Genetic basis of the association of resistance genes mef(I) (macrolides) and catQ (chloramphenicol) in streptococci

In streptococci mef(I) and catQ, two relatively uncommon macrolide and chloramphenicol resistance genes, respectively, are typically linked in a genetic module designated IQ module. Though variable, the module consistently encompasses, and is sometimes reduced to, a conserved ∼5.8-kb mef(I)-catQ fragment. The prototype IQ module was described in Streptococcus pneumoniae. IQ-like modules have subsequently been detected in Streptococcus pyogenes and in different species of viridans group streptococci, where mef(E) may be found instead of mef(I). Three genetic elements, one carrying the prototype IQ module from S. pneumoniae and two carrying different, defective IQ modules from S. pyogenes, have recently been characterized. All are integrative and conjugative elements (ICEs) belonging to the Tn5253 family, and have been designated ICESpn529IQ, ICESpy029IQ and ICESpy005IQ, respectively. ICESpy029IQ and ICESpy005IQ were the first Tn5253 family ICEs to be described in S. pyogenes. A wealth of new information has been obtained by comparing their genetic organization, chromosomal integration, and transferability. The origin of the IQ module is unknown. The mechanism by which it spreads in streptococci is discussed.

The macrolide resistance gene mef(I) and the chloramphenicol resistance gene catQ are relatively uncommon genetic determinants accounting respectively for macrolide and chloramphenicol resistance. In streptococci, however, they are typically associated and genetically linked, since neither has yet been detected without the other. This review illustrates current data on the association and discusses its genetic basis.

mef GENES
In streptococci, mef-class genes encode efflux-mediated macrolide resistance. Compared to erm-class genes encoding methylasemediated target site modification, mef genes usually produce lower-level resistance, only affecting 14-and 15-membered macrolides (M phenotype; Sutcliffe et al., 1996). In streptococci, mef genes include a number of subclasses, of which mef(A) and mef(E) are the most significant. mef(A), the first subclass to be discovered (Clancy et al., 1996), is widespread in Streptococcus pyogenes but is also common in Streptococcus pneumoniae. mef (E), discovered in S. pneumoniae (Tait-Kamradt et al., 1997), is frequently found in this and other Streptococcus species, but is uncommon in S. pyogenes (Klaassen and Mouton, 2005). mef(I), discovered in S. pneumoniae (Cochetti et al., 2005; accession no. AJ971089) and subsequently detected in S. pyogenes (Blackman Northwood et al., 2009) and viridans group streptococci (VGS; Brenciani et al., 2014), is the most widely investigated among the other mef subclasses; these include mef(O), mef(B), and mef(G), detected respectively in S. pyogenes (Sangvik et al., 2005), Streptococcus agalactiae (Cai et al., 2007), and group G streptococci (Amezaga and McKenzie, 2006;Cai et al., 2007). An msr-class gene with homology to msr(A), involved in macrolide efflux in Staphylococcus aureus (Ross et al., 1990), is located immediately downstream of the mef gene: it is usually designated msr(D), although a number of variants are associated with different mef genes. Cotranscription of mef(E) and msr(D) in S. pneumoniae suggested that the products of the two genes may act as a dual efflux system (Gay and Stephens, 2001).

cat GENES
The first and still predominant mechanism of bacterial resistance to chloramphenicol is enzymatic inactivation of the drug by different chloramphenicol acetyltransferases (CATs) encoded by cat genes. Two types of CATs have been recognized, A and B, each including a number of structurally and phylogenetically diverse groups (Schwarz et al., 2004). The catQ gene, encoding a type-A CAT assigned to a distinct group (Schwarz et al., 2004), was originally described (Rood et al., 1989) and then sequenced (Bannam and Rood, 1991;accession no. M55620) in Clostridium perfringens CW531. However, no further catQ gene from this species has been described or is available in GenBank, even though genes with 77-92% DNA identities to catQ from C. perfringens CW531 can be found in several other Clostridium species by in silico analysis. Different type-A CATs encoded by a number of cat genes have been described in streptococci (Trieu-Cuot et al., 1993;Schwarz et al., 2004). With regard to catQ, an early study reported a catQlike gene in a S. agalactiae strain (Trieu-Cuot et al., 1993). More recently, Mingoia et al. (2007) described catQ in two S. pneumoniae isolates and found that the catQ gene from S. pneumoniae Spn529 (the reference isolate) displayed 96.2% DNA identity www.frontiersin.org (accession no. AJ971089) to the original catQ from C. perfringens CW531.

THE IQ MODULE
When catQ was first detected in S. pneumoniae, it was found to be linked to mef(I) in an ∼15.1-kb DNA fragment, containing two identical tnp1 transposase genes at either end, that was designated 'IQ element' (Mingoia et al., 2007). Here, we prefer to call it 'IQ module,' since the term 'element' might be confusing. Two defective IQ modules were subsequently described in S. pyogenes: one was characterized by a shorter region upstream of mef(I), formed only by tnp1 and a truncated recombinase gene rec2; the other consisted of the sole mef(I)-catQ fragment (Del Grosso et al., 2011). Very recently, a number of variously defective IQ-like modules have been detected in different VGS species (Brenciani et al., 2014). An interesting characteristic of VGS modules is that they most often bear mef(E) instead of mef(I). Moreover, all IQ-like modules detected in VGS lack the right tnp1, i.e., the transposase gene downstream of catQ that is found in the IQ module of S. pneumoniae. Despite variability upstream of the mef gene and downstream of catQ, all reported IQ modules appear to share a conserved mef-catQ fragment [six open reading frames (ORFs), ∼5.8 kb; Figure 1].
The origin of the IQ module is unknown. The prototype IQ module of pneumococcal origin has been proved to be capable of undergoing excision in circular form, but was unable to be transferred to a recipient .

FROM THE 5216IQ COMPLEX TO A Tn5253 FAMILY ICE
The first attempt to clarify the genetic basis of the mef(I)-catQ association dates back to 2007, when Mingoia et al. (2007) identified in S. pneumoniae a composite structure, that they designated 5216IQ complex (∼30.5 kb); it consisted of two moieties, one represented by the IQ module and the other formed by fragments of transposons Tn5252 and Tn916. The Tn5252 fragment corresponds to the conjugal transfer-related (CTR) functional module of Tn5252 (Alarcon-Chaidez et al., 1997); the Tn916 fragment contains a silent tet(M), unexpressed because it lacks the promoter, the ribosome-binding site, and part of the leader peptide. As realized later  also thanks to the new notion of integrative and conjugative element (ICE; Seth-Smith and Wozniak and Waldor, 2010), the so-called 5216IQ complex is not in fact a genetic element (Varaldo et al., 2009), but rather part of a larger element, a Tn5253 family ICE.

GENETIC ELEMENTS CARRYING IQ MODULES
The genetic location of the IQ module has recently been investigated in three strains: one S. pneumoniae (Spn529, serotype 11A [Mingoia 07]) and two S. pyogenes (Spy029 and Spy005, respectively emm types 25.0 and 12.42 [Del Grosso 11]). All three IQ modules were found to be inserted in Tn5253 family ICEs, which were respectively designated ICESpn529IQ, ICESpy029IQ, and ICESpy005IQ . It is worth noting that this was the first time that Tn5253 family ICEs were described in natural isolates of S. pyogenes. The family  (Brenciani et al., 2014); (E) S. oralis (Brenciani et al., 2014); (F) S. mitis, S. sanguinis, S. parasanguinis (Brenciani et al., 2014). The ORFs are depicted as arrows pointing in the direction of transcription: the mef gene is diagonally striped (green, mef (I); blue, mef (E)); catQ is red; the transposase gene tnp1 is black; and the other ORFs are white.
prototype, Tn5253, is a long-established conjugative composite element resulting from the insertion of Tn5251 (virtually identical to Tn916) into Tn5252 (Ayoubi et al., 1991). Several studies have recently addressed the heterogeneity of Tn5253like composite elements (Croucher et al., , 2011Ding et al., 2009;Henderson-Begg et al., 2009;Mingoia et al., 2011;Wyres et al., 2013), whose integrase genes may belong to one of two types: int 5252 (the gene of the Tn5253 prototype) or int Sp23FST81 [first described in ICESp23FST81 ]. The complete sequence of Tn5253 is available (accession no. EU351020) and has been extensively analyzed (Iannelli et al., 2014).
ICESpn529IQ has been completely sequenced (accession no. HG965092): it is 59,466 bp long; its G+C content is 36%; and sequence analysis disclosed 66 ORFs. The IQ module spans from orf22 to orf36 (which correspond to the two identical tnp1 transposase genes located at either end of the module); catQ is orf23 and mef(I) is orf28. ICESpy029IQ and ICESpy005IQ have been characterized based on PCR mapping, partial DNA sequencing, and restriction analysis. The three ICEs are shown and compared in Figure 2.
ICESpn529IQ and ICESpy029IQ appear to share a very similar organization. Both carry the integrase gene int Sp23FST81 and an identical Tn916 fragment. Aside from the different size of their IQ modules, which nevertheless share the same position and insertion site in the respective ICE, the only noticeable difference between the two ICEs is a transposase gene tnp1 adjacent to int Sp23FST81 ; as a result ICESpy029IQ carries a third tnp1 in addition to those found at the ends of the IQ module. In contrast, ICESpy005IQ has the integrase gene int 5252 and no Tn916 fragment, and its defective IQ module is found in a different position compared to the IQ modules of ICESpn529IQ and ICESpy029IQ.
All three ICEs lack the linearized plasmid pC194, which contains the chloramphenicol resistance determinant cat pC194 and is usually a distinctive cargo of the Tn5252-like moiety of Tn5253 family ICEs. This might be a fitness-related characteristic: as if the ICE no longer needed the plasmid since chloramphenicol resistance is already assured by catQ. It is well known since early studies that the DNA region containing the linearized pC194 is flanked by direct repeats whose recombination may lead to spontaneous curing of the region (Ayoubi et al., 1991;Kiliç et al., 1994). Recent analysis of the sequence of Tn5253 (Iannelli et al., 2014) disclosed a 7,627-bp cat(pC194) flanked by two 1,169-bp direct FIGURE 2 | Schematic representation of ICESpn529IQ, completely sequenced, and of ICESpy 029IQ and ICESpy 005IQ, characterized by PCR mapping, partial DNA sequencing and restriction analysis. The ICE genomes are represented as white rectangles with some ORFs and some modules or fragments in evidence. The IQ module is depicted as a gray rectangle, the Tn916 fragment as a spotted rectangle, and the CTR module as a dotted rectangle. The ORFs are depicted as arrows pointing in the direction of transcription, including mef (I) (diagonally striped green), catQ (red), tnp1 (black), integrase genes (yellow), TA gene pair (purple), and silent tet (M) (blue). orf66 of ICESpn529IQ is white. The progressive number is reported only below selected ones of the 66 ORFs of ICESpn529IQ.
repeats, which contain a toxin-antitoxin gene pair (Palmieri et al., 2013) that has previously been detected in the pathogenicity island PPI-1 of S. pneumoniae strain TIGR4 (Khoo et al., 2007). In all three ICEs the lack of the integrated plasmid pC194 is associated with the presence in the corresponding region of a single copy of this toxin-antitoxin operon (TA pair), which likely represents the recombination site of cat(pC194) and might be the outcome of repair after plasmid excision. Of the three ICEs, only ICESpy005IQ has been detected as a free circular form, a critical condition for an ICE to undergo conjugal transfer (Wozniak and Waldor, 2010).

TRANSFERABILITY OF mef (I) AND catQ
The transferability of mef(I) and catQ varies with the transfer mechanism (conjugation or transformation) and the strain harboring the IQ module. Using S. pneumoniae donors, transfer of mef(I) and catQ has never been obtained to any recipient by conjugation, whereas mef(I)-and catQ-positive transformants have been obtained by transformation (Nielsen et al., 2010;Mingoia et al., 2014). It is worth noting that, in experiments using S. pneumoniae Spn529 as the donor, transformants were obtained at a very low frequency (around 10 −9 ) and only using a crude lysate as the transforming DNA: this suggests that DNA may undergo changes during the extraction process, which may account for the early negative results (Mingoia et al., 2007). Conversely, using S. pyogenes donors, conjugation assays yielded no transconjugants from Spy029, whereas transconjugants harboring the entire ICESpy005IQ were obtained at high frequencies from Spy005 (Del Grosso et al., 2011;Mingoia et al., 2014). Transformants were obtained with both S. pyogenes donors, but at a high frequency from strain Spy005 and at a very low frequency from strain Spy029.

SPREAD OF mef (I) AND catQ IN STREPTOCOCCI
It has been argued by analogy with other mef gene subclasses (especially in pneumococcal populations) that the spread of mef(I) -and of the IQ module, where it is consistently found -could largely depend on transformation mechanisms . In S. pneumoniae, the two major mef determinants, mef(E) and mef(A), are carried respectively by the mega element (Gay and Stephens, 2001) and Tn1207.1 (Santagati et al., 2000), whose relationship with the relevant mef gene is roughly comparable to that of the IQ module with mef(I). Similar to the IQ module, the mega element and Tn1207.1 are transferable by transformation rather than by conjugation. While the mega element has a variety of insertion sites in the pneumococcal chromosome and its dissemination is erratic, Tn1207.1 is integrated into the chromosome at a specific site and its dissemination is mainly clonal (Del Grosso et al., 2002). Nonclonal dissemination of the IQ module is suggested by the wide range of serotypes of the S. pneumoniae isolates that have so far been found to be mef(I)-positive (Cochetti et al., 2005;Reinert et al., 2008;Nielsen et al., 2010); moreover S. pyogenes strains Spy029 and Spy005 also belong to different emm types (Del Grosso et al., 2011). Altogether, these data suggest that the behavior of mef(I) is likely to be more similar to that of mef(E) than of mef(A).

CONCLUDING REMARKS
In streptococci, the macrolide resistance determinant mef(I) and the chloramphenicol resistance determinant catQ, both relatively uncommon within the respective class of resistance genes, are typically linked in a genetic module, designated IQ module. Although the IQ module is uncommon in streptococci, it is likely to be unique to them: it has not been reported outside the genus Streptococcus, and its origin is unknown. Though variable, the IQ module consistently encompasses, and may be reduced to, a conserved ∼5.8-kb DNA fragment spanning from mef(I) (or mef(E) in some VGS) to catQ. After early report in S. pneumoniae (Mingoia et al., 2007), IQ-like modules have been detected in S. pyogenes (Del Grosso et al., 2011) and VGS (Brenciani et al., 2014); moreover, Mingoia et al. (2014) have recently reported on ongoing studies of isolates of other Streptococcus species (S. agalactiae) harboring linked mef(I) and catQ genes.
In recent years, neither mef(I) nor catQ have been described in the documented absence of the other in streptococci. However, the association is unlikely to have been the rule in earlier investigations. Indeed, the S. agalactiae strain where a catQ-like gene was first described in a streptococcus was erythromycin susceptible (Trieu-Cuot et al., 1993). The same is probably true of clostridia: C. perfringens CW531, where catQ was originally detected and sequenced even though its genetic environment was not investigated, was reported to be resistant, besides chloramphenicol, to tetracycline, but not to erythromycin (Rood et al., 1989); and there is no evidence of associated mef genes in the sequenced clostridial genomes or contigs containing catQ-related genes.