|Home | About | Journals | Submit | Contact Us | Français|
We evaluated the in vitro activity of fosfomycin against a total of 192 CTX-M β-lactamase-producing Escherichia coli strains isolated in 70 Japanese clinical settings. Most of the isolates (96.4%) were found to be susceptible to fosfomycin. On the other hand, some of the resistant isolates were confirmed to harbor the novel transferable fosfomycin resistance determinants named FosA3 and FosC2, which efficaciously inactivate fosfomycin through glutathione S-transferase activity.
Clinical efficacy of an old antibiotic, fosfomycin, is being reassessed owing to its in vitro high potent activity against multidrug-resistant Gram-negative bacilli belonging to the family Enterobacteriaceae (5, 6, 8). In the present study, we investigated the prevalence of fosfomycin resistance among CTX-M extended-spectrum β-lactamase (ESBL)-producing Escherichia coli clinical isolates in Japan and clarified the molecular mechanisms underlying the fosfomycin resistance, with special focus on exogenous resistance determinants, like the FosATN protein (9).
A total of 192 CTX-M ESBL-producing E. coli isolates, which were collected from 70 medical facilities throughout Japan between 2002 and 2007, were retrospectively subjected to fosfomycin susceptibility testing with the agar dilution method according to the CLSI guideline (4). The result is shown in Fig. Fig.1.1. Most of the strains (96.4%) investigated were susceptible to fosfomycin (MIC, ≤64 μg/ml), while seven isolates (3.6%) showed nonsusceptibility to fosfomycin (MIC, ≥128 μg/ml). It seems likely that CTX-M-producing E. coli isolates that have acquired fosfomycin resistance are infrequent in Japan, and these data suggest the probable clinical efficacy of fosfomycin for the treatment of infectious diseases, like urinary tract infections (UTIs), caused by CTX-M-producing E. coli to some extent.
We evaluated the fosfomycin resistance mechanism of the 10 isolates and found reduced susceptibility to fosfomycin (MIC, ≥64 μg/ml) (Fig. (Fig.11 and Table Table1).1). The transmissibility of the fosfomycin resistance determinant in the 10 isolates was investigated, and it was found that the nature of fosfomycin resistance of three strains, 08-642, 06-607, and C316, was successfully transferred to a recipient E. coli strain. The cefotaxime resistance phenotype was cotransferred to a recipient strain with the fosfomycin resistance (Table (Table11).
The DNA fragments containing fosfomycin resistance determinants were cloned from the conjugative plasmids of E. coli 08-642, 06-607, and C316 strains and partially sequenced (Table (Table1).1). The fosfomycin resistance determinants and their genetic neighboring regions are shown in Fig. Fig.2.2. The KpnI ca. 8-kb fragment cloned from the transferable plasmid of E. coli 08-642 and the SacII ca. 10-kb fragment from that of E. coli 06-607 included the same nucleotide region flanked by IS26 (Fig. (Fig.2).2). The deduced amino acid sequences of one open reading frame (named fosA3) showed 70% identities to those of FosATN, the Mn(II)- and K+-dependent glutathione (GSH) S-transferase from Tn2921 of Serratia marcescens (2, 3) and 59% identities to that of FosAPA from Pseudomonas aeruginosa (Fig. (Fig.3)3) (1, 11). The fosA3 gene is likely to be responsible for fosfomycin resistance in strains 08-642 and 06-607.
The 1.8-kb region containing orf1 to Δorf3 at the 3′ end of fosA3 had 78% nucleotide identity with a part of the chromosome sequence of Klebsiella pneumoniae strain 342 (Fig. (Fig.2)2) (7). Moreover, this 1.8-kb homology region on the chromosome of K. pneumoniae strain 342 was close to the fosA gene. FosA of K. pneumoniae strain 342 has 80% amino acid identity to the FosA3 found in the present study. Although the precise physiological function of chromosomally encoded FosA proteins of K. pneumoniae remains to be determined, it is speculated that these proteins are the origin of a plasmid-mediated fosfomycin-modifying enzyme like FosA3.
Additionally, one open reading frame, named fosC2, was found in the fragment cloned from the conjugative plasmid of E. coli strain C316 (Table (Table1)1) (13). The amino acid sequence of FosC2 had 72%, 56%, and 51% identity to that of FosC found in Achromobacter xylosoxidans (GenBank accession no. DQ112222), FosATN, and FosAPA, respectively (Fig. (Fig.3).3). The fosC2 gene was the first gene cassette in a class 1 integron accompanied by dfrA17 and aadA5 (Fig. (Fig.22).
No transfer of fosfomycin resistance determinants was observed in the seven E. coli strains showing reduced susceptibility to fosfomycin (MIC, ≥64 μg/ml) (Table (Table1).1). Next, already-known chromosomally derived genes glpT, uhpT, uhpA, and murA, which are involved in fosfomycin resistance, were investigated (Table (Table1)1) (10, 12). The primers used in the present study are listed in Table Table2.2. Several outcomes supposed to be involved in fosfomycin resistance were observed in six of the strains, but no remarkable change was detected in strain 03-285 among the investigated genes. Although the extent to which the fosfomycin resistance conferred by chromosomally encoded factors described above remains controversial, these factors would partially explain the fosfomycin resistance in the clinical isolates.
Finally, we purified C-terminal histidine-tagged FosA3 and FosC2 with the combination of a pET29a vector and E. coli BL21(DE3)(pLysS) and determined the enzymatic characteristics through a bioassay. Assays were performed in 50 mM HEPES buffer (pH 7.8) containing 100 mM KCl, 0.05 mM MnCl2, 5 mM fosfomycin, 10 mM glutathione, and 10 μM purified protein in a final volume of 100 μl at 35°C for 30 min, and the reaction was quenched by adding methanol. Ten microliters of sample solution was added to a blank disc set on an agar plate inoculated with E. coli ATCC 25922, and the remaining antibacterial activity was measured as a growth inhibition zone. When a sample solution containing only fosfomycin and GSH was added, a 21-mm inhibitory zone was observed. When the same sample supplemented with FosA3 or FosC2 was added, no inhibitory zone was observed around the disc. No decrease in the growth inhibition zone was observed when the sample containing only fosfomycin and purified proteins were added. The consumption of GSH catalyzed by FosA3 and FosC2 was confirmed using Ellman's reagent. These results indicated that FosA3 and FosC2 inactivated fosfomycin by exerting GSH S-transferase activity, very similar to FosATN and FosAPA (1, 3).
In conclusion, we report here the prevalence of fosfomycin resistance among CTX-M-producing E. coli isolates in Japan, together with the emergence of two novel plasmid-borne fosfomycin-modifying enzymes, FosA3 and FosC2. The fosfomycin resistance rate in CTX-M-producing E. coli is still low (3.6%) in Japan, but the fosfomycin resistance genes were already indwelling in the transferable plasmid of ESBL-producing clinical isolates. Continuous monitoring will be necessary to prevent further dissemination of fosfomycin resistance genes, together with prudent use of fosfomycin in clinical settings.
The nucleotide sequences for fosA3 and fosC2 have been deposited in GenBank under accession numbers AB522970 and AB522969, respectively.
This study was supported by the Ministry of Health, Labor, and Welfare of Japan (grant no. H21-Shinkou-Ippan-008).
Published ahead of print on 19 April 2010.