Transformation-defective adenovirus 5 E1A mutants exhibit antioncogenic properties in human BLM melanoma cells

Adenoviral E1 A proteins exhibit a strong tumor-suppressive activity in human tumor cells. However, E1 A is capable of transforming rodent and human cells in cooperation with other oncoproteins, such as activated RAS. Thus, the therapeutic use of wild-type E1A harbors the principal risk of enhancing tumor malignancy. This prompted us to construct E1A 13S cDNA-derived mutants that were unable to transform baby mouse kidney cells in cooperation with E1B and to test their tumor-suppressive activity in BLM human melanoma cells. Anchorage-independent growth in soft agar was reduced for those cell lines expressing the E1AdelCR2 mutant, which lacks the entire conserved region 2 (CR2) sequences, or for cells expressing the E1AcR3Ex2 mutant, which contains CR3 plus exon 2 sequences. In contrast, cell lines expressing the entire E1A wild-type (E1AWT) or only the exon 2 sequences (E1AEx2) grew like the parental BLM cells. Moreover, inoculation of nude mice with BLM cells or cells expressing E1AEx2 revealed large tumors after 2 weeks. In contrast, tumors derived from E1AdelCR2- or E1ACR3Ex2-expressing cells exhibited a substantial delay in tumor growth accompanied by a loss of E1A expression in the outgrown tumors. Cell lines expressing E1AWT showed an intermediate phenotype. Thus, expression of CR3 plus exon 2 sequences is sufficient to enhance both the antioncogenic properties and the therapeutic safety of E1A in our system.

Expression of human p53 requires synergistic activation of transcription from the p53 promoter by AP-1, NF-kappaB and Myc/Max

Transcriptional control of p53 expression participates in the generation of appropriate levels of active p53 in response to mitogenic stimulation. This prompted us to study the role of a putative AP-1 and a NF-kappaB motif in the human p53 promoter for transcriptional regulation. We show that mutation of the AP-1 or the NF-kappaB motif abolishes transcription from the human p53 promoter in HeLa, HepG2 and adenovirus type 5 E1-transformed 293 cells. In comparison, mutation of the previously characterized Myc/Max/USF binding site in the human p53 promoter reduces the transcription rate fivefold. The AP-1 motif in the human p53 promoter binds c-Fos and c-Jun and the NF-kappaB motif binds p50(NF-kappaB) and p65RelA. The cooperative nature of transcriptional activation by these factors was documented by repression of c-fos or NF-kappaB1 translation: Pretreatment of the cells with a c-fos or p50(NF-kappaB1) antisense oligonucleotide suppresses transcription from the human p53 promoter completely. In addition, we show that (a) the level of endogenous p53 mRNA and (b) transcription from the strictly p53-dependent human mdm2 promoter are reduced in the presence of c-fos, c-jun, p50(NF-kappaB1), p65RelA or c-myc antisense oligonucleotides, underscoring the importance of these transcription factors for the expression of functional p53.

Two domains within the adenovirus type 12 E1A unique spacer have disparate effects on the interaction of E1A with P105-Rb and the transformation of primary mouse cells

Transformation of primary rodent cells by functions of the adenovirus type 12 (Ad12) early region 1 (E1) is reduced severalfold compared with transformation by E1 of Ad2. We analyzed whether the unique spacer region of Ad12 E1A that borders the conserved region (CR) 2 and represents an oncogenic determinant of Ad12 E1A is involved in this impaired transformation property, putatively by modulating transformation-relevant biological E1A functions. We show that a mutant (E1ASpm1) that lacks 12 amino-terminal residues of the spacer binds p105-Rb and p130 as Ad12 E1A wild type (E1Awt), whereas a second spacer mutant (E1ASpm2) that lacks an adjacent stretch of six alanines exhibits highly reduced binding to p105-Rb. The binding of this mutant to the p130 pocket protein is, however, little impaired. E1ASpm1 diminishes the formation of the p105-Rb-E2F complex more efficiently than E1Awt or, least efficient, E1ASpm2. These properties of the spacer mutants to target and to disintegrate the p105-Rb-E2F complex correspond with their ability to transform primary mouse cells in combination with E1B: E1ASpm1 (plus Ad12 E1B)-transfected cells could be easily established as cell lines, comparable to Ad12 E1Awt- or Ad2 E1Awt-transfected cells. In contrast, cells transfected with E1ASpm2 or Ad12 E1AdelCR2 (lacking the entire CR2) died within 6-10 weeks after replating, although foci were formed in all cases. Of note, the E1ASpm1-transformed cells grow as fast as the Ad2 E1Awt-transformed cells, with a doubling rate of 15 h, whereas the doubling of the Ad12 E1Awt-transformed cells takes approximately 120 h. Moreover, in the established cell lines, the affinity of E1ASpm1 to p105-Rb was higher than with that of E1Awt. Our data suggest the presence of a transformation-suppressing domain within the carboxyl-terminal 12 residues of the Ad12 E1A-unique spacer, whereas the hydrophobic stretch of six alanines in the spacer is required for stable transformation.

Assessment of resistance induction to cytosine arabinoside following transfer and overexpression of the deoxycytidylate deaminase gene in vitro

Recent attempts to protect hematopoietic progenitor cells from cytarabine (ara-C)-induced toxicity by transfer of the cytidine deaminase (CDD) gene resulted in efficient in vitro inducibility of ara-C resistance. Another enzyme involved in intracellular ara-CTP inactivation is the deoxycytidylate deaminase (dCMPD). We therefore transfected the human dCMPD cDNA gene into murine fibroblasts and investigated the relationship of forced dCMPD expression and resistance induction to ara-C. Several cell lines were established which demonstrated a 1.7-3.5-fold increase in cellular dCMPD activity and an up to 2-fold increase in the IC50 value of ara-C. However, increases in dCMPD activities did not show a positive linear correlation with the induction of ara-C resistance. In addition, CD34 + hematopoietic progenitor cells revealed the highest endogenous dCMPD enzyme levels among different human hematopoietic cells. Thus, despite the documented role for dCMPD in ara-CTP inactivation of certain cell types, these results suggest that the dCMPD gene may prove less useful than the CDD gene as a therapeutic target in attempts to attenuate ara-C-induced bone marrow toxicity.

Recombinant gene products of two natural variants of the human cytidine deaminase gene confer different deamination rates of cytarabine in vitro

The recent cloning of human cytidine deaminase (CDD) revealed two variants with a nonconservative amino acid deviation (Gln<-->Lys) at codon 27 within a region of structural homology to a core domain of bacterial CDDs. We here confirm the occurrence of both CDD sequences by cDNA cloning and show that at cytarabine (ara-C) concentrations of 1x10(-8) to 2x10(-2) M, the recombinant enzyme corresponding to the Lys-carrying natural variant (CDD-2) exerts a 1.3- to 3.3-fold higher in vitro deamination rate of ara-C than the Gln-carrying enzyme (CDD-1). These results suggest that this genetic polymorphism contributes to the different deamination phenotypes of ara-C observed in vivo, and that investigation of CDD allelotype frequencies and their correlation with ara-C resistance in patients with acute leukemia may be warranted. In addition, our data may be relevant to recently considered CDD gene transfer strategies for the detoxification of hematopoietic stem cells during high-dose therapy with cytosine nucleoside analogs.

Formation of the early-region-2 transcription-factor-1-retinoblastoma-protein (E2F-1-RB) transrepressor and release of the retinoblastoma protein from nuclear complexes containing cyclin A is induced by interferon alpha in U937V cells but not in interferon-alpha-resistant U937VR cells

We have analysed the different regulation of cell-cycle-relevant proteins by interferon alpha (IFN alpha) in IFN alpha-sensitive and resistant U937 leukemic cell lines. In contrast to the INF alpha-sensitive U937 variant cell line U937V, the IFN alpha-resistant derivative (U937VR) is insensitive to the antiproliferative activity of IFN alpha. As we found no differences between these cell lines concerning the induction by IFN alpha of the pathway involving tyrosine-protein kinases and the signal transducer and activator of transcription (Jak-Stat), we examined whether cell-cycle-regulating proteins are differently affected by IFN alpha in U937VR and U937VR cells. In U937V cells IFN alpha induced the formation of the complex between early-region-2 transcription factor 1 (E2F-1) and retinoblastoma protein (RB) which is known to repress transcription of E2F-1-inducible genes, necessary for cell cycle progression. Formation of this complex was not inducible by IFN alpha in U937VR cells, although the suitable binding partners (E2F-1 and under-phosphorylated RB) were present. Interestingly, treatment of nuclear extracts from logarithmically growing U937V and U937VR cells with an antiserum against cyclin A that disrupts cyclin-A-containing complexes, led to the formation of the E2F-1-RB complex, suggesting the presence of under-phosphorylated (active) RB, trapped in nuclear complexes that contain cyclin A. This suggestion was supported by combined immunoprecipitation/western blot experiments that revealed a physical interaction between phosphorylated as well as under-phosphorylated forms of RB and cyclin A complex(es) in U937V and U937VR cells. RB, especially the under-phosphorylated form, was released by treatment with IFN alpha from this complex(es) in the case of U937V cells but not U937VR cells. We conclude that the missing induction of the E2F-1-RB transrepressor by IFN alpha and the failure to release RB from cyclin-A-containing complexes might contribute to the resistance of U937VR cells to the antiproliferative effects of IFN alpha.

A cis-acting element 7 bp upstream of the ESF-1-binding motif is involved in E1A 13S autoregulation of the adenovirus 12 TS2 promoter

Transcription of the E1A gene of the highly oncogenic adenovirus 12 (Ad12) initiates at two start sites (TS1 and TS2). We have previously shown that the E2F and ATF motifs distal of TS1 co-operatively participate in E1A autostimulation from the TS1 promoter region. Here we report the identification of a second E2F-like target region (E2DFII) immediately upstream of the E1A-stimulating factor 1 binding site (ESF-1), important for 13S-mediated autoactivation from TS2. Reporter constructs lacking distinct TS2 cis-acting elements were analysed for their levels of CAT expression in the absence and presence of the E1A 13S protein in transient expression assays. In the absence of 13S, full promoter activity was observed only for a construct containing all elements (the E2F-like motif, and E-Box and the TATA element). Promoter activation increased significantly in Ad12 E1A-co-transfected cells. Induction by the 13S protein was also detected for the construct containing a non-functional ESF-1 sequence. Our results indicate that the E2F-like motif is responsible for activation medicated by the 13S protein from TS2, while ESF-1-or TATA-binding protein activity were not involved. Additionally, the TATA sequence appeared to be dispensable for transactivation. Gel-shift experiments using the E2F-like promoter element as a probe indicated the binding of an E2F-5 or E2F-5-like transcription factor to this region. We conclude that transcription through the TS1 as well as the TS2 promoter region is stimulated by the Ad 12 13S protein. Moreover, transfection of the construct including both TS1 and TS2 indicates an E2F-site-mediated synergism between both regions with respect ot E1A-induced transactivation.

Repression of c-Jun-induced mouse major histocompatibility class I promoter (H-2Kb) activity by the Adenovirus type 12-unique 52R E1A protein

Down-regulation of major histocompatibility (MHC) class I gene expression by protein products of the early region 1A (E1A), which might allow transformed cells to escape the host immune system, is discussed as one cause for the oncogenicity of Adenovirus (Ad) subtype 12-transformed cells. The MHC class I promoter is activated through several cellular-transcription factors among them AP-1, whose target sequences are located in the enhancers A and B, and NF kappa B. In this report we present evidence that the Ad12-unique 52R E1A protein inhibits c-Jun-induced activation of MHC class I gene expression. Repression occurs through both AP-1 recognition sequences with the AP-1 binding site of Enhancer A, which can be bound by c-Jun dimers in vitro, being the main target for c-Jun activation as well as 52R-mediated down-regulation. Furthermore our data revealed that both promoter elements of Enhancer A, the AP-1 and NF kappa B binding sites, are necessary for full promoter activity. As NF kappa B is down-regulated by the 266R protein of Ad12 E1A our results suggest a model in which two Ad12 E1A proteins co-operate in the repression of MHC class I gene expression.

Repression of the c-Jun trans-activation function by the adenovirus type 12 E1A 52R protein correlates with the inhibition of phosphorylation of the c-Jun activation domain

The early region 1A 52R polypeptide, a protein expressed exclusively by the in vivo oncogenic adenovirus subtype 12, represses the trans-activating function of the cellular transcription factor complex AP-1 consisting of c-Jun-c-Jun homodimers. In this report we demonstrate that the repression in vivo correlates with a direct physical interaction of the adenovirus protein with c-Jun in vitro. Interestingly, the 52R protein binds to the bZIP domain of c-Jun essential for dimerization and DNA binding but not to the c-Jun activation domain. This interaction does not prevent the promoter binding of c-Jun/AP-1. Moreover, the physical association between c-Jun and the TATA box-binding protein TBP is not disturbed by the 52R polypeptide. In fact, we show evidence that down-regulation of c-Jun activity by the adenoviral protein is due to the inhibition of phosphorylation of the c-Jun trans-activation domain. In vivo phosphorylation of the c-Jun activation domain is necessary for the interaction of c-Jun with specific cofactors such as CBP and therefore a prerequisite for the activation of target genes. Due to these results we propose a model in which the 52R protein represses the trans-activating function of c-Jun by preventing its phosphorylation through a specific kinase necessary for the activation of the cellular transcription factor.

Altered host range phenotype of the transformation-defective Ad12 mutant CS-1 is due to deletions in the E1 region

Although it grows well in bulk infection, human adenovirus type 12 (Ad12) does not plaque efficiently in Vero cells of simian origin. After long-term passage of the virus or after transfection of Ad12 DNA into these cells, however, transformation-defective, host-range mutants giving high plaque yields in Vero cells were isolated. The original mutants have deletions in both E1a and E1b as well as additions of viral sequences at the right terminus of the genome. We have constructed a recombinant virus (Ad12d169) carrying both E1 alterations of the original mutant CS-1 on the Ad12 wild-type background. Another mutant (Ad12mut2) has additional sequences at the right terminus and an intact E1 region. In plaque assays mutant virus Ad12d169 carrying E1a deletions has an about thousandfold higher efficiency in Vero cells than Ad12wt and Ad12mut2, mapping the enhanced replication in Vero cells to the deletions in E1. Mechanical models for the influence of the E1 deletions, for example by up-regulation of the E2-encoded DNA-binding protein, are discussed to explain the efficient replication of mutant adenoviruses in Vero cells under plaque assay conditions.

A new fast and easy method for the oligonucleotide affinity purification of DNA binding proteins

Images

Binding of chondroitin sulfate, dermatan sulfate and fat-storing cell-derived proteoglycans to rat hepatocytes

1. The interaction of isolated rat hepatocytes with exogenous 3H-labeled chondroitin-4-sulfate and dermatan sulfate and with biosynthetically 35S-labeled proteoglycans secreted by cultured rat liver fat-storing cells has been studied. 2. All ligands are bound by hepatocytes in a concentration-dependent manner. Scatchard-plot analysis of the data revealed the existence of high- and low-affinity binding modes. 3. The cell-bound exogenous [3H]glycosaminoglycans could be displaced by each unlabeled ligand and by heparin, whereas displacement of the endogenous material was less effective. 4. Binding of all ligands to hepatocytes increased with time. For the exogenous glycosaminoglycans the two- to threefold amount was retained at 37 degrees C as compared to 4 degrees C; it was markedly reduced by pretreatment of the cells with trypsin. 5. Degradation of the exogenous ligands could be detected neither for the cell-bound fraction nor for the free glycosaminoglycans in the culture medium. 6. The binding of the ligands to hepatocytes is viewed as a cell-matrix interaction. Its possible pathobiochemical relevance in liver fibrosis or neoplasia is discussed.