Anal. Chem. 1996, 68, 4300-4303
Determination of the pI of Human Rhinovirus Serotype 2 by Capillary Isoelectric Focusing Ursula Schnabel,† Franz Groiss,† Dieter Blaas,‡ and Ernst Kenndler*,†
Institute of Analytical Chemistry, University of Vienna, Wa¨ hringerstrasse 38, A-1090 Vienna, Austria, and Institute of Biochemistry, Medical Faculty, University of Vienna, Dr. Bohrg. 9/3, A-1030 Vienna, Austria
Capillary isoelectric focusing was applied to determine the pI value of human rhinovirus serotype 2 (HRV 2), a picornavirus of about 8 500 000 Da in size. Using fused silica capillaries dynamically coated with hydroxypropylmethyl cellulose (added at 0.08% to the catholyte), the virus zone failed to reach the steady state position in the pH gradient within times usually employed in focusing experiments, as the electroosmotic flow (EOF) pushed the analyte zone past the detector. Therefore, the residence time of the zones in the separation capillary was extended by applying hydrodynamic pressure at the detector side during focusing, thus pneumatically counteracting the EOF. After completion of focusing, the zones were mobilized by pressure maintaining the high voltage. For calibration of the pH gradient, low molecular mass pI marker substances were employed. Using the relation between the apparent pI value of the virus and the focusing time under counter pressure, the actual pI of HRV2 was determined as 6.8 by extrapolating to infinite time. The pI value is one of the best established parameters in the field of biochemistry to characterize and identify zwitterionic entities. This applies not only to proteins but also to such complex macromolecular assemblies as viruses and virus-antibody complexes.1-5 Classically, pI values are determined by isoelectric focusing in polyacrylamide gels. A serious drawback of this method lays in the comparable small pore size of the gel matrix, which prevents high molecular weight macromolecules from attaining their equilibrium position. Therefore, determination of the pI of such macromolecules is usually carried out in free solution using sucrose density gradients for stabilization of the pH gradient; however, the commercial focusing devices require large volumes of expensive ampholytes and considerable amounts of sample, which are often not available. Capillary isoelectric focusing (CIEF) in free solution offers many advantages over the classical gel format. One of the most important aspects is the absence of the rigid porous gel structure, which normally restricts the migration of large entities. Additionally, on-line monitoring is possible, thus avoiding time-consuming †
Institute of Analytical Chemistry. Institute of Biochemistry. (1) Brioen, P.; Thomas, A. A.; Boeye, A. J. Gen. Virol. 1985, 66, 609. (2) Colonno, R. J.; Callahan, P. L.; Leippe, D. M.; Rueckert, R. R.; Tomassini, J. E. J. Virol. 1989, 63, 36. (3) Emini, E. A.; Kao, S. Y.; Lewis, A. J.; Crainic, R.; Wimmer, E. J. Virol. 1983, 46, 466. (4) Emini, E. A.; Ostapchuk, P.; Wimmer, E. J. Virol. 1983, 48, 547. (5) Vrijsen, R.; Rombaut, B.; Boeye, A. J. Gen. Virol. 1983, 64, 2339. ‡
4300 Analytical Chemistry, Vol. 68, No. 23, December 1, 1996
detection procedures; further, full automation and quantitation are easy to perform. One of the fields of interest of one author of this article (D.B.) is the mechanism underlying antibody-mediated neutralization of human rhinoviruses (HRVs), the main causative agent of the common cold.6 These viruses belong to the family Picornaviridae, and are icosahedral particles, (8.3-8.7) × 106 Da in size; although classified as “small viruses”, they are huge when compared to usual biomolecules. The capsid is composed of 60 copies each of four different proteins which enclose the genomic positive strand RNA genome. HRVs are not stable at pH < 5 or elevated temperature; they are rapidly converted to another conformation with lower sedimentation value, modified antigenicity, and a lower isoelectric point.7 Similarly to these structural changes, attachment of neutralizing antibodies to the picornaviral particles might also lead to a change in isolectric point.1-5 In the course of experiments aimed to identify receptor binding sites on the surface of a minor receptor group rhinovirus (HRV2), we intended to monitor changes in the isoelectric point upon binding of particular antibodies and soluble viral receptor fragments. To downscale the amount of sample required for this task, to reduce the manipulation time required, and to increase the resolution of isoelectric focusing, experiments were conducted to investigate the applicability of CIEF for these analyses. As marker substances for calibration of the pH gradient we applied low molecular weight synthetic compounds introduced by Slais et al.,8,9 which are much better suited for CIEF than the normally used reference proteins. Compared to these synthetic pI markers, the relative molecular mass of the virus is extremely high. We thus had to take into account significant differences between their focusing properties, in that the dyes would most probably reach their equilibrium pI position in the capillary within a much shorter time than the virus. Therefore, it was anticipated that focusing time would be of considerable influence on the apparent pI of the large virus particle. We here report on CIEF of HRV2 in its native and its “denatured” conformations. We show that the virus asymptotically approaches the position corresponding to its isoelectric point; the value of the real pI can be exactly determined by extrapolation from the values obtained at various time points. A major problem related to extended focusing times is the electroosmotic flow (EOF). Chemically bonded coatings were not (6) Stott, E. J.; Killington, R. A. Annu. Rev. Microbiol. 1972, 26, 503. (7) Lonberg-Holm, K.; Noble-Harvey, J. J. Virol. 1973, 12, 819. (8) Slais, K.; Friedl, Z. J. Chromatogr. A 1994, 661, 249. (9) Caslavska, J.; Molteni, S.; Chmelik, J.; Slais, K.; Matulik, F.; Thormann, W. J. Chromatogr. A 1994, 680, 549. S0003-2700(96)00378-2 CCC: $12.00
© 1996 American Chemical Society
used here to supress the EOF because they are unstable at the pH conditions used (the catholyte filling the main part of the capillary has pH >12). EOF still persists in capillaries dynamically coated with various polymers.10-16 This phenomenon can cause migration of the analyte through the detector before it has attained its equilibrium position in the pH gradient. To overcome this problem, we applied a hydrostatic pressure directed against the EOF during focusing.17 Due to the variability of the EOF, the particular value of the pressure necessary to counterbalance the EOF exactly cannot be easily predicted; we therefore used a pragmatic approach by measuring the pI as a function of time until a plateau of the pI vs time curve was attained. EXPERIMENTAL SECTION Reagents. The pH gradient was established in 4% w/v Servalyte (pH 3-10) purchased from Serva. Chemicals used for the preparation of the electrolyte solutions were orthophosphoric acid (reagent grade, E. Merck, Darmstadt, Germany), sodium hydroxide (100 mmol/L, reagent grade, Fluka, Buchs, Switzerland), hydroxypropylmethyl cellulose (HPMC, Sigma, St. Louis, MO; viscosity of 2% solution at 25 °C, 4000 cP). To prepare the catholyte, the latter two solutions were mixed in a ratio of one volume of 100 mmol/L NaOH and four volumes of 0.1% w/v HPMC solution (NaOH/HPMC). Doubly distilled water was used throughout. As pI makers, low molecular mass nitrophenolic amine dyes were used; they were a kind gift from K. Slais, Czech Academy of Science, Brno, Czech Republic. These dyes have absorption maxima between 400 and 420 nm and were characterized in detail in refs 8 and 9. They were dissolved in the carrier ampholyte solution at a concentration of between 3 and 40 µg/mL. HRV2 (5 mg/mL in phosphate-buffered saline) was prepared and purified according to standard protocols.18 Apparatus. All experiments were performed using an HP3D capillary electrophoresis system (Hewlett-Packard, Waldbronn, Germany), consisting of a CE unit with built-in diode array detector (DAD) and an HP3D CE ChemStation for system control, data collection, and data analysis. The instrument was equipped with uncoated fused-silica capillaries from Polymicro Technologies Inc. (Bloomfield, NJ) of 75 µm i.d., a total length of 50.5 cm, and an effective length (distance between inlet and detector) of 42.5 cm. The capillaries were dynamically coated by addition of 0.08% HPMC to the catholyte. During the entire experiments, the capillary was thermostated at 25.0 °C. Method Parameters. Preceding injection, conditioning was done as follows: the capillary was flushed (at 950 mbar) with 100 mmol/L NaOH (5 min), H2O (2 min), and NaOH/HPMC mixture (7 min). The detector was placed on the cathodic side, with NaOH/HPMC as catholyte. The ampholyte/dye solution and the sample were inserted by pressure (50 mbar). To keep the virus (10) Hjerten, S.; Elenbring, K.; Kilar, F.; Liao, J.-L.; Chen, A. J.; Siebert, C. J.; Zhu, M.-D. J. Chromatogr. 1987, 403, 47. (11) Mazzeo, J. R.; Krull, I. S. Anal. Chem. 1991, 63, 2852. (12) Mazzeo, J. R.; Krull, I. S. J. Chromatogr. 1992, 606, 291. (13) Thormann, W.; Caslavska, J.; Molteni, S.; Chmelik, J. J. Chromatogr. 1992, 589, 321. (14) Chmelik, J.; Thormann, W. J. Chromatogr. 1993, 632, 229. (15) Molteni, S.; Thormann, W. J. Chromatogr. 1993, 638, 187. (16) Molteni, S.; Frischknecht, H.; Thormann, W. Electrophoresis 1994, 15, 22. (17) Nolan, J. A. Application Information A-1750, Beckman Instruments, Fullerton, CA, 1993. (18) Skern, T.; Sommergruber, W.; Blaas, D.; Pieler, C.; Kuechler, E. Virology 1984, 136, 125.
solution clean, it was not mixed with the carrier ampholyte but was injected between two ampholyte/dye plugs (70 s ampholyte/ dye, corresponding to about 9 cm plug length; 8 s virus sample, corresponding to 1 cm plug length; 70 s ampholyte/dye). As anolyte, phosphoric acid (10 mmol/L) was used. Focusing was at 20 kV; to counteract the electroosmotic flow, pressure of 11 mbar was applied from the cathodic side. After completion of focusing, mobilization past the detector was carried out with 30 mbar while the high voltage was maintained.19 Marker dyes were monitored at 410 nm, the virus at 260 nm. RESULTS AND DISCUSSION Usually, the pI of analytes is determined by mixing them with the ampholyte solution before application of the electric field. As focusing proceeds, the peaks of the analytes become increasingly sharper, which is indicative for the establishment of the steady state. However, as human rhinovirus is unstable at pH < 5, it was not mixed with the ampholytes in order to avoid exposure to zones of low pH for extended time periods; contact with low pH would be inevitable for those virus particles close to the anode which have to migrate in the pH gradient toward their steady state position. Virus was thus inserted into the capillary not too far away from the expected equilibrium position in the pH gradient in the capillary (that of moderate final pH), which can be calculated from the total length of the adjusted ampholyte zone. Under these conditions, the establishment of the steady state is not exclusively reflected in the sharpness of the peak, but rather in the constancy of its position in the gradient with focusing time. When virus was subjected to focusing together with the marker dyes for different time periods (Figure 1), it could clearly be seen that, after 10 min (a time interval which was found sufficient to reach the equilibrium for all proteins we have investigated previously20,21 ), it was far from its appropriate position in the pH gradient. This is evident from comparison with its position upon extension of focusing time. Whereas at 10 min the virus was found in the weakly alkaline region between markers 2 and 3, it had reached a position between markers 3 and 4 upon 20 min. Between 40 and 60 min, it somewhat moved toward marker 4, to be finally detected closer to marker 4 than to marker 3 upon completion of the experiment. At that position, the virus forms a very sharp peak, also indicating that steady state conditions have been attained. The markers were found to elute at essentially the same positions, regardless of the focusing time employed (compare panels a-d in Figure 1). It should be mentioned that, in the run shown in Figure 1b, the marker substance 5 is not visible in the electropherogram because it was moved out of the anodic end of the capillary due to the counter flow, which overcompensated here the seemingly lower EOF as compared to the other runs. For calibration of the pH gradient, a graph of the known pI values of the respective markers vs residence times upon mobilization was constructed. A linear relation between pI and residence time of the various pI markers should be obtained. As can be seen in Figure 2, the markers eluted at almost identical positions upon various focusing times under counter pressure, whereas the peak corresponding to HRV2 appeared to be shifted (19) Huang, T. L.; Shieh, P. C. H.; Cooke, N. Chromatographia 1994, 39, 543. (20) Groiss, F. Diploma work, University of Vienna, 1995. (21) Minarik, M.; Groiss, F.; Gas, B.; Blaas, D.; Kenndler, E. J. Chromatogr. A 1996, 738, 123.
Analytical Chemistry, Vol. 68, No. 23, December 1, 1996
4301
Figure 3. UV spectra of the virus sample (a) from the CIEF experiments (peak V in Figure 1d) and (b) as measured by CZE.
Figure 1. Electropherograms of HRV2 recorded after different focusing times: (a) 10, (b) 20, (c) 40, and (d) 50 min. The diode array detector was placed at the cathodic side of the capillary. Tracings are shown at 410 and 260 nm for a 10.0 min interval. Migration of the zones due to electroosmosis was counteracted by hydrodynamic pressure from the cathodic side (11 mbar). After completion of focusing, the zones are mobilized by pressure (30 mbar) applied from the anodic side, with voltage maintained. Ampholyte: pH 3-10, 4% w/v. Catholyte: sodium hydroxide (20 mmol/L), hydroxypropylmethyl cellulose (0.08% w/v) added as dynamic coating. Anolyte: phosphoric acid (10 mmol/l). pI values of the reference dyes: 1, 10.4; 2, 8.6; 3, 7.5; 4, 6.6; and 5, 5.3. The virus peak is indicated by V.
Figure 2. Residence times vs pI of the five reference dyes for different focusing times between 10 and 60 min. Data are from focusing experiments as shown in Figure 1. The pI values of the reference dyes are given in the legend to Figure 1. Values of the linear correlation coefficient are given in parentheses.
toward lower pH values upon extension of time (see Figure 1). Based on the known pI values8,9 of the marker substances, the linearity of the pH gradient was found to be reasonably good for short focusing times (up to 40 min), indicating that the marker substances reach their steady state position quickly; increasing the focusing time to 60 min led to a slight deviation from linearity, with correlation coefficients being still within an acceptable range (0.989 and 0.971 for 50 and 60 min, respectively). It should be mentioned here that the inconstancy of the EOF has a marginal influence on the accuracy of the determination of the pI value of the analytes, because the pI markers act as internal standards. 4302 Analytical Chemistry, Vol. 68, No. 23, December 1, 1996
The virus preparations used in this investigation were about 95% pure, as judged from SDS-polyacrylamide gel electrophoresis of denatured viral capsid proteins (data not shown). Nevertheless, a number of facts ascertain unequivocal identification of the virus in the CIEF experiments. Contaminants might be contained in the ampholytes, but they might also come from the tissue culture medium used to prepare the virus. Bovine serum albumin, a major component of the medium, has a pI of 4.9, which is clearly different from the one found for HRV2 (see Figure 1). Phenol red, an indicator added to monitor the pH of the tissue culture medium, is a strong anion and is expected to migrate out of the anodic side of the capillary. Finally, the spectra of HRV2 used as sample (measured by CZE in phosphate buffer, 10 mmol/L, pH 7.5) and of the peak tentatively identified as HRV2 in the electropherogram (V in Figure 1d) were found to be virtually identical (Figure 3). We are thus confident that the peaks labeled V in Figure 1 indeed represent isolelectrically focused HRV2. As mentioned before, exposure of human rhinoviruses to acidic medium or elevated temperature leads to conversion of the native structure to conformations with lower pI. Therefore, additional evidence for the identity of peak V in Figure 1 with HRV2 was obtained by focusing sample from virus which had been exposed to 56 °C for 30 min. Under these conditions, we consistently observed a decrease of the native virus peak, concomitantly with the appearance of a new peak focusing at the rear side of marker 5 (peak D in Figure 4). This peak clearly represents altered viral particles with a pI of about 4.4. In some experiments, the peak corresponding to native virus was not observed but was rather replaced by the peak representing denatured material (see above). This is clearly due to the well-known instability of human rhinoviruses. We currently do not know which conditions during the focusing procedure led to this apparent denaturation of the virus, with the peak found to be shifted from pH 6.8 to around 4.4. That the focused material was of viral origin was proven by injection of HRV2 metabolically labeled with 35S-methionine and scintillation counting of fractions collected after mobilization (not shown). To determine the exact pI of the virus, it is necessary to extrapolate to infinite focusing times. The relationship between apparent pI and time becomes obvious in the graphic representation (Figure 5). Whereas the virus exhibits an apparent pI of
Figure 4. CIEF of the “denatured” form of HRV2 (indicated by D). The viral sample was exposed to elevated temperature (56 °C) for 30 min. CIEF conditions were as in Figure 1.
about 8 after 10 min, it clearly shifts toward lower values upon increasing the focusing time. Extension of the time beyond 50 min does not lead to any substantial change in the apparent pI, indicative of the equilibrium being attained at a pI of 6.8. CONCLUSION We here demonstrate that CIEF in free solution enables the determination of the pI of viruses within short time, at extremely low sample demand, with on-line detection and with the potential of full automatization. Extended focusing times, as needed for the accurate determination of the pI values of large molecules or molecular assemblies such as viruses, can be realized in CIEF by the application of pressure at the cathodic side, which counteracts the electroosmotic flow. However, when dealing with viruses known to be unstable in nonphysiological solutions, the results have to be interpreted with caution: for reasons not fully understood at the present time, denaturation might occur under the conditions prevailing in the capillary under focusing conditions. Experiments using viruses with higher stability such as the closely related polioviruses will
Figure 5. Relation between the apparent pI of HRV2 and focusing time with applied counter pressure obtained by focusing similar to that shown in Figure 1.
reveal whether this effect is limited to rhinoviruses or is due to so far unknown interactions between large particles and the chemical environment of the separation system. ACKNOWLEDGMENT We thank L. Frasel and I. Go¨ssler for preparing the virus. This work was supported by the Austrian Science Foundation, Grant P-9999-MOB.
Received for review April 18, 1996. Accepted September 4, 1996.X AC9603789
X
Abstract published in Advance ACS Abstracts, October 15, 1996.
Analytical Chemistry, Vol. 68, No. 23, December 1, 1996
4303
