Φm46.1, the Main Streptococcus pyogenes Element Carrying mef(A) and tet(O) Genes

ABSTRACT Φm46.1, the recognized representative of the most common variant of mobile, prophage-associated genetic elements carrying resistance genes mef(A) (which confers efflux-mediated erythromycin resistance) and tet(O) (which confers tetracycline resistance) in Streptococcus pyogenes, was fully characterized. Sequencing of the Φm46.1 genome (55,172 bp) demonstrated a modular organization typical of tailed bacteriophages. Electron microscopic analysis of mitomycin-induced Φm46.1 revealed phage particles with the distinctive icosahedral head and tail morphology of the Siphoviridae family. The chromosome integration site was within a 23S rRNA uracil methyltransferase gene. BLASTP analysis revealed that the proteins of Φm46.1 had high levels of amino acid sequence similarity to the amino acid sequences of proteins from other prophages, especially Φ10394.4 of S. pyogenes and λSa04 of S. agalactiae. Phage DNA was present in the host cell both as a prophage and as free circular DNA. The lysogeny module appears to have been split due to the insertion of a segment containing tet(O) (from integrated conjugative element 2096-RD.2) and mef(A) (from a Tn1207.1-like transposon) into the unintegrated phage DNA. The phage attachment sequence lies in the region between tet(O) and mef(A) in the unintegrated form. Thus, whereas in this form tet(O) is ∼5.5 kb upstream of mef(A), in the integrated form, tet(O), which lies close to the right end of the prophage, is ∼46.3 kb downstream of mef(A), which lies close to the left end of the prophage.

The worldwide increase in the rates of erythromycin resistance in streptococci since the 1990s has prompted extensive epidemiological as well as molecular investigations that in the last few years have led to substantial progress in the knowledge of the mechanisms, determinants, and genetic elements involved (35). The increase in the rate of resistance is partly due to the further spread among streptococci of conventional ermclass gene-encoded methylases, the presence of which usually leads to coresistance to macrolide, lincosamide, and streptogramin B (MLS B ) antibiotics (the MLS B phenotype) (36). To an even greater extent, however, it has been due to the emergence of an active efflux-mediated mechanism, encoded by mef-class genes and associated with a pattern of low-level resistance affecting, among the MLS B antibiotics, only 14-and 15-membered macrolides (the M phenotype) (32). As a rule, a second efflux gene-an msr-class determinant usually designated msr(D)-is located immediately downstream of the mef gene.
mef(A) was discovered in Streptococcus pyogenes (16) and is by far the most common mef-class variant responsible for efflux-mediated erythromycin resistance in this species. However, the first mef(A)-carrying element, designated Tn1207.1, was detected in Streptococcus pneumoniae (30). Tn1207.1 (7,244 bp) contains eight open reading frames (ORFs), of which mef(A) and msr(D) are the fourth and the fifth; it is integrated at a specific site of the pneumococcal chromosome into celB, a late competence gene; and it has been reported to be transferable by transformation but not by conjugation. Whereas in S. pneumoniae Tn1207.1 is the sole recognized element carrying mef(A), in S. pyogenes it is not detected as such but is detected as part of larger and mobile composite elements. These elements differ depending on the tetracycline susceptibility or resistance of M-phenotype isolates of S. pyogenes (11,22); and they are all chimeric in nature; i.e., they result from a transposon (identical or related to Tn1207.1) inserted into a prophage (6,22). One of two closely related elements, Tn1207.3 (52,491 bp) (29) or ⌽10394.4 (58,761 bp) (5,6), is found in tetracycline-susceptible isolates integrated into the same chromosomal gene (comEC) and inserted into the same prophage (11), with the only difference being that in Tn1207.3, Tn1207.1 represents the left end of the element, whereas in ⌽10394.4 there is an additional left-hand region of ϳ6 kb. Since this region has been reported to be quite variable in size (6), Tn1207.3 could represent the end of the variability range in which it is completely lacking (22). In tetracyclineresistant isolates, tetracycline resistance is consistently mediated by the tet(O) determinant and mef(A) is linked to tet(O) in a mobile, phage-like element (21,22). In fact, there are a variety of related tet(O)-mef(A) phage-like elements, in which mef(A) is contained in a range of changeable and defective variants of Tn1207.1 (11,22). The most common such element, of which ⌽m46.1 is the typical representative, is not integrated within the comEC gene and has been transferred to an S. pyogenes recipient in mating experiments (21). An ϳ12-kb region of ⌽m46.1 encompassing the tet(O) gene and the Tn1207.1-related transposon has been sequenced (11).
In the study described here, we determined the site of inte-gration into the chromosome, the complete genome sequence and organization, and the ultrastructure of ⌽m46.1. The genome displayed the distinctive modular arrangement of tailed bacteriophages, and electron microscopic analysis confirmed that it has the distinctive morphology of Siphoviridae family bacteriophages. Phage DNA (55,172 bp) was present in the host cell both as a prophage and as free circular DNA. The sequences of the ORFs of ⌽m46.1 were compared with those from protein databases.

MATERIALS AND METHODS
Bacterial strain. S. pyogenes m46 was used in this study. The strain, originally collected as a throat clinical isolate belonging to M type 4 (34), is resistant to both erythromycin [MIC, 16 g ml Ϫ1 ; M phenotype; mef(A) genotype] and tetracycline [MIC, 64 g ml Ϫ1 ; tet(O) genotype] and was used in all previous experiments that led to the identification of the genetic linkage between the tet(O) and the mef(A) genes in a mobile, prophage-associated element eventually designated ⌽m46.1 (3,11,21,22).
Gene detection and amplification experiments. The principal primer pairs used in the PCR experiments are listed in Table 1. DNA preparation and amplification and electrophoresis of the PCR products were carried out by established procedures and according to the conditions recommended for the use of the individual primer pairs. The Ex Taq system (TaKaRa Bio, Shiga, Japan) was used when the expected sizes of the PCR products exceeded 3 kb.
Phage DNA sequencing and sequence analysis. Long PCR experiments, inverse PCR, and primer-walking techniques were used to obtain overlapping fragments of the ⌽m46.1 prophage with primers designed from ⌽10394.4 (EMBL accession no. NC_006086). Inverse PCR was carried out as described by Sambrook and Russell (28); HindIII-, MboI-, or SexAI-digested genomic DNA (all endonucleases were from Roche Applied Science, Basel, Switzerland) was ligated and used as the template in the PCR assays (Table 1). All PCR products used for sequence analysis were purified by using Montage PCR filter units (Millipore Corporation, Bedford, MA). Amplicons were sequenced (bidirectionally or by primer walking) with an ABI Prism sequencer (Perkin-Elmer Applied Biosystems, Foster City, CA) and dye-labeled terminators. The sequences were analyzed by using the Sequence Navigator software package (Perkin-Elmer Applied Biosystems). ORF analysis was performed by using the software ORF Finder available online (http://www.ncbi.nlm.nih.gov/projects/gorf/). The criteria used to designate a potential 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 by using the 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/). Electron microscopy. For phage induction, S. pyogenes strain m46 was treated with 0.2 g ml Ϫ1 mitomycin C (Sigma Chemical Co., St. Louis, MO) for 4 h at 37°C. The bacteria were then centrifuged at 8,000 ϫ g for 15 min, the supernatant was filtered through a 0.45-m-pore-size nylon membrane (Millipore) and centrifuged at 141,000 ϫ g for 2 h at 10°C, and the pellet was suspended in 1 ml of phage suspension buffer (0.15 M NaCl, 10 mM Tris HCl [pH 7.5], 5 mM MgCl 2 , 1 mM CaCl 2 ). A drop of sample was allowed to absorb to carbon-coated 300 mesh grids for 1 min, and 1 drop of water was used to wash the grids. The sample was stained with 4 drops of 3% aqueous uranyl acetate (Sigma) for 30 s. The excess liquid was wicked off and the grids were air dried. The samples were viewed under a CM10 transmission electron microscope (Philips, Eindhoven, The Netherlands) operating at 80 kV with a magnification of ϫ53,000.
Nucleotide sequence accession number. The complete genome sequence of ⌽m46.1, with its chromosome integration site, has been submitted to the EMBL database under accession no. FM864213.

RESULTS AND DISCUSSION
Bacteriophage ⌽m46.1 integration site. Use of the portion of ⌽m46.1 that included the tet(O) and mef(A) genes (11) as a query in a BLASTN analysis of the ϳ12-kb sequence-previously determined in our laboratory (EMBL accession no. AJ715499)-yielded a 308-bp region common to the Streptococcus agalactiae A909 genome sequence (EMBL accession no. CP000114) (33). In our sequence, this region is upstream of the Tn1207.1-like transposon. In the S. agalactiae A909 genome, the region (bases 652,826 to 653,124) carries the 3Ј end (36 bp) of the rumA gene, which encodes a 23S rRNA uracil methyltransferase, and part of the intergenic region (272 bp) between rumA (SAK_0718 locus) and the adjacent gene (SAK_0719 locus). It is worth noting that the S. agalactiae rumA gene is homologous to a 23S rRNA uracil methyltransferase gene detectable in all S. pyogenes genomes sequenced so far; the highest degree of homology (73%) was with the gene from S. pyogenes MGAS10750 (EMBL accession no. CP000262), a strain that also shares M type 4 with S. pyogenes m46 (9). PCR experiments were carried out with two primer pairs, one for the left junction and one for the right junction, to explore the possibility that this region was the ⌽m46.1 chromosome integration site (Fig. 1). The two primers for the left junction were LYT-for, internal to Spy1198, an ORF of the MGAS10750 genome just upstream of the 23S rRNA uracil methyltransferase gene (Spy1197), and MEFA2, internal to the mef(A) gene (Fig. 1A). The two primers for the right junction were TETO1, internal to the tet(O) gene, and THIOrev, internal to Spy1195 of the MGAS10750 genome, located downstream of Spy1197 (Fig. 1B). By pairing primers LYT-for and MEFA2, a 4,425-bp segment at the attL region was amplified from the lysogenic bacterial DNA template. This amplicon sequence was aligned with the ⌽m46.1 genome (see the next paragraph) and with the MGAS10750 genome. Positions 1,727 to 1,764 of the amplicon were found to be the same as positions 55,156 to 20 of the ⌽m46.1 genome, and positions 1,701 to 1,744 were found to be the same as positions 1,142,452 to 1,142,409 of the S. pyogenes MGAS10750 genome (Fig. 1A).
An almost completely overlapping 18-bp sequence was found in the ⌽m46.1 and MGAS10750 genomes. By pairing primers TETO1 and THIO-rev, a 6,573-bp segment at the attR region was amplified from the lysogenic bacterial DNA template. Similar alignment assays revealed that positions 5,298 to 5,357 of the amplicon were the same as positions 55,125 to 55,172 of the ⌽m46.1 genome and that positions 5,341 to 5,361 were the same as positions 1,142,426 to 1,142,393 of the MGAS10750 genome (Fig. 1B). The same 18-bp overlapping sequence found in the ⌽m46.1 and MGAS10750 genomes was also detected when the left junction was analyzed. This 18-bp sequence, shared by both the phage [between tet(O) and mef(A)] and the host bacterium (near the 3Ј end of the 23S rRNA uracil methyltransferase gene), should be the core site, i.e., the critical sequence where the site-specific recombination process presumably takes place. Definition of the core site enabled the integrative reaction and structure relationship between the phage genome and the host chromosome to be deduced according to the integration mechanism elucidated in bacteriophage lambda (12,13). Interestingly, the 23S rRNA uracil methyltransferase gene identified as the chromosome integration site of ⌽m46.1 in S. pyogenes m46 is the same as two integrated conjugative elements (ICEs), 2096-RD.2 and 6180-RD.1, in S. pyogenes strains MGAS2096 and MGAS6180, respectively (8).
Crucially, identification of the core site provided the knowledge that the phage attachment sequence (attP) falls right in the region between tet(O) and mef(A) in the previously sequenced ϳ12-kb DNA fragment of ⌽m46.1. Therefore, both that sequence and the previously disclosed tet(O)-mef(A) linkage (21) Table  2. The ORF sequences were compared with sequences from protein databases by using the BLASTP program. On the basis of these comparisons, most of the ORFs could be assigned to different modules, according to a modular organization typical of tailed phages (14). Interestingly, both ends of the ⌽m46.1 genome were represented by nonphage DNA.
(i) Initial portion (1 to 248 bp). The ⌽m46.1 prophage started with a short sequence (248 bp) that preceded the Tn1207.1-like transposon. The first 22 bp of this sequence restored the gene encoding the 23S rRNA uracil methyltransferase, whereas its first 222 bp displayed 88% homology with a sequence close to the left end of the pneumococcal mega element (EMBL accession no. AF274302).
(iii) Phage modules (7,390 to 48,807 bp). orf8 and orf9, which represent the beginning of the actual prophage-like region, probably corresponded to the left part (7,390 to 8,309 bp) of the lysogeny control module, which in ⌽m46.1 appears to be divided into two portions due to the insertion of a segment including a tet(O)-containing fragment from ICE 2096-RD.2 (8) and the mef(A)-containing Tn1207.1-like transposon. The DNA replication module (11,363 to 20,144 bp), spanning from orf14 to orf22, was highly conserved in the three mef(A)-carrying phage-like elements ⌽m46.1, Tn1207.3, and ⌽10394.4. The DNA modification module (22,539 to 26,115 bp) spanned from orf27 to orf30. The DNA packaging and head morphogenesis module (26,993 to 33,651 bp) spanned from orf33 to orf42. Inside but apparently alien to the module, orf35 and orf36 seemed to be related to a toxin-antitoxin (TA) system (26) and might contribute to the stable maintenance of ⌽m46.1 in the bacterial population. It is noteworthy that the two putative TA-related genes partially overlapped. The tail morphogenesis module (34,336 to 39,108 bp) was formed by four ORFs (orf45 to orf48). Just downstream of this module, orf49 and orf50 (39,387 to 43,656 bp) encoded proteins exhibiting low levels of amino acid identity to distinct portions (for the presence of a stop codon) of a PblB protein. Proteins PblA and PblB act as phage-borne virulence factors by promoting bacterial binding to human platelets (7). The host cell lysis module (44,227 to 46,099 bp) was formed by three ORFs (orf53 to orf55), all of which were shared by ICE Sde3396, a newly described genetic element from Streptococcus dysgalactiae subsp. equisimilis (17). Downstream of the host cell lysis module, orf56 and orf57 (46,322 to 48,807 bp) encoded two sitespecific recombinases (serine recombinases belonging to the resolvase family), i.e., enzymes that are usually found in the lysogeny control module (25). In ⌽m46.1, the two recombinase ORFs might have been separated from the rest of the lysogeny module by the insertion into the unintegrated phage DNA of a segment including the ICE 2096-RD.2 fragment and the Tn1207.1-like transposon.   017 to 52,098 bp). A ϳ3-kb fragment highly homologous (98%) to a region of 2096-RD.2, a 63-kb ICE-like element of S. pyogenes MGAS2096 harboring several antibiotic resistance genes (8), was found immediately downstream of the two site-specific recombinase-encoding ORFs. This region, which in the MGAS2096 genome spans from bases 1,101,217 to 1,103,555 (EMBL accession no. CP000261), in ⌽m46.1 includes three ORFs, orf58, orf59, and orf60, corresponding to Spy1150, Spy1149, and Spy1148, respectively, in the MGAS2096 genome. orf59 was the tetracycline resistance gene tet(O), whose presence in S. pyogenes we first demonstrated (21) and further investigated (11) in strain m46. It is worth noting that inverse PCR assays were needed to sequence the unknown DNA region upstream of the tet(O) gene (from approximately orf53 to orf58).
(v) The last three ORFs (52,351 to 55,079 bp). The final region of the ⌽m46.1 prophage included the last three ORFs (orf61, orf62, and orf63), previously designated orfA, orfB, and orfC, respectively (11). Immediately downstream of orf63, a 76-bp sequence displayed 91% homology with a sequence at the right end of the pneumococcal mega element. Consideration of this short area of homology and the one (222 bp) described above highlights an intriguing correspondence be-tween the ends of the ⌽m46.1 prophage and those of the mega element.
Comparative phage genomics. BLASTP analysis revealed that the proteins of ⌽m46.1 had very high levels of amino acid sequence similarity to the amino acid sequences of proteins from other prophages, namely, ⌽10394.4 of S. pyogenes and Sa04 of S. agalactiae. The two phages have quite a different prominence in the literature. ⌽10394.4, detected in the genome of S. pyogenes MGAS10394 (EMBL accession no. CP000003), has been the subject of extensive specific investigations (5,6,11,19,20,22) and is also found in GenBank under a separate accession number (GenBank accession no. AY445042). In contrast, Sa04, detected in the genome of S. agalactiae A909, is mentioned only in the deposited sequence of the entire bacterial genome (GenBank accession no. CP000114), but it has never been the subject of specific investigations and even went unmentioned in the paper in which the genome analysis of S. agalactiae A909 was described (33). The relationship of ⌽m46.1 with ⌽10394.4 and Sa04, documented as database matches in Table 2, is illustrated in Fig. 2 by the alignment of the ORF map of ⌽m46.1 with the maps of the two other prophages (only amino acid identities Ն70% are reported). (i) ⌽m46.1 and ⌽10394.4 prophage comparison. ⌽m46.1 and ⌽10394.4 shared three major areas of homology, namely (from left to right), two clusters of closely related genes and a third cluster of moderately related genes. The first two clusters were separated by an area of nonhomology represented in ⌽10394.4 by four ORFs, i.e., a restriction-modification cassette (orf16 to orf18) and orf19 (19), and in ⌽m46.1 by three unrelated ORFs (orf10 to orf12) (3). It is worth noting that while the DNA of tetracycline-resistant M-phenotype isolates, which typically carry ⌽m46.1 or a related tet(O)-mef(A) element, is usually digested by SmaI, tetracycline-susceptible M-phenotype isolates, which typically carry ⌽10394.4 or Tn1207.3, are SmaI nontypeable because a DNA-modifying methyltransferase encoded by the spyIM gene (orf16 in ⌽10394.4) acts on the SmaI recognition sequence and makes the DNA refractory to cleavage by SmaI (3,19,20).
In ⌽m46.1, the first cluster included part of the Tn1207.1like transposon (orf2 to orf7) and the left portion of the lysogeny module (orf8 and orf9). In ⌽10394.4, these ORFs corresponded to orf8 to orf15; the levels of amino acid identity were very high (mostly Ͼ90%). In ⌽m46.1 the second cluster spanned from orf13 to orf27, including the DNA replication module and the beginning of the DNA modification module. In ⌽10394.4, these ORFs corresponded to orf20 to orf33; the levels of amino acid identity were Ͼ90% for most correlated ORFs. The third cluster of ⌽m46.1 spanned from orf29 to orf46, including part of the DNA modification module, the DNA packaging and head morphogenesis module, and the beginning of the tail morphogenesis module. In ⌽10394.4, these ORFs corresponded to orf35 to orf51; the levels of amino acid identity were lower (Ͻ90%) than those for the first two clusters.
(ii) ⌽m46.1 and Sa04 prophage comparison. ⌽m46.1 and Sa04 shared three major areas of homology, namely (from left to right), a first cluster of moderately related genes, a second cluster of closely related genes, and a third cluster (in fact, a couple of ORFs) of moderately related genes.
In ⌽m46.1 the first cluster spanned from orf13 to orf27, corresponding in Sa04 to orf1 to orf14; the levels of amino acid identity were Ͻ90% for most correlated ORFs. The second cluster of ⌽m46.1 spanned from orf29 to orf46, corresponding in Sa04 to orf15 to orf31; the levels of amino acid identity were Ͼ90% for most correlated ORFs. Interestingly, in both ⌽m46.1 (orf35 and orf36) and Sa04 (orf20 and orf21), this second cluster of genes included two putative TA-related genes that were lacking in ⌽10394.4. It is remarkable that in the two bacteriophages the two couples of TA-related genes, even though they encoded proteins with no significant amino acid sequence identities, were found in the same position, i.e., immediately downstream of an ORF (orf34 in ⌽m46.1, orf19 in Sa04) encoding a large terminase subunit. In ⌽m46.1, the third area of homology was represented by orf56 and orf57 (encoding the two site-specific recombinases). In Sa04, these two ORFs corresponded to orf40 and orf41, with the levels of amino acid identity being 70% and 76%, respectively.
Phage ultrastructure. After induction with mitomycin C, electron microscopic analysis revealed phage particles with the typical icosahedral head and tail morphology of the Siphoviridae (Fig. 3), the most common phage family in streptococci (1). That the phage particles were indeed ⌽m46.1 is consistent with the findings of pulsed-field gel electrophoresis experiments, which demonstrated that it is the only prophage carried by S. pyogenes m46 that is inducible by mitomycin C (data not shown). On the other hand, the modular organization of the ⌽m46.1 genome was that typical of tailed phages (14), and a similar ultrastructure has been reported for ⌽10394.4 (6).
Concluding remarks. ⌽m46.1, whose complete sequence analysis and final characterization were the aim of this study, is the recognized representative of the most common variant of the so-called tet(O)-mef(A) elements, responsible for effluxmediated erythromycin resistance in tetracycline-resistant S. pyogenes isolates (35). In an early study (21), genes mef(A) and tet(O) were detected in S. pyogenes m46, were cotransferred to a susceptible recipient of the same species, and were found to be linked, with mef(A) being detected ϳ5.5 kb downstream of tet(O); a single new DNA insertion into the transconjugants with the mef(A) tet(O) genotype was consistent with a chromosomal location of the two genes. Subsequent investigations (11) demonstrated a variety of closely related tet(O)-mef(A) elements harboring a range of changeable and defective variants of Tn1207.1, of which the element detected and originally investigated in strain m46 was the most common. Mitomycin C induction experiments showed that the tet(O)-mef(A) elements were in fact prophages (22). The present study has conclusively clarified that the chromosome integration site is within the 23S rRNA uracil methyltransferase gene (near its 3Ј end). In the host cell, ⌽m46.1 exists not only as a prophage but also as free circular DNA. While in the latter form tet(O) is found ϳ5.5 kb upstream of mef(A), in the integrated form it is close to the right end of the prophage (orf59 of 63 ORFs), ϳ46.3 kb downstream of mef(A), which is close to the left end of the prophage (the third ORF). Accordingly, the designation "mef(A)-tet(O) elements" would be more appropriate than the one "tet(O)mef(A) elements" (11,22,35), which has so far been used to indicate these genetic elements.
It is well established that each prophage is a unique entity that not only shares blocks of sequences with different prophages but that also possesses unique sequences with no known homologies in current databases. This genetic mosaicism is a hallmark of tailed phages and reflects an unusually high degree of horizontal genetic exchange in phage evolution (23,24). Genome mosaicism is characterized by the presence of novel sequence joints, in which the similarity between two phages abruptly ceases; other phages and bacterial hosts may be the sources for such new sequences (14). ⌽m46.1 and ⌽10394.4, whose genomes are highly mosaic, appear to fit well into the general model. One explanation proposed for the origin of Major examples of such diversification (in ⌽m46.1 compared to ⌽10394.4) are the lack of the initial ϳ6-kb left-hand region, also lacking in Tn1207.3; orf1, in place of the initial portion of Tn1207.1; three new ORFs (orf10 to orf12) in place of the restriction-modification cassette; a putative TA system (orf35-orf36); the host cell lysis module (orf53 to orf55), shared by and possibly acquired from ICE Sde3396; the tet(O)-containing fragment (orf58 to orf60), shared by and possibly acquired from ICE 2096-RD.2; and the last three ORFs (orf61 to orf63), likely of chromosomal origin. In particular, as far as the two antibiotic resistance genes carried by ⌽m46.1 are concerned, the present findings support the hypothesis of the stepwise acquisition of mef(A) and tet(O) (11). These data are consistent with the current belief in the key role of bacteriophages in the evolution of important bacterial pathogens: on the one hand, by carrying a versatile range of new genetic information within and between bacterial species and on the other hand, by rearranging existing genetic information in unique combinations (10). This is particularly true of S. pyogenes, a species distinguished by a unique propensity to acquire and reshuffle phage-encoded virulence and resistance determinants (4). The highly mosaic genome of ⌽m46.1, in which different segments are related to distinct streptococcal phages, entails that phages of S. pyogenes continue to exchange genetic material, contributing to the extraordinary horizontal transfer of mobile genetic elements in this species. Such a formerly underestimated role of prophages in clonal diversification has even led some investigators to question the validity of the conventional associations of certain M serotypes with specific clinical manifestations of S. pyogenes infections and to suggest the need for a new classification scheme that better represents the genetic bases of S. pyogenes virulence (2).