Portal de Periódicos da CAPES

Structural Analysis of Xanthomonas XopD Provides Insights into Substrate Specificity of Ubiquitin-like Protein Proteases

2007; Elsevier BV; Volume: 282; Issue: 9 Linguagem: Inglês

10.1074/jbc.m608730200

1083-351X

Renée J. Chosed, D.R. Tomchick, Chad A. Brautigam, Sohini Mukherjee, Veera S. Negi, Mischa Machius, Kim Orth,

Genomics and Phylogenetic Studies

XopD (Xanthomonas outer protein D), a type III secreted effector from Xanthomonas campestris pv. vesicatoria, is a desumoylating enzyme with strict specificity for its plant small ubiquitin-like modifier (SUMO) substrates. Based on SUMO sequence alignments and peptidase assays with various plant, yeast, and mammalian SUMOs, we identified residues in SUMO that contribute to XopD/SUMO recognition. Further predictions regarding the enzyme/substrate specificity were made by solving the XopD crystal structure. By incorporating structural information with sequence alignments and enzyme assays, we were able to elucidate determinants of the rigid SUMO specificity exhibited by the Xanthomonas virulence factor XopD. XopD (Xanthomonas outer protein D), a type III secreted effector from Xanthomonas campestris pv. vesicatoria, is a desumoylating enzyme with strict specificity for its plant small ubiquitin-like modifier (SUMO) substrates. Based on SUMO sequence alignments and peptidase assays with various plant, yeast, and mammalian SUMOs, we identified residues in SUMO that contribute to XopD/SUMO recognition. Further predictions regarding the enzyme/substrate specificity were made by solving the XopD crystal structure. By incorporating structural information with sequence alignments and enzyme assays, we were able to elucidate determinants of the rigid SUMO specificity exhibited by the Xanthomonas virulence factor XopD. The small ubiquitin-like modifier (SUMO) 2The abbreviations used are: SUMO, small ubiquitin-like modifier; ULPs, ubiquitin-like protein proteases; SENPs, sentrin-specific proteases; T-SUMO, tomato SUMO; M-SUMO, mammalian SUMO; HA, hemagglutinin; RanGAP, Ran GTPase-activating protein; AtSUMO, A. thaliana SUMO; GST, glutathione S-transferase; DTT, dithiothreitol; RRL, rabbit reticulocyte lysate.2The abbreviations used are: SUMO, small ubiquitin-like modifier; ULPs, ubiquitin-like protein proteases; SENPs, sentrin-specific proteases; T-SUMO, tomato SUMO; M-SUMO, mammalian SUMO; HA, hemagglutinin; RanGAP, Ran GTPase-activating protein; AtSUMO, A. thaliana SUMO; GST, glutathione S-transferase; DTT, dithiothreitol; RRL, rabbit reticulocyte lysate. is a member of a large family of reversible post-translational modifiers. SUMO is structurally similar to ubiquitin. Like ubiquitin, SUMO utilizes a conjugation machinery (ubiquitin-activating enzyme, ubiquitin carrier protein, and ubiquitin-protein isopeptide ligase) to modify target proteins, but sumoylated proteins are not targeted for degradation by the proteasome (1Bayer P. Arndt A. Metzger S. Mahajan R. Melchior F. Jaenicke R. Becker J. J. Mol. Biol. 1998; 280: 275-286Crossref PubMed Scopus (324) Google Scholar). The conjugation machinery catalyzes the formation of a covalent isopeptide bond between the C-terminal glycine residue of SUMO and the ε-amino group of a lysine residue in the target protein. The removal of SUMO moieties from conjugated proteins by isopeptidases regenerates pools of processed SUMOs and unmodified target proteins. Yeast Ulp1 is the founding member of a growing family of these isopeptidases or desumoylating enzymes. Ubiquitin-like protein proteases (ULPs) constitute peptidase family C48 and include animals, plants, fungi, bacteria, and viruses (2Rawlings N.D. Morton F.R. Barrett A.J. Nucleic Acids Res. 2006; 34: D270-D272Crossref PubMed Scopus (468) Google Scholar). Besides being isopeptidases that remove SUMO from SUMO-conjugated target proteins, these enzymes are also proteases that cleave full-length SUMO to produce the mature form of SUMO (exposing its C-terminal Gly-Gly motif) (3Li S.J. Hochstrasser M. Nature. 1999; 398: 246-251Crossref PubMed Scopus (604) Google Scholar). Seven ULPs have been identified in humans. They are referred to as SENP1, -2, -3, -5, -6, -7, and -8 (4Melchior F. Schergaut M. Pichler A. Trends Biochem. Sci. 2003; 28: 612-618Abstract Full Text Full Text PDF PubMed Scopus (336) Google Scholar). All sentrin-specific proteases (SENPs) share the conserved C-terminal peptidase domain with yeast Ulp1, yet they differ in size, N-terminal regulatory domain sequence, substrate preference, and subcellular localization (4Melchior F. Schergaut M. Pichler A. Trends Biochem. Sci. 2003; 28: 612-618Abstract Full Text Full Text PDF PubMed Scopus (336) Google Scholar, 5Yeh E.T. Gong L. Kamitani T. Gene (Amst.). 2000; 248: 1-14Crossref PubMed Scopus (415) Google Scholar). SENP1 and SENP2 have been shown to process mammalian SUMO-1, -2, and -3 conjugates (6Gong L. Millas S. Maul G.G. Yeh E.T. J. Biol. Chem. 2000; 275: 3355-3359Abstract Full Text Full Text PDF PubMed Scopus (234) Google Scholar, 7Zhang H. Saitoh H. Matunis M.J. Mol. Cell. Biol. 2002; 22: 6498-6508Crossref PubMed Scopus (229) Google Scholar). SENP3 is localized to the nucleolus during interphase and is known to process mammalian SUMO-2 and SUMO-3 (8Nishida T. Tanaka H. Yasuda H. Eur. J. Biochem. 2000; 267: 6423-6427Crossref PubMed Scopus (143) Google Scholar). SENP5 exhibits isopeptidase activity for SUMO-2- and SUMO-3-modified proteins (9Gong L. Yeh E.T. J. Biol. Chem. 2006; 281: 15869-15877Abstract Full Text Full Text PDF PubMed Scopus (191) Google Scholar). SENP6 is found in the cytoplasm and is considered to be an Ulp2-like enzyme, i.e. it is thought to exhibit mainly peptidase activity as opposed to isopeptidase activity (10Kim K.I. Baek S.H. Jeon Y.J. Nishimori S. Suzuki T. Uchida S. Shimbara N. Saitoh H. Tanaka K. Chung C.H. J. Biol. Chem. 2000; 275: 14102-14106Abstract Full Text Full Text PDF PubMed Scopus (136) Google Scholar). SENP7 is not characterized in the literature to date. Unlike the other characterized SENPs, SENP8 removes Nedd8 (another ubiquitin-like protein) from target proteins (11Gan-Erdene T. Nagamalleswari K. Yin L. Wu K. Pan Z.Q. Wilkinson K.D. J. Biol. Chem. 2003; 278: 28892-28900Abstract Full Text Full Text PDF PubMed Scopus (157) Google Scholar). So far, five ULP1 genes have been identified in Arabidopsis thaliana. They are referred to as ULP1A, -B, -C, and -D, and ESD4, respectively (12Kurepa J. Walker J.M. Smalle J. Gosink M.M. Davis S.J. Durham T.L. Sung D.Y. Vierstra R.D. J. Biol. Chem. 2003; 278: 6862-6872Abstract Full Text Full Text PDF PubMed Scopus (361) Google Scholar, 13Murtas G. Reeves P.H. Fu Y.F. Bancroft I. Dean C. Coupland G. Plant Cell. 2003; 15: 2308-2319Crossref PubMed Scopus (178) Google Scholar). A. thaliana ESD4 (early in short days 4) was shown to act as a SUMO protease and is potentially involved in regulating time of flowering (13Murtas G. Reeves P.H. Fu Y.F. Bancroft I. Dean C. Coupland G. Plant Cell. 2003; 15: 2308-2319Crossref PubMed Scopus (178) Google Scholar). Recently, we characterized four ULPs from A. thaliana and showed that these proteases exhibit substrate specificity both for the processing of SUMO and for the cleavage of SUMO conjugates (14Chosed R. Mukherjee S. Lois L.M. Orth K. Biochem. J. 2006; 398: 521-529Crossref PubMed Scopus (54) Google Scholar). Another ULP, XopD (Xanthomonas outer protein D), is expressed as a secreted virulence factor by Xanthomonas campestris pv. vesicatoria (15Hotson A. Chosed R. Shu H. Orth K. Mudgett M.B. Mol. Microbiol. 2003; 50: 377-389Crossref PubMed Scopus (218) Google Scholar). This pathogen is the causative agent of bacterial spot disease in tomato and pepper plants (16Jones J.B. Stall R.E. Bouzar H. Annu. Rev. Phytopathol. 1998; 36: 41-58Crossref PubMed Scopus (141) Google Scholar). Upon contact, Xanthomonas employs an hrp- (hypersensitive response and pathogenicity) type 3 secretion system to inject effector proteins into host plant cells (17Alfano J.R. Collmer A. Annu. Rev. Phytopathol. 2004; 42: 385-414Crossref PubMed Scopus (601) Google Scholar, 18Ghosh P. Microbiol. Mol. Biol. Rev. 2004; 68: 771-795Crossref PubMed Scopus (325) Google Scholar). There are 20-25 different proteins that make up a typical type 3 secretion system apparatus, which forms a syringe-like structure projecting across the inner bacterial membrane, the peptidoglycan layer, the outer bacterial membrane, and the host cell wall and plasma membrane to reach the host cell cytoplasm (18Ghosh P. Microbiol. Mol. Biol. Rev. 2004; 68: 771-795Crossref PubMed Scopus (325) Google Scholar, 19Mudgett M.B. Annu. Rev. Plant Biol. 2004; 56: 509-531Crossref Scopus (244) Google Scholar). The effector proteins are transported through this syringe-like channel in an unfolded state to reach the host cell interior (18Ghosh P. Microbiol. Mol. Biol. Rev. 2004; 68: 771-795Crossref PubMed Scopus (325) Google Scholar), where they interact with specific targets to cause plant cell death (17Alfano J.R. Collmer A. Annu. Rev. Phytopathol. 2004; 42: 385-414Crossref PubMed Scopus (601) Google Scholar). In planta, XopD exhibits an isopeptidase activity that reduces the amount of SUMO protein conjugates (15Hotson A. Chosed R. Shu H. Orth K. Mudgett M.B. Mol. Microbiol. 2003; 50: 377-389Crossref PubMed Scopus (218) Google Scholar). In vitro, XopD exhibits peptidase and isopeptidase activities (15Hotson A. Chosed R. Shu H. Orth K. Mudgett M.B. Mol. Microbiol. 2003; 50: 377-389Crossref PubMed Scopus (218) Google Scholar). XopD appears to be plant-specific as evidenced by the fact that, in vitro, it is able to cleave and activate tomato SUMO (T-SUMO), but not mammalian SUMO (M-SUMO). Likewise, XopD can process plant sumoylated proteins, but not mammalian sumoylated proteins (15Hotson A. Chosed R. Shu H. Orth K. Mudgett M.B. Mol. Microbiol. 2003; 50: 377-389Crossref PubMed Scopus (218) Google Scholar). The activity of XopD is inhibited by mutation of the catalytic cysteine to alanine or serine or by treatment with the chemical inhibitors N-ethylmaleimide and iodoacetamide (15Hotson A. Chosed R. Shu H. Orth K. Mudgett M.B. Mol. Microbiol. 2003; 50: 377-389Crossref PubMed Scopus (218) Google Scholar). AvrXv4, another enzyme from Xanthomonas, was also shown to desumoylate in planta (20Roden J. Eardley L. Hotson A. Cao Y. Mudgett M.B. Mol. Plant-Microbe Interact. 2004; 17: 633-643Crossref PubMed Scopus (147) Google Scholar). Because bacteria do not encode either ubiquitin or SUMO signaling machineries, Xanthomonas appears to have usurped the activity of eukaryotic ULPs to aid in pathogenesis by disrupting host defense signaling in the infected plant cell (15Hotson A. Chosed R. Shu H. Orth K. Mudgett M.B. Mol. Microbiol. 2003; 50: 377-389Crossref PubMed Scopus (218) Google Scholar, 21Orth K. Curr. Opin. Microbiol. 2002; 5: 38-43Crossref PubMed Scopus (156) Google Scholar, 22Cornelis G.R. Nat. Rev. Mol. Cell Biol. 2002; 3: 742-752Crossref PubMed Scopus (347) Google Scholar). Unlike yeast Ulp1, which is promiscuous in its choice of SUMO substrate, XopD demonstrates a rigid specificity (15Hotson A. Chosed R. Shu H. Orth K. Mudgett M.B. Mol. Microbiol. 2003; 50: 377-389Crossref PubMed Scopus (218) Google Scholar). By comparing the sequences of several SUMOs from plants, human, and yeast, one can predict which residues in SUMO might dictate specificity. To gain further insight into the SUMO specificity of XopD, we determined the crystal structures of the catalytic core of XopD and the inactive C470A mutant. On the basis of the structural comparison of XopD with yeast Ulp1 and the analysis of the region surrounding the structurally conserved His-Asp-Cys catalytic triad, we identified residues that we believed to be responsible for the differing specificity of the two enzymes. On the basis of this information, we mutated yeast SUMO (Smt3), creating a substrate that could be cleaved by XopD, and similarly mutated T-SUMO, creating a substrate that could no longer be cleaved by XopD. Thus, using structural, enzymatic and bioinformatic analyses, we were able to identify the residues that dictate the specificity of the plant virulence factor XopD for its eukaryotic substrate plant SUMO. Cloning and Construction of Plasmids—XopD, T-SUMO, yeast ΔUlp1 (Ulp1-(403-621)), and hemagglutinin (HA)-tagged Ran GTPase-activating protein (RanGAP) were prepared as described by Hotson et al. (15Hotson A. Chosed R. Shu H. Orth K. Mudgett M.B. Mol. Microbiol. 2003; 50: 377-389Crossref PubMed Scopus (218) Google Scholar). A. thaliana SUMO (AtSUMO)-1, -2, -3, and -5; Smt3; and M-SUMO-1, -2, and -4 were cloned as described previously (14Chosed R. Mukherjee S. Lois L.M. Orth K. Biochem. J. 2006; 398: 521-529Crossref PubMed Scopus (54) Google Scholar). To generate glutathione S-transferase (GST)-tagged XopD constructs, XopD-(1-545), XopD-(285-545), and XopD-(335-520) were cloned into pGEX-rTEV (see supplemental Table S1 for details) (23Chook Y.M. Blobel G. Nature. 1999; 399: 230-237Crossref PubMed Scopus (291) Google Scholar). The N-terminal GST tag can be cleaved with tobacco etch virus protease or thrombin. The GST-tobacco etch virus protease expression vector was a gift from Yuh Min Chook. Smt3-HA was cloned from Saccharomyces cerevisiae cDNA. T-SUMO-HA was constructed by using pT7-LO (a gift from J. Clemens), and Smt3-HA was constructed by using pET15b (Novagen). These constructs contained an N-terminal HA tag directly following the Gly-Gly motif. To generate GST-SUMO-Gly-Gly-STOP constructs to be used in in vitro sumoylation assays, AtSUMO-1, -2, and -3; T-SUMO; Smt3; and M-SUMO-1, -2, and -4 were cloned into pGEX-rTEV (23Chook Y.M. Blobel G. Nature. 1999; 399: 230-237Crossref PubMed Scopus (291) Google Scholar). These SUMOs were constructed with an N-terminal GST tag and a stop codon after the Gly-Gly motif (see Ref. 14Chosed R. Mukherjee S. Lois L.M. Orth K. Biochem. J. 2006; 398: 521-529Crossref PubMed Scopus (54) Google Scholar for details). To generate His6-SUMO-GG-HA constructs to be used in in vitro peptidase assays, AtSUMO-1, -2, and -3; T-SUMO; Smt3; and M-SUMO-1, -2, and -4 were cloned into pET15b, and AtSUMO-5 was cloned into pT7-LOH (gift from J. Clemens). These SUMOs were constructed with an N-terminal His6 tag, and their C termini contained an HA tag directly following the Gly-Gly motif (see Ref. 14Chosed R. Mukherjee S. Lois L.M. Orth K. Biochem. J. 2006; 398: 521-529Crossref PubMed Scopus (54) Google Scholar for details). The catalytic C470A mutant of XopD and mutants of T-SUMO-HA and Smt3-HA were generated using the QuikChange™ site-directed mutagenesis kit (Stratagene). All mutants were confirmed by DNA sequence analysis (supplemental Table S1). Protein Expression and Purification—GST-tagged XopD constructs, GST-tagged SUMO constructs, and GST-tagged yeast Ulp1 were expressed in Escherichia coli BL21(DE3) cells and purified by GST affinity chromatography (24Zhou G. Bao Z.Q. Dixon J.E. J. Biol. Chem. 1995; 270: 12665-12669Abstract Full Text Full Text PDF PubMed Scopus (537) Google Scholar). Briefly, cells were grown to A600 = 0.6-0.8 in 2× yeast extract and tryptone medium and then induced with 400 μm isopropyl β-d-thioga-lactopyranoside (Roche Applied Science) for 4 h at 30 °C. The cells were lysed in phosphate-buffered saline (pH 8), 1% (v/v) Triton X-100 (Fisher), 0.1% (v/v) 2-mercaptoethanol (Bio-Rad), and 1 mm phenylmethylsulfonyl fluoride (Sigma) using a cell disrupter (EmulsiFlex-C5, Avestin Inc.). The lysates were incubated with glutathione-agarose beads, and bound protein was eluted with elution buffer (10 mm reduced glutathione, 50 mm Tris (pH 8), and 150 mm NaCl) as described previously (24Zhou G. Bao Z.Q. Dixon J.E. J. Biol. Chem. 1995; 270: 12665-12669Abstract Full Text Full Text PDF PubMed Scopus (537) Google Scholar). In Vitro Peptidase Assays—AtSUMO-1-HA, AtSUMO-2-HA, AtSUMO-3-HA, AtSUMO-5-HA, M-SUMO-1-HA, M-SUMO-2-HA, M-SUMO-4-HA, Smt3-HA (and all Smt3-HA mutants), T-SUMO-HA (and all T-SUMO-HA mutants), and mammalian HA-RanGAP were translated in vitro using the TnT coupled rabbit reticulocyte lysate system (Promega Corp.) with l-[35S]methionine (Amersham Biosciences). For each in vitro peptidase assay, 2 μl of the 35S-labeled translation reaction mixture was added to 18 μl of 0.5 mg/ml purified XopD for 1 h at 30 °C. Elution buffer without glutathione was used as a control. The mixtures were resolved by SDS-PAGE, and the gels were incubated with Amplify fluorographic reagent (Amersham Biosciences) and analyzed by autoradiography. T-SUMO-His6 (3.9 nmol) was incubated with submolar amounts of ΔXopD (0.13, 0.025, and 0.013 nmol) for 0, 1, 5, 10, 30, and 60 min at 30 °C. Reactions were terminated by addition of SDS sample buffer, resolved by SDS-PAGE, and analyzed by Coomassie Blue staining. In Vitro Sumoylation Assays—In vitro sumoylation of 35S-labeled mammalian HA-RanGAP with GST-AtSUMO-1, GST-AtSUMO-2, GST-AtSUMO-3, GST-AtSUMO-5, GST-T-SUMO, GST-Smt3, GST-M-SUMO-1, GST-M-SUMO-2, and GST-M-SUMO-4 was performed as described previously (14Chosed R. Mukherjee S. Lois L.M. Orth K. Biochem. J. 2006; 398: 521-529Crossref PubMed Scopus (54) Google Scholar, 25Desterro J.M. Rodriguez M.S. Hay R.T. Mol. Cell. 1998; 2: 233-239Abstract Full Text Full Text PDF PubMed Scopus (911) Google Scholar). In vitro peptidase assays using these sumoylated RanGAP proteins were performed as described above. Protein Expression and Purification for Crystallization Experiments—GST-XopD-(335-520) and GST-XopD-(335-520)(C470A) were expressed and purified as described above. After buffer exchange with 50 mm Tris-HCl (pH 8.0), 150 mm NaCl, and 0.1% (v/v) 2-mercaptoethanol by ultrafiltration, the GST moiety was removed by overnight incubation with thrombin (Novagen), followed by glutathione-agarose column purification. The protein was then concentrated using an Amicon ultrafiltration device, and the buffer was exchanged with 20 mm Tris-HCl (pH 7.5), 0.1 m KCl, and 1 mm dithiothreitol (DTT). The concentrated protein was subjected to ion-exchange chromatography (HiTrap Mono Q, Amersham Biosciences) and eluted with a linear gradient of 0.1-1.0 m KCl in 20 mm Tris-HCl (pH 7.5) and 1 mm DTT using anÁKTA FPLC system (Amersham Biosciences). At this stage, the protein was either used for initial crystallization screening or further purified for crystal optimization. For this purpose, the protein was subjected to size-exclusion chromatography (HiLoad 16/60 Superdex 75, Amersham Biosciences) in 20 mm Tris-HCl (pH 7.5), 75 mm KCl, and 0.5 mm DTT. The protein was then concentrated to 15 mg/ml for crystallization experiments. For the production of selenomethionine-labeled protein, the XopD expression vector was transformed into the methionine auxotroph E. coli strain B834 (Novagen). The cells were grown in M9 medium supplemented with 125 mg/liter each adenine, uracil, thymine, and guanosine nucleotide; 2.5 mg/liter thiamin; 4 mg/liter d-biotin; 20 mm glucose; 2 mm magnesium sulfate; and 50 mg/liter l-selenomethionine (Sigma). Selenomethionine-labeled XopD protein was purified as described above. All purified proteins were analyzed by SDS-PAGE and quantified using a modified version (26Bailey J.L. Techniques in Protein Chemistry. Elsevier, New York1967Google Scholar) of the Lowry procedure. Protein Crystallization—Wild-type XopD-(335-520) and XopD-(335-520)(C470A) crystals were grown by vapor diffusion in hanging-drop mode at 20 °C. Initial screening was performed with the Index crystallization kit (Hampton Research), yielding one hit. Optimized crystallization conditions used drops of 1 μl of highly purified XopD (15 mg/ml in 20 mm Tris-HCl (pH 7.5), 75 mm KCl, and 0.5 mm DTT) plus 1 μl of reservoir solution (1.4-1.6 m sodium potassium phosphate (pH 7.5-7.75)) suspended over 0.5 ml of reservoir solution. Crystals typically appeared in 2-7 days. Selenomethionine crystals were obtained under similar conditions that included 5 mm DTT or 10 mm 2-mercaptoethanol. Prior to x-ray diffraction experiments, crystals were cryoprotected by transfer into 1.6 m sodium/potassium phosphate (pH 7.75) supplemented with 15% (v/v) ethylene glycol. Crystals were flash-cooled in liquid propane and then stored in liquid nitrogen. Data Collection, Structure Determination, and Refinement—Diffraction data were collected at 100 K at beamline 19-BM of the Argonne National Laboratory Structural Biology Center at the Advanced Photon Source. Data sets were indexed, integrated, and scaled using the HKL2000 program package (27Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref PubMed Scopus (38526) Google Scholar). XopD-(335-520) crystals exhibited the symmetry of space group P41212 with unit cell parameters a = b = 92 Ä and c = 45 Ä, diffracted x-rays to a minimum Bragg spacing (dmin) of 1.8 Ä, and contained one molecule in the asymmetric unit (35% solvent). Phases were obtained from a crystal of the selenomethionyl-substituted protein by the single-wavelength anomalous diffraction method using x-rays with an energy near the selenium K-absorption edge. Location of heavy atom sites, phase calculation, and refinement were performed with the program Solve (28Terwilliger T.C. Berendzen J. Acta Crystallogr. Sect. D Biol. Crystallogr. 1999; 55: 849-861Crossref PubMed Scopus (3219) Google Scholar). Density modification was carried out with the program dm (29Cowtan K.D. Main P. Acta Crystallogr. Sect. D Biol. Crystallogr. 1996; 52: 43-48Crossref PubMed Scopus (288) Google Scholar). An initial model was constructed automatically using the program ARP/wARP (30Perrakis A. Morris R. Lamzin V.S. Nat. Struct. Biol. 1999; 6: 458-463Crossref PubMed Scopus (2563) Google Scholar) and contained 150 of 186 residues. Further model building was performed manually with the program O (31Jones T.A. Zou J.Y. Cowan S.W. Kjeldgaard Acta Crystallogr. Sect. A. 1991; 47: 110-119Crossref PubMed Scopus (13009) Google Scholar). This model served as the starting model for the refinement of the native XopD crystal structure. Refinement included simulated annealing, conjugate gradient minimization, and calculation of individual B-factors in the program CNS (Version 1.1) (32Brunger A.T. Adams P.D. Clore G.M. DeLano W.L. Gros P. Grosse-Kunstleve R.W. Jiang J.S. Kuszewski J. Nilges M. Pannu N.S. Read R.J. Rice L.M. Simonson T. Warren G.L. Acta Crystallogr. Sect. D Biol. Crystallogr. 1998; 54: 905-921Crossref PubMed Scopus (16957) Google Scholar). Water molecules were added where stereochemically reasonable using the program CNS. Final model refinement was performed in REFMAC (Version 5.2) (33Murshudov G.N. Vagin A.A. Dodson E.J. Acta Crystallogr. Sect. D Biol. Crystallogr. 1997; 53: 240-255Crossref PubMed Scopus (13853) Google Scholar) and included TLS refinement. The model that was fit to data obtained from wild-type XopD crystals included all residues except residues 335-337, 405-407, 431-433, 465-468, and 515-520 as well as 88 water molecules. XopD-(335-520)(C470A) crystals were isomorphous to the wild-type crystals; model building and refinement followed a protocol similar to that used for the wild-type structure. The model for the XopD-(335-520)(C470A) protein included all residues except residues 335, 466-467, and 514-520 as well as 96 water molecules. Data collection, phasing, and refinement statistics are detailed in Table 1.TABLE 1Data collection, structure determination, and refinementData collectionCrystalNativeSeMetaBijvoet pairs were kept separate for data processing.MutantSpace groupP41212P41212P41212Unit cell parameters (Ä)a = b = 91.618, c = 44.820a = b = 91.790, c = 45.487a = b = 91.953, c = 45.013Energy (eV)12,660.6312,648.9212,663.73Resolution range (Ä)32.04-1.95 (2.0-1.95)41.05-1.85 (1.89-1.85)25.3-1.80 (1.84-1.80)Unique reflections14,28217,00917,346Multiplicity7.6 (7.5)14.6 (10.6)7.6 (3.1)Data completeness (%)98.6 (99.9)98.9 (91.9)94.2 (77.4)Rmerge (%)bRmerge = 100 ∑h∑i|Ih,i - 〈Ih 〉|/∑h∑iIh,i, where the outer sum (h) is over the unique reflections and the inner sum (i) is over the set of independent observations of each unique reflection.6.1 (71.0)9.0 (76.8)7.5 (58.0)I/σ(I)31.2 (2.25)36.2 (2.55)30.6 (1.95)Wilson B-value (Ä2)37.823.430.5Phase determinationAnomalous scatterSelenium (3 of 5 possible sites)Figure of merit (35.0-1.85 Ä)0.218Refinement statisticsCrystalNativeMutantResolution range (Ä)32.0-1.95 (2.00-1.95)25.3-1.80 (1.86-1.80)No. of reflections (Rwork/Rfree)13,554/720 (982/49)17,318/1,346 (514/37)Atoms (non-H protein/solvent)1337/881396/96Rwork (%)23.8 (41.9)23.6 (38.57)Rfree (%)28.7 (48.6)25.8 (35.36)r.m.s.d. bond length (Ä)0.0170.012r.m.s.d. bond angle1.46°1.57°Mean B-value (Ä2)36.638.2Missing residues335-337, 405-407, 431-433, 465-468, 515-520335, 466-467, 514-520a Bijvoet pairs were kept separate for data processing.b Rmerge = 100 ∑h∑i|Ih,i - 〈Ih 〉|/∑h∑iIh,i, where the outer sum (h) is over the unique reflections and the inner sum (i) is over the set of independent observations of each unique reflection. Open table in a new tab Structure Figures—All figures of XopD and yeast Ulp1 structures were constructed using the MacPyMOL program (34DeLano W.L. The PyMOL Molecular Graphics System. DeLano Scientific, San Carlos, CA2002Google Scholar). Substrate Specificity of XopD—Previous results showing that XopD can process T-SUMO, but not M-SUMO-1, led to the hypothesis that XopD exhibits species specificity for its SUMO substrates (15Hotson A. Chosed R. Shu H. Orth K. Mudgett M.B. Mol. Microbiol. 2003; 50: 377-389Crossref PubMed Scopus (218) Google Scholar). Recently, we reported the characterization of several SUMOs and ULPs from A. thaliana and demonstrated that ULP1 shows varying specificity for AtSUMO-1, -2, -3, and -5 as well as for M-SUMO-1, -2, and -4 and yeast SUMO (Smt3) (14Chosed R. Mukherjee S. Lois L.M. Orth K. Biochem. J. 2006; 398: 521-529Crossref PubMed Scopus (54) Google Scholar). Sequence comparisons of these SUMOs and T-SUMO revealed a striking similarity (83% sequence identity) among AtSUMO-1, AtSUMO-2, and T-SUMO (supplemental Fig. S1). On the basis of this high sequence similarity, we hypothesized that XopD would be able to process both AtSUMO-1 and AtSUMO-2, but not AtSUMO-3 and AtSUMO-5, which show little sequence identity to T-SUMO, Smt3, or M-SUMOs. To explore this possibility, AtSUMO-1, -2, -3, and -5; M-SUMO-1, -2, and -4; and yeast Smt3 were used as substrates for in vitro peptidase assays (Fig. 1A). All SUMOs used in this peptidase assay were constructed with an HA tag at their C termini directly following their C-terminal Gly-Gly residues (SUMO-Gly-Gly-HA). Peptidase activity of the enzymes can be detected by their ability to remove the HA tag from the SUMOs, resulting in a cleavage product that migrates faster during SDS-PAGE. Both full-length XopD (XopD-(1-545)) and only the peptidase domain of XopD (XopD-(285-545)) were tested along with the peptidase domain (amino acids 403-621) of yeast Ulp1 (ΔUlp1) for their ability to process the panel of SUMO substrates. In all peptidase assays, molar excesses of enzyme to SUMO substrate were used to ensure cleavage of the substrates tested. XopD preferentially recognized T-SUMO and, as predicted, AtSUMO-1 and AtSUMO-2 (Fig. 1B). To demonstrate catalytic processing of substrate by XopD, we performed time course peptidase assays in which recombinant T-SUMO-His6 was incubated with submolar amounts of recombinant XopD-(285-545) over time. Supplemental Fig. S2 shows that T-SUMO-His6 was processed to completion within 10 min using 1:10 molar eq of XopD-(285-545) to T-SUMO. Using 1:100 molar eq of enzyme to substrate showed no product formation at 10 min, but close to 100% conversion at 1 h. To further understand how XopD recognizes its SUMO substrates, the XopD isopeptidase activity for various SUMO substrates was assessed. The isopeptidase activity of XopD was determined using an in vitro sumoylation system with each SUMO expressed as a GST fusion protein and in vitro translated 35S-RanGAP as the substrate (Fig. 2A). In this assay, a fraction of the translated 35S-RanGAP was modified by the endogenous rabbit reticulocyte lysate (RRL) sumoylation machinery with RRL SUMO. Purified recombinant GST-tagged AtSUMO-1, -2, -3, and -5; M-SUMO-1, -2, and -4; Smt3; and T-SUMO were conjugated to a fraction of the unmodified 35S-RanGAP and appeared as higher molecular mass bands on an SDS-polyacrylamide gel (Fig. 2B). Using this in vitro sumoylation system, the substrate specificity of the isopeptidase activity of XopD was consistent with its peptidase activity. XopD could process RanGAP that had been modified by GST-T-SUMO, GST-AtSUMO-1, and GST-AtSUMO-2, but not Ran-GAP that had been modified by GST-AtSUMO-3, GST-At- SUMO-5, the GST-M-SUMOs, or GST-Smt3 (Fig. 2B). As an internal control, each reaction contained RRL SUMO-Ran-GAP. Yeast ΔUlp1, but not XopD, was able to cleave this conjugated RanGAP (Fig. 2B). Thus, XopD can process T-SUMO, AtSUMO-1, and AtSUMO-2 in peptidase assays as well as in isopeptidase assays. Recently, Colby et al. (35Colby T. Matthai A. Boeckelmann A. Stuible H.P. Plant Physiol. 2006; 142: 318-332Crossref PubMed Scopus (144) Google Scholar) reported that XopD can process AtSUMO-3 conjugated to yeast proliferating cell nuclear antigen. Although we observed that XopD did not process AtSUMO-3 conjugated to RanGAP, we have shown previously that the regulatory domain of an ULP can change the activity of an enzyme (14Chosed R. Mukherjee S. Lois L.M. Orth K. Biochem. J. 2006; 398: 521-529Crossref PubMed Scopus (54) Google Scholar). In this case, it may allow XopD to recognize a SUMO substrate associated with different target proteins. On the basis of these data and sequence comparisons, we can recognize residues that appear to be important for the specificity of XopD for various SUMOs (Fig. 3). Using the nomenclature of Schechter and Berger (37Schechter I. Berger A. Biochem. Biophys. Res. Commun. 1967; 27: 157-162Crossref PubMed Scopus (4755) Google Scholar) for describing amino acids of substrates and proteases, we observed that residues at the P5-P7 positions (Met-Leu-His) are absolutely conserved in SUMO substrates cleaved by XopD in peptidase and isopeptidase reactions. T-