A Reversible Gel for Chiral Separations - Analytical Chemistry (ACS

The reversibility of the gel network offers potential advantages for chiral ... Gretchen Marie Peters , Luke P. Skala , Taylor N. Plank , Hyuntaek Oh ...
0 downloads 0 Views 146KB Size
Anal. Chem. 2004, 76, 4558-4563

A Reversible Gel for Chiral Separations Victoria A. Dowling, Joseph A. M. Charles, Emily Nwakpuda, and Linda B. McGown*

Department of Chemistry, P. M. Gross Chemical Laboratory, Box 90346, Duke University, Durham, North Carolina 27708

The use of a guanosine gel as a chiral selector in capillary electrophoresis is introduced. Guanosine gels are reversible organized media that are formed through the selfassociation of guanosine compounds. Their degree of organization and their physicochemical properties can be modulated through variations in guanosine monomer concentration, pH, temperature, and cation content. Baseline resolution of the D and L enantiomers of propranolol was achieved using a reversible biogel formed by 5′-guanosine monophosphate as the run buffer in capillary electrophoresis. Conditions were optimized to provide enantiomeric resolution of 2.1-2.3 in less than 5 min. The reversibility of the gel network offers potential advantages for chiral separations, including the possibility of using thermal or chemical dissociation of the gel network to remove the nucleoside monomers from the separated enantiomers, thereby eliminating the chiral selector as a source of physical contamination of the enantiomerically pure products and spectral background in UV absorbance detection. Numerous approaches to chiral separations in capillary electrophoresis have been described in recent years, primarily due to the demand for enantiomeric purity in many pharmaceutical products. The most commonly used chiral selectors are cyclodextrin compounds, although there is interest in a variety of other types of selectors as well, such as chiral surfactants and micelles, antibiotics, crown ethers, proteins, and polymeric phases.1 We describe here a new phase for chiral separations that employs a reversible biogel formed by 5′-guanosine monophosphate (5′-GMP). Guanosine gels (G-gels) are self-assembled networks of hydrogen-bonded guanine tetrads formed by guanosine nucleosides and nucleotides. Guanosine gelation, which was described as early as 1910,2 was long regarded primarily as a nuisance in DNA research and oligonucleotide synthesis. Over the years, however, interest arose in the unique aggregation behavior of G-gels.3-6 Among the DNA bases, gelation is unique to guanine, presumably because it alone has multiple hydrogen-bonding donor and * Corresponding author. Tel: (919) 660-1545. Fax: (919) 660-1605. E-mail: [email protected]. (1) Ward, T. J. Anal. Chem. 2002, 74, 2863. (2) Bang, I. Biochem. Z. 1910, 26, 293. (3) Gellert, M.; Lipsett, M. N.; Davies, D. R. Proc. Natl. Acad. Sci. U.S.A. 1962, 48, 2013. (4) Guschlbauer, W.; Chantot, J.-F.; Thiele, D. J. Biomol. Struct. Dyn. 1990, 8, 491. (5) Simonsson, T. Biol. Chem. 2001, 382, 621. (6) Chantot, J. F.; Sarocchi, M. T.; Guschlbauer, W. Biochimie 1971, 53, 347.

4558 Analytical Chemistry, Vol. 76, No. 15, August 1, 2004

acceptor sites.4 The building block of G-gels is the “guanosine tetrad”, which is formed through Hoogsteen (G:G) hydrogen bonding between each of four guanines and its neighbors. As the concentration of G-tetrads increases, the tetrads may dimerize to form octets, which can then stack upon themselves to form columnar structures7 that are stabilized by π-π stacking of the guanosines, as well as by the presence of certain cations.8 Alternatively, the monomeric nucleosides may self-assemble into a continuous helix to form a columnar structure,3 which gains added stability from the additional hydrogen-bonded network.3 As the concentration of the guanosine compound increases, the columnar structures can organize into more highly ordered, anisotropic liquid crystalline phases.9 The formation of G-gels is reversible, as evidenced by circular dichroism melting curves.3 Their degree of organization and their viscosity depend on monomer concentration, temperature, pH, and cation content,3,6-8 providing a variety of parameters that can be used to control their formation/disassembly and to reversibly modulate their properties. Interest in G-gels as biomaterials has focused primarily on the investigation of columnar G-wires and layered thin films of guanosine “nanoribbons” as molecular wires for nanoelectronics,10-16 although their potential use in a broader range of applications in materials science, sensors, and nanotechnology has been discussed.16 Due to their asymmetric nature, the columnar structures and the resulting liquid crystalline phases are expected to exhibit chiral selectivity. One such example has been reported, in which lipophilic derivatives of guanosine were shown to be enantioselective.17 In the present work, we demonstrate the use of a 5′GMP gel in capillary electrophoresis (CE) for the separation of the D and L enantiomers of the drug propranolol. This is, to our knowledge, the first use of G-gels for enantiomeric separations. (7) Spindler, L.; Olenik, I. D.; Copic, M.; Romih, R.; Cerar, J.; Skerjanc, J.; Mariani, P. Eur. Phys. J., E 2002, 7, 95. (8) Walmsley, J. A.; Burnett, J. F. Biochemistry 1999, 38, 14063. (9) Mariani, P.; Ciuchi, F.; Saturni, L. Biophys. J. 1998, 74, 430. (10) Calzolari, A.; Di Felice, R.; Molinari, E.; Garbesi, A. Appl. Phys. Lett. 2002, 80, 3331. (11) Marsh, T. C.; Henderson, E. Biochemistry 1994, 33, 10718. (12) Marsh, T. C.; Vesenka, J.; Henderson, E. Nucleic Acids Res. 1995, 23, 696. (13) Gottarelli, G.; Spada, G. P.; Garbesi, A. In Comprehensive Supramolecular Chemistry; Lehn, J. M., Ed.; Pergamon: New York, 1996; Vol. 9, p 483. (14) Samori, P.; Pieraccini, S.; Masiero, S.; Spada, G. P.; Gottarelli, G.; Rabe, J. P. Colloids Surf., B 2002, 23, 283. (15) Rinaldi, R.; Branca, E.; Cingolani, R.; Di Felice, R.; Calzolari, A.; Molinari, E.; Masiero, S.; Spada, G.; Gottarelli, G.; Garbesi, A. Ann. N. Y. Acad. Sci. 2002, 960, 184. (16) Davis, J. T. Angew. Chem., Int. Ed. 2004, 43, 668. (17) Andrisano, V. V.; Gottarelli, G.; Masiero, S.; Heijne, E. H.; Pieraccini, S.; Spada, G. P. Angew. Chem., Int. Ed. Engl. 1999, 38, 2386. 10.1021/ac0400010 CCC: $27.50

© 2004 American Chemical Society Published on Web 06/11/2004

EXPERIMENTAL SECTION Bare fused-silica capillaries (50-µm inner diameter) were purchased from Polymicro Technologies (Phoenix, AZ). All reagents were purchased from Sigma (St. Louis, MO). Gel solutions were prepared in 25 mM potassium or sodium phosphate buffer, pH 7.0, unless otherwise specified. The gel solutions were allowed to sit at room temperature overnight before use. Propranolol stock solutions contained 1.0-1.5 mg/mL DL-propranolol in 24.5 mM citrate-51.4 mM potassium phosphate buffer at pH 5.0. Samples containing 0.10-0.05 mg/mL propranolol were prepared by diluting the propranolol stock solution with the run buffer, which was 25 mM potassium or sodium phosphate buffer, pH 7.0, with KCl as specified, with or without 5′-GMP. Capillary electrophoresis experiments were performed on a Beckman Coulter P/ACE 5000 CE (Fullerton, CA). The instrument was set in forward polarity mode, with the cathode at the outlet. The total length of the capillaries was 37 cm, and the length to the detection window was 30 cm. Capillaries were contained in a temperature-controlled cartridge, with experimental temperatures ranging from 15 to 25 °C. Absorbance was detected at 214 nm. At the beginning of each day, the capillary was conditioned by rinsing at high pressure (20 psi) with 0.1 M NaOH for 10 min, deionized water for 5 min, and the run buffer for 10 min. Between runs, the capillary was high-pressure rinsed with 0.1 M NaOH for 3 min, deionized water for 2 min, and the run buffer for 3 min. Samples were introduced into the capillary using hydrodynamic injections at 0.5 psi for 1-5 s. Field strength ranged from 81 to 405 V/cm. The electrode reservoirs contained the run buffer. Circular dichroism spectra were collected using a Jasco J-710 spectropolarimeter. Scattered light intensity was measured using an SLM-Aminco model 8100 Series 2 spectrometer (Spectronics) with the excitation and emission wavelengths set to 400 nm. The scattered light intensity was measured at 90° to the incident beam.

RESULTS AND DISCUSSION Figure 1 (top) shows the CZE electropherogram of a racemate of DL-propranolol. As expected, there is no resolution of the enantiomers, which coelute at ∼9 min. Figure 1 (bottom) shows partial resolution of the enantiomers achieved under identical conditions except for the addition of 0.010 M 5′-GMP to the run buffer. Since G-gels absorb below 300 nm, they contribute a background to the CE signal across the electropherogram. The single peak at ∼12 min is a blank signal associated with changes in the gel phase upon injection of buffer, whether or not it contains propranolol. As shown in Figure 2, as the 5′-GMP concentration increases, the resolution of the enantiomers improves but the peaks become broader and the stability of the baseline decreases. The baseline may be affected not only by the increasing concentration of 5′GMP but also by increases in the organization of the gel, which could hinder flow of the gel run buffer through the capillary and exaggerate the perturbations due to sample injection. Based on the enantiomeric resolution and the quality of the baselines of the electropherograms, it was decided that 0.020 M 5′-GMP would be used in subsequent experiments. Decreasing pH below pH 7 also resulted in poor baseline stability (results not shown). Acidic pH favors gelation, which has an effect similar to increasing

Figure 1. Electropherograms of 0.10 mg/mL DL-propranolol. Run buffer: 25 mM potassium phosphate, pH 7.0, 0.020 M KCl, in the absence of GMP (top) and the presence of 0.010 M 5′-GMP (bottom); 1-s hydrodynamic injection, 135 V/cm, and 15 °C. Table 1. Resolution (Rs) at Varying Concentrations of Additional K+ (Added as KCl) in the Run Buffer Containing 0.020 M 5′-GMP in 25 mM Potassium or Sodium Phosphate Buffer at pH 7.0a [KCl]

Rs (K+)

Rs (Na+)

0 0.010 0.020 0.040

1.0 1.2 1.3 1.0

1.0 1.0 1.1

a DL-Propranolol (0.05 mg/mL) was introduced using a 1-s hydrodynamic injection. The field strength was 189 V/cm. The Rs values are the means of duplicate runs.

monomer concentration. The pH of the gel run buffer was therefore held at 7.0 to facilitate flow. The effects of temperature are shown in Figure 3. Resolution decreases when temperature is increased to 25 °C, which is generally the case for chiral separations, in which chiral selectivity is based on thermodynamically driven interactions. Additionally, there may be a decrease in the chiral structure of the gel as it begins to lose its organization at the higher temperature. Guanosine gels are stabilized by the presence of certain cations. For example, it has been shown that K+ can influence the organization of a 5′-GMP gel.8,9 Table 1 shows the resolution as a function of additional K+ in the run buffer for both potassium phosphate and sodium phosphate buffers. The K+ concentrations in Table 1 are in addition to the K+ or Na+ contributed from the Analytical Chemistry, Vol. 76, No. 15, August 1, 2004

4559

Figure 2. Electropherograms of 0.050 mg/mL DL-propranolol. Run buffer was (a) 0.010 M 5′-GMP, (b) 0.020 M 5′-GMP, and (c) 0.050 M 5′-GMP in 25 mM potassium phosphate, pH 7.0, 0.020 M KCl; 1-s hydrodynamic injection, 135 V/cm, and 20 °C.

buffer salt. In the sodium phosphate buffer, the resolution is relatively independent of K+ concentration, while in the potassium phosphate buffer, resolution passes through a maximum in the range of 0.01-0.02 KCl. These differences in behavior may be related to formation of different structures in the different buffers or to the smaller total K+ concentration in the sodium phosphate/ KCl solution. The cause of the decrease in resolution at 0.04 M KCl in the potassium buffer has not yet been determined and may be due to effects of the K+/GMP ratio on gel structure. Based on the results in Table 1, it was determined that a 25 mM potassium buffer, pH 7.0, with 0.02 M additional K+ would be optimal for a 0.02 M 5′-GMP gel for the separation of DL-propranolol. The effect of electric field strength is shown in Figure 4. The resolution increases with increasing field strength, most likely due to band broadening at lower field strengths as the sample plug spends increasing amounts of time in the capillary. 4560 Analytical Chemistry, Vol. 76, No. 15, August 1, 2004

The use of organic additives often enhances the resolution of enantiomers in CE.18-20 Figure 5 shows the separations achieved in the absence and presence of 5% (v/v) 2-propanol in the 5′-GMP run buffer. The addition of 2-propanol improves the quality of the baseline of the electropherogram. Table 2 shows the effect of 5% (v/v) 2-propanol on resolution. There is no effect on the resolution in the potassium phosphate buffer, but some improvement is seen in the sodium phosphate buffer. Ohm’s law plots are shown in Figure 6 for 5′-GMP run buffer with and without organic modifier. In both cases, the field strength used for the enantiomeric separations is within the linear range of the plot and the generated currents are reasonable. Joule (18) Armstrong, D. W.; Rundlett, K.; Reid, G. L. Anal. Chem. 1994, 66, 1690. (19) Liu, Z.; Zou, H.; Yem, N. J.; Zhang, Y. Electrophoresis 1999, 20, 2898. (20) Ward, T. J.; Dann, C.; Blaylock, A. J. Chromatogr., A 1995, 715, 337.

Figure 3. Electropherograms of 0.050 mg/mL DL-propranolol. Run buffer was 0.020 M 5′-GMP in 25 mM potassium phosphate, pH 7.0, 0.020 M KCl; 1-s hydrodynamic injection, 189 V/cm; (a) 15, (b) 20, and (c) 25 °C.

Table 2. Resolution (Rs) in the Absence and Presence of 2-Propanol in the Run Buffera % (v/v) 2-propanol

Rs (K+)

Rs (Na+)

0 5

1.3 1.3

1.0 1.2

a Other conditions as in Table 1 with 0.020 M KCl. The R values s are the means of duplicate runs.

heating is therefore not expected to be a significant consideration for the 5′-GMP run buffers. The mean resolution and mean migration time (taken at the midpoint between the two peaks) for nine consecutive separations of DL-propranolol on a single column were 2.2 ( 0.1 ((4.5%) and 4.77 ( 0.06 min ((1.3%), respectively. The fluctuations in resolu-

tion were randomly distributed, while there was a slight downward trend in migration time, particularly for the first two to three runs. The decrease in migration time may be due to modification of the fresh capillary over the course of the runs or to effects of minor changes in ambient temperature on the gel run buffer. The reversibility of the gel was confirmed using circular dichroism (CD) spectroscopy. The CD spectrum of a solution of the gel mobile phase was collected at 20 °C, then at 50 °C, and again at 20 °C. The experiment was also performed for the mobile phase with pH lowered to 5.0, at which there is a greater degree of gelation and the CD spectral features are more prominent. At both pHs, the CD spectral features were diminished upon heating to 50 °C and recovered upon cooling back to 20 °C, indicating reversibility. To get a qualitative idea of the region of gel organization in which the gel mobile phase falls, scattered light was measured Analytical Chemistry, Vol. 76, No. 15, August 1, 2004

4561

Figure 4. Electropherograms of 0.050 mg/mL DL-propranolol. Run buffer was 0.020 M 5′-GMP in 25 mM potassium phosphate, pH 7.0, 0.020 M KCl; 1-s hydrodynamic injection, 20 °C; (a) 135, (b) 189, and (c) 270 V/cm.

as a function of GMP monomer concentration. The results are shown in Figure 7. Since the dimensions of monomers and small G-tetrad aggregates are less than 1/20th of the wavelength of the scattered light (400 nm),21 they should behave like point scatterers. Therefore, the increase in scattered light intensity from 0 to 0.015 M GMP is attributed to increasing concentration of GMP monomers. The change in slope at 0.02 M GMP likely indicates the onset of aggregation of GMP to form G-tetrads. Since there is an excess of K+ in the solution, contributed from the KCl as well as the potassium phosphate buffer, the aggregates may be 2-plane G-quartets with the K+ centrally located between two G-tetrads. As concentration increases from 0.02 to 0.03 M GMP, the aggregates increase in size but not number, which would (21) Jurga-Nowak, H.; Banachowicz, E.; Dobek, A.; Patkowski, A. J. Phys. Chem. B 2004, 108, 2744

4562 Analytical Chemistry, Vol. 76, No. 15, August 1, 2004

explain the plateau. The resumption of increasing scattered light intensity with increasing concentration above 0.03 M GMP suggests further aggregation of the G-tetrads to form longer, columnar aggregates of G-quartets. The gel mobile phase contains 0.020 M GMP, which we tentatively identify as the approximate concentration at which aggregation of GMP to form G-quartet structures begins.21 Further work is needed to confirm the structure of the GMP mobile phase. CONCLUSION The 5′-GMP run buffer provides excellent resolution of the D and L enantiomers of propranolol in CE. The effects of monomer concentration, pH, temperature, and cation concentration on enantiomeric resolution are related to changes in gel structure and viscosity as well as more general electrophoretic effects. The most important feature that distinguishes G-gels from other

Figure 5. Electropherograms of 0.050 mg/mL DL-propranolol. Run buffer was 0.020 M 5′-GMP in 25 mM potassium phosphate, pH 7.0, 0.020 M KCl; 1-s hydrodynamic injection, 189 V/cm, 20 °C, (a) with 5% v/v 2-propanol and (b) without 2-propanol.

Figure 6. Ohm’s law plot. Run buffer was 0.020 M 5′-GMP in 25 mM potassium phosphate pH 7.0, 0.020 M KCl, with and without 5% 2-propanol; 20 °C.

reagents for chiral separations is the reversibility of guanosine gelation in response to temperature. In an appropriate platform, such as a microfluidic device, it might be possible to melt the reversible gel network following a chiral separation, releasing dissociated guanosine monomers that could then be diverted away from the sample stream prior to detection or collection of the enantiomers, or both. The result would be enantiomeric separations free of spectral interference and physical contamination by (22) Ludwig, M.; Kohler, F.; Belder, D. Electrophoresis 2003, 24, 3233. (23) Finn, M. G. Chirality 2002, 14, 534.

Figure 7. Scattered light intensity vs 5′-GMP concentration in 25 mM potassium phosphate buffer, pH 7.0, 0.020 M KCl. The excitation and emission monochromators were set at 400 nm. Scattered light was measured at 90° relative to the incident beam.

the chiral selector. The formation/dissociation cycle might be repeated indefinitely. Such capabilities would be a significant advance in pharmaceutical and agrochemical fields, which rely upon chiral separations for high-throughput screening of enantiomeric purity and of combinatorial libraries for discovery of asymmetric catalysts.22,23 Received for review January 2, 2004. Accepted April 29, 2004. AC0400010

Analytical Chemistry, Vol. 76, No. 15, August 1, 2004

4563