CsE

SHC

a

AA

KEVTCSU

1

ochona�aowmDfc

ac

0d

Journal of Chromatography A, 1216 (2009) 1232–1240

Contents lists available at ScienceDirect

Journal of Chromatography A

journa l homepage: www.e lsev ier .com/ locate /chroma

hromatographic and spectroscopic studies on the chiral recognition ofulfated �-cyclodextrin as chiral mobile phase additivenantiomeric separation of a chiral amine

hengli Ma, Sherry Shen, Nizar Haddad, Wenjun Tang, Jing Wang,eewon Lee, Nathan Yee, Chris Senanayake, Nelu Grinberg ∗

hemical Development Department, Boehringer Ingelheim Pharmaceuticals, Inc., 900 Ridgebury Road, Ridgefield, CT 06877-0368, USA

r t i c l e i n f o

rticle history:vailable online 11 December 2008

eywords:nantiomeric separationCD

a b s t r a c t

A fast enantiomeric separation of a chiral aromatic amine was achieved, using ultra high pressure liquidchromatography and highly sulfated beta-cyclodextrin (S-�-CD) as a chiral additive in the mobile phase.The stationary phase consisted of a core–shell support with a particle size of 2.7 �m. Under these con-ditions the base-line separation was obtained within 2.5 min. The influence of the concentration of theadditive, along with the thermodynamics of the separation, was studied. Vibrational circular dichroism

hermodynamicsonformational changesulfated beta-cyclodextrinltra high pressure liquid chromatography

(VCD) spectroscopy was applied to assess the absolute configuration of the two enantiomeric analytes,as well as the interaction of these enantiomers with the S-�-CD. The VCD results revealed that S-�-CDundergoes a temperature-induced conformational change. Further, VCD experiments indicate that theinteractions of the two enantiomers with the S-�-CD occur through an inclusion of the aromatic part ofthe analyte, as well as through electrostatic interaction between the protonated amine and the sulfategroups located at the narrow part of the S-�-CD. Molecular mechanics calculations performed according

sisten

to the VCD results are con

. Introduction

Cyclodextrins (CDs) are products of the enzymatic degradationf starch [1]. If the amylose fraction of starch is degraded by glu-osyltransferases, one or several turns of the amylose helix areydrolyzed off and their ends are joined together, producing cyclicligosaccharides called cyclodextrins. Because these enzymes areot specific, the hydrolysis produces a number of CDs with a vari-ble number of sugar units [2]. The most abundant are �-, �-, and-cyclodextrin (�-CD, �-CD and �-CD, respectively) with six, sevennd eight glucose rings, respectively. They are homochiral cyclicligosaccharides and have a doughnut-like annular structure withide and narrow hydrophilic ends, due to their secondary and pri-ary hydroxyl groups located at the two ends of the annulus [1,3].ue to these functional groups, CDs present an attractive platform

or chemical modifications starting from esters and ending with

arbamates [2–6].

Much of the interest in native and modified CDs stems from theirbility to form host–guest complexes by including large classes ofompounds inside their annuli. The application of cyclodextrin to

∗ Corresponding author.E-mail address: nelu.grinberg@boehringer-ingelheim.com (N. Grinberg).

021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2008.12.016

t with experimental data, providing further evidence of these interactions.© 2008 Elsevier B.V. All rights reserved.

the separation of enantiomers is a natural extension of their inclu-sion properties. Their chiral discrimination ability results from thedifference in the stability constant of each enantiomer upon com-plexation with the CDs. The interaction of each enantiomer occursthrough inclusion inside the CD annuli, along with hydrogen bond-ing, dipole–dipole interaction, etc., through the functional groupslocated at the wider side of the CD [2].

Separation of enantiomers using chiral mobile phases is a tech-nique which was widely used in the 1980s and early 1990s, dueto the lack of chiral stationary phases. The technique has severaladvantages compared to chiral stationary phases. For instance, itemploys achiral stationary phases, which are less expensive andmore rugged than a chiral stationary phase. At the same time itallows more flexibility because a column can be used in conjunc-tion with several additives at the same time or the chiral additivecan be eluted out of the column with a strong solvent and a subse-quent additive can be used. The disadvantage of the method stemsfrom the existence of a secondary chemical equilibrium which has agreat contribution to the band broadening of the two enantiomericanalytes and makes the mechanism of separation difficult to under-

stand.

The introduction of core–shell technology particles with all theadvantages of particles in the range of 2.7 �m particle size, alongwith a porous shell of 0.5 �m [7–10], revitalized chiral separationwith chiral mobile phases. Fast kinetics of mass transfer, along

gr. A 1

wstae

o[tdpitc

orbmsacltptittwttuc

amtmfaTb

2

2

As(uApppcmag

3fct

scaled by 0.97 and calculated IR and VCD intensities were con-verted to Lorentzian bands for comparison to experiment. Thefinal calculated spectra were composed by averaging the spectraof each conformer based on a Boltzmann distribution of corre-sponding energies. The experimental VCD and calculated VCD were

S. Ma et al. / J. Chromato

ith a fast distribution of the analyte/complex, minimizes theecondary chemical equilibrium and consequently brings substan-ial improvement to the band broadening. Such stationary phases,long with ultra high pressure HPLC, bring improved timeliness andfficiency to chiral separations with chiral mobile phases.

Sulfated �-CD (S-�-CD) represents a class of anionic derivativesf CDs, and has been used extensively in capillary electrophoresis11,12]. Stalcup and Gahm [13] used this selector as a chiral sta-ionary phase to enantioresolve 33 compounds, amongst which sixid not have an amine functional group. There are very few exam-les of the use of S-�-CD as an additive in chiral mobile phases. For

nstance, Ameyibor and Stewart [14] used S-�-CD as an additive inhe mobile phase for the enantiomeric separation of some selectedhiral drugs.

Vibrational circular dichroism (VCD) is an analog techniquef electronic circular dichroism extending to the mid-infraredegion where abundant information on molecular vibrations cane observed. By measuring the different interactions of a chiralolecule with left vs. right circularly polarized light, VCD shows

pecial sensitivity to chiral substances. VCD has become a uniquend routine tool to determine the absolute configuration of chiralompounds in conjunction with ab initio quantum mechanic calcu-ation [15,16]. In addition, VCD is extensively used to characterizehe structure of optically active macromolecules such as proteins,eptides, and polysaccharide and derivatives [17–20]. Comparedo its parent IR spectra, VCD shows more resolution with eitherts positive or negative peaks and hence reveals additional struc-ural information on chiral substances. Therefore, VCD can be usedo detect structural changes of chiral molecules, their interactionsith other substances, and resulting conformational changes on

he interacting molecules. For example, Setnicka et al. [21] studiedhe interaction between cyclodextrins with aromatic compoundssing VCD. Bose and Polavarapu [22] applied VCD to investigate theovalent binding between copper ions and cyclodextrins.

In this paper, we present the enantiomeric separation of anmine using core–shell technology with sulfate �-CD as the chiralobile phase selector and ultra high pressure LC. We first assess

he influence of the concentration of the additive and the ther-odynamics of the separation on chromatographic conditions. We

urther explore the molecular interactions between enantiomersnd S-�-CD using VCD spectroscopy as well as molecular modeling.he determination of the absolute configuration of each enantiomery VCD is presented as well.

. Experimental

.1. Chromatography

The chromatographic experiments were performed on a Waterscquity UPLC system equipped with a vacuum degasser, a high pres-ure binary pump, a column oven, and a photo diode array detectorWaters, Milford, MA, USA). All the separations were carried outsing two Halo 4.6 mm×150 mm columns (Halo C8 and Halo C18,dvanced Material Technology, Wilmington, DE, USA) with 2.7 �marticle size. The highly sulfated �-CD, acetonitrile and water wereurchased from Sigma–Aldrich (Milwaukee, WI, USA). The mobilehase consisted of a solvent mixture of 15% acetonitrile in waterontaining different amounts of S-�-CD, depending on the experi-ent. The mobile phase was premixed prior to the chromatography

nd pumped through the system at a flow rate of 1.5 ml/min whichenerated a back pressure of ∼11 000 p.s.i.

The column was equilibrated with mobile phase for at least0 min prior to each experiment. The same procedure was per-ormed each time the column temperature or the mobile phase washanged. The column void time, to, was determined by measuringhe first base-line perturbation. Each experiment was carried out

216 (2009) 1232–1240 1233

for three times at each mobile phase composition and temperaturechange. The retention factor, k′, was obtained using: k′ = (tR− to)/to,where tR is the retention time of the analyte. The selectivity factor,˛, was calculated as: ˛ = k′2/k′1, where k′1 and k′2 are the retentionfactors of the first and second eluted enantiomer, respectively.

2.2. Analytes

The two enantiomers of the chiral amine hydrochloride salt ofcompound 1 used in this study were synthesized in house. 1H NMR(400 MHz, CDCl3) ı: 11.91 (bs, 1H), 7.3 (m, 5H), 3.48 (m, 1H), 3.2(m, 2H), 2.8 (bd, 4H), 1.45 (d, J = 7.0 Hz, 3H), 1.28 (m, 6H). 13C NMR(125 MHz, CDCl3) ı: 143.0, 129.2, 127.5, 126.9, 58.1, 35.8, 22.1, 10.0.For chromatographic experiments, the samples were dissolved inmethanol at a concentration of ∼0.1 mg/ml.

2.3. Experimental VCD

The VCD spectra of the two enantiomers of compound 1 (Fig. 1)were measured in a CDCl3 solution at a concentration of 50 mg/mlin a 200 �m liquid IR cell. The spectra were collected for 4 h with4 cm−1 resolution on a ChiralIR VCD spectrometer (BioTools, Jupiter,FL, USA) equipped with a dual-PEM bench. The final VCD spectrawere corrected by subtracting the solvent (CDCl3) spectra measuredat the same conditions.

For the measurements of S-�-CD, the sample was dissolved ina H2O/acetonitrile-d3 (85/15, v/v) solution at the concentration of500 mg/ml. The complexes were made by mixing S-�-CD with eachenantiomer at a 1:1 molar ratio in a solvent of H2O/ACN-d3 (85/15,v/v) mixture that was the same as those used in chromatographicexperiments.

The VCD spectra of S-�-CD at different temperatures and itscomplexes with each enantiomer were collected using 6-�m bio-cell windows (BioTools) for 12 h with an 8 cm−1 resolution. Thenoise level was less than 1/100 of VCD signal. The final spectra werecorrected by subtracting the corresponding solvent’s spectra.

2.4. Calculated VCD and molecular modeling

The procedure for the calculation of theoretical VCD spectrabegins with the construction of a three-dimensional (3D) struc-ture of the one arbitrary enantiomer (we selected the S-enantiomerin this case) in HyperChem (Hypercube, Gainesville, FL, USA). The3D structure was subsequently used for conformer search at themolecular mechanic level to obtain all the conformers with low-est energies. The lowest energy conformers were selected withinan energy range of 5 kcal/mol. The final optimized geometriesand vibrational frequencies for IR and VCD intensities were deter-mined by using Gaussian 03 (Gaussian, Wallingford, CT, USA)at the density functional theory (DFT) level with a B3LYP func-tional and 6-31G(d) basis set [23]. The calculated frequencies were

Fig. 1. Structure of Compound 1: chiral center was labeled with asterisk (*).

1 gr. A 1

fit

eoaeaS

3

TcassiCcocbg

3

gcaddSes

cVtowo

Fcest

234 S. Ma et al. / J. Chromato

nally compared for the determination of the absolute configura-ion.

All computations on S-�-CD and its complexes with eachnantiomer were performed using Hyperchem software. Structureptimizations and energy minimizations were implemented usingmolecular mechanics force field, MM+ method. Docking of the

nantiomers was initially done to maximize the electrostatic inter-ction and hydrophobic interaction between the enantiomer and-�-CD.

. Results and discussion

The structure of the compound under study is presented in Fig. 1.he compound has a tertiary amine and an aromatic moiety. Thehiral center is in the alpha position to the aromatic. The tertiarymine group has three ethyl groups attached. It is expected thatteric hindrance will be a major factor impeding interaction at theseites. Because the structure of the compound shows few sites ofnteraction, we pursued the separation using a porous shell C8 and18 reversed-phase stationary phase along with a mobile phaseontaining S-�-CD as an additive. The tertiary amine with a pKa

f ∼10.8 is expected to be protonated at lower pHs. Under theseonditions the interactions with the S-�-CD could be assessed asoth electrostatic interactions between the amine and the sulfateroups of cyclodextrins and as inclusion inside the cavity of S-�-CD.

.1. Determination of the absolute configuration

In order to assess the order of elution during the chromato-raphic separation, the first step was to assess the absoluteonfiguration of each enantiomer of this compound. Here, wepplied experimental VCD and theoretical VCD calculations toetermine the absolute configuration of the two enantiomers. Theetails about the measurements and calculation were described inection 2. Based on the lowest free energies of the six conform-rs (see Fig. S1 in the Supplemental material), the final calculatedpectra were composed using a Boltzmann average.

In Fig. 2 the calculated VCD spectrum for the S-enantiomer isompared to those measured for both enantiomers. The strongest

CD bands correlate well with the S absolute configuration, while

hey are opposite to the R-enantiomer. The agreement betweenbserved and calculated VCD intensities for the model structureith an S configuration provides firm evidence for the assignment

f the absolute configuration.

ig. 2. Comparison of experimental VCD with calculated VCD of compound 1: thealculated VCD is labeled as solid curve and offset for clarity; the enantiomer thatluted first (enantiomer R) is labeled as solid curve; the enantiomer that elutedecond (enantiomer S) is labeled as dashed curve. The major peaks were correlatedogether through a dotted line between experimental and calculated VCD for clarity.

216 (2009) 1232–1240

3.2. Enantiomeric separation using S-ˇ-CD

3.2.1. Influence of concentration of sulfated ˇ-CD on theseparation

The presence of cyclodextrin as a chiral additive in the mobilephase induces a secondary chemical equilibrium. In general terms,under chromatographic conditions a chiral additive (L) is constantlypartitioned between the mobile phase and the stationary phaseaccording to the equilibrium:

LmK1←→Ls (1)

where the subscripts m and s refer to the species present in themobile phase and in the stationary phase, respectively. When aracemic mixture is introduced into the chromatographic system,each of the enantiomers (E) will undergo the following equilibria[2]:

Formation of diastereomers in mobile phase

Em + LmK2←→ELm (2)

Distribution of diastereomeric complex between mobile phase andstationary phase

ELmK3←→ELs (3)

Distribution of analyte between mobile phase and stationary phase

EmK4←→Es (4)

Formation of diastereomers in stationary phase

Es + LsK5←→ELs (5)

where K1, K2, K3, K4 and K5 are the equilibrium constants for eachprocess and can be expressed as:

K1 =[L]s

[L]m(6)

K2 =[EL]m

[E]m × [L]m(7)

K3 =[EL]s

[EL]m(8)

K4 =[E]s

[E]m(9)

K5 =[EL]s

[E]s × [L]s(10)

The overall distribution coefficient, D, for the enantiomerpresent in the chromatographic system is given by:

D = {[E]s + [EL]s}{[E]m + [EL]m}

(11)

which by substitution from Eqs. (6)–(10) leads to:

D = {K4(1+ K1K5[L]m}{1+ K2[L]m}

(12)

The retention factor, k′, is proportional to the distribution coef-ficient and can be written:

k′ = ˚ · D (13)

where k′ is the retention factor and � is the phase ratio. CombiningEq. (13) with (12) we obtain:

k′ = ˚K4(1+ K1K5[L]m)

1+ K2[L]m(14)

gr. A 1

e

(

(

rnCotpfcitrip

rf2

F1t

S. Ma et al. / J. Chromato

According to Eq. (14), two different cases can be primarilyncountered:

A) If the additive L is strongly absorbed in the stationary phases,K1 is very large and becomes a dominant factor in Eq. (14). Thusthe interaction will occur mainly in the stationary phase, andEq. (14) can be simplified as:

k = ˚K4(1+ K1K5[Lm]) (15)

An example of such a case is the separation of amino acidsusing copper complexes of long chain amino acids as chiraladditives under thin-layer chromatographic conditions [24,25].Under these circumstances, the additive (the copper complex)interacts so strongly with the stationary phases that it behaveslike a chiral stationary phase, leading to chiral separations.

B) If K2[L]m is large and K1 becomes fairly small, the interactionswill occur mainly in the mobile phase and Eq. (14) can be writ-ten:

k = ˚K4

1+ K2[L]m(16)

Under such conditions, the ligand interacts weakly with thestationary phase, and the complexation with analytes occursmainly in the mobile phase. Most enantiomeric separationsunder reversed-phase conditions with �-CD as chiral additivesbelong to this case [26–30,14,31–35].

Under the same mobile phase conditions used for the sepa-ation of the two enantiomers of compound 1, S-�-CD exhibitsearly no retention (elutes out close to the dead volume) on both8 and C18 stationary phases. This indicates that the adsorptionf this chiral additive on both stationary phases is minimal andhe complexation with the chiral additive occurs in the mobilehase. Since K4 is the same for both enantiomers, the limiting

actor that determines the enantiomeric separation at a certainoncentration of chiral additive will be K2, which represents thenteraction (affinity) between each enantiomer and the chiral addi-ive in the mobile phase. In addition, according to Eq. (16) theetention factor (k′) of each enantiomer will decrease upon increas-ng the concentration of the chiral additive [L] in the mobilehase.

In order to assess the influence of the chiral additive on the sepa-ation, the concentration of S-�-CD was varied in the mobile phaserom zero to 2.0 mM. An example of base-line separation under.5 min is presented in Fig. 3.

ig. 3. A chromatogram of the separation of compound 1 on UPLC: flow rate,.5 ml/min; mobile phase, 2 mM S-�-CD in H2O/acetonitrile (85/15, v/v); tempera-ure, 20 ◦C; column, Halo C8 with 2.7 �m particle size, 4.6 mm×150 mm dimension.

216 (2009) 1232–1240 1235

The influence of the concentration of the additive on the reten-tion factor (k′) of the two enantiomers, and on the selectivity, ˛, ispresented in Fig. 4a–c.

The data of Fig. 4 indicate that, upon increasing the concen-tration of the additive in the mobile phase, the retention factordecreases drastically from zero to 0.25 mM; it then remains con-stant for the rest of the range of the concentrations used in the study.Such results indicate that the complexation between the sulfated�-CD and each enantiomer occurs in the mobile phase, which is inaccordance with the predicted behavior of Eq. (16).

The enantioselectivity (˛) for a pair of enantiomers in the pres-ence of a chiral mobile phase additive can be written:

˛ = k′Rk′S= DR

DS= K4(R)(1+ K1(R)K5(R))

K4(S)(1+ K1(S)K5(S))× 1+ K2(S)[L]m

1+ K2(R)[L]m(17)

with k′R > k′S , and DR > DS. The subscripts R and S refer to the R- andS-enantiomers, respectively. The first term of Eq. (17) represents thecomplexation of the enantiomeric analytes in the stationary phase,while the second represents complexation in the mobile phase. Ifthe complexation of the two enantiomers with the chiral additiveoccurs in the mobile phase, Eq. (17) reduces to:

˛ = k′Rk′S= 1+ K2(S)[L]m

1+ K2(R)[L]m(18)

From Eq. (18), it can be clearly seen that if the concentrationof chiral additive in the mobile phase ([L]m) remains constant, theenantioselectivity essentially stems from the different affinity con-stants, (K2) of the two enantiomers with the chiral additive in themobile phase. Furthermore, increasing the concentration of the chi-ral additive in the mobile phase will lead to an increase in theenantioselectivity.

Indeed, the enantioselectivity, ˛, increases with the increase inconcentration of the chiral additive, as shown in Fig. 4c. Interest-ingly, the enantioselectivity is higher on the C8 stationary phasecompared to the C18 stationary phase. In order to understand thephenomenon we explored Fourier transformed attenuated totalreflection infrared spectroscopy (FT-ATR-IR) of the two stationaryphases. The spectra of the two stationary phases (see Fig. S2 inSupplemental material) show that while the bands at 2922 and2853 cm−1 (corresponding to the –CH2– stretching mode) have thesame absorbance, the band at 2962 cm−1 (corresponding to the–CH3 stretching mode), has smaller intensities for C18 comparedto C8. This indicates (Fig. S2) that the density of the C18 ligands onthe RP stationary phase is smaller than that of the C8. As a con-sequence there is a possibility that there may be residual silanolswhich may produce an achiral interaction which impedes the enan-tioselectivity on the C18 stationary phase.

3.2.2. Influence of temperature on separationFor an isobaric, isothermal chromatographic process, the effect

of temperature on the capacity factor can be expressed by the van’tHoff equation [36]:

ln k′ = −�H◦

RT+ �S◦

R+ ln ˚ (19)

where �H◦ and �S◦ are the standard enthalpy and entropy,respectively, of solute transfer from mobile phase to the station-ary phase, and ˚ is the phase ratio. According to Eq. (19) aplot of ln k′ vs. 1/T should yield a straight line with a slope of�H◦/R and an intercept of (�S◦/R + ln ˚). Deviation from linear-

ity can be indicative of a conformational change of the chiral phase[37–39].

The separation factor for a given pair of enantiomers (˛ = k′2/k′1)is a measure of the enantioselectivity and it represents the differ-ence in the free energy of the interactions of the two enantiomers

1236 S. Ma et al. / J. Chromatogr. A 1216 (2009) 1232–1240

tion f

w

l

−

Fig. 4. Influence of the concentration of S-�-CD on reten

ith the chiral phase.

n ˛ = −��G◦

RT= −��H◦

RT+ ��S◦

R(20)

A plot of ln ˛ vs. 1/T will yield a straight line with a slope of��H◦/R and an intercept of ��S◦/R, providing that the enan-

Fig. 5. van’t Hoff plots in k′ for the two enantiomer

actor, (a) Halo C8, (b) Halo C18 and selectivity factor (c).

tioselective interactions do not change over the temperature range

of the study.

In our study the temperature was varied between 10 and 60 ◦C.Plots of ln k′ vs. 1/T for the separation of the two enantiomers ofcompound 1 on both stationary phases, C8 and C18, are presentedin Fig. 5a and b.

s: (a) Halo C 8 column; (b) Halo C18 column.

S. Ma et al. / J. Chromatogr. A 1216 (2009) 1232–1240 1237

for (a

b1ceecll

a(paobtbetbtr

(a5osd

dttbatWtsat1

3.3. Investigation of enantiomeric recognition using VCDspectroscopy

As mentioned above, the complexation of the two enantiomerswith S-�-CD occurs mainly in the mobile phase. As a consequence

Fig. 6. van’t Hoff plot in selectivity ˛

Both plots of Fig. 5a and b indicated that the relationshipetween ln k′ and the temperature is non-linear. Plots of ln ˛ vs./T are presented in Fig. 6a and b. The plot generated from thehromatographic data on the C18 stationary phase show a lin-ar behavior (r2 > 0.999), indicating that temperature has the sameffect on both enantiomers (Fig. 6b). The plot generated from thehromatographic data on the C8 stationary phase showed non-inear behavior; however it was possible to divide the plot into twoinear regions with a change in slope at ∼40 ◦C (Fig. 6a).

Interestingly, the values for ��H◦ and ��S◦ in the high temper-ture region are almost twice as big as those at lower temperatureFig. 6). Such results may be interpreted as follows: at lower tem-eratures both the cyclodextrin and the enantiomeric analytesre solvated. As the temperature increases, a thermal desolvationf the sulfated �-CD occurs. As a consequence the hydropho-ic interaction between the two enantiomers and the interior ofhe cyclodextrin cavity, as well as the electrostatic interactionetween the tertiary amine of the enantiomers and the SO3

−, arenhanced. Thus, the ��H◦ becomes more negative and the interac-ion becomes entropically unfavorable. At the same time, as ��H◦

ecomes more negative, ��S◦ also becomes more negative inhe high temperature region (compared to the lower temperatureegion), as a result of enthalpy–entropy compensation.

In order to understand the change in slope in the van’t Hoff plotsFig. 6a), we studied the behavior of S-�-CD as a function of temper-ture. Thus, we performed VCD experiments using S-�-CD at 15 and5 ◦C, corresponding to two temperatures in the two linear regionsf the plot. The experiments were performed in the presence of aolvent mixture which mimics the mobile phase composition useduring the temperature study. The results are shown in Fig. 7.

The VCD spectra of Fig. 7 show that increasing temperature pro-uces a band broadening of the peak at 1366 cm−1 (correspondingo the free OH on the �-CD). The peak at 1300 cm−1 (correspondingo the free secondary OH) has a red shift toward a lower wave num-er when the temperature is increased from 15 to 55 ◦C. The peakt 1230 cm−1 (corresponding to the symmetric stretching vibra-ions of SO3

− [40]), shows two broad overlapping peaks at 15 ◦C.hen the temperature is increased to 55 ◦C, the two peaks merge

o a single sharper peak. The peaks at 1181 and 1156 cm−1 (corre-ponding to glucose ring vibrations) exhibit peak broadening anddecrease in intensity when the temperature is increased from 15

o 55 ◦C. Similar effects can be seen for the peaks at 1094, 1078 and062 cm−1 (corresponding to the glycosidic C O C bond).

) Halo C 8 and (b) Halo C18 column.

To explain these behaviors it should be considered that sul-fated �-CD is hydrated at lower temperatures. The free –OH groupsof the cyclodextrin are involved in hydrogen bonding with watermolecules. With increasing temperature, partial thermal desolva-tion of the cyclodextrin occurs and the free –OH groups start toform intramolecular H-bonds, as indicated in the VCD spectra by thebroadening and the red shift. At the same time, the sulfate groups,which were also hydrated at lower temperature, suffered a dehy-dration at higher temperature, leading to an electrostatic repulsion.Such behavior produces a peak sharpening in the spectra. More-over, the entire ring of sulfated �-CD relaxes at higher temperatures,looses part of its chirality, and consequently results in the observeddecrease in the intensity of the corresponding VCD bands. Thus, theS-�-CD undergoes a temperature-induced conformational changewhen the temperature is increased. To our knowledge this is the firstevidence of conformational changes of S-�-CD induced by temper-ature.

Fig. 7. VCD spectra of S-�-CD in H2O:acetonitrile (85:15, v/v) solution at differenttemperature: solid curve, 15 ◦C; dashed curve, 55 ◦C. Major peaks and correspondingvibrational modes are labeled in the figure.

1238 S. Ma et al. / J. Chromatogr. A 1

Fe1

wtmeU1pSic

the cyclodextrin and the tertiary amine moiety of the compound.

F(

F(

ig. 8. VCD spectra of R-enantiomer complex with S-�-CD (solid curve) and S-nantiomer complex with S-�-CD (dashed curve): close view in the region of600–1500 cm−1.

e pursued spectroscopic studies under conditions mimickinghe chromatographic mobile phase used in previous experi-

ents. Thus, using VCD, we further investigated the interaction ofach enantiomer of compound 1 with the sulfated cyclodextrin.pon complexation with the sulfated cyclodextrin, the peaks at589 cm−1 (corresponding to –CH aromatic ring stretching of com-

ound 1) show higher intensity for the R-enantiomer, relative to(Fig. 8). At the same time the peak at 1504 cm−1 shows higher

ntensity for the complex with the R-enantiomer compared to theomplex with the S-enantiomer, which also indicates a stronger

ig. 9. Comparison of the VCD spectra of S-�-CD (solid curve), S-�-CD complex with R-ena) close view in the region of 1450–1400 cm−1; (b) close view in the region of 1300–1200

ig. 10. Comparison of the VCD spectra of S-�-CD (solid curve), S-�-CD complex with R-ea) close view in the region of 1200–1100 cm−1 and (b) close view in the region of 1080–1

216 (2009) 1232–1240

interaction of the R-enantiomer with the S-�-CD, relative to theS-enantiomer.

Upon complexation with the two enantiomers, the peaks locatedin the VCD spectra at 1442 and 1423 cm−1 (corresponding tothe –CH2 deformation vibrations from the –CH2–O–SO3 groups)merged to a single broad peak, with a higher intensity for the Rcomplex relative to the S complex (Fig. 9a). This indicates that theR-enantiomer interacts more strongly with the S-�-CD relative tothe S-enantiomer. At the same time, the maximum at 1228 cm−1

(corresponding to the symmetrical stretching vibrations of the pri-mary sulfate groups, –CH2–O–SO3

−,) presents a single maximumfor S-�-CD. Upon interaction with the cationic amine moiety of thetwo enantiomers of compound 1, this peak splits into two over-lapping maxima at 1236 and 1222 cm−1, respectively [40] (Fig. 9b).Moreover, such a split indicates that there are two types of sulfategroups, one at a lower wavenumber (1222 cm−1) which stronglyinteracts with the amine moiety, and the other (1236 cm−1) whichweakly interacts with the amine part of the molecule of compound1. The intensity of the band corresponding to the R-enantiomer ishigher, relative to the S-enantiomer, indicating a stronger interac-tion of R with S-�-CD, relative to S (Fig. 9a and b). These resultssuggest that the interaction between the S-�-CD and compound 1occurs between the sulfate groups located at the narrower side of

The band at 1155 cm−1 (corresponding to the glucose ring vibra-tion) shows higher intensity for the complex with the R-enantiomerrelative to the S-enantiomer (Fig. 10a). At the same time the bandat 1063 cm−1 shows lower intensity relative to the uncomplexed S-

antiomer (dashed curve), and S-�-CD complex with S-enantiomer (dotted curve):cm−1.

nantiomer (dashed curve), and S-�-CD complex with S-enantiomer (dotted curve):050 cm−1.

S. Ma et al. / J. Chromatogr. A 1216 (2009) 1232–1240 1239

-enan

�weic

3e

tI�etlsgoSi

3pStit

t�awcgse�tcif

4

a

[

[[

Fig. 11. Optimized structure of S-�-CD complex with R

-CD for both diastereomeric complexes (Fig. 10b). The complexith the R-enantiomer shows lower intensity relative to the S-

nantiomer. Such behavior indicates that, upon formation of thenclusion complex with the cyclodextrin, the S-enantiomer exer-ises more strains on the cyclodextrin relative to the R-enantiomer.

.4. Molecular modeling of the complex of S-ˇ-CD with eachnantiomer of compound 1

The interactions between S-�-CD and each enantiomer were fur-her assessed using molecular modeling. Based on the VCD andR studies, each enantiomer of compound 1 was docked with S--CD such that the electrostatic interactions occurred betweenach enantiomer and S-�-CD through the protonated nitrogen ofhe compound and the negatively charged –SO3 groups of S-�-CDocated at the narrow part of the annuli. At the same time an inclu-ion occurred through the hydrophobic interaction with the phenylroup and the cavity of S-�-CD. In Fig. 11a and b, we present theptimized structure of the two complexes of each enantiomer with-�-CD. It can be observed that the nitrogen of each enantiomers structurally close to the sulfate groups, at a distance of 3.58 and

.63 ´̊A for the R- and S-enantiomers, respectively. In addition, thehenyl groups of both enantiomers are included in the cavity of-�-CD. The R-enantiomer is slightly closer to the interior wall ofhe annulus than the S-enantiomer. These results suggest that thenteraction between the R-enantiomer and the S-�-CD is strongerhan for the S-enantiomer.

The free energy obtained from molecular simulation showshat the R-enantiomer has a greater potential interaction with S--CD relative to the S-enantiomer. The energy values are 170.89nd 171.63 kcal/mol (1 cal = 4.18 J), respectively, for the complexith the R-enantiomer and the S-enantiomer. These values are

onsistent with the elution order observed in the chromato-raphic experiment, which shows that the R-enantiomer interactstrongly with S-�-CD and consequently elutes first. The differ-nce in the interaction energies between the two enantiomers,E = 0.74 kcal/mol, represents the energetic contribution to enan-

ioselectivity. Although the absolute energy difference does notorrelate well with that obtained from the chromatographic exper-ments, the elution order is consistent with the relative energiesrom the theoretical calculation.

. Conclusion

In conclusion, this report demonstrates the chiral separation ofchiral amine compound using S-�-cyclodextrin as a chiral addi-

[

[[

[

tiomer (a) and S-�-CD complex with S-enantiomer (b).

tive in the mobile phase. The interactions between the analyteswith the chiral mobile phases and the C8 and C18 stationary phaseswere further investigated using chromatographic methods as wellas VCD spectroscopy. It was found that S-�-CD undergoes a confor-mational change with increasing temperature. To our knowledge,this is the first report mentioning such a phenomenon. In addi-tion, VCD results reveal additional molecular-level information forthe complexation of enantiomers with compound 1. Particularly, itwas shown that the aromatic part of the molecule is included insidethe cyclodextrin through hydrophobic interactions with the inte-rior of the macrocycle. In addition, electrostatic interactions occurbetween the tertiary amine portion of the molecule and the sul-fated groups located at the narrow part of the rim of the sulfatedcyclodextrin.

Acknowledgement

We thank Dr. Joseph J. Destefano from Advanced Materials Tech-nology, Inc., (Wilmington, DE, USA) for supplying the Halo C18 andC8 stationary phases.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.chroma.2008.12.016.

References

[1] W. Saenger, Angew. Chem. Int. Ed. 19 (1980) 344.[2] N. Grinberg, T. Burakowski, A.M. Stalcup, in: Y. Kazakevich, R. LoBrutto

(Eds.), HPLC for Pharmaceutical Scientists, Wiley-Interscience, 2007, p.987.

[3] C.J. Easton, S.F. Lincoln, Modified Cyclodextrins. Scaffolds and Templates forSupramolecular Chemistry, Imperial College Press, London, 1999.

[4] I. Ciucanu, W.A. Konig, J. Chromatogr A. 685 (1994) 166.[5] G. Gubitz, M.G. Schmid, Biopharm. Drug Dispos. 22 (2001) 291.[6] W.A. Konig, S. Lutz, G. Wenz, Angew. Chem. Int. Ed. 27 (1988) 979.[7] F. Gritti, A. Cavazzini, N. Marchetti, G. Guiochon, J. Chromatogr. A 1157 (2007)

189.[8] F. Gritti, G. Guiochion, J. Chromatogr. A 1176 (2007) 107.[9] F. Gritti, G. Guiochon, J. Chromatogr. A 1169 (2007) 125.10] N. Marchetti, G. Guioshon, J. Chromatogr. A 1176 (2007) 206.

[11] C.E. Evans, A.M. Stalcup, Chirality 15 (2003) 709.12] S.R. Gratz, A.M. Stalcup, Anal. Chem. 70 (1998) 5166.13] A.M. Stalcup, K.H. Gahm, Anal. Chem. 68 (1996) 1369.

14] E. Ameyibor, J.T. Stewart, J. Liq. Chromatogr. Relat. Technol. 20 (1997)

855.15] T.B. Freedman, X. Cao, R.K. Dukor, L.A. Nafie, Chirality 15 (2003) 743.16] P.J. Stephens, F.J. Devlin, J.-J. Pan, Chirality 20 (2008) 643.

[17] T.B. Freedman, L.A. Nafie, T.A. Keiderling, Biopolymers 37 (1995) 265.18] T.A. Keiderling, Protein Sci. 2 (1993) 307.

1 gr. A 1

[

[

[

[[

[

[[

[

[[[[[[

[

[[

[

[Feibush, C. Moeder, G. Bicker, N. Grinberg, Anal. Chem. 69 (1997) 1999.

240 S. Ma et al. / J. Chromato

19] T.A. Keiderling, in: N. Berova, K. Nakanishi, R.W. Woody (Eds.), Circular Dichro-ism, second ed., Wiley-VCH, New York, 2000, p. 621.

20] S. Ma, S. Shen, H. Lee, N. Yee, C. Senanayake, L.A. Nafie, N. Grinberg, TetrahedronAsymmetry 19 (2008) 2111.

21] V. Setnicka, M. Urbanová, V. Král, K. Volka, Spectrochim. Acta. A: Mol. Biomol.Spectrosc. 58 (2002) 2983.

22] P.K. Bose, P.L. Polavarapu, Carbohydr. Res. 323 (2000) 63.23] M.J.T.G.W. Frisch, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, J.J.A.

Montgomery, T. Vreven, K.N. Kudin, J.C. Burant, J.M. Millam, S.S. Iyengar, J.Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson,H. Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J.E. Knox, H.P. Hratchian,J.B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J.Austin, R. Cammi, C. Pomelli, J.W. Ochterski, P.Y. Ayala, K. Morokuma, G.A.Voth, P. Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D. Daniels,M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K. Raghavachari, J.B. Foresman,J.V. Ortiz, Q. Cui, A.G. Baboul, S. Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A.Liashenko, P. Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-Laham,C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill, B. Johnson, W. Chen,

M.W. Wong, C. Gonzalez, J.A. Pople, Gaussian 03, Revision B., third ed., Gaussian,Inc., Pittsburgh, PA, 2003.

24] N. Grinberg, in: N. Grinberg (Ed.), Modern Thin-layer Chromatography, MarcelDekker, New York, 1990, p. 398.

25] K. Günther, J. Chromatogr. 448 (1988) 11.26] J.P. Foley, W.E. May, Anal. Chem. 59 (1987) 110.

[

[

216 (2009) 1232–1240

27] J.E. Hardcastle, M. He, B. Begum, R. Vermillion-Salsbury, J. Chromatogr. A 691(1995) 225.

28] D.G. Durham, Chirality 8 (1996) 58.29] P.K. Owens, A.F. Fell, M.W. Coleman, J.C. Berridge, Chirality 8 (1996) 466.30] E. Ameyibor, J.T. Stewart, J. Chromatogr. B 703 (1997) 273.31] P.K. Owens, A.F. Fell, M.W. Coleman, J.C. Berridge, Chirality 9 (1997) 184.32] J. Barbosa, I. Toro, V. Sanz-Nebot, J. Chromatogr. A 823 (1998) 497.33] A.G. León, A.I. Olives, M.A. Martín, B. del Castillo, J. Incl. Phenom. Macrocycl.

Chem. 57 (2007) 577.34] A.G. León, A.I. Olives, B. del Castillo, M.A. Martín, J. Chromatogr. A 1192 (2008)

254.35] C. Moeder, T. O’Brien, R. Thompson, G. Bicker, J. Chromatogr. A 736 (1996) 1.36] C.S. Horvath, W.R. Melander, in: E. Heftman (Ed.), Fundamentals and Appli-

cations of Chromatographic and Electrophoretic Methods, Elsevier ScientificPublishing Company, Amsterdam, 1983, p. A26.

37] M. See, D.R. Sidler, A.J. Simon, C.R. Middaugh, R. Thompson, L.J. August, H.J.Perpall, N. Grinberg, Chirality 11 (1999) 224.

38] T. O’Brien, L. Crocker, R. Thompson, K. Thompson, P.H. Toma, D.A. Conlon, B.

39] Y. Bereznitski, R. LoBrutto, R. Thompson, K. Thompson, P. Sajonz, L.S. Crocker, J.Kowal, D. Cai, M. Journet, T. Wang, J. Wyvratt, N. Grinberg, Enantiomer 7 (2002)305.

40] G. Zundel, Hydration and Intermolecular Interaction. Infrared Investigationswith Polyelectrolyte Membranes, Academic Press, New York, 1969, p. 46.