Divergent Evolution of Legionella RCC1 Repeat Effectors Defines the Range of Ran GTPase Cycle Targets

Abstract

Legionella pneumophila governs its interactions with host cells by secreting >300 different "effector" proteins. Some of these effectors contain eukaryotic domains such as the RCC1 (regulator of chromosome condensation 1) repeats promoting the activation of the small GTPase Ran. In this report, we reveal a conserved pattern of L. pneumophila RCC1 repeat genes, which are distributed in two main clusters of strains. Accordingly, strain Philadelphia-1 contains two RCC1 genes implicated in bacterial virulence, legG1 (Legionella eukaryotic gene 1), and ppgA, while strain Paris contains only one, pieG The RCC1 repeat effectors localize to different cellular compartments and bind distinct components of the Ran GTPase cycle, including Ran modulators and the small GTPase itself, and yet they all promote the activation of Ran. The pieG gene spans the corresponding open reading frames of legG1 and a separate adjacent upstream gene, lpg1975legG1 and lpg1975 are fused upon addition of a single nucleotide to encode a protein that adopts the binding specificity of PieG. Thus, a point mutation in pieG splits the gene, altering the effector target. These results indicate that divergent evolution of RCC1 repeat effectors defines the Ran GTPase cycle targets and that modulation of different components of the cycle might fine-tune Ran activation during Legionella infection.IMPORTANCELegionella pneumophila is a ubiquitous environmental bacterium which, upon inhalation, causes a life-threatening pneumonia termed Legionnaires' disease. The opportunistic pathogen grows in amoebae and macrophages by employing a "type IV" secretion system, which secretes more than 300 different "effector" proteins into the host cell, where they subvert pivotal processes. The function of many of these effector proteins is unknown, and their evolution has not been studied. L. pneumophila RCC1 repeat effectors target the small GTPase Ran, a molecular switch implicated in different cellular processes such as nucleocytoplasmic transport and microtubule cytoskeleton dynamics. We provide evidence that one or more RCC1 repeat genes are distributed in two main clusters of L. pneumophila strains and have divergently evolved to target different components of the Ran GTPase activation cycle at different subcellular sites. Thus, L. pneumophila employs a sophisticated strategy to subvert host cell Ran GTPase during infection.

Keywords: Acanthamoeba; Dictyostelium; Legionella; amoeba; bacterial evolution; effector protein; guanine nucleotide exchange factor; host-pathogen interaction; macrophage; microtubule; pathogen vacuole; phosphoinositide lipid; small GTPase; type IV secretion; vesicle trafficking.

FIG 1

Distribution of L. pneumophila RCC1 repeat genes. The phylogenetic tree of 59 L. pneumophila strains was inferred using the fast core genome multialigner Parsnp (https://github.com/marbl/parsnp). Circles at nodes represent bootstrap support, and the size of each circle is proportional to the corresponding bootstrap value. The scale bar represents the estimated number of substitutions per site. For each strain, the presence or absence of the RCC1 repeat effector genes lpp1959 (pieG) and lpg2224 (ppgA) was assessed by OrthoMCL (https://orthomcl.org), and all of the strains were found to be distributed in two main clusters. In strain C4S, a split fragment of gene lpp1959 is a pseudogene. Strain NCTC11404 does not harbor an lpg2224 gene (verified by blast search), and one of three lpg2224 copies in strain D-7630 is a pseudogene.

FIG 2

L. pneumophila RCC1 repeat effectors and their role in pathogen-host interactions. (A) Murine RAW 264.7 macrophages were infected (multiplicity of infection [MOI] 0.1) with L. pneumophila strain JR32, ΔicmT, ΔppgA, ΔlegG1, or ΔppgA-ΔlegG1, the cells were lysed, and intracellular replication at 37°C was assessed by CFU counting. (B) D. discoideum was infected (MOI 10) with L. pneumophila strain JR32, ΔicmT, ΔppgA, ΔlegG1, or ΔppgA-ΔlegG1 producing GFP (pNT28), and intracellular replication at 25°C was assessed by GFP fluorescence increase. Data show means and standard deviations of results from triplicates (one-way analysis of variance [ANOVA]; ***, P < 0.001). RFU, relative fluorescence units. (C) D. discoideum was infected (MOI 0.1) with wild-type L. pneumophila strain Paris or with the ΔdotA or ΔpieG mutant strain, and intracellular replication at 25°C was assessed by CFU counting in lysates of infected cells (3 independent experiments; two-way ANOVA; *, P < 0.1; **, P < 0.01; ***, P < 0.001). (D and E) Real-time fluorescence microscopy of LCV motility in D. discoideum producing calnexin-GFP (pCaln-GFP) infected (MOI 5, 1 to 2 h) with L. pneumophila JR32, ΔppgA, ΔlegG1, or ΔppgA-ΔlegG1 producing DsRed alone (pCR077) or together with M45-LegG1 (pER005) or M45-PpgA (pLS008) (D), or with the Paris wild-type strain or the ΔpieG mutant strain producing DsRed alone (pCR077) or together with M45-PieG (pLS033), M45-LegG1 (pER005), or M45-PpgA (pLS008) (E). LCV motility was recorded for 180 s with images taken every 10 s and quantified using ImageJ/Fiji software with the manual tracking plugin (n > 50/strain; 3 independent experiments; one-way ANOVA; **, P < 0.01; ***, P < 0.001; compared to the wild-type {D} or ΔpieG {E} strain). (F) D. discoideum producing GFP (pDM317) was infected (MOI 5, 1 h) with L. pneumophila JR32, ΔppgA, ΔlegG1, or ΔppgA-ΔlegG1 producing DsRed alone (pCR077) or together with M45-LegG1 (pER005) or M45-PpgA (pLS008) and was seeded in a culture inset 2-well dish (Ibidi) for 2 h. After removal of the inset, cell migration was analyzed by confocal microscopy at 0 and 3 h. Bars, 200 μm. (G) Cell migration was quantified using ImageJ/Fiji software. The cell density ratio represents the average fluorescent signal intensity in the 500 μm gap divided by the average fluorescent signal intensity in 250 μm on each side of the gap center (3 independent experiments; two-way ANOVA; **, P < 0.01; ***, P < 0.001; all groups compared to wild-type strain at 3 h [t3]). (H) D. discoideum producing GFP-PpgA (pLS078), GFP-LegG1 (pER017), or GFP-LegG1_ΔCAAX (pER016) was infected (MOI 5, 1 h) with L. pneumophila JR32 or ΔicmT producing DsRed (pSW001), and localization of RCC1 repeat effectors was analyzed by confocal microscopy. Bars, 1 μm.

FIG 3

PpgA requires functional RanGAP1 to inhibit yeast growth. (A) The localization of L. pneumophila RCC1 repeat effectors was analyzed by confocal microscopy in S. cerevisiae wild-type strain BY4741 producing GFP-PpgA (pLS120), GFP-LegG1 (pLS118), or GFP-PieG (pLS113). Bars, 1 μm. (B) Growth of yeast producing RCC1 repeat proteins was tested by dot spot assays. S. cerevisiae wild-type strain BY4741 containing an empty plasmid (pYEP351gal) or producing FLAG-PpgA (pLS085), FLAG-LegG1 (pLS084), or FLAG-PieG (pLS086) was spotted in 10-fold dilutions on SG plates without leucine and grown at the indicated temperatures for 5 (30°C and 37°C), 6 (25°C), or 7 (20°C) days (left panel). Colony size (at 30°C) was measured using ImageJ/Fiji (right panel; n = 20; one-way ANOVA; ***, P < 0.001). (C) S. cerevisiae mutant strains rna1-1 (RanGAP1), prp20-1 (RCC1), yrb1-51 (RanBP1), and yrb2Δ (RanBP2) containing an empty plasmid (pYEP351gal) or a plasmid producing FLAG-PpgA (pLS085), FLAG-LegG1 (pLS084), or FLAG-PieG (pLS086) were spotted in 10-fold dilutions on SG plates without leucine and grown at the indicated temperatures for 5 (30°C), 6 (25°C and 37°C), or 7 (20°C) days (left panel). Colony size was measured using ImageJ/Fiji (right panel) (n > 10; one-way ANOVA; ***, P < 0.001). (D) S. cerevisiae strain BY4741 or mutant strain rna1-1 containing the empty plasmids pYEP351gal and pRS316, or pYEP351gal and pRS316-RNA1 (RNA1), pRS316 and pLS085 (FLAG-PpgA), or pRS316-RNA1 (RNA1) and pLS085 (FLAG-PpgA) was spotted in 10-fold dilutions on SG plates without leucine and uracil and grown at the indicated temperatures for 3 to 6 days. Colony size was measured using ImageJ/Fiji (one-way ANOVA; n > 9; ***, P < 0.001). (E) S. cerevisiae wild-type strain BY4741 or mutant strain rna1-1 containing the empty backbone (pYEP351gal) or producing FLAG-PpgA (pLS085), FLAG-LegG1 (pLS084), or FLAG-PieG (pLS086) was grown in SG medium without leucine for 7 days at 20°C, and OD₆₀₀ was measured every hour. Data show means and standard deviations (for simplicity, only the standard deviations at 50 h, 100 h, 150 h, and 167 h are shown) of results from three independent experiments (two-way ANOVA; **, P < 0.01; ***, P < 0.001).

FIG 4

PpgA and LegG1 target distinct Ran GTPase cycle components and promote Ran activation. (A) S. cerevisiae BY4741 ectopically producing Strep-tagged RNA1 (pLS128) or RNA1-1 (pLS127) and FLAG-tagged PpgA (pLS085) or LegG1 (pLS084) was lysed, a co-IP with anti-FLAG was performed, and interactions were revealed by anti-Strep Western blotting. Yeast lysates were analyzed by Western blotting with anti-Strep (input control) and anti-FLAG (co-IP control). (B) Lysates of HEK 293T cells were incubated with Strep-LegG1_33–286 bound to Strep-Tactin resin, and interaction partners were identified by mass spectrometry. (C) Yeast two-hybrid assays using S. cerevisiae reporter strain AH109 containing plasmids encoding the GAL4 DNA-binding domain (BD; pGBKT7) alone or fused to RanBP10 (pLS213), RanGAP1 (pM1593), Ran_WT (pM1057), Ran_T24N (pM1059), Ran_Q69L (pM1058), or Ran_N122D (pM1286) and the GAL4 activation domain (AD) fused to LegG1 (pLS211). Transformants were spotted onto SD lacking His (−His) or SD plus His (+His) and incubated at 30°C for 5 days. (D) Yeast two-hybrid assay in reporter strain AH109 containing plasmids encoding the GAL4 DNA-binding domain (BD; pGBKT7) alone or fused to RanBP10 (pLS213) and the GAL4 activation domain (AD; pGADT7) alone or fused to LegG1 (pLS211), PpgA (pM1028) or PieG (pM1027). Transformants were spotted in 10-fold serial dilutions onto SD lacking His (−His) or SD plus His (SD+) and incubated at 30°C for 5 days. (E) Human A549 cells transfected for 48 h with 10 nM siRNA oligonucleotides targeting RanGAP1, RanBP10, or Arf1 (positive control) or with AllStars siRNA (“scrambled”) were infected (MOI 10) with GFP-producing L. pneumophila JR32 (pNT28). Intracellular bacterial replication was assessed by fluorescence increase with a fluorescence plate reader after 24 h and compared to the levels seen at 1 h. Means and standard deviations of results from three independent experiments are shown (one-way ANOVA; **, P < 0.01; all groups compared to scrambled). (F) The depletion efficiency of siRNA oligonucleotides targeting RanGAP1 or RanBP10 upon transfection of A549 epithelial cells for 48 h was assessed by Western blotting with the antibodies indicated. For each target protein four different oligonucleotides were used. Untreated cells were used as a negative control (“Untr.”), Qiagen AllStars unspecific oligonucleotides (“Scrambled,” “Scr.”) were used to control for off-target effects, and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) served as the loading control. Data are representative of results from two independent experiments. (G) HEK 293T cells were transfected for 24 h with constructs producing Strep-tagged PpgA (pLS229), LegG1 (pLS226), PieG (pLS230), or RCC1 (pLS231). Nontransfected cells (-) were taken along, and lysates were treated with GTPγS (positive control) or GDP (negative control). Ran(GTP) was precipitated with RanBP1-coupled agarose beads in cell lysates, and the amount of Ran(GTP), total Ran, and GAPDH (loading control) was assessed by Western blotting (top panel). The amount of Ran(GTP) relative to total Ran was quantified using ImageQuant TL (bottom panel). Means and standard deviations of results from three independent experiments are shown for the first four bands (one-way ANOVA; *, P < 0.05; **, P < 0.01; ***, P < 0.001; all groups compared to the nontransfected cells [-]).

FIG 5

PieG and the fusion protein Lpg1975-LegG1 interact with Ran and RanGAP1. (A) A point mutation (insertion of thymidine phosphate) in L. pneumophila lpg1975 (strain Philadelphia-1) causes a frameshift and fuses the open reading frames of lpg1975 and legG1 (lpg1976), resulting in a fusion gene with high homology to pieG (strain Paris). (B) A co-IP screen was performed with yeast coproducing FLAG-PpgA (pLS085), FLAG-LegG1 (pLS084), FLAG-PieG (pLS086), or FLAG-Lpg1975-LegG1 fusion (pLS087) and RNA1-Strep (pLS128), Strep-PRP20 (pLS185), Strep-YRB1 (pLS186), or Strep-GSP1 (pLS173). Yeast lysates were incubated with FLAG beads, and eluates were analyzed for co-IP of Strep-proteins by anti-Strep Western blotting. Yeast lysates were analyzed by Western blotting with anti-Strep (input control) and anti-FLAG (co-IP control). (C) Lysates of yeast harboring pYEP351gal-FLAG-ppgA (pLS085), pYEP351gal-FLAG-legG1 (pLS084), pYEP351gal-FLAG-lpg1975-legG1 fusion (pLS087), or pYEP351gal-FLAG-pieG (pLS086) were incubated with FLAG-agarose beads, and eluates were analyzed for the co-IP of endogenous GSP1 using a specific antibody. (D) HEK 293T cells were transfected for 24 h with plasmids harboring GFP-Lpg1975 (pLS233), GFP-LegG1 (pLS227), GFP-Lpg1975-LegG1 (Fusion) (pLS234), GFP-PpgA (pLS095), GFP-PieG (pLS093), or GFP-PieG_split (pLS235) and Strep-RanBP10 (pLS242). Cell lysates were analyzed by anti-GFP Western blotting (input control), or incubated with anti-Strep antibody and A/G agarose beads, and eluates were analyzed by Western blotting for co-IP of GFP fusion proteins using anti-GFP and anti-Strep (co-IP control). (E) Real-time fluorescence microscopy of LCV motility in D. discoideum producing calnexin-GFP (pCaln-GFP) (green) infected (MOI 5, 1 to 2 h) with L. pneumophila JR32 (WT) or with the ΔlegG1-ΔppgA mutant producing DsRed alone (pCR077) or together with M45-tagged PieG (pLS033), Lpg1975-LegG1 fusion protein (pLS026), LegG1 (pER005), or PpgA (pLS008). LCV motility was recorded for 180 s with images taken every 10 s and quantified using ImageJ/Fiji software with the manual tracking plugin (n > 50/strain; 3 independent experiments; one-way ANOVA; ***, P < 0.001; all groups compared to ΔlegG1-ΔppgA).

FIG 6

Ran GTPase cycle targets of L. pneumophila RCC1 repeat effectors. The different Ran GTPase cycle targets of the L. pneumophila RCC1 repeat effectors LegG1, PpgA and PieG (red) are indicated. Divergent evolution of L. pneumophila RCC1 repeat effectors defines the range of target components of the Ran GTPase cycle. Regardless of the distinct target, the L. pneumophila RCC1 repeat effectors promote the activation of Ran. Putative activation (>), inhibition (|), or binding (×) of Ran GTPase cycle components by L. pneumophila RCC1 repeat effectors is indicated with different arrow endings. For details see text.