ABSTRACT
We investigated the epidemiology and resistance mechanisms of ampicillin-sulbactam-nonsusceptible Escherichia coli, focusing on the role of the TEM-1 β-lactamase. We collected all nonduplicate E. coli clinical isolates at 10 Japanese hospitals during December 2014 and examined their antimicrobial susceptibility, β-lactamases, TEM-1 transferability, TEM-1 β-lactamase activity, outer membrane protein profile, membrane permeability, and clonal genotypes. Among the 329 isolates collected, 95 were ampicillin-sulbactam nonsusceptible. Of these ampicillin-sulbactam-nonsusceptible isolates, β-lactamases conferring resistance to sulbactam, such as AmpC, were present in 33%. Hyperproduction of sulbactam-susceptible β-lactamases, TEMs with a strong promoter, were rare (5%). The remaining 59 isolates (62%) had only sulbactam-susceptible β-lactamases, including TEM-1 with a wild-type promoter (n = 28), CTX-Ms (n = 13), or both (n = 17). All 45 transconjugants from 96 donors with TEM-1 had higher ampicillin-sulbactam MICs (4 to 96 mg/liter) than the recipient (2 mg/liter). In donors with only TEM-1, TEM-1 activity correlated with the 50% inhibitory concentration of sulbactam and ampicillin-sulbactam MICs. The decreased membrane permeation of sulbactam was associated with an increased ampicillin-sulbactam MIC. The reduced permeation was partly attributable to deficient outer membrane proteins, which were observed in 57% of the ampicillin-sulbactam-nonsusceptible isolates with only TEM-1 and a wild-type promoter. Sequence type 131 (ST131) was the most common clonal type (52%). TEM-1 with a wild-type promoter primarily contributed to ampicillin-sulbactam nonsusceptibility in E. coli, with the partial support of other mechanisms, such as reduced permeation. Conjugative TEM-1 and the clonal spread of ST131 may contribute to the prevalence of Japanese ampicillin-sulbactam-nonsusceptible isolates.
INTRODUCTION
Escherichia coli causes extraintestinal infections, including urinary tract infections, intra-abdominal infections, and bacteremia. Ampicillin-sulbactam (SAM) and amoxicillin-clavulanate (AMC) have wide spectra of activity against Gram-positive, Gram-negative, and anaerobic organisms. In Japan, SAM is commonly used in daily clinical practice, whereas intravenous AMC is not available. However, a decreasing rate of susceptibility to SAM among E. coli strains threatens its continued clinical use (1).
The TEM-1 β-lactamase belongs to group 2b in the Bush-Jacoby classification scheme and is inhibited by β-lactamase inhibitors, such as sulbactam and clavulanate (2). Therefore, E. coli isolates with TEM-1 are usually susceptible to SAM and AMC. However, hyperproduction of TEM-1 overcomes the inhibitory effects of sulbactam and clavulanate (3) and has been reported to be a common resistance mechanism against SAM and AMC in E. coli (4, 5). A strong promoter, such as the Pa/Pb promoter, can contribute to TEM-1 hyperproduction (6).
Other mechanisms that have been described include plasmid-mediated AmpC β-lactamase (p-AmpC), hyperproduction of the chromosomal AmpC β-lactamase (c-AmpC), OXA β-lactamase, and inhibitor-resistant TEM (IRT) β-lactamase (7). A deficiency of outer membrane proteins (OMPs), such as OmpF and OmpC, has been indicated to contribute to increases in the MIC for isolates with TEM-1 (3). However, the prevalence and contribution of OMP deficiency in SAM- or AMC-nonsusceptible E. coli remain unknown.
In addition to the horizontal gene transfer of blaTEM-1, which is usually associated with a transposon (8) and which is present on a plasmid (9), the dissemination of SAM-resistant E. coli is attributable to E. coli sequence type 131 (ST131), which has spread worldwide and is recognized as a main driver of fluoroquinolone resistance and extended-spectrum β-lactamase (ESBL) production (10, 11). ST131 strains frequently harbor blaTEM-1 (10). ST131 strains consist of 3 different clades (clades A, B, and C). In clade C, subclades C1 and C2 are associated with fluoroquinolone resistance. Subclade C2 and a specific lineage within C1, C1-M27, are associated with ESBLs (12, 13). The ST131 clade responsible for the spread of SAM resistance is not known.
To understand the mechanisms and epidemiology of SAM resistance and the increase in SAM-nonsusceptible E. coli, we investigated the β-lactamase genes, transferability of blaTEM-1, TEM-1 activity, OMPs, membrane permeability, and clonal genotypes among clinical E. coli isolates in Japan.
RESULTS AND DISCUSSION
We investigated a total of 329 clinical E. coli isolates that were consecutively collected by a Japanese multicenter surveillance program (14). Of these isolates, 95 isolates (29%) were nonsusceptible (intermediate, 60 isolates; resistant, 35 isolates) to SAM, and 61 isolates (19%) were nonsusceptible to AMC. The prevalence of SAM-nonsusceptible isolates is consistent with the values reported in previous studies from the Asia-Pacific region (30%) (15) and the United States (31%) (4).
Epidemiology of β-lactamase genes.The comparison of the β-lactamase genes between the SAM-nonsusceptible and SAM-susceptible isolates revealed a high prevalence of an acquired β-lactamase gene (91% and 28%, respectively; P < 0.01), especially blaTEM-1 with a P3 promoter, which weakly promotes blaTEM-1 expression, in SAM-nonsusceptible isolates (Table 1). This high prevalence of blaTEM-1 with a P3 promoter is consistent with that described in previous reports (4). The hyperproduction of sulbactam-susceptible β-lactamases (group A in Table S1 in the supplemental material), namely, TEMs with the Pa/Pb promoter, was rare (5%).
β-Lactamases and promoter types of blaTEMs detected in SAM-nonsusceptible and SAM-susceptible isolates
Twelve isolates (13%), including the 9 isolates lacking any acquired β-lactamase genes, were defined as c-AmpC hyperproducers with promoter mutations (Table S2), and their prevalence was similar to that in a Spanish study (5). The rarity of IRT (1%) and the relatively low frequencies of p-AmpC and blaOXA-1 (6% and 13%, respectively) in SAM-nonsusceptible isolates are consistent with the findings of previous studies (4, 16). These sulbactam-resistant β-lactamases together accounted for one-third of the SAM-nonsusceptible isolates in our study (group B in Table S1).
The β-lactamases that were not considered to confer SAM nonsusceptibility as a single resistance mechanism included blaTEM-1 with a P3 promoter (29%), the combination of blaTEM-1 and blaCTX-M (18%), and blaCTX-M only (14%) (group C in Table S1). The ubiquitous presence of the acquired β-lactamases or c-AmpC and the high frequency of blaTEM-1 in the SAM-nonsusceptible isolates support the notion that β-lactamases, especially blaTEM-1 with a P3 promoter, are important contributors to SAM nonsusceptibility.
Epidemiology of deficient OMPs in SAM-nonsusceptible isolates.We investigated the OMP profiles of SAM-nonsusceptible isolates by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Representative OMP profiles obtained by SDS-PAGE are shown in Fig. S1. We identified 38 isolates (40%) with deficient OMPs (deficient OmpF, 35 isolates; deficient OmpC; 3 isolates; both deficient OmpF and deficient OmpC, 3 isolates), and the sequence profiles of the deficient OMPs are shown in Table S3. Among 28 isolates with only blaTEM-1 with a P3 promoter, more than a half of the isolates (57%) had deficient OMPs (Table S1).
Transferability of blaTEM-1 by conjugation.A conjugation assay was performed to assess blaTEM-1-transferability and the influence of blaTEM-1 on SAM nonsusceptibility in all 96 blaTEM-1-positive isolates. We obtained 45 blaTEM-1-positive transconjugants, 1 of which harbored blaTEM-1 and blaCTX-M (Table S4). Plasmid replicon typing revealed that incompatibility group F (IncF) accounted for 89% of the transconjugants with blaTEM-1 (Table S5).
SAM MICs for donor-transconjugant pairs harboring TEM-1 with or without other β-lactamase genes.The SAM MICs for the recipient E. coli J53 (2 mg/liter) were elevated through acquisition of conjugative blaTEM-1 in all transconjugants (median MIC, 12 mg/liter; range, 4 to 96 mg/liter) (Fig. S2), suggesting the potential role of TEM-1 in SAM nonsusceptibility.
The higher MICs for donors with the combination of blaTEM-1 and blaCTX-M than for their transconjugants harboring only blaTEM-1 (Fig. S2) and the higher frequency of this combination in SAM-nonsusceptible isolates than in SAM-susceptible isolates (19% and 2%, respectively; P < 0.01) (Table 1) suggest that the combination of multiple β-lactamases, each of which alone could be inhibited by sulbactam (17), may increase the MIC of SAM.
SAM MICs and β-lactamase activity of donor-transconjugant pairs harboring only TEM-1.Of the 45 donor-transconjugant pairs, we selected the 32 pairs with only blaTEM-1 to further investigate the role of TEM-1. Acquisition of conjugative blaTEM-1 in transconjugants did not always increase the SAM MIC, as was observed for donors (Spearman’s rank correlation [rS] = 0.34, P = 0.06) (Fig. 1A).
TEM-1 activity and SAM MIC for 32 donor-transconjugant pairs with only TEM-1 and ATCC 35218. (A) SAM MICs (shown by the ampicillin MICs) for donors and transconjugants. The regression line was calculated from the data for all the donors and their transconjugants with TEM-1. When the SAM MICs for the donor and its transconjugant are equal, the corresponding dot should be on the line y = x (dashed line). (B, C) The TEM-1 activities and SAM MICs for donors (B) and their transconjugants (C). The regression lines were calculated from the data for donors or transconjugants with TEM-1 with a P3 promoter. (D) TEM-1 activities of donors and their transconjugants. The regression line was calculated from the data for all the transconjugants and their donors with TEM-1. When the TEM-1 activities of the donor and its transconjugant are equal, the corresponding dot should be on the line y = x (dash line). In two donors (arrows), MICs and TEM-1 activities were two times higher than those of their transconjugants (SAM MIC, 24 versus 8 mg/liter and 64 versus 8 mg/liter; TEM-1 activity, 17.7 versus 7.2 nmol/min/ml and 37.0 versus 10.4 nmol/min/ml), and their cefoxitin-cloxacillin disk test results were negative. Spearman’s rank correlation is indicated by rS. A transconjugant of ATCC 35218 could not be obtained.
We measured the β-lactamase activity of the 32 pairs (Fig. 1B and C). A previous study defined AMC-resistant Enterobacteriaceae with only blaTEM-1 whose β-lactamase activity was at least 2-fold higher than that of ATCC 35218, a reference E. coli strain harboring blaTEM-1, as TEM-1 hyperproducers and found that 8 of 10 isolates were TEM-1 hyperproducers (18). Among the 32 donors, this definition identified 2 TEM-1 hyperproducers (6%), which showed high SAM MICs (64 and 128 mg/liter). One of them possessed TEM-1 with the Pa/Pb promoter, which showed the highest β-lactamase activity and the highest MIC for both the donor and its transconjugant. Although the definition of TEM-1 hyperproducers is dependent on the cutoff value of β-lactamase activity, the frequency of TEM-1 hyperproducers among SAM-nonsusceptible isolates may be low in our study, as opposed to a previous study suggesting a predominance of TEM-1 hyperproducers among SAM-nonsusceptible isolates (4).
The positive correlation between the β-lactamase activity of TEM-1 with a P3 promoter and the SAM MICs (for donors, rS = 0.59 and P < 0.01; for transconjugants, rS = 0.82 and P < 0.01) is consistent with the findings of a study analyzing TEM-1-producing E. coli clinical isolates (19), in which some isolates did not conform to a quantitative relationship between β-lactamase activity and the AMC MIC. Likewise, in our study, donors likely showed a smaller correlation coefficient between β-lactamase activity and the SAM MICs than transconjugants (P = 0.07, Fisher’s z transformation), reflecting diverse host cell characteristics among donors which can affect SAM MICs, in contrast to the characteristics of the same host cells (J53) among transconjugants. This difference in host characteristics may also explain the discrepancy in the SAM MIC between donors and transconjugants mentioned above (Fig. 1A). The very low activity (<0.5 nmol/min/ml) of a basal level of the c-AmpC β-lactamase of J53 and the correlation of the β-lactamase activity between donors and transconjugants, except in 2 isolates (Fig. 1D), imply that the β-lactamase activity of donors primarily represents the activity of TEM-1. Therefore, in addition to TEM-1 activity, which showed a positive relationship with the SAM MIC, host factors other than β-lactamase may contribute to increased SAM MICs.
OMPs of donors with only TEM-1.SDS-PAGE of the 32 donors identified 17 donors with deficient OmpF and 7 donors with deficient OmpC (Fig. 1A). The ratio of the MICs for donors to those for transconjugants was higher in isolates with deficient OmpF than in isolates with wild-type OmpF, whereas this ratio was not associated with the OmpC profile (P = 0.03 and P = 0.12, respectively, Mann-Whitney test), implying that donors with deficient OmpF are likely to have a higher SAM MIC than their transconjugants regardless of TEM-1 activity. Therefore, deficient OmpF may be one of the host factors influencing the SAM MIC.
The concentration of sulbactam required to inhibit TEM-1 for determination of membrane permeability.To investigate membrane permeability in isolates with nonhyperproducing TEM-1, for which the SAM MIC can easily be influenced by host factors, we analyzed the permeability index (20). To determine the permeability index, we quantified the 50% inhibitory concentrations (IC50) of sulbactam, at which TEM-1 activity is reduced by half, for lysates (lysate-IC50) and intact cells (cell-IC50). After excluding the 2 TEM-1 hyperproducers, we selected 19 isolates from among the remaining 30 donors representing three groups according to their correlations between TEM-1 activity and the SAM MICs (Fig. S3), in addition to ATCC 35218.
The lysate-IC50 and cell-IC50 correlated with the TEM-1 activity of lysates without sulbactam (rS = 0.85 and P < 0.01 and rS = 0.57 and P < 0.01, respectively) (Fig. 2A and B), which confirmed the quantitative relationship between TEM-1 activity and the sulbactam concentration. In contrast to sulbactam’s direct inhibition of the released TEM-1 in the suspension of sonicated lysates, permeation of sulbactam into the periplasm is required for inhibition of TEM-1 in intact cells; thus, the various membrane permeabilities (e.g., OMPs and efflux pumps) among clinical isolates may affect the cell-IC50. This hypothesis is supported by the correlation observed between the cell-IC50, but not the lysate-IC50, and the SAM-MIC measured using intact cells (rS = 0.75 and P < 0.01 and rS = 0.31 and P = 0.18, respectively) (Fig. 2C and D).
The 50% inhibitory concentration of sulbactam and its relationships with TEM-1 activity and the SAM MIC for 19 clinical isolates and ATCC 35218. (A, B) TEM-1 activity and the 50% inhibitory concentration (IC50) of sulbactam for sonicated lysates (lysate-IC50) (A) and intact cells (cell-IC50) (B) harboring only blaTEM-1 with a P3 promoter. (C, D) Lysate-IC50 (C) and cell-IC50 (D) compared with the SAM MIC. The ampicillin MICs are shown. Spearman’s rank correlation is indicated by rS.
Membrane permeability and OMPs.For the selected 19 isolates and ATCC 35218, we determined the permeability index for sulbactam as the ratio of the cell-IC50 to the lysate-IC50 (20). Isolates with a deficient OmpF (10 isolates) had a higher permeability index, reflecting greater penetration difficulty, than those with wild-type OmpF (P = 0.03), although no such association with the permeability index was observed in isolates with a deficient OmpC (5 isolates) (P = 0.18) (Fig. 3A and B). Reguera et al. reported that an OmpF-deficient laboratory strain with a blaTEM-1-containing plasmid had a higher amoxicillin-sulbactam MIC and AMC MIC than a strain with wild-type OmpF and the same plasmid (3). Our results support their hypothesis that sulbactam penetrates bacterial cells via OmpF. β-Lactams, including ampicillin, have also been shown to diffuse principally through OmpF (21). As the permeability index should be determined by the combined effect of the influx and efflux of sulbactam, the impact of the efflux pump remains to be elucidated. In summary, deficient OmpF may play some part in the reduced membrane permeation of sulbactam.
Outer membrane proteins and the permeability index of sulbactam for 19 clinical isolates and ATCC 35218. The permeability indices of sulbactam for clinical isolates and ATCC 35218 (arrow) according to their OmpF (A) and OmpC (B) profiles by SDS-PAGE are shown. A higher permeability index corresponds to less sulbactam penetration of the outer membrane. The median permeability index is indicated by the horizontal line. The statistical analysis was performed with the Mann-Whitney U test.
Impact of the permeability index on the relationship between TEM-1 activity and the SAM MIC.We assessed whether the permeability index can explain the discrepancy between TEM-1 activity and the SAM MIC. Figure 1B was redrawn with the permeability data for 19 selected isolates and ATCC 35218 (Fig. 4A). Isolates with a high SAM MIC compared with their TEM-1 activity (isolates above the regression line) tended to have high permeability indices. To clarify these relationships, the ratio of the SAM MIC to TEM-1 activity was calculated (Fig. 4B) and was found to correlate with the permeability index. The MIC for one blaTEM-1-deficient isolate obtained by plasmid curation experiments suggests that the SAM MIC for the host cells, which may be partially influenced by membrane permeability, may have a basal impact on the SAM MIC for isolates with TEM-1 (Table S6). Therefore, the membrane permeation of sulbactam, in addition to TEM-1 activity, contributes to an increase in the SAM MIC. However, the relatively small correlation coefficient (rS = 0.59, P < 0.01) leads us to suspect the presence of other host factors, including the interaction between ampicillin and the cell wall and the membrane permeation of ampicillin.
Relationships between the permeability index and TEM-1 activity or the SAM MIC in 19 clinical isolates and ATCC 35218. (A) TEM-1 activities and SAM MICs (shown by the ampicillin MICs). Each plot is labeled with the permeability index. A higher permeability index corresponds to less sulbactam penetration of the outer membrane. (B) Permeability indices and ratio of the SAM MIC to TEM-1 activity. A high ratio implies a relatively high SAM MIC for TEM-1 activity. Spearman’s rank correlation is indicated by rS.
Sequence types and the characteristics of isolates.We previously analyzed the lineage of the study isolates by CH (fumC and fimH) typing and an ST131-specific PCR (14). blaTEM-1 was more common in the ST131 isolates than in the non-ST131 isolates (44% and 22%, respectively; P < 0.01), as previously reported (10), and was associated with SAM nonsusceptibility in the ST131 isolates (SAM-nonsusceptible isolates, 57%; SAM-susceptible isolates, 33%; P = 0.01). ST131 was also associated with TEM-1 with CTX-M and OXA-1 in SAM-nonsusceptible isolates (Table S1). Using the results of multiplex PCR in our previous study (22), we further classified ST131 into clades, each of which showed different characteristics (Table 2). Subclade C2 showed an association with SAM nonsusceptibility, and OXA-1 was the most common β-lactamase in SAM-nonsusceptible isolates of this subclade, followed by TEM-1 with CTX-M.
ST131 clades of the SAM-nonsusceptible and SAM-susceptible isolates
Our conjugation assay demonstrated that blaTEM-1, which was mostly on the IncF plasmid, could transfer in 47% of isolates. Although the conjugative blaTEM-1 was commonly present in ST131 isolates (26 isolates, 58%), it was distributed variably in ST131 clades (C1, non-M27, 12 isolates; C2, 5 isolates; A, 4 isolates) as well as non-ST131 clones (7 CH types other than ST131), indicating horizontal transfer by a plasmid or other mobile genetic elements, such as transposons (8).
Some potential limitations exist in this study. We examined OMPs by the use of SDS-PAGE and the permeability index, but we did not perform gene editing of OMPs; thus, the role of OMPs requires further clarification. Other host characteristics, such as efflux pumps or the permeation of ampicillin, were not evaluated. We did not confirm the links between specific plasmids of the transconjugants and blaTEM-1. We plan to identify whole-nucleotide sequences of these plasmids to clarify such links. The short duration of isolate collection may reflect a potential selection bias. To our knowledge, this is the first study to comprehensively analyze the roles of TEM-1 in the SAM MIC by blaTEM-1 epidemiology, TEM-1 activity, the inhibitory effect, the membrane permeation of sulbactam, and OMP profiles. We believe that this research provides insight into the resistance mechanisms of SAM nonsusceptibility.
Conclusions.In Japan, SAM-nonsusceptible E. coli is prevalent and threatens the optimal use of SAM. TEM-1, the most common β-lactamase, alone or in combination with decreased membrane permeability, overcame the inhibitory effect of sulbactam and contributed to an increase in the SAM MIC. The impaired permeation of sulbactam was associated with a deficient OmpF, which was observed in more than half of the SAM-nonsusceptible isolates with only TEM-1. Other possible mechanisms include CTX-M in combination with blaTEM-1 and sulbactam-resistant β-lactamases (c-AmpC, blaCMY-2, and blaOXA-1). The increase in SAM-nonsusceptible E. coli is associated with conjugative plasmids with blaTEM-1 and/or an ST131 pandemic clone carrying the above-mentioned β-lactamases. We must continue to monitor the abundance of β-lactamases and clones that contribute to resistance.
MATERIALS AND METHODS
In total, 329 nonduplicate clinical E. coli isolates consecutively collected at 10 hospitals from 5 prefectures in Japan (5 hospitals in Kyoto, 2 hospitals in Osaka, and 1 hospital each in Aichi, Hyogo, and Shiga) during December 2014 (14) were included. The most common source of the isolates was the urinary tract (214 isolates, 65%), followed by blood culture (39 isolates), intra-abdominal sites (19 isolates), the female genital tract (18 isolates), the skin and subcutaneous tissue or pus (17 isolates), the respiratory tract (15 isolates), and others (7 isolates). Susceptibility to the antimicrobials SAM and AMC was evaluated by microdilution at the reference laboratory using a Dry Plate Eiken (Eiken, Tokyo, Japan) following CLSI standards (23). E. coli isolates with an MIC of >8/4 mg/liter at a fixed ratio of 2:1 were defined as SAM nonsusceptible regardless of their susceptibilities to the other antimicrobials. The Ethics Committee of the Kyoto University Graduate School and Faculty of Medicine approved this study (E-2103) and waived the requirement to obtain informed consent from each patient.
β-Lactamase identification.Bacterial DNA was isolated using a QIAamp DNA minikit (Qiagen, Hilden, Germany). The presence of β-lactamase genes was defined by PCR amplification and sequencing of the blaTEM (24), blaSHV (25), blaCTX-M (26), blaGES (27), blaPER (27), blaVEB (27), and blaOXA (28, 29) genes and the six main groups of p-AmpC-type genes (30). c-AmpC promoter mutations were also examined by PCR and sequencing (31). c-AmpC hyperproducers were defined as SAM-nonsusceptible isolates with mutations in the promoter and/or attenuator region and a positive phenotypic test by the cefoxitin-cloxacillin disk method (32).
Conjugation.The transferability of blaTEM-1 was assayed by conjugation. Conjugation experiments were performed by the liquid mating method using azide-resistant E. coli J53 as the recipient and all the blaTEM-1-harboring clinical isolates, including blaCTX-M or blaCMY-2-coharboring isolates, as the donors, regardless of their SAM susceptibility. Transconjugants selected on Trypticase soy agar (TSA) containing azide (50 mg/liter) and ampicillin (50 mg/liter) were further differentiated from donors by phylogenetic grouping (33). The transfer of blaTEM-1 and other β-lactamase genes from the donors was confirmed by PCR. The SAM susceptibility of the transconjugant and its donor was tested by the Etest (Sysmex bioMérieux, Tokyo, Japan). Plasmid replicon typing of the transconjugants was performed as described previously (34, 35).
β-Lactamase activity assay.We measured the β-lactamase activity of the donors and transconjugants with only blaTEM-1. Cells were harvested from midgrowth Luria-Bertani broth cultures (5 ml) by centrifugation and resuspended in sodium phosphate buffer (0.1 M, pH 7) at an optical density at 600 nm (OD600) of 0.3. One milliliter of cell suspension was sonicated to release β-lactamase (six bursts of 10 s each on ice). ATCC 35218 was used to determine the appropriate duration of sonication to completely release β-lactamase (see Fig. S4 in the supplemental material). β-Lactamase activity was measured spectrophotometrically against 100 μM nitrocefin at 492 nm with an MTP-310 microplate reader (Corona Electric, Ibaraki, Japan). All assays were performed in triplicate at room temperature. As a negative control, an equivalent amount of buffer rather than cell lysate was added. ATCC 35218 was used as a positive control.
Assay of IC50 and permeability index.The cell suspension described above was divided into 2 portions (1 ml each). The first aliquot, which was directly used to determine the IC50 for TEM-1 in intact cells (cell-IC50), was incubated with sulbactam (6 to 8 different concentrations) at 37°C for 120 min as described previously (36). The cell suspension was washed twice. After sonication, the uninhibited β-lactamase activity against nitrocefin was measured. The second aliquot was sonicated to first disrupt the cells to determine the IC50 for the released TEM-1 (lysate-IC50). The lysate was incubated with sulbactam for 20 min. Then, β-lactamase activity was measured. A 4-parameter logistic model was used to determine the IC50 values of intact cells and lysates. The permeability index was defined as the ratio of the cell-IC50 to the lysate-IC50 (20). We used the drc package in R for the analysis (37).
OMP analysis.The coding sequences of ompF and ompC, amplified as previously described (38), were compared with those of the E. coli K-12 reference strain (GenBank accession number CP001637.1). OmpF and OmpC were identified by SDS-PAGE as described previously (39). Briefly, a membrane fraction was extracted from a cell suspension (7 ml) at an OD600 of 0.5 with sodium N-lauroyl sarcosinate. Then, the extracted protein was analyzed by SDS-PAGE. E. coli ATCC 25922 (wild type) was used as a control.
Clonal group identification.The genetic relationship between the E. coli isolates was determined by CH typing using fumC and fimH alleles, as described in our previous study (40). The ST131 status was also determined by PCR-based detection of ST131-specific single nucleotide polymorphisms in gyrB and mdh (41). To confirm the major ST131 clades, multiplex PCR was performed as described in our previous study (22).
Statistical analysis.Categorical variables were compared using Fisher’s exact test. Continuous variables were compared using the Mann-Whitney U test. The relationships between the continuous variables were represented by Spearman’s rank correlation (rS). The correlation coefficients were compared using Fisher’s z transformation. The results were considered significant at an alpha level of 0.05. We conducted our statistical analysis using Stata software, version 13 (StataCorp, College Station, TX, USA).
ACKNOWLEDGMENTS
We thank the members of the 89th Annual Meeting of the Japanese Association for Infectious Diseases–Drug-Resistant Bacteria Research Group (89JAID-BRG; Ooi Yukimasa, Tomohiro Higashiyama, Masaaki Sasano, Harumi Watanabe, Tsunehiro Shimizu, Akihiko Hayashi, Yasufumi Matsumura, Taro Noguchi, Miki Nagao, Masaki Yamamoto, Michio Tanaka, Takeshi Higuchi, Toru Kanahashi, Naohisa Fujita, Toshiaki Komori, Yukiji Yamada, Kohei Ueda, Go Yamamoto, Kunihiko Moro, and Shigenobu Yabu) for collection of the isolates used in this study.
FOOTNOTES
- Received 6 November 2018.
- Accepted 7 November 2018.
- Accepted manuscript posted online 19 November 2018.
Supplemental material for this article may be found at https://doi.org/10.1128/AAC.02366-18.
- Copyright © 2019 American Society for Microbiology.
