Genome sequence of a Lancefield group C Streptococcus zooepidemicus strain causing epidemic nephritis: new information about an old disease

Abstract

Outbreaks of disease attributable to human error or natural causes can provide unique opportunities to gain new information about host-pathogen interactions and new leads for pathogenesis research. Poststreptococcal glomerulonephritis (PSGN), a sequela of infection with pathogenic streptococci, is a common cause of preventable kidney disease worldwide. Although PSGN usually occurs after infection with group A streptococci, organisms of Lancefield group C and G also can be responsible. Despite decades of study, the molecular pathogenesis of PSGN is poorly understood. As a first step toward gaining new information about PSGN pathogenesis, we sequenced the genome of Streptococcus equi subsp. zooepidemicus strain MGCS10565, a group C organism that caused a very large and unusually severe epidemic of nephritis in Brazil. The genome is a circular chromosome of 2,024,171 bp. The genome shares extensive gene content, including many virulence factors, with genetically related group A streptococci, but unexpectedly lacks prophages. The genome contains many apparently foreign genes interspersed around the chromosome, consistent with the presence of a full array of genes required for natural competence. An inordinately large family of genes encodes secreted extracellular collagen-like proteins with multiple integrin-binding motifs. The absence of a gene related to speB rules out the long-held belief that streptococcal pyrogenic exotoxin B or antibodies reacting with it singularly cause PSGN. Many proteins previously implicated in GAS PSGN, such as streptokinase, are either highly divergent in strain MGCS10565 or are not more closely related between these species than to orthologs present in other streptococci that do not commonly cause PSGN. Our analysis provides a comparative genomics framework for renewed appraisal of molecular events underlying APSGN pathogenesis.

Figure 1. Genome atlas.

Data from outermost-to-innermost circles are in the following order. Genome size in megabase pairs (circle 1). Annotated CDSs encoded on the forward (circle 2) and reverse (circle 3) chromosomal strands are in dark and light blue respectively. Reference landmarks (circle 4) as labeled are, ribsomal RNAs in green, fimbrial operons in orange, hyaluronic acid capsule synthesis loci in gold, CRISPR/CAS phage immunity loci in light blue, streptolysin S (sag) operon in purple, and ISs/transposons in red. CDS percent G+C content (circle 5) with greater and lesser than average in red and blue, respectively. Net divergence of CDS dinucleotide composition (circle 6) from the average is in orange. TBLASTN comparison of gene content with nephritogenic GAS serotype M12 strain MGAS2096 (circle 7), with high similarity in red and low in and blue. TBLASTN comparison of gene content with other sequenced streptococcal species (circle 8), with high similarity in red and low in blue. Species-specific gene content (circle 9), products not present in the other streptococcal species are in black and products sharing less than 50% amino acid identity with the most similar streptococcal homologue are in gray.

Figure 2. Streptococcal multilocus genetic relationships.

Inferred sequences of the DNA replication and repair proteins DnaA, DnaE, DnaG, DnaI, DnaJ, DnaN, and DnaX were concatenated, aligned and used to infer genetic relationships among the streptococcal strains for which complete genome sequences are available. The genes encoding these proteins are conserved genus-wide and arrayed around the chromosome.

Figure 3. Genome comparisons.

(A) Gene content comparison. The inferred proteomes of S. zooepidemicus (GCS), GAS, and S. agalactiae (GBS) were compared pair wise to the translated genomes of each other using TBLASTN (cutoff, e = 10⁻⁹). The numbers of genes given for each section are color coded to match the respective genomes. The numbers for CDS shared in common in the intersections differ slightly due to variance in gene copies species-to-species, such as resulting from gene duplication and mobile genetic element transfer events. (B) Aligned GCS and GAS genomes. The nucleotide sequence of the MGCS10565 (GCS) and MGAS2096 (GAS) genomes were compared and regions sharing at least 60% identity over a window of 30 nucleotides are illustrated. Conserved regions oriented in the same direction are in red, and regions opposite in direction are in blue.

Figure 4. DNases genetic relationships.

Inferred products of all of the DNase genes present in the genomes of the 12 sequenced GAS strains, chromosomally encoded (shown in blue) and prophage encoded (shown in black) were aligned with the those present in the S. zooepidemicus MGCS10565 genome (shown in red) and genetic relationships were inferred using the unweighted pair group method with arithmetic mean (UPGMA). Each of the S. zooepidemicus DNases is an outlier relative to the GAS DNases of the same type, arguing for an independent evolutionary path and against very recent horizontal transfer between the species.

Figure 5. Schematic of three fimbrial operons.

Illustrated are the 3 putative fimbrial operons identified in the MGCS10565 genome. Each operon has a gene encoding a protein homologous to the fimbrial backbone/major subunit protein (red) and an ancillary/minor subunit protein (orange). The genes encoding the fimbrial structural proteins are flanked at the 5′ end (but oriented in the opposite direction) by genes encoding regulatory proteins (yellow) and at the 3′ end by genes encoding sortases of the C-family (green). Gene numbers are given in black. Homologues of the structural proteins and their percent amino acid identity are given in blue. The putative major subunit proteins all have homology to GAS T-antigens, as expected.

Figure 6. Schematic of collagen-like proteins.

Illustrated are 12 inferred proteins with collagen structural motifs encoded by the MGCS10565 genome. These proteins are composed of the following domains (from amino- to carboxy-terminus): SSP, secretion signal peptide; V, variable region; CL, collagen-like region, W, a proline-rich putative cell wall spanning region, and finally a tripartite cell wall anchor. Numbers in black below the schematics are the last amino acid residue of each the respective regions. The number of contiguous Gly-Xxx-Yyy repeats composing the CL-regions are given in red. Sites within the CL-regions matching integrin recognition sequences RGD and KGD are indicated by red and black bars, respectively. Prokaryotic analogs (GxPGER) of human collagen sequences mediating high-affinity integrin binding, are indicated by violet bars. CLPs 6, 9, and 10 have V-regions with fibronectin-binding domains.

Figure 7. Genetic relationships of virulence factors implicated in PSGN pathogenesis.

(A) Streptokinase, (B) GAPDH, and (C) Enolase. Inferred products for each of the virulence factors implicated in PSGN pathogenesis were aligned and relationships were inferred using the UPGMA method. GCS S. equi subsp. equi and zooepidemicus strains are shown in red, GAS S. pyogenes in blue, GBS S. agalactiae in green, GCS/GGS S. dysgalactiae subsp. equisimilis in purple, and other streptococcal species in black. These virulence factors are not more closely related between S. zooepidemicus and S. pyogenes than among the other streptococcal species.

Figure 8. Schematic of the SpeB encoding region.

Illustrated is an alignment of the SpeB encoding region of the S. pyogenes neprhitogenic serotype M12 strain MGAS2096 genome (shown in blue) with the corresponding region of the S. zooepidemicus strain MGCS10565 genome (shown in red). speB and several flanking genes (e.g. smeZ, sof, and sfbX) are not present in the S. zooepidemicus strain MGCS10565 genome.