Transcription factors DksA and PsrA are synergistic contributors to Legionella pneumophila virulence in Acanthamoeba castellanii protozoa

Abstract

The environmental bacterium Legionella pneumophila, an intracellular parasite of free-living freshwater protozoa as well as an opportunistic human pathogen, has a biphasic lifestyle. The switch from the vegetative replicative form to the environmentally resilient transmissive phase form is governed by a complex stringent response-based regulatory network that includes RNA polymerase co-factor DksA. Here, we report that, through a dysfunctional DksA mutation (DksA1), a synergistic interplay was discovered between DksA and transcription regulator PsrA using the Acanthamoeba castellanii protozoan infection model. Surprisingly, in trans expression of PsrA partially rescued the growth defect of a dksA1 strain. Whilst in trans expression of DksA expectantly could fully rescue the growth defect of the dksA1 strain, it could also surprisingly rescue the growth defect of a ΔpsrA strain. Conversely, the severe intracellular growth defect of a ΔdksA strain could be rescued by in trans expression of DksA and DksA1, but not PsrA. In vitro phenotypic assays show that either DksA or DksA1 was required for extended culturability of bacterial cells, but normal cell morphology and pigmentation required DksA only. Comparative structural modelling predicts that the DksA1 mutation affects the coordination of Mg²⁺ into the active site of RNAP, compromising transcription efficiency. Taken together, we propose that PsrA transcriptionally assists DksA in the expression of select transmissive phase traits. Additionally, in vitro evidence suggests that the long-chain fatty acid metabolic response is mediated by PsrA together with DksA, inferring a novel regulatory link to the stringent response pathway.

Keywords: Legionella pneumophila; biphasic lifestyle; host-microbe interaction; regulatory network; stringent response; transcription factor.

Fig. 1.. Structural comparison of DksA and DksA1 AlphaFold models. (a) Structural overlay of DksA (PDB ID: 5VSW) with both Lp02 DksA and DksA1 as modelled by AlphaFold. (b) Residue conservation of Lp02 DksA as plotted by the server ConSurf [5556]. Yellow indicates insufficient information in the multisequence alignment. Indicated on the structure are conserved residues important for binding ppGpp and allostery with RNAP (R97 to K104), in addition to contact with the active site (d77–d80). (c) Binding of DksA with the RNAP allosteric site. A DksA:RNAP experimental structure (PDB ID: 5VSW) and aligned AlphaFold models are shown as cartoons, whereas RNAP is drawn as a molecular surface. The structure of ppGpp bound in the complex is also shown. (d) Slice view of DksA showing the contact point of residues D80 and D77 with magnesium in the RNAP active site. (d) Interactions of DksA variants with ppGpp in the RNAP allosteric site are shown and not disturbed by the truncation in DksA1. (e) Interactions of DksA variants with magnesium in the active site of RNAP are shown. (f) AlphaFold predicts that the truncation in DksA1 drastically moves the position of a critical aspartate residue (D69/71/77).

Fig. 2.. Synergistic impacts of DksA, DksA1 and PsrA on virulence of L. pneumophila in the A. castellanii protozoa. Intracellular growth kinetics of L. pneumophila strains in A. castellanii protozoa at 25 °C. Strains tested were parental Lp02 and avirulent ΔdotA along with isogenic strains featuring single or combinatorial genetic mutations (ΔpsrA, ΔdksA and dksA1) with or without in trans expression of psrA, dksA or dksA1. Error bars represent sd across the mean of three biological replicates, with two technical replicates each. Samples below the detection limit (~10 c.f.u. ml⁻¹) were recorded as zero and omitted from the graph if they lacked countable colonies across all replicates.

Fig. 3.. DksA is dispensable for growth in human macrophages. Intracellular growth kinetics of L. pneumophila strains in U937-derived macrophages at 37 °C. Strains tested were parental Lp02 and avirulent ΔdotA along with isogenic strains featuring single or combinatorial genetic mutations (ΔpsrA, ΔdksA and dksA1) with or without in trans expression of psrA, dksA or dksA1. Error bars represent sd across the mean of three biological experiments with two technical replicates each. Experiments in panels a–c were conducted concurrently but presented separately for clarity.

Fig. 4.. L. pneumophila in vitro growth at 37 °C is impacted by DksA, PsrA and DksA1. In vitro growth kinetics of parental Lp02 and isogenic strains featuring single or combinatorial genetic mutations (ΔpsrA, ΔdksA and dksA1) with and without in trans expression of psrA, dksA or dksA1 genes. Plate-grown bacteria were suspended in BYE broth and normalized to initial OD₆₀₀=0.15, and growth at 37 °C was monitored for OD₆₀₀ hourly for 24 h. Graphs in panels a–f were experiments conducted concurrently but presented separately for clarity. Error bars are sd across the mean of three independent biological replicates with three technical replicates each.

Fig. 5.. L. pneumophila in vitro growth at 25 °C is impacted by DksA, PsrA and DksA1. In vitro growth kinetics of parental Lp02 and isogenic strains featuring single or combinatorial genetic mutations (ΔpsrA, ΔdksA and dksA1) with and without in trans expression of psrA, dksA or dksA1 genes. Plate-grown bacteria were re-suspended in BYE broth and normalized to initial OD₆₀₀=0.15, and growth at 25 °C was monitored for OD₆₀₀ every 2 h for 48 h. Graphs in panels a–f were experiments conducted concurrently but presented separately for clarity. Error bars are sd across the mean of three independent biological replicates with three technical replicates each.

Fig. 6.. L. pneumophila DksA and DksA1, but not PsrA, affect colony morphology. Parental Lp02 and isogenic strains featuring single or combinatorial genetic mutations (ΔpsrA, ΔdksA and dksA1) with and without in trans expression of psrA, dksA or dksA1 genes. Strains were grown on BCYE plates for 3 days and used to generate a bacterial suspension in BYE at OD₆₀₀=0.15. Ten microlitres of the suspension were spotted onto BCYE plates in triplicate; then, the plates were incubated at 37 °C at 5% CO₂ for 4 days prior to being photographed. Images are representative of one of three biological replicates.

Fig. 7.. L. pneumophila ΔdksA, but not dksA1 or psrA, causes filamentation. Parental Lp02 and isogenic strains featuring single or combinatorial genetic mutations (ΔpsrA, ΔdksA and dksA1) with or without in trans expression of psrA, dksA or dksA1 genes were grown on BCYE plates and used to inoculate a dilute suspension of OD₆₀₀=0.02 in BYE broth, then incubated overnight at 37 °C with aeration, to exponential (E) (~18 h, OD₆₀₀ =~0.5) or stationary (S; 24 h after EP) growth phases, and then live-mounted on a 2% agarose pad prepared in ddH₂O and imaged on an Axio Observer Z1 inverted microscope (Zeiss) equipped with a glycerol-immersion 150X objective. Representative microscopic images of bacteria from one of three biological replicates. The scale bar is 5 µm.

Fig. 8.. L. pneumophila ΔdksA, dksA1 and psrA contribute to cell length regulation. Parental Lp02 and isogenic strains featuring single or combinatorial genetic mutations (ΔpsrA, ΔdksA and dksA1) with or without in trans expression of psrA, dksA or dksA1 genes were grown from 3-day-old plates subcultured to BYE broth at OD₆₀₀=0.02, then incubated at 37 °C with aeration to (a) exponential growth phase (overnight, OD₆₀₀ ~0.5) or (b) stationary growth phase (24 h after exponential phase) and then imaged with an Axio Observer Z1 inverted microscope (Zeiss) equipped with a glycerol-immersion 150X objective, and individual bacterial cell lengths were quantified using the length tool in ImageJ. Where filamentation was evident, only individual cells for which the full length was visible were measured. Data represent >150 individual readings spread equally across three biological replicates. The bar on the graph and error bars represent the mean and 95% CI, respectively. Statistical significance determined by Student’s t-test (*P<0.05; **P<0.01; ***P<0.001; ****P<0.0001; ns, not significant, P>0.05).

Fig. 9.. Bacterial culturability in depleted nutrient media is dependent on DksA or DksA1, but not PsrA. Panels (a–f) show in vitro growth kinetics of Lp02 and isogenic strains featuring single or combinatorial genetic mutations (ΔpsrA, ΔdksA and dksA1) with or without in trans expression of psrA, dksA or dksA1 genes in BYE broth at 37 °C for 96 h with OD₆₀₀ values (dotted lines; left Y-axis) and titre enumerated by serial dilution and incubation on BCYE plates (solid lines; right Y-axis) at 24 h intervals. Data are presented as mean and sd of three biological replicates. Points were omitted if results could not be calculated (no colonies recovered, below the detection limit of ~100 c.f.u. ml⁻¹). Graphs in panels a–f were experiments conducted concurrently but presented separately for clarity.

Fig. 10.. Pigmentation is regulated by DksA, but not PsrA. Parental Lp02 and isogenic strains featuring single or combinatorial genetic mutations (ΔpsrA, ΔdksA and dksA1) with or without in trans expression of psrA, dksA or dksA1 genes were grown on BCYE plates for 3 days and used to generate a 5 ml bacterial suspension (BYE at OD₆₀₀=0.15) and then incubated at 37 °C for 30 h. (a) Qualitative observation of pigmentation. After 48 h of incubation, tubes were set aside for 24 h at room temperature to allow cells to sediment out and then photographed. Representative image of three replicates. (b) Quantification of pigment at 30 h. After 30 h of incubation, 1 ml of culture was centrifuged at 20,000 g. The supernatant was measured for pigment levels via absorbance at 550 nm, and then, the pellet was resuspended in the same volume of 1X PBS pH 7.0 to measure cell density via OD₆₀₀. Readings were taken with a plate reader in a 96-well plate relative to appropriate BYE or PBS blanks. Readings reflect Abs₅₅₀ /OD₆₀₀. Error bars represent sd across three biological replicates. Statistical significance calculated via Student’s t-test and relative to parental Lp02 pThy is indicated above bars (*P<0.05; **P<0.01; ***P<0.001; ****P<0,0001; ns, not significant, P>0.05).

Fig. 11.. PsrA and DksA influence LCFA-mediated growth inhibition. Parental Lp02 and isogenic strains featuring single or combinatorial genetic mutations (ΔpsrA and/or ΔdksA), with or without in trans expression of psrA or dksA genes, grown with and without PA supplementation (as 40 mM stock dissolved in ethanol) in BYE media normalized to 0.5% ethanol, in a 96-well microplate for 24 h with shaking at 37 °C. Inhibition was defined as growth, as assessed by OD₆₀₀ values, of <50% of that of the same strain in a BYE+0.5% ethanol-only control at the end of 24 h. Bars indicate the mean of the maximum concentration of PA where the indicated strain grew >50% of maximum, over three biological replicates with sd indicated. Dots in bars indicate individual replicate values. Panels a and b were conducted separately. Statistical analyses were done using Student’s unpaired t-test: ns, no significance; *=P<0.05.

Fig. 12.. Model of independent mechanisms of DksA and PsrA. A hypothetical promoter is repressed by both PsrA and DksA. (a) RNAP holoenzyme is targeted to −10 and −35 boxes by a sigma factor (dark grey RNAP component). PsrA (oval dimers) binds a defined site in the promoter and inhibits RNAP recruitment. (b) In a ΔdksA strain, only GreA (triangles) is available for RNAP, thus elevating transcription levels, but in trans psrA overexpression generates higher levels of PsrA, strengthening inhibition of RNAP recruitment to compensate. Conversely, (c) the ΔpsrA strain allows uninhibited RNAP-promoter binding, increasing transcription, but this is offset by overexpressing dksA in trans, increasing the proportion of RNAP bound to transcription-inhibiting DksA (octagons), thereby generating the compensatory effect. (d) In an alternative scenario in which a promoter is not subject to PsrA influence, RNAP binds to GreA or DksA interchangeably, with enhanced transcription initiation rates occurring whilst bound to GreA versus DksA.