Journal of General Virology (1992), 73, 2225-2234. Printed in Great Britain 2225

Characterization of two antigenic sites recognized by neutralizing monoclonal antibodies directed against the fusion glycoprotein of human respiratory syncytial virus

Juan Arbiza, 1 Geraldine Taylor, 2 Juan A. L6pez, ~ Julie Furze, z Sara Wyld, 2 Paul Whyte, z E. James Stott, 2 Gaii Wertz, 3 Wayne Suilender, 3 Michel Trudel 4 and Jose A. Melero 1.

1Servicio de Biologia Molecular, Centro Nacional de Microbiologia, Instituto de Salud "Carlos II1% Majadahonda, 28220 Madrid, Spain, 2AFRC Institute for Animal Health, Compton, Nr Newbury, Berkshire RG16 ONN, U.K., 3Department of Microbiology, University of Alabama Medical School, Birmingham, Alabama 35294, U.S.A. and 4Institut Armand-Frappier, UniversitO du QuObec, 531 boulevard des Prairies, C.P. 100, Laval, QuObec, Canada H7N 4Z3

Two antigenic sites recognized by neutralizing mono- clonal antibodies (MAbs) directed against the fusion (F) glycoprotein of human respiratory syncytial virus were mapped on the primary structure of the protein by (i) the identification of amino acid substitutions selected in antibody-escape mutants and (ii) the reactivity of synthetic peptides with MAbs. The first site contained several overlapping epitopes which were located within the trypsin-resistant amino-terminal third of the large F1 subunit. Only one of these epitopes was faithfully reproduced by a short synthetic peptide;

the others might require specific local conformations to react with MAbs. The second antigenic site was located in a trypsin-sensitive domain of the F1 subunit towards the carboxy-terminal end of the cysteine-rich region. One of these epitopes was reproduced by synthetic peptides. In addition, mutagenized F protein with a substitution of serine for arginine at position 429 did not bind MAbs to the second site. These results are discussed in terms of F protein structure and the mechanisms of virus neutralization.

Introduction

Human respiratory syncytial virus (RSV) is the major cause of lower respiratory tract infections in infants and young children (Mclntosh & Chanock, 1990). Early attempts at vaccination with formalin-inactivated virus failed to protect young vaccinees against infection by RSV during subsequent epidemics, despite the induction of a serum neutralizing antibody response. Moreover, vaccinated children experienced an exacerbated disease after a natural infection more frequently than control children who received an inactivated parainfluenza virus type 3 virus vaccine (Kim et al., 1969). Experimental live vaccines made with attenuated RSV also fail to induce protection in children (Belshe et al., 1982). However, seroepidemiological data indicate that high levels of naturally acquired neutralizing antibodies (maternally derived or generated after a natural infection) correlate with less severe infection or complete protection (Lam- precht et al., 1976; Glezen et aL, 1981, 1986; Hall et al., 1991).

The failure of formalin-inactivated RSV to induce a protective immune response might be due, at least in

part, to modification of selected epitopes on the fusion (F) or attachment (G) viral glycoproteins (Prince et al., 1986). Thus, a disproportionately high ratio of antibodies that bind/neutralize virus, as judged by ELISA, has been observed in the sera of children who received the inactivated vaccine (Murphy et al., 1986). However, the augmentation of disease observed in vaccinated children appears to involve cell-mediated immune mechanisms. These results emphasize the importance of a detailed characterization of RSV antigens in order to design and evaluate alternative safe vaccines.

The two major RSV external glycoproteins (G and F) represent presumptive targets of the protective immune response. Consistent with this hypothesis, polyclonal antisera raised against purified G or F protein (Walsh et al., 1987), or sera from animals inoculated with vaccinia virus recombinants expressing G or F antigens (Olmsted et al., 1986; Stott et al., 1987; Portela et al., 1989) neutralize RSV in tissue culture. Furthermore, the immunized animals are protected against a challenge with live RSV. The neutralizing and protective responses against the G protein are subgroup-specific, whereas the anti-F protein responses are broadly cross-reactive with

0001-0991 © 1992 SGM

2226 J. Arb i za and others

strains of the two (A and B) antigenic subgroups (Olmsted et al., 1986; Stott et al., 1987; Walsh et al., 1987).

Monoclonal antibodies (MAbs) with the highest neutralization index recognize epitopes of the RSV F glycoprotein (Trudel et al., 1987; Anderson et al., 1988; Beeler & van Wyke Coelingh, 1989; Garcia-Barreno et al., 1989). This protein is synthesized as an inactive precursor which is glycosylated and processed proteo- lytically during maturation to generate two subunits (F1 and F2) that remain linked by disulphide bridges (Gruber & Levine, 1983). Virus-binding competition assays between MAbs have identified several antigenic sites on the F protein molecule which are recognized by both the murine and the bovine immune systems (the latter being a natural host of bovine RSV, which is closely related to human RSV) (Taylor et al., 1992).

Several laboratories have located a few epitopes on the primary structure of the F protein by (i) using synthetic peptides to examine the binding of MAbs (Trudel et al.,

1987; Bourgeois et al., 1991; Martin-Gallardo et al., 1991) or polyclonal antibodies (Scopes et al., 1990), or (ii) isolating viruses that escape neutralization by anti-F protein MAbs in order to identify amino acid residues essential for epitope integrity (Lopez et al., 1990). Since both approaches have limitations (see Discussion), we have used the two complementary strategies to locate the epitopes recognized by several anti-F protein neutraliz- ing MAbs on the primary structure of the protein.

Methods

Cells and viruses. Wild-type and neutralization escape mutant viruses were grown in HEp-2 cells and purified from culture supernatants as described previously (Garcia-Barreno et al., 1988). The Long and A2 reference strains of RSV were plaque-purified before being used for the selection of MAb-resistant mutants. These were selected in two different ways.

(i) Escape mutants of the Long strain were isolated as described previously (Garcia-Barreno et al., 1989). Briefly, virus stocks were enriched in mutant viruses by four or five consecutive passages in the presence of the selecting antibody (47F, 7C2 or AK 13A2). The viruses were then plaque-purified in agar plates containing antibody. Several virus plaques were isolated, and their resistance to antibody neutral- ization was confirmed. A single plaque originating from each aliquot of the virus stock was chosen for further analysis.

(ii) Escape mutants of the A2 strain were selected for resistance to MAbs B4, 11, 19 or 20 by direct plaquing of the virus in agar plates containing antibody as described (Taylor et al., 1992).

Purification of'the F glycoprotein and trypsin treatment. The F protein was purified by immunoaffinity chromatography from extracts of HEp-2 cells infected with the Long strain using the method of Walsh et al. (1985) as described previously (Garcia-Barreno et al., 1989). The purified protein was digested with different amounts of trypsin for 4 h at 37 °C. Digestion was terminated by the addition of electrophoresis sample buffer (Studier, 1973) and boiling of the samples for 3 min. The

binding of MAbs to the trypsin fragments was tested by Western blotting (Towbin et al., 1979).

Antibody binding to R S V . The isolation and characterization of the MAbs used in this study have been documented (Taylor et al., 1984, 1992; Trudel et al., 1987; Garcia-Barreno et al., 1989). Antibody AK 13A2, raised against the Long strain F protein, was a generous gift from Dr. P. Coppe (Centre d'Economie Rurale, Marloie, Belgium). B4 is a bovine MAb produced by a heterohybridoma of bovine splenocytes and murine myeloma cells (Kennedy et al., 1988). All other MAbs were of murine origin. The antibodies were purified from ascitic fluids by Protein A-Sepharose chromatography and peroxidase-labelled as described (Garcia-Barreno et al., 1989).

The different viruses were tested for reactivity with anti-F protein MAbs in a direct ELISA using non-saturating amounts of each antibody previously titrated against the Long strain (Palomo et al., 1990). The competition between two antibodies for simultaneous binding to the virus was determined by using a peroxidase-labelled antibody mixed with increasing amounts of an unlabelled antibody (Garcia-Barreno et al., 1989). The capacity of an anti-idiotype rabbit antiserum, raised against MAb 47F, to inhibit the binding of MAbs to RSV was tested by ELISA as described (Palomo et al., 1990).

Peptides. The peptides indicated in Table 2 were synthesized according to the sequence of the Long strain F protein in an Applied Biosystems 430 instrument, using the solid-phase technology and t-Boc chemistry developed by Merrifield (1986). The peptides were cleaved off the resin with trifluoromethyl sulphonic acid, and purified from protecting groups and scavengers by Sephadex G-25 chromatography. The amino acid sequence of each peptide was confirmed by automated Edman degradation in an Applied Biosystems 477 protein sequencer.

The binding of MAbs to synthetic peptides or RSV cell lysate was tested by ELISA in PVC microtitre plates coated overnight with 1 to 2 I~g of peptide. PBS containing 5% pig serum was used as a blocking reagent to eliminate spurious cross-reactions (Taylor et al., 1992).

Sequencing o f f mRN As . HEp-2 cells were infected with the different viruses and harvested 30 to 40 h postqnfection, when c.p.e, was evident by the formation of syncytia. Total RNA was isolated by the isothiocyanate-CsCl method (Chirgwin et al., 1979) and poly(A) + RNA was selected by oligo(dT)-cellulose chromatography. These mRNA preparations were used for sequencing by the dideoxynucleo- tide chain termination method (Sanger et al., 1977) using reverse transcriptase and 5'-32p-labelled oligonucleotides, followed by a chase with terminal deoxynucleotidyl transferase (DeBorde et al., 1986). The primers used for sequencing were synthesized according to the reported sequence of the Long strain F protein gen e (Lopez et al., 1988) and their composition can be obtained from the authors upon request.

For mutants selected with MAb 11, cytoplasmic extracts were prepared from virus-infected HEp-2 cells treated with actinomycin D (Huang & Wertz, 1983). After centrifugation through CsC1, poly(A) + RNA was selected as above. Primer extension sequencing was performed using avian myeloblastosis virus reverse transcriptase (Air, 1979) and a synthetic oligonucleotide complementary to bases 893 to 906 in the strain A2 F protein gene sequence (Collins et al., 1984). The sequence between bases 758 and 847 was determined for the parental strain and each mutant selected with MAb 11.

Generation and expression o f mutated F glycoproteins. Plasmid pSLF88 (Lopez et al., 1988), which contains a cDNA copy of the entire F mRNA from the Long strain, was used as a template to mutagenize nucleotide 1298 by polymerase chain reaction (PCR) technology (Ho et al., 1989), as illustrated in Fig. 4(a). Oligonucleotides F1306 (5" GATTCCACTATTTTTATTGG, anti-mRNA sense) and F1287 (5' CCAATAAAAATAGTGGAATC, mRNA sense), which included the transversion C to A at position 1298 (shown in bold), were used in conjunction with oligonucleotides F11 (5' ACAATGGAGTTGCCA-

Epitopes on R S V fusion protein 2227

MAb used for selection

| m

l i !ii'"'ii.ilii ,mlLmr~l

=====.

Fig. 1. Binding of anti-F protein MAbs to RSV Long strain and escape mutant viruses. The MAbs were tested in a direct EL|SA using the purified viruses indicated at the top to coat microtitre plates. < 2 5 ~ (l-q), 25~ to 50~ (m) and > 5 0 ~ ( n ) of the absorbance values obtained with the Long strain.

ATCC, mRNA sense) and F1899-Eco R1 (5" C G G A A T T C C G A T A T - AACTATAAACTAGG, anti-mRNA sense) for the first amplification step. Oligonucleotides F743 (5' ACTACACCTGTAAGCACT) and F1899-Eco RI were used for the second amplification step.

The amplified DNA was digested with BclI and EcoRI and the intervening segment was isolated after electrophoresis in agar plates. This segment was ligated to the largest segment of the LF1 plasmid (Cristina et al., 1990) digested with the same enzymes. Bacterial colonies containing the recombinant plasmid (LF1/1298) were identi- fied by hybridization to radiolabelled Fl287 primer, and the mutated nucleotide was confirmed by direct sequencing of the insert. The recombinant LF1/1298 was used in the transient expression system of Fuerst et al. (1986), using CV1 cells infected with a vaccinia virus-T7 recombinant (10 p.f.u./cell) and subsequently transfected with plasmid DNA (1 ng/103 cells) by the lipofectin procedure (Bethesda Research Laboratories). The cells were incubated at 37 °C for 48 h, fixed with methanol:acetone and stained for immunofluorescence as described (Garcia-Barreno et al., 1988).

Results

Isolation and character&ation of neutralization escape mutants

The Long and A2 strains of human RSV were plaque- purified before being used to select viruses that escaped neutralization by MAbs 47F, AK13A2, 7C2, B4, 19 or 20 (see Methods). Each escape mutant was used as antigen in an ELISA to test its reactivity with a panel of anti-F protein MAbs (Fig. 1). The epitopes recognized by most MAbs had been previously grouped into five antigenic areas (Ia, Ib, II, III and IV) according to their

lOOb . ~ - ~ ' . "-.,~c~, + ~ .~ . /"--.,~c~,

~ii011l[[ 11 , [ i i97~I* I I I , I 1 ~ ~ i i i ! ~4~ 3 A 21 6 5 4 3 2 1 0 6 5 4 3 2 I1 ~ 47F

100

80

60

4O

20

0

I I I I I I I

ic) ~ . , ~ . ~'~z ° ~ '~e~ - - e I BCI ! ~-e~e~ ~*AK 13A2--

7C2 I I I I I I T 47F 6 5 4 3 2 1 0

d I I I I

~ _ __. . e ~ e _ - - - - - - e T C 2

. . . . • AKI3A2" e-~ ~/eB4

e~e49F --e47F I I I I 5 4 3 2

Antibody dilution (-log3)

Fig. 2. Competition for virus binding by MAbs and anti-idiotype antibodies. Peroxidase-labelled MAbs, denoted by an asterisk in (a) (47F), (b) (AK13A2) and (c) (49F), were tested in an ELISA against purified Long strain RSV in the presence of increasing concentrations of the unlabelled MAb shown. The results are expressed as a percentage of the value obtained with each labelled antibody in the absence of competitor. (d) Binding of each MAb to Long strain RS¥ in the presence of increasing amounts of an anti-idiotype antiserum raised against MAb 47F (Palomo et al., 1990).

competition for simultaneous binding to the virus (Garcia-Barreno et al., 1989). The epitopes recognized by MAbs AK13A2, B4 and 7C2 were included in antigenic area II owing only to their reactivity with escape mutants; similarly, the epitopes of MAbs 19 and 20 were included in antigenic area IV by the same criteria. The A2 strain reacted identically to the Long strain with the MAbs tested in Fig. 1 (not shown).

The mutations selected in the escape viruses affected only epitopes from the antigenic area which included the epitope recognized by the selecting MAb. For instance, mutant 4/4 did not react with any of the MAbs recognizing area II, whereas other mutants selected with the same MAb (11/3, 4, 5 and 7) reacted with MAbs 7C2 and B4, but not with MAbs 47F, 49F or AK13A2. Similarly, the mutants selected with MAbs 19 or 20 did not bind the MAbs recognizing epitopes grouped in antigenic area IV, except MAb 52F. However, in all

2228 J. Arbiza and others

(a) (b) 1 2 3 4 5 1 2 3 4

(c) 1 2 3

F1

5K

2K

(a) F2 F1

Nucleotide 1 L 1

Amino acid 1 ' A • l.d..

I o 20.5K t 0 19K

I---

V t899

q • 574

-4 26K

Fig. 3. ~denti~cati~n~fFpr~teintrypsinfragmentsreactingwithMAbs.(at~c)Severa~a~iqu~ts~fpuri~edEpr~tein(~5~tgeach)were incubated with 2 ~tg 0ane 1), 4 ixg (lane 2), 8 p.g (lane 3) or 16 ~tg (lane 4) of trypsin. Lane 5 is undigested F protein. The samples were separated by SDS-PAGE (Studier, 1973) and electrotransferred (Towbin et al., 1979) to Immobilon membranes. These were either stained with Coomassie blue (a) or developed with MAb AK 13A2 (b) or 19 (c). The positions of Mr markers and trypsin fragments are indicated to the left and right of each panel, respectively. (d) Diagram of the primary structure of the F glycoprotein denoting the hydrophobic regions ( / ) , the site of proteolytic processing (j,), the potential sites for N-glycosylation CA), the cysteine residues (O) and the amino acid residues changed in the neutralization escape mutants (~7) (see Table 1). The locations of the trypsin fragments recognized by different MAbs are shown below the diagram.

cases the mutant viruses did bind MAbs recognizing other antigenic areas.

The different reactivities of the MAbs from antigenic area II with the escape mutants indicated that their epitopes might overlap on the F protein molecule, but were not identical. To differentiate these epitopes further, we tested whether the corresponding MAbs would compete for simultaneous binding to the virus, and if their binding could be inhibited by an anti- idiotype antiserum raised against MAb 47F (Palomo et al., 1990). The results obtained (Fig. 2) indicated

extensive competition between these MAbs for virus binding; however, MAb AK 13A2 inhibited the binding of MAbs 47F and 49F in a non-reciprocal manner. In addition, the anti-idiotype antiserum inhibited the virus binding of MAbs 47F and 49F, but not that of MAbs AK13A2, 7C2 and B4 (Fig. 2d). Thus, the epitopes included in antigenic area II could be distinguished by at least one of the following criteria: (i) the reactivity of MAbs with escape mutants, (ii) the competition of MAbs for virus binding and (iii) the inhibition of virus binding by an anti-idiotype antiserum. Only the epitopes recog-

Epitopes on R S V fusion protein 2229

nized by MAbs 47F and 49F could not be distinguished by the above criteria, but these MAbs differ in both neutralizing capacity and susceptibility to denaturing agents. Thus, MAb 49F is a non-neutralizing antibody the binding activity of which is lost in Western blots, whereas MAb 47F is a highly neutralizing antibody which resists Western blot t reatment (Garcia-Barreno et al., 1989).

Location o f amino acid changes selected in neutralization escape mutants

As an initial step toward locating the epitopes recognized by the MAbs used in the selection of mutant viruses, the MAbs were tested against trypsin fragments of purified F protein by Western blotting (Fig. 3). Increasing amounts of trypsin generated smaller fragments of the F 1 subunit which were stained with Coomassie blue (Fig. 3a). Four F1 fragments of 30K, 20.5K, 19K and 15K were recognized by MAb AK13A2 (Fig. 3b). The 20-5K and 19K fragments have been mapped previously (Lopez et al., 1990) to the N-terminal end of the F1 subunit (Fig. 3d). MAbs 47F, 7C2 and B4 recognized the same set of fragments as A K 13A2 (not shown). Thus, the epitopes of antigenic area I I could be ascribed to amino acid sequences included within the N-terminal third of the F 1 fragment.

Antibody 19 reacted with a different set of F1

fragments (Fig. 3c). Only large fragments (26K and 22K), generated with low amounts of trypsin, reacted with MAb 19 (MAb 20 reacted less efficiently with the same set of fragments; data not shown). Thus, the epitopes recognized by MAbs 19 and 20 contain trypsin- sensitive amino acid sequences which were tentatively located within the C-terminal two-thirds of the F1 subunit (Fig. 3d), outside the region covered by the fragments recognized by MAb AK13A2. The N- terminal end of the 26K and 22K fragments could not be determined by direct protein sequencing owing to a low yield after trypsin treatment.

To identify the amino acid changes selected in the escape mutants, we sequenced F m R N A s obtained from cells infected with the different viruses. The F m R N A s of viruses selected with MAbs AK13A2, 7C2 and B4 were sequenced between nucleotides 420 and 920, which encode the trypsin-resistant fragments recognized by these MAbs (Fig. 3). Similarly, the F m R N A s of viruses selected with MAbs 19 and 20 were sequenced between nucleotides 1100 and 1680, which encode the region of the putative 26K trypsin-resistant f ragment recognized by these MAbs.

The changes detected in the different mutants are listed in Table 1, including two previously reported mutants selected with MAb 47F (Lopez et al., 1990). Also included in Table 1 is the sequence change observed in the prototype (C4389/1) of five mutants selected with

Table 1. Sequence changes selected in the neutralization escape mutants

MAb Nucleotide at position* Amino acid at position* used for Loss of binding selection Virus 583 659 786 797 816 827 828 1298 190 216 258 262 268 272 429 with antibodies

- Long and A2 C A U A A A A C Ser Asn Leu Asn Asn Lys Arg 11 C4389/1 U Ile NDt

47F 4 U Tyr 47F, 49F, AK13A2 7 U Ile 47F, 49F, AK13A2, 7C2, B4

AK 13A2 4/4 G U Asp Tyr 47F, 49F, AK 13A2, 7C2, B4 11/3 U Tyr 47F, 49F, AK13A2

4 U Tyr 47F, 49F, AKI3A2 5 U Tyr 47F, 49F, AK13A2 7 U Tyr 47F, 49F, AK13A2 4' G Glu 47F, 49F, AK13A2, 7C2, B4

7C2 1 G Glu 47F, 49F, AK13A2, 7C2, B4 4 A C Arg Ser 7C2

11 C Thr 47F, 49F, AK13A2, 7C2, B4 12 C Thr 47F, 49F, AK13A2, 7C2, B4

B4 61 : 16/7 C Thr 47F, 49F, AK13A2, 7C2, B4 61 : 16/8 C Thr 47F, 49F, AK13A2, 7C2, B4

19 C4848f A Ser 56F, 57F, 19, 20 C4848/1 A Ser 56F, 57F, 19, 20 C4848/3 A Ser 56F, 57F, 19, 20

20 C4902Wa A Ser 56F, 57F, 19, 20 C4902Wb A Ser 56F, 57F, 19, 20 C4902Wc A Ser 56F, 57F, 19, 20

* Only nucleotide (mRNA sense) and amino acid changes from the sequence of the Long and A2 strains of RSV are shown. t ND, Not done.

2230 J. Arbiza and others

(a) F 1287 BcZI

F-~ ~ EcoRI m s06

I BclI / \ iE oRI

F743 T7 Sp6

F 1899/EeoRl ~ @

BclI ~ EcoRl

T7 ~ Sp6

Fig. 4. Generation of a recombinant F glycoprotein mutated at nucleotide 1298, and reactivity with MAbs. (a) Scheme of the isolation of recombinant LF1]1298, which contains the RSV strain Long F glycoprotein cDNA with a single transversion of C to A at nucleotide 1298, cloned in the polylinker of pGEM4 (see Methods). (b to e) CVI cells were infected with vTF7-3 and transfected with LF1/1298 DNA. The cells were fixed 48 h later and processed for immunofluorescence with MAb 7C2 (a), 47F (b), AKI3A2 (d) or 19 (e).

MAb 11. This MAb competed for virus binding with other MAbs recognizing epitopes from antigenic area I I (Taylor et al., 1992) and selected mutants with a single transversion (A to U) at nucleotide 816, which changed the asparagine residue at position 268 to isoleucine. This change is identical to that found in mutant 7 selected with MAb 47F, which led to the loss of all the epitopes included in antigenic area I I (Fig. 1).

Four viruses selected with MAb AK13A2 (11/3, 4, 5 and 7) had a single transversion (A to U) at nucleotide 797 which changed the asparagine at position 262 to tyrosine. This change eliminated the binding sites for antibodies 47F, 49F and AK13A2 (see Fig. 1) and is identical to the change observed in mutant 4 selected with MAb 47F. A fifth virus selected with M A b AK13A2 (4/4) had, in addition, a transition (A to G) at

Epitopes on R S V fusion protein 2231

Table 2. Reactivity of MAbs with synthetic peptides

MAb

Peptide 7C2 47F AK 13A2 11 B4 19 20

25~273 <2.0* <2.0 <2.0 <2.0 <2.0 <2-0 <2-0 255-275 <2.0 <2.0 <2.0 <2-0 6.7 <2.0 <2.0 258-271 <2.0 <2.0 <2-0 <2-0 <2.0 <2.0 <2.0 417~32 <2.0 <2.0 <2.0 <2.0 <2.0 2.8 <2.0 422~38 <2-0 <2.0 <2.0 <2.0 <2.0 6-0 <2~0 435~50 <2-0 <2-0 <2.0 <2-0 <2.0 2.3 <2.0

RSVstrain A2 8.4 6.1 4.9 6.4 5.3 6.4 6.3

* log~0 titre of MAb binding to synthetic peptides dried onto wells or RSV antigen tested in an EL1SA.

nucleotide 659 which changed the asparagine at position 216 to aspartic acid. This second amino acid change led to the loss of all the epitopes from antigenic area II (Fig. 1). The last mutant selected with MAb AK 13A2 (4') had a single transition of A to G at nucleotide 827, leading to the replacement of the lysine at position 272 with glutamic acid and the loss of all the epitopes from area II.

All mutants selected with MAb 7C2, except mutant 4, contained single nucleotide changes (A to G or A to C) at positions 827 or 828 which changed Lys 272 to Glu or Thr, respectively. These changes eliminated the reacti- vity of all MAbs with the epitopes in antigenic area II. Mutant 4 had two nucleotide substitutions, at positions 583 (C to A) and 786 (U to C), which changed amino acids 190 (Ser to Arg) and 258 (Leu to Ser). The last mutant lost only the binding site for MAb 7C2, but retained its reactivity with all the other anti-F protein antibodies (Fig. 1).

The two mutants selected with MAb B4 had a single nucleotide transversion at position 828 (A to C) which changed Lys 272 to Thr. Thus, all the amino acid changes selected with MAbs recognizing epitopes from antigenic area II were clustered in a small segment of the F protein, between amino acids 262 and 272, except the changes at amino acids 258, 216 and 190, which were detected only in viruses with two amino acid substitutions.

All mutants selected with MAbs 19 or 20 contained a single C to A transversion at nucleotide 1298 which changed Arg 429 to Ser (Table 1). This amino acid change, located towards the C-terminal end of the cysteine-rich region of the F1 subunit (Fig. 3d), elimin- ated the reactivity of all the MAbs recognizing epitopes in antigenic area IV, except that of MAb 52F. The sequence results confirmed that the epitopes included in antigenic areas II and IV did not overlap in the primary structure of the F protein, as inferred from the results in Fig. 1 and 3.

Since the location on the F1 subunit of the trypsin fragments which reacted with MAbs 19 and 20 was deduced indirectly (Fig. 3), direct confirmation of the

involvement of amino acid 429 in these epitopes was sought. To this end, plasmid F88 (Lopez et al., 1988) was mutagenized at nucleotide position 1298 using PCR technology (Fig. 4a). A BclI-EcoRI fragment of plasmid LF1, which contained a Long strain F cD N A cloned into the pGEM4 vector (Cristina et al., 1990), was replaced by the in vitro mutagenized DNA. Recombinants with the expected nucleotide change were obtained, and this was confirmed by direct sequencing of the insert. One of these recombinants was used in the vaccinia virus-T7 transient expression system (Fuerst et al., 1986) to test the reactivity of the mutated F protein with MAbs. Fig. 4(b to e) shows the immunofluorescence of cells transfected with one of these recombinants. The trans- fected cells were stained with MAbs 47F, 7C2 and AK13A2, but not by MAb 19 (nor MAbs 56F and 20; data not shown). This result confirmed that a single change (Arg to Ser) at amino acid 429 of the F glycoprotein was sufficient to confer the antigenic changes observed in the escape mutants selected with MAbs 19 or 20.

Reactivity o f antibodies with synthetic peptides

Since the antibodies used to select the escape mutants reacted in Western blots with trypsin fragments of the F 1 subunit (Fig. 3), we tested whether synthetic peptides could reproduce the epitopes recognized by these antibodies.

Three peptides were synthesized with sequences corresponding to amino acids 250 to 273, 255 to 275 or 258 to 271 of the F1 subunit, which surrounded the positions changed in the mutants selected with MAbs recognizing antigenic area II. These peptides were tested in an ELISA for reactivity with MAbs (Table 2). Only antibody B4 (and another bovine MAb, B5; not shown) reacted with peptide 255 to 275. The MAb B4 titre with this peptide was similar to that obtained against an RSV- infected cell lysate. However, this antibody did not react with peptide 250 to 273 nor 258 to 271, which contained

2232 J. Arb i za and others

almost the entire amino acid sequence included in peptide 255 to 275. All other antibodies recognizing area II failed to react with any of the peptides.

Three other peptides, corresponding to the F1 se- quences between amino acids 417 and 432, 422 and 438, and 435 and 450 which surround position 429 changed in the escape mutants selected with MAbs 19 or 20, were also tested by ELISA (Table 2). Only MAb 19 reacted with peptide 422 to 438. The titre of MAb 19 with this peptide was similar to that obtained with the RSV- infected cell lysate.

Discussion

Several epitopes on the RSV F glycoprotein recognized by neutralizing antibodies were located using two complementary procedures: (i) isolation and sequencing of mutants which escaped neutralization by MAbs and (ii) synthesis of peptides with sequences containing the amino acids changed in the escape mutants. The results obtained identified two regions of the F protein primary structure in which the epitopes recognized by neutraliz- ing MAbs were located.

The first region mapped within a trypsin-resistant segment of the F1 subunit near its N terminus, and contained the overlapping epitopes recognized by MAbs 47F, 49F, 7C2, AK13A2, 11 and B4, included in antigenic area II (Fig. I and 3). Most amino acid changes found in mutants selected with these MAbs were clustered around amino acids 262 to 272. Since these MAbs reacted in Western blots with proteolytic frag- ments of the F 1 subunit, it was originally thought that they recognized linear epitopes determined by sequences of consecutive amino acids. However, it seems that some local conformation(s) is needed for the integrity of certain epitopes. For instance, the change at amino acid 216 (Asn to Asp), selected in mutant 4/4 resistant to MAb AK13A2, eliminated the reactivity with MAbs 7C2 and B4 (which selected changes at position 272). Conse- quently, some long-range effect of amino acid 216 in the structure adopted by the epitope recognized by MAb B4 in the F1 subunit is likely to occur.

The results obtained with the synthetic peptides were also indicative of conformational constraints on the epitopes in antigenic area II. The epitope recognized by MAb B4 was reproduced by peptide 255 to 275; however, other closely related peptides failed to react with this MAb. In addition, none of the peptides tested repro- duced other epitopes in antigenic area II. It is worth noting that the region of the F1 subunit containing these epitopes is resistant to high doses of trypsin, indicative of a particular three-dimensional conformation which might be preserved in Western blots, but not in short peptides.

Other studies have also identified synthetic peptides closely related to regions of antigenic area II which react with anti-F protein MAbs. Bourgeois et al. (1991) have reported the reactivity of one (RS-348) of nine MAbs with peptide 200 to 225 and Martin-Gallardo et al. (1991) have reported that peptides spanning residues 289 to 298 react with MAb L4. However, no data from mutants escaping neutralization by MAbs RS-348 and L4 has been published to confirm the epitope location deduced using the peptides.

It has been reported previously that MAb 7C2 used in this study reacts with peptides containing regions 221 to 232 and 275 to 288 of the FI subunit in a dot-blot assay (Trudel et al., 1991). However, this antibody failed to react with peptides containing those regions under the stringent ELISA conditions used in this study (not shown). Thus, caution should be taken in interpreting the results of antibody binding to synthetic peptides, which are prone to show cross-reactions (Trifilieffet al., 1991). The isolation and sequencing of neutralization escape mutants therefore are important to confirm the epitope location inferred from using synthetic peptides. In addition, the escape mutants can be used to locate conformational epitopes, not reproduced by peptides; however, long-range conformational effects of the mutations, similar to those reported for foot-and-mouth disease virus (Parry et al., 1990), should not be overlooked.

The second antigenic region (area IV) of the F protein identified in this study maps towards the end of the cysteine-rich region located in the middle of the F1 subunit. This region is very sensitive to trypsin digestion and is relatively near the amino acid sequence (479 to 490) which Scopes et al. (1991) found to react with human post-infection sera. Peptide 422 to 438 reacted strongly with MAb 19; however, MAbs 52F, 56F, 57F and 20 (having epitopes also included in area IV) did not react with either the synthetic peptides or the denatured F1 subunit. A recombinant F protein which incorporated a single amino acid substitution at position 429 (Arg to Ser) reproduced the antigenic changes observed in the viruses that escaped neutralization by MAbs 19 and 20.

The neutralization of RSV by the MAbs used in this study might be related to their capacity to inhibit the membrane fusion activity of the F glycoprotein (Garcia- Barreno et al., 1989; Taylor et al., 1984). By analogy with other paramyxoviruses (review by Morrison, 1988), it is assumed that the fusion activity of RSV depends upon the proteolytic processing of the F protein precursor. This modification generates the new N terminus of the F1 subunit, proposed to interact with lipid membranes through a short hydrophobic peptide. The antigenic areas of the F glycoprotein identified in this study are located distant to the fusion peptide in a linear map;

Epitopes on R S V fusion protein 2233

however, it is possible that the binding of antibodies to other regions of the F protein influence the activity of the fusion peptide. In this respect, mutants altered in the fusogenic activity of the influenza virus haemagglutinin ( D a n i e l s et al., 1985) h a v e b e e n m a p p e d ou t s ide the fus ion p e p t i d e o f the H A 2 subun i t . I t shou ld also be

e m p h a s i z e d t h a t in vivo p r o t e c t i o n by M A b s does n o t a lways cor re la te w i th t h e i r c a p a c i t y for in vitro neu t r a l i z a t i on .

I t is o f i n t e re s t t h a t the m u r i n e a n d b o v i n e M A b s r ecogn i z ing ep i topes in a n t i g e n i c a r ea I I o n the F p r o t e i n mo lecu l e were i so la ted i n d e p e n d e n t l y in severa l l abo ra - tories, a n d t h a t h u m a n p o s t - i n f e c t i o n sera c o n t a i n a n t i b o d i e s w h i c h c o m p e t e for v i rus b i n d i n g w i t h at

leas t some o f these M A b s ( P a l o m o et al., 1991). Thus ,

ep i topes in th is s i te o f the F p r o t e i n a re a n t i g e n i c in severa l hosts. I n a d d i t i o n , the ep i topes r ecogn ized by the a n t i b o d i e s r e cogn i z in g a reas I I a n d IV are c o n s e r v e d in a

large n u m b e r o f v i rus s t r a ins (no t shown) . Th es e resul ts are of specia l r e l evan ce for the des ign a n d e v a l u a t i o n of effect ive R S V vacc ines .

We thank Angel del Pozo for the art work. J. A. was the recipient of a training fellowship from the CEE and J.A.L. of a postdoctoral fellowship from "Instituto de Salud Carlos III". This research was supported by grants from Comision Interministerial de Ciencia y Tecnologia, Fondo de Investigaciones Sanitarias and the World Health Organization (J.A.M.), and by a British/Spanish Joint Research Programme (J.A.M. and G.T.).

References

AIR, G. M. (1989). Nucleotide sequence coding for the "signal peptide" and the N terminus of the hemagglutinin from an Asian (H2N2) strain of influenza virus. Virology 97, 468-472.

ANDERSON, L. J., BINGHAM, P. & HEIRHOLZER, J. C. (1988). Neutralization of respiratory syncytial virus by individual and mixtures of F and G protein monoclonal antibodies. Journal of Virology 62, 4235-4238.

BEELER, J. A. & VAN WYKE COELINGH, K. (1989). Neutralization epitopes of the F glycoprotein of respiratory syncytial virus: effect of mutation upon fusion function. Journal of Virology 63, 2941-2950.

BELSHE, R. B., VAN VORIS, L. P. & MUFSON, M. A. (1982). Parenteral administration of live respiratory syncytial virus vaccine: results of a field trial. Journal of Infectious Diseases 145, 311-319.

BOURGEOIS, C., CORVAISIER, C., BOUR, J. B., KOHLI, E. & POTHIER, P. (1991). Use of synthetic peptides to locate neutralizing antigenic domains on the fusion protein of respiratory syncytial virus. Journal of General Virology 72, 1051-1058.

CHIRGWIN, J. M., PRZYBYLA, A. E., MACDONALD, R. J. & RUTTER, W. J. (1979). Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18, 5294--5299.

COLLINS, P. L., HUANG, Y. T. & WERTZ, G. W. (1984). Nucleotide sequence of the gene encoding the fusion (F) glycoprotein of human respiratory syncytial virus. Proceedings of the National Academy of Sciences, U,S.A. 81, 7683-7687.

CRISTINA, J., LOPEZ, J. A., ALBO, C., GARCIA-BARRENO, B., GARCIA, J., M ELERO, J. A. & PORTELA, A. (1990). Analysis of genetic variability in human respiratory syncytial virus by the RNase A mismatch cleavage method: subtype divergence and heterogeneity. Virology 174, 126-134.

DANIELS, R. S., DOWNIE, J. C., HAY, A. J., KNOSSOW, M., SKEHEL, J. J., WANG, M. L. & WILEY, D. C. (1985). Fusion mutants of the influenza virus hemagglutinin glycoprotein. Cell 40, 431-439.

DEBORDE, D. C., NAEVE, C. W., HERLOCHER, M. L. & M.~gSAB, H. F. (1986). Resolution of a common RNA sequencing ambiguity by terminal deoxynucleotidyl transferase. Analytical Biochemistry 157, 275-282.

FUERST, T. R., N1LES, E. G., STUDIER, F. W. & MOSS, B. (1986). Eukaryotic transient-expression system based on recombinant vaccinia virus that synthesizes bacteriophage T7 RNA polymerase. Proceedings of the National Academy of Sciences, U.S.A. 83, 8122- 8126.

GARCIA-BARRENO, B., JORCANO, J. L., AUKENBAUER, T., LOPEZ- GALINDEZ, C. & MELERO, J. A. (1988). Participation of cytoskeletal intermediate filaments in the infectious cycle of human respiratory syncytial virus (RSV). Virus Research 9, 307-322.

GARCIA-BARRENO, B., PALOMO, C., PEI~IAS, C., DELGADO, T., PEREZ- BREgIA, P. & MELERO, J. A. (1989). Marked differences in the antigenic structure of human respiratory syncytial virus F and G glycoproteins. Journal of Virology 63, 925-932.

GLEZEN, W. P., PAREDES, P., ALLISON, J. A., TABER, L. H. & FRANK, A. L. (1981). Risk of respiratory syncytial virus infection for infants from low-income families in relationship to age, sex, ethnic group, and maternal antibody level. Journal of Pediatrics 98, 708-715.

GLEZEN, W. P., TABER, L. H., FRANK, A. L. & KASEL, J. A. (1986). Risk of primary infection and reinfection with respiratory syncytial virus. American Journal of Diseases of Children 140, 543-546.

GRUBER, C. & LEVINE, S. (1983). Respiratory syncytial virus polypeptides. III. The envelope-associated proteins. Journal of General Virology 64, 825-832.

HALL, C. B., WALSH, E. E., LONG, C. E. & SCHNABEL, K. C. (1991). Immunity to and frequency of reinfection with respiratory syncytial virus. Journal of Infectious Diseases 163, 693-698.

Ho, S. N., Hu~rr, H. D., HORTON, R. M., PULLEN, J. K. & PEASE, L. R. (1989). Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77, 51-59.

HUANG, Y. T. & WERTZ, G. W. (1983). Respiratory syncytial virus mRNA coding assignments. Journal of Virology 46, 667-672.

KENNEDY, H. E., JONES, B. V., TUCKER, E. M., FORD, N. J., CLARKE, S. W., FURZE, J., THOMAS, L. H. & SToa'r, E. J. (1988). Production and characterization of bovine monoclonal antibodies to respiratory syncytial virus. Journal of General Virology 69, 3023 3032.

KIM, H. W., CANCHOLA, J. G., BRANDT, C. D., PYLES, G., CHANOCK, R. M., JENSEN, K. & PARROTT, R. H. (1969). Respiratory syncytial virus disease in infants despite prior administration of antigenic inactivated vaccine. American Journal of Epidemiology 89, 422-434.

LAMPRECHT, C. L., KRAUSE, H. E. & MUFSON, M. A. (1976). Role of maternal antibody in pneumonia and bronchiolitis due to respiratory syncytial virus. Journal of Infectious Diseases 134, 211-217.

LOPEZ, J. A., VILLANUEVA, N., MELERO, J. A. & PORTELA, A. (1988). Nucleotide sequence of the fusion and phosphoprotein genes of human respiratory syncytial (RS) virus Long strain: evidence of subtype genetic heterogeneity. Virus Research 10, 249-262.

LOPEZ, J. A., PLEAS, C., GARCIA-BARRENO, B., MELERO, J. A. & PORTELA, A. (1990). Location of a highly conserved neutralizing epitope in the F glycoprotein of human respiratory syncytial virus. Journal of Virology 64, 927-930.

MClNIOSH, K. & CHANOCK, R. M. (1990). Respiratory syncytial virus. In Virology, pp. 1045-1072. Edited by B. N. Fields & D. M. Knipe. New York: Raven Press.

MARTIN-GALLARDO, A., FIEN, K. A., HU, B. T., FARLEY, J. F., SLID, R., COLLINS, P. L., HILDRETH, S. W. & PARADISO, P. R. (1991). Expression of the F glycoprotein gene from human respiratory syncytial virus in Escherichia coli: mapping of a fusion inhibiting epitope. Virology 184, 428-432.

MERRIFIELD, B. (1986). Solid phase synthesis. Science 232, 341-347. MORRISON, T. G. (1988). Structure, function and intraceUular

processing of paramyxovirus membrane proteins. Virus Research 10, 113-136.

MURPHY, B. R., PRINCE, G. A., WALSH, E. E., KIM, H. W., PARROT'r, R. H., HEMMING, V. G., RODRIGUEZ, W. J. & CHANOCK, R. M.. (1986). Dissociation between serum neutralizing and glycoprotein

2234 J, Arbiza and others

antibody responses of infants and children who received inactivated respiratory syncytial virus vaccine. Journal of Clinical Microbiology 24, 197-202.

OLMSTED, R. A., ELANGO, N., PRINCE, G. A., MURPHY, B. R., JOHNSON, P. R., Moss, B., CHANOCK, R. M. & COLLINS, P. L. (1986). Expression of the F glycoprotein of respiratory syncytial virus by a recombinant vaccinia virus: comparison of the individual contribu- tions of the F and G glycoproteins to host immunity. Proceedings of the National Academy of Sciences, U.S.A. 83, 7462-7466.

PALOMO, C., ALBAR, J. P., GARCIA-BARRENO, B. & MELERO, J. A. (1990). Induction of a neutralizing immune response to human respiratory syncytial virus with anti-idiotypic antibodies. Journal of Virology 64, 4199-4206.

PALOMO, C., GARCiA-BARRENO, B., PEI~IAS, C. & MELERO, J. A. (1991). The G protein of human respiratory syncytial virus: significance of carbohydrate side-chains and the C-terminal end to its antigenicity. Journal of General Virology 72, 669-675.

PARRY, N., FOX, G., ROWLANDS, n. , BROWN, F., FRY, E., ACHARYA, R., LOGAN, D. & STUART, D. (1990). Structural and serological evidence for a novel mechanism of antigenic variation in foot-and- mouth disease virus. Nature, London 347, 569-572.

PORTELA, A., RODRIGUEZ, J. R., RODRIGUEZ, D., RODRIGUEZ, F., GARCIA-BARRENO, B., MELERO, J. A. & ESTEBAN, M. (1989). Expression of the fusion (F) protein of human respiratory syn- cytial virus using an attenuated strain of vaccinia virus. In Genetics and Pathogenicity of Negative Strand Viruses, pp. 74-78. Edited by D. Kolakofsky & B. W. J. Mahy. Amsterdam: Elsevier/ North-Holland.

PRINCE, G. A., JENSON, A. B., HEMMING, V. G., MURPHY, B. R., WALSH, E. E., HORSWOOD, R. L. & CHANOCK, R. M. (1986). Enhancement of respiratory syncytial virus pulmonary pathology in cotton rats by prior intramuscular inoculation of formalin-inactiv- ated virus. Journal of Virology 57, 721-728.

SANGER, F., NICKLEN, S. & COULSON, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proceedings of the National Academy of Sciences, U.S.A. 74, 5463-5467.

SCOPES, G. E., WATT, P. J. & LAMBDEN, P. R. (1990). Identification of a linear epitope on the fusion glycoprotein of respiratory syncytial virus. Journal of General Virology 71, 53-59.

STOTT, E. J., TAYLOR, G., BALL, L. A., ANDERSON, K., YOUNG, K. K.-Y., KING, A. M. Q. & WERTZ, G. W. (1987). Immune and histopathological responses in animals vaccinated with recombinant vaccinia viruses that express individual genes of human respiratory syncytial virus. Journal of Virology 61, 3855-3861.

STUOIER, F. W. (1973). Analysis of bacteriophage T7 early mRNAs and proteins on slab gels. Journal of Molecular Biology 79, 237-248.

TAYLOR, G., STOTT, E. J., FERNIE, B. F., COTE, P. J., COLLINS, A. P., HUGHES, M. & JEBBETT, J. (1984). Monoclonal antibodies that protect against respiratory syncytial virus infection in mice. Immunology 52, 137 142.

TAYLOR, G., STOTT, E. J., FURZE, J., FORD, J. & SOPP, P. (1992). Protective epitopes on the fusion protein of respiratory syncytial virus recognized by murine and bovine monoclonal antibodies. Journal of General Virology 73, 2217-2223.

TRIFILIEFF, E., DUBS, M. C. & VAN REGENMORTEL, M. H. V. (1991). Antigenic cross-reactivity potential of synthetic peptides immobi- lized on polyethylene rods. Molecular Immunology 28, 889-896.

TOWBIN, H., STAEHELIN, T. & GORDON, J. (1979). Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proceedings of the National Academy of Sciences, U.S.A. 76, 4350-4354.

TRUDEL, M., NADON, F., SEGUIN, C., DIONNE, G. & LACROIX, M. (1987). Identification of a synthetic peptide as part of a major neutralization epitope of respiratory syncytial virus. Journal of General Virology 68, 2273-2280.

TRUDEL, M., STOTT, E. J., TAYLOR, G., OTH, D., MERCIER, G., NADON, F., SEGUIN, C., SIMARD, C. & LACROIX, M. (1991). Synthetic peptides corresponding to the F protein of RSV stimulate murine B and T cells but fail to confer protection. Archives of Virology 117, 59-71.

WALSH, E. E., BRANDRISS, M. W. & SCHLESINGER, J. J. (1985). Purification and characterization of the respiratory syncytial virus fusion protein. Journal of General Virology 66, 409-415.

WALSH, E. E., HALL, C. B., BRISELLI, M., BRANDRISS, M. W. & SCHLESINGER, J. J. (1987). Immunization with glycoprotein subunits of respiratory syncytial virus to protect cotton rats against viral infection. Journal of Infectious Diseases 155, 1198-1204.

(Received 20 March 1992; Accepted 11 May 1992)