Comparative analysis of Klebsiella pneumoniae genomes identifies a phospholipase D family protein as a novel virulence factor

Abstract

Background: Klebsiella pneumoniae strains are pathogenic to animals and humans, in which they are both a frequent cause of nosocomial infections and a re-emerging cause of severe community-acquired infections. K. pneumoniae isolates of the capsular serotype K2 are among the most virulent. In order to identify novel putative virulence factors that may account for the severity of K2 infections, the genome sequence of the K2 reference strain Kp52.145 was determined and compared to two K1 and K2 strains of low virulence and to the reference strains MGH 78578 and NTUH-K2044.

Results: In addition to diverse functions related to host colonization and virulence encoded in genomic regions common to the four strains, four genomic islands specific for Kp52.145 were identified. These regions encoded genes for the synthesis of colibactin toxin, a putative cytotoxin outer membrane protein, secretion systems, nucleases and eukaryotic-like proteins. In addition, an insertion within a type VI secretion system locus included sel1 domain containing proteins and a phospholipase D family protein (PLD1). The pld1 mutant was avirulent in a pneumonia model in mouse. The pld1 mRNA was expressed in vivo and the pld1 gene was associated with K. pneumoniae isolates from severe infections. Analysis of lipid composition of a defective E. coli strain complemented with pld1 suggests an involvement of PLD1 in cardiolipin metabolism.

Conclusions: Determination of the complete genome of the K2 reference strain identified several genomic islands comprising putative elements of pathogenicity. The role of PLD1 in pathogenesis was demonstrated for the first time and suggests that lipid metabolism is a novel virulence mechanism of K. pneumoniae.

Figure 1

K. pneumoniae Kp52.145 genome. A) Chromosome representation: Outermost layers in gray indicate the position of positive and negative strand CDSs. tRNAs are represented in green, while the four virulence-related genomic islands (GI) are in red, the locus coding for anaerobic metabolism in orange, the T6SS gene clusters in blue and capsular synthesis region in yellow. Inner circle represents G + C%. Detail of the T6SS locus III region containing the putative phospholipase (pld1 gene) (orange) and sel1 (pink) genes is shown enlarged. B) Plasmids representation: plasmid maintenance genes and IS sequences are shown in blue, proteins with unknown functions in gray, known functions in green, toxin-antitoxin systems in yellow and rmpA in red. CDSs, coding sequences.

Figure 2

Genomic islands (GIs) identified in the genome of Kp52.145. Positive and negative strand CDSs are represented in gray. Putative virulence related genes are highlighted in different colors. The size of each GI is given in kb. Panel A: GI-I, B: GI-II, C: GI-III and D: GI-IV. CDS, coding sequences.

Figure 3

Sequence alignment of Kp52.145 (Kp) PLD1 and its closest related sequences: putative phospholipase D family protein from Pseudomonas syringae (PS) and cardiolipin synthases from P. putida (PP), Staphylococcus aureus (SA), Bacilus subtilis (BS) and E. coli (EC). The rectangles indicate the phospholipase D active site regions and the asterisk points to the corresponding transposon insertion site in pld1 mutant strain.

Figure 4

PLD1 involvement in K. pneumoniae virulence. Mice survival after infection with K. pneumoniae Kp52.145 wild-type, pld1 mutant and complemented strains. Data are representative of seven mice per group from two independent experiments. Standard deviation is shown. PLD, phospholipase D.

Figure 5

PLD1 involvement in lipids metabolism. A) TLC lipid profiles of E. coli strain SD9 and SD9 + pPLD. Red circle indicates phosphatidyl glycerol. Blue circle represents the lipid present specifically upon pld1 expression. B) Mass spectrometry profiles of E. coli strain SD9 and SD9 + pPLD lipids. The red arrow points to m/z 789.4 exclusively found in wild-type strain. TLC, thin-layer chromatography.