Proc. Nati. Acad. Sci. USA Vol. 84, pp. 8355-8359, December 1987 Biochemistry A cellular protein, activating transcription factor, activates transcription of multiple ElA-inducible adenovirus early promoters (cellular transcription factor/DNA-binding proteins) KEVIN A. W. LEE*, TSONG-YUING HAI*, LAKSHMI SIVARAMANt, BAYAR THIMMAPPAYAt, HELEN C. HURSTf, NICOLAS C. JONESf, AND MICHAEL R. GREEN* *Department of Biochemistry and Molecular Biology, Harvard University, 7 Divinity Avenue, Cambridge, MA 02138; tDepartment of Microbiology and Immunology and Cancer Center, Northwestern University Medical School, 303 East Chicago Avenue, Chicago, IL 60611; and tGene Regulation Group, Imperial Cancer Research Fund, London, England WC2A 3PX Communicated by James C. Wang, August 17, 1987 (receivedfor review June 6, 1987) ABSTRACT We have examined the relationship between promoters, there is currently no evidence that ElA acts by sequence-specific DNA-binding proteins that activate transcrip- increasing the DNA-binding activity of factors required for tion of ElA-inducible adenovirus early promoters. Factors pre- E4 and E3 transcription (18, 19). For the EIB promoter (22) viously referred to as E4F1 and E2A-EF bind to the E4 and E2A and the cellular /-globin promoter (23), the "TATA" box has promoters, respectively. We demonstrate here that E4F1 and been implicated as an ElA-responsive promoter element. E2A-EF have identical DNA-binding specificity. Moreover, E4F1 Thus activation of a variety of ElA-inducible promoters and E2A-EF both activate transcription of the E4 and E2A appears to involve different cellular factors and may occur promoters in vitro. These findings demonstrate that E4F1 and through divergent pathways, ultimately linked by ElA. E2A-EF are the same factor, which we have designated activating Two independent studies have identified additional factors transcription factor, or ATF. In addition to the E4 and E2A that interact with early viral promoters. A factor referred to promoters, ATF binds to an important functional element of the as E4F1 binds to the E4 promoter and also interacts with the ElA-inducible E3 promoter. Interaction of a common activator EIA, E2A, and E3 promoters (18). Similarly, a factor referred protein, ATF, with multiple ElA-inducible early viral promoters, to as E2A-EF binds to the E2A, EIA, E3, and E4 promoters suggests a significant role for ATF in ElA-mediated transcrip- (17). We show here that E4F1 and E2A-EF have the same tional activation. DNA-binding specificity and that both factors activate tran- scription of the E4 and E2A promoters in vitro. These results The EIA gene ofadenovirus produces closely related 13S and demonstrate that E4F1 and E2A-EF are the same factor, 12S mRNAs that encode nuclear-localized phosphoproteins which we refer to as ATF, for activating transcription factor. with diverse transcriptional regulatory properties (1-3). The In addition to the E4 and E2A promoters, ATF interacts with EMA 13S gene product coordinately activates a set of viral an important functional element of the adenovirus E3 pro- early genes (EIA, EIB, E2A, E3, and E4) during a productive moter. The interaction of ATF with multiple ElA-inducible infection of permissive human cells (4-7). The ElA 12S gene promoters suggests a significant role for ATF in E1A- product encodes a transcriptional repression function that mediated transcriptional activation. appears to act through transcriptional enhancer elements (8-10). In addition to regulating viral transcription, ElA MATERIALS AND METHODS activates or represses transcription of a limited number of Plasmids. pE4WT contains the adenovirus type 5 genome cellular genes (11, 12), and activates polymerase III-depen- between map units 100 and 89, including the entire E4 gene dent promoters (for review, see ref. 7). cloned into pBR322 between the EcoRI and Pvu II sites. Despite numerous studies, several factors have contributed ,BA128 contains the entire human ,8-globin gene including 128 to the difficulties in elucidating the mechanism(s) by which EMA base pairs (bp) of 5' flanking sequences cloned between the activates transcription of early viral genes. (l) Unlike many EcoRI and Pvu II sites of pBR322. pBR730 contains adeno- other transcriptional regulatory proteins, EMA is not a se- virus type 5 sequences between map units 59.5 and 75.9 quence-specific DNA-binding protein (13), implying that ElA (containing the E2A promoter) cloned between the EcoRI and acts indirectly via interaction with cellular transcription com- BamHI sites of pBR322. ponents. (ii) Extensive mutagenesis studies have failed to In Vtro Transcription. Nuclear extracts were prepared identify common promoter target sequences that signal induc- from HeLa cells as previously described (18, 24). Transcrip- tion by EMA (see ref. 7). (iii) The biochemical activities and tion reactions and competition assays were done as described cellular components (transcription factors) required for tran- (18). RNA was prepared and analyzed by primer extension as scriptional activation by EMA protein are largely unidentified. described (10). Several groups have recently identified cellular factors that DNase I "Footprinting" Experiments. The 3' 32P end-labeled interact with ElA-inducible viral promoters (14-19). In one probes were prepared by isolating a DNA fragment containing case, a protein factor that interacts with the E2A promoter E4 promoter sequences between -138 and +250 and labeling at (E2F) is markedly increased in DNA-binding activity (or position -138 with reverse transcriptase. Footprinting assays amount) by ElA (15, 20). This suggests that activation ofE2F were done as described (18). Competitor DNA and labeled is required for ElA-mediated activation ofthe E2A promoter. probe (-1 ng per reaction) were added simultaneously to the However, E2F protein does not interact with all early viral binding reactions. DNase I digestions were terminated by the promoters (21), and consequently, activation of E2F cannot addition of 3 vol of 2 x proteinase K buffer (25) containing an entirely account for coordinate activation ofearly viral genes. additional 10 mM EDTA followed by proteinase K treatment, Other observations support this view. For the E4 and E3 purification of labeled DNA, and fractionation on denaturing polyacrylamide gels as described (18). The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" Abbreviations: Me2SO4, dimethyl sulfate; E2F, E2 promoter binding in accordance with 18 U.S.C. §1734 solely to indicate this fact. factor; ATF, activating transcription factor; ... .F, factor. 8355 Downloaded by guest on October 1, 2021 8356 Biochemistry: Lee et A Proc. Natl. Acad. Sci. USA 84 (1987) Gel Retardation Assays. Probes used for gel retardation nuclear extract were done as for DNase I footprinting assays are described in the figure legends. Labeled probes were experiments. Me2SO4 was diluted 10-fold in water, and 1 ,l prepared by end labeling gel-purified DNA fiagments or syn- of diluted solution was added to a 20-,lI binding reaction. thetic double-stranded oligonucleotides. Gel retardation assays After incubation at room temperature for 4 min, reactions were done as described (17) using either crude nuclear extract were stopped by addition of 50 ,u of stop solution (26), 180 from uninfected HeLa cells or a nuclear protein fraction ,u ofwater, and 750 /14 ofethanol, precooled on dry ice. DNA obtained as follows. Crude nuclear extract was loaded onto a was precipitated by centrifugation for 15 min at 4°C in a phosphocellulose (P11) column at 40 mM KCI followed by step Microfuge; the pellet was resuspended in 200 ,u of 0.3 M elution with 250 mM and 600 mM KCI. The 600 mM fraction sodium acetate (pH 7) and ethanol-precipitated once more. (containing most of the binding activity) was dialyzed against The pellet was resuspended in 100 ,u of 1 M piperidine and buffer D (24) containing 100 mM KCI and then loaded onto a incubated at 90°C for 30 min. Samples were Iyophilized three DEAE-Sephacel column equilibrated in buffer D. The flow- times and then fractionated on denaturing polyacrylamide through fraction contained most of the ATF-binding activity, gels. and we refer to this fraction as DEAE-fractionated nuclear extract. RESULTS Dimethyl Sulfate (Me2SO4) Footprinting Experiments. Bind- A Related Protein Factor Binds to the Adenovirus E4, E3, ing reactions using 10 ,ug of protein from DEAE-fractionated and E2A Promoters. Using a DNase I footprinting assay we B EXTRACT - .9 1 2 3 4 5 6 7 8 'Is_ A -39 - ._o=Im 1 2 3 4 5 6 7 8 9 10 E4 promoter * _ t- COMPLEX-. t t3 - -56& U_ _ _ F REE i FREN DA_. ft #6 *0wh COMPETITOR - E4P E2AP E2LS B3P COMPETITOR E2LS E2WT - MOL. EXCESS - 10 20 40 10 20 40 20 40 50 MOL.EXCESS 30 90 180 30 90 180 - C 1 2 3 4 5 6 7 8 9 10 11 12 13 1415 16 17 wIu 6W..* - - I :s i COMPLEX - i.l-4 .. 9 0 FREE - DNA --w COMPETITOR - E4P BP - MOL. EXCESS - 20 50 20 50 - 20 - 50 FIG. 1. A related factor binds to the E4, E3, and E2A promoters. Labeled DNA probes and unlabeled competitor DNA fragments were as follows: E4P (containing an E4F1 binding site), 27-bp synthetic oligonucleotide (oligo) containing positions -37 to -63 ofthe E4 promoter; E2AP (containing an E2A-EF binding site), 27-bp oligo containing positions -63 to -85 of the E2A promoter; E3P1 (containing an E3F2 binding site), 24-bp oligo containing positions -44 to -67 of the E3 promoter; E3P2 (containing an E3F3 binding site), 22-bp oligo containing positions -82 to -103 of the E3 promoter; E3P3 (containing an E3F4 binding site), 24-bp oligo containing positions -157 to -180 of the EB promoter; E3N, "nonsense" 22-bp oligo (containing no factor-binding sites) containing positions -107 to -128 of the E3 promoter; E2WT, 81-bp DNA fragment containing positions -17 to -98 of the E2A promoter; E2LS, equivalent to E2WT but lacking positions -74 to -85 of the E2A promoter (17); ,BP, 24-bp oligo containing sequences from the human ,B-globin pre-mRNA (5' GCCCT CTA1TITCCCACCCTTAGG 3').