Comparative phenotypic, proteomic and genomic approaches to assess lipopolysaccharide and outer membrane protein diversity among isolates of Yersinia ruckeri recovered from Atlantic salmon and rainbow trout

Yersinia ruckeri is the causative agent of enteric redmouth (ERM) disease in farmed salmonids. ERM disease is traditionally associated with rainbow trout (Oncorhynchus mykiss¸ Walbaum), but the incidence of the disease in Atlantic salmon (Salmo salar) has increased in recent years. Historically, motile (biotype 1), serotype O1 isolates of Y. ruckeri have been mostly responsible for ERM in rainbow trout worldwide but non-motile (biotype 2), serotype O1 isolates have become increasingly prevalent in this species over wide geographic areas since their first isolation in the UK in the 1980s. Yersinia ruckeri isolates responsible for infection of salmon have been less well characterised than those from rainbow trout and little is known about their diversity. The emergence of new pathogenic strains, together with vaccine breakdown in the field, has emphasised the need for greater knowledge about strain diversity which may lead to the development of improved vaccines for both species. In the present study, a unique and extensive strain collection encompassing 135 isolates of Y. ruckeri were characterised using complementary phenotypic, proteomic and genomic approaches. In the initial part of this thesis, 135 isolates recovered over a 14 year period in the UK from infected Atlantic salmon (109 isolates) and rainbow trout (26 isolates) were phenotypically characterised through biotype, serotype, and outer membrane protein (OMP) -type analysis. Atlantic salmon isolates represented a wider range of O-serotypes and associated lipopolysaccharide (LPS) types, and had more diverse OMP profiles than those from rainbow trout. Most significantly, a new O-serotype/LPS type was identified in 56 Atlantic salmon isolates; other Atlantic salmon isolates were represented by serotypes O1 (five isolates), O2 (34 isolates) and O5 (14 isolates). This new LPS type comprises a core polysaccharide region similar to that of serotype O1 but has a unique, previously unidentified O-antigen region. Atlantic salmon isolates could be assigned to one of four major OMP-types and to one of 11 OMP-sub types. Isolates recovered from rainbow trout were represented by the same non-motile clone that is responsible for the majority of ERM outbreaks in this species within the UK. This clone was not associated with any infected salmon. However, two isolates of the novel serotype O1/O5 were recovered from rainbow trout in 2010 and 2011. These data suggest that different Y. ruckeri strains are specifically adapted to cause disease in either Atlantic salmon or rainbow trout. The efficacy of current vaccine formulations against different clonal groups must be examined. Subsequently, an in-depth characterisation of the outer membrane (OM) proteome of isolates recovered from Atlantic salmon and rainbow trout was conducted. Outer membrane proteins are at the interface of host pathogen interactions, with important roles in adherence, evasion of host immune response, and transport. Using a bioinformatic prediction pipeline and four publicly available genomes, 141 proteins were confidently predicted to be OM associated. Subsequently, the OM proteomes of eight representative isolates (four from rainbow trout; four from Atlantic salmon) were analysed using a combination of gel-based and gel-free proteomic approaches. In total, 66 OMPs were identified through this combined approach, of which 28 were unique to the gel-free approach and 13 were unique to the gel-based approach. Further to this, the OM proteomes of these eight representative isolates were examined when cells were grown under conditions that aimed to mimic the in vivo and environmental conditions of Y. ruckeri. These included growing cells aerobically at 22°C and 28°C, anaerobically, under iron-depletion and in an artificial seawater medium at 22°C. In total, 76 OMPs were identified in all eight isolates under these growth conditions. Finally, a phylogenetic study was undertaken whereby the genomes of 16 representative isolates encompassing a range of biotypes, serotypes, host species (eight from rainbow trout, seven from Atlantic salmon and one from European eel), geographic locations and dates of isolation were considered. A phylogenetic species tree based on the concatenated sequences of 19 housekeeping genes revealed host specific lineages suggesting an earlier host-associated evolutionary split within Y. ruckeri. Subsequent analysis of the presence, absence and variation of the nucleotide and amino acid sequences of the 141 predicted OMPs revealed high levels of conservation (with 120 OMPs showing less than 1% nucleotide variation). One hundred and thirty proteins were identified in all 16 genomes examined. However, 11 proteins were not, and these included invasins, OmpE and proteins involved in pilus biogenesis. Further examination of the OMPs OmpA and OmpF, which were identified in the genomes of all 16 isolates, revealed variation in the surface exposed loop regions which may play a role in pathogenicity and/or host specificity. This study represents a comprehensive characterisation of Y. ruckeri isolates recovered from Atlantic salmon and rainbow trout using a range of molecular techniques, and reveals important adaptations that the bacteria may make in order to survive both inside and outside of the host. Importantly, this study provides comprehensive support for future work involving this fish pathogen