key: cord-103924-mhgnqi80
authors: Shin, Donghyuk; Bhattacharya, Anshu; Cheng, Yi-Lin; Alonso, Marta Campos; Mehdipour, Ahmad Reza; van der Heden van Noort, Gerbrand J.; Ovaa, Huib; Hummer, Gerhard; Dikic, Ivan
title: Novel class of OTU deubiquitinases regulate substrate ubiquitination upon Legionella infection
date: 2020-04-28
journal: bioRxiv
DOI: 10.1101/2020.04.25.060954
sha: 
doc_id: 103924
cord_uid: mhgnqi80

Legionella pneumophila is a gram-negative pathogenic bacterium that causes Legionaries’ disease. The Legionella genome codes more than 300 effector proteins able to modulate host-pathogen interactions during infection. Among them are also enzymes altering the host-ubiquitination system including bacterial ligases and deubiquitinases. In this study, based on homology-detection screening on 305 Legionella effector proteins, we identified two Legionella OTU-like deubiquitinases (LOT; LotB (Lpg1621/Ceg23) and LotC (Lpg2529), LotA (Lpg2248/Lem21) is already known). A crystal structure of LotC catalytic core (LotC14-310) was determined at 2.4 Å and compared with other OTU deubiquitinases, including LotB. Unlike the classical OTU-family, the structures of Legionella OTU-family (LotB and LotC) shows an extended helical lobe between the Cys-loop and the variable loop, which define a novel class of OTU-deubiquitinase. Despite structural differences in their helical lobes, both LotB and LotC interact with ubiquitin. LotB has an additional ubiquitin binding site (S1’) enabling specific cleavage of Lys63-linked poly-ubiquitin chains. By contrast, LotC only contains the S1 site and cleaves different species of ubiquitin chains. MS analysis of catalytically inactive LotB and LotC identified different categories of host-substrates for these two related DUBs. Together, our results provide new structural insights of bacterial OTU deubiquitinases and indicate distinct roles of bacterial deubiquitinases in host-pathogen interactions.

Ubiquitination, a well-studied post-translational modification system, regulates the fate of 41 various substrates by tagging them with ubiquitin (Yau and Rape, 2016 Linkage specificity of the OTU family relies on one of the following mechanisms: 1) additional 161 ubiquitin-binding domains, 2) ubiquitinated sequences in the substrates or 3) defined S1' or S2 162 ubiquitin-binding sites (Mevissen et al., 2013) . To define the minimal OTU domain for 163 biochemical and structural studies, we designed several constructs and tested their activity against 164 the di-Ub panel (Fig. 3a, b) . While LotC retained its activity with the predicted OTU domain (7-165 310), LotB lost its activity after deletion of 50 amino acids (300-350) located at the C-terminus 166 beyond the predicted OTU domain (11-283). Based on the LotB structure (PDB:6KS5, (Ma et al., 167 2020)), we assumed that this extra helical region might be required for the additional ubiquitin 168 binding site (S1') to accept the distal ubiquitin moiety from K63 Ub2 (Fig. 3c) . To understand the 169 detailed mechanism of linkage specificity of LotB and LotC at the molecular level, we determined 170 the crystal structure of the catalytic domain of LotC (LotC14-310) at 2.4 Å (Fig. 3d features in the S1 ubiquitin-binding site (Fig. 3c , d and Table 1 ). Whereas the overall fold of the 174 catalytic core of LotB and LotC resembles that of other OTU-deubiquitinases, both showed clear 175 differences in the helical arm region, which has been shown to interact with ubiquitin and it serves 176 as an S1 binding site (Mevissen et al., 2013) . The structure and sequence alignment with other OTUs clearly showed that both LotB and LotC contain a relatively long insertion between the Cys-178 loop and the variable loop, compared to other OTU members (Fig. 3e) . The typical length of the 179 helical lobe of the known OTUs is ranging from 50 to 60 amino acids (except Otubain family 180 which contain 110-120 amino acids), while LotB and LotC contain 183 and 210 amino acids, 181 respectively. Based on this observation, we wondered whether LotA, another Legionella OTU-182 deubiquitinase (Kubori et al., 2018) , also contains a longer insertion in the same region. Based on 183 the catalytic cysteine and histidine residues of the two OTU domains on LotA (Hermanns and  184 Hofmann, 2019), we analyzed the sequence and found that both OTU domains of LotA also contain 185 the longer insertion between the Cys loop and the variable loop (179 and 178 amino acids, 186 respectively; Fig. 3e ). Together, our results identify Lot-DUBs as a novel class of the OTU-family 187 with longer insertions in the helical lobe region ( Supplementary Fig.3a) . 188 189 Novel structural fold of S1-ubiquitin binding sites on Legionella OTUs 190 Both LotB and LotC have extended helices, specifically near the S1 ubiquitin binding site and we 191 wondered how these regions interact with ubiquitin. To address this, we performed ubiquitin 192 docking into both LotB and LotC, followed by molecular dynamics (MD) simulations for 600 ns 193 ( Fig. 4a- To gain better insights into the physiological roles of LotB and LotC, we decided to identify their 216 interacting proteins or substrates. First, to enrich for the interacting partners, catalytically inactive 217

LotB or LotC were expressed in cells and immuno-precipitated from cell lysates. Ubiquitin 218 (UBA52) is strongly enriched with both catalytically inactive LotB and LotC (Fig. 5a, c) . MS 219 analysis revealed that LotB mainly interacts with membrane protein complexes (COPB1, ATP5B, 220 ATP5H, COX5A, SEC61B). We also found interactions with some ER-resident proteins (Calnexin 221 (CANX), DDOST, STT3A). By contrast, most of the enriched proteins from the inactive LotC 222 pull-down were non-membrane-bound organelle-and ribosome-related proteins (RPS8, RPLP2, RPS27, RPLP1, RPL13) (Fig. 5a, c) . To further understand this, we sought to find the cellular 224 localization of both DUBs (Fig. 5b, d) . Consistent with the recent publication, LotB specifically 225 co-localized with the ER marker protein Calnexin, but not with other organelle markers (TOMM20 226 and GM130 for Mitochondria and Golgi, respectively, Fig. 5b) , and the OTU domain itself failed 227 to localize on the ER (Supplementary Fig. 4a ). By contrast, we could not find a specific cellular 228 localization of LotC (Fig. 5d) . Next, to gain more insights into the functional roles of LotB and 229

LotC, we decided to explore combinatorial ubiquitination events with other ubiquitin-related 230 LotC. Together, these findings establish important guidance on how to screen for more DUBs in 254 other pathogenic bacteria or viruses, how to characterize their physiological roles during infection. 255 We also showed that the two Legionella OTUs have different ubiquitin-binding modes 256 that enable them to cleave specific ubiquitin chains. With ubiquitin activity-based probes (Prg-, 257 VME-probes), we showed that LotB contains an extra ubiquitin-binding site (S1') and is specific 258 to K63-linked ubiquitin chains, wherease LotC cleaves different types of ubiquitin chains. 259

Interestingly, we observed a modification of LotB with NEDD8-Prg ABP. Further studies on 260 neddylated proteins with LotB will give us more insights into dual-activity of LotB. In contrast, 261 we could not see the modification between NEDD8-ABP and LotC and we reasoned that the Arg72 262 on ubiquitin, which is replaced by alanine in NEDD8, is important to locate the C-terminus of of GST-tagged protein was incubated 1 hour with glutathione-S-sepharose pre-equilibrated with 305 washing buffer (50 mM Tris-HCl pH 7.5, 500 mM NaCl, 2 mM DTT) and non-specific proteins 306

were cleared with washing. GST-proteins were eluted with elution buffer (50 mM Tris-HCl pH 307 8.0, 50 mM NaCl, 2 mM DTT, 15 mM reduced glutathione) and buffer exchanged to storage buffer 308 (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM DTT). For His-tagged proteins, the supernatant 309 was incubated with Ni-NTA pre-equilibrated with washing buffer (50 mM Tris-HCl, pH 7.5, 500 310 mM NaCl, 20 mM Imidazole) for 2 hours and eluted with elution buffer (50 mM Tris-HCl, pH 7.5, 311 500 mM NaCl, 300 mM Imidazole) and the buffer was exchanged to storage buffer. For LotC14-312 310, instead of using the elution buffer, glutathione beads were incubated with sfGFP-TEV protease 313 The program requires two protein structures as inputs, which were prepared by running the 364 refinement protocol before the docking step. We performed the local docking approach and 365 generated 100 independent structures for each complex. The complexes in this way were subject 366 to local refinement to remove remaining small clashes. The complexes were then clustered based 367 on the distance matrix of Cα atoms between the ligase and ubiquitin using the KMeans method. 368

The representatives of two major clusters in each case were selected based on the interface score 369 (I_sc), which represents the energy of the interactions across the interface of two proteins. These Tables:  445  Table1. TOP 5 Values are obtained from HHpred server (MPI Bioinformatics Toolkit) 447 Prg-ABP VME-ABP 180-S1' S1' S1' S1 S1' S1 Prg K63 S1 S1 S1' K48 S1' S1 S1' S1' S1'

S1' S1' S1 S1 Prg K63 S1 S1 LotC-host proteins Interactome  LotB-host proteins Interactome   MRPL12  NDUFS4  RPS12  SF3B5  SLC25A1   SERBP1  NPM1  SPTBN1  PRDX2  UBA52   RPS5  RPS8  RPS27  PPM1B  PCNA  NONO  DNAJA1  TUBA1B  PABPC1  PABPC4  ATP5B  SCRIB  UQCRC1  SLC3A2  COX5A  RBM10  ATP5H  COPB1  ATP5C1  PSMC5  PPIA  STK38  UBA52   RPL4  RPL5  RPL6  RPL7  RPL7A  RPL8  RPL10  RPL10A  RPL11  RPL13  RPL18  RPL18A  RPL23  RPL26  RPL27  RPL28   RPL29  RPL34  RPL35  RPS2  RPS3A  RPS5  RPS6  RPS8  RPS12  RPS13  RPS17  RPS18  RPS20  RPS25  RPS26  RPLP2  MRPL12   HSPA1B;1A  PSMD1  DNAJA1  TCP1  RAB7A  RAB11A;B  STK38L  RBM39  PSMA4  HIST1H  PSMC6  TRIM21  PHKB  VDAC2  SUN2  RIOK1  PSMC1  RPN1  UBA52   HNRNPC  NPM1  EEF1D  CCT5  RPLP1  RPLP2  YBX1  HSPA8  HSPA9  TUBB   Non-membrane-bounded  organelle   Structural constituent  of ribosome   C11orf84  EIF3I  ELMO2  ERH  KPRP  PHGDH  PHKA1  PHKA2  PHKB  PHKG2   DOCK4  EEF1A  EEF1D  EMD  GNB1  GOLGA3  HDX  HNRNPC  HNRNPH1  HNRNPK  HNRNPM  HSPA8   PSMC5  RUVBL1  RPS23  EEF1A  PSMD1  VPS33B   RPS9  EEF1D  HSP90AA1  SOD1  RPS14  CCT3  PPIA  HIST1H2B  ACTG1  GRAMD4  UBA52   ARCN1  ATP1B3  BOLA2  CANX  CCT3  CCT5  CKAP4  CLNS1A  CNOT1  COPB1  CSE1L  DDOST   ATP5A1  ATP5B  ATP5C1  ATP5H  ATP5J  ATP5O  BSG  CHCHD3  COX5A  CYB5R3  DNAJA1  HSPA1B  HSPA9  HSPD1   HSPE1  LRRC59  MYCBP  OGT  PHB  PHB2  SFXN1  SLC25A5  SLC25A6  STOML2  TUFM  UQCRC1  VDAC1   Mitochondrion   IPO8  JAK1  KPNB1  NONO  NPM1  OTUD4  PABPC1  PCNA  PGRMC1  PPIA  PPM1B  PRDX1  PRKDC  PRMT5  PSMC3  PSMC5  PSMD4  PSPC1  RAB11A  RAB7A  RIOK1  RNF219  RPN1  RPS18  RPS27  RPS5  RUVBL2  SEC61B   SLC3A2  SMN1  SNRPD2  SRRM2  STK38  STT3A  SUN2  TCP1  TFRC  TMEM33  TMPO  TNRC6B  TUBA1B  TUBB  TUBB2B 

Insights into catalysis and function of phosphoribosyl-linked serine ubiquitination

OTULIN Antagonizes LUBAC Signaling by Specifically Hydrolyzing Met1-Linked

Legionella translocates an E3 ubiquitin ligase that has 623 multiple U-boxes with distinct functions

LotA, a Legionella deubiquitinase, has dual catalytic 626 activity and contributes to intracellular growth

Distinct Deubiquitinase Class Important for Genome Stability

Crystal structures of two bacterial HECT-like E3 ligases in 633 complex with a human E2 reveal atomic details of pathogen-host interactions

A Compact Viral Processing Proteinase/Ubiquitin Hydrolase from the OTU Family

deubiquitinase Ceg23 regulates the association of Lys-63-linked polyubiquitin molecules 640 on the Legionella phagosome

Molecular basis of Lys11-polyubiquitin specificity in the deubiquitinase Cezanne

A Native Chemical Ligation Handle that

Enables the Synthesis of Advanced Activity-Based Probes: Diubiquitin as a Case Study

Polymorphic transitions in single crystals: A new molecular 655 dynamics method

Cellular quality control by the ubiquitin-proteasome system and 657 autophagy

The Molecular Basis for Ubiquitin and Ubiquitin-like Specificities in Bacterial Effector 660

Ubiquitination 662 independent of E1 and E2 enzymes by bacterial effectors

Molecular Characterization of LubX: Functional Divergence of the U-Box Fold by 666

Stop and Go Extraction Tips for Matrix-Assisted

Laser Desorption/Ionization, Nanoelectrospray, and LC/MS Sample Pretreatment in

The MEROPS 671 database of proteolytic enzymes, their substrates and inhibitors in 2017 and a comparison 672 with peptidases in the PANTHER database

Interferon-inducible antiviral effectors

RhoGDI using a family of "parallel" expression vectors

Serine Ubiquitination by Deubiquitinases DupA and DupB

Covalent inhibition of SUMO and 684 ubiquitin-specific cysteine proteases by an in situ thiol-alkyne addition

Decision-making in structure solution using Bayesian 688 estimates of map quality: the PHENIX AutoSol wizard

The 691 Perseus computational platform for comprehensive analysis of (prote)omics data

Deubiquitinase function of arterivirus papain-like

protease 2 suppresses the innate immune response in infected host cells

Evidence for Bidentate Substrate Binding as the Basis for 699 the K48 Linkage Specificity of Otubain 1

Insights into the ubiquitin transfer cascade catalyzed by the 702

CHARMM-GUI Membrane Builder toward realistic 705 biological membrane simulations

A Novel Method for High-Level Production 707 of TEV Protease by Superfolder GFP Tag

The increasing complexity of the ubiquitin code

Lupas 712 AN, Alva V

HHpred Server at its Core