Molecular diagnostic protocols for bovine genital campylobacteriosis using comparative genomics and virulence studies

Abstract

Bovine genital campylobacteriosis is difficult to identify due to problems with the isolation of its causative agent Campylobacter fetus subspecies venerealis. To optimise isolation methods, three selective culture and three transport and enrichment media (TEMs) were evaluated using a pure C. fetus subsp. venerealis culture, with and without simulated field contamination by Pseudomonas aeruginosa. For optimal C. fetus subsp. venerealis isolation and growth, TEMs should be returned from the field within 48h of collection and plated within 4h onto Campylobacter selective agar plates incubated for 48-72h at 37°C under microaerophilic conditions. Longer transport or TEM incubation resulted in overgrowth of P. aeruginosa limiting C. fetus subsp. venerealis recovery. In order to establish an isolate collection, the parA real-time PCR assay was used to screen bull preputial secretions prior to culturing positive samples. A more sensitive real-time PCR assay was designed from the ISCfe1 insertion element and used to assess additional preputial samples, compared to the parA real-time assay. A proportion of real-time negative samples was also cultured. Two conventional PCR assays, biochemical testing and sequencing of late real-time PCR products were applied to cultures. The parA real-time PCR detected 174 positives out of 717 samples with 44 Campylobacter-like organisms isolated (25.3%). Culture of 94 parA-negative and 40 untested samples yielded 9 Campylobacter-like isolates. ParA and ISCfe1 real-time PCR analysis of a further 266 samples facilitated the isolation of 17 Campylobacter-like cultures from 70 positives (24.3%). Fourteen cultures were obtained from either ISCfe1 (8/43; 18.6%) or parA (6/51; 11.8%) negative samples. ISCfe1 and parA sequences were obtained from isolates not confirmed by conventional methods suggesting that both parA and ISCfe1 may be found in organisms other than C. fetus subsp. venerealis. Molecular testing of cultures (n=84) confirmed the presence of these sequences in isolates identified as other Campylobacter spp. or Arcobacter-like organisms, while they were not always detected in all biochemically confirmed C. fetus subsp. venerealis isolates. The lack of specificity of all methods highlights the need for more reliable C. fetus subsp. venerealis-specific gene targets. The draft genomes of two C. fetus subsp. venerealis biovars aligned to C. fetus subsp. fetus genome provided 26 unique C. fetus subsp. venerealis regions. In order to identify new diagnostic targets, 13 putative conventional and one real-time PCR assays designed from contigs common to both C. fetus subsp. venerealis biovars were developed to screen the Campylobacter-like culture collection (n=84). Isolates biochemically identified as organisms other than C. fetus subsp. venerealis were positive on the new real-time and six conventional PCR assays, indicating that they were not suitable subspecies-specific markers. Seven assays detected isolates phenotyped as C. fetus subsp. venerealis except for two aerotolerant isolates recognized as C. fetus subsp. venerealis by other published methods. These seven gene targets could be further investigated for real-time PCR assay development as in silico BLAST analysis confirmed their conservation in the genomes of five of our C. fetus subsp. venerealis isolates. However, further bioinformatics analysis revealed that they were found in several C. fetus subsp. fetus strains, underlining that a pan-genomics study including multiple C. fetus genomes is needed to identify new C. fetus subsp. venerealis-specific core genes. Variation in putative virulence, type IV secretion system (T4S) and pathogenicity island (PAI) genes observed from the C. fetus subsp. venerealis biovars supported the hypothesis that differences in virulence may exist between strains. Nine T4S and PAI conventional PCR assays were used to characterize the Campylobacter-like isolate collection (n=84) and other field cultures (n=21). New putative virulence (n=11) and PAI (n=4) conventional PCR assays were designed from unique C. fetus subsp. venerealis regions, and also assisted in predicting the virulence status of each isolate. In silico analysis of seven T4S gene sequences from C. fetus field and reference strains was performed to investigate if previously reported amino acid consensus sequences were conserved. A range of isolates were positive on different assays, including C. fetus subsp. fetus, other Campylobacter spp. and Arcobacter-like organisms, underlining a shared virulence and PAI gene content between closely related bacteria. Results contributed to the selection of C. fetus subsp. venerealis candidates for testing in a parallel pregnant guinea pig model, but differences in virulence between strains in vivo did not correlate with differences in virulence and PAI gene profiles. Even though T4S consensus sequences were conserved for six of the seven genes examined, observed sequence differences may also influence the amino acid sequence of encoded proteins and thus C. fetus virulence in vivo. Whilst the genes assessed here may be involved in C. fetus subsp. venerealis virulence, other factors and potential interactions as well as differences in expression during animal infection are yet to be examined. A pan-genomic study including C. fetus and Campylobacter-like strains isolated here, in correlation with further animal and virulence experiments, will contribute to improving diagnostic methods for bovine genital campylobacteriosis.