A Ubiquitin Shuttle DC-UbP/UBTD2 Reconciles Protein Ubiquitination and Deubiquitination via Linking UbE1 and USP5 Enzymes

The ubiquitination levels of protein substrates in eukaryotic cells are delicately orchestrated by various protein cofactors and enzymes. Dendritic cell-derived ubiquitin (Ub)-like protein (DC-UbP), also named as Ub domain-containing protein 2 (UBTD2), is a potential Ub shuttle protein comprised of a Ub-like (UbL) domain and a Ub-binding domain (UBD), but its biological function remains largely unknown. We identified two Ub-related enzymes, the deubiquitinating enzyme USP5 and the Ub-activating enzyme UbE1, as interacting partners of DC-UbP from HEK 293T cells. Biochemical studies revealed that the tandem UBA domains of USP5 and the C-terminal Ub-fold domain (UFD) of UbE1 directly interacted with the C-terminal UbL domain of DC-UbP but on the distinct surfaces. Overexpression of DC-UbP in HEK 293T cells enhanced the association of these two enzymes and thus prompted cellular ubiquitination, whereas knockdown of the protein reduced the cellular ubiquitination level. Together, DC-UbP may integrate the functions of USP5 and UbE1 through interacting with them, and thus reconcile the cellular ubiquitination and deubiquitination processes.


Introduction
Ubiquitination is one of the common post-translational modifications of proteins in eukaryotic organisms [1]. By working as a versatile regulatory signal controlling protein stability, cellular localization and biological function, ubiquitination plays very important roles in gene regulation, cell cycle, cellular protein level, cell signaling and so on [2,3,4]. In these processes, ubiquitin is covalently attached to a target protein with the cascade participation of three enzymes, Ub-activating enzyme E1 (UbE1), Ub-conjugating enzyme E2 (UbE2) and Ub E3 ligase (UbE3) [1,5]. UbE1 (Uba1 in yeast) is a unique enzyme that universally activates Ub molecules for conjugating to a UbE2 and transferring to substrates assisted by one of the numerous UbE3 ligases.
Ubiquitination is regulated in multi-levels and aspects, and most importantly, this process can be reversed by deubiquitinating enzymes (DUB). DUB selectively removes Ub or edits the length or type of Ub chain on substrate [6]. There are five families of DUBs in eukaryotes, which may have different locations, targets or mechanisms, and their activities and specificities on substrates are extremely diverse [7,8]. The largest group ubiquitin-specific protease (USP) contains a catalytic domain usually consisting of a Cys box and a His box [9]. USP typically cleaves Ub conjugates from ubiquitinated protein substrates or unanchored Ub chains.
It is generally considered that the ubiquitination levels of protein substrates in cells are highly orchestrated with various protein cofactors [10,11], including shuttle factors like Rad23 and Dsk2. Dendritic cell-derived ubiquitin-like protein (DC-UbP), also named as Ub domain-containing protein 2 (UBTD2), is a novel Ub domain-containing protein firstly identified in dendritic cells and implicated in ubiquitination pathway [12]. Our previous work has elucidated the solution structure of the C-terminal part of DC-UbP (UbP_C), indicating that UbP_C is structurally comprised of a typical Ub-like (UbL) fold but lacks the conserved diglycine tail necessary to Ub modification [13]. The UbL structure also displays a positively-charged surface distinct from Ub molecule, suggesting that the UbL domain of DC-UbP may have its unique interacting partner and cellular function. We also solved the novel structure of the N-terminal part of DC-UbP (UbP_N) and found that it is potentially a Ub-binding domain (UBD) [14]. More importantly, the DC-UbP protein is a combination of UbL and UBD domains, which increase the possibility for DC-UbP to be involved in the ubiquitination process or other relevant pathways [15]. However, the detailed biological function of DC-UbP and its underlying mechanism are still unclear.
To unravel the biological function of DC-UbP in protein ubiquitination and delivery of ubiquitinated substrates, we firstly performed pull-down experiments to characterize its potential interacting partners that led to identify two enzymes, UbE1 and USP5, which function cooperatively in protein ubiquitination and deubiquitination. Then we confirmed their interactions in vitro and in cell model by biochemical methods. DC-UbP may play a role in mediating association of UbE1 and USP5 and thus modulating the ubiquitination levels of protein substrates in cells. Finally, a schematic model is proposed that DC-UbP participates in the delicate regulation of cellular ubiquitination and deubiquitination processes through linking the UbE1 and USP5 enzymes.

Materials and Methods
Plasmids, antibodies and reagents PCR-amplified cDNAs of human DC-UbP and its N-and Cterminal domains (residues 14-141, 129-234) were cloned into the vector pET22b(+), pGEX-4T-3 or pCMV-tag2B, respectively. UbE1 was cloned from mouse cDNA library and ligated into pET28a vector (Invitrogen) for purification. The FH (1-439) and SCCH (622-891) fragments of UbE1 were generated by PCR amplification and inserted into pET22b(+) vector. To get stable expression, the UFD domain of UbE1 (950-1058) was cloned into pGEX-4T-3 vector at the BamH I and Xho I sites. The cDNA encoding human USP5 was cloned from the cDNA library of HEK 293T cells; then it was ligated into pET-MG vector (BamH I/Xho I) for expression in E. coli or pcDNA3.1-Myc/His vector (BamH I/Xho I) for expression in mammalian cells. The cDNAs encoding the UBA12 (residues 625-749), UBA1 (631-692) and ZnF (169-289) domains of USP5 were generated by PCR amplification and inserted into pGBTNH or pET-M vector at the BamH I and Xho I sites, respectively. All constructs were verified by DNA sequencing.
The following antibodies were used: mouse anti-FLAG (Sigma), rabbit anti-FLAG (Sigma), mouse anti-Myc (Cell Signaling), mouse anti-UbE1 (Abcam), rabbit anti-USP5 (Proteintech), mouse anti-DC-UbP (Sigma), mouse anti-Ub (Santa Cruz), anti-His tag and mouse anti-GAPDH (Sigma). Ni-NTA (Qiagen) and glutathione-agarose beads (Amersham Biosciences) were utilized for protein purification or GST pull-down experiments. Protein A beads were the product of GE Healthcare and the anti-FLAG M2-Agarose were the product of Sigma. 15  Pull-down with cell lysates and LC-MS/MS analysis GST and GST-fused DC-UbP were expressed in E. coli and purified with glutathione-agarose column. HEK 293T cells (American Type Culture Collection) were cultured at 37uC in Dulbecco's modified Eagle's medium (DMEM, Gibco) supplemented with 10% (V/V) fetal bovine serum (Gibco), penicillin and streptomycin, and under a humidified atmosphere containing 5% CO 2 . The cells of 20610 cm plates were harvested and resuspended in the lysis Buffer A (20 mM HEPES, 150 mM NaCl, 1 mM EDTA, 10% glycerol, 1 mM DTT, 1% Triton X-100, pH 7.4) with 1 mM PMSF and Roche Protease Cocktail (1%). After incubating on ice for 15 min, the cell lysates were centrifuged for 10 min at 10,000 rpm and the clear supernatants were loaded onto the glutathione-agarose column (500 mL beads for each) pre-saturated with GST or GST-DC-UbP. After incubation for 3 h or overnight, the proteins were washed with the lysis buffer containing 350 mM NaCl and then eluted by a GSH elution buffer (50 mM Tris-HCl, 10 mM GSH, pH 8.0). The elutes were precipitated with 20% TCA, and the pellets were re-suspended with the loading buffer in 1/10 of the original volume for SDS-PAGE or silver staining. Different bands were selected and cut from the gel, washed and digested for 20 h by trypsin, and then the peptides were separated and detected by liquid chromatography (LC) coupled MS/MS analysis (LTQ, Thermo Finnigan; Agilent Technologies). The MS data were processed by using BIOWORKS software to give a peak list files, and the HUMAN (Version3.36) database was used for protein searching.

GST pull-down assay
The concentrations of the proteins were determined spectrophotometrically at 280 nm with the extinction coefficient of each protein. For GST pull-down assay, the purified GST or GSTfused UFD was mixed with glutathione-agarose beads (30 mL) for 1 h at 4uC. Then His-tagged DC-UbP or UbP_C was added and the mixtures were incubated for another 6 h at 4uC. The beads were washed three times with PBS buffer, and the proteins were eluted by GSH. The eluted proteins were subjected to SDS-PAGE (15% gel) and analyzed by immunoblotting with anti-His antibody. Proteins were also detected by Coomassie blue staining.

Transfection, immunoprecipitation and Western blotting
FLAG tagged DC-UbP or its fragments were co-transfected with Myc-USP5 into HEK 293T cells (Polyjet reagent) and cultured for another 36 to 48 h. The cell lysates were centrifuged and then the supernatants were subjected to SDS-PAGE. For immunoprecipitation, the clear supernatants were incubated with equal amount of monoclonal FLAG antibody or Myc antibody for above 3 h at 4uC. The bound beads were washed three times with the lysis buffer and once with PBS buffer, and then the proteins were released from the beads by glycine-HCl buffer (100 mM, pH 3.5). The samples were fractionated by SDS-PAGE and subjected to immunoblotting for USP5, DC-UbP and Ub, respectively. Endogenous UbE1 was detected directly by anti-UbE1 antibody. Co-transfection of pCMV-tag2B vector with Myc-USP5 was set as a control.

Isotope labeling of proteins and NMR titration
The 15 N-labeled DC-UbP, UbP_N and UbP_C proteins were prepared by using the M9 minimal medium containing 15 N-NH 4 Cl, and were purified by Ni + -NTA affinity column. The proteins were further purified via Superdex-75 column (GE Healthcare) and then desalted and lyophilized. For NMR titration, the samples (100 mM) were dissolved in a phosphate buffer (20 mM phosphate, 50 mM NaCl, pH 6.5) containing 10% D 2 O. All NMR spectra were acquired at 298 K on a Bruker Avance 600-MHz NMR spectrometer equipped with a TCI cryoprobe. 15 N-labeled protein or domain was titrated stepwise with unlabeled UbE1, USP5 or their various fragments (ZnF, UBA12 and UBA1 from USP5; FH, SCCH and UFD from UbE1). A series of 1 H-15 N HSQC spectra were obtained and the chemical shift changes were measured at different molar ratios. Data processing was followed as described previously [16,17].

In vitro deubiquitination assay
The deubiquitinating activity of USP5 was measured for cell lysates or the components from GST-DC-UbP pull-down by using Ub-AMC substrate as described previously [18,19]. All the experiments were performed in a buffer of 50 mM Tris (pH 8.0) 150 mM NaCl, 10 mg/mL ovalbumin and 1 mM DTT. Ub-AMC (250 nM) was incubated with various samples for measuring the hydrolytic activities. The fluorescence of AMC release was recorded on a Fluorescence Spectrophotometer (Varian Cary Eclipse) during the reaction process with an excitation at 380 nm and the emission at 460 nm.

In vitro ubiquitination assay
The cell lysates from 2610 cm plates were incubated with GST or GST-DC-UbP for 3,4 h, then eluted with 100 mL GSH buffer. The reactions of Ub activation and conjugation to E2 enzyme were carried out by mixing 10 mL of the pull-down components with 2 mM UbcH5C and 2 mM Ub and incubating in 20 mL of the reaction system (20 mM Tris, pH 7.5, 50 mM NaCl, 0.1 mM DTT, 4 mM ATP and 10 mM MgCl 2 ). The reaction lasted for 10 min at 25uC and was terminated with the sample loading buffer (with or without DTT). The samples were then subjected to SDS-PAGE, followed by anti-Ub antibody detection. The purified UbE1 (50 nM) or whole-cell lysates (1/10) was used as a control. K48-linked diUb was prepared according to the literatures [20,21]. The experiment for Ub conjugation to E2-25K was carried out under the similar condition, where DC-UbP (BSA as a control) was added in the reaction system for testing its potential effect on UbE1 activity and Ub conjugation.

Identification of the interacting partners for DC-UbP protein
Human DC-UbP/UBTD2 is a UbL domain-containing protein that is implicated in ubiquitination pathway [12], but its interacting partners and biological function remain largely unknown. To study its function, we firstly performed pull-down experiments on HEK 293T cell lysates with GST-fused DC-UbP protein. As compared with GST control, the sample pulled down by GST-DC-UbP exhibited mainly three additional protein bands at 100 kDa and around 250 kDa respectively in the silver staining gel (Fig. 1A). Combined with mass spectrometry (LC-MS/MS) analysis and database searches (Fig. S1), we revealed two important enzymes, ubiquitin-activating enzyme E1 (UbE1) and ubiquitin-specific protease 5 (USP5), in the bands around 250 kDa (Fig. 1A). Additionally, Ub molecules, polyUb chains or Ub precursors had also been identified together with GST-DC-UbP in these bands (Fig. S1). Interestingly, all proteins identified in the 250-kDa bands are involved in or related to the ubiquitination processes. These proteins could be detected with corresponding antibodies, and all these three antibodies displayed positive results to the high molecular-weight species (Fig. 1B), implying that DC-UbP interacts with UbE1 or USP5 and Ub plays a role in their associations. We presumed putatively that these proteins pulled down by GST-DC-UbP formed a large complex or assembly even under the SDS-PAGE condition. However, under the condition of 8 M urea, the band for this complex became very weak or almost undetectable (Fig. 1C). Note that the bands of associated proteins could also be blotted by the Ub antibody but disappeared in the presence of 8 M urea (Fig. 1D), suggesting the involvement of the Ub molecules or ubiquitinated species in the complex. Moreover, the respective bands of UbE1 (,120 kDa) and USP5 (,95 kDa) could also be detected with their respective antibodies in the sample treated with 8 M urea (Fig. 1D). Thus, the SDS-resistant DC-UbP complex or assembly formed at least by UbE1, USP5 and Ub is through non-covalent interactions, which is implicated in the ubiquitination and deubiquitination processes.

DC-UbP directly interacts with UbE1 or USP5
To verify the interactions among DC-UbP, UbE1 and USP5, we performed co-immunoprecipitation (co-IP) experiments by cotransfecting with FLAG-DC-UbP and Myc-USP5 (Fig. 2B). The result showed that DC-UbP could immunoprecipitate USP5 as well as endogenous UbE1, that is to say, the protein interacted with USP5 or UbE1. Since DC-UbP contains two domains ( Fig. 2A), we tested which domain interacts with UbE1 or USP5. The co-IP data indicated that the C-terminal part of DC-UbP (UbP_C, residues 129-234) specifically interacted with USP5 or UbE1 (Fig. 2B), whereas the N-terminal part (UbP_N, residues 14-141) did not. To further demonstrate which domain of DC-UbP binds with UbE1 or USP5, we performed NMR titration experiments on the 15 N-labeled DC-UbP protein. The backbone assignment of full-length DC-UbP was just derived from the assignments of the two individual domains [13,14]. With the stepwise addition of USP5 or UbE1 into the labeled DC-UbP, the intensities of 1 H-15 N cross peaks became weak or even disappearing in the spectra (Fig. S2), indicating that USP5 or UbE1 directly bound with DC-UbP. When the peak intensity (as indicated by peak height) of each amide of free DC-UbP was normalized as 1, the ratios of the peak intensities over those in DC-UbP alone at each titration step could be plotted against the residue number of DC-UbP (Fig. 2C, 2D) [16,17]. The mean peak intensities were 0.56 and 0.27 for the UBD and UbL domains of DC-UbP, respectively, at the 1:1 molar ratio of USP5 and DC-UbP (Fig. 2C). Similarly, the mean peak intensities for the UBD and UbL domains were 0.54 and 0.39 when titrated with UbE1 at the 1:1 molar ratio (Fig. 2D). The C-terminal UbL domain of DC-UbP exhibited a more intensive reduction in the peak intensities than the N-terminal UBD domain, suggesting that DC-UbP bound with USP5 or UbE1 preferentially on its C-terminal UbL domain.

The tandem UBA domains of USP5 is responsible for its binding with the UbL domain of DC-UbP
In USP5, both of its UBA and ZnF domains bind with Ub ( Fig. 3A) [19,22]. The C-terminal portion of DC-UbP is a Ub-like (UbL) domain, but it has a distinct surface with respect to Ub [13]. So we wondered whether this UbL domain could bind the tandem UBA12 domains or the ZnF domain of USP5. We labeled UbP_C (the UbL domain of DC-UbP) with 15 N and carried out NMR titration with USP5_UBA12 (residues 625-749) and USP5_ZnF (169-289) domains, respectively. Fig. 3B showed overlay of the 1 H-15 N HSQC spectra of UbP_C in the presence of different concentrations of USP5_UBA12. Some of the resonance peaks on the spectra of UbP_C shifted or disappeared with the stepwise addition of USP5_UBA12, suggesting that the UbL domain of DC-UbP interacted with the tandem UBA12 domains of USP5. However, titration with USP5_ZnF caused little chemical shift change (Fig. 3C), indicating that UbL did not interact with the ZnF domain of USP5. A rational reason is that the UbL domain of DC-UbP lacks the conserved C-terminal GG motif, which is necessary to the binding of Ub with the ZnF domain of USP5 [22]. On the other hand, binding of Ub with UBA is largely dependent on the hydrophobic surfaces [23,24], which is present either in Ub or in UbL of DC-UbP. We also determined the binding affinity of UbP_C with tandem USP5_UBA12 by NMR titration (Fig. 3D). The average K D value for the two UBA domains was around 400 mM. This affinity is weaker than those of Ub with USP5_UBA12 (K D = ,50 mM) and other UBA domains [24,25]. Fig. 3E showed the chemical-shift perturbation of UbP_C upon titrating with USP5_UBA12 at a molar ratio of 1:1. The residues on UbL exhibiting significant chemical-shift change or peak broadening were probably resided on the binding site. In combination of the chemical-shift perturbation data and interface analysis by HADDOCK program (Fig. S3) [26], we mapped the binding interface on the structure of the UbL domain (Fig. 3F). These residues including Phe195, Phe196, Val217 and Val218 are located at the b4 and b5 strands on the opposite surface of the positively-charged surface [13], which may contribute to the binding specificity of UbL with UBA12 of USP5.
To substantiate the essentiality of the key residues on the UbL surface of DC-UbP for binding with USP5_UBA, we generated three mutants, F195A, R199A and Q219A/I221A; and carried There exist two protein bands around 250 kDa from the protein species pulled down by GST-fused DC-UbP. By using LC-MS/MS assay, UbE1, USP5 and Ub have been identified from each band. GST was set as a control. The samples were eluted by Buffer A containing 350 mM NaCl. *, this band around 100 kDa was identified as P97/VCP. B, Immunoblotting detection of the identified proteins with antibodies against UbE1, USP5 or Ub. #, this high-molecular-weight species could also be detected by all the three antibodies applied. C, Dissociation of the highmolecular-weight species in the presence of 8 M urea. The 250-kDa bands were disappeared in the silver-staining gel. After TCA precipitation, the pellet was re-suspended with 8 M urea. D, Western blotting detection of the individual components in 8 M urea dissociated from the high-molecularweight species. The antibodies against UbE1, USP5 and Ub were applied. The 120-and 95-kDa proteins were UbE1 and USP5, respectively. The smear bands for Ub derivatives could not be detected due to very small amounts in the gel. doi:10.1371/journal.pone.0107509.g001 Coordination of UbE1 and USP5 by DC-UbP PLOS ONE | www.plosone.org Coordination of UbE1 and USP5 by DC-UbP PLOS ONE | www.plosone.org out NMR titration analysis. Here we used UBA1 for NMR titration, because we thought that individual UBA and tandem UBA12 bound with UbL in a similar manner. The result showed that, compared with wild-type UbL (Fig. 4A), the F195A mutant gave a reduced binding affinity considerably with UBA1 (Fig. 4B), whereas R199A lost its binding ability (Fig. 4C). However, as a control, the Q219A/I221A mutant did not show any affinity change (Fig. 4D). It suggests that residues Phe195 and Arg199 of UbL are important to interaction with the UBA domains of USP5.

The C-terminal UFD domain of UbE1 interacts with the UbL domain of DC-UbP
Mammalian UbE1 is also a large multi-domain enzyme with unknown structure, but the 3D structures of its orthologs have been solved [27,28]. The domain architecture of human UbE1 shows that it is mainly comprised of the adenylation domain (AD), the catalytic cysteine domains (FCCH, SCCH) and the C-terminal UFD domain (Fig. 5A) [29]. We purified three fragments of UbE1 and performed NMR titration experiments on DC-UbP and its separate domains. The individual UFD protein was not stable when purified in vitro, so we expressed and purified this domain as a GST-fused form for the experiments. The results showed that only the C-terminal UFD part bound with DC-UbP, whereas both the FH (FCCH+AD) and SCCH fragments did not (Fig. S4A-S4C). Similar with USP5 (Fig. 3), the UFD domain of UbE1 directly interacted with DC-UbP, preferentially on the UbL domain of DC-UbP but not on the N-terminal UBD domain (Fig. 5B). Further NMR titration experiments with individual parts of DC-UbP verified that only the C-terminal UbL domain bound with UFD (Fig. S4D, S4E). GST pull-down experiment confirmed that the UFD domain of UbE1 directly interacted with DC-UbP Coordination of UbE1 and USP5 by DC-UbP PLOS ONE | www.plosone.org (Fig. 5C, lane 2) and especially its C-terminal UbL domain (Fig. 5C, lane 5). We also attempted to map the interacting surface on the UbL structure from NMR titration data (Fig. 5B). The residues with significant intensity decreases include Val174, K178 and R179, and Leu210 and Val222. It is known that the UFD domain of yeast Uba1 contains an acidic surface that specially interacts with the positive charges on the surface of UbE2 (Fig. S5) [27]. Thus, we presumed that the UFD-domain binding site on the UbL domain is mainly located on the positively-charged surface including residues Lys178, Arg179 and Arg180 (Fig. 5D) [13], which is opposite to the UBA binding site on UbL (Fig. 3F). This provides a possibility that USP5 and UbE1 bind to the UbL domain of DC-UbP simultaneously but on the opposite position.

UbE1 and USP5 still reserve their respective catalytic activities when bound with DC-UbP
UbE1 and USP5 are two enzymes functioning in the ubiquitination and deubiquitination processes, respectively. Their catalytic activities can be monitored by several established methods in vitro [18,20]. We firstly examined whether the components pulled down by GST-DC-UbP from cell lysates still reserved the activity of catalyzing hydrolysis of Ub-AMC. The untreated cell lysates (with 10-fold dilution) were set as a positive control, while the GST pull-down sample was as a negative control. The increase of fluorescence intensity represents the hydrolysis of Ub-AMC. Fig. 6A displayed the time courses of fluorescence intensity changes upon addition of different pulldown samples. The data clearly demonstrated that the components pulled down by GST-DC-UbP (from cell lysates) could hydrolyze Ub-AMC, albeit the activity seemed to be lower than that of the untreated cell lysates. This indicates that the native USP5 enzyme pulled down by GST-DC-UbP from the cell lysates still reserves its catalytic activity on Ub-AMC. It also implies that binding with DC-UbP may have little impact on the USP5 activity. To verify this speculation, we purified the USP5 protein and monitored the effect of DC-UbP or its UbL domain on the enzymatic activity. The kinetic profiles showed that the deubiquitinating activity of USP5 was not affected significantly even in the presence of more than ten-fold excess of DC-UbP or UbP_C ( Fig. 6B and Fig. S6). Therefore, DC-UbP binding has no detectable effect on the catalytic activity of USP5.
We also measured the Ub activating activities of the components pulled down by GST-DC-UbP from cell lysates by detecting the Ub conjugation to UbcH5C (UbE2). As a positive control, the Coordination of UbE1 and USP5 by DC-UbP PLOS ONE | www.plosone.org purified UbE1 could activate Ub and conjugate Ub to UbcH5C to form a thioester linkage (Fig. 6C, lane 1). The activating activity could also be detected from the untreated cell lysates (with 10-fold dilution) (Fig. 6C, lane 3), suggesting that the endogenous UbE1 had its activity of catalyzing activation of Ub for transferring to UbE2. Besides UbE2 thioester, the DTT-resistant species, like diUb, tetraUb or other higher molecular-weight Ub chains could also be observed due to potential contaminant of E3 ligase from the cell lysates. For the components pulled down by GST-DC-UbP, this sample still had the activities of activation and conjugation of Ub to UbE2 (Fig. 6C, lane 7), whereas, under the same condition, the components from GST pull-down did not possess this action. These data further support the specific interaction between DC-UbP and UbE1, and more importantly, UbE1 reserves the activating activity in its DC-UbP-associated form. We then carried out the experiment on synthesis of diUb chain in vitro to examine whether the presence of excess DC-UbP affects the downstream of ubiquitination by UbE1 activity (Fig. 6D). By detecting formation of K48-linked diUb at different incubation time, it was obvious that DC-UbP had no detectable effect on the reaction process. Taken together, all the data demonstrate that UbE1 reserves its activating activity when it is bound with DC-UbP.

Overexpression of DC-UbP improves the association between UbE1 and USP5 in cell
Because association of DC-UbP did not affect the catalytic activities of UbE1 and USP5 in vitro, we wondered whether DC-UbP association had any effects on the cellular functions of these two enzymes. We previously observed that FLAG-tagged DC-UbP could immunoprecipitate both UbE1 and USP5 in HEK 293T cells (Fig. 2B). So we asked whether USP5 immunoprecipitate endogenous UbE1 under the condition of DC-UbP overexpression. When co-transfected with FLAG-DC-UbP, Myc-USP5 immunoprecipitated more UbE1 than that with the mock (Fig. 7A). We then detected total Ub conjugates of the cells transfected with FLAG-DC-UbP. The Ub conjugation level in the cell lysates was not changed obviously for co-transfection of Myc- Coordination of UbE1 and USP5 by DC-UbP PLOS ONE | www.plosone.org USP5 with FLAG-DC-UbP. However, after immunoprecipitation with the anti-Myc antibody, the Ub conjugation level was considerably increased as compared to the vector control (Fig. 7B), especially the conjugates with large molecular weights in the gel, suggesting formation of a functional complex mediated by DC-UbP. Because DC-UbP binds with USP5 and UbE1 directly, the increased amount of DC-UbP provides a mediator for association between UbE1 and USP5, in which some ubiquitinated conjugates are included. We then produced three mutants on the UbL surface of DC-UbP (F195A, R199A and F195A/R199A) that could destroy the interaction with USP5. Indeed, these mutations somewhat reduced the interactions with USP5 (Fig. 7C, 7D, middle panels); however, they had no significant effect on the association of USP5 with endogenous UbE1 (Fig. 7C, 7D,  bottom). This is probably that endogenous DC-UbP or other factors in cell are also involved in mediating the association between USP5 and UbE1. All these data suggest that DC-UbP mediates association between USP5 and UbE1 in cell through directly interacting with these two enzymes.

Overexpression of DC-UbP modulates the cellular ubiquitination level
To get information whether DC-UbP impacts on the cellular ubiquitination level, we overexpressed DC-UbP and detected the Ub conjugates in HEK 293T cells. When the cells were transfected with the plasmid for DC-UbP at an appropriate amount, the total level of the Ub conjugates increased considerably as compared with the vector control (Fig. 8A), whereas the levels of its interacting partners, USP5 and UbE1, remained unchanged. We then performed a dose-dependent assay, and found that the level of Ub conjugates increased gradually with the increase of DC-UbP transfection (Fig. 8B). However, the R199A and F195A/R199A mutants of DC-UbP lost this effect to increase the conjugation level (Fig. 8C). Further extensive study by dose-dependent experiment showed that the double mutant totally lost this modulating effect on the Ub conjugation level (Fig. 8D). Taken together, these results demonstrate that DC-UbP modulates the cellular ubiquitination levels depending on its cellular amount, which is probably through mediating the association between USP5 and UbE1. Figure 6. Effects of DC-UbP binding on the catalytic activities of UbE1 and USP5. A, Fluorescence assay for deubiquitinating activity of the components from GST-DC-UbP pull-down. The GSH conjugated beads were incubated with purified GST-DC-UbP and HEK 293T cell lysates and then eluted with GSH. The components that may contain the USP5 enzyme were subjected to activity assay with Ub-AMC (250 nM) as a substrate. The cell lysates was set as a positive control; and the GST pull-down components was a negative control. B, Fluorescence assay for deubiquitinating activities of purified USP5 in the presence of DC-UbP or UbP_C. USP5 (10 nM) was pre-incubated with 10 mM DTT for 15 min, then DC-UbP or UbP_C was added with 10-fold excess in amounts. C, In vitro ubiquitin conjugation assay for the components from GST-DC-UbP pull-down. The pull-down components were subjected to assaying the activity of UbE1 enzyme (lanes 7 and 8). Purified UbE1 and cell lysates were set as positive controls, and the GST pull-down components were a negative control. In this reaction, the UbcH5C-Ub (E2-Ub) conjugates were generated, while dissociated in the presence of DTT. D, Effect of purified DC-UbP on the diUb synthesis in vitro. The concentrations of UbE1 and E2-25K were 0.25 mM and 2 mM, respectively, while those of His 6 -DC-UbP and BSA were 80 mM. The reaction system (100 mL) included 0.25 mg Ub, 0.25 mg Ub-K48R and 2 mM ATP in a buffer of 50 mM Tris, 10 mM MgCl 2 , pH 8.0. doi:10.1371/journal.pone.0107509.g006

Knockdown of DC-UbP reduces the cellular ubiquitination level
When knockdown of DC-UbP in HEK 293T cells by siRNAs (Fig. 9A), the amount of UbE1 was decreased considerably as detected with the antibody against UbE1 (Fig. 9B). Similarly the USP5 level was also reduced as the anti-USP5 antibody indicated (Fig. 9C). This indicates that formation of the DC-UbP complex may be beneficial to the stabilities of both UbE1 and USP5. Intriguingly, the ubiquitinated conjugate level was also reduced in the cell lysates (Fig. 9D). It is possible that reduction of the UbE1 enzyme may result in decrease of the ubiquitination level and consequently the ubiquitinated conjugates in cells [30]. Although the amount of USP5 is simultaneously decreased when knockdown of DC-UbP in cell, we are still not clear whether this enzyme is contributable to regulating the ubiquitination level. Since USP5 is known to function mainly in hydrolyzing unanchored Ub chains or Ub precursors [9], it is reasonable to presume that reduction of USP5 may affect the Ub pool but not the ubiquitinated substrates.

Cross-talk between protein ubiquitination and deubiquitination
Protein ubiquitination and deubiquitination are two reversible processes that control the cellular homeostasis of various Ub forms, including free Ub, Ub precursors, unanchored Ub chains and ubiquitinated substrates [9]. Accumulating evidence shows that some ubiquitination enzymes can interplay with deubiquitinating enzymes and mediate the cross-talk between these two Figure 7. Effects of DC-UbP overexpression on the association of UbE1 and USP5. A, Co-IP for detecting the association of UbE1, USP5 and DC-UbP. Myc-USP5 was co-transfected with FLAG-DC-UbP or mock vector into HEK 293T cells and then precipitated with anti-Myc antibody. The endogenous UbE1 was monitored by anti-UbE1 antibody. B, As in (A), detected with Ub antibody. After IP, the Ub conjugate level and especially the 250-kDa species were increased when co-transfected with DC-UbP. *, IgG heavy chain. C, Effect of the F195A/R199A mutant of DC-UbP on the association of USP5 and UbE1. Due to partial removal of the FLAG tag during co-IP experiment, two bands for DC-UbP by an antibody against DC-UbP were observed. D, As in (C), effect of the DC-UbP mutants on the association of USP5 and UbE1. Myc-USP5 was co-transfected with FLAG-tagged WT or its mutants (F195A, R199A, and F195A/R199A) and analyzed by Western blotting. The gels were blotted with anti-Myc, anti-DC-UbP and anti-UbE1 antibodies, respectively. Due to proteolytic degradation, two bands were detected in the gel by anti-DC-UbP antibody. doi:10.1371/journal.pone.0107509.g007 opposite processes. The DUB ataxin-3 coordinates with the E2 enzyme Ube2w in regulating the ubiquitination cycle of CHIP [31]; while another DUB, OTUB1, interacts with the E2 enzyme UBC13 to inhibit the RNF168 pathway for DNA damage response [32,33]. Some DUB enzymes can also bind with an E3 ligase, such as USP8 with Nrdp1 [34] and USP7 with ICP0 [35], and regulate the stability and activity of the E3 ligase. More interestingly, the deubiquitinating enzyme A20 contains both Ub ligase and DUB domains that cooperatively down-regulates NF-kB signaling [36]. Our finding that DC-UbP/UBTD2 couples UbE1 and USP5 together to form a functional complex will exemplify the cross-talk between ubiquitination and deubiquitination processes that orchestrates the ubiquitin pool in cell.
Both UbE1 and USP5 are involved in the homeostasis of ubiquitin pool UbE1 (Uba1) is central to Ub conjugation since it initiates the first key step of ubiquitination process by catalyzing Ub activation [1]. UbE1 is relatively unique for Ub conjugation and highly conserved in all eukaryotes [37]. USP5 is an important DUB enzyme that mainly disassembles the unanchored polyUb chains or Ub precursors to regenerate free Ub molecules [38,39]. As its yeast homologue Ubp14 [40], USP5 can sequentially remove Ub Coordination of UbE1 and USP5 by DC-UbP from the proximal end of unanchored polyUb chains [41]. Thus, it is reasonable to deduce that both UbE1 and USP5 are contributable to maintaining the cellular homeostasis of Ub pool.
Both USP5 and UbE1 are large proteins with multi-domains involved in regulating cellular ubiquitination levels. Structurally, USP5 is constituted by two ZnF-UBP domains [42], a catalytic USP domain and two inserted UBA domains [22]. All the four Ub-binding domains participate in binding with polyUb chain in the cleavage process [22,43]. Our study suggests that DC-UbP binds to USP5 on the tandem UBA domains, implying that binding of DC-UbP to USP5 potentially regulates the deubiquitination process. On the other hand, the UbE1 enzyme utilizes ATP to activate the terminal glycine residue of Ub generating a covalent thioester linkage between 'activated' Ub and the UbE1 enzyme itself [44]. The UbE1 protein is mainly comprised of three domains, the adenylation domain that binds ATP and Ub, the catalytic cysteine domain that binds activated Ub, and the Cterminal Ub-fold domain (UFD) that recruits specific E2 conjugating enzymes [27]. Our finding that DC-UbP binds to UbE1 on the C-terminal UFD domain suggests that DC-UbP also participates in regulating the ubiquitination process. Thus, coordination of UbE1 and USP5 by DC-UbP is involved in modulating the Ub pool in cell.

DC-UbP links UbE1 and USP5 enzymes to modulate the ubiquitin pool
Ub is a small protein modifier for huge number of proteins involved in myriad biological processes, such as DNA repair, gene regulation, protein degradation, translocation, apoptosis, and immune response [1,2,5,45]. Maintaining the balance of the free Ub pool is critical to the eukaryotic cells [46,47]. USP5 is a common deubiquitinating enzyme mainly responsible for the cleavage of unanchored polyUb chains or the Ub precursor proteins, and thus plays a role in maintaining the homeostasis of free Ub pool. The free Ub molecules stored in the pool can then be utilized in the next cycle and activated by UbE1 to participate in the consequent processes, conjugation to UbE2 and transfer to specific substrate by E3 ligase. We revealed that DC-UbP directly interacts with USP5 and UbE1 to form a dynamic complex through transient interactions, which can improve their stabilities, mediate their associations and integrate their functions. Our finding suggests that DC-UbP is involved in modulating the Ub pool in cells, although the underlying molecular mechanism is not clearly understood. Thus, DC-UbP modulates the cellular levels of ubiquitinated conjugates by coordinating USP5 and UbE1. These two reverse processes, ubiquitination and deubiquitination, are coupled together and interplay with each other by assistance of DC-UbP (Fig. 10). In this model, UbE1 and USP5 can bind to the Coordination of UbE1 and USP5 by DC-UbP C-terminal UbL domain of DC-UbP but on the opposite surface ( Fig. 3F and Fig. 5D), forming a population of complex. This dynamic complex organized by DC-UbP can cooperatively integrate the functions of UbE1 and USP5, which may improve the efficiency of cellular deubiquitination and ubiquitination processes.
In conclusion, DC-UbP mediates USP5 and UbE1 to form a functionally dynamic complex that plays a regulatory role in facilitating the efficiency of ubiquitination and deubiquitination in cellular metabolism. As aberrant regulation of USP5 and UbE1 has been implicated in some diseases [30,48], the USP5 and UbE1 enzymes have been highlighted as new therapeutic targets, and several small molecule inhibitors have been developed [49,50]. Because DC-UbP can integrate the functions of these two enzymes, targeting DC-UbP by inhibitors might become a promising therapeutic strategy for the relevant diseases. Figure S1 LC-MS/MS analysis of different protein bands in SDS-PAGE gels from the GST-DC-UbP pulldown components. A, A representative profile for separating the peptides digested from a 250-kDa protein band. The RP-C18 column (0.15 mm6150 mm, Column Technology Inc.) and the Zorbax 300 SB-C18 peptide traps (Agilent Technologies) were used for separation and collection of the digested peptides. B, A representative profile of the mass spectrometry for one of the peptide fragments. The MS data were further processed using BIOWORKS software to give a peak list files and the protein database of HUMAN (Version3.36) were selected for searching. C, Top four proteins identified from MS analysis of the two bands around 250 kDa, which include UBTD2, USP5, UbE1 and UBB (or UBC, RPS27A). (TIF) Figure S2 NMR titration for charactering the interactions of DC-UbP with USP5 and UbE1. A, Chemical-shift assignment of the DC-UbP protein. The 1 H-15 N HSQC spectrum shows the resonance peaks of almost all amides of full-length DC-UbP. The assignment is derived from the chemical-shift assignments of the individual UbP_N (PDB: 2KSN) and UbP_C (PDB: 1TTN) fragments that have been completed. B, Overlay of the HSQC spectra of 15 N-labeled DC-UbP (100 mM) and addition of USP5 at different molar ratios. The peak broadening during USP5 titration indicates direct interaction between DC-UbP and USP5. C, As in (B), UbE1 titration. The peak broadening during UbE1 titration indicates direct interaction between DC-UbP and UbE1. (TIF) Figure S3 Ribbon representation of the structure of UbL-UBA1 complex as constructed by using HADDOCK method. The interfaces between UbL and UBA1 are shown. As known, for UBA1, the 666 MGV 668 loop between helix1 and helix2 contributes to the specific interaction. The residues Arg199, Gln219 and Ile221 are potentially located in the interface for the UbL domain of DC-UbP. Combined with mutagenesis, we thus characterized the residues Phe195 and Arg199 of UbL that are important to interacting with UBA1. (TIF) Figure S4 NMR titration for charactering the interactions of DC-UbP with various fragments of UbE1. A, Overlay of the HSQC spectra of 15 N-labeled DC-UbP (100 mM) and addition of FH (FCCH and AD) fragment at different molar ratios. There is no considerable peak change in the spectra during titration with the FH fragment. B, As in (A), SCCH titration. There is no considerable peak change in the spectra during titration with the SCCH fragment. C, As in (A), GST-UFD titration, except that the concentration of 15 N-labeled DC-UbP Coordination of UbE1 and USP5 by DC-UbP was 20 mM. With addition of GST-UFD, some peaks in the Cterminal part (UbL) of DC-UbP become weak or disappearing, indicating that UFD specifically binds to the UbL domain. D, Diagram of the peak intensity (height) changes of UbP_N titrated with GST-UFD at a molar ratio of 1:3 against residue number. E, As in (D), diagram of UbP_C titrated with GST-UFD at a molar ratio of 1:3. The lines indicate the mean peak intensities for the UBD (a.a.  and UbL (152-225) domains, respectively, suggesting that UFD specifically binds with UbL but not with UBD. (TIF) Figure S5 Structural model of the UFD domain of human UbE1 showing the electrostatic surface. The acidic residues on the surface corresponding to those in yeast Uba1 potentially binding to the positively-charged interface of Ubc1 (UbE2) are highlighted. The structure was generated by homology modeling using I-TASSER server and the electrostatic surface was displayed with MOLMOL. (TIF) Figure S6 Effect of the C-terminal part of DC-UbP (UbP_C) on the deubiquitinating activity of USP5. The fluorescence increases were monitored for the deubiquitinating activities of purified USP5 in the presence of different molar ratios of UbP_C/USP5. The concentrations of USP5 and Ub-AMC were 10 nM and 250 nM, respectively. (TIF)