The antibody response of cattle infected with ... - Semantic Scholar

Journal of General Virology (1999), 80, 237–243. Printed in Great Britain ...................................................................................................................................................................................................................................................................................

The antibody response of cattle infected with bovine immunodeficiency virus to peptides of the viral transmembrane protein Linda Scobie,1 Chris Venables,2 Kenneth Hughes,2 Michael Dawson2 and Oswald Jarrett1 1 2

Retrovirus Research Laboratory, Department of Veterinary Pathology, University of Glasgow, Bearsden, Glasgow G61 1QH, UK Department of Virology, Central Veterinary Laboratory, New Haw, Addlestone, Surrey KT15 3NB, UK

The development of the antibody response to peptides of the transmembrane glycoprotein of bovine immunodeficiency virus (BIV) was followed over a period of 50 weeks in six cattle experimentally infected with the BIVFL112 isolate. Antibody was detected by an enzyme immunoassay using either a linear or a cyclized peptide with structural features common to an immunodominant region of other lentiviruses. The assay was specific for BIV, detecting antibody in bovine sera to BIVFL112 or BIVR29 but not to six other common viruses of cattle. Antibody was present in the sera of all cattle inoculated with BIVFL112 within 4 weeks of infection, peaked between 10 and 30 weeks and persisted in most cattle during the 50 weeks of observation. These features indicate that this assay may be useful in identifying cattle infected with other strains of BIV in the field.

Introduction Bovine immunodeficiency virus (BIV) was first isolated in the USA from a dairy cow with lymphoproliferative lesions (Van der Maaten et al., 1972). Originally termed bovine visnalike virus, the virus has subsequently been confirmed as a lentivirus and its evolutionary relationship to the other members of the genus has been defined (reviewed by Gonda et al., 1994). The impact of BIV on the health of cattle is still controversial. One of the difficulties in assessing the role of BIV in bovine disease is inconsistency in the methods used to detect infected cattle. Seroepidemiological surveys have demonstrated that BIV infection occurs in many parts of the world, including the USA (Amborski et al., 1989 ; Cockerell et al., 1992 ; St Cyr Coats et al., 1994), Canada (McNab et al., 1994), the Netherlands (Horzinek et al., 1991), New Zealand (Horner, 1991) and France (Polack et al., 1996). However, the performance of some of the assays used has been unreliable and they may have given false-positive results. In addition, in certain assays the antibody response to some BIV antigens appears to decline to undetectable levels within a few months of infection (Isaacson et al., 1995), so that false-negative results may be obtained. Therefore, there is an urgent requirement for Author for correspondence : Linda Scobie. Fax j44 141 330 5602. e-mail l.scobie!vet.gla.ac.uk

0001-5755 # 1999 SGM

a reliable assay to detect antibodies to BIV that can be used to study the immune response of cattle to infection and to investigate the occurrence of the infection in cattle populations where the prevalence has been reported to be low (Horzinek et al., 1991). For this purpose, an assay that detects antibodies to all BIV strains would be essential. Enzyme immunoassays (EIA) for antibodies to the transmembrane (TM) peptides of several lentiviruses have proved useful for the detection of individuals infected with, for example, human immunodeficiency virus (HIV) (Gnann et al., 1987), feline immunodeficiency virus (FIV) (Fontenot et al., 1992 ; Sibille et al., 1995), equine infectious anaemia virus (EIAV) (Chong et al., 1991) and caprine arthritis–encephalitis virus (CAEV) (Bertoni et al., 1994). Accordingly, we chose to explore a similar strategy for the detection of BIV-infected cattle. The amino acid sequence for the derivation of the peptides was based on the sequence of BIVR . Three isolates are currently available for study : the #* original BIVR (Van der Maaten et al., 1972), BIVFL and #* ""# (Suarez et al., 1993). Of these, the full-length BIVFL %*" nucleotide sequence is available only for BIVR , although the #* surface envelope glycoprotein has been analysed for all three and found to have 50 % amino acid sequence identity (Suarez & Whetstone, 1995). The TM glycoprotein of the envelope of BIV appeared to be a suitable target for antiviral antibodies, since the TM proteins of all lentiviruses have common

CDH

L. Scobie and others

structural features and contain immunodominant B-cell epitopes (Gnann et al., 1987 ; Chong et al., 1991 ; Bertoni et al., 1994 ; Pancino et al., 1995). Within the extracellular ectodomain of TM are two conserved cysteine residues, which are separated by four to eight amino acids (Gallacher et al., 1989). The amino acid sequence between and flanking the conserved cysteine residues varies between different lentiviruses but is conserved between isolates of the same virus. The immunodominant domain forms a characteristic loop structure that is critical for peptide antigenicity and is stabilized by disulphide bonding between the cysteine residues (Fontenot et al., 1992). In BIV, the N-terminal portion of TM contains an immunodominant epitope that includes the region containing the conserved cysteine residues (Chen et al., 1994), but the antigenic properties of this region have not been fully characterized. In the present study, we established an immunoassay (patent applied for) to detect antibodies to the putative immunodominant determinant of the BIV TM peptide and defined the relevant epitope in this region by analysis of flanking peptides or peptides which did not contain a cysteine. The immunoassay was then used over a period of 50 weeks to determine the development and persistence of antibodies in cattle infected with BIVFL . ""#

Methods

Bovine sera. Sera from cattle CC10 and CC1360, which were experimentally infected with BIVR , were used as positive controls and #* serum from animal CC175, which was seronegative by indirect immunofluorescence assay (IFA) for BIVR , was used as the negative #* control. Sera were obtained over a period of 50 weeks from six cattle (CJ25, CJ29, CJ96, CJ28, CJ53 and CJ114) following subcutaneous inoculation with BIVFL and from three cattle (CJ118, CJ120 and CJ123) ""# which had received a pasteurized BIVFL inoculum (Venables et al., ""# 1997). All animals in the herd from which these cattle were derived were known to be seronegative for BIV on the basis of the IFA and free of proviral DNA as determined by PCR. However, some of the cattle were found to have antibodies to bovine syncytial virus (BSV) by IFA (Venables et al., 1997). To confirm the specificity of the antibody response to BIV, bovine sera were also tested that had been shown by IFA to contain antibodies to BSV, bovine leukaemia virus (BLV), bovine herpesvirus-1 (BHV-1), bovine viral diarrhoea virus (BVDV), respiratory syncytial virus (RSV) or parainfluenza virus type 3 (PIV3).

Synthetic peptides. Peptides were synthesized either in our own laboratory or by Alta Bioscience (University of Birmingham, UK) by using the available sequence of the TM glycoprotein of BIVR . The #*

amino acid sequence of the peptides is shown in Fig. 1. The nucleotide positions corresponding to the various peptides were : TMA1 and TMA2, 7326–7370 ; TMA3, 7251–7304 ; TMA4, 7311–7361 ; TMA5, 7350–7403 ; and TMA6, 7398–7451. Sequences were taken from GenBank accession no. M32690. TMA4 and TMA5 were designed to overlap TMA1\A2 and to include one of the conserved cysteines. TMA3 and TMA6 overlapped, and extended into, the region flanking the putative immunodominant domain (Fig. 1). The peptides were synthesized in the linear form and were cleaved from the resin by 95 % trifluoroacetic acid (TFA), 2n5 % water and 2n5 % ethane di-thiol, dried and washed in ether and then freeze-dried. Peptides were purified by HPLC on a Vydac 300A peptide column and then lyophilized. The linear peptide TMA2 was dissolved in 0n01 M ammonium acetate at a peptide concentration of 1 mg\ml and stirred for 3 h at room temperature to form the cyclic peptide (TMA1), before lyophilization.

EIA for antibodies to BIV TM peptides. A volume of 100 µl peptide (10 µg\ml in 0n05 M carbonate–bicarbonate buffer, pH 9n6) was adsorbed to each well of a microtitre plate (Immulon 2, Dynatech) overnight at 4 mC. The wells were washed three times with 200 µl wash buffer (WB ; 0n1 M Tris-buffered saline, pH 7n4, 0n1 % Tween-20) and then blocked with 4 % BSA in WB for 1 h at room temperature. Following three washes with WB, 100 µl aliquots of bovine sera diluted 1 : 50 in WB were incubated in the wells for 1 h at room temperature. After washing three times with WB, horseradish peroxidase (HRP)-conjugated sheep anti-bovine IgG (1 : 3000 in WB) (Sigma) was added for 1 h at room temperature. Following six washes with WB, the peroxidase reaction was visualized with the chromogen tetramethyl benzidine and the results were expressed as the absorbance at 650 nm. A positive result was indicated by an absorbance three times that of the appropriate negative control.

PCR amplification and sequencing of BIVFL112 TM protein. Genomic DNA from a cow infected with BIVFL was used for PCR ""# amplification of the template for sequencing. DNA was extracted from whole blood with a QIAamp blood kit (Qiagen). PCR amplification of the BIVFL env gene was carried out in 0n5 ml thin-walled microtubes ""# containing 250 ng genomic DNA, 2n5 mM MgCl , 15 pmol primer and # a Ready-To-Go PCR bead (Pharmacia Biotech) in a final volume of 25 µl. The positive-strand amplification primer was KEj1 and the negativestrand primer was KE-1 (Table 1). DNA was amplified for 38 cycles as follows : 95 mC for 1 min, 61n5 mC for 1 min and 72 mC for 2 min, with a final step of 72 mC for 10 min to allow extension of product. The product was then visualized on an ethidium bromide-stained agarose gel. A band of the correct molecular mass was excised and the PCR product was purified with GFX columns (Pharmacia Biotech). To the contents of an ABI PRISM dye terminator FS kit (Applied Biosystems), 60 ng PCR template and 3n2 pmol primer (Table 1) were added in a final volume of 20 µl. The reaction was amplified by 25 cycles of 96 mC for 30 s, 50 mC for 15 s and 60 mC for 4 min. The sequencing product was precipitated by the addition of 2 µl 3 M sodium acetate,

Fig. 1. Amino acid sequence and alignment of the BIV TM peptides TMA1–TMA6. The arrows indicate the ranges covered by peptides TMA4 and TMA5.

CDI

Antibody response to BIV

Table 1. Primers used for amplification of the template DNA and for sequencing of the TM region of the BIVFL112 env gene Primers were supplied by Gibco and HPLC-purified. Names including j indicate positive-strand primers and those including k indicate negative-strand primers. Primer name KEj1 KEj3 KEj4 KEj5 KEj6 KEk1 KEk3 KEk5 KEk6

Sequence TCC CCA GCT ATG GGT TTG T CAA CCG AAA GCC TCG AGC A TGT GGT GGT GAA TGA CAC A CAA TTC CTC CAT CTG TTT A TGC TCA AAC CCA GGA ACA AT GGG CGG AGT AAG CAA GCG TAA GC GCG TTT CAG GGC GGA GTA AGC GCG TGT TAA GGT AGG CGT TTC AGG CTC CCA CAG TCC ACC CTT TTA A

pH 4n6 and 50 µl 95 % ethanol and sequenced by using an ABI373 automated sequencer (Applied Biosystems). Sequence data were assembled using Seqman II (DNASTAR).

Results Specificity of the antibody response of BIV-infected cattle to TM peptides

The ability of the BIV TM peptides, TMA1 and TMA2, to bind sera from cattle known to be infected with BIV was tested in an EIA. Sera from two cattle, CC10 and CC1360, which had been inoculated with BIVR , reacted with both the cyclic #* TMA1 (Fig. 2) and the linear TMA2 peptide (data not shown), while sera from the control cattle CJ118, CJ120 and CJ123, which had received a pasteurized BIVFL inoculum, had no ""# detectable antibody response to BIV peptides TMA1 (Fig. 2) or TMA2 (data not shown), even 50 weeks after infection. To confirm that the TM peptides specifically detected antibodies to BIV, sera from cattle seropositive by IFA to six common bovine viruses were tested for reactivity with both the linear and cyclic TM peptides. None of these sera reacted with the peptides (Fig. 3). The possible cross-reactivity of the BIV TM peptides with antibodies to other lentiviruses was examined. Sera from animals naturally infected with EIAV, visna–maedi virus or CAEV gave no detectable response to BIV TMA1 or TMA2 peptides (data not shown). The sera from all of the cattle were tested for reexperimentally infected with BIVFL ""# activity to the FIV TM2 peptide (Avrameas et al., 1992 ; Sibille et al., 1995) and no response was detected (L. Scobie & A. Pacitti, unpublished data). Definition of the immunodominant region of BIV TM

To ensure that the peptide selected for use in the EIA contained the immunodominant epitope, adjacent peptides

within the putative principal immunodominant domain were tested for reactivity with sera from BIV-infected cattle. The relative positions of these peptides, TMA3, TMA4, TMA5 and TMA6, are shown in Fig. 1. TMA4 and TMA5 flanked TMA1\TMA2 and overlapped the cysteine residues, while TMA3 and TMA6 were outside the TMA1\TMA2 region. While the sera from the two BIVR -infected cattle and all the #* cattle infected with BIVFL reacted with the TMA1 and ""# TMA2 peptides, none of the sera showed any significant reactivity to peptides TMA3–TMA6. A representative example is shown in Fig. 4. Sera from the control cattle CJ118, CJ120 and CJ123 did not react with any of the peptides (data not shown). Antibodies to the BIV TM immunodominant determinant develop rapidly and persist during experimental infection of cattle

The EIA was then used to follow the development of the antibody response of cattle experimentally infected with BIV. Serum samples were obtained at intervals over a period of 50 weeks from six cattle inoculated subcutaneously with BIVFL ""# (Venables et al., 1997) and samples were analysed for antibody levels to the TM peptides. Antibodies were detected by EIA to both the linear TMA2 (Fig. 5 a) and cyclic TMA1 peptides (Fig. 5 b). For both peptides, antibodies were detected in all animals at 4 weeks after infection. Antibody levels to the cyclic TMA1 peptide appeared to be higher, and were maintained for longer, than those to the linear peptide. The antibody responses to the TMA2 peptide peaked during the period of 10–30 weeks after infection (Fig. 5) but were still substantial at the final sampling, 50 weeks after infection. The timing of the response to the peptides corresponded with the presence of proviral DNA in peripheral blood and seroconversion to BIV, as monitored by IFA (Venables et al., 1997). One animal, CJ25, mounted an immune response to TMA1 and TMA2 by 2 weeks after infection (Fig. 5), which declined during the experiment. This 19-month-old calf was seropositive by IFA and PCR-positive by 14 days after inoculation and BIV was recovered from peripheral blood mononuclear cells (Venables et al., 1997). The animal exhibited signs of ataxia during the experiment and, on necropsy, was found to have an extensive lymphadenopathy and encephalitis (Munro et al., 1998). However, it has not yet been established whether animal CJ25 was immunodeficient. The TM immunodominant domain is conserved between BIVR29 and BIVFL112

The level of the antibody response of cattle infected with BIVR to TMA1 and TMA2 peptides was comparable to #* (Fig. 5), suggesting that the those infected with BIVFL ""# peptides are efficiently recognized by antibodies against both strains of BIV. The basis of this cross-reactivity was examined transmembrane protein. by sequencing of the BIVFL ""# Comparison of the amino acid sequence of isolates BIVR and #* CDJ

L. Scobie and others 1·4

3·0

1·2

2·5

1·0 2·0 A650

A650

0·8

1·5

0·6 1·0

0·4

Time after infection (weeks)

Sera to selected bovine viruses

Controls

Fig. 2

CC1360

CC10

CC175

RSV

PIV3

BVDV

BHV-1

BSV –

50

BSV+

30

CC175

18

CC1360

6

CC10

0

BLV

0·5

0·2

Controls

Fig. 3

Fig. 2. Antibody response to the TMA1 cyclic peptide of three cattle that received BIVFL112 pasteurized inoculum. Sera of animals CJ118 (horizontally hatched bars), CJ120 (unfilled bars) and CJ123 (diagonally hatched bars) were analysed from 0 to 50 weeks after infection. The solid bars indicate sera from two cattle infected with BIVR29, CC1360 and CC10, used as positive controls and the negative control, CC175, from a cow seronegative for BIVR29 by IFA. Fig. 3. Reactivity of sera from selected bovine viruses to the TMA1 cyclic peptide. Sera tested were from cattle seropositive for BLV, BSV, BHV-1, BVDV, RSV and PIV3 and from an animal seronegative for BSV. CC10 and CC1360 are the positive control sera and CC175 is the negative control serum.

2·0

Discussion

1·8 1·6 1·4 A650

1·2 1·0 0·8 0·6 0·4 0·2 0

2

4

6

10

14

18

24

30

40

50

Time after infection (weeks) Fig. 4. Antibody response of sera from animal CJ53 to the overlapping peptides TMA1 (4), TMA2 ( ), TMA3 (>), TMA4 (i), TMA5 (M) and TMA6 ($). Sera were analysed from 0 to 50 weeks after infection.

showed 76 % similarity across the entire transBIVFL ""# membrane protein sequence. Comparison of the region of the TMA1\TMA2 peptide sequence showed 87 % identity between the two isolates and a conserved substitution of an arginine for a lysine and alteration of a valine to an alanine in the BIVFL isolate (Fig. 6). These changes did not appear to ""# affect the reactivity of the sera of cattle infected with either BIVR or BIVFL to the cyclic peptide TMA1 or the linear #* ""# peptide TMA2 in the EIA (Fig. 5). CEA

We have shown that a peptide based on the amino acid sequence of the TM glycoprotein of BIVR binds antibodies #* from cattle experimentally infected with either BIVR or #* BIVFL . The level of antibody response of cattle infected ""# with either isolate was similar. The specificity of the immunoassay appeared to be high, since the sera of cattle positive for several common bovine viruses did not react with the BIV peptide and there was no cross-reactivity with antibodies to the lentiviruses of four other species of domestic animals. The sensitivity of this test was not established, owing to the lack of a ‘ gold standard ’ assay of antiviral antibodies. The reactivity in the assay was restricted to an immunodominant region of TM, defined by peptides TMA1 and TMA2. Although the antibody response to the TMA1 cyclic peptide tended to be higher than that to the linear peptide TMA2, analysis of variance revealed that the differences between the absorbances for the linear and cyclic TM peptides were not significant (data not shown). None of the sera from BIV-infected cattle had detectable reactivity to any other peptide tested, suggesting that antibody binding to the BIV TMA1 and TMA2 peptides was dependent on the presence of both conserved cysteine residues. Conservation of this immunological determinant suggests that it has functional importance and indeed, it has been demonstrated for FIV that disruption of the determinant by mutagenesis of the conserved cysteines results in non-functional viral envelopes (Pancino et

Antibody response to BIV (a) 2·0 1·8 1·6 1·4

A650

1·2 1·0 0·8 0·6 0·4

4

6

10

14

18

24

30

40

50


CC175

2

CC1360

0

CC10

0·2

Controls

(b) 2·0 1·8 1·6 1·4

A650

1·2 1·0 0·8 0·6 0·4

4

6

10

14

18

24


30

40

50

CC175

2

CC1360

0

CC10

0·2

Controls

Fig. 5. Immune responses to the TMA2 linear peptide (a) and the TMA1 cyclic peptide (b) of six cattle infected with BIVFL112 from 0 to 50 weeks post-infection. Each bar represents an individual animal at each time-point indicated : CJ25 (filled bars), CJ29 (horizontally hatched bars), CJ96 (diagonally hatched bars), CJ28 (unfilled bars), CJ53 (dark grey bars) and CJ114 (cross-hatched bars). Pre-inoculation serum is indicated by 0 weeks. Positive control sera CC10 and CC1360 and the negative control serum CC175 are indicated by the dotted bars.

CEB

L. Scobie and others

Fig. 6. Comparison of TM glycoprotein sequences for isolates BIVR29 (GenBank accession no. M32690) and BIVFL112 (GenBank accession no. AJ011527). The BIVR29 TM protein sequence is translated from nucleotide 5367 to 8482. The sequence of the region corresponding to peptides TMA1/TMA2 is boxed.

al., 1995). The presence of both conserved cysteines may create a conformational constraint necessary for antibody recognition of the TM epitope. The value of an assay for anti-BIV antibodies in detecting infected cattle depends not only on the sensitivity and specificity of a diagnostic test but also the capacity of the test to detect antibodies in infected cattle throughout their lives. By studying the antibody response to BIV TM peptides in experimentally infected cattle over a period of 50 weeks, we established that antibody levels rise rapidly and persist at a high level over this time. Peptides derived from the immunodominant determinant of the TM glycoproteins of lentiviruses have been shown to be immunologically active for both FIV and HIV (Avrameas et al., 1992 ; Pancino et al., 1995 ; Gnann et al., 1987). In a similar study, Fontenot et al. (1992) have described seroconversion to an FIV TM determinant during the experimental infection of cats with FIV-Petuluma within 2 weeks after infection. Our finding, that antibodies to BIV TM peptides persist at a high level for 50 weeks in the sera of cattle experimentally infected with BIVFL , has consequences for ""# the identification of cattle infected with BIV, since previous studies indicate that antibodies to some of the structural proteins of the virus may decline to undetectable levels within a few months of infection (Isaacson et al., 1995). Antibodies to p26, the major BIV capsid protein, were found to persist for more than 2 years after inoculation of experimentally infected cattle (Whetstone et al., 1990, 1991). While anti-p26 antibodies were detected 2 weeks after infection and peaked at 6–8 weeks, two out of four of the cattle tested had no detectable antibody by 19–38 months post-infection, as determined by Western blot and IFA (Whetstone et al., 1990). In other studies, Isaacson CEC

et al. (1995) examined the response to the BIV Gag and Env proteins by using recombinant p26 and a 134 amino acid Nterminal portion of TM (gp45) (Chen et al., 1994). A strong antibody response to both Gag and Env was detected 4 weeks after infection of cattle with BIVR . While all cattle continued #* to react to Env for over 4 years after infection with BIV, the response to the Gag protein was lost after about 6 months (Isaacson et al., 1995). This suggests a greater persistence of reactivity to BIV Env protein in cattle. Most serological studies of BIV prevalence in naturally infected cattle have relied on assays based on the detection of antibodies to the Gag protein. This may have led to an underestimate of BIV infection rates and a misunderstanding of the role of BIV in bovine disease. The features of the antibody response to the TM peptides that we have described indicate that this assay would be useful in the detection of BIV-infected cattle in the field. The EIA is highly specific and cattle infected with both currently available isolates, BIVR and BIVFL , demonstrated a similar level of #* ""# response to the TM peptides. The antibody response to the TM is rapid and appears to persist throughout infection. Sequencing showed amino acid conservation between the two isolates at the site of the TMA1\TMA2 peptide. This confirms the value of this immunodominant epitope as a reagent for antibody detection. Sequencing of further isolates is required to examine this region in greater detail. We are currently validating the sensitivity of the TM EIA in comparison to previous methods of BIV detection, such as Western blot and PCR, and are examining the utility of this assay in assessing the seroprevalence of BIV infection in the field. The authors thank Drs C. Whetstone and D. Suarez for the provision of the FL112 isolate of BIV. This work was funded by the Ministry of Agriculture, Fisheries and Food.

References Amborski, G. F., Lo, J. L. & Seger, C. L. (1989). Serological detection of

multiple retroviral infections in cattle : bovine leukemia virus, bovine syncytial virus and bovine visna virus. Veterinary Microbiology 20, 247–253. Avrameas, A., Guillet, J. G., Chouchane, L., Mouraillon, A., Sonigo, P. & Strosberg, A. D. (1992). Localisation of three epitopes of the env

protein of feline immunodeficiency virus. Molecular Immunology 29, 565–572. Bertoni, G., Zhano, M.-L., Zanoni, R., Vogt, H.-R., Peterhans, E., Ruff, G., Cheevers, W. P., Sonigo, P. & Pancino, G. (1994). Antibody

reactivity to the immunodominant epitopes of the caprine arthritis– encephalitis virus gp38 transmembrane protein associates with the development of arthritis. Journal of Virology 68, 7139–7147. Chen, P., Liu, Z. Q. & Wood, C. (1994). Use of TrpE fusion protein to identify antigenic domains within the BIV envelope protein. Journal of Virological Methods 47, 331–343. Chong, Y.-H., Ball, J. M., Issel, C. J., Montelaro, R. C. & Rushlow, K. E. (1991). Analysis of equine humoral immune responses to the trans-

membrane envelope glycoprotein (gp45) of equine infectious anemia virus. Journal of Virology 65, 1013–1018.

Antibody response to BIV Cockerell, G. L., Jensen, W. A., Rovnak, J., Ennis, W. H. & Gonda, M. A. (1992). Seroprevalence of bovine immunodeficiency-like virus and

bovine leukemia virus in a dairy cattle herd. Veterinary Microbiology 31, 109–116. Fontenot, J. D., Hoover, E. A., Elder, J. H. & Montelaro, R. C. (1992).

Pancino, G., Camoin, L. & Sonigo, P. (1995). Structural analysis of the

principal immunodominant domain of the feline immunodeficiency virus transmembrane glycoprotein. Journal of Virology 69, 2110–2118. Polack, B., Schwartz, I., Berthelemy, M., Belloc, C., Manet, G., Vuillaume, A., Baron, T., Gonda, M. A. & Levy, D. (1996). Serologic

evidence for bovine immunodeficiency virus infection in France. Veterinary Microbiology 48, 165–173.

Evaluation of feline immunodeficiency virus and feline leukemia virus transmembrane peptides for serological diagnosis. Journal of Clinical Microbiology 30, 1885–1890.

St Cyr Coats, K., Pruett, S. B., Nash, J. W. & Cooper, C. R. (1994).

Gallacher, W. R., Ball, J. M., Garry, R. F., Griffin, M. C. & Montelaro, R. C. (1989). A general model for the transmembrane proteins of HIV

Bovine immunodeficiency virus : incidence of infection in Mississippi dairy cattle. Veterinary Microbiology 42, 181–189.

and other retroviruses. AIDS Research and Human Retroviruses 5, 431–440. Gnann, J. W., Nelson, J. A. & Oldstone, M. A. B. (1987). Fine mapping of an immunodominant domain in the transmembrane glycoprotein of human immunodeficiency virus. Journal of Virology 61, 2639–2641. Gonda, M. A., Luther, D. G., Fong, S. E. & Tobin, G. J. (1994). Bovine immunodeficiency virus : molecular biology and virus–host interactions. Virus Research 32, 155–181. Horner, G. W. (1991). Serological evidence of bovine immunodeficiency-like virus and bovine syncytial virus in New Zealand. Surveillance 18, 9. Horzinek, M., Keldermans, L., Stuurman, T., Black, J., Herrewegh, A., Sillekens, P. & Koolen, M. (1991). Bovine immunodeficiency virus :

immunochemical characterization and serological survey. Journal of General Virology 72, 2923–2928. Isaacson, J. A., Roth, J. A., Wood, C. & Carpenter, S. (1995). Loss of Gag-specific antibody reactivity in cattle experimentally infected with bovine immunodeficiency-like virus. Viral Immunology 8, 27–36. McNab, W. B., Jacobs, R. M. & Smith, H. E. (1994). A serological survey for bovine immunodeficiency-like virus in Ontario cattle and associations between test results, production records and management practices. Canadian Journal of Veterinary Research 58, 151–181.

Sibille, P., Avrameas, A., Moraillon, A., Richardson, J., Sonigo, P., Pancino, G. & Strosberg, A. D. (1995). Comparison of serological tests

for the diagnosis of feline immunodeficiency virus infection of cats. Veterinary Microbiology 45, 259–267. Suarez, D. L. & Whetstone, C. A. (1995). Identification of hypervariable and conserved regions in the surface envelope gene in the bovine lentivirus. Virology 212, 728–733. Suarez, D. L., VanDerMaaten, M. J., Wood, C. & Whetstone, C. A. (1993). Isolation and characterization of new wild-type isolates of

bovine lentivirus. Journal of Virology 67, 5051–5055. Van der Maaten, M. J., Boothe, A. D. & Seger, C. L (1972). Isolation of

a virus from cattle with persistent lymphocytosis. Journal of the National Cancer Institute 49, 1649–1657. Venables, C., Lysons, R., Horigan, M., Stagg, D. & Dawson, M. (1997).

Munro, R., Lysons, R., Venables, C., Horigan, M., Jeffrey, M. & Dawson, M. (1998). Lymphadenopathy and non-suppurative meningo-encepha-

Bovine immunodeficiency-like virus : inactivation in milk by pasteurisation. Veterinary Record 140, 275–277. Whetstone, C. A., VanDerMaaten, M. J. & Black, J. W. (1990). Humoral immune response to the bovine immunodeficiency-like virus in experimentally and naturally infected cattle. Journal of Virology 64, 3557–3561. Whetstone, C. A., VanDerMaaten, M. J. & Miller, J. M. (1991). A western blot assay for the detection of antibodies to bovine immunodeficiency-like virus in experimentally inoculated cattle, sheep, and goats. Archives of Virology 116, 119–131.

litis in calves experimentally infected with bovine immunodeficiency-like virus. Journal of Comparative Pathology 119, 121–134.

Received 16 June 1998 ; Accepted 30 September 1998

CED