SUG1, a putative transcriptional mediator and subunit of the PA700 proteasome regulatory complex, is a DNA helicase.

Yeast SUG1 was originally characterized as a transcriptional mediator for the GAL4 transactivator. A similar role in vertebrates was suggested by the ligand-enhanced interaction between mammalian homologues of yeast SUG1 and the ligand-dependent activating domain (AF-2) of nuclear receptors. SUG1 was also shown to be a component of the PA700 regulatory complex of the 26 S proteasome and a member of a large family of putative ATPases. However, no catalytic function has yet been attributed to SUG1. We show here that SUG1 is a 3′-5′ DNA helicase whose activity is dependent on an intact ATP binding domain. The sedimentation heterogeneity of mammalian SUG1 suggests that it may be associated with distinct protein complexes and therefore play multiple roles.

Yeast SUG1 was originally characterized as a transcriptional mediator for the GAL4 transactivator. A similar role in vertebrates was suggested by the ligandenhanced interaction between mammalian homologues of yeast SUG1 and the ligand-dependent activating domain (AF-2) of nuclear receptors. SUG1 was also shown to be a component of the PA700 regulatory complex of the 26 S proteasome and a member of a large family of putative ATPases. However, no catalytic function has yet been attributed to SUG1. We show here that SUG1 is a 3-5 DNA helicase whose activity is dependent on an intact ATP binding domain. The sedimentation heterogeneity of mammalian SUG1 suggests that it may be associated with distinct protein complexes and therefore play multiple roles.
Studies in transcription interference/squelching indicate that enhancement of transcription by nuclear receptors may require protein mediators (co-activators) linking ligand-activated nuclear receptors to the transcription machinery at promoters of target genes (1)(2)(3). Several putative transcription intermediary factors interacting in a ligand-dependent manner with the region containing the ligand-dependent transcription activation function-2 (AF-2) of nuclear receptors have been described recently (for review, see Ref. 4). Ligand-enhanced interaction between the AF-2 domain of nuclear receptors and either mouse (m)SUG1 (5) or its human homologue TRIP1 (6,7) has suggested that they may act as mediators in transcription activated by nuclear receptors. Moreover, yeast (y)SUG1, has been shown to suppress the effect of a mutation in the transcriptional activator GAL4 (8,9). Functional similarity between yeast and mammalian SUG1 is supported by the ability of mSUG1 to rescue a conditional SUG1 mutation in yeast complementation experiments (5). Yeast SUG1 was also found in purified RNA polymerase II (pol II) 1 holoenzyme complexes responsive in vitro to the transcriptional activators GAL-VP16 and GCN4 (10). However, this association is controversial as SUG1 does not appear to be present in the yeast holoenzyme preparation of Rubin et al. (11).
The sequence similarity between mSUG1, ySUG1, and TRIP1 is the highest in a region containing the consensus ATP binding site motif. This motif corresponds to an AAA module recently described for a large family of putative ATPases involved in a variety of cellular processes (for review, see Ref. 12). For instance, MSS1 and TBP1, which were originally identified as transcription factors, were later shown to be members of the PA700 proteasome regulatory complex of the 26 S proteasome (13)(14)(15)(16)(17). Similarly, Kominami et al. (18) and then Rubin et al. (11) found SUG1 in the yeast 26 S proteasome. This agrees with genetic evidence showing that yeast cells harboring a mutant allele of ySUG1 accumulate ubiquitinylated proteins normally degraded by 26 S proteasomes (19). A role for SUG1 in the regulation of the activity of the 26 S proteasome in vivo is further supported by the identical amino acid sequence of mSUG1 with the p45 subunit of the PA700 proteasome regulatory complex purified from human and bovine tissues, respectively ( (5,17,20). 2 Therefore, SUG1 may be involved in more than one cellular function.
Despite the presence of a putative ATPase domain, no catalytic function has yet been attributed to SUG1. We show here that, in accordance with sequence similarities found with the DExH/D (where x ϭ any amino acid) box subfamily of helicases (21), recombinant mouse SUG1 exhibits intrinsic 3Ј-5Ј DNA helicase activity that is abolished by a mutation in its putative ATP binding domain. That SUG1 helicase may have multiple roles is supported by the observation that nuclear hSUG1 sediments not only with the PA700 complex, but also in regions corresponding to higher and lower molecular weights.

EXPERIMENTAL PROCEDURES
SUG1 Sequence Homology with Helicases-A computer search (BLAST (22)) was conducted with the full-length amino acid sequences of mSUG1 and two characterized helicases, XPB and XPD, to identify proteins with structural homology. Sequence comparison of mSUG1 with consensus helicase domains and known helicase XPB was done manually.
Purification of Recombinant His-SUG1-Plated Sf9 cells (400 ϫ 10 6 ) infected at a multiplicity of infection of 2 plaque-forming units/cell with baculovirus encoding a His-tagged SUG1 cDNA in the pAcSG HisNTB vector 3 (PharMingen) were collected in growth medium, pelleted (3000 * This work was supported by grants from the Center National de la Recherche Scientifique, the Institute National de la Santé et de la Recherche Médicale, Center Hospitalier Universitaire Régional, the Ministère de la Recherche et de la Technologie, the Fondation pour la Recherche Médicale, the Collège de France, Brystol-Meyers Squibb, and the Association pour la Recherche contre le Cancer. 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  rpm, 10 min, 4°C) and resuspended in 100 pellet volumes of phosphatebuffered saline containing 30% glycerol. Cells were pelleted, Douncehomogenized in cell lysis buffer (20 mM Tris-HCl, pH 7.9, 150 mM NaCl, 0.1% Nonidet P-40, 5 mM DTT, 5 mM PMSF, and 1 ϫ protease inhibitor mixture (24)), and centrifuged at 12,000 ϫ g for 50 min to remove cellular debris. Supernatant dialyzed against load buffer (20 mM Tris-HCl, pH 7.9, 50 mM KCl, 0.1 mM EDTA, 5 mM DTT, 20% glycerol, 5 mM PMSF, and 1 ϫ protease inhibitor mixture) was loaded onto an equilibrated DEAE-Spherodex column washed with load buffer until protein elution was negligible (as determined by a 280-nm wavelength monitor). Protein was stepwise eluted from the column with 0.125, 0.25, 0.6, and 1 M KCl in load buffer and analyzed for SUG1 immunoreactivity by Western blot. The 0.25 M KCl fraction containing the immunoreactive SUG1 (see Fig. 2B) was dialyzed against Ni-NTA binding buffer (20 mM Tris-HCl, pH 7.9, 500 mM NaCl, 5 mM imidazole, 5 mM DTT, 5 mM PMSF, and 1 ϫ protease inhibitor mixture) prior to application onto an equilibrated nickel chelate (Ni-NTA) column. The column was washed extensively with wash buffer (binding buffer with 60 mM imidazole) and then eluted into 0.5 column volume fractions with 100 then 500 mM imidazole-containing binding buffer. Fractions were analyzed by Western blot and silver nitrate staining.
SDS-PAGE-Equal volumes (20 l for silver-stained and 5 l for Western blot analysis) of the 500 mM imidazole fractions 1, 2, 3, and 4 from the Ni-NTA column were subjected to SDS, 10% PAGE and either silver nitrate-stained (24) or analyzed by Western blot as described previously (5), except that goat anti-mouse IgG peroxidase was used as the secondary antibody.
Helicase and ATPase Assays-Heparin fraction 12 (5 l) containing concentrated TFIIH (25) and equal volumes (5 l) of His-SUG1 (Ni-NTA 500 mM imidazole fractions 1-4) were added to a bi-directional helicase assay (26). The Ni-NTA 500 mM imidazole elution fraction 2 (100 l) was immunodepleted of His-SUG1 with an N-terminal-specific anti-SUG1 mAb (2SU 1B8; Ref. 5) as follows: the fraction was precleared with 10 l of protein G-Sepharose in batch at 4°C for 1 h; the supernatant was then incubated with mAb bound to protein G-Sepharose at 4°C for 2 h; the supernatant was collected and analyzed for helicase activity (5 l) and SUG1 immunoreactivity; the His-SUG1 bound to the mAb/protein G-Sepharose was washed with load buffer containing 500 mM KCl; the wash was dialyzed and tested (5 l) for helicase activity and SUG1 immunoreactivity; His-SUG1 bound to the mAb/resin was also analyzed (10 l of slurry) for helicase activity and for SUG1 immunoreactivity. ATPase assays were performed as described previously (27).
Baculovirus Mutant Mouse SUG1-Mutant His-SUG1 (His-196) cDNA bearing amino acid substitution of the lysine in position 196 (5) (K in bold, Fig. 1B, Domain I) was cloned into pET15b vector, and the XbaI/BamHI fragment containing the His 6 tag was subsequently cloned into pVL1392 baculovirus expression vector (PharMingen). The mutant was expressed in and purified from Sf9 cells as described above for wild type (wt) His-SUG1.
Glycerol Gradients-High molecular weight marker proteins (Pharmacia Biotech Inc.) and 150 -200 l (1.5-2 mg of protein) of a HeLa cell nuclear extract (28) were separately centrifuged through a 20 -40% glycerol gradient as described (25). 20 l of each 140 l fraction were analyzed by Western blot for the presence of hSUG1, the large subunit of RNA pol II (mAb 7G5), TFIIH subunit p62 (mAb 3C9 (27)) and PA700 subunits S7 (MSS1; mAb 2SCO) and S4 (mAb 4SCO). The glycerol concentration in each fraction was determined to ensure that the gradients were linear (data not shown).
Antibody Production-Synthetic peptides MPDYLGADQRKTKE-DEKDDKPIC and PGGGKKDDEDKKKKYEPPVPC corresponding to hyperantigenic regions in the amino acid sequences of S7 and S4, respectively, were cross-linked to ovalbumin and injected into mice for the production of mAb 2SCO and 4SCO, respectively, as described previously (24).

RESULTS
Computer searches were performed with the full-length amino acid sequence of mSUG1 to identify proteins exhibiting sequence similarity. Apart from the sequence matches to members of the AAA ATPase proteins, all data bases screened (PIR, Prodom, TrEMBL) revealed a significant similarity between SUG1 and two regions within the Mycoplasma genitalium homologue of the Holiday junction helicase RuvB (29). These regions of similarity (data not shown) overlapped with domains I and Ia (47% identity and 58% similarity over 34 residues) and II (31% identity and 75% similarity over 16 residues) of the seven conserved domains found in a DNA helicase family of proteins containing the DExH/D box (21). Some of these DExH/D box-containing proteins have been shown to have helicase activity (e.g. the helicase subunits of the repair/transcription factor TFIIH, XPB and XPD (26,30,31)), and sequence analysis has revealed seven consecutive conserved domains; I, Ia (the putative ATP binding segment), II (the putative Mg 2ϩ binding segment), III, IV, V, and VI (of unknown function) (21). No significant sequence similarity between SUG1, XPD, and XPB could be found using the BLAST program. Interestingly, apart from homologues in other organisms, a search with the BLAST program did not match XPD and XPB with sequences of other known helicases. However, some similarities were uncovered by manual alignment of the DExH/D box helicase consensus motifs (21) of SUG1 and XPB (30,31) (Fig. 1B or XPD data not shown; see Ref. 30). Note that in SUG1 the sequence similar to the consensus motif of domain VI is located between domains IV and V (Fig. 1A).  (21)). Amino acids matching the motif are boxed. Invariant amino acids are represented by their single letter code, conserved charged/polar amino acids (S, T, D, E, N, Q, K, R) are indicated with "p", hydrophobic/nonpolar amino acids (I, L, V, M, F, Y, W) with "n," and nonspecific amino acids with "x". The lysine residue in the ATPase domain I (Lys-196 (K196)) which is mutated to histidine for the production of the mutant His-SUG1 (5) (mut, Fig. 3

) is indicated in bold.
To investigate whether mSUG1 possesses helicase activity, recombinant His-SUG1 was expressed in Sf9 insect cells and purified to near homogeneity by DEAE-Spherodex followed by nickel-chelate (Ni-NTA) affinity chromatography (lanes 1-4, Fig. 2A, silver-stained gel; this procedure was repeated independently at least seven times by two investigators (R. A. F. and M.R.) without significant variation in the results). Purified His-SUG1 activity was tested in a bi-directional DNA helicase assay; release of the 27-and 24-nucleotide fragment from the linear template corresponded to 5Ј-3Ј and 3Ј-5Ј helicase activity, respectively (26). SUG1 displayed a 3Ј-5Ј helicase activity, as it specifically catalyzed the displacement of only the 24nucleotide fragment (Fig. 2B, Helicase lanes 1-4). Included as a positive control (Fig. 2B), purified TFIIH exhibited both the 5Ј-3Ј and 3Ј-5Ј helicase activities expected of the subunits XPD/ ERCC2 (5Ј-3Ј) and XPB/ERCC3 (3Ј-5Ј), respectively (26,31). Note that the level of helicase activity correlated well with the amount of SUG1 present in each reaction (Fig. 2B, Western  blot lanes 1-4 ). Furthermore, immunodepletion of SUG1 using an anti-SUG1 mAb specifically removed the helicase activity (Fig. 2B, lane 5). Moreover, the helicase activity, as well as the SUG1 polypeptide, remained bound to the mAb/protein G-Sepharose (Fig. 2B, lane 7), even after high stringency salt washes (500 mM KCl; Fig. 2B, lane 6). Elution of His-SUG1 from the mAb using an excess of epitope peptide displayed dose-related 3Ј-5Ј helicase activity (data not shown). Taken together these data demonstrate that SUG1 possesses a specific 3Ј-5Ј DNA helicase activity.
Helicases couple the energy of ATP hydrolysis to the disruption of the hydrogen bonds formed between the Watson-Crick base pairs of duplex DNA (32). A requirement for ATP hydrolysis was tested by generating a His-SUG1 protein bearing a point mutation in the putative ATP binding domain. The lysine residue at position 196 in domain I (5) was mutated to histidine (Fig. 1B) and the recombinant mutant SUG1 protein expressed and purified as above. Unlike wt His-SUG1, equimolar amounts of the mutant protein had no detectable helicase or ATPase activity (Fig. 3). The purification of the mutant and wt His-SUG1 was repeated several times, and in all cases the helicase and ATPase activities of each mutant preparation were negative in comparison with equimolar amounts of wt His-SUG1, thus eliminating the possibility that the helicase activity associated with wt His-SUG1 could be due to a copurifying contaminant helicase. We conclude from these experiments that SUG1 is a 3Ј-5Ј DNA helicase whose activity requires an intact ATP binding domain.
Monomeric SUG1 has a molecular mass of 48 kDa and is known to dimerize in vitro (5). To determine the relative size of endogenous SUG1 in the cell nucleus, human HeLa cell nuclear extracts were centrifuged through a 20 -40% glycerol gradient. No further manipulations were performed on the crude extract to ensure that high molecular weight complexes remain intact (see Ref. 33). The sedimentation of HeLa cell SUG1 was compared with that of two other subunits of the PA700 complex (S7 (MSS1, molecular mass ϭ 49 kDa) and S4 (molecular mass ϭ 56 kDa) (14,34)) and two components of the RNA pol II holoenzyme (the largest subunit of RNA pol II (molecular mass ϭ 220 kDa) and the p62 subunit (molecular mass ϭ 62 kDa) of TFIIH (27)) as well as markers of known molecular mass (Fig.  4). Most of the hSUG1 protein co-sedimented with the two other subunits, S7 and S4, of the PA700 complex (ϳ700 kDa) (Fig. 4, fractions 11-15). However, unlike the PA700 subunits S7 and S4, a portion of hSUG1 sedimented in fractions corresponding to much higher molecular masses that also contained a portion of the 220-kDa subunit of RNA pol II and of the p62 subunit of TFIIH (Fig. 4, fractions 22-26). Approximately 15-20% of nuclear SUG1 was also present in the low molecular mass range fractions (Fig. 4, fractions 2-7), in which very little S7 or S4 could be detected, thus indicating the possible existence of nuclear SUG1 complexes smaller than 700 kDa. Similar results were obtained with three different nuclear extracts (data not shown). DISCUSSION SUG1 is characterized here as a 3Ј-5Ј DNA helicase based on the following biochemical evidence: (i) the helicase activity co-fractionates with His-SUG1 over DEAE and affinity chromatographic steps; (ii) the level of helicase activity correlates with the amount of stained and immunoreactive His-SUG1 present in the chromatographic fractions; (iii) monoclonal an- tibodies raised against peptides derived from the SUG1 sequence specifically remove the helicase activity from these fractions; (iv) the helicase activity remains bound to the mAbprotein G resin after high stringency washes (500 mM KCl) and can be eluted by addition of an excess of epitope peptide; (v) a single amino acid mutation in the ATP binding domain eliminates both ATPase and helicase activities from the purified mutated His-SUG1, making it very highly unlikely that a contaminating protein could be responsible for the helicase activity present in the wt His-SUG1 preparations. In this respect, it is also important to stress that the purifications of wt and mutated SUG1 have been repeated several times with similar results.
SUG1 shares sequence similarities with the known helicase protein XPB and with the M. genitalium homologue of the Escherichia coli Holiday junction helicase ruvB, consistent with consensus sequences identified in a comparison of other known helicases (Fig. 1). However, our present knowledge concerning the structure/function relationship for helicase proteins is limited. Although the putative helicase M. genitalium homologue of the E. coli Holiday junction helicase ruvB was identified by BLAST searches for proteins with sequence homology with SUG1, it required manual sequence alignment with DExH/D box-containing helicase consensus domains and functional analysis to convincingly define SUG1 as 3Ј-5Ј DNA helicase. Similarly, only two putative helicases were identified in BLAST searches using the full amino acid sequences of the known DNA helicase protein XPD and none using the fulllength XPB sequence. Moreover, SNF2 and MOT1, which contain the conserved helicase motifs, lack any detectable helicase activity (35,36). Thus, the domain(s), in addition to the ATP binding domain, which are required for helicase activity, remain(s) to be characterized.
DNA helicases are known to be involved in several cellular processes, including DNA transcription, replication, repair, and recombination (26,31,32,37), and are implicated in a number of genetically inherited diseases (Ref. 38 and references therein). It has been shown that SUG1 interacts directly with several transcription factors, including (i) activators, such as GAL4, VP16, and several nuclear receptors (Refs. 4 -6 and references therein) and (ii) basal factors such as TBP and TAF II 30 (5)(6)(7). We also found that SUG1 directly interacts with XPB and co-purifies through several steps with the transcription⅐DNA repair complex TFIIH. 3 This may account for its presence in some RNA pol II holoenzyme preparations (10) and its co-sedimentation through glycerol gradients with components of the holoenzyme. It is therefore conceivable that SUG1 helicase activity is involved in some aspect of transcription and/or DNA repair. However, SUG1 is also a component of the PA700 complex (see Ref. 39 for review), 2 a regulatory subunit of the 26 S proteasome which recognizes and possibly unfolds ubiquitinylated proteins for proteolysis (39). Interestingly, histone-ubiquitin conjugates have a tendency to concentrate in the nucleosomes of genes being transcribed (40), but are absent during mitosis (41). Thus, as a component of the PA700 complex, SUG1 may be involved in stripping ubiquitinylated proteins from DNA (32). Note, however, that the PA700 exists in cells separate from the 26 S proteasome (Fig. 4 and Ref. 39) and may have functions distinct from proteolysis. As SUG1 is also present in low molecular mass complexes distinct from the PA700 and RNA pol II holoenzyme, it may also be involved on its own, through its ligand-enhanced interaction with the nuclear receptors, in local alteration of the DNA topology, thereby participating in the remodeling of the chromatin template of target genes (23 and references therein).
In conclusion, the helicase function of SUG1 may be involved in multiple cellular functions contingent upon SUG1 proteinprotein interactions. Ultimately, an understanding of the physiological roles played by the helicase activity of SUG1 will require the production of conditional SUG1 mutants in animal cells. FIG. 4. Sedimentation of HeLa nuclear extract through a 20 -40% glycerol gradient indicates that endogenous hSUG1 is associated with multiple complexes within the cell nucleus. Upper panel, the relative sedimentation of the largest subunit of RNA pol II (RNA pol II), p62 subunit of TFIIH (p62), the S4 and S7 subunits of the PA700 complex, and SUG1 was determined by repeated immunostaining of two Western blots (fractions 2-15 and 16 -29, respectively). Positions of markers of known molecular mass are indicated at the top. Unlike other proteins, immunoreactive SUG1 was found in all tested fractions with relative peak concentrations in the molecular mass ranges of 67-140 kDa (fractions 2-7) and 669 kDa (fractions [11][12][13][14][15][16]. SUG1 was also present in very high molecular mass ranges (fractions [22][23][24][25][26]. Lower panel, fractions 10 -20 were re-run to clearly show that SUG1 co-sedimented with the PA700 complex, but was also present in fractions of higher and lower molecular mass.