Distribution of Insertion Sequences Associated with Tn1546-Like Elements among Enterococcus faecium Isolates from Patients in Korea

Abstract

The vanA gene cluster is carried as a part of Tn1546-like elements. The genetic diversity in Tn1546-like elements has been documented previously. The differences described thus far have included the integration of insertion sequence (IS) elements IS1216V, IS1251, IS1476, and IS1542. Among these, IS1216V has been reported to be widespread in VanA enterococci of diverse geographic areas, whereas IS1542 and IS1476 have been reported only in the United Kingdom and Canada, respectively. We investigated the distribution of ISs among 20 vanA-containing Enterococcus faecium isolates from human patients in nine different university hospitals in Korea. Pulsed-field gel electrophoresis (PFGE) was performed to identify the clonality of the isolates. Moreover, PCR amplification of the internal regions of Tn1546 was performed for structural analysis of the van gene, and both DNA strands of the PCR amplicons were directly sequenced by the dideoxy termination method. The PFGE patterns revealed a high degree of clonal diversity. Structural analyses of the van gene detected IS1542 and IS1216V in the genomes of all 20 isolates, whereas it did not detect IS1476 or IS1251 in the genomes of any of the isolates. In addition, IS19 was detected in the vanS-vanH intergenic region of one isolate. These data indicate that identification of the IS within a vanA gene cluster could be a useful tool in epidemiological investigations. In addition, the distribution of ISs associated with Tn1546-like elements among the Korean isolates is therefore similar to that among European vancomycin-resistant enterococci.

Recently, vancomycin-resistant enterococci (VRE) containing the vanA gene have been isolated from humans and animals worldwide (21). Epidemiologic studies of VRE indicate that there are geographic differences. In Europe, the evolution of VRE is related to the use of avoparcin, which has been used to promote animal growth, and this is consistent with the potential spread of VRE from animals to humans (16, 19). In contrast, in the United States, where avoparcin has not been approved for use, VRE have been isolated only from hospitalized patients and not from healthy individuals or animals.

In Korea, VRE were first detected in 1992 (10). Until 1997, the prevalence of VRE in Korea was low; however, it rapidly increased thereafter. Shin et al. (15) reported that only 8 VRE isolates among 5,275 enterococci were detected from 1995 to 1997. In contrast, from 1998 to 2000 the number of VRE significantly increased to 325 isolates among 5,705 enterococci. Moreover, until 1998 avoparcin had been used in Korea, as it had in Europe, to enhance animal growth. In this context, VRE in Korea have been isolated not only from hospitalized patients but also from animal feces and raw meat (23). Thus, it is important to understand the underlying molecular mechanisms for the dissemination of VRE in Korea.

Generally, the main mechanism for the dissemination of the van gene in enterococci could be the clonal dissemination of VRE. Recently, the horizontal transfer of a resistance gene cluster has also been regarded as an important mechanism in this. Pulsed-field gel electrophoresis (PFGE) has been widely carried out to obtain an understanding of the clonal dissemination of VRE, and recently, structural analyses of the van gene have been introduced to obtain an understanding of the horizontal transfer of the resistance gene cluster. It has been known that the vanA gene cluster is carried as a part of Tn1546-like elements, and this indicates that the horizontal transfer of Tn1546-like elements plays an important role in the spread of vanA-type VRE. Therefore, investigation of the genetic variations among Tn1546-like elements would be essential to providing an understanding of the mechanism of spread of VRE, particularly in the case of horizontal gene transfer. The majority of the variations comprise integration of insertion sequences (ISs) with or without a deletion at the insertion site, point mutations, and deletions (2). In this study, we therefore analyzed vanA-containing enterococci isolated from diverse geographic areas in Korea by PFGE and molecular analysis of the Tn1546-like elements to elucidate the mechanism of spread of VRE.

MATERIALS AND METHODS

Bacterial strains.

From 2000 to 2002, 20 clinical isolates of vanA-containing Enterococcus faecium were collected from nine different university hospitals in diverse geographic areas in Korea. Previously characterized VRE strain E. faecium BM4147 served as a control (1). Organisms were identified by conventional biochemical reactions, with the Vitek identification system (bioMérieux, Hazelwood, Mo.), and with the API 20 Strep system (bioMérieux).

PFGE.

PFGE was performed on a CHEF-DR III apparatus (Bio-Rad Laboratories, Richmond, Calif.), as described previously (11). After digestion with SmaI, genomic DNA was separated by electrophoresis, with ramped pulse times beginning with 5 s and ending with 30 s at 6 V/cm for 20 h. The banding patterns were interpreted by Dice analysis and analysis by the unweighted pair group method with arithmetic averages with Bio-Gene software (Vilber Lourmat Inc., Marne-la-Vallée, France).

DNA extraction and PCR.

Extraction of bacterial DNA was performed with a Qiagen DNeasy kit (Qiagen GmbH, Hilden, Germany), according to the instructions of the manufacturer. The vancomycin resistance genotypes were determined by PCR with primers specific for the vanA, vanB, vanC1, and vanC2-vanC3 gene sequences, as described previously (3, 6). For structural analysis of Tn1546-like elements, overlapping PCR amplification of internal regions of Tn1546 was performed. The primer sequences and target locations for specific Tn1546 regions are listed in Table 1. Primers ISV650F and ISV132R were designed according to the published sequence of IS1216V with the OLIGO program (version 6.0; National Biosciences Inc., Plymouth, Minn.). The melting temperatures of the individual primers were calculated by using the OLIGO program (National Biosciences Inc.). To determine the exact left end of the truncated Tn1546-like elements, genomic DNA from all isolates was amplified with a combination of Tn1546-derived primer 4511R and IS1216V-specific primers (primers ISV650F and ISV132R).

TABLE 1.

Nucleotide sequences of PCR primers used in this study

Gene specificity and primer	Sequences (5′→3′)	Location (positions)	Amplicon size (bp)	Reference or source
Tn1546-specific primers
42F	ATT TTC CTG ACG AAT CCC TCG	42-62	308	14
349R	TCG GAA AAC AAG GTG AGC TTA GA	349-327		14
164F	AAC CTA AGG GCG ACA TAT GGT G	164-185	758	14
921R	AAA AGG AGC CAC CAT CTA CCG	921-901		14
170F	AGG GCG ACA TAT GGT GTA ACA	170-190	1,744	5
1913R	CGT CCT GCC GAC TAT GAT TAT TT	1913-1891		5
949F	GCA TGT AGT GAT GAA ACA CCT AGC TGC	949-975	2,028	14
2976R	TGA AGA TGA ATG GAT ACT GGG GAC C	2976-2952		14
1871F	ACC GTT TTT GCA GTA AGT CTA AAT	1871-1894	1,856	5
3726R	AGC CCT AGA TAC ATT AGT AAT T	3726-3705		5
3514F	ACT GTA ATG GCT GGT GTT AAC	3514-3534	465	15
3978R	CAT AGT TAT CAC CCC TTT CAC TAT	3978-3956		5
3907F	ATG CTT ATA AAT TCG GCC C	3907-3925	888	5
4794R	ATC CAA TCC CCA AGT TTC CC	4794-4775		5
3992F	TTA TTG TGG ATG ATG AAC ATG	3992-4012	520	15
4511R	TCG GAG CTA ACC ACA TTC	4511-4494		15
4676F	AAC GAC TAT TCC AAA CTA GAA C	4676-4697	1,094	5
5769R	GCT GGA AGC TCT ACC CTA AA	5769-5749		5
5235F	ATA TCA CGT TGG ACA AAG C	5235-5253	1,801	15
7035R	TTA CGT CAT GCT CCT CTG AG	7035-7017		15
8082F	ACT TGG GAT AAT TTC ACC GG	8082-8101	424	14
8505R	TGC GAT TTT GCG CTT CAT TG	8505-8486		14
8448F	GAT GAA CGC TCT CAT CAT GC	8448-8467	691	1
9138R	TTC CTG AGA AAA CAG TGC TTC A	9138-9117		1
8544F	GCA TAT AGC CTC GAA TGG	8544-8561	1,037	15
9580R	TCG TCA AGC TTG ATC CTA C	9580-9562		15
10446F	AAT ACT GTT GGA GGC TTT CTT GG	10446-10468	132	14
10577R	GGT ACG GTA AAC GAG CAA TAA TAC G	10577-10553		14
IS1216V-specific primers
650F	ACC TTC ACG ATA GCT AAG GTT			This study
132R	AGG ATT ATA TAA GAA AAC CCG			This study

Sequence analysis.

PCR amplicons larger than that of the prototype vanA gene cluster were purified with GeneClean kits (Qbiogene Inc., Carlsbad, Calif.). The purified PCR products were directly sequenced by using an ABI Prism 3100 DNA sequencer (Applied Biosystems, Foster City, Calif.). DNA fragments amplified with a combination of a Tn1546-specific primer and IS1216V-specific primer were also purified and subsequently sequenced to determine the exact integration site and orientation of the IS1216V insertion. The DNASIS program for Windows (version 2.6; Hitachi Software Engineering, South San Francisco, Calif.) was used for sequence analysis.

RESULTS

PFGE.

All isolates were largely heterogeneous in nature. The isolates formed only one cluster when a similarity cutoff of 85% was used (Fig. 1).

FIG. 1.

Dendrogram produced following Dice analysis and analysis by the unweighted pair group method with arithmetic averages of the PFGE patterns of E. faecium isolates from hospitals.

Structural analysis of Tn1546-like elements by PCR mapping and sequence analysis.

All isolates were divided into three main types according to the distribution of ISs integrated into Tn1546 elements (Table 2 and Fig. 2). Type I was characterized by an IS1542 insertion in the orf2-vanR intergenic region and an IS1216V insertion in the vanX-vanY intergenic region. Type II was characterized by two copies of IS1216V at the left ends of Tn1546-like elements and in the vanX-vanY intergenic region as well as IS1542 in the orf2-vanR intergenic region. Type III was characterized by the presence of IS19 in the vanS-vanH intergenic region, in addition to IS1542 in the orf2-vanR intergenic region and IS1216V in the vanX-vanY intergenic region. No isolates were identical to the prototype strain.

TABLE 2.

Structural analysis of Tn1546-like elements by overlapping PCR

Tn1546-like element type	No. of isolates	PCR product with the following primer pairsa:
		42-349	164-921	170-1891	949-2976	1871-3726	3514-3978	3907-4794	3992-4511	4676-5769	5735-7035	8082-8505	8448-9138	8544-9580	10446-10577
BM4147	0	+	+	+	+	+	+	+	+	+	+	+	+	+	+
I	7	−	+	+	+	+	++	++	+	+	+	+	++	+	+
IIa	1	−	++	++	+	+	++	++	+	+	+	+	++	+	+
IIb	7	−	−	−	−	−	++	++	+	+	+	+	++	+	+
IIc	4	−	−	−	−	−	−	−	+	+	+	+	++	+	+
III	1	−	+	+	+	+	++	++	+	+	+	+	++	+	+

^a+ and ++, amplicons the same size as and larger than those obtained for E. faecium BM4147, respectively; −, absence of an amplicon with the particular pair of primers. The numbers in each primer pair correspond to the first numbers of the positions listed in Table 1.

FIG. 2.

Genetic maps of Tn1546 types of E. faecium isolates from Korean hospitals. The positions of genes and open reading frames (orf1 and orf2) and the direction of transcription are marked by open arrows at the top. Boxes with vertical lines represent IS elements. The positions of the first nucleotide upstream and the first nucleotide downstream from the IS insertion sites are depicted. Solid arrows indicate the transcriptional orientation of the inserted IS elements. Deletions are indicated by dotted lines.

IS1216V was present in the vanX-vanY intergenic regions of the genomes of all 20 isolates, but at various points of integration. Among these 20 isolates, the insertions in 15 isolates were accompanied by small deletions adjacent to the insertion site (Fig. 2). Also, IS1216V was present at the left ends of Tn1546-like elements of 12 isolates (type II), with or without large deletions encompassing the orf1 and/or orf2 region. IS1542 was detected in the orf2-vanR intergenic regions of all 20 isolates. The location of the IS1542 insertion in this region was identical to that described previously (4, 14, 22), corresponding to nucleotide 3932 of Tn1546 with an 8-bp duplication of the target sequence (CTATAATC). In type IIc isolates (Fig. 2), the 3′ end of IS1542 was deleted at various points by an IS1216V insertion. IS19 was detected in only one isolate and was integrated at nucleotide 5813 with an 8-bp duplication of the target sequence (GATGTATA).

DISCUSSION

After the first identification of VRE in Korea in 1992, several outbreaks of VRE associated with clonal spread were observed. Recently, however, the horizontal transfer of the resistance gene cluster has also been regarded as the main mechanism in the dissemination of the van gene (11, 15). VanB VRE were predominantly isolated in Korean hospitals in the initial years, whereas recent isolates (those recovered from 1998 to 2000) were mostly of the VanA type (11, 15). Therefore, structural analysis of the vanA gene cluster is critical to investigations of the epidemiology of vanA-containing enterococci from Korean hospitals.

The vanA gene cluster is carried as a part of Tn1546-like elements. The heterogeneity of Tn1546 has previously been reported and comprises point mutations, deletions, and the integration of IS elements. Among these variations, the presence of IS elements accounts for much of the heterogeneity. Until recently, IS1216V, IS1542, IS1251, and IS1476 were reported in VanA VRE. IS1216V is known to be ubiquitous in vanA elements (9, 20), whereas the other three IS elements appear to be geographically restricted. For example, IS1542 is frequently found in clinical and poultry VRE isolates from the United Kingdom and Ireland (14, 22). IS1251 and IS1476 have been reported in the vanA elements of enterococci from the United States (5, 7) and Canada (12), respectively.

In this study we characterized the structures of the Tn1546-like elements among E. faecium isolates from patients admitted to hospitals in different parts of Korea and typed the strains by PFGE. By using a combination of overlapping PCR and sequencing analysis, three main types of vanA gene clusters were identified according to the distributions of ISs, i.e., IS1216V, IS1542, and IS19, in the Tn1546-like elements. In contrast, IS1476 and IS1251 were not detected in the Korean isolates. IS1216V and IS1542 were identified in the genomes of all isolates from Korean hospitals. In type IIb and IIc isolates, IS1216V was inserted at the left ends of the vanA elements, with a deletion that included the orf1 and/or orf2 regions. IS1542 was found in the Tn1546-like element at exactly the same position described previously (4, 14, 22). Interestingly, the 3′ end of IS1542 belonging to type IIc isolates was deleted at various points by the IS1216V insertion. This finding suggests that IS1216V at the left end of Tn1546 was acquired later than IS1542. Moreover, importantly, to our knowledge, our study is the first to demonstrate the presence of IS19 in the vanS-vanH intergenic region of the vanA gene cluster. Perichon et al. (13) reported that IS19 was inserted in the d-Ala-d-Ala ligase gene of VanD strain E. faecium BM4416, resulting in inactivation of the ddl ligase. However, IS19 has never been documented in Tn1546-like elements.

The movement of ISs frequently causes structural alterations in Tn1546-like elements. Furthermore, several investigators have documented the functional changes associated with IS integration with or without the adjacent deletion, as in the loss of VanY activity by an IS1476 insertion (12) and inactivation of the ddl ligase by IS19 (13). In our study, it was unlikely that the integration of an IS would affect the function of the vanA gene cluster.

The genetic differences among Tn1546-like elements have been investigated in several previous studies (2, 8, 9, 17, 18, 20, 22). However, the Tn1546 subtypes of the enterococci investigated were not comparable, since various molecular techniques were used. Before the use of overlapping PCR in order to analyze the structures of Tn1546-like elements, we performed long-PCR amplification with subsequent restriction fragment length polymorphism (RFLP) analysis, as described previously (17). Eleven of the 20 isolates failed to yield a product by long PCR, since they were types IIb and IIc by the overlapping PCR. Failure of long-PCR amplification was explained by deletions in the proximal orf1 primer binding region of the vanA gene cluster. This result suggests that the long PCR with RFLP analysis is not suitable for typing of the VanA transposons of Korean isolates. By using overlapping PCR and sequencing analyses, the Tn1546 subtypes could be compared with those described previously. To our knowledge, type I in this study likely corresponds to type B of the European isolates reported by Willems et al. (20), except for the IS1542 insertion. Type II corresponds to type E. In particular, type IIc appears to be identical to subtype E13 (14, 20). All types of the Tn1546-like elements in this study harbored IS1542, which is restricted to Europe, mainly the United Kingdom. Thus, the distribution of ISs in Tn1546-like elements among Korean VRE isolates is similar to that among European VRE isolates rather than to that among American VRE isolates. The identification of ISs within the vanA gene cluster to analyze and compare the structures of Tn1546-like elements could be a useful tool in epidemiological studies.