Expression Cloning and Characterization of PREB (Prolactin Regulatory Element Binding), a Novel WD Motif DNA-Binding Protein with a Capacity to Regulate Prolactin Promoter Activity

Abstract

Previous studies have implied that a transcription factor(s) other than Pit-1 is involved in homeostatic regulation of PRL promoter activity via Pit-1-binding elements. One such element, 1P, was employed to clone from a rat pituitary cDNA expression library a novel 417-amino acid WD protein, designated PREB (PRL regulatory element binding) protein. PREB contains two PQ-rich potential transactivation domains, but no apparent DNA-binding motif, and exhibits sequence-specific binding to site 1P, to a site nonidentical to that for Pit-1. The PREB gene (or a related gene) is conserved, as an apparently single copy, in rat, human, fly, and yeast. A single approximately 1.9-kb PREB transcript accumulates in GH₃ rat pituitary cells, to levels similar to Pit-1 mRNA. PREB transcripts were detected in all human tissues examined, but the observation of tissue-specific multiple transcript patterns suggests the possibility of tissue-specific alternative splicing. RT-PCR analysis of human brain tumor RNA samples suggested region-specific expression of PREB transcripts in brain. Western and immunocytochemical analysis implied that PREB accumulates specifically in GH₃ cell nuclei. Transient transfection employing PREB-negative C6 rat glial cells showed that PREB is as active as, and additive with, Pit-1 in transactivation of a PRL promoter construct; and that PREB, but not Pit-1, can mediate transcriptional activation by protein kinase A (PKA). Expression in GH₃ cells of a GAL4-PREB fusion protein both strongly transactivated a 5XGAL indicator construct and yielded a further stimulation of expression of this construct by coexpressed PKA, implying that PREB can mediate both basal and PKA-stimulated transcriptional responses in pituitary cells. These observations imply that PREB will prove to play a significant transcriptional regulatory role, both in the pituitary and in other organs in which transcripts of its gene are expressed.

INTRODUCTION

The PRL gene is expressed in a cell type-specific fashion in pituitary lactotropic cells. The pituitary-specific transcription factor Pit-1 has been show to play an important role in expression by the pituitary of the PRL gene (and also of the genes for GH and TSH), both during development (1) and in the mature organism (2). The PRL promoter contains multiple binding sites for Pit-1, which have been implicated not only in basal PRL gene expression (3), but also in kinase-mediated hormonal regulation of gene expression of this gene. For example, the most proximal of these Pit-1 binding sites, spanning positions −66 to −32 and termed 1P, has been reported to direct a response to a number of external agents, including TRH and Ca²⁺ (4), cAMP (2, 5), and constitutively active forms of G-α_s (6) and G-α_q (7) and of two negatively acting G proteins, α_i and α_o (8). In addition, a more distal site, termed 3P, has been reported to direct a response to cAMP (9), while both site 3P (9) and the more distal site 4P (10) have been implicated in the Ras responsiveness of the PRL promoter.

The observation that treatment of GH₃ rat pituitary cells with activators of either protein kinase A or protein kinase C stimulated Pit-1 phosphorylation (11) suggested that Pit-1 phosphorylation might mediate many of the hormonal responses that localize to PRL promoter Pit-1 binding sites. However, reports that a number of these responses do not require Pit-1 phosphorylation (2, 12, 13) implied that factors other than Pit-1 are required for regulation of PRL gene expression directed by binding sites for Pit-1. Indeed, it has been known for some time that two ubiquitously expressed factors can bind to site 1P and activate expression via this site: Oct-1 (14) and thyrotroph embryonic factor (TEF) (15). However, since protein kinase A (PKA)-mediated phosphorylation of Oct-1 strongly decreases its binding, either to an octamer consensus site (16) or to site 1P (R. Ashton and C. Bancroft, unpublished observations), this protein is unlikely to direct kinase-mediated transcriptional activation. TEF mRNA levels are low in GH cells, and this protein is believed to exert its major action in thyrotroph development and regulation of TSHβ expression in this pituitary cell type (15). More recently, a possibly ubiquitous factor has been implicated in a Pit-1-independent transcriptional action of PKA that is mediated by PRL promoter site 1P in nonpituitary cells (5), but this factor has apparently not yet been further characterized. Finally, the studies described above of Ras responsiveness of the PRL promoter via Pit-1 binding sites (9, 10) have provided evidence that this response involves binding of an Ets-1-like factor within or adjacent to such a site.

To search for factors that might play a role in regulation of PRL gene expression via the proximal Pit-1 binding site 1P, we employed this DNA element as a probe to screen a bacteriophage cDNA expression library prepared from the GC rat pituitary cell line. We report here the identification and characterization of PREB (PRL regulatory element binding), a WD motif transcription factor. PREB binds specifically to PRL promoter site 1P but contains no discernable DNA-binding motif in its sequence; exhibits a nuclear accumulation in pituitary cells; and can both activate PRL promoter activity, alone or in the presence of Pit-1, and transmit the transcriptional action of PKA to this promoter.

We also report evidence that PREB is encoded by a single-copy gene in human and rat, while single-copy PREB gene homologs are also detected in Drosophila and yeast; and that PREB transcripts accumulate in multiple tissues in human, exhibiting patterns consistent with tissue-specific splicing. We propose that PREB represents a novel type of evolutionarily conserved transcription factor that may contribute to regulation of PRL gene expression in the mammalian pituitary, and that this protein and/or its isomers or analogs may also play a more widespread transcriptional role in various tissues and organisms.

RESULTS

Expression Cloning of PREB

We carried out Southwestern screening, in which a filter-bound, denatured-renatured expression library is screened with a concatenated probe, as described (17, 18). [³²P]-labeled PRL promoter element 1P DNA was employed to screen a GC rat pituitary cell (19)λ ZAPII cDNA library (Stratagene, La Jolla, CA). Screening of 1.5× 10⁶ plaques yielded several presumptive positive clones that bound to labeled site 1P. However, only one of these clones exhibited DNA-sequence specificity: binding to element 1P, but not to either a mutated element 1P or a canonical cAMP-response element (CRE) (data not shown). The properties of this clone, designated PREB (PRL regulatory element binding) protein, was further investigated. The pBLUESCRIPT vector containing PREB cDNA was excised in vivo, and the cloned cDNA was sequenced, yielding a 1.28-kb sequence containing a large (243-amino acid) open reading frame (Fig. 1, amino acids 175–417).

Figure 1.

Structure of PREB At the top is a schematic structure of the protein sequence predicted from the PREB cDNA clone, indicating the motifs identified: three WD repeats plus two PQ-rich regions. Shown below are the nucleotide sequence and the predicted amino acid sequence of PREB. Three sequences exhibiting extensive similarity to the consensus WD motif are underlined, and the P or Q residues in the two PQ-rich regions are circled. In the 3′-untranslated region, two ATTTA potential mRNA destabilizing elements and a potential polyadenylation signal are indicated by, respectively, underlining and capital letters. The PREB cDNA sequence has been deposited in the GenBank database under accession no. AF061817.

PREB Exhibits DNA Sequence-Specific Binding to PRL Promoter Element 1P

Electrophoretic mobility shift was employed to examine the DNA sequence- binding specificity of recombinant PREB(175–417) (Fig. 2A). PREB(175–417) plus the element 1P probe that was employed to clone PREB(175–417) yielded a major shifted band, plus weaker lower mol wt bands that probably correspond to partially degraded PREB(175–417) (lane 2), all of which were competed by excess cold element 1P (lanes 3–5). Use of element 1P mutants (illustrated in Fig. 2B) showed that PREB(175–417) binding to element 1P was also inhibited by an excess of the 1P^* mutant (lanes 9–11), but not of the ^*1P mutant (lanes 6–8). The inability of ^*1P to compete implies that the PREB binding site in element 1P is located at least partially on the 5′-side of this element (Fig. 2B). Since x-ray analysis has recently shown that the Pit-1 POU domain binds to element 1P as a dimer by contacting the bases bracketed in Fig. 2B (20), the ability of 1P^* to compete PREB binding implies that Pit-1 and PREB possess different (but possibly overlapping) binding sites within element 1P. The observation (Fig. 2A) that PREB(175- 417) binding was not competed by excess amounts of oligonucleotides corresponding either to other Pit-1 binding sites, PRL-3P (lanes 12–14) or GH-GHF1 (lanes 15–17), or to two unrelated DNA sequences, an SP1-binding site (lanes 18–20) or the PRL promoter CLE (CRE-like element) sequence (21) (lanes 21–23), further confirms the DNA binding specificity of PREB. Since no known DNA-binding motif can be detected in the PREB(175–417) amino acid sequence shown in Fig. 1, this specificity is likely to be conferred by a novel DNA-binding motif.

Figure 2.

The DNA Target Site for PREB A, Partially purified poly-His-tagged PREB(175–417) (∼40 ng) was incubated with a [³²]P-site 1P probe, plus or minus excess (50- to 250-fold) of the indicated competitors, and analyzed on polyacrylamide gels. Lane 1 received no added protein, and lanes 1–11 and 12–23 were analyzed on separate gels. To improve resolution, the free probe was run off the end of the gel. The single major shifted band observed with PREB(175–417) is indicated. B, Structure of site 1P and sequences mutated in competitors *1P and 1P* and region predicted to form part of the binding site for PREB. The bases shown by x-ray analysis to be contacted by Pit-1 POU domain dimers (20 ) are bracketed.

Isolation of the Full-Length PREB-Coding Sequence

Successive applications of 5′-RACE (rapid amplification of DNA ends) to GH₃ cell RNA, employing initially primers corresponding to the 5′-terminus of PREB(175–417), yielded an additional 636 bases of 5′-terminal DNA sequence, resulting in the PREB cDNA sequence shown in Fig. 1, containing in-frame methionine codons at positions 1 and 217. The ATG at base 217 is preceded by ACC, matching well the Kozak consensus (ACCAUGG), while the ATG at base 1 is preceded by the poorly matching sequence GGG. The sequence after each methionine encodes PREB containing, respectively, either 345 amino acids (∼ 38 kDa) or 417 amino acids (∼ 46 kDa). Putative polyadenylation (AATAAA) and AU-rich mRNA shortened half-life (ATTTA) elements in the cDNA 3′-UT are also noted.

Structural Motifs in PREB

Comparison of the PREB cDNA coding sequence with GenBank yielded no match with any previously cloned protein cDNA sequence, implying that PREB is a novel protein. However, portions of the PREB cDNA sequence are similar to several expressed sequence tags of unknown function posted on the Web by the Washington University-Merck EST Project, identified in human infant brain (R12741), and in various mouse tissues: AA111707 (thymus), AA184937 (lymph node), W30204 (19.5-day whole fetus), and AA250620 (tissue not specified).

BLASTN analysis of the PREB sequence revealed three regions (underlined in Fig. 1) exhibiting extensive (≥39%) identity to the WD repeat consensus (β-transducin motif) (22) (Fig. 3), identifying PREB as a member of the WD-repeat protein family (23). The highest degree of similarity between PREB and other known proteins was observed in the PREB region containing WD repeats I and II, which exhibits the following similarity with the corresponding region in two yeast WD-repeat gene regulators: HIR1 (24) (30% identity, 50% similarity), and Tup1 (25) (34% identity, 52% similarity).

Figure 3.

Sequence Alignments of WD Repeats in PREB Amino acid positions are shown on the right. Shown above the line are the three regions in PREB that exhibit the indicated identities (indicated by shading) to the β-transducin consensus (22 ) shown below the line.

PREB also contains two amino acid stretches (residues 86–134 and 223–279) that are unusually rich in both proline and glutamine. Proline-rich and glutamine-rich transactivation motifs have been identified in transcription factors (26). In addition, protein regions rich in both proline and glutamine have been shown to mediate the transcriptional repression function of the tumor repressor WT1 (27) and to form part of the transcriptional activation domain of YY1 (28). Either or both of the proline/glutamine-rich domains may thus be responsible for the transactivational capabilities of PREB described below.

PREB Is Encoded by an Evolutionarily Conserved Single-Copy Gene That Is Expressed Both in Rat Pituitary Cells and in Multiple Human Organs

Southern blot analysis with a cDNA probe was employed to investigate the structure and number of PREB genes in various organisms (Fig. 4). Digestion of rat DNA with EcoRI and HindIII or XhoI and PvuI yielded, upon hybridization with a PREB cDNA probe and hybridization, single or double bands, respectively. The double bands detected with the latter two enzymes probably correspond to cleavage within an intron, since the PREB cDNA sequence contains neither recognition site. Single bands were detected upon hybridization of PREB cDNA to human DNA digested with any of four enzymes. Thus, PREB is apparently a single-copy gene that is well conserved between rat and human. Restriction enzyme digestion of either Drosophila or yeast DNA also yielded either one or two bands that hybridized with rat PREB cDNA, implying that the PREB gene (or a related gene) has been highly conserved during evolution.

Figure 4.

Southern Blot Analysis of the PREB Gene in Various Organisms DNA isolated from the indicated organisms was digested with the indicated restriction enzymes and subjected to Southern blot analysis, employing a PREB cDNA probe. Hybridization was at reduced stringency (39 C), except for rat DNA, where 42 C was employed. The sizes of internal DNA mol wt markers are indicated.

We next investigated PREB gene expression, beginning with examination of accumulation of PREB mRNA in the GH₃ rat pituitary cells (Fig. 5). Northern analysis of total cellular RNA yielded a single PREB mRNA band comigrating with 18S ribosomal RNA (hence ∼1.9 kb in size). GH₃ cells thus produce a single predominant molecular form of PREB mRNA, with an apparent size in good agreement with that predicted by the PREB cDNA clone (Fig. 1). Moreover, it can be seen that Northern analysis of the mRNAs for PREB and Pit-1 mRNA under equivalent conditions of RNA input, probe amounts, and specific activities yielded a PREB mRNA band and a major Pit-1 mRNA band exhibiting comparable intensities. PREB and Pit-1 mRNA thus accumulate to similar levels in the GH₃ cells, consistent with an important role for PREB in these cells in regulation of PRL promoter activity.

Figure 5.

Northern Analysis of PREB mRNA in GH₃ Cells Upper panel, RNA was isolated from GH₃ cells, and two aliquots (15 μg) of the same sample were analyzed on the same gel, transferred to nitrocellulose, and hybridized with a cDNA probe for either PREB or Pit-1. To permit comparison of the signals obtained for the two mRNAs, cDNA probes of equal size were labeled to equivalent specific activities, and equal amounts of each probe were employed for hybridization. The positions of the two ribosomal RNA bands are indicated. Lower panel, Stained ribosomal RNAs from the same samples.

The isolation from various human and mouse organs of ESTs homologous to PREB mRNA, described above, suggests that the PREB gene may be expressed in multiple tissues. To investigate this question, a human multiple tissue Northern blot was probed with a rat PREB cDNA probe (Fig. 6A). PREB transcripts were detected in all tissues examined, with the strongest signals seen in heart, skeletal muscle, and pancreas. Significantly, three different bands were detected, corresponding to mRNA sizes of approximately 2.2, 1.9, and 1.5 kb. Most tissues yielded two bands, but different tissues yielded different patterns, suggesting the possibility of tissue-specific alternative splicing of the PREB gene transcript. Thus, all tissues examined except lung yielded the 2.2-kb species (the kidney band is detectable upon prolonged exposure), while the 1.5-kb species was detected only in heart, muscle, and pancreas, and the 1.9-kb species only in brain, placenta, and lung. Finally, lung and liver yielded only, respectively, the 1.9- and 2.2-kb species. The 1.9-kb mRNA band may correspond to the 1.9-kb PREB mRNA species we have detected in the rat GH₃ pituitary cells (see Fig. 5).

Figure 6.

Analysis of PREB mRNA in Various Tissues and Cell Lines A, PREB mRNA expression in various human organs. A human multiple tissue Northern blot (Clontech 7760, 2 μg polyA+ RNA from each tissue) was probed with a PREB cDNA probe (upper panel), then stripped and reprobed with β-actin cDNA (lower panel). In the upper panel, a shorter exposure of lane 1 is shown in lane 1′. Open circles, closed circles, and arrowhead mark bands corresponding to transcripts approximately 2.2, 1.9, and 1.5 kb, respectively, in size. B, PREB mRNA expression in cell lines. Poly(A)-enriched mRNA (1 μg) prepared from either GH₃ or C6 cells was subjected to Northern blot analysis with a PREB cDNA probe. EtBr staining indicated that each lane received equivalent amounts of undegraded ribosomal RNA (not shown). The PREB band in the GH₃ lane is purposefully overexposed to determine whether the C6 cell lane contains detectable PREB mRNA, and its size is estimated from migration of the 18S rRNA band. C, PREB mRNA expression in various human brain tumors and cells. Lanes 2–7, An aliquot of cDNA (2 μl) from each of the indicated brain tumors was PCR-amplified employing gene-specific primer sets for either PREB (top) or HGPRT (bottom). Shown are the sizes of an internal 1-kb DNA Ladder (GIBCO), and an arrow indicating the major product amplified with PREB primers. Insufficient amounts of cDNA were available for HGPRT analysis of the lane 3 sample. Lane 9 (top), PCR amplification products of PREB cDNA; lanes 10 and 11 (top), RT-PCR amplification products of 100 ng poly(A)-enriched mRNA from C6 cells. Astro, Astrocytoma; gliom, glioma; br tum, brain tumor (metastasis into brain); neuroep, neuroepithelioma; RT, reverse transcriptase.

For the characterization of the regulatory properties of PREB, it was desirable to attempt to identify a PREB-negative cell line, for use both as a negative control for studies of cellular PREB expression and as a recipient for transfection analysis of PREB activity. The widespread distribution of PREB transcripts among human organs suggested that it might be difficult to identify a PREB-negative cell line. However, we have found previously that the rat C6 glial cell line exhibits virtually undetectable expression of transfected PRL promoter constructs in the absence of cotransfected transcription factors (29), suggesting that this cell line is deficient in factors capable of stimulating PRL promoter activity. Northern analysis of C6 cell polyA⁺ RNA (Fig. 6B) yielded no detectable PREB mRNA band, under conditions where GH₃ cell mRNA yielded a strong signal for this mRNA; and RT-PCR analysis of C6 cell polyA⁺ RNA (Fig. 6C) yielded a readily detectable control HGPRT signal, but no detectable PREB signal. In addition, PREB protein was undetectable in the C6 cells, both by Western blot analysis (data not shown) or by immunocytochemistry (see below). Hence, the C6 cells were employed as a PREB-negative cell line in our further studies.

The above observations that PREB transcripts accumulate in a total brain preparation but not in glial cells suggested that PREB expression might be restricted to specific brain regions. To investigate this concept, primers specific for PREB or HGPRT were employed for RT-PCR analysis of cDNA prepared from various human brain tumors (Fig. 6C). An amplified PREB signal was detected in cDNA prepared from two astrocytomas (lanes 2 and 3) and a metastasis into the brain (lane 6). By contrast, no such signal was detected in cDNA from either of two gliomas (lanes 4 and 5) or from a neuroepithelioma (lane 7), even though an HGPRT signal was detectable in both glioma samples and in the neuroepithelioma sample. These observations with human tumor samples imply that PREB transcripts accumulate in astrocytes but not in glial or neuroepithelial cells, and thus that the PREB gene exhibits region-specific expression within the brain.

PREB Can Function as a Transcriptional Activator

To begin to investigate whether PREB might serve as a pituitary cell transcription factor, we investigated the expression and intracellular location of PREB protein in the GH₃ rat pituitary cells. Nuclear or cytosolic proteins isolated from an equal number of cells were subjected to Western blot analysis (Fig. 7A). Anti-PREB antiserum, but not control preimmune serum, detected a major approximately 45-kDa band that accumulates preferentially in nuclei. This observation implies that PREB is a nuclear protein in pituitary cells. In addition, the size of the protein detected suggests both that synthesis of this protein is initiated by the more N-terminal methionine encoded by the PREB cDNA sequence (i.e. amino acid 1 in Fig. 1) and that the entire PREB cDNA coding sequence has been cloned.

Figure 7.

Size and Nuclear Accumulation of PREB in GH₃ Cells A, Nuclei and cytosol were prepared, and the nuclei extracted with 0.42 mm KCl. Nuclear and cytosolic proteins from an equal number of cells were then analyzed on a Western blot, employing anti-PREB antiserum. Sizes of marker proteins on the same gel are indicated. B, The same antiserum was employed for immunocytochemistry of GH₃ cells or C6 cells. Ab, anti-PREB antiserum; Pre, preimmune serum from the same animal.

The intracellular location of PREB was examined further by immunocytochemical analysis (Fig. 7B). With anti-PREB antiserum, the pituitary GH₃ cells yielded a strong signal that was located specifically over the nuclei, while control rat glial C6 cells yielded only a faint diffuse background signal. Preimmune serum also yielded only a background signal with GH₃ cells (Fig. 7B) or C6 cells (data not shown). The observation with both techniques that PREB cross-reacting material exhibits a substantial nuclear accumulation in pituitary cells is consistent with a role for PREB as a cellular transcription factor.

To investigate directly a possible role for PREB as a PRL gene transcription factor, we examined the ability of this protein to transactivate PRL promoter activity in rat glial C6 cells which, as described above, do not exhibit detectable PREB expression. We first examined the ability of PREB to regulate expression of a construct (−1957)PRL-CAT (chloramphenicol acetyl transferase), which contains the first 1957 bp upstream of the PRL gene body and thus covers both the promoter and enhancer regions (30) (Fig. 8). As expected from previous studies (29), this PRL-CAT construct alone was inactive in the C6 cells, but was activated by coexpression of Pit-1. Coexpression of PREB was also observed to activate (−1957)PRL-CAT expression, showing that this protein can act as a transcriptional activator. Although in the experiment illustrated, PREB was more active than Pit-1, in other similar experiments, the two proteins exhibited equivalent activation of (−1957)PRL-CAT expression (data not shown). This result demonstrates that PREB can transactivate PRL promoter/enhancer activity. This region of the PRL-regulatory region contains at least seven binding sites for Pit-1 (30). The similar levels of activity observed for Pit-1 and PREB on this construct thus suggests that the PRL promoter/enhancer region may also contain multiple functional PREB-binding sites.

Figure 8.

PREB Can Transactivate Expression Directed by the PRL Promoter Plus Enhancer C6 rat glial cells (2.6 × 10⁶) were electroporated with (−1957)PRL-CAT (10 μg) plus RSV-β gal (2 μg), plus 5 μg either RcRSV (“None”), or RSV-Pit-1 or RSV-PREB, divided into three 60-mm dishes, incubated 2 days, and then assayed for CAT andβ -galactosidase activity. Shown is the mean ± sd of CAT activity, corrected for β-galactosidase activity, observed with triplicate dishes.

Since element 1P is the only presently known PREB-binding site in the PRL promoter (Fig. 2), we concentrated on this site in our further functional studies of PREB. To study element 1P in its natural context within the PRL promoter, we investigated the ability of exogenously expressed PREB to trans-regulate construct (−113)PRL-CAT. As illustrated in Fig. 9 (top), the only known PRL promoter elements in this truncated construct are the CLE, element 1P, and a TATA box. The observation that the CLE does not bind PREB (Fig. 2) implies that any effects of PREB (and/or Pit-1) on this PRL-CAT construct are mediated via element 1P.

Figure 9.

PREB and Pit-1 Yield Equivalent Activation of Expression of a PRL Promoter Construct Top, Structure of plasmid (−113)PRL-CAT, illustrating the positions of the known PRL promoter elements. Bottom, C6 cells (2.6 × 10⁶) were electroporated with (−113)PRL-CAT (10 μg) and RSV-β gal (2 μg), plus the indicated amounts of either RSV-Pit-1 or RSV-PREB, and treated thereafter as in Fig. 8. Shown is the mean ± sd of corrected CAT activity observed with triplicate dishes.

A comparison of the abilities of various inputs of RSV-based expression vectors for Pit-1 and PREB to transregulate (−113)PRL-CAT expression (Fig. 9) yielded equivalent stimulation by the two proteins. This result again demonstrates that PREB can act as a transcriptional activator of a PRL promoter-regulatory region. Furthermore, since Pit-1 is known to be a powerful transactivator of PRL gene expression (31, 32), the observation of equivalent activities for Pit-1 and PREB on two PRL-regulatory region constructs (see Figs. 8 and 9) suggests that PREB may also represent a significant regulator of expression of this gene.

The results of electrophoretic shift mobility analysis described above imply that the PREB- and Pit-1-binding sites within element 1P are centered over different regions. We thus investigated the ability of PREB to regulate PRL promoter activity in the presence of Pit-1 (Fig. 10). As before, transfection of an equal amount (2.5 μg) of an expression construct for either protein alone yielded similar levels of transactivation of (−113)PRL-CAT. Cotransfection of 2.5 μg of expression vectors for each protein yielded a level of transactivation that was approximately additive over that yielded by either expression vector alone, suggesting that these two proteins exert additive actions on element 1P. Doubling the input of each expression vector, to close to maximal activity levels (see Fig. 9), increased expression approximately 3-fold, again consistent with an approximately additive action of PREB and Pit-1. These observations suggest that, at least in the basal cellular state, these two proteins may exert actions on element 1P that are of approximately equal strength, but are largely independent.

Figure 10.

PREB and Pit-1 Exhibit Additive Stimulation of Expression of a PRL Promoter Construct C6 cells were electroporated with (−113)PRL-CAT and RSV-β gal as in Fig. 9, plus the indicated amounts of RSV-PREB and/or RSV-Pit-1 (“1” = 2.5 μg plasmid), and treated thereafter as in Fig. 8. Shown is the mean ± sd of corrected CAT activity observed with triplicate dishes.

PREB Can Mediate Transcriptional Stimulation by PKA in Either GH₃ Pituitary Cells or Heterologous C6 Cells

As described above, previous studies have implied that Pit-1 is not the direct functional target of PKA action on the PRL promoter. Inspection of the predicted PREB amino acid sequence revealed a number of potential PKA phosphorylation sites (data not shown). This, together with the observations described above that PREB can bind specifically to PRL element 1P and exhibits transcriptional activity, suggested that PREB may represent the cellular protein that directly mediates PKA action on the PRL promoter via element 1P. To begin to investigate this possibility in GH₃ cells, under conditions that are independent of both endogenous PREB and PREB-binding sites, we investigated the ability of a GAL4-PREB fusion protein to transmit PKA action to a cotransfected GAL4 indicator construct (Fig. 11). As expected, control GAL(1–147) was unable to transactivate 5XGAL4-CAT expression, in either the presence or absence of an RSV-PKA expression vector. In the absence of RSV-PKA, GAL4-PREB alone strongly transactivated 5XGAL4-CAT, demonstrating again that PREB contains a transcriptional activator domain. Cotransfection of RSV-PKA yielded an approximately 3-fold increase in the ability of GAL4-PREB to transactivate 5XGAL4-CAT. This observation demonstrates that PREB can support a PKA-mediated transcriptional response in pituitary cells, possibly in the absence of any change in its ability to bind DNA.

Figure 11.

A GAL4-PREB Fusion Protein Can Transmit Basal and PKA-Stimulated Reporter Gene Activation GH₃ rat pituitary cells were electroporated with 5XGAL4-CAT (10 μg) and RSV-β gal (2 μg), plus 5 μg of either pGAL4(1–147) or GAL4-PREB, plus 5 μg of either RcRSV (PKA−) or RSV-PKA (PKA+), and treated thereafter as in Fig. 8. Shown is the mean ± sd[ of corrected CAT activity observed with triplicate dishes.

We next investigated whether PREB can transmit a PKA transcriptional signal to the PRL promoter, in the absence of other pituitary cell signals. To do this, we examined the effect of expression of RSV-PKA in C6 glial cells on transactivation by either Pit-1 or PREB of indicator construct (−113)PRL-CAT (Fig. 12). As before, (−113)PRL-CAT alone exhibited minimal activity, which was only slightly increased by coexpression of PKA. Coexpression of Pit-1 strongly transactivated CAT activity. However, this activity was not further increased by coexpression of PKA, in agreement with previous observations (2). In contrast, the transactivation of CAT activity by PREB was strongly increased by coexpression of PKA. Thus, PREB (but not Pit-1) can support a PKA-mediated transcriptional response directed by element 1P in the context of the PRL promoter.

Figure 12.

PREB but not Pit-1 Can Mediate Regulation by PKA of Expression of a PRL Promoter Construct C6 cells were electroporated with (−113)PRL-CAT and RSV-β gal as in Fig. 9, plus 2.5 μg of either RcRSV (None), RSV-Pit-1, or RSV-PREB, plus 5 μg of either RcRSV (PKA−) or RSV-PKA (PKA+), and treated thereafter as in Fig. 8. Shown is the mean ± sd of corrected CAT activity observed with triplicate dishes.

DISCUSSION

We report here the identification of an 1805-bp rat cDNA encoding a 417- amino acid protein, PREB. This protein binds directly to a regulatory element in the PRL promoter, element 1P, previously shown to represent a target for the pituitary-specific transcription factor Pit-1, and can mediate both basal and PKA-activated expression of the PRL promoter.

Analysis of the primary sequence of PREB revealed that it is a novel protein. However, PREB contains three segments (Fig. 3, WD I, WD II, and WD III) possessing a significant degree of homology to the WD repeat consensus and is thus identified as a member of the WD-repeat protein family (23). It is interesting that the highest degree of similarity between PREB and known proteins occurs in segments WD I and WD II, which resemble regions in the two yeast WD-repeat gene regulators, HIR1 (24) and Tup1 (25). Two significant features are shared by most known members of the subset of WD-repeat proteins that are gene regulators [HIR1 and Tup1, plus the Drosophila proteins Groucho (33) and dTAF_II80 (34), yeast Met30p (35), and plant COP1 (36)]: 1) Of the proteins whose regulatory functions are known (all but dTAF_II80), each is a repressor of its target gene. 2) With one possible exception, none is a DNA-binding protein; instead they all repress their target genes by interacting with a DNA-bound transcription factor. The possible exception is COP1, which may bind directly to DNA via a novel zinc-binding motif (36). The first common feature above suggests that PREB might prove, under some conditions, to represent a repressor of target gene expression. The second common feature implies that PREB may represent the first member of a new family of mammalian WD-repeat gene regulatory transcription factors that act by binding directly to a DNA target.

PREB(175–417) exhibited DNA sequence-specific binding to site 1P. However, the amino acid sequence of this portion of PREB exhibits no homology to known DNA-binding motifs, implying that this region of PREB will prove to contain a DNA-binding region(s) representing a novel member of this class of regulatory domains. Analysis by electrophoretic mobility shift assay of interaction of PREB(175–417) with site 1P mutants (Fig. 3) implied that at least a portion of the PREB-binding site resides upstream of the well defined binding site for Pit-1 in this element (20). This observation suggests the possibility that PREB and Pit-1 can co-occupy element 1P, which could represent the structural basis for the ability of PREB and Pit-1 to exhibit additive activity on the PRL promoter via this element (see below).

A number of observations suggest that PREB plays a role as a pituitary cell transcription factor. Antiserum raised against recombinant PREB detects a major protein that localizes to the nuclei of GH₃ cells (Fig. 7). It is conceivable that the protein detected corresponds to a cross-reacting protein. However, the observation that the apparent size of the detected protein, 45 kDa, corresponds to that predicted for PREB, and that PREB mRNA is expressed in GH₃ cells (Fig. 5), strongly suggests that this protein represents endogenous PREB, and that PREB is thus a nuclear protein in pituitary cells. The observation that PREB exhibits sequence-specific binding to a specific element in the PRL promoter (Fig. 3) suggested that PREB might be capable of regulating transcription of this promoter. The ability of PREB to stimulate PRL promoter activity was demonstrated directly by the observation that this protein can stimulate activity of cotransfected PRL-CAT constructs in heterologous cells (Figs. 8–10 and 12). In these investigations, PREB yielded transactivation of a truncated PRL promoter that was equivalent to that yielded by Pit-1. The interexperimental variability in the relative basal transactivation activities of the two proteins on either the full-length PRL promoter/enhancer region (see Results text) or on this truncated promoter (e.g. compare Figs. 9 and 10 with Fig. 12) does not permit a more precise delineation of the relative activities of PREB and Pit-1 on PRL promoter activity. However, the observation that the two proteins exhibited approximately additive transactivational activities on the truncated PRL promoter (Fig. 10) is consistent with the concept that PREB can strongly regulate basal PRL promoter activity independently of Pit-1. Although these experiments, together with the transactivational properties of a GAL4-PREB fusion (Fig. 11), clearly imply that PREB can act as a transcriptional activator, it is not presently known which region of the protein is responsible for transactivation. However, as noted above, it seems possible that such a domain may encompass either or both of the PREB proline/glutamine-rich regions between residues 86–134 and 223–279. The observation that a region containing this type of motif has been shown previously to mediate the ability of the tumor repressor WT1 to repress transcription (27) suggests the interesting possibility that PREB may, in the appropriate context, act as a transcriptional repressor.

What might be the physiological role of PREB in pituitary cells? The ability of PREB to transactivate cotransfected PRL constructs in heterologous cells suggests a role for this protein in basal PRL gene expression. Similarly, the ability of PREB (Fig. 12), but not Pit-1 (Fig. 12; see also Refs. 2, 12), to mediate PKA stimulation of a cotransfected PRL-CAT construct in heterologous cells suggests that PREB may play a role in cAMP-mediated transcriptional actions of extracellular stimulators of PRL promoter activity such as PACAP, which is apparently completely dependent upon cellular PKA activity for its transcriptional activation of the PRL promoter (21). The ability of a PREB fusion protein to direct either basal or PKA-stimulated transcriptional effects to a heterologous target promoter in a pituitary cell context (Fig. 11) lends further support to this suggestion. However, it should be emphasized that a determination of whether PREB does play a significant physiological role in either basal or PKA-stimulated regulation of PRL gene expression will require further investigations, involving examination of the consequences of repression of PREB activity in pituitary cells.

The present studies imply that PREB has the capacity to mediate PKA action on the PRL promoter in the absence of Pit-1 (Fig. 12). This property and the observations that PREB exhibits sequence-specific binding to element 1P (Fig. 2) and that its mRNA sequences are expressed in multiple tissues (Fig. 6) are consistent with the functional properties of a factor detected by Gutierrez-Hartmann and co-workers (5) in HeLa cells, which has been reported to exhibit Pit-1-independent regulation of the PRL promoter via element 1P. Although the HeLa cell factor has apparently not yet been further characterized, it is thus possible that PREB corresponds to this factor. The studies to date of both PREB and the HeLa cell factor imply that PKA-mediated regulation of PRL promoter activity can occur in the absence of Pit-1. However, it seems possible that, although Pit-1 alone cannot mediate PKA transcriptional actions, an interaction between Pit-1 and PREB on PRL promoter element site 1P could yield an enhanced response by pituitary cells to PKA-mediated stimuli. A possible model for this type of cellular role for PREB would involve PREB activation via PKA-mediated phosphorylation. Although it is not yet known whether PREB can serve as a PKA substrate either in vitro or in vivo, the predicted sequence of this protein (Fig. 1) does contain a number of sequences resembling PKA phosphorylation consensus sites (37). Finally, a very recent report that the cofactor CBP (CREB-binding protein) can interact with Pit-1 during a PKA-mediated transcriptional response (38) suggests the interesting possibility of an anologous functional interaction between CBP and PREB.

A number of observations suggest that PREB may also play a transcriptional role in organs other than the pituitary. The observation that a GAL4-PREB construct can transmit both basal and PKA-stimulated transcriptional effects to a GAL4 promoter construct (Fig. 11) implies that PREB can exert these actions independently of its ability to bind to the PRL promoter. Consistent with a wider role for this protein, PREB transcripts were observed in all human tissues examined (Fig. 6A). It is of course conceivable that some of the PREB transcript size variants detected in specific tissues in this experiment arise from modifications in either transcript initiation site or poly(A) tail size. However, the tissue-specific transcript size patterns observed does suggest the possibility of alternative splicing of the initial PREB transcript, and thus of the production in different tissues of isoforms of PREB potentially possessing unique transcriptional properties. It is worth noting in this connection that the PREB cDNA sequence (Fig. 1) contains a 306-bp (102-amino acid) open reading frame that starts at position 1333 and thus slightly overlaps the terminus of the major PREB open reading frame, suggesting that domain shuffling arising from alternative splicing might generate a PREB isoform containing some or all of this peptide sequence at its N terminus. Finally, the detection of PREB transcripts in human brain tumors arising from astrocytes, but not from glial or neuroepithelial cells (Fig. 6C), suggests that the PREB gene exhibits region-specific brain expression, and thus the intriguing possibility that PREB serves a specific transcriptional function within restricted portions of the brain.

In summary, we have cloned from a pituitary cell cDNA library a transcription factor, termed PREB, that can bind to a regulatory element in the PRL promoter, and has the capacity to mediate both basal and PKA-stimulated activity, either of the PRL promoter, or as a fusion protein of a heterologous promoter. Further investigations should delineate the regions of PREB that direct binding to, and transcriptional activation of, the PRL promoter and also illuminate the physiological role of this novel transcription factor, both within the pituitary and in other tissues that express the PREB gene transcript.

MATERIALS AND METHODS

Imaging of Gels

Autoradiograms of gel blots or photographs of gels were recorded electronically using an AGFA (ARCUS II) Scanner (Agfa-Gevaert Group, Mortsel, Belgium).

Electrophoretic Mobility Shift Assay

Partially purified recombinant his-tagged PREB (PREB amino acids 175–417, preceded by six histidines) was prepared from Escherichia coli by solubilization of an insoluble pellet in binding buffer [5 mm imidazole, 0.8 m NaCl, 10 mm Tris (pH 7.9), 8 m urea], application to a nickel affinity column, gradual refolding in the column by progressive dilution of the urea as described (17), elution in buffer containing 0.5–1 m imidazole, followed by desalting and concentration in an Amicon (Beverly, MA) spin column. We have previously described the sequences of double-stranded oligonucleotides corresponding to PRL promoter sites 1P, ^*1P, 1P^*, and 3P (39) and the CLE (29). The sequence of the double-stranded SP1-binding site oligonucleotide probe (containing a 5′ SalI site) is 5′-TCGACGGGGCGGGGCC-3′, and of the double-stranded rat GH pGHF-1 site is 5′-TCGACTGGCTCCAGCCATGAATAAATGTATAGGGAAAG-3′. All procedures were performed at 4 C in the presence of protease inhibitors [1 mm phenylmethylsulfonyl fluoride (PMSF), 1 μg/ml aprotonin, and 5 μg/ml leupeptin]. The partially purified his-tagged PREB was then incubated 10 min at room temperature in 9 μl containing 10 mm Tris (pH 7.9), 60 mm KCl, 1 mm EDTA, 0.03% NP40, 4% Ficoll, 1 mm dithiothreitol, 5 μg poly(dI-dC), 1 μg BSA, with or without unlabeled DNA competitors, and then an additional 10 min after addition of 1 ng ³²P-end-labeled site 1P probe, followed by analysis on a 5% polyacrylamide gel in 0.25× Tris-borate-EDTA (TBE) at 4 C. The dried gel was then autoradiographed 1–3 h at −70 C with intensifying screens.

Cell Lines

GH₃ rat pituitary cells (19) were propagated in suspension culture as described (40). C6 rat glial cells (41) were maintained in monolayer culture in DMEM containing 5% FCS, 100 mg/ml streptomycin, and 100 U/ml penicillin at 37 C in 10% CO₂, and cultures were split 1:5 every other day.

Southern Blot Analysis

Standard procedures were employed to prepare genomic DNA from rat GH₃ cells and from Drosophila (provided by Dr. M. Frasch of Mount Sinai). Human DNA was provided by Dr. D. Bishop (Mount Sinai School of Medicine). Yeast genomic DNA was prepared employing a PureGene kit obtained from GENTRA (Minneapolis, MN). Southern blot analysis was performed as described previously (42), employing a ³²P-labeled probe prepared by either PCR or random primer labeling from a PREB cDNA template, followed by autoradiography for 3–16 h as described above.

Northern Blot and PCR Analysis

Total or poly(A)-enriched RNA was prepared from the GH₃ or C6 cell lines using the Oligotex column affinity purification system (Qiagen, Inc., Chatsworth, CA) without or with addition of oligo-d(T)-linked beads. For Northern analysis, RNA was analyzed by 1% formaldehyde agarose gel electrophoresis as described (43). After transfer of size-separated RNA to nitrocellulose, baking, and prehybridization (42), the nitrocellulose was hybridized for 15 h to a ³²P-labeled PREB cDNA probe prepared as described above, and the membranes were then autoradiographed for 1–3 days as described above. For PCR analysis of RNA samples, RNA was denatured (70 C, 5 min), annealed with oligo(dT), and subjected to first-strand synthesis in the presence of 100 U reverse transcriptase (SuperSCRIPT II), according to the directions of the manufacturer (Life Technologies, Inc., Gaithersburg, MD). The resultant cDNA, or cDNA prepared from tumor sample RNA, was then subjected to nested PCR using Taq polymerase (Perkin-Elmer Corp., Norwalk, CT). The first and second amplification employed primer sets corresponding, respectively, to positions 1081–1097/1776–1761 and 1300–1318/1654–1636 of the PREB cDNA sequence shown in Fig. 1. The first amplification began with a hot start (94 C, 5 min; 80 C, 5 min), followed by 35 cycles [94 C, 1 min; 41 C, 1 min; 72 C, 40 sec (10 min in the last cycle)]. The second amplification employed the same conditions, except that in each cycle a 60 C annealing temperature was employed instead of 41 C, and the 40-sec extension time was replaced by 30 sec.

Preparation of anti-PREB Antiserum

Inclusion bodies were prepared from E. coli expressing recombinant his-tagged PREB (PREB amino acids 175–417, preceded by six histidines), as described (43). Briefly, cells were lysed with lysozyme and deoxycholic acid in the presence of 50 mm PMSF, treated with DNAase (1 mg/ml) at room temperature 15–30 min, and subjected to centrifugation. The pellet was then suspended, washed with 6.5 m urea in 0.1 mm Tris, pH 8.5, after which PREB was extracted with elution buffer (8 m urea, 50 mm Tris, pH 8.0, 1 mm EDTA, 100 mm NaCl, 0.1 mm PMSF), and subjected to SDS-PAGE. A gel fragment containing the major PREB band was excised, frozen, ground with a mortar and pestle, and supplied frozen to Cocalico Biologicals, Inc. (Reamstown, PA) for preparation of antiserum in rabbits according to their standard protocol. Before use in Western blots and immunocytochemistry, either anti-PREB antiserum or preimmune serum from the same animal was preadsorbed with extracts of host E. coli. An equal volume of 2× SDS sample buffer was added to a pelleted 50-ml bacterial culture, mixed thoroughly, incubated at 65–70 C for 10 min, fractionated on a 4.5% SDS-PAGE minigel employing a 1.5-mm preparative comb, and transferred to nitrocellulose. After soaking the filter in a 5% solution of Carnation brand nonfat milk, the filter was incubated on a rocking platform either 1 h at room temperature or 4 C overnight, with anti-PREB antiserum diluted 1:250 into TBST buffer (20 mm Tris, pH 7.6; 137 mm NaCl, 0.1% Tween-20) containing 3% BSA, and the adsorbed antiserum was employed for analysis of PREB.

Western Blot Analysis of Cellular Fractions

Nuclei and cytosol (i.e. the soluble fraction of the postnuclear supernatant) were prepared from GH₃ cells as described (44), in the presence of the protease cocktail described above. Preliminary SDS-PAGE analysis revealed no gross degradation of proteins in either fraction (data not shown), and that, on a per cell basis, the cytosol contained a considerably higher protein content than the nucleus. Samples were subjected to SDS-PAGE, transferred to nitrocellulose on a semidry transfer apparatus (Hoeffer Scientific Instruments, San Francisco. CA), employing Towbin buffer (25 mm Tris, 192 mm glycine, 0.0372% SDS, 20% methanol) for 30–60 min at 100 mA. Filters were blocked for 1 h with 5% Carnation nonfat dry milk in TBST, incubated with anti-PREB antibody (1:250), followed by washing with TBST. The filter was then exposed to secondary antibody (conjugated to horseradish peroxidase and diluted 1:5000) in TBST containing 3% BSA for 30–60 min. After washing with TBST (four times for 10 min each time), immunoreactive proteins were visualized by enhanced chemiluminescence according to the directions of the manufacturer (Amersham Corp., Arlington Heights, IL).

Immunocytochemistry of Cultured Cells

Cells were plated on glass coverslips in serum-containing media the night before use (C6 cells, DMEM plus 5% FCS; GH₃ cells in Ham’s F10 plus 15% horse serum plus 2.5% FCS). For use with GH₃ cells, the coverslips were first coated with CellTak (14 μg/25 mm coverslip) (Collaborative Biomedical Products, Bedford, MA). The cells were washed twice with PBS, fixed 20 min at room temperature in 2% paraformaldehyde, incubated 20 min in blocking buffer (0.6% Tween 20 in DMEM containing 5% FCS) and then incubated overnight at 4 C with either 1:250 dilutions of preimmune serum or anti-PREB, each preadsorbed as described above. The dishes were then washed three times with PBS and incubated 30 min with rhodamine-labeled goat antirabbit IgG (American Qualex, LaMiranda, CA) at 1:500 in blocking buffer, rinsed three times, and mounted in Mowiol. Images were captured using a 40× oil objective on a Nikon (Melville, NY) inverted microscope with a Cohu (San Diego, CA) CCD4910 camera in conjunction with a Colorado video integrator unit. Digital data were processed with Metamorph software from Universal Imaging (Media, PA).

Plasmids

We have previously described construction of plasmids (−1957)PRL-CAT and (−113)PRL-CAT (45, 46). RSV-PKA was prepared by employing PCR primers to amplify the murine PKA catalytic subunit described by Uhler and McKnight (47), followed by HindIII/Xba digestion of the amplified product, and ligation into the corresponding site in plasmid RcRSV (Invitrogen, San Diego, CA). RSV-Pit-1 (32) and RSV-β gal were kindly supplied by Dr. H. Samuels (New York University). GAL4 constructs pSG424 (48) [referred to in the present paper as pGAL4(1–147)] and 5XGAL4-CAT (49) were kindly supplied by Dr. M. Ptashne (Sloan-Kettering Institute). GAL4-PREB was constructed by cloning the PREB coding sequence (amino acids 1–417) upstream of and in register with the GAL4(1–147) sequence in pSG424. RSV-PREB was constructed by cloning the PREB-coding sequence into the HindIII/XbaI site of RcRSV.

Transient Cotransfection Assays

For each treatment group, approximately 0.5 × 10⁸ GH₃ cells or 2 × 10⁶ C6 cells were subjected to electroporation. Preliminary experiments were performed with PRL-CAT constructs and a known PRL promoter regulator, Pit-1, to optimize transfection efficiency for each cell line (data not shown). For transfection, cells were resuspended in 0.75 ml DMEM containing 10% FBS plus the indicated plasmids and subjected to electroporation at 960 μFarads and either 300 V (C6 cells) or 240 V (GH₃ cells), divided among three 60-mm tissue culture dishes, and incubated 48 h as described above for each cell line. One day after transfection, each dish was examined microscopically, and any experiment that exhibited gross differences in cell survival among different treatment groups was discarded. Cells were then harvested with a rubber policeman, lysed by sonication (four times for 2 min each time) in 0.25 m Tris, pH 7.8, 10 mm EDTA, and heated 65 C for 10 min to inactivate deacetylases. Half of each cell extract was assayed for CAT activity as described previously (12), employing [³H]chloramphenicol (0.01 μCi/μl) and butyryl CoA (5 mg/ml) and a 4-h incubation, which yielded results in the linear range of the assay. The remainder of the cell extract was employed for assay of β-galactosidase activity as described (42). For each experimental condition, the average CAT assay result was divided by the average β-galactosidase activity under that condition, relative to a value of 1 assigned to the average β-galactosidase activity in the controls. It may be noted that this procedure always yielded a correction of ≤2%. Each experiment reported here has been repeated a total of at least three times, with results similar to those shown in the figures.

Acknowledgments

We thank Drs. S. Aaronson and A. Chan (Mount Sinai School of Medicine, New York, NY) for generously supplying cDNA samples prepared from mRNA isolated from various human brain tumors. We also thank Drs. M. Ptashne (Harvard University, Boston, MA) and H. Samuels (New York University, NY) for kindly supplying the plasmids indicated in the text, and Drs. M. Frasch and D. Bishop (Mount Sinai School of Medicine) for providing, respectively, Drosophila and a sample of human DNA.

This work was supported by NIH Grant GM-36847 and GM-056186 (to C.B.) and NIH Grant DK-19974 and Cancer Center Core Research Grant CA-11098 (to P.M.H.). M.S.F. was supported by NIH Cellular and Molecular Endocrinology Training Grant DK-07645.

^*

Present address: Department of Otolaryngology, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205.