cyclic amp and phorbol ester-stimulated transcription mediated by similar dna elements that bind...

Cyclic AMP and Phorbol Ester-Stimulated Transcription Mediated by Similar DNA Elementsthat Bind Distinct ProteinsAuthor(s): Paul J. Deutsch, James P. Hoeffler, J. Larry Jameson and Joel F. HabenerSource: Proceedings of the National Academy of Sciences of the United States of America,Vol. 85, No. 21 (Nov. 1, 1988), pp. 7922-7926Published by: National Academy of SciencesStable URL: http://www.jstor.org/stable/32540 .

Accessed: 02/05/2014 11:04

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

National Academy of Sciences is collaborating with JSTOR to digitize, preserve and extend access toProceedings of the National Academy of Sciences of the United States of America.

http://www.jstor.org

This content downloaded from 194.29.185.185 on Fri, 2 May 2014 11:04:36 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=nas

http://www.jstor.org/stable/32540?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Proc. Natl. Acad. Sci. USA Vol. 85, pp. 7922-7926, November 1988 Biochemistry

Cyclic AMP and phorbol ester-stimulated transcription mediated by similar DNA elements that bind distinct proteins

(transacting factors/signal transduction/chorionic gonadotropin a gene)

PAUL J. DEUTSCH*, JAMES P. HOEFFLER, J. LARRY JAMESON, AND JOEL F. HABENERt

Laboratory of Molecular Endocrinology, Massachusetts General Hospital, and Howard Hughes Medical Institute, Harvard Medical School, Boston, MA 02114

Communicated by Alexander Leaf, July 21, 1988

ABSTRACT cAMP and phorbol esters mediate cellular metabolism by the activation of distinct signal transduction pathways consisting of a cascade of sequential protein phos- phorylations. An important consequence of the activation of these pathways is the stimulation of gene transcription by way of interactions of speciflc proteins with DNA control elements. The 8-base-pair (bp) DNA consensus sequence TGACGTCA [cAMP response element (cAMP-RE)] has been shown to confer cAMP responsivity on transcription from various promoters, and the closely related 7-bp consensus sequence TGA- (C or G)TCA [phorbol 12-myristate 13-acetate response element (PMA-RE)] lends transcriptional responsiveness to phorbol esters. In the JEG-3 placental cell line we fimd that several variants of the cAMP-REs fused to a gonadotropin a promoter chloramphenicol acetyltransferase reporter gene mediate responsiveness to cAMP but not to phorbol esters. The PMA-RE is responsive to phorbol esters but also imparts submaximal sensitivity to cAMP in the JEG-3 cells and in the Hep G2 hepatoma cell line. The transcriptional activities of cAMP-RE and PMA-RE are markedly influenced by the composition of the neighboring bases, but different sequences are permissive for the activity of the cAMP-RE versus the PMA-RE. The two signaling agents together display a supraadditive effect on reporter genes containing active PMA-REs but not cAMP-REs. Gel-mobility-shift and UV cross-linking analyses show that distinct proteins bind to the two control elements. One protein of 38 kDa binds to the cAMP-RE and several proteins of 48-84 kDa bind to the PMA-RE.

Many genes that are transcriptionally regulated by cAMP, including rat somatostatin (1), human a-gonadotropin (2-4), and human vasoactive intestinal polypeptide (5), contain a conserved sequence in the 5' flanking region that is identical or similar to the palindromic octamer TGACGTCA [cAMP response element (cAMP-RE)]. A very similar symmetrical heptameric motif, TGACTCA [phorbol 12-myristate 13-acetate (PMA-RE)], or close relatives thereof, has been identified in many genes transcriptionally activated by phorbol esters (6-8). A protein (CREB) that binds the cAMP-RE present in the rat somatostatin gene (1) has been characterized in PC-12 cell nuclear extracts and shown to have a molecular mass of -43 kDa on NaDodSO4/PAGE (9). The HeLa nuclear protein AP-1, 44-47 kDa on NaDodSO4/ PAGE, has been shown to bind the related 7-base-pair (bp) motif (7, 8). The similar sizes of these proteins characterized from different species by different laboratories make defin- itive determination of their identity or nonidentity difficult without a more direct comparison.

Most, if not all, of the effects of cAMP in eukaryotic cells (10), including the activation of gene transcription (11), are mediated by cAMP-dependent protein kinase. Phorbol esters

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. ?1734 solely to indicate this fact.

mimic the action of a physiologic mediator of cell surface signals, diacylglycerol, by activating protein kinase C (12). It is likely that the transcriptional activities of the cAMP-RE and PMA-RE consensus sequences are mediated by protein phosphorylation initiated by the respective signal transduction events of adenylate cyclase activation, which increases intracellular cAMP, and phosphatidylinositol bisphosphate hydrolysis, which releases diacylglycerol.

Because of the obvious similarity in consensus sequences, the parallels in mechanism of action, and the fact that the two signal transduction pathways can interact at several levels (13-20), we sought to characterize in the same experimental system the capability of either consensus motif to respond to either second messenger analog.

In this report we compare the signal-responsive properties of synthetic cassettes, including the 8-bp consensus cAMP- RE in the context of several different flanking sequences, with the transcriptional enhancing properties of the 7-bp consensus PMA-RE when synthesized in the contexts of the identical neighboring bases. We chose sequence contexts that surround the octamer or heptamer in 5' flanking regions of native genes. We report here that the functional properties of these related DNA elements can be dissociated by the addition or subtraction of a single base and that different DNA-binding proteins interact with the 8-bp and 7-bp sequences.

MATERIALS AND METHODS Plasmid Constructions. Construction of the plasmid

palOOCAT containing sequence - 100 through + 44 of the human chorionic gonadotropin a (CGa) gene has been described (3). Synthetic regulatory sequences were synthesized on an Applied Biosystems (Foster City, CA) model 380A DNA synthesizer as complementary pairs with overhanging GATC 5' cohesive ends. Adjacent to the GATC ends were a C-G base pair on one terminus and an A-T base pair on the other terminus so as to reconstitute a BamHI or Bgl II site on the respective ends. Prehybridized oligonucleotide duplexes were cloned into the Bgl II site located at base - 100 of palOOCAT as described (3).

Cell Culture, DNA Transfection, and Chloramphenicol Acetyltransferase (CAT) Assays. JEG-3 (HTB36) and Hep G-2 cells (HB8065) were obtained from the American Type Culture Collection and grown on 60-mm plates in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. Cultures were transfected prior to confluence with 2

Abbreviations: PMA, phorbol 12-myristate 13-acetate; cAMP-RE, cAMP response element; PMA-RE, PMA response element; CG, chorionic gonadotropin; COL, collagenase; SV, simian virus; CAT, chloramphenicol acetyltransferase; 8-Br-cAMP, 8-bromoadenosine 3',5'-cyclic monophosphate. *Present address: Division of Endocrinology, Department of Med- icine, Cornell University Medical College, 1300 York Avenue, New York, NY 10021. tTo whom reprint requests should be addressed.

7922



Biochemistry: Deutsch et al. Proc. Natl. Acad. Sci. USA 85 (1988) 7923

,ug of plasmid and 8 ,ug of carrier pGEM (Promega Biotec, Madison, WI) DNA using the calcium phosphate precipita- tion method for Hep G-2 cells (21); this was followed 4 hr later by glycerol shock (22). Cells harvested from each 60-mm dish were resuspended in 100 /ul of 0.25 M Tris HCI (pH 7.8) and assayed for CAT activity as described (3).

DNA-Binding Protein Analyses. Whole cell extracts were prepared from JEG-3 cells by the method of Manley et al. (23). Protein concentrations were determined by the Brad- ford assay (24). Gel-mobility-shift analysis was performed as described (25, 26). UV cross-linking experiments were performed by the method of Chodosh et al. (27). To generate uniformly labeled probes, the synthetic oligonucleotide 5'- AAAGCCAGAGGTGTCTGAC(G)TCATGCTTTATAA- CATCCTCTTGATTAGCTA-3' was annealed to the 15-base primer 5'-TAGCTAATCAAGAGG-3'. The (G) represents the single base insertion in the cAMP-RE relative to the PMA-RE. Klenow polymerase (1 unit), 250 tuCi of [a-32P]- dCTP (3000 Ci/mmol; 1 Ci = 37 GBq), and 50 ,uM (each) dATP, dCTP, and BrdUrd triphosphate were then added to initiate coding strand synthesis. The labeled fragments were purified on 8% polyacrylamide gels.

RESULTS We investigated the phorbol ester responsiveness of the 8-bp cAMP-responsive DNA sequence TGACGTCA. The se-

quences including the 8-bp cAMP-RE and the surrounding bases present in the 5' flanking regions of several genes were synthesized and inserted just upstream of the promoter of the plasmid palOOCAT (3). This vector contains the human CGa transcriptional start site and the proximal promoter consisting of 100 bp of 5' flanking sequence. The plasmid palOOCAT contains no known cAMP-RE or PMA-RE and, when transfected into the placental line JEG-3, has barely detectable transcriptional activity even when 8-bromoadenosine 3',5'- cyclic monophosphate (8-Br-cAMP) (3) and/or the phorbol ester PMA are added (data not shown). Insertion of a synthetic copy of the 18-bp sequence including the 8-bp cAMP-RE (CGa-8) resulted in enhanced basal activity that was stimulated an additional 28-fold when 1 mM 8-Br-cAMP was added for the last 12 hr of the 36-hr transfection (Fig. 1A), similar to effects shown previously after the treatment of the cells with 1 mM 8-Br-cAMP for 24 hr (3).

In contrast to the cAMP responsiveness, there was no significant transcriptional stimulation of the reporter gene containing the CGa-8 element by PMA, either when added alone or with 8-Br-cAMP. Preliminary studies suggest that 12 hr of treatment with 100 nM PMA was maximally effective for stimulating activity using PMA-responsive reporter genes in JEG cells (data not shown). The bases that naturally surround a cAMP-RE consensus sequence in the 5' flanking regions of the rat somatostatin, bovine parathyroid hormone, and rat

INSERTED OLIGONUCLEOTIDE A 0 20 40 60 80 100 (name) (sequence)

CGax8 aaat TGACGTCA tggtaa

COL8 agcttga TGACGTCA gccg AGENT

SMS8 ttggc TGACGTCA gagagaga 14X NONE

29X PM

cAMP

PTH8 ~~~~~~~~~~~~~~~~~~cAMP PTH8 ggag TGACGTCA tctgta lox & PMA -

6X FIG. 1. cAMP and phorbol ester responsive-

22X ness of octamers and heptamers in various con- CORE8 (gatcc) TGACGTCA (agatc) 7 22 X texts. Sequences from the 5' flanking regions of

i |21X( the human gonadotropin (CGa), rat somatostatin (SMS), and bovine parathyroid hormone (PTH)

l I I I l genes, which included the cAMP-RE (A) or the 0 20 40 60 80 1 00 PMA-RE (B), are highlighted in bold, and the

CAT ACTIVITY indicated adjacent bases (lowercase letters) present in the native genes were synthesized as

INSERTED OLIGONUCLEOTIDE U complementary pairs of single-stranded deoxyri- 0 4 8 1 2 16 bonucleotides with terminal linker sequences. (name) (sequence) _ l - The sequences synthesized for B were exactly as

those for A, except that the central cytosine (or CGa7 aaat TGA GTCA tggtaa 3X guanine, on the antisense oligonucleotide) was

AGENT deleted. The COL8 sequence corresponds to a AGENT PMA-responsive oligonucleotide pair derived in

> 5X NONE E part from the collagenase gene (6) with the oblig- CQL7 agcttga TGA GTCA gccg _ 19X PMA 3 atory linkers and an inserted guanine nucleotide cAMP E to reconstitute the 8-bp sequence. The CORE8 CAMP oligonucleotide pair contains the 8-bp consensus

SMS7 ttggc TGA GTCA gagagaga 2X PA sequence alone flanked by the linker sequences. CAT fusion genes containing the indicated synthetic sequence inserted 5' to the promoter region

PTH7 ggag TGA GTCA tctgta 8X of palOOCAT were transfected into JEG-3 cells. ggag TGA GTCA tctgta Medium alone or PMA (100 nM) and/or 8-Br- 34X cAMP (cAMP, 1 mM) was added 24 hr after

CORE7 (gatcc) TGA GTCA (agatc) 2X transfection. Thirty-six hours after transfection, CORE7 (gatcc) TGA GTCA (agatc) L2X 3X cells were harvested and CAT activity was deter- _ _5X _ mined, expressed as units/Al of extract. Values

represent the average of three (where standard o 4 8 1 2 1 6 error bars are shown) or two independent trans-

CAT ACTIVITY fections.



7924 Biochemistry: Deutsch et al. Proc. Natl. Acad. Sci. USA 85 (1988)

glucagon genes were found to have a marked impact on cAMP responsiveness mediated by the cAMP-RE; the parathyroid hormone (Fig. 1A) and glucagon (ref. 35; P.J.D., unpublished data) contextual sequences largely negated the activity of the core 8-bp sequence. Thus, we investigated whether the cAMP-RE might exhibit PMA responsiveness when inserted in these various contexts or when the core cAMP-RE alone was inserted adjacent to the obligatory restriction enzyme linkers. None of these oligonucleotides, whether the context was permissive for cAMP responsiveness or not, conferred any detectable phorbol ester responsivity (Fig. 1). We also synthesized an oligonucleotide es- sentially identical to one previously shown to confer PMA- responsive activity on a heterologous promoter (6) except a central guanine nucleotide was inserted, reconstituting the octamer from the heptamer. This oligonucleotide, COL-8, includes several of the adjacent bases in the collagenase gene and an internal HindIII as well as Bgl II and BamHI sites to precisely reconstitute the 10 contextual bases that were shown to be permissive for the PMA-responsive activity of the heptameric PMA-RE. This sequence conferred consid- erable cAMP-responsive activity, but, as was observed with the other octameric cAMP-REs, it was not responsive to phorbol esters.

We wondered whether other contextual sequences would have parallel permissive or restrictive effects on the cAMP- RE and PMA-RE. We synthesized heptamer versions of each of the oligonucleotides cassettes displayed in Fig. 1A by omitting the central guanine in the synthetic sequences. These oligonucleotides were similarly inserted into the palO0- CAT vector and expressed in JEG-3 cells to test the relative cAMP and PMA responsiveness of the PMA-RE in the various contexts. Oligonucleotides COL-7 and PTH-7, containing the AP-1 consensus heptamer flanked by the COL and parathyroid hormone contexts, conferred 5- to 8-fold responsiveness to PMA (Fig. 1B). The PMA-RE was not able to confer significant PMA responsiveness in any of the other contexts, including CGa, somatostatin, glucagon, or the obligatory restriction enzyme linkers alone. These observa- tions indicate that deletion of one base can potentially convert a plasmid that responds by 11-fold to cAMP and not at all to PMA (e.g., COL-8) to one (COL-7) with diminished response to cAMP but that responds 5-fold to PMA. How- ever, for cAMP-RE and PMA-RE, certain adjacent contextual sequences negate the potential latent transcriptional activity of the core consensus sequence. Moreover, the restrictive contextual sequences vary between the cAMP-RE and PMA-RE. By comparing CGa-8 (Fig. LA) with the CGa-7 (Fig. 1B), it is clear that the native contextual bases in CGa allow the 8-bp sequence to function as a cAMP-RE, but the PMA-RE is inactive in the same context. In contrast, the glucagon context prohibits activity of both sequences (ref. 35; P.J.D., unpublished data), whereas the parathyroid hormone context is less permissive for activity of the cAMP-RE (Fig. 1A) sequence than the PMA-RE (Fig. 1B).

Several of the PMA-RE-containing oligonucleotides conferred 5- to 8-fold responsiveness to either cAMP or PMA, including the PMA-RE flanked by a few bases from the COL gene (Fig. 1A). Thus, the PMA-RE heptamer does not act as

a pure PMA-RE as had been shown previously (13) but, at least when paired with the CGa promoter in JEG-3 cells, serves as a cAMP-RE as well. The PMA-RE, however, is a weaker cAMP responder than is the cAMP-RE version (Fig. lB vs. Fig. 1A), especially if one compares the absolute levels of CAT activity after cAMP stimulation rather than only the fold stimulation. The contexts that permit the PMA-RE to function in a cAMP-responsive manner are precisely the ones that allow for PMA responsiveness (see COL-7 and PTH-7 in Fig. 1B). On the other hand, several sequences that are permissive for cAMP-responsive activity of the cAMP-RE (CGa, somatostatin, and the linkers alone in Fig. 1A) are not permissive for the cAMP responsiveness of the PMA-RE, suggesting that different proteins are involved in the cAMP responsiveness of the 7- and 8-bp sequences.

We added 8-Br-cAMP and PMA together to see if the effects were different from the addition of either compound alone. PMA had no significant effect on any of the cAMP- RE-containing plasmids, whether added by itself or with cAMP (Fig. 1A). However, cAMP and PMA added together clearly produced supraadditive effects on the activity of the PMA-REs (Fig. 1B). Certain of the sequences that conferred no statistically significant response to 8-Br-cAMP or to PMA alone gave detectable responses to the two agents added together (Fig. 1B, see CGa-7, SMS-7, and CORE-7). Both 8-Br-cAMP and PMA were added in all cases at concentrations that clearly gave maximal responses. Thus, PMA and 8-Br-cAMP act synergistically to stimulate transcription, a finding that indicates a convergence of actions somewhere in their respective pathways of gene activation.

Because the capability of PMA and cAMP to stimulate gene transcription from the simian virus 40 (SV40) and c-fos promoters, respectively, is dependent on cell type (14, 33), we wondered whether the signal-responsive properties of the PMA-REs and cAMP-REs studied in this investigation would vary with the cell employed for expression. To compare with JEG-3 cells, we chose the hepatic-derived Hep G-2 cell line, in which more dramatic responsiveness of the PMA-RE to PMA had been previously reported (6). The 7-bp and 8-bp motifs in the COL contexts, inserted in the palOOCAT reporter gene, actually functioned quite similarly in Hep G-2 cells (Table 1) compared with JEG-3 cells (Fig. 1 A and B). In Hep G-2 cells, as in JEG-3 cells, the COL-8 sequence acted as a cAMP-RE but conferred no responsiveness to PMA, and the COL-7 sequence conferred 3- to 4-fold cAMP responsiveness and PMA responsiveness. The COL-7-containing plasmids did have a 9-fold higher level of basal expression than did the COL-8 sequence in Hep G-2 cells, the converse of the situation in JEG-3 cells. But the general phenomenon of the 7-bp motif conferring equal PMA and cAMP responsiveness, although less cAMP responsiveness than the 8-bp sequence, occurs in the Hep G-2 and JEG-3 cell lines. The PMA responsiveness of the 7-bp motif is not being signifi- cantly blunted by performance of these experiments in 10% fetal calf serum, because, although serurn can down-regulate the PMA-induced protein kinase C pathway, repeating the experiments in 0.5% fetal calf serum reduces basal and signal-induced CAT activity proportionally and does not affect fold induction (data not shown). Moreover, the CGa

Table 1. Transcriptional activity of the 7-bp and 8-bp signal-responsive motifs in Hep G-2 cells Inserted oligonucleotide CAT activity, units/Al of extract

Name Sequence Basal + cAMP + PMA COL-7 agcttga TGA-GTCA gccg 3.7 ? 1.7 13.6 ? 0.4 12.4 ? 1.5 COL-8 agcttga TGACGTCA gccg 0.4 ? 0.05 3.4 ? 0.8 0.3 ? 0.2

The experiment was performed as described in the legend to Fig. 1 except that Hep G-2 cells were transfected, requiring a 4-hr glycerol shock (22). Bases in lowercase letters are those of the COL context.



Biochemistry: Deutsch et al. Proc. Natl. Acad. Sci. USA 85 (1988) 7925

sequence from - 100 to +44 in the parent vector does not seem to contain a latent cAMP-RE because when the SV40 enhancer is inserted just upstream of CGa-100 (J.L.J., unpublished observation) or CGa-118 (4), the resultant reporter plasmids exhibit no responsiveness to cAMP in JEG-3 cells. Thus, the AP-1 binding 7-bp motif, like the AP-2 enhancer (13), appears, at least when paired with the CGa promoter, to function as a (submaximal) cAMP-RE as well as a PMA-RE.

The similar sizes on NaDodSO4/PAGE of the proteins AP-1 (7, 8) and CREB (9), and the related signal responsiveness of the 8-bp and 7-bp motifs, prompted us to investigate interactions of DNA-binding proteins with the phorbol ester and cAMP-responsive DNA sequences. Gel-mobility-shift and photoaffinity cross-linking techniques were used to eluci- date the interactions of DNA-binding proteins with these signal transduction DNA elements. Whole cell extracts were prepared from JEG-3 cells, allowed to react with 32P-labeled oligonucleotide DNAs, and subjected to polyacrylamide gel electrophoresis on 6% nondenaturing gels (Fig. 2A). By utilizing separate unlabeled cAMP-RE or PMA-RE oligonucleotides as competitors, the specificity of the binding was determined. Binding to the labeled cAMP-RE was effectively inhibited by the unlabeled cAMP-RE-containing sequence but was unaffected by a 100-fold molar excess of PMA-RE- containing competitor DNA. These results corroborated the studies of transcriptional expression and suggested that separate trans-acting factors interacted with the two motifs.

COMP A w

o~~~~~~~~~~~~~~~7 L z CE~~~~~~LC

-O c c

0 : :

f-Brd~ ~~ ~~ 00000 iSf:S SSDt0X

o 0 0- COMP COMP

-8B4K

j ~~ ~ ~~ 1 i iOK

Labeled Labeled Labeled cAMP-RE cAMP-RE PMA-RE

FIG. 2. (A) Gel-mobility-shift analysis of DNA-binding proteins in JEG-3 extracts that specifically interact with cAMP- and PMA- responsive DNA elements. 32P-end-labeled double-stranded oligonucleotides (3 fmol, 3000 cpm/fmol) containing the cAMP-RE (COL-8) were incubated for 15 min with whole cell extracts (15 ,g of protein) of JEG-3 cells in 20 ,lI of binding buffer and were analyzed by electrophoresis on 6% nondenaturing polyacrylamide gels (25, 26). B, bound probe; F, free probe; COMP, competitor. (B) UV cross-linking of cAMP-RE- and PMA-RE-binding proteins. Experi- ments were performed with the protocol of Chodosh et al. (27), in which BrdUrd and [a-32P]dCTP were incorporated into the DNA probe by primed cDNA synthesis in M13. The labeled DNA (104 cpm) was incubated with 15 ,utg of total protein from JEG-3 whole cell extracts under the conditions used for gel-mobility-shift experiments followed by cross-linking with UV light and electrophoresis of products on NaDodSO4/PAGE (27).

The JEG-3 cell proteins that bind to these two distinct DNA elements were further characterized by cross-linking them to DNA probes containing either the cAMP-RE or PMA-RE (Fig. 2B) and resolving the labeled proteins by employing NaDodSO4/PAGE. By utilizing the cAMP-RE- containing sequence, binding of a single protein of 38 kDa was detected. Binding of the protein was inhibited specifically by a 100-fold molar excess of homologous unlabeled oligonucleotide but not by a similar amount of the oligonucleotide containing the PMA-RE.

When the labeled PMA-RE-containing probe was irradi- ated in the presence of JEG-3 whole cell extract, three distinct bands representing proteins of approximately 48, 50, and 84 kDa were noted. The predominant species were proteins of 48 and 50 kDa. Addition of a 100-fold molar excess of homologous unlabeled PMA-RE-oligonucleotide pre- vented the cross-linking of these proteins, whereas a similar amount of cAMP-RE-containing oligonucleotide had no effect on this binding. The PMA-RE-binding proteins identified in the present study appear to correspond to three of the proteins that were observed to bind to the sequence-specific affinity column utilized by Lee et al. (8) to purify AP-1. There is also a less distinct band(s) present in the 30- to 40-kDa range that specifically binds to the PMA-RE but has not consistently been seen in repeated experiments. Thus, it appears that a single protein of 38 kDa specifically binds to a cAMP response-conferring octamer sequence, whereas several proteins ranging from 48 to 84 kDa consistently bind to a PMA- (and cAMP-) response-conferring heptamer. The various proteins may be products of different genes or, alternatively, distinct posttranslationally modified versions of a single gene product. Clearly, however, the two motifs differing by 1 bp exhibit. quite different protein-binding properties consistent with their different functional properties.

DISCUSSION The assignment of a consensus sequence to a given set of transcription-stimulating and protein-binding properties of DNA sequences requires somewhat arbitrary decisions in assimilating data from multiple sources, as exemplified by the formulation of slightly different consensus rules from similar data (7, 8). Ideally, one would like there to be absolute rules. The nature, however, of transcriptionally productive interactions of eukaryotic DNA-binding proteins with DNA sequences appears to elude the assignment of simple restriction enzyme-like rules. The 8-bp sequence TGACGTCA will confer cAMP responsiveness, but only when flanked by certain contextual bases. Similarly, the PMA responsiveness conferred by the 7-bp sequence TGAGTCA depends on the DNA context but no simple consensus rule can be derived for the base immediately 5' or 3' to the heptamer in the sequences that impart PMA responsivity. Thus, at our current state of knowledge, one cannot predict with certainty the signal- responsive properties of a gene by mere inspection of its sequence. The presence of a given motif is necessary but may not be sufficient to constitute a transcriptionally responsive element.

Despite the complexities of the contextual rules, the functional properties of the octamer sequences versus the heptamer sequences clearly differed from each other, and the active members of each motif-containing group were quali- tatively similar. In the cell types examined, the active octamer-containing sequences were cAMP-responsive but not PMA-responsive. However, the active heptamers were capable of conferring cAMP and PMA responsiveness. The DNA-binding protein AP-2 that recognizes a quite different DNA motif is another protein that mediates gene activation by cAMP and phorbol esters (13). The version of the 7-bp



7926 Biochemistry: Deutsch et al. Proc. Natl. Acad. Sci. USA 85 (1988)

AP-1-binding motif present in the SV40 enhancer was previously shown to confer PMA but not cAMP responsiveness to the f-globin promoter in the HeLa TK- cells (13) in contrast to the cAMP and PMA responsiveness of the AP-2 motif in the same study and to the results of the current investigation with the SV40 motif. Many possibilities exist to explain the apparent discrepancy. The functional properties of the PMA-RE enhancer may depend on a combination of the promoter and the cell type. As a precedent, the 8-bp cAMP-RE acts as a 100-fold basal enhancer when paired with the CGa promoter but not with the SV40 promoter when expressed in JEG-3 cells and not with either promoter in BHK cells (3). These kinds of properties suggest cooperative interaction between the protein that binds to the enhancer and a protein(s) that binds to specific downstream promoter elements in which the latter protein may not be expressed in all cell types. The results of the current study suggest that the CGa promoter is indeed very receptive to cAMP-REs and PMA-REs in Hep G-2 as well as JEG-3 cells, perhaps consistent with the frequently observed ectopic expression of CGa in liver carcinomas (28).

The DNA elements involved in responses to the cellular signaling agents do not subdivide neatly into pure responders to one or the other agent. This situation differs from that for the estrogen- and glucocorticoid-responsive DNA elements, both of which have related sequences; specific synthetic sequences were shown to respond to either but not both hormones (29). The DNA-binding proteins that mediate steroid hormone responsiveness represent domains of the hormone receptor for the steroid hormones characterized to date (30-32). On the other hand, the characterized DNA- binding proteins mediating response to membrane receptor- interactive peptides and their corresponding second messen- gers have been distinct from and presumably distal to the receptors and other previously identified components of signaling cascades (7-9, 33), except for one report that the regulatory subunit of cAMP-dependent protein kinase type II is a DNA-binding topoisomerase (34). This fundamental difference between steroid hormones and peptide hormones may relate to the different specificities of their corresponding DNA-binding proteins.

The synergistic response of PMA and cAMP on the 7-bp motif but not the 8-bp motif implies that the two signal transduction pathways converge at one of several levels, some of which have precedents.

(i) Activation of protein kinase C (e.g., by phorbol esters) enhances cAMP accumulation (15-17) perhaps by way of phosphorylation of an inhibitory GTP-binding protein (18), phosphorylation of a stimulatory GTP-binding protein (16), or phosphorylation of the catalytic subunit of adenylate cyclase (19); (ii) cAMP activates protein kinase C or trans- locates it to the nucleus (20); (iii) cAMP-dependent protein kinase and protein kinase C phosphorylate and activate the same protein-e.g., AP-1; (iv) different proteins at the end of cAMP-dependent protein kinase A- and protein kinase C- mediated phosphorylation cascades (e.g., CREB and AP-1) bind to the 7-bp motif. The ability of cAMP alone to also activate the 7-bp motif but not for PMA to activate the 8-bp motif favors mechanism i, ii, or iii. The finding in the current study that the 7-bp and 8-bp motifs appear to bind different proteins (Fig. 2 A and B) favors model iv, although distinctly phosphorylated versions of the same protein may have quite different molecular masses. Clearly, these two potent DNA motifs differing by 1 bp have quite different transcriptional and protein-binding properties but share the property of cAMP responsiveness, reflecting the complex interactions between these parallel signal transduction pathways.

We thank Julia Lin, Frank Traynor, and Chris Albanese for expert experimental assistance, Janice Canniff for preparation of the manu- script, and Michael Kopczynski and Phil Behn for synthesis of oligonucleotides. This work was supported in part by National Institutes of Health Grant DK-25532.

1. Montminy, M. R., Sevarino, K. A., Wagner, J. A., Mandel, G. & Goodman, R. H. (1986) Proc. Nat!. Acad. Sci. USA 83, 6682-6686.

2. Silver, B. J., Bokar, J. A., Virgin, J. B., Vallen, E. A., Mil- sted, A. & Nilson, J. H. (1987) Proc. Nat!. Acad. Sci. USA 84, 2198-2202.

3. Deutsch, P. J., Jameson, J. L. & Habener, J. F. (1987) J. Biol. Chem. 262, 12169-12174.

4. Delegeane, A. M., Ferland, L. H. & Mellon, P. L. (1987) Mol. Cell. Biol. 7, 3994-4002.

5. Tsukada, T., Fink, J. S., Mandel, G. & Goodman, R. H. (1987) J. Biol. Chem. 262, 8743-8747.

6. Angel, P., Baumann, I., Stein, B., Delins, H., Rahmsdorf, H. J. & Herrlich, P. (1987) Mol. Cell. Biol. 7, 2256-2266.

7. Angel, P., Imagawa, M., Chiu, R., Stein, B., Imbra, R. J., Rahmsdorf, H. J., Jonat, L., Herrlich, P. & Karin, M. (1987) Cell 49, 729-739.

8. Lee, W., Mitchell, P. & Tjian, R. (1987) Cell 49, 741-752. 9. Montminy, M. R. & Bilezikjian, L. M. (1987) Nature (London)

328, 175-178. 10. Rubin, C. S. & Rosen, 0. M. (1975) Annu. Rev. Biochem. 4,

831-887. 11. Grove, J. R., Price, D. J., Goodman, H. M. & Avruch, J. A.

(1987) Science 238, 530-533. 12. Kikkawa, U., Takai, Y., Tanaka, Y., Miyuke, R. & Nishizuka,

Y. (1983) J. Biol. Chem. 258, 11442-11445. 13. Imagawa, M., Chiu, R. & Karin, M. (1987) Cell 51, 251-260. 14. Bravo, R., Neuberg, M., Burckhardt, J., Almendral, J., Wal-

lich, R. & Muller, R. (1987) Cell 48, 251-260. 15. Sudgen, D., Vanecek, J., Klein, D. C., Thomas, T. P. &

Anderson, W. B. (1985) Nature (London) 314, 359-361. 16. Rozengurt, E., Murray, M., Zachary, I. & Collins, M. (1987)

Proc. Natl. Acad. Sci. USA 84, 2282-2286. 17. Ritvos, O., Jalkanen, J., Huhtaniem, I., Stenman, Y.-H.,

Alfthan, H. & Ranta, T. (1987) Endocrinology 120, 1521-1526. 18. Katada, T., Gilman, A. G., Watanabe, Y., Buuer, S. & Jakobs,

K. H. (1985) Eur. J. Biochem. 151, 431-437. 19. Yoshimasa, T., Sibley, D. R., Bouvier, M., Lefkowitz, R. J. &

Caron, M. G. (1987) Nature (London) 327, 67-70. 20. Cambier, J. L., Newell, M. K., Justement, L. B., McGuire,

J. C., Leach, K. L. & Chen, Z. Z. (1987) Nature (London) 327, 629-632.

21. Graham, F. & van der Eb, A. (1973) Virology 52, 456-457. 22. Lopata, M. A., Cleveland, D. W. & Sollner-Webb, B. (1984)

Nucleic Acids Res. 12, 5707-5717. 23. Manley, J. L., Fire, A., Cano, A., Sharp, P. A. & Gefter,

M. L. (1980) Proc. Natl. Acad. Sci. USA 77, 5706-5710. 24. Bradford, M. M. (1976) Anal. Biochem. 72, 248-254. 25. Jameson, J. L., Jaffe, R. C., Deutsch, P. J., Albanese, C. &

Habener, J. F. (1988) J. Biol. Chem. 263, 9879-9886. 26. Singh, H., Sen, R., Baltimore, D. & Sharp, P. A. (1986) Nature

(London) 319, 154-158. 27. Chodosh, L. A., Carthew, R. W. & Sharp, P. A. (1986) Mol.

Cell. Biol. 6, 4723-4733. 28. Braunstein, G. D., Bridson, W. E., Glass, A., Hull, E. W. &

Mclntire, K. R. (1972)J. Clin. Endocrinol. Metab. 35,857-862. 29. Klock, G., Strahle, U. & Schutz, G. (1987) Nature (London)

329, 734-736. 30. Giguere, V., Hollenberg, S. M., Rosenfeld, M. G. & Evans,

R. M. (1986) Cell 46, 645-652. 31. Kumar, V., Green, S., Staub, A. & Chambon, P. (1986) EMBO

J. 5, 2231-2236. 32. Glass, C. K., Franco, R., Weinberger, C., Albert, V. R.,

Evans, R. M. & Rosenfeld, M. G. (1987) Nature (London) 329, 738-741.

33. Chiu, R., Imagawa, M., Imbra, R. J., Bockoven, J. R. & Karin, M. (1987) Nature (London) 329, 648-651.

34. Constantinou, A., Squinto, S. & Jungmann, R. A. (1985) Cell 42, 429-437.

35. Deutsch, P. J., Hoeffler, J. P., Jameson, J. L., Lin, J. C. & Habener, J. F. (1988) J. Biol. Chem., in press.



cyclic amp and phorbol ester-stimulated transcription mediated by similar dna elements that bind...

Documents