Adenovirus E1a proteins repress expression from polyomavirus early and late promoters

How the Rb tumor suppressor structure and function was revealed by the study of Adenovirus and SV40

The review recounts the history of how the study of the DNA tumor viruses including polyoma, SV40 and Adenovirus brought key insights into the structure and function of the Retinoblastoma protein (Rb). Knudsen's model of the two-hit hypothesis to explain patterns of hereditary and sporadic retinoblastoma provided the foundation for the tumor suppressor hypothesis that ultimately led to the cloning of the Rb gene. The discovery that SV40 and Adenovirus could cause tumors when inoculated into animals was startling not only because SV40 had contaminated the poliovirus vaccine and Adenovirus was a common cause of viral induced pneumonia but also because they provided an opportunity to study the genetics and biochemistry of cancer. Studies of mutant forms of these viruses led to the identification of the E1A and Large T antigen (LT) oncogenes and their small transforming elements including the Adenovirus Conserved Regions (CR), the SV40 J domain and the LxCxE motif. The immunoprecipitation studies that initially revealed the size and ultimately the identity of cellular proteins that could bind to these transforming elements were enabled by the widespread development of highly specific monoclonal antibodies against E1A and LT. The identification of Rb as an E1A and LT interacting protein quickly led to the cloning of p107, p130, p300, CBP, p400 and TRRAP and the concept that viral transformation was due, at least in part, to the perturbation of the function of normal cellular proteins. In addition, studies on the ability of E1A to transactivate the Adenovirus E2 promoter led to the cloning of the heterodimeric E2F and DP transcription factor and recognition that Rb repressed transcription of cellular genes required for cell cycle entry and progression. More recent studies have revealed how E1A and LT combine the activity of Rb and the other cellular associated proteins to perturb expression of many genes during viral infection and tumor formation.

E1A: tumor suppressor or oncogene? Preclinical and clinical investigations of E1A gene therapy

In the late 1980s, we have shown that the E1A gene can downregulate HER-2/neu overexpression, thus reversing the tumorigenic and metastatic phenotype. Further, E1A can function as a tumor suppressor gene by inducing apoptosis and inhibiting metastasis. At The University of Texas M. D. Anderson Cancer Center, we have been investigating the adenovirus type 5 E1A gene as a potential therapeutic gene in breast and ovarian cancer since 1995 by using cationic liposome as gene delivery system. In this chapter, we recount our development of E1A as a therapeutic gene.

Dominant negative mutants of Myc inhibit cooperation of both Myc and adenovirus serotype-5 E1a with Ras

We have used dominant negative Myc mutants to analyze the Myc and E1a mechanisms of cooperation with Ras. We show that mutants of Myc with an altered basic region (BR; RR366, 367EE) or deletion of the leucine zipper (LZ; delta aa 414-439), changes which modify the DNA binding domain, or with deletions in the Myc amino terminal conserved regions box 1 (dlMB1; delta aa 46-55) and box 2 (dlMB2; delta aa 132-140) inhibit cooperation of wt Myc and activated Ras to transform rat embryo fibroblasts (REF). Expression of the amino terminal 104 aa had no effect whereas wt Myc stimulated focus formation. Mutant dlMB1 cooperated with Ras with one half wt efficiency while dlMB2 was inactive. No mutant tested was toxic during neomycin cotransformation of REF to G418 resistance. Interestingly, these Myc mutants exerted a parallel inhibition of E1a-Ras cooperation to transform REF. This suggests that the Myc-Ras and E1a-Ras cooperation pathways intersect and require common protein factors. A Myc box 2 deletion mutant which is a wt transactivator of the Myc responsive ornithine decarboxylase promoter, but unlike the wt does not repress the adenovirus-2 core promoter (Li et al., 1994, EMBO J., 13:4070-4079), inhibits Myc-Ras and E1a-Ras cooperation. This suggests that a box 2-dependent step, potentially gene repression, is required for both the E1a- and Myc-Ras cooperation mechanisms.

Comparison between E1A gene from oncogenic and non-oncogenic adenoviruses in cellular transformation (Ad E1A conserved region)

All adenoviruses transform primary BRK cells in vitro, but only cells transformed by oncogenic adenoviruses are tumorigenic for immunocompetent animals. The transforming E1 regions of human Ad 2 and Ad 12 also differ from each other in the frequency in which they can transform BRK cells. We have investigated these properties which can be assigned to the specific domain of the E1A region. For this purpose, chimeric E1A regions between Ad 2 and Ad 12 have been constructed. The efficiency of cell transformation appeared to be determined by the encoding region. The promoter sequences were not important for an efficient cellular transformation although the E1B region cis activated in E1A transcription in both cell transformation and transient expression. We show that sequences located in the E1B promoter were responsible for this effect. In the encoding region the CR 1 domain was essential for the cell transformation frequency.

Phosphorylation of adenovirus E1A proteins by the p34cdc2 protein kinase

Adenovirus early region 1A (E1A) products are phosphorylated nuclear oncoproteins which appear to derive transforming activity largely through interactions with cellular proteins including the tumor suppressor p105/Rb-1 and cyclin A (p60cycA), a regulatory subunit associated with p34cdc2 and the related protein kinase p33cdk2. We have identified several sites of phosphorylation on E1A proteins previously and showed that phosphorylation at Ser-89 alters electrophoretic mobility significantly and affects E1A-mediated transforming activity to some extent. We now report that both Ser-89 and Ser-219, the major E1A phosphorylation site, were phosphorylated in vitro by p34cdc2 purified from HeLa cells. We also found that E1A proteins seemed to be phosphorylated at the highest levels in vivo in mitotic cells which express maximal levels of p34cdc2 kinase activity. Thus, in addition to forming complexes with p60cycA, a regulator of p34cdc2 and related kinases, and p105/Rb-1 which exhibits cell cycle-dependent phosphorylation, E1A proteins seem to be substrates for p34cdc2. These data suggested that a link could exist between phosphorylation, cell cycle progression, and the regulation of transforming activity of E1A proteins.

Cellular transformation by E1 genes of enteric adenoviruses

The ability of Ad40 and Ad41 E1A plus E1B genes to transform BRK cells was considerably lower than that of Ad5 and Ad12 corresponding genes. However, as for Ad5, the E1A genes of enteric adenoviruses could cooperate with an activated ras oncogene for full cell transformation and the Ad41 E1B could be complemented by E1A gene of Ad5 or Ad12 for cell transformation. Complementation studies suggested that the conserved region 1 of Ad41 E1A was responsible for this inefficient transformation. The Ad40- and Ad41-transformed cell lines exhibited a low level of major histocompatibility complex (MHC) class I antigens correlated to the low level of Ad12-transformed cells. Class I MHC antigen amounts expressed at the surface of the cells transformed by the weakly oncogenic Ad3 were between the high level of Ad5- and the low level of Ad12-transformed cells.

A novel function of the transforming domain of E1a: repression of AP-1 activity

Adenovirus E1a represses transcription of the collagenase gene via the phorbol ester-responsive element (collTRE). The mechanism involves inhibition of the trans-activating function of the transcription factor AP-1 without reduction of its synthesis and without any apparent change in DNA binding or composition. The ability of E1a to downmodulate AP-1 is a unique property among dominant oncogenes. This repression depends on conserved region 1, one of the transforming domains of E1a, indicating that it is an integral feature of adenovirus transformation.

Analysis of phosphorylation sites in the exon 1 region of E1A proteins of human adenovirus type 5

Early region 1A (E1A) of human adenovirus type 5 (Ad5) produces five mRNAs that encode proteins of 55, 171, 217, 243, and 289 residues. We have shown previously that the major products of 289 and 243 residues are phosphorylated at a minimum of three sites of which one, Ser-89, is located in the amino terminal half of the protein. In the present report we show that these E1A proteins are also phosphorylated at a second site in this region located at Ser-96. The 171 and 217 residue E1A species were also tentatively identified and, as predicted, neither contained the Ser-89 or Ser-96 sites but both appeared to be phosphorylated at the same sites as 289R and 243R toward the carboxy terminus. Studies with mutants in which Ser-89 or Ser-96 were converted to alanine residues indicated that phosphorylation of Ser-89 but not Ser-96 induces the major shift in gel mobility of E1A products. However, neither site appears to be of major importance in the regulation of E1A-mediated transactivation or transformation.

Adenovirus E1A products suppress myogenic differentiation and inhibit transcription from muscle-specific promoters

The primary function of the adenovirus E1A-region genes is to activate other adenoviral genes during a permissive viral infection by modifying the host cell transcriptional apparatus. Host cell immortalization, or transformation by the whole adenoviral early region, presumably results as a consequence of these modifications. Both transcriptional activation and transcriptional repression of non-adenoviral genes by the E1A proteins have been reported. It is currently not clear which, if either, of these activities contributes to host cell transformation and immortalization. Although there may be a physiological impact of some E1A-stimulated host cell genes, in many cases the functional significance is unclear. No common target sequences have been recognized in stimulated cellular genes and it has recently been proposed that in many cases, particularly involving newly transfected genes, available 'TATA-box' sequences may be the opportunistic beneficiaries of E1A assistance as a secondary consequence of E1A primary functions within the host cell nucleus. E1A-mediated transcriptional repression appears to be a more specific process insofar as common core elements are shared by the E1A-suppressed SV40, polyoma B, IgG heavy-chain and insulin enhancers. In the present communication we report that the complete myogenic programme of L8 and C2 myoblasts can be blocked by the introduction of constitutively expressing E1A genes, and show that the transcriptional induction of muscle-specific genes is inhibited. In particular, the promoter-inducing activities of well-defined elements that are required for the muscle-specific expression of the two sarcomeric alpha-actins, and which normally bind cellular trans-acting factors, become targets for E1A suppression. The results support the hypothesis that the suppression of differentiation by E1A products is effected by an E1A-mediated block in the transcriptional activation of cellular genes by specific developmentally regulated cis-acting promoter elements.

Use of deletion and point mutants spanning the coding region of the adenovirus 5 E1A gene to define a domain that is essential for transcriptional activation

To help in identifying functional domains within Ad5 E1A proteins, we have constructed a series of mutants that create deletions throughout these products. We have also produced several mis-sense point mutations in the unique 13 S mRNA region. These mutated E1A regions have been tested in plasmid form for their ability to activate transcription of an E3-promoted CAT gene. From the results, a major domain for transactivation has been identified. This begins between residues 138 and 147, ends between residues 188 and 204, and encompasses the unique 13 S region. This domain is sensitive to mis-sense mutations. Transactivation was unaffected by small deletions in the N-terminal half of E1A proteins between residues 4 and 138, but was destroyed when this whole region was deleted. The C-terminal 71 residues may affect transactivation, but the results with the mutant in which this region was deleted were variable. The results obtained with these mutants are discussed in relation to the transactivation obtained by J. W. Lillie et al. [(1987). Cell 50, 1091-1100] with a synthetic peptide similar to the domain described here.

Mutually exclusive interaction of the CCAAT-binding factor and of a displacement protein with overlapping sequences of a histone gene promoter

The sperm histone H2B-1 gene of the sea urchin Psammechinus miliaris contains two octamer sequences (ATTTGCAT) and two CCAAT motifs upstream of its TATA box. The CCAAT-binding factors present in nuclear extracts from testis and from blastula and gastrula embryos are indistinguishable by mobility shift and methylation interference analysis. However, there is a testis-specific octamer-binding factor in addition to the ubiquitous form. In DNAase I protection experiments, the CCAAT-binding factor of only the testis extract is able to interact with the sperm H2B promoter. In the two embryonic extracts a novel factor binds with high affinity to sequences overlapping the proximal CCAAT element, thus preventing the DNA interaction of the CCAAT-binding factor in the embryo where the sperm H2B gene is not expressed. This CCAAT displacement protein may therefore act as a repressor of sperm H2B gene transcription.