Targeting of the small GTPase Rab6A' by the Legionella pneumophila effector LidA

Abstract

When the bacterium Legionella pneumophila, the causative agent of Legionnaires' disease, is phagocytosed by alveolar macrophages, it delivers a large number of effector proteins through its Dot/Icm type IV secretion system into the host cell cytosol. Among those proteins is LidA, an effector that interacts with several host GTPases of the Rab family, including Rab6A', a regulator of retrograde vesicle trafficking within eukaryotic cells. The effect of LidA on Rab6A' function and the role of Rab6A' for L. pneumophila growth within host cells has been unclear. Here, we show that LidA preferentially binds Rab6A' in the active GTP-bound conformation. Rab6 binding occurred through the central region of LidA and followed a stoichiometry for LidA and Rab6A' of 1:2. LidA maintained Rab6A' in the active conformation by efficiently blocking the hydrolysis of GTP by Rab6A', even in the presence of cellular GTPase-activating proteins, suggesting that the function of Rab6A' must be important for efficient intracellular replication of L. pneumophila. Accordingly, we found that production of constitutively inactive Rab6A'(T27N) but not constitutively active Rab6A'(Q72L) significantly reduced the ability of L. pneumophila to initiate intracellular replication in human macrophages. Thus, the presence of an active pool of Rab6 within host cells early during infection is required to support efficient intracellular growth of L. pneumophila.

Fig 1

The central region of LidA binds Rab6A′ in a nucleotide-dependent manner. (Left) Schematic representation of LidA and the truncated variants used in this study. The numbers indicate amino acid residues. (Right) Pulldown of LidA variants by Rab6A′. Purified recombinant full-length LidA or LidA variants were incubated with beads coated with either constitutively active Rab6A′(Q72L) or inactive Rab6A′(T27N), and proteins pelleted by the beads were visualized by SDS-PAGE and Coomassie staining. The gel images are representatives of two repetitions.

Fig 2

Rab6A′ and Rab1 compete for LidA binding. (A) Rab1 interferes with LidA precipitation by Rab6A′. Agarose beads coated with either GST or GST-Rab6A′(Q72L) were incubated with LidA, and LidA binding to protein-immobilized beads was determined by SDS-PAGE and Coomassie staining. Rab1 was added to LidA either 30 min prior (t = −30 min) or 5 min after (t = +5 min) addition of Rab6A′(Q72L)-coated beads. SidM served as negative control. (B) Rab6A′ prevents Rab1 from binding to LidA. Complex formation between LidA and Rab1 was analyzed by isothermal titration calorimetry in either the absence (left graph) or presence (right graph) of a 2-fold molar excess of Rab6A′. (Top) Raw data for 14 injections of full-length LidA into the isothermal titration calorimetry cell containing wild-type Rab1. Each injection peak corresponds to the heat released during that injection. (Bottom) Scatter plot showing the binding isotherm for the Rab1-LidA interaction, with the integrated heat plotted against the stoichiometry.

Fig 3

LidA binds Rab6A′ with a 1:2 stoichiometry. (A) Gel permeation chromatogram showing LidA-Rab6A′ complex formation. Purified recombinant Rab6A′ and LidA were mixed at the indicated molar ratios and separated by gel filtration. The protein amount detected by absorbance (at 280 nm) is plotted against the elution volume. The amount of LidA and Rab6A′ within each fraction (0.5 ml) was determined by immunoblotting using protein-specific antibodies. The elution profiles of uncomplexed LidA and Rab6A′ are shown in each chromatogram. (B) Summary of the molecular mass of the LidA-Rab6A′ complex shown in panel A, plotted against the molar ratio of LidA to Rab6A′. (C) Results of the same experiment as shown in panel B, showing LidA-Rab6A′(Q72L) complex formation. (D) Gel permeation chromatogram of Rab6A′ variants. Purified recombinant Rab6A′ wild type (WT) or the indicated Rab6A′ point mutants were analyzed by gel permeation chromatography. Proteins within the eluted fractions were detected by absorbance (at a 280-nm wavelength).

Fig 4

LidA protects Rab6A′ from GAP-mediated inactivation. (A) GapCenA does not interact with Rab6A′. Beads coated with the indicated Rab6A′ proteins or with GST (control) were incubated with purified recombinant GapCenA, and retention of GapCenA by the beads was determined by SDS-PAGE and Coomassie staining. (B) Schematic representation of the GTP hydrolysis assay developed in this study. Rab6A′ loaded with [α-³²P]GTP was incubated with PNS from 293T cell lysate. Samples taken at the indicated time points were denatured in 1% SDS (final concentration), and radiolabeled nucleotides were separated by TLC. (C) Example of TLC plates, showing that PNS but not buffer triggered conversion of [α-³²P]GTP to [α-³²P]GDP by Rab6A′. (D) Quantification of [α-³²P]GDP formation by Rab6A′ incubated with PNS but not by Rab6A′ incubated with buffer or LidA alone. The amount of [α-³²P]GTP generated by Rab6A′ in the presence of PNS after 60 min was arbitrarily set to 100%. (E) LidA blocks GTP hydrolysis by Rab6A′ in a dose-dependent manner. [α-³²P]GTP-loaded Rab6 was incubated with LidA at the indicated molar ratios and with PNS, and generation of [α-³²P]GDP over time was monitored by TLC. The graphs shown in panels D and E are representatives of two independent experiments.

Fig 5

LidA targets Rab6A′ but not its GAP(s). (A) The inhibitory effect of LidA on Rab6A′ GTP hydrolysis is proteinase K sensitive. LidA pretreated with proteinase K (PK) was tested for its ability to prevent GTP hydrolysis by Rab6A′ in the presence of PNS. Note that pretreatment of PNS with the same amount of proteinase K did not affect its GAP activity. (B) Schematic representation of the experiment quantified in panel C, showing that pretreatment of PNS with LidA-coated beads did not affect the ability of PNS to trigger [α-³²P]GTP hydrolysis by Rab6A′. (D) The central region of LidA protects Rab6A′ from inactivation. [α-³²P]GTP-loaded Rab6A′ was incubated with PNS and the indicated LidA variants. Generation of [α-³²P]GDP after 30 min was determined by TLC. Each graph is representative of two independent experiments, and the amount of [α-³²P]GDP is displayed relative to the positive control, which was arbitrarily set to 100%.

Fig 6

Rab6 is required for efficient intracellular replication of L. pneumophila. (A) U937 cells transiently transfected with plasmids carrying the indicated Rab6A′ protein gene were challenged for 1 h with either L. pneumophila Lp02 or a lidA deletion strain at an MOI of 5. Extracellular bacteria were removed, and cells were incubated for an additional 12 h at 37°C. The efficiency of L. pneumophila replication vacuole formation was determined by counting the number of bacteria in at least 100 vacuoles. Results represent the means ± standard deviations of three independent experiments. ***, P < 0.001; **, P < 0.01 (t test). (B) Coprecipitation of Rab6A/A′ and LidA from L. pneumophila-infected cells. Human U937 macrophages were challenged with L. pneumophila Lp02 (WT) or a lidA deletion strain (ΔlidA). After 2 h or 6 h, the cells were mechanically lysed, LidA was precipitated using anti-LidA antibody-coated beads, and the amount of Rab6A/A′ bound to the beads was detected by immunoblotting using antibody directed against Rab6A/A′. The bottom panel shows the total amount of Rab6A/A′ within the PNS. Images are representative of two independent experiments. (C) LidA interferes with binding of cellular ligands to Rab6A′. Uncoated control beads or beads coated with either GST-Rab6A′(T27N) or GST-Rab6A′(Q72L) were incubated with postnuclear supernatant from transiently transfected 293T cells producing GFP-tagged R6IP1A or R6IP2B. Beads were washed, and proteins retained by the beads were detected by immunoblotting using anti-GFP antibody. Each panel is representative of three independent experiments.