Table 1.
Description of cases from whom Legionella was isolated during the 1976 Legionnaires' disease outbreak in Philadelphia.
Fig 1.
Genomic characteristics and core-SNP-based phylogenetic analyses of all L. pneumophila strains sequenced in the present study.
(A) Genomic characteristics of all sequenced strains are shown as orange bars, green bars, and red stacked bars representing genome size, core genes (2828 genes), and accessory genes outside of the core, respectively. (B) Maximum-likelihood tree based on 11,356 core SNPs identified in all genomes. The Philadelphia clade is outlined and also expanded, therefore branches are not to scale in the inset. Blue and green shaded boxes highlight the confirmed (-O), and potential (-P) outbreak isolate pairs, respectively. Units of branch length (“Tree Scale”) are in nucleotide substitutions per site.
Table 2.
Pairwise whole-genome SNP comparisons of Philadelphia clade L. pneumophila isolates.
Fig 2.
BRIG analysis plot comparing nucleotide content of all genomes analyzed in the current study.
As defined in the legend, clinical isolates are shown in red while environmental isolates are blue. From the innermost ring: first black ring, L. pneumophila str. Phildadelphia-4 used as reference; second black ring, G+C content; third multicolored ring, GC skew; black bars, loci identified on the outer black bars and labeled; light blue ring 1 through 11 represent E1-P, E2-N, E3-N, E4-N, E6-N, E7-O, E8-O, E9-O, E10-P, and E11-U, respectively; red ring 1 through 11 represent C1-S, C2-S, C3-O, C4-S, C5-P, C6-S, C7-O, C8-S, C9-S, C10-S, and C11-O, respectively; outer purple ring L. pneumophila str. Paris; the outer black ring/bars highlight regions of interest, e.g., L-1 is locus 1, L-2 is locus 2, VGR-2 is variable genomic region 2, etc.
Fig 3.
The pP36-Ph mobilizable genetic element.
Nucleotide and structural comparison of the chromosomally integrated element within L. pneumophila strains Philadelphia and Paris, and the potential episomal form are shown. Gene prediction for pP36-Ph is based on the current strain Paris annotations found at NCBI. The inner blue circle of the episome represents G+C content. Horizontal, solid grey strips represent identical, conserved sequence, while thin vertical black lines or solid black regions represent nucleotide differences (SNPs) between genomes. Thin horizontal black lines between solid grey regions represent gaps or deletions in the sequence. The nucleotide boundaries where pP36-Ph is integrated in the chromosome are shown above the sequence (relative to strain Philadelphia-1 and Paris), along with neighboring genes. A double jagged line represents additional internal sequence not shown.
Fig 4.
Mauve whole-genome alignment of L. pneumophila strains within the Philadelphia clade.
ProgressiveMauve was used to compare the fully assembled sequences of the Philadelphia historical Legionella strains as well as isolate E1-P. The minimum weight for pairwise LCBs (locally collinear blocks), which share common colors across genomes, was set to 100, otherwise, the program was run using default parameters as described in the Methods. The general clade organization, as well as the identity and location of the ~40-kb pP36-Ph and the ~45-kb pLP45 elements are shown. The general, expanded Philadelphia clade organization from Fig 1 is shown, therefore the phylogenetic distances are not to scale.
Table 3.
Nucleotide polymorphisms shared by all CDC Philadelphia strains relative to the NCBI Philadelphia-1 reference sequence.
Fig 5.
Selected genetic differences between the NCBI strain Philadelphia-1 reference sequence and all historical CDC Philadelphia isolates.
Solid grey strips, and vertical and horizontal black lines represent conserved sequence, nucleotide SNPs, and sequence gaps, respectively, as described in Fig 3. Nucleotide boundaries for the potential rtxA and mompS deletions of ~10,203 and ~1,032 bp, respectively, are given above the sequence representations relative to the NCBI strain Philadelphia-1 reference sequence. The blue arrow in the top pane represents the full length rtxA gene found in strains Philadelphia-1, -2, -3, and -4. The blue arrow in the bottom pane represents a tandem mompS paralog not identified in the NCBI strain Philadelphia-1 reference sequence.
Table 4.
Nucleotide polymorphisms shared by strains CDC Philadelphia-2, -3, -4, and ATCC Philadelphia-1 relative to the NCBI Philadelphia-1 reference sequence.
Fig 6.
Variable genomic region 1 and 2 (VGR-1, and -2) within L. pneumophila Philadelphia strains and additional ST36 genomes.
(A) VGR-1A is conserved in strain Philadelphia-1 (CDC) and in 19 of 22 additional ST36 strains, including strains C2-S, C3-O, C4-S, C5-P, C6-S, C7-O, C8-S, C9-S, C10-S, C11-O, E3-N, E4-N, E5-N, E6-N, E7-O, E8-O, E9-O, E10-P, E11-U, as well as L. pneumophila str. LPE509 and a sg12 strain (ATCC 43290), while VGR-1B is conserved in strains Philadelphia-2, -3, -4, E1-P, C1-S, and Paris. (B) VGR-2A is conserved in strain Philadelphia-1 (CDC) and in 21 of 22 additional ST36 strains, including C1-S, C2-S, C3-O, C4-S, C5-P, C6-S, C7-O, C8-S, C9-S, C10-S, C11-O, E2-N, E3-N, E4-N, E5-N, E6-N, E7-O, E8-O, E9-O, E10-P, and E11-U, while VGR-2B is conserved in strains Philadelphia-2, -3, -4, E1-P, and Paris. Solid grey strips, vertical, and horizontal black lines represent conserved sequence, SNPs, and sequence gaps, respectively, as in Fig 3. A Solid blue or green rectangle above the sequence delineates the boundaries of the VGR. Nucleotide boundaries and neighboring genes (in red) are relative to the genome immediately below these descriptions.
Table 5.
Nucleotide polymorphisms unique to the CDC Philadelphia strains relative to the NCBI Philadelphia-1 reference sequence.
Fig 7.
Gubbins-based recombinational analysis and phylogenetic tree reconstruction.
Regions of elevated SNP density, representing potential horizontal gene transfer events, were identified by the Gubbins algorithm and software package, masked in the original multiple sequence alignment, and then core SNPs were identified with kSNP v3, as detailed in the Methods. The maximum-likelihood tree shown was constructed using RAxML v8 and 307 core, non-recombinant SNPs with 1000 bootstrappings. Red blocks represent regions of elevated SNP density conserved in multiple strains and blue blocks represent elevated SNP density only found in a single strain. L. pneumophila str. Philadelphia-4 was used as the reference/outgroup, and VGR-1 and -2 are labeled. Gubbins was run with default parameters. A Gubbins-generated tree with 480 non-recombining SNPs also exhibited similar topology.