J Biol Chem, Vol. 275, Issue 5, 3100-3106, February 4, 2000

The Pit-1 Homeodomain and -Domain Interact with Ets-1 and Modulate Synergistic Activation of the Rat Prolactin Promoter^*

Andrew P. Bradford^§¶, Kelley S. Brodsky, Scott E. Diamond, Laura C. Kuhn, Yingmiao Liu^**, and Arthur Gutierrez-Hartmann^§**

From the Departments of  Obstetrics and Gynecology, ^§ Biochemistry and Molecular Genetics, and  Medicine, and ^** Program in Molecular Biology, University of Colorado Health Sciences Center, Denver, Colorado 80262

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Pit-1/GHF-1 is a pituitary-specific, POU homeodomain transcription factor required for development of somatotroph, lactotroph, and thyrotroph cell lineages and regulation of the temporal and spatial expression of the growth hormone, prolactin (PRL), and thyrotropin- genes. Synergistic interaction of Pit-1 with a member of the Ets family of transcription factors, Ets-1, has been shown to be an important mechanism regulating basal and Ras-induced lactotroph-specific rat (r) PRL promoter activity. Pit-1/GHF-2, an alternatively spliced isoform containing a 26-amino acid insert (-domain) within its transcription-activation domain, physically interacts with Ets-1 but fails to synergize. By using a series of Pit-1 internal-deletion constructs in a transient transfection protocol to reconstitute rPRL promoter activity in HeLa cells, we have determined that the functional and physical interaction of Pit-1 and Ets-1 is mediated via the POU homeodomain, which is common to both Pit-1 and Pit-1. Although the Pit-1 homeodomain is both necessary and sufficient for direct binding to Ets-1 in a DNA-independent manner, an additional interaction surface was mapped to the -domain, specific to the Pit-1 isoform. Thus, the unique transcriptional properties of Pit-1 and Pit-1 on the rPRL promoter may be due to the formation of functionally distinct complexes of these two Pit-1 isoforms with Ets-1.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Pit-1/GHF-1 is a POU homeobox transcription factor specifically expressed in the anterior pituitary that not only specifies somatotroph, lactotroph, and thyrotroph cell lineages but also regulates growth hormone (GH),¹ prolactin (PRL), and thyrotropin (TSH) gene expression (1-3). The critical importance of Pit-1 for the ontogeny of these cell fates, and for the cell-specific spatial and temporal expression of the GH, PRL, and TSH genes, has been well documented (3-5). However, since Pit-1 is expressed in somatotroph, lactotroph, and thyrotroph cells, yet each expresses a highly specialized and distinct peptide hormone, factors other than Pit-1 must be involved in the regulation of these pituitary-specific genes (6, 7). Combinatorial interactions between Pit-1 and other trans-acting factors also play a critical role in the regulation of PRL, GH, and TSH gene expression by hormones and growth factors. Thus, Pit-1 has been proposed to serve as a cell-specific signal integrator by functionally interacting with other transcription factors at composite or adjacent response elements (8, 9). Indeed, specific interaction of Pit-1 with Ets-1, the proto-typical member of the ETS family of transcription factors, is required not only for optimal lactotroph-specific basal rPRL promoter activity (10) but also mediates Ras-induced rPRL gene transcription (9, 11-13). The Ets-1/Pit-1 combination, acting via a composite DNA element, also provides a molecular mechanism to account for differential regulation of the Pit-1-dependent rPRL and rGH genes (10, 11).

Differential splicing of the Pit-1 transcript generates the functionally distinct Pit-1 isoform, which contains a 26-amino acid (aa) insertion (-domain) within the amino-terminal transcription activation domain (14-16). Pit-1 and Pit-1 exhibit distinct transcriptional properties with respect to the regulation of the GH, PRL, and TSH promoters (9, 10, 14-18). Specifically, Pit-1 inhibits both basal and Ras-stimulated rPRL promoter activity in GH4 rat pituitary cells and fails to synergize functionally with Ets-1 in a gene transfer reconstitution assay (9, 10, 17). Since both Pit-1 and Pit-1 have been shown to interact physically with Ets-1 (10), the precise mechanism underlying the differential regulation of the rPRL promoter by these two isoforms remains to be elucidated.

By using a transient transfection protocol to reconstitute rPRL promoter activity in a nonpituitary HeLa cell line, we have previously demonstrated (10) synergistic transcriptional activation by Pit-1 and Ets-1 and mapped the region of Ets-1 required for both functional and physical interaction with Pit-1. In vivo co-localization of transfected Ets-1 and Pit-1 has also been observed in HeLa cells, using fluorescence resonance energy transfer microscopy (19). In the studies reported here, we utilize the reconstitution system and a protein-protein interaction assay to identify key domains of Pit-1 and Pit-1 required for functional and physical combinatorial interactions with Ets-1. Our results show that Pit-1 and Pit-1 share a common primary carboxyl-terminal Ets-1 binding motif located within the homeodomain. However, Pit-1 contains a secondary amino-terminal Ets-1 interaction surface that maps to the unique -domain. Thus, discrete physical interactions of Pit-1 and Pit-1 with Ets-1 result in the formation of functionally distinct, isoform-specific transcriptional complexes that differentially regulate rPRL gene transcription.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Plasmid Constructs-- The reporter constructs pA3PRLluc and pCMV (CLONTECH) have been described previously (9, 20). Plasmid pSG5c-Ets-1 encodes the p68 chicken Ets-1 under control of the SV40 early promoter (21). Pit-1 constructs were cloned into the expression vector pCGN2 (22), introducing a hemagglutinin (HA) epitope tag (YPYDVPDYA) at the amino terminus. The Pit-1 expression plasmids, pRc.935Pit1wt FLAG and pRc.935Pit1A3 FLAG, encoding wild-type and a Pit-1 phosphorylation mutant, respectively (23, 24), were modified and generously provided by Dr. Fred Schaufele (University of California, San Francisco). GST-Pit-1 internal deletions (15) were constructed by polymerase chain reaction and cloned into a modified pGEX 2TK vector, pGexDFGK, incorporating the multicloning site derived from pCGN2, constructed and provided by Dr. David Gordon, University of Colorado Health Sciences Center.

The plasmids pCGN2 Pit-1-(199-291), pCGN2 Pit-1-(2-82), and pCGN2 Pit-1-(2-108) encode fragments of Pit-1, Pit-1, and the -domain, respectively, fused to an amino-terminal HA tag. The plasmids pGex Pit-1-(199-291), pGex Pit-1-(2-82), pGex Pit-1-(2-108), and pGex -(48-75), which encode fragments of Pit-1, Pit-1, and the -domain, respectively, fused to GST were constructed as follows. All mutant -domain constructs were constructed by polymerase chain reaction amplification of selected subregions of Pit-1 or Pit-1. Amplified DNA was initially subcloned into pCR 2.1 (Invitrogen). Commercially synthesized deoxyoligonucleotides (Life Technologies, Inc.) contained the following sequences each incorporating a NotI restriction site to facilitate subcloning: 5'Pit-1-(199-291), GCG GCC GCC AGG TCG GAG CTT TGT ACA AT; 3'Pit-1-(199-291), GCG GCC GCT TAT CTG CAC TCA AGA TGC TC; 5'Pit-1-(2-82), GAG CGG CCG CAG TTG CCA ACC TTT CAC CTC G; 3'Pit-1-(2-82), GCG GCC GCT CAT GGA AAC TTG TAA AGA CAA G; 5'Pit-1-(2-108), GAG CGG CCG CAG TTG CCA ACC TTT CAC CTC G; 3'Pit-1-(2-1-108), GCG GCC GCT CAT GGA AAC TTG TAA AGA CAA G; 5', GAT CCG CTG TCC CGT CTA TTT TGT CTT TGA TCC AAA CTC CTA AAT GTT TGC ACA CAT ATT TCT CGA TGA CAA CGA TGG GAA ATA CAG CTA; and 3', GAT CTA GCT GTA TTT CCC ATC GTT GTC ATC GAG AAA TAT GTG TGC AAA CAT TTA GGA GTT TGG ATC AAA GAC AAA ATA GAC GGG ACA GCG.

The presence of each introduced mutation and integrity of the Pit-1/Pit-1 subregions were verified by Sanger sequencing using the University of Colorado Health Sciences Cancer Center DNA Sequencing Core facility. The amplified subregions Pit-1-(199-291), Pit-1-(2-82), and Pit-1-(2-108) were excised from pCR2.1 by digestion with NotI and ligated either to the unique NotI site of pCGN2 or to the unique NotI site of pGexDFGK. Plasmids were re-sequenced, and those with correctly oriented inserts were retained as pCGN2 Pit-1-(199-291), pCGN2 Pit-1-(2-82), and pCGN2 Pit-1-(2-108) and also pGex Pit-1-(199-291), pGex Pit-1-(2-82), pGex Pit-1-(2-108), and pGex -(48-75). The amplified -domain (aa 48-75) was cloned into the unique BamHI site of pGEX 2TK (Amersham Pharmacia Biotech). -Domain fusion constructs were re-sequenced, and those with properly oriented inserts were retained as pGex -(48-75). A fortuitous mutant construct that introduced a stop codon after amino acid 62 was also retained as pGex -(48-62).

Cell Culture-- HeLa cells were maintained in Dulbecco's modified Eagle medium (DMEM, Life Technologies, Inc.) supplemented with 15% horse serum and 2.5% fetal calf serum (Life Technologies, Inc.). Cells were grown at 37 °C in 5% CO₂. Medium was changed 4-12 h prior to transfection, and cells were harvested at 50-70% confluency.

Electroporation-- Cells were harvested in 0.05% trypsin and 0.5 mM EDTA and resuspended in DMEM supplemented with 15% horse serum and 2.5% fetal calf serum. Aliquots of approximately 2-4 × 10⁶ cells in 200 µl of medium were added to plasmid DNA and transfected by electroporation (25) at 220 V and 500 microfarads using a Bio-Rad Gene Pulser with 4-mm gap cuvettes. All transfections included 0.3 µg pCMV (CLONTECH) as an internal control for transfection efficiency. Total DNA was kept constant, and nonspecific effects of viral promoters were controlled by using the appropriate empty vectors. Following transfection, cells were plated in DMEM with 15% horse and 2.5% fetal calf serum and incubated for 24 h. Electroporations were performed in triplicate for each condition within a single experiment, and experiments were repeated several times using different plasmid preparations of each construct.

Luciferase and -Galactosidase Assays-- Transfected cells were harvested in PBS (16 mM Na₂HPO₄, 4 mM NaH₂PO₄, 150 mM NaCl) containing 3 mM EDTA, and extracts were prepared by three sequential freeze-thaw cycles in 100 mM potassium phosphate, 1 mM dithiothreitol, pH 7.8. Cell lysis was increased by vortexing between cycles. Cell debris was pelleted by centrifugation at 10,000 × g for 10 min at 4 °C, and aliquots of the supernatant were used in subsequent assays. Luciferase was assayed as described previously (20). Samples were measured in duplicate using a Monolight 2010 Luminometer (Analytical Luminescence Laboratories, San Diego CA). -Galactosidase activity was determined spectrophotometrically using the chromogenic substrate o-nitrophenyl--D-galactopyranoside as described (20). Total luciferase light units were normalized to total -galactosidase activity. The normalized relative luciferase activity for each control was set to 1, and results were expressed as fold rPRL promoter activation.

Western Blotting-- Cell extracts were prepared from confluent 60-mm dishes. Cells were washed in cold PBS and harvested with Laemmli SDS sample buffer. Extracts (100 µg) were boiled for 5 min, and viscosity was reduced by shearing through a 22-gauge needle. Samples were resolved on 12% SDS-polyacrylamide gels and transferred to nitrocellulose in 192 mM glycine, 25 mM Tris, 10% methanol, at 100 mA for 16 h. Filters were blocked in 5% non-fat milk, 0.2% Tween 20, probed with monoclonal antibodies to HA (Babco, Berkeley, CA) or GST (Santa Cruz Biotechnology, Santa Cruz, CA), and developed using ECL (Amersham Pharmacia Biotech)) according to the manufacturer's directions.

GST Fusion Proteins-- Recombinant fusion proteins GST-Pit-1 and the GST-Pit-1 internal deletions and truncations were prepared from bacterial extracts (9, 26). Overnight cultures of Escherichia coli BL-21(DE3)pLysS (Stratagene, La Jolla, CA), transformed with plasmid pGexDFGKrPit-1 or the internal deletions and truncations, were diluted 1:10 in fresh Luria broth supplemented with ampicillin (50 µg/ml) and grown at 30 °C. Upon attaining an absorbance at 600 nm of 0.5 to 0.8, cultures were induced by addition of isopropyl--D-thiogalactopyranoside to a final concentration of 1 mM. Growth was continued for another 2 h at 30 °C. Bacterial cells were harvested by centrifugation at 5,000 × g for 5 min at 4 °C and resuspended in 1/10 volume of PBS containing 1% Triton X-100 and the recommended concentration of Complete^TM protease inhibitor mixture (Roche Molecular Biochemicals). Cells were lysed by sonication on ice for 10 s using a cell disruptor microprobe (Heat Systems-Ultrasonics, Plainville, NY) on maximum setting. The cells were then placed on a rotator for 30 min at room temperature for further lysis. Cellular debris was removed by centrifugation at 10,000 × g for 10 min at 4 °C. Supernatants were bound to glutathione-Sepharose (Amersham Pharmacia Biotech) for 1 h at 4 °C and washed extensively in PBS supplemented with Complete^TM protease inhibitors. Bound protein was analyzed by SDS-polyacrylamide gel electrophoresis and Coomassie Blue staining. Protein concentration was measured by the Bio-Rad assay (Bio-Rad).

In Vitro Binding Assays-- Ets-1 was synthesized and labeled with [³⁵S]methionine (NEN Life Science Products), using the TNT^TM coupled transcription-translation reticulocyte lysate system with T7 polymerase, according to the manufacturer's protocol (Promega, Madison, WI). Equal amounts (20 µg) of GST fusion proteins were bound to glutathione-agarose beads and suspended in binding buffer (40 mM HEPES, 100 mM NaCl, 5 mM MgCl₂, 0.5 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 0.05% Nonidet P-40, 1 mM dithiothreitol, pH 7.5) supplemented with Complete^TM protease inhibitors. ³⁵S-Ets-1 translated in vitro was incubated with immobilized GST, GST-Pit-1, or the GST-Pit-1 internal deletions and truncation mutants in a final volume of 0.5 mlof binding buffer containing 50 µg/ml ethidium bromide and mixed by rocking for 1 h at room temperature. Beads were collected by a rapid, 30-s centrifugation at 1,000 × g, and then washed five times for 5 min each in 0.5 ml of binding buffer containing 0.1% Triton X-100. Bound ³⁵S-Ets-1 was eluted by boiling in SDS sample buffer and analyzed by SDS-polyacrylamide gel electrophoresis and autoradiography (9). Bands were quantitated using a Molecular Dynamics Laser-scanning densitometer with Imagequant^TN software.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Functional Interaction of Ets-1 and Pit-1 Requires the Pit-1 Homeodomain-- Multiple regions of Pit-1, including the amino-terminal trans-activation domain (TAD) and bipartite DNA binding POU domain, have been implicated in physical and functional synergistic interactions with other transcription factors, including homeodomain proteins, nuclear hormone receptors, and co-activators or co-repressors (27-36). In order to map the domain of Pit-1 required for functional interaction with Ets-1, a series of internal deletions (37) were amino-terminally tagged with the hemagglutinin (HA) epitope by cloning into the pCGN2 expression vector (Fig. 1). We have previously documented that it is essential to achieve equivalent protein expression levels of Pit-1 mutants in order to evaluate their functional characteristics (38). The amino-terminal HA tag facilitates monitoring of Pit-1 construct expression by Western blotting, using a specific monoclonal antibody, and has no effect on Pit-1 transcriptional activation in pituitary or non-pituitary cell-types (Ref. 17 and data not shown).

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Structure of HA-tagged Pit-1 deletion constructs. The indicated Pit-1 deletions were cloned into pCGN2. HA, influenza hemagglutinin epitope (YPYDVPDYA); Pit-1 POU DNA binding domain-(132-273) consists of POU:POU-specific subdomain-(132-198) and homeo:POU-homeodomain-(214-273). Bold numbers indicate amino acids deleted in each construct.

We have developed a PRL promoter reconstitution protocol in HeLa cells transiently transfected with Ets-1 and Pit-1, which results in synergistic activation of rPRL gene transcription (10). This system provides a highly sensitive assay to map the domains of Pit-1 required for both basal activity and to mediate functional interaction with Ets-1. The HA-Pit-1 constructs depicted in Fig. 1 were transiently transfected into HeLa cells to determine the effects of the internal deletions on activation of the rPRL promoter (Fig. 2A). Input of plasmid DNA encoding the various HA-Pit-1 constructs was first optimized to yield equivalent levels of HA-tagged protein as determined by Western blotting of cell extracts (Fig. 2C), and these DNA amounts were scaled down proportionately to optimize transcriptional response and to avoid nonspecific squelching (Fig. 2, A and B). As reported previously (10), rPRL promoter activity in nonpituitary HeLa cells, which lack Pit-1, is very low compared with that in pituitary cells. Expression of intact wild-type Pit-1 activated the rPRL promoter approximately 250-fold over vector control (Fig. 2A). Deletions within the putative amino-terminal TAD of Pit-1-(2-40 and 2-80) revealed a progressive reduction in rPRL promoter trans-activation, exhibiting 60 and 40% activity of wild-type Pit-1, respectively, when expressed at comparable levels (Fig. 2A and Table I). Deletion of the "hinge region" (72-125), between the TAD and POU-specific domain, or of sequences between the POU-specific and homeodomains (200-211), did not significantly affect Pit-1 trans-activation of the rPRL promoter. The Pit-1 construct lacking the POU-specific region of the DNA binding domain (124-201) exhibited reduced, but substantial, transcriptional activity (i.e. 29% of wild-type Pit-1). In contrast, deletions within the homeodomain essentially abrogated Pit-1 trans-activation of the rPRL promoter, with 209-252 and 255-291 resulting in less than 1% of wild-type Pit-1 activity (Fig. 2A and Table I). These results are in general accord with the reported activity of these internal Pit-1 deletions on the rat growth hormone promoter (37).

View larger version (35K): [in this window] [in a new window]   Fig. 2.   Mapping of the functional domains of Pit-1 required for synergistic interaction with Ets-1. HeLa cells were co-transfected with 3 µg of pA3rPRLluc reporter and pCGN2 (50 ng), pCGN2rPit-1wt (50 ng), pCGN22-40 (12.5 ng), pCGN22-80 (12.5 ng), pCGN272-125 (12.5 ng), pCGN2124-201 (12.5 ng), pCGN2200-211 (25 ng), pCGN2209-252 (37.5 ng), or pCGN2255-291 (12.5 ng), as indicated, in the absence (A) or presence (B) of 5 µg of pSG5 cEts-1. Cells were harvested after 24 h and assayed for luciferase and -galactosidase as described under "Experimental Procedures." Results are expressed as fold activation relative to basal promoter activity and are the means ± S.E. of five experiments, each consisting of triplicate transfections. C, analysis of the Pit-1 internal deletions expression by Western blotting. HeLa cells were transfected with 4 µg of pCGN2, 4 µg of pCGN2rPit-1wt, 1 µg of pCGN22-40, 1 µg of pCGN22-80, 1 µg of pCGN272-125, 1 µg of pCGN2124-201, 2 µg of pCGN2200-211, 3 µg of pCGN2209-252, and 1 µg of pCGN2255-291 as indicated. Cells were harvested as described under "Experimental Procedures" and 100 µg extract analyzed by SDS-gel electrophoresis and probed with anti-HA antibody (1:1000) and developed by ECL (Amersham Pharmacia Biotech) according to the manufacturer's protocol.

We next examined the effect of the Pit-1 internal deletions on their functional interaction with Ets-1 (Fig. 2B). Transfection of Ets-1 alone activated the rPRL promoter 100-fold. However, as described previously (10), in the presence of both Pit-1 and Ets-1, rPRL promoter activity is synergistically enhanced to 2100-fold. Co-transfection of Ets-1 had no effect on expression levels of wild-type or mutant Pit-1 proteins, and Ets-1 expression was not altered in the presence of Pit-1 constructs (data not shown). As shown in Fig. 2B, significant functional synergism between Pit-1 and Ets-1 is also revealed by Pit-1 mutants bearing deletions within the TAD (2-40 and 2-80), the POU-specific domain (124-201), and in the linker region between the POU-specific and homeodomains (200-211), resulting in 840-, 1140-, 1170-, and 1195-fold activations, respectively. Deletion of the region between the TAD and the POU-specific domain, 72-125, resulted in synergistic transcriptional activation similar to that of wild-type Pit-1, when normalized for protein expression. In contrast, Pit-1 homeodomain deletions, 209-252 or 255-291, failed to synergize with Ets-1 and instead reduced the Ets-1 activation of the rPRL promoter from 100- to 48- and 31-fold, respectively (Fig. 2B). This inhibition of Ets-1 trans-activation of the rPRL promoter, by Pit-1 homeodomain mutations, was not simply a numerical average but rather was consistently observed in each of the 15 transfections compiled to generate Fig. 2B.

To evaluate and quantitate more accurately the effects of Pit-1 internal deletions on the Pit-1/Ets-1 synergistic response, we calculated a true synergy fold, as previously reported (9, 10, 39). The Pit-1/Ets-1 synergy fold is defined as rPRL promoter activity in the presence of a combination of Ets-1 plus Pit-1 construct, divided by the sum of the individual fold activations induced by Ets-1 alone and each Pit-1 construct alone (10). Thus, co-transfection of Ets-1 and wild-type Pit-1 resulted in levels of rPRL promoter activity approximately 6-fold greater than that predicted based on the sum of their individual responses (Table I). By contrast, a synergy fold value of 1 would reveal an additive response, thereby indicating a loss of functional interaction (39). The data in Table I show that internal deletions of Pit-1 spanning the TAD (2-40 and 2-80), the hinge region (72-125), the POU-specific (124-201), and linker (residues 200-211) domains all retain synergistic activation of the rPRL promoter in combination with Ets-1, with synergy fold values ranging from 3.3- to 8.9-fold. In contrast, deletions within the homeodomain, 209-252 or 255-291, not only abrogated Pit-1/Ets-1 synergy but actually showed apparent negative cooperativity, reflected by fractional (<1) synergy fold values of 0.47 and 0.3, respectively. Thus, the Pit-1 homeodomain is necessary for Ets-1/Pit-1 synergistic activation of the rPRL promoter, and deletions within it may confer a dominant-negative phenotype with respect to functional interaction with Ets-1.

Functional Interaction of Pit-1 and Ets-1 Is Independent of Pit-1 Phosphorylation-- Given that the Pit-1 homeodomain is required for the Pit-1/Ets-1 synergy, we focused on residues of this domain that may play a role in binding to Ets-1. Serine 115, threonine 219, and threonine 220 are three principal Pit-1 phosphorylation sites, with the latter two located at the amino terminus of the homeodomain, targeted by protein kinase A, protein kinase C, and during the cell cycle (24, 40, 41). Although phosphorylation of these sites is not required to mediate Pit-1-dependent hormone or growth factor activation of the rPRL promoter (13, 23, 36, 40), phosphorylation of Pit-1 has been reported to modulate binding to DNA (24, 40, 41). Furthermore, Pit-1 regulation of the c-fos promoter at the serum response element is dependent on phosphorylation of Pit-1 at these sites, suggesting that phosphorylation of the Pit-1 homeodomain may modulate formation of a ternary complex with serum response factor (42). Thus, to determine the role of Pit-1 phosphorylation in the functional interaction with Ets-1, HeLa cells were transiently transfected with either wild-type Pit-1 or mutant Pit-1(A3), in which the three principal phosphorylation sites are substituted by alanine (23, 24). Fig. 3A shows that Ets-1 alone resulted in 190-fold activation, Pit-1 wild-type in 430-fold, and both together resulted in a 3245-fold response (or a 5.2-fold synergy) of the rPRL promoter. Pit-1-(A3) resulted in a 310-fold activation of the rPRL promoter (Fig. 3A), showing that, consistent with previous reports (23, 40), mutation of the phosphorylation sites did not significantly affect basal trans-activation by Pit-1. In the presence of Ets-1, Pit-1-(A3) exhibited a synergistic 3460-fold activation of the rPRL promoter (or a 6.9-fold synergy), comparable to that of wild-type Pit-1. Fig. 3B shows that both forms of Pit-1 protein were expressed at equivalent levels and that expression of Pit-1 is not altered by co-transfection of Ets-1. Hence, neither phosphorylation nor the specific identities of serine 115 or the homeodomain amino acids threonine 219 and 220 of Pit-1 are required for functional interaction with Ets-1.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Functional interaction of Pit-1 phosphorylation site mutants with Ets-1. A, HeLa cells were transiently transfected with 3 µg of pA3PRLluc, 0.3 µg of pCMV,and 5 µg of pRc.935Pit1wt FLAG or pRc.935Pit1A3 FLAG mutant ± 5 µg of pSG5c-Ets-1 as indicated. PRL promoter activity was determined as in Fig. 2. Results are expressed as fold activation over control and represent the mean ± S.D. of 9 transfections. B, expression of Pit1wt or A3 mutant in transfected HeLa cells detected by Western blotting with anti-FLAG M2 antibody. Lanes 1-5 show extracts of cells transfected as in A.

Physical Interaction of Pit-1 and Ets-1-- The data from the above reconstitution system demonstrate a requirement for the Pit-1 homeodomain, spanning amino acids 209-291, for functional interaction with Ets-1, in order to mediate a synergistic activation of the rPRL promoter. Although these data are consistent with a direct interaction of the Pit-1 homeodomain with Ets-1, mutations in the homeodomain also render Pit-1 unable to bind DNA and/or enter the nucleus (37, 43). Thus, despite the inhibition of Ets-1 trans-activation by co-expression of 209-252 and 255-299, which suggests a functional interaction, the lack of synergy with Ets-1 may simply be due to the inability of Pit-1 homeodomain mutants to activate gene transcription. Our previous results have shown that Ets-1 is able to bind directly to intact Pit-1 or Pit-1, independent of the binding of either factor to their cognate DNA elements (10). To determine whether the Pit-1 homeodomain mediates a direct physical interaction with Ets-1, fusion proteins containing GST linked to either wild-type Pit-1, 209-255, or 255-291 were constructed (Fig. 4A), immobilized on glutathione-agarose, and used in protein-protein interaction assays with ³⁵S-labeled Ets-1 as described under "Experimental Procedures." Equivalent amounts of beads and bound Pit-1 fusion proteins, based upon protein determination and staining of samples resolved by SDS-polyacrylamide gel electrophoresis, were used, and incubations were carried out in the presence of ethidium bromide to block nonspecific protein-DNA interactions (44). Expression levels of intact fusion proteins were monitored by Western blotting with anti-GST antibodies (Fig. 4C). As shown in Fig. 4B, ³⁵S-Ets-1 binds specifically to immobilized intact GST-Pit-1 but not to GST alone. Comparison of bound ³⁵S-Ets-1 with the input of ³⁵S-Ets-1 indicates a significant portion (21%) is bound under these dilute solution conditions. GST-Pit-1 fusions with deletions within the homeodomain retain the ability to bind to ³⁵S-Ets-1, but the amount of ³⁵S-Ets-1 bound is considerably reduced (Fig. 4B). Densitometric scanning of the autoradiograph revealed that, relative to intact Pit-1 (100%), ³⁵S-Ets-1 binding to the 209-252 and 255-291 homeodomain mutants is reduced to 48 and 42%, respectively (Fig. 4B, lanes 4 and 5). This is consistent with the inhibition of Ets-1 transcriptional activity by these Pit-1 mutants (Table I) and suggests that deletions within the homeodomain of Pit-1 inhibit both physical and functional interaction with Ets-1.

View larger version (34K): [in this window] [in a new window]   Fig. 4.   Binding of Ets-1 to wild-type and homeodomain mutants of Pit-1. A, structure of Pit-1 GST fusion proteins expressed in E. coli and immobilized on glutathione-agarose. B, binding assays were performed as described under "Experimental Procedures." Aliquots (20 µl packed volume) of glutathione-Sepharose beads, bound to 20 µg of GST (lane 2), 20 µg of GST Pit-1 wt (lane 3), or 20 µg of GST Pit-1 internal deletions (lanes 4 and 5), were incubated with equal amounts of in vitro transcribed and translated 35S-labeled Ets-1. Lane 1 shows 10% of the amount of methionine-labeled Ets-1 added to each reaction. Ets-1·Pit-1 complexes were resolved by SDS-polyacrylamide gel electrophoresis and analyzed by autoradiography. C, expression of GST Pit-1 internal deletion fusion proteins. Fusion proteins were prepared and purified as under "Experimental Procedures." Aliquots (20 µg) of bound fusion proteins were eluted in Laemmli sample buffer and analyzed by Western blotting using an anti-Pit (amino acids 34-56) (Babco, Richmond, CA).

The Homeodomain of Pit-1 Is Necessary and Sufficient for Binding to Ets-1-- The reduction in Ets-1 binding resulting from deletion of either the amino-terminal (209-252) or carboxyl-terminal (255-291) regions of the Pit-1 homeodomain suggested that protein-protein interactions with Ets-1 encompass this entire subdomain. To address directly the role of the Pit-1 homeodomain in binding to ³⁵S-Ets-1, GST fusions encoding selected regions of Pit-1 and Pit-1 were constructed (Fig. 5A), expressed in E. coli, and used in ³⁵S-Ets-1 binding experiments. Equal amounts of immobilized Pit-1 fusion constructs were used. As shown in Fig. 5B, a fusion protein encoding the entire Pit-1 homeodomain, aa 199-291, exhibited full ³⁵S-Ets-1 binding (40% of input; lane 5) comparable to that of intact Pit-1 or Pit1 (each 40% of input; lanes 3 and 4). In contrast, a construct encoding the TAD of Pit-1, aa 2-82, exhibited only nonspecific binding to ³⁵S-Ets-1 (lane 6), since the level of interaction was equal to the background binding to GST alone (lane 2). Conversely, a GST fusion encoding the TAD of Pit-1, which contains the 26-amino acid -domain insert, showed significant binding to ³⁵S-Ets-1 (14% of input; lane 7), suggesting an additional Ets-1 interaction site in this alternatively spliced isoform. Finally, a GST construct encoding only the 26-amino acid -domain, -(48-75), bound to ³⁵S-Ets-1 at levels equal to that of the intact -TAD (14% of input; lane 8), confirming the presence of a second Ets-1-binding site and localizing it to within this 26-amino acid Pit-1-specific sequence. Truncation of the carboxyl-terminal half of the -domain, -(48-62), abolished specific binding to ³⁵S-Ets-1, implying that this region is critical for the interaction. Thus, the Pit-1 homeodomain (aa 199-291), common to Pit-1 and Pit-1, is both necessary and sufficient for direct DNA-independent binding to Ets-1. In addition, Pit-1 exhibits isoform-specific Ets-1 interactions via a second binding site located within the -domain of the TAD. Pit-1 is able to bind to ³⁵S-Ets-1 with an affinity equal to (Fig. 5B), if not greater than, that of Pit-1 (10). However, in contrast to Pit-1, Pit-1 fails to synergize with Ets-1 in HeLa cells and inhibits basal rPRL promoter activity in pituitary GH4 lactotrophs (10). Overexpression of Pit-1 in GH4 cells also antagonizes Ras activation of the rPRL promoter (9, 11). Pit-1 and Pit-1 are identical, except for the insertion of the -domain within the TAD. Thus, the data presented here suggest that the unique transcriptional properties of Pit-1 (10, 17) may be conferred, in part, by the presence of the secondary Ets-1-binding site in the -TAD, which functionally sequesters Ets-1 in an inhibitory conformation.

View larger version (44K): [in this window] [in a new window]   Fig. 5.   Analysis of Ets-1 binding to domains of Pit-1 and Pit-1. A, structure of GST-Pit-1 and GST-Pit-1 domain fusion proteins. B, aliquots, 20 µg, of the indicated Pit-1 or Pit-1 GST fusion constructs were bound to of glutathione-Sepharose beads (20 µl packed volume) and incubated with equal amounts of in vitro-transcribed and translated 35S-labeled Ets-1. Bound Ets-1 was assayed as in Fig. 3. Lane 1 indicates 10% of input of labeled Ets-1.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Combinatorial interactions of Pit-1 with other transcription factors play a central role in the establishment of somatotroph, lactotroph, and thyrotroph patterns of gene expression and in the regulation of GH, PRL, and TSH promoters by hormones and growth factors (27, 28, 30-34, 36). Members of the Ets family of transcription factors are key components of the regulation of rPRL gene expression that mediate both positive and negative transcriptional responses to hormones and growth factors acting via multiple signaling pathways (9-11, 19, 45-48). Interaction of Ets-1 and Pit-1 at a composite DNA-binding site provides a molecular mechanism to target the Ras signaling pathway to the rPRL promoter (9, 11, 49) and may account for the differential basal and hormone/growth factor-induced transcriptional regulation of the related rGH gene (10). In this study we have shown that binding of Ets-1 to Pit-1 and the consequent synergistic transcriptional activation of the rPRL promoter is mediated via the homeodomain of Pit-1. In contrast, the Pit-1 isoform-specific -domain provides a secondary Ets-1 interaction surface, resulting in formation of an inhibitory Pit-1·Ets-1 complex.

The POU domain of Pit-1 constitutes the bipartite DNA-binding motif comprised of a POU-specific domain and a POU homeodomain, both elements being required for high affinity, specific binding to Pit-1 DNA-binding sites (43, 50, 51). However, we have shown that that binding of Ets-1 to Pit-1 is independent of DNA, occurring in dilute solution and in the presence of ethidium bromide (Figs. 4 and 5). Internal deletion mutations within the Pit-1 homeodomain significantly reduce binding to Ets-1 but do not eliminate it (Fig. 4B). In contrast, these same deletions not only abrogate Pit-1/Ets-1 synergistic activation of the rPRL promoter but also inhibit activation by Ets-1 alone (Fig. 2B and Table I). Moreover, the region spanning the homeodomain alone, aa 209-291, binds Ets-1 as efficiently as does intact, wild-type Pit-1, indicating that the homeodomain is sufficient to bind Ets-1. This is consistent with the absolute requirement of the Pit-1 homeodomain for Ets-1·Pit-1 transcriptional synergy. These results imply that there are extensive protein-protein contacts involving both amino- and carboxyl-terminal portions of the homeodomain that contribute to physical and functional interactions with Ets-1.

The 26-amino acid -domain within the Pit-1 TAD endows the Pit-1 isoform with unique functional properties (9, 10, 15-17, 38, 52). Specifically, Pit-1 antagonizes Ets-1-dependent regulation of rPRL gene transcription, inhibiting basal rPRL promoter activity and blocking its activation by Ras in GH4 pituitary cells (9, 10, 17). We have previously shown that the unique transcriptional properties of Pit-1 versus Pit-1 on the rPRL promoter are due to the specific amino acid sequence of the -domain, rather than its function as a spacer (17). Here we show that whereas the common homeodomain provides a critical Ets-1 interaction surface, the -domain constitutes an additional Ets-1 binding region (Fig. 5B). This isoform-specific -domain interaction surface may affect either Ets-1 functions (see below) and/or the activity of co-repressors/co-activators. Indeed, we have recently shown that specific hydrophobic amino acids within the -domain modulate the activity of a histone deacetylase-containing co-repressor.² Thus, Pit-1 and Pit-1 are likely to form functionally distinct transcriptional complexes with Ets-1.

Structure-function analysis of many Ets proteins has revealed that DNA binding and trans-activation are auto-inhibited to various degrees (53). De-repression of these Ets auto-inhibitory functions is regulated by selective protein partnerships, which can be further modulated by signal-mediated phosphorylation (49, 53-56). The data presented here suggest that direct protein-protein interactions with the Pit-1 homeodomain, independent of DNA binding, alter the conformation of and activate Ets-1. Conversely, binding of the alternatively spliced Pit-1 isoform to Ets-1 may sequester the latter in an inactive or inhibitory complex. Moreover, the Pit-1·Ets-1 complex is refractory to MAPK-mediated activation of Ets-1 transcriptional potency (9, 17, 57), suggesting that interaction with the -domain renders the MAPK phosphorylation site within the Ets-1 TAD inaccessible.

Functional and physical interactions of Pit-1 with other transcription factors have been described in the regulation of several pituitary-specific genes, and protein-protein interactions have been mapped to distinct regions of Pit-1. Factors that interact with the Pit-1 TAD and hinge regions have included P-OTX, TR, and ER (Fig. 6). Pituitary OTX-related factor, P-OTX, was isolated using a yeast two-hybrid assay based on its interaction with the amino terminus of Pit-1 (aa 1-128) (33). Although intact Pit-1 binds directly to ER and TR, the precise domain(s) of Pit-1 required for these physical interactions have not been identified (58, 59). The Pit-1 region required for functional interaction with ER and TR has been mapped to Pit-1 residues 48-100, including a tyrosine-dependent synergy domain (aa 48-72) located within the TAD (28, 29, 32). Furthermore, disruption of this region by insertion of the -domain at residue 48 of the Pit-1 TAD interferes with the Pit-1/TR synergy (52).

View larger version (14K): [in this window] [in a new window]   Fig. 6.   Functional and physical interaction domains of Pit-1 and Pit-1. The diagram depicts the principal structure/function domains of Pit-1 and Pit-1. Numbering refers to Pit-1. POU, POU-specific domain; Homeo, POU homeodomain. The solid bold lines denote the regions of Pit-1/Pit-1 implicated in synergistic interactions with and binding to the indicated transcription factors. Broken line indicates the secondary Ets-1-binding site. Retinoid X receptor, RAR, and PPARR also interact physically with Pit-1, but their respective binding sites have not been defined.

The majority of Pit-1-interacting partners (Pit-1/Pit-1, Oct-1, P-Lim, VDR, CBP, nuclear receptor co-repressor, GATA-2, and Ets-1) binds to the DNA-binding region of Pit-1, involving either the POU-specific domain and/or the POU homeodomain (Fig. 6). Crystal structure of the bipartite Pit-1 DNA binding domain dimer revealed critical and direct contacts between the carboxyl terminus of the homeodomain of one partner and the amino terminus of the POU-specific domain of the other partner (51). Similarly, a pituitary LIM-homeodomain protein, P-Lim, interacts with Pit-1 via both the POU-specific and POU homeodomains (27). The POU-specific and POU homeodomains are also required for interactions with NCoR co-repressor and CBP co-activator complexes (34, 36). Proteins that interact with the Pit-1 POU homeodomain alone include Oct-1 (35), vitamin D receptor (VDR) (30), GATA-2 (22, 31), and Ets-1 (this report). Finally, synergistic interactions of Pit-1 with Zn-15 (60), retinoid receptors (RXR/RAR) (59), CCAAT/enhancer binding protein- (61), the neuronal-specific zinc finger protein, NZF-1 (62), and peroxisome proliferator-activated receptor (PPAR) have also been described, but the domains of Pit-1 required have not been identified.

The role of Pit-1 phosphorylation in regulating functional interactions is not clear. Although, Pit-1 phosphorylation is required for Pit-1 induction of the c-fos gene (42), the weight of current evidence indicates that phosphorylation of the Pit-1 homeodomain does not modulate protein-protein interactions or transcriptional synergy with other factors involved in the regulation of pituitary-specific gene expression (Fig. 3) (36, 61, 63). However, cAMP- or growth factor-induced phosphorylation of a CPB/p300 co-activator complex appears to modulate the balance of interactions of Pit-1 with either CBP co-activator or nuclear receptor co-repressor complexes (36). Interestingly, Ets-1 and Pit-1 bind to identical regions of CBP (34, 64), and formation of a transcriptional complex containing Ets-1, Pit-1, and CBP has been postulated (34). It is tempting to speculate that Ras/MAPK-mediated phosphorylation of Ets-1 (57, 65) alters the stability, composition, and/or activity of such multicomponent complexes, shifting the balance toward activation, whereas complexes containing the -domain may promote interaction with histone deacetylase-containing co-repressors. Thus, the unique transcriptional properties of Pit-1 and Pit-1 may be attributable to isoform-specific interactions or recruitment of cofactors by the -domain resulting in the formation of functionally distinct complexes.

	ACKNOWLEDGEMENTS

We thank Dr. Fred Schaufele for the Pit-1 phosphorylation site mutants, Dr. Bohdan Wasylyk for the Ets-1 construct, Dr. D. F. Gordon for pCGN2 and pGexDFGK vectors, and Dr. Michael Karin for internal Pit-1 deletion mutants. We also acknowledge D. F. Gordon, W. M. Wood, J. J. Tentler, T. A. Jackson, and R. E. Schweppe for critical reading and discussions of this manuscript. This work utilized the Tissue Culture and DNA Sequencing Core Facilities of the University of Colorado Cancer Center, supported by NCI Grant P30 CA46934 from the National Institutes of Health.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grants DK 46868 (to A. G.-H.) and DK 53496 (to A. P. B.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^¶ To whom correspondence should be addressed: Dept. of Obstetrics and Gynecology, University of Colorado Health Sciences Center, 4200 East Ninth Ave., B-198, Denver, CO 80262. Tel.: 303-315-4146; Fax: 303-315-8889; E-mail: Andy.Bradford@uchsc.edu.

² S. E. Diamond and A. Gutierrez-Hartmann, submitted for publication.

	ABBREVIATIONS

The abbreviations used are: GH, growth hormone; r, rat; PRL, prolactin; TSH, thyrotropin ; HA, hemagglutinin; DMEM, Dulbecco's modified Eagle's medium; MAPK, mitogen activated protein kinase; ER, estrogen receptor; TR, thyroid hormone receptor, VDR, vitamin D receptor; CBP, CREB-binding protein; PPAR, peroxisome proliferator-activated receptor; RAR, retinoic acid receptor; TAD, trans-activation domain; aa, amino acid; GST, glutathione S-transferase; PBS, phosphate-buffered saline.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES