OUP user menu

Characterization of new insertion-like sequences of Enterococcus hirae and their dissemination among clinical Enterococcus faecium isolates

Leonardo A. Sechi, Richard Franklin, Ilaria Duprè, Stefania Zanetti, Giovanni Fadda, Lolita Daneo-Moore
DOI: http://dx.doi.org/10.1111/j.1574-6968.1998.tb12944.x 165-172 First published online: 1 April 1998


Sequence analysis of different fragments that hybridized with a 4.5-kb EcoRI fragment originally cloned from Enterococcus hirae ATCC 9790 showed 66% homology to IS-like sequences found in staphylococci and lactococci. We tested several enterococcal ATCC strains and found that only E. hirae ATCC 9790 and Enterococcus faecium ATCC 19434 hybridized with the IS-like sequence. Moreover, we wanted to investigate the dissemination of this new IS among E. faecium strains. We analyzed 131 clinical E. faecium isolated in Italy and the USA for the presence of the IS and we found its presence in more than 63% of the isolates. The hybridization patterns obtained vary considerably between unrelated strains and allow further classification among ribotype-grouped species.

Key words
  • Insertion-like sequence
  • Enterococcus
  • DNA fingerprinting
  • Staphylococcus

1 Introduction

Repeated chromosomal elements are present in the genomes of all bacteria [1]. Our previous studies have characterized the rrn organization in Enterococcus hirae and in several enterococcal ATCC strains [2, 3]; rRNA operons and insertion (IS) sequences are the major recurrent chromosomal elements in prokaryotes. A 4.5-kb insert was previously found to hybridize with six differently sized fragments obtained from the chromosomal DNA of E. hirae digested with EcoRI [4]. In the Enterococcus genus only few IS sequences have been reported so far [58]. One of these IS sequences has been used to identify the clonal relationship of clinical isolates [8]. Here we report on the presence in enterococci of a new sequence homologous to IS1272 of Staphylococcus haemolyticus and its possible use in the molecular epidemiology of Enterococcus faecium strains [9, 10].

2 Materials and methods

2.1 Bacterial strains

E. faecium strains were obtained from different hospitals; 24 strains were isolated on the east coast of the United States in four different hospitals: five from St. Christopher's Hospitals for Children, Philadelphia (SC); four from the Children's Hospital of Philadelphia (CH); 14 from Thomas Jefferson University Hospital, Philadelphia (TJ); and one from New York University Hospital, New York City (SF). Five of the SC strains were described previously [11]. A second collection of 73 clinical isolates were from the Istituto di Microbiologia, Facoltà di Medicina e Chirurgia ‘Agostino Gemelli’, Università del Sacro Cuore, Rome (RM). Twenty strains were from the clinical Hospital of Sassari (SS); six were from the Main Hospital of Cagliari (CA); three were isolated in Genoa (GE); two were from Padua (PD); two were from Catania (CT); and one was collected in the Slovak Republic, Brno.

Enterococcal ATCC strains (E. hirae 9790, E. faecium 19434, E. avium 14025, E. gallinarum 49573, E. pseudoavium 49372, E. solitarius 49428, E. saccharolyticus 43076, E. seriolicida 49156, E. mundtii 43186, E. maleodoratus 43197 and E. durans 19432) were obtained from the American Type Culture Collection. Additional E. faecalis strains were JH2-2, obtained from Don Clewell (University of Michigan); 20 clinical E. faecalis isolates were from the Istituto di Microbiologia, Università Cattolica del Sacro Cuore, Rome.

2.2 DNA extraction and restriction analysis

Chromosomal DNA was extracted as previously published [2].

2.3 IS fingerprinting and ribotyping

For hybridization, the DNA of E. hirae and of all clinical isolates of E. faecium was transferred to supported nitrocellulose using a vacuum transfer device (ABN, Emeryville, CA) and a modification of the method of Southern [12]. Hybridization was carried out at 68°C and the blot was washed at 68°C with 0.1×SSC (1×SSC is 0.15 M NaCl, 0.015 M sodium citrate, pH 7.0) and 0.1% SDS.

A 1122-bp probe containing most of the IS-like sequence was obtained by partially digesting a 7.8-kb EcoRI fragment with the enzymes MboI and MspI (see Fig. 1).

Figure 1

Nucleotide comparison of the IS-like sequences of E. hirae and S. haemolyticus (upper sequence and lower sequence, respectively). The two open reading frames are shown starting at bases 3121 (ORF2) and 3984 (ORF1) in the E. hirae sequence. The 9-bp direct repeats (DR) are also indicated, as well as the inverted repeats (IR1 and IR2).

A 1.8-kb ApaI fragment [2, 3], containing the 3′ end of the 16S rRNA gene, the intergenic spacer and the 5′ end of the 23S rRNA gene, was used as a probe in the ribotyping.

Probes were labeled with the enhanced chemiluminescent gene labeling kit (Amersham Int., Amersham, UK). Autoradiography was carried out at room temperature using Kodak X-RP films. The 1.1-kb band was eluted from the low-melting-point agarose gel after electrophoresis.

2.4 Sequence analysis

Sequencing was performed by using the dideoxy chain-termination method of Sanger [13] with the Sequenase kit from US Biochemical Corporation with primers corresponding to either the regions flanking the pBluescript SK+/− multiple cloning region, or pJDC9 multiple cloning region, or primers designed from the previous sequence. All DNA analysis and homology studies were done using PC/Gene release 12.0 Intelligenetics (Mountain View, CA).

2.5 GenBank accession numbers

The nucleotide sequences of the 2.3-kb EcoRI-KpnI fragment and a partial sequence of the 7.8-kb EcoRI have been entered into GenBank [14] (accession numbers are U22540 and U22541, respectively).

3 Results and discussion

In order to clone the multiple chromosomal bands of E. hirae hybridizing with the 4.5-kb EcoRI probe [4], a genomic library for E. hirae was made by digesting the chromosomal DNA with EcoRI and cloning in the EcoRI site of pBluescript SK+/−. Escherichia coli DH5α was transformed with the ligation mixture. The 4.5-kb EcoRI fragment was subcloned into a 2.2-kb EcoRI-KpnI fragment and a 2.3-kb EcoRI-KpnI fragment. The 2.3-kb EcoRI-KpnI fragment was used as a probe to screen the transformants by colony hybridization. A colony which hybridized with the probe was isolated, and the DNA was analyzed by electrophoresis and found to contain a 7.8-kb EcoRI fragment. Partial sequencing of the clone revealed 99% identity with the 2.3-kb EcoRI-KpnI. A search of the GenBank data base [14] indicated that the 7.8-kb (position 2282–4262) fragment as well as the 2.3-kb fragment contained sequences similar to an IS-like sequence (GenBank accession number L14017) found in Staphylococcus aureus[9] and S. haemolyticus (IS1272, GenBank accession number U35635) [10] (Fig. 1) and lactococci [15]. The nucleotide identity was as high as 66% (Fig. 1). Sixty one percent nucleotide identity was also found with IS1182, which is part of transposon Tn5405 of S. aureus (GenBank accession number L43082) [16].

An open reading frame of 225 amino acid residues starting at position 3121 on the minus strand of the 7.8-kb EcoRI fragment was found to be 65% identical at the amino acid level to the open reading frame of S. aureus and S. haemolyticus, ORF2 (Fig. 1) [9, 10]. A second open reading frame (ORF1) starting at position 3984 on the minus strand of the 7.8-kb EcoRI fragment has 33% identity over 133 residues to the open reading frame of S. haemolyticus ORF1 (Fig. 1) [9, 10].

Direct repeats (DRs) (TCTGTTTTG) were found at positions 2378 and 3158 in the 7.8-kb EcoRI. Two inverted repeats (IRs) were found in the 7.8-kb fragment; the second (AAACGGATTATC) at positions 2832 and 3229 and the first (GCCACCTTTAT) at positions 2737 and 3140 (Fig. 1). We were not able to find homology with the two inverted repeats found in IS1272.

Insertion sequences can serve as sites for homologous recombination; consequently, duplications, deletions, and inversion arise through these structures [1, 17], as has been demonstrated by several works [6, 9, 10]. Archer et al. [10] reported the presence of a IS-like sequence responsible for an insertion-deletion event that involves the mecR1 region in S. aureus.

It seems that between staphylococci and enterococci there exists a well established exchange of genetic information. The IS-like sequence reported by Archer et al. [9] is widely spread in the Staphylococcus genus, this fact suggested a search for the IS-like sequence presence among other enterococci. The 7.8-kb EcoRI fragment was partially digested with MboI and MspI, after electrophoresis the 1.1-kb MboI-MspI fragment, which contains the 5′ end of the IS-like sequence, was used as a probe to screen several enterococcal species [3] for the presence of the IS-like sequence, but only E. hirae ATCC 9790 and E. faecium ATCC 19434 hybridized with the probe (data not shown). Strains of E. faecalis, including 20 clinical isolates, were examined, but none contained regions hybridizing to our IS probe (data not shown). The results obtained were confirmed by using also the 2.3-kb EcoRI-KpnI as a probe (which contains most of the 3′ portion of the IS).

DNA from 131 isolates of E. faecium was cut with XbaI and analyzed by hybridization with the 1.1-kb probe. Eighty-three clinical isolates out of 131 were positive on hybridization with the 1.1-kb IS probe and the 2.3-kb EcoRI-KpnI probe. Many of the patterns obtained showed a single band of hybridization, although in 39 strains (47%), 2–6 bands of differing intensity were observed.

We previously found that the best ribotyping profiles were obtained using the XbaI endonuclease [2, 3]. In order to evaluate the IS ability to differentiate among clinical isolates of E. faecium, the IS fingerprinting results were compared with the ribotyping patterns of the DNA from the same isolates cut with XbaI; the family patterns obtained were more heterogeneous than those obtained with ribotype, the IS-like probe was able to discriminate different clusters within the major part of the ribotyping family patterns. For instance, in Fig. 2 A the ribotypes of some E. faecium strains isolated in Philadelphia are represented, as well as the patterns obtained with the 1.1-kb IS-like fragment (Fig. 2B) for the isolates of Saint Christopher's Hospital of Philadelphia (ASC).

Figure 2

Computer-generated lane map (Image Master 1D, Pharmacia) combining the RFLP of 83 E. faecium isolates obtained with ribotyping (A) and IS fingerprinting (B). The origin of the specimens is also indicated (SS=Sassari, CA=Cagliari, RM=Rome, GE=Genoa, ATJ=Thomas Jefferson Hosp., ASC=Saint Christopher Hosp., PD=Padua, CT=Catania). The lanes are arranged according to the level of similarity of the RLFPs obtained by ribotyping and IS fingerprinting. Molecular mass marker: λ-HindIII (Boehringer Mannheim) is shown on both sides of the figure.

The IS fingerprinting provided more information about strains ASC1–ASC5, showing identity of ASC2 and ASC3 (but differences for ASC4 and ASC5 or ASC1) (Fig. 2B), in contrast the ribotype did not discriminate among ASC1, ASC2, ASC3 and ASC5. The IS patterns were in agreement with those found by Rupar et al. [11]. The five strains were divided into three different classes by IS fingerprinting whereas only two classes were differentiated by ribotyping. The IS patterns of the strains did not change after several subcultures. The results obtained indicate that IS fingerprinting could be used as an independent marker, since the relationship between ribotyping and IS profiles was not always evident. For instance, strains RM28, RM32, RM46 and RM53 had the same pattern with ribotyping whereas they were different with regard to IS fingerprinting (Fig. 2A,B). These strains were isolated from different patients independently and at different times.

In conclusion, we have cloned and sequenced two fragments of 2.3-kb KpnI-EcoRI and 7.8-kb EcoRI repeated several times in the E. hirae chromosome and found a high degree of homology with an IS-like sequence found in staphylococci [10]. The use of the 1.1-kb MboI-MspI fragment as a probe against the chromosomal DNA of several E. faecium clinical isolates allowed us to distinguish between different strains. The profile comparison with the ribotyping results revealed an independent level of differentiation of the IS fingerprinting. One of the disadvantages of using IS-like sequences is that they may not be present in all the strains tested.


This work was supported in part by Public Health Service Grant DE08942 from the National Institute of Health. We thanks Dr. J. Mortensen and Dr. H. Framow for providing the strains from Saint Christopher Hospital and Graduate Hospital (Philadelphia). We also thank Dr. G.D. Shockman for many helpful discussions.


View Abstract