- Email: [email protected]
Analysis of the isomeric composition of the proline peptide bond in an angiotensin-converting enzyme inhibitor using capillary electrophoresis
ANALYTICAL BIOCHEMISTRY Analytical Biochemistry 321 (2003) 167–173 www.elsevier.com/locate/yabio
Analysis of the isomeric composition of the proline peptide bond in an angiotensin-converting enzyme inhibitor using capillary electrophoresis Earle Stellwagena,* and Robin Ledgerb b
a Department of Biochemistry, University of Iowa, Iowa City, IA 52242, USA School of Pharmacy, University of Otago, Adams Building, Frederick Street, Dunedin, New Zealand
Received 14 March 2003
Abstract The cis and trans isomeric composition of a proline peptide bond can be determined by routine free-solution capillary electrophoresis measurements provided that one isomeric form is preferentially stabilized by a dissociable ionic group. This capability is illustrated using the angiotensin converting enzyme (ACE) inhibitor (S)-1-N-[1-(ethoxycarbonyl)-3-phenylpropyl]-L -ala-L -pro, which has the trade name enalapril. Electropherograms indicate that the two isomeric forms of enalapril can be separated with baseline resolution at 15 °C using capillary buffers having pH values in the dissociation ranges of the enalapril carboxyl group, pKcis and pKtrans of 2.6 and 3.1, and of the enalapril amine group, pKcis and pKtrans of 5.9 and 5.6. Such baseline resolution indicates that the isomeric composition does not change during analysis, facilitating measurement of the isomer composition of a sample prior to its injection into the capillary. Thus the effect of pH, ionic strength, or an aprotic solvent on the isomeric composition of enalapril can be measured under uniform analytical conditions. The trans isomer composition changes from 68% in the cationic form, pH <2, to 50% in the isoelectric form, pH 4.5, to 60% in the anionic form, pH >7. Addition of salt to the isoelectric form or addition of an aprotic solvent to any form prior to analysis increases the trans isomer composition. Similar analyses can be made using the alternative ACE inhibitors captopril and enalaprilat. Ó 2003 Elsevier Inc. All rights reserved. Keywords: Proline peptide isomerization; cis/trans Isomeric composition; Capillary electrophoresis; Enalapril
All peptide bonds populate two isomeric forms, termed the cis and trans isomers, which are in dynamic equilibrium. Steric considerations dictate that the peptide bonds generated by most amino acids are overwhelmingly in the trans isomer at equilibrium. However, peptide bonds whose amide group is contributed by proline have a significant population of cis isomers at equilibrium, due to the unique side chain structure of this amino acid [1]. Accordingly, proteins, peptides, and drugs containing such proline peptide bonds will significantly populate both isomers at equilibrium [2–4]. Since the interconversion between the isomers is relatively slow on the biological time scale [2,4,5], the efficacy of a drug containing proline peptide bonds may * Corresponding author. Fax: 1-319-335-7950. E-mail address: [email protected] (E. Stellwagen).
0003-2697/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0003-2697(03)00437-8
well be dependent on the equilibrium concentration of a particular isomer. NMR measurements have revealed that the cis and trans isomers of model peptides and drugs containing a C-terminal proline moiety have different net charges in limited pH ranges [3,4,6,7]. These different net charges may result from preferential stabilization of one of the isomers by noncovalent intramolecular interactions involving an ionizable group such as a carboxylate or an amine [3,4]. The difference in the net charge of these isomers should facilitate their resolution by capillary electrophoresis measurements, as has been illustrated using model dipeptides containing C-terminal proline residues [8–10]. However, the rapid exchange times exhibited by these model dipeptides required hardware modification of commercial capillary electrophoresis instruments to
168
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173
Fig. 1. Isomeric structures of enalapril.
adequately resolve their cis and trans isomers. By contrast, we find that the exchange times for the cis and trans isomers of drugs containing C-terminal proline residues are often sufficiently slow to utilize commercial instrumentation without need for hardware modification or for addition of components to the running buffer. We illustrate this capability using the angiotensinconverting enzyme (ACE)1 inhibitor enalapril, (S)-1-N[1-(ethoxycarbonyl)-3-phenylpropyl]-L -ala-L -pro. The structures of the isomeric forms of enalapril are shown in Fig. 1. It should be noted that each isomeric form can exist in three different charge states in aqueous solutions, a cationic form having a net charge of +1, an isoelectric form having a net charge of 0, and an anionic form having a net charge of )1.
Experimental
terization System. The capillary mounted in a thermostatted cassette was rinsed for at least 1 min with capillary buffer at 25 psi prior to injection of a sample. Each sample of enalapril containing about 0.2 mg/ml, about 0.4 mM, was pressure injected for 3 s at 25 psi. Electrophoresis was done at a constant voltage, E, selected not to exceed a current of 35 lA when using the coated capillary. Absorbance was constantly monitored at 214 nm. All measurements were obtained at a constant temperature of 15.0 °C unless noted otherwise. Electrophoretic mobility, l, having the units cm2 (Vs)1 , was calculated from the observed peak migration time, t, using the equation l ¼ LT LD =Et. Mobility values obtained using the uncoated capillary in the acidic pH range were corrected for electroosmotic flow using histamine as an internal reference compound. Peak areas were obtained using the integration routines encoded in the instrumental software. Overlapping peaks were analyzed by dropping a perpendicular from an observable valley.
Materials Energy minimization Enalapril maleate, captopril, N-acetyl-L -proline and valylproline were each purchased from Sigma, cacodylic acid was purchased from Fisher and malonic acid from Aldrich. Enalaprilat was a gift from Merck, Sharpe, and Dohme (Quimica de Puerto Rico Inc, Barcelonta, Puerto Rico). Stock solutions of enalapril contained approximately 2 mg/ml (4 mM as the maleate salt) in deionized water. Aliquots from this stock solution were diluted by an order of magnitude in various buffered solutions prior to injection into the capillary. A 75-lm-ID uncoated fused-silica capillary was purchased from Polymicro Technologies (Phoenix, AZ) and trimmed so that its total length, LT , was 39.7 cm and the length to the detection window, LD , was 29.4 cm. A 100-lm-ID polyacrylamide-coated fused-silica capillary was purchased from Beckman/Coulter (Fullerton, CA) and trimmed so that LT was 40.0 cm and LD was 29.8 cm. Each capillary was filled with deionized water, mounted in a thermostatted cassette, and stored at 4 °C. Capillary electrophoresis All electrophoretic measurements were obtained using a Beckman/Coulter P/ACE MDQ Molecular Charac1 Abbreviations used: ACE, angiotensin-converting MECC, micellar electrokinetic capillary chromatography.
enzyme;
The molecular mechanics of enalapril ions were optimized using HyperChem version 6.01 (Hypercube, Inc., Gainesville, FL). An MM+ method [11] used a Polak–Ribiere algorithm to a RMS gradient <0.1 kcal/ after first optimizing the intramolecular mol per A electrostatic or hydrogen-bonded ring. Little or no change was observed in the initial minima when models were perturbed and reoptimized. Energy minima are reported for molecules in the absence of solvent.
Results Capillary buffer pH Enalapril can be resolved into two electrophoretic components at 15 °C by free-solution electrophoresis using either a coated or an uncoated capillary and a variety of capillary buffers of different pH, as shown in Fig. 2A. The filled circles indicate the mobility of the faster component and the open circles the mobility of the slower component. Each electrophoretic component exhibits a biphasic transition from a common cationic form having a mobility of 1.0 104 cm2 (Vs)1 , to an isoelectric form having no mobility, to an anionic form having a mobility of 1:2 104 cm2 (Vs)1 .
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173
169
Table 1 Analysis of isomer titration curves Mobility
Carboxyl pK
Amine pK
Isomer
Faster Slower
3.11 2.55
5.63 5.93
trans cis
The data illustrated in Fig. 2 were analyzed under the assumption that the cationic form of enalapril has a mobility of 1.00 104 cm2 (Vs)1 , the neutral isoelectric form a mobility of zero, and the anionic form a mobility of 1:20 104 cm2 (Vs)1 . The apparent pK values have a fitting variability of 0.01 pH units.
correspond reasonably well with the published [12] pK values for enalapril, 2.97 and 5.35, noting that these latter values likely represent a nonequivalent mixture of the two isomers. Capillary buffer temperature
Fig. 2. Effect of capillary pH. Either an uncoated 75-mm capillary or a coated 100-mm capillary was filled with HCl, malonate, formate, or cacodylate buffers having the indicated pH values. An aliquot of a 0.4 mM solution of enalapril in water was pressure injected at zero time. All measurements were obtained at a capillary temperature of 15 °C. (A) Dependence of the mobility of each isomer on the pH of the capillary buffer. The filled circles denote the faster isomeric form and the open circles the slower form. The line connects the mobility values for the predominant isomeric form. (B) pH dependence of the difference in mobility of the two isomeric forms, where the difference in the mobility is the mobility of the faster form minus the mobility of the slower form. The line was constructed using the values listed in Table 1.
Each component exhibits a distinctive transition between these forms. This can perhaps be appreciated more clearly in Fig. 2B, in which the measurements shown in Fig. 2A are presented as the difference in the mobility of the two components, namely the mobility of the faster component minus the mobility of the slower component. It can be seen that the maximal mobility difference between the two components occurs at pH 2.8 and at pH 5.8 at 15 °C. One interpretation of the mobility measurements shown in Fig. 2 is that each of the two dissociable groups in enalapril, the proline carboxyl and the secondary amine, have different pK values in the two isomers. The experimental mobility values are well fit by the apparent pK values listed in Table 1, as shown by the curvilinear line in Fig. 2B. These apparent pK values
Electropherograms of enalapril obtained using an uncoated capillary filled with 20 mM cacodylate buffer, pH 6.18, are shown in Fig. 3. An uncoated capillary was used for these measurements because of the fragility of the capillary coating at elevated temperatures. Enalapril is separated into its two components with baseline resolution when electrophoresis is performed at 15 °C. As the electrophoresis temperature is raised, the resolution of the two components is diminished until only a single component is observed at 60 °C. Similar electropherograms were observed using an uncoated capillary filled with 50 mM formate buffer, pH 2.97. Similar observations have been made using micellar electrokinetic capillary chromatography (MECC) to resolve enalapril [13]. The electropherograms shown in Fig. 3 are characteristic of a temperature-dependent increase in the exchange rate between two components in dynamic equilibrium, such as the cis/trans isomerization of a proline peptide bond. The increase in the exchange rate over the temperature range shown in Fig. 3 is consistent with the known large activation energy for proline peptide isomerization [2,5,8–10]. Accordingly, we assume that the two electrophoretic components of enalapril represent its cis and trans proline peptide isomers. Isomer assignment NMR measurements of a variety of molecules containing a C-terminal proline peptide bond have demonstrated that the trans isomer predominates at low pH [3,4]. Based on these observations, it is commonly assumed that the trans isomer predominates in the acidic form of other molecules containing a C-terminal proline peptide bond. We assume that this pertains to enalapril also. As noted above, electropherograms of enalapril observed in the pH regions of maximal difference mobility
170
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173
Fig. 3. Effect of capillary temperature. An uncoated 75-lm capillary was filled with 20 mM cacodylate buffer, pH 6.18. An aliquot of a 0.5 mM solution of enalapril in water was pressure injected at zero time. Absorbance was continually monitored at 214 nm. The illustrated electropherograms were obtained at capillary temperatures maintained at 15, 30, 45, and 60 °C, reading upward.
at 15 °C, Fig. 2, exhibit baseline resolution of the peaks corresponding to the cis and trans proline peptide isomers. This indicates that no significant isomeric interchange occurs during electrophoresis at 15 °C. Accordingly, electrophoretic measurements at 15 °C indicate the isomeric composition of samples prior to their injection into the capillary. A solution of enalapril in water was adjusted to pH 1.62 and equilibrated to generate the cationic form. Aliquots of this sample were injected into either a coated capillary containing 20 mM cacodylate buffer, pH 5.82, or an uncoated capillary containing 50 mM formate buffer, pH 3.08, each maintained at 15 °C. Analysis of the resultant electropherograms indicated that about 68 1% of the enalapril was in the peak having the faster mobility. We therefore conclude that the faster peak measured under these electrophoretic conditions represents the trans proline peptide isomer. If the trans isomer is preferentially stabilized at low pH, then the carboxyl pK of the trans isomer should be higher than that of the cis isomer. Inspection of the values shown in Table 1 indicate that this is case, namely the carboxyl pK of the faster, trans, isomer, is higher than that of the slower, cis, isomer. Thus, the measured carboxyl pK values are consistent with the given isomer assignment.
Fig. 4. Effect of sample pH. Aliquots of enalapril were equilibrated in 10–50 mM solutions of HCl, malonate, formate, or cacodylate buffers of the indicated pH values prior to injection. The open circles denote measurements obtained using an uncoated capillary filled with 50 mM formate buffer, pH 3.08, maintained at 15 °C. The filled circles denote measurements obtained using a coated capillary filled with 20 mM cacodylate buffer, pH 5.83, maintained at 15 °C.
irrespective of whether the analysis was done using an uncoated capillary filled with formate buffer, pH 3.08, or a coated capillary filled with cacodylate buffer, pH 5.83. The results shown in Fig. 4 indicate that the trans isomer predominates in the cationic form of enalapril, both isomers are essentially equally populated in the isoelectric form and the trans isomer again predominates, albeit not as strongly, in the anionic form. The increase in the fractional composition of the trans isomer accompanying the amine deprotonation suggests that the cis isomer is preferentially stabilized in the isoelectric form. Such a preferential stabilization would be expected to raise the amine pK of the cis isomer relative to that of the trans isomer. This appears to be the case, as shown in Table 1. In contrast to the measurements shown in Fig. 4, prior HPLC measurements at pH 7.0 [14] and prior MECC measurements at pH 8.5–9.1 [13] indicate that the cis isomer dominates the anionic form. This conclusion was based on the assumption that the cis isomer has the greater hydrophobic surface, an assumption that we would question.
Sample buffer pH Sample ionic strength The protocol used to assign the trans isomer can be utilized to determine the isomeric content of any enalapril sample prior to injection. The relative area under the faster peak will always represent the fractional content of the trans isomer. Fig. 4 illustrates the dependence of the percentage of trans isomer on the pH of the sample prior to injection. It should be noted that comparable results were obtained
As shown in Fig. 5, addition of NaCl to the isoelectric form of enalapril at pH 4.5 prior to electrophoresis leads to a modest increase in the fractional composition of the trans isomer. However, at NaCl concentrations above 144 mM, the sum of the areas of both isomers diminishes, as also shown in Fig. 5, suggesting that enalapril precipitates at higher ionic strength.
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173
171
Table 2 Effect of miscible solvents 90% (v/v) Solvent component
% trans isomer
Acetonitrile Acetone Dimethylacetamide Water Propanol
73 70 68 59 58
Analyses were performed as described in the legend to Fig. 6.
Fig. 5. Effect of sample ionic strength. Samples of enalapril were equilibrated with the indicated concentrations of NaCl in 20 mM acetate buffer, pH 4.48. Aliquots of these solutions were pressure injected into a coated capillary filled with 20 mM cacodylate buffer, pH 5.82, and electrophoresis was performed at 15 °C. The filled circles indicate the fractional composition of the trans isomer and the open circles indicate the total area of both isomer peaks.
Sample solvent polarity As shown in Fig. 6, addition of the aprotic solvent acetonitrile, CH3 CN, to an aqueous-buffered solution of the isoelectric form of enalapril increases the fractional composition of the trans isomer to a limiting value of about 75%. A similar limiting value is attained when excess acetonitrile is added to aqueous-buffered solutions of the cationic and anionic forms of enalapril. Aliquots of an aqueous solution of enalapril were mixed with a variety of neat miscible solvents to give a series of enalapril solutions which were 90% (v/v) in the added solvent. The trans isomeric compositions of these
Fig. 6. Effect of sample solvent polarity. An unbuffered solution of enalapril maleate, pH 2.84, squares, or pH 4.46, circles, was mixed and equilibrated with indicated volume fractions of acetonitrile. Aliquots of these solutions were pressure injected into a uncoated capillary filled with 40 mM formate buffer, pH 2.59, and electrophoresis was performed at 15 °C.
samples are listed in Table 2. Addition of a protic solvent, either water or propanol, gives a trans isomeric composition of about 60%. This value is identical to the trans composition measured in an alternative protic solvent, methanol, by NMR [15]. Addition of an aprotic solvent, either acetonitrile, acetone, or dimethylacetamide, gives a trans isomeric composition of about 70%. This value is very similar to the trans composition, 75%, measured in an alternate aprotic solvent, dimethyl sufoxide, by NMR [16]. Other analytes Preliminary measurements indicate that the isomeric forms of the alternative ACE inhibitors captopril and enalaprilat can also be baseline resolved by routine freesolution capillary electrophoresis measurements at 15 °C.
Discussion The dependence of the trans isomeric composition of enalapril on sample pH, shown in Fig. 4, can be interpreted from two perspectives. One perspective considers that this pH dependence simply reflects the energetics of the enalapril isomers in the absence of any intramolecular interactions. As shown in Table 3, molecular mechanical calculations predict that the trans isomer would be dominant in the cationic form, less dominant in the anionic form, and isoenergetic with the cis form in the isoelectric form in the absence of intramolecular interactions. These predictions correlate with the observed pH-dependent changes in the trans isomeric composition illustrated in Fig. 4. An alternative perspective considers that the observed pH dependence reflects changes in intramolecular interactions within enalapril. Such interactions might include (i) a hydrogen bond involving the proline carboxyl group as the hydrogen donor and the peptide carbonyl as the hydrogen acceptor, (ii) an electrostatic interaction involving the proline carboxylate and the protonated amine, and (iii) a hydrogen bond involving the amine as the hydrogen donor and either the proline carboxyl or the carboxylate as the hydrogen acceptor. Interaction (i) would be limited to the cationic form of
172
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173
Table 3 Energy minimization calculations of enalapril Enalapril form
Net charge
Cationic
+1
Isoelectric
0
Anionic
)1
Intramolecular bond
None Hydrogen (OHO) Hydrogen (NHO) None Hydrogen (NHO) Electrostatic None Hydrogen (NHO)
Energy (kcal/mol) trans Isomer
cis Isomer
18.5 19.9 31.2 20.4 a a 18.0 35.1
25.5 a 36.7 21.0 23.8 58.3 20.8 24.5
The larger the energy, the less stable the structure. The letter a indicates that the given isomeric structure minimizes to the other isomer.
enalapril, interaction (ii) would be limited to the isoelectric form, and interaction (iii) could exist in all three forms. The energy minimization calculations shown in Table 3 predict that interaction (i) stabilizes the trans isomer, interaction (ii) stabilizes the cis isomer, and interaction (iii) stabilizes the trans isomer in the cationic form but the cis isomer in the isoelectric and anionic forms. This analysis largely correlates with the observed pH dependent changes in the trans isomeric content illustrated in Fig. 4. Accordingly, our present measurements cannot persuasively distinguish between these two perspectives. One promising avenue to distinguish between these two perspectives involves covalent modification. The trans isomeric content of acetylproline is reduced from 80 to 50% following dissociation of its proline carboxyl group [3]. However, the trans isomeric content remains at 80% following methylation of the proline carboxyl [3]. Since methylated acetylproline cannot form a proline carboxyl/peptide carbonyl hydrogen bond, the trans form of unmethylated acetylproline cannot be stabilized by this intramolecular interaction. We are currently investigating the effect of a series of analogous chemical modifications on the isomeric composition of enalapril.
Conclusions We have demonstrated that routine free solution capillary electrophoresis can be employed to measure the cis/trans isomeric composition of a proline peptide bond. Such measurements require the preferential stabilization of one isomeric form by a dissociable ionic group such as a carboxyl or an amine. The effect of changes in solution conditions such as pH, ionic strength, or solvent polarity on the isomeric composition can be examined by pretreatment of the sample prior to analysis. Such measurements should be made at the highest feasible voltage and lowest feasible temperature to minimize isomer exchange during analysis.
Acknowledgments R.L. offers his thanks to the Department of Biochemistry, University of Iowa for provision of facilities during his Study and Research Leave from the University of Otago. The authors are indebted to Dr. Nancy Stellwagen for generous access to her capillary electrophoresis apparatus.
References [1] V. Madison, J. Schellman, Location of proline derivatives in conformational space. I. Conformational calculations; optical activity and NMR experiments, Biopolymers 9 (1970) 511–567. [2] J.F. Brandts, H.R. Halvorson, M. Brennan, Consideration of the possibility that the slow step in protein denaturation reactions is due to cis–trans isomerism of proline residues, Biochemistry 14 (1975) 4953–4963. [3] C. Grathwohl, K. Wuthrich, NMR studies of the molecular conformations in the liner oligopeptides H-(L -Ala)n-L -Pro-OH, Biopolymers 15 (1976) 2043–2057. [4] C.A. Evans, D.L. Rabenstein, Nuclear magnetic resonance studies of the acid–base chemistry of amino acids and peptides. II. Dependence of the acidity of the C-terminal carboxyl group on the conformation of the C-terminal peptide bond, J. Am. Chem. Soc. 96 (1974) 7312–7317. [5] C. Grathwohl, K. Wuthrich, NMR studies of the rates of proline cis–trans isomerization in oligopeptides, Biopolymers 20 (1981) 2623–2633. [6] D.L. Rabenstein, A.A. Isab, Conformational and acid–base equilibriums of captopril in aqueous solution, Anal. Chem. 54 (1982) 526–529. [7] S.V.S. Mariappan, D.L. Rabenstein, Kinetics and thermodynamics of cis–trans isomerization of captopril and related compounds, J. Org. Chem. 57 (1992) 6675–6678. [8] A.W. Moore Jr., J.W. Jorgenson, Resolution of cis and trans isomers of peptides containing proline using capillary zone electrophoresis, Anal. Chem. 67 (1995) 3464–3475. [9] S. Ma, F. Kalman, A. Kalman, F. Thunecke, C. Horvath, Capillary zone electrophoresis at subzero temperatures. I. Separation of the cis and trans conformers of small peptides, J. Chromatogr. A 716 (1995) 167–182. [10] A.S. Rathore, C. Horvath, Capillary zone electrophoresis of interconverting cis–trans conformers of peptidyl-proline dipeptides: estimation of the kinetic parameters, Electrophoresis 18 (1997) 2935–2943.
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173 [11] N.L. Allinger, Conformational analysis. 130. MM2. A hydrocarbon force field utilizing V1 and V2 torsional terms, J. Am. Chem. Soc. 99 (1977) 8127–8134. [12] D.P. Ip, G.S. Brenner, Enalapril maleate, Anal. Profiles Drug Subst. 16 (1987) 207–243. [13] X.-Z. Qin, D.P. Ip, E.W. Tsai, Determination and rotamer separation of enalapril maleate by capillary electrophoresis, J. Chromatogr. 626 (1992) 251–258. [14] A. Kocijan, R. Grahek, D. Kocjan, L. Zupancic-Kralj, Effect of column temperature on the behavior of some angiotensin
173
converting enzyme inhibitors during high-performance liquid chromatographic analysis, J. Chromatogr. B 755 (2001) 229–235. [15] Y. Sakamoto, Y. Sakamoto, I. Oonishi, T. Ohmoto, Conformational analysis of enalapril (MK-421) in solution by proton and carbon-13 NMR, J. Mol. Struct. 238 (1990) 325–334. [16] A. Saran, S. Srivastava, E. Coutinho, Conformation and macromolecular interaction of ACE inhibitors enalapril and lisinopril by NMR, Eur. J. Pharm. Sci. 1 (1994) 181–188.
Analysis of the isomeric composition of the proline peptide bond in an angiotensin-converting enzyme inhibitor using capillary electrophoresis Earle Stellwagena,* and Robin Ledgerb b
a Department of Biochemistry, University of Iowa, Iowa City, IA 52242, USA School of Pharmacy, University of Otago, Adams Building, Frederick Street, Dunedin, New Zealand
Received 14 March 2003
Abstract The cis and trans isomeric composition of a proline peptide bond can be determined by routine free-solution capillary electrophoresis measurements provided that one isomeric form is preferentially stabilized by a dissociable ionic group. This capability is illustrated using the angiotensin converting enzyme (ACE) inhibitor (S)-1-N-[1-(ethoxycarbonyl)-3-phenylpropyl]-L -ala-L -pro, which has the trade name enalapril. Electropherograms indicate that the two isomeric forms of enalapril can be separated with baseline resolution at 15 °C using capillary buffers having pH values in the dissociation ranges of the enalapril carboxyl group, pKcis and pKtrans of 2.6 and 3.1, and of the enalapril amine group, pKcis and pKtrans of 5.9 and 5.6. Such baseline resolution indicates that the isomeric composition does not change during analysis, facilitating measurement of the isomer composition of a sample prior to its injection into the capillary. Thus the effect of pH, ionic strength, or an aprotic solvent on the isomeric composition of enalapril can be measured under uniform analytical conditions. The trans isomer composition changes from 68% in the cationic form, pH <2, to 50% in the isoelectric form, pH 4.5, to 60% in the anionic form, pH >7. Addition of salt to the isoelectric form or addition of an aprotic solvent to any form prior to analysis increases the trans isomer composition. Similar analyses can be made using the alternative ACE inhibitors captopril and enalaprilat. Ó 2003 Elsevier Inc. All rights reserved. Keywords: Proline peptide isomerization; cis/trans Isomeric composition; Capillary electrophoresis; Enalapril
All peptide bonds populate two isomeric forms, termed the cis and trans isomers, which are in dynamic equilibrium. Steric considerations dictate that the peptide bonds generated by most amino acids are overwhelmingly in the trans isomer at equilibrium. However, peptide bonds whose amide group is contributed by proline have a significant population of cis isomers at equilibrium, due to the unique side chain structure of this amino acid [1]. Accordingly, proteins, peptides, and drugs containing such proline peptide bonds will significantly populate both isomers at equilibrium [2–4]. Since the interconversion between the isomers is relatively slow on the biological time scale [2,4,5], the efficacy of a drug containing proline peptide bonds may * Corresponding author. Fax: 1-319-335-7950. E-mail address: [email protected] (E. Stellwagen).
0003-2697/$ - see front matter Ó 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0003-2697(03)00437-8
well be dependent on the equilibrium concentration of a particular isomer. NMR measurements have revealed that the cis and trans isomers of model peptides and drugs containing a C-terminal proline moiety have different net charges in limited pH ranges [3,4,6,7]. These different net charges may result from preferential stabilization of one of the isomers by noncovalent intramolecular interactions involving an ionizable group such as a carboxylate or an amine [3,4]. The difference in the net charge of these isomers should facilitate their resolution by capillary electrophoresis measurements, as has been illustrated using model dipeptides containing C-terminal proline residues [8–10]. However, the rapid exchange times exhibited by these model dipeptides required hardware modification of commercial capillary electrophoresis instruments to
168
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173
Fig. 1. Isomeric structures of enalapril.
adequately resolve their cis and trans isomers. By contrast, we find that the exchange times for the cis and trans isomers of drugs containing C-terminal proline residues are often sufficiently slow to utilize commercial instrumentation without need for hardware modification or for addition of components to the running buffer. We illustrate this capability using the angiotensinconverting enzyme (ACE)1 inhibitor enalapril, (S)-1-N[1-(ethoxycarbonyl)-3-phenylpropyl]-L -ala-L -pro. The structures of the isomeric forms of enalapril are shown in Fig. 1. It should be noted that each isomeric form can exist in three different charge states in aqueous solutions, a cationic form having a net charge of +1, an isoelectric form having a net charge of 0, and an anionic form having a net charge of )1.
Experimental
terization System. The capillary mounted in a thermostatted cassette was rinsed for at least 1 min with capillary buffer at 25 psi prior to injection of a sample. Each sample of enalapril containing about 0.2 mg/ml, about 0.4 mM, was pressure injected for 3 s at 25 psi. Electrophoresis was done at a constant voltage, E, selected not to exceed a current of 35 lA when using the coated capillary. Absorbance was constantly monitored at 214 nm. All measurements were obtained at a constant temperature of 15.0 °C unless noted otherwise. Electrophoretic mobility, l, having the units cm2 (Vs)1 , was calculated from the observed peak migration time, t, using the equation l ¼ LT LD =Et. Mobility values obtained using the uncoated capillary in the acidic pH range were corrected for electroosmotic flow using histamine as an internal reference compound. Peak areas were obtained using the integration routines encoded in the instrumental software. Overlapping peaks were analyzed by dropping a perpendicular from an observable valley.
Materials Energy minimization Enalapril maleate, captopril, N-acetyl-L -proline and valylproline were each purchased from Sigma, cacodylic acid was purchased from Fisher and malonic acid from Aldrich. Enalaprilat was a gift from Merck, Sharpe, and Dohme (Quimica de Puerto Rico Inc, Barcelonta, Puerto Rico). Stock solutions of enalapril contained approximately 2 mg/ml (4 mM as the maleate salt) in deionized water. Aliquots from this stock solution were diluted by an order of magnitude in various buffered solutions prior to injection into the capillary. A 75-lm-ID uncoated fused-silica capillary was purchased from Polymicro Technologies (Phoenix, AZ) and trimmed so that its total length, LT , was 39.7 cm and the length to the detection window, LD , was 29.4 cm. A 100-lm-ID polyacrylamide-coated fused-silica capillary was purchased from Beckman/Coulter (Fullerton, CA) and trimmed so that LT was 40.0 cm and LD was 29.8 cm. Each capillary was filled with deionized water, mounted in a thermostatted cassette, and stored at 4 °C. Capillary electrophoresis All electrophoretic measurements were obtained using a Beckman/Coulter P/ACE MDQ Molecular Charac1 Abbreviations used: ACE, angiotensin-converting MECC, micellar electrokinetic capillary chromatography.
enzyme;
The molecular mechanics of enalapril ions were optimized using HyperChem version 6.01 (Hypercube, Inc., Gainesville, FL). An MM+ method [11] used a Polak–Ribiere algorithm to a RMS gradient <0.1 kcal/ after first optimizing the intramolecular mol per A electrostatic or hydrogen-bonded ring. Little or no change was observed in the initial minima when models were perturbed and reoptimized. Energy minima are reported for molecules in the absence of solvent.
Results Capillary buffer pH Enalapril can be resolved into two electrophoretic components at 15 °C by free-solution electrophoresis using either a coated or an uncoated capillary and a variety of capillary buffers of different pH, as shown in Fig. 2A. The filled circles indicate the mobility of the faster component and the open circles the mobility of the slower component. Each electrophoretic component exhibits a biphasic transition from a common cationic form having a mobility of 1.0 104 cm2 (Vs)1 , to an isoelectric form having no mobility, to an anionic form having a mobility of 1:2 104 cm2 (Vs)1 .
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173
169
Table 1 Analysis of isomer titration curves Mobility
Carboxyl pK
Amine pK
Isomer
Faster Slower
3.11 2.55
5.63 5.93
trans cis
The data illustrated in Fig. 2 were analyzed under the assumption that the cationic form of enalapril has a mobility of 1.00 104 cm2 (Vs)1 , the neutral isoelectric form a mobility of zero, and the anionic form a mobility of 1:20 104 cm2 (Vs)1 . The apparent pK values have a fitting variability of 0.01 pH units.
correspond reasonably well with the published [12] pK values for enalapril, 2.97 and 5.35, noting that these latter values likely represent a nonequivalent mixture of the two isomers. Capillary buffer temperature
Fig. 2. Effect of capillary pH. Either an uncoated 75-mm capillary or a coated 100-mm capillary was filled with HCl, malonate, formate, or cacodylate buffers having the indicated pH values. An aliquot of a 0.4 mM solution of enalapril in water was pressure injected at zero time. All measurements were obtained at a capillary temperature of 15 °C. (A) Dependence of the mobility of each isomer on the pH of the capillary buffer. The filled circles denote the faster isomeric form and the open circles the slower form. The line connects the mobility values for the predominant isomeric form. (B) pH dependence of the difference in mobility of the two isomeric forms, where the difference in the mobility is the mobility of the faster form minus the mobility of the slower form. The line was constructed using the values listed in Table 1.
Each component exhibits a distinctive transition between these forms. This can perhaps be appreciated more clearly in Fig. 2B, in which the measurements shown in Fig. 2A are presented as the difference in the mobility of the two components, namely the mobility of the faster component minus the mobility of the slower component. It can be seen that the maximal mobility difference between the two components occurs at pH 2.8 and at pH 5.8 at 15 °C. One interpretation of the mobility measurements shown in Fig. 2 is that each of the two dissociable groups in enalapril, the proline carboxyl and the secondary amine, have different pK values in the two isomers. The experimental mobility values are well fit by the apparent pK values listed in Table 1, as shown by the curvilinear line in Fig. 2B. These apparent pK values
Electropherograms of enalapril obtained using an uncoated capillary filled with 20 mM cacodylate buffer, pH 6.18, are shown in Fig. 3. An uncoated capillary was used for these measurements because of the fragility of the capillary coating at elevated temperatures. Enalapril is separated into its two components with baseline resolution when electrophoresis is performed at 15 °C. As the electrophoresis temperature is raised, the resolution of the two components is diminished until only a single component is observed at 60 °C. Similar electropherograms were observed using an uncoated capillary filled with 50 mM formate buffer, pH 2.97. Similar observations have been made using micellar electrokinetic capillary chromatography (MECC) to resolve enalapril [13]. The electropherograms shown in Fig. 3 are characteristic of a temperature-dependent increase in the exchange rate between two components in dynamic equilibrium, such as the cis/trans isomerization of a proline peptide bond. The increase in the exchange rate over the temperature range shown in Fig. 3 is consistent with the known large activation energy for proline peptide isomerization [2,5,8–10]. Accordingly, we assume that the two electrophoretic components of enalapril represent its cis and trans proline peptide isomers. Isomer assignment NMR measurements of a variety of molecules containing a C-terminal proline peptide bond have demonstrated that the trans isomer predominates at low pH [3,4]. Based on these observations, it is commonly assumed that the trans isomer predominates in the acidic form of other molecules containing a C-terminal proline peptide bond. We assume that this pertains to enalapril also. As noted above, electropherograms of enalapril observed in the pH regions of maximal difference mobility
170
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173
Fig. 3. Effect of capillary temperature. An uncoated 75-lm capillary was filled with 20 mM cacodylate buffer, pH 6.18. An aliquot of a 0.5 mM solution of enalapril in water was pressure injected at zero time. Absorbance was continually monitored at 214 nm. The illustrated electropherograms were obtained at capillary temperatures maintained at 15, 30, 45, and 60 °C, reading upward.
at 15 °C, Fig. 2, exhibit baseline resolution of the peaks corresponding to the cis and trans proline peptide isomers. This indicates that no significant isomeric interchange occurs during electrophoresis at 15 °C. Accordingly, electrophoretic measurements at 15 °C indicate the isomeric composition of samples prior to their injection into the capillary. A solution of enalapril in water was adjusted to pH 1.62 and equilibrated to generate the cationic form. Aliquots of this sample were injected into either a coated capillary containing 20 mM cacodylate buffer, pH 5.82, or an uncoated capillary containing 50 mM formate buffer, pH 3.08, each maintained at 15 °C. Analysis of the resultant electropherograms indicated that about 68 1% of the enalapril was in the peak having the faster mobility. We therefore conclude that the faster peak measured under these electrophoretic conditions represents the trans proline peptide isomer. If the trans isomer is preferentially stabilized at low pH, then the carboxyl pK of the trans isomer should be higher than that of the cis isomer. Inspection of the values shown in Table 1 indicate that this is case, namely the carboxyl pK of the faster, trans, isomer, is higher than that of the slower, cis, isomer. Thus, the measured carboxyl pK values are consistent with the given isomer assignment.
Fig. 4. Effect of sample pH. Aliquots of enalapril were equilibrated in 10–50 mM solutions of HCl, malonate, formate, or cacodylate buffers of the indicated pH values prior to injection. The open circles denote measurements obtained using an uncoated capillary filled with 50 mM formate buffer, pH 3.08, maintained at 15 °C. The filled circles denote measurements obtained using a coated capillary filled with 20 mM cacodylate buffer, pH 5.83, maintained at 15 °C.
irrespective of whether the analysis was done using an uncoated capillary filled with formate buffer, pH 3.08, or a coated capillary filled with cacodylate buffer, pH 5.83. The results shown in Fig. 4 indicate that the trans isomer predominates in the cationic form of enalapril, both isomers are essentially equally populated in the isoelectric form and the trans isomer again predominates, albeit not as strongly, in the anionic form. The increase in the fractional composition of the trans isomer accompanying the amine deprotonation suggests that the cis isomer is preferentially stabilized in the isoelectric form. Such a preferential stabilization would be expected to raise the amine pK of the cis isomer relative to that of the trans isomer. This appears to be the case, as shown in Table 1. In contrast to the measurements shown in Fig. 4, prior HPLC measurements at pH 7.0 [14] and prior MECC measurements at pH 8.5–9.1 [13] indicate that the cis isomer dominates the anionic form. This conclusion was based on the assumption that the cis isomer has the greater hydrophobic surface, an assumption that we would question.
Sample buffer pH Sample ionic strength The protocol used to assign the trans isomer can be utilized to determine the isomeric content of any enalapril sample prior to injection. The relative area under the faster peak will always represent the fractional content of the trans isomer. Fig. 4 illustrates the dependence of the percentage of trans isomer on the pH of the sample prior to injection. It should be noted that comparable results were obtained
As shown in Fig. 5, addition of NaCl to the isoelectric form of enalapril at pH 4.5 prior to electrophoresis leads to a modest increase in the fractional composition of the trans isomer. However, at NaCl concentrations above 144 mM, the sum of the areas of both isomers diminishes, as also shown in Fig. 5, suggesting that enalapril precipitates at higher ionic strength.
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173
171
Table 2 Effect of miscible solvents 90% (v/v) Solvent component
% trans isomer
Acetonitrile Acetone Dimethylacetamide Water Propanol
73 70 68 59 58
Analyses were performed as described in the legend to Fig. 6.
Fig. 5. Effect of sample ionic strength. Samples of enalapril were equilibrated with the indicated concentrations of NaCl in 20 mM acetate buffer, pH 4.48. Aliquots of these solutions were pressure injected into a coated capillary filled with 20 mM cacodylate buffer, pH 5.82, and electrophoresis was performed at 15 °C. The filled circles indicate the fractional composition of the trans isomer and the open circles indicate the total area of both isomer peaks.
Sample solvent polarity As shown in Fig. 6, addition of the aprotic solvent acetonitrile, CH3 CN, to an aqueous-buffered solution of the isoelectric form of enalapril increases the fractional composition of the trans isomer to a limiting value of about 75%. A similar limiting value is attained when excess acetonitrile is added to aqueous-buffered solutions of the cationic and anionic forms of enalapril. Aliquots of an aqueous solution of enalapril were mixed with a variety of neat miscible solvents to give a series of enalapril solutions which were 90% (v/v) in the added solvent. The trans isomeric compositions of these
Fig. 6. Effect of sample solvent polarity. An unbuffered solution of enalapril maleate, pH 2.84, squares, or pH 4.46, circles, was mixed and equilibrated with indicated volume fractions of acetonitrile. Aliquots of these solutions were pressure injected into a uncoated capillary filled with 40 mM formate buffer, pH 2.59, and electrophoresis was performed at 15 °C.
samples are listed in Table 2. Addition of a protic solvent, either water or propanol, gives a trans isomeric composition of about 60%. This value is identical to the trans composition measured in an alternative protic solvent, methanol, by NMR [15]. Addition of an aprotic solvent, either acetonitrile, acetone, or dimethylacetamide, gives a trans isomeric composition of about 70%. This value is very similar to the trans composition, 75%, measured in an alternate aprotic solvent, dimethyl sufoxide, by NMR [16]. Other analytes Preliminary measurements indicate that the isomeric forms of the alternative ACE inhibitors captopril and enalaprilat can also be baseline resolved by routine freesolution capillary electrophoresis measurements at 15 °C.
Discussion The dependence of the trans isomeric composition of enalapril on sample pH, shown in Fig. 4, can be interpreted from two perspectives. One perspective considers that this pH dependence simply reflects the energetics of the enalapril isomers in the absence of any intramolecular interactions. As shown in Table 3, molecular mechanical calculations predict that the trans isomer would be dominant in the cationic form, less dominant in the anionic form, and isoenergetic with the cis form in the isoelectric form in the absence of intramolecular interactions. These predictions correlate with the observed pH-dependent changes in the trans isomeric composition illustrated in Fig. 4. An alternative perspective considers that the observed pH dependence reflects changes in intramolecular interactions within enalapril. Such interactions might include (i) a hydrogen bond involving the proline carboxyl group as the hydrogen donor and the peptide carbonyl as the hydrogen acceptor, (ii) an electrostatic interaction involving the proline carboxylate and the protonated amine, and (iii) a hydrogen bond involving the amine as the hydrogen donor and either the proline carboxyl or the carboxylate as the hydrogen acceptor. Interaction (i) would be limited to the cationic form of
172
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173
Table 3 Energy minimization calculations of enalapril Enalapril form
Net charge
Cationic
+1
Isoelectric
0
Anionic
)1
Intramolecular bond
None Hydrogen (OHO) Hydrogen (NHO) None Hydrogen (NHO) Electrostatic None Hydrogen (NHO)
Energy (kcal/mol) trans Isomer
cis Isomer
18.5 19.9 31.2 20.4 a a 18.0 35.1
25.5 a 36.7 21.0 23.8 58.3 20.8 24.5
The larger the energy, the less stable the structure. The letter a indicates that the given isomeric structure minimizes to the other isomer.
enalapril, interaction (ii) would be limited to the isoelectric form, and interaction (iii) could exist in all three forms. The energy minimization calculations shown in Table 3 predict that interaction (i) stabilizes the trans isomer, interaction (ii) stabilizes the cis isomer, and interaction (iii) stabilizes the trans isomer in the cationic form but the cis isomer in the isoelectric and anionic forms. This analysis largely correlates with the observed pH dependent changes in the trans isomeric content illustrated in Fig. 4. Accordingly, our present measurements cannot persuasively distinguish between these two perspectives. One promising avenue to distinguish between these two perspectives involves covalent modification. The trans isomeric content of acetylproline is reduced from 80 to 50% following dissociation of its proline carboxyl group [3]. However, the trans isomeric content remains at 80% following methylation of the proline carboxyl [3]. Since methylated acetylproline cannot form a proline carboxyl/peptide carbonyl hydrogen bond, the trans form of unmethylated acetylproline cannot be stabilized by this intramolecular interaction. We are currently investigating the effect of a series of analogous chemical modifications on the isomeric composition of enalapril.
Conclusions We have demonstrated that routine free solution capillary electrophoresis can be employed to measure the cis/trans isomeric composition of a proline peptide bond. Such measurements require the preferential stabilization of one isomeric form by a dissociable ionic group such as a carboxyl or an amine. The effect of changes in solution conditions such as pH, ionic strength, or solvent polarity on the isomeric composition can be examined by pretreatment of the sample prior to analysis. Such measurements should be made at the highest feasible voltage and lowest feasible temperature to minimize isomer exchange during analysis.
Acknowledgments R.L. offers his thanks to the Department of Biochemistry, University of Iowa for provision of facilities during his Study and Research Leave from the University of Otago. The authors are indebted to Dr. Nancy Stellwagen for generous access to her capillary electrophoresis apparatus.
References [1] V. Madison, J. Schellman, Location of proline derivatives in conformational space. I. Conformational calculations; optical activity and NMR experiments, Biopolymers 9 (1970) 511–567. [2] J.F. Brandts, H.R. Halvorson, M. Brennan, Consideration of the possibility that the slow step in protein denaturation reactions is due to cis–trans isomerism of proline residues, Biochemistry 14 (1975) 4953–4963. [3] C. Grathwohl, K. Wuthrich, NMR studies of the molecular conformations in the liner oligopeptides H-(L -Ala)n-L -Pro-OH, Biopolymers 15 (1976) 2043–2057. [4] C.A. Evans, D.L. Rabenstein, Nuclear magnetic resonance studies of the acid–base chemistry of amino acids and peptides. II. Dependence of the acidity of the C-terminal carboxyl group on the conformation of the C-terminal peptide bond, J. Am. Chem. Soc. 96 (1974) 7312–7317. [5] C. Grathwohl, K. Wuthrich, NMR studies of the rates of proline cis–trans isomerization in oligopeptides, Biopolymers 20 (1981) 2623–2633. [6] D.L. Rabenstein, A.A. Isab, Conformational and acid–base equilibriums of captopril in aqueous solution, Anal. Chem. 54 (1982) 526–529. [7] S.V.S. Mariappan, D.L. Rabenstein, Kinetics and thermodynamics of cis–trans isomerization of captopril and related compounds, J. Org. Chem. 57 (1992) 6675–6678. [8] A.W. Moore Jr., J.W. Jorgenson, Resolution of cis and trans isomers of peptides containing proline using capillary zone electrophoresis, Anal. Chem. 67 (1995) 3464–3475. [9] S. Ma, F. Kalman, A. Kalman, F. Thunecke, C. Horvath, Capillary zone electrophoresis at subzero temperatures. I. Separation of the cis and trans conformers of small peptides, J. Chromatogr. A 716 (1995) 167–182. [10] A.S. Rathore, C. Horvath, Capillary zone electrophoresis of interconverting cis–trans conformers of peptidyl-proline dipeptides: estimation of the kinetic parameters, Electrophoresis 18 (1997) 2935–2943.
E. Stellwagen, R. Ledger / Analytical Biochemistry 321 (2003) 167–173 [11] N.L. Allinger, Conformational analysis. 130. MM2. A hydrocarbon force field utilizing V1 and V2 torsional terms, J. Am. Chem. Soc. 99 (1977) 8127–8134. [12] D.P. Ip, G.S. Brenner, Enalapril maleate, Anal. Profiles Drug Subst. 16 (1987) 207–243. [13] X.-Z. Qin, D.P. Ip, E.W. Tsai, Determination and rotamer separation of enalapril maleate by capillary electrophoresis, J. Chromatogr. 626 (1992) 251–258. [14] A. Kocijan, R. Grahek, D. Kocjan, L. Zupancic-Kralj, Effect of column temperature on the behavior of some angiotensin
173
converting enzyme inhibitors during high-performance liquid chromatographic analysis, J. Chromatogr. B 755 (2001) 229–235. [15] Y. Sakamoto, Y. Sakamoto, I. Oonishi, T. Ohmoto, Conformational analysis of enalapril (MK-421) in solution by proton and carbon-13 NMR, J. Mol. Struct. 238 (1990) 325–334. [16] A. Saran, S. Srivastava, E. Coutinho, Conformation and macromolecular interaction of ACE inhibitors enalapril and lisinopril by NMR, Eur. J. Pharm. Sci. 1 (1994) 181–188.