Evaluating Chiral Separation Interactions by Use of Diastereomeric

Bertha C. Valle, Kevin F. Morris, Kristin A. Fletcher, Vivian Fernand, Drew M. Sword, ... Shahab A. Shamsi, Bertha C. Valle, Fereshteh Billiot, and Is...
0 downloads 0 Views 132KB Size
Anal. Chem. 1999, 71, 4044-4049

Evaluating Chiral Separation Interactions by Use of Diastereomeric Polymeric Dipeptide Surfactants Eugene Billiot, Stefan Thibodeaux,† Shahab Shamsi,‡ and Isiah M. Warner*

Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana 70803

Poly sodium N-undecyl leucine-leucine (poly SULL) is used as a diagnostic tool to investigate chiral molecular interactions via electrokinetic chromatography (EKC). Poly SULL has two chiral centers which are defined by two asymmetric carbons. Each chiral center of poly SULL can have two possible configurations (D or L). Consequently, four different optical configurations are possible within the surfactant molecule (L-L, D-D, L-D, and D-L). In this study, five chiral analytes of various charge states and hydrophobicities were used to investigate the role of electrostatic interactions and hydrophobicity on chiral recognition with polymeric dipeptide surfactants. These studies lead to a proposed hypothesis for interaction of the analytes with dipeptide surfactants. The hypothesis was tested and the contribution of the double chiral centers to this interaction was evaluated by use of two dipeptide surfactants in which one chiral amino acid is replaced by an achiral amino acid glycine, i.e., poly sodium N-undecyl L-leucine-glycine (poly L-SULG) and poly sodium N-undecyl L-glycine-leucine (poly L-SUGL). The results reported here provide new insights into the mechanism for chiral recognition of select chiral analytes by use of polymeric chiral surfactants. Chiral separations are essential to further growth within the pharmaceutical industry. This is because many drugs are chiral and the differences in physiological interactions of individual enantiomers are not completely understood. Consequently, there is great interest in understanding the kinds of diastereomeric interactions which lead to chiral recognition. One of the more recent tools for chiral separations, i.e., capillary electrophoresis (CE), has gained considerable popularity in recent years. For example, the number of publications on this subject has increased by more than an order of magnitude since 1990.1 Of the variations of CE used for chiral separations, capillary electrokinetic chromatography (EKC) is one of the most powerful. In EKC, enantiomeric separation is achieved by differential partitioning of analyte enantiomers between the chiral pseudo-stationary phase (CPSP) and the bulk phase. A variety of CPSPs have been used for enantiomeric separations in EKC. For example, cyclodextrins * To whom correspondence should be addressed. † Current address: Department of Chemistry, Purdue University, West Lafayette, IN, 47906. ‡ Current address: Department of Chemistry, Georgia State University, Atlanta, GA, 30303. (1) Gu ¨ bitz, G.; Schmid, M. G. J. Chromatogr., A 1997, 792, 179-225.

4044 Analytical Chemistry, Vol. 71, No. 18, September 15, 1999

(natural and derivatized) are the most common CPSPs used in EKC for enantiomeric separations.2-5 Other types of CPSPs of interest include macrocylic antibiotics,6,7 polysaccharides,8,9 proteins,10,11 crown ethers,12,13 calixarenes,14 and micelles.15-19 The use of chiral surfactants for enantiomeric separations has recently shown much promise in micellar electrokinetic chromatography (MEKC).19 One of the advantages of MEKC is the ability to separate neutral and charged species in a single run. Although the theoretical plate numbers in MEKC are not as high as could be achieved with capillary zone electrophoresis (CZE), column efficiency in MEKC is much higher than in high-performance liquid chromatography. The use of achiral polymeric pseudo-stationary phases for EKC was first introduced by Palmer et al. in 1992.20,21 This technique is similar to MEKC except that the micelles are replaced with covalently linked surfactant monomer units. Thus, polymeric surfactants or micelle polymers are produced. Polymeric surfactants have certain distinct advantages over conventional micelles in EKC. A major advantage of polymeric surfactants is the elimination of the dynamic equilibrium between monomer and micelle.22 This advantage results in the lack of a critical micelle concentration (CMC). Thus, the polymer can be used over a wider range of concentrations than the monomer, e.g., below the normal CMC of the unpolymerized surfactants. In addition, organic (2) Rawjee, Y. Y.; Staerk, D. U.; Vigh, G. J. Chromatogr., A 1993, 635, 291. (3) Rawjee, Y. Y.; Williams, R. L.; Vigh, G. J. Chromatogr., A 1993, 652, 233. (4) Biggin, M. E.; Williams, R. L.; Vigh, G. J. Chromatogr., A 1995, 692, 319. (5) Wang, F.; Khaledi, M. G. Anal. Chem. 1996, 68, 3460. (6) Gasper, M. P.; Berthod, A.; Nair, U. B.; Armstrong, D. W. Anal. Chem. 1996, 66, 2501. (7) Rundelett, K. L.; Armstrong, D. W. Anal. Chem. 1995, 67, 2008. (8) D’Hurst, A.; Verbeke, N. J. Chromatogr. 1992, 608, 275-287. (9) Soini, H.; Stefansson, M.; Riekkola, M. L.; Novotny, M. V. Anal. Chem. 1994, 66, 3477-3484. (10) Barker, G. E.; Russo, P.; Hartwick, R. A. Anal. Chem. 1992, 64, 3024. (11) Busch, S.; Kraak, J. C.; Poppe, H. J. Chromatogr., A 1993, 635, 11. (12) Kuhn, R.; Erni, F.; Bereuter, P.; Hausler, J. Anal. Chem. 1992, 64, 2815. (13) Kuhn, R.; Riester, D.; Fleckenstein, B.; Wiesmu ¨ ller, K. H. J. Chromatogr., A 1995, 716, 371. (14) Pena, M. S.; Zhang, Y.; Thibodeaux, S.; McLaughlin, M. L.; de la Pena, A. M.; Warner, I. M. Tetrahedron Lett. 1996, 37 (33), 5841. (15) Shamsi, S. A.; Warner, I. M. Electrophoresis 1997, 18, 853-872. (16) Nishi, H.; Terabe, S. J. Chromatogr., A 1996, 735, 3-27. (17) Smith, J. T.; Nashabeh, W.; El Rassi, Z. Anal. Chem. 1994, 66, 1119. (18) St. Claire, R. L., III Anal. Chem. 1996, 68, 569R-586R. (19) Williams, C. C.; Shamsi, S. A.; Warner, I. M. Adv. Chromatogr. (NY) 1997, 37, 363-419. (20) Palmer, C. P.; McNair, J. J. Microcolumn Sep. 1992, 4, 509-514. (21) Palmer, C. P.; Khaled, M. Y.; McNair, J. J. High Resolut. Chromatogr. 1992, 15, 509-514. (22) Palmer, C. P. J. Chromatogr., A 1997, 780, 75-92. 10.1021/ac990540i CCC: $18.00

© 1999 American Chemical Society Published on Web 08/14/1999

modifiers can be used without total disruption of the micelle.20-26 Finally, structural rigidity and purification of the micelle polymer allows improvement in the mass transfer rate, thereby reducing peak broadening. In 1994, Wang and Warner reported the use of a chiral polymeric surfactant for enantiomeric separation using EKC.27 Their polymer, poly sodium N-undecyl L-valine (poly L-SUV), was used as a pseudo-stationary phase for chiral separation of (()1,1′-bi-2naphthol (BOH) and D,L laudanosine. Over the past few years, Warner et al.28,29 and Dobashi et al.30 have extended the range of enantiomers separated with poly L-SUV using EKC. We have recently embarked on an extensive investigation of the use of various polymeric dipeptide surfactants as pseudostationary phases for chiral separations in EKC. For example, the dipeptide surfactant poly sodium N-undecyl L-valine-valine (poly L-SUVV) has been compared with the single amino acid surfactant poly L-SUV for the separation of four racemates (two cationic, one neutral, and one anionic).31 Of those four, three of the racemates showed significant improvement in chiral recognition with the dipeptide surfactant, as compared with the single amino acid surfactant. In a more recent study, the effect of amino acid order in dipeptide surfactant separation was investigated, and it was shown that the order of the amino acids in the dipeptide surfactant can produce a very dramatic effect on the chiral-recognition ability of the surfactant.32,33 The results of these previous studies have led us to systematically examine the effect of changes in optical configurations in dipeptide surfactants. Five analytes with varying degrees of hydrophobicity and charge states and eight polymeric surfactants were examined in this study. The primary surfactants used in this study were poly sodium N-undecyl (L-L)-leucine-leucine [poly (LL)-SULL], poly sodium N-undecyl (D-D)-leucine-leucine [poly (DD)-SULL], poly sodium N-undecyl (L-D)-leucine-leucine [poly (LD)-SULL], poly sodium N-undecyl (D-L)-leucine-leucine [poly (DL)-SULL], poly sodium N-undecyl D-leucine (poly D-SUL), and poly sodium N-undecyl L-leucine (poly L-SUL). To gain additional insights into the interactions responsible for enantiomeric separations by use of chiral polymeric surfactants, two other dipeptide surfactants containing only one chiral center, [poly sodium N-undecyl L-leucine-glycine (poly L-SULG) and poly sodium N-undecyl L-glycine-leucine (poly L-SUGL)] were also examined. Our rationale here is that the achiral amino acid, glycine, serves as a spacer for placing the chiral amino acid as the C-terminal or N-terminal amino acid for a given dipeptide surfactant. Thus, synergism between chiral centers on the same surfactant can be more easily discerned. Finally, enantiomeric separation of two cationic β-blockers, [propranolol (Prop) and alprenolol (Alp)], as (23) Palmer, C. P. J. Chromatogr., A 1995, 780, 297-304. (24) Ozaki, H.; Ichihara, A.; Terabe, S. J. Chromatogr., A 1995, 709, 3-10. (25) Yang, S. Y.; Bumgarner, M. G.; Khaledi, M. G. J. High Resolut. Chromatogr. 1995, 18, 443-445. (26) Palmer, C. P.; Terabe, S. Anal. Chem. 1997, 69, 1852-1860. (27) Wang, J.; Warner, I. M. Anal. Chem. 1994, 66, 3773-3776. (28) Wang, J.; Warner, I. M. J. Chromatogr., A 1995, 711, 297-304. (29) Agnew-Heard, K.; Sanchez Pena, M.; Shamsi, S.; Warner I. M. Anal. Chem. 1997, 69, 958-964. (30) Dobashi, A.; Hamada, M.; Dobashi, Y. Anal. Chem. 1995, 67, 3011-3017. (31) Shamsi, S.; Macossay, J.; Warner, I. M. Anal. Chem. 1997, 69, 2980-2987. (32) Billiot, E.; Macossay, J.; Thibodeaux, S.; Shamsi, S.; Warner, I. M. Anal. Chem. 1998, 70, 1375-1381. (33) Billiot, E.; Agbaria, R. A.; Thibodeaux, S.; Shamsi, S.; Warner I. M. Anal. Chem. 1999, 71, 1252-1256.

Figure 1. Structures of analytes examined in this study.

Figure 2. Proposed interactions of various chiral analytes with polymeric dipeptide surfactants on the basis of electrostatic and hydrophobic interactions.

well as three model atropisomers, [(()1,1′-bi-2-naphthol (BOH), (() 1,1′-bi-2-naphthyl-2,2′-diamine (BNA), and (() 1,1′-bi-2-naphthyl-2,2′-diyl hydrogen phosphate (BNP)], were compared using the aforementioned polymeric surfactants. The structures of the analytes examined in this study are shown in Figure 1. In addition, a schematic representation of the polymeric dipeptide surfactants and the proposed sites of preferential interaction of the analytes examined in this study are shown in Figure 2. EXPERIMENTAL SECTION Synthesis of Polymeric Surfactants. All surfactants in this study were synthesized using the procedure reported by Wang and Warner.27 Surfactant monomers were prepared by mixing the N-hydroxysuccinimide ester of undecylenic acid with the amino Analytical Chemistry, Vol. 71, No. 18, September 15, 1999

4045

acid or dipeptide to form the corresponding N-undecyl chiral surfactant. Polymerization was achieved by use of 60Co γ irradiation at concentrations above the CMC. All polymers used in this study were found to be greater than 99% pure as estimated from elemental analysis. Materials. The racemic mixtures and the pure optical isomers of 1,1′-bi-2-naphthol (BOH), 1,1′-bi-2-naphthyl-2,2′-diamine (BNA), 1,1′-bi-2-naphthyl-2,2′-diyl hydrogen phosphate (BNP), propranolol (Prop), and alprenolol (Alp) were purchased from Aldrich (Milwaukee, WI) and used as received. The 3-(cyclohexylamino)-1propanesulfonic acid (CAPS), and sodium borate were obtained from Fisher Scientific Company (Fair Lawn, NJ) and used as received. Chemicals used for the synthesis of surfactants included: N,N′-dicyclohexylcarbodiimide (DCC), N-hydroxysuccinimide, undecylenic acid, various amino acids, and the dipeptides. All were supplied by Sigma (St. Louis, MO) and used as received. Preparation of EKC Buffer Solutions. The background electrolyte (BGE) for separation of the binaphthyl derivatives was 50 mM sodium borate at pH 10.0. The BGE used for the cationic β-blockers was 50 mM sodium borate and 300 mM CAPS at pH 8.5. CAPS was added to minimize analyte interaction with the capillary wall. An appropriate amount of the polymeric surfactant was then added to the BGE and the pH readjusted with 1 M NaOH or 1 M HCl if necessary. Capillary Electrophoresis. The EKC experiments were conducted on a Hewlett-Packard 3D CE model no. G1600AX. An untreated fused silica capillary (effective length 55 cm, 50 µm i.d.) was purchased from Polymicro Technologies (Phoenix, AZ). The surfactants were added to the buffer solution and the solution filtered through a 0.45-µm membrane filter. Separations were performed at +30 kV, with UV detection at 215 nm. The temperature of the capillary was maintained at 25° C for the binaphthyl derivatives and 12 °C for Prop and Alp by the instrument thermostating system, which consisted of a Peltier element for forced-air cooling and temperature control. The binaphthyl derivatives, (BNP, BNA, and BOH), were prepared in 50:50 methanol/water, at concentrations of 0.1 mg/mL. Propranolol and Alp were also prepared in 50:50 methanol/water, but at a concentration of 2.5 mg/mL. Samples were injected for 5 s with 10 mbar of pressure. Prior to use, each new capillary was conditioned for 30 min with 1 M NaOH and then for 30 min with 0.1 M NaOH. Finally, the capillary was rinsed for 15 min with triply distilled deionized water. Prior to each run, the capillary was flushed with the EKC buffer for 2 min to condition and fill the capillary. RESULTS AND DISCUSSIONS Enantiomeric Separation of Alprenolol and Propranolol. The single amino acid surfactants of opposite optical configuration (poly L-SUL and poly D-SUL) were used to determine the elution order of the enantiomers. The R form of both Alp and Prop, which for identification purposes is at half the concentration of the S form, elutes first in the case of poly D-SUL (Figure 3a). When poly D-SUL is replaced with poly L-SUL, the order of elution of the enantiomers is reversed, i.e., the S form of both Prop and Alp elutes second (Figure 3b). A comparison of dipeptide surfactants with the same optical configuration at both chiral centers, such as poly (L-L) SULL and its antipode poly (D-D) SULL again show a reversal of enantiomeric order (Figure 3c-d). Further compari4046 Analytical Chemistry, Vol. 71, No. 18, September 15, 1999

Figure 3. Comparison of elution order and enantioseparation of alprenolol (Alp) and propranolol (Prop) [R(+)/S(-) ratio ) 2:1] obtained with various polymeric surfactants: (a) poly D-SUL, (b) poly L-SUL, (c) poly (D,D) SULL, (d) poly (L,L) SULL, (e) poly (D,L) SULL, (f) poly (L,D) SULL. EKC conditions: applied voltage + 30 kV, buffer solution prepared with 50 mM sodium borate and 300 mM CAPS at pH 8.5, column temperature 12 °C, 18 mM each of the polymeric surfactants at equivalent monomer concentrations.

sons of the electropherograms show that the elution order of the enantiomers is the same for poly D-SUL, poly (D-D) SULL, and poly (L-D) SULL, i.e., the S form of both Alp and Prop always elutes first. However, in the cases of poly L-SUL, poly (L-L) SULL, and poly (D-L) SULL, the S form elutes second. Thus, the elution order obtained in Figures 3 a-f suggests that chiral recognition for Alp and Prop occurs primarily at the C-terminal amino acid of the polymeric dipeptide surfactants, R2 of Figure 2. Since chiral recognition of these two β-blockers occurs at the outside (Cterminal) amino acid, it is reasonable to assume that Alp and Prop interact preferentially with the outside (C-terminal) amino acid. Furthermore, it is logical to assume that since the surfactants used in this study are anionic, and Alp and Prop are cationic under the experimental conditions used here, electrostatic attraction is the major factor responsible for the preferential interaction of these analytes with the outside (C-terminal) amino acid. Further evidence of the preferential interaction of Alp and Prop with the outside (C-terminal) amino acid can be seen by a comparison of the resolution of the enantiomers of Alp and Prop with dipeptide surfactants (poly L-SULG and poly L-SUGL) containing only one chiral center, Figure 4. Poly L-SUGL produced baseline resolution of the enantiomers of both Alp and Prop. In contrast, poly L-SULG provided no enantiomeric separation for Alp and only slightly resolved the enantiomers of the more hydrophobic analyte Prop (Figure 4). These results support the assertion that the cationic species examined in this report interact, for separation purposes, with the outermost amino acid. Enantiomeric Separation of Binaphthyl Derivatives. As with the β-blockers, the single amino acid surfactants poly L-SUL and poly D-SUL were also used to determine the elution order of the enantiomers of BNP, BOH, and BNA. The most hydrophobic of these three analogues is BNA which is neutral under the experimental conditions (50 mM sodium borate, pH 10) used here, while BOH is partially ionized (pKa1 ≈ 9.5) and BNP is completely

Figure 4. Comparison of the enantiomeric resolution of alprenolol (Alp) and propranolol (Prop) obtained with the single chiral center dipeptide surfactants poly L-SULG and poly L-SUGL.

Figure 6. Comparison of elution order and enantioseparation of 1,1′-bi-2-naphthol (BOH) [R(+)/S(-) ratio ) 2:1] obtained with various polymeric surfactants: (a) poly L-SUL, (b) poly D-SUL, (c) poly (L,L) SULL, (d) poly (D,D) SULL, (e) poly (L,D) SULL, (e) poly (D,L) SULL. EKC conditions are the same as those listed in Figure 5.

Figure 5. Comparison of elution order and enantioseparation of 1,1′-bi-2-naphthyl-2,2′-diamine (BNA) [R(+)/S(-) ratio ) 2:1] obtained with various polymeric surfactants: (a) poly L-SUL, (b) poly D-SUL, (c) poly (L,L) SULL, (d) poly (D,D) SULL, (e) poly (L,D) SULL, (f) poly (D,L) SULL. EKC conditions: applied voltage + 30 kV, buffer solution prepared with 50 mM sodium borate at pH 10.0, column temperature 25 °C, 5 mM each of the polymeric surfactants at equivalent monomer concentrations.

anionic. The S form of BNA, which is at half the concentration of the R form, eluted earlier than the R form with poly L-SUL (Figure 5a). Again, as expected, when poly L-SUL was replaced with poly D-SUL, the order of elution of the enantiomers of BNA was reversed (Figure 5b). Evidence of the preferential site of interaction for BOH and BNA can be seen in the elution order of the enantiomers with the various dipeptide surfactants. A comparison of dipeptide surfactants with the same optical configuration at both chiral centers [poly (L-L) SULL and its antipode poly (D-D) SULL] shows reversal of enantiomeric order for BNA, as shown in Figure 5cd. Evidence of the chiral interaction of BNA with the inside (Nterminal) amino acid is provided through a comparison of poly L-SUL, poly (L-L) SULL, and poly (L-D) SULL. The R form elutes last in all three cases. In the case of poly D-SUL, poly (D-D) SULL,

and poly (D-L) SULL, the R form elutes first. These results strongly suggest that chiral recognition of BNA is occurring primarily at the inside (N-terminal) amino acid, R1 of Figure 2. Similar trends were observed for BOH (Figure 6). The R form of BOH elutes last for poly L-SUL, poly (L-L) SULL and poly (L-D) SULL and first for poly D-SUL, poly (D-D) SULL, and poly (D-L) SULL. It is important to note here that although chiral recognition for BOH and BNA is occurring primarily at the inside (N-terminal) amino acid, BOH and BNA also interact to some extent with the outside (C-terminal) amino acid. This is evident from the observed decrease in chiral resolution of BOH and BNA for the D-L and L-D configurations as compared with the D-D and L-L configurations of poly SULL. As shown in Figures 5 and 6, the enantiomeric resolution is approximately the same for the D form as compared to the L form, for the L-L form as compared to the D-D form, and for the L-D form as compared to the D-L form. However, a comparison of dipeptide surfactants with the same optical configurations (i.e., L-L, D-D) and dipeptide surfactants with different optical configurations (i.e., L-D, D-L) shows a marked decrease in chiral resolution of both BNA and BOH. Since the chiral selectivities of the D and L forms are opposite, interaction of the analyte with two chiral centers of opposite configuration would tend to reduce the chiral selectivity. In view of the fact that BNA is neutral and BOH is only partially anionic under the conditions used, these analytes are more hydrophobic than BNP, which is completely ionized under these conditions (50 mM sodium borate, pH 10). Therefore, since BNP is less hydrophobic than BOH and BNA, different trends would be expected for BNP as compared with BOH and BNA. These predictions were verified by subsequent experiments on BNP. Examination of the data for BNP suggests that BNP does not penetrate as deeply into the hydrophobic core of the polymeric surfactant as BOH and BNA (Figure 7). As seen in Figure 7 a-d, the elution order is the same for poly L-SUL and poly (L-L) SULL Analytical Chemistry, Vol. 71, No. 18, September 15, 1999

4047

Figure 7. Comparison of elution order and enantioseparation of 1,1′-bi-2-naphthyl-2,2′-diyl hydrogen phosphate (BNP) [R(+)/S(-) ratio ) 2:1] obtained with various polymeric surfactants: (a) poly L-SUL, (b) poly D-SUL, (c) poly (L,L) SULL, (d) poly (D,D) SULL, (e) poly (L,D) SULL, (e) poly (D,L) SULL. EKC conditions are the same as those listed in Figure 5.

(R first, S last), while the opposite elution order is observed for poly D-SUL and poly (D-D) SULL, (S first, R last). In contrast, no resolution is observed for BNP with poly (L-D) SULL and poly (D-L) SULL, (Figure 7 e-f). One possible explanation for such behavior is that, since BNP is anionic under the experimental conditions used (pH 10), it does not penetrate as deeply into the core of the polymeric surfactant as the neutral, more hydrophobic binaphthyl derivatives (BOH and BNA). Thus, it is believed that BNP, under the experimental conditions used for these experiments, interacts approximately the same with both chiral centers (R1 and R2 in Figure 2) of the dipeptide surfactant. It is reasonable to assume that if an analyte interacts approximately equally with two chiral centers of equal chiral selectivity but opposite configuration, then no enantiomeric resolution would be expected, as is observed for BNP. Further evidence for the preferential site of interaction for BOH, BNA, and BNP can be gathered by an examination of the single chiral center dipeptide surfactants poly L-SULG and poly L-SUGL. In this scheme, the achiral amino acid, glycine, serves as a spacer for placing the chiral amino acid as the C-terminal or N-terminal amino acid within a given dipeptide surfactant. Note that no chiral recognition is observed for BOH or BNA for poly L-SUGL, while poly L-SULG was able to separate BOH and BNA very well (Figure 8). These data further support the hypothesis that BOH and BNA interact primarily with the inside (N-terminal) amino acid. We speculated earlier that since no chiral separation was observed for BNP with the polymeric dipeptide surfactants of opposite optical configuration [poly (L-D) SULL and poly (D-L) SULL], then BNP interacts with both chiral centers on the dipeptide surfactant. The results from our single chiral center dipeptide surfactant study are in agreement with our hypothesis. Both poly L-SULG and poly L-SUGL provided excellent resolution of BNP as shown in Figure 8. This suggests that BNP is not 4048 Analytical Chemistry, Vol. 71, No. 18, September 15, 1999

Figure 8. Comparison of the enantiomeric resolution of 1,1′-bi-2naphthol (BOH), 1,1′-bi-2-naphthyl-2,2′-diamine (BNA) and 1,1′-bi2-naphthyl-2,2′-diyl hydrogen phosphate (BNP) obtained with the single chiral center dipeptide surfactants (poly L-SULG and poly L-SUGL). Table 1. Relative Degree of Interaction of Analytes with the R Groups of the Dipeptide Surfactants on the Basis of Differences in Observed Resolution in the Presence of Diastereomeric Surfactants analyte

average Rs (LL & DD)

average Rs (DL & LD)

% interaction R1/R2a

BNA BOH BNP Prop Alp

3.4 2.6 3.3 1.7 1.2

1.7 1.3 0.0 1.2 0.5

75/25 75/25 50/50 15/85 29/71

a % interaction R /R ) eq 1/eq 2 for β-blockers, and eq 2/eq 1 for 1 2 binaphthyls. eq 1: % interaction with secondary amino acid ) 100 × {[Avg Rs(LL & DD) - Avg Rs(DL & LD)]/2}/(Avg Rs(LL & DD)). eq 2 % interaction with primary amino acid ) 100 - (% interaction with secondary amino acid).

penetrating as deeply into the core as BOH and BNA since poly L-SUGL was able to separate BNP, but not BOH or BNA. The differences in the observed enantiomeric resolutions of chiral analytes in the presence of diastereomeric surfactants can also be used to compute a rough estimate of the relative degree of interaction of the chiral analytes with each of the amino acids on the polar headgroup of the dipeptide surfactant. A rough estimate of the degree of interaction can be established by a comparison of the enantiomeric resolutions obtained with dipeptide surfactants with chiral centers of different optical configurations and those obtained with dipeptide surfactants with the same optical configurations. These results are shown in Table 1. They suggest that BNA and BOH both interact ∼25% with the outside (C-terminal) amino acid and ∼75% with the inside (N-terminal) amino acid. Since no enantiomeric separation was observed for BNP with poly (L,D) or (D,L) SULL, then the percent interaction can be assumed to be 50/50 with R1 and R2. This is confirmed by our computation in Table 1. Examination of Prop and Alp indicates that Prop interacts ∼15% with the inside (N-terminal) amino acid while Alp interacts ∼29% with the N-terminal amino acid. This indicates that factors other than hydrophobicity are also important in determining the preferential site of interaction since Prop is more hydrophobic than Alp. Steric factors are probably responsible for the observed discrepancy between Alp and Prop.

CONCLUSIONS The results of our studies suggest that polymeric diasteromeric dipeptide surfactants can be used to gain insight into some of the factors responsible for chiral discrimination with polymeric dipeptide surfactants. One of the major factors that determines chiral resolution when using polymeric dipeptide surfactants is the depth to which the analyte penetrates into the hydrophobic core of the surfactant. The depth of penetration of the analyte is governed by two major factors: the hydrophobicity of the analyte and electrostatic interactions. Examination of our data suggests that the more hydrophobic the analyte (e.g., BOH and BNA), the more it will interact with the inside (N-terminal) amino acid on the polar headgroup of the polymeric dipeptide surfactant. Thus, chiral selectivity will be governed primarily by the innermost amino acid. Conversely, if the analyte is relatively hydrophilic and/or cationic (e.g., Prop, Alp), it will interact primarily with the outside C-terminal amino acid. However, if the analyte is moderately hydrophobic (e.g., BNP), it may interact with both chiral centers on the polymeric dipeptide surfactant, and its chiral selectivity will thus be dependent on the optical configuration of both chiral centers. Furthermore, it is important to recognize that although enantiomers may associate preferentially with one of the chiral centers on the dipeptide, the interaction is not necessarily limited to that one chiral center. Analytes may interact with both chiral centers, which is evident from the decrease in resolution observed for all analytes in this study when poly (L-D) SULL and poly (D-L) SULL are compared with poly (L-L) SULL and poly (DD) SULL. While the method described in this manuscript can easily and unambiguously determine the preferential site of interaction,

factors other than hydrophobicity and electrostatics may also contribute to this preferential interaction. One factor may be the three-dimensional orientation of the surfactant-analyte complex. Studies are presently underway to examine the surfactant-analyte complex and to determine the role that rearrangement of the surfactant and/or analyte upon complexation plays in the preferential site of interaction. It is well-known that rearrangement of the selector-selectand upon complexation can play a major role in chiral selectivity. However, it is not known at this time whether this also plays a role in determining the preferential site of interaction. In summary, we believe that the insights into the chiral interactions, reported here, represent a new approach to examining chiral interactions and, thus, recognition mechanisms. Information gathered from our approach can extend beyond its use here with polymeric surfactants. For example, it may be possible to use this approach to examine schemes for asymmetric synthesis using an asymmetric catalyst. Finally, it should be noted that while the focus of this study is on hydrophobic and electrostatic interactions, we continue to develop this approach for studying hydrogen bonding and other forms of interactions. ACKNOWLEDGMENT This work was supported by Grant GM39844 from the National Institutes of Health. I. M. Warner also acknowledges the Philip W. West endowment for partial support of this research. Received for review May 19, 1999. Accepted June 1, 1999. AC990540I

Analytical Chemistry, Vol. 71, No. 18, September 15, 1999

4049