Human and swine hosts share vancomycin-resistant Enterococcus faecium CC17 and CC5 and Enterococcus faecalis CC2 clonal clusters harboring Tn1546 on indistinguishable plasmids

Abstract

VRE isolates from pigs (n = 29) and healthy persons (n = 12) recovered during wide surveillance studies performed in Portugal, Denmark, Spain, Switzerland, and the United States (1995 to 2008) were compared with outbreak/prevalent VRE clinical strains (n = 190; 23 countries; 1986 to 2009). Thirty clonally related Enterococcus faecium clonal complex 5 (CC5) isolates (17 sequence type 6 [ST6], 6 ST5, 5 ST185, 1 ST147, and 1 ST493) were obtained from feces of swine and healthy humans. This collection included isolates widespread among pigs of European Union (EU) countries since the mid-1990s. Each ST comprised isolates showing similar pulsed-field gel electrophoresis (PFGE) patterns (≤6 bands difference; >82% similarity). Some CC5 PFGE subtype strains from swine were indistinguishable from hospital vancomycin-resistant enterococci (VRE) causing infections. A truncated variant of Tn1546 (encoding resistance to vancomycin) and tcrB (coding for resistance to copper) were consistently located on 150- to 190-kb plasmids (rep(pLG1)). E. faecium CC17 (ST132) isolates from pig manure and two clinical samples showed identical PFGE profiles and contained a 60-kb mosaic plasmid (rep(Inc18) plus rep(pRUM)) carrying diverse Tn1546-IS1216 variants. The only Enterococcus faecalis isolate obtained from pigs (CC2-ST6) corresponded to a multidrug-resistant clone widely disseminated in hospitals in Italy, Portugal, and Spain, and both animal and human isolates harbored an indistinguishable 100-kb mosaic plasmid (rep(pRE25) plus rep(pCF10)) containing the whole Tn1546 backbone. The results indicate a current intra- and international spread of E. faecium and E. faecalis clones and their plasmids among swine and humans.

Fig. 1.

Computer analysis of SmaI-digested genomic DNAs of representative vancomycin-resistant E. faecium (A) and E. faecalis (B) isolates from humans and swine. PFGE subtypes were established by visual analysis following the criteria of Tenover et al. (48). Computer analysis was performed with the Fingerprinting II Informatix software package (Bio-Rad Laboratories, Hercules, CA). The acquisition of the image was performed by using a Gel Doc XR camera (Bio-Rad Laboratories Inc.). The image was normalized by using four reference lanes of the low-range PFGE marker (0.13 kb to 1,018.5 kb; New England BioLabs, La Jolla, CA) as an external DNA size marker. The phylogenetic tree was subsequently constructed by use of the Dice coefficient and UPGMA clustering (optimization, 0.5%; band tolerance, 1.5%; threshold cutoff value set at 82%) (36). Year, year of isolation.

Fig. 2.

ClaI-digested plasmid DNAs of VRE isolates recovered from hospitals and swine from different origins. Lane M, molecular marker ladder EcoT 14 I/BglII (Takara Bio Inc., Otsu, Japan); lane 1, E. faecalis ST6-CC2 (PFGE B5) from an outbreak strain recovered in Spain in 1999; lane 2, E. faecalis ST6-CC2 (PFGE B5) from a swine isolate recovered in Portugal in 2007; lane 3, E. faecium ST132-CC17 (PFGE 119.5) from a swine isolate recovered in Portugal in 2007; lane 4, E. faecium ST132-CC17 (PFGE 119) from a clinical isolate recovered in Portugal in 2002 that is representative of strains causing urinary tract infections in two unrelated patients during 2002.