OUT, a Novel Basic Helix-Loop-Helix Transcription Factor with an Id-like Inhibitory Activity*

Transcription factors belonging to the basic helix-loop-helix (bHLH) family are involved in various cell differentiation processes. We report the isolation and functional characterization of a novel bHLH factor, termed OUT. OUT, structurally related to capsulin/epi-cardin/Pod-1 and ABF-1/musculin/MyoR, is expressed mainly in the adult mouse reproductive organs, such as the ovary, uterus, and testis, and is barely detectable in tissues of developing embryos. Physical association of OUT with the E protein was predicted from the primary structure of OUT and confirmed by co-immunoprecipi-tation. However, unlike other bHLH factors, this novel protein failed to bind E-box or N-box DNA sequences and inhibited DNA binding of homo- and heterodimers consisting of E12 and MyoD in gel mobility shift assays. Dishes coated with type I collagen (IWAKI GLASS, Japan) were used to culture C2C12 cells. For the transient transfections, NIH3T3 and C2C12 cells were plated at densities of 5 3 10 4 cells/25-mm well and 5 3 10 4 cells/35-mm well, respectively, in DMEM supplemented with 10% FCS 24 h before transfection. Transfections were performed by the lipofection method using TransIT-LT1™ (Pan Vera Corp.) according to the manufacturer’s instructions. The total amount of DNA added to cells was adjusted to 1.2 m g/25-mm well and 2.0 m g/35-mm well by addition of appropriate empty vector. Electrophoretic Mobility Shift Assays— Electrophoretic mobility shift assays (EMSA) were performed essentially as described previously (60).

Transcription factors with a basic helix-loop-helix (bHLH) 1 motif have been demonstrated to play critical roles in cell fate determination and differentiation in a variety of tissues of both vertebrates and invertebrates (1,2). Examples include myogenic bHLH factors such as MyoD and myogenin in skeletal muscle development (1)(2)(3)(4), SCL/TAL1 in hematopoiesis (5,6), and neuronal factors such as Mash1 and neurogenin in neurogenesis (7)(8)(9)(10)(11). The bHLH motif consists of a short region rich in basic amino acids and two amphipathic helices separated by an intervening loop region (12). The bHLH proteins form homoor heterodimers through the helix-loop-helix (HLH) domains, enabling the basic regions to form a bipartite DNA-binding motif that recognizes so-called E-box sequences, CANNTG, commonly found in the promoter or enhancer regions of numerous developmentally regulated genes (12). Typically, tissuespecific class B bHLH factors, such as MyoD and neurogenin, dimerize with ubiquitously expressed class A bHLH factors and promote cell fate determination and differentiation into specific lineages (12). Class A bHLH factors are exemplified by socalled E proteins, such as E2A gene products E12 and E47.
Cell differentiation is a complex and well organized process in which cells respond to stimuli from the environment by carrying out a genetic program. It has been shown that bHLH factors directly or indirectly regulate expression in a gene activation network. The best studied system is skeletal muscle development. Four myogenic bHLH factors, MyoD, Myf-5, myogenin, and MRF4, participate in the development of mammalian skeletal muscles (1,2). Although all of them can induce skeletal muscle differentiation in a wide variety of non-muscle cell types (13)(14)(15)(16)(17), expression analyses (14,18) and gene-targeting experiments indicate differences in their positions in the genetic network for myogenesis (3, 4, 19 -25). MyoD and Myf-5 play redundant roles in establishing myoblast identity of mesodermal progenitors (19 -21). Subsequently, myogenin promotes differentiation of myoblasts to myotubes and their maturation (3,4,23). MRF4 functions during the differentiation process of myoblasts, together with MyoD, as well as in the terminal stage (25). Combinatorial orchestration of growth factors and other transcription factors such as MEF2 is also involved in this gene activation network, leading ultimately to muscle development (1,26,27). Similar cascades of bHLH factors have been also demonstrated in neurogenesis (9).
In addition to these genetic networks of "positive" bHLH factors, "negative" HLH or bHLH factors enable the proper execution of the cell differentiation control through functional modulation of bHLH factors (11,28). The Id proteins, inhibitors of DNA binding/differentiation, are negative regulators of bHLH factors (28,29). They possess HLH domains and heterodimerize with bHLH factors, but, due to a lack of the basic region, the resultant heterodimers have no DNA binding activity. As a consequence, cell differentiation is inhibited. Four Id proteins, Id1-Id4, are expressed in a wide range of embryonic tissues and are believed to be involved in the expansion of immature cell populations (28 -32). The HES proteins are repressive bHLH factors mainly expressed in the developing nervous system (11,33,34). A homodimer of HES binds to the E-box-related N-box sequence and actively represses transcription by recruiting a co-repressor through the WRPW domain present in the C terminus (11,(35)(36)(37). HES, like Id, also sequesters bHLH factors (34). ABF-1 (38)/musculin (39)/MyoR (40), which is expressed in activated B lymphocytes and muscle precursors, binds to the E-box sequence but does not activate transcription. Instead, it represses E-box-mediated transactivation by competing for binding sites with positive bHLH factors and through a transcriptional repressive domain. Mist1 (41), Twist (42), and Stra13 (43) are also repressive bHLH factors with multiple inhibitory mechanisms. Among them, Stra13 is an exception, because it possesses no DNA binding activity.
The importance of the negative regulation of bHLH factors in cell differentiation has been emphasized by loss-of-function mutants in Drosophila and mice. For example, Drosophila mutants defective in orthologues of Id and HES, emc (44) and hairy (45), respectively, show developmental defects in the formation of sensory hairs. Mice lacking Id2 show loss of lymph node and Peyer's patch development and a defect in development of natural killer cells (46). HES-1-deficient mice demonstrate a defect in neural tube closure and microphthalmia due to premature differentiation of neurons (47,48). Moreover, inactivation of twist results in defective dorso-ventral patterning due to disturbed gastrulation in Drosophila (49,50) and defects in cranial neural tube closure and mesodermal derivatives in mice (51).
In this study, we identified a novel bHLH factor, OUT, using PCR with degenerate primers. OUT is expressed mainly in the adult mouse reproductive organs and is barely detectable in the developing mouse embryo. In gel shift and oligonucleotide selection assays, OUT failed to bind DNA. In the presence of OUT, E12 and MyoD were prevented from homo-and heterodimer formation and failed to induce E-box-mediated transactivation. By deletion analyses, the bHLH and C-terminal regions were identified as important domains for the inhibitory action of the OUT protein. Furthermore, introduction of OUT in C2C12 myoblasts hampered their terminal differentiation. These functional characteristics indicated that OUT possesses an inhibitory activity similar to that of Id.
cDNA Cloning and 5Ј-RACE-The putative bHLH domain of a novel HLH factor was obtained by the reverse transcription-PCR (RT-PCR) method using the following two degenerate primers: MESO-S CCAA-  (53,54).
One g of total RNA of mouse mammary glands at 14 days postcoitus (d.p.c.) was reverse-transcribed with oligo(dT) primer (Life Technologies, Inc.) using Moloney murine leukemia virus reverse transcriptase (Superscript II, Life Technologies, Inc.) in a total volume of 20 l. One l of the product was subjected to PCR with an Ex-Taq kit (TaKaRa) in a thermal cycler (Takara). The PCR product was subcloned into XcmI-digested pKRX (55) and sequenced. Among 35 clones analyzed, 2 encoded an identical novel bHLH protein.
An oligo(dT)-primed cDNA library was constructed from 3 g of poly(A) ϩ mouse ovary RNA using a TimeSaver TM cDNA Synthesis Kit (Amersham Pharmacia Biotech) according to the manufacturer's instructions. Lambda ZAP II (Stratagene) was used as the vector. Plaque hybridization using the fragment obtained above as a probe was performed at high stringency, and two positive clones were identified from one million independent phage clones. The phage clones were converted to plasmids by the in vivo excision system. The longer clone, pBS-mOUT, was used in this study. The nucleotide sequence of the clone was determined on both strands using an ABI 377 autosequencer (Perkin-Elmer).
5Ј-Rapid amplification of cDNA ends (RACE) was performed with the 5Ј-RACE system (Life Technologies, Inc.) according to the manufacturer's instructions. Amplification was performed using a nested primer set, TGAGGCTGTAGGCCCTAGAGCAGGGACACAGTACCC and TGCCTCTGTGGCCTCCTGTGACATGCCGCTATCATG (corresponding to cDNA nucleotides (nt) 50 -85 and 91-126, respectively). The 5Ј-RACE product was subcloned into XcmI-digested pKRX, and the sequences of 6 clones were analyzed using an ABI 377 autosequencer (Perkin-Elmer).
Northern Blot and RT-PCR Analyses-Twenty g of poly(A) ϩ RNA of adult mouse organs was separated by electrophoresis on a 1.0% agarose-formaldehyde gel, transferred onto filters, and cross-linked in a UV chamber. A radioactive DNA probe for OUT was prepared by randomprimed labeling of a 1.4-kb PstI-XbaI fragment of the OUT cDNA (nt 398 -1791). Hybridization and washing were performed under high stringency conditions as described previously (56). The full-length glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA was adopted as a probe for an internal loading control. X-ray films were exposed with an intensifying screen at Ϫ80°C for 72 h for OUT and for 12 h for GAPDH.
For RT-PCR analyses, 5 g of total RNA from various adult mouse organs was reverse-transcribed with random hexamer (Takara) in a total volume of 20 l using a standard protocol. One l of the product was subjected to PCR amplification using the following two primers: GCCACAAGCTACATTGCCCACCTC and TCATTTGTTACCAAAAGC-TGGAGA (corresponding to cDNA nt 460 -483 and 709 -732, respectively). For an internal control, two primers corresponding to the ␤-actin gene were utilized.
In Situ Hybridization-In situ hybridization was performed with paraffin-embedded sections of the uterus and ovary at 7 d.p.c. essentially as described previously (56). 35 S-Labeled antisense and sense riboprobes were prepared by in vitro transcription with suitable RNA polymerases following linearization of pCMV-OUT (see below) with appropriate restriction enzymes. The probe spanned nt 103-732 of the OUT cDNA. The samples were hybridized and washed at high stringency and autoradiographed with the emulsion of NTB2 (Eastman Kodak Co.).
For CASTing assays, the OUT expression vector tagged with the FLAG epitope (MDYKDDDDK) was generated by subcloning the coding region of OUT (nt 103-732) downstream of the FLAG epitope sequence in pCMV-FLAG-2 vector (Sigma). For use in in vitro translation experiments, the FLAG-OUT fragment was transferred to pBluescript using the SacI-BamHI site, generating pBS-FLAG/OUT. Similarly, pBS-FLAG/MyoD was generated from pCMV-FLAG-2 MyoD (a gift from Shosei Yoshida).
To generate deletion mutants of OUT, fragments corresponding to amino acid sequences indicated in Fig. 7A were obtained by PCR. The sense and antisense primers contained KpnI and XbaI sites in their 5Ј ends, respectively. For deletion of the N-terminal portion, the sequence spanning nt 283-306 was included in respective sense primers to equalize the translation efficiency. For deletion of the C-terminal portion, a stop codon was included in antisense primers. Amplified fragments were digested with KpnI and XbaI and inserted into the KpnI-XbaI site of the pCMV vector.
The authenticity of plasmids constructed by PCR was verified by sequencing.
Cell Cultures and DNA Transfections-NIH3T3 fibroblast and C2C12 myoblast cell lines were purchased from American Tissue Culture Collection and provided by Shosei Yoshida (Kyoto University), respectively. They were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS) plus 100 units/ml penicillin and 100 mg/ml streptomycin. Dishes coated with type I collagen (IWAKI GLASS, Japan) were used to culture C2C12 cells. For the transient transfections, NIH3T3 and C2C12 cells were plated at densities of 5 ϫ 10 4 cells/25-mm well and 5 ϫ 10 4 cells/35-mm well, respectively, in DMEM supplemented with 10% FCS 24 h before transfection. Transfections were performed by the lipofection method using TransIT-LT1™ (Pan Vera Corp.) according to the manufacturer's instructions. The total amount of DNA added to cells was adjusted to 1.2 g/25-mm well and 2.0 g/35-mm well by addition of appropriate empty vector.
Electrophoretic Mobility Shift Assays-Electrophoretic mobility shift assays (EMSA) were performed essentially as described previously (60).  (67) and an oligonucleotide containing an N-box from the HES-1 promoter (34) were annealed and end-labeled with [␣-32 P]dCTP using the Klenow fragment of Escherichia coli DNA polymerase I. The core sequences within these oligonucleotides were CAGGTG (E2), CACCTG (MCK-R, MLCA and MLCC), CAGCTG (MLCB, Hen1 consensus sequence, CE-2 and TnI E-box), CAGATG (EF1 and 8701), CATCTG (RIPE3), and CACGAG/ CACAAG (HES-1 promoter). For the competition assays, we designed a mutant E-box in which each core sequence CANNTG was converted to ACNNGT. For the mutant N-box, CACGAG and CACAAG were replaced by CCCGAG and CCCAAG, respectively. The oligonucleotides were 22-26-mers. In vitro transcripts containing the 5Ј-7mGpppG cap (New England BioLabs Inc.) were prepared from the linearized plasmid templates using appropriate RNA polymerase. Transcripts then were translated into proteins in vitro using a rabbit reticulocyte lysate system (Promega) according to the manufacturer's instructions.
Each protein involved in the DNA-protein complex was identified in supershift assays using anti-E12 antibody (Santa Cruz Biotechnology) and anti-MyoD antibody (PharMingen). The sequence specificity was confirmed by competition assays using a 100-fold excess of non-labeled wild-type E-box or mutant E-box (ACNNGT) sequences. The molar ratio of OUT to other bHLH proteins tested was 1:1, unless otherwise indicated.
CASTing-Cyclic amplification and selection of targets (CASTing) was performed essentially as described previously (68). Briefly, an 84-mer oligonucleotide containing 14 randomized bases flanked by 4 restriction sites (HindIII and PstI sites at the 5Ј end and BamHI and EcoRI at the 3Ј end) and priming sequences for PCR (M13 forward and reverse sequences at the 5Ј and 3Ј ends, respectively) was synthesized (M13 forward-HindIII-PstI-N 14 -BamHI-EcoRI-M13 reverse). The oligonucleotides were converted to double-stranded DNA by ExTaq (Takara) using the M13 reverse primer at 72°C for 30 min. The FLAG epitopefused OUT protein or FLAG epitope-fused MyoD protein were co-translated in vitro with E12 in a rabbit reticulocyte lysate system (Promega), using in vitro transcripts of linearized pBS-FLAG/OUT, pBS-FLAG/ MyoD, and pCMV-E12. The efficiency of in vitro translation was confirmed by performing translation in the presence of [ 35 S]methionine and analyzing the products by SDS-PAGE. Five l of in vitro translation product was then incubated with the double-stranded oligonucleotides in the binding buffer (final composition, 20 mM HEPES (pH 7.9), 5% glycerol, 50 mM KCl, 1 mM EDTA, 1 mM dithiothreitol) at room temperature for 20 min. After the addition of 5 l of anti-FLAG M2 affinity gel (A1205, Sigma), the mixture was incubated at room temperature for 1 h. Then the gel was washed three times in 500 l of washing buffer (1ϫ phosphate-buffered saline containing 0.1% bovine serum albumin and 0.1% Nonidet P-40) and resuspended in 100 l of PCR reaction mixture containing M13 forward and M13 reverse primers. Ten cycles of PCR were carried out, with each cycle consisting of denaturation at 94°C for 1 min, annealing at 65°C for 1 min, and elongation at 72°C for 1 min. Ten l of the PCR product was subjected to the next round of CASTing. After six rounds, the selected DNA was purified and subcloned into pBluescript using the BamHI and HindIII sites. CASTing was carried out twice independently.
Co-immunoprecipitations-For co-immunoprecipitations, COS-7 cells were plated at a density of 2 ϫ 10 6 cells/150-mm dish 24 h before the transfection. Thirty g of DNA was transfected by the lipofection method. Cells were harvested after 48 h of incubation in DMEM with 10% FCS. Nuclear extract prepared as described previously (69) was dialyzed against buffer consisting of 20 mM HEPES (pH 7.9), 5% glycerol, 50 mM KCl, 1 mM EDTA, and 1 mM dithiothreitol. Then 20 g of each nuclear extract was incubated in the same buffer for 2 h at room temperature with protein G-Sepharose (Amersham Pharmacia Biotech) that had been coupled with monoclonal mouse anti-human Myc antibody (9E10, Santa Cruz Biotechnology) or isotype-matched monoclonal mouse IgG 1 antibody (PharMingen). The precipitates were separated by SDS-PAGE, transferred onto polyvinylidene difluoride membranes (Immobilon P, Millipore), and subjected to Western blot analysis. For E12 detection, a standard procedure employing anti-E12 antibody was used. For Myc epitope detection, to avoid detecting the antibody included in the immunoprecipitates, primary and secondary antibodies were mixed to form a complex, and the free secondary antibody was blocked with mouse serum. This mixture was used as a probe. Secondary antibodies were conjugated with horseradish peroxidase, and conjugated and enhanced chemiluminescence reagents (Renaissance R , NEN Life Science Products) were used for visualization.
Luciferase Assays-As reporter plasmids, we utilized pE7-␤A-luc (70) for E-box-mediated luciferase assays and tk-GALpx3-LUC or tk-LUC (58) for GAL4 binding assays. The CMV promoter-driven sea-pansy luciferase plasmid, pRL-CMV (Promega), was used as an internal control to normalize firefly luciferase activity. NIH3T3 fibroblasts plated as described above were transiently co-transfected with 70 fmol of each expression vector together with 70 fmol of reporter plasmid and 7 fmol of pRL-CMV per 25-mm well. Before transfection, the total amount of DNA per well was adjusted to 1.2 g by addition of the pCMV empty vector or pEF BOS empty vector. After 48 h of incubation in DMEM with 10% FCS, the cells were lysed, and the luciferase activities were measured using the Dual-Luciferase TM reporter assay system (Promega) according to the manufacturer's instructions with a Lumat LB 9507 (EG & G Berthold) luminometer. The firefly luciferase activities were corrected by the CMV promoter-driven sea-pansy luciferase activity.
Differentiation of C2C12 Myoblasts-The myoblast differentiation assays were performed as described previously (71). C2C12 myoblasts plated as described above were transiently transfected with 0.5 pmol of each expression vector together with 0.25 pmol of pNLS/lacZ per 35-mm well. The plasmid of pNLS/lacZ (a gift from Nobutake Akiyama, Kyoto University) encodes E. coli ␤-galactosidase with a nuclear localization signal (NLS). The total amount of DNA added to C2C12 cells was adjusted to 2.0 g by addition of empty pCMV vector. After induction of differentiation in DMEM with 2% horse serum for 96 h, cells were fixed in phosphate-buffered saline containing 4.0% (w/v) paraformaldehyde and stained with 5-bromo-4-chloro-3-indolyl-␤-D-galactoside (X-gal) and then with anti-troponin T (TnT) antibody (Sigma). Differentiation was evaluated by counting the number of TnT-positive cells relative to that of ␤-galactosidase-positive cells.
Nucleotide Sequence Accession Number-The nucleotide sequence of OUT was deposited in the GenBank TM data base with the accession number AF142405.

RESULTS
Isolation of a Novel bHLH Factor, OUT-In an effort to identify novel bHLH factors, we performed RT-PCR analyses using the total RNA of the mammary gland of pregnant mice at 14 d.p.c. as the source. The bHLH transcription factors comprise a very large family, and the amino acid sequences of the bHLH regions are not highly conserved. We therefore designed several degenerate primer sets targeting the conserved sequences within various bHLH subfamilies. By using primers MESO-S and MESO-AS, which were designed based on the sequences of the mesodermally expressed bHLH proteins paraxis (53) and scleraxis (54), we obtained PCR products with an appropriate size comparable to that of the bHLH region. Sequencing of these products revealed a novel bHLH sequence consisting of 144 nucleotides including the primer sequences at both ends.
Since a preliminary Northern blot analysis using this fragment as a probe revealed a transcript in the adult mouse ovary, we next constructed a mouse ovary cDNA library and screened it with the 144-bp fragment as a probe. Two positive clones were identified among approximately 1 ϫ 10 6 independent phage clones. Restriction enzyme and sequence analyses indicated that these clones were overlapping. The longer clone, bearing a 4.1-kb cDNA insert, was used for further analyses. Nucleotide sequence analysis revealed that this clone contained a 4100-bp cDNA with a single open reading frame of 630 bp (Fig. 1). The novelty of the gene was confirmed by homology searches against data bases. Two possible initiation codons were found, and both of them closely matched the Kozak con-sensus sequence (72). As shown below, the size of the cDNA insert was approximately 100 bp shorter than that of the transcript detected by Northern blot analysis. To obtain information about the 5Ј-terminal region of the mRNA, we performed 5Ј-RACE using two specific primers designed to hybridize near the 5Ј end of the cDNA and isolated fragments containing an additional 108-bp sequence (4 and 2 clones contained 107-and 108-bp inserts, respectively). This region contained 4 termination codons in the same reading frame (data not shown), demonstrating that translation is not initiated at a site upstream of those noted above. The cDNA sequence thus predicted 2 species of proteins consisting of 210 and 200 amino acids with calculated molecular masses of 22.9 and 21.9 kDa, respectively, depending on the translation initiation site (Fig. 1). In vitro translated products of this gene migrated on SDS-PAGE with apparent molecular weights consistent with the calculated masses of the two proteins (Fig. 4B, lanes 7-10 in the lower panel). Whether these isoforms have functional differences remains to be determined.
We designated this novel factor as OUT, on the basis of the main organs that express this gene, the ovary, uterus, and testis, as shown below.
OUT Is Related to Mesodermal bHLH Factors-Data base searching and motif analysis identified a bHLH motif spanning 56 amino acid residues in the middle of the OUT protein (amino acid residues 75 to 130) (Fig. 1), which closely conformed to the  Fig. 2. In the basic region of OUT, there are only a few basic amino acids, i.e. three arginine residues, although it preserves the motif of ERXR, which is a determinant of E-box recognition (81,82). In addition, an arginine residue positioned at the first consensus residue of the bHLH family is replaced by a serine residue in OUT. Of note, the basic region of OUT possesses one proline residue, as do the basic regions of repressive bHLH factors such as HES (33,34) and Stra13 (43), although the positions of the proline residues are not the same among them. A proline residue is also found in the corresponding region of the Id proteins (29 -32). The remainder of the sequence of OUT shows no apparent similarity to any previously described proteins or motifs. The nuclear localization of the OUT protein was verified by using a fusion protein between OUT and the green fluorescent protein (data not shown).
OUT Is Expressed in the Adult Reproductive Organs-To determine the expression pattern of OUT in the adult mouse tissues, Northern blot experiments were performed (Fig. 3A). Twenty g of poly(A) ϩ RNA from various adult mouse tissues were probed with a 1.4-kb radiolabeled fragment. This probe was designed to contain the 3Ј-half of the coding region and the following 1.0-kb 3Ј-untranslated region and to cover one of the putative splicing sites. As shown in Fig. 3A, a single transcript was detected, and its size was estimated to be 4.2 kb. This was about 100 bp longer than the cDNA isolated from the ovary cDNA library. The expression level was highest in the uterus, ovary, and testis, in that order. Faint expression was also noted in the lung, heart, intestine, and spleen.
We also analyzed the OUT expression by RT-PCR using the total RNA from the same set of tissues (Fig. 3B). The primer set was designed to cover one of the putative splicing sites and to give an expected product of 273 bp. Overall, the expression pattern obtained was identical to that seen with Northern blotting. With this method, in addition to the organs in which OUT was detected by Northern blot analysis, faint expression of OUT was detected in virtually all samples analyzed, includ-ing the mammary glands from which OUT was initially identified. No apparent fragment of any other size was present.
The reproductive organs are under the influence of hormone action, and the uterus, in particular, shows functional and morphological changes during pregnancy and delivery. To obtain more clues about OUT functions in vivo, we further analyzed the expression in the adult uterus according to the estrus cycle and gestational stages (Fig. 3C). OUT expression in the uterus was higher in the diestrus phase than in the estrus phase and reached a maximum at 7.5 d.p.c., thereafter declining toward the time of delivery. The level of OUT transcripts returned to the non-pregnant level 4 days after delivery.
To identify the cell types that express OUT in the uterus of the pregnant mouse, we next performed RNA in situ hybridization using an 35 S-labeled riboprobe (Fig. 3D). On the sections of the 7.5 d.p.c. uterus hybridized with the antisense OUT riboprobe, OUT expression was detected as double streaks that corresponded to the two layers of myometrium, the inner circular and the outer longitudinal muscle layers (83). On the serial section hybridized with the sense OUT riboprobe, no apparent signal was detected. As compared with control images hybridized with sense riboprobe, a faint signal also appeared to be present in the endometrium.
In contrast to the results in the adult, no detectable signal was observed in the developing embryos by Northern blot analysis using RNA from 7.5, 10.5, 11.5, 14.5, and 18.5 d.p.c. embryos and by whole mount in situ hybridization of 7.5, 8.5, and 9.5 d.p.c. embryos (data not shown). Furthermore, no OUT expression was detected in the uterus before puberty (data not shown).
OUT Does Not Bind DNA but Rather Inhibits DNA Binding of Other bHLH Proteins-Considering the deduced primary structure, the OUT protein was expected to be a transcription factor with a bHLH motif and to possess DNA binding activity specific for the E-box, which is a common target of the bHLH transcription factors (12). To test this, we performed electrophoretic mobility shift assays (EMSA) using 32 P-labeled EF1 oligonucleotide bearing the core sequence of CAGATG, one of the well known E-boxes (65). The proteins used were prepared by in vitro transcription of the template cDNA followed by in vitro translation in rabbit reticulocyte lysates. As shown in  7) and E12-MyoD heterodimers (lane 8). Similar results were obtained with 10 other oligonucleotides containing different E-box sequences that have been reported so far (data not shown, see under "Experimental Procedures"). In addition, there was no evident binding of OUT to an N-box, with which HES proteins, repressive bHLH factors, preferentially interact (data not shown, see under "Experimental Procedures"). Thus, we could not detect any DNA binding activity of OUT, but instead we found that it inhibited the DNA binding of other bHLH factors.
These functional features of OUT are reminiscent of those of the Id proteins, which lack the basic region but possess the HLH region (28,29). Id proteins form inactive heterodimers with bHLH factors and negatively regulate their function. Therefore, we next investigated the dose dependence of the inhibitory action of OUT using the E2 oligonucleotide, containing the core sequence of CAGGTG (61), in comparison with the inhibitory action of Id. As shown in Fig. 4B, the DNA binding activities of the E12-MyoD heterodimers and MyoD homodimers were attenuated, in a manner dependent on the dose of OUT protein added in the reaction mixtures (lanes 2 and 7-10). Meanwhile, Id2, one of the Id proteins (31), showed an activity similar to but stronger than that of OUT (lanes 2-6). Appropriate amounts of the respective proteins in each reaction were confirmed by SDS-PAGE of 35 S-labeled proteins as shown in the lower panel of Fig. 4B.
Cyclic amplification and selection of targets, CASTing, was next performed to identify OUT-binding DNA sequences, which might be different from the E-box or the N-box. FLAG epitopetagged OUT was co-translated with E12 in vitro. For a positive control, FLAG epitope-fused MyoD was used and similarly co-translated with E12. After incubation of in vitro translation product with double-stranded degenerate oligonucleotides, the mixture was precipitated with an anti-FLAG antibody, and the bound DNA was subjected to amplification by PCR. After six rounds of CASTing with the positive control, the bound DNA was detected by gel electrophoresis and subcloned into pBluescript. Sequence analyses indicated that all of 22 clones examined contained the E-box sequences, CANNTG (data not shown). However, no obvious DNA fragment was obtained from the tagged OUT-E12 complex, supporting the idea that OUT has no DNA binding activity (data not shown).
OUT Interacts Physically with Class A bHLH Factor E12-The results shown above suggested that the inhibitory effect of OUT is similar to that of Id. The main mechanism by which Id inhibits bHLH factors is to quench the activity of the E proteins. We therefore next explored protein-protein interaction between OUT and E12 using the co-immunoprecipitation method. Two kinds of Myc-tagged OUT expression vectors (a full-length and a mutant that lacks the bHLH region) were constructed and transfected into COS-7 cells together with pCMV-E12. As controls, Myc-tagged Id2 expression vectors (a full-length and a mutant lacking the HLH region of Id2) were prepared and analyzed. Nuclear extracts were subjected to the immunoprecipitation with anti-Myc antibody, and the precipitates were separated by SDS-PAGE and probed with anti-Myc or anti-E12 antibodies. The results are shown in Fig. 5. As anticipated, E12 was co-immunoprecipitated together with the Myc-tagged Id2 by anti-Myc antibody (Fig. 5, lane 2) but not with the HLH-deleted mutant Id2 (Fig. 5, lane 5). Similarly, an association of OUT with E12 was detected (Fig. 5, lane 8). E12 was not detected in the precipitate of the nuclear extract of cells transfected with the cDNA of E12 and mutant OUT lacking the bHLH region (Fig. 5, lane 11). The specificity of the immunoprecipitation was confirmed with an isotype-matched nonspecific mouse IgG 1 (lanes 3, 6, 9, and 12). These results indicate that the OUT protein forms a complex with the E12 protein by physical interaction through the bHLH domain. The smaller species of OUT products observed in in vitro translation (Fig. 4B, lower panel) was barely detectable in the expression system with COS-7 cells.
OUT Inhibits Transactivation Induced by bHLH Factors-To evaluate the effect of OUT on E-box-mediated transcription, luciferase assays were performed using NIH3T3 cells (Fig. 6). pCMV vectors expressing OUT, E12, MyoD, and Id2 were cotransfected with a reporter plasmid in various combinations indicated in Fig. 6. As anticipated, OUT failed to induce the transactivation over the basal level (lanes 2, 5, and 7). Next, the inhibitory effect of OUT on E12-MyoD-induced transactivation was studied at varying molar ratios and compared with the inhibitory effect of Id2. In the presence of Id2, E12-MyoDmediated luciferase activity was greatly reduced in a dose-dependent manner (lanes 6 and 8 -12). The reduction was about 50% even at the molar ratio of 0.125:1 (lanes 6 and 8) and about 80% at the molar ratio of 1:1 (lanes 6 and 11). OUT showed a similar effect, but the reduction was smaller at the same molar ratio; the reductions were about 50 and 70% at the molar ratios of 1:1 and 2:1, respectively (lanes 6 and 13-17). In addition, luciferase activity could be restored as the molar ratio of E12 and MyoD to OUT increased (data not shown).
To exclude the possibility that the inhibitory effect observed with OUT was the result of the overexpression of exogenous genes in NIH3T3 cells, we overexpressed the neural bHLH TAL2 (56) and placental bHLH Mash2 (84) in the same context, instead of OUT. As shown in Fig. 6B, although a 10 -20% reduction was caused by TAL2 at a 1:2 ratio (lane 16), no major inhibition of luciferase expression was caused by either TAL2 or Mash2 in the luciferase expression, excluding the above possibility.
These results demonstrated that the inhibitory effect of OUT on the E-box-mediated transactivation by bHLH factors was in accordance with the effects seen in the EMSA.
Delineation of the Functional Domain of OUT-To determine the region responsible for the inhibitory activity of OUT, various deletion mutants were constructed, as indicated in Fig. 7A. Each vector was co-transfected into NIH3T3 cells with a molar equivalent of pCMV-E12, pCMV-MyoD, and pE7-luc, and the luciferase activity was evaluated. Since mutants lacking the bHLH region exhibited no inhibitory activity (Fig. 7A rows 5,  11, and 12), the bHLH region was suggested to be essential for the function of OUT. However, the bHLH region alone did not reduce the induction of luciferase activity by E12-MyoD heterodimers ( Fig. 7A row 7). Additionally, inclusion of the whole N-terminal portion caused only a marginal inhibition (Fig. 7A  row 6). To investigate the effect of the C-terminal portion of the protein, we subsequently prepared the constructs consisting of the bHLH region and various parts of the C-terminal portion. These mutants showed inhibitory activity proportional to the length of their C-terminal regions (Fig. 7A rows 7-10). The results suggested that both the bHLH region and the C-terminal portion are essential for the inhibitory function of OUT. From the structural features of bHLH proteins (1, 2), it is most plausible that the bHLH region is the main functional domain that interacts with dimerization partners. On the other hand, the C-terminal portion of the protein is probably required to facilitate dimerization or to stabilize already formed heterodimers, as demonstrated for Id3 (85).
It has been reported that Stra13, an inhibitory bHLH factor, has no DNA binding activity but possesses a repressor domain (43). By using the GAL4 system (58, 59), we next attempted to elucidate whether OUT has a transcriptional repressor domain. The full-length and three portions (N-terminal region, bHLH region, and C-terminal region, indicated in Fig. 7B) of OUT were fused to the GAL4 DNA-binding domain (GAL4 DBD) under the control of the human elongation factor 1␣ promoter (pEF-BOS). These expression vectors were transfected into NIH3T3 cells with the firefly luciferase reporter plasmid carrying five repeats of GAL4-binding sites upstream of the thymidine kinase (tk) promoter. The reporter produced a  , lanes 7-10), which are probably due to alternative initiation of translation (see text). Two species of OUT proteins are also detected in translation products of OUT tagged with 6Myc in its C terminus (data not shown). The positions of each protein product are indicated on the left. Gels were dried and exposed to x-ray film for 48 h in both A and B.
high basal level of transcription activity. This activity was strongly suppressed by co-expression of transcriptional repressors KRAZ1 or KOX1, as reported (59), but not by the parental GAL4 DBD plasmid alone (Fig. 7B left panel, lanes 1-4). In this assay system, none of the OUT domains displayed an apparent repressive activity (Fig. 7B left panel, lanes 6 -8), although the full-length OUT showed a slight repression (Fig. 7B left panel,   lane 5). However, this repression was almost negligible as compared with the repressor activity induced by positive controls ( Fig. 7B left panel, lanes 3 and 4). Moreover, similar repression was detected with the other reporter plasmid lacking the GAL4 binding site (Fig. 7B right panel, lane 5), suggesting that the slight repression induced by the full-length OUT protein was due to a nonspecific effect on the transfected  6. OUT inhibits E-box-mediated transcription in luciferase assays. A, NIH3T3 fibroblasts were transiently transfected with pE7-␤A-luc reporter gene, cytomegalovirus (CMV) promoter-driven E12, and MyoD expression vectors (pCMV-E12 and pCMV-MyoD) with increasing amounts (molar ratio from 1:0.125 to 1:2) of cytomegalovirus promoter-driven OUT or Id2 expression vector (pCMV-OUT or pCMV-Id2) as indicated. The total amount of DNA used per well was adjusted to 1.2 g by addition of pCMV empty vector. Luciferase expression was evaluated as relative luciferase activity normalized by the sea-pansy luciferase activity produced by pRL-CMV which was co-transfected simultaneously. B, to evaluate the effect of overexpression of exogenous genes, pCMV-E12 and pCMV-MyoD were co-transfected into NIH3T3 fibroblasts with increasing amounts of pCMV vector carrying tal-2 or Mash2 cDNA as indicated. All assays were independently performed twice in triplicate. Error bars indicate S.E. FIG. 7. Deletion studies and evaluation of the transcriptional repressor activity of OUT. A, to delineate the functional domain of OUT, various deletion mutants of OUT were analyzed in the same E-box-mediated luciferase assay system used in Fig. 6. Schematic representations of deletion mutants are shown on the left. Id2 was used as a positive control. The results are expressed as relative luciferase activity. Error bars, S.E. B, the GAL4 binding assay was performed to identify a transcriptional repressor activity of OUT. Schematic representation of reporter plasmids Novel Inhibitory bHLH Factor 3518 cells. OUT thus seems to possess no apparent repressor domain, and inhibitory interaction with bHLH factors is the most plausible mechanism by which OUT exerts its inhibitory activity.
OUT Inhibits Differentiation of Myoblasts into Myotubes-C2C12 mouse myoblasts differentiate into myotubes under low serum conditions, providing a good experimental model system to examine the functions of bHLH transcription factors. We next used this system to ask if OUT indeed has an inhibitory effect on cell differentiation, similar to the effects of the Id proteins. OUT was introduced into C2C12 cells, and its effects on the terminal differentiation of muscles were compared with those of other HLH factors (Fig. 8). To identify the cells that incorporated exogenous DNA, a reporter plasmid containing the ␤-galactosidase gene with a nuclear localization signal was co-transfected. When the cells were transfected with myogenin, almost all (97.1%) of the ␤-galactosidase-positive cells differentiated into muscle cells, and more multinucleated cells were formed than when the cells were transfected with the other expression vectors (data not shown). In contrast, the terminal differentiation in cells transfected with OUT and Id2 was greatly suppressed, and only 40.3 and 36.6% of the cells differentiated into muscles, respectively. To evaluate the specificity of the inhibitory effect of OUT further, the neural bHLH factor TAL2 (56) and placental bHLH factor Mash2 (84) were heterologously expressed in C2C12 cells, and their effects on muscle differentiation were determined. In accordance with the results of the luciferase assays, myoblasts expressing Mash2 or TAL2 differentiated to muscles to an extent similar to the mocktransfected cells, demonstrating the specificity of the effect of OUT. In all experiments, ␤-galactosidase-negative cells differentiated to muscle cells to a similar extent, which provided an internal control of terminal differentiation of C2C12 myoblasts. Thus, the results indicated that OUT exerts an inhibitory effect similar to that of Id2 on the differentiation of C2C12 cells. DISCUSSION In this study we have described the molecular cloning and functional characterization of a novel mouse bHLH factor, termed OUT, that is expressed mainly in the adult reproductive organs. OUT was identified on the basis of its structural similarity to bHLH factors paraxis (53) and scleraxis (54), which are expressed in tissues of mesodermal origin. By using EMSA and E-box-mediated transactivation analyses, we demonstrated that OUT not only lacks DNA binding activity but also inhibits DNA binding of and transactivation by other bHLH factors. No obvious transcriptional repressor domain was found in the GAL4 binding assay. Physical interaction of OUT with bHLH factors was demonstrated by co-immunoprecipitation experiments. Moreover, OUT blocks the terminal differentiation of C2C12 myoblasts when exogenously introduced into the cells. These functional characteristics resemble those of the Id proteins, which are negative regulators of bHLH transcription factors (28,29). The Id proteins are HLH factors that can dimerize with bHLH factors. However, due to the lack of the basic region, these heterodimers have no DNA binding activity and inhibit the function of bHLH factors at the protein level. As demonstrated by comparison with Id2, OUT is a novel bHLH factor with an inhibitory function similar to that of the Id proteins, although OUT does have the basic region as well as the HLH region.
OUT shows a high degree of homology to bHLH factors that are expressed in tissues of mesodermal origin. Among them, capsulin (74,75), also known as epicardin (76) or Pod-1 (77), and ABF-1 (38), also known as musculin (39) or MyoR (40), are the most closely related bHLH factors, with 55.4% identity in the bHLH region at the amino acid level. As the identity between the bHLH regions of capsulin/epicardin/Pod-1 and ABF-1/musculin/MyoR is 96.5%, and these two factors form a subfamily within the bHLH factors. In this context, it is highly probable that OUT belongs to a new subfamily within the bHLH factors. Interestingly, ABF-1/musculin/MyoR has been reported to be a transcriptional repressor (38,40), whereas many of the bHLH factors induce or enhance expression of their target genes through E-box elements present in the promoter or enhancer regions of downstream genes. ABF-1/musculin/MyoR binds E-boxes as a homodimer or heterodimer with the E protein but fails to induce transactivation. Instead, it inhibits the transcriptional activation induced by other bHLH factors. Capsulin/epicardin/Pod-1 also binds DNA but is unable to induce the E-box-mediated transactivation, depending on the situation (74,75). Thus, OUT is closely related to the repressive bHLH factors not only in structure but also in function. However, the mechanism by which OUT inhibits the functions of bHLH factors is different from those of these repressive bHLH factors. As demonstrated in co-immunoprecipitation experiments, OUT is able to heterodimerize with E12 through the HLH region, but the resultant heterodimeric complexes are functionally inactive, being unable to bind DNA in EMSA. By titrating out E12 and MyoD, OUT inhibits the DNA binding of the E12-MyoD heterodimer. This feature distin- guishes OUT from these repressive bHLH factors.
Although less closely related to OUT, the HES proteins, Mist1, Twist, and Stra13, are also repressive bHLH factors and exert an inhibitory effect at least partly via a mechanism similar to Id (33,34,(41)(42)(43). These factors are functionally related to OUT but display repressive activities, also through alternative mechanisms. HES proteins bind weakly to the E-box sequence, and homodimers of HES prefer the N-box as a binding site (11,34). In the main mechanism employed by HES proteins, a WRPW motif in the C terminus recruits a co-repressor, such as Groucho or TLE, resulting in active suppression of the transcription of their downstream genes (11,(35)(36)(37). On the other hand, Mist1 and Twist repress the activities of myogenic bHLH factors by occupying specific E-box target sites and through their repressor regions, which are capable of inhibiting activators, in addition to the mechanism of titrating bHLH factors (41,42). Twist can also inhibit transactivation by MEF2 proteins, which are transcription factors containing the MADS domain, and regulate muscle-specific genes cooperatively with myogenic bHLH factors (42), whereas Mef2 is directly activated by Twist (86). Another repressive bHLH factor is Stra13, which is structurally highly related to HES (43). Although Stra13 can form dimers well with Mash1 and poorly with E proteins, it has no DNA binding activity. It possesses an ␣-helix-rich domain through which it directs repression of transcription (43). On the other hand, OUT contains no obvious repressor domain and no apparent WRPW motif, suggesting that OUT belongs to a different category from these repressive bHLH factors in terms of their inhibitory mechanisms.
The DNA binding activities of bHLH transcription factors are determined by amino acid residues that constitute the basic region. Crystallographic analyses of bHLH proteins indicate that the determinants of E-box recognition are the first glutamate and last arginine residues in the ERXR motif of Murre's consensus sequence (81,82). The glutamate residue, in particular, contacts cytosine and adenine bases (81,82). The replacement of this glutamate with other amino acid residues disturbs the DNA binding activity (87). The remaining amino acid residues in the region contribute to the DNA binding of bHLH factors by interacting with the phosphodiester backbone of DNA or by defining the specificity of interactions between the central dinucleotides of the E-box sequences and bHLH factors (81,82). OUT contains the motif ERXR in the basic region and was expected to be able to bind DNA through E-box sequences. OUT protein, however, failed to bind E-box or N-box sequences. In addition, no obviously bound DNA was recovered from the CASTing assay. What is the molecular basis for the inability of OUT to bind DNA? The one proline and relatively few basic amino acid residues in the basic region may account for this inability. Site-directed mutagenesis of the proline residue, however, indicated that its replacement with an arginine, asparagine, or glycine residue is not sufficient to restore the DNA binding activity of OUT in EMSA (data not shown). Alternatively, OUT may require an as yet unknown bHLH factor to form a functionally active heterodimer for binding to the E-box and for induction of transactivation.
The in vivo function of OUT remains to be determined at present. As OUT mRNA is barely detectable in the developing mouse embryo by Northern blot and whole mount in situ hybridization analyses, OUT appears not to be involved in organogenesis or cell differentiation during development. In the adult, however, OUT is expressed mainly in the reproductive organs, particularly in the uterus and ovary. This expression profile of OUT is distinct and contrasts with those of other bHLH factors reported so far. The other factors show embryonic expression in addition to expression in the adult organs and participate in morphogenesis and organogenesis of the developing embryo. The unique expression pattern of OUT suggests a role of OUT in relation to the reproductive organs under the regulation of sex hormones after sexual maturation, particularly in females. In support of this notion, Northern blot analyses indicate that OUT expression is maximal in early pregnancy and minimal around parturition. OUT expression recovers to non-pregnant levels 4 days after parturition. Additionally, in situ hybridization studies demonstrate that the myometrium is a predominant site of OUT expression. These results suggest that OUT is involved in the regulation or modulation of smooth muscle contraction of the uterus during pregnancy and particularly around the time of delivery. The physiological role of OUT is not clear in the ovary or other organs, including testis, mammary gland, lung, intestine, and pancreas.
The results presented here indicate that OUT has an inhibitory activity similar to those of the Id proteins, the mechanism of which distinguishes OUT from other bHLH factors reported so far. Further characterization will clarify the in vivo function of OUT and our understanding of the mechanisms underlying the functional regulation of the adult reproductive organs by bHLH factors.