Molecular Basis of Lys-63-linked Polyubiquitination Inhibition by the Interaction between Human Deubiquitinating Enzyme OTUB1 and Ubiquitin-conjugating Enzyme UBC13*

Background: A deubiquitinating enzyme OTUB1 inhibits Lys-63-linked ubiquitination by binding to a ubiquitin-conjugating enzyme UBC13. Results: A mechanism of human OTUB1-UBC13 interaction was revealed by human OTUB1-UBC13-MMS2 complex structure and structure-based mutagenesis. Conclusion: The atomic-level interactions presented by the OTUB1-UBC13-MMS2 complex structure are critical for Lys-63-linked ubiquitination inhibition. Significance: Learning how ubiquitination is regulated by the OTUB1-UBC13 interaction is crucial for understanding DNA damage response in biology. UBC13 is the only known E2 ubiquitin (Ub)-conjugating enzyme that produces Lys-63-linked Ub chain with its cofactor E2 variant UEV1a or MMS2. Lys-63-linked ubiquitination is crucial for recruitment of DNA repair and damage response molecules to sites of DNA double-strand breaks (DSBs). A deubiquitinating enzyme OTUB1 suppresses Lys-63-linked ubiquitination of chromatin surrounding DSBs by binding UBC13 to inhibit its E2 activity independently of the isopeptidase activity. OTUB1 strongly suppresses UBC13-dependent Lys-63-linked tri-Ub production, whereas it allows di-Ub production in vitro. The mechanism of this non-canonical OTUB1-mediated inhibition of ubiquitination remains to be elucidated. Furthermore, the atomic level information of the interaction between human OTUB1 and UBC13 has not been reported. Here, we determined the crystal structure of human OTUB1 in complex with human UBC13 and MMS2 at 3.15 Å resolution. The presented atomic-level interactions were confirmed by surface-plasmon resonance spectroscopy with structure-based mutagenesis. The designed OTUB1 mutants cannot inhibit Lys-63-linked Ub chain formation in vitro and histone ubiquitination and 53BP1 assembly around DSB sites in vivo. Finally, we propose a model for how capping of di-Ub by the OTUB1-UBC13-MMS2/UEV1a complex efficiently inhibits Lys-63-linked tri-Ub formation.

molecules, producing eight types of poly-Ub chains that play important roles for their specific cellular processes. For example, Lys-48-linked polyUb chains (Lys-48 chains) are the most abundant in vivo and served as the canonical signal for degradation by the proteasome (1). On the other hand, Lys-63-linked polyUb chains (Lys-63 chains) function primarily in contexts outside of targeting substrates to the proteasome (2). Lys-63 chain synthesis is solely dependent on an E2 enzyme UBC13 and its cofactor UEV1a or MMS2 (3)(4)(5).
In the course of DNA damage response induced by DNA double-strand break (DSB), Lys-63 chains signal sites of DNA damage to recruit a RAP80-BRCA1 complex and 53BP1, which play a critical role for DNA repair, cell cycle checkpoint control, and maintenance of genome stability (6 -18). DNA damage-dependent Lys-63-linked polyubiquitination is the downstream event after ATM-dependent phosphorylation of histone H2AX and MDC1 and initiated by the RNF8-dependent ubiquitination on damaged chromatin. Histone ubiquitination promotes the accumulation of a RING E3 enzyme RNF168 at DSB sites and elongates Lys-63 chains on chromatin-bound substrates that likely include histones H2A and H2AX in concert with UBC13 and UEV1a/MMS2. Defective DSB repair associated with mutations on RNF168 causes RIDDLE (radiosensitivity, immunodeficiency, dysmorphic features, and learning difficulties) syndrome because of the loss of the RNF168-dependent chromatin ubiquitination (13,19). RNF8 and RNF168 are served as positive regulators in the context of a ubiquitinationdependent process of DSB signaling and repair (11-13, 16 -18).
Typical negative regulators for ubiquitination-dependent signaling pathways are deubiquitinating enzymes (DUBs) that can remove Ub and/or Ub chains from substrates by hydrolysis (20). DNA damage-dependent ubiquitination by RNF168 and UBC13 has been found to be negatively regulated by a DUB OTUB1 (21). Surprisingly, this negative regulation, however, occurs independently of its DUB activity. Rather, OTUB1 can directly bind UBC13 and inhibit RNF168/UBC13-dependent ubiquitination in a non-canonical fashion. Furthermore, in our ubiquitination inhibition assay, we found that Lys-63-linked tri-Ub formation is strongly suppressed, whereas di-Ub formation slowly but steadily occurs. Although mechanisms for preventing di-Ub production have been recently illustrated by the crystal structures of worm OTUB1-human UBC13ϳUb and human OTUB1-UBCH5b complexes (22,23), a mechanism for suppressing tri-Ub production remains elusive. In addition, no atomic level information of the interaction between human OTUB1 and UBC13 is available because of the lack of a complex structure comprising human OTUB1 and UBC13.
In this study we determined the crystal structure of human OTUB1 in complex with human UBC13 and MMS2 at 3.15 Å resolution. The observed atomic level interactions between human OTUB1 and UBC13 were confirmed by structure-based mutagenic analyses using a surface-plasmon spectrometry, an in vitro ubiquitination assay, and an immunofluorescence assay of DNA damage-dependent responses in vivo. Finally, we propose two additional mechanisms to elucidate the inhibition of Lys-63 chains longer than di-Ub by OTUB1 based on our docking analysis of the OTUB1-UBC13-MMS2 complex and Ub chains.

Preparation of the OTUB1-⌬N-UBC13-MMS2 Complex-
The genes encoding human OTUB1-⌬N (residues 45-271) and UBC13 (residues 3-150) were amplified from the cDNAs obtained from Open Biosystems. The gene encoding human MMS2 (residues 6 -143) was PCR-amplified from a human cDNA library. The gene encoding yeast SUMO was PCR-amplified from a Saccharomyces cerevisiae cDNA library. The gene encoding SUMO was attached to the N terminus of the genes encoding OTUB1-⌬N, UBC13, and MMS2. The amplified genes were cloned into the pET26b expression vector with NdeI and XhoI sites to produce N-terminal His 6 -tagged SUMO fusion proteins and confirmed by DNA sequencing. Escherichia coli strain Rosetta TM (DE3) cells (Invitrogen) were transformed with the expression vector and cultured in LB containing 50 mg/liter kanamycin at 37°C. The expression was induced by the addition of 0.1 mM isopropyl-␤-D-thiogalactopyranoside at A 600 of ϳ0.5, and then the E. coli cells were incubated overnight at 15°C. The cells were collected by centrifugation at 8000 ϫ g for 15 min and disrupted by sonication in the sonication buffer (50 mM Tris-HCl buffer (pH 8.0) containing 150 mM NaCl, 20 mM imidazole, and 0.5% Triton X-100). The lysates were centrifuged at 30,000 ϫ g for 60 min, and the supernatants were then loaded onto a nickel-nitrilotriacetic acid (Qiagen) column that had been pre-equilibrated with the sonication buffer. The column was washed with the sonication buffer and then with 50 mM Tris-HCl buffer (pH 8.0) containing 150 mM NaCl and 20 mM imidazole. The His 6 -tagged SUMO fusion proteins were eluted with 50 mM Tris-HCl buffer (pH 8.0) containing 150 mM NaCl and 200 mM imidazole. The His 6-tagged SUMO was cleaved by Ulp1 protease, and the samples were dialyzed against 50 mM Tris-HCl buffer (pH 8.0) containing 1 mM DTT. The proteins were loaded onto a ResourceQ anion exchange column (GE Healthcare) pre-equilibrated with 50 mM Tris-HCl buffer (pH 8.0) containing 1 mM DTT and were eluted with a linear gradient of 0 -1 M NaCl. To prepare the OTUB1-⌬N-UBC13-MMS2 complex, a 1.5-fold molar excess of UBC13 was incubated at 4°C overnight with OTUB1-⌬N and MMS2. OTUB1-⌬N-UBC13-MMS2 complex was loaded onto a HiLoad Superdex 75 size-exclusion column (GE Healthcare) pre-equilibrated with 10 mM Tris-HCl buffer (pH 7.2), 50 mM NaCl, and 5 mM ␤-mercaptoethanol to remove unbound monomers.
Crystallography-The purified OTUB1-⌬N-UBC13-MMS2 complex was concentrated to 10 g/liter by using an Amicon Ultra-15 10,000 molecular weight cutoff filter (Millipore). Initial crystallization screening was performed using the sitting drop vapor diffusion method at 20°C, with a Mosquito R liquidhandling robot (TTP Lab Tech). We tested about 600 conditions using crystallization reagent kits supplied by Hampton Research and Qiagen, and initial hits were further optimized. The best crystals of the OTUB1-⌬N-UBC13-MMS2 complex were obtained at 20°C with the sitting drop vapor diffusion method by mixing 0.5 l of protein solution with an equal amount of precipitant solution containing 90 mM Tris-HCl buffer (pH 8.5), 19% PEG3350, and 10 mM EDTA sodium salt and equilibrating against 500 l of reservoir solution contain-ing 100 mM Tris-HCl buffer (pH 8.5) and 22% PEG3350. For data collection, the crystals were transferred to a cryostabilizing solution (50 mM Tris-Cl (pH 8.5), 24% PEG3350, and 20% ethylene glycol) and flash-frozen in a 100 K nitrogen stream. The diffraction data set was collected at the beamline BL41XU in SPring-8 (Hyogo, Japan) and was processed with the program HKL2000 (24) and the CCP4 program suite (25). The complex structure was determined by molecular replacement using the program MolRep (26). The crystal structures of the human OTUB1 (PDB code 2ZFY) and the human UBC13-MMS2 complex (PDB code 1J7D) were used as search models. Three OTUB1-⌬N molecules and nine UBC13-MMS2 complexes were found in the asymmetric unit. At first, the crystal apparently belonged to the space group P2 1 2 1 2 1 , with unit cell dimensions of a ϭ 102.1 Å, b ϭ 137.3 Å, c ϭ 257.2 Å. The model was built and refined using the programs COOT (27) and Refmac5 against the data processed with P2 1 2 1 2 1 . However, the R free value was not improved (around 35%), and the resultant 2F o Ϫ F c map was poor. We, therefore, suspected a pseudosymmetry and processed the data in the space group P2 1 with unit cell dimensions of a ϭ 102.1 Å, b ϭ 137.3 Å, c ϭ 257.1 Å, and ␤ ϭ 90.03°. This improved the R free values and the electron density map, and the final model consists of six OTUB1 molecules and 18 UBC13-MMS2 complexes. Refinement was carried out by using Refmac5, including temperature factor, positional, and TLS refinement, treating each protein molecule as one TLS group. The final models have excellent stereochemistry and R free values of 28.1% at 3.15 Å resolution. Data collection, phasing, and refinement statistics are shown in Table 1. All molecular graphics were prepared with the program PyMOL.
Surface Plasmon Resonance Spectrometry-GST-fused OTUB1, OTUB2, and UBC13 were purified by glutathione-Sepharose FF and ResourceQ anion exchange columns (GE Healthcare). Experiments were carried out on a Biacore T200 instrument (GE Healthcare) equilibrated at 25°C in HBS-P buffer including 10 mM HEPES (pH 7.4), 150 mM NaCl, and 0.05% surfactant P-20 using a Sensor Chip CM5 (GE Health-care). GST antibody (GE Healthcare) was covalently immobilized on the sensor chip at a density of about 15,000 resonance units (RUs), and GST-fused OTUB1, OTUB2 and UBC13 were captured on the sensor chip at a density of 500 -1000 RUs. OTUB1, UBC13, and UBC13-MMS2 were then injected for 60 s at a flow rate of 10 ml/min. Dissociation constants (K d ) were calculated by fitting to a 1:1 interaction model using Biacore T200 evaluation software (GE Healthcare).
Cell Culture and Plasmid Transfection-U2OS cells and 293T cells were grown in McCoy's 5A medium and Dulbecco's modified Eagle's medium with 10% FBS, respectively. Plasmid transfections were performed using Effectene (Qiagen) following the manufacturer's protocol with minor modifications. pcDNA3-FLAG OTUB1, pcDNA3-HA RNF168, and pCMV-MYC ubiquitin expression vectors were described previously (21). Point mutations and deletions were generated by using PCR.
Western Blotting-Cells were lysed in SDS sample buffer. Whole cell lysate were separated on SDS-PAGE (5-20% or 15% e-PAGEL, ATTO) and then transferred to nitrocellulose membranes by iBlot Gel Transfer System (Invitrogen). For Ub immunodetection, nitrocellulose membranes were incubated with 0.5% glutaraldehyde in PBS before blocking. Membranes were blocked with 5% Difco skim milk (BD Biosciences) in TBST and stained with the indicated primary antibodies followed by horseradish peroxidase (HRP)-conjugated secondary antibodies. Western Lightning ECL Pro (PerkinElmer Life Sciences) was used for chemiluminescent horseradish peroxidaseconjugated antibody detection. Chemiluminescent signals were detected with LAS4000 (Fuji Film).
Immunoprecipitation and Acid Extraction of Histones-10-cm dishes of 293T cells were grown, and each dish was transfected with 0.5 g of pcDNA3-OTUB1, 4 g of pcDNA3-HA RNF168. and 1 g of pCMV-MYC ubiquitin expression vectors using Effectene (Qiagen). After 24 h of post-transfection, cells were harvested and divided in two. One sample was lysed in ice-cold lysis buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.5% digitonin supplemented with EDTAfree protease and phosphatase inhibitor cocktails (Nacalai Tesque)). Cell lysates were sonicated and then clarified by centrifugation at 4°C for 30 min at 26,000 ϫ g. Clarified cell lysates were precleared using protein G-agarose (Sigma) and then subjected to immunoprecipitation with 20 l of anti-FLAG (M2) affinity gel (Sigma). Immunoprecipitations were carried out for 1 h at 4°C. Immunoprecipitates were washed three times with the lysis buffer followed by two washes with ice-cold PBS. The beads were then boiled in 80 l of 2ϫ SDS sample buffer. The other sample was lysed in nuclear preparation buffer (10 mM Tris-HCl (pH 7.6), 150 mM NaCl, 1.5 mM MgCl 2 , 0.65% Nonidet P-40, EDTA-free protease, and phosphatase inhibitor cocktails). Nuclei were collected by centrifugation (5 min, 920 ϫ g, 4°C). The nuclear pellet was extracted into 0.4 N H 2 SO 4 for 1 h at 4°C. The acid soluble fraction was collected after centrifugation (5 min, 7500 ϫ g, 4°C). Trichloroacetic acid was added to the supernatant to a final concentration of 20% and then incubated on ice for 1 h. Precipitates were collected by centrifugation (10 min, 15,000 ϫ g, 4°C). Precipitates were washed once with 0.1% HCl-acetone and twice with acetone and then resuspended in distilled H 2 O.
Immunofluorescence Microscopy-U2OS cells were grown on no. 1 glass coverslips (Fisher), and 0.1 g of pcDNA3 FLAG-OTUB1 expression vectors were transfected at 24-well format using Effectene (Qiagen). Cells were irradiated to 2-gray x-ray 1 day post-transfection and subjected to immunofluorescence 1 h post-IR. Cells were fixed using 3% paraformaldehyde and 2% sucrose in PBS for 15 min at room temperature and permeabilized with 0.5% Triton X-100 in PBS for 15 min at room temperature. After fixation, cells were washed with PBS 3 times and blocked with 2% BSA in PBS for 1 h. Cells were then stained with anti-FLAG and anti-53BP1 antibodies, which were diluted in 2% BSA in PBS. Cells were washed with PBS twice and then stained with Alexa Fluor 488 goat anti-mouse IgG or Alexa Fluor 555 goat anti-rabbit IgG (Invitrogen). Cells were washed with 2% BSA in PBS and PBS alone 2 times each. DNA was counterstained with DAPI (0.2 g/ml) in PBS, and samples were mounted with Prolong Gold mounting agent (Invitrogen). Confocal images were captured using an inverted microscope (TCS SP5, Leica) equipped with a 63ϫ oil immersion lens. For quantification of the immunostaining data, a minimum of 300 cells were analyzed per experiment. Images were acquired in LAS AF (Leica) format.
In Vitro Deubiquitination Assay-Assays were set up in a total volume of 20 l in 50 mM HEPES-KOH (pH 7.5), 25 mM NaCl, 5 mM MgCl 2 , 0.5 mM DTT. 1 g of Lys-48-linked tetra-Ub (Boston Biochem) was incubated with 0.9 g of recombinant OTUB1 proteins. Reaction mixtures were incubated at 37°C for 150 min.

Structure Determination of Human OTUB1-UBC13-MMS2
Ternary Complex-To reveal a mechanism of the interaction between human OTUB1 and UBC13 for the DUB activity-independent, non-canonical inhibition of ubiquitination, we tried crystallization of the full-length or N-terminal-truncated human OTUB1 (i.e. residues 45-271; referred to hereafter as OTUB1-⌬N) in complex with human UBC13 or UBC13(C87S/ K92R)ϳUb in the presence or absence of human MMS2 or UEV1a. C87S/K92R double mutation was introduced to UBC13 for preparing a stable oxyester-linked analog of the UBC13ϳUb intermediate without producing a ubiquitinated UBC13 via Lys-92. Most of the crystallization trials failed, but a trial using the N-terminal truncated OTUB1, UBC13, and MMS2 was successful. The crystal comprising the ternary complex belongs to the space group  3 ] 3 clusters that are related with each other by a pseudo 2 1 -fold symmetry along x and z axes of the unit cell. Therefore, the crystal of the ternary complex appears to belong to the space group P2 1 2 1 2 1 .
In this complicated packing of OTUB1-⌬N and the UBC13-MMS2 complex in the crystal, OTUB1-⌬N contacts four distinct UBC13-MMS2 complexes (Fig. 1a). For convenience, the UBC13-MMS2 complexes contacting ␣4, ␣8:␣9, ␣5:␣7:␣9 and the loop connecting ␣8 and ␣9 of OTUB1-⌬N are referred to as Complexes A, B, C, and D, respectively (Fig. 1a). To identify a biologically functional OTUB1-UBC13-MMS2 ternary complex, we first calculated a buried surface area at the interface between OTUB1-⌬N and UBC13 in Complex A, B, C, or D. The combination of OTUB1-⌬N and Complex C (721 Å 2 ) buries a sufficiently larger contact surface than other combinations (114, 129. and 202 Å 2 for Complex A, B, and D, respectively). Furthermore, the contact of Complex C with OTUB1-⌬N is mediated exclusively by UBC13, consistent with the fact that neither MMS2 nor UEV1a is involved in the interaction between OTUB1 and UBC13 (Fig. 1, a and b). These structural features suggest that the interactions between OTUB1-⌬N and UBC13 in Complex C should be physiologically functional ( Fig.  1 and supplemental S1).
Interactions between OTUB1 and UBC13-Interactions between OTUB1-⌬N and UBC13 in Complex C are mainly hydrophobic. The side chains of Phe-138, Val-173, Leu-177, and Met-211 in OTUB1 form a hydrophobic surface that interacts with the side chains of Pro-63, Met-64, Pro-97, and Ala-98 in UBC13 ( Fig. 1c and supplemental S2). Furthermore, the side chain of Phe-133 in OTUB1 is accommodated in a pocket formed by the aliphatic portions of the Lys-10, Glu-11, and Arg-14 side chains in UBC13. In addition, the N atom of Lys-213 in OTUB1 hydrogen bonds with the O␦ atom and main-chain CO of Asp-93 in UBC13. The O⑀ atom of Glu-209 in OTUB1 hydrogen bonds with the O␥ atom of Ser-96 in UBC13. The Ser-96 -Pro97-Ala98 tripeptide sequence is conserved among OTUB1 binding E2s (i.e. UBC13 and UBE2D and UBE2E families) (28,29) but not among other E2s (supplemental S3), suggesting that UBC13, UBE2D, and UBE2E families share a common strategy for OTUB1 binding.
To test whether these interactions actually contribute to the specific binding between OTUB1 and UBC13 in solution, we replaced Phe-133, Phe-138, Met-211, and Lys-213 by Ala in the full-length OTUB1 and measured the affinities by using surface plasmon resonance spectrometry (supplemental S4 and Table 2). As expected, these mutations reduced the affinities 33-50-fold, clearly indicating that the abovementioned interactions between OTUB1 and UBC13 occur in solution and contribute to the binding between them. Met-211 is conserved in both OTUB1 and OTUB2, and Phe-138 is perfectly conserved in OTUB1 and mostly in OTUB2 (supplemental S5). On the other hand, Phe-133 is conserved in OTUB1 but is replaced by His in OTUB2, and Lys-213 is conserved in OTUB1 but not in OTUB2. Therefore, Phe-133 and Lys-213 of OTUB1 may con-tribute to the specific binding between OTUB1 and UBC13. Our surface plasmon resonance analyses show that OTUB1 binds UBC13 with a 40-fold higher affinity than OTUB2.
Structure-based Mutational Analysis of the Inhibition of Lys-63 Chain Synthesis by OTUB1 in Vitro-We next asked if these atomic-level interactions for the binding between OTUB1 and UBC13 are critical for the inhibition of Lys-63 chain synthesis in vitro. Lys-63 chain formation was reconstituted with E1, Lys-63-only Ub (i.e. K6R/K11R/K27R/K29R/ K33R/K48R), and UBC13/MMS2 or UBC13/UEV1a in the presence or absence of the wild type or mutant OTUB1, and then products were analyzed by Western blotting for Ub. In this experiment, DUB-inactive mutation C91S was additionally introduced to OTUB1 to avoid a nonspecific DUB activity affecting the result. In the presence of C91S mutant of OTUB1, UBC13/UEV1a-or UBC13/MMS2-dependent Lys-63 chain formation was inhibited (Fig. 2, a and b, and supplemental S6,  compare lanes 2 and 3). In striking contrast, C91S/F133A, C91S/F138A, and C91S/F133A/F138A/M211A mutants of OTUB1 showed no inhibitory effect (Fig. 2, a and b, and supplemental S6, lanes 4 -6), whereas C91S/M211A mutant had a marginal inhibitory effect (Fig. 2, a and b, and supplemental S6, lane 7). Note that proper folding of F133A, F138A, M211A, and F133A/F138A/M211A mutants of OTUB1 was confirmed by their retaining full DUB activities (Fig. 3). These results indicate that the observed atomic-level interactions between OTUB1 and UBC13 are critical for the inhibition.
Structure-based Mutational Analysis of the Inhibition of DNA Damage Response by OTUB1 in Vivo-All the above data prompted us to confirm the significance of the OTUB1-UBC13 interaction through Phe-133, Phe-138, and Met-211 of OTUB1 at the cellular level. Binding between UBC13 and the wild type or mutant OTUB1 in cells was assessed by a co-immunoprecipitation experiment. As shown in Fig. 4a, the wild type FLAG-OTUB1 can be efficiently co-precipitated with UBC13. In contrast, F133A, F138A, and F133A/F138A/M211A mutants of FLAG-OTUB1 cannot be co-precipitated with UBC13, whereas M211A mutant can be fairly weakly co-precipitated. These results are consistent with the abovementioned Lys-63 ubiquitination inhibition analyses. We next analyzed the impact of F133A, F138A, and M211A mutations of OTUB1 on
Altogether, we concluded that OTUB1 inhibits E2 activity of UBC13 and the downstream UBC13/RNF168-mediated DNA damage response depending on the affinity between OTUB1 and UBC13 through specific interactions observed in the present OTUB1-⌬N-UBC13-MMS2 ternary complex structure.
OTUB1 Delays di-Ub Production and Completely Blocks Tri-Ub Production in Vitro-Our in vitro ubiquitination inhibition assay shows that Lys-63-linked di-Ub formation is inhibited in the presence of OTUB1 in a short time scale (i.e. ϳ30 min) (Fig. 2b). Furthermore, more detailed time course analysis with time points of 1, 2, 4, 6, and 10 min clearly demonstrates that di-Ub formation is sufficiently delayed in the presence of OTUB1 (Fig. 5).
Recently, the crystal structures of the UbϳUBCH5b-OTUB1-Ub and UbϳUBC13-OTUB1ϳUb complexes have been reported (22,23). The N-terminal conserved region of  OTUB1 (i.e. residues 22-46), which is required for the E2 inhibition, folds into an ␣-helix and interacts with Ile-44-centered patch of the donor Ub charged to UBCH5b and UBC13. For the inhibition of UBCH5b (and possibly, other OTUB1-interacting E2s), the edge of the helix partly shields a thioester linkage between the active site Cys of E2 and the terminal Gly of the donor Ub, thus likely interfering with a transesterification reaction to form an isopeptide linkage between the donor and acceptor Ub molecules. For UBC13 inhibition, it is likely that the N-terminal helix of OTUB1 sterically interferes with the binding between UBC13 and UEV1a/MMS2, which provides a binding site for the acceptor Ub and thus is required for Lys-63 chain formation. Interfering with the transesterification reaction and/or UEV1a/MMS2 binding by OTUB1 can explain how OTUB1 delays Lys-63-linked di-Ub production.
On the other hand, in the detailed time course analysis, we found that the amount of di-Ub gradually increased, whereas no tri-Ub could be detected. Even in a longer time scale (i.e. ϳ2 h), tri-Ub formation is completely blocked in the presence of OTUB1 despite di-Ub being sufficiently produced (Fig. 2b).
These results indicate that OTUB1 suppresses Lys-63-linked tri-Ub formation more strongly than Lys-63-linked di-Ub formation, which postulates another additional mechanism for the inhibition of Lys-63 chain synthesis for the complete blocking of tri-Ub production, as discussed below.

DISCUSSION
A molecular mechanism of Lys-63 chain synthesis by the UBC13-MMS2/UEV1a complex has been proposed on the basis of the crystal structure of yeast UBC13ϳUb-MMS2 binary complex. The UBC13-MMS2 complex forms two adjacent Ub binding sites termed "acceptor site" and "donor site." The donor Ub charged to the active site Cys by a thioester bond binds to the donor site, and then the acceptor Ub is recruited to the acceptor site. This places Lys-63 of the acceptor Ub in the active site, near the thioester bond between the donor Ub and UBC13. The N atom of Lys-63 of the acceptor Ub can then attack the thioester and form an isopeptide bond between the terminal carboxyl group of the donor Ub and Lys-63 of the acceptor Ub. After that, the donor Ub becomes the acceptor Ub for the next round of Lys-63-linked isopeptide formation, although it is not clear whether the donor Ub moves to the acceptor site on the same UBC13-MMS2/UEV1a complex or binds to the acceptor site on another complex. In either case, binding of the distal Ub moiety in the nascent di-Ub to the acceptor site of the UBC13-MMS2/UEV1a complex for tri-Ub chain formation should be a step to be potentially inhibited by OTUB1.
A docking model of the OTUB1-UBC13ϳUb(donor)-MMS2-Ub(acceptor) quaternary complex, which was built by superposing the present OTUB1-⌬N-UBC13-MMS2 complex on the reported UBC13ϳUb-MMS2 and UbϳUBCH5b-OTUB1-Ub complexes (C␣ root mean square deviation values of 2.65 and 2.60 Å over 280 and 362 residues, respectively) enabled us to propose the following models for the inhibition of Lys-63 chain synthesis by OTUB1 (Fig. 6). The OTUB1-UBC13-MMS2 ternary complex remains on the nascent di-Ub in the elongation cycle of Lys-63 chain synthesis, likely due to the enhanced affinity through the interaction with the charged Ub and behaves as a capping protein for the nascent di-Ub. This capping can sterically block the binding of the acceptor Ub at the distal terminus to the acceptor site of a new UBC13ϳUb-MMS2 complex. Additionally, the modeled N-terminal helix of