A tissue-specific transcription factor containing a homeodomain specifies a pituitary phenotype

Pituitary cell phenotypes involve cell-specific Pit-1 mRNA translation and synergistic interactions with other classes of transcription factors

Development of the anterior pituitary gland involves proliferation and differentiation of ectodermal cells in Rathke's pouch to generate five distinct cell types that are defined by the trophic hormones they produce. A detailed ontogenetic analysis of specific gene expression has revealed novel aspects of organogenesis in this model system. The expression of transcripts encoding the alpha-subunit common to three pituitary glycoprotein hormones in the single layer of somatic ectoderm on embryonic day 11 established that primordial pituitary cell commitment occurs prior to formation of a definitive Rathke's pouch. Activation of Pit-1 gene expression occurs as an organ-specific event, with Pit-1 transcripts initially detected in anterior pituitary cells on embryonic day 15. Levels of Pit-1 protein closely parallel those of Pit-1 transcripts without a significant lag. Unexpectedly, Pit-1 transcripts remain highly expressed in all five cell types of the mature pituitary gland, but the Pit-1 protein is detected in only three cell types--lactotrophs, somatotrophs, and thyrotrophs and not in gonadotrophs or corticotrophs. The presence of Pit-1 protein in thyrotrophs suggests that combinatorial actions of specific activating and restricting factors act to confine prolactin and growth hormone gene expression to lactotrophs and somatotrophs, respectively. A linkage between the initial appearance of Pit-1 protein and the surprising coactivation of prolactin and growth hormone gene expression is consistent with the model that Pit-1 is responsible for the initial transcriptional activation of both genes. The estrogen receptor, which has been reported to be activated in a stereotypic fashion subsequent to the appearance of Pit-1, appears to be capable, in part, of mediating the progressive increase in prolactin gene expression characteristic of the mature lactotroph phenotype. This is a consequence of synergistic transcriptional effects with Pit-1, on the basis of binding of the estrogen receptor to a response element in the prolactin gene distal enhancer. These data imply that both transcriptional and post-transcriptional regulation of Pit-1 gene expression and combinatorial actions with other classes of transcription factors activated in distinct temporal patterns, are required for the mature physiological patterns of gene expression that define distinct cell types within the anterior pituitary gland.

An insulin-induced DNA-binding protein for the human growth hormone gene

The control of gene transcription is usually mediated by transacting transcriptional factors that bind to upstream regulatory elements. As insulin regulates transcription of the growth hormone (GH) gene, we tested nuclear extracts from unstimulated and insulin-stimulated Chinese hamster ovarian (CHO) cells for binding to four human GH (hGH) gene promoter oligonucleotide fragments identified as target-binding sequences by DNAse I footprinting. Using a mobility shift assay, an insulin-induced DNA-binding protein was identified. This protein binds to two upstream overlapping oligonucleotide sequences. Binding activity is present at low levels in unstimulated CHO cells and is stimulated by insulin treatment with a time course suggesting that protein synthesis is required. Incubation of the cells with cycloheximide and puromycin confirmed that de novo protein synthesis is necessary for the increased binding activity. Competition with excess unlabeled specific competitor oligonucleotides prevented binding, while unrelated similar-sized oligonucleotides failed to compete for binding, indicating that the observed DNA-protein complex formation is specific. A protein of approximately 70-80 kD was detected by gradient gel electrophoresis. In conclusion, insulin-mediated DNA-protein binding has been identified on the upstream hGH promoter, suggesting a trans-active role for insulin in mediating polypeptide hormone gene expression.

Identification of a mammalian nuclear factor and human cDNA-encoded proteins that recognize DNA containing apurinic sites

Damage to DNA can have lethal or mutagenic consequences for cells unless it is detected and repaired by cellular proteins. Repair depends on the ability of cellular factors to distinguish the damaged sites. Electrophoretic binding assays were used to identify a factor from the nuclei of mammalian cells that bound to DNA containing apurinic sites. A binding assay based on the use of beta-galactosidase fusion proteins was subsequently used to isolate recombinant clones of human cDNAs that encoded apurinic DNA-binding proteins. Two distinct human cDNAs were identified that encoded proteins that bound apurinic DNA preferentially over undamaged, methylated, or UV-irradiated DNA. These approaches may offer a general method for the detection of proteins that recognize various types of DNA damage and for the cloning of genes encoding such proteins.

DBP, a liver-enriched transcriptional activator, is expressed late in ontogeny and its tissue specificity is determined posttranscriptionally

The full-length cDNA for a transcriptional activator, DBP, that binds to the D site of the albumin promoter has been cloned. DBP belongs to a family of related transcription factors including Fos, Jun, CREB, and C/EBP, which share a conserved basic domain. However, unlike most other members of this family, DBP does not contain a "leucine zipper" structure. Among several rat tissues tested, significant levels of its protein are only observed in liver; yet, with the exception of testis, DBP mRNA is present in all of the examined tissues. DBP as well as its mRNA accumulate to significant levels only in adult animals. During chemically induced liver regeneration, DBP expression is rapidly down-regulated, suggesting that DBP may be involved in the proliferation control of hepatocytes. This cell growth-dependent expression of DBP, in contrast to its tissue specificity, appears to be controlled at the level of mRNA accumulation.

Transcriptional stimulation via SC site of Bombyx sericin-1 gene through an interaction with a DNA binding protein SGF-3

Three protein binding sites have been identified in the upstream region of the sericin-1 gene. Two of them, SA and SC sites, have been known as putative cis-acting elements. Using synthetic oligonucleotides of these binding sites, it was found that silk gland factor-1 (SGF-1) binds to the SA site, and silk gland factor-3 (SGF-3) binds to the SC site but not to a mutated SC site, SCM. Tissue distribution of the two factors was different. SGF-3 is present abundantly in the middle silk gland (MSG) where the sericin-1 gene is transcribed specifically but is also present in other cell types, though in a much less concentration. SGF-1 is observed very abundantly in the two parts of silk gland, MSG and posterior silk gland (PSG), but not in other cells. Templates containing multimerized SA or SC sites at -39 of the sericin-1 gene promoter were tested in MSG nuclear extracts. The SC multimer strongly activated transcription, while the mutant SCM multimer did not. The SA multimer also gave a slight stimulation of transcription. These results suggest that SGF-3 stimulates transcription through an interaction with the SC site, and SGF-1 does so weakly through the SA site.

Sp1 can displace GHF-1 from its distal binding site and stimulate transcription from the growth hormone gene promoter

DNase I footprinting experiments showed that binding activities of Sp1 and of GHF-1 to its distal site on the human growth hormone gene promoter are mutually exclusive. The kinetics of GHF-1 binding were indicative of positive cooperativity. The Sp1 site did not affect promoter activity in cell-free transcription. Still, Sp1 could compensate partially for the decreased stimulation of transcription seen at low GHF-1 concentrations.

New type of POU domain in germ line-specific protein Oct-4

Members of a family of murine octamer-binding proteins interact specifically with the octamer motif, a transcription regulatory element found in the promoter and enhancer regions of many genes. Oct-4 is a maternally expressed protein that is also present in the pre-implantation mouse embryo. Although many regulatory proteins are expressed in post-implantation embryos, transcription factors regulating pre-implantation processes have remained elusive. The Oct-4 gene is therefore a prime candidate for an early developmental control gene. Here we report the complementary DNA cloning of the mouse Oct-4 gene, and the characterization of the encoded protein(s) by sequential in vitro transcription, translation, DNA-binding and protease-clipping analysis. Deletion analysis shows that the DNA-binding activity is mediated by a POU domain encoded in an open reading frame corresponding to a 324-amino-acid protein. Sequence comparison with known POU domains reveals that Oct-4 is a novel member of the POU-family.

The ubiquitous octamer-binding protein(s) is sufficient for transcription of immunoglobulin genes

All immunoglobulin genes contain a conserved octanucleotide promoter element, ATGCAAAT, which has been shown to be required for their normal B-cell-specific transcription. Proteins that bind this octamer have been purified, and cDNAs encoding octamer-binding proteins have been cloned. Some of these proteins (referred to as OTF-2) are lymphoid specific, whereas at least one other, and possibly more (referred to as OTF-1), is found ubiquitously in all cell types. The exact role of these different proteins in directing the tissue-specific expression of immunoglobulin genes is unclear. We have identified two human pre-B-cell lines that contain extremely low levels of OTF-2 yet still express high levels of steady-state immunoglobulin heavy-chain mRNA in vivo and efficiently transcribe an immunoglobulin gene in vitro. Addition of a highly enriched preparation of OTF-1 made from one of these pre-B cells or from HeLa cells specifically stimulated in vitro transcription of an immunoglobulin gene. Furthermore, OFT-1 appeared to have approximately the same transactivation ability as OTF-2 when normalized for binding activity. These results suggest that OTF-1, without OTF-2, is sufficient for transcription of immunoglobulin genes and that OTF-2 alone is not responsible for the B-cell-specific regulation of immunoglobulin gene expression.

Expression of GHF-1 protein in mouse pituitaries correlates both temporally and spatially with the onset of growth hormone gene activity

The relationship between expression of the pituitary-specific transcription factor, GHF-1, and activation of the growth hormone and prolactin genes during mouse anterior pituitary development was investigated. While GHF-1 transcripts were detected within 24 hr of the first observable events in anterior pituitary differentiation, no GHF-1 protein could be detected until about 3 days later. The appearance of GHF-1 protein showed good temporal and spatial correlation with activation of the growth hormone gene. Prolactin gene expression, on the other hand, was observed transiently during embryonic day 16 in two different populations of cells, of which the major one does not contain GHF-1 or growth hormone. These results suggest that expression of GHF-1 is controlled both transcriptionally and posttranscriptionally. The spatial and temporal correlation between the appearance of GHF-1 protein and growth hormone gene activation suggests that GHF-1 is responsible for this very last step in the specialization of somatotrophic cells.

HNF-1 shares three sequence motifs with the POU domain proteins and is identical to LF-B1 and APF

The coordinate expression of genes during development and differentiation is thought to be accomplished by common transcription factors operating on the promoters of families of coexpressed genes. HNF-1 is a transcriptional factor involved in the expression of genes in the liver and was originally defined as playing a major role in coordinating the expression of the linked fibrinogen genes. We have isolated cDNA clones for HNF-1 using oligonucleotides prepared to the sequence of the purified protein. The sequence of HNF-1 shares homeo domain, as well as short acidic and basic sequences with the POU family of transcriptional activators. Peptides from the protein interacting with the albumin proximal element, or B box (APF), and the factor interacting with the alpha 1-antitrypsin promoter (LF-B1) are found in the predicted sequence of HNF-1. HNF-1 mRNA is not present in the dedifferentiated hepatoma variant, C2, but reappears upon selection for gluconeogenesis coincident with the re-expression of liver-specific genes. Finally, the mRNA is not present in somatic cell hybrids in which liver-specific gene expression is extinguished. In contrast to earlier published results, we find that in addition to being present in the liver, HNF is expressed in the kidney, intestine, and spleen, but not in other tissues. This pattern of expression mirrors the complex pattern of expression of many genes, such as alpha-fetoprotein, alpha 1-antitrypsin, and fibrinogen, whose promoters contain HNF-1 sites. These data indicate that HNF-1 is a more broadly acting transcription factor than has been indicated by previous work.

The ubiquitous octamer-binding protein(s) is sufficient for transcription of immunoglobulin genes

All immunoglobulin genes contain a conserved octanucleotide promoter element, ATGCAAAT, which has been shown to be required for their normal B-cell-specific transcription. Proteins that bind this octamer have been purified, and cDNAs encoding octamer-binding proteins have been cloned. Some of these proteins (referred to as OTF-2) are lymphoid specific, whereas at least one other, and possibly more (referred to as OTF-1), is found ubiquitously in all cell types. The exact role of these different proteins in directing the tissue-specific expression of immunoglobulin genes is unclear. We have identified two human pre-B-cell lines that contain extremely low levels of OTF-2 yet still express high levels of steady-state immunoglobulin heavy-chain mRNA in vivo and efficiently transcribe an immunoglobulin gene in vitro. Addition of a highly enriched preparation of OTF-1 made from one of these pre-B cells or from HeLa cells specifically stimulated in vitro transcription of an immunoglobulin gene. Furthermore, OFT-1 appeared to have approximately the same transactivation ability as OTF-2 when normalized for binding activity. These results suggest that OTF-1, without OTF-2, is sufficient for transcription of immunoglobulin genes and that OTF-2 alone is not responsible for the B-cell-specific regulation of immunoglobulin gene expression.

Growth hormone gene regulation: a paradigm for cell-type-specific gene activation

The growth hormone (GH) gene is specifically expressed within specialized cells of the anterior pituitary. The central role in GH gene activation is played by GHF-1, a homeodomain protein that is itself specifically expressed in the anterior pituitary.

Transcriptional and posttranscriptional regulation of the rat prolactin gene by calcium

The rat prolactin gene is expressed at a high basal level in the pituitary tumor GH3 cell line. Culturing GH3 cells in a low-Ca2+, serum-free medium (SFM) depresses prolactin mRNA levels, and subsequent addition of Ca2+ to the SFM results in a specific, gradual, and sustained increase in prolactin mRNA levels. We have now examined whether the observed increase in prolactin mRNA levels can be attributed solely to an increase in the transcriptional rate of the prolactin gene. Treatment of GH3 cells in SFM with 0.4 mM CaCl2 for 24 to 48 h increased cytoplasmic prolactin mRNA levels by 5- to 10-fold, whereas the transcriptional rate of the prolactin gene was increased by less than twofold over values for SFM controls. Prolactin mRNA levels increased progressively during the 24-h period after Ca2+ addition, whereas prolactin gene transcription never exceeded a twofold increase over values for SFM controls. The activities of nuclear extracts from control and Ca2(+)-induced cells were examined in an in vitro transcription assay. The two extracts directed transcription from the prolactin promoter and the adenovirus major late promoter equally well. Cycloheximide had no effect on the ability of Ca2+ to increase or maintain prolactin mRNA levels. In dactinomycin mRNA clearance experiments, prolactin mRNA was cleared at the same rate in the absence and presence of Ca2+. These results demonstrate that although Ca2+ has a small effect on the transcriptional rate of the prolactin gene, Ca2+ produces a significant increase in prolactin mRNA levels by acting at a posttranscriptional site(s). Furthermore, Ca(2+) appears to increase prolactin mRNA levels by posttranslational modification of a stable protein, probably at a nuclear site.

Chromosomal translocation t(1;19) results in synthesis of a homeobox fusion mRNA that codes for a potential chimeric transcription factor

The gene (E2A) for enhancer binding transcription factors E12 and E47 maps to the t(1;19) chromosomal translocation breakpoint in pre-B cell leukemias. Altered E2A transcripts lacking sequences coding for the helix-loop-helix DNA binding motif were detected in several t(1;19)-carrying cell lines. Fusion cDNAs that crossed the t(1;19) breakpoint were cloned and shown to code for an 85 kd protein consisting of the amino-terminal two-thirds of E2A fused to a chromosome 1-derived protein. The fusion protein has the features of a chimeric transcription factor in which the DNA binding domain of E2A is replaced by the putative DNA binding domain of a homeoprotein from chromosome 1 for which the name Prl (pre-B cell leukemia) is proposed. Identical E2A-prl mRNA junctions were detected by PCR in three t(1;19)-carrying cell lines, indicating that the fusion transcripts and predicted chimeric protein are a consistent feature of this translocation.

Differential transcriptional activation by Oct-1 and Oct-2: interdependent activation domains induce Oct-2 phosphorylation

The ubiquitous Oct-1 and lymphoid Oct-2 POU homeodomain transcription factors bind to the same DNA sequence but differ in their activation potential. Oct-2 is a positive, negative, or neutral regulator of beta-globin transcription depending on the position and sequence of multimerized binding sites. To activate transcription, Oct-2 relies on two interdependent nonacidic domains, an N-terminal glutamine-rich region and a C-terminal serine-, threonine-, and proline-rich region. Oct-1 also contains a functional glutamine-rich region but fails to activate beta-globin transcription in our assay because the Oct-1 C-terminus is inactive, indicating that differential activation by Oct-1 and Oct-2 is determined by the combination of multiple activation domains. Oct-2 displays a unique phosphorylation pattern that is absent from molecules lacking one or the other activation domain, suggesting the activation domains have a role in inducing protein phosphorylation.

A novel octamer binding transcription factor is differentially expressed in mouse embryonic cells

We have identified a novel octamer binding factor (Oct-3) in P19 embryonal carcinoma cells. Oct-3, which recognizes the typical octamer motif (ATTTGCAT) as well as the AT-rich sequence TTAAAATTCA, is present in P19 stem cells but disappears when the cells are induced to differentiate by retinoic acid (RA). Cloned cDNA corresponding to Oct-3 encodes a protein of 377 amino acids. Oct-3 has a conserved POU domain, but the remaining part is distinct from other POU domain-containing proteins such as Oct-1 and Oct-2. mRNA of 1.5 kb coding for Oct-3 is abundant in P19 stem cells but is dramatically repressed during RA-induced differentiation. Repression of the 1.5 kb mRNA is rapid and specific to RA. In mouse, oct-3 mRNA is undetectable in all the adult organs examined. The N-terminal proline-rich region of Oct-3, when fused to the DNA binding domain of c-Jun, functions as a transcriptional activating domain. We suggest that Oct-3 is a novel octamer binding transcription factor that is developmentally regulated during mouse embryogenesis.

The rat albumin promoter: cooperation with upstream elements is required when binding of APF/HNF1 to the proximal element is partially impaired by mutation or bacterial methylation

We have characterized in the accompanying paper (P. Herbomel, A. Rollier, F. Tronche, M.-O. Ott, M. Yaniv, and M. C. Weiss, Mol. Cell. Biol. 9:4750-4758, 1989) six different elements in the albumin promoter. One of them, the proximal element (PE), is the binding site for a strictly liver specific factor, APF/HNF1. This binding site contains a bacterial DAM DNA methylase methylation target sequence which, when methylated, decreases the affinity of the protein for this element. When the different albumin promoter constructions were prepared in an Escherichia coli deoxyadenosine methylase-negative strain, the respective contributions of the elements to the overall promoter activity were strikingly different. An intact proximal element plus the TATA box gave almost full transcriptional activity in transient transfection experiments and only in differentiated hepatoma cells of line H4II, whereas the distal elements (distal element III [DEIII], the NF1-binding site DEII, and the E/CBP-binding site DEI) had become essentially dispensable. Mutations affecting the CCAAT box showed only a two- to threefold decrease. When PE was methylated, mutated, or replaced by the homologous element from the alpha-fetoprotein gene, activity in the context of the short promoter (PE plus the TATA box) was abolished. However, activity was restored in the presence of the upstream elements, showing that cooperation with factors binding to the CCAAT box and distal elements favors the functional interaction of the liver-specific APF/HNF1 factor with lower-affinity binding sites.

Transcriptional and posttranscriptional regulation of the rat prolactin gene by calcium

The rat prolactin gene is expressed at a high basal level in the pituitary tumor GH3 cell line. Culturing GH3 cells in a low-Ca2+, serum-free medium (SFM) depresses prolactin mRNA levels, and subsequent addition of Ca2+ to the SFM results in a specific, gradual, and sustained increase in prolactin mRNA levels. We have now examined whether the observed increase in prolactin mRNA levels can be attributed solely to an increase in the transcriptional rate of the prolactin gene. Treatment of GH3 cells in SFM with 0.4 mM CaCl2 for 24 to 48 h increased cytoplasmic prolactin mRNA levels by 5- to 10-fold, whereas the transcriptional rate of the prolactin gene was increased by less than twofold over values for SFM controls. Prolactin mRNA levels increased progressively during the 24-h period after Ca2+ addition, whereas prolactin gene transcription never exceeded a twofold increase over values for SFM controls. The activities of nuclear extracts from control and Ca2(+)-induced cells were examined in an in vitro transcription assay. The two extracts directed transcription from the prolactin promoter and the adenovirus major late promoter equally well. Cycloheximide had no effect on the ability of Ca2+ to increase or maintain prolactin mRNA levels. In dactinomycin mRNA clearance experiments, prolactin mRNA was cleared at the same rate in the absence and presence of Ca2+. These results demonstrate that although Ca2+ has a small effect on the transcriptional rate of the prolactin gene, Ca2+ produces a significant increase in prolactin mRNA levels by acting at a posttranscriptional site(s). Furthermore, Ca(2+) appears to increase prolactin mRNA levels by posttranslational modification of a stable protein, probably at a nuclear site.

Regulation of cell-type-specific transcription and differentiation of the pituitary

The transcription of rat prolactin and growth hormone genes in vitro requires a pituitary transcription factor, specific to certain cell types in the pituitary, which currently appears to be the PUF-I/Pit-1/GHF-1 protein. This factor binds to cis-regulatory elements in the 5' region of both genes and exerts a positive influence on transcription initiation presumably by interacting with general transcription factors. The PUF-I/Pit-1/GHF-1 transcriptional regulatory protein probably has an important role in not only the differentiation of the pituitary lactotroph/somatotroph cell lineage; it is also expressed in the early development of the nervous system but its function there is less well documented. It appears to be one member of a family of trans-activator proteins involved in differential gene expression in several cell types.

Binding of a Drosophila POU-domain protein to a sequence element regulating gene expression in specific dopaminergic neurons

Several genes have been identified that affect the specification of unique neuronal identities during development of the Drosophila central nervous system. At least two of these genes, fushi tarazu (ftz) and even-skipped (eve) share a common structural motif, the homoeobox, which encodes a sequence-specific DNA-binding domain (homoeodomain). A family of related proteins has been recently characterized in mammals and nematodes that contain a diverged homoedomain as part of a structure referred to as a POU domain. Mammalian genes encoding POU domains have region-specific patterns of expression in the central nervous system, indicating a potential role for them in the regulation of neuronal development. The nematode POU-domain gene, unc-86, is involved in the determination of neuroblast lineages leading to serotonergic and dopaminergic neurons. We have now identified a Drosophila gene, Cfla, encoding a sequence-specific DNA-binding protein containing a highly conserved POU domain. The Cfla gene product binds to a DNA element required for expression of the dopa decarboxylase gene (Ddc) in selected dopaminergic neurons, implying that it functions as a neuron-specific transcription factor.

Characterization of transgenic livestock production

The objective of transgenic livestock improvement projects is to develop and bring to market superior breeding stock, as well as germplasm for the artificial insemination and embryo transfer industries. Livestock animal biotechnology programs hold the promise of achieving, in a single generation, improvements in commercially important livestock species previously possible only through long-term traditional selective breeding practices or by chance mutation. Transgenic farm animals harboring growth hormone or metabolically related structural genes have been created. Studies of these animals demonstrate the effects of inadequate regulation of transgene expression. Research continues to explore the intricacies of developmental regulation of such genes and phenotypic consequences of mammalian gene transfer. Ultimately, genetically engineered livestock will provide producers with the benefit of increased production efficiencies while the consumer will have healthier animal food products. Conceivably, products will be produced with lower levels of fat, cholesterol, feed additives and pharmaceutical residues from animals with altered carcass composition that will result in greater nutritional benefit for the consumer.

A two-base change in a POU factor-binding site switches pituitary-specific to lymphoid-specific gene expression

The structurally related POU homeo domain proteins Pit-1 and Oct-2 activate pituitary- and lymphoid-specific transcription, respectively, by binding to similar AT-rich motifs in their target genes. In this study we identify bases critical for recognition and activation by Pit-1 and examine how small differences in Pit-1 and Oct-2-binding sites can impart differential transcriptional responses in pituitary and B-lymphoid cells. Scanning mutagenesis of Pit-1 response elements in both the rat prolactin and growth hormone genes reveals a critical binding motif recognized in an identical manner by the native Pit-1 protein and cloned Pit-1 gene product. This motif, ATTATTCCAT, differs by only two bases from the octamer element, ATTTGCAT, required for Oct-2-dependent activation of immunoglobulin genes. Cross recognition of Pit-1 and Oct-2 sites by both factors can be demonstrated in competitive binding assays, in which an oligometric Pit-1 site from the prolactin gene is converted to an Oct-2 site by a double point mutation. In contrast to the binding data, no cross activation of transcription is detectable in cultured cell lines. When inserted immediately 5' to a prolactin TATA box, the wild-type prolactin element enhances transcription strongly in pituitary cells but is inactive in B cells, whereas the octamer variant of the prolactin site activates expression in B cells but is silent in pituitary lines. Both elements are nonfunctional in heterologous cell lines that lack Pit-1 and Oct-2.(ABSTRACT TRUNCATED AT 250 WORDS)

Dissection of functional domains of the pituitary-specific transcription factor GHF-1

The specific expression of growth hormone (GH) in the somatotrophic cells of the anterior pituitary is largely attributable to a short promoter in the 5' flanking region of the GH gene. This promoter contains two binding sites for the transcription factor GHF-1, the expression of which is also specific to cells of the somatotrophic lineage and correlates with activation of the GH gene in the developing mouse pituitary. Various studies indicate that GHF-1 is the main determinant of cell type-specific expression of the GH gene. GHF-1 is a member of the POU-domain class of proteins that each contain two highly conserved sequence motifs, the homoeodomain and the POU-specific domain. Here we report that the GHF-1 homoeodomain is sufficient for sequence-specific DNA binding, although its activity is stimulated by the POU-specific domain, which does not interact directly with the DNA. Transcriptional activation is mediated by a separate domain rich in hydroxylated amino-acid residues. Similar sequences are present in other cell type-specific transcription factors.

A regulatory hierarchy for cell specialization in yeast

The specialized sets of genes that determine different cell types in yeast are controlled by combinations of DNA-binding proteins some of which are present only in certain cell types whereas others are present in all cell types. Final differentiation requires an inductive signal that triggers both gene transcription and cell-cycle arrest. Synthesis of the proteins coded by the 'master regulatory' mating-type locus is regulated so as to generate a heterogeneous mitotic cell population containing a stem-cell lineage.

Herpes simplex virus Vmw65-octamer binding protein interaction: a paradigm for combinatorial control of transcription

The transcriptional status of a given viral or cellular gene is determined both by the availability of functional transcription factors and by the combination and spatial arrangement of the cis-acting elements to which they bind. While differential gene expression can be achieved to some extent by the interaction of different factors with different genes it is clear that in some cases the same factor is required for the regulation of genes that are not coordinately expressed. How a transcription factor achieves selective function in the absence of selective binding is a question that remains largely unanswered. The recent advances in understanding how both protein-protein and protein-DNA interactions are required to mediate the induction of herpes simplex virus immediate early (HSV IE) gene expression have highlighted a novel mechanism for combinatorial control of transcription which has significant implications for the differential control of cellular gene expression.

SCIP: a glial POU domain gene regulated by cyclic AMP

We have isolated cDNA clones encoding SCIP, a POU domain gene expressed by myelin-forming glial of the central and peripheral nervous systems. In purified Schwann cells cultured in the absence of neurons, expression of SCIP is suppressed. This suppression is relieved by cAMP, and induction of SCIP mRNA by this second messenger precedes cAMP induction of myelin-specific genes. Similarly, SCIP expression in vivo precedes full expression of myelin-specific genes in developing oligodendrocytes and Schwann cells. The sequence of the SCIP POU domain is identical to that of Tst-1, a recently identified member of a family of POU domain genes expressed by restricted subsets of neurons. Our results demonstrate that SCIP is also expressed by myelin-forming glia and suggest that it plays a central role in the progressive determination of these cells and their commitment to myelination.

The rat albumin promoter: cooperation with upstream elements is required when binding of APF/HNF1 to the proximal element is partially impaired by mutation or bacterial methylation

We have characterized in the accompanying paper (P. Herbomel, A. Rollier, F. Tronche, M.-O. Ott, M. Yaniv, and M. C. Weiss, Mol. Cell. Biol. 9:4750-4758, 1989) six different elements in the albumin promoter. One of them, the proximal element (PE), is the binding site for a strictly liver specific factor, APF/HNF1. This binding site contains a bacterial DAM DNA methylase methylation target sequence which, when methylated, decreases the affinity of the protein for this element. When the different albumin promoter constructions were prepared in an Escherichia coli deoxyadenosine methylase-negative strain, the respective contributions of the elements to the overall promoter activity were strikingly different. An intact proximal element plus the TATA box gave almost full transcriptional activity in transient transfection experiments and only in differentiated hepatoma cells of line H4II, whereas the distal elements (distal element III [DEIII], the NF1-binding site DEII, and the E/CBP-binding site DEI) had become essentially dispensable. Mutations affecting the CCAAT box showed only a two- to threefold decrease. When PE was methylated, mutated, or replaced by the homologous element from the alpha-fetoprotein gene, activity in the context of the short promoter (PE plus the TATA box) was abolished. However, activity was restored in the presence of the upstream elements, showing that cooperation with factors binding to the CCAAT box and distal elements favors the functional interaction of the liver-specific APF/HNF1 factor with lower-affinity binding sites.

Lineage-restricted expression of homeobox-containing genes in human hematopoietic cell lines

We investigated the role of homeobox-containing genes in human hematopoiesis because homeobox genes (i) control cell fate in the Drosophila embryo, (ii) are expressed in specific patterns in human embryos, and (iii) appear to function as transcription factors that control cell phenotype in other mammalian organs. Using four homeobox probes from the HOX2 locus and a previously undescribed homeobox cDNA (PL1), we screened mRNAs from 18 human leukemic cell lines representing erythroid, myeloid, and T- and B-cell lineages. Complex patterns of lineage-restricted expression are observed: some are restricted to a single lineage, while others are expressed in multiple lineages. No single homeobox gene is expressed in all types of hematopoietic cells, but each cell type exhibits homeobox gene expression. HOX2.2 and -2.3 homeobox-containing cDNAs were cloned from an erythroleukemia cell (HEL) cDNA library, while the homeobox cDNA PL1 was isolated from a monocytic cell (U-937) library. Differentiation of HEL and K-562 cells with various inducers results in modulation of specific homeobox transcripts. In addition, HOX2.2 is expressed in normal bone marrow cells. We have demonstrated (i) lineage-restricted expression of five homeobox genes in erythroid and monocytic cell lines; (ii) expression of additional homeobox genes in other cell lineages (HL-60 and lymphoid cells); (iii) expression of one homeobox gene in normal marrow cells; and (iv) modulation of expression during differentiation. These data suggest that these genes play a role in human hematopoietic development and lineage commitment.

5' flanking DNA sequences direct cell-specific expression of rat tyrosine hydroxylase

Tyrosine hydroxylase (TH) is selectively expressed in catecholaminergic neurons and in chromaffin cells of the adrenal medulla. Constructs in which 5' flanking sequences of the rat TH gene directed expression of bacterial chloramphenicol acetyltransferase (CAT) were transfected into cell lines and assayed for transient expression of CAT. In most nonexpressing cell lines, CAT levels were less than 5% of that found in a TH-positive pheochromocytoma line (PC8b). In two lines described here, a rat anterior pituitary cell line (GH4) and a rat fibroblast line (Fr3T3), CAT expression reached 12 and 20%, respectively, of the PC8b level. Greater than 90% of the PC8b activity was lost when sequences between -212 and -187 (in relation to the transcriptional initiation site) were deleted. Further deletions that removed the cyclic AMP response element (CRE) (-45) and the TATA box at -29 reduced transcriptional activity to background in all three lines. These data suggest that 212 nucleotides of the 5' sequence are sufficient for pheochromocytoma expression and that information between -212 and -187, which includes an AP1 site (-206 to -200), is essential for full transcriptional activity. In addition, sites for other protein transcription factors (AP2, POU/Oct, SP1, and CRE) reside between -221 and -38 and are largely conserved between the human and rat gene.

Binding of a pancreatic nuclear protein is correlated with amylase enhancer activity

The mouse amylase gene Amy-2.2 is expressed at high levels specifically in the acinar cells of the pancreas. The region between -172 and -110 of this gene includes sequence elements common to pancreas-specific genes. Nuclear proteins with specific affinity for this region were partially purified from rat pancreas. The consensus element of another pancreas-specific gene, elastase 1, competes for protein binding to the amylase sequences. Binding was localized by DNase I protection to the sequence -156 to -122. Site-directed mutagenesis of this sequence resulted in concomitant loss of protein binding and enhancer activity. Photo-affinity labelling of pancreatic nuclear extracts identified one predominant binding protein with a molecular weight of approximately 75 kDa. The data indicate that binding of this nuclear protein is essential for the enhancer activity of this pancreas-specific element.

The Oct-1 homoeodomain directs formation of a multiprotein-DNA complex with the HSV transactivator VP16

The herpes simplex virus transactivator VP16 participates in the formation of a multiprotein-DNA complex with the ubiquitous octamer-motif-binding factor Oct-1. Complex formation is dependent on specific amino acids in the Oct-1 homoeodomain which are in positions analogous to positive control mutations in helix 2 of the lambda phage repressor helix-turn-helix motif, indicating that this structure is an ancient target for protein-protein interactions mediating transcriptional control.

The liver-specific transcription factor LF-B1 contains a highly diverged homeobox DNA binding domain

The nuclear protein LF-B1 (also referred to as HNF-1) is a transcription activator required for the expression of several liver-specific genes. LF-B1 has been purified to homogeneity from rat liver nuclear extracts. The sequence of the protein has been partially determined and, subsequently, overlapping cDNA clones containing the entire open reading frame of LF-B1 were isolated. The full-length cDNA encodes a 628 amino acid protein and directs the synthesis in vitro of a protein capable of binding DNA with the same specificity as LF-B1. The cDNA was recombined into a vaccinia virus vector and active LF-B1 was obtained from infected HeLa cells. Addition of the vaccinia recombinant protein to rat spleen extracts results in activation of transcription of an LF-B1-dependent promoter. The DNA binding domain of LF-B1 is located in the amino-terminal part of the protein and displays distant structural similarity to the homeobox domain. The distribution of LF-B1 mRNA is restricted to liver, which correlates with the tissue-specific expression of its target genes.

Dexamethasone control of growth hormone mRNA levels in GH3 pituitary cells is cycloheximide-sensitive and primarily posttranscriptional

To clarify the mechanism of growth hormone (GH) gene activation by glucocorticoids in GH3 pituitary cells, GH mRNA accumulation in nuclear and cytoplasmic compartments was measured in the presence and absence of cycloheximide. In dexamethasone-treated cells, levels of GH mRNA were increased in the nucleus by 6 h and in the cytoplasm by 12 h. Dexamethasone treatment caused a 5- to 24-fold rise in total GH mRNA levels by 48-72 h. The differential elevation of nuclear levels of GH mRNA relative to the amount of cytoplasmic GH mRNA persisted for 48 h. A transient accumulation of GH mRNA in the nucleus was followed by a brief rise in cytoplasmic GH mRNA levels in GH3 cells treated simultaneously with dexamethasone and cycloheximide. In GH3 cells pretreated for 2 h with cycloheximide, the rise in nuclear and cytoplasmic GH mRNA levels mediated by dexamethasone was blocked completely. Levels of glucocorticoid receptor were unaffected by cycloheximide. These data suggest that the stimulation of GH mRNA levels by glucocorticoids is initiated within the nucleus and that cycloheximide-sensitive events are essential for this stimulation to occur. To assess the importance of GH gene transcriptional activation by glucocorticoids, nuclear transcription run-on reactions and assays of GH promoter activity in an aminoglycoside 3'-phosphotransferase (Neo) fusion gene within stably transformed GH3 cells were performed. Evidence for a weak, transient transcriptional activation of the GH gene by dexamethasone in nuclear run-on assays was obtained. Consistent with this idea, a 30-72 h exposure to dexamethasone raised levels of Neo mRNA in GH-Neo GH3 cell transformants by less than or equal to 2-fold. We conclude that glucocorticoid stimulation of GH mRNA in GH3 cells requires ongoing protein synthesis and can occur largely independently of GH gene transcriptional activation.

A protein that binds to a cis-acting element of wheat histone genes has a leucine zipper motif

The structure and function of transcription factors of higher plants was studied by isolating cDNA clones encoding a wheat sequence-specific DNA binding protein. A hexameric nucleotide motif, ACGTCA, is located upstream from the TATA box of several plant histone genes. It has been suggested that this motif is essential for efficient transcription of the wheat histone H3 gene. A wheat nuclear protein, HBP-1 (histone DNA binding protein-1), which specifically binds to the hexameric motif, has previously been identified as a putative transcription factor. A cDNA clone encoding HBP-1 has been isolated on the basis of specific binding of HBP-1 to the hexameric motif. The deduced amino acid sequence indicates that HBP-1 contains the leucine zipper motif, which represents a characteristic property of several eukaryotic transcription factors.

A purified Drosophila homeodomain protein represses transcription in vitro

even-skipped (eve) is a homeodomain-encoding gene that is a genetically defined repressor of Ultrabithorax (Ubx), fushi-tarazu (ftz), and wingless (wg). Here we report that purified eve protein represses transcription in vitro at the Ubx promoter, in a DNA binding site-dependent manner. eve protein represses transcription when bound either upstream or downstream of the RNA start site or when DNA binding sites are in either orientation. We also show that eve represses expression from the Ubx promoter in Drosophila tissue culture cells, again in a binding site-dependent manner. Deletion of eve DNA binding sites does not alter transcription in the absence of eve, and so repression is not likely to be the result of eve competitively inhibiting an activator protein from binding to the same DNA element. Instead, we propose that eve protein is probably interfering with the function of proteins bound at other locations in the promoter. The biochemical demonstration that a Drosophila homeodomain protein can directly regulate RNA synthesis strengthens the view that this class of regulators act as transcription factors to control development.

Transcriptional regulation in mammalian cells by sequence-specific DNA binding proteins

The cloning of genes encoding mammalian DNA binding transcription factors for RNA polymerase II has provided the opportunity to analyze the structure and function of these proteins. This review summarizes recent studies that define structural domains for DNA binding and transcriptional activation functions in sequence-specific transcription factors. The mechanisms by which these factors may activate transcriptional initiation and by which they may be regulated to achieve differential gene expression are also discussed.

Inhibition of rat growth hormone promoter activity by site-specific DNA methylation

The effect of methylation on rat growth hormone (rGH) promoter activity was determined in GH3 cells by measuring rGH-Neo and rGH-CAT fusion gene expression with or without prior in vitro treatment with the site-specific DNA methyltransferases, M-BsuE and M-HhaI. To assay for rGH-promoter-specific effects of DNA methylation, RSV-Neo and RSV-CAT activities with or without M-BsuE, M-HhaI and M-HpaII treatment were measured in parallel cultures of GH3 cells. GH1-Neo and RSV-Neo fusion gene expression was inhibited by in vitro methylation from 44 to 83% as measured by the number of Geneticin-resistant GH3 cell colonies. Methylation of the GH1 promoter by M-BsuE exhibited some selective inhibition of Neo expression as determined by colony numbers, although extensive methylation of non-promoter DNA in GH1-Neo and RSV-Neo by M-HhaI and M-HpaII also inhibited Neo expression. Southern blot analysis of genomic DNA isolated from the Geneticin-resistant GH3 cells indicated that Geneticin-resistance was accompanied by demethylation of the BsuE (ThaI) sites in stably incorporated GH1-Neo DNA but not RSV-Neo DNA. Transient expression of the CAT gene in GH3 cells was selectively inhibited by 60% upon methylation of two BsuE (ThaI) sites in the GH1 promoter of GH1-CAT by M-BsuE. The data demonstrate, for the first time, to our knowledge, a direct effect of DNA methylation on the activity of the rat growth hormone promoter.

Expression of a large family of POU-domain regulatory genes in mammalian brain development

A novel region referred to as the POU-domain is present in two tissue-specific transcription factors, Pit-1 and Oct-2, that activate expression of genes specifying pituitary and lymphocyte phenotypes. We report the identification of multiple new members of a large family of POU-domain genes expressed in adult brain, and document that all the known mammalian POU-domain genes, including Pit-1 and Oct-2, are expressed widely in the developing nervous system.

A pituitary POU domain protein, Pit-1, activates both growth hormone and prolactin promoters transcriptionally

The anterior pituitary gland provides a model for investigating the molecular basis for the appearance of phenotypically distinct cell types within an organ, a central question in development. The rat prolactin and growth hormone genes are expressed selectively in distinct cell types (lactotrophs and somatotrophs, respectively) of the anterior pituitary gland, reflecting differential mechanisms of gene activation or restriction, as a result of the interactions of multiple factors binding to these genes. We find that when the pituitary-specific 33-kD transcription factor Pit-1, expressed normally in both lactotrophs and somatotrophs, is expressed in either the heterologous HeLa cell line or in bacteria, it binds to and activates transcription from both growth hormone and prolactin promoters in vitro at levels even 10-fold lower than those normally present in pituitary cells. This suggests that a single factor, Pit-1, may be capable of activating the expression of two genes that define different anterior pituitary cell phenotypes. Because a putative lactotroph cell line (235-1) that does not express the growth hormone gene, but only the prolactin gene, appears to contain high levels of functional Pit-1, a mechanism selectively preventing growth hormone gene expression may, in part, account for the lactotroph phenotype.

Cell-specific expression of the prolactin gene in transgenic mice is controlled by synergistic interactions between promoter and enhancer elements

Prolactin gene expression is restricted to the lactotrophic and somatomammotrophic cells of the anterior pituitary. In transgenic mice, a fusion gene consisting of 3 kb of prolactin 5'-flanking region fused to a firefly luciferase or human growth hormone (hGH) reporter gene is expressed at high levels with the strict tissue and cell-type specificity that is characteristic of the endogenous prolactin gene. High levels of expression require two cis-acting regions: a distal enhancer (-1.8 to -1.5 kb) and a proximal region (-422 to +33 bp). Each of these regions alone can direct low levels of fusion gene expression to prolactin-producing cell types in transgenic mice, but a synergistic interaction between these regions is necessary for high levels of expression. The ontogeny of the prolactin transgene expression closely follows the appearance of high levels of a POU homeo-domain transcription factor, Pit-1, that has been shown previously to bind structurally related sequences in both the distal enhancer and proximal regions and to activate the expression of the prolactin gene in vitro. Unexpectedly, transgenes containing the distal enhancer removed from its normal context are expressed in both the prolactin-producing lactotrophs and the thyroid-stimulating hormone (TSH)-producing thyrotrophs, thereby suggesting that sequences flanking this enhancer are necessary to restrict expression to the correct cell type within the pituitary. These data indicate that distinct processes of gene activation and restriction are necessary for the fidelity of cell-type-specific expression within an organ.

Transcriptional activation by the Antennapedia and fushi tarazu proteins in cultured Drosophila cells

Drosophila homeodomain proteins bind to specific DNA sequences in vitro and are hypothesized to regulate the transcription of other genes during development. Using a cotransfection assay, we have shown that homeodomain proteins encoded by the homeotic gene Antennapedia (Antp) and the segmentation gene fushi tarazu, as well as a hybrid homeodomain protein, are activators of transcription from specific promoters in cultured Drosophila cells. Sequences downstream of the Antp P1 and Ultrabithorax transcription start sites mediate the observed activation. A TAA-rich DNA sequence to which the Antp protein binds in vitro is sufficient to confer regulation on a heterologous promoter. The results demonstrate that homeodomain proteins are transcriptional regulators in vivo and that in cultured cells, different homeodomain-containing proteins can act upon a common sequence to modulate gene transcription.

Cloning of cDNA for the major DNA-binding protein of the erythroid lineage through expression in mammalian cells

Genes expressed in erythroid cells contain binding sites for a cell-specific factor believed to be an important regulator for this haematopoietic lineage. Using high-level transient expression in mammalian cells, we have identified complementary DNA encoding the murine protein. The factor, a new member of the zinc-finger family of DNA-binding proteins, is restricted to erythroid cells at the level of RNA expression and is closely homologous between mouse and man.

Cell-specific transcriptional control of the mouse DNA-binding protein mC/EBP

The mRNA encoding the mouse homolog of C/EBP, a rat DNA-binding protein that participates in activating a number of genes in hepatocytes, is present in liver cells at a far higher concentration than in most other cells, including spleen, kidney, muscle, and the majority of the brain. However, fat cells and intestinal cells contain 25-50% as much mRNA as liver cells. "Run-on" experiments show that the basis for the restricted cellular distribution of the mouse C/EBP mRNA is transcriptional regulation of the gene. We also show that disruption of cell-cell contacts incident to liver cell dispersion results in a prompt and extensive reduction in mouse C/EBP transcription as we had earlier shown to be the case for a group of 10 genes transcribed in a hepatocyte-specific fashion. In contrast, breaking cell contacts and plating the hepatocytes in culture leads to a prolonged increase in transcription of the Jun-B gene that encodes a widely distributed transcription factor. These results illustrate that the regulation of expression of a mammalian regulatory protein with limited tissue distribution is controlled at the level of transcription and depends on cell contacts.