Tn5253 Family Integrative and Conjugative Elements Carrying mef(I) and catQ Determinants in Streptococcus pneumoniae and Streptococcus pyogenes

Abstract

The linkage between the macrolide efflux gene mef(I) and the chloramphenicol inactivation gene catQ was first described in Streptococcus pneumoniae (strain Spn529), where the two genes are located in a module designated IQ element. Subsequently, two different defective IQ elements were detected in Streptococcus pyogenes (strains Spy029 and Spy005). The genetic elements carrying the three IQ elements were characterized, and all were found to be Tn5253 family integrative and conjugative elements (ICEs). The ICE from S. pneumoniae (ICESpn529IQ) was sequenced, whereas the ICEs from S. pyogenes (ICESpy029IQ and ICESpy005IQ, the first Tn5253-like ICEs reported in this species) were characterized by PCR mapping, partial sequencing, and restriction analysis. ICESpn529IQ and ICESpy029IQ were found to share the int_Sp23FST81 integrase gene and an identical Tn916 fragment, whereas ICESpy005IQ has int₅₂₅₂ and lacks Tn916. All three ICEs were found to lack the linearized pC194 plasmid that is usually associated with Tn5253-like ICEs, and all displayed a single copy of a toxin-antitoxin operon that is typically contained in the direct repeats flanking the excisable pC194 region when this region is present. Two different insertion sites of the IQ elements were detected, one in ICESpn529IQ and ICESpy029IQ, and another in ICESpy005IQ. The chromosomal integration of the three ICEs was site specific, depending on the integrase (int_Sp23FST81 or int₅₂₅₂). Only ICESpy005IQ was excised in circular form and transferred by conjugation. By transformation, mef(I) and catQ were cotransferred at a high frequency from S. pyogenes Spy005 and at very low frequencies from S. pneumoniae Spn529 and S. pyogenes Spy029.

INTRODUCTION

The linkage between two specific resistance genes, one encoding macrolide efflux [mef(I), a subclass of the mef gene], and one encoding acetyltransferase-mediated chloramphenicol inactivation (catQ), was first described in Streptococcus pneumoniae (1). The novelty of the gene combination was further enhanced by the fact that mef(I) was a newly detected mef subclass with an adjacent new msr(D) gene variant (2), and catQ had until then been described only in Clostridium perfringens (3, 4), even if a catQ-like gene was previously reported in a single isolate of Streptococcus agalactiae (5). mef(I) and catQ were found in the so-called IQ element, a module-containing two identical tnp1 transposase genes at either end, which represented a moiety of a composite structure (5216IQ complex, ∼30.5 kb); the other moiety was formed by fragments of transposons Tn5252 and Tn916, with Tn916 containing a silent tet(M) gene (1). No transfer of erythromycin or chloramphenicol resistance was obtained in repeated conjugation and transformation assays.

Subsequently, two defective IQ elements were described in Streptococcus pyogenes, one characterized by a considerably shorter region upstream of mef(I) comprising tnp1 and a truncated recombinase gene, rec2, and one reduced to the sole mef(I)-catQ fragment (6). By conjugation, while the former defective IQ element proved nontransferable, the latter could be transferred to an S. pyogenes recipient. Apart from its transferability, the latter defective IQ element was quite different from that of the S. pneumoniae 5216IQ complex, since it appeared to be enclosed in a putative genetic element of ~50 kb, was located on the opposite side of the Tn5252 fragment, and lacked the Tn916 fragment (6).

New data and the notion of the integrative and conjugative element (ICE) (7), which has become established over the past few years, largely replacing the earlier concept of conjugative transposons (8, 9), prompted us to undertake a better characterization of the nature and organization of the genetic elements carrying mef(I) and catQ in S. pneumoniae and S. pyogenes. In recent years, considerable progress has been made in the understanding of those large genetic elements, a distinct family of ICEs, whose prototype is Tn5253, a conjugative composite structure resulting from the insertion of Tn5251 (virtually identical to Tn916) into Tn5252 (10). The complete sequence of Tn5253 has become available and has lately been analyzed (11), and several studies have recently stressed the heterogeneity of Tn5253-like composite elements and their integrase genes (int₅₂₅₂ or int_Sp23FST81) (12–17).

The present study was aimed at identifying and characterizing the genetic elements exhibiting the mef(I)-catQ gene combination in one strain of S. pneumoniae and two of S. pyogenes. The relevant elements from all three strains were found to be Tn5253 family ICEs: two shared the int_Sp23FST81 integrase gene and an identical Tn916 fragment, one bore the int₅₂₅₂ integrase gene and lacked any Tn916 fragment, and all lacked the linearized pC194 plasmid that is usually a distinctive cargo of Tn5253-like ICEs. The conjugation and transformation experiments disclosed significant differences in the abilities of the three strains to transfer the mef(I) and catQ genes.

MATERIALS AND METHODS

Bacterial strains.

Three Streptococcus isolates (one of S. pneumoniae and two of S. pyogenes) carrying mef(I) and catQ were studied. The S. pneumoniae isolate (strain Spn529) was the one in which the mef(I)-catQ linkage and the IQ element were originally described (1). The two S. pyogenes isolates were those for which two different types of defective IQ elements were later described in that species: Spy029 (formerly called MB56Spyo029) and Spy005 (formerly called MB56Spyo005) (6). Spn529, Spy029, and Spy005 exhibited the M phenotype of macrolide resistance, with erythromycin MICs of 8, 16, and 16 μg/ml, respectively. The chloramphenicol MICs for the same strains were 16, 64, and 64 μg/ml, respectively.

PCR amplification experiments.

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

TABLE 1.

Oligonucleotide primer pairs used

Procedure gene	Primer designation	Sequence (5′ to 3′)	Reference or source	Product sizea
Integrase genes
int5252	INTF	AGTAACGTGGACTAAGATG	16	1,090
	INTR	ACTACTGGTGATGAAGCG	16
intSp23FST81	intICE-F	TATCAGCGTGTCCTAATCG	16	934
	intICE-R	AATCAACAGGGACTACCG	16
Resistance genes
mef(I)b	MEFA1	AGTATCATTAATCACTAGTGC	42	348
	MEFA2	TTCTTCTGGTACTAAAAGTGG	42
catQ	catQ1	TAGAAAGCCATACTTTGAGC	6	536
	catQ2	CATGATGCACCTGTACAGAC	6
Mapping of ICESpn529IQ and ICESpy029IQc
intSp23FST81	intICE-F	TATCAGCGTGTCCTAATCG	16	4,310
orf6	MobR	TGATGGACAAAAACAGATGG	16
orf4	RelF	GTTCCTTCAGTGTTTCAATC	16	8,985
orf13	primF	CTGTTCAAGAAAATAGAGAGAG	This study
orf13	primR	GCATCATCACAATCTACCGC	This study	8,419
orf18	M6	GTTAGAACCATTAAGTGAGC	This study
orf18	R18	GTAACTTTCTAAATCTTGGTC	This study	6,465
catQ	catQ1	TAGAAAGCCATACTTTGAGC	6
orf54	TraG	TAGAGTTCCAAAACGGGC	This study	6,123
repA	NewORF6α	ATGCTCTTGACAATGTTCGC	16
Mapping of ICESpy005IQd
int5252	INTF	AGTAACGTGGACTAAGATG	16	6,962e
umuC	13R	CATTTGATGTAAAGACTCGC	16
umuC	13F	CAATAGTGAAGTGAACCC	16	13,010/4.7 kbf
orf52	primF	CTGTTCAAGAAAATAGAGAGAG	This study
orf52	primR	GCATCATCACAATCTACCGC	This study	8,287
orf47	M6	GTTAGAACCATTAAGTGAGC	This study
orf47	R18	GTAACTTTCTAAATCTTGGTC	This study	26,616/9.0 kbf
virB4	26R	GTGAGAGACTACGGTATGG	This study
orf19	28F	GATGACTTACTGAACCTG	This study	8,030e
traG	21R	TTAGCCTCCTACTTGGTCG	16
orf12	CTRfor	GCCATTCGTTCCTGTTGTTC	This study	5,362/>10 kbf
orf2	NewORF6α	ATGCTCTTGACAATGTTCGC	16
Detection of the IQ element in ICESpy005IQ
orf12g	CTRfor	GCCATTCGTTCCTGTTGTTC	This study	6.1 kb
catQ	catQ1	TAGAAAGCCATACTTTGAGC	6
mef(I)	MEFA2	TTCTTCTGGTACTAAAAGTGG	42	6.3 kb
orf2g	NewORF6α	ATGCTCTTGACAATGTTCGC	16
ICESpy005IQ junctions
Spy0956h	Spy0956L	ATCATGGTTCGCAAGGTTTC	This study	1,679
int5252	INTR	ACTACTGGTGATGAAGCG	16
repA	ICE-RepA	AAGCGAACATTGTCAAGAGC	This study	1,735
Spy0955h	Spy0955R	ATCTTTGACCGTTTGACAGG	This study
Integration site of ICESpn529IQ and ICESpy029IQ
repA	ICE-RepA	AAGCGAACATTGTCAAGAGC	This study	1,730
spr1212i	RPLJ	AGCGAAAATCTTGAACGAC	16
Circular form of the IQ element from ICESpn529IQ
catQ	catQ1	TAGAAAGCCATACTTTGAGC	6	2.5 kb
orf35c	NF2	GATGATTTGAATGGTGTTCC	This study
Circular form of the IQ elements from ICESpy029IQ and ICESpy005IQ
catQ	catQ1	TAGAAAGCCATACTTTGAGC	16	NOj
mef(I)	MEFA2	TTCTTCTGGTACTAAAAGTGG	42
ICESpn529IQ and ICESpy029IQ circular forms
intSp23FST81	ICE-5′	GGTGTACCAGAAATTACGGG	This study	NOj
repA	ICE-RepA	AAGCGAACATTGTCAAGAGC	This study
ICESpy005IQ circular form
int5252	INTR	ACTACTGGTGATGAAGCG	16	2.5 kb
repA	ICE-RepA	AAGCGAACATTGTCAAGAGC	This study

^aUnless otherwise specified, values are reported as bp.

^bOriginally proposed for mef(A) and mef(E) (42) and likewise targeting mef(I) (2).

^cExcept for integrase and resistance genes, and unless otherwise specified, the designations are according to the ORF numbering of ICESpn529IQ.

^dExcept for the integrase gene, and unless otherwise specified, the designations are according to the ORF numbering of Tn5253.

^eAmplicon also used for RFLP analysis.

^fThe first value was expected from the reported sequence of Tn5253, and the second value was consistent with the structure of ICESpy005IQ.

^gAccording to the ORF numbering of Tn5253.

^hFrom the S. pyogenes MGAS2096 genome.

ⁱFrom the S. pneumoniae R6 genome.

^jNO, not obtained.

Detection and characterization of genetic elements.

The demonstration of Tn5253-like ICEs was obtained by the detection of the integrase gene (int₅₂₅₂ or int_Sp23FST81) and other functional genes involved in ICE maintenance and conjugative transfer, using previously described PCR-based strategies (16). The mef(I)- and catQ-carrying ICE of Spn529 was completely sequenced. To characterize the two S. pyogenes ICEs, the primary approach was usually based on PCRs and mapping. DNA sequencing, or, in some instances, restriction analysis, were resorted to when special issues had to be elucidated or when amplification reactions yielded unexpected amplicon sizes.

DNA sequencing and sequence analysis.

PCR products were purified using Montage PCR filter units (Millipore Corporation, Bedford, MA). Sequencing was carried out, bidirectionally or by primer walking, using ABI Prism (PerkinElmer Applied Biosystems, Foster City, CA) with dye-labeled terminators. The sequences were analyzed using the Sequence Navigator software package (PerkinElmer). Open reading frames (ORFs) were predicted using the ORF Finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html) and NEBcutter version 2.0 software products. The criteria to design a potential new ORF were the existence of a start codon and a minimum coding size of 50 amino acids. 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) (see http://www.ncbi.nlm.nih.gov).

Experiments using restriction enzymes.

Some details about the structures of the genetic elements investigated were determined by PCR-restriction fragment length polymorphism (RFLP) analysis. Different endonucleases (HindIII, XbaI, EcoRI, and BamHI; Roche Applied Science, Basel, Switzerland) were used, depending on the region analyzed. Two primer pairs (INTF/13R and 28F/21R) used to analyze two functional regions of the genetic element from strain Spy005 are detailed in Table 1.

Conjugation experiments.

Filter mating experiments were performed as described elsewhere (18). The three strains being examined (S. pneumoniae Spn529, S. pyogenes Spy029, and S. pyogenes Spy005) were used as donors. Six rifampin- and fusidic acid-resistant (RF) derivatives of erythromycin- and chloramphenicol-susceptible strains of different Streptococcus species were used as recipients: S. pneumoniae strain R6RF (2), S. pyogenes strain 12RF (18), Streptococcus agalactiae strain 1357RF (19), Streptococcus oralis strain 1235RF and Streptococcus gordonii strain 1435RF (20), and Streptococcus dysgalactiae subsp. equisimilis strain 5381RF, an RF derivative obtained for this study from a recently investigated strain (21). The transconjugants were selected on plates containing erythromycin (1 μg/ml) or chloramphenicol (5 μg/ml), in addition to rifampin (10 μg/ml) and fusidic acid (10 μg/ml). To rule out the contribution of transformation to the genetic exchange during conjugation, matings were carried out in the presence of 10 mg/ml DNase I (Sigma-Aldrich Co., St. Louis, MO) (22). The transfer of mef(I) and catQ was verified by PCR, and the transferred element was confirmed by PCR mapping in the representative transconjugants. The frequency of transfer was expressed as the number of CFU of the transconjugants per donor. The transfer experiments were done at least three times.

Transformation experiments.

The laboratory strain S. pneumoniae Rx1 was used as the recipient in the transformation assays. The experimental procedure, including the preparation of competent cells, was as described previously (23). The transforming DNA was a crude lysate or was extracted using the GenElute bacterial genomic DNA kit (Sigma-Aldrich). Transformants were selected by plating the transformation mixture onto selective plates containing erythromycin (1 μg/ml) or chloramphenicol (5 μg/ml). The transfer of mef(I) and catQ and their linkage were verified by PCR. The transformation frequency was expressed as the number of CFU of the transformants per recipient.

Nucleotide sequence accession numbers.

The complete nucleotide sequence of ICESpn529IQ was submitted to the EMBL sequence database and assigned accession no. HG965092. The accession no. of the sequences referred to in this study are as follows: S. pneumoniae strain R6 genome, AE007317; S. pneumoniae strain TIGR4 genome, AE005672; 5216IQ complex from S. pneumoniae Spn529, AJ971089; S. pneumoniae strain ATCC 700669 genome harboring ICESp23FST81, FM211187; Tn5253 from S. pneumoniae DP1322, EU351020; S. pyogenes MGAS2096 genome, CP000261; and the Streptococcus parasanguinis FW213 genome harboring pathogenicity island FWisland_1, CP003122.

RESULTS

Evidence of Tn5253-like ICEs containing the IQ element in S. pneumoniae strain Spn529 and S. pyogenes strains Spy029 and Spy005.

A number of indications suggested that the mef(I)- and catQ-carrying elements were likely to be Tn5253 family ICEs. The three strains being studied were thus subjected to PCR analysis for integrase and other scaffold genes of Tn5253 family ICEs and for their linkage with the respective IQ element. Positive PCRs for int_Sp23FST81 were obtained with S. pneumoniae Spn529 and S. pyogenes Spy029, whereas strain Spy005 yielded a positive reaction for int₅₂₅₂. A linkage between the IQ element and a scaffold gene was demonstrated by positive PCRs using the primer pairs R18/CATQ1 (strains Spn529 and Spy029) and MEFA2/NewORF6α (strain Spy005) (Table 1). These results established that in all three strains, the IQ element was contained in a Tn5253-like ICE. The three ICEs were designated ICESpn529IQ, ICESpy029IQ, and ICESpy005IQ for strains Spn529, Spy029, and Spy005, respectively.

Characterization and organization of ICESpn529IQ.

ICESpn529IQ is the complete organization, revised in the light of new pieces of knowledge and the new notion of ICE, of the mef(I)- and catQ-carrying structure harbored by S. pneumoniae Spn529. In this isolate, early studies led to the identification of the 5216IQ complex (1), which is in fact a part of ICESpn529IQ. ICESpn529IQ was completely sequenced; it was found to be 59,466 bp in size and have a G+C content of 36%, and its sequence analysis disclosed 66 ORFs. Its genetic organization was characterized using ICESp23FST81 (13) as the genetic reference (Fig. 1A). While the two elements had virtually identical scaffolds, the cargo genes of ICESp23FST81, such as a uvrD helicase gene involved in the SOS response, a lantibiotic synthesis gene cluster, and pC194 plasmid, were missing in ICESpn529IQ, whose genetic organization is summarized below.

FIG 1.

Schematic representation of ICESpn529IQ, as determined by complete DNA sequencing, and of ICESpy029IQ and ICESpy005IQ, as determined by PCR mapping, partial sequencing, and PCR-RFLP analysis. (A) ICESpn529IQ and ICESpy029IQ, containing int_Sp23FST81, are shown, with ICESp23FST81 as the genetic reference. (B) ICESpy005IQ, containing int₅₂₅₂, is shown, with Tn5253 as the genetic reference. The ICE genomes are represented as thick bars, where the black portions were sequenced and the white portions were PCR mapped. The cargo regions are depicted as green rectangles (diagonally striped in the case of IQ-like elements). The conjugal transfer-related functional module (CTR), shared by all Tn5253 family ICEs, is depicted as a pink rectangle. The gray areas between the ICEs indicate areas of homology: darker gray areas denote >90% DNA identities between the sequenced regions and the lighter gray areas denote indistinguishable structures based on PCR mapping. The ORFs are schematically represented as arrows pointing in the direction of transcription, including integrase genes (black), TA gene pairs (yellow), cat_pC194 (brown), tet(M) (purple), catQ (light blue), mef(I) (red), and tnp1 (orange). The other ORFs of ICESpn529IQ are white. The progressive number is reported only above selected ORFs (out of the 66 ORFs) of ICESpn529IQ. The horizontal thin bars indicate the amplicons, each with the relevant primer pair, allowing PCR mapping. The sets of amplicons plotted below ICESpn529IQ and Tn5253 were used to investigate the structures of ICESpy029IQ and ICESpy005IQ, respectively. The amplicons obtained using primer pairs 13F/primF, R18/26R, and CTRfor/NewORF6α displayed unexpected sizes (see text and Table 1). The two primer pairs plotted below ICESpy005IQ allowed a definition of the location of the defective IQ element in the ICE.

The region orf1 to orf19 (bp 1 to 20061) includes genes usually found in the scaffolds of Tn5253-like pneumococcal elements. In particular, orf1 encodes the integrase, having 99% amino acid identity to the site-specific integrase of ICESp23FST81. orf2 is identical to the second ORF of ICESp23FST81. orf3 encodes a transcriptional regulator having 99% amino acid identity to orf3 of Tn5252, involved in the regulation of the conjugative transposition of the element (24). The orf4-orf6 cluster is formed by three conjugative machinery genes encoding relaxase (orf4) and mobilization proteins (orf5 and orf6), whose deduced amino acid sequences show the highest identities (96%, 93%, and 99%, respectively) to the corresponding ORFs of ICESp23FST81. In S. pneumoniae ATCC 700669 and in ICESp23FST81, the same cluster is found in the pathogenicity island PPI-1 (13). The product of orf7 shows 99% amino acid identity to a protein encoded by a gene of the pathogenicity island FWisland_1 of S. parasanguinis (25). orf8 and orf9 code for a toxin-antitoxin (TA) system. The two genes are closely related to the bicistronic operon pezT-pezA, with 92% and 98% identities between the respective deduced amino acid sequences, described in the pathogenicity island PPI-1 of S. pneumoniae TIGR4 (26). ORFs from orf10 to orf19 encode proteins with various degrees of amino acid identity (70% to 98%) to the ICESp23FST81 ORFs.

The region orf20 to orf21 (bp 20062 to 22775) is formed by two adjacent genes whose deduced amino acid sequences show 99% and 98% identities, respectively, to the adjacent genes spr0602 and spr0601 of the S. pneumoniae R6 genome and SP0687 and SP0686 of the S. pneumoniae TIGR4 genome. The duplication of these two chromosomal genes in ICESpn529IQ is one reason why during an investigation of the 5216IQ complex environment, this area was thought to belong to the chromosome (1).

The region orf22 to orf36 (bp 22776 to 37886) corresponds to the IQ element previously described in the same pneumococcal isolate. In the ORF order of ICESpn529IQ, the two tnp1 transposase genes, located at either end of the IQ element, are orf22 and orf36, catQ is orf23, and mef(I) is orf28.

The region orf37 to orf45 (bp 37887 to 43889) corresponds to the Tn916 fragment previously described in the same isolate. The silent tet(M) gene, which is unexpressed because it lacks the promoter, the ribosome-binding site, and part of the leader peptide (1), is orf37, and int₉₁₆ is orf45.

The region orf46 to orf55 (bp 43890 to 53150) is the cluster originally regarded as the Tn5252 fragment of the 5216IQ complex. It corresponds to the conjugal transfer-related functional module (CTR) of Tn5252 involved in the production of a type IV secretion system, delivering protein and DNA substrates to target cells, generally by a contact-dependent mechanism (27).

The region orf56 to orf66 (bp 53151 to 59466) includes a number of genes with high identities (>90%) to the chromosomal genes of S. pneumoniae R6 and/or TIGR4. Their detection was one reason why this area was considered to fall within the chromosome during the investigation of the 5216IQ complex environment (1). In fact, in a later study, the same chromosomal genes were reported to be duplicated in an ICE remnant in the TIGR4 genome (13). orf64 and orf65 (the third and second to last ORFs, respectively) are two adjacent and conserved genes that are consistently found in pneumococcal ICEs: they code for a cytosine methyltransferase and a replication initiator protein, respectively, which show 97% and 100% identities, respectively, to the deduced amino acid sequences of the corresponding ORFs (again, the third and second to last) of ICESp23FST81.

In parallel, we determined the chromosomal integration site of ICESpn529IQ. By combining PCR and sequencing assays, the integration site was found to be the same as that of ICESp23FST81, i.e., near the 3′ end of gene rplL (13), which is consistent with the site specificity of the int_Sp23FST81-encoded integrase (16).

Structure of ICESpy029IQ.

ICESpy029IQ displayed a structure closely comparable to that of ICESpn529IQ (Fig. 1A). This was first shown by PCR mapping; further confirmation was provided by partial sequencing of the amplicons containing major functional genes, such as those coding for the integrase, the TA operon, and the replication initiator protein. Moreover, the region of ICESpy029IQ where the defective IQ element is inserted was analyzed by partial sequencing of the amplicon obtained by pairing primers R18 and catQ1 (Table 1).

The product of int_Sp23FST81 from ICESpy029IQ showed 99% and 98% identities to the deduced amino acid sequences of the corresponding integrase genes of ICESpn529IQ and ICESp23FST81, respectively. The products of the TA-coding genes displayed 100% identities to the deduced amino acid sequences of ICESpn529IQ orf8 and orf9. The defective IQ element was inserted in the same position as the pneumococcal IQ element, i.e., next to a couple of genes corresponding to orf20 and orf21 of ICESpn529IQ. Aside from the shorter IQ element, the only noticeable difference from ICESpn529IQ was the presence of a tnp1 transposase gene adjacent to int_Sp23FST81, so that ICESpy029IQ carried a third tnp1 in addition to the two found at the ends of the defective IQ element.

The chromosomal integration site of ICESpy029IQ was near the 3′ end of gene rplL, in line with the site specificity of the int_Sp23FST81-encoded integrase.

Structure of ICESpy005IQ.

Given the presence of int₅₂₅₂, ICESpy005IQ was characterized using Tn5253 as the genetic reference (Fig. 1B). Compared to Tn5253, ICESpy005IQ carried a defective IQ element [reduced to the mef(I)-catQ fragment] but lacked both cargo regions represented by Tn916/Tn5251 and the linearized pC194 plasmid. This was first demonstrated by PCR mapping; further confirmation was provided by partial sequencing of all relevant PCR products. The amplicons obtained by pairing primers INTF and 13R and 28F and 21R (Table 1) were also examined by RFLP analysis.

The products of int₅₂₅₂ and xis₅₂₅₂ from ICESpy005IQ showed 99% and 100% identities, respectively, to the deduced amino acid sequences of the corresponding genes from Tn5253. The products of TA-coding genes displayed 92% and 96% identities to the deduced amino acid sequences of ICESpn529IQ orf8 and orf9, respectively.

By PCR mapping, the amplicon obtained using primer pair CTRfor/NewORF6α, targeting orf12 and orf2, respectively, of Tn5253 (Table 1), was much larger (>10 kb) than expected (5,362 bp). This was due to the insertion of the defective IQ element into the distal region of the Tn5252-like structure of ICESpy005IQ, as confirmed by positive PCRs using the primer pairs CTRfor/CATQ1 and MEFA2/NewORF6α (Table 1).

In S. pneumoniae, the site-specific integration of Tn5253 is in the rbgA gene (spr1043, according to ORF numbering in the S. pneumoniae R6 genome) (11, 16, 22, 28). Therefore, we investigated whether the same site specificity also applied to ICESpy005IQ. A homolog of the pneumococcal rbgA gene was identified in all S. pyogenes genomes available in GenBank: its highest DNA identity was to gene Spy0956 of S. pyogenes MGAS2096 (29), whose product, a GTP-binding protein, is the same as that of gene spr1043 of S. pneumoniae R6. The MGAS2096 genome was thus used as the reference sequence to determine the chromosomal junctions of ICESpy005IQ. The primer pairs Spy0956L/INTR and ICERepA/Spy0955R (Table 1) were used to detect the left and right junctions, respectively. The sequencing of the resulting amplicons (1,679 bp and 1,735 bp, respectively) disclosed that ICESpy005IQ was integrated in a chromosomal gene corresponding to the Spy0956 gene of the MGAS2096 genome (at base 917261).

Search for IQ element and ICE circular forms.

PCR experiments using outward-directed primer pairs (CATQ1/NF2 for S. pneumoniae Spn529 and CATQ1/MEFA2 for S. pyogenes Spy029 and S. pyogenes Spy005) (Table 1) were performed to seek IQ element circular forms in the relevant strains. A circular form was detected only for the IQ element of S. pneumoniae Spn529.

The circular forms of ICESpn529IQ, ICESpy029IQ, and ICESpy005IQ were sought by a similar approach using outward-directed primer pairs: ICE-5′/ICERepA for the first two ICEs and INTR/ICERepA for the third (Table 1). A circular form was detected for ICESpy005IQ, whereas no circular forms were detected for the other two ICEs.

Transferability of the mef(I) and catQ determinants.

In conjugative transfer experiments, mef(I) and catQ proved to be nontransferable to any recipient from S. pneumoniae Spn529 and S. pyogenes Spy029, in line with the early negative results obtained in intraspecies mating assays (1, 6). In contrast, ICESpy005IQ was consistently transferred to all recipients, usually at high frequencies, from S. pyogenes Spy005 (Table 2).

TABLE 2.

Conjugal transfer of ICESpy005IQ from the S. pyogenes donor Spy005 to erythromycin- and chloramphenicol-susceptible recipients

Streptococcus species recipient	Strain	MIC (μg/ml)a		Transfer frequencyb	Transconjugants
					Resistance genotype	MIC (μg/ml)a
		ERY	CHL			ERY	CHL
S. pyogenes	12RF	≤0.125	0.5	2.8 × 10−6	mef(I) catQ	32	32
S. agalactiae	1357RF	0.25	2	7.6 × 10−3	mef(I) catQ	16	32
S. pneumoniae	R6RF	≤0.125	1	1.7 × 10−4	mef(I) catQ	4	16
S. dysgalactiae	5381RF	≤0.125	2	4.0 × 10−6	mef(I) catQ	16	32
S. gordonii	1435RF	≤0.125	2	7.7 × 10−5	mef(I) catQ	16	32
S. oralis	1235RF	≤0.125	1	2.8 × 10−5	mef(I) catQ	8	16

^aERY, erythromycin; CHL, chloramphenicol.

^bThe reported transfer frequencies are those obtained by selecting with chloramphenicol; comparable values were obtained with erythromycin.

In the transformation experiments, the transformants were obtained from both S. pyogenes donors, although at a much higher frequency from Spy005 (usually around 10⁻⁶) than from Spy029 (around 10⁻⁹). Using S. pneumoniae Spn529 as the donor, the transformants were obtained at a low frequency (around 10⁻⁹), and only using a crude lysate as the transforming DNA. All transformants were resistant to erythromycin (M phenotype) and chloramphenicol, and PCR experiments confirmed that they bore mef(I) and catQ, linked at the expected distance.

DISCUSSION

The special feature shared by the three Tn5253 family ICEs of this study, and the reason why they were examined together, is carriage of the resistance genes mef(I) and catQ, which in the original report were described as being linked in the so-called IQ element (∼15.1 kb) in S. pneumoniae Spn529 (1). In the early descriptions of S. pyogenes Spy029 and Spy005 (6), the two resistance genes were linked in two different IQ elements, both of which were defective compared to the original pneumococcal IQ element. In this study, all three IQ elements were found to be carried by a Tn5253 family ICE. This finding strengthens and expands our understanding of such ICEs. To our knowledge, this is the first time that Tn5253 family ICEs, which are very common in S. pneumoniae, were described in natural isolates of S. pyogenes. The conjugal transfer of Tn5253 (or related variants) from S. pneumoniae to S. pyogenes and other Streptococcus species recipients has been obtained in the laboratory, as documented in previous reports (11, 22, 28). Remarkably, the two integrase genes known to be associated with the Tn5252-like moiety of such ICEs in S. pneumoniae, i.e., int₅₂₅₂ and int_Sp23FST81, were both detected in the S. pyogenes ICEs, with int₅₂₅₂ in ICESpy005IQ and int_Sp23FST81 in ICESpy029IQ.

A conserved mef-catQ fragment, most often with the mef gene subclass mef(E) instead of mef(I), was recently detected in a number of variously defective IQ elements from different species (Streptococcus mitis, S. oralis, Streptococcus sanguinis, and S. parasanguinis) of viridans group streptococci (30). In studies currently in progress in our laboratory, the same mef(I)-catQ fragment has also been detected in other Streptococcus species (M. Mingoia, unpublished data). This suggests that the DNA fragment spanning from the mef gene [mef(I) or possibly mef(E)] to the catQ gene (six ORFs, ∼5.8 kb) is a very conserved region, a sort of mef-catQ cassette that may be found alone or inside a variable structure represented by a more or less defective IQ element in different Streptococcus species, but that so far has not been reported outside this genus. Although apparently confined to streptococci, mef(I) (2, 31, 32) and catQ (4, 5) do not appear to be widespread in these organisms.

It is intriguing that all three Tn5253 family ICEs investigated, in which chloramphenicol resistance is provided by catQ, lack the linearized pC194 plasmid carrying the chloramphenicol resistance determinant cat_pC194, which is commonly regarded as typical cargo of the Tn5252-like moiety of Tn5253 family ICEs. As shown by early studies (10, 33), the cat_pC194-containing DNA region is flanked by direct repeats, the recombination of which may lead to spontaneous curing of the region. Very recently, the issue was thoroughly investigated by Iannelli et al. (11), who described an Ωcat(pC194) (7,627 bp), flanked by two 1,169-bp direct repeats. This region has also been reported as an example of the so-called unconventional circularizable structures, i.e., particular genetic structures which, though lacking their own recombinase genes, can be excised in circular form thanks to extensive flanking direct repeats (34).

The direct repeats flanking the cat_pC194-containing region in Tn5253 family ICEs generally contain the genes coding for a TA system (11, 34). This TA operon thus appears to be typically associated with the Tn5252-like moiety of Tn5253 family ICEs. The single copy of the TA operon found in ICESpn529IQ (orf8-orf9), ICESpy029IQ, and ICESpy005IQ is likely to represent the recombination site of Ωcat(pC194) and might be the outcome of a repair after pC194 excision (11). Several TA systems are found in streptococcal genetic elements. In S. pneumoniae, the TA operon of ICESpn529IQ has been described and characterized in the pathogenicity island PPI-1 of strain TIGR4 (26), but several additional TA systems, most not yet characterized in terms of structure and function, have recently been recognized in this species (35). In S. pyogenes, the best investigated TA system is the one harbored by the pSM19035 plasmid (36), of which a TA-containing fragment is also found in an ICE (ICESp1116) that is responsible for erm(B)-mediated inducible resistance to erythromycin in this species (37). Another S. pyogenes TA system has been detected in a prophage (Φm46.1) carrying the mef gene subclass mef(A) and tet(O) (38).

Transfer experiments demonstrated significant differences among the three strains and relevant ICEs. While ICESpy005IQ was transferred at high frequencies in both intra- and interspecific matings, no conjugal transfers were obtained with the closely related ICESpn529IQ and ICESpy029IQ. These differences were consistent with the finding that only ICESpy005IQ was excised in circular form, a critical condition for an ICE to undergo conjugal transfer (9). An excision defect might be the reason why ICESpn529IQ and ICESpy029IQ are incapable of conjugal transfer. A parallel dissimilarity between S. pyogenes Spy005, on the one hand, and S. pyogenes Spy029 and S. pneumoniae Spn529, on the other, was also observed in transformation experiments, since transformants carrying the mef(I)-catQ cassette were obtained at a high frequency from S. pyogenes Spy005 but at very low frequencies from the two other strains. The fact that transformants were obtained from S. pneumoniae Spn529 using only a crude lysate as the transforming DNA suggests possible alterations of Spn529 genomic DNA during the extraction process and is likely to account for early negative results (1). In another study, mef(I) and catQ were also cotransferred from a pneumococcus by transformation but not by conjugation (32). Transformation might be important in IQ element spread: in S. pneumoniae, Tn1207.1 (39) and the mega element (40), both transferable by transformation rather than by conjugation, play comparable roles in the spread of other mef determinants, i.e., mef(A) and mef(E), respectively. Whereas Tn1207.1 is integrated at a specific site of the pneumococcal chromosome (into the celB gene), and mef(A) dissemination is mainly clonal, the mega element has a variety of insertion sites into the pneumococcal chromosome, and its dissemination is more erratic (41). There are no data about the chromosomal insertion of IQ-like elements, which are however characterized by two different insertion sites in the three ICEs investigated here. Nonclonal dissemination of IQ-like elements is suggested by the fact that the pneumococcal isolates so far shown to be mef(I) positive all belonged to different serotypes (2, 31, 32); on the other hand, S. pyogenes Spy029 and Spy005 belong to different emm types (6). In these respects, the spread of mef(I) appears to be more reminiscent of mef(E) than of mef(A).