Affinity Monoliths Generated by In Situ Polymerization of the Ligand

Anal. Chem. 2001, 73, 5126-5132

Affinity Monoliths Generated by In Situ Polymerization of the Ligand R. Hahn,† A. Podgornik,‡ M. Merhar,‡ E. Schallaun,† and A. Jungbauer*,†

Institute for Applied Microbiology, University of Agricultural Sciences, A-1190 Vienna, Austria, and BIA Separations d.o.o., Sl-1000 Ljubljana, Slovenia

An affinity monolith with a novel immobilization strategy was developed leading to a tailored pore structure. Hereby the ligand is conjugated to one of the monomers of the polymerization mixture prior to polymerization. After the polymerization, a monolithic structure was obtained either ready to use for affinity chromatography or ready for coupling of additional ligand to further increase the binding capacity. The model ligand, a peptide directed against lysozyme, was conjugated to glycidyl methacrylate prior to the polymerization. With this conjugate, glycidyl methacrylate, and ethylene dimethacrylate, a monolith was formed and tested with lysozyme. A better ligand presentation was achieved indicated by the higher affinity constant compared to a conventional sorbent. Affinity chromatography has been employed with great success for analytical and preparative applications whenever complex mixtures of biological samples had to be separated.1-6 The quality of the separation and the respective reliability depend on the optimum chromatographic material. Mostly, the chromatographic matrixes for this purpose are porous matrixes. By choosing the grade of porosity, the chemical nature of the surface of the matrixes, or both, the chromatographic processes can be influenced and tailored for the respective separation problem. For generation of a suited affinity sorbent, an affinity ligand must be immobilized onto the surface of the chromatography material. Usually as a solid matrix, particulate materials with a certain particle size and particle size distribution are used.7,8 The bond between the ligand and the solid matrix has to be at least as stable as the material to survive the separation conditions. Furthermore, if the immobilization of a ligand has to be performed chemically, * Corresponding author: (fax) 0043-1-36006 1249; (e-mail) jungbaue@ hp01.boku.ac.at. † University of Agricultural Sciences. ‡ BIA Separations d.o.o. (1) Chaiken, I. M.; Wilchek, M.; Parik, I. Affinity Chromatography and Biological Recognition; Academic Press: New York, 1983. (2) Cuatrecasas, P.; Wilchek, M.; Anfinsen, C. B. Proc. Natl. Acad. Sci. U.S.A. 1968, 61, 636-643. (3) Turkova, J. Affinity Chromatography; Elsevier: Amsterdam, 1978. (4) Wilchek, M.; Chaiken, I. Methods Mol. Biol. 2000, 147, 1-6. (5) Scouten, W. H. Chemical Analysis: Affinity Chromatography; John Wiley & Sons: New York, 1981. (6) Lowe, C. R. An Introduction to Affinity Chromatography; Elsevier: Amsterdam, 1979. (7) Porath, J. J. Chromatogr. 1981, 218, 241-259. (8) Narayanan, S. R. J. Chromatogr., A 1994, 658, 237-258.

5126 Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

the reaction conditions have to be chosen in such a way that the binding properties of the ligand are not adversely affected. Typically, the immobilization of a ligand has been performed during a multistep procedure, in particular for organic matrixes. Chemical activation of the carrier is the first step with most organic chromatographic substances. Then a spacer having a reactive or activatable group is introduced. When the spacer does not have a reactive group, a second step for activation has to be performed. If bifunctional reagents are used, spacer introduction and activation of the spacer are reduced to a single step. Finally, the ligands are coupled and the remaining reactive groups are blocked.9 Immobilization of polypeptides on chromatographic media occurs in a random manner.10 The immobilization can only imperfectly be influenced by adjusting the reaction conditions. It may happen that the amino acids responsible for the affinity interaction between the ligand immobilized on the surface of the matrix and the molecule to be separated are not available anymore for biospecific interaction. Furthermore, due to the multipoint attachment, the three-dimensional structure of the ligand bound to the matrix may be altered in such a way that the binding pocket will be deformed so that the substance to be separated cannot bind anymore.11 Therefore, the immobilization of an affinity ligand must be carefully designed to achieve the optimal performance. Besides the immobilization chemistry, the immobilization strategy must also consider the porous structure of the matrixes in order to place the ligand in the optimal physical position. When small peptide ligands, which became very popular due to combinatorial chemistry12-15 and biological peptide libraries,16 are immobilized, the issue of immobilization strategy is even more critical compared to conventional macromolecular ligands. The likelihood of covalent modification of a critical amino acid during immobilization is higher compared to macromolecules whenever a random coupling procedure is applied. Macromolecular ligands (9) Hermanson, G. T.; Mallia, A. K.; Smith, P. K. Immobilized Affinity Ligand Techniques; Academic Press: San Diego, 1992. (10) Turkova, J. J. Chromatogr., B 1999, 722, 11-31. (11) Walters, R. R. Anal. Chem. 1985, 57, 1099A-1101A, 1102A-1106A. (12) Amatschek, K.; Necina, R.; Hahn, R.; Schallaun, E.; Schwinn, H.; Josic, D.; Jungbauer, A. J. High Resolut. Chromatogr. 2000, 23, 47-58. (13) Fodor, S. P.; Read, J. L.; Pirrung, M. C.; Stryer, L.; Lu, A. T.; Solas, D. Science 1991, 251, 767-773. (14) Houghten, R. A. Proc. Natl. Acad. Sci. U.S.A. 1985, 82, 5131-5135. (15) Huang, P. Y.; Carbonell, R. G. Biotechnol. Bioeng. 1999, 63, 633-641. (16) Huang, P. Y.; Baumbach, G. A.; Dadd, C. A.; Buettner, J. A.; Masecar, B. L.; Hentsch, M.; Hammond, D. J.; Carbonell, R. G. Bioorg. Med. Chem. 1996, 4, 699-708. 10.1021/ac0103165 CCC: $20.00

© 2001 American Chemical Society Published on Web 10/03/2001

Figure 1. Schematic drawing of the preparation of an affinity monolith produced by in-situ polymerization.

extend the molecule into the porous space. This does not happen with small peptides. They may be shielded by the fibrous structure forming the backbone of the matrix. Another difficulty may arise when the active part of the ligand cannot be reached by the binding partner in the mixture to be separated. The ligand may be immobilized in pores that are not accessible by the protein. Either the diffusion is restricted or the superficial area of the pore is smaller than the protein.17 In contrast to ion-exchange chromatography, long-action forces are not present in affinity matrixes and cannot enhance effective diffusion. Monolithic stationary phases based on polyacrylamide,18-21 polymethacrylate,22-25 or aged silica26-28 exhibit more efficient mass-transfer properties compared to conventional chromatographic media. The lower mass-transfer resistancy has been explained by a reduced diffusional length and convective transport.29,30 In this work, we introduce a novel stationary phase which combines the efficiency of monoliths with the selectivity of affinity ligands, providing a new dimension in bioaffinity chromatography. (17) Petropoulos, J. H.; Liapis, A. I.; Kolliopoulos, N. P.; Petrou, J. K.; Kanellopoulos, N. K. Bioseparation 1990, 1, 69-88. (18) Hjerten, S.; Li, M.; Mohammed, J.; Nakazato, K.; Pettersson, G. Nature 1992, 356, 810-811. (19) Liao, J. L.; Hjerten, S. J. Chromatogr. 1988, 457, 175-182. (20) Hjerten, S.; Liao, J.; Zhang, R. J. Chromatogr. 1989, 473, 273-275. (21) Liao, J. L. In Advances in chromatography; Brown, P. R., Grushka, E., Eds.; Marcel Dekker: New York, 2000; Vol. 40, pp 467-502. (22) Strancar, A.; Koselj, P.; Schwinn, H.; Josic, D. Anal. Chem. 1996, 68, 34833488. (23) Tennikova, T.; Svec, F.; Belenkii, B. G. J. Liq. Chromatogr. 1990, 13, 63. (24) Strancar, A.; Barut, M.; Podgornik, A.; Koselj, P.; Josic, D.; Buchacher, A. LC-GC Int. 1998, (Oct). (25) Tennikova, T. B.; Freitag, R. J. High Resolut. Chromatogr. 2000, 23, 2738. (26) Minakuchi, H.; Nakanishi, K.; Soga, N.; Ishizuka, N.; Tanaka, N. Anal. Chem. 1996, 68, 3498-3501. (27) Minakuchi, H.; Nakanishi, K.; Soga, N.; Ishizuka, N.; Tanaka, N. J. Chromatogr., A 1997, 762, 135-146. (28) Cabrera, K.; Lubda, D.; Eggenweiler, H.; Minakichi, H.; Nakanishi, K. J. High Resolut. Chromatogr. 2000, 23, 93-99. (29) Hahn, R.; Jungbauer, A. Anal. Chem. 2000, 72, 4853-4858. (30) Iberer, G.; Hahn, R.; Jungbauer, A. LC-GC 1999, 17, 998-1005.

The affinity ligand, in our case a peptide directed against lysozyme,31 was conjugated to one monomer of the polymerforming constituents prior to the polymerization. The affinity peptide was reacted with glycidyl methacrylate (GMA), and a block polymer was formed by copolymerization of this conjugate with GMA and ethylene dimethacrylate (EDMA). The influence of the amount of conjugate, the porogen composition and the polymerization conditions was studied in respect to pore size and binding capacity. The purification of lysozyme was tested with the optimized affinity monolith. MATERIALS AND METHODS Production of Conjugate. The peptide with the sequence H2N-YLSYPLTFGA (molecular weight 1131) was synthesized on a synthesizer 431 from Applied Biosystems (Foster City, CA) using Fmoc chemistry.32 Larger quantities were purchased from Novabiochem (San Diego, CA). A 200-mg sample of the peptide was reacted with 2.1 mL of GMA in a buffer containing 1 mL of 100 mM Na2CO3, pH 8.5, and 2 mL of ethanol. The solution was vigorously stirred for 18 h at room temperature. The reaction was followed by analytical reversed-phase chromatography with a Luna C18 (3 µm) column (Phenomenex, Torrance, CA) performed on a HP-1090 station (Agilent Technologies, Vienna, Austria). The column dimension was 4.6 × 10 mm, and the flow rate was 1 mL/ min. The peptide-GMA conjugate was purified on a Luna C 18 (15 µm) column (Phenomenex). The column dimension was 21.2 × 250 mm, and the flow rate was 30 mL/min. Buffer A was 0.1% aqueous trifluoroacetic acid (TFA), and buffer B was acetonitrile containing 0.1% TFA. A 10-mL sample of the conjugate solution was loaded and eluted with a linear gradient from 5 to 40% buffer B over 50 min. For column regeneration, the gradient was raised to 60% B. The purification was carried out on a preparative HPLC station, consisting of two P-3500 pumps, an LCC(31) Welling, G. W.; van Gorkum, J.; Damhof, R. A.; Drijfhout, J. W.; Bloemhoff, W.; Welling-Wester, S. J. Chromatogr. 1991, 548, 235-242. (32) Fields, G. B.; Noble, R. L. Int. J. Pept. Protein Res. 1990, 35, 161-214.

Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

5127

Figure 2. Preparation of the peptide-GMA conjugate. (A) Analytical RP chromatogram of the reaction: (1) GMA, (2) peptide, (3) peptideGMA substituted on the N-terminus, (4) peptide-GMA substituted on a tyrosine residue, and (5) peptide substituted with 2 GMA residues. (B) Preparative RP-HPLC: (1) GMA, (2) peptide, (3) peptide-GMA, and (4) peptide-2GMA. (C) Mass spectrogram of peak 3 of the preparative RP-HPLC chromatogram. (D) Mass spectrogram of peak 4 of the preparative RP-HPLC chromatogram.

500 Plus controller, two gradient mixers, and an UV-M detector at 214 nm (Amersham Pharmacia Biotech, Uppsala, Sweden). The eluates were evaporated and freeze-dried. The molecular mass of the eluted fractions was determined by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (ThermoBioanalysis Ltd., Santa Fe, NM). Preparation of Monoliths. Monoliths were synthesized by radical copolymerization of GMA and EDMA in the presence of pore-forming solvents, cyclohexanol (CyOH) and dodecanol (DoOH), as described by Tennikova et al.23 A total of 5, 15, and 30% (weight) of the standard amount of GMA was substituted by the peptide-GMA conjugate. Polymerizations were carried out at 65, 62, 59, and 57 °C. Commercially available CIM epoxy disks (Convective Interaction Media) were obtained from BIA Separations (Ljubljana, Slovenia). For immobilization of peptides, monoliths were washed with the coupling buffer, 100 mM Na2CO3, pH 9.5. The peptide was dissolved at a concentration of 10 mg/mL and allowed to couple for 48 h. All monoliths had a height of 3 mm and a diameter of 12 mm. To measure pore size distribution, dried monoliths were immersed in liquid mercury. The volume of the intruded mercury is a direct measure for the penetrated pore volume. The experiments were performed on a mercury porosimeter Pascal 440 (ThermoQuest Italia S.p.A., Rodano, Italy). The porosimeter is a closed system. To avoid any contamination, the system was placed in a dedicated room. Scanning electron 5128

Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

microscopy (SEM) was performed with a DSM 990 digital scanning electron micrograph from Zeiss (Oberkochen, Germany). Affinity Chromatography Conditions. All experiments were performed on an A¨ kta explorer 100 (Pharmacia). For pulse response experiments, 50 µL of a 1% acetone solution was injected. The flow rate was 0.94 mL/min, corresponding to a linear flow velocity of 50 cm/h. Breakthrough curves were performed with lysozyme at a concentration of 0.2 mg/mL dissolved in 20 mM Tris-HCl at pH 7.5. A 1 M NaCl solution in Tris buffer was applied for elution of the bound protein. Protein was loaded until full breakthrough was achieved. For testing selectivity, yeast extract was diluted to a conductivity of 2.5 mS/cm, the pH was adjusted to 7.5, and the solution was spiked with lysozyme to a final concentration of 0.2 mg/mL. Ten milliliters of the sample was loaded onto one monolith after equilibration with 20 mM Tris buffer, pH 7.5. Elution was carried out with 1 M NaCl in Tris buffer. Fractions were analyzed by SDS-PAGE on an Xcel Minicell system from Novex (San Diego, CA). The separation was performed on 4-25% Tris-glycine Novex Precast gels. Proteins were stained with Coomassie Brilliant Blue R 250, 0.1% (w/v) in 10% acetic acid, 40% methanol. Destaining was performed with 8% acetic acid and 25% ethanol. Adsorption Isotherm. The amount of bound protein was determined by breakthrough curves. Amounts of 0.1, 0.2, 0.5, 1.0,

Figure 3. Pore size distribution of affinity monoliths measured by mercury intrusion porosimetry. (A) Influence of the conjugate concentration at a polymerization temperature of 59 °C. (B) Influence of the temperature at a conjugate concentration of 15%. (C) Influence of the dodecanol concentration at a conjugate concentration of 30% and a polymerization temperature of 57 °C. The standard is a commercially available monolith polymerized at 65 °C at a dodecanol concentration of 6%.

and 2.0 mg of lysozyme/mL in 20 mM Tris-HCl buffer were applied. The equilibrium concentration q in the stationary phase was calculated using the following expression:

q)

1 (C (V - V0)) Vc 0 e

∫

Ve

V0

Figure 4. Pulse response experiments of affinity monoliths polymerized at different temperatures.

the void volume, and Ve is the volume of the applied protein solution. From the analysis, an adsorption isotherm was calculated and fitted by the experimental data with the Langmuir adsorption isotherm

q)

qmaxc KD + c

C(V) dV

where Vc is the bed volume, C0 is the feed concentration, V0 is

where q is the bound protein in the stationary phase, qmax is the maximum binding capacity, KD is the equilibrium dissociation Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

5129

Figure 5. Scanning electron micrographs of affinity monoliths. (A) Influence of the conjugate concentration on the bed structure. (B) Influence of the dodecanol (DoOH) concentration on the bed structure.

constant, and c is the equilibrium concentration in the mobile phase. RESULTS AND DISCUSSION The critical issue since invention of affinity chromatography was the strategy of immobilization to achieve the maximum possible ligand concentration and an optimal ligand utilization.33 To reach this goal, the ligand has been immobilized through various activated groups. Many attempts were made to increase them. Increasing the pore size improves the mass-transfer properties and accessibility but reduces the physical stability of the matrix as well as the surface and consequently the binding capacity. Ideally the backbone of the matrixes or parts of the backbone should serve as the recognition site. In our approach, we intended to combine the excellent mass-transfer properties of monoliths with the high selectivity of affinity chromatography. A strategy with high capacity was used where the affinity ligand was directly polymerized into the monolithic block. Preparation of Conjugate. The strategy is outlined in Figure 1. According to this strategy, GMA was reacted with our model peptide at room temperature as described in the Material and Methods section. The reaction was monitored by analytical reversed-phase chromatography, which indicated that several byproducts have been formed (Figure 2A). Free GMA, free peptide, single-conjugated peptide, and double-conjugated peptide were detected. To achieve a high yield of conjugated peptide, the reaction conditions were optimized. A 70-fold molar excess of GMA over the peptide and a reaction time of 18 h gave the best results. A higher excess is not beneficial. The components reach the limit of solubility, and additionally, the tyrosine residue will also be conjugated. An increase of peak 5, the peptide containing two GMA residues, was observed with increasing GMA content. The peak denoted as 4 in the chromatogram was attributed to a peptide conjugated with one GMA moiety through one of the tyrosine residues. This was made visible by the identical mass observed also for the peptide conjugated through the N-terminus. Welling et al..31 showed that this model peptide should be immobilized through the N-terminus. Therefore, it is not desirable to have significant contamination of peptides conjugated through an internal amino acid. The reaction of an epoxy group with the hydroxy group of the tyrosine side chain is very slow; thus, mainly (33) Wilchek, M.; Miron, T.; Kohn, J. Methods in Enzymology: Affinity Chromatography; Academic Press: New York, 1984.

5130

Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

R-amino conjugated peptides were obtained. With preparative HPLC, the double-conjugated peptide and the unreacted compounds were separated (Figure 2B). The various fractions were further analyzed by MALDI-TOF mass spectrometry (Figure 2C and D). The theoretical molecular weight of the single- and doubleconjugated peptide could be identified. The peptide has a molecular mass of 1131 Da and GMA, of 142 Da. The sodium peaks are shown in Figure 2C and D. The obtained pure peptide conjugate, containing one GMA moiety linked through the N-terminus, was then used as one of the three components for formation of an affinity monolith. Polymerization of Monoliths. GMA, EDMA, and the peptide-GMA conjugate were reacted in various ratios with different amounts of solvents responsible for the pore formation. The reactions were also performed at different temperatures. From all monoliths, the porosity was determined by mercury intrusion porosimetry. In Figure 3, the effects of peptide-GMA conjugate concentration, temperature, and dodecanol concentration are shown. The pore radius decreases with increasing conjugate concentration (Figure 3A). Due to the small pores in the case of the highest conjugate concentration, an unacceptable highpressure drop was achieved. Therefore, further optimization steps of the polymerization conditions were performed. As shown by Svec and Frechet,34 for preparation of GMA-EDMA monoliths, the polymerization temperature and the solvent composition, the so-called porogenes, influence the pore size. The effect of temperature on the pore size is shown in Figure 3B. At lower temperatures, monoliths with larger pores were obtained, demonstrating the same trends as in the case of monoliths without copolymerized conjugate. A polymerization temperature of 57 °C and a conjugate concentration of 30% resulted in a same pore size distribution as typical for commercially available epoxy monoliths. To achieve the same pore size distribution with 30% conjugate, however, even lower temperatures should be applied. Unfortunately, in this case, the polymerization proceeded only partially; therefore, a different approach is required to obtain larger pores. The effect of dodecanol concentration is shown in Figure 3C. The polymerization temperature was kept low at 57 °C, but the dodecanol concentration was increased to create larger pores resulting in a reduced back pressure. The high back pressure of the monoliths with small pores is not the only problem. This was made visible by pulse response methods. A (34) Svec, F.; Frechet, J. M. J. Chem. Mater. 1995, 7, 707-715.

Figure 7. Adsorption isotherms of affinity monoliths with lysozyme. For the epoxy CIM disk KD ) 6.3 × 10-9 mol and for the disk with copolymerized peptide KD ) 1.4 × 10-9 mol.

Figure 8. SDS-PAGE of eluted fractions when yeast extract spiked with lysozyme was loaded onto an affinity monolith and elution was carried out with NaCl. Columns: (1) molecular weight marker; (2) sample; (3) and (4) flow-through; (5) washing; (6) eluate.

Figure 6. Breakthrough curves of affinity monoliths with different conjugate concentrations, with further immobilized peptide at different flow velocities. Standard is a commercially available monolith.

small pulse of 1% acetone was injected, and the effluent was monitored at 280 nm. Splitted peaks indicate irregularities in the bed structure. In Figure 4, it can be clearly seen that splitted peaks disappeared with lower polymerization temperatures. The change of the pore distribution by the different polymerization conditions was also corroborated by the structure made visible by SEM (Figure 5). With increasing conjugate concentration, the monolith gets more dense compared to a commercially available reference

material. The effect of dodecanol is even more striking (Figure 5B). Without dodecanol, almost no pores are visible. Irregularities within the bed, suspected due to splitted peaks observed with pulse response experiments, could not be revealed, although samples were cut from different sites of the monolith. Characterization of Chromatographic Properties. To examine the capacity, breakthrough curves were performed with lysozyme. Inhomogeneous monoliths indicated by splitted peaks in the pulse response experiments showed biphasic breakthrough curves (Figure 6). Due to the inhomogeneous structure of the monoliths produced at 62 and 59 °C, an early breakthrough was observed at higher flow velocities. Monoliths produced at a polymerization temperature of 57 °C gave reproducible steep breakthrough curves. As mentioned, these monoliths also showed Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

5131

pulse response profiles without any splitted peaks or significant tailing. Finally, a conjugate concentration of 30%, a dodecanol concentration of 8%, and a polymerization temperature of 57 °C were found to be the most appropriate polymerization conditions to obtain affinity monoliths with the best chromatographic properties. To further improve the binding characteristics of the affinity monolith, additional ligand was immobilized by reacting the peptide with the free epoxy groups. An adsorption isotherm of such an affinity monolith was compared to a conventionally produced monolith, where the peptide was reacted with the epoxy groups after the polymerization (Figure 7). The novel monolith has a higher binding capacity and the dissociation constant is 4 times smaller. This means that more peptide is exposed in a way accessible for the target protein, and consequently, the binding capacity is higher for low protein concentrations. A wide linear range in the adsorption isotherm makes such an affinity monolith better suited for analytical applications. An almost linear adsorption is observed up to a solidphase concentration of 6 mg/mL. With conventional application, this value is below 2 mg/mL. The peptide utilization during the polymerization procedure and the immobilization are almost ideal. This was confirmed by amino acid analysis. A total of 288 mg of conjugate was copolymerized, forming a 4-mL monolith. An aliquot was subjected to amino acid analysis. A total of 99% of the theoretical amount of amino acids was found. From this point of view, the ligand can be fully exploited. Additionally, a higher binding capacity was achieved by the novel polymer. Under equilibrium conditions, 12 mg of lysozyme could be bound onto 1 mL of solid support. Finally, the monolith was tested with a crude solution. Yeast extract was spiked with lysozyme and chromatographed. The SDS-PAGE of the eluted lysozyme is shown in Figure 8. Due to the high selectivity of the peptide, a pure protein could be recovered.

5132

Analytical Chemistry, Vol. 73, No. 21, November 1, 2001

CONCLUSION An immobilization strategy was developed where the ligand was copolymerized with the other monomers to a monolith. A careful optimization of the polymerization conditions led to a stable monolith with increased capacity and better ligand presentation. An advantage of the in situ polymerization is the possibility to form an affinity monolith of different geometry. Here we think of a stationary phase placed on chips or other analytic formats. This is not so easy with a conventional gel or monolith. The polymer can be either formed to a capillary or a more complex geometry used in biosensor technology. The demonstrated strategy exemplified by a peptide-protein interaction could in principle be extended to all other small ligands containing a functional group that can be conjugated with GMA. If multiple binding sites, e.g., NH2 groups and SH groups, are present, these additional groups must be selectively protected to achieve the conjugation through one favored group. The GMA conjugate must be soluble in solvents applied for this kind of polymerization. The polymerization reaction itself does not differ from the conventional formation of a polymethacrylate. After immobilization, additional functional groups are available to introduce a second ligand or charges that could be used for generation of electroendosmosis in capillary electrochromatography. ACKNOWLEDGMENT The work has been supported by the EUREKA project FAST 1 and a grant of the Austrian Research Foundation and OCTAPHARMA, Vienna.

Received for review March 15, 2001. Accepted August 16, 2001. AC0103165