CcrB typing tool: an online resource for staphylococci ccrB sequence typing

Abstract

Sir, In 2006, we published a manuscript in this journal describing the allelic variation in the ccrAB locus in a representative collection of methicillin-resistant Staphylococcus aureus (MRSA) based on DNA sequencing of internal fragments of both genes. This work confirmed the very close relationships among ccrAB alleles associated with SCCmec types I–IV and VI, which was found to be independent of the MRSA lineage, geographic origin or isolation period. Moreover, particularly for the ccrB gene, SCCmec types II and IV, both defined by ccrAB allotype 2, could be discriminated. This method provides a significant improvement in ccrAB typing resolution since these SCCmec types have different epidemiological characteristics: type II is mostly found among hospital-acquired MRSA (e.g. ST5/ USA100 and ST36/EMRSA-16 clones), whereas type IV is mostly found among community-acquired MRSA (e.g. ST80, ST30 and ST1 clones). Based on these observations and since SCCmec types are defined based on the ccrAB allotype and the genetic organization of the mecA locus, we have proposed that sequencing an internal fragment of ccrB could be used as a SCCmec typing strategy, either as a first-line assay or as a confirmation tool for SCCmec type assignments. From a practical perspective, ccrB sequence typing can be easily incorporated in other widespread sequence-based MRSA typing strategies, such as multilocus sequencing (MLST) and spa typing. Under this rationale, we have developed an online resource for storage and automatic analysis of ccrB internal sequences obtained using our previously published protocol. The so-called ‘ccrB typing tool’ was launched in late October 2007 and is freely available at http://www.ccrbtyping.net. A detailed tutorial is available online as well the contacts of the site developers. Users can access the ccrB typing online database either anonymously or as registered users; registration is required for submission of data to the public database or to create a personal and private online database of ccrB alleles. Users can paste ccrB internal sequences in the FASTA format, which can be automatically trimmed to fix the sequence length for analysis at 455 bp. Then, after an automatic multiple sequence alignment to the database known ccrB alleles, the user’s sequence is either assigned to a ccrB allele (based on a 100% homology) or to a new one, if a homology between 90% and 100% is found to any of the available alleles. If a new allele is found, the most similar allele is indicated and, after submission to the public database, an allele number is assigned. Based on this assignment, a prediction of the ccrAB allotype and SCCmec type is also outputted. The user can also check all outputs by inspecting a graphical display of the multiple sequence alignment and the reconstruction of neighbour-joining or average-distance trees available through a Java applet. Users can also select subsets of private and public sequences to run the multiple sequence algorithm and visualize the resulting trees. If users choose to submit their data to the public database, the submission process is validated by a curator that checks for data consistency and quality. If a new ccrB allele is found, users are requested to upload both trace files. Upon development of the ‘ccrB typing tool’, we have deposited all sequences described in Oliveira et al. and also all ccrB sequences available at GenBank (www.ncbi.nlm.nih.gov, last accessed on 16 November 2007) covering the same 455 bp used in the ccrB typing tool. Besides ccrB sequences for S. aureus, 13 sequences for coagulase-negative staphylococci (CoNS), such as Staphylococcus epidermidis, Staphylococcus hominis, Staphylococcus saprophyticus and Staphylococcus warneri, were inserted. Altogether, as of 15 November 2007, 96 ccrB internal sequences were made available, which were assigned to 17 alleles (Figure 1). In spite of the increased size of the collection and the extension to staphylococcal species other than S. aureus, the conclusions obtained for the well-defined MRSA collection are still valid. This is particularly relevant if one takes into account that 45 sequences were assigned to SCCmec type IV, the most variable structural type, and that there is a great diversity in the SCCmec elements circulating in the CoNS population. Among the 96 ccrB isolates, five were described for methicillin-susceptible strains (i.e. SCCmec negative). Although one sequence was assigned to a new cluster (ccrB allele 700), the remaining sequences were clustered in previously existing groups, suggesting that ccrB typing might also be useful for the characterization of other SCC elements. In conclusion, ccrB typing is indeed a promising SCCmec typing strategy since there is robust correlation among ccrB allelic clusters and SCCmec types.