Epidermal Growth Factor (EGF) Receptor Gene Transcription

6 downloads 0 Views 7MB Size Report
We have studied in vitro transcription of the human epidermal growth factor (EGF) receptor proto-onco- gene using nuclear extracts of A431 human epidermoid.
Vol. 263, No.13,Issue of May 5, pp. 6329-6336, 1988 Printed in U.S.A.

THEJOURNALOF BIOLOGICAL CHEMISTRY

Epidermal Growth Factor (EGF) Receptor Gene Transcription REQUIREMENT FOR S p l AND AN EGFRECEPTOR-SPECIFIC

FACTOR* (Received for publication, December 10, 1987)

Ryoichiro Kageyama, Glenn T. Merlino, and Ira Pastan From the Laboratory of Molecular Biobgy, National Cancer Institute, National Institutesof Health, Bethesda, Maryland 20892

We have studiedin vitro transcription of the human chloramphenicol acetyltransferase (CAT) reporter gene indiepidermal growth factor (EGF) receptor proto-oncocated that the region between residue -151 and residue -20 gene using nuclearextracts of A431 human epidermoid (measure related to the start of translation) is necessary for carcinoma cells, which overproduce theEGF receptor. maximal promoter function (18). Exonuclease I11 protection With the in vitro system we found that Spl and other assays showed that there are at least eight specific A431 trans-acting factors bound to the EGF receptor pro- nuclear protein binding sites, including the sites containing moter regions and are required for maximal expres- GC boxes in the promoter region (18).In this study, we have sion.FractionationshowedthataDEAE-Sepharose developed an in vitro transcription system for the EGFrecepfraction (BA) contained a novelfactor, which specifi- tor gene using A431 nuclear extract to characterize the relacally stimulated EGF receptor transcription5- to 10- tionship between these nuclear factors and transcriptional fold. The molecular mass of the native formofthe factor is about 270-kDa based on its migration on activities and tounderstand the mechanism of transcriptional Sephacryl 5-300. This factor may activate transcrip- activation. The invitro transcription system (19,20) hasbeen tion of the proto-oncogene through a weak or indirectsuccessfully used to analyze and purify transcription factors APsuch as Spl (21), MLTF (22), CTF (23), TFIIB (24), and interaction with the DNA template. 1 (25, 26). Here we show that Spl and other DNA-binding proteins are required for the optimal expression of the EGF receptor, and we also demonstrate that a novel trans-acting The epidermal growth factor (EGF)’ receptor mediates the factor obtained from A431 nuclear extracts specifically stimmitogenic activity of EGF andplays an important role in the ulates EGF receptor transcription, suggesting that it might regulation of cellular growth (1).The receptor consists of be a key factor for the transcription activation of the cellular three domains: EGF-binding, transmembrane, and cyto- oncogene. plasmic tyrosine kinase domains (2-4). The transmembrane EXPERIMENTAL PROCEDURES and cytoplasmic kinase domains of the EGF receptor have extensive homology with the avian erythroblastosisvirus erbB Pbmids-Promoter deletion mutants were made as follows. For oncogene product (5, 6). In addition, the EGF receptor is pGER-6, -7, -9, -10, and -15, pERcat 5 (18)was cut with PuuII, TaqI, overproduced in many tumors, including A431 human epider- AuaII, BglI, and AuaI, respectively, and treated with Klenow fragment moid carcinoma cells (7-9), squamous carcinoma cells (10- or T4 DNA polymerase. These blunt-endedDNAs weredigested with 12), and some glioblastomas (13, 14). These findings suggest HindIII, and thefragments of the EGF receptor promoter region were These fragments and the 250-bp HindIII-EcoRI fragment that the constitutive activation of the mitogenic pathway of excised. excised from pSV2cat were ligated into the H i d 1 and EcoRI sites growth factors and receptors is involved with oncogenic prop- of pGEM-4 (Promega Biotec). pGER-16, -17, -18, and -19 were made erties. Recently, direct evidence that the EGF receptor can from pGER-15 using Bal31. The boundaries of the promoter confunction as an oncogene product was obtained by infecting tained in these plasmids were determined by dideoxy sequencing. The cells with a retrovirus containingthe EGFreceptor and show- control SV40 template was constructed by ligating the 325-bp PuuIIing the production of EGF-dependent tumors (15). Although HindIII fragment and the 250-bp HindIII-EcoRI fragment excised pSV2cat (27) into theHindII and EcoRI sites of pGEM-4. the overproduction of the EGF receptor protein generally from Preparation and Fractionation of A431 Nuclear Extract-The correlates with the level of transcription, it does not always buffer we used was HM, which consists of20mM Hepes, pH 7.9, 1 accompany gene amplification (13, 16). Therefore,it is likely mM MgCl,, 2 mM dithiothreitol, 17% glycerol, and KC1 as indicated. that trans-acting factors cause transcriptional activation of During the fractionation, we also added 0.5 mM phenylmethylsulfonyl fluoride. A431 nuclear extract was prepared according to Wildeman the EGF receptor gene. The promoter of the EGF receptor gene lacks a character- et al. (28). At the final step, the extract was dialyzed against 0.1 M about 5 h. We usually obtained 100-150 mg of nuclear extract istic TATA box and CAAT box, but contains multiple GC HM from 30 g of A431 cells. boxes and multiple transcription initiation sites (17). TranNext the extracts were subjected to several columns, as described sient transfection analysis of this promoter linked to the by Dynan and Tjian(29), with the following modifications. First, the extracts (250 mg) were applied to a heparin-agarose column (30 ml) equilibrated in 0.1 M HM, and the column was washed with 0.1 M the payment of page charges. This article must therefore be hereby HM (fraction A), 0.4 M HM (fraction B), and1M HM (fraction C). marked “aduertisement” in accordance with 18 U.S.C. Section 1734 The 0.1 M fraction was frozen and stored at -80 “C. This fraction solely to indicate this fact. was used in the transcription assay because it stimulated the reaction. has been submitted The 0.4 and 1 M fractions were dialyzed against 0.1 M HM for 3 h. The nucleotide sequence(s)reported in this paper to the GenBankTM/EMBL Data Bank with accessionnumber(s) The 0.4 M fraction (75 mg) was next applied to a DEAE-Sepharose 503206. CL-GB column (30 ml) equilibrated in 0.12 M HM, and the column The abbreviations used are: EGF, epidermal growth factor; bp, was washed with 0.12 M HM (fraction BA), 0.25 M HM (fraction BB), base pair; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonicacid; and 0.5 M HM (fraction BC). The protein-containing fractions from SDS,sodium dodecyl sulfate, PIPES, 1,4-piperazinediethanesulfonic each eluate were pooled, adjusted to 0.33 mg/ml ammonium sulfate, acid CAT, chloramphenicol acetyltransferase. and incubated at 4 “C for 1 h. After centrifugation each precipitate

* The costs of publication of this article were defrayed in part by

6329

6330

EGF Receptor Transcription Factors

was suspended in 500-600 pl of 0.12 M HM and dialyzed against the same buffer. From 250 mg of nuclear extract the final yields of the fractions BA, BB, and BC were usually 4, 12, and 3 mg, respectively. The sample from the 0.12 M eluate was then subjected to Sephacryl S-300 column or oligonucleotide affinity chromatography. The S-300 column (40 ml) was equilibrated in 0.12 M HM and run a t 18 ml/h. Each fraction(300 pl) from the S-300 column was concentrated about 2.5-fold by using Centricon 30 (Amicon) and assayed for the transcriptional and DNA binding activities. The purified S p l was kindly provided by Drs. J. T. Kadonaga and R. Tjian (University of California a t Berkeley). Oligonucleotide Affinity Chromatography-The oligonucleotide affinity column was prepared basically according to Kadonaga and Tjian (30). The hybridized oligonucleotides (Fig. 6) were coupled to CNBr-activated Sepharose 4B (Pharmacia LKB Biotechnology Inc.). The fraction BA (2 mg) was applied to the column (0.6 ml) equilibrated in 0.12 M HM, and thecolumn was washed with 0.12 M HM (fraction BAl), and 0.5 M HM (fraction BA2). The 0.5 M fraction was dialyzed against 0.12 M HM for 3 h at 4 ‘C, and both fractions were frozen and stored a t -80 “C. For further purification of BA2, the 0.5 M fraction was diluted to 0.1 M KC1 with HM buffer, reapplied, washed with 0.25 M HM, and eluted. The eluates were analyzed by 10% SDS-polyacrylamide gel. In Vitro Transcription and Competition Analysis-The final concentrations of the transcription reactions (25-100 pl) were 3 mM Hepes, pH 7.9, 80 mM KCl, 1.2 mM M&lz, 1.8 mM spermidine, 1.5 mM dithiothreitol, 12%glycerol, 500 p~ each of ribosomal nucleotide triphosphate, and 20 or 40 pg/ml DNA templates. We used one-tenth of the molar ratio of SV40 early promoter as a control. Under these conditions, we did not observe any competition between the EGF receptor and SV40 promoters. The transcription reactions contained 7.5 pl (75 pg) of nuclear extract in a final volume of 25 pl, 15 pl (30 pg) of heparin-agarose fraction A, 50 p1 (150 pg) of B, and 7 pl (5 pg) of C in a final volume of 100 pl, or 10 p1 (20 pg) of fraction A,25 pl (150 pg) of DEAE fraction BA, 20 pl (300 pg) of BB, and 17 pl (100 pg) of BC in a final volume of 100 pl. For the assay of S-300 fractions, we used 5 pl of fraction A, 20 pl of BB, 15 pl of BC, and 43 pl of each fraction. For the assay of the purified factor of fraction BA2, we used 4 or 6 pl. After incubation for 1 h a t 30 “C, the reaction was stopped by the addition of 4 volumes of 0.3 M sodium acetate, 0.4% SDS, and 1mM EDTA, extracted with phenol/chloroform and with chloroform, and precipitated with 2.5 volumes of ethanol. The pellet was dissolved in 200 pl of 2 M LiCl and precipitated again with 400 pl of ethanol by incubating on ice for 30 min. The pellet was suspended in 100 p1 of 10 mM Tris, pH 7.5, 100 mM NaCl, and 10 mM MgCl,, and treated with 13 pg of RNase-free DNase I (Bethesda Research Laboratories) for 10 min at 37 “C. The reaction was stopped by the addition of 100 pl of 10 mM EDTA, 0.2% SDS, and 150 mM NaCl. After phenol/ chloroform extraction and ethanol precipitation, the pellet was dissolved in 10 pl of 40% formamide, 0.4 M NaCl, 40 mM PIPES, pH 6.4, and 1 mM EDTA, containing 0.1 pmol of 5’-end labeled CAT primer. The CAT primer is a synthetic single-stranded 24-mer which hybridizes to the region between residues 4920 and 4943 of pSV2cat (27). After hybridization for 3 h a t 42 “C, the reverse transcriptase reaction was carried out according to Maniatis et al. (31). The primerextended product was electrophoresed on 5% polyacrylamide, 7 M urea gel. For competition analysis, we used a 143-bp AuaII fragment excised from pGER-7, an 89-bp HindIII-AuaI fragment, and a 66-bp HindIII-AuaI fragment excised from pGER-10 as competitors. The amounts of the comDetitors in the reaction were indicated in the legend of Fig. 3. DNase I Footnrintiw-The Hind11 framentcontaining the EGF receptor promoter r e g i k was excised from pGER-9 to make DNA probes. For the noncoding strand probe, the Hind11 fragment was treated with calf intestinal alkaline phosphatase,followed by phenol/ chloroform extraction and ethanol precipitation. The fragment was 5’-end-labeled with T4 polynucleotide kinase and [-p3’P]ATP. For the coding strand probe, the HindIII fragment was treated with Klenow fragment, [a-32P]dCTP, and [cx-~*P]~GTP in the presence of cold0.3 mM dATP. Both of the labeled Hind111 fragments were digested with BglI, and the135-bp fragment was purified from acrylamide gel. For the SV40 probe, the HindIII fragment containing the promoter regionwas excised from the SV40 template DNA, 5’-end-labeled, digested with PstI, and purified from polyacrylamide gel. DNase I footprinting reactions were carried out according to Dynan and Tjian (32). The DNA binding reactions contained 5 ng of the end-labeled

probe, 1 pg of sonicated calf thymus DNA, and protein sample in a final volume of 50 pl. Calf thymus DNA was omitted for the assay of purified factors. RESULTS

Deletion Analysis of the EGF Receptor Promoter in VitroNuclear extracts were prepared from A431 epidermoid carcinoma cells and tested for their ability to synthesize RNA from an EGF receptor promoter in uitro. Several methods were used to study in uitro transcription; the clearest results were obtained using a primer extension assay. DNA fragments containing the EGF receptor promoter were ligated to the CAT reporter gene (Fig. 1).The CAT gene was truncated in order to reduce the background (see “Experimental Procedures”). A CAT-specificprimer was usedto detect transcripts initiating within the EGFreceptor promoter. As a control, we used a plasmid containing the SV40 promoter ligated to the CAT gene so that we could simultaneously measure its transcript with the same CAT-specific primer. As shown in Fig. 2, we detected one prominent transcript

a pGEM4

promoter

-771

pGER-6 -569

pGER-7 -384 -151

-

-105 -96

-

-81 -65 -54

pGER-9 pGER-10 pGER-15 pGER-16

-

pGER-17

-

pGER-19

pGER-18

+

b

-300 CTCCTCCTCCCGCCCTGCCTCCCGCGCCTCGGCCCGCGCGAGCTAGACGT

-200 G T G T G A G C G C C C G C C C G C C G A G G C G G C C G G A G T C C C G A G C T A G C C C C G G

L

150

10

G G C G C C G C C G C C C A G A C C G G A C G A C A G G C C A C C T C G T C G C G T C C G CCGA

kl-

-100 G L ~ ~ ~ X C G C C A A ~ ~ C A A C C A CC ACC GGG C C C C C T G A C T

+

“L 18

4- 19

~ 5 0 CCGTCCAGTATTGATCGGGAGAGCCGGAGCGAGCTCTTCGGGGAGCAGCG

FIG. 1. Structure of various deletion mutants and the sequence of the EGF receptor promoter. a, the structures of pGER plasmids are shown. The EGF receptor (EGFR) promoter region (solid box) and the CAT gene (stippbd box) are cloned into pGEM-4. The start site(-48) and direction of the transcription are indicated by the upper arrow. The asterisk shows the 5’-end-labeled site of the CAT primer (lower arrow). b, the nucleotide sequence of the EGF receptor promoter is shown from -300 to -1. The regions protected by DNase I footprinting are underlined (see Fig. 5). S p l recognition sites are shown by an open box (18).The thick (-48) and thin (-255) arrows above the sequence indicate the major and minor in uitro transcriptioninitiation sites, respectively. The arrows under the sequence show the end pointsof the deletion mutants determined by DNA sequencing.

EGF Receptor Transcription Factors

6331 TABLE I

123 4 5 6 789

In vitro and in vivo transcription activity of the deletion mutants of the EGF receptor promoter The in vitro transcriptional activity relative to pGER-6 was measured by densitometric scanning of the autoradiograph shown in Fig. 2. The in vivo activity except for pGER-18 was determined by Johnson et al. (IS), using a transient transfection assay of monkey CV-1 cells. ND, not determined. Terndate

PGER-6 (-771) PGER-9 (-384) PGER-10 (-151) 59 (-105) pGER-15 pGER-16ND (-96) pGER-17 (-81) pGER-18 (-65) PGER-19 (-54)

*e--

- EGFR

a

%

%

100 100 100 60 30 10 0 0

100 105 105

b

1234

567

-

(-527) All

(-384) All

a

ND 0 ND

C

FIG. 2. In vitro transcription of the deletion mutants of EGF receptor (EGFR)promoter. In vitro transcription activity was analyzed by using an A431 nuclear extract. Each reaction contained 1 pgof deletion mutant DNA and 0.1pg of SV40DNA. Lane 1, PGER-B; lane 2, pGER-7; lam 3,PGER-9; lone 4, PGER-IO; lane 5, PGER-15; lam 6, PGER-16; lam 7,PGER-17; lone 8, PGER-1% lam 9, pGER-19. The primer extension product expected for SV40 is 149 bases and for EGF receptor is 117 bases.

that corresponds to initiation at a site 48 bp upstream of the start of translation (position -48 is a minor in vivo start site in A431 cells). In vivo, the EGF receptor promoter has multiple transcription start sites, and the major site is at -255 (17). A longerexposure indicated that weak transcription started at -255 in vitro as well (data not shown); we could not detect other transcription initiations which can be seen in vivo. To determine the regions which are required for the EGF receptor expression in vitro, we tested the transcriptional activities of several deletion mutants. As shown in Fig.2, deletion of the region upstream from -151 did not have any effecton transcription from the -48 start site. However, deletion of the region from -151 to -105 resulted in a 40% decrease in activity. Additional deletion of sequences from -105to -81 gave a further decrease, and a template beginning at -81 had extremely low activity (see Table I). Thus, about 100 bp upstream of the transcription initiation site was required in vitro for optimal transcriptional efficiency. Transient expression assays using various CAT constructs also indicated that the same regions were important for EGF receptor gene expression in vivo (18; Table I). Because the in vitro system showed parallel results to the in vivo transient expression experiments, we decided to characterize EGF receptor-specific transcription factors by monitoring transcription from the major in vitro start site. Competition Analysis of in Vitro Transcription-Previous data using exonuclease I11 footprinting showed that there were several A431 nuclear protein binding sites in the EGF receptor promoter region (18; see Fig.16).To determine if the binding factors were necessary for transcription, we carried out competition assays. For competition we used three different DNA fragments from the EGF receptor promoter (Fig. 3):

In vivo

In vitro

w e

8 9 1011

(-151)(-105) B AI

(-20)

H

w b

c

FIG.3. Competition analysis of in vitro transcription. Competition analysis of in vitro transcription was carried out using 1pg of pGER-7 (lanes 1-4) or 1 pg of pGER-10 (lones 5-12) in a final volume of 25 pl. No competitors were added in lanes 2, 6, and 8. In lanes 4,5, and I l , 140.90, and 90 ng of HaeIII-digested @Xfragments were added, respectively. The added competitor DNA fragments and their molar ratio to the template are indicated above each lane. The structures of the competitor DNA fragments are outlined in the bottom diagram. AI, AvaI; AII,AvaII; B, BglI; H,HindIII. -527 to -384 (a), -151 to -105 (b), and -105 to -20 (c). DNA fragment c contains two binding sites for nuclearfactors; DNA fragment b contains one binding site. Furthermore, both come from regions important for EGF receptor expression (Fig. 2). Fragment a is a control fragment because its sequences are unnecessary for in vitro transcription according to the deletion analysis shown in Fig. 2. We also usedHaeIIIdigested 4X fragments as another type of negative control. As shown in Fig. 3, a 5-fold molar excessof fragment a did not compete more than control 4X fragments. On the other hand, fragment 6 or fragment c competed effectively evenat a 2.5-fold molar excess. These results suggestthat trans-acting transcription factors bind to the regions between -151 and -105 and between -105 and -20 and are critical for in vitro transcription.

EGF Receptor Transcription Factors

6332

often stimulated transcription, especially whenthe activity of fraction B was low, we added fraction A to this assay system to allow further characterization of fraction B. Fraction B was next applied to DEAE-Sepharose CL-GB column. Three fractions were collected and the activity of each fraction was analyzed. Fig. 4c shows that all three fractions (BA,BB, and BC) were necessaryformaximal expression of both the EGF receptor promoter and the SV40 promoter. Without the 0.25 M KC1 fraction (BB) or 0.5 M KC1 fraction (BC), both SV40 and EGF receptor transcripts were greatly reduced (lanes 3 and 4). In contrast, the absence of the 0.12 M KC1 fraction (BA) didnot influence the transcripA431 NuclearExtract tion of SV40 as much as that of the EGF receptor ( l a n e 5). We consistently observed that theaddition of fraction BA to Hapafin AQarrwo BB and BC stimulated EGF receptor transcription around 5to 10-fold, whereasSV40 transcription was enhanced at most 0.1M 0.4M 1M A B 2-fold. These data raise the possibility that fraction BA conC tains an EGF receptor-specific transcription factor. A random transcription assay usingcalf thymus DNA as a template showed that fraction BC contained most of the RNA 0.12M 0.25M 0.5M polymerase I1 activity. There was still some transcription BA BB Bc without fraction BC, but it was probably due to thecontamination of RNApolymerase I1 in the otherfractions (especially ( POI 11 1 in fraction A). It was not due to transcription by the other Oligo Affinity RNA polymerasesbecause 1 pg/ml a-amanitin completely inhibited transcription (Fig. 4,b and c). Althoughfraction BB 0.12M OSM has not been wellcharacterized,we speculatethat this fraction BAl as well as fraction BC contains general transcription factors, BA2 because transcription from the SV40 and the EGF receptor template was similarly impaired in the absence of fractions BB and BC. DNase Z Footprinting of DEAE Fractions-To investigate the relationship between the DNA-binding factors and transcriptional activity, we carried out DNase I footprinting analyses using the DEAE fractions. Fig. 5a shows that fraction C BA possessed a factor which bound to both the noncoding U and coding strands of the region between -105 and -81; this region was important for transcription according to the deletion analysis (Fig. 2). We previously found that Splbound weakly to the region between -105 and -81 (18).This suggested that fraction BA contained Spl or Spl-like factors. However, fraction BA did not stimulate SV40 transcription in our system, although Spl is known to stimulate SV40 early transcription about 10- to 20-fold (21). This raised at least two different possibilities. One is that the DNA-binding factor in fraction BAwas different from Spl and specificallystimulated the EGF receptor transcription. The other is that Spl was supplied by SV40 SV40 another fraction and stimulation of EGF receptor-specific transcription was due to a different protein in fraction BA. To resolve this problem, we first analyzed all three DEAE EGm fractions (BA, BB,and BC) by the DNase I footprinting assay using a more sensitive probe that contained the SV40 promoter region. As shown inFig. 5b, fraction BB also protected the GC box regions, suggestingthat this BB-containing transcription assay system possessed enough Spl without fraction BA. 1 2 3 4 5 1234567 DEAE fraction BB showed protection of two regions of the FIG. 4. Fractionation of A431 nuclear extract. a, the A431 nuclear extract was fractionated according to this scheme. DEAE EGF receptor promoter: one around -130 and the other at fraction BC contained most of the RNA polymerase I1 activity. b, the -40 (see Fig. lb), both were previously detected by the exotranscriptional activities of the heparin-agarose fractions were tested nuclease I11 protection assay as specific nuclear binding sites using pGER-10 and SV40 templates. Lane 1 contained 1 pg/ml a- (18). The deletion mutant which does not contain the site amanitin. The fractions contained in the reactions are shown above around -130 (pGER-15) showed a 40% decrease in activity each lane. c, the transcriptional activities of the DEAE-Sepharose CL-GB fractions were analyzed using pGER-6 and SV40 templates. (Table I), suggesting that a nuclear factor in fraction BB Lane 1 contained 1pg/ml a-amanitin. Thefractions containedin the binds to this region and stimulates transcription. Purification and Characterization of the DNA-bindingFacreactions are shown above each lane.

Fractionation of A431 Nuclear Extract: Evidencefor an EGF Receptor-specific Factor in DEAE Fraction BA-To characterize the trans-acting factors A431 nuclear extracts were fractionated by using several column chromatography techniques basically accordingto Dynan and Tjian (29) (Fig. 4a). For the first step we used a heparin-agarosecolumn. As shown in Fig. 4b, the 0.4 M KC1 fraction (B) was sufficient for EGF receptor transcription. Because the 0.1 M KC1 fraction (A)

a

1 -

I

I

b

P

-

-

EGF Receptor Transcription Factors

6333

has been successfully used for purification of several DNAbinding transcription factors. We used short double-stranded oligonucleotides consisting of a 30-bp segment of the BAbinding site to make an affinity resin (Fig. 6a). Fraction BA was applied to this affinity column, and the 0.12 M KC1 flow-through fraction (BA1) and a 0.5 M KC1 -140 BB eluted fraction (BA2) were collected (Fig. 4a). Fraction BA2 was reapplied to the affinity column, washedwith 0.25 M KCl, -120 and eluted. After each cycle the eluate was analyzed by SDSpolyacrylamide gel electrophoresis along with Spl. As shown in Fig.66, both Spl and the factor in BA2 contained two -100 prominent bands with the same molecular mass of about 95 BA kDa. We also compared the DNA binding activity of both factors. We first determined the unit concentration of each sample by the DNase I footprinting assay using the SV40 promoter -80 region as a probe (one unit isdefined as theamount of protein required to bind the GC boxes and completely protect them from digestion by DNase I). Thenwe used equal units of both factors on a DNase I footprinting assay using the EGF receptor promoter DNA. As shown in Fig. 6c, both Spl and the factor in BA2 showed similar protection patterns at all con-60 centrations tested, indicating that they have the same affinity for the EGF receptor promoter region. Therefore, we concluded that the DNA-binding protein in fraction BA2 is identical or very similar to Spl. We next added the purified factor of fraction BA2 to thein uitro transcription system containing BB and BC to test whether the purified protein is able to replace fraction BA. However, it did not show any stimulation of either EGF BB receptor transcription or SV40 transcription (Fig. 7a). We thus concluded that fraction BB contains a saturating level -40 of Spl, and the EGF receptor-specific stimulating activity of fraction BA was not due to Spl but caused by a different 1 2 3 4 5 6 7 8 9 1 0 factor. Our deletion and competition analyses shown above, however, indicated that Spl is also required for EGF receptor b BA BB BC transcription. To test directly the Spl dependency, we depleted Spl from fraction BB by using an oligonucleotide affinity column. When we used the Spl-depletedfraction BB, addition of the purified Spl stimulated SV40 and EGF receptor transcription approximately 2- to 4-fold (Fig. 76). These data thus demonstrate that Spl is required for optimal transcription of EGF receptor. .. cu Further Characterization ofFraction BA-To obtain further evidence that EGF receptor-specific stimulating factor is dif"ad ferent from Spl, we applied fraction BA to a Sephacryl S-300 1 2 3 4 sizing column and tested fractions for binding activity and in FIG. 5. DNase I footprinting analysis. a, the DNA-binding uitro transcription activity (Fig. 8).Because the fractions were activities of heparin-agarose fractionB and the DEAE fractions were diluted and the protection patterns somewhat weak, we quananalyzedby using a noncoding (lanes 1- 5) or coding (lanes 6-10) strand of the EGF receptor probe. All reactions contained 5 ng of the titated their binding activity by calculating the ratio of the end-labeledprobe. Lanes 1 and 6 show the control DNase digestion densities of the DNase I-hypersensitive band (A) and the pattern with no protein factors added. The reactions in the other protected band (B) (Fig. 86). As shown in Fig. Sa, the peak lanes contained 50 pg of each fraction as indicated above the lanes. fraction of Spl-like DNA-binding activity was around fraction The lanes G + A and C show the sequence ladders produced fromthe 58, which corresponds to a molecular mass of 500 kDa. This noncoding DNA probe. The nucleotide residues ofthe EGF receptor promoter are indicated on the left of the autoradiograph, and the agrees well with the native molecular mass of Spl reported regions protected by DEAE fractions BA and BB are shown on the by Briggs et al. (21). However, this fraction did not show the right. b, the DNA-bindingactivities of DEAE fractions were analyzed strongest stimulatory activity when in uitro transcription of by using an SV40 probe. Lane 1 shows the control DNase digestion the EGF receptor template was measured. The peak of EGF patternwithnoproteinadded. The reactions in the other lanes receptor-specific in uitro transcription activity was around contained about10 pg of each fraction as indicated abovethe lanes. fraction 63, which corresponds to a molecular mass of about 270 kDa (Fig.&).This experiment was repeated with identical tor in Fraction BA-To prove whether the DNA-binding results. Therefore, we conclude that fraction BA contained factor in fraction BA is Spl, we further purified the factor an EGF receptor-specifictranscription factor which is differusing oligonucleotide affinity chromatography. This method ent from Spl.

-

1

-

I

-

-

1

-

J!

EGF Receptor Transcription Factors

6334

a FIG.6. Purification of the DNAbindingproteinin fraction BA. a, the sequences of the two oligonucleotides, 40-mer and IO-mer, are shown. The hybridized segment of the oligonucleotides corresponds to theregion from -108 to -79 of the EGF receptor promoter. The hybridized oligonucleotides were coupled to CNBr-activated Sepharose 4B. b, eluates from oligonucleotide affinity step were analyzed by SDS-polyacrylamide gel. Lanes 1, 2, and 3 show the 0.5 M KC1 fractions of the first, second, and thirdaffinity step,respectively. Lane 4 shows purified Spl. After electrophoresis, gel was stained with silver. The sizes of the protein markers are shown on the left. c, the DNA-binding activities of a purified factor in BA2 (lanes 2-4) and Spl (lanes 5-7) were analyzed by DNase I footprinting assay using a coding strand probe of EGF receptor promoter. Lanes 1 and 8 show the control DNase digestion pattern with no protein factors added. The unit amounts used are shown above each lane. Theunit concentration was determined as described in the test. The protected region is shown on the right.

C 5'

3'

CACGACGCGACCGCCCGAGTCCCCCGCCTCGCCGCCAACGC GOCGGGCTCAGGGGCGGAGCGGCGGTGCG 3' 5'

b

-

200

-

116 97.4

-

66.2

-

42.7

I'

Iw

1

a

b

4-

2

3

4

major start site showed only a weak initiation in the in vitro experiments. This type of difference between in vivo and in vitro studies maybecommonfor promoters that do not EGFR SV40 contain a TATA box. For example, the murine dihydrofolate n n reductase gene (33) and the hamster 3-hydroxy-3-methylglu-+ -+ -+ taryl-coenzyme A reductase gene (34), neither of which has a TATA box, also showed different efficiency of utilization of multiple transcription start sites between the in vivo and in I . vitro experiments.This indicates that unknown factors equiv"sV40 alent to aTATA-binding protein or TFIID (35) might be lost during the preparation of the nuclear extracts, or that a more highly organizedstructure might be required to mimic in vivo transcription initiation. However, becauseboth in uiuo and in vitrodeletion analyses showed that the sameregions are important for the optimal expression of the EGF receptor 12 1 2 34 gene, the transcription initiating at -48 may be predominant FIG.7. I n vitro transcriptional analysis of purified Spl. a, in vivo as well with the mutants that do not contain the in in vitro transcriptional analysis was done by using pGER-6 in the vivo major start site. absence ( h n e 1 ) or presence (lane2 ) of purified Spl from BA2. Both Spl Is Required for EGF Receptor Transcription-Previreactions contained fractions A, BB, and BC. b, in vitro transcripously we found that regionsbetween-151 and -105 and tional analysis was done by using Spl-depleted fraction BB in the absence (lanes 1 and 3) or presence (lams 2 and 4 ) of purified Spl between -105 and -81 were necessary foroptimal expression from BA2. of the EGF receptor promoter using promoter CAT constructs in transient expression assays, and that specific nuclear factors bound to those regions (18).The in uitro deletionanalyses DISCUSSION and competition experiments described here clearly showed We have developeda cell-free transcription system for the that nuclear factors that bind to the above regionsare imporEGF receptor gene promoter using A431 nuclear extracts and tant for transcription. This indicates that atleast two differshowed that several transcription factors positively regulate ent DNA-binding proteins positively regulate EGF receptor EGF receptor expression in vitro. Two of these factors have transcription in vitro as well as in vivo. Purification by olibeen characterized. One is similar or identical to Spl. The gonucleotideaffinity chromatographyrevealed that one of the other is a novel factor whichspecifically stimulates EGF factors is identical or very similar to Spl, and addition of the receptor transcription about 5- to 10-fold. purified factor stimulated EGF receptor transcription. Thus, Transcription Initiation Sites-Although the EGF receptor we conclude that Spl is required for optimal EGF receptor promoter has multiple transcription start sites in vivo,we transcription. In our in vitrotranscription system, the purified found only one predominant site (-48) and one weak site Spl showed a weak stimulation of SV40 transcription because (-255) in the in vitro transcription assay. The in vitro major Spl-depleted fraction BB still contained some Spl-like activstart site is a minor initiation site in uiuo, whereas the in vivo ity even after two cycles of oligonucleotide affinity chroma-

m l

ba

6335

FIG.8. Analysis of DNA-binding and transcriptional activities of the 5-300 column. a, DEAE fraction BA was applied to a Sephacryl S-300column, and the DNA-binding and transcriptional activities were analyzed. The DNA-binding activity (e) was measured as described in the text. The transcriptional activity(0)was estimated by densitometric scanning. I and 2 on the right ordinate mean no stimulationand 2-fold stimulation of the EGF receptor transcription compared to SV40 transcription, respectively. b, DNase I footprinting analysis was carried out by using the coding strand probe. Lane I shows the control DNase digestion pattern with no protein factors, and lane 2 contained 50 pg of fraction BA. The reactions of the other lanes contained 30 pl of each S300 fractions. A, DNase I-hypersensitive band; B, protected band. c, in vitro transcription analysis was done by using pCER-6 as a template.All reactions contained fractions A, BB, and BC, and 43 pl of each S-300 fraction was used for the assay.

b

C

-

- BA

BA 50 58 70

61 63 65 67

"n I

"

.

"

9 . w

.I)$"

SV40

ww-EGFR

.Y

"s

I

i"

1

2

3

tography. Thus, Spl dependency of SV40 and EGF receptor transcriptions should be greater than observed in our assay system. It was previously shown that Spl also binds to the promoter region of the c-Ha ras proto-oncogene (36). Therefore, it would be interesting to know whether the overproduction or activation of Spl or Spl-like factors might be involved in carcinogenesis by activating several oncogenes at the same time. In our transcription system containing untreated fraction BB, Spl did not stimulate SV40 transcription because fraction BB contained a saturating level of Spl. However, we did not detect the Spl-likeDNA-bindingactivity of fraction BB with EGF receptor probes, but we did with an SV40 probe. This is probably because the binding affinity of Spl to this EGF receptor promoter site is weak compared to the GC boxes of SV40 (18) and because fraction BB contains less Spl than fraction BA. EGF Receptor-specific Transcription Factor-Sephacryl S-

4

5

1

2

3

4

5

6

300 chromatography experiments showed that fraction BA contained a novel EGF receptor-specific transcription factor, which we call ETF, in addition to Spl (Fig. 8).The molecular mass of the native form of ETF was estimated to be about 270 kDa. This isclearly different from Spl, which has a molecular massof about 500 kDa (21) (Fig. 8). Because ETF specifically stimulates EGF receptor transcription up to 10fold, we speculate that it is a key factor for transcriptional activation of cellular oncogenes. In this context, ETF also might be an oncogene product. Recently, it was found that the jun oncoprotein and AP-1 recognized essentially the identical sequence, suggestingthat the jun oncoprotein is derived from the normal AP-1 transcription factor (37). Thus, it is possible that ETFis activated or increased in carcinoma cells such as A431 cells. However,analysis of intriguing questions such as whether the amount or activity of ETF correlates to the transcriptional level of oncogenes and whether ETF is able to activate other cellular oncogenes awaits further purification of this factor.

EGF Receptor Tramscription Factors

6336

ETF did not appear to bind the DNA template strongly because we could not detect any specific DNA-binding activity other than Spl-like activity with fraction BAby DNase I footprinting assay. The mechanism of howthis factor discriminates between the SV40 and EGF receptor promoters is not yet understood. A similar type of transcriptional stimulation was reported by Spangler et al. (38) for adenovirus EIA protein. They showed that purified EIA protein specifically stimulated in vitro transcription from adenovirus promoters, although EIA by itself is not a DNA-binding factor. It has been postulated that EIA might increase the concentration or activity of a cellular transcriptional factor (39). Recently, Reinberg et al. (40) characterized general transcription factors and proposed that a transcription factor TFIIA, which does not show any specific DNA-binding activity by DNase I footprinting assay, does associate with the DNA template and induces stable binding of TFIID toa TATA box. Our preliminary studiesshowed that ETFcan bind to theoligonucleotide affinity column in low (0.12 M) KC1 but not at 0.25 M KC1, whereas S p l can bind to the column with 0.25 M KCl. This result suggests that ETF is able to associate weakly with the DNA template. Therefore, it is possible that thisfactor shows a weak interaction with the DNA template, recognizes a specific sequence, and determines the promoter selectivity. However, whether the factor associates directly with a specific sequence in the DNA template and/or whether there is any interaction between ETF and other transcriptionfactors such as Splremains to be determined. By using the in vitro transcription system described here with purified factors, it may be possible to begin to understand the mechanism of regulation of EGF receptor transcription. Thesestudies should also help analyze the mechanism of transcriptionalactivation of other oncogenes with similar promoter structures. Acknowledgments-We thank Dr. J. T. Kadonaga and Dr. R. Tjian for providing purified Spl; Dr. J. N. Brady for critical discussion and technical advice; Dr. M. M. Gottesman, Dr. S. Adhya, Dr. P. Marino, and Dr. A.C. Johnson for useful discussion; Dr. C. Wu and Dr. F. Amano for technical advice; B. Lovelace and A. Harris for excellent maintenance of cell culture; C. Parkison for sequencing; S. Neal for photography; and Dr. P. Rossi for CAT-specific primer. REFERENCES

1. Savage, C. R., Jr., and Cohen, S. J. (1972)J. Biol. Chem. 247, 7609-7611 2. Das, M., Miyakawa, T., Fox, C. F., Pruss, R. M., Aharonov, A., and Herschman, H. R. (1977)Proc. Natl. Acad. Sci. U. S. A. 74,2780-2794 3. SahYoun, N.9 Hock, R. A.9 and Hollenberg, M- D. (1978)PrOC. Natl. Sci. Acad. U. S. A. 75. 1675-1679 4. Ushiro, H., and Cohen, S. (1980)J. Bwl. Chem. 225,8363-8365 5. Downward, J., Yarden, Y., Mayes, E., Scrace, G., Toffy, N., Stockwell, P., Ullrich, A., Schlessinger, J., and Waterfield, M. D. (1984)Nature 307, 521-527 6. Yamamoto, T., Nishida, T., Miyajima, N., Kawai, S., Ooi, T., and Toyoshima, K. (1983)Cell 35, 71-78 7. Merlino, G. T., Xu, Y.-h., Ishii, S., Clark, A. J. L., Semba, K., Toyoshima, K., Yamamoto, T., and Pastan, I. (1984)Science 224,417-419 8. Ullrich, A., Coussens, L., Hayflick, J. S., Dull, T. J., Gray, A., Tam, A. W., Lee, J., Yarden, Y., Libermann, T. A., Schlessinger, J., Downward, J., Mayes, E. L. V., Whittle, N., Waterfield,

M. D., and Seeburg, P. H. (1984)Nature 309,418-425 9. Lin, C. R., Chen, W. S., Kruiger, W., Stolarsky, L. S., Weber, W., Evans, R. M., Verma, I. M., Gill, G. N., and Rosenfeld, M. G. (1984)Science 224,843-848 10. Cowley, G., Smith, J. A., Gusterson, B., Hendler, F., and Ozanne, B. (1984)Cancer Cells 1,5-10 11. Merlino, G. T., Xu, Y.-h., Richert, N., Clark, A. J. L., Ishii, S., Banks-Schlegel, S., and Pastan, I. (1985)J. Clin.Znuest. 75, 1077-1079 12. Yamamoto, T., Kamata, N., Kawano, H., Shimizu, S., Kuroki, T., Toyoshima, K., Rikimaru, K., Nomura, N., Ishizaki, R., Pastan, I., Gamou, S., and Shimizu, N. (1986)Cancer Res. 46, 414-416 13. Xu, Y.-h., Richert, N., Ito, S., Merlino, G. T., and Pastan, I. (1984)Proc. Natl. Acad. Sci. U. S. A. 81, 7308-7312 14. Libermann, T. A., Nasbaum, H. R., Razon, N., Kris, R., Lax, I., Soreq, H., Whittle, N., Waterfield, M.D., Ullrich, A., and Schlessinger, J. A. (1985)Nature 313, 144-147 15. Velu, T. J., Beguinot, L., Vass, W.C., Willingham, M. C., Merlino, G. T., Pastan, I., and L o w , D. R. (1987)Science 238, 14081410 16. King, C. R., Kraus, M.H., Williams, L. T., Merlino, G. T., Pastan, I. H., and Aaronson, S. A. (1985)Nucleic Acids Res. 13,8477-8486 17. Ishii, S., Xu, Y.-h., Stratton, R. H., Roe, B. A., Merlino, G. T., and Pastan, I. (1985)Proc. Natl. Acad. Sci. U. S. A. 82,49204924 18. Johnson, A. C., Ishii, S., Jinno, Y., Pastan, I., and Merlino, G. T. (1988)J. Biol. Chem., in press 19. Manley, J. L., Fire, A., Cano, A., Sharp, P. A., and Gefter, M. L. (1980)Proc. Natl. Acad. Sci. U. S. A. 77,3855-3859 20. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G. (1983)Nucleic Acids Res. 11, 1475-1489 21. Briggs, M. R., Kadonaga, J. T., Bell, S. P., and Tjian, R. (1986) Science 234,47-52 22. Chodosh, L. A., Carthew, R. W., and Sharp, P. A. (1986)Mol. Cell. Bwl. 6, 4723-4733 23. Jones, K. A., Kadonaga, J. T., Rosenfeld, P. J., Kelly, T. J., and Tjian, R. (1987)Cell 48,79-89 24. Reinberg, D., and Roeder, R. G. (1987)J.Biol. Chem. 262,33103321 25. Angel, P., Imagawa, M., Chiu, R., Stein, B., Imbra. R. J., Rahmsdorf, H. J., Jonat, C., Herrlich, P., and Karin, .M. (1987)Cell 49,729-739 26. Lee, W., Mitchell, P., and Tjian, R. (1987)Cell 49,741-752 27. Gorman, C. (1985)in DNA Cloning ZZ (Glover, D.M., ed) pp. 143-190,IRL Press, Wash., D. C., Oxford 28. Wildeman, A. G., Sassone-Corsi, P., Grundstroem, T., Zenke, M., and Chambon, P. (1984)EMBO J. 3.3129-3133 29. Dynan, W. S., and Tjian, R. (1983)Celi 32, 669-680 30. Kadonaga, J. T., and Tjian, R. (1986)Proc. Natl. Acad. Sci. U. S. A. 83,5889-5893 31. Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982)Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York 32. Dynan, W. S., and Tjian, R. (1983)Cell 35,79-87 33. Farnham, P. J., and Schimke, R. T. (1986)Mol.Cell. Bid. 6, 3293-3An1 " " ""

34. Osborne, T. F., Gil, G., Brown, M. S., Kowal, R. C., and Goldstein, J. L. (1987)Proc. Natl. Acad. Sci. U. S. A. 84,3614-3618 35. Sawadogo, M., and Roeder, R. G. (1985)Cell 43, 165-175 36. Ishii, S., Kadonaga, J. T., Tjian, R., Brady, J. N., Merlino, G. T., and Pastan, I. (1986)Science 232, 1410-1413 37. Struhl, K. (1987)Cell 50,841-846 38. Spangler, R., Bruner, M., Dalie, B., and Harter, M.L. (1987) Science 237,1044-1046 39. Kovesdi, I., Reichel, R., and Nevins, J. R. (1986)Cell 45, 219228 40. Reinberg, D., Horikoshi, M., and Roeder, R. G. (1987)J. Bwl. Chem. 262,3322-3330