Phylogenetic reconstruction of the Legionella pneumophila Philadelphia-1 laboratory strains through comparative genomics

Abstract

Over 20 years ago, two groups independently domesticated Legionella pneumophila from a clinical isolate of bacteria collected during the first recognized outbreak of Legionnaires' disease (at the 1976 American Legion's convention in Philadelphia). These two laboratory strains, JR32 and Lp01, along with their derivatives, have been disseminated to a number of laboratories around the world and form the cornerstone of much of the research conducted on this important pathogen to date. Nevertheless, no exhaustive examination of the genetic distance between these strains and their clinical progenitor has been performed thus far. Such information is of paramount importance for making sense of several phenotypic differences observed between these strains. As environmental replication of L. pneumophila is thought to exclusively occur within natural protozoan hosts, retrospective analysis of the domestication and axenic culture of the Philadelphia-1 progenitor strain by two independent groups also provides an excellent opportunity to uncover evidence of adaptation to the laboratory environment. To reconstruct the phylogenetic relationships between the common laboratory strains of L. pneumophila Philadelphia-1 and their clinical ancestor, we performed whole-genome Illumina resequencing of the two founders of each laboratory lineage: JR32 and Lp01. As expected from earlier, targeted studies, Lp01 and JR32 contain large deletions in the lvh and tra regions, respectively. By sequencing additional strains derived from Lp01 (Lp02 and Lp03), we retraced the phylogeny of these strains relative to their reported ancestor, thereby reconstructing the evolutionary dynamics of each laboratory lineage from genomic data.

Figure 1. Legionella pneumophila Philadelphia-1 laboratory strains.

(A) Presumed phylogeny of the laboratory strains of L. pneumophila Philadelphia-1 based on the parallel experimental methods used to generate each strain. Both the Lp01 and JR32 lineages were derived from a clinical isolate, L. pneumophila Philadelphia-1, collected during the first recognized outbreak of Legionnaires' disease at a 1976 convention of the American Legion in Philadelphia, Pennsylvania. Lp01 was subsequently used to derive a spontaneous thymidine auxotroph Lp02. An avirulent dotA mutant, Lp03, reportedly derived from Lp02, is commonly used in studies as a translocation-deficient, intracellular replication defective control. (B) Out of a total of 700 publications, an estimate of the relative number which reference the Lp01/Lp02 lineage, JR32, or both. (See Materials and Methods).

Figure 2. Short, GC-rich direct repeats flank each of the large genomic regions lost in the Lp01 and JR32 lineages.

(A) The 45.4 kb lvh region that is deleted in the Lp01/Lp02/Lp03 lineage is flanked by a set of 69% GC, 49 nt direct repeats that likely facilitated intrachromosomal recombination and loss of the lvh region, which includes a type IVA secretion system as well as a restriction endonuclease, lpg1237, and its cognate methytransferase, lpg1236. (B) The 64.2 kb tra region is flanked by a set of 67% GC, 43 nt direct repeats that share no sequence homology with the lvh repeat. The genomic sequence of JR32 is consistent with an intrachromosomal recombination event between these repeats and subsequent loss of the intervening tra region.

Figure 3. Phylogeny of the Legionella pneumophila Philadelphia-1 laboratory strains as determined by whole-genome Illumina sequencing.

Illumina sequencing was performed on the Philadelphia-1 progenitor and each of its laboratory derivatives, JR32, Lp01, Lp02, and Lp03. Each genome was aligned to the published L. pneumophila Philadelphia-1 sequence (GenBank accession AE017354.1). The identification of single nucleotide polymorphisms (SNPs) and insertions and deletions (INDELs) in each laboratory strain was used to determine genetic distances with maximum likelihood between each strain. To explain these relationships, intermediate ancestral strains were proposed, as represented by open circles. A previously sequenced Lp01 derivative (Lp01^JK) containing luxN and ftsE mutations, but not lpg0716 and lpg0718 mutations indicates that luxN and ftsE emerged first within the Lp01 lineage . Not displayed: a putative gene conversion event that emerged independently in only the Lp01^CR and JR32 strains (AGAT 608,162–608,165 GAAG). (For details of each exact isolate sequenced, see materials and methods: JR32^AE, Lp01^CR, Lp01^JK, Lp02^AE, Lp03^AE).