The Max transcription factor network: involvement of Mad in differentiation and an approach to identification of target genes

The small bHLHZip protein, Max, was originally identified through its interaction with Myc family proteins and appears to be an obligate partner for Myc function. Max has now been found to interact with at least two other proteins, Mad and Mxi1. These also belong to the bHLHZip class but are otherwise unrelated to Myc. Mad has been shown to abrogate the positive transcriptional activity of Myc and to inhibit Myc in co-transformation assays. This suggests that Mad may antagonize Myc function. Mad is rapidly induced upon differentiation, a time when Myc is frequently down-regulated. We show here evidence for Mad expression upon differentiation of myeloblasts, monoblasts, and keratinocytes. Mad:Max complexes are detected during differentiation and appear to replace the Myc:Max complexes present in proliferating cell populations. Since these complexes appear to form even in the presence of Myc, there may exist mechanisms that act to inhibit Myc:Max, or to promote Mad:Max, complex formation. We speculate that Max complex switching causes a change in the transcriptional activity of groups of target genes. Mad is not induced in all differentiating cell types, suggesting that other, possibly tissue-restricted, proteins might act in similar switch mechanisms to effect changes in transcriptional programs. We have also developed an approach to identification of the gene targets for Myc:Max complexes. By employing an immunoisolation procedure, we have begun characterization of several clones whose expression levels correlate with those of c-myc. Further identification of Myc-regulated genes may allow us to determine the molecular mechanism by which Myc governs cell proliferation and differentiation.

Thyroid hormone receptor can modulate retinoic acid-mediated axis formation in frog embryogenesis

Thyroid hormone receptor acts as a hormone-dependent transcriptional transactivator and as a transcriptional repressor in the absence of thyroid hormone. Specifically, thyroid hormone receptor can repress retinoic acid-induced gene expression through interactions with retinoic acid receptor. (Retinoic acid is a potent teratogen in the frog Xenopus laevis, acting at early embryonic stages to interfere with the formation of anterior structures. Endogenous retinoic acid is thought to act in normal anterior-posterior axis formation.) We have previously shown that thyroid hormone receptor RNA (alpha isotype) is expressed and polysome-associated during Xenopus embryogenesis preceding thyroid gland maturation and endogenous thyroid hormone production (D. E. Banker, J. Bigler, and R. N. Eisenman, Mol. Cell. Biol. 11:5079-5089, 1991). To determine whether thyroid hormone receptor might influence the effects of retinoic acid in early frog development, we have examined the results of ectopic thyroid hormone receptor expression on retinoic acid teratogenesis. We demonstrate that microinjections of full-length thyroid hormone receptor RNA protect injected embryos from retinoic acid teratogenesis. DNA binding is apparently essential to this protective function, as truncated thyroid hormone receptors, lacking DNA-binding domains but including hormone-binding and dimerization domains, do not protect from retinoic acid. We have shown that microinjections of these dominant-interfering thyroid hormone receptors, as well as anti-thyroid hormone receptor antibodies, increase retinoic acid teratogenesis in injected embryos, presumably by inactivating endogenous thyroid hormone receptor. This finding suggests that endogenous thyroid hormone receptors may act to limit retinoic acid sensitivity. On the other hand, after thyroid hormone treatment, ectopic thyroid hormone receptor mediates teratogenesis that is indistinguishable from the dorsoanterior deficiencies produced in retinoic acid teratogenesis. The previously characterized retinoic acid-responsive gene, Xhox.lab2, can be induced by thyroid hormone in embryos ectopically expressing thyroid hormone receptor and is less responsive to retinoic acid in such embryos. The fact that both thyroid hormone and retinoic acid can affect overlapping gene expression pathways to produce abnormal embryonic axes and can regulate the same early-expressed gene suggests a model in which thyroid hormone receptor blocks retinoic acid receptor-mediated teratogenesis by directly repressing retinoic acid-responsive genes.

A switch from Myc:Max to Mad:Max heterocomplexes accompanies monocyte/macrophage differentiation

Mad is a basic-helix-loop-helix-zipper protein that heterodimerizes with Max in vitro. Mad:Max heterodimers recognize the same E-box-related DNA-binding sites as Myc:Max heterodimers. However, in transient transfection assays Myc and Mad influence transcription in opposite ways through interaction with Max; Myc activates while Mad represses transcription. Here, we demonstrate that Mad protein is induced rapidly upon differentiation of cells of the myeloid lineage. The Mad protein is synthesized in human cells as a 35-kD nuclear phosphoprotein with an extremely short half-life (t1/2 = 15-30 min) and can be detected in vivo in a complex with Max. In the undifferentiated U937 monocyte cell line Max was found complexed with Myc but not Mad. However, Mad:Max complexes began to accumulate as early as 2 hr after induction of macrophage differentiation with TPA. By 48 hr following TPA treatment only Mad:Max complexes were detectable. These data show that differentiation is accompanied by a change in the composition of Max heterocomplexes. We speculate that this switch in heterocomplexes results in a change in the transcriptional regulation of Myc:Max target genes required for cell proliferation.

Binding of myc proteins to canonical and noncanonical DNA sequences

Using an in vitro binding-site selection assay, we have demonstrated that c-Myc-Max complexes bind not only to canonical CACGTG or CATGTG motifs that are flanked by variable sequences but also to noncanonical sites that consist of an internal CG or TG dinucleotide in the context of particular variations in the CA--TG consensus. None of the selected sites contain an internal TA dinucleotide, suggesting that Myc proteins necessarily bind asymmetrically in the context of a CAT half-site. The noncanonical sites can all be bound by proteins of the Myc-Max family but not necessarily by the related CACGTG- and CATGTG-binding proteins USF and TFE3. Substitution of an arginine that is conserved in these proteins into MyoD (MyoD-R) changes its binding specificity so that it recognizes CACGTG instead of the MyoD cognate sequence (CAGCTG). However, like USF and TFE3, MyoD-R does not bind to all of the noncanonical c-Myc-Max sites. Although this R substitution changes the internal dinucleotide specificity of MyoD, it does not significantly alter its wild-type binding sequence preferences at positions outside of the CA--TG motif, suggesting that it does not dramatically change other important amino acid-DNA contacts; this observation has important implications for models of basic-helix-loop-helix protein-DNA binding.

Expression of two distinct homologues of Xenopus Max during early development

The Max protein belongs to the basic region-helix-loop-helix-leucine zipper family of transcriptional regulators. Max heterodimerizes with Myc family proteins to form sequence-specific DNA-binding complexes. In order to elucidate the potential role of Myc and Max during amphibian embryogenesis, we have isolated and analyzed the expression of two Xenopus Max complementary DNAs: XMax1 and XMax2. Comparison of XMax1 and XMax2 with their mammalian counterparts demonstrates a strikingly high degree of conservation at both the nucleotide and amino acid levels, with the exception of a 24-residue deletion in both XMax proteins within their COOH termini. In addition, the two Xenopus Max proteins differ in that XMax2 contains a unique 27-amino acid insertion that interrupts the COOH-terminal end of the zipper domain; XMax1 lacks this insertion. Despite these differences, both XMax1 and XMax2 can form complexes with either Xenopus or human c-Myc proteins. Analysis of XMax expression during embryogenesis reveals that both mRNA and protein are expressed throughout early development, including the egg, 32-cell stage, and midblastula transition. Although the expression of XMax1 RNA appears to predominate at all stages examined, the ratios of XMax1 to XMax2 protein vary during development as well as between different tissue culture cell lines, suggesting a potential for cell type-specific regulation. Our results demonstrate the presence of Xenopus Max throughout frog development, raising the possibility that Myc and Max could function as a complex even during early embryogenesis.

Mad: a heterodimeric partner for Max that antagonizes Myc transcriptional activity

Myc family proteins appear to function through heterodimerization with the stable, constitutively expressed bHLH-Zip protein, Max. To determine whether Max mediates the function of regulatory proteins other than Myc, we screened a lambda gt11 expression library with radiolabeled Max protein. One cDNA identified encodes a new member of the bHLH-Zip protein family, Mad. Human Mad protein homodimerizes poorly but binds Max in vitro, forming a sequence-specific DNA binding complex with properties very similar to those of Myc-Max. Both Myc-Max and Mad-Max heterocomplexes are favored over Max homodimers, and, unlike Max homodimers, the DNA binding activity of the heterodimers is unaffected by CKII phosphorylation. Mad does not associate with Myc or with representative bHLH, bZip, or bHLH-Zip proteins. In vivo transactivation assays suggest that Myc-Max and Mad-Max complexes have opposing functions in transcription and that Max plays a central role in this network of transcription factors.

Dephosphorylation of the retinoblastoma protein during differentiation of HL60 cells

Immunoprecipitated retinoblastoma protein from HL60 cells migrated as a series of bands during electrophoresis. The heterogeneity appeared to be generated by phosphorylation of the retinoblastoma protein. Treatment of the cells with the phorbol ester, tetradecanoyl phorbol acetate (TPA), resulted in both a loss of the heterogeneity of the pRB species and a significant decrease in the level of pRB phosphorylation. These changes accompanied differentiation of the HL60 cells into macrophages. Treatment of the cells with dibutyryl cAMP also resulted in dephosphorylation of pRB as well as cell cycle arrest, although no recognizable differentiation occurred. These results are consistent with a model in which TPA and dibutyryl cAMP dependent pathways can activate pRB by altering its phosphorylation.

Regulation of Myc: Max complex formation and its potential role in cell proliferation

The myc family of proto-oncogenes encodes short-lived nuclear phosphoproteins (Myc) involved in the control of cell proliferation and differentiation. Here we discuss the evidence for Myc's involvement in normal and abnormal cell proliferation and review recent information on Max, a novel protein that forms a sequence-specific DNA-binding complex with Myc. The properties of the Myc: Max heterodimeric complex suggest a model for how Myc may function in the cell.

Myc and Max proteins possess distinct transcriptional activities

The Myc family proteins are thought to be involved in transcription because they have both a carboxy-terminal basic-helix-loop-helix-zipper (bHLH-Z) domain, common to a large class of transcription factors, and an amino-terminal fragment which, for c-Myc, has transactivating function when assayed in chimaeric constructs. In addition, c-, N- and L-Myc proteins heterodimerize, in vitro and in vivo, with the bHLH-Z protein Max. In vitro, Max homodimerizes but preferentially associates with Myc, which homodimerizes poorly. Furthermore Myc-Max heterodimers specifically bind the nucleotide sequence CACGTG with higher affinity than either homodimer alone. The identification of Max and the specific DNA-binding activities of Myc and Max provides an opportunity for directly testing the transcriptional activities of these proteins in mammalian cells. We report here that Myc overexpression activates, whereas Max overexpression represses, transcription of a reporter gene. Max-induced repression is relieved by overexpression of c-Myc. Repression requires the DNA-binding domain of Max, whereas relief of repression requires the dimerization and transcriptional activation activities of Myc. Both effects require Myc-Max-binding sites in the reporter gene.

Mitosis-specific phosphorylation of the nuclear oncoproteins Myc and Myb

The c-myc and c-myb proto-oncogenes encode phosphorylated nuclear DNA binding proteins that are likely to be involved in transcriptional regulation. Here we demonstrate that both Myc and Myb proteins are hyperphosphorylated during mitosis. In the case of Myb, hyperphosphorylation is accompanied by the appearance of three M phase-specific tryptic phosphopeptides. At least one of these phosphopeptides corresponds to a phosphopeptide generated after phosphorylation of Myb in vitro by p34cdc2 kinase. By contrast, the mitotic hyperphosphorylation of Myc does not correlate with the appearance of unique phosphopeptides, suggesting that M phase and interphase sites may be clustered within the same peptides. In addition Myc does not appear to be a target for p34cdc2 phosphorylation. The hyperphosphorylated forms of Myc and Myb from mitotic cells are functionally distinct from the corresponding interphase proteins in that the former have reduced ability to bind nonspecificially to double-stranded DNA cellulose. Furthermore, mitotic Myb binds poorly to oligodeoxynucleotides containing an Myb response element. We surmise that the decreased DNA binding capacity of hyperphosphorylated Myb and Myc during M phase may function to release these proteins from chromatin during chromosome condensation.

Phosphorylation of casein kinase II by p34cdc2 in vitro and at mitosis

In human epidermal carcinoma A431 cells, the beta subunit of casein kinase II is phosphorylated at an autophosphorylation site and at serine 209 which can be phosphorylated in vitro by p34cdc2 (Litchfield, D. W., Lozeman, F. J., Cicirelli, M. F., Harrylock, M., Ericsson, L. H., Piening, C. J., and Krebs, E. G. (1991) J. Biol. Chem. 266, 20380-20389). Given the importance of p34cdc2 in the regulation of cell cycle events, we were interested in examining the phosphorylation of casein kinase II during different stages of the cell cycle. In this study it is demonstrated that the extent of phosphorylation of serine 209 in the beta subunit is significantly increased relative to phosphorylation of the autophosphorylation site when chicken bursal lymphoma BK3A cells are arrested at mitosis by nocodazole treatment. This result suggests that serine 209 is a likely physiological target for p34cdc2. In addition, the alpha subunit of casein kinase II also undergoes dramatic phosphorylation with an associated alteration in its electrophoretic mobility when BK3A cells or human Jurkat cells are arrested with nocodazole. Phosphopeptide mapping studies indicate that p34cdc2 can phosphorylate in vitro the same peptides on the alpha subunit that are phosphorylated in cells arrested at mitosis. These phosphorylation sites were localized to serine and threonine residues in the carboxyl-terminal domain of alpha. Taken together, the results of this study indicate that casein kinase II is a probable physiological substrate for p34cdc2 and suggest that its functional properties could be affected in a cell cycle-dependent manner.

Mapping of MAX to human chromosome 14 and mouse chromosome 12 by in situ hybridization

The protein encoded by the MAX gene is a member of the class of basic region-helix-loop-helix-zipper proteins and has been demonstrated to associate with N-, L-, and c-Myc proteins both in vitro and in vivo. Heterodimers formed between c-Myc and Max proteins have been shown to possess sequence-specific DNA-binding activity. Here we report the mapping of the MAX gene to a single region on human chromosome 14 (bands q22-q24) and to mouse chromosome 12 (region D). Chromosome abnormalities linked to several neoplasms have been previously associated with this region on human chromosome 14. The mapping results also confirm a region of homology between human chromosome 14q22-24 and mouse chromosome 12 region D.

Thyroid hormone receptor transcriptional activity is potentially autoregulated by truncated forms of the receptor

ErbA/thyroid hormone receptor is a nuclear receptor that can affect transcription from promoters containing a thyroid hormone response element (TRE) in a thyroid hormone (T3)-dependent manner. We reported earlier that the thyroid hormone receptor is expressed in embryonic avian erythroid cells as a nested set of four proteins with a common C terminus. The full-length receptor is capable of both high-affinity binding to thyroid hormone and specific binding to DNA. We now report that the two smallest ErbA forms, which contain the hormone-binding domain but lack the N-terminal DNA-binding domain, have the same affinity for T3 as does full-length ErbA but are incapable of specific DNA binding. In transactivation assays, these N-terminally truncated proteins are able to specifically suppress both transcriptional repression and hormone-dependent transcriptional activation by the full-length ErbA. We also find that retinoic acid-dependent transactivation by retinoic acid receptors is inhibited by the truncated ErbA proteins. Furthermore, the smaller ErbA forms inhibit binding to TREs by full-length ErbA in vitro. Results from experiments involving site-specific mutagenesis of a conserved region within the hormone-binding domain of the smaller ErbA proteins indicate that the suppressive effect of the smaller receptor forms is independent of hormone binding and that this region is important in mediating protein-hormone as well as protein-protein interactions. We have also found that full-length ErbA homodimers can be detected only in the presence of a specific DNA-binding site. However, no association between full-length and the N-terminally truncated non-DNA-binding ErbA proteins could be detected, indicating that the complex either is unstable or does not form. Our results suggest that inhibition of receptor function occurs through transient formation of heterodimers which lack DNA-binding activity or by competition for factors which positively affect DNA binding by the full-length protein. This finding raises the possibility that thyroid hormone receptor transcriptional activity is autoregulated by means of alternative receptor translation products acting in a dominant negative manner.

Myc and Max function as a nucleoprotein complex

The Myc family of oncoproteins are thought to regulate proliferation and differentiation in a wide variety of cell types. Recent studies show that Myc proteins form sequence-specific DNA-binding complexes with Max, a new member of the helix-loop-helix leucine zipper protein class. The properties of the Myc-Max complex suggest a mechanism for Myc's function in both normal and neoplastic cell behavior.

Transcriptional activities of the Myc and Max proteins in mammalian cells

The myc family of oncogenes exhibit deregulated expression in a host of neoplasias. Though the molecular function of the Myc protein in both normal and tumorigenic cells has remained uncertain, it has been postulated to play a role in gene transcription on the basis of amino acid homologies with known transcription factors such as MyoD (Lüscher & Eisenman, 1990). We report here the direct testing of full-length Myc and its dimerization partner, Max, on the transcriptional activity of reporter genes bearing Myc/Max binding sites. Such reporter constructs display an endogenous level of activity in transient transfections which is dependent on the presence of the CACGTG sequence. Exogenous expression of myc results in modest activation of reporter gene transcription. Similar overexpression of max results in a repression of reporter gene activity, an effect which is reversed by co-expression with c-myc. Max repression is dependent on an intact DNA binding region, while Myc activation depends on both the N-terminal activation and the C-terminal dimerization domains. These results suggest a model in which Max homodimers can act as as repressors, and Myc-Max heterodimers as activators, of potential target genes.

Myc and Max associate in vivo

Max is a helix-loop-helix zipper protein that associates in vitro with Myc family proteins to form a sequence-specific DNA-binding complex. We show here, by means of a coimmunoprecipitation assay with anti-Myc and anti-Max antibodies, that Myc and Max are associated in vivo and essentially all of the newly synthesized Myc can be detected in a complex with Max. This complex possesses specific DNA-binding activity for CACGTG-containing oligonucleotides. Although Max itself is a highly stable protein, Myc is rapidly degraded during or after its association with Max. In vivo Max is shown to be a nuclear protein phosphorylated by casein kinase II, and alternatively spliced forms of Max are expressed in cells. Furthermore, the levels of Max expression are equivalent in quiescent, mitogen-stimulated, and cycling cells. We conclude that the highly regulated rate of Myc biosynthesis is likely to be a limiting step in the formation of Myc:Max complexes.

The thyroid hormone receptor gene (c-erbA alpha) is expressed in advance of thyroid gland maturation during the early embryonic development of Xenopus laevis

The c-erbA proto-oncogene encodes the thyroid hormone receptor, a ligand-dependent transcription factor which plays an important role in vertebrate growth and development. To define the role of the thyroid hormone receptor in developmental processes, we have begun studying c-erbA gene expression during the ontogeny of Xenopus laevis, an organism in which thyroid hormone has well-documented effects on morphogenesis. Using polymerase chain reactions (PCR) as a sensitive assay of specific gene expression, we found that polyadenylated erbA alpha RNA is present in Xenopus cells at early developmental stages, including the fertilized egg, blastula, gastrula, and neurula. By performing erbA alpha-specific PCR on reverse-transcribed RNAs from high-density sucrose gradient fractions prepared from early-stage embryos, we have demonstrated that these erbA transcripts are recruited to polysomes. Therefore, erbA is expressed in Xenopus development prior to the appearance of the thyroid gland anlage in tailbud-stage embryos. This implies that erbA alpha/thyroid hormone receptors may play ligand-independent roles during the early development of X. laevis. Quantitative PCR revealed a greater than 25-fold range in the steady-state levels of polyadenylated erbA alpha RNA across early stages of development, as expressed relative to equimolar amounts of total embryonic RNA. Substantial increases in the levels of erbA alpha RNA were noted at stages well after the onset of zygotic transcription at the mid-blastula transition, with accumulation of erbA alpha transcripts reaching a relative maximum in advance of metamorphosis. We also show that erbA alpha RNAs are expressed unequally across Xenopus neural tube embryos. This differential expression continues through later stages of development, including metamorphosis. This finding suggests that erbA alpha/thyroid hormone receptors may play roles in tissue-specific processes across all of Xenopus development.

Negative charge at the casein kinase II phosphorylation site is important for transformation but not for Rb protein binding by the E7 protein of human papillomavirus type 16

The human papillomavirus E7 protein is phosphorylated at the two serines in positions 31/32, which are part of a consensus sequence for casein kinase II (CKII). In this study, we have investigated the effect of CKII phosphorylation site mutations, all of which lead to unphosphorylated E7 proteins. The replacement of the two serines by uncharged alanine residues drastically reduced the ability of E7 to cotransform primary cells with ras, whereas negatively charged aspartic acid at the same positions produced only a slight effect. This difference was not reflected in the p105Rb binding or the E2 promoter transactivation capability of these two mutants. Mutations that changed the CKII consensus without altering the serine residues also resulted in a loss of phosphorylation and transformation. This indicated that negative charge at positions 31/32 provided either by phosphorylation or by a negatively charged amino acid is necessary for efficient transformation without significantly affecting p105Rb binding or transactivation.

A role for the p34cdc2 kinase and phosphatases in the regulation of phosphorylation and disassembly of lamin B2 during the cell cycle

While the p34cdc2 kinase is considered to be a critical regulator of mitosis, its function has not yet been directly linked to one of the key events during the onset of mitosis: nuclear envelope breakdown. Here we show that a major structural protein of the nuclear envelope, lamin B2, is phosphorylated by p34cdc2. Results from two-dimensional phosphopeptide mapping experiments demonstrate that the p34cdc2-specific phosphopeptides represent both mitotic and interphase specific phosphorylations of lamin B2 and include the major interphase phosphorylation site. In mitotic cells we detected two distinct forms of lamin B2 which differ in electrophoretic mobility and in degree of phosphorylation. The phosphorylation pattern of lamin B2 generated in vitro by p34cdc2 was more closely related to the less phosphorylated mitotic lamin B2, suggesting that another kinase(s) in addition to p34cdc2 is involved in generating the mitotic phosphorylation pattern. In addition, we show that treatment of interphase cells with okadaic acid, a potent phosphatase inhibitor, leads to the acquisition of mitosis-specific phosphopeptides and can reversibly increase the detergent-solubility of lamin B2. However, the M-phase-like phosphorylation of lamin B2 in itself is not sufficient to induce its disassembly from the nuclear lamina suggesting that an additional event(s) besides phosphorylation is required.

Max: a helix-loop-helix zipper protein that forms a sequence-specific DNA-binding complex with Myc

The myc protooncogene family has been implicated in cell proliferation, differentiation, and neoplasia, but its mechanism of function at the molecular level is unknown. The carboxyl terminus of Myc family proteins contains a basic region helix-loop-helix leucine zipper motif (bHLH-Zip), which has DNA-binding activity and has been predicted to mediate protein-protein interactions. The bHLH-Zip region of c-Myc was used to screen a complementary DNA (cDNA) expression library, and a bHLH-Zip protein, termed Max, was identified. Max specifically associated with c-Myc, N-Myc, and L-Myc proteins, but not with a number of other bHLH, bZip, or bHLH-Zip proteins. The interaction between Max and c-Myc was dependent on the integrity of the c-Myc HLH-Zip domain, but not on the basic region or other sequences outside the domain. Furthermore, the Myc-Max complex bound to DNA in a sequence-specific manner under conditions where neither Max nor Myc exhibited appreciable binding. The DNA-binding activity of the complex was dependent on both the dimerization domain and the basic region of c-Myc. These results suggest that Myc family proteins undergo a restricted set of interactions in the cell and may belong to the more general class of eukaryotic DNA-binding transcription factors.

Sequence-specific DNA binding by the c-Myc protein

While it has been known for some time that the c-Myc protein binds to random DNA sequences, no sequence-specific binding activity has been detected. At its carboxyl terminus, c-Myc contains a basic--helix-loop-helix (bHLH) motif, which is important for dimerization and specific DNA binding, as demonstrated for other bHLH protein family members. Of those studied, most bHLH proteins bind to sites that contain a CA- -TG consensus. In this study, the technique of selected and amplified binding-sequence (SAAB) imprinting was used to identify a DNA sequence that was recognized by c-Myc. A purified carboxyl-terminal fragment of human c-Myc that contained the bHLH domain bound in vitro in a sequence-specific manner to the sequence, CACGTG. These results suggest that some of the biological functions of Myc family proteins are accomplished by sequence-specific DNA binding that is mediated by the carboxyl-terminal region of the protein.

Myb DNA binding inhibited by phosphorylation at a site deleted during oncogenic activation

The c-Myb nuclear oncoprotein is phosphorylated in vitro and in vivo at an N-terminal site near its DNA-binding domain by casein kinase II (CK-II) or a CK-II-like activity. This in vitro phosphorylation reversibly inhibits the sequence-specific binding of c-Myb to DNA. The site of this phosphorylation is deleted in nearly all oncogenically activated Myb proteins, resulting in DNA-binding that is independent of CK-II. Because CK-II activity is modulated by growth factors, loss of the site could uncouple c-Myb from its normal physiological regulator.

Mutational analysis of the carboxy-terminal casein kinase II phosphorylation site in human c-myc

Myc proteins are phosphorylated within two critical regions by casein kinase II (CKII): the central acidic domain and a carboxy-terminal region bordering the basic region-helix-loop-helix segment. In order to test whether the carboxy-terminal phosphorylation site was functionally important we introduced three types of mutations into this region. Two of the mutations would be expected to prevent phosphorylation and minimize negative charge while the third introduced a permanent negative charge. The Myc CKII site mutants were cloned into a retroviral vector and were shown to be efficiently expressed in several different cell types. In one mutant we directly demonstrated loss of the phosphorylation site. When the Myc mutants were used in a cooperative transformation assay of Rat-1 cells with the bcr-abl oncogene we were unable to detect a significant difference in transformation efficiency between wild-type Myc and any of the mutants. While the CKII site is non-functional in this assay, the high levels of Myc produced may have overridden potential CKII regulation.

FH3, a v-myc avian retrovirus with limited transforming ability

We have isolated a new acute avian transforming virus which contains the oncogene myc. This virus, designated FH3, was isolated after injection of a 10-day-old chick embryo with avian leukosis virus. While FH3 shares many properties with other v-myc-containing avian retroviruses, it also has several unique properties. The primary target for transformation in vitro is chicken macrophages; infection of chicken fibroblasts does not lead to complete morphological transformation. FH3 also exhibits a limited host range, in that Japanese quail macrophages and fibroblasts are infected but are not completely transformed. FH3 induces in vivo a limited tumor type if injected into 10-day-old chick embryos; only a cranial myelocytoma, which does not appear to be metastatic, can be detected. The v-myc gene of FH3 is expressed predominantly as a P145 Gag-Myc protein which is encoded by a ca. 8-kilobase genomic RNA. This FH3-encoded polyprotein is localized in the nucleus of all infected cells, whether or not they are transformed.

A second c-myb protein is translated from an alternatively spliced mRNA expressed from normal and 5'-disrupted myb loci

The major protein encoded by the c-myb oncogene in many species has been identified as an unstable, nuclear DNA-binding protein with an apparent molecular mass of 75 to 80 kilodaltons (p75c-myb). Recently, an alternatively spliced form of c-myb-encoded mRNA has been identified in murine cells containing either normal or rearranged c-myb genes. This mRNA includes a new exon, termed E6A, formed through use of cryptic splice sites located in the large intron between c-myb exons vE6 and vE7. E6A is predicted to contribute an internal 121-residue in-frame insertion into a region C terminal of the DNA-binding domain the c-myb-encoded protein. Here we report the identification of an 85-kilodalton (p85c-myb-E6A) protein as the translation product of the alternatively spliced E6A c-myb mRNA. This protein as well as p75c-myb were precipitated with anti-Myb antibodies raised against the conserved DNA-binding region of c-Myb. Proteolytic mapping studies showed that the two proteins are highly related but not identical. However, only the p85 protein reacted with an antiserum prepared against the E6A region expressed in bacteria, demonstrating that p85 but not p75 contains E6A sequences. In addition, the mobilities of both p85 and p75 were increased in myeloid tumor cell lines containing proviral integrations upstream of the 5' coding exons of v-myb, indicating that both proteins are truncated forms of c-Myb expressed from the same disrupted allele. p75c-myb and p85c-myb-E6A were indistinguishable with respect to nuclear localization and protein half-life. Furthermore, both forms of Myb were synthesized continuously throughout the cell cycle in 70Z ore-B cells. The contribution of the E6A domain to c-myb function remains to be elucidated.

The E7 protein of human papillomavirus type 16 is phosphorylated by casein kinase II

The E7 protein of human papillomavirus type 16 (HPV16) transforms cultured cells and cooperates with the ras or fos oncogenes in the transformation of primary cells. In this study we have investigated the phosphorylation of E7. When we immunoprecipitated E7 from CaSki cells with a rabbit polyclonal antiserum to a bacterial fusion protein (trpE-E7), we found that E7 was phosphorylated at serine residues contained in five characteristic thermolysin peptides. Immunoprecipitated E7, and fusion proteins harboring the E7 protein from various HPV types, could all be specifically phosphorylated in vitro by the ubiquitous, growth factor-activated casein kinase II (CKII). Comparative peptide mapping showed that the sites of in vivo and in vitro phosphorylation are the same. CKII was shown previously to specifically phosphorylate serine or threonine residues within a cluster of acidic amino acids. The E7 protein contains such a sequence between amino acids 30 and 37. When a synthetic peptide corresponding to this region of E7 was phosphorylated by CKII in vitro, its thermolysin digestion products were the same as those in the phosphorylated E7 protein. We conclude that E7 is phosphorylated in vivo only at serines within the predicted CKII site and that CKII, or a CKII-like enzyme, participates in the reaction. Both the E1A and SV40 large T proteins contain similar CKII consensus sites proximal to the regions required for their associations with the retinoblastoma gene product (p105Rb). Thus it is conceivable that CKII phosphorylation can modulate the interaction between the transforming proteins and the retinoblastoma gene product.

Myc oncoproteins are phosphorylated by casein kinase II

Casein kinase II (CK-II) is a ubiquitous protein kinase, localized to both nucleus and cytoplasm, with strong specificity for serine residues positioned within clusters of acidic amino acids. We have found that a number of nuclear oncoproteins share a CK-II phosphorylation sequence motif, including Myc, Myb, Fos, E1a and SV40 T antigen. In this paper we show that cellular myc-encoded proteins, derived from avian and human cells, can serve as substrates for phosphorylation by purified CK-II in vitro and that this phosphorylation is reversible. One- and two-dimensional mapping experiments demonstrate that the major phosphopeptides from in vivo phosphorylated Myc correspond to the phosphopeptides produced from Myc phosphorylated in vitro by CK-II. In addition, synthetic peptides with sequences corresponding to putative CK-II phosphorylation sites in Myc are subject to multiple, highly efficient phosphorylations by CK-II, and can act as competitive inhibitors of CK-II phosphorylation of Myc in vitro. We have used such peptides to map the phosphorylated regions in Myc and have located major CK-II phosphorylations within the central highly acidic domain and within a region proximal to the C terminus. Our results, along with previous studies on myc deletion mutants, show that Myc is phosphorylated by CK-II, or a kinase with similar specificity, in regions of functional importance. Since CK-II can be rapidly activated after mitogen treatment we postulate that CK-II mediated phosphorylation of Myc plays a role in signal transduction to the nucleus.

Detection of a Myc-associated protein by chemical cross-linking

A single nuclear protein (Myc-associated protein) can be specifically cross-linked to avian Myc proteins by treatment of nuclei or cells with the reversible cross-linker dimethyl 3,3'-dithiobis-propionimidate. Myc-associated protein has a molecular weight of approximately 500,000, is not detectably phosphorylated and, in contrast to Myc, has a long apparent half-life of greater than 3 h.

c-erbA encodes multiple proteins in chicken erythroid cells

To identify and characterize the proteins encoded by the erbA proto-oncogene, we expressed the C-terminal region of v-erbA in a bacterial trpE expression vector system and used the fusion protein to prepare antiserum. The anti-trp-erbA serum recognized the P75gag-erbA protein encoded by avian erythroblastosis virus and specifically precipitated six highly related proteins ranging in size from 27 to 46 kilodaltons from chicken embryonic erythroid cells. In vitro translation of a chicken erbA cDNA produced essentially the same pattern of proteins. Partial proteolytic maps and antigenicity and kinetic analyses of the in vivo and in vitro proteins indicated that they are related and that the multiple bands are likely to arise from internal initiations within c-erbA to generate a nested set of proteins. All of the c-erbA proteins are predominantly associated with chicken erythroblast nuclei. However, Nonidet P-40 treatment resulted in extraction of the three smaller proteins, whereas the larger proteins were retained. During differentiation of erythroid cells in chicken embryos, we found maximal levels of c-erbA protein synthesis at days 7 to 8 of embryogenesis. By contrast, c-erbA mRNA levels remained essentially constant from days 5 to 12. Together, our results indicate that posttranscriptional or translational mechanisms are involved in regulation of c-erbA expression and in the complexity of its protein products.

c-myc and c-myb protein degradation: effect of metabolic inhibitors and heat shock

The proteins encoded by both viral and cellular forms of the c-myc oncogene have been previously demonstrated to have exceptionally short in vivo half-lives. In this paper we report a comparative study on the parameters affecting turnover of nuclear oncoproteins c-myc, c-myb, and the rapidly metabolized cytoplasmic enzyme ornithine decarboxylase. The degradation of all three proteins required metabolic energy, did not result in production of cleavage intermediates, and did not involve lysosomes or ubiquitin. A five- to eightfold increase in the half-life of c-myc proteins, and a twofold increase in the half-life of c-myb proteins was detected after heat-shock treatment at 46 degrees C. In contrast, heat shock had no effect on the turnover of ornithine decarboxylase. Heat shock also had the effect of increasing the rate of c-myc protein synthesis twofold, whereas c-myb protein synthesis was decreased nearly fourfold. The increased stability and synthesis of c-myc proteins led to an overall increase in the total level of c-myc proteins in response to heat-shock treatment. Furthermore, treatments which reduced c-myc and c-myb protein turnover, such as heat shock and exposure to inhibitors of metabolic energy production, resulted in reduced detergent solubility of both proteins. The recovery from heat shock, as measured by increased turnover and solubility, was energy dependent and considerably more rapid in thermotolerant cells.

Translocated c-myc genes produce chimeric transcripts containing antisense sequences of the immunoglobulin heavy chain locus in mouse plasmacytomas

Immunoglobulin heavy chain gene antisense transcripts contribute to the expression of translocated c-myc genes in several murine plasma cell tumors. These novel, chimeric transcripts comprise 5-50% of steady-state c-myc mRNA. Two transcripts isolated as cDNA clones use the normal splice donor and acceptor sites within the c-myc first intron. Another cDNA clone has the potential for encoding two types of c-myc proteins. The significance of immunoglobulin heavy chain gene antisense transcripts and transcriptional competence of the immunoglobulin heavy chain locus for c-myc expression is discussed.

A non-AUG translational initiation in c-myc exon 1 generates an N-terminally distinct protein whose synthesis is disrupted in Burkitt's lymphomas

The c-myc gene comprises three exons with a single large AUG-initiated open reading frame extending from exon 2 through exon 3. Exon 1 lacks any AUG codons. Cells from a wide range of species produce two c-myc proteins that, while highly related, do not appear to arise from posttranslational interconversion. To understand the origin of the two proteins, we mapped them and analyzed the in vitro protein-coding capacity of c-myc cDNAs. Our findings show that the two proteins are derived from alternative translational initiations at the exon 2 AUG and at a non-AUG codon near the 3' end of exon 1, resulting in the production of proteins with distinct N termini. In Burkitt's lymphomas, the removal or specific mutation of exon 1 in c-myc translocations correlates with suppression of synthesis of the larger protein, and thus may contribute to the oncogenic activation of c-myc.

Expression of the c-myc proto-oncogene during development of Xenopus laevis

We isolated and characterized Xenopus laevis c-myc cDNAs from an oocyte-specific library. These cDNA clones encompass 2.35 kilobases of the X. laevis c-myc RNA and contain the entire coding domain of 1,257 nucleotides of the 419-amino acid-long X. laevis c-myc protein. The 2.7-kilobase X. laevis c-myc mRNA is expressed in the oocyte, maintained in the egg, and is present throughout the early cleavage stages of embryogenesis. At the time of transcriptional activation in the embryo the c-myc RNA levels show a significant decline and then reaccumulate continuously throughout the remainder of premorphogenic development. At the early neurula stage of embryogenesis the pattern of c-myc RNA expression is elevated in the mesoderm with respect to the endoderm and ectoderm. In the adult X. laevis the c-myc mRNA is expressed in some (e.g., skin, muscle) but not all differentiated tissues. The X. laevis c-myc protein migrates as a doublet of 61,000- and 64,000-dalton species. Both species are phosphorylated in oocytes and somatic cells, exhibit extremely short half-lives of less than 30 min, and are localized to the nuclear fraction of somatic cells. By contrast, the oocyte protein shows both cytoplasmic and germinal vesicle distribution and appears to be stable.

Proteins expressed by the c-myc oncogene in lymphomas of human and avian origin

We have prepared antisera against synthetic peptides corresponding to the C-terminal region of the avian and human myc oncogene coding sequences. Immunoprecipitates from avian and human cells show two major proteins which, by the criteria of hybrid-selected translation, transfection, and peptide-blocking assays, are the c-myc protein products. These proteins are phosphorylated nuclear proteins which are tightly bound to the nuclear matrix-lamin and which have a short half-life. Analysis of avian and human lymphoma cell lines containing rearranged c-myc alleles show significant changes in the ratio of the two proteins although only the avian lymphomas have increased quantities of c-myc protein.

Delineation of transcriptional control signals within the Moloney murine sarcoma virus long terminal repeat

We identified three distinct elements within the Moloney murine sarcoma virus long terminal repeat that control transcription. The phenotypes of unidirectional deletion mutants of the long terminal repeat were assayed in microinjected frog oocytes and in transfected mouse fibroblasts. Steady-state levels of RNA bearing the same 5' terminus as the authentic Moloney murine sarcoma viral transcripts were measured by primer extension in assays that included a pseudo-wild-type internal reference. Mutant phenotypes define the boundaries of three functional elements. A region between 21 and 31 base pairs upstream from the mRNA cap site contains AT-rich sequences that function to establish the transcription start site. A second control element, termed the distal signal, lies between 31 and 84 base pairs upstream of the mRNA cap site. A CAT box consensus sequence is located at the 5' boundary of the distal signal. Additional components of the distal signal include a hexanucleotide sequence that is repeated four times. The distal signal augments transcription efficiency in oocytes but contributes only weakly to long terminal repeat-mediated expression in mouse fibroblasts. A third transcriptional control element lies between 156 and 364 base pairs upstream of the mRNA cap site. This element includes the 75-base-pair repeats previously identified as the Moloney murine sarcoma virus enhancer. In contrast to the distal signal, the Moloney murine sarcoma virus enhancer is crucial for significant expression in mouse fibroblasts but does not contribute to transcriptional expression in frog oocytes.

c-myc oncogene protein synthesis is independent of the cell cycle in human and avian cells

Several lines of evidence suggest a role for the myc oncogene in cell proliferation. Most recently, mitogenic stimulation of quiescent lymphoid, fibroblast and epithelial cells has been demonstrated to lead to a sharp increase in c-myc RNA levels. To determine how c-myc expression is linked to the cell proliferative cycle, we have used centrifugal elutriation to enrich for populations of avian and human cells at different stages of the cell cycle. Centrifugal elutriation is a counterflow centrifugation method that separates cells on the basis of volume, a parameter correlating well with progression through the cell cycle. Using myc-specific anti-peptide antibodies, we show here that the synthesis, half-life and modification of c-myc proteins are constant throughout the cell cycle of normal and transformed cells.

V-myc- and c-myc-encoded proteins are associated with the nuclear matrix

A series of extraction procedures were applied to avian nuclei which allowed us to define three types of association of v-myc- and c-myc-encoded proteins with nuclei: (i) a major fraction (60 to 90%) which is retained in DNA- and RNA-depleted nuclei after low- and high-salt extraction, (ii) a small fraction (1%) released during nuclease digestion of DNA in intact nuclei in the presence of low-salt buffer, and (iii) a fraction of myc protein (less than 10%) extractable with salt or detergents and found to have affinity for both single- and double-stranded DNA. Immunofluorescence analysis with anti-myc peptide sera on cells extracted sequentially with nucleases and salts confirmed the idea that myc proteins were associated with a complex residual nuclear structure (matrix-lamin fraction) which also contained avian nuclear lamin protein. Dispersal of myc proteins into the cytoplasm was found to occur during mitosis. Both c-myc and v-myc proteins were associated with the matrix-lamin, suggesting that the function of myc may relate to nuclear structural organization.

Proteins encoded by the human c-myc oncogene: differential expression in neoplastic cells

To examine myc protein products in the wide variety of human tumor cells having alterations of the c-myc locus, we have prepared an antiserum against a synthetic peptide corresponding to the predicted C-terminal sequence of the human c-myc protein. This antiserum (anti-hu-myc 12C) specifically precipitated two proteins of 64 and 67 kilodaltons in quantities ranging from low levels in normal fibroblasts to 10-fold-higher levels in Epstein-Barr virus-immortalized and Burkitt's lymphoma cell lines, to 20- to 60-fold-higher levels in cell lines having amplified c-myc. The p64 and p67 proteins were found to be highly related by partial V8 proteolytic mapping, and both were demonstrated to be encoded by the c-myc oncogene, using hybrid-selected translation of myc-specific RNA. In addition, the p64 protein was specifically precipitated from cells transfected with a translocated c-myc gene. Both p64 and p67 were found to be nuclear phosphoproteins with extremely short half-lives. In tumor cell lines having alterations at the c-myc locus due to amplification or translocation, we observed a significant change in the expression of p64 relative to p67 when compared with normal or Epstein-Bar virus-immortalized cells.

New findings on the congenital transmission of avian leukosis viruses

RAV-0, an endogenous avian leukosis virus, does not undergo congenital transmission in infected K28 chickens. In contrast, avian leukosis viruses of exogenous origin undergo highly efficient congenital transmission. The relative abilities of endogenous and exogenous viruses to undergo congenital transmission appear to be determined by the p27 capsid proteins of these viruses.

Proteins encoded by v-myc and c-myc oncogenes: identification and localization in acute leukemia virus transformants and bursal lymphoma cell lines

We have prepared an antiserum against a synthetic dodecapeptide whose sequence corresponds to the C terminus of the MC29 v-myc protein. This antiserum (anti-v-myc 12C) specifically precipitates the known gag-myc fusion proteins produced by the defective leukemia viruses MC29, CMII, and OK10, but does not react with gag-precursor or product proteins. In addition, proteins of 62 kd and 61/63 kd are precipitated by anti-v-myc 12C from OK10 and MH2 transformants, respectively. The serum also recognizes comigrating 62 kd proteins from three chicken bursal lymphoma cell lines and from the products of in vitro translation of c-myc-specific mRNA. All of these myc-related proteins are phosphorylated and all appear to be localized in the cell nucleus. In uninfected quail cells, anti-v-myc 12C also recognizes a candidate c-myc protein of 60 kd, which does not appear to be phosphorylated and is present in low levels relative to v-myc and lymphoma c-myc proteins.

Nuclear location of the putative transforming protein of avian myelocytomatosis virus

The putative transforming protein of avian myelocytomatosis virus MC29 is a 110,000 dalton (P110gag-myc) polyprotein comprised of sequences derived from both the gag region and the MC29-specific myc region. Two approaches have been taken to determine the location of the MC29 gag-related proteins in transformed cells: subcellular fractionation and immunofluorescence. Analysis of subcellular fractions of MC29-transformed cells by immunoprecipitation indicates that the majority of the gag-myc polyprotein is found in the nuclear fractions of Q8 cells (a nonproducer line of MC29-transformed quail embryo fibroblasts) and nonproducer cells derived from a liver tumor of MC20-infected quail. This is in contrast to the distribution of gag-related helper virus proteins lacking myc, which are found only in nonnuclear fractions of superinfected Q8 cells. The purity of unlabeled nuclei was assessed by electron microscopy and enzyme assays, revealing little contaminating material from other subcellular fractions. Immunofluorescence experiments using monospecific anti-gag serum showed specific, intense immunofluorescence in the nuclei of fixed Q8 cells. In contrast, the majority of P75gag-erb, a candidate transforming protein produced by avian erythroblastosis virus (AEV), is absent from the nuclei of nonproducer AEV-transformed chick embryo fibroblasts. The nuclear association of the MC29 transforming protein may be related to some of the unique properties of MC29-transformed cells.

Synthesis and processing of polymerase proteins of wild-type and mutant avian retroviruses

We have studied the biosynthesis of avian retrovirus proteins related to reverse transcriptase in permissive avian embryonic cells. Analysis of immune precipitates from avian sarcoma virus (ASV)-infected cells demonstrated the presence of the 180,000-dalton gag-pol "read-through" protein (Pr180gag-pol) and a 130,000-dalton polypeptide (Pr130gag-pol). Pr130gag-pol was found, in serological and peptide mapping studies, to consist primarily of sequences related to reverse transcriptase and the gag-encoded protein p15. Pr180gag-pol was found to be phosphorylated, whereas Pr130gag-pol was not. In addition, only Pr180gag-pol but not Pr130gag-pol was susceptible to cleavage with the virion protease p15. Although the structure of Pr130gag-pol would suggest that it is generated by removal of a portion of the gag region from Pr180gag-pol, an analysis of labeling kinetics has failed to demonstrate unequivocally whether Pr130gag-pol is a cleavage product of Pr180gag-pol or a primary translation product. We were repeatedly unable to detect either Pr180gag-pol or Pr130gag-pol in virus particles released from the cell, whereas both beta and alpha subunits were readily observed. Several presumed intermediates between Pr130gag-pol and the beta subunit of reverse transcriptase were also observed in virions. These studies indicate cleavage of polyemrase precursors at the time of virus budding. On the basis of these data, we present a processing scheme for the generation of reverse transcriptase subunits. We have also examined reverse transcriptase biosynthesis in cells producing two mutants that fail to package the enzyme. Previous work showed that integrated proviruses of both mutants are missing DNA sequences in pol: one mutant, PH9 (Mason et al., J. Virol. 30:132-140, 1979), contains a deletion near the 3' end of pol, whereas the other, SE52d (linial et al., Virology 87:130-141, 1978), may have inserted a host cell sequence near the 5' end of pol. Neither mutant synthesized Pr180gag-pol or Pr130gag-pol, but instead produced novel proteins comprised of sequences shared with gag proteins plus a region antigenically related to reverse transcriptase. Both proteins were defective as precursors to reverse transcriptase. Whereas Pr180gag-pol and Pr130gag-pol were precipitated by an antiserum raised against p32 (a virion protein derived from the portion of the beta subunit removed during processing of beta to alpha [Schiff and Grandgenett, J. Virol. 28:279-291, 1978]), the novel protein synthesized by PH9 ws not precipitated. This suggets that the alpha subunit is generated by a COOH-terminal cleavage of the beta subunit.

Identification of a Rous sarcoma virus transformation-related protein in normal avian and mammalian cells

Avian sarcoma viruses (ASV) contain a gene (src) whose protein product mediates sarcomagenic transformation. This product is a 60,000-Mr phosphoprotein designated pp60src. We have found that normal uninfected frog, chicken, rat, and human cells contain a 60,000-Mr phosphoprotein related to the product of the ASV src gene and have designated that protein pp60. A phosphoprotein of similar size was not detectable in Drosophila cells. The pp60 proteins were detected by immunoprecipitation with rabbit antitumor serum containing broad spectrum antibodies to pp60src. Peptide maps of [35S]methionine-labeled pp60 and pp60src indicated major similarities as well as some differences in amino acid composition. Peptide maps of the 32P-labeled proteins demonstrated that the phosphopeptides of all endogenous pp60 molecules tested were identical. However, some differences were noted between the phosphopeptide patterns of pp60 and viral pp60src. The kinase activity associated with pp60src was measured in the immunocomplex and resulted in the transfer of radioactive phosphorus from [gamma-32P]ATP to the immunoglobulin heavy chain as well as to an 80,000-Mr phosphoprotein. The pp60 of chicken, rat, and human origin also contained an associated kinase activity. These results are consistent with the notion that the pp60 molecules are the protein products of endogenous sarc sequences found in vertebrate cells.