Identification of a novel isopeptidase with dual specificity for ubiquitin- and NEDD8-conjugated proteins.

Covalent conjugation of proteins by ubiquitin or ubiquitin-like molecules is an important form of post-translational modification and plays a critical role in many cellular processes. Similar to the concept of phosphorylation and dephosphorylation, these conjugates are regulated by a large number of deconjugating enzymes. Here, we report the cloning of a 2,141-base pair DNA fragment from human placenta cDNA library by a strategy that involves expressed sequence tag data base searching, polymerase chain reaction, and rapid amplification of cDNA ends. Nucleotide sequence analysis revealed that the cloned cDNA contains an open reading frame of 1,143 base pairs encoding a novel protease, USP21, which is composed of 381 residues with a calculated molecular mass of 43 kDa. The human USP21 gene is located on chromosome 1q21 and encodes a member of the ubiquitin-specific protease family with highly conserved Cys and His domains. The activity and specificity of USP21 were determined by using a COS cell expression system in vivo. We showed that USP21 is capable of removing ubiquitin from ubiquitinated proteins as expected. Furthermore, USP21 is capable of removing NEDD8 from NEDD8 conjugates but has no effect on Sentrin-1 conjugates. As expected from its biochemical activity, overexpression of USP21 has a profound growth inhibitory effect on U2OS cells. Thus, USP21 is the first ubiquitin-specific protease shown to have dual specificity for both ubiquitin and NEDD8 and may play an important role in the regulation of cell growth.

(E1) to form a high energy E1-Ub thiol ester. The activated Ub is then transferred to the SH group of a Cys residue in the active site of a Ub-conjugating enzyme (E2 or Ubc). Finally, in the presence of a Ub-protein ligase (E3), the C-terminal Gly residue of Ub is conjugated via an isopeptide bond to the ⑀-amino group of a Lys residue on the target protein. Sequential conjugation of Ub to a previously conjugated Ub by linking the ⑀-amino group of one Ub lysine residue to the C terminus of another Ub results in the formation of polyubiquitin chain, which targets proteins for degradation by the 26 S proteasome (3,4). Regulation of these processes requires precise timing and specific recognition of different substrates at appropriate cellular milieu, thought to be mediated by a combinatorial usage of different E2 and E3 enzymes and by a large family of deubiquitinating enzymes (5,6).
Deubiquitinating enzymes are ubiquitin-specific thiol-proteases that cleave either linear ubiquitin precursor proteins or ubiquitin conjugates. Some deubiquitinating enzymes appear to perform an editing function, which controls the fidelity of the conjugation process, thus preventing inappropriate degradation of cellular proteins. Two families of deubiquitinating enzymes have been identified on the basis of in vitro activities and/or sequence identity. The ubiquitin C-terminal hydrolases (UCHs) are small (approximately 25 kDa) thiol proteases that share amino acid sequence identity and cleave esters and amids from the C terminus of ubiquitin (5). Unrelated in sequence to the UCHs are the ubiquitin-specific proteases (UBPs), a large family of proteins differing greatly in length but characterized by sequence similarity in several regions: the Cys box, the His box, and six other blocks of amino acid sequence identity (5).
In recent years, ubiquitin-like proteins, such as UCRP/ ISG15, Sentrin-1, and NEDD8, have greatly expanded the scope of post-translational protein modification (7)(8)(9)(10)(11)(12)(13). These ubiquitin-like proteins are translated in precursor form, with one or more amino acids following a Gly-Gly dipeptide that forms the C terminus of the mature protein. The proteases specific for Sentrin-1 and Smt-3 are structurally different from the Ubiquitin-specific UBPs and the UCHs (14,15). UCH-L3, originally identified as a ubiquitin-specific hydrolase, has been shown by our laboratory to possess a C-terminal hydrolase activity for NEDD8 (8,16,17). However, the isopeptidase responsible for NEDD8 deconjugation has not been identified. Here, we report the identification of a novel protease, USP21 (ubiquitin-specific protease 21), which is not only specific for ubiquitin conjugates but also functions as a NEDD8-specific isopeptidase.

EXPERIMENTAL PROCEDURES
Identification of USP21-To clone USP21, we used a similar strategy as previous described (14,18,19). We first performed a TBLASTN search of the human EST data base using the conserved sequences previously identified as the Cys and His boxes (5). This search identified * This work was supported in part by National Institutes of Health Grants GM-57502 and DK-56298, the DREAM project, and an American Heart Association Established Investigator award. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF233442.
‡ To whom correspondence should be addressed: Div. of Molecular Medicine, Dept. of Internal Medicine, University of Texas-Houston Health Science Center, 6431 Fannin, Suite 4.200, Houston, TX 77030. Tel.: 713-500-6660; Fax: 713-500-6647; E-mail: eyeh@heart.med.uth. tmc.edu. 1 The abbreviations used are: Ub, ubiquitin; E1, ubiquitin-activating enzyme; E2, ubiquitin-conjugating enzyme; E3, ubiquitin-protein ligase; UCH, ubiquitin C-terminal hydrolase; UBP, ubiquitin-specific protease; EST, expressed sequence tag; PCR, polymerase chain reaction; RACE, rapid amplification of cDNA ends; RH, RGS-His. 59 positive EST clones. After further analysis, three EST clones (R25480, F11440, and W79945) were found to have highly conserved Cys box sequences differing from known UBPs. Further search of the human EST data base using the nucleotide sequence of R25480 revealed two additional EST clones (D31406 and DS1281) that partially overlap with R25480. PCR amplification was then used to confirm that both D31406 and R25480 were from the same cDNA fragment. RACE was performed to extend the cDNA sequence toward the 5Ј direction. Finally, using a pair of redesigned 5Ј and 3Ј primers compatible with the sequence revealed in the 5Ј extension and cloned PCR fragment, we amplified a 2195-base pair cDNA fragment by PCR reaction.
PCR, 5Ј-RACE, and Sequence Analysis-The nested primers were synthesized on the basis of the information obtained from the positive EST clones. These primers were used to amplify the novel protease gene fragments by PCR from a human placenta cDNA library. Both PCR and RACE were performed as described previously (14,18,19). The nucleotide sequences were determined using dye terminator sequencing and an automated sequencer from Applied Biosystems Inc. (Foster City, CA).
Northern Blot Analysis of Human USP21-To study the expression pattern of the USP21 gene, a commercially available Northern blot (CLONTECH, Palo Alto, CA) containing ϳ2 g of poly(A) ϩ RNA from eight different tissues was used. The full-length coding region for UNSP1 was used as template. The cDNA probe was labeled with [␣-32 P]dCTP using the random primers DNA labeling system (Life Technologies, Inc.), and unincorporated nucleotide was removed by chromatography on a NucTrap column (Stratagene, La Jolla, CA). Prehybridization and hybridization were performed under conditions recommended by the supplier. The blot was dried and autoradiographed with an x-ray film and intensifying screen at Ϫ70°C for 3 days.
Transfection of COS-M6 Cells-COS-M6 cells were grown to approximately 80% confluency in Dulbecco's modified Eagle's medium containing 2 mM L-glutamine, 1% penicillin/streptomycin, and 10% fetal bovine serum at 37°C. Cells were transfected using LipofectAMINE (Life Technologies, Inc.). 5 g of plasmid and 25 g of lipid were added to 500 l of Dulbecco's modified Eagle's medium (serum-and antibiotic-free) and then incubated for 30 min at room temperature. After this time, the lipid/DNA mix was diluted 5-fold with Dulbecco's modified Eagle's medium (containing 10% fetal bovine serum, 1% antibiotics) for 2 days.
Western Blot Analysis-The cells were harvested in lysis buffer, and the proteins were separated on a 15% acrylamide gel and subsequently transferred to a nitrocellulose polyvinyl difluoride membrane. After blocking for 1 h in 5% skim milk, the filters were incubated with anti-RH antibody (used at dilution 1:4000; Qiagen) for 1 h. After five 5-min washes in TBST (0.5 N NaCl, 20 mM Tris, pH 8.0, 0.1% Tween 20), filters were incubated for 1 h at 25°C with goat anti-mouse secondary antibody conjugated to horseradish peroxidase (1:2000). Blots were developed by the ECL system (Amersham Pharmacia Biotech) according to the manufacturer's instructions.
Site-directed Mutagenesis-The cysteine residue of USP21 that lies within the conserved UBP Cys domain (Cys 37 ) was mutated to alanine using deoxyoligonucleotides and the Stratagene QuikChange TM sitedirected mutagenesis kit according to the manufacturer's specifications. The plasmid pcDNA3-USP21 was used as a template. Mutation was confirmed through DNA sequencing.
Cell Growth Assay-U2OS cells (8 ϫ 10 5 ) were plated in a 10-cm dish and transfected by LipofectAMINE either with 5 g of control empty pcDNA3 vector or of the same vector containing the USP21 sequence in the same orientation. After 24 h, the cells were washed twice with phosphate-buffered saline and incubated with fresh 10% serum-containing medium. After incubation for another 24 h, cells were washed twice with phosphate-buffered saline, rinsed off the plates with 0.05% trypsin, 0.2% EDTA in phosphate-buffered saline, collected in 10% serum-containing medium, and then split at different dilutions and subcultured into medium containing 500 g/ml G418. After 14 days of incubation drug-resistant colonies were counted after trypan blue staining.

RESULTS AND DISCUSSION
Molecular Cloning of the Human USP21 cDNA-Using a combined PCR cloning and EST data base search technique described previously (14,18,19), a 2141-base pair cDNA fragment was cloned from a human placenta cDNA library. The nucleotide and deduced amino acid sequence of the cloned cDNA are shown in Fig. 1. The cloned cDNA has the following features: 1) a translation initiation codon (ATG) is located at nucleotide positions 727-729, preceded by a 726-base pair 5Јuntranslated sequence. The ATP triplet is preceded by an inframe stop codon at nucleotide positions 514 -516. (2) A translation termination codon TGA is located at nucleotide positions 1869 -1872, followed by a 3Ј-untranslated region of 269 nucleotides. At the 3Ј terminus a potential polyadenylation signal, AATAAA, is present at nucleotide 2128. (3) The cloned fragment encodes a 381-amino acid polypeptide, called USP21 (approved by the Human Genome Nomenclature Committee). The calculated molecular mass for USP21 is 43,031 Da. The amino acid sequence of USP21 was compared with sequences in several data bases using the BLAST network service at the National Center for Biotechnology Information. The results indicated that USP21 is a novel member in the ubiquitin-specific protease family as shown in Fig. 2. Similar to DUB-1, a deubiquitinating enzyme with growth-suppressing activity, USP21 contains all eight conserved sequence motifs originally identified in the yeast ubiquitin-specific proteases (5). These domains include a conserved cysteine residue and two histidine residues, respectively, that presumably form part of the active site of these thiol proteases (Fig. 2).
Expression of USP21 Transcripts in Human Tissues-To study the expression of USP21, we designed a probe that has little homology to the other UBP genes. On a Northern blot analysis, a single transcript of approximately 2.2 kilobases was detected in all the tissues, indicating that the isolated cDNA was likely to be full length (Fig. 3). In addition, USP21 was expressed at different levels in the tissues examined. The highest level of USP21 expression was found in the heart, pancreas, and skeletal muscle. USP21 messages could also be detected in the brain, placenta, liver, and kidney but are very low in the lung.
Chromosomal Localization of the Human USP21 Gene by EST Mapping-The human USP21 cDNA sequence was used to search the human EST data base using the BLASTN program, and 42 positive EST clones were identified. All of these EST clones were used to check the Human Gene Map data base. 11 of them (DDBJ/EMBL/GenBank TM accession numbers T33337, T33359, T16677, R54635, H42874, W74194, R45549, R56564, H17877, N64752, and R76797) have been mapped to chromosome 1 between D1S484 and D1S426 microsatellite markers at 173-181 centimorgans (National Center for Biotechnology Information). Thus, the USP21 gene is located in chromosome 1q21.
Comparison between USP21 with DUB, Another UBP Family Member-We first tested the activity of USP21 in vivo using DUB-1 as a comparison. DUB-1 contains all eight conserved sequence motifs as shown in Fig. 2A, and the glutathione S-transferase-DUB-1 fusion protein has been shown to cleave ubiquitin from a ubiquitin-␤-galactosidase protein in vitro (20). A COS cell expression system was used to demonstrate the activity of USP21 in vivo. Briefly, plasmid containing Histagged ubiquitin was introduced into COS-M6 cells by liposome-mediated transfection as described previously (8,9). Total cell lysates were prepared 16 h after transfection for Western blot analysis using anti-RH antibody. As shown in Fig.  4, lysates prepared from COS cells expressing RH-tagged Ub revealed a ladder of ubiquitin monomer, multimers, and ubiquitin-conjugated proteins (lanes 2 and 5). When RH-tagged Ub was co-expressed with HA-tagged USP21 in the COS cells, the higher molecular weight ubiquitin conjugates were completely removed (lane 3). The disappearance of the high molecular weight ubiquitin conjugates also coincided with the accumulation of either a 38-kDa band (lane 3) or free ubiquitin monomers (lane 6). However, when RH-tagged Ub was co-expressed with HA-tagged DUB-1 in COS cells, no significant change in the pattern of ubiquitination was observed (lane 4). These data suggest that USP21 is capable of deconjugating most ubiquitinated proteins, whereas DUB-1 may be more limited in its substrate selection. Both USP21 and DUB-1 have a conserved Cys box (72% identity) as shown in Fig. 2. The Cys 60 in DUB-1 has been shown to be required for ubiquitin-specific thiol protease activity (20). When the conserved cysteine residue (Cys 37 ) of USP21 was mutated to alanine (C37A), the activity of USP21 was completely abolished (Fig. 4, lane 7), suggesting that the activity of USP21 is critically dependent on the conserved cysteine residue, as are other UBP family members.
USP21 Is Specific for NEDD8, but Not for Sentrin Conjugates-We next tested the activity of USP21 against other ubiquitin-like proteins using the COS cell expression system described above. As shown in Fig. 5, when RH-tagged NEDD8 was expressed in COS cells, a series of bands in a ladder-like pattern representing NEDD8-conjugated proteins were observed (lanes 1 and 4). However, when RH-tagged NEDD8 was co-expressed with HA-tagged USP21 in the COS cells, the higher molecular weight NEDD8 conjugates were completely removed (lane 5). The disappearance of the high molecular weight NEDD8 conjugates also coincided with the accumulation of free NEDD8 monomers (lane 5). As expected, the (C37A)USP21 point mutant also lost its activity against NEDD8 conjugates (lane 6). Furthermore, USP21 has no effect on Sentrin conjugates, demonstrating that USP21 is not a generalized isopeptidase for all ubiquitin-like proteins (lanes 10 -12). These results are consistent with the previous observation that the Sentrin-specific protease, SENP1, is structurally different from UBPs (14). NEDD8 is 60% identical and 80% homologous to ubiquitin, whereas sentrin is only 18% identical and 48% similar to ubiquitin. Thus, it is not entirely surprising that NEDD8 and ubiquitin could share similar UBPs. To test whether other UBPs may also be active against NEDD8 conjugates, we tested the effect of DUB-1 on NEDD8 (lane 2) or Sentrin conjugates (lane 8). As shown in Fig. 5, DUB-1 has no activity against either NEDD8 or Sentrin conjugates. This observation suggests that USP21 is a unique UBP with dual specificity for NEDD8 and ubiquitin.
USP21 Inhibits U2OS Cell Growth-Naviglio et al. (21) have previously shown that UBPY(USP8), another UBP family member, inhibits U2OS cell growth. We asked whether USP21 also have similar biological property. As shown in Fig. 6, USP21 inhibited cell growth by up to 80% as compared with control in two separate experiments. In contrast, SENP1, a Sentrin-specific protease, only showed 15% inhibition of cell growth. These data suggest that USP21 has a profound effect on cell growth regulation similar to UBPY. In conclusion, USP21 is an novel member of the UBP family that appears to possess dual specificity for both NEDD8 and ubiquitin conjugates and may play important role in the regulation of cell growth.