Practical Agar-Based Disk Potentiation Test for Detection of Fosfomycin-Nonsusceptible Escherichia coli Clinical Isolates Producing Glutathione S-Transferases

Abstract

The number of reports concerning Escherichia coli clinical isolates that produce glutathione S-transferases responsible for fosfomycin resistance (FR-GSTs) has been increasing. We have developed a disk-based potentiation test in which FR-GST producers expand the growth inhibition zone around a Kirby-Bauer disk containing fosfomycin in combination with sodium phosphonoformate (PPF). PPF, an analog of fosfomycin, is a transition-state inhibitor of FosA^PA, a type of FR-GST from Pseudomonas aeruginosa. Considering its mechanism of action, PPF was expected to inhibit a variety of FR-GSTs. In the presence of PPF, zone enlargement around the disk containing fosfomycin was observed for FosA3-, FosA4-, and FosC2-producing E. coli clinical isolates. Moreover, the growth inhibition zone was remarkably enlarged when the Mueller-Hinton (MH) agar plate contained 25 μg/ml glucose-6-phosphate (G6P). When we retrospectively tested 12 fosfomycin-resistant (MIC, ≥256 μg/ml) E. coli clinical isolates from our hospital with the potentiation test, 6 FR-GST producers were positive phenotypically by potentiation disk and were positive for FR-GST genes: 5 harbored fosA3 and 1 harbored fosA4. To identify the production of FR-GSTs, we set the provisional cutoff value, 5-mm enlargement, by adding PPF to a fosfomycin disk on the MH agar plates containing G6P. Our disk-based potentiation test reliably identifies FR-GST producers and can be performed easily; therefore, it will be advantageous in epidemiological surveys and infection control of fosfomycin-resistant bacteria in clinical settings.

INTRODUCTION

The increased emergence of multidrug-resistant (MDR) pathogenic bacteria is becoming a serious public health concern, as MDR bacteria limit the choice of antimicrobials available for treatment (1). In such a situation, “old” antimicrobials, such as colistin and fosfomycin, have been reintroduced into clinical practice to overcome the difficulties posed by MDR pathogens (2). However, some clinically isolated bacteria have already developed resistance to these reintroduced antimicrobials (3, 4). We have previously demonstrated that CTX-M-type extended-spectrum β-lactamase (ESBL)-producing Escherichia coli isolates have already acquired resistance to fosfomycin, although the incidence of resistance is still low (3.6%) (5). In addition, we revealed that some ESBL-producing fosfomycin-resistant E. coli clinical isolates can inactivate fosfomycin by producing glutathione S-transferases (GSTs), such as FosA3 and FosC2 (Fig. 1) (5). These fosfomycin resistance determinants were located on transferable plasmids and were mostly linked with CTX-M-type ESBL genes (5, 6).

FIG 1.

Predicted amino acid sequence of fosfomycin resistance determinants. *, amino acid residues conserved among the eight fosfomycin resistance determinants; colons and dots, amino acid substitutions that result in homologous amino acid residues. The resistance determinants (GenBank accession no.) are FosA^PA (AAT49669), FosA^TN (AAA98399), FosA2 (ACC85616), FosA3 (BAJ10054), FosA4 (AB908992), open reading frame 1 (ORF1) (AAP50248), FosC (AAZ14834), and FosC2 (BAJ10053). The box indicates the conserved Thr9 residue in the fosfomycin resistance glutathione S-transferases.

Since our first report of fosfomycin-resistant E. coli organisms producing FosA3 in 2010, reports from East Asian countries have identified many FosA3 producers in E. coli isolates from clinical specimens (7), healthy individuals (6), consumable animal products (8, 9), and domestic animals (10). Further spread of the gene encoding FosA3 would be a serious public health concern because of the global distribution of CTX-M-type ESBL producers in a variety of settings (11).

To identify FosA3-producing E. coli organisms and prevent their further spread, it is important to develop a specific detection method. PCR is the most common technique used to detect specific antibiotic resistance determinants, but its availability is generally limited to highly advanced facilities, such as research laboratories and university hospitals. Hence, a simple yet cost-effective method, such as the ESBL confirmation test, which uses commercially available antibiotic disks in combination with a potent inhibitor, would be a preferred tool for screening bacteria with specific antibiotic resistance mechanisms in clinical microbiology laboratories (12). The aim of this study was to develop a simple and cost-effective detection method based on the standard disk diffusion test in order to identify E. coli isolates producing fosfomycin resistance-mediating glutathione S-transferases (FR-GSTs), such as FosA3.

MATERIALS AND METHODS

Bacterial strains.

A total of 25 E. coli clinical isolates from our laboratory stock, including seven fosA3-positive isolates, one fosC2-positive isolate, one fosA4-positive isolate, and 16 isolates without any of these three genes, were used to collect basic data to develop the disk-based potentiation test described below. All the isolates were nonsusceptible to fosfomycin (MIC, ≥128 μg/ml). Detailed characterizations were reported for some of these isolates in our previous studies (5, 6).

PCR.

The presence of the FR-GST genes, fosA3, fosA4, and fosC2, was confirmed by PCR using the primer sets described in Table 1. The detailed genotypes of fosA3/fosA4 were determined by PCR and nucleotide sequencing analyses using primers designed on the basis of the sequences deposited in GenBank (accession no. AB522970) (Table 1). The PCR conditions were 30 cycles of 98°C for 10 s, 55°C for 30 s, and 72°C for 1 min.

TABLE 1.

Primers used in this study

Primer	Sequence
fosA3/4-forward	5′-TGA ATC ATC TGA CGC TGG-3′
fosA3/4-reverse	5′-TCA ATC AAA AAA GAC CAT C-3′
fosC2-forward	5′-CGT TCC GTG GAG TTC TAT AC-3′
fosC2-reverse	5′-CTT GAT AGG GTT TAG ACT TC-3′
fosA3/4En-forward	5′-CGA TCA CAG TTT ACA ACA GG-3′
fosA3/4En-reverse	5′-GGC TAT CTT GCT CAG CTC TA-3′

Susceptibility testing.

The MIC of fosfomycin (Wako Pure Chemical Industries, Ltd., Osaka, Japan) for each strain was determined using the agar dilution method according to the Clinical and Laboratory Standards Institute (CLSI) guidelines (13). E. coli strain ATCC 25922 was used as the control.

Disk potentiation test.

The E. coli isolates were adjusted to a McFarland standard of 0.5 according to the CLSI guideline and inoculated onto a Mueller-Hinton (MH) agar plate (Nippon Becton, Dickinson Co., Ltd., Tokyo, Japan) and an MH agar plate containing 25 μg/ml glucose-6-phosphate (G6P) (MH-G6P plate) (Sigma-Aldrich, St. Louis, MO). The addition of G6P to the MH agar plates is specified in the CLSI guidelines when determining the MIC of fosfomycin for E. coli (12). Two Kirby-Bauer (KB) disks containing fosfomycin (50 μg/disk) and G6P (5 μg/disk) (Eiken Chemical Co., Ltd., Tokyo, Japan) were placed on the plates. Sodium phosphonoformate (PPF) (Sigma-Aldrich), a candidate inhibitor of FR-GSTs, was dissolved in water to a concentration of 50 mg/ml. The inhibitor (20 μl, 1 mg in total) was applied to one of the two fosfomycin disks and incubated for 18 h at 37°C, and the diameter of the growth inhibition zone (i.e., the area where bacterial growth was completely inhibited) around each disk was measured. The diameter of the growth inhibition zone around a fosfomycin disk with PPF was compared with that around a fosfomycin disk. To test the antibacterial activity of sodium PPF itself, PPF (1 mg) was added to a blank disk and incubated, and the inhibition zone was measured.

Retrospective screening with the potentiation test.

Sixteen E. coli clinical isolates that had been predicted to be ESBL producers (cefotaxime MIC, >2 μg/ml) with MIC values that were >16 μg/ml for fosfomycin were selected from the E. coli laboratory collection stocked in the clinical microbiology laboratory of Nagoya University hospital (a 1,000-bed tertiary care national university hospital). The characteristics of these bacterial isolates were preliminarily determined using the MicroScan WalkAway system (Siemens Healthcare Diagnostics, Tokyo, Japan) as part of the routine microbiology laboratory workup. These 16 isolates were subjected to fosfomycin susceptibility testing using the agar dilution method. Those isolates that were found to be nonsusceptible to fosfomycin (MIC, ≥128 μg/ml) were further subjected to the disk potentiation test and genetic characterization for genes mediating fosfomycin resistance.

Nucleotide sequence accession number.

The nucleotide sequence of fosA4 has been recorded in GenBank under accession no. AB908992.

RESULTS

In the present study, we developed a practical disk potentiation test to macroscopically detect the production of FR-GSTs in bacterial cells using PPF. The results for two representative E. coli isolates, one fosA3 positive and the other FR-GST negative, are shown in Fig. 2A. Both strains showed resistance to fosfomycin (MIC, ≥512 μg/ml), resulting in almost no growth inhibition zone around the fosfomycin disks. Following the addition of PPF, a 4-mm expansion of the growth inhibition zone around the fosfomycin disk on the MH agar plate was observed for the fosA3-positive strain (Fig. 2A, upper left), while a 13-mm-greater expansion of the growth inhibition zone was found on the MH-G6P plate (Fig. 2A, upper right). The addition of G6P to the MH agar plate resulted in the formation of larger growth inhibition zones around the fosfomycin disks. For a fosfomycin-resistant strain not harboring any FR-GST genes, no zone enlargement was observed with PPF; however, in fact, a slight reduction in the size of the zone was observed on both the MH and MH-G6P plates (Fig. 2A, lower left and right). To test the antibacterial activity of PPF alone, 1 mg PPF was added to a blank disk. For both strains, no growth inhibition zone was observed around the PPF disk, indicating that PPF by itself had no apparent suppressive effect on bacterial growth (data not shown). Taken together, the expansion of the growth inhibition zone by PPF, especially on MH-G6P plates, is a good indicator of FR-GST production, probably via the inactivation of FR-GST activity.

FIG 2.

(A) Potentiation by PPF in growth inhibition zone diameters around a fosfomycin disk. PPF was added to a fosfomycin disk (right side in each panel). Shown are a fosA3-positive strain on an MH plate (upper left) and MH-G6P plate (upper right), and an FR-GST-negative strain on an MH plate (lower left) and MH-G6P plate (lower right). (B) The results of fosC2-positive (left) and fosA4-positive strains on MH-G6P plates (right). FR-GST, fosfomycin resistance-mediating glutathione S-transferase; PPF, phosphonoformate; MH, Mueller-Hinton; G6P, glucose-6-phosphate.

To confirm the reliability of the developed method, we tested the remaining 23 fosfomycin-intermediate and -resistant E. coli isolates from our laboratory stock (6 fosA3-positive, 1 fosC2-positive, 1 fosA4-positive, and 15 isolates that were negative for these three genes), except the two representative strains shown in Fig. 2A. The fosA3-positive fosfomycin-resistant isolates showed an enlargement of the growth inhibition zone of 11 to 14 mm around the fosfomycin disk containing PPF on MH-G6P plates (Fig. 3). The fosfomycin-intermediate and -resistant (MIC range, 128 to ≥512 μg/ml) isolates not harboring any FR-GST genes showed either no increase or a slight decrease in diameter (up to 4 mm) on MH-G6P plates (Fig. 3), although it is unclear why a slight reduction in zone diameter was observed with PPF. Compared with the MH plates, a considerable expansion of the growth inhibition zones on the MH-G6P plates was observed for all fosA3-positive strains (Fig. 3). Thus, MH-G6P agar plates are useful for identifying FosA3 production. This disk potentiation test, which uses PPF as an inhibitor of FR-GST and MH-G6P plates, makes it possible to easily identify FosA3-producing strains among fosfomycin-intermediate and -resistant E. coli clinical isolates with high reliability.

FIG 3.

Summary of changes in growth inhibition zone diameter with PPF. The y axis represents the enlargement of the growth inhibition zone (mm) by PPF. FR-GST, fosfomycin resistance-mediating glutathione S-transferase; PPF, phosphonoformate; MH, Mueller-Hinton; G6P, glucose-6-phosphate.

We applied our method to a fosC2-positive E. coli isolate (5) and observed an 11-mm expansion of the growth inhibition zone (Fig. 2B, left). In addition, an expansion of the growth inhibition zone (15 mm) was observed for one E. coli isolate producing FosA4 (Fig. 2B, right), which is a newly identified variant of FosA3; FosA4 shares 94% amino acid identity with FosA3 (Fig. 1).

Finally, to validate whether our test works well for identifying FR-GST producers, we retrospectively screened 16 E. coli clinical isolates that were preliminarily classified as probable ESBL producers, with an MIC of >16 μg/ml for fosfomycin, in our clinical microbiology laboratory. Among these 16 isolates, 12 were fosfomycin resistant (MIC, ≥256 μg/ml) and 4 were fosfomycin susceptible (MIC, ≤64 μg/ml). Thus, 12 fosfomycin-resistant isolates were subjected to the disk-based potentiation test. We observed that six isolates exhibited an expansion of their growth inhibition zones (7 to 13 mm) with PPF, but the remaining six isolates showed no enlargement of their growth inhibition zones with PPF. PCR and sequencing analyses were performed to confirm the presence of FR-GST genes in the 12 isolates. The six isolates that showed positive results in the disk potentiation test also gave positive results for FR-GST genes: 5 harbored fosA3 and 1 harbored fosA4. The remaining six isolates that showed negative results in the potentiation test were also negative for the fosA3, fosA4, and fosC2 genes.

DISCUSSION

Rigsby et al. (14) previously demonstrated the crystal structure of an FR-GST, namely, FosA^PA (PA1129) from Pseudomonas aeruginosa, in a complex with PPF (14). PPF bound to FosA^PA via interaction with MnII(+) and Thr9 at the active site of FosA^PA and behaved as a transition-state inhibitor analogous to fosfomycin. Notably, the Thr9 residue is conserved among all types of FR-GSTs (Fig. 1), including FosA3, which is active in the presence of MnII(+), similar to FosA^PA (5, 15). Therefore, we hypothesize that PPF may be an effective inhibitor of not only FosA^PA but also other FR-GSTs, such as FosA3, FosA4, and FosC2. We then used this agent to develop a practical disk potentiation test for detecting FR-GST producers. In addition, we found that supplementing the agar plates with G6P (Fig. 2 and 3) enabled us to obtain definitive results from our potentiation test. Although the commercially available fosfomycin disks in Japan contain G6P (5 μg/disk), adding G6P to the agar plates used in the test enhances the inhibitory effects caused by PPF. The addition of G6P may have accelerated the influx of fosfomycin into the bacterial cells by stimulating the expression of the UhpT transporter (16). In a disk diffusion test, the CLSI guidelines recommend using standardized fosfomycin disks containing 200 μg fosfomycin and 50 μg G6P, which are 4- and 10-fold higher, respectively, than the amounts used in this study. The use of disks containing such high dosages, especially of G6P, may eliminate the need to add G6P to the MH agar plates in our potentiation test. Nonetheless, in this study, the use of MH-G6P plates and disks containing 50 μg fosfomycin enabled us to reliably identify FR-GST producers among fosfomycin-nonsusceptible E. coli clinical isolates.

PPF seems to be a common inhibitor of all types of FR-GSTs known to date, which generally retain the Thr9 residue and are probably active in the presence of MnII(+), as demonstrated by the resistance determinants FosA3 and FosA^PA (5, 15) (Fig. 1). Thus, our test appears to be capable of identifying any type of FR-GSTs, as shown in Fig. 1, in which the active site contains the key residue Thr9 and MnII(+), and of successfully detecting the production of FosA3, FosA4, and FosC2. However, because the number of the isolates used in the present study was limited, further evaluation will be required to determine whether our test can be used to identify all types of FR-GST producers. Recently, it was reported that the fosA3 gene has been detected in Klebsiella pneumoniae clinical isolates (7). Future studies are required to evaluate the capability of the test developed here to detect FR-GST producers other than E. coli.

To identify FR-GST production, we established a 5-mm enlargement as the provisional cutoff value for the expansion of the diameter of the zone around the fosfomycin disk containing PPF on an MH-G6P plate compared with the diameter of the growth inhibition zone around the disk containing fosfomycin alone. This cutoff value was determined by taking the average minus 3 standard deviations of the values obtained for the changes in the diameter of the growth inhibition zones for 15 FR-GST-positive strains analyzed in this study. Using our cutoff value, the sensitivity and specificity of our test for identifying FR-GST producers were both 100%.

In this study, we used the potentiation test that we had developed to retrospectively screen for FR-GST producers among E. coli clinical isolates whose susceptibility to fosfomycin had preliminarily been determined through an automated system during routine work in the clinical microbiology laboratory. Using this test, we were able to classify FR-GST producers among the fosfomycin-resistant isolates. Although the mechanisms underlying fosfomycin resistance in those isolates that gave negative results in the potentiation test have not been fully characterized in the present study, the resistance of these isolates to fosfomycin does not depend on the production of FR-GSTs. Our test is simple and cost-effective; thus, it can be incorporated into the practice of screening for FR-GST producers in clinical microbiology laboratories.

In conclusion, our new method is simple, highly sensitive, specific, and feasible for routine use in clinical microbiology laboratories for identifying fosfomycin-resistant E. coli isolates producing FR-GSTs, such as FosA3, FosA4, and FosC2. Because FosA3-producing E. coli organisms have been identified in food animals and pets in China (8–10), fosfomycin-resistant E. coli should be assessed using the potentiation test in order to identify FR-GST-positive microbes in livestock farming environments at the early stage of emergence. We believe that our test will contribute significantly to both epidemiological analyses and better infection controls of fosfomycin-resistant bacteria in clinical settings.