Composite Structure of Streptococcus pneumoniae Containing the Erythromycin Efflux Resistance Gene mef(I) and the Chloramphenicol Resistance Gene catQ

Abstract

In recent years mef genes, encoding efflux pumps responsible for M-type macrolide resistance, have been investigated extensively for streptococci. mef(I) is a recently described mef variant detected in particular isolates of Streptococcus pneumoniae instead of the more common mef(E) and mef(A). This study shows that mef(I) is located in a new composite genetic element, whose sequence was completely analyzed and the left and right junctions determined, demonstrating a unique genetic organization. The new composite structure (30,505 bp), designated the 5216IQ complex, consists of two halves: a left one (15,316 bp) formed by parts of the known transposons Tn5252 and Tn916, and a right one (15,115 bp) formed by a new fragment, designated the IQ element. While the defective Tn916 contained a silent tet(M) gene, the IQ element, ending with identical transposase genes on both sides and containing the mef(I) gene with an adjacent new msr(D) gene variant and a catQ chloramphenicol acetyltransferase gene, was completely different from the genetic elements carrying other mef genes in pneumococci. This is the first report demonstrating catQ in S. pneumoniae and showing its linkage with a mef gene. Analysis of the chromosomal region beyond the left junction revealed an organization more similar to that of S. pneumoniae strain TIGR4 than to that of strain R6. The 5216IQ complex was apparently nonmobile, with no detectable transfer of erythromycin resistance being obtained in repeated transformation and conjugation assays.

In streptococci, active efflux-mediated erythromycin resistance was established in 1996 and associated with the M phenotype, i.e., a low-level resistance pattern affecting only 14- and 15-membered macrolides among macrolide-lincosamide-streptogramin B antibiotics (28). Recently, a trend towards increasing rates of erythromycin resistance, noted in many countries among Streptococcus pyogenes (13) and Streptococcus pneumoniae (20) populations as well as with other streptococci (4, 18, 23, 31), has been correlated with a worldwide emergence of this efflux mechanism.

The efflux pump responsible for M-type resistance is encoded by the mef gene. The first mef variant to be discovered in S. pneumoniae was called mef(E) (29) and was subsequently shown to be carried by a ca. 5.5-kb nonconjugative element (macrolide efflux genetic assembly [mega] element) containing five open reading frames (ORFs), of which mef(E) is the first (16). Another mef variant, originally described for S. pyogenes and called mef(A) (8), has been shown to be carried (i) in S. pneumoniae by a ca. 7.2-kb nonconjugative transposon (Tn1207.1) containing eight ORFs, of which mef(A) is the fourth (25), and (ii) in S. pyogenes by a variety of larger mobile elements, where it is part of an element identical (6, 24) or related (17) to Tn1207.1. In all such instances, an msr gene encoding an ATP-dependent efflux pump—now designated msr(D), albeit different variants received different appellations—was found downstream of mef. It has been suggested that, msr(D) being cotranscribed with mef, the proteins encoded by the two genes might act as a dual efflux system (16) in which the expression of both genes would be inducible by erythromycin (2) and that the msr(D)-encoded pump would be capable of functioning independently of the one encoded by mef (2, 14).

More recently, new mef alleles have been found in S. pyogenes (23), in group G beta-hemolytic streptococci (4), and in S. pneumoniae (12). In particular the last, designated mef(I), exhibited comparable homologies with the mef(E) gene of the mega element and the mef(A) gene of the Tn1207.1 transposon and was associated with no PCR evidence of the other ORFs of either mega or Tn1207.1 (12). In the present study, aimed at investigating the genetic location of mef(I), we show that this new mef variant—adjacent to a new msr variant undetected with the primers used originally (12) and linked to a catQ chloramphenicol resistance gene—is contained in a new 15-kb genetic element (designated IQ) inserted into a defective Tn916 transposon (10), which in turn is inserted into a defective Tn5252 transposon (1, 5) to form a new composite structure (30.5 kb in size) designated the 5216IQ complex.

MATERIALS AND METHODS

Bacteria.

As reported elsewhere (12), the two strains originally shown to share mef(I) also shared an identical susceptibility pattern (resistance to erythromycin and azithromycin; susceptibility to rokitamycin, telithromycin, clindamycin, tetracycline, penicillin, and levofloxacin), an identical serotype (11A) and pulsed-field gel electrophoresis type, and a new sequence type (ST1774). Although isolated in different areas of Italy, one (Spn529) from the upper respiratory tract in 2003 and one (Spn832) from bronchial aspirate in 2004, the two isolates were indistinguishable. The present study was essentially conducted with Spn529, Spn832 being used in control or confirmatory experiments.

PCR amplification experiments.

The primer pairs used in PCR experiments are listed in Table 1. DNA preparation and amplification and electrophoresis of PCR products were carried out by established procedures and following recommended conditions for the use of individual primer pairs. The Ex Taq system (TaKaRa Bio, Shiga, Japan) was used in amplification experiments expected to yield PCR products exceeding 3 kb in size.

TABLE 1.

Oligonucleotide primer pairs used^k

Gene or region	Primer designation	Primer sequence (5′-3′)	Reference or source	Product size(s) (bp)
orf24	TN6	GCTGAATGAATGTTTGATGG	15	858/5,208a
spr1199	GS13	GACTTGACAACCTGAAGTAG	15
spr1206	GD1	CAGGAACCATTCTAGCTTGG	15	194/11,639b
IRc	TN4	AGGCTTTACGAGCATTTAAG	15
orf9	SG3	GAATCTTTAGCCAGCGGTATC	15	1,778/10,104d
mef(I)	MEFA2e	TTCTTCTGGTACTAAAAGTGG	27
mef(I)	MEFA2inv	CCACTTTTAGTACCAGAAGAA	27	3,663f
IRg	M2h	TCATAGTTTCGGTTCCTC	This study
IRi	DR3	TAGAAATCGTCCATAACC	This study	ca. 2,500
nf10	INV4	GACAACGGAAACTCTATGCC	This study
nf10	DR4	GCATAGAGTTTCCGTTGTCC	This study	ca. 8,000
IRj	INV5	AGGCTTTCCGTGATTTGAG	This study

^aThe smaller product size was expected from the reported right junction in transposon Tn2009 (15); the larger was consistent with the organization of the 5216IQ complex (EMBL accession no. AJ971089).

^bThe smaller product size was expected from the reported left junction in transposon Tn2009; the larger was consistent with the organization of the 5216IQ complex.

^cFrom the intergenic region between orfI and the left end of Tn916.

^dThe smaller product size was expected from the reported organization of transposon Tn2009; the larger was consistent with the organization of the 5216IQ complex.

^eOriginally proposed for mef(A) and mef(E) (27) and likewise targeting mef(I) (12).

^fConsistent with the genetic organization of the composite 5216IQ complex.

^gFrom the intergenic region upstream of the 5′ end of the mef(I) gene.

^hDesigned from the region upstream of the mef(E) coding sequence in the mega element (16) (GenBank accession no. AF274302).

ⁱFrom the intergenic region between nf9 and nf10.

^jFrom the intergenic region between catQ and the right tnp1.

^kThe last three primer pairs were used in inverse-PCR assays.

Inverse PCR.

Inverse PCR was carried out as described by Sambrook and Russell (22) to analyze unknown DNA regions. HindIII-, NcoI-, or MunI-restricted genomic DNA (the endonucleases were from Roche Applied Science, Basel, Switzerland) were ligated and used as templates in PCR assays using primer pair MEFA2inv/M2, INV4/DR3, or INV5/DR4, respectively (Table 1).

DNA sequence analysis.

All PCR products used for sequence analysis were purified using Montage PCR filter units (Millipore Corporation, Bedford, MA). Amplicons were sequenced (bidirectionally or by primer walking) using ABI Prism (Perkin-Elmer Applied Biosystems, Foster City, CA) with dye-labeled terminators. Sequences were analyzed using the Sequence Navigator software package (Perkin-Elmer Applied Biosystems). ORF analysis was performed using the online available software NEB cutter V2.0 (http://tools.neb.com/NEBcutter2/index.php), and sequence similarity and conserved domain searches were carried out using tools (BLAST and CDART) available online at the National Center for Biotechnology Information of the National Library of Medicine (Bethesda, MD) (http://www.ncbi.nlm.nih.gov).

Transformation and conjugation experiments.

Transfer experiments were carried out as described elsewhere (12), using well-known laboratory strains as recipients: S. pneumoniae Rx1 in transformation assays and S. pneumoniae R6 and Enterococcus faecalis JH2-2 in conjugation assays.

Nucleotide sequence accession number.

The complete nucleotide sequence of the 5216IQ complex (30,505 bp) has been submitted to the EMBL sequence database and assigned accession no. AJ971089.

RESULTS AND DISCUSSION

Analysis of the DNA regions flanking mef(I).

Inverse-PCR experiments aimed at analyzing DNA regions flanking mef(I) using primer pair MEFA2inv/M2 yielded no amplification following HindIII-restricted DNA fragment ligation, whereas an amplicon was obtained with undigested DNA. This indicated the unpredicted occurrence of a sequence targeted by primer M2 downstream of mef(I). Sequencing demonstrated that the amplicon (3,663 bp) contained the last 816 bp of mef(I) (EMBL accession no. AJ971089), followed by an msr(D) gene variant with 96.2% and 95.9% homologies with the msr(D) pneumococcal genes of the mega element and of Tn1207.1, respectively (Table 2). The incomplete homology between the new msr(D) variant and the previous ones involved the sequences targeted by the primers used originally (3, 14), thus accounting for the lack of PCR evidence of the msr(D) gene initially noted for strain Spn529 (12). In contrast, sequencing confirmed the lack of orf3, orf4, and orf5 of the mega element, in line with the negative results of PCR assays (12).

TABLE 2.

Characteristics of the sequence of the IQ element from strain Spn529^a

Coordinates (bp)	ORF	Description and homology
15,434-16,579	tnp1	Domain structure similarity with transposase proteins belonging to the ISL3 family (74.2% homology with ISSha1 of Staphylococcus haemolyticus, DDBJ accession no. AP006716)
16,940-17,359	hol	Belonging to phage holin family proteins (70.1% homology with the corresponding sequence of Dehalococcoides ethenogenes, REFSEQ accession no. YP_181786)
17,361-17,930	nf3	Structure homology with N-acetylmuramoyl-l-alanine amidase family proteins (76.0% homology with the corresponding sequence of Bacillus cereus ATCC 10987, GenBank accession no. AE017194)
18,715-20,267	rec1	Putative DNA recombinase (90.1% homology with the corresponding sequence of B. cereus ATCC 10987)
20,231-20,680	nf5	Hypothetical protein (92.8% homology with the corresponding sequence of B. cereus ATCC 10987)
20,684-22,255	rec2	Putative DNA recombinase (92.2% homology with the corresponding sequence of B. cereus ATCC 10987)
22,687-23,904	mef(I)	Homologies: 93.6% with mef(E) of the mega element (GenBank accession no. AF274302), 91.4% with mef(A) of Tn1207.1 (GenBank accession no. AF227520), 92.3% with a mef variant detected in S. pyogenes in Norway (GenBank accession no. DQ016305), and 92.1% with a mef variant of group G streptococci (EMBL accession no. AJ617704)
24,019-25,482	msr(D)	Homologies: 96.2% with msr(D) (mel) of the mega element, 95.9% with msr(D) (orf5) of Tn1207.1, and 93.0% with an msr(D) variant of group G streptococci (EMBL accession no. AM084232)
25,614-26,081	nf9	Hypothetical protein
26,199-26,780	nf10	Domain structure similarity with oxidoreductase family proteins
26,767-27,267	nf11	Hypothetical protein belonging to an uncharacterized protein family (UPF0157) (92.4% homology with the corresponding sequence of Bacteroides ovatus, EMBL accession no. AJ557257)
27,589-28,248	catQ	Chloramphenicol acetyltransferase (96.5% homology with the corresponding catQ gene of Clostridium perfringens, GenBank accession no. M55620)
28,952-30,097	tnp1	Domain structure similarity with transposase proteins of the ISL3 family

^aThe IQ element (15,115 bp) forms the right portion of the 5216IQ complex (accession no. AJ971089).

Finding of part (6,003 bp) of the Tn916 transposon upstream of the DNA region containing mef(I).

In previous unrelated PCR experiments aimed at investigating the tet(M) gene in streptococcal populations, strain Spn529, being tetracycline susceptible (MIC, 0.5 μg/ml), was run among negative controls but unexpectedly yielded a positive reaction. By analogy to Tn2009, a composite element recently discovered in S. pneumoniae and formed by a Tn916-like transposon containing the mega element (15), we investigated whether the DNA fragment carrying mef(I) and mrs(D) in strain Spn529 was also inserted in a Tn916-like element.

By pairing primers SG3 (targeting the orf9 gene of Tn916) and MEFA2, an amplicon (10,104 bp) was obtained, which confirmed that the new DNA fragment was inserted into Tn916 but in a different fashion compared to Tn2009. Its sequencing showed that in strain Spn529 the DNA fragment was integrated upstream of tet(M), at nucleotide 11905 of the Tn916 sequence (GenBank accession no. U09422) (Fig. 1), causing the loss of the tet(M) promoter. A 125-bp deletion (at positions 12087 to 11963) included the ribosome-binding site and 67 of the 87 nucleotides of the tet(M) leader peptide (orf12). The tet(M) coding sequence was complete and had 98.1% homology with the tet(M) gene of Tn916 and 99.1% with the tet(M) gene of Tn2009 (GenBank accession no. AY466395). Based on the well-known regulation of tet(M) expression in Tn916 (26), the lack of the promoter and of most of the tet(M) regulatory region is consistent with tetracycline susceptibility of strain Spn529. All of the eight ORFs described downstream of tet(M) in Tn916 (int, xis, orf5, orf8, orf7, orf10, orf9, and orf6) were detected in other PCR and sequencing experiments, whereas the Tn916 ORFs upstream of tet(M) were not (Fig. 1).

FIG. 1.

ORF map of the 5216IQ complex from S. pneumoniae strain Spn529 (accession no. AJ971089). Light-gray arrows indicate ORFs from transposon Tn5252. Dark-gray arrows indicate ORFs from transposon Tn916. White arrows indicate ORFs from the new fragment containing mef(I) (checkered), msr(D) (striped), and catQ (spotted). Black arrows indicate chromosomal ORFs. Thin arrows below the ORF map indicate the positions and directions of the primers used in inverse-PCR experiments and of the major primers used to investigate the genetic organization of the new element.

Finding of part (9,313 bp) of the Tn5252 transposon between the Tn916 region and the left junction of the 5216IQ complex.

To define the insertion of the new element containing the defective Tn916 transposon and the new efflux gene variants [mef(I) and msr(D)] into the chromosome, we used primer pairs GD1/TN4 and TN6/GS13 (15), where TN4 and TN6 target the two ends of Tn916 and GD1 and GS13, respectively, spr1206 and spr1199, i.e., two chromosomal ORFs located approximately 10 kb apart in the genome of strain R6 (GenBank accession no. AF008493 and AF008492). Sequence assays showed that the right end of amplicon TN6/GS13 (5,208 bp) overlapped with the left end of amplicon GD1/TN4 (11,639 bp) (Fig. 1), thus not only revealing that spr1206 and spr1199 were both upstream of the 5216IQ complex but also disclosing a chromosomal organization largely different from that of strain R6. Sequencing of the DNA fragment encompassed by the regions targeted by primers TN4 and TN6 (14,664 bp) revealed the presence, downstream of the int gene of Tn916, of a sequence (9,313 bp) displaying 96.5% homology to a fragment of the streptococcal conjugative transposon Tn5252 (1, 5), the defective Tn916 being inserted at nucleotide 239 of the reported sequence of Tn5252 (GenBank accession no. AF295925). While in the latter sequence this fragment contains nine ORFs (designated orf28 to orf20), the corresponding sequence in our element contains equally nine ORFs (designated orfA to orfI), apparently resulting from rearrangements (point mutations, premature stops, frameshifts, etc.) of those of Tn5252 (Fig. 1) (see the deposited sequences of the 5216IQ complex and Tn5252 for details). The left junction of the 5216IQ complex was found upstream of this sequence, as demonstrated by the detection of an ORF homologous (96.3%) to spr1199 of the genome of strain R6 (Fig. 1).

Analysis of the chromosome region upstream of the left junction of the 5216IQ complex.

Upstream of the left junction of the 5216IQ complex, sequence analysis confirmed a chromosomal organization quite different from that of strain R6. Upstream of spr1199, a DNA fragment (2,960 bp) showing a chromosomal organization similar (93.9% homology) to that of S. pneumoniae TIGR4, a well-established virulent isolate (30) (GenBank accession no. AE005672), was detected. This fragment contained five ORFs, homologous to TIGR4 ORFs SP1332 (93.0%), SP1333 (99.1%), SP1334 (93.6%), SP1335 (92.0%), and SP1336 (94.1%). Remarkably, primers GD1 and TN6, designed from the sequences of spr1206 of the R6 chromosome and of orf24 of Tn916, respectively (15), matched the TIGR4 chromosomal ORFs SP1332 (93.1% homologous to spr1206) and SP1336 (albeit with 4 different bases out of 22 in the targeted sequence), respectively.

IQ element and right junction of the 5216IQ complex.

Completing the characterization of the 5216IQ complex still required analysis of its right portion, where early inverse-PCR and sequencing experiments had demonstrated an msr(D) gene variant downstream of mef(I).

Sequencing of the region encompassed by the last detectable base of Tn916 and the sequence targeted by primer MEFA2 revealed six ORFs upstream of the mef(I) gene, designated tnp1 (1,146 bp), hol (420 bp), nf3 (570 bp), rec1 (1,553 bp), nf5 (450 bp), and rec2 (1,572 bp) (Fig. 1). tnp1 encoded a transposase belonging to the ISL3 family, hol a protein belonging to the holin family, nf3 and nf5 two hypothetical proteins, and rec1 and rec2 two putative DNA recombinases. The complex formed by rec1, nf5, and rec2 had 89.8% homology to a sequence occurring in the genome of Bacillus cereus (Table 2).

The region downstream of the msr(D) gene variant and the right junction of the 5216IQ complex were determined by sequencing the amplicons obtained in inverse-PCR experiments with primer pairs INV4/DR3 and INV5/DR4. Sequencing of amplicon INV4/DR3 revealed the presence of four ORFs: three, designated nf9 (468 bp), nf10 (582 bp), and nf11 (501 bp), encoded hypothetical proteins; and the fourth (660 bp) was a cat (chloramphenicol acetyltransferase) gene displaying 96.5% homology with catQ of Clostridium perfringens (Fig. 1; Table 2). Accordingly, strains Spn529 and Spn832 were chloramphenicol resistant (MIC, 16 μg/ml). It is worth noting that upstream of the catQ-like gene an 18-bp sequence that overlapped with that targeted by primer M2 (except for 4 bases) was found, thus accounting for the above-mentioned 3,663-bp amplicon obtained in early PCR experiments with primer pair MEFA2inv/M2. Sequencing of amplicon INV5/DR4 revealed a transposase gene identical to the above-mentioned transposase gene tnp1 downstream of catQ. The new fragment, ending with identical transposase genes on both sides, was designated the IQ element. The right junction of the 5216IQ complex was discovered after the right tnp1: next to a no-homology intergenic DNA segment of 408 bp there was a 24-bp sequence that was identical to a chromosomal sequence shared by both S. pneumoniae R6 (REFSEQ accession no. NC_003098) and S. pneumoniae TIGR4 (GenBank accession no. AE005672) and was followed by an ORF displaying 98.1% homology with the same ORF designated spr0601 in R6 and SP0686 in TIGR4 (Fig. 1). The sequence of tnp1 has been submitted to the ISFinder database (http://www-is.biotoul.fr), where it was designated ISSpn3.

Transfer experiments.

No detectable transfer of erythromycin resistance was obtained in repeated transformation and conjugation assays using both Spn529 and Spn832 as donors.

Conclusions.

This study shows that mef(I), a macrolide efflux resistance mef gene variant recently detected in S. pneumoniae (12), is carried by a new, apparently nonmobile, composite structure that we designated the 5216IQ complex. Its complete sequencing and the analysis of its left and right junctions revealed a unique genetic organization, identical in both strains Spn529 and Spn832. The 5216IQ complex (30,505 bp) is composed of two halves of comparable sizes: (i) a left one (15,316 bp) formed by parts of known transposons, namely, 9,313 bp of Tn5252 and 6,003 bp of Tn916; and (ii) a right one (15,115 bp) formed by the IQ element. In the assembly of the left half, the ORFs of Tn916 lying upstream of the tet(M) gene have apparently been lost, and the same is true of most of Tn5252. It is worth noting that the occurrence in Tn916-like elements not only of a regular tet(M) gene encoding tetracycline resistance, as long established (10), but also of a silent tet(M) gene associated with a tetracycline-susceptible phenotype is consistent with recent findings for erm(B)-carrying elements from both S. pneumoniae (11, 21) and S. pyogenes (7). The IQ element, ending with identical transposase genes on both sides, contained the mef(I) gene, a new msr(D) gene variant, and a catQ chloramphenicol resistance gene. To our knowledge, this is the first report demonstrating catQ in S. pneumoniae and showing its linkage with a mef gene. The IQ element—whose genetic organization is totally unlike those of the genetic elements carrying other mef genes, i.e., mef(E) (15, 16) or mef(A) (25), in pneumococci—should probably be considered a novel transposon, although it does not appear to be mobile in the genetic context of the pneumococcus. The right junction of the 5216IQ complex was just upstream of the spr0601 gene of the S. pneumoniae R6 chromosome (designated SP0686 in TIGR4). The left junction was just downstream of the spr1199 gene of the R6 chromosome; however, further sequencing revealed a chromosomal organization more similar to that of strain TIGR4 than to that of strain R6. Such extensive genome rearrangements are consistent with the well-established recombination-mediated genetic plasticity that is a distinctive feature of S. pneumoniae (9). The apparent nonmobility of the 5216IQ complex is consistent with the loss of conjugation-related ORFs of both Tn916 (10) and Tn5252 (19). On the other hand, owing to the greater similarity to TIGR4 than to R6, a lack of zones of homology with the pneumococcal chromosome could account for the unsuccessful transformation experiments, considering that the standard Rx1 recipient has the same derivation (from Avery's historical D39) as R6.