Transcriptional Activation of Mouse Mast Cell Protease-7 by Activin and Transforming Growth Factor-β Is Inhibited by Microphthalmia-associated Transcription Factor*

Previous studies have revealed that activin A and transforming growth factor-β1 (TGF-β1) induced migration and morphological changes toward differentiation in bone marrow-derived cultured mast cell progenitors (BMCMCs). Here we show up-regulation of mouse mast cell protease-7 (mMCP-7), which is expressed in differentiated mast cells, by activin A and TGF-β1 in BMCMCs, and the molecular mechanism of the gene induction of mmcp-7. Smad3, a signal mediator of the activin/TGF-β pathway, transcriptionally activated mmcp-7. Microphthalmia-associated transcription factor (MITF), a tissue-specific transcription factor predominantly expressed in mast cells, melanocytes, and heart and skeletal muscle, inhibited Smad3-mediated mmcp-7 transcription. MITF associated with Smad3, and the C terminus of MITF and the MH1 and linker region of Smad3 were required for this association. Complex formation between Smad3 and MITF was neither necessary nor sufficient for the inhibition of Smad3 signaling by MITF. MITF inhibited the transcriptional activation induced by the MH2 domain of Smad3. In addition, MITF-truncated N-terminal amino acids could associate with Smad3 but did not inhibit Smad3-mediated transcription. The level of Smad3 was decreased by co-expression of MITF but not of dominant-negative MITF, which resulted from proteasomal protein degradation. The changes in the level of Smad3 protein were paralleled by those in Smad3-mediated signaling activity. These findings suggest that MITF negatively regulates Smad-dependent activin/TGF-β signaling in a tissue-specific manner.

Mast cells play a crucial role in inflammatory and immediate allergic responses. The multivalent binding of an antigen to receptor-bound IgE provides the trigger for activation of mast cells, leading to secretion of granules containing mast cell proteases (MCPs) 1 (1,2). MCPs are serine proteases, and nine mouse MCPs (mMCPs) have been described to date (3). Mouse MCP-1, -2, -4, -5, and -9 are chymases, and mMCP-6 and -7 are tryptases (3). Expression of mMCPs in mast cells varies between tissues, e.g. mmcp-7 mRNA is present in differentiated mast cells of the skin but not in mast cells of the intestinal mucosa (4), whereas mmcp-2 mRNA is expressed in mast cells located in the mucosa of the stomach (5). These observations suggest that gene expression of mMCPs is regulated by the local environment (6), but currently little information is available on regulation of mMCP expression by local growth and differentiation factors (6).
Activin and transforming growth factor-␤ (TGF-␤) are members of the TGF-␤ superfamily and are multipotent growth and differentiation factors (7,8). These peptides are not only among the most potent cellular growth inhibitors, but they also regulate other diverse biological processes including early embryonic patterning and cell fate determination. Activin and TGF-␤ bind to two different types of serine/threonine kinase receptors, termed type I and type II. Type I receptor is activated by type II receptor upon ligand binding and mediates specific intracellular signals. Smads are the central signal mediators of the TGF-␤ superfamily (9,10). Smad2 and -3 directly interact with type I receptors for activin and TGF-␤ and become activated through phosphorylation of the C-terminal SSXS motif. Smad2 and -3 then form heteromeric complexes with common partner Smad, Smad4, and translocate into the nucleus. Nuclear Smad complexes bind to transcriptional activators or co-repressors and regulate transcription of target genes (9,10).
Previous studies have shown that expression of activin A, a homodimer of inhibin/activin ␤ A , is up-regulated in immune cells, including monocytes (8,11), macrophages (12), and mast cells (13), in response to activation. Activin A induced migration of human monocytes (14) and increased gene expression and the activity of matrix metalloproteinase-2 in mouse peritoneal macrophages (12). Furthermore, activin A and TGF-␤ 1 also induced migration at lower concentrations and morphological changes and gene expression of mmcp-1 at higher concentrations in mouse mast cell progenitors (15,16).
In this report, we show that activin and TGF-␤ up-regulate the mmcp-7 gene, which is transcriptionally regulated. Smad3 but not Smad2 is responsible for the transcriptional activation. Microphthalmia-associated transcription factor (MITF), which is a tissue-specific transcription factor predominantly expressed in mast cells, melanocytes, and heart and skeletal muscle (17)(18)(19), accelerates the breakdown of Smad3 protein by the proteosomal system. As a result, the Smad3-mediated transcription is inhibited in the presence of MITF. Our results suggest that the tissue-specific and negative regulation of Smad3 activity by MITF plays a role in the discrete control of activin and TGF-␤ activities in mast cells.

EXPERIMENTAL PROCEDURES
Materials-Recombinant human activin A was kindly provided by Dr. A. F. Parlow through the National Pituitary and Hormone Distribution Program at NIDDK, National Institutes of Health. Purified TGF-␤ 1 was purchased from BD Biosciences. Stem cell factor (SCF) was purchased from PeproTech EC (London, UK). Pokeweed mitogen was purchased from Seikagaku Kogyo (Tokyo, Japan). Mouse monoclonal anti-FLAG antibody (M2), anti-HA antibody (12CA5), and anti-Myc antibody (9E10) were obtained from Sigma, Roche Applied Science, and Santa Cruz Biotechnology, respectively. Lactacystin and MG-132 were obtained from Calbiochem.
Cell Culture and Transfection-Bone marrow-derived cultured mast cell progenitors (BMCMCs) were cultured from the bone marrow cells of BALB/c mice as described previously (16). More than 95% of the trypan blue-excluding viable cells were mast cells on the basis of staining with acid toluidine. L17 cells, a derivative of the mink lung epithelial cell line (Mv1Lu) (35), obtained from Dr. J. Massagué, were cultured as described previously (21,36). HepG2, COS7, and HEK293 cells were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum. NIH3T3 cells were cultured in Dulbecco's modified Eagle's medium with 10% calf serum. For transient transfection, cells in 24-or 6-well plates were transfected using PolyFect transfection reagent (Qiagen).
RNA Isolation, Competitive RT-PCR-Total RNA from BMCMCs was isolated using an RNA isolation kit (RNeasy, Qiagen) according to the manufacturer's protocol. Recovered RNA was reverse-transcribed as described previously (12). To examine the mRNA levels semi-quantitatively, competitive RT-PCR was performed as described previously (16). For the mmcp-7 gene transcript (GenBank TM accession number NM031187), oligonucleotides encoding positions from 177 to 198 and 812 to 789 were used as the PCR primers. A competitor for mMCP-7 was made by a deletional mutation of the mMCP-7 PCR product, which deleted a cDNA segment between positions 199 and 493. The competitor was made by overlap extension PCR of the native PCR product, followed by purification using a Suprec-02 column (Takara, Tokyo, Japan). The PCR primers and competitor for G3PDH were as described previously (12). A constant amount (2 ϫ 10 Ϫ5 amol for mMCP-7 and 5 ϫ 10 Ϫ4 amol for G3PDH) of the competitor template was co-amplified with the specific primers with reverse-transcribed samples or varying amounts of the target cDNA standard. The PCR products were separated on 2% agarose gels in 1ϫ TAE buffer and visualized with ethidium bromide.
Real Time Quantitative RT-PCR-Total RNA from BMCMCs was reverse-transcribed as described previously (12). PCR was performed in a total volume of 25 l with SYBR Green PCR Master Mix (Applied Biosystems), 2 l of cDNA, and 200 nM of each primer. PCR primers were designed with the computer program Primer Express (Applied Biosystems), using parameters recommended by the manufacturer. For the mmcp-7 and g3pdh gene transcripts (GenBank TM accession numbers NM031187 and M32599, respectively), oligonucleotides encoding positions from 361 to 379 and 425 to 404 for mMCP-7 and those from 740 to 759 and 812 to 789 for G3PDH were used as the PCR primers. Reactions were carried out in an ABI-prism 7700 sequence detector (Applied Biosystems), using the following conditions: an initial denaturation step consisted of 2 min at 50°C and 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and 1 min at 60°C. Levels of mMCP-7 and G3PDH expression in each sample were determined by using the relative standard curve method. A relative amount of DNA of mMCP-7 was expressed as a ratio to G3PDH DNA, and the mMCP-7 level in BMC-MCs without TGF-␤ 1 , activin A, and SCF was 1.
Reporter Assay-Luciferase assays were conducted as described previously (36,37). Cells were transiently transfected with the indicated expression vectors, reporter construct, or a plasmid expressing ␤-galactosidase (pCMV-␤Gal). Equal amounts of DNA were transfected in each experiment and adjusted with pcDNA1 and pcDNA3, and cells were harvested 40 h after transfection. Luciferase activity was normalized to ␤-galactosidase activity, and the luciferase activity in the cell lysate transfected with empty vector was set at 1.
Sequential Immunoprecipitation and Immunoblot-COS7 cells were transfected with HA-tagged Smad and Myc-tagged MITF. Twenty seven hours after transfection, cells were treated with or without 10 M lactacystin or 50 M MG-132 for 16 h. The cells were lysed in lysis buffer (20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1% Nonidet P-40, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 1% aprotinin). After 30 min on ice, cell debris was removed by centrifugation at 600 ϫ g for 5 min at 4°C, and the supernatant was immunoprecipitated overnight at 4°C with anti-HA antibody (12CA5) and incubated with protein G-agarose beads for 1 h at 4°C. The beads were washed four times with lysis buffer, followed by elution in SDS-PAGE sample buffer. The immunoprecipitates were subjected to Western blotting with anti-HA antibody (12CA5) or anti-Myc antibody (9E10) as primary antibody, and the bands were visualized with ECL Plus reagent (Invitrogen).
Mammalian Two-hybrid Assay-Mammalian two-hybrid assays were performed using the checkmate mammalian two-hybrid system (Promega), according to the manufacturer's protocol. In brief, HEK293 cells were co-transfected with the plasmid of interest, a plasmid expressing ␤-galactosidase (pCMV-␤Gal). Equal amounts of DNA were transfected in each experiment and adjusted with pBIND or pACT. Twenty seven hours after transfection, cells were treated with or without 10 M lactacystin for 16 h. Luciferase activity was normalized to ␤-galactosidase activity, and the luciferase activity in the cell lysate transfected with empty vector was set at 1.

Mouse mcp-7 Gene Is Transcriptionally Activated by
Smad3-To investigate the effect of activin and TGF-␤ on BMCMCs, gene transcripts of mmcp-1, -7, and -9 were first examined by RT-PCR, which was conducted under nonsaturating conditions for the PCR products. The band intensities of mMCP-1 and mMCP-7 were reproducibly increased by treatment with activin A or TGF-␤ 1 (data not shown). Activin A-and TGF-␤ 1 -induced gene expression of mMCP-1 has been described elsewhere (16). To characterize mmcp-7 gene induction in response to activin A and TGF-␤ 1 in detail, we semi-quantitatively measured levels of mmcp-7 gene transcript by competitive RT-PCR. Treatment of BMCMCs with activin A or TGF-␤ 1 for 24 h clearly increased the gene transcript level of mmcp-7 (Fig. 1A). Consistent with a previous study (38), treatment with SCF also increased the mRNA level of mmcp-7. Co-treatment with SCF and activin A or TGF-␤ 1 resulted in further accumulation of mmcp-7 mRNA (Fig. 1A). Real time quantitative RT-PCR analyses also revealed the stimulatory effect of activin A and TGF-␤ 1 on mmcp-7 gene transcript (Fig. 1B).
To examine whether the mmcp-7 gene induction by activin A and TGF-␤ 1 is transcriptionally regulated, we performed transcriptional activation assays using the 5Ј-flanking region of mmcp-7 fused to the luciferase reporter gene (25). As shown in Fig. 2A, transfection of type II (T␤RII) and type I (ALK5) TGF-␤ receptor into HepG2 cells did not affect luciferase expression, whereas transfection of a constitutively active type I TGF-␤ receptor (ALK5(TD)) (31) increased luciferase expression 5.6-fold. Consistent with expression of a series of TGF-␤ receptor isoforms, expression of a constitutively active type I activin receptor (ALK4(TE)) (21) but not of a type II (ActRII) or type I (ALK4) activin receptor increased luciferase expression 6.3-fold. Transcriptional assays in NIH3T3 cells and L17 cells, a derivative of mink lung epithelial cells (35), also showed increased luciferase expression only in cells expressing constitutively active type I receptors (ALK5(TD) and ALK4(TE)) ( Fig.  2, B and C). These results suggest that TGF-␤-and activininduced mmcp-7 mRNA is transcriptionally regulated and that the receptor activation is responsible for transcriptional activation of mmcp-7.
Expression of WT-MITF also decreased Smad3-induced transcription of mmcp-7 in a dose-dependent manner (Fig. 4A). In contrast, overexpression of mi-MITF, a dominant-negative mutant that deletes one of four consecutive Arg in the basic domain (6), had a weak effect on inhibition of the induced luciferase expression (Fig. 4A). We also examined the effect of MITF expression on transcriptional activation of AR3-luc, which contains regulatory sequences from the Xenopus mix2 gene (27) and is activated by co-expression of the transcription factors FoxH3 and Smad2 (36,37). Similar to the inhibitory effect on mmcp-7 transcription, WT-MITF inhibited FoxH3-dependent and Smad2-mediated AR3 transcription, whereas mi-MITF hardly affected transcriptional activity (Fig. 4B). These results suggest that functional MITF acts as a negative regulator for Smad2 and -3 signaling. Transcriptional inhibition of WT-MITF was not a nonspecific event, since overexpression of WT-MITF but not mi-MITF increased luciferase expression when the 5Ј-flanking region of mmcp-6 fused to the luciferase gene was used as a reporter gene (Fig. 4C). In addition, when the 5Ј-flanking region of mmcp-9 fused to the luciferase gene was used, overexpression of ALK5(TD) and Smad3 neither increased luciferase expression nor affected WT-MITF-induced transcription of mMCP-9-luc (Fig. 4D). MITF has several isoforms. The mouse MITF gene contains at least five isoformspecific first exons, and exons 2-9 of all isoforms of MITF are the same (18,40,41). Two isoforms of MITF, MITF-mc (19) and MITF-E (18), inhibited Smad3-mediated mMCP-7 transcription and FoxH3-dependent and Smad2-mediated AR3-luc transcription (data not shown).
MITF Binds to and Induces Breakdown of Smad Proteins-To explore the molecular basis of the inhibitory effect of MITF on Smad2 and -3 signaling, we examined the expression of Smads in response to co-expression of MITF in COS7 cells. Protein expression levels of Smad2, -3, and -4 were significantly decreased by co-expression of WT-MITF (Fig. 5, A-C). In contrast, mi-MITF hardly affected the protein level of Smad2, -3, or -4. The WT-MITF-induced decreases in Smad protein level were independent of TGF-␤ signaling; expression of ALK5(TD) had no effect on decreases in Smad2 and -3 when co-expressed with WT-MITF. The decreased protein levels of Smad associated with co-expression of WT-MITF were not nonspecific events, because the protein level of MEK1, a mitogenactivated protein kinase kinase, was relatively unchanged in response to WT-MITF expression (Fig. 5D).
Next, we examined whether the decreased protein levels of Smad2 and -3 due to co-expression of WT-MITF resulted from proteasomal breakdown. Treatment with lactacystin (10 M), which inhibits activity of 20 S proteasome (42), effectively blocked the decreases in Smad2 and -3 proteins caused by co-expression of WT-MITF, resulting in constant protein levels of Smad2 and -3 irrespective of co-expression of WT-MITF (Fig.  6A). In addition, decreases in protein expression of Smad2 and -3 by co-expression of WT-MITF were also blocked by treatment with another proteasome inhibitor, MG-132 (50 M) (Fig. 6B). These results suggest that expression of WT-MITF but not mi-MITF induces proteasome-dependent degradation of Smad2 and -3.
Smad signaling is modulated by association with transcriptional regulators (9, 10). We examined whether Smads associate with MITF using mammalian two-hybrid assays in HEK293 cells and by co-immunoprecipitation assays in COS7 cells. As shown in Fig. 7A, no significant interactions between Smad and MITF were detected in the absence of lactacystin. In contrast, in the presence of lactacystin, both Smad2 and -3 interacted with WT-and mi-MITF (Fig. 7A). When lysates of COS7 cells expressing HA-tagged Smad2 or -3 and Myc-tagged MITF in the presence of lactacystin were incubated with anti-HA antibody, the immunoprecipitates contained Myc-tagged WT-and mi-MITF, which was consistent with the results of mammalian two-hybrid assays (Fig. 7B). In addition, association between mi-MITF and Smad2 and -3 was detected in the absence of lactacystin (Fig. 7B).
Next, we examined the region of Smad2 and -3 responsible for functional and physical interaction with MITF (Fig. 8). Transfection of the MH2 domain but not the MH1 and linker region of Smad3 increased luciferase expression, when mMCP-7-luc was used as a reporter construct (Fig. 8B, right panel). Also, expression of the MH2 domain but not the MH1 and linker region of Smad2 induced transcription of FoxH3-dependent AR3-luc (Fig. 8B, left panel). Co-expression of WT-MITF effectively inhibited the transcription induced by the MH2 domain of Smad2 and -3, suggesting that at least the MH2 region of Smad2 and -3 functionally interacted with MITF (Fig.  8B). When expression of Smad proteins was examined, not only the protein level of the MH2 domain but also that of the MH1 and linker region of Smad2 and -3 was decreased by expression of WT-MITF but not by that of mi-MITF (Fig. 8C). When association between Smad3 and MITF was examined by coimmunoprecipitation assays, complex formation of the MH1 and linker region of Smad3 and WT-MITF was detected in a lactacystin-dependent manner, whereas no interaction was observed between the MH2 domain of Smad3 and MITF. In addition, the MH1 and linker region of Smad3 formed a complex with mi-MITF even in the absence of lactacystin (Fig. 8D).
The region of MITF necessary for functional and physical interaction with Smad3 was also examined (Fig. 9). Truncation of the C-terminal 80 amino acids of WT-MITF did not affect the inhibitory effect of MITF on Smad3-mediated mMCP-7 transcription (Fig. 9B). However, the truncation of a further 43 amino acids resulted in the loss of the inhibitory effects. In addition, truncation of the N-terminal 127 amino acids also destroyed its ability to inhibit Smad3 signaling. Similar effects of deletion mutants of MITF were also detected on FoxH3-dependent and Smad2-mediated AR3-luc transcriptional activation (data not shown). These results suggest that at least two regions of MITF are crucial for functional interaction with Smad3 signaling. The protein level of Smad3 was decreased by co-expression of full-length MITF but not by partial MITFs that did not inhibit Smad3-mediated transcription (Fig. 9, B and C), and similar regulation of Smad3 protein level by MITF was observed in the MH1 and linker region of Smad3 and in the MH2 domain of Smad3 (Fig. 9C). Co-immunoprecipitation assays revealed that MITF-(128 -419) stably associated with Smad3, irrespective of lactacystin treatment, whereas no significant interaction between MITF-(1-298) and Smad3 was observed (Fig. 9D). DISCUSSION In the present study, we show that activin A and TGF-␤ 1 increase gene expression of mmcp-7 in mast cell progenitors, which is transcriptionally mediated by Smad3. The Smad3mediated transcriptional activation is negatively regulated by MITF, a tissue-specific transcription factor of the bHLHLZ family. Breakdown of Smad3 protein, but not complex formation of Smad3 and MITF, is essential for inhibition of Smad3mediated signaling by MITF. Recently, we have shown increased expression of activin A in mast cells in response to activation (13) and modulation of migration and morphological changes toward differentiation by activin A and TGF-␤ 1 in mast cell progenitors (16), suggesting that activin/TGF-␤ positively regulate the functions of mast cells as effector cells in the immune system. The present data demonstrate another function of these potent growth/differentiation factors. In addition, functional interactions between Smad and MITF suggest a lineage-specific regulation of Smad signaling.
Several pathways are known to regulate Smad signaling (9, FIG. 7. Association of Smad2/3 and MITF. A, interaction between Smad2/3 and MITF was examined by mammalian two-hybrid analysis. After transfection, HEK293 cells were cultured in the presence or absence of lactacystin. Luciferase activity was normalized to ␤-galactosidase activity, and luciferase activity in the cell lysates in the absence of MITF and Smad and of lactacystin was set to 1. A representative result from three independent experiments is shown. Data are expressed as the mean Ϯ S.D. (n ϭ 3). B, the interaction of HA-tagged Smad2/3 with Myc-tagged MITF was examined by immunoprecipitation (IP) followed by Western blotting (WB) in COS7 cells. 10). Modulation of turnover of Smad protein is one method of regulation of Smad signaling in the cytoplasm as well as in the nucleus (10,43). Ubiquitin ligases forming complexes with Smad2 and -3 were involved in ubiquitination and proteasomal degradation of Smad2 and -3 (44 -46). Expression of Jab1, identified as the Jun-activating domain binding protein, accelerated breakdown of Smad4 (47). Phosphorylation of Smad2 by ERK1 stabilized Smad2 protein and increased transcriptional activity of Smad2 (37). In contrast, calmodulin blocked Smad2 phosphorylation by ERK1 and decreased Smad2 protein level (37). In the present study, the inhibition of Smad3-mediated transcriptional activation of mmcp-7 was paralleled by the protein level of Smad3, and the decrease in Smad3 protein by expression of MITF was blocked by inhibition of proteasomal activity. These results suggest that MITF regulates Smad3 signaling by controlling the turnover of Smad3 protein.
There are at least two regions of MITF responsible for inhibition of Smad3-mediated signaling, amino acids 1-127 and 341-419 (Fig. 9A). These regions are located neither in the basic helix-loop-helix domain nor in the leucine zipper domain, and currently these are not regions with a known motif. The involvement of multiple regions of MITF suggests that multiple regulatory pathways are involved in the inhibition of Smad3mediated transcription. Alternatively, it is possible that truncation of MITF causes conformational changes leading to loss of the inhibitory activity on Smad3-mediated transcription.
MITF associates with the MH1 and linker region of Smad3. In addition, amino acids 298 -419 of MITF, which contains neither a basic helix-loop-helix domain nor a leucine zipper domain (Fig. 9A), was suggested to be the region responsible for Smad binding. The MH1 domain of Smad3 is responsible for physical interaction with TFE3, one of the members of the bHLHLZ family (48,49), and the leucine zipper domain of TFE3 interacts with the MH1 and linker region of Smad3 (49). These results suggest that the structural organization of the complex of Smad3 and MITF is distinct from that of Smad3 and TFE3. Both Smad3 and MITF can associate with a transcriptional co-activator, CBP/p300 (50 -52). However, it is unlikely that complex formation of Smad3 and MITF is bridged by CBP/p300, because the MH2 domain of Smad3 and amino acids 123-127 of MITF are the regions responsible for CBP/p300 binding (50 -52). Both regions were dispensable for complex formation of Smad and MITF.
Although the MH2 domain of Smad3 could not form a complex with MITF, the protein level of the MH2 domain of Smad3 was decreased, and the MH2 domain-induced mMCP-7 transcription was inhibited by co-expression of MITF. In contrast, N-terminal amino acid-truncated MITF formed a complex with Smad3, but neither decreased Smad3 protein level nor inhibited Smad3-mediated transcriptional activation. These results suggest that complex formation of Smad and MITF was neither necessary nor sufficient for breakdown of Smad proteins. In fact, mi-MITF formed a complex with Smad3 but did not decrease the Smad3 protein level. Smad3-mediated transcriptional activation paralleled the protein level, suggesting that Smad regulation by MITF is achieved through modulation of the Smad protein level. MITF could form a complex with LEF-1, a signal mediator of the Wnt pathway, but the association was not sufficient for efficient transactivation of genes activated by the Wnt pathway (53). In addition, both Max, another member of the bHLHLZ family, and TFE3 bound to the MH1 region of Smad3 in vitro, but Max and TFE3 had opposite effects on Smad3-mediated transcriptional activation (49,54).
It is possible that complex formation of Smad with a bHLHLZ transcription factor does not necessarily correlate with transcriptional regulation, and that the bHLHLZ transcription factor regulates Smad activity through distinct mechanisms.
Yeast two-hybrid analyses revealed that MITF associated with Ubc9, a ubiquitin-conjugating enzyme (55,56). Multiubiquitination of protein for proteasomal degradation is achieved through an enzymatic cascade, ubiquitin-activating enzyme, ubiquitin-conjugating enzyme, and ubiquitin-protein ligase. Thus, we tested whether Ubc9 expression affects the amount of Smad protein. Neither Smad protein level nor Smad-mediated transcriptional activation was affected by overexpression of Ubc9 (data not shown). In addition, the effect of Ubc9 expression was not detected on MITF-induced breakdown of Smad protein and blockage of Smad signaling (data not shown). Molecules other than Ubc9 would be responsible for MITF-induced degradation of Smads. mMCP-7 is expressed predominantly in a subpopulation of differentiated mast cells that reside in numerous connective tissue sites (4,57,58). mMCP-7 is a tryptase that degrades fibrinogen (59). When mast cells are activated after IgE/antigen stimulation, preformed granule mediators including mMCP-7 are released, leading to vasodilation, edema, and an influx of hemopoietic cells into the inflammatory site (1). However, in general, significant amounts of cross-linked fibrin and aggregated platelets are not detected at the site of inflammation (60), which is suggested to result from inactivation of fibrinogen by mMCP-7 at the endothelium/blood barrier (59). Taken together, the role of mMCP-7 and the stimulatory effects of activin A and TGF-␤ 1 on migration and differentiation of mast cell progenitors (16) suggest that these potent growth/ differentiation factors act as positive regulators of the effector cells in the immune system.
Previous studies (4,38,58,61) revealed that patterns of gene expression of mmcp-7 partly overlapped with, but were distinct from, those of the other mmcps. DNA microarray analyses revealed that SCF selectively up-regulated gene expression of mmcp-7 and -4 in BMCMCs from C57BL/6 mice but not in those from Rac2-null mice (38). During the culture of BMCMCs in interleukin-3-rich medium, which was WEHI-3 myelomonocytic cell-conditioned medium (62), mmcp-7 was transiently expressed, in contrast to the expression of mmcp-5 and -6 (4). Cultured mast cells from MITF-null mice did not express mmcp-4, -5, and -6 (61), whereas they did express mmcp-7 (25). Furthermore, mmcps were expressed differently in mast cells during infection with Strongyloides venezuelensis (5) and Trichinella spiralis (63). Therefore, the gene regulation of mmcp-7 revealed in this study may be unique and not applicable to that of the other mmcps.
The present study revealed the inhibitory effect of MITF on Smad3-mediated mmcp-7 gene transcription, which was achieved by breakdown of Smad3. By a similar mechanism, MITF also negatively regulated FoxH3-dependent and Smad2mediated signaling. Although precise mechanism by which MITF increases breakdown of Smad protein is not yet clear, our results suggest that MITF acts as a negative regulator of activin/TGF-␤ signaling. MITF is a transcription factor predominantly expressed in mast cells, melanocytes, and heart and skeletal muscle, which controls transcription of lineage-specific genes. Therefore, gene products regulated by MITF may be involved in Smad protein degradation; the results that dominant-negative MITF, mi-MITF, did not inhibit Smad-mediated transcriptional activation support this possibility. In addition to the direct regulation of gene transcription as a transcription factor, MITF may also maintain tissue-specific functions through proteasome-dependent degradation of widely expressed molecules. In turn, numerous cross-talks between the Smad pathway and the pathways mediated by ubiquitously expressed molecules and by tissue-specific molecules may explain the diverse functions of activin and TGF-␤.