Veterinary Microbiology 63 (1998) 85±97

Detection of antibodies to bovine viral diarrhoea

virus (BVDV) and characterization of
genomes of BVDV from Brazil
ClaÂudio Wageck Canalb, Marc Strassera, Christian Hertiga,
Aoi Masudab,c, Ernst Peterhansa,*
a
Institute of Veterinary Virology, University of Berne, Laenggass-Strasse 122, CH-3012, Bern, Switzerland
b
Centro de Biotecnologia, Universidade Federal do Rio Grande do Sul, Av. Bento Gonc,alves 9500,
C.P. 15005, CEP 91501-970, Porto Alegre, RS, Brazil
c
Departamento de Biotecnologia, Universidade Federal do Rio Grande do Sul, Av. Bento Gonc,alves 9500,
C.P. 15005, CEP 91501-970, Porto Alegre, RS, Brazil

Received 20 October 1997; accepted 16 June 1998

Abstract

An ELISA for the detection of antibodies to bovine viral diarrhea virus (BVDV) was developed
based on antigens derived from a genotype I BVDV strain isolated in Switzerland. Using
monoclonal antibodies we showed that this antigen contained the conserved non-structural protein
NS3 whereas it essentially lacked the more strain-specific E2 surface glycoprotein. This ELISA has
a sensitivity of 97.5% and a specificity of 99.2% as compared to the serum neutralization test
(SNT). Preliminary experiments showed that this ELISA reliably detects antibodies to BVDV
strains circulating in Brazil. Serum samples obtained from 430 adult cattle on 19 farms of the State
of Rio Grande do Sul (Brazil) and one farm from Corrientes (Argentina) were tested for antibodies
by means of this ELISA. We found antibodies in 56%15.1% of the cattle sera tested, which
indicates that, in Brazil, the prevalence of infection with BVDV is similar to that found in Europe
and the USA. Our sequence analysis of two BVDV isolates showed that BVDV of both genotypes I
and II circulate in Brazil. # 1998 Elsevier Science B.V. All rights reserved.

Keywords: Bovine viral diarrhea virus; Pestivirus; ELISA; Reverse transcription; Polymerase chain reaction;
Epidemiology; Phylogenetic analysis

* Corresponding author. Tel.: +41-31-631-24-13; fax: +41-31-631-25-34; e-mail: peterhans@ivv.unibe.ch

0378-1135/98/$ ± see front matter # 1998 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 1 1 3 5 ( 9 8 ) 0 0 2 3 2 - 6
86 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97

1. Introduction

The genus Pestivirus of the Flaviviridae comprises the three single-stranded, positive-
polarity RNA viruses bovine viral diarrhea virus (BVDV), classical swine fever virus
(CSFV), and border disease virus (BDV) (Horzinek, 1991). Pestiviruses cause
economically significant diseases in ruminants and pigs. BVDV and BDV are endemic
world-wide, whereas CSFV has actually been eradicated in some countries (Dahle and
Liess, 1992).
Exposure of cattle to BVDV may result in either acute transient or persistent infection.
The latter is initiated in the early stages of pregnancy by infection of the developing fetus
and results in immunotolerance specific to the infecting viral strain. Animals infected in
utero may be born normal and play an important role in the epidemiology of BVDV
because they shed large amounts of virus, thereby serving as a source of infection for
susceptible animals. In these, infection is transient, being terminated by a vigorous
immune response. The main clinical manifestations of transient BVDV infection are mild
diarrhea, fever and respiratory symptoms (Radostits and Littlejohns, 1988), although
recently fatal thrombocytopenia has also been reported (Hibberd et al., 1993; Ridpath
et al., 1994; Pellerin et al., 1994).
The antibody titer induced by transient infection decreases very slowly (Brownlie
et al., 1987) and correlates with a lifelong immunity to BVDV. Therefore, the prevalence
of antibody-positive animals reflects the proportion of animals previously exposed to
BVDV (Houe, 1995). Antibodies to BVDV may be detected by several methods,
including the serum neutralization test (SNT) and ELISA. Since BVDV strains are highly
heterogeneous, particularly in the surface glycoproteins targeted by neutralizing
antibodies, antibodies to certain strains may actually escape detection in neutralization
assays. Therefore, assays based on antibodies to more conserved viral antigens, in
particular the non-structural protein NS2/3, are more suitable (Kwang et al., 1991;
Lecomte et al., 1991; Paton et al., 1991; Petric et al., 1992). In this work, we have
developed a simple ELISA that detects antibody to NS3 but not to the more strain-
specific surface glycoprotein E2. Antigen prepared from a German BVDV isolate was
shown to reliably detect antibodies to BVDV strains circulating in Brazil. Genomic
analysis revealed the presence of BVDV strains of genotypes I and II in the Brazilian
cattle population.

2. Materials and methods

2.1. Cells and viruses

The cytopathic A1138/69 BVDV strain was propagated in embryonic bovine turbinate
cells, cultured in Earle's minimal essential medium (Gibco BRL, Grand Island, NY) with
1% non-essential amino acids, antibiotics (Penicillin, 100 I.U./ Streptomycin, 100 mg/ml)
(Seromed, Biochrom KG, Berlin, Germany), 1.25% L-Glutamine (Gibco BRL) and 2%
fetal calf serum free of anti-BVDV antibodies and BVDV (Seromed). Uninfected cell
cultures were used as negative controls.
 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97 87

The non-cytopathic Soldan BVDV strain was isolated from an animal with mucosal
disease-like symptoms that had died in Eldorado do Sul (State of Rio Grande do Sul) in
1992. The non-cytopathic BR275 strain was isolated from the serum of an antibody
negative animal originating from the CapaÄo do LeaÄo province. The non-cytopathic strains
were demonstrated by immunoperoxidase staining of infected bovine embryonic
turbinate cells (Meyling, 1983).

2.2. Preparation of antigen for ELISA

Cells infected with the cytopathic A1138/69 BVDV strain showing an approximately
80% cytopathic effect were harvested after three cycles of rapid freezing and thawing
followed by centrifugation at 48 000g for 4 h. Cells infected with the non-cytopathic
BR275 and Soldan BVDV strains were harvested at the same time post-infection and
processed by the same method. Pellets containing virus and cell debris were resuspended
in phosphate buffered saline (PBS, pH 7.2) with 2% Tween 20 and incubated by gentle
shaking on a minishaker at 378C for 2 h and then sonicated for 30 s (output 6.5, pulse
mode 60%) (Branson sonifier, Branson, USA). The suspension was clarified by
centrifugation at 1000g for 30 min and the supernatant was used as the antigen for
coating ELISA plates. Antigen containing viral glycoprotein was prepared as previously
described for rabies G protein using solubilization buffer containing 300 mM NaCl,
68 mM n-octyl-b-D-glucopyranoside (OGP), 50 mM Tris±HCl, pH 7.4 (Grassi et al.,
1989). Antigens were stored in aliquots at ÿ808C.

2.3. Monoclonal antibodies (McAb)

McAB C16, directed against NS3, and antibodies CT2, CT3, CT6 and CT9, directed
against E2 were kindly supplied by Dr. V. Moennig, Hannover, Germany. McAb C16 was
raised using the NADL BVDV strain as the antigen and McAbs CT2, CT3, CT6, CT9
were obtained using the A1138/69 BVDV strain as the antigen for immunization (Peters
et al., 1986).

2.4. Elisa protocols

Virus and negative control antigens were diluted to 4±8 mg/ml total protein in 0.1 M
Na2CO3±NaHCO3 buffer, pH 9.6 and used for coating of microtitration plates (100 ml/
well) (Maxisorp, A/S Nunc, Kamstrub, Denmark). After an incubation period of 14 h at
48C in a humid chamber, plates were washed three times with washing buffer (0.1 M
NaCl; 0.02 M Tris; 0.05% Tween 20; pH 7.4). Monoclonal antibodies were diluted in the
corresponding washing buffer and incubated at room temperature for 90 min. Bound
antibody was detected using goat anti-mouse IgG (heavy chain specific) peroxidase
labeled conjugate (Kirkegaard and Perry Labs., Gaithersburg, USA). Serum diluted 1:10
in washing buffer containing 1% skimmed milk powder was added to the plates. After
incubation at room temperature in a humid chamber for 90 min the plates were washed as
described above. Goat anti-bovine IgG (heavy chain specific) peroxidase labelled
(Kirkegaard and Perry Labs.) was used to detect bound bovine antibodies. Plates were
88 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97

incubated in a humid chamber for 90 min. After washing three times in washing buffer
the chromogen solution 2,20 -azinodi (ethylbenzothiazolinesulfonic acid) (ABTS)
(Boehringer Mannheim, Germany) was added and plates were incubated for 20 min. at
room temperature. Optical density was measured by an ELISA reader (Titertek multiscan
MC) at a wavelength of 405 nm. Sera reaching more than 50% of a positive control serum
were considered positive. The negative cutoff limit was set at less than 20% of a positive
control serum.

2.5. Serum neutralization test (SNT)

The test was carried out according to a standard microtiter procedure (Steck
et al., 1980) using embryonic bovine turbinate cells, doubling dilutions of heat
inactivated sera and 100 TCID50 of the A1138/69 virus. The serum neutralizing titer
is expressed as the reciprocal of the serum dilution to inhibit 50% of the viral cytopathic
effect (cpe).

2.6. Cattle sera

Serum samples obtained from 940 cattle from Switzerland were used to establish the
correlation of antibody levels between SNT and ELISA.
The 430 bovine serum samples used in the prevalence study for Brazil were collected
from adult animals in abattoirs in Rio Grande do Sul State, except for 10 sera which were
collected in Corrientes (Argentina). The 12 samples from the Aratiba region were
obtained from breeding stock.

2.7. Statistical analyses of ELISA results

The confidence interval for the Brazilian prevalence of anti-BVDV antibodies was
corrected for clustering and determined according to the method of Donald and Donner
(1987). The software package Statistica (StatSoft, Oklahoma, USA) was used for
calculation.

2.8. RNA extraction and RT±PCR

Trizol LS Reagent (Gibco BRL) was used to isolate RNA from 250 ml of cell culture
supernatants or 50 ml of serum. The cDNA was synthesized according to standard
procedures (Sambrook et al., 1989) using antisense primer P43 (50 -AGACGAATTC-
TCAACTCCATGTGCCATGTAC-30 ), nucleotide position 395±374 of the published
NADL sequence (Collett et al., 1988; the Eco RI site added is underlined) and Superscript
II reverse transcriptase (Gibco BRL) in a 20 ml reaction. PCR was performed according to
standard procedures using primer P43 and sense primer P41 (50 -TTCATCTAGA-
GAGGCTAGCCATGCCCTTAG-30 ), nucleotide positions 98±117 of the published
NADL sequence (Collett et al., 1988; the Xba I restriction site added is underlined).
 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97 89

2.9. Sequencing and phylogenetic analysis

The amplification products were purified from a 1.5% agarose gel with the Qiaquick
Gel Extraction Kit (Qiagen). Both strands were sequenced using the ABI PRISM Dye
Terminator Cycle Sequencing Reaction Kit (Perkin±Elmer, Foster, CA) with primers P41
and P43. The extension products were purified on Centri-Sep columns (Princeton
Separations, Adelphia, NJ) and loaded into an ABI PRISM 310 Genetic Analyser. The
GCG software package (Genetics Computer Group, Madison, WI) was used for sequence
analysis, which included a 247 bp fragment of A1138/69, BR275, Soldan and published
sequences of genotype I and II BVD viruses (Renard et al., 1987; Collett et al., 1988;
Deng and Brock, 1992; Pellerin et al., 1994). Distances between sequences were
calculated by means of the Jukes±Cantor index. The phylogenetic tree was constructed
using the unweighted pair-group arithmetic averaging method.

3. Results

3.1. Analysis of the ELISA antigens by means of monoclonal antibodies

Since our aim was to develop an ELISA capable of detecting antibodies to a wide
variety of BVDV strains, we analyzed the antigen preparations for the presence of
conserved and variable BVDV proteins. Fig. 1 shows that antigen prepared by
solubilization of BVDV-infected cells by Tween 20 contained the conserved non-
structural protein NS3; however, it contained only very minor amounts of E2 ± the
neutralizing antibodies' main target ± which is known to vary between the individual viral
strains (Becher et al., 1994). In contrast, antigen prepared by incubation with the non-
ionic detergent OGP contained little NS3 but was rich in E2.

3.2. Validation of the ELISA by means of sera obtained from Swiss cattle

In initial experiments we determined the cut-off value of the ELISA using 357 Swiss
cattle sera and taking neutralization titers of >25 as positive. To standardize the ELISA
we used a positive control serum. Fig. 2 shows that the ELISA allowed a clear distinction
of positive and negative sera, with only three false positive, three false negative and one
neutralization-positive serum in the indeterminate range. Based on these results, we
determined the negative range as being 20% of the positive control serum and the
positive range as 30%. Taking a further 583 sera diluted 1:5, 1:25 and 1:125, we
observed concordant results (i.e. SNT‡/ELISA‡ or SNTÿ/ELISAÿ) in 923 of these sera
(Fig. 3). Fifteen serum samples (1.6%) were negative in ELISA but positive in SNT. We
were able to isolate BVDV from the buffy coat of four of these samples. Two samples
(0.2%) were ELISA-positive but SNT-negative. These samples showed a positive
reaction when tested by SNT using the R56/74 BVDV strain, a virus isolated from a cow
in Switzerland (Steck et al., 1980). Taking the serum neutralization test as the gold
standard, the newly established ELISA has a sensitivity of 97.5% and a specificity
of 99.2%.
90 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97

Fig. 1. Analysis of ELISA antigens prepared from the A1138/69 BVDV strain, prepared with Tween 20 (black
bars) or OGP (white bars) as described in Section 2. Plates were incubated with monoclonal antibodies and a
positive and negative control serum, respectively. Monoclonal antibody C 16 is directed against NS3, the other
monoclonal antibodies detect the surface glycoprotein E2. Bars indicate the reactivity of antibodies against
antigenic determinants, presented by the different preparations of ELISA antigens.

Fig. 2. Comparison of serum neutralization test with ELISA using 357 Swiss cattle sera. Sera diluted in five-
fold steps starting with 1:5 were tested for antiviral antibody in serum neutralization and ELISA. Sera with a
neutralization titer of 25 were rated negative, those with a titer of >25 were rated positive. Sera with an
extinction of 20% of that of the positive control serum were rated negative in the ELISA, those reacting >30%
positive in the ELISA.
 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97 91

Fig. 3. Validation of the Tween 20 ELISA using serum neutralization as the reference standard. A titer of 25 was
considered positive in serum neutralization. 940 serum samples were used (ˆ100%). A titer of 20 was considered
positive. In 98.2% of the sera tested, a concordance between the two tests was observed. In 1.6%, the reaction was
negative in ELISA and positive in SNT, whereas in 0.2% the reaction was positive in ELISA and negative in SNT.

To assess whether the ELISA based on antigen from the German A1138/69 BVDV
isolate was suitable for detecting antibodies in cattle infected with BVDV strains
circulating in Brazil, we tested its performance using a panel of randomly chosen sera
from Brazilian cattle. Since the antigenic properties of BVDV strains circulating in Brazil
had not been determined, we prepared antigen from two BVDV strains isolated in Brazil
and compared the reaction of the sera with plates coated with BVDV antigens prepared
from the German and the two Brazilian viral strains. Fig. 4 shows that there was complete
agreement in the results obtained in all three assays, which indicates that antigen prepared
from the German A1138/69 isolate is indeed suitable for the demonstration of antibodies
in BVDV strains circulating in Brazil.

3.3. The prevalence of antibodies in cattle sera from Brazil

430 cattle sera obtained from cattle of 19 farms of the Rio Grande do Sul State and from one
farm in the Corrientes State of Argentina were analyzed by means of the A1138/69 ELISA
(Table 1). The average prevalence of antibodies to BVDV was 56.0%, with a confidence
interval of 15.1%, adjusted for clustering (Donald and Donner, 1987), and varying from
0% (47 sera from Santa VitoÂria do Palmar 2) to 100% (12 sera from Aratiba).

3.4. Sequence analysis of two BVDV strains isolated in Brazil

The last two-thirds of the 50 UTR of the genome of the A1138/69 BVDV strain and the
Brazilian BR275 and Soldan isolates were sequenced directly from RT±PCR reactions.
92 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97

Fig. 4. Comparison of ELISA using antigen prepared from the A1138/69 BVDV strain and the Brazilian Soldan
and BR275 strains. Antigens were prepared as described in Section 2 and plates were coated with the A1138/69
(diamonds), BR275 (squares), and Soldan (triangles) BVDV strains.

The results of sequencing and the dendrogram (Fig. 5) showed these viruses to be of the
two different genotypes described for BVDV. The BR275 isolate and the A1138/69 strain
cluster with type Ia strains. With the exception of 1 base, BR275 is identical to the SD-1
strain, whereas A1138/69 relates to the Oregon strain. However, the Soldan isolate
clusters with type II viruses and is more distantly related to the strains in the group
previously described.

4. Discussion

BVDV has a worldwide distribution and previous reports showed infection with this
virus to also occur in Brazil (Brown et al., 1989; Rweyemamu et al., 1990). Since BVDV
strains are genetically highly heterogeneous, particularly in the genes encoding the
surface glycoproteins responsible for the induction of neutralizing antibodies, we decided
to use in our study an assay that detects antibodies to conserved antigens such as the non-
 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97 93

Table 1
Number of negative and positive sera for each farm

Region Negative Positive

Barra do Ribeiro 10 11
CandelaÂria 6 15
Aratiba 0 12
CapaÄo do LeaÄo 1 24
Carazinho 2 4
Corrientes (Argentina) 0 10
Eldorado do Sul 9 28
Livramento 36 11
Mostardas 4 19
Pedro OsoÂrio 1 21
Piratini 1 0 2
Piratini 2 6 19
Rio Grande 1 18 17
Rio Grande 2 2 12
Rio Grande 3 1 19
Santa VitoÂria do Palmar 1 14 13
Santa VitoÂria do Palmar 2 47 0
SaÄo Lourenc,o do Sul 1 3 2
SaÄo Lourenc,o do Sul 2 1 14
SaÄo Sepe 16 0
Total 177 253

The farms are identified by the region of origin of the sera and the numbers following the names are used to
differentiate different farms from the same region.

structural proteins of the virus. The broad reactivity of the ELISA antigen was validated
using antigen prepared from the cytopathic A1138/69 BVDV strain, originally isolated in
Germany from a cow suffering from mucosal disease (Frey and Liess, 1971). This strain
is routinely used in our laboratory in SNT for the detection of antibodies to BVDV.
Potentially, an ELISA based on conserved antigens of this viral strain should permit the
detection of antibodies induced by a wider variety of antigenically different BVDV
strains. The evaluation of the antigens prepared by solubilization with Tween 20
demonstrated the presence of the NS3 protein, which, due to its antigenically conserved
and immunodominant properties, is well suited for serology of BVDV (Kwang et al.,
1991; Lecomte et al., 1991; Paton et al., 1991; Petric et al., 1992). Interestingly, antigen
prepared by solubilization with Tween 20 was essentially free of the more variable E2
surface glycoprotein whereas solubilization with OGP resulted in antigen rich in E2 but
poor in NS3. Preliminary results suggest that the OGP prepared antigen may indeed be
suitable for detecting strain-specific antibodies (Strasser et al., unpublished). The absence
of E2 from the antigen prepared by solubilization with Tween 20 may well be important
because it minimizes strain-specific differences in the sensitivity of antibody detection.
The Tween 20-prepared antigen of the German BVDV isolate was demonstrated to have a
high sensitivity and specificity regarding the detection of antibodies present in randomly
selected cattle sera from Switzerland. During the validation of the ELISA we made an
observation which is most likely to be related to the unique biology of the interaction of
94 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97

Fig. 5. Dendrogram of the nucleotide sequence of part of the 50 untranslated regions of BVDV strains. The
horizontal branch lengths in the clusters are proportional to the similarity between the sequences.

BVDV with its bovine host. At first sight, those sera which were negative in ELISA and
positive in SNT may reflect a lower sensitivity of the ELISA as compared to SNT.
However, the isolation of non-cytopathic BVDV from some of these sera may indicate
that these animals were in fact persistently infected. The neutralizing antibodies found in
these sera may therefore, indicate that, at some stage, these animals were superinfected
with BVDV strains carrying in their E2 glycoproteins antigenic determinants different to
those of the persisting strains. Such superinfection is, however, expected to stimulate only
a weak immune response to NS3 because this protein is more conserved among BVDV
strains. The two sera which were positive in ELISA but negative in SNT may, in contrast,
reflect infection with a viral strain antigenically different from A1138/69. This
interpretation is supported by the positive result of SNT when we used a Swiss BVDV
isolate.
 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97 95

Although the amino acid sequences of certain regions of NS3 are completely conserved
among different BVDV strains, other regions display differences (Hertig et al., 1995).
Since the antigenic properties of NS3 of BVDV strains circulating in Brazil were not
known, we carried out a comparison of the A1138/69 ELISA with ELISAs based on
antigen from two BVDV strains isolated in Brazil prepared according to the same
protocol. The complete agreement between these assays enabled us to determine the
prevalence of BVDV infections. Although the confidence interval of this study is rather
wide (15%), the prevalence of 56% found in the serum samples available to us suggests
that BVDV infections are as widespread in southern Brazil as they are in Europe
(Harkness et al., 1978; Meyling, 1983; Houe and Meyling, 1991).
Interestingly, we observed a different antibody prevalence on individual farms, the
extremes being 0% (i.e., no evidence of previous contact with BVDV) and 100%. This
situation clearly differs from that in Switzerland where, although the overall antibody
prevalence is similar, differences between individual farms are smaller (unpublished
observation). The more heterogeneous antibody prevalence on Brazilian in contrast to
Swiss farms may be explained by differences in management practice. In Switzerland,
farms are comparatively small (average number of animals 28.2) whereas farms in Brazil
are considerably larger (average number of animals 72.8). Consequently, exchange and
trade of animals between farms can be expected to be more frequent in Switzerland than
in Brazil. In addition, in Switzerland it is common practice to send heifers originating
from different farms away to spend the summer on common grazing pastures in the Alps.
This increases the risk of animals contracting the infection from persistently infected
animals, which results in a more homogeneous prevalence of antibody-positive animals
on individual farms.
The analysis of part of the 50 untranslated region of the two Brazilian isolates revealed
that the Soldan strain clusters with the group of BVDV genotype II viruses. Interestingly,
the Soldan strain considerably broadened the heterogeneity of the BVDV genotype II viruses,
with a drop in the degree of similarity from approximately 95% to 88%. It will be interest-
ing to analyze additional isolates from the geographical area of origin of this isolate in order to
obtain evidence that genotype II viruses are indeed genetically more heterogeneous than
indicated by the nucleotide sequences of the previous isolates. The second strain analyzed,
BR 275, is very closely related to SD-1, a non-cytopathic North American strain for which
the full-length sequence was published in 1992 (Deng and Brock, 1992).
Future work will aim at obtaining more precise information on the epidemiology of
BVDV in southern Brazil and on the genetic heterogeneity of the viral strains circulating
in that region. Such investigations are of interest not least because of the possibility that
pestiviruses may also be present among artiodactyla indigenous to South America.

Acknowledgements

We thank JuÈrg RuÈfenacht and Patrick Schaller for the statistical and epidemiological
analysis, H. Stalder and K.L. Haag for the phylogenetic tree, Dr. P.M. Roehe for the
Soldan isolate, Dr. A. Zaha and I. Vaz Junior for the Brazilian cattle sera, and Ruth
Parham for linguistic improvements.
96 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97

C.W. Canal had a scholarship from CNPq-RHAE and the Institute of Veterinary
Virology, University of Bern, Switzerland. Financial support was provided by FAPERGS,
CNPq). E. Peterhans was supported by the Swiss National Science Foundation (Grant no.
31-39733.93).

References

Becher, P., Shannon, A.D., Tautz, N., Thiel, H.-J., 1994. Molecular characterization of border disease virus, a
pestivirus from sheep. Virology 198, 542±551.
Brown, C.C., Olander, H.J., Castro, A.E., Behymer, D.E., 1989. Prevalence of antibodies in goats in north-
eastern Brazil to selected viral and bacterial agents. Trop. Anim. Health. Prod. 21, 167±169.
Brownlie, J., Clarke, M.C., Howard, C.J., 1987. Pathogenesis and epidemiology of bovine virus diarrhea virus
infection of cattle. Ann. Rech. VeÂt. 18, 157±166.
Collett, M.S., Larson, R., Gold, C., Strick, D., Anderson, D.K., Purchio, A.F., 1988. Molecular cloning and
nucleotide sequence of the pestivirus bovine viral diarrhea virus. Virology 165, 191±199.
Dahle, J., Liess, B., 1992. A review on classical swine fever infections in pigs: epizootiology, clinical disease
and pathology. Comp. Immun. Microbiol. Infect. Dis. 15, 203±211.
Deng, R., Brock, K.V., 1992. Molecular cloning and nucleotide sequence of a Pestivirus genome, noncytopathic
bovine viral diarrhea virus strain SD-1. Virology 191, 867±879.
Donald, A., Donner, A., 1987. Adjustments to the Mantel±Haenszel chi-square statistic and odds ratio variance
estimator when the data are clustered. Statistics Medicine 6, 491±499.
Frey, H.R., Liess, B., 1971. Vermehrungskinetik und Verwendbarkeit eines stark zytopathogenen VD-MD-
Virusstammes fuÈr diagnostische Untersuchungen mit der Mikrotiter-Methode. Zbl. Vet. Med. B 18,
61±71.
Grassi, M., Wandeler, A.I., Peterhans, E., 1989. Enzyme-linked immunosorbent assay for determination of
antibodies to the glycoprotein of rabies virus. J. Clin. Microbiol. 27, 899±902.
Harkness, J.W., Sands, J.J., Richards, M.S., 1978. Serological studies of mucosal disease virus in England and
Wales. Res. Vet. Sci. 24, 98±103.
Hertig, C., Stalder, H., Peterhans, E., 1995. Genetic heterogeneity within the coding regions of E2 and NS3 in
strains of bovine viral diarrhea virus. Gene 153, 191±195.
Hibberd, R.C., Turkington, A., Brownlie, J., 1993. Fatal bovine viral diarrhea virus infection of adult cattle. Vet.
Rec. 132, 227±228.
Horzinek, M.C., 1991. Pestiviruses ± taxonomic perspectives. Arch. Virol. Suppl. 3, 1±5.
Houe, H., 1995. Epidemiology of bovine viral diarrhea virus. Vet. Clin. North Am Food Anim. Pract. 11(3),
521±547.
Houe, H., Meyling, A., 1991. Prevalence of bovine virus diarrhea (BVD) in 19 Danish dairy herds and
estimation of incidence of infection in early pregnancy. Prev. Vet. Med. 11, 9±16.
Kwang, J., Bolin, S., Littledike, E.T., Dubovi, E.J., Donis, R.O., 1991. Expression of the p80 region of bovine
viral diarrhea virus and identification of specific antibodies to this recombinant protein in bovine sera.
Biochem. Biophys. Res. Commun. 178, 1326±1334.
Lecomte, C., Vandenbergh, D., Vanderheijden, N., De Moerlooze, L., Pin, J.J., Chappuis, G., Desmettre, P.,
Renard, A., 1991. Identification and production of pestivirus proteins for diagnostic and vaccination
purposes. Arch. Virol. (Suppl. 3), 149±156.
Meyling, A., 1983. An immunoperoxidase technique Detection of BVD virus in serum of clinically and
subclinically infected cattle. Proc. 3rd Int. Symp. Vet. Lab. Diag. pp. 179±184.
Paton, D.J., Edwards, S., Wensvoort, G., 1991. An ELISA detecting antibody to conserved pestivirus epitopes.
J. Virol. Methods 31, 315±324.
Pellerin, C., Van den Hurk, J., Lecomte, J., Tijssen, P., 1994. Identification of a new group of bovine viral
diarrhea virus strains associated with severe outbreaks and high mortalities. Virology 203, 260±268.
Peters, W., Greiser-Wilke, I., Moennig, V., Liess, B., 1986. Preliminary serological characterization of bovine
viral diarrhoea virus strains using monoclonal antibodies. Vet. Microbiol. 12, 195±200.
 C.W. Canal et al. / Veterinary Microbiology 63 (1998) 85±97 97

Petric, M., Yolken, R.H., Dubovi, E.J., Wiskerchen, M., Collett, M.S., 1992. Baculovirus expression of pestivirus
non-structural proteins. J. Gen. Virol. 73, 1867±1871.
Radostits, O.M., Littlejohns, I.R., 1988. New concepts in the pathogenesis, diagnosis and control of diseases
caused by the bovine viral diarrhea virus. Can. Vet. J. 29, 513±528.
Renard, A., Dina, D., Martial, J.A., 1987. Complete nucleotide sequence of bovine viral diarrhea virus genome
and its fragment, useful for making antigenic proteins useful in therapy and diagnosis. European Patent
Application No. 0208672.
Ridpath, J.F., Bolin, S.R., Dubovi, E.J., 1994. Segregation of bovine viral diarrhea virus into genotypes. Virology
205, 66±74.
Rweyemamu, M.M., Fernandez, A.A., Espinosa, A.M., Schudel, A.A., Lager, I.A., Mueller, S.B., 1990.
Incidence, epidemiology and control of bovine virus diarrhoea virus in South America. Rev. Sci. Tech. 9,
207±221.
Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring
Harbour Laboratory Press, Cold Spring Harbour, NY.
Steck, F., Lazary, S., Fey, H., Wandeler, A., Huggler, C., Oppliger, G., Baumberger, H., Kaderli, R., Martig, J.,
1980. Immune responsiveness in cattle fatally affected by bovine virus diarrhea-mucosal disease. Zbl. Vet.
Med. B 27, 429±445.