The neutralization site on the E2 glycoprotein of Venezuelan equine encephalomyelitis (TC-83) virus is composed of multiple conformationally stable epitopes

142, 347-356

VIROLOGY

(1985)

The Neutralization Site on the E2 Glycoprotein of Venezuelan Equine Encephalomyelitis (TC-83) Virus Is Composed of Multiple Conformationally Stable Epitopes JOHN Division

of

T. ROEHRIG’

AND

JAMES

H. MATHEWS

Vector-Borne Viral Diseases, Center for Iqfedious Diseases, Centem for Disease Public Health Service, U 5’. Department of Health and Human Services, Post O&e Box 2087, Fort CoUins, Cobmd.o 80522 Received November 1, 198.4; accepted

Control,

December 21, 1984

The neutralization (N) site on the gp56 (E2) surface glycoprotein of the TC-83 vaccine strain of Venezuelan equine encephalomyelitis (VEE) virus has been characterized using monoclonal antibodies. Five new epitopes (E2d.“) were identified three of which could be mapped into the critical N site by using a competitive binding assay (CBA). Antibodies reactive with these three epitopes had either N or N and hemagglutinationinhibition activity. All epitopes contained within this N site elicited monoclonal antibodies that could protect mice from peripheral virus challenge. Antibodies reactive with the N site on other subtypes of VEE virus (IC and II) bound to, but failed to neutralize, TC-83 virus. Epitopes defined by these antibodies could be located outside of the N site on TC-83 virus by CBA. Antigenic activity of all epitopes except E2” was resistant to treatment with 2% SDS, 3% b-mercaptoethanol, or cleavage with Staph&COCCUS aureus V8 protease. Those antibodies which defined epitopes located within the N site of TC-83 with CBA bound the same V8 fragments in immunoblots. Those antibodies which defined epitopes not located within the N site bound a different set of fragments than neutralizing antibodies. These results indicate that there is a specific N site on the E2 of VEE virus which undergoes significant antigenic drift while maintaining structural and functional integrity. INTRODUCTION

with the E2 (Pedersen and Eddy, 1974; France et aL, 1979; Kinney et al, 1983). Evolutionary changes in the antigenic and biochemical structure of these proteins have been investigated (Kinney and Trent, 1982; Kinney et uL, 1983). These investigators concluded that the E2 was chemically the most variable structural protein when analyzed by tryptic peptide mapping. The C and El were more conserved. These results were consistent with the observation that the E2 elicited virus type-specific neutralizing antibody (France et cd, 1979; Kinney et aL, 1983). Monoclonal antibodies have been used to analyze the antigenic structure of a number of viruses. Our laboratory and others have been applying hybridoma technology to the characterization of the antigenic structure of the envelope glycoproteins of the alphaviruses in an at-

The alphavirus, Venezuelan equine encephalomyletis (VEE) virus is the etiologic agent of a severe encephalitic disease (Kubes and Rios, 1939; Sellers et aL, 1965). The virion contains single-stranded infectious RNA enclosed in an icosahedral nucleocapsid which is composed of multiple copies of a single capsid protein species (30,000 Da, C) (Pederson and Eddy, 1974). The nucleocapsid is enclosed in an envelope containing spikes composed of heterodimers of El (50,000 Da, gp50) and E2 (56,000 Da, gp56). For convenience and consistency, we will henceforth refer to these glycoproteins as E2 and El. The functions of virus infectivity and hemagglutination (HA) have been associated i Author addressed.

to whom

requests

for

reprints

should

be

347 0042-6822/85

$3.00

348

ROEHRIG

AND

tempt to better understand virus antigenic variation, virus virulence, and the antiviral host immune response. The antigenic structure of the El of western equine encephalitis (WEE) virus (Hunt and Roehrig, 1985); the El and E2 of the Sindbis (SIN) virus (Roehrig et al, 1980, 1982b; Schmaljohn et al, 1982; Chanas et aL, 1982; Clegg et a& 1983; Righi et al, 1983; Schmaljohn et a& 1983); the E2 of Semliki Forest virus (Boere et ak, 1983); and the El and E2 of VEE virus (Roehrig et aL, 1982b; Mathews and Roehrig, 1982) have been investigated. The results of these studies have made it increasingly apparent that the antigenicity of the El envelope glycoprotein depends on proper protein conformation (Roehrig et aZ., 1982a; Schmaljohn et al, 1982; Clegg et ak, 1983, Schmaljohn et al, 1983; Hunt and Roehrig, 1985. Some epitopes expressed on the surface of the infected cell are hidden or “cryptic” on the mature virion (Schmaljohn et aZ., 1982; Hunt et d, 1985). Protection studies using anti-E2 VEE virus monoclonal antibodies have identified a “critical” neutralization (N) site probably identical to the N activity identified with monospecific polyclonal antiserum (Pedersen and Eddy, 19’74; France et aL, 1979; Kinney et al, 1983). Antibodies specific for an epitope (E2”) located within this site also had hemagglutination-inhibition (HI) activity (Roehrig et aL, 198213). To date, no extensive studies have been published defining the relationship that protein conformation of the E2 may have to expression of its antigenicity. In the context of developing effective genetically engineered alphavirus vaccines, such a relationship must be understood. In this study, we report the identification of five new epitopes on the E2 of VEE virus. With these antibodies, we have discovered that the critical N site is composed of more than one epitope and that only those epitopes that map into this site by competitive binding assays have HA activity or react with neutralizing antibody. Unlike the El epitopes, these epitopes on VEE viruses are conformationally very stable, maintaining antigenicity after boiling in

MATHEWS

SDS and p-mercaptoethanol, or cleavage with Staphylococcus aureus V8 protease. Antigenic drift of these epitopes among VEE viruses of different subtypes was investigated. Some of these antibodies which define highly conserved E2 epitopes are efficient in protecting animals from challenge with virulent VEE virus and could possibly be used as models for a genetically engineered, cross-protective, artificial VEE vaccine. MATERIALS

AND

METHODS

Cells and viruses. The Sp2/0-Ag14 nonsecreting myeloma cell line was used for all fusions and cultivated as previously described (Roehrig et a& 1980). Viruses were grown in BHK-21 cells and plaqued in Vero cells. The representative VEE virus strains and their respective subtypes used in this study were TC-83 (IA); Trinidad Donkey (TRD; IA); PTF-39 (PTF; IB); P676 (IC); Mena II (IE); and Everglades virus, FE3-7c (EVE; II). All viruses were plaque purified once from stocks maintained at the Division of VectorBorne Viral Diseases, Centers for Disease Control, Fort Collins, Colorado. Virus puri$katim. Viruses were grown in BHK-21 cells and purified as previously described (Obijeski et al, 1974). Hybrids production. The protocol for cell fusion was a modification of the technique of Galfre et al (1977) and has been previously published (Roehrig et al, 1980). In all cases, purified virus used as immunogen was first inactivated by incubation in 4% Zwittergent (Calbiochem, La Jolla, Calif.) in 0.2 M Tris, pH 8.0, at 37” for 30 min. Live virus killed animals in the initial inoculations. Fifty micrograms of inactivated antigen mixed 1:l (v/v) in Freund’s complete adjuvant was administered subcutaneously to Balb/c/AnNCrLBr mice every other week for a total of three injections. Four days before fusion, animals were boosted intraperitoneally (ip) with 50 pg purified, live virus. Following fusion using PEG 1500, hybridomas surviving cultivation in hypoxanthine, aminopterin, and thymidine were screened for antibody by enzyme-linked immuno-

N SITE

OF

sorbent assay (ELISA) using purified virus as antigen and rabbit anti-mouse alkaline phosphatase (Roehrig et al, 1980). Cells producing antivirus antibodies were cloned in soft agar, and cloned cells were injected into pristane-primed mice to produce ascitic fluids. Hybridomu characterization and antibody pur@catim. Antibodies were characterized as to subclass by microimmunodiffusion using rabbit anti-mouse subclass antisera (Litton Bionetics, Charleston, S. C.) and lo-fold concentrates of hybridoma cell culture supernatant. Monoclonality was confirmed by isoelectric focusing (Nicolatti et aL, 1979). All antibodies demonstrated the microheterogeneity characteristic of monoclonal antibodies. Antibodies were purified from ascitic fluids by 50% ammonium sulfate precipitation and chromatography on protein A-Sepharose (Ey et c& 1978). Preparation of alkaline phosphatase-antibody conjugates followed the procedure of Voller et al, (1976). Antigen specijkity determinations. Specificity of antibodies for homologous or heterologous viruses was defined by either S. aureus protein A radioimmune precipitation (RIP) or immunoblot ELISA. For the RIP, virus-infected BHK-21 cells were labeled with [%S]methionine (1212 Ci/mmol, New England Nuclear, Boston, Mass.) from 8 to 20 hr postinfection. Infected cells were lysed with cell lysis buffer (0.2 M Tris, pH 8.0; 0.65% Triton X-100, 0.15 M NaCl) and, after pelleting the nuclei, the cell lysate was used as antigen. Antigen-antibody complexes formed after a 2-hr incubation at RT were removed with with S. aureus protein A (Pansorbin, Calbiochem, La Jolla, Calif.), precipitated antigens were analyzed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) on 10% gels, and protein bands were visualized by fluorography (Laemmli, 1970; Bonner and Laskey, 1974). Nonspecific precipitation of El and capsid were often observed. Such ambiguities in specificity could usually be cleared up by immunoblot. Immunoblot analysis was a modification of the procedures developed by Towbin et

VEE

VIRUS

349

a& (1979). Purified virus or V8 fragments were dissociated and proteins were separated by SDS-PAGE. Separated proteins were electrophoretically blotted to nitrocellulose for 3 hr at 4” and 60-V constant voltage in blotting buffer (25 mM Tris, pH 8.3, 192 mAf glycine, 20% methanol). Following blotting, bands could be visualized on the nitrocellulose by staining with 0.1% amido black in 45% methanol (Towbin et al, 1979) or in the gel by staining with silver (Wray et al, 1981). Nonreacted sites on the nitrocellulose were blocked by overnight incubation at RT with 3% BSA in 0.01 M Tris, pH 7.4, and 0.9% NaCl. Blocked blots were incubated with alkaline phosphatase-monoclonal antibody conjugate diluted 1:200 in PBS with 0.5% Tween 20. After 2 hr at RT, excess antibody was removed by three rinses in PBS-Tween 20. After 2 hr at RT, excess antibody was removed by three rinses in PBS-Tween 20 and one rinse in PBS. Antigen bands were visualized by incubation in 0.2 M Tris, pH 8.0, with 0.7 mg/ml a-naphthol phosphate, and 1.7 mg/ml fast violet (Kreuzer et aZ., 1975). Developed blots were recorded by photographing; however, the faint appearance of many bands sometimes yielded poor photographic exposures. In these cases, blots were aligned with their silverstained gels, and antibody reactive bands were noted. Serologti assays. N tests were done using 50-100 PFU/test in Vero cells, and 70% endpoints were recorded (Monath 1976). HI tests were done at pH 5.9 with goose erythrocytes and purified virus using the procedure of Casals and Brown (1954). Purified virus antigen was standardized to 4-8 HA units/test. An alkaline-phosphatase-based ELISA was used in all cross-reactivity binding assays and competitive binding assays. This procedure was a modification of that described by Voller et a& (1981) and has been previously described in detail (Roehrig et al., 1980). All serologic assays were performed with known quantities of purified antibodies to facilitate antibody titration comparisons. Competitive binding assay (CBA). The techniques for the CBA and antibody af-

350

ROEHRIG

AND

finity determinations have been published previously (Frankel and Gerhard, 1979; Roehrig et aL, 198213). Affinities of purified antibodies were approximated by endpoint ELISA using 5 pg of purified virus/well. Only antibodies of similar affinities were used in the CBA. Proteolytic digestion using 5’. aureus V8 protease. The VS-protease-digestion procedure was a modification of the protocol of Cleveland et al, (1977). Briefly, 80 pg of purified virus (0.26 mg/ml) was incubated for 3 hr at 37” with 30 pg/ml V8 protease in proteolysis buffer (125 mM Tris-HCl, pH 6.8, 0.25% SDS, 10% glycerol). The digestion was stopped by boiling in SDS-PAGE loading buffer (2% SDS, 10% glycerol, 3% b-mercaptoethanol). Digests were analyzed by electrophoresis on 15% polyacrylamide gels for 3.5 hr at 40 mA constant current. Bands were visualized by silver staining, and molecularweight estimations were made by coelectrophoresis with marker proteins of known molecular weight (BSA, 66,000 Da; egg albumin, 45,000 Da; carbonic anhydrase, 30,000 Da; soybean trypsin inhibitor, 20,000 Da; and cytochrome c, 12,000 Da; Sigma Chemical, St. Louis, MO.). Protection of mice bp passive transfer of monoclonal antibodies. The conditions and characteristics for passive antibody protection studies have been published in detail elsewhere (Mathews and Roehrig, 1982). RESULTS

Isolation and characterization of hybridom~.~. Five fusions were performed using mice individually immunized with prototypes of any of one of the documented VEE virus subtypes IA, IB, IC, IE, and II. Seventy-five stable hybridomas were isolated, characterized as to class, isotype, antigen specificity, homologous HI and N activity, and virus cross-reactivity. No anti-IE hybridomas were used for these experiments. Results are presented for representative antibodies defining eight unique epitopes (Table 1). Identification of gp56” (E2”), gp56b (E2b), and gp56’ (E2e) has been previously published (Roehrig

MATHEWS

et al, 1982b). Antibody specificities for protein components of homologous and heterologous viruses were determined by RIP or immunoblot analysis with homologous and heterologous antigens. In all cases, these antibodies were specific for the E2 for both homologous and heterologous viruses. It was of some interest to note the variations in reactivity patterns for some of the antibodies tested by RIP. Representative results for these reactivities are shown in Figs. 1 and 2. Anti-EZ’, E2d, and E2g monoclonal antibodies precipitated intracellular ~65 and E2 with all viruses tested. These antibodies could also recognize SDS and P-mercaptoethanol-denatured E2 from purified viruses in immunoblots (Fig. 2). A different pattern of reactivity was noted for anti-E2”, E2’, and E2h monoclonal antibodies. These antibodies were able to recognize the E2 in immunoblots with purified viruses (data not shown), but the antigenic structure of the E2 was apparently different in TC-83infected cells because these antibodies could primarily react only with the ~65 of TC-83 virus (Fig. 1, lane G). They were reactive with both ~65 and E2 of all other viruses tested (Fig. 1, lanes H, I, and J). Biological activities of the monoclonal antibodies. We used monoclonal antibodies in the N and HI test with homologous and heterologous viruses to determine the interaction of these defined epitopes in virus infectivity (Table 1). Antibody specific for the E2” epitope was the most potent neutralizer of virus infectivity. This epitope has been previously defined as being in the critical N site on VEE virus (Roehrig et ok, 1982b; Mathews and Roehrig, 1982). Antibodies reactive with the newly identified epitopes also had N and HI activity. A unique observation was the ability of certain epitopes to selectively lose biological function. For example, the antibody defining the E2” site bound quite well to viruses of the IA, IB, and IC subtypes, but only efficiently neutralized the infectivity of the IC virus. Antibodies defining the E2d site bound well to subtype IA, IB, and II viruses but only efficiently neutralized the subtype II virus. Similarly, some antibodies could have both HI and

N SITE

OF

VEE

TABLE CHARACTERISTICS

VIRUS

351

1 OF E2

EPITOPES VEE

Virus

Representative antibody

Subclass

TM.3

5B4D-6’

IgG2A

TC-83

PTF

EVE

2A4B-12

lA4A-1

lA6C-3

Epitope specificity E2’

IgGZB

ELISA

E2b

IgG2A

lA3A-5

IgGPA

lA4D-1

IgG2A

<40 ND ND

<40 ND ND

HI N

1280 <4 <4

1280 c4 <4

1280 %4 <4

<40 ND ND

<40 ND ND

640 ND ND

<40 ND ND

HI N

5120 28192 4096

1280 28192 38192

5120 38192 38192

2560 38192 4096

640 128 2048

<40 r4 <4

1280 28192 28192

5120 c4 32

5120

5120

8 64

4 16

160 <4 <4

<40 <4 <4

<40 <4 <4

5120 1024 8192

5120

2560 32 64

2560 16 16

1280 1024 2048

800 c4 c4

800 <4 <4

<40

1280 1024 256

5120 512 512

1280 4 512

640 4 4

<40 <4 64

<40 c4 s4

2560 128 2048

1280 64 2048

1280 <4 1024

1280 16 512

1280

640 %4 4

1280 16 2048

1280 16 2048

1280

2560

640 4 1024

4 32

ELISA

1280 HI N

PTF

lA3A-9

IgG2A

E2O

8 512

ELISA

5120 HI N

PTF

lA3B-7

IgG2A

E2”

8 256

ELISA

1280 HI N

’ Antibodies ‘Reciprocal ‘Not done;

were purified and endpoint titers. antibodies negative

used

at a concentration

in ELISA

were

II EVE

<40 ND ND

ELISA

E2’

IE Mena II

<40 ND” ND

HI N TRD

ID 3880

<40 64 <4

ELISA

E2’

IC P676

<40 <4 <4

HI N P676

IB PTF

1280’ 64 64

ELISA

E2’

IA TRD

HI N ELISA

E2’

IgG2A

IA TC-83

Test

subtype

4 256 of 1.0 mg/ml

not tested

N activity with some viruses but have altered N or HI activity with other viruses. Anti-E2’ antibody had both N and HI activity on TRD and PTF virus, but only N activity on TC-83 and P6’76 viruses. It had very little biologic activity on 3880 virus, even though it bound quite well to this virus (Table 1). Spatial arrangement of epitopes. We performed CBAs on TC-83, P676, and EVE viruses using these representative antibodies in an attempt to ascertain the spatial arrangement of these epitopes and to determine if this arrangement affects the expression of biological function. The results indicated that the E2 of TC-83 could be divided into four CBA groups

in HI, N, and

in HI or N with

heterologous

4 2048

4 4

4 1024

4 16

1280 4 a8192

ELISA. viruses.

(Table 2 and Fig. 3). Those antibodies that competed well with anti-E2” antibody (e.g., anti-EZ’, EZg, E2’) had either N or N and HI activity. Those antibodies which mapped epitopes distal from the E2” (e.g., anti-EZb, EZd, and E27 were able to bind to virus, but had neither HI nor N activity on TC-83 virus. If the CBA was repeated using either P676 or EVE, it could be demonstrated that the E2” and EZd now mapped proximal to the E2” site on the homologous viruses (Table 2, Fig. 3). With these homologous viruses (P676 and EVE), antibodies specific for these epitopes acquired N and HI activity (Table 1). The expression of N and HA by a particular epitope appeared, therefore, to corre-

352

ROEHRIG

ABCDEFfiHI

AND

J

P6b

E2 El

FIG. 1. antibodies antibodies lysates of

Specificity testing of anti-E2 monoclonal by radioimmune precipitation. Monoclonal were reacted with [%]methionine-labeled virus-infected cells and precipitated with Staph&nmccus aureus as described under Materials and Methods. The precipitated proteins were analyzed by SDS-PAGE. Lane A contains marker proteins of TC-33 virus. The following are precipitations with anti-E2B monoclonal antibody (lA3A-9) and cell lysates from lane (B) TC-33, (C) PTF, (D) P676, (E) 3890, and (F) Mena II infected cells. The following are precipitations with anti-E2h monoclonal antibody (lA3B-7) and cell lysates from (G) TC-83, (H) PTF, (I) P676, and (J) Mena II infected cells.

late to the proximity the epitope had to the E2”. Linhq7e of epbpe.9 &terminQd bg partial proteolgsis. To corroborate the hypothesis that proximity to the E2” was necessary for expression of N or HA, we produced peptides of various VEE viruses using partial proteolysis with V8 protease (Fig. 4). We hypothesized that if antibodies which mapped distal from the E2” in CBAs recognized peptides other than those recognized by anti-E2” antibodies, this would be more direct evidence that these epitopes were not physically linked in the protein. All of these antibodies except for anti-E9 were similar to anti-ET antibody in that they could recognize antigen in immunoblots (data not shown). Because of the instability of the E2d, anti-ET (lA4A-1) and anti-ET (lA3A-5) were chosen to investigate this hypothesis. A representative immunoblot using anti-ET and V8 digests of subtypes IA, IB, IC, ID, IE, and II VEE viruses is shown in Figs. 4B.

MATHEWS

Anti-E2” and anti-E2” MAbs recognize the same peptides in P676 virus V8 digests (Fig. 5). This result indicates that the E2” and E2” epitopes are closely linked on the P676 virus E2. With TC-83 virus, however, while anti-E2” MAbs recognize peptides from V8 digests, anti-ET MAbs cannot react with any V8 peptide. Most probably, then, E2” and E2” epitopes are not closely linked on the TC-83 virus E2. Protective capacity of the anti-E2 mow clonal antibodies. Because some of these newly identified epitopes appeared to function in virus infectivity and were more conserved throughout the VEE complex than the E2’, their potential as possible models for a cross-protective genetically engineered artificial vaccine was considered. We were interested in determining how antibodies specific for these epitopes would protect animals from a lethal challenge with VEE virus. We had previously demonstrated that anti-E2 antibody was extremely effective in protecting animals from VEE virus challenge, requiring as little as 5 pg to abrogate disease (Mathews and Roehrig, 1982). In this study, mice

A

6

C

D

E

F

-E2 -El

FIG. 2. Immunoblot specificity testing with antiE2 monoclonal antibody (lA4A-1). SDS-PAGE gel analysis showing reactions with (A) EVE, (B) 3939, (C) P676, (D) PTF, (E) TRD, and (F) TC-93 purified viruses following immunoblot analysis as described under Materials and Methods.

N SITE

OF

VEE

TABLE MAPPING

353

2

OF E2 EPITOPES BY COMPETITIVE Alkaline

BINDING

phosphatase

ASSAY

conjugate

epitope

specificity

Antibody competitor

E2”

E2j

E2”

E2O

E2”

E2*

E2d

E2”

E2” E2f E2” E2’ E2’ E2* E2d E2”

lA4A-1” lA4D-1 lA3B-7 lA3A-9 5B4D-6 2A4B-12 lA6C-3 lA3A-5

+* + + + + -

+ + + + + -

+ + + + + + -

+ +

+ + + + +

-

-

+ + + + -

-

-

-

+ + + + + -

+ + +

+ + +

P676

E2” E2h E2’ E2”

lA4A-1 lA3B-7 lA3A-9 lA3A-5

+ + + +

ND” ND ND ND

+ + + +

+ + + +

NRd NR NR NR

NR NR NR NR

NR NR NR NR

+ + + +

EVE

E2c E2h E2d

lA4A-1 lA3B-7 lA6C-3

+ + +

NR NR NR

+ + +

NR NR NR

NR NR NR

NR NR NR

+ + +

NR NR NR

(160640);

“-”

Antigen TC-83

Epitope specificity

VIRUS

a Antibody was purified and used at a concentration of 1 mg/ml. *Reciprocal endpoint titers were converted to ‘I+/-” using the following (G40). ‘Not done. d Not reactive in ELISA.

were passively immunized iv with 20 pg/ animal of purified antibody. Twenty-four hours later, the animals were challenged with 100 ip LDm of virus. Survival was monitored (Table 3). Protection was related to the N pattern of these antibodies with heterologous viruses (see Table 1). It is of interest that a vaccine constructed with the E2g and E2h epitopes would probably confer protection against all of these VEE subtypes. DISCUSSION

In our previous studies with VEE virus, we identified three epitopes on the E2 glycoprotein (Roehrig et al, 1982b). Of these epitopes, only antibodies specific for the E2” and N or HI activity and were also able to protect animals from a lethal challenge with virulent VEE virus (Mathews and Roehrig, 1982). In this study, we have identified five new E2 epitopes (E2d-h). Antibodies specific for each of the

scheme:

“+”

new epitopes identified possessed either N or HI and N activity with various VEE viruses. This level of biological activity was less than that demonstrated by antiE2” monoclonal antibody; however, most were efficient in blocking virus infectivity. It was of interest to note that antibodies could maintain similar binding characteristics to two different viruses in ELISA, but vary in their ability to either block hemagglutination or infectivity of the virions. Examples of this variation was the antibody specific for E2” epitope. Anti-E2e antibody reacted to identical titer with TC-83, TRD, PTF, and P676 virus when measured in ELBA. A similar cross-reactivity analysis using either HI or N revealed that this antibody had HI or N activity only for P676 virus. Similar ‘variations in reactivity were evident for antiE2d, E2’, E2%, and E2h antibodies. Antibodies specific for epitopes E2’, E2g, and E2h demonstrated differential reactivities with TC-83 and TRD viruses. One

ROEHRIG

354

Lost E2

rl

AND

MATHEWS

IWctivities: a,b,d

E2=

,EtE,

E2 a,b,e,f,g

E2d

E2h

FIG. 3. CBA maps of the E2 for subtypes IA (TC83), IC (P676), and two (EVE) VEE viruses as determined with the antibodies listed in Table 2. Lost reactivities are those epitopes that cannot be identified on each respective virus.

antigenic difference between TC-83 and TRD viruses had been previously identified in the E2” epitope (Roehrig et cd, 1982b). Anti-E2’ and anti-EZg antibodies had HI activity with TRD virus and not with TC83 virus. Anti-E2g and E2h antibodies were lo-fold more reactive in N tests with TRD viruses than with TC-83 virus. Another difference between TC-83 and the other

FIG. 4. SDS-PAGE and immunoblot analysis of V8 fragment with anti-E20 (lA4A-1) monoclonal antibody. (A) Purified viruses were digested with V8 protease and analyzed by SDS-PAGE on 15% gels and silver stained. Lane (A) TC-83, (B) TRD, (C) PTF, (D) P676, (E) 3886, (F) Mena II, and (G) EVE digests. (B) Immunoblot with fragments from lane (A) TC-83, (B) PTF, (C) P676, (D) 3880, (E) Mena II, and (F) EVE digests. The numbers signify the migration of molecular-weight markers in kilodaltons.

FIG. 5. Fragment cross-reactivities of anti-ET and E2e monoclonal antibodies. Silver-stained V8 digest of (A) TC-83 virus and (B) P676 virus. Open arrows indicate fragments reactive with anti-E2 (lA4A-l), closed arrows indicate fragments reactive with antiE2” (lA3A-5) antibodies following immunoblot analysis.

VEE viruses was the differential precipitation of TC-83 E2 in the RIP (Fig. 1). The reduced reactivity of intracellular E2 with anti-E2e, E2’, and E2h antibodies compared to p65 reactivity seemed to indicate that these epitopes were fully expressed on the mature virion (antibodies specific for these epitopes neutralized virus) and the E2 precursor, ~65, but were altered on the intracellular form of the E2. The results presented here also indicate that the antigenic reactivity of these E2 epitopes was particularly stable to conformational changes in protein because reactivity was maintained after denaturation with SDS and B-mercaptoethanol, or cleavage with V8 protease. Only the E2d epitope lost reactivity after denaturing. Therefore, the reactivity of these epitopes does not appear to rely on charge

N SITE

OF

interactions or disulfide bonding. This result is radically different from that found with the El of WEE virus. In that case, the antigenic reactivity of all but one epitope was destroyed by treatment with 0.5% SDS (Hunt and Roehrig, 1985). Similar observations have been reported for the El of SIN virus (Roehrig et al, 1982a; Schmaljohn et aL, 1983). The results from the CBAs and immunoblots with V8 protease virus digests demonstrate that there was a critical spatial site on the E2 that was involved in HA or virus infectivity. For our purposes, we have considered the E2” epitope to be the hub of this critical site due to the extremely high HI and N activities demonstrated by anti-ET antibodies. Antibodies that compete with anti-E2” antibodies usually had HI and N activity regardless of which virus antigen was used in the CBA (Tables 1 and 2; Fig. 3). Those antibodies which failed to compete with anti-E2’ antibodies usually had no NI or N activity. The exception to these observations was anti-E2” monoclonal antibodies: this epitope was present only on TC-83 virus and probably represented an unusual reactivity, peculiar for this virus. The result that anti-E2” and anti-E2” reacted with the same V8 fragments for P676 virus but not TC-83 virus, indicated that these epitopes were not linked on TC-83 virus but were linked on P676 virus. The observation that there is a structurally critical domain on the glycoprotein that is involved in virus adsorption to susceptible cells has been demonstrated for influenza virus (Wilson et al, 1984). The observation that was most interesting in this study was how glycoproteins could vary dramatically in their antigenic reactivities (Table 1) and yet probably conserve overall structure-function interactions at the N site. The identification of antigenically stable, cross-reactive protective epitopes (E2e, E26, and E2h) was a promising observation within the context of designing a genetically engineered artificial vaccine (Table 3). Current efforts in this laboratory are centering on preparing, isolating, and N-terminal sequencing antibody-reactive

VEE

355

VIRUS

TABLE

3

PROTECTIONFROM VEE VIRUS CHALLENGE FOLLOWING PASSIVE TRANSFER OF ANTI-ES MONOCLONAL ANTIBODIES Challenge virus Antibody

Epitope

3B4C-4

E2C

lA4A-1 lA3A-9 lA3B-7 PBS

E2C E2’ E2h -

TRD lO/lO” lO/lO lO/lO lo/lo o/10

P676 Mena II lO/lO ND* lO/lO lO/lO 2/10

o/10 ND lO/lO lO/lO o/10

EVE lO/lO ND NR” 10110 o/10

o Survivors/total NIH outbred Swiss mice immunized with 20 gg/mouse of purified antibody and challenged with 100 ip LDw virus. *Not done. ’ Not reactive.

fragments similar to the ones reported here. This information, when taken in the context of the derived amino acid sequence from the cloned E2 gene, should allow us to model the glycoprotein and target our efforts to develop improved antiviral prophylaxis. ACKNOWLEDGMENTS

The authors thank Mrs. Susan Ure for her excellent secretarial assistance in the preparation of this manuscript and Mr. R. M. Kinney for his assistance in virus purification. REFERENCES

BOERE, W. A. M., BENAISSA-TROUW,~B. J., HARMSEN, M., KRAAIJEVELD, C. A., and SNIPPE, H. (1983). Neutralizing and non-neutralizing monoclonal antibodies to the E2 glycoprotein of Semliki Forest virus can protect mice from lethal encephalitis. J. Gen

Viral

64.1405-1408.

BONNER, W. M., and LASKEY, R. (1974). A film method for tritium-labeled proteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem. 46, 83-88.

CASALS, J., and BROWN, L. V. (1954). Hemagglutination with arthropod-borne viruses. J. Exp. Med 99,429-499.

CHANAS, A. C., GOULD, E. A., CLEGG, J. C. S., and VARMA, M. G. R. (1982). Monoclonal antibodies to Sindbis virus glycoprotein El can neutralize, enhance infectivity, and independently inhibit haemagglutination or haemolysis. J. Gen Viral 58, 37-46.

CLEGG, J. C. S., CHANAS, A. C., and GOULD, E. A. (1983). Conformational changes in Sindbis virus

356

ROEHRIG

AND MATHEWS

El glycoprotein induced by monoclonal antibody binding. J. Gen viral 64. 1121-1126. CLEVELAND, D. W., FISCHNER, S. G., KIRSCHNER, M. W., and LAEMMLI, Il. K. (19’77). Peptide mapping by limited proteolysis in sodium dodecyl sulfate and analysis by gel electrophoresis. J. Biol Chum. 252, 1102-1106. EY, P. L., PROWSE, S. J., and JENKIN, C. R. (1978). Isolation of pure IgGl, IgGLa, and IgG2b immunoglobulins from mouse serum using protein A-sepharose. Immunochemistry 15,429-436. FRANCE, J. K., WYRICK, B. C., and TRENT, D. W. (1979). Biochemical and antigenic comparisons of the envelope glycoproteins of Venezuelan equine encephalomyelitis virus strains. J. Gen lrird 44, 725-740. FRANKEL, M. E., and GERHARD, W. (1979). The rapid determination of binding constants for antiviral antibodies by a radioimmunoassay. An analysis of the interaction between hybridoma proteins and influenza virus. Mel Immund 16,101-105. GALFRE, G., HOWE, S. C., MILSTEIN. C., BUTCHER, G. W., and HOWARD, J. C. (1977). Antibodies to major histocompatibility antigens produced by hybrid cells. Nature (London) 266, 550-552. HUNT, A. R., and ROEHRIG, J. T. (1985). Biochemical and biological characteristics of epitopes on the El glycoprotein of western equine encephalitis virus. virdogy 142, 334-346. KINNEY, R. M., and TRENT, D. W. (1982). Conservation of tryptic peptides in the structural proteins of viruses in the Venezuelan equine encephalitis complex. nh 121, 345-362. KINNEY, R. M., TRENT, D. W., and FRANCE, J. K. (1983). Comparative immunological and biochemical analyses of viruses in the Venezuelan equine encephalitis complex. J. Gen ViroL 64, 135-147. KREUZER, K., PRATT, D., and TORRIANI, A. (1975). Genetic analysis of regulatory mutants of alkaline phosphatase of E. Coli Genetics 81,459-468. KUBES, V., and RIOS, F. A. (1939). The causative agent of infectious encephalomyelitis in Venezuela. Scitmce

(Washingtun,

D. C) X3.20-21.

LAEMMLI, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (Lcnzdan) 277, 680-685. MATHEWS, J. H., and ROEHRIG, J. T. (1982). Determination of the protective epitopes on the glycoproteins of Venezuelan equine encephalomyelitis virus by passive transfer of monoclonal antibodies. J. Immunol 129,2763-2767. MONATH, T. P. (1976). Togaviruses, Bunyaviruses, and Colorado tick fever. In “Handbook of Clinical Microbiology” (N. R. Rose and H. Friedman, eds), pp. 456-462. American Society for Microbiology, Washington, D. C. NICOLATPI, R. A., BRILES, D. E., SCHROER, J., and DAVIE, J. M. (1979). Isoelectric focusing of immunoglobulins: Improved methodology. J. Immund Methods 33, 101-115.

OBJJESKI, J. F., MARCHENKO, A. T., BISHOP, D. H. L., CANN, B. W., and MURPHY, F. A. (1974). Comparative electrophoretic analysis of the virus proteins of four rhabdoviruses. J. Ga vird 22, 21-22. PEDERSEN, C. E., and EDDY, E. A. (1974). Separation, isolation, and immunochemical studies of the structural proteins of Venezuelan equine encephalomyelitis virus. J. fir& 14, 740-744. RIGHI, M., RADAELLI, A., CASTAGNOLI, P. R., LIBOI, E., and DE GIUL MORGHEN, D. (1983). Identification by monoclonal antibodies of a new epitope in the glycoprotein complex of Sindbis virus. J. fiti Methods 6, 203-214. ROEHRIG, J. T., CORSER, J. A., and SCHLESINGER, M. J. (1980). Isolation and characterization of hybrid cell lines producing monoclonal antibodies directed against the structural proteins of Sindbis virus. Virw 101, 41-49. ROEHRIG, J. T., GORSKI, O., and SCHLESINGER,M. J. (1982a). Properties of monoclonal antibodies directed against the glycoproteins of Sindbis virus. J. Gen

ViroL

59,421-425.

ROEHRIG, J. T., DAY, J. W., and KINNEY, R. M. (1982b). Antigenic analysis of the surface glycoproteins of a Venezuelan equine encephalomyelitis virus (TC-83) using monoclonal antibodies. virolog 118.269-278. SCHMALJOHN, A. L., JOHNSON, E. D., DALRYMPLE, J. M., and COLE, G. A., (1982). Nonneutralizing monoclonal antibodies can prevent lethal alphavirus encephalitis. Nature (London) 297, 70-72. SCHMAIJOHN, A. L., KOKUBUN, K. M., and COLE, G. A. (1983). Protective monoclonal antibodies define maturational and pH-dependent antigenic changes in Sindbis virus El glycoprotein. Virdogy 130,144-154.

SELLERS, R. F., BERGOLD, G. H., SUAREZ, 0. M., and MORALES, A. (1965). Investigations during Venezuelan equine encephalitis outbreaks in Venezuela-1962-1964. Amer. .I !Frop. Med H?/s. 14, 460469. TOWBIN, H., STAEHEXIN, T., and GORM)N, J. (1979). Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: Procedure and some applications. Proc Nd Acud Sci USA 76,4350-4354.

VOLLER, A., BIDWELL, D., and BARTL~XT, A. (1976). Microplate enzyme immunoassay for the immunodiagnosis of virus infections. In “Handbook of Clinical Immunology” (N. R. Rose and H. Friedman, eds.), pp. 506-512. American Society for Microbiology, Washington, D. C. WILSON, I. A., NIYAN. H. L., HOUGHTEN, R. A., CHERENSON.A. R., CONROILY, M. L., and LERNEX, R. A. (1984). The structure of an antigenie determinant in a protein. Cell 37.767-778. WRAY, W., BOULIKAS, T., WRAY, V. P., and HANCOCK, R. (1981). Silver staining of proteins in polyacrylamide gels. And Biockem 118,197~203.