Protein Analysis by Capillary Zone Electrophoresis Utilizing a

Department of Agricultural and Industrial Biotechnology, University of Verona, Strada Le Grazie No. 15, Verona 37134, Italy, and Department of Chemist...
0 downloads 0 Views 74KB Size
Download PDF
Anal. Chem. 2001, 73, 3862-3868

Protein Analysis by Capillary Zone Electrophoresis Utilizing a Trifunctional Diamine for Silica Coating Cecilia Gelfi,*,† Agnese Vigano`,† Marilena Ripamonti,† Pier Giorgio Righetti,‡ Roberto Sebastiano,§ and Attilio Citterio§

Institute of Advanced Biomedical Technologies, CNR, L.I.T.A., Via Fratelli Cervi 93, Segrate 20090 (Milano), Italy, Department of Agricultural and Industrial Biotechnology, University of Verona, Strada Le Grazie No. 15, Verona 37134, Italy, and Department of Chemistry, Politecnico of Milano, Via Mancinelli 7, Milano 20131, Italy

A novel method is here reported for the analysis of mixture of proteins with pI ranging from pH 3-9.5 in an ample pH interval (pH 2.5-9.0) without adsorption onto the naked silica wall. It consists of treating the capillary surface at alkaline pH, typically 9.0, with small amounts (2-4 mM) of a quaternarized piperazine derivative: (Nmethyl-N-ω-iodobutyl)-N′-methylpiperazine (Q-PzI). It appears that this compound is able to dock onto the wall via trifunctional links: a salt bridge via the quaternary nitrogen, a hydrogen bond via the tertiary nitrogen, and finally, a covalent link via the terminal iodine in the butyl chain and a neighboring ionized silanol. This last reaction seems to be completed in a few minutes of incubation of the capillary at room temperature. Because the compound is permanently affixed to the wall, its presence is not needed during protein/peptide separations. By properly dosing the level of Q-PzI in the preconditioning step, it is possible to strongly reduce the electroendoosmotic flow (EOF), zero it, or reverse it. Unlike dynamic coatings with oligoamines, which are most effective only at acidic pH values and are required as additives during separations, Q-PzI is effective in an ample pH interval (pH 2.5-9.0) and is not needed during the CZE analysis. A broad pI (pH 3-10) protein mix can be separated according to protein mobility in free phase, suggesting a strong modulating capacity of the functionalized wall. The same separation is not obtained in capillaries permanently coated with neutral, hydrophilic polymers (such as polyacrylamide), even if the quality of a single protein/peptide profile in Q-PzI-conditioned capillaries is equivalent to those obtained in capillaries permanently coated. Although there is strong indirect evidence of the ability of Q-PzI to alkylate the silica wall, to which it is then irreversibly bound, such an alkylation event does not occur with proteins on potentially reacting sites, such as the free -SH of Cys or the -OH group of Tyr, as demonstrated by incubating them overnight in a large molar excess at strongly alkaline pH values and analyzing such proteins by MALDI-TOF mass spectrometry. 3862 Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

Recently, capillary zone electrophoresis (CZE) has become a powerful tool for analyzing biopolymers and has been accepted as a high-performance separation technique applicable to a variety of molecules, including DNA fragments, proteins, and peptides. The performance of this technique is often compared to HPLC when small molecules are subjected to analysis but, in the field of protein separation, the performance of CZE must be compared with that of conventional gel electrophoresis techniques. Polyacrylamide gel electrophoresis involves many steps of manual handling, including gel preparation, sample loading, staining, and destaining. For quantitation of stained proteins, additional manual steps, such as densitometry or image analysis of the gels, are required. A potential alternative could be the use of the CZE technology; however, in protein analysis, significant tailing or irreversible adsorption onto the inner wall of fused-silica capillaries was observed, because a number of interactions, in general of cooperative nature, are established between the inner wall of the capillary and those macromolecules, compromising resolution and detection. In the proteomic era it represents a crucial problem to overcome. Several approaches have been proposed so far: (a) to control the pH or ionic concentration of the background electrolytes;1,2 (b) to modify the capillary by dynamic coating3-9 or physical absorption;10-14 and (c) to modify the capillary permanently by covalent bonding. * To whom correspondence should be addressed. Tel: +39-02-26423365. Fax: +39-02-26423302. E-mail: [email protected]. † Institute of Advanced Biomedical Technologies. ‡ University of Verona. § Politecnico of Milano. (1) McCornic, R. M. Anal.Chem. 1988, 60, 2322-2328. (2) Green, J. S.; Jorgenson, J. W. J. Chromatogr. 1989, 478, 63-70. (3) Muijselaar, W. G. H. M.; Bruijn, C. H. M. M.; Everaerts, F. M. J. Chromatogr. 1992, 605, 115-123. (4) Swedberg, S. A. Anal. Biochem. 1990, 185, 51-56. (5) Towns, J. K.; Regnier, F. E. Anal. Chem. 1991, 63, 1126-1132. (6) Strege, M. A.; Lagu, A. L. Anal. Biochem. 1993, 210, 402-410. (7) Yao, Y. J.; Li, S. F. Y. J. Chromatogr. A 1994, 663, 97-104. (8) Oda, R. P.; Madden B. J.; Morris, I. C. Spelsberg T. C.; Landers, J. P. J. Chromatogr. A 1994, 680, 341-351. (9) Wiktorowicz, J. E.; Colbrun, J. C. Electrophoresis 1990, 11, 769-733. (10) Preisler, J.; Yeung, E. S. Anal. Chem. 1996, 68, 2885-2889. (11) Iki, N.; Yeung, E. S. J. Chromatogr. A 1996, 731, 273-282. (12) Towns, J. K.; Regnier, F. E. J. Chromatogr. 1990, 516, 69-78. (13) Erim, F. B.; Cifuentes, A.; Poppe, H.; Kraak, J. C. J. J. Chromatogr. A 1995, 708, 356-361. (14) Morand, M.; Blass, D.; Kenndler, E. J. Chromatogr. B 1997, 691, 192196. 10.1021/ac001478o CCC: $20.00

© 2001 American Chemical Society Published on Web 07/10/2001

The last procedure can be divided into two methods: fixing a hydrophilic layer to the capillary wall by covalent bonding15-20 or fixing a stable cross-linked hydrophilic layer with another stable layer between the capillary wall and the hydrophilic layer, such as epoxy resin or highly cross-linked poly(styrene-divinylbenzene).21-23Procedure (a) is not very suitable for protein analysis, because choosing an extreme pH condition may denature these macromolecules; procedure (b), utilizing oligoamine additives in the background electrolyte, presents a strong drawback in the alkaline pH range as a result of deprotonation of the oligoamino backbone.24 Because the vast majority of protein separations occur in the alkaline range, such a wall conditioning cannot be utilized for protein separations. Procedure (c) avoids this problem, because analysis can be performed in a wide pH interval, but the drawback is represented by the long coating procedures, and quite often, unsatisfactory separations are obtained because of hydrophobic interactions with the wall. Recently, our group proposed a new chemical derivative that is most efficiently able to quench and control EOF along the pH scale,25 thus overcoming the long procedures that are required for covalent coating and the use of viscous polymers, which are not suitable for automation in microchips and capillary arrays. The present work describes the use of this new compound, (N-methylN-ω-iodobutyl)-N′-methylpiperazine, for CZE analysis of a number of proteins with pIs ranging from pH 3 to 9.0 and the comparison of protein separations in covalently coated capillaries. EXPERIMENTAL SECTION Reagents. Ovalbumin, human thrombin, human transferrin, superoxide dismutase (SOD), bovin serum albumin, soybean trypsin inhibitor, horse spleen ferritin, pig pepsin, β-lactoglobulin, human myoglobin, horse myoglobin, bovine and human carbonic anidrase, tripsinogen, β-casein, human albumin, and lactic dehidrogenase were from Sigma Aldrich (St. Louis, MO), phosphoglucomutase was from Boehringer (Mannheim, Germany), broad pI calibration kit (pH 3-10) was from Pharmacia Biotech (Uppsala, Sweden). Fused-silica capillaries (50 µm i.d. × 375 µm o.d.) were from Polymicro Technologies (Phoenix, AZ). Acrylamide, sodium tetraborate, vinylmagnesiumbromide, and all other analytical-grade chemicals were from Sigma Aldrich. (N-methyl-N-ωiodobutyl)-N′-methylpiperazine (QPzI) was synthesized as already described.24 Procedures. Capillary Treatment. The capillaries were washed with 1 M NaOH for 30 min, then rinsed in water for 30 min. The silica wall was treated for 30 min with a 2 mM solution of quaternarized piperazine [(N-methyl-N-×9dω-iodobutyl)-N′-methylpiperazine] (Q-PzI) in 25 mM sodium tetraborate buffer, pH 9.0, (15) Hjerte´n, S.; Kiessling-Johansson, M. J. Chromatogr. 1991, 550, 811-822. (16) Hjerte´n, S. J. Chromatogr. 1985, 347, 191-198. (17) Cobb, K. A.; Dolnik, V.; Novotny, M. Anal. Chem. 1990, 62, 2478-2483. (18) Nashabeh, W.; El Rassi, Z. E. J. Chromatogr. 1991, 559, 367-383. (19) Chiari, M.; Dell’Orto, N.; Gelain, A. Anal.Chem. 1996, 68, 2731-2736. (20) Gelfi, C.; Curcio, M.; Righetti, P. G.; Sebastiano, S.; Citterio, A.; Ahmadzadeh, H.; Dovichi, N. Electrophoresis 1998, 19, 1677-1682. (21) Liu, Y.; Fu, R.; Gu, J. J. Chromatogr. A 1996, 723, 157-167. (22) Ren, X.; Shen, Y.; Lee, M. J. Chromatogr. A 1996, 741, 115-122. (23) Huang, X.; Horvath, C. J. Chromatogr. A 1997, 788, 155-164. (24) Olivieri, E.; Sebastiano, R.; Citterio, A.; Gelfi, C.; Righetti, P. G. J. Chromatogr. A 2000, 894, 273-280. (25) Sebastiano, R.; Gelfi, C.; Righetti, P. G.; Citterio, A. J. Chromatogr. A 2000, 894, 53-61.

followed by a 3-min washing with 25 mM tetraborate, pH 9.0 (25). The protocol was repeated at least three times until the EOF was completely stabilized. The sample analysis was always performed using the following procedure: flushing with the modifier for 1.5 min and washing for 3 min with 25 mM tetraborate, pH 9.0, as running buffer. After each analysis, the capillary was washed with 50 mM CTAB in tetraborate buffer, pH 9.0. Covalently Coated Capillary. After the washing step with NaOH followed by water and methanol, the 50-µm capillary wall was covalently functionalized by the Grignard reaction as previously described.20 Five percent acrylamide monomer solution, after degassing and the addition of TEMED and ammonium persulfate, was allowed to react with the wall. After polymerization, an extensive washing step was performed for removing the unreacted monomers. EOF Measurements. Measurements of electrophoretic mobility (µ) were performed by considering the migration time of acrylamide under constant voltage

µ ) LtLd/tmV

where Lt and Ld refer to the total length and the length to detector of the capillary respectively, tm is the time for the sample to migrate to the detector under the applied voltage (V). The 1.4 M acrylamide solution was injected by pressure for 1 s, 5 psi/s, 25 °C, and 400 V/cm was applied. The detector was set at 214 nm. Run-to-run and day-to-day EOF values were determined by performing five replicate injections on successive days. Protein Separations. Protein samples were dissolved in 25 mM tetraborate buffer, pH 9.0, at 1 mg/mL final concentration. The sample was injected by pressure for 2 s, 5 psi/s, and run at 200 V/cm. Detection was at 214 nm. Run-to-run migration times were determined by performing five replicates. The single proteins were analyzed, for both a coated and a Q-PzI-treated capillary, in a 37cm-long capillary, 50 µm i.d.; the protein mixture was analyzed in a 77-cm-long capillary, 50 µm i.d., for the Q-PzI-treated capillary and the covalently coated control. Mass Spectrometry. Instrument. All measurements were performed using a TofSpec-2E MALDI-TOF mass spectrometer (Micromass, Manchester, U.K.), equipped with a pulsed nitrogen laser (337 nm; pulse width, 3 ns). Spectra of the proteins were recorded in the positive linear mode, with an acceleration voltage of 20 kV. Sample Preparation. Matrix solutions were freshly prepared when needed by dissolving 10 mg/mL sinapinic acid in a solvent mixture composed of 60% water (0.1% TFA) and 40% acetonitrile. Sample solution (2 µL) was mixed with an equivalent volume of matrix solution, and 1 µL of the resulting mixture was deposited on the target plate of the instrument and allowed to dry at room temperature. RESULTS AND DISCUSSION First of all, we had to assess how the addition of the diamine (Q-PzI) would modulate the EOF of the naked silica wall. The capillary wall was equilibrated with increasing amounts of Q-PzI, up to 2 mM concentration, and the EOF was measured by using acrylamide as a neutral marker. The equilibration step for each Q-PzI concentration was repeated until the EOF values were Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

3863

Figure 1. Assessment of EOF after treating the capillary wall with Q-PzI at increasing molarities, as described in the Experimental Section. Acrylamide was used as neutral marker. The equilibration step for each Q-PzI concentration was repeated until the EOF values were completely stabilized. Note the reversal of EOF above 1.5 mM Q-PzI.

completely stabilized. It can be appreciated from Figure 1 that, starting from 0.1 mM up to 1.5 mM Q-PzI at 25 °C, the EOF decreases by 1 order of magnitude; at higher molarities, up to 2 mM Q-PzI, the EOF was reversed. It can thus be perceived that the EOF value can be tuned by using the diamine to the point of reaching a zero-value and even of reversing the flux. The ability of Q-PzI to modulate EOF to a point of reaching a zero-value and even reversing it (above a minimum critical concentration, tentatively set between 1.5 and 2 mM Q-PzI) suggests progressive alkylation of silanols on the wall, which would ultimately leave an excess of positive charges on it that are able to reverse the EOF flux. The alkylation was performed at 25 °C, which is the temperature utilized for biomolecule analysis. Such behavior could be obtained only by covalently affixing the immobiline chemicals to the silica wall via binding a mixed polymer containing the neutral monomer (acrylamide) and either a number of positively or negatively charged immobilines (acrylamido weak acids and bases).26 A number of proteins (as listed in Table 1) were injected into the wall-modified capillary, and the resolution was compared with a capillary covalently coated with a neutral polymer (polyacrylamide).20 The sample injections were from the cathode to the anode, in the same direction as the flux. The separation conditions were as described in the Experimental Section. After each injection, the capillary was washed with a cationic detergent (CTAB). The detergent cannot interact with the positive wall, but it can with proteins eventually adsorbed to the silica surface. This procedure was introduced to guarantee a constant EOF value that is not influenced by the material eventually pasted to the wall. The table lists the pI values, the transit times at the detector window, the area log of all of the species that were analyzed, and the normalized values in order to account for transit time variations27 We can observe that the areas of the protein peaks were quite comparable with both coatings; in the case of soybean trypsin inhibitor, ferritin, and β-casein, the area logs were slightly (26) Capelli, L.; Ermakov, S. V.; Righetti, P. G. J. Biochem. Biophys. Methods 1996, 32, 109-124. (27) Altria, K. D.; Rudd, D. R. Chromatographia, 1995, 41, 325-31.

3864

Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

lower for the Q-PzI-treated capillary. Conversely, in the case of pepsin, ovalbumin, and myoglobin, better recoveries were experienced in the case of Q-PzI treatment. As we expected, the migration time does not correlate with pI because of weak interactions between the positively modified wall and the protein. Such behavior is more evident in the case of strongly acidic proteins presenting negative charges at pH 9 and migrating against the positively charged silica surface. Figure 2 shows the profiles of a 4 different proteins injected in a covalently coated vs a Q-PzI-treated capillary. In the case of protein separations performed in covalently coated capillaries, the running times are longer than those observed in Q-PzI-treated capillaries, and the peaks are broad, whereas the peak areas, as shown in Table 1, remain substantially the same, indicating the absence of, or at least the same, sample adsorption to the wall. To evaluate the stability of the wall treatment at different pH values, the capillary, after wall modification and EOF measurement, was equilibrated with buffers of different pHs and different ionic strengths, as adopted for protein separations in many reports. The following systems were used: 25 mM imino diacetic acid (IDA), pH 3.2; 25 mM acetate buffer, pHs 4 and 5; 25 mM MES, pH 6.0; 25 mM tetraborate buffer, pH 9.0. The EOF values where evaluated for each buffer, as shown in Figure 3. It is seen that a substantial reversed EOF flow persists at every pH, indicating that the modifier must be bound to the wall, since in no case it is present in the background electrolyte. If we consider the dynamic coating procedures, there is no doubt that oligoamines, especially those containing four or more nitrogens, such as spermine or TEPA (tetraethylene pentamine), are the best quenchers of any protein binding to the silica wall.28Their efficacy, though, must obey to the following rules: (a) the oligoamine has to be present at all times in the background electrolyte, since its omission will result in an immediate leaching of the modifier from the wall; (b) such oligoamines rapidly loose their effect as the buffer pH adopted in protein separation is raised above neutrality, as a result of progressive deprotonation of the amines in their backbone. Alternatively, quaternary ammonium diamino alkanes can be utilized, but the compound must be present in the background electrolyte.8 From this point of view, the quaternarized piperazine here described is quite exceptional on both accounts. First of all, we use it only to condition the wall and never add it to the background electrolyte used in analyte (protein or any other molecule) separation. Second, Q-PzI is just as effective at basic as well as neutral and acidic pH value. The utilization of a Q-PzI-treated capillary in a broad range of pHs differentiates it strongly from other reagents, such as MICROCOAT9, which can be used only below pH 7.0, thus limiting the application range. Such a unique behavior can only be explained by hypothesizing that Q-PzI is able to bind covalently to the silanols on the silica surface, although we failed up to the present to obtain direct evidence for this phenomenon. The following indirect evidence, however, supports such a hypothesis. Q-PzI reverses the EOF not only at alkaline, but also at acidic, pH values, in both cases by a preconditioning step, that is, by not being present during the actual separation run. Figure 4 shows the electropherogram of myoglobin analyzed at different pH values. The protein was dissolved in water to 1 (28) Verzola, B.; Gelfi, C.; Righetti, P. G. J. Chromatogr. A 2000, 868, 85-99.

Table 1. Elution Times and Recoveries of a Number of Proteins in Polymer-Coated vs Q-Pzl-Treated Capillaries covalently coated protein

pI

time min

pepsin β-lactoglobulin horse ferritin soybean trypsin inhibitor ovalbumin β-casein thrombin human albumin human transferrin superoxide dismutase bovine albumin myoglobin lactic dehydrogenase

3.02 4 4.1 4.2 4.63 4.8 4.8 4.8/5.6 5.6/6.2 5.86 6 7 8/6.65

21.2 24.53 34.61 23.54 35.87 24.72 40.45 24.78 45.41 54.65 23.64 25.3 unrevealed

a

Q-PzI-treated

area log

area/time log

time min

area log

area/time log

mMa

7 6.97 7.04 6.82 6.43 7.26 6.57 7.54 6.77 7.28 6.84 6.4 unrevealed

5.68 5.58 5.49 5.44 4.86 5.87 4.96 6.15 5.11 5.51 5.45 5 unrevealed

31.43 6.84 12.85 5.56 6 2.43 3.09 8.66 8.35 8.12 9.11 6.89 6.39

8.05 6.99 6.51 6.47 7.17 6.85 6.77 7.46 6.88 6.98 6.87 7.34 6.8

6.55 6.12 5.4 5.65 6.4 6.47 6.29 6.53 5.96 6.07 6 6.5 6.28

0.125 2 1 0.5 2 2 2 2 2 2 2 2 2

Molarity of Q-PzI used for conditioning the wall.

Figure 2. Separation of a number of protein markers, injected in a covalently coated (A) and in a Q-PzI-treated (B) capillary. Capillary length, 37 cm, 50 µm i.d. Separation conditions: run at 200 V/cm; sample injection by pressure for 2 s, 5 psi/s; detection at 214 nm. In both cases, the running buffer was 25 mM sodium tetraborate, pH 9.0. Note the much shorter running times and sharper zones in the 2 mM Q-PzI capillary.

mg/mL final concentration and injected by pressure as indicated previously. The capillary was first equilibrated with 25 mM iminodiacetic acid, pH 3.2, as a buffer, since the EOF value shown in Figure 3 indicated the presence of Q-PzI bound to the wall. The capillary was then treated as previously described with Q-PzI, and the protein was injected from the anode to the cathode according to protein mobility at a given pH. The electrophoretic

profile obtained at pH 3.2 in IDA buffer indicates the absence of protein absorption to the wall. When running the same protein in 25 mM acetate buffer, pH 4, the time at the detector appeared slightly longer as a result of the decrement on EOF at this pH, but still protein adsorption was not observed, and the area log was the same as in IDA buffer. When the same separation was repeated in 25 mM acetate buffer, pH 5, the mobility further Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

3865

Figure 3. Profiling the EOF along the pH scale of a Q-PzI-treated 37-cm-long capillary. EOF measurements were performed in buffers at the following pHs: 25 mM imino diacetic acid (IDA), pH 3.2; 25 mM acetate buffer, pHs 4 and 5; 25 mM MES, pH 6.0; and 25 mM tetraborate, pH 9.0. The EOF measurements were performed as described in the Experimental Section.

Figure 4. Electropherograms of myoglobin analyzed in the various buffers of Figure 3 at pHs 3.2, 4, 5, and 9.0, respectively. The protein was dissolved in water at 1 mg/mL final concentration and injected by pressure.

decreased as the net positive charge of the protein diminished in its approach to the pI value, but the area log was always the same, thus indicating the absence of adsorption. Finally, when myoglobin 3866

Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

was injected from the cathode to the anode according to its mobility at given pH and in the same direction as the EOF induced by Q-PzI treatment in a 25 mM tetraborate buffer, pH 8.5, the uppermost elution profile having a sharper peak was obtained. Figure 5A shows the separation of a mixture of 9 proteins with pI ranging from pH 3 to 10. It can be appreciated that the separation occurs according to the protein mobility and EOF. The first protein eluted is, thus, the moderately alkaline horse myoglobin acidic and basic band, then the acidic one, for which the mobility is influenced by the positive charge docked into the wall. The last group of protein is represented by the most alkaline ones, first the lentil lectins with pIs ranging from pH 8.15 to 8.65 just before the neutral marker then trypsinogen (pI 9.5), which overlaps with the neutral acrylamide marker. Peak identification was performed by direct spotting of the pure protein into the calibration pI mixture. The negative peak disturbance is associated only with the standard protein mix utilized and is not present in a blank injection as a separation artifact. Figure 5B shows the same protein mixture injected in a covalently coated capillary. In this case, we observe lower peak resolution. In principle, a mixture of proteins can be separated in free phase according to their titration curve, in absence of molecular sieving. Of course, the operative pH must be in the region of maximal mobility divergence. In the capillary, the EOF can tune the protein mobility, but we could not separate the protein mixture in a covalently coated capillary with negligible EOF value, not in the presence of quaternarized amino alkanes with a residual EOF (data not shown). This suggests that the alteration in EOF is important but not the only parameter involved in the modified-wall enhancement of resolution. We hypothesize the quaternary ammonium docked onto the wall can interact both with silanols25 and proteins via weak ion interactions, improving dramatically protein resolution without any additive. Last, we had addressed the question of a potential alkylation reaction of Q-PzI toward proteins. If, as we believe (see also the Discussion Section), this compound acts by covalently binding to the wall silanols. In principle, it could also react with proteins, notably with their free -SH groups. We have investigated this potential reaction, as shown in Figure 6A,B When Q-PzI was incubated overnight with superoxide dismutase (SOD) at a pH of either 8.5 or 10, two different patterns could be seen. At pH 8.5, essentially no reaction could be detected. The incubated SOD was identical to the control (in both panels, the m/z 15764 peak representing the complex of SOD with sinapinic acid). In contrast, at pH 10, a minute reaction channel could be discerned at m/z 15724, a Mr value of which suggests the presence of just a monocalculated species at a level of barely 10%. We conclude that, in both cases, the extent of reaction is so minute as to ward off any alarm concerning any potential alkylation of proteins during their separations in a capillary lumen, separations that typically occur not only at lower pH values, but for extremely short times as well. Why, then, should Q-PzI react with silanols on the silica wall and not with the analogous groups in proteins? The only explanation we have is that the iodine in the butyl chain, per se, is quite unreactive and becomes reactive only when it is in close proximity to the deprotonated silanols. This can only occur, apparently, on the silica wall, for which we have hypothesized a peculiar anchoring of Q-PzI: first, a docking driven by the positive

Figure 5. Separation of protein mixture with pI ranging from pH 3-10 (A) in Q-PzI-treated capillary 77 cm long, 50 µm i.d.; (B) covalently coated capillary, 77 cm long, 50 µm, i.d. Separation conditions: 250V/cm, sample injection by pressure for 5 s, running in tetraborate buffer, pH 9.0. (1) Horse myoglobin, (2) bovine carbonicanhydrase B, (3) human carbonicanhydrase B, (4) β-lactoglobulin A, (5) soybean trypsin inhibitor, (6) lentil-lectin pI 8.15 (7) lentil-lectin pI 8.55, (8) lentil-lectin pI 8.65, and (9) trypsinogen

Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

3867

Figure 6. MALDI TOF mass spectrometry results of 1 mg/mL SOD (a) and after overnight incubation with 10 mM Q-PzI at pHs 8.5 (b) and 10 (c).

charge on the quaternarized nitrogen, followed by additional hydrogen bonding of the tertiary nitrogen with nonionized silanols. Once Q-PzI is pasted to the wall, the extreme proximity of the reactants fires the reaction of the ω-iodine residue. CONCLUSIONS Reported here is a simple coating procedure for CZE of proteins based on the reaction of the new compound Q-PzI with the silica wall. The reaction occurs spontaneously at room temperature, at alkaline pH, at low Q-PzI concentration, and the binding of the quaternarized piperazine derivative to the wall seems to be stable at all pH intervals tested. The Q-PzI wall treatment allows the separation of a number of native proteins without adsorption onto the wall surface, as is typical of covalently affixed hydrophilic polymers Complex mixtures of proteins with different pI values, from pH 3.0 to 10, are separated, and the analysis is possible in buffers covering the pH range from pH 3.2 to 9.0, indicating the wide stability of this wall treatment. In addition, the absence of an additive in the background electrolyte

3868

Analytical Chemistry, Vol. 73, No. 16, August 15, 2001

will allow analyzing CZE separated protein by mass spectrometry with direct interface. The Q-PzI represents a novel approach for capillary modification and EOF flow control and offers significant utility for current applications requiring protein separations by CZE, such as protein-ligand binding studies or protein structurefunction studies. ACKNOWLEDGMENT Supported in part by a grant from MURST (Coordinated Project 40%, A New Approach in Proteome Analysis, 2000) and by ASI (Agenzia Spaziale Italiana, grant no. I/R/28/00) to P.G.R. We are much indebted to Dr. Mahmoud Hamdan, head of the mass spectrometry lab at GlaxoWellcome (Verona) for the MS analyses of Figure 6.

Received for review December 14, 2000. Accepted May 29, 2001. AC001478O