MNP Ecology Center

Bacteriocin Diversity

2.3.1 Colicins

Surveys of colicin diversity in different collections of E. coli all give rise to similar results - only a small fraction of the known colicins are present in a given collection and, in general, different colicins are detected in different collections (Riley and Gordon 1996; Table 2.1). There are a few exceptions to this general trend - colicins E1 and Ia are often observed (Riley and Gordon 1996; Table 2.1). Colicin Ia is encoded on a conjugative plasmid and therefore is able to transfer among E. coli lineages, so the fact that it is one of the more commonly observed colicins could be expected. However, colicin B is also borne on a conjugative plasmid and is the most common colicin produced by the isolates from mammals, yet it is uncommon in human isolates (Table 2.1).

The available data also suggest that a single colicin type dominates in any particular population of E. coli. This observation is expected, based on our understanding of the dynamics of colicin-producing, resistant and sensitive cells. As the frequency of cells resistant to the dominant colicin type in a population increases, that of the dominant producer population will decline, thereby providing the opportunity for the invasion of a different type of producer to which the fraction resistant to the original colicin is susceptible. Thus, theory predicts that there should be a continual flux in the relative frequency of different colicin types in a population of E. coli. There is some

"Identity of the colicin produced was not determined

"Identity of the colicin produced was not determined evidence to support this prediction. Colicin E2 is rare among strains in the collection of E. coli recovered from Australian mammals (Table 2.1). However, in a study of E. coli isolated from a single Australian population of house mice, 25% of the colicinogenic strains produced colicin E2. Over a 7-month period, there was a significant decline in the frequency of E2 and a concomitant increase in the frequency of resistance to E2 (Gordon et al. 1998). Additional evidence comes from the distribution of colicin D. Overall, colicin D is rare in E. coli isolated in Australia (Table 2.1) but, among mammals, colicin D was detected only from a single population of mountain brush-tailed possum (Trichosurus canis).

The ecological niche of a bacterial strain may also play a role in determining bacteriocin diversity. Two new bacteriocins, alvecin A and B (Wertz and Riley 2004), have recently been described from genomic species 2 of the genus Hafnia. PCR screening for the presence of alvecins A and B revealed that these bacteriocins were most frequently produced by strains isolated from mammals but were not detected in bacteriocinogenic strains isolated from reptiles (unpublished data). The bacteriocinogenic isolates from reptiles appear to produce several novel, as yet uncharacterised bacteriocins which appear to be absent in isolates from mammals (unpublished data).

The frequency of bacteriocin production in E. coli varies depending on the genetic group membership of the producing strains, as does the type of bacteriocin produced by a strain. In the collection of isolates from mammals, colicin Ia is significantly more prevalent among group B2 strains (20%), less common in B1 strains (8%), uncommon in D strains (4%), and absent in group A strains. In the collection of isolates from humans, however, the frequency of colicin Ia producers is independent of a strain's E. coli group membership. In both the human and mammal E. coli collections, colicin E1 is significantly co-associated with K1, and a strain which is K1 positive is five times more likely to harbour the colicin E1 plasmid than if it is not.

2.3.2 Microcins

All of the seven microcins screened for were detected in E. coli isolated from humans and all, but one, were detected in the isolates from mammals (Table 2.2). As was the case for colicins, the distribution of a particular microcin varies based upon a strain's group membership. Microcin V is encoded on a conjugative plasmid and, in human isolates, its prevalence does not vary with a strain's E. coli group membership. However, V is not randomly distributed among E. coli genotypes. Thus, among the group B2 strains isolated from humans, microcin V is never detected in a strain encoding for either of the adhesins focG or iha, the toxin hylA, or the secreted protein she. Therefore, microcin V is absent from the 69% of the B2 strains which possess one or more of these traits but present in the 30% of the B2 strains which possess none of these traits.