Hyaluronan- and RNA-binding deubiquitinating enzymes of USP17 family members associated with cell viability

Cloning of USP17 subfamily members in human chorionic villi and embryonic carcinoma cell lines

To identify putative deubiquitinating enzymes that contain a conserved Asp (I) domain in humans, the GenBank database was accessed and searched using the BLAST algorithm at the NCBI, as described previously [22]. In this study, we obtained multiple cDNAs including DUB-3 [21] and USP17 [23] from human chorionic villi tissues and various cancer cell lines by RT-PCR. We performed RT-PCR, sequenced PCR products three times independently in order to exclude the possibility of PCR errors, and classified the genes as USP17 subfamily members (USP17K, USP17L, USP17M, and USP17N) due to the high homology to known sequences [20, 21]. To characterize the genetic properties of human USP17 genes, the complete nucleotide sequences of USP17K, USP17L, USP17M, and USP17N [GenBank: AF544011, AF544012, AY188990, and AY533200, respectively] were analyzed. Nucleotide and deduced amino acid sequences were analyzed with the SeqEd V1.0.3 program and Clustal method of the MegAlign program, a multiple alignment program of the DNASTAR package. The full-length cDNAs for USP17K, USP17L and USP17M consist of 1593 nucleotides in open reading frame and the deduced amino acid sequence of 530 amino acids with a calculated molecular mass of 58 kDa (Additional File 1). The full-length cDNA for USP17N consists of 1197 nucleotides in open reading frame and the deduced amino acid sequence of 398 amino acids with a calculated molecular weight of 44.5 kDa (Additional File 1). The conserved Cys, His and Asp domains, which are thought to form the active site of the deubiquitinating enzymes, are also found in the USP17K, USP17L, USP17M and USP17N (Fig. 1); these conserved domains have been previously shown in other deubiquitinating enzymes including USP36 [24], DUB-1 [25], and DUB-2 [26]. Sequence analysis for amino acids of novel USP17 subfamily members indicates that USP17K, USP17L and USP17M have high degree of homology (>95%), and USP17N is identical with USP17C when compared with each other and USP17 family members (Table 1). Phylogenetic analysis for nucleotide sequences and deduced amino acid sequences of novel genes revealed that all novel genes were grouped into the USP17 subfamily (Fig. 2) and they were characterized in this report.

Examination of the deduced amino acid sequence of these enzymes by Pfam [27] revealed that novel subfamily members of USP17 contain the putative conserved hyaluronan binding motifs (Fig. 3) -(R/K)X₇(R/K)- [28]. These putative motifs are located between amino acid residues Arg 401 to Lys 410, and Lys 445 to Lys 453 (Fig. 3). In addition, "RGG" amino acid sequences predicted for the novel RNA binding motif are found at the C-terminus. Interestingly, USP17N lacks of these binding motifs (Fig. 3), suggesting the unique functional roles for these motifs. Even though it has been suggested that human USP17 subfamily members are homologous to murine DUB subfamily members due to the fact that they are cytokine-inducible and regulate cell growth and survival [20], hyaluronan and RNA binding motifs are not found in members of murine DUB subfamily (data not shown).

Expression of transcripts for novel USP17 subfamily members in human tissues and cancerous cells

To investigate the expression pattern of transcripts for novel USP17 subfamily members in human tissues and cancerous cells, Northern blot and RT-PCR analyses were performed. As shown in Figure 4A, Northern blot analysis revealed that transcripts for novel USP17 subfamily members were expressed in the heart, liver, and pancreas at a high level. Interestingly, we obtained multiple bands having different sizes (1.7 kb, 1.6 kb, 1.3 kb, 1 kb) using a 5' 500 bp probe of USP17L (Probe: N-500). We also detected 1.6 kb and 1.7 kb major bands of USP17 subfamily members using a probe designed for a 3' 1 kb probe of USP17L (Probe: C-1000) (Fig. 4A). It is possible that other minor mRNA transcripts around 1.3 kb and 1 kb, which have not been detected with DUB-3 probe [21], may be derived from alternative splicing, promoter usage, or poly (A+) usage. The presence of a few different sizes of transcript indicates that novel genes belong to the USP17 subfamily and that there are at least several members of this family present. In addition to various human tissues, RT-PCR analysis showed that the expression of transcripts for novel USP17 subfamily members was detected in various human cancer cell lines including HeLa, HEC-1A, NCCIT, and fibrosarcoma cells (Fig. 4B).

Novel USP17 subfamily members as a deubiquitinating enzyme

To determine whether novel USP17 subfamily proteins have ubiquitin specific protease enzymatic activity, deubiquitinating enzyme assays were performed. For in vitro enzyme assay, novel USP17 subfamily members were expressed as a GST fusion protein. As shown by immunoblotting analysis (Fig. 5), four GST fusion proteins demonstrated a strong efficient cleavage of Ub-Met-β-gal (Fig. 5; lanes 2, 4, 6, and 8). We generated the Cys conserved domain mutant for each of novel USP17 subfamily proteins (C89S) using a site-directed mutagenesis method. As expected, cells with the empty vector alone (Fig. 5, lane 1) and mutants (Fig. 5; lanes 3, 5, 7 and 9) failed to cleave Ub-Met-β-gal.

For in vivo deubiquitinating enzyme assay, HA-tagged pMT123-ubiquitin was transiently expressed in human endometrial carcinoma cells (HEC-1A) with or without pcDNA3-myc-USP17L and pcDNA3-myc-USP17L (C89S). Then, we analyzed deubiquitinating enzyme activity by immunoblotting using a monoclonal anti-HA antibody. The expression bands of polyubiquitinated proteins were nearly invisible in cells expressing pcDNA3-myc-USP17L (Fig. 6; lane 4), but not in cells expressing pcDNA3-myc-USP17L (C89S) (Fig. 6; lane 5). The molecular weight for a mutant form of USP17L (C89S) is bigger than wild type USP17L, suggesting that there may be a conformational change due to the amino acid substitution. In addition, other USP17 subfamily members showed the deubiquitinating enzyme activity in vivo (data not shown). Taken together, we concluded that novel USP17 subfamily proteins have deubiquitinating enzyme activity.

Hyaluronan/RNA-binding properties of USP17 subfamily proteins

Since it is not known whether hyaluronan-binding sites within USP17 subfamily proteins are indeed functional, we further examined whether USP17 subfamily proteins actually binds hyaluronan. The cetylpyridinium chloride (CPC)-precipitation assay has been used for investigating binding affinity [29] and the assay in our study demonstrated that USP17 subfamily proteins interact with hyaluronan based on the comparison of binding intensity between endogenous and exogenous hyaluronans (Fig. 7A, lanes 4 and 5). However, USP17N did not interact with hyaluronan (Fig. 7A, lane 5), indicating that the C-terminal domain of USP17 subfamily proteins is required for binding with hyaluronan.

Since it has been reported that the HABP4_PAI-RBP1 domain, which is involved in mRNA stability [30], interacts with both hyaluronan and RNA [17], we additionally investigated the activity of USP17 subfamily proteins for binding with RNA. We performed the CPC-precipitation assay and found that USP17 subfamily proteins also bind to RNA in a dose-dependent manner (Figs. 7A and 7B). Interestingly, USP17N did not interact with RNA as shown with hyaluronan (Fig. 7A, lane 6), indicating that the C-terminal domain of USP17 subfamily proteins is also required for binding with RNA.

Localization study of USP17L

In order to investigate the intracellular localization of USP17 subfamily proteins, the GFP-tagged USP17L or RFP-tagged USP17N, or both were transfected into HeLa cells. Intriguingly, USP17L proteins are found in the nucleus and are concentrated at the nucleoli (Figs. 8E and 8H) similar to intracellular hyaluronan [31]. However, USP17N proteins are localized both in the nucleus and cytoplasm (Figs. 8F and 8I). A merged format shows that USP17L and USP17N are localized in different regions (Figs. 8G and 8J). After 48 hours of transfection, cells expressing GFP-tagged USP17L showed the onset of irregular nuclear division and irregular morphology of overall cell shape compared with controls (data not shown). In contrast, cells expressing RFP-tagged USP17N did not show any sign of apoptosis, suggesting that hyaluronan binding ability and the localization of USP17 subfamily proteins may be critical for regulating apoptosis leading to cell death.

Regulation of cell viability

Since it has been suggested that intracellular hyaluronan binding proteins are involved in tumor formation [32, 33], we next investigated whether the USP17 subfamily proteins are involved in the regulation of cell viability and cell cycle for cancer cells. In order to investigate the cellular roles of USP17 subfamily proteins, myc-tagged USP17L was transfected into HEC-1A and HeLa cells to establish stable cell lines. However, we observed transfected cells undergoing apoptosis (Fig. 9) and approximately 3 weeks later, all cells died. However, the transfection of either USP17L (C89S) or USP17N into the same cell lines did not result in any sign of apoptosis (Fig. 9). This indicates that USP17 subfamily proteins containing hyaluronan binding motifs and their substrate(s) may regulate cell viability. Based on the observation of cell death for USP17L-transfected cells, we performed quantitative real-time PCR to compare fold change for the expression of caspase 3 as an indicator of apoptosis. The expression of caspase 3 showed more than two fold increase in USP17L-transfected cells compared with controls (Fig. 10). However, the expression of caspase 3 in USP17L (C89S)-transfected cells and in USP17N-transfected cells was similar to one in controls (Fig. 10). All reactions were performed triplicate.

In order to confirm the apoptosis induced by USP17 subfamily proteins, USP17L was stably expressed by the Tet-on system. The morphology of cells undergoing apoptosis was similar to that of ones transfected with pcDNA3-myc-USP17L (data not shown). Therefore, we harvested the USP17L-expressing cells at 1 week intervals for 3 weeks and analyzed through FACS analysis. The rate of apoptosis for cells expressing either USP17L or USP17L (CS) was analyzed by double-staining with annexin V and PI (Fig. 11). The cell stages of apoptosis were differentiated depending on the location in the quadrant (annexin V +, PI -; early stage) and (annexin V +, PI +; late stage) [34]. The USP17L-expressing cells by the Tet-on system undergoing the early stage apoptosis showed about 2.5 fold increase compared with control HeLa cells and those undergoing late stage apoptosis showed about 1.5 fold increase (Fig. 11). However, the significant change was not observed in cells expressing USP17L (CS) or USP17N (Fig. 11). This experiment was performed triplicate.

Specimens and transfection

Chorionic villi samples used were described previously [36]. Various human cancer cells (HeLa, HEC-1A, NCCIT, and fibrosarcoma cells) were cultured in DMEM (Gibco BRL, Rockville, MD, USA) supplemented with 10% fetal bovine serum (Gibco BRL, Rockville, MD, USA). Transfection was performed with LipofectAMINE reagent (Gibco BRL, Rockville, MD, USA) according to manufacturer's instruction.

cDNA synthesis and cloning for USP17 subfamily members

TRIzol reagent (Gibco BRL, Rockville, MD, USA) was used for total RNA isolation from obtained chorionic villi, HeLa, HEC-1A, NCCIT, and fibrosarcoma cells. cDNA synthesis from obtained total RNA was in accordance with the protocol of SuperScript Preamplification system (Gibco BRL, Rockville, MD, USA). The full-length cDNAs were amplified by RT-PCR in human chorionic villi tissues, HeLa, HEC-1A, NCCIT, and fibrosarcoma cells. Primers (Forward: 5'-CGGGATCCATGGAGGACG-3', Reverse: 5'-GGAATTCCACTGAGATCACT-3') used for isolating putative DUB genes are made on the basis of published sequences [GenBank: XM_252830]. Primers used were synthesized commercially and gel-purified prior to use (Bioneer, Daejeon, Korea). To express GST fusion proteins, pGEX-2TK-USP17K, pGEX-2TK-USP17L, pGEX-2TK-USP17M, and pGEX-2TK-USP17N were cloned by inserting each USP17 subfamily member into pGEX-2TK vector (Amersham Life Science, Uppsala, Sweden). BLAST programs on the web site established by the National Center of Biological Information [37] were used for similarity searches in the GenBank, and Pfam [27] program was used for prediction of possible motifs and domains in the protein sequences.

Site-directed mutagenesis of USP17 subfamily members

Missense mutation constructs for pGEX-2TK-USP17K (C89S), pGEX-2TK-USP17L (C89S), pGEX-2TK-USP17M (C89S), and pGEX-2TK-USP17N (C89S) were derived from each wild type construct by PCR-based site-directed mutagenesis, using a QuikChange™ Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA, USA) according to manufacturers' instructions. The two primers; forward primer (5'-GGGAAATACCCGCTACGTGAACGCTTC-3') and reverse primer (5'-GAAGCGTTCACGTAGCGGGTATTTCC-3') were used for replacing cysteine with serine at amino acid 89 of USP17 subfamily proteins. Primers used were synthesized commercially and gel-purified prior to use (Bioneer, Daejeon, Korea).

Northern blot analysis

As described previously [36], Northern blot analysis was carried out to confirm the expression level of USP17 transcripts using a human tissue blot purchased from Clontech. The immobilized nucleic acids were hybridized with radiolabeled DNA probes. Probes used for this study was the 5' 500 bp and 3' 1 kb of USP17L. They were labeled by random primers with [³²P]dCTP (Amersham Biosciences, Buckinghamshire, England). Hybridization was performed in a bag containing 5 × SSC, 1 × Denhardt's solution, 100 μg/ml of denatured salmon sperm DNA, 0.1% SDS and 50% formamide at 60°C and it was washed in 0.1 × SSC and 0.2% SDS, the membrane was exposed to Kodak X-ray films with an intensifying screen for 24 hours at -70°C.

Deubiquitinating enzyme assays

Deubiquitinating enzyme assays were previously described [9]. For the in vitro deubiquitination assay, pGEX-2TK-USP17K, pGEX-2TK-USP17K (C89S), pGEX-2TK-USP17L, pGEX-2TK-USP17L (C89S), pGEX-2TK-USP17M, pGEX-2TK-USP17M (C89S), pGEX-2TK-USP17N, and pGEX-2TK-USP17N (C89S) were co-transformed into E. coli BL21 competent cells (Stratagene, La Jolla, CA, USA) with pACYC184-Ub-β-gal. Plasmid-bearing E. coli BL21 cells were grown in LB medium and induced with isopropyl-1-thio-β-D-galactopyranoside (IPTG) for 4 hours at 37°C. Cells were harvested and were lysed in cracking buffer (0.01 M phosphate [pH 7.4]. 8 M Urea, 1% SDS, and 1% β-mercaptoethanol). These whole cell lysates centrifuged for 10 min at 14,000 rpm and the supernatant was loaded onto a 7.5% SDS-PAGE gel. The immunoblotting was performed with a rabbit anti-β-gal antiserum (ICN, Costa Mesa, CA, USA) or a rabbit anti-GST antiserum (Upstate Biotechnology, Lake Placid, NY, USA).

The effect of USP17 expression on the ubiquitin-proteasome system in vivo was analyzed by transfecting pcDNA3-myc-USP17L or pcDNA3-myc-USP17L (C89S) along with pMT123-HA-ubiquitin into HEC-1A cells. After 24 hours of transfection, cells were harvested and lysed. 40 μg of total proteins was loaded in each lane of a 10% SDS-PAGE for immunoblotting analysis using an anti-HA antibody (Roche Applied Science, Mannheim, Germany). Equal loading was verified by immunoblotting against α-myc (Santa Cruz Biotechnology, Santa Cruz, CA, USA).

Real-time PCR analysis and the Tet-on system

After the transfection of pcDNA3-myc-USP17L, the expression of caspase 3 as a marker for apoptosis was analyzed by real-time PCR (Opticon2, MJ Research, Ramsey, MN, USA). Product specificity was examined by melting curve analysis and agarose gel electrophoresis after each real-time PCR reaction. To normalize the expression data of the genes, GAPDH was used (Forward primer: 5'-ACTGGTGCTGCCAAGGCTGT-3', Reverse primer: 5'-TCCACCACCCTGTTGCTGTA-3'). Quantification was performed using an internal cDNA standard curve. Fold changes of caspase 3 expression in USP17L-overexpressed cells were calculated relative to the levels in control cells.

For the long-term (3 weeks) observation and FACS analysis, USP17L was expressed in HeLa cells using the Tet-on system (Gibco BRL, Rockville, MD, USA) by manufacturer's instruction. Expression of USP17L was confirmed by real-time PCR analysis.

FACS analysis

For apoptosis analysis, HeLa-Tet-on cells expressing USP17L were harvested in PBS and resuspended 100 μl of binding buffer (10 mM Hepes/NaOH, pH 7.4, 140 mM NaCl, 2.5 mM CaCl₂). 5 μl annexin V-FITC (BD Biosciences, Bedford, MA, USA) and 10 μl propidium iodide (PI, 50 μg/ml, BD Biosciences, Bedford, MA, USA) were added and incubated 15 minutes in the dark. Finally 400 μl of binding buffer was added and analyzed by FACSVantage™ SE system (BD Biosciences, Bedford, MA, USA). Cells in stage of apoptosis were differentiated depending on the location in quadrant (annexin V +, PI -; early stage) and (annexin V +, PI +; late stage) [34].

Localization study

USP17L and USP17N were subcloned into pEGFP-C1 and pDsRed-C1 (Clontech, Palo Alto, CA, USA), respectively, and transfected into HeLa cells using ExGen500 (Fermentas, Hanover, MD, USA). After 48 hours, the localization of GFP-tagged USP17L and RFP-tagged USP17N was observed using a confocal microscopy (Radiance 2100 2Q, Bio-Rad, Hercules, CA, USA). In order to observe the morphology of nuclei for HeLa cells, DAPI staining was performed. After the fixation, 5 μl of DAPI staining solution (Roche Applied Science, Mannheim, Germany) was added in wells containing HeLa cells, which express either GFP-tagged USP17L or RFP-tagged USP17N. Stained cells were observed using Nikon Eclipse E600 microscopy.

Hyaluronan- and RNA-binding assays

These assays were previously described [29]. 9 μg of pcDNA3-myc-USP17L or pcDNA3-myc-USP17N was transiently transfected into 1 × 10⁶ HeLa cells in 100Φ dish by ExGen500 transfection reagent (Fermentas, Hanover, MD, USA). The supernatant from lysates was divided into 100 μl aliquots in Eppendorf tube and incubated at room temperature with 50 μg of hyaluronan (H5388, Sigma, St. Louis, MO, USA). For RNA binding assay, this procedure was repeated with using total RNA of HeLa cells instead of hyaluronan.