An Upstream NF-Y-binding Site Is Required for Transcriptional ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY

Vol. 269, No. 40, Issue of October 7, pp. 25042-25048, 1994 Printed in U.S.A.

@ 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

An Upstream NF-Y-binding Site Is Required for Transcriptional Activation from thehst Promoter in F9 Embryonal Carcinoma Cells* (Received for publication, April 5 , 1994, and in revised form, June 28, 1994)

Shahid Hasan*§, Toshiaki Koda*,and Mitsuaki Kakinumaa From the $Section of Bacterial Infection, Institute of Immunological Science a n d the §Section of Environmental Medicine, Graduate School of Environmental Science, Hokkaido University, Sapporo 060, J a p a n

Expression of hst (k-FGF, FGF-4), a member of the One such gene whose expression is modulated on differenfibroblastgrowthfactorgenefamily, is restricted to tiation of F 9 by treatment with retinoic acid/Bt,cAMP is hst early stages of developing embryos and to embryonal (also referred toas k-FGF or FGF-4). The hst gene was initially carcinoma cells. In F9, which is a prototype of embryo- discovered through transformation of NIW3T3 cells with hunal carcinomacells expressinghst, the expressionof hst man DNA from various sources (4-6). The Hst protein is a gene is positively regulated by a downstream octamer member of the fibroblast growth factor (FGF) family, which also motif that functions as an enhancer. We have investi- includes acidic FGF (FGFl), basic FGF(FGFZ), Int-2 (FGFS), gated, by chloramphenicol acetyltransferase (CAT) re- FGF5 (71, FGF6 (81, and KGF (9). Transcription of murine hst F9, the cis-actingregula- has been detected in the preimplantation and early postimporter fusion gene analysis in hst promoterregionthat toryelementwithinthe plantation embryos (10,11). In cultured cells, expressionof the interacts with this enhancer. Electrophoretic mobility hst gene is limited to undifferentiated embryonal carcinoma shift assay and methylation interference analysis of F9 with retinoic acid/ cells including F9. Differentiation showed that the hst promoter contains, in a segment Bt,cAMP results in the repression of hst expression (12, 13). termed Y, thesequence B’-CTGA!ITGGCA-3’, which The temporal expression of hst gene, therefore, allows us to closely resembles the consensus binding motif for the CCAAT-binding factor NF-Y. Deletions or mutations in study transcriptional modulation in molecular details. that an enhancer Previously, we (14)and others (15) reported this element substantially reduced expression of hstCAT constructs. The nuclear factor binding to Y the seg- element residing in the third exon is essential for positively ment of the hst promoter was indistinguishable from regulating transcription from the hst promoter in F9. SubseNF-Y, as inferred from interactions with specific anti- quently, the functional elementof this enhancer was delimited is known to NF-Ymonoclonalandpolyclonalantibodies. We con- to an octamer motif(16, 17). The octamer sequence clude that the expression of thehst gene in F9 is posi- bind in vitro to both Octl, a ubiquitous factor, and Oct3, a factor tively regulatedby the coordinated interaction betweenpresent in undifferentiated embryonal carcinoma cell lines and an NF-Y-bindingsite and an octamer motif. in early stages of mouse embryo (18). Although the octamer motif plays the roleof an enhancer forthe hst promoter in F9, to gain a better insight into the mechanisms of hst gene expression in F9, finer analysis of the hst promoter region beTranscriptional regulation is one of the crucial steps deter- comes essential. We report here the characterization of a funcmining which gene should be selectively switched on or off. tional NF-Y (CP-1, CBF) (19-21) binding element present in Considerable evidence implies that this is dictated by transthe hst promoter ina segment that we refer toas Y. acting factors that interact with cis-acting elements within EXPERIMENTALPROCEDURES transcription units andthat i t i s t h e combined action of such Plasmid Constructs and Oligonucleotides-DNA fragments used in that confers ona given sequence-specific DNA-binding proteins this studywere derived fromthe 5’-promoterregion of human hst gene. gene its particular pattern of transcriptional activity (1). A F9 is an embryonal carcinoma cell line resembling the inner 277-bpRmaIIApaI (WA, -172/105) fragment was cloned at the HindIII site in pBluescript KS+, utilizing a HindIII linker. From this cell masses of blastocysts both morphologically and antigeni- fragment, RmaIISmaI ( W S , -172/-98) was generated and subcloned cally (2). Differentiation of F9 can be induced by treatment with between HindIII and SmaI sites in pBluescript KS+. The oligonucleoretinoic acid and dibutyryl cyclic AMP (Bt,cAMP)’ resulting in morphologically altered parietal endoderm-like cells (3). Therefore, this system has allowed the analysisof the mechanism(s1 of gene regulation during the process of cell differentiation.

* 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. 1 To whom correspondence should be addressed: Section of Bacterial Infection, Institute of ImmunologicalScience,HokkaidoUniversity, Kita-15, Nishi-7, Kita-ku, Sapporo 060, Japan. Tel.:81-11-707-6835; Fax: 81-11-707-6835. The abbreviations used are: Bt,cAMP, dibutyryl cyclic AMP;bp, base pair(s); CAT, chloramphenicol acetyltransferase; EMSA, electrophoretic mobility shift assay; FGF, fibroblast growth factor; dF9, differentiated F9.

tides corresponding to a portion ofWA were synthesized on a DNA synthesizer (-1, model 380B). Sequences of their sense strands were as follows: Y oligomer corresponds to 5’-CCCCCGGCCCTGATTGGCAGGCGG-3’(-153/-130); Ym with mutation in CCAATbox, 5‘CCCCCGGCCCTtcgTGGCAGGCGG-3‘.The other oligonucleotideused inthis study was Ea oligomer, 5”AAACATTTTTCTGATTGGTTAAAAGTTGAGT-3’, present in Ea gene promoter (19). pY is a 71-bp BamHUSaZI fragment that contained Y oligomer sequence subclonedat the SmaI site of pBluescript Ks+. A mutant hst promoter construct was created by annealing mutagenizing oligonucleotides, Ym and antisense Ym, to WA cloned in pBluescript KS+. Polymerase chain reaction was performedwith one of the polylinker primers using either Tag (Promega) or Vent polymerase (New England Biolabs Inc.). The amplified products were annealed and reamplified (double polymerase chain reaction) to give a full-length product between two polylinker primers, digested with HindIII, and cloned in pBluescript KS+. Resulting plasmids were sequenced to confirm mutations, digested with HindIII, and finally cloned upstream of

25042

25043

A Functional NF-Ybinding Site in hstPromoter

Human -302 AGTGCAAGAGGCAAAACTGGCTGAAAAGCAGAAGTGTAGGAGCCGCCAAGG~GGGACG~CAGGTCCGTGGGCCGGGCGGAGCCAAG~T-GGGGGCCGGG

*

***

*

**

**

* *

************** ******* ** * ** ************* ** **** ***

Mouse - 2 1 3 GTCCCGTAAGGAAGGAGCACAGGAGATGCCCTGGGGAGCAGAGAGCCAAGGGGCGGGATC~CA~TTCGAGTGCGG~GGAGCC~GAGT~GGGGTTGGG

m aI

Y ~""~_"""_"_"""-

Human -200 GTCCCTCCAGGTGGCACTCG-CGGCGCTAGTCCCCAGCCTCCTCCCTTCCCCCGGCCCTGATTGGCAGGCGGCCTGCGACCAGCC~GAACGCCACAGCGC

*** ********* ** * * * *** ** *** ******

********* *****************

****

Mouse -163 GTCTCTCCAOGTGACAGTAGCCACCGCCA~CCGCGCCTCC-----TCCCCCGGC~TGATTG~AGGC~CT----------------------GCGC Sma I

TATA

Human -100 CCCGGGCGCCCAGGAGAACGCGAACGGCCCCCGCGGGAGCG~GAGTAGGAGG~GCGCCGGGCTATATATATAGCGGCTCGGCCTCGGGCGGGCCTG-1

***** * ******

*** * * * ********* ** . . . . . . . . . . . . . . . . . . . . . .

*********

* **** *

*

** *

Mouse - 8 9 CCCGG-CTCCCAGG------ CGACCGACGCCCCGCGGG-GAGGCGAGTAGGAGG~GCGCCGG-CTA~ATATACCACTGCTCCGGAGGGCTGGGCGC-1 FIG.1. Comparison of the human and mouse hst promoters. Nucleotide sequences of human and mouse hst promoters upstream of transcriptional initiation site (-1)are compared. Stars between the two sequences indicate conserved nucleotides.Dashes have been introduced to maximize homology. A TATA box-like sequenceis indicated with a bold line. The dashed line represents the segment Y inwhich an NF-Y-binding site was characterized in this study. Two restriction sites, RmaI and Smal, in the human hst promoter region are indicated above the sequence.

the CAT gene in pSVOOCAT as described previously (14). In some CAT constructs, a DNA fragment referred to as 53, which spans 3,548/3,717 of the hst gene and contains an octamer sequence, was inserted at the 3'-end of the CAT gene. RA-00,Ym-00, and -64-00 were CAT constructs without fragment 53. "53, Ym-53, and -64-53 werethose with fragment 53. -64-00 and -64-53 wereconstructed by the insertion of a polymerase chain reaction-generated 44/92 fragment instead ofFUA. 00-53 was constructed by the insertion of only fragment 53 downstream of the CAT gene in pSVOOCAT. Cell Culture and Preparation of Nuclear Extracts-F9 was cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and Jurkat in RPMI 1640, 10% fetal calf serum. For differentiation of F9, 3 x lo5 cells were seeded in 100-mm tissue culture plates coated with 0.1% gelatin. Next day the medium was replaced with fresh Dulbecco's modified Eagle's medium, 10% fetal calf serum M retinoic acid and 1l l l ~ Bt,cAMP and then cultured containing 2 x further for 3-4 days. Nuclear extracts were prepared accordingto Schreiber et al. (221, and protein concentrations were established by Bradford's method (23). Antibodies-Anti-NF-YAmonoclonalantibodyYAla, affinity-purified polyclonal anti-NF-YB, and anti-NF-YB rabbit serum were generously donated by Dr. Diane Mathis, Strasbourg, France, and have been described elsewhere (24). Electrophoretic Mobility Shift Assays (EMSAsi-Binding reactions were performedessentially as described by Pascal and Tjian (25).Typically, 5-7 pg of crude nuclear extracts were incubated in thepresence of 2 pg of poly(d1-dC)in afinal volume of 20 1.11.The mixture was incubated on ice for 15 min and then added with 0.2-0.5 ng of [y-32P]ATP5'-endlabeled and electrophoretically purified probes. Further incubation at room temperature was carried out for 30 min, and then the samples were electrophoresed on 4% native polyacrylamide gels at 4 "C in a circulating buffer containing 0.5 x TBE (1x TBE = 0.089 M Tris, 0.089 M boric acid, 0.02 m M EDTA, pH 8 ) and 0.05% (vh) Nonidet P-40. For supershift, antibody and nuclear extracts were added together. Unless stated otherwise, 200-fold molar excess of specific or nonspecific competitors was added prior to addition of probes. After electrophoresis, gels were dried and subjected to autoradiography. Methylation Interference Analysis-End-labeled DNA was partially methylated with dimethyl sulfate according to Maxam and Gilbert (26), incubated with the nuclear extracts, and then electrophoresed on a native polyacrylamide gel as describedabove.Bound and free DNA were eluted from the polyacrylamide gel, purified on DEAE-cellulose (DESl), and finally cleaved with 1M piperidine at 90 "C for 40 min prior to analysis on a sequencing gel. W Induced Cross-linking of Oligonucleotide-Protein ComplexesW induced cross-linking was performed essentially as described by Cooney et al. (27) except that [y-32PlATP5'-labeled Y oligomer probe was used. Protein-DNA complexes were cut out from the gel and treated with a solution containing 3% SDS, 0.375 M Tris-HCl (pH 6.8), and 50 mM dithiothreitol for 5 min at 37"Cbefore layering them on a 10% SDS-polyacrylamide gelalong with molecular weight markers. The gel was dried and subjected t o autoradiography. CATAssay-For transfection with chloramphenicol acetyltransferase constructs into F9, 1 x lo6 cells were seeded in tissue culture plates coated with 0.1% gelatin. Transfections were performed with 20 pg of respective reporter plasmids and 1pg of reference plasmid, pRSV-

Lac& by the standard calcium phosphate precipitation method (5). After 36-48 h,which included twowashes with phosphate-bufferedsaline (pH 7.2) followed by a medium change, cells wereharvested by scraping in a buffer containing 40 mM Tris-HC1 (pH 7.5), 10 mM EDTA, and 150 mM NaCl and thensubjected to five freeze-thaw cycles. CAT activity was determined as described previously (14) and normalized to p-galactosidase activity to eliminate differences in transfection efficiency. RESULTS DNA-binding Proteins Interact with the RrnaIlSrnaI (-1 72 1 -98) Region of the hst Promoter-Fig. 1 shows the promoter regions of human and mouse hst genes. Previously, we narrowed the promoter region of human hstto a 277-bp fragment, WA (-172/105) that, operating concurrently with an octamer element (ATGCAAAT, 3656/3663) residing in the third exon, w a s sufficient to confer the promoter activity in F9. Furthermore, we observed that deletion of t h e 5"region (-172/-65) of this promoter substantially decreased its strength suggesting that important regulatory sequences were confined withinthe deleted fragmentof the hst promoter (17). Since DNA-binding proteins may represent trans-acting factors that affect the transcription of the hst gene, we set out to search for proteins that interact withthe 5'-end of the region of interest. Utilizing the unique SmaI restriction site at position -98, a 75-bp fragment RmaIISmaI ( W S ;-172/-98) was isolated and used as a probe in EMSA. When the W S probe was incubated with nuclear extract from F9, a single protein-DNA complex, complex 11, was observed (Fig. 2). Nuclear extract fromdF9 yielded two additional complexes (I and 111). All these complexes were disrupted by the addition of a 200-fold molar excess of unlabeled W S (lanes 3 and 8) or an oligonucleotide Y (lanes 5 and l o ) , which was designed on the basis of t h e DNA contact points that were established for complexesI and I1 observed with theW S probe (see below). On the other hand, none of t h e complex was disrupted when a HinfUApaI (32/105) fragment derived from a neighboring region wasused as a nonspecific competitor (lanes 4 and 9) or with a mutant Y oligomer, Ym (lanes 6 and 11). CCAAT Box i n Reverse Is the Site for Protein-DNA Interaction i n R m a I l S m aFragment-Protein I binding sites in t h e WS probe in t h e above complexes were precisely mapped by methylation interference analysis.The DNA probe was end-labeled on each end and partially methylated with dimethyl sulfate. Binding reactions were performed with F9 and dF9 nuclear proteins. Analysis of the methylated protein-DNA complex I1 (Fig. 3) revealedthat methylation of guanine and adenineresidues at -141 and -138 on coding strand and at -139, -140, -145, and -146 on noncoding strand interfered with the binding of the protein. The same contact points were also observed

A Functional NF-Y-binding Site in hstPromoter

25044

-

NE probe cold

--

F9

dF9

F9 R/S R/S

/A

Y

Ym

-

R / S H/A

Y

Ym

probe antisense

-

R/S

sense

-133

I1

>

Ji4

T

T I11

>

A+ G

T C

7; -148

FIG.3. Methylation interference analysis of the protein-DNA complex established with WS. Partially methylated probe WS bound to complex I1 and in the free state as shown Fig.in2 was recovered from the gel and treated as described under “Experimental Procedures.” Protein-DNA contact points are indicated by arrowheads. L,F,and C represent G + A ladder, free probe, and probe bound in the complex, respectively. FIG.2. Formation of specific protein-DNA complexes with W S (-172“) probe in EMSA. Binding reactions and electrophoresis were performed a s described under “Experimental Procedures.” Endlabeled W S fragment from the hst promoter was incubated with 5 pg of nuclear extracts from F9 (lanes 2-6) and dF9 (lanes 7-11).Competitions, both specific (lanes 3,5,8,and 1 0 ) and nonspecific (lanes 4,6,9, and 11 ), were performed at a 200-fold molar excess. The positions of complexes observed with R/S are shown to the left of the autoradiograms. Lane 1 represents migration of protein-free DNA. NE, nuclear extract; cold, unlabeled competitors; F, free probe.

for complex I (data not shown). These DNA contact points enabled us to recognize a CCAAT box. Complex I11 did not show any methylated guanine or adenine residue interfering with binding (data not shown). The ProteinBinding to theCCAAT Box in F9 Nuclear Extract Is NF-Y-The IUS region of the hst promoter shared a 9-bp perfect homology with the consensus sequence of the Y-box found in the promoters of major histocompatibility complex class I1 genes (19). Since NF-Y is known to be the nuclearfactor that interacts with the Y-box, it became imperative to determine if the nuclear factor($ binding to the IUS fragment was NF-Y or related factor(s). For this purpose, both strands of a n oligonucleotide, Y, which included the sequences pinpointed by methylation interference analysis, along with a mutant oli-

gomer, Ym, were synthesized and annealed,respectively. Also, both strands of a n oligonucleotide corresponding to the NF-Ybinding site in the E a promoterweresynthesized and annealed. The resulting double-stranded oligonucleotides were then either used as probes or competitors in mobility shift experiments. As can be seen in Fig. 4 A , when Y oligomer was used as a probe, a single protein-DNAcomplex, termed W, was observed with F9 nuclear extract (Fig. 4 A , lane 2). However, when this probe was incubated with nuclear extract from dF9, another protein-DNA complex, YL, migrating witha faster mobility was also observed. The specificity of W and YL was confirmed by competing with unlabeled wild-type Y oligomer or with its mutant version, Ym. As anticipated, while Y oligomer inhibited theformation of both the complexes, Ym did not (Fig. 4 A , compare lane 5 with lane 3 for F9 andlane 12 with lane 10 for dF9). Next, we wished to see if the formation of W and YL could also be inhibited by competing with the unlabeled Ea oligomer. At a 200-fold molar excess, like Y, Ea completely inhibited the formation of W (Fig. 4 A , lanes 4 and 111. However, Ea was unable to inhibit the binding of YL to the same extent as thewild-type Y oligomer (Fig. 4 A , compare lanes 10 and 111. The fact that NF-Y-binding E a oligomer efficiently disrupted the formation of W strongly suggested that thiscomplex con-

A Functional NF-Ybinding Site

in hst Promoter C

B

A

-D

25045

ii-b"

..

UY

*.-3

-My

,

. "

-B

",3Y

-~"----J

.-

'

L
LL?

.IOl".'2~>'3

FIG.4. Oligonucleotide probe, Y, derived from FUS contains an NF-Y-bindingsite. A, EMSA was performed with double-stranded Y oligomer probe under the conditions described in the legend to Fig.2. Sources of nuclear extracts are F9(lanes 2-8)and dF9 (lanes 9-15). For competitions, Y(lanes 3 and IO),mutant Y ( Y m )(lanes 5 and I2),and Ea(lanes 4 and II) oligomers were included. For supershift experiments, 0.6 pl ofYAla, monoclonal antibody against NF-YA(lanes 6 and 13), and 0.06 p1 of a n affinity-purified polyclonal antibody specific for NF-YI3 (lanes 7 and 1 4 )or a1:20 dilution of antiserum alsospecific for NF-YB (lanes 8 and 15) were used.B, for supershift experiments,0.1 p1 of affinity-purified polyclonal antibody specific for NF-YB (lanes 1 , 3 , 5 , and 7 )or a 1:lO dilution of antiserum (lanes 2 , 4 , 6 , and 8 ) was used. Y or E a probes are shown above the respective lanes. C , binding of NF-Y to itscognate binding site in E a oligomer probe. End-labeled E a probe was incubated with nuclear extracts from F9 (lanes 2-7)and dF9(lanes 8-13), and complexes were competed or supershifted as described inA (above). Lane I in panels A and C represents migration of protein-free DNA. Position of the supershifted complex is shown by a n arrow, and positions of W, NF-Y, and YL are shown to the left of the panels. NE, nuclear extract; cold, unlabeled competitors; F,free probe; Ab, antibody; Y A l a , monoclonal antibody against NF-YA A.P and cs, affinity-purified polyclonal antibody and rabbit serum against NF-YE!, respectively.

sisted of NF-Y. In order to establish this, monoclonal and polyclonal antibodies against NF-Y were included in the reaction mixtures before W and YL were separated byEMSA. The monoclonal antibody YAla, which binds to theglutamine-rich activation domain of the NF-YA subunit, supershiftedW (Fig. 4 A , lanes 6 and 13) but had no influence on YL (Fig. 4 4 , lane 13).Again, at appropriate dilutions, addition of affhity-purified polyclonal anti-NF-YB or anti-NF-YB rabbit serum to the reaction mixtures inhibited the formation of YU and primarily led to aggregated material at the top of the gel (Fig. 4 B , lanes 1-4). In contrast, the binding or mobility ofYL was not affected (Fig. 4 B , lanes 3 and 4 ) . As a control, a similar mobility shift assay wasperformed with theE a oligomer as the probe. As can be seen in Fig. 4C and Fig. 4B, lanes 5-8, results similar to those obtained withthe Y oligomer probe were observed except that a complex equivalent toYL was notobserved with nuclear extract from dF9 (compare Fig. 4C, lane 8 with Fig. 4 A , lane 9). These results demonstrate thatW is NF-Y and suggest that YL may represent an independentDNA-binding protein. We also employed U V induced cross-linking experiments to determine the polypeptide compositions of YU and YL. W yielded three cross-linked polypeptide bands with mobilities corresponding to approximately 47, 45, and 37 kDa when this complex was resolved on a denaturing SDS-polyacrylamide gel (Fig. 5, lanes 3, 5,and 7). The mobilities of these bands were similar to those reported for the cross-linked subunits of NF-Y yielded asingle cross-linked (28). YL, on theotherhand, polypeptide band with a mobility of approximately 70-80 kDa (lanes 2 and 4 ) . An approximately 200-kDa band in lane2 was not reproducible and probably represented an experimental artifact. hst Promoter Activity Is Attributable to the NF-Ybinding Site-In order to evaluate the functional significance of the

NF-Y-binding site that we characterized, CAT reporter plasmids having mutations in the CCAAT core sequence of the hst promoter were constructed. These promoter constructs were transfected into F9, and their CAT activities were compared with the level obtained in parallel transfections with the parent, FtA-53 (Fig. 6, lane 4 ) , to give a measure of the promoter activity. Mutation in theCCAAT sequence of the NF-Y-binding site in segment Y Cym-53) reduced the promoter activity by about 70% (lane 2). Deletion of Y (-64-53), as expected, brought thepromoter activity to the basallevel (lane 6). These results clearly demonstrate that theNF-Y-binding site in segment Y of the hstpromoter and theoctamer element present in the noncoding region of the third exon are the essential activating elements of the hst gene. NF-Y Associates with a Differentiation-inducedNuclear Factor-The implication of NF-Y as a nuclear factor that positively regulates theexpression from the hstpromoter in F9led us tohypothesize that in differentiated cell lines a differentiation-induced nuclear factor may associate with NF-Y. This assumption was tempting for two reasons. First, methylation interference results hadshown that theDNA contact pointsfor complex I were identicalto thoseestablished for NF-Y (data not shown). Second, the intensity of NF-Y (complex 11) from the dF9 nuclear extract was conspicuously lower than that observed with nuclear extract from F9, although equal amounts of nuclear extracts from both F9 and dF9 were used (Fig. 2, compare lane 2 with lane 7). To test this hypothesis, the R/S probe was incubated with dF9 nuclear extract in thepresence of the monoclonal antibody YAla. As can be seen in Fig. 7A, complex I, like NF-Y (complex 11), was supershifted. We then considered the possibility that the flankingsequences may be required for stabilization of complex I in vitro since unlabeled Y oligomer efficiently disrupted complex I, but when used as a

in hst Promoter

A Functional NF-Y-binding Site

25046

Cell

-

Band

control

YL

Lane

1

2

F9

dF9

Jurkat YU

3

YL 4

YU 5

YL 6

YU M.W.

7

+ 200 +- 97

FIG.5. UV cross-linking of nuclear proteins from FS, dF9, and Jurkat. U V irradiated W and YL complexes were separated by EMSA, cut from the gel, and analyzed on SDS-polyacrylamide gel electrophoresis. Lane 1 is a sample derived from a portion of the EMSA gel unrelated to complex. Lanes 3, 5, and 7 represent W from Jurkat, dF9, and F9, respectively; lunes 2 and 4 represent YL from Jurkat and dF9,respectively. In lune 6, a faint U V cross-linked complex (supposedly YL from F9 nuclear extract) was analyzed. Molecular mass markers are indicated by arrows. Open arrowheads indicate the position of cross-linked complexes.

Plasmid

Y

C- 6 8 I

L 43 I 29

CAT

Relative CAT oct A c t i v i t y ".

I.

Yrn-00

2. Yrn-53 3. R A - 0 0

0.31

aL--

0.16 1

5. - 6 4 - 0 0

r).>---_"_"-

0.19

6. - 6 4 - 5 3

0.1 0

7. 0 0 - 5 3

0.00

8. pSV2CAT

0.07

FIG.6. NF-Y-bindingsite is required to stimulate transcription from the hst promoter. CAT activities of a series of hst-CAT expression Procedures." The thin layer chromatogram shown is acetylation constructs transfected into F9 were determined as described under "Experimental of [dichloroacetyl-"C]chloramphenico1with 40 pl of cellular extractsof transiently transfectedF9. Relative CAT activities are the values corrected for transfection efficiencies by expression of cotransfected pRSV-LacZ. 53 represents a 3"fragment of the hst gene that contains the octamer sequence. The cross in Y represents mutations in this segment. Each transfection was performed as four independent experiments, and the chromatogram representsa typical transfection.

probe i t did not yield this complex (Fig. 4A). To test this possibility, a 41-mer oligomer bearing the core CCAAT sequence (-159/-119) was used as a probe in EMSA; still thiscomplex did not appear (data notshown). Then we designed a probe, pY, in which the Y oligomer was flanked with sequences derivedfrom the pBluescript KS+ plasmid (see "Experimental Procedures"). When this probe was incubated with nuclear extracts from F9 or dF9, protein-DNA complexes similar to those established earlier with theW S probe (Figs. 2and 7A) were seen (Fig. 7 B ) . Complex I, in addition to being inhibited by competing with Y and Ea oligomers (Fig. 7B, lanes 8 and 9),was again supershifted with YAla (lane 11 ). From these results we conclude that, in complex I, NF-Y is associated with a differentiation-

induced nuclear factor and that this interaction most likely involves a protein-protein interaction between NF-Y and a nonDNA-binding protein. DISCUSSION

Although the enhanceractivity of an octamer element residing in the noncoding region of the third exon of hst has been demonstrated (14-16,29), the mechanism(s)of hst activation in F9 has not been established as yet. In the present study, we have characterized the functional region of the hst promoter. The sequence harboring the CCAATbox in hst promoter shared a perfect homology with 9 bp of the CCAAT-bearing consensus Y-box sequence found in the promoters of major

A Functional NF-Y-binding Sitein hst Promoter

25047

I * NF-Y

FIG.7. NF-Y, in dF9, is complexed with a differentiation-inducednuclear factor. EMSA and supershift experiments were performed as described in the legends to Figs. 2 and 4, respectively, withend-labeled €US as probe ( A ) and witha 71-bp pY probe ( B ) (seetext). Source of nuclear extracts and competitors is indicated above the lanes. Arrowheads represent the position of complex I and arrows the position of supershifted complexes. YAla, monoclonal antibody against NF-Y. Abbreviations are explained in the legend to Fig.4.

NF-Y

F

histocompatibility complex class I1 genes (19). NF-Y, which interacts with theY-box, is required not only for the accurate and efficient transcription of major histocompatibility complex class I1 genes but also for the transcription from several promoters such as the rat albumin (24,30) anda collagen promoters (31).Here, we detected a major protein-DNA complex, W, binding with highspecificity to Y oligomer spanning -1531-130 of the hst promoter. Formation of W was completely inhibited by competing with the NF-Y-binding Ea oligomer (Fig. 4A). More importantly, monoclonal antibody YAlasupershifted this complex but hadno influence on YL, confirming the specificity of this supershift (Fig. 4A). Also, affinity-purified polyclonal anti-NF-YE3 and anti-NF-YB serum prevented the formation of this complex (Fig. 4, A and B ) .Furthermore, threecross-linked polypeptides observed for W (Fig. 5) are in agreement with those reported earlier for NF-Y (28). These data strongly suggest that NF-Y is the predominant nuclear factor binding to segment Y of the hst promoter. Complex I, visible with the lUS probe, is unique to nuclear extract from dF9 (Figs. 2 and 7). The DNA contact points established for this complex were identical to those for NF-Y (data not shown). When this complex was challenged with the monoclonal antibody YAla, it supershifted (Fig. 7A). Furthermore, when a single copyof the Y oligomer flanked by sequences derived from a plasmid was used as a probe, a similar

f

supershift wasobserved (Fig. 7 B ) .These findings show that, in complex I, NF-Y is associated with a differentiation-induced nuclear factor and suggest that this association may involve a protein-protein interaction between NF-Y and a non-DNAbinding protein. YL, observed with theY oligomer probe, also appears tobe a differentiation-induced nuclear factor for the following reasons. First, it competed in a sequence-specific manner (for example compare lane 12 with lane 10 in Fig. 4A).Second, a U V induced cross-linking of this complex yielded a single polypeptide band with a mobility of approximately 70-80 kDa (Fig. 5). Third, when F9was induced todifferentiate withretinoic acid Bt'cAMP, YL appeared after 2 days of differentiation.' Attempts to purify these factors are inprogress and will help US to elaborate further. The results presented in this report have demonstrated that NF-Y motif is an essential activating elementof the hst promoter in F9. The residual activity observed with either the intact NF-Y or the octamer binding site alone is below 30% of the response with both intact promoter and 3"enhancer element (Fig. 6).This line of evidence suggests that a concerted action of these elements and their respective transcription factors is required for maximum induction of hst gene expression.

S.Hasan, unpublished data.

25048

A Functional NF-Y-binding Site hst in

The differential expression of Oct3 in F9 may lead one to hypothesize that itcould primarily be responsible for modulating the expression of hst in F9. However, we have shown that the expression of oct3 cDNA in HeLa or PYS-2 cells, which do not express hst, did not rescue the expression from the hst promoter (17).Similarly, Shimazaki et ul. (32) have shown that transfection of Oct3 expression plasmid in HeLa cells did not rescue a P19 (embryonal carcinoma cell)specific enhancer. Oct3 has also been shownto play a suppressive role in F9 (33). These lines of evidence indirectly suggest the involvement of additional factor(s) that may selectively recruit Oct1/0ct3 to interact with NF-Y or bridge these with the general transcriptional machinery of the hst in F9. Such a mechanism has been demonstrated for differential promoter activation by Octl through selective corecruitment with herpesvirus truns-activator VP16 (34). Alternatively, a protein-protein interaction between a differentiation-induced nuclear factor and NF-Y may mask the transcriptional activation domain of NF-Y without altering its DNA binding activity inamanner analogous t o the binding of GAL80, a negative regulator, to the transcriptional activation domain of GAL4 (35). Work is in progress to characterize the nuclear factor that associates with NF-Y. This will help in elaborating on the possible mechanisms by whichtranscription of the hst gene is modulated. In summary, we have demonstrated that, in addition to a distal octamer enhancer element, a functional NF-Y binding site in the proximal promoter region is required for full activation of the hst gene in F9. Our study now provides a useful model that may enable one to study the roles of NF-Y- and octamer-binding proteins in developmental regulation. Acknowledgments-We are grateful to Drs. D. Mathis, C. Benoist, and R. Mantovani for kindly providing us with monoclonal and polyclonal antibodies against NF-Y. We thank Dr. K. Yoshida for critical reading of the manuscript and valuable suggestions, Dr. A. Sasaki for the generous gift of some of the plasmids used in thisstudy, K. Yamaguchi for oligonucleotidesynthesis, and C. Oshiba for technical assistance. S. H. is indebted to the members of The Graduate School of Environmental Science forsupport and encouragement.

4. Sakamoto, H., Mori, M., Taira, M., Yoshida, T., Matsukawa, S., Shimizu, K, Sekiguchi, M., Terada, M., and Sugimura, T.(1986)Proc. Natl. Acad. Sci. U.S.A. 83, 3997-4001 5. Koda, T.,Sasaki,A., Matsushima, S., and Kakinuma, M. (1987)Jpn. J. Cancer Res. 78,325328 6. Delli-Bovi, F!, and Basilico, C. (1987)Proc. Natl. Acad. Sci. U.S. A. 134,56605664 7. Burgess, W. H., and Maciag, T.(1989)Annu. Rev. Biochem. 68,575-606 8. Manes, I., Adelaide, J., Raybaud, F., Mattei, M. G., Coulier, F.,Planche, J., de-Lapeyriere, O., and Birnbaum, D. (1989)Oncogene 4,335-340 9. Finch, P. W., Rubin, J. S., Miki, T., Ron, D., and Aaronson, S. A. (1989)Science 246,752-755 10. Niswander, L., and Martin, G. R.(1992)Deuelopment (Comb.)114,755-768 11. Drucker, B. J., and Goldfarb, M. (1993)Mech. Deu. 40, 155-163 12. Yoshida, T.,Muramatau, H., Muramatsu, T., Sakamoto, H., Katoh, O., Sugimura, T., and Terada, M. (1988)Biochem. Biophys. Res. Commun. 167, 61W25 13. Velcich, A., Delli-Bovi, P., Mansukhani, A,, Ziff, E. B., and Basilico, C. (1989) Oncogene 6,31-37 14. Sasaki, A., Kubo, M., Hasan, S., Yano, Y , and Kakinuma, M. (1991)Jpn. J. Cancer Res. 82, 1191-1195 15. Curatola, A. M., and Basilico, C. (1990)Mol. Cell. B i d . 10, 2475-2484 16. Schoorlemmer, J., and Kruijer, W. (1991)Mech. Deu. 36, 75-86 17. Koda, T.,Hasan, S., Sasaki, A,, Arimura, Y., and Kakinuma, M. (1994)FEBS Lett. 342,71-75 18. Scholer, H. R.,Hatzopoulos,A. K., Balling, R., Suzuki, N., and Gruss, P. (1989) EMBO J. 8,2543-2550 19. Dorn, A., Durand, B., Marfing, C., Le Meur, M., Benoist, C., and Mathis, D. (1987)Proc. Natl. Acad. Sci. U.S. A. 84,6249-6253 20. Chodosh, L. A., Baldwin, A. S., Carthew, R. W., and Sharp,P. A. (1988)Cell 63, 11-24 21. Hatamochi, A., Golumbek, F! T., Van SchaRingen, E., and de Crombrugghe, B. (1988)J. Biol. Chem. 263,5940-5947 22. Schreiber, E., Matthias, P., Muller, M. M., and Schaffner, W. (1989)Nucleic Acids Res. 17, 6419 23. Bradford, M. M. (1976)Anal. Biochem. 72,246-254 24. Mantovani, R., Pessara, U.,Tronche, F., Li, X-Y., Knapp, A., Pasquali, J., Benoist, C., and Mathis, D. (1992)EMBO J. 11, 3315-3322 25. Pascal, E.,and Tjian, R. (1991)Genes & Deu. 5,1646-1656 26. Maxam, A,, and Gilbert, W. (1980)Methods Enzymol. 66,499-560 27. Cooney,A. J., %ai, S. Y., and %ai, M. (1993)in Dunscription Factors-A Practical Approach (Rickwood, D., and Hames, B. D., eds) pp. 49-79, IRL Press at Oxford University Press, Oxford 28. Li, X-Y., HooR van Huijsduijnen, R., Mantovani, R., Benoist, C., and Mathis, D. (1992)J . B i d . Chem. 267,8984-8990 29. Ma, Y. Rosfjord, E., Huebert, C., Wilder, P., Tiesman, J., Kelly, D., and Rizzino, A. (1992)Deu. Biol. 164,45-54 30. Ravmondiean. M.. Cereehini. S.. and Yaniv. M. (1988)Proc. Nutl. Acad. Sci. U . S . i. 86; 757-761 . . 31. Maity, S. N.,Golumbek, P. T., Karsenty, G., and de Crombrugghe, B. (1988) Science 241,582485 32. Shimazaki, T., Okazawa, H., Fujii, H., Ikeda,M., Tamai, K., McKay, R.D. G., Muramatsu. M.. and Hamada. H. (1993)EMBO J. 12.4489-4498 ~, 33. Lenardo, M. J., Staudt, L., Robbins, P., Kuang, A., Mulligan, R. C., and Baltimore, D. (1989)Science 243, 544-546 34. Clearlv. M. A,, Stern.. S.,. Tanaka, M., and Herr, W. (1993)Genes & Deu. 7, 72-83 35. Ma, J., and Ptashne, M. (1987)Cell 60,137-142 ~~~~~~

REFERENCES 1. Lewin, B. (1990)Cell 61, 1161-1164 2. Howe, C. C., and Solter, D. (1981)Deu. Bid. 84,239-243 3. Strickland, S., Smith, K. K., and Marotti, K R. (1980)Cell 21,347355

Promoter

~~~

~~~

~

~