- Email: [email protected]
Protective monoclonal antibodies define maturational and pH-dependent antigenic changes in sindbis virus E1 glycoprotein
130, 144-154 (1983)
VIROLOGY
Protective ALAN
Monoclonal Antibodies Define Maturational and pH-Dependent Antigenic Changes in Sindbis Virus El Glycoprotein
L. SCHMALJOHN: Department
KRISTINA
M. KOKUBUN,
AND
GERALD
A. COLE
of Microbiology, University of Maryland School of Medicine, 660 West Redwood Street, Baltimore, Maryland 21201 Received April
12, 1983’; accepted July 11, 1983
Monoclonal (MC) antibodies specific for either the El or E2 glycoproteins of Sindbis virus (SIN) were used to probe for differences in the surface topography of SIN epitopes between infected cells and mature virions. Employing an enzyme-linked immunosorbent assay (ELISA) in which binding of individual peroxidase-labeled MC antibodies to immobilized (solid-phase) detergent-disrupted SIN was inhibited specifically by one or more unlabeled antibodies, viral epitopes could be grouped into six spatially distinct antigenic sites-five on El, designated a through e, and one site on E2. All six sites were represented on the surfaces of SIN-infected cells as shown by the complement (C)-dependent lysis mediated by antibodies of the corresponding epitope specificities. In contrast, virus-neutralizing (NT) activity was restricted to antibodies specific for epitopes on E2 and on site c of El, irrespective of the presence of added C’ and an antiserum against mouse immunoglobulins. That El sites a, b, d, and e became inaccessible to antibody binding was shown by a competitive-inhibition ELISA. Whereas all MC antibodies were inhibited from binding to solid-phase SIN when premixed with detergent-treated virions, only those having NT activity could be competitively inhibited by intact virions. Sites El-d and Ele could be exposed not only by detergent disruption but also by lowering the virion pH from 7.2 to 6.0. These collective results indicate that a majority of immunologically relevant El epitopes present on SIN-infected cell surfaces become cryptic during SIN maturation and, except at low pH, remain undetectable on virion surfaces. INTRODUCTION
Sindbis virus (SIN), like other members of the Alphmkm genus, matures by budding at host-cell plasma membranes. Alphavirus structure and morphogenesis are reasonably well understood from the comprehensive studies of SIN and Semliki Forest virus (reviewed by Schlesinger and Klilrianen, 1980; Garoff et al, 1982). The fully assembled SIN virion, approximately 65 nm in diameter, is a relatively simple particle which contains only three structural proteins, a lipid envelope, and a single 42 S molecule of genomic RNA. Two of the viral proteins, designated El and E2, are envelope glycoproteins, each approximately 50,000 Da (50K), that together form the external virion spikes. The internal r To whom reprint
requests should be addressed.
0042-6822/83 $3.00 Copyright All rights
0 1983 by Academic Press, Inc. of reproduction in any form reserved
nucleocapsid contains the genomic RNA within an icosahedral shell composed of 30K capsid (C) protein subunits. Following glycosylation and translocation of the envelope proteins to the cell surface, nucleocapsids align along the cytoplasmic face of the plasma membrane and virus budding occurs. This final stage of morphogenesis is a highly ordered process; host proteins are excluded from the viral envelope and the three structural proteins are incorporated into virions apparently in equimolar amounts. When viral structure and morphogenesis are viewed in the context of protective antibody responses to SIN and other alphavirus infections, the nature and locations of viral epitopes which induce such antibodies are of particular relevance. Since El and E2 are present on plasma 144
ANTIGENIC
CHANGES
membranes of SIN-infected cells as well as on virion envelopes (Sefton et aZ., 1973; Smith and Brown, 19’77), protective antibodies specific for either glycoprotein might be expected to bind at both locations. Epitopes on the surfaces of virions, especially those to which binding by antibodies results in neutralization of infectivity in vitro, are usually considered the “targets” of humoral immune responses which mediate host recovery (Casals, 1963). However, antibodies induced by alphavirus epitopes expressed exclusively on host-cell surfaces also appear to be important in protective immunity. As reported recently, monoclonal (MC) antibodies several against El of SIN, although lacking neutralizing (NT) activity against SIN in vitro, were capable of passively conferring protection to adult mice against an otherwise lethal intracerebral SIN challenge. Protection by these antibodies correlated with their ability to mediate complement (C’)dependent lysis of SIN-infected cells in vitro (Schmaljohn et aL, 1982). Based on current knowledge of alphavirus morphogenesis, the existence of such anti-El antibodies would not have been anticipated since the only covalent modification of SIN envelope glycoproteins thought to occur at the plasma membrane is the cleavage of PE2, the precursor of E2 (see reviews cited above). It was therefore surprising to find that certain El epitopes were demonstrable in assays for cytolytic but not NT antibodies. That certain antiEl MC antibodies might bind to the surfaces of virions without reducing their infectivity seemed unlikely after it was found that these antibodies had no NT activity in the presence of c’ and anti-immunoglobulin antibodies. Data presented below establish that several immunologically important El epitopes are present on infected-cell surfaces and on detergent-solubilized viral proteins but are undetectable on virion surfaces. MATERIALS
AND
Virus propagation Sindbis virus, strain
METHODS
and puti$cation. AR339, was main-
IN SINDBIS
VIRUS
145
tained at -80” as an infected suckling mouse brain suspension (seed virus) and was propagated in BHK-21 (hamster kidney) cells. Growth medium for BHK-21 cells consisted of RPM1 1640 containing 10% fetal bovine serum (FBS), penicillin (100 U/ml), streptomycin (100 pg/ml), gentamycin (40 pg/ml), and 10 mM HEPES. Subconfluent monolayer cultures were infected with seed virus at an m.o.i. of 0.1 PFU/cell in a small volume of growth medium which, following a 60-min adsorption at 37”, was replaced with medium containing 2% FBS. After cultures were incubated for 24-36 hr at 37”, the medium was harvested and clarified by centrifugation .for 10 min at 10,000 g. Virus was concentrated by precipitation with polyethylene glycol, layered onto a 15-40% potassium tartrate gradient, and centrifuged to equilibrium, usually overnight, at 22,000 rpm in a Beckman SW 27.1 rotor. The virus band was collected, assayed for protein content (BioRad protein assay), and used immediately or stored at -80”. Mwrmclmal antibodies. Infectious SIN, either as purified virus or as a suckling BALB/c mouse brain suspension, was used to immunize BALB/c mice for each of four fusions between immune spleen cells and the P3 X 63Ag8 (IgGl, kappa secretor) myeloma cell line. The fusion, selection, and cloning procedures were the same as those described for anti-Dengue virus MC antibodies by Gentry et al. (1982). The specificity of each MC antibody was determined by solid-phase radioimmunoassay using purified individual proteins of SIN (Dalrymple et aL, 1976) and subsequently confirmed by immunoprecipitation of 3H-labeled virion proteins followed by SDSpolyacrylamide gel electrophoresis, as described by Smith and Pifat (1982). Immunoglobulin heavy-chain isotypes of MC antibodies were determined by immunodiffusion; culture supernates from each hybrid cell line were tested against immunoglobulin class- and subclass-specific (IgM, IgG3, IgGl, IgGBb, IgGBa, and IgA) antisera (Litton Bionetics). MC antibodycontaining ascitic fluids were prepared in Pristane-primed BALB/c mice and used in all other assays because of the greater an-
146
SCHMALJOHN.
KOKUBLJN.
tibody concentrations (l-10 mg/ml) in ascitic fluids. Neutralization assays. Plaque-reduction neutralization (PRNT) assays of suckling mouse brain seed virus were performed as previously described (Schmaljohn et aL, 1982) by preincubating SIN (50-100 PFU/ 0.1 ml) with dilutions of antibodies and then measuring residual infectious virus by plaque assay on BHK-21 cells. All PRNT assays shown were performed in the presence of added 5% fresh guinea pig serum (Cedarlane) as a source of c’, for some assays, a rabbit antiserum against mouse immunoglobulins (Cappel) was added (final dilution of 1:200) and the mixture of virus and antibodies was incubated for an additional 45 min at 37” before adsorption onto BHK-21 monolayers. PRNT assays were also performed with Vero (monkey kidney) cells to assure that results were not unique to the BHK-21 cell assay system; no significant differences were observed between the two cell types. Chromium-release assaya To demonstrate C-dependent antibody-mediated cytolysis, Vero cells were infected (or mock infected) with seed virus at an m.o.i. of 1 and incubated for 14 hr at 37”. Cells were trypsinized, washed, labeled with 51Cr (150 &i/lo7 cells), washed, and seeded into flatbottomed 96-well tissue culture plates (Costar) as previously described (Schmaljohn et aL, 1982). Each well received a lOO~1 aliquot containing 2.5 X lo4 cells in RPM1 1640 supplemented with 10% FBS. Plates were incubated for 60 min at 37” to assure regeneration of cell-surface viral antigens that may have been removed by trypsinization. Rabbit C’ (Cedarlane Low-Tox M) was added in 50-~1 aliquots, followed by dilutions of heat-inactivated MC antibodies, also in 50-~1 aliquots. The final serum C’ content in each well was 2% (v/v). All tests and controls were done in triplicate. Plates were incubated for 3 hr at 37” and then 100 ~1 of supernate was removed from each well and counted in a gamma counter; SEM for triplicate wells did not exceed 5% of the mean counts per minute (cpm). The percent specific 51Cr release was calculated for infected and uninfected Vero cells as % specific release = (test,rm - control,,,)
AND
COLE
X lOO/maximum releasable,,, where the denominator is the number of cpm released from cells treated with 1% Triton X-100. Complement controls, which spontaneously released 13 and 23% of maximum releasable cpm for uninfected and SIN-infected Vero cells, respectively, were used in the above formula for all test systems containing C’. Medium controls spontaneously released 12 and 18%, respectively, for uninfected and SIN-infected Vero cells. The lytic titer of each antibody was taken as the highest dilution that caused at least 10% specific release; specific release values for normal ascitic fluids and anti-nucleocapsid MC antibodies never exceeded 3%. Qualitatively, results of cytolysis assays were not dependent upon the host-cell species nor upon the C’ source because similar results were obtained with BHK-21 cells and guinea pig C’ (Schmaljohn et ul, 1982). Antibody quantitation. Antigen-containing plates for enzyme-linked immunosorbent assay (ELISA) were prepared by disrupting gradient-purified SIN (400-900 pg/ ml) with 1.0% Triton X-100, diluting to 10 pg/ml in phosphate-buffered saline (PBS), pH 7.2, and then distributing 50 &well into 96-well polystyrene plates (Costar 3590); plates were evaporated to dryness overnight in a biological safety cabinet. The antibody assay was essentially the same as that described by Kennett (1980). Briefly, wells were “blocked” with 5% FBS and washed; dilutions of antiviral antibodies in PBS containing 5% FBS were added, in duplicate, in volumes of 50 pi/well. Plates were incubated overnight at 4°C and then washed four times with PBS containing 0.02% Tween-20 and 0.5% FBS. Horseradish peroxidase-conjugated goat immunoglobulin (Cappel), reactive against all mouse immunoglobulin classes, was diluted 1:lOOO in the above washing solution and added (50 ~1) to each well. Plates were incubated for 90 min at room temperature with constant rocking and then washed four times. Freshly prepared o-phenylenediamine substrate solution was added to each well in a volume of 100 ~1 and plates were rocked for 30 min at room temperature. Reactions were stopped by the addition of 100 ~1 of 0.1 M NaF and the optical
ANTIGENIC
CHANGES
density (OD) for each well was read at 450 nm with a Dynatech MicroELISA II automated plate reader. The ELISA titer for each antibody was taken as the dilution of antibody resulting in an OD of 0.5 which, by comparison with mouse immunoglobulin standards, corresponded to approximately 5 ng of bound antibody per well; therefore, an ELISA titer of 104.’ represented (for high-affinity antibodies) approximately 1 mg/ml of specific antibody in the undiluted sample.
Competitive inhibition by unlabeled antibodies. A competitive-binding ELISA, similar to that described by Roehrig et al (1982), was used to determine whether different MC antibodies bound to the same or to spatially separate sites on the viral proteins. Selected MC antibodies were directly conjugated with horseradish peroxidase, using a two-step glutaraldehyde coupling procedure (Engvall, 1980). Unlabeled MC antibodies were then tested individually for their abilities to bind solidphase SIN antigen and competitively inhibit subsequent binding of each peroxidase-labeled antibody. Unlabeled antibodies were added to antigen wells first, and plates were incubated for 90 min at room temperature before dilutions of labeled MC antibodies were added; another 90-min incubation followed. Plates were washed and substrate added as described above. Two antibodies were presumed to bind to the same or to topographically related antigenie sites if the unlabeled antibody caused at least 50% inhibition of binding of the labeled antibody.
Competitive inhibition
by viral antigen.
A competitive-inhibition ELISA was used to determine whether the binding of MC antibodies to solid-phase (detergent-disrupted) SIN could be inhibited by preincubating them with either intact or detergent-disrupted virions in suspension. Potential inhibitors consisted of (1) intact virus, fresh from an equilibrium gradient, diluted to 50 Mg/ml in PBS, pH 7.2, (2) the same virus preparation disrupted by incubation for 30 min in 1.0% Triton X-100 followed by dilution to 50 pg/ml in PBS, (3) PBS diluent, and (4) PBS-plus-Triton diluent. For each MC antibody, a single
IN SINDBIS
VIRUS
147
dilution series was prepared in PBS containing 5% FBS; an aliquot (75 ~1) of each dilution was then mixed with an equal volume of intact SIN, disrupted SIN, or control diluent. After 30 min incubation at room temperature with constant rocking, each mixture was assayed in duplicate for antibody binding to solid-phase SIN as described above. A modification of the above competitiveinhibition assay was used to determine the effect of pH on the binding of MC antibodies to intact SIN. The buffers and the general approach were those used routinely for arbovirus hemagglutination-inhibition (HI) assays (Clarke and Casals, 1958). Briefly, a series of 0.2 Msodium phosphate buffers was prepared. Each MC antibody was diluted to approximately 0.5 pg/ml in every buffer of the series. Each buffered antibody of the series was then mixed with an equal volume (75 ~1) of either gradientpurified SIN diluted to 50 pg/ml in 0.05 M borate-saline-albumin, pH 9, or with diluent alone. The resulting pH of each mixture was measured and is shown under Results. After a 30-min incubation at room temperature, each buffered antibody (control) or virus-antibody mixture was assayed in duplicate by ELISA, as above, for its capacity to bind to solid-phase SIN. RESULTS
Antigenic Sites on SIN Glycoproteins To determine whether MC antibodies bound to the same or to different topographical sites on El and E2, we tested, by ELISA, the ability of each MC antibody to competitively inhibit the binding of selected peroxidase-labeled MC antibodies to solid-phase SIN (Table 1). Assuming that an unlabeled antibody would inhibit subsequent binding of a labeled antibody if the two bound to identical or adjacent epitopes, we were able to distinguish five topographically distinct antigenic sites on El (designated a through e) and a single site on E2 (Table 1). Three of the El sites, a, b, and c, were related in that the antibody (No. 38) defining site b competitively inhibited binding (and was similarly inhibited in the reciprocal assay) by those MC
148
SCHMALJOHN,
KOKUBUN, TABLE
AND
COLE
1
TOPOGRAPHICALLY DISTINCT ANTIGENIC SITES ON SINDBIS VIRUS GLYCOPROTEINS DEMONSTRATED BY COMPETITIVE-INHIBITION ELISA Unlabeled MC antibody
Peroxidase-labeled
MC antibodies competitively by unlabeled antibodies
inhibited
Specificity
No.
7
38
33
35
45
1
16
31
49
Antigenic site designation
El El El
7 20 40
i-i+ + +
+ + +
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
El-a El-u El-a
El
38
+
0
0
0
0
0
0
El-b
El
33
0
0
0
0
0
0
El-c
El El
35 45
0 0
0 0
0 0
El+ +
+ +
0 0
0 0
0 0
0 0
El-d El-d
El El El El
1 16 31 26
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
El+ + + +
+ + + +
+ + + +
0 0 0 0
El-e El-e El-e El-e
E2 E2 E2 E2 E2 E2
49 50 18 23 10 30
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
1+ + + + + +
‘0
,x;
E2 E2 E2 E2 E2 E2
Notes. An excess of each unlabeled MC antibody was added to wells containing solid-phase SIN antigen, followed by the addition of horseradish peroxidase-labeled MC antibodies. Combinations designated positive (+) reproducibly exhibited at least 50% inhibition of binding of the labeled MC antibody as measured by ELISA, while less than 20% competitive inhibition was observed in those designated negative (0). Patterns defining topographically distinct antigenic sites are enclosed by boxes. Each of the five spatially distinct antigenic sites on El was given a small-letter designation, and the site hound by each MC antibody is shown in the last column.
antibodies defining sites a and c. Sites a and c were separate from one another, and sites d and e were also topographically distinct. In control combinations, binding of each peroxidase-labeled MC antibody was competitively inhibited by the homologous unlabeled antibody. Each of the antigenic sites defined by competitive inhibition may contain multiple discrete epitopes too closely linked to be distinguished (Yewdell and Gerhard, 1981), and nonidentical antibodies may bind to the same site. For example, some but not all MC antibodies to the single E2 site reacted with a neuroadapted variant of SIN (Schmaljohn et al, 1982; Cole et al, 1982). Similarly, MC antibodies to site El-
e exhibited three distinct reaction patterns when tested against other alphaviruses (data not shown).
L?issociation of NT and Cgtolytic Activities All El- and ES-specific MC antibodies shown in Table 2 were C-fixing isotypes (IgG2a or IgG2b) and were able to mediate C-dependent lysis of SIN-infected cells at antibody dilutions exceeding 10p3. Uninfected cells were not lysed and infected cells were lysed only in the presence of added c’. In contrast, PRNT assays, which were also performed with added C, revealed that with one exception (MC No. 33) NT activity was restricted to antibodies with E2 spec-
ANTIGENIC
CHANGES
IN SINDBIS
TABLE PROPERTIES
OF MONOCLONAL
Log,, antibody MC antibody
149
VIRUS
2
ANTIBODIES
TO SINDBIS VIRUS
titer
ELISA
80% PRNT
El-a El-b El-c El-d El-d El-e El-e El-e
4.2 3.9 4.4 4.4 4.6 5.2 4.7 4.5
3.5 3.5 a4.0 3.5 33.0 4.0 3.0 3.5
10 10 160 640 160 320 320 640
IgG2a IgG2a IgG2b IgG2a IgG2a IgG2a IgG2a IgG2b
E2 E2 C
5.0 4.3 3.5
5.0 5.0
85.0 84.0
12,800 6,400
IgG2a IgG2b IgG3
No.
Specificity
7 38 33 35 45 16 1 31 49 50 3
C’-Dependent cytolysis
HI titer
Immunoglobulin isotype
Notes. Aseitic fluids containing MC antibodies were tested for reactivity against SIN by ELISA, plaquereduction neutralization (PRNT), C-dependent lysis of infected cells, and hemagglutination-inhibition (HI). The antigenic site (Table 1) bound by each anti-El MC antibody is shown. An anti-capsid (C) antibody is shown for comparison. PRNT assays were performed in the presence of added C’, which caused a modest augmentation in preexisting NT activities of anti-E2 and anti-El MC antibodies of IgG2a and IgG2b isotypes (data not shown) but did not alter the undetectable NT activities of MC antibodies to sites El-a, -b, -d, and -e. Cytolysis titers represent the highest antibody dilutions effecting at least 10% specific 51Cr release from SIN-infected cells; the lytic antibodies shown caused 40-60% specific release at higher antibody concentrations (i.e., 10e2 dilution) but did not lyse uninfected cells and did not lyse infected cells in the absence of C’. All El- and EZ-specific MC antibodies shown above are protective in viwo (Schmaljohn et aL, 1982; Cole et al, 1982).
ificity. Mixtures of two or more non-NT anti-El antibodies were also ineffective in PRNT assays (data not shown). Therefore, MC antibodies to four topographically distinct antigenic sites on El mediated C’dependent lysis of SIN-infected cells but exhibited no NT activity. All anti-El and anti-E2 antibodies were able to bind, in ELISA, to purified, disrupted SIN in solid phase (Table 2). Hemagglutination-inhibiting (HI) activity was associated with all NT anti-E2 antibodies tested (Table 2 and data not shown) and with anti-El antibodies to sites El-c, Eld, and El-e (Table 2).
Effect-sof Anti-Immunoglobulins tivity
on NT Ac-
From the data in Table 2, it was not clear whether non-NT anti-El antibodies failed to bind infectious virions or whether
they bound with no resulting loss of viral infectivity. The latter possibility was tested by adding to the PRNT assay both C’ and a rabbit antiserum against mouse immunoglobulins. While the addition of the second reagent clearly augmented the activities of NT antibodies (anti-E2 or the single NT anti-El), it did not enable the non-NT anti-El antibodies to reduce viral infectivity (Table 3).
Cryptic Vi&n
Epitvpes
The absence of NT activity still did not exclude totally the possibility that certain anti-El antibodies might bind to virion surfaces. We therefore attempted to competitively inhibit the binding of MC antibodies to solid-phase SIN antigen by preincubating different amounts of each MC antibody with either intact or detergent-disrupted SIN. As shown in Fig. 1 and
150
SCHMALJOHN, TABLE
KOKUBUN,
3
EFFECT OF ANTI-MOUSE IMMUNOGLOBULINS ON SINDBIS VIRUS NEUTRALIZATION BY MC ANTIBODIES
80% PRNT titer Antibody 7 38 33 1 31 49 23
Specificity
-RAMIg
El-u El-b El-c El-e El-e E2 E2
+RAMIg
AND
COLE
termined whether intact virions in suspension would be bound by MC antibodies at different pH values. As exemplified in Fig. 2, the binding of each MC antibody (in buffered diluent alone) to solid-phase SIN was relatively unaffected by changes of pH. Similarly, the binding of NT antibodies to intact virions, as measured by competitive inhibition, was not influenced by pH (MC 49 in Fig. 2; MC 50 and MC 33 not shown). In contrast, MC No. 16 (Fig.
Notes. Plaque-reduction neutralization (PRNT) titers represent the highest dilution of ascitic fluid (containing l-10 mg/ml of specific antibody as measured by ELISA) effecting 380% reduction of SIN plaques on BHK-21 cells. Tenfold dilutions of MC antibodies were incubated with SIN in the presence of added C’ and in the presence (+) or absence (-) of rabbit antiserum reactive with mouse immunoglobulins (RAMIg) before viral plaque assay. MC 23 was an IgG3 (non-C’-fixing) isotype and exhibited the greatest augmentation of NT activity in the presence of both RAMIg and C’. Non-NT anti-El antibodies did not acquire NT activity even in the presence of C’ and RAMIg.
summarized in Table 4, all NT antibodies tested (including the one anti-El with NT activity) were competitively inhibited by both intact and disrupted virions; non-NT anti-El antibodies, however, were inhibited only by disrupted virions. These results, together with results from NT and cytolysis assays (Table 2), demonstrated that epitopes in four distinct antigenic sites on El became cryptic, i.e., inaccessible to antibodies, during viral maturation at host-cell surfaces.
pH-Dependent Antigenic Virion Surface
Changes in the
Since hemagglutination is a pH-dependent property of purified virions (Mooney et aL, 1975), the observation that certain anti-El antibodies (MC No. 16, for example) inhibited hemagglutination appeared discrepant with their failure to bind to virion surfaces. To resolve this issue, we de-
MC 38 IEll 0 21) -LOG,0
2.5
3.0
ANTIBODY
3.5
4.0
4.5
DILUTION
FIG. 1. Competitive inhibition of MC antibodies by intact or Triton-disrupted Sindbis virions. Dilutions of each MC antibody were mixed with PBS, pH 7.2 (o), with gradient-purified SIN in PBS (0), or with Triton X-lOO-disrupted SIN in PBS (A). The PBSTriton controls (not shown) were indistinguishable from PBS controls. After incubation, mixtures were assayed by ELISA for binding to solid-phase Tritondisrupted SIN. MC No. 49 (top), typical of all NT antibodies tested, was competitively inhibited by either intact or disrupted SIN. MC No. 38 (bottom), typical of anti-El antibodies that lacked NT activity, was inhibited only by Triton-disrupted SIN. Therefore, the epitope bound by MC No. 38 was undetectable on the surfaces of intact virions but could be exposed by treating virions with a nonionic detergent.
ANTIGENIC TABLE
CHANGES
4
COMPETITIVE INHIBITION OF MC ANTIBODIES BY INTACT OR DETERGENT-DISRUPTED SINDBIS VIRIONS % Inhibition of binding by: MC antibody No.
Specificity
7 38 33 35 16 49 50 23 18
El-u El-b El-c El-d El-e E2 E2 E2 E2
Intact SIN 0 0 85 0 0 95 100 95 95
Disrupted SIN (Triton X-100) 95 95 95 95 95 95 90 90 90
IN SINDBIS
151
VIRUS
same epitopes are cryptic, i.e., inaccessible to antibodies, in virions. On the plasma membranes of infected cells, epitopes in all five El antigenic sites (Table 1) were accessible to antibodies as shown by the abilities of MC antibodies to mediate Cdependent cytolysis (Table 2). In contrast, only one El epitope (defined by MC No. 33) was demonstrable on the surfaces of infectious virions. MC antibodies to other El
1.5m I ID-
MC49IE2)
Notes. MC antibodies were preincubated with suspensions of either intact or Triton X-IWtreated SIN and then assayed for their binding to solid-phase SIN, as detailed under Materials and Methods. Percent inhibition, rounded to the nearest 51, was determined from the displacement of the binding curve for each MC antibody compared to the control (antibody + PBS), which was taken as 190% binding (see Fig. 1). MC antibodies to El sites a, b, d, and e did not bind to virion surfaces but did bind to solubilized virion proteins.
2) as well as MC 45 and MC 35 (not shown), all anti-El MC antibodies with HI activity (Table 2), exhibited a pH-dependent ability to bind to (and thus be competitively inhibited by) intact virions. Thus, the expression on virion surfaces of epitopes recognized by such antibodies, like the property of SIN hemagglutination (Mooney et aZ., 1975), occurred only below pH 6.4. These epitopes were present almost exclusively at antigenic sites El-d and El-e (Table 1); antibodies to sites a and b bound poorly to virions at pH 6 (data not shown) and the NT antibody to site c (MC 33) bound virion surfaces at pH ‘7 (Table 4) as well as at pH 6. DISCUSSION
We have shown that epitopes on the El glycoprotein of SIN are present on the surfaces of infected cells but that many of the
E ; :
OS
= oii
MC16
62
6.4
6.6
IEll
6%
7.0
7.2
PH
FIG. 2. Effect of pH on the binding of MC antibodies to Sindbis virions. A predetermined dilution of each MC antibody was incubated with buffered diluent alone (0) or with gradient-purified SIN (0) at each pH indicated. Subsequent binding of antibodies to solid-phase SIN antigen was demonstrated by ELISA. Results showed that pH changes alone had no pronounced effects on the binding of either MC antibody to solid-phase antigen from detergent-disrupted virions (0). In the competitive-inhibition assay (0) MC No. 49 (top panel) bound to and therefore was competitively inhibited by virions (in suspension) at all pH values. In contrast, MC No. 16 exhibited a pHdependent ability to bind to SIN. Conformational changes in the viral glycoproteins that occurred below pH 6.4 caused a formerly hidden El epitope to become accessible to antibody binding.
152
SCHMALJOHN,
KOKUBUN,
epitopes did not bind to the surfaces of intact virions as evidenced by their inabilities to reduce viral infectivity even in the presence of C’ (Table 2) or antiimmunoglobulins (Table 3) and also by their failure to bind intact virions in competitive-inhibition assays (Fig. 1 and Table 4). In parallel assays, all NT antibodies tested mediated cytolysis, exhibited enhanced NT activity in the presence of antiimmunoglobulins, and bound to virion surfaces in competitive-inhibition assays. These results are consistent with a recent report (Gates et a!, 1982) that SIN-immune serum contains a population of antibodies that can be much more efficiently adsorbed by infected cells than by virions. A previously described population of anti-El antibodies capable of binding to virion surfaces without reducing infectivity (Symington et aL, 19’77; Chanas et ak, 1982) is not represented in our current battery of SIN-specific MC antibodies. The observation that non-NT anti-El antibodies did not bind to virion surfaces has added significance in view of our previous observation that the same non-NT antibodies, when passively administered to adult mice, can prevent lethal alphavirus encephalitis (Schmaljohn et aL, 1982; Cole et aL, 1982). Protection in vivo cannot be easily attributed to virolysis (Stollar, 1975) or to opsonization of virions and is more likely the result of non-NT antibodies binding to infected-cell surfaces. The exceptional anti-El MC antibody (No. 33, Table 2) with NT activity illustrates that NT epitopes are not restricted to E2. Similarly, Chanas et al. (1982) described an anti-El MC antibody with NT activity. In the competitive-inhibition assays used to topographically map El and E2 epitopes, the NT epitopes on E2 were spatially unrelated to the NT epitope on El (Table 1); instead, the El NT epitope was proximal to an El epitope defined by a non-NT antibody. The molecular events that lead to sequestration or masking of El epitopes have not been determined. Because all cryptic El sites could be exposed by disrupting virions with a nonionic detergent (Fig. 1 and Table 4) and two were exposed at pH
AND
COLE
6 (Fig. 2), it is unlikely that El epitopes became cryptic as the sole result of either glycosylation or disulfide-bond formation. Our results are therefore different but not in conflict with those of Kaluza et al. (1980) who found antigenic differences between the glycosylated and nonglycosylated forms of both El and E2 of Semliki Forest virus and also described (Kaluza and Pauli, 1980) the conformational change associated with chemical reduction of El. The antigenic changes in El described herein also appear to be very different from the reported conformational change, marked by decreased reactivity with MC antibodies (Roehrig et aL, 1982), that accompanies exposure of El to SDS, a strong ionic detergent. The expression of El epitopes at pH 6 (Fig. 2) is probably due to a conformational change in El. The evidence for this eonformational change is most easily derived from the observation by Dalrymple et al. (1976) that purified El binds to red cell membranes at pH 5.8 but not at pH 7.0; virions also bind in a pH-dependent fashion to protein-free model liposomes (Mooney et al., 1975) and therefore a change in El rather than in the erythrocyte probably occurs. As recently shown by Edwards et al. (1983) concurrent changes in E2 may also take place. While the acquisition of hemagglutinating activity which accompanies the conformational change in El is well known, the demonstration of an antigenic change in the virion surface (Fig. 2) at pH 6 is a novel finding. The biological significance of the pH-dependent alterations in alphavirus surfaces may lie in the proposed endosomal route of entry of these viruses into the host-cell cytoplasm (Helenius et aL, 1980). Intermolecular interactions may also influence El epitope expression. The virion form of E2 is derived from a larger precursor, PE2, which is cleaved at or near the plasma membrane (Schlesinger and Kailriiinen, 1980; Garoff et aL, 1982). That El-PE2 interactions may be weaker than El-E2 interactions has been suggested by crosslinking studies (Rice and Strauss, 1982) and by analysis of plasma membranes recovered from phagocytized latex
ANTIGENIC
CHANGES
beads (Scheefers et aL, 1980). Some epitopes on the plasma membrane form of El may therefore become inaccessible to antibodies due to increased El-E2 associations and may be further masked during budding as the more tightly packed virions are formed. The observation that NT anti-E2 antibodies had HI activity (Table 2) was initially surprising because El appears to serve functionally as the hemagglutinin (Dalrymple et aL, 1976). But because of the intimate El-E2 association described above, anti-E2 MC antibodies might inhibit hemagglutination either by aggregating El-E2 complexes or by sterically inhibiting the El-erythrocyte interaction. In this context the alphavirus HI assay actually may be biased toward measuring NT antibodies because the initial antigen-antibody incubation is performed at pH 9, while certain of the non-NT HI antibodies (No. 16, for instance) may effectively bind antigen only at the low pH (5.8-6.2) encountered when erythrocytes are added. That HI antibodies but not NT antibodies crossreact among different alphaviruses (Casals, 1963) probably reflects the general finding that antigenically (Dalrymple et al, 1976) and by sequence analysis (Rice and Strauss, 1981) El is more highly conserved among alphaviruses than is E2. The above data demonstrate that the surface of Sindbis virus is antigenically different from the individual proteins that comprise the surface; epitopes in four distinct antigenic sites on the El glycoprotein, though present on infected-cell surfaces and solubilized virions, are undetectable on the surfaces of infectious virions. The virion surface, however, is not static; antigenic changes accompany the pH-dependent acquisition of HA activity. These observations, together with the recent recognition that non-NT antibodies can mediate protection in vivo (Schmaljohn et aZ., 1982), have considerable implications for the assessment of alphavirus immunity and the design of alphavirus vaccines. ACKNOWLEDGMENTS We thank Dr. Joel M. Dalrymple for invaluable discussions and for generously providing hybridoma cell lines for use in these studies. This work was sup-
IN SINDBIS
VIRUS
153
ported by Contract C-8042 from the U. S. Army Research and Development Command and by USPHS Grant NS 17741. REFERENCES CASALS, J. (1963). Relationships among arthropodborne animal viruses determined by cross-challenge tests. Amer. J. Trap. Med Hyg. 12, 587-596. 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. 58,3’746. CLARKE, D. H., and CASALS, J. (1958). Techniques for hemagglutination and hemagglutination-inhibition with arthropod-borne viruses. Amer. J. Twp. Meat Hyg. 7, 561-573. COLE, G. A., SCHMALJOHN, A. L., and DALRYMPLE, J. M. (1982). Protection against lethal Sindbis virus infection with specific monoclonal antibodies. In “Viral Disease in South-East Asia and the Western Pacific” (J. S. MacKenzie, ed.), pp. 541-545. Academic Press, New York. DALRYMPLE, J. M., SCHLESINGER, S., and RUSSELL, P. K. (1976). Antigenic characterization of two Sindbis envelope glycoproteins separated by isoelectric focusing. Virology 69, 93-103. EDWARDS, J., MANN, E., and BROWN, D. T. (1983). Conformational changes in Sindbis virus envelope proteins accompanying exposure to low pH. J. ViroL 45, 1090-1097. ENGVALL, E. (1980). Enzyme immunoassay ELISA and EMIT. In “Methods in Enzymology” (H. Van Vunakis and J. J. Langone, eds.), Vol. 70, pp. 419-439. Academic Press, New York. GAROFF, H., KONDOR-KOCH, C., and RIEDEL, H. (1982). Structure and assembly of alphaviruses. Curr. Top. Microbial. Immunol. 99, l-50. GATES, D., BROWN, A., and WUST, C. J. (1982). Comparison of specific and cross-reactive antigens of alphaviruses on virions and infected cells. Irlfect. Immun 35.248-255. GENTRY, M. K., HENCHAL, E. A., MCCOWN, J. M., BRANDT, W. E., and DALRYMPLE, J. M. (1982). Identification of distinct antigenic determinants on Dengue-2 virus using monoclonal antibodies. Amer. J. Trap. Med Hyg. 31, 548-555. HELENIUS, A., KARTENBECK, J., SIMONS, K., and FRIES, E. (1980). On the entry of Semliki Forest virus into BHK-21 cells. J. Cell Biol. 84, 404-420. KALUZA, G., Roar, R., and SCHWARZ, R. T. (1980). Carbohydrate-induced conformational changes of Semliki Forest virus glycoproteins determine antigenicity. Virology 102, 286-299. KALUZA, G., and PAULI, G. (1980). The influence of intramolecular disulfide bonds on the structure and
154
SCHMALJOHN.
KOKUBUN,
function of Semliki Forest virus membrane glycoproteins. VZroZogy 102, 300-309. KENNEW, R. H. (1980). Enzyme-linked antibody assay with cells attached to polyvinyl chloride plates. In “Monoclonal Antibodies” (R. H. Kennett, T. J. McKearn, and K. B. Bechtol, eds.), pp. 376-377. Plenum, New York. MOONEY, J. J., DALRYMPLE, J. M., ALVING, C. R., and RUSSELL, P. K. (1975). Interaction of Sindbis virus with liposomal mode1 membranes. J. ViroL 15,225231. RICE, C. M., and STRAUSS, J. H. (1981). Nucleotide sequence of the 26s mRNA of Sindbis and deduced sequence of the encoded virus structural proteins. Proc. Nat. Acad Sci. USA 78, 2062-2066. RICE, C. M., and STRAUSS, J. H. (1982). Association of Sindbis virus glycoproteins and their precursors. J. Mol. Biol 154, 325-348. ROEHRIG, J. T., DAY, J. W., and KINNEY, R. M. (1982a). Antigenic analysis of the surface glycoproteins of a Venezuelan equine encephalitis virus (TC-83) using monoclonal antibodies. firology 118, 269-278. ROEHRIG, J. T., GORSKI, D., and SCHLESINGER, M. J. (198213). Properties of monoclonal antibodies directed against the glycoproteins of Sindbis virus J. Gen Viral. 59, 421-425. SCHEEFERS, H., SCHEEFERS-BORCHEL, U., EDWARDS, J., and BROWN, D. T. (1980). Distribution of virus structural proteins and protein-protein interactions
AND
COLE
in plasma membrane of baby hamsters’ kidney cells infected with Sindbis or vesicular stomatitis virus. Proc. Nat. Ad Sci USA 77, 7277-7281. SCHLESINGER, M. J., and KKHRIHNEN, L. (1980). Translation and processing of alphavirus proteins. In “The Togaviruses: Biology, Structure, Replication” (R. W. Schlesinger, ed.), pp. 371-392. Academic Press, New York. 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. SEFTON, B. M., WICKUS, G. G., and BURGE, B. W. (1973). Enzymatic iodination of Sindbis virus proteins. J. ViroL 11, 730-735. SMITH, J. F., and BROWN, D. T. (1977). Envelopment of Sindbis virus: Synthesis and organization of proteins in cells infected with wild type and maturation-defective mutants. J. Vird 22, 662-678. SMITH, J. F., and PIFAT, D. Y. (1982). Morphogenesis of sandfly fever viruses (Bunyaviridae family). virology 121, 61-81. STOLLAR, V. (1975). Immune lysis of Sindbis virus. Virology 66, 620-624. SYMINGTON, J., MCCANN, A. K., and SCHLESINGER, M. J. (1977). Infectious virus-antibody complexes of Sindbus virus. I$ect. Immun 15, 720-725. YEWDELL, J. W., and GERHARD, W. (1981). Antigenic characterization of viruses by monoclonal antibodies. Annu. Rev. Microbial 35, 185-206.
VIROLOGY
Protective ALAN
Monoclonal Antibodies Define Maturational and pH-Dependent Antigenic Changes in Sindbis Virus El Glycoprotein
L. SCHMALJOHN: Department
KRISTINA
M. KOKUBUN,
AND
GERALD
A. COLE
of Microbiology, University of Maryland School of Medicine, 660 West Redwood Street, Baltimore, Maryland 21201 Received April
12, 1983’; accepted July 11, 1983
Monoclonal (MC) antibodies specific for either the El or E2 glycoproteins of Sindbis virus (SIN) were used to probe for differences in the surface topography of SIN epitopes between infected cells and mature virions. Employing an enzyme-linked immunosorbent assay (ELISA) in which binding of individual peroxidase-labeled MC antibodies to immobilized (solid-phase) detergent-disrupted SIN was inhibited specifically by one or more unlabeled antibodies, viral epitopes could be grouped into six spatially distinct antigenic sites-five on El, designated a through e, and one site on E2. All six sites were represented on the surfaces of SIN-infected cells as shown by the complement (C)-dependent lysis mediated by antibodies of the corresponding epitope specificities. In contrast, virus-neutralizing (NT) activity was restricted to antibodies specific for epitopes on E2 and on site c of El, irrespective of the presence of added C’ and an antiserum against mouse immunoglobulins. That El sites a, b, d, and e became inaccessible to antibody binding was shown by a competitive-inhibition ELISA. Whereas all MC antibodies were inhibited from binding to solid-phase SIN when premixed with detergent-treated virions, only those having NT activity could be competitively inhibited by intact virions. Sites El-d and Ele could be exposed not only by detergent disruption but also by lowering the virion pH from 7.2 to 6.0. These collective results indicate that a majority of immunologically relevant El epitopes present on SIN-infected cell surfaces become cryptic during SIN maturation and, except at low pH, remain undetectable on virion surfaces. INTRODUCTION
Sindbis virus (SIN), like other members of the Alphmkm genus, matures by budding at host-cell plasma membranes. Alphavirus structure and morphogenesis are reasonably well understood from the comprehensive studies of SIN and Semliki Forest virus (reviewed by Schlesinger and Klilrianen, 1980; Garoff et al, 1982). The fully assembled SIN virion, approximately 65 nm in diameter, is a relatively simple particle which contains only three structural proteins, a lipid envelope, and a single 42 S molecule of genomic RNA. Two of the viral proteins, designated El and E2, are envelope glycoproteins, each approximately 50,000 Da (50K), that together form the external virion spikes. The internal r To whom reprint
requests should be addressed.
0042-6822/83 $3.00 Copyright All rights
0 1983 by Academic Press, Inc. of reproduction in any form reserved
nucleocapsid contains the genomic RNA within an icosahedral shell composed of 30K capsid (C) protein subunits. Following glycosylation and translocation of the envelope proteins to the cell surface, nucleocapsids align along the cytoplasmic face of the plasma membrane and virus budding occurs. This final stage of morphogenesis is a highly ordered process; host proteins are excluded from the viral envelope and the three structural proteins are incorporated into virions apparently in equimolar amounts. When viral structure and morphogenesis are viewed in the context of protective antibody responses to SIN and other alphavirus infections, the nature and locations of viral epitopes which induce such antibodies are of particular relevance. Since El and E2 are present on plasma 144
ANTIGENIC
CHANGES
membranes of SIN-infected cells as well as on virion envelopes (Sefton et aZ., 1973; Smith and Brown, 19’77), protective antibodies specific for either glycoprotein might be expected to bind at both locations. Epitopes on the surfaces of virions, especially those to which binding by antibodies results in neutralization of infectivity in vitro, are usually considered the “targets” of humoral immune responses which mediate host recovery (Casals, 1963). However, antibodies induced by alphavirus epitopes expressed exclusively on host-cell surfaces also appear to be important in protective immunity. As reported recently, monoclonal (MC) antibodies several against El of SIN, although lacking neutralizing (NT) activity against SIN in vitro, were capable of passively conferring protection to adult mice against an otherwise lethal intracerebral SIN challenge. Protection by these antibodies correlated with their ability to mediate complement (C’)dependent lysis of SIN-infected cells in vitro (Schmaljohn et aL, 1982). Based on current knowledge of alphavirus morphogenesis, the existence of such anti-El antibodies would not have been anticipated since the only covalent modification of SIN envelope glycoproteins thought to occur at the plasma membrane is the cleavage of PE2, the precursor of E2 (see reviews cited above). It was therefore surprising to find that certain El epitopes were demonstrable in assays for cytolytic but not NT antibodies. That certain antiEl MC antibodies might bind to the surfaces of virions without reducing their infectivity seemed unlikely after it was found that these antibodies had no NT activity in the presence of c’ and anti-immunoglobulin antibodies. Data presented below establish that several immunologically important El epitopes are present on infected-cell surfaces and on detergent-solubilized viral proteins but are undetectable on virion surfaces. MATERIALS
AND
Virus propagation Sindbis virus, strain
METHODS
and puti$cation. AR339, was main-
IN SINDBIS
VIRUS
145
tained at -80” as an infected suckling mouse brain suspension (seed virus) and was propagated in BHK-21 (hamster kidney) cells. Growth medium for BHK-21 cells consisted of RPM1 1640 containing 10% fetal bovine serum (FBS), penicillin (100 U/ml), streptomycin (100 pg/ml), gentamycin (40 pg/ml), and 10 mM HEPES. Subconfluent monolayer cultures were infected with seed virus at an m.o.i. of 0.1 PFU/cell in a small volume of growth medium which, following a 60-min adsorption at 37”, was replaced with medium containing 2% FBS. After cultures were incubated for 24-36 hr at 37”, the medium was harvested and clarified by centrifugation .for 10 min at 10,000 g. Virus was concentrated by precipitation with polyethylene glycol, layered onto a 15-40% potassium tartrate gradient, and centrifuged to equilibrium, usually overnight, at 22,000 rpm in a Beckman SW 27.1 rotor. The virus band was collected, assayed for protein content (BioRad protein assay), and used immediately or stored at -80”. Mwrmclmal antibodies. Infectious SIN, either as purified virus or as a suckling BALB/c mouse brain suspension, was used to immunize BALB/c mice for each of four fusions between immune spleen cells and the P3 X 63Ag8 (IgGl, kappa secretor) myeloma cell line. The fusion, selection, and cloning procedures were the same as those described for anti-Dengue virus MC antibodies by Gentry et al. (1982). The specificity of each MC antibody was determined by solid-phase radioimmunoassay using purified individual proteins of SIN (Dalrymple et aL, 1976) and subsequently confirmed by immunoprecipitation of 3H-labeled virion proteins followed by SDSpolyacrylamide gel electrophoresis, as described by Smith and Pifat (1982). Immunoglobulin heavy-chain isotypes of MC antibodies were determined by immunodiffusion; culture supernates from each hybrid cell line were tested against immunoglobulin class- and subclass-specific (IgM, IgG3, IgGl, IgGBb, IgGBa, and IgA) antisera (Litton Bionetics). MC antibodycontaining ascitic fluids were prepared in Pristane-primed BALB/c mice and used in all other assays because of the greater an-
146
SCHMALJOHN.
KOKUBLJN.
tibody concentrations (l-10 mg/ml) in ascitic fluids. Neutralization assays. Plaque-reduction neutralization (PRNT) assays of suckling mouse brain seed virus were performed as previously described (Schmaljohn et aL, 1982) by preincubating SIN (50-100 PFU/ 0.1 ml) with dilutions of antibodies and then measuring residual infectious virus by plaque assay on BHK-21 cells. All PRNT assays shown were performed in the presence of added 5% fresh guinea pig serum (Cedarlane) as a source of c’, for some assays, a rabbit antiserum against mouse immunoglobulins (Cappel) was added (final dilution of 1:200) and the mixture of virus and antibodies was incubated for an additional 45 min at 37” before adsorption onto BHK-21 monolayers. PRNT assays were also performed with Vero (monkey kidney) cells to assure that results were not unique to the BHK-21 cell assay system; no significant differences were observed between the two cell types. Chromium-release assaya To demonstrate C-dependent antibody-mediated cytolysis, Vero cells were infected (or mock infected) with seed virus at an m.o.i. of 1 and incubated for 14 hr at 37”. Cells were trypsinized, washed, labeled with 51Cr (150 &i/lo7 cells), washed, and seeded into flatbottomed 96-well tissue culture plates (Costar) as previously described (Schmaljohn et aL, 1982). Each well received a lOO~1 aliquot containing 2.5 X lo4 cells in RPM1 1640 supplemented with 10% FBS. Plates were incubated for 60 min at 37” to assure regeneration of cell-surface viral antigens that may have been removed by trypsinization. Rabbit C’ (Cedarlane Low-Tox M) was added in 50-~1 aliquots, followed by dilutions of heat-inactivated MC antibodies, also in 50-~1 aliquots. The final serum C’ content in each well was 2% (v/v). All tests and controls were done in triplicate. Plates were incubated for 3 hr at 37” and then 100 ~1 of supernate was removed from each well and counted in a gamma counter; SEM for triplicate wells did not exceed 5% of the mean counts per minute (cpm). The percent specific 51Cr release was calculated for infected and uninfected Vero cells as % specific release = (test,rm - control,,,)
AND
COLE
X lOO/maximum releasable,,, where the denominator is the number of cpm released from cells treated with 1% Triton X-100. Complement controls, which spontaneously released 13 and 23% of maximum releasable cpm for uninfected and SIN-infected Vero cells, respectively, were used in the above formula for all test systems containing C’. Medium controls spontaneously released 12 and 18%, respectively, for uninfected and SIN-infected Vero cells. The lytic titer of each antibody was taken as the highest dilution that caused at least 10% specific release; specific release values for normal ascitic fluids and anti-nucleocapsid MC antibodies never exceeded 3%. Qualitatively, results of cytolysis assays were not dependent upon the host-cell species nor upon the C’ source because similar results were obtained with BHK-21 cells and guinea pig C’ (Schmaljohn et ul, 1982). Antibody quantitation. Antigen-containing plates for enzyme-linked immunosorbent assay (ELISA) were prepared by disrupting gradient-purified SIN (400-900 pg/ ml) with 1.0% Triton X-100, diluting to 10 pg/ml in phosphate-buffered saline (PBS), pH 7.2, and then distributing 50 &well into 96-well polystyrene plates (Costar 3590); plates were evaporated to dryness overnight in a biological safety cabinet. The antibody assay was essentially the same as that described by Kennett (1980). Briefly, wells were “blocked” with 5% FBS and washed; dilutions of antiviral antibodies in PBS containing 5% FBS were added, in duplicate, in volumes of 50 pi/well. Plates were incubated overnight at 4°C and then washed four times with PBS containing 0.02% Tween-20 and 0.5% FBS. Horseradish peroxidase-conjugated goat immunoglobulin (Cappel), reactive against all mouse immunoglobulin classes, was diluted 1:lOOO in the above washing solution and added (50 ~1) to each well. Plates were incubated for 90 min at room temperature with constant rocking and then washed four times. Freshly prepared o-phenylenediamine substrate solution was added to each well in a volume of 100 ~1 and plates were rocked for 30 min at room temperature. Reactions were stopped by the addition of 100 ~1 of 0.1 M NaF and the optical
ANTIGENIC
CHANGES
density (OD) for each well was read at 450 nm with a Dynatech MicroELISA II automated plate reader. The ELISA titer for each antibody was taken as the dilution of antibody resulting in an OD of 0.5 which, by comparison with mouse immunoglobulin standards, corresponded to approximately 5 ng of bound antibody per well; therefore, an ELISA titer of 104.’ represented (for high-affinity antibodies) approximately 1 mg/ml of specific antibody in the undiluted sample.
Competitive inhibition by unlabeled antibodies. A competitive-binding ELISA, similar to that described by Roehrig et al (1982), was used to determine whether different MC antibodies bound to the same or to spatially separate sites on the viral proteins. Selected MC antibodies were directly conjugated with horseradish peroxidase, using a two-step glutaraldehyde coupling procedure (Engvall, 1980). Unlabeled MC antibodies were then tested individually for their abilities to bind solidphase SIN antigen and competitively inhibit subsequent binding of each peroxidase-labeled antibody. Unlabeled antibodies were added to antigen wells first, and plates were incubated for 90 min at room temperature before dilutions of labeled MC antibodies were added; another 90-min incubation followed. Plates were washed and substrate added as described above. Two antibodies were presumed to bind to the same or to topographically related antigenie sites if the unlabeled antibody caused at least 50% inhibition of binding of the labeled antibody.
Competitive inhibition
by viral antigen.
A competitive-inhibition ELISA was used to determine whether the binding of MC antibodies to solid-phase (detergent-disrupted) SIN could be inhibited by preincubating them with either intact or detergent-disrupted virions in suspension. Potential inhibitors consisted of (1) intact virus, fresh from an equilibrium gradient, diluted to 50 Mg/ml in PBS, pH 7.2, (2) the same virus preparation disrupted by incubation for 30 min in 1.0% Triton X-100 followed by dilution to 50 pg/ml in PBS, (3) PBS diluent, and (4) PBS-plus-Triton diluent. For each MC antibody, a single
IN SINDBIS
VIRUS
147
dilution series was prepared in PBS containing 5% FBS; an aliquot (75 ~1) of each dilution was then mixed with an equal volume of intact SIN, disrupted SIN, or control diluent. After 30 min incubation at room temperature with constant rocking, each mixture was assayed in duplicate for antibody binding to solid-phase SIN as described above. A modification of the above competitiveinhibition assay was used to determine the effect of pH on the binding of MC antibodies to intact SIN. The buffers and the general approach were those used routinely for arbovirus hemagglutination-inhibition (HI) assays (Clarke and Casals, 1958). Briefly, a series of 0.2 Msodium phosphate buffers was prepared. Each MC antibody was diluted to approximately 0.5 pg/ml in every buffer of the series. Each buffered antibody of the series was then mixed with an equal volume (75 ~1) of either gradientpurified SIN diluted to 50 pg/ml in 0.05 M borate-saline-albumin, pH 9, or with diluent alone. The resulting pH of each mixture was measured and is shown under Results. After a 30-min incubation at room temperature, each buffered antibody (control) or virus-antibody mixture was assayed in duplicate by ELISA, as above, for its capacity to bind to solid-phase SIN. RESULTS
Antigenic Sites on SIN Glycoproteins To determine whether MC antibodies bound to the same or to different topographical sites on El and E2, we tested, by ELISA, the ability of each MC antibody to competitively inhibit the binding of selected peroxidase-labeled MC antibodies to solid-phase SIN (Table 1). Assuming that an unlabeled antibody would inhibit subsequent binding of a labeled antibody if the two bound to identical or adjacent epitopes, we were able to distinguish five topographically distinct antigenic sites on El (designated a through e) and a single site on E2 (Table 1). Three of the El sites, a, b, and c, were related in that the antibody (No. 38) defining site b competitively inhibited binding (and was similarly inhibited in the reciprocal assay) by those MC
148
SCHMALJOHN,
KOKUBUN, TABLE
AND
COLE
1
TOPOGRAPHICALLY DISTINCT ANTIGENIC SITES ON SINDBIS VIRUS GLYCOPROTEINS DEMONSTRATED BY COMPETITIVE-INHIBITION ELISA Unlabeled MC antibody
Peroxidase-labeled
MC antibodies competitively by unlabeled antibodies
inhibited
Specificity
No.
7
38
33
35
45
1
16
31
49
Antigenic site designation
El El El
7 20 40
i-i+ + +
+ + +
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
El-a El-u El-a
El
38
+
0
0
0
0
0
0
El-b
El
33
0
0
0
0
0
0
El-c
El El
35 45
0 0
0 0
0 0
El+ +
+ +
0 0
0 0
0 0
0 0
El-d El-d
El El El El
1 16 31 26
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
El+ + + +
+ + + +
+ + + +
0 0 0 0
El-e El-e El-e El-e
E2 E2 E2 E2 E2 E2
49 50 18 23 10 30
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
0 0 0 0 0 0
1+ + + + + +
‘0
,x;
E2 E2 E2 E2 E2 E2
Notes. An excess of each unlabeled MC antibody was added to wells containing solid-phase SIN antigen, followed by the addition of horseradish peroxidase-labeled MC antibodies. Combinations designated positive (+) reproducibly exhibited at least 50% inhibition of binding of the labeled MC antibody as measured by ELISA, while less than 20% competitive inhibition was observed in those designated negative (0). Patterns defining topographically distinct antigenic sites are enclosed by boxes. Each of the five spatially distinct antigenic sites on El was given a small-letter designation, and the site hound by each MC antibody is shown in the last column.
antibodies defining sites a and c. Sites a and c were separate from one another, and sites d and e were also topographically distinct. In control combinations, binding of each peroxidase-labeled MC antibody was competitively inhibited by the homologous unlabeled antibody. Each of the antigenic sites defined by competitive inhibition may contain multiple discrete epitopes too closely linked to be distinguished (Yewdell and Gerhard, 1981), and nonidentical antibodies may bind to the same site. For example, some but not all MC antibodies to the single E2 site reacted with a neuroadapted variant of SIN (Schmaljohn et al, 1982; Cole et al, 1982). Similarly, MC antibodies to site El-
e exhibited three distinct reaction patterns when tested against other alphaviruses (data not shown).
L?issociation of NT and Cgtolytic Activities All El- and ES-specific MC antibodies shown in Table 2 were C-fixing isotypes (IgG2a or IgG2b) and were able to mediate C-dependent lysis of SIN-infected cells at antibody dilutions exceeding 10p3. Uninfected cells were not lysed and infected cells were lysed only in the presence of added c’. In contrast, PRNT assays, which were also performed with added C, revealed that with one exception (MC No. 33) NT activity was restricted to antibodies with E2 spec-
ANTIGENIC
CHANGES
IN SINDBIS
TABLE PROPERTIES
OF MONOCLONAL
Log,, antibody MC antibody
149
VIRUS
2
ANTIBODIES
TO SINDBIS VIRUS
titer
ELISA
80% PRNT
El-a El-b El-c El-d El-d El-e El-e El-e
4.2 3.9 4.4 4.4 4.6 5.2 4.7 4.5
3.5 3.5 a4.0 3.5 33.0 4.0 3.0 3.5
10 10 160 640 160 320 320 640
IgG2a IgG2a IgG2b IgG2a IgG2a IgG2a IgG2a IgG2b
E2 E2 C
5.0 4.3 3.5
5.0 5.0
85.0 84.0
12,800 6,400
IgG2a IgG2b IgG3
No.
Specificity
7 38 33 35 45 16 1 31 49 50 3
C’-Dependent cytolysis
HI titer
Immunoglobulin isotype
Notes. Aseitic fluids containing MC antibodies were tested for reactivity against SIN by ELISA, plaquereduction neutralization (PRNT), C-dependent lysis of infected cells, and hemagglutination-inhibition (HI). The antigenic site (Table 1) bound by each anti-El MC antibody is shown. An anti-capsid (C) antibody is shown for comparison. PRNT assays were performed in the presence of added C’, which caused a modest augmentation in preexisting NT activities of anti-E2 and anti-El MC antibodies of IgG2a and IgG2b isotypes (data not shown) but did not alter the undetectable NT activities of MC antibodies to sites El-a, -b, -d, and -e. Cytolysis titers represent the highest antibody dilutions effecting at least 10% specific 51Cr release from SIN-infected cells; the lytic antibodies shown caused 40-60% specific release at higher antibody concentrations (i.e., 10e2 dilution) but did not lyse uninfected cells and did not lyse infected cells in the absence of C’. All El- and EZ-specific MC antibodies shown above are protective in viwo (Schmaljohn et aL, 1982; Cole et al, 1982).
ificity. Mixtures of two or more non-NT anti-El antibodies were also ineffective in PRNT assays (data not shown). Therefore, MC antibodies to four topographically distinct antigenic sites on El mediated C’dependent lysis of SIN-infected cells but exhibited no NT activity. All anti-El and anti-E2 antibodies were able to bind, in ELISA, to purified, disrupted SIN in solid phase (Table 2). Hemagglutination-inhibiting (HI) activity was associated with all NT anti-E2 antibodies tested (Table 2 and data not shown) and with anti-El antibodies to sites El-c, Eld, and El-e (Table 2).
Effect-sof Anti-Immunoglobulins tivity
on NT Ac-
From the data in Table 2, it was not clear whether non-NT anti-El antibodies failed to bind infectious virions or whether
they bound with no resulting loss of viral infectivity. The latter possibility was tested by adding to the PRNT assay both C’ and a rabbit antiserum against mouse immunoglobulins. While the addition of the second reagent clearly augmented the activities of NT antibodies (anti-E2 or the single NT anti-El), it did not enable the non-NT anti-El antibodies to reduce viral infectivity (Table 3).
Cryptic Vi&n
Epitvpes
The absence of NT activity still did not exclude totally the possibility that certain anti-El antibodies might bind to virion surfaces. We therefore attempted to competitively inhibit the binding of MC antibodies to solid-phase SIN antigen by preincubating different amounts of each MC antibody with either intact or detergent-disrupted SIN. As shown in Fig. 1 and
150
SCHMALJOHN, TABLE
KOKUBUN,
3
EFFECT OF ANTI-MOUSE IMMUNOGLOBULINS ON SINDBIS VIRUS NEUTRALIZATION BY MC ANTIBODIES
80% PRNT titer Antibody 7 38 33 1 31 49 23
Specificity
-RAMIg
El-u El-b El-c El-e El-e E2 E2
+RAMIg
AND
COLE
termined whether intact virions in suspension would be bound by MC antibodies at different pH values. As exemplified in Fig. 2, the binding of each MC antibody (in buffered diluent alone) to solid-phase SIN was relatively unaffected by changes of pH. Similarly, the binding of NT antibodies to intact virions, as measured by competitive inhibition, was not influenced by pH (MC 49 in Fig. 2; MC 50 and MC 33 not shown). In contrast, MC No. 16 (Fig.
Notes. Plaque-reduction neutralization (PRNT) titers represent the highest dilution of ascitic fluid (containing l-10 mg/ml of specific antibody as measured by ELISA) effecting 380% reduction of SIN plaques on BHK-21 cells. Tenfold dilutions of MC antibodies were incubated with SIN in the presence of added C’ and in the presence (+) or absence (-) of rabbit antiserum reactive with mouse immunoglobulins (RAMIg) before viral plaque assay. MC 23 was an IgG3 (non-C’-fixing) isotype and exhibited the greatest augmentation of NT activity in the presence of both RAMIg and C’. Non-NT anti-El antibodies did not acquire NT activity even in the presence of C’ and RAMIg.
summarized in Table 4, all NT antibodies tested (including the one anti-El with NT activity) were competitively inhibited by both intact and disrupted virions; non-NT anti-El antibodies, however, were inhibited only by disrupted virions. These results, together with results from NT and cytolysis assays (Table 2), demonstrated that epitopes in four distinct antigenic sites on El became cryptic, i.e., inaccessible to antibodies, during viral maturation at host-cell surfaces.
pH-Dependent Antigenic Virion Surface
Changes in the
Since hemagglutination is a pH-dependent property of purified virions (Mooney et aL, 1975), the observation that certain anti-El antibodies (MC No. 16, for example) inhibited hemagglutination appeared discrepant with their failure to bind to virion surfaces. To resolve this issue, we de-
MC 38 IEll 0 21) -LOG,0
2.5
3.0
ANTIBODY
3.5
4.0
4.5
DILUTION
FIG. 1. Competitive inhibition of MC antibodies by intact or Triton-disrupted Sindbis virions. Dilutions of each MC antibody were mixed with PBS, pH 7.2 (o), with gradient-purified SIN in PBS (0), or with Triton X-lOO-disrupted SIN in PBS (A). The PBSTriton controls (not shown) were indistinguishable from PBS controls. After incubation, mixtures were assayed by ELISA for binding to solid-phase Tritondisrupted SIN. MC No. 49 (top), typical of all NT antibodies tested, was competitively inhibited by either intact or disrupted SIN. MC No. 38 (bottom), typical of anti-El antibodies that lacked NT activity, was inhibited only by Triton-disrupted SIN. Therefore, the epitope bound by MC No. 38 was undetectable on the surfaces of intact virions but could be exposed by treating virions with a nonionic detergent.
ANTIGENIC TABLE
CHANGES
4
COMPETITIVE INHIBITION OF MC ANTIBODIES BY INTACT OR DETERGENT-DISRUPTED SINDBIS VIRIONS % Inhibition of binding by: MC antibody No.
Specificity
7 38 33 35 16 49 50 23 18
El-u El-b El-c El-d El-e E2 E2 E2 E2
Intact SIN 0 0 85 0 0 95 100 95 95
Disrupted SIN (Triton X-100) 95 95 95 95 95 95 90 90 90
IN SINDBIS
151
VIRUS
same epitopes are cryptic, i.e., inaccessible to antibodies, in virions. On the plasma membranes of infected cells, epitopes in all five El antigenic sites (Table 1) were accessible to antibodies as shown by the abilities of MC antibodies to mediate Cdependent cytolysis (Table 2). In contrast, only one El epitope (defined by MC No. 33) was demonstrable on the surfaces of infectious virions. MC antibodies to other El
1.5m I ID-
MC49IE2)
Notes. MC antibodies were preincubated with suspensions of either intact or Triton X-IWtreated SIN and then assayed for their binding to solid-phase SIN, as detailed under Materials and Methods. Percent inhibition, rounded to the nearest 51, was determined from the displacement of the binding curve for each MC antibody compared to the control (antibody + PBS), which was taken as 190% binding (see Fig. 1). MC antibodies to El sites a, b, d, and e did not bind to virion surfaces but did bind to solubilized virion proteins.
2) as well as MC 45 and MC 35 (not shown), all anti-El MC antibodies with HI activity (Table 2), exhibited a pH-dependent ability to bind to (and thus be competitively inhibited by) intact virions. Thus, the expression on virion surfaces of epitopes recognized by such antibodies, like the property of SIN hemagglutination (Mooney et aZ., 1975), occurred only below pH 6.4. These epitopes were present almost exclusively at antigenic sites El-d and El-e (Table 1); antibodies to sites a and b bound poorly to virions at pH 6 (data not shown) and the NT antibody to site c (MC 33) bound virion surfaces at pH ‘7 (Table 4) as well as at pH 6. DISCUSSION
We have shown that epitopes on the El glycoprotein of SIN are present on the surfaces of infected cells but that many of the
E ; :
OS
= oii
MC16
62
6.4
6.6
IEll
6%
7.0
7.2
PH
FIG. 2. Effect of pH on the binding of MC antibodies to Sindbis virions. A predetermined dilution of each MC antibody was incubated with buffered diluent alone (0) or with gradient-purified SIN (0) at each pH indicated. Subsequent binding of antibodies to solid-phase SIN antigen was demonstrated by ELISA. Results showed that pH changes alone had no pronounced effects on the binding of either MC antibody to solid-phase antigen from detergent-disrupted virions (0). In the competitive-inhibition assay (0) MC No. 49 (top panel) bound to and therefore was competitively inhibited by virions (in suspension) at all pH values. In contrast, MC No. 16 exhibited a pHdependent ability to bind to SIN. Conformational changes in the viral glycoproteins that occurred below pH 6.4 caused a formerly hidden El epitope to become accessible to antibody binding.
152
SCHMALJOHN,
KOKUBUN,
epitopes did not bind to the surfaces of intact virions as evidenced by their inabilities to reduce viral infectivity even in the presence of C’ (Table 2) or antiimmunoglobulins (Table 3) and also by their failure to bind intact virions in competitive-inhibition assays (Fig. 1 and Table 4). In parallel assays, all NT antibodies tested mediated cytolysis, exhibited enhanced NT activity in the presence of antiimmunoglobulins, and bound to virion surfaces in competitive-inhibition assays. These results are consistent with a recent report (Gates et a!, 1982) that SIN-immune serum contains a population of antibodies that can be much more efficiently adsorbed by infected cells than by virions. A previously described population of anti-El antibodies capable of binding to virion surfaces without reducing infectivity (Symington et aL, 19’77; Chanas et ak, 1982) is not represented in our current battery of SIN-specific MC antibodies. The observation that non-NT anti-El antibodies did not bind to virion surfaces has added significance in view of our previous observation that the same non-NT antibodies, when passively administered to adult mice, can prevent lethal alphavirus encephalitis (Schmaljohn et aL, 1982; Cole et aL, 1982). Protection in vivo cannot be easily attributed to virolysis (Stollar, 1975) or to opsonization of virions and is more likely the result of non-NT antibodies binding to infected-cell surfaces. The exceptional anti-El MC antibody (No. 33, Table 2) with NT activity illustrates that NT epitopes are not restricted to E2. Similarly, Chanas et al. (1982) described an anti-El MC antibody with NT activity. In the competitive-inhibition assays used to topographically map El and E2 epitopes, the NT epitopes on E2 were spatially unrelated to the NT epitope on El (Table 1); instead, the El NT epitope was proximal to an El epitope defined by a non-NT antibody. The molecular events that lead to sequestration or masking of El epitopes have not been determined. Because all cryptic El sites could be exposed by disrupting virions with a nonionic detergent (Fig. 1 and Table 4) and two were exposed at pH
AND
COLE
6 (Fig. 2), it is unlikely that El epitopes became cryptic as the sole result of either glycosylation or disulfide-bond formation. Our results are therefore different but not in conflict with those of Kaluza et al. (1980) who found antigenic differences between the glycosylated and nonglycosylated forms of both El and E2 of Semliki Forest virus and also described (Kaluza and Pauli, 1980) the conformational change associated with chemical reduction of El. The antigenic changes in El described herein also appear to be very different from the reported conformational change, marked by decreased reactivity with MC antibodies (Roehrig et aL, 1982), that accompanies exposure of El to SDS, a strong ionic detergent. The expression of El epitopes at pH 6 (Fig. 2) is probably due to a conformational change in El. The evidence for this eonformational change is most easily derived from the observation by Dalrymple et al. (1976) that purified El binds to red cell membranes at pH 5.8 but not at pH 7.0; virions also bind in a pH-dependent fashion to protein-free model liposomes (Mooney et al., 1975) and therefore a change in El rather than in the erythrocyte probably occurs. As recently shown by Edwards et al. (1983) concurrent changes in E2 may also take place. While the acquisition of hemagglutinating activity which accompanies the conformational change in El is well known, the demonstration of an antigenic change in the virion surface (Fig. 2) at pH 6 is a novel finding. The biological significance of the pH-dependent alterations in alphavirus surfaces may lie in the proposed endosomal route of entry of these viruses into the host-cell cytoplasm (Helenius et aL, 1980). Intermolecular interactions may also influence El epitope expression. The virion form of E2 is derived from a larger precursor, PE2, which is cleaved at or near the plasma membrane (Schlesinger and Kailriiinen, 1980; Garoff et aL, 1982). That El-PE2 interactions may be weaker than El-E2 interactions has been suggested by crosslinking studies (Rice and Strauss, 1982) and by analysis of plasma membranes recovered from phagocytized latex
ANTIGENIC
CHANGES
beads (Scheefers et aL, 1980). Some epitopes on the plasma membrane form of El may therefore become inaccessible to antibodies due to increased El-E2 associations and may be further masked during budding as the more tightly packed virions are formed. The observation that NT anti-E2 antibodies had HI activity (Table 2) was initially surprising because El appears to serve functionally as the hemagglutinin (Dalrymple et aL, 1976). But because of the intimate El-E2 association described above, anti-E2 MC antibodies might inhibit hemagglutination either by aggregating El-E2 complexes or by sterically inhibiting the El-erythrocyte interaction. In this context the alphavirus HI assay actually may be biased toward measuring NT antibodies because the initial antigen-antibody incubation is performed at pH 9, while certain of the non-NT HI antibodies (No. 16, for instance) may effectively bind antigen only at the low pH (5.8-6.2) encountered when erythrocytes are added. That HI antibodies but not NT antibodies crossreact among different alphaviruses (Casals, 1963) probably reflects the general finding that antigenically (Dalrymple et al, 1976) and by sequence analysis (Rice and Strauss, 1981) El is more highly conserved among alphaviruses than is E2. The above data demonstrate that the surface of Sindbis virus is antigenically different from the individual proteins that comprise the surface; epitopes in four distinct antigenic sites on the El glycoprotein, though present on infected-cell surfaces and solubilized virions, are undetectable on the surfaces of infectious virions. The virion surface, however, is not static; antigenic changes accompany the pH-dependent acquisition of HA activity. These observations, together with the recent recognition that non-NT antibodies can mediate protection in vivo (Schmaljohn et aZ., 1982), have considerable implications for the assessment of alphavirus immunity and the design of alphavirus vaccines. ACKNOWLEDGMENTS We thank Dr. Joel M. Dalrymple for invaluable discussions and for generously providing hybridoma cell lines for use in these studies. This work was sup-
IN SINDBIS
VIRUS
153
ported by Contract C-8042 from the U. S. Army Research and Development Command and by USPHS Grant NS 17741. REFERENCES CASALS, J. (1963). Relationships among arthropodborne animal viruses determined by cross-challenge tests. Amer. J. Trap. Med Hyg. 12, 587-596. 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. 58,3’746. CLARKE, D. H., and CASALS, J. (1958). Techniques for hemagglutination and hemagglutination-inhibition with arthropod-borne viruses. Amer. J. Twp. Meat Hyg. 7, 561-573. COLE, G. A., SCHMALJOHN, A. L., and DALRYMPLE, J. M. (1982). Protection against lethal Sindbis virus infection with specific monoclonal antibodies. In “Viral Disease in South-East Asia and the Western Pacific” (J. S. MacKenzie, ed.), pp. 541-545. Academic Press, New York. DALRYMPLE, J. M., SCHLESINGER, S., and RUSSELL, P. K. (1976). Antigenic characterization of two Sindbis envelope glycoproteins separated by isoelectric focusing. Virology 69, 93-103. EDWARDS, J., MANN, E., and BROWN, D. T. (1983). Conformational changes in Sindbis virus envelope proteins accompanying exposure to low pH. J. ViroL 45, 1090-1097. ENGVALL, E. (1980). Enzyme immunoassay ELISA and EMIT. In “Methods in Enzymology” (H. Van Vunakis and J. J. Langone, eds.), Vol. 70, pp. 419-439. Academic Press, New York. GAROFF, H., KONDOR-KOCH, C., and RIEDEL, H. (1982). Structure and assembly of alphaviruses. Curr. Top. Microbial. Immunol. 99, l-50. GATES, D., BROWN, A., and WUST, C. J. (1982). Comparison of specific and cross-reactive antigens of alphaviruses on virions and infected cells. Irlfect. Immun 35.248-255. GENTRY, M. K., HENCHAL, E. A., MCCOWN, J. M., BRANDT, W. E., and DALRYMPLE, J. M. (1982). Identification of distinct antigenic determinants on Dengue-2 virus using monoclonal antibodies. Amer. J. Trap. Med Hyg. 31, 548-555. HELENIUS, A., KARTENBECK, J., SIMONS, K., and FRIES, E. (1980). On the entry of Semliki Forest virus into BHK-21 cells. J. Cell Biol. 84, 404-420. KALUZA, G., Roar, R., and SCHWARZ, R. T. (1980). Carbohydrate-induced conformational changes of Semliki Forest virus glycoproteins determine antigenicity. Virology 102, 286-299. KALUZA, G., and PAULI, G. (1980). The influence of intramolecular disulfide bonds on the structure and
154
SCHMALJOHN.
KOKUBUN,
function of Semliki Forest virus membrane glycoproteins. VZroZogy 102, 300-309. KENNEW, R. H. (1980). Enzyme-linked antibody assay with cells attached to polyvinyl chloride plates. In “Monoclonal Antibodies” (R. H. Kennett, T. J. McKearn, and K. B. Bechtol, eds.), pp. 376-377. Plenum, New York. MOONEY, J. J., DALRYMPLE, J. M., ALVING, C. R., and RUSSELL, P. K. (1975). Interaction of Sindbis virus with liposomal mode1 membranes. J. ViroL 15,225231. RICE, C. M., and STRAUSS, J. H. (1981). Nucleotide sequence of the 26s mRNA of Sindbis and deduced sequence of the encoded virus structural proteins. Proc. Nat. Acad Sci. USA 78, 2062-2066. RICE, C. M., and STRAUSS, J. H. (1982). Association of Sindbis virus glycoproteins and their precursors. J. Mol. Biol 154, 325-348. ROEHRIG, J. T., DAY, J. W., and KINNEY, R. M. (1982a). Antigenic analysis of the surface glycoproteins of a Venezuelan equine encephalitis virus (TC-83) using monoclonal antibodies. firology 118, 269-278. ROEHRIG, J. T., GORSKI, D., and SCHLESINGER, M. J. (198213). Properties of monoclonal antibodies directed against the glycoproteins of Sindbis virus J. Gen Viral. 59, 421-425. SCHEEFERS, H., SCHEEFERS-BORCHEL, U., EDWARDS, J., and BROWN, D. T. (1980). Distribution of virus structural proteins and protein-protein interactions
AND
COLE
in plasma membrane of baby hamsters’ kidney cells infected with Sindbis or vesicular stomatitis virus. Proc. Nat. Ad Sci USA 77, 7277-7281. SCHLESINGER, M. J., and KKHRIHNEN, L. (1980). Translation and processing of alphavirus proteins. In “The Togaviruses: Biology, Structure, Replication” (R. W. Schlesinger, ed.), pp. 371-392. Academic Press, New York. 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. SEFTON, B. M., WICKUS, G. G., and BURGE, B. W. (1973). Enzymatic iodination of Sindbis virus proteins. J. ViroL 11, 730-735. SMITH, J. F., and BROWN, D. T. (1977). Envelopment of Sindbis virus: Synthesis and organization of proteins in cells infected with wild type and maturation-defective mutants. J. Vird 22, 662-678. SMITH, J. F., and PIFAT, D. Y. (1982). Morphogenesis of sandfly fever viruses (Bunyaviridae family). virology 121, 61-81. STOLLAR, V. (1975). Immune lysis of Sindbis virus. Virology 66, 620-624. SYMINGTON, J., MCCANN, A. K., and SCHLESINGER, M. J. (1977). Infectious virus-antibody complexes of Sindbus virus. I$ect. Immun 15, 720-725. YEWDELL, J. W., and GERHARD, W. (1981). Antigenic characterization of viruses by monoclonal antibodies. Annu. Rev. Microbial 35, 185-206.