Natural variation of large plasmids in bacterial populations

Abstract

Plasmids encode an extraordinary range of adaptive functions, including antibiotic resistances, virulence factors and degradative enzymes. Plasmids also mediate horizontal gene transfer, which impacts evolution of bacterial chromosomes. Despite their important ecological and evolutionary roles, we lack a comprehensive picture of the variation and evolution of plasmids themselves. In this dissertation, I investigated the natural variation of large plasmids on two different scales: 1. individual plasmid sequences from Gram-negative and -positive bacteria and 2. total plasmid content of E. coli and Salmonella reference collections.

For the sequence-based studies, I sequenced four large plasmids from E. coli, Staphylococcus and Corynebacterium. Whereas 98% of the E. coli plasmid was highly similar to other plasmid sequences, 40-60% of the three Gram-positive plasmids had no apparent similarity to known sequences. The E. coli plasmid’s backbone, which encodes “core” functions such as replication, was almost identical to that of a Salmonella enterica sv Typhimurium plasmid. Subsequent analysis identified this backbone in five other enterobacterial plasmids. All seven plasmids differed in accessory gene content, leading to variation in plasmid-encoded phenotypes.

For the collection-based studies, I analyzed 228 E. coli and Salmonella strains from four reference collections. Plasmids, especially large plasmids, were abundant. Replicon typing identified two common families, IncF and IncI1, by the presence of characteristic backbone genes. At least 20% of strains with large plasmids in each collection were untypable, suggesting variation in plasmid backbones. Restriction fragment length polymorphism (RFLP) analysis showed high variation in large plasmid genomes. The majority of E. coli and Salmonella large plasmids (87.5% and 56%, respectively) had unique RFLP patterns. The lack of widespread RFLP patterns suggested that the prevalence of IncF and IncI1 is due to similar backbone genes on otherwise different plasmids. The only exception was the IncFII virulence plasmid pSLT, which was detected in multiple Typhimurium strains from different hosts and geographic locations, suggesting that virulence plasmids evolve differently from other plasmids.

These results show that plasmid evolution is a significant force shaping the horizontal gene pool. Rather than existing as static elements, plasmids are dynamic and flexible genetic scaffolds driving gene flux in host bacteria.