May 9, 1988 - HURK,"13 LORNE A. BABIUK,13* AND MICHAEL J. P. LAWMAN'13. Department of ... biological function(s) of the glycoproteins of BHV-1. Al- though the ...... Choo, K. H., G. Filby, S. Greco, Y.-F. Lau, and Y. W. Kan. 1986.
JOURNAL OF VIROLOGY, Nov. 1988, p. 4239-4248 0022-538X/88/114239-10$02.00/0 Copyright © 1988, American Society for Microbiology
Vol. 62, No. 11
Expression of Bovine Herpesvirus 1 Glycoproteins Transfected Murine Cells
gI
and glll in
DAVID R. FITZPATRICK,' TIM ZAMB,2 MICHAEL D. PARKER,'3 SYLVIA VAN DRUNEN LITTEL-VAN HURK,"13 LORNE A. BABIUK, 13* AND MICHAEL J. P. LAWMAN'13
DEN
Department of Veterinary Microbiology, University of Saskatchewan,' and the Veterinary Infectious Disease Organization,3* Saskatoon, Saskatchewan, Canada, S7N OWO, and Department of Veterinary Sciences, University of Nebraska, Lincoln, Nebraska 685832 Received 9 May 1988/Accepted 19 July 1988
Genes encoding two of the major glycoproteins of bovine herpesvirus 1 (BHV-1), gI and gIll, were cloned into the eucaryotic expression vectors pRSVcat and pSV2neo and transfected into murine LMTK- cells, and cloned cell lines were established. The relative amounts of gI or gIII expressed from the two vectors were similar. Expression of gI was cell associated and localized predominantly in the perinuclear region, but nuclear and plasma membrane staining was also observed. Expression of gI was additionally associated with cell fusion and the formation of polykaryons and giant cells. Expression of gIII was localized predominantly in the nuclear and plasma membranes. Radioimmunoprecipitation in the presence or absence of tunicamycin revealed that the recombinant glycoproteins were proteolytically processed and glycosylated and had molecular weights similar to those of the forms of gI and gIII expressed in BHV-1-infected bovine cells. However, both recombinant glycoproteins were glycosylated to a lesser extent than were the forms found in BHV-1-infected bovine cells. For gI, a deficiency in N-linked glycosylation of the amino-terminal half of the protein was identified; for glll, a deficiency In 0-linked glycosylation was implicated. The reactivity pattern of a panel of gI- and gIII-specific monoclonal antibodies, including six which recognize conformation-dependent epitopes, was found to be unaffected by the glycosylation differences and was identical for transfected or BHV-l-infected murine cells. Use of the transfected cells as targets in immune-mediated cytotoxicity assays demonstrated the functional recognition of recombinant gI and glll by murine antibody and cytotoxic T lymphocytes. Immunization of mice with the transfected cells elicited BHV-1-specific virus-neutralizing antibody, thus verifying the antigenic authenticity of the recombinant glycoproteins and the important role of gI and gIll as targets of the immune response to BHV-1 in this murine model system.
Bovine herpesvirus 1 (BHV-1) specifies four major glycoproteins, tentatively designated gI, gIl, glll, and gIV, which are homologous to the herpes simplex virus (HSV) glycoproteins gB, gE, gC, and gD, respectively (50; T. Zamb et al., manuscript in preparation). Of these glycoproteins, gI, gIll, and gIV have been identified as the major immunogens recognized by sera from cattle infected with BHV-1 (49). Furthermore, immunization with any of these three glycoproteins, individually or in combination, has been shown to induce significant protection against BHV-1 infection in cattle (3). However, little is known of the cellular immune responses to BHV-1 glycoproteins, which are likely to be the most important responses mediating the observed protection (3, 42). In particular, the induction of major histocompatibility complex antigen-restricted T lymphocytes by the glycoprotein immunogens would be a prerequisite for all of the acquired cell-mediated immune defense mechanisms (42). A second area which is poorly understood concerns the biological function(s) of the glycoproteins of BHV-1. Although the homology of gI, gII, gIII, and gIV to the HSV glycoproteins noted above is well cstablished at the nucleotide and amino acid sequence levels (Zamb et al., in preparation), it has not yet been directly demonstrated that the glycoproteins of BHV-1 possess homologous functions. The HSV glycoprotein-associated activities of virus attachment (15), virus penetration (16, 43), cell fusion (31, 36), virus assembly (1), complement factor C3b binding (12), and immunoglobulin G (IgG) Fc binding (23) are intimately *
associated with the pathogenesis of infection, and it is therefore important to determine whether such functions are conserved in BHV-1. To analyze the immunobiology of the major BHV-1 glycoproteins in more detail, we have initiated studies to produce each glycoprotein by recombinant DNA, mammalian cell-based expression systems. The establishment of stable mammalian cell lines which constitutively express authentic individual BHV-1 glycoproteins would be particularly useful in studies to determine the biological function(s) of each glycoprotein and to dissect the specificities of individual cellular immune defense mechanisms directed against BHV-1. In this report, we describe the derivation of transfected, cloned murine cell lines expressing BHV-1 gI and gIll. Subcellular localization of expression was determined by immunocytochemistry. Structural and antigenic characteristics of the recombinant glycoproteins were examined by comparison with the glycoproteins produced in BHV-1infected bovine cells, by analysis of the reactivities of a panel of glycoprotein-specific monoclonal antibodies, and, for gI, by testing for the biological function of cell fusion. Use of the cell lines in preliminary analysis of the specificity of antibody- and cell-mediated immune responses to BHV-1 in a mouse model system is described. MATERIALS AND METHODS Reagents and media. Restriction enzymes, T4 DNA polymerase, T4 DNA ligase, calf intestinal alkaline phosphatase, phosphorylated BglII linker oligonucleotides, deoxynucleo-
Corresponding author. 4239
4240
FITZPATRICK ET AL.
J. VIROL.
These manipulations placed the gI gene start codon approximately 130 bp downstream of the SV40 early promoter and A. pRSVgI approximately 100 bp downstream of the transcriptional t t t t start site associated with this promoter (11; Fib. 1B). FolB/Ba E TG> A lowing the gI gene stop codon were approximately 430 bp of 4 + ________ lnoncoding BHV-1 DNA, 170 bp of noncoding TnS DNA, the B. pSV2gI L7YZIZIZm + B/Ba sequences encoding the SV40 small t-antigen intron, and the 4 E B/Ba SV40 polyadenylation signals (45). ATG TAG The complete coding sequence of BHV-1 gIll was excised + + I kb from a subclone of pSD113 as a 2,400-bp BamHI-EcoRI E-C. pRSVgIII fragment (32; Zamb et al., in preparation), treated with T4 t t t t DNA polymerase, ligated to BglII linkers, digested with E IB/Ba E BglII, and then cloned into pRSVcat and pSV2neo as ATG lo described for BHV-1 gI (Fig. 1C and D). In the pRSVgIII ______> zIzzzIzD. pSV2gIII construction, the gIll gene start codon was placed approxi4 4 B/Ba E E mately 140 bp downstream of the RSV promoter and 110 bp FIG. 1. Structure of gI and gIll expression plassmid construcdownstream of the transcriptional start site associated with tions. The origins of DNA sequences included in Ithe expression this promoter. Approximately 850 bp of DNA lay between plasmids are represented as follows: Ffiz, pBR3222; rmr, RSV; the glll stop codon and the vector-associated polyadenylaE, SV40; _, BHV-1; ED, TnS (20, 45). The start and stop tion signals. In the pSV2gIII construction, the gIII start codons of gI and gIll are indicated (Zamb et al., in pr(eparation), and codon was positioned approximately 170 bp downstream of ATG
---.
T G3 A 4
E
E
B
-*
B
TAG
E
B
the direction of transcription from the RSV and SV40 promoters are arrowed (11, 55). Restriction endonuclease cleaNiage sites: E, EcoRI, B, BglII; B/Ba, BgllI-BamHI sites destroyec I by ligation,
~re
side triphosphates, and protein A-Sepharose we purchased from Pharmacia, Dorval, Quebec, Canada, and used as recommended by the manufacturer. Other clhemicals and reagents for DNA manipulations, transfections ,, and protein analysis were purchased from Sigma Chemiical Co., St. Louis, Mo., and used in the standard methods described by Maniatis et al. (30) and Davis et al. (9), excep t when noted otherwise below. Cell culture media, fetal b)ovine serum (FBS), G418 and other cell culture reagents Mvere obtained from GIBCO/BRL, Burlington, Ontario, Canaida. Antibodies, wheat germ agglutinin, avidin-biotin immuinoperoxidase staining kits, and other reagents for enzyme innmunoassays were purchased from Dimension Laboratories, Mississauga, Ontario, Canada, and used as recommended by the manufacturer. Radioisotopically labeled compounds and reagents for fluorography were purchased from Amersh; am, Oakville, Ontario, Canada. Plasmid constructions. The complete coding sequence of BHV-1 gI was excised from a subclone of pSDI L06 (32; Zamb et al., in preparation) and inserted into the expr -ession vector pRSVcat (20) in place of the cat gene by li, gation of the 3,300-base-pair (bp) BglII-BamHI gI gene fragiment into the HindIII-HpaI sites of pRSV cat after these si ites had been converted into a unique BgII cloning site by blunt-end repair, BglII linker addition, and BglII dige -stion. These manipulations removed the normal viral promc)ter upstream of the gI gene and placed the start codon off the gI gene approximately 100 bp downstream of the R'ous sarcoma virus (RSV) promoter and approximately 70 bp) downstream of the transcriptional start site associated with this promoter (55; Fig. 1A). Approximately 480 bp lay betwe en the gI stop codon and the simian virus 40 (SV40)-based pollyadenylation signals remaining in the expression vector afte r the removal of the cat gene. A polyadenylation signal of ] BHV-1 origin approximately 30 bp downstream of the gI stop codon (Zamb et al., in preparation) was retained in this cons truction. The gI gene was similarly subcloned into the expr ession vector pSV2neo (45) in place of the neo gene by li igation of the 3,300-bp BglII-BamHI gI gene fragment into the HindIIISmaI sites of pSV2neo after these sites had be en converted into a unique BglII cloning site as described for pRSVcat.
the SV40 early promoter and 140 bp downstream of the
transcriptional start site. Following the glll stop codon was approximately 800 bp of BHV-1 DNA, plus the Tn5 and SV40 sequences noted above for pSV2gI. Plasmid DNA was prepared for transfection by equilibrium banding in CsCl-ethidium bromide gradients and sterilized by ethanol precipitation. Cells and virus. Madin-Darby bovine kidney (MDBK) and murine 3T3 cells were cultured in Eagle minimal essential medium supplemented with 10% FBS. Murine LMTK- and L929 cells and cultured in Dulbecco modified Eagle medium supplemented with 5% FBS. Virus stocks of BHV-1 P8-2 were grown in MDBK or Georgia bovine kidney cells as previously described (4). Virus stocks of vaccinia virus WR were grown in BS-C-1 cells as previously described (29). Transfections. LMTK- cells were transfected with expression plasmid constructions by a modified calcium phosphate precipitation procedure. LMTK- cells at approximately 50% confluency were rinsed and incubated at 37°C in fresh growth medium for 3 h before transfection. Calcium phosphate precipitates of plasmid DNA were prepared as previously described (22, 54), with pSV2neo DNA incorporated into each precipitate as a cotransfecting selectable marker. Control precipitates were prepared with pSV2neo or salmon sperm DNA only. Medium was removed from the cells, and the DNA precipitates were added and adsorbed for 45 min at room temperature. Growth medium was then added, and adsorption continued at 37°C in a 4% CO2 atmosphere (7). After 4 h the medium was removed, and the cells were exposed to 20% glycerol shock (13) for 2 min at room temperature and then incubated at 37°C in growth medium supplemented with 8 mM sodium butyrate (19). After 16 to 24 h the supplemented medium was removed and replaced with growth medium for 48 h. The cells were then passaged in selective growth medium containing 400 ,ug of G418 per ml, which was replaced every 3 to 5 days. Resistant colonies appeared in 10 to 14 days at a frequency of approximately 1i-' by this method. The colonies derived from each transfection were pooled and cloned by limiting dilution at least once before screening. Immunocytochemistry assay and ELISA. G418-resistant LMTK- cell clones were seeded onto glass chamber slides (Miles Laboratories, Rexdale, Ontario, Canada) and 96-well plastic tissue culture plates (Nunclon, Roskilde, Denmark), which had been precoated with 2 ,ug of poly-L-lysine hydro-
VOL. 62, 1988
EXPRESSION OF BHV-1 gI AND glll GENES IN MURINE CELLS
bomide per cm2, and grown to confluency. For BHV-1infected control cells, MDBK or LMTK- cells were similarly seeded onto poly-L-lysine-coated slides and plates, grown to 80% confluency, and then infected with BHV-1 at a multiplicity of infection of 1. After 1 h of adsorption at 37°C, fresh medium containing 2% FBS was added, and incubation was continued for a further 12 to 18 h, for MDBK cells, or for a few minutes, for infected LMTK- cells. Transfected LMTK- cell clones and control cells were either fixed and permeabilized with methanol at -20°C for 15 min and then washed in Hanks balanced salt solution (HBSS) or, for surface expression studies, washed in HBSS without being fixed. Nonspecific binding sites were blocked by adding heat-inactivated normal equine serum diluted 1:75 in HBSS and incubating the mixture at room temperature for 1 h. The blocking solution was removed, and biotinylated wheat germ agglutinin or monoclonal antibodies specific for gI and glll (52) were diluted 1:1,000 in HBSS and added to the slides and plates, which were incubated at room temperature for 1 h. The slides and plates were then processed with an avidin-biotin-enhanced immunoperoxidase assay kit specific for mouse IgG (Vector Laboratories, Burlingame, Calif.) as recommended by the manufacturer, up to the final substrate development step. For slides, the final substrate was 50 mM Tris hydrochloride (pH 7.5)-0.01% H202-1.7 mM NiCl2-1 mg of 3,3'-diaminobenzidine tetrahydrochloride per ml. The substrate reaction was stopped after 5 min of incubation at room temperature by rinsing the slides in tap water. For enzytne-linked immunosorbent assays (ELISAs) the final substrate was 0.1 M citric acid (pH 4.0)-0.015% H202-1 mg of ABTS [2,2'-amino-di-(3-ethylbenzthiazoline sulfonate)] (6) per ml. The ELISA substrate reactions were stopped after a 5- to 10-min incubation at room temperature by addition of sodium dodecyl sulfate (SDS) to a final concentration of 5%, and the A405 of each well was read in a plate reader. Radioimmunoprecipitation. To radiolabel cellular proteins, clones of transfected LMTK- cells at approximately 80% confluency were incubated at 37°C for 6 h in methionine-free Dulbecco modified Eagle medium supplemented with 2% FBS. For glycosylation inhibition studies, tunicamycin was included at this point at a final concentration of 2 ,ug/ml. After 6 h of incubation, [35S]methionine was added to a final concentration of 50 ,uCi/ml, and the cells were then incubated for an additional 18 h. BHV-1-infected MDBK cells were radiolabeled by a similar method, as previously described (47). Radiolabeled cells were harvested by scraping, washed with HBSS, and suspended in modified RIPA buffer (50 mM Tris hydrochloride [pH 8.0], 150 mM NaCl, 1% sodium deoxycholate, 1% Nonidet P-40, 0.1% SDS, 1 mM phenylmethylsulfonyl fluoride). After incubation on ice for 15 min, the cell suspensions were sonicated and then centrifuged at 75,000 x g for 1 h at 4°C. The supernatants were collected, gI- or gIll-specific monoclonal antibody ascites fluid (49) was added to a final dilution of 1:20, SDS was added to a final concehtration of 0.2 to 0.5%, and the samples were incubated for 16 to 18 h at 4°C on a rocking platform. Coated protein A-Sepharose beads were prepared by swelling lyophilized protein A-Sepharose beads in modified RIPA buffer at a concentration of 10 mg/ml for 1 h at 4°C on a rocking platform, then adding rabbit IgG antimouse IgG to a final concentration of 800 ,ug/ml, and incubating the mnixture for a further 16 to 18 h. After incubation, unbound rabbit IgG anti-mouse IgG was removed from the coated protein ASepharose beads by three washes with modified RIPA
4241
buffer. Approximately 10 mg of coated protein A-Sepharose beads was added to each mixture of radiolabeled cell lysate plus monoclonal antibody, and the samples were incubated at 4°C on a rocking platform. After 3 to 4 h, the samples were washed four times with modified RIPA buffer, suspended in reducing sample buffer ( 62 mM Tris hydrochloride [pH 6.8], 2% SDS, 5% 2-mercaptoethanol, 10% glycerol, 0.01% bromophenol blue), and boiled for 4 min. Samples were separated by electrophoresis in SDS-10% polyacrylamide gels and fluorographed (50). Antibody complement cytotoxicity. Transfected murine clones were seeded into 96-well round-bottom plastic tissue culture plates at a density of 2 x 103 cells per well and incubated for 24 h at 37°C in growth medium containing 1.5 pRCi of Na251CrO4 per well. The plates were washed three times, and control, gI-specific, or gIII-specific monoclonal antibodies were added at various dilutions in Dulbecco modified Eagle medium containing 2% FBS and 1 ,ug of actinomycin D per ml. The transfected cells, like all normal nucleated cells, are resistant to complement attack in the absence of metabolic inhibitors such as actinomycin D (6; M. Campos and D. R. Fitzpatrick, unpublished observations). After 2 h of incubation at 37°C, freshly thawed rabbit
complement (Cedar Lane, Hornby, Ontario, Canada), at various dilutions, was added. Control wells for calcjilation of total releasable radiolabel received 3% Triton X-100 instead of complement. After 90 min of incubation at 37°C, 50% of the supernatant fluid from each well was harvested and counted, and the specific release was calculated as previously described (14, 34). Cytotoxic T-ceHl cytotoxicity. C3H/HeJ (H-2k) or Balb/c (H-2d) mice were immunized intraperitoneally with approximately 108 PFU of BHV-1 at 8 and 11 weeks of age. At 3 weeks after the second immunization, the spleens were excised and cell suspensions were prepared by gentle homogenization. The suspensions were treated with 0.83% ammonium chloride to remove erythrocytes, washed, counted, scored for viability, and seeded into 6-well tissue culture plates at a concentration of approximately 2 x 106 cells per well in RPMI 1640 medium containing 10% FBS, 25 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), and 5 x 10-5 M 2-mercaptoethanol. The cells were restimulated with 2 x 106 PFU of BHV-1 per well and incubated at 37°C in a humidified 5% CO2 atmosphere for 6
days. L929 and 3T3 cells to be used as targets were suspended in RPMI 1640 medium and infected with BHV-1 or vaccinia virus at a multiplicity of infection of 5 for 1 h at 37°C. Infected targets, uninfected controls, and transfected cells were then labeled with Na251CrO4 for 1 h at 37°C. The labeled target cells were washed three times with RPMI 1640 medium containing 5% FBS, 25 mM HEPES, and 5 x 10-5 M 2-mercaptoethanol and then seeded into U-bottom microdilution plates at 104 cells per well. Restimulated effector cells were washed, counted, scored for viability, and added to the plates containing radiolabeled targets at various effector-to-target-cell ratios, with quadruplicate wells used for each variable. The plates were incubated for 7 h at 37°C in a 5% CO2 atmosphere, supernatant fluids were harvested and counted, and specific cytotoxicity values were calculated as previously described (27). Immunizations with transfected cells and antibody titer determinations. C3H/HeJ mice were immunized intraperitoneally with 1065 transfected cells suspended in 0.5 ml of HBSS, without adjuvant, at 6, 10, and 14 weeks of age. Pooled sera were obtained at 5, 8, 11, and 15 weeks of age
4242
FITZPATRICK ET AL.
from groups of five identically immunized mice. BHV-1specific antibody levels were measured by virus neutralization assay and ELISA as previously described (52). RESULTS Expression of recombinant gI and glll in transfected murine LMTK- cells. Approximately 120 limit-diluted clones from transfections of the four expression constructions described above, plus negative control clones derived from a transfection conducted with pSV2neo alone, were screened for expression of BHV-1 gI or glll by ELISA and an immunocytochemistry assay. The use of unfixed or methanol-fixed and permeabilized cells in each assay revealed surface or surface plus intracellular glycoprotein expression, respectively. ELISA was used to compare the relative amount of surface and intracellular gI or glll expression by clones derived from a single transfection and by clones derived from transfections with the different expression vectors. For 17 clones positive for gI expression and 35 clones positive for gIll expression, a similar range and distribution of ELISA readings was obtained with either pRSV- or pSV2based constructions (data not shown). Immunocytochemistry revealed that expression of gI was localized predominantly intracellularly in a perinuclear region which probably corresponds to the Golgi apparatus and/ or rough endoplasmic reticulum of these cells, as evidenced by the similar localization of wheat germ agglutinin (Fig. 2G and H). However, cell surface expression of gI was also visible (Fig. 2B), and some nuclear-membrane localization was manifest as faint rings outlining the nuclei (Figs. 2E and H), which were not detectable in negative controls (Fig. 2D). In addition, clones expressing gI exhibited a high degree of cell fusion, polykaryon formation, nucleus fusion, and giantcell formation (Fig. 2E, J, and K), which was not apparent in clones expressing gIl or negative control clones. Expression of gIll was localized predominantly in the nuclear and plasma membranes, although diffuse cytoplasmic staining was also evident (Fig. 2C, F, and I). The subcellular distributions of recombinant gI and gIll are similar to those observed for these glycoproteins in BHV-1-infected bovine cells (37), although the perinuclear accumulation of gI in the transfected murine cells appears to be greater than that observed in infected bovine cells. Comparison of recombinant gI and glll produced in transfected murine cells with gI and gIll produced in BHV-1infected bovine cells. Radioimmunoprecipitation of gl from BHV-1-infected bovine cells revealed three major protein bands of molecular weight (MW) approximately 130,000, 75,000, and 55,000 (Fig. 3, lane 2), which correspond, respectively, to the intact uncleaved glycoprotein and the two cleavage fragments which are linked by disulfide bonding in the mature nondenatured molecule (50). Only the last two cleavage fragments were precipitated from two clones of murine cells transfected with gI expression plasmids, indicating that proteolytic cleavage of gI occurred to completion in these cells (Fig. 3, lanes 3 and 4). In addition, the larger of the two fragments produced in the transfected murine cells was slightly lower in MW than the equivalent fragment produced in infected bovine cells. Identical results were obtained with a number of other clones positive for gI expression (data not shown). Radioimmunoprecipitation of gIll from infected bovine cells yielded two major bands of MW approximately 99,000 and 73,000 (Fig. 4, lane 2). These correspond, respectively,
J. VIROL.
to the mature glycosylated glll and its partially glycosylated precursor form (50). Only the former band was precipitated
from clones of murine cells transfected with the glll expression plasmids, suggesting that the precursor form(s) of glll is more efficiently processed to mature molecules in the murine cells. As observed for gI, recombinant glll had a slightly lower MW than the mature form of glll produced in infected bovine cells (Fig. 4, lanes 3 and 4). These results were also verified by analysis of a number of other clones positive for gIll expression (data not shown). Analysis of the proteins precipitated from cells treated with an N-linked glycosylation inhibitor, tunicamycin, was conducted to compare the N- and 0-linked glycosylation patterns of the recombinant and infected-cell glycoproteins. Radioimmunoprecipitation with gI-specific antibodies yielded a single band of MW approximately 105,000 from both infected bovine cells and gI-transfected murine cell clones, although additional partially glycosylated products of MW approximately 45,000 to 50,000 also accumulated in the transfected cells (Fig. 3, lanes 7 to 9). The slightly higher MW of the 105,000-MW band in SV2gI-transfected cells (Fig. 3, lane 9) is an artifact of sample volume differences and was not observed in other experiments. The 105,000MW band corresponds to the nonglycosylated, uncleaved form of gI, which accumulates owing to the dependence of gI proteolytic cleavage on N-linked glycosylation and/or associated function(s) which are blocked by tunicamycin (47). The identical MW of this band in both infected bovine cells and transfected murine cells indicates that no 0-linked oligosaccharides are added to gI in either cell type and suggests that the MW differences described above for untreated cells may be due to differences in N-linked glycosylation.
Radioimmunoprecipitation of gIll from tunicamycintreated, BHV-1-infected bovine cells yielded two bands of MW approximately 80,000 and 57,000 (Fig. 4, lane 7). These correspond to a glycosylated form of gIll, containing only 0-linked oligosaccharides, and its nonglycosylated precursor (47). Only an MW 70,000 band was precipitated from the tunicamycin-treated, gIlI-transfected murine cell clones, suggesting that any precursor forms of gIll are rapidly processed in these cells and that the amount of 0-linked oligosaccharides added to gIll is lower than that added in infected bovine cells (Fig. 4, lanes 8 and 9). The antigenic structure of the recombinant gI and gIII produced in the murine cell clones was analyzed with a panel of gl- and gill-specific monoclonal antibodies, the majority of which have been mapped to different epitopes on these glycoproteins (51). Relative antibody reactivity was assessed by ELISA and immunocytochemistry on both fixed and unfixed cells and, for selected monoclonal antibodies, by radioimmunoprecipitation and/or flow cytometry. The immunocytochemistry results for methanol-fixed and permeabilized cells are representative of all the assays used and are shown in Table 1. The reactivity pattern of the entire monoclonal antibody panel was identical for the recombinant and viral forms of gI and gIll, including two gI-specific and four gIll-specific antibodies which do not recognize denatured forms of these glycoproteins (51, 52). These results suggest that the primary, secondary, and/or tertiary structures of the recombinant glycoproteins, in the vicinity of the epitopes recognized by this panel of monoclonal antibodies, are indistinguishable from those of the glycoproteins produced in BHV-1-infected bovine cells. Recognition of gI and gIII by antibody and cell-mediated cytotoxic immune defense mechanisms. The antibody comple-
*~ ~ .K
EXPRESSION OF BHV-1 gI AND glll GENES IN MURINE CELLS
VOL. 62, 1988
t
IW6.
t
~
-4
~
~
~
..
I
t.
A
k;ks A~~ *
* ,.
G
4243
'
9¶
7.*
U
H
a
t J *t i
...
:'f'~
I
K
'i4IiII
FIG. 2. Immunocytochemistry of transfected murine cells expressing BHV-1 gl or gill. LMTK cells transfected with pSV2neo (A, D, and G), pSV2gl (B, E, H, J, and K), or pSV2gIII (C, F, and 1). Identical results were obtained with cells transfected with pRSV-based constructions. Live unfixed cells (panels A, B, and C) or methanol-fixed and permeabilized cells (panels D to K) were treated with monoclonal antibodies specific for gl (panels A, B, D, E, H, J, and K) or gill (panels A, C, D, F, and 1) or with biotinylated wheat germ agglutinin (panel G) and then subjected to an avidin-biotin-enhanced immunoperoxidase staining procedure as described in Materials and Methods.
J. VIROL.
FITZPATRICK ET AL.
4244
1 2 3 4 5 6 7 8 9 10
200'-
11697p.
-200 '116 -97
6604_m_
43N
-43
FIG. 3. Immunoprecipitation of gl from BHV-1-infected bovine cells and transfected murine cells. Lanes: 1 to 5, lysates from untreated [35S]methionine-labeled cells; 6 to 10, lysates from tunicamycin-treated and [35S]methionine-labeled cells. Uninfected MDBK cells (lanes 1 and 6), BHV-1-infected MDBK cells (lanes 2 and 7), a pRSVgl-transfected, cloned LMTK- cell line (lanes 3 and 8), a pSV2gI-transfected, cloned LMTK- cell line (lanes 4 and 9), and a pSV2neo-transfected, cloned LMTK- cell line (lanes 5 and 10) are shown. Radiolabeled cell lysates were immunoprecipitated with a gI-specific monoclonal antibody (1F8 [52]), separated on SDSpolyacrylamide gels, and fluorographed. ment cytotoxicity results shown in Table 2 indicate that gI and glll are expressed on the surfaces of transfected murine cell clones at a level and in a manner which is recognized by complement-fixing gI- or glll-specific monoclonal antibodies and which thereby renders the cells susceptible to attack complement. The lower levels of lysis of cells expressing gI are due primarily to the higher spontaneous release of radioactive label from unstable fusing cells and polykaryons. Table 3 shows the results obtained in cytotoxic T-cell cytotoxicity assays with transfected murine cell clones expressing gI or gIll as targets. In experiment 1, splenic
1
2 3
4
5
6 7 8
116-
;_"
(epitope)"
Reactivity wt denatured
protein"i
Reactivity with following murine cellsb: LMTK--BHV LMTK--gI LMTK-gIII
lBlO (g1-I) 3F3 (gl-ll) lE11 (gI-III) 1F8 (gI-IVa) 5G2 (gI-IVb) 5G11 (gI-IVc) lF1O (gI-V) 2C5 (gI-V) 1B4 (gI-?)
++ ++ ++ ++ ++ ++ ++
1F9 (gI-?)
-
3H7 (glll-Ia) lCll (glIl-Ib) 3E3 (glll-Il) 3F12 (glll-Ill) 1E2 (gllI-IV) 3G8 (g1II-V)
++ ++ ++ ++ ++
+ ++ ++ ++ ++ ++ ++ + + + ++ ++ ++ ++ +
-
++ ++ ++ ++
1D6 (gIII-VI) 2A11 (glIl-VII) lF1l (glll-VIII)
++
+ ++ ++ ++ ++ ++ ++ + + + -
-
++ ++
++ ++ + ++ ++ ++ ++
' Data derived from references 51 and 52. For definition of symbols. see footnote b. b Reactivity as measured by immunocytochemistry of methanol-fixed and permeabilized cells. Identical results were obtained for cells transfected with pRSV- or pSV2-based expression plasmids. Reactivity score: + +, positive, +, weakly positive; -, negative.
.197 -66
TABLE 2. Antibody complement cytotoxicity of transfected murine cell clones expressing gI or glll
9 10
4116
A_ 97CNW
Monoclonal antibody antibody
lymphocytes from mice immunized and restimulated with BHV-1 recognized and lysed histocompatible cells infected with BHV-1. A portion of this activity was nonspecific natural killer cell-like cytotoxicity, as evidenced by the lysis of vaccinia virus-infected targets and nonhistocompatible targets; however, the marked restriction of cytotoxicity which occurred when nonhistocompatible target cells were used provided proof of the involvement of cytotoxic, major histocompatibility complex-restricted T lymphocytes. The results of experiment 2 confirm the above findings and establish the optimum effector-to-target-cell ratio for measurement of specific cytotoxicity as 50:1. Experiment 3 demonstrates that recombinant gI and glll expressed by the transfected cell lines are recognized by a significant proportion of BHV-1-specific cytotoxic T lymphocytes. Note that the lysis of the negative control vaccinia virus-infected L929
-200
200'
TABLE 1. Reactivity of BHV-1-specific monoclonal antibodies with BHV-1-infected or transfected murine cells
so
66-
% Specific releasea
43'
Target cells
.43
FIG. 4. Immunoprecipitation of glll from BHV-1-infected bovine cells and transfected murine cells. Lanes: 1 to 5, lysates from untreated [35S]methionine-labeled cells; lysates from tunicamycintreated and [35S]methionine-labeled cells. Uninfected MDBK cells (lanes 1 and 6), BHV-1-infected MDBK cells (lanes 2 and 7), a pRSVgIII-transfected, cloned LMTK- cell line (lanes 3 and 8), a pSV2glII-transfected, cloned LMTK- cell line (lanes 4 and 9), and a pSV2neo-transfected, cloned LMTK- cell line (lanes 5 and 10) are shown. Immunoprecipitations were conducted as described in the legend to Fig. 3, except that a gIll-specific monoclonal antibody (1D6 [52]) was used.
Negative control monoclonal
antibody' LMTK--pSV2neo LMTK--pRSVgl
LMTK--pSV2gl LMTK--pRSVgIlI LMTK--pSV2gIIl "
0.8 0.6 2.3 0.3 0.6
gilI
gI-specific monoclonal
specific
antibody (lE11)
monoclonal
0.0 8.3 25.0
0.0
-
51.7 47.6
-
antibody
(11D6) -C -
Spontaneous release in the presence of complement alone did not exceed
17% of the total releasable radiolabel.
b BHV-1 glV-specific monoclonal antibody 3D9 (52). -, Not done.
EXPRESSION OF BHV-1
VOL. 62, 1988
gI
AND glll GENES IN MURINE CELLS
4245
TABLE 3. Lysis of transfected murine cells expressing gI or glll by BHV-1 specific cytotoxic T lymphocytes % Specific release" with following effector cell' and effector-to-target-cell ratio: Expt
Target cells
1
2
3
C3H/HeJ
BALB/c (50:1)
100:1
50:1
25:1
5:1
1:1
L929 L929-BHV L929-vaccinia virus 3T3 3T3-BHV
-' -
-
-
-
-
5.4 62.4 9.6 3.4 18.4
-
-
-
-
-
-
-
3.4 45.6
L929-vaccinia virus L929-BHV 3T3-BHV
12.3 72.0 20.3
8.8 63.6 14.3
6.4 34.3 9.1
2.1 21.0 10.2
2.4 14.2 0.8
-
LMTK--pSV2neo LMTK--pRSVgI LMTK--pSV2gI LMTK--pRSVgIII LMTK--pSV2gIII L929-BHV L929-vaccinia virus L929
-
1.1 25.2 0 23.1 14.0 63.1 18.0 0
-
-
-
-
-
-
-
-
6.1 11.6
-
-
-
-
-
-
-
-
-
aSpontaneous release from targets did not exceed 25% of the total releasable radiolabel. b Effector spleen cells were restimulated in vitro with BHV-1 for 6 days. -, Not done.
cells in experiment 3 is abnormally high compared with the results of experiments 1 and 2. We have also verified the specific recognition of gI and glll by using recombinant vaccinia virus-infected targets which express these glycoproteins (M. J. P. Lawman et al., manuscript in preparation). The different levels of lysis for pRSV- versus pSV2-based transfected cells, particularly for the gI-expressing cells, does not correlate with the comparable total expression of the recombinant glycoproteins as measured by radioimmunoprecipitation and ELISA and may therefore reflect quantitative and/or qualitative differences in the amount of processed antigen(s) which is produced by the different transfected cell lines and recognized by the cytotoxic effector cells in this assay. Immunogenicity of transfected cells in mice. Histocompatible mice immunized with transfected cells in the absence of adjuvant produced detectable BHV-1-specific antibody after only one immunization (Table 4). Both ELISA and virusneutralizing antibody levels were significantly boosted by secondary but not by tertiary immunization. The induction of comparable antibody levels with cells expressing gI or gIll under the control of different enhancer-promoter units corroborates the data above, which suggest that the SV40 and RSV elements are quantitatively equivalent expression units
for these glycoproteins in LMTK- cells. The induction of significant levels of virus-neutralizing antibody supports the reactivity and cytotoxicity data which indicate that the recombinant glycoproteins are antigenically authentic. DISCUSSION In this report we have described the derivation of cloned murine cell lines expressing two of the major glycoproteins of BHV-1: gI and glll. Two different eucaryotic expression vectors were used for each glycoprotein gene owing to the existence of conflicting reports concerning the relative efficiency of the SV40- and Rous sarcoma virus-based enhancer-promoter units in murine cells (8, 19) and owing to our intention to express these glycoproteins in a number of mammalian cell lines in which cell specific factors may effect the rate of expression from different enhancer-promoter units (18). Although we did not quantitate the gene copy number, transcription rate, or other factors which may contribute to expression, we found that the final rate of expression of mature gI or gIll from the pRSV cat- and pSV2neo-based constructions was similar in LMTK- cells. This observation was consistent for a number of clones derived from several transfections with either glycoprotein.
TABLE 4. Serologic responses of mice immunized with transfected LMTK- cells expressing gl or gilll Cell 0
LMTK--pSV2neo
LMTK--pRSVgI LMTK--pSV2gI LMTK--pRSVgIII LMTK--pSV2gIII
