Abstract
Campylobacter is the most common cause of bacterial gastroenteritis in the developed world, with approximately 70,000 cases reported in the UK per annum. It is well accepted that Campylobacter spp. form biofilms which aid its survival in both the environment and the host. The formation of biofilms in poultry processing plants are of particular concern, as they are potential sources of contamination between meat batches, and facilitate the transmission of the pathogen through the human food chain. However, despite the importance of biofilms, the molecular mechanisms and metabolic pathways associated with biofilm formation in Campylobacter have not been well elucidated. Here, 30 C. jejuni strains were isolated from commercial chicken meat and assayed for their motility and ability to form biofilms, using crystal violet staining at 37ᵒC and 42ᵒC. Only five of the 30 isolates were able to form biofilms, with more complex biofilm phenotypes observed at 37ᵒC. Although all isolates were motile, a weak correlation between motility and biofilm formation was observed, indicating that motility is not essential to the phenotype. Ten isolates were selected, representing the five most competent, and five poorest biofilm formers. These isolates were screened for their virulence profiles using Galleria mellonella and adhesion and invasion of Caco-2 models. No correlation between the ability to form biofilms and virulence phenotypes was observed. A competent biofilm former (isolate CJP13) was selected and a mariner transposon mutant library was constructed in this strain. Over 3,000 of the resulting transposon mutants were individually screened for their ability to form biofilms. Thirteen of the 3,000 transposon mutants showed reduced ability to form biofilms across two independent biological replicates. Of those, individual knock-out mutants of Cj0080, Cj1623 (memP), hydA and trbJ and complemented mutants were constructed in CJP13 and NCTC11168 strains. All mutants showed reduced ability to form biofilms compared to wild type strains, although the NCTC11168 ΔmemP mutant showed the most significant reduction, with almost no biofilm ability observed (p<0.001). Complemented mutants presented a mixed ability to restore biofilm formation to wild type levels. Next-Generation Sequencing (NGS) and subsequent pangenome analysis revealed genes which were differentially present/absent in competent and poor biofilm genomes, two of which are involved with sialic acid synthesis and transport. Phylogenetic analysis revealed CJP17 and CJP19 strains (competent and poor biofilm formers respectively) to be almost genetically identical, with three gene mutations in the CJP17 genome. One such mutation is predicted to cause truncation of pflA, which is suggested to be the cause of reduced motility in this strain compared to CJP19. Despite this mutation, CJP17 displayed a competent biofilm phenotype, suggesting the mechanisms involved in biofilm formation are motility independent. The panel of 10 isolates were subjected to Biolog phenotypic array analysis to study the ability of Campylobacter to metabolise 95 different carbon substrates. Competent biofilm formers were able to significantly metabolise several carbon sources more readily; D-ribose and L-lyxose when using lag phase to define utilisation, and L-lactic acid when using max utilisation and max slope to represent substrate utilisation parameters. However, varying concentrations of L-lactic acid failed to induce biofilm formation in chicken isolates when added to complex media. The studies reported here demonstrate significant differences in the metabolism and genetic composition between poor and competent biofilm isolates. Moreover, this work provides evidence that multiple C. jejuni genes are responsible for the biofilm phenotype in currently circulating C. jejuni isolates. This study suggests that the role of membrane proteins, such as memP, is key in the formation of biofilm in NCTC11168, but less so for CJP13, indicating strain-specific mechanisms for C. jejuni biofilm formation. The continuation of genomic and metabolic analysis of biofilm formation is required to facilitate the development of novel control strategies aimed at mitigating biofilm formation in poultry processing plants, to prevent subsequent human infection.