Role of conjugative elements in the evolution of the multidrug-resistant pandemic clone Streptococcus pneumoniaeSpain23F ST81

Abstract

Streptococcus pneumoniae is a human commensal and pathogen able to cause a variety of diseases that annually result in over a million deaths worldwide. The S. pneumoniae(Spain23F) sequence type 81 lineage was among the first recognized pandemic clones and was responsible for almost 40% of penicillin-resistant pneumococcal infections in the United States in the late 1990s. Analysis of the chromosome sequence of a representative strain, and comparison with other available genomes, indicates roles for integrative and conjugative elements in the evolution of pneumococci and, more particularly, the emergence of the multidrug-resistant Spain 23F ST81 lineage. A number of recently acquired loci within the chromosome appear to encode proteins involved in the production of, or immunity to, antimicrobial compounds, which may contribute to the proficiency of this strain at nasopharyngeal colonization. However, further sequencing of other pandemic clones will be required to establish whether there are any general attributes shared by these strains that are responsible for their international success.

FIG. 1.

Circular diagram representing the S. pneumoniae ATCC 700669 chromosome, arranged to have the origin of replication at the top, as indicated by the GC deviation (innermost graph). The outer rings show the arrangement of coding sequences on the two strands of the genome, colored according to annotated function (see the key). The first inner ring indicates the major variable regions as blue blocks: moving clockwise from the origin of replication, these represent the prophage remnant, cps locus, PPI-1, the Na⁺-dependent ATP synthase island, ICESp23FST81, the pclA gene cluster, φMM1-2008, and the psrP gene cluster. The red blocks demarcate loci identified as having atypical nucleotide composition by the Alien Hunter algorithm (71). Moving inward, the rings show the position of IS elements (pink if intact, brown if pseudogene), RUP repeats (blue), and BOX A, B, and C repeat modules (red), respectively. The black graph indicates sequence GC content.

FIG. 2.

(A) Representation of ICESp23FST81. “Cargo” genes are colored according to the scheme detailed in Fig. 1 and are labeled with their putative function, where one can be assigned. The division of the ICE into Tn5252 and Tn916-type elements is indicated by the bars at the top and bottom of the diagram. (B) Comparison of the ICE insertion sites in the ATCC 700669 and G54 genomes with the corresponding loci in the TIGR4 and D39 genomes, which lack intact ICE. The ICE are represented as pink blocks: Tn5252-type elements are represented by blocks above the scale line, while Tn916-type elements are represented by blocks below the scale line. Red bands indicate BLAST matches between genomes in the same direction, whereas blue bands indicate matches in opposite directions. The intensity of the band represents the strength of the match. The region of each genome shown is bounded at the 5′ end by ICE-derived sequences and at the 3′ end by rplL, which streptococcal ICEs frequently insert directly upstream of. The ICE remnants in the TIGR4 and D39 genomes, apparently derived from the distal region of the Tn5252-type element, are marked. The alignment shows that there is a remnant in D39 directly adjacent to rplL, at the point where the elements usually insert, but the two remnants in the TIGR4 genome are further removed upstream.

FIG. 3.

Comparison of streptococcal integrative and conjugative elements. Genes likely to be part of the conjugative machinery of the element, on the basis of conservation or functional assignment, are colored pink. Zeta toxin-epsilon antitoxin systems are colored gray, antibiotic resistance genes are white, and other “cargo” genes are green. Bands are colored as in Fig. 2, with the intensity of the color indicating the strength of the BLAST match. Unlike the S. suis and S. pneumoniae elements, the S. agalactiae and S. dysgalactiae Tn5252-like elements do not contain a Tn916-type component, which carries the tetM gene responsible for tetracycline resistance in the other ICE.

FIG. 4.

Comparison of PPI-1 of ATCC 700669 with ICESp23FST81 found in the same strain. Three regions of apparently conserved similarity are observed: the zeta toxin-epsilon antitoxin system, a group of Tn5252 conjugative transfer genes (encoding ORF 9, ORF10, and relaxase proteins), and a short stretch of DNA adjacent that forms the 3′ border of PPI-1, shortly upstream of the cell division gene ftsW in the pneumococcal chromosome. The GC content of this region is shown, with the line across the graph indicating the average for the region (33.71% GC). The vertical dotted lines on the graph delimit the extent of PPI-1.

FIG. 5.

Alignment of PPI-1 from complete and draft pneumococcal genome data. Variation in the regions intervening between the relaxase and 3′ end of PPI-1 (these boundaries both have sequence homology with Tn5252-type ICE) appears to be due to horizontal gene transfer, indicating that conjugative elements may contribute to the diversity found within this island through homologous recombination-mediated exchange. There is also variation putatively resulting from degeneration of the original conjugative element insertion: the PPI-1 of S. pneumoniae G54 has lost the pezAT toxin-antitoxin system and much of the cluster of conjugative transfer genes (although it has retained a CDS encoding a MobA-domain protein, which are typically associated with autonomously mobile elements, indicated in pink), while that of 18-BS74 appears to have lost all vestiges of the original element's conjugative machinery.

FIG. 6.

Lantibiotic synthesis gene clusters and predicted structures. The color scheme is the same as that used in Fig. 1. (A) Gene cluster encoded on the PPI-1 of 23-BS72. Each structural gene appears to be associated with its own synthetase. The ABC transporter is thought to be responsible for the self-immunity of the producer. The likely sequences of the mature peptides are shown aligned with other dimeric lantibiotics, with the lanthionine rings (formed by the dehydration of cysteine and serine or threonine residues) conserved between haloduracin and lacticin 3147 indicated by the curved lines above the alignment. The conservation of the functionally most important residues with these other bacteriocins indicates the 23-BS72 lantibiotic is likely to be bactericidal. (B) Alignment of the PPI-1 of S. pneumoniae 23-BS72 and ATCC 700669, revealing a ∼6-kb deletion in the gene cluster of the latter. (C) Gene cluster encoded on ICESp23FST81. This is predicted to produce a monomeric lantibiotic, most similar to lichenicidin and mersacidin, produced by Bacillus species. The known ring structure of mersacidin is indicated, again showing that the functional, ring-forming residues are largely conserved in this new lantibiotic. It is the third ring (the most strongly conserved between the three sequences) that is thought to be most important for the bactericidal effects of mersacidin, resulting from the inhibition of peptidoglycan transglycosylation.