Journal of General Virology (1991), 72, 557-565. Printed in Great Britain 557

Identification of linear epitopes on Semliki Forest virus E2 membrane protein and their effectiveness as a synthetic peptide vaccine

Alies Snijders, ~* Barry J. Benaissa-Trouw, ~ Tom A. M. Oosterlaken, ~ Wouter C. Puijk, 2 Willem P. A. Posthumus, 2 Rob H. Meloen, 2 Wil A. M. Boere, 1 Janine D. Oosting, l Cornelis A. Kraaijeveld I and Harm Snippe ~

1Eijkman-Winkler Laboratory o f Medical Microbiology, Academic Hospital Utrecht, H P G04.614, Postbox 85500, 3508 GA Utrecht and 2Central Veterinary Institute, Postbox 65, 8200 A B Lelystad, The Netherlands

Semliki Forest virus (SFV) infection of mice was used as a model to study the applicability of synthetic peptides containing only linear epitopes as viral vaccines. The identification of linear epitopes with vaccine potential on the E2 membrane protein of SFV was based on the binding of SFV-specific antibodies to a set of overlapping synthetic hexapeptides (Pepscan) representing the whole E2 amino acid sequence. The 14 available E2-specific monoclonal antibodies which were protective in vivo proved to be unsuitable for the identification of linear epitopes because they recog- nized only conformational epitopes, as indicated by their lack of reactivity with unfolded, reduced E2 protein on immunoblots. Three epitopes were detected with polyclonal anti-SFV serum at amino acid posi-

tions 135 to 141,177 to 185 and 240 to 246 of the E2 protein. Synthetic peptides containing these epitopes were coupled to a carrier protein and tested as a vaccine. Mice immunized with the peptide containing amino acids 240 to 255 of protein E2 were protected against a challenge with virulent SFV but protection of mice immunized with the peptides containing amino acids 126 to 141 or 178 to 186 was only marginally better than that of controls. The prechallenge sera of most peptide-immunized mice reacted with SFV- infected cells but none of these sera neutralized the virus in vitro. However, protection of mice correlated well with SFV-specific antibody titre, suggesting antibody-mediated protection.

Introduction

Chemically synthesized oligopeptides have become popular tools for the characterization of the interaction between protein antigens and the immune system. They allow the fine mapping, on proteins of infectious agents, of areas that are critical for induction of protective humoral and cellular immune responses (B and T cell epitopes). This technique led to the concept of synthetic immunogens for vaccination against infectious diseases (Arnon, 1972; Lerner, 1982). The advantages of such vaccines would be enormous in terms of safety, stability, and ease and cost of production. Although synthetic peptides have the disadvantage of containing only linear epitopes, the feasibility of synthetic viral vaccines is suggested by the great number of studies reporting successful in vitro neutralization of virus infectivity by antipeptide antibodies for, among others, tobacco mosaic virus (Anderer & Schlumberger, 1965), poliovirus (Emini et al., 1984), foot-and-mouth disease virus (FMDV) (Bittle et al., 1982; Pfaffet al., 1982; Geysen et al., 1985), rhinovirus (Francis et al., 1987; McCray &

Werner, 1987), hepatitis B virus (HBV) (Neurath et al., 1986) and human immunodeficiency virus (Chanh et al., 1986; Ho et al., 1988). In contrast to the in vitro results, there are only a few reports of full protection of test animals by peptide vaccines, i.e. for FMDV (Bittle et al., 1982; Dimarchi et al., 1986) and HBV (Itoh et al., 1986; Thornton et al., 1987). For other viruses, like murine hepatitis virus (Talbot et al., 1988), Friend murine leukaemia virus (Bayer & Hunsmann, 1987), influenza virus (MUller et al., 1982) and herpes simplex virus (Eisenberg et al., 1985; Weijer et al., 1988), the neutralizing and protective properties of antipeptide antibodies have been weak or absent. Furthermore, a problem in the development of peptide vaccines is that for many viral diseases it is not possible to measure protection, owing to the lack of a suitable animal model.

To gain more insight into the applicability of peptide vaccines for viral diseases, we used Semliki Forest virus (SFV) infection of mice as a model. This system enabled us to test the effectiveness of peptide vaccines in a host that is highly susceptible to the virus. SFV is an alphavirus in the family Togaviridae. The SFV particle

0000-9959 © 1991 SGM

558 A . S n i j d e r s a n d o t h e r s

consists of an RNA-containing capsid surrounded by an envelope in which the viral spikes, composed of trimers consisting of the three membrane proteins, E 1, E2 and E3 (Vogel et al., 1986), are anchored. The nucleotide sequence of the genes encoding the structural proteins of SFV (capsid and membrane proteins) is known for the (avirulent) prototype strain of SFV (Garoff et al., 1980), making it possible to synthesize peptides representing regions of these SFV proteins. Virulent strains of SFV cause lethal encephalitis in mice. An important role for antibodies in protection against SFV is indicated by the results of previous studies (Boere et al., 1983, 1984) which demonstrated that both El- and E2-specific monoclonal antibodies (MAbs) can protect mice against an otherwise lethal infection with SFV, and that protection can even be conferred by non-neutralizing antibodies.

Selection of peptides with vaccine potential was based on the identification of linear antibody-reactive epitopes on the E2 membrane protein. The epitopes recognized by protective, SFV-specific MAbs would be the most attractive candidates for peptide vaccines but only applicable if they are linear epitopes. To determine the ability of the available E2-specific antibodies to interact with linear structures, these antibodies were tested for their reaction with unfolded E2 (produced by treatment with SDS and 2-mercaptoethanol). Effective unfolding of the SFV E1 and E2 proteins by reduction with 2- mercaptoethanol would be expected owing to the large number of cysteine residues in the external domains of E 1 and E2 (16 and 12 respectively) at conserved positions (Dalgarno et al., 1983), suggesting many disulphide bonds. To map linear epitopes, a complete set of overlapping hexapeptides (Pepscan) representing the E2 membrane protein sequence was synthesized simulta- neously and tested in a subsequent ELISA with SFV- specific antibodies; peptides containing linear epitopes were tested as vaccines. The possibility of finding protective, linear epitopes on the SFV E2 membrane protein had been suggested by the results of Grosfeld et al. (1989), which showed that certain cloned fragments of E2 can confer partial to complete protection against lethal SFV challenge.

Methods

Virus strains. The prototype strain of SFV (Henderson et al., 1970), obtained from Dr H. Garoff (Garoffet al., 1980), and the virulent strain ofSFV, SF/LS 10 C1/A (Bradish et al., 1971), were grown and purified as described previously (Snijders et al., 1989; Kraaijeveld et al., 1979).

Preparation o f ant i-SFV immune sera. The immune mouse serum used in the Pepscan was prepared against prototype strain SFV passaged once in L cells (resulting in an SFV titre of l09 p.f.u./ml). Male BALB/c mice were injected intraperitoneally (i.p.) with l0 s p.f.u. SFV in 0.5 ml phosphate-buffered saline (PBS) and boosted twice (2 and 6 weeks

later) with 5 x 106 p.f.u. SFV by i.p. injection. Sera were obtained 1 week after the last booster. This anti-SFV serum also contained antibodies to calf serum, due to its presence in the culture medium. Therefore, a second anti-SFV mouse serum was prepared, using a brain suspension from BALB/c mice containing the prototype strain of SFV, as described (Snijders et al., 1989). The latter serum was used for ELISA and immunoblot assays.

SFV-specific MAbs. Mouse MAbs, UM 4.2, UM 8.22, UM 8.77, UM 8.73, UM 8.48, UM 5.1, UM 8.62, UM 8.107, UM 8-20, UM 8.47, UM 8.64, UM 8.115 and UM 8.139, prepared against the avirulent strain of SFV, MRS MP 192/7, are described by Boere et al. (1984, 1986). Mouse MAbs UM 1.1, UM 1-2, UM 1.3, UM 1.4, UM 1-10, UM 1.13, UM 1.26 and UM 1.27 were prepared by T. A. M. Oosterlaken against the prototype strain of SFV. The procedure was similar to that for the other MAbs except that SP2/0 cells were used for fusion and bybridomas were selected for production of SFV-neutralizing, instead of SFV- binding, antibodies.

Preparation o f antipeptide immune sera. Peptides containing amino acid sequences corresponding to regions of the membrane protein E2, with a cysteine residue added to either the N or C terminus were synthesized by Dr J. van Zoelen (Hubrecht Laboratory, Utrecht, The Netherlands) on a Biosearch SAM-2 synthesizer by the solid-phase method of Barany & Merrifield (1980). The amino acid composition of the peptides was confirmed by analysis of the total hydrolysate. Peptides were treated with 2-mercaptoethanol, purified by gel filtration and coupled to the carrier protein, keyhole limpet haemocyanin (KLH), by the method of Lee et al. (1980), using 7-maleimidobutyric acid N-hydroxysuccinimide ester (GMBS; Calbiochem) as the cou- pling agent. After treatment of KLH with GMBS, which resulted in between eight and 20 maleimide groups per 100K of KLH, a 1-2-fold molar excess of reduced peptide was added to the pretreated KLH.

Each peptide-KLH conjugate (1 mg containing about 0.1 to 0.3 mg peptide) was emulsified in Freund's complete adjuvant (FCA) and injected intramuscularly into a New Zealand White rabbit. A booster injection with 1 mg peptide-KLH conjugate emulsified in Freund's incomplete adjuvant (FIA) was given 3 weeks later. Immune serum was obtained 2 weeks after the booster. Control sera were obtained by preimmunization bleeding of the same rabbits.

Groups of inbred male BALB/c mice were injected i.p. with 50 ~tg peptide-KLH emulsified in FCA. Booster injections were given after 2 and 4 weeks with the same dose of peptide-KLH in FIA (protocol 1). Alternatively, the groups of BALB/c mice were injected subcutaneous- ly (s.c.) with a mixture of 50 ~tg peptide-KLH and 50 ~tg adjuvant Quil A (Superfos Biosector) and received one s.c. booster injection with the same dose of peptide-KLH and Quil A 5 weeks later (protocol 2). Immune sera were obtained about 2 weeks after the last booster; sera from the mice of each group were pooled.

Solid-phase ELISA. Peptides (1 p.g/well) were coated in high-binding microtitre plates (Nunc; Maxisorb) by drying peptide solution overnight at 37 °C. An SFV prototype preparation (obtained by infecting L cells at an m.o.i, of 0.5 and harvesting the supernatant after an infection period of 18 to 20 h) was similarly coated by drying. A coat of SFV-infected L cell monolayers was obtained by virus inoculation at an m.o.i, of 40 and subsequent incubation for 6.5 h at 37 °C (Boere et al., 1986). The cells were fixed with 0.05~ glutaraldehyde and washed. Sera were diluted in PBS, 0.5 ~ Tween, supplemented where necessary with 3 ~ dried milk proteins (Protifar; Nutricia) or 5 ~ calf serum. Horseradish peroxidase (HRPO)-coupled goat anti-rabbit Ig and goat anti-mouse Ig were obtained from Nordic.

S D S - P A G E and immunoblot analysis. SDS-PAGE and blotting (40 mA, 2 to 3 h) were performed by standard techniques. Blots were preincubated overnight in'PBS containing 3 ~ bovine serum albumin.

Synthetic peptide vaccines to S F V 559

T a b l e 1. Characteristics and reactivities o f E2-specific MAbs

Antibody In vivo MAb* isotype Neutralization1" protection:~

Binding to E2 on immunoblot§

Non-reduced reduced Cloned Epitope location

UM 4.2 ~ IgG2a 0 + + - - UM 8.221 IgG2a 0 + + + - - UM 8.771 IgG2b 0 + + - - UM 8.73 IgG2a 0 + + - - UM 8.48 IgG2a 1-5 + + - - UM 5 . 1 2 IgG2a 5-3 + + + + - UM 8.622 IgG2a 6.0 + + + - - UM 8.107 IgG1 0 - + + + + +

UM 1.1 IgG2a 5.6 (5-0) + + + UM 1 .33 IgG2a 1.7 (0-7) + + UM 1.103 IgG2a 1.4 (1-5) + + UM 1 . 4 4 IgG2a 4.6 (4.2) + + + UM 1-274 IgG2a 2-5 (2-6) + + UM 1.13 IgG2a 5.8 (5-2) + + + UM 1.26 IgG2a 4.7 (5.2) + + +

169 to 231 of E2

* MAbs marked with the same number are mutually competing. t Log10 of the reciprocal ascites dilution giving 50% plaque reduction of the virulent (or the prototype) strain

of SFV. + + , Full protection after i.v. injection of 1 gg purified MAb against i.p. challenge with virulent SFV; + ,

full protection with 100 ~tg MAb; +, partial protection with 100 p.g MAb; -, no protection. § Immunoblot assays as described in Fig. 1. Tested cloned fragments of E2 are described in the text.

The subsequent ELISA was performed in PBS with 0-5% Tween and 3% dried milk proteins.

Antibody-dependent complement-mediated cytotoxicity (ADCMC). ADCMC of SFV-infected L cells by the concerted action of antibodies and 10% guinea-pig complement was performed essentially as described (Boere et al., 1985) except that antiserum and complement dilutions were added 4 h after infection of the 51Cr_labelled L cells. The supernatant was harvested 4 h later.

Neutralization and protection studies. The 50 % plaque reduction assay for both strains of SFV was performed as described previously (Boere et al., 1984).

Passive protection studies were performed by injecting groups of five or six male BALB/c mice intravenously (i.v.) with immune serum or purified MAb (diluted in 0.5 ml PBS), followed 3 h (MAb) or 24 h (serum) later by an i.p. challenge with 16 p.f.u, virulent SFV. Control mice received the same dose of a corresponding control serum or a control MAb before challenge. To quantify protection, mice were observed for 21 days.

Protection by vaccination was studied in groups of between five and 10 male BALB/c mice immunized with peptide-KLH conjugate (as described for antipeptide sera). Control mice received only KLH (or PBS) and adjuvant. All mice were challenged i.p. 2 or 3 weeks after the last booster with 400 p.f.u. (mice immunized by protocol 1) or 16 p.f.u. (mice immunized by protocol 2) virulent SFV.

Results

Reaction o f SFV-specific antibodies with reduced and non-reduced E2 protein

T o d e t e r m i n e t h e c a p a c i t y o f E2 - spec i f i c M A b s to r e a c t

w i t h l i n e a r e p i t o p e s , t h e e f fec t o f u n f o l d i n g a n d

1 2 3 4 5 6 7 8 9 10 1112

E2--

E l - -

C - -

E2/E1

Fig. 1. Immunoblot of non-reduced and reduced SFV proteins after reaction with SFV-specific antibodies. Whole SFV (about 10 ixg/lane) was heated in sample buffer containing 1% SDS without (lanes 1, 3, 5, 7, 9 and 11) or with (lanes 2, 4, 6, 8, 10 and 12) 5% 2-mercaptoethanol. The proteins were separated on 10% SDS-polyacrylamide gels and blotted onto nitrocellulose. The blots were cut into strips and pairs of strips were subjected to an indirect ELISA using either a 100-fold dilution of polyclonal mouse serum against SFV (lanes 1 and 2) or 100- fold diluted ascites fluid containing MAbs UM 4.2 (lanes 3 and 4), UM 5.1 (lanes 5 and 6), UM 8-107 (lanes 7 and 8), UM 1.27 (lanes 9 and 10) or UM 8.139 (lanes 11 and 12) and HRPO-coupled goat anti-mouse IgG.

r e d u c t i o n o f t h e E 2 p r o t e i n o n t h e b i n d i n g o f t h e s e

a n t i b o d i e s i n i m m u n o b l o t a s s a y s w a s i n v e s t i g a t e d .

F i f t e e n d i f f e r e n t M A b s f r o m l 0 d i f f e r e n t c o m p e t i t i o n

g r o u p s w e r e t e s t e d ( T a b l e 1; F ig . 1). Al l 14 p r o t e c t i v e

560 A. Snijders and others

2.0 (a)

1.0 O -= ._= 0 "~ 0.5

0 0-5

0

(b)

100 200 300 400 Amino acid no.

Fig. 2. Antigenic profiles (Pepscan) for the E2 membrane protein of SFV. Overlapping hexapeptides homologous to amino acids 1 to 6, 2 to 7, 3 to 8 etc. of the E2 membrane protein of SFV were synthesized on a solid support. The support-coupled peptides were subjected to indirect ELISA using 500-fold dilutions of hyperimmune serum raised in mice against whole SFV of the prototype strain (a), the same serum after adsorption to purified intact prototype SFV (b), or MAb UM 5.1 ascites fluid (c), and a subsequent incubation with a dilution of HRPO-coupled goat anti-mouse Ig. Results are shown as vertical lines proportional to the extinction value obtained in the ELISA against the location of the NHz-terminal amino acid of each peptide within the E2 sequence.

MAbs bound to SDS-treated, non-reduced E2 on immunoblots, but weakly (MAb UM 5.1) or undetect- ably (the other 13 MAbs) to SDS-treated, reduced E2. In contrast, MAb UM 8.107 bound the non-reduced and reduced forms of the E2 protein equally well; however, this MAb does not neutralize the virus in vitro and is non- protective in vivo. Similar results were obtained with five different El-specific MAbs (not shown). These results indicate that protective and, especially, neutralizing SFV-specific MAbs generally recognize conformational epitopes that are destroyed by reduction of the disulphide bonds. non-reduced hybrid protein expressed from plasmid pSH3-2 in Escherichia coli, which contains amino acids 1 to 350 of E2 (Snijders et al., 1989), was bound only by the non-protective MAb UM 8-107. When tested with smaller cloned fragments of E2, the epitope recognized by MAb UM 8.107 could be localized to amino acids 169 to 231 (results not shown).

Identification of epitopes on the E2 membrane protein using polyclonal anti-SFV serum

Although the tested panel of protective, SFV-specific antibodies recognized only conformational epitopes, there might still be other protective antibodies present in anti-SFV serum that recognize linear epitopes. Using the published nucleotide and amino acid sequences for the structural proteins of the prototype strain of SFV (Garoff et al., 1980), a complete set of overlapping hexapeptides, corresponding to amino acids 1 to 6, 2 to 7, 3 to 8 etc. of the E2 membrane protein of SFV, was synthesized on a solid support as described (Geysen et al., 1984). The

support-coupled peptides were subjected to an indirect ELISA with hyperimmune serum raised in mice against L cell-derived SFV of the prototype strain (Fig. 2a), or with the same serum after absorption with a concen- trated suspension of purified intact prototype strain SFV (Fig. 2b). This resulted in the identification of three major continuous regions (linear epitopes) on the E2 membrane protein recognized by antibodies in poly- clonal anti-SFV serum. These continuous epitopes were located at amino acids 135 to 141,177 to 185 and 240 to 246.

Recognition of hexapeptides by SFV-specific MAbs

The support-coupled peptides were also subjected to an indirect ELISA with eight different E2-specific MAbs, UM 4.2, UM 8.22, UM 8-77, UM 8-73, UM 8.48, UM 5-1, UM 8.62 and UM 8.107 (Table 1), which recognized at least five different epitopes (Boere et al., 1984, 1986). The result of the Pepscan of E2 with MAb UM 5-1 is shown in Fig. 2 (c). The results obtained with these MAbs were negative, as would be expected from their lack of binding to reduced E2, but even the non-protective UM 8-107, which did recognize reduced E2 and cloned fragments of E2 on immunoblots, did not bind any of the hexapeptides in the Pepscan.

Reactivities o f antipeptide antibodies

Peptides containing the epitopes identified with poly- clonal anti-SFV serum were synthesized, coupled to the carrier protein KLH and injected with adjuvant (FCA or Quil A) into mice and also, to obtain large quantities of serum, into rabbits. The resulting antipeptide sera were tested for their binding and neutralizing capacities. Measured by indirect ELISA both rabbit and mouse antipeptide antibodies bound very well and specifically to homologous peptide (Table 2). The antibody titre varied between peptides but mouse and rabbit sera against the same peptide had comparable antibody titres, although on the basis of body weight the mice received a five- to 10-fold higher dose of peptide-KLH conjugate. Rabbit and mouse antibodies to peptide 240 to 255 were clearly cross-reactive with SFV because they reacted quite well with SFV-infected cells (Table 2; Fig. 3) but the cross-reactivity with SFV of the antibodies raised to the other peptides was rather weak. However, for three of the rabbit antipeptide sera, the E2 specificity could be confirmed by their binding to the E2 membrane protein of SFV on immunoblots (Table 2). The antibody titres of individual sera of mice immunized with peptide 126 to 141 or 178 to 186 (measured by ELISA on SFV- infected L cells) varied from < 2 to 3, and for peptide 240 to 255 from 3 to > 3.5. We chose to measure antibody

Synthetic peptide vaccines to S F V 561

Table 2. Reactivities of antipeptide sera

Antibody titre:~ Antiserum Binding Lysis

against Immunized Homologous SFV-infected to E2 on in Neutralization peptide* animal'~ peptide SFV L c e l l s immunoblot§ ADCMC (~)]t titre¶

126 to 141 m 4.5 <2 <0.5 r 4-5 <2 <2 + <0.5

134 to 141 r >5.5 2.5 2-5 + + 15 <0-5 178 to 186 m 4.0 2 <0-5

r 3.5 <2 <2 - <0.5 240 to 255 m 5.0 3.5 <0.5

r > 5-5 3 3.5 + + 40 < 0.5

* Numbers indicate the amino acid positions of the region of E2 present in the peptide. t The antisera were either pooled prechallenge sera of six mice immunized by protocol 1 (m), or individual rabbit sera (r).

- Log10 of the highest serum dilution that resulted in a ratio of/> 1.8 between the extinction values obtained for immune and control serum in an indirect ELISA on the indicated specific coat and, concomitantly, in a difference > 0.10 between the extinction value obtained with immune serum at that dilution on the indicated specific coat and the extinction value on the corresponding control coat (heterologous peptide, calf serum or uninfected L cells).

§ Binding of 100-fold dilutions of rabbit antipeptide sera to E2 on immunoblots in an indirect ELISA performed as described in Fig. 1.

H ADCMC of SFV-infected L cells using 30-fold diluted rabbit antipeptide serum. Less than 5 % lysis was measured in controls of infected cells with antiserum and inactivated complement and 7.5% with control serum and active complement.

¶ Logl0 of the reciprocal serum dilution that gave a 50~ plaque reduction of prototype SFV.

1.01

O

~ 0.5 U~

I I I

•

a N o

\ .

2 3 4

I i t

(b)

\

I I I

2 3 4 Reciprocal serum dilution (log~0)

Fig. 3. Antipeptide and antiviral reactivity of rabbit serum against peptide 240 to 255 measured by indirect ELISA. The extinction value obtained in ELISA on a coat of peptide 240 to 255 (a) or on a coat of SFV-infected L cells (b) was plotted against the loglo of the reciprocal serum dilution, using rabbit serum against peptide 240 to 255 (O), preimmune serum from the same rabbit (O), mouse serum against SFV (prototype strain) grown in L cells (A), mouse serum against L cells (Z~), mouse serum against SFV obtained from mouse brain (V) or normal mouse serum (V) and HRPO-eoupled goat anti-rabbit or goat anti-mouse Ig.

titres to SFV main ly on SFV-infected L cells because much smaller amounts of SFV were needed for that method and the results on SFV or SFV-infected cells seemed comparable (Table 2). Fur thermore , the two rabb i t ant ipept ide sera wi th the highest an t ibody titre on (fixed) SFV-infected cells proved able to cause comple- ment -media ted lysis ( A D C M C ) of SFV-infected cells

(Table 2) in vitro in the presence of 10~o guinea-pig complement , showing that the an t ipept ide ant ibodies also reacted with unfixed SFV-infected cells. However,

in spite of the cross-reactivity with SFV, none of the ant ipept ide sera (Table 2) showed any detectable capacity to neutral ize SFV in vitro, whereas the poly- clonal mouse an t i -SFV sera had neutra l iza t ion titres of 4.5 (for the serum against L cell-grown SFV) and 3-6 (for the serum against bra in-passaged SFV) in the same test.

Protection of peptide-immunized mice

The same antibody-selected peptides that were used to raise ant ipept ide sera were (after conjugat ion to K L H ) tested as a vaccine against lethal SFV infect ion in male BALB/c mice. Of 21 mice immunized with the K L H - coupled peptide con ta in ing amino acids 240 to 255 of E2,

20 were protected against a challenge with otherwise lethal doses of vi rulent SFV (Table 3), as were some of the mice immunized with the KLH-coupled peptide 126 to 141 or 178 to 186. A high dose of v i ru lent SFV was needed to study protect ion of mice i mmun ized by protocol 1 using F C A as the adjuvant , otherwise substant ia l number s of control mice (which received only FCA, or K L H and F C A ) survived after challenge. Even after challenge with as high a dose as 400 p.f.u, non- specific survival sometimes occurred (results not shown). Us ing i m m u n i z a t i o n protocol 2 with Quil A as the adjuvant , no non-specific survival was observed after challenge with 16 p.f.u. SFV, the 10 LDs0 dose for na ive male BALB/c mice, as de te rmined previously (Kraaije-

veld et al., 1979). All surviving mice immunized by protocol 1 had

neutra l iz ing ant ibodies in their serum 3 weeks after the

562 A. Snijders and others

Table 3. Survival of peptide-immunized mice*

Immunization of mice Antibodies to SFV:[: Infection of mice

KLH- Serum conjugated Adjuvant Booster obtained Antibody Challenge i.p. Injected

peptide~" used Route (day) (day) titre§ (day) dose (p.f.u.) Survival

rate

Mean survival

time (days)

PBS FCA/FIA i.p. 14,28 41 < 2.0 42 400 KLH FCA/FIA i.p. 14,28 41 < 2'0 42 400 126 to 141 FCA/FIA i.p. 14,28 41 < 2.0 42 400 178 to 186 FCA/FIA i.p. 14,28 41 2.0 42 400 240 to 255 FCA/FIA i.p. 14,28 41 3.5 42 400 KLH Quil A s.c. 36 48 < 2-0 72 16 126 to 141 Quil A s.c. 36 48 < 2.0 72 16 178 to 186 Quil A s.c. 36 48 2.5 72 16 240 to 255 Quil A s.c. 36 48 > 3-5 72 16 KLH Quil A s.c. 34 49 < 2.0 64 16 240 to 255 Quil A s.c. 34 49 4.3 64 16

0/6 1/6 2/6 3/6 6/6 0/6 0/6 1/6 5/5 O/lO 9/10

5-3 6.4 7.2 8'0

5"3 5'2 6-2

6'7 10"0

* Results of three separate experiments are given. t Numbers indicate the amino acid position of the E2 region present in the peptide.

Results of pooled sera are given. Antibody titres were measured by indirect ELISA against SFV-infected L cells. § Titre as defined in Table 2.

challenge (results not shown), whereas their prechallenge sera had no neutralizing activity, indicating that some virus replication probably occurred to create an immuno- genic dose. However, only one of the five surviving mice immunized with peptide 240 to 255 by protocol 2 had neutralizing activity in its serum after challenge (results not shown). The antibody titre on SFV-infected cells of the individual prechallenge sera of the peptide-immu- nized mice (protocol 2) were also determined. The titres were < 2 for peptide 126 to 141-immunized mice and > 3.5 and 4.3 for peptide 240 to 255-immunized mice, and varied from < 2 to 3 for mice immunized with peptide 178 to 186. The prechallenge sera of individual surviving mice had SFV-specific antibody titres at least equal to, but usually higher than, those of non-surviving mice in the same group (results not shown). Immunized mice which did not survive infection had a longer mean survival time than control mice that did not survive (Table 3).

Protection by antipeptide serum

To test whether antipeptide immune serum was able to produce protective immunity against SFV in naive mice, a passive transfer experiment was performed: a group of 16 donor mice was immunized s.c. with KLH-conjugated peptide 240 to 255 and Quil A (by protocol 2). Sera were collected 2 weeks after the last booster injection and pooled; control serum was obtained from mice immu- nized similarly with KLH or Quil A alone. The undiluted pooled sera were injected i.v. into naive male BALB/c mice (0-5 ml per mouse). Blood was taken from two mice from each group of recipients 24 hours later. The SFV-

specific antibody titre of serum in immune serum recipients was 3.1 compared to 3.8 for the donor serum. Subsequently, all recipient mice were challenged i.p. with 16 p.f.u. SFV. The presence of viraemia was determined 24 h after infection in four mice from each group. Recipients of immune serum showed no detect- able viraemia in contrast to that observed in all control animals tested (Table 4). However, of six recipients of immune serum, three died having lived longer (9.0 days compared to 7.0 days) than control mice. The combined results suggest that non-neutralizing antibodies to SFV induced by peptide 240 to 255 mediate protection against virulent SFV in vivo.

Discussion

In this study, we used SFV infection of mice as a model to analyse the applicability of viral synthetic peptide vaccines. The choice of SFV infection of mice as a model for these vaccines was based on the fact that virulent strains of SFV cause a lethal encephalitis in mice which can be completely prevented by administration of antibodies (Boere et al., 1984, 1985). Peptides to be tested as a vaccine were selected on the basis of their reactivity with antiviral antibodies. Of the selected peptides (coupled to KLH), one proved able to protect mice against a challenge with virulent SFV.

Using a Pepscan of the E2 membrane protein, three linear epitopes were identified by the reactivity of antibodies in polyclonal SFV-specific serum (Fig. 2). It is likely that more linear epitopes would have been detected if the peptides in the Pepscan of E2 had been

Synthetic peptide vaccines to S F V 563

Table 4. Reduction of mortality and viraemia by passively transferred antipeptide serum

Specificity Antibody titre* Viraemial" Survival Mean of donor to SFV in serum 24 h after rate survival

serum 24 h after transfer infection time (+ S.D.) (days)

KLH <2.0 2.6 + 0-5 1/6 7-0 + 1-2 KLH-240 to 255 3.1 <0-7 3/6 9.0 + 1.0

* Mean value of the antibody titres for the sera of two mice in each group (determined as described in Table 2 by indirect ELISA on SFV-infected L cells).

t Viraemia (mean loglo p.f.u. + S.D.) was determined by plaque titration of the sera of four mice of each group.

longer, as indicated by the lack of reactivity of MAb UM 8-107 with the hexapeptides in the Pepscan although it did recognize a linear epitope on E2 (Table 1). Recently, Ariel et al. (1990) showed that a greater number of linear determinants could be detected with 20 amino acid-long overlapping synthetic peptides of the SFV E2 protein. However, for the protective MAbs the lack of binding to the peptides of the Pepscan is fully explained by the lack of binding to unfolded E2.

These findings are in agreement with those of others that determinants recognized by neutralizing antibodies on the membrane proteins of SFV (Greiser-Wilke et al., 1989; Wust et al., 1989) and Sindbis virus (Roehrig et al., 1982; Stanley et al., 1985) are conformational. Further- more, most epitopes on protein antigens are of a discontinuous nature (Benjamin et al., 1984) and the frequency of peptide-reactive antibodies is low after immunization with native protein (Jemmerson, 1987). We would like to emphasize the strong correlation found between the protective capacity of the antibody and the conformational nature of the epitope. This correlation indicates that the chances of identifying protective linear epitopes with antiviral antibodies are low. This will probably be true in many viral systems.

Carrier-coupled peptides incorporating the identified epitopes were tested as vaccines. One of them (peptide 240 to 255) proved effectively protective, causing mice immunized with this peptide to be protected against a lethal SFV challenge although the peptides induced only non-neutralizing antibodies; the two other peptides showed very limited protective capacity. These results are in agreement with those of Grosfeld et al. (1989) who showed that some fragments of E2 from SFV (amino acids 115 to 150, 218 to 288 and 289 to 352), selected on the basis of a hydrophilicity pattern and produced as hybrid proteins with fl-galactosidase in E. coli, are recognized by anti-SFV antibodies and can confer partial to complete protection although these fragments induce only non-neutralizing, SFV-specific antibodies. Our studies differ in some respects. The cloned frag- ments of E2 used by Grosfeld et al. (1989) are

considerably larger than our synthetic peptides and could, as they suggest, still retain some of the spatial conformation of the native protein. This might explain why they identify the fragment containing amino acids 289 to 352 of E2 as being the most effective in the induction of a protective response, whereas we failed to identify a linear epitope located within the 289 to 352 amino acid region. Our approach, however, led to a more precise localization of linear epitopes.

In our study, peptide 240 to 255, which showed the strongest protective capacity, induced SFV-specific antibodies to a high titre, whereas the sera of mice immunized with peptide 126 to 141 or 178 to 186, which showed only limited protection, were hardly cross- reactive with SFV. The antibody titre of the pooled prechallenge sera of the groups of peptide-immunized mice (measured on SFV-infected cells) correlated well with the survival rate of each group (Table 3). A good correlation was also found between the SFV-specifiC antibody titre of individual prechallenge sera and individual survival (results not shown). These combined results indicate that the protection induced by the peptides is antibody-mediated, in agreement with the way in which they were selected. However, passive protection experiments with sera indicated that high doses of antibody are needed for protection (Table 4). The mechanism of peptide-induced protection may involve Fc receptor-mediated uptake of opsonized SFV by non-permissive mouse macrophages and antibody- dependent lysis of SFV-infected cells by complement, as described for protection by non-neutralizing, protective SFV-specific MAbs (Boere et al., 1985, 1986).

In conclusion, we have identified three linear epitopes on E2 of SFV by the reaction of anti-SFV antibodies with a Pepscan of E2. A short synthetic peptide containing one of these linear epitopes could induce protective immunity against virulent SFV in mice by inducing high levels of non-neutralizing antibodies; peptides contain- ing the other linear epitopes were only slightly protec- tive. Our results show that, although protective, SFV- specific antibodies generally recognize epitopes with a

564 A. Snijders and others

discontinuous conformational nature and, although antipeptide antibodies are generally inferior in inducing protection compared to antiviral antibodies, a peptide containing a linear epitope selected with anti-SFV antibodies proved, after coupling to a carrier and application with adjuvant, effectively protective as a synthetic peptide vaccine.

We would like to thank Hetty Vernooy, Geert Ekstijn and Martin Becker for technical assistance and Peter Capel for critically reading the manuscript. This work was supported by grant 178BT/93-138/X from the Dutch Organization for Scientific Research (NWO).

References

ANDERER, F. A. & SCHLUMBERGER, H. D. (1965). Properties of different artificial antigens immunologically related to tobacco mosaic virus. Biochimica et biophysica aeta 97, 503-509.

ARIEL, N., LEHRER, S., ELHANATY, E., SABO, T., BRODT, P., LACI-LMI, B., KATZ, D., LEVIN, R., GROSFELD, H., VELAN, B. d~ SHAFFERMAN, A. (1990). Serologically defined linear epitopes in the E2 envelope glycoprotein of Semliki Forest virus. Archives of Virology 113, 99- 106.

ARNON, R. (1972). Synthetic vaccines - dream or reality. In Immunity of Viral and Rickettsial Diseases, pp. 209-222. Edited by A. Kohn & A. M. Klinkenberg. New York: Plenum Press.

BARANY, G. & MERRIFIELD, R. B. 0980). Solid-phase peptide synthesis. In The Peptides, vol. 1, pp. 1-284. Edited by E. Gross & J. Meienhofer. New York: Academic Press.

BAYER, H. & HUNSMANN, G. (1987). Synthetic vaccines against Friend murine leukaemia virus-induced erythroleukaemia: in vivo and in vitro studies with synthetic oligopeptides and sequence-specific antisera. Journal of General Virology 68, 515-522.

BENJAMIN, D. C., BERZOFSKY, J. A., EAST, I. J., GURD, F. R. N., HANNUM, C., LEACH, S. J., MARGOLIASH, E., MICHAEL, J. G., MILLER, A., PRAGER, E. M., REICHLIN, M., SERCARZ, E. E., SMITH- GILL, S. J., TODD, P. E. & WILSON, A. C. (1984). The antigenic structure of proteins: a reappraisal. Annual Review of Immunology 2, 67-101.

BITI'LE, J. L., HOUGHTEN, R. A., ALEXANDER, J., SHINNICK, T. M., SUTCLIFFE, J. G., LERNER, R. A., ROWLANDS, D. J. & BROWN, F. (1982). Protection against foot-and-mouth disease by immunization with a chemically synthesized peptide predicted from the viral nucleotide sequence. Nature, London 298, 30-33.

BOERE, W. A. i . , BENAISSA-TRouw, B. J., HARMSEN, M., KRAAIJE- VELD, C. A. 8z SNIPPE, H. (1983). Neutralizing and non-neutralizing monoclonal antibodies to the E2 glycoprotein of Semliki Forest virus can protect mice from lethal encephalitis. Journal of General Virology 64, 1405-1408.

BOERE, W. A. M., VINJI~, J., BENAISSA-TRouw, B. J., KRAAIJEVELD, C. A. & SNIPPE, H. (1984). Identification of distinct antigenic determinants on Semliki Forest virus by using monoclonal antibo- dies with different antiviral activities. Journal of Virology 52, 575- 582.

BOERE, W. A. M., BENAISSA-TROUW, B. J., HARMSEN, M., ERICH, T., KRAAIJEVELD, C. A. & SNIPPE, H. (1985). Mechanisms of mono- clonal antibody-mediated protection against virulent Semliki Forest virus. Journal of Virology 54, 546-551.

BOERE, W. A. M., HARMSEN, M., KRAAIJEVELD, C. A. & SNIPPE, H. (1986). Antigenic differences between virulent and avirulent strains of Semliki Forest virus detected with monoclonal antibodies. Archives of Virology 88, 105-112.

BRADISH, C. J., ALLNER, K. & MABER, H. B. (1971). The virulence of original and derived strains of Semliki Forest virus for mice, guinea- pigs and rabbits. Journal of General Virology 12, 141-160.

CHANH, T. C., DREESMANN, G. R., KANDA, P., LINETrE, G. P., SPARROW, J. T., HO, n. D. & KENNEDY, R. C. (198b). Induction of anti-HIV neutralizing antibodies by synthetic peptides. EMBO Journal 5, 3065 3071.

DALGARNO, L., RICE, C. M. & STRAUSS, J. H. (1983). Ross River virus 26S RNA: complete nucleotide sequence and deduced sequence of the encoded structural proteins. Virology 129, 170-187.

DIMARCHI, R., BROOKE, G., GALE, C., CRACKNELL, V., DOLL, T. & MOWAT, N. (1986). Protection of cattle against foot-and-mouth disease by a synthetic peptide. Science 232, 639-641.

EISENBERG, R. J., CERIN1, C. P., HEILMAN, C. J., JOSEPH, A. D., DIETZSCHOLO, B., GOLUB, E., LONG, n., PONCE DE LEON, i . & COHEN, G. H. (1985). Synthetic glycoprotein D-related peptides protect mice against herpes simplex virus challenge. Journal of Virology 56, 1014-1017.

EMINI, E. A., JAMESON, B. A. t~ WIMMER, E. (1984). Identification of a new neutralization antigenic site on poliovirus coat protein VP2. Journal of Virology 52, 719-721.

FRANCIS, M. J., HASTINGS, G. Z., SANGAR, D. V., CLARK, R. P., SYRED, A., CLARKE, B. E., ROWLANDS, D. J. • BROWN, F. (1987). A synthetic peptide which elicits neutralizing antibody against human rhinovirus type 2. Journal of General Virology 68, 2687-2691.

GAROFF, H., FRISCHAUF, A.-M., SIMONS, K., LEHRACH, H. & DELIUS, H. (1980). Nucleotide sequence of cDNA coding for Semliki Forest virus membrane glycoproteins. Nature, London 288, 236-241.

GEYSEN, H. M., MELOEN, R. H. & BARTELING, S. J. (1984). Peptide synthesis used to probe viral antigens for epitopes to a resolution of a single amino acid. Proceedings of the National Academy of Sciences, U.S.A. 81, 3998-4002.

GEYSEN, H. M., BARTELING, S. J. & MELOEN, R. H. 0985). Small peptides induce antibodies with a sequence and structural require- ment for binding antigen comparable to antibodies raised against the native protein. Proceedings of the National Academy of Sciences, U.S.A. 82, 178-182.

GREISER-WILKE, I., MOENNIG, V., KAADEN, O.-R. & FIGUEIREDO, L. T. M. (1989). Most alphaviruses share a conserved epitopic region on their nucleocapsid protein. Journal of General Virology 70, 743 748.

GROSFELD, H., VELAN, B., LEITNER, M., COHEN, S., LUSTIG, S., LACHMI, B.-E. & SHAFFERMAN, A. (1989). Semliki Forest virus E2 envelope epitopes induce a nonneutralizing humoral response which protects mice against lethal challenge. Journal of Virology 63, 341 6- 3422.

HENDERSON, B. E., METSELAAR, D., KIRYA, G. B. & TIMMS, G. L. (1970). Investigations into yellow fever virus and other arboviruses in the northern regions of Kenya. Bulletin of the Worm Health Organization 42, 787-795.

Ho, D D., KAPLAN, J. C., RACKAUSKAS, I. E. & GURNEY, M. E. (1988). Second conserved domain of gp 120 is important for HIV infectivity and antibody neutralization. Science 239, 1021-1023.

ITOH, Y., TAKAI, E., OHNUMA, H., KITAJIMA, K., TSUDA, F., MACHIDA, A., MISHIRO, S., NAKAMURA, T., MIYAKAWA, Y. & MAYUMI, M. (1986). A synthetic peptide vaccine involving the product of the pre-S(2) region of hepatitis B virus DNA: protective efficacy in chimpanzees. Proceedings of the National Academy of Sciences, U.S.A. 83, 9174-9178.

JEMMERSON, R. (1987). Antigenicity and native structure of globular proteins: low frequency of peptide reactive antibodies. Proceedings of the National Academy of Sciences, U.S.A. 84, 9180-9184.

KRAAIJEVELD, C. A., HARMSEN, M. & KHADER BOUTAHAR-TROUW, B. (1979). Cellular immunity against Semliki Forest virus in mice. Infection and Immunity 23, 213-218.

LEE, A. C. J., POWELL, J. E., TREGEAR, G. W., NIALL, H. D. & STEVENS, V. C. (1980). A method for preparing fl-hCG COOH peptide-carrier conjugates of predictable composition. Molecular Immunology 17, 749-756.

LERNER, R. A. (1982). Tapping the immunological repertoire to produce antibodies of predetermined specificity. Nature, London 299, 592-596.

McCRAY, J. & WERNER, G. (1987). Different rhinovirus serotypes neutralized by antipeptide antibodies. Nature, London 329, 736-738.

Synthetic peptide vaccines to S F V 565

Mi)LLER, G. M. SHAPIRA, M. & ARNON, R. (1982). Anti-influenza response achieved by immunization with a synthetic conjugate. Proceedings of the National Academy of Sciences, U.S.A. 79, 569-573.

NEURATH, A. R., KENT, S. B. H., PARKER, K., PRINCE, A. M., STRICK, N., BROTMAN, B. & SPROUL, P. (1986). Antibodies to a synthetic peptide from the pre-S 120-145 region of the hepatitis B virus envelope are virus-neutralizing. Vaccine 4, 35-37.

PFAFF, E., MUSSGAY, M., B6HM, H. O., SCHULZ, G. E. & SCHALLER, H. (1982). Antibodies against a preselected peptide recognize and neutralize foot-and-mouth disease virus. EMBO Journal 1, 865-874.

ROEHRIG, J. T., GORSKI, D. & SCHLESlNGER, M. J. (1982). Properties of monoclonal antibodies directed against the glycoproteins of Sindbis virus. Journal of General Virology 59, 421-425.

SNI/DERS, A., BENAISSA-TROuW, B. J., OOSTING, J. D., SNIPPE, H. & KRAAIJEVELD, C. A. (1989). Identification of a DTH-inducing T-cell epitope on the E2 membrane protein of Semliki Forest virus. Cellular Immunology 123, 23-35.

STANLEY, J., COOPER, S. J. & GRIFFIN, D. E. (1985). Alphavirus neurovirulence: monoclonal antibodies discriminating wild-type from neuroadapted Sindbis virus. Journal of Virology 56, 110-119.

TALBOT, P. J., DIONNE, G. & LACROIX, M. (1988)~ Vaccination against lethal coronavirus-induced encephalitis with a synthetic decapeptide homologous to a domain in the predicted peplomer stalk. Journal of Virology 62, 3032-3036.

THORNTON, G. B., MILICH, D., CHISARI, F., MITAMURA, K., KENT, S. B., NEURATH, R., PURCELL, R. & GERIN, J. (1987). Immune responses in primates to the pre-S2 region of hepatitis B surface antigen: identification of a protective determinant. In Vaccines, "87, pp. 77-80. Edited by R. M. Chanock, R. A. Lerner, F. Brown & H. Ginsberg. New York: Cold Spring Harbor Laboratory.

VOGEL, R. H., PROVENCHER, S. W., VON BONSDORFF, C.-H., ADRIAN, A. & DUBOCHET, J. (1986). Envelope structure of Semliki Forest virus reconstructed from cryo-electron micrographs. Nature, London 320, 533-535.

WEIJER, W. J., DRIJFHOUT, J. W., GEERLINGS, H. J., BLOEMHOEF, FEIJLBRIEF, W. M., BOS, C. A., HOOGERI-IOUT, P., KERLING, K. E. T., POPKEN-BOER, T., SLOPSEMA, K., WlLTERDINK, J. B., WELLING, G. W. & WELL1NG-WESTER, S. (1988). Antibodies against synthetic peptides of herpes simplex virus type 1 glycoprotein D and their capability to neutralize viral infectivity in vitro. Journal of Virology 62, 501-510.

WUST, C. J., WOLCOTI" NICHOLAS, J. A., FREDIN, D., DODD, D. C., BRIDEAU, R. J., LEVELY, M. E. & BROWN, A. (1989). Monoclonal antibodies that cross-react with the E1 glycoprotein of different alphavirus serogroups: characterization including passive protection in vivo. Virus Research 13, 101-112.

(Received 8 October 1990; Accepted 29 November 1990)