Apoptosis and interferons: role of interferon-stimulated genes as mediators of apoptosis.

IFNs are a family of cytokines with pleiotropic biological effects mediated by scores of responsive genes. IFNs were the first human proteins to be effective in cancer therapy and were among the first recombinant DNA products to be used clinically. Both quality and quantity of life has been improved in response to IFNs in various malignancies. Despite its beneficial effects, unraveling the mechanisms of the anti-tumor effects of IFN has proven to be a complex task. IFNs may mediate anti-tumor effects either indirectly by modulating immunomodulatory and anti-angiogenic responses or by directly affecting proliferation or cellular differentiation of tumor cells. Both direct or indirect effects of IFNs result from induction of a subset of genes, called IFN stimulated genes (ISGs). In addition to the ISGs implicated in anti-viral, anti-angiogenic, immunomodulatory and cell cycle inhibitory effects, oligonucleotide microarray studies have identified ISGs with apoptotic functions. These include TNF-alpha related apoptosis inducing ligand (TRAIL/Apo2L), Fas/FasL, XIAP associated factor-1 (XAF-1), caspase-4, caspase-8, dsRNA activated protein kinase (PKR), 2'5'A oligoadenylate synthetase (OAS), death activating protein kinases (DAP kinase), phospholipid scramblase, galectin 9, IFN regulatory factors (IRFs), promyelocytic leukemia gene (PML) and regulators of IFN induced death (RIDs). In vitro IFN-alpha, IFN-beta and IFN-gamma induced apoptosis in multiple cell lines of varied histologies. This review will emphasize possible mechanisms and the role of ISGs involved in mediating apoptotic function of IFNs.

Organizations and promoter analyses of the human and the mouse genes for PACT, the protein-activator of the interferon-induced protein kinase, PKR.

PACT is an activator of the protein kinase, PKR. Here we report the isolation and the characterization of the mouse Pact gene. It contains eight exons ranging in size from 79 to 630 bp spanning a region of 18 kb with the largest and smallest introns being 3700 and 500, respectively. The human PACT gene, as analyzed from sequence available in the GenBank database, has a very similar organization. The 5' flanking regions of both mouse and human PACT genes are devoid of TATA boxes but are rich in GC boxes. Although there are putative binding sites of numerous transcription factors on both promoters, their organizations and identities are different. For examining promoter activities, about 2 kb of DNA 5' to the transcription start sites of both genes was cloned upstream of a reporter luciferase gene. Transient transfection assays demonstrated that both promoters are strong. Deletion analyses revealed that most of the positive cis-elements lie within 400 bp upstream of the transcription start sites of both mouse and human PACT genes.

A specific isozyme of 2'-5' oligoadenylate synthetase is a dual function proapoptotic protein of the Bcl-2 family.

2-5(A) synthetases are a family of interferon-induced enzymes that polymerize ATP into 2'-5' linked oligoadenylates that activate RNase L and cause mRNA degradation. Because they all can synthesize 2-5(A), the reason for the existence of so many synthetase isozymes is unclear. Here we report that the 9-2 isozyme of 2-5(A) synthetase has an additional activity: it promotes apoptosis in mammalian cells. The proapoptotic activity of 9-2 was isozyme-specific and enzyme activity-independent. The 9-2-expressing cells exhibited many properties of cells undergoing apoptosis, such as DNA fragmentation, caspase activation, and poly ADP-ribose polymerase and lamin B cleavage. The isozyme-specific carboxyl-terminal tail of the 9-2 protein was shown, by molecular modeling, to contain a Bcl-2 homology 3 (BH3) domain, suggesting that it may be able to interact with members of the Bcl-2 family that contain BH1 and BH2 domains. Co-immunoprecipitate assays and confocal microscopy showed that 9-2 can indeed interact with the anti-apoptotic proteins Bcl-2 and Bclx(L) in vivo and in vitro. Mutations in the BH3 domain that eliminated the 9-2-Bcl-2 amd 9-2-Bclx(L) interactions also eliminated the apoptotic activity of 9-2. Thus, we have identified an interferon-induced dual function protein of the Bcl-2 family that can synthesize 2-5(A) and promote cellular apoptosis independently. Moreover, the cellular abundance of this protein is regulated by alternative splicing; the other isozymes encoded by the same gene are not proapoptotic.

A new pathway of translational regulation mediated by eukaryotic initiation factor 3.

We report a new pathway of translation regulation that may operate in interferon-treated or virus-infected mammalian cells. This pathway is activated by P56, a protein whose synthesis is strongly induced by interferons or double-stranded RNA. Using a yeast two-hybrid screen, we identified the P48 subunit of the mammalian translation initiation factor eIF-3 as a protein that interacts with P56. The P56-P48 interaction was confirmed in human cells by co-immunoprecipitation assays and confocal microscopy. Gel filtration assays revealed that P56 binds to the large eIF-3 complex that contains P48. Purified recombinant P56 inhibited in vitro translation of reporter mRNAs in a dose-dependent fashion, and that inhibition was reversed by the addition of purified eIF-3. In vivo, expression of transfected P56 or induction of the endogenous P56 by interferon caused an inhibition of overall cellular protein synthesis and the synthesis of a transfected reporter protein. As expected, a P56 mutant that does not interact with P48 and eIF-3 failed to inhibit protein synthesis in vitro and in vivo.

Different subcellular localizations for the related interferon-induced GTPases, MuGBP-1 and MuGBP-2: implications for different functions?

The guanylate-binding proteins (GBPs) are a family of 65-67-kDa proteins induced by both type I and type II interferons (IFN). Members of the GBP family of GTPases are among the most abundant IFN-gamma-induced proteins. GBPs contain an unusual GTP binding site, which is consistent with GBP hydrolysis of GTP to both GDP and GMP. In addition, six of the eight known GBPs have a carboxy-terminal CaaX motif for the addition of isoprenyl lipids. Despite their abundance, however, little is known about the biologic function or cellular location of GBPs. We report here on studies to localize both a newly identified murine GBP (MuGBP-2) and its closely related family member, MuGBP-1. In both IFN-treated macrophages and fibroblasts, MuGBP-2 is found in both a granular distribution throughout the cytoplasm and localized to vesicle populations of heterogeneous sizes. The localization of MuGBP-2 to vesicles is dependent on its isoprenylation. Despite a high degree of sequence identity and the presence of an identical CaaX sequence, MuGBP-1 has a very homogeneous cytoplasmic distribution and fails to localize to intracellular vesicles. The different intracellular distribution of these two closely related family members suggests differential function(s).

Novel functions of interferon-induced proteins.

Interferons are important cytokines which regulate antiviral, cell growth, immune modulatory and anti-tumor functions. These pleiotropic effects of interferons are brought about by a large number of cellular proteins, the interferon-inducible proteins. Investigation of the biochemical and cellular activities of some of these proteins have revealed new pathways of regulation of cellular RNA and protein metabolism, growth and differentiation, apoptosis and signal transduction. In this article we discuss recent findings on the novel activities of a selected number of interferon-induced proteins.

Induction of the human protein P56 by interferon, double-stranded RNA, or virus infection.

P56 is the most abundant protein induced by interferon (IFN) treatment of human cells. To facilitate studies on its induction pattern and cellular functions, we expressed recombinant P56 as a hexahistidine-tagged protein in Escherichia coli and purified it to apparent homogeneity using affinity chromatography. A polyclonal antibody raised against this recombinant protein was used to show that P56 is primarily a cytoplasmic protein. Cellular expression of P56 by transfection did not inhibit the replication of vesicular stomatitis virus and encephalomyocarditis virus. P56 synthesis was rapidly induced by IFN-beta, and the protein had a half-life of 6 h. IFN-gamma or poly(A)(+) could not induce the protein, but poly(I)-poly(C) or an 85-bp synthetic double-stranded RNA efficiently induced it. Similarly, infection of GRE cells, which are devoid of type I IFN genes, by vesicular stomatitis virus, encephalomyocarditis virus, or Sendai virus caused P56 induction. Surprisingly, Sendai virus could also induce P56 in the mutant cell line P2.1, which cannot respond to either IFN-alpha/beta or double-stranded RNA. Induction of P56 in the P2.1 cells and the parental U4C cells by virus infection was preceded by activation of IRF-3 as judged by its translocation to the nucleus from the cytoplasm.

Characterization of the interaction between the interferon-induced protein P56 and the Int6 protein encoded by a locus of insertion of the mouse mammary tumor virus.

For determining cellular functions of the interferon-inducible human cytoplasmic protein P56, we undertook a Saccharomyces cerevisiae two-hybrid screen that identified Int6 as a P56-interacting protein. That the interaction also occurs in human cells was confirmed by coimmunoprecipitation and the observed cytoplasmic displacement of nuclear Int6 upon coexpression of P56. Because Int6 has been claimed to be both a cytoplasmic and a nuclear protein, we investigated the structural basis of this discrepancy. By mutational analyses, we showed that the Int6 protein contains a bipartite nuclear localization signal and a nuclear export signal at the far end of the amino terminus. The 20 amino-terminal residues of Int6, when they were attached to a different nuclear protein, were sufficient to translocate that protein to the cytoplasm. Within this region, replacement of any of the three leucine residues with alanine destroyed the function of the export signal. The specific domain of P56 that is required for its interaction with Int6 was mapped using the yeast two-hybrid assay and a mammalian coimmunoprecipitation assay. Both assays demonstrated that the C-terminal region of P56 containing three specific tetratricopeptide motifs is required for this interaction. In contrast, removal of an internal domain of P56 enhanced the interaction, as quantified by the two-hybrid assay.

Transcriptional induction of the p69 isoform of 2',5'-oligoadenylate synthetase by interferon-beta and interferon-gamma involves three regulatory elements and interferon-stimulated gene factor 3.

The 2',5'-oligoadenylate synthetases are key enzymes that mediate antiviral actions of interferon (IFN). The mRNAs for the intermediate isoforms (p69) of human 2',5'-oligoadenylate synthetase are rapidly induced 10- to 20-fold in HT1080 glioma cells by IFN-beta and induced 3-fold at 24 h by IFN-gamma. Induction is mediated by three regulatory elements, an IFN-stimulated response element and two identical sites resembling interferon response factor binding sites that are located within 300 bp of the transcriptional start site. Maximal induction requires all three elements, yet mutation in the most distal IRF-1-like site diminishes transcription only slightly. Mutation in the ISRE substantially decreases constitutive expression but does not abrogate the response to IFNs. Simultaneous mutation in all three elements abolishes responsiveness to both IFN-beta and IFN-gamma. Both constitutive and IFN-beta-induced expression from the p69 promoter is blocked in mutant cell lines deficient in components of the transcription factor, interferon-stimulated gene factor 3, suggesting that it is the primary factor controlling IFN-beta induced expression of this gene.

Production, purification, and characterization of recombinant 2', 5'-oligoadenylate synthetases.

2',5'-Oligoadenylate [2-5(A)] synthetases are a family of interferon-induced enzymes that polymerize ATP into 2'-5'-linked oligoadenylates in the presence of double-stranded RNA (dsRNA), their cofactor. The 2-5(A) molecules, in turn, activate the latent ribonuclease RNase L by promoting its dimerization. The 2-5(A) synthetase pathway has been implicated in interferon's antiviral and anticellular activities. In addition to their interesting cellular properties, these enzymes are also enzymologically interesting because they are the only known template and primer independent nucleotide (DNA or RNA)polymerases that synthesize 2'-5'-linked oligonucleotides. Moreover, their mode of activation by dsRNA remains unknown. In the past, biochemical and structure-function studies have been hampered by the lack of a convenient system for expressing recombinant 2-5(A) synthetases. These proteins are toxic to mammalian cells, probably because of RNase L activation, and proteins produced in bacteria do not have full enzymatic activity. To circumvent these problems, we have developed a baculovirus-insect cell system for high-yield expression of the small and medium isozymes. Here, methods are described for the production, purification, and characterization of the mouse small (9-2) (S. K. Ghosh, J. Kusari, S. K. Bandyopadhyay, H. Samanta, R. Kumar, and G. C. Sen, 1991, J. Biol. Chem. 266, 15293-15299) and human medium (P69) (I. Marie and A. G. Hovanessian, 1992, J. Biol. Chem. 267, 9933-9939) 2-5(A) synthetase isozymes and their mutants using the insect cell system. We also report methods for studying 2-5(A) synthetase-dsRNA interactions and protein-protein interactions among the subunits of the two isozymes.

Restoration of interferon responses of adenovirus E1A-expressing HT1080 cell lines by overexpression of p48 protein.

We have previously shown that both alpha interferon (IFN-alpha) and IFN-gamma signaling pathways are blocked in HeLa cells expressing the adenovirus E1A proteins (G. T. Leonard and G. C. Sen, Virology 224:25-33, 1996). Here, we report that in two other E1A-expressing cell lines derived from the HT1080 cells, neither IFN-alpha nor IFN-gamma could induce the transcription of genes containing the IFN-stimulated response element (ISRE). In contrast, IFN-gamma-mediated signaling to the gamma-activated sequence was unimpaired in these cells. This dichotomy was due to a lowered level of functional p48 protein but not of STAT1 protein in the E1A-expressing HT1080 cells. When p48 was overexpressed in those cells by stably transfecting a p48 expression vector, both types of IFN could effectively induce the transcription of ISRE-driven genes. Consequently, IFN-alpha was highly effective in inhibiting the replication of encephelomyocarditis virus in the E1A-expressing cells, which also overexpressed p48. These results reinforce the general conclusion that adenovirus E1A proteins block IFN signaling pathways by lowering the functional levels of one or more components of the trans-acting complexes that activate the transcription of IFN-stimulated genes.

Effects of mutating specific residues present near the amino terminus of 2'-5'-oligoadenylate synthetase.

In this study, we investigated the role of specific amino acid residues present near the amino terminus of the 9-2 isozyme of 2'-5'-oligoadenylate synthetase. In vitro expression of deletion mutants showed that residues 1-9 are required for enzyme activity. Within this region, residues 3, 7, and 8 were found to be conserved among all known isozymes of 2'-5'-oligoadenylate synthetase. Mutation of these residues singly or in combination resulted in partial or total loss of enzyme activity. Substitution of the proline residue at position 7 by different residues caused a partial or complete loss of activity. The properties of the inactive P7Q mutant were further explored by expressing the protein in bacteria. The bacterially expressed protein was also enzymatically inactive. The mutant protein could bind the substrate ATP and the activator double-stranded RNA normally. Oligomerization properties of the protein were examined by an affinity-based interaction assay and by glycerol gradient centrifugation; there was no detectable difference between the wild type and the P7Q mutant. These results demonstrated the importance of the proline residue at position 7 in conferring enzyme activity to the protein without affecting its other properties.

Enzymatic activity of 2'-5'-oligoadenylate synthetase is impaired by specific mutations that affect oligomerization of the protein.

Previous studies from our laboratory have shown that deletion of residues 321 to 344 of the 9-2 isozyme of 2'-5'-oligoadenylate (2-5(A)) synthetase causes a loss of its enzyme activity (Ghosh, S. K., Kusari, J., Bandyopadhyay, S. K., Samanta, H., Kumar, R., and Sen, G. C. (1991) J. Biol. Chem. 266, 15293-15299). Sequence comparison of this region among the different isozymes of 2-5(A) synthetases revealed that the residues at positions 330 to 333 are highly conserved. Alanine-scanning mutagenesis of these residues demonstrated that the residues present at 331, 332, and 333 are important for activity but the proline at position 330 was dispensable. The triple mutant containing Ala residues at 331, 332, and 333 was completely inactive. Different double mutants were slightly active, and the three single mutants were partially active. The triple mutant was further characterized for delineating the nature of its defect. The mutant protein was enzymatically inactive irrespective of whether it was synthesized in rabbit reticulocyte lysate, Escherichia coli or Trichoplusia ni insect cells. It could bind double-stranded RNA and ATP as efficiently as the wild type protein. It was, however, defective in oligomerization. Gel filtration and sedimentation velocity analyses of in vitro synthesized proteins revealed that the wild type protein, but not the triple mutant, formed tetramers. The tetrameric fraction, but not the monomeric fraction of the wild type protein was enzymatically active. The failure of the triple mutant to participate in homomeric protein-protein interaction was confirmed by in vivo assays in insect cells. These results indicate that tetramerization of the protein is required for the enzymatic activity of the small 2-5(A) synthetases.

Effects of adenovirus E1A protein on interferon-signaling.

We have previously shown that adenovirus E1A proteins can block interferon-alpha (IFN-alpha)-signalling. In the current study, we examined if the same is true for IFN-gamma signaling. Cotransfection experiments showed that both 289R and 243R forms of E1A could block the expression of an IFN-gamma-inducible reporter gene. Similarly, in an E1A-expressing HeLa cell line IFN-gamma failed to induce the synthesis of IRF-1 mRNA. This failure was due to a block in activation of the crucial trans-acting factor, GAF, which in turn was due to the lack of IFN-gamma-activated tyrosine phosphorylation of the STAT1 alpha protein in E1A-expressing cells. The above defect could be attributed to a reduced level of STAT1 alpha protein. The level of p48 protein, which is required for IFN-alpha signaling, was also lowered. However, the level of lak-1 protein, one of the tyrosine kinases necessary for both IFN-alpha and IFN-gamma signalling, was comparable in the E1A-expressing and the control cells. These results indicate that the observed inhibition of IFN signalling in E1A-expressing cells is a consequence of a lower abundance of the necessary trnas-acting factors.

Analysis of the negative transcriptional regulatory element in the angiotensin-converting enzyme gene.

We have characterized the sequence requirements and the protein binding properties of the previously identified transcriptional negative element present in the rabbit angiotensin-converting enzyme (ACE) gene. DNase footprinting experiments revealed that within the negative element (-715 to -610) several regions interact with proteins present in the nuclear extracts of ACE-expressing and -nonexpressing cell lines. Transfection analysis using the heterologous beta-actin promoter and mutated negative elements demonstrated that the SP1 site, the collagen-silencer-like sequence, and the inverted repeat elements are dispensable for their functioning. Deletion of the region between -692 to -668, however, completely eliminated the activity of the negative element, and mutation of the synapsin-silencer-like sequence present within this region vastly reduced it. This region (-692 to -668) by itself, when present in two copies, could effectively repress the activity of the beta-actin promoter. The same point mutations in the silencer element that destroyed its action on the beta-actin promoter greatly increased the transcriptional efficiency of the native ACE promoter. Electrophoretic mobility shift assay using the -692 to -668 ACE silencer sequence demonstrated the formation of a DNA/protein complex. UV cross-linking of the components of this complex revealed the presence of one prominent protein of approximately 21.5 kDa. This protein may be responsible for mediating the transcriptional-repressing activity of the ACE negative element. Homology between the ACE silencer and neuronal silencer consensus sequence, together with the promoter- and tissue-independent function of the the ACE silencer, suggests this element may bind a member of a large family of common negative regulatory transcription factors.

The cyclic AMP response elements of the genes for angiotensin converting enzyme and phosphoenolpyruvate carboxykinase (GTP) can mediate transcriptional activation by CREM tau and CREM alpha.

The potential of the CREM family of proteins to activate transcription of the genes encoding the testis-specific isozyme of angiotensin converting enzyme (ACET) and the gluconeogenic enzyme, phosphoenolpyruvate carboxykinase (GTP) (PEPCK) (EC 4.1.1.32) were investigated. Both CREM tau and CREM alpha bind efficiently to the putative cyclic AMP response element (CRE) present in the ACET gene (CRET) and to the CRE in the PEPCK gene. In HepG2 cells, the CRE was required for the strong stimulation by CREM tau of the expression of a chimeric PEPCK (-210 to +73)-chloramphenicol acetyl transferase (CAT) gene. The CRE could be mutated to the CRET sequence without losing the stimulatory effects of CREM tau. However, a similar chimeric gene driven by the regulatory region of the ACET gene, which contains the CRET site, could only be stimulated by CREM tau when its imperfect TATA element was mutated to an authentic TATA. Surprisingly, CREM alpha, an alleged inhibitor of CRE-mediated transcription, stimulated the expression of both PEPCK-CAT and ACET-CAT genes in HepG2 cells, a process which required the presence of the CRE and the CRET sites, respectively. In contrast, when the same CRE elements were used to drive the transcription of a chimeric gene containing the thymidine kinase promoter linked to the CAT structural gene, CREM alpha inhibited its expression in HepG2 and JEG3 cells. The expression of the same chimeric gene, however, was stimulated by CREM alpha in F9 embryonal carcinoma cells. These results demonstrated that the nature of the transcriptional effects of CREM isoforms on CRE-mediated transcription depends on the specific gene, the specific cell type and the promoter context of the CRE site.

Transcriptional induction by double-stranded RNA is mediated by interferon-stimulated response elements without activation of interferon-stimulated gene factor 3.

Many genes induced by type I interferons (IFNs) are also induced by double-stranded (ds)RAN. In this study, we investigated the mechanism of this induction process. Using cell lines from which the type I IFN genes have been deleted, we established that induction by dsRNA of the IFN-inducible 561 gene is direct and not mediated by the intermediate synthesis of IFN. Unlike 561 mRNA, the IFN-inducible 6-16 mRNA was induced poorly by dsRNA. Transfection studies demonstrated that the sequence difference between the core IFN-stimulated response elements (ISREs) of these two genes is not responsible for their differential inducibility by dsRNA. A point mutation in the 561 ISRE that abolished its response to IFN-alpha also made it unresponsive to dsRNA, thus demonstrating that the ISRE is the relevant cis-acting element for dsRNA signaling. The roles of different known ISRE-binding protein and tyrosine kinases in transducing the signal elicited by dsRNA were evaluated in genetically altered cell lines. dsRNA failed to induce 561 mRNA in cells expressing an anti-sense RNA for interferon regulatory factor 1, whereas it was induced strongly in cells expressing the corresponding sense mRNA. 561 mRNA was also induced strongly by dsRNA, but not by IFN-alpha, in mutant cell lines that do not express functional tyrosine kinases Tyk2 or JAK1 or ISRE binding protein, p48, or STAT2, all of which are required for IFN-alpha signaling. However, in cells devoid of functional STAT1, which is also required for IFN-alpha signaling, the induction of 561 mRNA by dsRNA was very low. Expression of transfected STAT1 alpha protein, but not of STAT 1beta protein, in these cells greatly enhanced the dsRNA inducibility of the 561 gene. These studies indicated that the major ISRE-mediated signaling pathway used by dsRNA requires interferon regulatory factor 1 and STAT alpha. This pathway, however, does not require the other known cytoplasmic components used for IFN-alpha signaling.

The interferon-inducible double-stranded RNA-activated protein kinase self-associates in vitro and in vivo.

The interferon-inducible double-stranded (ds) RNA-activated protein kinase (PKR) exhibits antiviral, anticellular, and antitumor activities. The mechanisms of its enzymatic activation by autophosphorylation and of the observed transdominant inhibitory phenotype of enzymatically inactive mutants have invoked PKR dimerization. Here we present direct evidence in support of PKR-PKR interaction. We show that radiolabeled PKR can specifically interact with matrix-bound unlabeled PKR in the absence of dsRNA. The self-association activity resides, in part, in the N-terminal region of 170 residues, which also constitutes the dsRNA-binding domain (DRBD). DRBD can bind to matrix-bound PKR or to matrix-bound DRBD. Dimerization of DRBD was directly demonstrated by chemical crosslinking. Affinity chromatography and electrophoretic mobility supershift assays demonstrated that mutants that fail to bind dsRNA can still exhibit protein-protein interaction. The PKR-PKR interaction could also be observed in a two-hybrid transcriptional activation assay in mammalian cells and consequently is likely to be an important feature of PKR activity in vivo.

Activation of interferon-inducible 2'-5' oligoadenylate synthetase by adenoviral VAI RNA.

2'-5' oligoadenylate (2-5(A)) synthetase and protein kinase, RNA activated (PKR) are the only two known enzymes that bind double-stranded RNA (dsRNA) and get activated by it. We have previously identified their dsRNA binding domains, which do not have any sequence homology. Here, we report a profound difference between the two enzymes with respect to the structural features of the dsRNA that are required for their activation. The adenoviral virus-associated type I (VAI) RNA cannot activate PKR, although it binds to the protein and thereby prevents its activation by authentic dsRNA. In contrast, we observed that VAI RNA can both bind and activate 2-5(A) synthetase. Mutations in VAI RNA, which removed occasional mismatches present in its double-stranded stems, markedly enhanced its 2-5(A) synthetase-activating capacity. These mutants, however, are incapable of activating PKR. Other mutations, which disrupted the structure of the central stem-loop region of the VAI RNA, reduced its ability to activate 2-5(A) synthetase. These debilitated mutants could bind to the synthetase protein, although they fail to bind to PKR.

Transcriptional repression of heterologous constitutive and inducible promoters by the negative element of the rabbit angiotensin-converting enzyme gene.

We previously identified a transcriptional negative element (NE) present in the rabbit angiotensin-converting enzyme (ACE) gene. Here, we report that the NE can also repress transcription driven by the strong constitutive promoters of the human beta-actin gene and SV40 in both ACE-expressing and nonexpressing cell lines. The extent of repression was influenced by the relative positions of the NE and the SV40 promoter and enhancer. The NE could also repress transcription driven by interleukin-1 and interferon-alpha-inducible promoters. Finally, transcription from a TATA-less promoter was equally repressed by the NE. Taken together, these results suggest that the NE of the rabbit ACE gene can function as a universal transcriptional silencer element.