Capillary Electrochromatography of Proteins on an Anion-Exchanger

Capillary Electrochromatography of Proteins on an Anion-Exchanger Column .... Theory and practical understanding of the migration behavior of proteins...
0 downloads 0 Views 191KB Size
Download PDF
Anal. Chem. 2000, 72, 3022-3029

Capillary Electrochromatography of Proteins on an Anion-Exchanger Column Jian Zhang, Xian Huang, Shuhong Zhang, and Csaba Horva´th*

Department of Chemical Engineering, Yale University, New Haven, Connecticut 06520

Capillary electrochromatography (CEC) of proteins was carried out using 50-µm-i.d. fused-silica capillaries packed with 5-µm silica beads having strong anion-exchanger functions attached to hydrophilic spacers at the chromatographic surface. The siliceous microspheres and the capillary innerwall were treated first with a heterobifunctional silanizing agent and reacted subsequently with a vinyl monomer containing quaternary ammonium groups to form a “tentacular” anion exchanger. A mixture of bovine carbonic anhydrase, r-lactalbumin, soybean trypsin inhibitor, and ovalbumin was separated using CEC by isocratic elution in the codirectional mode with aqueous phosphate buffer, pH 7.0, containing sodium chloride. The retention mechanism of isocratic CEC for proteins on the anion-exchanger column was illustrated by the results of a study on the effect of salt concentration on the separation. The potential of CEC for protein separation with high resolution was also demonstrated by electrochromatograms of conalbumin and hemoglobin variants. The results shed light on the mechanism of protein separation by isocratic CEC, which is believed to be a combination of chromatographic retention by electrostatic interactions and electrophoretic migration. Assuming that the contributions of the two mechanisms to the overall migration velocity are additive, an electrochromatographic resolution equation was derived and compared to the resolution equation in HPLC to reveal the constituents responsible for the enhancement of resolution by CEC with respect to that in HPLC. The advantage of CEC was also examined by comparing peak capacities in CEC on an isocratic platform with peak capacities obtained with isocratic and gradient elution HPLC. Capillary electrochromatography (CEC) has been shown to have the potential of achieving high peak efficiency due to the favorable properties of electrosmotic flow1-12 in terms of band * To whom all correspondence should be addressed: (tel) (203) 432-4357; (fax) (203) 432-4360; (e-mail) [email protected]. (1) Pretorius, V.; Hopkins, B. J.; Schieke, J. D. J. Chromatogr. 1974, 99, 23. (2) Knox, J. H.; Grant, I. H. Chromatographia 1991, 32, 317. (3) Knox, J. H. J. Chromatogr. 1994, 680, 3. (4) Yamamoto, H.; Baumann, J.; Erni, F. J. Chromatogr. 1992, 593, 313. (5) Yan, C.; Schaufelberger, D.; Erni, F. J. Chromatogr. 1994, 670, 15. (6) Yan, C.; Dadoo, R.; Zhao, H.; Zare, R. N. Anal. Chem. 1995, 67, 2026. (7) Yan, C. U.S. Patent Appl. 08/142, 917, 1993. (8) Boughtflower, R. J.; Underwood, T.; Paterson, C. J. Chromatographia 1995, 40, 329. (9) Dittmann, M. M.; Rozing, G. P. J. Chromatogr., A 1996, 744, 63.

3022 Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

spreading. So far, in most applications of the technique, neutral and less polar compounds have been separated under conditions similar to those employed in reversed-phase HPLC. However, for CEC to become a widely used analytical technique it is also necessary to offer a means to separate charged biomacromolecules with a selectivity different from that obtained in HPLC or CZE. Nonetheless, this aspect of CEC has received scant attention, probably because it requires more novel separating systems especially tailored to take advantage of the peculiar features of CEC. Furthermore, an understanding of the separation mechanism greater than presently available is needed to exploit the full potential of CEC. The most widely used columns in CEC are packed with alkylsilica stationary phases. These sorbents have an abundance of silanol groups, which are usually negatively charged in contact with neutral or alkaline mobile phases. They are responsible for the generation of electrosmotic flow in high electric field. In reversed-phase chromatography, the retention is brought about by solvophobic interactions between the hydrocarbonaceous functions of the stationary phase and the sample components.13 Addition of acids and bases to the mobile phase has been proposed14,15 to improve the separation of basic compounds that may strongly interact with the negatively charged silanol groups. The employment of additives more than often imposes difficulties on the use of CEC with mass spectrometric detection. This should give another impetus to the development of novel stationary-/ mobile-phase systems for CEC. Ion-exchange chromatography has been used to separate peptides and proteins since the early 1950s.16,17 The use of inert polar supports such as cellulose, cross-linked dextrin, and agarose with aqueous eluents allowed the chromatography of complex biomolecules without untoward degradation. At the advent of protein HPLC, silica-based ion exchangers were widely employed. The silica surface was first coated with a water-soluble polymer such as poly(ethylene imine)18,19 or poly(aspartic acid),20 and then the layer was cross-linked in order to mask the surface silanol groups by forming an insoluble polymer layer with fixed charges (10) Choudhary, G.; Horva´th, Cs. J. Chromatogr., A 1997, 781, 161. (11) Poppe, H. J. Chromatogr., A 1997, 778, 3. (12) Rathore, A. S.; Horva´th, Cs. J. Chromatogr., A 1996, 743, 231. (13) Vailaya, A.; Horva´th, Cs. J. Chromatogr., A 1998, 829, 1. (14) Corradini, D.; Kalghatgi, K.; Horva´th, Cs. J. Chromatogr., A 1996, 728, 225. (15) Lurie, I. S.; Conver, T. S.; Ford, V. L. Anal. Chem. 1998, 70, 4563. (16) Sober, H. A.; Peterson, E. A. J. Am. Chem. Soc. 1954, 76, 1711. (17) Moore, S.; Stein, W. H. Adv. Protein Chem. 1956, 11, 191. (18) Vanecek, G.; Regnier, F. E. Anal. Biochem. 1982, 121, 156. (19) Alpert, A. J.; Regnier, F. E. J. Chromatogr. 1979, 185, 375. (20) Alpert, A. J. J. Chromatogr. 1983, 266, 23. 10.1021/ac000114t CCC: $19.00

© 2000 American Chemical Society Published on Web 06/17/2000

at the chromatographic surface. “Tentacle-like” ion exchangers were also introduced21 in order to enhance the number of available ionic and ionogenic groups on the stationary phase by grafting flexible functionalized polymer chains to the silica surface. In certain cases, the selectivity and efficiency of the tentacle-like ion exchangers were substantially increased with respect to those of conventional stationary phases. Ion-exchange chromatography of proteins is most commonly carried out by gradient elution with increasing salt concentration in the eluent. However, in the present investigation we have found, in agreement with other recent reports,22-24 that our CEC system facilitates protein separation on an isocratic platform. Some of the theoretical foundations of this possibility are discussed in an attempt to establish a mechanistic model for the separation of charged biomacromolecules. EXPERIMENTAL SECTION Materials. Fused-silica capillary tubing with polyimide outer coating, having 50-µm i.d. and 375-µm o.d. was purchased from Quadrex Scientific (New Haven, CT). 3-(Trimethoxysilyl)propyl methacrylate was purchased from Polysciences (Warrington, PA); 2,2-diphenyl-1-picrylhydrazyl hydrate (DPPH), and 3-(methacryloylamino)propyltrimethylammonium chloride (50 wt %) were from Aldrich (Milwaukee, WI); and monobasic, dibasic, and tribasic sodium phosphate, dimethylformamide (DMF; 99.9%), and hydrochloric acid (37.5%) were from J. T. Baker (Phillipsburg, NJ). Bovine carbonic anhydrase (BCA), R-lactalbumin (R-LAC), soybean trypsin inhibitor (STI), ovalbumin (OVA), conalbumin, and hemoglobin were purchased from Sigma (St. Louis, MO). Dimethyl sulfoxide (DMSO) was purchased from Burdick & Jackson (Muskegon, MI). Phosphoric acid (85%), sodium hydroxide (98.8%), and potassium persulfate were of analytical reagent grade from Mallinckrodt (Paris, KY). Reagent grade methanol and acetone were purchased from Fisher (Fair Lawn, NJ). Water was purified and deionized with a NanoPure system (Barnstead, Boston, MA). Preparation of the Stationary Phase. The chemical reactions involved in the functionalization process are shown schematically in Figure 1. One gram of Spherisorb S5-W bare silica beads, 5 µm, 80 Å (Phase Separations, Hauppauge, NY) was dispersed in 10 mL of 1 M HCl solution in a 20-mL glass vial which was capped and heated at 80 °C for 40 min. Subsequently, the vial was cooled to room temperature. The liquid was removed and the beads were washed with water, methanol, and acetone. After drying at 80 °C for 60 min, the beads were suspended in 9 mL of DMF, stirred for 5 min, and sonicated for 30 min. Then 1 mL of 3-(trimethoxysilyl)propyl methacrylate and 0.1% (w/v) DPPH were added; the suspension was purged with nitrogen for 15 min and heated for 6-8 h at 90 °C with the cap closed. The beads were subsequently separated by centrifugation from the silanizing solution and washed with DMF, acetone, methanol, and water. Then to the vial containing the silanized beads, 9 mL of water, 1 mL of 3-(methacryloylamino)propyltrimethylammonium chloride, and 0.1% (w/v) K2S2O8 were added. After the suspension was purged with nitrogen for 10 min, the vial was heated at 70 °C for 6-8 h. (21) Mu ¨ ller, W. J. Chromatogr. 1990, 510, 133. (22) Huang, X.; Zhang, J.; Horva´th, Cs. J. Chromatogr., A 1999, 858, 91. (23) Gusev, I.; Huang, X.; Horva´th, Cs. J. Chromatogr., A 1999, 855, 273. (24) Zhang, S.; Huang, X.; Zhang, J.; Horva´th, Cs. J. Chromatogr., A, in press.

Figure 1. Schematic illustration of the synthesis of the chromatographic surface. In the first step, DPPH is used to inhibit the free radical polymerization of the heterobifunctional silanizing agent.

After cooling the product to room temperature and washing the beads with water, they were ready for packing. Functionalization of the Capillary Inner Wall. The pretreatment and the silanization of the inner wall of the fused-silica capillary were described previously.25 The silanized capillary was filled with an aqueous solution containing 10% (v/v) 3-(methacryloylamino)propyltrimethylammonium chloride and 0.1% (w/v) K2S2O8. After sealing both ends, the capillary was heated in the oven at 90 °C for 6 h. Then it was cooled to room temperature, flushed with deionized water and methanol, and blown dry with nitrogen. Column Packing Procedure. A previously described6,10 packing procedure was modified as shown in Figure 2. Typically a capillary of 400 mm length was employed. The capillary was tapped in a vial containing 5-µm of dry silica particles to fill a ∼0.5 mm length into the capillary. An Archer Torch model B (Radio Shack, New Haven, CT) microtorch fueled with butane was used to make a retaining frit at the end of the column by sintering the silica particles at elevated temperature. The permeability and stability of the frit were tested at a pressure drop of ∼4000 psi (1 psi ) 6903 Pa) with a ConstaMetric III metering pump (Thermo Separation Products, San Jose, CA). A 5% (w/v) slurry of the functionalized silica beads was made in deionized water, and the slurry was sonicated for 30 min. Then, the capillary was connected to a cylindrical stainless steel reservoir (30 × 4.7 mm) which was (25) Huang, X.; Horva´th, Cs. J. Chromatogr., A 1997, 788.

Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

3023

Figure 2. Illustration of the steps involved in the fabrication of packed capillary columns for CEC.

filled with the slurry and connected to the metering pump. Deionized water was employed as the packing solvent, and the flow rate was set at 0.4 mL/min. The front of the packings gradually extended from the retaining frit toward the inlet. The length of the packed segment in the following experiments was 260 mm. Once the front of the packings reached the length, the flow rate of the packing solvent was turned down to zero in order to let the pressure gradually release. Thereafter the column was disconnected from the reservoir and connected to another metering pump. Deionized water was pumped through the column at the same flow rate until the pressure stabilized and kept for 4-5 h. Afterward another retaining frit was made by sintering the front of the packings with butane flame while keeping the packing solvent pumped through. A 4-5-mm-long detection window was also formed during sintering because of the burnt off area on the polyimide outer layer. The column was stabilized for another 2 h, and then the flow rate of the packing solvent was turned down to zero to gradually release the pressure in the column. Instrumentation. A model HP3DCE capillary electrophoresis unit (Hewlett-Packard, Wilmington, DE) with both the inlet and outlet vials pressurized with nitrogen up to 12 bar (1 bar ) 1.0 × 105 Pa) was used with a model P150 Hewlett-Packard personal computer. Windows 95 (Microsoft, Redmond, WA) and Chemstation V. 4.01 were installed to control the instrument functions and to process the data. The temperature of cartridge was set at 20 °C in all experiments, and UV adsorbance at 200 or 214 nm was recorded. Capillary Electrochromatography. In all experiments, the mobile phases were made by dissolving appropriate amounts of NaCl in 5 mM aqueous sodium phosphate buffer, pH 7. The practical concentrations of NaCl in the mobile phases ranged from 50 to 200 mM. The capillary columns were 340 mm long with a 260-mm-long packed segment. Sample solution containing 1.0 mg/ mL of each protein in deionized water was injected over 6 s at -8 kV. A solution of 0.1% (v/v) DMSO (the EOF marker) in deionized water was injected over 3 s at -3 kV. To equilibrate the column, 12 bar nitrogen pressure was applied to the column outlet to force the buffer solution through the column until a stable baseline was obtained. Then the column was flushed electrokinetically at -15 kV, with both ends under 12 bar nitrogen pressure in order to suppress bubble formation until the baseline stabilized. After the electrochromatographic run of a protein sample, the column was rinsed with 1 M NaCl for 3 min and equilibrated with the buffer solution before injection of the next sample. RESULTS AND DISCUSSION Description of the Stationary Phase. At present, silica-based stationary phases, which were developed originally for use in 3024

Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

reversed-phase HPLC, are employed most commonly in CEC. Their properties are known; they are readily available and have an abundance of silanol groups at the surface that are dissociated in contact with neutral and alkaline mobile phases so that upon applying high electric field to the column a cathodic electrosmotic flow is generated. Recent studies27 have demonstrated that CEC has significantly higher column efficiency than HPLC for neutral compounds by virtue of employing electrosmotic flow instead of viscous flow. It was also reported28 recently that, by using an amine-modulated buffer at acidic pH, peptides can be separated on alkylated silica stationary phases by CEC. However, silanophilic interactions may impede the separation of proteins on siliceous stationary phases at neutral or basic pH of the mobile phase that is necessary for obtaining sufficiently high EOF in the column. To circumvent such difficulties, a new type of stationary phase was developed in our laboratory.22-24 The chromatographic surface is functionalized with alkyl chains carrying quaternary ammonium groups that are positively charged over a wide pH range. In chromatography of proteins, the use of mild conditions is preferred in order to minimize denaturation. Therefore, the use of eluents having high organic strength and/or extreme pH has to be avoided. In the present study, a siliceous ion exchanger was chosen because of the rigidity of the silica support to withstand high packing pressures, the simplicity in column fabrication, and the availability of established surface modification techniques.22-26 Silica- or polymer-based ion exchangers for protein separation often have a hydrophilic stratum over the surface of the support. Figure 1 shows the process of functionalization of the siliceous support and the formation of the chromatographic surface of our strong anion exchanger. It is noted that in this particular case the functional groups serve as an umbrella over the underlying silica base. The positively charged quaternary ammonium groups are attached to the flexible oligomeric chains. Thus, this stationary phase is similar to the tentacle-like ion exchangers developed a decade ago21 for protein separation by HPLC using wide-pore silica support. Since the mean pore diameter of our silica support was 80 Å, most of the small pores of the support were plugged during the process of functionalization and became inaccessible to proteins. Thus, we believe that it behaved as a pellicular stationary phase.29 Generation of EOF. Upon applying high electric field, the positively charged surface of our packed column generated anodic EOF. The electrosmotic mobility was measured with DMSO as the neutral marker. The pH dependence of the electrosmotic mobility served as a diagnostic tool for the column as illustrated in Figure 3. It shows that with successfully prepared column packings the electrosmotic mobility is about the same in the pH range from 2.5 to 7.0 at (1.61 ( 0.02) × 10-8 m2 V-1 s-1 using 5 mM phosphate buffer containing 200 mM NaCl. The constancy of EOF in CEC is due to an effective masking of residual silanols at the chromatographic surface, a uniform surface coverage, and the stability of the stationary phase proper. When the magnitude (26) Ericson, C.; Liao, J.-L.; Nakazato, K.; Hjerte´n, S. J. Chromatogr., A 1997, 767, 33. (27) Wen, E.; Asiaie, R.; Horva´th, Cs. J. Chromatogr., A 1999, 855, 349. (28) Walhagen, K.; Unger, K. K.; Olsson, A. M.; Hearn, M. T. W. J. Chromatogr., A 1999, 853, 263. (29) Horva´th, Cs.; Preiss, B. A.; Lipsky, S. R. Anal. Chem. 1967, 39, 1422.

10 8

Table 1. Isoelectric Points and Molecular Massess of the Proteins Investigated

Figure 3. EOF mobility against pH. Plots illustrate the difference between an acceptable and a defective column.

Figure 4. Effect of salt concentration on migration velocity of DMSO, BCA, R-LAC, OVA, and STI. Conditions: column, 50 µm i.d. × 340/ 260 mm, packed with strong anion exchanger; mobile phase, 50, 100, and 200 mM NaCl in 5 mM phosphate buffer, pH 7.0; applied voltage, -15 kV; injection, 6 s at -8 kV; UV detection, 200 nm.

of the EOF declines with increasing pH, it indicates that at least one of the above-mentioned three criteria is not met. It follows from Figure 1 that negatively charged sample components will electrostaticaly bind to the quaternary ammonium groups at the chromatographic surface. For this reason, relatively high salt concentration is needed to bring about the elution of the oppositely charged migrants. The magnitude of EOF also depends on the electrolyte concentration in the eluent, which has an inverse relationship to the thickness of the double layer.10 The electrosmotic velocity and the overall migration velocities of the proteins are plotted against the salt concentration in Figure 4. It is seen that the EOF only slightly changed with the salt concentration in the range from 50 to 200 mM. The reasons for the relatively low sensitivity of electrosmotic flow to salt concentration in packed columns are yet to be elucidated. Effect of Salt on the Separation of Proteins. The effect of salt concentration on the separation of proteins in CEC was investigated to gain information on the role of electrostatic interactions in the retention. The four proteins whose pI values and molecular masses are listed in Table 1 were subjected to CEC. The resulting chromatograms obtained at three salt concentrations

protein

pI

M (kDa)

bovine carbonic anhydrase (BCA) R-lactalbumin (R-LAC) soybean trypsin inhibitor (STI) ovalbumin (OVA) conalbumin hemoglobin

6.2 5.0 4.5 4.7 6.24, 6.68, 7.17 6.8

29 14 22.7 45 86 17.1 (× 4)

are shown in Figure 5. The peak efficiencies for each protein with the 340/260-mm-long column are evaluated and listed in Table 2. It is seen that the column efficiency obtained in the separation by CEC was on the average of 80 000 theoretical plates, which is considered to be quite high in protein chromatography. As seen from the electrochromatogram in Figure 5a, BCA and R-LAC elute when the mobile phase contains 50 mM NaCl. It is because their pI values are close to 7.0 at which pH they are not or only weakly bound to the stationary phase. The more strongly bound OVA and STI cannot be eluted at such low salt concentration, i.e., low eluent strength. The increase of salt concentration to 100 mM has just brought about the elution of OVA. The separation of all components was completed by isocratic elution with 200 mM NaCl in the eluent. The plots of the migration velocities of each protein against the salt concentration are shown in Figure 4. It is seen that the electrosmotic velocity is largely invariant with changing salt concentration, but the migration velocities of proteins increase significantly when the salt concentration is increased from 100 to 200 mM NaCl. The strong increase in migration velocities is believed due to an attenuation of protein binding to the chromatographic surface by electrostatic forces with increasing salt concentration. As estimated from their pI values and as shown in Figure 5c, the proteins were eluted in order of the strength of their electrostatic interaction with the quaternary ammonium groups of the stationary phase. Thus, the debinding behavior of the proteins in our CEC system is similar to that usually exhibited in ion-exchange chromatography. Furthermore, recent results24 obtained by similar CEC systems based on solvophobic instead of electrostatic interactions suggest a common mechanistic framework for the two approaches to protein separation by CEC. It is based on a putative dual mechanism arising from the interplay between chromatographic retention and electrophoretic migration. By using a sufficiently strong eluent, the chromatographic retention of the proteins is attenuated and their migration is initiated. In the process, the proteins are crudely separated by chromatography while the charged sample components are also subjected to electrophoresis. The electrochromatograms in Figure 5 also show that all the proteins are eluted before the EOF marker. It is because in our CEC system the electrophoretic migration of the proteins is codirectional with the EOF. Once the protein molecules are not bound by the chromatographic surface, they are driven both by the anodic electrophoretic force, which is separative, and by the electrosmotic force, which is not separative and measured with DMSO as the EOF marker. The resulting cocurrent movement of the sample components makes the proteins migrate faster than Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

3025

Table 2. Column Efficiencies Measured at Three Salt Concentrations with the Standard Proteinsa no. of theoretical plates NaCl (mM)

BCA

R-LAC

OVA

STI

50 100 200

90 300 133 800 16 100

86 800 78 000 39 200

86 600 92 600

102 900

a Conditions: column, 50 µm i.d. × 340/260 mm, packed with the strong anion exchanger; mobile phase, 50, 100, and 200 mM NaCl in 5 mM phosphate buffer, pH 7.0; applied voltage, -15 kV; injection, 6 s at -8 kV; UV detection, 200 nm.

Figure 6. Isocratic CEC of chicken egg white conalbumin variants. Conditions: column, 50 µm i.d. × 340/260 mm, packed with the strong anion exchanger; mobile phase, 200 mM NaCl in 5 mM phosphate buffer, pH 7.0; applied voltage, -15 kV; injection, 6 s at -8 kV; UV detection, 200 nm.

Figure 5. Isocratic CEC of standard proteins. Conditions: column, 50 µm i.d. × 340/260 mm; mobile phase, pH 7.0, 5 mM phosphate buffer in addition of (a) 50, (b) 100, and (c) 200 mM NaCl; applied voltage, -15 kV; injection, 6 s at -8 kV; UV detection, 200 nm. Sample: BCA, R-LAC, OVA, and STI.

the neutral marker. However, it should be noted that it is purely fortuitous that in this case all proteins migrate ahead of the EOF marker. CEC of Protein Variants. The potential of CEC for separation of protein mixture under conditions discussed above was explored first by separating the four variants of conalbumin from chicken egg white:30,31 one iron-free, two monoferric, and one diferric. The electrochromatogram depicted in Figure 6 was obtained with our 3026

Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

anion-exchanger column at pH 7.0. It shows that the peak efficiencies are rather high (∼55 000 theoretical plates with the 340/260-mm-long column) considering the low molecular diffusivities of the proteinaceous sample components. The pattern of the electrochromatogram is very similar to that obtained for isoelectric focusing of conalbumin variants.30,31 Another illustration of the separating potential of CEC is given in Figure 7 , which shows the separation of two components in bovine hemoglobin: oxyhemoglobin and methemoglobin. The electrochromatogram was obtained under essentially the same conditions as those described in Figure 6. The high resolution revealed in Figures 6 and 7 clearly demonstrates the potentially important role of CEC in bioanalytical chemistry. Effect of Salt Concentration on Protein Retention. If chromatographic elution and electrophoretic migration take place independently in CEC, the overall velocity of a charged sample component is given32,33 as

ucec ) (ueof + ueph)/(1 + k′lc)

(1)

or (30) Kila´r, F. In Handbook of Capillary Electrophoresis; Landers, J. P., Ed.; CRC Press: Boca Raton, FL, 1994; p 95. (31) Richards, M. P.; Huang, T. L. J. Chromatogr., B 1997, 690, 43. (32) Lelie`vre, F.; Yan, C.; Zare, R. N.; Gareil, P. J. Chromatogr., A 1996, 723, 145. (33) Grego, A. L.; Gonza´lez, A.; Maria, M. L. Crit. Rev. Anal. Chem. 1996, 26, 261.

Figure 8. Illustration of the overall migration velocity, ucec, of charged sample components in co-directional and counter-directional CEC systems. It is assumed that the overall migration velocity, ucec, is the sum of contributions by EOF and electrophoretic migration; i.e., the use of eq 1 is appropriate.

Figure 7. Isocratic CEC of bovine hemoglobin containing 75% methemoglobin and the rest oxyhemoglobin. Conditions: column, 50 µm i.d. × 340/260 mm, packed with the strong anion exchanger; mobile phase, 200 mM NaCl in 5 mM phosphate buffer, pH 7.0; applied voltage, -15 kV; injection, 6 s at -8 kV; UV detection, 200 nm.

ucec ) (1 + k′e)/(1 + k′lc)ueof

Rs )

xN (ui/uj) - 1 4 (ui/uj) + 1

(5)

(2)

where k′lc is the chromatographic retention factor, ueof is the electrosmotic velocity, ueph is the electrophoretic velocity of charged sample components in the column, and k′e is the velocity factor.12 The migration behavior of charged sample components in CEC system is illustrated in Figure 8. To represent the dependence of k′lc on the salt concentration in CEC of proteins, we assume that the concentration of eluting salt on the retention behavior of proteins is described by the threeparameter equation,34 which was developed for electrostatic interaction chromatography, as

log k′lc ) A - B log ms + Cms

electrophoretic mobilities. In the following, the resolution35 is expressed by using the velocity frame as

where ui and uj are the overall migration velocities of the two peaks i and j under consideration. Substituting eq 1 into eq 5, the resolution in CEC is expressed as

xN λi,jγi,j - 1 4 λi,jγi,j + 1

(6)

ueof + ueph,i 1 + k′e,i ) ueof + ueph,j 1 + k′e,j

(7)

1 + k′lc,j 1 + k′lc,i

(8)

Rs )

where

λi,j )

(3) and

where ms is the molality of the salt in the eluent and B is the electrostatic interaction parameter and depends on the characteristic charge of the protein and the salt counterions. If hydrophobic interactions are negligible, that is, the C term is very small, eq 3 can be simplified to the expression as

log k′lc ) A - B log ms

(4)

Equation 4 was used to estimate k′lc in isocratic electrostatic interaction chromatography for three model proteins at different salt concentrations based on literature data34 and the results are listed in Table 3. The k′lc data illustrate that, at low salt concentration, the sample components were well separated while at high salt concentration, the difference between the retention factors is greatly reduced. Effect of Chromatographic Retention and Electrophoretic Migration on the Resolution. The resolution, Rs, is the ultimate measure of the separation of two sample components in chromatography. It is of interest to express Rs as a function of the two governing parameters of CEC: chromatographic retention and (34) Melander, W. R.; Rassi, Z. E.; Horva´th, Cs. J. Chromatogr. 1989, 469, 3.

γi,j )

It is seen that λi,j is the ratio of overall migration rates of the two peaks i and j in the absence of chromatographic retention; consequently, it represents for the contribution of electrophoretic migration to the resolution in CEC. When λi,j ) 1.0, eq 7 becomes the well-known equation of resolution of chromatography. On the other hand, γi,j is the ratio of the chromatographic retention factors of the two peaks i and j; thus it expresses the contribution of chromatographic retention to the resolution in CEC. The ratio of the respective resolutions measured by CEC and HPLC with a given column under otherwise identical conditions, which we shall call the resolution ratio, is expressed as follows

Rs,CEC ) Rs,HPLC

(x )(

)( )

NCEC λi,jγi,j - 1 γi,j + 1 NHPLC λi,jγi,j + 1 γi,j - 1

(9)

Equation 9 shows the contributions to the resolution ratio by the (35) Horva´th, Cs.; Melander, W. R. Theory of Chromatography. In Chromatography; Heftmann, E., Ed.; Elsevier: Amsterdam, 1983; pp A27-A135.

Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

3027

Table 3. Estimated Values of the Chromatographic Retention Factors of r-Chymotrypsinogen (CHY), Cytochrome c (CYT), and Lysozyme (LYS) at Different Molalities of NaCl in the Range from 0.05 to 0.4 m by Using Eq 4a klc′ NaCl (m)

CHY

CYT

LYS

0.05 0.10 0.15 0.20 0.25 0.30 0.40

4.83 0.71 0.22 0.10 0.05 0.03 0.02

9.89 1.95 0.74 0.39 0.22 0.15 0.08

38.11 5.25 1.60 0.72 0.37 0.21 0.10

a Parameter values at pH 7.8 are taken from the literature.34 Parameters: CHY, A ) -2.93, B ) 2.78; CYT, A ) -2.06, B ) 2.35; LYS, A ) -2.15, B)2.87.

column efficiency, by the electrophoretic migration, and by the chromatographic retention. As seen from eq 9, the resolution ratio can be expressed by three terms as in the classical resolution equation of chromatography. The resemblance is not unexpected, but it cannot extend to the term representing electrophoretic migration. To gain further insight in the mechanism of protein separation by CEC, as described in this paper, we shall analyze the effects of electrophoretic migration and column efficiency on the resolution ratio. In this endeavor we assume that both HPLC and CEC exhibit the same chromatographic retention mechanism. First we examine the contribution of electrophoretic migration to the resolution ratio. In packed columns, the effective electrophoretic mobilities of charged sample components are lower than in open tubes, due to the tortuosity and porosity of the column and the relatively low effective diffusivities of biomacromolecules36,37 in porous media. Figure 9 shows plots of the simulated resolution ratio against salt concentration. It is seen by using data estimated from the literature26,34 that at sufficiently high salt concentration the resolution ratio becomes negative, indicating a reversal of the elution order of the two sample components under consideration. This is because the chromatographic retention is greatly reduced by increasing the ionic strength of the eluent so that the electrophoretic process, which has a selectivity of opposite sign in this case, becomes the dominant separation mechanism (Note that we depart from the convention that the selectivity is always positive.). Selectivity change will be limited, however, when the effective electrophoretic mobilities of the sample components are much smaller that the electrosmotic mobility, i.e., when λi,j is close to unity. However, as described previously in the literature,38 the architecture of duplex CEC column, which consists of an open and packed segment, allows for an additional degree of freedom to manipulate the selectivity by varying the length ratio of the two segments. Assuming a column efficiency of N ) 4000 in isocratic HPLC and using the retention factors in Table 3, the resolutions of the three model proteins at the NaCl concentration ms ) 0.40m are (36) Frey, D.; Schweinheim, E.; Horva´th, Cs. Biotechnol. Prog. 1993, 9, 273. (37) Park, Y. G. Korean J. Chem. Eng. 1999, 16, 128. (38) Rathore, A. S.; Horva´th, Cs. Anal. Chem. 1998, 70, 3271.

3028

Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

Figure 9. Graph illustrating plots of the resolution ratio, Rs,CEC/ Rs,HPLC, against the salt molalities in the eluent with λi,j as the parameter. For the two proteins under consideration, R-chymotrypsinogen and cytochrome c, the retention factors at different salt molalities were obtained from Table 3 for the estimation of γi,j. The effective electrophoretic mobilities under CEC conditions of the two proteins were used to estimate the values of λi,j as 0.795, 0.854, and 0.955 for 100%, 50%, and 10% of their electrophoretic mobilities in CZE,26 respectively.

estimated as Rs,CYT/CHY ) 0.883 and Rs,LYS/CYT ) 0.333. This means that they cannot be baseline separated at unit resolution. On the other hand, in CEC it is rather easy to obtain even for proteins column efficiencies on the order of 104 theoretical plates as shown by the experimental data in Table 2. For such column efficiencies, eq 9 predicts at least 3-fold higher resolution ratios. In this case, one can predict that Rs,CYT/CHY and Rs,LYS/CYT will be greater than unity; i.e., baseline separation of the components will be obtained. Peak Capacity in Various Elution Modes. High peak capacity and high speed of analysis are very much in demand in the development of a new analytical separation process for use in life sciences and biotechnology. The peak capacity in isocratic elution chromatography is given39 approximately by the expression as,

n ) 1 + (xN/4) ln(Vz/Vo)

(10)

whereas in gradient elution chromatography40 the following relationship applies

n ) (xN/4)[(Vz/Vo) - 1]

(11)

where Vo and Vz are the retention volumes of the first and last peaks. The ratio Vz/Vo in eqs 10 and 11 is termed as the elution window. In practice of HPLC, proteins are separated by gradient elution rather than isocratic elution in order to narrow the elution window and improve the peak capacity. As seen in Figure 5, proteins are separated in isocratic CEC at high salt concentration with a narrow elution window. In (39) Giddings, J. C. Anal. Chem. 1967, 39, 1027. (40) Horva´th, Cs.; Lipsky, S. R. Anal. Chem. 1967, 39, 1893.

It should be recognized that other factors such as the phase ratio of the column have to be considered. Evidently, a quantitative treatment of the separation process in CEC requires a highly elaborated study.

Figure 10. Schematic illustration of the peak capacity against the elution window for isocratic and gradient elution in HPLC and isocratic elution in CEC. The elution window is expressed by the ratio Vz/Vo, where Vo and Vz are the retention volumes of the first and last peaks. Column efficiency in use for the plot: N ) 4000 in HPLC isocratic and gradient elution; N ) 10 000 and 100 000 in CEC isocratic elution.

isocratic elution chromatography, the elution window Vz/Vo is given by

Vz/Vo )

1 + k′lc,z 1 + k′lc,0

(12)

It is seen from Table 3 that the elution window narrows as the eluent strength increases. If the peak capacity can be estimated by eq 10 in CEC, Figure 10 shows the comparison of peak capacities in isocratic, gradient elution chromatography and isocratic CEC at different column efficiencies. It is seen that when the elution window is sufficiently narrow, CEC results in peak capacities higher than those obtained in gradient and isocratic elution HPLC. Moreover, the critical value of the elution window below which CEC shows advantages over the other two elution modes becomes wider with an increase in the column efficiency.

CONCLUSIONS The results of this study demonstrate that the introduction of novel stationary phases, which are tailored to the separation problems at hand, can be of advantage in striving for rapid and high-resolution separation of proteins by CEC. Furthermore, they give some insight into the mechanism of the electrochromatographic separation process and provide an explanation why protein separation by CEC can be carried out with isocratic elution. The stationary phase, which has positive charges attached to oligomeric chains that are in turn anchored to the surface of the support, has been successfully employed for the separation of proteins. Since the oligomeric chains are neutral and hydrophilic, electrostatic interactions between the proteins and the quaternary ammonium functions dominated the chromatographic retention at neutral pH where the proteins were negatively charged. The result concerning the mechanistic aspects of this study can be summarized as follows. For the isocratic separation of proteins, the eluent strength of the mobile phase is kept at a relatively high level so that the magnitude of the chromatographic retention factors of the sample components is relatively small. Therefore, they will elute in a narrow elution window and are ultimately separated by the column that has high intrinsic efficiency due to the use of electrosmotic flow.27 ACKNOWLEDGMENT This work was supported by Grant GM 20993 from the National Institutes of Health, U.S. Department of Health and Human Services, and a grant from the National Foundation of Cancer Research. Received for review February 3, 2000. Accepted April 9, 2000. AC000114T

Analytical Chemistry, Vol. 72, No. 14, July 15, 2000

3029