key: cord-0755459-5dshegp6
authors: Deregt, Dirk; Babiuk, Lorne A.
title: Monoclonal antibodies to bovine coronavirus: Characteristics and topographical mapping of neutralizing epitopes on the E2 and E3 glycoproteins
date: 1987-12-31
journal: Virology
DOI: 10.1016/0042-6822(87)90134-6
sha: 10f13e39553ffb0656e2286446fda5ac13dc3c8b
doc_id: 755459
cord_uid: 5dshegp6

Abstract Monoclonal antibodies to the Quebec isolate of bovine coronavirus were produced and characterized. Monoclonal antibodies to both the E2 and the E3 glycoproteins were found to efficiently neutralize virus in vitro. None of the monoclonal antibodies directed against the E1 glycoprotein neutralized virus infectivity. Neutralizing monoclonal antibodies to the E2 glycoprotein were all found to immunoprecipitate gpl90, gpl00, and their intracellular precursor protein gp170. Neutralizing monoclonal antibodies to the E3 glycoprotein immunoprecipitated gp124 and showed differential reactivity to its precursor proteins gp59 and gpl 18. These monoclonal antibodies also showed differential reactivity to an apparent degradation product of E3. Neutralizing monoclonal antibodies to E2 bound to two distinct nonoverlappig antigenic domains as defined by competitive binding assays. Neutralizing monoclonal antibodies to the E3 glycoprotein also bound to two distinct antigenic sites as defined by competitive binding assays plus a third site which overlapped these regions. Other results indicated that one domain on the E3 glycoprotein could be further subdivided into two epitopes. Thus four epitopes could be defined by E3-specific monoclonal antibodies.

Bovine coronavirus (BCV), an enteric virus, is composed of four structural proteins . Three of these proteins are analogous to the mouse hepatitis virus (MHV) A59 strain ; E1, a transmembrane protein of 23K-26K (molecular weight, MW) ; E2 the large peplomeric protein of coronaviruses, which exists in both uncleaved and cleaved forms of approximately 180K-190K and 90K-120K, respectively ; and N, the nucleocapsid protein (50K-54K) and its trimer (160K) (Sturman and Holmes, 1977 ; Sturman et al., 1985 ; Storz et al., 1981 ; King and Brian, 1982 ; Hogue et al., 1984 ; Robbins et a!, 1986 : Deregt et al., in press ) . The fourth protein of BCV, E3, is a disulfide-linked dimer with a molecular weight of approximately 124K-140K, reducible to two apparently identical subunits of 62K-65K (King and Brian, 1982 ; Deregt et al., in press ) . This protein has been identified as the BCV hemagglutinin and does not appear on MHV-A59 (Sturman and Holmes, 1977 ; Hogue et a!, 1984 ; King at al., 1985) .

BCV (Quebec isolate) intracellular precursor proteins have been recently identified by us (Deregt et a/., in press ) . The precursor protein to the BCV E2 protein (gpl 90/gp 100) is a glycoprotein of 170K (MW), gpl70 . This protein is apparently further glycosylated to yield r A portion of this work was presented at the Third International Coronavirus Symposium at Asilomar (1986) . 2 To whom requests for reprints should be addressed .

gpl90/E2 before subsequent proteolytic cleavage to yield gpl00 . Two presumptive comigrating gplOO/E2 species are believed to be the result of cleavage . The precursor proteins to the BCV E3 glycoprotein (gp124/gp62) are gp59 (monomer) and gpl18 (disulfide-linked dimer) . The primary precursor was found to be gp59 which apparently undergoes rapid dimerization to produce gpl 18 before further glycosylation to gp 124/E 3 .

In this report, monoclonal antibodies to BCV were characterized as to their specificities and their abilities to neutralize virus infectivity. Monoclonal antibodies to both BCV E2 and E3 glycoproteins were found to efficiently neutralize virus infectivity in vitro in the absence of complement . Also, in this report, neutralizing epitopes of the E2 and the E3 proteins were topographically mapped and further characterized .

The Quebec isolate of BCV (obtained from S . Dea, Department of Pathology and Microbiology, Faculty of Veterinary Medicine, University of Montreal, St . Hyacinthe, Quebec) was used in these studies (Dea et al ., 1980) . Virus was propagated in Madin-Darby bovine kidney (MDBK) cells and purified as described previously (Deregt et al., in press ) .

BCV infectivity was assayed in MDBK cells and plaques were allowed to develop under 0 .8% agarose as previously described (Deregt et al., in press) . To quantify the virus-neutralizing ability of monoclonal antibodies 100 PFU of virus in 250 µl MEM was incubated with an equal volume of various dilutions of heat-inactivated (56° for 30 min) mouse ascites fluid for 1 hr at 37°. The virus-ascites fluid mixtures were then allowed to adsorb to confluent MDBK monolayers for 1 hr . Controls included ascites fluid containing monoclonal antibodies to bovine herpes virus-1 . Neutralization endpoint titers were expressed as the reciprocal of the highest dilution of monoclonal antibody which gave a 50% reduction in plaque numbers .

Production and conjugation of monoclonal antibodies BALB/c mice were immunized with either purified whole virus or SDS-denatured virus preparations (Cianfriglia et at., 1983) . The production of monoclonal antibodies is described elsewhere (Deregt et al., in press) . Monoclonal isotypes were determined by an enzyme immunoassay of hybridoma supernatants (Hyclone Laboratories, Logan, UT) . Total IgG concentration of ascites fluids was determined by radial immunodiffusion (ICN Immuno-biologicals, Lisle, IL) .

Monoclonal antibodies were conjugated to horseradish peroxidase (HRP) by the method of Nakane and Kawaoi (1974) as modified by van den Hurk and Kurstak (1980) .

An enzyme-linked immunosorbent assay (ELISA) for antibody-binding curves and titration of ascites fluids was performed . Ninety microliters of purified BCV (7 .5 µ/ml) in coating buffer (0 .5 M NaHC0 3/Na 2 C0 3 , pH 9 .6) was applied to each well of polystyrene microtiter plates (Immulon 2 ; Dyntech Laboratories, Inc ., Fisher Scientific, Edmonton, Alberta) and incubated overnight at 4°. The plates were washed five times in washing buffer [0 .01 M PBS, pH 7 .2, containing 0 .05% Tween 20, (PBST)] and then 150 µl of blocking buffer, PBST containing 10% heat-inactivated horse serum (GIBCO), was added to each well and incubated for 1-2 hr at room temperature to block unreacted sites . The plates were then washed as before with PBST and 75 µl of monoclonal antibody serially diluted in PBST containing 1% horse serum (PBSTS) was added and incubated for 2 hr at room temperature . After washing the plates to remove unbound antibody, 75 µl of a 1 :1600 dilution of HRP-conjugated goat anti-mouse MONOCLONAL ANTIBODIES TO BOVINE CORONAVIRUS 4 1 1 IgG and IgM (Boehringer-Mannheim, Dorval, Quebec) diluted in PBSTS was then added and incubated for 1 hr at room temperature . After final washing of the plates in PEST, 75 µl of substrate solution containing 1 mg/ml of recrystallized 5-aminosalicylic acid and 0 .005% hydrogen peroxide in 0 .01 M phosphate buffer, pH 6 .0, was added to each well . The enzyme reaction was terminated after 30 min of incubation at room temperature by the addition of 75 µl of 1 .0% sodium azide (Saunders et al., 1977) . The optical density at 490 nm (0D,90 ) was determined immediately after the addition of sodium azide using a microELISA reader (MR580, Dynatech) . Endpoint titers were determined from the point at which the antibody-binding curve crossed the average absorbance attributed to nonspecific binding of the conjugated antibody for each plate . For experiments using saturating amounts of HRPconjugated goat anti-mouse IgG and 19M, dilutions as low as 1 :50 were used and incubated for 3-4 hr . The enzyme reaction with substrate was terminated after 4 or 6 min before substrate became limiting .

For competitive binding assays, the procedure was the same as first described above except that HRPconjugated monoclonal antibodies at concentrations of 1 :400-1 :6400 (equivalent to approximately 60-90% saturation levels for different HRP-conjugated monoclonal antibodies) in PBSTS were used in place of HRP-conjugated goat anti-mouse IgG and IgM . The competition curves were calculated as described by Kimura-Kurado and Yasui (1983) using the formula [100(A -n)]l(A -8)] where A is the OD in the absence of competing antibody, B is the OD in the presence of homologous antibody at 1 :10 (or 1 : 100) dilution, and n is the OD in the presence of competitor at 1 :101_1 :107 dilutions . Thus, by definition, self-competition at low dilution of homologous unlabeled antibody was equal to 100% and the measurement of competition by heterologous unlabeled antibody was a relative measure . Competition at the plateau was regarded as strong (+) if greater than 70%, significant (±) if between 30 and 50% and negative (-) if less than 30% . Actual selfcompetitions were calculated using the formula [100(A -B)]/A, where A and B in this instance were corrected for nonspecific binding as measured by the binding of anti-mouse-conjugated antibody to wells incubated without unlabeled competitor antibody .

Confluent monolayers of MDBK cells were washed once with MEM . Virus, at a multiplicity of infection of 5-10 PFU/cell, was then adsorbed for 1 hr at 37°.

After adsorption the inoculum was removed and replaced with MEM + 2% FBS . This medium was removed and replaced with 50 µCi/ml of [ 35 S]methionine in methionine-deficient medium 8-14 hr postadsorption . At 24-36 hr postadsorption the cells were washed and harvested in ice-cold PBS and after pelleting were prepared for radioimmunoprecipitation or electrophoresis .

For pulse-chase experiments the cells were starved for methionine for 2 hr and then labeled with 200 µCi/ ml of [35 Sjmethionine in methionine-deficient medium beginning at 18-23 hr for different experiments . After a 15-min pulse, the cells were washed with MEM or MEM containing 10 times the normal concentration of methionine and then further incubated with this same medium for 60 min before harvesting .

The procedure used for radioimmunoprecipitation was essentially as that described previously (Deregt et al ., in press ) .

Purified BCV was solublized with various sample buffers (see PAGE, below) and proteins were fractionated in a 10% polyacrylamide gel . Proteins were then transferred to nitrocellulose paper (Bio-Rad) byelectroblotting at 120 V (constant) for 8-10 hr at 4° in Trisglycine methanol buffer (20 mM Tris-hydrochloride, pH 8 .3, 190 mM glycine, 20% methanol) .

After transfer, blots were blocked with 3% skim milk powder (SMP) in 0 .01 M PBS overnight at 4°. Blots were cut into strips and monoclonal antibody [1 :2 dilution of hybridoma supernatants or ascites fluids (1 :200 dilution)] or rabbit anti-BCV polyclonal serum (1 :100 dilution) in PBS containing 0 .05% Tween 20 (PBST) and 1% SMP (PBSTS) was incubated with individual strips in an incubation tray (Bio-Rad) and rocked for 2 hr at room temperature . Strips were then washed with PBST for 30-60 min at room temperature and then incubated with a 1 :2000 dilution of HRP-conjugated goat anti-mouse IgG and IgM or a 1 :5000 dilution of HRP-conjugated goat anti-rabbit IgG (Boehringer-Mannheim) in PBSTS for 90 min at room temperature . Strips were again washed and developed by reaction with 0 .05% 4-chloro-1 -naphthol (Bio-Rad) substrate for 15 min-1 hr . Alternatively, the substrate 0 .02% dianisidine dihydrochloride (Sigma) was used in a slightly modified procedure (Dergt et al,, in press) .

Samples were resuspended in sample buffer (Laemmli, 1970) with or without 2-mercaptoethanol, boiled for 2 min, and stored at -20° or analysed immediately on various percentages of polyacrylamide gels .

Alternatively, samples were resuspended in a sample buffer containing urea, without 2-mercaptoethanol, pH 6 .8 (Sturman, 1980) , and incubated at 37° for 30 min . Electrophoresis and fluorography were performed as previously described (Deregt et al., in press) .

From a total of 105 hybridoma cell lines, 15 monoclonal antibodies were chosen for detailed characterization, 7 were specific for the E2 protein, and 4, 3, and 1 for the E3, El, and N proteins, respectively (Fig . 1) . These antibodies were chosen based on their overall superior reactivity in assays as well as our particular interest in analyzing BCV glycoproteins . All monoclonal antibodies were of the IgG immunoglobulin class . The predominant isotypes were IgG28 and IgG, (six of each) . ELISA titers of ascites fluids ranged from 10 45 to 1075 . Total ascites IgG, as determined by ra- Analysis was done in polyacrylamide gels ranging from 7-12 .596 . For lane F, reactivity was detected using HRP-conjugated rabbit anti-mouse IgG and IgM antibody and the substrate used was dianisidine dihydrochloride . MW, positions of molecular weight standards . Note that the protein gp340, a probable aggregate of gp170/pE2, was more prominant in this analysis then was usually observed . Its MW was calculated from other gels by use of thyroglobulin (MW 330K, Pharmacia) as a MW standard (see also Fig . 2 ) . dial immunodiffusion, ranged from less than 1 to 32 mg/ml .

To determine which monoclonal antibodies could neutralize virus infectivity in vitro without the aid of complement, heat-inactivated ascites fluids and virus were mixed and incubated in a plaque reduction assay (Table 1) . Monoclonal antibodies to both the E2 and the E3 proteins were found to neutralize virus infectivity with plaque reduction (50%) titers of up to 130,000 and 50,000, respectively . None of the monoclonal antibodies to the El or the N protein neutralized virus infectivity in vitro .

Patterns of reactivity of E2-and E3-specific monoclonal antibodies (a) Monoclonal antibodies to the E2 protein . Upon initial screening, by radioimmunoprecipitation, most hybridoma supernatants reactive to E2 were observed to bind to both gp100/E2 and gp170 (designated as pE2, for precursor to E2) as well as to higher molecular weight species, which were apparently aggregated forms of E2 or pE2 . Reactivity to gpl90/E2 was sometimes difficult to assess when regularly labeled lysates were used, because this protein often occurred in small amounts (Fig . 1) . However, in pulse-chase experiments, got 90/E2 occurs quite prominently prior to cleavage to go 100/E2 (Deregf et al., in press) . Thus to Reciprocal of the highest dilution of antibody (ascites fluid) which gave a 50% reduction in plaque number . fected and mock-infected cells, respectively, were pulsed for 1 5 min and chased with MEM for 60 min before immunoprecipitation with a representative E2-specific mAB (BB7-14) . Proteins were solubilized in Laemmli sample buffer without 2-mercaptoethanol before analysis in a 6% polyacrylamide gel . MW, positions of molecular weight standards . Note that lane A represents virus that had undergone two freeze-thaw cycles after solubilization because a portion of this virus was used for a previous analysis . Aggregation of gp100/E2 was apparent as reflected in the relative prominance of the got 90/ E2 band compared with the previous analysis (not shown) and the appearance of band a . Band a has a MW of approximately 300K when gp340 was used as a molecular weight standard and is a presumptive aggregate (trimer) of gpl OO/E2 . assess the binding of selected E2 monoclonal antibodies to these related proteins, lysates were prepared from [ 35 S]methionine pulse-chased cells and used for radioimmunoprecipitation, All of these monoclonal antibodies immunoprecipitated gp100/E2, gp190/E2, gpl70/pE2, and gp340, an apparent aggregate of pE2 (Fig . 2B ) .

(b) Monoclonal antibodies to the E3 protein . (i) Differential reactivity of E3-specific monoclonal antibodies to an apparent degradation fragment of gpl 24/E3 . Upon initial characterization by radioimmunoprecipitation it was noted that monoclonal antibodies to the E3 protein reacted differentially to gp124/E3 and a 64K (MW) polypeptide (Fig . 3A, lanes A-C) . This 64K polypeptide (gp64) was found to be glycosylated (data not shown) and virus specific by its absence in mock-infected cell lysates (Fig . 313, Monoclonal antibodies BD9-8C, KD9-40, and KC4-3 reacted with gp64 (KC4-3 with apparent higher avidity then with gpl 24/E3 itself) while monoclonal antibody (mAB) HC10-5 did not . This differential reactivity did not c hange. by increasing (10-fold) or decreasing (5fold) the amount of mAB HC10-5 used, nor did the exclusion of SDS in the incubation mixture have any effect on HC10-5 immunoprecipitation . Further, it was possible to immunoprecipitate gp64 from supernatants of HC10-5 incubation mixtures with mAB BD9-8C, indicating that the differential reactivity was, indeed, due to binding (i .e ., mABs, BD9-8C, KD9-40, and KC4-3) or lack of binding of this polypeptide (i .e ., mAB HC 10-5 1 data not shown) .

It had further been noted that when immunoprecipi-DEREGT AND BABIUK tates of mAB BD9-8C and purified virus preparations containing gp64 were reduced with 2-mercaptoethanol, a polypeptide of approximately 33K (MW) appeared (Fig . 3B, lane D' ) . This polypeptide was also glycosylated (data not shown) . It was suspected that the 33K polypeptide (gp33) was derived by reduction of gp64 and that the latter protein was a disulfidelinked dieter (similar to gpl24/E3) . If this were true, gp64 should be absent under reducing conditions . Since gpl 24/E3 is itself reduced to gp62/E3, the presence or absence of gp64 would normally be difficult to assess under reducing conditions due to a lack of resolution between these proteins . However, when immunoprecipitates of mAB KC4-3 (which demonstrated greater reactivity to gp64 than to gpl24/E3) were examined, an assessment could be made . Under reducing conditions gp64 was absent and apparently replaced by gp33 (compare Fig . 3B , lanes F and F') . This indicated that gp64 was a disulfide-linked dimer of probably identical gp33 subunits . Since these polypeptides have no apparent role in the genesis of gp124/E3, as observed in pulse-chase experiments (Deregt et al., in press), they were concluded to be degradation products of gp124/E3 and designated gp64/dE3 and gp33/dE3) .

(ii) Differential reactivity to gp59/pE3 and gp118/ pE3 by monoclonal antibodies to E3 . The differential reactivity of E3-specific monoclonal antibodies to the E3 degradation product, gp64, suggested that these monoclonal antibodies might also bind differentially to the E3 precursor monomer gp59-and precursor dimer gpl 18 (designated as pE3) . Thus to determine the binding ability of E3 monoclonal antibodies to these proteins, BCV-infected and mock-infected cells were pulsed with [ 35 S]methionine and chased with MEM, and lysates were prepared and used for immunoprecipitation . The results show that although all E3 monoclonal antibodies could bind to gpl 1 8/pE3 (Fig . 4A,  lanes A, C, E, G) , only mAB BD9-8C was able to immunoprecipitate the monomer gp59/pE3 (Fig . 4A, lane  A) . Monoclonal antibody KC4-3 revealed the same weak binding for gpl 18/-E3 as it did for gp124/E3 . This reaction was better observed after a 30-min pulse with [35 S]methionine (Fig . 4B, lane C) . The same results were obtained when the concentration of monoclonal antibody was increased (10-fold) or decreased (5-fold) . When SDS was excluded from the lysates of pulsed cells a minor amount of what was apparently gp59 in mABs HC10-5 and KD9-40 immunoprecipitations was detected along with apparent nonspecific binding to p53, the nucleocapsid protein . This was observed in lanes A-D when the gel shown in Fig . 4B was exposed 10 times longer (data not shown) . We could not distinguish whether this was due to a very weak reaction with gp59, nonspecific binding, or a slight amount of reduction of gp118, after immunoprecipitation, during sample preparation . Nevertheless, we considered it as a negative result . Based on their differential reactivities with gp124/E3, gp64/dE3, gp59/pE3, and gpl 18/pE3, summarized in Table 2 , the E3 monoclonal antibodies defined four epitopes on the E3 protein .

To determine the nature of BCV epitopes, the binding of monoclonal antibodies in Western immunoblotting assays was studied . These results are shown in Fig . 5 and summarized in Table 3 . All monoclonal antibodies to the E2 protein bound to SDS-denatured antigen in Western immunoblots (Figs . 5E and 5F) . In contrast, only one E2-specific monoclonal antibody (BB10-27) bound to antigen denatured by SDS plus 2-mercaptoethanol although a weaker reaction (than with antigen denatured with SDS alone) was observed (Fig . 5E') E3 also bound to SDS-denatured antigen (Fig . 5G ) . However, mABs HC 10-5 and KC4-3 showed relatively weak reactions compared to mABs BD9-8C and KD9-40 (data not shown) . None of the E3 monoclonal antibodies could react with antigen denatured by SDS plus-2-mercaptoethanol (Fig . 5G') . Thus all E2 and E3 epitopes represented by this panel of monoclonal antibodies, with the exception of that represented by mAB Clone designation neutralization° SDS + 2-ME Determined by studying the reactivity of the antibody in immunoblotting assays .° +, 5096 plaque reduction at ascites dilution of more than 1 :1000 (see Table 2 for actual titers) .°A ntigen boiled in the presence of SDS or SDS plus 2-mercaptoethanol (2-ME) prior to PAGE . B131 0-27 (E2), were apparently dependent upon disulfide linkages for their antigenic integrity . Rabbit polyclonal anti-BCV serum also showed an apparent lack of reaction to reduced species of E2 and E3 . However, weak reactions may have been obscured by background reactions in these experiments (Fig . 5B') . Transfer of these reduced proteins was confirmed when [ 35 S]methionine-labeled virus was immunoblotted (data not shown) .

Two of three mABs specific to El proteins, mABs AE12-11 (to gp26/E1) and CC7-3 (to gp26/E1 and gp25/E1), reacted with antigen denatured under both nonreducing and reducing conditions (Fig . 5C') as did mAB MD8-3 (N specific) (Fig . 5D') . The El monoclonal antibodies also reacted with apparent aggregates of El (most notably, species a) (Fig . 5C') . There was some distortion in the gel front for reduced proteins which caused the E 1 proteins to migrate with apparent lower mobilities (Figs . 5B' and 5C') . The extraneous bands observed were apparently the result of some protein degradation or aggregation (Fig . 5, particularly  lanes B and B' ) .

Binding characteristics of E2 and E3 monoclonal antibodies used in competitive binding assays (CBA) Binding characteristics of E2 and E3 monoclonal antibodies to BCV were determined by parallel ELISA titrations using a single antigen preparation . Binding curves of monoclonal antibodies used in CBA were constructed and all monoclonal antibodies apparently DEREGT AND BABIUK saturated BCV at a dilution of 1 :10 3 (data not shown) . This was confirmed by CBA which typically demonstrated maximum levels of competition over three log o dilutions . Thus, the same amount of input antibody was represented on the plateaus of the binding Curves .

Absorbances reached at the plateau, reflecting the amount of monoclonal antibody bound to BCV, were used as an approximate measure of relative avidity of monoclonal antibodies for their respective proteins (Frankel and Gerhard, 1979 ; Stone and Nowinski, 1980 ; Bruck et at., 1982) . However, since anti-IgG, antibodies were significantly underrepresented in the HRP-conjugated probe, as determined by later titration, comparisons were not made between E2 monoclonal antibodies across isotypes at limiting dilutions of probe . The results suggested that mABs JB9-3 and HF8-8 had significantly lower relative avidities for the E2 protein than did HE7-3 and HB10-4 . When binding was measured using saturating or near saturating levels of the probe, similar adsorbances were reached at the plateau for E2 monoclonal antibodies HE7-3 (IgG,) and BB7-14 (IgG 2b), suggesting that their avidities were similar.

Avidities of the anti-E3 monoclonal antibodies (all of IgG 2a isotype) appeared similar in this assay . However, since the E3 protein is a disulfide-linked dimer of apparently identical subunits, variation in the number of epitopes (one or two) or type of antibody binding (monovalent versus bivalent) were possibilities that would make binding comparisons meaningless at limiting dilutions of the probe . Thus binding was also measured at saturating levels of the probe . Again very similar absorbances were reached at the plateau for E3 monoclonal antibodies indicating that approximately the same amount of antibody binding occurred for the same amount of input antibody .

To determine the topography of epitopes on the E2 and E3 glycoproteins involved in neutralization, monoclonal antibodies described in Table 1 (with the exception of J89-3 which induced an insufficient amount of ascites and BB10-27 which was inefficient at neutralization and whose epitope could be distinguished by other means) were conjugated to HRP and used as probes in CBAs . Conjugation had no apparent adverse effect on the binding properties of these monoclonal antibodies and endpoint titers ranged from approximately 10°4 -10 5 . Binding of HRP-conjugated monoclonal antibody was inhibited from approximately 70 to greater than 990 (>85% for all monoclonal antibodies, except HF8-8, 73%) in the presence of saturating unlabeled homologous antibody . As a measure of avidity these results were in general agreement with those derived from the binding curves .

The results of typical competition binding experiments are shown in Figs . 6 and 7 and summarized in Tables 4 and 5 . For each experiment a monoclonal antibody to another BCV protein was used as a "competitor" control . Two nonoverlapping antigenic regions were defined by E2-specific neutralizing monoclonal antibodies (Fig . 6) . Monoclonal antibodies HF8-8, HB10-4, HE7-3, andJB5-6 demonstrated reciprocal competition and defined one region which was designated A(E2) (Figs . 6A-6D) . The monoclonal antibody BB7-14 defined a second region which was designated B(E2) (Fig . 6E) . The mAB JB9-3, an antibody of apparent low avidity, showed intermediate levels of competition with the BB7-14 probe . However, since a reciprocal competition experiment could not be done because of lack of conjugated (JB9-3) probe this monoclonal antibody was not assigned to a particular antigenic group . Note. Summary of data is presented in Fig . 6 . Percentage competition at the plateau : +, >70% ; ± . 30--50% ; the competitive binding experiments delineated one distinct antigenic region defined by HC10-5 and KD9-40 monoclonal antibodies and another by BD9-8C and these were designated A(E3) and C(E3), respectively . The antigenic site defined by KC4-3, which overlapped these two regions, was designated B(E3) . Since other data presented in this report indicate that TAB HC10-5 and KD9-40 bind to different epitopes, the antigenic region defined by these monoclonal antibodies was further subdivided and designated as A1(E3) and A2(E3) epitopes, specified by HC10-5 and KD9-40, respectively .

The results presented show that both BCV E2 and E3 glycoproteins can induce the production of antibodies that can efficiently neutralize virus infectivity in the absence of complement . Previous studies have shown the E2 peplomeric protein, the protein responsible for cell membrane fusion and cellular attachment, to be the major neutralizing antigen of coronaviruses, while antibodies to the El glycoprotein had little if any neutralizing ability in the absence of complement (Collins et at, 1982 ; Vautherot and Laporte, 1983 ; Talbot et al., 1984 : Wege et al ., 1984 Laude at al., 1986 ; Jimenez at al., 1986) . The coronaviruses MHV-A59, infectious bronchitis virus, and transmissible gastroenteritis virus (TGEV), some of the best-studied coronaviruses, all apparently lack an analogous protein to the BCV and human coronavirus (HCV-043) E3 glycoprotein (Sturman and Holmes, 1977 ; Rottieret al., 1981 ; Cavanagh, 1981 ; Stern et al., 1982 ; Garwes and Pocock, 1975 ; Hogue et al., 1984 ; Hogue and Brian, 1986 ; Deregt at al., in press ) . There may be an analogous protein on MHV-JHM ; however, the identity of this protein remains unclear (Wege et al., 1979 ; Siddell at al., 1981 ; Siddell, 1982 ; Taguchi at al ., 1986) . Since the E3 protein can induce efficient neutralizing antibodies, this suggests that this protein has an important function in bovine coronavirus infectivity . However, the exact nature of this function remains to be determined .

Monoclonal antibodies directed against E2 defined three antigenic regions on the BCV protein . Two of these regions can induce the production of highly efficient neutralizing antibodies and their antigenic integrity was dependent upon disulfide linkages . At least six major antigenic sites on the E2 protein of MHV-JHM were identified by competitive binding assays in previous studies . Three of these sites were involved in neutralizing virus (Talbot et al., 1984 ; Wege at al., 1984) . Four major antigenic sites were found on TGEV E2 protein by CBA . Two of these sites were responsible for most of the neutralization mediating determinants and up to six neutralizing epitopes could be distinguished by various other criteria in another study (Delmas at al., 1986 ; Jimenez et at, 1986) . These studies employed, in general, a larger panel of monoclonal antibodies to the E2 protein than reported here . Thus whether more neutralizing epitopes on the BCV E2 protein sites could have been identified had a larger panel of monoclonal antibodies been used remains to be determined . Wege et al. (1984) could distinguish three classes of monoclonal antibodies by their pattern of immunoprecipitation of E2 (gp170 and gp98) and pE2 (gp150) proteins . Only monoclonal antibodies that could bind all three proteins in the immunoprecipitation assay had high neutralizing titers . Similarly, all neutralizing monoclonal antibodies to the BCV E2 protein reported here immunoprecipitated the analogous proteins gpl90/ E2, gpl OO/E2, and gpl70/pE2 . Three antigenic domains designated A, B, and C representing four distinct epitopes could be defined on the E3 protein by the present panel of E3-specific monoclonal antibodies . Differential reactivity of E3 monoclonal antibodies to gpl 24/E3 and gp64/dE3, in radioimmunoprecipitation assays (RIPA), distinguished three epitopes : epitope Al and epitope B (defined by mABs HC10-5 and KC4-3, respectively) from each other and from the epitopes A2 and C (defined by mABs KD9-40 and BD9-8C, respectively) . The differential avidity of mAB KC4-3 for gpl24/E3 and gp64/ dE3 in this assay was considered significant and suggested that an apparent change in epitope conformation occurs when gp124/E3 is proteolytically cleaved to gp64/dE3, one which allows better binding of KC4-3 under the conditions of the assay . Assay conditions were likely important for mAB KC4-3 binding since although KC4-3 bound poorly to gp124/E3 in RIPA, it BD9-8C C appeared to bind as well as other monoclonal antibodies to E3 in antibody binding assays . Two patterns of reactivity to gp59/pE3 and gpl 18/ pE3 were observed by E3 monoclonal antibodies . These patterns distinguished epitope C defined by mAB BD9-8C from epitope A2 defined by mAB KD9-40 (and from epitopes A and B) because only mAB BD9-8C bound the precursor monomer gp59 .

Epitope C defined by mAB BD9-8C exists apparently complete on gp59/E3 which indicates the existance of two epitopes on each dimer (gpl 18/pE3 and gpl 24/ E3) . Other E3 mABs bound only the dimeric proteins indicating that epitope (Al, A2, and B) conformation was dependent on protein dimerization . However, there may be two A1, A2, and B epitopes on each E3 dimer since the behavior of these monoclonal antibodies (HC10-5, KD9-40, and KC4-3, respectively) was similar to that shown by mAB BD9-8C in antibody binding assays .

Results obtained by Western immunoblotting suggested that epitopes Al and B defined by mAB HC 10-5 and KC4-3, respectively, might be partially altered by SDS denaturation as indicated by a weak response in this assay while epitopes A2 and C, defined by mA8 KD9-40 and BD9-8C, respectively, were apparently fully resistant . Alternatively, this may indicate slight avidity differences between the E3 monoclonal antibodies . All E3 epitopes were, however, apparently destroyed when SIDS and 2-mercaptoethanol were used, suggesting that disulfide linkages are essential for their antigenicity . E3 occurs as gp62 (monomer) under reducing conditions, thus the loss of antigenicity under these conditions would be expected for epitopes Al, A2, and B, as the defining monoclonal antibodies do not bind the precursor monomer gp59 in radioimmunoprecipitation assays . Conversely, since mAB BD9-8C (epitope C) can bind the monomer gp59/pE3 it might have been expected to react with gp62/E3 in Western immunoblots . However, since this monoclonal antibody did not react with gp62/E3 and since intermolecular disulfide bonds are apparently not required for this epitope's antigenicity (as indicated by the reactivity of mAB BD9-8C for gp59/pE3) this suggests that an intramolecular disulfide bond may be involved in preservation of this epitope's conformation .

Competitive binding assays indicate that the E3 epitopes are in close spatial proximity to each other (with possible overlap) as antigenic region A (containing epitopes Al and A2) overlaps with B and B with C . Results presented indicate that epitopes A2, B, and C are located on gp64/dE3 which remains ill-defined, but apparently contains at least one of the intermolecular disulfide bond(s) bridging gp62/E3 monomers . The location of epitope A1 is speculative with regard to gp64/dE3, although its apparent proximity to A2 suggests that this epitope may be destroyed in some manner upon cleavage of gpl24/E3 .

The important antigenic region on the E3 protein involved in neutralization is apparently region A which probably contains or is closely adjacent to a biologically important domain on this protein . We have not yet examined whether these monoclonal antibodies can inhibit BCV hemagglutination . Thus, whether this biologically important domain is the same domain for erythrocyte binding (King et al., 1985) remains to be determined . Of greater interest, however, is the determination of the actual function of this protein in bovine coronavirus-cell interactions . Such studies are presently in progress .

Monoclonal antibodies define eight independent antigenic regions on the bovine leukemia virus (BLV) envelope glycoprotein

Structural poiypeptides of coronavirus (BV . 1

Simple immunization protocol for high frequency of soluble antigen-specific hybridomas

Monoclonal antibodies to murine hepatitis virus-4 (strain JHM) define the viral glycoprotein responsible for attachment and cell-cell fusion

Bovine coronavirus isolation and cultivation in continuous cell lines

Anugenic structure of transmissible gastroenteritis virus . II . Domains in the peplomer giycoprotein

Structural proteins of bovine coronavirus and their intracellular processing

The rapid determination of binding constants for antiviral antibodies by a radioimmunoassay . An analysis of the interaction between hybridoma proteins and influenza virus

The polypeptide structure of transmissible gastroenteritis virus . 1. Geri . Viral

Structural proteins of human respiratory coronavirus OC43

Antigenic relationships among proteins of bovine coronavirus, human respiratory coronavirus OC43, and mouse hepatitis coronavirus A59

Critical expitopes in transmissible gastroenteritis virus neutralization

Topographical analysis of antigenic determinants on envelope glycoprotein V3 (E) of Japanese encephalitis virus, using monoclonal antibodies

Bovine coronavirus structural proteins

Bovine coronavirus hemagglutinin protein

Cleavage of structural proteins during the assembly of the head of bacteriophage T4

Antigenic structure of transmissible gastroenteritis virus . I . Properties of monoclonal antibodies directed against virion proteins

Peroxidase-labelled antibody : A new method of conjugation

RNA-binding proteins of coronavirus MHV : Detection of monomeric and multimeric N protein with an RNA overlay-protein blot assay

Viral protein synthesis in mouse hepatitis virus strain A59-infected cells : Effect of tunicamycin

Application of the indirect enzyme-labelled antibody microtest to the detection and surveillance of animal diseases

Coronavirius JHM : Trypic peptide finger-printing of virion proteins and intracellular polypeptides

Intracellular protein synthesis

Structural analysis of virion proteins of the avian coronavirus infections bronchitis virus

Topological mapping of murine leukemia virus proteins by competition-binding assays with monoclonal antibodies

On enteropathogenic bovine coronavirus

Characterization of a coronavirus . II . Glycoproteins of the viral envelope : Tryptic peptide analysis

Isolation of coronavirus envelope glycoproteins and interaction with the viral nucleocapsid

Proteolytic cleavage of the E2 glycoprotein of murine coronavirus : Activation of cell-fusing activity of virions by tryspin and separation of two different 90K cleavage fragments

Characterization of a variant virus isolated from neural cell culture after infection of mouse coronavirus JHMV

Topographical mapping of epitopes on the glycoproteins of murine hepatitis virus-4 (JHM) : Correlation with biological activities

Characterization of Epstein-Barr nuclear antigen (EBNA) . I . A new technique for the detection of EBNA or anti-EBNA antibodies and its applicability to the study of chromosome-EBNA interactions

Utilization of monoclonal antibodies for antigenic characterization of coronavirus

Hybridoma antibodies to the murine coronavirus JHM : Characterization of epitopes on the peplomer protein (E2)

Structural polypeptides of the murine coronavirus JHM

The authors thank James E . Gilchrist and Elaine Gibbons for their valued assistance with hybridoma production, Dr . Sylya van Drunen Littel-van den Hurk for her helpful advice and assistance with HRP conjugation of monoclonal antibodies, and Irene Kosokowsky and Marilee Hagen for preparation of this manuscript . This work was supported by grants from the Medical Research Council (Canada) and Farming of the Future (Alberta) . Published with permission of the Director as journal series No . 56 .