Sde proteins coordinate ubiquitin utilization and phosphoribosylation to establish and maintain the Legionella replication vacuole

Abstract

The Legionella pneumophila Sde family of translocated proteins promotes host tubular endoplasmic reticulum (ER) rearrangements that are tightly linked to phosphoribosyl-ubiquitin (pR-Ub) modification of Reticulon 4 (Rtn4). Sde proteins have two additional activities of unclear relevance to the infection process: K63 linkage-specific deubiquitination and phosphoribosyl modification of polyubiquitin (pR-Ub). We show here that the deubiquitination activity (DUB) stimulates ER rearrangements while pR-Ub protects the replication vacuole from cytosolic surveillance by autophagy. Loss of DUB activity is tightly linked to lowered pR-Ub modification of Rtn4, consistent with the DUB activity fueling the production of pR-Ub-Rtn4. In parallel, phosphoribosyl modification of polyUb, in a region of the protein known as the isoleucine patch, prevents binding by the autophagy adapter p62. An inability of Sde mutants to modify polyUb results in immediate p62 association, a critical precursor to autophagic attack. The ability of Sde WT to block p62 association decays quickly after bacterial infection, as predicted by the presence of previously characterized L. pneumophila effectors that inactivate Sde and remove polyUb. In sum, these results show that the accessory Sde activities act to stimulate ER rearrangements and protect from host innate immune sensing in a temporal fashion.

Fig. 1. The Sde DUB domain is required for efficient phosphoribose-ubiquitination of Rtn4.

A Domain structure of Sde family proteins. Catalytically inactive point mutations are shown in red. B Percent poly-Ub positive LCVs of L. pneumophila strains at noted timepoints post-infection. C Quantification of Rtn4-positive LCVs. BMDMs were challenged with indicated strains for the noted infection times, followed by fixation, permeabilization with 1% Triton X-100 and probing as described in Methods. 0 MPI was immediately after 5 min centrifugation of bacteria onto BMDM. Noted strains are wild type, Lp02; Δsde, ΔsidEΔsdeCΔsdeB-A; Vector or pV, pJB908att-empty (Supplementary Table S1). For each experiment, n > 80 LCVs (B) or n > 50 LCVs (C) per experiment were evaluated and data were determined from 3 biological replicates (mean ± SEM; two-way ANOVA with Tukey’s multiple comparison (B, C); ns (non-significant); exact P values are displayed over data, and are found in Source Data File. For each replicate, n value is shown in Source Data File. D–I Representative micrographs of Rtn4 recruitment, scale bar = 5 µm. Scalebar determined for each image independently and pasted on each image. Macrophages were challenged with either WT containing pV or noted mutant strains for 1 h, fixed, permeabilized and probed with anti-L. pneumophila (Alexa Fluor 594 secondary, red), anti-Rtn4 (Alexa Fluor 488 secondary, green), and Hoechst (nucleus, blue). J Cartoon of strategy to identify pR-Ub modified Rtn4. K Immunoblot image of anti-Rtn4 immunoprecipitates by immunoprobing with anti-HA. To right of immunoblots are Rtn4 isoforms and modification status. HEK293T cells were challenged with noted L. pneumophila strains for either 10 or 180 min. Modified Rtn4 substrates were immunoprecipitated using crosslinked anti-Rtn4, fractionated on 7.5% SDS-PAGE, and probed with anti-HA. Lanes: WT, Lp02; vector, Δsde; psdeA, Δsde expressing SdeA; psdeA C118S, Δsde expressing SdeA DUB mutant; psdeB, Δsde expressing SdeB; psdeB C118S, Δsde expressing SdeB DUB mutant; MOCK, uninfected. Masses are apparent molecular weights as noted in kDal. Source data are provided as a Source Data file.

Fig. 2. The ART domain blocks DUB activity and promotes hyper-polyubiquitination of the LCV.

A SdeC WT or its catalytic mutant derivatives (1 μg) were adsorbed to Ni-NTA resin for 1 h at 4 °C, then incubated with either K63 or K48-linked polyUb_3-7 (2 μg) for 2 h at 37 °C. SdeC resin was removed and polyUb_3-7 was then incubated with: human recombinant Polyhistidine-otubain 1, isoform 1 (100 nM); recombinant USP2 catalytic domain (50 nM); recombinant Polyhistidine-CYLD (50 nM); or recombinant SdeC DUB CD 1-192 (50 nM) at 37 °C for 2 h. Cleavage was monitored by immunoblotting with anti-Ub (FK1). Shown are data from one of two independent replicates performed on separate days with similar results. Molecular weights are shown to left of each blot using protein standards of apparent molecular masses 25 kDal and 75 kDal. B Quantification of polyUb intensity associated with individual LCVs of the noted strains at 20 MPI. Data shown are means ± 95% confidence intervals, analyzed by one-way ANOVA with Dunnett’s multiple comparison. Data are pooled from 3 biological replicates. WT, n = 21; C118S, n = 23; E859, n = 17; H416, n = 28. The dotted lines represent background level of polyUb intensity about LCVs. C PolyUb intensity of LCVs harboring Δsde expressing SdeC PDE^- (H416A) mutant at 20 or 60 MPI. Data are shown in median ± 25%ile; Mann–Whitney 2-tailed test. The dotted lines represent background level of polyUb intensity about LCVs. Data are pooled from 3 biological replicates. 20 MPI, n = 28; 60 MPI n = 25. All P values are found in Source Data File. D–H Examples of polyUb-positive LCVs. Shown are images from one of three biological replicates with similar results. BMDMs were challenged by 5 min centrifugation at an MOI of 1, incubated with noted L. pneumophila strains, fixed, permeabilized with 1% Triton X100, and probed with anti-polyUb (Alexa Fluor 488 secondary, green), anti-L. pneumophila (Alexa Fluor 594 secondary, red), and Hoechst (blue). Strains used: WT, Lp02; Δsde, ΔsidE ΔsdeC ΔsdeB-A; Δsde psdeC_WT, Δsde expressing SdeC WT; Δsde psdeC_C118S, Δsde expressing SdeC DUB mutant; Δsde psdeC_E859A, Δsde expressing SdeC ART mutant; Δsde psdeC_H416A, Δsde expressing SdeC PDE mutant. Scale bar: 5μm. Source data are provided as a Source Data file.

Fig. 3. Phosphoribosyl modification of Ub4 occurs at Arg42.

a SdeC(DUB^-) incubation with Ub₄ results in a mass increase of 848.21 AMU. Samples were subjected to LC-MS analysis and the deconvoluted masses of the peaks for each sample are displayed. Top: untreated Ub₄. Bottom: Ub₄ incubated with SdeC(DUB^-) and NAD. b Sde modification of Ub₄ occurs exclusively on Arg42. Representative extracted ion chromatograms (EICs) are shown for the pR-modified R42 tryptic fragment, as well as tryptic fragments containing R54 and R72. The unmodified R42 fragment could not be found in SdeC-treated samples. c MS/MS analysis of the pR-modified tryptic fragment. Observed diagnostic ions that confirm specificity to R42 are shown in bold. Blue: b ions. Red: y ions. Source data are displayed. Underlying data are available on ProteomeXchange repository (www.proteomexchange.org/).

Fig. 4. Sde-mediated modification of polyUb prevents its recognition by p62.

A Space filling model of Ub. R42 residue (blue letter) and I44 residue are indicated with blue arrows (from reference^,). B Interface residues between UBA domain and Ub, showing possible involvement of Ub R42 in interacting with UBA domain (from Refs. ^,). C, D Ub modification at R42 blocks p62 binding (Methods). Ni-NTA-p62 was incubated with Ub_3-7 that had been treated with SdeC WT or its catalytic mutant derivatives and allowed to bind (In). Unbound Ub_3-7 was removed (FT), the beads were washed, and then eluted with 250 mM imidazole (E: elution). The p62-bound poly-Ub in the eluate was fractionated by SDS-PAGE followed by probing with anti-Ub. Relative loading: In, input (2.5%); FT, flow-through (2.5%); E, eluate (25%). Molecular weight markers are from protein standards of 25 kDal and 75 kDal apparent molecular weights. Shown are data from one of two independent experiments. E Surface plasmon resonance (SPR) quantitation of p62 binding to Ub derivatives. Ub₁ or Ub₄ was incubated with SdeC WT or its derivatives immobilized on Ni-NTA resin, the supernatant containing modified Ub derivatives was incubated with 6xHis-tagged p62 bound to a Ni-NTA chip, and binding affinity was measured by plasmon resonance. F, G Sensorgrams of Ub₄ or Ub₁ binding to p62. Binding of Ub₄ to p62 yields a K_D = 0.97 × 10⁻⁷ M. H–K Sensorgrams of Ub₄, modified as noted in panels (red balls: phosphoribosyl addition; green balls: ADP-ribosyl addition). Reactions were formed in the presence of Ub-propargylamide (DUB inhibitor). All sensorgrams are source data. All other data are available as a Source Data file. Source data are provided as a Source Data file.

Fig. 5. Transient blockade of p62 recruitment to the LCV in the presence of Sde family members.

A, B Examples of punctate (A) or enveloped (B) p62 morphology localized around LCVs after challenge of BMDMs with noted L. pneumophila strains. The scale bar represents 10 µm. C, E, G, J The Percentage of circumferential p62 associated with LCVs at 0 (immediate bacterial contact; C) or 20 (E) or 60 (G) or 180 (J) MPI. 100 LCVs per replicate were counted in n = 3 biological replicates (mean ± SEM; one-way ANOVA with Dunnett’s multiple comparison; ns (non-significant). P values are displayed over data. D, F, H, K Quantification of mean p62 intensity/pixel around LCVs at 0 (D) or 20 (F) or 60 (H) or 180 (K) MPI of vacuoles analyzed in (C, E, G, J). Image analysis performed on same coverslips at corresponding timepoints. Number of cells analyzed (n) for each strain at each timepoint is provided in Source Data File. At least 70 vacuoles per replicate were quantified and data were pooled from 3 biological replicates (mean ± SD; one-way ANOVA with Dunnett’s multiple comparison (D, F, H) and Student’s two tailed t-test (K); ns (non-significant). Exact P values are displayed over data. I A representative micrograph of p62 association with LCVs at 180 MPI, scale bar 10 µm. L Kinetics of blockade against p62 recruitment. All strains described in Supplementary Table S1 (pV: pJB908att-empty). Data acquired from (C), (E), (G) and (J), and shown as mean ± SEM. All P values are found in Source Data File. Source data are provided as a Source Data file.

Fig. 6. ART activity is sufficient to block p62 recruitment to the replication vacuole.

A Quantification of individual LCVs enveloped by p62 at 0 MPI. BMDMs were challenged with the noted Legionella strains. 100 LCVs per replicate were counted in n = 3 biological replicates (mean ± SEM; one-way ANOVA with Dunnett’s multiple comparisons; ns (non-significant). Exact P values are displayed over data. B Quantitation of mean p62 intensity/pixel associated with individual LCVs. Images of individual vacuoles were grabbed and pixel intensities of p62 staining about regions of interest were determined (Methods). At least 198 LCVs were counted from the replicates performed in (A). Data are shown as means ± SD (one-way ANOVA with Dunnett’s multiple comparison; ns (non-significant). Exact P values are displayed over data. All P values are found in Source Data File. C Examples of recruited p62 to LCVs harboring Δsde expressing SdeC catalytic mutant derivatives. All strains described in Supplementary Table S1 (pV: pJB908att-empty). Scale bar: 10 μm. Source data are provided as a Source Data file.

Fig. 7. Sde accessory activities drive Rtn4 rearrangements and block autophagic recognition.

The DUB activity provides fuel for pR-Ub modification of Rtn4, driving rapid ER-associated rearrangements. A Bacteria are internalized, Sde family proteins are translocated. B Polyubiquitination occurs about the replication vacuole. C Digestion of polyUb provides Ub fuel for ART/PDE domain to catalyze pR-Ub conjugation to substrates such as Rtn4. D Rtn4-pR-Ub modified protein accumulates about the replication vacuole polyUb is modified by phosphoribosylation. The pR-modification of the polyUb at R42 residue stabilizes polyubiquitination about LCVs, simultaneously preventing degradation by host cell-derived DUBs and blocking recognition by the autophagic machinery.