronic effects in the course of

biochemical reactions has been pointed out for many enzymes, such as ribozymes8-10, serine

proteases11, lysozyme12, or ribonuclease9,10,13. Most of the studies in the field of stereoelectronic

effects have been devoted to six-membered rings, and the state-of-the-art for such systems has been

the subject of many review articles1-7. On the other hand, to the best of our knowledge, no unifying

model regarding the influence of stereoelectronic effects on the conformation of heterocyclic five-

membered rings at the monomeric or oligomeric level has yet come out in the literature. This is

possibly due to the fact that, in contrast with rather unflexible and consequently conformationally

biased six-membered rings, saturated five-membered rings, such as the pentofuranose moiety in

nucleosides and nucleotides, are generally involved in complex conformational equilibria.

Some early experimental studies have shown in a qualitative manner how the anomeric and

gauche effects control the bias of conformational equilibria of the pentofuranose moiety in

nucleos(t)ides and of other saturated five-membered rings. A significant progress has however only

been accomplished during the last decade with the development of new methodologies in our

laboratory which give accurate estimates of the magnitude of energetics of stereoelectronic effects

driving the two-state North �South pseudorotational equilibrium in nucleos(t)ides14-45.

At the oligomeric level, recent investigations46-54 on modified oligonucleotides have been

conducted as a part of the pursuit for developing antisense compounds with increased nuclease-

resistance compared with the parent DNA and RNA, and with improved hybridization with the target

RNA. The analysis of the thermodynamic properties and of the three-dimensional structure55-61 of

these oligonucleotides has shown that a particular modification of one of the three components of the

constituent nucleotides, i.e. the nucleobase, the sugar moiety and the phosphate backbone alters both

their stability and overall structure. The actual participation of stereoelectronic effects to the observed

structural changes has been adressed only qualitatively46,49,62-64.

The potential power of the application of the gauche / anomeric engineering strategy to design

new oligonucleotides endowed with therapeutic properties is evident. A detailed appraisal of the

qualitative and quantitative results from the early as well as from state-of-the-art studies on the

manifestation, magnitude and origin of stereoelectronic effects in the pentofuranose ring in

nucleosides and nucleotides is therefore required in order to understand its role in the self

organization of DNA, RNA and their analogues. This is the purpose of the present monograph, which

is organized as follows: (i) We first introduce the concept of stereoelectronic effects by summarizing

the most important findings of conformational studies on pyranose derivatives (Section 1). (ii) The

main results of qualitative conformational studies on nucleosides and nucleotides showing the

influence of the nature of the substituents at C1'-C4' on the conformation of the constituent

pentofuranose sugar are subsequently presented and analyzed in detail (Section 2). (iii) Sections 3 - 7



9

are devoted to the quantitation of the thermodynamics of stereoelectronic effects driving

pseudorotational equilibria in nucleos(t)ides using the methodology (described in Section 3)

developed by us in Uppsala. In Sections 3 - 7, we show how specific conformational preferences of

nucleos(t)ides can be engineered by altering either the nature or the configuration of one of their

constituents, i.e. heterocyclic nucleobase, pentofuranose moiety or C2' and C3' substituents, or the pH

of the medium. Evidences regarding the transmission of information between the basic components

within the nucleotidyl unit are presented in Sections 4 - 7. (iv) We finally present recent works, both

from our laboratory and others, on the impact of the anomeric and gauche effects on the

threedimensional structure and biological function of oligonucleotides (Section 8).

Stereoelectronic effects can be classified into a few categories: (i) The original Edward-

Lemieux effect65-67 which designates the preference of the electronegative methoxy or acetoxy group

at C2 in 2-substituted tetrahydropyran for the axial over the equatorial orientation. (ii) The generalized

anomeric effect (GAE), which is an extension of the original Edward-Lemieux effect in the case of

any acyclic or cyclic R-X-A-Y system. (iii) The reverse anomeric effect allows to rationalize a

conformational preference opposite to that expected on the basis of the (generalized) anomeric effect,

i.e. an increased tendency of the anomeric substituent in R-X-A-Y fragment to assume an equatorial

orientation in comparison with its steric effect. (iv) The gauche effect refers to a stereoelectronic

preference for conformations in which the best donor lonepair or bond is antiperiplanar to the best

acceptor bond.

1.1 Equatorial monosubstituted cyclohexane

Except in rare cases (such as bulky V-shaped alkyl groups68), the most stable conformation of

monosubstituted cyclohexane (1) is the chair form E1 in which the substituent X takes up an

equatorial orientation (Fig 1A). This is owing to the fact that in the axial chair conformation A1, X

exerts energetically unfavourable steric repulsions with the axially oriented H3 and H5. The tendency

of X in cyclohexane to adopt an equatorial orientation is dictated by its effective size and can be

estimated from the Gibbs free-energy (i.e. ∆G°steric(cylcohex.)) of the A1 � E1 equilibrium (Eq 1):

∆G°steric(cyclohex.) = -RT ln [E1]/[A1] ..... Eq 1.

In a study69 based on temperature-dependent 13C-NMR spectra of CFCl3-CDCl3 solutions, it

has been shown that ∆G°steric(cyclohex.) at 300 K is ≈ -7.3 kJmol-1 (X = Me), -7.5 kJmol-1 (X = Et)

and -9.3 kJmol-1 (X = iPr).

1.2 The Edward-Lemieux effect

The acid hydrolysis of methyl β-pyranosides of glucose, mannose and galactose, in which the

anomeric methoxy group is equatorial, is faster than that of the α-counterparts, in which it is axial,

which has led Edward65 to propose that the axial orientation of the anomeric substituent in

pyranosides is more stable than the equatorial one. This is opposite to what one would predict on the

basis of simple steric effect rule (vide supra). Edward has attributed this unusual conformational

behaviour to the destabilization of the equatorial conformer by electrostatic repulsions between the

anomeric substituent and the ring dipole induced by the presence of the endocyclic oxygen atom. In

his studies on various acetylated α- versus β-aldohexopyranoses, Lemieux66,67,70 also found that the

acetoxy substituent at C2 adopts preferentially an axial orientation. He introduced the term "anomeric


10

effect" to designate this preference. An extension of the original65-67,70 Edward-Lemieux effect to the

analysis of the conformational equilibrium of 2-methoxytetrahydropyran (2) allows to predict

correctly71 that the axial chair A2 is more stable than its equatorial counterpart E2 (Fig 1B). The Gibbs

free-energy (∆G°heterocycle) of the A2 � E2 conformational equilibrium in 2 is calculated according to

Eq 2: ∆G°heterocycle = -RT ln [E2]/[A2] ..... Eq 2

Figure 1. Axial � Equatorial anomeric conformational

equilibrium in six-membered rings. (A) The drive of the A1 �

E1 equibrium in monosubstituted cyclohexane (1) toward the

equatorial conformer E1 minimizes steric repulsions of the

anomeric group with H3 and H5. (B) The O-C-Ome anomeric

effect in 2-methoxytetrahydropyran (2) pushes its conformational

equilibrium toward the A2 form in spite of unfavourable steric

repulsions, which are reduced in the E2 chair.

The original definition of the anomeric effect

therefore implies that the stereoelectronic component of the O-C-O effet prevails over the

counteracting steric effect. The Gibbs free-energy of the anomeric effect (∆∆G°AE, kJmol-1) in 2 is

estimated using Eq 3:

∆∆G°AE = ∆G°heterocycle - F * ∆G°steric(cyclohex.) - 0.08 ..... Eq 3

where F (= 1.53) stands for Franck's correction factor72 and accounts for the fact that in 2 steric

repulsions generated by the axial orientations of X, H3 and H5 are greater than in the parent

monosubstituted cyclohexane (1), owing to the shorter C-O bond in 2 compared with C-C bond in (1).

1.3 ∆∆G° versus ∆∆H° estimates for the anomeric effect

The question of which thermodynamic quantity [i.e. either enthalpy (∆∆H°), or Gibbs free-

energy (∆∆G°), Table 1] is adequate to estimate the magnitude of the anomeric effect that drives the

conformational equilibrium of 2-substituted tetrahydropyran has been adressed in many studies71,75,91,

92. An initial work by Booth et al 91 based on temperature-dependent 13C-NMR spectra showed that -

T∆S° overrides the negligible ∆H° in the drive of the conformational equilibrium of 2-

methoxytetrahydropyran (2) in CDCl3-CFCl3. It was therefore suggested that the stabilization of A2

over E2 form of 2 was the result of the entropy term only. Lemieux71 however pointed out that such a

negligible ∆H° value is only obtained when 2-methoxytetrahydropyran is solvated by hydrogen-

bonding or polar solvents, and he showed that in CCl4 the origin of the anomeric effect is mainly

enthalpic. The relative preference69 for equatorial over axial chair forms of monosubstituted

cyclohexane (1) carrying either Me, Et or iPr group is strongly temperature-dependent, as the result of

entropy effects. It is therefore preferable to estimate the magnitude of the anomeric effect in terms of

∆∆H°ΑΕ (Eq 4) rather than ∆∆G°AE (Eq 3) in order to eliminate any misleading entropy-based

contribution92:m∆∆H°AE = ∆H°heterocycle - ∆H°steric(cyclohex.) ..... Eq 4

where ∆H°steric(cyclohex.) and ∆H°heterocycle represent the ∆H° contributions to ∆G° of the

conformational equilibria for substituted cyclohexane (Fig 1A) and tetrahydropyran (Fig 1B),

respectively. Juaristi et al 86 have recently estimated the strength of the enthalpic S-C-P(O) anomeric

effect in 2-(diphenylphosphinoyl)-1,3-dithiane by replacing ∆H°steric(cyclohex.) in Eq 4 by

F*∆H°steric(cyclohex.) in order to take into consideration Franck's argument72 regarding different

H X

H

OMe

O

H

H X

O

OMe

(A)

(B)

1: A1 chair 1: E1 chair

2: A2 chair 2: E2 chair

ΔGo

steric (cyclohex.) = -AX

ΔGo

heterocycle



11

effective steric sizes of the anomeric substituents in substituted cyclohexane and tetrahydropyran

(Table 1).

Enthalpic anomeric stabilizations have been found experimentally for substituted heterocyclic

six-membered rings, such as 2-substituted tetrahydropyrans, 2-carbomethoxy- and 2-cyano-

piperidines87 and they have been extensively reviewed4,6 (Table 1).

1.4 The generalized anomeric effect

The concept of generalized anomeric effect93-96 results from an extension of the original

Edward-Lemieux effect to describe the preference for gauche over trans orientation of R-X with

respect to A-Y in R-X-A-Y fragments. R-X-A-Y may be either acyclic or constitute a part of a ring

other than the original substituted tetrahydropyran. X designates an element with at least one pair of

non-bonding electrons, R and A have intermediate electronegativities and Y is more electronegative

than A. Typically1-6, R stands for C or H, A is a tetrahedral (anomeric) center such as C or Si, X is

either O, N or S whereas Y designates F, Cl, Br, O, N or P. Further manifestations and extensions of

the generalized anomeric effect include the “benzylic anomeric effect” 97 which allows to explain the

preferred perpendicular orientation of C-X bond with respect to the plane of the benzene ring (X =

S(O)Me, SO2Me, Cl, etc) in ArCH2X compounds. The preference of chlorine and methoxy in α-

chloro and α-methoxycyclohexanone oximes for an axial orientation has been attributed to the

“vinylogous anomeric effect” 98, whereas the “syn anomeric effect” is responsible for the relatively

stable synperiplanar orientation of the nitrogen lonepair with respect to the C-F bond in

fluoromethylamine, as shown by ab initio calculations99. When R-X-A-Y is part of a heterocycle and

both X and Y possess lonepairs of electrons, the conformation of the heterocycle is driven by the fine

balance between the endo and exo anomeric effects3, which operate in opposite directions: The endo

anomeric effect involves one of the lonepairs of the endocyclic atom X, which tends to be

antiperiplanar with respect to the exocyclic A-Y bond, whereas for the exo anomeric effect it is one of

the electron pairs of the exocyclic Y atom which adopts an anti orientation with respect to the

endocyclic R-X bond. Depending on the relative magnitudes of endo and exo anomeric effects, the

overall estimate (i.e. endo + exo ) for the magnitude of the anomeric effect may be either positive or

negative3,4,6.

1.5 Influence of the nature and configuration of substituents on the anomeric effect

The magnitude of the anomeric effect depends on the nature of the anomeric group, or other

substituents and on their relative configurations. It is also modulated by the polarity and nature of the

solvent.

Conformational studies on acetylated and benzoylated β-D-ribo and xylopyranose derivatives

have shown that the magnitude of the anomeric effect is regularly increased as the electron-

withdrawing character of the anomeric substituent increases, H100,101 < OMe102 < OAc103 < OBz104 <

Halogen101. Similarly, other experimental studies have shown that the preference of the anomeric

substituent X for axial orientation in 2-substituted tetrahydropyran decreases steadily when X is

changed: Br73,74, Cl73,74 > OMe76 > OH80 > (CH2)2N81 > MeHN81. The relative


12

Table 1. The magnitude of the anomeric effect in heterocyclic 6-membered ringsa

Axial Equatorial

Z

Y

Y

R

R'

XZ

R'

R

X

ΔGo

heterocycle

R R' Y Z X ∆∆G°AE1 (ref.) b,c ∆∆G°AE2 b,d ∆∆H°AE (ref.) e

H H O CH2 Br f ≈ 9.6 (73) / 13.4 (74) 10.8 l -

H H O CH2 Cl f ≈ 9.6 (73) / 11.3 (74) 10.8 l 8.9 m (75)

H H O CH2 MeO 8.0 (76) 10.3 3.1 m (75)

H H O CH2 EtO 7.5 (76) 9.8 -

H H O CH2 Me3CO 5.8 (76) 8.1 -

H H O CH2 PhO 6.7 (77) 8.3 -

H H O CH2 AcO 5.8 (78) 7.6 -

H H O CH2 MeS 6.2 (79) 8.5 -

H H O CH2 HO 3.9 (80) 6.2 2.5 m (75)

H H O CH2 (CH2)2N 4.1 (81) 7.6 -

H H O CH2 MeHN 1.5 (81) 4.4 ≈ 0 m (75)

H H O CH2 MeO2C -0.5 (82) 2.4 -

Me H O CH2 HO 5.0 (78) 7.3 -

Me H O CH2 MeO 7.7 (83) 10.0 -

Me H O CH2 EtO 7.2 (83) 9.5 -

Me H O CH2 AcO g 5.7 (84) 7.6 -

Me H O CH2 MeO2C h 0.1 (85) 3.0 -

Me H O CH2 Cl f 11.3 (74) 12.3 -

Me H O CH2 Br f 13.4 (74) 14.6 -

Me H O CH2 I 13.0 (74) 14.0 -

H Me O CH2 MeO 7.2 (83) 9.5 -

H Me O CH2 EtO 7.0 (83) 9.2 -

H Me O CH2 AcO g 6.0 (84) 7.8 -

H Me O CH2 MeS 5.6 (83) 7.9 -

H Me O CH2 Me3CS 5.8 (83) 8.1 -

H Me O CH2 MeO2C h 0.1 (85) 3.0 -

H H S S POPh2 i 14.2 n (86)

H H O CH2 COOMe j 0.5 n (6,87)

H H NH CH2 COOMe j 4.5 n (6,87)

H H O CH2 CN k 2.3 n (6,87)

H H NH CH2 CN k 10.1 n (6,87)

a In kJ mol-1. Other data on the anomeric effect can also be found in ref.3. NMR spectra have been recorded in CCl4

unless stated otherwise. ∆G°

steric(cyclohex.) values (Eq 1, Fig. 1A) are taken from refs. 88-90. b ∆∆G°

AE1 and

∆∆G°

AE2 values represent the free-energy of the Y-C-X anomeric effect in each compound. c ∆∆G°

AE1 values have been

estimated using the equation: ∆∆G°

AE1 = ∆G°

heterocycle - ∆G°

steric(cyclohex.) where ∆G°

heterocycle is the free-energy of

the Axial � Equatorial equilibrium of the heterocyclic six-membered ring (Eq 2, Fig 1B). ∆G°

steric(cyclohex.) is the

estimate for the steric effect of X in the substituted cyclohexane counterpart. For (CH2)2N, ∆G°

steric(cyclohex.) has been



13

assumed equal to that of (CH3)2N. See ref. 3 and the refs. indicated in parentheses in col. 6 for the original ∆G°

heterocycle

values. d ∆∆G°

AE2 has been calculated according to Eq 3. See ref. 3 and the refs. indicated in parentheses in col. 6 for

the original ∆G°

heterocycle values. e ∆∆H°

AE values designate the enthalpy of the Y-C-X anomeric effect in each

compound. f Neat. g In acetic acid. h In methanol. i In CDCl3. j In ether/toluene-d8. k In CFCl3/CDCl3.

l

Using ∆G°

heterocycle values from ref. 73. m In CFCl3/CDCl3. ∆∆H°

AE has been estimated from the simple relation:

∆∆H°

AE = ∆H°

heterocycle - ∆H°

steric(cyclohex.) (Eq 4) where ∆H°

heterocycle is the enthalpy of the Axial � Equatorial

equilbrium of the heterocyclic six-membered ring. ∆H°

steric(cyclohex.) is the enthalpy of the steric effect of X in the

substituted cyclohexane counterpart. n ∆∆H°

AE has been derived from the equation: ∆∆H°

AE = ∆H°

heterocycle - F x

∆G°

steric(cyclohex.), by analogy with the method described in footnote d for free-energy estimates of the anomeric effect

(see ref. 6 and 86 for the values of F).

populations of chair conformations with axially or equatorially oriented anomeric substituent in

pyranose derivatives are also controlled by the electronegativity of the substituents at C4 and C53.

1.6 Effect of the polarity of the solvent on the anomeric effect

The influence of the polarity of the solvent upon the magnitude of the anomeric effect has

been experimentally evidenced in numerous NMR studies on substituted heterocyclic 6-membered

rings such as 2-hydroxy-80,84, 2-alkoxy-71,83,105 and 2-alkylthio-tetrahydropyran79,83, 2-alkoxy-1,3-

dioxane83, 5-halogen-, 5-methoxy- and 5-ethoxy-2-isopropyl-1,3-dioxane95. Polar solvents stabilize

the relatively more polar conformation, i.e. the equatorial chair, more efficiently than the apolar ones.

Therefore the anomeric effect is weakened as the dielectric constant of the solvent increases. In the

case of other substituents at C2 in tetrahydropyran, such as N(CH2)281, the modulation of the bias of

the conformational equilibrium is much reduced. The influence of the solvent on the magnitude of the

anomeric effect should however be considered with caution, especially when there are no accurate

estimates for the actual change of the size (or bulkiness, i.e. A-value) of the anomeric group as a

function of the polarity of the solvent.

1.7 The reverse anomeric effect

An electronegative anomeric group in substituted tetrahydropyrans preferentially adopts an axial

orientation, despite unfavorable steric interactions, owing to the anomeric effect. However, it has been

claimed that if this electronegative substituent is positively charged, the bias of the conformational

equilibrium is just opposite to that predicted in terms of anomeric effect, i.e. the anomeric group

prefers the equatorial orientation more than expected on the basis of steric effect alone 7,106-110. This

tendency has been attributed to the reverse anomeric effect (see Section 4.8 for a comparison with the

pentofuranose system with the protonated aglycone30). The concept of reverse anomeric effect has

been introduced by Lemieux106 to explain the preference of pyridinium in N-(tetra-O-acetyl-α-D-

glucopyranosyl)-pyridinium or 4-methylpyridinium bromides and N-(tri-O-acetyl-α-D-2-deoxy-2-

iodo-mannopyranosyl)-pyridinium perchlorate in aqueous solution for an equatorial orientation, as

shown by his conformational analysis based on the interpretation of vicinal proton-proton coupling

constants. A reverse anomeric effect has also been noted in the case of other pyranosides107,108 with a

pyridinium group at C2. However, whether the preference of pyridinium for the equatorial orientation

is the result of (i) a specific stereoelectronic interaction, or (ii) simply the necessity of the anomeric


14

substituent to avoid steric repulsions disfavouring an axial position as a result of a possible increase of

steric bulk of the anomeric group in the protonated (P) compared to neutral (N) state, or (iii) a specific

solvation of the cation, is still a matter of debate7, especially since no reverse anomeric effect has

been observed for the corresponding glucopyranosylammonium ions, in which the bulkiness of the

anomeric group is comparatively significantly reduced111.

The increase (experimentally evidenced by the change of vicinal 3JHH by 2-3 Hz) of the

population of chair conformers with equatorial anomeric substituents in N-(tetra-O-acetyl-α-D-

glucopyranosyl)imidazolium94, N-(tetra-O-acetyl-α-D-mannopyranosyl) imidazolium94, and N-(tri-O-

acetyl-α-D-xylopyranosyl)imidazolium107 in comparison with their neutral imidazole counterparts in

lipophilic media (i.e. in CDCl3 and Me2CO-d6) has also been attributed to the reverse anomeric

effect. In this context, it is important to point out that the positive charge on the remote nitrogen in

imidazolium does not increase dramatically the effective size of imidazolium in comparison with

imidazole, as shown recently by Perrin et al. 112 with help of a 1H-NMR titration method at low

temperature. However, the situation in the lipophilic medium changes completely when the coupling

constants of the nonacetylated parent compounds are measured in aqueous solution 113: No significant

change (well within the experimental accuracy) in 3JHH is found between their P and N states.

Another investigation by Perrin's laboratory114 has shown that the tendency of imidazolium in N-(D-

glucopyranosyl)imidazolium and its tetraacetate to adopt an axial orientation in DMSO-d6, CD3OD

and aqueous solution is slightly greater than that of imidazole in the neutral counterparts. This result,

which is in agreement with what one would expect basing on the simple anomeric effect rule, violates

the concept of the reverse anomeric effect. It also contradicts the initial results of Lemieux94. On the

other hand, Thatcher et al110 have shown that in the case of N-(tri-O-acetyl-α-D-

xylopyranosyl)imidazole in CDCl3 there is a clear enhancement (by ≈ 35 %) of the population of

equatorial chair 1C4 upon addition of trifluoroacetic acid (TFA) with respect to the neutral

counterpart, and this enhancement is not the result of the change of the ionic strength of the solution

induced by addition of TFA. This work confirms the initial findings of Paulsen et al107 and supports

the existence of a reverse anomeric effect for the xylopyranose derivative.

1.8 Hyperconjugation as the origin of the anomeric effect

As yet, no consensus has been reached regarding the origin of the anomeric effect in the

hexose system, but recent evidences from our lab on the pentose systems in nucleosides and

nucleotides point to the orbital mixing as the most probable reason for the anomeric effect (see

Section 4.8). Two models have received outstanding recognition among all those proposed for the

origin of the anomeric effect:

(i) The anomeric effect has often been interpreted as the result of a stabilizing

hyperconjugative interaction115,116 between the filled orbital of one of the electron lonepairs (nX) of

the endocyclic heteroatom X with the antibonding orbital (σ∗A-Y

) of the A-Y bond.

(ii) Alternatively, the anomeric effect can also be attributed to the attempt of the system to

minimize electrostatic repulsions65,117,118 between the dipole formed by the A-Y bond and the dipole

induced by the presence of the heteroatom X. (iii) A few other explanations have also been proposed,

such as Eliel's119 "rabbit-ear effect", which also involves electrostatic interactions, however this time



15

directly between lonepairs of electrons. Finally, 4e- orbital mixing destabilizing effects120 have also

been advocated.

Lucken originally explained116 the unusually low nuclear quadrupole resonance frequencies of

35Cl in X-C-Cl fragments of α-halogenoethers by an overlap between a p-type orbital of the

heteroatom X and the antibonding orbital (σ*C-Cl

) of the C-Cl bond. His original interpretation has

given rise to the very popular molecular orbital overlap model which is currently used to explain the

anomeric effect [Fig 2(A-C2)]. In α-halogenoethers, the preference of the halogen atom for an axial

orientation is concomittant with a characteristic shortening of the C-O bond and a lengthening of the

C-Cl bond115 compared with tetrahedral values (Fig 2E).

Figure 2. The rationalization of the anomeric effect in

6-membered rings with help of the delocalization115 and

molecular orbital overlap116 models. (A) Overlap of an

electron lonepair orbital of the endocyclic oxygen with

the σ*C-Cl antibonding orbital of C-Cl bond in 2-

chlorotetrahydropyran (using sp3 hybridized oxygen

lonepairs). (B) The magnitude of the anomeric effect is

proportional to S2 (S is the overlap between the oxygen

lonepair orbital and the σ*C-Cl orbital) and inversely

proportional to the energy difference between both

orbitals, i.e. ∆E(σ*C-Cl - nO)4,5. The nO σ*C-Cl

interaction results in the formation of two new orbitals.

(C1) and (C2) Orientation of the electron lonepairs of the

endocyclic oxygen [either sp3 hybridized, i.e. 1nsp3 and

2nsp3 in (C1) or sp2 hybridized, i.e. 1nsp2 (p-type) and 2nsp2

(s-type) in (C2)] with respect to the exocyclic C-Cl bond.

The nO σ*C-Cl overlap is maximal in the axial conformer

[A(C1) in (C1) or A(C2) in (C2)] in which the torsion

angle β [defined as 1nsp3-O1-C2-Cl in (C1) or 1nsp2 (p-

type)-O1-C2-Cl in (C2)] is closest to 180°, whereas in the

other conformer [E(C1) in (C1) or E(C2) in (C2)] the

oxygen lonepairs are either gauche (for both nsp3 orbitals

in E(C1) and for the 2nsp2 (p-type) orbital in E(C2)) or cis

(in the case of the 2nsp2 (s-type) orbital in E(C2)) to the

exocyclic C2-Cl bond. The anomeric effect therefore

favours the axial over the equatorial conformer. When the

oxygen lonepairs orbitals are sp2 hybridized as in (C2),

the magnitude of the anomeric effect is expected to be reduced in comparison with a situation in which lonepairs are sp3

hybridized as in (C1), owing to the fact that β is closer to 180° in A(C1) than in A(C2). (D) The anomeric effect has been

alternatively explained by the hyperconjugation of one of the endocyclic oxygen lonepairs to the exocyclic C-Cl bond,

which corresponds to the double bond � no-bond resonance in terms of valence bond theory. (E) Shortening of the

endocyclic bond C2-O and concomitant lengthening of the exocyclic bond C2-F in the axial conformer A3 of tri-O-

benzoyl-2-fluorotetrahydropyran compared with the equatorial counterpart E4 of tri-O-acetyl-2-fluorotetrahydropyran,

supporting the hyperconjugative origin of the anomeric effect101.

The hyperconjugation model (Fig 2D), initially proposed by Romers et al115 allows to explain the O-

C-O anomeric effect in 2-substituted-tetrahydropyran (as well as the generalized anomeric effect in

other systems) by the delocalization of one of the lonepairs of the endocyclic oxygen into the

O

O

Cl

H1

Cl

H1

Cl

H1

Cl

O

O

OAc

AcO

H1

AcO

BzO

BzO

OBz

F

F

Cl

Cl

(A)

O

1nsp3

2nsp3

β

α

1nsp3

O

2nsp3

β

α

β

1nsp2 (p-type)

OO2nsp2 (s-type)

β

2nsp2 (s-type)

1nsp2 (s-type)

1nsp3

2nsp3

2nsp3

1nsp3

nOHyperconjugative

stabilization

ΔE (σ*C-Cl - nO)

AE α S2 /ΔE (σ*C-Cl - nO)

σ*C-Cl

σ*C-Cl

O

Cl

O

Cl−

(B)

(C1)

(D)

(C2)

(E)

A2 E2

A(C2) E(C2)

1.339 Å

1.367 Å

1.398 Å

1.406 Å

A(C1) E(C1)


16

antibonding orbital of the exocyclic C-O bond. This hyperconjugation, which corresponds to a

double-bond � no-bond resonance in terms of valence-bond theory (Fig 2D), is most efficient when

the lonepair orbital and the antibonding orbital of the exocyclic C-O bond are in the same plane, i.e.

when the lonepair adopts an antiperiplanar orientation with respect to the exocyclic C-O bond [Fig 2,

(C1) & (C2)]. This is only achieved in the axial conformer A2, which is therefore more stable than the

equatorial counterpart E2 (Fig 2A). The energy stabilization resulting from the n → σ* interaction is

inversely proportional to the energy difference [∆E(σ*C-Cl

- nO)] between both orbitals and proportional

to square of their overlap (Fig 2B). The electron-donating abilities of orbitals decrease in the order:

nC- > nN > nO > σC-S > σC-H > σC-C > σC-O > σC-F, whereas the electron accepting capability of

orbitals is reduced in the order: σ∗C-Cl > σ∗C-S > σ∗C-F > σ∗C-O > σ∗C-C > σ∗C-H 4,5.

Delocalization is expected to result in the lengthening of the exocyclic C-O bond, whereas the

concomittant shortening of the endocyclic C-O bond and opening of the O-C-O bond angle with

respect to "standard" (tetrahedral) values are observed due to the reduced and enhanced p-character of

the endocyclic and exocyclic C-O bonds, respectively.

The hyperconjugation model is therefore strongly supported by the various examples of

structural data found in the literature, such as in the case of pyranose derivatives, which show

significant changes of the lengths of the exocyclic C-O (or C-Y in the general case) and endocyclic C-

O bonds as well as of the value of O-C-O (or O-C-Y) bond angle (Fig 2E)1,121,122 in chair conformers

with axially versus equatorially oriented anomeric groups.

1.9 The nature of the electron lonepairs

The question of the adequate hybridization state (i.e. sp2 versus sp3) of the electron lonepairs

orbitals has given rise to some controversy. The use of sp3 hybridized lonepairs allows to predict

correctly conformational preferences induced by the generalized anomeric effect1. However,

experimental and theoretical studies123-126 have led to the conclusion that the nonbonding oxygen

lonepairs are nonequivalent: One of them occupies a higher energy orbital with high p-type character,

whereas the other occupies an orbital of lower energy with predominant s-type character. In recent

studies on the anomeric effect117,127,128, the sp2 hybridization model of the electron lonepairs orbitals

has therefore gained widespread recognition.

Basing on the results of their statistical analysis of X-ray crystal structures of COCOC

molecular fragments, Cossé-Barbi and Dubois have shown129 that the five-membered furanose ring

orients polar anomeric groups in axial or pseudoaxial position (95%) much more efficiently than the

six-membered pyranose ring (56%). They argued that this stronger anomeric effect in furanose

compared with pyranose is the result of the better ability of one of the lonepairs of the endocyclic

oxygen to participate in stabilizing hyperconjugation, which can be easily understood with help of the

sp2 (not sp3) hybridization model of lonepairs. Indeed, sp2 hybridization makes the nO → σ

*

C-X

interaction more efficient in furanose than in pyranose130, since it results in a perfect antiperiplanar

orientation of the p-type lonepair orbital of the endocyclic oxygen atom with respect to the exocyclic

C-X bond in the former but not in the latter [compare panel (C2) in Fig 2 with panel (C) in Fig 9]. If

the lonepairs of the endocyclic oxygen were sp3 hybridized, the situation would be just reverse, i.e.



17

the overlap between nO and σ

*

C-X orbital would be optimal in the pyranose system but not for the

furanose counterpart [compare panel (C1) in Fig 2 with panel (B) in Fig 9].

1.10 Dipole-dipole (electrostatic) interactions as origin of the anomeric effect

Already in 1955, Edward65 proposed that the destabilization of the E2 chair of 2-

methoxytetrahydropyran (2) (with equatorially oriented 2-OCH3 group) with respect to the axial

counterpart A2 is the result of electrostatic repulsions between 2-OCH3 and the ring dipole. The ring

dipole is itself the resultant of individual dipole moments from endocyclic C-O bonds and the

lonepairs of the endocyclic oxygen. More recently, the electrostatic model has been refined and the

anomeric effect has been attributed to electrostatic repulsions117,118 between the exocyclic C-X dipole

and the ring dipole in the E2 chair (Fig 3A). The repulsion between two dipoles is function of their

magnitudes as well as the distance and the angle between them74,131. In the equatorial conformer E2,

both dipoles are nearly parallel and repel each other much more than in the axial form A2 where the

angle between them is nearly 90°. As a result, the E2 conformer is destabilized by electrostatic

repulsions compared to the A2 counterpart and the anomeric group 2-OCH3 preferentially adopts an

axial orientation.

The experimentally found (Section 1.6) dependence of the magnitude of the anomeric effect

upon the polarity of the solvent has been considered as the best evidence for the electrostatic origin of

the anomeric effect: Polar solvents stabilize the more polar conformation (i.e. the equatorial chair E2)

more efficiently than solvents with lower dielectric constants, therefore the strength of the anomeric

effect will be reduced in polar compared to apolar solvents. In a recent calorimetric study of the heats

of hydrolysis of the anomeric forms of 4,6-dimethyl-2-methoxytetrahydropyran, Wiberg and

Marquez118 have also shown that the nature of the solvent strongly influences ∆H° and ∆G° of the

isomerization equilibrium, which led them to conclude that the anomeric effect is mainly induced by

the electrostatic repulsion between the C-O dipoles, the magnitude of which decreases when the

dielectric constant of the solvent is increased.

However, there are also known examples of an increase in the strength of the anomeric effect

induced by an increased polarity of the solvent, such as for trans-2,5-bis(trimethylsiloxy)-1,4-

dioxane132. This particular result cannot be explained by the electrostatic model. Instead, it has been

rationalized in terms of hyperconjugative interactions (Section 1.8). More importantly, in contrast

with the hyperconjugation model, the electrostatic model does not allow to explain the changes of

bond lengths and bond angles that are commonly associated with the existence of the anomeric effect

in heterocyclic 6-membered rings (Section 1.8).

The question of the relative importance of hyperconjugative interactions versus dipole-dipole

repulsions as the origin of anomeric effect in non-polar solvents has been raised in a recent

comparative conformational study of 2-methoxy-1,3-dimethylhexahydropyrimidine (3) and 2-

methoxy-1,3-dioxane (4)117. It was argued that since the dipole moment of CH3-NH2 is smaller than

that of CH3-OH, substitution of nitrogen in 3 for oxygen in 4 should lead to increased electrostatic

repulsions in 4 compared to 3. Conversely, as evident from the lower ionization potential of CH3NH2

in comparison with that of CH3OH, replacing oxygen in 4 by nitrogen in 3 should raise the energy of

the filled nN lonepair orbital in the later with respect to that of nO orbital in the former and therefore


18

strengthen the n→σ* hyperconjugative interactions in 3 compared to 4. As a result, if the anomeric

effect operates mainly through dipole-dipole (electrostatic) interactions, then replacing oxygen by

nitrogen should reduce the strength of the anomeric effect in 3 compared to 4. Conversely, if

hyperconjugative interactions predominate, then the strength of the anomeric effect should be

increased in 3 compared to 4. Perrin's conformational analysis based on temperature-dependent 1H

and 13C NMR spectra showed that the proportion of axial conformer is nearly the same in 3 and 4.

However, after correction of the results for the steric effects induced by the presence of methyl groups

in 3 (supported by molecular mechanics and AM1 calculations), it turned out that the anomeric effect

is weaker in 3, and it was therefore concluded that the anomeric effect is mainly the result of

electrostatic interactions. Additionally, it was also suggested that the bond length changes that are

often invoked as a key evidence for the existence of hyperconjugative interactions (Section 1.8) can

also be explained in terms of dipole-dipole interactions. A subsequent reinvestigation of the origin of

the anomeric effect by Salzner128 based on ab initio geometry optimizations (at HF/6-31G* level) of

various conformers of hexahydropyrimidine (5), 2-hydroxypiperidine (6) and 2-

hydroxyhexahydropyrimidine (7) and on the NBO analysis of the resulting wave functions showed

that indeed the magnitude of the anomeric effect in 6 and 7 is much reduced in comparison with the

oxygen analogs. However, it was also shown in that work that the total energies of the conformers do

not correlate with total dipole moments, which discredited the predominance of electrostatic

repulsions. Instead, the weaker anomeric effects in 6 and 7 in comparison with the oxygen

counterparts were interpreted as the results of the competition between OH/NH bond repulsions and

hyperconjugative stabilizations.

1.11 Alternative explanations for the origin of the anomeric effect

According to Eliel's original rabbit-ear effect model119

, the anomeric effect in R-X-A-Y

fragment arises from repulsions between electron lonepairs of X and Y heteroatoms (Fig 3B). This

model has been used in particular to explain the drive of the conformation of 2-

alkoxytetrahydropyran76,95 and tripepiperideine133 by the O-C-O and N-C-N anomeric effects,

respectively. Alternatively, it has also been suggested by Box120 that four electrons orbital mixing

destabilizing interactions allow to account for X-C-Y anomeric effect controlling the conformation of

glycosides: according to this model, the occupied (2+2) electron lonepair orbitals of X and Y

heteroatoms interact to form two new orbitals, occupied by the four electrons (Fig 3C). However, the

energy destabilization of the system induced by the newly formed high-energy orbital is greater than

the stabilization gained by the formation of the new low-energy orbital, therefore such lonepair-

lonepair interactions, on the overall, are destabilizing, and the system attempts to adopt a

conformation in which there are fewer lonepair orbital interactions. As a result, the anomeric

substituent adopts preferentially an axial orientation.



19

1.12 The gauche effect

The "gauche effect"1,134-143 designates the tendency of R-X-Y-R' and X-C-C-Y fragments

(where X and Y represent electronegative atoms or groups) to adopt preferentially gauche over

antiperiplanar orientations in spite of unfavourable steric repulsions and electrostatic interactions (i.e.

the relative energy difference between trans and gauche conformers, ∆Et-g, is negative, as depicted in

Fig 4). For 1,2-disubstituted ethane in the gas phase, there is a linear relationship between ∆Et-g and

the sum of Huggins electronegativities of both substituents135. 1,2-dimethoxyethane exists

preferentially in a gauche conformation in solution, but the stabilization of the gauche conformer is

much reduced in the gas phase144. On the other hand, ab initio calculations predict almost identical

energies for the gauche and trans rotamers142. A study of the energy relationship of the isomers of

1,2-difluorethylene, 1,2-difluorodiazene and several other halogen- and oxygen-substrituted

ethylenes145 show that the cis isomer has the lower energy, which is generally referred to as the cis

effect.

Figure 3. The anomeric effect in 2-substituted

tetrahydropyran as the result of electrostatic interactions.

Lonepairs on the endocyclic and exocyclic oxygen atoms are

represented assuming sp3 hybridization. (A) Destabilization

of the equatorial chair E2 with respect to the axial counterpart

A2 by dipole-dipole repulsions65,117,118

. (B) Prediction of the

relative stabilities of B1 - B6 conformers of 2-

methoxytetrahydropyran (2) using Eliel's rabbit-ear model95

of interaction between the electron lonepairs of the

endocyclic and exocyclic oxygens. ("black" lonepair: in the

front, "empty": at the back, "grey": in the plane of the

drawing). The equatorial conformers B1 - B3 as well as the

axial forms B4 - B5 are destabilized owing to one or two (for

the B2 form only) repulsion(s) between lonepairs of the

endocyclic and exocyclic oxygens. The axial conformer B6,

in which the anomeric group OMe adopts an axial

orientation, is the most stable, since only in this situation

there is no destabilizing lonepair-lonepair interaction. (C)

The four electrons interaction model120

. The interaction

between two filled lonepairs orbitals of the endocyclic [nO

(endo)] and exocyclic [nO (exo)] oxygens produces two new orbitals. This interaction is destabilizing, since the stabilization

(∆E1) of the system gained from the formation of [nO(endo) + nO(exo)] orbital is weaker than the corresponding

destabilization (∆E2) owing to the newly formed high-energy [nO(endo) - nO(exo)] orbital .

“Attractive” and “repulsive” gauche effects define a greater preference of a X-C-C-Y fragment for

gauche and anti orientations, respectively, than that expected on the basis of steric effects and

electrostatic interactions alone. It has been shown138-140 that the extent of stabilization of the 1,2-

diaxial versus 1,2-diequatorial conformers of 1,2-trans-disubstituted cyclohexane is dictated by the

fine balance between steric, electrostatic and gauche interactions between vicinal X and Y

substituents: For strongly electronegative X and Y substituents (i.e. O/O, F/Cl, F/Br, F/I or F/O pairs),

a strong attractive gauche effect operates, whereas when the electronegativity of X and Y is

intermediate (such as O/Cl, O/I or Cl/I pairs), the observed conformational preferences can be

explained by steric and electrostatic repulsions alone. In contrast, for S/S, S/Cl, S/Br or S/O

O

O

XX

A2 chair E2 chair

O OO

O O O

O O O

O O O

Me

Me

Me

MeMe

Me

(A)

(B) B1 B2 B3

B4 B5 B6

(C)

ΔE2

nO (endo) - nO (exo)

ΔE1

nO (endo) + nO (exo)

ΔE2 > ΔE1

nO (endo)nO (exo)


20

disubstitution, the population of gauche rotamers is less than expected (owing to the repulsive gauche

effect). In 5-substituted-1,3-dioxan, conformers with axially oriented C5 substituent are preferred as

the result of the attractive gauche effect146, whereas in 5-methoxy- and 5-methylthio-1,3-dithianes

repulsive S/O and S/S gauche effects are operating147.

1.13 The energetics of the gauche effect

As stated above, the gauche effect implies the electronic energy preference of a gauche

rotamer over the corresponding anti rotamer1. The experimental energetic relationship of various

rotamers for 1,2-difluoroethane (as derived from Raman148,149, infrared149, NMR spectroscopy150 and

electron diffraction151-153 studies) or 1,2-dimethoxyethane (by NMR spectroscopy144) in the gas

phase has been well worked out. These experimental works have shown that the gauche rotamer has a

lower energy than the anti counterpart by 2.4 - 3.4 kJmol-1 (in the case of difluoro substitution),

which compares well with results obtained from theoretical studies141,154 (i.e. ab initio calculations at

the MP2/6-311++G** and TZ+2D+P levels).

Figure 4. Plot of the relative energy of 1,2-

disubstituted ethane conformers as a function of the

ΦXCCY torsion angle, showing the stabilization [∆E(t-

g) < 0] of gauche+ and gauche-

rotamers over the trans

counterpart, as the result of the attractive X-C-C-Y

gauche effect.

1.14 Possible origins of the gauche effect

The origin of the gauche effect is still a matter of controversy: (i) It has been shown that the

preferred gauche orientation in ethylene glycol142,155 results from hydrogen bonding. (ii) The

repulsive gauche effects for some 1,2-disubstituted cyclohexane derivatives (e.g. for S/S

disubstitution pattern) have been attributed to the through-space interaction (or "hockey sticks effect")

between lonepair orbitals on the X and Y heteroatoms in the X-C-C-Y fragment138,156,157. Such an

overlap between two doubly-occupied lonepairs is destabilizing: It results in the formation of a

bonding orbital and its antibonding counterpart, however the destabilizing influence of the latter is

greater than the stabilizing effect of the former (Fig 5, Panel A1). On the other hand, Hoffmann158

suggested that the lonepair interaction does not only take place through-space but it has also a

significant through bond component (Fig 5, Panels A2 and A3). The actual repulsive or attractive159

character of the gauche effect is the result of the fine balance between through-space and through-

bond contributions. Hoffmann's analysis is supported by the results of a photoelectron spectroscopic

study of 3,7,9-trihetero derivatives of bicyclo[3.3.1]nonane, showing the preponderant influence of

the through-bond component160.

(iii) σ →σ∗ orbital interactions between the best donor σ orbital and the best acceptor σ*

orbital have also been suggested136,161. The efficiency of these interactions is proportional to the

0 60 120 180 240 300 360

0

1

2

3

4

Φ X-C-C-Y

( ° )

ΔE

(t-g)

Y

X

Y X

Y

X

Y

X

X Y

Y

X

X

Y

gauche −

gauche +

eclipsed

cis

trans

eclipsed

cis

Relative Energy



21

square of the overlap between the donor and acceptor orbitals and inversely proportional to the

difference between their energies. An example of σC-H →σ∗C-F orbital overlap in the gauche

conformer of 1,2-difluoroethane is shown in Fig 5B-C.

Figure 5. (A1)-(A3) Through-space (A1) and through-bond (A2 and A3) interactions138,156-160

between lonepair orbitals

(Φ1 and Φ2) on the X and Y heteroatoms in the X-C-C-Y

fragment, as the origin of attractive or repulsive gauche

effects. Through-space (A1) four electron interactions

result in the formation of a bonding orbital (Φ1 + Φ2),

which is stabilizing (∆E1 < 0) whereas the antibonding

counterpart (Φ1 - Φ2) is destabilizing (∆E2 > 0). ∆E2 >

∆E1 and therefore the system is destabilized on the

overall. (Φ1 + Φ2) can in turn interact with σC-C [through-

bond component of the origin of the gauche effect, (A3)].

This four-electron interaction is again destabilizing since

∆E4 > ∆E3. On the other hand, the concomittant two-

electron interaction between (Φ1 - Φ2) and σ*C-C is

stabilizing, as shown in Panel (A2). (B) and (C): σC-H

→σ*C-F orbital overlap to explain the gauche effect in

1,2-difluoroethane136,161

. The σ →σ* overlap is most

efficient when it involves the best donor and best acceptor

σ and σ* orbitals, respectively. In the case of the trans

conformer, the σ orbital is σC-F, which is much poorer

donor than σC-H in the corresponding gauche conformer

[Section 1.8]. Therefore, the σC-H →σ*C-F interaction

stabilizes the gauche conformer with respect to the trans

counterpart, and by bond-no bond resonance the olefin (3)

is envisioned as a canonical form. (D) The gauche effect

is shown in relation with bent bonds143,162

. In the trans

form, the C-C bond paths are bent in the opposite

direction (as shown by the dashed lines), which leads to a

reduced overlap and a weaker C-C bond. On the contrary,

in the gauche rotamer, bond paths are bent essentially in

the same direction. Therefore the trans rotamer is

destabilized with respect to the gauche counterpart.

Wolfe originally defined the gauche effect134 as "a tendency to adopt that structure which has

the maximum number of gauche interactions between the adjacent electron pairs and/or polar bonds",

which he rationalized in terms of a nuclear-electron attraction prevailing over nuclear-nuclear and

electron-electron repulsions. This definition enables to predict correctly the preferred gauche

orientation of two vicinal electron lonepairs R and R' in a R-X-Y-R' system or of X-C and C-Y polar

bonds in X-C-C-Y fragment, but not the antiperiplanar orientation136 of a lonepair with respect to a

vicinal polar bond, i.e. owing to the anomeric effect (Sections 1.2 - 1.6). It is therefore preferable to

define1 gauche effect as "a stereoelectronic preference for conformations in which the best donor

lonepair or bond is antiperiplanar to the best acceptor bond". (iii) The preferential gauche arrangement

in 1,2-difluoroethane has also been explained by the fact that in this rotamer, the C-C bond paths (i.e.

the regions with maximum electron density connecting both carbon atoms163) are bent in the same

direction, whereas in the relatively less stable trans form bending occurs in opposite directions, which

results into a reduced overlap and a weaker C-C bond143 (Fig 5D). This theory is supported by the

C C

F

F

C C

FF

(D)

H

H

F

H

H

F

H

H

F

H

F

H

F

H

H

H

F-

trans (1) gauche (2) (3)

(B)

trans (1) gauche (2)

H

F

H H

H

F

H

H

H F

H

F

180ο

60ο

(C)

YX

YX

X

Y

YX YX

YX YX

ΔE1

ΔE2

Φ1

Φ2

Φ1 +Φ2

Φ1 - Φ2

Φ1 +Φ2

Φ1 - Φ2

σ

σ*

(A1)

(A2)

(A3)

ΔE3

ΔE4


22

comparison of geometrical parameters for the gauche and anti rotamers of 1,2-difluoroethane, as

obtained from an experimental study based upon high-resolution infrared spectroscopy162 or from ab

initio calculations (using hybrid-DFT164 or RHF/6-311++G** level149) .

2. Stereoelectronic effects in nucleosides and nucleotides

2.1 Structure of nucleic acids

Endowed with unique capabilities, such as the storage of the genetic information, induction of

cellular differentiation, propagation of tumor viruses as well as splicing and autocatalysis, deoxyribo

(DNA) and ribonucleic acids (RNA) are polymers built up of monomers, the nucleosides, which are

covalently linked through 3'→5' phophodiester linkages165. Nucleosides consist of a pentofuranose

sugar and a heterocyclic nucleobase (adenin-9-yl, guanin-9-yl, cytosin-1-yl, thymin-1-yl, uracil-1-yl or

modified C-166 or N-heterocycle165) connected by the covalent C1'-N1/9 or C1'-C (glycosyl) bond. In

natural DNA and RNA, the D-sugar enantiomer and β-glycosyl linkage are preferred over the L- and

α-counterparts. The conformation along the phosphate backbone in nucleotides can be fully

defined167,168 by the torsion angles α, β, γ, δ, ε and ζ (Section 7.1).

The stability of the three-dimensional structure of nucleic acids is usually attributed to only a

few types of forces: (i) Strong intermolecular H-bonds between complementary nucleobases; (ii)

Intramolecular base-base stacking interactions; (iii) Electrostatic interactions and hydrophobic forces

between adjacent base pairs and across the nucleotide chain165, and hydration. The purpose of the

next sections is to present recent progress in our understanding of the interplay of stereoelectronic

gauche and anomeric effects in the drive of the conformation of simple nucleosides and nucleotides

and their analogs as well as the self-organization of oligonucleotides.

2.2 The pseudorotation concept

Saturated heterocyclic six-membered rings are much less flexible than their five-membered

counterparts169. The activation energy barrier for the conversion of one chair form of cyclohexane

into the other is170,171 about 42 kJmol-1. 1H- and 13C NMR spectra of 2-methoxy-1,3-

dimethylhexahydropyrimidine (3)117 measured in various solvents have shown that the ∆G* value for

the ring inversion between two chair forms is ≈ 37 ± 2 kJmol-1 (the two interconverting chair

conformers could be isolated at low temperature). On the other hand, the activation energy barrier for

the interconversion between the two preferential puckering modes, i.e. N- and S-type pseudorotamers,

of the pentofuranose moiety in purine nucleosides is much reduced in comparison, i.e. below 20 - 25

kJmol-1 (vide infra).

The pseudorotation concept has been introduced by Kilpatick et al. 172 to describe the

continuous interconversions between an infinite number of indefinite puckered forms of the

cyclopentane ring. Pseudorotation173 allows cyclopentane to relieve strains which would be induced

by 120° bond angles and eclipsed methylene groups if it would adopt a planar geometry.

A barrier to planarity of cyclopentane of 22 kJmol-1 has been reported176. The concept of

pseudorotation has been applied for the first time to furanose by Hall et al. 177 in their conformational

analyses of pentofuranosyl fluorides.



23

Kilpatrick originally described172 the puckered states of cyclopentane by sequential out-of-

plane displacements (of the carbon atoms) from the least square plane of the unpuckered ring. The

displacement zi of the ith ring carbon in a particular puckered conformer was calculated using Eq 5:

zi = (2/5)1/2 q cos 2 (Φ + 2πi/5) ..... Eq 5

where, q and Φ represent the maximum puckering amplitude and the phase angle of pseudorotation of

the pseudorotamer, respectively. Eq 5 holds exactly in the case of the infinitesimal diplacement in an

equilateral pentagon. Cremer and Pople have later on devised a generalized set of puckering

coordinates valid for equilateral and non-equilateral five-membered rings178. This formalism is

however quite cumbersome since it requires the

Figure 6. The pseudorotation wheel (E =

envelope; T = twist) for pentofuranosyl D-

nucleosides. The hyperspace of geometries

accessible to N- and S-type pseudorotamers is

within the shaded areas for β-D-dNs, β-D-rNs

and β-D-arabinonucleosides (-1° < PN < 34°,

137° < PS < 194°, 30° < Ψm(N), Ψm(S) <

46°)174 and within the empty oval shaped

areas for α-D-dNs, α-D-rNs, α-D-xylo- and α-

D-arabinofuranosyl nucleosides175 (-18° < PN

< 19°, 168° < PS < 224°, 28° < Ψm(N), Ψm(S)

< 49°).

knowledge of 15 cartesian

coordinates. Altona et al. have solved

this problem by proposing an alternative description179-181 of a puckered geometry of the

pentofuranose ring in nucleic acid derivatives by two parameters P and Ψm, related in turn to the five

endocyclic torsion angles νi (i = 0...4) (Eq 6a): νi = Ψm cos (P+4π(i-2)/5) ..... Eq 6a

P, the phase angle of pseudorotation (0° < P < 360°), shows which part of the ring is mostly

puckered, whereas Ψm, the maximum puckering amplitude, provides information about the largest

deviation of the endocyclic torsions from zero. The ensemble of puckered forms that may be adopted

by the pentofuranose moiety in nucleos(t)ides is represented in the form of the pseudorotation cycle

(Fig 6). In nucleosides, the endocyclic torsions νi (i = 0...4) are defined as follows: ν0 [C4'-O4'-C1'-

C2'], ν1 [O4'-C1'-C2'-C3'], ν2 [C1'-C2'-C3'-C4'], ν3 [C2'-C3'-C4'-O4'] and ν4 [C3'-C4'-O4'-C1']. P

and Ψm are derived from Eqs 6b and 6c:

tan P = [(ν4 + ν1) - (ν3 + ν0)] / [2ν2 (sin 36° + sin 72°)] .....Eq 6b

Ψm = ν2 / cos P .....Eq 6c

R

O

R'

O

R

N

H

H R'

HOH2C

OHOH2C

R

N

H

H

R'

N

HOH2CH

H

H

H

H

H

H

H

H

O

H

H

HR'

R

CH2OHN 0E

144º

3E

108º 0

1T

0º

T

306º

234º

198º

T

4

3T

0

4T

3

4T

3

2T

1

2T

T1

0

0E

3E

1E

4E

2E

1E

4E

2E

324º

288º

252º

216º

180º

72º

36º40

ο

90º

40ο

342º

162º

54º

18º

126º

20ο

0ο

2

3

270º

2

1

4

020

ο

T

West (W)

South (S)

East (E)

North (N)


24

For practical purposes, Eq 6a may be regarded as nearly exact for equilateral five-membered

rings, since it reproduces the values of the endocyclic torsion angles of a cyclopentane pseudorotamer

with an accuracy better than 0.1°182. In a study of 178 β-D-furanosides, the endocyclic torsion angles

could be calculated using Eq 6a with an r.m.s. error of 0.4 - 0.9° 183,184. For non-equilateral rings, a

more accurate expression (r.m.s. = 0.2 - 0.4°) for the dependence of the endocyclic torsions on P and

Ψm is given by Eq 7183-186:

νi = ai * Ψm cos (P+εi+4π(i-2)/5) ..... Eq 7

where, ai and εi represent correction factors that are included to overcome differences in endocyclic

bond lengths.

Altona's model has not only been used to describe the conformation of the pentofuranose

moiety in nucleosides and nucleotides187-226 but also the puckered states of cyclopentane179, the

pyrrolidine ring in L-proline and in its derivatives182,227, the ring D in steroids180, and other five-

membered rings115,228-231.

Ab initio calculations232 have shown that pseudorotation of cyclopentane occurs without any

substantial change in potential energy whereas in unsymmetrically substituted saturated five-

membered rings potential energy thresholds induced by the presence of the exocyclic substituents

limit the flexibility of the system and lead to prefered puckering modes233.

O

OR'B

O

OR'

B

RO

OR

O

B

HO

HO

O

OR'

B

O

OR'

BRO

OR

North (N) β-D-sugar (C3'-endo-C2'-exo)

South (S) β-L-sugar (C2'-endo-C3'-exo)

(iv)

O

B

O

B

HO

OH

(iii)

HO

North (N) β-L-sugar(C2'-exo-C3'-endo)

South (S) β-D-sugar (C2'-endo-C3'-exo)

HO

North (N) α-L-sugar(C2'-exo-C3'-endo)

South (S) α-L-sugar (C2'-endo-C3'-exo)

South (S) α-D-sugar (C2'-endo-C3'-exo)

North (N) α-D-sugar (C3'-endo-C2'-exo)

(i) (ii)

B = adenin-9-yl, guanin-9-yl, cytosin-1-yl, thymin-1-yl or uracil-1-yl (in RNA)

D-series L-series

O

B

OH

HO

Scheme 1: The D- and L- mirror image relationship for the two-state dynamic N � S sugar equilibrium in α-D-

dNs (i) , α-L-dNs (ii), β-D-dNs (iii) and β-L-dNs (iv). In L-nucleosides234, the N sugar is redefined as being the

form with maximal negative value for the endocyclic torsion [C1'-C2'-C3'-C4']. Note that in α-D-dNs (i) and α-L-

dNs (ii), the aglycone B becomes more pseudoaxial as the anomeric effect becomes stronger in the S-type

conformation, whereas in β-D-dNs (iii) and β-L-dNs (iv), this is achieved in the N-type conformation. Hence, the

sign for the energetic drive of the anomeric effect in α-nucleosides is opposite to that of β-counterparts (Sections 4

and 5).



25

2.3 The two-state N � S equilibrium in β-D-nucleosides

In D-nucleosides, in the typical N (P = 0°) and S (P = 180°) conformations, the [C1'-C2'-C3'-

C4'] torsion angle has a maximal positive and negative value, respectively, whereas in the L-

enantiomers the signs, as defined by Hoffmann and Altona234, are opposite as a result of the mirror-

image relationship of D- and L-enantiomers (Scheme 1).

A perusal174 of 178 crystal structures of β-D-ribo-, 2'-deoxyribo- and arabinonucleosides

shows that the phase angle values of their puckered pentofuranose moieties are not evenly distributed

over the whole 0 - 360° range but are instead clustered in the North (N, -1° < P < 34°, centered around

C3'-endo) and South (S, 137° < P < 194°, centered around C2'-endo) regions, while Ψm is within the

30° - 46° range with an average value of 38.6 ± 3° (Fig 6). The N and S regions are nearly equally

populated for ribonucleosides whereas for 2'-deoxyribonucleosides there is a clear preference for S-

type geometries (≈ 3:1 for the ratio of S versus N pseudorotamer populations). Only a few E-type

pseudorotamers and no W-type conformers were found among these crystal structures, which is

consistent with an analysis based upon simple steric effects: W-type conformations are energetically

highly disfavoured owing to the pseudoaxial orientation of both the nucleobase and the 5'CH2OH

group and to the fact that the C2' and C3' substituents are eclipsed. Analogously, the energy

destabilization of E-type conformations can be attributed to the eclipsed orientation of the C2' and C3'

substituents. 1H-NMR studies in aqueous solution193-204,206-208,212,216,219,220,222,224-226,235-240, have

shown that the puckered geometries of the furanose moiety in nucleos(t)ides interconvert rapidly (in

the NMR timescale) since only time-average chemical shifts and coupling constants can be extracted

from the spectra. This means that the activation energy barrier between the interconverting

pseudorotamers is significantly reduced in comparison with heterocyclic six-membered rings, which

often adopt a single chair conformation (vide infra).

However, two distinctly identifiable and dynamically interconverting N and S conformations have

been observed in some B � Z DNA241,242, A � Z RNA243,244 or A-form � B-form lariat

RNA245,246 transitions.

The NMR results studies as well as de Leeuw et al's statistical analysis of the distribution of P

values of the crystal structures of nucleos(t)ides (vide supra) suggest that the conformation of the

pentofuranose moiety of nucleos(t)ides in aqueous solution is adequately described by a two-state N

� S equilibrium model, since no third state is found yet. This is also supported by the following

observations, based upon the results of our own conformational studies on nucleosides and

nucleotides14-44: (i) PN, PS, Ψm(N) and Ψm(S), as obtained from pseudorotational analyses of

experimental vicinal proton-proton coupling constants (3JHH) for β-D-2',3'-dideoxynucleosides19 are

virtually the same in the analyses which have been performed using the 3JHH data at any single

temperature or in other analyses based upon the whole set of temperature-dependent 3JHH coupling

constants (i.e. at seven temperatures within the 283 K - 353 K range). (ii) Van't Hoff plots of the


26

ratios of mole fractions of S- and N-type sugars (xS/xN) as a function of the inverse of temperature

give straight lines for all nucleosides with correlation coefficients ≥ 0.95. (iii) We have recently

shown that ∆H˚, ∆S° and ∆G° of the N � S equilibrium in β-30,32,36, and α-nucleosides36,37 are pD-

dependent (Sections 4 - 6). The shift of the N � S equilibrium in β- (or α-) nucleosides toward more

N- (or S-type) sugars at acidic pDs upon protonation of the nucleobase, and toward S-type (or N-type)

conformers at alkaline pDs upon deprotonation, results from the transmission of pD-tunable electronic

character of the nucleobase to steer the sugar conformation through modulable anomeric and gauche

effects. The pDs at the inflection points of the plots of pD-dependent ∆G° values gave the pKas of the

nucleobases, and these values were virtually the same as those published247 or independently

determined through monitoring of pD-dependent chemical shifts of non-exchangeable aromatic and

anomeric protons30,32,36,37. Raman spectroscopy study on A and T containing DNAs248 also support

that the two-state N � S equilibrium model also holds for the pentofuranose sugar moiety.

Throughout this whole book and in all papers cited from this laboratory, the signs of the

thermodynamic quantities have arbitrarily chosen in such a way that the positive values indicate the

drive of the N � S equilibrium to N, whereas the negative values describe the drive to S.

Interconversions between N- and S-type puckered furanose conformers most likely occur

along the pseudorotational cycle (i.e. via puckered geometries) rather than through a planar

intermediate, which is disfavoured as the result of high strain energies181,249,250. The fact that only a

few E-type puckered geometries and no W-type pseudorotamer were identified among the crystal

strutures of 178 β-D nucleos(t)ides suggests that N- to S- interconversion proceeds via an E-type

puckered geometry intermediate, not through the W region.

The intrinsic flexibility of the pentofuranose moiety is absent in cyclic nucleosides such as

2',3'-O-, 8,2'-O-, 8,3'-O-anhydronucleosides or in cyclic nucleotides165. This results in a reduced

puckered amplitude, for instance in 2',3'-cyclic nucleotides251-253 or to unusual puckering modes of

3',5'-cyclic nucleotides, both in solution254 and in the solid state255-257.

The effects of packing forces and hydrogen bonds upon the observed puckering modes of

nucleosides in the solid state are well known165. Solution and solid state conformations of

pentofuranose differ dramatically for some nucleosides: Adenosine258 crystallizes as the N-type

conformer, whereas its hydrochloride salt crystallizes as the S-type form259 and our NMR studies30

show a clear preference (≈ 67% at room temperature) for S-type conformers at neutral pD, and the

pseudorotational equilibrium is shifted to more N-type sugars at acidic pD (≈ 55% South conformer at

298 K). 3'-azidothymidine17 in D2O is involved in an equipopulated two-state N � S equilibrium,

despite the fact that it crystallizes in the S-type puckered conformers: P = 175° with Ψm = 32° and P =

215° with Ψm = 36°260-265. 3'-fluorothymidine crystallizes266,267 in the form of S-type conformers,

however the average phase angle value for these conformers (P ≈ 171°, Ψm ≈ 34°) deviates from that

found for the preferred S-type sugar geometry (≥ 90%, P ≈ 160°, Ψm ≈ 34°) by 1H NMR studies17,26.

For all β-D-2',3'-dideoxynucleosides (ddNs), the pentofuranose sugar is mostly puckered as the N-



27

type conformer in D2O solution19 (≥ 75% at 298K), whereas in the solid state, it adopts S-type

geometries (P = 194°, Ψm = 37° for ddA268, P = 208° and Ψm = 34° for ddC269).

An integrated conformational study on 4'-thiothymidine and (E)-5-(2-bromovinyl)-2'-deoxy-4'-

thiouridine in solution (based on the interpretation of 3JHH with help of the pseudorotation concept

and on the analysis of nOe enhancements) and in the solid state (from X-ray crystallography) has been

recently reported by our lab270. It was found that the crystal structures of both thymidine derivatives

are grossly similar with a 4'-thiofuranose moiety in S-type conformation (P = 180° and Ψm = 48° for

both compounds). Owing to the nonequilateral nature of 4'-thionucleosides, the pseudorotation

equation (Eq 7) was reparametrized basing on ab initio calculations. The solution structure of the

major conformer of 4'-thiothymidine and its derivative were similar to the crystal structure, showing

that the presence of S4' vis-a-vis O4' does not result in any additional flexibility of the thiofuranose

ring. The similarity of the bias of the N � S equilibrium in thionucleosides and thymidine show that

the magnitude of interplay of gauche and anomeric effects is comparable in both, thereby showing the

insignificant effect of S vis-a-vis O. It is also noteworthy that the bond lengths of C1'-S and C4'-S are

essentially the same as found in many other natural C-nucleosides271-282 and N-nucleosides174,283,

suggesting that the identification of the anomeric effect on the basis of unequal bond lengths of C4'-

X4' versus C1'-X4' is an oversimplification5.

2.4 The two-state N � S equilibrium in α-nucleosides

A recent perusal175 of the few crystal structures available in the literature for α-nucleosides

has shown that their constituent pentofuranose moieties also tend to adopt predominantly

conformations belonging to the C2'-endo and C3'-endo domains (-18° < PN < 19°, 168° < PS < 224°,

28° < Ψm(N), Ψm(S) < 49°) with a few exceptions found in the C4'-endo region. Therefore, the

hypothesis of the two-state N � S equilibrium has also been retained during our conformational

studies on α-nucleosides (Section 5). In that work175, basing on the results of potential energy

calculations on both α-purine and α-pyrimidine nucleosides, it was suggested that interconversion

between N- and S-type α-sugars proceeds through the O4'-exo rather than through the higher O4'-endo

activation energy barrier, in contrast with what has been claimed for β-counterparts (vide infra).

2.5 The two-state N � S equilibrium in carbocyclic nucleosides

The chemical and enzymatic vulnerability of the glycosyl bond of natural nucleosides has

inspired the synthesis of a wide gamut of carbocyclic nucleosides284-291 in which a cyclopentane ring

replaces the ribose moiety. However, these more stable carbocyclic nucleosides are generally less

potent284 than their natural counterparts. This is attributed to the more flexible nature of the

cyclopentane ring where both anomeric and gauche effects inherent to the natural nucleosides17,19-

30,32,33,35-38,40-44 are lacking. Efforts to reduce the flexibility of the cyclopentane ring have resulted in

the design of conformationally constrained carbocyclic nucleosides292,293 (see Sections 8.2 and 8.3).

The X-ray crystal structures of four conformationally unconstrained carbocyclic

nucleosides294-297 are so far known, i.e. for (-)-aristeromycin294 (P = 89.0° and Ψm = 40.8), (+)-

carbathymidine295 (P = 118.6° and Ψm = 41.2°), 1-(c-2-fluoro-t-4-hydroxy-c-3-hydroxymethyl-r-1-

cyclopentyl)uracil296 (P = 95.7° and Ψm = 43.1°) and the carbocyclic analogue of 2'-


28

deoxyguanosine297 (P = 293.3° and Ψm = 35.8°). The crystal structures59 of 6'-α-methylcarbocyclic

thymidine in four different forms (P = 147.4° and Ψm = 47.6°; P = 141.9° and Ψm = 49.1°; P = 153.6°

and Ψm = 44.5°; P = 111.2° and Ψm = 37.6°) and of 6'-α-hydroxycarbocyclic thymidine (P = 127.9°

and Ψm = 44.2°; P = 141.1° and Ψm = 45.6°; P = 140.0° and Ψm = 48.2°; P = 123.4° and Ψm = 49.0°)

as constituents in modified oligo-DNA duplexes have also been determined using X-ray

crystallography.

These data suggest that the cyclopentyl rings in carbocyclic nucleosides are prone to adopt a

greater variety of conformations than the pentofuranosyl counterpart in the corresponding natural N-

nucleosides. The greater flexibility of carbocyclic nucleosides compared with natural pentofuranosyl

nucleosides may be attributed to the absence of O4' oxygen in the former, whereas in the latter O4' is

involved in stereoelectronic interactions with the glycosyl nitrogen (i.e. the O4'-C1'-N1/9 anomeric

effect) and with O2' and/or O3' (i.e. through gauche effects). However, any possible contribution of

different solvation and electrostatic potential in the cyclopentyl compared with pentofuranosyl

analogues cannot be excluded. The conformational variations of cyclopentane moieties in carbocyclic

nucleosides in the solid state clearly suggested that a knowledge of their dependable solution

conformations is highly desirable in order to compare the solution and the solid state structures.

Hence, a modified Karplus equation correlating the 3JHH values to the corresponding proton-proton

torsion angle specifically in carbocylic nucleosides is now available (Eq 8b in Section 3.4.2).

∆G° of the protonation � deprotonation equilibrium drives the two-state N � S equilibrium

of the furanose moiety in N-30,36,37 and C-nucleosides32 through the modulation of the interplay of

gauche and anomeric effects (Sections 4 - 6). It has been found that the electronic nature of the

aglycone is reflected, through the anomeric effect, in ∆G° of the N � S equilibrium of the constituent

pentofuranose moiety. In fact, the pKa value of the constituent nucleobase can be independently

measured from a pD-dependent change of ∆G° of the N � S equilibrium, which is closely similar to

the pKa value obtained from the pD-dependent chemical shift plots. We have examined41 3JHH

of aristeromycin and its 2' and 3'-deoxy derivatives at pD ≈2 and ≈7 in the 278 K - 358 K range to

examine whether the protonation � deprotonation equilibrium of the adenin-9-yl base also drives the

pseudorotational equilibrium of their constituent cyclopentane moieties. As expected, temperature-

dependent 3JHH values are virtually unchanged from acidic to neutral pDs (± 0.15 Hz, Table 1 in ref.

41) suggesting that the change of the electronic nature of the aglycone in aristeromycin has no effect

on the drive of the constituent cyclopentane conformation, thereby establishing the role of the

anomeric effect in the drive of the sugar conformation in N- and C-nucleosides17,19-30,32,33,35-38,40-42.

The comparison of the Energy barrier of the pseudoroation in natural pentofuranosyl

nucleoside versus the carbocyclic counterpart.

Our ab initio calculations (Fig 7) at HF/3-21G* level with the GAUSSIAN 94 program298 on

various pseudorotamers (with defined P values in the 0 - 360° range at a constant Ψm for all

conformers of a particular compound) of aristeromycin (8), 2'-deoxyaristeromycin (9), 3'-

deoxyaristeromycin (10), 2',3'-dideoxyaristeromycin (11) and their natural pentofuranosyl

counterparts have allowed us to further shed light on the validity of the two-state N � S (or W)



29

equilibrium model for the cyclopentane ring in carbocylic nucleosides (as suggested by the presence

of two major energy wells, vide infra) and to assess the effect of the substitution of CH2 for O4' upon

the energy barrier of pseudorotation in the modified versus natural nucleosides. This comparison is

intended to improve our understanding of the energy penalty of stereoelectronic gauche and anomeric

effects in natural nucleosides and nucleotides.

(i) 2',3'-dideoxynucleosides constitute the system of choice to analyze and compare the relative

flexibility of sugar conformation in carbocyclic versus natural pentofuranosyl nucleosides since 2'-

and 3'-OH groups are absent, which prevents any H-bonding interaction between themselves or with

the nucleobase [see subsections (iii) and (iv) below]. Both for 2',3'-dideoxyaristeromycin (11) and β-

D-ddA (30), N-type pseudorotamers (P = 0°) are slightly preferred over the S-type counterparts (P =

180°) [by ≈ 0.7 kcalmol-1 for 11 and ≈ 1.2 kcalmol-1 for 30 (Compare Panels (A) and (B) in Fig 7].

However, whereas for β-D-ddA, the activation energy barrier (∆∆G°‡) in the E region for the N to S

interconversion is ≈ 7.6 kcalmol-1 (see Section 2.6 for an upper limit of energy of activation of

pseudorotation found experimentally) the same transition in 2',3'-dideoxyaristeromycin is more facile

(∆∆G°‡ ≈ 3.7 kcalmol-1) which uniquevocally shows that the O4'-C1'-N9 anomeric effect efficiently

hampers free pseudorotation in the former. If we now for the sake of convenience consider that 2',3'-

dideoxyaristeromycin and β-D-ddA (30) are isosteric, then a simple subtraction of their above ∆∆G°‡ values gives us an idea of the influence of the anomeric effect on the energy barrier of the

pseudorotation in β-D-ddA (30), which is 3.9 kcalmol-1. For nearly all optimized pseudorotamers of

11 and 30, the nucleobase is in anti orientation and β is found in the anti range while γ+ rotamers are

preferred [Panels (I) and (J) in Fig 7].

(ii) For β-D-dA (37), among all possible pseudorotamers, N and S-type sugar conformations

(both having the same energy) are energetically preferred as shown by the energy profile [Panel (D) in

Fig 7]. The activation energy barrier from N to S through E is reduced (∆∆G°‡ ≈ 4.7 kcalmol-1) in

comparison with β-D-ddA (7.6 kcalmol-1, see panel (B) in Fig 7), showing the effect of 3'-OH. The

constituent nucleobase remains in the anti orientation and γt, ε3t and βt rotamers (see the legend of

Fig 7 for definition of the ε3 torsion angle) are preferred throughout the whole pseudorotational cycle

[Panel (L) in Fig 7]. For 2'-deoxyaristeromycin (9), among all pseudorotamers, the S- or W-type

conformers (210° < P < 240°) are energetically preferred by ≈ 3 kcalmol-1 over the N-type

counterparts (330° < P < 60°) [Panel (C) in Fig 7]. However, in contrast with the case of β-D-dA, the

interconversion between the energetically preferred N- and S-/W-type conformations takes place

through a very small barrier in the E-region (∆∆G°‡ ≈ 1.1 kcalmol-1) which is consistent with the

absence of the O4'-C1'-N9 anomeric and [O3'-C3'-C4'-O4'] gauche effects, as discussed for 2',3'-

Chatt

opadhya

ya e

t al,

"S

tere

oel

ectr

onic

Eff

ects

in N

ucl

eosi

des

& N

ucl

eoti

des

and t

hei

r S

truct

ura

l Im

pli

cati

ons"

,

Dep

t of

Bio

org

anic

Chem

istr

y, B

ox 5

81, U

ppsa

la U

niv

ersi

ty, S

-75123 U

ppsa

la, S

wed

en, V

er 1

60205 j

yoti

@boc.

uu.s

e

30

Bas

eO

OR

O

OH

OH

Bas

e

Bas

eO

OH O

HO

H

O

OH

Bas

e

OB

ase

OH

OH

O

OR

'

Bas

eO

R

N

NNN

NH2

NN

N

NNH2

O

NH

NO

O

NH

NO

O

H3C

NH

NO

O

F

Bas

eO

OH

O

P

OH

−O O

Bas

eO

OH

O

P

O

−O O

H2

CC

H3

Bas

eO

OH

O

P

OH

−O O

OH

Bas

eO

OH

O

P

O

−O O

H2

C

CH3

OH

H

NN

N

NH

O

OO

H OH

OH

H

NN

N

N

NH2

OO

H OH

OH

H

N

N

N

NH2

O

OH O

HO

H

NN

H

NH2

OO

OH OH

OH

RN

NR

'

O

OO

OH O

HO

H

AO

OH

OH

NH

NNN

O

NH2

16:

R =

Me;

R' =

OM

e; R

" =

H

15:

R =

R"

= H

; R

' = O

Me

1

3:

R =

R"

= H

; R

' = O

H

14:

R =

H;

R' =

R"

= O

H

28: α

-L-d

C

cyto

sin

-1-y

l (C

)

Bas

e =

aden

yl-

9-y

l (A

)

25: α

-D-d

C (

R =

R' =

H)

9:

2'-d

eoxy-a

rist

ero

myci

n

(R

= O

H;

R' =

H)

10:

3'-d

eoxy-a

rist

ero

myci

n

(

R =

H;

R' =

OH

)

hyp

oxan

thin

-9-y

l (I

)

17: α

-D-d

dA

23:

3',5

'-d

iOM

e-α

-D-d

A (

R =

R' =

Me)

18: α

-D-d

dG

21

: α

-D-d

A (

R =

R' =

H)

22:

3'-O

Me-α

-D-d

A (

R =

Me;

R' =

H)

19: α

-D-d

dC

20: α

-D-d

dT

24

: α

-D-d

G (

R =

R' =

H)

27: α

-L-d

A

26

: α

-D-T

(R

= R

' = H

)

gu

anin

-9-y

l (G

)

52: β

-D-C

54: β-D

-U

34: β-D

-dd

T (

R =

H)

31: β-D

-dd

G (

R =

H)

29: α

-L-T

33: β-D

-dd

C (

R =

H)

50: β

-D-A

32:

5'-O

Me-β

-D-d

dG

(R

= M

e)

51: β

-D-G

47: β

-L-d

G

46: β

-L-d

A

53: β

-D-r

T

30: β-D

-dd

A (

R =

H)

γ

β

55:

5-F

-β-D

-U

35: β-D

-dd

U (

R =

H)

49: β

-L-T

imid

azo

l-

1-y

l (I

m)

thym

in-1

-yl

(T)

ura

cil-

1-y

l (U

)5

-F-u

raci

l-1

-yl

[5-F

-dU

(4

5)

& 5

-F-U

(5

5)]

Na+

64

: β

-D-d

AM

P

65: β-D

-dG

MP

66: β-D

-dC

MP

67: β-D

-TM

P

68: β-D

-dU

MP

ε

δ

β

74: β-D

-AM

P

48: β

-L-d

C

72: β-D

-TM

PE

t

γ

ζ

69: β-D

-dA

MP

Et

70: β-D

-dG

MP

Et

71: β-D

-dC

MP

Et

β

76: β-D

-CM

P

78: β-D

-rU

MP

α

77: β-D

-rT

MP

75: β-D

-GM

P

Na+

Na+

α

6

β

9

Na+

δ ζ

79: β-D

-rA

MP

Et

80: β-D

-rG

MP

Et

81: β-D

-rC

MP

Et

82: β-D

-rT

MP

Et

83: β-D

-rU

MP

Et

73: β-D

-dU

MP

Et

1 28

8

ε

75

3

7

4

5

9

1

57:

Fo

rmyci

n A

56:

Fo

rmyci

n B

58:

9-d

eaza

-A

6 3

1

42

5

3

721

9

6

8

2

4

5

6

5

3

60: ψ

-U:

R =

R' =

H12

6

59: ψ

-iso

C

44

61:

1-M

e-ψ

-U:

R

= M

e, R

' = H

3

63: β-D

-3'-d

A6

2:

1,3

-diM

e-ψ

-U:

R

= R

' = M

e

8:

Ari

ster

om

yci

n (

R =

R' =

OH

)

HO

OR

H

R'

R"

Bas

eO

OR

'

OR

AO

H RR

'

12:

R =

R' =

R"

= H

H6'

H6"

39

: 3

',5'-d

iOM

e-β-D

-dA

(

R =

R' =

Me)

37: β

-D-d

A (

R =

R' =

H)

38

: 3

'-O

Me-β

-D-d

A

(

R =

Me;

R' =

H)

41: β-D

-dG

(R

= R

' = H

)

42: β-D

-dC

(R

= R

' = H

)

NH

NNN

O

45

: 5

-F-β

-D-d

U (

R =

R' =

H)

44

: β-D

-dU

(R

= R

' = H

)

43: β-D

-T (

R =

R' =

H)

40: β-D

-dIm

(R

= R

' = H

)

36: β-D

-dd

I (R

= H

)1

1:

2',3

'-d

ideo

xy-a

rist

ero

myci

n

(

R =

R' =

H)

59

b: ψ

-iso

C, H

Cl

Chatt

opadhya

ya e

t al,

"S

tere

oel

ectr

onic

Eff

ects

in N

ucl

eosi

des

& N

ucl

eoti

des

and t

hei

r S

truct

ura

l Im

pli

cati

ons"

,

Dep

t of

Bio

org

anic

Chem

istr

y, B

ox 5

81, U

ppsa

la U

niv

ersi

ty, S

-75123 U

ppsa

la, S

wed

en, V

er 1

60205 j

yoti

@boc.

uu.s

e

31

O

R F3"

H2"

B

H2'

H3'

H4'

H5"

H5'

H1'

O

OR1

H3"

OR

H2'

F3'

H4'

H5"

H5'

H1'

BO

OM

MT

r

F3"

H2"

B

HO

H3'

H4'

H5"

H5'

H1'

O

OH H

3"

H2"

B

F2'

H3'

H4'

H5"

H5'

H1'

O

OH

H3"

H2'

B

F2"

H3'

H4'

H5"

H5'

H1'

O

OR

F3'

H2"

B

H2'

H3"

H4'

H5"

H5'

H1'

CO

OC

H3

CO

OC

H3

HA

F

HB

O

CO

OH

CO

OH

HA

F

HB

R1

R3

R2

R4

Cl

Cl

Cl

Cl

Cl

Cl

F

F

F

F2

H4

FF

F

F

F

F

F

O

OH O

H

F2'

B

F2"

H3'

H4'

H5"

H5'

H1'

N

H2ax

H2eq

H4ax

H4eq

F

R

H

HN

H6ax

H6eq

H4ax

H4eq

F

H

H

CO

OH

N

H2ax

H2eq

H4ax

H4eq

F

CO

O

H

Me

Me

Me

FH

O

Hax

Me

FM

e

Me

HB

HA

OH

O

OH O

H

H2'

B

F2"

H3'

H4'

H5"

H5'

H1'

O

OR

F3'

H2"

B

H2'

H3"

H4'

H5"

H5'

H1'

G =

guan

in-9

-yl

96

: B

= U

(F

2'd

dU

) 88

: R

= O

H, B

= T

(F

LT

)

115

: X

= N

O2

92

: R

= R

1 =

H, B

= A

(F

XA

)

90

: R

= N

H2, B

= T

(A

FL

T) 99

91

: R

= O

H, B

= U

(F

3"d

dU

)

98

:R =

H, B

= U

(F

3'd

dU

)

93

: R

= H

; R1 =

MM

Tr,

B

= A

(F

XA

5)

97

: B

= U

(F

2"d

dU

)

89

: R

= O

Tr,

B =

T (

FL

T5)

104

C =

cyto

sin-1

-yl

A =

aden

in-9

-yl

T =

thym

in-1

-yl

103

: R1 =

FA

; R2 =

FB

;

R3 =

HA

; R4 =

FB

; 100

94

: R

= C

S-O

Ph;

R1 =

MM

Tr,

B =

A (

FX

A25)

101

: R1 =

FA

; R2 =

HB

;

R3 =

HA

; R4 =

FB

;

95

: B

= T

(F

3A

T)

102

: R1 =

HA

; R2 =

HB

;

R3 =

FA

; R4 =

FB

;

U =

ura

cil-

1-y

l

85

: B

= G

(diF

G)

86

: B

= T

(diF

T)

110

109

112

: B

= C

(F

2"C

)

87

: B

= C

(diF

C)

105

: R

= H

106

: R

= C

OO

H

108

84

: B

= A

(diF

A)

107

O

OH X

TH5"

H5'

114

: X

= O

Me

116

: X

= O

CF3

111

:R =

H, B

= A

(F

3'd

dA

)

113

: X

= N

H2

Chatt

opadhya

ya e

t al,

"S

tere

oel

ectr

onic

Eff

ects

in N

ucl

eosi

des

& N

ucl

eoti

des

and t

hei

r S

truct

ura

l Im

pli

cati

ons"

,

Dep

t of

Bio

org

anic

Chem

istr

y, B

ox 5

81, U

ppsa

la U

niv

ersi

ty, S

-75123 U

ppsa

la, S

wed

en, V

er 1

60205 j

yoti

@boc.

uu.s

e

32

(A)

2',3'-did

eo

xya

riste

rom

ycin

(11

)

P (o

)

-60

060

120

180

240

E (kcal mol-1

)

048

12

16

(C)

2'-d

eoxyari

ste

rom

ycin

(9

)

P (o

)

-60

060

120

180

240

E (kcal mol-1

)

048

12

16

(E)

3'-de

oxya

riste

rom

ycin

(10

)

P (o

)

-60

060

120

180

240

E (kcal mol-1

)

048

12

16

(G)

ari

ste

rom

ycin

(8)

P (o

)

-60

060

120

180

240

E (kcal mol-1

)

048

12

16

(B) β

-D-d

dA

(30

) P (o

)

-60

060

120

180

240

E (kcal mol-1

)

048

12

16

(D) β-D

-dA

(37)

P (o

)

-60

060

120

180

240

E (kcal mol-1

)

048

12

16

(F) β

-D-3

'-d

A (63

) P (o

)

-60

060

120

180

240

E (kcal mol-1

)

048

12

16

(H) β-D

-A (50

)

P (o

)

-60

060

120

180

240

E (kcal mol-1

)

048

12

16

(I)

2',3

'-d

ide

oxya

riste

rom

ycin

(11

)

P (o

)-60

060

120

180

240

Torsion angle (o)

-240

-180

-120

-600

60

χ

γ

β

(J) β-D

-dd

A (30

) P (o

)-60

060

120

180

240

Torsion angle (o)

-240

-180

-120

-600

60

χ

γ

β

(K)

2'-d

eoxya

riste

rom

ycin

(9

)

P (o

)-60

060

120

180

240

Torsion angles (o)

-240

-210

-180

-150

-120

-90

-60

χ

γβε 3

(L) β

-D-d

A (37

)

P (o

)-60

060

120

180

240

Torsion angle (o)

160

170

180

190

200

χγ

βε 3

(N) β

-D-3

'-d

A (63) P (o

)

-60

060

120

180

240

Torsion angles (o)

-240

-180

-120

-600

60

d(H-bond) (Å)

2345

d(2'-OH-N

3)

d(5'-OH-O

4')

χ

γβε 2

(M)

3'-d

eo

xya

riste

rom

ycin

(10

)

P (o

)

-60

060

120

180

240

Torsion angles (o)

-240

-180

-120

-600

603456

d(2'-OH-N

3)

χ γβ

ε 2

d(H-bond) (Å)

(P) β

-D-A

(50)

P (o

)-60

060

120

180

240

Torsion angles (o)

-240

-180

-120

-60

d(H-bond) (Å)

246

d(3'-OH-O

2')

χγ

β

ε 2

ε 3

d(2'-OH-O

3')

d(5'-OH-O

4')

d(2'-OH-N

3)

(O)

ari

ste

rom

ycin

(8

)

P (o

)-60

060

120

180

240

Torsion angles (o)

-240

-180

-120

-600

60

d(H-bond) (Å)

246

χ γβ

ε 2

d(2'-OH-N

3)

d(2'-OH-O

3')

d(3'-OH-O

2')

ε 3

Fig

ure

7.

The

plo

ts o

f th

e en

ergy (

E),

of

var

ious

tors

ion a

ngle

s an

d o

f in

tera

tom

ic d

ista

nce

s fo

r a

b i

nit

io o

pti

miz

ed p

seud

oro

tam

ers

of

2',3

'-d

ideo

xyar

iste

rom

yci

n (11

) [P

anel

s (A

) an

d

(I)]

, 2

'-d

eoxyar

iste

rom

yci

n (9

) [P

anel

s (C

) an

d (K

)],

3'-d

eoxyar

iste

rom

yci

n (10

) [P

anel

s (E

) an

d (M

)],

aris

tero

myci

n (8

) [P

anel

s (G

) an

d (O

)] a

nd

of

thei

r p

ento

fura

no

syl

counte

rpar

ts

β-D

-dd

A (30

) [P

anel

s (B

) an

d (J

)], β

-D-d

A (37

) [P

anel

s (D

) an

d (L

)], β

-D-3

'-d

A (63

) [P

anel

s (F

) an

d (N

)] a

nd

β-D

-A (50

) [P

anel

s (H

) an

d (P

)] a

s a

funct

ion o

f th

e p

has

e an

gle

of

pse

ud

oro

tati

on (

P).

The

calc

ula

tio

ns

hav

e b

een p

erfo

rmed

by c

onst

rain

ing o

nly

tw

o e

nd

ocy

clic

to

rsio

ns

(ν0 a

nd

ν4)

to t

he

app

rop

riat

e val

ues

in o

rder

to

sw

eep

P f

rom

0 t

o 3

60°

in 3

0°

(fo

r 8

- 10

, 37

, 50

and

63

or

20°



33 (for 11 and 30) intervals at a common Ψm value [41° for 8 - 11 (as in the crystal structure292 of 8), 40° for 30 and 36° (as

in the crystal structure of adenosine259) for 37, 50 and 63] whereas the rest of the molecule has been freely optimized. For

all conformers, the starting geometries used as input for the optimization with GAUSSIAN program298 were as follows (ε2

and ε3 designate the [2'-OH-O2'-C2'-C1'] and [3'-OH-O3'-C3'-C4'] torsion angles, respectively): χ = 120°, β = γ = 180° for

2',3'-dideoxyaristeromycin (11); χ = -170°, β = γ = 180° for β-D-ddA (30); χ = 120°, β = γ = ε3 = 180° for 2'-

deoxyaristeromycin (9); χ = -170°, β = γ = ε3 = 180° for β-D-dA (37); χ = 120°, β = γ = ε2 = 180° for 3'-

deoxyaristeromycin (10); χ = -170°, β = γ = ε2 = 180° for β-D-3'-dA (63); χ = 120°, β = γ = ε2 = 180° and ε3 = -60° for

aristeromycin (8) and χ = -170°, β = γ = ε2 = 180° and ε3 = -60° for β-D-A (50).

dideoxyaristeromycin under the subsection (i). It is also noteworthy that whereas the values of the γ, ε3

and β torsion angles (in the anti domain) do not change drastically from one pseudorotamer of 2'-

deoxyaristeromycin to the other, the nucleobase adopts a syn orientation in the high energy W-type

pseudorotamers while for all other sugar conformations it is anti [Panel (K) in Fig 7]. (iii) For 3'-

deoxyaristeromycin (10), the W-type (P = 240°) pseudorotamers are preferred among all possible

geometries and they are more stable by ≈ 6 kcalmol-1 than the N-type counterparts giving rise to a

local energy minimum at P ≈ 330° [Panel (E) in Fig 7]. The interconversion between N- and S-type

pseudorotamers is possible, provided that the energy barrier in the E-region (∆∆G°‡ ≈ 4.3 kcalmol-1) is

overcome. It is noteworthy that the preference for W-type (P = 240°) geometries of 3'-deoxyaristeromycin results from the gauche+ orientation of 2'-OH hydroxy group (ε2 ≈ 60°) which

enables its hydrogen-bonding interaction [d(2'-OH-N3) ≈ 2.5 Å] with the nucleobase in antiperiplanar

orientation [Panel (M) in Fig 7]. On the other hand, the relatively higher energy of other

pseudorotamers in the W-region (P = 270° and P = 300°) is associated with a syn orientation of the

nucleobase. The energy profile presented in Panel (F) for the natural counterpart β-D-3'-dA (63) is far

more complex than that of 3'-deoxyaristeromycin and suggests a strong preference for E-type (P = 90°)

or S-type (P = 180° or 210°) pseudorotamers (by ≈ 5 - 5.5 kcalmol-1) over the local energy minimum in the N-region (P = 0°). The preference for E-type (ε2

+ with β-) and S-type (ε2+ with β+) pseudorotamers

is owing to a stabilizing H-bonding interaction between 2'-OH proton and N3 [d(2'-OH-N3) ≈ 2.0 Å] and

between 5'-OH and O4' [d(5'-OH-O4') ≈ 2.0 Å] (for E-type conformations) [Panel (N) in Fig 7].

(iv) For aristeromycin (8), the S/W type conformers (120° < P < 240°) are preferred over the N-

type (330° < P < 30°) geometries by ≈ 10 kcalmol-1 [Panel (G) in Fig 7]. The N- to S/W-type

geometries interconversions preferably take place through the activation energy barrier in the E (∆∆G°‡

≈ 4.8 kcalmol-1 from N to S; ≈ 13.0 kcalmol-1 from S to N) rather than the W (∆∆G°‡ ≈ 5.8 kcalmol-1

from N to S; ≈ 14.1 kcalmol-1 from S to N) region of the pseudorotation cycle. The increased stability

of S/W-type pseudorotamers with respect to other geometries is due to 2'-OH....N3 and 3'-OH....O2' hydrogen-bonding interactions which are permitted by ε2

+ and ε3t orientations [Panel (O) in Fig 7]. On

the other hand for P values in the range 270° < P < 90°, a single H-bonding interaction is possible

between 2'-OH and O3'. As for 2'-deoxy and 3'-deoxyaristeromycin (see (ii) and (iii)), in the relatively

unstable conformers in the W-region (P = 270° and P = 300°), a syn orientation of the nucleobase is

observed, whereas for the energy minima the nucleobase is anti. In fact, the energy profile [Panel (G)

in Fig 7] and the dependence of the glycosyl torsion angle χ on the value of P appear to be very similar.

The plot presented Panel (H) in Fig 7 shows that the energy of various pseudorotamers of β-D-A (50)


34

is much less sensitive to the phase angle value than for the aristeromycin counterpart [Panel (G) in Fig

7], possibly because of the fact that all pseudorotamers are stabilized by the H-bonding interaction

between 2'-OH and O3' [Panel (P) in Fig 7]. The absolute energy minimum is found for N-type

conformations of β-D-A (P = 0°) which are preferred over secondary minima in the E/S-regions at P =

120° (by 0.9 kcalmol-1) and P = 210° (by 2.3 kcalmol-1). The activation energy barriers between N-

and E- (P = 120°) (through P = 60°: 5.5 kcalmol-1 and 4.6 kcalmol-1) and from E- to S- (P = 210°)

(through P = 150°: 3.5 kcalmol-1 and 2.2 kcalmol-1) are reduced in comparison with what was found

for aristeromycin. For all pseudorotamers, γt rotamers are preferred and the nucleobase is in anti orientation with ε2

t for 270° < P < 90° and ε2- for 90° < P < 270°. The preference for P = 0° among all

pseudorotamers in the N region correlates nicely with the fact that in this situation ε2 is in the maximal

trans orientation. However, this trans orientation does not result into an H-bond between 2'-OH and N3 as shown by the fact that the d(2'-OH-N3) distance is greater than 4 Å at any P value. On the other hand,

the presence of the secondary energy minimum at P = 120° may be attributed to a hydrogen-bond

between 5'-OH and O4' permitted by the β- orientation.

In summary, the energy plots in Fig 7 show the energy penalties induced by some specific

conformational properties of the cyclopentane ring owing to the absence of O4' in carbocyclic

nucleosides vis-a-vis natural pentofuranosyl nucleosides, in which O4' is the cause of both gauche and

anomeric effects. Work is now in progress in our laboratory to perform the above calculations at a

higher basis set in order to delineate the influence of hydration on the pseudorotational energy and

conformational profile in the carbocylic as well as in pentofuranosyl nucleosides.

2.6 Energy barriers of the pseudorotation cycle of β-D-nucleosides

Early quantum chemical, CNDO or classical potential energy calculations250,299-302 and more

recent calculations using consistent force field method303 (permitting bond stretching and bond angle

bending) on unsubstituted ribo and deoxyribofuranose have shown that the N-type and S-type sugars

are preferred among all pseudorotamers, and that they have almost the same energy. Calculations on 1-

amino-ribose, -2-deoxyribose or -3-deoxyribose304,305 and on nucleosides themselves306 have shown

that the activation energy barrier for N- to S- interconversions is clearly greater in the W (≈ 24 kJ/mol

for 2-deoxyribofuranose, ≈ 31 kJ/mol for ribofuranose) than in the E region (≈ 7.5 kJ/mol for 2-

deoxyribofuranose, ≈ 16 kJ/mol for ribofuranose).

The measurement307 of 13C longitudinal relaxation times of single tertiary carbons showed that

the energy of activation of the N- to S-type sugar interconversion in purine nucleosides is ≈ 20 ± 2

kJ/mol, whereas for pyrimidines the internal motions are slow compared with the rotational diffusion of the whole molecule, as a result of a possible H-bonding between the 5'CH2OH group and the base,

and the authors suggested that this might raise the apparent activation energy of pseudorotation for

uridine and cytosine above 25 kJ/mol.

A recent temperature-dependent 2H and 13C relaxation study from our laboratory308 on

selectively deuterated thymidines and allofuranoses as well as their comparisons with the

conformationally constrained analogues and abasic sugars did not allow us to determine the activation

energy barrier of pseudorotation because the internal motions are heavily coupled with the overall

molecular reorientations, which prevents dissection of the observed activation energy barrier of 20-23



35

(±0.9) kJmol-1 into contribution from pseudorotational interconversions, rotation around the glycosyl

and C4'-C5' torsions and overall tumbling. Nevertheless, this experimental estimate308 gives an upper

limit for the pseudorotational barrier (see Appendix for detailed discussion on the estimates of the

barrier).

2.7 Steric effect of the nucleobase on the sugar conformation

The inspection of the three molecular fragments in nucleos(t)ides (i.e. the aglycone, the sugar

and the phosphate) allows us to estimate the bias of the two-state N � S pseudorotational equilibrium

of the constituent pentofuranose moiety on the basis of various steric and stereoelectronic effects.

The N-nucleobase influences the pentofuranose conformation through its inherent steric effect

and counteracting stereoelectronic interactions within O4'-C1'-N1/9 fragment. In β-D-nucleosides,

among all pseudorotamers, steric repulsions penalize O4'-exo (W-type) conformation304-306,309 most

owing to the 1,3-diaxial orientation of the nucleobase and 5'CH2OH group and to the eclipsed

arrangement of C2' and C3' substituents (Fig 6). In O4'-endo (E-type) pseudorotamers, the steric

repulsions between the nucleobase and the 5'CH2OH group are minimized, however the substituents at

C2' and C3', just as in the case of W-type conformers, are in the unfavorable eclipsed orientation (Fig.

6). Conformational analyses of nucleos(t)ides in aqueous solution are performed on the basis of a two-

state N �S equilibrium model (vide supra). In terms of steric effect alone, for β-D-nucleosides, S-type

pseudorotamers are energetically favoured in comparison with N-type counterparts, since the

pseudoequatorially oriented nucleobase in the former exerts less steric repulsions with the other

substituents on the pentofuranose moiety than when it is pseudoaxial in the later (Figs 6 and 8A).

In α-D-nucleosides, the nucleobase and the 5'CH2OH group are on opposite faces of the

pentofuranose sugar, therefore their steric interactions will be minimal in comparison with the situation

in β-nucleosides. In N-type conformers, the nucleobase and 3'-OH are pseudoequatorial while 2'-OH is

pseudoaxial, whereas the opposite is true for S-type pseudorotamers. Both in W- and E-type

conformers, 2'-OH and 3'-OH are pseudoaxial, whereas the nucleobase is pseudoequatorial in W-type

geometry and pseudoaxial in the E-type counterpart. In α-D-2'-deoxynucleosides, the nucleobase exerts

stronger steric repulsions with 3'-OH and H4' in N-type than in S-type pseudorotamers. Therefore, in

solution, taking into consideration of a two-state N � S equilibrium model, S-type pseudorotamers

would be energetically favoured over the N-type counterparts if the steric effect alone were considered.

Potential energy calculations on α- and β-2'-deoxy (or ribo) adenosine and thymidine have

shown that syn conformations of the nucleobase in the α-anomers are energetically less favourable over

the entire range of phase angle values, and that the energy barrier of anti � syn interconversions is

higher in the α-ribo than in the α-2'-deoxy series. Highly hypothetical syn orientations of a nucleobase

in α-2'-deoxynucleosides may be stabilized by hydrogen-bonding interactions between N3 (or O2) and

3'-OH.

From the previous observations, it can be derived that as the bulk of the nucleobase is

increased, one expects the bias of the two-state N � S pseudorotational equilibrium to be shifted

toward more S-type and N-type conformations in β- and α-nucleosides, respectively.


36

2.8 O4'-C1'-N1/9 stereoelectronic effect (the anomeric effect)

The anomeric effect in heterocyclic six-membered rings is commonly explained in terms of

stabilizing hyperconjugative interactions (molecular orbital overlap model)115,116 or, alternatively, by

destabilizing dipole-dipole repulsions65,117,118 (Section 1.3 and Table 1 for the energetics of the

anomeric effect in 2-substituted hexopyranoses and derivatives).

Similarly, the O4'-C1'-N1/9 anomeric effect in nucleosides and nucleotides may be explained

either (i) by stabilizing nO4'

→σ∗C1'-N1/N9 orbital interactions between the orbital of one of the

endocyclic O4' electron lonepairs (nO4'

) and the antibonding orbital of the C1'-N1/9 glycosyl bond

(σ∗C1'-N1/N9

) (Fig 9), or (ii) by destabilizing electrostatic repulsions between two dipoles [i.e. the dipole

of the furanose ring, on one hand (which is itself the resultant of C4'-O4', O4'-C1' individual dipole

moments and of the dipole induced by O4' lonepairs) and the dipole oriented from C1' to N1/9, on the

other (Fig 8B-C)].

In β-D-nucleosides, O4'-C1'-N1/9 stereoelectronic interactions are most efficient in W-type

sugar geometries, where one of the O4' lonepairs is in optimal antiperiplanar orientation with respect to

the C1'-N1/9 bond, and at the same time dipole-dipole repulsions are minimal since the angle between

both dipoles is nearly 90°. However, for the steric reasons stated above, West-type conformers have

neither been found in crystal structures of β-D-nucleos(t)ides nor in the conformational equilibria in

solution. Conversely, in E-type pseudorotamers, both O4' lonepairs are gauche with respect to the C1'-

N1/9 bond, which makes hyperconjugation (or molecular orbital overlap) least efficient;

Simultaneously, dipole-dipole repulsions are maximal, owing to the fact that both dipoles are nearly

parallel.

Figure 8: Steric effect of the base and

rationalization of the anomeric effect in

nucleos(t)ides in terms of electrostatic repulsions,

as examplified for β-D-dA (37). (A) The drive of

the two-state N �S equilibrium toward S- over

N-type pseudorotamers by the steric effect of the

nucleobase. (B) & (C) Stronger electrostatic

repulsions between the pentofuranose ring dipole

(black arrow) and the C1'-N9 dipole (white

arrow) in S-than in N-type sugars. Both dipoles

are nearly parallel in the S-type pseudorotamers

whereas they are nearly perpendicular in N-type

sugars (β << α, Panel C).

The anomeric effect is therefore often

invoked as one of the factors responsible

for the activation energy barrier

encountered in the East region of the pseudorotational cycle for β-D-nucleosides, which, besides the

stronger barrier in the West region, also hampers free pseudorotation of the constituent pentofuranose.

The comparison of the geometries of N- and S-type pseudorotamers of β-D-nucleosides shows

that dipole-dipole repulsions (Fig 8B-C) and O4'-C1'-N1/9 hyperconjugative interactions (Fig 9) are

OHOH2C

O

HOH2C

OH

HO

NN

N

N

N

N

N

N

NH2

NH2

H1'

N

C2'

C4'

H1'

N

C2'

C4'

O4'C

O

OH

HO

NN

N

N9

N

N

N

N9

NH2

NH2

H

HO

HH

H

HH

H

H

CHO

HH

(B)

(C)

(A)

αβ



37

reduced and enhanced, respectively, in the former compared to the latter. Therefore, O4'-C1'-N1/9

stereoelectronic interactions drive the two-state N � S pseudorotational equilibrium in β-D-

nucleosides, in aqueous solution, toward N-type conformations.

Figure 9. Rationalization of the O4'-C1'-

N1/9 anomeric effect in nucleos(t)ides in

terms of molecular orbital overlap and

hyperconjugation, as examplified for β-D-

dA (37). The O4' lonepairs orbitals are

represented using either the sp2 [i.e. higher

energy 1nsp2(p-type)

lonepair with

predominant p-type character and lower

energy 2nsp2(s-type)

lonepair with

predominant s-type character] or sp3 [i.e. 1n

sp3 and 2nsp3

lonepairs with the same

energy] hybridization models123-126,129,130

(Section 1.9). (A) Molecular orbital overlap

model116: The overlap (i.e. nO4' σ*C1'-N9)

of a lonepair orbital of O4' [1nsp2 (p-type)]

with the σ*C1'-N9 antibonding orbital of the

glycosyl bond results in the stabilization of

N- over S-type pseudorotamers and shifts

the pseudorotational equilibrium toward N.

(B) & (C) Influence of sp3 (i.e. 1nsp3

and

2nsp3

, Panel (B)) versus sp2 (1nsp2

(p-type)

and 2nsp2

(s-type), Panel (C)) hybridization of

the O4' lonepairs orbitals on their relative

orientation with respect to the glycosyl bond

and on the efficiency of the O4'-C1'-N9 stereoelectronic interactions. For a typical N-type pseudorotamer (the angles in the

projections are calculated for P = 0° and Ψm = 40°, assuming perfect trigonal symmetry), the nO4' σ*C1'-N9 interaction is

possible due to the near antiperiplanar orientation of either 1nsp3

(β1 ≈ 132°) or 1n

sp2 (p-type) (β2 ≈ 162°) with respect to

σ*C1'-N9 orbital, whereas in the S-type sugar counterpart (the angles are calculated for P = 160°, Ψm = 40°) the efficiency of

the interaction is much reduced as the result of the more accute β1 and β

2 angles compared to those in the N-type sugar. (D)

Double-bond � No-bond resonance resulting from hyperconjugation of one of the O4' lonepairs to the glycosyl C1'-N9

bond115. This model is consistent with the shortening of O4'-C1' bond compared to O4'-C4' bond observed in the crystal

structures174,310,311 of β-nucleos(t)ides. (E) This overlap, which results in the formation of a stabilizing (occupied) and a

destabilizing (unoccupied) orbitals, becomes more efficient as the square of the overlap (S) between both orbitals increases

and as the difference (∆E) between their respective energies decreases4,5. A maximal interaction requires an antiperiplanar

orientation of nO4' relative to the C1'-N9 bond.

The significant change (Fig 2E)1,121,122 of the endocyclic and exocyclic C-O bond lengths in

combination with concommitant closing or opening of the corresponding bond angles in crystal

structures of chair conformers of various 2-substituted-tetrahydropyrans with axially versus

equatorially oriented anomeric substituents is considered as a strong evidence for the hyperconjugative

origin of the anomeric effect in such systems. A perusal174,310,311 of the crystal structures of β-D-N-

nucleosides also shows that the C1'-O4' bond is 0.03 - 0.04 Å shorter than the C4'-O4' bond, which

OHOH2C

O

HOH2C

OH

HO

N

N

N

NH2

C2'

N9

H1'O4'

C4'

C2'

N9

H1'

O4'

C4'

C2'

N9

H1'O4'

C4'

C2'

N9

H1'

O4'

C4'

N9

OHOH2C

HOO

HOH2C

HO

(A)

(B)

(C)

(D)

β1 = 132ο

β1 = 96ο

β2 = 162ο

β2 = 126ο

σ*C1'-N9

δ+

δ−

2nsp3

1nsp2 (p-type)

2nsp2 (s-type)

1nsp2 (p-type)

1nsp3

1nsp3

2nsp3

2nsp2 (s-type)

1nsp2 (p-type)

1nsp2 (p-type)

(E)nO4'

σ*C1'-N9

ΔE(σ*C1'-N9 - nO4')

Stabilization due to anomeric effect (AE)

N

N

N

NH2

N9

N

N

N

NH2

N9

NN

N

NH2

N9

N-type sugar S-type sugar


38

goes hand in hand with a possible hyperconjugative origin for the anomeric effect in nucleosides (Fig

9).

2.9 Effect of the aglycone on the conformation of nucleos(t)ides and oligos

History of the discovery of stereoelectronic effects in nucleos(t)ides: The first set of eight

papers qualitatively indicating the influence of the electronic character of the C1'-aglycone on the

conformation of the constituent pentofuranose sugar in nucleos(t)ides were the result of almost

simultaneous independent observations: (i) That the UV spectrum of 1-methylcytosine and cytosine

nucleosides varies with the nature of the carbohydrate component is the first milestone paper that

established the cross-talk between the aglycone and the constituent sugar312. (ii) Guschlbauer et al

showed313 that the confomation of the ribose moiety changes with the protonation of the guanin-9-yl

base at N7 in 2'-GMP. (iii) As shown314 by Sarma et al., the sugar conformation is indeed different in

oxidized and reduced β-nicotinamide mononucleotides. (iv) Remin and Shugar showed315 that the

protonation of cytidine or arabinocytidine changes the 3J1'2' value by about 0.2 Hz. (v) Altona and

Sundaralingam did an important qualitative correlation on the nature of sugar pucker depending on the

nature of the aglycone basing on their X-ray crystal structures data181.

(vi) Subsequently, Altona and Sundaralingam extended their X-ray correlation data to the

coupling constant correlation, showing again the qualitative dependence of 3J1'2' with the nature of the

aglycone316. (vii) The preference of S versus N conformer is pH-dependent, as demonstrated for the

first time by Guschlbauer et al on guanosine phosphates317. (viii) A similar observation was also made

by Hruska et al on 2'-O-methyladenosine318.

Further qualitative evidences:

In 1987, we first showed14 basing on thermodynamic estimates on a set of four isomeric 2'/3'-

deoxynucleosides that the net result of the gauche effect and the anomeric effect is of major importance

in determining the overall furanose conformation. Beside the subsequent quantitative works from our

labs (Sections 4 - 6), many qualitative studies have been conducted to understand how the electronic

nature,309,319-323 protonation state198, bulkiness or substitution pattern207,239,324-327 and configuration

of the nucleobase175,328-346 or of the C1 substituent234 modulate the sugar conformation in nucleosides

and nucleotides as well as in furanosides347-351. The conformational analysis46 in solution of a dimer

containing a 4'-oxofuran derivative352, based on pseudorotational analysis of vicinal 3JHH, has shown

that the modified nucleoside adopts almost exclusively (89 %) S-type puckered geometries, as a result

of the cooperative drive by the O5'-C4'-O4' anomeric effect and the [O3'-C3'-C4'-O4'] gauche effect

(vide infra).

2.9.1 Configuration-dependent sugar conformation in furanosides

At the sugar level, spectroscopic studies based upon the interpretation of 3JHH, 3JCH and 3JCC

coupling constants347,349,353 have shown that the α- and β-anomers of D-ribofuranose (and of their 1-

methyl derivatives) adopt preferentially S- and N-type puckered geometries, respectively, owing to the

O1-C1-O4 anomeric effect, which places the anomeric group in pseudoaxial or nearly pseudoaxial



39

orientation. The contribution of the exo-anomeric effect to the relative stabilities and preferred

conformations of α- and β-D-erythrofuranose and -threofuranose, as obtained from ab initio

calculations, has also been adressed348.

2.9.2 Effect of electron-withdrawing nucleobases on pseudorotamer populations

1H-NMR studies have shown that in 2-substituted tetrahydropyran, the magnitude of the O-C-O

anomeric effect increases as the anomeric group becomes more electronegative (Section 1.5). In

nucleosides and nucleosides, similar trends have also been found:

(i) Egert et al., in their integrated approach321 consisting of an analysis of crystallographic data

and 1H-NMR spectra as well as quantum-chemical calculations on 5-substituted uridines, have shown

that electron-withdrawing substituents at C5 induce a lengthening and shortening of N1-C1' and C1'-

O4' bonds, respectively, in comparison with the reference compound, uridine. For electron-donating

substituents, the opposite situationwas observed. They invoked a possible charge transfer from one of

the O4' lonepairs to the C5 substituent to rationalize these tendencies. We, on the other hand, argue that

electron-withdrawing (or electron-donating) substituents at C5 are expected to strengthen (or weaken)

the nO4' →σ∗C1'-N9 interactions (i.e. the anomeric effect), which would in turn also lead to the

experimentally observed lengthening (shortening) and shortening (lengthening) of C1'-N1 and O4'-C1'

bond lengths, respectively. In that work, the concomitant opening (closing) of the glycosyl torsion

angle χ in 5-substituted uridines with electron-withdrawing (electron-donating) substituents at C5,

respectively, with respect to uridine, was explained by the fact that small χ values favour nO4' →πC5-C6

and nO4' →π*C5-C6 interactions, which are destabilizing and stabilizing, respectively. For electron-

withdrawing at C5, the stabilizing nO4' →π*C5-C6 interactions are strengthened while the destabilizing

nO4' →πC5-C6 interactions is weakened (in comparison with the reference compound uridine), leading

to the overall stabilization of the system, therefore such substituents favour small values of χ. In

contrast, electron-donating groups at C5 will induce a strengthening of the destabilizing nO4' →πC5-C6

interactions and a weakening of the nO4' →π*C5-C6 interactions (i.e. an overall destabilization of the

system) and in turn greater χ values will be found in comparison with uridine. A qualitative correlation

between the preference of the sugar moiety in 5-substituted uridine derivatives for N-type versus S-type

geometries and the value of χ was also proposed.

(ii) In their conformational study on 5-substituted uridine derivatives in aqueous solution by

1H-NMR spectroscopy, Uhl et al.322 have shown that the population of N-type pseudorotamers (% N)

increases from 44% to ≈ 90% going from NH2 to NO2 substituent at C5, and a plot of % N as a

function of the Hammett constant, σp, of the substituent at C5, gave a straightline. This is also

consistent with our proposition that electron-withdrawing (electron-donating) groups at C5 are

expected to strengthen (weaken) nO4' →σ∗C1'-N9 interactions (i.e. the anomeric effect) and therefore

stabilize (destabilize) the N- over S-type pseudorotamers. This is clearly evident from our quantitative

analysis showing that the protonation of the nucleobase in adenosine, guanosine and cytidine

nucleosides and in their deoxy derivativies strengthen the anomeric effect (i.e. increase of N-type

conformation) , whereas it is weakened by the deprotonation (i.e. decrease of N-type conformation),

compared to the neutral state (see Section 4 for detailed study).


40

(iii) We have quantitatively shown that the sugar conformation can be steered by altering the

electronic character of the nucleobase as well as by the change of the electronegativity of the sugar

substituent (see section 4). Subsequently, Rosemeyer and Seela have confirmed this by showing225 that

the extent of the preference for N-type pseudorotamers of 7-substituted 7-deaza-2'-deoxyadenosines

increases as the Hammett contant σm of the 7-substituent increases. This result is consistent with the

expected increase in the strength of the O4'-C1'-N9 anomeric effect as the glycosyl nitrogen becomes a

better electron-acceptor.

2.9.3 Effect of the protonated nucleobase on pseudorotamer populations

The protonation of a nucleobase in a nucleotide directly controls its hydrogen-bonding

capabilities, therefore the overall three-dimensional structure of oligonucleotides is dictated by the pH

of the medium. For instance, the local DNA triple helix formed354,355 by the binding of a natural

homopyrimidine oligonucleotide to a target DNA duplex is much more stable in the acidic solution

than at neutral pH, owing to the fact that Hoogsteen base-pairing with the third pyrimidine strand

requires that the cytosin-1-yl nucleobases be protonated. However, substitution of cytidine in the

Hoogsteen strand for 5-methylcytosine or 5-bromouridine allows to increase its affinity for the DNA

duplex under physiological pH356,357. pH-dependent conformational transitions and stabilities of C.A

and G.A mismatches in DNA358-362 have been extensively studied. It has also been suggested363,364

that although the secondary structure of the oligoDNA d(A+-G)10 is presumably helical, it is stabilized

not by stacking bases or hydrogen-bonding base pairs but instead by ionic bonds between positively

charged 2'-deoxyadenosine residues and distal negatively charged phosphates. The formation of

unusual parralel double-stranded DNA duplex365-367 or four-strand tetrads (the i-motif)368,369 with two

parallel stranded base-paired duplexes at low pH has also been experimentally evidenced.

The acid-base character of nucleobases in nucleosides and nucleotides varies widely165,247,370-

381: Whereas adenosine165 (pKa ≈ 3.5) and cytidine165 (pKa ≈ 4.2) can be easily protonated in the

acidic solution at N1 and N3, respectively, uridine165 (pKa ≈ 9.4), 5-fluoro-2'-deoxyuridine382-384 (pKa

≈ 7.8) and thymidine165 (pKa ≈ 9.9) are deprotonated at N3 in alkaline solution. Guanosine is either

protonated379 at N7 (pKa ≈ 1.9) in the acidic medium or deprotonated165 at N1 in the alkaline solution

(pKa ≈ 9.4). pKa values for C-nucleosides are known for formycin A385,386 (pKa = 4.4 (protonation at

N3) and 9.6 (deprotonation at N7)), formycin B387,388 (pKa = 8.8 (deprotonation at N1) and 10.4

(deprotonation at N7)), pseudoisocytidine389 (pKa = 3.7 (protonation at N1) and 9.0 (deprotonation at

N3)) and pseudouridine390 (pKa = 9.0 (mixed deprotonation at N1 and N3)). We have recently reported

for the first time34 the pKa values corresponding to protonation at N3 in 9-deazaadenosine (pKa = 6.0)

and formycin B (pKa = 1.3). Most pKa values of nucleobases in nucleos(t)ides are rather different from

the physiological pH, therefore one might conclude that pH-induced conformational transitions are not

likely to occur in DNA and RNA near physiological pH (≈ 6.8 - 7.3), unless the pKa of a particular

nucleobase changes drastically as a result of change of the microenvironment359,360,363,369,396-398. This

can be examplified by the pKa values of adenin-9-yl and cytosin-1-yl moieties, which are significantly

different at some oligonucleotides than those found in their monomers: For example, the pKa of

adenin-9-yl in the A25 residue (located close to the cleavage site in a lead-dependent ribozyme) is

6.5397,398 which is unusually high compared with the pKa of adenin-9-yl in adenosine (3.5). Clearly,



41

any slight deviation of the medium from the physiological pH may impose an ionic character on

nucleobases with standard pKa values. Additionally, transition or soft metal ions, by binding to the

nucleobase391-395 may also impose a change in the ionic character of the nucleobase, equivalent to the

effect of its protonation or deprotonation (depending upon which of the sites in a nucleobase a metal

ion prefers to bind!).

The affinity of a pyrimidine oligonucleotide399 cytidine residues for a target DNA duplex

increases by 10-fold when the pH is changed from 7.6 to 5.8, whereas the equilibrium binding constant

for another pyrimidine oligonucleotide in which 5-methylcytosin-1-yl nucleotides have been

incorporated increases by a factor 20 upon the same change in pH. An analysis of these data in terms of

two-state model revealed that the above pyrimidine oligonucleotides with cytidine and its 5-methyl

counterpart form triple helical structures with apparent pKas of 5.5 [(C+GC) triplets] and 5.7

[(m5C+GC) triplets], respectively. These pKa values are respectively 1.2 and 1.3 unit higher than those

known for the nucleoside counterparts, 2'-dC (pKa = 4.3) and 5-methyl-2'-dC (pKa = 4.4), respectively.

From the above studies397-399, it is clear that some specific adenin-9-yl and cytosin-1-yl nucleotides,

under certain folded structural states would form the protonated species at a pH close to the

physiological pH. Since the magnitude of the O4'-C1'-N1/9 anomeric effect is enhanced in the P

compared to the N state of the nucleobase in nucleosides (see the sections below), those residues are

more prone to take a N-type conformation, which in turn may dictate the local phosphate backbone to

adopt a specific conformation, which together may constitute a recognition element for a ligand

binding.

Owing to the key role played by protonation of the constituent nucleobases in the biological

function of DNA and RNA, it appeared necessary to obtain reliable estimates of the actual magnitude

of the anomeric effect in the N, P and D states of the nucleobases in nucleosides and nucleotides

(Sections 4 - 6).

As stated in Section 2.9, protonation of the nucleobase in various purine and pyrimidine

nucleosides and nucleotides, such as adenin-9-yl at N1 in 2'-O-methyladenosine318 or arabinoadenosine

derivatives198, guanin-9-yl at N7 in 2'-, 3'- and 5'-phosphates of guanosine313,317, cytosin-1-yl at N3 in

arabinocytidine or its methyl derivatives315 results in the shift of the N � S pseudorotational

equilibrium of the constituent pentofuranose moiety toward more N-type pseudorotamers (as

experimentally evidenced by the change in 3JHH coupling constants), in which the nucleobase adopts a

pseudoaxial orientation.

In six-membered rings, on the other hand, it has been shown that the preference of imidazolium

or pyridinium as anomeric groups in 2-substituted pyranose derivatives for equatorial positions is

greater than that of their neutral counterparts94,106-108. This has been attributed to the reverse anomeric

effect (Section 1.7). If the reverse anomeric effect were to play any role in the drive of the sugar

conformation in β-D-nucleosides at acidic pH, one should observe an increase in the population of S-

type pseudorotamers, in which the nucleobase adopts a pseudoequatorial orientation, in the acidic

compared to the neutral pH. Since the above qualitative studies198,313,315,317,318,400 as well as our

recent quantitative works30,32,36,37,44 (described in detail in Sections 4 - 6 and 8.8) show opposite

trends, i.e. a greater preference for N-type conformations in the P compared with the N state, one can

therefore conclude that no reverse anomeric effect operates in the protonated pentofuranosyl β-


42

nucleosides. Instead, the shift of the two-state N � S pseudorotational equilibrium in nucleosides

toward N-type conformers at acidic pH can be easily explained in terms of pD-dependent magnitude of

the anomeric effect318: As the nucleobase becomes protonated, a partial positive charge is created at

the glycosyl nitrogen. As a result, nO4' →σ∗C1'-N9 interactions become more efficient. Owing to the

strengthening of the anomeric effect, the tendency of the nucleobase to adopt pseudoaxial orientations

in N-type pseudorotamers is greater in protonated nucleosides than in their neutral counterparts.

Protonation of guanosine at N7 promotes delocalization of the lonepair of the glycosyl nitrogen

N9, making it slightly positively charged, which has a certain azine-type character, and therefore its

15N chemical shift resonates downfield by 6.7 ppm compared to N9 in neutral guanine moiety401.

Similarly, upon protonation of N3 in 1-methylimidazole, the delocalization of the N1 lonepair results

in its partial positive character, which promtes its 7.4 ppm downfield shift compared to the counterpart

in the neutral form402. The delocalization of the N9 lonepair in N1-protonated adenosine to stabilize

N1H+ species also results in partly positively charged N9, which resonates downfield by 6.6 ppm in

comparison with N9 in neutral adenine403. On the other hand, in cytidine the chemical shift of N1 is

hardly affected (the downfield shift is only 1.1 ppm404) upon protonation of cytosin-1-yl at N3.

Upon its deprotonation at N3 in 3'-O-methylarabinouridine405, uracil-1-yl tends to prefer more

pseudoequatorial orientations than in the N state, as reflected in the increase of the population of S-type

pseudorotamers. This can be attributed to the fact that deprotonation of uracil-1-yl reduces the ability

of the glycosyl nitrogen to be involved in nO4' →σ∗C1'-N9 interactions, i.e. a weakening of the

anomeric effect, simply because of the electron-rich character of the conjugate base of uracil moiety.

2.9.4 Effect of base-modifications on the stability of nucleic acids

The modification of the chemical nature of nucleobases affects the overall stability of

DNA/RNA heteroduplexes has been reviewed47,49. However, the influence of the O4'-C1'-N1/9

anomeric effect (Section 4) as a result of base modification, or its modulation by the change of the

aromatic character by the change of pH of the medium (i.e. protonation-deprotonation equilibrium) or

by its binding to any specific ligand (i.e. association-dissociation equilibrium) may affect the stability

of the duplexes, which has not been addressed hithertofore in the literature.

The stabilization of DNA/RNA duplexes (with respect to parent unmodified duplexes), as a

result of modification of some constituent nucleobase(s) in the oligodeoxyribonucleotide chain, may be

attributed to either of the following events: (i) Increased stacking interactions in the case of 5-propynyl

dU406, 5-(amino-ethyl-3-acrylimido) dU49,407, 7-halo-7-deaza408,409 and 7-propyne-7-deaza410 purines

modifications, or (ii) to the shielding of the negative phosphate charges by 5-amino-hexyl-substituted

pyrimidines411, or (iii) to the possibility to form an additional hydrogen-bond in the case of 2-

aminoadenosine412,413. On the other hand, the destabilization of the heteroduplex DNA/RNA upon

modification of the nucleobase has been explained either by (i) loss of hydrophobic interactions (upon

substitution of thymin-1-yl by uracil-1-yl)414, or (ii) the inability of the thymin-1-yl nucleobase to adopt

anti orientation around the glycosyl torsion when it is substituted at C6414, or (iii) to the loss of

hydrogen-bonding sites in O2 or O4 substituted thymin-1-yl415.

2.10 The gauche effects in α- and β-nucleosides



43

In addition to O4'-C1'-N1/9 stereoelectronic interactions and the associated steric effect of the

nucleobase, the gauche effect also plays an important role to explain the preferred conformation of the

pentofuranose moiety in all nucleosides and nucleotides. The origin and the energetics of the gauche

effect in simple 1,2-disubstituted ethanes have been discussed in Sections 1.12 - 1.14. In this section,

we only present the relevant 1,4-gauche interactions which are involved in the drive of the sugar

conformation in nucleosides and nucleotides.

2.10.1 2'-Deoxynucleosides

In α- and β-2'-deoxynucleosides, the gauche effect of [O4'-C4'-C3'-O3'] fragment stabilizes

pseudorotamers in which the O4'-C4' and O3'-C3' bonds are in a gauche orientation (i.e. W- and S-

type), over those where they are in a trans arrangement (i.e. E- and N-type)304,305. However, as

discussed in Section 2.3-2.5, W-type pseudorotamers are neither observed in the solid state nor in

solution, owing to the high energy penalty resulting from the eclipsed orientation of C2' and C3'

substituents, therefore the [O4'-C4'-C3'-O3'] gauche effect alone drives the sugar conformation of

nucleos(t)ides preferentially toward S-type pseudorotamers. In α-2'-deoxynucleosides, since O4'-C1'-

N1/9 stereoelectronic interactions cooperate with the [O4'-C4'-C3'-O3'] gauche effect, one expects that

the conformation of the pentofuranose sugar will be biased toward S-type conformations. The results of

our recent study37 on the thermodynamics of the N � S equilibrium in α-D/L-2'-deoxynucleosides are

in agreement with this simple reasoning (Section 5). However, we have also shown that the magnitude

of stereoelectronic interactions in α-series is in general considerably reduced than in the β-

counterparts. In β-dNs, the anomeric effect drives the two-state N � S pseudorotational equilibrium in

solution toward N-type forms, whereas the [O4'-C4'-C3'-O3'] gauche effect favours S-type geometries

(vide supra). The experimentally observed preference of β-dNs and β-2'-deoxy mono- and

oligonucleotides for S-type conformations165 therefore suggests that the [O4'-C4'-C3'-O3'] gauche

effect is the predominant factor controlling the conformation of the constituent sugar moieties,

prevailing over the counteracting anomeric effect. This is consistent with our recent estimates of the

magnitudes of anomeric and gauche effects in such systems20,30.

2.10.2 Ribonucleosides and nucleotides

In ribonucleosides and nucleotides, three additional gauche effects determine the preferred

orientations within [O2'-C2'-C1'-O4'], [O2'-C2'-C1'-N1/9] and [O2'-C2'-C3'-O3'] fragments. Both in α-

and β-ribonucleos(t)ides, among all possible pseudorotamers, W-type geometries are the most favoured

by the [O4'-C4'-C3'-O3'] and [O4'-C1'-C2'-O2'] gauche effects, since only in that situation both

fragments are in a gauche orientation. Conversely, E-type pseudorotamers are the most disfavoured,

owing to the trans orientation of O4'-C4' and O4'-C1' bonds with respect to C3'-O3' and C2'-O2',

respectively. However, W-type pseudorotamers are energetically penalized (vide supra ), and in

aqueous solution, the sugar moiety is involved in a two-state N � S equilibrium. The [O4'-C4'-C3'-

O3'] and [O4'-C1'-C2'-O2'] fragments are in a trans and a gauche orientation in N-type pseudorotamers,

respectively, whereas in the S-type counterparts, this opposite is true, therefore the [O4'-C4'-C3'-O3']

and [O4'-C1'-C2'-O2'] gauche effects cancel each other. The [O2'-C2'-C1'-N1/9] gauche effect drives

the pseudorotational equilibrium of the sugar moiety in β-ribonucleosides toward S-type


44

pseudorotamers, whereas in α-ribonucleosides C1'-N1/9 and C2'-O2' bonds are in gauche orientation

both in N- and S-type conformations. The [O4'-C4'-C5'-O5'] gauche effect is one of the factors (besides

the steric or hydrogen-bonding interactions between 4'-CH2OH group and the nucleobase) that controls

the preferred conformation across the γ torsion angle.

Several models have been proposed to account for the gauche effects within X-C-C-Y

fragments, X and Y being electronegative elements or groups (Section 1.14). A simple explaination is

based on σ → σ* interactions between best donor σ and best acceptor σ* orbitals as shown for 2'-

deoxynucleosides in Fig 10 (A & B). Possible contributions from bond-bending, or a combination of

through-space and through bonds interactions to the mechanism of the gauche effect can however not

be excluded (Section 1.14).

In all studies on the effect of 1,4-gauche interactions upon the sugar conformation in

nucleosides and nucleotides, the estimates obtained for the magnitudes of these gauche effects actually

correspond to overall strength, consisting of steric, stereoelectronic and electrostatic components.

2.11 The gauche effects of sugar substituents and the self-organization of DNA/RNA

2.11.1 Studies on nucleosides

1H- and 13C-NMR studies on N-nucleosides17,46,193,195-199,201-204,206,208,

213,216,219,222,226,235,416-423, and on C-nucleosides224,237 as well as theoretical works304,305 have

qualitatively shown that the preferred conformation of the pentofuranose moiety in nucleosides,

nucleotides and their derivatives is strongly affected by the electronic nature and relative configuration

of the substituents at C2' and C3' positions.

We, on the other hand, have shown in our recent quantitative studies, that as the nucleobase in

β-D-N- and C-nucleosides becomes electron-deficient in the protonated state, the strength of O4'-C1'-

N1/9 stereoelectronic interactions increases, whereas in the deprotonated state, the N9/1 becomes more

electron-rich resulting in the weakening of O4'-C1'-N1/9 stereoelectronic interactions. This has been

discussed in details in Sections 4-6.

Systematic analyses of the conformational preferences of 2'-deoxy-2'-substituted uridine199,235

and adenosine206,417,418 derivatives have allowed us to derive linear relationships between the

population of the N-type pseudorotamer and the electronegativity of the 2'-substituent: As the 2'-

substituent becomes more electronegative, the population of N-type pseudorotamers linearly increases

as a result of the increased strength of the 2'-substituent gauche effects. 2'-thionucleosides416 prefer

more S-type geometries than in 2'-deoxynucleosides. Similarly, the two-state N � S equilibrium in 2'-

methylthionucleosides421 is strongly (> 70 %) biased toward S-type conformations in CD3OD, and the

effect of 2'-SMe has been attributed both to its reduced electronegativity (i.e. resulting in weaker [S2'-

C2'-C1'-O4'] and [S2'-C2'-C1'-N1/9] gauche effects) and increased steric bulk (resulting in the

destabilization of N-type pseudorotamers).

The population of S-type pseudorotamers in 3'-deoxy-3'-substituted arabinofuranosyladenine201

is also dictacted by the tuning of the gauche effect of [X3'-C3'-C4'-O4'], which is in turn controlled by

the electronegativity of the 3'-substituent (X). Similarly, substitution of 3'-OH in β-D-2'-



45

deoxynucleosides by a more electronegative 3'-OPO2213, 3'-NO2

422 or 3'-F17,219,424 results in the

increased preference of the sugar moiety for S-type conformations. From a comparison of the

conformational properties of natural and 3'-modified thymidine dimers it has been found423 that S-type

pseudorotamers are more and more stabilized as the 3'-substituent is changed in the following order: 3'-

NH2 < 3'-N3 < 3'-O-H < 3'-O-P < 3'-O-S. The two-state N � S equilibrium in 2',3'-cis-fused furano-

and pyrrolidino-β-D-nucleosides is driven to S- (in case of C3'-O or C3'-N substitution) or N-type

sugars (for C2'-O or C2'-N).

Figure 10. Rationalization of the [X3'-C3'-

C4'-O4'] gauche effect in a 2',3'-dideoxy-3'-

substituted-β-D-nucleoside in terms σ σ* interactions between best donor (σC3'-H3') and

best acceptor (σ*C4'-O4') orbitals. The σ σ*

interactions stabilize the S- over N-type

pseudorotamers [Panel (A)], owing to the fact

that in the former the C3'H3' and C4'O4'

bonds are almost antiperiplanar, and therefore

the torsion angle between σC3'-H3' and σ*C4'-

O4' orbitals is much reduced in the S-type

sugar geometries (β ≈ -37°) compared to the

N-type counterparts (β ≈ -93°) [Panel (B)]

(Torsions angles have been calculated from

the values of the endocyclic torsion angle ν3

from ab initio optimized (at HF/6-31G*)

geometries of typical N- (PN = 21°; Ψm(N) =

35°) and S-type (PS = 143°; Ψm(S) = 34°)

pseudorotamers of 3'-fluorothymidine,

assuming simple trigonal symmetry. (C) The

extent of the energy stabilization of S-type

pseudorotamers through [X3'-C3'-C4'-O4']

gauche effect is proportional to the square of

the overlap between σC3'-H3' and σ*C4'-O4'

orbitals (i.e. S2) and inversely proportional

between the difference between their energies.

The sugar moiety prefers S- (or N) type conformers more in the furano than in the pyrrolidino

analog, owing to the stronger gauche effect of [O4'-C4'-C3'-O3'] (or [O4'-C1'-C2'-O2'] and [O2'-C2'-

C1'-N]) in the former compared with [O4'-C4'-C3'-N3'] (or [O4'-C1'-C2'-N2'] and [N2'-C2'-C1'-N]) in

the latter419,420. The gauche effect of the fluorine substituent, due to its high electronegativity, has a

profound stereoelectronic effect on the stereochemical orientation of the neighbouring groups, thereby

fluorine substituent governs the overall conformation of the sugar ring199,206,208,222.The sugar moieties

in 2'-α-fluoro-2',3'-β-D-dideoxyuridine and 3'-β-fluoro-2',3'-β-D-dideoxyuridine adopt exclusively N-

type conformations, owing to the cooperative drive of the [F2"(α)-C2'-C1'-O4'] and [F3'(β)-C3'-C4'-

O4'] gauche effects, respectively with the O4'-C1'-N1/9 anomeric effect. In contrast, as a result of

configuration-dependent gauche effect, the two-state N � S equilibrium in 2'-β-fluoro-2',3'-β-D-

dideoxyuridine and 3'-α-fluoro-2',3'-β-D-dideoxyuridine are strongly biased to the S-type conformers

because of the predominance of the [F2'(β)-C2'-C1'-O4'] and [F3"(α)-C3'-C4'-O4'] gauche effects,

respectively, over the anomeric effect226. 3'-methyl-thymidine in a TT dimer adopts preferentially N-

HOH2CO

HOH2C

Base

Base

H

H4'

C5'O4'

C2'

C2'

H4'

C5' H3'

X

C1'

C1'

σC3'H3'

α S2 / ΔE (σ*C4'O4' - σC3'H3')

σ*C4'O4'

Stabilization due to gauche effect (GE)

ΔE (σ*C4'O4' - σC3'H3')

O

O4'

X

X

σ*C4'O4'

σC3'H3'

σ*C4'O4'

σC3'H3'

H3'

XσC4'O4'

σ*C4'O4'

σ C3'H3'σC3'H3' β1 = -93ο

H3'

β1 = -37ο

σC4'O4'

σC4'O4'

σ*C4'O4'

where S = overlap between σ*C4'O4' and σC3'H3'GE

North, trans [O4'-C4'-C3'-X]

HF/6-31G* geometry; X = F HF/6-31G* geometry; X = F

South, gauche [O4'-C4'-C3'-X]

(A)

(B)

(C)


46

type conformations425, owing to the lack of [O3'-C3'-C4'-O4'] gauche effect, whereas 3'-OH in the

thymidine counterpart drives the sugar conformation toward S-type pseudorotamers.

2.11.2 Studies on oligonucleotides

The application of the gauche effect concept in the design of oligonucleotides with specific

conformational properties will be discussed in detail in Section 8.1.

3. Methods to quantitate stereoelectronic effects in nucleos(t)ides

In our laboratory, a new strategy has been recently elaborated to obtain reliable accurate

estimates for the magnitudes of the O4'-C1'-N1/9 anomeric effect and the gauche effects of [O4'-C4'-

C3'-O3'], [O4'-C1'-C2'-O2'] and [O2'-C2'-C1'-N1/9] fragments controlling the sugar conformation in

nucleosides and nucleotides14-45. Our method is based on two essential steps: (i) The initial calculation

of the thermodynamics (i.e. ∆H°, ∆S° and ∆G°) of the two-state N � S equilibrium in 12 - 83 (Section

3.1-3.7). (ii) We have subsequently correlated the experimental ∆H° values with the structural features

of 12 - 83 either through semi-quantitative regression analysis or pairwise comparisons (Section 3.10

and following).

3.1 Thermodynamics of the two-state N � S equilibrium

∆H°, ∆S° and ∆G° of the two-state N � S equilibria of abasic sugars 12 - 1620,37, α-D-2',3'-

dideoxynucleosides (α-D-ddNs) 17 - 2036, α-D-2'-deoxynucleosides (α-D-dNs) 21 - 2637, α-L-2'-

deoxynucleosides (α-L-dNs) 27 - 2937, β-D-2',3'-dideoxynucleosides (β-D-ddNs) 30 - 3620,36, β-D-2'-

deoxynucleosides (β-D-dNs) 37 - 4520,30, β-L-2'-deoxynucleosides (β-L-dNs) 46 - 4937, β-D-

ribonucleosides (β-D-rNs) 50 - 5520,30, β-D-ribo-C-nucleosides (β-D-C-rNs) 56 - 6224,25,32, β-D-3'-dA

6327,426, 3'-monophosphates of β-D-2'-deoxynucleosides (β-D-dNMPs) 64 - 6823, 3'-ethylphosphates

of β-D-2'-deoxynucleosides (β-D-dNMPEts) 69 - 7323, 3'-monophosphates of β-D-ribonucleosides (β-

D-rNMPs) 74 - 7828 and 3'-ethylphosphates of β-D-ribonucleosides (β-D-rNMPEts) 79 - 8328 have

been derived in two steps: (i) Vicinal proton-proton coupling constants (3JHH) extracted from their 1H-

NMR spectra have been initially translated into the parameters defining the geometry of the N- [PN and

Ψm(N)] and S-type [PS and Ψm(S)] sugar pseudorotamers as well as their mole fraction at the

equilibrium using the PSEUROT203,209,427 program. (ii) Van't Hoff analysis of the temperature-

dependent mole fractions of the N- and S-type conformations has subsequently allowed to derive ∆H°,

∆S° and ∆G° values of their N � S equilibria at each pD, and in the case of nucleosides exhibiting pD-

dependent conformational preferences, we have subsequently estimated the values of ∆H°, ∆S° and

∆G° in each of their P, N and deprotonated (D) states from a nonlinear fitting procedure.

3.2 1H-NMR spectra [temperature, pD, ligand-dependent spectra of nucleos(t)ides]

One-dimensional 1H-NMR spectra of 5 - 20 mM D2O solutions of 12 - 83 have been recorded

between 278 K and 358 K (in 5 K or 10 K steps) at one or several pDs within the 0.5 - 12.0 range at

600 MHz or 500 MHz using Bruker DRX 600, DRX 500 and AMX 500 spectrometers.

For abasic sugars 12 - 16, β-D-ddU (35), β-D-ddI (36) and mononucleotides 64 - 73 and 75 -

83, the spectra have been recorded at one (neutral) pD only. At neutral pD, 50% of all molecules of 3'-



47

monophosphates β-D-dNMPs 64 - 68 and β-D-rNMPs 74 - 78 are expected to have a net charge of -1

(pKa ≈ 1.5 for the first dissociation of the monophosphate moiety165) whereas the remaining 50% will

carry a charge of -2 (pKa ≈ 6.7 for its second dissociation165). However, the corresponding 3'-

ethylphosphates β-D-dNMPEts 69 - 73 and β-D-rNMPEts 79 - 83 carry a single net negative charge at

the neutral pD. In order to assess any possible influence of the ionization state of the phosphate moiety

upon the bias of the two-state N � S equilibrium in β-D-rNMPs compared with their β-D-rNMPEts

counterparts, we have also recorded 1H-NMR spectra of β-D-AMP (74) at several pDs in the range

from 6.5 to 8.4 (0.5 pD unit resolution). Under this pD range, only the ionization state of the phosphate

moiety is expected to change, not that of the constituent adenin-9-yl, since its pKa is ≈ 3.5 (section

2.8). We found that 3JHH and 3JHP remain constant throughout the whole 6.5 - 8.4 pD range, showing

that the ionization state of the 3'-monophosphate moiety does not affect the preferred conformation of

the pentofuranose moiety. For L-nucleosides 27 - 29 and 46 - 49 and for D-nucleosides 22, 23, 38 and

39, the 1H-NMR spectra have been recorded at two or three pDs (one in each of their P, N and/or D

states).

For all 1H-NMR experiments, 8 - 32 scans were typically recorded using a spectral width of ≈

10 ppm. The FIDs were processed using a slight gaussian apodization in order to enhance resolution,

and the final spectra consisted of 64 K datapoints. The pD values correspond actually to pH* values,

since they have been obtained simply by reading the values displayed on a pH meter (equipped with a

calomel electrode calibrated with pH 4 and 7 standard buffers in H2O) without any further correction

for the deuterium isotope effect. The pD of each sample has been adjusted by the simple addition of

microliter volumes of D2SO4 or NaOD solutions (typically 0.1 - 0.5 N). The small (< 0.1 ppm) and

irregular change in the chemical shifts of all resonances from 278 K to 358 K suggested that

aggregation was negligible. The 1H resonances were assigned through decoupling experiments and

one-dimensional nOe difference experiments. 1H-1H coupling constants (i.e. geminal 2JHH and vicinal

3JHH) have been verified with help of the DAISY428 simulation and iteration program package.

Identical values have been found at 1mM and 20 mM concentrations.

The errors on experimental 3JHH coupling constants (± 0.1 Hz in most cases, except for some

α-D-ddNs and β-D-ddNs at acidic pDs owing to the near isochronocity of some of H2', H2", H3' and

H3" multiplets36) have been estimated from the fit of simulated (or iterated) spectra to the

corresponding experimental 1H-NMR spectra. The influence of these errors upon the geometries and

relative populations of N- and S-type pseudorotamers engaged in the two-state equilibrium in aqueous

solution has been investigated during the pseudorotational analyses with a slightly modified version of

PSEUROT program (vide infra).

3.3 Pseudorotational analyses of 3JHH with PSEUROT and some practical hints

Some practical hints to perform dependable PSEUROT calculations on natural and unnatural

systems for quantitative estimation of the thermodynamics of intramolecular stereoelectronic effects

are perhaps important for new users. Before discussing in details the steps involved in performing

dependable PSEUROT calculations, we would like to urge a potential user of PSEUROT to consider

the following guidelines in addition to those described in the manual on PSEUROT 6.0 (Prof C.

Altona, Gorlaeus Laboratories, University of Leiden, 2300 Leiden, The Netherlands).


48

3.3.1 Incorporation of coupling constant errors in PSEUROT calculations

Measurements of experimental coupling constants are prone to errors of at least 0.1 Hz. In the

original PSEUROT program this is not taken into consideration. We have therefore modified version

5.4 of PSEUROT to accomodate such error by generating datasets of coupling constants (subsequently

used to perform individual calculations with PSEUROT) with a gaussian distribution around the

experimental values (see our website:

http://bioorgchem.boc.uu.se/WWW_secret/manual_pages/local/ps54ran.html). Readers are directed to

Section 3.6 for details.

3.3.2 Parameters to be fixed or optimized during PSEUROT calculations

PSEUROT allows to perform a multilinear fit of the experimental coupling constants to P and

Ψm of the N and S conformers and their mole fractions in either of the following ways:

(i) All parameters can be freely optimized during the calculation. This can be done only when

the number of observables (i.e. temperature-dependent coupling constants) is greater than the number

of unknowns (P and Ψm of the N and S conformers and mole fraction xN or xS of a pseudorotamer).

This is a very important consideration! It is also noteworthy that this is also the reason why

pseudorotational calculations on 2'- or 3'-dNs are more relibale than on the ribo counterparts.

(ii) In the case of two-state N �S equilibrium strongly biased (i.e. ≥ 70%) to either of the N- or

S-type pseudorotamers, it is advised to obtain as much information as possible on the geometry of the

major conformer, and also the evidence of why two-state equilibrium is a valid concept on the

unnatural compound in question. This can be achieved by performing a set of PSEUROT runs in which

P(minor) and Ψm(minor) of the minor pseudorotamer are constrained to different values in each

calculation. P(minor) and Ψm(minor) are chosen in such a way that over the whole set of calculations the

ensemble of P(minor) and Ψm(minor) values that one can reasonably expect to be accessible to the minor

pseudorotamer is surveyed. The hyperspace of "reasonable" geometries that can be adopted by the

minor pseudorotamer may be assessed basing on either a perusal of crystal structures of the compound

of interest or some of its derivatives, or on the results of ab initio calculations at a higher basis set such

HF/6-31G* (even semiemperical or molecular mechanics based geometry can show some guidelines

when it is not possible to perform ab intio calculations). Further hints can also be found in the manual

of PSEUROT version 6.0, p. 24, which can be obtained from Prof. Altona's group.

(iii) In contrast, when there is no clear preference for either N- or S-type pseudorotamers, a

PSEUROT calculation is performed in which Ψm(N) and Ψm(S) are fixed to an identical value during

each calculation, and this value is incremented in the following runs with PSEUROT in such a way that

all possible "reasonable" values are surveyed throughout the whole set of calculations.

(iv) In the case of conformationally constrained compounds, it is required that the coupling

constants show some variation as a function of temperature. If not, it is impossible to derive

thermodynamics of stereoelectronic effects for these compounds.

3.3.3 Electronegativity of the substituents on each HCCH fragment



49

In the earlier versions of PSEUROT (e.g. 3B), the generalized equation (referred to as the EOS

equation, Eq 8 in Section 3.4) was based on Huggins group electronegativities. In more recent versions,

the Karplus equations incorporate the values of substituent parameters λ429,430 to take into account the

effect of the electronegativity of the substituents on the HCCH fragment of the coupling protons, which

present several advantages over the Huggins electronegativities (see Section 3.4 for details). In a

situation where λ is not known for a particular substituent (which is very often the case in unnatural

nucleosides used in the antisense work, for instance!), it is advised to use as a starting point the value

for the chemical group (among the 50 presented in the article429) whose chemical nature is closest of

that of the group of interest in the unnatural nucleoside. An alternative would consist in performing a

series calculations in which hypothetical λ values (between the limiting values in the scale 0 to 1.4) are

successively used in order to assess the effect of the unknown λ value on the thermodynamics of the N

� S equilibrium. For compounds with a known pKa value, one can determine the pH-dependent

thermodynamics of the N � S equilibrium using a series of λ values, and figure out which one gives

the correct pKa value of the aglycone or of any other ionizeable substituent. In our recent studies on the

pD-dependent modulation of the thermodynamics of the N � S equilibrium in N- and C-nucleosides

(Sections 4 - 6) through tunable gauche and anomeric effects, it was necessary to assess the influence

of the change of the electron-density of the glycosyl nitrogen upon its λ value. Our unpublished results

(see Section 3.7(c) for details) show that the effect of a particular λ(N1/9) value chosen within the 0.0 -

1.4 range on ∆H° and ∆S° of the N � S pseudorotational equilibrium is not significant.

3.3.4 Priority rule to number the substituents on the HAC1C2HB fragment

The original convention431 proposed by Altona's group to differentiate the substituents on the

HAC1C2HB fragment according to their relative orientations with respect to the coupling protons HA

and HB is better shown under the form of Scheme 2 (see also p. 14 in the manual of PSEUROT version

6.0).

The substituent S1 on C1 is said to be "positive" (ζ = +1 in the Karplus equation) since the

projected valency angle between HA and S1, counting clockwise from HA, amounts to ≈ +120°.

Conversely, S2 is a "negative" substituent, since the projected valency angle between HA and S2,

counting anticlockwise from HA, amounts to ≈ -120° (ζ = +1 in the Karplus equation). Analogously, S3

and S4 are "positive" and "negative" substituents, respectively, owing to the fact that the projected

valency angles between HB and S3, on one hand, and between HB and S4, on the other (counting

clockwise and anticlockwise from HB, respectively) are ≈ +120° and ≈ -120°.

3.3.5 Translation of HCCH

torsion angles into endocyclic

torsion angles

In the second step of

PSEUROT (Step 4a in Scheme 3

and step 2 in Scheme 4), proton-

proton (HCCH) torsion angles are

translated into the corresponding

HA

(-) S2 S1 (+)

(+) S3 HB

S4 (-)

HB

(-) S4 S3 (+)

(+) S1 HA

S2 (-)

S2

S3

C2

S4

HB

C1

S1

HA

Φ

α1 = +120°α2 = -120°

α3 = +120°α4 = -120°

Φ

Scheme 2. Definition of "positive" and "negative" substituents on the HAC1C2HB fragment


50

endocyclic torsion angles (ν0 .... ν4) via linear relationships such as: Φ = Aνi + Bi (Eq 11 in Section

3.5). For natural pentofuranosyl nucleosides, A and B have been determined for the β-D-ribose, β-D-

2'-deoxyribose, β-D-arabinose, β-D-lyxose, β-D-xylose, α-D-lyxose and α-L-xylose configurations

from linear regressions based on torsion angles extracted from crystal structures174,182,205 (see also the

discussion in the manual of PSEUROT version 6.0). For other modified nucleosides, A and B pairs can

be derived from correlation plots of HCCH versus νi torsion angles as extracted from ab initio

optimized geometries (e.g. at HF/3-21G* or higher basis sets) of various pseudorotamers of the

nucleoside of interest with selected P values (for instance, 12 structures with 0° < P < 330° at

30° resolution and a common puckering amplitude). Care is to be taken so as to ensure that each

pseudorotamer optimized ab initio will have the desired phase angle value, i.e. by constraining the

values of two endocyclic torsions (e.g. ν0 and ν4) during the calculation.

3.3.6 General operational conditions for PSEUROT

In our conformational studies on nucleosides and nucleotides, 3JHH coupling constants have

been interpreted in terms of a two-state N � S equilibrium with help of the program

PSEUROT203,209,427. The validity of the two-state equilibrium has been experimentally evidenced by

the analysis of the distribution of sugar conformations in the X-ray crystal structures of nucleosides, as

well as by the results of NMR studies in solution both in our own lab and elsewhere32,36,37,174, 230,241-

244,245,248 (Section 2.3 and the following).

PSEUROT Program

Experimental 3JHH

Coupling Constants for

a nucleoside X

Karplus-Altona Equation

AH

, BH parameters

from ab initio

calculations or

crystal structure

Endocyclic Torsion Angle (ν0 - ν4)

Pseudorotational Concept

Phase Angle of Pseudorotation (P) and Maximum Puckering Amplitude (Ψm) and Equilibrium Populations for X

(5a)

(4a)

(3a)

X-ray Crystal Structure of X

Phase Angle of Pseudorotation (P) and Maximum Puckering Amplitude (Ψm)

(6)(7)

Φ(HCCH) torsion angles


(2)

Experimental 3JHH in

constrained nucleosides

Corresponding Φ(HCCH) torsions angles from X-ray

structures

A New 7 Parameter Karplus

type equation

(9)(10)


Phase Angle of Pseudorotation (P) and Maximum Puckering Amplitude (Ψm) and Equilibrium Populations

(3b)

(4b)

(5b)


Experimental 3JHH

Coupling Constants

(8)

(11) (12)

Ab Initio Calculations

on X

The two-state N S equilibrium is a valid model for

pseudorotational analyses of 3JHH

of the nucleoside X

(1)

Scheme 3: Iterative structure elucidation of nucleosides using NMR-PSEUROT203,209,427 analyses of 3JHH coupling

constants, ab initio calculations and comparisons. Ab initio calculations on aristeromycin (8), 2'-deoxyaristeromycin (9) and

3'-deoxyaristeromycin (10)41 allowed us to validate the two-state N � S equilibrium model (Steps 1 & 2) in such

compounds and to derive AH and BH values required for step 4a (see below). Three translation steps are involved in

PSEUROT: (i) Experimental 3JHH are first used to calculate ΦHCCH torsion angles (Step 3a) using Eqs 8a-9a; (ii) ΦHCCH

are used to calculate the endocyclic torsion angles (νi) (Step 4a). AH and BH are published174 for β-D-dNs/rNs. For α-D-

dNs, 120° was subtracted from AH and BH known for Φ1'2' (Φ1'2") in parent β-D-dNs. For α-/β-L-dNs, AH and BH are the



51

same (but with opposite signs) as those of α-/β-D counterparts234. For β-D-dNMPs/dNMPEts, we used the same values as

for β-D-dNs. For β-D-rNMPs/rNMPEts and β-D-C-rNs we assumed AH and BH identical to those of β-D-rNs. For β-D-

ddNs, AH and BH were derived from regression analysis of HCCH torsion angles versus the corresponding νi (taken from

crystal structures268,269,432-434). For α-D-ddNs, we have subtracted 120° from the values found for Φ1'2' (Φ1'2") for the

parent β-D-ddNs. (iii) PN, PS, Ψm(N), Ψm(S) and xS are finally calculated (Step 5a) from all νi values using Eq 6. For

aristeromycin, we found41 that the geometry of the cyclopentyl ring in the solid state (steps 6 and 7) is different (step 8)

from the solution state geometry derived from PSEUROT analyses (based on the original431 Karplus-Altona Eqs 8a-9a). To

verify whether this was due to a poor parametrization of Eqs 8a-9a for H2CCH2 fragments, we specifically reparametrized

them (Eq 8b-9b, steps 9 & 10) for carbocyclic nucleosides. PSEUROT analyses (steps 3b - 5b) based on Eqs 8b-9b

produced the same geometries (step 11) for the cyclopentyl ring in aristeromycin, its 2' and 3'-deoxy derivatives as the initial

analyses (steps 3a-5a), however the r.m.s. errors were reduced when Eqs 8b-9b were used.

Five parameters are necessary to define the position of the equilibrium at a certain temperature or pD:

The phase angles of the N- (PN) and S-type (PS) pseudorotamers and their respective puckering

amplitudes [Ψm(N) and Ψm(S)] as well as the mole fraction of one of the conformers (xN or xS). In the

case of 12 (10 experimental 3JHHs), of α-D-ddNs 17 - 20 and β-D-ddNs 30 - 36 (8 3JHHs) and of 13,

15 and 16 (7 3JHHs), the system is overdetermined and PN, PS, Ψm(S), Ψm(S) and xS can in principle

be estimated simultaneously from a single set of coupling constants. For α-D-dNs 21 - 26, α-L-dNs 27

- 29, β-D-dNs 37 - 45, β-L-dNs 46 - 49, β-D-dNMPs 64 - 68, β-D-dNMPEts 69 - 73 and β-D-3'-dA

(63), five 3JHHs are available at each temperature. Since there are as many unknowns as knowns, if the

calculation is performed without constraining one of the unknowns to some value estimated by another

method (for instance from X-


52

PSEUROT Program

Experimental 3JHH

Coupling Constants

Karplus-Altona Equation

Proton-Proton Torsion Angle

AH , BH parameters from ab initio calculations or crystal structure


Pseudorotational Concept

Phase Angle of Pseudorotation (P) and

Maximum Puckering Amplitude (Ψm)

and Equilibrium Populations

Experimental 3JHF

Coupling Constants

Proton-Fluorine Torsion Angle


3JHF and Corresponding Proton-Fluorine

Torsion Angles from Conformationally

Fixed Compounds

AF , BF parameters from ab initio calculation or crystal structure

Proton-Fluorine Torsion Angle

Limiting 3JHF

Coupling Constants

Extrapolation of experimental 3JHF to pure major conformer

New Karplus-Type Equation

for 3JHF Coupling Constants

(9)

(8)(6)

(7)

(5)

(10)

(3)

(2)

(1)

(4)

Scheme 4: The "PSEUROT+JHF" program39, as an extension of the original PSEUROT program for pseudorotational

analysis of 3JHF in combination with 3JHH coupling constants. We first performed pseudorotational analyses of 3JHH

coupling constants of monofluoronucleosides to find out P, Ψm and xS (at various temperatures) of the N and S forms

(Steps 1 - 3). From the plot of temperature-dependent xS as a function of temperature-dependent experimental 3JHF, we

derived limiting 3JHF values for the major N or S pseudorotamer of each nucleoside (step 5). The corresponding HCCF

torsion angles for all 3JHF coupling constants were calculated from the corresponding endocyclic torsion angles (step 4),

basing upon AF and BF values (Eq 11 applied to ΦHCCF torsion angles) derived from ab initio optimized geometries of our

monofluoronucleosides. A dataset of 57 pairs of the above limiting (3JHF , ΦHCCF) from our monofluoronucleosides and

from conformationally constrained cyclic fluorinated organic compounds (step 6), which were added in order to better

define the values of 3JHF for ΦHCCF around ± 90° and 180°, was subsequently used to parametrize a new Karplus-type

equation specifically for 3JHF (Step 7), which includes a term accounting for HCC and FCC bond angle changes. The

validity of Eq 8c was proven by the fact that pseudorotational analyses based on 3JHF alone (steps 8 - 10) of a set of other

monofluoro nucleosides yielded nearly the same values for P, Ψm of the N and S pseudorotamers as our initial analyses

which relied exclusively upon 3JHH data (Steps 1 - 3). Eq 8c was finally used in combination with the Karplus-Altona

equation430 to derive for the first time accurate estimates for P, Ψm of the N and S pseudorotamers and xS (at various

temperatures) for difluoronucleosides 84 - 87 (steps 1 - 3 and 8 - 10).

ray crystal structures or ab initio calculations), the accuracy of the results will strongly depend upon the

accuracy of the experimental 3JHH coupling constants. For all other compounds, only 3 (or 4 in the



53

case of abasic sugar 14) 3JHHs are available at each temperature and the iteration with PSEUROT is

only possible if at least two (or one) parameter(s) are (is) fixed.

Calculations with PSEUROT are based on three principal steps (Schemes 3 & 4): (i)

Translation of experimental 3JHH coupling constants into the corresponding proton-proton torsion

angles (Φ) [Step 3a in Scheme 3, step 1 in Scheme 4] with help of the generalised Karplus-Altona

equation. In some modified nucleosides it might be necessary to reparametrize this Karplus-Altona

equation to take account of the influence of a specific substituent on the bond length, bond angle,

through-space transmission effect (i.e. the Barfield effect) etc; (ii) ΦHCCH are themselves used to

calculate the corresponding endocyclic torsion angles of the pentofuranose moiety (ν0 ... ν4) (Step 4a

in Scheme 3, step 2 in Scheme 4); (iii) PN, PS, Ψm(N) and Ψm(S) of the N-type and S-type

pseudorotamers as well as xS are ultimately derived from all endocyclic torsion angles obtained in step

(ii) using the law of pseudorotation (Eq 6, Step 5a in Scheme 3, step 3 in Scheme 4).

3.4 Generalised Karplus-type equation

Karplus-type equations (Step 3a in Scheme 3, Step 1 in Scheme 4) allow to translate vicinal

coupling constants into the torsion angle between the coupling nuclei, and their use in conformational

analysis is well known215,217,435-437.

3.4.1 EOS Karplus-Altona equation for 3JHH

The generalised EOS equation (Eq 8a) developed by Altona's group431 describes the

dependence of a vicinal 3JHH coupling constant upon the corresponding proton-proton torsion angle,

the Electronegativity and relative Orientations of the Substituents in the H-C-C-H fragment (EOS).

3J

HH = P1 cos2Φ+ P2 cosΦ + P3 + ∑

−=

Δ

m1i

χi(g) [P4 + P5 cos2 (ζiΦ + P6 |∆χi(g)|)] ... Eq 8a

Φ designates the H-C-C-H torsion angle. The first three-terms in Eq 8a show the dependence of 3JHH

upon the H-C-C-H torsion angle alone, using the same formalism as that originally proposed by

Karplus et al438. The introduction of a phase-shifted cosine square function in the remaining terms

allows to take into consideration the influence of the electronegativity of the non-hydrogen substituents

on the H-C-C-H fragment and of their relative orientations with respect to the coupling protons. It is

assumed that the effects of m substituents are additive. The value of ζi (± 1) reflects the orientation of

the substituent i with respect to the coupling protons. Δχi(g) represents the difference between the

group electronegativities of the substituent i and of hydrogen, which is used as a reference, in the

Huggins scale439. Δχi(g) is calculated according to Eq 9a: Δχi (g) = Δχi(α) - P7 ∑−=

Δ

n1j

χj (β-substituent)

..... Eq 9a. Δχi(g) takes into account both the electronegativity of the α-substituent [Δχi(α)] and

the electron-withdrawing or donating effect of n β-substituents [Δχj (β−substituent)]. No term in Eq 8a

takes into account a possible effect of H-C-C bond angles and C-C bond distances440,441, or the

possible influence of through space orbital interactions 442-444. P1 - P7 have been optimized separately

for HCCH, H2CCH and H2CCH2 fragments using a set of 315 experimental 3JHH coupling constants

and Φ values (as derived from molecular mechanics calculations), of conformationally constrained six-


54

membered rings, which implies that the torsion angles are clustered into the gauche and trans regions.

Eq 8a allows to predict the experimental 3JHH with an accuracy (r.m.s.) of 0.48 Hz.

Recently, Eq 8a has been reparametrized430 using a set of 299 pairs of (3JHH, Φ). In that work,

Huggins Δχi(g) values (Eq 9a) were replaced by empirical substituent parameters (λ) and the overall

r.m.s. error between experimental and theoretical 3JHH values dropped down from 0.48 Hz to 0.36 Hz.

Additionally, it was found that a separate parametrization for HCCH, H2CCH and H2CCH2 fragments

was no longer necessary.

3.4.2 Reparametrized EOS equation for carbocyclic nucleosides

We have recently determined41 the solution conformations of aristeromycin (8), 2'-

deoxyaristeromycin (9) and 3'-deoxyaristeromycin (10) from pseudorotational analyses of 3JHH

coupling constants using PSEUROT version 3B203-203 program, which is based on the Karplus-Altona

EOS equation (Eq 8a). We found serious discrepancies between the X-ray crystal structure (P = 89°,

Ψm = 41°) of aristeromycin (8) and its structure calculated by NMR-PSEUROT conformational

analysis (35° < P [3/4

T - 0/4

T] < 65°, 35° < Ψm < 45°) (128° < P [1E] < 131°, 34° < Ψm < 36°), as

well as relatively high errors in the NMR-PSEUROT analyses for aristeromycin and its 2'-deoxy and

3'-deoxy derivatives [∆Jmax ≤ 1.6 Hz (i.e. maximal difference between experimental and PSEUROT-

calculated 3JHH) and r.m.s. error ≤ 0.7 Hz]. These observations have prompted us to reparametrize (Eq

8b) the Karplus equation implemented in the PSEUROT program by using torsion angles derived from

solid state geometries of conformationally constrained nucleosides and their corresponding

experimental 3JHH.

3JHH

= 13.41 cos2Φ -0.98 cosΦ + 1.37

+ ∑−=

Δ

m1i

χi(g) [0.20 - 2.26 cos2 (ζiΦ + 0.39 |∆χi(g)|)] ......Eq 8b

where, Δχi (g) = Δχi(α) + 0.071 ∑−=

Δ

n1j

χj (β-substituent) ..... Eq 9b

The χ2 value for the fitting of the experimental ΦHH vs 3JHH

data using Eq. 8b is 7.2 Hz2,

which corresponds to an r.m.s. error of 0.40 Hz. For comparison, the r.m.s. error of the original

Haasnoot-Altona's equation (Eq 8a) was in the range 0.36 - 0.51 Hz, depending upon the substitution

patterns of H-C-C-H fragments (0.48 Hz in the case where all H-C-C-H fragments are considered

together for a common set of parameters). With the help of our new Karplus-type equation [Eq. 8b],

the difference between the experimental and calculated 3JHH couplings, ∆Jmax, was below 0.5 Hz for

40 data points, and in the range from 0.8 -1.3 Hz for five data points. It should also be noted that in our

approach we have a common equation for all H-C-C-H fragments. Moreover, the small P7 value

indicates the reduced influence of β-substituents. Therefore, for practical reasons, β-substituents may

even be excluded when the determination of the electronegativities of the substituents is performed.

The results of the PSEUROT analyses performed with the standard Haasnoot-Altona Karplus

equation (EOS equation: Eq 8a) are also very comparable in terms of geometry with those based on our

reparametrized equation (Eq. 8b). Both series of PSEUROT analyses suggest that the predominant

conformation of the cyclopentane ring in carbocyclic nucleosides is defined by 128° < P < 140° for

aristeromycin (8), 105° < P < 116° for 2'-deoxyaristeromycin (9) and 118° < P < 127° for 3'-



55

deoxyaristeromycin (10), with the puckering amplitude in the range from 34° to 40°. However,

PSEUROT analyses based on our Karplus equation produced a smaller r.m.s. error by ≤ 0.14 Hz and

∆Jmax error by ≤ 0.5 Hz than those performed with the standard Haasnoot-Altona's equation.

3.4.3 Karplus equation for interpretation of 3JHF

In the case of difluoronucleosides 84 - 87, the conformation of the constituent sugar moiety

cannot be determined on the basis of the unique experimental 3JHH available, i.e. 3J3'4' . However, four

additional vicinal coupling constants, i.e. proton-fluorine couplings 3JHF, can also be extracted from

their 1H-NMR spectra. It was clear to us that these 3JHF could also be used to find out the preferred

conformation of the sugar moiety in 84 - 87 in solution through pseudorotational analyses, provided

that we have at our disposal an accurate Karplus-type equation to translate them into the corresponding

HCCF torsion angles. It is this observation that has prompted us to parametrize39 a new seven-term

Karplus equation specifically for 3JHF coupling constants using a dataset consisting of

monofluoronucleosides and conformationally constrained fluorinated organic compounds (Eq 8c) [see

Section 3.8 for details]. Eq 8c has been constructed basing upon Eq 8a, however two significant

improvements have been made in comparison with the original formalism: (i) We have used

λ substituent parameters to account for the electronegativity of the substituents on the H-C-C-F

fragment of interest, in view of Altona's latest work430 and (ii) we have incorporated a cosine squared

term that allows to reproduce the experimentally observed variation in 3JHF coupling constants as a

function of the HCC (aHCC) and HCF (aFCC) bond angle values.

3JHF = 40.61 cos2Φ - 4.22 cosΦ + 5.88 + Σ λi [-1.27 - 6.20 cos2(ξi Φ + 0.20 λi)]

- 3.72 [(aFCC + aHCC)/2 - 110]

....Eq. (8c)

Using Eq 8c, we have been able to elucidate the conformation of a series of mono and

difluorinated nucleosides using a combination 3JHH and 3JHF coupling constants (Section 3.8). Our

study has shown that the geometries of the N- and S-type pseudorotamers derived from calcualtions

with PSEUROT based on either type of coupling constants were nearly identical.

3.4.4 Refined Karplus equation for 3JHH based on Fourier formalism

The formalism proposed in Eq 8a suffers itself from the following limitations. (i) It implies a

strict additivity of the effects of the substituents, in spite of the fact that linear correlations between the

value of 3JHH and the sum of the substituent electronegativities for monosubstituted ethanes break

down for 1,1-disubstituted ethanes carrying highly electronegative substituents. (ii) It poorly

reproduces the small experimental 3JHH values (< 1.2 Hz) for torsion angles in the ≈ ± 90° range. As a

result, Eq 8a has only been incorporated in the earlier versions of the PSEUROT program (version 3B).

In more recent versions (including version 5.4427, which has been used to perform the pseudorotational

analyses for 12 - 83), the dependency of 3JHH upon the HCCH torsion angle Φ and the nature of the

substituents is formulated as a truncated Fourier series429,445,446 in Φ (up to 3Φ) with coefficients

expanded as a Taylor series in the empirical substituent parameter λ, as shown in Eq 10a-b. 3J

HH = C0+ C1 cosΦ + C2 cos2Φ + C3 cos3Φ + S2 sin2Φ .... Eq 10a

where C0 - C3 and S2 are calculated according to Eq 10b, as follows:


56

C0 = 7.01 - 0.58 Σi λi - 0.24 (λ1λ2 + λ3λ4); C1 = -1.08; C2 = 6.54 - 0.82 Σi λi + 0.20 (λ1λ4 + λ2λ3);

C3 = -0.49; S3 = 0.68 Σi (ζiλ2i) ... Eq 10b

The new formalism is clearly advantageous: Terms reflecting the individual influence of each

substituent are still included, as in Eq 8, however new λ cross-terms are also introduced to take into

account possible interdependent substituent effects. Additionally, the influence of β-substituents upon 3JHH is incorporated in the value of empirically optimized λ parameters. The λ scale has been

developed429 through a least-squares fitting procedure, basing upon the change in the value of vicinal

3JHH as a function of the substitution pattern in mono- and 1,1-disubstituted ethanes, using two

references, i.e. λ(H) = 0.0 and λ(OR) = 1.4.

During the pseudorotational analyses performed on experimental 3JHH for 12 - 83, the

following λ values429,446 for the substituents at C1' - C4' have been used: λ(H) = 0.0; λ(C1') = 0.62 in

nucleosides or 0.67 in abasic sugars; λ(C2' or C3' deoxy) = 0.67; λ(C2' or C3' ribo) = 0.62; λ(C4') =

0.62; λ(C5') = 0.68; λ(O4') = 1.27; λ(OH) = 1.26; λ(OMe) = 1.27; λ(OPO3H-/OPO3Et-) = 1.27; λ(C-

aglycone at C1') = 0.45. For the glycosyl nitrogen of the nucleobase in nucleosides, we have used λ =

0.58 at all pD values.

3.5 Translation of experimental 3JHH and 3JHF into pseudorotational parameters

Proton-proton torsion angles (Φ) derived from step 3a in Scheme 3 and Step 1 in Scheme 4 can

in turn be used to calculate endocyclic torsion angles (ν0 .... ν4) of the pentofuranose moiety in 12 - 83,

using simple linear relationships (Eq 11, Step 4a in Scheme 3 and Step 2 in Scheme 4): Φ = Aνi + Bi

.... Eq 11

In the case of β-D-dNs and β-D-rNs, the values of A and B parameters, taken from the

literature174, have been determined from linear regressions, basing on torsion angles extracted from the

crystal structures of nucleosides. For α-D-dNs, we have subtracted 120° from the values published for

Φ1'2' (and Φ1'2") in the parent β-D-dNs, whereas for α- and β-L-dNs, A and B have been obtained by

reversing the signs of both A and B values used for their α- and β-D counterparts, respectively, as

suggested by Altona et al234. For β-D-dNMPs and β-D-dNMPEts, we have used the same A and B

values as for β-D-dNs, whereas A and B for β-D-rNMPs, β-D-C-rNs and β-D-rNMPEts have been

assumed identical to those known for β-D-rNs. For β-D-ddNs, A and B values have been determined

from regression analysis of some proton-proton torsion angles Φ as a function of the corresponding

νi values. Φ and νi values have been extracted from the crystal structures of β-D-ddA268, β-D-ddC269,

β-D-ddT432, β-D-ddU433 and β-D-2',3'-dideoxyribavirin434. Finally, for α-D-ddNs, we have subtracted

120° from the values found for Φ1'2' (Φ1'2") for the parent β-D-ddNs. Step 3 in PSEUROT translate the

values of the endocyclic torsion angles νi found in Step 2 into PN, Ψm(N) of the N-type pseudorotamer

and PS, Ψm(S) of the S-type sugar via the law of pseudorotation (Eq 6a).

3.6 Principle of iterations with PSEUROT

PSEUROT iterates the values of PN, Ψm(N), PS, Ψm(S), and of the mole fractions of the

conformers (xS) using a Netwon-Raphson minimization procedure, in such a way that a best fit is

obtained between experimental 3JHH and those back-calculated using the best fit PN,S, Ψm(N,S) and xS



57

values. Initially, PSEUROT calculates theoretical 3JHH values using the values of P and Ψm of both N-

and S-type pseudorotamers and their respective populations defined in the user's input. The

experimental and calculated 3JHH are compared. In the following iteration steps, random changes are

made in P and Ψm values, and/or in the populations, depending upon the user's input, which specifies

the parameters to be optimized or constrained during the calculation. The discrepancy between

experimental and calculated 3JHH is monitored and optimized during the iteration procedure. When the

best fit is found, the optimal PN, Ψm(N), PS and Ψm(S) of the N- and S-type conformers and their

relative populations are printed out, together with the error analysis, which shows both individual

differences between experimental and calculated 3JHH as well as an overall root mean square error

(r.m.s.).

Incorporation of the error in experimental 3JHH during the PSEUROT calculation: We have

modified the original version 5.4 of the PSEUROT program 203,209,427 to assess the propagation of

errors in experimental 3JHH throughout the calculations and during the subsequent treatment of the

results. Our modified program32 retains all features of the original version 5.4, all changes are

additions. The estimated error, expressed as standard deviation (σ), of each 3JHH and the desired

number of sets of randomly varied 3JHH to be generated and subsequently analyzed by pseudorotational

analyses, are included in the input file. Typically, for each set of contrained P and Ψm values (vide

infra), 1000 data sets are generated and individually analyzed. Each generated dataset contains

randomly varied 3JHHs but over all data sets, each 3JHH is normally distributed around its experimental

value with the given σ. The output from our modified program consists of statistical data [average, σ

and skew of the calculated geometrical parameters, of the mole fractions and of the "randomized" 3JHHs] and of the results from all the individual pseudorotational analyses (the calculated P and Ψm

values and mole fractions). We have discarded results which fall outside given ranges for: (i) The

individual errors between experimental and calculated coupling constants, i.e. Jcalc-Jexp, (ii) the root

mean square error (r.m.s.) in 3JHH, (iii) PN, PS, Ψm(N) and Ψm(S). PN, PS, Ψm(N) and Ψm(S) values

are considered as "reasonable" when they are within the ranges found for the sugar moieties in the

crystal structures of nucleosides and nucleotides, i.e. typically174: -40° < PN < 40°, 120° < PS < 200°,

30° (or 25° for α-nucleosides175) < Ψm(N) < 45° and 30° (or 25° for α-nucleosides175) < Ψm(S) < 45°.

Typically, Jcalc-Jexp and r.m.s. in 3JHH are considered as acceptable when they do not exceed Jcalc-Jexp

and r.m.s., respectively, of the best fit analyses by more than 0.1 Hz.

Chatt

opadhya

ya e

t al,

"S

tere

oel

ectr

onic

Eff

ects

in N

ucl

eosi

des

& N

ucl

eoti

des

and t

hei

r S

truct

ura

l Im

pli

cati

ons"

,

Dep

t of

Bio

org

anic

Chem

istr

y, B

ox 5

81, U

ppsa

la U

niv

ersi

ty, S

-75123 U

ppsa

la, S

wed

en, V

er 1

60205 j

yoti

@boc.

uu.s

e

58

Tab

le 2.

∆H° a

nd -

T∆

S° c

ontr

ibuti

on

s (k

Jmol-

1)a

-c t

o ∆

G2

98 o

f th

e tw

o-s

tate

N �

S p

seudoro

tati

onal

equil

ibri

um

in a

bas

ic s

ugar

s 12 -

16 a

nd

nucl

eosi

des

17 -

63 w

ith f

ull

y p

roto

nat

ed, neu

tral

and f

ull

y d

epro

tonat

ed n

ucl

eobas

es a

nd t

he

corr

espondin

g p

Ka

val

ues

d,e

Com

pound

Full

y p

roto

nat

ed n

ucl

eobas

e pK

a fr

om

: N

eutr

al n

ucl

eobas

e pK

afro

m:

Full

y d

epro

tonat

ed n

ucl

eobas

e

Δ

HP;�

-Τ

ΔS

P;�

Δ

GP;2

98

∆G

° d

δ

1H

e

ΔH

�

-ΤΔ

S N

;� Δ

G N

;29

8

∆G

° d

δ

1H

e

ΔH

D;�

-Τ

ΔS

D;�

ΔG

D;2

98

12

-

- -

- -

0.4

(0

.3)

-0.3

(0

.3)

0.1

(0

.4)

- -

- -

-

13

-

- -

- -

-4.1

(0

.3)

1.5

(0

.3)

-2.6

(0

.4)

- -

- -

-

14

-

- -

- -

0.4

(0

.1)

1.5

(1

.0)

1.9

(1

.0)

- -

- -

-

15

-

- -

- -

-4.6

(0

.4)

1.2

(0

.3)

-3.4

(0

.5)

- -

- -

-

16

-

- -

- -

-4.2

(0

.4)

0.9

(0

.3)

-3.3

(0

.5)

- -

- -

-

α-D

-dd

A (

17

) -1

.7 (

0.1

) 0

.2 (

0.1

) -1

.5 (

0.1

) -

3.7

-1

.7 (

0.1

) 0

.2 (

0.1

) -1

.5 (

0.1

) -

- -

- -

α-D

-dd

G (

18

)

-8.7

6

.0

-2.7

2

.7

2.7

-0

.4 (

0.2

) -0

.9 (

0.2

) -1

.2 (

0.1

) -

9.7

-0

.4 (

0.2

) -0

.9 (

0.2

) -1

.2 (

0.1

)

α-D

-dd

C (

19

)

-2.9

(0

.2)

1.4

(0

.2)

-1.5

(0

.1)

- 4

.1

-2.9

(0

.2)

1.4

(0

.2)

-1.5

(0

.1)

- -

- -

-

α-D

-dd

T (

20

)

- -

- -

- -1

.1 (

0.1

) 0

.6 (

0.1

) -0

.5 (

0.1

) 1

0.1

9

.7

-0.5

(0

.1)

0.3

(0

.1)

-0.2

(0

.1)

α-D

-dA

(2

1)

-5

.0 (

0.4

) 2

.1 (

0.3

) -2

.8 (

0.1

) -

3.6

-5

.0 (

0.4

) 2

.1 (

0.3

) -2

.8 (

0.1

) -

- -

- -

3'-O

Me-

α-D

-dA

(2

2)

-5.8

(1

.5)

0.8

(1

.5)

-4.9

(0

.3)

- -

-6.4

(0

.8)

1.9

(0

.7)

-4.5

(0

.3)

- -

- -

-

3',5

'-d

iOM

e-α

-D-d

A (

23

) -5

.3 (

1.3

) 0

.8 (

1.2

) -4

.7 (

0.3

) -

3.8

-5

.8 (

0.8

) 1

.6 (

0.7

) -4

.1 (

0.3

) -

- -

- -

α-D

-dG

(2

4)

-10

.7 (

2.0

) 6

.4 (

2.0

) -4

.4 (

0.3

) 2

.6

2.5

-4

.5 (

0.5

) 2

.7 (

0.5

) -1

.9 (

0.1

) -

9.5

-3

.4 (

0.5

) 1

.4 (

0.5

) -1

.9 (

0.1

)

α-D

-dC

(2

5)

-7.1

(0

.3)

2.5

(0

.8)

-4.3

(0

.2)

4.1

4

.2

-7.1

(0

.3)

4.0

(0

.5)

-3.1

(0

.2)

- -

- -

-

α-D

-T (

26

) -

- -

- -

-4.0

(0

.2)

2.0

(0

.5)

-2.1

(0

.2)

9.8

9

.8

-4.0

(0

.2)

2.7

(0

.4)

-1.1

(0

.1)

α-L

-dA

(2

7)

-5

.5 (

0.9

) 2

.7 (

0.9

) -2

.8 (

0.3

) -

- -6

.0 (

0.5

) 3

.2 (

0.6

) -2

.9 (

0.4

) -

- -

- -

α-L

-dC

(2

8)

-7.3

(0

.6)

3.0

(0

.6)

-4.2

(0

.2)

- -

-7.2

(0

.5)

4.1

(0

.6)

-3.1

(0

.2)

- -

- -

-

α-L

-T (

29

) -

- -

- -

-4.2

(0

.4)

2.1

(0

.5)

-2.1

(0

.2)

- -

-3.4

(0

.2)

2.5

(0

.4)

-1.0

(0

.1)

β-D

-dd

A (

30

) 9

.2 (

0.1

) -5

.0 (

0.1

)

4.1

(0

.1)

3.6

3

.8

3.5

(0

.1)

-0.9

(0

.1)

2.6

(0

.1)

- -

- -

-

Chatt

opadhya

ya e

t al,

"S

tere

oel

ectr

onic

Eff

ects

in N

ucl

eosi

des

& N

ucl

eoti

des

and t

hei

r S

truct

ura

l Im

pli

cati

ons"

,

Dep

t of

Bio

org

anic

Chem

istr

y, B

ox 5

81, U

ppsa

la U

niv

ersi

ty, S

-75123 U

ppsa

la, S

wed

en, V

er 1

60205 j

yoti

@boc.

uu.s

e

59

Tab

le 2

(C

onti

nued

)

Com

pound

Full

y p

roto

nat

ed n

ucl

eobas

e pK

a fr

om

: N

eutr

al n

ucl

eobas

e pK

afro

m:

Full

y d

epro

tonat

ed n

ucl

eobas

e

Δ

HP;�

-Τ

ΔS

P;�

Δ

GP;2

98

∆G

° d

δ

1H

e

ΔH

�

-ΤΔ

S N

;� Δ

G N

;29

8

∆G

° d

δ

1H

e

ΔH

D;�

-Τ

ΔS

D;�

ΔG

D;2

98

β-D

-dd

G (31

) 2

3.6

-1

7.1

6

.7

2.5

2

.5

3.4

(0

.4)

-0.3

(0

.2)

2.9

(0

.1)

9.6

9

.6

1.4

(0

.1)

0.5

(0

.1)

1.8

(0

.1)

5'-O

Me-β

-D-d

dG

(32

) 2

3.4

-1

6.8

6

.8

2.5

2

.5

4.3

(0

.5)

-0.6

(0

.5)

3.6

(0

.4)

9.7

9

.4

2.3

(0

.1)

1.0

(0

.1)

3.3

(0

.1)

β-D

-dd

C (33

) 9

.6 (

0.2

) -4

.9 (

0.1

) 4

.6 (

0.1

) 4

.2

4.3

6

.6 (

0.1

) -3

.0 (

0.1

) 3

.5 (

0.1

) -

- -

- -

β-D

-dd

T (34

) -

- -

- -

5.4

(0

.1)

-2.2

(0

.1)

3.2

(0

.1)

9.8

9

.9

3.5

(0

.1)

-1.1

(0

.1)

2.4

(0

.1)

β-D

-dd

U (35

) -

- -

- -

5.7

(0

.3)

-2.1

(0

.4)

3.6

(0

.5)

- -

- -

-

β-D

-dd

I (36

) -

- -

- -

6.2

(0

.2)

-2.9

(0

.3)

3.3

(0

.4)

- -

- -

-

β-D

-dA

(37

) -0

.7 (

0.1

) -0

.4 (

0.1

) -1

.1 (

0.1

) 3

.5

3.6

-3

.9 (

0.2

) 1

.8 (

0.2

) -2

.1 (

0.3

) -

- -

- -

3'-O

Me-β

-D-d

A (38

) -1

.4 (

0.6

) -0

.7 (

0.6

) -2

.2 (

0.2

) -

- -4

.6 (

0.4

) 1

.4 (

0.4

) -3

.2 (

0.2

) -

- -

- -

3',5

'-d

iOM

e-β

-D-d

A (39

) -1

.0 (

0.6

) -0

.7 (

0.6

) -1

.7 (

0.2

) -

3.5

-3

.6 (

0.4

) 1

.5 (

0.4

) -2

.1 (

0.2

) -

- -

- -

β-D

-dIm

b (40

) 0

.1 (

0.1

) -0

.2 (

0.1

) -0

.1 (

0.1

) 6

.0

6.0

-2

.2 (

0.1

) 0

.8 (

0.1

) -1

.4 (

0.1

) -

- -

- -

β-D

-dG

(41

) 2

.1 (

0.1

) -2

.2 (

0.2

) -0

.1 (

0.2

) 2

.3

2.2

-2

.8 (

0.2

) 1

.1 (

0.2

) -1

.7 (

0.3

) 9

.5

9.5

-4

.9 (

0.2

) 2

.2 (

0.2

) -2

.7 (

0.3

)

β-D

-dC

(42

) 0

.0 (

0.1

) -0

.8 (

0.1

) -0

.8 (

0.1

) 4

.2

4.3

-0

.7 (

0.1

) -0

.5 (

0.1

) -1

.3 (

0.1

) -

- -

- -

β-D

-T (43

) -

- -

- -

-1.4

(0

.2)

0.1

(0

.1)

-1.3

(0

.2)

9.7

9

.8

-1.9

(0

.2)

0.3

(0

.2)

-1.6

(0

.3)

β-D

-dU

(44

) -

- -

- -

-0.6

(0

.2)

-0.4

(0

.1)

-1.1

(0

.2)

9.5

9

.4

-1.3

(0

.2)

-0.1

(0

.3)

-1.4

(0

.4)

5-F

-β-D

-dU

(45

) -

- -

- -

-0.8

(0

.1)

-0.4

(0

.1)

-1.2

(0

.1)

7.8

7

.7

-1.1

(0

.1)

-0.4

(0

.1)

-1.5

(0

.1)

β-L

-dA

(46

) -1

.2 (

0.6

) 0

.1 (

0.6

) -1

.1 (

0.1

) -

- -3

.9 (

0.3

) 1

.8 (

0.4

) -2

.2 (

0.2

) -

- -

- -

β-L

-dG

(47

) 1

.3 (

0.8

) -1

.6 (

0.8

) -0

.3 (

0.1

) -

- -2

.7 (

0.3

) 0

.9 (

0.4

) -1

.8 (

0.2

) -

- -4

.3 (

0.4

) 1

.7 (

0.4

) -2

.6 (

0.2

)

β-L

-dC

(48

) -0

.2 (

0.3

) -0

.6 (

0.3

) -0

.8 (

0.1

) -

- -0

.7 (

0.3

) -0

.5 (

0.3

) -1

.2 (

0.1

) -

- -

- -

β-L

-T (49

) -

- -

- -

-1.2

(1

.1)

0.0

(1

.0)

-1.2

(1

.5)

- -

-2.3

(0

.7)

0.6

(0

.6)

-1.7

(0

.9)

β-D

-A (50

) -0

.2 (

0.1

) -0

.4 (

0.2

) -0

.5 (

0.2

) 3

.5

3.5

-4

.4 (

0.2

) 2

.6 (

0.1

) -1

.8 (

0.2

) -

- -

- -

β-D

-G (51

) 5

.4 (

0.2

) -4

.2 (

0.5

) 1

.5 (

0.5

) 2

.1

2.1

-3

.3 (

0.2

) 1

.8 (

0.2

) -1

.5 (

0.3

) 9

.5

9.6

-7

.6 (

0.5

) 4

.8 (

0.2

) -2

.8 (

0.5

)

Chatt

opadhya

ya e

t al,

"S

tere

oel

ectr

onic

Eff

ects

in N

ucl

eosi

des

& N

ucl

eoti

des

and t

hei

r S

truct

ura

l Im

pli

cati

ons"

,

Dep

t of

Bio

org

anic

Chem

istr

y, B

ox 5

81, U

ppsa

la U

niv

ersi

ty, S

-75123 U

ppsa

la, S

wed

en, V

er 1

60205 j

yoti

@boc.

uu.s

e

60

a T

he

erro

rs a

re s

ho

wn i

n p

aren

thes

es.

-T

ΔS

° a

nd

ΔG

°ar

e at

29

8 K

exce

pt

for

17

, 1

8,

30

-3

2 (

28

8 K

) o

win

g t

o d

eco

mp

osi

tio

n a

t ac

idic

pD

36,3

7.

b

∆H

°,

-TΔ

S° a

nd

ΔG

° a

re f

rom

ref.

20

(fo

r 1

2 -

14

), r

ef.

37

(1

5,

16

, 2

1 -

29

, 3

7 -

39

, 4

6 -

49

), r

ef.

36

(1

7 -

20

, 3

0 -

34

), r

ef.4

47

(3

5),

ref

. 3

0 (

40

- 4

5,

50

- 5

5),

ref

. 3

2 (

56

- 6

2),

ref

. 2

7(a

nd

our

unp

ub

lish

ed r

esult

fo

r

acid

ic p

Ds)

fo

r 6

3.

Eal

ier

esti

mat

es f

or

30

- 3

4 i

n t

he

N s

tate

wer

e p

ub

lish

ed20.

The

larg

est

dif

fere

nce

bet

wee

n t

hem

and

the

val

ues

in t

able

1 (

fro

m r

ef.

36

) is

1.4

kJ/

mo

l fo

r Δ

HN

;�

of

33

, w

hic

h i

s nea

rly w

ithin

the

sum

of

the

erro

rs o

f th

e es

tim

ates

. c

The

pla

teau

s in

the

P,

N a

nd

D s

tate

s fo

r ∆

H°,

-TΔ

S° a

nd

ΔG

° o

f 1

8,

20

, 2

4 -

26

, 3

0 -

34

, 3

7,

40

- 4

5,

50

- 6

1

and

63

res

ult

fro

m t

he

fit

of

exp

erim

enta

l p

D-d

epen

den

t val

ues

to

Eq

12

. F

or

22

, 2

3,

27

- 2

9,

38

, 3

9,

46

- 4

9,

the

erro

rs a

re t

ho

se o

f in

div

idual

ΔH

°,

-TΔ

S°

and

ΔG

° a

t a

cert

ain p

D

(ref

. 3

7).

d T

hes

e p

Kas

wer

e ca

lcula

ted

fro

m p

lots

of

exp

erim

enta

l ∆

G°

ver

sus

pD

and

fro

m H

ill

plo

ts o

f p

D v

ersu

s lo

g(∆

∆G

° tot -

∆∆

G°

/ ∆

∆G

°) (

refs

. 3

0,

32

-37

). F

or

18

, 3

1 a

nd

32

at a

cid

ic p

Ds,

the

pK

a c

alcu

late

d f

rom

pD

-dep

end

ent

1H

chem

ical

shif

ts (

col.

6)

was

use

d a

s a

const

rain

t in

the

det

erm

inat

ion o

f ∆

H°,

-TΔ

S° a

nd

ΔG

° i

n P

, N

and

D s

tate

s.

e T

hes

e

(aver

age)

pK

a

val

ues

co

resp

ond

to

the

pD

s at

the

infl

ecti

on p

oin

t(s)

of

plo

ts o

f 1H

chem

ical

shif

ts v

ersu

s p

D (

29

8 K

fo

r 1

7 -

20

, 2

88

K f

or

30

- 3

4)

and

fro

m t

he

corr

esp

ond

ing H

ill

plo

ts.

Com

pound

Full

y p

roto

nat

ed n

ucl

eobas

e pK

a fr

om

: N

eutr

al n

ucl

eobas

e pK

afro

m:

Full

y d

epro

tonat

ed n

ucl

eobas

e

Δ

HP;�

-Τ

ΔS

P;�

Δ

GP;2

98

∆G

° d

δ

1H

e

ΔH

�

-ΤΔ

S N

;� Δ

G N

;29

8

∆G

° d

δ

1H

e

ΔH

D;�

-Τ

ΔS

D;�

ΔG

D;2

98

β-D

-C (

52

) 5

.2 (

0.2

) -3

.3 (

0.4

) 1

.9 (

0.4

) 4

.0

4.1

2

.3 (

0.1

) -0

.8 (

0.5

) 1

.5 (

0.5

) -

- -

- -

β-D

-rT

(5

3)

- -

- -

- 1

.3 (

0.1

) -1

.4 (

0.3

) -0

.1 (

0.3

) 9

.9

9.8

-0

.2 (

0.1

) -0

.2 (

0.3

) -0

.4 (

0.3

)

β-D

-U (

54

) -

- -

- -

2.0

(0

.2)

-1.7

(0

.3)

0.3

(0

.4)

9.6

9

.4

0.3

(0

.1)

-0.2

(0

.1)

0.1

(0

.1)

5-F

−β

-D-U

(5

5)

- -

- -

- 2

.3 (

0.1

) -1

.8 (

0.2

) 0

.5 (

0.3

) 8

.0

7.6

0

.8 (

0.1

) -0

.6 (

0.3

) 0

.2 (

0.3

)

Fo

rmyci

n B

(5

6)

-0.5

(0

.1)

-0.8

(0

.1)

-1.3

(0

.1)

1.4

1

.3

-8.1

(0

.1)

4.8

(0

.1)

-3.3

(0

.1)

8.9

8

.8

-8.8

(0

.1)

5.3

(0

.1)

-3.5

(0

.1)

Fo

rmyci

n A

(5

7)

-2.4

(0

.1)

0.4

(0

.1)

-2.0

(0

.1)

4.5

4

.4

-8.1

(0

.5)

4.7

(0

.4)

-3.4

(0

.1)

- 9

.5

-8.1

(0

.5)

4.7

(0

.4)

-3.4

(0

.1)

9-d

eaza

-A (

58

) -7

.4 (

0.2

) 3

.7 (

0.3

) -3

.6 (

0.1

) 5

.9

6.0

-1

4.2

(0

.4)

9.1

(0

.4)

-5.0

(0

.1)

- -

- -

-

Ψ-i

soC

(5

9)

4.2

(0

.1)

-3.8

(0

.2)

0.5

(0

.1)

3.6

3

.6

-1.9

(0

.1)

0.6

(0

.1)

-1.4

(0

.1)

9.2

9

.0

-7.9

(0

.2)

4.9

(0

.2)

-3.0

(0

.1)

Ψ-U

(6

0)

- -

- -

- 0

.7 (

0.2

) -1

.4 (

0.2

) -0

.6 (

0.1

) 9

.4

9.1

-4

.5 (

0.1

) 2

.2 (

0.1

) -2

.3 (

0.1

)

1-M

e-Ψ

-U (

61

) -

- -

- -

1.1

(0

.1)

-1.8

(0

.1)

-0.7

(0

.1)

9.9

9

.7

-3.0

(0

.1)

1.5

(0

.1)

-1.5

(0

.1)

1,3

-diM

e-Ψ

-U (

62

) -

- -

- -

2.0

(0

.1)

-2.3

(0

.1)

-0.3

(0

.1)

- -

- -

-

β-D

-3'-d

A (

63

) 6

.9 (

0.2

) -3

.2 (

0.3

) 3

.8 (

0.1

) 3

.4

3.5

1

.3 (

0.1

) 0

.5 (

0.1

) 1

.8 (

0.1

) -

- -

- -



61

3.7 Estimation of ∆H°, ∆S° and ∆G° of the N� S equilibrium

3.7.1 Methodology

Each calculation performed with PSEUROT (using a single set of experimental temperature-

dependent 3JHH at one particular pD) produces one set of temperature-dependent mole fractions of

the N- and S-type pseudorotamers.

Table 3. ∆H° and -T∆S° contributionsa to ∆G298 of the two-state N�S pseudorotational

equilibrium in mononucleotides 64 - 83

Compound ΔH° ∆S° -T∆S298 ∆G298 % S278 % S358 ∆ %S

β-D-dAMP (64) -5.4 (0.3) -9.9 (0.7) 3.0 (0.2) -2.4 (0.4) 76 65 -11

β-D-dGMP (65) -4.1 (0.2) -6.5 (0.7) 1.9 (0.2) -2.2 (0.3) 73 64 -9

β-D-dCMP (66) -3.2 (0.1) -6.8 (0.5) 2.0 (0.1) -1.2 (0.2) 64 56 -8

β-D-TMP (67) -2.6 (0.1) -4.3 (0.4) 1.3 (0.1) -1.3 (0.2) 65 59 -6

β-D-dUMP (68) -2.7 (0.1) -3.8 (0.8) 1.1 (0.2) -1.6 (0.3) 67 61 -6

β-D-dAMPEt (69) -5.5 (0.2) -7.9 (0.7) 2.4 (0.2) -3.1 (0.3) 81 71 -10

β-D-dGMPEt (70) -4.8 (0.2) -7.0 (0.7) 2.1 (0.2) -2.7 (0.3) 77 68 -9

β-D-dCMPEt (71) -3.6 (0.2) -5.3 (0.9) 1.6 (0.3) -2.0 (0.3) 72 64 -8

β-D-TMPEt (72) -3.0 (0.1) -3.2 (1.1) 1.0 (0.3) -2.0 (0.3) 71 65 -6

β-D-dUMPEt (73) -2.7 (0.2) -2.3 (0.6) 0.7 (0.2) -2.0 (0.3) 71 65 -6

β-D-AMP (74) -4.9 (0.4) -9.1 (0.7) 2.7 (0.2) -2.2 (0.5) 74 63 -11

β-D-GMP (75) -4.5 (0.2) -10.8 (1.6) 3.2 (0.5) -1.3 (0.5) 66 55 -11

β-D-CMP (76) 0.8 (0.2) -0.6 (1.8) 0.2 (0.5) 1.0 (0.6) 40 42 2

β-D-rTMP (77) -0.9 (0.2) -1.6 (0.9) 0.5 (0.3) -0.4 (0.3) 55 53 -2

β-D-UMP (78) 1.2 (0.2) 2.1 (1.2) -0.6 (0.4) 0.6 (0.4) 43 46 3

β-D-AMPEt (79) -6.9 (0.8) -13.6 (1.1) 4.1 (0.3) -2.8 (0.9) 79 66 -13

β-D-GMPEt (80) -5.8 (0.4) -12.3 (1.1) 3.7 (0.3) -2.1 (0.5) 74 62 -12

β-D-CMPEt (81) -1.5 (0.2) -5.3 (1.0) 1.6 (0.3) 0.1 (0.4) 50 47 -3

β-D-rTMPEt (82) -2.5 (0.3) -5.3 (1.2) 1.6 (0.4) -0.9 (0.5) 61 55 -6

β-D-UMPEt (83) -1.6 (0.1) -3.1 (0.6) 0.9 (0.2) -0.7 (0.2) 58 54 -4

a ΔH° (kJmol-1), ΔS° (Jmol-1K-1), -TΔS° (298 K, kJmol-1) and ΔG° (298 K, kJmol-1) have been taken from ref. 23

for 64 - 67, from ref. 447 for 68 and 73 and from ref. 28 for 74 - 83. For 69 - 72, the values in Table 3 are refined

(using their sodium salts) in comparison with the original data in ref. 23 (ammonium salts).

At each pD, the input file for the PSEUROT calculations has been prepared in such a way

that (i) the hyperspace of geometries that is accessible to the constrained P and Ψm values of the

minor pseudorotamer and/or to the constrained Ψm of both N- and S-type conformations was well

covered, and (ii) the errors in the experimental 3JHH were taken into account by generating ≈ 1000

sets of "randomized" temperature-dependent coupling constants for each geometrical constraint.

Therefore, the total number of calculations at each pD was typically ≈ 5000 - 20000. The mole

fractions from each calculation are used to construct a van't Hoff plot. The average slopes and


62

intercepts from ≈ 5000 - 20000 van't Hoff plots are used to calculate ∆H° and ∆S° (and their errors)

of the N � S equilibrium of 12 - 83 at the particular pD. The free-energy ∆G° (T), at the

temperature T, of the two-state N � S equilibrium has been calculated either by (i) adding ∆H° and

-T∆S° contributions or (ii) directly from the average [lnaverage (xS / (1 - xS))] of the 5000 - 20000

individual ln (xS / (1 - xS)). The error on ∆G° (T) is given by: σ∆G° = [σ2

∆H° + σ2

-T∆S°] (case (i)) or

σ∆G° = -R.T.lnaverage (xS / (1 - xS)) (case (ii)).

∆H°, ∆S° and ∆G° of the two-state N � S equilibrium of the sugar moiety in compounds 12 - 16, β-

D-ddU (35), β-D-ddI (36) and mononucleotides 64 - 83 (at a single neutral pD), in L-nucleosides

(27 - 29 and 46 - 49) and in D-nucleosides 22, 23, 38 and 39 (at two or three pDs, one in each of

their P, N and D states) are presented in Tables 2 and 3. For all other nucleosides, the

conformational analyses has been performed over part of or the entire 0.5 - 12.0 pD range, and we

only report in Table 2 the limiting values of the thermodynamics of their two-state N � S equilibria

in each of the P, N and D states, which have been derived according to the procedure described in

the present section.

3.7.2 Accuracy of thermodynamics

The error on ∆H°, -TΔS° and ΔG° values corresponds to the standard deviation of the

average of the corresponding individual ∆H°, -TΔS° and ΔG° values. For all compounds, the

standard deviations on ∆H°, -TΔS° and ΔG° values are typically ≈ 0.5 - 1.0 kJ/mol, which have

been estimated during pseudorotational analyses in which the 3JHH error of ≈ 0.1 Hz for all

compounds except for some36 ddNs and abasic sugars (0.2 Hz) was taken into consideration. The

experimental accuracy of pD-dependent ΔH°, -TΔS° and ΔG° values were evident from the fact that

they gave the pKa values of the nucleobase within an accuracy ±0.2 pD unit in average (in

comparison with the literature values) with some exceptions where the ΔG° change as a function of

pD was less than 1.0 kJ/mol (such as pyrimidine nucleosides).

3.7.3 Influence of λN1/9 on the thermodynamics

In their recent investigation of the effect of solvent, pH, temperature and concentration upon

the values of the substituent parameters, λ, Altona et al have shown430 that the 3JHH coupling

constant to methyl in ethylamine and propylamine in D2O increases by 0.16 Hz and 0.26 Hz,

respectively, in going from pD 12.5 (NH2) to pD 7.5 (NH3+), therefore it was suggested that the λ

value of NH2 decreases from 1.10 to 0.82 upon protonation.

In view of this observation, we have examined the influence of the selected λ value for the

glycosyl nitrogen N9 of the nucleobase in β-D-dG (41), in each of its P (pD = 1.0), N (pD = 7.5)

and D (pD = 11.4) states, on ∆H°and ∆S° contributions to the free-energy ∆G° of its two-state N �

S equilibrium. At each pD, we have performed four calculations with PSEUROT. The input for

each of these calculations only differs in the value of λ(N9) which has been taken successively as

0.3, 0.58, 0.88 and 1.2. The individual ∆H° and ∆S° values at each pD for the twelve calculations

are collected in Table 4.

A perusal of the data in Table 4 suggests the followings:



63

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0

-2.5

-2.0

-1.5

-1.0

-0.5

77

73

69

65

60

55

50

pD

pD0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0

-3.0

-2.0

-1.0

0.0

1.0

2.0

77

69

60

50

40

31(E)

β-D-U (54 )

β-D-rT (53 )β-D-A (50 )

β-D-G (51 )

β-D-C (52 )

(D)

β-D-dG (41 )

β-D-dA (37 )

β-D-dC (42 ) β-D-dU (44 )

β-D-T (43 )

β-D-dImb (40 )

5-F-β-D-dU (45 )

5-F-β-D-U (55 )

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0

-4.0

-2.0

0.0

92

83

69

50

pD

Formycin B (56 )

Formycin A (57 )

9-deaza-A (58 )

Ψ -isoC (59 )

Ψ -U (60 )

1-Me-Ψ -U (61 )

(F)

1,3-diMe-Ψ -U (62 )

Figure 11 (Cont'd): Experimental ∆G° values (298 K) of N �S equilibria in α-D-ddNs 17 - 20 [Panel (A)], α-D-

dNs 21 and 24 - 26 [Panel (B)], β-D-ddNs 30 - 34 [Panel (C)], β-D-dNs 37, 40 - 45 [Panel (D)], β-D-rNs 50 - 55 [Panel

(E)] and β-D-C-rNs 56 - 62 [Panel (F)] as a function of pD. The sigmoidal plots (except for α-D-ddA, α-D-ddC, α-D-

dA and 1,3-diMe-Ψ-U, pD-independent values) were obtained by fitting the experimental data to Eq 12, giving the

pKa(s) of the nucleobases at the inflection point(s) and limiting ∆G° values in the P, N and D states (Table 2).


64

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

40

35

31

27

23

20

17

14

12

10

8

7

6

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0

-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

77

73

69

65

60

55

50

45

pD

α-D-ddA (17 ) & α-D-ddC (19 )

α-D-ddG (18 )

α-D-ddT (20 )

(A)

0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0

-4.5

-4.0

-3.5

-3.0

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

88

86

83

80

77

73

69

65

60

55

50

pD

α-D-dA (21 )

α-D-dG (24 )

α-D-dC (25 )

α-D-T (26 )

(B)

pD

β-D-ddA (30 )

β-D-ddG (31 )

β-D-ddC (33 )

β-D-ddT (34 )

5'-OMe-β-D-ddG (32 )

(C)

Figure 11 (See the legend p. 74)



65

(i) The largest influence of a change of λ(N9) upon ∆H° is found at alkaline pD [∆H°λ=1.2 -

∆H°λ=0.3 = -1.3 kJmol-1], which is nearly within the sum (0.9 kJmol-1) of the standard deviations of

∆H° values at both pDs. At pD 1.0 and 7.5 ∆H° is virtually pD-independent.

(ii) By comparing the ∆H° values that have been calculated at pD 1.0 and 7.5 using 0.58 as

the value for λ(N9), we find that the N-type pseudorotamers are stabilized by ∆∆H°(P-N, λ=0.58] =

4.6 kJmol-1 in going from the N to the P state of guanin-9-yl in β-D-dG, as the result of the

strenghtening of the anomeric effect (see the discussion in section 4.8).

(iii) If we now assume that λ(N9) is reduced from 0.58 at pD 7.5 to 0.3 at pD 1.0, as suggested by

Altona et al430, and if we compare again the corresponding ∆H° values obtained with this new λ

value, we find that the extent of the stabilization of N-type conformers at pD 1.0 with respect to pD

7.5 is the same as in (i), i.e. ∆∆H°(P-N, λ(pD 1.0)=0.3] = 4.6 kJmol-1.

(iv) If we assume that λ(N9) is not affected by the deprotonation of guanin-9-yl, we find that

S-type pseudorotamers are more stabilized at pD 11.4 than at pD 7.5 by ∆∆H°(D-N, λ=0.58] = -2.2

kJmol-1.

(v) If we now assume that λ(N9) is increased from 0.58 to 0.88 in going from pD 7.5 to pD

11.4, the stabilization of S-type conformers is even greater, i.e. ∆∆H°(D-N, λ(pD 11.4)=0.88] = -2.7

kJmol-1. In conclusion, the extent of the stabilization or destabilization of N-type pseudorotamers

upon protonation and deprotonation of guanin-9-yl in β-D-dG is independent of the value of λ(N9)!

Note however that this conclusion has been only validated for our system, and should be carefully

checked for any other unproven system.

Table 4. Influence of the value of the substituent parameter for the glycosyl nitrogen λ(N9)

in β-D-dG (41) in each of the P, N and D states of the constituent guanin-9-yl upon the

thermodynamicsa of the two-state N �S equilibrium of the constituent sugar moiety

P state (pD = 1.0) N state (pD = 7.5) D state (pD = 11.4)

λ(N9) ΔH°P -TΔS°P ΔG°P ΔH°N -TΔS°N ΔG°N ΔH°D -TΔS°D ΔG°D

0.3 1.8 (0.9) -1.8 (0.9) 0.0 (0.1) -2.6 (0.3) 1.1 (0.3) -1.5 (0.2) -4.6 (0.4) 2.2 (0.4) -2.4 (0.2)

0.58 1.8 (0.9) -2.0 (1.0) -0.2 (0.1) -2.8 (0.3) 1.0 (0.4) -1.8 (0.2) -5.0 (0.4) 2.3 (0.4) -2.7 (0.2)

0.88 1.8 (1.0) -2.2 (1.0) -0.4 (0.1) -3.0 (0.4) 1.1 (0.4) -1.9 (0.2) -5.5 (0.4) 2.5 (0.4) -2.9 (0.2)

1.2 1.8 (1.0) -2.2 (1.0) -0.5 (0.1) -3.2 (0.4) 1.5 (0.3) -1.7 (0.1) -5.9 (0.5) 2.9 (0.5) -3.0 (0.2)

a In kJmol-1. -T∆S° and ∆G° in the P, N and D states are given at 298 K. All thermodynamic parameters have

been calculated using our methodology for each particular value of λ(N9) (Section 3). Standard deviations of

each thermodynamic quantity take into account the influence of the uncertainty (± 0.1 Hz) in experimental 3JHH, as discussed in Section 3.6.


66

3.7.4 Influence of the nature of the aglycone dictates the thermodynamics

In the case of α-D-ddA (17), α-D-ddC (19) and α-D-dA (21), ∆H°, ∆S° and ∆G° were found

to be virtually pD-independent and the values reported in Table 2 for each of the P, N and D states

have been calculated by averaging the experimental data at all pDs.

On the other hand, the experimental ∆H°, ∆S° and/or ∆G° values of the N � S equilibria in

α-D-ddG (18), α-D-ddT (20), α-D-dG (24), α-D-dC (25), α-D-T (26), β-D-ddNs 30 - 34, β-D-dNs

37 and 40 - 45, β-D-rNs 50 - 55, β-D-ribo-C-nucleosides 56 - 61 and β-D-3'-dA (63) are dictated by

the pD of the aqueous solution30,32,36,37,426, as shown by the sigmoidal plots in Fig 11. The curves

through the experimental points in Fig 11 result from a nonlinear least-squares fitting of the

experimental data to the Henderson-Hasselbach equation (Eq 12), which gives limiting values of the

thermodynamics of two-state N � S equilibria in each of the P and D states and the pKa(s) of the

constituent nucleobase at the inflection point(s). [Note that the errors on the limiting ΔH°, -TΔS°,

and ΔG° values reported in Table 4 actually correspond to standard deviations of averages of

individual experimental ∆H°, ∆S° and ∆G° values at several pDs on the plateau in this particular

state, whereas the standard deviation for ∆H°, ∆S° and ∆G° values at each pD are higher (typically ≈

0.5 - 1.0 kJmol-1), as shown in the Tables of pD-dependent thermodynamics in the original

publications.]

pD = pKa + ]AH[

]A[log

+

= pKa + α

α−1log ..... Eq 12

In Eq 12, α represents the fraction of the protonated species, and it has been calculated from

the change in experimental ∆H°, ∆S° and ∆G° value at a given pD relative to the reference neutral

state divided by the total change in their respective values between the N state and the P or D state.

The pKa(s) of the nucleobase in 18, 20, 24 - 26, 30 - 34, 37, 40 - 45, 50 - 61 and 63 has been

determined either (i) directly from the nonlinear fit of the pD-dependent experimental ∆G° values to

Eq 12, as discussed above, or (ii) from Hill plots of pD as a function of the logarithm of the ratio of

the protonated to the unprotonated species. The Hill plots gave straight lines with Pearson's

correlation coefficients typically above 0.9 and slopes close to 1, which is characteristic of a

protonation � deprotonation equilibrium involving a single protonation site (see Table 2, cols. 5

and 10 for the pKa(s) of nucleobases derived from pD-dependent ∆G° values). The pKa values

determined from pD-dependent ∆G° values of the N � S equilibrium in nucleosides fit very well to

the literature values as well as those obtained from pD-dependent proton chemical shifts.

3.8 New Karplus equation to interpret 3JHF coupling constants

Substitution of a hydroxyl group or hydrogen in nucleosides for a fluorine atom does not

change significantly the steric effect of the functionality, however, the strongly electronegative

fluorine atom involved in powerful gauche interactions governs the overall conformation of the

sugar moiety199,206,210,222,448-449 Fluorine has been widely incorporated into nucleosides to design

new therapeutic agents (e.g. FLT (88): anti-HIV activity450; diFC (87): cytotoxic against Chinese

hamster ovary and tumor cells451; F2"C (112): inhibits the growth of human lymphoblastic cell

lines452) with specific puckering modes453-457. Incorporation of 2'-deoxy-2'-fluoro nucleosides in



67

DNA458 stabilizes the antisense DNA:RNA duplex, which adopts an A-form conformation, owing

to the stabilization of N-type sugars by the [F2'-C2'-C1'-O4'] gauche effect459. 2'-fluoro-2'-

deoxyguanosine in a hexadeoxynucleotide induces a A-DNA � Z-DNA conformational

transition460. Incorporation of 2'-fluoro-2'-deoxyarabinothymidine (S-type conformation) in the

Dickerson-Drew dodecamer stabilizes the duplex, but 2'-fluoro-2'-deoxyribothymidine destabilizes

it by inducing the formation an A-B junction in the middle of the duplex64. When nucleosides are

highly substituted (such as difluoronucleosides 84 - 87), it is not to assess the conformation of their

sugar moieties on the basis of a single or a few 3JHH (in 84 - 87, only 3J3'4' is available). 3JHF

coupling constants have only been qualitatively used to assess the structure and conformation of

pyranose derivatives461-463 and fluoronucleosides204,235 owing to the fact that no Karplus-type

equation allows to translate quantitatively experimental 3JHF data into HCCF torsion angles (ΦHF).

Several simplified equations such as 3JHF = A cos2 ΦHF + B cos ΦHF + C taking into account only

the torsion angle dependence were parametrized464-468.

3.8.1 Dataset of (3JHF,ΦHF) pairs for monofluoronucleosides

We have parametrized a Karplus equation for the interpretation of 3JHF coupling constants

using the strategy depicted in Scheme 4. We have initially performed pseudorotational analyses of

temperature-dependent 3JHH for monofluoronucleosides (88 - 98) using the PSEUROT203,427

program (steps 1 - 3 in Scheme 4). In the second step in PSEUROT (step 2 in Scheme 4), νi torsion

angles are translated into the corresponding HCCH torsion angles using simple a relationships:

ΦHH = AH νi + BH. AH and BH were obtained from ab initio calculations using the GAUSSIAN 94

program298 on 88, 90 - 92 and 95 - 98 (Table 3 and experimental Section in ref. 39). The plots of

ΦHH versus νi for 88, 90 - 92 and 95 - 98 as extracted from their ab initio optimized geometries

gave straight lines with correlation coefficients above 0.95.

From the results of these initial pseudorotational analyses, the following conclusions could

be drawn: (i) All compounds are strongly conformationally biased either to the N- or S-type sugar

geometries. (ii) In all nucleosides, the presence of strongly electronegative fluorine atom at C2' or

C3' on the α- or β-face makes the C-F bond adopt a gauche orientation with C1'-O4' or C4'-O4'. (iii)

This gauche effect is stronger over any other stereoelectronic effect and is responsible for the bias of

the sugar conformational equilibrium toward N- or S-type pseudorotamers, which is evidenced by

the following observations: In F3"ddU (91) and F3'ddU (98), the same anomeric effect stabilizes N-

type conformers. However, S-type pseudorotamers are preferred in 91 owing to predominant [F3"-

C3'-C4'-O4'] gauche effect favouring S-type sugars, whereas 98 is locked in N-type conformation

owing to cooperative anomeric effect and [F3'-C3'-C4'-O4'] gauche effect. As in F3"ddU (91), the

pentofuranose sugar in FLT (88) and its derivatives 89 and 90 prefers S-type conformations owing

to the stronger [F3"-C3'-C4'-O4'] gauche effect. In F2'ddU (96), the [F2'-C2'-C1'-N1] fragment is in

gauche orientation both in N- and S-type pseudorotamers, and the stabilization of S- over N-type

sugar is owing to the stronger [F2'-C2'-C1'-O4'] gauche effect over the anomeric effect. In F2"ddU

(97) the combined influence of the [F2"-C2'-C1'-O4'] gauche effect and of the anomeric effect

overrides the counteracting [F2"-C2'-C1'-N1] gauche effect and N-type conformers are preferred.


68

Basing on the results (PN, PS, Ψm(N), Ψm(S) and temperature-dependent xS) of these

pseudorotational analyses, we have built a dataset composed of 29 pairs of (3JHF,ΦHF) for the

major pseudorotamers of 88 - 98 in two steps (steps 4 and 5 in Scheme 4): (i) From linear

relationships ΦHF = AF νi + BF (based on torsion angles extracted from ab initio optimized

geometries of 88, 90 - 92 and 95 - 98, vide supra) we calculated ΦHF torsion angles from the

endocyclic torsions νi for each HCCF fragment in the major pseudorotamers of 88 - 98. (ii) We

found a linear correlation between temperature-dependent xS and 3JHF values, which allowed us to

calculate limiting 3JHF values for pure N- or S-type (major) pseudorotamers in 88 - 98 (Table 4 in

ref. 39).

An analysis of the statistical distribution of the 29 pairs of (3JHF,ΦHF) for 88 - 98 revealed

that the torsion angles are not distributed evenly over the whole 0 - 360° range, instead most of

them are found in the cis and trans regions. Additionally, neither torsion angles around ≈ ± 90° nor 3JHF above 45.3 Hz were present in this dataset. It is however known that the limiting 3JHF

coupling constants can be ≈ 45 - 47 Hz for the trans substitution pattern of decalins469,470.

Therefore a parametrization based on this dataset would not have allowed to define precisely the

position of the minima (≈ ± 90°) and maxima (≈ 180°) of the Karplus curve in these regions. We

have therefore extended (step 6 in scheme 4, Ref. 39) our dataset by incorporating 28 additional

pairs of (3JHF,ΦHF) from conformationally constrained (cyclic) compounds 99 - 109466,467,469-471.

ΦHF were extracted from the structures of 99 - 109, whose geometry was optimized ab initio with

GAUSSIAN 94298. Among these 28 pairs, 3 (entries #31, 33, 42 in Table 4 in ref. 39) allowed to

precise the (3JHF,ΦHF) values around ± 90° while 9 (entries #43, 44, 47, 48, 51 - 54, 57) were found

in the trans region (with 3JHaxF = 46 Hz in 109).

3.8.2 Parametrization of the Karplus equation

A perusal of the 57 pairs of (3JHF,ΦHF) in Table 4 (in ref. 39) shows variations of up to 16

Hz between the value of 3JHF either in the cis or trans region for the same torsion angle in different

compounds (compare 3JH2"F3" in F3AT (95) and 3JH3'F2' in F2'ddU (96), on one hand; 3JH2'F3" in

AFLT (90) and 3JH4'F3' in FXA (92)). This can be attributed to different substitution patterns on

H2"-C2'-C3'-F3" in 95 and H3'-C3'-C2'-F2' in 96, in one hand, and H2'-C2'-C3'-F3" in 90 versus

H4'-C4'-C3'-F3' in 92, in the other. This led us to conclude that a simple 3-term Karplus-type

equation (i.e. of the form 3JHF = A cos2 ΦHF + B cos ΦHF + C) will not reproduce our experimental

3JHF accurately on the basis of the torsion angle value alone. It has been suggested that the

electronegativity of the substituents on the HCCF fragment affects the value of the 3JHF coupling

constant461,472. In order to account for the influence of the nature and configuration of the

substituents on the HCCF fragments on the value of 3JHF, we have first parametrized (step 7 in

scheme 4) a Karplus-type equation, whose form is that originally proposed by Altona et al (Eq

8a)431 (Section 3.4). In Eq 8a, we have however used λ substituent parameters429 to account for the

effect of the electronegativity of the substituents on the H-C-C-F fragments instead of the original

Huggins group electronegativities (Δχi(g) ).

Through a MonteCarlo fitting procedure, we optimized the values of P1 - P6. Large

discrepancies up to 10 Hz between experimental 3JHF and those back-calculated with help of Eq 8a



69

were found, suggesting that its form is not adequate. By examining the variations in 3JHF values for

a specific ΦHF value, we then realized that the coupling constants are also dictated by the values of

HCC and FCC bond angles, as suggested in early qualitative studies466-468. After examining several

forms of equations to reproduce the effects of bond angle, we found that it is best reflected in a

cosine square function of the torsion angle (Eq 8c, Section 3.4). The r.m.s of the fit of our

experimental data to the coupling constants calculated using Eq 8c is 1.38 Hz and the largest

individual error is 2.9 Hz (Table 4 in ref. 39). Thus, the accuracy of our newly developed Karplus

equation (r.m.s. ≈ 1.4 Hz for 3JHF in the trans region ≈ 46 - 47 Hz, i.e. ≈ 3 % error) is comparable to

that used commonly to interpret 3JHH data431 (r.m.s. ≈ 0.4 - 0.5 Hz for 3JHH in the trans region ≈ 8 -

9 Hz, i.e. ≈ 4 % error). Other combinations of parameters in Eq 8c led to similar r.m.s error,

however, plots of 3JHF for particular substitution patterns (in particular highly substituted, with

substituents of high electronegativity) a function of torsion angle showed negative (below -0.5 Hz) 3JHF values for ΦHF ≈ ± 90° (Fig. 1 in ref. 39).

3.8.3 Pseudorotational analyses of 3JHF in fluoronucleosides to validate the Karplus

equation.

In order to demonstrate that Eq 8c can be used as a reliable tool in conformational analysis of

fluoronucleosides, we have subsequently proceeded in four steps:

Plots of 3JHF versus 3JHH for qualitative conformational analysis of 88 - 98: We have first

constructed plots of 3JHF versus 3JHH data for different P values in the 0° to 360° range (at different

Ψm) for mononucleosides 88 - 98 (Figs 3 and 4 in Ref. 39). The position of the experimental

temperature-dependent pairs of (3JHF, 3JHH) for 88 - 98 suggested a high preference of their

pentofuranose moieties for N- and S-type pseudorotamers and the extent of this preference was in

good agreement with that obtained from our initial pseudorotational analyses based on 3JHH data

only.

Pseudorotational analyses on 88 - 98 based on a combination of 3JHF and 3JHH data (Steps 8 -

10 in scheme 4): We have slightly modified (see the experimental Section on "PSEUROT+JHF" in

Ref. 39 and our web site) the original PSEUROT version 3B program in order to make it possible to

perform a pseudorotational analysis using either (i) temperature-dependent 3JHH data or

temperature-dependent 3JHF data alone or (ii) a combination of 3JHH and 3JHF coupling constants.

(i) The results (PN, PS, Ψm(N), Ψm(S) values and temperature-dependent xS as well as r.m.s.

error and individual errors in 3JHF or 3JHH) of the pseudorotational analyses performed on 88 - 98

are compiled in Table 5 in Ref. 39. The comparison of the results from analyses based on 3JHH or

3JHF data only shows that the largest difference between P values is 35° (for AFLT), whereas for

Ψm values it is at most 13° (for AFLT), and for xS values we found a largest error of 13 % unit (for

F2'ddU). The r.m.s. error of both types of analysis is similar and ≤ 0.8 Hz. The largest individual

error between experimental and back-calculated 3JHH and 3JHF is 1.2 Hz and 2.0 Hz, respectively.

(ii) The analyses based on 3JHH and 3JHF data have been performed in two ways. During the

multilinear fitting process with PSEUROT+JHF, the errors in 3JHH and 3JHF data have been either

attributed the same "scale factor (1.0) or in a second series of analyses, the error in 3JHF was scaled


70

down to 20 % of that in 3JHH in order to account for the fact that the former are in average ≈ 5 times

greater than the latter. The introduction of scale factors did not affect the geometry of the

pseudorotamers in the final output of the PSEUROT calculation. The analyses based on 3JHH and

3JHF show that the P and Ψm values obtained from the analyses described in step (i) and (ii) are

comparable r.m.s. errors below 1.2 Hz were found.

Qualitative and quantitative pseudorotational analyses on monofluoronucleosides 111, 112

and difluoronucleosides 84 - 87: In order to further verify the validity of Eq 8c, we have applied it

to the conformational analysis of another set of monofluoro (111 and 112) and difluoronucleosides

(84 - 87) through qualitative plots of 3JHF as a function of 3JHH data (vide supra, Fig 5 in Ref. 39)

for phase angles values from 0° to 360° or to the quantitative pseudorotational analysis using our

modified PSEUROT+JHF program (Table 5 in Ref. 39). The position of the experimental pairs of

(3JHF, 3JHH) data suggest that the pentofuranose sugar adopts almost exclusively (or predominantly)

N-type conformations in F3'ddA (or in F2"C). For 84 - 87, conformational analyses based on the

single 3JH3'H4' are at best misleading. Our qualitative plots for 84 - 87 show for the first time that

the conformation of their sugar moieties is strongly biased toward N-E-type pseudorotamers (50° <

P < 90°), as suggested by the clusering of all experimental coupling constants in this region. Our

more elaborate pseudorotational analyses on 3JHH and/or 3JHF data for 84 - 87, 111 and 112 precises

the conclusions drawn on the basis of the above plots. Note that the results of the pseudorotational

analyses are again virtually independent of the type of vicinal coupling constants used (3JHH and/or

3JHF) (i) The sugar moiety in F3'ddA (111) is indeed locked in N-type conformation (≥ 94 % at 298

K) with PN in the range from 11° to 20° while Ψm is between 34° and 37°. The r.m.s. of all analyses

is ≤ 0.7 Hz with -1.3 Hz as the largest error in 3JHH and 3JHF in analyses performed using 3JHH and

3JHF data. (ii) Similarly, N-type pseudorotamers are strongly preferred in F2"C (112) (by > 73%)

with PN ≈ 36° and Ψm(N) ≈ 33° - 37°. The r.m.s. error is ≤ 1.7 Hz for analyses based on all coupling

constants with 3.2 Hz as the largest individual error in 3JHF. (iii) In difluoronucleosides 84 - 87,

C4'-exo to O4'-endo-C4'-exo pseudorotamers (61° < P < 76°, Ψm ≈ 35° - 46°) are favoured by 78 -

94 %. The r.m.s. of all calculations is below 1.5 - 2.0 Hz, while individual errors are at most 3.4 Hz

for 3JHF data.

In conclusion, this is the first report of an accurate Karplus-type equation for the

interpretation of 3JHF coupling constants in terms of torsion angles that can be successfully used to



71

define the solution conformational characteristics and relative populations of conformers engaged in

an equilibrium either in nucleosides, nucleotides or organic compounds in general.

4. The quantitation of the stereoelectronic effects in nucleos(t)ides

A study of our first experimental thermodynamic estimates14 of pseudorotations, as early as

1987, on 9-(2'-deoxy-β-D-threo-ribofuranosyl)-adenine and 9-(3'-deoxy-β-D-threo-ribofuranosyl)-

adenine and their comparison with the natural 2'-deoxyadenosine and 3'-deoxyadenosine revealed

that the net result of the gauche effect and the anomeric effect is of major importance in determining

the overall furanose conformation in nucleosides. This has subsequently led us to examine the

interplay of gauche and anomeric effects determining overall sugar conformation in nucleosides, in

general, more in details. The strength of the gauche and anomeric effects operating in 12 - 83 is

reflected in the value of ∆H° contribution to ∆G° of their N � S equilibrium (N � E in 12 and

1420). Our ∆H° estimates correspond to the overall contribution (i.e. stereoelectronic and

counteracting steric effect) of each molecular fragment. -T∆S° of the N �S equilibrium in 12 - 83

shows the difference in the entropy between both pseudorotamers at the temperature T. The effect of

a possible change in the solvation and/or bulk of the nucleobases or other substituents at C2'-C5' in

17 - 83 at different pDs (and/or owing to different configuration at C1') on the modulation of ∆H°

and/or -T∆S° of their N � S equilibria in either of their P or D states compared with the N state

cannot be assessed in a straightforward quantitative manner.

4.1 Quantitation of the anomeric and gauche effects by regression analyses

To quantitate the gauche and anomeric effects that drive the sugar conformation in 12 - 73,

we have initially correlated their structural features with the experimental values of their two-state N

� S equilibria via the regression analyses (A) - (E) (using the program SYSTAT473 (Table 5)).

4.1.1 Stereoelectronic effects in neutral β-D-Ns

The 26 ΔH°N values of the N � E equilibrium (in 12 and 14) or N � S equilibrium in 13, 15

and 16, β-D-ddNs 30 - 35, β-D-dNs 37 - 39 and 41 - 45, β-D-rNs 50 - 55 and β-D-3'-dA (63) have

been dissected into following components: (i) The (steric + gauche) effect of 5'CH2OH; (ii) The

effect of 5'CH2OMe compared with 5'CH2OH; (iii) The [HO2'-C2'-C1'-O4'] gauche effect; (iv) and

(v) The [O2'-C2'-C1'-N9(purine)] and [O2'-C2'-C1'-N1(pyrimidine)] gauche effects for purine and

pyrimidine ribonucleosides, respectively; (vi) The [HO3'-C3'-C4'-O4'] gauche effect; (vii) The

[MeO3'-C3'-C4'-O4'] gauche effect and (viii) - (xiii) the overall (steric + stereoelectronic) effects of

adenin-9-yl, guanin-9-yl, cytosin-1-yl, thymin-1-yl, uracil-1-yl, 5-fluorouracil-1-yl. The ΔH°N

values of β-D-ddI (36) and β-D-dImb (40) were not used, since no other nucleoside has

hypoxanthin-9-yl or imidazol-1-yl as nucleobase. The Pearson's correlation coefficient of regression

analysis (A), performed using these 26 ΔH°N values, is 0.993 and the standard error of estimates 0.6

kJmol-1. The largest individual errors between experimental and back-calculated [using estimates

from regression (A)] ΔH°N values was found for 12 (-1.2 kJmol-1) and β-D-ddA (30) (1.0 kJmol-1).


72

Table 5. Estimationa of the strengths of the anomeric and gauche effects driving the sugar

conformation in 12 - 73 from regression analysis with the SYSTAT program473

Regression model (A) (B) (C) (D) (E)

SE 5'CH2OH 1.6 (0.4) 1.6 (0.3) 0.1 (0.3) 1.6 (0.3) 1.7 (0.3)

SE 5'CH2OMe 0.7 (0.4) 0.7 (0.4) 0.5 (0.4) 0.9 (0.6) 0.6 (0.4)

AE (A) 0.9 (0.4) 0.9 (0.3) -1.7 (0.3) 1.2 (0.3) 1.0 (0.3)

AE (G) 2.0 (0.5) 2.0 (0.4) -0.6 (0.4) 1.5 (0.4) 2.0 (0.4)

AE (C) 4.4 (0.5) 4.3 (0.4) -3.2 (0.3) 3.7 (0.4) 3.9 (0.4)

AE (T) 3.5 (0.5) 3.5 (0.4) -0.6 (0.3) 2.1 (0.4) 3.6 (0.4)

AE (U) 4.1 (0.5) 4.1 (0.4) - 3.5 (0.6) 4.0 (0.4)

AE (5-FU) 4.1 (0.6) 4.1 (0.5) - 3.2 (0.8) 4.0 (0.5)

GE[HO3'-C3'-C4'-O4'] -6.3 (0.3) -6.4 (0.2) -3.9 (0.2) -5.4 (0.4) -6.3 (0.2)

GE[MeO3'-C3'-C4'-O4'] -6.7 (0.5) -6.7 (0.4) -4.7 (0.4) -6.9 (0.6) -6.7 (0.4)

GE[-1/-2RO3PO3'-C3'-C4'-O4'] - - - - -8.3 (0.3)

GE[O2'-C2'-C1'-O4']] 5.1 (0.7) 5.1 (0.6) - 4.2 (1.0) 5.1 (0.6)

GE[O2'-C2'-C1'-N9(pur.)] -5.9 (0.8) -5.9 (0.7) - -5.6 (1.1) -6.0 (0.7)

GE[O2'-C2'-C1'-N1(pyr.)] -2.5 (0.8) -2.4 (0.7) - -1.6 (1.1) -2.3 (0.7)

Multiple R 0.993 0.993 0.992 0.974 0.992

σ estimates 0.6 0.5 0.4 1.0 0.5

a See the text for the assumptions used to build each regression model. All estimates are given in kJmol-1 (their standard

deviations are indicated in parentheses). SE 5'CH2OH designates the overall (steric + gauche) effect of 5'CH2OH

substituent, whereas SE 5'CH2OMe represents the additonal stabilization of N-type conformations by 5'CH2OMe in

comparison with 5'CH2OH. AE (A), AE (G), AE (C), AE (T), AE (U) and AE (5-FU) give the overall strength of the

(steric + stereoelectronic) effect of adenin-9-yl, guanin-9-yl, cytosin-1-yl, thymin-1-yl, uracil-1-yl and 5-fluoro-uracil-1-

yl in 17 - 73. GE[HO3'-C3'-C4'-O4'], GE[MeO3'-C3'-C4'-O4'] and GE[O2'-C2'-C1'-O4'] are estimates of the strength of

the [HO3'-C3'-C4'-O4'], [MeO3'-C3'-C4'-O4'] and [O2'-C2'-C1'-O4'] gauche effects. GE[-1/-2RO3PO3'-C3'C4'-O4'] in

regression analysis (E) represents the magnitude of the [RO3PO3'-C3'C4'-O4'] gauche effect in β-D-dNMPs 64 - 68 (R =

H, charge = -1 and -2 in 1:1 ratio, Section 7) and β-D-dNMPEt (R = Et, charge is -1) 69 - 73, which has been assumed

to be the same for each nucleobase and independent of the nature of R. GE[O2'-C2'-C1'-N9(pur.)] and GE[O2'-C2'-C1'-

N1(pyr.)] represent the strength of the [O2'-C2'-C1'-N9(pur.)] and [O2'-C2'-C1'-N1(pyr.)] gauche effects in purine and

pyrimidine ribonucleos(t)ides, respectively. The multiple R is the Pearson's correlation coefficient. "σ estimates"

corresponds to the average standard deviation of all the estimated stereoelectronic effects for each regression analysis.

4.1.2 Effect of the 5'CH2OH versus 5'CH2OMe

According to regression analysis (A), the steric and [O5'-C5'-C4'-O4'] gauche effects of

5'CH2OH drive the conformation of abasic sugars and nucleosides toward the N [∆H°(5'CH2OH) =

1.6 kJmol-1]. Substitution of 5'CH2OH in 15, 31 and 38 for 5'CH2OMe in 16, 32 and 39 slightly

reinforces the drive of sugar conformation toward N-type forms [i.e. ∆H°(5'CH2OMe) >

∆H°(5'CH2OH)]. The stronger influence of 5'CH2OMe than 5'CH2OH might be attributed to the

increased steric bulk of OMe compared with OH. ∆H°(5'CH2OH) is stronger (by 1.2 kJmol-1) than

the experimental ΔH°N value of 12, where only 5'CH2OH controls the sugar conformation. This

might suggest that the effect of 5'CH2OH is stronger in β-D-nucleosides than in 12, owing to the

proximity of the nucleobase in the former. However, the contribution of 5'CH2OH to the drive of



73

the sugar conformation in β-D-nucleosides is much weaker than that of all other stereoelectronic

effects.

4.1.3 Effect of the nucleobase

The nucleobase promoted drive of the N � S equilibrium in nucleos(t)ides is the result of

two counteracting contributions: stereoelectronic interactions within the O4'-C1'-N1/9 fragment, i.e.

nO4' →σ∗C1'-N1/N9 interactions and/or electrostatic repulsions between the dipole of the furanose

ring and the C1'-N1/9 dipole, and the inherent steric effect of the nucleobase (Section 2.7).

Stereoelectronic interactions are expected to drive the sugar conformation toward N-type

pseudorotamers, in which the nucleobase is pseudoaxial, whereas the steric effect stabilizes S-type

forms in which the nucleobase adopts a pseudoequatorial orientation (Section 2.7). The fact that

ΔH° estimates resulting from regression analysis (A) for the overall effects of the nucleobases in β-

D-ddNs, β-D-dNs, β-D-rNs and 63 have been found positive shows that the C4'-O4'-C1'-N1/9

fragments adopt preferentially a gauche over trans orientation owing to predominating

stereoelectronic over steric interactions, therefore an anomeric effect is operating. The magnitude of

the overall effect of the nucleobase is dictated by its electronic and steric nature, which are as

follows: adenin-9-yl (0.9 kJmol-1) < guanin-9-yl (2.0 kJmol-1) < thymin-1-yl (3.5 kJmol-1) < uracil-

1-yl (4.1 kJmol-1) ≈ 5-fluoro-uracil-1-yl (4.1 kJmol-1) < cytosin-1-yl (4.4 kJmol-1). The effects of

thymin-1-yl, uracil-1-yl, 5-fluorouracil-1-yl and cytosin-1-yl are stronger than those of adenin-9-yl

and guanin-9-yl because in the former, N1, part of the electron-deficient pyrimidine ring, is more

predisposed to accept electrons through nO4'→σ∗C1'-N orbital interactions than N9 in the latter

which is part of the electron-rich imidazole ring. Similar correlations between the strength of the

anomeric effect involving a substituent X at C2 in heterocyclic six-membered rings and the

electronegativity of X have already been established (Section 1.5). However, it seems that 5-

fluorouracil-1-yl behaves as an exception, since its overall effect is the same as that of uracil-1-yl,

although one would expect fluorine to reinforce the electron-withdrawing character of its glycosyl

nitrogen N1 and therefore its nO4' →σ∗C1'-N1 stereoelectronic interactions.

4.1.4 Gauche effects

The conformation of the pentofuranose moiety in β-D-dNs is driven toward the S owing to

the fact that the [HO3'-C3'-C4'-O4'] (or [MeO3'-C3'-C4'-O4'] in 38 and 39) gauche effect prevails

over the counteracting overall effect of their nucleobases, as evident from the comparison of their

respective ∆H° estimates. The magnitudes of the [HO3'-C3'-C4'-O4'] and [MeO3'-C3'-C4'-O4']

gauche effects is almost the same, which is consistent with the comparable electronegativities of 3'-

OH and 3'-OMe26. In β-D-rNs compared with β-D-dNs, two additional gauche effects operate

within [HO2'-C2'-C1'-O4'] and [HO2'-C2'-C1'-N1/9] fragments. Our regression analysis shows that

∆H° of the [HO3'-C3'-C4'-O4'] and [HO2'-C2'-C1'-O4'] gauche effects in β-D-rNs almost cancel

each other, which is consistent with the fact that the experimental ΔH°N values of the two-state N �

S equilibria in abasic sugars 12 and 14 are identical. Therefore, the actual preference (as reflected in

their respective experimental ΔH°N values) of β-D-rNs for N- or S-type conformations can be

attributed to the fine balance of the overall effect of their nucleobases and of the [HO2'-C2'-C1'-

N1/9] gauche effect. In purine nucleosides β-D-A and β-D-G, N9 is part of the π-electron-rich


74

imidazole ring, and its ability to participate to a gauche interaction with O2' is much greater than

that of N1 in the π-electron deficient pyrimidine counterparts 52 - 55. This is reflected in the

stronger [HO2'-C2'-C1'-N9(purine)] gauche effect compared to the [HO2'-C2'-C1'-N1(pyrimidine)]

gauche effect. In sharp contrast, the ability of the purine nucleobase in β-D-A and β-D-G to shift the

pseudorotational equilibrium of the constituent pentofuranose sugar toward N-type conformations is

clearly reduced with respect to that of the pyrimidine ring in 52 - 55. Consequently, ΔH°N values of

the N � S equilibria in purine and pyrimidine β-D-rNs reflect the stabilization of S- and N-type

pseudorotamers, respectively. In β-D-3'-dA, no [HO3'-C3'-C4'-O4'] gauche effect is operating and

the pseudorotational equilibrium is driven toward N-type forms owing to the cooperativity of the

effects of the nucleobase, the 5'CH2OH substituent and of the [HO2'-C2'-C1'-O4'] gauche effect,

which override the counteracting [O2'-C2'-C1'-N9(purine)] gauche effect.

4.1.5 Energetic equivalence of mirror-image β-D-dNs and β-L-dNs

The mirror image relationship of the natural DNA and of the unnatural β-L-counterpart 474-

476 has been proven by the identical X-ray crystal structures of D- and L-d(CGCGCG)

hexamers474,477. The oligomerization of activated guanosine mononucleotides in the presence of

poly(C) template is easily achieved if both the template and the monomers are of the same

handedness, whereas it is much less efficient with the racemic D/L mixture of the monomer, since

monomers of opposite handedness act as chain terminators due to their incorporation at the 2'(3')

end of the oligomer478. The comparison of the thermodynamics of the two-state N � S equilibria in

β-D-dNs 37 and 41 - 43 and in their β-L-dNs counterparts 46 - 49 (either with neutral, fully

protonated or deprotonated nucleobases) shows that D- and L-nucleosides cannot be energetically

distinguished within the timeframe and accuracy of our method based on pseudorotational analyses

of 3JHH, since they exhibit virtually identical pD-dependent conformational preferences (Table 2).

The largest individual difference (0.8 kJmol-1) is found between ΔH°N values of β-D-dG (41) and

β-L-dG (47) with fully protonated guanin-9-yl and it is within the accuracy (σ = 0.8 kJmol-1) of the

experimental ΔH°N value of β-L-dG. This suggests that the magnitudes of the stereoelectronic

anomeric and gauche effects that drive the sugar conformation in D- and L-nucleosides are identical.

In order to improve the statistical significance of the estimates derived from the regression analysis

(A), we have performed a second multilinear regression analysis [regression (B)] basing on an

extended dataset consisting of 30 experimental ΔH°N values (i.e. 26 from the dataset used for

regression (A) as well as the 4 ΔH°N values for β-L-dNs 46 - 49). The estimates for the gauche and

anomeric effects from regression (B) are virtually identical (within ± 0.1 kJmol-1) to those derived

from regression (A) (Table 5, col. 2). The Pearson correlation coefficients of regressions (B) and (A)

are nearly identical (0.993 versus 0.994). The standard error of estimates determined from

regression (B) is 0.5 kJmol-1 [as compared with 0.6 kJmol-1 for regression (A)]. The largest error

between experimental and theoretical ΔH°N values [i.e. calculated using the estimates from

regression (B) or (A)] for nucleosides is found for β-D-ddA (1.0 kJmol-1). For all β-L-dNs

[regression (B)], the error is within 0.3 kJmol-1.

4.1.6 3'-phosphate has stronger gauche effect than 3'-hydroxy



75

The magnitudes of the gauche and anomeric effects operating in all β-D/L-nucleosides in β-D-

dNMPs 64 - 68 and β-D-dNMPEts 69 - 73 (Table 5) have been derived from the regression analysis

(E), which was performed using 40 experimental ΔH°N values assuming identical strengths for the

[-1/-2HO3PO3'-C3'-C4'-O4'] and [-1EtO3PO3'-C3'-C4'-O4'] gauche effects in 64 - 68 and 69 - 73.

The estimates resulting from regression analysis (E), its correlation coefficient and the standard

error on the estimates are virtually the same as of regression (B) (Table 5). The average [3'-

phosphate-C3'-C4'-O4'] gauche effect is about -8.3 kJmol-1, i.e. -2.0 kJmol-1 stronger than the

[HO3'-C3'-C4'-O4'] counterpart.

4.1.7 Stereoelectronic effects in neutral α-D-N

In order to quantitate the anomeric and gauche effects driving the conformation of α-

nucleosides in the N state and compare them with those obtained for β-nucleosides [regression

(B)]), we have performed the regression analysis (C) using 17 ΔH°N values for abasic sugars 12, 13,

15 and 16, α-D-ddNs 17 - 20, α-D-dNs 21 - 26 and α-L-dNs 27 - 29. These ΔH°N values have been

correlated with the magnitudes of: (i) the effect of 5'CH2OH; (ii) the additional influence of

5'CH2OMe compared with 5'CH2OH; (iii) the [HO3'-C3'-C4'-O4'] gauche effect, (iv) the [MeO3'-

C3'-C4'-O4'] gauche effect and the overall effects of (v) adenin-9-yl, (vi) guanin-9-yl, (vii) cytosin-

1-yl and (viii) thymin-1-yl. Since, as in the β-anomers, the ∆H° values of the two-state N � S

equilibrium in each of the α-D-/-L- pairs are identical [within the accuracy of the individual

estimates, i.e. with 1.0 kJmol-1 as largest deviation for ΔH°N of α-D-dA (21) versus α-L-dA (27)],

we have considered that the anomeric and gauche effects operate with an identical strength in both

α-D- and -L-series. The correlation coefficient of the regression analysis (C) is 0.992 [0.993 for

regression (B)] and the standard error of the estimates is 0.4 kJmol-1 (0.5 kJmol-1 for regression

(B)]. As a result, the residual errors between theoretical ∆H° values for 12, 13 and 15 - 29

[calculated using the estimates derived from the regression analysis (C)] and the corresponding

experimental ΔH°N values are well within the accuracy of the later.

4.1.8 Weakening of the effects of 5'CH2OH and 5'CH2OMe in α-nucleosides

The comparison of the data in cols. 3 and 4 for regressions (B) (β-nucleosides) and (C) (α-

nucleosides) in Table 5 suggests that in α-nucleosides, the contribution of the effects of 5'CH2OH

and 5'CH2OMe to the drive of the two-state N �S equilibrium is reduced in comparison with the β-

counterparts. This is possibly due to the fact that in the former the 5'-substituent and the nucleobase

are on opposite faces of the pentofuranose ring whereas they are both on the β-face in the later.

4.1.9 Weakening of the effect of the nucleobase in α-nucleosides

The data in Table 5 indicate that the magnitude of the overall effect of the nucleobase is

weaker in α- compared with β-nucleosides by 1.2 kJmol-1 for cytosin-1-yl, 1.4 kJmol-1 for guanin-

9-yl and 2.9 kJmol-1 for thymin-1-yl. Only the overall effect of adenin-9-yl has a comparable

magnitude in the α- and β-series (0.9 kJmol-1 in the β-anomers versus slightly stronger -1.7 kJmol-1

in the α-counterparts). The weakening of the anomeric effect in α-nucleosides implies that the

configuration of the nucleobase at C1' presumably affects the ability of one of the lonepairs of the


76

O4' oxygen to be engaged in molecular orbital interactions with the σ∗C1'-N1/N9

antibonding orbital.

This may also suggest that the anomeric effect is the result of predominant molecular orbital

interactions over electrostatic repulsions. One may possibly invoke a sp2 hybridization of the O4'

lonepairs orbitals, and an unfavourable orientation of the occupied lobe of the 1nsp2 (p-type) lonepair

with respect to the σ∗C1'-N1/9

orbital in the α-series only (Fig 12). We have shown that in β-

nucleosides the pyrimidine nucleobases drive the two-state N � S pseudorotational equilibrium

toward N-type conformations more efficiently than the purine counterparts (Section 4.4). In the α-

anomers, the ability of the nucleobase to stabilize S-type pseudorotamers is only slightly increased

when the base is changed in the order guanin-9-yl, thymin-1-yl, adenin-9-yl, cytosin-1-yl. The fact

that thymin-1-yl has an intermediate effect cannot be easily explained.

4.1.10 Weaker 3'-hydroxy gauche effect in α- compared with β-D-Ns

The [HO3'-C3'-C4'-O4'] and [MeO3'-C3'-C4'-O4'] gauche effects drive the pseudorotational

equilibrium of the pentofuranose moiety both in α- and β-nucleosides toward S-type conformations,

however their magnitude is much reduced (by 2.4 and 2.0 kJmol-1, respectively) in the α-anomers

compared with the β-counterparts (Table 5, cols 2 and 4). The fact that the nucleobase and the 3'-

substituents do not influence the drive of the pseudorotational equilibrium in α- and β-nucleosides

with the same efficiency is further evidenced by the results of the regression analysis (D), which has

been performed using a dataset of 43 experimental ΔH°N values, including abasic sugars 12 - 16, β-

D-ddNs 30 - 35, β-D-dNs (except 40), β-D-rNs, β-D-3'-dA (63) and β-L-dNs, α-D-ddNs 17 - 20, α-

D-dNs 21 - 26 and α-L-dNs 27 - 29, basing upon the assumption that the magnitudes of the effects

of all nucleobases as well as the gauche effects are independent of the configuration of the sugar

moiety at C1' (for the effects of the nucleobase, we have only assigned the opposite sign in the α-

series compared to the β-counterparts during the fitting procedure) (Table 5, col. 5). The multiple

Pearson's correlation coefficient is 0.974 [0.993 for regression (B)]. The standard error of the

estimates derived from regression (D) is twice higher (1.0 kJmol-1) than for those derived from

regression (B) (0.5 kJmol-1). Whereas the estimates of the anomeric and gauche effects determined

through the regression analyses (B) and (C) allow to reproduce the experimental ΔH°N values of β-

D- and α-D-nucleosides, respectively, fairly accurately, some theoretical ∆H° values calculated on

the basis of the estimates derived from regression (D) are much less accurate, and individual

residual errors between them and the corresponding experimental ΔH°N are -2.1 kJmol-1 for α-D-

ddA, 1.9 kJmol-1 for α-D-T, 1.7 kJmol-1 for α-L-T, 1.8 kJmol-1 for β-D-ddT, ±1.3 kJmol-1 for β-L-

dA, β-D-dA and β-D-ddC.

4.1.11 Limitations of regression analysis to quantitate stereoelectronic effects

For the regression analyses (A) - (E), the following approximations have been made:

(i) The magnitude of the effect of each nucleobase has been assumed to be identical in β-D-

ddNs, β-D-dNs and β-D-rNs (as well as in 63 for adenin-9-yl). This means that any potential

modulation of the electron-withdrawing character of the glycosyl nitrogen and electron-donating

ability of O4' by the presence of 2'- and/or 3'-OH(Me) substituents in β-D-dNs and β-D-rNs in



77

comparison with β-D-ddNs counterparts has been neglected. In the case of the glycosyl nitrogen,

this hypothesis is supported by the fact that the change of the pKa value of the nucleobase in β-D-

ddNs compared with the corresponding β-D-dNs and β-D-rNs is rather small (≤ 0.4 pD unit), as

discussed in Section 4.8.

(ii) It has also been assumed that 5'CH2OH and 5'CH2OMe exert the same (steric + gauche)

effect upon the conformation of the sugar moiety in abasic sugars 12 - 16 and in all nucleosides,

independently of their configuration at C1'. Any possible interaction (H-bonding or steric) between

the nucleobase and 5'CH2OH or 5'CH2OMe (section 4.8) in β-D-nucleosides has been neglected.

Figure 12: Possible origin for the

weaker O4'-C1'-N1/9 anomeric effect

in α-D-ddNs 17 - 20 (for the orbital

overlap in β-nucleosides, see Fig. 9,

sections 2.8 & 4). Our interpretation

is based on the fact that the anomeric

effect results from predominant nO4'

σ*C1'-N9 orbital mixing116 over

weaker electrostatic repulsions

(section 4.8). The O4' lonepairs

orbitals are represented using either

the sp2 [i.e. higher energy 1nsp2(p-

type) lonepair with predominant p-

type character and lower energy 2n

sp2(s-type) lonepair with

predominant s-type character] or sp3

[i.e. 1nsp3 and 2n

sp3 lonepairs with

the same energy] hybridization

models123-126,129,130 (section 1.9). (A) & (B) Relative orientations of sp2-hybridized O4' lonepairs with respect to the

C1'-N9 bond in N- (P = 0°, Ψm = 40°) and S-type (P = 160°, Ψm = 40°) sugars of α-D-ddA. β1/2 were calculated

assuming a perfect trigonal symetry. Neither in N- (β1 ≈ 42°) nor in S-type (β1 ≈ 6°) sugars is the 1nsp2 (p-type) orbital

antiperiplanar with the C1'-N9 bond, therefore the anomeric effect is at most weak (in favour of N-type forms!), and

clearly much weaker than in the β-counterparts (i.e. β1 ≈ 162°, Panel (C) in Fig. 9). This model is consistent with our

experimentally observed weaker anomeric effect in α-D-ddNs than in β-D-ddNs (Table 5). (C) & (D) Relative

orientations of sp3-hybridized O4' lonepairs with respect to C1'-N9 bond in the same N- and S-type sugars. The 2nsp3

lonepair and C1'-N9 bond assume an antiperiplanar orientation (β1 ≈ -144°) in the S-type form, but they are

perpendicular in N-type form. Therefore, if O4' lonepairs were sp3 hybridized, one would expect a stabilization of the S-

over N-type form to a similar extent as the stabilization of N-type sugars in β-nucleosides, which is not consistent with

our thermodynamic data for α- versus β-anomers.

(iii) The [HO3'-C3'-C4'-O4'] gauche effect has been attributed the same strength in β-D-dNs,

β-D-rNs and abasic sugars 13 and 14. We have also assumed that the [HO2'-C2'-C1'-O4'] gauche

effect has an equal strength in β-D-rNs, β-D-3'-dA (63) and abasic sugar 14, and that the [MeO3'-

OHOH2C O

HOH2C

N

N

N

N9

N

N

N

NH2

NH2

N9

OHOH2C O

HOH2C

N

N

N

N9

N

N

N

NH2

NH2

N9

C2'

H1'

N9

O4'

C4'

C2'

H1'

N9

O4'

C4'

C2'

H1'

N9

O4'

C4'

C2'

H1'

N9

O4'

C4'

(A)

1nsp2 (p-type)

σ*C1'-N9

(C)

σ*C1'-N9

(B)

β1 = 42ο

β1 = 6ο

(D) β2 = -108ο

β2 = -144ο


1nsp2 (p-type)

2nsp2 (s-type) 2

nsp2 (s-type)

1nsp2 (p-type)

1nsp3

2nsp3

1nsp3

2nsp3

2nsp3

1nsp3

1nsp3

2nsp3


78

C3'-C4'-O4'] gauche effect exerts the same influence upon the pentofuranose conformation in abasic

sugars 15, 16 and adenosine derivatives 22, 23, 38 and 39.

The implications of these assumptions are two-fold: (a) We have not taken into

consideration any possible mutual compensation of different gauche effects in β-D-dNs, β-D-rNs

and abasic sugars 13 - 16 in comparison with β-D-ddNs and 12, respectively. (b) The magnitude of

the above gauche effects has been considered independent of the nature of the nature of the C1'

substituent (either a hydrogen in abasic sugars 12 - 16 or a nucleobase in nucleosides).

4.2 Quantitation of the anomeric and gauche effects by pairwise comparisons

Owing to the above simplifying approximations, the estimates obtained from our regression

analyses can only be used for semi-quantitative purposes. In order to refine these estimates for a

specific class of compounds (or even a single compound), we have also performed a set of pairwise

comparisons between ∆H° of compounds that differ only in one structural feature, in order to assess

the stereoelectronic contribution of that particular feature to the drive of the two-state N � S

pseudorotational equilibrium of the pentofuranose moiety in the compound(s) of interest. The

strategy adopted to perform these pairwise comparisons is depicted in Fig 13, and the resulting

estimates for the anomeric and gauche effects are compiled in Tables 6 - 8. In the following

sections, all discussions are based upon the results of our pairwise comparisons.

4.3 Strengths of ∆H° and -T∆S° to ∆G°of pseudorotation of neutralnucleosides

The rather limited experimental dataset of Altona et al.316,203 showed that the populations of

N- and S-type pseudorotamers of β-D-dA and of other β -D-dNs in the N state remain virtually

unchanged as the temperature increases from 291 K (75 % S) to 333 K (73 % S). They interpreted

this result by assuming that the stabilization of S-type pseudorotamers in. these compounds does not

have an enthalpic origin, but is instead the result of the larger entropy of the S-compared to the N-

type pseudorotamers

The comparison of ∆H° and -T∆S° contributions to ∆G° of the N � S equilibrium in neutral

β-D-dNs and β-L-dNs, derived from our pseudorotational analyses of 3JHH over a wider

temperature range (278 - 358 K) than that used by Altona et al., leads us to conclude that it is the

enthalpy factor, not the entropy, that is mainly responsible for the overall conformational preference

of sugar moieties in β-D-nucleosides.

This conclusion is supported by the following observations:

(i) For β-D/L-dA, 3'-OMe-β-D-dA, 3',5'-diOMe-β-D-dA, β-D-dImb, β-D/L-dG and β-D/L-

T, ∆H° overrides -T∆S°, resulting in the overall stabilization of S-type sugars at 298 K (i.e. negative

∆G° value).

(ii) For β-D/L-dC, β-D-dU and 5-F-β-D-dU, both ∆H° and -T∆S° contributions stabilize S-

type conformations and have nearly the same magnitude. Therefore, in the N state of all β-D/L-dNs,

-T∆S° is never found stronger than ∆H°.

Chatt

opadhya

ya e

t al,

"S

tere

oel

ectr

onic

Eff

ects

in N

ucl

eosi

des

& N

ucl

eoti

des

and t

hei

r S

truct

ura

l Im

pli

cati

ons"

,

Dep

t of

Bio

org

anic

Chem

istr

y, B

ox 5

81, U

ppsa

la U

niv

ersi

ty, S

-75123 U

ppsa

la, S

wed

en, V

er 1

60205 j

yoti

@boc.

uu.s

e

79

O

OH

BP/N/D

O

OH

BP/N/D

HO

O

OH

BP/N/D

HO

OH

O

OH

BP/N

MeO

O

OMe

BP/

MeO

O

OH

MeO

BP/N

O

OH

HO

BP/N/D

O

OMe

MeO

BP/N

O

OH

BP/N/D

O

OH

HO

O

OH

HO

OH

O

OH

MeO

O

OMe

MeO

O

OH

BP/N/D

-1/-2HO3PO

OH

O

OH

BP/N/D

-1EtO

3PO

OH

O

OH

BP/N/D

-1/-2HO3PO

O

OH

BP/N/D

-1EtO

3PO

O

OMe

BP/N/D

O

OH

BP/N/D

OH

O

OH

C-B

P/N/D

HO

OH

O

OH

ΔΔ

Ho

17

ΔΔ

Ho

18

ΔΔ

Ho

13

ΔΔ

Ho

14

ΔΔ

Ho

2ΔΔ

Ho

3ΔΔ

Ho

4ΔΔ

Ho

5ΔΔ

Ho

6

ΔΔ

Ho

30

ΔΔ

Ho

15

ΔΔ

Ho

16

ΔΔ

Ho

9

ΔΔ

Ho

10

ΔΔ

Ho

24

ΔΔ

Ho

20

ΔΔ

Ho

23

ΔΔ

Ho

26

ΔΔ

Ho

28

ΔΔ

Ho

8

ΔΔ

Ho

11

ΔΔ

Ho

27

ΔΔ

Ho

29

ΔΔ

Ho

1d

ΔΔ

Ho

1a

ΔΔ

Ho

1b

ΔHo

1c

17 -

20

23

12

13

14

15

16

30, 31 &

33 -

35

50 -

55

38

63

56 -

62

69 -

73

74 -

78

79 -

83

37, 41 -

45

21 &

24 -

26

22

39

32

64 -

68

ΔΔ

Ho

12

ΔΔ

Ho

7

Fig

ure

13

. E

stim

ates

for

the

anom

eric

and

gau

che

effe

cts

in 1

2 -

83 f

rom

pai

rwis

e co

mpar

isons

Chatt

opadhya

ya e

t al,

"S

tere

oel

ectr

onic

Eff

ects

in N

ucl

eosi

des

& N

ucl

eoti

des

and t

hei

r S

truct

ura

l Im

pli

cati

ons"

,

Dep

t of

Bio

org

anic

Chem

istr

y, B

ox 5

81, U

ppsa

la U

niv

ersi

ty, S

-75123 U

ppsa

la, S

wed

en, V

er 1

60205 j

yoti

@boc.

uu.s

e

80

Tab

le 6

. E

stim

ates

(∆

∆H°)

for

the

mag

nit

udes

of

the

effe

cts

of

the

nucl

eobas

es a

nd v

ario

us

gau

che

effe

cts

that

dri

ve

the

sugar

confo

rmat

ion i

n

com

pounds

12 -

16 a

nd i

n α

- an

d β

-N-n

ucl

eosi

des

17 -

63 f

rom

pai

rwis

e co

mpar

isonsa

Nucl

eobas

e A

G

Im

C

T

U

5-F

-U

Sta

te

P

N

P

N

D

P

N

P

N

N

D

N

D

N

D

∆∆

H°1

b

0.4

1

.0

- -

- -

- -

- -

- -

- -

-

∆∆

H°1

c -

- -0

.2

0.9

0

.9

- -

- -

- -

- -

- -

∆∆

H°1

d

0.5

0

.6

- -

- -

- -

- -

- -

- -

-

∆∆

H°2

b

8

.8

3.1

2

3.2

3

.0

1.0

-

- 9

.2

6.2

5

.0

3.1

5

.3

- -

-

∆∆

H°3

3

.4

0.2

6

.2

1.3

-0

.8

4.2

1

.9

4.1

3

.4

2.7

2

.2

3.5

2

.8

3.3

3

.0

∆∆

H°4

-0

.6

-4.8

5

.0

-3.7

-8

.0

- -

4.8

1

.9

0.9

-0

.6

1.6

-0

.1

1.9

0

.4

∆∆

H°5

3

.2

0.0

-

- -

- -

- -

- -

- -

- -

∆∆

H°6

3

.2

0.6

-

- -

- -

- -

- -

- -

- -

∆∆

H°1

0

-9.9

-7

.4

-21

.5

-6.2

-6

.3

- -

-9.6

-7

.3

-6.8

-5

.4

-6.3

-

- -

∆∆

H°1

1

0.5

-0

.5

3.3

-0

.5

-2.7

-

- 5

.2

3.0

2

.7

1.7

2

.6

1.6

3

.1

1.9

∆∆

H°1

2

-0.7

-0

.7

- -

- -

- -

- -

- -

- -

-

∆∆

H°1

3

-2.1

-2

.1

-9.1

-0

.8

-0.8

-

- -3

.3

-3.3

-1

.5

-0.9

-

- -

-

∆∆

H°1

4

-0.9

-0

.9

-6.2

-0

.4

0.7

-

- -3

.0

-3.0

0

.1

0.1

-

- -

-

∆∆

H°1

5

-1.2

-1

.8

- -

- -

- -

- -

- -

- -

-

∆∆

H°1

6

-1.1

-1

.6

- -

- -

- -

- -

- -

- -

-

∆∆

H°1

7

-3.3

-3

.3

-2.0

-4

.1

-3.0

-

- -4

.2

-4.2

-2

.9

-3.5

-

- -

-

∆∆

H°1

8

-0.8

-1

.4

- -

- -

- -

- -

- -

- -

-

∆∆

H°1

9

-7.1

-5

.7

- -

- -

- -

- -

- -

- -

-

a

A,

G,

Im,

C,

T,

U a

nd

5-F

-U r

epre

sent

aden

in-9

-yl,

guan

in-9

-yl,

im

idaz

ol-

1-y

l, c

yto

sin-1

-yl,

thym

in-1

-yl,

ura

cil-

1-y

l an

d 5

-flu

oro

-ura

cil-

1-y

l, r

esp

ecti

vel

y.

P,

N a

nd

D d

eno

te

the

pro

tonat

ed,

neu

tral

and

dep

roto

nat

ed s

tate

s o

f 1

7 -

63

, re

spec

tivel

y.

All

est

imat

es (

kJm

ol-1)

wer

e ca

lcula

ted

by s

ub

trac

ting

ΔH°

, Δ

H° N

or Δ

H° D

val

ues

(T

able

1)

of

two

com

po

und

s am

ong 1

2 -

63

, ac

cord

ing t

o t

he

stra

tegy d

epic

ted

in F

ig 1

3.

∆∆

H° 1

a (0

.4 k

Jmo

l-1),

∆∆

H° 7

(-4

.5 k

Jmo

l-1),

∆∆

H° 8

(4

.5 k

Jmo

l-1)

and

∆∆

H° 9

(-0

.5 k

Jmo

l-1)

are

const

ant

fact

ors

cal

cula

ted

by s

ub

trac

ting Δ

H° N

of

abas

ic s

ugar

s 1

2 -

16

. b

∆∆

H° 2

w

as a

lso

cal

cula

ted

by s

ub

trac

ting Δ

H° N

o

f 1

2 f

rom

ΔH° N

of β

-D-d

dI

(36

) (5

.8 k

Jmo

l-1).



81

(iii) In the N state of β-D-rT, β-D-U and 5-F-β-D-rU, opposing ∆H° (driving to the N) and -

T∆S° nearly cancel each other, resulting in nearly unbiased equilibrium. However, in the N states of

β-D-G and β-D-A, ∆H° (stabilizing S-type sugars) prevails over the opposing -T∆S°, and S-type

sugars are favoured at 298 K. In the N state of β-D-C, ∆H° also prevails over counteracting -T∆S°,

but it stabilizes N-type sugars at 298 K.

Table 7. Quantitation of various gauche effects (∆∆H°) driving the sugar conformation in

mononucleotides 64 - 83 from pairwise comparisonsa

Nucleobase adenin-9-yl guanin-9-yl cystosin-1-yl thymin-1-yl uracil-1-yl

∆∆H°20 -8.9 -7.5 -9.8 -8.0 -8.4

∆∆H°21 -9.0 -8.2 -10.2 -8.4 -8.4

∆∆H°22 -1.5 -1.3 -2.5 -1.2 -2.1

∆∆H°23 -1.6 -2.0 -2.9 -1.6 -2.1

∆∆H°24 -0.1 -0.7 -0.4 -0.4 0.0

∆∆H°25 -0.5 -1.2 -1.5 -2.2 -0.8

∆∆H°26 -2.5 -2.5 -3.8 -3.8 -3.6

∆∆H°27 -1.4 -1.0 2.1 0.5 1.1

∆∆H°28 -2.0 -1.3 -2.3 -1.6 -2.8

∆∆H°29 0.5 -0.4 4.0 1.7 3.9

a All estimates (kJmol-1) have been calculated by subtracting values ΔH°N (Table 2 & 3) of two compounds among β-D-

ddNs 30 , 31 and 33 - 35, β-D-dNs 37 and 41 - 44, β-D-rNs 50 - 54 and mononucleotides 64 - 83 according to the

nomenclature shown in Fig 13.

Table 8. Quantitation of the enthalpy (∆∆H°30) of the (steric + stereoelectronic) effect of

the C-nucleobase, in each of its protonated (P), neutral (N) and deprotonated (D) states,

driving the sugar conformation in C-nucleosides 64 - 83a

Compound P N D

Formycin B (56) -0.9 -8.5 -9.2

Formycin A (57) -2.8 -8.5 -8.5

9-deaza-A (58) -7.8 -14.6 -

Ψ-isoC (59) 3.8 -2.3 -8.3

Ψ-U (60) - 0.3 -4.9

1-Me-Ψ-U (61) - 1.2 -3.4

1,3-diMe-Ψ-U (62) - 1.6 -

a ∆∆H°30

(kJmol-1) has been calculated by subtracting ΔH°N (0.4 kJmol-1, Table 2) of abasic sugar 14 from

those of a particular C-nucleoside among 56 - 62.

(iv) The situation is far less complicated in neutral β-D-ddNs 30 - 36, since for all of them,

∆H° (driving to N) always outweighs the counteracting -T∆S° contribution, which results in a high

preference for N-type pseudorotamers at room temperature. For abasic sugars 13, 15 and 16, the

two-state N � S pseudorotational equilibrium is driven toward S-type conformations mainly by


82

∆H°, whereas for 12 ∆H° and -T∆S° contributions cancel each other. In contrast, in 14, the

predominance of N-type conformations results from the cooperative drive of the prevailing -

T∆S° contribution and the minor ∆H° term.

4.4 Nucleobase-dependent anomeric effects in neutral nucleosides

4.4.1 Effect of the 5'CH2OH versus 5'CH2OMe

The strengthening of the overall effect of 5'CH2OMe moiety in abasic sugar 16 in

comparison with 5'CH2OH in 15 is evident from the subtraction of the ΔH°N value of 15 from that

of 16 (i.e. ∆∆H°1a = 0.4 kJmol-1, Fig 13 and Table 6). The comparison of the effect of 5'CH2OMe

in nucleosides 39 and 32 with the influence of 5'CH2OH in their counterparts, 38 and 31 supports

the results of regression analysis (A) regarding the slightly stronger ∆H°(5'CH2OMe) with respect to

∆H°(5'CH2OMe) since ∆∆H°1b (1.0 kJmol-1) and ∆∆H°1c (0.9 kJmol-1) in neutral nucleosides are

indeed slightly larger than ∆∆H°1a for abasic sugars.

4.4.2 The effect of the nucleobase is sugar-dependent

In β-D-ddNs 30, 31 and 33 - 36, an estimate for the overall effect of the constituent

nucleobases can be easily obtained by subtracting (∆∆H°2) the contribution of the 5'CH2OH effect

(i.e. reflected in the experimental ΔH°N value of 12) from experimental ΔH°N values of 30, 31 and

33 - 36 (Table 6). The comparison of ∆∆H°2 values shows that pyrimidines are able to shift the

pseudorotational equilibrium in β-D-ddNs toward N-type conformations more efficiently than

purines, as already suggested by the results of the regression analysis (A), since ∆∆H°2 increases in

the order: guanin-9-yl (3.0 kJmol-1) ≈ adenin-9-yl (3.1 kJmol-1) << thymin-1-yl (5.0 kJmol-1) ≈

uracil-1-yl (5.3 kJmol-1) < cytosine-1-yl (6.2 kJmol-1) .

In β-D-dNs 37 - 45, beside the effect of the nucleobase, both the influence of 5'CH2OH and

the gauche effect of [HO3'-C3'-C4'-O4'] fragment contribute to the drive of the sugar conformation.

Therefore, in order to obtain an estimate for the overall effect of the nucleobase in this series, it is

necessary to subtract (∆∆H°3 in Table 6) from their experimental ΔH°N values an estimate for the

combined effects of the 5'CH2OH and 3'-OH gauche effect. Such an estimate has been found in the

experimental ΔH°N value of abasic sugar 13 (-4.1 kJmol-1). The comparison of ∆∆H°3 values for

the nucleobases in β-D-dNs shows that, as in the β-D-ddNs series, pyrimidines exert a stronger

influence than for purines, i.e. ∆∆H°3 values are in the order: adenin-9-yl (0.2 kJmol-1) < guanin-9-

yl (1.3 kJmol-1) < imidazol-1-yl (1.9 kJmol-1) < thymin-1-yl (2.7 kJmol-1) < 5-fluorouracil-1-yl

(3.3 kJmol-1) ≈ cytosin-1-yl (3.4 kJmol-1) ≈ uracil-1-yl (3.5 kJmol-1). The fact that the effect of

imidazolyl-1-yl is slightly stronger than that of adenin-9-yl and guanin-9-yl is consistent with its

reduced steric bulk in comparison with adenin-9-yl and guanin-9-yl. The subtractions of ΔH°N of

abasic sugar 15 from that of 3'-OMe-β-D-dA (38) (i.e. ∆∆H°5 = 0.0 kJmol-1), on one hand, and of

ΔH°N of 16 from that of 3',5'-diOMe-β-D-dA (39), on the other (i.e. ∆∆H°6 = 0.6 kJmol-1) shows

that the effect of adenin-9-yl is the same in 38, 39 and β-D-dA (37) (0.2 kJmol-1, see above) itself.



83

For each nucleobase, the comparison of ∆∆H°2 with ∆∆H°3 suggests that in β-D-dNs the effect of

the nucleobase is much weaker than in the β-D-ddNs counterparts. The weaker anomeric effect in

dNs compared to ddNs is explained as follows: (i) The electron-withdrawing character of 3'-OH

reduces the electron density around O4', making O4' lonepair less available for nO4'

→σ∗C1'N1/9

interactions in β-D-dNs compared with β-D-ddNs (Fig 14). If the nucleobase is

substituted for a hydrogen atom, no anomeric effect is operating, therefore the electron density

around O4' is maximal and as a result σC3'H3'

→σ∗

C4'O4' interactions are weaker (because of higher

difference between their energy levels) in abasic sugar 13 compared to β-D-dNs. Similarly as 3'-OH

in dNs is substituted for 3'-phosphate, the 3'-gauche effect is enhanced. This means that the strength

of the 3'-gauche effect increases as an electron-withdrawing group is placed at either C1' or C3'

(Section 4.5). (ii) Conversely, σC3'H3'

→σ∗

C4'O4' interactions in β-D-dNs may result in the lowering of

O

HOH2C

O

HOH2C

OH

OH

N

N

N

NH2

N9

NN

N

NH2

N9

H3'

H3'

σ*C4'O4'

σC3'H3'σ*C4'O4'

σC3'H3'

σ*C1'-N


1nsp2 (p-type)

1nsp2 (p-type)

σ*C1'-N9

ΔE1 σC3'H3'

σ*C4'O4'

ΔE2

ΔE(GE)

ΔE(AE)

1nsp2 (p-type)

ΔE

___________

__

___________________________________________________________

___________ E1

E2

E3

E4

E5

E6___________________________________

The relative energies of orbitals (E1 - E6) are based on their relative donor-acceptor abilities, and they are constantly modulated by the nature of eachsubstituent at the sugar as well as on the aglycone in free, ionic and complexe state

Since nO4' is a better donor than σC3'H3' and σ*C4'O4' is a better acceptor than σ*C1'N9, therefore ΔE(GE) > ΔE(AE)

(A)

(B)

Figure 14. The interplay of the O4'-C1'-N9 anomeric effect and the [HO3'-C3'-C4'-O4'] gauche effect in β-D-dNs drives the N

� S equilibrium (the 2'-gauche effect in rNs is also a participant in this stereoelectronic interplay, but it is not shown here in

order to retain clarity in the diagram). (A) 1nsp2 →σ∗C1'N1/9 orbital overlap favours N-type sugars (i.e. anomeric effect),

whereas σC3'H3' →σ∗C4'O4' interactions (i.e. [HO3'-C3'-C4'-O4'] gauche effect) stabilize S-type conformations. The

differences in the electronic characters of the substituent at C1', C2' and C3' however decide the donor-aceptor capabilities of

the occupied and vaccant orbitals (i.e. the interplay of the gauche and anomeric effects) (B) The energy difference between

σC3'H3' and σ∗C4'O4' (∆E1) is smaller (hence a better mixing of the orbitals) than that between 1nsp2 and σ∗C1'N1/9 (∆E2),

therefore the gauche effect is more efficient than the anomeric effect This however is dictated on a case-to-case basis by the

nature of a set of substituents at the endocyclic carbons that are present on the sugar ring in a specific nucleoside. The reason

why we do not take 5'CH2OH into consideration is because it is viryually a free rotor where steric and gauche effect have been

found to have negligible contribution to the drive of the sugar conformation (vide supra).


84

the energy level of the nO4' orbital. Since the efficiency of nO4'

→σ∗C1'N1/9

interactions is inversely

proportional to the difference between their energies, the anomeric effect is weakened in the dNs

compared to ddNs. This means that the modulations of the gauche effect by the anomeric effect, and

of the anomeric effect by the gauche effect are mutually interrelated and interdependent.

In β-D-rNs, the sugar conformation is driven by the interplay of the following

stereoelectronic effects: (a) the effect of the nucleobase, (b) the effect of 5'CH2OH, (c) the gauche

effect of [HO2'-C2'-C1'-O4'], (d) the gauche effect of [HO3'-C3'-C4'-O4'] and (e) the gauche effect

of [HO2'-C2'-C1'-N1/9]. The [HO3'-C3'-C2'-O2'H] gauche effect is negligible, since in N- and S-

type sugars, O3'-C3' is gauche with respect to O2'-C2'. In order to quantitate the effect of the

nucleobase in β-D-rNs, we subtracted from their ΔH°N values the contributions from the effects

detailed in (b) - (e). The combined influence of 5'CH2OH and of the [HO3'-C3'-C4'-O4'] and [HO2'-

C2'-C1'-O4'] gauche effects can be accounted for by subtracting (∆∆H°4) ΔH°N of 14 from that of a

β-D-rN. ∆∆H°4 values increase in the following order: adenin-9-yl < guanin-9-yl < thymin-1-yl <

uracil-1-yl ≈ 5-fluorouracil-1-yl ≈ cytosin-1-yl. In order to obtain an estimate for the overall effect

of the nucleobase in each β-D-rN, it was then necessary to subtract the strength of the [O2'-C2'-C1'-

N1/9] gauche effect from ∆∆H°4. Estimates for the strength of the [O2'-C2'-C1'-N1/9] gauche effect

(∆H°GE[O2'-C2'-C1'-N1/9]) in neutral β-D-rNs (Section 4.6) are as follows: -7.9 kJmol-1 in β-D-A, -

6.7 kJmol-1 in β-D-G, -4.3 kJmol-1 in β-D-C, -4.1 kJmol-1 in β-D-rT and -3.7 kJmol-1 in β-D-U.

The subtraction of ∆H°GE[O2'-C2'-C1'-N1/9] from ∆∆H°4 gives the effect of each nucleobase (AErN)

in β-D-rNs (kJmol-1): 3.0 in β-D-G ≈ 3.1 in β-D-A < 5.0 in β-D-rT ≈ 5.3 in β-D-rU < 6.2 in β-D-C.

The comparison of ∆∆H°2 and AErN values in the neutral state shows that the effect of a nucleobase

is the same in β-D-ddNs and β-D-rNs, owing to the cancellation of the [HO2'-C2'-C1'-O4'] and

[HO3'-C3'-C4'-O4'] gauche effects in the latter.

The fact that the overall effect of the nucleobase in β-D-nucleosides is characteristic of its

electronic and steric nature has three important implications in understanding structure-activity of

nucleic acids:

(i) The change of the electronic character of purine or pyrimidine heterocycles upon their

protonation or deprotonation is expected to result into the modulation of the anomeric effect, which

in turn should be reflected in the shift of bias of the two-state N � S equilibrium toward more N-

type sugars (upon protonation owing to the enhanced anomeric effect) or S-type pseudorotamers

(upon deprotonation due to the weaker anomeric effect). Qualitative studies on the preferred

conformation of the sugar moiety in various adenosine, guanosine and cytidine nucleosides and

nucleotides as a function of pD, as discussed in Section 2.9 are consistent with this simple

reasoning. We, on the other hand, have quantitated for the first time the actual magnitude of the

modulation of the anomeric effect in α- and β-nucleosides upon the change of the pD of the aqueous

solution (Sections 4-6).

(ii) Divalent metal cations are essential in natural processes involving nucleic acids, such as

the replication, transcription and translation of the genetic code479. In the presence of Mg2+

cofactor, the EcoRV restriction endonuclease is able to cleave DNA at one particular recognition



85

site with high specificity, whereas when Mg2+ is replaced by Mn2+, both the activity and specificity

of the enzyme are dramatically reduced. Numerous studies have been undertaken to establish the

sites of coordination of metal ions in nucleosides and nucleotides as well as the relative stabilities of

the complexes247,391,393,480-495. Whereas hard metal ions such as Mg2+ mainly bind to the phosphate

oxygen moiety of nucleotides, there is also spectroscopic evidence (NMR and UV) that many

transition metal ions such as Mn2+, Co2+, Ni2+, Cu2+ coordinate to both the phosphate oxygens and

the nucleobase nitrogens of adenosine and guanosine nucleotides either directly or via a coordinated

water molecule487. At the other end of the affinity scale, softer Pt in anticancer drug cis-

Pt(NH3)2Cl2 binds to DNA, by preferential coordination at N7 of the constituent adenin-9-yl and

guanin-9-yl nucleobases393. A slow titration with HgClO4 has shown that the binding of Hg2+ to N3

of thymin-1-yl in the AT-tract of DNA causes the disruption of the Watson-Crick base-pairing, and

is therefore responsible for a conformational transition from a B-type DNA to a new bulge-

containing conformer. As in the case of protonation (Section 4.8), one expects that this metallation

at N7 or N3 will enhance the strength of the anomeric effect of adenin-9-yl, guanin-9-yl or thymin-

1-yl which in turn should be reflected into an increased preference of the sugar moieties in the

respective adenosine, guanosine and thymidine residues for N-type pseudorotamers. The results of

recent studies on the conformational changes induced by the interaction of metal ions with the

nucleobase and the phosphate moieties in guanin-9-yl nucleotides are discussed in Section 8.7.

(iii) The overall effect of a N-nucleobase in a nucleoside upon the conformation of the

constituent sugar moiety consists of the counteracting contributions of (i) stereoelectronic

interactions within O4'-C1'-N1/9 fragment and (ii) the counteracting effect of the nucleobase.

Therefore, to obtain an estimate for the stereoelectronic anomeric effect it is necessary to subtract

from the overall effect of the nucleobase the steric component. Since it is possible to modulate the

relative importance of both contributions by changing the electronic character of the nucleobase, it

will be possible to engineer a nucleoside in which the nucleobase will steer the sugar conformation

through its inherent steric effect alone, and such a nucleoside can therefore be used as a reference

point further on (see our pD-dependent conformational studies on C-nucleosides in Section 6.1

which we used as the basis for the quantitation of the anomeric effect of adenin-9-yl in β-D-dA (37)

and β-D-A (50) and of guanin-9-yl in β-D-dG (41) and β-D-G (51)).

4.5 The gauche effect of the 3'-substituent in neutral dNs

4.5.1 The influence of the C1'-substituent

The subtraction of ΔH°N of abasic sugar 12 from that of 13 yields an estimate for the

strength of the [HO3'-C3'-C4'-O4'] gauche effect in 13 (∆∆H°7 = -4.5 kJmol-1 in Table 6). In order

to compare the magnitude of the [HO3'-C3'-C4'-O4'] gauche effect in 13 and in β-D-dNs, we have

also subtracted ΔH°N of each β-D-ddN 30, 31 and 33 - 35 from that of each β-D-dN counterpart 37

and 41 - 44 (∆∆H°10 in Table 6). ∆∆H°10 values are within the range from -7.4 kJmol-1 to -6.2

kJmol-1, showing that the [HO3'-C3'-C4'-O4'] gauche effect is much more efficient in β-D-dNs than

in abasic sugar 13. This suggests that as O4' is involved in nO4' →σ∗

C1'-N1/N9 stereoelectronic


86

interactions, the ability of O4'-C4' bond to adopt a gauche orientation with respect to O3'-C3' is

much greater than in 13. It is however difficult to establish a straightforward quantitative correlation

between the magnitude of the effect of the nucleobase and the corresponding strength of the [HO3'-

C3'-C4'-O4'] gauche effect in β-D-dNs 37 and 41 - 44. The small ∆∆H°9 value (-0.5 kJmol-1) on the

other hand, shows that the [MeO3'-C3'-C4'-O4'] gauche effect in 15 is only slightly more efficient

than the [HO3'-C3'-C4'-O4'] gauche effect in 13. The conclusion remains the same in the β-D-dNs

series, as shown by the small ∆∆H°12 value (-0.7 kJmol-1) reflecting the small additional stabilition

of S-type conformations in 3'-OMe-β-D-dA (38) in comparison with β-D-dA (37).

4.5.2 3'-substituent electronegativity dictates 3'-gauche effect

A comparative conformational analysis of 2'-deoxy-2'-substituted uridine235,199 and adenosine

206,417,418 derivatives has afforded linear relationships between the population of the N-type

pseudorotamer and the electronegativity of the 2'-substituent (Section 2.11). Various hypotheses

regarding the possible origin of the gauche effects in nucleos(t)ides have been discussed (See

Section 1.12-1.14 for the general discussion about the gauche effect in any system). We have

mentioned in above that substitution of 3'-OH in abasic sugar 15 and in β-D-dA (37) for 3'-OMe in

16 and 38 results in a slight increase in the preference of their sugar moieties for S-type

conformations (as experimentally evidenced by the slightly negative ∆∆H°9 and ∆∆H°12 values,

respectively, Table 6). This has been attributed to the modulation of the [MeO3'-C3'-C4'-O4']

gauche effect compared to [HO3'-C3'-C4'-O4'] gauche effect in the latter compared to the former.

The next logical step consisted in addressing the following questions: (i) Is it possible to

modulate the thermodynamics of the N � S equilibrium in a particular 3'-substituted-β-D-ddN (X =

3'-substituent) simply by altering the nature of its 3'-substituent and conversely (ii) is it possible to

predict quantitatively the strength of the [X3'-C3'-C4'-O4'] gauche effect driving the conformation

of any 3'-substituted-β-D-ddN using a simple procedure?

These questions have led us to perform a systematic comparative study26

on the preferred

conformations of the pentofuranose moieties in a series of 3'-substituted-β-D-ddT derivatives [X =

H (34), NH2 (113), OH (43), OMe (114), NO2 (115), OPO3H- (67), F (88)] with the same thymin-9-

yl nucleobase. These 3'-substituted-β-D-ddT derivatives only differ in the nature of 3'-X, which

varies from poorly (i.e. NH2 in 113) to strongly electronegative (i.e. F in 88). 3'-NH2 in 113 should

be half-protonated at ca. pD 7.0 since its pKa is expected to be very similar to that of 2'-NH2 in the

respective nucleoside (for instance, the pKa value of 2'-NH2 in 2'-NH2-β-D-dU has been reported to

be 6.2496,497).

The pseudorotational analyses of temperature-dependent 3JHH extracted from 1H-NMR

spectra of 88 and 113 - 115 and subsequent estimation of ΔH°N, ΔS°N and ΔG°N values of their N �

S equilibria have been performed using the same procedure as for β-D-ddT (34), β-D-T (43) and β-

D-TMP (67) (Section 3.1-3.7, Table 9 for ΔH°N, ΔS°N and ΔG°N values of 43, 67, 88 and 113 - 115.

From the initial comparison of ΔH°N, ΔS°N and ΔG°N values of 34, 43, 67 and 88, 113 - 115, the

following conclusions can be drawn: (i) ΔH°N is the main determinant to the drive of the sugar

conformation since it prevails over the counteracting (in 34, 43, 67 and 88) or cooperative (in all



87

other compounds) -ΔS°N contribution to ΔG°N. (ii) In β-D-ddT (34), the N � S equilibrium is

driven almost exclusively toward N-type conformations (16 % S at 278 K) by the cooperative

effects of the nucleobase and of 5'CH2OH, owing to the absence of any electron-withdrawing

substituent at C3'. (iii) In

contrast, the regularly

increasing stabilization of S-

type pseudorotamers in the

order 3'-NH2-β-D-T (113, 20

% at 278 K), β-D-T and β-D-

TMP (43 and 67, 64 % at 278

K), 3'-OMe-β-D-T (114, 74

% at 278 K), 3'-NO2-β-D-T

(115, 81 % at 278 K), 3'-F-β-

D-T (88, 91 % at 278 K), can

be attributed unambiguously

to the increasing preference

for gauche orientation within

[X3'-C3'-C4'-O4'] fragment,

which progressively prevails

over the anomeric effect.

ΔH°N reflects the strength of

the combined influence of the

(steric + stereoelectronic)

contributions of this gauche

effect, increases in the order:

NH2 (113) < OH (43) < OMe

(114) < NO2 (115) < OPO3H-

(67) < F (88).

The [X3'-C3'-C4'-O4'] gauche effect (ΔH°GE) in 43, 67, 88 and 113 - 115 has been

subsequently estimated by subtracting from their ΔH°N values that of β-D-ddT (34) (Table 9), which

allows to account for the combined effects of thymin-9-yl and of 5'CH2OH groups in 43, 67, 88 and

113 - 115. It should be however noted that ΔH°GE values consist of three intractable components: (i)

the stereoelectronic gauche effect, (ii) a term for the electrostatics and solvation, and (iii) the steric

effect of the 3'-substituent.

We have examined the dependence of ΔH°GE on the group electronegativity498-500,501 of 3'-X

using the electronegativity scales developed by Marriott (χMarriott)502, Mullay (χMullay)503,504 and

Inamoto (ιInamoto)505-508 and their coworkers. Marriott's electronegativity scale is based on ab initio

calculations with the GAUSSIAN program298 (at HF/6-31G* level). The actual electronegativity of

X in H-X corresponds to the atomic electron population on H in H-X, as determined from a

Mulliken populations analysis of geometrically optimized H-X. Mullay's group electronegativities

(I)

ΔHo

GE (kJ mol

-1)

-12 -10 -8 -6 -4 -2

Group electronegativity of 3'-substituent χ or τ

0

1

2

3

4

5

(III)

(II)

88

116

114

43

67

115

113

88

116

114

43

67

115

113116

88 114

43

67

115

113

Figure 15: Correlation plots of the group electronegativity of the 3'-X substituent

in 3'-substituted-β-D-ddT derivatives [X = H (34), NH2 (113), OH (43), OMe

(114), NO2 (115), OPO3H- (67), F (88)] as a function of the strength (ΔH°GE) of

the [X3'-C3'-C4'-O4'] gauche effect. The group electronegativity has been

expressed in the scales developed by Mullay (χMullay)503,504 [....., graph (I)],

Inamoto (ιInamoto)505-508 [___, graph (II)] and Marriott (χMarriott)502 [----,

graph (III)]. All plots show straightlines with high Pearson's correlation

coefficients [R = -1.0, -0.98 and -0.96 for graphs (I) - (III), respectively]. In a

recent study42, we have determined the thermodynamics of the N �S

equilibrium [∆H° = 0.0 kJmol-1 (σ = 0.4) , ∆S° = 7.5 Jmol-1K-1 (σ = 1.5) and

∆G° (298 K) = -2.3 kJmol-1 (σ = 0.6)] in 3'-OCF3-β-D-ddT (116) using the

procedure described in Section 3.1-3.7]. ΔH°GE (-5.4 kJmol-1) of the [OCF3-C3'-

C4'-O4'] gauche effect in 116 was subsequently calculated by subtracting ∆H° of

β-D-ddT (34) from its own. From the equations of the straightlines (I) [χMullay

= 2.63 - 0.19ΔH°GE], (II) [ιInamoto = 2.25 -0.08 ΔH°GE] and (III) [χMarriott =

0.26 - 0.022 ΔH°GE], the group electronegativity of 3'-OCF3 in 116 has been

determined: χMullay = 3.66 ± 0.05, ιInamoto = 2.64 ± 0.10 and χMarriott = 0.38

± 0.05.


88

are based on modified Slater effective nuclear charges, effective principal quantum numbers,

fractional p characters, assuming charge conservation and electronegativity equalization within each

bond in the group. Inamoto's scale ι is an inductive substituent parameter scale, which is derived

from the corresponding group electronegativities showing a strong correlation with trans H,H

coupling constant in monosubstituted ethene fragments. The plots of the ΔH°GE values as a function

of χMarriott502, χMullay

503,504 and ιInamoto505-508 gave all straight lines with high correlation

coefficient (> 0.96) (Fig 15). From the equations of these simple correlation plots, we have

subsequently been able to estimate the group electronegativity of 3'-OPO3H- in β-D-TMP (67), i.e. :

χMarriott = 0.44, χMullay = 4.12 and ιInamoto = 2.8.

Table 9. The enthalpy and entropy contributionsa to the N � S pseudorotational equilibrium of the

pentofuranose moiety in 3'-substituted-β-D-ddT derivatives 43, 67, 88 and 113 – 115 (see ref 26)

Compound Group electronegativity (χ or ι) of

3'-substituent d

ΔH°N ΔS°N −ΤΔS°N ΔG°N Δ%S b

358-278 ΔH°GE

c

Marriott Mullay Inamoto

3'-NH2-ddT (113) 2.6 -1.9 0.6 3.2 +5 -2.8 0.33 3.15 2.47

T (43) -1.8 -0.9 0.3 -1.5 -4 -7.2 0.43 3.97 2.79

3'-Ome-ddT(114) -2.1 1.1 -0.3 -2.4 -4 -7.5 0.44 4.03 2.82

3'-NO2-ddT (115) -2.4 3.7 -1.1 -3.5 -3 -7.8 0.4 4.08 2.75

TMP (67) -2.6 -4.3 1.3 -1.3 -3 -8.0 0.44

e

4.12e 2.8

e

3'-F-ddT (88) -5.9 -2.3 0.7 -5.2 -6 -11.3 0.52 4.73 3.1

a In kJmol-1. ΔS°N and ΔG°N are given at 298K (see the methodology described in Section 3). b Δ %S (358-278)

indicates the change in the population of the S-type pseudorotamer as the temperature is raised from 278 K to 358 K,

and shows the influence of both the gauche effect enthalpy and entropy contributions. At a particular temperature T, the

population of the S-type species is calculated from the corresponding free energy values as follows: %S (T) = 100 * [exp

(- ΔGT/ RT)] / [exp (- ΔGT/ RT) +1]. c ΔH°GE has been calculated by subtracting ΔH°N characterizing the N �S

equilibrium of the sugar moiety in 43, 67, 88 and 113 - 115 from that obtained for β-D-ddT (34) (Table 2). d

The

group electronegativities of the 3'-substituents NH2 (113), OH (43), MeO (114), NO2 (115) and F (88) are given

according to ref. 502-508. e For β-D-TMP (67), the group electronegativities of the OPO3H- substituent have

been back-calculated from its ΔH°GE value using the graphs shown in Fig 15.

S-type and the N-type pseudorotamer populations. Our NMR analysis of a set of 3′-substituted

thymidine derivatives [Scheme 1: for 1 – 7]26

, in which the gauche effect of 5′-CH2OH and the

anomeric effect of the nucleobase remain as the constant factor, has clearly shown that the

conformational preference for the S-type sugar conformation linearly increases with the increase of

the strength of the electronegetivity of the 3′-substitutent. It has been envisioned that as the



89

electronegetivity of the 3′-substituent increases [H < NH2 < OCF3 < OH < OCH3 < NO2 < F] the

C3′-H3′ becomes further polarised, the difference in the energy levels of the donor σC3′-H3′

and

acceptor σ*

C4′-O4′

orbitals decreases, facilitating the σC3′-H3′

�σ*

C4′-O4′

orbital mixing, thereby

enhancing the 3′-gauche effect promoted S-type conformational preference.

We argued that since we already have a dependable experimental means for the estimation

of K298

NMRG° Δ

ref, we could challenge the theory with this by calculating the total energy difference

(ΔE) of N-type and S-type geometries (ΔES-N) of 1 – 7 (Figure 1) using high level ab initio

calculations. The Gaussian calculations of both N- and S-type pseudorotamers of potential anti-HIV

2′,3′-dideoxy 3′-substituted thymidine derivatives 1 – 7 have been performed using systemetic

variation of different basis functions at Hatree-Fock level. With the lower basis functions such as 6-

31G* or even with 6-311+G* basis sets (Table 1), we failed to observe any colinearity of the ab

initio calculated ΔES-N in the gas phase with our K298

NMRG° Δ . This pilot study showed the necessity of

introducing diffuse function with higher basis set in order to reduce the basis set superposition

error6, which also illustrated the caveats of using a low-level ab initio method for structural

calculation of flexible biomolecules.

The use of higher basis set, such as HF/6-311++G**, showed a marked improvement of

correlation of ΔES-N with the experimental K298

NMRG° Δ for 1 – 7 (Table 1). Thus a plot of ΔES-N (gas

phase HF/6-311++G**) as a function of K298

NMRG° Δ showed a Pearson’s correlation coeffecient of 0.92

(Graph I in Figure 1A). In order to mimick the solvation behaviour of the experimental NMR

analyses, the solution phase (ε = 78.0) ab initio calculations was performed at HF/6-311++G** level

using Onsager solvation model. Remarkably, this gave even better correlation compared to the gas

phase calculations as evidenced from the Graph II in Figure 1A, giving Pearson’s correlation

coeffecient of 0.97, as a result of improved colinearity of the energies from the solution phase ab

initio calculations with the experimental NMR data. Moreover, the single point ab initio

calculations at B3LYP/6-311++G** level of theory in solution phase with Onsager solvation model

have been performed using the optimised geometry at HF/6-311++G** level for 1 – 7 (Table 1).

The plot of ΔES-N calculated from this solution phase B3LYP/6-311++G**//HF/6-311++G**

calculation as a function of experimental K298

NMRG° Δ gives the Pearson’s correlation coeffecient as

0.97 (Figure 1B).

This straightforward correlation of our experimental NMR findings with the present

theoretical ab initio calculations justifiably points to the following observations: (i) it is now

possible to predict the K298

NMRG° Δ of nucleosides quite dependably by simply knowing the ab initio

calculated ΔES-N; (ii) the substituent-dependent steric and stereoelectronic effects on the bias of the

two-state1,2b-e

N � S equilibrium in nucleosides can also be easily predicted from the ab initio

calculations provided a sufficiantly large basis set is used, and (iii) the comparison of Pearson’s

correlation coeffecients in Figure 1 indicates the necessity of mimicking the solvation behaviour of

the experimental NMR measurement condition in these high level ab initio calculations. It is quite

likely that this correlation between the theory and the experiment would give much deeper insight

into the molecular orbital basis of the role of the stereoelectronic forces in modulating the


90

conformation of nucleosides and nucleotides as well as shed light in their ubiquitous self-assembly

process governing the chemistry of life in general.

4.5.3 Stronger gauche effect in nucleosides than in 1,2-difluoroethane

In 1,2-difluoroethane, an energy stabilization of 2.4 -3.4 kJ/mol has been observed in favour

of the gauche rotamer over the trans counterpart (section 1.13) but in 3'-substituted-2',3'-

dideoxynucleosides we have found that the [X3'-C3'-C4'-O4'] gauche effect ranges from -2.8 kJ/mol

for X = NH2 to -11.3 kJ/mol for X = F (Fig. 15, Table 9). This can be understood in terms of

dihedral constraints imposed on the endocyclic torsions owing to the ring closure nature of the five-

membered pentofuranose nucleosides in contrast to 1,2-difluoroethane where only constraint is

imposed by favourable orbital overlap or bond-bending in the gauche compared to the trans rotamer.

4.6 The 2'-OH effect in ribonucleos(t)ides is nucleobase-dependent

The subtraction of ΔH°N of a particular β-D-dN 37 or 41 - 44 from that of its β-D-rN

counterpart (50 - 54) yields an estimate (∆∆H°11, Table 6) for the combined strength of the [O2'-

C2'-C1'-O4'] and [O2'-C2'-C1'-N1/9] gauche effects, i.e. the overall effect of 2'-OH in β-D-rNs in

comparison with β-D-dNs. ∆∆H°11 values are as follows (in kJmol-1): adenin-9-yl (-0.5) = guanin-

9-yl (0.5) < uracil-1-yl (2.6) ≈ thymin-1-yl (2.7) ≈ cytosin-1-yl (3.0) ≈ 5-fluorouracil-1-yl (3.1).

Thus, the ability of 2'-OH to drive the sugar conformation toward S-type pseudorotamers is greater

in the purine β-D-rNs with respect to the pyrimidine β-D-rNs.

(34)

ΔGNMR

-6 -4 -2 0 2 4

ΔE(S -N)

-16

-12

-8

-4

0

4

(I)

(II)

(34)

ΔGNMR

-6 -4 -2 0 2 4

ΔE(S -N)

-16

-12

-8

-4

0

4(113)

(43)

(114)

(115)

(116)

(88)

(34)

(113)

(113)

(88)

(88)

(43)

(43)

(114)

(114)

(115)

(115)

(116)

(116)

(A) B)

Figure 16. Panel (A) shows the plot of the free energy of the N � S pseudorotational equilibrium (K298

NMRG° Δ , in kJ

mol-1

) as a function of ab initio calculated energy diffrenence between the S- and the N-type pseudorotamers (ΔES-N, in

kJ mol-1

) at HF/6-311++G** level for 2′,3′-dideoxy 3′-substituted (X) thymidine derivatives [with X = H (34), NH2

(113), OH (43), OCH3 (114), NO2 (115), OCF3 (116), F (88)], both in the gas phase [■, dotted line, graph I] as well as

in the solution [�, solid line, graph II] gives straight lines with slope = 1.03 (σ = 0.11), intercept = -1.22 (σ = 0.31) and

R = 0.92 for I and slope = 1.50 (σ = 0.09), intercept = -4.92 (σ = 0.30) and R = 0.97 for II respectively. Panel (B) shows

the similar plot ofK298

NMRG° Δ (in kJ mol

-1) as a function ΔES-N (kJ mol

-1) derived from the single point solution phase

DFT calculations at B3LYP/6-311++G**//HF/6-311++G** level, giving the straight line with slope = 1.41 (σ = 0.11),

intercept = -4.17 (σ = 0.35) and R = 0.97. The fact that ∆ES-N for 116 is clearly smaller than that of 88 supports our

experimental result of the reduced efficiency of the [3'-OCF3-C3'-C4'-O4'] gauche effect in the former in comparison

with the [F3'-C3'-C4'-O4'] gauche effect in the latter.



91

2'-OH stabilizes N-type pseudorotamers in pyrimidine β-D-rNMPs 76 - 78 (by 1.7 - 4.0

kJmol-1, i.e. ∆∆H°29 in Table 7) and in β-D-rNMPEts 81 - 83 (by 0.5 - 2.1 kJmol-1, i.e. ∆∆H°27 in

Table 7) in comparison with the 2'-deoxycounterparts β-D-dNMPs 66 - 68 and β-D-dNMPEts 71 -

73. In contrast, there is virtually no enthalpy stabilization of the S-type pseudorotamers in purine β-

D-rNMPs 74 and 75 and in β-D-dNMPs 64 and 65, as shown by small ∆∆H°29 values (Table 7).

Finally, in β-D-AMPEt (79) and β-D-GMPEt (80), 2'-OH even destabilizes N-type conformations in

comparison with β-D-dAMPEt (69) and β-D-dGMPEt (70) (by ∆∆H°27 = -1.4 and -1.0 kJmol-1,

respectively). Thus, as in the ribonucleosides, the ability of 2'-OH to stabilize N-type

pseudorotamers is much greater in the pyrimidine than in the purines series.

The subtraction (∆∆H°8 = 4.5 kJmol-1, Table 6) of ΔH°N of abasic sugar 13 from that of 14

yields an estimate for the combined influence of the [HO2'-C2'-C1'-O4'] and [HO2'-C2'-C3'-O3']

gauche effects, the later being negligible (because of gauche orientation in both N and S

conformers). ∆∆H°8 has the same value, but the opposite sign, as ∆∆H°7, showing that the [HO2'-

C2'-C1'-O4'] and [HO4'-C4'-C3'-O3'] gauche effects are cancelling each other, which is consistent

with the results of regression (A) and also with the fact that the enhancement or reduction of

electron-density of O4' owing to variable electronic character of the nucleobase will affect both

[HO2'-C2'-C1'-O4'] and [HO3'-C3'-C4'-O4'] in a similar manner.

Figure 17. The correlation

plot of the overall effect

(AEddN) of adenin-9-yl in β-

D-ddA (30), guanin-9-yl in β-

D-ddG (31), cytosin-1-yl in β-

D-ddC (33), thymin-1-yl in β-

D-ddT (34) and uracil-1-yl in

β-D-ddU (35) as a function of

the overall effect of 2'-OH in

β-D-rNs 50 - 54, β-D-rNMPs

74 - 78 and β-D-rNMPEts 79 -

83 (in the N state). The effect

of the C1'-aglycone in β-D-ddNs was estimated from ∆∆H°2 (Table 6 and Fig 13). The overall effect of 2'-OH was

estimated from ∆∆H°11 for β-D-rNs, ∆∆H°29 for β-D-rNMPs and ∆∆H°27 for β-D-rNMPEts (Table 7 and Fig 13). The

Pearson's corelation coefficient of the straight line is 0.87.

Therefore, if one assumes that the strengths of the [HO3'-C3'-C4'-O4'] and [HO2'-C2'-C1'-

O4'] gauche effects are identical but with the opposite signs (as suggested by the identical ΔH°N

values of 12 and 14), then the nucleobase-dependent -∆∆H°10 values give estimates for the

magnitudes of the [HO2'-C2'-C1'-O4'] gauche effect in β-D-rNs (Table 6). Subtraction of -∆∆H°10

from ∆∆H°11 affords the nucleobase-dependent strength of the [O2'-C2'-C1'-N1/9] gauche effect in

β-D-rNs 50 - 54, i.e.: -7.9 kJmol-1 in β-D-A, -6.7 kJmol-1 in β-D-G, -4.3 kJmol-1 in β-D-C, -4.1

kJmol-1 in β-D-rT and -3.7 kJmol-1 in β-D-U, showing that in purine nucleosides this gauche effect

is much more efficient than in the pyrimidine counterparts.

The interdependency of the overall effect of 2'-OH in β-ribonucleosides and ribonucleotides

is also evidenced by the fact that a plot of the effect of the nucleobase in β-D-ddNs 30 - 35 as a

-4 -2 0 2 4 6

AE ddN (kJ mol-

1

)

2

4

6

8

79 75 7480

50 & 51

82

83

53

54

76

77

52

78

81


92

function of ∆∆H°11 (for β-D-rNs 50 - 54), ∆∆H°27 (for β-D-rNMPEts 79 - 83) and ∆∆H°29 (for β-

D-rNMPs 74 - 78) in Fig. 17 gives a straigthline with correlation coefficient of 0.87. As 2'-OH

drives the pseudorotational equilibrium to the N, the strength of the anomeric effect of the

nucleobase increases in a cooperative manner.

4.7 3'-gauche effect modulation by 2'-OH in ribonucleos(t)ide

In ribonucleos(t)ides, 2'-OH influences the drive of the sugar conformation through the [O2'-

C2'-C1'-O4'] and [O2'-C2'-C1'-N1/9] gauche effects. Unlike in β-D-dNs, in β-D-rNs a possible

contribution of H-bonding142 interaction between 2'- and 3'-substituents to the preference for gauche

orientation within [HO3'-C3'-C4'-O4'] fragment cannot be excluded. It has been for instance shown

by 1H-NMR509,510 that there is an intramolecular water bridge between the vicinal 2'-OH and 3'-

phosphate in cAMP in which the 2'-OH hydrogen accepts the lonepair of water oxygen.

We have attempted to delineate27 the magnitudes of the gauche effects involving 2'-OH in β-

D-A (50) in the N state by comparing the conformational preferences of its constituent

pentofuranose sugar with those in β-D-ddA (30), β-D-dA (37), β-D-3'-dA (63), β-D-dAMP (64), β-

D-AMP (74) and abasic sugars 12 - 14 through a set of pairwise subtractions of the ΔH°N values of

their N � S equilibria (Table 6). The main results of this work are as follows:

(i) As discussed in Section 4.6, in abasic sugars 13 and 14, the [HO3'-C3'-C4'-O4'] and

[HO2'-C2'-C1'-O4'] gauche effects cancel each other, their magnitudes being respectively ∆∆H°7 = -

4.5 and ∆∆H°8 = 4.5 kJmol-1.

(ii) The comparison of ∆∆H°10 (-7.4 kJmol-1) and ∆∆H°19 (-5.7 kJmol-1) shows that the

strength of the [HO3'-C3'-C4'-O4'] gauche effect is reduced by 1.7 kJmol-1 in β-D-A in comparison

with β-D-dA, which can be attributed to the effect of 2'-OH in the former.

(iii) The strengths of the [O4'-C4'-C3'-O3'PO3H-1/-2] gauche effect in β-D-dAMP and β-D-

AMP have been estimated by ∆∆H°20 (-8.9 kJmol-1) for the former and by subtracting (-6.2 kJmol-

1) ΔH°N value of β-D-3'-dA from that of β-D-AMP, respectively. Thus the [O4'-C4'-C3'-O3'PO3H-]

gauche effect is less efficient by 2.7 kJmol-1 in β-D-AMP than in β-D-dAMP, showing the effect of

2'-OH in the former.

The fact that 2'-OH weakens the [O3'-C3'-C4'-O4'] gauche effect less efficiently in β-D-A

than in β-D-AMP [compare (i) and (ii)] can be understood as a consequence of the greater freedom

of the lonepair of O2' to participate in the [O2'-C2'-C1'-N9] and [O2'-C2'-C1'-O4'] gauche effects in

β-D-AMP than in β-D-A, where it acts as a donor in the intramolecular hydrogen-bonding

interaction between 2'-OH and 3'-OH511,512. Hydrogen bonding of 2'-OH with N3513 may induce the

strengthening of the [O2'-C2'-C1'-N9] gauche effect in purine β-D-rNs.

In A-type RNA, pseudorotamers are inherently stabilized when 2'-OH is involved in

hydrogen bonding interaction with the nucleobase, and the substitution of 2'-OH in hammerhead

ribozyme for NH2 or F changes the ratio of N- versus S-type pseudorotamers as well as influences

the hydrogen bonding capabilities514-516 of the sugar.



93

(iv) The overall stereoelectronic effect of 2'-OH has first been estimated in β-D-3'-dA by

subtracting (-2.2 kJmol-1) ΔH°N of β-D-ddA from its own. For β-D-A and β-D-AMP, the strength

of the 2'-OH effect is reflected in the values of ∆∆H°11 (-0.5 kJmol-1) and ∆∆H°29 (0.5 kJmol-1),

respectively. Thus the ability of the effect of 2'-OH to counteract the [O3'-C3'-C4'-O4'] gauche

effect by stabilizing N-type conformations increases in the order: β-D-3'-dA < β-D-A < β-D-AMP.

The 2'-OH stabilization of N-type conformations in β-D-AMP compared with β-D-A and β-D-3'-dA

can be attributed either to a more efficient [O2'-C2'-C1'-O4'] gauche effect or to the weakening of

the [O2'-C2'-C1'-N9] gauche effect.

4.8 Drive of pseudorotation in β-nucleosides by the nature of the nucleobase

4.8.1 The two-state N � S equilibrium is evidenced by pKa values from ∆G°

For all β-D-nucleosides, either in the 2',3'-dideoxy (30 - 34), 2'-deoxy (37 and 40 - 45) or

ribo (50 - 55) series, the plots of ∆H°, -T∆S° and ∆G° values as a function of pD show that the bias

of their two-state N � S equilibrium can be successfully modulated by the protonation and/or

deprotonation of the nucleobase (Fig. 11 and Table 2). The origin of this modulation will be

discussed for each series in the next paragraph. (In β-L-nucleosides, similar trends are observed

since the thermodynamics of the N � S equilibrium in each of the D- and L-enantiomers are

virtually the same owing to the mirror-image relationship, as discussed in Section 4.1(e)). The value

of the pD at the inflection point of the plot of pD-dependent ∆G° values for each compound has

been determined by fitting the experimental data to the Henderson-Hasselbach equation, as

discussed in Section 3.7. These pKa(s) values are virtually identical to the known pKa(s) of the

constituent nucleobases in 30 - 34, 37, 40 - 45 and 50 - 55, as found in the literature165,247,379,382-384

and to other estimates of pKa values independently derived by us from Hill plots of pD-dependent

1H chemical shifts of the anomeric and aromatic protons of the same nucleosides (± 0.4 pD unit,

compare cols. 5 and 6, on one hand, cols. 10 and 11, on the other, in Table 2). This result

unequivocally validates the two-state N �S equilibrium model.

4.8.2 Identical pKas of the nucleobases in 2',3'-dideoxy, 2'-deoxy and ribo series

The pKa values of cytosin-1-yl and thymin-1-yl (determined from pD-dependent 1H

chemical shifts) in β-D-ddC/T, β-D-dC/T and β-D-C/rT are the same (± 0.1 from the average

values), showing the lack of influence of 2'- and/or 3'-OH on their electronic character (See Section

8.8 for the discussion on 3',5'-bisphosphate of guanosine). For adenin-9-yl and guanin-9-yl, the

electron-withdrawing effect of 2'-/3'-OH is only partly reflected in the slight decrease of the pKa

value in the order ddN > dN > rN (i.e. by 0.3 and 0.4 pD unit, respectively). This tendency should

however be considered with caution, since the accuracy of the estimates of the pKa values

determined from pD-dependent chemical shifts is ≈ 0.1 pD unit.


94

Table 10. The relative populations of syn versus anti rotamers around the glycosidic torsion in

α-D-ddNs 17 - 20, β-D-ddNs 30 - 34, α-D-dNs 21 - 26, and β-D-dNs 37 - 39 and 41 - 43 at various

pDs a from 1D-nOe difference experimentsb.

Compound

pD

ηH1' (%)

(H6/H8 irradiated)

ηH6/H8 (%)

(H1' irradiated)

(ηH1' + ηH6/H8) / 2 % syn

rotamers

α-D-ddA (17) 7.3 2.3 1.9 2.1 20

β-D-ddA (30) 7.0 2.3 2.4 2.4 22

α-D-ddG (18) 7.5 1.8 1.8 1.8 17

β-D-ddG (31) 2.1 1.4 1.5 1.5 14

7.4 1.9 2.4 2.2 21

11.5 2.8 2.8 2.8 26

5'-OMe-β-D-ddG (32) 2.1 1.6 - 1.6 15

6.6 1.7 1.4 1.6 21

11.7 1.3 1.3 1.3 12

α-D-ddC (19) 7.1 2.4 - 2.4 22

β-D-ddC (33) 7.0 1.5 - 1.5 14

α-D-ddT (20) 7.2 3.7 3.2 3.5 33

β-D-ddT (34) 7.0 2.1 2.0 2.1 20

α-D-dA (21) 2.1 2.5 2.8 2.7 25

6.5 5.4 4.0 4.7 44

3'-OMe-α-D-dA (22) 1.6 0.6 0.7 0.7 7

6.9 1.2 0.9 1.1 10

3',5'-diOMe-α-D-dA (23) 1.6 1.0 0.9 1.0 10

6.7 0.9 1.1 1.0 10

β-D-dA (37) 1.6 3.5 4.2 3.9 36

6.9 5.8 5.9 5.9 55

3'-OMe-β-D-dA (38) 1.6 4.1 4.2 4.2 39

7.0 5.5 6.5 6.0 56

3',5'-diOMe-β-D-dA (39) 2.2 2.4 2.4 2.4 22

7.3 1.4 1.2 1.3 12

α-D-dG (24) 1.8 3.8 2.3 3.1 29

7.3 4.7 2.3 3.5 33

11.6 6.1 7.7 6.9 65

β-D-dG (41) 1.8 3.5 3.1 3.3 31

7.7 3.8 4.0 3.9 36

11.1 4.5 4.6 4.6 43

α-D-dC (25) c 2.0 0.9 - 0.9 8

7.1 2.4 - 2.4 22

β-D-dC (42) d 7.3 2.8 2.8 2.8 26

α-D-T (26) 7.1 3.0 - 3.0 28

11.6 3.4 - 3.4 32

β-D-T (43) 6.7 3.7 3.0 3.4 32

11.7 3.1 2.4 2.8 26

a 1D-nOe difference experiments were performed in D2O [298 K except for all guanin-9-yl nucleosides (288 K) due to

their decomposition above 288 K at acidic pDs]. b We used the method of Rosemeyer to calculate the population of

syn rotamers around the glycosidic torsion from homonuclear 1H nOes517. c Selective saturation of H1' could not be

performed at acidic pD for α-D-dC and at neutral pD for α-D-ddC and β-D-ddC due to near isochronous H1' and H5 . d Could not be measured in the acidic solution due to the fact that H1' and H5 in β-D-dC have the same chemical

shift at pD < 3 at 298 K.



95

4.8.3 Anomeric effect in β-D-ddNs is modulated by the nature of the nucleobase

A comparative analysis of opHΔ , -T o

pSΔ and opGΔ values of 30 - 34 and of the corresponding

oN

HΔ , -T oNSΔ , o

NGΔ and o

DHΔ , -T o

DSΔ and oD

GΔ values (Table 2) shows that opHΔ and o

DHΔ prevail over

the counteracting -T opSΔ and -T o

DSΔ which results into the overall stabilization of N-type sugars in

the P and D states (i.e. positive opGΔ and o

DGΔ ). Protonation of the nucleobase in 30 - 34 steadily

shifts the N � S pseudorotational equilibrium toward more N-type conformation than in the N state.

When the nucleobase is fully protonated, a plateau in thermodynamic values in the P state is reached

[i.e. oNP

G−

ΔΔ (kJmol-1) 1.5 (30), 3.8 (31), 3.2 (32), 1.1 (33)]. In contrast, deprotonation shifts the N

� S equilibrium toward more S-type sugars (with pseudoequatorially oriented nucleobase) than in

the N state [i.e. oND

G−

ΔΔ (kJmol-1) = -1.1 (31), -0.3 (32) and -0.8 (34)]. We have compared the

preferred conformation of the sugar moiety and of the nucleobase in β-D-ddG (31) and 5'-OMe-β-

D-ddG (32) over the whole ≈ 2.0 - ≈ 12.0 pD range in order to examine whether 5'-OH and guanin-

9-yl, by forming an H-bond or guanin-9-yl alone, by adopting different orientations around the

glycosyl torsion at different pDs, contribute to the above oNP

G−

ΔΔ and oND

G−

ΔΔ values. ∆H° of 31 and

32 are nearly identical at any pD suggesting that 5'CH2OH and N3 in 31 do not form a hydrogen

bond. Slightly greater o

NHΔ and

o

NHΔ in 32 than in 31 can be attributed to the increased steric

bulk 5'CH2OMe compared to 5'CH2OH. However, on the overall, N-type sugars are more preferred

at 298 K in 32 than 31 (see their Error! No topic specified. and oD

GΔ values), due to an entropy effect.

[In the P state, slightly larger opHΔ for 31 than for 32 can be easily explained: (i) 1H-NMR spectra of

31 and 32 were recorded at pD 1.9 and 2.0, respectively, but not at lower pDs due to rapid

decomposition. (ii) The standard deviations of ∆H°, ∆S° and ∆G° values at pD 1.9 for 31 (4.5

kJmol-1) and at pD 2.0 for 32 (3.6 kJmol-1) are higher than at other pDs, because 1H-NMR spectra

could only be recorded in a narrow temperature range (274 - 303 K). Thus ΔH°, -TΔS° and ΔG° of

31 and 32 are inevitably less accurate in the P than in the N and D states.] Our 1D-nOe difference

experiments in the P, N and D state of 31 and 32 show that guanin-9-yl assumes an anti orientation

in both compounds at any pD and this preference is almost independent of the pD and of the nature

of 5'-substituent. Moreover, the nucleobases in β-D-ddA (30), β-D-ddC (33) and β-D-ddT (34)

adopt almost exclusively anti orientations (78 - 86 % at room temperature) at neutral pD (Table 10).

The shift of the N � S equilibrium in the protonated and deprotonated states of β-D-ddNs toward

more N- and S-type sugars, respectively, compared with the N state is therefore neither the result of

any interaction between 5'-OH and the nucleobase nor induced by a different orientation of the

nucleobase around the glycosyl torsion at different pDs.

4.8.4 The orbital mixing as the origin of the O4'-C1'-N1/9 anomeric effect

The modulation of the thermodynamics of the two-state N � S equilibrium in β-D-ddNs 30

- 34 by the pD of the aqueous solution can be rationalized in terms of pD-dependent magnitude of

the O4'-C1'-N1/9 anomeric effect: As the nucleobase becomes protonated, a partial positive charge

is created at the glycosyl nitrogen. As a result, nO4' →σ∗C1'-N9 orbital mixing becomes more

favourable because of the lowering of the antibonding σ* orbital (Section 2.8). This favourable


96

orbital mixing prompts the nucleobase to adopt a pseudoaxial orientation, which in turn is achieved

in the N-type pseudorotamers. Conversely, as the nucleobase becomes deprotonated in the alkaline

pD, the ability of the glycosyl nitrogen to participate in O4'-C1'-N1/9 stereoelectronic interactions is

reduced, which is evident by the shift of N � S equilibrium to more S-type (with pseudoequatorial

aglycone) in comparison with the N state. Our interpretation is based on the assumption that the

steric bulk of the nucleobase remains to be comparable in the P, D and N states. Of course we

expect a change in the hydration chracteristics in the various cationic or anionic state vis-a-vis N

state, but that cannot be estimated experimentally with the present state-of-the art. Our results also

tend to suggest that the contribution of electrostatic repulsions (see Section 2.8, Fig 8) between the

dipole of the pentofuranose moiety and the C1'-N1/9 dipole to the overall strength of the O4'-C1'-

N1/9 anomeric effect in nucleosides and nucleotides is rather small in comparison with that of

hyperconjugative interactions (Figs 8 and 9).

Indeed, one can reasonably assume that the partial positive charge created at the glycosyl

nitrogen upon protonation of the nucleobase in β-D-ddNs will weaken the C1'-N1/9 dipole, which

in turn should lead to the reduction of the electrostatic repulsions with the O4' lonepair in the P

compared with N state of 30 - 33. Therefore, if the origin of the O4'-C1'-N1/9 anomeric effect in

nucleosides and nucleotides was mainly electrostatic, the preference of the nucleobase for

pseudoaxial orientation in the N-type pseudorotamers should be much reduced in the P state than in

the N state owing to the weakening of the anomeric effect in the former, but our results clearly show

the opposite, i.e. a higher preference for the N-type sugars at acidic compared with neutral pDs.

Similarly, upon deprotonation of the nucleobase in β-D-ddNs 31, 32 and 34, owing to the

presumably higher electron-density at the glycosyl nitrogen, one may also suggest that dipole-dipole

repulsions should be reinforced in comparison with the N or P state, and that the anomeric effect

should be more efficient in the D compared with the N or P state, which is again in contradiction

with the experimentally observed tendency of destabilization of N-type sugars in the alkaline

compared with the neutral solution (see above, section 4.8).

The strengthening of the effect of the nucleobase upon its protonation (or weakening upon

its deprotonation) can be estimated by directly subtracting oN

HΔ from opHΔ (or from o

DHΔ ). Thus

oNP

H−

ΔΔ (kJmol-1) = 3.0 (33, cytosin-1-yl) < 5.7 (30, adenin-9-yl) < 20.2 (31, guanin-9-yl) and 20.1

(32, guanin-9-yl) whereas oND

H−

�ΔΔ (kJmol-1) = -2.0 (31, guanin-9-yl) and (32, guanin-9-yl) ≈ -1.9

(34, thymin-1-yl). The subtraction of ∆∆H°

2 values in the P and N states (or N and D states) gives

identical values as oNP

H−

�ΔΔ and oND

H−

ΔΔ . Thus purine nucleosides respond more to the protonation

of their nucleobase than the pyrimidine counterparts, which is consistent with the fact that upon

addition of one equivalent trifluoroacetic acid the chemical shift of the glycosyl nitrogen is more

affected in the purine than in the pyrimidine series (∆δ (downfield shift of N9) = 6.6 ppm in N1H+-

adenosine and 6.7 ppm in N7H+-guanosine401,403 >> ∆δ (downfield shift of N1) = 1.1 ppm in

N3H+-cytosine404).



97

4.8.5 No reverse anomeric effect in pentofuranosyl nucleosides

If the reverse anomeric effect were playing any role in the drive of the sugar conformation in

pentofuranosyl β-nucleosides at acidic pD, one should observe an increase in the population of S-

type pseudorotamers, in which the nucleobase adopts a pseudoequatorial orientation, in the acidic

compared to the neutral pD. Our results show the opposite trends, therefore the pD-dependent

conformational preferences of β-D-nucleosides can be attributed to the resulting modulation of the

anomeric effect, not a reverse anomeric effect, as suggested in the case of imidazolium or

pyridinium derivatives of pyranose (Section 1.7).

4.8.6 Variable tunablity of anomeric effect in β-D-ddNs, β-D-dNs and β-D-rNs

In protonated β-D-dA, 3'-OMe-β-D-dA and 3',5'-diOMe-β-D-dA, and deprotonated β-D-dU

and 5-F-β-D-dU, both ∆H° and -T∆S° stabilize S-type conformations either with the same

magnitude or slightly stronger ∆H°. In the D state of β-D-dG and β-D-T, ∆H° overrides the

counteracting -T∆S°, resulting in the overall stabilization of S-type sugars at 298 K. Only in

protonated β-D-dC, -T∆S° is slightly stronger than ∆H° and determines the drive of the N �S

equilibrium toward S-type sugars, whereas in protonated β-D-dImb and β-D-dG, ∆H° (driving to N)

and -T∆S° (driving to S) nearly cancel each other, therefore N and S-type sugars have almost the

same population at 298 K. The balance of ∆H° and -T∆S° terms of the N � S in β-D-rNs is also

dictated by the nature of the nucleobase and the pD: In the D state of β-D-U and 5-F-β-D-rU,

opposing ∆H° (driving to the N) and -T∆S° nearly cancel each other, resulting in nearly unbiased

equilibrium. However, in the D state of β-D-G, ∆H° (stabilizing S-type sugars) prevails over the

opposing -T∆S°, and S-type sugars are favoured at 298 K. In the P state of β-D-G and β-D-C, ∆H°

prevails over counteracting -T∆S°, stabilizing N-type sugars at 298 K. In the P state of β-D-A and D

state of β-D-rT, small negative ∆H° and -T∆S° values favour slightly S-type conformations at 298

K.

In agreement with our observations in the case of β-D-ddNs 30 - 34, as the nucleobase in β-

D-dNs and β-D-rNs is protonated, the two-state N � S equilibrium of the constituent sugar moieties

is also shifted toward N-type conformations [ oNP

G−

ΔΔ (kJmol-1) = 1.0 (37 and 38), 0.4 (39), 1.3 (40),

1.6 (41), 0.5 (42) for β-D-dNs and 1.3 (50), 3.0 (51), 0.4 (52) for β-D-rNs] and conversely, as it is

deprotonated, S-type sugars are further stabilized in comparison with the N state oND

G−

ΔΔ (kJmol-1) =

-1.0 (41), -0.3 (43 - 45) for β-D-dNs and -1.3 (51), -0.3 (53 and 55), -0.2 (54) for β-D-rNs]. The

comparison of oNP

G−

ΔΔ and oND

G−

ΔΔ values for each nucleobase in β-D-ddNs (30, 31, 33 and 34), β-

D-dNs (37, 41 - 43) and β-D-rNs (50 - 53) shows that the change in the ratio of N- and S-type

pseudorotamers in the P and D states in comparison with the N state is much reduced in the 2'-

deoxy and ribo series than in the 2',3'-dideoxy counterparts. This is owing to the fact that both the

overall entropy of the system and the enthalpy of the N � S equilibrium are much less affected by

the pD of the aqueous solution in (37, 41 - 43) compared with (30, 31, 33 and 34) and (50 - 53).

At any pD, the nucleobase in β-D-dNs 37 - 39 and 41 - 43 adopts preferentially an anti

orientation around the glycosyl torsion (except β-D-dA (37) and 3'-OMe-β-D-dA (38) at neutral pD


98

since ≈ 1:1 ratio of syn and anti rotamers was found, Table 10), as suggested by our 1D nOe

difference experiments37. These observations let us propose that N3 (in purines) or O2 (in

pyrimidines) does not form a hydrogen bond with 5'-OH. 37 - 39 all have rather similar ∆H°, -T∆S°

and ∆G° values at any pD, in spite of the fact that in 39 the shift from less to more syn rotamers

occurs in going from the N to the P state, not reverse. The fact that 37 and 39 behave in the same

way also implies that in 37 no H-bond between 5'-OH and adenin-9-yl is responsible for the

positive oNP

G−

ΔΔ and oNP

H−

ΔΔ values.

The modulation of the overall stereoelectronic forces (i.e. anomeric effect + gauche effects)

upon protonation and/or deprotonation of the constituent nucleobase in β-D-dNs and β-D-rNs is

reflected in the change in the enthalpy of the N � S equilibrium in the P and/or D compared with N

state, which has been estimated by subtracting their oN

HΔ values from the corresponding opHΔ and

oD

HΔ values, respectively :

oNP

H−

ΔΔ (kJmol-1) = 0.7 (42, cytosin-1-yl) < 2.3 (40, imidazol-1-yl) ≈ 2.6 (39, adenin-1-yl) <

3.2 (37, 38, adenin-1-yl) < 4.9 (41, guanin-9-yl) in β-D-dNs and 2.9 (52, cytosin-1-yl) < 4.2 (50,

adenin-1-yl) < 8.7 (51, guanin-9-yl), in β-D-rNs whereas oND

H−

ΔΔ (kJmol-1)= -0.3 (45, 5-fluoro-

uracil-1-yl) ≈ -0.5 (43, thymin-1-yl) ≈ -0.7 (44, uracil-1-yl) < -2.1 (41, guanin-9-yl) in β-D-dNs and

-1.5 (55, 5-fluoro-uracil-1-yl) = -1.5 (53, thymin-1-yl) ≈ -1.7 (54, uracil-1-yl) < -4.3 (51, guanin-9-

yl) in β-D-rNs (for oND

H−

ΔΔ , the < sign signifies a reduced ability of the deprotonated nucleobase to

destabilize N-type conformations with respect to the N state). [Note that the same values can be

derived from the subtraction of ∆∆H°3 values in the N state from those in the P and D states,

respectively, for β-D-dNs and from the subtraction of ∆∆H°4 values in the N state from those in

the P and D states, respectively, for β-D-rNs (Table 6, Fig 13)]. Thus, in all ddNs (Section 4.8),

dNs and rNs, pyrimidine nucleobases are much less able to transmit the free-energy of its

protonation � deprotonation equilibrium than the purine counterparts. The smaller value of oNP

H−

Δ

of 40 in comparison with those of 37 - 39 and 41 shows the effect of the electron-withdrawing

character of the fused pyrimidine ring upon the imidazole moiety in the purine nucleosides.

oNP

H−

ΔΔ and oND

H−

ΔΔ increase in the following order: β-D-dNs < β-D-rNs << β-D-ddNs

(Table 2). The reduced modulation in the dNs compared to the ddNs is presumably the result of the

opposing influence of the [HO3'-C3'-C4'-O4'] gauche effect in the former, which is also in

agreement with the fact that ∆∆H°2 values (reflecting the overall effect of the nucleobase in for β-D-

ddNs) are much larger than the corresponding ∆∆H°3 values (showing the effect of the base in β-D-

dNs) at any pD (Fig 13, Table 6).

As O4' is involved in the [HO3'-C3'-C4'-O4'] gauche effect, its electrostatic potential is

changed because of σC3'-H3' → σ∗C4'-O4' participation and its ability to be participate to O4'-C1'-N1/9

stereoelectronic nO4' → σ∗C1'-N interactions is reduced. In Table 6, ∆∆H°10

values show that as the

nucleobase is protonated, the strength of the [HO3'-C3'-C4'-O4'] gauche effect in β-D-dNs

increases, whereas it is reverse when the nucleobase is deprotonated.

Again, this is explained by the fact that as the O4' becomes more positively charged owing to

the enhanced nO4' → σ∗C1'-N interactions with the protonated aglycone, the energy level of σ∗C4'-O4'



99

is lowered (in comparison with the neutral state), which renders the σC3'-H3' → σ∗C4'-O4' participation

more efficient in the protonated nucleosides.

The modulation of the thermodynamics of N � S equilibrium in β-D-rNs is much smaller

than ddNs counterparts [the relative degree of pD-dependent flexibility follows the following order:

β-D-dNs < β-D-rNs << β-D-ddNs, Table 2] because of the interplay of three gauche effects ([O3'-

C3'-C4'-O4'], [O2'-C2'-C1'-O4'] and [O2'-C2'-C1'-N1/9]) in the former modulating the strength of

the anomeric effect.

Since the [O3'-C3'-C4'-O4'] and [O2'-C2'-C1'-O4'] gauche effects presumably cancel each

other (compare ∆H° of 12 and 14 in Table 2), the greater flexibility of rNs than dNs as a function of

pD may be attributed to the [O2'-C2'-C1'-N1/9] gauche effect. The magnitude of the 2'-OH effect in

β-D-rNs as a function of pD can be estimated by subtracting opHΔ (or o

NHΔ or o

DHΔ ) values of β-D-

dNs counterparts (∆∆H°11

in Table 6) from those of β-D-rNs. The ability of 2'-OH to drive the

pentofuranose conformation toward N-type pseudorotamers increases as follows: adenin-9-yl =

guanin-9-yl <<< uracil-1-yl ≈ thymin-1-yl ≈ cytosin-1-yl ≈ 5-fluorouracil-1-yl (in the N state),

adenin-9-yl << guanin-9-yl << cytosin-1-yl (in the P state) and guanin-9-yl <<< uracil-1-yl ≈

thymin-1-yl ≈ 5-fluorouracil-1-yl (in the D state) (see section 4.6 for correlation of the anomeric

effect and the overall 2'-OH effect in ribonucleosides and nucleotides, Table 6, Figs 13 and 17).

Therefore, in purines 2'-OH stabilizes more S-type conformations than in the pyrimidines at any pD.

This is owing to the reduced [O2'-C2'-C1'-N1(pyrimidine)] gauche effect in the former in

comparison with the [O2'-C2'-C1'-N9(purine)] gauche effect in the latter. In going from the P to the

N and D state, 2'-OH becomes steadily more efficient to drive the sugar conformation in rNs toward

S-type conformations, which are reflected in the values of ∆∆H°11

(Table 6, Figs 13 and 17).

4.8.7 Correlation of the electronic nature of aglycone with pseudorotational state

The monitoring of chemical shift of aromatic and anomeric protons is a dependable marker

to assess the protonation �deprotonation equilibrium, which when measured as a function of pH

yields the pKa

518. For each β-D-ddN 30 - 34, β-D-dN 37 and 40 - 45 and β-D-rN 50 - 55, we have

plotted the change in the chemical shift of the aromatic proton(s) of the constituent nucleobase as a

function of the ∆G° of the two-state N � S equilibrium over the whole pD range (Fig 18) to

examine if there is an interdependency. A straightline could be fitted through all experimental points

of each correlation plot. The Pearson's correlation coefficients (R) are larger than 0.97 for all

nucleosides except for β-D-T (43, R = 0.90) and β-D-U (54, R = 0.95) owing to the errors involved

in the determination of their pD-dependent ∆G° values as a result of the limited change in 3J

HH over

the 278 K - 358 K temperature range. This shows that the electronic changes that take place in the

aglycone as a result of pD-dependent change in the protonation � deprotonation equilibrium is

indeed transmitted to alter the ∆G° of the two-state N � S equilibrium over the whole pD range.

Interestingly, this transmission of the tunable electronic character of the nucleobase to steer the

sugar conformation in β-D-nucleosides takes place through pD-dependent change of the strength of


100

ΔGo

(kJ/mol)

-3 -2 -1 0

δH8

(288 K, ppm)

8.1

8.2

8.3

8.4

8.5

8.6

ΔG o (kJ/mol)

-3 -2 -1

δH8

(288 K, ppm)

7.6

8.0

8.4

8.8

9.2

ΔGo

(kJ/mol)

-3 -2 -1 0

δH6

(298 K, ppm)

7.6

7.7

7.8

7.9

8.0

8.1

α-D-ddA (17)

(A )

α-D-ddG (18) α-D-ddC (19)

(B ) (C )

ΔGo

(kJ/mol)

-0.9 -0.6 -0.3 0.0

δH6

(298 K, ppm)

7.2

7.3

7.4

7.5

7.6

7.7

ΔGo (kJ/mol)

-5 -4 -3 -2 -1

δH8

(288 K, ppm)

8.3

8.4

8.5

8.6

ΔGo

(kJ/mol)

-5 -4 -3 -2

δH8

(288 K, ppm)

8.0

8.4

8.8

9.2

α-D-ddT (20)

(D )

α-D-dA (21) α-D-dG (24)

(E ) (F)

ΔGo

(kJ/mol)

-4.4 -4.0 -3.6 -3.2

δH6

(298 K, ppm)

7.7

7.8

7.9

8.0

8.1

8.2

ΔGo (kJ/mol)

-2.0 -1.5 -1.0

δH6

(298 K, ppm)

7.6

ΔGo

(kJ/mol)

1 2 3 4 5 6

δH8

(298 K, ppm)

8.1

8.2

8.3

8.4

8.5

8.6

8.7

α-D-dC (25)

(G )

α-D-dT (26) β-D-ddA (30)

(H ) (I)

ΔGo

(kJ/mol)

2 4 6

δH6

(288 K, ppm)

8.0

8.4

8.8

9.2

ΔG o (kJ/mol)

3 4 5 6

δH8

(288 K, ppm)

7.6

8.0

8.4

8.8

9.2

ΔGo

(kJ/mol)

2 3 4 5 6

δH6

(298 K, ppm)

7.8

7.9

8.0

8.1

8.2

8.3

8.4

β-D-ddG (31)

(J)

5'-OMe-β-D-ddG (32) β-D-ddC (33)

(K ) (L)

ΔGo

(kJ/mol)

2 3 4

δH8

(298 K, ppm)

7.4

7.5

7.6

7.7

7.8

ΔG o (kJ/mol)

-2 -1

δH8

/δH2 (p

pm)

8.0

8.2

8.4

8.6

ΔGo

(kJ/mol)

-1 0

δHa

/δHb

/δHc

(ppm)

6.4

7.2

8.0

8.8

9.6

β-D-ddT (34)

(M )

β-D-dA (37) β-D-dImb (40)

(N ) (O )

ΔGo

(kJ/mol)

-3.2 -2.4 -1.6 -0.8 0.0

δH8

(ppm)

7.5

7.8

8.1

8.4

8.7

9.0

9.3

ΔGo (kJ/mol)

-1.2 -1.0 -0.8

δH6

(ppm)

7.7

7.8

7.9

8.0

8.1

ΔGo

(kJ/mol)

-1.8 -1.6 -1.4 -1.2

δH6

(ppm)

7.35

7.40

7.45

7.50

7.55

7.60

7.65

β-D-dG (41)

(P )

β-D-dC (42) β-D-T (43)

(Q ) (R )

Figure 18 (See the legend p. 116)



101

ΔGo

(kJ/mol)

-1.5 -1.4 -1.3 -1.2 -1.1 -1.0

δH6

(ppm)

7.5

7.6

7.7

7.8

7.9

ΔGo

(kJ/mol)

-2.1 -1.4 -0.7 0.0

δH8

/δH2

(ppm)

8.0

8.1

8.2

8.3

8.4

8.5

8.6

8.7

β-D-dU (44)

(S)

β-D-A (50)

(U)

ΔGo

(kJ/mol)

-3 -2 -1 0 1

δH8

(ppm)

8.0

8.5

9.0

ΔGo (kJ/mol)

1.4 1.5 1.6 1.7 1.8 1.9 2.0

δH6

(ppm)

7.6

7.8

8.0

8.2

ΔGo

(kJ/mol)

-0.4 -0.3 -0.2 -0.1

δH6

(ppm)

7.4

7.5

7.6(V) (W) (X)

ΔGo

(kJ/mol)

-4.0 -3.6 -3.2 -2.8 -2.4 -2.0

δH8

(ppm)

7.9

8.0

8.1

8.2

8.3

8.4

8.5

ΔGo (kJ/mol)

-5 -4 -3

δH8

(ppm)

7.9

8.0

8.1

8.2

8.3

8.4

8.5

8.6

ΔGo

(kJ/mol)

-4 -3 -2 -1 0 1

δH6

(ppm)

7.6

7.7

(b) (c) (d)

ΔGo

(kJ/mol)

-2 0

δH1'

(ppm)

4.5

4.6

4.7

ΔGo (kJ/mol)

-2 -1

δH6

(ppm)

7.2

7.6

8.0

(e) (f)

ΔGo

(kJ/mol)

-1.6 -1.5 -1.4 -1.3 -1.2 -1.1

δH6

(ppm)

7.6

7.7

7.8

7.9

8.0

8.1

(T)

5-F-β-D-dU (45)

ΔGo

(kJ/mol)

0.0 0.1 0.2 0.3 0.4

δH6

(ppm)

7.4

7.5

7.6

7.7

7.8

7.9

ΔGo (kJ/mol)

0.1 0.2 0.3 0.4 0.5 0.6

δH6

(ppm)

7.6

7.8

8.0

ΔGo

(kJ/mol)

-4 -3 -2 -1

δH2

(ppm)

8.0

8.4

8.8(Y) (Z) (a)

β-D-G (51) β-D-rT (53)β-D-C (52)

β-D-U (54)

Formycin B (56)

5-F-β-D-U (55)

9-deaza-A (58)Formycin A (57) ψ-isoC (59)

1-Me-ψ-U (61)ψ-U (60)

Figure 18: Correlation plots of the chemical shifts of H2/H6/H8/H1' in α-D-ddNs 17 - 20 [Panels (A)-(D)], α-D-dNs

21, 24 - 26 [Panels (E)-(H)], β-D-ddNs 30 - 34 [Panels (I)-(M)], β-D-dNs 37 and 40 - 45 [Panels (N)-(T)], β-D-rNs 50 -

55 [Panels (U)-(Z)] and β-D-C-rNs 56 - 61 [Panels (a)-(f)] (at 298 K unless stated otherwise) as a function of pD-

dependent ∆G° of their N � S equilibrium. The Pearson's correlation coefficients (R) of the straight lines are > 0.96 for

all nucleosides except β-D-T (R = 0.90), β-D-U (R = 0.95), α-D-ddA (R = 0.62), α-D-dA (infinite slope) and formycin

B (R ≈ 0.92) owing to the errors involved in the determination of their pD-dependent ∆G° values as a result of the

limited change in 3JHH over the 278 K - 358 K range.


102

anomeric and gauche effects. This change of N � S equilibrium further predisposes the

conformation of the phosphate backbone as discussed in section 7. Thus the single-stranded

nucleotide acts like a wire (Nucleotide wire)38,44.

5. Comparison of stereoelectronic effects in α- and β-D-nucleosides

β-D-nucleosides are solely chosen by Nature as the ubiquitous building blocks for the

storage of information in DNA and RNA. Only a few α-nucleosides are found in Nature519-522. The

biological (e.g. antitumor, bacteriostatic, cytostatic) activities of some α-nucleosides have been

reported523-530. Séquin531 has postulated that a double helix featuring stabilizing hydrogen bonds

through Watson-Crick base pairing and base-base stacking as in the natural DNA helix can be

formed between a chain of α-nucleotides and its complementary α-counterpart (chains of opposite

polarity) or its complementary β-counterpart (chains of the same polarity). NMR experiments on α-

hexadeoxyribonucleotides532,533 have subsequently confirmed Séquin's predictions. It has been

shown recently that native β-anomeric sequences containing a single α-anomeric nucleotide

(inserted via 3'-3' or 5'-5' phosphodiester linkage) form duplexes whose structures emulate canonical

B-DNA534, with subtle differences in stability and local structure dictated by the nature of the

nucleobase in the α-nucleotide. A possible relation between the natural selection of β- over α-

nucleosides in DNA and a lack of conformational variability of the pentofuranose moieties in the

latter has been proposed basing on the relative energies of various pseudorotamers in both series175.

However, no experimental comparison of the thermodynamics of the N � S equilibrium in α-

versus β-nucleosides has supported this theoretical work.

We have assessed36,37 for the first time how the interplay of anomeric and gauche effects in

α-nucleosides dictates their sugar conformation. We have initially36 considered only α-D-ddNs,

owing to the fact that only the nucleobase and 5'CH2OH drive their sugar conformation. We have

subsequently37 turned our attention to α-D-dNs in order to analyze the impact of the presence of 3'-

OH on the drive of their sugar conformation in comparison with the dideoxy counterparts. In these

works, we have uniquely shown that the change of the configuration at C1' results in a drastic

modulation of the strengths of stereoelectronic forces as well as of the pD-dependent flexibility of

the sugar conformation in the α- versus β-nucleosides.

5.1 The relative magnitude of the anomeric and gauche effects in Neutral state

5.1.1 Anti orientation of the nucleobase in α/β-D-ddN and α/β-D-dN

Our 1D-nOe difference experiments in the N state of α-D-ddNs 17 - 20 and α-D-dNs 21 - 26

show that except in the case of α-D-dA (44 % syn) and α-D-dG (65 % syn in the D state), the

constituent nucleobases in all α-nucleosides compounds adopt preferentially (67 - 93 %) an anti

orientation around the glycosyl torsion (at 298 K or 288 K) (Table 10). Additionally, the extent of

the stabilization of anti over syn rotamers is nearly independent of the configuration of the sugar

moiety at C1' in α- versus β-D-ddNs and in α- versus β-D-dNs. There are only a few exceptions: In

3'-O-Me-α-D-dA (22), adenyl-9-yl is almost exclusively anti whereas in 3'-OMe-β-D-dA (38) the



103

population of syn rotamers is higher at any pD, and for α-D-dG (24) (in the D state), in which

guanin-9-yl adopts preferentially a syn orientation whereas in β-D-dG (41) the anti and syn rotamers

are almost equipopulated. These results have three important implications: (i) The observed

differences in the drive of the N � S equilibrium as a function of temperature and change of pD

(see below) in α-D-ddNs compared to their β-counterparts is not correlated with the preferred

orientation of the nucleobase around the C1'-N1/9 bond, since it is independent of the configuration

at C1'. (ii) In α-D-dNs, the anti orientation of the nucleobase prevents it from forming a hydrogen

bond with 3'-OH. (iii) In spite of the fact that syn orientations of adenyl-9-yl are more stable in α-D-

dA (21) than in 3'-OMe-α-D-dA (22) and in 3',5'-diOMe-α-D-dA (23) , the three compounds all

have pD-independent ∆H° values (see below) which discards any effect of the orientation of the

nucleobase in the former.

5.1.2 The balance of ∆H° and -T∆S° in neutral α- and β-D-N

In neutral α-D-ddNs 17 - 20, α-D-dNs 21 - 23, 25, 26 and α-L-dNs 27 - 29, the ∆H°

contribution (favouring S-type conformations) to the free-energy ∆G° of the two-state N � S

equilibrium prevails over the counteracting -T∆S°; Only in the case of α-D-ddG (18) they

cooperate, with slightly stronger -T∆S° than ∆H°. However, the absolute contribution of -T∆S° to

the free-energy ∆G° of the N � S equilibrium in α-nucleosides is negligible (0.2 kJmol-1) only in

the case of α-D-ddA (17). Unlike in dNs and rNs, where the bias of the N � S equilibrium can be

explained by the complex interplay of gauche and anomeric effects, the conformation of the sugar

moiety in ddNs is almost exclusively driven by the overall effect of the nucleobase, the contribution

of the 5'CH2OH group being minimal. Therefore, α-D-ddNs and β-D-ddNs constitute the systems of

choice to examine whether the magnitude of the O4'-C1'-N1/9 anomeric effect is dictated by the

configuration of the pentofuranosyl moiety at C1'. In the β-series (30 - 36), the effect of the

nucleobase drives toward N-type sugars (and cooperates with the much smaller effect of 5'CH2OH),

as evident from the positive ∆H° values (Table 2, Section 4). In α-D-ddNs 17 - 20, owing to the

change of configuration at C1', oN

HΔ values of the N � S equilibrium stabilize S-type conformations,

in agreement with the fact that nO4' →σ∗

C1'-N1/N9 stereoelectronic interactions, favouring S- over

N-type conformations, prevail over the counteracting steric effect of the nucleobase (driving to the

N) as found in the β-anomers. Since oN

HΔ values are the main contribution to oN

G values, S-type

sugars are stabilized at room temperature (vide supra). However, the comparison of oN

GΔ and oN

HΔ

values of the N �S equilibrium in β-D-ddNs 30, 31, 33 and 34 and in α-D-ddNs 17 - 20 shows that

the magnitude of the free-energy and enthalpy stabilization of the N-type sugars in the former is ≈ 2

- 6 times and ≈ 2 - 8 times greater, respectively, than the stabilization of S-type conformers in the

later, depending on the nature of the nucleobase. This means that the pentofuranose moiety in β-D-

ddNs is more prone to adopt multiple conformations as a result of the change of the temperature

than the α-counterparts.

In α-D/L-dNs 21 - 29, oN

HΔ as well as the overall oN

GΔ stabilize more S-type conformations

than in the parent α-D-ddNs (Table 2). This is in agreement with the observation that in the former

both the effect of the nucleobase and the [O3'-C3'-C4'-O4'] gauche effect are expected to prefer S-


104

type pseudorotamers, whereas in the later, only the effect of the nucleobase is operating. Larger

negative oN

HΔ values for α-D/L-dNs 21 - 29 than for β-D/L-dNs 37 - 39, 41 - 43 and 46 - 49 are

consistent with the fact that in the former both the [O3'-C3'-C4'-O4'] gauche effect and the effect of

the nucleobase stabilize the S-type sugars whereas in the latter the gauche effect counteracts the

effect of the nucleobase driving the conformation toward N-type geometries. However, the

additional enthalpy or free-energy stabilization of S-type conformations in the α-series is less than

or equal to -2.6 kJmol-1 for almost all compounds. oN

HΔ of α-D-dC (25) appears to be the only

exception, since its sugar conformation is driven much more efficiently toward S-type geometries

than that of β-D-dC (42) (Table 2).

In view of the above remarks and since both 5'CH2OH and the steric bulk of the nucleobase

are conserved structural features in both α-and β-D-ddNs, it is reasonable to attribute the lack of

flexibility of the sugar conformation in the former compared with the latter to much less efficient

O4'-C1'-N1/9 anomeric effect in the α-series, owing to the change of configuration at C1'.

5.1.3 Weakening of 5'-substituent effect in α- compared with β-nucleosides

The weakening of the 5'CH2OMe effect in the α- compared to the β-series is shown by the

smaller value of ∆∆H°1d [for 3'-OMe-α-D-dA (22)] compared with ∆∆H°1b [for 3'-OMe-β-D-dA

(38)].

5.1.4 3'-gauche effect weakens anomeric effect in α-D-dN compared to α-D-ddN

The subtraction (∆∆H°13

) of oN

HΔ of abasic sugar 12 from that of a particular α-D-ddN 17 -

20 provides an estimate for the magnitude of the overall effect of the nucleobase in α-D-ddNs. Thus

the ability of the nucleobase to promote the stabilization of S-type conformations increases as

follows (in kJmol-1): guanin-9-yl in 18 (-0.8) < thymin-1-yl in 20 (-1.5) < adenin-9-yl in 17 (-2.1) <

cytosin-1-yl in 19 (-3.3). The subtraction (∆∆H°14

) of oN

HΔ of 13 from that of an α-D-dN 21 or 24 -

26 gives the strength of the effect of the nucleobase in α-D-dNs, assuming that the magnitude of the

[HO3'-C3'-C4'-O4'] gauche effect in α-D-dNs and in abasic sugar 13 is the same. The ability of

∆∆H°14 to stabilize S-type pseudorotamers increases as follows (in kJmol-1): thymin-1-yl (0.1) <

guanin-9-yl (-0.4) < adenin-9-yl (-0.9) < cytosin-1-yl (-3.0). Thus, upon substitution of H3" in α-D-

ddN 17 - 20 for 3'-OH in α-D-dNs 21 and 24 - 26 the effect of each nucleobase is slightly reduced in

the latter compared to the former. This may originate from the change of the electrostatic potential

around O4' in the 2'-deoxy compared to the 2',3'-dideoxy counterparts, as it becomes also involved

in the [HO3'-C3'-C4'-O4'] gauche effect in the former. A similar trend has also been noticed in

going from β-D-ddNs to β-D-dNs. (Section 4). However the weakening of the effect of the

nucleobase in the β-D-dNs (calculated by subtracting ∆∆H°3 from ∆∆H°2) is much more

pronounced than in the α-D-dNs (subtraction of ∆∆H°14 from ∆∆H°13) [the weakening is: 2.9

kJmol-1 for β-D-dA compared to β-D-ddA but only by 1.2 kJmol-1 for α-D-dA compared α-D-ddA,

1.7 kJmol-1 for β-D-dG compared to β-D-ddG but only by 0.4 kJmol-1 for α-D-dG compared α-D-

ddG, 2.8 kJmol-1 for β-D-dC compared to β-D-ddC but only by 0.3 kJmol-1 for α-D-dC compared



105

α-D-ddC, 2.3 kJmol-1 β-D-T compared to β-D-ddT but only 1.6 kJmol-1 for α-D-T compared α-D-

ddT].

5.1.5 Weakening of the 3'-gauche effect in α-D/L-dN compared with β-D/L-dN

∆∆H°7 gives an estimate for the [HO3'-C3'-C4'-O4'] gauche effect in abasic sugar 2 (-4.5

kJmol-1) (Table 6). In order to assess the contribution of the [HO3'-C3'-C4'-O4'] gauche effect to the

drive of the sugar conformation in α-D-dNs 21 and 24 - 26, we have subtracted (∆∆H°17) from their

oN

HΔ values those of the parent α-D-ddNs 17 - 20. ∆∆H°17 is weakest in α-D-T (20) (-2.9 kJmol-1)

and strongest in α-D-dC (19) (-4.2 kJmol-1). Thus the ability of the gauche effect of [HO3'-C3'-C4'-

O4'] to stabilize S-type conformations in α-D-dNs is of the same order of magnitude as in abasic

sugar 2 but it is also much reduced than in the β-D-dNs counterparts (compare ∆∆H°17 and ∆∆H°10

in Table 6). A qualitative trend between the reduced magnitudes of the effect of the nucleobase and

of the [HO3'-C3'-C4'-O4'] gauche effect in α- compared to β-D-dNs emerges from the comparison

of ∆∆H°17 and ∆∆H°10, on one hand, of ∆∆H°14 and ∆∆H°3, on the other. The comparison of

∆∆H°18 and ∆∆H°12 values shows that both in α- and in β-nucleosides, the substitution of 3'-OH for

3'-OMe results in a similar small stabilization of S-type conformations.

5.2 The relative magnitude of the anomeric and gauche effects in the ionic states

5.2.1 Virtualy identical pKa values of the nucleobase in α- and β-nucleosides

The pKa values, derived from pD-dependent 1H chemical shifts (Section 3.7) of the

constituent nucleobases in α-D-ddNs versus β-D-ddNs counterparts, on one hand, in α-D-dNs

versus β-D-dNs, on the other, are almost identical (within ± 0.2 and ± 0.3 pD unit in the ddN and

dN series, respectively) (Table 2). This suggests that the electronic character of the nucleobase

remains unchanged as the configuration of the pentofuranose at C1' is inverted. It also implies that

the influence of the configuration of the sugar moiety at C1' upon the magnitude of the O4'-C1'-

N1/9 anomeric effect cannot be attributed to the modulation of the electronegativity of the glycosyl

nitrogen in the α- compared to the β-series.

5.2.2 Predominant enthalpy over entropy in the ionic states of α-D-ddN and -dN

As in the neutral solution, in the P state of α-D-ddA (17), α-D-ddG (18) and α-D-ddC (19), opHΔ s responsible for the free-energy stabilization of S-type pseudorotamers, since it overrides the

counteracting -T opSΔ term. In the D state, enthalpy and entropy have nearly the same strength, but

whereas they counteract each other in α-D-ddT (resulting in nearly unbiased pseudorotational

equilibrium), in α-D-ddG they both stabilize S-type sugars. In the P and D states, as in the N state,

the enthalpy or free-energy stabilization of S-type conformers in α-D-ddNs is much reduced

compared to the corresponding preference for N-type forms in the β-D-ddNs counterparts. Both in

the P and D states of α-D-dNs 21 - 26, ∆H° prevails over the opposing -T∆S° as reflected in the

overall preference for S- over N-type sugars at room temperature.

5.2.3 Weaker anomeric effect in α-D-ddN gives poorer flexibility


106

In going from the N to the P state of α-D-ddA (17) and α-D-ddC (19) and from the N to the

D state of α-D-ddG (18), both ∆H° and -T∆S° contributions to ∆G° of their N � S equilibria remain

unchanged and the preference of their constituent sugar moieties for S-type conformations remains

the same (Table 2). This means that the free-energy of their protonation � deprotonation

equilibrium is not transmitted to enhance (for 17 and 19) or weaken (for 18) the N-aglycone effect in

the P state (D for 18) compared to the neutral pD. This is in sharp contrast with what has been found

for the β-D-ddNs counterparts (Section 4.8). pD-dependent conformational preferences in α-D-

ddNs have only been observed upon protonation of guanin-9-yl in α-D-ddG (18) [ oNP

G−

ΔΔ = -1.5

kJmol-1] owing to the enhancement of the N-aglycone substituent effect [ oNP

H−

ΔΔ = -8.3 kJmol-1] and

upon deprotonation of thymin-1-yl oND

G−

ΔΔ = 0.3 kJmol-1] owing to the weakening N-aglycone

substituent effect oND

H−

ΔΔ = 0.6 kJmol-1]. These ∆∆H° and ∆∆G° values are however much smaller

than in the β-counterparts, confirming our above conclusion on relative ability of nucleobase in α-

and β-anomers to steer the sugar conformation.

5.2.4 The interplay of pD-independent ∆H° and ∆S° in α-D-dN

In all α-D-dNs [except α-D-dG (24)], protonation and deprotonation of the nucleobase has

almost no effect on the enthalpy of the two-state N � S equilibrium, showing that the efficiencies of

the effect of the N-aglycone and of the [HO3'-C3'-C4'-O4'] gauche effect are insensitive to the pD.

This result is in agreement with our finding for α-D-ddNs, and it is in sharp contrast with the

situation in β-D-dNs, in which a clear pD-dependent modulation of the anomeric effect has been

observed, as shown by non negligible ∆∆H° values (Section 4.8).

However, whereas in α-D-dA (21) the preference for S-type conformations remains the same

at all pDs, owing to the fact that the entropy term is a constant factor, in the pyrimidine nucleoside

α-D-dC (25), we have observed an entropy stabilization of S-type conformers in the P compared to

the N state [ oNP

)S−

Δ−( T = -1.5 kJmol-1]. Conversely, in α-D-T (26), as thymin-9-yl becomes

deprotonated, N-type pseudorotamers become less unstable than in the neural solution, which is also

the result of an entropy effect [ oND

)S−

Δ−( T = 0.7 kJmol-1]. A comparison of oNP

G−

ΔΔ and oND

G−

ΔΔ values

for α-D-dC, α-D-T and their β-D-counterparts shows that in fact, the entropy in the former

modulates more efficiently the bias of the two-state N � S equilibrium than the combined pD-

dependent enthalpy and entropy in the latter.

In α-D-dG (24), as guanin-9-yl is protonated, both the effect of the N-aglycone [ oNP

)S−

Δ−( T = -

6.2 kJmol-1) and the counteracting -T∆S° contributions [ oNP

)S−

Δ−( T = 3.7 kJmol-1] are stronger than

in the N state, and further stabilize S- and N-type sugars, respectively. On the overall, since oNP

)H−

(Δ

> Δ oNP

)S−

Δ−( T , the population of S-type pseudorotamers is higher in the P state than in the N state.

The strengthening of the N-aglycone effect in going from the N to the P state is however much

weaker in α-D-dG (24) than in α-D-ddG (18), showing the effect of 3'-OH in the former (compare

their oNP

H−

ΔΔ values).

5.2.5 Poor correlation of the nature of aglycone with pseudorotation in α-D-N



107

The correlation plots of the pD-dependent 1H chemical shifts as a function of ∆G° of the N

� S equilibria in α-D-ddNs (17 - 20), α-D-dA (21), α-D-dG (24), α-D-dC (25) and α-D-T (26) all

give straight lines (Fig 18, Panels (A) - (H)). The Pearson's correlation coefficients for the plots are

above 0.96 for α-D-ddNs and above 0.99 for the corresponding α-D-dNs, except for α-D-ddA (17,

R = 0.62) and α-D-dA (21), straight line with infinite slope) owing to virtually pD-independent ∆G°

values. For α-D-ddNs, α-D-dA (21) and α-D-dG (24), the slopes of the plots are much larger than in

the case of β-anomers, showing the less efficient transmission of the pD-tunable electronic character

of the nucleobase to drive the sugar conformation. Conversely, the pD-dependent entropy alone is

more efficient to drive the sugar conformation in α-D-dC (25) and α-D-T (26) than the combined

pD-dependent enthalpy and entropy in β-D-dC (42) and β-D-T (43), since, in absolute value, the

slopes of the correlation plots in the former are much smaller than those of the later.

6. Quantitation of the anomeric effects in C- and N-nucleosides

We herein report on our initial24,25 and revised32,33 pD-dependent conformational studies on

C-nucleosides 56 - 62. Basing on the results of these works, we have developed a new method for

the estimation25,33 of the stereoelectronic nO4'

→σ∗C1'-N9 anomeric effect in β-D-A, β-D-dA, β-D-G

and β-D-dG through a set of pairwise comparisons of ∆H° of their N � S equilibria with those of

purine C-nucleosides, which were used as reference points for the quantitation of the counteracting

steric effect of the N-nucleobases.

6.1 The anomeric effect in C-nucleosides

The interplay of nO4'→σ∗C1'-N9 interactions and of the counteracting steric effect determines

the overall effect of a nucleobase upon the sugar conformation in an N-nucleoside. The relative

importance of both contributions can be assessed by monitoring the magnitude of the enthalpy of the

C1'-substituent effect (∆H°C1'-subst.) in comparison with ∆H° of all steric and stereoelectronic forces

driving the N � S equilibrium: If the nucleobase acts as a C-substituent, it preferentially takes up a

pseudoequatorial orientation, which in turn shifts the N � S equilibrium to the S (i.e. negative

∆H°C1'-subst.

value), whereas predominant stereoelectronic interactions stabilize N-type sugars in

which the nucleobase is pseudoaxial (i.e. positive ∆H°C1'-subst.

value). If both terms are of equal

strength and cancel each other, ∆H°C1'-subst.

is zero.

The strength of the (stereoelectronic) anomeric effect alone can only be estimated when the

magnitude of the counteracting steric term is known. If one can engineer a nucleoside X in which

the nucleobase (isosteric to that of the N-nucleoside in which we want to quantitate the anomeric

effect) drives the N � S equilibrium exclusively through its steric effect (stereoelectronic

interactions being minimal), ∆H° of the N � S equilibrium in X can be used as a reference point for

the steric effect of the nucleobase in the N-nucleoside. Making use of this observation, we have

seeked for a family of nucleosides in which the effect of the nucleobase exerts its influence as much

as possible through its steric effect, while the opposing stereoelectronic component remains

negligible.

This has led us to turn our attention to C-nucleosides, since we expected that substitution of

the glycosyl nitrogen in N-nucleosides for C1' in C-nucleosides would result in minimal


108

stereoelectronic interactions between O4' and the nucleobase. In the crystal structures of N-

nucleosides, the O4'-C1' bond is shorter (by ≈ 0.03 Å) than the C4'-O4' bond, supporting the

existence of the O4'-C1'-N1/9 anomeric effect (Section 2.8). In contrast, no such clear trend is

observed in the crystal structures of C-nucleosides: in general the difference between O4'-C4' and

O4'-C1' bond lengths is much reduced (∆ ≈ 0.01 Å) compared with N-nucleosides: For 4-thio-ψ-

uridine271 (three molecular structures in the crystalline cell), ψ-isocytidine hydrochloride272, (α-D)

epishowdomycin monohydrate273, 3'-deoxyformycin A hydrochloride274, 5'-chloro-3',5'-dideoxy β-

L-formycin A monohydrate275, 1-benzyl-2'-deoxyshowdomycin276, oxoformycin B277, formycin A

monohydrate278 and α-pseudouridine monohydrate279, C4'-O4' and C1'-O4' bonds have nearly the

same length (∆ ≈ 0.01 Å), whereas C4'-O4' is longer than C1'-O4' in formycin A hydrobromide

monohydrate280 (C4'-O4' = 1.47 Å, C1'-O4' = 1.41 Å), formycin B277 (C4'-O4' = 1.46 Å, C1'-O4' =

1.42 Å) and its hydrochloride281 (C4'-O4' = 1.46 Å, C1'-O4' = 1.43 Å). ψ-uridine crystallizes in the

form of two structures282. In one of them C4'-O4' (1.45 Å) is longer than C1'-O4' (1.42 Å), but in

the other it is the opposite (C4'-O4' = 1.42 Å, C1'-O4' = 1.44 Å).

We have shown that in β-D-rNs 50 - 55, the modulation of the bias of the N � S

equilibrium by the pD of the D2O solution can be attributed to the resulting fine tuning of the effect

of the nucleobase (or to the anomeric effect alone, assuming that the steric effect is a constant factor

at any pD) and of the [HO2'-C2'-C1'-N1/9] gauche effect (Section 4.8). In ribo C-nucleosides 56 -

62, the [HO2'-C2'-C1'-N1/9] gauche effect is absent. Therefore, any differences observed in the

conformational preferences of the sugar moiety in 56 - 62 at various pDs is owing to the different

steric and electronic nature of their C1'-aglycones.

6.1.1 Effect of the C1'-pyrimidine aglycone on the conformation of the sugar

C-nucleosides differ from natural N-nucleosides by the unique carbon-carbon link between

C9 of purines or C5 of pyrimidines and C1' of the pentofuranose sugar. Many of them have been

isolated as antibiotics, and have antiviral and/or anticancer activity166,535-540. Their presence in

tRNAs is absolutely vital to the biochemical function541,542. N-nucleosides are known to carry the

genetic information, but our quantitative and other qualitative studies have also shown that their

conformation, and in turn, biological function, can be engineered by changing the nature of either

the nucleobase, the C2'-C4' substituents or the electronic character of the endocyclic O4' atom upon

their substitution. Very little is known on how the absence of a glycosyl nitrogen in C-nucleosides

affects their conformation and to which extent the stereoelectronic partnership of their nucleobase

and pentofuranose moieties is affected by this modification. It is necessary to examine carefully

whether the absence of N1/9 in the C1'-aglycone really results in the complete cancellation of its

stereoelectronic interaction with O4', as suggested by part of the above data on X-ray crystal

structures of C-nucleosides.

The conformational analysis Ψ-isocytidine (59), its hydrochloride (59b), 1-Me-Ψ-uridine

(61) and 1,3-diMe-Ψ-uridine (62) was initially24 performed in neat D2O solution using the

methodology described in Section 3. The thermodynamics (Table 11) of the N �S equilibria in 59,

59b, 61 and 62, suggest that: (i) Only in Ψ-isocytidine (59), ∆H°, driving to S-type sugars, prevails



109

clearly over the counteracting -T∆S°, therefore at 298 K, S-type sugars are preferred. (ii) In the

hydrochloride (59b), ∆H° (driving to the N) and counteracting -T∆S° almost cancel each other, and

the pseudorotational equilibrium is unbiased. (iii) In 1-Me-Ψ-uridine (61) and 1,3-diMe-Ψ-uridine

(62), -T∆S° is the predominant factor driving the conformation to the S over counteracting ∆H°. We

have estimated the effect of C1'-aglycones in 59, 59b, 61 and 62 by subtracting (i.e. ∆∆H°30 in Fig

13) ∆H° of 14 from their ∆H°, which gives (in kJmol-1): -2.5 (59) < 0.5 (61) < 1.6 (62) < 3.5 (59b).

To calculate ∆∆H°30 the strength of the [O3'-C3'-C4'-O4'] and [O2'-C2'-C1'-O4'] gauche effects and

of the 5'CH2OH substituent effect is assumed to be the same in C-nucleosides and in 14. Only

isocytosin-5-yl in 59 adopts preferentially a pseudoequatorial orientation in S-type sugars, however

this preference is rather small. The comparison of ∆∆H°30 values for 59, 61, 62 and 59b suggests

that only in 59, isocytosin-5-yl really acts as a C-substituent to drive the sugar conformation toward

S-type pseudorotamers. In all other cases, ∆∆H°30 is slightly positive, which indicates a drive

toward N-type conformation, and suggests that the steric effect of the nucleobase (stabilizing S-type

sugars) is not the predominant factor contributing to its overall effect. This prompts us to suggest

the existence of stereoelectronic interactions (see below) between the pentofuranose sugar and the

nucleobase, i.e. the nO4'

→σ∗

C1'-C5(sp2) anomeric effect by analogy with what we found for the N-

nucleosides counterparts (Section 4).

Table 11. The ΔH˚ and ΔS˚ of N � S equilibrium in C-nucleosides 56 - 59, 61 and 62 from our

initial studies at a single pD in native D2O solutiona

Compounds ΔH˚ ΔS˚ -TΔS˚ ΔG298 %S278 b %S358 b ∆% S b (358-278 K)

formycin (A) (57)c -7.7 (0.7) -14.7 (0.9) 4.4 -3.3 83 69 -14

formycin (B) (56)c -7.1 (0.6) -13.4 (0.8) 4.0 -3.1 81 68 -13

9-deaza-A (58)c -5.2 (1.2) -5.4 (2.0) 1.6 -3.6 83 75 -8

Ψ-isoC (59)d -2.1 (0.3) -3.0 (2.0) 0.9 -1.2 63 58 -5

Ψ-isoC hydrochloride (59b)d 3.9 (0.2) 12.5 (0.6) -3.7 0.2 45 55 +10

1-Me-Ψ-U (61)d 0.9 (0.2) 5.1 (0.8) -1.5 -0.6 55 58 +3

1,3-diMe-Ψ-U (62)d 2.0 (0.2) 7.9 (0.5) -2.4 -0.4 52 57 +5

a ΔH°, -TΔS° (at 298 K) and ΔG298 are given in kJ/mol, ΔS° in J/molK. b ∆% S (358-278 K) = %S358 - %S278 shows

the change of population of South type sugar owing to the net result of ΔH° and -TΔS° contribution as a function of

temperature c Data taken from ref. 25. d Data taken from ref.24.

The remarkable shift of the N � S equilibrium of 59b with respect to that of 59 to the N

upon protonation of isocytosin-5-yl is easily explained on the basis of the results of our ab initio

calculations on 5-methyl-isocytosine and N1H+-5-methyl-isocytosine suggesting that C5=C6 has a

higher double bond character (since it is shorter) in 59b than in 59. In 59, C5=C6 π-electrons will

easily be delocalized in the pyrimidine ring, and much less accessible for an interaction with nO4'

than in 59b, where the protonated guanidine skeleton results in the more localized C5=C6 π-

electrons, leading to a stronger steroelectronic anomeric effect nO4' →σ∗C1'-C5(sp2) .


110

The main results of this work are as follows: (i) Pyrimidine C1'-aglycones in 59, 61 and 62

do not provide good reference points for the estimation of the stereoelectronic O4'-C1'-N1 anomeric

effect in the pyrimidine N-nucleosides, since they do not act exclusively as C-substituents, as shown

by negative ∆∆H°30 values, owing to non negligible nO4' →σ∗

C1'-C5(sp2) stereoelectronic

interactions. (ii) Protonation of isocytosin-5-yl in 59b results in the strengthening of these

stereoelectrionic interactions, as in the case of the anomeric effect in N-nucleosides.

6.1.2 Effect of the C1'-purine aglycone on the drive of the sugar conformation

Our initial conformational analyses25 on purine C-nucleosides formycin B (56), formycin A

(57) and 9-deaza-A (58) have also been performed in D2O solution at neutral pD (see the next

sections for the discussion of the protonation state of 9-deaza-A in this work). The resulting

thermodynamics of their N � S equilibria are presented in Table 11, and their comparison suggests

the main following conclusions: (i) The pseudorotational equilibrium in 56 - 58 is driven toward the

S-type conformation by the ∆H° term, which prevails over the counteracting -T∆S° contribution to

the free-energy and therefore S-type puckered geometries are favoured at room temperature. (ii) The

C1'-aglycone in 56 - 58, in sharp contrast with the situation in the pyrimidine C-nucleosides

counterparts 59 and 61 - 62, acts as an efficient C-substituent pushing the pseudorotational

equilibrium toward S-type conformations. In the purine series, under the present experimental

conditions, stereoelectronic interactions nO4' and the σ*C1'-C9(sp2) orbital are rather inefficient in

comparison with the counteracting steric effect of the C1'-aglycone (see the next paragraph for the

discussion on the case of 9-deaza-A). The strength of the overall C1'-substituent effect upon the

drive of the sugar conformation in 56 - 58 has been estimated by subtracting from their ∆H° values

that of the abasic sugar 14, which yields: ∆∆H°30 = -8.1 kJmol-1 (formycin A) ≈ -7.5 kJmol-1

(formycin B) > -5.6 kJmol-1 (9-deaza-A). The ">" sign indicates the increased ability of the

nucleobase to shift the pseudorotational equilibrium toward S-type conformation through their steric

effect. Therefore, in this work, the C1'-aglycone in formycin B (or formycin A) is the most

pseudoequatorially oriented among those of 56 - 58 and it represents the best reference point for the

estimation of the steric effect of adenin-9-yl in β-D-dA (37) and β-D-A (50) or guanin-9-yl in β-D-

dG (41) and β-D-G (51) (see below).

6.1.3 pD-tunable anomeric effect in C-nucleosides

The thermodynamics of the N � S equilibrium in formycin A (56), formycin B (57), 9-

deaza-A (58), Ψ-isoC (59), Ψ-U (60), 1-Me-Ψ-U (61) and 1,3-diMe-Ψ-U (62) have been

subsequently determined32 over part or the whole 0.5 - 12.0 pD range in order to examine whether,

as suggested above, stereoelectronic nO4' →σ∗C1'-C5/9(sp2) interactions also participate to the drive

of the conformation of the pentofuranose sugar, in a similar manner to the anomeric effect operating

in N-nucleosides (Section 4). We argued, that if an anomeric effect would also exist in C-

nucleosides, it should also be reflected in the shift of the bias of their two-state N �S equilibria



111

toward more N- and S-type conformations owing to increased and decreased electron-withdrawing

character of the nucleobase in the P and D states, respectively, compared with the N state.

6.1.4 Transmission of the nature of the C-aglycone drives the N�S equilibrium in C-

nucleosides

In this revised study, we found that ∆H°, -T∆S° and ∆G° values of the N � S equilibrium in

all C-nucleosides 56 - 61, including the purine derivatives 56 - 58, are pD-dependent (Table 232,

Fig. 11, Panel (F)). Just as in the case of β-D-rNs, the energetics of the protonation � deprotonation

equilibrium of the nucleobase are transmitted to steer the sugar conformation through the

modulation of its overall effect. This means that our previous hypothesis of the minimal

contribution of the anomeric effect to the drive of the sugar conformation in purine C-nucleosides

56 - 58 was an oversimplification. Assuming that the steric effect of the nucleobase is constant over

the whole pD range, it is possible to attribute the pD-dependent conformational preferences of the

pentofuranose sugar in 56 - 61 to the tuning of the nO4' →σ∗

C1'-C5(sp2) stereoelectronic interactions.

The experimental pD-dependent ∆H°, -T∆S° and ∆G° values for 56 - 61 have been fitted to

the Henderson-Hasselbach equation to give the limiting values in the P, N and D states and the pKa

of the constituent nucleobases, as described in Section 3. These pKa values are nearly identical to

those found in the literature34,385-390, and to our independent estimates derived from pD-dependent

chemical shifts of aromatic and anomeric protons (Table 2). The efficient communication of

stereoelectronic information between constituent nucleobase and pentofuranose is further evidenced

by the fact that the plots (Fig 18) of the chemical shifts of the aromatic protons as a function of pD-

dependent ∆G° values all give straight lines with high correlation coefficients (> 0.96 except for

formycin B, owing to the smaller change in ∆G° values in the alkaline pD range, Table 2). In each

of the P, N and D states, the pentofuranose moieties in purine C-nucleosides 56 - 58 prefer more S-

type conformations than in the pyrimidine counterparts, in agreement with what we found for β-D-

dNs and β-D-rNs (Table 2).

6.1.5 Estimates for the thermodynamics of the N � S equilibrium

For formycin A (57) and formycin B (56), Ψ-isoC (59), 1-Me-Ψ-U (61) and 1,3-diMe-Ψ-U

(62), almost identical oN

HΔ , -T oNSΔ and o

NGΔ values were found in our earlier24,25 and latest32 studies

[the largest difference is 1.0 kJmol-1 for oN

HΔ of formycin B, which is within the standard deviation

of both estimates]. For 9-deaza-A (58), the comparison of our earlier estimates (Table 11) and those

from the updated study (Table 2) suggested that in the earlier work, the pD of the aqueous solution

was in the acidic range, since in the P state (pD ≤ 5.2), opHΔ , -T o

pSΔ and opGΔ are respectively: -7.4 (σ

= 0.8), 3.7 (σ = 0.6) and -3.6 (σ = 0.4) kJmol-1 whereas in the N state oN

HΔ , -T oNSΔ and o

NGΔ are -

14.2 (σ = 1.5), 9.1 (σ = 1.1) and -5.0 (σ = 0.4) kJmol-1 (Table 2).The above standard deviations,

taken at a single pD in the original publication32, are higher than the errors shown in Table 2. The

difference of 2 kJmol-1 between opHΔ and -T o

pSΔ values of the earlier and latest work is within the

sum of the standard deviations of the estimates.

6.1.6 Enhanced anomeric effect upon protonation of the aglycone


112

In the P state, formycin B (56), formycin A (57) and 9-deaza-A (58) prefer S-type

conformations, owing to cooperative opHΔ and -T o

pSΔ values (in 56), or to predominant opHΔ over the

counteracting -T opSΔ (in 57 and 58). In contrast, in the P state of protonated Ψ-isoC, o

pHΔ is slighlty

stronger than the opposing -T opSΔ contribution, which results in the slight stabilization of the N-type

conformers. Protonation of formycin B at N3387,388, formycin A at N3385,386, 9-deaza-A at N3 (as

shown by our pD-dependent 13C chemical shifts in ref. 34) and of Ψ-isoC at N1389 shifts their N �

S equilibria toward more N-type forms until full protonation of the heterocycle is achieved, which is

reflected in the plateau of ∆G°, ∆H° and -T∆S° values in the P state (Table 2, Fig. 11, Panel (F)).

The additional stabilization of N type sugars in the P compared to the N state is given by oNP

G−

ΔΔ

values (in kJmol-1): 2.0 (for 56), 1.4 (57), 1.4 (58) and 1.9 (59). The corresponding oNP

H−

ΔΔ values

(kJmol-1) are as follows: 7.6 (56), 5.7 (57), 6.8 (58) and 6.1 (59). The change of the steric effect of

the nucleobase in 56 - 59 upon protonation cannot account for these oNP

H−

ΔΔ and oND

G−

ΔΔ values: As

the nucleobase is protonated, its steric effect will presumably increase, therefore this should shift the

pseudorotational equilibrium toward more S-type conformations (with pseudoequatorially oriented

nucleobase), however we observe the opposite. This validates our hypothesis of the tuning of nO4'

→σ∗

C1'-C5(sp2) stereoelectronic interactions by the pD of the aqueous solution.

6.1.7 Weaker anomeric effect upon deprotonation of the aglycone

Deprotonation of formycin A at N7 does not affect the bias of its N � S equilibrium

compared with the N state. However, deprotonation of Ψ-isoC (59), at N3, of Ψ-U (60) at N1 and

N3, and of 1-Me-Ψ-U (61) at N3 results into the increased population of S-type sugars in

comparison with the N state, until a plateau in the D state is reached when the nucleobase is fully

deprotonated. The additional stabilization of S-type sugars in the D state compared to the N state is

reflected in the oND

G−

�ΔΔ values (in kJmol-1): -1.6 (for 59), -1.7 (60) and -0.8 (61). The

corresponding oNP

G−

ΔΔ values (kJmol-1) are as follows: -1.6 (59), -1.7 (60) and -0.8 (61).

6.1.8 Quantitation of the anomeric effect in C-nucleosides

We have estimated the enthalpy of the stereoelectronic interactions operating in C-

nucleosides 56 - 62 by subtracting ∆H° of abasic sugar 14 from their ∆H° values (∆∆H°30). The

results of these subtractions are compiled in Table 8. The perusal of ∆∆H°30 values shows that as

purine C-nucleosides 56 - 58 become protonated, the contribution of stereoelectronic interactions to

the overall effect of the C1'-aglycone increases, whereas the effect of deprotonation is insignificant.

For pyrimidine C-nucleosides 60 - 62 in the N state, the steric effect is overriden by the anomeric

effect, whereas in the D state it is reverse. Finally, in protonated Ψ-isoC (59), nO4' →σ∗C1'-C5/9(sp2)

interactions predominate, but as the base becomes neutral and deprotonated, their magnitude is

steadily reduced and they become overriden by the counteracting anomeric effect.

6.1.9 Comparison of pD-induced flexibility in C- and N-nucleosides

The extent of the stabilization of the N-type (or S-type) pseudorotamers in 56 - 61 upon

protonation (deprotonation) is dictated by the electronic character of the nucleobase: oNP

G−

�ΔΔ values



113

are of the same order of magnitude in purine C-nucleosides 56 - 58 and in the N-counterparts β-D-A

(50) and β-D-G (51). However, oND

G−

ΔΔ is negligible for 56 and 57 whereas it is -1.3 kJmol-1 in β-D-

G (51). At the other end of the flexibility scale, oNP

G−

ΔΔ for Ψ-isoC (59) is ≈ 5 times larger than for

β-D-C (52) and oND

G−

ΔΔ for Ψ-U (60) is ≈ 8.5 times larger than for β-D-U (54).

6.1.10 Correlation of the effect of the aglycone and its electronic nature

Formycin A (57) and 9-deaza-A (58) differ only in the nature of the fused five-membered

ring. The more efficient deactivating pyrrazolo ring in the former compared with pyrrolo ring in the

later is evident from the lower pKa value of the nucleobase in 57 with respect to 58, and results in

stronger nO4' →σ∗C1'-C5(sp2) stereoelectronic interactions in 57 than in 58 (i.e. less negative ∆G° and

∆H°, Table 2). Although in the N state, the anomeric effect in formycin B (56) and formycin A (57)

has the same strength (compare their oN

GΔ and oN

HΔ values), in the P state the stabilization of the

N3H+ charge is more efficient through delocalization in the amidine moiety in the later than in the

pyrimidone ring in the former, therefore the fused pyrrazolo ring will be more electron-deficient in

56 than in 57 and conversely nO4' →σ∗C1'-C5(sp2) stereoelectronic interactions are stronger in 56 than

in 57. Whereas oN

GΔ and oN

HΔ of formycin B and A are the same, they drive the sugar conformation

in Ψ-U (60) more to the N than in Ψ-isoC (59), showing that replacing the amidine function for an

amide function in the six-membered ring of 60 compared with 59 has much more effect on the sugar

conformation than when this change is made in the remote fused pyrimidone ring of 56 compared

with pyrimidine in 57, which are both not directly attached at C1'. The comparison of ∆G° and ∆H°

of the N � S equilibrium in Ψ-isoC (59) and Ψ-U (60) shows that they stabilize more S-type

conformations in in the former than in the latter, owing to more electron-rich pyrimidine ring in 59

than in 60.

Interestingly, the latest study from our laboratory40 on the preferred conformation of the

constituent pentofuranose sugar in benzene, pyridine and pyrimidine C-nucleosides in aqueous

solution has allowed to establish a clear qualitative correlation between the electronic character of

the C-aglycone and the magnitude of the O4'-C1'-C(aglycone) anomeric effect (Table 12): As the

electron-deficient character of the C-aglycone increases (i.e. in the benzene derivatives, in the order:

anilino (119) < benzyl (117) < α-naphthyl (118)), the nO4'

→σ∗C1'-C(aglycone,sp2) stereoelectronic

interactions are strengthened and counteract more and more efficiently the steric effect, as indicated

by the more positive ∆H° for the drive of the two-state N � S equilibrium toward N-type

conformation.

In the pyridine derivatives, it was shown that the effect of a substituent in the para position

(with respect to the sugar moiety) upon the electronic character of the pyridine ring is negligible,

since negative ∆H° values of the N � S equilibrium in compounds 120 and 122 stabilize S-type

pseudorotamers to the same extent as C1'-benzyl in C-nucleoside 117, owing to relatively weak nO4'

→σ∗C1'-C(aglycone,sp2) orbital mixing. In contrast, it was also found that the inductive (-I) effect of the

ortho substituent promotes a strong O4'-C1'-C(aglycone) anomeric effect, which cancels the

counteracting steric effect, as evident from negligible ∆H° values for compounds 121 and 123.


114

6.2 Quantitation of anomeric effect in N-nucleosides using C-nucleoside as reference

The drive of the N � S equilibrium in β-D-rNs 50 - 55 toward S-type conformations is the

result of the interplay of six steric and stereoelectronic forces: (i) The effect of 5'CH2OH, (ii) The

stereoelectronic component to the overall effect of the nucleobase (i.e. nO4'

→σ∗

C1'-N9

stereoelectronic

interactions), (iii) The counteracting steric effect of the nucleobase, (iv) the gauche effect of [HO3'-

C3'-C4'-O4'], (v) the gauche effect of [O2'-C2'-C1'-O4'] and (vi) the gauche effect of [O2'-C2'-C1'-

N9] (Section 4). Therefore, the anomeric effect of adenin-9-yl in β-D-A (50) or guanin-9-yl in β-D-

G (51) can be estimated using Eq 13:

AE of adenin-9-yl in β-D-A (50) or guanin-9-yl in β-D-G (51) =

∆H°(β-D-A or β-D-G) - [∆H°(14) + ∆H°GE[O2'-C2'-C1'-N9] + (∆H°ref. C-nucl. - ∆H°(14))] .....Eq 13

In Eq 13, ∆H°GE[O2'-C2'-C1'-N9] represents the strength of the [O2'-C2'-C1'-N9] gauche effect in β-D-A

(50) or β-D-G (51). The subtraction of ∆H°(14) from the experimental ∆H° value of 50 or 51 [∆H°(β-

D-A or β-D-G)] gives an estimate for the resultant of the effect of the nucleobase (stereoelectronic +

steric) and the ∆H°GE[O2'-C2'-C1'-N9] gauche effect in 50 or 51.

Our preliminary conformational study25 suggested that among purine C-nucleosides 56 - 58,

the nucleobase in formycin B (56) and A (57) prefers more pseudoequatorial orientations in S-type

conformations than 9-deaza-adenin-9-yl in 9-deaza-A (58), which was experimentally evidenced by

the larger negative ∆H° values for 56 and 57 in comparison with 58 (Table 11). Therefore, we have

initially quantitated25 the anomeric effect of adenin-9-yl in β-D-A (50) and of guanin-9-yl in β-D-G

(51) in their N state using the average (-7.4 kJmol-1, Table 11) of the ∆H° values of the N � S

equilibrium in formycin B (56) and A (57) as ∆H°ref. C-nucl. in Eq 13. In that work, ∆H°GE[O2'-C2'-C1'-

N9] (-6.3 kJmol-1) was derived from a regression analysis similar to regression (A), based on a total

set of 30 compounds, including β-D-dAMP (64), β-D-AMP (74), β-D-dAMPEt (69), β-D-dGMP

(65), β-D-GMP (75), β-D-dGMPEt (80), β-D-dCMP (66), β-D-CMP (76), β-D-dCMPEt (71), β-D-

dUMP (68), β-D-UMP (78), β-D-TMP (67) and β-D-TMPEt (72). Eq 13 was therefore rewritten as

Eq 13a:

AE of adenin-9-yl in β-D-A (50) or guanin-9-yl in β-D-G (51) in the N state =

∆H°(β-D-A or β-D-G) - [0.4 - 6.3 -7.8] .....Eq 13a

Using -4.6 kJmol-1 and -3.2 kJmol-1 as estimates for ∆H°(β-D-A or β-D-G) of β-D-A (50) and

β-D-G (51) (these values are in agreement with ∆H° values in the N state in Table 2, within the

error of the estimates), we found that the strength of the AE of adenin-9-yl in β-D-A (50) is slightly

weaker (9.1 kJmol-1) than that of guanin-9-yl in β-D-G (51) (10.5 kJmol-1).

In Eq 13a, we considered that formycin A and B in their N state constitute the best reference

points for the quantitation of the steric effect of an N-nucleobase. However, as discussed in Section

6.1, we have recently experimentally evidenced32 that the extent of the stabilization of S-type

conformations in 56 - 58 is dictated by the protonation state of the constituent nucleobase, therefore

the most pseudoequatorially oriented nucleobase is 9-deaza-adenin-9-yl in the N state of 9-deaza-A



115

(58) ( oN

HΔ = -14.2 kJmol-1, in the pD range from 8.8 to 12.0, Table 2), not those of formycin A and

B in the N state. Taking this fact into account, we have recently revised the estimates from Eq 13a

using Eq 13b: AE of adenin-9-yl in β-D-A (50) or guanin-9-yl in β-D-G (51) at a certain pD:

pD = ∆H°pD(β-D-A or β-D-G) - [∆H°(14) + ∆H°GEpD[O2'-C2'-C1'-N9] + ( oN

HΔ (58) - ∆H°(14))] .... Eq 13b

O

OH

OH OH

O

OH

OH OH

O

OH

OH OH

NH2

O

OH

OH OH

N

F

O

OH

OH OH

N

Br

HN

O

O

OH

OH OH

NH

OO

OH

OH OH

117 118 119

121

120

122 123

1234

56

1

2345

61

23

456

1

3

6

24

5

123

45

6

OH

OH OH

124

NH

H N

N

O

O

Table 12: Dependence of the Thermodynamicsa of the Two State N � S equilibrium upon the

Electronic Naturea of the C-aglycone.

Benzene-

derivativesb

Pyridine-

derivativesb

Pyrimidine-

derivativesb

(117) (118) (119) (120) (121) (122) (123) (59) (60)

ΔH° -5.4

(0.8)

0.4

(0.6)

-7.3

(1.0)

-4.4

(0.6)

0.4

(0.5)

-5.6

(0.9)

-0.1

(0.5)

-1.8

(0.3)

0.6

(0.2)

ΔS° -8.7

(2.7)

2.7

(1.7)

-13.1

(1.0)

-4.4

(2.0)

5.2

(1.7)

-7.4

(2.0)

0.7

(1.7)

-2.0

(1.1)

4.0

(1.1)

-ΤΔS° 2.6

(0.8)

-0.8

(0.5)

3.9

(0.8)

1.3

(0.6)

-1.5

(0.5)

2.2

(0.6)

-0.2

(0.5)

0.4

(0.3)

-1.2

(0.3)

∆G298 -2.8

(0.5)

-0.4

(0.5)

-3.4

(0.4)

-3.1

(0.4)

-1.1

(0.1)

-3.4

(0.5)

-0.3

(0.2)

-1.4

(0.2)

-0.6

(0.1)

%S298 76 53 80 78 60 80 53 63 57

Actual

AE c -5.8 0.0 -7.7 -4.8 0.0 -6.0 -0.5 -2.2 0.3

a The ΔH°, -ΤΔS° (at 298 K) and ΔG298 are in kJ/mol. The standard deviations (σ) are in parentheses. For 59 (pD =

7.0) and 60 (pD = 6.9), data are taken from ref.32. b In this work, the steric contribution of the substituents in ΔH˚

could not be dissected because of the unavailability of the corresponding saturated system. c The actual anomeric effect

(AE) was obtained by a simple subtraction (i.e. ∆∆H°30

) of the ΔH° of the N � S pseudorotational drive of 1-deoxy-β-

D-ribopentofuranose (14) (ΔH˚ = 0.4 kJ/mol)20 from the ΔH° of a specific C-nucleoside.


116

(A)

pD

0 2 4 6 8 10 12

ΔΔH1

1

-1.0

-0.5

0.0

0.5

1.0

(B)

pD

0 2 4 6 8 10 12

ΔΔH11

-4

-2

0

2

4

(C)

pD

0 2 4 6 8 10 12

ΔΔH10

-10

-9

-8

-7

(D)

pD

0 2 4 6 8 10 12

ΔΔH1

0

-25

-20

-15

-10

-5

(E)

pD

0 2 4 6 8 10 12

ΔΔH10

+ ΔΔH11

-10

-9

-8

-7

(F)

pD

0 2 4 6 8 10 12

ΔΔH1

0 + ΔΔH1

1

-20

-15

-10

-5

(G)

pD

0 2 4 6 8 10 12

AE (A)

16

18

20

22

24

(H)

pD

0 2 4 6 8 10 12

AE (G)

10

15

20

25

30

35

40

(I)

pD

0 2 4 6 8 10 12

AE (dA)

14

15

16

17

18

19

(J)

pD

0 2 4 6 8 10 12

AE (dG)

12

14

16

18

20

22

Figure 19. Estimates (kJmol-1) for gauche and anomeric effects driving the sugar conformation in β-D-dA (37), β-D-

A (50), β-D-dG (41) and β-D-G (51) from pairwise comparisons (Fig 13, Table 6, Section 6.2). The pD-dependent

strength of the 2'-OH effect (i.e. [HO2'-C2'-C1'-N9] + [HO2'-C2'-C1'-O4'] gauche effects) in β-D-A [Panel (A)] and β-

D-G [Panel (B)] was estimated from ∆∆H°11. The pD-dependent strength of the [HO3'-C3'-C4'-O4'] gauche effect in β-

D-dA [Panel (C)] and β-D-dG [Panel (D)] is reflected in ∆∆H°10 values. The strength of the pD-dependent [O2'-C2'-

C1'-N9] gauche effect in β-D-A [Panel (E)] and β-D-G [Panel (F)] was calculated by adding the respective ∆∆H°10 and

∆∆H°11 at each pD. The modulation of the strength of the anomeric effect of adenin-9-yl in β-D-A and β-D-dA and of

guanin-9-yl in β-D-G and β-D-dG by the pD of the aqueous solution is shown in Panels (G), (I), (H) and (J)

respectively.

Eq 13b allows to calculate the AE in 50 and 51 as a function of pD. ∆H°pD(β-D-A or β-D-G)

represents the enthalpy of the N � S equilibrium of 50 or 51 at a certain pD in the range 0.6 - 12.0.

∆H°GEpD[O2'-C2'-C1'-N9] represents the magnitude of the [O2'-C2'-C1'-N9] gauche effect in 50 or 51

at a certain pD, which has been calculated using a procedure consisting of four steps: (i) We have



117

first determined the strength of the overall 2'-OH effect (i.e. [O2'-C2'-C1'-O4'] + [O2'-C2'-C1'-N9]

gauche effects) in 50 and 51 by subtracting from their experimental ∆H°

values at each pD in the 0.6

- 12.0 range those of their 2'-deoxy counterparts β-D-dA (37) and β-D-dG (40), respectively

(∆∆H°11 in Fig 13 and Table 6, Fig 19 for the plot of the change of ∆∆H°11 as a function of pD). (ii)

We have subsequently estimated the change in the strength of the [O3'-C3'-C4'-O4'] gauche effect in

β-D-dA (37) and β-D-dG (40) by subtracting from their experimental ∆H° values those from their

2',3'-dideoxy counterparts β-D-ddA (30) and β-D-ddG (31) (∆∆H°10 in Fig 13, Table 6 and Fig 19

for the plot at each pD). (iii) An estimate for the [O2'-C2'-C1'-O4'] gauche effect in 50 and 51 has

been subsequently obtained from pD-dependent -∆∆H°10 values (Table 6, Fig 13 and Fig 19)

assuming that it has an equal magnitude (and opposite sign) to the [O4'-C4'-C3'-O3'] gauche effect

in β-D-dNs. (iv) By adding the pD-dependent values for ∆∆H°10 and ∆∆H°11, the actual magnitude

of the [O2'-C2'-C1'-N9] gauche effect has been quantitated (Fig 19).

On the basis of these estimates, Eq 13b has been used to give the magnitude of the

stereoelectronic O4'-C1'-N9 anomeric effect of adenin-9-yl in 50 and of guanin-9-yl in 51 over the

whole pD range from 0.6 to 12.0 (Fig 19). Thus, the AE varies from 23.4 to 17.7 kJmol-1 from pD

1.2 to 7.0 for adenosine (50) and changes from 37.5 to 15.6 kJmol-1 from pD 0.6 to 11.6 for

guanosine (51).

In β-D-dA (37) and β-D-dG (41), the drive of the N � S equilibrium toward S-type

conformations is the result of the interplay of only four stereoelectronic forces: (i) The effect of

5'CH2OH, (ii) The stereoelectronic component to the overall effect of the nucleobase (i.e. nO4'

→σ∗

C1'-N9 stereoelectronic interactions), (iii) The counteracting steric effect of the nucleobase, (iv)

the gauche effect of [HO3'-C3'-C4'-O4'] fragment. In order to quantitate the stereoelectronic AE of

adenin-9-yl in 37 and of guanin-9-yl in 41, (ii), it is therefore necessary to subtract from the ∆H°

values of their two-state N �S equilibrium the contributions from (i), (iii) and (iv), according to Eq

14:

AE of 9-adeninyl in β-D-dA (37) or 9-guaninyl in β-D-dG (41) at a certain pD = ∆H°pD(β-D-dA or β-

D-dG) - [∆H°(13) + ( oN

HΔ (58) - ∆H°(14))]..... Eq 14

In Eq 14, ∆H°pD(β-D-dA or β-D-dG) represents the experimental ∆H° value of the N � S

equilibrium in β-D-dA (37) and β-D-dG (41) at the pD of interest, whereas ∆H°(13) and ∆H°(14)

denote the experimental ∆H° value for abasic sugars 13 and 14, respectively. The subtraction of

∆H° value of 13 from ∆H°pD(β-D-dA or β-D-dG) (referred as by ∆∆H°3 in Table 6) gives an estimate

for the pD-dependent overall effect of the nucleobase in 37 or 41. During this quantitation, the same

reference point (9-deazaadenin-9-yl in 9-deaza-A (58) in the N state) has been used to estimate the

steric effect of the nucleobase in 37 and 41. The AE of adenin-9-yl in β-D-dA (37) is weakened

from 18.0 to 14.8 kJmol-1 in going from pD 0.9 to pD 7.0, whereas the AE of guanin-9-yl in β-D-

dG (41) is reduced from 20.7 kJmol-1 to 13.8 kJmol-1 from pD 0.9 to pD 11.6.

An alternative strategy based on the cancellation of the stereoelectronic interactions between

the cyclopentane ring and the nucleobase in a carbocylic nucleoside to estimate the anomeric

effect of the nucleobase in N-nucleosides


118

In a recent work543, an alternative strategy has been developed to estimate the strength of the

stereoelectronic anomeric effect in purine nucleosides. The thermodynamics of the two-state N �S

equilibrium (suggested by ab initio calculations) in the pyrazolo[4,3-c]pyridine-carbaribo-C-

nucleoside (124) were estimated using our methodology. It was found that the cyclopentane ring in

124 adopts preferentially S-type conformation (∆H° = -11.6 kJmol-1) as does anionic 9-

deazadenosine (∆H° = -14.2 kJmol-1)32. The steric effect (∆H°steric = -12.0 kJmol-1) of the C1'-

pyrazolo[4,3-c]pyridine aglycone in 124 was estimated by subtracting ∆H° of 14 from that of 124.

The C-nucleobase in 124 was assumed to be isosteric to adenin-9-yl in β-D-ddA, β-D-dA and β-D-

rA and to guanin-9-yl in β-D-ddG, β-D-dG and β-D-rG. Subtraction of ∆H° of 12 and of ∆H°steric

from ∆H° of β-D-ddA and β-D-ddG yielded estimates for the anomeric effect (16.4 kJmol-1) of

adenin-9-yl and guanin-9-yl in these ddNs. The anomeric effect of adenin-9-yl in β-D-dA (14.7

kJmol-1) and of guanin-9-yl in β-D-dG (16.4 kJmol-1) was subsequently estimated by subtracting

∆H° of 124 and the strength of the [HO3'-C3'-C4'-O4'] gauche effect from ∆H° of each dN. The

anomeric effect of adenin-9-yl in β-D-A (13.9 kJmol-1) and of guanin-9-yl in β-D-G (15.3 kJmol-1)

was calculated by subtracting ∆H° of 124 and the strength of the [HO2'-C2'-C1'-N9] gauche effect

from ∆H° of each rN. These estimates are of the same order of magnitude as our own33, except for

adenosine, where a difference of 3.8 kJmol-1 has been found. This means that the 9-deazaadenin-9-

yl aglycone in the D state indeed takes up a maximal pseudoequatorial orientation in 57 and hence

serves a correct reference point for the estimation of the steric of purine N-nucleobase.

7. The interdependency of the sugar and phosphate conformation

The sugar moiety, the nucleobase and the phosphate backbone constitute the three structural

elements making ploynucleotide chains. We have discussed in Sections 2 - 6 how in nucleosides the

change of the electronic character (for instance upon the change of the pD of the aqueous solution or

complexation with metal ions) and steric bulk of the substituents at C1' - C5' allows to engineer

certain conformational preferences of the constituent pentofuranose sugar, as the result of tunable

stereoelectronic gauche and anomeric effects. The phosphate backbone is not itself an isolated

structural element, as suggested by the qualitative correlation between preferred orientation around

the ε torsion and sugar puckering modes in mononucleotides and RNA trimers. We report the results

of our investigations23,28 that have adressed the following fundamental questions: Does the sugar

conformation dictate the phosphate backbone torsions? Is there any preferred phosphate torsion that

steers the sugar conformation in a certain manner? Are there any correlated interdependencies of

endocyclic sugar torsions with the preferred phosphate torsions? In order to further understand the

forces that govern the stabilization of the tertiary structure of oligo- and polynucleotides, we have

examined and dissected the nature of fundamental intranucleotidyl interactions that contribute to the

drive of the sugar-phosphate backbone in DNA and RNA using simple model systems, i.e.

nucleoside 3'-ethylphosphates, both in the 2'-deoxy (69 - 73) and ribo series (79 - 83) in which the

effect of internucleotidyl base-base stacking on the drive of the sugar-phosphate backbone is

completely eliminated.



119

7.1 Methods to assess the preferred conformation across the phosphate backbone

The dependence of 3JC2'P3', 3JC4'P3' and 3JH3'P3' coupling constants on the torsion angle ε

(Fig. 20) is described by the following Karplus-type equations223,435,544-546 (Eqs 15a - 15c): 3J

H3'P3' = 15.3 cos2(ε + 120)- 6.2 cos(ε + 120) + 1.5 ..... Eq 15a

3JC4'P3'

= 9.1 cos2(ε)- 1.9 cos(ε) + 0.8 ..... Eq15b

3JC2'P3'

= 9.1 cos2(ε - 120)- 1.9 cos(ε - 120) + 0.8 ..... Eq 15c

In Eqs 15a - 15c, a perfect trigonal symmetry for the position of C2', C4' and H3' with respect

to the C3'-O3' bond is assumed. In our interpretation23,28 of the experimentally measured 3JH3'P3'

,

3JC4'P3'

and 3JC2'P3'

coupling constants for nucleosides 3'-monophosphates 64 - 73 and their 3'-

ethylphosphates 74 - 83, we have considered a two-state εt � ε- equilibrium for the following

reasons: (i) ε+ rotamers are not found in crystal structures of nucleotides, suggesting that this state is

forbidden435 on account of steric and electrostatic repulsions547-550 between O4' and phosphoryl

oxygen. (ii) Additionally, for ε+ rotamers, one expects 3JH3'P3' ≈ 23 Hz223,435, but our experimental

values23,28 are ≈ 7 - 8 Hz, therefore it is likely that the population of ε+ rotamers is negligible.

A rough estimate for the population of ε- rotamers can also be obtained551 from 4JH2'P3' using

Eq 16: x(ε-) = 4JH2'P (obs) / 4JH2'P (-) .... Eq 16, Where 4JH2'P (obs) represents the

experimental value of 4JH2'P coupling constant for the compound of interest and 4JH2'P (-)

designates the corresponding coupling constant in a pure gauche- (ε-) conformation (i.e. 2.3 Hz551).

The experimental time-averaged 3JH5'P5', 3JH5"P5' and 3JC4'P5' coupling constants can be used

to estimate the population of the staggered rotamers around the C5'-O5' bond via simple linear

relationships435 (Eqs 17a - 17b):

% βt = 100 x [25.5 - (3JH5'P5' - 3JH5"P5')] / 20.5 .... Eq 17a

% βt = 100 x (3JC4'P5' - 0.73 / 10.27) .... Eq 17b

When P5'-O5'-C5'-C4'-H4' lie in the same plane and form a W-type conformation551-553 [i.e.

(βt, γ+) rotamer], the 3JH4'P5' coupling constant is approximately 3 Hz and can be used as a marker

to recognize this conformer, since in all other cases the coupling constant will be much reduced. 3JH4'H5' and 3JH4'H5" are translated into the populations of the staggered γ+ [x(γ+)], γ- [x(γ-)]

and γt [x(γt)] rotamers554 using Eqs 18a - 18d which have been parametrized on the basis of

crystallographic data: 3JH4'H5' = 2.4 x(γ+) + 10.6 x(γ-) + 2.6 x(γt) .... Eq 18a

3JH4'H5" = 1.3 x(γ+) + 3.8 x(γ-) + 10.5 x(γt) .... Eq 18b

x(γ+) + x(γ-) + x(γt) = 1 .... Eq 18c

% γ+ = 100 x [13.3 - (3JH4'H5' + 3JH4'H5")] / 9.7 .... Eq 18d

Using Eq. 19, the value of the torsion angle δ [C5'-C4'-C3'-O3'] can be derived from that of ν3

for each of the N- and S-type pseudorotamers, since ν3 is itself related to the respective PN, Ψm(N),

PS and Ψm(S) values via Eq. 6a. Finally, owing to the low natural abundance of 17O (0.04 %), the

preferred α and ζ conformations in nucleotides cannot be determined from spin-spin coupling

constants: δ ≈ ν3 + 120° .... Eq 19


120

7.2 No correlation of sugar and phosphate conformation in 2'-dN-3'-ethylphosphates

We have elucidated33 the conformational preferences of the pentofuranose moieties in β-D-

dNMPs (64 - 68) and their 3'-ethylphosphates (69 - 73) using the methodology described in Section

3. The thermodynamics of the N � S equilibrium in 64 - 73 in the neutral solution are compiled in

Table 3. In order to assess the preferred orientation around the ε torsion in β-D-dNMPEts 69 - 73 in

the 278 K - 358 K range, we have developed the program epsilon, which allows to translate

experimental 3JC2'P3', 3JC4'P3' and 3JH3'P3' coupling constants into the geometries and relative

populations of the rotamers engaged in the two-state εt � ε- equilibrium. These vicinal coupling

constants did not significantly change (≤ 0.5 Hz) as the temperature was raised from 278 K to 358

K, as the result of nearly temperature-independent mole fractions of the εt and ε- rotamers (≈ 1:1

ratio). Similarly, 3JCH2P3' and 3JCH3P3' were nearly the same over the whole temperature range,

suggesting that the population of βt rotamers (≈ 50%) is not affected by the temperature. The

comparative analysis of ∆H°, -T∆S° and ∆G° values of 64 - 73 in Table 3 leads to the main

following conclusions: (i) In all cases, ∆H° prevails over the counteracting -T∆S° term and it is the

main factor responsible for the overall (∆G°) stabilization of S-type pseudorotamers at 298 K. (ii) S-

type pseudorotamers in β-D-dNMPs 64 - 68 are slightly more preferred than the β-D-dNs 37 and 41

- 44 counterparts as shown by more negative ∆G° values for the former in comparison with the

latter. This can be attributed to the slightly more electronegative 3'-OPO3H- group in the latter with

respect to 3'-OH in the former. The additional stabilization of S-type pseudorotamers in 64 - 68

through [-1/-2HO3PO3'-C3'-C4'-O4'] gauche effect is shown by the ∆∆H°22 values in Table 7 (Fig

13) which are slightly nucleobase-dependent and within the range from -2.5 to -1.2 kJmol-1. (iii)

Similarly, S-type pseudorotamers are slightly more favoured in β-D-dNMPEts 69 - 73 than in β-D-

dNs 37 and 41 - 44. This stabilization throught the stronger [-EtO3PO3'-C3'-C4'-O4'] gauche effect

is shown in the ∆∆H°23 values in Table 7 (Fig 13). (iv) For each nucleobase, the [-1/-2HO3PO3'-C3'-

C4'-O4'] gauche effect in 64 - 68 has been estimated by subtracting from their experimental ∆H°

values those of the corresponding β-D-ddNs 30, 31 and 33 - 35, yielding: ∆∆H°20 (kJmol-1) = -8.9

(adenin-9-yl), -7.5 (guanin-9-yl) = -9.8 (cytosin-1-yl), -8.0 (thymin-1-yl), -8.4 (uracil-1-yl). Thus

∆∆H°20 varies by ≈ ± 1kJmol-1 from the average value depending upon the nature of the

nucleobase. (v) Using the same strategy, we have quantitated the [-EtO3PO3'-C3'-C4'-O4'] gauche

effect in 69 - 73: ∆∆H°21 (kJmol-1) = -9.0 (adenin-9-yl), -8.2 (guanin-9-yl) = -10.2 (cytosin-1-yl), -

8.4 (thymin-1-yl), -8.4 (uracil-1-yl). Thus the comparison of ∆∆H°20 with ∆∆H°21 values shows that

the stereoelectronic gauche effects within [-1/-2HO3PO3'-C3'-C4'-O4'] and [-EtO3PO3'-C3'-C4'-O4']

fragments operate with nearly the same magnitude in the 2'-deoxynucleotides 64 - 68 and 69 - 73.

This is also proven by the negligible ∆∆H°24 values in Table 7.

We have also estimated the magnitudes of the gauche and anomeric effects operating in all β-

D/L-nucleosides (i.e. the dataset used for regression analysis (B))



121

as well as β-D-dNMPs 64 - 68 and β-D-dNMPEts 69 - 73 (Table 5, Section 4.1) from regression

(E), which has been performed using a dataset consisting of 40 experimental ∆H° values assuming

identical strengths for

the [-1/-2HO3PO3'-C3'-

C4'-O4'] and [-

1EtO3PO3'-C3'-C4'-O4']

gauche effects in 64 -

68 and 69 - 73. The

estimates resulting from

regression analysis (E),

its correlation

coefficient and the

standard error on the

estimates are virtually

the same as of

regression (B). The

average [3'-phosphate-

C3'-C4'-O4'] gauche

effect is about ≈ -8.3

kJmol-1, i.e. -2.0 kJmol-

1 stronger than the

[HO3'-C3'-C4'-O4']

counterpart.

Thus this work

shows that: (i) there is no straightforward correlation between preferred orientation across C3'-O3'

bond and most stable sugar puckering mode in the 2'-deoxynucleotides, (ii) the additional

stabilization of S-type pseudorotamers in 64 - 68 and 69 - 73 with respect to β-D-dNs 37 and 41 -

44 is nearly the same owing to equally efficient 3'-gauche effects.

7.3 Interaction of 2'-OH with vicinal 3'-phosphate in ribonucleotides

The thermodynamics28 of the N � S equilibrium in β-D-rNMPs 74 - 78 and β-D-rNMPEts

79 - 83 based upon pseudorotational analyses of temperature-dependent 3JHH coupling constants are

presented in Table 3. The pairwise comparison of ∆H° values of 74 - 83 with those of nucleosides

and other 2'-deoxynucleotides leads to following conclusions: (i) For all ribonucleotides (except β-

D-CMPEt (81), for which they are of equal strength and cancel each other), ∆H° contribution to the

free-energy ∆G° of the N � S equilibrium stabilizes S-type conformations over the counteracting

entropy, which prefers N-type sugars. (ii) The N � S equilibrium in β-D-rNMPs 74 - 78 is driven

more toward S-type conformations than in the parent β-D-rNs 50 - 55, as evident from negative

∆∆H˚25 values. ∆∆H˚25 (kJmol-1) is in the range -1.2 < ∆∆H˚25 < -0.5 for purines and -2.2 <

∆∆H˚25 < -0.8 for pyrimidines. This can be attributed to the stronger [-1/-2HO3PO3'-C3'-C4'-O4']

O

O

O

P

O

-O O

OH

O

PO

-O

C4'

PO3

H5"H5'

C4'

O3P

H5"H5'

C4'

PO3

H5" H5

'

C2' C4'

H3'

PO3

C2' C4'

H3'

O3P

PO3

C2' C4'

H3'

O5'

H5' H5"

O4' C3'

H4'

O5'

H5' H5"

C3' H4'

O4'

O5'

H5' H5"

H4' O4'

C3'

ON

OH

O

O

ON

OH

NH

NN

N

O

NH2

β

χ

α

εδ

γ

A

C

ζ

5'

3'

(n-1)

(n+1)

(n)

-------------------------------------------------

------------------------------------------------------

εt

ε-

Torsion angle

γt

ε+

εt

= 180º

γ−

ε-

= -60º

γ (O5'-C5'-C4'-C3')

β (P5'-O5'-C5'-C4')

βt

= 180º

γt

= 180º

ε+

= +60º

γ-

= -60º

γ+

γ+

= +60º

Torsion angle ε (C4'-C3'-O3'-P)

β-

= -60º β+

= +60º

Torsion angle

Figure 20. The description of the phosphate backbone conformation in

nucleosides and nucleotides by the α [(n-1)O5'-P5'-O5'-C5'], β [P5'-O5'-C5'-C4'], γ

[O5'-C5'-C4'-C3'], δ [C5'-C4'-C3'-O3'], ε [C4'-C3'-O3'-P3'] and ζ [C3'-O3'-P3'-

O5'(n+1)] torsion angles.


122

gauche effect in β-D-rNMPs compared with the [HO3'-C3'-C4'-O4'] gauche effect in the former.

(iii) In β-D-rNMPEts 79 - 83, S-type conformations are also more stable than in β-D-rNs 50 - 55 by

∆∆H˚26 = -2.5 kJmol-1 for purines and ∆∆H˚26 ≈ -3.7 kJmol-1 for pyrimidines, owing to the

stronger [-EtO3PO3'-C3'-C4'-O4'] gauche effect in 79 - 83 in comparison with the [HO3'-C3'-C4'-

O4'] gauche effect in 50 - 55. (iv) The [-EtO3PO3'-C3'-C4'-O4'] gauche effect in β-D-rNMPEts is

clearly stronger (by -2.8 to -1.3 kJmol-1) than the [-1/-2HO3PO3'-C3'-C4'-O4'] gauche effect in β-D-

rNMPs, as shown by ∆∆H˚28 values, whereas in the 2'-deoxy counterparts, they are the same (as

shown by nearly identical ∆∆H˚20 and ∆∆H˚21). (v) The overall 2'-OH effect drives the sugar

conformation toward S and N-type geometries in purine and pyrimidine β-D-rNMPEts, respectively,

as shown by the comparison of their conformational preferences with the β-D-dNMPEts

counterparts (i.e. ∆∆H˚27). In contrast, 2'-OH drives the conformation of the sugar moiety in β-D-

rNMPs to more N in comparison with β-D-dNMPs, as shown by negligible or positive ∆∆H˚29

values.

(vi) As the temperature is increased from 278 K to 358 K, the population of S-type conformers

decreases and the population of εt rotamers in β-D-rNMPEts 79 - 83 increases in a coopeative

manner, as shown by the concomittant change in 3JC2'P3', 3JC4'P3' and 3JH3'P3' coupling constants.

Table 13.The thermodynamics of the two-state N �S and εt � ε- pseudorotational equilibria in β-

D-rNMPEts 79 - 83 showing the identical values (within experimental error) for ΔG298 of N � S

pseudorotational and εt � ε- equilibria (see Fig 21 for the correlation plot).

Estimation of the drive of εt � ε- conformational equilibria derived from

temperature-dependent 3JHP and 3JCP

ΔHºε

a ΔSºε

a -TΔS˚ b ΔG

298

%ε-278 c

%ε-358 c Δ%ε- d

β-D-AMPEt (79) -6.6 (0.9) -14 (3) 4.2 -2.4 76 63 -13

β-D-GMPEt (80) -5.8 (0.9) -12 (3) 3.6 -2.2 74 62 -12

β-D-CMPEt (81) -2.8 (0.9) -9 (4) 2.7 -0.1 53 46 -7

β-D-rTMPEt (82) -3.8 (0.8) -8 (3) 2.4 -1.4 66 58 -8

β-D-UMPEt (83) -3.3 (0.7) -7 (2) 2.1 -1.2 64 57 -7

a ΔH˚ (kJ mol-1) and ΔS˚ (J mol-1 K-1) are the average values (standard deviations are given in brackets) and

were calculated from individual van't Hoff plots using populations of N and S pseudorotamers from several

individual PSEUROT analyses. ΔHºε and ΔSº

ε were calculated from 15 van't Hoff plots using populations of εt

and ε- rotamers. The signs of thermodynamic parameters are arbitrarily chosen in such a way that the positive

values indicate the drive of N � S and εt �ε- equilibria to N and εt, whereas the negative signs describe the

drive to S and ε-, respectively. b -TΔS˚ (kJ mol-1) term is given at 298 K. c The population of the S and ε-

conformers were calculated using the relation: %S (T) or %ε- (T) = 100 * [exp (- ΔGT/ RT)] / [exp (- ΔGT/ RT)

+1]. d Δ%ε- = %ε-358 - %ε-278.

(vii) The interdependency of the conformation of the sugar and phosphate moieties in β-D-

rNMPEts 79 - 83 is evidenced by the fact that the free-energies (∆G298) of the two-state N � S and

εt � ε- equilibria are the same (Table 2 and Table 13), within the experimental error of our

measurements and calculations (±0.5 kJ/mol). This is further evidenced by a simple correlation plot

of the temperature-dependent populations of ε- rotamers as a function of the temperature-dependent



123

population of the S sugar pseudorotamers (Fig 21, Panels (A1) - (E1)) and of ∆G° of the εt � ε-

equilibrium versus ∆G° of the N � S equilibrium for all β-D-rNMPEts (Fig 21, Panels (A2) - (E2)),

showing their straightforward correlation with the correlation coefficients between 0.79 and 0.98 (in

pyrimidine nucleotides, the relatively smaller values for the correlation coefficients are the result of

the limited change in ∆G° or xε- or xS as a function of temperature in comparison with the purines

counterparts).

(viii) The unique interaction of 2'-OH with the lonepair of the vicinal heteroatom (e.g. O3')

is able to act as a molecular switch between (N,εt) � (S,ε-) conformational equilibria in 79 - 83. An

alternative H-bond directly between 2'-OH and the 3'-phosphate oxygen is ruled out because they

are too far away from each other, even on rotation around α and ζ torsions. This 2'-OH....O3'

interaction stabilizes the S and ε- conformers ("On-Off" switch) in a cooperative manner over N and

εt, and it is experimentally evidenced by the fact that the difference in the chemical shifts of the

CH2 protons of 3'-ethylphosphate in 79 - 83 is steadily reduced as the temperature is increased (until

β-D-AMPEt (79)

% S

64 68 72 76 80 84

% ε

−

56

60

64

68

72

76

80

ΔGo

(N/S)

-3.2 -2.8 -2.4 -2.0 -1.6 -1.2

ΔG

o(ε

t/ε

−)

-3.5

-3.0

-2.5

-2.0

(A1) (A

2)

β-D-AMPEt (79)

% S

60 64 68 72 76

% ε

−

60

64

68

72

76 (B1)

β-D-GMPEt (80)

ΔGo

(N/S)

-2.8 -2.4 -2.0 -1.6 -1.2

ΔG

o(ε

t/ε

−)

-3.0

-2.5

-2.0

-1.5 (B2)

β-D-GMPEt (80)

β-D-CMPEt (81)

% S

46 48 50

% ε

−

44

48

52

56

ΔGo

(N/S)

-0.5 0.0 0.5

ΔG

o(ε

t/ε

−)

-1.0

-0.5

0.0

0.5(C

1) (C

2)

β-D-CMPEt (81)

% S

52 56 60 64

% ε

−

56

60

64

68

72

(D1)

β-D-rTMPEt (82)

ΔGo

(N/S)

-1.0 -0.5

ΔG

o(ε

t/ε

−)

-2.0

-1.5

-1.0(D

2)

β-D-rTMPEt (82)

β-D-UMPEt (83)

% S

54 57 60

% ε

−

54

57

60

63

66

ΔGo

(N/S)

-0.8 -0.6 -0.4

ΔG

o(ε

t/ε

−)

-1.5

-1.2

-0.9

-0.6(E1) (E

2)

β-D-UMPEt (83)

Figure 21. The correlation plots of the temperature-dependent population of the ε- phosphate backbone

rotamers versus the population of S-type conformers (at 298 K) [Panels (A1) - (E1)] and of ∆G°(εt/ε-) of

εt � ε- equilibrium versus ∆G°

(N/S) of N �S equilibrium [Panels (A2) - (E2)] in β-D-rNMPEts 79 - 83.

All plots show straight lines (s = slope, i = intercept, R = Pearson's correlation coefficient): For β-D-

AMPEt, s = 1.02, i = -5.78, R = 0.98 (A1) and s = 1.01, i = 0.62, R = 0.95 (A2), For β-D-GMPEt, s =

0.95, i = 4.10, R = 0.98 (B1) and s = 0.94, i = -0.25, R = 0.95 (B2), For β-D-CMPEt, s = 1.44, i = -19.9,

R = 0.79 (C1) and s =1.39, i = -0.25, R = 0.80 (C2), For β-D-rTMPEt, s = 1.37, i = -16.73, R = 0.93

(D1) and s = 1.30, i = -0.30, R = 0.86 (D2), For β-D-UMPEt, s = 1.96, i = -49.5, R = 0.94 (E1) and s =

1.94, i = 0.12, R = 0.85 (E2).


124

both protons become isochronous) owing to the increase in the population of N,εt-type conformers

in which no H-bond between 2'-OH and the vicinal phosphate occurs. The N � S equilibrium is

unbiased (ΔG298 = 0.1 kJ mol-1) for β-D-CMPEt (81) because of the strongly opposing anomeric

effect. (ix) The in-line attack of 2'-OH to the vicinal phosphate in the RNA self-cleavage

reaction10,555, giving a 2',3'-cyclic phosphate via a transient trigonal bipyramidal phosphorane as

intermediate, requires a S,ε- conformational state246,556,557. This cleavage is more facile whith

cytosin-1-yl as the nucleobase at the cleavage site than any other nucleobase558, presumably owing

to a smaller activation energy barrier for N- to S-type sugar interconversion. We, on the other hand,

have found that ∆G° of the (N,εt) � (S,ε-) conformational equilibrium in pyrimidine β-D-rNMPEts

is much reduced (≈ 0 kJmol-1 for β-D-CMPEt) than for the purine counterparts. (x)

Owing to the internucleotidyl stacking, the preferred conformational state in RNA-RNA559 or RNA-

DNA560 duplex is (N,εt). However, when these stacking interactions are absent such as in the single

stranded hairpin loop, (S,ε-) and gg C4'-C5' conformers are favoured, as in our model systems β-D-

rNMPEts 79 - 83.

8. Application of stereoelectronic effects in oligonucleotides

8.1 Design of antisense oligonucleotides via the gauche engineering

What is the structure of DNA-RNA hybrid?

The target for the antisense strand is the RNA. The mechanism of antisense effects of the

antisense NA involve either RNase H mediated cleavage of the RNA strand in the hybrid duplex

duplex561, or the physical blocking of the translation machinary562. Much is understood about the

RNase H promoted RNA excission from the hybrid DNA-RNA duplex than the physical blocking.

In order to be able to optimally design the antisense strand using the RNase H promoted RNA

excission of the DNA-RNA hybrid, it is important to understand first the structure of the DNA-

RNA hybrid both in the solution and in the solid state. The summary of various studies performed

in many labs can be divided into three categories, and they are as follows: (i) The RNA and DNA

strands of the chimeric duplex are similar to the corresponding A-RNA and the B-DNA563,564. (ii)

The RNA strand of the duplex is similar to A-RNA and the DNA strand is somewhat intermediate

between A- and B-forms of DNA560,563,564. (iii) In one crystalline duplex, the conformation of both

ribose and deoxyribose sugars were found to be in C3'-endo (i.e. N-type) conformation, but the

sugar residues in the DNA strand underwent a conformational transition to C2'-endo (i.e. S-type) in

solution565. The occurrence of all the above variations in the DNA-RNA hybrid clearly shows that

the RNA sugars are invariably C3'-endo, but the DNA sugars are flexible and can indeed take up

various conformations (O4'-endo to C1'-exo to C2'-endo, 72° < 180°).

Configuration-dependent drive of the sugar conformation

The effects of covalently linked substituents at C2', C3' and C5' upon the conformation of

the sugar moieties of the resulting nucleosides and the overall stability in modified oligonucleotides

have been extensively reviewed46,49. It has been demonstrated first by Remin et al 405 that inversion



125

of configuration at C2' from ribo to ara-cytidine induces further stabilization of the S-type

conformation (40% S for cytidine and 50% S for ara-C at 24°C in D2O at pH 7). This preference for

S-type conformation is dramatically increased as the hydroxyl groups are ionized at pH ≥ 13

because of intramolecular 5'-OH...O2'- H-bond (45%S for cytidine and 90%S for ara-C). Consistent

with this study, it has been found226 that the electronegative substituents such as 2'-F at C2' on the

α-face (i.e. 2'F-2'-deoxyribo analog) enhances the preference of the sugar moiety for N-type

pseudorotamers, whereas C2'-F in the β-face (i.e. 2'F-2'-deoxyara-analog) enhances the preference

of the sugar moiety for S-type pseudorotamers, thereby showing that the gauche effect is indeed

configuration-dependent.

The engineered conformational preference of nucleosides preorganizes the oligo conformation.

It has been shown64 that two South thymidines in a 12mer duplex greatly stabilized the

double helix, whereas two North thymidines destabilized it by inducing an A-B junction in the

middle of the duplex. The introduction of the North nucleosides into oligo-DNA will induce an

RNA-type geometry, and consequently improves its binding capability with the complementary

target RNA (but not the RNase H cleavage property; see below!). Consequently, the incorporation

of 2'F-2'-deoxyribonucleosides into a DNA strand458 stabilizes the DNA:RNA heteroduplex by

≈2°C per modification. The modified antisense oligodeoxynucleotide adopts an A-form

conformation, as a result of the drive of the sugar conformation toward N-type pseudorotamers by

the [F2'-C2'-C1'-O4'] gauche effect458,459.

The antisense NA-RNA hybrid should mimic the DNA-RNA hybrid conformation for good RNase H

cleavage

Consistent with the studies of Remin et al 405 that the ara-nucleosides have intrinsic

preference for the South-type conformation compared to ribo counterpart, fluoroarabinonucleic

acids (2'F-ANA)-RNA and arabinonucleic acid(ANA)-RNA duplexes showed a close CD

resemblance566 to the native DNA-RNA counterpart (suggesting that 2'F-ANA and ANA have

preorganized structures similar to DNA, and their hybrids with RNA have very similar A-like

helical conformations as the native DNA-RNA hybrid), but their melting studies showed that the

former had higher thermodynamic stability compared to latter. Consequently, the 2'F-ANA-RNA

duplex was cleaved by RNase H much faster than the ANA-RNA duplex. The susceptibility of the

2'F-ANA-RNA hybrids to RNase H was found to be very similar to that observed for native DNA-

RNA and DNA-thioate-RNA hybrids. This study showed that lower thermal stability of the ANA-

RNA hybrids translated into poorer digestion by RNase H although they had similar A-type helix,

whereas 2'F-ANA-RNA, native DNA-RNA and DNA-thioate-RNA have also similar conformation

but higher stabilities than ANA-RNA duplex, therefore they all showed faster cleavage. In

contradistinction, 2'F-2'-deoxyribonucleosides has a predominant N-type conformation as the native

ribonucleosides, and therefore 2'F-RNA has a preorganized RNA-type structure, which forms A-

type hybrid with a target RNA as the native RNA-RNA duplex, and, as expected, none of them

serves as a substrate to RNase H.


126

The antisense NA-RNA heteroduplex is more efficiently stabilized by 2'F-2'-deoxyribo- than

by 2'-O-methyl ribonucleosides567, which is consistent with the stronger gauche effect induced by

the former, owing to the higher electronegativity of fluorine compared with hydroxy458. The

presence of North 2'-fluoro-2'-deoxyguanosine in a modified hexadeoxynucleotide enables the

transition between A-DNA conformation at low ionic strength and Z-DNA at high salt

concentrations460. 2'-O-methyl-ribonucleosides567 when incorporated into an oligodeoxynucleotide

improves its binding affinity with the complementary RNA strand, but incorporation of 2'-SMe-46,49

or 2'-amino-2'-deoxyriboucleosides497 into the antisense strand results in the destabilization of the

resulting antisense-NA-RNA duplex because of the propensity of these modified nucleosides to

adopt S-type conformations, owing to the poorer electronegativity of 2'-SMe or 2'-NH2 group

comapred to 2'-OH.

Requirements for the design of the antisense strand for RNase H cleavage of the hybrid

The above discussion point to two distinct requirements for the design of the antisense

strand complementary to RNA target: (i) The antisense strand should mimic the preorganized

conformation of a native DNA strand in order to form a hybrid duplex with a target RNA with a

structure closely similar to the native DNA-RNA hybrid, and (ii) the affinity of this preorganized

antisense strand to the native RNA should be at least as high as the native DNA-RNA strand in

order to be a good substrate to RNase H. In contradistinction to these requirements, the necessity

for the antisense strand to mimic the preorganized conformation of a native DNA strand is often

overlooked. A major consideration is still given49 in the design of antisense strand is the increase of

the thermal stability of the antisense NA-RNA hybrid. Despite this, the great majority of the sugar

and backbone-modified nucleoside building blocks (for example, -S-CH2-O-CH2- or -CH2-NCH3-

O-CH2- backbone) have a preferred N-type conformation, thereby their antisense oligo is expected

to have a preferred RNA-type, not the DNA-type, preorganized structure. The increase of melting

point observed with the hybrid of these antisense NA with RNA target is therefore presumably

owing to RNA-RNA duplex type structure, which are not as good substrates for RNAse H as DNA-

RNA duplex. Regarding the relative stability of various types of duplexes, it is noteworthy that it

increases in the following order: DNA-DNA < DNA-RNA < RNA-RNA.

These studies are consistent with our own studies568 with ten different 5'-fluorophore

tethered DNA-RNA hybrids as substrates for RNAse H: We showed that among all these ten

different 5'-tethered chromophores, 5'-phenazine and 5'-phenanthridene tethered DNA-RNA hybrids

had the shortest half-life of digestion by RNase H because these antisense oligos showed higher

thermal stability (5-10°C higher than the native counterpart) and least deviation of the global

structure from the native heteroduplex.

Requirements for 3'-modification for engineering antisense DNA strand (the Gauche engineering)

It has been shown by our lab in 199426 that the distinct conformational preferences observed

for the pentofuranosyl moieties in various 3'-substituted (X)-2',3'-dideoxynucleosides [X = H, NH2,



127

OH, OMe, NO2, OPO3H-1/-2 and F] are closely related to the strength of the 3'-gauche effect exerted

by the 3'-substituent, X. This work demonstrated that the magnitude of the enthalpy of the 3'-gauche

effect increases with the increase of the electronegativity of the 3'-substituent. Thus, as the

electronegativity of the 3'-substituent decreases [F > OPO3H-1/-2 > NO2 > OMe > OH > NH2], the

preference for the North-type pseudorotamer increases [F < OPO3H-1/-2 < NO2 < Ome < OH <

NH2]. Thus 3'-amino-2',3'-dideoxynucleosides have been shown to possess a preferred N-type

conformation26,569 which indeed is conserved in the sugar puckers (66 of the 72 nucleotides per

asymmetric unit adopting C3'-endo conformation) in the N3' → P5' phosphoramidate 12mer DNA

duplex as evident by the A-type duplex conformation62. Moreover, the CD-spectra of the N3' P5'

phosphoramidate 12mer DNA duplex are consistent with adoption of an A-RNA conformation570.

8.2 Fused nucleosides to engineer preorganized DNA structures

We have seen above that the engineered conformational preference of the flexible

pentofuranose ring in nucleosides is capable of preorganizing the conformation of the

oligonucleotide duplex. This can be simply achieved by engineering the stereoelectronics of the

nucleosides, namely the gauche and the anomeric effects as described in the earlier chapters.

Clearly, such stereoelectronic engineering gives us the possibility to steer a specific conformation

depending upon the conformation of the target strand. The challenge in the stereoelectronic

engineering is to direct the sugar conformation to a specific low energy minima and stabilize it by

overcoming the intrinsic temptation to fall into other energy minima since the interconversions

between various pseudorotamers take place through a low energy barrier. Alternatively, one can

introduce appropriate rigidity to the pentofuranose or to the cyclopentane system in a nucleoside

analogue and trap it into one specific conformation. Introduction of these fused systems in oligos

have produced preorganization and appropriate rigidity of the single strand, which has been found to

play a critical role for the formation of stable duplexes. Several such systems have been so far

designed, and the readers are directed to a recent review article49.

8.3 Enzyme recognition by fused carbocyclic nucleosides of fixed conformation

Recently, a conformationally constrained 2E (N)-methano-carbocyclic AZT mimic [(N)-

4',6'-methano-carba-3'-azidothymidine-5'-triphosphate] was shown to be equipotent to AZT-5'-

triphosphate in inhibiting the target enzyme, reverse transcriptase, whereas the antipodal 3E

carbocyclic analogue [(S)-1',6'-methano-carba-3'-azidothymidine-5'-triphosphate] was inactive571.

When carbocyclic nucleosides are incorporated into a modified oligodeoxynucleotide59,572,

they are able to mimic the conformation of the neighbouring nucleotides owing to the relatively

more flexible nature of the cyclopentane ring (see below). In contrast, the incorporation of the

conformationally constrained 2E carbathymidine analogue [(N)-4',6'-methano carbathymidine] was

able to force a conformational change in the oligonucleotide that resulted in an enhanced stability of

DNA-RNA heteroduplexes573,574.

8.4 The conformational transmission in the self-cleaving Lariat-RNA


128

Small synthetic lariat RNAs 129a, 130a and 131a have been found557 to undergo site

specific self-cleavage to give an acyclic branched-RNA with 2',3'-cyclic phosphate and a

5'-hydroxyl termini as in 129b, 130b and 131b, which are reminiscent of the products formed in

some catalytic RNAs. Noteworthy is the fact that in these self-cleavage reactions of 129a, 130a and

131a one specific phosphodiester bond between G and C residues underwent intramolecular

transesterification reaction amongst all other phosphates moieties in the molecule (akin to the

splicing of pre-mRNA to give functional mRNA), suggesting that the self-cleaving transesterified

G C phosphodiester has enormously different chemical reactivity in comparison with all other

phosphates in the molecule. These lariat-RNAs 125, 126, 129a, 130a and 131a are much smaller

than the natural catalytic RNAs such as the hammerhead ribozyme (k = ~1 min-1 at 37 °C), and

their rate of the self-cleavage is also much slower (k = 0.25 x 10-4 min-1 for lariat hexamer 129a,

and 0.16 x 10-3 min-1 for lariat heptamer 130a at 22 °C). We have shown that the trinucleotidyl

loop in the tetrameric and pentameric lariat-RNAs245 126 is completely stable whereas the

tetranucleotidyl or pentanucleotidyl loop in the hexameric 129a or heptameric 130a lariat-RNA

RNA246,556,575 does indeed have the required local and global conformation promoting the self-

cleavage. It has been also shown that simple 2' 5' or 3' 5'-linked cyclic RNAs, 127 and 128,

respectively, are completely stable and their structures are considerably different from the self-

cleaving lariat-RNAs such as 129a or 130a. In our search to explore the optimal structural

requirement for the self-cleavage reaction of RNA, we showed that the unique 3'-ethylphosphate

function in 131a (in which the branch-point adenosine has a 2' 5'-linked tetranucleotidyl loop and

a 3'-ethylphosphate moiety, mimicking the 3'-tail of the lariat-hexamer 129a) is the key structural

feature that orchestrates its self-cleavage reaction (k = 0.15 x 10-4 min-1 at 19 °C) compared to the

stable 2' 5'-linked cyclic RNA 127. A comparative study of the temperature dependence of the N

� S equilibrium for the lariat-tetramer 131a and the 2' 5'-linked cyclic tetramer 127 shows that the

A1 residue in the former is in 92% S-type conformation at 20 °C, whereas it is only in 55% S in the

latter. This displacement of the N � S pseudorotational equilibrium toward the S-type geometry is

due to the enhanced gauche effect of the 3'-OPO3Et- group at the branch-point adenosine in 131a

compared to 3'-OH group in 127. This 3'-OPO3Et- group promoted stabilization of the S geometry

at the branch-point by ΔH ≈ 4 kcal.mol-1 in 131a is the conformational driving force promoting its

unique self-cleavage reaction. The comparison of ΔH° and ΔS° of the N � S pseudorotational

equilibria in 131a and 127 (see Table 14) clearly shows the remarkable effect of the

3'-ethylphosphate group in 131a in being able to dictate the conformational changes from the sugar

moiety of the branch-point adenosine to the entire molecule (conformational transmission). Thus

the S conformation in A1, U2 and C6 sugar moieties is clearly thermodynamically more stabilized

while it is considerably destabilized in G3 owing to the 3'-ethylphosphate group in 131a compared

to 127. This is a clear experimental evidence of conformational transmission through a crankshaft

motion originating from the gauche effeect exerted by the 3'- ethylphosphate moiety at the A1 sugar.

It is interesting to note that the magnitude of enthalpy and entropy for the North to South transition

of the A1 sugar in 127 is comparable to the enthalpy and entropy of transition between the A- and



129

B-form of the lariat hexamer556,557. This self-cleaving tetrameric lariat-RNA 131a to 131b is the

smallest lariat-RNA molecule hitherto known to undergo the self-cleavage reaction, and hence it is

the simplest model of the active cleavage site of the natural self-cleaving catalytic RNA.

8.5 The conformational transmission by dihydrouridine in RNA

It has been recently shown576 that S-type conformations of the pentofuranose sugar in

dihydrouridine (D) 3'-monophosphate are more stable (by ≈ 1.5 kcalmol-1) than in the

corresponding β-D-UMP (78). This effect is amplified for the D residue (5.3 kcalmol-1) in the

ApDpA trimer (compared to ApUpA), and is also further transmitted to the neighbouring 5'-

terminal A (3.6 kcalmol-1). This observation is consistent with the fact that the C5-C6 saturation of

the uracil-1-yl aglycone to the dihydrouracil-1-yl moiety reduces the electron-withdrawing character

of the former, and hence weakens the n →σ* interactions of O4'-C1'-N1 moiety. This is consistent

P

O

O

N

O

N

O-

O

O

O

NO

P- O

O

OO

O

O

OPO

P- O

O

HO

ON

O-

OH

N

N

HN

HO

HO

O

O

O

NH2

N

N

NH2

H2NN

H

N N

O

PO

H2N

O

N

O

N

O-

O

O

O

N

O

P- O

O

O

OO

O

P- O

O

HO

O

N

HN

HO

O

O

NN

NH2

H2N

N

H

N

N

O

N

O

P

O

O-

O

NH

HO

O

O

OH

N

O

HO

O P

O O-

NO

P O

O

N

O-

O

O

N

O

O

P- O

OH

O

O

OP

- O

O

HO

O

N

N

HN

HO

O

O

ON

N

NH2

H2NN

H

N

N

O

O

P- O

O

O

HO

O

N

N

H2N

O

P O

HO

N

O-

O

O

O

NO

P- O

O

O

O

O

P

- O

O

HO

O

N

N

HN

HO

O

O

ON

N

NH2

H2NN

H

N

N

O

O

P- O

O

O

HO

O

N

N

H2N

O

P

128

C6

G3

A1

U2

U2

C6

G3

A1

127

126125

A1

G3

U2

C4

G3

A1

U2

C5

U4


130

P- O

O

PO

NH2

O

O

NH2

O

N

N

O-

O

O

O

N

NO

P OO

PO

O-

P

- O

O

N

O-

O

O

O

O

NO

P- O

O

O

O

O

O

NH

HO

P

- O

O

HO

O

N

O

N

HN

HO

O

O

O

O O

HO

N

N

N

NH2

H2NN

H

N

N

O

O

P- O

O

O

HO

O

N

N

H2N

O

OHO

O

P

O

O-

O

NO

O

O

PO

PO

- OOH

O

N

O-

N

O

HN

HO

ONH

O

O

HO

O

N

NNH2

H2NN

H

N

N

O

O

P

- O

O

O

O

HO

ON

N

H2N

O

O

HO

N

OHO

O

P

O

O-

NO

O

HOO

N

HN

O

O

P- O

O

PO

NH2

O

O

NH2

O

N

N

O-

O

O

O

N

NO

P OO

PO

O-

P

- O

O

N

O-

O

O

O

O

NO

P- O

O

O

O

O

O

NH

HO

P

- O

O

HO

ON

O

N

HN

O

O

O

O

O O

HO

N

N

N

NH2

H2NN

H

N

N

O

O

P- O

OOH

HO

O

N

N

H2N

O

OHO

O

P

O

O-

O

NO

O

OPO

PO

- OOH

O

N

O-

N

O

HN

O

ONH

O

O

HO

O

N

NNH2

H2NN

H

N

N

O

O

P

- O

O

O

O

HO

ON

N

H2N

O

O

HO

N

OHO

O

P

O

O-

NO

OHHO

O

N

HN

O

O

P O

O

N

O-

O

O

O

NO

P- O

O

O

O

O

P

- O

O

HO

O

N

N

HN

HO

O

O

ON

N

NH2

H2NN

H

N

N

O

O

P- O

O

O

HO

O

N

N

H2N

O

P OCH3CH2O

PO

-

O

N

O-

O

O

O

NO

P- O

O

O

O

O

P

- O

O

OH

O

N

N

HN

O

O

O

ON

N

NH2

H2NN

H

N

N

O

O

P- O

OOH

HO

O

N

N

H2N

O

OCH3CH2O

PO

-

O

130a

131a

130b

131b

129b

at 22o C

k = 0.25 x 10-4

m-1

k = 0.16 x 10-3

m-1

k = 0.15 x 10-4

m-1

Cleavage

site

Cleavage

site

Cleavage

site

Some Synthetic Self-cleaving RNAs

at 19o C

at 22o C

129a



131

Table 14: The thermodynamicsa of for the N S equilibrium in 127, 129a & 131a

Compound Residue ΔH°

(kcal.mol-1)

ΔS°

(cal.K-1.mol-1)

-TΔS°b

(kcal.mol-1)

ΔG°

(kcal.mol-1)

Lariat- A1 -5.8 ± 1.3 -15.9 ± 3.8 4.7 ± 1.1 -1.1 ± 2.4

3'-Et-Phos U2 -7.1 ± 2.2 -21.3 ± 6.4 6.3 ± 1.9 -0.8 ± 4.1

RNA G3 12.3 ± 2.8 36.2 ± 8.2 -10.6 ± 2.4 1.7 ± 5.2

131a C6 1.2 ± 0.8 3.2 ± 2.5 -0.9 ± 0.7 0.3 ± 1.6

3'-OH A1 -1.8 ± 0.5 -5.5 ± 1.7 1.6 ± 0.5 -0.2 ± 1.1

RNA U2 -1.9 ± 1.2 -2.6 ± 3.8 0.8 ± 1.1 -1.2 ± 2.3

127 G3 6.1 ± 1.1 17.0 ± 3.2 -5.0 ± 0.9 1.2 ± 2.0

C6 6.6 ± 1.1 18.7 ± 3.4 -5.5 ± 1.0 1.1 ± 2.1

Hexameric A1 -1.2 ± 1.1 -0.7 ± 3.6 0.2 ± 1.0 -1.0 ± 2.2

Lariat- U2 -2.2 ± 1.0 -4.3 ± 3.0 1.3 ± 0.9 -1.0 ± 1.9

RNA G3 7.3 ± 1.2 22.1 ± 3.9 -6.5 ± 1.1 0.8 ± 2.4

129a C6 5.0 ± 1.1 16.2 ± 3.4 -4.8 ± 1.0 0.3 ± 2.1

a 90% confidence. b The -TΔS term is computed for 293 K.

with our observation that the strength of the anomeric effect of an aglycone increases as it becomes

more electron-withdrawing, and decreases as it becomes electron-rich. The transmission of the

stabilization of S-type conformations from the D nucleotide moiety to the 5'-terminal A moiety in

the ApDpA trimer most probably takes place through a crankshaft motion across the sugar-

phosphate backbone. As it has been stated above, the earlier example of this type of stereoelectronic

transmission through the crankshaft motion across the sugar-phosphate backbone has been

documented for the tetranucleotidyl lariat-RNA loop, in which placement of a 3'-ethylphosphate

moiety at the branch point not only stabilized the S-type conformation of the constituent sugar of the

nucleotide at the branch-point but also affected the conformation of all four other loop

nucleotides246,556,556.

Similar modifications of the electronic nature of five most common aglycones, say by

methylation or by any hypermodification (as found in tRNAs) or by complexation with any ligand

such as metal ion, drug or polypeptide will profoundly affect the local structure of the DNA and

RNA, which may itself locally alter the hydration pattern, H-bonding or electrostatics, and any one

of these, in turn, may serve as a recognition signal for a specific interaction and function.

8.6 Preferential recognition of 3'-anthraniloyladenosine by Elongation Factor Tu

The functional role of transfer RNA (tRNA) as adapter in translating the genetic code is well

recognized. The protein biosynthesis occurs via stepwise manner – (i) Initiation, (ii) Chain

elongation and (iii) Termination. The initiation step requires the assembly of a ribosome-mRNA-

tRNA ternary complex. During this process, amino acids are linked to tRNA by aminoacyl-tRNA

synthetases to form aminoacyl-transfer RNA (aa-tRNA). The binding of the aa-tRNA to the

ribosome is catalysed by the elongation factor Tu (EF-Tu), a GTP binding protein. The EF-Tu, in its


132

active form (EF-Tu.GTP), can specifically recognize the aa-tRNAs from the non-aminoacylated

tRNAs.

This precise recognition process depends on the complimentary structural features of the aa-

tRNA scaffold. From this viewpoint, the development of tRNA-mimics is an interesting approach to

understand the functional aspects of this biological process. Small molecule can mimic only a part

of a large tRNA. There are three such molecules, having tRNA mimicry, found in literatures – (i)

antibiotic puromycin [(a) Welch, M.; Majerfeld, I.; Yarus, M. Biocehmistry 1997, 36, 6614. (b)

Monroe, R. E.; Marcker, K. A. J. Mol. Biol. 1967, 25, 347], which corresponds to the 3'-end of a

tyrosinylated tRNA and found to be a powerful inhibitor in protein biosynthesis. However, it failed

to interact with EF-Tu or aaRS (ii) micro-tRNA, CCAOH trinucleotides, which has been shown to

mimic tRNA for aaRS recognition and can even be aminoacylated [Jovine, L.; Djordjevic, S.;

Rhodes, D. J. Mol. Biol. 2000, 301, 401] and (iii) 3'-O-anthraniloyladenosine (used in our present

study, see below). The differential recognition of 3'-O-anthraniloyladenosine (having 2'-endo sugar

conformation and puromycin (having 3'-endo sugar conformation) towards EF-Tu has enormous

biological significance.

Anthranilic acid charged yeast tRNAPhe or E. coli tRNAVal are able to form a stable

complex with EF-Tu*GTP, hence the

2'- and 3'-O-anthraniloyladenosines

and their 5'-phosphate counterparts

have been conceived to be the smallest

units that are capable to mimic

aminoacyl-tRNA577-582. Since the 3'-

O-anthraniloyladenosine (134) also

binds more efficiently to EF-Tu*GTP

complex compared to its 2'-isomer

(132), we delineated the stereoelectronic features that dictate the conformation of 3'-O-

anthraniloyladenosine (134) and its 5'-phosphate (135) vis-a-vis their 2'-counterparts (132 and 133)

as well as address how their structures and thermodynamic stabilizations are different from

adenosine and 5'-AMP. It has been found43 that the electron-withdrawing anthraniloyl group exerts

gauche effects of variable strengths depending upon whether it is at 2' or at 3' because of either

participation or absence of the [O2'-C2'-C1'-N9] gauche effect, thereby steering the pseudorotation

of the constituent sugar moiety either to the N-type or S-type conformation.

The 3'-O-anthraniloyladenosine 5'-phosphate has a relatively more stabilized S-type conformation

(ΔG° = -4.6 kJ/mol) than 3'-O-anthraniloyladenosine itself (ΔG° = -3.9 kJ/mol), whereas the ΔG°

for 2'-O-anthraniloyladenosine and its 5'-monophosphate are respectively -0.9 and -1.8 kJ/mol,

suggesting that the 3'-gauche effect of 3'-O-anthraniloyl group is stronger than that of 2'-O-

anthraniloyl in the drive of the sugar conformation. Since the EF-Tu can specifically recognize the

aminoacylated-tRNA from the non-charged tRNA, we have assesssed the free-energy (ΔG°) for this

recognition switch to be at least ≈ -2.9 kJ/mol by comparison of ΔG° of N � S pseudorotational

O

ROH2CO

ROH2C

O

HO

NN

N

NN

N

N

NH2

NH2

N

O

OH

O

H2N

O

NH2

132: R = H

133: R = OPO3H-1/-2

134: R = H

135: R = OPO3H-1/-2



133

equilibrium 3'-O-anthraniloyladenosine 5'-phosphate and 5'-AMP. The 3'-O-anthraniloyladenosine

and its 5'-phosphate are much more flexible than the isomeric 2'-counterparts as evident from the

temperature-dependent coupling constant analysis. The relative rate of the transacylation reaction of

2'(3')-O-anthraniloyladenosine and its 5'-phosphate is cooperatively dictated by the two-state N �S

pseudorotational equilibrium of the sugar, which in turn is controlled by a balance of the

stereoelectronic 3'- and the 2'-gauche effects as well as by the pseudoaxial preference of the 3'-O or

2'-O-anthraniloyl group. The reason for the larger stabilization of the 2'-endo conformer for 3'-O-

anthraniloyladenosine and its 5'-phosphate lies in the fact that the C3'-O3' bond takes up an optimal

gauche orientation with respect to C4'-O4' bond dictating the pseudoaxial orientation of 3'-

anthraniloyl residue, which can be achieved only in the S-type sugar conformation with adenin-9-yl

and the 2'-OH groups in the pseudoequatorial geometry, compared to the preferred C3'-endo sugar

with pseudoaxial aglycone and 2'-OH found in 3'-terminal adenosine moiety in the helical 3'-CCA

end of uncharged tRNA.

8.7 Conformational changes in nucleotides induced by interaction with metal ions

(i) Effect of Zn2+ or Hg2+ binding to the nucleobase on the sugar conformation in β-D-dG (41)

It has been shown by us30 that protonation at N7 in β-D-dG (41) shifts the N �S

pseudorotational equilibrium toward more N by ≈ 15% with respect to the neutral state, which was

attributed to the strengthening of the nO4' →σ∗

C1'-N9 anomeric effect at acidic pD owing to the

reduced electron-density in the imidazole ring compared with the neutral pD (Section 4.8). In

contrast, it has been found395 that binding of Zn2+ with N7 of the constituent guanin-9-yl in β-D-2'-

dG has a negligible effect on the population of the preferred S-type pseudorotamers. Similarly, upon

addition of 0.2 equivalent Hg2+, only a slight shift (by ≈ 4 % unit) of the two-state N � S

pseudorotational equilibrium toward more N-type has been found. These results suggest that

binding of Zn2+ or Hg2+ with N7 do not modulate efficiently the strength of the O4'-C1'-N9

anomeric effect. This is in agreement with the

observation395 that the resulting perturbation

(with respect to the uncomplexed/neutral

state, Section 2.9.3) of the electronic character

of the imidazole moiety is much reduced than

when N7 becomes protonated: Indeed, the

15N chemical shift of N7 in β-D-2'-dG

changes much less upon its complexation with Zn2+ (∆δ15N(N7) = 3.4 ppm) than upon its

protonation401 [∆δ15N(N7) = 46 ppm]. In the same study, it was also noted that hard Mg2+ did not

form any stable complex with N7 of the guanin-9-yl nucleobase in β-D-2'-dG.

(ii) Conformational changes observed across the sugar-phosphate backbone and in the

pentofuranose sugar in 5'- and 3'-methylmonophosphates of β-D-dG as the result of the interaction

of the nucleobase/phosphate moiety with Mg2+, Zn2+ and Hg2+

As an extension of the above work, Polak and Plavec have recently shown583 that the preferential

binding of Mg2+ with the phosphate oxygen atoms and C6=O carbonyl group in 5'- and 3'-

O

O

HO

O

P

MeO

-O

NH

NN

N

O

NH2

O

OH

O

NH

NN

N

O

NH2

PO

O- OMe

136: MepdG 137: dGpMe


134

methylmonophosphates of β-D-dG (i.e. MepdG (136) and dGpMe (137), respectively) does not alter

the bias of the N �S pseudorotational equilibrium toward S-type conformations (At 298K, 68% S

for MepdG and 75% S for dGpMe).

On the other hand, the interactions of Zn2+ or Hg2+ with N7 resulted in the small shift of the sugar

conformation in MepdG and dGpMe toward N (by ≈ 1 - 5%), as the result of the slight

strengthening of the anomeric effect (vide supra and Section 4.8). It was also observed that the

binding of the softer Zn2+ causes a larger shift of the syn � anti equilibrium toward anti than

complexation with hard Mg2+. Whereas the preferred orientation around the C4'-C5' (γ torsion

angle, in dGpMe) and O5'-C5' (β torsion angle, in MepdG) remain virtually unchanged in the metal

ion complexed state in comparison with the free state, interaction of hard Mg2+ or softer Zn2+ or

Hg2+ resulted in a small shift of the εt �ε- equilibrium toward εt rotamers. Thus, this work has

qualitatively shown that the small change of the electronic character of the nucleobase in dGpMe

upon the complexation of N7 with a soft metal ion resuts in relatively minor strengthening of the

O4'-C1'-N9 anomeric effect pushing the sugar conformation toward slightly more N-type forms, and

this additional

preference for N-

type sugars is in

turn transmitted to

control the

conformation of the

phosphate-

backbone by further

stabilizing εt

rotamers.

(iii) The change of

the electronic

character upon

Pt2+

binding to

guanin-9-yl nucleotide is transmitted to drive the conformation of the local sugar-phosphate

backbone

The anticancer drug cisplatin interferes with replication and transcription processes to form

crosslinks as major lesions by reacting with the N7 of guanines and adenines in DNA. The dynamic

microstructure alteration as a result of cisplatin binding to N7 of guanines in DNA has been recently

assessed45 through the multinuclear temperature-, pD- and concentration-dependent NMR study of

the effect of Pt2+ complexation to 2'-deoxyguanosine 3',5'-bisethylphosphate (138) and its ribo

analogue (140). Compounds 138 and 140 serve as models mimicking the central nucleotide moiety

in a trinucleoside diphosphate in order to shed light on how the nature and strength of

intramolecular stereoelectronic effects change as a result of Pt2+ binding in complete absence of

O

O

O

P

O

-

O O

R

O

P

O

-

O

NH

NN

N

O

NH2

H2C

CH3

CH2

H3C

Pt

NH3H3N

N7N7

O

O

O

P

O

-O O

R

O

P

O

-O

H2C

CH3

CH2

H3C

β

χ

α

ε

δ

γ

ζ

ζ−1

ε−1

α+1

β+1

138: EtpdGpEt (R = H)140: EtpGpEt (R = OH)

139: cis-(NH3)2Pt(EtpdGpEt)2141: cis-(NH3)2Pt(EtpGpEt)2

142: EtpdRpEt (R = H)143: EtpRpEt (R = OH)



135

intramolecular base-base stacking interactions. The study revealed that the N �S pseudorotational

equilibrium shifts towards more N-type conformers by Δ∆G° = 2.0 kJmol-1 and 2.6 kJmol-1,

respectively, thereby showing that free-energy of complexation is transmitted to drive the sugar

conformation. The increase in the population of N-type conformers was rationalized with the

strengthening of the anomeric effect in both 2'-deoxy and ribo nucleotides upon the formation of Pt-

N7 bond which promotes nO4' →σ∗

C1'-N9 orbital interactions due to the reduction of π-electron

density in imidazole part of guanine.

The additional stabilization of N-type conformers in Pt2+ complexes of ribonucleotide is due

to the tuning of gauche effect of [N9-C1'-C2'-O2'] fragment, which is absent in 2'-deoxyribo

counterparts. The platination of N7 favors N1 deprotonation in 139 and 141 by ΔpKa of 0.7 and 0.9

units in comparison to parent nucleotides 138 and 140, respectively. The N�S pseudorotational

equilibrium in 138 – 141 showed classical sigmoidal dependence as a function of pH with pKa

values at the inflection points.

Upon N1 deprotonation, the N �S pseudorotational equilibrium was shifted toward more S-

type conformations by Δ∆G°D-N = -0.3 kJmol-1 in 138, -0.9 kJmol-1 in 139, -0.4 kJmol-1 in 140 and

-2.5 kJmol-1 in 141 which showed that the thermodynamics of N1 deprotonation in guanin-9-yl

nucleosides is transmitted to drive N �S equilibrium towards S–type pseudorotamers. The

anomeric effect is weakened upon N1 deprotonation due to the increased π-electron density in the

imidazole part of the guanin-9-yl moieties and the population of S-type pseudorotamers increases.

The Pt2+-complex bound to N7 has been found to promote the redistribution of the π-electronic

density from anionic deprotonated N1 which results in the larger increase in the population of S-

type conformers in 139 and 141 in comparison to parent nucleotides 138 and 140. The formation of

Pt-N7 bond in bifunctional complexes 139 and 141 simultaneously causes a shift of syn �anti

equilibrium towards anti by 43 and 63 %, and the increase of the population of εt conformers by 20

and 32 % at 278K, respectively. Only minor conformational redistributions along β, γ, β+ and ε-

torsion angles have been observed which suggests weak conformational cooperativity between the

shift in the N �S pseudorotational equilibrium towards N-type conformers and the conformational

changes across phosphate torsion angles other than ε as a result of platination of guanin-9-yl. In

comparison to nucleotide phosphodiesters, apurinic 3',5'-bisethylphosphate sugars 142 and 143

showed no interaction with Pt2+ and therefore no conformational changes.

It is thus clear that any intermolecular interaction between nucleic acid and a ligand (other

than soft metal ion as described above) is expected to produce effects similar to partial electron-

deficiency as arising from the protonation of the aglycone (such as in the case of a lone-pair

donating aglycone involved in H-bonding or through-space charge transmission in stacked

complexes with intercalated drugs), or to partial electron-enrichment as arising from the

deprotonation of the aglycone (such as ionization of the aglycone by the basic sites of polypeptides

in general or a phosphate moiety in the vicinity or by a relatively hard metal ion).


136

8.8 RNA as molecular wire

The intrinsic dynamics and architectural flexibility of nucleic acids resulting into specific function

are the result of cooperative interplay of pentofuranose, nucleobase and phosphodiester moieties.

We have shown in Sections 4 - 7 that the interplay of various stereoelectronic gauche and anomeric

effects and associated steric effects energetically drives the sugar conformation, which in turn is

dictated by the electronic nature of the aglycone and other substituents on the sugar ring.

In our latest work44a,b, we have employed guanosine 3'-ethylphosphate, 5'-methylphophate

[MepGpEt (144a)] and adenosine 3',5'-bisphosphate [MepApEt (144b)], as a model mimicking the

central nucleotide moiety in a trinucleoside diphosphate in order to shed light on how the strength of

intramolecular stereoelectronic effects is modulated by the change of protonation �deprotonation

equilibrium of the aglycone in the complete absence of any intramolecular base-base stacking (For

our work on 2'-deoxy23 and ribonucleoside28 3'-ethylphosphates at the neutral pD, see Section 7).

The work uniquely shows a complete interdependency of conformational preference of sugar and

phosphate backbone in 144a and 144b as the protonation � deprotonation equilibrium of the

aglycone changes as a function of pD. Moreover, similar thermodynamic studies (only at two

extreme pDs i.e. at the neutral and the protonated states) of cytosine 3',5'-bisphosphate [MepCpEt

(144c)] have been performed (unpublished work) in order to compare with 144a and 144b to

demonstrate the tunibility of nucleobases in 144a – c.

In oligonucleotides the protonation of nucleobase directly affects their hydrogen-bonding

capabilities and therefore induces a change in overall three-dimensional structure (Section 2.9).

O

O

H

H

OH

O

H

H OH

NH

NN

N

O

NH2

N

NN

N

NH2

R

P

H3CH2CO

O−

OO

P

H3CH2CO

O−

O

O

R

O−

P

OCH3

O

O

O−

P

H3CO

O

N

N

NH2

O

(144b) R =

(145) R = H

(144c) R =

(144a) R =

South: 3'-exo-2'-endoNorth: 3'-endo-2'-exo

0° < P < 36°

Ψm

= 38.6 ± 3° 144° < P < 190°

Ψm

= 38.6 ± 3°

Schematic representation of the dynamic two-state North (N) � South (S) pseudorotational

equilibrium [P = phase angle and Ψm = puckering amplitude] of the constituent β-D-pentofuranose

moiety in gaunosine 3',5'-bisphosphate [MepGpEt (144a)], adenosine 3',5'-bisphosphate [MepApEt

(144b)] and cytidine 3',5'-bisphosphate [MepCpEt (144c)], as well as their apurinic counterpart

[Mep(ab)pEt (145)].



137

In these studies we have shown that changing the electronic character of aglycone through

protonation at basic site (N7 of guanin-9-yl, N1 of adenin-9-yl and N3 of cytosin-1-yl) not only

modulates the shift of N �S pseudorotational equilibrium of its constituate sugar by strengthening

the anomeric effect but is also transmitted to the sugar-phosphate backbone to influence the

conformational characteristics of backbone in 144a – c, which mimic the model of trinucleoside (3'-

5') diphosphate. This tunable transmission which has been observed, compared to the abasic

counterpart [Mep(ab)pEt] 145, to be much stronger at the 3'-phosphate backbone compared to 5'-

end justifies RNA as "molecular wire".

(i) pD-dependent shift of N � S pseudorotational equlibrium

In this endeavour, the mole fractions of N- and S-type conformers have been plotted as the

function of temperature in the van't Hoff type analyses to obtain ΔH° and ΔS° and subsequently to

calculate the total change in free energy at 298 K ( o K298

ΔG ) of the N � S pseudorotational

equlibrium at each of the seven pDs ranging from 6.7 to 1.0 (Table 15 and 16).

As the medium becomes more acidic the anomeric effect becomes more strengthened (See

Section 4.8) by enhancing the nO4' → σ∗C1'-N

orbital interaction owing to the reduced electron

density at N9 and due to which the N �S pseudorotational equilibrium for MepGpEt (144a) and

MepApEt (144b) is gradually shifted towards N-type pseudorotamers [79% S (for 144a) and 76%

S (for 144b) in the neutral state to 55% S (for 144a) and 67% S (for 144b) in the protonated state

respectively] as reflected from change of o K298

ΔG from -3.3 kJ mol

-1 (for 144a) and –2.8 kJ mol

-1

(for 144b) at pD 6.7 to -0.1 kJ/mol and -1.7 kJ mol-1

(for 144b) at pD 1.0 [Tables 15 and 16]. The

plots of pD-dependent o K298

ΔG values of the N �S equilibrium in MepGpEt (144a) as well as

MepApEt (144b) [Panel (B) in Fig 22] have the typical sigmoidal shapes, as found for β-D- 2',3'-

dideoxy, 2'-deoxy, ribo N- or C-nucleosides and some α-D-2',3'-dideoxy and 2’-deoxy-nucleosides

(Section 3.7, Fig 11). The value of the pD at the inflection point of the plots of pD-dependent o

K298

ΔG of N �S equilibrium in 144a and 144b were found to be 2.4 ± 0.1 and 3.8 ± 0.2

respectively, which are identical to the pKa of the constituent guanin-9-yl and adenine-9-yl

nucleobases, as determined independently from the plots of pD-dependent H8 chemical shift for

144a and that of both H8 and H2 chemical shifts for 144b [Panel (A) in Fig 22] [see the legend of

Fig 22 for details, Eq. 12 and Section 3.7.4]. As a control experiment, the difference in 3JHH values

between neutral and acidic pDs at 298K was found to be negligible in apurinic phosphodiester 145,

showing that in the absence of aglycone, the N � S equilibrium is unbiased and remains unchanged

at all pDs compared to 144a and 144b.


138

Table 15. Pseudorotational analyses and EPSILON calculations using temperature-dependent 3JHH

and 3JCPa at seven different pDs (1.0-6.7), and determination of pD-dependent thermodynamics of

the two-state N →← S and εt � ε- equilibrium in MepGpEt (144a).

Thermodynamics of N →← S

equilibrium

Thermodynamics of

εt ε- equilibrium

pD

o K298

ΔG =

-R*298*ln (xN / xS) %S

o K298

ΔG =

-R*298*ln (xεt/xε

-) %ε-

1.0b

-0.0 (0.1) -b 0.3 (0.2) -

b

1.6 -0.3 (0.2) 53 0.2 (0.2) 48

2.0 -0.7 (0.2) 58 0.0 (0.5) 50

2.4 -1.4 (0.2) 64 -1.2 (0.2) 62

2.7 -2.3 (0.2) 72 -1.7 (0.2) 66

3.0 -2.4 (0.2) 72 -1.8 (0.2) 67

6.7 -3.2 (0.2) 78 -2.0 (0.2) 69

a For the experimental procedure used to calculate the values, see Section 3 and ref. 44a. b

No calculations performed due to unavailability of 3JHH owing to spectral overlap. ΔG° (at

298 K) value have been extrapolated.

Table 16. Pseudorotational analyses and EPSILON calculations using temperature-dependent 3JHH

and 3JCPa at nine different pDs (1.0-7.9), and determination of pD-dependent thermodynamics of

the two-state N →← S and εt � ε- equilibrium in MepApEt (144b).

Thermodynamics of N →← S

equilibrium

Thermodynamics of

εt � ε- equilibrium

pD

o K298

ΔG =

-R*298*ln (xN / xS) %S

∆Go (at 298 K) =

-R*298*ln (xεt/xε

-) %ε-

1.0 -1.7 (0.2) 67 -1.5 (0.2) 65

2.0 -2.0 (0.2) 69 -1.6 (0.2) 66

2.5 -2.0 (0.2) 69 -1.6 (0.2) 66

3.0 -2.1 (0.2) 70 -1.6 (0.2) 66

3.5 -2.2 (0.2) 71 -1.7 (0.2) 66

4.0 -2.4 (0.2) 73 -1.8 (0.2) 67

4.5 -2.7 (0.2) 75 -1.9 (0.2) 68

4.8b

-2.8 (0.2) -b -2.0 (0.2) -

b

7.9 -2.8 (0.2) 76 -1.9 (0.2) 68

a For the experimental procedure used to calculate the values, see Section 3 and ref. 44b.

b

No calculations performed due to unavailability of 3JHH owing to spectral overlap. ΔG° (at

298 K) value have been extrapolated.



139

(ii) pD-dependent shift of the εt � ε- equlibrium

The temperature dependent 3JH3'P3',

3JC4'P3', 3JC2'P3' have been used to calculate the bias of

conformational equlibrium across C3'-O3' bond (ε torsion). Since the ε+ conformer is energetically

forbidden and is not found in crystal data we have interpreted the experimental coupling constants

in terms of the two-state εt � ε- equilibrium (Section 7.1). Logarithm of ratio of the resulting mole

fractions of ε- and εt were plotted as a function of temperature in a van't Hoff type analyses to obtain

ΔH° and ΔS° and subsequently to calculate the total change in free energy at 298 K ( o K298

ΔG ) of

εt � ε- equlibrium at each of the seven pDs ranging from 6.7 to 1.0 (Table 15). 69% (for 144a) and

68% (for 144b) in neutral state to 48% (for 144a) and 65% (for 144b) in protonated state

respectively. The corresponding shift of o K298

ΔG is from -2.1 kJ mol

-1 (for 144a) and -1.9 kJ mol

-1

(for 144b) in neutral state to +0.3 kJ mol-1

(for 144a) and -1.5 kJ mol-1

(for 144b) in protonated

state [Tables 15 and 16]. Interestingly, the plots of pD-dependent o K298

ΔG values of the εt �ε-

equlibrium in MepGpEt (144a) as well as in MepApEt (144b) also give a sigmoidal curve [Panel

(C) in Fig. 22], as found for the plot of the corresponding o K298

ΔG values of the N � S equilibrium

of the constituent pentofuranose sugar. The values of pD at the inflection point (i.e. pD = 2.3 for

144a and pD = 3.7 for 144b) of the graphs shown Panel (C) in Fig 22, and is nearly identical (i.e.

within the accuracy of the measurements) to the pKa values (2.4 for 144a and 3.9 for 144b

respectively) of the guanin-9-yl nucleobase in MepGpEt. Thus, it can be concluded that the pD-

dependent reorientation of the 3'-ethylphosphate moiety across the C3'-O3' bond (reflected

in o K298

ΔG values of the εt � ε- equlibrium) is directly dictated by the pKa of the constituent C1’-

aglycone in MepGpEt (144a) and MepApEt (144b).

(A)

pD

0 1 2 3 4 5 6 7 8

δH

8.0

8.2

8.4

8.6

8.8

9.0

9.2

(B)

pD

0 1 2 3 4 5 6 7 8

ΔG

o(N/S) at 298 K

-3.6

-3.0

-2.4

-1.8

-1.2

-0.6

0.0

0.6

(C)

pD

0 1 2 3 4 5 6 7 8

ΔG

o(ε

t/ε

−) at 298 K

-2.4

-1.8

-1.2

-0.6

0.0

0.6

δH8G of MepGpEt

δH8A of MepApEt

δH2A of MepApEt

MepGpEt

MepApEt

MepGpEt

MepApEt

Figure 22. Panel (A) shows the plots of H8 chemical shift of the constituent guanine-9-yl nucleobase (δH8G in ppm) in

144a as well as H8 and H2 chemical shift of the constituent Adenin-9-yl nucleobase (δH8A and δH2A respectively in

ppm) in 144b, as a function of pD at 298K. Panels (B) and (C) show the plots of the experimental ΔG° (in kJ mol-1

) of

the N � S equilibrium of the constituent pentofuranose moiety and that of ε

t

� ε− equilibrium of the 3'-ethylphosphate

group in 144a and 144b respectively, as a function of pD at 298K. The following pKa values have been obtained: 2.4 ±

0.1 (for δH8 in 144a), 4.0 ± 0.1 and 3.9 ± 0.02 (for δH8 and δH2 respectively in 144b) in Panel (A); 2.4 ± 0.1 (for

144a) and 3.8 ± 0.2 (for 144b) in Panel (B); 2.3 ± 0.4 (for 144a) and 3.7 ± 0.2 (for 144b) in Panel (C).

These results constitute the first experimental evidence for transmission of electronic

information to steer the conformation of the 3'-phosphate as a result of change of electronic

properties of the nucleobase (protonation/deprotonation) via the tuning of the sugar conformation in

a nucleoside 3',5'-bisphosphates (such as in 144a and 144b) (see the correlation plots below under

subsection (iii)). In the case of our reference compound, i.e. apurinic phosphodiester [Mep(ab)pEt


140

(145)] no change in 3JHP and 3JCP coupling constant values has been observed compared to that of

purinic phophodiester 144a and 144b over the whole pD range which allows us to conclude that the

conformation across C3'-O3' remains unchanged over the whole pD range in abasic compound

(145) since there is no nucleobase to be protonated in 145.

(iii) The cooperative shift of the (N,εt) �(S,ε-) equilibrium as the result of protonation of guanin-9-

yl in MepGpEt (144a) and MepApEt (144b) are evidenced by [∆G°(N/S) vs ∆G°(ε t/ε-)] or [∆G°(N/S)

or (ε t/ε-) vs δH8] correlation plots

The correlation plots of pD-dependent o K298

ΔG values of the N�S equilibrium of the

pentofuranose sugar as a function of pD-dependent o K298

ΔG values of the εt � ε- equilibrium of the

3'-phosphate moieties in 144a and 144b give a straight line with a high Pearson correlation

coefficient [R = 0.98 for both 144a and 144b, Panel (A) in Fig 23]. This means that as the

constituent guanin-9-yl in 144a and adenine-9-yl in 144b are gradually protonated in the acidic

medium, the modulation of the strength of the anomeric effect shifts the N� S equilibrium toward

N, which in turn dynamically shifts the εt �ε- equilibrium toward more εt in comparison with the

neutral pD.

(A)

ΔGo

(εt

/ε−

) at 298 K

-2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5

ΔG

o(N/S) at 298 K

-4

-3

-2

-1

0

1

2

(C)

δH

8.0 8.2 8.4 8.6 8.8 9.0

ΔG

o(ε

t/ε

−) at 298 K

-3.6

-2.4

-1.2

0.0

1.2

(B)

δH

8.0 8.2 8.4 8.6 8.8 9.0

ΔG

o(N/S) at 298 K

-3.6

-2.4

-1.2

0.0

1.2 δH8G of MepGpEt

δH8A of MepApEt

δH2A of MepApEt

MepGpEt

MepApEt

δH8G of MepGpEt

δH8A of MepApEt

δH2A of MepApEt

Figure 23. Panel (A) shows the plot of o

K298

ΔG (in kJ mol-1

) of the N � S pseudorotational equilibrium [ΔG

o(N/S) at 298

K] as a function of the o

K298

ΔG (in kJ mol

-1) of ε

t

� ε− equilibrium [ΔG

o(εt/ε−) at 298 K] of the 3'-ethylphosphate group at

298 K for 144a and 144b; R = 0.98 (for 144a) and 0.98 (for 144b). Panels (B) shows the plot of o

K298

ΔG (in kJ mol

-1)

of the N � S pseudorotational equilibrium as a function of the chemical shift of the constituent guanine-9-yl nucleobase

(δH8G in ppm) in 144a as well as H8 and H2 chemical shift of the constituent adenin-9-yl nucleobase (δH8A and δH2A

respectively in ppm) in 144b at 298 K; R = 1.00 (for 144a), 0.94 and 0.98 (for δH8A and δH2A respectively in 144b).

Panels (C) shows the plot of o

K298

ΔG (in kJ mol

-1) of the ε

t

� ε− equilibrium

as a function of the chemical shift of the

constituent guanine-9-yl nucleobase (δH8G in ppm) in 1 as well as H8 and H2 chemical shift of the constituent adenin-

9-yl nucleobase (δH8A and δH2A respectively in ppm) in 2 at 298 K; R = 0.99 (for 144a), 0.94 and 0.97 (for δH8A and

δH2A respectively in 144b).

An additional evidence for the transmission of the free-energy of the

protonation�deprotonation equilibrium of the nucleobase to steer phosphate conformation through

the change of the sugar conformation moiety is that the plots of the pD-dependent o K298

ΔG values

of the N�S equilibrium [Panel (B) in Fig 23] or of the pD-dependent o K298

ΔG values of the εt �ε

-

equilibrium [Panel (C) in Fig 23] in 144a as well as in 144b as a function of the pD-dependent

chemical shift of aromatic protons (H8 in 144a as well as both H8 and H2 in 144b) give straight

lines with high correlation coefficients (R ≥ 0.94).



141

Figure 24. The demonstration of single stranded RNA as molecular wires using the model MepGpEt (144a).

Transmission of the free energy of the protonation � deprotonation equilibrium of the guanin-9-yl group in 144a drives

the sugar-phosphate conformations through three consecutive stereoelectronic tunings (Newman projections: a - c; AE =

anomeric effect, GE = gauche effect). Appropriate orbital overlap and the energy difference between the donor and

acceptor orbitals (see Fig 25), dictated by various sugar atoms and substituents, drive the sugar-phosphate conformation

through the interplay of gauche and anomeric effects. All orbitals are shown by straight arrows. Smaller curved arrows

show the preferred torsional orientation, whereas the larger curved arrows indicate the mixing of donor and acceptor

orbitals.

The energy stabilization through either GE or AE, manifested by the interaction between the participating

donor and the acceptor orbitals, is directly proprtional to the square of the overlap between the donor and the acceptor

orbitals (i.e. S2) as well as inversely proportional to their energy differences (stabilization by AE or GE = S

2/ΔE). When

the aglycone is protonated at N7 (Panel A), the N-type pseudorotamers are preferred owing to the strengthening of the

AE(O4'-C1'-N9), as shown by the overlap between the 1nsp2 orbital of one of the O4' lone pairs [1nsp2 (p-type, O4')]

and the antibonding orbital of the C1'-N9 bond [σ*C1’-N9]. This is illustrated through a Newman projection (Panel 3a),

where O4' lone pair orbitals [preferentially the higher energy 1nsp2 (p-type), not the lower energy 2nsp2 (S-type)] and

the nO4' (P-type) → σ∗C1'-N9 overlap stabilizes the N-type over the S-type sugars. For an N-type sugar, nO4' →

σ∗C1'-N9 interactions are possible owing to a near antiperiplanar orientation of 1nsp2 (p-type) with respect to the

σ*C1'-N9 bond [Φ(1nsp2(p-type)-O4'-C1'-N9) ≈ 159°], which takes place when the aglycone is pseudoaxial. In

contrast, they are much weaker when the aglycone is pseudoequatorial in the S-type sugars owing to relatively large

O

HOH2C

O

N

N9

H3'

P

O

-O O

H3CH2C

O

HOH2C

O

N

N9

H3'

P

O

-O

O

CH2CH3

+H+

-H+

1nsp2 (p-type, O4')

σC3'H3'

σ*C1'-N9

σ*C4'O4'

σ*O3'P3'

1nsp2 (p-type, O3')

εt

1nsp2 (p-type, O4')

σC3'H3'

σ*C1'-N9

σ*C4'O4'

σ*O3'P3'

1nsp2 (p-type, O3')

ε-ζ

-

α-

ζ-

α-

H+

H4'

C5'O4'

C2'

C2'

O4'

C5'

H3'

C1'

C1'

C2'

N9 (N7H+)

H1'

O4'

C4'

C2'

N9

H1'

O4'

C4'

+H+

-H+

+H+

-H+

OPO3CH2CH3 OPO3CH2CH3

+H+

-H+

Φ = -101

o

σ* C4'O4'

σ C3'H3'

Φ = -45

o

σ* C4'O4'

σ C3'H3'

Φ = 159o

2nsp2 (s-type)

1nsp2 (p-type)

Φ = 128ο

2nsp2 (s-type)

1nsp2 (p-type)

(a) AE (O4'-C1'-N9)

(b) GE (O4'-C4'-C3'-O3')

(c) AE (O3'-P3'-OEt)

Φ = -156

o

2nsp2 (s-type)

1nsp2 (p-type)

OCH2CH3

O3'C3'

O-

O

H3'

H4'

Φ = -154

o

2nsp2 (s-type)

1nsp2 (p-type)

OCH2CH3

O3'C3'

O-

O

Change of Electronic

Character of the Aglycone

One-way Transmission

OFFON

ON OFF

εt : ON ε

- : OFF

(A) (B)


142

deviation from antiplanarity, as indicated by smaller Φ values [Φ(1nsp2(p-type)-O4'-C1'-N9) ≈ 128°]. On the other

hand, in the neutral state (Panel 3B) the S-type pseudorotamers are relatively preferred owing to the absence of an O4'-

C1'-N9 anomeric effect as well as due to the stabilization through the O3'-C3'-C4'-O4' gauche effect (i.e. the overlap of

σC3'-H3' with σ*C4'-O4'. In the Newman projection (Panel 3b), the most efficient overlap of the best donor (σC3'-H3') and

best acceptor orbitals (σ*C4'-O4') is indicated by a preference for the gauche orientation [smaller value for Φ(σC3'-H3'-

C3'-C4'-σ*C4'-O4') ≈ -45°] in the S-type sugar over the trans conformer [higher value of Φ(σC3'-H3'-C3'-C4'-σ*C4'-O4') ≈ -

101°] in the N-type sugar. Thus, in the S-type conformation, a reduction of charge density at the O3' lone pair [1nsp2(p-

type)] takes place owing to its maximal interaction with σ*C4'-O4' (presumably a combination of through-bond and

through-space effects are involved in this process), thereby weakening the O3'-P3'-OCH2CH3 anomeric effect in the S/ε-

state. Owing to the nO4' → σ∗C1'-N9 overlap (Panel 3A), the O4' lone pair is relatively more delocalized in the N-type

conformation, which places the O3'-C3'-C4'-O4' fragment in the trans orientation, preventing the gauche effect from

being fully operational (the reverse is true when the anomeric effect is weakened in the S-type conformation, Panel 3B).

Since the 3'-GE is not fully operational in the N-type sugar conformation, the charge density at the O3' lone pair [1nsp2

(p-type, O3')] is fully available to act as a donor and interact through the anomeric effect with the antibonding σ*P3'-

O(ester) orbital [AE(O3'-P3'-OCH2CH3], when C3'-O3' is in εt, O3'-P3' in ζ

−

, and P3'-O5' in α−

conformations. The

Newman projection (Panel 3c) shows that the overlap between the O3' lone pair orbitals and the σ*P3'-O(ester) orbital [nO3'

→ σ* P3'- O(ester) orbital mixing] stabilizes εt over ε

-. This is not only due to an antiperiplanar orientation of 1nsp2(p-

type) with respect to the P3'-O(ester) bond, as Φ(1nsp2(p-type)-O3'-P3'-OCH2CH3) is nearly the same for the two cases,

but largely owing to the greater electron density availible at nO3', arising from the absence of 3'-GE in the N-type sugar.

Therefore, these works constitute the first quantitative evidence for the transmission of the

electronic character of the nucleobase to drive the 3'-phosphate conformation in a ribonucleoside

3',5'-bis-phosphate as the direct consequence of the concomittant modulation of the bias of the

pseudorotational equilibrium of the constituent sugar moiety.

(iv) The mechanism of the transmission of anomeric effect of the nucleobase to 3’-gauche effect and

further on to steer the 3’O-P-O(ester) anomeric effect

In our earlier work on 2'-deoxy23 and ribonucleoside28 3'-ethylphosphates at the neutral pH

(Section 7), we found a non-equivalent methylene protons of the 3'-ethyphosphate moiety, which

turned out to be a temperature-dependent feature. At higher temperature, this non-equivalency

disappeared and the methylene protons showed NMR time average chemical shifts with similar

multiplicity as those of 2'-deoxy counterparts (owing to the absence of 2'-OH promoted hydrogen

bonding). This observation demonstated that the nonequivalency of the methylene protons in

ribonucleotides was owing to the 2'-OH promoted hydrogen bonding with the vicinal O3'. We

attributed the concerted sugar-phosphate backbone conformational change solely to the 2'-OH

effect, the strength of which was modulated by temperature (the “on-off switch”). In our work with

pD-dependent conformational studies with MepGpEt (144a) and MepApEt (144b), we found

similar temperature-dependent pattern and intensity of multiplicities of methylene protons of the 3'-

ethylphosphate moiety at all pD as found for the neutral state of guanosine 3'-ethylphosphate (80)

and adenosine 3'-ethylphosphate (79) respectively, thereby suggesting that the 2'-OH hydrogen-

bonding remains the same in 144a and 144b over the whole pD range (pD 1.0 to 6.7) at room

temperature (298 K). This means that all changes of free-energies observed at 298 K (Table 15) for

144a and 144b as a function of pD is attributed to the free-energy changes of the



143

protonation�deprotonation equilibrium of the aglycones (i.e. guanin-9-yl and adenin-9-yl) to drive

the sugar-phosphate backbone in a concerted manner. As the pD-tunable change of the electronic

character of the nucleobase tunes the strength of the anomeric effect, an increased preference of the

N-type sugar conformation is imposed because of enhanced nO4'→σ∗

C1'-N orbital interaction,

which, in turn, affects the strength of the [O3'-C3'-C4'-O4'] gauche effect by retuning the energy

levels of the donor and aceeptor orbitals in the σC3'-H → σ∗C4'-O4' interaction. The extent of σC3'-H

→ σ∗C4'-O4' participation influences the electron density at O3', which in turn modulates the O3'-

P3'-O anomeric effect (a concerted transmission). Readers are directed to Fig 24 for a detailed

discussion of the molecular orbital diagram based interpretation of the concerted pD-dependent

change of sugar-phosphate backbone conformation (taking 144a as a model). We also suggest that

the readers take a close look at Fig 25 for the corresponding energy diagram, which show that the

energy levels of the orbitals involved in the anomeric and gauche effects are based on their relative

acceptor/donor abilities in a purely qualitative manner. In Fig 25, various donor and acceptors

orbitals are connected by the dotted lines to show their interdependent modulation and steering of

different orbital mixing. This translates itself in terms of relative strength of gauche and anomeric

effects and the preferred confomational states which make the RNA to act as a molecular wire. The

electron flow is unidirectional from the nucleobase to the phosphate, which is relected by the fact

that the hybrid orbital produced by the O3'-P3'-O(ester) anomeric effect is at a lower energy level

than the corresponding hybrid orbital resulting from O4'-C1'-N anomeric effect.

It is noteworthy that the proof of the one-way (i.e. from nucleobase to phosphate through

sugar) transmission of electronic information in MepGpEt (144a) is evident by the fact that the pKa

value of the guanin-9-yl aglycone is 2.4, which is the same as the 2',3'-dideoxuguanosine (pKa 2.5),

2'-deoxyguanosine (pKa 2.3) and guanosine (pKa 2.1) within the error limits of our experimentals.

In fact, the same pKa value has been found for each nucleobase in corresponding nucleosides

and nucleotides in either 2',3'-dideoxy, 2'-deoxy and ribo configuration as well as in MepGpEt

(144a) and MepApEt (144b), as discussed in Section 4.8, show the minimal influence of the

sbstituents at C2'/3'/5' on the electronic character of their constituent nucleobase at C1'. The final

proof of the operation of O3'-P3'-O anomeric effect could only be experimentally obtained if we

could only measure the ζ and α torsions across the 3'-phosphate backbone and the preferential O3'-

P3'-O bond angle.

(v) pD-dependent conformational change across β torsion is negligible

The conformational equilibrium across the torsion angle β [C4'-C5'-O5'-P5'] for 144a has been

calculated by using temperature dependent 3JC4'P5' as well as the sum of 3JH5'P5' and 3JH5''P5'

(Section 7.1). The analysis of both sets of data provided comparable results with discrepancies

below 3%, which is within the accuracy of Karplus equation used, showing that the population of βt

.

Chatt

opadhya

ya e

t al,

"S

tere

oel

ectr

onic

Eff

ects

in N

ucl

eosi

des

& N

ucl

eoti

des

and t

hei

r S

truct

ura

l Im

pli

cati

ons"

,

Dep

t of

Bio

org

anic

Chem

istr

y, B

ox 5

81, U

ppsa

la U

niv

ersi

ty, S

-75123 U

ppsa

la, S

wed

en, V

er 1

60205 j

yoti

@boc.

uu.s

e

144

AN

AP

CP

CN

PN

BP

BN

DP

DN

N1

P1

Neu

tral

Sta

te

ΔE

[O3

'-P

3'-

O(e

ste

r) A

E]

ΔE

[O4

'-C

1'-

N9

AE

]

σ*C

1'-

N9

σ*C

4'O

4'

1n

sp2

(p

-typ

e, O

3')

σ*P

3'-

O(e

ste

r)σ

C3'H

3'

ΔE

(GE

)

ΔΔ

E(G

E)

ΔΔ

E(A

E)

1n

sp2

(p

-typ

e, O

4')

ΔE

Proto

nate

d S

tate

F

igu

re 2

5.

The

rela

tive

do

no

r-ac

cep

tor

abil

itie

s o

f var

ious

orb

ital

s (i.e

. th

e re

lati

ve

emp

iric

al e

ner

gie

s) a

re m

od

ula

ted

by t

he

nat

ure

of

each

sub

stit

uen

t at

the

sugar

and

o

f th

e p

ho

sphat

e b

ackb

one

as w

ell

as o

n th

e ag

lyco

ne

in fr

ee,

ionic

an

d co

mp

lex st

ate.

S

ince

th

e el

ectr

onic

st

ate

of

the

agly

cone

mo

dula

tes

the

sugar

confo

rmat

ion w

hic

h i

n t

urn

mo

dula

tes

the

pho

sphat

e to

rsio

n,

we

hav

e p

lace

d t

he

1

nsp

2 (p

-type,

O3’) o

rbit

al a

t a

rela

tivel

y l

ow

er e

ner

gy l

evel

than

1

nsp

2 (p

-type,

O4’).

As

the

O4

’-C

1’-

N9

ano

mer

ic e

ffec

t st

arts

op

erat

ing,

the

stre

ngth

of

the

[O3

’-C

3’-

C4

’-O

4’]

gau

che

effe

ct (σ

C3’-

H3’ →

σ*

C4’-

O4’)

co

unte

ract

s th

e an

om

eric

eff

ect

ow

ing t

o t

he

rela

tivel

y l

ow

er e

ner

gy o

f σ

*C

4’-

O4’

[AN b

eco

mes

AP s

tate

when

neu

tral

nucl

eob

ase

(N)

bec

om

es p

roto

nat

ed (

P)]

. A

s th

e nucl

eob

ase

in t

he

(N)

stat

e ta

kes

up

the

(P)

stat

e in

eit

her

Mep

Gp

Et

(14

4a

) o

r in

Mep

Ap

Et

(14

4b

), σ

*C

1’-

N9 b

eco

mes

a b

ette

r ac

cep

tor

and

the

O4

’-C

1’-

N9

ano

mer

ic e

ffec

t is

str

ength

ened

, an

d t

hat

mak

es [

O3

’-

C3

’-C

4’-

O4

’] g

auch

e ef

fect

mo

re e

ffec

tive

[σC

3’-

H3’

(BN b

eco

mes

BP s

tate

) →

σ*

C4’-

O4’, i.e

. m

ore

eff

ecti

ve

orb

ital

mix

ing o

f BP w

ith A

P,

see ΔΔ

H°

10 i

n F

ig 1

3 a

nd

Tab

le 6

). H

ow

ever

, th

e an

om

eric

eff

ect

is s

tro

nger

than

the

gau

che

effe

ct,

ther

efo

re w

e se

over

all

stab

iliz

atio

n o

f m

ore

N-t

yp

e su

gar

s (ΔΔ

E(G

E) <

ΔΔ

E(A

E), T

able

15

)

in t

he

P s

tate

co

mp

ared

to

in N

sta

te.

The

O3

’-P

3’-

O(e

ster

) an

om

eric

eff

ect

is w

eaker

in t

he

S-t

yp

e p

seud

oro

tam

ers

(at

aro

und

neu

tral

pH

, i.e.

in C

N s

tate

) th

an i

n t

he

N-t

yp

e co

unte

rpar

ts (

at a

round

aci

dic

pH

, i.e.

in C

P s

tate

) b

ecau

se t

he σ

C3’-

H3’

orb

ital

over

lap

s w

ith t

he σ

*C

4’-

O4’

(i.e

. [O

3’-

C3

’-C

4’-

O4

’ gau

che

effe

ct]

is s

tro

nger

in

the

form

er s

tate

), r

educi

ng t

he

elec

tro

n d

ensi

ty a

t O

3’

(B).

This

mea

ns

that

the

1

nsp

2 (p

-type,

O3’) l

onep

air

is r

elat

ivel

y l

ess

avai

lab

le i

n S

-typ

e co

nfo

rmat

ion (

at n

eutr

al

pH

) to

inte

ract

wit

h t

he σ

*P

3’-

O(e

ster

) th

an i

n t

he

N-t

yp

e co

nfo

rmat

ion (

at a

cid

ic p

H),

whic

h s

ho

ws

that

the

O3

’-P

3’-

O(e

ster

) an

om

eric

eff

ect

is w

eaker

in S

-typ

e

confo

rmat

ion.



145

rotamer is pD independent. Hence, the transmission of the information on the change of the

electronic character from nucleobase to 5'-phosphate ends at the γ torsion Similarly, the change of βt

rotamer is pD independent and that of γ torsion is very much within the limit of error of the analyses

in 144b (even less than 144a).

(vi) Conformational equilibrium across C1'-N9 (χ) bond

The populations of the syn and anti conformations involved in the two-state syn �anti

conformational equilibrium across the glycosyl bond in 144a have been calculated from nOe

enhancements experimentally measured at H1’ upon saturation of H8 using the semi-quantitative

equation derived by Rosemeyer and Seela517. The change in population of anti conformer from

neutral (6.7) to acidic (1.0) pDs is observed to be very small viz. 49% to 63% respectively.

(vii) 1H and 31P Chemical shift correlations

The correlation plot between the change in population of S-type sugar conformation owing to the

stronger anomeric effect in 144a upon protonation at N7 of guanin-9-yl and the resultant pD-

dependent H8 chemical shift has been shown in Panel (B) of Fig 26. Inspection of Panels (A) - (D)

in Fig 26 also shows that the percentage conformational change in either 3'-sugar-phosphate

backbone conformation (ε- conformer) or the 5'-end (population of γ+ conformer) in 144a has been

nicely corroborated with the gradual downfield chemical shift of H8 of heterocyclic base due to

protonation. However, the change of γ torsion as a function of pD is very much within the limit of

error of the analyses in 144b (even less than 144a).

Figure 26. (A) The plot of percentage population of γ+ conformer (calculated w.r.t. original assignment, taking H5'

downfield and H5" upfield) at 298K as a function of H8 chemical shift of guanin-9-yl in 144a at different pDs ranging

from 1.6 to 6.7 showing the straight line (R = 0.91) with a slope of 7.37 (σ = 0.99) and an intercept of 11.88 (σ = 8.37).

(B) The plot of %S at 298K as a

function of H8 chemical shift of

guanin-9-yl in 144a at six pD

values ranging from 1.6 to 6.7

showing the straight line (R =

1.00) with a slope of -27.54

(σ = 0.39) and an intercept of

299.00 (σ = 3.31). (C) The plot of

percentage population of ε-

conformer at 298K in 144a as a

function of H8 chemical shift of

its guanin-9-yl at six pDs ranging

from 1.6 to 6.7 showing the

straight line (R = 0.98) with a

slope of -25.18 (σ = 1.43) and an

intercept of 273.27 (σ = 12.05). (D) The plot of percentage

population of βt conformer in 144a at 298K as a function of H8 chemical shift of guanin-9-yl at six pDs ranging from

1.6 to 6.7 showing the straight line (R = 0.96) with a slope of -2.93 (σ = 0.25) and an intercept of 104.25 (σ = 2.07).

Thus the tranmission of the electronic character of the nucleobase to drive the phosphate backbone

conformation via tuning of the pentofuranose conformation has been quantitatively established.

Moreover, the plots in Panels (B) and (C) of Fig 23 and in Panels (A) and (B) of Fig 27 indicate a

direct correlation between 31P3' chemical shift respectively with both ∆G°298Κ (N � S) and

(A)

δH8 (ppm)

7.8 8.0 8.2 8.4 8.6 8.8 9.0

%γ

+

68

70

72

74

76

78

80

(B)

δH8 (ppm )

7.8 8.0 8.2 8.4 8.6 8.8 9.0

%S

50

55

60

65

70

75

80

85

90

(C)

δH8 (ppm)

7.8 8.0 8.2 8.4 8.6 8.8 9.0

%ε

−

45

50

55

60

65

70

75

80

(D)

δH8 (ppm)

7.8 8.0 8.2 8.4 8.6 8.8 9.0

%β

t

77

78

79

80

81

82


146

∆G°298Κ (εt � ε-) over the entire pD range for 144a and 144b which can be considered as another

proof of the transmission in terms of conformational energetics from base to phosphate backbone

through nucleotidyl wire in our trinucleotide model system.

(A)

ΔGo

(N/S) at 298 K

-3.6 -3.0 -2.4 -1.8 -1.2 -0.6 0.0

δ3

1

P

-1.65

-1.60

-1.55

-1.50

-1.45

-1.40

(B)

ΔGo

(εt

/ε−

) at 298 K

-2.4 -1.8 -1.2 -0.6 0.0 0.6

δ3

1

P

-1.65

-1.60

-1.55

-1.50

-1.45

-1.40MepGpEt

MepApEt

MepGpEt

MepApEt

Figure 27. Panel (A) shows the plot of ΔG° (in kJ mol-1

) of the N � S pseudorotational equilibrium as a function of the

phosphorous chemical shift (δ31

P in ppm) for MepGpEt (144a) and MepApEt (144b) at 298 K; R = 0.97 (for 144a) and

0.96 (for 144b). Panel (B) shows the plot of ΔG° (in kJ mol-1

) of the εt

� ε− equilibrium

as a function of the

phosphorous chemical shift (δ31

P in ppm) for 144a and 144b at 298 K; R = 0.95 (for 144a) and 0.91 (for 144b).

(viii) Tunibility of aglycones and tranmission of the electronic character to drive the phosphate

backbone conformation

This aglycone dependent conformational transmission of sugar-phosphate backbone via

pentofuranose depends upon the tunibility of aglycone vis-à-vis conformational modulation of sugar

geometry. Our control studies with MepCpEt (144c) at neutral and protonated state showed that the

relative conformational tunibility is in order MepGpEt (144a) > MepApEt (144b) > MepCpEt

(144c) [Fig 28].

(A)

0.944

0.215 0.198

0

0.3

0.6

0.9

1.2

ΔΔδ(P-N)

ppm

EtpGpMe

EtpApMe

EtpCpMe

(B)3.2

1.1 0.9

0

1

2

3

4

ΔΔG(P-N) (N/S)

kJ mol-

1

(C)

0.0

0.4

2.3

0

0.5

1

1.5

2

2.5

ΔΔG(P-N) (εt/ε

−

)

kJ mol-

1

Figure 28. The relative conformational tunibility of MepGpEt (144a), MepApEt (144b) and MepCpEt (144c) between

neutral (N) and protonated (P) states at 298 K. Panel (A) shows the relative change of chemical shift of aromatic protons

[ΔΔδ (P-N)] in 144a – c; Panel (B) and (C) show the relative change of free energy of the N � S pseudorotational

equilibrium [ΔΔG°(P-N) (N/S), in kJ mol-1

] and that of εt

� ε− equilibrium [ΔΔG°(P-N) (ε

t/ε−

), in kJ mol-1

] in 144a – c. The

relative order of tunibility is 144a > 144b > 144c

8.9 The importance of O4' in the self-organization of oligo-DNA

In order to understand the implication of the endocylic oxygen (i.e. O4') in the origin of the

stereoelectronic gauche and anomeric effects, we have determined584 the solution conformation of a

12mer oligo-DNA in which a specific sugar has been replaced by a carbocyclic analogue,



147

k1

k-1

Watson-CrickbasepairedDuplex (Ia)

HoogsteenbasepairedDuplex (Ib)

Aristeromycin (i.e. A6 residue). This

modified oligo-DNA, 5'-

d(C1G2C3G4A5A6T7T8C9G10C11G12)-3'

(Duplex I) with the sugar moiety of A6

replaced by aristeromycin has been compared with the natural unmodified counterpart, the

Dickerson-Drew dodecamer (5'-d(CGCGAATTCGCG)-3').

Table 16. Thermodynamics of duplex to single strand melting by UV spectroscopy for the native

and modified dodecamersa from the plot of 1/Tm vs ln(concentration) at 8, 12, 16, 20 and 24 mM

concentration (at least two measurements have been made at each concentration).

Native dodecamer

(kcalmol-1)

Modified dodecamer

(kcalmol-1) ∆H° -82 (± 1) -79 (±1)

-T∆S° -64 (±1) -63 (±1)

∆G°298 -18 (±1) -16 (±1)

a See ref. 585 for an independent measurement.

Table 17. Thermodynamicsa of the exchange process between Hoogsteen and Watson-Crick

basepaired duplexes.

kJmol-1 (Ia)(Ib) 1k⎯⎯ →⎯−

(Ib)(Ia) 1k⎯⎯→⎯

Ea 155 ±13 167 ±14

∆Gθ 117 124

∆Hθ‡

153 ± 13 164 ± 14

-T∆Sθ‡

-36 -40

a The rate constants (k1 and k-1) for the two-state exchange process between Watson-Crick basepaired duplex (Ia) and

Hoogsteen basepaired duplex (Ib) were calculated from the initial slope of build-up rates of NOESY and ROESY

exchange crosspeaks at different mixing times at 7 different temperatures. The rate constants (k1 and k-1) were then used

in an Arrhenius plot to calculate the energy of activation (Ea) which in turn gives the enthalpy (∆Hθ‡

) and enthropy

(∆Sθ‡

) contributions to the free energy of activation (∆Gθ) (see ref. 586 and 587).

The duplex (I) was found to exist in dynamic equilibrium between two different

interconverting conformations, (Ia) and (Ib), with equilibrium constant K = k1 / k-1 = 0.56 ± 0.08

in the temperature range of 287 - 308K. Thus the free-energy of stabilization (ΔΔG°298) of Watson-

Crick basepaired duplex (Ia) realtive to Hoggsteen basepaired duplex (Ib) is 1.4 kJmol-1, which is

calculated from -RT* lnK. In the (Ia)-form, the duplex adopts a canonical B-type conformation

where all basepairs are of the Watson-Crick type, whereas in the (Ib)-type structure, the basepairs

formed between A6 and T7 are of the Hoogsteen type (the N1 of the adenin-9-yl base in

aristeromycin is looking at the major groove), and the rest of the molecule however adopts a B-type

conformation. The thermodynamics and the kinetics of the Watson-Crick basepaired (Ia) duplex �

Hoogsteen basepaired (Ib) duplex equilibrium have also been investigated, and are shown in Tables

16 and 17, respectively. It can be seen that the difference between the thermodynamics of the self-

assembly of the modified DNA duplex and the native counterpart is negligible, well within the error


148

limits. This means that there is virtually no energy penalty for replacement of O4' by CH2, which is

evident from the fact that the equilibrium populations of Watson-Crick and Hoogsteen duplexes are

nearly the same. This suggests that there is no preference for furanose-based duplex over the

cyclopentane-based duplex in energetic terms. However, the largest structural implication in the

cyclopentane-based duplex (Ib) is the change of the local structure around the Hoogsteen

basepaired junction compared to the Watson-Crick basepaired counterpart in that the N1 of the

adenin-9-yl moieties in aristeromycin (i.e. A6) is now turned toward the major groove compared to

N7 in the corresponding Watson-Crick basepaired duplex (Ia), which has considerable implication

regarding the hydration behaviour as well as in the ligand binding capability of DNA duplex (Ib)

comapred to (Ia), which is under investigation in the author's laboratory.

9. References

(1) Kirby, A. J. In The Anomeric Effect and Related Stereoelectronic Effects at Oxygen; Springer Verlag: Berlin,

1983; Vol. 15.

(2) Deslongchamps, P. InStereoelectronic Effects in Organic Chemistry; Pergamon Press: Oxford, 1983.

(3) Tvaroska, I.; Bleha, T. In Advances in Carbohydrate Chemistry and Biochemistry; R. S. Tipson and D.

Horton, Ed.; Academic Press: San Diego, 1989; Vol. 47; pp 45.

(4) Juaristi, E.; Cuevas, G. Tetrahedron 1992, 48, 5019.

(5) Thatcher, G. R. J. In The Anomeric Effect and Associated Stereoelectronic Effects; G. R. J. Thatcher, Ed.;

American Chemical Society: Washington, DC, 1993; Vol. 539; pp 6.

(6) Graczyk, P. P.; Mikolajczyk, M. In Topics in Stereochemistry; Eliel, E. L. and Wilen, S.H., Ed.; John Wiley

& Sons, Inc.: New York, 1994; Vol. 21; pp 159.

(7) Perrin, C. L. Tetrahedron 1995, 51, 11901.

(8) Taira, K.; Uebayasi, M.; Maeda, H.; Furukawa, K. Protein Engineering 1990, 3, 691.

(9) Storer, J. W.; Uchimaru, T.; Tanabe, K.; Uebayasi, M.; Nishikawa, S.; Taira, K. J. Am. Chem. Soc. 1991,

113, 5216.

(10) Zhou, D.-M.; Taira, K. Chem. Rev. 1998, 98, 991.

(11) Bizzozero, S. A.; Dutler, H. Bioorg. Chem. 1981, 10, 46.

(12) Post, C. B.; Karplus, M. J. Am. Chem. Soc. 1986, 108, 1317.

(13) Taira, K. Bull. Chem. Soc. Jpn. 1987, 60, 1903.

(14) Koole, L. H.; Buck, H. M.; Nyilas, A.; Chattopadhyaya, J. Can. J. Chem. 1987, 65, 2089.

(15) Koole, L. H.; Buck, H. M.; Bazin, H.; Chattopadhyaya, J. Tetrahedron 1987, 43, 2989.

(16) Koole, L. H.; Moody, H. M.; Buck, H. M.; Grouiller, A.; Essadiq, H.; Vial, J.-M.; Chattopadhyaya, J. Recl.

Trav. Chim. Pays-Bas 1988, 107, 343.

(17) Plavec, J.; Koole, L. H.; Sandström, A.; Chattopadhyaya, J. Tetrahedron 1991, 47, 7363.

(18) Plavec, J.; Buet, V.; Grouiller, A.; Koole, L.; Chattopadhyaya, J. Tetrahedron 1991, 47, 5847.

(19) Plavec, J.; Koole, L. H.; Chattopadhyaya, J. J. Biochem. Biophys. Meth. 1992, 25, 253.

(20) Plavec, J.; Tong, W.; Chattopadhyaya, J. J. Am. Chem. Soc. 1993, 115, 9734.

(21) Plavec, J.; Garg, N.; Chattopadhyaya, J. J. Chem. Soc. Chem. Comm. 1993, 1011.

(22) Plavec, J.; Fabre-Buet, V.; Uteza, V.; Grouiller, A.; Chattopadhyaya, J. J. Biochem. Biophys. Meth. 1993, 26,

317.

(23) Plavec, J.; Thibaudeau, C.; Viswanadham, G.; Sund, C.; Chattopadhyaya, J. J. Chem. Soc. Chem. Com. 1994,

781.

(24) Thibaudeau, C.; Plavec, J.; Watanabe, K. A.; Chattopadhyaya, J. J. Chem. Soc. Chem. Com. 1994, 537.

(25) Thibaudeau, C.; Plavec, J.; Chattopadhyaya, J. J. Am. Chem. Soc. 1994, 116, 8033.

(26) Thibaudeau, C.; Plavec, J.; Garg, N.; Papchikhin, A.; Chattopadhyaya, J. J. Am. Chem. Soc. 1994, 116, 4038.

(27) Plavec, J.; Thibaudeau, C.; Chattopadhyaya, J. J. Am. Chem. Soc. 1994, 116, 6558.

(28) Plavec, J.; Thibaudeau, C.; Viswanadham, G.; Sund, C.; Sandström, A.; Chattopadhyaya, J. Tetrahedron

1995, 51, 11775.

(29) Plavec, J.; Thibaudeau, C.; Chattopadhyaya, J. Pure Appl. Chem. 1996, 68, 2137.

(30) Thibaudeau, C.; Plavec, J.; Chattopadhyaya, J. J. Org. Chem. 1996, 61, 266.

(31) Chattopadhyaya, J. Nucleic Acids Res. Symposium Ser. 1996, 35, 111.

(32) Luyten, I.; Thibaudeau, C.; Sandström, A.; Chattopadhyaya, J. Tetrahedron 1997, 53, 6433.

(33) Luyten, I.; Thibaudeau, C.; Chattopadhyaya, J. J. Org. Chem. 1997, 62, 8800.



149

(34) Luyten, I.; Thibaudeau, C.; Chattopadhyaya, J. Tetrahedron 1997, 53, 6903.

(35) Thibaudeau, C.; Chattopadhyaya, J. Nucleosides & Nucleotides 1997, 16, 523.

(36) Thibaudeau, C.; Földesi, A.; Chattopadhyaya, J. Tetrahedron 1997, 53, 14043.

(37) Thibaudeau, C.; Földesi, A.; Chattopadhyaya, J. Tetrahedron 1998, 54, 1867.

(38) Thibaudeau, C.; Chattopadhyaya, J. Nucleosides & Nucleotides 1998, 17, 1589.

(39) Thibaudeau, C.; Plavec, J.; Charropadhyaya, J. J. Org. Chem. 1998, 63, 4967.

(40) Luyten, I.; Matulic-Adamic, J.; Beigelman, L.; Chattopadhyaya, J. Nucleosides & Nucleotides 1998, 17,

1605.

(41) Thibaudeau, C.; Kumar, A.; Bekiroglu, S.; Matsuda, A.; Marquez, V. E.; Chattopadhyaya, J. J. Org. Chem.

1998, 63, 5447.

(42) Thibaudeau, C.; Nishizono, N.; Sumita, Y.; Matsuda, A.; Chattopadhyaya, J. Nucleosides & Nucleotides, 18,

1035-1053 (1999).

*(43) Acharya, P.; Nawrot, B.; Sprinzl, M.; Thibaudeau, C.; Chattopadhyaya, J. J. Chem. Soc. Perkin Trans. II,

p1531-1536 (1999).

*(44) (a) Acharya, P.; Trifonov, A.; Thibaudeau, C.; Földesi, A.; Chattopadhyaya, J. Angew Chem. Int. Ed. 38,

3645-3650 (1999); (b) Velikian, I., Acharya, P.; Trifonov, A.; Földesi, A.; Chattopadhyaya, J. J. Phys. Org.

Chem.. 13, 300-305 (2000).

*(45) Polak, M.; Plavec, J.; Trifonova, A.; Földesi, A.; Chattopadhyaya, J. JC.S. Perkin I, p2835-2843 (1999).

(46) Griffey, R. H.; Lesnik, E.; Freier, S.; Sanghvi, Y. S.; Teng, K.; Kawasaki, A.; Guinosso, C.; Wheeler, P.;

Mohan, V.; Cook, P. D. In Carbohydrate Modifications in Antisense Research; New Twists on Nucleic

Acids; Structural Properties of Modified Nucleosides Incorporated into Oligonucleotides V.S. Sanghvi &

P.D. Cook, Ed.; American Chemical Society: Washington DC, 1994; Vol. 580; pp 212,

(47) de Mesmaeker, A.; Häner, R.; Martin, P.; Moser, H. E. Acc. Chem. Res. 1995, 28, 366.

(48) Crooke, S. T. Med. Res. Rev. 1996, 16, 319.

(49) Freier, S. M.; Altmann, K.-H. Nucleic Acids Res. 1997, 25, 4429.

(50) Egli, M. Angew. Chem. Internat. Edit. 1996, 35, 1894.

(51) Zong, G.; Matsukura, M. Int. J. Immunotherapy 1990, 6, 187.

(52) Uhlmann, E.; Peyman, A. Chem. Rev. 1990, 90, 543.

(53) Altmann, K.-H.; Bévierre, M.-O.; De Mesmaeker, A.; Moser, H. E. Bioorg. & Med. Chem. Lett. 1995, 5, 431.

(54) Schmit, C.; Bévierre, M.-O.; De Mesmaeker, A.; Altmann, K.-H. Bioorg. & Med. Chem. Lett. 1994, 4, 1969.

(55) Brown, S. C.; Thomson, S. A.; Veal, J. M.; Davis, D. G. Science 1994, 265, 777.

(56) Betts, L.; Josey, J. A.; Veal, J. M.; Jordan, S. R. Science 1995, 270, 1838.

(57) Lubini, P.; Zürcher, W.; Egli, M. Chem. & Biol. 1994, 1, 39.

(58) Boggon, T. J.; Hancox, E. L.; McAuley-Hecht, K. E.; Connolly, B. A.; Hunter, W. N.; Brown, T.; Walker, R.

T.; Leonard, G. A. Nucleic Acids Res. 1996, 24, 951.

(59) Portmann, S.; Altmann, K.-H.; Reynes, N.; Egli, M. J. Am. Chem. Soc. 1997, 119, 2396.

(60) Gao, X.; Jeffs, P. W. J. Biomol. NMR 1994, 4, 17.

(61) Blommers, M. J. J.; Pieles, U.; de Mesmaeker, A. Nucleic Acids Res. 1994, 22, 4187.

(62) Tereshko, V.; Gryaznov, S.; Egli, M. J. Am. Chem. Soc. 1998, 120, 269.

(63) Berger, I.; Tereshko, V.; Ikeda, H.; Marquez, V. E.; Egli, M. Nucleic Acids Res. 1998, 26, 2473.

(64) Ikeda, H.; Fernandez, R.; Wilk, A.; Barchi Jr., J. J.; Huang, X.; Marquez, V. E. Nucl. Acids Res. 1998, 26,

2237.

(65) Edward, J. T. Chem. and Ind. 1955, 1102.

(66) Lemieux, R. U.; Chu, P. In Abstracts of Papers, 133rd Meeting of the American Chemical Society,; American

Chemical Society: San Francisco, CA, 1958; på 31N.

(67) Lemieux, R. U. In Molecular Rearrangements, Pt. 2; de Mayo P.., Ed.; Interscience Publishers: New York,

1964; pp 709.

(68) Biali, S. E. J. Org. Chem. 1992, 57, 2979.

(69) Booth, H.; Everett, J. R. J. Chem. Soc. Perkin II 1980, 255.

(70) Lemieux, R. U.; Kullnig, R. K.; Bernstein, H. J.; Schneider, W. G. J. Am. Chem. Soc. 1958, 80, 6098.

(71) Lemieux, R. U. In Profiles, Pathways and Dreams; J. I. Seeman, Ed.; American Chemical Society:

Washington DC, 1990; pp 89.

(72) Franck, R. W. Tetrahedron 1983, 39, 3251.

(73) Booth, G. E.; Ouellette, R. J. J. Org. Chem. 1966, 31, 544.

(74) Anderson, C. B.; Sepp, D. T. J. Org. Chem. 1967, 32, 607.

(75) Booth, H.; Khedhair, K. A.; Readshaw, S. A. Tetrahedron 1987, 43, 4699.

(76) de Hoog, A. J.; Buys, H. R.; Altona, C.; Havinga, E. Tetrahedron 1969, 25, 3365.

(77) Pierson, G. O.; Runquist, O. A. J. Org. Chem. 1968, 33, 2572.

(78) Anderson, C. B.; Sepp, D. T. Chem. & Ind. (London) 1964, 2054.

(79) de Hoog, A. J.; Havinga, E. Recl. Trav. Chim. Pays-Bas 1970, 89, 972.

(80) El-Kafrawy, A.; Perraud, R. C. R. Acad. Sci. Ser. C. 1975, 280, 1219.

(81) Barbry, D.; Couturier, D.; Ricard, G. J. Chem. Soc. Perkin II 1982, 249.

(82) Eliel, E. L.; Hargrave, K. D.; Pietrusiewicz, K. M.; Manoharan, M. J. Am. Chem. Soc. 1982, 104, 3635.


150

(83) Eliel, E. L.; Giza, C. A. J. Org. Chem. 1968, 33, 3754.

(84) Anderson, C. B.; Sepp, D. T. Tetrahedron 1968, 24, 1707.

(85) Anderson, C. B.; Sepp, D. T. J. Org. Chem. 1968, 33, 3272.

(86) Juaristi, E.; Cuevas, G. Tetrahedron Lett. 1992, 33, 2271.

(87) Booth, H.; Dixon, J. M.; Khedhair, K. A. Tetrahedron 1992, 48, 6161.

(88) Schneider, H.-J.; Hoppen, V. J. Org. Chem 1978, 43, 3866.

(89) Eliel, E. L.; Allinger, N. L.; Angyal, S. J.; Morrison, G. A. In Conformational Analysis; Wiley: New York,

1965.

(90) Jensen, F. R.; Bushweller, C. H.; Beck, B. H. J. Am. Chem. Soc. 1969, 91, 344.

(91) Booth, H.; Grindley, T. B.; Khedhair, K. A. J. Chem. Soc. Chem. Comm. 1982, 1047.

(92) Booth, H.; Khedhair, K. A. J. Chem. Soc. Chem. Comm. 1985, 467.

(93) Eliel, E. L. Acc. Chem. Res. 1970, 3, 1.

(94) Lemieux, R. U. Pure Appl. Chem. 1971, 25, 527.

(95) Eliel, E. L. Angew. Chem. 1972, 84, 779.

(96) Bailey, W. F.; Eliel, E. L. J. Am. Chem. Soc. 1974, 96, 1798.

(97) Penner, G. H.; Schaefer, T.; Sebastian, R.; Wolfe, S. Can. J. Chem. 1987, 65, 1845.

(98) Denmark, S. E.; Dappen, M. S.; Sear, N. L.; Jacobs, R. T. J. Am. Chem. Soc. 1990, 112, 3466.

(99) Irwin, J. J.; Ha, T.-K.; Dunitz, J. D. Helv. Chim. Acta 1990, 73, 1805.

(100) Luger, P.; Kothe, G.; Vangehr, K.; Paulsen, H.; Heiker, F. R. Carbohydr. Res. 1979, 68, 207.

(101) Paulsen, H.; Luger, P.; Heiker, F. R. In The Anomeric Effect, Origin and Consequences; W.A. Szarek and D.

Horton, Ed.; American Chemical Society: Washington DC, 1979; Vol. 87; pp 63.

(102) Durette, P. L.; Horton, D. Carbohyd. Res. 1971, 18, 403.

(103) Durette, P. L.; Horton, D. J. Org. Chem. 1971, 36, 2658.

(104) Durette, P. L.; Horton, D. Carbohydr. Res. 1971, 18, 389.

(105) Praly, J.-P.; Lemieux, R. U. Can. J. Chem. 1987, 65, 213.

(106) Lemieux, R. U.; Morgan, A. R. Can. J. Chem. 1965, 43, 2205.

(107) Paulsen, H.; Györgydeák, Z.; Friedmann, M. Chem. Ber. 1974, 107, 1590.

(108) Hosie, L.; Marshall, P. J.; Sinnott, M. L. J. Chem. Soc. Perkin Trans II 1984, 1121.

(109) Perrin, C. L. Pure Appl. Chem. 1995, 67, 719.

(110) Vaino, A. R.; Chan, S. S. C.; Szarek, W. A.; Thatcher, G. R. J. J. Org. Chem. 1996, 61, 4514.

(111) Perrin, C. L.; Armstrong, K. B. J. Am. Chem. Soc. 1993, 115, 6825.

(112) Perrin, C. L.; Fabian, M. A.; Armstrong, K. B. J. Org. Chem. 1994, 59, 5246.

(113) Finch, P.; Nagpurkar, A. G. Carbohydr. Res. 1976, 25, 275.

(114) Fabian, M. A.; Perrin, C. L.; Sinnott, M. L. J. Am. Chem. Soc. 1994, 116, 8398.

(115) Romers, C.; Altona, C.; Buys, H. R.; Havinga, E. In Topics in StereochemistryE.L. Eliel and N. L. Allinger,

Ed.; Wiley Interscience: New York, 1969; Vol. 4; pp 39.

(116) Lucken, E. A. C. J. Chem. Soc. 1959, 2954.

(117) Perrin, C. L.; Armstrong, K. B.; Fabian, M. A. J. Am. Chem. Soc. 1994, 116, 715.

(118) Wiberg, K. B.; Marquez, M. J. Am. Chem. Soc. 1994, 116, 2197.

(119) Hutchins, R. O.; Kopp, L. D.; Eliel, E. L. J. Am. Chem. Soc. 1968, 90, 7174.

(120) Box, V. G. S. Heterocycles 1990, 31, 1157.

(121) Fuchs, B.; Schleifer, L.; Tartakovsky, E. Nouv. J. Chim. 1984, 8, 275.

(122) Schleifer, L.; Senderowitz, H.; Aped, P.; Tartakovsky, E.; Fuchs, B. Carbohydr. Chem. 1990, 206, 21.

(123) Sweigart, D. A. J. Chem. Educ. 1973, 50, 322.

(124) David, S.; Eisenstein, O.; Hehre, W. J.; Salem, L.; Hoffmann, R. J. Am. Chem. Soc. 1973, 95, 3806.

(125) Laing, M. J. Chem. Educ. 1987, 64, 124.

(126) Brundle, C. R.; Turner, D. W. Proc. Roy. Soc. 1968, A307, 27.

(127) Reed, A. E.; Schleyer, P. v. R. Inorg. Chem. 1988, 27, 3969.

(128) Salzner, U. J. Org. Chem. 1995, 60, 986.

(129) Cossé-Barbi, A.; Dubois, J.-E. J. Am. Chem. Soc. 1987, 109, 1503.

(130) Cossé-Barbi, A.; Watson, D. G.; Dubois, J. E. Tetrahedron Lett. 1989, 30, 163.

(131) Salzner, U.; von Ragué Schleyer, P. J. Org. Chem. 1994, 59, 2138.

(132) Fuchs, B.; Ellencweig, A.; Tartakovsky, E.; Aped, P. Angew. Chem. Internat. Edit. 1986, 25, 287.

(133) Kessler, H.; Möhrle, H.; Zimmermann, G. J. Org. Chem. 1977, 42, 66.

(134) Wolfe, S. Acc. Chem. Res. 1972, 5, 102.

(135) Phillips, L.; Wray, V. J. Chem. Soc. Chem. Comm. 1973, 90.

(136) Brunck, T. K.; Weinhold, F. J. Am. Chem. Soc. 1979, 101, 1700.

(137) Juaristi, E. J. Chem. Educ. 1979, 56, 438.

(138) Zefirov, N. S.; Gurvich, L. G.; Shashkov, A. S.; Krimer, M. Z.; Vorob'eva, E. A. Tetrahedron 1976, 32, 1211.

(139) Zefirov, N. S. Tetrahedron 1977, 33, 3193.

(140) Zefirov, N. S.; Samoshin, V. V.; Subbotin, O. A.; Baranenkov, V. I.; Wolfe, S. Tetrahedron 1978, 34, 2953.

(141) Wiberg, K. B.; Murcko, M. A.; Laidig, K. E.; MacDougall, P. J. J. Phys. Chem. 1990, 94, 6956.



151

(142) Murcko, M. A.; DiPaola, R. A. J. Am. Chem. Soc. 1992, 114, 10010.

(143) Wiberg, K. B. Acc. Chem. Res. 1996, 29, 229.

(144) Abe, A.; Inomata, K. J. Mol. Struct. 1991, 245, 399.

(145) Craig, N. C.; Piper, L. G.; Wheeler, V. L. J. Phys. Chem. 1971, 75, 1453.

(146) Kaloustian, M. K.; Dennis, N.; Mager, S.; Evans, S. A.; Alcudia, F.; Eliel, E. L. J. Am. Chem. Soc. 1976, 98,

956.

(147) Eliel, E. L.; Juaristi, E. J. Am. Chem. Soc. 1978, 100, 6114.

(148) Harris, W. C.; Holtzclaw, J. R.; Kalasinsky, V. F. J. Chem. Phys. 1977, 67, 3330.

(149) Durig, J. R.; Liu, J.; Little, T. S.; Kalasinsky, V. F. J. Phys. Chem. 1992, 96, 8224.

(150) Hirano, T.; Nonoyama, S.; Miyajima, T.; Kurita, Y.; Kawamura, T.; Sato, H. J. Chem. Soc. Chem. Comm.

1986, 606.

(151) Fernholt, L.; Kveseth, K. Acta Chem. Scand. 1980, A34, 163.

(152) van Schaick, E. J. M.; Geise, H. J.; Mijlhoff, F. C.; Renes, G. J. Mol. Struct. 1973, 16, 23.

(153) Freisen, D.; Hedberg, K. J. Am. Chem. Soc. 1980, 102, 3987.

(154) Dixon, D. A.; Smart, B. E. J. Phys. Chem. 1988, 92, 2729.

(155) Kazerouni, M. R.; Hedberg, L.; Hedberg, K. J. Am. Chem. Soc. 1997, 119, 8324.

(156) Zefirov, N. S. Zhur. Org. Khim. 1970, 6, 1761.

(157) Zefirov, N. S.; Blagoveshchensky, V. S.; Kazimirchik, I. V.; Surova, N. S. Tetrahedron 1971, 27, 3111.

(158) Hoffmann, R. Acc. Chem. Res. 1971, 4, 1.

(159) Epiotis, N. D.; Sarkanen, S.; Bjorkquist, D.; Bjorkquist, L.; Yates, R. J. Am. Chem. Soc. 1974, 96, 4075.

(160) Gleiter, R.; Kobayashi, M.; Zefirov, N. S.; Palyulin, V. A. Dokl. Chem. 1977, 235, 396.

(161) Dionne, P.; St-Jacques, M. J. Am. Chem. Soc. 1987, 109, 2616.

(162) Craig, N. C.; Chen, A.; Suh, K. H.; Klee, S.; Mellau, G. C.; Winnewisser, B. P.; Winnewisser, M. J. Am.

Chem. Soc. 1997, 119, 4789.

(163) Runtz, G. R.; Bader, R. F. W.; Messer, R. R. Can. J. Chem. 1977, 55, 3040.

(164) Muir, M.; Baker, J. Molec. Phys. 1996, 89, 211.

(165) (a) Saenger, W. In Principles of Nucleic Acid Structure; Springer Verlag: Berlin, 1988. (b) ibid, pp93-96 for

anomeric effect across the 3'O-P-O-ester fragment in the sugar-phosphate backbone and references cited

therein.

(166) Watanabe, K. A. In Chemistry of Nucleosides and Nucleotides; L. B. Townsend, Ed.; Plenum Press: New

York, 1994; Vol. 3; p 421.

(167) IUPAC-IUB Eur. J. Biochem. 1983, 131, 9.

(168) IUPAC-IUB Pure Appl. Chem. 1983, 55, 1273.

(169) Angyal, S. J. In The Composition of Reducing Sugars in Solution, R.S. Tipson and D. Horton, Ed.; Academic

Press: London, 1984; Vol. 42; p 15.

(170) Squillacote, M.; Sheridan, R. S.; Chapman, O. L.; Anet, F. A. L. J. Am. Chem. Soc. 1975, 97, 3244.

(171) Anderson, J. E. Top. Curr. Chem. 1974, 45, 139.

(172) Kilpatrick, J. E.; Pitzer, K. S.; Spitzer, R. J. Am. Chem. Soc. 1947, 69, 2483.

(173) Strauss, H. L. Ann. Rev. Phys. Chem. 1983, 34, 301.

(174) de Leeuw, H. P. M.; Haasnoot, C. A. G.; Altona, C. Isr. J. Chem. 1980, 20, 108.

(175) Swarna Latha, Y.; Yathindra, N. Biopolymers 1992, 32, 249.

(176) Carreira, L. A.; Jiang, G. J.; Person, W. B.; Willis, J. N. J. J. Chem. Phys. 1972, 56, 1440.

(177) Hall, L. D.; Steiner, P. R.; Pedersen, C. Can. J. Chem. 1970, 48, 1155.

(178) Cremer, D.; Pople, J. A. J. Am. Chem. Soc. 1975, 97, 1354.

(179) Geise, H. J.; Altona, C.; Romers, C. Tetrahedron Lett. 1967, 15, 1383.

(180) Altona, C.; Geise, H. J.; Romers, C. Tetrahedron 1968, 24, 13.

(181) Altona, C.; Sundaralingam, M. J. Am. Chem. Soc. 1972, 94, 8205.

(182) Haasnoot, C. A. G.; de Leeuw, F. A. A. M.; de Leeuw, H. P. M.; Altona, C. Biopolymers 1981, 20, 1211.

(183) Diez, E.; Esteban, A. L.; Bermejo, F. J.; Altona, C.; de Leeuw, F. A. A. M. J. Mol. Struct. 1984, 125, 49.

(184) de Leeuw, F. A. A. M.; van Kampen, P. N.; Altona, C.; Diez, E.; Esteban, A. L. J. Mol. Struct. 1984, 125, 67.

(185) Diez, E.; Esteban, A. L.; Bermejo, F. J.; Rico, M. J. Phys. Chem. 1980, 84, 3191.

(186) Diez, E.; Esteban, A. L.; Guilleme, J.; Bermejo, F. L. J. Mol. Struct. 1981, 70, 61.

(187) Davies, D. B.; Sadikot, H. Biopolymers 1983, 22, 1843.

(188) Sarma, R. H.; Lee, C.-H.; Evans, F. E.; Yathindra, N.; Sundaralingam, M. J. Am. Chem. Soc. 1974, 96, 7337.

(189) Lüdemann, H.-D.; Röder, O.; Westhof, E.; Goldammer, E. v.; Müller, A. Biophys. Struct. Mechanism 1975,

1, 121.

(190) Rossi, C.; Pichi, M. P.; Tiezzi, E. Magn. Reson. Chem. 1990, 28, 348.

(191) Davies, D. B.; Danyluk, S. S. Biochemistry 1974, 13, 4417.

(192) Davies, D. B.; Danyluk, S. S. Biochemistry 1975, 14, 543.

(193) Plach, H.; Westhof, E.; Lüdemann, H.-D.; Mengel, R. Eur. J. Biochem. 1977, 80, 295.

(194) de Leeuw, H. P. M.; de Jager, J. R.; Koeners, H. J.; van Boom, J. H.; Altona, C. Eur. J. Biochem. 1977, 76,

209.

(195) Westhof, E.; Plach, H.; Cuno, I.; Lüdemann, H.-D. Nucleic Acids Res. 1977, 4, 939.


152

(196) Ekiel, I.; Darzynkiewicz, E.; Dudycz, L.; Shugar, D. Biochemistry 1978, 17, 1530.

(197) Klimke, G.; Cuno, I.; Lüdemann, H.-D.; Mengel, R.; Robins, M. J. Z. Naturforsch. 1979, 34c, 1075.

(198) Ekiel, I.; Remin, M.; Darzynkiewicz, E.; Shugar, D. Biochim. Biophys. Acta 1979, 562, 177.

(199) Guschlbauer, W.; Jankowski, K. Nucleic Acids Res. 1980, 8, 1421.

(200) Guschlbauer, W. Biochim. Biophys. Acta 1980, 610, 47.



(203) Haasnoot, C. A. G.; de Leeuw, F. A. A. M.; de Leeuw, H. P. M.; Altona, C. Org. Magn. Reson. 1981, 15, 43.

(204) Lipnick, R. L.; Fissekis, J. D. Biochim. Biophys. Acta 1980, 608, 96.

(205) de Leeuw, F. A. A. M.; Altona, C. J. Chem. Soc. Perkin II 1982, 375.

(206) Uesugi, S.; Kaneyasu, T.; Imura, J.; Ikehara, M.; Cheng, D. M.; Kan, L.-S.; Ts'o, P. O. P. Biopolymers 1983,

22, 1189.

(207) Yamamoto, Y.; Yokoyama, S.; Miyazawa, T.; Watanabe, K.; Higuchi, S. FEBS Lett. 1983, 157, 95.

(208) Cheng, D. M.; Kan, L.-S.; Ts'O, P. O. P.; Uesugi, S.; Takatsuka, Y.; Ikehara, M. Biopolymers 1983, 22, 1427.

(209) de Leeuw, F. A. A. M.; Altona, C. J. Comp. Chem. 1983, 4, 428.

(210) Chiasson, J. B.; Jankowski, K.; Brostow, W. J. Bioelectricity 1983, 2, 93.

(211) Westhof, E.; Sundaralingam, M. J. Am. Chem. Soc. 1983, 105, 970.

(212) Birnbaum, G. I.; Brisson, J. R.; Krawczyk, S. H.; Townsend, L. B. J. Am. Chem. Soc. 1985, 107, 7635.

(213) Rinkel, L. J.; Altona, C. J. Biomol. Struct. Dynamics 1987, 4, 621.

(214) Orbons, L. P. M.; van der Marel, G. A.; van Boom, J. H.; Altona, C. Eur. J. Biochem. 1987, 170, 225.

(215) Hosur, R. V.; Govil, G.; Miles, H. T. Magn. Reson. Chem. 1988, 26, 927.

(216) Swapna, G. V. T.; Jagannadh, B.; Gurjar, M. K.; Kunwar, A. C. Biochem. Biophys. Res. Commun. 1989, 164,

1086.

(217) Altona, C. In Theor. Biochem. & Mol. Biophys.; D. L. Beveridge and R. Lavery, Ed.; Adenine Press: 1990;

Vol. 1 (DNA); pp 1.

(218) Pieters, J. M. L.; de Vroom, E.; van der Marel, G. A.; van Boom, J. H.; Koning, T. M. G.; Kaptein, R.;

Altona, C. Biochemistry 1990, 29, 788.

(219) Jagannadh, B.; Reddy, D. V.; Kunwar, A. C. Biochem. Biophys. Res. Commun. 1991, 179, 386.

(220) Everaert, D. H.; Peeters, O. M.; de Ranter, C. J.; Blaton, N. M.; van Aerschot, A.; Herdewijn, P. Antivir.

Chem. Chemotherapy 1993, 4, 289.

(221) Wijmenga, S. S.; Mooren, M. M. W.; Hilbers, C. W. In NMR of Macromolecules; G. C. K. Roberts, Ed.;

Oxford University Press: Oxford, 1993; pp. 217-288.

(222) van den Boogaart, J. E.; Kalinichenko, E. N.; Podkopaeva, T. L.; Mikhailopulo, I. A.; Altona, C. Eur. J.

Biochem. 1994, 221, 759.

(223) Mooren, M. M. W.; Wijmenga, S. S.; van der Marel, G. A.; van Boom, J. H.; Hilbers, C. W. Nucleic Acids

Res. 1994, 22, 2658.

(224) Verberckmoes, F.; Esmans, E. L. Spectrochim. Acta 1995, 51A, 153.

(225) Rosemeyer, H.; Seela, F. J. Chem. Soc. Perkin Trans II 1997, 2341.

(226) Barchi, J. J. J.; Jeong, L.-S.; Siddiqui, M. A.; Marquez, V. E. J. Biochem. Biophys. Meth. 1997, 34, 11.

(227) de Leeuw, F. A. A. M.; Altona, C.; Kessler, H.; Bermel, W.; Friedrich, A.; Krack, G.; Hull, W. E. J. Am.

Chem. Soc. 1983, 105, 2237.

(228) Church, T. J.; Carmichael, I.; Serianni, A. S. J. Am. Chem. Soc. 1997, 119, 8946.

(229) Haasnoot, C. A. G.; Liskamp, R. M. J.; van Dael, P. A. W.; Noordik, J. H.; Ottenheijm, H. C. J. J. Am. Chem.

Soc. 1983, 105, 5406.

(230) Pavelcik, F.; Luptakova, E. Collect. Czech. Chem. Commun. 1990, 55, 1427.

(231) Varela, O.; Marino, C.; Cicero, D. An. Assoc. Quim. Argent. 1991, 79, 107.

(232) Cremer, D.; Pople, J. A. J. Am. Chem. Soc. 1975, 97, 1358.

(233) Pitzer, K. S.; Donath, W. E. J. Am. Chem. Soc. 1959, 81, 3213.

(234) Hoffmann, R. A.; van Wijk, J.; Leeflang, B. R.; Kamerling, J. P.; Altona, C.; Vliegenthart, J. F. G. J. Am.

Chem. Soc. 1992, 114, 3710.

(235) Cushley, R. J.; Codington, J. F.; Fox, J. J. Can. J. Chem. 1968, 46, 1131.

(236) Fang, K. N.; Kondo, N. S.; Miller, P. S.; Ts'o, P. O. P. J. Am. Chem. Soc. 1971, 93, 6647.

(237) Schleich, T.; Blackburn, B. J.; Lapper, R. D.; Smith, I. C. P. Biochemistry 1972, 11, 137.

(238) Tran-Dinh, S.; Cachaty, C. Biochim. Biophys. Acta 1973, 335, 1.

(239) Allore, B. D.; Queen, A.; Blonski, W. J.; Hruska, F. E. Can. J. Chem. 1983, 61, 2397.

(240) West, R. T.; Garza, L. A. I.; Winchester, W. R.; Walmsley, J. A. Nucleic Acids Res. 1994, 22, 5128.

(241) Feigon, J.; Wang, A. H.-J.; van der Marel, G. A.; van Boom, J. H.; Rich, A. Nucleic Acids Res. 1984, 12,

1243.

(242) Tran-Dinh, S.; Taboury, J.; Neumann, J.-M.; Huynh-Dinh, T.; Genissel, B.; Langlois d'Estaintot, B.; Igolen, J.

Biochemistry 1984, 23, 1362.

(243) Davis, P. W.; Hall, K.; Cruz, P.; Tinoco Jr., I.; Neilson, T. Nucleic Acids Res. 1986, 14, 1279.

(244) Davis, P. W.; Adamiak, R. W.; Tinoco, I. J. Biopolymers 1990, 29, 109.



153

(245) Agback, P.; Sandström, A.; Yamakage, S.-i.; Sund, C.; Glemarec, C.; Chattopadhyaya, J. J. Biochem.

Biophys. Meth. 1993, 27, 229.

(246) Agback, P.; Glemarec, C.; Yin, L.; Sandström, A.; Plavec, J.; Sund, C.; Yamakage, S.-i.; Viswanadham, G.;

Rousse, B.; Puri, N.; Chattopadhyaya, J. Tetrahedron Lett. 1993, 34, 3929.

(247) Izatt, R. M.; Christensen, J. J.; Rytting, J. H. Chem. Rev. 1971, 71, 439.

(248) Brahms, S.; Fritsch, V.; Brahms, J. G.; Westhof, E. J. Mol. Biol. 1992, 223, 455.

(249) Westhof, E.; Sundaralingam, M. J. Am. Chem. Soc. 1980, 102, 1493.

(250) Sasisekharan, V. Jerus. Symp. Quant. Chem. Biochem. 1973, 5, 247.

(251) Coulter, C. L. J. Am. Chem. Soc. 1973, 95, 570.

(252) Reddy, B. S.; Saenger, W. Acta Cryst. 1978, B34, 1520.

(253) Saenger, W.; Eckstein, F. J. Am. Chem. Soc. 1970, 92, 4712.

(254) Lee, C.-H.; Sarma, R. H. J. Am. Chem. Soc. 1976, 98, 3541.

(255) Ku Chwang, A.; Sundaralingam, M. Acta Cryst. 1974, B30, 1233.

(256) Coulter, C. L. Acta Cryst. 1969, B25, 2055.

(257) Varughese, K. I.; Lu, C. T.; Kartha, G. J. Am. Chem. Soc. 1982, 104, 3398.

(258) Lai, T. F.; Marsh, R. E. Acta Cryst. 1972, B28, 1982.

(259) Shikata, K.; Ueki, T.; Mitsui, T. Acta Cryst. 1973, B29, 31.

(260) Gurskaya, G. V.; Tsapkina, E. N.; Skaptasova, N. V.; Lindeman, A. A.; Struchkov, Y. T. Dokl Akad. Nauk.

USSR 1986, 291, 854.

(261) Birnbaum, G. I.; Giziewicz, J.; Gabe, E. J.; Lin, T.-S.; Prusoff, W. H. Can. J. Chem. 1987, 65, 2135.

(262) Camerman, A.; Mastropaolo, D.; Camerman, N. Proc. Natl. Acad. Sci. USA 1987, 84, 8239.

(263) van Roey, P.; Salerno, J. M.; Duax, W. L.; Chu, C. K.; Ahn, M. K.; Schinazi, R. F. J. Am. Chem. Soc. 1988,

110, 2277.

(264) Parthasarathy, R.; Kim, H. Biochem. Biophys. Res. Commun. 1988, 152, 351.

(265) Dyer, I.; Low, J. N.; Tollin, P.; Wilson, H. R.; Howie, R. A. Acta Cryst. 1988, C44, 767.

(266) van Roey, P.; Schinazi, R. F. Antiviral Chem. & Chemotherapy 1990, 1, 93.

(267) Camerman, N.; Mastropaolo, D.; Camerman, A. Proc. Natl. Acad. Sci. USA 1990, 87, 3534.

(268) Chu, C. K.; Bhadti, V. S.; Doboszewski, B.; Gu, Z. P.; Kosugi, Y.; Pullaiah, K. C.; van Roey, P. J. Org.

Chem. 1989, 54, 2217.

(269) Birnbaum, G. I.; Lin, T.-S.; Prusoff, W. H. Biochem. Biophys. Res. Commun. 1988, 151, 608.

(270) Koole, L. H.; Plavec, J.; Liu, H.; Vincent, B. R.; Dyson, M. R.; Coe, P. L.; Walker, R. T.; Hardy, G. W.;

Rahim, S. G.; Chattopadhyaya, J. J. Am. Chem. Soc. 1992, 114, 9936.

(271) Barnes, C. L.; Hawkinson, S. W.; Wigler, P. W. Acta Cryst. 1980, B36, 2299.

(272) Birnbaum, G. I.; Watanabe, K. A.; Fox, J. J. Can. J. Chem. 1980, 58, 1633.

(273) Barrett, A. G. M.; Broughton, H. B. J. Org. Chem. 1984, 49, 3673.

(274) Mc Kenna, R.; Neidle, S.; Serafinowski, P. Acta Cryst. 1987, C43, 2358.

(275) Mc Kenna, R.; Neidle, S.; Serafinowski, P. Acta Cryst. 1990, C46, 2448.

(276) Mumper, M. W.; Aurenge, C.; Hosmane, R. S. J. Biomol. Struct. Dynamics 1994, 11, 1107.

(277) Koyama, G.; Nakamura, H.; Umezawa, H.; Iitaka, Y. Acta Cryst. 1976, B32, 813.

(278) Prusiner, P.; Brennan, T.; Sundaralingam, M. Biochemistry 1973, 12, 1196.

(279) Rohrer, D.; Sundaralingam, M. J. Am. Chem. Soc. 1970, 92, 4950.

(280) Koyama, G.; Umezawa, H.; Iitaka, Y. Acta Cryst. 1974, B30, 1511.

(281) Singh, P.; Hodgson, D. J. Acta Cryst. 1977, B33, 3189.

(282) Hempel, A.; Lane, B. G.; Camerman, N. Acta Cryst. 1997, C53, 1707.

(283) Krol, M. C.; Huige, C. J. M.; Altona, C. J. Comp. Chem. 1990, 11, 765.

(284) Marquez, V. E. In Advances in Antiviral Drug Design; E. de Clercq, Ed.; JAI Press, Inc.: Greenwich, CT,

1996; Vol. 2; pp 89.

(285) Borthwick, A. D.; Biggadike, K. Tetrahedron 1992, 48, 571.

(286) Marquez, V. E.; Lim, M.-I. Med. Res. Rev. 1986, 6, 1.

(287) Altmann, K.-H.; Imwinkelried, R.; Kesselring, R.; Rihs, G. Tetrahedron Lett. 1994, 35, 7625.

(288) Shealy, Y. F.; O'Dell, C. A. Tetrahedron Lett. 1969, 27, 2231.

(289) Marumoto, R.; Yoshioka, Y.; Furukawa, Y.; Honjo, M. Chem. Pharm. Bull. 1976, 24, 2624.

(290) Kinoshita, K.; Yaginuma, S.; Hayashi, M.; Nakatsu, K. Nucleosides & Nucleotides 1985, 4, 661.

(291) Popescu, A.; Hörnfeldt, A.-B.; Gronowitz, S.; Johansson, N. G. Nucleosides & Nucleotides 1995, 14, 1639.

(292) Siddiqui, M. A.; Ford, H. J.; George, C.; Marquez, V. E. Nucleosides & Nucleotides 1996, 15, 235.

(293) Rodriguez, J. B.; Marquez, V. E.; Nicklaus, M. C.; Mitsuya, H.; Jr., B. J. J. Med. Chem. 1994, 37, 3389.

(294) Kishi, T.; Muroi, M.; Kusaka, T.; Nishikawa, M.; Kamiya, K.; Mizuno, K. Chem. Pharm. Bull. 1972, 20, 940.

(295) Kálman, A.; Koritsánszky, T.; Béres, J.; Sági, G. Nucleosides & Nucleotides 1990, 9, 235.

(296) Biggadike, K.; Borthwick, A. D.; Exall, A. M.; Kirk, B. E.; Roberts, S. M.; Youds, P.; Slawin, A. M. Z.;

Williams, D. J. J. Chem. Soc. Chem. Comm. 1987, 255.

(297) Marumoto, R. AIDS-Iyakuhin Kaihatsu-Kenkyu-Hokuku (The Japan Health Sciences Foundation) 1991, 3.

(298) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.;

Keith, T.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.;


154

Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C.

Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox,

D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-Gordon, M.; Gonzalez, C.; Pople, J. A.

Gaussian 94 Revision C.3; Gaussian, Inc.: Pittsburgh PA, 1995.

(299) Govil, G.; Saran, A. J. Theor. Biol. 1971, 33, 399.

(300) Lugovskoi, A. A.; Dashevskii, V. G. Mol. Biol. (USSR) 1972, 6, 354.

(301) Lugovskoi, A. A.; Dashevskii, V. G.; Kitaigorodskii, A. I. Mol. Biol. (USSR) 1972, 6, 361.

(302) Lugovskoi, A. A.; Dashevskii, V. G.; Kitaigorodskii, A. I. Mol. Biol. (USSR) 1972, 6, 494.

(303) Levitt, M.; Warshel, A. J. Am. Chem. Soc. 1978, 100, 2607.

(304) Olson, W. K.; Sussman, J. L. J. Am. Chem. Soc. 1982, 104, 270.

(305) Olson, W. K. J. Am. Chem. Soc. 1982, 104, 278.

(306) Pearlman, D. A.; Kim, S.-H. J. Biomol. Struct. Dynamics 1985, 3, 99.

(307) Röder, O.; Lüdemann, H.-D.; von Goldammer, E. Eur. J. Biochem. 1975, 53, 517.

(308) Plavec, J.; Roselt, P.; Földesi, A.; Chattopadhyaya, J. Mag. Res. Chem. 1998, 36, 732.

(309) Jalluri, R. K.; Yuh, Y. H.; Taylor, E. W. In The Anomeric Effect and Associated Stereoelectronic Effects; G.

R. J. Thatcher, Ed.; American Chemical Society: Washington, DC, 1993; Vol. 539; pp 277.

(310) Allen, F. H.; Davies, J. E.; Galloy, J. J.; Johnson, O.; Kennard, O.; MacRae, C. F.; Mitchell, E. M.; Mitchell,

G. F.; Smith, J. M.; Watson, D. G. J. Chem. Inf. Comput. Sci. 1991, 31, 187.

(311) Lo, A.; Shefter, E.; Cochran, T. G. J. Pharm. Sc. 1975, 64, 1707.

(312) Fox, J. J.; Shugar, D. Biochim. Biophys. Acta 1952, 9, 369.

(313) Tran-Dinh, S.; Thiéry, J.; Guschlbauer, W.; Dunand, J.-J. Biochim. Biophys. Acta 1972, 281, 289.

(314) Sarma, R. H.; Mynott, R. J. J. Am. Chem. Soc. 1973, 95, 1641.

(315) Remin, M.; Shugar, D. J. Am. Chem. Soc. 1973, 95, 8146.

(316) Altona, C.; Sundaralingam, S. J. Am. Chem. Soc. 1973, 95, 2333.

(317) Tran-Dinh, S.; Guschlbauer, W. Nucleic Acids Res. 1975, 3, 873.

(318) Hruska, F. E.; Wood, D. J.; Singh, H. Biochim. Biophys. Acta 1977, 474, 129.

(319) Seela, F.; Becher, G.; Rosemeyer, H.; Reuter, H.; Kastner, G.; Mikhailopulo, I. A. Helv. Chim. Acta 1999, 82,

105.

(320) Hillen, W.; Gassen, H. G. Biochim. Biophys. Acta 1978, 518, 7.

(321) Egert, E.; Lindner, H. J.; Hillen, W.; Böhm, M. C. J. Am. Chem. Soc. 1980, 102, 3707.

(322) Uhl, W.; Reiner, J.; Gassen, H. G. Nucleic Acids Res. 1983, 11, 1167.

(323) O'Leary, D. J.; Kishi, Y. J. Org. Chem. 1994, 59, 6629.

(324) Birnbaum, G. I.; Blonski, W. J. P.; Hruska, F. E. Can. J. Chem. 1983, 61, 2299.

(325) Cadet, J.; Ducolomb, R.; Taieb, C. Tetrahedron Lett. 1975, 40, 3455.

(326) Birnbaum, G. I.; Hruska, F. E.; Niemczura, W. P. J. Am. Chem. Soc. 1980, 102, 5586.

(327) Bergström, D. E.; Zhang, P.; Toma, P. H.; Andrews, P. C.; Nichols, R. J. Am. Chem. Soc. 1995, 117, 1201.

(328) Post, M. L.; Birnbaum, G. I.; Huber, C. P.; Shugar, D. Biochim. Biophys. Acta 1977, 479, 133.

(329) Sundaralingam, M. J. Am. Chem. Soc. 1971, 93, 6644.

(330) Rohrer, D. C.; Sundaralingam, M. J. Am. Chem. Soc. 1970, 92, 4956.

(331) Remin, M.; Ekiel, I.; Shugar, D. Eur. J. Biochem. 1975, 53, 197.

(332) Seto, H.; Otake, N.; Yonehara, H. Tetrahedron Lett. 1972, 38, 3991.

(333) Ruzic-Toros, Z. Acta Cryst. 1979, B35, 1277.

(334) Crowfoot Hodgkin, D.; Lindsey, J.; Sparks, R. A.; Trueblood, K. N.; White, J. G. Proc. Roy. Soc. 1962,

A266, 494.

(335) Brink-Shoemaker, C.; Cruickshank, D. W. J.; Crowfoot Hodgkin, D.; Kamper, M. J.; Pilling, D. Proc. Roy.

Soc. 1964, A278, 1.

(336) Hawkinson, S. W.; Coulter, C. L.; Greaves, M. L. Proc. Roy. Soc. 1970, A318, 143.

(337) Savage, H. F. J.; Lindley, P. F.; Finney, J. L.; Timmins, P. A. Acta Cryst. 1987, B43, 280.

(338) Lenhert, P. G. Proc. Roy. Soc. 1968, A303, 45.

(339) Hamor, T. A.; O'Leary, M. K.; Walker, R. T. Acta Cryst. 1977, B33, 1218.

(340) Shefter, E.; Kotick, M. P.; Bardos, J. J. Pharm. Sci. 1967, 56, 1293.

(341) Ide, H.; Shimizu, H.; Kimura, Y.; Sakamoto, S.; Makino, K.; Glackin, M.; Wallace, S. S.; Nakamuta, H.;

Sasaki, M.; Sugimoto, N. Biochemistry 1995, 34, 6947.

(342) Gutowski, G. E.; Chaney, M. O.; Jones, N. D.; Hamill, R. L.; Davis, F. A.; Miller, R. D. Biochem. Biophys.

Res. Commun. 1973, 51, 312.

(343) Armstrong, V. W.; Dattagupta, J. K.; Eckstein, F.; Saenger, W. Nucleic Acids Res. 1976, 3, 1791.

(344) Cline, S. J.; Hodgson, D. J. Biochim. Biophys. Acta 1979, 563, 540.

(345) Post, M. L.; Huber, C. P.; Birnbaum, G. I.; Shugar, D. Can. J. Chem. 1981, 59, 238.

(346) Piper, I. M.; MacLean, D. B.; Faggiani, R.; Lock, C. J. L.; Szarek, W. A. Can. J. Chem. 1985, 63, 2915.

(347) Serianni, A.; Barker, R. J. Org. Chem. 1984, 49, 3292.

(348) Serianni, A. S.; Chipman, D. M. J. Am. Chem. Soc. 1987, 109, 5297.

(349) Kline, P. C.; Serianni, A. S. J. Am. Chem. Soc. 1990, 112, 7373.



155

(350) Raap, J.; van Boom, J. H.; van Lieshout, H. C.; Haasnoot, C. A. G. J. Am. Chem. Soc. 1988, 110, 2736.

(351) Ellervik, U.; Magnusson, G. J. Am. Chem. Soc. 1994, 116, 2340.

(352) Teng, K.; Cook, P. D. J. Org. Chem. 1994, 59, 278.

(353) Cyr, N.; Perlin, A. S. Can. J. Chem. 1979, 57, 2504.

(354) Le Doan, T.; Perrouault, L.; Praseuth, D.; Habhoub, N.; Decout, J.-L.; Thouong, N. T.; Lhomme, J.; Hélène,

C. Nucleic Acids Res. 1987, 15, 7749.

(355) Moser, H. E.; Dervan, P. B. Science 1987, 238, 645.

(356) Povsic, T. J.; Dervan, P. B. J. Am. Chem. Soc. 1989, 111, 3059.

(357) Xodo, L. E.; Manzini, G.; Quadrifoglio, F.; van der Marel, G. A.; van Boom, J. H. Nucleic Acids Res. 1991,

19, 5625.

(358) Brown, T.; Leonard, G. A.; Booth, E. D.; Kneale, G. J. Mol. Biol. 1990, 212, 437.

(359) Boulard, Y.; Cognet, J. A. H.; Gabarro-Arpa, J.; Le Bret, M.; Sowers, L. C.; Fazakerley, G. V. Nucleic Acids

Res. 1992, 20, 1933.

(360) Wang, C.; Gao, H.; Gaffney, B. L.; Jones, R. A. J. Am. Chem. Soc. 1991, 113, 5486.

(361) Macaya, R. F.; Gilbert, D. E.; Malek, S.; Sinsheimer, J. S.; Feigon, J. Science 1991, 254, 270.

(362) Gao, X.; Patel, D. J. J. Am. Chem. Soc. 1988, 110, 5178.

(363) Dolinnaya, N. G.; Fresco, J. R. Proc. Natl. Acad. Sci. USA 1992, 89, 9242.

(364) Dolinnaya, N. G.; Braswell, E. H.; Fossella, J. A.; Klump, H.; Fresco, J. R. Biochemistry 1993, 32, 10263.

(365) Robinson, H.; van der Marel, G. A.; van Boom, J. H.; Wang, A. H.-J. Biochemistry 1992, 31, 10510.

(366) Liu, K.; Miles, H. T.; Frazier, J.; Sasisekharan, V. Biochemistry 1993, 32, 11802.

(367) Jaishree, T. N.; Wang, A. H.-J. Nucleic Acids Res. 1993, 21, 3839.

(368) Gehring, K.; Leroy, J.-L.; Guéron, M. Nature 1993, 363, 561.

(369) Leroy, J.-L.; Gehring, K.; Kettani, A.; Guéron, M. Biochemistry 1993, 32, 6019.

(370) Garcia, B.; Palacios, J. C. Ber. Bunsenges. Phys. Chem. 1988, 92, 696.

(371) Taylor, R.; Kennard, O. J. Mol. Struct. 1982, 78, 1.

(372) Jardetzky, C. D.; Jardetzky, O. J. Am. Chem. Soc. 1960, 82, 222.

(373) Angell, C. L. J. Chem. Soc. 1961, 504.

(374) Tsuboi, M.; Kyogoku, Y.; Shimanouchi, T. Biochim. Biophys. Acta 1962, 55, 1.

(375) Katritzky, A. R.; Waring, A. J. J. Chem. Soc. 1962, 1540.

(376) Lord, R. C.; Thomas, G. J. J. Spectrochim. Acta 1967, 23A, 2551.

(377) Clauwaert, J.; Stockx, J. Z. Naturforsch. 1968, 23B, 25.

(378) Wagner, R.; von Philipsborn, W. Helv. Chim. Acta 1970, 53, 299.

(379) Christensen, J. J.; Rytting, H.; Izatt, R. M. Biochemistry 1970, 9, 4907.

(380) Poulter, C. D.; Anderson, R. B. Tetrahedron Lett. 1972, 36, 3823.

(381) Dunn, D. B.; Hall, R. H. In Handbook of Biochemistry and Molecular Biology; G. D. Fasman, Ed.; C.R.C.

Press: Cleveland, Ohio, 1975; Vol. I; pp 65.

(382) Kremer, A. B.; Mikita, T.; Beardsley, G. P. Biochemistry 1987, 26, 391.

(383) Wempen, I.; Fox, J. J. J. Am. Chem. Soc. 1964, 86, 2474.

(384) Cushley, R. J.; Wempen, I.; Fox, J. J. J. Am. Chem. Soc. 1968, 90, 709.

(385) Ward, D. C.; Reich, E.; Stryer, L. J. Biol. Chem. 1969, 244, 1228.

(386) Cho, B. P.; McGregor, M. A. Nucleosides & Nucleotides 1994, 13, 481.

(387) Koyama, G.; Umezawa, H. J. Antibiotics 1965, A18, 175.

(388) Robins, R. K.; Townsend, L. B.; Cassidy, F.; Gerster, J. F.; Lewis, A. F.; Miller, R. L. J. Heterocycl. Chem.

1966, 3, 110.

(389) Chu, C. K.; Wempen, I.; Watanabe, K. A.; Fox, J. J. J. Org. Chem. 1976, 41, 2793.

(390) Chambers, R. W. Prog. Nucleic Acids Res. Mol. Biol. 1966, 5, 349.

(391) Smith, R. M.; Martell, A. E.; Chen, Y. Pure Appl. Chem. 1991, 63, 1015.

(392) Marzilli, L. G.; de Castro, B.; Solorzano, C. J. Am. Chem. Soc. 1982, 104, 461.

(393) Marzilli, L. G.; Kline, T. P.; Live, D.; Zon, G. In Metal-DNA Chemistry; T. D. Tullius, Ed.; American

Chemical Society: Washington DC, 1989; pp 119.

(394) InInteractions of Metal Ions with Nucleotides, Nucleic Acids, and Their Constituents; A. Sigel and H. Sigel

ed.; M. Dekker, Inc.: New York, 1996; Vol. 32 (Metal Ions in Biological Systems).

(395) Polak, M.; Plavec, J. Nucleosides & Nucleotides 1998, 17, 2011.

(396) Puglisi, J. D.; Wyatt, J. R.; Tinoco, I. J. Biochemistry 1990, 29, 4215.

(397) Legault, P.; Pardi, A. J. Am. Chem. Soc. 1994, 116, 8390.

(398) Legault, P.; Pardi, A. J. Am. Chem. Soc. 1997, 119, 6621.

(399) Singleton, S. F.; Dervan, P. B. Biochemistry 1992, 31, 10995.

(400) Tran-Dinh, S.; Guschlbauer, W.; Guéron, M. J. Am. Chem. Soc. 1972, 94, 7903.

(401) Remaud, G.; Zhou, X.-x.; Welch, C. J.; Chattopadhyaya, J. Tetrahedron 1986, 42, 4057.

(402) Schuster, I. I.; Roberts, J. D. J. Org. Chem. 1979, 44, 3864.

(403) Remaud, G.; Welch, C.; Zhou, X.-x.; Chattopadhyaya, J. Nucleosides & Nucleotides 1988, 7, 167.

(404) Markowski, V.; Sullivan, G. R.; Roberts, J. D. J. Am. Chem. Soc. 1977, 99, 714.

(405) Remin, M.; Darzynkiewicz, E.; Dworak, A.; Shugar, D. J. Am. Chem. Soc. 1976, 98, 367.


156

(406) Froehler, B. C.; Wadwani, S.; Terhorst, T. J.; Gerrard, S. R. Tetrahedron Lett. 1992, 33, 5307.

(407) Manoharan, M., unpublished results.

(408) Ramzaeva, N.; Seela, F. Helv. Chim. Acta 1996, 79, 1549.

(409) Seela, F.; Thomas, H. Helv. Chim. Acta 1995, 78, 94.

(410) Buhr, C. A.; Wagner, R. W.; Grant, D.; Froehler, B. C. Nucleic Acids Res. 1996, 24, 2974.

(411) Hashimoto, H.; Nelson, M. G.; Switzer, C. J. Am. Chem. Soc. 1993, 115, 7128.

(412) Lesnik, E.

(413) Gryaznov, S.; Schultz, R. G. Tetrahedron Lett. 1994, 35, 2489.

(414) Sanghvi, Y. S.; Hoke, G. D.; Freier, S. M.; Zounes, M. C.; Gonzalez, C.; Cummins, L.; Sasmor, H.; Cook, P.

D. Nucleic Acids Res. 1993, 21, 3197.

(415) Dunkel, M.; Cook, P. D.; Acevedo, O. L. J. Heterocycl. Chem. 1993, 30, 1421.

(416) Imazawa, M.; Ueda, T.; Ukita, T. Chem. Pharm. Bull. (Tokyo) 1975, 23, 604.

(417) Uesugi, S.; Miki, H.; Ikehara, M.; Iwahashi, H.; Kyogoku, Y. Tetrahedron Lett. 1979, 42, 4073.

(418) Ikehara, M. Heterocycles 1984, 21, 75.

(419) Koole, L. H.; Wu, J.-C.; Neidle, S.; Chattopadhyaya, J. J. Am. Chem. Soc. 1992, 114, 2687.

(420) Xi, Z.; Glemarec, C.; Chattopadhyaya, J. Tetrahedron 1993, 49, 7525.

(421) Fraser, A.; Wheeler, P.; Dan Cook, P.; Sanghvi, Y. S. J. Heterocycl. Chem. 1993, 30, 1277.

(422) Hossain, N.; Papchikhin, A.; Garg, N.; Fedorov, I.; Chattopadhyaya, J. Nucleosides & Nucleotides 1993, 12,

499.

(423) Glemarec, C.; Reynolds, R. C.; Crooks, P. A.; Maddry, J. A.; Akhtar, M. S.; Montgomery, J. A.; Secrist, J. A.

I.; Chattopadhyaya, J. Tetrahedron 1993, 49, 2287.

(424) van Aerschot, A.; Everaert, D.; Balzarini, J.; Augustyns, K.; Jie, L.; Janssen, G.; Peeters, O.; Blaton, N.; de

Ranter, C.; de Clerq, E.; Herdewijn, P. J. Med. Chem. 1990, 33, 1833.

(425) Sanghvi, Y. S.; Vasseur, J.-J.; Debart, F.; Cook, P. D. Collect. Czech. Chem. Commun. 1993, 58, 158.

(426) Thibaudeau, C.; Chattopadhyaya, J. unpublished result

(427) de Leeuw, F. A. A. M.; van Wijk, J.; Altona, C. PSEUROT Version 5.4, Leiden University, Leiden,

Netherlands 1992,

(428) Hägele, G. DAISY-Program-Package for Simulation and Automated Analysis of High Resolution NMR

Spectra on Bruker's Aspect X32 and Silicon Graphics Indy computers 1991-1996.

(429) Altona, C.; Ippel, J. H.; Westra Hoekzema, A. J. A.; Erkelens, C.; Groesbeek, M.; Donders, L. A. Magn.

Reson. Chem. 1989, 27, 564.

(430) Altona, C.; Francke, R.; de Haan, R.; Ippel, J. H.; Daalmans, G. J.; Westra Hoekzema, A. J. A.; van Wijk, J.

Magn. Reson. Chem. 1994, 32, 670.

(431) Haasnoot, C. A. G.; De Leeuw, F. A. A. M.; Altona, C. Tetrahedron 1980, 36, 2783.

(432) Harte, W. E. J.; Starrett, J. E. J.; Martin, J. C.; Mansuri, M. M. Biochem. Biophys. Res. Commun. 1991, 175,

298.

(433) van Roey, P.; Salerno, J. M.; Chu, C. K.; Schinazi, R. F. Proc. Natl. Acad. Sci. USA 1989, 86, 3929.

(434) Sanghvi, Y. S.; Hanna, N. B.; Larson, S. B.; Robins, R. K.; Revanker, G. R. Nucleosides & Nucleotides 1987,

6, 761.

(435) Lankhorst, P. P.; Haasnoot, C. A. G.; Erkelens, C.; Altona, C. J. Biomol. Struct. Dynamics 1984, 1, 1387.

(436) Imai, K.; Osawa, E. Magn. Reson. Chem. 1990, 28, 668.

(437) For a recent review, see the contributions in Magn. Res. Chem. 1996, 34, S1-S186.

(438) Karplus, M. J. Chem. Phys. 1959, 30, 11.

(439) Huggins, M. L. J. Am. Chem. Soc. 1953, 75, 4123.

(440) Karplus, M. J. Am. Chem. Soc. 1963, 85, 2870.

(441) Barfield, M.; Smith, W. B. J. Am. Chem. Soc. 1992, 114, 1574.

(442) Marshall, J. L.; Walter, S. R.; Barfield, M.; Marchand, A. P.; Marchand, N. W.; Segre, A. L. Tetrahedron

1976, 32, 537.

(443) Jaworski, A.; Ekiel, I.; Shugar, D. J. Am. Chem. Soc. 1978, 100, 4357.

(444) de Leeuw, F. A. A. M.; van Beuzekom, A. A.; Altona, C. J. Comp. Chem. 1983, 4, 438.

(445) Diez, E.; San-Fabian, J.; Guilleme, J.; Altona, C.; Donders, L. A. Mol. Phys. 1989, 68, 49.

(446) Donders, L. A.; de Leeuw, F. A. A. M.; Altona, C. Magn. Reson. Chem. 1989, 27, 556.

(447) Plavec, J. Ph.D. Thesis, Department of Bioorganic Chemistry, Uppsala University, Sweden 1995.

(448) Blandin, M.; Son, T.-D.; Catlin, J. C.; Guschlbauer, W. Biochim. Biophys. Acta 1974, 361, 249.

(449) Blandin, M.; Jankowski, K. Rev. Latinoameric. Quim. 1983, 14, 45.

(450) Huang, J.-T.; Chen, L.-C.; Wang, L.; Kim, M.-H.; Warshaw, J. A.; Armstrong, D.; Zhu, Q.-Y.; Chou, T.-C.;

Watanabe, K. A.; Matulic-Adamic, J.; Su, T.-L.; Fox, J. J.; Polsky, B.; Baron, P. A.; Gold, J. W. M.; Hardy,

W. D.; Zuckerman, E. J. Med. Chem. 1991, 34, 1640.

(451) Heinemann, V.; Hertel, L. W.; Grindey, G. B.; Plunkett, W. Cancer Res. 1988, 48, 4024.

(452) Brox, L. W.; LePage, G. A.; Hendler, S. S.; Shannahoff, D. H. Cancer Res. 1978, 34, 1838.

(453) Tsuchiya, T. Adv. Carbohydr. Chem. Biochem. 1990, 48, 91.



157

(454) Marquez, V. E.; Lim, B. B.; Barchi, J. J. J.; Nicklaus, M. C. In Nucleosides and Nucleotides as Antitumor

and Antiviral Agents; C. K. B. Chu, D.C., Ed.; Plenum Press: New York, 1993; pp 265. (455) Herdewijn, P.; Pauwels, R.; Baba, M.; Balzarini, J.; de Clerq, E. J. Med. Chem. 1987, 30, 2131. (456) Herdewijn, P.; van Aerschot, A.; Kerremans, L. Nucleosides & Nucleotides 1989, 8, 65. (457) Martin, J. A.; Brushnell, D. J.; Duncan, I. B.; Dunsdon, S. J.; Hall, M. J.; Machin, P. J.; Merret, J. H.; Parkes,

K. E. B.; Roberts, N. A.; Thomas, G. J.; Galpin, S. A.; Kinchington, D. J. Med. Chem. 1990, 33, 2137. (458) Kawasaki, A. M.; Casper, M. D.; Freier, S. M.; Lesnik, E. A.; Zounes, M. C.; Cummins, L. L.; Gonzalez, C.;

Cook, P. D. J. Med. Chem. 1993, 36, 831. (459) Monia, B. P.; Lesnik, E. A.; Gonzalez, C.; Lima, W. F.; McGee, D.; Guinosso, C. J.; Kawasaki, A. M.; Cook,

P. D.; Freier, S. M. J. Biol. Chem. 1993, 268, 14514. (460) Fazakerley, G. V.; Uesugi, S.; Izumi, A.; Ikehara, M.; Guschlbauer, W. FEBS Lett. 1985, 182, 365. (461) Phillips, L.; Wray, V. J. Chem. Soc. (B) 1971, 1618. (462) Hall, L. D.; Manville, J. F. Chem. Comm. 1968, 37. (463) Foster, A. B.; Hems, R.; Hall, L. D. Can. J. Chem. 1970, 48, 3937. (464) Gopinathan, M. S.; Narasimhan, P. T. Molec. Phys. 1971, 21, 1141. (465) Govil, G. Molec. Phys. 1971, 21, 953. (466) Ihrig, A. M.; Smith, S. L. J. Am. Chem. Soc. 1970, 92, 759. (467) Williamson, K. L.; Li Hsu, Y.-F.; Hall, F. H.; Swager, S.; Coulter, M. S. J. Am. Chem. Soc. 1968, 90, 6717. (468) Williamson, K. L. W.; Li, Y.-F.; Hall, F. H.; Swager, S. J. Am. Chem. Soc. 1966, 88, 5678. (469) Berthelot, J.-P.; Jacquesy, J.-C.; Lesivalles, J. Bull. Soc. Chim. Fr. 1971, 5, 1896. (470) Lankin, D. C.; Chandrakumar, N. S.; Rao, S. N.; Spangler, D. P.; Snyder, J. P. J. Am. Chem. Soc. 1993, 115,

3356. (471) Emsley, J. W.; Phillips, L.; V., W. In Progress in Nuclear Magnetic Resonance Spectroscopy; J. W. Emsley,

Feeney, J., Sutcliffe, L.H., Ed.; Pergamon Press: New York, 1977; Vol. 10(3/4); pp 229. (472) Abraham, R. J.; Cavalli, L. Molec. Phys. 1965, 9, 67. (473) Wilkinson, L. SYSTAT: The System for Statistics Systat Inc.: Evanston, IL, 1989. (474) Urata, H.; Shinohara, K.; Ogura, E.; Ueda, Y.; Akagi, M. J. Am. Chem. Soc. 1991, 113, 8174. (475) Urata, H.; Ogura, E.; Shinohara, K.; Ueda, Y.; Akagi, M. Nucl. Acids Res. 1992, 20, 3325. (476) Hashimoto, Y.; Iwanami, N.; Fujimori, S.; Shudo, K. J. Am. Chem. Soc. 1993, 115, 9883. (477) Doi, M.; Inoue, M.; Tomoo, K.; Ishida, T.; Ueda, Y.; Akagi, M.; Urata, H. J. Am. Chem. Soc. 1993, 115,

10432. (478) Joyce, G. F.; Visser, G. M.; van Boeckel, C. A. A.; van Boom, J. H.; Orgel, L. E.; van Westrenen, J. 1984,

310, 602. (479) Eichhorn, G. L. In Inorganic BiochemistryElsevier/North Holland: New York, 1973; pp 1210. (480) Buchanan, G. W.; Stothers, J. B. Can. J. Chem. 1982, 60, 787. (481) Dabrowiak, J. C.; Greenaway, F. T.; Longo, W. E.; van Husen, M.; Crooke, S. T. Biochim. Biophys. Acta

1978, 517, 517. (482) Buncel, E.; Boone, C.; Joly, H. Inorg. Chim. Acta 1986, 125, 167. (483) Shankar Nandi, U.; Wang, J. C.; Davidson, N. Biochemistry 1965, 4, 1687. (484) Kan, L. S.; Li, N. C. J. Am. Chem. Soc. 1970, 92, 4823. (485) Young, P. R.; Shankar Nandi, U.; Kallenbach, N. R. Biochemistry 1982, 21, 62. (486) Beveridge, S. J.; Walker, W. R. Aust. J. Chem. 1974, 27, 2563. (487) Sigel, H. In Metal-DNA Chemistry; T. D. Tullius, Ed.; American Chemical Society: Washington DC, 1989;

pp 159. (488) Frøystein, N. Å.; Sletten, E. J. Am. Chem. Soc. 1994, 116, 3240. (489) Simpson, R. B. J. Am. Chem. Soc. 1964, 86, 2059. (490) Kajiwara, T.; Ito, T. J. Chem. Soc. Chem. Comm. 1994, 1773. (491) Kajiwara, M.; Sugiyama, M.; Kimura, E. J. Chem. Soc. Chem. Comm. 1994, 1747. (492) Chu, G. Y. H.; Mansy, S.; Duncan, R. E.; Tobias, R. S. J. Am. Chem. Soc. 1978, 100, 593. (493) Eichhorn, G. L.; Clark, P. J. Am. Chem. Soc. 1963, 85, 4020. (494) Gruenwedel, D. W.; Cruikshank, M. K. Biochemistry 1990, 29, 2110. (495) Sarker, M.; Chen, F.-M. Biophys. Chem. 1991, 40, 135. (496) Verheyden, J. P. H.; Wagner, D.; Moffatt, J. G. J. Org. Chem. 1971, 36, 250. (497) Aurup, H.; Tuschl, T.; Benseler, F.; Ludwig, J.; Eckstein, F. Nucleic Acids Res. 1994, 22, 20. (498) Wells, P. R. Prog. Phys. Org. Chem. 1968, 6, 111. (499) Bratsch, S. G. J. Chem. Educ. 1985, 62, 101. (500) Sanderson, R. T. J. Chem. Educ. 1988, 65, 112. (501) Bratsch, S. G. J. Chem. Educ. 1988, 65, 223. (502) Marriott, S.; Reynolds, W. F.; Taft, R. W.; Topsom, R. D. J. Am. Chem. Soc. 1984, 49, 959. (503) Mullay, J. J. Am. Chem. Soc. 1984, 106, 5842. (504) Mullay, J. J. Am. Chem. Soc. 1985, 107, 7271. (505) Inamoto, N.; Masuda, S. Chem. Lett. 1982, 1003. (506) Inamoto, N.; Masuda, S. Chem. Lett. 1982, 1007.


158

(507) Inamoto, N.; Masuda, S.; Tori, K.; Yoshimura, Y. Tetrahedron Lett. 1978, 46, 4547.

(508) Inamoto, N.; Masuda, S. Tetrahedron Lett. 1977, 3287.

(509) Bolton, P. H.; Kearns, D. R. Biochim. Biophys. Acta 1978, 517.

(510) Bolton, P. H.; Kearns, D. R. J. Am. Chem. Soc. 1979, 101, 479.

(511) Birnbaum, G. I.; Giziewicz, J.; Huber, C. P.; Shugar, D. J. Am. Chem. Soc. 1976, 98, 4640.

(512) Ducruix, A.; Riche, C.; Pascard, C. Acta Cryst. 1976, B32, 2467.

(513) Sundaralingam, M.; Rao, S. T. Int. J. Quantum Chemistry: Quantum Biology Symposium 1983, 10, 301.

(514) Williams, D. M.; Pieken, W. A.; Eckstein, F. Proc. Natl. Acad. Sci. USA 1992, 89, 918.

(515) Pieken, W. A.; Olsen, D. B.; Benseler, F.; Aurup, H.; Eckstein, F. Science 1991, 253, 314.

(516) Kulinska, K.; Laaksonen, A. J. Biomol. Struct. Dynamics 1994, 11, 1307.

(517) Rosemeyer, H.; Tóth, G.; Golankiewicz, B.; Kazimierczuk, Z.; Bourgeois, W.; Kretschmer, U.; Muth, H.-P.;

Seela, F. J. Org. Chem. 1990, 55, 5784.

(518) Sowers, L. C.; Fazakerley, G. V.; Kim, H.; Dalton, L.; Goodman, M. F. Biochemistry 1986, 25, 3983.

(519) Kaplan, N. O.; Ciotti, M. M.; Stolzenbach, F. E.; Bachur, N. R. J. Am. Chem. Soc. 1955, 77, 813.

(520) Suzuki, S.; Suzuki, K.; Imai, T.; Suziki, N.; Okuda, S. J. Biol. Chem. 1965, 240, 554.

(521) Suzuki, K.; Nakano, H.; Suzuki, S. J. Biol. Chem. 1967, 242, 3319.

(522) Dinglinger, F.; Renz, P. Hoppe-Seyler's Z. physiol. Chem. 1971, 362, 1157.

(523) LePage, G. A.; Junga, I. G. Molec. Pharmacol. 1967, 2, 37.

(524) LePage, G. A. Can. J. Biochem. 1968, 46, 655.

(525) Perry, A.; LePage, G. A. Cancer Res. 1969, 29, 617.

(526) Nakai, Y.; LePage, G. A. Cancer Res. 1972, 32, 2445.

(527) Tamaoki, T.; LePage, G. A. Cancer res. 1975, 35, 1015.

(528) Bobek, M.; Bloch, A. J. Med. Chem. 1975, 18, 784.

(529) Falke, D.; Range, K.; Arendes, J.; Müller, W. E. G. Biochim. Biophys. Acta 1979, 563, 36.

(530) Christensen, L. F.; Broom, A. D.; Robins, M. J.; Bloch, A. J. J. Med. Chem. 1972, 15, 735.

(531) Séquin, U. Experientia 1973, 29, 1059.

(532) Morvan, F.; Rayner, B.; Imbach, J.-L.; Chang, D.-K.; Lown, J. W. Nucleic Acids Res. 1987, 15, 4241.

(533) Morvan, F.; Rayner, B.; Imbach, J.-L.; Lee, M.; Hartley, J. A.; Chang, D.-K.; Lown, J. W. Nucleic Acids Res.

1987, 15, 7027.

(534) Aramini, J. M.; van de Sande, J. H.; Germann, M. W. Biochemistry 1997, 36, 9715.

(535) Townsend, L. B. In Handbook of Biochemistry and Molecular Biology; G. D. Fasman, Ed.; C.R.C. Press:

Cleaveland, Ohio, 1975; Vol. I; pp 271-401.

(536) Hanessian, S.; Pernet, A. G. Adv. Carbohydr. Chem. Biochem. 1976, 33, 111.

(537) Daves, G. D.; Cheng, C. C. Prog. Med. Chem. 1976, 13, 303.

(538) Hacksell, U.; Daves, G. D. Prog. Med. Chem. 1985, 22, 1.

(539) Daves, G. D. Acc. Chem. Res. 1990, 23, 201.

(540) Burchenal, J. H.; Ciovacco, K.; Kalaher, K.; O'Toole, T.; Kiefner, R.; Dowling, M. D.; Chu, C. K.; Watanabe,

K. A.; Wempen, I.; Fox, J. J. Cancer Res. 1976, 36, 1520.

(541) Cortese, R.; Kammen, H. O.; Spengler, S. J.; Ames, B. N. J. Biol. Chem. 1974, 249, 1103.

(542) Samuelsson, T.; Boren, T.; Johansen, T. I.; Lustig, F. J. Biol. Chem. 1988, 263, 13692.

(543) Polak, M.; Mohar, B.; Kobe, J.; Plavec, J. J. Am. Chem. Soc. 1998, 120, 2508.

(544) Davies, D. B.; Sadikot, H. Org. Magn. Reson. 1982, 20, 180.

(545) Lankhorst, P. P.; Haasnoot, C. A. G.; Erkelens, C.; Westerink, H. P.; van der Marel, G. A.; van Boom, J. H.;

Altona, C. Nucleic Acids Res. 1985, 13, 927.

(546) Plavec, J.; Chattopadhyaya, J. Tetrahedron Lett. 1995, 36, 1949.

(547) Dhingra, M. M.; Saran, A. Biopolymers 1982, 21, 859.

(548) Sundaralingam, M. Biopolymers 1969, 7, 821.

(549) Yokohama, S.; Inagaki, F.; Miyazawa, T. Biochemistry 1981, 20, 2981.

(550) Damha, M. J.; Ogilvie, K. K. Biochemistry 1988, 27, 6403.

(551) Blommers, M. J. J.; Nanz, D.; Zerbe, O. J. Biomol. NMR 1994, 4, 595.

(552) Hruska, F. E.; Mak, A.; Singh, H.; Sugar, D. Can. J. Chem. 1973, 51, 1099.

(553) Sarma, R. H.; Mynott, R. J.; Wood, D. J.; Hruska, F. E. J. Am. Chem. Soc. 1973, 95, 6457.

(554) Haasnoot, C. A. G.; de Leeuw, F. A. A. M.; de Leeuw, H. P. M.; Altona, C. Recl. Trav. Chim. Pays-Bas

1979, 98, 576.

(555) Uhlenbeck, O. C. Nature 1987, 328, 596.

(556) Rousse, B.; Puri, N.; Viswanadham, G.; Agback, P.; Glemarec, C.; Sandström, A.; Sund, C.; Chattopadhyaya,

J. Tetrahedron 1994, 50, 1777.

(557) Rousse, B.; Sund, C.; Glemarec, C.; Sandström, A.; Agback, P.; Chattopadhyaya, J. Tetrahedron 1994, 50,

8711.

(558) Ruffner, D. E.; Stormo, G. D.; Uhlenbeck, O. C. Biochemistry 1990, 529, 10695.

(559) Varani, G.; Cheong, C.; Tinoco, I. Biochemistry 1991, 30, 3280.



159

(560) Maltseva, T. V.; Agback, P.; Repkova, M. N.; Venyaminova, A. G.; Ivanova, E. M.; Sandström, A.; Zarytova,

V. F.; Chattopadhyaya, J. Nucleic Acids Res. 1994, 22, 5590.

(561) Stein, C. A.; Subasinghe, C.; Shinozuka, K.; Cohen, J. S. Nucl. Acids Res. 1988, 16, 3209.

(562) Knudsen, H.; Nielsen, P. E. Nucl. Acids Res. 1996, 20, 494.

(563) Chou, S.-H.; Flynn, P.; Reid, B. Biochemistry 1989, 28, 2435.

(564) Katahira, M.; Lee, S. J.; Kobayashi, Y.; Sugeta, H.; Kyogoku, Y.; Iwai, S.; Ohtsuka, E.; Benevides, J. M.;

Thomas, G. J. J. J. Am. Chem. Soc. 1990, 112, 4508.

(565) Wang, A. H.-J.; Fujii, S.; van Boom, J. H.; van der Marel, G. A.; van Boeckel, S. A. A.; Rich, A. Nature

1982, 299, 601.

(566) Damha, M. J.; Wilds, C. J.; Noronha, A.; Brukner, I.; Borkow, G.; Arion, D.; Parniak, M. A. J. Am. Chem.

Soc. 1998, 120, 12976.

(567) Lesnik, E. A.; Guinosso, C. J.; Kawasaki, A. M.; Sasmor, H.; Zounes, M.; Cummins, L. L.; Ecker, D. J.;

Cook, P. D.; Freier, S. M. Biochemistry 1993, 32, 7832.

(568) Puri, N.; Chattopadhyaya, J. Tetrahedron 1999, submitted,

(569) Ding, D.; Gryaznov, S. M.; Lloyd, D. H.; Chandrasekaran, S.; Yao, S.; Ratmeyer, L.; Pan, Y.; Wilson, W. D.

Nucl. Acids Res. 1996, 24, 354.

(570) Gryaznov, S. M.; Lloyd, D. H.; Chen, J.-H.; Schultz, R. G.; DeDionisio, L. A.; Ratmeyer, L.; Wilson, W. D.

Proc. Natl. Acad. Sci. USA 1995, 92, 5798.

(571) Marquez, V. E.; Ezzitouni, A.; Russ, P.; Siddiqui, M. A.; Ford, H. J.; Feldman, R. J.; Mitsuya, H.; George,

C.; Barchi, J. J. J. J. Am. Chem. Soc. 1998, 120, 2780.

(572) Denisov, A. Y.; Bekiroglu, S.; Maltseva, T. V.; Sandström, A.; Altmann, K.-H.; Egli, M.; Chattopadhyaya, J.

J. Biochem. Biophys. Meth. 1998, 16, 569.

(573) Altmann, K.-H.; Kesselring, R.; Francotte, E.; Rihs, G. Tetrahedron Lett. 1994, 35, 2331.

(574) Marquez, V. E.; Siddiqui, M. A.; Ezzitouni, A.; Russ, P.; Wang, J.; Wagner, R. W.; Matteuci, M. D. J. Med.

Chem. 1996, 39, 3739.

(575) Agback, P.; Glemarec, C.; Sandström, A.; Yin, L.; Plavec, J.; Sund, C.; Yamakage, S.-i.; Viswanadham, G.;

Rousse, B.; Puri, N.; Chattopadhyaya, J. In Proceedings of the 8th Conversations; Adenine Press, New York:

1994; pp 293.

(576) Dalluge, J. J.; Hashizume, T.; Sopchik, A. E.; Closkey, J. A.; Davis, D. R. Nucleic Acids Res. 1996, 24, 1073.

(577) Nawrot, B.; Milius, W.; Ejchart, A.; Limmer, S.; Sprinzl, M. Nucl. Acids Res. 1997, 25, 948.

(578) Taiji, M.; Yokohama, S.; Miyazawa, T. J. Biochem. 1985, 98, 1447.

(579) Sprinzl, M.; Kucharzewski, M.; Hobbs, J. B.; Cramer, F. Eur. J. Biochem. 1977, 78, 55.

(580) Janiak, F.; Dell, V. A.; Abrahamson, J. K.; Watson, B. S.; Miller, D. L.; Johnson, A. E. Biochemistry 1990,

29, 4268.

(581) Potts, R. O.; Ford, N. C. J.; Fournier, M. J. Biochemistry 1981, 20, 1653.

(582) Limmer, S.; Vogtherr, M.; Nawrot, B.; Hillenbrand, R.; Sprinzl, M. Angew. Chem. Internat. Edit. 1997, 36,

2485.

(583) Polak, M.; Plavec, J. Eur. J. Inorg. Chem. 1999, 547.

(584) Isaksson, J.; Maltseva, T.; Agback, P.; Kumar, A.; Zamaratski, E.; Chattopadhyaya, J. 1999, Manuscript in

preparation.

(585) Fujiwara Nucl. Acids Res. Symposium Ser. 1997, 37, 51.

(586) Overmars, F. Ph.D. Thesis Thesis, Leiden University, The Netherlands, 1997.

(587) Mc Connell, H. M. J. Chem. Phys. 1958, 28, 430.


160

(10.4) Latest articles (1998 – 2002) on the aspects of Anomeric (AE)

and Gauche (GE) Effects and discussions/comments on these works

(10.4.1) Review: Perrin, C. L. Acc. Chem. Res. 2002, 35, 28.

A detailed discussion on the stabilities of product formed on hydrolysis of five- and seven-

membered cyclic amides. Product studies showed that the substantial cleavage of the bond that is

antiperiplannar to only one lone pair of electrons and syn to the other with transition-state

stabilization of 2 kcal mol-1

.

(10.4.2) NMR titration studies (Perrin, C. L. J. Am. Chem. Soc. 1999, 121, 6911) showed the

absence of reverse AE (see also Kirby, A. J. J. Chem. Soc., Chem. Commun. 1998, 1695, Szarek, A.

J. Org. Chem. 2001, 66, 1097, Pinto, B. M. J. Org. Chem. 2000, 65, 220) in cationic N-

(glycopyranosyl)imidazoles and their tetra-O-acetyl derivatives (ΔΔGβ-α is negative, thereby

showing greater preference for the axial position of a protonated imidazolyl group than of an neutral

group).

(10.4.3) The net effect (both steric and stereoelectronic) of substitution (X, where X = OH, O-alkyl,

O-acetyl and F) in a hexapyranoside (J. Chem. Soc., Perkin Trans. 2, 2002, 337) indicates reduced

reactivity (in terms of the rate of acid-catalysed hydrolysis) at the anomeric center compared to the

parent tetrahydropyranyl acetal.

(10.4.4) Ab initio (MP2/6-31G*) studies of pseudorotation and conformational stabilities of

pyrrolidine (Carballeria, L.; Perez-Juste, I. J. Chem. Soc., Perkin Trans. 2, 1998, 1339) showed that

pseudorotation path is preferred for inter-conversion between the N-H axial and equatorial form

with a barrier ~0.6 kcal mol-1

supporting the experimental microwave spectroscopy results.

(10.4.5) Computational studies (Senyurt, N.; Aviyente, V. J. Chem. Soc., Perkin Trans. 2, 1998,

1463) by ab initio (HF/6-31G*) of AE in 2-[(4-substituted phenyl)seleno]-1,3-dithianes with NO2,

H, CH3, OCH3 and N(CH3)2 as substituents showed the preference for axial conformer with

enhanced electron withdrawing groups. NBO analyses showed that delocalization involving the

sulfur and selenium lone-pairs and the σ*C2-Se plays and important role in stabilizing the axial

conformer.

(10.4.6) NMR and ab initio studies (Serianni, A. S. J. Am. Chem. Soc. 1997, 119, 8933) for 2-

deoxy-β-D-glycero-tetrofuranose showed that it favors the S-form in solution (89% 4T3, 11% E2).

Protonated form showed exclusive preference for S-form (E3) with higher energy barrier than the

neutral form.

(10.4.7) Conformational studies with constrained nucleosides (Lowary, T. L. J. Org. Chem. 2001)

showed that furanose ring conformation in all compounds is locked either into E3 or 0E. It has also

been shown that a nucleoside containing a conformationally locked furanose ring does influence the

conformation of adjacent neighbor.

(10.4.8) Comparison of conformatinal analysis by theoretical studies (at various level with ab initio

and DFT methods) with that of experimentally measured by NMR for Methyl 3-O-Methyl-α-D-

arabinofuranoside (Lowary, T. L. J. Am. Chem. Soc. 2001, 123, 8811) has been performed.

Moreover, computational studies (at HF/6-31G* and B3LYP/6-31G*) of conformational analyses of

Methyl-α-D-arabinofuranoside (Lowary, T. L. J. Am. Chem. Soc. 1999, 121, 9682) showed 3E as the

lowest energy N-type conformer and either 2E or E1 (depending upon the level of theory used) as

lowest energy S-type conformer which is also the global minimum.



161

(10.4.9) Conformational analyses of carba sugar analogues of methyl α-D-arabinofuranosides and

methyl β-D-arabinofuranosides have been reported (Lowary, T. L. J. Org. Chem. 2001, 66, 8961). It

showed that furanose ring conformation in these compounds is biased towards N-type.

(10.4.10) The magnitude of the one-bond coupling constant between C1 and H1 in 2,3-anhydro-O-

furanosides has been shown (Lowary, T. L. J. Org. Chem. 2001, 66, 4549) to be sensitive to the

stereochemistry at the anomric center. A panel of 24 compounds was studied and in case where

anomeric hydrogen is trans to the epoxide moiety, 1JC1-H1 = 163 – 168 Hz; whereas for cis

orientation of this hydrogen with respect to the oxirane ring, 1JC1-H1 = 171 – 174 Hz. However, for

2,3-anhydro-S-glycosides, the 1JC1-H1 is insensitive to the C1 stereochemistry.

(10.4.11) NMR and molecular modeling studies of 8-Aza-3-deazaguanine (Plavec, J. J. Chem. Soc.,

Perkin Trans. 2, 2001, 1433) showed stabilization of N-type conformers by ΔΔHo of 3.1 kJ mol

-1,

which has been attributed to the strengthening of O4'-C1'-N9 anomeric effect (strength of 19.5 kJ

mol-1

).

(10.4.12) NMR and ab initio studies (Plavec, J. J. Chem. Soc., Perkin Trans. 2, 2000, 255) showed

The S-C-N anomeric effect is stronger in purine than in pyrimidine 4'-thionucleosides (thymine <

cytosine < guanine < adenine), which is opposite to natural 4'-oxonucleosides. However, the

strength of AE in 4'-thionucleosides is weaker compared to that in natural nucleosides.

(10.4.13) Steric and stereoelectronic effects of 7- and 8-substituted 7-deaza-2'-deoxy-adenosine and

-guanosine on the sugar pucker as well as conformation about C4'-C5' bond have been studied

(Seela, F. J. Chem. Soc., Perkin Trans. 2, 1997, 2341): (i) higher electron-attracting effect of the

substituents drives N/S equilibrium of the pentofuranose towards N-type conformation and (ii)

higher electron-withdrawing effect of the 7-substituents, higher γ+ across C4'-C5'.

(10.4.14) Conformational analysis of 2-halocyclohexanones by NMR, theoretical and solvation

studies (Yoshinaga, F. J. Chem. Soc., Perkin Trans. 2, 2002, 1494) showed axial conformation of 2-

fluoro compound in vapour phase is stabilized (ΔEeq – ax = 0.45 kcal mol-1

) whereas equatorial

conformer predominates for other haloketones (ΔEeq – ax = 1.05, 1.50 and 1.90 kcal mol-1

for chloro,

bromo and iodo compounds respectively).

(10.4.15) Detailed MO calculations of cyclohexane, 1,3-Dioxane, 1,3-Oxathiane and 1,3-Dithiane

showed (Alabugin, I. V. J. Org. Chem. 2000, 65, 3910) the importance of hyperconjugative

interaction (through the balance of three effects: σC-X → σ*C-Heq, σ*C-Heq → σC-X and nP → σ*C5-Heq

interactions) to explain the relative elongation of equatorial C5-H bonds. The role of nP → σ*C5-Heq

interaction is important in dioxane. In diathiane, distortion of the ring by long C – S bonds increases

preferential overlap of σC5-Heq and σ*C-S orbitals.

(10.4.16) Recently a report has appeared (S.F. Wnuk, D.R. Companioni, V.Neschadimenko, and

M.J. Morris, J. Org. Chem. (Web: http://dx.doi.org/10.1021/jo020428b) on the competition

between steric and -fluorine effects, including the impact of fluorine regio- and stereochemistry, on

radical deoxygenations of 2'(3')-O-phenoxythiocarbonyl (PTC) esters of fluoropentofuranosyl-

adenine nucleosides. Thus it has been found that steric effects are decisive for determination of the

stereoslectivity of transfer of deuterium from tributyltin hydride to fluoropentofuranosyl radicals

generated from these adenine nucleoside derivatives. In all cases, deuterium abstraction occurs at

the less hindered α face of the sugar ring trans to the heterocyclic base. However, this α face

stereoselectivity is enhanced by the anti effect of a vicinal fluorine substituent with an arabino or

xylo orientation (on the α face of the ring). A smaller anti effect is still apparent with a vicinal


162

fluorine on the α face (ribo orientations). Complex stereoelectronic/steric interactions might be

involved with these furanose rings that have electronegative (F, O, N) substituents.

(10.5) The sensitivity of the RNase H discriminates the local structure changes owing to

conformational transmission induced by 3'-endo sugar constrained nucleotides in the

antisense strand of the antisense-RNA hybrid duplex.

RNase H an ubiquitous enzyme cleaves the RNA in a DNA/RNA hybrid. Based on the

dissociation constant determined by competitive inhibition analysis, the RNase H can bind to all

duplexes, irrespective of their conformational preorganization [DNA/RNA > RNA/RNA >

DNA/DNA]. The catalytic cleavage by the enzyme however demands a flexible hybrid duplex

structure whose overall geometry should be close to an A type helix. For example, in the native

DNA-RNA duplex, the conformation of the DNA strand is B type (all nucleotides are in O4’-endo

conformation) whereas the RNA strand adopts a conformation, which is very similar to single

stranded A-RNA type (all nucleotides are in C3’-endo conformation). It has emerged that the RNase

H cleavage site retains this B-type-DNA/A-type RNA conformation in order to be substrate for

cleavage reaction by RNase H. This conformational criteria has been so far difficult to achieve with

the modified residues in the antisense strand at the cleavage site. We have thus witnessed the

emergence of the gapmer or the mixmer strategies, which not only allows the cleavage of the

complementary RNA strand, but also enhances the thermodynamic stability of the hybrid duplex.1

The tolerance of the modified nucleotide residues in the antisense strand in the above two strategies

are very sensitive to the type of modifications used, both because of conformational pre-

organization and the intrinsic flexibility required of the hybrid duplex for catalytic cleavage by

RNase H.

Most of the North conformationally constraind nucleosides upon introduction (full

modification or partial modification) to Antisense OligoNucleotide (AON) render their hybrid

duplex with RNA (AON/RNA) to a rigid RNA/RNA type duplex, which were found to be

insensitive towards the RNase H promoted cleavage, although the stability of the resulting hybrid

duplexes are enhanced1. The reasons for this RNase H insensitivity to conformationally-constrained

containing antisense complex with complementary RNA (both in the gapmer as well as in the

mixmer strategies) is beginning to be understood. Remarkably, some conformational alterations

brought about by the modifications is transmitted to the neighbouring nucleotides, which enzyme

can sense, but spectroscopic techniques such as NMR or CD can not detect2. A slow consensus is

however emerging as to how far the rigidity of the N-type conformationally constrained nucleotides

in the AON can propagate to alter the conformational characters of the neighboring S-type residues

in the hybrid duplex. This information is important because the alteration of S-type character in the

neighbors to an N-type forces those nucleotides in the AON/RNA duplex to adopt RNA/RNA type

conformation, which prevents it from the RNase H cleavage. The uniquness of RNase H promoted

cleavage of AON/RNA hybrids holding varying number of conformationally constrained North-East

(N/E)-type nucleotides, can be employed to address these issue which we have shown recently with

the oxetane incorporated nucleotides in the antisense-RNA hybrid duplex 2.

We have introduced single, double, triple N/E conformationaly constrained nucleoside, [1-

(1',3'-O-anhydro-β-D-psico-furnosyl)thymine] (T), to a 15 mer AON and targeted to a 15 mer RNA.

CD failed to detect any structural perturbances of the modified hybrids compared to the native

counterpart; however, the RNase H cleavage pattern clearly showed that the local conformational

changes spanning a total of 5 nucleotides including the modification towards the 5'-end of AON (3'-

end of RNA). This was evident from the fact that the 5 nucleotide region of the RNA strand,

beginning from the nucleotide opposite to the T modification, became completely inactive to the

catalytic cleavage reaction by RNase H2. Since the site just after the 5 nucleotides were accessible

for enzymatic cleavage (Figure 10), and the fact that the binding and cleavage sites are different for

RNase H3, it is evident that the structurally altered duplex region was suitable for enzyme binding

but not for cleavage. By suitably placing the just three T modifications we have shown that the



163

single cleavage site can be engineered in the 15mer AON/RNA hybrid duplex (Figure 1). This work

clearly shows that owing to the N-type or N/E-type conformational constrain introduced at the

oligonucleotide level, leads to microenvironmental conformational alterations in the neighboring 4

nucleotides toward the 5’-end. This conformational transmission can be effectively mapped by

RNase H, when CD and NMR fails to show any conformational heterogeinity in the local structure.

There is only limited information available regarding the tolerance of RNase H towards local

conformational transmission of AONs holding other conformationally constrained nucleotides. A

recent report4 outlined the RNase H tolerence of AON/RNA hybrids modified with North-

conformationally constrained LNA monomers. Introduction of one LNA itself was found to alter

global helical structure5 (detected by CD) and three LNA monomers were sufficient to create a

RNA/RNA type AON/RNA hybrid duplex. None of these modifications are however reported to

elicit any RNase H response, although considerable enhnacement in the thermodynamic stability

was observed. Mixmer AONs with 3-5 deoxynucleotide gaps were found to be also insensitive

towards RNAse H promoted cleavage once hybridized to RNA4. This clearly shows that

microstructural alterations brought by the LNA modification propogates beyond 5 nucleotids in

contradistinction to the oxetane modifications.

An important question in this context is how far down the polynucleotide chain the

conformational transmission propagates in the AON/RNA hybrid duplex2? Clearly, an appropriate

answer should make it possible to use of the LNAs both as Tm enhancer and a substrate for RNAse

H, as it has been achieved in a recent systematic study with a series of LNA incorporated AON-

RNA duplexes in which the size of the deoxynucleotide gap (required for RNase H cleavage) in the

mixmer as well as in the gapmer has been varied5. This study has shown that the gapmer/mixmer of

LNA/RNA duplexes having 6 - 10 deoxynucleotide gaps was cleaved by RNase H. Quite expectedly2,

it was also found that the maximum cleavage efficiency was observed as the size of the

deoxynucleotide gap was increased to 10, presumably because it adopted a substantial degree of

DNA/RNA type structure5. This shows that the conformational transmission of LNA stretches upto 7

nucleotides to make it conform to RNA/RNA type structure, which is resistant to RNase H, whereas

for oxetane modifications, we have shown it stretches upto 5 nucleotides2

2’-O-methylnucleosides (3J1’2’ = 5.2 Hz,

3J3’4’ = 4.7 Hz,) show a slightly preferred N-type

conformation in the North-South pseudorotational equilibrium. This slight preference for N-type

conformation is attributed to the steric repulsion between the aglycone and the 2’-O-methyl group in

the C2’-endo (South form)6. In the oligomer there would be additional steric clash of 2’-O-alkyl

group with the 3’-phosphate, which drives it to more North conformation. Therefore, unlike the

other North-constrained nucleotides (LNA, oxetane), the conformational influence by 2’-O-alkyl

nucleotides, which can be detected by RNase H, to the neighbours are less pronounced, and it was

found to be sequence dependent7. This is clearly visible in the RNase H digestion pattern of 2’-O-

alkyl modified AON/RNA gapmer hybrids, where the microstructural alterations brought by the

modification, which blocks the cleavage was found to vary from 3 to 5 nucleotides including the

modification7. This means that the conformational transmission of the North-type 2’-O-

alkylnucleotides in the AON/RNA/RNase H ternary complex is less pronounced than in other

conformational constrained counterparts. Unlike the North constrained nucleotides, the

conformational status of 2’-O-methylnucleosides and its analogs in the antisense strand or in the

hybrid duplex with RNA can be envisioned as sensitive to both sequence make-up as well as

whether they are in the complex form with RNase H or not. It has been recently shown that the

modified AONs containing acyclic interresidue units (the 2’,3’-secouridine or a butanediol linker in the

modified AONs containing 2’F-ANA) supports RNase H-promoted cleavage of complementary RNA

(Damha et al., J. Am. Chem. Soc. 125, 654-661 (2003). These conformationally labile AONs shows

some remarkable benefits in the enzymatic recognition of the AON/RNA hybrids, because of the

flexible nature of their sugar-phosphate backbone conformation, which is consistent with our

observation of the oxetane-modified AON/RNA recognition and cleavage by RNase H. It should be

noted that the introduction of a 2’,3’-secouridine moitey reduces the Tm by ca 10ºC whereas a butanediol

linker reduces the Tm by ca 6ºC.


164

(A) AON (1) AON (2) AON (4) AON (5)AON (3) AON (6)

5'-r(G A A G A A A A A A U G A A G)-3'3'-d(C T T C T T T T T T A C T T C)-5'

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

11

1213


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1112


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

12


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

8


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

8

7

AON (2)

AON (3)

AON (4)

AON (5)

AON (6)

8 1011

12139765

AON (1)

9

10

7

Figure 10. (A) The PAGE analysis of RNase H hydrolysis of the hybrid duplexes, AON (1)-

(6) / RNA. Time after the

addition of the enzyme is shown on the top of each gel lane. The length and sequence of 5’-32

P-labeled RNAs formed up

on enzymatic cleavage are shown on the left and right side of the gel and it was deduced by comparing the migration of

products with those oligonucleotides generated by parial digetion of the target RNA by Snake venom phosphodiesterase

(SVPDE). (B) RNase H cleavage cleavage pattern of the hybrid duplexexs. Long and short arrows represents major and

minor sits respectively (after 2h of incubation). Boxes represents the parts of the RNA sequence insensitive towards

RNase H cleavage.

References:

1. (a) M. Manoharan, Biochim Biophys Acta., 1489, 117(1999). (b) P. Herdewijn, Biochim

Biophys Acta., 1489, 167(1999).

2. (a) P. I.Pradeepkumar, E. Zamaratski, A. Földesi, J. Chattopadhyaya. Tetrahedron Lett., 41,

8601 (2000). (b) P. I. Pradeepkumar, E. Zamaratski, A. Földesi, J. Chattopadhyaya. J. Chem.

(B)



165

Soc., Perkin Trans. 2, 3, 402 (2001). (c) P. I. Pradeepkumar, J. Chattopadhyaya. J. Chem. Soc.,

Perkin Trans. 2, 11, 2074 (2001).

3. S. Kanaya, Enzymatic activity and protein stability of E. coli ribonuclease HI, in

Ribonucleases H, Crouch, R. J., editors; INSERM: Paris, Chapter 1, pp. 1-37 (1998).

4. J. Kurreck, E. Wyszko, C. Gillen, V. A. Erdmann, Nucleic Acids Res., 30, 1911(2002).

5. K.Bodensgaard, M. Petersen, S.K. Singh, V.K. Rajwanshi, R. Kumar, J. Wengel, J.P.

Jacobsen. Chem. Eur. J, 6, 2687 (2000).

6. G. Kawai, Y. Yamamoto, T. amimura, T. Masegi, M. Sekine, T. Hata, T. Iimori,

T.Watanabe, T. Miyazawa, S. Yokoyama, Biochemistry, 31, 1040 (1992).

7. (a) H. Inoue, Y. Hayase, S. Iwai, E. Ohtuka., FEBS Lett., 215, 327 (1987). (b) Y-T.Yu, M-D.

Shu, J. A. Steiz, RNA., 3, 324 (1997). (c) B.Larrouy, C. Boiziau, B.Saproat, J-J.Toulme,

Nucleic Acids Res., 23, 3434 (1995).

(10.6) The influence of flouro substitution at the sugar moiety in modulating the furanose

conformation of flourinated nucleosides.

The furanose conformation of a nucleoside is determined by interplay of stereoelectronic

gauche and anomeric effects (vide infra) tuned by different sugar substituents. The strength of the

gauche and anomeric effects is determined by the electronic nature of the nucleobase, and the

electronegetivity, position, configuration of substituents in the furanose moiety. Thus highly

electronegative flouro substitution in the sugar moiety can predominantly drive its overall North

(N)� South (S) equilibrium. The various flourinated nucleosides have therefore been designed and

analysed for their ability to conformationally preorganize the furanosyl moiety in solution.

Marquez et al. have performed a systematic study of various fluorinated dideoxy uridines by

NMR and pesudorotational analysisis1. In 2'-flouro-2', 3'-didexy-ara-uridine (2'F-up, 1) and 3'-

flouro-2', 3'-didexy uridine (3'F-down, 4) the F-gauche effect dominates over the opposing O4'-C1

'-

N anomeric effect, thereby the furanose ring adopts the South.type conformation in solution.

However in the 2'F-down (2) and 3'F–up (3) dideoxy uridines, the F-gauche effect along with the

anomeric effect plays in tandem to lock the sugar conformation in to the Nothern hemisphere (N-

type) of the pseudorotational cycle. It is worth noting that the 2'F-down dideoxy uridine (2)

exclusively prefers the N-type conformation at 25 oC with a pseudorotatonal phase angle of 19

o.

In the case of the 2'-flouro-2'-deoxyarabinoribofuranosyl thymine (2'-F up, 5) the gauche

effect of O4'-C1

'-C2'-F and F-C2

'-C1

'-N fragments prevails over the anomeric effect and the sugar

adopts the C2'-endo/ C1'-exo conformation in solution.

2,3 The 2

'-flouro-2'-deoxyribofuranosyl

thymine (2'F-down, 6) prefers the N-type sugar conformation due to the combined influence of

gauche and anomeric effects. However, in the X-ray and NMR structure of the oligonucleotide

hybrid duplexes containing the 2'F-up thymidine units showed the O4'-endo conformation for the

furanose ring.4-6

This is explained on the basis of steric clash between the 2'F atom and the C6-

carbon in the South sugar.


166

O

H3'' H''

HO

NH

N

O

O

FH3'

O

H3'' F

HO

NH

N

O

O

H3'H2'

O

H3'' H2''

HO

NH

N

O

O

FH

'O

F H2''

HO

NH

N

O

O

H3'H2'

J3'-4' = 5.2 Hz

J3''-4' = 8.6 Hz

J1'-2'' = 3.3Hz

P = 131o

South -type

J3'-4' = 11.3 Hz

J3''-4' = 4.8 Hz

J1'-2' = 0.9 Hz

P = 19o

North -type

J3''-4' = 2.4 Hz

J1'-2' = 2.1 Hz

J1'-2'' = 8.2 Hz

P = 27o

North -type

J3'-4' = 1.15 Hz

J1'-2' = 9.0 Hz

J1'-2'' = 5.7 Hz

P = 169o

South -type

Barchi, J.J.; Jeong, L.K.; Siddiqui, M.A.; Marquez, V.E. J. Biochem. Biophys. Methods.1997, 34, 11.

2', 3' monoflorinated dideoxyuridines

1 2 34

O

HO H2''

HO

NH

N

O

O

F

J3'-4' = 4.5 Hz

J1'-2'' = 4 Hz

South -type

5

O

HO F

HO

NH

N

O

O

J3'-4' = 11.8 Hz

North -type

6

2' monoflorinated thymidines

O

HO

HO

F

N

NN

N

NH2

O

HO F

HO

N

NN

N

NH2

O

H3'' F

HO

N

NN

N

NH2

O

H3''

HO

F

N

NN

N

NH2

H3'

Ikeda, H.; Fernandez, R.; Wilk, A.; Barchi, J.J.;

Huang, X.; Marquez, V.E. Nucleic. Acids. Res.

1998, 26, 2237.

Wilds, J.C.; Damha, M.J. Nucleic. Acids. Res.

2000, 28, 3625.

H3'

South -type

P = 130o

North -type

J1'-2' = 2.82 Hz

P = 0o

South -type North -type

J1'-2' = 3.3Hz

P = 130o

J1'-2' = 0 Hz

P = 0o

2' monoflorinated deoxy and dideoxy adenosines

Ford, H.; Lan Mu, F.D.; Siddiqui, M.A.; Nicklaus, M.C.; Anderson, L.; Marquez, V.E.; Barchi, J.J.

Biochemistry. 2000, 39, 2581.

O

HO

HO

F

6 7 89

10

NH

N

N

O

NH2N

Br

O

HO

HO

F

N

N

N

NH2

NH2N

Br

11

2' monoflorinated deoxy 3-bromopyrazolo[3,4,-d] pyramidines

J3'-4' = 7.2 Hz

J1'-2'' = 6.3 Hz

J3'-4' = 7.5 Hz

J1'-2'' = 5.98 Hz

North -type North -type

P = -2.1o

P = -2.1o

He, J.; Mikhailopulo, I.; Seela, F. J. Org. Chem. 2003, 68, 5519.

To unravel the structural preference of North-constrained deoxy and dideoxy 2'F-adenosine

at the active site of adenosine deaminase, a detailed conformational analysis of various fluorinated

adenosine analogues has been reported recently.7 In the case of 2'-flouro-2',3'-



167

dideoxyarabinofuranosyl adenosine (F-up, 8) and in the corresponding ribo analogue (F-down, 9)

only O4'-C1

'-C2'-F is operative along with the anomeric effect. In 8 this F-.gauche effect

predominates over the anomeric effect and the sugar adopts the C1'-exo conformation (P =130o,

81% S-type). The assistance of anomeric effect in 9 locks the sugar conformation in to an exclusive

N-type (P= 0o, 99% N-type). The situation is more complex in the case of 2'-flouro -2'-

deoxyarabinofuranosyl adenosine (6, F-up) and 2'-flouro -2'-deoxyribofuranosyl adenosine (7, F-

down) owing to the presence of 3'-OH. Here the additional O4'-C4

'-C3'-O3' gauche effect plays a

crucial role in dictating the sugar conformation. In 2'F-down compound 6 the 2'F -gauche effect and

anomeric effect overcome the O4'-C4

'-C3'-O3' gauche effect and the major pesudorotomer adapts

the N-type conformation with 76% population. While in 2'F-up compound 7 the sugar adapts

predominantly the C1'-exo pseudorotomer owing to the dominal influence of gauche over anomeric

effect (P =130o, % of S-type = 64%)

Sofar the studies on the stereoelectronic effects in nucleosides and nucleotides showed that 2' (or

2") and/or 3' (or 3") F/O mediated GE is stronger than AE. A recent report from Seela's group showed,8

for the first time, that the AE can be stronger than the F/O mediated GE in dictating the furanose

conformation of the nucleosides. Thus, they have shown that they could overcome the strong GE of 2'-F

substituent in the β-face (ara configuration) by the introduction of the extra nitrogen atom at the 6 position

of a pyrimidine moity (6-azauracil-1-yl moiety) or at the 8 position of the purine moiety [3,4-d]pyrimidine-

1-yl aglycone (3-bromo or 3-H gave almost identical N/S population). The two nucleosides, 3-

bromopyrazolo [3,4-d]pyrimidine-2'-deoxy-2"-flouro and its ara counterpart 10 and 11 showed exclusive

N-type conformation in solution and solid state (with pseudorotational phase angle –2.10),8 which is

unexpected in view of the earlier studies with 2'(β)-flouro substituted nucleosides. The strong electron

wihdrawing nature of nucleobase in these nuclosides enhances the nO4' → σ*C1'-N1 interaction

(anomeric effect) and it prevails over the gauche interactions results in the existence of C3'-endo

(N-type) pseudorotomer in solution.

References:

1. Barchi, J.J.; Jeong, L.K.; Siddiqui, M.A.; Marquez, V.E. J. Biochem. Biophys.

Methods.1997, 34: 11.

2. Ikeda, H.; Fernandez, R.; Wilk, A.; Barchi, J.J.; Huang, X.; Marquez, V.E. Nucleic. Acids.

Res.1998, 26, 2237.

3. Wilds, J.C.; Damha, M.J. Nucleic. Acids. Res.2000, 28, 3625.

4. Berger, I.; Tereshko, V.; Ikeda, H.; Marquez, V.E.; Egli, M. Nucleic. Acids. Res.1998, 26,

2473.

5. Trempe. J.F.; Wilds, C.J.; Denisov, A.Y.; Pon, R.T.; Damha, M.J.; Gehring, K. J. Am.

Chem. Soc.2001, 123, 4896.

6. Denisov, A.U.; Noronha, A.M.; Wilds, C.J. Trempe, J-F.; Pon, R.T.; Gehring, K.; Damha,

M.J. Nucleic. Acids. Res. 2001, 29, 4284.

7. Ford, H.; Lan Mu, F.D.; Siddiqui, M.A.; Nicklaus, M.C.; Anderson, L.; Marquez, V.E.;

Barchi, J.J. Biochemistry. 2000, 39, 2581.

8. He, J.; Mikhailopulo, I.; Seela, F. J. Org. Chem. 2003, 68, 5519.

stereoelectronic effects in nucleosides and nucleotides and their

Documents