We report the complete sequence of a blaCTX-M-14
–carrying IncK plasmid, pCT, from an E. coli
isolate from a cattle farm in the United Kingdom (16
). Within the 115 putative CDs, there is an absence of any genes known to play a role in determining virulence of the host and the absence of any other antimicrobial drug resistance genes except for blaCTX-M-14
. Therefore, the persistence and spread of pCT cannot be attributed to coselection associated with pressure from non–β-lactam antimicrobial drugs. This finding suggests that pCT persistence and dissemination have been driven by either constant β-lactam exposure or that pCT can remain stable within a population in the absence of any antimicrobial drug selective pressure. There has been much speculation about the role of the type IV pilus and its shufflon found in plasmids and possible role(s) in adhesion of the host bacterium to surfaces and eukaryotic epithelial cells in vitro and in biofilm formation (34
). The thin pilus might also aid conjugation in a liquid environment by anchoring donor and recipient cells (35
). This system may have played a role in the persistence of E. coli
C159/11, which originally was found to remain within slurry and on the floor of cow sheds during the longitudinal farm study from which it was identified (16
). These attributes of the type IV pilus might contribute to the persistence of pCT within bacteria isolated from the UK farm and in animals and humans throughout the world.
Two other features of pCT are of interest. The first, and most notable, is a putative RNA polymerase sigma factor (CDS pCT_066) within the sequence, which shares homology with genes found in only 4 closely related plasmids (pO26vir; pO113, TP113, and pSERB1) and has limited identity to homologue SigB in Yersinia frederiksenii. Other weak protein matches show some homology to the extracytoplasmic function sigma factors, small regulatory proteins divergent in sequence to most of the other sigma factors and involved in global gene regulation. Both examples are chromosomally encoded. Although sigma factors of this group have previously been noted on plasmids, scant information has been published about their role or function.
The second feature of interest is that in large stably maintained conjugative plasmids, such as R64, functional large addiction operons such as ParA/B or kor/mck usually are present; however, these are lacking in pCT. Despite the apparent lack of stability or persistence genes, pCT has remained stable in a population in the absence of selective pressure for prolonged periods (N.G. Coldham et al., unpub. data).
Comparison of the genome of pCT with other blaCTX-M
–encoding plasmids showed no conserved regions outside the β-lactamase gene. Therefore, no single feature of the plasmid backbone appears responsible for the spread of blaCTX-M
genes, and the acquisition of these genes is unlikely to have been a single event. Homology was highest between pCT and 4 plasmids (pO26_vir, pO113, pSERB1, and TP113). pO26_vir was identified in a Shiga toxin–producing E. coli
strain 026:HII and encodes several virulence genes not found on pCT, including genes for the production of a hydrolase, catalase, and a hemolysin transport protein. pO113 was isolated from another hemolysin-producing EHEC O113:H21 E. coli
sample from a patient in Australia (36
). The finding that pCT is most closely related to 2 plasmids that carry an array of virulence genes is of concern because of the potential for recombination between these plasmids, creating mobile elements carrying virulence genes and the blaCTX-M-14
The genome sequencing of pCT enabled development of PCRs that amplified discrete regions of the pCT sequence, thereby enabling rapid identification of other pCT-like plasmids that share these loci. pCT-like plasmids were identified in bacteria isolated from 2 other UK farms in 2006 and 2008 and, most recently, from human clinical isolates in the United Kingdom (M. Stokes et al., pers. comm.).
Four human clinical isolates from Spain also carried pCT-like plasmids, with all 6 pCT regions amplified by PCR, which had the same insertion sites for blaCTX-M-14
. These data show the ability of a large conjugative plasmid to transfer between bacteria isolated from humans and animals, facilitating the movement of blaCTX-M-14
between these niches. Since 2000, when CTX-M-14 was identified in bacteria from Spain, it has become one of the most commonly detected enzymes isolated from human and animal isolates in Spain (24,37
). Previous studies conducted in hospitals in Spain examined an association between blaCTX-M-14
and IncK plasmids. Valverde et al. (11
) isolated an IncK plasmid, pRYC105, from many lineages of E. coli
from community-acquired infections and the environment in different geographic regions of Spain. These authors hypothesized that pRYC105 shared identity with the plasmid isolated in the United Kingdom by Liebana et al. (16
), and the present study has confirmed this hypothesis by showing that pRYC105 is pCT-like.
Human clinical isolate E. coli
8 CH13, described in 2002 and isolated in 1998 from China, contained pOZ174, which encodes blaCTX-M-14
); as with pRYC105, we showed that pOZ174 is pCT like. Furthermore, our data suggest that pCT has persisted since 1998 and is distributed across Europe, Asia, and Australia in diverse E. coli
lineages isolated from humans and animals. Because CTX-M-14 is the most frequently identified ESBL in Spain and China, further investigation using this molecular test will determine whether pCT is the dominant vector of blaCTX-M-14
in these areas and whether pCT has disseminated to other ecosystems. The identical insertion site for blaCTX-M-14
in each of the pCT-like plasmids investigated in our study suggests a single capture of this β-lactamase gene onto the plasmid backbone and subsequent spread of the plasmid.
The alignment and analysis of nikB from pCT-like plasmids were also used to determine how related the plasmids are and demonstrated sequence identity of >98%. These sequences clustered with pCT within a phylogenetic tree, which indicated less sequence divergence than with other IncI complex non–pCT-like plasmids. Design of the pCT–specific PCRs distributed throughout the plasmid and sequencing of nikB amplicons provided a useful and rapid tool in first identifying pCT-like plasmids. Relaxase or nikB typing also would provide a suitable locus in recently developed plasmid multilocus sequence typing. These assays can now be used to screen CTX-M-14–producing bacteria for other pCT-like plasmids.
The sequence of pCT enabled an understanding of its backbone and seems to suggest that, apart from plasmid replication and transfer functions, the only known gene that confers a selective advantage on this plasmid is blaCTX-M-14. Subsequent PCRs successfully indicated that pCT-like plasmids are distributed over several countries worldwide in bacteria isolated from humans and animals. This approach can be applied to the study of other plasmids of clinical relevance and facilitate better trace analyses of horizontally acquired antimicrobial drug resistance or virulence genes. Additionally, use of this method may lead to identification of new vectors and increase understanding of the interaction among bacteria isolated from humans, animals, and the environment.