Coronaviruses are enveloped single-stranded positive-sense RNA viruses. The IBV genome is 27.569 kbp in length and contains four coronavirus structural protein genes, spike (S), envelope (E), membrane (M) and nucleocapsid (N), plus non-structural protein genes 3 and 5 and its viral replicase gene (Rep). The order of these genes is 5′ Rep-S-3-E-M-5-N 3′ (Boursnell et al. 1987; Spaan, Cavanaugh & Horzinek 1988). IBV S protein undergoes post-translational cleavage to S1 and S2 subproteins. The S1 protein is responsible for attachment to the host cell membrane and also has epitopes to which neutralising and haemagglutination-inhibiting antibodies bind (Kant et al. 1992; Koch et al. 1990). Two hypervariable regions (HVRs) have been demonstrated within the 50–159 residue regions of the S1 gene. HVR1 lies within region 38–51 whilst HVR 2 is located between amino acids 99 and 115 (Cavanaugh, Davis & Mockett 1988; Cavanaugh et al. 1992; Wang, Xu & Collisson 1997). A third HVR between amino acids 274 and 387 has also been described (Lee & Jackwood 2001). The S1, S2 and M proteins have all been shown to induce cell-mediated immunity with the N gene proposed as carrying the predominant T cell epitope (Ignjatovic & Galli 1994). The high degree of variability within the S1 gene has been attributed to the lack of proofreading by the viral RNA-dependent RNA polymerase and to recombination events during replication (Toro, Van Santen & Jackwood 2012).

The emergence of variation within the IBV population has been widely described since 1951, when Jungherr, Chomiak and Luginbuhl (1956) demonstrated that the Connecticut isolate neither cross-neutralised nor cross-protected against the traditional H120 Massachusetts (Mass) isolate of the 1940s. The first published reports of IBV from South Africa were by Morley and Thompson in 1984. This was confirmed as an unknown variant by Cook et al. in 1999. These viruses were associated with swollen-head syndrome observed in infected flocks throughout southern Africa. Novel IBV has been described in Nigeria (De Wit, Cook & Van der Heijden 2011) as well as Zimbabwe, where a QX-like IBV strain has been reported (Jackwood 2012).

The South African poultry industry is the single largest agricultural subsector, accounting for an annual turnover of R31.7 billion per annum (South African Poultry Association 2012) with flock sizes ranging from several hundred to large complexes of 600 000 birds. Perdue and Seal (2000) listed the estimated losses per flock from an IBV infection as 10% to 20% of market value. This can devastate a poultry farm, where margins are relatively small. The economic impact of this disease is thus of great importance. Currently available research findings do not describe variation over a large collection of South African IBV isolates and thus do not adequately describe the degree of variation within the South African broiler flock. The core question revolves around the extent of variation present in the South African IBV population. Objectives of this study included identifying variation within IBV isolates, describing the correlation between serotyping and genotyping, as a method of variation identification and describing variation within the KwaZulu-Natal IBV isolates over time.

The standard virus isolation procedures are based on the published laboratory manual for the isolation, identification and characterisation of avian pathogens in 9 to 10-day-old embryonated chicken eggs via the allantoic sac inoculation route (Gelb 1989). Dead embryos were opened after candling; diagnosis of the cause of death was conducted and in the case of evidence that the embryo was infected with IBV, the allantoic fluid was collected. Typical signs of IBV infection in the embryo, namely stunting, urate deposits in the mesonephros and clubbing of the down feathers, were noted. In the cases where no pathological signs were seen in the embryo, the allantoic fluid was still collected and reinoculated for an additional passage. A total of three passages were performed until the sample was confirmed negative. Inoculations with supporting IBV clinical signs were confirmed for the presence of IBV through a chloroform inactivation test as well as diagnostic polymerase chain reaction (PCR) assay. The chloroform inactivation test differentiates between enveloped and non-enveloped viruses, and in combination with embryo pathology and haemagglutination tests a diagnosis of IBV can be made. PCR-confirmed IBV-positive allantoic fluid isolates were then stored at -70 °C until further processing.

Complementary DNA (cDNA) reactions were diluted by adding 20 µL PCR-grade water (Roche). PCR was conducted with 10 µL 2X Thermo Scientific Phusion Flash High-Fidelity PCR master mix (Inqaba Biotech, Pretoria), 2 µL Onderstepoort (OP) IBV S1 forward primer, 2 µL OP IBV S1 reverse primer, 1 µL PCR-grade H₂O and 5 µL cDNA. The target region was amplified in a Veriti thermal cycler (Life Technologies) using a cycling profile of 98 °C for 10 s, 30 cycles of 98 °C for 5 s, 50 °C for 15 s and an elongation phase of 72 °C for 2 min. A final elongation step of 72 °C for 4 min was added. OP IBV S1 FOR: 5′-GAACAAAASACNGACTTAG-3′, OP IBV S1 REV: 5′-CCATAACTAACATAAGGRCAA-3′.

PCR products were electrophoretically separated on a 1% agarose gel stained with ethidium bromide. A KAPA™ universal ladder was included to identify the 1700 bp S1 gene amplicon. Bands of the correct size were excised and DNA purified using QiaQuick PCR purification kit (Qiagen) following the recommended procedure.

Purified 1.7 kb PCR amplicons were submitted to Inqaba Biotech (Pretoria) for Sanger sequencing. Results were edited using Chromas Lite and then aligned using BioEdit version 7.1.11 (Hall 1999). The Basic Local Alignment Search Tool (BLAST) server was used to retrieve similar sequences selected from Genbank for comparative analysis, together with important reference viruses. Pairwise similarity was calculated using BioEdit (Hall 1999) and phylogenetic trees were constructed in MEGA 5.2 (Tamura et al. 2011). Full S1 gene sequences were deposited in Genbank under the accession numbers KJ200273–KJ200291. Partial sequences are available from the corresponding author on request.

Phylogenetic relationships were inferred using the neighbour-joining tree inference method with a Jukes-Cantor model of sequence evolution using MEGA 5.2 (Tamura et al. 2011). One thousand bootstrap samplings were performed to assign confidence values to branching orders.

Forty-one samples produced a band of approximately 1700 base pairs (data not shown), of which full S1 gene sequences could be assembled for 19 isolates, with the remainder generating either partial forward or reverse sequences. Despite numerous attempts on the service provider’s part to troubleshoot sequencing reactions, they were unable to generate the complementary forward or reverse reactions for 30 of the amplicons submitted. The reasons for this are unknown, since the exact same forward and reverse primers were used for sequencing as for RT-PCR, and numerous re-submissions of fresh amplified RT-PCR product in sufficient molar quantities were made. Phylogenetic analysis of the full S1 gene sequences are presented in Figure 1. Partial sequence trees are not shown, but were used to genotypically classify the isolates (results are presented in Table 1). Isolates were classified as either QX-like (n = 4) or Mass-type (n = 16).

S1 gene nucleotide sequence comparisons (data not shown) revealed variability in QX-like isolates CK/ZA/3665/11, CK/ZA/4916/11, CK/ZA/4871/12 and CK/ZA/6720/12 when compared to a standard H120 Mass serotype with a high number of base changes observed at positions 95–119, 142–148,165–172 and 695–708. This was evident in the translated amino acid sequences (Figure 2), and notably between residues 58 and 62 of these isolates where ISSES is replaced with STNHT.

FIGURE 2: Multiple amino acid sequence alignment for full S1 proteins. Hypervariable regions are indicated in grey. The first sequence is used as a reference and identities are plotted as ‘.’.

FIGURE 2 Continues: Multiple amino acid sequence alignment for full S1 proteins. Hypervariable regions are indicated in grey. The first sequence is used as a reference and identities are plotted as ‘.’.

In this QX-like group, residues 77–85 contained the motif KDVYNQSVA compared to the Mass grouping with a motif of HGGRVVNA. A large insertion was also detected between residues 122 and 129 with a motif of SSGSGSCP, which is not present in any of the Mass-type isolates. Two point mutations within CK/ZA/3665/11 resulted in the replacement of asparagine with histidine at position 24 and then aspartic acid with tyrosine at position 248. Variation was observed between the two 2011 and the 2012 isolates within the QX-like group; the protein sequence at position 18–22 of the 2012 isolates was FALWK and CALCS as compared to the 2011 isolates at the same position. The most variable portion of the amino acid sequence for this group occurred at residue 400, where two of four isolates had consensus and variation between asparagine, histidine and tyrosine was seen.

The remaining isolates were closely related to reference sequence H120 (Mass serotype). Single-point mutations were observed at various points across the isolates. No variation in nucleotide sequence is seen after base pair 792, demonstrating greater conservation in the second half of the gene.

Pairwise similarities were calculated between the study isolates (Table 2) and selected serotype sequences obtained from Genbank (Table 3). Pairwise comparisons against reference sequences support the description of two distinct groups of isolates. QX-like isolates shared

97.2% – 98.4% nucleotide sequence identities, corresponding to 97.6% – 98.8% identity at the amino acid level. These QX-like isolates only shared 77.7% – 79% nucleotide sequence and 85.9% – 88.6% amino acid sequence identities with the Mass-like isolates. Sequence variation within the Mass-like group was less, at 98.9% – 100% and 99% – 100% nucleotide and amino acid similarities respectively (Table 2). Four isolates (CK/ZA/0890/81, CK/ZA/6689/12, CK/ZA/5315/11 and CK/ZA/2287/11) were very closely related. These sequences differed by 0.3% or four point mutations at positions 294, 302, 374 and 483.

FIGURE 1: Phylogenetic tree of full S1 gene nucleotide sequences (1665 bp) with isolates sequenced in this study highlighted.

The remainder of the Mass-type South African isolates shared high degrees of similarity with two point mutations common to all isolates when compared to the reference strain. The replacement of thymine with cysteine at residue 11 and thymine with adenine at residue 138 was found amongst all Mass isolates. The highest frequency of variation in this case was found between base pairs 706 and 818, with 10-point mutations found in this region.

These two distinct groupings were also evident when local strains were compared to international reference strains (Table 3). CK/ZA/3665/11, CK/ZA/4916/11, CK/ZA/4871/12 and CK/ZA/6270/12s shared the highest nucleotide and amino acid homology with IBV QX-like Chinese TC07-1 and UK 3355/09 strains. The South African Mass-type viruses shared 99.2% – 99.7% nucleotide sequence identity and 96.5% – 96.7% amino acid similarity with isolate Mass Avail.

When compared against the year or area of isolation it was found that the highly similar Mass grouping originated from KZN between April 2011 and September 2011 and then again in October 2012. One isolate from the North West province from July 2011 and two from the Western Cape collected in October 2011 also fell within this group. Strong correlations were observed between the isolates that originated outside KZN. Despite the geographic areas being far apart there was movement of eggs and embryos between provinces during the study period. The variability of 0.4% observed between CK/ZA/6743a/11 and CK/ZA/6743b/11 is interesting, since these samples were collected from the same farm on the same day but originated from separate chicken houses or units. The existence of an insertion at 673 and point mutations at 4, 6, 296 and 879 demonstrate the existence of variability and selection of subpopulations on a single farm. The selective criteria for these subpopulations are unknown, but variables such as maternally derived antibody levels and timing of vaccination could apply greater selective pressure for the emergence of variants of the epitopes found on vaccine-derived viruses (De Herdt et al. 2001; Meulemans et al. 2001; Mondal & Naqi 2001). An alternative hypothesis is that the individual house conditions with regard to air quality and the degree of damage to the chicken’s tracheal epithelium and cilia affect the selection of S1 protein subpopulations.

The use of especially the live day-old H120 vaccine and the close similarity of Mass-type isolates to H120 must not be overlooked as the most likely influence on the prevalence of the H120 group. This implies that the most likely contributors to IBV serotypes within these flocks are Mass-type live vaccines. However, there is the possibility that either through genetic drift or quasi-species selection, field-adapted virulent Mass-type IBV was able to proliferate in suitable hosts. The basis of this statement is that all the samples collected were from clinical cases of respiratory disease and the presence of suggestive pathology on postmortem examination. Recombination and reversion to virulence within Mass-type IBV vaccines have been described and field experience over the years of use has found that poor vaccination technique, concurrent viral or bacterial infections as well as on-farm management conditions can lead to IBV pathology in light of vaccine usage (Cavanaugh, Davis & Cook 1992).

Variability within the QX-like genotype was greater at 97.7% – 98.5% than that seen in the Mass grouping at 99.0% – 99.7%. A temporal trend in genetic drift was observed. This observation is in contrast to evolutionary patterns within the Mass group and supports the existence of genetic drift within this IBV subpopulation. The greatest similarity was noted when comparing it to the QX-like strains of European origin (97%) as opposed to the Asian QX-like strains (94%), which provides circumstantial evidence of the spread of QX-like IBV from Asia into Europe and then downward into Africa.

The results of the KZN isolates would support the high host adaptability of the QX-like variant and identify it as the leading ‘variant’ within KZN poultry flocks.

Shaw, Britton and Cavanaugh (1996) investigated Belgian nephrotropic IBV strains and reported that the S1 gene differed by 21% – 45% from vaccine strains whilst the S2 gene only had 10% – 12% differentiation. The 793/B strain was shown to display between 21% and 25% variation amongst field isolates, whilst still retaining the same serotype (Adzhar & Gough 1997). The CK/CH/LDL/971 strain of IBV was shown to have < 80% homology between the vaccine and pathogenic strain as well as between itself and other IBV types (Liu et al. 2009). Host-driven selection has been postulated and can occur as demonstrated during egg passage of several 793/B isolates; although this selection did not alter the pathogenicity, it does illustrate the quasi-species theorem regarding IBV populations (Cavanaugh et al. 2005). Jackwood, Hilt and Callison (2003) utilised real-time PCR techniques and melting curves to visualise the existence of quasi-species within the IBV Beaudette strain. Re-isolated virus has been demonstrated to contain subpopulations, again supporting the quasi-species theorem (McKinley, Hilt & Jackwood 2008).

The correlation between QX-like serotypes and the actual sequence was inconsistent, as only 4 of the 11 serotype QX-like isolates were classified within the QX-like genotype. All seven remaining Mass serotypes genotyped into the same group. By comparison, when examining partial S1 gene sequence data, none of the QX-like partial sequences were serotyped as QX-like. Seven serotype QX-like isolates genotyped as Mass, whilst the two isolates that did not fall into a known serotype at the time were split between the Mass and QX-like genotype. The above results cast doubt on the correlation between serotype and genotype results. It has been described that small sequence variations may alter the virus neutralising epitope and thereby the serotype of the virus (Cavanaugh et al. 1992). An alternative hypothesis could be that coinfection with multiple strains of IBV is responsible for clinical disease, and between the culture or serum neutralisation test and the subsequent successive passages and RT-PCR selection for host-specific strains there is selection for an alternate dominant strain (Liu et al. 2009; Xu et al. 2007). Comparing these protein sequences to the isolates that both genotyped and serotyped as Massachusetts, there is no obvious genetic marker that serotyped as QX-like in contrast to the group that serotyped as Mass-like.

Boursnell, M.E.G., Brown, T.D.K., Foulds, I.J., Green, P.F., Tomley, F.M. & Binns, M.M., 1987, ‘Completion of the sequence of the genome of the coronavirus avian infectious bronchitis virus’, Journal of General Virology 68, 57–77. http://dx.doi.org/10.1099/0022-1317-68-1-57

Cavanaugh, D., 2001, ‘A nomenclature for avian coronavirus isolates and the question of species status’, Avian Pathology 30, 109–115. http://dx.doi.org/10.1080/03079450120044506

Cavanaugh, D., Davis, P.J. & Cook, J.K.A., 1992, ‘Infectious bronchitis virus: Evidence for recombination within the Massachusetts serotype’, Avian Pathology 21(3), 401–408. http://dx.doi.org/10.1080/03079459208418858

Cavanaugh, D., Davis, P.J., Cook, J.K.A., Li, D., Kant, A. & Koch, G., 1992, ‘Location of the amino acid differences in the S1 spike glycoprotein subunit of closely related serotype of infectious bronchitis virus’, Avian Pathology 21(1), 33–43. http://dx.doi.org/10.1080/03079459208418816

Cavanaugh, D., Davis, P.J. & Mockett, A.P.A., 1988, ‘Amino acids within the hypervariable region 1 of avian coronavirus IBV (Massachusetts serotype) spike glycoprotein are associated with neutralization epitopes’, Virus Research 11, 141–150. http://dx.doi.org/10.1016/0168-1702(88)90039-1

Cavanaugh, D. & Naqi, S.A., 1997, ‘Infectious bronchitis’, in B.W. Calnek (ed.), Diseases of poultry, 10th edn., vol. 18, pp. 511–527, Iowa State University Press, Ames.

Cavanaugh, D., Picault, J.P., Gough, R.E., Hess, M., Mawditt, K. & Britton, P., 2005, ‘Variation in the spike protein of the 793/B type of infectious bronchitis virus, in the field and during ultimate passage in chickens and embryonated eggs’, Avian Pathology 34(1), 20–25. http://dx.doi.org/10.1080/03079450400025414

Cook, J.K.A., 1983, ‘Isolation of a new serotype of infectious bronchitis-like virus from chickens in England’, Veterinary Record 112,104–105. http://dx.doi.org/10.1136/vr.112.5.104

Cook, J.K.A., Huggins, M.B. & Ellis, M.M., 1991, ‘Use of an infectious bronchitis virus and Escherichia coli model infection to assess the ability to vaccinate successfully against infectious bronchitis in the presence of maternally-derived immunity’, Avian Pathology 20, 619–626. http://dx.doi.org/10.1080/03079459108418801

Cook, J.K.A., Orbell, S.J., Woods, M.A. & Huggins, M.B., 1999, ‘Breadth of protection of the respiratory tract provided by different live-attenuated infectious bronchitis vaccines against challenge with infectious bronchitis viruses of heterologous serotypes’, Avian Pathology 28, 477–485. http://dx.doi.org/10.1080/03079459994506

De Herdt, P., Ducatelle, R., Uyttebroek, E., Sneep, A. & Torbeyns, R., 2001, ‘Infectious bronchitis serology in broilers and broiler breeders: Correlations between antibody titres and performance in vaccinated flocks’, Avian Diseases 45, 612–619. http://dx.doi.org/10.2307/1592902

De Wit, J.J., Cook, J.K.A. & Van der Heijden, H.M.J.F., 2011, ‘Infectious bronchitis virus variants: A review of the history, current situation and control measures’, Avian Pathology 40(3), 223–235. http://dx.doi.org/10.1080/03079457.2011.566260

Dhinakar Raj, G. & Jones, R.C., 1997, ‘Infectious bronchitis virus: Immunopathogenesis of infection in the chicken’, Avian Pathology 26, 677–706. http://dx.doi.org/10.1080/03079459708419246

Gelb, J. Jnr., 1989, ‘Infectious bronchitis’, in H.G. Purchase, L.H. Arp, C.H. Domermuth & J.E. Pearson (eds.), A laboratory manual for the isolation and identification of avian pathogens, 3rd edn., pp. 124–127, American Association of Avian Pathologists, Kendall Hunt, Dubuque.

Hall, T.A., 1999, ‘BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT’, Nucleic Acid Symposium Series 41, 95–98.

Ignjatovic, J. & Galli, L., 1994, ‘Immune responses to structural proteins of avian infectious bronchitis virus’, Avian Pathology 24, 313–332. http://dx.doi.org/10.1080/03079459508419072

Jackwood, M.W., 2012, ‘Review of infectious bronchitis virus around the world’, Avian Diseases 56, 634–641. http://dx.doi.org/10.1637/10227-043012-Review.1

Jackwood, M.W., Hilt, D.A. & Callison, S.A., 2003, ‘Detection of infectious bronchitis virus by real-time reverse transcriptase-polymerase chain reaction and identification of a quasispecies in the Beaudette strain’, Avian Diseases 47, 718–724. http://dx.doi.org/10.1637/6075

Jungherr, E.L., Chomiak, T.W. & Luginbuhl, R.E., 1956, ‘Immunologic differences in strains of infectious bronchitis virus’, Proceedings of the 60th Annual Meeting of the US Livestock Association, Chicago, pp. 203–209.

Kant, A., Koch, G., Van Roozelaar, D.J., Kusters, J.G., Poelwijk, F.A.J. & Van der Zeijst, B.A.M., 1992, ‘Location of the antigenic sites defined by neutralizing monoclonal antibodies on the S1 avian infectious bronchitis virus glycopolypeptide’, Journal of General Virology 73, 591–596. http://dx.doi.org/10.1099/0022-1317-73-3-591

Koch, G., Hartog, L., Kant, A. & Van Roozelaar, D.J., 1990, ‘Antigenic domains on the peplomer protein of avian infectious bronchitis virus: Correlation with biological functions’, Journal of General Virology 71, 1929–1935. http://dx.doi.org/10.1099/0022-1317-71-9-1929

Lee, C. & Jackwood, M.W., 2001, ‘Origin and evolution of Georgia 98 (GA98), a new serotype of avian infectious bronchitis virus’, Virus Research 80, 33–39. http://dx.doi.org/10.1016/S0168-1702(01)00345-8

Liu, S., Zhang, X., Gong, L., Yan, B., Li, C., Han, Z. et al., 2009, ‘Altered pathogenicity, immunogenicity, tissue tropism and 3-7 kb region sequence of an avian infectious bronchitis coronavirus strain after serial passage in embryos’, Vaccine 27, 4630–4640. http://dx.doi.org/10.1016/j.vaccine.2009.05.072

McKinley, E.T., Hilt, D.A. & Jackwood, M.W., 2008, ‘Avian coronavirus infectious bronchitis attenuated live vaccines undergo selection of subpopulations and mutations following vaccination’, Vaccine 26, 1274–1284. http://dx.doi.org/10.1016/j.vaccine.2008.01.006

McKinley, E.T., Jackwood, M.W., Hilt, D.A., Kissinger, J.C., Robertson, J.S., Lemke,C. et al., 2011, ‘Attenuated live vaccine usage affects accurate measures of virus diversity and mutation rates in avian coronavirus infectious bronchitis virus’, Virus Research 158, 225–234. http://dx.doi.org/10.1016/j.virusres.2011.04.006

Meulemans, G., Boschmans, M., De Caesstecker, M., Van den Berg, T.P., Denis, P. & Cavanaugh, D., 2001, ‘Epidemiology of infectious bronchitis virus in Belgian broilers: A retrospective study, 1986 to 1995’, Avian Pathology 30, 411–421. http://dx.doi.org/10.1080/03079450120066412

Mondal, S.P. & Naqi, S.A., 2001, ‘Maternal antibody to infectious bronchitis virus: Its role in protection against infection and development of active immunity to vaccine’, Veterinary Immunology and Immunopathology 79, 31–40. http://dx.doi.org/10.1016/S0165-2427(01)00248-3

Morley, A.J. & Thompson, D.K., 1984, ‘Swollen-head syndrome in broiler chickens’, Avian Diseases 28, 238–243. http://dx.doi.org/10.2307/1590147

Perdue, M.L. & Seal, B.S., 2000, ‘Impact of avian viruses’, Viral Ecology 12, 549–591. http://dx.doi.org/10.1016/B978-012362675-2/50016-1

Shaw, K., Britton, P. & Cavanaugh, D., 1996, ‘Sequence of the spike protein of the Belgian B164S isolate of nephropathogenic infectious bronchitis virus’, Avian Pathology 25(3), 607–611. http://dx.doi.org/10.1080/03079459608419165

Spaan, W., Cavanaugh, D. & Horzinek, M.C., 1988, ‘Coronaviruses: Structure and genome expression’, Journal of General Virology 69, 2939–2952. http://dx.doi.org/10.1099/0022-1317-69-12-2939

Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M. & Kumar, S., 2011, ‘MEGA5: Molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony method’, Molecular Biology and Evolution 28, 2731–2739. http://dx.doi.org/10.1093/molbev/msr121

Toro, H., Van Santen, V.L. & Jackwood, M.W., 2012, ‘Genetic diversity and selection regulates evolution of infectious bronchitis virus’, Avian Diseases 56, 449–455. http://dx.doi.org/10.1637/10072-020212-Review.1

South African Poultry Association, 2012, SAPA Industry Profile, viewed 15 September 2013, from http://www.sapoultry.co.za/industry_profile.php

Wang, L., Xu, Y. & Collisson, E.W., 1997, ‘Experimental confirmation of recombination upstream of the S1 hypervariable region of infectious bronchitis virus’, Virus Research 49, 139–145. http://dx.doi.org/10.1016/S0168-1702(97)01466-4

Xu, C., Zhao, J., Hu, X. & Zhang, G., 2007, ‘Isolation and identification of four infectious bronchitis virus strains in China and analyses of their S1 glycoprotein gene’, Veterinary Microbiology 122, 61–71. http://dx.doi.org/10.1016/j.vetmic.2007.01.006