Anal. Chem. 2000, 72, 1263-1267
Degree of Solute Inclusion in Native β-Cyclodextrin: Chromatographic Approach Eric Peyrin,†,* Yves Claude Guillaume,‡ and Anne Ravel†
Laboratoire de Chimie Analytique, Faculte´ de Pharmacie, Domaine de la Merci, 38700 La Tronche, France and Laboratoire de Chimie Analytique, Faculte´ de Me´ decine et Pharmacie, Place Saint-Jacques, 25030 Besanc¸ on Cedex, France
The reversed-phase liquid chromatography retention and separation of a series of D,L dansyl amino acids were investigated over a wide range of salting-out agent (sucrose) concentrations using native β-cyclodextrin as a chiral stationary phase. An original treatment was developed to determine the number of sucrose molecules (n) excluded from the solute-β-cyclodextrin cavity interface when the analyte transfer occurred. Using the n values, the relative degrees of compound inclusion were calculated and correlated to the steric bulkiness of the solute. Thermodynamic parameter variations are discussed in relation to the inclusion degree of the dansyl amino acids. This numerical approach is a valuable tool to explore the steric effects implied in the host-guest complex formation. There are two general approaches to the direct liquid chromatographic separation of chiral compounds: (i) the use of an achiral stationary phase in conjunction with a chiral mobile phase and (ii) the use of a chiral stationary phase (CSP) with an achiral mobile phase. Both methods are based on the temporary formation of diastereoisomeric adsorbates of different stability, with the most stable one being preferentially retained. Several chiral stationary phases have been used over the past few years to separate enantiomers including Pirkle type phases;1,2 chiral ligandexchange phases;3 cellulose and derivatives;4 immobilized proteins such as human serum albumin,5-7 bovine serum albumin,8 or chymotrypsin;9 crown-ether;10 or, more recently, antibiotic phases.11 However, the most frequent CSP used is the cyclodextrin (CD) stationary phase. To investigate the theoretical aspects of the solute interaction with the β-CD, several methods have been proposed such as * Corresponding author: (phone) 333.81.66.55.46; (fax) 333.81.66.55.27; (e-mail)
[email protected]. † Laboratoire de Chimie Analytique, Faculte ´ de Pharmacie. ‡ Laboratoire de Chimie Analytique, Faculte ´ de Me´decine et Pharmacie. (1) Pirkle, W. H.; Hyun, M. H.; Banks, B. J. Chromatogr. 1984, 316, 585. (2) Wainer, I. W.; Alembik, M. C. J. Chromatogr. 1986, 367, 59. (3) Gubitz, G. J. Liq. Chromatogr. 1986, 9, 519. (4) Okamoto, Y.; Aburatani, R. Polym. News 1989, 14, 295. (5) Peyrin, E.; Guillaume, Y. C.; Guinchard, C. Anal. Chem. 1997, 69, 4979. (6) Peyrin, E.; Guillaume, Y. C.; Guinchard, C. Anal. Chem. 1998, 70, 4235. (7) Peyrin, E.; Guillaume, Y. C. Chromatographia 1998, 48, 431. (8) Allenmark, S.; Bomgren, B.; Boren, H. J. Chromatogr. 1982, 237, 473. (9) Wainer, I. W.; Jadaud, P.; Schombaum, G. Chromatographia 1988, 25, 903. (10) Castelnovo, P. Chirality 1993, 5, 181. (11) Armstrong, D. W.; Yubing, Y.; Zhou, Y.; Bagwill, C.; Chen, J. R. Anal. Chem. 1994, 66, 1473. 10.1021/ac990705n CCC: $19.00 Published on Web 01/29/2000
© 2000 American Chemical Society
nuclear magnetic resonance,12 the solubility technique,13 fast-atom bombardment mass spectrometry,14 spectroscopic and photophysical methods,15 thermal analysis,16 potentiometric titration,17 or high-performance liquid chromatography.18-20 In the chromatographic reversed-phase mode, it is known that the inclusion process is the main phenomenon which determines chiral recognition with a native β-CD.21 As well, it has been demonstrated that the hydrophobic effect constitutes one of the main driving forces in this complex formation.22 However, few studies have specifically examined this hydrophobic contribution to the soluteβ-CD inclusion and its chiral discrimination. The aim of this paper was to investigate the role of this effect in the D,L dansyl amino acid-β-CD interaction by varying the concentration of a saltingout agent, i.e., sucrose, in the mobile phase. Using a model previously established to describe the hydrophobic interactions in various chromatographic systems,23-25 the number of sucrose molecules (n) excluded from the solute-β-CD interface (when the compound was transferred from the mobile phase to the hydrophobic cavity) was calculated. This n value was treated as a specific marker of the inclusion degree of the D/L enantiomer in the hydrophobic cavity. Thermodynamic variations were obtained from van’t Hoff plots and were discussed relative to this inclusion degree. THEORY In the reversed-phase application, the separation mechanism on β-CD is thought to be the result of the formation of inclusion complexes in which the solute is included in the hydrophobic cavity of the CD. The formation of this complex may be caused (12) Djedaini, F.; Perly, B. J. Pharm. Sci. 1991, 80, 1157. (13) Frank, S. G.; Kavaliunas, D. R. J. Pharm. Sci. 1983, 72, 1215. (14) Juo, C. G.; Shiu, L. L.; Shen, C. K. F.; Luh, T. Y.; Her, G. R. Rapid Commun. Mass Spectrom. 1995, 9, 604. (15) Andersson, T.; Sundahl, M.; Westman, G.; Wennerstrom, O. Tetrahedron Lett. 1994, 35, 7103. (16) Morin, N.; Guillaume, Y. C.; Peyrin, E.; Rouland, J. C. Anal. Chem. 1998, 70, 2819. (17) Valsami, G. N.; Koupparis, M. A.; Macheras, P. E. Pharm. Res. 1992, 9, 94. (18) Armstrong, D. W.; Demond, W. J. Chromatogr. Sci. 1984, 22, 411. (19) Ward, T. J.; Arrmstrong, D. J. J. Liq. Chromatogr. 1986, 9, 407. (20) Stalcup, A. M.; Chang, S. S.; Armstrong, D. W.; Pitha, J. J. Chromatogr. 1990, 513, 181. (21) Armstrong, D. W.; Han, Y. I.; Han, S. M. Anal. Chim. Acta 1988, 208, 275. (22) Janado, M.; Yano, Y.; Umura, M.; Kondo, Y. J. Solution Chem. 1995, 24, 587. (23) Peyrin, E.; Guillaume, Y. C.; Morin, N.; Guinchard, C. Anal. Chem. 1998, 70, 2812. (24) Peyrin, E.; Guillaume, Y. C. Anal. Chem. 1999, 71, 1496. (25) Guillaume, Y. C.; Peyrin, E. Anal. Chem. 1999, 71, 1326.
Analytical Chemistry, Vol. 72, No. 6, March 15, 2000 1263
by a variety of interactions such as a release of high-energy water or modifier during the complex formation, hydrogen bonding with the OH groups at the periphery of the cavity, van der Waals interactions, and a hydrophobic effect depending on the cavity size of the CD or the size of the guest molecule as well as the solvent composition. The Gibbs free energy of the inclusion complex, ∆G°Inc, can be divided into hydrophobic, ∆G°IncH, and nonhydrophobic, ∆G°IncNH, contributions. It has been known for many years that other factors which do not contribute to the chiral recognition could participate in retention. These possible mechanisms which can compete with inclusion-complex formation include the binding to the unreacted silanol groups, the external adsorption to the OH groups of the CD, and the hydrophobic effect with the spacer chain.20 The nonspecific Gibbs free energy, ∆G°NInc, can be described by the sum of the hydrophobic, ∆G°NIncH, and the nonhydrophobic, ∆G°NIncNH, contributions. The total Gibbs free energy of the solute transfer from the mobile to the stationary phases, ∆G°, can be obtained by combining the ∆G°Inc and ∆G°NInc Gibbs free energies. Thus, ∆G° is equal to the sum of the ∆G°IncH, ∆G°NIncH, and ∆G°NH terms, where ∆G° NH corresponds to all the nonhydrophobic contributions implied in the solute transfer. Previous papers have described the variation of ∆G°H, i.e., hydrophobic Gibbs free energy, when a salting-out agent concentration varies in the mobile phase. Using the adsorption isotherm of Gibbs, the following equation is obtained23-25
(
)
d∆G°H d ln c
T
) RTλ
of the solute.27 So, N is linked to the asa by
N ) β(asa)
(2)
N ) β′(asaR)
ln k′ ) (N + n)ln 1/c + κ
(3)
where N and n are the number of excluded sucrose molecules for the nonspecific process and the inclusion-complex formation, respectively, and κ is a constant. It is accepted that the hydrophobic effect is correlated with the accessible surface area (asa) (26) Back, J. F.; Oakenfull, D.; Smith, M. B. Biochemistry 1979, 18, 5191.
1264
Analytical Chemistry, Vol. 72, No. 6, March 15, 2000
(5)
where asaR is the asa value of the R chain and β′ a constant. Combining eqs 3 and 5 gives:
ln k′ ) (β′(asaR) + n)ln 1/c + κ
(6)
As the enantioselectivity R ) k′D/k′L, the following relation is obtained
ln R ) (nD - nL)ln 1/c + κ′ where φ represents the phase ratio (volume of the stationary phase divided by the volume of the mobile phase). Back et al.26 have shown that the contribution of sucrose at a concentration of 1 M to hydrogen bonding and electrostatic interactions can be considered negligible in comparison with its effect on hydrophobic interaction. Thus, ∆G° NH can be assumed to be constant over the sucrose concentration range studied (0.10.5 M). By integration of eq 1 and combination with eq 2, the following relation is obtained
(4)
where β is the proportionality constant. As the dansyl amino acids differ only by the R chain (Figure 1), the following relation can be obtained
(1)
where RT is the thermal energy, c the sucrose concentration in the mobile phase, and λ the excess of sucrose molecules at the solute-stationary phase interface. It is also well-known that the retention factor for a solute is related to ∆G°. This relation is represented by the following equation
∆G° ) RT(ln φ - ln k′)
Figure 1. Dansyl amino acid structures.
(7)
with κ′ constant. EXPERIMENTAL SECTION Apparatus. The HPLC system consisted of a HPLC waters pump 501 (Saint Quentin, Yvelines, France), an Interchim Rheodyne injection valve model 7125 (Montluc¸ on, France) fitted with a 20-µl sample loop a Merck 2500 diode array detector (Nogent sur Marne, France). An Interchim 125 mm × 4 mm β-CD column (5 µm, particle size) was used with controlled temperature in an Interchim oven TM N°701 (Montluc¸on, France). The mobile phase flow rate for all experiments was 1 mL/min. Reagents. All the D,L dansyl amino acids were obtained from Sigma Aldrich (Saint-Quentin, France). The chemical structures of these compounds are given in Figure 1. Fresh samples were prepared daily at a concentration of 20 mg/L. Sodium nitrate (Merck, Nogent sur Marne, France) was used as a dead-time (27) Chothia, C.; Janin, J. Nature (London) 1975, 256, 705.
Figure 2. Plots of ln k′ against ln 1/c for D (A) and L (B) dansyl phenylalanine and D (C) and L (D) dansyl valine at T ) 15 °C. Figure 3. Plots of n (T ) 15 °C) against asaR (Å2) for the D (]) and L (O) enantiomers.
marker. Sodium hydrogen phosphate and sodium dihydrogen phosphate were supplied by Prolabo (Paris, France). Water was obtained from an Elgastat option water purification system (Odil, Talant, France) fitted with a reverse osmosis cartridge. The mobile phase consisted of phosphate buffer 0.1 M at pH ) 6.0. The variation range of the sucrose (Prolabo) was 0.1-0.5 M. Each mobile phase was allowed to stand at ambient temperature, and its pH was measured after 1, 2, and 4 h. No pH fluctuations were observed, and the pH of each mobile phase was within 0.5% of the desired value. Twenty microliters of each solute (or a suitable mixture) was injected, and the retention times were measured. Temperature Studies. Compound retention factors were determined over the temperature range 15-40 °C. The chromatographic system was allowed to equilibrate at each temperature for at least 1 h prior to each experiment. To study this equilibrium process, the retention time of the D-dansyl norvaline was measured every hour for 7 h and again after 20, 21, and 23 h. The maximum relative difference of the retention time of this compound between these different measurements was always 0.4%, making the chromatographic system sufficiently equilibrated for use after 1 h. RESULTS AND DISCUSSION Determination of the Degree of Solute Inclusion. The retention factor values were determined for a wide range of sucrose concentrations at a column temperature equal to 15 °C. All the experiments were repeated three times. The coefficients of variation of the k′ values were 0.99. This good correlation between the theoretical and experimental values was considered adequate to verify the model. It is generally accepted for the native β-CD that a solute must have at least two rings, i.e., a naphthyl moiety, for chiral
Table 1. Relative Degrees of Inclusion for the D and L Enantiomers of the Dansyl Amino Acids at a Column Temperature Equal to 15 °C compounds
nL/n0
nD/n0
nL/nD
Dns alanine Dns valine Dns norvaline Dns leucine Dns phenylalanine Dns tryptophan
0.67 0.66 0.66 0.66 0.65 0.30
0.95 0.88 0.87 0.83 0.79 0.32
0.71 0.75 0.76 0.80 0.82 0.94
recognition due to the fact that the internal diameter of the β-CD is approximately the same size as that of the naphthyl moiety. Small compounds containing no ring or only one ring, for example, have too much freedom and move and rotate in the β-CD cavity. Chiral recognition is obtained when the conformation and movement of a complexed solute is restricted such as in the case of molecules containing two or more rings. Thus, it is reasonable to expect that the n values are the reflection of the exclusion of the sucrose molecules from the naphthyl moiety when the solute transfer from the mobile phase to the interior of the CD cavity takes place. This n value was treated as an inclusion marker of the naphthyl moiety of the dansyl amino acid in the CD cavity. The n values were plotted against the asaR values.28 Figure 3 shows these plots. A linear correlation was obtained corresponding to
nD ) 4.7 × 10-3asaR - 2.80 r2 ) 0.99
(8)
nL ≈ -1.90
(9)
when the dansyl tryptophan molecule was excluded from the plots (Figure 3). For asaR ) 0, the theoretical n value (n0) representing the complete inclusion without steric hindrance of the R chain was equal to -2.80. The n/n0 ratio was introduced to describe the relative degree of inclusion of an analyte in the cavity (Table 1). In the same way, the relative inclusions nL/nD between L and D enantiomer were listed in Table 1. The naphthyl moiety inclusion process under the hydrophobic effect dependence was largely influenced by the steric hindrance (28) Wimley, W. C.; Creamer, T. P.; White, S. H. Biochemistry 1996, 35, 5109.
Analytical Chemistry, Vol. 72, No. 6, March 15, 2000
1265
caused by the R group. For the D enantiomer, this inclusion decreased linearly (Figure 3 and eq 8) the larger the R group, the weaker the possibility of the molecule to be completely included in the CD cavity (see the nD/n0 values in Table 1). On the other hand, the inclusion degree of the L enantiomer was approximately constant for all L solutes. This indicated that the chiral recognition further decreased when the R chain increased. The theoretical asaR value which corresponds to no chiral recognition is 191 Å2. This represents a theoretical linear chain of 6 atoms of carbon, which has a molecular length of around 7.8 Å.29 It is exactly the value of the internal diameter of the native β-CD. This observation confirms that a large R group (>7.8 Å) restricted the inclusion process of the D enantiomer in the same manner as it restricted the L-enantiomer inclusion and thus hindered the chiral discrimination. These results would suggest that the R chain of the L enantiomer must be near and interact with the mouth of the cavity whatever its length (eq 9) in such a way that it limits the stability of the complex. In the case of the D enantiomer, this same substitute was spatially more favorably placed for the complete inclusion. However, when the R chain length increased, this benefit was progressively canceled and did not discriminate for R equal to the internal diameter of the CD cavity. Figure 3 shows that there was a modification in the inclusion mechanism for dansyl tryptophan in comparison with the mechanism for the other compounds. There was a weak difference in the inclusion process between D and L enantiomer given a weak selectivity value (see the nL/nD values in Table 1). The R chain of the dansyl tryptophan containing two rings would appear to be included in the cavity in the same way as the naphthyl moiety. Thus, there was a weak difference between enantiomer conformations, and the chiral discrimination between D and L enantiomers was reduced. Effects of Temperature on the Enantioselectivity and Thermodynamic Parameter Variations. To study the influence of temperature on the R values, the same experiments were carried out at T equal to 20, 25, 30, 35, 40, and 45 °C. The separation-factor values for all enantiomer pairs were calculated from the k′ values. The difference of free energy ∆∆G° between the D and L enantiomers can be broken down into enthalpic ∆∆H° and entropic ∆∆S° terms to give the van’t Hoff equation:30
ln R ) -∆∆H°/RT + ∆∆S°/R
(10)
If there is no change in the enantioselective interactions in relation to temperature, then a plot of ln R vs 1/T would be linear with a slope of -∆∆H°/R and an intercept of ∆∆S°/R. These plots were determined for different sucrose concentrations. The van’t Hoff plots were all linear for L and D dansyl amino acids. The correlation coefficients for the linear fits were in excess of 0.98. The typical standard deviations of slope and intercept were, respectively, 0.008 and 0.08. Figures 4 and 5 show the variation in ∆∆H° and ∆∆S° with asaR and c. The negative values of the ∆∆H° and ∆∆S° terms demonstrated that the enantioselectivity was controlled enthalpically. This was consistent with results (29) Wanwimolruk, S.; Birkett, D. J.; Brooks, P. M. Mol. Pharmacol. 1983, 24, 458. (30) Guillaume, Y. C.; Guinchard, C. J. Phys. Chem. 1997, 101, 8390.
1266
Analytical Chemistry, Vol. 72, No. 6, March 15, 2000
Figure 4. Variation in ∆∆H° (kJ/mol) as a function of asaR (Å2) and c (M) for all the D,L-enantiomer pairs.
Figure 5. Variation in ∆∆S° (J/mol/K) as a function of asaR (Å2) and c (M) for all the D,L-enantiomer pairs.
reported in the literature for various chromatographic enantioseparations.31,32 The ∆∆H° values decreased when c increased. When the sucrose concentration increased in the mobile phase, the D enantiomer transfer was more enhanced than the L enantiomer transfer and a strong enthalpic effect occurred in the CD cavity. In the same manner, the entropic term decreased while c increased due to the great loss of the degree of freedom when the D-enantiomer transfer in the cavity increased. The variations in the thermodynamic parameters with asaR followed the relative degree of inclusion of the solute. The larger the bulkiness, the weaker the ∆∆H° and ∆∆S° variations in relation to the steric hindrance determined by the R chain. For example, the dansyl phenylalanine molecule had a ratio nL/nD equal to 0.82 at T ) 15 °C with a ∆∆H° value equal to - 4.8 kJ/ mol at c ) 0.25 M. In the case of the dansyl valine molecule, nL/ nD ) 0.75 with a greater variation in the ∆∆H° value equal to -6.5 kJ/mol (at the same T and c values). The ∆∆H° and ∆∆S° variations for dansyl tryptophan demonstrated a change in comparison with the other solute variations, as shown by the break at the top of Figures 4 and 5. This confirmed on the basis of the results described above that the mechanism of inclusion in the CD cavity was different for this molecule. (31) Thompson, R. A.; Ge, Z.; Grinberg, N.; Ellison, D.; Tway, P. Anal. Chem. 1995, 67, 1580. (32) Peyrin, E.; Guillaume, Y. C.; Guinchard, C. J. Chromatogr. Sci. 1998, 36, 97.
CONCLUSION This paper described an original approach to assessing the contribution of the hydrophobic interaction to the dansyl amino acid inclusion process in the cylodextrin cavity. Using a simple physicochemical model, it was possible to determine the number of sucrose molecules which were excluded from the solutecylodextrin interface when the compound was transferred from the mobile to the stationary phase. This number was correlated to the accessible surface area of the R chain of the dansyl amino acid, indicating that the inclusion process was largely dependent
on steric hindrance. This numerical treatment could be applied to other host-guest complex types where steric effects play a preponderant role. ACKNOWLEDGMENT We thank Mireille Thomassin for her technical assistance. Received for review June 29, 1999. Accepted December 3, 1999. AC990705N
Analytical Chemistry, Vol. 72, No. 6, March 15, 2000
1267
