Cointegration as a mechanism for the evolution of a KPC-producing multidrug resistance plasmid in Proteus mirabilis

The incidence and transmission of Klebsiella pneumoniae carbapenemase (KPC) producing plasmids have been well documented. However, the evolutionary dynamics of KPC plasmids and their fitness costs are not well characterized. Here, two carbapenemase-producing plasmids from Proteus mirabilis, pT18 and pT211 (both carrying bla(KPC-2)), were characterized through whole genome sequencing. pT211 is a 24.2 kbp N-type plasmid that contains bla(KPC-2) and a single copy of the IS6-family insertion sequence IS26. pT18 is a 59 kbp cointegrate plasmid comprised of sequences derived from three different plasmids: a close relative of pT211 (containing bla(KPC-2)), an FII-33 plasmid (bla(TEM-1B), bla(CTX-M-65), rmtB and fosA3) and a rolling-circle plasmid. The segments of pT18 derived from each of the different plasmids are separated by copies of IS26, and sequence analysis indicated that pT18 was likely generated by both conservative and replicative IS26-mediated cointegrate formation. pT18 and pT211 were transferred into Escherichia coli DH5α separately to assess the impact of plasmids on host fitness. Only DH5α harbouring pT18 grew slower than the wild type in antibiotic-free media. However, in sub-inhibitory concentrations of fosfomycin and amikacin, cells containing pT18 grew faster than the wild type, and the minimum concentrations of fosfomycin and amikacin required to observe an advantage for plasmid-carrying cells were 1/3 and 1/20 the DH5α MIC, respectively. This study highlights the importance of the role of cointegrate plasmids in the dissemination of antibiotic resistance genes between pathogenic bacterial species, and highlights the importance of sub-inhibitory concentrations of antibiotics to the persistence of such plasmids.