Conceived and designed the experiments: AD ABC RGD GM. Performed the experiments: AD ABC. Analyzed the data: AD ABC RGD GM. Wrote the paper: AD ABC RGD GM.
Current address: Biobanc, Hospital Clínic-IDIBAPS, Barcelona, Spain
The authors have declared that no competing interests exist.
USP25m is the muscle isoform of the deubiquitinating (DUB) enzyme USP25. Similarly to most DUBs, data on USP25 regulation and substrate recognition is scarce.
Ubiquitin (Ub) modifies protein architecture when covalently attached to its substrates. Besides being the main tag for sending misfolded proteins to the proteasome, Ub also plays a relevant role in protein-protein interaction and modulation of catalytic activity or protein fate
Although many studies have investigated the activation of Ub and its transfer to substrates
UBDs are found not only in proteins that interact with ubiquitinated substrates, but also in ubiquitinating or deubiquitinating enzymes. The deubiquitinating enzymes (DUBs) hydrolyze the Ub moieties conjugated to substrates and thus, process newly synthesized Ub, recycle Ub, or edit polyUb chains
Data on the substrate specificity and physiological function of most DUBs, including USP25, are still scanty.
We aimed to identify structural domains relevant for USP25m regulation. By
USP25m sequence (1125 aa) alignments revealed five highly conserved distinct motifs (I to V), embedded in two domains (USP1 and USP2) characteristic of the ubiquitin-specific protease family (UBPs, USPs in humans)
A. Sequence homologies revealed five highly conserved USP motifs (I to V) in two domains (USP1 and USP2) that catalogue USP25m as a deubiquitinating enzyme. Cys-178 is the putative active site of the enzyme, since it is conserved in all analyzed members of the family. B. Deubiquitinating activity assays in
A. USP25m contains one UBA and two UIM (USP25_1, USP25_2) domains, as shown by alignments with other UBAs or UIMs. B. Schematic representation of the USP25m C-terminal and UBD deletion mutants: ΔUBA (Δ19-58 aa, inclusive), ΔUIM1 (Δ96-115 aa, inclusive), ΔUIM2 (Δ121-141 aa, inclusive), ΔUBA-UIM1 (Δ19-115 aa, inclusive), ΔUBA-UIM1-UIM2 (Δ19-141 aa, inclusive), ΔUIM1-UIM2 (Δ96-141 aa, inclusive). The constructs bearing serial deletions of USP25m at the C-terminus are also shown (E679X, E769X, Q863X, E1020X). C. Deubiquitinating activity assays indicated that UBDs were not required to cleave off ubiquitin (left upper panel). The mutant USP25mE679X was unable to hydrolyze Ub from the Ub-βgal substrate, indicating that the region between the amino acids 679 and 769 was required for enzymatic activity (right upper panel). The empty GST vector and the full length USP25m were respectively used as negative and positive controls. The expression level of each USP25m mutant was comparable (lower panels).
To assess whether the UBA and UIM domains contribute to USP25m deubiquitinating activity, we co-expressed GST epitope-tagged deletion mutants of USP25m, which lacked one or several of the UBDs (
USP enzymes are usually proteins of high molecular weight, which stretch at the N- and/or the C-terminus of the USP catalytic domains. These extensions have been proposed to be involved in substrate recognition, regulation of the catalytic activity or subcellular localization. USP25 stretches more than 450 amino acids at the C-terminus, including the muscle-specific peptide (introduced by alternative spliced exons 19a and 19b, see
In contrast with the results obtained with the UBD mutants, the analysis of the serial truncated proteins at the C-terminus of the USP25m protein clearly showed that mutant E679X was incapable of cleaving off the ubiquitin moiety of the Ub-β-gal protein, whereas mutants E769X, Q863X and E1020X still retained the enzymatic activity (
As UBDs have also been involved in shifts in subcellular localization, we asssessed whether the wild-type USP25m and UBD-deleted constructs, either in their catalytically active or inactive forms, showed different localizations. No change in the distribution pattern was observed in any condition, indicating that the UBA and UIM domains were not required for targeting USP25 to its localization (
The dynamic nature of the Ub-pathway requires the formation of complexes in which enzymes and cofactors are transiently recruited, not only E2 and E3 ligases but also DUBs
A. Coimmunoprecipitation assays after co-expressing two differently tagged forms (cMyc- or GFP-) of either the wild-type USP25m or the C178S mutant, showed that USP25 dimerized
Native gel electrophoresis followed by western blot immunodetection confirmed that USP25 was included in high molecular weight complexes (>250 kDa, data not shown). As non-denaturing conditions were used to detect protein complexes, the dimerization (oligomerization) of USP25 could either be direct or require some other substrate/partners.
Many E3 ligases and some DUBs undergo post-translational modifications, such as ubiquitination or sumoylation, which modulate the recognition of their substrates
A. Immunodetection of cell lysates expressing Myc-tagged USP25m showed one additional high molecular-weight band. This band was stronger when co-expressing His(6x)-Ub, suggesting that it corresponded to mono-ubiquitinated USP25m. The high molecular weight bands were stronger when co-expressing His(6x)-Ub and the catalytically inactive mutant USP25mC178S. The lower histogram shows the percentage of non-modified versus mono-Ub-conjugated USP25m. B. The same experiment was performed co-expressing His(6x)-Ub with all the UBD USP25m deletion mutants, in combination or not with the C178S mutation. Again, the ubiquitinated band was much visible in the C178S version of the mutants. C. Ni2+ pull-down assays to purify His(6x)Ub-conjugated proteins confirmed that USP25m was ubiquitinated. All the mutant constructs were tested, confirming that monoubiquitination (and multi- or poly-ubiquitination) did not depend on UBDs, neither on the presence of the C-terminus. The ratio output/input is 4. (Output samples were eluted at pH 4.5, which could account for the slight variation in the apparent protein molecular weight compared to inputs). D. Protein stability of the USP25m full-length and mutant constructs. Cells were grown in standard conditions (−), or treated with MG132 (+). Immunodetection of α-tubulin was used as a loading control.
To examine the possible involvement of UIMs in USP25m ubiquitination
To further study USP25m ubiquitination, we performed a Ni2+ pull-down assay in cells co-expressing the different USP25 mutants together with His-tagged Ubiquitin. We recovered ubiquitinated USP25m proteins in all UBD deletion mutants (
To identify the lysine residue involved in the mono-ubiquitination, we co-transfected cells with the mutant USP25mC178S with His(6x)Ub, enriched the lysate in USP25m forms by immunoprecipitation with an anti-cMyc antibody, and analysed the obtained bands by LC-ESI-QTOF mass spectrometry. One Ub-modified peptide appeared consistently, indicating that K99 was the most likely acceptor site (
Peptide (+Ubiquitin modification) | Protein identified |
KYVDPSR | USP25 |
TPTEVWR | USP25 |
YNDIAVTK | USP25 |
AIKLEYAR | USP25 |
YLSYGSGPK | USP25 |
TEIENDTR | USP25 |
DSRNPYDR | USP25 |
FLAVGVLEGK | USP25 |
VLEASAIAENK | USP25 |
TLLEQFGDR | USP25 |
YLFALLVGTSK | USP25 |
AVEILKDAFK | USP25 |
HQQTFLNQLR | USP25 |
AEEETDEEKPK | USP25 |
AQFLIQEEFNK | USP25 |
LEFPQVLYLDR | USP25 |
FEFNQALGRPEK | USP25 |
ETGITDEEQAISR | USP25 |
LAQEDTPPETDYR | USP25 |
DSNGNLELAVAFLTAK | USP25 |
LNEQAAELFESGEDR | USP25 |
ETGPQLVGIETLPPDLR | USP25 |
IHNKLEFPQVLYLDR | USP25 |
SGQEHWFTELPPVLTFELSR | USP25 |
LRESETSVTTAQAAGDPEYLEQPSR | USP25 |
YISVGSQADTNVIDLTGDD |
USP25 (Ubiquitinated) |
EGIPPDQQR | Ubiquitin |
ESTIHLVLR | Ubiquitin |
IQDKEGIPPDQQR | Ubiquitin |
TITLEVEPSDTIENVK | Ubiquitin |
Taken together, our results strongly suggest that: i) USP25m was ubiquitinated and underwent autodeubiquitination, ii) UIM1, UIM2 and UBA domains promoted, but were not strictly required for monoubiquitination, iii) the C-terminal region is relevant for the protein stability and, when deleted, USP25m is polyubiquitinated and targeted for proteasome degradation, and iv) the preferential target lysine for ubiquitination is K99.
Given that USP25 was also reported to be a target for SUMO
Although the targets of most DUBs are unkown, USP25 is a DUB that specifically recognizes and binds its substrates in physiological conditions. We previously reported that the muscle-specific isoform USP25 interacted with MyBPC1, and that the DUB activity of USP25m rescued this substrate from proteasome degradation. This recognition was highly specific and depended on the peptide encoded by the muscle-specific exons 19a and 19b, as the ubiquitous USP25 isoform was unable to rescue this substrate
Given the reported relevance of UBDs in the regulation of protein folding and modular domain interactions, we were prompted to test the effect of the absence of UBA and/or UIM domains of USP25m in the rescue of MyBPC1 from proteasome degradation. As a positive control, the expression of the wild-type USP25m rescued MyBPC1 to the levels attained with the MG132 proteasome inhibitor (
A. MyBPC1 is differentially rescued from proteasome degradation depending on the presence of the distinct UBDs. Transfection of MyBPC1 with the empty GFP vector was used as the negative control, and addition of MG132 was used as a positive control. B. Relative quantification of the MyBPC1 rescue by different USP25m mutants. α-tubulin was used for normalization of protein concentration (data not shown) and USP25m expression levels were used to normalize for transfection efficiency. The rescue achieved by the wild-type USP25m was considered as the reference (value of one). At least three different replicates were used for quantification. Asterisks indicate statistical significance (p<0.05, Mann-Whitney test). C. The catalytically inactive C178S and the K99R mutants behaved similarly and are unable to rescue MyBPC1 from proteasome degradation in a time-course experiment when new protein synthesis is inhibited. The rescue achieved by expression of the wild-type USP25m was used as a control. D. The MyBPC1 levels (normalized by α-tubulin expression) were quantified and expressed relatively to those observed at time 0 h (30 h post-transfection, before cycloheximide treatment), which were considered 100%. The values corresponded to a minimum of three different replicates in several independent experiments. Asterisks indicate statistical significance (p<0.05, Mann-Whitney test). CHX- cicloheximide.
As aforementioned, previous reports showed that sumoylation of USP25 occurred at K99, and this modification inhibited USP25m deubiquitinating activity on tetraubiquitin chains
As expected, the USP25mC178S was not able to rescue MyBPC1 in a time-course experiment, whereas the expression of the wild-type USP25m raised the half-life of MyBPC1, as its levels were steadily maintained through time when protein synthesis was inhibited (
As DUBs are the least known members of the UPS, we studied the physiological function of USP25 by domain dissection. We particularly focussed in the three predicted UBDs, as these motifs are usually clustered in the same protein and confer subtle differences in the interaction with ubiquitinated substrates. By generating serial and combinatorial deletions, we assessed USP25 protease activity on a recombinant substrate, and showed that all UBD deletion mutants were catalytically active. We concluded that these domains were not strictly required for ubiquitin recognition or the deubiquitinating activity.
Increasing evidence support that ubiquitin-pathway enzymes (E2–E3 ligases, and more recently, DUBs) form cooperative complexes
One of the reported functions of UBA and UIM sequences is the promotion of ubiquitination of the protein in which they are embedded, thus facilitating autoregulation
Concerning the ubiquitination state and fate of the wild-type protein and the UBD mutants, we surmised that it corresponded mainly to mono- and multi- ubiquitinated forms, not related to protein degradation, as they were stable through time under our conditions. In contrast, the modification of the C-terminus mutants was compatible with polyubiquitination, as their protein levels were increased when the proteasome was inhibited, pointing to the relevance of the last 106 amino acids in USP25m stability. Polyubiquitination did not appear to be related to the catalytic activity of USP25m as: i) truncated mutants E1020X, Q863X and E769X were enzymatically active but degraded by the proteasome, and ii) of the two catalytically inactive E679X and C178S, the former was polyubiquitinated and degraded, whereas the latter was monoubiquitinated and this modification was not related to degradation. Therefore, autodeubiquitination does not seem to be required for USP25m stability.
Finally, we assessed the contribution of the UBD deletion mutants to the recognition of the USP25m specific substrate MyBPC1, considering that the requirements for the interaction with a specific physiological substrate might be different from those of a synthetic polyubiquitin substrate. None of the UBDs was critical for enzyme-substrate interaction, as all the mutants rescued the substrate from proteasome degradation. However, the effects were distinct depending on the domains deleted or preserved. The analysis of the contribution of the single and combined domains suggest that the UBA domain negatively modulated the USP25 function mainly by interaction with the UIM1 domain. The effect of the two UIM domains on the substrate rescue appeared to fit an additive/synergical mode of action. Deletion of the three UBDs would effectively remove all these regulatory domains, including those involved in SUMO modification and the target K99. Given that the overexpression of this UBD-deleted USP25 construct caused increased rescue of MyBPC1, we interpreted that the lack of these regulatory domains allowed USP25m free (non-regulated) access to its substrate. UBDs then would mostly contribute to the enzyme regulation in response to cellular requirements rather than to strict substrate recognition.
Ubiquitin and SUMO pathways may engage in cross-talk, determining opposite fates or functions of a particular substrate, and even compete for the same residues
A. Post-translational modifications of USP25m concerning monoubiquitination in Lys99 (K99) and sumoylation in K99 and K141. U: Ubiquitin; S: SUMO; UBA: Ubiquitin associated domain; SIM: SUMO interacting motif; UIM: Ubiquitin interacting motif; USP: Ubiquitin specific protease catalytic domains; CC: coiled coil domain; 19a19b: peptide encoded by the muscle-specific exons. B. Model on the regulation of USP25m activity through alternative and mutually exclusive conjugation of SUMO (inhibiting) and ubiquitin (activating) on the same lysine residue (K99) (see
Further regulation of the enzyme activity would rely on autodeubiquitination (either intra- or inter- molecularly in a dimer/complex), which would make this lysine residue available for alternative modifications, thus allowing the shift between the enzymatic activity states (
According to this model, regulation and integration of cell signals would be exerted through the N-terminus of USP25, where the SIM, UBDs and the preferential sites for SUMO and ubiquitin conjugation are clustered. In this context, the deletion of the three UBDs would remove all the regulating domains of the enzyme and permit free access to the substrate, which would explain the higher rescue obtained for this mutant. Indeed, we have previously showed that the recognition of the specific substrate MyBPC1, was dependent not on UBDs but on the peptide encoded by exons 19a19b
Deubiquitinating enzymes have to integrate cellular signals and promote dynamic interactions with their substrates, similarly to what occurs with E2–E3 ligases. Modification of a single target residue in USP25 by SUMO (inhibiting) or ubiquitin (activating), combined with the cluster of SIM and UBDs domains in the same molecule, provides new insights and open new avenues for the study of DUB regulation concerning substrate recognition and catalytic activity. To illustrate this statement, the USP25 closest homolog (sharing 52% of amino acid identities) is USP28
The USP25m protein sequence was analyzed using the InterPro (
Mutants USP25mC178S and USP25mK99R were generated by site directed mutagenesis to serine using the QuickChange Site-Directed Mutagenesis Kit (Stratagene). Expression constructs with the full-length USP25m cloned in pGEX-4-T1 (GE Healthcare), pcDNA3 (Invitrogen) and pEGFP-C2 (Clontech), were used to generate by PCR the UBD deletion mutants of USP25m (ΔUBA, ΔUIM1, ΔUIM2, ΔUBAUIM1, ΔUBAUIM1UIM2 and ΔUIM1UIM2). The Accuprime
The ubiquitin-specific protease activity of USP25m and of all the mutant constructs was analyzed as described elsewhere
HEK293T cells were seeded on 100 mm tissue culture dishes (2×105 cells/dish). After 16 h, cells were transiently co-transfected with cMyc-USP25m and GFP-USP25m, either full-length or the deletion mutants at the N- and C- terminus, using Lipofectamine 2000 (Invitrogen). Cells were collected 42 h postransfection, resuspended in lysis buffer (0.5% Nonidet P-40, 50 mM TrisHCl pH 7.5, 1 mM EDTA, 150 mM NaCl and protease inhibitor cocktail (Roche) and lysed by sonication. Protein extracts were recovered after removal of cellular debris by centrifugation, incubated at 4°C with 2 µg of anti-cMyc mAb (Santa Cruz Biotechnology) during 4 hours with end-over-end mixing. The protein-antibody complexes were removed with 1 hour incubation at 4°C with protein G-Sepharose beads (Amersham GE-Healthcare). After washing, bound proteins were eluted from the beads by boiling 5 min with protein loading buffer, loaded onto 8% SDS-PAGE gels and analysed by Western Bloting using anti-GFP pAb (1∶1000, Santa Cruz Biotechnology), anti-cMyc mAb (1∶1000, Santa Cruz Biotechnology).
Human embryonic kidney (HEK)293T cells were plated (2×106) in 10 cm Petri dishes. After 12 hours, they were co-transfected with 6 µg of a construct expressing His(6x)-Ub (kindly provided by Dr. M. Rodriguez) and 6 µg of pcDNA-Myc-USP25m, or pcDNA-Myc-USP25mC178S, either in their full length version or with the UBD-deletion mutants (ΔUBA, ΔUIM1, ΔUIM2, ΔUBAUIM1, ΔUBAUIM1UIM2 and ΔUIM1UIM2), or the C- terminal (679X, 769X and 1020X) deletion mutants, using Lipofectamine 2000 (Invitrogen). Forty-eight hours post-transfection, cells were washed with PBS and resuspended in 1.4 ml of denaturing lysis buffer pH 8.0 (50 mM sodium-phosphate buffer pH 8.0, 8 M urea, 300 mM NaCl, 0.5% Triton X-100, with 10 mM iodoacetamide and 10 mM NEM, freshly added) and stored at −80°C. Cleared cell lysates were loaded onto 8% SDS-PAGE gels and analyzed by Western blotting with anti-cMyc monoclonal antibody (1∶1000, Santa Cruz Biotechnology).
For the Ni2+ pull-down assay, cell lysates obtained as described above, were incubated with 80 µl of His-Select Nickel Affinity Gel (Sigma-Aldrich) during 3 h at room temperature. After 3 washes with the following buffer at pH 6.3 (50 mM sodium-phosphate buffer pH 6.0, 8 M urea, 300 mM NaCl), samples were eluted by boiling 5 minutes in 100 µl of protein-loading buffer (60 mM TrisHCl pH 6.8, 10% glycerol, 2% SDS, 0.1% bromophenol blue and 10% β-mercaptoethanol) and loaded onto 8% SDS-PAGE gels. After blotting, the proteins were detected by Western as stated above.
For further assessment of ubiquitination, cell lysates were incubated at 4°C with 2 µg of anti-cMyc mAb (Santa Cruz Biotechnology) during 4 hours with end-over-end mixing. The protein-antibody complexes were removed by one hour incubation at 4°C with protein G-Sepharose beads (Amersham GE-Healthcare). After thorough washing, bound proteins were eluted from the beads by boiling 5 min with protein loading buffer and loaded onto 8% SDS-PAGE gels. Bands were excised after Coomassie-Blue R250 staining and trypsinized. Tryptic peptides were analyzed in MALDI-TOF/TOF (4700 Proteomics Analyzer, Applied Biosystems) and/or in LC-ESI-QTOF (Q-TOF Global, Micromass-Waters) mass spectrometers and submitted using a MASCOT database search engine against non-redundant NCBi or SwissProt databases.
HEK293T cells were seeded on 24-well plates (2×105 cells/well). After 12 hours, cells were transiently co-transfected with constructs expressing HA-MyBPC1 and GFP-USP25m (full-length, or the corresponding deletion mutants), using Lipofectamine 2000 (Invitrogen). When stated, the proteasome inhibitor MG132 (10 µM, Sigma) was added to the medium during the last 16 hours of culture and collected 48 hours postransfection. Inhibition of new protein synthesis was achieved by adding cycloheximide (CHX, 150 µmol/ml, Sigma) to the medium 30 h postransfection and cells were collected immediately or after 4, 16 or 24 hour treatment. Cells were washed with PBS and recovered with 250 µl of protein loading buffer. Samples were loaded onto 8% SDS-PAGE gels and analyzed by western blotting using anti-HA monoclonal antibody (1∶1000, Santa Cruz Biotechnology) and anti-GFP polyclonal antibody (1∶1000, Santa Cruz Biotechnology) to assess the expression levels of MyBPC1 and USP25m, respectively. Films were scanned and quantified using QuantityOne software (Bio-Rad).
In silico predictions of functional domains and secondary structure of USP25m. Localization of the predicted Ubiquitin Binding Domains (one UBA and two UIMs), the catalytic deubiquitinating domains (USP), the peptides encoded by the muscle-specific alternatively spliced exons (19a and 19b), and several potential sumoylation sites and phosphorylation sites. In silico searches used the InterPro (
(6.61 MB TIF)
UBDs do not alter USP25m subcellular localization. USP25m localization was monitored by immunohistochemistry using a polyclonal antibody against USP25. Localization of full length USP25m and deletion mutants is predominantly cytosolic, with certain accumulation in the perinuclear region. Transfection of full length USP25m, or the deletion mutants, does not affect distribution of Ub, as assessed by immunodetection with an anti Ub antibody.
(2.19 MB DOC)
USP25 is sumoylated, phosphorylated and acetylated. A. USP25m is sumoylated. USP25m and all the UBD deletion mutants display an extra higher molecular weight band (asterisk) after in vitro sumoylation assays with SUMO-1 (middle lanes) and SUMO-2 (right lanes). In the case of USP25m lacking both UBA and UIM1, the band corresponding to SUMO-USP25m is weaker (two asterisks). Note that the absence of all three UBDs rendered similar levels of USP25m sumoylation to that of the full-length protein. B. USP25m is phosphorylated. Myc-tagged USP25m and USP25mC178S were immunoprecipitated with Myc antibodies and detected in Western blots with pan-anti-Phospho-Ser and pan-anti-phospho-Tyr. Bands appearing at the size corresponding to USP25m indicate that USP25m is phosphorylated both in serine(s) and threonine(s) (1st and 2nd panel, middle lane). This band also appears when expressing USP25mC178S, indicating that USP25m phosphorylation occurs irrespectively of its catalytic activity (1st and 2nd panel, right lane). Membranes were stripped and detected with a Myc antibody to confirm that the band corresponded to USP25m (3rd panel). Immunoprecipitation inputs were assessed with antibodies against phosphorylated AKT and Myc as phosphorylation and transfection controls respectively (4th and 5th panels). C. USP25m is acetylated. Myc-tagged USP25m and USP25mC178S were immunoprecipitated with Myc antibodies and detected in Western blots with pan-anti-acetylated-Lys. Bands appearing at the USP25m size indicate it is acetylated, both WT and C178S (upper panel). The same membrane was stripped and detected with anti-Myc to confirm the identity of the bands (2nd panel). Immunoprecipitation inputs were assessed with antibodies against acetylated p53, Myc and α-Tubulin as acetylation, transfection and loading controls, respectively (3rd, 4th and 5th panels).
(1.01 MB DOC)
We are indebted to Dr. M.S. Rodríguez and his group from CIC BioGUNE (Bilbao, Spain) for providing reagents and helpful suggestions. We thank the Serveis Científico-Tècnics de la Universitat de Barcelona for the Sequencing and Confocal microscopy facilities, and particularly, to the Proteomics platform, a member of ProteoRed.