Search

Nuclear factor kappa B (NF-kappa B) is a ubiquitous transcription factor which binds to decameric DNA sequences (kappa B sites) and regulates transcription of multiple genes. The activity of NF-kappa B is regulated by an inhibitor protein, I kappa B, which sequesters NF-kappa B in the cytoplasm. Release of I kappa B and subsequent nuclear translocation of NF-kappa B generally require activating signals. However, in mature murine B cells, the DNA-binding activity of NF-kappa B is constitutively nuclear and activates the Ig kappa gene, a marker for mature B cells. To understand the basis for the constitutive NF-kappa B activation, we examined the properties of NF-kappa B and I kappa B in both pre-B and mature B cells, the regulated and constitutive states, respectively. We found that expression of I kappa B alpha and p105, members of the I kappa B family, and Rel, a member of the NF-kappa B family, is augmented in mature B cells. Both I kappa B alpha and p 105 are associated with NF-kappa B proteins and sequester most of these proteins in the cytoplasm of mature B cells. However, rapid I kappa B alpha dissociation and degradation lead to continuous nuclear translocation of a small fraction of NF-kappa B proteins, which represent the constitutively active NF-kappa B in mature B cells. We estimate that the protease activity is at least 35-fold greater in mature B cells than in pre-B cells. Rapid degradation of I kappa B alpha is directly involved in constitutive NF-kappa B activation, because stabilization of I kappa B alpha by a protease inhibitor causes loss of NF-kappa B activity in mature B cells. These results provide evidence that continuous and rapid degradation of I kappa B alpha in the absence pf external stimuli is causally involved in the constitutive activation of NF-kappa B in mature murine B cells.

Transcription factor NF-kappa B regulates the expression of a plethora of genes. The activity of NF-kappa B proteins is regulated by I kappa B proteins. We report that induction of I kappa B alpha, a member of the I kappa B family of proteins, is preceded by activation of NF-kappa B complex. The promoter of the I kappa B alpha gene contains a kappa B site that is directly involved in its induction by the NF-kappa B complex. Degradation of I kappa B alpha protein precedes activation of NF-kappa B DNA binding activity, whereas newly synthesized I kappa B alpha protein inhibits NF-kappa B activity. If the degradation of I kappa B alpha is prevented, the induction of DNA binding activity of NF-kappa B complex is severely curtailed. These data suggest the existence of an autoregulatory loop whereby I kappa B alpha regulates the activity of transcription factor NF-kappa B, which in turn regulates the I kappa B alpha activity.

DNA-binding complexes which are temporally regulated in developing rat brain have been identified by their ability to interact with the kappa B enhancer sequence in electrophoretic mobility-shift assays. These complexes, referred to as developing-brain factors (DBFs) 1 and 2, are abundant in nuclear extracts from developing rat brain through postnatal day 2 and decline to nearly undetectable levels by postnatal day 7. DBFs were not detected in extracts from cultured cell lines or in tissues other than the developing brain. The highest level of DBF DNA-binding activity was observed in developing cortex, and the lowest in cerebellum. In UV crosslinking experiments, a labeled kappa B oligonucleotide probe crosslinked 110- and 115-kDa proteins from DBF complexes. DBFs are likely to be involved in the regulation of transcriptional events which take place during brain development.

The Fos family of proteins now includes seven members: the retroviral proteins FBR-v-Fos and FBJ-v-Fos and the cellular proteins c-Fos, FosB, FosB2, Fra1, and Fra2. Four proteins (FBR-v-Fos, FBJ-v-Fos, c-Fos, and FosB) transform established rodent fibroblast cell lines, while three (FosB2, Fra1, and Fra2) do not. As all family members display sequence-specific DNA-binding activity as part of a heterodimeric complex with Jun proteins, other features must account for the differences in transforming potential. We demonstrate here that all transforming members have a C-terminal transactivation domain that is lacking in nontransforming members. The nontransforming proteins Fra1 and Fra2 can be converted to transforming proteins by fusion of a transactivation domain from either FosB or VP16. We also demonstrate that differences in the basic region-leucine zipper domain affecting either the affinity or sequence specificity of DNA binding are not determinants of the difference in transforming potential among members of the Fos family. The results further define the functional requirements for transformation by Fos proteins and suggest that the subunit composition of AP1 complexes is an important determinant of mitogenic signalling capability.

The c-Rel protein is able to associate in vitro and in vivo with the TATA-binding protein (TBP) of the TFIID complex. Coexpression of TBP with c-Rel augments transactivation from the kappa B site in Drosophila Schneider cells. DNA-binding mutants of TBP not only fail to cooperate, but they repress transactivation by c-Rel. There may be a direct communication between kappa B enhancer binding proteins and basal transcription factors which leads to enhanced transcription.

Nerve growth factor (NGF) is the best understood of a class of trophic proteins that are important for the survival of neurons and the elaboration of their characteristic processes. Here we demonstrate that RINm5F, a rat insulinoma cell line representing an early stage in pancreatic beta cell differentiation, expresses both the Trk and p75 NGF receptors and responds to NGF by extending neurite-like (neurofilament-containing) processes. NGF treatment of RINm5F cells also induces the expression of genes normally responsive to NGF in neurons, including the NGF-1A gene. Inasmuch as pancreatic beta cells arise from the embryonic endoderm, these results suggest that NGF may play a wider role during development than previously thought-a role not restricted to cells of neuroectodermal origin--and that endocrine and neuronal cells share a developmental pathway. The specific effect of NGF on an early pancreatic beta cell line also suggests that this neurotrophic factor might form the basis of a therapeutic treatment for some types of diabetes by inducing the proliferative differentiation of islet cells.

Overexpression of some members of the Fos gene family, including FosB, leads to transformation of established rodent fibroblasts. We have previously shown that transformation by FosB requires the presence of a C-terminal transcriptional activation domain. We now report that transformation by FosB also requires an intact DNA-binding domain composed of the functionally bipartite basic region and leucine zipper as well as sequences present in the N terminus that serve a regulatory function. Deletion of the N-terminal sequences results in proteins impaired in transcriptional activation and transformation. This region does not itself function as a transcriptional activation domain but instead regulates the transactivation functions present in the FosB-Jun complex. The requirement for this N-terminal region can be abolished by the presence of a strong constitutive activation domain. The primary sequence of the region that we have defined is highly conserved in the Fos family of proteins, suggesting functional conservation.

The activity of the rel/NF-kappa B/dorsal family of kappa B site binding proteins is regulated by I kappa B proteins. The ankyrin repeat motif identified I kappa B family members, which include I kappa B alpha (pp40/MAD-3), I kappa B gamma, and bcl3, directly associated with kappa B site binding proteins, resulting in specific DNA-binding inhibition of rel, p50, or p65 dimers. We report that I kappa B gamma, containing eight ankyrin repeats, mediates a reversible inhibition of (p50)2-DNA complex but cannot displace preformed DNA-protein complexes. I kappa B alpha and bcl3, on the other hand, can displace preformed DNA-protein complexes. I kappa B alpha specifically displaces (p65)2 or p50/p65-DNA complexes but requires the C-terminal 37 amino acids in addition to the ankyrin repeat domain. Human bcl3 protein specifically displaces (p50)2-DNA complexes. Because I kappa B alpha and bcl3 can displace preformed (p65)2 or (p50)2-DNA complexes, respectively, we propose that they can act as repressors or antirepressors of NF-kappa B-induced gene expression.

Fos and Jun oncoproteins form a complex that regulates transcription from promoters containing AP-1 binding sites. These two proteins, like other transcriptional activators, are likely to stimulate transcription through direct and/or indirect interactions with members of the basal transcriptional machinery. The ability of c-Fos and c-Jun proteins to interact directly with the TATA box-binding protein (TBP), the general transcription factor required for initiating the assembly of transcription complexes, was investigated. Using co-immunoprecipitation and protein-protein association assays, we show that both c-Fos and c-Jun bind specifically and stably to TBP. Mutational analysis demonstrates that both the basic region and leucine zipper domains of c-Fos and c-Jun are necessary and sufficient for stable association with TBP. A 51-residue region from the conserved C-terminal region of TBP, previously shown to be the binding site for the viral activator protein E1A, interacts with c-Fos and c-Jun proteins. We propose that c-Fos and c-Jun proteins function as transcriptional activators, in part by recruiting TBP to form complexes to initiate RNA synthesis.

The bcl-3 gene product, overexpressed in chronic lymphocytic leukemia (CLL) patients with the translocation t(14;19), is a member of the I kappa B family. The bcl-3 protein is able to inhibit the DNA binding and trans-activation of authentic NF-kappa B heterodimers p50-p65 and p49-p65, as well as p50 and p49 homodimers. The bcl-3 protein does not inhibit either the DNA-binding activity of the Rel protein or its ability to trans-activate genes linked to the kappa B site. A human 37-kD protein (I kappa B alpha), identified previously as a member of the I kappa B family, is also unable to inhibit DNA-binding activity of the Rel protein. However, unlike bcl-3, the 37-kD (I kappa B alpha) protein has no effect on the DNA-binding activity of p50 or p49 homodimers. Two dimensional phosphotryptic peptide maps of the human bcl-3 and the human 37-kD (I kappa B alpha) proteins reveal that the phosphopeptides from the 37-kD (I kappa B alpha) protein are nested within the bcl-3 protein. Furthermore, bcl-3 antisera immunoprecipitates an in vitro-radiolabeled 37-kD (I kappa B alpha) protein. Proteins of 56 and 38 kD can be identified in HeLa cells stimulated with PMA and immunoprecipitated with bcl-3 antisera. Comparison of tryptic peptide maps of the bcl-3 protein synthesized in vitro, and p56 and p38 from HeLa cells, shows that they are all structurally related. Removal of the amino-terminal sequences of the bcl-3 protein generates a protein that inhibits the DNA binding of the p50-p65 heterodimer but, like the 37-kD (I kappa B alpha) protein, is no longer able to inhibit the binding of the p50 and p49 homodimers with kappa B DNA. We propose that the bcl-3 and 37-kD (I kappa B alpha) proteins are related and are members of the I kappa B family.

We have explored the use of primary myoblasts as a somatic tissue for gene therapy of acquired and inherited diseases where systemic delivery of a gene product may have therapeutic effects. Mouse primary myoblasts were infected with replication-defective retroviruses expressing canine factor IX cDNA under the control of a mouse muscle creatine kinase enhancer and human cytomegalovirus promoter. The infected myoblasts were injected into the hindlegs of recipient mice and levels of secreted factor IX protein were monitored in the plasma. We report sustained expression of factor IX protein for over 6 months without any apparent adverse effect on the recipient mice.

Retroviral vectors provide a safe and efficient method of introducing genes of therapeutic interest into dividing cells. The principle limitation of these vectors in the past has been poor gene expression in vivo. This problem has been overcome recently through the use of tissue-specific enhancers in commonly used retroviral vectors. In this review we discuss both the relevant biology and some of the practical applications of retroviral vectors in gene therapy.

Grafts of various types of cells have been performed using bioresorbable polymer matrices. These synthetic fibers are degraded by hydrolysis into normal metabolic intermediates and induce a number of events that are conductive to healing and/or repair, the most important of which may be angiogenesis. The use of biodegradable meshes to deliver genetically altered cells was studied. A beta-galactosidase gene was inserted into Long-Evans rat bone marrow stromal (BMS) cells or fibroblasts derived from C57BL/6J mouse embryos using the retroviral vector LNL-SLX beta gal. Expression was monitored using X-gal staining. X-gal+ cells from monolayer cultures were seeded onto either polyglycolic acid (PGA) or polyglactin (PGL) biodegradable meshes and grown to confluence. Two types of grafts were performed: (1) embryonic C57BL/6J mouse fibroblasts (EMF) into either nude mice or adult C57BL/6J mice, and (2) Long-Evans rat BMS into Long-Evans rats. Beta-Galactosidase activity was found for up to 152 days for EMF in nude mice, 123 days for EMF in adult C57BL/6J mice, and 90 days for grafts of syngeneic BMS cells into Long-Evans rats. Noninfected cells grafted using the same methods did not stain with X-gal.

Mutants in the leucine zipper and basic regions of mouse c-jun were tested for transformation in chicken embryo fibroblast cultures. Reduction or elimination of the ability of Jun to dimerize or to bind to DNA severely decreased transformation. A chicken v-jun gene from which the major transactivation domain was deleted also failed to transform. We conclude that an intact leucine zipper, basic region and transactivation domain are required for Jun-induced oncogenic transformation. Coexpression of chicken c-Fos increased formation of transformed foci by Jun proteins of moderate to low oncogenic potency but had no effect on highly transforming Jun. Chicken c-Fos could also transform chicken embryo fibroblasts on its own, albeit after prolonged culture and at a low efficiency.

Fos and jun heterodimers activate the transcription of genes containing an AP-1 site. The activity of Fos and Jun proteins is regulated by post-translational modification. The activity of the rel/NF-kappa family of transcriptional factors is regulated by their sequestration in the cytoplasm in association with the inhibitor protein, I kappa B. An ankyrin repeat motif in I kappa B proteins is required for their direct association with rel/NF-kappa B.

To understand the mechanism by which pp40/I kappa B beta inhibits DNA binding activity of the rel/NF-kappa B family of transcription factors, we have investigated the role of ankyrin repeats on the biological function of pp40 by deleting or mutating conserved residues. We show that (i) ankyrin repeats alone are not sufficient to manifest biological activity but require the C-terminal region of the pp40 protein; (ii) four out of the five ankyrin repeats are essential for inhibiting the DNA binding activity; (iii) pp40 mutants that do not inhibit DNA binding of rel protein also do not associate with rel; (iv) although pp40 can associate with the p65 and p50 subunits of NF-kappa B, pp40 inhibits the DNA binding activity of only the p50-p65 heterodimer and the p65 homodimer; and (v) pp40 inhibits the transcription of genes linked to kappa B site; however, mutants that do not affect DNA binding have no effect. We propose that the ankyrin repeats and the C-terminal region of pp40 form a structure that associates with the rel homology domain to inhibit DNA binding activity.

We have used retroviral vectors to introduce human or canine factor IX cDNAs into cultured primary murine and canine myoblasts and into the established murine myoblast cell line C2C12. In all cases, the stably infected cells produced biologically active factor IX in culture and secreted detectable amounts into the culture medium both before and after differentiation of the cells into myotubes. Myoblasts and differentiated myotubes are therefore capable of performing all the posttranslational modifications of the coagulation factor required for biological activity. We have grafted the genetically modified myoblasts into skeletal muscles of nude mice and have detected stable levels of circulating human factor IX for up to two months after grafting. We propose that grafting genetically modified primary myoblasts or established myoblast cell lines into skeletal muscle may represent a useful approach to gene therapy for a variety of genetic diseases, including intrinsic muscle disease and defects in circulating proteins as in the hemophilias.

Two functionally distinct proteins derived from the FosB gene by alternative splicing have recently been described. FosB protein transforms fibroblasts efficiently, whereas FosB2 protein, a carboxy-terminally truncated form of FosB, does not, despite the fact that both proteins can participate in high-affinity, sequence-specific DNA binding as part of a heterodimeric complex with c-Jun protein. We show here that the functional difference between these proteins is the result of the presence of a potent proline-rich transcriptional activation domain in the carboxy-terminal amino acids unique to FosB. This conclusion is supported by three lines of evidence: (1) Mutations in the carboxy-terminal region of FosB that impair transcriptional activation also reduce transforming potential, despite the fact that DNA binding as part of a complex with c-Jun is not affected; (2) the carboxy-terminal region unique to FosB functions as an activation domain when fused to the DNA-binding domain of GAL4; and (3) transforming potential can be conferred on FosB2 by fusing any of several different well-characterized trans-activation domains. These results identify an additional functional requirement for transformation by Fos proteins and have implications for the mechanism(s) of mitogenic signaling by the AP-1 transcription complex.

A cDNA corresponding to the 2.6 kb NF-kappa B mRNA species present in a variety of lymphoid cell lines has been molecularly cloned. The deduced 607 amino acid sequence is identical to the sequence of the C-terminal region of 110 kd NF-kappa B protein. A 70 kd protein can be identified in lymphoid cells using antibodies raised against the C-terminal region of p110 NF-kappa B. Comparison of the two-dimensional tryptic peptide maps of the 70 kd protein expressed in cells and the in vitro translated product encoded by the cDNA display extensive homology. The 70 kd protein expressed in bacteria prevents sequence-specific DNA binding of p50-p65 NF-kappa B heterodimer, p50 homodimer, and c-rel. p70 also interferes with transactivation by c-rel and prevents its nuclear translocation. The 70 kd protein, predominantly found in lymphoid cells, is a new member of the I kappa B family of proteins and is referred to as I kappa B gamma.

We have studied the phosphorylation of the nuclear oncoprotein Fos by cyclic AMP-dependent protein kinase (PKA). We demonstrate that the human c-Fos protein, phosphorylated either in vitro with purified PKA or in vivo in JEG3 cells following treatment with forskolin, has similar phosphotryptic peptide maps. Serine 362, which constitutes part of a canonical PKA phosphorylation site (RKGSSS), is phosphorylated both in vivo and in vitro. A mutant of Fos protein in which serine residues 362 to 364 have been altered to alanine residues is not efficiently phosphorylated in vitro. Furthermore, Fos protein in which serines 362 to 364 have been altered to alanine shows increased transforming potential. We propose that phosphorylation of Fos by PKA is an important regulatory step in controlling its activity in normal cell growth and differentiation.

We have used gene transfer vectors derived from a replication-defective mutant of herpes simplex virus type 1 (HSV-1) expressing the hepatitis B virus surface antigen (HBsAg), Escherichia coli beta-galactosidase (beta-gal), or canine factor IX (cFIX) from the immediate early promoter of human cytomegalovirus (hCMV) to infect mouse liver by direct injection or through the portal vein. By either route, high levels of transgene expression were demonstrated by the detection of immunoreactive HBsAg or cFIX in the circulation and by histochemical detection of beta-gal activity in situ. The results were striking in that the serum level of cFIX reached 10% of the normal murine levels. Although the level of transgene expression from the hCMV promoter was transient, a significant number of persistent vectors could be rescued from the livers of recipient mice up to 2 months after inoculation. Replacement of the hCMV promoter with the HSV-1 latency-associated transcript (LAT) promoter resulted in reduced but prolonged expression of both HBsAg and cFIX. The very high level of factor IX expression suggests that clinically useful gene transfer may eventually be feasible through direct vector delivery to the liver.

The product of the proto-oncogene Jun inhibits myogenesis. Constitutive expression of Jun in myoblasts interferes with the expression and the function of MyoD protein. In transient transfection assays Jun inhibits transactivation of the MyoD promoter, the muscle creatine kinase enhancer, and a reporter gene linked to MyoD DNA-binding sites. Conversely, MyoD suppresses the transactivation by Jun of genes linked to an AP-1 site. We demonstrate that both in vivo and in vitro MyoD and Jun proteins physically interact. Mutational analysis suggests that this interaction occurs via the leucine zipper domain of Jun and the helix-loop-helix region of MyoD.

The product of the proto-oncogene fos is a nuclear phosphoprotein. We have investigated if nuclear location of Fos protein is mediated by a nuclear targeting sequence. We show that the cytoplasmic chicken pyruvate kinase protein translocates to the nucleus if fused to the 22 amino acid Fos basic region (amino acids 139-161). However, the nuclear localization of the Fos protein is not affected if the basic region is deleted. We propose that Fos protein contains multiple nuclear targeting sequences which may act independently or in concert with one another.

The product of proto-oncogene Rel associates with a number of cellular proteins. We have studied the effect of one of them, a phosphoprotein of 40 kD (pp40), on the DNA-binding activity of the Rel protein. We demonstrate that purified pp40 not only inhibits the binding of Rel, but also NF-kappa B (p50-p65) heterocomplex to DNA. Additionally, I kappa B beta, but not I kappa B alpha, also prevented the binding of Rel to the kappa B site. I kappa B beta and pp40 are related proteins because (1) they share a number of common tryptic peptides, (2) their inhibitory effect on DNA binding can be abolished by preincubation with pp40-specific antiserum, and (3) labeled I kappa B beta can be immunoprecipitated with pp40 antibodies. pp40 is part of the Rel complex present in the cytoplasm and nuclear extracts of WEHI-231 cells. The activity of pp40 to inhibit the DNA binding of Rel and NF-kappa B is modulated by phosphorylation.

We present evidence that retinoic acid can down-regulate transcriptional activation by the nuclear protooncogene c-jun. All three members of the retinoic acid receptor (RAR) subfamily (RAR alpha, RAR beta, and RAR gamma) can repress transcriptional induction of the human collagenase gene or a heterologous promoter that contains the collagenase promoter AP-1-binding site. In contrast, the retinoid X receptor fails to repress Jun/AP-1 activity, demonstrating a significant difference between the two regulatory systems through which retinoids exert their transcriptional control. Analysis of RAR alpha mutants in transfection studies reveals that the DNA-binding domain is important for the inhibition of Jun/AP-1 activity, even though the RAR does not bind the collagenase AP-1 site. Rather, gel-retardation assays reveal that bacterially expressed full-length RAR alpha inhibits binding of Jun protein to target DNA. These data suggest that the RAR alpha may form a nonproductive complex with c-Jun and provides a simple mechanisms by which retinoic acid may limit cell growth and possibly malignant progression.

Two forms of FosB transcript and their products can be identified in mouse NIH 3T3 cells following serum induction. The larger RNA codes for a 338-amino acid protein, whereas the smaller RNA results from the removal of an additional 140 nucleotides from FosB mRNA by alternative splicing. This alternative splicing event places a stop codon following the "leucine zipper" region and results in a shorter protein (FosB2) of 237 amino acids that lacks 101 amino acids at the carboxyl terminus. FosB2 is able to form heterodimers with c-Jun and bind to an AP-1 site but is not able to activate the transcription of promoters containing AP-1 sites. Furthermore, FosB2 can not only suppress the transcriptional activation by c-Fos and c-Jun of promoters containing an AP-1 site but also interferes with the transforming potential of viral and cellular Fos proteins. We propose that FosB2 protein functions as a trans-negative regulator.

Toward the goal of gene therapy, we have been attempting to establish model somatic cell systems with the potential of sustained expression of the foreign gene. We report here that long-term expression of foreign genes in mouse embryo fibroblast implants can be achieved if a housekeeping gene promoter is used to drive transcription. Specifically, we have shown that in implants containing a beta-galactosidase gene linked to either an immediate early promoter of cytomegalovirus or a dihydrofolate reductase (DHFR) gene promoter, only the DHFR promoter allows long-term expression. We propose that choice of the promoter manifests significant influence on the long-term expression of genes introduced in fibroblast implants by retroviral vectors.

Somatic cell gene therapy for the correction of many human genetic diseases is now technically possible. We review several methods of gene transfer that have been successfully used in animal studies, and discuss the promise and potential limitations of these methods in the treatment of human genetic diseases.

We show that the product of the protooncogene c-rel is a constituent of an NF-kappa B-like complex that binds to the kappa B site originally identified in the enhancer of immunoglobulin kappa light chain gene. c-rel protein synthesized in bacteria binds to the kappa B site in a sequence-specific manner. The rel-kappa B complex can be disrupted by incubation with anti-rel antibodies. The rel protein can form oligomers. The c-rel protein can activate transcription from promoters containing kappa B sites; v-rel, on the other hand, suppresses the transcription of genes linked to kappa B sites. Thus, v-rel may interfere with the normal transcriptional machinery of the cell by acting as a dominant negative mutant.

We have studied the regulation of protooncogene fos following serum induction. We show that un- or hypo-phosphorylated form of transcription factor cyclic AMP response element binding (CREB) protein represses the transcription of fos promoter. The negative regulation by CREB is alleviated if it is phosphorylated at serine 133 by the catalytic subunit of protein kinase A (PKA). A DNA binding mutant of CREB is unable to suppress transcription of the fos promoter. However, mutation in the cyclic AMP responsive element (CRE) at -60 or AP-1 binding site at -290, known to bind to CREB, does not appear to be involved in repression. Serum induction of dyad symmetry element (DSE) linked reporter gene is also repressed by unmodified CREB, which can be relieved following phosphorylation by PKA. We propose that posttranslational modification of CREB regulates serum inducibility of fos promoter.

Proto-oncogene fos encodes a nuclear phosphoprotein of 380 amino acids that can modulate the transcription of other genes either by transactivation or by transrepression. The v-Fos protein (381 amino acids) shares the first 332 amino acids with the c-Fos protein (with five single amino-acid changes), but differs at the C terminus. We have previously reported that the c-Fos protein undergoes more extensive post-translational modification than v-Fos (refs 9, 10). The major modification of the c-Fos protein involves serine phosphoesterification of sites in the extreme C terminus. We therefore argued that modification of the C-terminal region of the c-Fos protein may be involved in its ability to transrepress transcription without compromising its ability to transactivate other genes. Here we show that mutant c-Fos protein which is hypophosphorylated at its C terminus is unable to repress transcription of the c-fos promoter following induction with serum or tetraphorbol acetate. The C-terminal phosphorylation-deficient mutant is, however, fully competent to activate transcription of promoters containing a phorbol response element. The requirement for phosphorylation can be offset by the introduction of a net negative charge in the C terminus of the Fos protein.

Phenotypic revertants of Finkel-Biskis-Riley (FBR)-murine sarcoma virus-transformed rat fibroblasts were isolated on the basis of their adherence to plastic tissue culture dishes in the absence of divalent cations. Some revertants had sustained deletions or inactivating mutations of the v-fos gene. However, two revertants expressed a functional v-fos gene at levels equal to that in the transformed parental cells, and therefore phenotypic reversion was due to mutations in nonviral genes. These revertants were considered nontransformed according to four criteria: (i) they were flat and had a nontransformed morphology, (ii) they were contact inhibited when grown to confluence, (iii) they did not display anchorage-independent growth in soft agar, and (iv) they did not form tumors in nude mice. Somatic-cell hybrids between the revertants and the transformed parental cells were nontransformed, suggesting that the revertants had sustained an activating mutation of a gene capable of suppressing transformation. The expression of c-jun, junB, and junD was not altered in the revertants, and they could not be transformed by transfection with a c-jun expression vector. The revertants were resistant to transformation by an activated c-Ha-ras gene but were susceptible to transformation by simian virus 40. Our results demonstrate the existence of a class of revertants that harbor genes capable of suppressing transformation by v-fos and some other oncogenes. This contrasts with previously described revertants of transformation by v-fos that contain recessive mutations.

We have shown that the murine c-rel protein can act as a transcriptional transactivator in both yeast and mammalian cells. Fusion proteins generated by linking rel sequences to the DNA-binding domain of the yeast transcriptional activator GAL4 activate transcription from a reporter gene linked in cis to a GAL4 binding site. The full-length mouse c-rel protein (588 amino acids long) is a poor transactivator; however, the C-terminal portion of the protein between amino acid residues 403 to 568 is a potent transcriptional transactivator. Deletion of the N-terminal half of the c-rel protein augments its transactivation function. We propose that c-rel protein has an N-terminal regulatory domain and a C-terminal transactivation domain which together modulate its function as a transcriptional transactivator.

We present evidence that the glucocorticoid receptor (GR) and transcription factor Jun/AP-1 can reciprocally repress one another's transcriptional activation by a novel mechanism that is independent of DNA binding. Overexpression of c-Jun prevents the glucocorticoid-induced activation of genes carrying a functional glucocorticoid response element (GRE). Conversely, GR is able to repress AP-1-mediated transcriptional activation. Mutant analysis reveals that the ligand binding and DNA binding domains of GR and the region including the leucine zipper of c-Jun are required for repression. Gel retardation analysis demonstrates that bacterially expressed c-Jun disrupts GR-GRE complexes. These data indicate that members of two distinct classes of transcription factors can oppose one another's activity through a mechanism likely involving protein-protein interactions.

The products of the Jun and Fos proto-oncogenes form a heterodimer that binds to and activates transcription from 12-O-tetradecanoylphorbol-13-acetate-responsive promoter elements (TGACTCA) and AP-1-binding sites (TGACATCA). These two proteins belong to a family of related transcription factors which contain similar domains required for protein dimerization and DNA binding but display different protein and DNA binding specificities. The basic region, required for DNA binding, is followed by a leucine zipper structure, a domain that mediates protein-protein interactions. To assess the role of these two domains in three related proteins, Fos, Jun, and CREB, we carried out extensive domain-swapping analysis. We found that (i) dimers formed by two Jun leucine zipper-containing proteins were unable to bind DNA as efficiently as a Fos-Jun combination, regardless of the source of the basic region; (ii) the Fos leucine zipper was unable to form either homo- or heterodimers with a chimeric protein containing a Fos leucine zipper; (iii) the Fos basic region was capable of binding to an AP-1 site; (iv) replacement of the Jun amino terminus with that of CREB had little effect on dimerization, whereas replacement with the amino terminus of Fos disrupted both protein-protein and protein-DNA interactions; (v) changes in relative affinities of the Fos and Jun basic regions for the AP-1 element were dependent on the secondary contributions of amino-terminal residues; and (vi) the Fos-Jun chimeric constructs cooperated in transcriptional transactivation of the Jun promoter in NIH 3T3 cells.

Primary skin fibroblasts from hemophilic dogs were transduced by recombinant retrovirus (LNCdF9L) containing a canine factor IX cDNA. High levels of biologically active canine factor IX (1.0 micrograms per 10(6) cells per 24 hr) were secreted in the medium. The level of factor IX produced increased substantially if the cells were stimulated by basic fibroblast growth factor during infection. Additionally, we also report that endothelial cells transduced by this virus can produce high levels of biologically active factor IX. We propose that skin fibroblasts and endothelial cells from hemophilia B dogs may serve as potential venues for the development and testing of models for treatment of hemophilia B by retrovirally mediated gene replacement therapy.

We have shown that the transcriptional activity of the protooncogene jun (c-jun) promoter is repressed by a transcription factor, the cAMP response element-binding protein (CREB). This repression can be alleviated when CREB is phosphorylated by the catalytic subunit of protein kinase A. Repression cannot be alleviated by a mutant CREB deficient in the protein kinase A phosphorylation site (M1 CREB Ser-133----Ala), suggesting that phosphorylation of CREB at this site is essential for the relief of repression. Repression by CREB requires its binding to the c-jun promoter. In NIH 3T3 cells stably expressing CREB, c-jun is no longer induced by serum, but this repression can be relieved by treatment of the cells with forskolin, an agonist of the adenylate cyclase pathway. Thus, CREB has a dual function, that of a repressor in the absence of phosphorylation and an activator when phosphorylated by protein kinase A.

Murine cell lines releasing helper-free recombinant retroviruses containing human alpha-globin and mouse/human hybrid beta-globin genes were generated. The expression of the hybrid beta-globin gene but not the human alpha-globin gene was regulated appropriately in infected mouse erythroid leukemia (MEL) cells. Murine bone marrow cells were infected by coculture with virus-producing cells and transplanted into lethally irradiated syngeneic recipients. Greater than 90% of the spleen colonies (12-15 days), which are derived from hemopoietic multipotential stem cells, showed proviral integration. Various levels of expression of the transduced globin genes were detected in all of the provirus-positive spleen colonies. Proviral sequences and transcripts from the transduced globin genes could also be detected in a few long-term reconstituted recipients in an observation period of 10 months after transplantation.

Jun and Fos nuclear oncoproteins form a complex that regulates transcription from promoters containing activator protein AP-1 binding sites. The leucine-zipper and basic-region domains of both Fos and Jun are necessary for formation of the heterodimer that binds to DNA. Reciprocal mutations in the basic region of Fos or Jun can influence the binding of the heterodimer to DNA, implying a symmetrical binding site. DNA-binding mutants of Jun exhibit increased affinity for Fos and are capable of suppressing wild-type Fos-Jun DNA-binding activity. In contrast, mutations in the basic domain of Fos, which prevent binding to DNA in association with Jun, do not significantly diminish the ability of the wild-type heterodimer to bind to DNA. These dominant negative mutants are functional in vivo and can be exploited to study the role of Fos and Jun in normal and transformed cells.

Proto-oncogene products c-Fos and c-Jun form a complex which binds with high affinity to the 12-O-tetradecanoylphorbol-13-acetate (TPA) response DNA element and which stimulates transcription of phorbol ester- inducible genes. We have previously identified, by screening a lambda gt11 expression library, murine protein mXBP, which binds to a sequence which overlaps the 3' end of the murine class II major histocompatibility complex A alpha gene X box, a conserved transcription element found upstream of all class II genes. Here, we demonstrate that the target sequence for mXBP is a consensus cyclic AMP response element (CRE). mXBP is a member of the leucine zipper family of DNA-binding proteins and has significant homology to oncoproteins c-Fos and c-Jun. The inferred amino acid sequence of mXBP shows near identity to human CRE-BP1, except it does not contain an internal proline-rich domain. Immunoprecipitation and glutaraldehyde cross-linking studies show that mXBP/CRE-BP2 can form a complex with c-Jun. Complex formation is dependent on intact leucine zipper domains in both proteins. mXBP-c-Jun complexes can coexist with c-Fos-c-Jun complexes and can bind with high affinity to CRE, but not to TPA response DNA element, sequences. These results suggest that changes in the expression of mXBP/CRE-BP2, c-Fos, and c-Jun, which alter the ratio of mXBP-c-Jun to c-Fos-c-Jun complexes, would affect the relative expression of cyclic AMP and phorbol ester-responsive genes. This provides support for a combinatorial model of gene regulation, whereby protein-protein interactions which alter the DNA binding specificity of protein complexes can expand the flexibility of cellular transcriptional responses.

The product of the jun proto-oncogene has been identified as one form of the transcription factor AP-1. The p55fos protein associates with jun/AP-1 by means of a heterodimer which requires intact 'leucine zipper' domains of both proteins. The fos/jun heterodimer binds to and activates transcription from TPA-responsive promoter elements (TGACTCA), which represent one final target of the protein kinase C pathway. The other main signal transduction pathway, initiated by the activation of the adenylate cyclase, involves the transcription factor CREB. The promoter element recognized by CREB, a cyclic AMP responsive element (CRE), consist of a palyndromic sequence similar to a TRE (TGACGTCA). We show that jun efficiently trans-activates CRE sequences and that fos and jun efficiently bind and cooperate in activating CRE promoter elements. The similarity between TRE and CRE sequences may involve an interplay in transcriptional regulation and 'cross-talk' between components of the two major signal transduction pathways.

We generated recombinant baculoviruses that contained the human fos gene and that, upon infection of insect cells, synthesized fos protein. The quantity of fos protein produced was at least 10 to 20 times higher than that observed in any mammalian cells reported so far. The fos protein made in insect cells manifested most of the characteristics of mammalian fos protein, which include (i) 55-kilodalton size, (ii) nuclear localization, (iii) phosphoesterification at serine residues, (iv) identical 35S tryptic peptide maps, (v) ability to make heterodimers with the nuclear jun oncoprotein, and (vi) cooperation with the jun protein to bind to a 12-O-tetradecanoyl-phorbol-13-acetate-responsive element. A 100- to 150-fold purification of the fos protein from infected insect cells was achieved in a single step by immunoaffinity chromatography. Availability of authentic fos protein made by baculoviral vectors in insect cells should allow a more rigorous analysis of its biochemical and biological properties.

Two nuclear oncoproteins, fos and jun (AP-1), cooperate in forming a very stable heterodimeric complex that binds to the AP-1 site on DNA with high affinity. The 'leucine zipper' domain of both fos and jun is necessary for the formation of this heterodimer. Mutations of single residues within the leucine zipper domain have no effect on protein complex formation. However, results from mutagenesis of the first leucine of the heptad repeat in either fos or jun basic regions and alteration of the spacing between the basic and leucine zipper domains indicate that the basic region of fos plays a crucial role in determining the DNA binding affinity of the transcriptional complex. Mutations of the basic amino acids in fos protein prevent binding to the tumour promoter response element (TRE) in the presence of wild-type jun protein. Thus fos protein appears to be dominant in jun-fos binding to DNA, even though fos alone cannot bind to TRE. Mutants in the basic region of fos and jun can be exploited as dominant-negative mutants to ablate the normal fos cellular function.

We have isolated the Xenopus laevis homologue of the protooncogene fos. Nucleotide sequence analysis of the complementary DNA revealed a coding domain for a 370-amino acid protein which shares extensive sequence homology with the avian and mammalian fos protein. During Xenopus development, the fos mRNA is present in the oocyte and egg stages and declines to relatively low levels by the mid-blastula transition. Transient expression of fos can be seen during the late neurula and again at the tadpole stage. The fos mRNA is uniformly distributed in the vegetal and animal poles of the oocytes. In a Xenopus kidney cell line, the fos gene can be transcriptionally activated by serum growth factors and 12-O-tetradecanoylphorbol-13-acetate, and the fos mRNA can be superinduced by cycloheximide. The X. laevis fos protein migrates as a series of bands around Mr 55,000 and is localized in the nucleus. Both the Xenopus fos RNA and fos proteins exhibit a short half-life analogous to that observed with their mammalian counterpart. The Xenopus fos protein readily associates with murine jun (AP-1) oncoprotein, and the complex binds efficiently to AP-1 recognition elements. Furthermore, the Xenopus fos protein collaborates with mammalian jun protein to activate transcription from 12-O-tetradecanoylphorbol-13-acetate-responsive element. The differential expression of fos during Xenopus development and its ability to regulate transcription in association with jun suggest a role during embryogenesis.