The Basic Helix-Loop-Helix Protein, SHARP-1, Represses Transcription by a Histone Deacetylase-dependent and Histone Deacetylase-independent Mechanism*

Many aspects of neurogenesis and neuronal differentiation are controlled by basic helix-loop-helix (bHLH) proteins. One such factor is SHARP-1, initially identified on the basis of its sequence similarity to hairy. Unlike hairy , and atypically for bHLHs, SHARP-1 is expressed late in development, suggestive of a role in terminal aspects of differentiation. Nevertheless, the role of SHARP-1 and the identity of its target genes remain unknown. During the course of a one-hybrid screen for transcription factors that bind to regulatory domains of the M 1 muscarinic acetylcholine receptor gene, we iso- lated the bHLH transcription factor SHARP-1. In this study, we investigated the functional role of SHARP-1 in regulating transcription. Fusion proteins of SHARP-1 tethered to the gal4 DNA binding domain repress both basal and activated transcription when recruited to either a TATA-containing or a TATAless promoter. Furthermore, we identified two independent repression domains that operate via distinct mechanisms. Repression by a domain in the C terminus is sensitive to the histone deacetylase inhibitor trichostatin A, whereas repression by the bHLH domain is insensitive to TSA. Furthermore, overexpression of SHARP-1 represses transcription from the M 1 promoter. This study represents the first report to

Structurally, bHLH proteins share a number of common features. The HLH domain mediates homomeric or heteromeric dimerization (7), and the adjacent basic region mediates DNA binding. Three groups of bHLH proteins can be recognized, according to the target binding site they recognize (8 -10). Class A and Class C bHLHs function as transcriptional activators and repressors, respectively, whereas class B bHLH proteins can be either activators or repressors.
Recently, the cDNA for SHARP-1, a novel bHLH protein, was isolated on the basis of its homology to Hairy and Enhancer of Split (11). However, sequence alignment showed that SHARP-1 is only distantly related to these proteins, exhibiting 37-42% sequence identity within the bHLH domain. Unlike most bHLH proteins, SHARP-1 is not expressed in neuronal progenitor cells or early differentiating neurons but is restricted to a subset of neurons of the postnatal central nervous system (11), suggestive of a role in terminal neuronal differentiation rather than in neural determination. Unlike all other HAIRY/E(spl)/HES proteins, SHARP-1 lacks the hallmark WRPW domain, which binds the co-repressor GROUCHO (or GROUCHO-like proteins) and is required for both transcriptional repression and suppression of neurogenesis. Absence of the WRPW motif suggests that SHARP-1 functions by recruiting transcriptional machinery other than GROUCHO. All class B bHLH proteins contain an arginine at position 13 in the basic region, essential for these proteins to bind to class B sites. The presence of an arginine at this position in SHARP-1 suggests that it belongs in this group, but because this group contains both activators and repressors, SHARP-1 function cannot be predicted on the basis of protein sequence.
At present, almost nothing is known about the role of bHLH proteins in differentiated neurons, and in common with many other bHLH proteins, no target genes of SHARP-1 are known. In the present study, we ascribe a transcriptional function to SHARP-1 and identify the M 1 muscarinic acetylcholine receptor gene as a target gene. We show that SHARP-1 is able to repress transcription of both TATA-containing and TATAless promoters when recruited via a Gal4 DNA binding domain (DBD). Repression occurs when SHARP-1 is bound either proximally or more distally to the promoter. Furthermore, we show that repression by SHARP-1 is bimodal. One mode of repression requires the bHLH domain and is insensitive to the histone deacetylase inhibitor, TSA, whereas the other is mediated via the C-terminal domain and represses transcription through a TSA-sensitive mechanism. We also show that overexpression of SHARP-1 represses transcription of a reporter construct containing the M 1 promoter. These results show that within the HAIRY-related/HES family, SHARP-1 is unique in its com-bination of presumed biological function and transcriptional mechanism.
Yeast One-hybrid Screening-pBM2389 ϩ417/ϩ166 M 1 was transformed into the yeast strain SFY526 (16). This yeast strain was then transformed with DNA from an adult rat brain cDNA yeast expression library (CLONTECH), using the protocol of Schiestl and Gietz (17), and transformants were grown on complete supplement mixture ϪHis/ ϪLeu/ϪTrp (Bio 101, Vista, CA), containing 5 mM 3-amino-1,2,4-triazole (Sigma) to select for interactions. Library candidates were tested for their ability to specifically activate the M 1 -containing reporter plasmid by retransforming back into SFY526. Library plasmids producing interacting proteins were sequenced for identification.
DNA Transfections-Qiagen column-purified DNA was transfected into cells using Tfx 50 (Promega) according to the manufacturer's instructions. Briefly, cells were plated onto 10-mm wells to a density of 50%. For IMR32 and 3T3 cells, Immunoprecipitation Assay-Neuro 2a cells were plated onto 10-cm plates to a density of 50%. Cells were incubated for 3-4 h with 10 g of DNA and 22.5 l of Tfx 50 in a final volume of 4.8 ml. Cells were harvested after 2 days into 1 ml of 1ϫ phosphate-buffered saline containing 0.5% Nonidet P-40 and protease inhibitors Block (Roche Molecular Biochemicals), sonicated for 90 s, and centrifuged at maximum speed for 10 min at 4°C. The supernatant was precleared for 2 h with 80 l of protein G-Sepharose. For each immunoprecipitation, half of the total sample was incubated with 3 l of Gal4 DBD antiserum (Santa Cruz Biotechnology) overnight. Beads were added, and samples were incubated for an additional 2 h. Samples were washed four times with 20 mM Tris, pH 8.0, 1 mM EDTA, 100 mM NaCl, 0.5 mM Nonidet P-40, 10% glycerol, and 0.1% SDS. Proteins were eluted with 15 l of loading dye. Samples were run on a 10% SDS-polyacrylamide gel electrophoresis gel and blotted onto a Hybond Cϩ nylon membrane (Amersham Pharmacia Biotech). The membrane was subjected to Western blot analysis using a 1/1000 dilution of c-myc antiserum (Santa Cruz Biotechnology). 1 Promoter-In our previous studies, we have shown that transcription of the M 1 muscarinic acetylcholine receptor gene is regulated by several domains within the first exon (12,18). In particular, the region between ϩ166 and ϩ412 (relative to the transcription start site) appears to contain both enhancer and repressor elements. To identify transcription factors that bind to this region of the M 1 gene we used the yeast one-hybrid approach (19,20). Using the ϩ166/ϩ417 domain as bait to screen an adult rat brain cDNA library, we isolated two independent positive clones (Fig. 1a, FIG. 1. SHARP-1 binds to the M 1 promoter. a, an adult rat brain cDNA library was screened using the region of the M 1 gene between ϩ166 and ϩ417 as bait and the yeast strain SFY526. Candidate clones were retransformed into SFY526-pBM2389 ϩ417/ϩ166 M 1 and plated onto ϪTrp/ϪHis/ϪLeu plus 10 mM 3-amino-1,2,4-triazole (3AT). Two positive clones were isolated (number 3 and 10), which failed to grow with either empty reporter vector, pBM2389, or with empty expression vector, pGAD10, showing that growth was dependent upon the interaction of the expressed library protein with the M 1 bait. b, digestion of the isolated library clones with BglII showed that both clones contain a cDNA of 1.5 kDa. Sequencing of the two clones showed that they both encompass the entire SHARP-1 open reading frame and also 130 base pairs of the 5Ј untranslated region. colonies 3 and 10). Digestion of the isolated clones showed that they both contain an insert of 1.5 kilobases (Fig. 1b). Sequencing of the inserts in these clones showed that they were identical and contained the entire open reading frame of the previously identified transcription factor SHARP-1 (11). Growth of yeast on plates containing 10 mM 3-amino-1,2,4triazole was dependent upon the binding of SHARP-1 to the M 1 sequence, because no growth was seen in the absence of either bait or SHARP-1 (Fig. 1a).

SHARP-1 Binds to the M
SHARP-1 Has High Homology to SHARP-2, Stra13, and DEC1-SHARP-1 was originally isolated during the course of a search for mammalian bHLH proteins expressed in differentiated neurons (11). Although SHARP-1 was isolated by homology to hairy and Enhancer of Split, sequence alignment with these proteins shows that they are quite distantly related, sharing only 37-42% homology within the bHLH domain (11).
As a first step toward identification of a function for SHARP-1, we carried out a data base search for proteins with homology to SHARP-1 and identified three proteins: SHARP-2, Stra13, and DEC1 ( Fig. 2). SHARP-2 is a bHLH protein isolated in the same screen as that used to identify SHARP-1 (11), and its function is also unknown. Stra13 was isolated as a retinoic acid-inducible gene in mouse P19 embryonic carcinoma cells and has been shown to be able to repress the thymidine kinase promoter when fused to Gal4 DBD (21). Finally, DEC1 is a protein that was cloned by subtractive hybridization to identify mRNAs expressed in cAMP-differentiated human embryo chondrocytes (22). Again, no function for DEC1 has been reported. Inspection of amino acid sequences shows that SHARP-2, Stra13, and DEC1 contain 411 or 412 amino acids, of which 366 are conserved, showing a sequence identity between them of 89%, suggesting that they are, in fact, rat, mouse, and human homologues. SHARP-1 is more divergent and contains only 253 amino acids. The highest sequence identity is seen in the bHLH domain and in helices 3 and 4 (also called Orange domain (23,24)), whereas within the C-terminal domain only two stretches of 8 and 11 amino acids are conserved.
SHARP-1 Homodimerizes-All bHLH proteins dimerize to bind DNA (25). Because SHARP-1 was identified in the present study using a yeast one-hybrid screen, it seemed likely that SHARP-1 could either homodimerize or heterodimerize with a yeast partner. To distinguish between these possibilities, we carried out an immunoprecipitation assay using differentially tagged recombinant SHARP-1. Neuro 2a cells were transfected with either a combination of myc-tagged SHARP-1 (pMT SHARP-1) and myc-tagged Gal4 DBD (pMT G4) (Fig. 3, lanes 1  and 3) or a combination of myc-tagged SHARP-1 (pMT SHARP-1) and a myc-tagged fusion of Gal4 DBD and SHARP-1 (pMT G4 SHARP-1) (Fig. 3, lanes 2 and 4). Cell extracts were immunoprecipitated with Gal4 DBD antiserum and subjected to polyacrylamide gel electrophoresis, and the results were visualized by Western blot analysis using c-myc antiserum. An antibody to Gal4 was able to co-immunoprecipitate SHARP-1 only in the presence of a GAL4-SHARP-1 fusion protein (Fig. 3,  compare lanes 4 and 3), demonstrating that SHARP-1 is able to homodimerize.
Expression of SHARP-1 in Different Cell Lines-It has been shown previously that expression of SHARP-1 is largely restricted to differentiated neurons in the postnatal central nervous system, predominantly in the cerebellum and hippocampus, although it is also detectable at a reduced level in heart, muscle, and lung (11). We examined expression of SHARP-1 in different cell lines and cerebellum using reverse transcription-PCR. As seen in Fig. 4a, SHARP-1 is highly expressed in IMR32 cells, a human neuroblastoma cell line that also expresses M 1 , Neuro 2a cells, and NB4 1A3, two mouse M 1 FIG. 2. SHARP-1 is closely related to SHARP-2/Stra13/DEC1. Shown is a sequence alignment of SHARP-1 with other bHLH proteins identified in a data base search for proteins related to SHARP-1. SHARP-2 (11), Stra13 (21), and DEC1 (22) share 89% sequence identity between them and appear to be homologues derived from rat, mouse, and human, respectively. SHARP-1 is highly conserved within the bHLH domain with SHARP-2/Stra13/DEC1 (95% sequence identity) and within helix III and IV, also called Orange domain (60% sequence identity). Outside of these domains, SHARP-1 diverges from the other three proteins, apart from two stretches of 8 and 15 amino acids, respectively, within the C-terminal domain. non-expressing neuroblastoma cell lines. Low levels of expression were detected in the 3T3 fibroblast cell line. PCR was carried out using hprt primers as a cDNA loading control (Fig.  4b).
SHARP-1 Acts as Transcriptional Repressor-bHLH proteins can act as transcriptional activators or repressors (reviewed in Ref. 1). Because the transcriptional function of SHARP-1 is unknown, we assessed its ability to (a) regulate transcription from both TATA-containing and TATAless promoters, (b) regulate transcription when bound either proximally or distally, and (c) regulate basal and activated transcription. IMR32, 3T3, and Neuro 2a cells were transfected with plasmids expressing SHARP-1 fused to Gal4 DBD with each of the reporter genes showed in Fig. 5. pTRE UAS TATA contains seven TRE and five Gal4 binding sites 21 base pairs upstream of the E1b TATA box. pGL3 UAS TRE TATA contains five Gal4 binding sites (placed 350 base pairs from the TATA box) and seven TRE upstream of the TATA box. In pGL3 TRE UAS Inr, the TATA box from pGL3 TRE UAS TATA, was replaced by the adenovirus major late promoter initiator. Expression values of all reporter constructs were normalized to expression in the presence of Gal4 DBD alone. SHARP-1 was able to repress transcription of a TATA-containing promoter when bound proximally to the transcription start site in all cell lines (Fig. 5a, left). SHARP-1 was also able to repress activated transcription by Tet-VP16 (activation domain of the herpes simplex virus transcriptional activator VP-16 fused to the binding domain of the tetracycline-responsive factor) in all cell lines (Fig. 5a, right). In addition, SHARP-1 was able to repress both basal and activated transcription from a TATA-containing promoter when bound distally to the transcription start site (Fig.  5b). We also tested the ability of SHARP-1 to regulate transcription from a TATAless promoter. As can be seen in Fig. 5c, SHARP-1 can repress basal and activated transcription from a TATAless promoter. We therefore conclude that SHARP-1 acts as a repressor of both basal and activated transcription of both TATA-containing and TATAless promoters. In the case of a TATA-containing promoter, repression is evident when SHARP-1 is bound either proximally or distally to the promoter, although the degree of repression is more marked when SHARP-1 is tethered proximally.
SHARP-1 Represses Transcription through Two Independent Domains-To map the domain(s) responsible for the repression function of SHARP-1, we generated deletion mutants of the Gal4-SHARP-1 fusion protein. The ability of these fusion proteins to repress transcription was analyzed using the reporter gene driven by a TATA-containing promoter with five Gal4 binding sites proximal to the transcription start site (pTRE UAS TATA) in Neuro 2a cells (Fig. 6). Results were normalized to expression of the reporter gene in the presence of Gal4 DBD alone. Western blot analysis showed that all constructs were expressed at similar levels (data not shown). Deletion of the C-terminal domain of SHARP-1 (to give pMT G4 NbHO-SHARP-1) slightly relieved repression by SHARP-1, but the C-terminal domain (residues 174 -253) of SHARP-1 fused to Gal4 DBD (pMT G4 C-SHARP-1) was able to repress transcription as robustly as full-length SHARP-1. Therefore, it would appear that SHARP-1 must contain at least two independent repression domains, one in the C-terminal domain and another in the remaining fragment. To map the second repression domain of SHARP-1, more deletion mutants were examined. Deletion of the Orange domain and C-terminal domain to give pMT G4 NbH-SHARP-1 still gave robust repression, but further deletion of the bHLH domain to leave only the N-terminal domain (pMT G4 N-SHARP-1) led to relief of most of the repression activity, suggesting that the bHLH domain also mediates repression. This was confirmed by analysis of two further constructs. Fusion of the Orange domain and flanking sequence (residues 103-173) and the Gal4 DBD (pMT G4 O-SHARP-1) showed some degree of repression, but fusion of the bHLH domain (residues 43-102) and Gal4 DBD (pMT G4 bH-SHARP-1) indicated that the bulk of repression in this second region was mediated by the bHLH domain. In summary, we identified two independent repression domains in SHARP-1, one in the bHLH domain and the other in the C terminus.
SHARP-1 Represses Transcription through Two Different Mechanisms-Recent studies have shown that many transcriptional repressors exert their action through recruitment of histone deacetylase activity (see Ref. 26 for review). To test whether SHARP-1 represses transcription through such a mechanism, we treated Neuro 2a cells with the histone deacetylase inhibitor TSA (27) and examined the effect on SHARP-1-mediated repression (Fig. 7). For each concentration of TSA used, expression values of the reporter gene were normalized to those obtained in the presence of Gal4 DBD alone, and results were expressed as fold over untreated cells. Repression by full-length SHARP-1 is partially relieved by TSA, because expression of the reporter gene was derepressed 4-fold in the presence of 100 nM TSA. Deletion of the C-terminal domain of SHARP-1 (to give pMT G4 NbHLHO-SHARP-1) showed that repression mediated by the bHLH domains was much less sensitive to TSA, resulting in a 1.6-fold derepression by 100 nM TSA. However, repression by the C-terminal domain alone was relieved by 6.5-fold with 100 nM TSA, and 2-fold derepression could be seen in the presence of 10 nM TSA. These results show that the C-terminal domain of SHARP-1 represses transcription via a mechanism that is likely to involve histone deacetylase activity but that the bHLH domain represses transcription in a histone deacetylase-independent manner. , and mouse cerebellum (lane 6). Lane 7 shows the water negative control, and the two outer lines correspond to a 1-kilobaseϩ DNA marker ladder (Life Technologies, Inc). PCR products were electrophoresed through a 2% MetaPhor agarose gel. b, PCR was performed on the same cDNA samples as above, using primers for hypoxanthine-guanine phosphoribosyl transferase gene hprt 231 s and hprt 567a to verify that equivalent amounts of cDNA had been assayed.

SHARP-1 Is a Transcriptional Repressor
functional effect of SHARP-1 on M 1 expression, IMR32, 3T3, and Neuro 2a cells were transfected with a reporter vector containing the region of the M 1 gene between Ϫ372 and ϩ602 (relative to the transcription start site). This construct has been shown before to be capable of driving expression in IMR32 cells, a neuronal cell line that expresses the M 1 gene (18). The same construct does not drive expression in the non-M 1 -expressing neuronal cell line Neuro 2a and drives only low levels of expression in 3T3 cells (18). Overexpression of SHARP-1 (pSHARP-1 myc) had no effect on expression of the promoterless reporter vector pGL3 basic but reduced expression driven by the M 1 promoter by 55% in IMR32 cells (Fig. 8a). A similar effect was seen in 3T3 cells (Fig. 8b). No effect was seen in Neuro 2a, the neuronal cell line that does not express M 1 (data not shown). These results show that SHARP-1 is able to repress expression of the M 1 gene in an M 1 -expressing cell line. DISCUSSION bHLH proteins are key players that regulate many aspects of development and differentiation in all tissues and phyla. To date, no target genes or function of SHARP-1 has been identified. SHARP-1 is unusual in two respects. First, SHARP-1 is related to, but distinct from, HAIRY/E(spl)/HES bHLH proteins. Second, expression of SHARP-1 appears to be restricted to postnatal neurons of the central nervous system, rather than neural progenitors, implying a role in late neuronal differentiation rather than neurogenesis. These features suggested that SHARP-1 may affect transcriptional regulation and target promoters distinct from those used by other members of the HAIRY/E(spl)/HES bHLH family.
SHARP-1 contains an Arg in position 13 of the basic region present in all class B bHLH proteins. This residue is essential for class B proteins to bind the consensus sequence CA(C/T)GTG (9). Sequence analysis of the region between ϩ166 and ϩ417 of the M 1 gene does not indicate the presence of any known recognition consensus motif (class A, B, or C) for bHLH (8 -10), suggesting that SHARP-1 may recognize a novel binding site. Gel electrophoresis mobility shift assays have failed to demonstrate an ability of Stra13 to bind to either an E-box or N-box (21). Because SHARP-1 and SHARP-2/Stra13/ DEC1 are all highly conserved within their basic regions or presumptive DNA binding domain, it is possible that both SHARP-1 and SHARP-2 recognize a common binding site, distinct from class A, B, or C sites.
By recruiting SHARP-1 to heterologous promoter constructs through the Gal4 DBD, we have shown that SHARP-1 acts as a transcriptional repressor and furthermore that SHARP-1 is able to repress both basal and activated transcription (Fig. 5).
It is well documented that arrangement of basal promoter elements can profoundly influence the response of a promoter to different factors. The evolutionarily conserved Kruppelassociated box present in the N-terminal regions of most Kruppel-class zinc finger proteins specifically silences the activity of promoters whose initiation is dependent on the presence of a TATA box (28), whereas initiator-containing promoters are relatively unaffected. Similarly, Oct2 isoforms repress transcription only from TATA-containing promoters (29). Here, we show that SHARP-1 is more promiscuous and can repress transcription driven from both TATA-containing and TATAless promoters (Fig. 5). In addition to promoter type, the relative position of its DNA binding site to the promoter can also influence repressor action. This can be seen in the case of the neuronrestrictive silencer factor (20, 30), which represses transcrip- tion when recruited distally but can activate transcription when recruited proximally (31). 2 Here again, SHARP-1 appears to act as a more global repressor and is capable of repressing transcription from a TATA-containing promoter when recruited either proximally or distally. Nevertheless, repression is more marked when SHARP-1 is tethered proximal to the TATA box (Fig. 5).
Transcriptional repression can take many forms (32). Several repressors function by deacetylating the N-terminal tails of histones that are thought to render chromatin inaccessible to the transcriptional machinery. Such is the case of the Mad family of bHLH-ZIP proteins, whose members bind to mSIN3A, an adapter molecule that links histone deacetylases to DNAbound transcription factors (33,34). Other repressors appear to interact directly with the transcriptional machinery itself, for example Kruppel, which at high concentrations homodimerizes and becomes a potent repressor by interacting directly with TFIIE (35,36). Others such as retinoblastoma can interact both with general transcription factors or with histone deacetylase (37)(38)(39)(40). Repression by SHARP-1 appears to be bimodal. One mode of repression is mediated via the bHLH domain and is TSA-insensitive, whereas the other mode is mediated by the C-terminal domain and is TSA-sensitive (probably due to recruitment of HDAC). Interestingly, Stra13 has been shown to repress expression of the c-myc gene through an HDAC-independent pathway and to negatively autoregulate its own promoter through an HDAC-dependent mechanism (41). The region responsible for the HDAC-independent repression of Stra13 has not been mapped, but glutathione S-transferase pull-down studies showed that residues 111-343 are required for interaction with HDAC-1, Sin3, and NcoR (41), whereas functional analysis of mutated Stra13 demonstrated that residues 147-354 were required for Stra13 repression of VP16activated transcription. As shown in Fig. 2, this region contains the C-terminal domain and part of the Orange domain of SHARP-2/Stra13/DEC1. We have mapped the domain responsible for the TSA-sensitive component of SHARP-1 repression to the C-terminal domain (residues 174 -253). Within this region, SHARP-1 and Stra13 display pockets of homology, principally within two stretches of 8 (residues 205-211) and 15 amino acids (residues 221-236), suggesting that these may mediate the interaction with HDAC. The HDAC-independent repression domain of SHARP-2/Stra 13/DEC1 has not been reported, but the GROUCHO-independent mode of repression by HAIRY/E(spl) requires the bHLH and Orange domains (23). Although SHARP-1 and SHARP-2 are almost identical in their bHLH domains (95% over 59 residues), they diverge more markedly from HAIRY (37% over 59 residues). It remains to be seen whether the same domains or mechanisms are involved in mediating HDAC-independent repression by SHARP-1, SHARP-2/Stra13/DEC1, and HAIRY/E(spl) proteins.
Bimodal repression is not uncommon. Repression by the neuron-restrictive silencer factor (20,  a C-terminal domain that recruits a novel corepressor Co-REST (45). Repression by retinoblastoma is also mediated via an HDAC-1-dependent and HDAC-1-independent mechanism (37, 38 -40). In the latter case, some promoters such as theadenovirusmajorlatepromoterarerepressedbytheHDAC-1dependent arm, whereas transcription activated by the SV40 enhancer or driven by the thymidine kinase promoter is repressed by the HDAC-insensitive arm. It may be that the breadth and selectivity of repressor action of bimodal repressors such as SHARP-1 is enhanced by their ability to bring one or both repression mechanisms to bear upon different promoters.
Although a function for SHARP-1 is unknown, its presence in terminally differentiated neurons suggests that it may be involved in regulating gene expression in differentiated neurons. Consistent with this hypothesis is our finding that SHARP-1 can bind to elements within the 5Ј untranslated region of the M 1 muscarinic receptor gene (Fig. 1) (12, 18). To date, the gene structures for three of the family of muscarinic receptor genes have been identified: the rat M 4 (47,48), the chicken M 2 (49), and the rat M 1 genes (12). Although the expression patterns of the M 1 and M 4 genes are very similar, transcription of the two genes appears to be controlled by different mechanisms. Previously, we and others have shown that the M 4 gene contains a constitutively active core promoter that is silenced, at least in non-neuronal cells, by the transcriptional repressor REST/neuron-restrictive silencer factor (50,51). On the other hand, we have previously shown that transcription of the M 1 gene is controlled by several elements within the 5Ј untranslated region acting both in a positive manner and in a negative manner (18). Within this fragment, we have identified a polypyrimidine/polypurine tract (from ϩ412 to ϩ485) and a conserved region across species (from ϩ504 to ϩ602) that act in concert to enhance expression in M 1 -expressing cells (18). Interestingly, this gene is expressed by differentiated neurons, and its expression increases in the first postnatal weeks (46,52) with an ontogenic profile similar to SHARP-1 (11). We show that overexpression of SHARP-1 represses expression of a reporter gene driven by the M 1 promoter in M 1 -expressing cells, suggesting that SHARP-1 can modulate levels of M 1 expression. By advancing our knowledge of the molecular mechanisms employed by SHARP-1 and identifying a potential target gene, these studies provide a platform for understanding the role of SHARP-1 in regulating neuronal gene expression.