V irus

ELSEVIER Virus Research 31(1994) 265-273 Research

Molecular basis of antigenic variation in infectious bursal disease virus

Vikram N. Vakharia a,b~*, Junkun He a,b, Basheer Ahamed b, David B. Snyder b+

a Center for Agricultural Biotechnology of Maryland Biotechnology Institute, University of Maryland, College Park, MD 20742, USA b VA-MD Regional College of Veterinary Medicine, University of Maryland,

College Park, MD 20742, USA

(Received 25 August 1993; revised 22 October 1993; accepted 29 October 1993)

Abstract

Four antigenically different strains of infectious bursal disease virus (IBDV), character- ized by their reactivities with a panel of neutralizing monoclonal antibodies (MAbs), were selected to determine the molecular basis of antigenic variation. The large genome segment A, encoding the structural proteins of the U.S. variants GLS, DS326, E/Del and the vaccine strain D78, was cloned and sequenced. Comparison of the deduced amino acid sequences of the U.S. variants with other IBDV strains showed that most of the amino acid substitutions occur in the central region between residues 212 to 332, especially in the two hydrophilic regions between residues 212 to 223 and residues 314 to 324 of VP2 protein. By comparing the amino acid sequences of these variant viruses and their reactivities with IBDV specific MAbs, the putative amino acids involved in the formation of virus-neutraliz- ing epitopes were identified. Comparison of the D78 versus PBG98 sequence showed that Gin at position 249 (Gln249) appears to be critical in binding with MAb B69. Similarly, comparison of the U.S. variant sequences with other serotype 1 sequences showed unique substitution(s) at residue Glu321 in GLS, residues Ile286, Asp318, Glu323 in E/Del, and residues Glu311 and Gln320 in DS326, which could be potential residue(s) involved in the recognition of MAb57, MAb67, and MAb179 epitopes, respectively. Comparison of the serotype 1 and serotype 2 sequences revealed that serotype 2 OH strain lacks the conserved amino acid sequence motif, S-W-S-A-S-G-S, found in all virulent strains, as well as the second hydrophilic peak region, indicating a possible role of these residues in the serotype specificty or the pathogenic&y of the virus. Phylogenetic analysis of the IBDV proteins

* Corresponding author. Fax: + 1 (301) 935-6079.

0168-1702/94/$07.00 0 1994 Elsevier Science B.V. All rights reserved SSDI 0168-1702(93)E0088-F

266 V.N. Vakharia et al. /virus Research 31 (1994) 265-273

indicated that the U.S. variants are antigenically different from geographically distant European viruses.

Key words: Infectious bursal disease virus; Amino acid sequence; Antigenic variation; Phylogenetic analysis

Infectious bursal disease virus (IBDV) is responsible for a highly contagious immunosuppressive disease in young chickens which causes significant losses to the poultry industry worldwide (reviewed in Kibenge et al., 1988). There are two serotypes of IBDV (McFerran et al., 1980). Serotype 1 viruses are pathogenic to chickens and differ markedly in their virulence (Winterfield and Thacker, 1978) whereas serotype 2 viruses, isolated from turkeys, are avirulent for chickens (Ismail et al., 1988; Kibenge et al., 1991).

IBDV is a member of the Birnaviridae family and its genome consists of two segments of dsRNA (Dobos et al., 1979). The smaller segment B (= 2800 bp) encodes VPl, the putative dsRNA polymerase (Azad et al., 1985; Spies et al., 1987), whereas the larger segment A ( = 3300 bp) encodes a 1 lo-kDa precursor protein in a single large open reading frame (polyprotein ORF) which is processed into mature VP2, VP3 and VP4 proteins (Hudson et al., 1986). Segment A can also encode putative VP5, a 17-kDa protein of unknown function, from a small ORF partly overlapping with the polyprotein ORF (Bayliss et al., 1990; Kibenge, et al., 1990). VP2 and VP3 are the major structural proteins of the virion. VP2 has been identified as the major host-protective antigen of IBDV and it contains the antigenic region responsible for the induction of neutralizing antibodies (Becht et al., 1988; Heine et al., 1991). VP3 is considered to be a group-specific antigen because it is recognized by monoclonal antibodies (MAbs) directed against VP3 from strains of both serotype 1 and 2 (Becht et al., 1988). VP4 appears to be a virus-coded protease and it is involved in processing of the precursor polyprotein (Jagadish et al., 1988).

Antigenic variation among isolates of IBDV has been reported from several laboratories (Jackwood and Saif, 1987; Snyder et al., 1988b; Van der Mare1 et al., 1990). The use of a select panel of MAbs, raised against various strains of IBDV, has led to the identification of naturally-occurring GLS, DS326 and Delaware variant viruses in the U.S.A. Substantial economic losses have been sustained due to the emergence of these antigenic variants (Delaware and GLS) in the field (Snyder et al., 1992).

In recent years, the complete nucleotide sequences of the large segment A of five serotype 1 IBDV strains [002-73 (Hudson et al., 1986) Cu-1, PBG98, 52/70 (Bayliss et al., 1990) STC (Kibenge et al., 1990), and serotype 2 OH strain (Kibenge et al., 1991)] have been determined. In addition, the VP2 gene of virulent Japanese IBDV strains (Lin et al., 1993) and Delaware variants A and E (Lana et al., 1992; Heine et al., 1991) has been sequenced. However, none of the U.S. IBDV variants have been completely cloned and characterized.

To determine the molecular basis of antigenic variation in IBDV, the genomic

KN. Vakharia et al. /l&us Research 31 (1994) 265-273 267

Table 1 Antigenic characterization of various IBDV strains by their reactivities with a panel of neutralizing MAbs

Virus strains Classification Reactivities with MAbs

B69 R63 179 8 10 57 67

D78 Classic + + + + f - - PBG98 Classic - + + + + - - STC Classic + + + + + - - 52/70 Classic + + + + + - - OH (serotype 2) Classic + + + + - - -

E/Del Variant - + + + - - + GLS Variant - - + -I- + + _

DS326 Variant - - - + + + -

segment A of four IBDV strains: GLS, DS326, Delaware variant E (E/Del) and D78 was cloned and characterized by sequencing. By comparing the deduced amino acid sequences of these strains with other serotype 1 and 2 sequences published previously, the putative amino acid residues involved in the binding with various neutralizing MAbs were identified, and the phylogenetic relationship of IBDV structural proteins was examined.

GLS, DS326 and STC strains of IBDV were propagated in the bursa of specific pathogen-free chickens (SPAFAS, Inc., Norwich, CT, USA). Tissue culture adapted E/Del-22, D78 and OH (serotype 21 strains of IBDV were propagated in primary chicken embryo fibroblast cells derived from lo-day-old emb~onated eggs (SPAFAS, Inc.) and purified as described (Snyder et al., 1988a). The MAbs against various strains of IBDV were produced and characterized using protocols previ- ously outlined (Snyder et al., 1988a,b). MAbs B69 and R63 were prepared against the D78 strain, whereas MAbs 8, 10, 57 and 179 were prepared against the GLS strain. In addition, a new MAb 67 was prepared which was neutralizing and specific for the E/Del strain. Identification of IBDV antigens by modified antigen capture ELISA (AC-ELISA) was carried out as described (Snyder et al., 1992).

Various strains of IBDV were characterized by their reactivities with a panel of neutralizing MAbs, as shown in Table 1. A plus sign indicates a relative titer level of 1 or greater on a scale of O-9 and a negative sign indicates a relative titer of less than 1 (based on the optical density readings, Snyder et al., 1988a). All standard serotype 1 viruses reacted with MAbs B69, R63, 179 and 8, except PBG98 (a British vaccine strain, Intervet, U.K.) which did not react with MAb B69. In contrast, all the U.S. variant viruses lack the virus-neutralizing B69 epitope. In addition, GLS and DS326 variants lack an R63 epitope but share a common epitope defined by the MAb 57. Thus, on the basis of the reactivities with various MAbs, these viruses were antigenically grouped as classic, GLS, DS326 and E/Del variants.

Complementary DNA clones, containing the entire coding region of the large RNA segment of various IBDV strains, were prepared using standard cloning procedures and methods previously described (Vakharia et al., 1992; Vakharia et

268 V.N. Vakharia et al. / virus Research 31 (1994) 265-273

al., 1993). The complete nucleotide sequence of these cDNA clones was deter- mined by the dideoxy method using a Sequenase DNA sequencing kit (U.S. Biochem. Corp., Columbus, OH). DNA sequences and deduced amino acid se-

__________

_._.______

________._ _________.

Fig. 1. Comparison of the deduced amino acid sequences of the structural proteins (VP2, VP3 and VP41

of ten IBDV strains. Dashes indicate amino acid identity and crosses denote a region where the

sequence was not determined. Filled bar indicates a gap in the sequence and verticle arrowheads mark

the possible cleavage sites of VP2/VP4 and VP4/VP3. The two hydrophilic peaks in the variable

region are overlined.

ViN. Vakharia et al. /‘virus Research 31 (1994) 265-273 269

Table 2 Percent amino acid sequence fiomology of large ORF of segment A of ten IBDV strains

Strain GLS DS326 E/Del D78 h-1 PBG98 52/70 STC 002-73 OH

GLS DS326 98.7 E/Del 98.4 98.3 D78 98.5 98.1 97.9 Co-1 98.6 98.2 98.0 99.6 PBG98 98.5 98.1 97.9 99.5 99.5

52/70 98.1 98.1 97.9 98.4 98.5 98.3 STC 97.7 98.0 97.5 98.4 98.5 98.3 98.3 002-73 97.0 97.1 96.7 97.6 97.7 97.6 97.3 97.4 OH 90.0 90.0 89.7 90.2 90.3 90.2 89.8 90.3 90.1

quences were analyzed by a PC/GENE software package (Int~lligenetics, Inc.). The nucleotide sequence data of the GLS strain has been deposited with GenBank Data Libraries and has been assigned an accession number M97346.

Comparisons of the nucleotide sequence of the GLS strain (3230 bp long) with eight serotype 1 and one serotype 2 IBDV strains exhibit 2 92% and 2 82% sequence homology, respectively, indicting that these viruses are closely related. It is interesting to find that there are only 6-9 base substitutions between D78, PBG98, and Cul strains which corresponds to a difference of about 0.2-0.3% (results not shown). Fig. 1 and Table 2 show a comparison of the deduced amino acid sequences and percent homology of the large ORF of segment A of the ten IBDV strains, including four IBDV strains used in this study. These comparisons show that the proteins are highly conserved. The degree of difference in the amino acid sequence ranges from 0.4% for the D78 versus Cu-1 comparison and 10.3% for the serotype 1 (E/Del) versus serotype 2 (OH) comparison (Table 2).

In Fig. 1, ~ignments of the deduced amino acid sequences of the large ORF (1012 residues) of ten IBDV strains (including four used in this study) show that most of the amino acid changes occur in the central variable region between residues 213 and 332 of VP2 protein, as shown earlier by Bayliss et al. (1990). It is interesting to note that all the U.S. variants (GLS, DS326 and E/Del) differ from the other strains in the two hydrophilic regions which are overlined in Fig. 1 (residues 212-223 and residues 314-324). These two hydrophilic regions have been shown to be important in the binding of neutralizing MAbs and hence may be involved in the formation of a virus-neutralizing epitope (Heine et al., 1991). Recently, we demonstrated that the ~nfo~ation-dependent MAbs B69, R63, 8, 179, 10, and 57 (see Table 1) i~unoprecipitate the VP2 protein (Snyder et al., 1992). In addition, E/Del specific MAb 67 also binds to VP2 protein (unpublished observations). Therefore, to identify the amino acids involved in the formation of virus-neutralizing epitopes, and hence the antigenic variation, we compared the amino acid sequences of the VP2 protein of classic and variant viruses. The deduced amino acid positions in VP2 which correlate with the binding of various MAbs are depicted in Table 3. Comparison of the D78 sequence with the PBG98

270 V.N. Vakharia et al. /Virus Research 31 (1994) 265-273

Table 3

Amino acid positions in VP2 which correlate with binding of various MAbs

Mab Virus Amino acid position

B69

179

67

10

D78

PBG98

STC

52/70

GLS

GLS

DS326

E/Del

D78

E/Del

A/Del

E/Del

fbursa)

GLS

E/Del

DS326

D78

76 249 280 326

GUY Gln ASIl Ser

Ser Aw Thr LeU

Ser Gin Am Ser

Ser Gin Asn Ser

Ser LYS Asn Ser

222 253 269 284

Thr His Ser Thr

Ser Gin Thr Ala

Ths Gin Thr Ala

Pro His Thr Thr

213 286 309 318

Asn Ile Ile Asp

Asp Ile LYS Asp Asn Ile LYS Asp

269 284 286 318 321 323

Ser Thr Thr GUY Glu Asp Thr Ala Ile Asp Ala Glu

Thr Ala Thr GUY Glu Asp Thr Thr Thr GUY Ala Asp

311

Glu

Lys GIU

Glu

323

Glu

Ghl

Glu

320

Gln

Leu

Gin

Gin

sequence shows only four amino acid substitutions at positions 76, 249, 280 and 326. Although the U.S. variants do not differ from the D78 sequence at positions 280 and 326, these viruses fail to bind to MAb B69. In addition, the STC and 52/70 strains differ from the D78 sequence at position 76 but these viruses do bind to MAb B69. This implies that Gln at position 249 (Gln249) may be involved in the binding with MAb B69. It should be noted that all U.S. variant viruses have a Gln -+ Lys substitution at this position and ,hence escape the binding with neutralizing MAb B69. Similarly, comparison of the GLS sequence with the DS326 sequence in the variable region shows six amino acid substitutions at positions 222, 253, 269, 284, 311 and 320. However, other strains of IBDV that do bind to MAb 179 have amino acid substitutions at positions 222, 253, 269 and 284 that are conservative in nature. Therefore, this suggests that Glu311 and Gin320 may be involved in the binding with MAb 179. Again, comparison of GLS and DS326 sequences with all other IBDV sequences shows a unique Ala -+ Glu substitution at position 321, suggesting the contribution of this residue in the binding with MAb 57. Since MAb 57 does not compete with MAb R63, it is conceivable that Ala321 may contribute to the binding with MAb R63. Similarly, comparison of E/Del sequence with other sequences shows five unique substitutions at positions 213, 286, 309, 318 and 323. However, comparison of this E/Del sequence (from tissue culture derived virus) with previously published VP2 A/ Del and E/Del sequences (bursa derived virus) suggests the involvement of Ile286, Asp318 and Glu323 in the binding with MAb 67 since residues at positions 213 and 309 are not substituted in

EN. Vakharia et al. /I&us Research 31 (1994) 265-273 271

A/Del and E/Del sequences, respectively (Heine et al., 1991; Lana et al., 1992; Vakharia et al., 1992). Finally, comparison of the GLS sequence with the E/Del sequence in the variable region shows six amino acid substitutions at positions 269, 284, 286, 318, 321 and 323, of which three (positions 269, 284 and 321) are GLS-specific and the other three are E/Del-specific. Again, the DS326 and D78 strains that do bind to MAb 10 have amino acid substitutions at positions 269, 284 and 321. Therefore, this suggests that one or more amino acids at positions 286, 321 and 323 may be involved in the binding with MAb 10. The precise binding of these MAbs can only be ascertained by site-directed mutagenesis studies.

Comparisons of the amino acid sequence also show a striking difference between serotype 1 and serotype 2 sequences. In the serotype 2 OH strain, there is an insertion of an amino acid residue at position 249 (serine) and a deletion of a residue at position 680. Previously, it has been shown that serotype 2 viruses are naturally avirulent and do not cause any pathological lesions in chickens (Ismail et al., 1988). Thus, these subtle changes in the structural proteins of serotype 2 OH strain may play an important role in the pathogenicity of the virus. Moreover, it has been hypothesized that an amino acid sequence motif, S-W-S-A-S-G-S, (re- sidues 326 to 332) is conserved only in virulent strains and could be involved in virulence (Heine et al., 1991). This sequence motif was also conserved in various pathogenic strains of IBDV isolated in Japan (Lin et al., 1993). Comparison of the amino acid sequences in this heptapeptide region reveals that non-pathogenic serotype 2 OH strain has three substitutions, whereas mildly pathogenic strains of serotype 1 (D78, Cu-1, PBG98 and 002-73) have one or two substitutions in this region. Since most of the amino acid residues causing antigenic variation reside in this region, it is possible that these residues may play an important role in the formation of virus-neutralizing epitopes, as well as serotype specificity.

To evaluate the antigenic relatedness of structural proteins of various IBDV strains, a phylogenetic tree was constructed, based on the large ORF sequences of ten IBDV strains, including the U.S. variant strains examined in this study (Fig. 2). Three distinguishable lineages were formed. The first one, which is most distant

GLS

DS326

E/Del

D78

cu-1

PBG98

52l70

STC

002-73

OH

Fig. 2. A phylogenetic tree for the IBDV structural proteins using the PAUP (phylogenetic analysis using parsimony) version 3.0 program (Illinois Natural History Survey, Champaign, IL).

from the others, is the serotype 2 OH strain, and the second one is the geographi- cally distant Australian serotype 1 strain (002-73). The third lineage consists of four distinct groups. The first and second group include highly pathogenic strains, namely, standard challenge (STC) strain from U.S. and the British field strain (52/70). The third group comprises all the European strains: the vaccine strains D78 (Holland), PBG98 (U.K.), and mildly pathogenic strain Cu-1 (Germany). The fourth group consists of the U.S. variant strains in which E/Del forms a different subgroup. The groups formed by the phyIogenetic analysis correlate very well with the MAbs reactivity patterns (see Table 11. As shown in Fig. 2, all the U.S. variant viruses which lack the 369 epitope form a distinct group, whereas all the classic viruses containing a B69 epitope form another group (except PBG98). In addition, closely related GLS and DS326 strains containing a common MAb 57 epitope and lacking an R63 epitope could be separated from the other variant E/Del strain.

In conclusion, we have identified the putative amino acid residues within the centraI variable region of the VP2 protein that can cause antigenic variation in IBDV. Since these residues may be involved in the formation of virus-neutralizing epitopes, this information would greatly aid in the development of recombinant IBDV vaccines for chickens.

We thank Dr. Jerome C. Regier for phylogenetic analysis and Gerard H. Edwards, Peter K. Savage and Stephanie A. Mengel-Whereat for technical assis- tance. This research was supported in part by grants from Intervet International B.V., U.S. Department of Agriculture (#90-34116-5399) and Maryland Agricul- tural Experiment Station. Scientific article no. A6496, contribution no. 8703, of the Maryland Agricultural Experiment Station.

References

Azad, A.A., Barrett, S.A. and Fahey, K.J. (1985) The characterization and molecular cloning of the double-stranded RNA genome of an Australian strain of infectious bursal disease virus. Virology 143, 35-44.

Bayliss, C.D., Spies, U., Shaw, K., Peters, R.W., Papageorgiou, A., Miiller, H. and Boursnell, M.E.G. (1990) A comparison of the sequences of segment A of four infectious bursal disease virus strains and identification of a variable region in VP2 J. Gen. Virol. 71, 1303-1312.

Becht, H., Miiller, H. and Miiller, H.K. (1988) ~mparative studies on structural and antigenic properties of two serotypes of infectious bursal disease virus. J. Gen. Virol. 69, 631-640.

Dobos, P., Hill, B.J., Hallet, R., Kells, D.T.C., Becht, H. and Teninges, D. (1979) Biophysical and biochemical characterization of five animal viruses with bisegmented double-stranded genomes. J. Virol. 32, 593-605.

Heine, H.G., Haritou, M., Failla, P., Fahey, K. and Azad, A. (1991) Sequence analysis and expression of the host-protective immunogen VP2 of a variant strain of infectious bursai disease virus which can circumvent vaccination with standard type I strains. J. Gen. Virol. 22, 1835-1843.

KN. Vakharia et al. /virus Research 31 (1994) 265-273 273

Hudson, P.J., McKern, N.M., Power, B.E. and Azad, A.A. (1986) Genomic structure of the large RNA segment of infectious bursal disease virus. Nucleic Acids Res. 14, 5001-5012.

Ismail, N.M., Saif, Y.M. and Moorhead, P.D. (1988) Lack of pathogenicity of five serotype 2 infectious bursal disease viruses. Avian Dis. 32, 757-759.

Jackwood, D.J. and Saif, Y.M. (1987) Antigenic diversity of infectious bursal disease viruses. Avian Dis. 31, 766-770.

Jagadish, M.N., Staton, V.J., Hudson, P.J. and Azad, A.A. (1988) Birnavirus precursor polyprotein is processed in Escherichia coli by its own virus-encoded polypeptide. J. Virol. 62, 1084-1087.

Kibenge, F.S.B., Dhillon, AS. and Russell, R.G. (1988) Biochemistry and immunology of infectious bursal disease virus. J. Gen. Virol. 69, 1757-1775.

Kibenge, F.S.B., Jackwood, D.J. and Mercado, C.C. (1990) Nucleotide sequence analysis of genome segment A of infectious bursal disease virus. J. Gen. Virol. 71, 569-577.

Kibenge, F.S.B., McKenna, P.K. and Dybing, J.K. (1991) Genome cloning and analysis of the large RNA segment (Segment A) of a naturally avirulent serotype 2 infectious bursal disease virus. Virology 184, 437-440.

Lin, Z., Kato, A., Otaki, Y., Nakamura, T., Sasmaz, E. and Ueda, S. (19931 Sequence comparisons of a highly virulent bursal disease virus prevalent in Japan. Avian Dis. 37, 315-323.

Lana, D.P., Beisel, C.E. and Silva, R.F. (1992) Genetic mechanisms of antigenic variation in infectious bursal disease virus: analysis of a naturally occurring variant virus. Virus Genes 6, 247-259.

McFerran, J.B., McNulty, M.S., McKilhop, E.R., Connor, T.J., McCracken, R.M., Collins, D.S. and Allan, G.N. (1980) Isolation and serological studies with infectious bursal disease viruses from fowl, turkeys and ducks: demonstration of a second serotype. Avian Pathol. 9, 395-404.

Snyder, D.B., Lana, D.P., Cho, B.R. and Marquardt, W.W. (1988al Group and strain-specific neutral- ization sites of IBDV defined with monoclonal antibodies. Avian Dis. 32, 527-534.

Snyder, D.B., Lana, D.P., Savage, P.K., Yancey, F.S., Mengel, S.A. and Marquardt, W.W. (1988b) Differentiation of infectious bursal disease viruses directly from infected tissues with neutralizing monoclonal antibodies: evidence of a major antigenic shift in recent field isolates. Avian Dis. 32, 535-539.

Snyder, D.B., Vakharia, V.N. and Savage, P.K. (1992) Naturally occurring-neutralizing monoclonal antibody escape variants define the epidemiology of infectious bursal disease virus in the United States. Arch. Virol. 127, 89-101.

Spies, U., Miiller, H. and Becht, H. (1987) Properties of RNA polymerase activity associated with infectious bursal disease virus and characterization of its reaction products. Virus Res. 8, 127-140.

Vakharia, V.N., Ahamed, B. and He, J. (1992) Use of polymerase chain reaction for efficient cloning of dsRNA segments of infectious bursal disease virus. Avian Dis. 36, 736-742.

Vakharia, V.N., Snyder, D.B., He, J., Edwards, G.H., Savage, P.K. and Mengel-Whereat, S.A. (1993) Infectious bursal disease virus structural proteins expressed in a baculovirus recombinant confer protection in chickens. J. Gen. Virol. 74, 1201-1206.

Van der Marel, P., Snyder, D. and Liitticken, D. (19901 Antigenic characterization of IBDV field isolates by their reactivity with a panel of monoclonal antibodies. Dtsch. Tierarztl. Wschr. 97, 81-83.

Winterfield, R.W. and Thacker, H.L. (1978) Immune response and pathogenicity of different strains of infectious bursal disease virus applied as vaccines. Avian Dis. 5, 253-260.