Legionella effector LnaB is a phosphoryl-AMPylase that impairs phosphosignalling

a, Native PAGE analyses of Ub, ADPR^R42-Ub and PR^R42-Ub after treatments by the cell lysates of the wild-type and sidE-deleted strains of L. pneumophila. b, In vitro ARH assays of MavL toward ADPR^R42-Ub. The ADP-ribosyl hydrolysation of ADPR^R42-Ub was examined by the mobility shifts on native PAGE gels and Coomassie blue staining. c, Interactions of MavL with polyUb and Ub. d, 2mFo-DFc and mFo-DFc electron density maps of Ub in the MavL–Ub complex. The 2mFo-DFc map is shown as blue meshes and contoured at 1.0 σ. The mFo-DFc map is shown as meshes coloured in green (positive) and red (negative) and contoured at 3.0 σ. e,f, Detailed interactions of MavL with Ub in the MavL–Ub complex structure. The mFo-DFc electron density map of the interacting residues (f) was calculated with the residues and loop regions removed and is shown as meshes coloured in green (positive) and red (negative) and contoured at 3.0 σ. g, Structural superimposition of MavL with Thermomonospora curvata PARG¹³ (PDB ID: 3SIG). The ADPR ligand in the PARG structure is represented as sticks in cyan. The catalytic pocket of MavL is highlighted within the dashed box. h, Close-up view of the ADPR-binding pockets of MavL and PARG. The catalytic residues of MavL and PARG are shown as sticks. i, Interactions of the catalytic pocket of MavL with the ADPR ligand that was modelled from PARG, according to the structural superimposition of MavL and PARG (g). j,k, Mutagenetic analyses of the Ub-interacting residues of MavL on its interactions with Ub (j) and its ARH activity (k). All uncropped and unprocessed scans of blots and gels are listed in Supplementary Fig. 1.

a, Mass spectrometric analysis of Ub, ADPR^R42-Ub and the MavL-produced Ub from ADPR^R42-Ub. The peak of the MavL product overlapped with that of Ub. b, In vitro ARH assays of MavL toward ADPR–actin. The ADP-ribosyl hydrolysation of ADPR^R42-Ub and ADPR–actin was examined by SDS–PAGE and immunoblotting with the anti-ADPR antibody. c,d, In vitro ARH assays of MavL toward ADPR–PARP1 (c) and ADPR–CycT1 (d). Recombinant auto-ADP-ribosylated PARP1 protein (ADPR–PARP1), and ADPR–CycT1 that was immunoprecipitated from H₂O₂-treated 293T cells were incubated with MavL. *, nonspecific band (c). e, Interactions of the inactive D323A mutant of MavL with ADP-ribosylated Ub and polyUb. Ub and polyUb immunoprecipitated from 293T cells were ADP-ribosylated by the SdeA mART ___domain and then incubated with the FLAG-tagged D323A mutant that was also expressed and immunoprecipitated from 293T cells. The interactions were examined by immunoprecipitation with the anti-HA antibody. f, MavL had no effects on the total ADP-ribosylation level of proteins in cells. The total ADP-ribosylation level in the cells was examined by immunoblotting with an anti-ADPR antibody. A PARP1 inhibitor significantly reduced the total ADP-ribosylation level in cells. g, MavL is not a NADase and had no effects on the NAD level in cells. Plasmids containing MavL or the NADase toxin TNT⁵⁰ of Mycobacterium tuberculosis were transfected into 293T cells. The total amount of NAD/NADH in the cells was examined with a NAD/NADH detection kit and BioTek Microplate Reader. Data represent the mean ± SD of four replicates, shown as black dots (n = 4). The P values were calculated by one-way, two-tailed ANOVA. A post hoc Dunnett’s multiple comparisons test was further conducted to compare differences between the two groups in GraphPad Prism 8.0. NS, not significant P > 0.05, **P < 0.01, ****P < 0.0001. h, Schematic diagram of the ARH activity of MavL. i, Detection of ADPR^R42-Ub in the L. pneumophila-infected cells. The cell lysates of U937 cells infected with the wild-type Lp02 strain or the ΔmavL strain were subjected to 14% SDS–PAGE and immunoblotting with the anti-ADPR and anti-Ub antibodies.

a, Representative screening results of the effectors responsible for producing ADPR^R42-Ub from PR^R42-Ub. PR^R42-Ub was treated with the cell lysates of 293T cells transfected with the plasmids encoding the indicated effectors and analysed by SDS–PAGE. b, Other representative screening results of the effectors responsible for producing ADPR^R42-Ub from PR^R42-Ub. c, Discrimination of PR^R42-Ub and ADPR^R42-Ub by 13% phosphate-affinity Phos-tag SDS–PAGE. d, GFP-LnaB immunoprecipitated from 293T cells was incubated with PR^R42-Ub with the ligand mixture of AMP, ADP, ATP and NAD. Effects of depletion of each ligand in the reactions were examined by immunoblotting. e, Immunoblotting analysis of the treatment of PR^R42-Ub by recombinant GST–LnaB protein that was purified from E. coli. f, GST and GST-tagged LnaB purified from E. coli were incubated with 293T cell lysates. After pull-down, GST or GST-tagged LnaB reacted with PR^R42-Ub in the presence of ATP. ADPR^R42-Ub was examined by immunoblotting with the anti-ADPR antibody. g, PR^R42-Ub was incubated with recombinant LnaB protein purified from E. coli and the proteinase K-treated or untreated 293T cell lysates. The samples were analysed by SDS–PAGE. h, Recombinant GST–LnaB protein purified from E. coli was incubated with the 293T cell lysates. After GST pull-down, the samples were analysed by SDS–PAGE and silver staining. The band of actin on the gel is highlighted with an arrow. i, List of the proteins identified by mass spectrometry analysis of the protein band bound by LnaB on SDS–PAGE in (h). j, Interactions of the LnaB fragments with endogenous actin in 293T cells. k, The activities of the LnaB fragments. PR^R42-Ub was treated by the cell lysates of 293T cells expressing GFP-tagged LnaB fragments. l, LnaB and LHLS inhibit yeast growth. LnaB HE/AA, the double mutant of H305A and E309A of LnaB; LHLS HE/AA, the double mutant of H287A and E291A of LHLS.

a, GST pull-down assay of the interactions of the LnaB HE/AA mutant and G-actin. b, The AMPylase activities of the LnaB mutants H305A, E309A, H305A/E309A and D312A. The recombinant LnaB mutant proteins purified from E. coli were incubated with PR^R42-Ub, actin and ATP. PR^R42-Ub and ADPR^R42-Ub were examined by 13% Phos-tag SDS–PAGE. c, The AMPylase activities of the LnaB mutants E156A, E276A, E281A and Y348A. d, Mass spectrometric analyses of the R42-containing peptide (EGIPPDQQRLIFAGK) in PR^R42-Ub and the LnaB-produced ADPR^R42-Ub. e, Determination of the phosphoribosylation site on PR^R42-Ub that was generated by SdeA in vitro and used as the substrate of LnaB by ETD MS/MS. The MS/MS spectrum of the phosphoribosylated peptide EGIPPDQQR^PRLIFAGK is shown as indicated. f, Schematic diagram of the activities of LnaB and MavL. g, The reversal of the SdeA-catalysed noncanonical ubiquitination by LnaB and MavL. *, a nonspecific band of the crosslinked PR-Ub. h, Localization of LnaB on the Legionella-containing vacuoles during infection. U937 cells were infected with the Lp02 and Lp03 strains that were transformed with the pJB908 plasmid containing the LnaB gene with an N-terminal HA tag for 3 h and examined by immunofluorescence with Zeiss LSM710 confocal microscope. The scale bar represents 5 μm. i,j, Double deletion of mavL and lnaB genes inhibited the auto-ubiquitination of TRAF6 (i) and degradation of IκBα (j) during infection. k, Effects of the gene deletion of mavL and lnaB on the ubiquitination of RIPK1 in 293T cells infected with the indicated L. pneumophila Lp02 strains. Ubiquitination of RIPK1 in the cells was analysed by immunoblotting after immunoprecipitation. l, Effects of the gene deletion of mavL and lnaB on the ubiquitination of RIPK3 in U937 cells infected with the indicated Lp02 strains. *, the heavy chain of antibody.

a, Recombinant GST-tagged LnaB and its truncated mutant (residues 1–327), which were purified from E. coli, were incubated with the Acanthamoeba castellanii cell lysate. The GST pull-down assays were analysed by SDS–PAGE and immunoblotting with an anti-α-actin antibody recognizing the amoebic actin. b, The actin-contained cell lysates of A. castellanii amoebae activated the AMPylase activity of LnaB. c, The intracellular growth of the L. pneumophila strains in amoebae and macrophage cells. Acanthamoeba castellanii and U937 cells were infected with the indicated L. pneumophila strains. Experiments were performed in triplicate and repeated three times with similar results. error bars represent s.d. (n = 3). d, Multiple-sequence alignment of the residues around the two conserved catalytic consensuses of LnaB family members. 40 homologous sequences were selected for analysis. e,f, Interactions of the fragments of LnaB (e,f), LHLS (e), LHEI (e) and LHBA proteins (f), respectively, with endogenous actin in transfected 293T cells. g,h, The actin-activated AMPylase activities of the LnaB homologues. PR^R42-Ub was incubated by recombinant LnaB, LHLS (g), LHEI (g) and LHBA proteins (h), respectively, with actin and ATP. *, the degraded fragments of GST-LHEI. i, The AlphaFold2-predicted structure of LnaB. The prediction quality is indicated by the pLDDT (predicted Local Distance Difference Test) scores ranging from blue for high confidence to red for low confidence. The catalytic ___domain of LnaB is highlighted by a black dashed box. j, Sequence coverage scores of LnaB in the structure prediction. k, The pLDDT scores of the five AlphaFold2-predicted LnaB models. l, Comparison of the AlphaFold2-predicted structures of the catalytic domains of LHLS, LHBA and LHEI with that of LnaB.

a, Cryo-EM data processing flowchart and local resolution distributions of the reconstructions of the wild-type LHLS–actin–AMP (left) and LHLS(H287A)–actin–ATP (right) complexes. Representative 2D class averages of the particles obtained from both datasets are shown as indicated. The orientation distributions of the particles are calculated in cryoSPARC. The number of particles for each viewing angle is shown in the heat map. The FSC curves of the final reconstructions are displayed as indicated. The average resolutions of wild-type LHLS–actin-AMP and LHLS(H287A)–actin–ATP complexes were estimated at 2.7 Å and 3.0 Å, respectively, using an FSC threshold of 0.143. b,c, Cryo-EM density maps of the wild-type LHLS–actin–AMP (b) and LHLS(H287A)–actin–ATP (c) complexes at contour levels 0.82 and 0.47. d,e, Densities of AMP in the wild-type LHLS–actin–AMP complex (d) and of ATP in the LHLS(H287A)–actin–ATP complex (e) in experimental (contour level, AMP: 0.76, ATP: 0.508) and deepEMhancer (contour level, AMP: 0.0157, ATP: 0.0377) sharpened maps. The magnesium ion was shown as a sphere in green (e). f, Cryo-EM structure of the LHLS(H287A)–actin–ATP complex. ATP is shown as sticks in grey. g, Structural comparison of the wild-type LHLS structure (cyan) determined by cryo-EM with the AlphaFold2-predicted structures of LHLS (light purple) and LnaB (grey). The C-terminal flexible regions of LHLS and LnaB are not included in structural comparison. h, Detailed interactions of ATP with the catalytic pocket of LHLS in the LHLS(H287A)–actin–ATP complex. Hydrogen bonds are indicated as blue dashed lines.

a, MS/MS analysis of the 9-amino acid tyrosine-phosphorylated peptide that was used as the substrate of LnaB in Fig. 5c. b, MS detection of the 9-amino acid tyrosine-phosphorylated peptide and the LnaB-produced ADP-Y peptide after AMPylation. Extracted ion chromatograms of the protonated peptides are shown with peak intensities, which indicate the relative amounts of the pY-containing (m/z = 559.72) and the ADP-Y-containing (m/z = 724.25) peptides. The modified tyrosine was labelled as Y^ADP in the peptide. c, MS/MS analysis of the 13-amino acid tyrosine-phosphorylated peptide derived from the FcγRII-B1 receptor that was used as the substrate of LnaB. d, MS/MS analysis of the LnaB-produced ADP-Y peptide of FcγRII-B1. e, MS detection of the 13-amino acid tyrosine-phosphorylated peptide of FcγRII-B1 and the LnaB-modified ADP-Y peptide of FcγRII-B1. f, MS/MS analysis of the 12-amino acid threonine-phosphorylated peptide SLPpTPPTREPKK of Tau, which was used as the substrate of LnaB in Fig. 5e. g, MS detection of the 12-amino acid threonine-phosphorylated peptide of Tau and the 12-amino acid ADP-T peptide after modification by LnaB.

a, MS/MS analysis of the 10-amino acid threonine-phosphorylated peptide APKpTPPSSGE of Tau, which was used as the substrate of LnaB. b, MS detection of the 10-amino acid threonine-phosphorylated peptide of Tau and the 10-amino acid ADP-T peptide after modification by LnaB. Extracted ion chromatograms of the protonated peptides are shown with peak intensities, which indicate the relative amounts of the pT-containing (m/z = 525.73) and the ADP-T (m/z = 690.26) peptides. c, MS/MS analysis of AMPylation on the 10-amino acid threonine-phosphorylated peptide of Tau by LnaB. d, MS/MS analysis of the 8-amino acid serine-phosphorylated peptide NNKGpSAAW of TAK1, which was used as the substrate of LnaB in Fig. 5f. e, MS detection of the 8-amino acid serine-phosphorylated peptide of TAK1 and the 8-amino acid ADP-S peptide after modification by LnaB. f, AMPylation of the okadaic acid- stimulated 293T cell lysates by FLAG-tagged LHEI proteins immunoprecipitated from the LHEI-overexpressed cells. The samples were analysed by SDS–PAGE and immunoblotting. *, nonspecific bands. g, Screening of AMPylation of various kinases by LnaB during infection. 293T cells expressing NPM-ALK, GFP-TBK1, 3×FLAG-IKK or 3×FLAG-AKT1 were infected with the wild-type Lp02 or ΔlnaB strain of L. pneumophila for 6 h. After immunoprecipitation, AMPylation was analysed by immunoblotting. h, MS/MS analysis of the activation loop peptide LIKDDEpYNPCQGSK of Fgr during infection of the L. pneumophila ΔlnaB strain.

a, 293T cells expressing C- terminal FLAG-tagged Fgr or Src were infected with the wild-type Lp02 or ΔlnaB strain of L. pneumophila. After immunoprecipitation with the anti-FLAG antibody, AMPylation of the kinases were analysed by SDS–PAGE and immunoblotting. b, AMPylation of an amoebic Fgr homologue by LnaB. 293T cells were co-transfected with the LnaB variants and the FLAG-tagged Fgr homologue from A. castellanii. AMPylation of the amoebic Fgr homologue was analysed by immunoblotting after anti-FLAG immunoprecipitation. c, AMPylation of endogenous Src by LnaB during infection. 293T cells were infected with the indicated L. pneumophila strains. AMPylation of Src was analysed by SDS–PAGE and western blotting after immunoprecipitation with the anti-Src antibody. d, 293T cells expressing the C-terminal HA-tagged Fgr variants were infected with the wild-type Lp02 stain of L. pneumophila. AMPylation of Fgr variants was analysed by western blotting after immunoprecipitation. e, The phosphorylation of Y416 and AMPylation of Src in 293T cells transfected with the LnaB variants were analysed by western blotting with an antibody specific for pY416 and anti-AMP antibody, respectively. f, Mass spectrometric analyses of the Y416-containing peptide (LIEDNEYTARQGAK) in p-Src and the LnaB-produced ADP-Src. g, MS/MS analysis of the AMPylation site on Src during infection. The fragment ions b7, b8, b10, b13 and y8, y10, y11, y12, y13 have a mass increase of 409 Da, corresponding to the addition of one ADP group, whereas the fragment ions b2 to b5 and y6, y7 do not exhibit this mass shift, confirming LnaB-catalysed AMPylation on pY416 of Src. h, Phosphorylation of endogenous Fgr in infected cells. U937 cells were infected with the indicated L. pneumophila strains at an MOI of 20 for 1 and 2 h. After immunoprecipitation with the anti-Fgr antibody, phosphorylation of Y412 of Fgr was analysed by western blotting with an antibody specific for pY412 (48984S, CST). i, 293T cells expressing the C-terminal HA-tagged Fgr Y523F were infected with the indicated L. pneumophila strains. Phosphorylation of Y412 and AMPylation of Fgr were analysed by western blotting with an anti-pY412 antibody (70926S, CST) and anti-AMP antibody, respectively.