cations in a molecular funnel: vibrational spectroscopy of...

DOI: 10.1002/cphc.201200810

Cations in a Molecular Funnel: Vibrational Spectroscopy ofIsolated Cyclodextrin Complexes with Alkali MetalsFrancisco G�mez,[a] Paola Hurtado,[a] Ana R. Hortal,[a] Bruno Mart�nez-Haya,*[a] Giel Berden,[b]

and Jos Oomens[b, c]

1. Introduction

Cyclodextrins (CD) are central macrocycles in modern supra-molecular chemistry.[1–4] The most common native cyclodex-trines are made of six (aCD), seven (bCD) or eight (gCD) a-d-glucopyranosyl units connected by a-(1!4)-glycosidic linkag-es. Their general three-dimensional architecture is of funnel-like geometry, with a narrower opening (primary face) shapedby the H-bonding network built by the -CH2OH side groups(one group per glucose unit), and with a broader opening (sec-ondary face) with shorter -OH hydroxylic side groups (twogroups per glucose unit). This is illustrated in Figure 1, whichshows the most stable C6 symmetric conformation of the iso-lated aCD. While the two rims of the native cyclodextrines arestrongly polar, the interior of the cavity is hydrophobic. Thismakes CDs versatile substrates for inclusion chemistry in aque-ous solution and constitutes the basis of numerous applica-tions.[4–7] The complexation of a given CD/guest pair is deter-mined by the balance between the inter- and intramolecularinteractions (dispersive attractions, hydrogen bonding and hy-drophobic effects) and the steric effects due to guest size andgeometry.

Fundamental insights into the mechanisms driving inclusionchemistry can be obtained by elucidating the structure ofhost–guest complexes under well-defined environmental con-

ditions in the gas phase. The strength of this microscopic“bottom-up” approach relies on the close link with quantumchemistry derived from the accurate definition of the molecu-lar system. This kind of studies has benefited from remarkabledevelopments in modern spectroscopy and mass spectrometrytechniques.[8, 9] Benchmark investigations of isolated cyclodex-trin complexes, while relatively scarce, have led to significantadvances in the understanding of the conformational con-

The benchmark inclusion complexes formed by a-cyclodextrin(aCD) with alkali-metal cations are investigated under isolatedconditions in the gas phase. The relative aCD-M+ (M = Li+ ,Na+ , K+ , Cs+) binding affinities and the structure of the com-plexes are determined from a combination of mass spectrome-try, infrared action spectroscopy and quantum chemical com-putations. Solvent-free laser desorption measurements reveala trend of decreasing stability of the isolated complexes withincreasing size of the cation guest. The experimental infrared

spectra are qualitatively similar for the complexes with thefour cations investigated, and are consistent with the bindingof the cation within the primary face of the cyclodextrin, aspredicted by the quantum computations (B3LYP/6-31 + G*).The inclusion of the quantum-chemical cation disrupts the C6symmetry of the free cyclodextrin to provide the optimum co-ordination of the cations with the -CH2OH groups in C1, C2 orC3 symmetry arrangements that are determined by the size ofthe cation.

Figure 1. a) Top view of the free a-CD in its most stable B3LYP/6-31 + G*conformation. b) Side view of the same conformation with a schematic indi-cation of the possible binding positions of the alkali cations on the primaryside (a,b), on the secondary side (c,d) or inside the cavity (e).

[a] Dr. F. G�mez, Dr. P. Hurtado, Dr. A. R. Hortal, Prof. B. Mart�nez-HayaDepartment of Physical, Chemical and Natural SystemsUniversidad Pablo de Olavide, 41013 Seville (Spain)

[b] Dr. G. Berden, Prof. J. OomensFOM Institute for Plasma Physics RijnhuizenEdisonbaan 14, 3439 MN Nieuwegein (The Netherlands)

[c] Prof. J. Oomensvan’t Hoff Institute for Molecular SciencesUniversity of Amsterdam, Science Park 9041098 XH Amsterdam (The Netherlands)

� 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim ChemPhysChem 2013, 14, 400 – 407 400

CHEMPHYSCHEMARTICLES

straints and interactions responsible for the selective inclusionof specific guests. Lebrilla and coworkers explored the gas-phase chemistry of cyclodextrins with mass spectrometric iontrapping techniques.[9] Their studies focused on specific as-pects such as enantiomeric selectivity[11, 10] or the stabilizationof zwitterion forms of aminoacids upon complexation.[12] Morerecent studies have employed amino acids fragmentation massspectrometry to investigate cyclodextrin complexes with smallorganic molecules.[13–15] A number of computational studieshave served as support to the experimental investigations,leading to the general picture that isolated native cyclodextrinsadopt symmetric conformations with extensive hydrogenbonding networks in their primary and secondary hydroxylrims.[15–23] The coordination of a guest with the cyclodextrinmust then compensate for the disruption of such a set of intra-molecular hydrogen bonds.

This work focuses on the gas-phase complexes formed byaCD with the metal cations of the alkali series. These com-plexes constitute valuable benchmark models to probe theconformational properties of cyclodextrins. In addition, theyare helpful to explore potential selective effects induced bythe presence of alkali cations bound to the substrate on theformation of complexes with anionic organic guests.[24] Thestudy intends to provide new insights into some fundamentalissues that remain unresolved and have been under discussionin recent years,[25–29] such as the basic question of the cationbinding site that provides the optimum coordination with thecyclodextrin and its relation with the relative cation affinities ofthis substrate. The possible binding sites of the cation are rep-resented schematically in Figure 1 and are discussed through-out the paper. The investigation combines laser desorption/ionization applied to solvent-free samples, infrared spectrosco-py of complexes isolated in an ion trap and quantum chemicalcomputations. Details of the methodology are provided in thefollowing Section.

Methods

MALDI Mass Spectrometry and Sample Preparation

Matrix-assisted laser desorption/ionization (MALDI) experimentswere performed with an UltrafleXtreme spectrometer (Bruker-Dal-tonics) in the positive ion reflectron mode. For each spectrum,2000 laser shots were averaged at a 1 kHz repetition rate (355 nmnanosecond pulses from a Nd:YAG laser). Mass calibration was per-formed with a mixture of polar polymers (polydispersed polyethy-lene glycol with average molecular weights of ca. 600 and 1000,Sigma–Aldrich).

The determination of gas-phase host/guest affinities with MALDI isnot a trivial task. Since solvent effects can modify the relative in-tensities of different cyclodextrin complexes observed in MALDIspectra, solvent-free approaches (SF-MALDI) become essentialwhen dealing with gas-phase affinities.[30, 31] Under solvent-free con-ditions, the interaction between the cyclodextrin and the cationtakes place primarily in the gas phase, after the initiation of thelaser desorption process. In a typical SF-MALDI experiment, the cy-clodextrin, the cation precursors and the MALDI matrix are mixedas solid fine powders, without the addition of any solvent. Well-de-fined relative molar ratios of the cation precursors are required to

evaluate the relative affinity of the host for the cations present inthe sample. It is as well crucial that the density of cations in thedesorption plume resembles their original molar ratio in thesample. In our experiments, this is achieved by combining pairs ofalkali salts of similar lattice energy in each sample; for instance, LiI/NaBr, NaI/KCl and KI/CsBr. For the present experiments, equimolaramounts of a-cyclodextrin, the two alkali salts and the 2,5-DHBmatrix (2,5-dihydroxybenzoic acid) were mixed in the samples. Inthis way, the relative gas-phase stabilities of the aCD-M+ com-plexes were obtained from the intensities of the correspondingpeaks in the mass spectra.

For the sake of comparison, a MALDI experiment with the conven-tional dried-droplet method for sample preparation was also car-ried out. In this case, the aCD was co-dissolved in water/methanolwith equimolar amounts of the four alkali chlorides and an excessof the 2,5-DHB matrix. A 2 mL aliquot of the solution was thenspotted on the sample plate and the solvent was allowed to evap-orate. Since the lattice energy of the chlorides decreases with thesize of the alkali, a greater concentration of the heavy cations inthe plume is expected. Furthermore, the detection of aCD-M+

complexes formed in solution or during the sample precipitationprocess cannot be ruled out. Such effects hinder the determinationof accurate gas-phase affinities with the dried-droplet MALDImethod.

ESI-FTICR IRMPD Spectroscopy

Infrared multiple photon dissociation (IRMPD) spectroscopy wasperformed with a Fourier-transform ion cyclotron resonance (FT-ICR) mass spectrometer coupled to the free electron laser FELIX.[32]

The aCD-M+ complexes (M = Li, Na, K, Cs) were produced by elec-trospray ionization (ESI) of 0.1 mm solutions of the a-cyclodextrinand the corresponding alkali chloride salt in water/methanol sol-vent. Ions were accumulated in a hexapole ion trap and then pulseinjected into the ICR cell, where they were mass isolated at roomtemperature and irradiated with typically 25 FELIX macro-pulses.Each macro-pulse is approximately five microseconds long, has anenergy of about 35 mJ, and consists of a train of micropulses witha repetition frequency of 1 GHz. The spectral bandwidth of the ra-diation amounts to 0.5 % of the central wavelength. A more in-volved description of typical experimental procedures can befound in ref. [33].

If the laser infrared wavelength is in resonance with a vibrationalmode of the complex, multiple photon absorption occurs leadingto the eventual dissociation of the parent ion. The main fragmenta-tion pathway observed for the complexes with K+ and Cs+ corre-sponds to the detachment of the cation from the aCD. Fragmenta-tion of the aCD-Li+ and aCD-Na+ complexes proceeds via cleav-age of glycosidic bonds resulting in cationized fragments lackingone or several sugar units, similar to those found with non-reso-nant far infrared IRMPD in a previous study.[34] IRMPD spectra wereobtained by plotting the fragmentation yield, defined as the sumof all fragment ions divided by the sum of all ions, as a function ofinfrared wavenumber. Since K+ is not easily quantified in the FT-ICR spectrometer, the IRMPD spectrum of aCD-K+ was recorded bymonitoring the depletion of the parent aCD-K+ ions as a functionof the irradiation wavenumber. In the 970–1200 cm�1 range, theIRMPD spectra have been recorded with a reduced laser energy of10 mJ per macropulse in order to prevent strong depletion of theparent ions. All spectra have been linearly corrected for the wave-length dependent variations in laser pulse energy.


CHEMPHYSCHEMARTICLES www.chemphyschem.org

www.chemphyschem.org

Quantum Chemical Calculations

Simulated annealing with the universal force field was appliedto generate an ensemble of molecular structures of the aCD-M+ complexes. The Materials Studio package was employedfor this purpose.[35] The conformations of lowest energy wereoptimized with density functional theory (DFT) employing theB3LYP functional and the 6-31 + G* basis set. For the Cs+

cation, the core electrons were substituted by the Stuttgart/Dresden effective core potentials.[36] The calculations were car-ried out with the Gaussian 09 code.[37] The theoretical harmon-ic IR spectra shown in this work were generated by convolut-ing the calculated vibrational stick spectrum with a Lorentzianline-shape function with a full-width-at-half-maximum of15 cm�1. For comparison with experiment, the computed vi-brational frequencies for all the complexes were scaled bya factor of 0.985, which provided a good match with the domi-nant band in the IRMPD spectra related to the C�O stretchingvibrational modes.

2. Results and Discussion

2.1. MALDI Experiments

The SF-MALDI spectra recorded in this investigation are shownin Figure 2. In each individual experiment, two alkali cationsare brought into competitive interaction with the cyclodextrin,namely, Li+ vs Na+ , Na+ vs K+ and K+ vs Cs+ . The intensitiesof the aCD-M+ complexes in the mass spectra reveal a cleartrend of decreasing affinity with increasing cation size. Onquantitative grounds, the recorded signals lead to relativeabundances of the complexes of 1.00:0.30:0.10:0.05 for the Li :-Na:K:Cs series. It is interesting to notice that these alkali-cationaffinities are in qualitative agreement with those determinedfor cyclofructans with dissociation experiments in an iontrap.[38] Previous electrospray ionization experiments assigninga larger affinity of the b-cyclodextrin molecule towards sodiumwith respect to all the other alkaline cations[27] are likely to re-flect solvent effects rather than a merely gas-phase behavior.While not of relevance to the present study, it can be noticedthat a small peak at m/z = 985.3 is observed in the SF-MALDIspectrum of the aCD/Li+/Na+ mixture, which corresponds tothe deprotonated CD bound to two Li+ cations, hence thesingly charged adduct [aCD-H + 2 Li]+ .

Figure 2 also shows that the dried-droplet MALDI methodleads to completely different results in comparison to solventfree MALDI. In fact, it can be observed that the intensity of theK+ and Cs+ complexes dominate the dried-droplet MALDImass spectrum. This can be attributed to solvent effects (de-tection of ions preformed in solution) and, more importantly,to the enhanced concentration of heavy cations in the plumedue to the smaller lattice energy of the chloride precursors.

Finally, it is also worth remarking that the cation affinitiesobtained with the SF-MALDI method follow the same trend asthose derived in earlier studies for the complexes formed bylinear polyethers with the same alkali cations.[31] For theselatter systems, the higher charge density of the smaller cations

and their tighter coordination with the oxygen atoms of thepolymer were responsible for the greater affinities under sol-vent-less environmental conditions. The question arises ofwhether cyclodextrins display sufficient flexibility for coordina-tion as to lead to a similar rationale of their preferential bind-ing of the smaller cations. The conformational aspects behindthis behavior are explored in the remainder of the paper.

2.2. Conformation of the aCD-M+ Complexes

The key questions faced in the present study are the determi-nation of the optimum binding site of the alkali cations (seeFigure 1), the rearrangement demanded to the cyclodextrinsubstrate to accommodate the coordination of its oxygenatoms with the cation, and the potential occurrence of cationsize effects that may enhance the affinity for a specific alkali.We analyze these aspects at this point in the light of the pre-dictions of the DFT study.

The computational exploration of the conformational land-scape of the isolated aCD led to a lowest energy conformer ofC6 symmetry with extended hydrogen-bonding networks alongboth rims of the cyclodextrin. This conformer is represented inFigure 1 and is in agreement with the one postulated in previ-ous studies with different levels of theory.[18] It can be appreci-

Figure 2. Top: Solvent-free MALDI measurements of aCD complexed withalkali metals. In each spectrum, two cations compete for binding with thecyclodextrin and a trend is observed of decreasing stability of the complexeswith increasing cation size. Bottom: Mass spectra obtained with convention-al dried-droplet MALDI, yielding stronger signals for the complexes of thelarger cations. This latter finding does not reflect the actual gas-phase affini-ties but it is a result of the greater density achieved for those cations in thedesorption plume (see text for details).




ated that the structure of the cyclodextrin is stretched into itscharacteristic funnel-like geometry by the concerted H-bond-ing network formed between the longer -CH2OH side groups.This molecular structure can be considered a reference for theevaluation of the conformational changes that the cyclodextrinmust undergo upon complexation with the alkali cations.

The preferential binding site of the alkali cations in the cy-clodextrin complexes may a priori be within its primary or sec-ondary faces, as well as towards the cyclodextrin wall or in theinner part of the cavity. In each case a different type of coordi-nation arrangement with the hydroxyl side groups and/or withthe oxygen atoms of the sugar cycles or of the glycosidicbonds can be expected to take place. We will see that unspe-cific binding of the cation on the outside of the cavity, withoutalteration of the H-bonding network of the substrate, is notenergetically favored. In contrast, the general prediction of thepresent computational survey is that the alkali cations bind onthe primary face of the cyclodextrin through multiple coordi-nation with the oxygen atoms of the -CH2OH side groups. Itwill be shown that such prediction is corroborated by the vi-brational spectroscopy experiments.

We will first outline the main trends found in the B3LYP/6-31 + G* study taking as reference the results for the aCD-Na+

complex. Figure 3 depicts the lowest energy B3LYP conforma-tions for this complex, which are labelled Na1-Na6 in order ofincreasing free energy. For each of them, a number of closelying stable conformations resulting from minor structural var-iations of the cyclodextrin backbone and the side groups werealso found which will not be discussed here. In the four aCD-Na+ conformations of lowest energy (Na1-Na4), the cation islocated on the primary face of the cyclodextrin. The most

stable structure (conformer Na1) corresponds to a C2 symmetryarrangement in which the cation builds a coordination withfour oxygen atoms of the hydroxyl side groups. The nexthigher conformations feature a three-fold coordination, withthe cation either centered or displaced to the side of the pri-mary face (Na2, Na3 and Na4, 2, 8 and 11 kJ mol�1, respective-ly). Incidentally, the cation-centered C3 symmetry arrangementof the Na3 conformation is qualitatively similar to the onefound in the alkali complexes of cyclofructans.[38] No furtherprototypical conformations of the aCD-Na+ complex werefound within 50 kJ mol�1 above Na1. Conformers Na5 and Na6represent the most stable types of structures related to thebinding of the cation on the secondary face of the cyclodex-trin. In conformer Na5 (64 kJ mol�1), the cation launches itselfon the cyclodextrin wall, where it forms a tight five-fold coordi-nation with three hydroxylic and two glycosidic oxygen atoms.In conformer Na6 (132 kJ mol�1), of C2 symmetry, the cation liesat the center of the secondary face and the coordination in-volves four hydroxyl groups. In both Na5 and Na6, a markeddeformation is induced on the cyclodextrin substrate, whichexplains the high values of the relative free energies of theseconformers.

A similar type of binding hierarchy as the one just describedfor aCD-Na+ was found for the rest of alkali cations, althoughwith some differences in the energetic ordering of the con-formers. Figure 4 shows the lowest energy conformers for theLi+ , K+ and Cs+ complexes. It can be appreciated that the sizeof the cation determines the coordination number and sym-metry of the complex. For Li+ , a threefold-coordinated struc-ture with the cation displaced to the side of the primary face(Li1, qualitatively similar to conformer Na2) is most stable. The

threefold-coordinated arrange-ments Li2 and Li3, similar to Na3and Na4, in which the cation oc-cupies a more centered positionwithin the primary face, liesomewhat higher in energy (12and 18 kJ mol�1, respectively). In-terestingly, the fourfold-coordi-nated conformation Li4, similarto Na1, is even less stable(29 kJ mol�1). This latter trend re-verses with increasing cationsize, as we have seen above foraCD-Na+ . For aCD-K+ , the moststable structure K1 is also a four-fold-coordinated arrangement ofC2 symmetry similar to Na1. Thethreefold-coordinated structureK2, similar to Na2, is significantlyhigher in energy (16 kJ mol�1).We note in passing that the C3conformer (similar to Li1 andNa3) is not stable in the aCD-K+

complex. Finally, the larger sizeof the Cs+ cation increases thecoordination number of the

Figure 3. Low-energy conformations of the isolated aCD-Na+ complex at the B3LYP/6-31 + G* level. In conformersNa1–Na4, the cation binds within the primary face of the cyclodextrin. In the higher energy conformers Na5 andNa6, the cation binds within the secondary face. The values in brackets are the relative free energies of each con-former in kJ mol�1. The solid-line square drawn on the most stable Na1 structure highlights the fourfold metal–oxygen coordination framework.




most stable conformer Cs1 to five, although the fourfold-coor-dinated C2 conformer Cs2 lies only 3 kJ mol

�1 higher in energy.It is worth mentioning that stable inclusion complexes were

found for Na+ , K+ and Cs+ , in which the cyclodextrin substrateessentially maintains the C6 structure and H-bonding networksof its non-complexed form, and the cation coordinates fromthe inner part of the cavity with all six oxygen atoms of theCH2OH primary groups. This type of conformation is illustratedfor aCD-Cs+ in Figure 4. These highly symmetric conforma-tions have relative free energies of 36 kJ mol�1 for Cs+ ,30 kJ mol�1 for K+ and 72 kJ mol�1 for Na+ above the corre-sponding most stable conformations Cs1, K1 and Na1, respec-tively. The comparably greater stability found in the K+ com-plex arises from the more favourable matching between thesize of the cation and the diameter of the C6 ring arrangementdefined by the OH groups. For Cs+ , a stable C6 conformationwas also found with the cation bound from the outer part ofthe cyclodextrin (Cs3 in Figure 4). That conformation is in factmore stable (23 kJ mol�1) than the inclusion complex with theCs+ cation inside the cavity (Cs4). This can again be interpret-ed as a size effect, as for Na+ and K+ that type of conforma-tion was found to be unstable and to relax systematicallybringing the cation inside the cavity through the primary face.

The B3LYP results can be employed to rationalize the relativeabundances of the aCD-M+ complexes in the SF-MALDI experi-ments described in Section 2.1. Considering the most stableconformations predicted for each cation (Figure 4), the B3LYP/6-31 + G* evaluation of the free energies for the gas-phasecation exchange reactions aCD-M+ + Li+!aCD-Li+ + M+ indi-cates that these processes are exothermic by 92 and167 kJ mol�1 for M = Na, K, respectively. The analogous exo-

thermicity for the Cs+ complexcannot be determined accuratelyfrom the present computationssince core potentials are usedfor the calculation in that case.Nonetheless, stability for thecomplexes in the order Li>Na>K>Cs can safely be assumedfrom the B3LYP energetics,which is in agreement with thetrend of relative affinities derivedfrom the SF-MALDI experiments.

2.3. IRMPD Spectroscopy ofthe aCD-M+ Complexes

IRMPD spectroscopy experi-ments were performed to char-acterize the vibrational modes ofthe isolated aCD-M+ complexes.The resulting spectra are expect-ed to provide sensitive probesto determine the most stablebinding site of the alkali cationsand the overall conformationalstructure of the complexes.

Figure 5 shows the IRMPD spectra recorded for the four aCD-M+ complexes in the 900–1500 cm�1 range. The vibrationalmodes observed within this spectral region correspond to C�Cstretching (900–950 cm�1), to C�O stretching in the sugarrings, the glycosidic bonds, and the C�O�H side groups (950–1170 cm�1), and to an extended series of complex concertedmotions involving C�O�H and R�C�H bending, as well as CH2twisting, wagging and rocking (1200–1500 cm�1).

Figure 4. Lowest energy conformers of the isolated aCD-M+ complexes (M = Li, K, Cs) at the B3LYP/6-31 + G*level. In all cases, the cations binds on the central part of the primary face and coordinates with the oxygenatoms of the -CH2OH groups. For conformers Cs3 and Cs4, top and side views are shown for visualization of thein-cavity or out-of-cavity location of the Cs+ cation. The values in brackets are the relative free energies of eachconformer in kJ mol�1. The triangle, square or pentagon drawn on the most stable structures Li1, K1 and Cs1 high-light the metal–oxygen coordination framework in each case.

Figure 5. Experimental IRMPD spectra of the four aCD-M+ isolated com-plexes scoped in this study. The bands observed correspond to C�C stretch-ing modes (900–950 cm�1), C�O stretching modes (970–1200 cm�1), andcomplex concerted motions involving C�O�H and R�C�H bending, as wellas CH2 twisting, wagging and rocking (1200–1500 cm

�1). The intensity in theregion of the strong C�O stretching band has been scaled by a factor 0.5for a better visualization of the weaker bands.




It can be appreciated that the IRMPD spectra of the fourcomplexes share similar features. In particular, the envelope ofthe intense C�O stretching band displays a progression of sev-eral maxima at roughly the same positions in all cases. Thehigher wavenumber range associated with the bending mo-tions and the CH2 deformation modes also displays a similarband progression in the four spectra. This observation alreadysuggests that the four aCD-M+ complexes are stabilized inconformationally close structures with similar docking sites ofthe cation. Hence, no qualitative structural changes appear totake place as a function of cation size. It is shown below thatthe experimental spectra are consistent with cation binding inall cases within the primary face of the cyclodextrin as predict-ed by the computed thermochemistry.

Figure 6 compares the experimental IRMPD spectrum of theaCD-Na+ complex with the harmonic IR spectra predicted bythe B3LYP computation for the Na1, Na2, Na5 and Na6 con-formers discussed above. The B3LYP spectrum of the lowestenergy conformer Na1 reproduces virtually all the fine detailsof the IRMPD bands and provides an excellent match of theexperimental spectrum. The only relevant difference is foundfor the positions of the higher frequency bands (>1200 cm�1),

which are blue-shifted by about 25 cm�1 in the computation.This suggests that the frequency scaling of the computed vi-brational modes involving motions of the light H atoms de-mands a slightly smaller scaling factor for comparison with ex-periment than those of the C�C and C�O stretching modes.This is actually not unusual since, formally, a different scalingfactor is associated with any given vibrational mode.[42] The IRspectrum of the Na2 conformer is also consistent with experi-ment, although the relative intensities of the peaks within theC�O band and the CH2 vibrational progressions are not so ac-curately reproduced. The B3LYP spectra of conformers Na3 andNa4 (not shown) provide a similar level of agreement with themeasured IRMPD spectrum. Hence, a significant (not dominant)contribution to the experiment from conformers Na2, Na3 andNa4 would be consistent with the recorded IRMPD spectrum.

On the other hand, the present experiments do rule out anyappreciable population of the conformers associated witha binding of the cation on the secondary face of the cyclodex-trin (e.g. , Na5 and Na6). The B3LYP spectra for this type of con-formers display marked differences with the IRMPD spectrum,which are particularly apparent for the C�C and C�O stretch-ing bands. For instance, the C�C stretching band is blue-shift-ed for Na5 and red-shifted for Na6 in comparison to experi-ment. Moreover, the envelope of the C�O stretching band issignificantly more congested for both of these conformersthan found in the experiment, leading to peak positions incon-sistent with the measurement.

The analysis of the experimental IRMPD spectra for the othercations explored, Li+ , K+ and Cs+ , and their comparison withthe B3LYP computation led to a similar conclusion of preferen-tial binding of the cation on the primary face of the cyclodex-trin. Figure 7 shows the excellent agreement between themeasured spectrum and the computational IR spectrum pre-dicted for the corresponding lowest energy conformer depict-ed in Figure 4. In contrast, binding of the cation on the secon-dary face was systematically found not to be compatible withthe observed IRMPD vibrational bands.

3. Conclusions and Final Remarks

The present experimental and computational study concludesthat alkali cations bind preferentially on the primary face ofthe cyclodextrin, with a weak or moderate distortion of thesubstrate backbone or the H-bonding network along the sec-ondary face. The optimum coordination of the primary hydrox-yl groups with the cations and the relative stability of the C1,C2, C3 or C6 symmetry arrangements of the complexes are de-termined by the size of the cation, in a qualitatively similarway as in other cyclic sugars[38] and even in polyethers.[39, 40]

This scenario lays down a relatively simplified framework forthe rationalization of the interaction of the alkali cations withthe aCD. The stronger binding of the smaller alkali cations inthe isolated aCD-M+ complexes results from their highercharge density and tighter interaction with the oxygen atomsof the primary -CH2OH groups. In fact, the average O�M+ dis-tance within the coordination shell of the cation increases sys-tematically with cation size, from 1.9 � in the most stable Li+

Figure 6. Comparison of the IRMPD spectrum measured for the isolatedaCD-Na+ complex with the B3LYP IR spectra predicted for the conformersNa1–Na4 depicted in Figure 3. The vertical grid lines help one follow theband assignments. The intensity in the region of the strong C�O stretchingband (970–1200 cm�1) has been scaled in all spectra by a factor of 0.5 fora better visualization of the weaker bands.




complex to 3.3 � in the corresponding Cs+ complex. Theshorter coordination distances and the higher charge densitycompensate for the reduced coordination number found inthe complexes of the smaller cations.

Also importantly, the binding of the cation on the primaryface of the cyclodextrin leads to a significant exposure of itspositive charge both to the interior of the substrate and to theexternal environment. Since the funnel-like structure of the CDis preserved, access to the cavity through the open secondaryface remains facile. Hence, the aCD-M+ complex has the po-tential to constitute a specific inclusion substrate for anionicguests (e.g. carboxylates),[24] with the cation acting as dockingsite for the negatively charged group of the guest anion. Alsoimportantly, the exposure of the cation to the exterior of thecyclodextrin allows for its partial solvation and can be expect-ed to contribute to the stability of the complex in aqueous orpolar environments. It should be remarked that hydration ef-fects are likely to alter the relative stability of the aCD-M+

complexes with respect to the gas-phase. On the one hand,the larger cations bound to the cyclodextrin are more exposedto the solvent, so that efficient hydration can take place with-out alteration of the optimum metal-oxygen coordination in

the complex. On the other hand, the hydration energy of theuncomplexed cations is greater for the smaller cations. Both ofthese effects tend to enhance the stability of the complexeswith increasing cation size. The situation is in fact similar tothe one found for the inclusion complexes of other macrocy-cles.[39, 41] The balance between solvation effects and CD-M+ in-trinsic binding can be expected to shift the preference of cy-clodextrins towards the cations heavier than Li+ in polar solu-tions. This would be consistent with the greater abundances ofthe complexes with Na+ and K+ among the alkali series arisingfrom the electrospray of aqueous solutions of cyclodextrinsand other oligosaccharides. Extension of the present experi-ments to microsolvated CD-M+ ···n H2O clusters, probing alsothe O�H stretching vibrational modes of the substrate and sol-vent, could provide relevant information on this topic.

Acknowledgements

The research leading to these results is supported by the Pro-gramme Consolider-Ingenio of the Ministry of Science and Inno-vation of Spain (MICINN), through project CSD2009-00038. B.M.H.acknowledges funding through projects P07-FQM-02600 (Juntade Andalucia-FEDER) and CTQ2012-32345 (MICINN). The presenttransnational collaboration was supported by the EuropeanCommunity Seventh Framework Programme (FP7/2007–2013,Grant No. 226716). The skillful assistance of Dr. B. Redlich andothers of the FELIX staff is gratefully acknowledged.

Keywords: alkali-metal cations · cyclodextrin · inclusioncomplexes · laser spectroscopy · MALDI

[1] J. Szejtli, Chem. Rev. 1998, 98, 1743 – 1753.[2] M. V. Rekharsky, Y. Inoue, Chem. Rev. 1998, 98, 1875 – 1917.[3] L. X. Song, L. Bai, X. M. Xu, J. He, S. Z. Pan, Coord. Chem. Rev. 2009, 253,

1276 – 1284.[4] E. M. M. del Valle, Proc. Biochem. 2004, 39, 1033 – 1046.[5] K. Uekama, F. Hirayama, T. Irie, Chem. Rev. 1998, 98, 2045 – 2076.[6] A. R. Hedges, Chem. Rev. 1998, 98, 2035 – 2044.[7] M. E. Davis, M. E. Brewster, Nat. Rev. Drug Discovery 2004, 3, 1023 – 1035.[8] J. P. Schermann, Spectroscopy and Modelling of Biomolecular Building

Blocks, Elsevier, Amsterdam, 2008.[9] C. B. Lebrilla, Acc. Chem. Res. 2001, 34, 653 – 661.

[10] G. Grigorean, S. Gronert, C. B. Lebrilla, Int. J. Mass Spectrom. 2002, 219,79 – 87.

[11] S. Ahn, J. Ramirez, G. Grigorean, C. B. Lebrilla, J. Am. Soc. Mass Spectrom.2001, 12, 278 – 287.

[12] S. Ahn, X. Cong, C. B. Lebrilla, S. Gronert, J. Am. Soc. Mass Spectrom.2005, 16, 166 – 175.

[13] Z. Yu, M. Cui, C. Yan, F. Song, Z. Liu, S. Liu, G. Zhang, H. Zhang, J. MassSpectrom. 2007, 42, 1106 – 1110.

[14] X. Ma, Z. Wei, X. Xiong, Y. Jiang, J. He, S. Zhang, X. Fang, X. Zhang, Ta-lanta 2012, 93, 252 – 256.

[15] Z. Li, E. P. A. Couzijn, X. Zhang, J. Phys. Chem. B 2012, 116, 943 – 950.[16] Y. Fu, L. Liu, Q. Guo, J. Inclusion Phenom. Macrocyclic Chem. 2002, 43,

223 – 229.[17] L. Liu, Q. X. Guo, J. Incl. Phenom. Macrocyclic Chem. 2004, 50, 95 – 103.[18] C. P. A. Anconi, C. S. Nascimento, J. Fedoce-Lopes, H. F. Dos Santos, W. B.

DeAlmeida, J. Phys. Chem. A 2007, 111, 12127 – 12135.[19] A. Karpfen, E. Liedl, W. Snor, P. Weiss-Greiler, H. Viernstein, P. Wolschann,

J. Inclusion Phenom. Macrocyclic Chem. 2007, 57, 35 – 38.[20] A. Karpfen, E. Liedl, W. Snor, H. Viernstein, P. Weiss-Greiler, P. Wolschann,

Monatsh. Chem. 2008, 139, 363 – 371.

Figure 7. Comparison of the IRMPD spectra measured for the isolated aCD-M+ (M = Li, K, Cs) complexes with the B3LYP IR spectra predicted for thelowest energy conformers depicted in Figure 4. The intensity in the regionof the strong C�O stretching band (970–1200 cm�1) has been scaled in allspectra by a factor of 0.5 for a better visualization of the weaker bands.



http://dx.doi.org/10.1021/cr970022chttp://dx.doi.org/10.1021/cr970022chttp://dx.doi.org/10.1021/cr970022chttp://dx.doi.org/10.1021/cr970015ohttp://dx.doi.org/10.1021/cr970015ohttp://dx.doi.org/10.1021/cr970015ohttp://dx.doi.org/10.1016/j.ccr.2008.08.011http://dx.doi.org/10.1016/j.ccr.2008.08.011http://dx.doi.org/10.1016/j.ccr.2008.08.011http://dx.doi.org/10.1016/j.ccr.2008.08.011http://dx.doi.org/10.1016/S0032-9592(03)00258-9http://dx.doi.org/10.1016/S0032-9592(03)00258-9http://dx.doi.org/10.1016/S0032-9592(03)00258-9http://dx.doi.org/10.1021/cr970025phttp://dx.doi.org/10.1021/cr970025phttp://dx.doi.org/10.1021/cr970025phttp://dx.doi.org/10.1021/cr970014whttp://dx.doi.org/10.1021/cr970014whttp://dx.doi.org/10.1021/cr970014whttp://dx.doi.org/10.1038/nrd1576http://dx.doi.org/10.1038/nrd1576http://dx.doi.org/10.1038/nrd1576http://dx.doi.org/10.1021/ar980125xhttp://dx.doi.org/10.1021/ar980125xhttp://dx.doi.org/10.1021/ar980125xhttp://dx.doi.org/10.1016/S1387-3806(02)00558-4http://dx.doi.org/10.1016/S1387-3806(02)00558-4http://dx.doi.org/10.1016/S1387-3806(02)00558-4http://dx.doi.org/10.1016/S1387-3806(02)00558-4http://dx.doi.org/10.1016/S1044-0305(00)00220-8http://dx.doi.org/10.1016/S1044-0305(00)00220-8http://dx.doi.org/10.1016/S1044-0305(00)00220-8http://dx.doi.org/10.1016/S1044-0305(00)00220-8http://dx.doi.org/10.1016/j.jasms.2004.10.007http://dx.doi.org/10.1016/j.jasms.2004.10.007http://dx.doi.org/10.1016/j.jasms.2004.10.007http://dx.doi.org/10.1016/j.jasms.2004.10.007http://dx.doi.org/10.1002/jms.1249http://dx.doi.org/10.1002/jms.1249http://dx.doi.org/10.1002/jms.1249http://dx.doi.org/10.1002/jms.1249http://dx.doi.org/10.1016/j.talanta.2012.02.029http://dx.doi.org/10.1016/j.talanta.2012.02.029http://dx.doi.org/10.1016/j.talanta.2012.02.029http://dx.doi.org/10.1016/j.talanta.2012.02.029http://dx.doi.org/10.1021/jp210329ahttp://dx.doi.org/10.1021/jp210329ahttp://dx.doi.org/10.1021/jp210329ahttp://dx.doi.org/10.1023/A:1021242310024http://dx.doi.org/10.1023/A:1021242310024http://dx.doi.org/10.1023/A:1021242310024http://dx.doi.org/10.1023/A:1021242310024http://dx.doi.org/10.1021/jp0762424http://dx.doi.org/10.1021/jp0762424http://dx.doi.org/10.1021/jp0762424http://dx.doi.org/10.1007/s10847-006-9166-2http://dx.doi.org/10.1007/s10847-006-9166-2http://dx.doi.org/10.1007/s10847-006-9166-2http://dx.doi.org/10.1007/s00706-007-0811-2http://dx.doi.org/10.1007/s00706-007-0811-2http://dx.doi.org/10.1007/s00706-007-0811-2www.chemphyschem.org

[21] V. Jim�nez, J. B. Alderete, J. Phys. Chem. A 2008, 112, 678 – 685.[22] R. V. Pinjari, V. V. Gobre, S. P. Gejji, J. Mol. Struct. THEOCHEM 2010, 957,

77 – 83.[23] J. K. Khedkar, V. V. Gobre, R. V. Pinjari, S. P. Gejji, J. Phys. Chem. A 2010,

114, 7725 – 7732.[24] I. V. Terekhova, A. O. Romanova, R. S. Kumeev, M. V. Fedorov, J. Phys.

Chem. B 2010, 114, 12607 – 12613.[25] S. Lee, T. Wyttenbach, M. T. Bowers, Int. J. Mass Spectrom. Ion Processes

1997, 167, 605 – 614.[26] K. P. Madhusudanan, J. Mass Spectrom. 2003, 38, 409 – 416.[27] S. Reale, E. Texeid�, F. de Angelis, Ann. Chim. 2005, 95, 375 – 381.[28] S. S. Choi, H. M. Lee, S. Jang, J. Shin, Int. J. Mass Spectrom. 2009, 279,

53 – 58.[29] M. T. Cancilla, S. G. Penn, J. A. Carroll, C. B. Lebrilla, J. Am. Chem. Soc.

1996, 118, 6736 – 6745.[30] S. Trimpin, A. Rouhanipour, R. Az, H. J. R�der, K. Mllen, Rapid Commun.

Mass Spectrom. 2001, 15, 1364 – 1373.[31] A. R. Hortal, P. Hurtado, B. Mart�nez-Haya, A. Arregui, L. BaÇares, J. Phys.

Chem. B 2008, 112, 8530 – 8535.[32] D. Oepts, A. F. G. van der Meer, P. W. van Amersfoort, Infrared Phys. Tech-

nol. 1995, 36, 297 – 308.[33] N. C. Polfer, J. Oomens, Phys. Chem. Chem. Phys. 2007, 9, 3804 – 3817.[34] Y. Xie, C. B. Lebrilla, Anal. Chem. 2003, 75, 1590 – 1598.[35] Accelrys Materials Studio 4.4. Accelrys, Inc. San Diego CA, USA.[36] D. Andrae, U. H�ußermann, M. Dolg, H. Stoll, H. Preuß, Theor. Chim.

Acta 1990, 77, 123 – 141.

[37] Gaussian 09 (Revision A.1), M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E.Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Men-nucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian,A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara,K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.Kitao, H. Nakai, T. Vreven, J. A. Montgomery, Jr. , J. E. Peralta, F. Ogliaro,M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Ko-bayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyen-gar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B.Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin,K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg,S. Dapprich, A. D. Daniels, . Farkas, J. B. Foresman, J. V. Ortiz, J. Cio-slowski, D. J. Fox, Gaussian, Inc. , Wallingford CT, 2009.

[38] C. Wang, S. H. Yang, J. Wang, P. Kroll, K. A. Schug, D. W. Armstrong, Int.J. Mass Spectrom. 2010, 291, 118 – 124.

[39] B. Mart�nez-Haya, P. Hurtado, A. R. Hortal, S. Hamad, J. D. Steill, J.Oomens, J. Phys. Chem. A 2010, 114, 7048 – 7054.

[40] P. Hurtado, A. R. Hortal, F. G�mez, S. Hamad, B. Mart�nez-Haya, Phys.Chem. Chem. Phys. 2010, 12, 13752 – 13758.

[41] D. Feller, J. Phys. Chem. A 1997, 101, 2723 – 2731.[42] S. Z. Fairchild, C. F. Bradshaw, W. Su, S. K. Guharay, Appl. Spectrosc. 2009,

63, 733 – 741.

Received: September 27, 2012Published online on December 13, 2012


http://dx.doi.org/10.1021/jp073011ohttp://dx.doi.org/10.1021/jp073011ohttp://dx.doi.org/10.1021/jp073011ohttp://dx.doi.org/10.1016/j.theochem.2010.07.010http://dx.doi.org/10.1016/j.theochem.2010.07.010http://dx.doi.org/10.1016/j.theochem.2010.07.010http://dx.doi.org/10.1016/j.theochem.2010.07.010http://dx.doi.org/10.1021/jp102304jhttp://dx.doi.org/10.1021/jp102304jhttp://dx.doi.org/10.1021/jp102304jhttp://dx.doi.org/10.1021/jp102304jhttp://dx.doi.org/10.1021/jp1063512http://dx.doi.org/10.1021/jp1063512http://dx.doi.org/10.1021/jp1063512http://dx.doi.org/10.1021/jp1063512http://dx.doi.org/10.1016/S0168-1176(97)00105-5http://dx.doi.org/10.1016/S0168-1176(97)00105-5http://dx.doi.org/10.1016/S0168-1176(97)00105-5http://dx.doi.org/10.1016/S0168-1176(97)00105-5http://dx.doi.org/10.1002/jms.453http://dx.doi.org/10.1002/jms.453http://dx.doi.org/10.1002/jms.453http://dx.doi.org/10.1002/adic.200590057http://dx.doi.org/10.1002/adic.200590057http://dx.doi.org/10.1002/adic.200590057http://dx.doi.org/10.1016/j.ijms.2008.10.007http://dx.doi.org/10.1016/j.ijms.2008.10.007http://dx.doi.org/10.1016/j.ijms.2008.10.007http://dx.doi.org/10.1016/j.ijms.2008.10.007http://dx.doi.org/10.1021/ja9603766http://dx.doi.org/10.1021/ja9603766http://dx.doi.org/10.1021/ja9603766http://dx.doi.org/10.1021/ja9603766http://dx.doi.org/10.1002/rcm.372http://dx.doi.org/10.1002/rcm.372http://dx.doi.org/10.1002/rcm.372http://dx.doi.org/10.1002/rcm.372http://dx.doi.org/10.1021/jp802089rhttp://dx.doi.org/10.1021/jp802089rhttp://dx.doi.org/10.1021/jp802089rhttp://dx.doi.org/10.1021/jp802089rhttp://dx.doi.org/10.1016/1350-4495(94)00074-Uhttp://dx.doi.org/10.1016/1350-4495(94)00074-Uhttp://dx.doi.org/10.1016/1350-4495(94)00074-Uhttp://dx.doi.org/10.1016/1350-4495(94)00074-Uhttp://dx.doi.org/10.1039/b702993bhttp://dx.doi.org/10.1039/b702993bhttp://dx.doi.org/10.1039/b702993bhttp://dx.doi.org/10.1021/ac026009whttp://dx.doi.org/10.1021/ac026009whttp://dx.doi.org/10.1021/ac026009whttp://dx.doi.org/10.1007/BF01114537http://dx.doi.org/10.1007/BF01114537http://dx.doi.org/10.1007/BF01114537http://dx.doi.org/10.1007/BF01114537http://dx.doi.org/10.1016/j.ijms.2010.01.014http://dx.doi.org/10.1016/j.ijms.2010.01.014http://dx.doi.org/10.1016/j.ijms.2010.01.014http://dx.doi.org/10.1016/j.ijms.2010.01.014http://dx.doi.org/10.1039/c0cp00595ahttp://dx.doi.org/10.1039/c0cp00595ahttp://dx.doi.org/10.1039/c0cp00595ahttp://dx.doi.org/10.1039/c0cp00595ahttp://dx.doi.org/10.1021/jp9700185http://dx.doi.org/10.1021/jp9700185http://dx.doi.org/10.1021/jp9700185www.chemphyschem.org

cations in a molecular funnel: vibrational spectroscopy of...

Documents