Journal of Chromatography,
435 (1988) 149-158
Elsevier Science Publishers B.V., Amsterdam -
Printed in The Netherlands
CHROM. 20 091
HIGH-PERFORMANCE LIQUID CHROMATOGRAPHIC SEPARATION AND IMMUNOLOGICAL CHARACTERIZATION OF SOLUBLE BOVINE VIRAL DIARRHEA VIRUS ANTIGEN
JOY G. MOHANTY* and YOUSSEF ELAZHARY Virology Section, Department of Pathology Hyacinthe. Quebec J2S 7C6 (Canada)
and Microbiology.
University of Montreal.
3200 Sicotte St., St.
(Received September 17th, 1987)
SUMMARY
Sodium dodecyl sulphate-polyacrylamide gel electrophoresis and Western blot analysis of the protein extracts from bovine viral diarrhea virus (BVDV, Singer strain) infected primary calf testicle cells (soluble antigen) showed the presence of four virus specific polypeptides of 105, 90, 84 and 67 kiloDaltons (kD) the 84-kD being the most abundant. Anion-exchange high-performance liquid chromatography (HPLC) of soluble antigen separated the virus specific polypeptides in individual peaks while the gel permeation HPLC collected all of them in a single protein aggregate peak of 290 kD. Except for the 84-kD polypeptide peak in anion-exchange HPLC, all peak fractions were found to be heterogeneous in nature having more than one polypeptide. Analysis of the antisera raised against the peaks having antigen activity showed that antisera against the 84-kD polypeptide peak did not neutralise BVDV while those against the fractions containing the 90- and 105-kD polypeptides neutralised the virus.
INTRODUCTION
Bovine viral diarrhea virus (BVDV), a single stranded RNA virus, has been identified as the causative agent of virus induced diarrhea mucosal disease in cattlelp3. BVDV is an enveloped virus which upon infection expresses virus specific proteins on an infected cell surface4*5. Soluble protein extract from BVDV infected cells has been shown to be an effective immunogen since it has been reported to protect against BVDV infection upon immunization in cattle6. Thus BVDV infected cell extract has been termed soluble antigen4g6T7. An attempt to separate individual BVDV (cytopathic Singer and non-cytopathic NY-l strains) antigenic proteins from the soluble antigen by high-performance liquid gel permeation chromatography (HPLGPC) has been reported by Coria et al.‘. HPLGPC of soluble antigen prepared from both the cytopathic and non-cytopathic BVDV infected cells showed similar separation profiles each consisting of seven peaks. Rabbit antisera against two peaks with molecular weights 240 and 140 0021-9673/88/$03.50
0
1988 Elsevier Science Publishers B.V.
J. G. MOHANTY,
150
Y. ELAZHARY
kiloDaltons (kD) could neutralize BVDV. However, the analysis of the homogeneity and nature of the proteins in these two peak fractions was not reported. A second attempt to isolate BVDV antigen proteins from the lectin purified BVDV (strain Ug-59) has been reported by KarsnHs et al.* in which the detergent solubilised virus was subjected to anion-exchange fast protein liquid chromatography (FPLC). Karsnas et al.* used a low-UV absorbing non-ionic detergent (Berol 172) to solubilise BVDV virions and to improve the subsequent chromatographic step. Again no further analysis of the peak fractions was described. In this report we demonstrate for the first time the heterogeneous nature of the HPLC-purified immunoreactive peaks following anion-exchange HPLC as well as HPLGPC of the soluble antigen. We also report the characterization of some of the proteins in these peak fractions according to the viral neutralization properties of their respective antibodies. To reduce the hydrophobic interactions of the membrane proteins during HPLC separation, a non-ionic detergent Berol 185 is used in the mobile phase. This detergent is similar to Berol 172 and has been used earlier in the separation of platelet membrane glycoproteins by FPLC9. EXPERIMENTAL
Cells and virus
BVDV of Singer strainlo was used in this study. Primary cells were prepared from calf testis and cultured in Eagles minimal essential medium (EMEM) containing 10% fetal bovine serum (free from mycoplasma and BVDV) and antibiotics. Chemicals
All chemicals unless otherwise stated were purchased from Sigma (St. Louis, MO, U.S.A.). The detergent Berol 185 was a kind gift from Berolkemi Inc., (Mississauga, Canada). Protein molecular weight standards used to calibrate the HPLGPC column were from Biihringer Mannheim (Canada). Preparation of soluble antigen
Soluble antigen was prepared as described earlier by Coria et a1.7. In brief, stock BVDV (Singer strain) prepared in primary calf testicle cells was used to infect the same cells for the preparation of soluble BVDV antigen. Infected cells [90-100% cytopathic effect (CPE)] were separated from the culture media by low-speed centrifugation (ca. 1000 g). The cells were lysed by sonication in hypotonic medium containing 1% Triton X-100 (TXlOO) and 1 mM phenylmethylsulfonyl fluoride (PMSF) at 4°C. The insoluble cell debris was pelleted at ca. 100000 g and the supernatant used as the source of crude soluble BVDV antigen. Detergent was removed from the supernatant by the solid phase extraction with SM-2 Biobeads (Bio-Rad Labs.). Following incubation with Biobeads overnight at 4°C the beads were removed by filtration through a 0.45~pm filter. The filtrate was further clarified by centrifugation at 100000 g and the supernatant (soluble antigen) stored at -70°C until use. Soluble proteins from non-infected calf testicle cells were also prepared in a similar manner and used as control. Anion-exchange
HPLC and HPLGPC
Anion-exchange chromatography
of soluble antigen
of soluble antigen was performed on DEAE
HPLC AND IMMUNOLOGY
OF VIRUS ANTIGENS
151
TSK-SPW (7.5 cm x 0.75 cm I.D.) column connected to a Waters HPLC system. The HPLC system consisted of two pumps (Model 510) a 254 nm detector (Model 440) with a flow cell, a Digital Professional 350 computer and a Digital Model LA50 printer. For anion-exchange HPLC of soluble antigen, proteins in buffer containing 0.02 M ethanolamine and 0.5% Berol 185, pH 9.2 (buffer A) was injected on to the column and the separation was performed by using a computer program to form a gradient of buffer A to buffer B (buffer A with 0.5 A4 sodium chloride). For HPLGPC, a TSK G3000SW (60 cm x 0.75 cm I.D.) column pre-equilibriated in 0.1 M dipotassium hydrogenphosphate, 0.2 M sodium chloride and 0.5% Berol 185, pH 7.0 (mobile phase) was used. The separation range of this column is 30 to 300 kD. Proteins in 100 ~1 of the mobile phase were injected on to the column and chromatography was run for 80 min at a flow-rate of 0.5 ml/min. For both the anion-exchange HPLC and HPLGPC, protein peaks were monitored at 254 nm and the chromatograms stored in the computer memory. The chromatograms were then reprocessed in the computer and printed. The peak fractions were collected manually and lyophilized to dryness prior to further analysis. Removal of salt and detergent from HPLC,fractions It was necessary to remove salt and detergent present in the HPLC fractions
before they could be analysed by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE). Each of the lyophilised HPLC fractions were reconstituted in cu. 1 ml of water and then dialysed against 500 ml of 0.01 M Tris-HCl, pH 7.5, at 4°C overnight with three changes of buffer. The dialysed fractions were again lyophilised and treated with cu. 1 ml of cold (- 20°C) acetone for 15 min followed by centrifugation (13 000 g, 1 min) to remove the detergent. The protein precipitates thus obtained were dissolved in SDS-PAGE sample buffer and electrophoresed. Measurement
of protein
content
in HPLC fractions
Protein content of the HPLC fractions were measured by using bicinchoninic acid (BCA) protein assay reagent from Pierce (Rockford, IL, U.S.A.) along with bovine serum albumin as standard according to the Company protocol. SDS-PAGE
and Western
blot analysis
Protein samples containing no detergent (soluble antigen, soluble protein from non-infected cells or the HPLC fractions with salt and detergent removed) were lyophilized to dryness, reconstituted in 50 ~1 of sample buffer (2% SDS, 10% glycerol and 0.1 M Tris-HCl, pH 6.8), heated on a boiling water bath for 5 min and then electrophoresed as according to Laemmli l l . For the purpose of Western blot analysis, each sample was electrophoresed in duplicate sets of wells and the gel was cut into two halves. One half of the gel was stained with Coomassie Blue to visualize total protein bands and the other half of the gel was subjected to electroblot on to a sheet of nitrocellulose paperi2. Following electrotransfer, the nitrocellulose sheet was treated with 1% gelatin in phosphate buffered saline (PBS) to block the sites unoccupied by polypeptide bands from SDS-PAGE gel. The BVDV specific polypeptide bands on the nitrocellulose sheet were then visualised by sequentially reacting it with a polyclonal bovine anti-BVD serum, rabbit anti-bovine IgG peroxidase conjugate and c1chloronanhthol’2.
IS2
J. G. MOHANTY.
Y. ELAZHARY
Preparution of’unti-BVD and antisera against HPLC,fructions Calves (cu. 400 Ibs) free from BVD virus and antibodies to the virus were injected intramuscularly twice at three-week intervals with inactivated BVDV in complete adjuvant. The virus was chemically inactivated by a similar procedure as reported earlier for foot and mouth disease virus 13. The titre of the serum tested by indirect immunofluorescence was 1:5120 and by seroneutralisation with 100 plague forming units (PFU) of Singer strain BVDV was at least 1:1280. Antisera against anion-exchange HPLC fractions were raised in guineapigs whereas those against HPLGPC fractions were raised in calves as described above. For this purpose, antigens were injected three times with two-week intervals in complete adjuvant. For comparison purposes antisera against the soluble antigen was also prepared in guineapigs as well as calves. Seroneutralisation titre of these sera were determined as above. RESULTS
Ident$cation qf B VD V spec$c polypeptides in soluble antigen Prior to the separation of BVDV specific polypeptides in soluble antigen, their nature and size were analysed by SDS-PAGE followed by Western blot analysis. Total proteins in soluble antigen and non-infected calf testicle cell extract (control sample) were electrophoresed on a 5510”/0 acrylamide gradient gel in two sets of
Fig. I. SDS-PAGE (lanes A and B, Coomassie Blue stained) and U’estern blot (lanes C and D) of noninfected calf testicle cell extract (lanes A and C) and soluble antigen (lanes B and D). Bands a, b, c and d in lane D correspond to 105, 90, X4 and 67 kD respectively. Molecular weight standards used were rabbit IgG (I 50 kD), bovine serum albumin (67 kD), chicken egg albumin (43 kD) and carbonic anhydrase (30 kD). K = kiloDaltons (kD).
HPLC
AND
IMMUNOLOGY
OF VIRUS
153
ANTIGENS
lanes. Following electrophoresis, proteins in one set of lanes in one half of the gel were visualised by staining them with Coomassie Blue while the virus specific polypeptides were detected by immunoblotting the other half of the gel on to a nitrocellulose sheet. As shown in Fig. 1, anti-BVD serum recognised four specific polypeptides a, b, c and d in the soluble antigen. Polypeptide band c showed a sharp and prominent band corresponding to a molecular weight of 84 kD. By carefully aligning the Coomassie Blue stained gel and the corresponding immunoblot, a faint but sharp band m (Fig. 1) in the Coomassie gel coincided with the band c in the immunoblot. Bands a and b showed very broad bands corresponding to 105 and 90 kD respectively whereas band d corresponded to 67 kD. Anion-exchange HPLC of soluble antigen Based upon the results described above, separation of polypeptides a, b, c and d in soluble antigen was attempted by anion-exchange HPLC on a column having DEAE functional group. Soluble antigen proteins in a buffer containing 0.02 M ethanolamine and 0.5% Berol 185, pH 9.2 (buffer A) was injected on to the column and the bound proteins eluted by the gradient O-O.5 A4 sodium chloride in the same buffer at a flow-rate of 1 ml/min. Buffer B consisted of buffer A and 0.5 M sodium chloride. Following injection, an initial 10 min of isocratic run in buffer A was maintained to detect any unbound material from the column. This was followed by 30 min of gradient elution for the anion-exchange separation. As shown in the upper chromatogram of Fig. 2, three sharp peaks were observed in the initial lo-min period, whereas more than six peaks clustered together were observed in the 15-30 min period and no peak was detected in the last part of the gradient (3045 min period). !.A! 0 0
100 100 0
7
TIME
-3
(min)
L!!
\
:
1:
f
18 :i
:: 43
19 59
100 100
60
0
8
0 0 10 I
11 35
40
45
50
55
Fig. 2. Anion-exchange HPLC of soluble antigen on DEAE TSK 5PW column with a flow-rate of ml/min. The gradient profiles of the 46-min (A) and 60-min (B) runs are described over the space available in the chromatograms. Numbers on the peaks correspond to the fractions collected.
1
154
J. G. MOHANTY.
Y. ELAZHARY
For a better separation of peaks, the region around the clustered peaks was expanded by decreasing the slope of the gradient. Following this slow gradient, a steeper gradient was applied to run directly up to 100% buffer B, which was maintained for a further lo-min period. In this manner, as shown in the lower chromatogram of Fig. 2, two peaks were eluted in the initial lo-min perod (being unbound to the column), seven peaks were eluted with the gradient up to 43% of buffer B and two peaks were eluted during the lo-min period of 100% buffer B alone. Peaks l-l 1 were collected at the points of inflection in the chromatogram and analysed further. The protein content of each peak in one separation was measured as described in the Experimental section. The recovery of protein from the column was 60-70% (data not shown). The SDS-PAGE and Western blot analysis of peak fractions l-6 and 7-l 1 are presented in Figs. 3 and 4 respectively. As shown in Fig. 3A, hardly any band was detected by Coomasie Blue for the peak fractions l-6. However, as expected when silver stained, many bands were detected. Interestingly, peak fraction 4 showed one single band corresponding to 84kD in the silver stained portion as well as its immunoblot (Fig. 3B and C). Peak fractions 5 and 7 contained BVDV specific polypeptides corresponding to 90 kD band b and 105 band a respectively. Although in the chromatogram, peak 7 was larger than peak 10, the latter had a higher number of protein bands (Fig. 4). However, Peak 10 showed no positive band in its immunoblot counterpart. Analysis of Figs. 3 and 4 suggest that most of the prominent bands recognized by anti-BVD serum were present in peak fractions 4, 5 and 7.
Fig. 3. SDS-PAGE (A, Coomassie Blue stained; B, silver stained) and Western blot (C) analysis of soluble antigen (lane g) and the anion-exchange HPLC peak fractions 1-6 (lanes a-f respectively). Molecular weight standards were myosin (200 kD), /%galactosidase (I I6 kD), phosphorylase B (92.5 kD), bovine serum albumin (67 kD) and chicken egg albumin (43 kD). K = kiloDaltons (kD).
HPLC
AND
IMMUNOLOGY
OF VIRUS
ANTIGENS
155
Fig. 4. SDS-PAGE (A, Coomassie Blue stained) and Western blot (B) analysis of soluble antigen (lane f) and anion-exchange HPLC peak fractions 7-l I (lanes a-e respectively). Molecular weight standards were the same as in Fig. 3. K = kiloDaltons (kD).
Based upon this result, the immunogenicity of the peak fractions 4, 5 and 7 were tested by raising antisera against them in guineapigs. The antiserum against peak fraction 4 did not neutralise the virus, whereas the titre of seroneutralisation for the antisera against peak fractions 5 and 7 were 1:20 and 1:80 respectively. This suggests that the peak fraction 7 contains the most important BVDV specific polypeptide( But as mentioned above, this peak fraction also contained a number of other polypeptides not recognisable by anti-BVD serum.
HPLGPC of soluble antigen and analysis of its peak fractions About 0.8 mg of soluble antigen proteins in 100 ~1 of the mobile phase was injected on to the HPLGPC column and the chromatography was run for 80 min at a flow-rate of 0.5 ml/min. Chromatographic separation pattern of soluble antigen was reproducible and the chromatogram consisted of seven major peaks and five minor peaks (Fig. 5). Protein molecular weight standards, catalase (240 kD), lactate dehydrogenase (140 kD), enolase (67 kD), myokinase (25 kD) and cytochrome c (12.5 kD) were run on the column in the same condition to calibrate the column. The peak fractions were collected, lyophilised to dryness, reconstituted in same volume (1 ml) of water and their protein content was measured as described in the Experimental section. The calculated molecular weights of the twelve peaks and their protein content are presented in Table I. About 88% of the total protein loaded on to the column could be recovered in fractions. Since the expected BVDV specific polypeptides a, b, c and d are greater than 25 kD the peak fractions 8-12 were ignored and the rest of the peaks were analysed further by SDS-PAGE and Western blot technioue.
J. G. MOHANTY,
156
Y. ELAZHARY
MINUTES
Fig. 5. HPLGPC of soluble antigen (1 ml injection) on TSK G3000SW column with 0.1 M K2HP04 phosphate buffer, pH 7.0, containing 0.2 M sodium chloride and 0.5% Berol 185 as the mobile phase. Flow-rate was 0.5 ml/min. Peak numbers l-5 were expanded ten times (IO X) to visualise them clearly.
As can be seen in the Fig. 6, the Western blot analysis shows that the antiBVD serum detected BVD virus specific polypeptides of 105 and 84 kD in peak fraction 4 which also contained non-antigenic polypeptides of other molecular weights. It is important to note that although peak 4 corresponded to a molecular weight of 290 000 Daltons in HPLGPC, it was resolved into several small sized polypeptides in SDS-PAGE (Fig. 6). Immunogenicity qf HPLGPC peak ,fraction 4 It has been noted above that the peak fraction from the anion-exchange
HPLC containing the 105-kD polypeptide was immunogenic in guineapigs and its antiserum neutralised BVD virus. Thus we were interested to see if peak fraction 4 from HPLGPC, containing the 105-kD polypeptide, would stimulate immune response in calves as well. For this purpose, two calves bearing numbers 33 and 35 were injected as follows. For each injection, calf 33 received the whole of peak fraction 4 containing
I
TABLE PROTEIN Peak No.
I 2 3 4 5 6 7 8 9 10 II 12
CONTENT
AND
SIZES OF HPLGPC
Molecular weight (Daltons)
Protein content (mg)
> 300 000
0.05 0.05 0.1 0.1 0.1 0.17 0.05 0.01 0.01 0.00 0.02 0.04
> 300 000 > 300 000 290 000 150 000 78 000 60 000 25000 x 5200* x 4200* z 2800* z 1650* l
MOLECULAR
Extraoolated
values.
PEAKS
HPLC
AND
IMMUNOLOGY
OF VIRUS
157
ANTIGENS
a b c rl’e
f B h
Fig. 6. SDS-PAGE (A, silver stained) and Western blot (B) analysis of soluble HPLGPC peak fractions l-7 (lanes bh respectively). Molecular weight standards Fig. 3. K = kiloDaltons (kD).
antigen (lane a) and were the same as in
about 100 pg protein obtained by HPLGPC of 0.8 mg of soluble antigen while calf 35 received about 0.8 mg of soluble antigen itself reconstituted in mobile phase buffer of HPLGPC. Seroneutralisation titre of the sera from calf 33 and the control calf 35 were 1:128 and 1:256, respectively. DISCUSSION
Prior to HPLC, soluble antigen was analysed by SDS-PAGE and Western blot techniques to identify the BVDV specific polypeptides. Although the SDS-PAGE patterns of non-infected calf testicle cell extract and the soluble antigen were similar due to the overabundance of cell proteins, the corresponding Western blots demonstrated the presence of four BVDV specific polypeptides of 105, 90, 84 and 67 kD in the soluble antigen (Fig. 1). Recently, Donis and Dubovi14 reported the resence of about six viral specific polypeptides in the range of 48 to 165 kD in the soluble antigen by radioimmunoprecipitation followed by SDS-PAGE (reducing condition) and fluorography. Only the 80-kD polypeptide reported by them would correspond to the prominent 84-kD polypeptide band in the present study. The reason for the discrepancy in sizes between these studies may be due to the different methods of sample preparation and/or the different techniques of immuno detection used. The anion-exchange HPLC of the proteins in soluble antigen separated the BVDV specific polypeptides 84, 90 and 105 kD in three separate peaks (Figs. 2-4) while HPLGPC collected all of them in a single 290-kD protein aggregate peak (Figs. 5 and 6). It was interesting to note that the 84-kD polypeptide could be separated in a single homogenous peak from the rest of the proteins in the anion-exchange HPLC
158
J. G. MOHANTY,
Y. ELAZHARY
of soluble antigen (Fig. 3). The 67-kD BVDV specific polypeptide could not be located in any of the peak fractions possibly due to its low concentration. All the peaks of HPLGPC and anion-exchange HPLC except peak 4 in the latter were observed to contain more than one polypeptide (Figs. 3,4 and 6). A similar phenomenon was also observed by McGregor et ~1.~ during separation of platelet membrane glycoproteins. It could similarly be argued that the 240- and 140-kD BVDV specific peaks reported by Coria et al7 for HPLGPC of soluble antigen were also protein aggregates and not single homogeneous polypeptides. The reason for such protein aggregation during HPLC could be due to strong interactions among certain polypeptides which could not be minimised under the conditions of chromatography used in this study. Attempts were made to examine the Western blot positive peak fractions for their immunogenicity in animals. Interestingly, the antisera raised against the peak fractions containing 90- and 105kD polypeptides neutralised BVDV while the antiserum against the 84-kD polypeptide did not. Perhaps the 84-kD polypeptide did not have a virus neutralising epitope and thus it may be a constituent polypeptide of the viral nucleocapsid. Moreover, the seroneutralisation titre of the serum raised against the fraction containing 105-kD polypeptide was higher than the serum against peak fraction with 90-kD polypeptide. These results suggested that possibly the 105 and the 90-kD polypeptides together constituted a better BVDV neutralising epitope. Infact, the antiserum raised against the 290-kD protein aggregate peak from HPLGPC of soluble antigen which contained both the 105- and 90-kD polypeptides had a BVDV neutralising titre quite close to that of the serum raised against the soluble antigen itself (see Results section). The 105- and 90-kD polypeptides may constitute the cell binding proteins of the virus and further studies are underway in an attempt to identify the virus receptor using these proteins. ACKNOWLEDGEMENTS
We are grateful to Dr. Brian G. Talbot and Dr. G. Khittoo for their valuable suggestions during the preparation of this manuscript. REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14
F. Fener, Intervirology, 7 (1976) 44. C. P. Peter, D. E. Tyler and F. K. Ramsey, J. Am. Vet. Med. Assoc., 150 (1967) 46. R. G. Thomson and M. Savan, Can. J. Comp. Med. Vet. Sci., 27 (1963) 217. D. E. Gutekunst and M. A. Malmquist, Arch. Gesamte Virusforsch., 15 (1965) 159. A. F. Purchio, R. Larson and M. Collett, J. Virol., 50 (1984) 666. A. L. Fernelius, L. G. Classik and R. L. Smith, Am. J. Vet. Res., 32 (1971) 1963. M. F. Coria, M. J. F. Schmerr, A. W. McClurkin and S. R. Bolin, Am. J. Vet. Res., 45 (1984) 2129. P. Karsnls, J. Moreno-Lopez and T. Kristiansen, J. Chromatogr., 266 (1983) 643. J. L. McGregor, P. Clezardin, M. Manach, S. Gronlund and M. Dechavanne, J. Chromatogr., 326 (1985) 179. A. W. McClurkin, E. C. Pirtle, M. F. Coria and R. L. Smith, Arch. Gesamte Virusforsch., 45 (1974) 285. U. K. Laemmli, Nature (London), 227 (1970) 680. H. Towbin, T. Staehelin and 1. Gordon, Proc. Natl. Acad. Sci. U.S.A., 76 (1979) 4350. H. G. Bahnemann, Arch. Virol., 47 (1975) 47. R. 0. Donis and E. J. Dubovi, Virology, 158 (1987) 168.