A structural comparison of lipopolysaccharide biosynthesis loci of Legionella pneumophila serogroup 1 strains

Abstract

Background: The lipopolysaccharide (LPS) is the major immuno-dominant antigen of all Legionella species including L. pneumophila. Its diversity is the basis for the classification of L. pneumophila into serogroups and monoclonal subgroups and is thought to be involved in strain specific virulence. The understanding of the genetic basis of the LPS-antigen is incomplete. Thus, we analyzed the genetic locus involved in LPS-biosynthesis of L. pneumophila serogroup 1 (Sg1) strains with the focus on strain specific gene composition.

Results: The LPS-biosynthesis loci of 14 L. pneumophila Sg1 strains comprise two distinct regions: A 15 kb region containing LPS-biosynthesis genes that can be found in all L. pneumophila strains and a Sg1-specific 18 kb region. The 15 kb region is highly conserved among Sg1 strains as reflected by high homologies of single ORFs and by a consistent ORF arrangement. In contrast, the Sg1 specific 18 kb region is variable and partially disrupted by phage related genes. We propose that the region spanning from ORF 6 to ORF 11 of the Sg1-specific region is likely involved in late LPS-modification. Due to the high variability of this small region and various combinations of single ORFs within this region a strain specific LPS-structure could be synthesized including modifications of legionaminic acid derivates.

Conclusions: Our data clearly demonstrate that the gene structure of the LPS-biosynthesis locus of L. pneumophila Sg1 strains show significant interstrain variability. These data can be used for further functional analysis of the LPS synthesis to understand pathogenesis and reactivity with monoclonal antibodies. Moreover, variable but strain specific regions can serve as basis for the development of novel genotyping assays.

Figure 1

Structural representation of the LPS-biosynthesis locus. Shown are the LPS-biosynthesis loci of 14 L. pneumophila Sg1 strains and the corresponding monoclonal subgroup (in brackets). Strains Alcoy 2300/99, Corby and L10/23, and Paris and Philadelphia 1, respectively had the same genetic structure and monoclonal subtype and were therefore shown in one scheme. The numbering of ORFs was adopted by [21]. A: shows the Sg1-specific 18 kb region (ORFs 1-13) and B: shows the 15 kb region (ORFs 14-28). The direction of transcription is indicated by arrowheads. The filled black arrows indicate transposases/phage-related proteins. Grey shades and hatched patters serve to distinguish ORFs. Asterisk in Uppsala 3, Philadelphia 1 and Paris represents a partial ORF 2 duplication (ORF 2 like) as described by [46]. Underlined ORFs 7–11 in strain 130b represent an inversion. Görlitz 6543 carries a truncated lag-1 marked with ^†.

Figure 2

Dendrogram of variable ORFs. Multiple amino acid based cluster analysis using UPGMA (BioNumerics, Applied Maths NV, Belgium). The phylogenetic trees of gene lag-1 and of the ORFs 6, 7 and 8 are shown. ORF 9 is identical to the phylogenetic tree of ORF 8 and is therefore not shown. Similarity values and branch distances were depicted in percentages [%]. The strain-specific mAb-subgroup is indicated in brackets. The mutated start codon of lag-1 of Görlitz 6543 was neglected for similarity analysis and is indicated with ^†.