The modification of proteins by SUMO (small ubiquitin-related modifier) plays important roles in regulating the activity, stability and cellular localization of target proteins. Similar to ubiquitination, SUMO modification is a dynamic process that can be reversed by SENPs [SUMO-1/sentrin/SMT3 (suppressor of mif two 3 homologue 1)-specific peptidases]. To date, six SENPs have been discovered in humans, although knowledge of their regulation, specificity and biological functions is limited. In the present study, we report that SENP7 has a restricted substrate specificity, being unable to process SUMO precursors and displaying paralogue-specific isopeptidase activity. The C-terminal catalytic domain of SENP7 efficiently depolymerized poly-SUMO-2 chains but had undetectable activity against poly-SUMO-1 chains. SENP7 also displayed isopeptidase activity against di-SUMO-2- and SUMO-2-modified RanGAP1 (Ran GTPase-activating protein 1) but had limited activity against SUMO-1-modified RanGAP1. in vivo, full-length SENP7 was localized to the nucleoplasm and preferentially reduced the accumulation of high-molecular-mass conjugates of SUMO-2 and SUMO-3 compared with SUMO-1. Small interfering RNA-mediated ablation of SENP7 expression led to the accumulation of high-molecular-mass SUMO-2 species and to the accumulation of promyelocytic leukaemia protein in subnuclear bodies. These findings suggest that SENP7 acts as a SUMO-2/3-specific protease that is likely to regulate the metabolism of poly-SUMO-2/3 rather than SUMO-1 conjugation in vivo.

INTRODUCTION

SUMOs (small ubiquitin-related modifiers) are covalently linked to lysine side chains in target proteins and confer altered properties on the modified proteins, which are involved in a variety of processes such as nuclear transport, transcription, recombination and chromosome segregation [14]. Three members of the SUMO family have been identified in vertebrates: SUMO-1 {also known in humans as SMT3C (suppressor of mif two 3 homologue 1 C), PIC1, GMP1 [GAP (GTPase-activating protein)-modifying protein 1], sentrin and Ubl1 (ubiquitin-like protein 1)}, SUMO-2 (also known as SMT3B and sentrin-3) and SUMO-3 (also known as SMT3A and sentrin-2). The mature forms of SUMO-2 and SUMO-3 are 97% identical in sequence with each other, but only 50% identical with SUMO-1 [5,6].

SUMOs are conjugated to target proteins through a cascade of reactions that typically involved three enzymes: an activating enzyme (E1), a conjugating enzyme (E2) and, usually, a SUMO ligase (E3) [7]. In the first step, E1 [a heterodimer containing SAE1 (SUMO-activating enzyme 1) and SAE2 subunits] utilizes ATP to adenylate the C-terminal glycine of SUMO. The formation of a thioester bond between the C-terminal glycine of SUMO and a cysteine in SAE2 is accompanied by the release of AMP. The second step is a transesterification reaction, in which SUMO is transferred from E1 to a cysteine residue within the SUMO-specific E2-conjugating enzyme, Ubc9 (ubiquitin-conjugating enzyme 9). Finally, the thioester-linked Ubc9–SUMO conjugate catalyses the formation of an isopeptide linkage between the C-terminal carboxyl group of SUMO and the ε-amino group of lysine in the substrate protein. Many sumoylation sites are found within a consensus motif ΨKXE/D, (where Ψ is a large hydrophobic amino acid, K is a lysine residue, X is any amino acid and E/D is a glutamic acid or aspartic acid residue), although modification at non-consensus sites has been reported [8]. A number of SUMO E3 enzymes such as Ran-binding protein 2 and the PIAS [protein inhibitor of activated STAT (signal transducer and activator of transcription)] family of proteins have been identified (reviewed in [9]). These enzymes can enhance the conjugation reaction and are likely to provide substrate specificity in the transfer of the SUMO to the lysine residue of target proteins. SUMO-2 and SUMO-3, similar to Smt3p, contain SUMO consensus motifs at their N-termini that can be utilized to form polymeric SUMO chains in vitro and in vivo [1012]. There is clear evidence that SUMO-2/3 chain formation might be important for the regulation of some target proteins in vivo [1314]. In mammalian cells, the kinetochore protein CENP-E was found to be modified specifically by SUMO-2/3 and to possess poly-SUMO-2/3-chain binding activity that is essential for kinetochore localization [15].

Sumoylation is a highly dynamic and fully reversible modification. The family of SENPs (SUMO-1/sentrin/SMT3-specific peptidases) are cysteine proteases that cleave the isopeptide bond between SUMO and the bound proteins. These proteases are also responsible for the processing of the SUMO precursors to reveal the di-glycine motif that is conjugated to acceptor lysine residues in target proteins (reviewed in [16]). In yeast, there are two SUMO-specific proteases, Ulp1 (Ubl-specific protease 1) and Ulp2, which have been characterized and detected at the nuclear pore and nucleoplasm respectively [1719]. In humans, six SENPs have been reported. All share a conserved C-terminal sequence of ∼200 residues within which are the conserved amino acids (cysteine, histidine and aspartic acid) that form the catalytic triad. SENP1 (also designated SuPr-2), SENP2 [also designated AXAM2 or SMT3IP2 (SMT3-specific isopeptidase 2)], SENP3 (also designated SMT3IP1) and SENP5 have been shown to function as SUMO-specific proteases. SENP1 is a nuclear protease that deconjugates a large number of sumoylated proteins [20]. SENP2 is a nuclear envelope-associated protease, although differential splicing generates SENP2 proteins that can be cytoplasmic nuclear-pore localized and nuclear-body localized [21,22]. SENP3 and SENP5 are nucleolar SENPs with a preference for SUMO-2/3 [23]. SENP6 is a nuclear protease and appears to be specific for the deconjugation of isopeptide-linked SUMO-2 and SUMO-3, but is unable to deconjugate isopeptide-linked SUMO-1 and does not process SUMO-1, SUMO-2 or SUMO-3 precursors [24].

Previous studies showed that SENP7 fails to process SUMO precursors and implied that its natural substrates may not be SUMO conjugates [25,26]. In the present study, we report the characterization of SENP7. We found that SENP7 does not process the precursor forms of SUMO-1, SUMO-2 and SUMO-3, but can remove SUMO-2 from RanGAP1 (Ran GTPase-activating protein 1)–SUMO-2 and efficiently deconjugate poly-SUMO-2 chains. However, SENP7 has negligible activity against RanGAP1–SUMO-1 and poly-SUMO-1 chains in vitro. in vivo, SENP7 localized to the nucleoplasm and displayed preferential isopeptidase activity against SUMO-2/3 compared with SUMO-1 conjugates. The ablation of SENP7 expression with siRNA (small interfering RNA) resulted in the accumulation of high-molecular-mass SUMO-2 forms and in the increased accumulation of PML (promyelocytic leukaemia) protein in subnuclear bodies. These findings suggest that SENP7 acts as a SUMO-2/3-specific protease that is likely to regulate the metabolism of poly-SUMO-2/3 rather than SUMO-1 in vivo.

EXPERIMENTAL

Plasmid construction

The cDNAs for SUMO-1, SUMO-2 and SUMO-3 were subcloned into a pcDNA3-HA vector (Invitrogen) as described previously [10]. Full-length human SENP7 cDNA was amplified from a human fetal-brain cDNA library by PCR. The PCR product was then subcloned into pcDNA3 vector (Invitrogen) and pEGFP [enhanced GFP (green fluorescent protein)]-C1 vector (Clontech) using standard techniques. The above plasmids were transfected into COS-7 cells using FuGENE 6 (Roche Applied Science) according to the manufacturer's instructions.

Full-length SUMO-1 (residues 1–101), SUMO-2 (residues 1–103) and SUMO-3 (residues 1–95) were cloned into pHisTEV vector (generously given by Dr Huanting Liu, Centre for Biomolecular Sciences, University of St Andrews, St Andrews, Scotland, U.K.) between the BamHI and HindIII sites. For bacterial expression of SENP7C (catalytic domain of SENP7; residues 733–1050), pLous3-SENP7C was constructed by PCR amplification of SENP7 cDNA. The PCR product was cleaved with EcoRI and SalI and inserted into the expression vector pLous3 (generously given by Professor Jim Naismith, School of Chemistry, University of St Andrews, St Andrews, Scotland, U.K.).

All clones were verified by automated DNA sequence analysis and shown to be identical with those previously reported for SENP7, SUMO-1, SUMO-2 and SUMO-3 (GenBank® accession numbers NP_065705, AAH53528, AAH68465 and NP_008867 respectively). SENP7 and SENP7M (mutant SENP7C; C992S) were generated using a PCR-based mutagenesis [27] and verified by DNA sequence analysis (DNA Sequencing Unit, University of Dundee, Dundee, Scotland, U.K.).

Protein expression and purification

SUMO-1, SUMO-2 and SUMO-3 precursors were expressed from pHisTEV vector in Escherichia coli BL21 (DE3) cells and purified by Ni-NTA (Ni2+-nitrilotriacetate)-affinity chromatography and gel filtration (Superdex™75; Amersham Pharmacia) as described previously [28]. Recombinant SENP7C was expressed as a His6-MBP (maltose-binding protein)-tagged fusion protein in E. coli BL21 (DE3) cells by induction with 0.1 mM isopropyl β-D-thiogalactoside at 20 °C for 12 h. His6-MBP-tagged SENP7C was purified by Ni-NTA-affinity chromatography. The His6-MBP tag was cleaved from the fusion protein by treatment with His6-tagged TEV (tobacco etch virus) protease (generated in house as described in [29]) for 2 h at 20 °C, and SENP7C was further purified by Ni-NTA-affinity chromatography and gel filtration (Superdex™200; Amersham Pharmacia). SENP7M was expressed and purified as described above. Protein purity was evaluated by SDS/PAGE (10% gel) and CBB (Coomassie Brilliant Blue) staining. The concentrations of the purified proteins were determined by the Bradford method (Bio-Rad Laboratories).

in vivo analysis of SENP7 deconjugation activity

COS-7 cells were maintained at 37 °C in Dulbecco's Eagle's medium supplemented with 10% (v/v) fetal calf serum and 1% (v/v) penicillin/streptomycin. COS-7 cells were cultured to 50–60% confluency and transfected with expression constructs for SENP7 or SENP7M and either HA (haemagglutinin)-SUMO-1, HA-SUMO-2 or HA-SUMO-3 using FuGENE™ 6 reagent according to the manufacturer's instructions (Roche). Controls were performed with identical conditions and empty plasmids. The total amount of DNA was equalized to 3 μg with pcDNA3.

At 48 h after transfection, cell extracts were separated by SDS/PAGE (4–12% gel) and proteins were transferred on to PVDF membranes (Millipore), which were blocked with 5% (v/v) non-fat milk in TPBS [PBS containing 0.1% (v/v) Tween-20] for 1 h. Membranes were incubated for an additional 1 h at room temperature (20 °C) with the primary antibody (see below) in TPBS containing 5% (v/v) non-fat milk, washed three times with TPBS and incubated for a further 1 h in TPBS containing 5% (v/v) non-fat milk containing the appropriate secondary antibody. After three washes with PBS, membranes were processed with ECL® chemiluminescence detection reagents (Amersham Biosciences). Primary antibodies used in Western blotting were: mouse monoclonal anti-HA antibody 12CA5 (used at 1:2000 dilution; BAbCO); antigen-affinity-purified sheep polyclonal anti-SENP7 antibody [used at 1:3000 dilution; produced in-house (see below)]; and mouse monoclonal anti- β-actin antibody (used at 1:30000 dilution; Sigma). Chicken anti-PML antibody (used at 1:1000 dilution to visualize PML subnuclear bodies) was a gift from Dr Valérie Lallemand-Breitenbach and Dr Hugues de Thé (Centre National de la Recherche Scientifique, Paris), and mouse monoclonal 5E10 anti-PML antibody (used at 1:200 dilution) was provided by Professor Roel van Driel (Swammerdam Institute for Life Sciences, University of Amsterdam, Amsterdam, The Netherlands).

The SENP7-specific polyclonal antibody was generated by immunizing sheep with MBP-SENP7C. After two further injections of MBP-SENP7C, blood from the animal was collected and, after clotting, the serum was diluted with PBS and passed through a column of MBP-SENP6 coupled to NHS (N-hydroxysuccinimide)-activated Sepharose beads (GE Healthcare; prepared according to the manufacturer's instructions) to remove cross-reacting antibodies. The unbound material was then passed through a second column containing identically coupled MBP-SENP7C. The MBP-SENP7C column was washed with 40 ml of 0.5 M NaCl and 10 mM Tris/HCl (pH 7.5), and the anti-SENP7 antibody was eluted with 0.1 M glycine/HCl (pH 2.25). Fractions of 500 μl were collected into tubes containing 50 μl of 1 M Tris/HCl (pH 8.0). Fractions containing antibody were pooled, and BSA was added to 1 mg/ml, glycerol to 10% (v/v) and sodium azide to 0.1% (w/v).

siRNA methods

SENP7 expression was ablated with either four individual siRNAs (sense-strand sequences: SENP7-1, 5′-CAAAGUACCGAGUCGAAUAUU-3′; SENP7-2, 5′-GAUAAUGAUCUACGUACUAUU-3′; SENP7-3, 5′-GGGCAGCAGUAGCUAGUUAUU-3′; SENP7-4, 5′-UAGCAGUGAUUGUGGAGUAUU-3′) or a pool containing equal amounts of each double-stranded oligoribonucleotide (ON-TARGETplus™; Dharmacon). siRNAs (50 nM final concentration) were introduced into HeLa cells (generously given by Professor Birgit Lane, College of Life Sciences, University of Dundee, Dundee, Scotland, U.K.) using Oligofectamine™ (Invitrogen) according to the manufacturer's instructions. Cells were analysed after 48 h of incubation.

In vitro SUMO-processing assay

To assay SUMO-processing activity, full-length SUMO-1, SUMO-2 and SUMO-3 precursors (25 μM) were incubated either in the absence (negative control) or presence of purified SENP7C (5 μM) or SENP1C (0.5 μM; positive control) at 37 °C for 1 h in a reaction mixture containing 50 mM Tris/HCl (pH 7.5), 150 mM NaCl and 5 mM 2-mercaptoethanol. Reactions were terminated by adding 2× protein sample buffer [100 mM Tris/HCl (pH 6.8), 20% (v/v) glycerol, 4% (w/v) SDS, 3% (v/v) dithiothreitol and 0.002% (w/v) Bromophenol Blue], and reaction products were fractionated by SDS/PAGE (4–12% gel) followed by CBB R-250 staining.

In vitro isopeptidase assays with RanGAP1–SUMO-1/2, SUMO-2 dimer and poly-SUMO-1/2

All isopeptidase assays were performed in a buffer containing 50 mM Tris/HCl (pH 7.5), 2 mM MgCl2, 150 mM NaCl and 5 mM 2-mercaptoethanol.

The C-terminal fragment of RanGAP1 (residues 418–587) modified by SUMO-1, -2 and -3 was prepared as described previously [29]. The isopeptide-linked di-SUMO-2 (for clarity, the isopeptide-linked SUMO-2 dimer is referred to as di-SUMO-2 in the present study) was produced by self-conjugation in vitro [10] and purified by gel filtration (Superdex™75; Amersham Pharmacia) according to the manufacturer's instructions. To make poly-SUMO-1, a SUMO-1 mutant (D15V) was generated by PCR-based mutagenesis. Poly-SUMO-1-D15V and poly-SUMO-2 were prepared as described previously [10].

Isopeptidase assays were performed in 50 μl reactions containing substrates at the indicated concentration (see the Results section) and a range of concentrations of SENP7C (7.5 nM to 2 μM) for 1 h at 37 °C. Reactions were terminated by adding 2× protein sample buffer. Samples were fractionated by SDS/PAGE (4–12% gel) followed by CBB staining.

RESULTS

Expression and purification of SENP7C

On the basis of sequence alignment, SENP7 is a SUMO-specific protease. All SENPs share a conserved catalytic domain (∼200 residues) that contains a characteristic catalytic triad of cysteine, histidine and aspartate residues [16]. Interestingly, 50–200 amino acid insertions split the SENP7 and SENP6 catalytic domains into two halves (Figure 1A). To establish biochemical activity of SENP7, we cloned the SENP7 gene and expressed and purified the catalytic domain SENP7C. SENP7C (residues 733–1050) was expressed in bacteria as a His6-MBP fusion protein and purified by Ni-NTA-affinity chromatography. The His6-MBP tag was cleaved by TEV protease, and SENP7C was purified by exclusion from Ni-NTA agarose and gel filtration. CBB staining revealed that the purified SENP7C was essentially homogeneous (Figure 1B).

Purification of SENP7

Figure 1
Purification of SENP7

(A) Sequence alignment of the catalytic domain of SENP7 with other members of the SUMO-specific protease family. Sequences were aligned using ClustalW [35]. Identical residues are indicated by asterisks (*), highly conserved residues by colons (:), loosely conserved residues by points (.) and residues of the catalytic triad are boxed. (B) Expression and purification of the catalytic domain of SENP7. Samples were taken from various steps during the purification of SENP7C. Molecular mass markers (kDa) are shown on the left, with sample loading as follows: lane 1, total bacterial extract from non-induced cells; lane 2, total cell lysate from induced cells; lane 3, proteins in the pellet; lane 4, proteins in the supernatant; lane 5, His6-MBP-SENP7C purified from the Ni-NTA column; lane 6, cleavage of fusion protein by TEV protease overnight at 4 °C; lane 7, purified SENP7C.

Figure 1
Purification of SENP7

(A) Sequence alignment of the catalytic domain of SENP7 with other members of the SUMO-specific protease family. Sequences were aligned using ClustalW [35]. Identical residues are indicated by asterisks (*), highly conserved residues by colons (:), loosely conserved residues by points (.) and residues of the catalytic triad are boxed. (B) Expression and purification of the catalytic domain of SENP7. Samples were taken from various steps during the purification of SENP7C. Molecular mass markers (kDa) are shown on the left, with sample loading as follows: lane 1, total bacterial extract from non-induced cells; lane 2, total cell lysate from induced cells; lane 3, proteins in the pellet; lane 4, proteins in the supernatant; lane 5, His6-MBP-SENP7C purified from the Ni-NTA column; lane 6, cleavage of fusion protein by TEV protease overnight at 4 °C; lane 7, purified SENP7C.

SENP7 does not act as a SUMO C-terminal hydrolase

SUMO-1, SUMO-2 and SUMO-3 are all expressed as inactive precursors that require precise cleavage by SUMO proteases to reveal the C-terminal di-glycine motif. Previous research has demonstrated that SENP1, SENP2, SENP3 and SENP5 possess hydrolase activity and can discriminate between SUMO paralogues as processing substrates [16,30]. To examine SENP7 processing activity, recombinant purified SUMO-1, SUMO-2 and SUMO-3 precursors (25 μM) were incubated with purified SENP7C (5 μM) and SENP1C (0.5 μM), and reaction products were analysed by SDS/PAGE (4–12% gel) followed by staining with CBB. Consistent with the previous findings [16], we observed efficient SENP1-mediated cleavage of the SUMO-1 and SUMO-3 precursors, with lower activity towards the SUMO-2 precursor. However, we found that even at a 10-fold higher concentration, SENP7C does not detectably cleave SUMO-1, SUMO-2 and SUMO-3 precursors during a 1-h incubation at 37 °C (Figure 2A). These results suggested that, similarly to SENP6, SENP7 does not function as a general SUMO-processing enzyme.

Processing and deconjugation activities for SENP7

Figure 2
Processing and deconjugation activities for SENP7

(A) Assays for SUMO-processing activities by SENP7C (5 μM) and SENP1C (0. 5 μM) using SUMO-1, SUMO-2 and SUMO-3 precursors (25μM). SENP7C does not have C-terminal hydrolase activity against SUMO-1, SUMO-2 and SUMO-3 precursors, whereas SENP1C is active for all SUMO precursors. (B) Deconjugation-activity assays for SENP7C (100 nM) and SENP1C (10 nM) using RanGAP1–SUMO-1/2 (20μM). SENP7C does not remove SUMO-1 from RanGAP1–SUMO-1; however, it does remove SUMO-2 from RanGAP1–SUMO-2 with lower efficiency compared with SENP1C. (C) Deconjugation of RanGAP1–SUMO-1/2 and SUMO-2 dimer with a range of concentrations of SENP7C (7.5 nM–2 μM) at 37 °C. Reactions were stopped after 1 h with loading buffer and analysed by SDS/PAGE (4–12% gel) followed by CBB staining.

Figure 2
Processing and deconjugation activities for SENP7

(A) Assays for SUMO-processing activities by SENP7C (5 μM) and SENP1C (0. 5 μM) using SUMO-1, SUMO-2 and SUMO-3 precursors (25μM). SENP7C does not have C-terminal hydrolase activity against SUMO-1, SUMO-2 and SUMO-3 precursors, whereas SENP1C is active for all SUMO precursors. (B) Deconjugation-activity assays for SENP7C (100 nM) and SENP1C (10 nM) using RanGAP1–SUMO-1/2 (20μM). SENP7C does not remove SUMO-1 from RanGAP1–SUMO-1; however, it does remove SUMO-2 from RanGAP1–SUMO-2 with lower efficiency compared with SENP1C. (C) Deconjugation of RanGAP1–SUMO-1/2 and SUMO-2 dimer with a range of concentrations of SENP7C (7.5 nM–2 μM) at 37 °C. Reactions were stopped after 1 h with loading buffer and analysed by SDS/PAGE (4–12% gel) followed by CBB staining.

SENP7 is an isopeptidase with preference for SUMO-2 conjugates in vitro

To assess the isopeptidase activity of SENP7, model substrates were generated by conjugating the C-terminal fragment of RanGAP1 (residues 418–587) to SUMO-1 or SUMO-2. The RanGAP1–SUMO conjugates were purified by affinity chromatography and gel filtration, and the purified proteins were used as substrates (20 μM) to evaluate isopeptidase activity. Consistent with earlier findings [29], SENP1 (10 nM) can efficiently remove both SUMO-1 and SUMO-2 from RanGAP1 (Figure 2B). Although SENP7C (100 nM) did not deconjugate SUMO-1 from RanGAP1, it could partially deconjugate SUMO-2 from RanGAP1–SUMO-2 (Figure 2B). To evaluate the deconjugating activity of SENP7C, 20 μM RanGAP1–SUMO-1 or RanGAP1–SUMO-2 was incubated with a range of concentrations of SENP7C. Even at a high concentration of SENP7C (2 μM), only limited deconjugation of RanGAP1–SUMO-1 was observed (Figure 2C). Although SENP7C was capable of cleaving RanGAP1–SUMO-2, the reaction did not appear to proceed very efficiently, as 120 nM SENP7C was required to cleave 50% of the substrate (Figure 2C). As an alternative isopeptidase substrate, di-SUMO-2 was generated by conjugation in vitro and purification by gel filtration. SENP7C robustly cleaved di-SUMO-2 (20 μM), achieving 50% cleavage with 15 nM SENP7C (Figure 2C). Thus SENP7C has limited isopeptidase activity against RanGAP1–SUMO-1 and increased activity against RanGAP1–SUMO-2, but efficiently cleaves di-SUMO-2.

SENP7 deconjugates polymeric chains of SUMO-2 but not SUMO-1

It has been reported that, similar to ubiquitin, SUMO-2/3, but not SUMO-1, can form polymeric chains, either as homopolymers or conjugated to protein substrates [10]. The apparent preference of di-SUMO-2 compared with RanGAP1–SUMO-2 as a substrate for the isopeptidase activity of SENP7C suggested that the preferred substrates for SENP7 might be poly-SUMO chains. Although SUMO-1 does not usually form poly-SUMO-1 chains, we generated a D15V version of SUMO-1 that is capable of forming poly-SUMO-1 chains. This enabled us to compare directly the ability of SENP7 to depolymerize poly-SUMO-1 and poly-SUMO-2 chains. Equal amounts of poly-SUMO-1 or poly-SUMO-2 chains were incubated with SENP7C, SENP7M (a catalytically inactive form) or SENP1C. It is clear that SENP7C is capable of deconjugating poly-SUMO-2 chains, but has no detectable activity against poly-SUMO-1 chains (Figure 3A). As expected, SENP1 efficiently deconjugated both poly-SUMO-1 and poly-SUMO-2 chains (Figure 3A). To enable a direct comparison of the substrate preference of SENP7, 0.85 μg/μl poly-SUMO-1-D15V or poly-SUMO-2 chains were incubated with a range of concentrations of SENP7, and the reaction products were analysed by SDS/PAGE (4–12% gel) followed by CBB staining. Even at a high concentration (2 μm), SENP7C had no detectable activity against poly-SUMO-1 chains but was very active in depolymerizing poly-SUMO-2 chains (Figures 3B and 3C). At high concentrations, SENP7C is capable of completely depolymerizing poly-SUMO-2 chains to monomeric SUMO-2. Thus SENP7 discriminates between SUMO-1 and SUMO-2 in vitro, raising the possibility that SENP7 may differentially regulate SUMO-2/3 compared with SUMO-1 conjugates in vivo.

Comparison of the substrate preference of SENP7 against poly-SUMO-1 and poly-SUMO-2 chains

Figure 3
Comparison of the substrate preference of SENP7 against poly-SUMO-1 and poly-SUMO-2 chains

(A) Equal amounts of poly-SUMO-1 or poly-SUMO-2 chains (0.85 μg/μl) were incubated with SENP7C (100 nM) or SENP1C (10 nM). SENP1 effectively deconjugates both poly-SUMO-1 and poly-SUMO-2. However, SENP7C (100 nM) is less active, only partially deconjugating poly-SUMO-2 chains. SENP7C has no detectable activity against poly-SUMO-1 chains. (B) Deconjugation of poly-SUMO-1 chains with different concentrations of SENP7C (7.5 nM–2 μM). (C) The same reaction as described in (B), but substituting poly-SUMO-2 chains for poly-SUMO-1 chains.

Figure 3
Comparison of the substrate preference of SENP7 against poly-SUMO-1 and poly-SUMO-2 chains

(A) Equal amounts of poly-SUMO-1 or poly-SUMO-2 chains (0.85 μg/μl) were incubated with SENP7C (100 nM) or SENP1C (10 nM). SENP1 effectively deconjugates both poly-SUMO-1 and poly-SUMO-2. However, SENP7C (100 nM) is less active, only partially deconjugating poly-SUMO-2 chains. SENP7C has no detectable activity against poly-SUMO-1 chains. (B) Deconjugation of poly-SUMO-1 chains with different concentrations of SENP7C (7.5 nM–2 μM). (C) The same reaction as described in (B), but substituting poly-SUMO-2 chains for poly-SUMO-1 chains.

SENP7 has SUMO-2/3 isopeptidase activity in vivo

To establish that SENP7 is active in vivo, the cDNA encoding full-length SENP7 was inserted into the eukaryotic expression vector pcDNA3 and transfected into COS-7 cells along with expression constructs for HA-SUMO-1, HA-SUMO-2 or HA-SUMO-3. As a control, the HA versions of SUMO were also co-transfected with an empty expression construct or a construct encoding SENP7M, for which the catalytic cysteine residue was mutated to a serine residue. After 48 h, SUMO-modified substrates in cell extracts were revealed by Western blotting with an anti-HA monoclonal antibody. Equal expression levels of SENP7 and its mutant were verified by anti-SENP7 Western blotting (Figure 4A). Transfection of HA-SUMO-1, HA-SUMO-2 or HA-SUMO-3 resulted in the accumulation of a number of high-molecular-mass SUMO-modified species. Co-transfection with wild-type SENP7 led to a reduction in the level of high-molecular-mass SUMO-2 and SUMO-3 conjugates, but not SUMO-1 conjugates. As expected, Cys992 is required for desumoylation, since expression of SENP7M failed to promote the reduction of either SUMO-2 or SUMO-3 conjugates (Figure 4A).

Activity of SENP7 in vivo

Figure 4
Activity of SENP7 in vivo

(A) SENP7 acts as a SUMO-2/3-specific protease in vivo. COS-7 cells were co-transfected with either empty pcDNA3 (−) or plasmids expressing HA-SUMO-1, HA-SUMO-2, HA-SUMO-3, SENP7 or SENP7M as indicated. Western blots using anti-HA, anti-SENP7 and anti-β-actin antibodies were performed on the cell lysates as detailed in the Experimental section. (B) Subcellular localization of SENP7. Full-length human SENP7 cDNA or its inactive mutant SENP7M was subcloned into pEGFP-C1 vector (Clontech) and transfected into COS-7 cells. The localization of the fusion protein (green) was determined by fluorescence microscopy. Cells were also stained with an antibody to visualize PML subnuclear bodies (red) and with DAPI (4′,6-diamidino-2-phenylindole) to visualize nuclei (blue).

Figure 4
Activity of SENP7 in vivo

(A) SENP7 acts as a SUMO-2/3-specific protease in vivo. COS-7 cells were co-transfected with either empty pcDNA3 (−) or plasmids expressing HA-SUMO-1, HA-SUMO-2, HA-SUMO-3, SENP7 or SENP7M as indicated. Western blots using anti-HA, anti-SENP7 and anti-β-actin antibodies were performed on the cell lysates as detailed in the Experimental section. (B) Subcellular localization of SENP7. Full-length human SENP7 cDNA or its inactive mutant SENP7M was subcloned into pEGFP-C1 vector (Clontech) and transfected into COS-7 cells. The localization of the fusion protein (green) was determined by fluorescence microscopy. Cells were also stained with an antibody to visualize PML subnuclear bodies (red) and with DAPI (4′,6-diamidino-2-phenylindole) to visualize nuclei (blue).

As most SUMO conjugates are known to be nuclear, it was of interest to determine the subcellular localization of SENP7. Therefore we fused the cDNA for SENP7 in-frame with the cDNA for EGFP, and an expression construct for the EGFP–SENP7 fusion protein was transfected into COS-7 cells. The localization of the fusion protein was determined by fluorescence microscopy using a DeltaVision® microscope (Applied Precision) and shown to be diffusely nuclear with some subnuclear punctate staining (Figure 4B). Cells were also stained with an anti-PML antibody (used at 1:200 dilution) for 1 h at 20 °C. Although the bulk of SENP7 was not co-localized to PML bodies, a small proportion of SENP7 appeared to localize with PML in subnuclear bodies. The distribution of the catalytically inactive SENP7M fused to EGFP was indistinguishable from that of wild-type SENP7 (Figure 4B).

SENP7 deletion results in the accumulation of SUMO-2 conjugates, but not SUMO-1 conjugates, in vivo

To examine the biological function of SENP7 in vivo, its expression was suppressed by siRNA-mediated gene silencing in human HeLa cells. At 48 h after siRNA treatment with a pool of siRNAs (Dharmacon) that targeted SENP7 mRNA or a non-targeting pool of siRNAs, cell extracts were prepared and analysed by Western blotting. SENP7 expression was efficiently reduced in cells treated with the targeting siRNA, but not the non-targeting siRNA (Figure 5A). The same cell extracts were also analysed by Western blotting with antibodies to SUMO-1 and SUMO-2/3. The ablation of SENP7 expression resulted in the accumulation of high-molecular-mass SUMO-2 species with only a small increase in SUMO-1-staining species. As SUMO modification is known to contribute to the metabolism and localization of PML subnuclear bodies, cells treated with either an siRNA targeted against SENP7 or a non-targeting siRNA were incubated with an antibody to PML, and PML localization was determined by immunofluorescence. Cells treated with the SENP7 siRNA displayed both an increase in the number of PML bodies (Figure 5C) and an increase in the intensity of PML subnuclear-body staining (Figure 5B). Thus SENP7 appears to regulate both SUMO-2/3 and PML metabolism in vivo.

Ablation of SENP7 expression in vivo

Figure 5
Ablation of SENP7 expression in vivo

(A) Knockdown of SENP7 leads to the accumulation of SUMO-2 conjugates in vivo. HeLa cells were transfected with a pool of siRNAs that targeted SENP7 mRNA (siS7) or with control siRNAs (siNT). At 48 h after siRNA treatment, cell lysates were analysed by Western blotting with anti-SENP7, anti-SUMO-1 and anti-SUMO-2 antibodies. (B) SENP7 regulates PML metabolism in vivo. Cells treated with either siRNAs against SENP7 or control siRNAs were stained with a mouse anti-PML 5E10 antibody, and PML localization (green) was determined by immunofluorescence microscopy with an anti-PML antibody. DNA was stained with DAPI (blue). (C) Knockdown of SENP7 leads to increased numbers of PML bodies in HeLa cells.

Figure 5
Ablation of SENP7 expression in vivo

(A) Knockdown of SENP7 leads to the accumulation of SUMO-2 conjugates in vivo. HeLa cells were transfected with a pool of siRNAs that targeted SENP7 mRNA (siS7) or with control siRNAs (siNT). At 48 h after siRNA treatment, cell lysates were analysed by Western blotting with anti-SENP7, anti-SUMO-1 and anti-SUMO-2 antibodies. (B) SENP7 regulates PML metabolism in vivo. Cells treated with either siRNAs against SENP7 or control siRNAs were stained with a mouse anti-PML 5E10 antibody, and PML localization (green) was determined by immunofluorescence microscopy with an anti-PML antibody. DNA was stained with DAPI (blue). (C) Knockdown of SENP7 leads to increased numbers of PML bodies in HeLa cells.

DISCUSSION

In the present study, we provide an analysis of the biochemical and biological activities of the SUMO-specific protease SENP7. We find that SENP7 appears to function primarily as an isopeptidase and has no detectable capacity to process endoproteolytically the natural precursors of SUMO-1, SUMO-2 and SUMO-3. In surveying the biochemical activities of the SENPs, the inability of SENP7 to process SUMO precursors had been noted [26]. As an isopeptidase, the preferred substrate of SENP7 appears to be polymeric chains of SUMO-2/3, although it is capable of deconjugating SUMO-2/3 from model substrates such as RanGAP1, albeit with reduced efficiency. It appears to have undetectable activity against either poly-SUMO-1 chains or SUMO-1 conjugated to model substrates. Thus SENP7 is a SUMO-2/3-specific protease with a distinct preference to cleave the isopeptide bond that links two SUMO-2/3 molecules, While the present manuscript was in preparation, the structure of SENP7 was reported [31]. This revealed that SENP7 was structurally related to the previously characterized SENP1 and SENP2, but had a number of unique structural features. This study also concluded that SENP7 had a preference for poly-SUMO chains, but the structural basis for this preference was not evident.

in vivo, the co-expression of SENP7 with HA-SUMO-2 or HA-SUMO-3 leads to the reduction of high-molecular-mass SUMO-2/3 species. This effect is highly specific as it is not observed when a catalytically inactive version of SENP7 is co-expressed with HA-SUMO-2 or HA-SUMO-3. Co-expressed SENP7 does not influence the modification profile of co-transfected HA-SUMO-1, consistent with the lack of activity of SENP7 against SUMO-1-modified substrates or poly-SUMO-1 in vitro. Consistent with these overexpression studies, siRNA-mediated ablation of SENP7 expression leads to a substantial accumulation of high-molecular-mass SUMO-2/3-modified species, with a small accumulation of high-molecular-mass SUMO-1-modified species. Whereas the in vitro analysis indicated that SENP7 was highly selective for SUMO-2/3 and had little activity against SUMO-1, the in vivo observations can be explained by the presence of poly-SUMO-2/3 chains that are terminated by SUMO-1 [32]. If the poly-SUMO-2/3 component of this mixed chain is cleaved by SENP7, this would account for the apparent loss of the high-molecular-mass SUMO-1 material. Analysis of PML protein expression by immunofluorescence indicates that the intensity of PML staining increases, as does the number of PML bodies per nucleus, in the absence of SENP7. As SUMO-2/3 modification of PML influences the metabolism of PML protein by the recruitment of the SUMO-specific ubiquitin E3 ligase Rnf4 (ring finger protein 4) [33,34], this indicates that SENP7 is probably a key regulator of PML turnover. SENP6 has a similar substrate specificity and effect on PML bodies to that observed above for SENP7 [24,31], suggesting that the two SUMO proteases do not have redundant functions, and that the ablation of expression of one of these proteases cannot be complemented by the other.

We thank Dr Heidi Mendoza (Department of Biochemical Medicine, Ninewells Hospital and Medical School, University of Dundee) and Dr Barbara Ink (GlaxoSmithKline) for providing the original SENP7 constructs.

Abbreviations

     
  • CBB

    Coomassie Brilliant Blue

  •  
  • DAPI

    4′,6-diamidino-2-phenylindole

  •  
  • EGFP

    enhanced green fluorescent protein

  •  
  • GAP

    GTPase-activating protein

  •  
  • HA

    haemagglutinin

  •  
  • MBP

    maltose-binding protein

  •  
  • Ni-NTA

    Ni2+-nitrilotriacetate

  •  
  • PML

    promyelocytic leukaemia

  •  
  • RanBP

    Ran-binding protein

  •  
  • SAE

    SUMO-activating enzyme

  •  
  • SENP

    SUMO-1/sentrin/SMT3-specific peptidase

  •  
  • SENP7C

    catalytic domain of SENP7

  •  
  • SENP7M

    mutant SENP7C (C992S)

  •  
  • siRNA

    small interfering RNA

  •  
  • SMT3

    suppressor of mif two 3 homologue 1

  •  
  • SMT3IP

    SMT3-specific isopeptidase

  •  
  • SUMO

    small ubiquitin-related modifier

  •  
  • TEV

    tobacco etch virus

  •  
  • TPBS

    PBS containing 0.1% (v/v) Tween-20

  •  
  • Ubc9

    ubiquitin-conjugating enzyme 9

  •  
  • Ubl

    ubiquitin-like protein, Ulp, Ubl-specific protease

AUTHOR CONTRIBUTION

Lin Nan Shen, Marie-Claude Geoffroy, Ellis Jaffray and Ronald Hay designed the experiments, Lin Nan Shen carried out the biochemical analysis, Marie-Claude Geoffroy performed the in vivo analysis of SUMO and PML metabolism, Ellis Jaffray carried out the siRNA experiments, and Lin Nan Shen and Ronald Hay wrote the paper.

FUNDING

This work was supported by Cancer Research UK (grant number C434/A7794), the BBSRC (Biotechnology and Biological Sciences Research Council) [grant number BB/F010125/1] and the European Union (RUBICON Network of Excellence) (all to R. T. H.).

References

References
1
Li
T.
Evdokimov
E.
Shen
R. F.
Chao
C. C.
Tekle
E.
Wang
T.
Stadtman
E. R.
Yang
D. C.
Chock
P. B.
Sumoylation of heterogeneous nuclear ribonucleoproteins, zinc finger proteins, and nuclear pore complex proteins: a proteomic analysis
Proc. Natl. Acad. Sci. U.S.A.
2004
, vol. 
101
 (pg. 
8551
-
8556
)
2
Vassileva
M. T.
Matunis
M. J.
SUMO modification of heterogeneous nuclear ribonucleoproteins
Mol. Cell. Biol.
2004
, vol. 
24
 (pg. 
3623
-
3632
)
3
Vertegaal
A. C.
Ogg
S. C.
Jaffray
E.
Rodriguez
M. S.
Hay
R. T.
Andersen
J. S.
Mann
M.
Lamond
A. I.
A proteomic study of SUMO-2 target proteins
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
33791
-
33798
)
4
Zhao
Y.
Kwon
S. W.
Anselmo
A.
Kaur
K.
White
M. A.
Broad spectrum identification of cellular small ubiquitin-related modifier (SUMO) substrate proteins
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
20999
-
21002
)
5
Su
H.
Li
S.
Molecular features of human ubiquitin-like SUMO genes and their encoded proteins
Gene
2002
, vol. 
296
 (pg. 
65
-
73
)
6
Hay
R. T.
SUMO: a history of modification
Mol. Cell
2005
, vol. 
18
 (pg. 
1
-
12
)
7
Hay
R. T.
Protein modification by SUMO
Trends Biochem. Sci.
2001
, vol. 
26
 (pg. 
332
-
333
)
8
Rodriguez
M. S.
Dargemont
C.
Hay
R. T.
SUMO-1 conjugation in vivo requires both a consensus modification motif and nuclear targeting
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
12654
-
12659
)
9
Melchior
F.
Schergaut
M.
Pichler
A.
SUMO: ligases, isopeptidases and nuclear pores
Trends Biochem. Sci.
2003
, vol. 
28
 (pg. 
612
-
618
)
10
Tatham
M. H.
Jaffray
E.
Vaughan
O. A.
Desterro
J. M.
Botting
C. H.
Naismith
J. H.
Hay
R. T.
Polymeric chains of SUMO-2 and SUMO-3 are conjugated to protein substrates by SAE1/SAE2 and Ubc9
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
35368
-
35374
)
11
Bencsath
K. P.
Podgorski
M. S.
Pagala
V. R.
Slaughter
C. A.
Schulman
B. A.
Identification of a multifunctional binding site on Ubc9p required for Smt3p conjugation
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
47938
-
47945
)
12
Bylebyl
G.R.
Belichenko
I.
Johnson
E. S.
The SUMO isopeptidase Ulp2 prevents accumulation of SUMO chains in yeast
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
44113
-
44120
)
13
Li
Y.
Wang
H.
Wang
S.
Quon
D.
Liu
Y. W.
Cordell
B.
Positive and negative regulation of APP amyloidogenesis by sumoylation
Proc. Natl. Acad. Sci. U.S.A.
2003
, vol. 
100
 (pg. 
259
-
264
)
14
Fu
C.
Ahmed
K.
Ding
H.
Ding
X.
Lan
J.
Yang
Z.
Miao
Y.
Zhu
Y.
Shi
Y.
Zhu
J.
, et al. 
Stabilization of PML nuclear localization by conjugation and oligomerization of SUMO-3
Oncogene
2005
, vol. 
24
 (pg. 
5401
-
5413
)
15
Zhang
X. D.
Goeres
J.
Zhang
H.
Yen
T. J.
Porter
A. C.
Matunis
M. J.
SUMO-2/3 modification and binding regulate the association of CENP-E with kinetochores and progression through mitosis
Mol. Cell
2008
, vol. 
28
 (pg. 
729
-
741
)
16
Mukhopadhyay
D.
Dasso
M.
Modification in reverse: the SUMO proteases
Trends Biochem. Sci.
2007
, vol. 
32
 (pg. 
286
-
295
)
17
Li
S. J.
Hochstrasser
M.
A new protease required for cell-cycle progression in yeast
Nature
1999
, vol. 
398
 (pg. 
246
-
251
)
18
Li
S. J.
Hochstrasser
M.
The yeast ULP2 (SMT4) gene encodes a novel protease specific for the ubiquitin-like Smt3 protein
Mol. Cell. Biol.
2000
, vol. 
20
 (pg. 
2367
-
2377
)
19
Takahashi
Y.
Mizoi
J.
Toh
E. A.
Kikuchi
Y.
Yeast Ulp1, an Smt3-specific protease, associates with nucleoporins
J. Biochem. (Tokyo)
2000
, vol. 
128
 (pg. 
723
-
725
)
20
Gong
L.
Millas
S.
Maul
G. G.
Yeh
E. T.
Differential regulation of sentrinized proteins by a novel sentrin-specific protease
J. Biol. Chem.
2000
, vol. 
275
 (pg. 
3355
-
3359
)
21
Zhang
H.
Saitoh
H.
Matunis
M. J.
Enzymes of the SUMO modification pathway localize to filaments of the nuclear pore complex
Mol. Cell. Biol.
2002
, vol. 
22
 (pg. 
6498
-
6508
)
22
Itahana
Y.
Yeh
E. T. H.
Zhang
Y.
Nucleocytoplasmic shuttling modulates activity and ubiquitination-dependent turnover of SUMO-specific protease 2
Mol. Cell. Biol.
2006
, vol. 
26
 (pg. 
4675
-
4689
)
23
Gong
L.
Yeh
E. T.
Characterization of a family of nucleolar SUMO specific proteases with preference for SUMO-2 or SUMO-3
J. Biol. Chem.
2006
, vol. 
281
 (pg. 
15869
-
15877
)
24
Mukhopadhyay
D.
Ayaydin
F.
Kolli
N.
Tan
S. H.
Anan
T.
Kametaka
A.
Azuma
Y.
Wilkinson
K. D.
Dasso
M.
SUSP1 antagonizes formation of highly SUMO2/3-conjugated species
J. Cell Biol.
2006
, vol. 
174
 (pg. 
939
-
949
)
25
Drag
M.
Mikolajczyk
J.
Krishnakumar
I. M.
Huang
Z.
Salvesen
G. S.
Activity profiling of human deSUMOylating enzymes (SENPs) with synthetic substrates suggests an unexpected specificity of two newly characterized members of the family
Biochem. J.
2008
, vol. 
409
 (pg. 
461
-
469
)
26
Mikolajczyk
J.
Drag
M.
Békés
M.
Cao
J. T.
Ronai
Z.
Salvesen
G. S.
Small ubiquitin-related modifier (SUMO)-specific proteases: profiling the specificities and activities of human SENPs
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
26217
-
26224
)
27
Brøns-Poulsen
J.
Petersen
N. E.
Hørder
M.
Kristiansen
K.
An improved PCR-based method for site directed mutagenesis using megaprimers
Mol. Cell. Probes
1998
, vol. 
12
 (pg. 
345
-
348
)
28
Shen
L. N.
Dong
C.
Liu
H.
Naismith
J. H.
Hay
R. T.
The structure of SENP1–SUMO-2 complex suggests a structural basis for discrimination between SUMO paralogues during processing
Biochem. J.
2006
, vol. 
397
 (pg. 
279
-
288
)
29
Shen
L.
Tatham
M. H.
Dong
C.
Zagórska
A.
Naismith
J. H.
Hay
R. T.
SUMO protease SENP1 induces isomerization of the scissile peptide bond
Nat. Struct. Mol. Biol.
2006
, vol. 
13
 (pg. 
1069
-
1077
)
30
Cheng
J.
Bawa
T.
Lee
P.
Gong
L.
Yeh
E. T.
Role of desumoylation in the development of prostate cancer
Neoplasia
2006
, vol. 
8
 (pg. 
667
-
676
)
31
Lima
C. D.
Reverter
D.
Structure of the human SENP7 catalytic domain and poly-SUMO deconjugation activities for SENP6 and SENP7
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
32045
-
32055
)
32
Matic
I.
van Hagen
M.
Schimmel
J.
Macek
B.
Ogg
S. C.
Tatham
M. H.
Hay
R. T.
Lamond
A. I.
Mann
M.
Vertegaal
A. C.
in vivo identification of human small ubiquitin-like modifier polymerization sites by high accuracy mass spectrometry and an in vitro to in vivo strategy
Mol. Cell. Proteomics
2008
, vol. 
7
 (pg. 
132
-
144
)
33
Tatham
M. H.
Geoffroy
M. C.
Shen
L.
Plechanovova
A.
Hattersley
N.
Jaffray
E. G.
Palvimo
J. J.
Hay
R. T.
RNF4 is a poly-SUMO-specific E3 ubiquitin ligase required for arsenic-induced PML degradation
Nat. Cell Biol.
2008
, vol. 
10
 (pg. 
538
-
546
)
34
Lallemand-Breitenbach
V.
Jeanne
M.
Benhenda
S.
Nasr
R.
Lei
M.
Peres
L.
Zhou
J.
Zhu
J.
Raught
B.
de Thé
H.
Arsenic degrades PML or PML-RARα through a SUMO-triggered RNF4/ubiquitin-mediated pathway
Nat. Cell Biol.
2008
, vol. 
10
 (pg. 
547
-
555
)
35
Thompson
J. D.
Higgins
D. G.
Gibson
T. J.
CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice
Nucleic Acids Res.
1994
, vol. 
22
 (pg. 
4673
-
4680
)