Determinants of Thyrotrope-specific Thyrotropin Promoter Activation

Thyrotropin (TSH) β is a subunit of TSH, the expression of which is limited to the thyrotrope cells of the anterior pituitary gland. We have utilized the thyrotrope-derived TtT-97 thyrotropic tumors to investigate tissue-specific expression of the TSHβ promoter. TSHβ promoter activity in thyrotropes is conferred by sequences between −270 and −80 of the 5′-flanking region. We have recently reported that the proximal region from −133 to −100 (P1) is required for promoter expression in thyrotropes. This region interacts with the pituitary-specific transcription factor Pit-1. While Pit-1 appears necessary for TSHβ promoter activity in thyrotropes, this transcription factor is not alone sufficient for promoter activity in pituitary-derived cells. In this report, we have generated a series of promoter mutations in the P1 region to identify additional protein-DNA interactions and determine their functional significance. We have found that Pit-1 interacts with the distal portion of the P1 region, and a second protein interacts with the proximal segment of this region. Each protein is able to independently interact with the TSHβ promoter, but neither alone can maintain promoter activity. Both proteins appear to be necessary for full promoter activity in thyrotropes. Southwestern analysis with the proximal segment of the P1 region (−117 to −88) reveals interaction with a 50-kDa protein. Interestingly, this protein is not found in the pituitary-derived GH3 cells and may represent a thyrotrope-specific transcription factor. Further characterization of this newly identified DNA-binding protein will further our understanding of the tissue-specific expression of the TSHβ gene.

nologically unique and has expression limited to thyrotropes. Gene expression is affected by trans-acting factors that bind directly to cis-acting elements within promoter regions of specific genes (2). Transfection experiments in thyrotrope cells have shown that the cell-specific activity of the mouse TSH␤ promoter is localized between Ϫ270 and Ϫ80 of the 5Ј-flanking region (3,4). DNase I protection studies using nuclear protein extracts from a TSH␤-expressing mouse thyrotropic tumor, TtT-97, have identified four cis-acting elements in the Ϫ270 to Ϫ80 region: D1 (Ϫ253 to Ϫ222), D2 (Ϫ196 to Ϫ176), P1 (Ϫ133 to Ϫ100) and P2 (Ϫ86 to Ϫ64) (4,5). We have been studying thyrotrope-specific expression of the TSH␤ promoter and factors involved in this expression.
Three proteins (Pit-1/GHF-1, thyrotrope embryonic factor (TEF), and mLIM-3) have been shown to interact with the TSH␤ promoter in the region from Ϫ270 to Ϫ80 (6 -11). Pit-1 is a well characterized POU-homeodomain, pituitary-specific transcription factor (12)(13)(14). Pit-1 protein expression is limited to thyrotropes, lactotropes, and somatotropes in the anterior pituitary gland (14). Pit-1 has been shown to be required for efficient transcription of the growth hormone and prolactin genes (15,16) as well as autoregulation of its own expression (17,18). Pit-1 has also been shown to mediate thyrotropinreleasing hormone (TRH) and cAMP response of TSH␤ promoter activity (7,8,19). Pit-1 interacts with three of the four cis-acting elements in the Ϫ270 to Ϫ80 region of the TSH␤ promoter, D1, P1, and P2 (5). Studies with the rat and human promoters suggested that the cognate D1 and P1 regions confer Pit-1-mediated stimulation with TRH and cAMP (8,19). We have recently shown that the P1 region is necessary for basal activity of the TSH␤ promoter in thyrotropes (20). A mutation in this region, which disrupted Pit-1 binding, decreased basal promoter activity to a level seen with a promoter deletion to Ϫ80. While Pit-1 is necessary for basal activity of the TSH␤ promoter in thyrotropes, it does not appear to be sufficient to stimulate promoter activity when transfected into cells that lack Pit-1 (5,8,9,19,21). A thyrotrope-specific splice variant of Pit-1, Pit-1T (10,20), stimulates the TSH␤ promoter through the P1 region, but its role in cell-specific promoter activity is unclear.
TEF is a member of the leucine zipper (bZIP) gene family of transcription factors (11). TEF mRNA appears in the rat anterior pituitary gland on embryonic day 14, before TSH␤ mRNA. TEF interacts with a region of the mouse TSH␤ promoter which is similar to the D2 thyrotrope extract protected region of the mouse promoter (4). TEF stimulates TSH␤ promoter activity in CV-1 cells, but TEF mRNA expression in adult tissues is not restricted to thyrotropes, raising questions as to its role as a thyrotrope-specific transcription factor. Recently, a study of mouse pituitary development has shown that TEF may be responsible for the initiation of the thyrotrope phenotype, but Pit-1 and perhaps other factors are necessary for the persistence of mature thyrotropes (22). mLIM-3 is a member of the LIM homeodomain family of transcription factors and appears to be pituitary-specific (23,24). Bach and colleagues (23) have shown that mLIM-3 mRNA is present in developing and adult mouse pituitaries as well as a number of pituitary-derived cell lines, including the thyrotrope-derived aTSH cells. They also showed that in combination with Pit-1, mLIM-3 stimulated mTSH␤ promoter activity in CV-1 cells and interacted with the Ϫ120 to Ϫ60 region of the promoter in gel retardation assays.
The P1 region of the mouse TSH␤ promoter is necessary for cell-specific basal promoter activity and Pit-1, which interacts with this region, does not appear to be sufficient for full promoter activity observed in thyrotrope cells. We recently observed that the DNase I protected footprints in this region differed at the 3Ј end between recombinant Pit-1 and the TtT-97 thyrotrope extract (20), which suggests that other thyrotrope proteins may be interacting with the mTSH␤ promoter. Steinfelder et al. (7) demonstrated that a human TSH␤ promoter fragment (Ϫ128 to Ϫ61) yeilded five distinct protein-DNA complexes with a thyrotrope extract by gel mobility shift analysis. None of these complexes were seen using non-pituitary HeLa cell extract. Furthermore, they proved that four of these complexes contained Pit-1 and one complex appeared to be due to an unrelated thyrotrope protein. These data suggest that more than one protein may be interacting with the TSH␤ promoter in this functionally critical region. In this report, we have undertaken a detailed mutagenesis analysis of the P1 region and identified a second protein that interacts with the proximal portion of this region and appears to be necessary for basal activity of the mTSH␤ promoter in thyrotropes.

Nuclear Extract Preparation and DNase I Protection Analysis-Nu-
clear extracts were prepared from TtT-97 thyrotropic tumors, ␣TSH and GH3 cells as described previously (25). The Ϫ392 to ϩ40 mTSH␤ promoter fragment was excised from pGEM7zfϩ with EcoRI and Mlu1 and gel-purified. The fragment was selectively labeled at the EcoRI end (ϩ40 position) with reverse transcriptase using [␣-32 P]dATP and [␣-32 P]dTTP. DNase I protection assays were carried out as described previously (20). 20 g of bovine serum albumin, 120 g of recombinant glutathione S-transferase/rPit-1 fusion protein bacterial extract (10), and 75 g of TtT-97 or 60 g of GH3 nuclear protein extract were used for the various assays.
In Vitro Mutagenesis of the TSH␤ Promoter-Site-directed mutagenesis of the P1 region extending from Ϫ133 to Ϫ86 was carried out in the context of the Ϫ392 to ϩ40 mTSH␤ promoter region in pSELECT (Promega, Madison, WI) as described previously (20). Briefly, for each mutated segment a 30-base pair oligonucleotide was generated with 4 -6 altered bases in the center. Mutagenesis was carried out with the Altered Sites System (Promega). The Ϫ392 to ϩ40 fragment was then excised with BamHI and HindIII and ligated into pGEM7zfϩ and pA3luc (3) for DNase I protection analysis and gene transfer studies, respectively. Sequences were verified in each plasmid by the chain termination method of Sanger (26). Wild-type mTSH␤ promoter sequence and respective mutations are noted in Fig. 1.
Gene Transfer Studies-Transient transfection assays in TtT-97 thy-rotropic tumor cells have been outlined previously (3). Briefly, 20 g of the various TSH␤ promoter-luciferase plasmids and 1 g of pCMV␤gal as an internal control for transfection efficiency were co-transfected by electroporation into 5-10 million TtT-97 cells. Cells were incubated in 4 ml of Dulbecco's modified Eagle's medium supplemented with charcoalstripped 10% fetal calf serum (Life Technologies, Inc.) at 37°C for 18 h. Cells were harvested, subjected to freeze-thaw extraction, and assayed for luciferase and ␤-galactosidase activity as described previously (10). Luciferase activity was corrected for ␤-galactosidase activity as an internal control for efficiency of each transfection. Statistical analysis was performed by one way analysis of variance. Pairwise multiple comparisons were made by the Student-Newman-Keuls test (27). Southwestern Blot Analysis-Southwestern blot analysis was carried out as described previously (5). Briefly, nuclear extract proteins (approximately 50 g) were separated on a denaturing 10% polyacrylamide gel. After electrotransfer, nitrocellulose filters were treated with guanidine hydrochloride to renature the proteins, then treated with a solution containing 5% non-fat dry milk and 1% bovine serum albumin to block nonspecific interactions with the labeled probe. Binding reactions were carried out with a [␣-32 P]dCTP-labeled probe (3 ϫ 10 6 cpm/ ml) corresponding to the wild-type or M7 mutation equivalent Ϫ117 to Ϫ88 region of the TSH␤ promoter in 10 mM Tris-HCl, pH 7.5, 0.5% non-fat milk, 0.5% bovine serum albumin, 50 mM NaCl, 5 mM MgCl 2 , 1 mM EDTA, 5 mM dithiothreitol, 0.1% Triton X-100, 5% glycerol, 200 M ZnSO 4 , and 10 g/ml each of native and denatured salmon testes DNA. Binding reactions were carried out overnight at 4°C. Filters were washed twice for 20 min at 4°C in binding buffer minus salmon testes DNA and autoradiographic exposures were carried out at Ϫ70°C overnight.

DNase I Protection Footprinting of the TSH␤ Promoter and
Mutations-We first compared protein-DNA interaction of recombinant Pit-1 protein, TtT-97, and GH3 nuclear extracts with an expanded region of the TSH␤ promoter from Ϫ390 to ϩ40 ( Fig. 2A). The 5Ј (distal) extent of the P1 footprint appears to be identical between Pit-1 and the nuclear extracts. The 3Ј (proximal) end, however, shows a clear extension of approximately 14 base pairs with the TtT-97 nuclear extract (lane 4) which is not seen with the recombinant Pit-1 protein (lane 3). This extension is similar in two nuclear extract preparations from two different TtT-97 thyrotrope tumors and redefines the P1 region to be Ϫ133 to Ϫ86. Interestingly, this footprint extension is not seen using an extract from the pituitary-derived, non-thyrotrope GH3 cells (lane 5) which suggests that factor(s) involved in this extension are thyrotrope-specific. The other regions of the TSH␤ promoter are also redefined as D1 (Ϫ295 to Ϫ222), D2 (Ϫ196 to Ϫ170), and P2 (Ϫ80 to Ϫ62) using the TtT-97 nuclear protein extract. A new region D0 (Ϫ330 to Ϫ300) is also noted, which is protected by TtT-97 extract but not recombinant Pit-1 protein or GH3 nuclear extract. Fig. 2B shows DNase I protection with recombinant Pit-1 and TtT-97 extract protein on the eight different P1 region mutations. Mutations P1M1, P1M2, P1M6, and P1M8 appeared to have no effect on either Pit-1 or TtT-97 protein binding. In contrast, mutations P1M3 and P1M4 disrupted recombinant Pit-1 binding as well as the distal portion of the footprint generated with TtT-97 extract, but the proximal P1protected region (PP-1) was preserved with the TtT-97 extract only, suggesting that a protein other than Pit-1 in the thyrotrope nuclear extract interacts with the proximal portion of the P1 region. The P1M5 mutation abrogated binding of Pit-1 and TtT-97 extract. This mutation corresponds to a previously reported random mutation (20) which also disrupted Pit-1 and TtT-97 extract binding. The P1M7 mutation disrupted binding of TtT-97 extract to the proximal P1 region (PP-1), but this mutation had no effect on Pit-1 binding or the binding of Pit-1 in the TtT-97 extract. These data suggest that two separate thyrotrope proteins interact with the P1 region of the mTSH␤ promoter and that each protein interacts independently of the other.

Activity of the TSH␤ Promoter and Mutations in TtT-97
Thyrotropes-In order to evaluate the functional effect of the P1 region TSH␤ promoter mutations in thyrotropes, we transiently transfected luciferase reporter plasmids containing the wild-type and mutated promoters within the context of the Ϫ390 to ϩ40 region into TtT-97 cells. Results of the various mutant promoter activities compared with wild-type are shown in Fig. 3. The activity of a Ϫ80 to ϩ40 TSH␤ promoter plasmid, which lacks the P1 region is shown for comparison. All of the mutations affect activity of the TSH␤ promoter in TtT-97 cells, which may reflect a greater sensitivity of these mutations in a functional assay compared with the protein-DNA interaction data displayed in Fig. 2B. However, the relative promoter activity of the mutants appears to mirror the ability of proteins to interact with this region. The P1M1 and P1M2 mutations, which do not affect protein interaction, show modest decreases of promoter activity (24 and 29% decrease,respectively) compared with the wild-type promoter. The P1M3 and P1M4 mutations, which disrupt binding of Pit-1 but not the protein to the proximal region (PP-1), greatly reduce promoter activity 70 and 72%, respectively. The P1M5 mutation interferes with interaction of both proteins and reduces activity 63%. Interestingly, the P1M6 mutation, which is well within the P1 region, does not affect protein-DNA interaction. Promoter function is modestly affected (38% decrease) which is significantly different than the M3, M4, and M5 mutations (p Ͻ 0.05). The P1M7 mutation is the only mutation that interferes with interaction in the PP-1 region and not Pit-1, and its promoter activity is profoundly reduced by 78% compared with the wild-type promoter. This reduction is equivalent to that seen with a deletion of the entire P1 region (Ϫ80 deletion, Fig. 2B). Finally, the P1M8 mutation does not appear to affect protein-DNA interaction. Functionally, however, promoter activity is moderately reduced by 43%. This data may reflect the differences in sensitivity between the functional and protein-DNA interaction assays. Yet this activity is significantly different than the re- ductions seen by the mutations which interfere with protein-DNA interactions (M3, M4, M5, M7) (p Ͻ 0.05). These data suggest that protein-DNA binding and promoter function correlate in this system and interaction of each protein with the P1 region is necessary for activity of the TSH␤ promoter and that Pit-1 or the other thyrotrope protein(s) alone is not sufficient to mediate this basal activity in thyrotropes.
Southwestern Analysis of Thyrotrope Nuclear Extract-To determine whether other proteins in thyrotrope nuclear extracts directly interact with the PP-1 region of the TSH␤ promoter, we performed Southwestern blot analysis of thyrotropederived TtT-97 nuclear extracts, as well as somatotropederived GH3 nuclear extracts which did not protect the proximal extension of the P1 region and thyrotrope-derived ␣TSH nuclear extracts which lack Pit-1 protein (5). A duplexed oligonucleotide lacking the Pit-1 binding region (Ϫ117 to Ϫ88) interacted with a 50-kDa protein in the thyrotrope-derived TtT-97 and aTSH nuclear extracts, but not the somatotropederived GH3 nuclear extract (Fig. 4, lanes 1-3). As expected, recombinant Pit-1 protein (31-33 kDa) did not interact with this oligonucleotide (data not shown). A duplexed oligoucleotide (Ϫ117 to Ϫ88) with changes corresponding to the M7 mutation ( Fig. 1) did not interact with the 50-kDa protein (Fig. 5, lanes  4 -6), suggesting that this protein is involved in TSH␤ promoter activity in thyrotropes. DISCUSSION In this report we have used mutagenesis, DNA footprinting, and gene transfer studies to dissect the functionally important P1 region of the mTSH␤ promoter. These results strongly suggest that more than one protein interacts within this region, a predictable conclusion when one considers the size of the native thyrotrope-derived footprint (47 base pairs). Many lines of evidence indicate that the pituitary transcription factor Pit-1 or one of its isoforms is one of the factors that interacts with this region and appears necessary for activation of the TSH␤ promoter. Pit-1 interacts with three regions in the Ϫ390 to ϩ40 segment of the mTSH␤ promoter which confers cell-specific activity, D1 (Ϫ295 to Ϫ222), P1 (Ϫ133 to Ϫ86), and P2 (Ϫ80 to Ϫ62) (Ref. 5, Fig. 2A). Several reports have shown that Pit-1 is necessary for TRH-and cAMP-mediated stimulation of the TSH␤ promoter, occurring through the D1-and P1-related regions in the rat and human promoters respectively (6 -9, 19, 28). Kim et al. (9) propose that Pit-1 acts together with an AP-1-like factor, which interacts with a TGGGTCA motif at Ϫ1 to ϩ6 of the hTSH␤ promoter to mediate TRH stimulation. Steinfelder et al. (19) observed that mutations of either Pit-1 site in the P1 equivalent region of the hTSH␤ promoter reduced activity by approximately 50% in GH3 cells, while forskolin and TRH stimulation was reduced by more than 60% in the up-stream Pit-1 site (P1M3 equivalent) and only 20 -30% in the downstream site (P1M6 equivalent). Lin and colleagues (22) mutated two putative Pit-1 binding sites, the P1M3 equivalent at Ϫ122 to Ϫ116 and a proximal site at Ϫ76 to Ϫ69. This double mutant promoter was no longer stimulated by Pit-1 in CV-1 cells. Basal activity in pituitary-derived cells was not studied.
We have recently shown that a mutation in the P1 region of the mTSH␤ promoter, which disrupted Pit-1 binding, abrogated basal activity of the promoter in TtT-97 thyrotropes, but had no effect in GH3 somatotropes (20), suggesting that both Pit-1 and the P1 region are necessary for cell-specific basal activity of the mTSH␤ promoter only in thyrotropes. However, introduction by gene transfer of Pit-1 into various Pit-1-deficient cell types has little or no effect on TSH␤ promoter activity (5,8,9,19,21), indicating that while Pit-1 may be necessary for TSH␤ promoter activity, it is not sufficient for promoter activity.
In this study, we have identified a segment of the P1 region, the proximal P1 region (PP-1), which interacts with a non-Pit-1 protein and is separately and equally critical to the region that interacts with Pit-1 for activity of the mTSH␤ promoter in thyrotropes. Systematic scanning mutagenesis of this region has revealed separate contact areas for these two distinct proteins. While all of the mutations affected promoter activity in thyrotropes, they appeared to segregate into mutations which did not affect protein-DNA interactions (less than 45% reduction in activity) and mutations which disrupted binding (greater than 60% reduction). The reduction of promoter activity seen with mutations that did not affect protein-DNA interactions may reflect sensitivity differences between functional and structural assays. P1M1 and P1M2 mutations (Fig. 1) appear to have little effect on protein binding and modest decreased function of the mTSH␤ promoter. Interestingly, the P1M2 mutation changes the final T to a G in the Pit-1 binding site AATNCAT. P1M3 and P1M4 mutations both affect Pit-1 interaction with the P1 region, but do not alter protein binding at the PP-1 region. These mutations cause a greater reduction in basal activity of the mTSH␤ promoter than the mutations which do not affect protein-DNA interaction. The P1M3 mutation significantly alters the Pit-1 binding sequence, while the P1M4 mutation does not directly alter this sequence but creates changes seen with a previously described mutation (20) that also affected Pit-1 interaction with the promoter. The P1M4 mutation does change the AA preceeding the Pit-1 binding site to CC. This A/T-rich region preceeding the core binding site is necessary for high-affinity Pit-1 interaction as well as promoter function (20,29,30). The P1M5 mutation appears to alter interaction of both proteins with the promoter. Promoter activity is reduced greater than 60%, but not to the level seen with the M3, M4, and M7 mutations which disrupt individual protein binding. It appears that this mutation affects binding of both proteins, but Dnase I protection (Fig. 2B) shows that Pit-1 and TtT-97 footprints are slightly different than the bovine serum albumin control and may reflect partial protein-DNA interaction. Surprisingly, the P1M6 mutation, while altering a putative Pit-1 binding sequence (AAANCAT), has no effect on protein-DNA interaction and modestly affects promoter function. This data implies that the upstream Pit-1 binding element is the critical element for protein-DNA interaction. The P1M7 mutation does not affect Pit-1 interaction with the promoter, but the novel protein no longer binds to the PP-1 region. This is the only mutation that affects binding in the PP-1 region and yet does not alter Pit-1 interaction with the promoter. Basal activity of the promoter is greatly reduced by this mutation, suggesting that the two factors can interact with the promoter independently of binding by the other, but both are necessary FIG. 4. Southwestern blot of nuclear extract proteins. 50 g of each protein extract was size-separated on a 10% SDS-acrylamide gel and transferred to nitrocellulose. After denaturation-renaturation with guanidine hydrochloride, binding reaction was carried out with a [␥-32 P]dATP end-labeled duplex oligonucleotide corresponding to the Ϫ177 to Ϫ88 wild-type (lanes 1-3) and M7 mutated (lanes 4 -6) region of the mTSH␤ promoter. TtT-97 (lanes 1 and 4), ␣TSH (lanes 2 and 5), and GH3 (lanes 3 and 6) are noted. Molecular size standards are noted at the left.
for full basal activity of the promoter. These systematic mutations clearly illustrate that two separate proteins are interacting at overlapping regions on the mTSH␤ promoter and that neither factor alone is capable of conferring full promoter activity in thyrotropes. Identification of the factor that interacts with the PP-1 region, and the nature of its functional interaction with Pit-1 will greatly increase our understanding of thyrotrope-specific activity of the TSH␤ promoter.
Using Southwestern blot analysis with a duplex oligonucleotide corresponding to the PP-1 segment of the P1 region (Ϫ117 to Ϫ88), we have shown that a 50-kDa protein in thyrotropederived cells interacts with this region. Examination of the sequence in this PP-1 region reveals an AGATAA motif at Ϫ98 to Ϫ93. The M7 mutation (Fig. 1), which disrupts this AGATAA sequence, abrogates both functional and structural interaction of the TSH␤ promoter and the 50-kDa protein. This sequence resembles a consensus binding site for the GATA family of transcription factors designated as GATA-1 through GATA-6 (31-34). Steger et al. (34) have shown that the pituitary-derived ␣T3-1 gonadotrope cell line contains mRNA for mGATA-2 as well as an mGATA-4 related factor, which has a different message size than the previously reported mGATA-4 (35). They also demonstrated by gel mobility shift analysis that two proteins interacted with a human ␣-subunit gene GATA-containing element (Ϫ224 to Ϫ97), and a mutation of this GATA element decreased transcriptional activity of the ␣-subunit promoter 2.5-fold in ␣T3-1 cells. In this report, we show that a mutation of this GATA element (M7) reduces transcriptional activity of the TSH␤ promoter 5-fold in thyrotropes. Since the proximal portion of the P1 element in the mTSH␤ promoter contains this consensus AGATAA sequence, and GATA proteins are present in pituitary-derived cell lines, a GATA-related protein may be present in thyrotropes and involved in expression of the TSH␤ gene in thyrotropes. Furthermore, the 50-kDa protein identified by Southwestern blot analysis is consistent with sizes of GATA factors reported previously (34). Two groups have recently described another homeodomain transcription factor, mLIM-3 (24) or P-Lim (23). mLIM-3 is a LIM (Lin-11, Isl-1, Mec-3) homeodomain protein which has a cysteine-rich domain which contains two adjacent zinc-coordinated structures, referred to as the LIM domain. This transcription factor mRNA has been identified in adult mouse pituitary as well as pituitary-derived GH3, GH4C1, and ␣T3 cell lines (24). In the developing mouse pituitary, detection of mRNA for mLIM-3 was coincident with the appearance of Rathke's pouch at embryonic day 9 (e9) (23). In the adult mouse, mRNA was confined to the anterior pituitary gland and neurointermediate lobe and not in other areas of the brain or the liver (23,24). Bach et al. (23) performed gene transfer studies in CV-1 cells by co-transfection of the mTSH␤ promoter (Ϫ1.2 kb), cytomegalovirus promoter-directed mLIM-3, and CMV-directed rPit-1. While mLIM-3 alone did not stimulate the TSH␤ promoter, the combination of Pit-1 and mLIM-3 cooperatively stimulated the promoter (50-fold) greater than Pit-1 alone (10-fold). However, mLIM-3 alone stimulated the ␣-subunit promoter. They extended their studies to protein-DNA interaction which showed that a truncated version of mLIM-3 interacted with the Ϫ120 to Ϫ60 region of the mTSH␤ promoter, but the full-length protein did not. The Ϫ120 to Ϫ60 region of the mTSH␤ promoter contains the proximal portion of the P1 footprint (Ϫ133 to Ϫ98) as well as the P2 footprint (Ϫ86 to Ϫ64). Both of these regions are AT-rich and could conceivably interact with a homeodomain protein such as mLIM-3. mLIM-3 is reported to contain 400 amino acids and has a predicted protein size of 50 kDa. Interestingly, both groups reported that mLIM-3 mRNA was present in GH3 cells, and our data indicate that the 50-kDa protein is not present in GH3 cells by footprint or Southwestern blot analysis. This suggests that the 50-kDa protein we have identified is unlikely to be mLIM-3. Therefore, mLIM-3 may be interacting with the mTSH␤ promoter in the P2 region. Alternatively, mLIM-3 protein may not be sufficiently expressed in GH3 cells to be detectable by footprint or Southwestern blot analysis.
In summary, we have identified a 50-kDa protein in thyrotropes that appears to be necessary for basal activity of the mTSH␤ promoter. Further characterization of this factor will lead to a better understanding of TSH␤ gene regulation and cell-specific gene expression.