Evidence of Nosocomial Infection in Japan Caused by High-Level Gentamicin-Resistant Enterococcus faecalis and Identification of the Pheromone-Responsive Conjugative Plasmid Encoding Gentamicin Resistance

Abstract

A total of 1,799 Enterococcus faecalis isolates were isolated from inpatients of Gunma University Hospital, Gunma, Japan, between 1992 and 1996. Four hundred thirty-two (22.3%) of the 1,799 isolates had high-level gentamicin resistance. Eighty-one of the 432 isolates were classified and were placed into four groups (group A through group D) with respect to the EcoRI restriction endonuclease profiles of the plasmid DNAs isolated from these strains. The 81 isolates were isolated from 36 patients. For 35 of the 36 patients, the same gentamicin-resistant isolates were isolated from the same or different specimens isolated from the same patient at different times during the hospitalization. For one other patient, two different groups of the isolates were isolated from the same specimen. Groups A, B, C, and D were isolated from 5, 14, 12, and 6 patients, respectively. The strains had multiple-drug resistance. The restriction endonuclease digestion patterns of the E. faecalis chromosomal DNAs isolated from isolates in the same group were also identical. The patients who had been infected with the gentamicin-resistant isolates from each group were geographically clustered on a ward(s). These results suggest that the isolates in each group were derived from a common source and had spread in the ward. The gentamicin-resistant isolates exhibited a clumping response upon exposure to pheromone (E. faecalis FA2-2 culture filtrate). The gentamicin resistance transferred at a high frequency to the recipient E. faecalis isolates by broth mating, and the pheromone-responsive plasmids encoding the gentamicin resistance were identified in these isolates.

Enterococcus strains have become a significant cause of nosocomial infections (15, 17, 18, 22, 27). Of the members of the genus Enterococcus, E. faecalis and E. faecium are commonly isolated from humans. These two organisms account for 85 to 95 and 5 to 10% of the strains isolated from clinical infections, respectively. The Enterococcus strains isolated from clinical infections have multiple-drug resistance. The multiple-drug resistance of the enterococci provides these organisms with a selective advantage in the hospital environment. Outbreaks of nosocomial infections caused by enterococcal strains resistant to various drugs have been reported previously (9, 10, 16–18, 23, 28, 29).

In a study of clinical isolates from patients in Gunma University Hospital in Gunma, Japan, enterococci were found to be the second most common among the gram-positive bacteria, after Staphylococcus aureus (unpublished data). Of the clinical E. faecalis isolates, most (about 80%) were resistant to tetracycline. Between 30 and 40% of the isolates were resistant to gentamicin or erythromycin. Ampicillin- or vancomycin-resistant strains were not isolated (14, 24). Certain E. faecalis conjugative plasmids confer a mating response to the small sex pheromones secreted by potential recipient cells (1–4, 8, 11). This mating signal induces the synthesis of a surface aggregation substance that facilitates the formation of mating aggregates and plasmid transfer (2–4, 7, 11, 25). Most (60%) of the drug-resistant strains exhibit a clumping response with a culture filtrate of a plasmid-free E. faecalis recipient strain (24), suggesting that the strains harbor a pheromone-responding plasmid.

To our knowledge, there is no report concerning nosocomial infection caused by enterococci in Japan. In this report, we describe nosocomial infections in Gunma University Hospital caused by high-level gentamicin-resistant isolates of E. faecalis and isolation of the pheromone-responsive plasmids from the isolates.

MATERIALS AND METHODS

Bacteria, media, and reagents.

A total of 1,799 clinical isolates of E. faecalis were obtained from multiple sites or specimens from 1,412 patients who had been admitted to Gunma University Hospital between 1992 and 1996. The sites or specimens included urine, pus, exudate, sputum, vagina, abscess, decubitus ulcer, bile, and blood. E. faecalis was identified with the API Strep 20 system (bioMerieux S. A., Marcy l’Etoile, France). E. faecalis FA2-2 (rifampin resistant [Rif^r], fusidic acid resistant [Fus^r]) (5), JH2SS (streptomycin resistant [Str^r], spectinomycin resistant [Spc^r]) (26), OG1RF (Rif^r Fus^r) (20), OG1-10 (Str^r) (10), and OG1X (13) were used as recipient strains. Unless otherwise indicated, the media used throughout this study were nutrient broth no. 2 (Oxoid, Basingstoke, Hants, England) supplemented with glucose (0.2%) and Tris-HCl (0.1 M; pH 7.7) (N2GT broth), antibiotic medium 3 (Difco Laboratories, Detroit, Mich.), and Todd-Hewitt broth (Difco Laboratories). The antibiotic concentrations used in the selective plates were as follows: erythromycin, 12.5 μg/ml; streptomycin, 500 μg/ml; spectinomycin, 250 μg/ml; tetracycline, 12.5 μg/ml; kanamycin, 500 μg/ml; gentamicin, 500 μg/ml; fusidic acid, 25 μg/ml; rifampin, 25 μg/ml; vancomycin, 3 μg/ml; chloramphenicol, 12.5 μg/ml; ampicillin, 12.5 μg/ml. Gentamicin resistance levels were determined by the agar dilution method. Overnight cultures of the strains grown in Todd-Hewitt broth were diluted 100 times with fresh broth. One loopful of each dilution was plated on agar plates containing drug. The drugs used were diluted by the agar dilution method. The plates were incubated for 18 h at 37°C.

Isolation and manipulation of plasmid DNA.

Plasmid DNA was isolated by the alkaline lysis method (21). Plasmid DNA was treated with restriction enzymes and was submitted to agarose gel electrophoresis for the analysis of DNA fragments. Restriction enzymes were obtained from Nippon Gene (Toyama, Japan), New England Biolabs, Inc. (Beverly, Mass.), and Takara (Tokyo, Japan) and were used in accordance with the suppliers’ specifications. Agarose was obtained from Wako Chemicals, Osaka, Japan.

Mating procedures.

Broth matings were performed as described previously (8, 11) with a donor/recipient ratio of 1:10. Overnight cultures of 0.05 ml of the donor and 0.5 ml of the recipient were added to 4.5 ml of fresh broth, and the mixtures were incubated at 37°C with gentle agitation for 4 h and then vortexed. Portions of the mixed culture were then plated onto a solid medium with the appropriate selective antibiotics. Colonies were counted after 48 h of incubation at 37°C.

Pulsed-field gel electrophoresis of chromosomal DNA.

For restriction endonuclease digestion of chromosomal DNA, small slices of the agarose plugs were placed into a mixture of 270 μl of distilled water, 30 μl of 10× reaction buffer, and 50 U of SmaI (New England BioLabs), and the mixture was incubated at 25°C overnight. After digestion, the plugs were washed for 1 h at room temperature. The slices were placed in wells of a 1.2% SeaPlaque GTG agarose gel (FMC, Rockland, Maine) made with 0.5× TBE (10× TBE is 0.89 M Tris, 0.89 M boric acid, and 0.025 M EDTA), and the wells were sealed with the same agarose. The gels were electrophoresed with a clamped homogeneous electric field (CHEF-DR II; Bio-Rad Laboratories, Richmond, Calif.) and were then stained with ethidium bromide and photographed with a UV light source.

Clumping assay.

Detection of aggregation (clumping) was carried out as described previously (7, 8, 11). The pheromone corresponded to a culture filtrate of plasmid-free strain FA2-2. Generally, 0.5 ml of a culture filtrate obtained from late-logarithmic-phase, growing cells was mixed with 0.5 ml of fresh N2GT broth and 20 μl of an overnight culture of the cells to be tested for the pheromone response. The mixtures were cultured for 4 h at 37°C with gentle shaking and were then examined for clumping.

RESULTS

High-level gentamicin-resistant clinical E. faecalis isolates.

A total of 1,799 clinical isolates of E. faecalis were obtained from 1,412 patients who had been admitted to Gunma University Hospital between 1992 and 1996. Four hundred thirty-two (24%) of the 1,799 isolates were high-level gentamicin resistant (MIC, more than 500 μg/ml). The plasmids isolated from 432 gentamicin-resistant isolates were analyzed by agarose gel electrophoresis. The plasmid DNA isolated from each isolate was digested with EcoRI, and the digested DNA was submitted to agarose gel electrophoresis. Eighty-one isolates isolated from 36 patients were classified into four groups (groups A to D) with respect to the EcoRI restriction profiles of their plasmids. The EcoRI restriction profiles of the plasmids from each group are presented in Fig. 1. Figure 2 presents the groups of the isolates, the wards, case numbers for the patients who had been infected with the gentamicin-resistant isolates, the length of each patient’s hospitalization, the times when the E. faecalis strains were isolated during the hospitalization, the specimens from which the E. faecium strains were isolated, and the results of the specimen cultures. For 35 of the 36 patients examined, the same gentamicin-resistant E. faecalis isolates were isolated from the same or different specimens isolated from the same patient at different times during the hospitalization. An abscess from patient 31 on the first surgical ward contained both group B and group D isolates. Patients who had been infected with the gentamicin-resistant strains from each group were geographically clustered on a ward(s). The group A isolates were isolated from patients on the internal medicine ward. Group B isolates were isolated from patients on the second surgical ward, the internal medicine ward, and the first surgical ward. The group C isolates were isolated from patients on the second surgical ward. The group D isolates were isolated from patients on the first surgical ward. All patients were hospitalized for more than 2 months.

FIG. 1.

Agarose gel electrophoresis of EcoRI-digested plasmid DNAs isolated from group A, B, C, and D strains. Bacteriophage λ DNA digested with HindIII was used as a molecular size marker.

FIG. 2.

Nosocomial infections of inpatients due to high-level gentamicin-resistant E. faecalis. Horizontal lines, the lengths of hospitalization; triangle, time when gentamicin resistant E. faecalis was isolated during hospitalization; black triangle, specimen contained mixed culture; shaded triangle, specimen contained E. faecalis and E. faecium; open triangle, specimen contained only E. faecalis. Abbreviations: a, abscess; d, decubitus ulcer; e, exudate; g, gall; s, sputum; u, urine.

Restriction endonuclease digestion patterns of E. faecalis chromosomal DNA.

The patterns obtained by pulsed-field gel electrophoresis were used to compare the gentamicin-resistant E. faecalis strains. The restriction endonuclease digestion patterns of the chromosomal DNAs from E. faecalis strains in the same group were identical (data not shown). Comparison of the restriction endonuclease digestion patterns of the chromosomal DNAs from the E. faecalis strains in the four groups showed three different patterns (Fig. 3). Group B and group D isolates had identical restriction endonuclease digestion patterns.

FIG. 3.

Pulsed-field gel electrophoresis of SmaI-digested chromosome DNAs isolated from group A, B, C, and D strains. A bacteriophage λ DNA ladder was used as a molecular size marker.

The specimens and prognosis for the patients.

The high-level gentamicin-resistant E. faecalis isolates were isolated from different specimens derived from the same patient (Fig. 2). When the same specimens derived from the same patient were counted as one specimen, 58 specimens were derived from 36 patients. Of the 58 specimens, 17 were urine, 21 were sputum, 10 were abscess, 5 were pus of a decubitus ulcer, 4 were exudate, and 1 was gall. Among the 58 samples or specimens, 15 contained only E. faecalis and the other 43 specimens contained mixtures of isolates. Nine specimens contained coagulase-negative Staphylococcus, eight contained Pseudomonas aeruginosa, and five contained S. aureus. The others contained a variety of other bacterial species. Each specimen contained each bacterium at more than 10⁶ organisms per milliliter of specimen. At least five patients who had severe underlying disease (patients 8, 13, 24, 25, and 30) died from their infections. The complication for patients 8, 13, 24, and 25 was pneumonia, and the complication for patient 30 was a postoperative wound infection. Sputum from patient 8 contained E. faecalis and P. aeruginosa. The urine from patient 13 contained only gentamicin-resistant E. faecalis and the sputum contained P. aeruginosa, methicillin-resistant S. aureus, and E. faecalis. The sputum from patient 24 contained E. faecalis and P. aeruginosa. The sputum from patient 25 contained E. faecalis and E. faecium. The pus from patient 30 contained S. aureus, coagulase-negative S. aureus, and E. faecalis.

Pheromone response and the conjugative transfer of gentamicin resistance.

Each of the 37 gentamicin-resistant isolates from 36 patients was examined for its response to the culture filtrate of E. faecalis FA2-2. All of the strains exhibited a clumping response upon exposure to pheromone (E. faecalis FA2-2 culture filtrate), indicating that the isolates contained the pheromone-responsive plasmid. To examine the transferability of the gentamicin resistance trait, mating experiments were performed in broth. One isolate from each group was selected for the experiments. The gentamicin resistance of each isolate was transferred to the recipient E. faecalis FA2-2 strain at a frequency of about 10⁻³ to 10⁻¹ per donor cell (Table 1). When the transconjugants were selected on a plate containing kanamycin, kanamycin-resistant transconjugants were obtained from group B and group C isolates (Table 1). The EcoRI restriction profiles of plasmid DNAs isolated from the transconjugants are presented in Fig. 4. The EcoRI restriction endonuclease profiles of the plasmids isolated from the gentamicin-resistant transconjugants of group A and group D strains were identical to those of the wild-type isolates. The kanamycin and gentamicin resistance plasmids were identified in a group B isolate (Fig. 4, lanes 4 and 5). The kanamycin resistance plasmid was identified in a group C isolate (Fig. 4, lane 8). Plasmid DNAs isolated from the gentamicin-resistant transconjugants of group B and group D isolates exhibited the same EcoRI restriction profile (Fig. 4, lanes 5 and 10). Each of the transconjugants also exhibited a clumping response upon exposure to a culture filtrate of E. faecalis FA2-2.

TABLE 1.

Conjugative transfer of gentamicin or kanamycin resistance by broth mating^a

Group	Patient infected with strain	Drug resistance pattern	Selective drug	Transfer frequency (no. of trans- conjugants/per donor cell)	Drug resistance patterns of transconjugants (%)
A	1	Cm Em Gm Tc	Gm	1.6 × 10−3	Gm (100)
B	15	Cm Em Gm Km Sm Tc	Gm	4.8 × 10−1	Gm (100)
			Km	9.3 × 10−1	Gm Km (40)
					Km (60)
C	17	Cm Em Gm Km Sm Tc	Gm	1.7 × 10−2	Gm Km (100)
			Km	3.9 × 10−2	Gm Km (60)
					Km (40)
D	31	Cm Em Gm Sm Tc	Gm	2.7 × 10−1	Gm (100)

^aThe mating experiment was performed between the gentamicin-resistant donor strain and recipient FA2-2 (Rif^r Fus^r) in broth. The transconjugants were selected on agar plates containing gentamicin or kanamycin plus rifampin and fusidic acid for counterselection of the donor strain. Abbreviations: Cm, chloramphenicol; Em, erythromycin; Gm, gentamicin; Tc, tetracycline; Km, kanamycin; Sm, streptomycin. 

FIG. 4.

Agarose gel electrophoresis of EcoRI-digested plasmid DNAs isolated from a wild-type strain in each group and the transconjugants. Lanes: 1, group A strain; 2, Gm^r transconjugant of group A strain; 3, group B strain; 4, Km^r transconjugant of group B strain; 5, Gm^r transconjugant of group B strain; 6, group C strain; 7, Gm^r Km^r transconjugant of group C strain; 8, Km^r transconjugant of group C strain; 9, group D strain; 10, Gm^r transconjugant of group D strain.

The gentamicin-resistant transconjugants of groups A, B, and D and the kanamycin-resistant transconjugants of groups B and C underwent high-frequency transfer from E. faecalis FA2-2 to JH2SS (frequency, about 10⁻² to 10⁻¹). The group C kanamycin resistance plasmid was isolated, but a gentamicin resistance plasmid was not isolated even after repeated transfer experiments between FA2-2 and JH2SS (data not shown), suggesting that a nontransferable gentamicin resistance plasmid was mobilized by the transferable kanamycin resistance plasmid.

DISCUSSION

A total of 1,799 E. faecalis isolates were isolated from 1,412 patients in Gunma University Hospital between 1990 and 1996. Four hundred thirty-two of the 1,799 isolates had high-level gentamicin resistance. By using the drug resistance pattern and the restriction endonuclease digestion patterns of the plasmid DNA and the chromosomal DNA as epidemiological markers for strain identity, 81 of the high-level gentamicin-resistant strains were classified and were placed into four groups. Strains belonging to each of the four groups were isolated from individual patients in the three different wards. The three wards are located on different floors of the same building. The first surgical ward is located on the third floor, the second surgical ward is located on the fourth floor, and the internal medicine ward is located on the fifth floor. The same nurses work on a single ward. These results suggest that the isolates in each group were derived from a common source and were spread from patient to patient. At present, we do not know the method of transient carriage in the nosocomial transmission of the gentamicin-resistant isolates.

Many reports have described the nosocomial transmission of enterococcus. Among the early studies, the report of Zervos et al. (28, 29) demonstrated the patient-to-patient transmission and interhospital spread of a high-level gentamicin-resistant E. faecalis strain. The E. faecalis strain with high-level gentamicin resistance, which was a common phenotype in their study hospital, was recovered from the environment and from the hands of the health care personnel (28, 29), which suggested that transient carriage of the organism on the hands might be the mode of transmission (29). High-level gentamicin-resistant E. faecalis is also common in Japan (12, 14, 24). Of the 1,412 E. faecalis strains which were isolated between 1992 and 1996 in Gunma University Hospital, 315 (22.3%) strains had high-level gentamicin resistance. The frequency of isolation of high level-gentamicin-resistant E. faecalis strains in Gunma University Hospital has increased, and such strains accounted for 17.0% of E. faecalis strains in 1992 and for 30% in 1996 (unpublished data). To our knowledge, there has been no report describing nosocomial enterococcus infections in Japan.

It has been shown that infection with high-level gentamicin-resistant E. faecalis is associated with prior antimicrobial therapy, perioperative antibiotic prophylaxis, presurgical procedures, and longer hospitalizations (28, 29). In our study, all of the patients were hospitalized for more than 2 months and were administered antibiotics (data not shown). Twenty-seven of the 36 patients were in the surgical wards and had had surgery.

There have been reports that the pheromone-responsive plasmids of E. faecalis encode gentamicin resistance (6, 10, 14, 19, 24). In this study, pheromone-responsive plasmids which encode high-level gentamicin resistance were identified among strains in groups A, B, and D, and the plasmids transferred to a recipient strain at a high frequency by broth mating. There is a possibility that the pheromone-responsive plasmids might play a role in the spread of the gentamicin resistance in clinical E. faecalis isolates.

The group B isolates harbored two plasmids. One plasmid encoded kanamycin resistance and the other plasmid encoded gentamicin resistance. The group D isolates harbored one plasmid which encoded gentamicin resistance. The gentamicin resistance plasmids of group B and group D isolates had the same EcoRI restriction endonuclease profiles. The restriction endonuclease digestion patterns of the chromosomal DNAs of group B and group D isolates were also identical. These results imply that the group D isolate resulted from the segregation of the gentamicin resistance plasmid from the group B isolate or that the group B isolate resulted from the transfer of the kanamycin resistance plasmid from another strain to a group D isolate.