Diversity of Tn1546 Elements in Clinical Isolates of Glycopeptide-Resistant Enterococci from Scottish Hospitals

Abstract

The Tn1546-related elements of 48 Van glycopetide-resistant enterococci were compared. Ten distinct Tn1546 types were identified with variation primarily due to IS1542 and IS1216V-like insertions. Clonal isolates frequently differed in their Tn1546 type, indicating instability of Tn1546-related elements. A putative hybrid promoter was identified, generated upstream of vanR by the insertion of IS1542. The presence of this hybrid promoter was associated with constitutive expression of the van genes and elevated teicoplanin resistance.

While all VanA phenotype glycopeptide-resistant enterococci (GRE) share the same basic Tn1546 structure, as described for Enterococcus faecium BM4147 (2), considerable diversity has now been identified within Tn1546-related elements. This variation, in the form of point mutations, insertion sequence (IS) elements, and deletions, has been exploited in several epidemiological studies (8, 10, 13), often in combination with well-established methods (e.g., pulsed-field gel electrophoresis [PFGE] and ribotyping). However, the possible transient nature of insertion sequences has led to questions about the suitability of this type of analysis in epidemiological studies (4, 15). In this study, we have compared 48 VanA enterococcal isolates by PFGE and by molecular analysis of Tn1546-related elements.

Forty-eight clinical isolates of enterococci, collected from eight hospitals in Scotland over a 5-year period (1995 to 1999) were confirmed as vanA positive by PCR (5) and identified with the AP120 Strep system (BioMerieux). Sixty-nine percent of isolates were E. faecium, and the remainder were E. faecalis. Vancomycin and teicoplanin MICs were determined by incorporation of the antimicrobial agents into Mueller-Hinton agar (Oxoid) and inoculation of plates with approximately 10⁴ CFU per inocula. All 48 isolates displayed resistance levels that were typical of the VanA phenotype. Vancomycin and teicoplanin MICs ranged from 64 to 1,024 mg/liter and 8–128 mg/liter, respectively.

PFGE analysis and interpretation were performed as previously described (9, 12). Following digestion with SmaI, genomic DNA was separated by electrophoresis for 24 h at 200 V with 5- to 40-s pulse times. Discounting those isolates belonging to a previously described outbreak strain of VanA E. faecium (3), the 48 isolates were largely heterogeneous in nature, with only small discrete clusters of related isolates identified. No GRE isolated from different geographic regions of Scotland shared PFGE patterns.

Template DNA for PCRs was prepared by the guanidium thiocyanate extraction method (11). Long-template PCR (L-PCR; Expand long-template PCR system; Boehringer Mannheim) was used according to the manufacturer's instructions. The inverted repeat (IR)-specific primers described previously (16) enabled amplification of the Tn1546-related elements and subsequent restriction fragment length polymorphism (RFLP) analysis by ClaI digestion. Five of the 48 isolates failed to yield an L-PCR product with the Tn1546-IR primer, suggesting that they lacked at least one of the IRs. The loss of the left IR has been previously described and is often associated with the presence of IS1216V-like elements alone or in combination with a truncated IS3-like element (6, 13). The reason for the L-PCR failure in these five isolates was not ascertained. The Tn1546-related elements of the remaining 43 isolates were assigned to 10 distinct types on the basis of ClaI RFLP analysis. Fourteen isolates harbored Tn1546 elements that were indistinguishable from the prototype Tn1546 element by ClaI RFLP analysis. The remaining 29 isolates harbored nonprototype elements. Different Tn1546 types were evident in clonally related isolates, indicating instability within Tn1546-related elements. Such instability has been previously described (15).

Using the primers listed in Table 1, all nonprototype Tn1546-related elements were further studied by PCR and by hybridization analysis with the ECL (enhanced chemiluminescence) random prime labeling and detection system (Amersham Life Sciences Ltd.). PCR products of interest were sequenced in both directions by the dideoxy method on an ABI Prism automated sequencer. All 29 nonprototype elements harbored IS1542 within the orf2-vanR intergenic region (nucleotide position 3932) and an IS1216V-like element within the vanX-vanY intergenic region (nucleotide position 8839). Both IS elements have been described previously at the same nucleotide positions (13, 15). In addition, all nonprototype Tn1546-related elements had considerable variation within the orf1-orf2 region. Hybridization analysis, performed following BamHI digestion of Tn1546-related elements, confirmed that this variation was due at least in part to the insertion of an IS1216V-like element within the orf1-orf2 region. The precise location of the insertion was not ascertained.

TABLE 1.

Primers used in the analysis of Tn1546-related elements^a

Primer	Sequence	Coordinates	Reference
Amplification
Tn1546-IR	5′-GGAAAATGCGGATTTACAACGCTAAG	13–38; 10839–10814	16
orf1-f	5′-GATGGTGGCTCCTTTTCC	906–923
orf1-r	5′-CCATTGTTGGCAGCAGAC	1597–1580
vanR-f	5′-GCCGAAAAGCCCTTTGATA	3841–3859
vanR-r	5′-TCGTTCTAGTTTGGAATAGTCG	4699–4678
vanS-f	5′-TAGGGTAGAGCTTCCAGCGA	5752–5771
vanH-r	5′-ATTATCGTTGCCATAACGCC	6115–6096
vanX-f	5′-GATGAACGCTCTCATCATGC	8448–8467
vanY-r	5′-TTCCTGAGAAAACAGTGCTTCA	9138–9117
vanX1	5′-ACTTGGGATAATTTCACCGG	8082–8101	8
vanX2	5′-TGCGATTTTGCGCTTCATTG	8505–8486	8
Sequencing
1216V-A	5′-TTGAAGATGTAAGGCAGAGC	8727–8746
1216V-B	5′-AGGCTGAACTGCCTGTTGAC	8857–8838
1216V-C	5′-GATCGCATAGAGGGGTGGTA	8802–8821
1542-A	5′-TAGCTGTTCGAGCGAGTTCA	980–961b
1542-B	5′-TCGGCAATTTCATGTTCATC	4022–4003
Hybridization
1216V-D	5′-GTACAGACCGAAAACCCGAA	88–107c
1216V-E	5′-GCAATTTCAGCAGGATGTGA	714–695c

^aCoordinates are based on the published Tn1546 sequence (GenBank accession number M97297) unless stated otherwise. Sequencing primers were used in the sequencing of the IS1542 and IS1216V-like elements within the orf2-vanR and vanX-vanY intergenic regions, respectively. Hybridization primers 1216V-D and 1216V-E were used to obtain IS1216V-like-specific PCR product from which labeled probes were produced. Primers without a reference were designed by Primer3 (http://www.genome.wi.mit.edu.cgi-bin/primer). 

^bCorresponds to the published sequence of IS1542 (GenBank accession number AF114715). 

^cCorresponds to the published sequence of the IS1216V-like element (GenBank accession number AF093508). 

The point mutation at Tn1546 nucleotide position 8234 (G→T) within the vanX gene was screened for as described previously (10). This point mutation, previously associated with porcine isolates of VanA GRE (14), was not evident in any of the isolates studied.

Sequencing of orf2-vanR intergenic regions harboring IS1542 identified a putative hybrid promoter. The -10 TATAAT box that forms part of the native vanR promoter proposed by Holman et al. (7) is duplicated by the 8-bp target site duplication generated by IS1542 insertion at nucleotide position 3932. This duplicated -10 box forms a putative promoter sequence in conjunction with an outwardly directed -35 box (TTTACA ) located within the inverted repeat of IS1542. The impact of the IS1542 insertion and the resulting hybrid promoter on the expression of glycopeptide resistance was assessed by growth curve analysis following glycopeptide challenge and by VanX d,d-dipeptidase enzyme assays (1). Growth curves were consistent with induced expression of van genes, irrespective of whether encoded by a prototype or nonprototype Tn1546 element. However, VanX enzyme assays revealed significant constitutive expression of the van genes of nonprototype Tn1546-related elements. Specific activities in the absence of induction were, on average, 10-fold greater than the background expression from prototype Tn1546 elements. Constitutive expression, which we propose is mediated by the IS1542-generated hybrid promoter, supplemented rather than replaced the inducible van gene expression. In addition, nonprototype Tn1546 elements conferred higher levels of teicoplanin resistance than did prototype Tn1546 elements. While this phenomenon was partly medium dependent (Table 2), the consistency of teicoplanin MICs conferred by prototype Tn1546 elements suggested that the elevated resistance was not solely attributable to the media used. The reason for the medium dependency is unclear.

TABLE 2.

Summary of teicoplanin MIC data for VanA isolates with prototype and nonprototype Tn1546 elements

Medium used for sensitivity testinga	Prototype Tn1546			Nonprototype Tn1546b
	MIC50c	MIC90c	Rangec	MIC50	MIC90	Range
Iso-Sensitest agar	8	32	4–32	64	128	16–128
Mueller-Hinton agar	16	64	4–64	32	64	16–128
Brain heart infusion agar	16	32	8–32	256	512	32–1024

^aThe control organism, E. faecalis NCTC 775, was consistently susceptible on all media tested (MIC, 0.5 mg/liter). 

^bDefined as having IS1542 and IS1216V-like insertions within the orf2-vanR and vcnX-vanY intergenic regions, respectively. 

^cExpressed in milligrams per liter. 

In conclusion, this study has revealed considerable diversity within the Tn1546-related elements of VanA GRE in Scotland. The types of variation witnessed were consistent with the findings of previous studies. Many IS elements have the potential to form hybrid promoters, owing to the presence of outwardly directed -35 regions within their inverted repeats. This study describes the first such case within the van gene cluster. We propose that constitutive expression of the van genes from the hybrid promoter results in the elevated teicoplanin resistance conferred by nonprototype Tn1546 elements. The elevation in teicoplanin resistance and not vancomycin resistance could potentially reflect different abilities of the glycopeptide agents to act against residual d-Ala-d-Ala-terminating precursors.