molecular properties of a representative glycine-rich sequence of elastin – bocvggvgoet: a...

www.elsevier.com/locate/theochem

Journal of Molecular Structure: THEOCHEM 812 (2007) 25–37

Molecular properties of a representative glycine-rich sequenceof elastin – BocVGGVGOEt: A combined FTIR experimental

and quantum chemical investigation

Giuseppe Lanza *, Anna M. Salvi, Antonio M. Tamburro

Dipartimento di Chimica, Universita della Basilicata, via Nazario Sauro 85, 85100 Potenza, Italy

Received 25 November 2006; received in revised form 8 February 2007; accepted 9 February 2007Available online 20 February 2007

Abstract

The intriguing relationship between molecular geometry and the vibrational frequencies of the protected BocValGlyGlyValGlyOEtpeptide has been studied by density functional calculations and FTIR spectroscopy. B3LYP/6-31G* data show several folded structuresstabilised by labile intramolecular hydrogen bonds, lying very close in energy. The great fluxionality of the peptide chain has beenascribed to the low steric encumbrance of the substituents in the Ca of glycine residues. In any case, the most stable conformationsinvolve large pseudocycles (C13 or C14), H-bonded with the carbonyl of terminal groups (urethane or ester). FTIR spectra of theNH and ND isotopomers of the BocValGlyGlyValGlyOEt peptide recorded in the solid state and in solution of chloroform, trifluoro-ethanol (TFE), or dimethyl-sulfoxide (DMSO) have been interpreted using B3LYP data as well as FTIR data of closely related mole-cules. The frequency and intensity of normal modes associated with mC@O and mN–H stretching are modulated by hydrogen bondingthus providing a direct correspondence with the given structures. FTIR data in the solid state support a single folded structure stabilisedby intramolecular hydrogen bonding and by electrostatic interactions with surrounding peptides. Experimental data in CHCl3 and TFEsolutions also support intramolecular hydrogen bonding. However, the concomitant presence of various folded structures in rapid equi-librium hampers a good separation of bands. Conversely, the substantial hydrogen bonding between >N–H groups of the peptide andDMSO solvent molecules seems to favour an extended structure.� 2007 Elsevier B.V. All rights reserved.

Keywords: Peptide conformation; Hydrogen bond; DFT; Vibrational frequencies; FTIR

1. Introduction

Considerable scientific interest in polypeptides is cur-rently fuelled by the desire to rationalise the conformationand function of the parent proteins. Peptides show rich var-iegate structural features (extended, a-helices, 310-helices,b-turns, etc.) depending on the nature of their amino acidcomponents. Almost all spectroscopies have been largelyapplied to gain information about chain conformations,both in solution and solid state, and thus a considerablebody of empirical structure-spectroscopic parameters cor-

0166-1280/$ - see front matter � 2007 Elsevier B.V. All rights reserved.

doi:10.1016/j.theochem.2007.02.019

* Corresponding author. Tel.: +39 097 120 2226; fax: +39 097 120 2226.E-mail address: [email protected] (G. Lanza).

relation has been established [1–7]. Nevertheless, anexhaustive understanding of the relationship between thestructure and spectroscopic properties is far from beingaccomplished. In this regard, electronic structure studiescan be of great relevance to the structural biochemistrycommunity in providing complementary and independentinformation on stability, geometry, and several spectro-scopic parameters.

Polypeptide sequences of the type ‘‘YGlyGlyZGly’’(Y,Z = Val, Leu, or Ala) frequently repeat themselves inelastin, the protein that provides elasticity in vertebrates[8]. These peptides exhibit conformations that are stronglydependent on the physical state and on the nature of thesolvent in which they are dissolved [9–11]. Indeed, variouskinds of b-turns, half turns, polyprolines II, and extended

mailto:[email protected]

26 G. Lanza et al. / Journal of Molecular Structure: THEOCHEM 812 (2007) 25–37

structures have been proposed. The large variety of possi-ble structures prompted us to investigate molecular proper-ties of the representative BocValGlyGlyValGlyOEt peptideby a combined theoretical quantum chemical and experi-mental FTIR approach. Particular attention has beendevoted to subtle details of hydrogen bond networks,which often determine conformation preference [12]. Tothis purpose, the geometry of a large number of conformershas been determined at the DFT-B3LYP level and com-puted vibrational frequencies have been compared withvarious experimental spectra recorded in the solid stateand in different kind of solvents. This work builds uponab initio calculations of the electronic structure of BocVal-GlyGlyValGlyOEt peptide combined with X-ray photo-electron spectroscopic (XPS) measurements (Fig. 1) [13,14].

2. Methods

2.1. Synthesis and FT-IR spectra

The synthesis of the protected peptide BocVGGVGOEtwas described elsewhere [13]. The peptide was dissolved indeuterated water/DMSO solution, lyophilised and thendried in order to prepare the N-D isotopomer (BocVGGV-GOEt-d5).

IR spectra were recorded on a Jasco FT/IR 460 spectro-photometer, using a resolution of 2 cm�1 and 200 scans.The spectra were taken at room temperature in the solidphase (KBr pellet) and in CHCl3, TFE and DMSO solutions(deuterated solvents were used for BocVGGVGOEt-d5)using a cell with either CaF2 or NaCl windows. The sampleconcentration was varied from 0.1% to 0.02% and no aggre-gation effects were observed. Because of the very low solubil-ity in chloroform the spectra were recorded in a saturatedsolution using a 1 mm pathlength cell.

2.2. Computational details

Computations were performed at the density functionallevel employing the hybrid B3LYP functional and the6-31G* basis set for all atoms. According to recent studieson various peptides, the B3LYP/6-31G* electronic

Fig. 1. Molecular structure of the extended, Z conformer of B

structure approach is sufficient for taking into accountstructural, energetic and vibrational frequencies [15,16].

The geometries of all the structures involved were fullyoptimised using gradient techniques. The nature of the moststable stationary points was determined by evaluations of ahessian matrix and of the associated harmonic vibrationalfrequencies. The vibrational frequencies computed at theB3LYP/6-31G* level allows an exhaustive rationalisationof experimental FTIR spectra. All the calculations were per-formed using the Gaussian-03 program [17].

3. Results and discussion

3.1. Geometrical structure of BocVGGVGOEt

The BocVGGVGOEt molecule can adopt several con-formations, depending on the formation of intramolecularhydrogen bonds between various >NH protons and car-bonyl oxygens belonging to two distinct amino acid resi-dues of the polypeptide backbone [12]. In the simplestcase, the molecular chain assumes a linear disposition(extended, Z) with all torsion bond angles (/i, wi, and xi)near to 180�. Alternatively, the BocVGGVGOEt moleculecan adopt different types of helices (Ra, La, d, 310, etc.) andseveral types of b-turns depending on the /i and wi torsionangles. These conformers are all potential candidates asground state of the BocVGGVGOEt molecule. However,a full characterisation of all minima on the Born–Oppen-heimer surface is beyond the scope of the present investiga-tion. Therefore, we limited our analysis to chemicallysignificant cases:

(i) the right- and left-handed a-helices (Ra and La);(ii) the types I and II b-turn conformations having V1–G2

at the corners;(iii) the types I and II b-turn conformations having G2–G3

at the corners;(iv) the b-turn types I and II conformations having G3–V4

at the corners.

Types I and II b-turn conformations defined on theportion V4–G5 cannot produce hydrogen bond enclosure

ocVGGVGOEt peptide and definition of dihedral angles.

G. Lanza et al. / Journal of Molecular Structure: THEOCHEM 812 (2007) 25–37 27

larger than C7 cycles, therefore, they were not examined.Because of several possible minima, the final structureachieved in the geometry optimisation depends on thestarting conformation, therefore, the choice of initial /i,and wi sets is a critical point. Furthermore, in some casesthe final geometry differs substantially from the initialone and stable unconventional conformations are reached.

Starting with standard torsion bond angles for the aboveconsidered conformations, full geometry optimisationshave been performed and results are reported in Figs. 2and 3, Table 1 and Supplementary materials. For all cases,

Fig. 2. Optimised structures of the BocVGGVGOEt peptide for the 310 helix (>NH� � �OC< bond distance (A) and the related N–H–O bond angle (deg) of t

the peptide bonds are almost planar, within about 15�because of their partial double bond character. For thisreason the xi torsion angles have not been further scrutin-ised and we have devoted our attention only to the (/i,wi)dihedral space.

Allowing the backbone angles to relax from the fully-extended structure (/i = wi = xi = 180�) the linearity ofthe chain is maintained after geometry optimisation. The/i and wi rotational parameters of the glycine residuesundergo to minor changes (<3.1�) while a consistent rear-rangement in the Ca of valine residues occurs in order to

A), b-turn I 0 V1–G2 (B), b-turn I V1–G2 (C) and b-turn II V1–G2 (D). Thehe atoms involved in hydrogen bonding are also reported.

Fig. 3. Optimised structures of the BocVGGVGOEt peptide for the b-turn I G2–G3 (E), b-turn I G3–V4 (G), C13 (I), and inverted-C14 (J) conformations.The b-turn II G2–G3 (F) and b-turn II G3–V4 (H) conformations are reported in Figure S1. The >NH� � �OC< bond distance (A) and the related N–H–Obond angle (deg) of the atoms involved in hydrogen bonding are also reported.


minimise backbone-sidechain repulsive interactions (Table1).

The initial structure of standard Ra-helix conformationis characterised by two C13-pseudocycles with the forma-tion of two hydrogen bonds involving Boc-V4 and V1–G5

residues. Nevertheless, after geometry optimisation themolecular shape changes substantially, thus leading to theformation of three interconnected C10 pseudocyclesinvolving hydrogen bonds between the Boc-G3, V1–V4

and G2–G5 residues (structure A in Fig. 2). This structureis better described as a slightly distorted 310 helix with /i

and wi backbone angles close to standard ones (�60� and�30�, respectively) [1].

Analogously, geometry optimisation starting from theLa-helix conformation leads to the formation of an addi-tional C10 pseudocycle inside one of the two C13. Thusthe CO group of the urethane is involved in a bifurcatedhydrogen-bond (structure B in Fig. 2). The >CO� � �HN<distances suggest that the hydrogen bond formed withthe C10 pseudocycle is stronger than the C13 one and the

structure is better described as b-turn. The (/1,w1) and(/2,w2) torsion angles are close to the standard values ofthe b-turn I 0 ((60�, 30�) and (90�, 0�)) hereafter this con-former will be indicated as b-turn I 0 V1–G2.

The initial geometry of the b-turn I V1–G2 conformerhas been adapted from the experimental X-ray structureof the BocVGGOH tripeptide which shows an intramolec-ular hydrogen bond between carbonyl oxygen of the ure-thane group and the amide hydrogen of the G3 residue[18]. The final optimised structure (structure C in Fig. 2)maintains the initial pseudocycle C10, however, two addi-tional C7 pseudocycles (c-turns) involving G2–V4 andG3–G5 residues are formed. The couple of (/3,w3) and(/4,w4) torsion angles (Table 1), related to the c-turncycles, are in full agreement with standard values forinverse c-turn (�75�, 75�) [1]. It is interesting to note thegood agreement of computed and experimental dihedralangles (Table 1) for the BocVGG- b-turn I framework.This confirms the reliability of B3LYP/6-31G* level inderiving peptide geometrical structures.

Table 1Optimised torsion angles (deg) and relative stability (kcal/mol) for the selected BocVGGVGOEt conformers

Rotationalparameters

Extendeda 310 helixAb

V1-G2 G2-G3 G3-V4

b-turn I0

Bcb-turn I Cd b-turn II

Deb-turn IEf

b-turn IIFe

b-turn IGf

b-turn IIHe

C13 (I) Inverted-C14 (J)

DE 0.0 1.0 9.1 0.02 7.2 1.7 1.5 �1.5 3.4 �2.8 �3.1/1 130.3 �66.6 53.0 �77.5 [�61.4] �56.9 �82.4 �81.0 �128.6 �121.3 �79.2 �102.9w1 162.5 �31.9 44.9 �11.9 [�31.8] 133.6 80.8 84.8 153.9 150.9 �12.5 126.3x1 178.6 178.6 173.6 170.8 [�177.9] �171.8 �162.8 �175.1 177.7 175.0 165.6 173.5/2 177.5 �61.2 65.2 �87.4 [�91.3] 99.5 �74.9 �55.4 179.6 �147.5 �77.2 �111.7w2 177.5 �18.7 �1.4 �1.1 [11.5] �18.9 �4.4 133.9 �160.9 �153.3 72.6 �117.5x2 �179.4 175.9 �169.5 �178.3 [179.7] �173.4 168.9 �172.4 �169.0 179.3 �174.6 174.9/3 176.9 �71.6 104.7 �81.4 �81.2 �76.2 108.9 �72.0 �70.2 139.1 �78.8w3 177.3 �3.7 55.5 53.0 65.6 �13.1 �31.2 �12.4 106.2 �61.5 75.0x3 169.2 170.6 �176.1 �178.5 �177.3 176.3 �178.7 177.0 �178.5 174.2 �173.2/4 �133.9 �81.2 55.4 �86.2 �84.4 �84.2 �86.3 �110.4 47.0 �88.8 69.8w4 162.8 �21.1 38.8 78.8 76.9 79.3 81.9 15.8 43.3 78.4 �66.6x4 178.9 172.4 177.2 �175.7 �176.4 �173.7 �176.5 176.8 176.6 �168.6 170/5 177.2 �76.2 96.5 �166.5 �165.2 �164.4 169.2 116.0 �165.8 �176.0 �71.2w5 �179.8 �17.7 8.0 175.7 175.5 175.8 176.4 �153.5 �169.5 173.8 172.9

a The rotational parameters of the fully extended structure (/i = wi = xi = 180�) were adopted in starting geometry.b Standard /i (�63�), wi (�49�), and xi (180�) dihedral angles of right-handed a-helix were adopted in starting geometry.c Standard /i (60�), wi (52�), and xi (180�) dihedral angles of left-handed a-helix were adopted in starting geometry.d Starting structure was adapted from experimental X-ray geometry of BocVGGOH (in square brackets) [18].e For all the b-turn II conformers the initial geometry were adapted from the corresponding b-turn I optimised structure.f In starting geometry the standard (/2,w2), (/3,w3) and (/3,w3), (/4,w4) torsion angles were adopted for b-turn I (G2–G3) and b-turn I (G3–V4)

structures, respectively. All remaining /i, wi and xi were place to be 180�.

Boc V1 -G2

G3

V4

+G5EtO


The initial geometry of the b-turn II V1–G2 conformerhas been adapted from the optimised b-turn I V1–G2 con-former with appropriate (/1,w1) and (/2,w2) angles. Aftergeometry optimisation (structure D in Fig. 2) the C10 (Boc-G3) and the two C7 (G2–V4 and G3–G5) hydrogen bondingenclosures persist and the (/1,w1) and (/2,w2) torsionangles are close enough to ideal b-turn II values((�60�, 120�) and (80�, 0�)).

The b-turn I G2–G3 conformation shows a C10 and twoC7 hydrogen bonds enclosures involving the V1–V4, Boc-G2 and G3–G5 residues, respectively (structure E inFig. 3). The (/2,w2) and (/3,w3) rotational parametersare compatible with those expected for the b-turn I stan-dard values ((�60�,�30�) and (�90�, 0�)) [12]. Also the(/1,w1) and (/4,w4) torsion angles are close to thoseexpected for the inverse c-turn conformation (�75�, 75�).

In analogy to the b-turn I G2–G3 conformation, thatwas used to derive the initial geometry, the b-turn II G2–G3 final structure is characterised by two inverse c-turnlinkages involving the Boc-G2 and G3–G5 residues mergedwith the main b-turn II V1–V4 hydrogen bonding enclosure(structure F in Figure S1).

Upon folding the BocVGGVGOEt at the G3–V4 cornerswith standard b-turn I and II angles, energy minimisationresults in a C10 pseudocycle with the formation of a hydro-gen bond between the >CO and >NH groups of G2 and G5

residues, respectively (structures G in Fig. 3 and H in Fig-ure S1). In both cases the C10 enclosure is accompanied bystronger inverted-C14 hydrogen bonding involving the car-bonyl of the terminal ester group and the amide hydrogen

of the G2 unit. The (/3,w3) and (/4,w4) torsion angles inboth b-turn I and II cases are compatible with standardvalues. The overall structures seem to be b-hairpins. Infact, two adjacent antiparallel strands are linked by a shortloop as in the following structure:

During various attempts, carried out to get the above-mentioned structures, many others stationary points havebeen found. Generally they are located at higher energy,however, there are two geometries more stable than allstructures above discussed. These structures (I and J inFig. 3 and Table 1) are characterised by a strong hydrogenbonding involving urethane and ester >CO and >NHgroup of the V4 or G2 residues, respectively, thus leadingto the formation of C13 and inverted-C14 pseudocycles.For both structures two secondary C7 c-turn linkagesinvolving V1–G3/G3–G5 or G2–V4/G3–G5 residues alsoaccompany the main hydrogen bonding enclosure. Rota-tional parameters of both inverted-C14 and C13 structuresare not coherent with those of any type of b-turn becauseof the concomitance of the others c-turn folding ‘‘inside’’the main pseudocycle. In fact, the torsion angles (/3,w3)and (/4,w4) of the inverted-C14 isomer are very close to


those expected for inverse and classic c-turns, respectively,while the (/2,w2) and the (/4,w4) dihedral of the C13 iso-mer indicate two inverse c-turns.

The absence of few sterically demanding groups on theCa of glycines, allows sidechain–sidechain and sidechain–backbone repulsive interactions minimisation for a largevariety of conformations. The stabilities of different con-formers are in the range of 10 kcal/mol and the five moststable conformers are within 3 kcal/mol. In addition, thelarge C13 or C14 enclosures minimise the ‘‘ring strain’’and pose these two conformers (I and J structures) as thebest candidates for the ground state of the BocVGGV-GOEt peptide. At room temperature (298 K) the energydifference between the inverted-C14 and C13 conformers(about 0.3 kcal/mol) is too small to establish a conforma-tion preference (Boltzmann population is 1:0.60) and ingas-phase or in solutions of low dielectric and weakly coor-dinating solvents both structures should be in equilibrium.

Classical b-turns I or II, 310 helix, hairpins, andextended structures have quite similar stabilities eventhough they differ substantially in the shape. They lie fewkcal/mol above the C13- and C14-type conformers andsuggest an extraordinary fluxionality of the peptide chain.Even though these structures are discarded as the groundstate they might facilitates the C13- and C14-typeinterconversion.

Present conclusions on the BocVGGVGOEt peptide arein general agreement with the available structural data ofrelated peptides.

Firstly, the involvement of the >CO of the protectivegroups in forming strong hydrogen bonding has beenalready observed for the tripeptide BocVGGOH (the mostclose X-ray structure available at the present) [18].

Secondly, the formation of b-turn-like structures is theleading factor in conformation preference for this class ofpeptide while c-turns are relegated to a secondary role.The BocVGGOH molecule shows a type I b-turn confor-mation resulting from the interaction between CO of theurethane group and the amide hydrogen of the G3 residue.This 4 fi 1 b-turn cannot occur in the BocVGOEt dipep-tide (only 3 fi 1 c-turns interactions are possible) and inthe solid state this molecule assumes an S-shape conforma-tion with no intramolecular hydrogen bond enclosures [18].Analogously, the solid state of the free peptide HVGGOHexhibits an extended b-sheet arrangement (the 4 fi 1b-turns interactions cannot occur) [19] while NMR, CDand theoretical data on the free pentapeptideHVGGVGOH indicates the formation of stable 4 fi 1hydrogen bonding [10,12]. Furthermore, CD and 1HNMR data of a large series of protected tetra- and penta-peptides including BocGLGGVOMe, BocGVGGLOMe,BocGVGGOEt, and BocVGGVOMe, give evidence of flex-ible b-turns as the dominant feature whose stability isfound to decrease by increasing the number of repetitiveunits [9].

Third, the extraordinary fluxionality of the BocVGGV-GOEt peptide, due to the three few sterically demanding

glycines, changes substantially as one of them (G2) isreplaced by a constrained geometry residue, such as proline[20]. In fact, 1H NMR, CD and FTIR measurements of theVPGVG repeating sequence with the Boc-, benzylesters,Ac- and methylamides end groups suggest the existenceof an equilibrium between a c-turn structure and a b-turnstructure in the Pro-Gly segment resulting in a single struc-ture that combines flexibility with strong conformationalpreferences [20].

3.2. Bond length and bond angles

Corresponding bond lengths in the 10 isomers have val-ues very close to each other, in agreement with the generalview that bond distances are slightly perturbed by confor-mation modifications (Tables 2 and S1). The small varia-tions observed are caused by hydrogen bond’s formation/breaking and by some associated structural relaxation(few units at the third decimal place). Thus, the mean bondlengths for various amino acid residues, within the±0.005 A range, are: RN–H = 1.014 A; RCO–N = 1.361 A;RC@O = 1.228 A; RN–Ca (Gly) = 1.449 A; RN–Ca(Val) =1.463 A; RCa–CO(Gly) = 1.531 A; RCa–CO(Val) = 1.544 A.The electron-releasing character of the iso-propyl groupsubstituent in the Ca of valine residues induces a lengthen-ing (�0.013 A) of the N–Ca and Ca–CO bonds as com-pared to analogous distances in glycine residues. Both theRN–H and RC@O bond lengths of group involved in hydro-gen bonds generally are longer (�0.005 A) than the corre-sponding >NH and >CO ‘‘free’’ groups thus indicating aslight bond strength reduction. Note that the RC@O inthe terminal ester group (1.214 A) is significantly shorterthan the homologous bond distance in amides indicativeof a greater bond strength.

Comparing theoretical bond distances in BocVGGV-GOEt peptide with the corresponding bond lengths ofsingle unit of glycine (RN–H = 1.019, RN–Ca = 1.451,RCa–CO = 1.525, RC@O = 1.211 A) and valine (RN–

H = 1.019, RN–Ca = 1.460, RCa–CO = 1.529, RC@O =1.213 A), a lengthening of the RC@O bonds (�0.02 A) isobserved. The RN–Ca bonds in single aminoacids undergoto minor variations upon peptide formation but they areconsiderably longer (�0.1 A) than RCO–N. These featuresare explained by the partial reduction of the C@O dou-ble-bond character and the formation of a partial double-bond character of the CO–N in the peptide.

Theoretical bond distances are in reasonable agreementwith experimental parameters derived from the X-raystructure of the BocVGGOH peptide recently reported[18] even though they are generally slightly longer (maxi-mum percent difference is 3, Table 2). Certainly, the adop-tion of a very large basis set and a rigorous treatment ofelectronic correlation effects should improve the compari-son [16]. It is worthy of note that important structural fea-tures theoretically derived as the Ca–CO and N–Ca

elongations on passing from glycine to valine residues,compare well to experimental data (Table 2).

Table 2Calculated equilibrium bond lengths (A) for the two most stable BocVGGVGOEt conformers

Bond lengtha C13 (I) Inverted-C14 (J) Averageb Experimentc Difference

C(tBu)–C(tBu) 1.531 1.531 1.531 1.505 1.7C(tBu)–C(tBu) 1.532 1.533 1.532 1.506 1.7C(tBu)–C(tBu) 1.531 1.533 1.532 1.509 1.5C(tBu)–O 1.479 1.469 1.476 1.489 0.9O–C 1.347 1.355 1.348 1.339 0.7C@O 1.226 1.225 1.227 1.224 0.2CO–N 1.371 1.366 1.369 1.362 0.5N–H (V1) 1.011 1.013 1.012N–Ca 1.464 1.459 1.460 1.463 0.2Ca–C(iPr) 1.549 1.551 1.551 1.535 1.0C(iPr)–C(iPr) 1.537 1.537 1.536 1.510 1.7C(iPr)–C(iPr) 1.536 1.534 1.536 1.514 1.5Ca–CO 1.536 1.537 1.540 1.537 0.2C@O 1.236 1.228 1.230 1.227 0.2CO–N 1.356 1.361 1.360 1.320 3.0N–H (G2) 1.011 1.016 1.014N–Ca 1.459 1.447 1.450 1.451 0.1Ca–CO 1.542 1.530 1.534 1.511 1.5C@O 1.224 1.231 1.228 1.242 1.1CO–N 1.363 1.363 1.359 1.329 2.3N–H (G3) 1.016 1.011 1.013N–Ca 1.454 1.460 1.454 1.449 0.3Ca–CO 1.534 1.540 1.539C@O 1.232 1.236 1.229CO–N 1.361 1.354 1.359N–H (V4) 1.018 1.020 1.016N–Ca 1.468 1.474 1.465Ca–C(iPr) 1.539 1.551 1.547C(iPr)–C(iPr) 1.535 1.537 1.536C(iPr)–C(iPr) 1.537 1.535 1.536Ca–CO 1.545 1.545 1.547C@O 1.235 1.227 1.228CO–N 1.354 1.366 1.361N–H(G5) 1.018 1.020 1.016N–Ca 1.448 1.440 1.444Ca–CO 1.516 1.531 1.522C@O 1.212 1.219 1.214OC–O 1.345 1.332 1.340O–C(Et) 1.449 1.458 1.452C–C(Et) 1.517 1.516 1.516

Average, experimental and percent differences between average and experimental bond lengths.a Labelling refers to Fig. 1.b Average bond angles for all the presently considered conformations.c Experimental bond angles derived from the BocVGGOH X-ray structure, see Ref. [18].


For all the structures presently analysed, the terminalgroup bond angles (tBu–O–CO = 121.6 ± 0.8�, O–CO–NH 109.7 ± 0.5�, Ca–CO–O 111.0 ± 2.0� and CO–O–Et116.4 ± 0.7�) are almost unaffected by chain conformation(Tables 3 and S2). Analogously, all the Ca–CO–NH(115.5 ± 2.0�) and the CO–NH–Ca(Gly) (121.2 ± 2.0�)backbone angles spread in a narrow range because of thegreat rigidity of the trigonal planar structure around car-bonyl carbons and nitrogens. Conversely, the CO–NH–Ca angles in valine residues (122.1 ± 3.5�) undergo greatervariations because the iso-propyl substituent induces a con-sistent rearrangement to minimise repulsive interactionsbetween iso-propyl side chain and the backbone. TheNH–Ca–CO bond angles for all residues undergo signifi-cant adjustments markedly depending on the adopted con-

formation (111.5 ± 5.5�). This can be associated with ahigher flexibility of the tetrahedral arrangement versusthe trigonal planar of nitrogen and carbonyl backboneatoms. These findings are in agreement with a well-estab-lished (at both theoretical and experimental levels) correla-tion between the NH–Ca–CO bond angles and the /i,wi

torsion angles. It has been shown that changes in theNH–Ca–CO bond angles are largely independent of thetype of amino acid residue, and vary with the backboneorientation in a concerted way [21].

3.3. Vibrational frequencies

The detailed one-to-one comparison of theoretical andexperimental frequencies is not straightforward [22,23].

Table 3Calculated equilibrium bond angles (deg) for the two most stable BocVGGVGOEt conformers

Bond anglea C13 (I) Inverted-C14 (J) Averageb Experimentc Difference

tBu–O–CO (Boc) 121.8 121.5 121.6 120.1 1.3O–CO–NH (Boc) 109.7 109.1 109.7 110.9 1.1CO–NH–Ca (V1) 121.2 122.3 121.5 118.6 2.4NH–Ca–CO (V1) 112.8 107.5 109.7 112.0 2.0Ca–CO–NH (V1) 117.5 115.6 115.4 117.8 2.0CO–NH–Ca (G2) 121.9 122.0 121.6 122.5 0.7NH–Ca–CO (G2) 111.1 110.4 112.9 115.5 2.2Ca–CO–NH (G2) 113.3 115.7 115.8 116.6 0.7CO–NH–Ca (G3) 123.0 121.2 121.6 123.1 1.2NH–Ca–CO (G3) 111.3 111.8 113.6 113.5 0.1Ca–CO–NH (G3) 116.3 114.2 115.6CO–NH–Ca (V4) 120.4 126.2 122.7NH–Ca–CO (V4) 107.0 111.1 109.9Ca–CO–NH (V4) 114.5 116.0 115.1CO–NH–Ca (G5) 120.5 119.0 120.5NH–Ca–CO (G5) 109.8 113.2 111.8Ca–CO–O (OEt) 110.6 110.4 111.0CO–O–Et (OEt) 115.9 117.6 116.4

Average, experimental and percent differences between average and experimental bond angles.a Labelling refers to Fig. 1.b Average bond angles for all the presently considered conformations.c Experimental bond angles derived from the BocVGGOH X-ray structure, see Ref. [18].


Actually vibrational frequencies derived by electronicstructure computations correspond to the isolated mole-cule, i.e. in vacuum. Conversely, FTIR spectra have beenrecorded in the solid state or in solution. In a simplifiedview two effects should be considered for molecules in thesolid state: (i) the gas-phase equilibrium structure mightdiffer from the solid state one; (ii) in addition to the intra-molecular hydrogen bonding in the solid state theBocVGGVGOEt molecules could be held together by anetwork of intermolecular hydrogen bonds between the>C@O and >N–H of amides, urethane and ester functions.Therefore, theoretical data cannot be directly comparedwith experimental solid state spectrum and a more mean-ingful comparison could be done with spectra recorded inapolar solvents. CHCl3 is a solvent with a very low polarity(e = 4.9), and therefore it is reasonable to assume that thepeptide maintains a structure close to that inferred fromgas-phase calculations. To gain additional information onhow the secondary structure of the BocVGGVGOEt pep-tide is influenced by solvents, FTIR measurements in theamide I and II regions (amide A region is hidden by strongsolvent absorptions) were carried out in higher polarity sol-vents such as TFE and DMSO (e = 26.67 and 46.7, respec-tively). Furthermore, TFE is a protic solvent and can beengaged in hydrogen bonding as donor or acceptor with>CO and >N–H amide groups, respectively [24]. TheDMSO molecules can only accept hydrogen bonds from>N–H groups of the peptides [2,25]. Therefore, bothTFE and DMSO can destroy, completely or in part,the intramolecular hydrogen-bonded c- and b-turnpseudocycles.

The evaluation of energy second derivatives reveals thatthe two most stable conformers C13 and inverted-C14

(structures I and J in Fig. 3) are actually minima on theBorn–Oppenheimer surface. Computed IR spectrum inthe harmonic approximation for the C13 and inverted-C14 conformers of both NH and ND isotopomers arereported in the Figs. 4 and 5 and compared with the exper-imental spectra in the solid state and in chloroform solu-tion. A rapid inspection of the figures (Figs. 4 and 5 forBocVGGVGOEt and BocVGGVGOEt-d5 peptides,respectively) shows a satisfactory agreement betweenexperimental and calculated IR frequencies. Theoreticalfrequencies are slightly overestimated but these differencescould be greatly reduced if a more flexible basis set andmore reliable electronic correlation treatments are consid-ered [16]. Nevertheless, it is reasonable to assume that theshift is constant for a homologous series of vibrations, thusimportant information can be derived from theoreticaldata of amide, terminal urethane and ester groups.

The calculated IR spectrum of both inverted-C14 andC13 conformers in the 3460–3625 cm�1 region exhibits fivenormal modes that are assigned to the mN–H stretching(Fig. 4) [26]. Two of these vibrations are located at higherenergy than the highest intensity components in both con-formers [27]. Normal mode eigenvectors analysis showsthat they correlate with the >N–H stretching of the aminoacid residues not involved in hydrogen bonding i.e. the V1

and G3 residues for inverted-C14 and V1 and G2 residuesfor C13. The N–H bonds weaken upon hydrogen bond for-mation and related mN–H stretching frequencies areobserved at lower values [27]. On the basis of computed fre-quency separations and IR intensity, for the C13 con-former the amide A signal should consist of a lowintensity peak at high frequency followed (�80 cm�1) byan intense band (Fig. 4). Conversely, for the inverted-C14

Abs

orba

nce

1000150020002500300035004000

Wavenumber (cm-1)

νN-H (amide A)

νC-H

νC=O(amide I)

δN-H(amide II)

δC-H

νC-N(amide III)

IR I

nten

sity

G5*V1V4*

G3

G2*

G3*

V1G2 V4*

G5*

# #

#

inverted-C14

C13

amide B

Fig. 4. Experimental FTIR spectra (in KBr pellet, black line, and in CHCl3 solution, red line), computed IR spectra for C13 and inverted-C14 isomers andassignment for the BocVGGVGOEt peptide. The ‘‘#’’ indicates absorptions due to solvent. The ‘‘*’’ indicates residues which >NH group is involved inhydrogen bonding. The insert (blue line) shows the theoretical spectrum convoluted as a sum of Gaussian functions. (For interpretation of the referencesto color in this figure legend, the reader is referred to the web version of this paper.)

Abs

orba

nce

1000150020002500300035004000

Wavenumber (cm-1)

νN-D (amide A)

νC-H

νC=O(amide I')

δN-D(amide II')

δC-H

IR I

nten

sity

G5*V1 V4*G3

G2*

#

V1G2G3*V

4*

G5*

##

C13

inverted-C14

Fig. 5. Experimental FTIR spectra (in KBr pellet, black line, and in CDCl3 solution, red line), computed IR spectra for C13 and inverted-C14 isomers andassignment for the BocVGGVGOEt-d5 peptide. The ‘‘#’’ indicates absorptions due to solvent or contamination of non-deuterated BocVGGVGOEtpeptide. The ‘‘*’’ indicates residues which >ND group is involved in hydrogen bonding. The insert (blue line) shows the theoretical spectrum convoluted asa sum of Gaussian functions. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)


conformer, two band envelops (G3/V1/G2 and V4/G5) withcomparable intensity and separated by 70 cm�1 are theo-retically predicted. Experimental IR spectrum recorded inCHCl3 shows two bands with comparable intensity centredat 3442 and 3330 cm�1 (Table 4 and Fig. 4) in qualitative

agreement with that calculated for the most stableinverted-C14 conformer.

In the solid state spectrum, the amide A band shifts to3289 cm�1 because of crystal packing forces [28]. Theabsence of evident peak splitting and the huge intensity

Table 4Relevant FTIR spectroscopic data for the NH and ND isotopomers of the BocValGlyGlyValGlyOEt peptide recorded in the solid state and in solution ofchloroform, trifluoroethanol, and dimethyl-sulfoxide

BocVGGVGOEt Amide A Amide I Amide II

KBr 3289 1735 1700 1674 1629 1554 1525CHCl3 3442 3330 1741 1675 1510TFE 1734 1667 1540DMSO 1749 1701 1670 1535

BocVGGVGOEt-d5 Amide A (N-D) Amide I 0 Amide II 0

KBr 2463 2417 1736 1694 1666 1626 1455 1420CDCl3 2540 2480 1740 1674 1447 1408TFE 1732 1662 1464 1414DMSO 1747 1695 1665 1447 1408

0.0

0.1

0.2

0.3

0.4

0.5

Abs

orba

nce

0.0

0.4

0.8

1.2

1.6

2.0

Abs

orba

nce

0

200

400

600

IR I

nten

sity

(K

m/m

ol)

14001500160017001800

Wavenumber (cm-1)

ester

urethaneδN-H (amide II)

δC=O (amide I) free H-bond

SOLID

CHCl3 TFE

DMSO

G5*

BO

C

V1

V4

G2*G3*

V4*G5*G2*

G3

V1

Fig. 6. Experimental (in the solid state, and in CHCl3, TFE and DMSOsolutions) and computed IR spectra of the BocVGGVGOEt peptide in the1800–1400 cm�1 region. The ‘‘*’’ indicates residues which >CO group isinvolved in hydrogen bonding.


with respect to the amide I absorption (amide A/amide Iarea ratio is 1.2) suggest the occurrence of additional fea-tures: (i) the Fermi resonance of the mC@O + dN–H com-bination powered by amide A modes [29–31]; and (ii)vibrational self-trapping [32]. There is ample experimentaland theoretical evidences for the occurrence of these phe-nomena in IR spectra of several solid state peptides[26–32].

The peak at 3083 cm�1 observed in solid state spectrumdoes not find a counterpart in the CHCl3 solution spectrumor in theoretical data. Generally, this band is indicated asamide B or sometime amide A 0 and originates from Fermiresonance between the first overtone of dN–H and themN–H vibrations [33] and partly from vibrational self-trap-ping phenomenon [32].

In the spectra of the deuterated BocVGGVGOEt-d5

peptide, mN-D stretches shift downward by about1000 cm�1 (amide A(N-D) absorption occur at�2500 cm�1) with a substantial intensity reduction in fullagreement with computed values (Fig. 5). Computed fre-quency separation between the two >N-D groups decreases(50 cm�1) and, in agreement with experimental data, thereis still a clear amide A(N-D) peak split in the CDCl3 solu-tion with a reduced gap between the two maxima (2540 and2480 cm�1). The amide A(N-D) band of the solid statesample is still located downward (50 cm�1) relative to thecorresponding signal in CDCl3 solution. In spite of thereduced separation between the two groups, the amideA(N-D) peak shows a doublet structure (the two maximaare at 2463 and 2417 cm�1) thus confirming that the bandat 3289 cm�1 of the non-deuterated sample contains animportant contribution from the Fermi resonance (themC–O + dN-D combination band would occur at�3100 cm�1). Analogously, the FTIR spectra of theBocVGGVGOEt-d5 peptide do not show any amideB(N-D) signal because the Fermi resonance is hampered(the first dN-D overtone would occur at �2900 cm�1).

Theoretical data show the six mC@O frequencies in the1730–1830 cm�1 region for both the inverted-C14 andC13 conformers (Figs. 4–7). These values slightly shiftdownward (<6 cm�1) in deuterated BocVGGVGOEt-d5

peptide. Eigenvectors analysis (Table 5) for the C13 con-

former indicates localisation of each normal mode on spe-cific C@O group. Conversely, for the inverted-C14conformer there is strong coupling of various oscillatorsto form complex collective vibrations. This coupling mayoccur through bond interactions between adjacent amidegroups, through hydrogen bond interactions, and/orthrough interactions between the oscillating electroniccharge densities of adjacent peptide groups [3,34–38]. Nev-ertheless, based on the dominant character, the mC@Ostretching can be ordered, in decreasing wavenumber,according to the following series:

Conformer inverted-C14 G5(ester) > BOC(urethane) > V1 > V4 > G2 > G3.Conformer C13 G5(ester) > G2 > BOC(urethane) > G3 > V1 > V4.


The mC@O of the terminal G5-OEt group always lies athighest wavenumber, well separated from the remainingmodes, thus indicating an high intrinsic stability of the ter-minal C@O ester group in agreement with the shorterRC@O bond distance found in all conformers (Tables 3and S1).

The relative frequencies of the other five mC@O stretch-ing strongly depend on the adopted conformation and inparticular on their possible involvement in hydrogen bond-ing [1–6,39–42]. In fact, hydrogen-bonded >C@O groups(labelled by ‘‘*’’ in Figs. 6 and 7) are shifted downwardbecause of the slight RC@O bond lengthening (�0.005 A)discussed in the geometrical structure section.

The amide I spectrum of the BocVGGVGOEt peptide inthe solid state shows four fairly well separated bands at1735, 1700, 1674 and 1629 cm�1 (Figs. 4 and 6 and Table4). On the basis of the present theoretical data for the moststable inverted-C14 conformer and of the experimentallyfound frequencies for ester carbonyl (1735 cm�1) and alkylurethane (1720–1680 cm�1) [40,41], the two high frequency

0.0

0.1

0.2

0.3

0.4

0.5

Abs

orba

nce

0.0

0.1

0.2

0.3

0.4

0.5

Abs

orba

nce

0

200

400

600

IR I

nten

sity

(K

m/m

ol)

14001500160017001800

Wavenumber (cm-1)

esterurethane

νC=O (amide I') H-bond free

SOLID

CDCl3TFE-d3

DMSO-d6

#

δN-D(amide II')

G5*

BO

C

V1

V4G2*

G3*

V4* G

5*

G2*

G3V

1

Fig. 7. Experimental (in the solid state, and in CDCl3, TFE-d3 andDMSO-d6 solutions) and computed IR spectra of the deuteratedBocVGGVGOEt-d5 peptide in the 1800–1400 cm�1 region. The ‘‘*’’indicates residues which >CO group is involved in hydrogen bonding. The‘‘#’’ indicates absorption due to contamination of non-deuteratedBocVGGVGOEt peptide.

bands are assigned to ester and urethane groups, respec-tively. The agreement between the empirical sequence andthat inferred theoretically for the inverted-C14 conformerexcludes the C13-like conformer as a candidate for thesolid state structure since in this case a different sequenceof vibrations is predicted. The peaks centred at 1674 and1629 cm�1 are assigned to free carbonyls and carbonylsinvolved in hydrogen bonds, respectively. In fact, the>C@O bands of simpler alkyl amides and free amide car-bonyls in peptides fall in the 1670–1690 cm�1 frequencyrange while the >CO groups involved in strong c-turn orb-turn hydrogen-bonded peptides or strong intermolecularhydrogen bonding occur in the range 1620–1650 cm�1

[40,41]. This assignment is in excellent agreement with pres-ent theoretical data for the inverted-C14 conformer whichindicate a frequency reduction upon hydrogen bonding for-mation. However, the computed separation cannot accountfor the experimental differentiation due to the enhancedintramolecular hydrogen bonding generally occurring inthe solid state [2].

These results are in agreement with our previous XPSresults on BocVGGVGOEt [14], in which spectral featuressupported the most stable conformation (inverted-C14(J)).

The amide I signals for spectra recorded in the CHCl3and TFE solutions are similar each to others and consistof a low intense peak at high frequency due to the mC@Oof the terminal G5-OEt ester group and of a large andintense peak which contains the five remaining mC@Ostretches (Figs. 4 and 6 and Table 4).

In the weakly interacting CHCl3 solvent the large bandstarts at 1720 and finishes at 1630 cm�1 thus indicating the

Table 5Harmonic frequency (cm�1) and normal mode composition (percentage)a

of the C@O stretching for the inverted-C14 and the C13 conformations ofthe BocVGGVGOEt peptide

Frequency Boc V1 G2 G3 V4 G5

Inverted-C14

1798 4 1 0 0 6 681777 35 20 2 1 6 21774 26 17 6 3 14 31772 3 21 0 4 27 41755 0 2 36 10 7 01741 0 1 16 41 3 0

C13

1828 0 0 0 0 1 831794 0 0 62 0 0 01773 67 0 0 1 0 01753 0 0 0 56 7 01740 0 10 0 7 48 11732 0 53 0 0 10 0

Contributions of the remaining groups have omitted for clarity.a The square of the normal mode coefficients has been normalised to

100.


presence of both weakly solvated and/or c-turns and b-turns pseudocycles. The absence of any clear separationbetween the bands suggests a high fluxionality of the pep-tide chain and possibly the presence of small percentagesof other conformers in rapid equilibrium as expected fromtheoretical data. On going to the spectrum recorded inTFE, there is a clear shift of amide I band maxima at lowerwavenumber thus indicating a more extensive involvementof >C@O groups in hydrogen bonding possibly includingalso TFE solvent molecules. The component of the bandbelow 1650 cm�1 suggests again the presence of weak c-turns and b-turns pseudocycles.

The amide I spectrum of the BocVGGVGOEt peptidedissolved in DMSO shows three well resolved bands at1749, 1701 and 1670 cm�1. The absence of any signal at1635 cm�1 doubtless indicates that all carbonyl amidegroups are not involved in hydrogen bonding since all the>NH groups of the peptide are probably engaged withDMSO molecules. In this case the BocVGGVGOEt mole-cule adopts an extended structure [2].

In all presently considered cases, the amide I 0 bands ofthe BocVGGVGOEt-d5 peptide are very similar to thenon-deuterated ones, both in terms of band maxima andband profiles (Fig. 7). The small shift of maxima at lowerwavelength (<8 cm�1) indicates that NH fi ND substitu-tion has a minor effect on amide I bands in agreement withcomputation (Figs. 6 and 7 and Table 4).

Below the mCO stretching, computational data revealfive vibrations having a high intensity mainly associatedwith the dN–H in-plane bending, the amide II absorptions(Figs. 4–7). These modes are grouped in two sets. Eigenvec-tor analysis shows that the set at higher energy is due to thethree >N–H groups involved in hydrogen bonds (G2, V4,G5 for the inverted-C14 conformer and G3, G5, V4 forthe C13 conformer), while the remaining set representsthe bending of free >N–H groups. The picture inferredfrom theoretical data is in agreement with the empiricalcorrelation between amide II band and hydrogen bondingpresence.

The form of the large amide II absorption for theBocVGGVGOEt peptide in the solid state suggests twomain components centred at �1550 and �1530 cm�1

(Fig. 6). These absorptions are diagnostic for >N–Hgroups engaged in hydrogen bonding and ‘‘free’’ amidegroups, respectively. The low resolution of spectral featuresin solutions hampers a detailed theoretical/experimentalcomparison. However, changes in frequency and shape ofamide II band indicate some structural modifications.

The amide II mode shifts from 1550 to 1455 cm�1 onNH fi ND exchange of backbone peptide groups. How-ever, this region is very complex for analysis in detail sinceit also contains contributions from the C–H bending ofCH2 and CH3 groups.

The amide III normal modes (mC–N and dN–H motionscoupled with (C)Ca–H bending motion) are computed inthe 1330–1230 cm�1 frequency range while for the deuter-ated peptide (amide III 0) they lies in the 1040–960 cm�1

region. IR intensity of these modes is very low, thereforethey are hidden by other absorptions and a useful theoret-ical/experimental comparison is hampered. Nevertheless,experimental Raman measurements reveal amide III andamide III 0 bands in the 1320–1220 cm�1 range and at990 cm�1, respectively, in good agreement with the presentdata [1,43].

4. Conclusions

A combined DFT-B3LYP theoretical analysis andexperimental FTIR study has been carried out to probemolecular properties of the BocVGGVGOEt peptide. Thelarge portion of the potential energy surface scrutinisedreveals several stable structures close in energy. They easilyinterconvert with each other thus inferring a great fluxio-nality to the peptide chain. The great flexibility of the pep-tide chain has been ascribed to the low steric encumbranceof glycine residues which allows easily rotation around /and w dihedral angles. All structures are stabilised by intra-molecular hydrogen bonds from amide protons to carbonyloxygens. The a-helices are not so stable and spontaneouslyconvert to 310 helices or distorted helices with b-turnsinside. However, even these helices are less stable thanother structures like classic b-turns or distorted folded con-formers. The two most stable structures are characterisedby a C14 and C13 pseudocycles involving hydrogen bond-ing of the terminal ester and urethane carbonyls, respec-tively. In both cases, the main hydrogen bond enclosureis accompanied by c-turns pseudocycles.

The calculated harmonic vibrational frequencies and thedipole derivative for the two most stable conformers haveprovided guidance for a rational and critical analysis ofthe FTIR spectra (of both NH and ND isotopomers)recorded in the solid state and in solution of CHCl3,TFE and DMSO. It is seen that the amide A and amideI bands are sensitive to hydrogen bonding and to transitiondipoles coupling thus providing information about the con-formation of the peptide chain.

The FTIR spectra in the solid state present well-resolvedbands thus suggesting the presence of a preferred singleconformation (the inverted-C14 one) as previously foundby XPS [14]. The single envelope amide I band, found inCHCl3 and TFE solutions, suggests rapid interconversionof various folded conformers stabilised by weakly hydro-gen-bonded b- or c-turns. In DMSO, it is likely that thepeptide chain assumes an extended conformation withoutany intramolecular hydrogen bonding.

As far as elastin is concerned, present results are in linewith and extend the recently reported relationship betweenstructure and elasticity of the protein [44,45]. In particular,an equilibrium between families of folded structures andfamilies of extended ones has been proposed to be at theorigin of the high entropy of the native, relaxed state ofthe elastomeric protein. That a similar equilibrium is alsopossible for a short elastin building block such asBocVGGVGOEt peptide is a further evidence of the


already demonstrated intrinsic self-similarity (fractality) ofelastin [46].

Acknowledgements

A.M.T. acknowledges the European Community (GrantElastage no. 018960). A.M.T. and A.M.S. acknowledge theMIUR (COFIN 2004).

Appendix A. Supplementary data

Figure S1 and Tables S1, S2; a complete list of Cartesiancoordinates of all structures presently analysed (24 pages).Supplementary data associated with this article can befound, in the online version, at doi:10.1016/j.theochem.2007.02.019.

References

[1] E. Vass, M. Hollosi, F. Besson, R. Buchet, Chem. Rev. 103 (2003)1917.

[2] T. Schirmeister, A. Breuning, A. Murso, D. Stalke, M. Mladenovic,B. Engels, A. Szeghalmi, M. Schmitt, W. Kiefer, P. Jurgen, J. Phys.Chem. A 108 (2004) 11398.

[3] A.V. Mikhonin, S.A. Asher, J. Phys. Chem. B 109 (2005) 3047.[4] J. Kubelka, R.A.G.D. Silva, T.A. Keiderling, J. Am. Chem. Soc. 124

(2002) 5325.[5] C. Kolano, J. Helbing, M. Kozinski, W. Sander, P. Hamm, Nature

444 (2006) 469.[6] M. Aschi, A. Mollica, G. Lucente, M.P. Paradisi, F. Mazza, J. Mol.

Struct. 785 (2006) 176.[7] F. Schotte, M. Lim, T.A. Jackson, A.V. Smirnov, J. Soman, J.S.

Olson, G.N. Phillips Jr., M. Wulff, P.A. Anfinrud, Science 300 (2003)1944.

[8] L. Debelle, A.M. Tamburro, Int. J. Biochem. Cell Biol. 31 (1999) 261.[9] A.M. Tamburro, V. Guantieri, L. Pandolfo, A. Scopa, Biopolymers

29 (1990) 855.[10] M.A. Castiglione Morelli, M. De Biasi, A. De Stradis, A.M.

Tamburro, J. Biomol. Struct. Dyn. 11 (1993) 181.[11] A. M Tamburro, A. Pepe, B. Bochicchio, D. Quagliano, I.P.

Ronchetti, J. Biol. Chem. 280 (2005) 2682.[12] H. Broch, M. Moulabbi, D. Vasilescu, A.M. Tamburro, Int. J.

Peptide Protein Res. 47 (1996) 394.[13] R. Flamia, P.A. Zhdan, M. Martino, J.E. Castle, A.M. Tamburro,

Biomacromolecules 5 (2004) 1511.[14] R. Flamia, G. Lanza, A.M. Salvi, J.E. Castle, A.M. Tamburro,

Biomacromolecules 6 (2005) 1299.

[15] A. Perczel, A. McAllister, P. Csaszar, I.G. Csizmadia, J. Am. Chem.Soc. 115 (1993) 4849.

[16] A. Perczel, O. Farkas, I. Jakli, I. Topol, I.G. Csizmadia, J. Comput.Chem. 24 (2003) 1026.

[17] M.J. Frisch, et al., Gaussian 03, Revision B.04, Gaussian, Inc.,Pittsburgh PA, 2003.

[18] M. Martino, A. Bavoso, M. Saviano, B. Di Blasio, A.M. Tamburro,J. Biomol. Struct. Dyn. 15 (1998) 861.

[19] V. Lalitha, R. Murali, E. Subramanian, Int. J. Peptide Protein Res. 27(1986) 472.

[20] O. Arad, M. Goodman, Biopolymers 29 (1990) 1652.[21] I. Jakli, A. Perczel, O. Farkas, A.G. Csaszar, C. Sosa, I.G. Csizmadia,

J. Comput. Chem. 21 (2000) 626.[22] P. Bour, D. Michalik, J. Kapitan, J. Chem. Phys. 122 (2005) 144501.[23] K.J. Jalkanen, M. Elstner, S. Suhai, J. Mol. Struct.: THEOCHEM

675 (2004) 61.[24] H. Reiersen, A.R. Rees, Protein Eng. 13 (2000) 739.[25] M.F. DeCamp, L. DeFlores, J.M. McCracken, A. Tokmakoff, J.

Phys. Chem. B 109 (2005) 11016.[26] J. Bandekar, S. Krimm, Biopolymers 27 (1988) 885.[27] N.G. Mirkin, S. Krimm, J. Phys. Chem. A 108 (2004) 5438.[28] S. Krimm, J. Bandekar, Adv. Protein Chem. 38 (1986) 181.[29] S. Krimm, A.M. Dwivedi, J. Raman Spectr. 12 (1981) 133.[30] I.G. Rubtsov, J. Wang, R.M. Hochstrasser, J. Phys. Chem. A 107

(2003) 3384.[31] O. Hritzova, D. Koscik, Collect. Czech. Chem. Commun. 59 (1994)

951.[32] J. Elder, P. Hamm, Phys. Rev. B 69 (2004) 214301.[33] T. Hasegava, J. Umemura, N. Yamada, J. Mol. Struct. 735–736

(2005) 63.[34] R. Schweitzer-Stenner, J. Phys. Chem. B 108 (2004) 16965.[35] P. Bour, T.A. Keiderling, J. Phys. Chem. B 109 (2005) 5348.[36] R. Huang, J. Kubelka, W. Barber-Armstrong, R.A.G.D. Silva, S.M.

Decatur, T.A. Keiderling, J. Am. Chem. Soc. 126 (2004) 2346.[37] P. Bour, J. Kubelka, T.A. Keiderling, Biopolymers 53 (2000) 380.[38] P. Hamm, S. Woutersen, Bull. Chem. Soc. Japan 75 (2002) 985.[39] W.K. Surewicz, H.H. Mantsch, D. Chapman, Biochemistry 32 (1993)

389.[40] G. Jalsovszky, S. Holly, M. Hollosi, J. Mol. Struct. 348 (1995) 329.[41] E. Vass, S. Holly, Z. Majer, J. Samu, I. Laczko, M. Hollosi, J. Mol.

Struct. 408/409 (1997) 47.[42] M.T. Nguyen, K. De Wael, T. Zeegers-Huyskens, J. Phys. Chem. 99

(1995) 9739.[43] R. Tuma, P.E. Prevelige Jr., G.J. Thomas Jr., Proc. Natl. Acad. Sci.

95 (1998) 9885.[44] B. Bochicchio, N. Floquet, A. Pepe, A.J.P. Alix, A.M. Tamburro,

Chem. Eur. J. 10 (2004) 3166.[45] A.M. Tamburro, B. Bochicchio, A. Pepe, Biochemistry 42 (2003)

13347.[46] A.M. Tamburro, A. De Stradis, L. D’Alessio, J. Biomol. Struct. Dyn.

12 (1995) 1161.

http://dx.doi.org/10.1016/j.theochem.2007.02.019

http://dx.doi.org/10.1016/j.theochem.2007.02.019

molecular properties of a representative glycine-rich sequence of elastin – bocvggvgoet: a...

Documents