Anal. Chem. 1999, 71, 3992-3999
Characterization and Thermodynamic Studies of the Interactions of Two Chiral Polymeric Surfactants with Model Substances: Phenylthiohydantoin Amino Acids H. Hyacinthe Yarabe,† Shahab A. Shamsi,‡ and Isiah M. Warner*,†
Department of Chemistry, Louisiana State University, Baton Rouge, Louisiana 70803, and Department of Chemistry, Georgia State University, Atlanta, Georgia 30303
Analytical ultracentrifugation is used for determination of the molecular weights and the sedimentation coefficients of poly(sodium undecanoyl-L-valinate) (PSUV) and poly(sodium undecanoyl-L-threoninate) (PSUT) at different temperatures. Plots of absorbance as a function of radius indicates that both PSUV and PSUT are highly monodispersed. A method for evaluating the partial specific volumes using density measurements is presented. The partial specific volumes of PSUV are slightly higher than those of PSUT. In addition, the temperature dependence of the retention factor in electrokinetic chromatography was used to estimate the enthalpy, the entropy, and the Gibbs free energy of the surfactant/analyte complexes. Five phenylthiohydantoin-DL-amino acids were separated and each enantiomeric pair was completely resolved. Comparison of the thermodynamic values obtained with PSUV vs PSUT using a van’t Hoff relationship suggests that PSUT, with a less favorable free energy change (i.e., less negative ∆(∆G)), generates a more positive entropy change, hence slightly less chiral resolution. During the past decade, there has been considerable interest in the use of polymeric surfactants. This interest in micellar polymerization arises because polymeric surfactants have provided some distinct advantages such as stability, rigidity, and controllable size over nonpolymeric (conventional) micelles. Recently, both achiral1-6 and chiral7-12 polymeric surfactants have been used as pseudostationary phases for analytical separation techniques such as electrokinetic capillary chromatography (EKC). However, their †
Louisiana State University. Georgia State University. (1) Palmer, C. P.; McNair, H. M. J. Microcolumn Sep. 1992, 4, 509. (2) Palmer, C. P.; Khaled, M. Y.; McNair, H. M. J. High Resolut. Chromatogr. 1992, 15, 756. (3) Palmer, C. P.; Terabe, S. Kuromatogurafi 1995, 16, 98. (4) Palmer, C. P.; Terabe, S. J. Microcolumn Sep. 1996, 8, 115. (5) Shamsi, A. S.; Akbay, C.; Warner, I. M. Anal. Chem. 1998, 70, 3078-3083. (6) Tanaka, N.; Nakagawa, K.; Hosoya, K.; Palmer, C.; Kunugi, S. J. Chromatogr., A 1998, 802, 23. (7) Wang, J.; Warner, I. M. Anal. Chem. 1994, 66, 3773. (8) Shamsi, S. A.; Macossay, J.; Warner, I. M. Anal. Chem. 1997, 69, 2980. (9) Shamsi, S. A.; Warner. I. M. Electrophoresis 1997, 18, 853. (10) Agnew-Heard, K. A.; Sanchez Pena, M.; Shamsi, S. A.; Warner, I. M. Anal. Chem. 1997, 69, 958.
fundamental properties and aggregation behavior is poorly understood. In addition, there are no studies on the thermodynamics of the interaction of ligand/micelle complexes; information that is essential for rational design of micelle polymers for interaction with a specific solute. Our research efforts have been centered around the synthesis and characterization of both chiral and achiral polymeric surfactants for EKC. The analytical characterization, in particular the determination of molecular weight of these polymeric micelles, is difficult because they often have a poorly defined conformation in solution, including a capacity to trap and entrain surrounding solvent molecules. This results in high exclusion volumes and, hence, nonideal thermodynamic behavior.13 Gel permeation chromatography (GPC) has been used to characterize similar polymers in the past;14-16 however, the nonavailability of structurally similar standards for calibration negatively affects its reliability. The recent coupling of GPC with a light-scattering detector for size determination has partially solved this problem. However, for most macromolecules, sample clarification still remains a problem due to light-scattering interferences from dust particles. Analytical ultracentrifugation provides numerous advantages over GPC because it is a diffusional process based on rigorous thermodynamics as well as mass, energy, and momentum balances. Therefore, the technique is well suited for study of nonideal species such as the polymeric surfactants used in this study. Sedimentation equilibrium and sedimentation velocity are the two usual modes of analytical ultracentrifugation. In the thermodynamically ideal case, the following equation17 has been used in sedimentation equilibrium experiments to calculate the molecular weight of a polymer:
‡
3992 Analytical Chemistry, Vol. 71, No. 18, September 15, 1999
(11) Billiot, E.; Macossay, J.; Thibodeaux, S.; Shamsi, S. A.; Warner, I. M. Anal. Chem. 1998, 70, 1375-1381. (12) Billiot, E.; Haddadian, F.; Thibodeaux, S.; Shamsi, S. A.; Warner, I. M. Anal. Chem. In press. (13) Flory, P. J. Principles of Polymer Chemistry; Cornell University Press: Ithaca, NY, 1953. (14) Doabashi, A.; Hamada, M.; Dobashi, Y. Anal. Chem. 1995, 67, 3011-3017. (15) Palmer, C. P.; Terabe, S. Anal. Chem. 1997, 69, 2772. (16) Leydet, A.; Elhachemi, H.; Boyer, B.; Lamaty, G.; Roque, J. P.; Scols, D.; Snoeck, R.; Andrei, G.; Ikeda, S.; Neyts, J.; Witvroww, M.; Declerq, E. J. Med. Chem. 1996, 39, 1626. (17) Fujita, H. In Foundations of Ultracentrifugal Analysis; Elving, P. J., Winefordner, J. D., Eds.; Wiley and Sons: New York, 1975. 10.1021/ac990212i CCC: $18.00
© 1999 American Chemical Society Published on Web 08/17/1999
ln
Aa Mω2r(1 - FV)(r2b - r2a) ) Ab 2RT
(1)
where Aa and Ab, respectively, represent the absorbance at the two radii, ra (meniscus) and rb (any position in the cell); F is the solvent density, ω is the angular velocity, R is the gas constant, T is the temperature (in kelvin), and M is the molar mass of the solute. The sedimentation coefficient is determined in the sedimentation velocity mode. It is an important parameter to measure because it relates to a distribution of molecular weight.18,19 When sedimentation velocity processes are controlled by a high centrifugal force, diffusional processes become negligible. Hence, the solute particles move rapidly toward the bottom of the centrifugal cell. In addition, the centrifugal force on the surface is partially counterbalanced by the buoyant force of the displaced solvent. Thus, the net sedimentation behavior of molecules in a centrifugal field is described by the Svedberg equation,20 i.e.
s)
M(1 - VF) ν ) 2 Nf ωr
(2)
where ν is the rate of sedimentation, s is the sedimentation coefficient, r is the radial distance from the center of rotation, N is Avogadro’s number, and f is the frictional coefficient. Terabe et al.21 were the first to use the thermodynamic parameters of micellar solubilization in MEKC to establish a relationship between the molecular structure of achiral solutes and the resulting thermodynamic quantities. Recently, a similar approach was employed by Peterson and Foley22 to show that MEKC is a useful and dependable technique for studying the thermodynamics of chiral solute/chiral micelle interactions. Phenylthiohydantoin (PTH)-DL-amino acids are important amino acid derivatives for determining amino acid sequences of peptides and proteins. As a result, a number of researchers have studied the enantiomeric separation of these compounds. Different chromatographic techniques including HPLC23,24 and GC25 have been used. The method of MEKC for PTH-amino acid sequencing often suffers from poor peak shapes. However, the addition of urea and methanol has improved these peak shapes in a few applications.26-29 Mixed micelles have also been employed for separation of PTH-amino acids, although a long separation time was required.30,31 More recently, MEKC has been used to attempt (18) Lechner, M. D.; Kehrhahn, J. H.; Machte, W. Polym. Rep. 1993, 34, 2448. (19) Ikeda, S.; Kakiuchi, K. J. Colloid. Interface Sci. 1967, 23, 134. (20) Hansen, J. C.; Lebowitz, J.; Demeler, B. Biochemistry 1994, 33, 13155. (21) Terabe, S.; Katsura, T.; Okada, Y.; Ishihama, Y.; Otsuka, K. J. Microcolumn Sep. 1993, 5, 23. (22) Peterson, A. G.; Foley, J. P. J. Microcolumn Sep. 1996, 8, 427. (23) Tambe, S. A. J. Chromatogr., A 1996, 740, 284. (24) Maguire, J. H. J. Chromatogr. 1987, 387, 453. (25) Eto, S.; Noda, H. J. Chromatogr. 1992, 579, 253. (26) Kurosu, Y.; Murayama, K.; Shindo, N.; Shisa, Y.; Ishioka, N. J. Chromatogr., A 1996, 752, 279. (27) Otsuka, K.; Terabe, S. Electrophoresis 1990, 11, 982. (28) Otsuka, K.; Kawahara, J.; Tatekawa, K. J. Chromatogr. 1991, 569, 209. (29) Otsuka, K.; Kaoru, K.; Higashimori, M.; Terabe, S. J. Chromatogr., A 1994, 680, 317. (30) Otsuka, K.; Kawakami, H.; Tamaki, W.; Terabe, S. J. Chromatogr., A 1995, 716, 319. (31) Otsuka, K.; Terabe, S. J. Chromatogr. 1990, 515, 221.
Figure 1. Chemical structures of the repeating units of the chiral polymeric surfactants. Poly(sodium undecanoyl-L-valinate) (A) and poly(sodium undecanoyl-L-threoninate) (B).
to achieve complete separation of PTH-amino acids and applied to the area of protein sequencing.32,33 Polymeric surfactants may offer significant advantages for enantiomeric separation of PTHamino acids since they do not require the addition of copious amounts of urea and organic solvents to the buffer. The study reported here employed analytical ultracentrifugation and densitometry for characterization of the solution behavior of polymeric surfactants. In addition, measurements of thermodynamic parameters (i.e., enthalpy, entropy, and free energy exchange) of the polymeric surfactant/analyte complex are performed using chiral EKC in order to better understand the mechanism of micelle polymer/analyte interactions. Two polymeric surfactants, poly(sodium undecanoyl-L-valinate) (PSUV) and poly(sodium undecanoyl-L-threoninate) (PSUT) (Figure 1) are used as pseudostationary phases to evaluate the differences in thermodynamic quantities due to a change in the structural feature of the PTH-DL-amino acids (Figure 2). As shown in Figure 1, two main structural differences can be observed for PSUV and PSUT. The former polymeric surfactant contains only one chiral center while the latter contains two chiral centers. Furthermore, PSUT possesses an hydroxyl group on the side chain which is capable of hydrogen bonding. Both polymers are composed of surfactants that exist predominantly as monovalent ions at pH >6 due to a single ionizable carboxylic group. EXPERIMENTAL SECTION Materials. The analytes (()-PTH phenylalanine, R-aminocaprylic acid, valine, and tryptophan were purchased from Sigma (St. Louis, MO). The PTH-norvaline was obtained from Aldrich Chemical Co. (Milwaukee, WI). All analytes were used as received. The reagents H3BO3, Na2HPO4, and triethylamine, all purchased from Sigma, were of analytical reagent grade and used as received for preparation of background electrolytes solutions. Synthesis. The monomeric amino acid surfactants of sodium N-undecenoyl-L-valinate (SUV) and sodium N-undecenoyl-L-thre(32) Cardinael, P.; Ndzie, E.; Petit, S.; Coquerel, G.; Combret, Y.; Combret, J. C. J. High Resolut. Chromatogr. 1997, 20, 560. (33) Kurosu, Y.; Murayama, K.; Shindo, N.; Shisa, Y.; Satou, Y.; Ishioka, N. J. Chromatogr., A 1997, 771, 311.
Analytical Chemistry, Vol. 71, No. 18, September 15, 1999
3993
Figure 2. Chemical structures of the chiral PTH-amino acids employed in this study.
oninate (SUT) were synthesized according to a procedure developed in our laboratory and reported elsewhere.7 Polymerization of a 100 mM solution of both surfactants was achieved by 60Co γ-irradiation (70 krad/h), for 168 h (total dose, 3-4 Mrad). Proton NMR was used to follow the polymerization process. The disappearance of the olefinic signals at 5.8 and 5.0 ppm of both SUV and SUT indicated complete polymerization. The broadening of the upfield peaks was an additional confirmation of polymerization. Both polymers were dialyzed against bulk water using a regenerated cellulose membrane with a 2000 molecular weight cutoff. The dialyzed products were then lyophilized to obtain the final products. The polymers were found to be 99+% pure as determined by elemental analysis. Analytical Ultracentrifugation. These measurements were performed using an Optima XLA analytical ultracentrifuge from Beckman Instruments, Inc. (Palo Alto, CA). The instrument has a high-intensity xenon flash light source and a grating monochromator that continuously scans from 190 to 800 nm. The detection system was set to measure the absorbance at 230 nm. The flash lamp illuminated only a selected sample during scanning. A toroidally curved holographic diffraction grating was used to select the wavelength and to collimate the beam of light. Four-sector cells were used, and data were acquired every 10 µm in replicates of 10. Data were digitized and displayed as absorption as a function of radial distance. For sedimentation equilibrium measurements, sample volumes were 110 mL while the solvent volumes were 125 mL. For sedimentation velocity measurements, the sample and solute volumes were 400 and 425 mL, respectively. Data were collected at 25 °C and at speeds of 20 000, 25 000, and 30 000 rpm. The temperature of the rotor was controlled thermoelectrically to within (0.5 °C. All samples had a polymer concentration of 0.1 g/L. The absorbance vs the distance from the center of rotation to any position in the sample column was collected at 720-min intervals. Successive scans of the cell were compared graphically using the XL-A software to ensure that the samples reached equilibrium. Densitometer. A high-precision densitometer (model DMA58), purchased from Anton Paar USA (League City, TX), was used to perform density measurements. Air and water were used for calibration. The precision of the temperature-controlled system was better than (0.005 °C. 3994 Analytical Chemistry, Vol. 71, No. 18, September 15, 1999
Preparation of PTH-Amino Acid Standard. All stock solutions of PTH-amino acids standards (1 mg/mL) were prepared using a mixture of 90% (v/v) methanol and 10% (v/v) triply deionized water (U.S. Filter, Lowell, MA). These samples served as stock solutions. The model test solution, which contained a mixture of five PTH-amino acids with a final concentration of 0.2 mg/mL, were prepared from the respective stock solution in 50% (v/v) methanol and 50% (v/v) triply deionized water. Electrokinetic Chromatography Instrumentation. The EKC data were collected by use of a Beckman (Fullerton, CA) PACE (model 5510) CE instrument. Separations were performed with uncoated fused-silica capillaries of 50 µm i.d., which were purchased from Polymicro Technologies (Phoenix, AZ). The total capillary length was 47 cm. The effective capillary length (to the detector) was 40 cm. A new capillary was flushed with 1 M NaOH for 120 min followed by a 10-min rinse with triply deionized water before use. Thereafter, the capillary was routinely conditioned by successive flushing with 0.1 M NaOH (1 min), triply deionized water (0.5 min), and operating electrolyte (3 min) before sample injection. Samples were then introduced using pressure injection for a period of 2 s. The separation was initiated by applying 25 kV between the two capillary ends, which were immersed in vials containing the operating buffer. This procedure resulted in improved peak shapes, and the migration time reproducibility was 1.2% RSD for the analytes. The EKC conditions were as follows: 275 mM H3BO3, 20 mM Na2HPO4, 10 mM triethylamine (TEA), buffered at pH 7.05. The concentration of each polymeric surfactant was 1.7% (w/v), i.e., 53 mM equivalent monomer concentration. The UV absorption detection was performed at 254 nm. Thermodynamic Calculations in EKC. In EKC, the capacity factor k is defined as
k)
tre - taq taq(1 - (tre/tmc))
(3)
where tre, taq, and tmc are the retention times of one of the enantiomers, the bulk solution, and the micelle, respectively. However, it should be noted that the negatively charged polymeric surfactant migrates at a velocity much larger than the normal (nonpolymeric micelle) toward the anodic (injection end). As tmc approaches infinity, the term (1 - tre/tmc) in eq 3 is negligible and reduces to the following form:
k ) (tre - taq)/taq
(4)
The capacity factor (k) is related to the distribution coefficient, K, by
k ) K(Vmc/Vaq) ) Kβ
(5)
where Vmc and Vaq are the volumes of the micelles and the aqueous phase, respectively. The phase ratio (β) can be calculated by using21,22
β)
V(Csurf - cmc) 1 - V(Csurf - cmc)
(6)
Table 1. Molecular Weights, Sedimentation Coefficients, and Partial Specific Volumes sedimentation coeff (10-13 svg)
mol wt
partial specific vol (mL/g)
T (°C)
PSUV
PSUT
PSUV
PSUT
PSUV
PSUT
20 25 30 35 40
9984 ( 251 9987 ( 215 9723 ( 205 10230 ( 183 9304 ( 175
11252 ( 415 15049 ( 306 18320 ( 384 20403 ( 329 17554 ( 498
0.67 ( 0.03 0.82 ( 0.05 0.78 ( 0.04 0.89 ( 0.02 1.02 ( 0.02
1.03 ( 0.03 1.58 ( 0.04 1.69 ( 0.03 1.83 ( 0.05 1.54 ( 0.04
0.804 ( 0.005 0.806 ( 0.003 0.814 ( 0.002 0.813 ( 0.005 0.817 ( 0.005
0.772 ( 0.002 0.776 ( 0.002 0.779 ( 0.004 0.781 ( 0.002 0.786 ( 0.004
where Csurf is the concentration (mM) of the surfactant, V is the partial specific volume, and cmc is the critical micelle concentration. Polymeric surfactants do not have a cmc. Therefore, eq 6 may be approximated to the simplified eq 7:
β ) VCsurf/(1 - VCsurf)
(7)
The distribution coefficient at various temperatures can be described by the van’t Hoff equation:
ln K ) -
∆H° ∆S° + RT R
(8)
where ∆H° is the enthalpy change associated with micellar solubilization or the transfer of the solute from the aqueous phase to the micelle, ∆S° is the corresponding entropy change, R is the gas constant, and T is the absolute temperature. Combining eqs 5, 7, and 8 results in an experimental equation that can be used to calculate enthalpy and entropy changes due to micellar solubilization of analytes from the temperature dependence of the retention factor, i.e.
ln k ) -
∆H° ∆S° + + ln β RT R
Figure 3. Plot of 1/F (mL/g) as a function of W [(weight solvent)/ (weight solvent + weight solute)]. The solutes are poly(sodium undecanoyl-L-valinate) (A) and poly(sodium undecanoyl-L-threoninate) (B). The solvent is 100 mM NaCl.
(9)
where F is the solution density and W is the solvent weight fraction. The partial specific volumes of both PSUV and PSUT are obtained as the y-intercept of the (1/F) vs W plot (Figure 3). Table 1 includes the partial specific volumes of the two polymeric surfactants as a function of the temperature (20-40 °C). It is interesting to note that at each temperature the partial specific volumes of PSUV are always higher than those of PSUT. This
indicates that PSUT, which has an hydroxyl group in the polar head, has a more compact structure than PSUV. This is likely due to hydrogen bonding between the hydroxyl group of the PSUT polymeric micelle and the amide proton. Furthermore, the temperature dependence of the partial specific volumes is negligible. Determination of Molecular Weights and Sedimentation Coefficients by Analytical Ultracentrifugation. Analytical ultracentrifugation allows the determination of molecular weights and sedimentation coefficients of our polymers. An equilibrium concentration distribution of macromolecules can be obtained in the cell if the centrifugal force is small enough to allow the process of diffusion to oppose the process of sedimentation.17 For an ideal homogeneous sample, the concentration distribution is an exponential function of the buoyant mass of the molecules. According to eq 2, the molecular weight of each polymer can be determined from the partial specific volume and the slope of the line generated by a plot of ln A vs r. However, it should be noted that this relationship only holds for monodispersed samples. For polydispersed species, the plot deviates from linearity.35 Therefore, one must analyze the curvature of such a plot in order to obtain the average molecular weights at distance r.
(34) Durchschlag, H. In Thermodynamic data for biochemistry and biotechnology; Hinz, H. J., Eds.; Springer-Verlag: New York, 1986.
(35) Correia, J. J.; Shire, S.; Yphantis, D. A.; Schuster, T. M. Biochemistry 1985, 24, 3292-3297.
RESULTS AND DISCUSSION Determination of Partial Specific Volume. The exact volume of a particle is a difficult quantity to measure. Therefore, one often uses partial specific volume (V), which is defined as the increase in volume when 1 g of the dry solute is dissolved in a large volume of the solvent. The V values are determined by plotting the inverse of the density (1/F) of the solutions as a function of the weight fraction (W) of the polymers according to the following equation:34
(1/F) 1 ) V + W∂ F ∂W
(10)
Analytical Chemistry, Vol. 71, No. 18, September 15, 1999
3995
Figure 4. Plots and residuals of absorbance vs radius for poly(sodium undecanoyl-L-valinate) (A) and poly(sodium undecanoyl-Lthreoninate) (B) in aqueous 0.1 M NaCl solution. Wavelength λ, 230 nm; speed, 25 000 rpm; temperature, 25 °C. All data were truncated to avoid the two meniscuses and to improve the fits.
From the analytical ultracentrifugation measurements, molecular weights and sedimentation coefficients were calculated for our system (Table 1). Figure 4 illustrates the equilibrium distribution of PSUV and PSUT at 25 °C. The data points measured around the meniscus and the cell bottom were truncated to give a more representative fit. The residuals at the top of each plot indicate how well the data points correlate with the fitting function. A straight regression line for ln A vs r (data not shown) was observed. This occurs only for highly monodispersed polymers. As shown in Table 1, the molecular weights and sedimentation coefficients of PSUV remained almost constant from 20 to 35 °C. In contrast, the molecular weights and the sedimentation coefficient of PSUT increased roughly by a factor of 2 in the same temperature range. Perhaps, hydrogen bonding of the hydroxyl group of the threonine is the cause of this apparent dimerization. Both the molecular weights and the sedimentation coefficients of PSUV decrease between 35 and 40 °C, possibly because of a nonlinear deviation of the solution viscosity. This suggests that the polymers either undergo a conformational change or aggregate above 35 °C. Thermodynamic Studies of the Analyte Solubilizations. In EKC, chiral recognition requires interaction between a chiral analyte and a chiral pseudostationary phase, leading to the formation of transient adsorbates. For enantiomeric separation to occur, there must be an adequate difference in free energy between the transient adsorbates. Previously, we optimized the EKC conditions for enantioseparation by varying the concentration and type of polymeric surfactants, buffer pH, and background electrolyte type and composition (buffer and ionic strength).36 The optimum conditions for enantiomeric resolution of PTH-amino acids were found to be 53 mM for both PSUV and PSUT, 275 mM boric acid, 20 mM dibasic sodium phosphate, and 10 mM triethylamine at pH 7.2. Figure 5 compares the electropherograms for the separation of 10 enantiomers of 5 PTH-amino acids using PSUV and PSUT. It should be noted that the elution times are slightly shorter when the PSUT polymeric surfactant is used as the background electrolyte. At the pH of this study, both polymeric surfactants (PSUV and PSUT) possess similar charge. Therefore, (36) Shamsi, S. A.; Yarabe, H. H.; Warner, I. M. Manuscript in preparation.
3996 Analytical Chemistry, Vol. 71, No. 18, September 15, 1999
Figure 5. Electropherograms of a test mixtures of PTH-amino acids. Conditions: 275 mM sodium phosphate dibasic, 20 mM boric acid, 10 mM triethylamine (pH 7.0), 25 kV applied voltage; 50 µA current; 25 °C temperature. (A) 53 mM equivalent monomer concentration of PSUV is added to the buffer. (B) 53 mM equivalent monomer concentration of PSUT is added to the buffer. Peak identifications: (1) (()-PTH-valine, (2) (()-PTH-norvaline, (3) (()-PTH-phenylalanine, (4) (()-PTH-tryptophan, and (5) (()-PTH-R-aminocaprylic acid.
the differences in their electrophoretic mobility are due to the differences in their molar mass. PSUT, which has higher molar mass than PSUV, will be swept faster by the electroosmotic flow toward the detector end than PSUV. This explains why the elution time is shorter with PSUT compared to PSUV. One of the goals of this study is to evaluate ∆H° and ∆S° as the analyte moves from the aqueous phase into the polymeric surfactant phase. Therefore, van’t Hoff plots have been used to determine ∆H° and ∆S° (eq 9) and are shown in Figure 6. The enthalpy (∆H°) and the entropy (∆S°) of the analyte transfer from the aqueous phase to the micellar core of the polymerized surfactants were computed by using the slope (-∆H°/R) and the y-intercept (∆S°/R + ln β) of the ln k vs 1/T plots. It is clear that the enthalpies, which are directly proportional to the slopes, do not differ significantly among the five analytes. However, considerable differences in retention factors are observed. Table 2 provides a comparison of the thermodynamic results (∆H°, ∆S°, ∆G°) for the PTH-amino acids as they interact with PSUV and PSUT. As shown in Table 2, ∆H° for all five pairs of enantiomers are negative for both PSUV and PSUT. This suggests that the movement of the analytes from the aqueous phase into the micelle polymer phase is thermodynamically favored. In contrast, ∆S° was negative for most enantiomers (except for Dand L-PTH-R-aminocaprylic acid) when PSUV was used in the buffer. However, when PSUT was used in the buffer, ∆S° was positive for all analytes except for D- and L-PTH-tryptophan. The negative ∆S° values indicate a more organized system, possibly due to electrostatic interactions between the polar anionic analytes and the micelle headgroup. Therefore, the analytes, pushed away from the chiral center of the polymeric surfactant, exist in an “adsorbed” state near the hydrophobic micellar core. In contrast, when ∆S° values are positive, the analytes are in a “dissolved” state and freely migrate between the core of the polymeric surfactant and the polar chiral headgroup. In all cases, the polymeric surfactant (PSUT), with the less favorable enthalpy changes (less negative ∆H°), generates the more positive entropies (∆S°).
Figure 6. The van’t Hoff plots of PTH-amino acids for poly(sodium undecanoyl-L-valinate) (A) and poly(sodium undecanoyl-L-threoninate) (B) for the second enantiomers of the PTH-amino acid standards. The conditions and numbering were as shown for Figure 5. Table 2. Enthalpies and Entropies of Transfer (Micelle Solubilization) for the PTH-Amino Acidsa ∆H° (kJ/mol) enantiomers
PSUV
PUST
PTH-valine
solute
1
-10.56
-8.36
PTH-norvaline
2 1
-11.12 -11.05
-8.84 -8.75
PTH-phenylalanine
2 1
-11.81 -11.85
-9.42 -9.54
PTH-R-aminocaprylic acid
2 1
-12.38 -10.55
-9.94 -7.26
PTH-tryptophan
2 1
-10.78 -14.24
-7.40 -11.93
2
-14.84
-12.34
a
∆(∆H)° (kJ/mol) PSUV -0.56 -0.76 -0.53 -0.25 -0.63
PSUT
∆S° (J/mol K) PSUV
PSUT
-4.06
2.05
-5.36 -4.24
0.93 2.10
-6.16 -3.37
0.41 3.10
-4.61 7.88
2.17 17.75
7.37 -10.51
17.55 -3.94
-11.82
-4.71
-0.48 -0.66 -0.40 -0.14 -0.38
∆(∆S)° (J/mol K) PSUV
PSUT
-1.29
1.11
-1.93
1.69
-1.24
0.93
-0.51
-0.20
-1.32
-0.77
Conditions: 2% (w/v) polymers, 275 mM boric acid, 20 mM sodium phosphate dibasic, 10 mM triethylamine, pH 7.0, 25 kV.
It should be noted that PSUT is more polar than PSUV due to its hydroxyl group. Moreover, PSUT has higher molecular weight than PSUV. As a result, a stronger steric repulsion between the PSUT headgroup and the analytes minimizes the polar hydrogenbonding interactions. Therefore, the analytes reside primarily between the hydrophobic core and the polar head. Hence, less binding and slightly lower chiral resolution are observed when PSUT is used as a pseudostationary phase. Our chiral EKC data are consistent with the partial specific volumes of the micelle
polymers (see Table 1). Pseudostationary phases with higher partial specific volumes can be pictured as a more diffuse environment where the analytes can move freely. This favors larger differences in free energy between the pseudostationary phase and the analytes. This may be a contributing factor as to why PSUV, which possesses a larger partial specific volume than PSUT, provides better enantiomeric resolution (Table 3). In the temperature range of this study, the low-molecular-weight PSUV provides better separation than the high-molecular-weight PSUT. Analytical Chemistry, Vol. 71, No. 18, September 15, 1999
3997
Table 3. Comparison of Distribution Coefficients, Change in Gibbs Free Energy, Resolution, and Selectivities for the PTH-Amino Acidsa K (25 °C) solute
∆G° (kJ/mol)
enantiomers
PSUV
PSUT
PSUV
PSUT
PTH-valine
1
47
39
-9.31
-8.99
PTH-norvaline
2 1
50 56
42 47
-9.47 -9.75
-9.13 -9.40
PTH-phenylalanine
2 1
61 86
50 72
-9.91 -10.81
-9.54 -10.49
PTH-R-aminocaprylic acid
2 1
92 195
76 166
-10.96 -12.97
-10.61 -12.72
PTH-tryptophan
2 1
204 96
171 84
-13.07 -10.98
-12.80 -10.75
2
106
90
-11.20
-10.89
a
∆(∆G°) (kJ/mol)
resolution (25 °C)
R (25 °C)
PSUV
PSUT
PSUV
PSUT
PSUV
PSUT
-0.16
-0.14
3.37
2.46
1.071
1.065
-0.17
-0.15
3.47
3.17
1.075
1.070
-0.15
-0.11
2.84
2.37
1.067
1.053
-0.09
-0.08
2.44
1.90
1.040
1.030
-0.22
-0.16
4.32
3.72
1.100
1.076
Conditions: 2% (w/v) polymers, 275 mM boric acid, 20 mM sodium phosphate dibasic, 10 mM triethylamine, pH 7.0, 25 kV.
Figure 7. Enthalpy-entropy compensation plot for the enantiomers of the PTH-amino acids. Solutes are numbered as in Figure 5. Poly (sodium undecanoyl-L-valinate) (A) and poly(sodium undecanoyl-Lthreoninate) (B).
The measurement of ∆H° and ∆S° values using PTH-amino acids as model analytes facilitates determination of the retention mechanism. Plots of ∆H° as a function of ∆S° are linear (figure not shown), indicating a compensation behavior. Compensation behavior is observed whenever identical reactions of molecules differ in their respective enthalpies in such a manner that the plot of the enthalpy as a function of entropy is linear. 37-39 The plots of ∆(∆H°) as a function of ∆(∆S°) for the enantiomeric pairs of PTHamino acids (Figure 7) also support a compensation behavior indicating similar retention mechanisms for these enantiomeric solutes. Table 3 provides a summary of the distribution coefficients and the Gibbs free energies of the analytes at 25 °C, the resolution (37) Mukerjee, P.; Ko, J. J. Phys. Chem. 1992, 96, 6090. (38) )Melander, W. R.; Horvath, C. In High-Performance Liquid Chromatography: Advances and Perspectives; Horvath, C., Ed.; Academic Press: New York, 1980; Vol. 2. (39) Melander, W. R.; Campbell, D. E.; Horvath, C. J. Chromatogr. 1978, 158, 215.
3998 Analytical Chemistry, Vol. 71, No. 18, September 15, 1999
as well as selectivities (R) for the enantiomeric pairs. A threepoint interaction is possible for all analytes. It is possible that the aromatic rings of the solutes are included in the hydrophobic core of the micelle polymer, while hydrogen bonding could occur between the amino acid groups and the polar headgroups of the micelle polymer. Adequate selectivity will depend also on how deep the hydrophobic part of the solutes penetrates the core, permitting the polymeric surfactant and the solute to interact. A critical factor in selectivity was the degree of alkylation of the amino acid groups. PTH-R-aminocaprylic acid, which has the highest degree of alkylation, gave the lowest selectivity and highest equilibrium constants. The highest partition coefficient obtained for PTH-R-aminocaprylic acid is due to hydrophobicity of this amino acid. Obviously, PTH-R-aminocaprylic acid is the most hydrophobic of the solutes. By consequent, it has a higher affinity for the polymeric surfactant phases. It was interesting to note that the above analyte also has the lowest ∆(∆G). This explains the observed low selectivity. The hydrogen-bonding capability of the analyte plays an important role in selectivity as well. PTH-amino acid, tryptophan, which has an amide proton on its aromatic ring, has the highest ∆(∆G) and highest selectivity. For the polymer PSUV, the Gibbs free energies are greater than that of PSUT. Hence, better selectivity is obtained when PSUV is used in the buffer. As expected, the distribution coefficients were less for the less hydrophobic PTH-amino acid (valine) than for the more hydrophobic PTH-amino acid (R-aminocaprylic acid). CONCLUSIONS The partial specific volumes and the size of the polymeric surfactants were determined at various temperatures. The lowmolecular-weight PSUV provides better resolution than the highmolecular-weight PSUT in the temperature range of this study. Differences in enthalpy and entropy of solute transfer from the aqueous phase to the micellar phase were observed between each polymeric surfactant and the chiral PTH amino acid used as model analytes. Linear ∆(∆H°) vs ∆(∆S°) plots of the different pairs of enantiomers suggest that each follow similar retention mechanisms. Once the analytes enter into the polymeric phase, they are subjected to an electrostatic force between the core of the
micelle and its polar heads. The polymer PSUT is more polar and bulkier than the polymer PSUV. Thus, we speculate that this polymer complexes with the analytes in a manner that removes them from the chiral center (possibly the analytes solubilize somewhere between the hydrophobic core and the polar head) of the micelle polymer. This results in lower resolution when PSUT is used in the buffer. Future investigations in this laboratory will involve NMR and luminescence spectroscopies to investigate the interaction of PTH-amino acids with PSUV and PSUT.
ACKNOWLEDGMENT I.M.W. acknowledges support from the National Institute of Health (Grant GM-39844) and partial support from the Philip W. West endowment. The authors also thank Dr. M. McCarroll for stimulating and fruitful discussions. Received for review February 23, 1999. Accepted June 28, 1999. AC990212I
Analytical Chemistry, Vol. 71, No. 18, September 15, 1999
3999
