tant for the function of the promoter of the epidermal growth factor (EGF) receptor gene and to characterize the factors that bind to these sites, the promoter ...
Val. 263,No. 12,Issue of April 25,pp. 5693-5699, 1988 Printed in U.S.A.
THEJOURNAL OF BIOLOGICAL CHEMISTRY
Epidermal Growth Factor Receptor Gene Promoter DELETION ANALYSIS AND IDENTIFICATION OF NUCLEAR PROTEIN BINDING SITES* (Received for publication, April 20, 1987)
Alfred C. Johnson$, ShunsukeIshiiQ, Yoshihiro Jinno, Ira Pastan, and Glenn T. Merlin0 From the Laboratory of Molecular Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892
To determine the location sites of that may be impor- without rearrangement, and increased expression without tant for the function of the promoter of the epidermal gene amplification (16). Regulation of expression of the EGF growth factor (EGF) receptor gene and to characterizereceptor protooncogene is undoubtably necessary to ensure the factors that bind to these sites, the promoter region proper growth control in many cell types. was analyzed by deletion analysis, exonuclease I11 proThe EGF receptor gene promoter has been identified, isotection andgel retardationassays with crude andfrac- lated, and characterized (17). The promoter region of the EGF tionatednuclear extracts and DNase Ifootprinting receptor gene has several special features. It lacks a characusing purifiedSpl. Transfection of chimeric chloram- teristic TATA box and CAATbox andcontains multiple phenicol acetyltransferase plasmidscontainingvartranscription start sites; it is very G + C rich (88%);and in ious deletions of the EGF receptor gene promoter into A431 cells that overexpress the EGF receptor there is a DNase CV- 1 cells indicated that the region between -178 and -16 (initiator ATG is +1) is sufficient for promoter I hypersensitive site that is situated close to the promoter activity. ExonucleaseI11 protectionassays revealed the region (4-6).Previously, we determined that theEGF receptor presence ofeight specific nuclear protein binding sites promoter contains five GC boxes (17) similar to those known in the region between-481 and -16. Gel retardation to bind the SV40 transcription factor Spl (18-21). Spl has assays confirmed that multiple protein binding sites been shown to bind to GC boxes in several viral gene proexist in this region (-481 to -16) and quantitatively moters (22-24). Cellular gene promoters, including c-Ha-rasl agree with exonuclease I11 protection. DNase I foot- and mouse dihydrofolate reductase, also have GC boxes printing using purified Spl showed thatthis transcrip- known to bind Spl (25-27). In thisstudy, we have characterized the EGF receptor gene tion factor can bind to four sites (-457 to -440, -365 to -286, -214 to -200, and -110 to -84) in the EGF promoter by deletion analysis using an i n vivo transfection receptor gene promoter and therefore may play a role assay, by exonuclease I11 protection, gel retardation, and by in its regulation. DNase I footprinting using purified Spl. We have identified sites in the promoter that bind nuclear proteins, and regions that maybe important in regulating EGF receptor gene expression. Epidermal growth factor (EGF)’ is a potent mitogen that EXPERIMENTALPROCEDURES acts via a specific membrane receptor to stimulate cell growth (1-3). The EGF receptor is a transmembrane glycoprotein Materials that has three domains: 1)a hydrophobic membrane-spanning T4 polynucleotide kinase, T, DNA polymerase, T, DNA ligase,and domain; 2) an amino portion that is glycosylated and binds E G F and 3) a carboxyl portion which is responsible for EGF- poly(dI. dC) were all purchased from Pharmacia LKB Biotechnology Inc. DNase I was obtained from Cooper Biomedical. Exonuclease 111 dependent tyrosine kinase activity (4-6). The EGF receptor was bought from Boehringer Mannheim. Heparin-agarose was purtransmembrane and kinase domains are homologous to the chased from Bethesda Research Laboratories (BRL). Restriction viral oncogene erb B; therefore, the EGF receptor is consid- enzymes were obtained from BRL, New England Biolabs, and Pharmacia LKB Biotechnology Inc. Radiolabeled [32P]NTPsor dNTPs ered to be the c-erb B gene product (7, 8). The level of EGF receptor gene expression varies in normal were purchased from Amersham Corp. and [‘4C]chloramphenicolwas and transformed cell types (9). The EGF receptor gene is bought from DuPont-New England Nuclear. amplified in A431 human epidermoid carcinoma cells (10-12), Methods several human squamous carcinoma cell lines (13, 14), and Construction of Deletion Mutants-The pEP5 plasmid containing some glioblastomas (15). Examples have been found in malig- a 5.3-kb EcoRI fragment was digested with SstI (17). T, DNA polymnant cells in which the EGFreceptor gene has been subjected erase (3’-5’ exonuclease activity) was used to remove 3”singleto gene amplification with rearrangement, gene amplification stranded ends. The fragments were subjected to HindIII linker liga-
* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The nucleotide sequence($ reported in this paper has been submitted to the GenBankm/EMBL Data Bank withaccessionnumber(s) 503730. $ Supported by Grant PF-2787 from the American Cancer Society. Present address: Inst. of Physical and Chemical Research, Koyadai 3-1-1Yatabe, Tsukuba, Ibaraki, 305 Japan. ’The abbreviations used are: EGF, epidermal growth factor; kb, kilobase; bp, base pair.
tion and digestion. The fragments were electrophoresed on an agarose gel, and a 1.1-kb fragment was isolated. This fragment was cloned into thepSVOCAT HindIII site(28).This resulted in a new construct, PERCAT-1, in which the EGF receptor gene promoter region now controlled the expression of the bacterial chloramphenicol acetyltransferase gene. Similar strategies, beginning with the 1.1-kb fragment, were used to construct smaller deletion mutants as shown in Fig. 1 using the specific enzymes indicated at the 5’ end of each construct. Transfection Assay-Plasmid constructs were transfected into African green monkey CV-1, or KB3-1 human epidermoid carcinoma cells that were seeded at 4-5 X lo6 cells/lOO-mm dish 24 h prior to addition of the CaP0,-DNA precipitate (29). One milliliter of the
5693
EGF Receptor Gene Promoter
5694
precipitate containing 20 pg of pERCAT plasmid DNA was added to the medium; the cells were incubated for 4 h at 37 “C and then exposed to glycerol for 3 min (CV-1 cells) or 30 s (KB cells) (30). Cells were harvested 48 h after transfection, and cellular extracts were prepared by freeze-thawing three times and by sonication. After a brief centrifugation to remove cell debris, extracts were incubated with [‘4C]chloramphenicoland acetyl-coenzyme A for 30 min and chloramphenicol acetyltransferase activity measured by thin-layer chromatography as described previously (28, 31). Preparation of Extracts and DNA Labeling-Nuclear protein extracts were prepared from A431 and KB3-1 cells as described previously (32). Extracts were fractionated on a heparin-agarose column and protein fractions pooled according to eluting salt concentration (see figure legends). DNA fragments were end-labeled with [y-”P] ATP andTI polynucleotide kinase. Exonuclease ZIZ Protection Assay-DNA fragments labeled with 32Pwere incubated for 15 min with nuclear extract followed by digestion with exonuclease 111(16,000 units/ml) at 30 “C asdescribed previously (33). This procedure was modifiedin that theexonuclease 111 concentration was approximately 5-fold higher, the time of digestion was increased from 10 to 60 min, and additional nonspecific DNAs (poly(dI.dC), 0.2 units/ml; X HindIII fragments, 0.15 pg/ml; and pBR322, 8 pg/ml) and bovine serum albumin (60 pg/ml) were added. These conditions gave the best digestion of the EGF receptor GC-rich promoter fragments. In competition assays, the nuclear extracts were incubated for 10 min a t 24 “Cwith competitor fragments before addition of labeled fragment. Gel Retardation Assay-DNA fragments end labeled with [r-”P] ATPand T, polynucleotide kinase were incubated with nuclear extracts at room temperature for 15 min in the presence of 50 mM Tris, pH 7.5, 10 mM EDTA, 12.5 A2,/ml poly(dI.dC), 250 pg/ml bovine serum albumin, and 5% glycerol. Samples (20 pl) were loaded onto a 5% polyacrylamide gel and electrophoresed for 120 min at 150 V using TBE (89 mM Tris, 89 mM boric acid, 2 mM EDTA) as running buffer. After electrophoresis, gels were dried and exposed to XAR-2 film with intensifying screens. DNase Z Footprinting-DNase I footprinting was performed according to Dynan et al. (18, 19). S p l was purified as described previously (34, 35). Twenty-five and fifty ng of purified S p l (>go%) was incubated with an end-labeled PstI-Hind111EGF receptor gene promoter fragment and then digested with DNase I. After digestion, samples were subjected to electrophoresis on an 8% denaturing polyacrylamide gel. RESULTS
Deletion Analysis of Promoter Region-To identify regions required for EGF receptor gene expression, a series of deletion fragments of the EGFreceptor gene promoter were ligated 5’ to the bacterial chloramphenicol acetyltransferase gene and tested for promoter activity by transfection intoAfrican green monkey kidney cells (CV-1) and human KB epidermoid carcinoma cells. The structure of the various deletions and their relative promoter activities in the transfection assay (compared to PERCAT-1) are shown in Fig. 1. The results of this analysis in CV-1 cells show that the5’-most region, -1109 to -911 (PERCAT-4) could be removed without substantially altering promoter activity. However, removal of the region between -1109 and -850 (pERCAT-5), -1109 and -569 (pERCAT-7), or -1109 and -481 (PERCAT-8) resulted in a 46-68% decrease in CAT expression. Furthertruncation to -384 (PERCAT-9) or -178 (PERCAT-10) restored the activity to above PERCAT-1 levels. Truncation to -105 (PERCAT-15) reduced chloramphenicol acetyltransferase expression to approximately 68%. Another plasmid, PERCAT-14, was constructed which only contained the sequences from -167 to -105. When this 62-bp DNA fragment was used as a promoter, chloramphenicol acetyltransferase was expressed at a level that was64% of the original PERCAT-1 activity. Two constructions lacking this region (PERCAT-12 and pERCAT-13) exhibited markedly reduced promoter activity. These data indicate that the region from
pERCAT-4 PERCAT-5 PERCAT-6
-911 i82 P
-771 I
T
pERCAT-7
121 115
I
32
0
pERCAT-8
-481
PERCAT-10 -1109
121 - l W A F
PERCAT-12 1 -771 I
54
45
-=
PERCAT-9
PERCAT-13 PERCAT-14 PERCAT-15
I I -I
-167’
i
H
-105
H
-167H-105 -105 H
120* 14 31 64* 68*
FIG. 1. Deletion analysis of the EGF receptor gene promoter. The series of chimeric plasmids ( h u e ) were constructed and transfected into CV-1 cells (see “Methods”). Promoter activity was determined and is expressed relative to PERCAT-1 which contains a 1.1-kb promoter DNA fragment and is corrected for transfection efficiencies using RSV#?-gal as an internal control. A schematic depicting the PERCAT-1 essential components is shown above the constructs. H3, HindIII; S, StuI; S2, SstII; B2, BglII; P, PuuII; T, TagZ; D, DdeI; A2, AuaII; Bl, BglI; Al, AuaI. Position -16 refers to 16 bp upstream of the translational start site (ATG). P,transcriptional start sites. 6, GC boxes with the hexanucleotide sequence CCGCCC. *, indicates the presence of upstream non-EGF receptor transcriptional start sites detected by primer extension (data not shown). CAT, chloramphenicol acetyltransferase.
-167 to -105 may be important for maximal promoter function. CV-1 cells express very low levels of EGF receptor. To determine if other cell types exhibited a similar pattern upon DNA transfection, human KB epidermoid carcinoma cells, known to express moderately high levels of EGF receptor (9) were utilized. The patternof chloramphenicol acetyltransferase expression resulting from transfection of the various deletion constructions into KBcells generally resembled that found in CV-1 cells (data not shown). Exonuclease Protection-To determine possible sites where nuclear proteins might bind to the EGF receptor promoter, various promoter DNA fragments were prepared, end-labeled and subjected to digestion with exonuclease 111 in with 32P, the presence or absence of extracts containing nuclear protein. Fragments chosen for analysis lie within one of the regions that the deletion analysis indicated was important for promoter activity (-481 to -16). Extracts from A431 cells, KB3-1, and HeLa cells were employed becausethese cell lines express the EGF receptor gene. All three types of extracts gave similar results. Exonuclease digestion was carried out as described by Wu (33) except that additional types of DNA were added to reduce nonspecific interactions, and bovine serum albumin was added as aprotein stabilizer. Fig. 2 shows a typical experiment in which a 465-bp fragment (from -481 to -16) labeled at one 5‘ end was studied. As the concentration of A431 extract was increased, a number of new bands appeared above the background pattern and did not appear in heat denatured samples as indicated by the arrows. The location of these bands within the 465-bp fragment was determined by comparison with DNA sequencing reaction markers. Each DNA fragment was also labeled at theother 5’ end, and an identical analysis was performed. In this way the location of each protected site was confirmed and the boundaries of each site were determined (data not shown). Additional support for the location of these sites was obtained by
EGF Receptor Gene Promoter
A
MAP POSITION
B
I
5695 0.3MHA
O.6MHA
-
/*
MAP POSITION -480 -420
-420 -325
-325
c 4-
-258 t
-258
-206
-163
4-
-206 4-
-12f
-IO€ -163
-8:
-139
FIG. 2. Exonuclease 111 footprinting of the EGF receptor gene promoter region. The 465-bp DdeIHind111 fragment 32P-end-labeledat the HindIIIsite was incubated in the presence or absence of nuclear extract followedby exonuclease I11 digestion as described under "Methods." Samples were electrophoresed on a 5% denaturing gel (two loadings). A, second loading. B, first loading. After electrophoresis, the gel was subjected to autoradiography a t -70 "C.Lanes 1 and 14, pBR322 DNA MspI-digested labeled fragments; lane 2, no extract; lanes 3-5,5,10, and 20 pg of A431 crude nuclear extract, respectively; lanes 6-8,5,10, and 20 pg of 0.3 M heparinagarose (0.3MHA) fraction, respectively; lane 9,20 pg of heat-denatured 0.3 M heparin-agarose fraction; lanes 1012,5, 10, and 20 pg of 0.6 M heparin-agarose fraction (O.GMHA),respectively; lane 13,20 pg of heat-denatured 0.6 M heparin-agarose fraction. Arrows indicate regions of the gel where protected bands appear. Two loadings were performed to separate more clearly different regions of the gel.
end labeling smaller DNA fragments and performing identical analyses (data not shown). To determine if the binding sites detected by exonuclease digestion were specificallyprotected, various unlabeled DNA fragments from either theregion under study or other regions wereused to compete for protein binding. Fig. 3 shows a typical result where eight of the protected sites were more susceptible to exonuclease I11 digestion when specific competition fragments were added to the binding reactions. Much less competition was detected when either a 340-bp BamHIHindIII fragment of pBR322 (Fig. 3) or a mixture of 4x174 fragments (not shown) were added. In addition, the factors responsible for exonuclease protection of the eight specific sites described above werefound to be sensitive to heating at 80 "C for 5 min (Fig. 2), reinforcing the idea that protected sites are the result of protein binding. Identical results to
these with A431 cell extracts were obtained with HeLa and KB3-1 cells. To characterize furtherthe components responsible for exonuclease protection, the A431 nuclear extract was fractionated on a heparin-agarose column. As shown in Figs. 2 and 3, most of the factors responsible for specificbinding sites appear in the 0.6 M NaCl fraction. Two prominent bands appear in the 0.3 M NaCl fraction. The sizes of the protected bands were determined, and the results of more than forty experiments with various fragments are summarized in Fig. 4. Only the specific binding sites in the 465-bp fragment are shown. These sites are considered specific based on extract titration, heat denaturation, competition, and fractionation. Sequences that were found to be protected included three of the five original GC boxes. No binding factors were found in the 1.0 M NaCl fraction. Gel Retardation-To support the data obtained by exonu-
Receptor
5696
EGF
Gene Promoter
A
B
MAP
0.3MHA
-
O.6MHA
L
L "
+
+
"
L "
MAP POSITION -400
I
-420
I
-325
I
-250
a a
-206 a
!I=.
,
-=t
I
3
.
..
-126
I
"-
-
m
.-,
-106
m "
"
-206
" "
-03 D
i
-50
';.
\
-163
-b
E -139
I -126
3
4
5
6
7 0
9 1011?21314
1
2
3 4
5
6
7 8
91011121314
FIG.3. Competition analysis using exonuclease I11 protection assays. In competition assays (lanes 4-5, 7-9,and 11-13), 20 pg of A431 nuclear extract (crude or fractionated) was incubated for 10 min a t 24 "C with unlabeled specific competitor (465 bp of DdeI-Hind111 EGF receptor gene promoter fragment) or nonspecific competitor (340 bp of BamHI-Hind111 pBR322 fragment) before addition of the labeled fragment. As in Fig. 2: A , second loading; and B, first loading of the gel. Lanes 1 and 14, pBR322 mspI markers; lane 2, no extract; lane 3, crude extract, no competitor; lane 4, crude extract plus 100-fold molar excess competitor; lane 5,crude extract plus 200-fold molar excess nonspecific competitor; lane 6, 0.3 M heparin-agarose fraction (0.3MHA), no competitor; lanes 7 and 8, 0.3 M heparin-agarose fraction plus 50- and 100-fold molar excess competitor, respectively; lane 9, 0.3 M heparin-agarose fraction plus 200-fold molar excess nonspecific competitor; lane 10, 0.6 M heparin-agarose (0.6MHA) fraction, no competitor; lanes 11 and 12, 0.6 M heparin-agarose fraction plus 50- and 100-fold molar excess competitor, respectively; lane 13, 0.6 M heparin-agarose fraction plus 200-fold molar excess nonspecific competitor. Arrows indicate bands that arespecifically competed in this analysis.
clease I11 protection, gel retardation assays were performed using three fragments that span the -481 to -16 region. As shown in Fig. 5A, using a fragment from -481 to -384, two major retarded bands are detected using either the crude (lane 2 ) or the 0.6 M heparin-agarose fraction (lane 1 0 ) which are competed with the specific DNA fragment (lanes 3,4,11, and 12) but not with a nonspecific fragment (lanes 5 and 13). A weak band that is not competed as well as the major band is produced by a factor located in the 0.3 M heparin-agarose fraction (lanes 6-9). When a fragment from -384 to -178 was used in this assay, three retarded bands were detected with the crude nuclear extract (Fig. 5 B ) . Two of these bands
were produced in the 0.6 M heparin-agarose fraction and one by the 0.3 M heparin-agarose fraction (Fig. 5B, lanes 6 and 10). The bands competed out with specific fragments (lanes 3, 4, 8, 9, 11, and 12) but not with nonspecific fragments (lanes 5, 9,and 13). A similar result was obtained when the fragment from -150 to -16 wasused in this assay. Two retarded bands that were specificallycompeted were detected in the crude nuclear extract (Fig. 5C, lanes 2-5). These appear to result from factors contained in the 0.6 M heparin-agarose fraction (Fig. 5C, lanes 9-13); an additional band is detected with this fragment using the 0.3 M heparin-agarose fraction (Fig. 5C, lane 6).This band also is specifically competed for
5697
EGF Receptor Gene Promoter FIG. 4. Summary of exonuclease 111 protection and DNase I footprinting of the EGF receptor gene promoter region. Nucleotide sequence of a portion of the EGF receptor gene promoter region is shown with the position of nuclear protein binding sitesindicated. This figure summarizes 40 experiments typical of Figs. 2 and 3. Asterisks indicate the end of the region analyzed. The translational start site is located at nucleotide + l . Arrows indicate transcriptional start sites. The footprint for site 3 covers a broad region from -290 to -331 and contains two potential recognition sites. We have not precisely determinedthelocation of site 4 and show the most probable recognition site.
00000 0 3M BINDING SITES EXONUCLEASE Ill PROTECTION
-
0 6M BINDING SITES EXONUCLEASE 111 PROTECTION
Three other regions that may be involved in EGF receptor gene regulation were also identified. Deletion of the regions -911 to -850 and -771 to -569 consistently decreased EGF receptor expression by about 45-70%. Additional deletion of the region between -481 and -384 restored activity to the level of PERCAT-1 (-1109 to -16). This 1.1-kb portion of the promoter region appears to contain regions that are involved in positive and negative regulation of EGF receptor gene expression. The EGF receptor gene promoter contains five CCGCCC sequences (referred to as GC boxes, Fig. 1) which have been shown to bind the transcription factor Spl in otherpromoters DISCUSSION (18-27). Recentlv, a new consensus seauence for S D bindine ~ We have previously identified a 1-kilobase pair 5'-flanking ( G G ~ ~ c ~ G G G C ) A TAAT has been determined andthereare nufragment asthe promoter region of the EGFreceptor gene by TA primer extension, S1 nuclease mapping, and in vitro transcrip- merous potential Spl binding sites in the EGF receptor gene tion analysis (17). In the current study, we have examined promoter region (34, 35). The promoter also contains four the promoter region in more detail using deletion analysis, nearly identical pyrimidine sequences that conform to a exonuclease I11 protection and DNase I footprinting. TCCTCC motif (-366 to -354, -345 to -336, -322 to -313, Deletion Analysis-In the first approach, various portions and -303 to -290). This motif is found in some other proof the promoter regionwere deleted from the PERCAT-1 moters to be associated with S1 nuclease sensitivity (36, 37). plasmid insert and their resulting chloramphenicol acetylExonuclease 111Protection-To evaluate the sequences imtransferase activities determined by transfection into CV-1 portant for promoter activity, nuclear factor binding sites and KB cells. Several interesting regions were defined by this were investigated by incubating end-labeled DNA fragments analysis. Removal of sequences -167 to -105 with or without from the promoter region (-481 to -16) with nuclear extracts retention of upstream sequences -167 to -1109 or -167 to of HeLa or A431 cells and then digesting with exonuclease -771 reduced promoter activity (PERCAT-12, -13, and -15; Fig. 1). These data suggest that region -167 to -105 is 111. To enhance the exonuclease digestion, the amount of important for maximal transcription. Therefore, a plasmid, exonuclease I11 was increased 5-fold and incubation time PERCAT-14, was constructed which only contains sequences increased from 10 to 60 min. These conditions gave the -167 to -105. This plasmid has promoter activity that is maximum amount of digestion of theEGF receptor gene greater than half that of the PERCAT-1 plasmid containing promoter fragments. We have attempted to completely digest the promoter fragments with exonuclease I11 under differing a 1.1-kilobase pair promoter fragment (Fig. 1). Analysis of deletion mutants PERCAT-10, -14, and -15 by conditions without success. We have changed salt conditions, primer extension of RNAs isolated from CV1 cells transfected temperature, enzyme concentration,and nonspecific DNA with the plasmid constructs, showed that several transcripts concentration without totally digesting theEGF receptor C did not initiate within the EGF receptor promoter (data not promoter fragments. This maybe due to the high G shown). In all these constructs, the major in vivo transcrip- content of these fragments which may form secondary structional start site located at -255 was removed. (In Fig. 1 an tures that are resistant to exonuclease I11 digestion. The asterisk shows which plasmids gave these results.) When the amount of nuclear extract, nonspecific DNA,and exonuclease pSVOCAT vector is used in the primer extension assay, we I11 were titrated to give the results presented here. Increasing the time of digestion with exonuclease I11 or the amount of detect similar primer extended products beginning in the vector (data not shown). The pSVOCAT plasmid yields very exonuclease I11 to greater than 16,000units/ml did not change low chloramphenicol acetyltransferaseactivity indicating the digestion patterns shown in Figs. 2 and 3. We have also these aberrant startsites do not produce RNA that is trans- used this exonuclease assay to identify the Spl and TATA lated into protein. On this basis it seems unlikely that the binding regions of an SV40 promoter fragment. With SV40 aberrant start sites in plasmids PERCAT-10, -14, and -15 DNA, complete digestion of unprotected regions was observed yield significant chloramphenicol acetyltransferase activity. confirming that the structure of the EGF receptor gene prothis assay (lanes 7-9). These results arein general agreement with the results of the exonuclease I11 protection assays. DNase Z Footprinting with Spl-The results of the exonuclease I11 assay raise the possibility that the factor Spl can bind to the EGFreceptor gene promoter. To test thishypothesis directly, purified S p l was bound to an end-labeled 725base pair PstI-Hind11 promoter fragment and DNase I footprinting performed as described by Dynan et al. (18, 19). As shown in Fig. 6, S p l protects four regions of the promoter. This experiment shows directly that purified S p l can interact in vitro with the promoter region of the EGFreceptor gene.
-
+
5698
EGF Receptor Gene Promoter
.......
A
.. . . . . . .
B
0,
....
. .,... . .
,.
C
0.
I
”
1
2 3
G-++-
4 5 6
7 8
9 1 0 11 1 2 1 3
F ~ G5.. Gel retardation analysis of the EGF receptor promoter region. DNA fragments: A , -481 to -384; B, -384 to -178; and C, -150 to -16; were end-labeled and incubated with: lane 1 , no extract; lane 2, 25 pg of crude extract; lane 3, 25 pg of crude extract with 25-fold molar excess specific competitor; lane 4, 25 pg of crude extract with 50-fold molar excess specific competitor; lane 5.25 pg of crude extract with 100-fold molar excess nonspecific competitor, lane 6, 20pgof0.3 M heparin-agarose fraction; lane 7, 20pgof 0.3 M heparin-agarose fraction with 25-fold molar excess specific competitor; lane 8,20 pg of 0.3 M heparin-agarose fraction with 50-fold molar excess specific competitor; lane 9, 20 pg of heparin-agarose fraction with 100-fold molar excess nonspecific competitor; lane 10, 20 pg of 0.6 M heparin-agarose fraction; lane 11,20 pg of 0.6 M heparin-agarose fraction with 25-fold molar excess specific competitor; lane 12, 20 pg of 0.6 M heparin-agarose fraction with 50-fold molar excess specific competitor; and lane 13,20 pg of 0.6 M heparin-agarose fraction with 100-foldmolar excess nonspecific competitor. Samples were subjected to native gel electrophoresis as described under “Methods.”
moter is responsible for incomplete digestion shown in Figs. 2 and 3. Several criteria were used to distinguish specific nuclear binding. The sites summarized in Fig. 4adhered to these criteria. Addition of increasing amounts of nuclear extract, crude or fractionated, resulted inthe appearance of a number of new bands that increase proportionally to the added extract. Furthermore, when the extract is heat denatured prior to incubation, these bands do not appear (Fig. 2, lanes 9 and
12345 FIG.6. DNase I footprint analysis of EGF receptor promoter region using purified Spl. An 8%polyacrylamide sequencing gel was usedto analyze the partial DNase I digest of the 32P-endlabeled EGF receptor promoter fragment in the presence of purified Spl (90% pure). The PERCAT-1 plasmid was 5’ end-labeled a t the Hind111 linker site and digested with PstI. The 725-bp PstI-HindIIIlabeled fragment was isolated by gel electrophoresis and incubated with Spl followed by partial digestion with DNase I as described previously (7,8). Lune 1, G, refers to theguanine marker obtained by partial digestion with dimethyl sulfate of the same end-labeled fragment. Lunes 2 and 5 (-) show the DNase I cleavage pattern in the absence of Spl. Lunes 3 and 4 show the DNase I cleavage pattern in the presence of25 and 50 ng of Spl, respectively. The protected regions are indicated by brackets to theright of the footprint.
13).Also, if an excess of unlabeled EGF receptor gene promoter fragment is added to the extract,eight of these bands are no longer protected from digestion (Fig. 3). The eight bands are protected, however, when an excess of an unlabeled nonspecific pBR322 fragment (Fig. 3, lanes 5, 9, and 13) or 4x174 HaeIII fragments (data not shown) are used. These data indicate that these eight sites (-457 to -440, -417 to -402, -305 to -286, -214 to -200, -185 to -170, -139 to -126,-110 to -84, and -39 to -23) are binding specific nuclear factors from the extracts. Gel Retardation-The exonuclease I11 protection assays
EGF Receptor Gene Promoter indicate that eight specific nuclear factor binding sites are located in the -481 to -16 region of the EGF receptor gene promoter. To support this data, gel retardation assays were performed. The results inFig. 5 show that eight retarded bands are detected when fragments are used that span this region. The location of these bandsagree with the exonuclease I11 protection data. Furthermore, upon fractionation, the expected number of specific bands appear in the 0.3 and 0.6 M heparin-agarose fractions, two and six, respectively. These results fully support the data acquired with exonuclease I11 protection assays. Spl Binding Sites-The EGF receptor gene promoter region contains five GC boxes (17) andnumerous potential sites to which Spl might possibly bind (18-27). Furthermore, four of the eight protected sites mapped by exonuclease I11 protection contain potential Spl binding sites. To determine if Spl does bind to the EGF receptor gene promoter, a DNase I protection experiment was carried out using Spl that was approximately 90% pure. Four regions were identified that can bind Spl and one of these(site 4)may be two sites separated by a DNase I hypersensitive site (Fig. 4). Site 3 covers a broad region(-290 to -331) andcontains two possible recognition sites (Fig. 4). Three of the four sites do not include the original GC box, CCGCCC, but are located very near this sequence. These S p l binding sites vary in affinity with site 1 having the weakest affinity while site 4 appears to be the strongest. The DNA sequences to which S p l binds in this promoter is consistent with the consensus Spl binding sequence, GGGGCGGGGC(3435). TA A TAAT We have identified several nuclear factor binding sites in the EGF receptor gene promoter. In addition to four Spl binding sites, the major exonuclease I11 protected site (-139 to -126) was located in a fragment that was shown to be necessary for optimal EGF receptor gene promoter activity. Thus, thefactor binding to thissite may beimportant in EGF receptor regulation. The DNA sequence of this binding site or the sequences of the other non S p l binding sites do not match the binding site sequence of any factors identified to date. Experiments are underway to purify and characterize factors binding to these sites. Acknowledgments-We are indebted to Drs. James Kadonaga and Robert Tjian for Spl DNase I footprinting and helpful discussion. We are also grateful to Drs. Carl Wu, Pamela Marino, and Ryoichiro Kageyama for their useful suggestions and support in this effort. We would like to thank Althea Gaddis for editorial assistance and Ray Steinberg and Steve Neal for photography. REFERENCES 1. Cohen, S. (1962) J . Biol. Chem. 2 3 7 , 1555-1562 2. Savage, C. R., Jr., and Cohen, S. 0.(1972) J. Biol. Chem. 2 4 7 , 7609-7611 3. Schlessinger, J., Schreiber, A.B., Levi, I., Libermann, T., and Yarden, Y. (1983) Crit. Reu. Biochem. 1 4 , 93-111 4. Ushiro, H., and Cohen, S. (1980) J. Bid. Chern. 255,8363-8365 5. Sahyoun, N., Hock, R.A., and Hollenberg, M.D. (1978) Proc. Natl. Acad. Sci. U. S. A. 7 5 , 1675-1679 6. Das, M., Miyakawa, T. C., Fox, F., Pruss, R. M., Aharonov, A.,
and Herschman, H. R. (1977) Proc. Natl. Acad.Sci. U. S. A. 74,2790-2794 7. Yamamoto, T., Nishida, T., Miyajima, N., Kawai, S., Ooi, T., and Toyoshima, K. (1983) Cell 35,71-78 8. Downward, J., Yarden, Y., Mayes, E., Scrace, G., Totty, N., Stockwell, P., Ullrich, A., Schlessinger, J., and Waterfield, M. D. (1984) Nature 3 0 7 , 521-527 9. Xu, Y.-h., Richert, N., Ito, S., Merlino, G. T., and Pastan, I. (1984) Proc. Natl. Acad. Sci. U. S. A. 8 1 , 7308-7312 10. 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 11. 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 3 0 9 , 418-425 12. 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 13. 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. 4 6 , 414-416 14. Merlino, G . T., Xu, Y.-h., Richert, N., Clark, A. J. L., Ishii, S., Banks-Schlegel, S., and Pastan, I. (1985) J. Clin.Znuest. 7 5 , 1077-1079 15. Libermann, T. A., Nusbaum, H. R., Razon, N., Kris, R., Lax, I., Soreq, H., Whittle, N., Waterfield, M. D., Ullrich, A,, and Schlessinger, J. (1985) Nature 3 1 3 , 144-147 16. King, C.R., Kraus, M.H., Williams, L. T., Merlino, G. T., Pastan, I., 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. 8 2 , 49204924 18. Dynan, W. S., and Tjian, R. (1983) Cell 3 2 , 669-680 19. Dynan, W. S., and Tjian, R. (1983) Cell 3 5 , 79-87 20. Gidoni, D., Dynan, W. S., and Tjian, R. (1984) Nature 312,409413 21 Gidoni, D., Kadonaga, J. T., Berrera-Saldana, H., Takahashi, K., Chambon, P., and Tjian, R. (1985) Science 230,511-517 22 Jones, K. A., Kadonaga, J. T., Luciw, P. A., and Tjian, R. (1986) Science 2 3 2 , 755-759 23. Jones, K. A., Yamamoto, K.R., and Tjian, R. (1985) Cell 4 2 , 559-572 24. Jones, K. A., and Tjian, R. (1985) Nature 3 1 7 , 179-182 25. Ishii, S., Kadonaga, J. T., Tjian, R., Brady, J., Merlino, G. T., and Pastan, I. (1986) Science 2 3 2 , 1410-1413 26. Dynan, W. S., Saffer, S., Lee, W. S., and Tjian, R. (1985) Proc. Natl. Acad. Sci. U. S. A. 8 2 , 4915-4919 27. Dynan, W. S., Sazer, S., Tjian, R., and Schmike, R. T. (1986) Nature 3 1 9 , 246-248 28. Gorman, C. M., Moffat, L.F., and Howard, B. H. (1982) Mol. Cell. Biol. 2 , 1044-1051 29. Graham, F., and van der Eb, A. (1973) Virology 52,456-457 30. Parker, B. A., and Stark, G . R. (1979) J. Virology 3 1 , 360-369 31. Gorman, C.M., Merlino, G. T., Willingham, M.C., Pastan, I., and Howard, B. H. (1982) Proc. Nutl. Acad. Sci. U. S. A. 7 9 , 6777-6781 32. Wu, C. (1984) Nature 309,229-234 33. Wu, C. (1985) Nature 317,84-87 34. Briggs, M. R., Kadonaga, J. T., Bell, S. P., and Tjian, R. (1986) Science 234,47-52 35. Kadonaga, J. T., andTjian, R. (1986) Proc. Nutl. Acad.Sci. U. 5'. A. 8 3 , 5889-5893 36. McKeon, C., Schmidt, A., and de Crombrugghe, B. (1984) J . Bid. Chem. 2 5 9 , 6636-6640 37. Yu, Y.-T., and Manley, J . L. (1986) Cell 45, 743-751
