Synergistic contribution of the Legionella pneumophila lqs genes to pathogen-host interactions

Abstract

The causative agent of Legionnaires' disease, Legionella pneumophila, is a natural parasite of environmental protozoa and employs a biphasic life style to switch between a replicative and a transmissive (virulent) phase. L. pneumophila harbors the lqs (Legionella quorum sensing) cluster, which includes genes encoding the autoinducer synthase LqsA, the sensor kinase LqsS, the response regulator LqsR, and a homologue of HdeD, which is involved in acid resistance in Escherichia coli. LqsR promotes host-cell interactions as an element of the stationary-phase virulence regulatory network. Here, we characterize L. pneumophila mutant strains lacking all four genes of the lqs cluster or only the hdeD gene. While an hdeD mutant strain did not have overt physiological or virulence phenotypes, an lqs mutant showed an aberrant morphology in stationary growth phase and was defective for intracellular growth, efficient phagocytosis, and cytotoxicity against host cells. Cytotoxicity was restored upon reintroduction of the lqs genes into the chromosome of an lqs mutant strain. The deletion of the lqs cluster caused more-severe phenotypes than deletion of only lqsR, suggesting a synergistic effect of the other lqs genes. A transcriptome analysis indicated that in the stationary phase more than 380 genes were differentially regulated in the lqs mutant and wild-type L. pneumophila. Genes involved in protein production, metabolism, and bioenergetics were upregulated in the lqs mutant, whereas genes encoding virulence factors, such as effectors secreted by the Icm/Dot type IV secretion system, were downregulated. A proteome analysis revealed that a set of Icm/Dot substrates is not produced in the absence of the lqs gene cluster, which confirms the findings from DNA microarray assays and mirrors the virulence phenotype of the lqs mutant strain.

FIG. 1.

The L. pneumophila lqs cluster and model of the Lqs quorum sensing circuit. (A) The lqs (Legionella quorum sensing) cluster consists of four ORFs, designated lqsA (lpg2731), lqsR (lpg2732), hdeD (lpg2733), and lqsS (lpg2734). The 5′ flanking region contains two ORFs, lpg2730 (dsbB) and lpg2731 (cycB), encoding electron transfer proteins. An operon including ORFs lpg2735 to lpg2738 (hemC, hemD, hemX, and hemY), involved in heme biosynthesis, is located in the 3′ flanking region. (B) Model of the Lqs quorum sensing circuit. (C) The two AflIII restriction sites in the lqs locus delineate the cloned genomic lqs fragment (6,790 bp), and the BsgI and BbvCI restriction sites were used to replace a 4,482-bp lqs fragment with a Km resistance cassette to construct the suicide vector for allelic exchange. Arrows indicate the oligonucleotide binding sites used for PCR screening of the genomic regions of the wild type (5,118 bp), the lqs deletion mutant (1,907 bp), and three Δlqs chromosomal complementation strains (CC-1, CC-9, and CC-17) (D).

FIG. 2.

Effects of the L. pneumophila lqs genes on growth in broth and morphology. (A) L. pneumophila wild-type strain JR32 (•, ○) or lqs mutant strains (▴, ▵) were inoculated at an OD₆₀₀ of 0.1, and the growth of the strains in AYE medium was followed for 21 h at 37°C (gray line) or 28 h at 30°C (black line). Similar results were obtained in two or three independent experiments. (B) Cell morphology of L. pneumophila wild-type strain JR32, the lqs mutant, and an lqsR mutant strain grown for 8 h (exponential phase), 21 h (early stationary phase), and 30 h (late stationary phase) in AYE medium at 37°C. Representative transmission electron micrographs (upper row) or confocal microscopy pictures (lower three rows) of dividing cells (8 h); single coccoid, rod-shaped bacteria (21 h and 30 h; wild type and ΔlqsR); or filamentous bacteria (21 h and 30 h; Δlqs) are shown.

FIG. 3.

L. pneumophila lacking the lqs cluster is less cytotoxic for amoebae and macrophages. (A) Cytotoxicity of L. pneumophila against A. castellanii was assayed by flow cytometry at 24 h postinfection (MOI = 100) using the wild-type strain JR32, an lqs mutant strain, a chromosomally complemented lqs mutant (CC9), or an icmT mutant strain. Live versus dead A. castellanii bacteria were scored by a population shift toward smaller, more granular cells (forward scatter [FSC] versus sideward scatter [SSC]; upper) and quantified by PI staining (PI versus counts; lower). The data shown are representative of at least two independent experiments. (B) The viability of A. castellanii (left) or murine RAW264.7 (right) macrophages infected with L. pneumophila wild-type strain JR32 or the lqs or icmT mutant strain was assayed by an Alamar Blue dye reduction assay at 19 h postinfection (MOIs of 100, 10, and 1). The data shown are representative of at least three independent experiments. Statistically significant (P < 0.05; unpaired Student t test) differences are marked by asterisks (*, wild type versus mutants; **, Δlqs versus ΔicmT).

FIG. 4.

L. pneumophila Δlqs is impaired for intracellular growth in A. castellanii and efficient phagocytosis. Intracellular replication of GFP-labeled L. pneumophila strains within A. castellanii over the period of 5 days was assayed by flow cytometry by gating on the amoebae (A) or on the bacteria (B) (black bars, wild-type JR32; dark gray bars, ΔlqsR; light gray bars, Δlqs; white bars, ΔicmT) released from infected amoebae (means and standard deviations for triplicates). (C) Intracellular replication of L. pneumophila strains was determined by counting CFU after infecting A. castellanii at an MOI of 1 with wild-type strain JR32 (black triangles), an hdeD mutant (gray triangles), an lqsR mutant (gray circles), an lqs mutant (white circles), or an icmT mutant (black squares). At the time points indicated, appropriate dilutions of supernatants of the infected amoebae were plated. (D) Phagocytosis of GFP-labeled L. pneumophila strains (wild-type JR32, ΔhdeD, ΔlqsR, Δlqs, or ΔicmT) by A. castellanii infected at an MOI of 50 for 1 h was assayed by flow cytometry. Similar results were obtained in two (A, B) and at least three (C, D) independent experiments.

FIG. 5.

Effects of the lqs genes on pH and salt sensitivity of L. pneumophila in AYE medium. L. pneumophila wild-type strain JR32 or the hdeD, lqsR, or lqs mutant strain grown for 21 h in AYE medium was spotted in triplicates at the dilutions indicated onto CYE agar plates titrated to the pH values indicated (A) or spotted onto CYE agar plates containing 100 mM NaCl or no additional salt (control) (B). One representative dilution series is shown for each strain, and similar results were obtained in three independent experiments.

FIG. 6.

The lqs genes regulate lqsR promoter activity. lqsR promoter activity was quantified by β-galactosidase activity (bars) at the OD₆₀₀s indicated (symbols) in wild-type L. pneumophila (black bars, •), an lqs mutant (white bars, ▴), or an lqsR mutant (gray bars, ▪) harboring plasmid pTS-14, which contains the lacZ gene under the control of the lqsR promoter. Similar results were obtained in at least two independent experiments.

FIG. 7.

Classification and graphic representation of the lqs mutant strain transcriptome. DNA microarray analysis of the stationary-phase L. pneumophila lqs mutant or the wild-type strain JR32 revealed 190 or 196 genes that were induced or repressed, respectively, at least twofold. The genes induced (A) or repressed (B) in the lqs mutant strain compared to the levels in wild-type L. pneumophila were classified into functional groups and further divided into subgroups (y axis) based on their annotation or prediction by use of the InterPro database. The percentages of genes regulated are indicated with a bar graph (x axis) and represented as a pie graph (inset). The analysis is based on Table S1 in the supplemental material.

FIG. 8.

Comparative 2-DE analysis of soluble intracellular proteins extracted from wild-type L. pneumophila or an lqs mutant strain grown in AYE to stationary phase at 37°C. (A) Proteins were separated in the first dimension by isoelectric focusing on IPG strips with nonlinear pH gradients from 3 to 10, followed by SDS PAGE on 13% polyacrylamide gels in the second dimension. The gels were stained with Coomassie brilliant blue G-250. (B) The resulting spot patterns were analyzed with the Proteomeweaver software program. About 30 spots with increased intensities in either wild-type L. pneumophila (red circles) or the lqs mutant strain (blue circles) were picked, processed, and analyzed using MALDI-TOF MS-based sequencing. The protein spots identified are listed in Table S2 in the supplemental material. Asterisks indicate proteins regulated with the same pattern as in DNA microarray experiments. The gels shown are representatives of two independent experiments.