Analysis of Common Cold Virus (Human Rhinovirus Serotype 2) by

Anal. Chem. 1999, 71, 2028-2032

Analysis of Common Cold Virus (Human Rhinovirus Serotype 2) by Capillary Zone Electrophoresis: The Problem of Peak Identification Vadim M. Okun,† Bernhard Ronacher,‡ Dieter Blaas,‡ and Ernst Kenndler*,†

Institute of Analytical Chemistry, University of Vienna, Wa¨ hringerstrasse 38, A-1090 Vienna, Austria, and Institute of Biochemistry, University of Vienna, Dr. Bohr Gasse 9/3, A-1030 Vienna, Austria

Different preparations of human rhinovirus serotype 2 (HRV2), a common cold virus, were analyzed by capillary zone electrophoresis (CZE) in untreated fused-silica capillaries using borate buffer (100 mmol/L, pH 8.3) and sodium dodecyl sulfate (10 mmol/L) as additive to prevent wall adsorption. The electropherograms showed one major peak at 205- and 254-nm detection wavelengths. The identity of the peak as originating from native virus was confirmed by several indirect methods. Heating to 56 °C is known to lead to release of the genomic RNA from the viral capsid; this treatment resulted in the disappearance of the major peak and the emergence of a new predominant peak that was identified as RNA by enzymatic digestion. As expected, RNase treatment of the unheated sample remained without effect as the viral genome is inaccessible in the native viral shell. A monoclonal, virus-aggregating antibody was used for immunodepletion of native virus; again, the major peak disappeared upon removal of viral aggregates by centrifugation prior to CZE analysis. In combination, these results allowed for the unambiguous identification of the main peak as native HRV2 and of the minor peaks as contaminants present in various amounts in the different viral preparations. It is demonstrated that CZE allows for an extremely easy and rapid assessment of conformational state and purity of virions in a given viral preparation. Since the first report on the possible feasibility of protein separation using capillary electrophoresis (CE) by Hjerten,1 considerable efforts have been put forth to characterize biological macromolecules such as proteins and oligonucleotides by this methodology. Although early work demonstrated that proteins as ionogenic substances with different functionalities (hydrogenbonding regions, hydrophobic patches, etc.) interact with the capillary wall, thus deteriorating the separation, many authors proposed different approaches to overcome this problem (for a review see, e.g., ref 2). Since protein-capillary wall interactions could be avoided (at least in most cases) by one of numerous †

Institute of Analytical Chemistry. Institute of Biochemistry. (1) Hjerten, S. Chromatogr. Rev. 1967, 9, 122. (2) Regnier, F. E.; Lin, S. In High Performance Capillary Electrophoresis; Khaledi, M. G., Ed.; John Wiley: New York, 1998; Vol. 146, pp 683-727. ‡

2028 Analytical Chemistry, Vol. 71, No. 10, May 15, 1999

techniques, various classes of proteins were studied by CE including human serum proteins,3-5 highly basic proteins,6 immunoglobulins,7 and many others. Furthermore, various proteinprotein and protein-hapten interactions were characterized by the recently developed affinity mode of CE (for reviews, see refs 8 and 9), thus enabling the development of different CE-based enzyme immunoassays.10,11 Moreover, the method allowed evaluating affinity binding constants.12 A theoretical basis for the prediction of protein peak shapes was also established.13-17 Surprisingly, no attempts have been made to characterize by CE more complex biological macromolecules such as, for example, viruses (except in refs 18 and 19), where the principle applicability of CE was demonstrated for tobacco mosaic virus). This is despite the fact that CE can provide conditions offering to a considerable extent the in vivo situation where molecules are free in solution. Applying capillary isoelectric focusing (CIEF) we have recently pioneered this field20 for the determination of the pI value of human rhinovirus serotype 2 (HRV2). HRVs, the main causative agents of common cold infections, are icosaedral particles of ∼35 nm in diameter. They are assembled from 60 copies each of four different polypeptides (VP1-VP4) and an RNA genome resulting in a molecular mass of ∼8 × 106 Da. Despite (3) Hjerten, S.; Valtcheva, L.; Elenbring, K.; Eaker, D. J. Liq. Chromtogr. 1989, 12, 2471-2499. (4) Hjerten, S.; Zhu, M. D. Proteins Biol. Fluids 1985, 33, 537. (5) Hjerten, S. Electrophoresis 1990, 11, 665. (6) Huang, X.; Horvath, C. J. Chromatogr., A 1997, 788, 155-164. (7) Harrington, S. J.; Varro, R.; Li, T. M. J. Chromatogr. 1991, 559, 385-390. (8) Rippel, G.; Corstjens, H.; Billiet, H. A. H.; Frank, J. Electrophoresis 1997, 18, 2175-2183. (9) Heegaard, N. H. H.; Nilsson, S.; Guzman, N. A. J. Chromatogr., B 1998, 715, 29-54. (10) Schmalzing, D.; Nashabeh, W. Electrophoresis 1997, 18, 2184-2193. (11) Bao, J. J. J. Chromatogr., B 1997, 699, 463-480. (12) Rundlett, K. L.; Armstrong, D. W. Electrophoresis 1997, 18, 2194-2202. (13) Minarik, M.; Groiss, F.; Gas, B.; Blaas, D.; Kenndler, E. J. Chromatogr., A 1996, 738, 123-128. (14) Minarik, M.; Gas, B.; Rizzi, A.; Kenndler, E. J. Capillary Electrophor. 1995, 2, 89-96. (15) Stedry, M.; Gas, B.; Kenndler, E. Electrophoresis 1995, 16, 2027-2033. (16) Gas, B.; Stedry, M.; Rizzi, A.; Kenndler, E. Electrophoresis 1995, 16, 958967. (17) Radko, S. P.; Chrambach, A.; Weiss, G. H. J. Chromatogr., A 1998, 817, 253-262. (18) Hjerten, S.; Elenbring, K.; Kilar, F.; Liao, J.-L.; Chen, A. J.; Siebert, C. J.; Zhu, M.-D. J. Chromatogr. 1987, 403, 47-61. (19) Grossman, P. D.; Soane, D. S. Anal. Chem. 1990, 62, 1592-1596. (20) Schnabel, U.; Groiss, F.; Blaas, D.; Kenndler, E. Anal. Chem. 1996, 68, 4300-4303. 10.1021/ac981324x CCC: $18.00

© 1999 American Chemical Society Published on Web 04/09/1999

such extremely large molecular mass, it was possible to focus the viral particles and to mobilize them together with pI markers enabling evaluation of the isoelectric point. These highly ordered structures easily undergo conformational modifications in vitro upon exposure to low pH or elevated temperature. Even prolonged storage at temperatures above -80 °C might damage the virus. Native virions exhibit a sedimentation coefficient of 150S in sucrose density gradients whereas structural modification results in 135S particles. Upon release of the RNA, the remaining particle sediments with 80S. In vivo, these conformational changes take place during the infectious pathway and result in the transfer of the genomic RNA into the cytosol where replication takes place (for a recent review see, e.g., ref 21). Virus of high purity and homogeneity is required for structural analyses such as X-ray crystallography. However, rhinoviruses are difficult to prepare in large amounts, and a rapid and sensitive analysis method requiring only minute amounts of sample is thus highly desirable. In this work, we report on the systematic investigation of the applicability of CE for the analysis and characterization of virus preparations. We demonstrate that CE allows easy assessment of the purity of rhinoviral preparations within very short analysis times and evaluation of the conformational state of viral molecules. This first demonstration of the utility of CE for the analysis of large macromolecular complexes such as rhinoviruses paves the way for applications in quality control and in various other analyses of a large number of different viral systems. EXPERIMENTAL SECTION Apparatus. All experiments were performed on an automated HP3D capillary electrophoresis system (Hewlett-Packard, Waldbronn, Germany). Fused-silica capillaries were purchased from Polymicro Technologies (Phoenix, AZ) with 50-µm i.d. Total length was 50.0 cm and effective length 41.5 cm. The capillaries were packed in a standard HP cassette, thermostated at 20 °C during all experiments. In all experiments, the positive polarity mode with 25 kV was used (detector placed at the cathodic side of the capillary). The detector signals were recorded at 205 and 254 nm. In some cases, the fast spectral scanning mode was used to aid in peak identification. Data collection, storage, and analysis were performed using an HP Chemstation (Hewlett-Packard). Reagents. HRV2 as originally obtained from the American Type Culture Collection (ATCC) was propagated in Rhino-HeLa cells (Flow Laboratories). Virus was routinely purified from ∼5 L of suspension cultures by following standard protocols (e.g. ref 22). The background electrolyte (BGE) was 100 mmol/L boric acid adjusted to pH 8.3 with 1 mol/L NaOH and containing 10 mmol/L sodium dodecyl sulfate (SDS), that is, above its critical micellar concentration. Monoclonal antibody 3B10 raised against native HRV2 was purified from tissue culture supernatants.23 RNase 1-A was from Sigma. All other chemicals were purchased from E. Merck (Darmstadt, Germany) and were used without further purification. Reagent solutions were prepared in deionized water. Buffers were (21) Rueckert, R. R. In Virology; Fields, B. N., Knipe, D. M., Eds.; Raven Press: New York, 1996; pp 609-654. (22) Skern, T.; Sommergruber, W.; Blaas, D.; Pieler, C.; Kuechler, E. Virology 1984, 136, 125-132. (23) Hewat, E. A.; Marlovits, T. C.; Blaas, D. J. Virol. 1998, 72, 4396-4402.

filtered through a 0.45-µm cellulose nitrate membrane. All solutions were centrifuged for 2 min in an Eppendorf centrifuge at full speed prior to CE analysis. Procedures. New capillaries were conditioned by flushing with 100 mmol/L hydrochloric acid, followed by water, 1 mol/L NaOH, and water for 10 min each. Between the analyses, capillaries were washed with 100 mmol/L NaOH, water, and BGE for 2 min each by applying ∼950 mbar pressure. Injection was performed at 50 mbar pressure for 9 s. Samples were diluted with buffer solution which corresponds to 1:10 diluted BGE; unless stated otherwise, SDS was omitted from the samples. The sample buffer was supplemented with o-phthalic acid (20 µg/mL) as an internal standard (IS). Enzyme Treatment. Virus solution (20 µL at ∼0.25 mg/mL, native or heat-denatured, as estimated from Coomassie brilliant blue staining of the viral proteins after separation on a denaturing polyacrylamide gel) was incubated with 2 µL RNase 1-A (2 mg/ mL in 100 mmol/L borate buffer, pH 8.3) for 1 h at 37 °C prior to analysis. Immunodepletion. To a virus solution (40 µL at ∼0.1 mg/ mL in sample buffer with IS) ∼0.8 µg of 3B10 in the same buffer but without IS was added. The sample was incubated at room temperature for 1 h and centrifuged at 8000 rpm for 10 min, and the supernatant was analyzed by CE. To precipitate any virus left in the supernatant, more 3B10 (4 µg) was added and incubation, centrifugation, and CE analysis was repeated. RESULTS AND DISCUSSION Background Electrolyte Selection. Protein sorption on the inner capillary wall is the most recognized problem in CE analysis of large biomolecules. SDS is known to improve significantly the peak shape in protein CE analyses.24 It binds to hydrophobic regions of the protein resulting in a net negative charge of the formed complex. This leads to electrostatic repulsion from the negatively charged capillary wall and thus prevents sorption. Although SDS is recognized as a denaturing agent in biochemistry,25 recent work,26,27 including our own,28 has shown that even bioaffinity reactions can be performed in the presence of SDS provided its concentration is low. Furthermore, HRV2 does not become denatured in the presence of SDS but is rather stabilized against structural modifications induced by heat or low pH.29,30 As BGE, we used borate buffer at pH 8.3 with 10 mmol/L SDS as additive, which maintains EOF at a relatively high level and provides conditions that are well tolerated by the virus. CE Analyses of Viral Preparations. Virus used in the present work was propagated in HeLa cells and purified from 5-L suspension cultures either from the cell pellet or from the cell supernatant by poly(ethylene glycol) (PEG) precipitation followed by differential ultracentrifugation and rate zonal sucrose density gradient sedimentation according to established protocols.22 Briefly, virus pellets recovered from a high-speed centrifugation (24) Strege, M. A.; Lagu, A. L. Anal. Biochem. 1993, 210, 402-410. (25) Tanford, C. The hydrophobic effect: formation of micelles and biological membranes; Wiley: New York, 1980. (26) Steinmann, L.; Caslavska, J.; Thormann, W. Electrophoresis 1995, 16, 19121916. (27) Steinmann, L.; Thormann, W. Electrophoresis 1996, 17, 1348-1356. (28) Okun, V. M.; Bilitewski, U. Electrophoresis 1996, 17, 1627-1632. (29) Lonberg Holm, K.; Noble, H. J. J. Virol. 1973, 12, 819-826. (30) Lonberg Holm, K.; Yin, F. H. J. Virol. 1973, 12, 114-123.

Analytical Chemistry, Vol. 71, No. 10, May 15, 1999

2029

Figure 1. CE separation of three different batches of purified HRV2. Virus recovered from cell supernatant (A); virus recovered from cell pellet, lower fraction of sucrose gradient (B); and higher fraction of the sucrose gradient (C). Peaks marked with asterisks were present only in sample B (see also other figures). Virus concentration, ∼0.25 mg/mL in 1:10 diluted BGE without SDS added. Capillary, 50-µm i.d. × 50.0 cm, uncoated; effective length 41.5 cm. Buffer, 100 mmol/L borate, pH 8.3, containing 10 mmol/L SDS. Voltage, 25 kV. IS, o-phthalic acid 20 µg/mL.

of the cell lysate or of the resuspended PEG precipitate were treated with RNase and trypsin to digest contaminating material, adjusted to 1% sarcosyl, and layered on top of preformed 7.545% sucrose density gradients. After centrifugation for 4 h in a Beckman SW28 swing out rotor, the opalescent bands that formed approximately in the middle of the gradient were recovered by puncture with a needle. Sucrose was diluted with phosphatebuffered saline (PBS) and virus was again pelleted by a high-speed centrifugation. Viral samples were then suspended in ∼200 µL of PBS. In total, three different virus batches were analyzed. These included virus recovered from the cell supernatant (A) and virus extracted from the cell pellet. In the latter case, a lower (B) and an upper (C) fraction were taken from the sucrose density gradient. As seen in Figure 1, all electropherograms revealed one major peak with a migration time of ∼3.6 min which exhibited absorbance at 254 nm, indicating the presence of nucleic acid. We thus assumed this peak to correspond to native virus. Note that the minor peaks present in the three samples were different. Assuming that these peaks represent contaminants or degradation/denaturation products of native virus, the batch that had been obtained from the cell supernatant (Figure 1A) appears to be of highest quality (the major peak occupied ∼90% of the total peak area) as compared to the samples from the lower (Figure 1B) and the upper (Figure 1C) fraction taken from the sucrose density gradient. In the latter case, the major peak occupied only ∼15% of the total area; this batch was thus considered to be the less pure and was not used in further experiments. Peak Identification. For the confirmation of peak identity in CE of viruses, direct methods such as mass spectrometry cannot be applied. Therefore, identification is only possible by indirect means; e.g., appropriate precolumn reactions might be chosen such as to lead to the disappearance or reappearance of a particular peak. In the present work, the following methods were used for this purpose: (i) heat denaturation of native virus, (ii) enzymatic treatment of native and heat-denatured virus, (iii) biospecific interaction of virus with a monoclonal antibody. (i) Heat Denaturation. As pointed out above, native virus readily changes its conformation upon heating to 56 °C for 10 2030 Analytical Chemistry, Vol. 71, No. 10, May 15, 1999

Figure 2. CE separation of heat-denatured HRV2. Viral samples as used in Figure 1 were heated at 56 °C for 10 min prior to analysis. Other conditions were as in Figure 1, except SDS (10 mmol/L) was present in the sample during heating.

min or to 50 °C for 15 min29,30 to a structure sedimenting with 80S; concomitantly, the innermost capsid protein VP4 and the genomic RNA are lost. Two representative samples (as in Figure 1A and B) were thus heated to 56 °C for 10 min to completely convert 150S particles to 80S particles. In these cases, SDS had to be added to the sample (at 10 mmol/L concentration, the same as in the BGE), otherwise aggregation of the virus leads to noninterpretable results. CE analysis revealed that the peak at 3.6 min completely disappeared (Figure 2). Concomitantly, a peak migrating at 4.8 min and barely detectable in one of the untreated samples (see Figure 1A) was strongly increased in the heated samples; absorbance at 254 nm indicates that this material either contains or is nucleic acid. Because heating is known to cause the release of the genomic RNA, it is thus most probable that this peak represents viral RNA. Small amounts of RNA might also be present in the original samples of native virus arising either from viral degradation or from cellular components. Other peaks that appeared upon heating (Figure 2) might represent partially degraded virus and/or incompletely denatured viral proteins. These peaks were highly reproducible and their origin is now under investigation. Comparison between samples prior to heating (Figure 1B) and after heating (Figure 2B) shows that three peaks with migration times of 3.0, 3.5, and 3.9 min, which were present only in this particular preparation, remained unchanged (asterisks). This might be taken to indicate that they do not represent viral material but are most probably contaminants originating from cells or culture media. To confirm the tentative assignment of the peak with migration time 4.8 min to RNA, the samples were incubated with RNase prior to analysis. (ii) Enzymatic Treatment. As long as the genomic RNA remains buried inside the viral capsid it is inaccessible to the degradative action of RNase. In contrast, RNA released upon heating of the virus should be degraded by the enzyme. Samples were thus incubated with RNase prior and after heating and analyzed by CE. As seen in Figure 3 for the unheated virus samples, the position and area of the major peak remained unchanged in both samples of native virus upon RNase treatment (compare to Figure 1). This clearly indicates lack of degradation. However, the minor peak with absorption at 254 nm (migration

Figure 3. CE separation of RNase-treated native HRV2. Viral samples as in Figure 1 were treated by RNase prior to CE analysis. For details, see Experimental Section. Other conditions were as in Figure 1.

Figure 5. Pelleting of native HRV2 upon low-speed centrifugation of very small volumes. A 40-µL aliquot of a virus solution containing o-phthalic acid as IS was centrifuged at 13 000 rpm or 8000 rpm, respectively, for the times indicated, and residual virus in the supernatant was determined as the ratio of peak height of the viral peak and IS.

Figure 4. CE separation of RNase-treated heat-denatured HRV2. Viral samples were heated to 56 °C for 10 min as in Figure 2 and treated by RNase prior to analysis. For details, see Experimental Section. Other conditions as in Figure 1.

time 4.8 min) disappeared completely, strongly suggesting that it corresponds to free RNA (compare Figure 1A and Figure 3A). The minor peaks seen in Figure 1B (migration times 3.0, 3.6, and 4.1 min) did not change upon RNase incubation (Figure 3B). This again suggests that they represent impurities present only in this particular virus sample and cannot be RNA. Heated virus samples were then treated with RNase prior to CE analysis. As depicted in Figure 4, the major peak, tentatively assigned to viral RNA (Figure 2), disappeared in both cases. The emergence of several small peaks suggests the formation of RNA degradation products with specific absorption at 254 nm. This is an additional confirmation of the correct assignment of the peak migrating at 4.8 min to viral RNA. It should be pointed out that the peak of the enzyme could be easily identified by its absence of absorption at 254 nm. (iii) Biospecific Identification of the Major Peak. One of the most accurate identifications of a given component is by a change of the corresponding peak upon an appropriate biospecific reaction. From several specific monoclonal antibodies reacting with HRV2, mAB 3B10, which neutralizes viral infectivity by causing aggregation of multiple virions,23 was chosen. Aggregates can be easily removed from the sample by centrifugation. To establish centrifugation conditions for the removal of viral aggregates without pelleting single virions, a very small sample

Figure 6. CE analysis after immunodepletion of native HRV2 with monoclonal antibody 3B10. A mixture of HRV2 (as in Figure 1B) and excess amount of monoclonal antibody 3B10 was incubated, and precipitates were removed by centrifugation at 8000 rpm. For the details, see Experimental Section. Other conditions as in Figure 1.

volume (40 µL) was centrifuged for various time periods in an Eppendorf centrifuge. All experiments described below were performed with a sample of the viral batch shown in Figure 1B. The centrifuge was set at 13 000 rpm, and the supernatant was analyzed by CE after 0, 15, 25, and 35 min of centrifugation. The ratios between the peak heights of virus and internal standard were calculated from the resulting electropherograms. From the decrease of this value with time (Figure 5), it becomes clear that virus pellets almost quantitatively upon centrifugation for more than 25 min at 13 000 rpm. Most of the contaminants (except one with migration time 3.0 min) remained in solution, indicating that they had much lower sedimentation rates than the virus (data not shown). When the virus pellet was redissolved and analyzed by CE, the main peak again became noticeable at the same migration time (3.6 min) together with the contaminant mentioned above (migration time 3.0 min). Other contaminants were absent, indicating that they had remained in the supernatant (data not Analytical Chemistry, Vol. 71, No. 10, May 15, 1999

2031

shown). The experiment was then repeated at 10 000 rpm and yielded a similar result. Only at 8000 rpm did the virus remain in solution after centrifugation for up to 40 min. Therefore, separation of free virus from virus complexed to antibody was carried out by centrifugation for 10 min at 8000 rpm in the following experiments. Immunodeplition of Native Virus by Monoclonal Antibody 3B10. About 5 µg of HRV2 was sequentially incubated with increasing amounts of 3B10, for 1 h at room temperature in a final volume of 42 µL, and the eventually formed precipitate was removed by centrifugation at 8000 rpm as detailed above. From CE analysis of the supernatant, it could be seen that the peak of native virus decreased in parallel with increasing amounts of 3B10; finally it completely disappeared. Upon addition of excess 3B10, a peak corresponding to free antibody became visible (Figure 6). All other peaks of viral contaminants did not change (compare to Figure 1B). Bearing in mind the high specificity of the monoclonal antibody used we conclude that the major peak with migration time 3.60-3.67 min indeed represents the native virus. CONCLUSIONS It was demonstrated that extremely large biomolecular complexes such as viruses can be successfully analyzed by CE. For obvious reasons, direct identification methods are not applicable

2032 Analytical Chemistry, Vol. 71, No. 10, May 15, 1999

for viral samples. Therefore, the heat lability of HRV2 and the specificity of a monoclonal antibody for native virus were used to confirm the identity of the viral peak. The peak appearing upon heat treatment was identified as free viral RNA by RNase digestion. CE proved to be extremely useful for the rapid and reliable analysis of viral samples and presents itself as the method of choice for quality control of HRV preparations and possibly for other nonenveloped viruses. Furthermore, it offers great potential for the analysis of conformational states of the viral capsid. The present work also paves the way for the analysis and characterization of complexes between virus and antibodies or receptor fragments. ACKNOWLEDGMENT This work was supported by Grant P12269-MOB from the Austrian Science Foundation and a Lise-Meitner Postdoctoral fellowship (Project M448-CHE) to V.M.O. We thank I. Go¨sler for preparing HRV2.

Received for review December 1, 1998. Accepted February 20, 1999. AC981324X