Stra13 homodimers repress transcription through class B E-box elements.

A mammalian basic helix-loop-helix protein known variably as Stra13, Sharp2, and Dec1 has been implicated in cell activation, proliferation, and differentiation. Indeed, Stra13 null mice develop age-induced autoimmunity as a result of impaired T-lymphocyte activation, leading ultimately to the accumulation of autoreactive T-cells and B-cells. Stra13 is expressed in embryonic as well as adult tissues derived from neuroectoderm, mesoderm, and endoderm and has been associated with response to hypoxia, suggesting a complex role for this protein and the highly related Sharp1/Dec2 protein in homeostatic regulation. Whereas Stra13 is known to regulate many important cellular functions and is known to cross-regulate biological responses to other basic helix-loop-helix containing transcription factors, including c-Myc and USF, it is unclear if this protein binds directly to DNA. Indeed, the basic domain of Stra13 contains a proline residue at an unprecedented position. Herein, we have determined that Stra13 binds with high affinity to CACGTG class B E-box elements as a homodimer with preference for elements preceded by T and/or followed by A residues. In addition, transient transfection experiments reveal that Stra13 represses transcription when bound to these and related sites. Our data suggest that Stra13 regulates cellular functions through antagonism of E-box activator proteins and also through active repression from E-box elements.

Basic helix-loop-helix proteins represent a large and diverse class of transcription factors implicated in cell fate specification, cell proliferation, apoptosis, metabolism, and cell activation (1). The Stra13, Sharp2, Dec1 basic helix-loop-helix transcription factor has been identified in a number of biological contexts (for simplicity we will refer to this protein as Stra13).
For example, Stra13 was identified as a retinoic acid-inducible gene that promotes neuronal differentiation in P19 embryonal carcinoma cells (2). In addition, Stra13 and a related protein, Sharp1/Dec2, were identified in a degenerate PCR screen for bHLH 1 proteins expressed in the adult rat brain (3). Interestingly, Stra13 and Sharp1/Dec2 were both induced as immediate early genes in cultured PC12 pheochromocytoma cells treated with nerve growth factor, and Stra13 was rapidly induced by glutamatergic stimulation throughout the rat cerebral cortex (3). This gene was also identified as a cAMP-inducible transcript in differentiating chondrocytes (4) and was later found to be cAMP-inducible in many cell types (5). More recent work has described Stra13 induction following T-cell activation (6), tyrosine kinase receptor signaling (7), hypoxia (8 -10), and even serum starvation (11). Taken together, these data indicate that Stra13 expression is closely associated with activation and stress in many cell types.
Recently, Sun et al. (6) have used gene targeting to generate Stra13 Ϫ/Ϫ mice. Surprisingly, homozygous Stra13 mutant mice are born and survive to adulthood. However, aging Stra13 mutant mice develop an autoimmune disorder. This effect has been traced to impaired CD4ϩ T-cell activation, with reduced interleukin-2 production, reduced clonal expansion, impaired T-cell differentiation, and reduced clearance of activated lymphocytes (6). Despite indications that Stra13 may regulate activation and stress in a number of cell types, Stra13-binding DNA elements have yet to be identified. Consequently, the effect of this transcription factor on its theoretical DNA target(s) is also unknown. The putative Stra13 DNA-binding/ dimerization domain is somewhat related to bHLH domains of enhancer of split family (E(spl))/Hes transcriptional repressor proteins. However, Stra13 contains a proline residue in a distinct location within the basic domain. Therefore it is not clear whether Stra13 even binds directly to DNA. In addition, a fusion construct between the GAL4 DNA-binding domain and Stra13 can repress transcription from GAL upstream activator sequence sites through the recruitment of a histone deacetylase and perhaps through direct effects on the basal transcription factor TFIIB (2,11). However, it is not clear whether Stra13 functions as a DNA-binding repressor, a co-repressor, or perhaps even a transcription activating protein that is behaving inappropriately when fused to GAL4 (12). Indeed, the conformation of a GAL4-Stra13 fusion protein on DNA would be dramatically different from the conformation of Stra13 bound through its own putative DNA-binding domain. Consequently, the biochemical mechanism by which Stra13 functions remains unknown. Here we report identification of the Stra13 DNA target site, the well described CACGTG E-box element, with preference for this site preceded by T and/or followed by an A residue. In addition we have determined that Stra13 can function as a homodimer to repress transcription from these sites. These data identify Stra13 as a transcriptional repressor protein that functions to regulate E-box elements through cross-interference with E-box-binding transcriptional activators that bind such sites and through direct transcriptional repression from the class B E-box and related sites.
Plasmid Vectors-Murine Stra13 was cloned from mammary gland cDNA through a combination of reverse transcriptase PCR and high stringency hybridization based on the identification of an expressed sequence tag sequence, which at the time represented a fragment from a novel bHLH domain containing cDNA. 2 The Stra13 constructs described in the following section were subcloned into pcDNA3 for expression in transfected COS-7 and HC11 cells (Invitrogen). FLAG-Stra13 and Stra13-Myc were created using PCR to add an in-frame epitope tag at the N-and C-terminal ends, respectively, of the Stra13 open reading frame. The FLAG epitope tag is DYKDDDDK, which is recognized by the M2 monoclonal antibody (Sigma), whereas the Myc epitope tag is EQKLISEED, and is recognized by the 9E10 monoclonal and A-14 rabbit polyclonal antibodies (Santa Cruz Biotechnology).
The Stra13 C-terminal deletion mutants were created using pBSK-FLAG-Stra13 as a substrate for exonuclease III (Exo-Size deletion kit, New England Biolabs). For this purpose, a double-stranded oligonucleotide was inserted downstream of the Stra13 stop codon. It contained a unique HindIII restriction site (exonuclease III-sensitive restriction site), a unique SpI restriction site (exonuclease III-resistant restriction site), and three frame stop codons. The precise extent of each deletion was determined by nucleotide sequencing analysis. Based on their distribution, particular deletion mutants were selected for further studies.
The epitope-tagged Stra13 basic domain mutant FLAG-Stra13(⌬basic) was created by linking through a chimeric XbaI site two PCR-generated Stra13 cDNA fragments corresponding, respectively, to FLAG-tagged amino acids 1-49 and amino acids 67-136 of the Stra13 open reading frame. The resulting linked mutant cDNA fragment in which the basic domain is specifically deleted from the Stra13 sequence was used to replace the corresponding wt N-terminal Stra13 sequence through fusion of the ⌬basic N-terminal fragment with the Stra13 C terminus using an internal XhoI. FLAG-Stra13(acidic) was created by inserting a double-stranded oligonucleotide coding for an in-frame peptide rich in acidic residues (EEEDDEEE, see Krylov et al. (15)) in the unique XbaI site of FLAG-Stra13(⌬basic). For luciferase reporter vectors pGL3-(tCACGTGa) 3 and pGL3-(tCACGTGa) 2 -SV40, oligonucleotide probe 1 (see below) was inserted at a unique NheI site of the Promega pGL3 and pGL3-SV40 vectors. The sequence of pGL3-(tCACGTGa) 3 from SacI to XhoI is as follows: GAGCTCTTACGCGTG-CTAGCAATCCTTGTGTCACGTGACGTCTAGCAATCCTTGTGTCAC-GTGACACAAGGATTGCTAGCAATCCTTGTGTCACGTGACGTCTA-GCCCGGGCTCGAG. The sequence of pGL3-(tCACGTGa) 2 -SV40 from SacI to XhoI is as follows: GAGCTCTTACGCGTGCTAGCAATCCTTG-TGTCACGTGACACAAGGATTGCTAGACGTCACGTGACACAAGGA-TTGCTAGCCCGGGCTCGAG. Mutated E-box versions of pGL3-(tCA-CGTGa) 2 -SV40 (i.e. pGL3-(tCAGGTGa) 2 -SV40, pGL3-(tCATATGa) 2 -SV40, and pGL3-(tCACGGAa) 2 -SV40) were created by double-stranded oligonucleotide replacement of the SacI-XhoI fragment of the multiple cloning site of pGL3-SV40. The oligonucleotides were designed to identically reproduce the sequence of pGL3-(tCACGTGa) 2 -SV40 except for the mutated nucleotides at the designated positions of the two E-boxes. pGL3-(tCACGTGa) 3 -TK was created by subcloning the thymidine kinase promoter into the blunt-ended HindIII site of pGL3-(tCACGTGa) 3 .
Binding Site Selection and EMSA Assays-The PCR-based site selection protocol was performed as previously described (16). For electrophoretic mobility shift assays (EMSA), 5 g of each complementary single-stranded oligonucleotide were annealed in a final volume of 50 l of 50 mM Tris-HCl (pH 8), 10 mM MgCl 2 , 50 mM NaCl. 0.4 g of annealed oligonucleotides was labeled with [ 32 P]dCTP for 1 h at room temperature with Klenow (Invitrogen), separated from free nucleotides by passage through NICK columns (Amersham Biosciences), and eluted into 300 l of Tris/EDTA for a probe concentration of 1.3 ng/l. Cold competitors (40 ng/l) were prepared in the same manner using Klenow and non-radioactive dNTPs. Proteins were synthesized using 1 g of linearized DNA template in the TNT Quick transcription/translation kit (Promega) according to manufacturer's specifications. Protein-DNA complexes were formed by incubation of protein (4 l from a 50-l reaction) with 2.5 ng of radiolabeled double-stranded oligonucleotide in a final volume of 20 l (10 mM Tris, pH 8, 40 mM KCl, 6% glycerol, 1 mM dithiothreitol, 0.05% Nonidet P-40) with a total of 500 ng of poly(dI/dC), 2 g of BSA, and 0.1 g of salmon sperm DNA. Binding reactions were performed at room temperature for 20 min. For supershift experiments, 1 g of mouse anti-FLAG monoclonal antibody (Sigma) was added to binding reactions. DNA-protein complexes were resolved by electrophoresis on 5% acrylamide gel, after which gels were dried and exposed to BioMax film with an intensifying screen at Ϫ70°C. Double-stranded oligo probe 2 (CTAGACGTCACGTGCCACAAGGATTGCTAG), probe M1 (CTAGACGTCAGGTGACACAAGGATTGCTAG with optimal flanking residues), probe A1 (CTAGTTGGCAGGTGCCAAAAGGATT-GCTAG without flanking T and A residues), and probe M2 (CTA-GACGTCATATGACACAAGGATTGCTAG) were used for Fig. 2. In data not shown oligo probe 1 (CTAGACGTCACGTGACACAAGGATT-GCTAG) was found to bind Stra13 with equal if not higher affinity compared with oligo probe 2. Additional probes with multiple mutations in the CACGTG consensus did not bind Stra13 homodimers nor did they compete Stra13 off of probe 2 (data not shown).
Immunoprecipitation/Western Blotting-1 g of primary antibody was added to soluble protein lysates and incubated at 4°C for 1 h (COS-7 lysates) or overnight (HC11 lysates). Secondary goat antimouse IgG coupled to agarose beads (Sigma catalog no. A-6531) was added to the immunoprecipitation reaction, which was then incubated for an additional 30 min at 4°C. Beads were washed six times using 1ϫ Triton X-100 lysis buffer (without protease or phosphatase inhibitors) and resuspended by boiling in sample buffer containing 40 mM dithiothreitol. The resulting denatured protein samples were first separated electrophoretically on SDS-PAGE gels and then transferred onto nitrocellulose membranes. Western blots were blocked in 5% BLOTTO (10 mM Tris, pH 7.5, 150 mM NaCl, 0.05% (v/v) Tween 20, and 5% (w/v) skim milk powder) and then probed with the appropriate primary rabbit polyclonal antibody (0.5 g/ml) in 1% BLOTTO for 1 h. The membranes were then rinsed (4 ϫ 15 min) in 1% BLOTTO followed by incubation with peroxidase-conjugated goat anti-rabbit (Jackson ImmunoResearch Laboratories) at a dilution of 1:10 4 for 1 h in 1% BLOTTO. Finally, the membranes were rinsed with 1% BLOTTO (4 ϫ 15 min) followed by immunodetection using Amersham ECL Western blotting detection reagents according to the manufacturer's guidelines.
Luciferase/␤-Galactosidase Assays-For experiments shown in Figs. 3 and 4, COS-7 cells or HC11 cells were transfected with 100 ng (Fig. 3, A-D), 25 ng (Fig. 3E), or 50 ng (Fig. 4) of luciferase reporter vector along with increasing amounts of pcDNA3 vector expressing either wt or mutant Stra13. In each case the observed luciferase activity for each transfected sample was corrected according to the ␤-galactosidase activity generated from a co-transfected Rous sarcoma virus-or cytomegalovirus-lacZ vector. The average corrected luciferase activity and standard deviation of the duplicate samples for each test condition were plotted. The average activity observed in the absence of co-transfected Stra13-expressing construct was given a relative value of 100%. For each experiment, levels of ectopic Stra13 protein were determined through immunoprecipitation/Western blot analysis on duplicate sam-2 B. St-Pierre and S. E. Egan, unpublished data. ples transfected with 4 g (COS-7) or 10 g (HC11) of Stra13 expression vector. Each experiment shown was a representative of at least three independent results. To determine the luciferase activity of each sample, 10 l of soluble protein lysate was mixed with 25 l of Promega luciferase reagent, and the light-emitting activity of the sample was immediately determined for 20 s using a Berthoid Lumat LB9501 luminometer. To determine the ␤-galactosidase activity from a sample, 20 l of soluble protein lysate was mixed with 200 l of reaction buffer (400 g/ml o-nitrophenyl-␤-D-galactopyranoside, 0.1 M NaPO 4 (pH 7.5), 10 mM KCl, 1 mM MgCl 2 , and 50 mM ␤-mercaptoethanol). Color development was allowed to proceed for 15-60 min at 37°C and was quantified using a SpectraMax250 Elisa plate reader (Molecular Devices).

RESULTS AND DISCUSSION
Stra13 Homodimerizes in Vivo-In a screen to clone bHLHencoding genes expressed in the mouse mammary gland, we identified a cDNA that was subsequently identified in a number of tissues and cell lines. 2 This gene, Stra13, represents the founding member of a novel family of genes encoding bHLH proteins with an unusual basic domain sequence (2)(3)(4)17). bHLH proteins typically form homodimers and/or heterodimers in order to bind DNA target elements (1). In vitro, Stra13 has previously been shown to form homodimers and heterodimers with other bHLH proteins, including E47 and Mash1 (2,17). To test whether Stra13 forms homodimers in vivo, we transfected COS-7 cells with two different epitope-tagged versions of Stra13 (Fig. 1A). In lysates from cells cotransfected with Myc epitope-tagged Stra13 and FLAG epitope-tagged Stra13, anti-FLAG immunoprecipitates contained Myc-tagged Stra13. In control lysates from cells singly transfected with Myc-tagged Stra13, anti-FLAG antibodies could not immunoprecipitate Myc-tagged Stra13. These data reveal that Stra13 homodimerizes in transfected COS-7 cells. Since HLH domains are typically responsible for dimerization of bHLH transcription factors, we tested if an N-terminal fragment of Stra13 (FLAG-Stra13-(1-143)) containing little more than the bHLH domain was sufficient for homodimerization. As expected anti-FLAG antibodies immunoprecipitated Myc-tagged Stra13 from lysates of cells cotransfected with FLAG-Stra13-(1-143) and Stra13-Myc but not from lysates of cells transfected with either cDNA alone (Fig. 1B). Next, we tested whether the basic portion of the bHLH domain was necessary for homodimerization by removing it or by replacing it with an acidic domain (15). Basic domain deleted and acidic HLH variant Stra13 proteins were both able to dimerize with wild type Myc-tagged Stra13 (Fig. 1C). Finally, we determined the subcellular localization of FLAG-tagged wild type Stra13, FLAG-tagged Stra13-(1-143), FLAG-tagged Stra13⌬basic, and the FLAG-tagged Stra13 acidic variants in transfected COS-7 cells and HC11 mouse mammary epithelial cells. In each case Stra13 protein was found in the nucleus (Fig. 1D).
Stra13 Homodimers Bind to the Class B E-box, CACGTG-Most bHLH proteins recognize short DNA elements that are related to the E-box CANNTG. However, as Stra13 contains a proline residue at an unusual position in its basic domain, it is not clear whether this protein binds directly to DNA. We used a PCR-based binding site selection protocol to test for a Stra13 homodimer binding site in vitro (16). Because the bHLH domain of Stra13 is related to enhancer of split family (E(Spl))/ Hes bHLH domains and these proteins were previously shown to bind N-box elements (CANNAG) (18, 19), we first established very low stringency conditions under which in vitro translated Stra13 could bind to N-box elements (specifically CACGAG). This occurred in the absence of BSA and poly(dI/ dC) nonspecific competitors. This low affinity Stra13 N-box complex was used as a molecular weight marker to purify specific Stra13 complexes from EMSA gels in the early rounds of our site selection protocol. After three rounds of Stra13 binding, EMSA gel purification of bound oligonucleotides, and PCR amplification we had enriched high affinity Stra13-binding sites to a level where we would detect radioactive Stra13-DNA complexes on the EMSA gel. The Stra13bound DNA pool was amplified by PCR, cloned, and sequenced (16). This revealed enrichment for one type of 6-nucleotide site in 23 of 26 clones that contained the predicted 70-bp insert (Table I). Interestingly, the Stra13 binding site identified in our screen was CACGTG, which is a class B E-box element (20). In addition, the CACGTG sequence was frequently preceded by a T residue and followed by an A residue, although these are not absolutely required for high affinity binding (see below).
To test for the importance of individual bases on Stra13 binding to the E-box identified in our screen, we performed EMSA assays to test for binding to tCACGTGc sites (probe 2) under high stringency conditions in the presence of BSA and poly(dI/dC) ( Fig. 2A) (21). This element bound a protein or proteins present in the control in vitro transcription/translation reaction. However, when the in vitro transcription/translation reaction was programmed with a Stra13-FLAG epitopetagged cDNA and this was incubated with probe 2, a novel Stra13-containing complex was formed that could be super-

Stra13 Binds CACGTG Elements and Represses Transcription
shifted through inclusion of anti-FLAG monoclonal antibody. The anti-FLAG antibody did not supershift control complexes. We next tested for formation of Stra13 complexes on related sites. For example, we determined whether Stra13 bound to tCAGGTGa and gCAGGTGc to test for the importance of residues within the selected E-box consensus and for the importance of flanking T and A residues that were selected in our screen. Interestingly, Stra13 formed a specific complex with tCAGGTGa as determined by the supershift of this complex in the presence of anti-FLAG antibody in the binding reaction. This site has a single mutation in the core E-box but with optimal flanking residues. In contrast, Stra13 did not form complexes with gCAGGTGc, which has the same E-box core mutation but without flanking T and A nucleotides. Next, we studied Stra13-DNA interactions by testing whether various E-box elements could compete Stra13 off of probe 2, tCACGTGc. Unlabeled excess tCACGTGc (probe 2), tCACGTGa (probe 1, not shown), gCACGTGc (probe B1, not shown), and gCACGTGt (probe B2, not shown) diminished radioactive Stra13-probe 2 complex formation ( Fig. 2B and data not shown). Interestingly, excess unlabeled CAGGTG elements including tCAGGTGa (mutant probe M1), gCAGGTGc (probe A1), and gCAGGTGg (probe A2) did not compete Stra13 from labeled probe 2 ( Fig. 2B and data not shown). In addition, the single mutant gCACGCGa element (hairy probe) (data not shown) and mutant E-box elements that differ from the CACGTG consensus by 2 nucleotides could not compete with probe 2 for binding to Stra13 (Fig. 2B and data not shown). Thus, Stra13 binds with high affinity to CACGTG elements. It can also bind to single mutant CAGGTG elements, although only when these elements are surrounded by optimal flanking residues ( Fig. 2A) and only with significantly lower affinity than CACGTG elements (this element did not compete for Stra13 binding in Fig. 2B).
Stra13 Is a Transcriptional Repressor of Class B E-box Elements-In previous work, GAL4-Stra13 fusion proteins were shown to repress transcription from upstream activator sequence elements that bind GAL4 (2). To determine whether Stra13 activates or represses transcription from its cognate DNA target site, three copies of the tCACGTGa element were inserted upstream of the minimal and thymidine kinase promoters in the luciferase vectors pGL3 and pGL3-TK, respectively. These constructs were transfected into COS-7 cells that do not express detectable Stra13 protein and also into HC11, a mouse mammary epithelial line that expresses endogenous Stra13. 2 The addition of E-box elements dramatically increased the luciferase activity generated by these vectors in both lines (Fig. 3, A and C), presumably because of the expression and effect of endogenous E-box-binding transcription-activating proteins. Cotransfection of FLAG-Stra13 repressed expression of either promoter in a dose-dependent fashion (Fig. 3, A and C). Interestingly, the N-terminal FLAG-Stra13-(1-143) bHLH domain construct was as active as full-length Stra13 in repressing transcription in these assays, suggesting that repression was occurring as a result of Stra13 proteins competing with endogenous E-box activator proteins for access to CACGTG elements (Fig. 3, A and C). To confirm that repression by Stra13 involved competition with endogenous E-box binding transcription factors for target CACGTG elements, we tested whether FLAG-Stra13(⌬basic) and FLAG-Stra13(acidic) mutants could function as repressors in this system. Both basic domain mutants were completely unable to repress transcription from the CACGTG elements at concentrations where wild type Stra13 inhibited transcription by ϳ90% (2 g) (Fig. 3, B and D). This

FIG. 2. Stra13 binds to CACGTG E-box elements.
A, electrophoretic mobility shift assay using radiolabeled tCACGTGc-, tCAG-GTGa-, or gCAGGTGc-containing probes in the presence or absence of FLAG-Stra13 protein. The formation of specific FLAG-Stra13-DNA complexes was confirmed by supershifting of complexes in the presence of anti-FLAG antibody (indicated with an *). B, the affinity of FLAG-Stra13 for tCACGTGc elements was tested by competition using excess non-radiolabeled tCACGTGc, tCAGGTGa, gCAGGTGc, or tCATATGa sequences as indicated.

FIG. 3. Competitive repression by Stra13 through class B (CACGTG) E-box elements. COS-7 cells (A, B, and E) or HC11 cells (C and D)
were transfected with luciferase reporter vector along with increasing amounts of a pcDNA3 vector expressing either wt Stra13 or mutant Stra13, or both, as indicated. The observed luciferase activity for each transfected sample was corrected according to the ␤-galactosidase activity generated from a co-transfected lacZ vector (see "Experimental Procedures"). E, COS-7 cells were transfected with pGL3-(tCACGTGa) 3 luciferase reporter vector in the presence or absence of 0.1 g of pcDNA3-FLAG-Stra13 vector. To test the ability of the Stra13 basic domain mutants to act as Stra13 Binds CACGTG Elements and Represses Transcription 46548 effect was associated with disruption of DNA binding as both proteins were stable (Fig. 1C) and found in the nucleus (Fig. 1D).
To test whether the basic domain mutant proteins could function to sequester wild type Stra13 away from E-box elements, we cotransfected Stra13 together with FLAG-Stra13(⌬basic) or FLAG-Stra13(acidic). The repressor function of wild type Stra13 was blocked by coexpression of either of these dimerization-competent mutant proteins (Fig. 3E).
We next tested whether Stra13 could actively repress transcription in the context of a complex promoter. Insertion of two copies of the tCACGTGa element upstream of the SV40 promoter in the pGL3-SV40 luciferase vector did not affect gene expression in COS cells (Fig. 4A). Co-transfected Stra13 inhibited transcription of this CACGTG-containing promoter in a dose-dependent manner but did not repress the parental pGL3-SV40 reporter (Fig. 4A). To test whether specific E-box elements were required for Stra13-mediated transcriptional repression we inserted three Stra13 mutant binding sites, tCAGGTGa, tCATATGa, and tCACGGAa, into the pGL3-SV40 reporter. Interestingly, the single nucleotide mutant site tCAGGTGa was still Stra13-responsive (Fig. 4B), consistent with the fact that tCAGGTGa elements bound Stra13 in vitro ( Fig. 2A). In contrast, the double nucleotide mutant elements tCATATGa and tCACGGAa that did not bind Stra13 in vitro were not responsive to Stra13 in vivo. Repression from E-box elements required an intact basic domain in Stra13, because it was not observed when FLAG-Stra13(⌬basic) or FLAG-Stra13(acidic) mutants were co-transfected with pGL3-(tCACGTGa) 2 -SV40-Luc (Fig. 4C). In addition, this effect represented active repression because in contrast to the competitive repression observed on minimal promoters (Fig. 3) repression of the SV40 promoter required sequences in the C-terminal 268 residues of Stra13 (data not shown). To define sequences in Stra13 responsible for active repression of the SV40 promoter in this context, we generated and cotransfected a series of C-terminal truncation mutants (Fig. 4, D and E). The C-terminal boundary of the transcriptional repression domain was mapped to lie between residues 322 and 333, because C-terminal truncation mutants including FLAG-Stra13-(1-333) were potent repressors, whereas FLAG-Stra13-(1-322) and smaller C-terminal truncation mutants had lost this activity (Fig. 4D). The inability of FLAG-Stra13-(1-322) to repress transcription was not caused by a lack of expression of this mutant protein (Fig. 4D).
Stra13 is a bHLH protein associated with cell activation and stress in many tissues. Indeed, this protein is induced in activated neuronal cells, chondrocytes, T-cells, fibroblasts, and a number of cancer cell lines. Recent genetic analysis has revealed that Stra13 is required for T-cell activation and regulation of lymphocyte clearance (6). In addition, Stra13 has recently been implicated in hypoxia-induced repression of adipogenesis (22). Despite the importance of Stra13 in these biological processes, the biochemical mechanism by which it functions remains unknown. For example, it had yet to be determined whether Stra13, which includes a conserved proline residue at an unprecedented site within its basic domain, binds directly to DNA. We have used a non-biased screen to determine that Stra13 binds to a specific DNA element, the CACGTG class B E-box element, with preference for sites preceded by T and followed by A. Following submission of this work, Zawel et al. (23) have recently reported that human Stra13 binds CACGTG elements in vitro and can repress transcription from reporters containing these sites. Such sites regulate expression of many genes associated with proliferation, differentiation, and cell activation. We have also determined that Stra13 can actively repress transcription of promoters that contain class B E-box elements. The domain of Stra13 responsible for transcriptional repression from CACGTG elements maps to sequences in the N-terminal 333 amino acids of Stra13. This result is consistent with data from Boudjelal et al. (2) who used GAL4-Stra13 chimeras to repress transcription from upstream activator sequence sites and to map a repression domain between amino acids 147 and 354. In addition, Sun and Taneja (11) have determined that HDAC1 and NcoR bind to Stra13 sequences between amino acids 111 and 343. Interestingly, the Stra13 deletion mutant Stra13-(1-322) that failed to repress transcription (Fig. 4) still bound HDAC1, suggesting that Stra13 may mediate transcriptional repression via multiple mechanisms (data not shown).
These results are important given that Stra13 binds to and antagonizes USF proteins (24) and represses expression of c-Myc (11), two classes of bHLH-ZIP transcription activating proteins that bind to CACGTG E-box elements. Stra13 also represses expression of its own promoter through a mechanism requiring histone deacetylase activity (11). Interestingly, the Stra13 promoter contains three CACGTG elements located 261, 1125, and 2901 bases upstream of the first transcribed nucleotide (25). Recent work has identified the PPAR␥2 gene as a target of Stra13-mediated transcriptional repression during hypoxia-induced suppression of adipogenesis (22). Interestingly, the promoter element that responds to Stra13-mediated repression contains binding sites for C/EBP bZIP proteins but no obvious E-box elements like those identified in our study. In addition, the bHLH domain of Stra13 was sufficient to repress expression of PPAR␥2. Perhaps Stra13 represses PPAR␥2 expression and adipogenesis through direct inhibition of C/EBP function. Future studies will be necessary to resolve the mechanism by which Stra13 represses expression of Stra13 and PPAR␥2 promoters and to identify E-box elements in the genome that are subject to regulation by Stra13 and the Stra13related protein, Sharp1/Dec2. Indeed, the Stra13-binding/responsive E-box elements identified in this study are likely to represent critical nodes in a competitive network of activator and repressor proteins controlling cell activation, proliferation, and stress.