Structure and expression of the human L-myc gene reveal a complex pattern of alternative mRNA processing

Epigenetic landscape of small cell lung cancer: small image of a giant recalcitrant disease

Small cell lung cancer (SCLC) is a particular subtype of lung cancer with high mortality. Recent advances in understanding SCLC genomics and breakthroughs of immunotherapy have substantially expanded existing knowledge and treatment modalities. However, challenges associated with SCLC remain enigmatic and elusive. Most of the conventional drug discovery approaches targeting altered signaling pathways in SCLC end up in the 'grave-yard of drug discovery', which mandates exploring novel approaches beyond inhibiting cell signaling pathways. Epigenetic modifications have long been documented as the key contributors to the tumorigenesis of almost all types of cancer, including SCLC. The last decade witnessed an exponential increase in our understanding of epigenetic modifications for SCLC. The present review highlights the central role of epigenetic regulations in acquiring neoplastic phenotype, metastasis, aggressiveness, resistance to chemotherapy, and immunotherapeutic approaches of SCLC. Different types of epigenetic modifications (DNA/histone methylation or acetylation) that can serve as predictive biomarkers for prognostication, treatment stratification, neuroendocrine lineage determination, and development of potential SCLC therapies are also discussed. We also review the utility of epigenetic targets/epidrugs in combination with first-line chemotherapy and immunotherapy that are currently under investigation in preclinical and clinical studies. Altogether, the information presents the inclusive landscape of SCLC epigenetics and epidrugs that will help to improve SCLC outcomes.

Family matters: How MYC family oncogenes impact small cell lung cancer

Small cell lung cancer (SCLC) is one of the most deadly cancers and currently lacks effective targeted treatment options. Recent advances in the molecular characterization of SCLC has provided novel insight into the biology of this disease and raises hope for a paradigm shift in the treatment of SCLC. We and others have identified activation of MYC as a driver of susceptibility to Aurora kinase inhibition in SCLC cells and tumors that translates into a therapeutic option for the targeted treatment of MYC-driven SCLC. While MYC shares major features with its paralogs MYCN and MYCL, the sensitivity to Aurora kinase inhibitors is unique for MYC-driven SCLC. In this review, we will compare the distinct molecular features of the 3 MYC family members and address the potential implications for targeted therapy of SCLC.

L-myc genotype is associated with different susceptibility to lung cancer in smokers

We have shown that L-myc genotype is associated with the risk of esophageal cancer from smoking and heavy drinking. In this study, we have analyzed the relationship between the L-myc genotypes and lung cancer risk from smoking in 191 Japanese lung-cancer patients and 241 non-cancer controls. The odds ratios (ORs) were markedly higher in SS and LS genotypes than in LL genotype; age-sex-adjusted ORs were 3.19, 2.30 and 0.92, respectively. This result suggests that the L-myc polymorphism may affect the induction of lung cancer by smoking. The OR for smoking in SS-genotype patients diagnosed within 2 years was higher than that in other SS patients, suggesting that smoking-related lung cancer in SS genotype might exhibit a poorer prognosis.

Different susceptibility of each L-myc genotype to esophageal cancer risk factors

To understand the relationship between the L-myc genotypes and esophageal cancer risk, a polymerase chain reaction-based restriction fragment length polymorphism analysis was performed on 91 Japanese patients with esophageal cancer and 241 non-cancer outpatients. No significant difference in the distribution of genotypes was observed between patients and controls; 18.7% LL genotype, 56.0% LS and 25.3% SS among patients, and 24.5%, 55.6% and 19.9%, respectively, among controls. Frequency of the s-allele in patients (0.533) was slightly higher than in controls (0.477), but the difference was not statistically significant. However, the odds ratios (ORs) for smoking or heavy drinking were markedly higher in SS and LS genotypes than in LL genotype; age-sex-adjusted ORs for smoking was 7.57 in the SS genotype, 6.40 in the LS genotype and 1.77 in the LL genotype. Age-sex-adjusted ORs for heavy drinking were 19.78, 18.20 and 7.40, respectively. The age-sex-adjusted ORs for both factors combined were 12.77, 18.45 and 1.44, respectively. These results suggested that the L-myc polymorphism might modify the effects of lifestyle factors on esophageal cancer risk.

L-myc restriction fragment length polymorphism in Japanese patients with esophageal cancer

L-myc polymorphism is a representative genetic trait related to an individual's susceptibility to several cancers. However, there have been no reports concerning the association between esophageal cancer and L-myc polymorphism. To analyze the distribution of polymorphism in Japanese patients with esophageal cancer, a molecular genotyping method using a polymerase chain reaction-based restriction fragment length polymorphism (PCR-RFLP) was used. Based on an analysis of 65 Japanese patients with esophageal cancer and 107 healthy control subjects, a significant difference was observed in either the distribution of genotypes (P=0.012) or of allele frequencies between the two groups (P=0.004). The relative risk of esophageal cancer for genotypes including the shorter allele was 2.9 compared to the longer allele homozygote. Furthermore, the patients with S-allele had a tendency for poor prognosis among those with three genotypes. A significant difference between the distribution of genotypes and the incidence of lymph node metastasis was found based on the clinicopathological features of the cancers. These results suggest that L-myc polymorphism may be implicated as a genetic trait affecting an individual's susceptibility to esophageal cancer, at least among Japanese patients.

Diverse karyotypic abnormalities of the c-myc locus associated with c-myc dysregulation and tumor progression in multiple myeloma

Translocations involving c-myc and an Ig locus have been reported rarely in human multiple myeloma (MM). Using specific fluorescence in situ hybridization probes, we show complex karyotypic abnormalities of the c-myc or L-myc locus in 19 of 20 MM cell lines and approximately 50% of advanced primary MM tumors. These abnormalities include unusual and complex translocations and insertions that often juxtapose myc with an IgH or IgL locus. For two advanced primary MM tumors, some tumor cells contain a karyotypic abnormality of the c-myc locus, whereas other tumor cells do not, indicating that this karyotypic abnormality of c-myc occurs as a late event. All informative MM cell lines show monoallelic expression of c-myc. For Burkitt's lymphoma and mouse plasmacytoma tumors, balanced translocation that juxtaposes c-myc with one of the Ig loci is an early, invariant event that is mediated by B cell-specific DNA modification mechanisms. By contrast, for MM, dysregulation of c-myc apparently is caused principally by complex genomic rearrangements that occur during late stages of MM progression and do not involve B cell-specific DNA modification mechanisms.

MYCL genotypes and loss of heterozygosity in non-small-cell lung cancer

Some studies have suggested that the S allele of the MYCL oncogene, which results from an intragenic EcoRI restriction fragment length polymorphism (RFLP), may be associated with cancer susceptibility. In addition, this allele has also been linked to metastases and adverse survival in certain cancers, although studies of lung cancer patients from different populations have yielded controversial results. We studied 108 cases of surgical resected non-small-cell lung cancer (NSCLC) and found no evidence that MYCL genotypes were associated with tumour progression or a worse prognosis. However, the presence of loss of heterozygosity (LOH) at this chromosome 1p32 locus correlated significantly with regional lymph node involvement, as well as advanced TNM stage. These data indicate the existence of a chromosome 1p candidate tumour-suppressor gene(s), possibly in linkage disequilibrium with the EcoRI RFLP in specific populations, which appears to play a role in determining tumour progression in NSCLC. Refined mapping of the critical region of loss should help attempts to identify and clone the candidate gene.

Expression and activity of L-Myc in normal mouse development

To determine the role of L-Myc in normal mammalian development and its functional relationship to other members of the Myc family, we determined the normal patterns of L-myc gene expression in the developing mouse by RNA in situ hybridization and assessed the phenotypic impact of L-Myc deficiency produced through standard gene targeting methodology. L-myc transcripts were detected in the developing kidney and lung as well as in both the proliferative and the differentiative zones of the brain and neural tube. Despite significant expression of L-myc in developing mouse tissue, homozygous null L-myc mice were found to be viable, reproductively competent, and represented in expected frequencies from heterozygous matings. A detailed histological survey of embryonic and adult tissues, characterization of an embryonic neuronal marker, and measurement of cellular proliferation in situ did not reveal any congenital abnormalities. The lack of an apparent phenotype associated with L-Myc deficiency indicates that L-Myc is dispensable for gross morphological development and argues against a unique role for L-Myc in early central nervous system development as had been previously suggested. Although overlapping expression patterns among myc family members raise the possibility of complementation of L-Myc deficiency by other Myc oncoproteins, compensatory changes in the levels of c- and/or N-myc transcripts were not detected in homozygous null L-myc mice.

Differential expression of myc, max and RB1 genes in human gliomas and glioma cell lines

Deregulated expression of myc proto-oncogenes is implicated in several human neoplasias. We analysed the expression of c-myc, N-myc, L-myc, max and RB1 mRNAs in a panel of human gliomas and glioma cell lines and compared the findings with normal neural cells. The max and RB1 genes were included in the study because their protein products can interact with the Myc proteins, being thus putative modulators of Myc activity. Several gliomas contained c/L-myc mRNAs at levels higher than those in fetal brain, L-myc predominantly in grade II/III and c-myc in grade III gliomas. High-level N-myc expression was detected. In one small-cell glioblastoma and lower levels in five other gliomas. In contrast, glioma cell lines totally lacked N/L-myc expression. The in situ hybridisations revealed mutually exclusive topographic distribution of myc and glial fibrillary acidic protein (GFAP) mRNAs, and a lack of correlation between myc expression and proliferative activity, max and RB1 mRNAs were detected in most tumours and cell lines. The glioma cells displayed interesting alternative splicing patterns of max mRNAs encoding Max proteins which either suppress (Max) or augment (delta Max) the transforming activity of Myc. We conclude that (1) glioma cells in vivo may coexpress several myc genes, thus resembling fetal neural cells; but (2) cultured glioma cells expression only c-myc; (3) myc, max and RB1 are regulated independently in glioma cells; and (4) alternative processing of max mRNA in some glioma cells results in delta Max encoding mRNAs not seen in normal fetal brain.

B-myc inhibits neoplastic transformation and transcriptional activation by c-myc

B-myc is a recently described myc gene whose product has not been functionally characterized. The predicted product of B-myc is a 168-amino-acid protein with extensive homology to the c-Myc amino-terminal region, previously shown to contain a transcriptional activation domain. We hypothesized that B-Myc might also function in transcriptional regulation, although its role in regulating gene expression is predicted to be unique, because B-Myc lacks the specific DNA-binding motif found in other Myc proteins. To determine whether B-Myc could interact with the transcriptional machinery, we studied the transcriptional activation properties of a chimeric protein containing B-Myc sequences fused to the DNA-binding domain of the yeast transcriptional activator GAL4 (GAL4-B-Myc). We found that GAL4-B-Myc strongly activated expression of a GAL4-regulated reporter gene in mammalian cells. In addition, full-length B-Myc was able to inhibit or squelch reporter gene activation by a GAL4 chimeric protein containing the c-Myc transcriptional activation domain. We also observed that B-Myc dramatically inhibited the neoplastic cotransforming activity of c-Myc and activated Ras in rat embryo cells. Because B-Myc inhibits both neoplastic transformation and transcriptional activation by c-Myc, we suggest that the transforming activity of c-Myc is related to its ability to regulate transcription. Whether B-Myc functions biologically to squelch transcription and/or to regulate transcription through a specific DNA-binding protein remains unestablished.

Differential patterns of DNA binding by myc and max proteins

c-myc, N-myc, and L-myc genes are subject to highly variable degrees of tissue-specific regulation. Their aberrant expression has also been implicated in the pathogenesis of a variety of malignant tumors. The recently identified max protein dimerizes with c-myc to promote its sequence-specific DNA binding. max exists in two forms (long and short) that differ by virtue of a 9-amino acid insertion/deletion at the N terminus. We tested recombinant myc and max proteins for binding to six oligonucleotides containing related c-myc sites. Each myc protein, alone and in association with max proteins, manifested a unique pattern of DNA binding. Phosphorylation of both max proteins was observed when they were incubated in a rabbit reticulocyte lysate. This strongly affected DNA binding by max(long) but not by max(short). Our results point to the existence of specific DNA binding preferences for each of the myc proteins. The 9-amino acid segment that distinguishes max(long) from max(short) appears to serve a regulatory function that provides additional control over DNA sequence recognition.

B-myc inhibits neoplastic transformation and transcriptional activation by c-myc

B-myc is a recently described myc gene whose product has not been functionally characterized. The predicted product of B-myc is a 168-amino-acid protein with extensive homology to the c-Myc amino-terminal region, previously shown to contain a transcriptional activation domain. We hypothesized that B-Myc might also function in transcriptional regulation, although its role in regulating gene expression is predicted to be unique, because B-Myc lacks the specific DNA-binding motif found in other Myc proteins. To determine whether B-Myc could interact with the transcriptional machinery, we studied the transcriptional activation properties of a chimeric protein containing B-Myc sequences fused to the DNA-binding domain of the yeast transcriptional activator GAL4 (GAL4-B-Myc). We found that GAL4-B-Myc strongly activated expression of a GAL4-regulated reporter gene in mammalian cells. In addition, full-length B-Myc was able to inhibit or squelch reporter gene activation by a GAL4 chimeric protein containing the c-Myc transcriptional activation domain. We also observed that B-Myc dramatically inhibited the neoplastic cotransforming activity of c-Myc and activated Ras in rat embryo cells. Because B-Myc inhibits both neoplastic transformation and transcriptional activation by c-Myc, we suggest that the transforming activity of c-Myc is related to its ability to regulate transcription. Whether B-Myc functions biologically to squelch transcription and/or to regulate transcription through a specific DNA-binding protein remains unestablished.

Activation domains of L-Myc and c-Myc determine their transforming potencies in rat embryo cells

Members of the Myc family of proteins share a number of protein motifs that are found in regulators of gene transcription. Conserved stretches of amino acids found in the N-terminal transcriptional activation domain of c-Myc are required for cotransforming activity. Most of the Myc proteins contain the basic helix-loop-helix zipper (bHLH-Zip) DNA-binding motif which is also required for the cotransforming activity of c-Myc. L-Myc, the product of a myc family gene that is highly amplified in many human lung carcinomas, was found to cotransform primary rat embryo cells with an activated ras gene. However, L-Myc cotransforming activity was only 1 to 10% of that of c-Myc (M. J. Birrer, S. Segal, J. S. DeGreve, F. Kaye, E. A. Sausville, and J. D. Minna, Mol. Cell. Biol. 8:2668-2673, 1988). We sought to determine whether functional differences between c-Myc and L-Myc in either the N-terminal or the C-terminal domain could account for the relatively diminished L-Myc cotransforming activity. Although the N-terminal domain of L-Myc could activate transcription when fused to the yeast GAL4 DNA-binding domain, the activity was only 5% of that of a comparable c-Myc domain. We next determined that the interaction of the C-terminal bHLH-Zip region of L-Myc or c-Myc with that of a Myc partner protein, Max, was equivalent in transfected cells. A Max expression vector was found to augment the cotransforming activity of L-Myc as well as that of c-Myc. In addition, a bacterially synthesized DNA-binding domain of L-Myc, like that o c-Myc, heterodimerizes with purified Max protein to bind the core DNA sequence CACGTG. To determine the region of L-Myc responsible for its relatively diminished cotransforming activity, we constructed chimeras containing exons 2 (constituting activation domains) and 3 (constituting DNA-binding domains) of c-Myc fused to those of L-Myc. The cotransforming potencies of these chimeras were compared with those of full-length L-Myc of c-Myc in rat embryo cells. The relative cotransforming activities suggest that the potencies of the activation domains determine the cotransforming efficiencies for c-Myc and L-Myc. This correlation supports the hypothesis that the Myc proteins function in neoplastic cotransformation as transcription factors.

Activation domains of L-Myc and c-Myc determine their transforming potencies in rat embryo cells

Members of the Myc family of proteins share a number of protein motifs that are found in regulators of gene transcription. Conserved stretches of amino acids found in the N-terminal transcriptional activation domain of c-Myc are required for cotransforming activity. Most of the Myc proteins contain the basic helix-loop-helix zipper (bHLH-Zip) DNA-binding motif which is also required for the cotransforming activity of c-Myc. L-Myc, the product of a myc family gene that is highly amplified in many human lung carcinomas, was found to cotransform primary rat embryo cells with an activated ras gene. However, L-Myc cotransforming activity was only 1 to 10% of that of c-Myc (M. J. Birrer, S. Segal, J. S. DeGreve, F. Kaye, E. A. Sausville, and J. D. Minna, Mol. Cell. Biol. 8:2668-2673, 1988). We sought to determine whether functional differences between c-Myc and L-Myc in either the N-terminal or the C-terminal domain could account for the relatively diminished L-Myc cotransforming activity. Although the N-terminal domain of L-Myc could activate transcription when fused to the yeast GAL4 DNA-binding domain, the activity was only 5% of that of a comparable c-Myc domain. We next determined that the interaction of the C-terminal bHLH-Zip region of L-Myc or c-Myc with that of a Myc partner protein, Max, was equivalent in transfected cells. A Max expression vector was found to augment the cotransforming activity of L-Myc as well as that of c-Myc. In addition, a bacterially synthesized DNA-binding domain of L-Myc, like that o c-Myc, heterodimerizes with purified Max protein to bind the core DNA sequence CACGTG. To determine the region of L-Myc responsible for its relatively diminished cotransforming activity, we constructed chimeras containing exons 2 (constituting activation domains) and 3 (constituting DNA-binding domains) of c-Myc fused to those of L-Myc. The cotransforming potencies of these chimeras were compared with those of full-length L-Myc of c-Myc in rat embryo cells. The relative cotransforming activities suggest that the potencies of the activation domains determine the cotransforming efficiencies for c-Myc and L-Myc. This correlation supports the hypothesis that the Myc proteins function in neoplastic cotransformation as transcription factors.

Complex intrachromosomal rearrangement in the process of amplification of the L-myc gene in small-cell lung cancer

The L-myc gene was first isolated from a human small-cell lung cancer (SCLC) cell line on the basis of its amplification and sequence similarity to c-myc and N-myc. A new mechanism of L-myc activation which results from the production of rlf-L-myc fusion protein was recently reported. On the basis of our earlier observation of a rearrangement involving amplified L-myc in an SCLC cell line, ACC-LC-49, we decided to investigate this rearrangement in detail along with the structure of L-myc amplification units in five additional SCLC cell lines. We report here the identification of a novel genomic region, termed jal, which is distinct from rlf and is juxtaposed to and amplified with L-myc during the process of DNA amplification of the region encompassing L-myc. Long-range analysis using pulsed-field gel electrophoresis revealed that the amplified L-myc locus is involved in highly complex intrachromosomal rearrangements with jal and/or rlf. Our results also suggest that the simultaneous presence of rearrangements both in rlf intron 1 and in regions immediately upstream of L-myc may be necessary for the expression of rlf-L-myc chimeric transcripts.

Complex intrachromosomal rearrangement in the process of amplification of the L-myc gene in small-cell lung cancer

The L-myc gene was first isolated from a human small-cell lung cancer (SCLC) cell line on the basis of its amplification and sequence similarity to c-myc and N-myc. A new mechanism of L-myc activation which results from the production of rlf-L-myc fusion protein was recently reported. On the basis of our earlier observation of a rearrangement involving amplified L-myc in an SCLC cell line, ACC-LC-49, we decided to investigate this rearrangement in detail along with the structure of L-myc amplification units in five additional SCLC cell lines. We report here the identification of a novel genomic region, termed jal, which is distinct from rlf and is juxtaposed to and amplified with L-myc during the process of DNA amplification of the region encompassing L-myc. Long-range analysis using pulsed-field gel electrophoresis revealed that the amplified L-myc locus is involved in highly complex intrachromosomal rearrangements with jal and/or rlf. Our results also suggest that the simultaneous presence of rearrangements both in rlf intron 1 and in regions immediately upstream of L-myc may be necessary for the expression of rlf-L-myc chimeric transcripts.

High frequency of myelomonocytic tumors in aging E mu L-myc transgenic mice

Transgenic mice that contain constructs of the L-myc gene under the transcriptional control of the immunoglobulin heavy chain enhancer (E mu) develop thymic hyperplasia and are predisposed to T cell lymphomas. Here we describe a second form of malignancy that occurs in aging E mu L-myc transgenic mice. The mean latency period for the development of this malignancy is longer compared with the E mu L-myc T cell lymphomas but the overall incidence is increased threefold. The histopathological morphology is that of a highly malignant mesenchymal neoplasm that closely resembles human fibrous histiocytoma. The tumor cells were classified as myelomonocytic on the basis of several lineage-specific markers and the lack of rearrangements of the immunoglobulin heavy chain and the T cell receptor beta loci. Cultured tumor cells produce macrophage colony-stimulating factor (M-CSF) protein and express the M-CSF receptor, suggesting the involvement of an autocrine loop in this malignancy. Similar to the E mu L-myc T cell lymphomas, these tumors show high-level transgene expression but no detectable levels of endogenous c-myc mRNA, directly implicating the deregulated expression of L-myc in the generation of this malignancy. E mu L-myc myelomonocytic tumors show consistent trisomy of chromosome 16, implicating this as a secondary event in the development of this tumor. In the light of recent findings that L-myc is expressed in human myeloid leukemias and in several human myeloid tumor cell lines, the results described here might implicate L-myc in the development of naturally occurring myeloid neoplasias.

Spontaneous deletions in Ig heavy chain genes: flanking sequences influence splice site selection

The cell line G403.4.7, isolated as a spontaneous variant of the MPC-11 derived myeloma G403.4, produces a truncated gamma 2b HC protein, but no light chain (LC), and a single gamma 2b specific transcript of 2.4kb. This gamma 2b transcript consists of the VDJ and CH1 exons, the CH1 to Hinge (Hi) intervening sequence (IVS) and HI exon, part of the IVS between the two membrane exons M1 and M2, and most of the membrane 3' untranslated (UT) region. Even though the mature mRNA contains intronic sequences, it is abundant in the cytoplasm. Analysis of the gamma 2b genomic organization reveals that this unusual transcript results in part from two genomic deletions of 2.5kb and 588bp and in part from an altered splicing pattern. This altered splicing pattern is probably a consequence of the sequence alterations resulting from the genomic deletions. Analysis of these events provides some interesting insights into the mechanism of splice site selection and the evolution of introns and exons.

An analysis of vertebrate mRNA sequences: intimations of translational control

Five structural features in mRNAs have been found to contribute to the fidelity and efficiency of initiation by eukaryotic ribosomes. Scrutiny of vertebrate cDNA sequences in light of these criteria reveals a set of transcripts--encoding oncoproteins, growth factors, transcription factors, and other regulatory proteins--that seem designed to be translated poorly. Thus, throttling at the level of translation may be a critical component of gene regulation in vertebrates. An alternative interpretation is that some (perhaps many) cDNAs with encumbered 5' noncoding sequences represent mRNA precursors, which would imply extensive regulation at a posttranscriptional step that precedes translation.

Testis-specific expression of the human MYCL2 gene

We have characterized the expression of MYCL2, an intronless X-linked gene related to MYCL1. RNase protection analysis of a panel of human normal and tumor tissues has revealed that MYCL2 is expressed almost exclusively in human adult normal testis; much lower levels of transcript were detected in one human lung adenocarcinoma. No MYCL2 transcript was found in human testis RNA obtained from second trimester fetuses. This observation suggests a germ cell rather than somatic cell origin of the transcript and possible developmental regulation of MYCL2. Northern blot analysis of poly(A)+ RNA from adult human normal testis with an antisense riboprobe revealed a transcript of approximately 4.8-kb, which is in agreement with the size predicted from the MYCL2 nucleotide sequence. Antisense transcripts were found spanning regions of MYCL2 corresponding to all three exons of MYCL1. No sizable open reading frame was seen for the MYCL2 antisense transcripts suggesting that they may represent either regulatory sequences or an intron of a gene encoded by the complementary strand. RNase protection assays and the 5' RACE protocol (Rapid Amplification of cDNA Ends) were used to address the localization of the transcription start site of the MYCL2 sense transcript and different putative promoters and transcription regulatory elements have been identified.

Contrasting expression patterns of three members of the myc family of protooncogenes in the developing and adult mouse kidney

The myc family of protooncogenes encode similar but distinct nuclear proteins. Since N-myc, c-myc, and L-myc have been found to be expressed in the newborn kidney, we studied their expression during murine kidney development. By organ culture studies and in situ hybridization of tissue sections, we found that each of the three members of the myc gene family shows a remarkably distinct expression pattern during kidney development. It is known that mesenchymal stem cells of the embryonic kidney convert into epithelium if properly induced. We demonstrate the N-myc expression increases during the first 24 h of in vitro culture as an early response to induction. Moreover, the upregulation was transient and expression levels were already low during the first stages of overt epithelial cell polarization. In contrast, neither c-myc nor L-myc were upregulated by induction of epithelial differentiation. c-myc was expressed in the uninduced mesenchyme but subsequently became restricted to the newly formed epithelium and was not expressed in the surrounding loose mesenchyme. At onset of terminal differentiation c-myc expression was turned off also from the epithelial tubules. We conclude that N-myc is a marker for induction and early epithelial differentiation states. That the undifferentiated mesenchyme, unlike stromal cells of later developmental stages, express c-myc demonstrates that the undifferentiated mesenchymal stem cells are distinct from the stromal cells. The most astonishing finding, however, was the high level of L-myc mRNA in the ureter, ureter-derived renal pelvis, papilla, and collecting ducts. In the ureter, expression increased, rather than decreased, with advancing maturation and was highest in adult tissue. Our results suggest that each of the three members of the myc gene family are involved in quite disparate differentiation processes, even within one tissue.

RNA processing is a limiting step for murine tumor necrosis factor beta expression in response to interleukin-2

We have previously reported that tumor necrosis factor beta (TNF beta) expression is induced by interleukin-2 (IL-2) in the murine lymphocytic T-cell line CTLL-2. In this study, we have characterized the nuclear and cytoplasmic TNF beta transcript and assessed their role in TNF beta gene expression. A unique feature of TNF beta expression was the accumulation of nuclear precursors, which reflected a slow nuclear RNA processing. As a consequence, there was a delay in the appearance of cytoplasmic messengers after the transcriptional induction of TNF beta by IL-2. We also found that two messengers, the fully spliced messenger and an intron 3-retaining messenger, were exported to the cytoplasm and actively translated. The same pattern of expression was observed in concanavalin A-stimulated splenocytes, although the level of expression was much lower than in CTLL-2 cells. The simple genetic structure and the high level of accumulation of nuclear precursors make TNF beta a particularly attractive model system to use for studies of RNA processing and cytoplasmic transport of partially spliced messengers.

RNA processing is a limiting step for murine tumor necrosis factor beta expression in response to interleukin-2

We have previously reported that tumor necrosis factor beta (TNF beta) expression is induced by interleukin-2 (IL-2) in the murine lymphocytic T-cell line CTLL-2. In this study, we have characterized the nuclear and cytoplasmic TNF beta transcript and assessed their role in TNF beta gene expression. A unique feature of TNF beta expression was the accumulation of nuclear precursors, which reflected a slow nuclear RNA processing. As a consequence, there was a delay in the appearance of cytoplasmic messengers after the transcriptional induction of TNF beta by IL-2. We also found that two messengers, the fully spliced messenger and an intron 3-retaining messenger, were exported to the cytoplasm and actively translated. The same pattern of expression was observed in concanavalin A-stimulated splenocytes, although the level of expression was much lower than in CTLL-2 cells. The simple genetic structure and the high level of accumulation of nuclear precursors make TNF beta a particularly attractive model system to use for studies of RNA processing and cytoplasmic transport of partially spliced messengers.

A single amino acid substitution results in a retinoblastoma protein defective in phosphorylation and oncoprotein binding

We have previously identified a small-cell lung cancer cell line (NCI-H209) that expresses an aberrant, underphosphorylated form of the retinoblastoma protein RB1. Molecular analysis of RB1 mRNA from this cell line revealed a single point mutation within exon 21 that resulted in a nonconservative amino acid substitution (cysteine to phenylalanine) at codon 706. Stable expression of this mutant RB1 cDNA in a human cell line lacking endogenous RB1 demonstrated that this amino acid change was sufficient to inhibit phosphorylation. In addition, this cysteine-to-phenylalanine substitution also resulted in loss of RB1 binding to the simian virus 40 large tumor and adenovirus E1A transforming proteins. These results confirm the importance of exon 21 coding sequences and suggest that the cysteine residue at codon 706 may play a role in achieving a specific protein conformation essential for protein-protein interactions.

N-myc mRNA forms an RNA-RNA duplex with endogenous antisense transcripts

Nuclear runoff transcription studies revealed nearly equivalent sense and antisense transcription across exon 1 of the N-myc locus. Antisense primary transcription initiates at multiple sites in intron 1 and gives rise to stable polyadenylated and nonpolyadenylated transcripts. This pattern of antisense transcription, which is directed by RNA polymerase II, is independent of gene amplification and cell type. The nonpolyadenylated antisense transcripts have 5' ends which are complementary to the 5' ends of the N-myc sense mRNA. We determined, by using an RNase protection technique designed to detect in vivo duplexes, that most of the cytoplasmic nonpolyadenylated antisense RNA exists in an RNA-RNA duplex with approximately 5% of the sense N-myc mRNA. Duplex formation appeared to occur with only a subset of the multiple forms of the N-myc mRNA, with the precise transcriptional initiation site of the RNA playing a role in determining this selectivity. Cloning of each strand of the RNA-RNA duplex revealed that most duplexes included both exon 1 and intron 1 sequences, suggesting that duplex formation could modulate RNA processing by preserving a population of N-myc mRNA which retains intron 1.

N-myc mRNA forms an RNA-RNA duplex with endogenous antisense transcripts

Nuclear runoff transcription studies revealed nearly equivalent sense and antisense transcription across exon 1 of the N-myc locus. Antisense primary transcription initiates at multiple sites in intron 1 and gives rise to stable polyadenylated and nonpolyadenylated transcripts. This pattern of antisense transcription, which is directed by RNA polymerase II, is independent of gene amplification and cell type. The nonpolyadenylated antisense transcripts have 5' ends which are complementary to the 5' ends of the N-myc sense mRNA. We determined, by using an RNase protection technique designed to detect in vivo duplexes, that most of the cytoplasmic nonpolyadenylated antisense RNA exists in an RNA-RNA duplex with approximately 5% of the sense N-myc mRNA. Duplex formation appeared to occur with only a subset of the multiple forms of the N-myc mRNA, with the precise transcriptional initiation site of the RNA playing a role in determining this selectivity. Cloning of each strand of the RNA-RNA duplex revealed that most duplexes included both exon 1 and intron 1 sequences, suggesting that duplex formation could modulate RNA processing by preserving a population of N-myc mRNA which retains intron 1.

Studies of the L-myc DNA polymorphism and relation to metastasis in Norwegian lung cancer patients

We studied 83 lung cancer patients and 129 controls for the EcoRI polymorphism of the L-myc gene. No association was found between the L-myc RFLP and increased risk of metastasis, either to lymph nodes or metastasis to other organs. There was no difference in survival time between the three different genotypes. The S-allele of the L-myc RFLP has been correlated to increased metastasis in lung cancer. We found no tendency towards a higher frequency of this allele in the cohort of patients with positive family history compared to the patients with no known first degree relatives with cancer. A higher frequency of the S-allele in the adenocarcinomas compared to other histological groups was found, although this difference was not statistically significant. No difference in the gene frequency of the L-myc RFLP was found between the lung cancer patients and the normal controls. These results are in contrast with a previous report. Possible explanations for the discrepancies are discussed.

Organization and expression of the chicken N-myc gene

We cloned the chicken N-myc gene and analyzed its structure and expression. We found that it consisted of three exons with coding regions in exons 2 and 3. Comparison to mammalian N-myc genomic sequence indicated that nucleotide sequences of the 5'-flanking region, noncoding exon 1, and introns were not conserved, but coding and 3' noncoding sequences showed significant homology to mammalian N-myc. Alignment of deduced amino acid sequences of chicken and mammalian N-myc proteins revealed nine conserved domains interrupted by different lengths of nonhomologous sequences. Two of the domains were specific to N-myc proteins, and the other seven were common to c-myc proteins. Northern blot (immunoblot) and in situ hybridization analyses of 3.5-day-old chicken embryos revealed that high-level expression of the N-myc gene was confirmed to certain tissues, e.g., the central nervous system, neural crest derivatives, and mesenchyme of limb buds. In the beak and limb primordia, N-myc expression in the mesenchyme was higher toward the distal end, suggesting possible involvement in positional assignment of the tissue within the rudimentary structures.

Organization and expression of the chicken N-myc gene

We cloned the chicken N-myc gene and analyzed its structure and expression. We found that it consisted of three exons with coding regions in exons 2 and 3. Comparison to mammalian N-myc genomic sequence indicated that nucleotide sequences of the 5'-flanking region, noncoding exon 1, and introns were not conserved, but coding and 3' noncoding sequences showed significant homology to mammalian N-myc. Alignment of deduced amino acid sequences of chicken and mammalian N-myc proteins revealed nine conserved domains interrupted by different lengths of nonhomologous sequences. Two of the domains were specific to N-myc proteins, and the other seven were common to c-myc proteins. Northern blot (immunoblot) and in situ hybridization analyses of 3.5-day-old chicken embryos revealed that high-level expression of the N-myc gene was confirmed to certain tissues, e.g., the central nervous system, neural crest derivatives, and mesenchyme of limb buds. In the beak and limb primordia, N-myc expression in the mesenchyme was higher toward the distal end, suggesting possible involvement in positional assignment of the tissue within the rudimentary structures.

v-maf, a viral oncogene that encodes a "leucine zipper" motif

We have molecularly cloned the provirus of the avian musculoaponeurotic fibrosarcoma virus AS42. Nucleotide sequence analysis of a biologically active clone of AS42 showed that this virus encodes a viral oncogene, maf. The deduced amino acid sequence of the v-maf gene product contains a "leucine zipper" motif similar to that found in a number of DNA binding proteins, including the gene products of the fos, jun, and myc oncogenes. However, unlike these oncogenes, the cellular maf gene was not transcriptionally activated by growth stimulation of cultured cells.

DNA-binding domain of human c-Myc produced in Escherichia coli

We have identified the domain of the human c-myc protein (c-Myc) produced in Escherichia coli that is responsible for the ability of the protein to bind sequence-nonspecific DNA. Using analysis of binding of DNA by proteins transferred to nitrocellulose, DNA-cellulose chromatography, and a nitrocellulose filter binding assay, we examined the binding properties of c-Myc peptides generated by cyanogen bromide cleavage, of mutant c-Myc, and of proteins that fuse portions of c-Myc to staphylococcal protein A. The results of these analyses indicated that c-Myc amino acids 265 to 318 were responsible for DNA binding and that other regions of the protein (including a highly conserved basic region and a region containing the leucine zipper motif) were not required. Some mutant c-Mycs that did not bind DNA maintained rat embryo cell-cotransforming activity, which indicated that the c-Myc property of in vitro DNA binding was not essential for this activity. These mutants, however, were unable to transform established rat fibroblasts (Rat-1a cells) that were susceptible to transformation by wild-type c-Myc, although this lack of activity may not have been due to their inability to bind DNA.

A transfected L-myc gene can substitute for c-myc in blocking murine erythroleukemia differentiation

We investigated the ability of the proto-oncogene L-myc to substitute for c-myc in blocking murine erythroleukemia differentiation. Murine erythroleukemia cells (line C19) were transfected with recombinant plasmids containing genomic and cDNA fragments of the L-myc gene driven by a Moloney murine leukemia virus long terminal repeat. Clones expressing constitutive high levels of L-myc failed to differentiate in response to the chemical inducer N,N'-hexamethylene bisacetamide (HMBA). The block to differentiation correlated with the level of L-myc expression. Furthermore, transfected clones grown in the presence of inducer for an extended period of time showed an increased level of L-myc expression. These results suggest that functional domains of the c-myc gene involved in differentiation are located in the discrete regions of homology between the c- and L-myc genes.

Isolation of human cDNA clones of ski and the ski-related gene, sno

cDNA clones of ski and the ski-related gene, sno, were obtained by screening human cDNA libraries. The predicted open reading frame of h-ski could encode a protein of 728 amino acid residues. The h-ski protein is highly homologous with the v-ski protein. The overall homology between h-ski and v-ski is 91% at the amino acid level. DNA sequencing analysis revealed two types of cDNA clones from the sno (ski-related novel gene) gene, possibly due to alternative splicing. The first type, named snoN (non Alu-containing), encoded a protein of 684 amino acid residues. The second type, named snoA (Alu-containing), encoded a protein of 415 amino acid residues. The first 366 amino acid residues of snoN and snoA are the same, but subsequent amino acids show divergence. Several transcripts of h-ski (6.0, 4.7, 3.8, 3.0, 2.1 and 1.8 kb) were detected. The mRNAs of h-sno were 6.2, 4.4 and 3.2kb.

DNA-binding domain of human c-Myc produced in Escherichia coli

We have identified the domain of the human c-myc protein (c-Myc) produced in Escherichia coli that is responsible for the ability of the protein to bind sequence-nonspecific DNA. Using analysis of binding of DNA by proteins transferred to nitrocellulose, DNA-cellulose chromatography, and a nitrocellulose filter binding assay, we examined the binding properties of c-Myc peptides generated by cyanogen bromide cleavage, of mutant c-Myc, and of proteins that fuse portions of c-Myc to staphylococcal protein A. The results of these analyses indicated that c-Myc amino acids 265 to 318 were responsible for DNA binding and that other regions of the protein (including a highly conserved basic region and a region containing the leucine zipper motif) were not required. Some mutant c-Mycs that did not bind DNA maintained rat embryo cell-cotransforming activity, which indicated that the c-Myc property of in vitro DNA binding was not essential for this activity. These mutants, however, were unable to transform established rat fibroblasts (Rat-1a cells) that were susceptible to transformation by wild-type c-Myc, although this lack of activity may not have been due to their inability to bind DNA.

A transfected L-myc gene can substitute for c-myc in blocking murine erythroleukemia differentiation

We investigated the ability of the proto-oncogene L-myc to substitute for c-myc in blocking murine erythroleukemia differentiation. Murine erythroleukemia cells (line C19) were transfected with recombinant plasmids containing genomic and cDNA fragments of the L-myc gene driven by a Moloney murine leukemia virus long terminal repeat. Clones expressing constitutive high levels of L-myc failed to differentiate in response to the chemical inducer N,N'-hexamethylene bisacetamide (HMBA). The block to differentiation correlated with the level of L-myc expression. Furthermore, transfected clones grown in the presence of inducer for an extended period of time showed an increased level of L-myc expression. These results suggest that functional domains of the c-myc gene involved in differentiation are located in the discrete regions of homology between the c- and L-myc genes.

Deregulated expression of human c-jun transforms primary rat embryo cells in cooperation with an activated c-Ha-ras gene and transforms rat-1a cells as a single gene

While the ability of the retroviral oncogene V-jun to transform chicken cells led to its discovery, the oncogenic potential of its cellular homologue, c-jun, which encodes a transcription factor, is unknown. We isolated a 1070-base-pair cDNA clone containing the unmutated entire open reading frame of c-jun from a human small cell lung cancer line. This cDNA as well as a 5.6-kilobase normal human genomic DNA fragment containing the c-jun gene were placed under the control of retroviral long terminal repeats and introduced into primary rat embryo cells (RECs), with or without other oncogenes, and into an immortal rat fibroblast cell line, Rat-1a, as a single gene. In Rat-1a cells the expression of human c-jun mRNA was associated with the ability to clone in soft agarose and form tumors in nude mice. When the c-jun cDNA or genomic DNA constructs were introduced into RECs, no foci of transformed cells were seen with c-jun alone or c-jun cotransfected with deregulated c-myc or L-myc protooncogenes. However, cotransfection of the c-jun constructs with an activated human c-Ha-ras gene led to foci of transformed cells which gave rise to immortalized cell lines that cloned in soft agarose and formed tumors in nude mice. Furthermore, formation of foci of transformed RECs by the c-jun/ras combination was augmented 3-fold by the tumor promoter phorbol 12-tetradecanoate 13-acetate. We conclude that deregulated expression of human c-jun can participate in malignant transformation of normal mammalian cells.

Multiple mechanisms for transcriptional regulation of the myc gene family in small-cell lung cancer

The molecular mechanisms reported to regulate the expression of myc family genes are multiple and complex and include gene amplification, transcriptional activation, transcriptional attenuation, and mRNA stability. We have investigated which of these mechanisms are responsible for the extreme variation in myc gene family mRNA levels observed in human small-cell lung cancer cell lines. In addition to gene amplification, a block to nascent mRNA chain elongation, causing attenuation of transcription, is an important regulatory mechanism controlling the steady-state levels of c-myc and L-myc mRNA. The loss of transcriptional attenuation is correlated with overexpression of these two genes in cell lines which do not show gene amplification. Expression of c-myc mRNA appears to be dependent on promoter activity and attenuator function. In contrast, regulation of expression of the N-myc gene does not involve transcriptional attenuation; steady-state mRNA levels are correlated with promoter activity as well as gene amplification. We conclude that transcriptional regulation of each member of the myc gene family is accomplished by a different assortment of complex mechanisms, including gene copy number, promoter activation, and transcriptional attenuation. Interference at multiple points in this complex regulatory process appears to be an important mechanism by which small-cell lung cancer and other human tumors evade growth control.

Identification of the human c-myc protein nuclear translocation signal

We identified and characterized two regions of the human c-myc protein that target proteins into the nucleus. Using mutant c-myc proteins and proteins that fuse portions of c-myc to chicken muscle pyruvate kinase, we found that residues 320 to 328 (PAAKRVKLD; peptide M1) induced complete nuclear localization, and their removal from c-myc resulted in mutant proteins that distributed in both the nucleus and cytoplasm but retained rat embryo cell cotransforming activity. Residues 364 to 374 (RQRRNELKRSP; peptide M2) induced only partial nuclear targeting, and their removal from c-myc resulted in mutant proteins that remained nuclear but were cotransformationally inactive. We conjugated synthetic peptides containing M1 or M2 to human serum albumin and microinjected the conjugate into the cytoplasm of Vero cells. The peptide containing M1 caused rapid and complete nuclear accumulation, whereas that containing M2 caused slower and only partial nuclear localization. Thus, M1 functions as the nuclear localization signal of c-myc, and M2 serves some other and essential function.

The human L-myc gene encodes multiple nuclear phosphoproteins from alternatively processed mRNAs

The human proto-oncogene L-myc generates at least four different mRNAs by alternative RNA processing. We have identified two phosphorylated L-myc proteins with molecular masses of 60,000 and 66,000 daltons [p60L-myc(human) and p66L-myc(human)] in a small-cell carcinoma line expressing high levels of L-myc mRNA. These proteins have a short half-life and are localized to the nuclear matrix fraction, as previously reported for the c-myc and N-myc proteins. In vitro translation experiments demonstrated that both the p60 and p66 species are encoded by a 3.9-kilobase (kb) mRNA which retains intron 1, while only the p60 protein is translated from a 3.6-kb L-myc mRNA which has had intron 1 removed. While L-myc proteins [p32L-myc(human) and p37L-myc(human)] could be synthesized in vitro from 2.2-kb mRNA templates, no such proteins were detected by immunoprecipitation in vivo. These observations suggest that alternative RNA processing of the L-myc transcript could play a role in determining the steady-state levels of the p60L-myc and p66L-myc proteins.

Isolation of human cDNA clones of myb-related genes, A-myb and B-myb

cDNA clones of the myb-related genes A-myb and B-myb were obtained by screening human cDNA libraries. The predicted open reading frame of B-myb could encode a protein of 700 amino acid residues. Although the C-terminal end has not been cloned yet, an almost entire coding region of A-myb, which is 745 amino acid long, was determined. The A-myb and B-myb proteins are highly homologous with the myb protein in three regions. Domain I, which is 161 amino acid long, is well conserved in the myb gene family. The homology between human-myb and A-myb in domain I is 90% at the amino acid level. Domain II, which is about 85 amino acid long, is less well conserved. Although it is a short stretch, domain III is found in the C-terminal region. The mRNAs of A-myb and B-myb were 5.0 and 2.6 kb, respectively. The mRNA expression pattern of the myb gene family in various tumors is presented.

Identification of the human c-myc protein nuclear translocation signal

We identified and characterized two regions of the human c-myc protein that target proteins into the nucleus. Using mutant c-myc proteins and proteins that fuse portions of c-myc to chicken muscle pyruvate kinase, we found that residues 320 to 328 (PAAKRVKLD; peptide M1) induced complete nuclear localization, and their removal from c-myc resulted in mutant proteins that distributed in both the nucleus and cytoplasm but retained rat embryo cell cotransforming activity. Residues 364 to 374 (RQRRNELKRSP; peptide M2) induced only partial nuclear targeting, and their removal from c-myc resulted in mutant proteins that remained nuclear but were cotransformationally inactive. We conjugated synthetic peptides containing M1 or M2 to human serum albumin and microinjected the conjugate into the cytoplasm of Vero cells. The peptide containing M1 caused rapid and complete nuclear accumulation, whereas that containing M2 caused slower and only partial nuclear localization. Thus, M1 functions as the nuclear localization signal of c-myc, and M2 serves some other and essential function.

The human L-myc gene encodes multiple nuclear phosphoproteins from alternatively processed mRNAs

The human proto-oncogene L-myc generates at least four different mRNAs by alternative RNA processing. We have identified two phosphorylated L-myc proteins with molecular masses of 60,000 and 66,000 daltons [p60L-myc(human) and p66L-myc(human)] in a small-cell carcinoma line expressing high levels of L-myc mRNA. These proteins have a short half-life and are localized to the nuclear matrix fraction, as previously reported for the c-myc and N-myc proteins. In vitro translation experiments demonstrated that both the p60 and p66 species are encoded by a 3.9-kilobase (kb) mRNA which retains intron 1, while only the p60 protein is translated from a 3.6-kb L-myc mRNA which has had intron 1 removed. While L-myc proteins [p32L-myc(human) and p37L-myc(human)] could be synthesized in vitro from 2.2-kb mRNA templates, no such proteins were detected by immunoprecipitation in vivo. These observations suggest that alternative RNA processing of the L-myc transcript could play a role in determining the steady-state levels of the p60L-myc and p66L-myc proteins.

Leader length and secondary structure modulate mRNA function under conditions of stress

Simian virus 40-based plasmids that direct the synthesis of preproinsulin in cultured monkey cells were used to study the effects of mRNA structure on translational efficiency. Lengthening the leader sequence enhanced translation in this system. The enhancement was most obvious when an unstructured sequence (two, four, or eight copies of the oligonucleotide AGCTAAGTAAGTAAGTA) was inserted upstream from a region of deliberate secondary structure; the degree of enhancement was proportional to the number of copies of the inserted oligonucleotide. Lengthening the leader sequence on the 3' side of a stem-and-loop structure, in contrast, did not offset the potentially inhibitory effect of the hairpin structure. Both the facilitating effect of length and the inhibitory effect of secondary structure were demonstrated most easily under conditions of mRNA competition, which was brought about by an abrupt shift in the tonicity of the culture medium. These experiments suggest a simple structural basis for the long-recognized differential response of viral and cellular mRNAs to hypertonic stress. The fact that the translatability of structure-prone mRNAs varies with changes in the environment may also have general implications for gene expression in eucaryotic cells.

Multiple mechanisms for transcriptional regulation of the myc gene family in small-cell lung cancer

The molecular mechanisms reported to regulate the expression of myc family genes are multiple and complex and include gene amplification, transcriptional activation, transcriptional attenuation, and mRNA stability. We have investigated which of these mechanisms are responsible for the extreme variation in myc gene family mRNA levels observed in human small-cell lung cancer cell lines. In addition to gene amplification, a block to nascent mRNA chain elongation, causing attenuation of transcription, is an important regulatory mechanism controlling the steady-state levels of c-myc and L-myc mRNA. The loss of transcriptional attenuation is correlated with overexpression of these two genes in cell lines which do not show gene amplification. Expression of c-myc mRNA appears to be dependent on promoter activity and attenuator function. In contrast, regulation of expression of the N-myc gene does not involve transcriptional attenuation; steady-state mRNA levels are correlated with promoter activity as well as gene amplification. We conclude that transcriptional regulation of each member of the myc gene family is accomplished by a different assortment of complex mechanisms, including gene copy number, promoter activation, and transcriptional attenuation. Interference at multiple points in this complex regulatory process appears to be an important mechanism by which small-cell lung cancer and other human tumors evade growth control.

Leader length and secondary structure modulate mRNA function under conditions of stress

Simian virus 40-based plasmids that direct the synthesis of preproinsulin in cultured monkey cells were used to study the effects of mRNA structure on translational efficiency. Lengthening the leader sequence enhanced translation in this system. The enhancement was most obvious when an unstructured sequence (two, four, or eight copies of the oligonucleotide AGCTAAGTAAGTAAGTA) was inserted upstream from a region of deliberate secondary structure; the degree of enhancement was proportional to the number of copies of the inserted oligonucleotide. Lengthening the leader sequence on the 3' side of a stem-and-loop structure, in contrast, did not offset the potentially inhibitory effect of the hairpin structure. Both the facilitating effect of length and the inhibitory effect of secondary structure were demonstrated most easily under conditions of mRNA competition, which was brought about by an abrupt shift in the tonicity of the culture medium. These experiments suggest a simple structural basis for the long-recognized differential response of viral and cellular mRNAs to hypertonic stress. The fact that the translatability of structure-prone mRNAs varies with changes in the environment may also have general implications for gene expression in eucaryotic cells.