The available Cfv genomic sequence information was aligned to the complete Cff genome sequence 82–40 in order to identify targets for the diagnostics for detecting Cfv. Based on the genome size estimates of Cfv [6, 24] and the completed Cff genome size, it is estimated that approximately 72% of the Cfv genome has been sequenced (unpublished, Prof Daniel Sanchez, Universidad Nacional de San Martin, Argentina). The ordering of available genome segments generally aligned well with the Cff genome as shown in Figure 1 and made evident a suite of Cfv specific contigs. This suite of contigs housed a large range of type IV secretion factors, and plasmid/phage like proteins. A number of potential virulence factors were clearly identified as shared between Cfv and Cff. These virulence factors include an outer surface array membrane protein, chemotaxis types, motility associated, regulatory and secretion systems. The existence of these classes of genes with roles in the infection process, but not showing sub species specificity, is consistent with a two-tier infection model. Surface/membrane components provide necessary (but not sufficient) structural components for attachment to host cells. Specific components that complete the features of the surface/membrane structures are required for infection. Fouts et al., (2005) found that many genes involved in host colonization were conserved across the Campylobacter genus. Variations that were species specific were evident for a lipo-oligosaccharide locus, a capsular (extracellular) polysaccharide locus, and a novel Campylobacter putative licABCD virulence locus (not found in available Cfv). These observations are consistent with the suggestions that interactions between a pathogen's surface-exposed proteins and host cells represent a pivotal step in pathogenesis and virulence . In pathogens several of the key players are proteins involved in adhesion, invasion, secretion, signalling, annulling host responses, toxicity, motility and lipoproteins .
Motility and chemotaxis genes have been found conserved among related Campylobacter species with flagella implicated in adhesion, protein secretion, invasion and virulence in pathogenic C. jejuni [1, 27–30]. Biosynthesis of flagella requires the involvement of more than 40 structural and regulatory proteins including a type III secretion system for flagellar assembly [28, 30–32]. The Cff flhA gene based on genome alignments was found to be absent in the available Cfv sequence contigs, and coincided with the ordered alignment gap/non-sequenced section relative to Cff. However, one chemotaxis regulatory protein campy.fasta.screen.Contig1091 orf6 appears to be absent in Cff (Additional file 1). We identified a lower complement of homologues associated with motility in Cff (n = 41) compared with the other Campylobacter spp. (n = 55–66) , however, the analysis of the incomplete Cfv genome identified a higher number of homologues (n = 46) than the total Cff sequence. PCR assays based on a subset of flagellar genes (flgH, flhF, fliH, flhA and fhlB), demonstrated conservation of these sequences at least among the members of our panel of C. fetus strains including both subspecies (although flhA could not be identified in the available Cfv contigs). An additional assay designed to amplify the flaB sequence of the Cfv AZUL-94 strain did not amplify other Cfv biovar venerealis strains but did amplify Cfv intermedius and the Cff isolates. We have not confirmed if this is attributed to flaB sequence variation or an absence of the gene in different geographical Cfv biovar venerealis strains, this gene has been targeted however for genotyping studies in other Campylobacter species . This study does confirm that the complete Cfv genome may harbour more flagellar/motility homologues than Cff. Virulent C. jejuni harbours more flagellar genes than less virulent species C. coli, C. lari and C. upsaliensis .
Adherence of other Campylobacter species to gut epithelial cells is mediated by multiple adhesins including cadF (C
adhesion to fibronectin); , PEB1 protein (putative binding component of an ABC transporter), , JlpA (jejuni lipoprotein A),  and a 43-kDa major outer membrane protein , confirmed as conserved in C. jejuni, C. lari, C. upsaliensis and C. coli genomes . Cfv homologues for PEB1 and fibronectin-binding (FN-binding) proteins were confirmed with the remaining 3 absent in the genome contigs currently available. However, only the PEB1 protein was identified in the complete Cff genome sequence 82–40. Fibronectin is known to enhance C. fetus attachment  however in the absence of an identified C. fetus cadF homologue, it appears that the adherence mechanisms in C. fetus may differ from other Campylobacter species. In the case of C. fetus subsp. venerealis, this is perhaps not surprising as Cfv colonise the genital tract and not the intestinal tract, thus perhaps novel adhesins will be identified with completion of a Cfv genome sequence.
Toxin sequences, two component regulatory systems, plasmids and type IV secretion systems have also been recognised as components in pathogenic Campylobacter spp. . Three cytolethal distending toxin (cdt) subunits A, B and C are confirmed as conserved across the four Campylobacter species (C. jejuni, C.lari, C. coli, C. upsaliensis) and C. fetus [22, 23]. In addition, the presence of cdt genes is linked to C. jejuni, C coli and C. fetus pathogenesis, where cdt negative strains were found to be less efficient during adherence and invasion in vitro [22, 39]. A similar survey of C. fetus will assist to confirm if cdt positivity is associated with an increase in pathogenicity. Two-component regulatory (TCR) systems are commonly used by bacteria to respond to specific environmental signals such as temperature . Five TCR systems (pairs of adjacent histidine kinase and response regulator genes) have been identified as conserved across Campylobacter species and confirmed in C. fetus subspecies.
The type IV secretory genes, which are possibly involved in conjugative plasmid transfer or the secretion of virulence factors [1, 18, 41], were absent in the Cff genome and unique to Cfv. A large proportion of Cfv subspecies specific ORFs (30%) were harboured in the Cfv contig specific regions. C. upsaliensis and C. jejuni are known to harbour plasmids and evidence does suggest that these plasmids can play a role in pathogenesis. One basic difference between the list of genes absent in Cff and present in Cfv is that many of them are in common to genes present on the plasmids of these related Campylobacter. The type IV secretion system is also found in C. jejuni, C. lari and C. coli plasmid sequence. The unique Cfv genome sequences also harboured many phage-like derived genes. The presence of type IV secretion system has also been described by Abril et al, 2007 , of which the putative VirB6 protein gene was found to be truncated by the insertion element (IScfe1). It is possible that contigs within this Cfv unique 80 Kb suite of contigs represent a number of extrachromosomal DNA plasmids. A wider survey of C. fetus isolates and the presence of plasmids (type IV secretion systems) and phage genes will assist to confirm our observations.
This analysis has provided diagnostic markers to discriminate the Campylobacter subspecies Cfv and Cff, which can be investigated for more general applicability for field use. Most of the Cfv assays based on the incomplete AZUL-94 genome sequence, showed amplification preference for Cfv biovar venerealis strains. The Cfv biovar intermedius strains were negative in all but one assay, which was otherwise positive for Cfv AZUL-94 strain only. Curiously, one of the assays designed to Cfv AZUL-94 strain virB9 (type IV Secretion gene) did not amplify other Cfv biovar venerealis isolates but did amplify biovar intermedius and the Cff strains tested here. However, as described above the Cff genome sequence (Strain 82–40) does not appear to have type IV secretion genes. A confounding factor in interpreting this data is that different Cff strains may also possess putative plasmid-borne genes and these may potentially be shared between subspecies and Cfv biovars. The Cfv AZUL-94 strain could also either consist of a mix of the 2 biovars or represent a novel strain of Cfv. However, assays based on putative plasmid-borne genes have previously demonstrated inconsistencies when applied for subspecies identification in some regions . The parA (plasmid partitioning protein gene),  assay target is thought to be plasmid borne, however evidence for plasmids containing parA in Cfv has not been confirmed to date [19, 42]. Very little research has been undertaken to compare the Cfv biovars and the diagnostic targets reported here now need to be further tested in multiple field strains to assess the stability of these markers and therefore the genomic regions in Cfv. However, the results presented do suggest that the Cfv research community could benefit from the generation of full genome sequence from both biovars as well as isolates from different geographical continents. Our results also demonstrated putative plasmid sequences are present in Cfv, absent in Cff, suggesting plasmid profiling and sequencing from C. fetus subspecies, biovars and strains will assist to confirm our findings.