Effect of Amino Acid Order, S - ACS Publications - American Chemical

and Dobashi.2 Since these initial studies, several other papers exploring the potential of ...... The resolutions are 0, 1.2, 4.5, and 8.7 for poly L-...
14 downloads 0 Views 125KB Size
Download PDF
Anal. Chem. 2000, 72, 1740-1748

Examination of Structural Changes of Polymeric Amino Acid-Based Surfactants on Enantioselectivity: Effect of Amino Acid Order, Steric Factors, and Number and Position of Chiral Centers Eugene Billiot† and Isiah M. Warner*

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

In this study, a large number of polymeric chiral surfactants were examined and their performances in terms of enantiomeric resolution compared for a variety of chiral analytes. The surfactants investigated in this study include all possible dipeptide combinations of the L-form of alanine, valine, leucine, and the achiral amino acid glycine (except glycine-glycine). Also included in this study were the single amino acid surfactants of alanine, valine, and leucine as well as the single chiral center dipeptide surfactant poly(sodium undecyl-L-leucine-β-alanine) (poly L-SULβA). Several different aspects of polymeric dipeptide surfactants, as they pertain to chiral separations, are examined. Some of the factors investigated in this report include the effect of position and number of chiral centers, amino acid order, and steric effects. The use of polymeric chiral surfactants for enantiomeric separations using capillary electrokinetic chromatography (EKC) has been reported by Wang and Warner1 in 1994 and also by Hara and Dobashi.2 Since these initial studies, several other papers exploring the potential of polymeric chiral surfactants for enantiomeric separations with EKC have been reported. In a subsequent paper, Wang and Warner investigated the use of polymeric chiral surfactants in combination with γ-cyclodextrin (γ-CD).3 A synergistic effect was observed for the chiral separation of four enantiomeric pairs with γ-CD in combination with poly(sodium undecyl-D-valine) (poly D-SUV). Later, two publications (one by our group and another by Dobashi et al.) extended the range of chiral analytes to be separated with poly L-SUV.4,5 In the next logical progression of this work, Shamsi et al. compared the single amino acid polymeric surfactant poly L-SUV to the polymeric dipeptide surfactant poly(sodium undecyl-(L,L)valine-valine) (poly (L,L) SUVV).6 The polymeric dipeptide surf† Present address: Department of Physical and Life Sciences, Texas A&M UniversitysCorpus Christi, 6300 Ocean Drive, CS 212, Corpus Christi, TX, 78412. (1) Wang, J.; Warner, I. M. Anal. Chem. 1994, 66, 3773-3776. (2) Hara, S.; Dobashi, A. Jpn. Pat. 04 149 205, 1993; Chem. Abstr. 1993, 118, p39405z. (3) Wang, J.; Warner, I. M. J. Chromatogr. A 1995, 711, 297-304. (4) Agnew-Heard, K.; Sanchez Pena, M.; Shamsi, S.; Warner I. M. Anal. Chem. 1997, 69, 958-964. (5) Dobashi, A.; Hamada, M.; Dobashi, Y. Anal. Chem. 1995, 67, 3011-3017.

1740 Analytical Chemistry, Vol. 72, No. 8, April 15, 2000

actant poly (L,L) SUVV demonstrated a significant improvement in chiral selectivity for three out of the four analytes examined when compared to the single amino acid surfactant poly L-SUV. To better understand the synergistic effect observed with the dipeptide surfactants compared to the single amino acid surfactants, our research group initiated an extensive study using a large group of polymeric dipeptide surfactants. In the first report investigating the synergistic effect of dipeptide surfactants, the effect of amino acid order was examined.7 The primary surfactants used in that study were poly(sodium undecyl-(L,L)-leucine-valine) (poly (L,L) SULV) and poly(sodium undecyl-(L,L)-valine-leucine) (poly (L,L) SUVL). The results of that study showed that the amino acid order has a dramatic effect on the chiral selectivity of the surfactant. In a more recent report by our group, a proposed structure of the polymeric dipeptide surfactants was discussed.8 It was postulated that the conformation of the dipeptide surfactants in solution is dependent on two primary factors (hydrophobic and steric). On the basis of hydrophobic interactions alone, both of the two R-groups would tend to face the hydrophobic core of the polymeric surfactant rather than be exposed to the bulk aqueous phase. However, due to steric considerations, the smallest R-group may be forced to twist toward the aqueous phase. There are two major implications of the proposed structure as far as chiral recognition is concerned. First, if the larger of the two amino acids in the dipeptide surfactant is in the N-terminal position, then the second (C-terminal) amino acid can act as a “finger” to help hold the analyte, restricting its movement. Second, if the larger amino acid is in the C-terminal position, it could block access to the first chiral center resulting in a significant decrease in the chiral selectivity of the surfactant for large bulky analytes such as the binaphthyl derivatives investigated in that report. In a subsequent paper, the preferential binding site of various analytes was investigated with diastereomeric dipeptide surfactants.9 In that study, it was determined that one of the major factors (6) Shamsi, S.; Macossay, J.; Warner, I. M. Anal. Chem. 1997, 69, 2980-2987. (7) Billiot, E.; Macossay, J.; Thibodeaux, S.; Shamsi, S.; Warner, I. M. Anal. Chem. 1998, 70, 1375-1381. (8) Billiot, E.; Agbaria, R. A.; Thibodeaux, S.; Shamsi, S.; Warner I. M. Anal. Chem. 1999, 71, 1252-1256. (9) Billiot, E.; Thibodeaux, S.; Shamsi, S.; Warner I. M. Anal. Chem. 1999, 71, 4044-4049. 10.1021/ac9908804 CCC: $19.00

© 2000 American Chemical Society Published on Web 03/10/2000

which determines chiral selectivity with polymeric dipeptide surfactants is the depth of penetration of the analyte into the hydrophobic core of the surfactant. The depth of penetration determines which part of the polar headgroup on the polymeric dipeptide surfactant with which the analyte will preferentially interact. In this study, 19 polymeric chiral surfactants are examined with a variety of chiral analytes. The purpose of this study was to gain deeper insight into the factors governing the enantioselectivity of polymeric amino acid-based surfactants. The major interactions which govern the enantioselectivity and the binding of the analyte to the surfactants are hydrophobic/hydrophilic interactions,5,10-11 electrostatic forces,5 hydrogen bonding,5,11-14 and steric factors.10,13,15 The major factor (in the absence of electrostatic attraction) governing the binding of the analyte to the surfactant and the preferential site of interaction of the analyte to the polar headgroup is hydrophobic/hydrophilic interactions. As stated earlier, the hydrophobic forces will determine the depth of penetration of the analyte into the core of the surfactant. The hydrophobic/hydrophilic interactions will also govern the orientation of the analyte into the hydrophobic core. The preferred orientation of the analyte will be with the more polar region of the molecule facing the bulk aqueous phase and the hydrophobic portion of the analyte directed toward the hydrophobic core. The hydrophobic interactions along with steric considerations also play a major role in the preferred configuration of the dipeptide surfactant in solution. Since amino acid-based surfactants do not possess very strong π-characteristics, the major attractive force (absent electrostatic attraction) of the analyte to the polar headgroup of the surfactant is hydrogen bonding. The enantioselectivity at the preferential site of interaction is then governed primarily by hydrogen bonding and steric factors near the stereogenic center of the surfactant, as well as the analyte. The results of this study yields insight into the role of three out of four of the major interactions involved in chiral selectivity of polymeric amino acid-based surfactants (hydrophobic/hydrophilic interactions, electrostatic forces, and steric factors). EXPERIMENTAL SECTION 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), alprenolol (Alp), oxprenolol (Oxp), temazepam (Temaz), lorazepam (Loraz), oxazepam (Oxaz), glutethimide (Glut), aminoglutethimide (Amino), and trifluoranthrylethanol (TFAE) were purchased from Aldrich (Milwaukee, WI). The structures of all the chiral analytes examined in this study are shown in Figure 1. Tris(hydroxymethyl)aminomethane (TRIS), 3-(cyclohexylamino)1-propanesulfonic acid (CAPS), and sodium borate were obtained (10) Dobashi, A.; Ono, T.; Hara, S.; Yamaguchi, J. Anal. Chem. 1989, 61, 19841986. (11) Tambute, A.; Begos, A.; Lienne, M.; Caude, M.; Rosset, R. J. Chromatogr. 1987, 396, 65-81. (12) Tickle, D. C.; Okafu, G. N.; Camilleri, P.; Jones, R. F. D.; Kirby, A. J. Anal. Chem. 1994, 66, 4121. (13) Cohen, S. A.; Paulus, A.; Karger, B. L. Chromatographia 1987, 24, 15-24. (14) Dobashi, A.; Ono, T.; Hara, S.; Yamaguchi, J. J. Chromatogr. A 1989, 480, 413-420. (15) Peterson, A. G.; Ahuja, E. S.; Foley, J. P. J. Chromatogr. B 1996, 683, 1528.

Figure 1. Structure of analytes examined in this study.

from Fisher Scientific Co. (Fair Lawn, NJ) and used as received. Chemicals used for the synthesis of surfactants included N,N′dicyclohexylcarbodiimide, N-hydroxysuccunimide, undecylenic acid, various amino acids, and the dipeptides. All were supplied by Sigma (St. Louis, MO) and used as received. Synthesis of Polymeric Surfactants. All surfactants in this study were synthesized using the procedure reported by Wang and Warner.1 Surfactant monomers were prepared by mixing the N-hydroxysuccinimide ester of undecylinic acid with the amino acid or dipeptide to form the corresponding N-undecylenyl chiral surfactant. Polymerization was achieved by 60Co γ-irradiation. All polymers used in this study were found to be 99% pure or better as estimated from elemental analysis. All the surfactants examined in this study and their abbreviations are listed in Table 1. In addition, a diagram depicting the general structure of the dipeptide surfactants is shown in Figure 2. Capillary Electrophoresis Procedure. The EKC experiments were conducted on a Hewlett-Packard 3DCE model no. G1600AX. An untreated fused silica capillary (effective length 55 cm, 50 µm i.d.) was purchased from Polymicro Technologies (Phoenix, AZ). Separations were performed at +30 kV, with UV detection at 215 nm. The temperature of the capillary was maintained at 25 °C for BOH, BNA, and BNP and 12 °C for the rest of the analytes by the instrument thermostating system, which consists of a Peltier element for forced air cooling and temperature control. All samples were prepared in 50:50 methanol/H2O. The concentration of some of the samples (BOH, BNA, BNP, TAFE, Temaz, Loraz, and Oxaz) was 0.1 mg/mL. The concentrations of the other analytes (Alp and Oxp) were 0.5 mg/mL, and they were 0.2 mg/mL for Prop, Amino, and Glut. The samples were injected for 5 s with 10 mbar of pressure. Prior to use, the new capillary Analytical Chemistry, Vol. 72, No. 8, April 15, 2000

1741

Table 1. List of Surfactants Examined in This Study and the Abbreviations Used To Describe Them surfactant sodium N-undecylalanine sodium N-undecylvaline sodium N-undecylleucine sodium N-undecylglycine-alanine sodium N-undecylglycine-valine sodium N-undecylglycine-leucine sodium N-undecylalanine-glycine sodium N-undecylalanine-alanine sodium N-undecylalanine-valine sodium N-undecylalanine-leucine sodium N-undecylvaline-glycine sodium N-undecylvaline-alanine sodium N-undecylvaline-valine sodium N-undecylvaline-leucine sodium N-undecylleucine-glycine sodium N-undecylleucine-alanine sodium N-undecylleucine-valine sodium N-undecylleucine-leucine sodium N-undecylleucine-β-alanine

abbreviations used Ala Val Leu GA GV GL AG AA AV AL VG VA VV VL LG LA LV LL LβA

L-SUA L-SUV L-SUL L-SUGA L-SUGV L-SUGL

(L,L) SUAG (L,L) SUAA (L,L) SUAV (L,L) SUAL (L,L) SUVG (L,L) SUVA (L,L) SUVV (L,L) SUVL (L,L) SULG (L,L) SULA (L,L) SULV (L,L) SULL L-SULβA

Figure 2. General structure of dipeptide surfactants.

was conditioned for 30 min with 1 M NaOH followed by 30 min of 0.1 M NaOH. Then, the capillary was rinsed for 15 min with deionized water. Prior to each run, the buffer was pressure injected through the column for 2 min to condition and fill the capillary. Preparation of EKC Buffer Solutions. The background electrolyte (BGE) for the benzodiazepams (Temaz, Loraz, and Oxaz) was 25 mM sodium borate and 25 mM TRIS at pH 8.5. For Amino and Glut the BGE was 50 mM TRIS at pH 9.2, and for TFAE the BGE was 30 mM sodium borate at pH 10. The BGE for the binaphthyl derivative experiments was 10 mM sodium borate and 100 mM TRIS at pH 10.0. The BGE for the cationic β-blockers was 50 mM sodium borate and 300 mM CAPS at pH 8.5. CAPS was added to minimize capillary wall interaction. An appropriate amount of the polymeric surfactants was then added to the BGE and the pH readjusted with 1 M NaOH or 1 M HCl if necessary. We deliberately chose buffer conditions that would minimize run time. Most of the analytes (7 out of 12) eluted in less than 10 min. Of the remaining 5, 4 eluted in 12 min or less and 1 (Prop) eluted in approximately 17 min. 1742

Analytical Chemistry, Vol. 72, No. 8, April 15, 2000

RESULTS AND DISCUSSIONS In this study, a large number of polymeric chiral surfactants are examined and their performances in terms of enantiomeric resolution compared for a variety of chiral analytes. We compare resolution values rather than the separation factors because we feel a comparison of resolution values, in this case, is more beneficial and illuminating than comparing separation factors. For instance, it is possible to have a very large separation factor and no chiral resolution if the efficiency is very poor. In addition, our studies have shown that, for all practical purposes, the separation efficiency is the same when comparing single analytes with various polymeric amino acid-based surfactants. Thus, resolution values and separation factors follow the same trends. A list of all of the analytes and their corresponding resolution values for various polymeric surfactants is given in Table 2. (A) Effect of Number and Position of Chiral Centers in Dipeptide Surfactants on Enantioselectivity. The role of the second chiral center in enantiomeric separations with polymeric dipeptide surfactants was investigated by comparing dipeptide surfactants with only one chiral center to various dipeptides surfactants with two chiral centers. The single chiral center dipeptide surfactants (SCCDSs) which were examined in this investigation are poly L-SUAG, poly L-SUGA, poly L-SUVG, poly L-SUGV, poly L-SULG, poly L-SUGL, and poly L-SULβA. The performance (in terms of enantiomeric resolution) of these polymeric surfactants was compared to the corresponding two chiral center dipeptide surfactants (TCCDSs) with alanine in place of the achiral amino acids glycine or β-alanine. The TCCDSs examined in this section are poly (L,L) SUAA, poly (L,L) SUVA, poly (L,L) SUAV, poly (L,L) SULA, and poly (L,L) SUAL. As discussed in the introduction, a previous study examined the effect of optical configuration order in diastereomeric dipeptide surfactants.9 The results of that study suggested that one of the factors which determines the enantioselectivity of chiral analytes with polymeric dipeptide surfactants is the depth of penetration of the analyte into the core of the polymeric surfactant. The depth of penetration of the analyte determines which part of the polar headgroup (N-terminal or C-terminal amino acid) with which the analyte will preferentially interact. The depth of penetration of the analyte is generally governed by two main factors. These factors are hydrophobicity and electrostatic interactions. The use of SCCDSs can also be used to gain insight into the preferential site of interaction. With SCCDSs, the depth of penetration of the analyte into the hydrophobic core of the dipeptide surfactant is particularly important since only one of the amino acids is chiral. When the inside (N-terminal) amino acid (R1 of Figure 2) is achiral, little or no chiral separation would be expected if the analyte penetrates deep into the core of the surfactant. Conversely, if the analyte interacts preferentially at the aqueous interface, then chiral selectivity will be dependent on the chirality of the outside (C-terminal) amino acid (R2 of Figure 2). (1) Binaphthyl Derivatives. A comparison of the enantiomeric separation of the binaphthyl derivatives with SCCDSs is in agreement with the previously mentioned optical configuration study. The data suggest that BOH and BNA interact preferentially with the inside (N-terminal) amino acid and BNP interacts closer to the bulk aqueous phase with both of the chiral centers on

Table 2. Resolution Values for the Various Analytes with the Different Polymeric Surfactants Examined in This Study

Ala Val Leu GA GV GL AG AA AV AL VG VA VV VL LG LA LV LL LBA

BNP

BOH

BNA

TFAE

AMINO

GLUT

OXP

ALP

PROP

TEMAZ

OXAZ

LORAZ

0.00 0.00 1.08 0.00 1.87 4.64 0.00 1.19 0.54 1.18 2.78 4.47 1.92 0.00 8.06 8.68 6.94 4.74 8.26

6.92 5.09 5.54 1.32 0.64 0.00 5.94 6.09 0.88 0.75 3.95 5.15 3.39 2.23 6.17 6.44 5.19 3.36 5.03

3.86 5.72 5.73 0.65 0.00 0.00 3.00 2.93 1.72 1.96 3.40 3.94 4.29 4.19 4.42 4.72 6.37 6.24 3.31

0.00 2.28 1.88 0.00 0.00 0.00 0.00 0.65 1.48 1.39 0.76 1.43 1.75 1.44 0.00 0.00 1.02 0.00 0.00

3.84 5.28 5.04 1.19 2.09 2.70 1.51 2.19 3.05 3.00 2.48 4.97 6.37 6.72 2.29 4.17 8.41 7.52 2.68

0.93 0.48 0.00 0.77 1.14 1.39 1.80 1.84 1.28 0.61 1.79 2.07 1.44 0.96 1.48 1.39 1.40 0.75 1.12

1.00 1.31 1.93 1.29 1.46 1.82 0.00 0.80 1.06 1.20 0.00 0.91 1.10 1.27 0.00 1.01 1.11 1.75 0.00

0.51 0.00 1.57 1.18 1.13 1.94 0.00 0.66 0.79 1.79 0.00 1.25 1.50 2.14 0.00 1.05 0.82 1.54 0.00

0.64 0.00 1.53 1.71 1.37 2.62 0.00 1.20 1.39 2.02 0.00 1.60 2.02 2.91 0.60 1.07 1.00 1.82 0.00

0.00 2.29 2.68 0.84 1.17 2.69 1.10 0.00 1.06 3.77 0.00 0.87 0.69 2.11 0.64 2.29 3.36 3.85 0.72

1.33 0.61 1.53 0.64 0.62 0.85 1.19 0.00 0.00 2.24 0.58 0.81 0.59 0.00 0.61 0.62 1.43 0.00 0.00

0.00 0.92 1.10 0.61 0.00 0.00 1.56 1.56 0.71 2.52 2.99 2.65 1.00 0.00 1.62 1.45 2.51 1.53 1.65

dipeptide surfactants. When the N-terminal amino acid of the SCCDS is chiral, a significant improvement in resolution is observed for BOH and BNA compared to the SCCDS when the inside (N-terminal) amino acid is achiral. A comparison of poly L-SUAG to poly L-SUGA reveals that the resolution is about 4 times greater for BOH and BNA for poly L-SUAG compared to the SCCDS when the inside amino acid is achiral, (i.e. poly L-SUGA). The SCCDS poly L-SUAG separated the enantiomers of BOH and BNA with a resolution of around 6 and 3, respectively. In contrast, the SCCDS poly L-SUGA yielded resolutions of about 1.3 and 0.7, respectively. The same trends are also observed when valine or leucine are the N- or C-terminal amino acid of a SCCDS. When the N-terminal amino acid of the SCCDS is chiral, such as in poly L-SUVG, poly L-SULβA, and poly L-SULG, the enantiomeric resolution is much greater (> 6 times) for BOH and BNA than the corresponding SCCDSs with an achiral N-terminal amino acid, i.e., poly L-SUGV and poly L-SUGL. Further evidence that BOH and BNA bind preferentially to the inside (N-terminal) amino acid is seen in a comparison of the TCCDSs to the SCCDSs. No dramatic difference in chiral selectivity is observed for BOH and BNA when the inside amino acid is chiral and the chirality of the outside amino acid is either chiral or achiral. The resolution is always less when the outside amino acid is achiral, but in general the difference is relatively small compared to other factors such as amino acid order (which will be discussed later) and when the inside (N-terminal) amino acid is achiral. The resolutions of BNA and BOH are about the same for the TCCDS poly (L,L) SUAA as it is with the SCCDS poly L-SUAG. The resolution for BNA with both surfactants is around 3 and approximately 6 for BOH. This trend is also observed for BOH and BNA with poly (L,L) SUVA compared to poly L-SUVG and with poly (L,L) SULA compared to poly L-SULβA and poly L-SULG. Because BNP is anionic under the experimental conditions used while BOH and BNA are essentially neutral, BNP is more hydrophilic than BOH and BNA. The difference in hydrophobicity results in BNP interacting with the polar headgroup of the dipeptide surfactant closer to bulk aqueous phase (i.e. closer to

the C-terminal amino acid) than BOH and BNA. Evidence of this is seen in a comparison of the SCCDSs by varying the position (N-terminal or C-terminal) of the achiral amino acid. The difference in resolution for BNP with poly L-SUVG and poly L-SUGV is not as dramatic as compared to BOH and BNA. Poly L-SUVG separated BNP with a resolution of 2.8, while poly L-SUGV separated the enantiomers of BNP with a resolution of around 1.9. It can also observed that poly L-SUGL was not able to separate the enantiomers of BOH and BNA while BNP was separated with a resolution of almost 5. These results clearly suggest that BNP does interact strongly with the outside (C-terminal) amino acid while little or no interaction with the C-terminal amino acid was observed with BOH and BNA. (2) Propranolol, Alprenolol, and Oxprenolol. A comparison of the SCCDSs poly L-SUAG and poly L-SUGA shows that when the outside (C-terminal) amino acid is achiral (i.e. poly L-SUAG), no chiral resolution was observed for all three analytes Alp, Prop, or Oxp. However, when the outside amino acid is chiral (i.e. poly L-SUGA), a resolution of at least 0.7 or better is achieved for all three enantiomeric pairs. The same trend is observed in a comparison of poly L-SUVG and poly L-SUGV. While no resolution of the enantiomers of Alp, Prop, and Oxp was observed with poly L-SUVG, when the second amino acid is chiral (poly L-SUGV), enantiomeric separation of all three β-blockers is achieved. Only one exception to this general trend of no chiral separation with the β-blockers when the C-terminal amino acid is achiral was observed. This exception was with Prop and the dipeptide surfactant poly L-SULG. Propranolol was separated with a resolution of around 0.6 with poly L-SULG. Examination of the data suggest that the β-blockers investigated in this study bind preferentially to the outside C-terminal amino acid. Electrostatic attraction between these analytes and the anionic C-terminal amino acid on the dipeptide surfactant is believed to be a major factor in binding of the β-blockers to the polar headgroup of the surfactants under study. Alprenolol, Prop, and Oxp are cationic under the experimental conditions used (pH 8.5). The values of the pKa’s for Prop, Alp, and Oxp are ∼9.2. Although electrostatic attraction plays the key role in the preferAnalytical Chemistry, Vol. 72, No. 8, April 15, 2000

1743

ential site of interaction with these analytes, hydrophobicity also plays an important role. The significance of hydrophobicity is observed with the chiral separation of Prop with the SCCDS poly L-SULG. As stated earlier, with the exception of Prop with poly L-SULG, none of these analytes show any sign of enantiomeric separation when the outside C-terminal amino acid is achiral. Note that this exception only occurs when leucine is the N-terminal amino acid and only with Prop. The reason for this apparent anomaly is believed to be due to hydrophobicity. Propranolol, with two aromatic rings, is the most hydrophobic of the β-blockers examined compared to the other two β-blockers (Alp and Oxp) in this study with only one aromatic ring. The dipeptide surfactants containing leucine are also more hydrophobic than the equivalent dipeptide surfactants containing alanine or valine. Thus, the hydrophobic interactions of Prop with poly L-SULG would be greater than the other surfactants and analytes discussed in this section. Therefore, Prop penetrates deeper into the core and interacts sufficiently with the N-terminal amino acid to enable some degree of chiral separation. Another possible explanation for the apparent anomaly with Prop could be steric factors. Propranolol is the largest and most sterically hindered of the β-blockers examined in this study, and poly L-SULG is more sterically hindered than poly L-SUAG and poly L-SUAV. Therefore, an increase in stereoselectivity of the analyte and the pseudostationary may be responsible for the enantiomeric separation observed with poly L-SULG and Prop. In addition to binding to the outside C-terminal amino acid, it appears that the enantiomeric separation of these analytes is favored by SCCDSs over TCCDSs. When the amino acid in the C-terminal position is held constant and the amino acid in the N-terminal position is varied (either glycine or alanine), the SCCDS produce better separation of the enantiomers of Oxp, Alp, and Prop than the dipeptide surfactants which contain two chiral centers. For example, poly L-SUGA separates the enantiomers of Oxp, Alp, and Prop better than poly L-SUAA. The resolution values for Alp, Oxp, and Prop with poly L-SUGA are 1.29, 1.18, and 1.71, respectively, and 0.80, 0.66, and 1.20, respectively, for poly L-SUAA. In addition, poly L-SUGV yields resolutions as good or better than poly L-SUAV, and poly L-SUGL is better than poly L-SUAL for the enantiomeric separation of these analytes. Since both chiral centers are of the same optical configuration, the decrease in resolution is assumed to be due to steric factors. The R-group attached to the inside N-terminal amino acid decreases the chiral selectivity of the dipeptide surfactant. (3) TFAE, Glutethimide, and Aminoglutethimide. No significant differences were observed in the enantiomeric resolution of Glut or Amino with position of the achiral amino acid using SCCDS. The enantiomeric resolutions with poly L-SUAG and poly L-SUGA were about 1.5 and 1.2 for Amino and 1.8 and 0.8 for Glut, respectively. Similarly, poly L-SUVG and poly L-SUGV yielded resolutions of 2.5 and 2.1 for Amino and 1.8 and 1.1 for Glut, respectively. In addition, poly L-SULβA, poly L-SULG, and poly L-SUGL separated the enantiomers of Amino with resolutions of 2.7, 2.3, and 2.7, respectively. These results suggest that Amino and Glut interact with both chiral centers of the dipeptide surfactant rather than interacting preferentially with one over the other. 1744

Analytical Chemistry, Vol. 72, No. 8, April 15, 2000

A comparison of the SCCDSs to the TCCDSs reveals that the second chiral center does not appear to play a significant role in the enantiomeric resolution of Glut. The SCCDS poly L-SUAG has approximately the same resolution for Glut (Rs ) 1.80) as the TCCDS poly L-SUAA (Rs ) 1.84). However, a relatively small decrease in resolution is observed with the TCCDS poly L-SUVA compared to the SCCDS poly (L,L) SUVG and with the TCCDS poly (L,L) SUAL compared to the SCCDSs poly L-SULβA, poly L-SULG, and poly L-SUGL. In contrast, the enantiomeric separation of Amino does appear to be slightly better with TCCDS compared to SCCDS especially with TCCDS with the larger of the two amino acids in the inside (N-terminal) position. Poly (L,L) SUVA separated the enantiomers of Amino better than the SCCDS poly L-SUVG and poly L-SUGV, (Rs ) 4.97, 2.48, and 2.09 respectively). Similarly, poly (L,L) SULA resolved the enantiomers of Amino better than the SCCDSs poly L-SULβA, poly L-SULG, and poly L-SUGL. However, not much of an improvement occurs when the amino acid order of the TCCDS is such that the larger of the two amino acids is in the outside (C-terminal) position. The enantiomeric resolution of Amino with the TCCDS poly (L,L) SUAV is not significantly different than the SCCDS poly L-SUVG and poly L-SUGV (Rs ) 3.05, 2.48, 2.09, respectively). Likewise, the SCCDSs poly L-SULβA, poly L-SULG, and poly L-SUGL separated the enantiomers of Amino with resolutions not substantially different from the TCCDS poly (L,L) SUAL. Only one of the SCCDS examined in this report showed any sign of chiral recognition of TFAE. Poly L-SUVG separated the enantiomers of TFAE with a resolution of 0.8. Since the enantiomers of TFAE were separated with just one of the SCCDS, no conclusions with respect to preferential site of interaction can be inferred. Also, since poly L-SUVG was the only SCCDS able to enantiomerically resolve TFAE and four out of the five TCCDS examined in this section were able to separate the enantiomers of TFAE, it appears that the enantiomeric separation of TFAE is favored by TCCDS over SCCDS. (4) Temazepam, Oxazepam, and Lorazepam. The enantiomeric resolutions of Temaz and Oxaz do not follow any definite trends with respect to position of chiral center with SCCDSs or with respect to number of chiral centers. The TCCDS poly (L,L) SUAA was not able to resolve the enantiomers of Temaz or Oxaz while both of the SCCDSs poly L-SUAG and poly L-SUGA did. A comparison of the SCCDS reveals that poly L-SUVG was not able to separate Temaz, but the other SCCDS (poly L-SUGV) yielded slightly higher resolutions than the two TCCDSs poly (L,L) SUVA and poly (L,L) SUAV. In contrast, both of the SCCDSs (poly L-SUVG and poly L-SUGV) were able to separate the enantiomers of Oxaz but only one of the TCCDS (poly (L,L) SUVA) could. Since the enantiomeric separations of Temaz and Oxaz do not follow any particular trends with respect to position of chiral center, no inference can be made as to the preferential site of interaction of these analytes to the polar headgroup of the dipeptide surfactants. On the other hand, the preferential site of interaction of Loraz appears to be with the inside N-terminal amino acid. When the inside (N-terminal) amino acid of the SCCDS is chiral (i.e. poly L-SUAG) an improvement in enantiomeric resolution of Loraz is observed compared to the SCCDS with an achiral N-terminal

amino acid (i.e. poly L-SUGA) (Rs ) 1.56 and 0.61, respectively). More conclusively, the SCCDSs poly L-SUGV and poly L-SUGL were not able to resolve the enantiomers of Loraz. In contrast, the SCCDSs with chiral N-terminal amino acids (poly L-SUVG, poly L-SULβA, and poly L-SULG) were all able to baseline resolve the enantiomers of Loraz. A comparison of the enantiomeric resolution of Loraz with TCCDS to SCCDS shows that the second chiral center does not play a significant role when the inside amino acid of the SCCDS is chiral and the larger of the two amino acids in the TCCDS is in the N-terminal position. The SCCDS poly L-SUAG separated the enantiomers of Loraz, as well as the TCCDS poly (L,L) SUAA. Both surfactants separated the enantiomers of Loraz with a resolution of 1.56. A slight increase in resolution occurs with the SCCDS poly L-SUVG compared to the TCCDS poly (L,L) SUVA, and a similar small increase in the enantiomeric separation of Loraz is also observed with the SCCDSs poly L-SULβA and poly L-SULG compared to the TCCDS poly (L,L) SULA. No consistent trends are observed when the larger of the amino acids of the TCCDS is in the C-terminal position. (B) Effect of Amino Acid Order. The order of amino acids in dipeptide surfactants has been shown to have a significant effect on the chiral selectivity, as well as the physical properties of the surfactant.7,8 It was reported that the final conformation of the polymeric dipeptide surfactant is governed by the hydrophobicity of the R-groups on the amino acids and steric factors. On the basis of hydrophobic interactions, the two hydrophobic groups of the dipeptide would tend to face the inner core of the micellar structure rather than be exposed to the bulk aqueous phase. However, the packed configuration of the dipeptide would not allow this preferred conformation to occur due to steric hindrance. Therefore, the smallest R-group would be forced to twist toward the aqueous phase. It was further shown that the preferred configuration of dipeptide surfactants for the enantiomeric separation of large hydrophobic bulky analytes such as binaphthyls was with the larger of the two amino acids in the inside (N-terminal) position. This section reports on the ongoing study of the effect of amino acid order on enantiomeric separation. (1) BNP, BOH, and BNA. The previously mentioned studies7,8 were conducted with BNP and BOH. This section also examines BNA. Since the effect of amino acid order was discussed in great detail in the aforementioned references, the discussion here will be brief. The only new point of interest is to show that BNA follows the same trend as BNP and BOH. The effect of amino acid order on chiral selectivity is clearly demonstrated in a comparison of poly (L,L) SUVA with poly (L,L) SUAV. Poly (L,L) SUVA, for instance, was able to separate the enantiomers of BNP and BOH with resolutions of around 4.5 and 5.0, respectively. In contrast, with poly (L,L) SUAV resolutions of about 1.0 or less are observed. Although the differences in resolution were not as great with BNA, the resolutions are at least twice as good for BNA when the inside (N-terminal) amino acid is larger than the outside (Cterminal) amino acid, i.e., poly (L,L) SUVA compared to poly (L,L) SUAV (Rs ) 3.94 and 1.72, respectively). Similarly, poly (L,L) SULA separated the enantiomers of BOH, BNP, and BNA better than poly (L,L) SUAL as did poly (L,L) SULV compared to poly (L,L) SUVL. In all of the cases observed, the preferred configuration of the dipeptide surfactants for the enantiomeric separation of the

binaphthyl derivatives is when the larger of the two amino acids is in the first position (R1 of Figure 2). (2) Propranolol, Alprenolol, and Oxprenolol. In the previous section it was stated that the preferred configuration of dipeptide surfactants for the enantiomeric separation of large bulky analytes such as the binaphthyl derivatives was with the larger of the two amino acids in the inside (N-terminal) position. The opposite trend is observed for the less bulky, less hydrophobic β-blockers examined in this section. With the exception of Alp and Prop with poly (L,L) SUAV and poly (L,L) SUVA, the enantiomeric separation of the β-blockers examined in this section appear to be favored by TCCDSs with the larger of the two amino acids in the outside (C-terminal) position. All three analytes (Oxp, Alp, and Prop) are separated better with poly (L,L) SUAL (Rs ) 1.20, 1.79, and 2.02, respectively) compared to poly (L,L) SULA (Rs ) 0.00, 0.00, and 0.60, respectively). The same trends are observed with poly (L,L) SUVL as compared to poly (L,L) SULV. (3) TFAE, Glutethimide, and Aminoglutethimide. When the larger of the two amino acids of the dipeptide surfactant is in the N-terminal position, an improvement in enantiomeric resolution of Amino and Glut is observed compared to the corresponding dipeptide surfactants with the reverse amino acid order. Poly (L,L) SUVA separated the enantiomers of Amino and Glut (Rs ) 4.97 and 2.07, respectively) better than poly (L,L) SUAV (Rs ) 3.05 and 1.28, respectively). Likewise poly (L,L) SULA separated the enantiomers of Amino and Glut better than poly (L,L) SUAL, as did poly (L,L) SULV compared to poly (L,L) SUVL. While the differences in resolution are not always dramatic, it appears that the enantiomeric separation of Amino and Glut is favored by dipeptide surfactants with the larger of the two amino acids in the inside (N-terminal) position. A comparison of the effect of amino acid order on the enantiomeric resolution of TFAE yielded mixed results. Poly (L,L) SUAV and poly (L,L) SUVA separated the enantiomers of TFAE with approximately the same resolution (Rs ) 1.48 and 1.43, respectively). Similarly, no significant difference in resolution of TFAE was observed with poly (L,L) SUVL compared to poly (L,L) SULV. However, poly (L,L) SUAL separated TFAE with a resolution of approximately 1.4 while poly (L,L) SULA showed no sign of enantiomeric resolution of TFAE. Unlike Amino and Glut, the enantiomeric separation of TFAE follows no definite trends with respect to amino acid order. (4) Temazepam, Oxazepam, and Lorazepam. Similar to the previous section, examination of the effect of amino acid order on the enantiomeric separation of Temaz, Ozax, and Loraz reveals no definite trends with respect to maino acid order. Poly (L,L) SUVA (with the larger of the two amino acids on the inside (Nterminal) position) separated the enantiomers of Oxaz and Loraz (Rs ) 0.81 and 2.65, respectively) better than poly (L,L) SUAV, which has the reverse amino acid order (Rs ) 0.00 and 0.71, respectively). However, poly (L,L) SUVA and poly (L,L) SUAV separated the enantiomers of Temaz with no significant difference in resolution (Rs ) 0.87 and 1.06, respectively). However, an improvement in chiral selectivity of all three analytes (Temaz, Oxaz, and Loraz) is observed with poly (L,L) SULV (Rs ) 3.36, 1.43, and 2.51) as compared to poly (L,L) SUVL (2.11, 0.00, and 0.00, respectively). A reversal of this trend is seen in a comparison Analytical Chemistry, Vol. 72, No. 8, April 15, 2000

1745

of poly (L,L) SUAL to poly (L,L) SULA. The dipeptide surfactant with the smaller of the two amino acids in the N-terminal position (i.e. poly (L,L) SUAL) yielded higher resolution values (Rs ) 3.77, 2.24, and 2.52, respectively) for all three enantiomeric pairs compared to poly (L,L) SULA (Rs ) 2.29, 0.62, and 1.45, respectively), which has the larger of the two amino acids in the N-terminal position. While the enantiomeric separation of the benzodiazepams studied in this section do not appear to follow consistent trends with respect to amino acid order, all three enantiomeric pairs are relatively consistent with respect to individual amino acid pairs. In other words, all three benzodiazepams were enantiomerically resolved better with poly (L,L) SUAL compared to poly (L,L) SULA and with poly (L,L) SULV compared to poly (L,L) SUVL. (C) Investigation of Steric Effects. This section deals with the effect of steric factors on enantiomeric resolution with polymeric dipeptide surfactants using a series of surfactants with varying size of amino acids in the first and/or second position of the dipeptide surfactant. The effects of steric factors are also investigated by comparing various single amino acid surfactants and dipeptide surfactants with the same amino acid in both positions such as poly (L,L) SUAA, poly (L,L) SUVV, and poly (L,L) SULL. (1) BNP, BOH, and BNA. An examination of the steric effects on the enantiomeric separation of BNP, BOH, and BNA reveals that the resolution of BOH and BNA increases significantly as the size of the inside (N-terminal) amino acid is increased. The resolution of BOH is 0.6, 0.9, 3.4, and 5.2 for poly L-SUGV, poly (L,L) SUAV, poly (L,L) SUVV, and poly (L,L) SULV, respectively. An even greater enhancement in resolution is observed for BNA with the same surfactants. The resolution of BNA went from 0 for poly L-SUGV to 1.7, 4.3, and 6.4 for poly (L,L) SUAV, poly (L,L) SUVV, and poly (L,L) SULV, respectively. The same general trends are observed for the other surfactants examined in this study. The only exception to this trend is poly (L,L) SUAA for BOH. Slightly different trends are observed for BNP. When alanine is held constant in the outside (C-terminal) position, the same general trends are observed as with BOH and BNA. As the size of the inside amino acid is increased, the resolution of BNP also increases. The resolutions are 0, 1.2, 4.5, and 8.7 for poly L-SUGA, poly (L,L) SUAA, poly (L,L) SUVA, and poly (L,L) SULA, respectively. Interestingly, as the size of the outside amino acid increases from valine to leucine, the effect of the size of the inside amino acid on the chiral selectivity of BNP changes. When valine is held constant in the outside (C-terminal) position, the resolution values for BNP with poly L-SUGV, poly (L,L) SUAV, poly (L,L) SUVV, and poly (L,L) SULV are 1.87, 0.54, 1.92, and 6.94, respectively. As can be seen, the resolution of BNP is approximately 1.9 for both poly L-SUGV and poly (L,L) SUVV. When leucine is held constant in the outside (C-terminal) position, the resolution values for BNP with poly L-SUGL, poly (L,L) SUAL, poly (L,L) SUVL, and poly (L,L) SULL are 4.64, 1.18, 0.00, and 4.74, respectively. Interestingly, of those four surfactants, poly L-SUGL and poly (L,L) SULL separated BNP the best with approximately the same resolution values. The contribution of steric factors on the chiral separation of BNP can also be examined when the inside (N-terminal) amino acid is achiral and the size of the outside (C-terminal) amino acid is varied. The resolution of BNP goes from zero for poly L-SUGA 1746

Analytical Chemistry, Vol. 72, No. 8, April 15, 2000

to 1.9 and 4.6 for poly L-SUGV and poly L-SUGL, respectively. The dependence on the size of the R-group is not as clear with BOH and BNA in comparing the SCCDSs. The resolution of BOH is 5.9, 4.0, and 6.3 for poly L-SUAG, poly L-SUVG, and poly L-SULG, respectively. The resolution drops for poly L-SUVG compared to the other two surfactants when glycine is held constant in the second position. When glycine is held constant in the second position, the resolution of BNA is 3.0, 3.4, and 4.4 for poly L-SUAG, poly L-SUVG, and poly L-SULG, respectively. The reason for these trends are unclear. This behavior will be investigated further with other dipeptide surfactants. Another interesting trend to note is the differences in enantiomeric resolution of the binaphthyl derivatives when the size of the amino acid in the first (N-terminal) position is held constant and the size of the amino acid in the second position is varied, compared the case when the amino acid in the second position is held constant and the size of the amino acid in the first position is varied. Relatively small changes in resolution are observed with increasing size of amino acid in the second position, with the exception of BOH when alanine is held constant in the first position. This is in contrast to the relatively large changes observed in resolution when the amino acid in the second position is held constant and the size of the amino acid in the first position is increased. It can be inferred from these data that the enantiomeric separations of these compounds are more affected by changes in the inside (N-terminal) amino acid than the size of the amino acid on the C-terminal position. Also, the fact that relatively small differences in resolution are observed when the size of the amino acid in the N-terminal position is held constant is further evidence of the preferential binding of these analytes to the inside amino acid. The effect of steric factors is further illustrated in a comparison of the single amino acid surfactants poly L-SUA, poly L-SUV, and poly L-SUL. The enantiomeric separation of BNP shows that only poly L-SUL, the largest of the three single amino acid surfactants, was capable of separating BNP. Similarly, an increase in enantiomeric resolution is observed for BNP with increasing bulkiness of the dipeptide surfactants with the same amino acid in both positions (i.e. poly (L,L) SUAA, poly (L,L) SUVV, and poly (L,L) SULL). The resolution values were 1.19, 1.92, and 4.74, respectively. It can be inferred from these results that the enantiomeric separation of BNP is favored by an increase in steric factors. The results for BOH and BNA are more ambiguous. The enantiomeric separation of BOH seems to be favored by the less sterically hindered surfactants poly L-SUA and poly (L,L) SUAA while BNA shows a general trend of improved resolution with an increase in steric factors. (2) Propranolol, Alprenolol, and Oxprenolol. The effects of steric factors on the enantiomeric separation of Prop, Alp, and Oxp are examined in this section. The enantiomeric separation of Oxp is clearly favored by an increase in steric factors on the outside (C-terminal) position of the polymeric dipeptide surfactant. The enantiomeric resolution of Oxp goes from zero for poly L-SUAG to 0.8, 1.1, and 1.2 for poly (L,L) SUAA, poly (L,L) SUAV, and poly(L,L) SUAL, respectively. The same general trend is observed when valine is held constant in the first (N-terminal) position and the size of the outside (C-terminal) amino acid is varied (i.e. poly (L,L) SUVA, poly (L,L) SUVV, and poly(L,L) SUVL).

The resolution values were 0.91, 1.10, and 1.27, respectively. Similar results were also observed for poly (L,L) SULA, poly (L,L) SULV, and poly (L,L) SULL (Rs ) 1.01, 1.11, and 1.75, respectively). As the size of the outside amino acid increases, so does the enantiomeric resolution of Oxp. The enantiomeric resolution of Alp and Prop also appears to be favored by an increase in size of the R-group on the C-terminal amino acid. The only exception observed to this general trend is when leucine is held constant in the inside (N-terminal) position. A small decrease in enantiomeric resolution of Alp and Prop is observed in going from poly (L,L) SULA to poly (L,L) SULV. No definite trends are observed when the outside (C-terminal) amino acid is held constant and the size of the inside (N-terminal) amino acid is varied. The fact that the size of the amino acid on the N-terminal position has little effect on resolution supports the earlier statement that the β-blockers bind preferentially to the C-terminal amino acid. Alprenolol and Prop also follow the same general trend as Oxp with a few minor exceptions. Interestingly, these exceptions only occur when valine is the N-terminal amino acid. An examination of steric effects on the enantiomeric separation of the β-blockers with the single amino acid surfactants poly L-SUA, poly L-SUV, and poly L-SUL, as well as the TCCDS with the same amino acids such as poly (L,L) SUAA, poly (L,L) SUVV, and poly (L,L) SULL, further suggests that the enantiomeric separation of Oxp is favored by an increase in steric factors. The enantiomeric resolution of Oxp for the single amino acid surfactants poly L-SUA, poly L-SUV, and poly L-SUL is approximately 1.0, 1.3, and 2.0, respectively. An increase in resolution is observed with an increase in size of R-group on the amino acid. Poly (L,L) SUAA separated the enantiomers of Oxp with a resolution of about 0.8 followed by a resolution of 1.1 and 1.8 for poly (L,L) SUVV and poly (L,L) SULL, respectively. While the enantiomeric separation of Alp and Prop appear to be favored by an increase in steric factors, the results are not as definitive. (3) TFAE, Glutethimide, and Aminoglutethimide. Examination of the data for the enantiomeric separation of TFAE strongly suggests the enantiomeric separation of TFAE is favored by a moderate increase in steric factors near the stereogenic center(s) of the polymeric surfactants examined in this report. As can be seen from the results in Table 2, poly L-SUAG showed no sign of enantiomeric separation, while poly (L,L) SUAA and poly (L,L) SUAV had resolutions of 0.7 and 1.5, respectively. This is followed by a slight drop in resolution with poly (L,L) SUAL. The resolution values for poly (L,L) SUVG, poly (L,L) SUVA, poly (L,L) SUVV, and poly (L,L) SUVL are 0.76, 1.43, 1.75, and 1.44, respectively. A drop in resolution is again observed when leucine is the C-terminal amino acid. In the case of poly (L,L) SULG, poly (L,L) SULA, poly (L,L) SULV, and poly (L,L) SULL, only poly (L,L) SULV was able to resolve the enantiomers of TFAE. A drop in resolution, when the more sterically hindered amino acid leucine is present, is also observed when the amino acid in the outside (C-terminal) position is held constant and the size of the inside amino acid is varied. An increase in resolution can be seen in going from the dipeptide surfactant with alanine in the N-terminal position compared to the dipeptide surfactant with valine in that same position. In other words, poly (L,L) SUVA separated the enantiomers of TFAE better than poly (L,L) SUAA

(Rs ) 1.43 and 0.65, respectively), as did poly (L,L) SUVV compared to poly (L,L) SUAV (Rs ) 1.75 and 1.48, respectively) and poly (L,L) SUVL compared to poly (L,L) SUAL (Rs ) 1.44 and 1.39, respectively). However, when leucine is the N-terminal amino acid, a drop in resolution occurs. Poly (L,L) SULA did not resolve the enantiomers of TFAE at all, nor did poly (L,L) SULL. A small drop in resolution also occurs with poly (L,L) SULV as compared to poly (L,L) SUVV. While some of the differences are not great, they are consistent. As with TFAE, the enantiomeric separation of Amino appears to be favored by a moderate increase in steric factors near the stereogenic center (at the preferential site of interaction). An increase in the resolution with increasing size of the C-terminal amino acid is observed with poly L-SUAG, poly (L,L) SUAA, and poly (L,L) SUAV (Rs ) 1.51, 2.19, and 3.05, respectively). However a very slight drop in resolution occurs with poly (L,L) SUAL (Rs ) 3.00). Similar results were also obtained with most of the other surfactants examined in this study. The only exception to this trend is with poly (L,L) SULA compared to poly (L,L) SUVA. A decrease in enantiomeric resolution with increasing size of the outside amino acid also occurs with Glut. The enantiomeric resolution of Glut is approximately the same with poly L-SUAG and poly (L,L) SUAA (Rs ) 18.0 and 1.84, respectively) followed by a decrease in resolution with poly (L,L) SUAV and poly (L,L) SUAL (Rs ) 1.28 and 0.61, respectively). While the same general trend is observed with poly L-SUVG, poly (L,L) SUVA, poly L-SUVV, and poly (L,L) SUVL (Rs ) 1.79, 2.07, 1.44, and 0.96, respectively), the enantiomeric separation of Glut appears to be relatively unaffected by the size of the amino acid in the C-terminal position with poly L-SULG, poly (L,L) SULA, poly L-SULV, and poly (L,L) SULL (Rs ) 1.48, 1.39, 1.40, and 0.75, respectively). The resolution of Glut also appears to be relatively unaffected by the size of the amino acid in the N-terminal position when valine and leucine are held constant in the outside (C-terminal) position. However, when alanine is held constant in the outside (C-terminal) position, an increase in resolution is observed with increasing size of the N-terminal amino acid for poly L-SUGA, poly (L,L) SUAA, and poly L-SUVA (Rs ) 0.77, 1.84, and 2.07, respectively), while a decrease in resolution is observed for poly (L,L) SULL (Rs ) 0.75). The overall effect of steric factors on the resolution of Glut appears to be that the enantiomeric separation of Glut is favored by a decrease in steric factors near the stereogenic center of the surfactant near the preferential site of interaction. (4) Temazepam, Oxazepam, and Lorazepam. The benzodiazepams examined in this study do not follow any definite trends with respect to steric factors. Since the data concerning the effect of steric factors on the enantiomeric separation of Temaz, Loraz, or Oxaz are inconsistent, no conclusions can be inferred about the effect of steric factors for any of the benzodiazepams examined in this report. CONCLUSIONS A summary of our results is given in Table 3. The results of these studies clearly suggest that BOH and BNA interact primarily with the inside (N-terminal) amino acid with little or no interaction with the outside C-terminal amino acid. In contrast, BNP binds closer to the bulk aqueous phase and interacts more with the outside (C-terminal) amino acid than BOH and BNA. Examination of dipeptide order indicates that the preferred configuration of Analytical Chemistry, Vol. 72, No. 8, April 15, 2000

1747

Table 3. Summary of Effectsa increase in steric factors analyte

N-terminal amino acid

BOH BNA BNP Oxp Alp Prop TFAE Amino Glut Temaz Oxaz Loraz

+ + ? NSD NSD NSD Mod Mod ? ? ? ?

C-terminal amino acid ?, NSD ?, + + + Mod Mod ?, + ? ?

posn of chiral center with SCCDS

preferential interaction site

preferential amino acid order

N-terminal amino acid

C-terminal amino acid

N-terminal amino acid

C-terminal amino acid

larger amino acid N-terminal position

larger amino acid C-terminal position

+ + + ? NSD NSD ? ? +

+ + + ? NSD NSD ? ? -

+ + + ? + + ? ? +

+ + + ? + + ? ? -

+ + + ? + + ? ? ?

+ + + ? ? ? ?

a Key: + ) positive interaction; - ) negative interaction; ? ) results vary; ?, - ) possible negative interaction; ?, + ) possible positive interaction; Mod ) separation favored by moderate interaction; NSD ) no significant difference.

the surfactant for the enantiomeric separation of the binaphthyl derivatives is when the larger of the two amino acids is in the first position. Investigation of steric effects suggests that the enantiomeric separations of BOH and BNA are more affected by changes on the inside (N-terminal) amino acid than the size of the amino acid on the C-terminal position. The resolution of BOH and BNA increases significantly as the size of the inside (N-terminal) amino acid is increased while BNP shows a somewhat different trend with increasing size of the outside (C-terminal) amino acid. It can be inferred from the results of these studies, as shown in Table 3, that the preferential site of interaction for the cationic β-blockers Alp, Oxp, and Prop with polymeric dipeptide surfactants is with the outside (C-terminal) amino acid. It can be further inferred that the enantiomeric separations of these analytes are favored by an increase in steric factors on the outside (C-terminal) amino acid. However, a decrease in enantiomeric resolution is observed with an increase in steric factors on the inside (Nterminal) amino acid. A summary of the analyte/dipeptide interactions for TFAE, Amino, Glut, Temaz, Oxaz, and Loraz is also shown in Table 3. It appears that the enantiomeric separation of TFAE and Amino is favored by a moderate increase in steric factors near the stereogenic center(s) of the polymeric surfactants. A general trend of increasing resolution was also observed for Temaz with an increase in steric factors at the C-terminal amino acid. However, the trends observed were not entirely consistent. We can also infer from the data that the enantiomeric separation of Glut decreases with an increase in steric factors at the C-terminal amino acid of TCCDSs. Due to inconsistent trends, no inference could be made about the effect of steric factors at the N-terminal amino acid for Glut and Temaz. Also due to inconsistent trends, the effects of steric factors at both the C- and N-terminal amino acids on the enantiomeric separation of Oxaz and Loraz are unclear. Similarly,

1748 Analytical Chemistry, Vol. 72, No. 8, April 15, 2000

no other relevant information about the analyte-dipeptide interactions is available for Temaz, Oxaz, and TFAE. The preferential binding sites of Amino, Glut, and Loraz were determined by the comparison of the enantiomeric resolution observed with the SCCDS varying the position of the chiral amino acid. Amino and Glut show no significant difference in resolution with position of the chiral amino acid. Therefore, Amino and Glut were determined to interact with both amino acids of the dipeptide. However, since the enantiomers of Loraz were separated better with SCCDSs with a chiral N-terminal amino acid compared to the SCCDSs with an achiral N-terminal amino acid, the preferential site of interaction was determined to be with the N-terminal amino acid. The preferred amino acid order for Amino and Glut was determined to be with the larger of the two amino acids in the N-terminal position. Not listed in the table is the effect of the number of chiral centers on enantiomeric separation. No significant difference was observed in the enantiomeric resolution of Glut and Loraz with TCCDSs compared to SCCDSs. In contrast, the enantiomeric separation of Amino was favored by TCCDSs. The results of these studies yield insight into the chiral interactions of polymeric dipeptide surfactants. These results should be useful in the design of more efficient surfactants, as well as aid in future development of models to predict which surfactants would have the best chance of separating particular types of chiral compounds. ACKNOWLEDGMENT This work was supported by a grant from the National Institute of Health (GM39844). I.M.W. also acknowledges the Philip W. West endowment for partial support of this research. Received for review August 4, 1999. Accepted January 12, 2000. AC9908804