Transcriptional activation of the immediate early gene pip92 by serum ...

Vol. 269,No. 37, Issue of September 16, pp. 23163-23170,1994 Printed in U.S.A.

THE JOWAL OF BIOLOGICAL CHEMIST~Y 0 1994 by The American Society for Bioehemistry and Molecular Biology, Inc.

Transcriptional Activation of the Immediate Early Gene pip92 by Serum GrowthFactors Requires Both Ets and CArG-like Elements* (Received for publication, May 25, 1994)

Branko V. Latinki6 and LesterF. LauS From the Departmentof Genetics, University of Illinois College of Medicine, Chicago, Illinois 60612

The most extensive analysis of the mechanism of transcrippip92 is an immediate early gene that is transcriptionally activated in mouse 3T3 fibroblasts upon treatment tional activation of immediate early genes has focused on the with serum growth factors or thetumor promoter 12-0- c-fos protooncogene. The response of c-fos to serum, purified tetradecanoylphorbol-13-acetate.Here we show that a growth factors, and the protein kinase C activator TPA' is 73-base pair pip92 promoter fragment located between mediated through a short sequence known as the SRE (4,5). base pairs 1231 and 1158 upstream of the transcription The inner core of the SRE is a sequence known as the CArG start site is sufficient to mediate transcriptional activa- box, characterized by the consensus sequence CC(A/T),GG, tion. This promoter fragment contains two binding sites which interacts with a 67-kDa protein, SRF (6). Binding of SRF for transcription factors of the Ets family and a low af- t o the SRE recruitsa 62-kDa protein, ~ 6 (where 2 TCF ~ ~ finity binding site for the serum response factor. The stands for ternary complex factor), to an Ets-like binding site minimal sequence that mediates serum induction in- (CAGGAT) that lies immediately upstream of the CArG box, cludes at least one copy of the Ets-binding site and a low thus resulting in a ternary complex (7). The S R F . S R E . P ~ ~ ~ ' ~ affinity binding site for the serum response factor. This ternary complex is necessary for induction by serum (7) and sequence can interact with at least two proteins in a (8) of c-fos in 3T3 fibroblasts. Recent experiments indicate TPA fibroblast nuclear extract that have binding character6 might 2 be ~Elkl ~(9) or SAP-1(101, both istics of an Ets family protein and a serum response that theidentity of ~ factor-like protein. These proteins can bind the pip92- of which belongto the Etsfamily of transcription factors. Neiinducible element simultaneously, thus forming a ter- ther Elklnor SAP-l can bind the c-fos SRE alone, but requires nary complex. Furthermore, the same elementinteracts the prior interaction of SRF with the SRE to form a ternary with recombinant serum response factor and Elkl pro- complex. This ternary complex can form in vitro with the Etsteins individually as well as simultaneously to form a like binding site being present in either orientation and at ternary complex. This modeof ternary complex forma- varying distances from the CArG box (11). tion is in contrastto the one seen in thepromoter of the Footprinting analyses indicated that the occupancy of the c-fos protooncogene, where formation of the ternary c-fos SRE by protein factors is not altered by growth factor complex is dependent on the prior assembly of the se- action in vivo(12). Moreover,the S R F . S R E . P ternary ~ ~ ~ ~ comrum response factor-DNA binary complex. The activa- plex can be detected in extracts from quiescent as well as tion of a number of immediate early genes appears to be stimulated cells in vitro (13, 14). Thus, transcriptional activamediated through a ternary complex involving memberstion upon growth factor treatment is most likely mediated of the Ets family of transcription factors and the serum through post-translational modification, possibly through response factor. We propose that different mechanisms phosphorylation of ~ 6 2 ~ ~ / Eby l kthe l mitogen-activated procan lead to the formation of such ternary complexes. ~ the ' ~ tein kinases (13). These observations point t o ~ 6 2 as primary target of regulation in mitogen-induced transcription of c-fos. In addition, phosphorylation of SRF might also play a Many extracellular signaling agents, including growth fac- role in the modulation of ternary complex formation (15, 16). Can the mechanism that activates c-fos transcription be tors that mediate growth and differentiation, initiate the activation of a genomic program in target cells. This response in- generalized for other immediate early genes? Several immedivolves the rapid and transient activation of a set of immediate ate early genes appear to be regulated by growth factors early genes, including those that encode nuclear proteins that through mechanisms similar t o that regulating c-fos, involvfunction as transcription regulators, secreted proteins that ing SRE-SRF interaction and potentially ternary complex for' ~ . genes include zifz68 (171, krox20, function as cytokines, and cytoplasmic proteins that regulate mation with ~ 6 2 ~ These intracellular signaling (1-3). Understanding how these imme- and p-actin (18, 19). However, another group of immediate diate early genes are coordinately activated is an important early genes, including cyr61, JE,and c-jun, do not seem t o be step toward understanding how extracellular signals regulate regulated by the same mechanism (20-22). Their response elements lack either an Ets site, as in cyr61, or both the CArG cellular functions, including growth and differentiation. box and the Ets site, as inJE and c-jun. To extend current in* This work was supported by National Institutes of Health Grant formation on the regulation of immediate early genes, we R01-CA52220. The costs of publication of this articlewere defrayed in have studied the transcriptional activation of p i p 9 2 in NIH part by the payment of page charges. This article must therefore be 3T3 fibroblasts. hereby marked "aduertisement"in accordance with 18 U.S.C. Section p i p 9 2 ( c h x l ) was cloned independently from serum-stimu1734 solely to indicate this fact. lated Balblc 3T3 fibroblasts (23) andfrom activated T-lymphoThe nucleotide seguence(s) reported in this paper hasbeen submitted to the GenBankmIEMBL DataBank with accession number(s) L26490. cytes treated with cycloheximide (24). Its humanhomolog was $ Established Investigator of the American Heart Association. To whom correspondence should be addressed: Dept. of Genetics, UniverThe abbreviations used are: TPA, 12-O-tetradecanoylphorbol-13sity of Illinois College of Medicine, 808 South Wood St., Chicago, IL acetate;SRE, serum response element; SRF,serum response factor;kb, 60612. Tel.: 312-996-6978; Fa:312-996-7034. kilobase(s); bp, base pair(s).

23163

23164

~ u n s c ~ i p t ~ oActivation nul of pip92 by Serum Growth Factors

recently cloned from the myeloid leukemia cell line HL-60 induced to differentiate by "PA (25). pip92 is rapidly and transiently induced by stimulation with serum growth factors in fibroblasts and by treatment with nerve growth factor or upon membrane depolarization in PC12 pheochromocytoma cells (23, 26, 27). Transcription of pip92 is activated within 2 min of serum stimulation, reaches a peak level by 10 min, and is repressed to a low level by 1 h after stimulation (26). pip92 encodes a short-lived, proline-rich cytoplasmic protein that does not share significant sequence similarity with any known protein (23). Here we show that theserum response ofpip92 is mediated through a composite response element that consists of binding sites for Ets- and SRF-like factors. The sequences of these binding sites are similar to, but distinct from, those found in the c-fos SRE. This pip92 response element can form a ternary complex with recombinant SRF and Elkl proteins in vitro and forms a ternary complex with proteins in fibroblast nuclear extracts that have binding characteristics of SRF- and Ets-like factors. Unlike the c-fos SRE, the pip92 response element can interact independently with recombinant Elkl protein without the prior formation of the SRFaDNA binary complex. E ~ E ~ PROCEDURES ~ N T ~ Cell Culture and 72-ansfection.s-NIH 3T3 cells were grown in Dulbecm's modified Eagle's medium supplemented with 10% calf serum (Hyclone Laboratories), 100 pg/ml streptomycin, 100 unitdml penicillin, and 2 m~ glutamine (Life Technologies, Inc.). Transfections were performed by the calcium phosphate precipitation method (28). To prepare stably transfected cell lines, 20 pg of test plasmid was cotransfected with 2 pg of pSVneo(29) per 100-mm dish (5 x lo5cells), Oneday after transfection, cells were split into selective medium containing 400 pglml G418 (Life Technologies, Inc.). Well defined colonies, which appear after 10-14 days, were pooled by trypsinization. Typically, several thousand independent colonies were pooled for each transfection to eliminate position effects. Pooled transfectants were further expanded in selective medium. Upon reaching confluence, cellswere switched to medium with 0.5% calf serum for 2 days and stimulated with the following agents: serum (20%) for 30 min, TPA (100 ng/ml) for 30 min, forskolin (25 PM) for 30 min, platelet-derived growth factor (50 ng/ml; Collaborative Research) for 30 min, cycloheximide (10 pg/ml) for 2 h, and serum (20%)in the presence of 10 pg/ml cycloheximide for2 h. Two independent pools (from twotransfections) of -1761/cat, -1270/cat and -1231/-1158 tk cat transfectants were analyzed. In every case, results obtained from the two pools were indistinguishable. Transient transfections were analyzed by chloramphenicol acetyltransferase assays as described (28) using cell lysates normalized for transfection efficiency by P-galactosidase assays based on enzyme expressed from a cotransfected control plasmid (20). For each construct, three t o five sets of transfections were analyzed. Isolation of pip92 Gene and Construction of Test PZasmids--All recombinant DNA manipulations were performed by conventional techniques (28). A Balb/cmousegenomic DNA library constructed in EMBL4 A phage from EcoRI-digestedDNA (size >12 kb;a generous gift of A. Lanahan and D. Nathans, Johns Hopkins University School of Medicine) was screened with the pip92 cDNA as a probe. Two independently isolated clones contain -10 kb of 5'-noncoding regionin addition to the coding regionand 3 kb of 3'-noncoding region as determined by physical mapping.A 3.5-kb SphI-Hind111fragment ofpip92 (see Fig. 1) was subcloned into pBluescript SK, and both strands were sequenced with Sequenase (U. S. Biochemical Corp.) by the chain termination method 130). To make pip92cat fusion constructs, a 1.8-kbSp~I-Hind111fragment (see Fig. 1)was cloned into pCAT Basic (Promega) and used to derive 5'-deletion constructs by restriction enzyme digestions: Bsu361for -1270/cat, NueI for -1lllicat, BglII for -6801cat, SmuI for -287/cat, Af7II for -179/cat, andNheI for -8Wcat. For tk cat fwion constructs, the following DNAfragments were cloned into the BarnHI site of the tk cat vector (31):Bsu361-Aut11 for the -12701-970 construct, ApaI-NurI for the -12311-1158 construct, and NueI-Aut11 for the -11111-970construct. Other plasmids were made by using synthetic double-stranded oligonucleotides:DP, 5'-gatCCGCCGGAAGCCCTCATCCGG~GTg;P, 5'-gatCCGCCGTCAGCCCTCATCCGGAAGTg; D, Ei'-gatCCGCCGGAAGCCCTCATCCGTCAGTg; C, 5'-gatcTTCTTATATGGGCATg; DCP, 5'-

~tCCGCCGGAAGCGTTCTTATATGGGCATCC~AAGTg; and CP, 5'-

gatCCGCCGTCAGCGTTCTTATATGGGCATCCGGAAGTg (lower-case letters represent nucleotides that are not pip92-derived). The structures of all constructs were verified after cloning by sequencing. The Cltk cat construct contains three copies of CArG-like sequences. All other test plasmids contain a single copy of test elements in the native orientation. The -12311-1158** construct was made by cloning oligonucleotides (differing from the -1231/-1158 sequence in a TG .-+ CA change in theCArG-like sequenceTCTTATATGG)into the tkcat vector (see Fig. 4C). RNA Isolation and Analysis-Total cellular RNAs were isolated as described (32). RNA (15-30 pg/sample) was analyzed by S1 nuclease protection assay (see below). For thepip92kat fusion transcripts shown in Fig. 3, we used a synthetic oligonucleotide with the following sequence as probe: 5'-GTCGACGGTATCGATAAGCTTACAGCAGCAGGTACTCTGCTAGACGTTCGGGCGAGGGCAGCAAGCCG-3'. The fusion transcript initiated at the major initiation site yields anS1 nuclease-protected product that is 10 nucleotides smaller than the probe. For pip92kat protection assays in Fig. 2 B , a single-stranded probe was prepared using -17Wcat as a template and the PEl primer (33). The 3'-end of this 333-nucleotide-long probe was generated by M I 1 restriction. The protection product of the correctly initiated transcript is 161 nucleotides long. Double-stranded probes for tk cat and neo RNA were prepared and used exactly as described (31). The probe for the endogenouspip92 mRNA was a 416-bp double-strandedAfllI-XhoIfragment 5'-end-labeled a t the XhoI site (bp 235 of cDNA). A 1-kb s2P-5'end-labeled ladder (Life Technologies, Inc.)was used as a size marker. S1 nuclease digestions and analysis on denaturing polyacrylamide gels were performed as described (281, except that the annealing temperature was 51 "C. E l e e ~ r o p h o r e t ~ ~ o ~Shift i Z i t yAssays-Nuclear extracts were prepared from NIH 3T3 cells (quiescent or stimulated with serum or TPA for 15 min) as described (34) with the following modifications:b d e r s A and C were supplemented with 1mM NaVO,, 5 mM NaF, and the protease inhibitors phenylmethylsulfonyl fluoride (0.5 mM) and leupeptin (0.5 pglml). BufferA contained, in addition, 20 mM ~-glycerophosphate and 10 mM p-nitrophenyl phosphate. SRF and Elkl were produced in vitro in a coupled transcription-translation system, as recommended by the manufacturer (Promega),using pT7AATG (35)and pT7(1"428) (11) as templates. These plasmids have the same translational control signals, so SRF and Elkl are translated with similar efficiency as determined by [35SlMetincorporation (data not shown). Binding reactions were performedin 60 mM salt (NaCl fornuclear extracts and KC1 for in vitro translated SRF and Elkl), 12 m~ HEPES, pH 7.9, 12% glycerol, 0.1% Nonidet P-40, 3 mM spermidine, 2 pg of Sau3AI-digested pBluescript (Stratagene) as a nonspecific competitor, and 0.2 ng of end-labeled C P probe (4-5 x 10,000 cpm). Reactions wereincubated at 25 "C for 15 min and analyzed on 4% acrylamide (40:l acrylamide/ bisacrylamide)gels run in0.5 x TBE buffer (28). Competitionsincluded 20- and 200-fold molar excesses of unlabeled double-stranded oligonucleotides. The sequence of cyr61 CArG box oligonucleotideshas been described (20). RESULTS

Structure of pip92 Gene-We isolated a 15-kb clone from a Balb/c mouse genomic library that contains the pip92 structural gene and 10 kb of 5'-flanking DNA. The genomic sequence ofpip92 including 1761 bp upstream of the primary transcription start site (23) was determined and is shown in Fig. 1. The pip92 gene is intronless. Several potential binding sites for transcription regulators in the 5'4anking region were noted: two high affinity Ets sites (361, an SP1 site (371, and a CAMP response element-binding protein site (38) (Fig.1).Although a CArG box was not found, two sequences that deviate from the CC(APT),GG consensus sequence each by a 1-bp change were present at bp -1136 to -1127 and bp -1187 to -1178. Such CArG-like sequences have low affinities for SRF (39). I n ~ ~ c i ~ xL pe r e s ~ M ~ oend ~ ~ t~ ~e dr o pip92 ~ g ~Promoter-As determined by S1 nuclease protection assay of the pip92 mRNA, induction of pip92 in 3T3 fibroblasts was strongest upon treatment of cells with serum or TPA, weaker induction was seen upon treatment with platelet-derived growth factor or forskolin, an agent that elevates intracellular c M P levels

23165

Danscriptional Activation of pip92 by Serum Growth Factors -1761

GCATGCAAGCAGGTAAAATACAAAAACAAAAAAAGGAGAAA

-1720

AGAAGGAGAAAGAAAAGACACTCAGAAGCAGTGCAGGACAGCTGAAACTATAAAGGTGATATCTTGTTTC~C~

-1634

AAAACAAAAAACAAAAATTAAT~GGCCAGCAAGGTGGCTCAGCAGGTAAAGGGACGTGCCGCCAACACTGAGTACACGGGTAG

-1548

TTTCCCCAGGAAGACGGGAAGAAAGAACAGCCCAAGAACTGAGTACTTTGTCCTCTACATATGTGAGTGCAGGAGGGTTTCGAAGA

-1462

TGTTTTTCCGTAACTCCACAGGGTGCGAATCTCACTCCACCCGGACTCATCCCGCATGACGGTGCCGCAGCGTGTGAGAGCATGAG

-1376

-1204

AGAAGGAAACATCCAGTTCTCTCCGGGACGCCGTGCAGACTGTCCAAAGAACTGTAGCACCAAACAGACTCGAACGACCCGAGACA Ets-D ACACTTAGTTAGGCTCCTGAGGGCTGGATTGGAGGGGCGTGCTCGCGATGGAAGGGGCCCCGCCGGMGCCCTCTCCTCGGGCTAA CArG-like CArG-like Ets-P GACCGGCTGCCCTTGTTCTTATATGGGCATCCGGAAGTGG~TGCGGCGCCAACCTACGTCATTACATCC~TTTAGCCAGGGAA

-1118

GTCGGCCGGCTGCAGTGGCGGGAAGTCTGCCCGCTGGCAGGGGGTGTTTCTTAAAGGGCCTGCACAGCAAATGAAAGGGGTTATTT

-1032

GAATTTTTCTCCCTTCTGCAACTAGAAGCAAG~TTTCTTGGGAAGTGATGCAAAGTGGAACCGACGTCTTTTTTTCTTTTCTTTG

-1290

-946

TTACTGTGTGTGTGTGAGACATCGTATAGAAGTCTGTTCTCTTGTACCGCGTGGGACCGGTGGGACCGGGAATCGAACTGAGTTCG

-860

TTAGGCTTGGCAACAAGTGGCTCCACCCACTGTGCCATCTCGACCCAGATATCTGTCTTTCTTGGAGTAATTCCGTTCACGAGGGC

-774 quence of pip92 gene. Boldface underlined sequences include two Ets protein- -688 binding sites, twoCArG-like elements, a n SP1 site, and a CAMPresponse element-602 bindingprotein(CREB)-bindingsite. - 516 Numbers at the leftreferto the first nucleotide of the lines, with the transcrip-430 tion start site marked as +l. The initiation methioninecodon is boldface and un- 344 derlined;thepolyadenylationsignalis shown in lower-case letters, andthe -258 poly(A) tail of the mRNA follows the un-172 derlined GGTGT sequence (23).

ATGAAGGCAGAAAACTGCTCTATTGACAAGGGAAGATCGCAGGCCCTATGGACATTGGGACTTCCATGCCACTGGGTTGCGG~

FIG.1. Complete nucleotide se-

-86 +1 87 173 259 345 431 517 603 689 775 861 947 1033 1119 1205 1291 1377 1463 1549 1635 1721

CTGGGAGATCTTTAACGCTGAAAGTAAAGAGCAGTCGTGAGGTGATGGATATCTGTGGTTACAGGGTACAGTGTCCGTGCCCACTT TCCACCCATGTCCAGAAAGGTGGAGATAGAAGGGCTGCAGGCTTTAGTTGATCCTTTCGCCATAGTATAATCATATAATCATTCGT TTAGTAAATATTGGTCAGGTGGCTGCCTATTTCTCACCTGGAAAACTGTATCTTCTATGGAGTGGAGGTGAGAGATGAGAGACATT GAATAAAAAGAGAAAGTTCAACACACTGTAGTAACTGGGTGTAGTAACTGCTTAATAAGTAGCAGCGTTTTTACTGTTC~T SP1 ATTAATTAGCAGGCTGGTAGGCAAGATGAATGCCCTCCCTCGGGGAATGTACCCCCCGGGTTTTCCCCACTGCAGTCAGATTA~

CCGACTATCCGTCTCGTTGGTGGGCCTCATCCTCTGGCCCGTTGGTGACTCTCTGAGAACCCAGAACTTAGTGTCTTAAGGTGCAG TGGGGAAGTTCTTCCTTGCCAGAAGTTAAGTCCTTTTTCCTTCGAGGCGTAAATTTATCCCTATTGCTTTAAGGACTCGCTAGCCA CREB CTGTGGAAAGCCCGGATGAGGCCCTCTGGTTGGGCGTGGTGTTGCAGTGACGTCACTCTCAG~TTAAAAGGCTCCGGCTTGCT GCCCTCGCCCGAACGTCTAGCAGAGTACCTGCTGCTGTAAGCTTGTCCTCTCGCCTGCACCGCCCGTCTTAACCCATTCTCGACTT AACTACTCTCGTCGAACAAGCATGGAAGTACAGAAAGAAGCGCAGCGCATCATGACTCTGTCGGTA~GAACATGTACCACTCTCG CATGCAGCGAGGTGGCTTGCGACTCCACCGGAGTCTGCAGCTATCCCTCGTTATGCGCAGCGCTCGACAGCTCTACCTCTCAGCCA AGGTAGAAGCCCACCAGCCCGAGTTCCCGCCATCCCGCAGGGCTCTTGACCCTCGCCTGCACCCGCCGCGGGAAGCCGAACTTGCA GTGGAAGTAGCGTCCCCCGAAGCCGTGCAGCCTCCGGAGCCCATGGATACGCAAGAGGAAGTGCTGCGAGTCCAGGAGACCCCTGC GCTCTGTGACCCGCCCCCCGCTAGAGTCAGCCGCAAGCGCCGGAGCAGCAGCGATTTGAGCGACAGTAGTGATGCCGGACTGGTAC CAAGCAAGAAGGCCCGTCTAGAAGAGGTGGAGGGGGAGGCGACGTCGGAGGTTCCCGATCGCCTGCAGCTTCCTCCGGCACAAAGC GAAGGTGCCTTCCCTAACCTCGCCCGCGTCCTCCAAAGGCGCTTCTCCAGTCTCCTGAACTGTGGACCCGCCGTGCCCCCGCCGAC GCCCCCCACGTGCGAGGCCAAGCCAGCCTGCCGCCCGGCCGACAATATGCTCAACGTGCTGGTGCGAGCTCTGGTGGCCTTCTGAG AGCTCTGGTGGCTTCTTTCGAGCGGCGCCACCGGAGCGGAGAACGCACACCCGAGGCGAAGGCCGGCGGGCGCCGTGAAGAAGAGC CGCGGCCCGAGCTGCCGAGAGGCCAGGGCAAGGACTGAGGAGCGAGGGGCGCGGGCGCCTTCTCCCAGACGTGCGTCCATAGGTGC TATTAAAGGACTGTCCCTTCCTTGGCTTGGACAAGGGACACCTAGATCTTGAATCTCAGGGTCGAACTCTCTAGGGGCCAGGCTGC CCTTTCAAGGCCGTTTCACTACCATTCGCGTTTCGGCCCCTACAAGTGGGCACGCTTGTGCAAGCGGTCAGAGTTGCGTCATGGGA CAGACGCGGGTGCTTCCTGTTGCCTTGCGTGGGTGTGGGGCCTGGGAGGAGGCCAGGGTGTGGACCCGCCCTAGGGACTGGGAAGT GACTTGAGTCACCTCGCCCCCACAGGCTGCTGTGGGTGAGCCTGAACTGAACCAATCAAATCTGCGCAGAGTTGAAGTGGCTGGAG ACCCCGGGACTGGTCAACCTAGATGATCGCCTGGCGTGGACCACCGCGGGACGGGTGGGCCGCTGGTCGTAGTTGCTGCCGTAGAC ACAGCTTCTTCGCGCAGGAAAG~TTTTTTTTTTACCAGCGTGTTTAAGAAAGTCTGTTTACTTTTCCCACGGTGCGTTGTTTA ATTAGCAACTACCTGGAGTTTTACAATGTCAGCTAGGAAaataaaGACCATC~AAAACGCGCGGCTTGTCTGGCTTGCTTCC CTGGTGGGGGAGGGGTGGGGTTAAGGGGTCAGCCTGCTCCCTAAGCTCCGGAACTGCTTTGACTGGGATGCGGAGCCAGAGGGAGG CCGGGCCCATTTCTTTGCCACAAGTTTACCTGCTTAAGCTAGCCCTGTAATTTCGGAGTGCCTTAGAAATTAACAAACTCTGCCGG AATAGTGTGTGGAAAGCTT

(Fig. 2 A ) (23).pip92 w a s also induced by cycloheximide and B -17611rat A was superinduced by serum in the presenceof cycloheximide (Fig. 2 A ) . Q S T P F P I a s T P ~ c s + c To dissect the sequences required for transcriptional activation of pip92, w e prepared constructs that contain various probe re-> gions of the pip92 promoter upstream of the cut gene (pip92kat constructs). These constructs were stably transfected NIH into pip 92 -> 3T3 cells, and pools of several thousand clones were selected and analyzed. Transcription initiated from the pip92kat conC - 92/Cal structs w a s detected by S 1 nuclease protection assay.As a control forRNA recovery and copy numbers, mRNA derived from the cotransfected neomycin resistance gene (neo) w a s measured in parallel S 1 nuclease protection reactions. W h e n 1761 bp of 5"flankingsequence of the pip92 promoter FIG.2. Correct regulation by pip92 promoter. A, endogenous w a s linked to the cut gene (-1761/cat), strong activation w a s pip92 mRNA was detected by S1 nuclease protection assay with a detected upon stimulation with serum and TPA (Fig. 2B). 416-bp double-stranded probe. The protected product is230 nucleotides Whereas weak induction was seen with platelet-derivedgrowth long. B, pip92/cat mRNA from the -1761/cat stable line was analyzed 2 B ) , this constructwas with a single-stranded probe. The protected product is 161nucleotides factor or forskolin stimulation (Fig. long. p . probe; t , tRNA control;Q, quiescent; S, serum (30m i d ; T,TPA inducible by cycloheximide treatment alone (data not shown). (30 min); P,platelet-derived growth factor (30 min); F , forskolin (30 These results indicate that the 1761-bp pip92 promoter frag- min); C, cycloheximide (2h); S+C, serum in thepresence of cycloheximimics mide (2 h). The pip92 probe and protected products are indicated by ment can mediate inducibility in a manner that closely arrows. the endogenous gene. "

23166

Danscriptional Activation

FIG.3. 5"Deletion analysis of pip92 promoter. RNAs from stable lines carrying the indicated test plasmids were subjected to S1 nuclease protection analysis with a 68-nucleotide probe for pip92Icat mRNA and with a double-stranded probe for neo mRNA as a control. The 92/cat probe, 92hat-protected product (58nucleotides), and neo-protected products are indicated by arrows. Lanes with samples from cells treated with different agents are labeled a s described in the legend to Fig. 2. We have consistently observed a slight increase inneo mRNA levels in the samples treated withcycloheximide (31).

of pip92 by Serum Growth

Factors

4-

probe

4-

92cat

4-

- Neo

We carried out a 5'-deletion analysis of the pip92 promoter scription factors. Two high affinity binding sites for the Ets by removing successively more upstream sequences from the family of proteins were noted at bp -1175 t o -1168 and bp -1761/cat construct. The numericdesignation of each construct -1229 to -1212; this fragment also included a CArG-like sereferstothe nucleotide to which the 5"deletion extends. quence (Fig. 4C). To test thepossible roles of these Ets-binding promoter fusions Whereas -1270lcat retained responsiveness to serum, TPA, sites inpip92 induction, we constructed tk cat and cycloheximide, -llll/cat had largely lost responsivenessto with synthetic oligonucleotides bearing either the pip92 proserum andTPA, rendering i t inducible by only 2-3-fold (Fig. 3). moter-proximal (P/tk cat) or promoter-distal(D/tk cat) Etsbinding site or both (DP/tk cat)(Fig. 5 0 . These constructswere Similar results were obtained with the -680/cat and -287/cat constructs (Fig. 3 and data not shown). Deletion to bp -179 and virtually uninducible by serum or TPA (Fig. 5). These results -89 retained only low basal activity, with virtually no induc- indicate that the two Ets-binding sites within bp -1231 to -1158 are by themselves unable to mediate strong serum inibility by any agent tested(Fig. 3). The -llll/cat and-287/cat constructs retainedmost of the superinduction by serum in the duction; additional sequences within this fragment are likely to presence of cycloheximide (Fig. 3). These data suggest that thebe required (seebelow). Serum andTPA Response ofpip92 Requires a t Least One Ets primary sequence responsible for serum induction of pip92 is located between bp -1270 and -1111 and that an additional Site and CArG-like Sequence-Immediately upstream of the a CArG-like sequence, TCTweakly inducibleelement islocated within the-287/-179 frag- proximal Ets-bindingsiteis ment. In addition, these results suggest that elements that TATATGG.We tested if this sequence plays any role in the mediate superinduction by serum in thepresence of cyclohex- serum induction ofpip92 by fusing it either by itself or with the imide are separable from the sequences required for serum and pip92 Ets-binding sites upstreamof the tk cat basalpromoter TPA inducibility. (Fig. 6B). These testplasmids weretransiently transfected into 73-bp Sequence between bp -1231 a n d -1158 Is Sufficient to NIH 3T3 cells and analyzed by chloramphenicol acetyltransA with three copies of the pip92 Mediate Serum InductionSinceour 5"deletion analysis indi- ferase activity assays. construct cated that the-1270/-1111 region is important for serum and CArG-like sequence (C/tk cat) was virtually uninducible by TPA response, we tested its ability to confer inducibility on a serum or TPA (Fig. 6A). In contrast, a CArG-like box with heterologous promoter. We cloned pip92 promoter fragments either proximal (CP/tk cat)or both proximal and distal(DCP/tk upstream of the herpes simplex virus thymidine kinase gene cat) Ets-binding sites was strongly inducible by serum or TPA basal promoter driving the expression of the cat gene. The (Fig. 6A). The levels of induction of CP/tk cat or DCP/tk cat resulting constructs were analyzed by S1 nuclease assay of were similar to those of the -1231/-1158-bp DNA fragment RNAisolated from stably transfectedNIH 3T3 cells (Fig. 4 B ) or (Figs. 6A and 4 A , respectively). DCP/tk cat wasalso found to be by chloramphenicol acetyltransferase activity assays per- serum-responsive by S1 nuclease protection assay (data not formed on extracts from transiently transfectedcells (Fig. 4A). shown). These results suggest that the Ets andCArG-like sites are We observed a good agreement between results obtained by chloramphenicol acetyltransferase assays and those obtained sufficient for serum andTPAinducibility ofpip92.However, we cannot exclude the possibility that the remaining sequence by S1 nuclease protection assays of mRNA. Our study showed that pip92promoter fragments containing within the-1231/-1158 fragment plays a role in thisinducibileither bp -1270 to -970 (-1270/-970) or bp -1231 to -1158 ity. To test this possibility, we introduced two point mutations context of the -1231/-1158 construct (-1231/-1158) were capable of strong induction by serum or into CArG-like sites in the TPA as judged by both S1 nuclease andchloramphenicol acetyl- (see Fig. 4 0 . As shown in Fig. 6C, this construct (-12311 transferase activity assays (Fig. 4). In contrast, -1111/-970 tk -11589 was uninducible by either serum or TPA and had residual activity similar to that of the DP/tk cat construct. cat was notinducible by serum (Fig. 4B ). Both the -1270/-970 and -1231/-1158 constructs displayed inducibility by cyclohex- These results show that the pip92 serum and TPA response imide, but only the -1270/-970 construct could be superin- element consists of the CArG-like sequence and a t least one duced by serum in the presence of cycloheximide (Fig. 4B). Ets-binding site. The spacing of Ets and CArG-like sites apThese results demonstrate that the fragment between bp-1231 peared to be unimportant (Fig. 6A), in agreement with a preand -1158 (Fig. 4C) contains sequences that arenecessary and vious report (11). sufficient to mediate serum and TPA induction. These results Formation of Ternary Complex on pip92 Response Elementalso indicate thatsequences that mediatecycloheximide induc- The pip92 CArG-like sequence is not expected to bind SRF with ibility and superinduction are separable. high affinity. Indeed, we found that theaffinity of this sequence Ets Sites of pip92 Are Insufficient for Serum Inducibilityfor recombinant SRF synthesized by in vitro translation was Since the 73-bp DNA fragment from bp -1231 to -1158 of the -10-fold lower than that of the consensussequence CC(N pip92 promoter could mediate inducibility (Fig. 41, we exam- T),GG (data not shown), in agreement with previous reports ined this sequence for possible binding sites for known tran- (39). The low affinity of the pip92 CArG-like sequence may

of pip92 by Serum Growth Factors

Danscriptional Activation

A

-1111/

"-

2311-1158 -12701-970 -1231/-1158 -12701-970 Q S T

Q

S

23167

T

p

s

0

t

s+c 0

C

S

"Cs+cS C

T

rn

uc cat --5 *

e220

I

B

Ets D CArG-like Ets P -1231 CCGCCGGAAGCCCTCTCCTCGGGCTAAGACCGGCTGCCCTTGTTCTTATATGGGCATCCGGAAGTGGTTTGCGG -1158 -1231 CA confer serum and TPAresponsiveness to heterologous promoter. A, extracts from NIH 3T3 FIG.4.73-bppip92promoter fragment can cells transiently transfected with the -1270/-970 or -1231/-1158 tk catplasmid together with pPGKPGAL (20) were normalized by P-galactosidase assays and tested for chloramphenicol acetyltransferase activity by a thin-layer chromatography-based assay. Q, extract from quiescent cells; S, extract from cells treated with serum for 4 h; T,extract from TPA-treated cells (4 h). B, RNA from NIH 3T3 cells stably transfected with the -12701-970, -123V-1158, or -1111/-970 construct was subjected to S1 nuclease analysis with double-stranded tk cat and neo probes. Lanes are labeled as described in the legend to Fig. 2. m , DNA size marker (1-kb ladder). tk cat andneo protections are indicated by urrows. C , shown is the sequence of the -1231/-1158 fragment. Ets sitesand the CArG-like box are underlined. The nucleotide changes in the-1231/-1158** construct are indicated. A

Dltkcat Q

S

T

FIG.5. Ets sites are insufficient for seruminducibility. A, extracts from NIH 3T3 cells transiently transfected with D/tk cat, P/tk cat, or-DPItk cat together with pPGKpGAL were normalized by P-galactosidaseand assays analyzed by chloramphenicol acetyltransferase assays. B, RNAs from stablelines transfected with P/tk cat or DP/tk cat were analyzed by S1 nuclease protection assays with tWcat and neo probes. Lanes are labeled as described in thelegend to Fig. 2. C , shown is a schematic presentation of the test plasmids used; all have 25-bp inserts and are identical except for the designated mutation, made by a 2-bp change from the wild-type sequence (see "Experimental Procedures").

B

Pltkcat Q

-

S

T

DPltkcat Q

S

T

Pttk cat Q

f! c-tkcat

:!

;,

m SRF --- > Elkl --->

FIG.7. pip92 CArG-like and Ets sitessupport formationof ternary complex. A, nuclear extracts from quiescent NIH 3T3 cells (lane 1 )or from cells stimulated by 20% serum for 15 min (lanes2-10) or in vitro translated SRF (lane 11) was incubated with end-labeled CP oligonucleotides, and the resultingprotein-DNA complexes were resolved by nondenaturing polyacrylamide gel electrophoresis. Unlabeled oligonucleotides used ascompetitors are indicated above the lanes.Lunes 3, 5, 7, and 9 contain a 20-fold molar excess, and lanes 4, 6 , 8, and 10

23 169

the CArG-like box.We expect that the ternary complex may form at thepip92 Ets andCArG-like sites from either Elkl-or SRF-like binary complexes. This expectation is supported by our results with Elkl and SRF proteins synthesized i n vitro (Fig. 7 B ) . Unlike c-fos, pip92 ternary complex assembly does not seemto have the strict requirement for the prior formation of an SRF.DNA binary complex. It would be interesting to see whether a ternary complex may form by the recruitmentof SRF by a p62TCF.DNAbinary complex. pip92 was independently cloned from activated T-lymphocytes treated with cycloheximide (24). In addition, we found pip92 mRNA to be strongly induced by cycloheximide in fibroblasts (Fig. 2 A ) . Our promoter analysis shows that cycloheximide-mediated inducibility is not confined to one region of the pip92 promoter, but could be mediated by a t least two separable elements (Fig. 3). This is not surprisingsince protein synthesis inhibitors, including cycloheximide and anisomycin, are known to have complex post-transcriptional effects as well as both positive and negative transcriptional effects on immediate early gene expression (42). An Ets site in thec-fos SRE hasbeen shown to be essential for TPA induction in Balb/c 3T3 cells (8).Other TPA response elements can be composed of tandem Ets sites or of Ets and AP-1 sites (43-45). We note that pip92 is induced strongly by TPA, in addition, the humanhomolog ofpip92 was cloned from TPA-treated HL-60 cells (23). The presence of two high affinity Ets sites in its promoter may explain this strong response to TPA. In contrast, the immediate early gene cyr61 is poorly inducible by TPA (46) and has no Ets site in thevicinity of its high affinity CArG box (20). Thus, the spectrum of signals to which an immediate early gene could respond may be determined by the natureof the otherfactors that interactwith SRF (8, 11). Another immediate early gene studied in our laboratory, nur77, a member of the steroidlthyroid receptor gene superfamily, is regulated inresponse to serumby an element similar to the one seen in pip92 (31). The nur77 serum response element contains a low affinity CArG-like box and a high affinity Ets site (31). In addition, it has been shown that a ternary complex can form with a wide variety of Ets sites present in either orientation and at variable distance from the CArG box-SRF binary complex (11).Taken together, these results suggest that a subset of immediate early genes may be regulated through Ets and CArG sites with varying affinity for their cognate transcription factors. One potential biological role of these varying sites could be in providing differential transcriptional regulation indiverse cell types. For example, composite response elements thatconsist o f high affinity Ets sites anda low affinity CArG box (like those of pip92 and nur77) may function as serum response elements in cells that are rich in both SRF- and Ets-like factors, whereas in cells poor in SRF and rich in Ets family members, they may function as SRFindependent enhancers. This hypothesis is testable in future work addressing thefunction of various EtdCArG elements in gene activation in diverse cell types. Acknowledgments-We thank Drs. A. Lanahan and D. Nathansfor a Balb/c genomic library, Dr. A. Nordheim for gifts of antibodies, and Dr. R. ”reisman for SRF and Elkl plasmids. We also thank Marija Zeremski for help with electrophoretic mobility shift assays and Steve Hayes for technical assistance.

II, contain a 200-fold molar excess. Specific complexes are labeled I, and III in the decreasing order of mobility. B, in vitro translated SRF and Elkl (1 pl of translation productisample) were incubated with the CP probe. The type of DNA template present in the reticulocyte lysate is indicated above the lanes.

23170

Danscriptional Activation of pip92 by Serum Growth Factors REFERENCES

1. Lau, L. F., and Nathans, D. (1991) Mol. Aspects Cell. Regul. 6, 257-293 2. Herschman, H. R. (1991)Annu. Reu. Biochem. 80,281319 3. Sun, H., Charles, C. H., Lau, L. F., and Tonks, N. K. (1993) Cell 75, 487-493 4. Treisman, R. (1990) Semin. Cancer Biol. 1,47-58 5. Treisman, R. (1992) D a d s Biochem. Sci. 17,423-426 6. Treisman, R. (1987) EMBO J. 6,2711-2717 7. Shaw, P. E., Schroter, H., and Nordheim, A. (1989) Cell 56, 563-572 8. Graham, R., and Gilman, M. (1991) Science 261,18%192 9. Hipskind, R. A., Rao, V. N., Mueller, C. G., Reddy, E. S., and Nordheim, A. (1991) Nature 354,531-534 10. Dalton, S., and Treisman, R. (1992) Cell 88, 597-612 11. Treisman, R., Marais, R., and Wynne, J. (1992)EMBO J. 11,46314640 12. Herrera, R. E.,Shaw, P. E., and Nordheim, A. (1989) Nature 340,68-70 13. Marais, R., Wynne, J., and Treisman, R. (1993) Cell 73, 381-393 14. Zinck, R., Hipskind, R. A,, Pingoud, V., and Nordheim, A. (1993)EMBO J. 12, 2377-2387 15. Rivera, V R., Miranti, C. K., Misra, R. P., Ginty, D. D., Chen, R. H., Blenis, J., and Greenberg, M. E. (1993) Mol. Cell. B i d . 13,626M273 16. Liu,S. H., Ma, J. T., Yueh, A. Y., Lees-Miller, S. P., Anderson, C. W., and Ng, S. Y. (1993)J. Biol. Chem. 268,21147-21154 17. Christy, B., and Nathans, D. (1989) Mol. Cell. Biol. 9, 4889-4895 18. Chavrier, P., Janssen Timmen, U., Mattei, M. G., &rial, M., Bravo, R., and Charnay, P. (1989) Mol. Cell. B i d . 9, 787-797 19.Kawamoto,T., Makino, K., Orita, S., Nakata, A., and Kakunaga, T. (1989) Nucleic Acids Res. 17, 523-537 20. LatinkiC, B. V., O'Brien, T. P., and Lau, L. F. (1991) Nucleic Acids Res. 19, 3261-3267 21. Freter, R.R., Irminger, J.-C., Porter, J.A,, Jones, S. D., and Stiles, C. D. (1992) Mol. Cell. Biol. 12, 52884300 22. Han, T.-H., Lamph, W.W., and Prywes, R. (1992)Mol. Cell. Biol.12,4472-4477 T.P., and Lau. t.F. (1990) Mol. Cell. 23. Charles, C. H., Simske,. J. S... .O'Brien, . Biol. .lo, 6769-6774 24. Coleclough, C., Kuhn, L., and Lekovits,I. (199O)Proc.Natl. Acad.Sei. U.S. A.

87, 1753-1757 25. Shimizu, N., Ohta, M., Fujiwara, C., Sagara, J., Mochizuki, N., Oda, T., and Utiyama, H. (1991) J. Biol. Chem. 286,12157-12161 26. Lau, L. F.,and Nathans,D. (1987)Proc.Natl. Acad. Sei.U. S. A. 84,1182-1186 27. Bartel, D.,.'F Sheng, M., Lau, L. F., and Greenberg, M. E. (1989) Genes & Deu. 3,304313 28. Sambrook, J., Fritsch, E. F.,and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual,2nd Ed., Cold Spring Harbor, Cold Spring Harbor, NY 29. Southern, P. J., and Berg, P. (1982)J. Mol. Appl. Genet. 1,327-341 30. Sanger, F., Nicklen, S.,and Coulson, A. R. (1977)Proc. Natl. Acad.Sci. U.S. A. 74,5463-5467 31. Williams, G. T.,and Lau, L. F. (1993)Mol. Cell. B i d . 13, 6124-6136 32. Chomczynski, P., and Sacchi, N. (1987)Anal. Biochem. 162, 156159 33. Yoon, J. K., and Lau, L. F. (1993) J. Biol. Chem. 288,9148-9155 34. Andrews, N. C., and Faller, D. V. (1991) Nucleic Acids Res. 19,2499 35. Norman, C., Runswick, M., Pollock, R., and Treisman, R. (1988) Cell 55,9891003 36. Nye, J. A., Petersen, J. M., Gynther, C. V., Jonsen, M. D., and Graves, B. J. (1992) Genes & Deu. 6,975-990 37. Kadonaga, J. T., Carner, K. R., Masiarz, F. R., and Tjian, R. (1987) Cell 61, 1079-1090 38. Montminy, M. R., and Bilezikjian, L. M. (1987)Nature 328,175-178 39. Leung, S., and Miyamoto, N. G. (1989) Nucleic Acids Res.17,1177-1195 40. Almendral, J. M., Seamier, D., MacDonald-Bravo, H., Burckhardt, J., Perera, J., and Bravo, R. (1988) Mol. Cell. Biol. 8,2140-2148 41. Janknecht, R., and Nordheim, A. (1992) Nucleic Acids Res. 20,3317-3324 42. Edwards, D. R., and Mahadevan, L. C. (1992)EMBO J. 11,2415-2424 43. Wasylyk, C., Gutman, A., Nicholson, R., and Wasylyk, B. (1991) EMBO J. 10, 1127-1134 44. Wasylyk, B., Wasylyk, C.,Flores, P., Begue, A,, Leprince, D., and Stehelin, D. (1990) Nature 346, 191-193 45. Buttice, G., and Kurkinen, M. (1993) J. Biol. Chem. 268,7196-7204 46. O'Brien, T. P.,Yang, G. P., Sanders, L., and Lau, L. F. (1990)Mol. Cell. Biol. 10, 3569-3577