A 2 gtl 1 cDNA library prepared from bovine leukae- mia virus (BLV)-producing ovine cells was screened with a cocktail of anti-BLV gp51 monoclonal antibo-.
Journal of General Virology(1992), 73, 2457-2461. Printedin Great Britain
2457
Mapping of sequential epitopes recognized by monoclonal antibodies on the bovine leukaemia virus external glycoproteins expressed in Escherichia coli by means of antipeptide antibodies J o z e f Ban, i* Stefan Czene, 1 Cestmir Altaner, a lsabelle Callebaut, 2 Viktor Krchnak, 3 M a l i k Merza, 4 Arsene Burny, 2 Richard Kettmann 2 and Daniel Portetelle 2 1Department of Molecular Virology, Cancer Research Institute, Slovak Academy of Sciences, Spitalska 21, 81232 Bratislava, Czechoslovakia, ZMolecular Biology and Microbiology Units, Faculty of Agronomy, B-5030 Gembloux, Belgium, 3Research Institute for Feed Supplements and Veterinary Drugs, 25449 Jilove u Prahy, Czechoslovakia and 4Biochemistry, Biomedical Center, S-75123 Uppsala, Sweden
A 2 gtl 1 cDNA library prepared from bovine leukaemia virus (BLV)-producing ovine cells was screened with a cocktail of anti-BLV gp51 monoclonal antibodies (MAbs). Four recombinant phages with inserts of about 2-5 kbp were isolated. One, 2 BLV-gp51-1, was sequenced and shown to encode the C-terminal part of gpS1 and all of gp30. This insert was subcloned into pEV-vrfl and expressed in Escherichia coli N-4830-1 cells. The BLV product and a series of antipeptide antibodies were used to localize the sequential epitopes
defined on BLV envelope glycoprotein gp51 by their reactivity with MAbs. Epitope B was localized to amino acids 180 to 205, B' to residues 195 to 205, D and D' to residues 218 to 237, and A to amino acids 249 to 260. All the mapped sequential epitopes were localized in the C-terminal half of BLV gp51. The results of epitope mapping with bacterially produced gp51 confirm the map obtained using native viral glycoprotein.
Retroviral envelope glycoproteins are known to play a critical role as targets for the host immune responses. Binding of these molecules to specific cell surface receptors is the first event leading to retrovirus infection. Identification and mapping of B cell epitopes on viral glycoproteins have contributed to a better understanding of their three-dimensional (3D) organization and to the definition of important regions likely to be useful for the development of subunit vaccines and diagnostic reagents. In the case of bovine leukaemia virus (BLV), the aetiological agent of enzootic bovine leukosis (Burny et al., 1990), the molecular dissection of the external glycoprotein, gp51, with a panel of monoclonal antibodies (MAbs) has allowed the definition of eight distinct antigenic sites (A to H) and three overlapping sites (B', D' and F') (Bruck et al., 1982a; unpublished results for F'). Gp51 carries three conformational epitopes (F, G and H), localized to its amino-terminal part (Portetelle et al., 1989a). These represent major determinants involved in the biological activities of the virus (infectivity and syncytium induction) (Bruck et al., 1982b), and are important target epitopes for the serological diagnosis of BLV infection (Portetelle et al., 1989b). Antibodies directed against these epitopes behave like protective
antibodies in BLV vaccination experiments (Portetelle et al., 1991; Altaner et al., 1991). Attempts to localize these conformational epitopes precisely either by comparisons of the amino acid sequences of BLV variants (Portetelle et al., 1989a; Mamoun et al., 1990) or by the use of synthetic peptides (Portetelle et al., 1989c) have been unsuccessful. However, sequential epitopes A, B, B', D, D' and E have recently been localized to the C-terminal half of native gp51 by using synthetic peptides (Callebaut et al., 1991 a). In this study, we have designed another strategy to refine the mapping of these sequential epitopes by using a recombinant unglycosylated gp51 polypeptide backbone and a series of antipeptide antibodies. The prototype BLV variant was grown in chronically infected foetal lamb kidney (FLK) cells; we used a clone (Altaner et al., 1985) derived from the FLK cell line initially established by Van der Maaten & Miller (1976) as a high virus producer. To isolate cDNAs encoding BLV gp51, a 2 gtll cDNA library was constructed according to the method of Young & Davis (1983). Recombinant phages were screened for their reactivity with a cocktail of MAbs recognizing sequential epitopes. After incubation for 4 h at 42 °C, nitrocellulose membranes soaked in 10 mM-IPTG were placed on the
0001-0964© 1992SGM
2458
Short communication
top of the agar layer, and the plates were incubated at 37 °C for 15 h. The filters were then washed in a solution of TBS buffer (10 mM-Tris-HC1 pH 7.8, 150 mM-NaCI) containing 1 0 ~ skimmed milk, and incubated for 2 h at room temperature with MAbs. After washing with TBST buffer (TBS containing 0.5 ~ Tween 20), secondary antimouse immunoglobulins conjugated with alkaline phosphatase (75 ng/ml) were added, and the filters were incubated for 2 h at room temperature. After washing with TBST buffer, plaques were visualized using a mixture of nitroblue tetrazolium chloride and 5-bromo-4chloro-3-indolyl phosphate toluidine salt at a 1 : 1 molar ratio as the chromogenic substrate. Immunoscreening of approximately 5 x 105 recombinant phages yielded four recombinant clones (2 BLVgp51-1, -2, -3 and -4), which were purified further. To confirm the binding specificity of one of these recombinant products to anti-gp51 MAbs, the lysate from lysogenic Y1089 cells infected with 2 BLV-gp51-1 was tested in a Western blot with a mixture of MAbs directed against the BLV gp51 sequential epitopes. A fusion protein, gp51-fl-galactosidase, of about 150K was specifically detected. To determine which epitopes were expressed on the recombinant product, each phage clone was tested separately with each MAb that recognized a sequential or conformational epitope on native gp51. For all four recombinant phages, a positive immunoreaction was detected only when MAbs directed against sequential epitopes A, B, B', D or D' were used. The four re¢ombinants isolated have an identical insert of about 2.5 kbp which hybridizes to a BLV envspecific probe (data not shown). To determine precisely the BLV information present in the BLV-gp51-1 recombinant phage, its EcoRI insert was subcloned into pBluescriptlI Sk ÷ and sequenced (data not shown). This cDNA, containing nucleotides 5214 to 7719 of the BLV provirus (Rice et al., 1987), encodes the C-terminal 101 amino acids of gp51 (residues 167 to 268) and the complete gp30 product (214 amino acids). Expressed in bacteria, it should give an unglycosylated product of about 32K. To express the BLV-gp51-1 product alone, unfused to fl-galactosidase, the EcoRI c D N A insert was subcloned in pEV-vrfl (Crowl et al., 1985) and expressed in Escherichia coli N-4830-1 (2 ci857) cells. The recombinant plasmid and its product were named pBLV-gp51-1 and Cgp51-gp30, respectively. The bacterial lysate was then tested in a Western blot. A protein of about 32K was detected only when MAbs directed against sequential epitopes A, B, B', D and D' were used (Fig. 1). We then took advantage of a panel of synthetic peptides covering almost the entire sequence of gp51 and the corresponding rabbit antipeptide antibodies (Calle-
M
1 2 3 4 5 6 7 8 9 10 11 N
1213
Fig. 1. Western blot reactivity of MAbs with Cgp51-gp30 (lanes 1 to 11) and virion gp51 (lanes 12 and 13). Bacterial lysate from cells expressing Cgp51-gp30 was subjected to 10~ SDS-PAGE and transferred to nitrocellulose.Lane M, cocktailof anti-gp51 MAbs with sequential epitope specificity; lanes 1 to 11, reactivity of anti-gp51 MAbs specific for epitopes A, B, B', C, D, D', E, F, F', G and H, respectively,with Cgp51-gp30 (reactivity against epitope A is weakly visible); lane N, cocktailof anti-p24 MAbs (negativecontrol); lanes 12 and 13, reactivity of MAbs specificfor epitopes B and D' with virion gp51. Table 1. Identification of gp51 peptides and their corresponding antipeptide antibodies Rabbit serum
Amino acids
R99 R21 R2659 R93 R91 R2664 R2651
168-180 169-188 195-205 218-237 249-268 255-268 260-268
Sequence LLNQTARAFPDCA LNQTARAFPDCAICWEPSPP VYNKTISGSGP NSSSFNTTQGWHHPSQRLLF PISLVNLSTASSAPPTRVRR STVSSAPPTRVRR SAPPTRVRR
baut et al., 1991a, b). The rabbit antipeptide sera directed against peptides derived from the C-terminal part of gp51 (Table 1) were used to map epitopes on the recombinant Cgp51-gp30 product in the lysate of cells transformed with pBLV-gp51-1. Two experimental approaches based on competitive solid-phase sandwich ELISA were used. In both systems the recombinant Cgp51-gp30 product was linked to a MAb of particular epitope specificity coated on the solid phase.
Short communication
2459
Table 2. Relative binding (°//o)of rabbit antipeptide antibodies to the recombinant Cgp51gp30 product adsorbed to anti-gp51 MAbs Rabbit antiserum directed against peptide Coated MAb
R99 168-180
A
81" 100 89 100 91
B B'
D D'
R21 169-188 100 10 94 99 94
R2659 195-205 76 19 IO 77 78
R93 218-237
R91 249-268
R2664 255-268
R2651 260-268
38t 84 90 73 94
14 80 87 85 100
77 75 86 70 86
92 79 100 10 lO
* Mean value of three individual ELISA experiments. t Relative binding representative of specific competition between MAbs and rabbit antipeptide antibodies for binding to the Cgp51-gp30 product is underlined.
In the first approach, the rabbit antipeptide sera were tested for their ability to bind to the MAb-Cgp51-gp30 complex; the final immune complex was assayed by using anti-rabbit immunoglobulins conjugated with peroxidase. Briefly, each microplate well was coated with 300 ng MAb and Cgp51-gp30 from the bacterial lysate was then adsorbed to the MAbs. After incubation and washing, dilutions of 1 : 1000 of each rabbit antibody were added and the adsorbed rabbit antibodies were finally visualized using a peroxidase-anti-rabbit immunoglobulin conjugate. It was observed that when the Cgp51-gp30 product was adsorbed with MAb A, the binding of rabbit antipeptide sera R2664 and R91 but not that ofsera R99, R21, R2659, R93 or R2651 was prevented (Table 2). Considering that sera R91, R2664 and R2651 recognize peptides 249 to 268,255 to 268 and 260 to 268, our results led us to map epitope A to the region between amino acids 249 and 260; epitope A was initially mapped to peptide 249 to 268 (Callebaut et al., 1991a). In the same way, epitopes D and D' were localized in the region of amino acids 218 to 237 based on the inhibition of binding of rabbit antipeptide 218 to 237 serum (R93) to Cgp51-gp30 pre-adsorbed with MAb D or D'. Epitope B' was found to map to a narrow region of amino acids defined by the rabbit antipeptide 195 to 205 serum (R2659). In the case of locating epitope B, it appeared that this epitope is defined by a stretch of amino acids encompassing two peptides recognized by the rabbit antipeptide 169 to 188 serum (R21) and the rabbit antipeptide 195 to 205 serum (R2659). No inhibition reaction occurred with rabbit antipeptide 168 to 180 serum (R99). Consequently, epitope B was localized between amino acids 180 and 205, instead of amino acids 195 and 205 (Callebaut et al., 1991a); it should be noted that the antipeptide 169 to 188 serum was raised after this previous study and the data suggest
steric hindrance between antibodies reacting with the region encompassing amino acids 180 to 205. In the second approach, immunological reactions were performed in the following order. First, anti-gp51 MAbs with different epitope specificities were coated onto wells and the recombinant Cgp51-gp30 product was added. After overnight incubation at 4°C, rabbit antipeptide serum (dilution 1 : 1000) and MAbs (200 ng) conjugated with peroxidase were tested for their mutual competition for binding to the MAb-Cgp51-gp30 complex. The presence of bound peroxidase-MAb conjugates was revealed by determination of peroxidase activity after addition of the reagents o-phenylenediamine and H202. These data are summarized in Table 3 and are in agreement with those obtained by the first approach. The influence of distance between epitopes was more pronounced in this test; the proximity of epitopes D and D' to epitopes B and B' on one side and epitope A on the other can be deduced from the data. These results also confirm that epitopes B and B' are located very close to each other. 2 gtl 1 libraries have already been used for the epitope mapping of many recombinant unglycosylated products (Wang & Strauss, 1991). In this study a 2 gt 11 library was used to isolate BLV c D N A s encoding sequential epitopes of BLV gp51. Indeed, the sequential epitopes of gp51 on the hybrid protein expressed in bacteria (the Cterminal portion of gp51 and complete gp30) were successfully determined by competitive ELISA. Apparently the presence of gp30 and the absence of glycosylation did not influence epitope mapping. The mapping of epitopes on the recombinant Cgp51gp30 product expressed in bacteria confirmed and refined the mapping achieved with native viral glycoprotein (Portetelle et al., 1989 a; Callebaut et al., 1991 a). The absence of the nucleotide sequence encoding epitope E was obviously the reason why this sequential epitope could not be localized. Site-directed mutagenesis of
2460
Short communication
Table 3. Percentage competition between rabbit antipeptide antibodies and peroxidase-MAb conjugates for binding to each adsorbed MAb-recombinant Cgp51-gp30 complex Coated MAb
Competing peroxidase-MAb conjugate
168-180
169-188
195-205
A
A
100"
B
10
B' D D'
11 10 15
95 98~ 14 14 15
86 55 97 11 17
A
12 89 53 10 12
14 100 50 ll 14
B B' D D'
11 50 98 12 10
D
A B B' D D'
D'
A B B' D D'
B
B B" D D' B'
A
Competing rabbit antiserum directed against peptide 218-237
249-268
255-268
260-268
92 12 11 96 94
95 12 17 14 13
100 14 13 16 22
90 18 10 13 10
l0 97 96 14 13
16 90 65 97 99
54 94 54 10 13
87 92 55 10 17
17 90 50 12 12
13 85 96 13 10
15 96 100 16 14
10 56 97 99 96
55 50 97 12 13
99 57 94 12 12
17 50 93 17 12
12 13 12 93 95
11 95 54 95 92
12 51 89 89 95
15 13 14 100 96
56 16 23 97 97
87 12 11 94 92
15 11 10 97 94
11 13 10 95 93
11 90 60 95 91
10 61 93 92 95
16 13 16 99 100
55 12 14 95 98
93 13 13 95 98
12 18 17 95 97
* Mean value of three individual ELISA experiments. t Results indicating specific competition are underlined.
recombinant gp51 D N A might be one of the best approaches for achieving this goal, especially in the case of epitope B. The data from this study confirm our previous results showing that gp51 can be divided into two different parts: the N-terminal part, apparently highly structured and dependent upon accurate glycosylation for the biological activity of the important conformational epitopes F, G and H, and the C-terminal part, highly glycosylated but carrying sequential epitopes obviously devoid of post-translational glycosylation (Portetelle et al., 1989a; Callebaut et al., 1991a). We are now using the accumulated data in an attempt to improve the molecular model of the oligomeric viral glycoproteins further. The authors would like to thank Mrs M. Zatkova for her technical assistance and Mrs M. Nuttinck for preparing the manuscript. This research was supported by grants from the Slovak Academy of Sciences and by Belgian Funds (Fonds Canc6rologique de la CGER, P61e d'Attraction Interuniversitaire). Josef Ban is a fellow of the Belgian
SPPS. Isabelle Callebaut and Richard Kettmann are, respectively, Research Assistant and Research Director of the Belgian National Fund for Scientific Research (FNRS).
References ALTANER, C., BAN, J., ZAJAC,V., KETrMANN,R. & BURNY, A. (1985). Isolation and characterization of cell clones producing various amounts of bovine leukosis virus. Folia biologica 31, 107-116. ALTANER, C., BAN, J., ALTANEROVA,V. & JANIK, V. (1991). Protective vaccination against bovine leukaemia virus by means of cell-derived vaccine. Vaccine 9, 889-895. BRUCK, C., MATHOT,S., PORTETELLE,D., BERTE, C., FRANSSEN,J. D., HERION, P. & BURNY, A. (1982a). Monoclonal antibodies define eight independent antigenic regions on the bovine leukemia virus (BLV) envelope glycoprotein gp51. Virology 122, 342-352. BRUCK, C., PORTETELLE, D., BURNY, A. & ZAVADA, J. (1982b). Topographical analysis by monoclonal antibodies of BLV-gp51 epitopes involved in viral functions. Virology 122, 353-362. BURNY,A., CLEUTER,Y., KETTMANN,R., MAMMERICKX,M., MARBAIX, G., PORTETELLE,n., VANDEN BROEKE,A., WILLEMS,L. & THOMAS, R. (1990). Bovine leukemia: facts and hypotheses derived from the study of an infectious cancer. In Retrovirus Biology and Human Disease, pp. 9-25. Edited by R. Gallo & F. Wong-Staal. New York: Marcel Dekker.
Short communication
CALLEBAUT,I., BURNY, A., KRCHNAK, V., GRAs-MASSE,H., WATHELET, B. & PORTETELLE,D. (1991 a). Use of synthetic peptides to map sequential epitopes recognized by monoclonal antibodies on the bovine leukemia virus external glycoprotein. Virology 185, 48-55. CALLEBAUT,I., BURNY,A. & PORTETELLE,D. (1991 b). Iodoacetamide treatment of bovine leukemia virus glycoprotein gp51 enhances the Western blotting reactivity of antipeptide antibodies. FEBS Letters 292, 148-150. CROWL, R., SEAMANS,C., LOMEDICO, P. & MCANDREW, S. (1985). Versatile expression vectors for high-level synthesis of cloned gene products in Escherichia coll. Gene 30, 31 38. MAMOUN,R. Z., MORISSON,M., REBEYROTTE,N., BUSETrA,B., COUEZ, D., KETTMANN, R., HOSPITAL, M. & GUILLEMAIN, B. (1990). Sequence variability of bovine leukemia virus env gene and its relevance to the structure and antigenicity of the glycoproteins. Journal of Virology 64, 4180-4188. PORTETELLE, O., COUEZ, D., BRUCK, C., KETTMANN,R., MAMMERICKX,M., VANDER MAATEN,M., BRASSEUR,R. & BURNY,A. (1989a). Antigenic variants of bovine leukemia virus (BLV) are defined by amino acid substitutions in the NH2 part of the envelope glycoprotein gp51. Virology 169, 27-33. PORTETELLE,D., MAMMERICKX,M. & BURNY,A. (1989b). Use of two monoclonal antibodies in an ELISA test for the detection of antibodies to bovine leukemia virus envelope glycoprotein gp51. Journal of Virological Methods 23, 211-222.
2461
PORTETELLE,D., DANDOY,C., BURNY, A., ZAVADA,J., SIAKKOU,H., GRAS-MASSE, H., DROBECQ, H. & TARTAR, A. (1989c). Synthetic peptides approach to identification of epitopes on bovine leukemia virus envelope glycoprotein gp51. Virology 169, 34-41. PORTETELLE, D., LIMBACH, K., BURNY, A., MAMMERICKX, M., DESME'I-rRE, PH., RIVIERE, M., ZAVADA,J. & PAOLETTI,E. (1991). Recombinant vaccinia virus expression of the bovine leukaemia virus envelope gene and protection of immunized sheep against infection. Vaccine 9, 194-200. RICE, N., STEPHENS,R. & GILDEN, R. (1987). Sequence analysis of the bovine leukemia virus genome. In Enzootic Bovine Leukosis and Bovine Leukemia Virus, pp. 115-144. Edited by A. Burny & M. Mammerickx. Boston: Martinus Nijhoff. VAN DER MAATEN, M. & MILLER, J. (1976). Replication of bovine leukemia virus in monolayer cell cultures. Bibliotheca haematologica 43, 360-362. WANG,K. S. & STRAUSS,J. H. (1991). Use of a lambda gtl 1 expression library to localize a neutralizing antibody-binding site in glycoprotein E2 of Sindbis virus. Journal of Virology 65, 7037-7040. YOUNG, R. A. & DAVIS, R. W. (1983). Efficient isolation of genes by using antibody probes. Proceedings of the National Academy of Sciences, U.S.A. 80, 1194-1198.
(Received 9 March 1992; Accepted 11 May 1992)
