Genotypic Diversity, Antimicrobial Resistance, and Virulence Factors of Human Isolates and Probiotic Cultures Constituting Two Intraspecific Groups of Enterococcus faecium Isolates

Bacterial isolates.A total of 128 E. faecium isolates were collected in the framework of a European Union-funded research project, PROSAFE (48), including 37 isolates from human feces, 79 isolates from different human clinical samples, and 12 isolates commercially used as probiotics, 6 of which were isolated from products with probiotic claims (41) while the remaining 6 were received directly from the manufacturer or the depositor (Table 1). A representative subset of these isolates has been deposited in the BCCM/LMG Bacteria Collection (Ghent University, Ghent, Belgium [http://www.belspo.be/bccm/lmg.htm]). The isolates were routinely grown on Columbia agar (Becton-Dickinson, Sparks, MD) supplemented with 5% defibrinated horse blood at 37°C for 24 h.

Identification and typing.For identification purposes, the isolates were first subjected to fluorescent AFLP (FAFLP) analysis as described below.

For FAFLP analysis, total chromosomal DNA was prepared using a modification of the method described by Pitcher et al. (31). Template preparation was carried out as described previously (19). Essentially, purified genomic DNA was digested by two restriction enzymes, a 4-bp cutter EcoRI and a 6-bp cutter TacI. Small double-stranded DNA molecules (adaptors; 15 to 20 bp) containing one compatible end were ligated to the corresponding “sticky ends” of the restriction fragments. These adaptors served as binding sites for selective amplification with the primer combination E01/T01 (primers extended with an additional A) (19). PCR products were separated according to their lengths on a high-resolution polyacrylamide gel using a DNA sequencer (ABI 377). Fragments that contained an adaptor specific for the restriction half-site created by the 6-bp cutter were visualized due to the 5′ end labeling of the corresponding primer with the fluorescent dye 6-carboxyfluorescein. The resulting electrophoretic patterns were numerically analyzed with Bionumerics software version 4.01 (Applied Maths, Belgium), using the Dice coefficient and unweighted-pair group method using average linkages cluster analysis.

The intraspecific diversity among the collection of E. faecium isolates was also studied by PFGE, as previously described (7). Briefly, bacterial cells from an overnight culture were embedded in low-melting-point preparative agarose (Bio-Rad Laboratories, Nazareth, Belgium). After cell wall and protein digestion, the plugs were digested overnight with 30 U of SmaI (MBI Fermentas, St. Leon-Rot, Germany) at 25°C. PFGE was performed with a 1% agarose gel by using a CHEF Mapper apparatus (Bio-Rad Laboratories) in 0.5× Tris-borate-EDTA buffer at 14°C at 6 V/cm. A linearly ramped switching time from 5 to 35 s was applied for 24 h. The DNA band profiles were stained with ethidium bromide, and the image was digitized with the Gel Doc 1000 System (Bio-Rad Laboratories). Conversion, normalization, and further analysis of the DNA band patterns were performed using GelCompar software version 4.0b (Applied Maths, Kortrijk, Belgium) as described previously (33). Similarity between PFGE patterns was expressed using the Dice band-based correlation coefficient.

Phenotypic and genotypic characterization of antimicrobial susceptibility.Antimicrobial susceptibility testing of the isolates was performed using broth microdilution following CLSI guidelines (5) to determine the MICs of the following agents (with the concentration ranges [mg/liter] tested given in parentheses): penicillin (0.032 to 64), ampicillin (0.032 to 64), ampicillin/sulbactam (sulbactam was tested as a fixed concentration of 8 mg/liter; 0.032 to 64), vancomycin (0.125 to 256), teicoplanin (0.125 to 256), gentamicin (1 to 2,048), streptomycin (2 to 4,096), erythromycin (0.016 to 32), clindamycin (0.032 to 32), quinupristin-dalfopristin (Q/D; tested as a 30:70 ratio; 0.032 to 64), oxytetracycline (0.063 to 128), chloramphenicol (0.125 to 256), and linezolid (0.016 to 32). For cotrimoxazole (tested as a 1:19 ratio; 0.25 to 512), MIC breakpoints according to CLSI guidelines for Staphylococcus (5) were used. For fusidic acid (0.063 to 128), breakpoints as defined by Toma and Barriault (43) were used, while for trimethoprim (0.25 to 512), European Food Safety Authority guidelines (9) were followed. According to the CLSI guidelines, enterococci are considered naturally resistant to clindamycin. Occasionally, however, strains with MICs in the susceptibility range have been observed (personal observation). Except for sulbactam and linezolid (Pfizer), teicoplanin and Q/D (Sanofi-Aventis), and erythromycin (Abbott), all tested antibiotics originated from Sigma.

For isolates displaying phenotypic resistance to one or more of the tested agents, the presence of the following acquired (and potentially transferable) resistance genes was investigated: the tet(M), tet(L), tet(K), tet(O), tet(P), tet(Q), tet(T), tet(S), tet(W), and tet(M) group; van(A), van(B), erm(A), erm(B), erm(C), cat(pC194), cat(piP501), aad(E), aph(2″)-aac(6′), aad(E)-aph(A), vat(D), and vat(E). For this purpose, DNA was isolated using the DNeasy tissue kit (Qiagen), and amplification of the corresponding gene fragments was performed in a DNA Engine Thermal Cycler “PTC-200” (MJ Research), as previously described (22). The following positive control strains were used: E. faecium UW 1342 for vat(D) (52); E. faecium UW 1965 for vat(E), erm(B), aad(E), aad(E)-aph(A), cat(pC194), and cat(piP501) (50, 51, 53, 55); Staphylococcus aureus 694/01 for erm(A), erm(C), tet(K), tet(M), and aac(A)-aph(D) (39); E. faecium UW 1873 for tet(L) (54); E. faecium BM4147 for van(A) (3, 29); E. faecium V583 for van(B) (29, 35); and Streptococcus pyogenes A498 for tet(T) (4). For the remaining genes, the following control plasmids were used: pGEM-tet(O) for tet(O) (2); pJIR667 for tet(P) (23a); pBT-1 for tet(Q) (27); pVP2 for tet(S) (30); and pGEM-tetW for tet(W) (2). The amplification products were detected by electrophoresis in a 1.5% agarose gel and subsequent ethidium bromide staining.

In vitro transfer experiments were performed by conjugation (filter mating) as previously described (22). Possible transconjugants were identified in several steps, selecting for the selective and nonselective markers. Probiotic E. faecium isolates representing non-wild-type isolates with acquired antibiotic resistance(s) were used as donors, whereas the well-documented strains E. faecium 64/3 and E. faecalis JH2-2 were chosen as recipients. Possible transconjugants were further characterized by MIC determination, PCR-based detection of resistance genes, PFGE, and (GTG)₅-PCR (15, 21).

Multiplex PCR virulence genes.Multiplex PCR for the detection of the virulence genes asa1, gelE, cylA, esp, and hyl was performed as described previously (49). Briefly, each 50-μl PCR mixture consisted of 5 μl of bacterial suspension; 0.1 μM primers for the detection of asa1, gelE, and hyl; 0.2 μM primers for the detection of cylA and esp; 25 μl HotStarTaq Master Mix (Qiagen, Germany); and an additional 1.0 mM MgCl₂. PCR was performed in a GeneAmp PCR System 9600 (Perkin Elmer, Wellesley, MA). An initial activation step at 95°C for 15 min, during which the HotStarTaq DNA polymerase was activated, was followed by 30 cycles of denaturation (94°C for 1 min), annealing (56°C for 1 min), and extension (72°C for 1 min), followed by 1 cycle consisting of 10 min at 72°C. The PCR products were electrophoresed in a 1.5% pronarose D1 gel (SphaeroQ, Burgos, Spain) for 1 h at 150 V in 0.5× Tris-borate-EDTA containing 0.05 mg/liter ethidium bromide (positive and negative controls were included in each set of amplifications) (49). A 100-bp DNA ladder (Invitrogen, Merelbeke, Belgium) was used as molecular size marker.

Statistical analysis.Student's t test was used for statistical analysis. A P value of <0.05 was considered statistically significant.

Identification and typing.All strains were identified at the species level as E. faecium by FAFLP (data not shown). The dendrogram obtained from numerical analysis of digitized FAFLP generated with primer combination E01/T01 clearly showed the presence of two intraspecific genomic groups (denoted I and II) among the set of 128 E. faecium isolates. The delineation of these two groups is shown for a selection of isolates in Fig. 1. A total of 87 E. faecium isolates belonged to FAFLP group I, whereas 41 E. faecium isolates belonged to FAFLP group II. FAFLP group I consisted of 86 isolates of human origin (25 fecal and 61 clinical) and 1 probiotic culture. In FAFLP group II, 30 isolates were of human origin (12 fecal and 18 clinical), and 11 isolates were received as probiotic cultures.

• Open in new tab
• Download powerpoint

FIG. 1.

Dendrogram based on numerical analysis of FAFLP patterns obtained with the primer combination E01/T01 with the corresponding SmaI PFGE patterns of selected E. faecium isolates. The dendrogram was constructed using the unweighted-pair group method using average linkages and the band-based Dice similarity coefficient. P, probiotic; PH, probiotic human consumption; PHA, probiotic human and animal consumption; HB, human blood isolate; HC, human clinical isolate; HF, human fecal isolate.

Cluster analysis and visual inspection of the PFGE profiles revealed high variability within the two genomic FAFLP groups. Based on the criterion that isolates exhibiting a maximum of six band position differences (42) in their respective PFGE patterns belonged to the same PFGE group, a total of 25 PFGE groups containing more than 1 isolate and 59 single isolates were recognized. In case no differences in band number and position were observed upon visual inspection, isolates were considered indistinguishable. The PFGE patterns of a selection of isolates are shown in Fig. 1. A total of 13 groups (i.e., groups 13, 17, 19, 21, 24, 25, 26, 31, 32, 37, 50, 51, and 98) contained only clinical isolates, 3 groups (i.e., groups 8, 20, and 56) contained only fecal isolates, and 5 groups (i.e., groups 6, 9, 22, 23, and 95) contained both clinical and fecal isolates. High similarity was observed between a number of isolates used commercially as probiotics. These isolates belonged to PFGE groups 57, 62, 63, and 76, of which group 63 contained six isolates received from five different depositors (i.e., D04, D05, D27, D42, and D50) with indistinguishable SmaI macrorestriction profiles (i.e., PRSF-E133, PRSF-E132, PRSF-E164, PRSF-E105, PRSF-E117, and PRSF-E107). In PFGE groups 62 and 57, a probiotic isolate clustered with a clinical isolate (i.e., PRSF-E103 with PRSF-E093 and PRSF-E104 with PRSF-E086) but could not be considered indistinguishable based on two band differences (Fig. 1). In PFGE group 76, fecal isolate PRSF-E051 clustered with the two probiotic isolates PRSF-E106 and PRSF-E134, but these isolates were also not considered indistinguishable (Fig. 1).

Phenotypic and genotypic characterization of antimicrobial susceptibilities.Using broth microdilution, phenotypic resistances to 13 antimicrobial agents were determined, after which the genetic basis of the observed resistance was investigated (Table 2). All isolates were susceptible to linezolid, but only 15 (12%) out of 128 isolates were susceptible to all agents tested. Overall, the highest phenotypic resistance frequencies were observed for erythromycin (62 of 128 isolates; 48%) and oxytetracycline (59 of 128 isolates; 46%). In Fig. 2, the distribution of antimicrobial resistances in FAFLP groups I and II is depicted. Resistance to gentamicin, penicillin, vancomycin, and teicoplanin was observed only in group I, and at least twice as many isolates (P < 0.001) in group I were resistant to the other tested antibiotics (Q/D, ampicillin, ampicillin/sulbactam, streptomycin, erythromycin, trimethoprim, cotrimoxazole, oxytetracycline, and chloramphenicol) compared to the isolates in group II. On the other hand, a larger number of group II isolates (31 out of 41 isolates; 76%) were resistant to fusidic acid compared to group I (50 out of 87 isolates; 57%), although the difference was not statistically significant (P = 0.09).

• Open in new tab
• Download powerpoint

FIG. 2.

Antibiotic resistances in FAPLP groups I and II. PEN, penicillin; AMP, ampicillin; ASU, ampicillin-sulbactam; GEN, gentamicin (only high-level resistance was reported); STR, streptomycin (only high-level resistance was reported); VAN, vancomycin; TPL, teicoplanin; ERY, erythromycin; FUS, fusidic acid; TMP, trimethoprim; SXT, cotrimoxazole; OTE, oxytetracycline; CMP, chloramphenicol.

The majority of the isolates (87 out of 113 isolates; 77%) displaying phenotypic resistance also possessed the corresponding antibiotic resistance gene (Table 2). Overall, out of the 59 tetracycline-resistant E. faecium isolates, the tet(M) and tet(L) genes were detected in 57 (97%) and 41 (69%) human isolates, respectively. Out of the 53 tetracycline-resistant isolates in group I, 51 (96%) carried the tet(M) gene and 37 (69%) carried the tet(L) gene. In group II, all six tetracycline-resistant isolates carried the tet(M) gene and four (67%) carried the tet(L) gene. The tet(K) gene was detected in only one human isolate belonging to group I, which also harbored the tet(L) gene. In all isolates displaying phenotypic resistance to chloramphenicol, streptomycin, and gentamicin, corresponding genes could be detected (Table 2). On the other hand, 11 nonfecal isolates and three fecal isolates possessed a cat gene, while none of these were classified as phenotypically resistant to chloramphenicol. All 18 vancomycin- and teicoplanin-resistant isolates of group I carried the van(A) gene. In addition, one isolate with phenotypic resistance to vancomycin but not to teicoplanin also contained the van(B) gene. The following genes were not detected: the tet(O), tet(P), tet(Q), tet(T), tet(S), tet(W), and tet(M) group; erm(C); and vat(E).

Among the probiotic isolates, 11 out of 12 (92%) displayed phenotypic resistance to fusidic acid. Two out of these 11 isolates, PRSF-E105 and PRSF-E140, were also resistant to erythromycin, but only isolate PRSF-E140 carried an erm(B) gene. Using filter mating, transfer of erm(B) from PRSF-E140 to recipients E. faecium 64/3 and E. faecalis JH2-2 could not be detected (data not shown).

Multiplex PCR virulence genes.A total of 12 out of 128 (9%) E. faecium isolates were positive for the enterococcal surface protein gene esp, including the fecal isolate PRSF-E149 from a hospitalized patient, which also contained the hyaluronidase gene, hyl. Of the remaining 11 (9%) esp-positive E. faecium isolates, 7 were blood isolates and 4 were wound isolates. More specifically, a total of 7 out of 87 (8%) isolates in FAFLP group I demonstrated the presence of the esp gene, of which 1 isolate also harbored the hyl gene, whereas in group II, the number of isolates positive for esp was slightly higher (5 out of 41; 12%). None of the probiotic isolates possessed any of the virulence factors tested. Table 2 gives an overview of the presence of virulence genes in the E. faecium isolates tested.