Novel Function of the Transactivation Domain of a Pituitary-specific Transcription Factor, Pit-1
2002; Elsevier BV; Volume: 277; Issue: 47 Linguagem: Inglês
10.1074/jbc.m202991200
ISSN1083-351X
AutoresMasahiko Kishimoto, Yasuhiko Okimura, Kazuhiro Yagita, Genzo Iguchi, Mariko Fumoto, Keiji Iida, Hidesuke Kaji, Hitoshi Okamura, Kazuo Chihara,
Tópico(s)Genetic Syndromes and Imprinting
ResumoPit-1 stimulates the expression of growth hormone, prolactin, and thyrotropin β subunit genes. Consequently, abnormality of the Pit-1 gene results in combined pituitary hormone deficiency (CPHD). In this study, we analyzed the function of Pit-1 with a mutation (proline to leucine at codon 24) in the transactivation domain, P24L, which has a normal POU domain important for binding to DNA, because this mutation had been reported in a patient with CPHD. We found that codon 24 proline in the transactivation domain as well as the POU domain of Pit-1 was crucial to recruit coactivator CREB-binding protein (CBP) in the cultured cells. P24L completely lost the responsiveness to cAMP to stimulate the expression of the Pit-1-targeted genes. Furthermore, CBP and Pit-1, but not P24L, markedly enhanced the expression of the Pit-1-targeted gene to cAMP, and adenovirus E1a that binds to CBP and abrogates its function blocked the induction by cAMP of Pit-1-stimulated gene transcription in the pituitary-derived GH3 cells. These results suggest that CBP and proline at codon 24 in the transactivation domain of Pit-1 are important for the cAMP-induced activation of Pit-1-targeted genes. However, P24L maintained basal transcriptional activity, suggesting that CBP is unlikely to be an essential coactivator for Pit-1. Pit-1 stimulates the expression of growth hormone, prolactin, and thyrotropin β subunit genes. Consequently, abnormality of the Pit-1 gene results in combined pituitary hormone deficiency (CPHD). In this study, we analyzed the function of Pit-1 with a mutation (proline to leucine at codon 24) in the transactivation domain, P24L, which has a normal POU domain important for binding to DNA, because this mutation had been reported in a patient with CPHD. We found that codon 24 proline in the transactivation domain as well as the POU domain of Pit-1 was crucial to recruit coactivator CREB-binding protein (CBP) in the cultured cells. P24L completely lost the responsiveness to cAMP to stimulate the expression of the Pit-1-targeted genes. Furthermore, CBP and Pit-1, but not P24L, markedly enhanced the expression of the Pit-1-targeted gene to cAMP, and adenovirus E1a that binds to CBP and abrogates its function blocked the induction by cAMP of Pit-1-stimulated gene transcription in the pituitary-derived GH3 cells. These results suggest that CBP and proline at codon 24 in the transactivation domain of Pit-1 are important for the cAMP-induced activation of Pit-1-targeted genes. However, P24L maintained basal transcriptional activity, suggesting that CBP is unlikely to be an essential coactivator for Pit-1. A pituitary-specific transcription factor, Pit-1, also known as GHF-1, is a member of the homeobox POU family of DNA-binding proteins (1Ingraham H.A. Chen R. Mangalam H.J. Elsholtz H.P. Flynn S.E. Lin C.R. Simmons D.M. Swanson L. Rosenfeld M.G. Cell. 1988; 55: 519-529Google Scholar, 2Bodner M. Castrillo J.L. Theill L.E. Deerinck T. Ellisman M. Karin M. Cell. 1988; 55: 505-518Google Scholar). It is involved in the development of the three pituitary cell types, somatotrophs, lactotrophs, and thyrotrophs, as well as the gene expression of growth hormone (GH), 1The abbreviations used are: GH, growth hormone; GHRH-R, growth hormone-releasing hormone receptor; TSH β, thyrotropin β subunit; CPHD, combined pituitary hormone deficiency; CBP, CREB-binding protein; CREB, cAMP-response element-binding protein; CMV, cytomegalovirus; GFP, green fluorescent protein; EGFP, epidermal growth factor protein; HA, hemagglutinin; RSV, Rous sarcoma virus; CPT, chlorophenylthio. prolactin (PRL), GH-releasing hormone receptor (GHRH-R), thyrotropin β subunit (TSH β), and Pit-1 itself by binding to the specific DNA elements of these genes (3Dolle P. Gastrillo J.L. Theill L.E. Deerinck T. Ellisman M. Karin M. Cell. 1990; 60: 809-820Google Scholar, 4Simmons D.M. Voss J.W. Ingraham H.A. Holloway J.M. Broide R.S. Rosenfeld M.G. Swanson L.W. Genes Dev. 1990; 4: 695-711Google Scholar, 5Ryan A.K. Rosenfeld M.G. Genes Dev. 1997; 11: 1207-1225Google Scholar, 6Rhodes S.J. Chen R. DiMattia G.E. Scully K.M. Kalla K.A. Lin S. Yu C. Rosenfeld M.G. Genes Dev. 1993; 7: 913-932Google Scholar, 7Li S. Crenshaw III, E.B. Rawson E.J. Simmons D.M. Swanson L.W. Rosenfeld M.G. Nature. 1990; 347: 528-533Google Scholar, 8Lin C. Lin S.C. Chang C.P. Rosenfeld M.G. Nature. 1992; 360: 765-768Google Scholar, 9Petersenn S. Rasche A.C. Heyens M. Schulte H. Mol. Endocrinol. 1998; 12: 233-247Google Scholar, 10Iguchi G. Okimura Y. Takahashi T. Mizuno I. Fumoto M. Takahashi Y. Kaji H. Abe H. Chihara K. J. Biol. Chem. 1999; 274: 12108-12114Google Scholar). Therefore, abnormalities of the Pit-1 gene result in GH, PRL, and TSH deficiencies. To date, at least 16 different mutations of the human Pit-1gene related to combined pituitary hormone deficiency (CPHD) have been reported (11–24). Most mutations are seen in the POU domain of Pit-1, which is important for its ability to dimerize and bind to DNA (25Sock E. Enderich J. Rosenfeld M.G. Wegner M. J. Biol. Chem. 1996; 271: 17512-17518Google Scholar). However, two mutations located in the transactivation domain have also been reported in patients with CPHD. One is a C to T mutation in codon 14 of the Pit-1 gene, resulting in an amino acid change from proline (P) to leucine (L) (P14L) (19Fofanova O.V. Takamura N. Kinoshita E. Yoshimoto M. Tsuji Y. Peterkova V.A. Evgrafov O.V. Dedov I.I. Goncharov N.P. Yamashita S. Am. J. Med. 1998; 77: 360-365Google Scholar) and the other is a Cys to Thr mutation in codon 24 leading to a corresponding change from Pro to Leu (P24L) (20Ohta K. Nobukuni Y. Mitsubuchi H. Fujimoto S. Matsuo N. Inagaki H. Endo F. Matsuda I. Biochem. Biophys. Res. Commun. 1992; 189: 851-855Google Scholar). Because the patients with these mutations were both heterozygous for the mutation, it is believed that P14L and P24L may dominantly inhibit the transcriptional activity of the wild type Pit-1 (21Tatsumi K. Amino N. Growth Horm. IGF Res. 1999; 9: 18-23Google Scholar, 22Parks J.S. Abdul-Latif H. Kinoshita E. Meacham L.R. Pfaffle R.W. Brown M.R. Horm. Res. 1993; 40: 54-61Google Scholar, 23Pfaffle R. Kim C. Otten B. Wit J.M. Eiholzer U. Heimann G. Parks J. Horm. Res. 1996; 45: 25-28Google Scholar). However, the precise mechanism causing CPHD in patients with these mutations has not been clarified. In the study presented here, we showed that P14L exerts its transcriptional effect to a similar extent to the wild type. This suggests that P14L may not be responsible for CPHD. On the other hand, P24L showed reduced transcriptional activity and could be a cause of CPHD. In addition, P24L could not recruit its coactivator, CREB-binding protein (CBP) into the complex containing Pit-1 in the cultured cells. Furthermore, although P24L did not lose its transcriptional activity completely, it did not respond to stimulation by cAMP. Although it remains possible that non-CBP proteins are responsible for this phenomenon, CBP as well as Pit-1 appears to be involved in cAMP-mediated enhancement of the gene expression via Pit-1- binding DNA elements. Fetal calf serum and Dulbecco's modified Eagle's medium were obtained from Invitrogen (Tokyo, Japan). COS7 cells were maintained in Dulbecco's modified Eagle's medium with 10% (v/v) calf serum. All culture medium contained penicillin G (100 units/ml) and kanamycin (100 μg/ml). Human Pit-1 cDNA was subcloned into an expression vector (pcDNA3.1), which was named pcDNA3.1-wild type-Pit-1. The mutant forms of Pit-1 (P14L, P24L, and E250X) expression vectors were constructed with a site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. Briefly, Pfu Turbo DNA polymerase was used to react 50 ng of template DNA (pcDNA3.1-wild type-Pit-1) with mutant sense primer (5′-GATACCTTTATACTTCTGAATTCTG-3′) and mutant antisense primer (5′-CAGAATTCAGAAGTATAAAGGTATC-3′) for P14L, with mutant sense primer (5′-CTGCAACTCTGCTTCTGATAATGC-3′) and mutant antisense primer (5′-GCATTATCAGAAGCAGAGTTGCAG-3′) for P24L, and with mutant sense primer (5′-GGATGGCTGAATAACTGAATCTGG-3′) and mutant antisense primer (5′-CCAGATTCAGTTATTCAGCCATCC-3′) for E250X. This reaction involved 30 s of denaturation at 95 °C and 15 cycles consisting of 30 s of denaturation at 95 °C, 1 min of annealing at 55 °C, and 5 min of extension at 68 °C. After digestion of the nonmutated parental DNA template with DpnI, the mutant forms of Pit-1 expression vectors (named pcDNA3.1-P14L, pcDNA3.1-P24L, and pcDNA3.1-E250X) were successfully transformed. The correct sequence was confirmed by DNA sequencing. E250X is a previously reported mutant form of Pit-1 that has completely lost its transcriptional activity because of the defect in the binding activity to DNA (16Irie Y. Tatsumi K. Ogawa M. Kamijo T. Preeyasombat C. Suprasongsin C. Amino N. Endocr. J. 1995; 42: 351-354Google Scholar). The 1.8-kb rat GH and 0.6-kb ratPRL 5′-flanking regions were inserted upstream of the luciferase reporter gene and these plasmids were named GH-Luc and PRL-Luc, respectively. cDNAs for wild type Pit-1, P14L, P24L, the transactivation domain containing the codons from the first to the 128th, and the POU domain containing the codons from the 129th to the terminal codon of the wild type Pit-1 were inserted in-frame into theXhoI-KpnI sites in the multiple cloning site of pEGFPC3 (Clontech Laboratories, Inc. Palo Alto, CA) and named wild type pPit-1-EGFP, pP14L-EGFP, pP24L-EGFP, pTRANS-EGFP, and pPOU-EGFP, respectively. Wild type Pit-1, P14L, and P24L cDNA were inserted in-frame into BglII-KpnI sites in the multiple cloning site of the pFLAG-CMVTM-6a expression vector (Sigma) and these plasmids were named wild type pPit-1-FLAG, pP14L-FLAG, and pP24L-FLAG, respectively. DNA fragments encoding the transactivation domain and the POU domain were inserted in-frame into the BglII-KpnI sites in the multiple cloning site of the pFLAG-CMVTM-6a expression vector to produce pTRANS-FLAG and pPOU-FLAG, respectively. Expression vectors for adenovirus E1a, which binds with CBP and abrogates its function (pcDNA3.1-E1a), and the mutant form of E1a, which cannot bind with CBP (pcDNA3.1-Δ-E1a), were constructed by inserting DNA fragments of E1a and Δ2–36 E1a to pcDNA.3.1, respectively. In all the transient expression experiments using luciferase assay, plasmid was transfected to cells in 35-mm dishes, unless otherwise indicated, using LipofectAce (Invitrogen). First, 2 μg of GH-Luc or PRL-Luc were transfected into COS7 cells with 0.3 μg of pcDNA3.1, pcDNA3.1-wild type-Pit-1, pcDNA3.1-P14L, pcDNA3.1-P24L, or pcDNA3.1-E250X to evaluate the transcriptional activity of P14L and P24L. To clarify whether P14L and P24L have a dominant negative effect, 0.3 μg of pcDNA3.1-wild type-Pit-1, pcDNA3.1-P14L, or pcDNA3.1-P24L were transfected with 0.3 μg of pcDNA3.1-wild type-Pit-1 and 2 μg of PRL-Luc. In the experiment for dose dependent activation of PRL-Luc by Pit-1, varying amounts (0–0.3 μg) of wild type Pit-1 or mutant forms of Pit-1 (P14L and P24L) were used. To compare the function of P24L to activate Pit-1-targeted genes in response to cAMP with that of wild type Pit-1, 2 μg of PRL-Luc or 1P-Luc, which contains seven Pit-1-responsive elements derived from a Pit-1-binding DNA element, 1P, of the rat PRL gene, was transfected into the COS7 cells with 0.3 μg of empty expression vector (pcDNA3.1), pcDNA3.1-wild type-Pit-1, or pcDNA3.1-P24L. Twenty-four hours after transfection, 1.0 mm CPT-cAMP was added to the medium of the transfected cells. To assess the effect of exogenous CBP on Pit-1-targeted gene expression, 2 μg of 1P-Luc with or without 0.5 μg of CBP expression vector was transfected into the COS7 cells with 20 ng of empty expression vector (pcDNA3.1), pcDNA3.1-wild type-Pit-1, or pcDNA3.1-P24L. We used a minimal dose of Pit-1 expression vectors to assess the effect of CBP more clearly. Twenty-four hours after transfection, 1.0 mm CPT-cAMP was added to the medium of the transfected cells. To clarify the effect of endogenous CBP on the Pit-1-targeted gene expression, 2 μg of 1P-Luc was transfected with 1 μg of pcDNA3.1, pcDNA3.1-E1a, or pcDNA3.1-Δ-E1a into the GH3 cells in 35-mm dishes. Twenty-four hours after transfection, 1.0 mm CPT-cAMP was added to the medium of the transfected cells. In all the experiments, 20 ng of pRL-CMV containing the cDNA encoding Renilla luciferase (Promega, Tokyo, Japan) was co-transfected in each transfection to normalize the luciferase activity. In each of the experiments, cells were harvested 48 h after transfection, and luciferase activities were measured with a Turner design luminometer TD-20/20 using the dual luciferase assay system (Promega). Values are expressed as multiples of induction relative to the basic luciferase activity when only an empty expression vector was co-transfected and they represent the mean ± S.E. of at least three determinations. Three μg of pcDNA3.1-wild type-Pit-1, pcDNA3.1-P14L, or pcDNA3.1-P24L were introduced into COS7 cells in 100-mm dishes using LipofectAce with or without 3 μg of pcDNA3.1-wild type-Pit-1. An empty expression vector (pcDNA3.1) was also co-transfected to fix the total amount of these plasmids at 6 μg. Forty-eight hours after transfection, proteins were extracted from the cells and immunoprecipitated with anti-Pit-1 polyclonal antibody, Pit-1 (X-7) (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). These samples were resolved by SDS-PAGE and anti-Pit-1 antibody was used for Western blotting. As a negative control, 6 μg of pcDNA3.1 alone was transfected and analyzed. One μg of pEGFPC3, pPit-1-EGFP, pP14L-EGFP, pP24L-EGFP, pTRANS-EGFP, or pPOU-EGFP was introduced into COS7 cells in 35-mm dishes with the aid of LipofectAce. Forty-eight hours after transfection, fluorescence images were examined. Ten μg of pcDNA3.1-wild type-Pit-1 or pcDNA3.1-P24L were introduced into COS7 cells in 100-mm dishes with 20 μg of PRL-Luc and 10 μg of RSV-HA-CBP (hemagglutinin (HA)-tagged CBP expression vector). Forty-eight hours after transfection, proteins were extracted from the cells, and divided into two equal parts, one part for immunoprecipitation with anti-HA antibody (Santa Cruz Biotechnology) and the other part for immunoprecipitation with anti-Pit-1 antibody. These samples were resolved by SDS-PAGE and anti-Pit-1 antibody was used for Western blotting. Furthermore, whether the transactivation domain alone or the POU domain alone can recruit CBP was examined. Ten μg of PRL-Luc and 5 μg of RSV-HA-CBP were introduced into COS7 cells in 100-mm dishes with 5 μg of wild type pPit-1-FLAG, pP14L-FLAG, pP24L-FLAG, pTRANS-FLAG, or pPOU-FLAG using LipofectAce. In the experiment to assess whether P24L blocks the interaction of wild type Pit-1 with CBP, 10 μg of PRL-Luc and 5 μg of RSV-HA-CBP were introduced into COS7 cells in 100-mm dishes with 5 μg of wild type pPit-1-FLAG, 5 μg of pP24L-FLAG, 5 μg of pPit-1-FLAG + 5 μg of pP24L-FLAG, or 10 μg of pPit-1-FLAG. pFLAG-CMVTM-6a was used to fix the total amounts of plasmids transfected. In both experiments, proteins were extracted from the cells 48 h after transfection and each sample was divided into two equal parts, one for immunoprecipitation with anti-HA antibody and the other for immunoprecipitation with anti-FLAG M2 monoclonal antibody (Sigma). These samples were then resolved by SDS-PAGE and anti-FLAG M2 monoclonal antibody was used for Western blotting. Expression vectors for fusion protein of Pit-1 or Pit-1 domains with the GAL4 DNA binding domain, pGAL4-Pit-1, pGAL4-TRANS, pGAL4-POU, and pGAL4-P24L were constructed by inserting wild type Pit-1, the transactivation domain of wild type Pit-1, the POU domain of wild type Pit-1, and P24L cDNA in-frame into BamHI-XbaI sites in the multiple cloning site of the pCMV-BD vector (Stratagene), respectively. Expression vectors for the fusion protein of CBP with the NF-κB transactivation domain, pAD-CBP, pAD-CBP-(1–319), pAD-CBP-(320–420), pAD-CBP-(421–1457), pAD-CBP-(1458–1891), and pAD-CBP-(1892–2441), were constructed by inserting cDNAs coding full-length CBP, CBP-(1–319), CBP-(320–420), CBP-(421–1457), CBP-(1458–1891), and CBP-(1892–2441) in-frame into BamHI-HindIII sites in the multiple cloning site of the pCMV-AD vector (Stratagene), respectively. Two μg of reporter plasmid, pFR-Luc (Stratagene), which contains GAL4-responsive elements and 1 μg of pGAL4-Pit-1 were transfected into the COS7 cells in 35-mm dishes with 1 μg of pCMV-AD, pAD-CBP, pAD-CBP-(1–319), pAD-CBP-(320–420), pAD-CBP-(421–1457), pAD-CBP-(1458–1891), or pAD-CBP-(1892–2441). Furthermore, 2 μg of pFR-Luc and 1 μg of pAD-CBP were transfected into the COS7 cells in 35-mm dishes with 1 μg of pGAL4-TRANS, pGAL4-POU, or pGAL4-P24L. In all the experiments, 20 ng of pRL-CMV was co-transfected to normalize the luciferase activity. Cells were harvested 48 h after transfection and values are expressed as multiples of induction relative to the basic luciferase activity when pFR-Luc was transfected with pGAL4-Pit-1 and pCMV-AD. Transcriptional activity of P14L and P24L for the expression of GH and PRL genes was investigated in COS7 cells (Pit-1 deficient). P24L activated GH-Luc containing the 1.8-kb rat GH 5′-flanking region and PRL-Luc containing the 0.6-kb rat PRL 5′-flanking region up to only 69 and 56%, respectively, of activation by wild type Pit-1. On the other hand, P14L activated GH-Luc and PRL-Luc up to a similar level to wild type Pit-1 (Fig. 1, A andB). Furthermore, whereas P14L showed a similar dose-dependent enhancement of its power as a transcription factor to the wild type Pit-1, the dose-dependent enhancement of that of P24L was obviously weak (Fig. 1 C). However, whereas one mutant form of Pit-1, E250X, which cannot bind to DNA, completely lost its transcriptional activity (16Irie Y. Tatsumi K. Ogawa M. Kamijo T. Preeyasombat C. Suprasongsin C. Amino N. Endocr. J. 1995; 42: 351-354Google Scholar), P24L did not completely lose it (Fig. 1, A–C). Because both of the probands with P14L or P24L were found to be heterozygous for these mutations (19Fofanova O.V. Takamura N. Kinoshita E. Yoshimoto M. Tsuji Y. Peterkova V.A. Evgrafov O.V. Dedov I.I. Goncharov N.P. Yamashita S. Am. J. Med. 1998; 77: 360-365Google Scholar,20Ohta K. Nobukuni Y. Mitsubuchi H. Fujimoto S. Matsuo N. Inagaki H. Endo F. Matsuda I. Biochem. Biophys. Res. Commun. 1992; 189: 851-855Google Scholar), it has been hypothesized that P14L and P24L might inhibit the activity of the wild type Pit-1 (21Tatsumi K. Amino N. Growth Horm. IGF Res. 1999; 9: 18-23Google Scholar, 22Parks J.S. Abdul-Latif H. Kinoshita E. Meacham L.R. Pfaffle R.W. Brown M.R. Horm. Res. 1993; 40: 54-61Google Scholar, 23Pfaffle R. Kim C. Otten B. Wit J.M. Eiholzer U. Heimann G. Parks J. Horm. Res. 1996; 45: 25-28Google Scholar). To test the ability of these mutants to interfere with the transactivation of Pit-1-targeted gene expression by wild type Pit-1, the expression vector for P14L or P24L was transfected into COS7 cells together with the PRL-Luc and wild type Pit-1 expression vector. Whereas P14L, like wild type Pit-1, enhanced the transcriptional effect of the wild type in an additive manner, the effect of P24L on the activity of the wild type Pit-1 was very weak. However, neither P14L nor P24L dominantly inhibited the activity of the wild type Pit-1 (Fig. 2 A). To rule out the possibility that the differences in expression levels of wild type Pit-1 and mutant forms of Pit-1 (P14L and P24L) could explain the minor effect of P24L on basal transcription rates and the absence of a dominant negative effect of P14L and P24L on wild type Pit-1, we examined protein levels of wild type Pit-1 and mutant forms of Pit-1 using Western blot analysis. The Pit-1 protein levels were similar to each other whether 3 μg of wild type Pit-1, P14L, or P24L expression vector was transfected (Fig. 2 B, lanes 1–3). Furthermore, the levels of Pit-1 protein were similar to each other whether 3 μg of wild type Pit-1, P14L, or P24L expression vector was co-transfected with 3 μg of wild type Pit-1 expression vector (Fig.2 B, lanes 4–6). These results suggested that the reduced activity of P24L and the absence of a dominant negative effect of P14L and P24L were not explained by the difference in the levels of wild type and mutant forms of Pit-1 protein. Chimera constructs of GFP with wild type Pit-1, P14L, P24L, the transactivation domain alone, or the POU domain alone of the wild type Pit-1 (pPit-1-EGFP, pP14L-EGFP, pP24L-EGFP, pTRANS-EGFP, and pPOU-EGFP) were produced. These constructs were transiently transfected into COS7 cells, and the fluorescence image of the expressed fusion proteins were analyzed. Cells transfected with only the GFP expression vector were also analyzed as a control. In the control cells, GFP was distributed homogeneously in both the cytoplasm and the nucleus (Fig. 3 A). On the other hand, nuclear accumulation of wild type-Pit-1-GFP, P14L-GFP, and P24L-GFP was observed in all the cells analyzed (Fig. 3,B–D). Furthermore, nuclear accumulation of POU-GFP was observed in all the cells analyzed (Fig. 3 F). On the other hand, TRANS-GFP was distributed homogeneously in both the cytoplasm and nucleus (Fig. 3 E), suggesting that the POU domain but not the transactivation domain is involved in nuclear localization. A basic cluster RKRKRR is present in the POU homodomain, one of the subdomains of the POU domain. Because the basic cluster RKRKRR, which is highly conserved in POU proteins, has been identified as NLS in Tst-1/Oct6, a member of POU protein family (25Sock E. Enderich J. Rosenfeld M.G. Wegner M. J. Biol. Chem. 1996; 271: 17512-17518Google Scholar), it appears to work as NLS in Pit-1 as well as in Tst-1/Oct6. Because P24L showed normal nuclear translocation and contained normal POU domain that is important for dimerization and DNA binding (26Jacobson E.M. Li P. Leon-del-Rio A. Rosenfeld M.G. Aggarwal A.K. Genes Dev. 1997; 11: 198-212Google Scholar), we assumed that some transcriptional cofactors may be related with the reduced activity of P24L. Because CBP is involved in the activities of many transcription factors including Pit-1 (28Xu L. Lavinsky R.M. Dasen J.S. Flynn S.E. Mclerney E.M. Mullen T.M. Heinzel T. Szeto D. Korzus E. Kurosawa R. Aggarwal A.K. Rose D.W. Glass C.K. Rosenfeld M.G. Nature. 1998; 395: 301-306Google Scholar, 29Kamei Y. Heinzel T. Torchia J. Kurokawa R. Gloss B. Lin S.C. Heyman R.A. Rose D.W. Glass C.K. Rosenfeldt M.G. Cell. 1996; 85: 403-414Google Scholar, 30Kwok R.P.S. Lundland J.R. Chrivia J.C. Richard J.P. Bachinger H.P. Brenann R.G. Roberts S.G.E. Green M.R. Goodman R.H. Nature. 1994; 370: 223-226Google Scholar, 31Eckner R. Yao T.P. Oldread E. Livingston D.M. Genes Dev. 1996; 10: 2478-2490Google Scholar, 32Zhang J.J. Vinkemeier U. Gu W. Charavarti D. Horvath C.M. Darnell J.E. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 15092-15096Google Scholar, 33Perkins N.D. Felzien L.K. Betts J.C. Leung K. Beach D.H. Nabel G.J. Science. 1997; 275: 523-527Google Scholar), we investigated whether the impaired function of P24L is linked to the inability to recruit CBP. The whole cell extract prepared from COS7 cells, which had been transfected with the expression vector for wild type Pit-1 or P24L in addition to the expression vector for HA-tagged CBP (RSV-HA-CBP) and PRL-Luc, was immunoprecipitated with anti-HA antibody. Whether or not CBP is present in the complex containing wild type Pit-1 or P24L was analyzed by Western blotting with anti-Pit-1 antibody (Fig.4 A, lanes 1 and2). Although wild type Pit-1 was co-immunoprecipitated with CBP, P24L was not co-immunoprecipitated with CBP, suggesting that the ability to recruit CBP was decreased by the substitution of Leu for Pro at the 24th amino acid of Pit-1. Furthermore, whether or not the transactivation domain alone or the POU domain alone could recruit CBP was tested. The whole cell extract prepared from COS7 cells, which had been transfected with RSV-HA-CBP and PRL-Luc in addition to the FLAG-tagged expression vectors, pPit-1-FLAG, pP14L-FLAG, pP24L-FLAG, pTRANS-FLAG, or pPOU-FLAG, was divided into two equal parts. One part was used for immunoprecipitation with anti-HA antibody and the other for immunoprecipitation with anti-FLAG antibody and both were analyzed by Western blotting with anti-FLAG antibody (Fig. 4 B). pPit-1-FLAG, pP14L-FLAG, pP24L-FLAG, pTRANS-FLAG, or pPOU-FLAG appeared to be equally transfected (Fig. 4B, lanes 6–10; immunoprecipitated with anti-FLAG antibody and Western blotted with anti-FLAG antibody). However, neither P24L, the transactivation domain, nor the POU domain were detected in the complex containing CBP (Fig.4 B, lanes 1–5; immunoprecipitated with anti-HA antibody and Western blotted with anti-FLAG antibody). Therefore, the transactivation domain, especially proline at codon 24 of Pit-1 seemed to be important to make a complex containing CBP in the cultured cells (Fig. 4 B, lanes 3 and 8). Furthermore, in contrast to the result of a previous study (28Xu L. Lavinsky R.M. Dasen J.S. Flynn S.E. Mclerney E.M. Mullen T.M. Heinzel T. Szeto D. Korzus E. Kurosawa R. Aggarwal A.K. Rose D.W. Glass C.K. Rosenfeld M.G. Nature. 1998; 395: 301-306Google Scholar), the POU domain was not able to recruit CBP (Fig. 4 B, lanes 5 and10). It was surprising because a previous report (28Xu L. Lavinsky R.M. Dasen J.S. Flynn S.E. Mclerney E.M. Mullen T.M. Heinzel T. Szeto D. Korzus E. Kurosawa R. Aggarwal A.K. Rose D.W. Glass C.K. Rosenfeld M.G. Nature. 1998; 395: 301-306Google Scholar) suggested that the POU domain was sufficient to make a complex with the CBP fragment, although the binding activity appeared to be weak. Our result suggested that the transactivation domain alone or the POU domain alone are insufficient to make a high affinity complex that contains CBP in vivo cell culture. In addition, whether or not P24L blocks the interaction of wild type Pit-1 with CBP was also examined (Fig. 4 C). The whole cell extract prepared from COS7 cells, which had been transfected with RSV-HA-CBP and PRL-Luc in addition to expression vectors for FLAG-tagged wild type Pit-1, FLAG-tagged P24L, or both of them, was divided into two equal parts. One part was used for immunoprecipitation with anti-HA antibody and the other for immunoprecipitation with anti-FLAG antibody and both were analyzed by Western blotting with anti-FLAG antibody (Fig.4 C). Although P24L could not make a complex containing CBP in the cultured cells, it did not block the interaction of wild type Pit-1 with CBP (Fig. 4 C, lanes 2 and3, immunoprecipitated with anti-HA antibody and Western blotted with anti-FLAG antibody). It is possible that loss of the nuclear localization is responsible for the disturbed interaction of the transactivation domain and CBP in the co-immunoprecipitation experiment. Therefore, we performed mammalian two-hybrid assays. When pGAL4-Pit-1-producing fusion protein of the GAL4 DNA binding domain with Pit-1 and the pAD-CBP coding NF-κB activation domain and CBP were co-transfected into the COS7 cells, the pFR-Luc that contains the GAL4-binding elements was activated (Fig.5). However, when pAD-CBP was co-transfected with pGAL4-P24L, pGAL4-TRANS producing a fusion protein of GAL4 and the transactivation domain, pGAL4-POU producing a fusion protein of GAL4 and the POU domain, pFR-Luc was not activated (Fig. 5). This result indicated that the transactivation domain alone could not interact with CBP, and was in accordance with the result from immunoprecipitation. On the other hand, pAD-CBP- (1–319), pAD-CBP-(320–420), pAD-CBP-(421–1457), pAD-CBP-(1458–1891), or pAD-CBP-(1892–2441) were co-transfected with pGAL4-Pit-1 and pFR-Luc to identify domains in CBP that interact with Pit-1. In accordance with the previous reports using the glutathione S-transferase pull-down assay (27Lill N.L. Grossman S.R. Ginsberg D. DeCaprio J. Livingston D.M. Nature. 1997; 387: 823-827Google Scholar, 34Cohen L.E. Hahimoto Y. Zanger K. Wondisford F. Radovick S. J. Clin. Invest. 1999; 104: 1123-1128Google Scholar, 35Zanger K. Cohen L.E. Hashimoto K. Radovick S. Wondisford F.E. Mol. Endocrinol. 1999; 13: 268-275Google Scholar), the regions containing the CH1 or CH3 domains seemed to be important for CBP to interact with Pit-1, because pAD-CBP-(320–420) and pAD-CBP-(1458–1891) could activate pFR-Luc (Fig. 5). CBP has been reported to be involved in cAMP-regulated gene expression via Pit-1-binding DNA elements (28Xu L. Lavinsky R.M. Dasen J.S. Flynn S.E. Mclerney E.M. Mullen T.M. Heinzel T. Szeto D. Korzus E. Kurosawa R. Aggarwal A.K. Rose D.W. Glass C.K. Rosenfeld M.G. Nature. 1998; 395: 301-306Google Scholar, 34Cohen L.E. Hahimoto Y. Zanger K. Wondisford F. Radovick S. J. Clin. Invest. 1999; 104: 1123-1128Google Scholar,35Zanger K. Cohen L.E. Hashimoto K. Radovick S. Wondisford F.E. Mol. Endocrinol. 1999; 13: 268-275Google Scholar). Therefore, we examined whether wild type Pit-1 and the mutant form (P24L) Pit-1 differ in their activation of the target genes after cAMP stimulation. Whereas cAMP stimulated the transcription of thePRL gene by wild type Pit-1, it did not affect the transcription by P24L (Fig.6 A). In addition, 1P-Luc containing seven Pit-1-responsive elements derived from a Pit-1 binding DNA element, 1P, of the rat PRL gene was also used because there were reports that the proximal region of the PRL gene contains a cAMP responsive element-like DNA element as well as Pit-1 binding sites (36Liang J. Kim K.E. Schoderbek W.E. Maurer R.A. Mol. Endocrinol. 1992; 6: 885-892Google Scholar, 37Peers B. Monget P. Nalda M.A. Voz M.L. Berwaer M. Belayew A. Martial J.A. J. Biol. Chem. 1991; 266: 18127-18134Google Scholar). cAMP stimulated the transcription of 1P-Luc by wild type Pit-1, but it did not affect the transcription by P24L (Fig.6 B). These results suggest that for Pit-1 making a complex containing CBP is associated with the cAMP-mediated signal pathway that stimulates the expression of the targeted genes. Besides, co-expression study revealed that CBP and cAMP enhanced the gene transcription by wild type Pit-1, but did not enhance the transcription by P24L (Fig.7 A). Furthermore, in the pituitary-derived GH3 cells, co-expression study revealed that adenovirus E1a, which binds with CBP and abrogates its function, suppressed the basal and cAMP-induced expression of 1P-Luc, whereas Δ-E1a that cannot bind to CBP did not (Fig. 7 B).Figure 7CBP is a mediator in the cAMP signal transduction pathway to activate Pit-1-targeted gene transcription. A, 2 μg of 1P-Luc containing seven Pit-1-binding elements was transfected into the COS7 cells in 35-mm dishes with or without 0.5 μg of CBP expression vector, in addition to 20 ng of empty expression vector (pcDNA3.1), pcDNA3.1-wild type-Pit-1, or pcDNA3.1-P24L. A minimal dose of Pit-1 expression vector was used to assess the effect of CBP more clearly. Empty expression vectors were used to fix the total amount of plasmids transfected. Twenty-four hours after transfection, 1.0 mmCPT-cAMP was added to the medium of the transfected cells. CBP and cAMP enhanced the gene transcription by wild type Pit-1, but did not enhance the transcription by P24L. B, 2 μg of 1P-Luc with 1 μg of pcDNA3.1, pcDNA3.1-E1a, or pcDNA3.1-Δ-E1a was transfected into the GH3 cells in 35-mm dishes. Twenty-four hours after transfection, 1.0 mm CPT-cAMP was added to the medium of the transfected cells. Wild type E1a, which binds to CBP and abrogates its function, suppressed the basal and cAMP-induced expression of 1P-Luc, whereas Δ-E1a, which is unable to bind to CBP did not. Experiments were performed in triplicate and data show the mean ± S.E. stimulation of the reporter construct. Twenty ng of pRL-CMV was also co-transfected to normalize the luciferase activity.View Large Image Figure ViewerDownload (PPT) This is the first functional analysis of Pit-1 with mutations in the transactivation domain found in patients with CPHD (19Fofanova O.V. Takamura N. Kinoshita E. Yoshimoto M. Tsuji Y. Peterkova V.A. Evgrafov O.V. Dedov I.I. Goncharov N.P. Yamashita S. Am. J. Med. 1998; 77: 360-365Google Scholar, 20Ohta K. Nobukuni Y. Mitsubuchi H. Fujimoto S. Matsuo N. Inagaki H. Endo F. Matsuda I. Biochem. Biophys. Res. Commun. 1992; 189: 851-855Google Scholar). The case reported by Fofanova et al. (19Fofanova O.V. Takamura N. Kinoshita E. Yoshimoto M. Tsuji Y. Peterkova V.A. Evgrafov O.V. Dedov I.I. Goncharov N.P. Yamashita S. Am. J. Med. 1998; 77: 360-365Google Scholar) was a 3-year-old girl who showed total GH/PRL and partial TSH deficiency without mental retardation. All six exons of the Pit-1 gene and its promoter region were directly sequenced by the authors and it was found that the patient was heterozygous for a mutation at codon 14 of exon 1 (CCT to CTT) resulting in the substitution of Leu for Pro. However, her mother, her maternal aunt, and grandmother were phenotypically normal, although they had the same heterozygous mutation. The need to thoroughly assess genomic imprinting was emphasized by the authors. The patient presented by Ohta et al. (20Ohta K. Nobukuni Y. Mitsubuchi H. Fujimoto S. Matsuo N. Inagaki H. Endo F. Matsuda I. Biochem. Biophys. Res. Commun. 1992; 189: 851-855Google Scholar) was a 5-year-old boy of short stature without mental retardation. Both of the plasma GH and PRL levels were undetectable either before or after the provocative stimulation tests. The basal TSH level was also low. TRH caused a weak TSH response. DNA sequence of the Pit-1 gene showed that the patient had a heterozygous mutation at codon 24 in exon 1 (CCT to CTT) resulting in substitution of Leu for Pro. Neither phenotype nor genotype of the family was described. As mentioned above, because both of the patients had only one mutant allele, it had been predicted that P14L and P24L were likely to possess the dominant negative effect (21Tatsumi K. Amino N. Growth Horm. IGF Res. 1999; 9: 18-23Google Scholar, 22Parks J.S. Abdul-Latif H. Kinoshita E. Meacham L.R. Pfaffle R.W. Brown M.R. Horm. Res. 1993; 40: 54-61Google Scholar, 23Pfaffle R. Kim C. Otten B. Wit J.M. Eiholzer U. Heimann G. Parks J. Horm. Res. 1996; 45: 25-28Google Scholar). However, neither of these mutations is likely to show dominant negative activity judging from our data. Especially, P14L showed normal transcriptional activity and therefore P14L did not seem to be responsible for CPHD, although it remains possible that this mutant lead to CPHD through an unknown mechanism in vivo. On the other hand, P24L could be a cause of CPHD in view of a previous report of monoallelic expression of thePit-1 gene (38Okamoto N. Wada Y. Ida S. Koga R. Ozono K. Chiyo H. Hayashi A. Tatsumi K. Hum. Mol. Genet. 1994; 3: 1565-1568Google Scholar). It was possible that only the mutant allele was expressed in this patient through a genomic imprinting mechanism rather than some other genomic or nongenomic etiology. Because CBP is associated with so many different transcription factors, mutations in the gene encoding CBP cause developmental disorder (known as the Rubinstein-Taybi syndrome) comprised of multiple abnormalities, broad thumbs and halluces, mental retardation, growth retardation, developmental delay, microcephaly, and craniofacial abnormalities (39Rubinstein J.H. Am. J. Med. Genet. 1990; 6: 3-16Google Scholar,40Petrij F. Giles R.H. Dauwerse H.G. Saris J.J. Hennekam R.C. Masuno M. Tommerup N. von Ommen G.J. Goodman R.H. Peters D.J. Breuning M.H. Nature. 1995; 376: 348-351Google Scholar). Rubinstein-Taybi syndrome is believed to be a haploinsufficient disorder (41Goodman R.H. Smolik S. Genes Dev. 2000; 14: 1553-1577Crossref Google Scholar) and homozygous pathological mutation of CBP is lethal (42Oike Y. Takakura N. Hara A. Kaname T. Akizuki M. Yamaguchi Y. Yasue H. Araki K. Yamamura K. Suda T. Blood. 1999; 93: 2771-2779Google Scholar). On the other hand, CBP appears to be haplosufficient for GH, PRL, and TSH gene expression, because there have been no reports of Rubinstein-Taybi syndrome with deficiencies of these hormones (43Olson D.P. Koenig R.J. J. Clin. Endocrinol. Metab. 1997; 82: 3264-3266Google Scholar). However, it is possible that if only the mutant form of Pit-1, which cannot recruit CBP, is expressed, not only the expression of the Pit-1-targeted genes, but also the development of somatotrophs, lactotrophs, and thyrotrophs might be impaired. Previously, we showed that Pit-1-binding DNA elements mediate the transcriptional response to cAMP through a mechanism that does not require inducible phosphorylation of Pit-1 (44Okimura Y. Howard P.W. Maurer R.A. Mol. Endocrinol. 1994; 8: 1559-1565Google Scholar). Although major phosphorylation sites of Pit-1 are serine 115 and threonine 220 (45Kapiloff M.S. Farkash Y. Wegner M. Rosenfeld M.G. Science. 1991; 253: 786-789Google Scholar), the mutations of those amino acids did not affect the transcriptional function in response to cAMP (44Okimura Y. Howard P.W. Maurer R.A. Mol. Endocrinol. 1994; 8: 1559-1565Google Scholar). This finding suggested the cAMP effects on the Pit-1- dependent gene expression might involve other proteins that associate with Pit-1 (44Okimura Y. Howard P.W. Maurer R.A. Mol. Endocrinol. 1994; 8: 1559-1565Google Scholar). Because the present study demonstrated that mutation of Pro at codon 24 of Pit-1 completely disrupted the transcriptional response to cAMP, and that adenovirus E1a that binds to CBP and abrogates its function blocked the stimulatory effect of cAMP on Pit-1-dependent gene transcription in the pituitary-derived GH3 cells, we speculated that the inability of the mutant form of Pit-1 (P24L) to recruit CBP is involved in the phenomenon (Fig. 8). However, how CBP is associated with the cAMP-mediated activation of Pit-1-targeted genes still remains to be clarified (28Xu L. Lavinsky R.M. Dasen J.S. Flynn S.E. Mclerney E.M. Mullen T.M. Heinzel T. Szeto D. Korzus E. Kurosawa R. Aggarwal A.K. Rose D.W. Glass C.K. Rosenfeld M.G. Nature. 1998; 395: 301-306Google Scholar, 35Zanger K. Cohen L.E. Hashimoto K. Radovick S. Wondisford F.E. Mol. Endocrinol. 1999; 13: 268-275Google Scholar, 41Goodman R.H. Smolik S. Genes Dev. 2000; 14: 1553-1577Crossref Google Scholar). It was very interesting that the basal activity of P24L as a transcription factor was not disrupted completely. This suggests that other coactivators for Pit-1 that function independently of CBP may exist. This functional analysis of the naturally occurring mutation in the transactivation domain of Pit-1 provided an interesting insight into the mechanism by which Pit-1 stimulates gene expression. However, the structure and function of the transactivation domain of Pit-1 remains to be clarified, although those of the POU domain have been extensively studied using NMR, crystallographic methods, and mutational analysis (26Jacobson E.M. Li P. Leon-del-Rio A. Rosenfeld M.G. Aggarwal A.K. Genes Dev. 1997; 11: 198-212Google Scholar, 46Liang J. Moye-Rowley S. Maurer R.A. J. Biol. Chem. 1995; 270: 25520-25525Google Scholar, 47Andersen B. Rosenfeld R.G. Endocr. Rev. 2001; 22: 2-35Google Scholar). In conclusion, it was shown first that the transactivation domain, especially Pro at codon 24 of Pit-1 is important to recruit CBP in the cultured cells. Second, CBP is a mediator in the cAMP signal transduction pathway to activate Pit-1-targeted gene transcription. And third, CBP does not appear to be an essential coactivator for Pit-1 to maintain basal transcription, suggesting that some other unknown coactivators functioning independently of CBP may exist. We thank Dr. Fukamizu for the gift of RSV-HA-CBP, Dr. Yang Shi for the gift of E1a expression vector, and Chika Ogata for excellent technical assistance.
Referência(s)