OUP user menu

Antibiotic resistance and genotypic characterization by PFGE of clinical and environmental isolates of enterococci

Giordano Dicuonzo, Giovanni Gherardi, Giulia Lorino, Silvia Angeletti, Fabrizio Battistoni, Lucia Bertuccini, Roberta Creti, Roberta Di Rosa, Mario Venditti, Lucilla Baldassarri
DOI: http://dx.doi.org/10.1111/j.1574-6968.2001.tb10758.x 205-211 First published online: 1 July 2001


Fifty-four Enterococcus faecalis and 20 Enterococcus faecium isolates from clinical and non-human sources in Rome, Italy, were characterized by antibiotic resistance and pulsed field gel electrophoresis (PFGE). Resistance to vancomycin, teicoplanin, ampicillin, and ciprofloxacin was more frequent in E. faecium than in E. faecalis, whereas high-level resistance to aminoglycoside was found primarily in E. faecalis. Multi-resistance was found primarily among clinical isolates, but was also observed among environmental isolates. Common genotypes shared among clinical and environmental isolates were observed, however, the majority of isolates occurred as unique, source-specific clones. Several PFGE types were associated with shared features in their antibiotic resistance patterns; evidences of clonal spread between and within wards were also noted. This is the first report indicating clonal relatedness between human and environmental enterococci isolated in Italy.

  • Pulsed field gel electrophoresis
  • Antibiotic susceptibility
  • Enterococcus

1 Introduction

Enterococcus faecalis and Enterococcus faecium account for greater than 95% of enterococcal infections in humans [1]. Enterococci are listed as the third/fourth cause of nosocomial infections [2] and there has been a rapid increase of glycopeptide and high-level aminoglycoside-resistant strains [3,4]. Fortunately, vancomycin-resistant enterococci still remain scarce in Italy [5]. Enterococci are also commonly isolated from non-human sources [6]. Elucidating the genetic relationships between human and non-human enterococcal isolates provides valuable information concerning the epidemiology of enterococcal infections. Pulsed field gel electrophoresis (PFGE) has gained wide acceptance for establishing clonal relatedness within many bacterial species including E. faecalis and E. faecium [79].

In this study we determined the prevalence of different species of clinical and environmental enterococcal isolates in Rome, Italy. We compared the antibiotic resistance pattern and PFGE profiles of these strains and documented numerous clonal types, some of which were found both in human and environmental isolates. Certain clonal types were associated with common features in drug resistance patterns.

2 Materials and methods

2.1 Isolates

Sixty-four isolates of enterococci were recovered from 20 different wards of three hospitals in Rome during a 3-year period (1997–1999) and 26 isolates were recovered from environmental sources in the same period. The initial isolation from environmental samples was done on bile-esculin azide agar incubated at 35–37°C for 24 h in 5% CO2; three colonies were randomly picked from each plate and identified to species level. Five out of the 64 human isolates were from the normal flora of healthy individuals, three being isolated from the oral cavity of healthy carriers, and two being part of the vaginal flora. The remaining 59 strains causing disease were from urinary tract infections (UTI) (12 isolates), from biliary stents (21), from the blood or endovascular devices of patients with endocarditis or endovascular infections (19), from surgical wound infections (two), from abscess (1), from ascites (one), from urinary catheter (one), and from a subhepatic drain (two). The 26 environmental isolates were recovered from seawater (eight), wastewater (12), and wells (six) in the same city. Seawater specimens were collected far from wastewater outlet. Wastewater samples were in the catchment area of the hospitals considered.

2.2 Identification

Isolates were identified to the species level according to standard biochemical tests. In cases of uncertain identification by the routine phenotypic tests PCR was performed using E. faecalis- and E. faecium-specific primers as previously described [10]. PCR results were resolved by electrophoresis on a 2% agarose gel containing 0.5 μg μl−1 of ethidium bromide.

2.3 Antimicrobial susceptibility tests

Fifty-four E. faecalis and 20 E. faecium isolates from clinical and environmental sources were tested by using the Kirby–Bauer or broth microdilution system according to Swenson et al. [11,12]. Antibiotics tested were ampicillin, vancomycin, teicoplanin, tetracycline, ciprofloxacin, chloramphenicol; high-level resistance (HLR) to gentamicin and streptomycin was also assessed. Isolates showing intermediate levels of susceptibility were classified as resistant.

2.4 PFGE analysis

Chromosomal SmaI restriction patterns were determined as previously described [13] using the 74 isolates for which antibiotic susceptibility tests were performed. Interpretative criteria of strain relatedness were those followed by Tenover et al. [14].

2.5 Cluster analysis

PFGE types were analyzed with Bionumerics software for Windows, version 2.5 (Applied Maths). The DNA banding patterns were normalized with lambda concatemer ladder standards. Comparison of the banding patterns was performed by the unweighted pair group method with arithmetic averages and with the Dice similarity coefficient. A tolerance of 1.5% in band position was applied during comparison of the DNA patterns.

3 Results and discussion

The majority of clinical isolates were recovered from UTI (12 isolates), from biliary stents (21 isolates), and from patients with endocarditis or endovascular infections (19 isolates). Among the 64 human isolates of enterococci the most prevalent species found was represented by E. faecalis (Table 1); of the 26 environmental isolates, E. faecalis, E. faecium, and Enterococcus hirae were the predominant species. PCR allowed species identification in 10 cases (five E. faecium and five E. faecalis) where biochemical tests were unclear (data not shown).

View this table:
Table 1

Identification of the 90 species of enterococci isolated from human and environmental sources

Source (no. of isolates)SpeciesNo. of isolates
Human isolates (64)E. faecalis48
E. faecium14
Enterococcus gallinarum1
Enterococcus avium1
Environmental isolates (26)E. faecalis8
E. faecium8
E. hirae7
Enterococcus casseliflavus2
Enterococcus durans/hirae1

Table 2 summarizes the antibiotic resistance profiles found among the E. faecalis and E. faecium strains analyzed in this study. Only four isolates (two E. faecalis and two E. faecium) were found to be vancomycin-resistant, three of these being human isolates. The environmental E. faecium resistant to vancomycin was also the only teicoplanin-resistant strain found in this survey; it is noteworthy that this strain was also multi-drug-resistant. Our data confirm the low incidence of glycopeptide resistance in Italy among clinical enterococci, and are in agreement with results described in a previous surveillance study [5].

View this table:
Table 2

Antibiotic-resistant profiles among clinical and environmental E. faecalis and E. faecium isolates

Species (no. of isolates)Antibiotica(No. of resistant isolates)(%)Clinical isolatesbEnvironmental isolatesc
E. faecalis (54)Va(2)(4)2400
E. faecium (20)Va(2)(10)18114
  • aAccording to National Committee for Clinical Laboratory Standards (NCCLS) definitions. Va, vancomycin; Tec, teicoplanin; Amp, ampicillin; Te, tetracycline; Cip, ciprofloxacin; Chl, chloramphenicol; HLSm and HLGm, HLR to streptomycin and gentamicin, respectively. Isolates showing intermediate levels to antibiotics were classified as resistant.

  • bForty-six and 13 clinical isolates of E. faecalis and E. faecium, respectively, were tested for antimicrobial resistance.

  • cEight and seven environmental isolates of E. faecalis and E. faecium, respectively, were tested for antimicrobial resistance.

Multi-drug-resistant isolates were found from both human and environmental sources (Tables 2 and 3). Three ampicillin-resistant environmental isolates, all genomically unrelated E. faecium strains, were found. The results of this survey indicate that multi-drug resistance is common among clinical isolates of enterococci, and in agreement with a previous report [15] it is found less frequently among a variety of non-clinical human and environmental aquatic sources. Interesting to note that four out of seven E. faecium aquatic isolates were resistant to at least three different antibiotics versus only two of eight E. faecalis multi-resistant environmental isolates (Table 3). We did not observe a high prevalence of HLR among environmental isolates, in contrast to a previous report [16]; however, it must be underlined that the number of strains examined in this study is much smaller than that in the referenced study.

View this table:
Table 3

Genotypic characterization of 54 E. faecalis (EFS) and 20 E. faecium (EFM) isolates in correlation to the site of infection, ward, and antibiotic resistance profiles

No. sampleSite of infectionWardPFGE patterndAntibiotypea
2-EFSbiliary stentI1TeR, CipR, ChlR, GmR, SmR
59-EFSbiliary stentI1AmpR, TeR, ChlR, GmR, SmR
3-EFSbiliary stentI1.1AmpR, TeR, CipR, ChlR, GmR, SmR
7-EFSbiliary stentI1.2TeR, CipR, ChlR, GmR, SmR
8-EFSbiliary stentI1.3sc
9-EFSbiliary stentI1.3s
40-EFSpharingeal swabII (carrier)2TeR, CipR
17-EFSUTIbII3CipR, ChlR,
4-EFSbiliary stentI4CipR, ChlR, GmR
21-EFSUTIII4.1TeR, ChlR, GmR
10-EFSbiliary stentI5s
27-EFSvaginaI (carrier)6s
11-EFSbiliary stentI7CipR
30-EFSoral tractcarrier9s
31-EFSoral tractcarrier9s
29-EFSwastewaterenvironment9.1TeR, CipR, ChlR, SmR
13-EFSbiliary stentIV9.2AmpR, TeR, CipR
44-EFSvaginaV (carrier)9.3TeR, ChlR
47-EFSV–P shuntbVI9.3TeR, CipR
12-EFSbiliary stentI9.4s
52-EFSbloodVIII11TeR, ChlR, VaR
38-EFSUTII14.1TeR, CipR, ChlR
20-EFSUTIII14.2TeR, CipR, ChlR, SmR
42-EFSascitisXI14.3TeR, CipR, ChlR, GmR, SmR
45-EFSbloodXII15AmpR, CipR
15-EFSUTIII15.1TeR, CipR
19-EFSUTIII16TeR, ChlR, GmR, SmR
26-EFSAVIbXIII16AmpR, TeR, CipR, GmR, SmR
64b-EFSbiliary stentI16TeR, CipR, ChlR, GmR, SmR
36-EFSwoundXIV17TeR,CipR, ChlR, GmR, SmR
39-EFSbloodXIV17TeR, CipR, ChlR, GmR, SmR
55-EFSbloodVIII17.1TeR, CipR, ChlR, GmR, SmR
43-EFSUTIXV17.2TeR, CipR, ChlR, GmR, SmR
51-EFSbloodXII17.3AmpR, TeR, CipR, ChlR, GmR, SmR
28-EFSwastewaterenvironment17.4TeR, CipR, ChlR
1-EFSbiliary stentI18TeR, CipR, ChlR, GmR, SmR
23-EFSbloodXVII18TeR, CipR, ChlR, GmR, SmR
65b-EFSsubhepatic drainXIX18TeR, CipR, ChlR, GmR, SmR
18-EFSUTIII20TeR, CipR, ChlR
69-EFSUTIXXI23AmpR, TeR, CipR, ChlR
5-EFMbiliary stentIAAmpR, TeR, CipR
63-EFMbiliary stentIAAmpR, TeR, CipR
6-EFMbiliary stentIBAmpR, TeR, GmR, SmR
64a-EFMbiliary stentIE2AmpR, CipR
65a-EFMsubhepatic drainXIXE3AmpR, TeR, CipR, ChlR, SmR
78-EFMbloodXIXE4AmpR, TeR, CipR, SmR
60-EFMbiliary stentIFCipR, GmR
61-EFMbiliary stentIFCipR
62-EFMbiliary stentIGVaR, CipR
66-EFMwellenvironmentHTeR, CipR
70-EFMwastewaterenvironmentICipR, ChlR
72-EFMseawaterenvironmentKVaR, TecR, AmpR, TeR, CipR, ChlR
73-EFMseawaterenvironmentLAmpR, CipR, ChlR
74-EFMseawaterenvironmentMAmpR, TeR, CipR, ChlR
75-EFMwastewaterenvironmentNTeR, CipR, ChlR
76-EFMbloodVIIIOAmpR, TeR, CipR, ChlR
  • aAccording to NCCLS definitions. Te, tetracycline; Cip, ciprofloxacin; Chl, chloramphenicol; Va, vancomycin; Amp, ampicillin; Gm and Sm, high levels to gentamicin and streptomycin, respectively. R, resistance; S, susceptibility.

  • bUTI, urinary tract infection; AVI, aortic valve infection; VDI, vascular device infection; V–P shunt, ventricular–peritoneal shunt.

  • cs, susceptible to all antibiotics.

  • dIsolates differing by ≤6 bands from subtype 1 for each group (type) were assigned a common type.

Fifty-four E. faecalis isolates and 20 E. faecium isolates were genotyped by PFGE to explore clonal relatedness of strains from environment and humans (Figs. 1 and 2). Among molecular methods used for subtyping bacterial species, PFGE is considered one of the most reliable due to its discriminatory power, sensitivity and reproducibility [79]. Several reports described molecular typing methods useful for epidemiologic surveillance on vancomycin-resistant enterococci [17,18]; the same methods, other than PFGE, proved to be non-appropriate or less reliable for accurate clonality studies on vancomycin-sensitive enterococci [79,17,18]. For these reasons we chose PFGE to genetically compare our clinical and environmental isolates of enterococci which only rarely showed glycopeptide resistance. Wide genotypic variability was found among the clinical and environmental isolates of E. faecalis and E. faecium. Twenty-four different PFGE types and 42 subtypes among E. faecalis strains, and 15 pulsed types with 18 subtypes for E. faecium isolates were identified (Figs. 1 and 2, Table 3). The majority of isolates clustered as source-specific clones, although examples of common PFGE types shared among clinical and environmental E. faecalis isolates occurred.

Figure 1

Genetic relatedness of the 54 E. faecalis examined, based on the PFGE banding patterns of the isolates. Strain codes and PFGE subtypes are depicted. Cases I–X indicate multiple-strain genetic clusters encountered among study E. faecalis. Environmental isolates are marked with an asterisk.

Figure 2

Genetic relatedness of the 20 E. faecium studied. Three multiple-strain genetic groups, defined as I, II, and III, were found. Strain codes, PFGE types, and dendrogram with percentages of similarity are also reported. Environmental isolates are marked with an asterisk.

Among E. faecalis isolates, 10 of the 24 PFGE types were shared between at least two or more isolates and 14 unique PFGE types were found (Fig. 1). All the 46 E. faecalis clinical isolates grouped into one of 19 PFGE types, whereas eight different genotypes characterized the eight environmental isolates (Fig. 1, Table 3). Types 3, 9, and 17 were shared between clinical and environmental E. faecalis isolates, with one isolate recovered from wastewater sharing the identical PFGE subtype 3 with two urinary isolates (Fig. 1, Table 3). This suggests the spread of environmental strains into human facilities and vice versa. Twenty-one out of 46 E. faecalis isolates recovered from humans clustered in only four PFGE types (1, 9, 14, 17), representing 44% of all human E. faecalis strains examined (Fig. 1, Table 3).

More genomic variability was seen among the 20 E. faecium isolates that were subjected to PFGE (Table 3). This was especially apparent with environmental isolates. Each of the seven environmental E. faecium isolates displayed a unique PFGE profile, while eight PFGE types were shared among the 13 clinical E. faecium isolates (Fig. 2, Table 3).

Consistencies of antibiotic resistance patterns were observed within several PFGE types. Among E. faecalis, a majority of isolates within PFGE types 1, 16, 17, and 18 were characterized by HLR, and also were resistant to tetracycline, ciprofloxacin, and chloramphenicol. PFGE types 3, 9, and 15 were uniformly sensitive to high-level aminoglycosides (Table 3).

Interestingly, PFGE type 1 was exclusive of biliary stent group, and accounted for six of the 13 isolates (Table 3); as all six isolates were recovered from patients hospitalized in the same ward, undergoing the same invasive procedure, this is suggestive of a ward-associated clonal spread [19]. E. faecalis isolates recovered from UTI were mainly PFGE types 3 and 14 (Table 3). Several examples of isolates within the same PFGE type from different wards of the same or different hospitals were found.

To our knowledge, this is one of the few studies examining the clonal relatedness between clinical and environmental enterococci. Little data have been reported from Italy regarding antibiotic resistance of enterococci from environmental sources; genetic relationship of Italian clinical and environmental enterococci has not been previously described. We have found genetically related clusters of strains from different settings. Establishing environmental origins and direction of spread of such strains will require further investigation.


We gratefully acknowledge Dr. Bernard Beall, CDC, Atlanta, USA, for helpful comments and suggestions. We thank Manola Valente for the technical support in the cluster analysis. This work was partially supported by the Italian Ministry of Health, Project 1%‘Interactions between opportunistic pathogens and biomaterials in the pathogenesis of prosthesis infections’ (ICS 080.1/RS 98.43) to L.B., and by ‘Finanziamento progetti di Ateneo 60%’, University ‘La Sapienza’, Rome, Italy, to G.L.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
View Abstract