THEJOURNALOF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry andMolecular Biology, Inc.

Vol. 268,No. 5, Issue of February 15,pp. 3791-3796,1993 Printed in U.S.A.

Evidence for the Involvement of at Least Two Distinct Transcription Factors, One of Which Is Liver-enriched, for the Activation of the Phosphoenolpyruvate Carboxykinase Gene Promoterby CAMP* (Received for publication, August 26, 1992)

William J. Roeslerz, Pam J. McFie, and DebraM. Puttick From the Department of Biochemistry, University of Saskatchewan, Saskatoon, Saskatchewan, Canada S7N 0 WO

A detailed analysisof the promoter sequence require-enhancers, exerting theireffects onto heterologous promoters ments for the cAMP induction of transcription of the independent of their orientation or location relative to the gene for phosphoenolpyruvate carboxykinase (GTP) transcription start site. Subsequently, proteins which bound (EC 4.1.1.32)was undertaken.It was determined that to these CREs were identified.While a multitude of such the CAMP-responsive unit of this promoter consisted proteins have been identified, only two of these have actually of two independently weak or inactive components; the been shownto be activated by protein kinaseA (PKA). CREB, typical cAMP response element (CRE) sequence located or CRE-binding protein, is phosphorylated at a specific serine a t position -85 and a region of the promoter extending residue by protein kinaseA whichincreases its transactivation from -300 to -230 which contained multiple binding 4). sites for liver-enriched nuclear proteins. A previous ability without altering its DNA binding properties (3, More recently a second protein, ATF-1, has been identified study had indicated that multiple binding sites are critical for the activity of this latter region (Liu, J., which also appears capable of mediating the induction of Park, E. A,, Gurney, A. L., Roesler, W. J., and Hanson, transcription by cAMP (5).ATF-1 containsa consensus phosR. W. (1991)J. Biol. Chem. 266, 19095-19102). In phorylation site for PKA and possesses extensive homology the present study, we extend this observation by dem-with CREB around this phosphorylation site. Additionally, onstrating that the activity of this region can be effec- the coexpression of ATF-1 and PKA catalytic subunit synertively substituted for by three copies of just one subsite gistically activated a promoter containinga CRE. Both CREB present in that region. This would suggest that the and ATF-1 are members of the ATF/CREBfamily of proteins, binding of three molecules of a single liver-enriched are capable of forming heterodimers, and are bothexpressed factor is what forms this component of the cAMP re- in a wide variety of cell types (5). sponse unit. The other component of the cAMP reTranscription of the gene encoding the cytosolic form of sponse unit, the CRE, has been shown previously to be PEPCK is rapidly and stronglyinduced by cAMP in ratliver bound by these same liver-enriched factorsas well as (6). The PEPCK promoter contains a CRE sequence at poby cAMP response element binding protein, leading to sition -85 in the 5“flanking sequence, and mutagenesis of some debateas to the identityof the protein mediating the CRE results in the promoter losing most of its cAMP the cAMP response through this element. By two difresponsiveness (1). DNase I footprinting studies have demferent experimental approaches, we show that neither (7). However,a onstrated that CREB binds to the CRE CCAATJenhancer binding protein nor D-site binding protein is the likely mediator of the cAMP response number of experimental findings indicate that the CRE in through theCRE, while cAMP response element bind- the PEPCK promoter is not solely responsible for the acute sensitivity of the promoter to CAMP. First,when one or two ing protein isa possible candidate. copies of an oligonucleotide containing the CRE are linked to truncated, heterologous promoters, the hybrid promoter is only minimally activated by cAMP (1, 8). However, when a Theability of cAMPtostimulatethetranscription of larger portion (-416/-62) of the PEPCK promoterwas fused specificgenes hasbeenthe subject of extensive research. to a truncated promoter, the response to cAMP was more Initially, regions of specific gene promoters which mediated a robust (8). These findingssuggested that a region of the cAMP response, termed cAMP response elements (CREs),’ promoter in addition to the CREwas required for the cAMP were identified (reviewed in Ref. 2). In many of these pro- responsiveness. A clue as to which sites in addition to the CRE mediated moters,theCREconsists of a palindromic sequence5’when it was noted that proteins, TGACGTCA-3’. In general, these CREs behave like typical the full cAMP response came which were enriched in liver nuclear extracts, bound notonly * This work was supported by operating grants from the Saskatch- at theCREbut also at several sitesupstream,clustered ewan Health Research Board and the Medical Research Council of together in a region of the promoter extendingfrom -300 to Canada. The costs of publication of this article were defrayed in part -230 (9). To date, three proteins - C/EBP (7, lo), LAP (I), by thepayment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 and DBP (11)have been shown to bind to theseregions, and all transactivate the PEPCK promoter as demonstrated by solely to indicate this fact. cotransfection assays in HepG2 cells. As mentioned above, $ T o whom correspondence should be addressed Tel.:306-9664375; Fax: 306-966-8718. CREB also binds to the CRE but not to the upstream, liverThe abbreviations used are: CRE, cAMP response element; specific binding sites in the PEPCK promoter (7). Site-diPEPCK, phosphoenolpyruvate carboxykinase; CAT, chloramphenicol acetyltransferase; C/EBP, CCAAT/enhancer binding protein; DBP, rected mutagenesis of several of these upstream sitesseverely of PKA, D-site binding protein; LAP, liver activator protein; CREB, cAMP reduced the abilityof CAMP,or the catalytic subunit response element binding protein; a-subunit, a-subunit of glycopro- to activate the PEPCK promoter(1). tein hormones; PKA, protein kinase A; bp, base pair(s). In the present study, we characterize in more detail the

3791

3792

Several Proteins Purticipute

CAMP in Induction

of PEPCK Promoter

synergistic activities of the CRE and P3/P4 region of the PEPCK promoter which make up the cAMP response unit. Our findings indicate that the full cAMP responsiveness of the PEPCK promoteris achieved by and requires the combination of a single CRE element, which may act as a binding for a site for a CREB-like protein, plus multiple binding sites liver-enriched transcription factor.

activity by the overexpression of the catalytic subunit of PKA, which has been shown to mimic, albeitin a more robust manner, theeffect of adding exogenous cAMP analogs to the media (1). The increasedinduction achievedprovides the sensitivity necessary to precisely analyze the role of various promoter regions in the cAMPresponse. Identification and Characterization of Two Functionally Distinct Regions of the PEPCK Promoter Which Make Up the EXPERIMENTALPROCEDURES Complete CRE-We carried out a series of experiments to define more preciselythe requirementsfor the cAMP responMaterials-DNA-modifying enzymes were purchased from Boehringer Mannheim, New England Biolabs, Pharmacia LKB Biotechsiveness of thePEPCKpromoter.Initially, we tested 5’ nology Inc., and United States Biochemicals. Poly(d1-dC) . (dI-dC) deletion mutants of the PEPCK promoterfor their ability to waspurchased from Pharmacia. [Y-~’P]ATP (3000 Ci/mmol) and beactivated by PKA (Fig. 1). Inagreementwithearlier [a~etyl-~HIacetyl CoA (10 Ci/mmol) were purchased from Du Pontfindings, themagnitude of activationdropssharply when New England Nuclear. Tissue culture supplies were from GIBCO. sequences containing the P3/P4 region are deleted (compare HepG2 cells (ATCC HB8065) were acquired fromAmerican Type Culture Collection. Oligonucleotides were synthesized and gel-puri- activation by PKA of p-355 PCK-CAT with that of p-200 fied by the Regional DNA Synthesis Laboratory, University of Cal- PCK-CAT) (1).The CRE remaining in the 5’ deletion mugary, Calgary, Alberta. tants below -200 confers a small level of cAMP responsiveTransfection Experiments”HepG2 hepatoma cells were grown in ness to the truncated promoters. Interestingly, 5’the deletion Dulbecco’s modified essentialmediumcontaining 10% fetal calf mutant -68 PCK-CAT, which lacks the CRE, was consistserum (GIBCO). Four h before transfection, the cells were plated to approximately 30% confluency in 10-cm plates. DNA transfections ently observed to be inhibited by the overexpression of the were carried outby the calcium phosphate-precipitation procedure as catalytic subunit of PKA (Fig. 1). described inSambrook et al.(12). pRSV-@-galactosidase(the @The data above clearly indicate that the region extending galactosidase gene driven by the Rous sarcoma virus promoter) was from -355 to -200 is required for the full cAMP responsivecotransfected inall experiments to correct for transfection efficiency. ness of the PEPCK promoter in addition to thewell characThe amount of DNA used in the various transfection experiments terized CRE. In fact, the 5’ deletion analysis suggests that it was maintained a t 25 pg per plate by the addition of the plasmid pTZ18R. The construction of expression vectors pMSV-C/EBP (13), is this region that is the more powerful in conferring cAMP responsiveness. T o address this question,we created thevecpCMV-DBP (141, pRSV-CREB (7), and pMt-C (PKA expression vector) (15) have all been previously described. Cell extracts were tor pDR4, which links the-3551-200 region to the-68 PCKprepared 48 h after transfection. @-Galactosidaseassays were carried CAT vector, and measured its ability to be activated by the out as described (12)andCATassays were performed using the catalytic subunit of PKA. This promoter chimera displayed [~cetyl-~H]coenzyme A-based assay (16).The differences in transfec- only weak cAMP enhancer activity (Fig. I), similar to the tion efficiency were corrected for by dividing the CAT activityby the -109 deletion mutantwhich can be viewed as containingonly @-galactosidase activity. the CRE “linked to the -68 5‘ deletion promoter. These data DNase ZFootprinting-The preparation of DNA probes and DNase in the I footprintingconditions havebeendescribedpreviously (9). All suggested thatthestrong,intactcAMPenhancer DNase I footprinting reactions contained 0.5 pg of poly(d1-dC) .(dI- PEPCK promoter actually consisted of two spatially sepadC). A description of the productionof recombinant CREB, C/EBP, rated promoter regionswhich independently display weak and DBP hasbeen detailed previously (9, 11). cAMP enhancer activity. T o test this hypothesis, we next Construction of CAT Vectors-The construction of the5’-end -109 5’ deletions (-490, -355, -200, -109, and -68 PCK-CAT) and internal linkedthe -355/-200 promoterfragmenttothe deletion mutant pDR4, of the PEPCK promoter, linked to the CAT deletion mutant to form pDR6. This vector now contained structural gene, have been described previously (7). Internal deletion both of the weak cAMP enhancers, the CRE and the P3/P4 mutant pDR6 was constructed by ligating an NdeIIBsu361 fragment region, similar to the intact promoter except that promoter (-355 to -200) of the PEPCK promoter to the -109 PCK-CAT 5’ sequences from -200 to -109 were deleted (Fig. 1). This

deletion mutant which had been linearized with XbaI. Several other vectors were constructed which contained one, two, or three copies of double-stranded syntheticoligonucleotides linked either to -68 PCKCAT or -109 PCK-CAT. The two double-stranded oligonucleotides - 355 PCK-CAT used contained PEPCK promoter sequences containing the cAMP response element or site P3 (see Ref. 9 for description). paPEPCKCAT was synthesized by site-directed mutagenesis using a modification of the method of Kunkel (17) as described by Liu et al. (18). -200 PCK-CAT Uracil-containing -490 PCK-CAT was used as the template, and the mutagenic oligonucleotide used contained the sequence 5’- IO9 PCK-CAT AAGGCCGGCCatTgACGTCAtAGGCGAGCCTCC-3’, with thelowercuse letters indicating the mismatched nucleotides. For all of the constructs made, the identification and verification of positive clones, -68PtK-CAT including number and orientation of inserts, was determined using restriction analysis and/or dideoxy sequencing. pDR4

Fold Induction by PKA P3IP4 r q i o n

CUE I

I

A 4

59L5

3 5 0.7

2 + 0.3

(3+ 0 . 4 ) -3 2 2 0.2

RESULTS

The region termed P3/P4 contains multiple binding sites 6 3 2 10 pDR6 for proteins enriched in liver nuclear extracts including C/ EBP, LAP, and DBP (1,7,9), and mutations in three different FIG. 1. Identification and characterization of the two comareas of this cluster results in a substantial reduction in the ponents of the CAMPresponse unit of the PEPCK promoter. cAMP responsiveness of the PEPCK promoter(1).It should HepG2 cells were transfected with 7 pg of the indicated CATvector, 1 pg of pRSV-@gal,in the presenceor absence of 5 pg of pMt-C per be noted that CREB binds only to the CRE, whereas the The location of the CRE and P3/P4 region in the vectors are liver-enriched proteins bind to both the CRE and the P4/P3 plate. indicated. The 5’ deletion endpoints are indicated in the plasmid region. name to the left or, as in the case of pDR4 and pDR6, shown in the All of the experiments in this paper indirectly measure theschematic drawingof the vector. Values shownare theaverage S.E. cAMP responsiveness by measuring the induction of promoter from three independent experiments.

*

Several Proteins Participate

in cAMP Induction

construct was activated 63-fold by PKA, similar to that observed using the -355 5’ deletionpromoter (Fig. 1). This finding indicated that these two regions function synergistically to mediate the cAMP responsiveness of the PEPCK promoter. Additionally, it indicated that the sequences between -200 to -109 play little if any function in the cAMP response, in agreement with site-directed mutagenesis analysis of the PEPCK promoter(1). The P3/P4 region,which above has beenshown to act synergistically with the CRE to mediate the cAMP responsiveness of the PEPCK promoter, ais155-bp fragment which contains several binding sites for C/EBP (7) and DBP (ll), with the binding affinity at site P3being the highest for both proteins. Liu et al. (1) showed that all cAMP inducibility of the PEPCK promoter could be abolished by creating a mutant which had both CRE and P3 sequences mutated. Thus, we nextwantedtoestablishwhethertheP3 sequencealone, linked incis to the CRE, could mediate thesynergistic activity displayed by the -355/-200 fragment. A synthetic doublestranded oligonucleotide representing the P3 sequence was ligated to the -109 deletion mutant, and clones containing one, two, and threecopies of the oligonucleotide were isolated and named p-l09/P3X1, -X2, and -X3, respectively (Fig. 2). One copy of the oligonucleotide linked to the-109 promoter, which contains the CRE,failed t o synthesize a strong cAMP enhancer (Fig. 2). However, two copies produced a moderate enhancer, and three copiesproduced a robust response to PKA, being even larger than that observed using the intact promoter (see Fig. 1).These results demonstrate that, while the P3 site alone is not sufficient for the synergistic cAMP inducible activityof the -355/-200 fragment, multiplecopies of the P3 oligonucleotide can mimic this activity. This suggests that perhaps the binding of a single protein to multiple sites in the P3/P4 region is what is required for the activity of this componentof the cAMP response unit of the promoter. Wenextwantedtoestablishwhetherornotthesame protein mediated the cAMP enhancer activity of each component of the cAMP enhancer unit. If it was the same protein, then the PKA inducibility of a construct containing either multiple copies of the CRE or the P3 site should be similar. T o answer this question, one, two, or three copies of either the CRE sequence or the P3 sequence were ligated to -68 PCK-CAT, and then tested for their ability t o be activated by PKA. One and twocopies of the CRE displayed little activity when fused to the -68 5’ deletion promoter (Fig. 3), similar to previous findings (1, 8) . When three copies of the CRE were present, a 22-fold activation by PKA was observed. Fold Induction b y PKA CRE

- I 0 9 PCK-CAT

p- 109/P3XI

p- 109/P3x2

p- 1 0 9 / ~ 3 x 3

2 2 0.3

0.2

142 3

16

FIG. 2. Three copies of the subsite P 3 can effectively mimic the activity of the P3/P4 region component of the cAMP response unit when linked to the CRE. HepG2 cells were transfected as described in the legend to Fig. 1. The indicated number of copies of asynthetic double-stranded oligonucleotide containing PEPCK promoter sequences from -263 to -225 were ligated to the -109 5’ deletion mutant of the PEPCK promoter as described under “Experimental Procedures.” Values shown are the averages f S.E. of three independent experiments.

of

PEPCK Promoter

3793 Fold Induction bv P K I

- 6 8 PCK-CAT

-6EXREXI

(4% 0.2)

Ezl-dkl

I LO2

-6BlCREXZ

-6WCREX3

-68IP3XI

1203

CREICD.E~cF?f

@”dk

22i 3

1ro.3

-68lPJXZ

250.4

-681PJX3

420 5

FIG. 3. Evidence for a functional distinction between the CRE and P3 element of the PEPCK promoter. HepG2 cells were transfected as described in the legend to Fig. 1. Details on the construction of the vectors shown are described under “Experimental Procedures.” Values shown are the averages f S.E. of three independent experiments. Value shown in parens reflects a -fold decrease in CAT activity produced by PKA.

Such cooperativity between tandemly arranged binding sites is a common observation(19-21). However, it shouldbe noted that the-fold activation producedwas still significantly lower than that produced by the intact PEPCK promoter, or other promoter constructswhich contain both the CRE and the P3/ P4 region (see Figs. 1 and 2). When the P3 sequence was similarly tested, a different observation was made. No such synergism was detected, with the -68/P3X3 construct displaying only a 4-fold induction by PKA. The data in Fig. 3 indicate that a distinct functional difference exists between the CRE and the P3 element, implying further that distinct proteins bind to these two sequences. Apparently, the protein binding to the P3 site is unable to independently mediatea strong induction in response to PKA overexpression, but can only act synergistically with the protein bindingat the CRE to mediate this induction. The above data suggested that a liver-enriched transcription factor(s) was involved in the cAMP responsiveness of the PEPCK promoter. Since at least one of these factors has been shown to be present atmuch reduced concentrations in transformed cells in culture compared to fully differentiated liver cells (13), we examined whether overexpression of these factors along with the catalytic subunit of PKA would result in the synergistic activation of the PEPCK promoter. Two liver-enriched transcription factors, C/EBP and DBP, were shown to transactivate the PEPCK promoter HepG2 in cells, with DBP providingahigher level of induction (Table I), similar to previous findings (11).When the catalytic subunit of PKA, which by itselfprovided a 59-fold induction of PEPCK promoter activity,was coexpressed with either DBP or C/EBP, a synergistic effect on promoter activity was observed (Table I). The activation by DBP went from 104-fold in the absenceof PKA overexpression to 432-fold when PKA was coexpressed; the inductionby C/EBP went from 6- to 98fold (Table I). In contrast, the activityof the SV40 promoter (pSV-CAT) was essentially unaffected by coexpression of either DBP, CIEBP, or PKA (Table I). CREB, but Not C/EBP or DBP, Is Capable of Mediating a c A M P Response through the CRE-There has been uncertainty as to the identity of the transcription factor which binds to the CRE in the PEPCK promoter and assists in mediating the cAMP response. This uncertainly arose from the observation thatsome of the liver-enriched transcription

3794

Several Proteins Participate in CAMPInduction of PEPCK Promoter

TABLE I Synergistic activation of PEPCKprornoter in HepG2cells by coexA pression of liver-enriched transcription factors and protein kinase HepG2 cells were transfected with 7 pg of CAT plasmid, 1 pg of pRSV-pgal, and either 5 pg of pCMV-DBP, 5 pg of pMT-C, or 10 pg of pMSV-C/EBP. The total amount of DNA per transfection was kept constant by the addition of pTZ19R. Transfections, CAT, and @galassays were carried out as described under “Experimental Procedures.” Values shown are theaveraged f S.E. of three independent experiments.

Footprinting studies showed that C/EBP binds with much lower affinity to the CREs in the a-gene than to the CRE in the PEPCK promoter (data not shown). This suggests that the nucleotide sequence differences within and/or surrounding the palindromic CRE in the a-subunit gene negatively impact on C/EBP bindingaffinity. Based on this finding, we decided to use site-directedmutagenesis to mutate the CRE in the PEPCK gene to the CRE present in the a-gene, along with a few contexual nucleotides. If our hypothesis above was correct, and CREB binding at Stimulation of CAT activity Proteids) overexpressed the CRE was necessaryfor the cAMP response, then we p-490 PCK-CAT pSV-CAT should observe little change in the induction of this mutant -fold promoter by PKA, since the CREs in the a-subunitgene are Catalytic subunit of PKA 59 f 5 2 0.3 powerful cAMP enhancers (22, 23) and bind CREB.2 AlterDBP 104 f 23 1 0.2 natively, if a liver-enrichedtranscription factorwas mediating C/EBP 6 f 2 1 f 0.3 the cAMPresponse at the CRE, then a reduction in the PKAPKA + DBP 432 f 61 2 f 0.2 PKA + CIEBP induction might be expected since a lower affinity binding 98 -+ 19 3 f 0.4 site for C/EBP and DBP (see Fig. 5) was introduced. To verify that the mutations incorporated into the CRE did in fact result indecreased binding affinityfor C/EBP and DBP, we carried out titration footprint experiments (Fig. 5 ) . Both C/EBP and DBP clearly bind to the mutated CRE in the aPEPCK promoter with a lower affinity compared to wild type (Fig. 5 , A and B ) . It was approximated that 13x and lOOx more of the recombinant C/EBP and DBP, respectively, was required to produce anequivalentfootprint over the mutated CRE compared with the wild type CRE. That the mutations hada greater effect on DBP binding predictable, is Control DBP C/EBP CREE since several studies have shown that DBP hasa more strin(11, 14, 25). FIG. 4. CREB, but not C/EBP or DBP, can mediate PKA gentbinding specificity comparedtoC/EBP induction of the PEPCK promoter through the CRE compo- Conversely, there was no apparent change in the binding nent. HepG2 cells were transfected as described in the legend to Fig. affinity of recombinant CREB for the mutated CRE in the 1. The CAT vector used was the -109 5‘ deletion mutant of the aPEPCK promoter compared to thewild type sequence (Fig. PEPCK promoter, which contains the CRE but lacks all other upstream elements (see Fig. 1).The PKA induction of this truncated 5C). Not only did the DNase I-hypersensitive bands within promoter was tested in the absence of transcription factor coexpres- the CRE region disappear with equivalent concentrations of sion (control)or in the presence of DBP, C/EBP,or CREB coexpres- CREB used, but there was also the simultaneous appearance sion. Values shown are the averages -+ S.E. of three independent in both probesof the DNase I-hypersensitive band at the top experiments. edge of the footprintas CREB binding occurred (Fig. 5 C ) . The wild type (-490 PCK-CAT) and aPEPCK promoters factors (C/EBP, DBP, LAP), in addition to CREB, also bind were then compared with respect to thelevel of transactivaby to the CRE (1, 7, 10, 11).Additionally, a t least two of them tion producedby C/EBP, DBP, CREB, and induction PKA in HepG2 cells. Both promoters responded similarly to have been shown to be able to transactivate the promoter through the CRE. We decided to carry out a simple experi- PKA and CREBcoexpression (Table II), consistent with the ment to address this question. HepG2 cells were transfected CREB titration footprint data shown above. However, the with the -109 PCK-CAT reporter plasmid, which contains level of transactivation produced by DBP was significantly diminished in the paPEPCK-CAT construct compared to wild only the CRE element, with or without expression vectors encoding the catalytic subunitof PKA and C/EBP, DBP,or type, againlikely reflecting the lowered DBP binding affinity CREB. Since there is only a single binding site, the CRE, for displayed by the mutant promoter (see Fig. 5 B ) . The inducthese proteins in the promoter of the reporter plasmid, only tion produced by C/EBP was also lower for paPEPCK-CAT the transcription factor which is capable of mediating PKA- compared to thewild type vector. It should be noted that the induced stimulation of transcription shouldshow an additive basal activity of the aPEPCK promoterwas not significantly or synergistic activation when coexpressedwith PKA( 5 ) .The different from thewild type promoter (data not shown). results of this experiment,shown inFig. 4, indicate that,while DISCUSSION DBP, C/EBP, and CREB all are capable of transactivating thePEPCKpromoter,CREBaloneisableto producea The aimsof this study were to identify and characterize in synergistic activation in the presenceof PKA. These results detail the components of the complete cAMP response unit are consistent with our hypothesis (11) that, while C/EBP in the PEPCK promoter and to identify (or rule out) tranand DBP are capable of transactivating the PEPCK promoterscription factorswhich bind to thewell characterized CRE to through the CRE, in vivo most of their effects are exerted mediate the cAMP response. Our results show that two disthrough the P3/P4 region, leaving the CRE available for the tinct regions of the promoter are required; the highly conprotein which mediates the cAMPresponse. served CRE at position-85 (24) and a region of the promoter We further verified this hypothesis by capitalizing on the extending from -300 to -230 which contains several binding finding that C/EBPdoes not bind to all CREsequences with sites for liver-enriched proteins (9) . Evidence presented in equal affinity. The promoter for the gene coding for the a- this paper also support the hypothesis that CREB, but not subunit of glycoprotein hormones (a-gene) contains two iden- C/EBP or DBP, is capable of mediating the cAMPresponse tical, tandemly arranged CREs extending from nucleotides through the CRE, and thatmultiple binding sites for a liver-146 to -111 (22, 23), that differ only slightly in nucleotide * W. J. Roesler and P. McFie, unpublished observations. sequence from the CRE presentin the PEPCK promoter(24).

*

*

Several Proteins Participate cAMP in Induction

of PEPCK Promoter

3795

TABLE I1 The nPEPCK promoteris transactioated to a lesser degree by DBP and CIEBP thanthe wild-type promoter but its response to protein kinase A is unaltered HepG2 cells were transfected as described in the legend to Table I. Values shown are the averages f S.E. of three independent experiments. Stimulation of CAT activity Proteins overexpressed ~ 4 9 PCK-CAT 0 paPEPCK-CAT

-fold

Catalytic subunit of PKA

DBP C/EBP CREB

60 f 5 94 f 16 6 f l 3 f 0.4

52 2 7 20 r 5 3 f 0.5 3 f 0.2

phorylated a transcription factor. This factor,bound at a CRE, would interact favorably with the general transcription apparatus assembled at the transcription start site and act to increase initiationfrequency of RNA polymerase 11. However, it isnow apparent that this scenario greatly is oversimplified. It is now known that there is a large family of CREB/ATF proteins (27, 28), although only CREB and ATF-1have been demonstrated toconfer PKA induction (3-5). Other proteins outside of this family of proteins have also been implicated 8 a as being mediators of cAMP induction (29, 30). There have " also been several proteins recently identified which appear to inhibit the activity of CREB (31, 32). Moreover, in the case of some promoters, cAMP responsiveness involves multiple promoter elements. For example, the c-fos promoter contains four different sequence elements which can mediate induction by cAMP independently (33). The human prolactinpromoter requires two cis-acting elements for cAMP responsiveness, one of which binds the pituitary-specific factor Pit-1 (30). C Interestingly,concatenatedPit-1bindingsitesare able to wild type a confer cAMP responsiveness to a truncated thymidinekinase promoter butonly in pituitary cells. However, whether in fact Pit-1 wasindeed mediatingthecAMP response was not clearly established. In the case of the PEPCK promoter, previous work (1)as region, well as the present study indicate that the P3/P4 which binds liver-enriched nuclear proteins, plays a critical role in the cAMP response. This region has also been implicated in determining the tissue-specific and developmental expression pattern of this gene. McGrane et al. (34) showed that mice bearing a PEPCK promoter/bovine growth hormone transgene constructwhich lacks the P3/P4region show no expression of the transgene. When this region is present k" " *I in the promoter of the transgene, mice are produced which express growth hormone in the propertissues, Le. in theliver, to a lesser degree in kidney, and with the correct developFIG. 5. Converting the CRE in the PEPCK promoterto the mental expression pattern (34, 35). These findings suggest CRE present in the a-subunit gene resultsa in lower binding that the P3/P4 region plays a multifunctional role in the affinity for C/EBP and DBP. A 279-bp Bsu36I-BglII fragment, containing promoter sequences from -205 to +73, was prepared from promoter, encompassing tissue-specific, developmental, and either the wild type promoter or (uPEPCK promoter by labeling on hormonal controlof promoter activity. the coding strand. The probes were incubated with increasing conThe P3/P4 region of the promoter containsmultiple bindcentrations of either recombinant C/EBP ( A ) , DBP ( B ) , or CREB ing sites for DBP and C/EBP (7, 11), as well as LAP (1). ( C ) and then subjected to DNase I digestion as describedunder Mutations in three different subsites of this region severely "Experimental Procedures." reduces the level of induction of PEPCK promoteractivity by cAMP (1).Interestingly, three copies of an oligonucleotide enriched protein(s) make up the other portionof the cAMP representing the single binding site P3 were required to be enhancer unit of the PEPCK promoter. linked to a promoter construct containing the CRErestore to Upon the discovery in several promoters of the highly inducibility by PKA to the level observed in the full length conserved palindromic CRE (reviewed in Ref. 2), followed by promoter (Fig. 2). Taken together, these data suggest that t h e identification and characterization of CREB (3, 4, 26), it multiple protein binding sites arerequired for the activity of at first appeared that the mechanism whereby cAMP induced this componentof the cAMPresponse unit. Since all three of transcription of genes was going to be straightforward. The the liver-enriched proteins mentioned above bind to the P3 model that emerged was that, as the concentration of cAMP subsite site, it impossible is to exclude any oneof these as the in thecell increased, PKA was activated which in turn phos- protein responsible for mediating the cAMP effect in con"

1

3 796

Several Proteins Participate in cAMP Induction

of

PEPCK Promoter

junction with the CRE. While we believe that the data in Fig. It isclear from the above discussion that a complete under2 is consistent with the hypothesis that only one member of standing of the mechanism of action of cAMP on PEPCK the liver-enriched proteins capable of binding to site P3 is gene transcription will require advancing from a qualitative involved in mediating the cAMPresponse, it is possible that to a quantitative analysis of transcriptional processes. For the requirement for multiple binding sites is to allow for a example, data presented in this paper suggest that CREB is response through the CRE, unique combination of these factors to bind,e.g. two C/EBP sufficient formediating the cAMP but it is not clear that CREB is necessary for this response. molecules and one DBP. One of our immediate goals is to identify the liver-enriched factor(s) involved in this process. Several possible lines of investigation, including the use of Recently, we have observed that the PEPCK promoteris not inhibitors of specific transcription factors, determination of of specific proteins for variouspromoter induced by coexpression of PKA in HeLa cells,2 despite the the binding affinities of the concentrationsof each of these (5). Since sites, and determination fact that HeLa cells express CREB and ATF-1 HeLa cells would not contain liver-enriched proteins, this notproteins in thenucleus, should provide valuable insights into only supports our view of the importance of such factors for the relative roles of individual transcription factors for the regulation of promoter activity. cAMP inducibility of the PEPCK promoter, but may also providea cell background in which toidentifythe liverAcknowledgments-We thank G. Stanley McKnight for the PKA specific factor(s) involved. expression vector, Steve McKnight for the C/EBP expression vector, The identity of the protein mediating the cAMP effect Chris Mueller for the DBP expression vector, John Nilson for the through the CRE firstbecame an issue of debate when it was vector containing the promoterof the gene encoding the a-subunitof observed that C/EBP (7), and more recently DBP (ll),could glycoprotein hormone, and Richard Hanson for the 5' deletion mutants of the PEPCK promoter. We thank Joe Angel for critically bind to the CRE and transactivate the promoter through this site. Subsequently, it was also shownthat if the CREsequence reading this manuscript. was mutated to a C/EBP binding site having perfect dyad REFERENCES symmetry, the mutated promoter maintained its cAMP in1. Liu, J., Park, E. A., Gurney, A. L., Roesler, W. J., and Hanson, R.W. (1991) J. Biol. Chem. 2 6 6 , 19095-19102 ducibility (1).However, since the binding of CREB to this 2. Roesler, W. J., Vandenbark,G. R., and Hanson, R. W. (1988) J. Biol. Chem. mutated sequence was not eliminated but only reduced in its 263,9063-9066 3. Gonzalez, G. A., and Montminy, M. R. (1989) Cell 59,675-680 binding affinity by approximately 4-fold, it was impossible to 4. Yamamoto. K. K.. Gonzalez. G. A., Menzel,. P... Rivier, J.. and Montminy, establish C/EBP as the cAMP mediating protein at the CRE M. R. (1990) Cell 60,611-617 5. Rehfuss, R. P., Walton, K. M., Loriaux, M. M., and Goodman, R. H. (1991) (1). In the present paper, we haveusedtwo different apJ. Biol. Chem. 2 6 6 , 18431-18434 proaches which together suggest that CREB, but not C/EBP 6. Lamers, W. H., Hanson, R. W., and Meisner, H. M. (1982) Proc. Natl. Acad. Sci. U. S. A. 79,5137-5141 or DBP, is capable of mediating the cAMP response through 7. Park, E. A., Roesler, W. J., Liu, J., Klemm, D. J., Gurney, A. L., Thatcher, J. D., Shuman, J., Friedman, A., and Hanson, R.W. (1990) Mol. Cell. the CRE component. First, we observed that CREB, but not Biol. 10,6264-6272 D B P or C/EBP, was able to cooperatewith PKA to induce a 8. Short, J. M., Wynshaw-Boris, A., Short, H. P., and Hanson, R. W. (1986) J. Biol. Chem. 261,9721-9726 PEPCK promoter construct containing only the CRE (Fig. 9. Roesler, W. J., Vandenbark,G. R., and Hanson, R. W. (1989) J. Biol. Chem. 4). Second, we showed that when the CREwas mutated to an 264,9657-9664 10. Trus, M., Benvenisty, N., Cohen, H., and Reshef, L. (1990) Mol. Cell. Biol. alternate CRE sequence to which DBP and C/EBP bound 10,2418-2422 with lower affinitybut left CREB'sbindingaffinityun11. Roesler. W. J.. McFie. P. J.. and Dauvin, C. (1992) J. Biol. Chem. 2 6 7 , 21235-21243 changed, the ability of PKA to induce transcription of this 12. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: A mutant promoterwas unaffected (Table 11). Laboratory Manual, Ed 2, Cold Spring Harbor Laboratory, Cold Spring Harbor, N Y However, it must be acknowledged that DBP and C/EBP 13. Friedman, A. D., Landschultz, W. H., and McKnight, S. L. (1989) Genes & Deu. 3 , 1314-1322 can, at least under certain conditions, transactivate through 14. Mueller, C. R., Maire, P., and Schibler, U. (1990) Cell 6 1 , 279-291 the CRE (Refs 7 and 11and Fig. 4). Since we have presented 15. Mellon, P. L., Clegg, C. H., Correll, L. A., and McKnight,G. S. (1989) Proc. Natl. A d . Sci. U. S. A. 86,4887-4891 evidence in support of the view that CREB may mediate the Nielsen, D. A., Chang, T.-S., and Shapiro,D. J. (1989)Anal. Biochen. 1 7 9 , cAMP response through this element, one of several situa- 16. 14-92 tions must exist to explain these findings. One, itmay be that 17. K i i k $ T. A. (1985) Proc. Natl. Acad. Sci. U. S. A. 82,488-492 18. Liu, J., Roesler, W. J., and Hanson,R. W. (1990) Biotechniques 9,738-742 in uiuo, DBP, C/EBP, and/or other liver-enriched proteins 19. Ondek, B., Shepard, A,, and Herr, H. (1987) EMBO J. 6,1017-1025 Schirm, S., Jiricny, J., and Schaffner, W. (1987) Genes & Deu. 1, 65-74 select the P3/P4 region a t which to bind, leaving CRE avail- 20. Fromental, C., Kanno, M., Nomiyama, H., and Chambon, P. (1988) Cell 21. able for binding by CREB. As elaborated on previously (ll), 54,943-953 Silver, B. J., Bokar, J. A., Virgin, J. B., Vallen, E. A,, Milsted, A., and 22. this scenario impliesthat mechanisms exist whereby a specific Nilson, J. H. (1987) Proc. Natl. Acad. Sa. U. S. A. 8 4 , 219&2202 protein can be selected for binding at a given site, perhaps 23. Deutsch, P. J., Jameson, J. L., and Habener, J. F. (1987) J. Biol. Chem. 262,12169-12174 from guidance by an adjacently bound protein (36). Second, 24. Bokar, J. A,, Roesler, W. J., Vandenbark, G. R., Kaetzel, D. M., Hanson, it may be that CREB, C/EBP, and DBP compete for binding R. W., and Nilson, J. H. (1988) J. Biol. Chem. 263,19740-19747 van Ooij, C., Snyder, R. C., Paeper, B. W., and Duester,G. (1992) Mol. Cell. 25. at the CRE, and thatof the three, CREB hasa significantly Biol. 12,3023-3031 26. Montminy, M. R., and Bilezik'ian, L. M. (1987) Nature 3 2 8 , 175-178 higher binding affinity for this sequence. We are currently 27. Hai, T., Liu, F., Coukos, W. j., and Green, M.R. (1989) Genes & Deu. 3 , carryingoutScatchardanalysistodeterminethebinding 9nnn-m~ affinities of these proteins for the CRE sequence. Third, it 28. Habener, J. (1990) Mol. Endocrinol. 4,1087-1094 29. Imagawa, M., Chiu, R., and Karin, M. (1987) Cell 5 1 , 251-260 may be that a binding switch occurs at the CRE, depending 30. Peers, B., Monget, P., Nalda, M. A,, Voz, M. L., Benvaer, M., Belayew, A., and Martial, J. A. (1991) J. Biol. Chem. 2 6 6 , 18127-18194 on the physiological signals being transmitted into the cell. 31. Foulkes, N. S., Borrelll, E., and Sassone-Corsl, P. (1991) Cell 6 4 , 739-749 Under this scenario, CREB might be the binding protein at 32. Walton, K. M., Rehfuss, R. P., Chrivia, J. C., Lochner, J. E., and Goodman, R. H. (1992) Mol. Endocrinol. 6,647-655 the CRE under conditionswhere cAMP levels are elevated in 33. Fisch, T.M., Prywes, R., Simon, M. C., and Roeder, R. G. (1989) Genes & the cells, whereas at other times DBP and/or C/EBP might Deu. 3 , 198-211 McGrane, M. M., Yun, J. S., Roesler, W. J., Park, E. A., Wagner, T. E., 34. be bound. Such a "protein switching" mechanism could be and Hanson, R. W. (1990) J. Reprod. Fertil. 4 1 , suppl., 17-23 regulated by phosphorylation of the transcription factors in- 35. McGrane, M. M., devente, J.,Yun, J., Bloom, J., Park,E., Wynshaw-Boris, A,, Wagner, T., Rottman,F. M., and Hanson, R. W. (1988) J. Biol. Chem. volved and would clearly allow for the fine-tuning of gene 263,11443-11451 Lamb, P., and McKnight, S. L. (1991) Trends Biochem. Sci. 16,417-422 36. expression. I " " I-"-