Sequence Determinants of the Intrinsic Bend in the Cyclic AMP Response Element†Leslie S. Sloan and Alanna Schepartz*Department of Chemistry, Yale UniVersity, New HaVen, Connecticut 06520-8107ReceiVed August 13, 1997; ReVised Manuscript ReceiVed January 6, 1998ABSTRACT: The cyclic AMP response element (CRE site, ATGACGTCAT) is the DNA target fortranscription factors whose activities are regulated by cyclic AMP (1). Recently, we discovered that theCRE site is bent by 10-13° toward the major groove (2). Little or no bend is detected in the relatedAP-1 site (ATGACTCAT), which differs from the CRE site by loss of a single, central, C‚G base pair (2,3). Here we describe experiments designed to identify which base pairs within the CRE site induce thebent structure in an attempt to understand the origins of the dramatically different conformations of theCRE and AP-1 sites. Our data indicate that the intrinsic CRE bend results from distortion within theTGA sequence found in each CRE half site (ATGAC). These two TGA sequences are located in phasewith one another in the CRE sequence but are not (completely) in phase in the AP-1 sequence. Thisdifference in phasing leads to the overall difference in bend as detected by gel (2) and cyclization methods(S. C. Hockings, J. D. Kahn, and D. M. Crothers, unpublished results; M. A. Fabian and A. Schepartz,unpublished results). Our results confirm earlier predictions of altered structure within TG steps, provideinsight into the structural reorganizations induced in DNA by bZIP proteins, and lead to a revision of therelationship between the structures of the free and bZIP-bound forms of the CRE and AP-1 sites.The possibility that certain DNA molecules might possessB-form structures that differed from that of the Watson-Crick model (4) was recognized by the discovery of bentDNA in 1982 (5). It was discovered that the kinetoplastminicircles of the Leishmania tarentolae parasite containedtracts of four to five dA‚dT base pairs (called A-tracts)repeated in phase with the DNA helical repeat (6, 7).Fragments of these minicircles displayed anomalously lowmobilities in nondenaturing gels when compared withreference DNA sequences of equivalent length. Subsequentexperimentation revealed that the anomalous migration ofminicircle DNA resulted from a bend of approximately 17°associated with each A-tract (8). When located along thesame face of the DNA double helix, the bends produced byeach A-tract added constructively to generate a DNAfragment with a significant overall bend. A-tract bendinghas been studied extensively (for reviews, see refs 9-13)since its discovery because of the significance of DNAstructure and structural dynamics for at least three vitalbiological functions: transcriptional regulation (14), controlof replication and recombination (15, 16), and the packagingof DNA into nucleosomes (10, 17, 18). Although a numberof intrinsically bent DNA sequences unrelated to A-tractshave been identified (14, 19-26), the magnitudes of thebends in these sequences are, with two exceptions (24, 26),small relative to the bend produced by a single A-tract (∼17°)(8).Recently, we discovered that the cyclic AMP responseelement (CRE site, ATGACGTCAT), the DNA target fortranscription factors whose activities are regulated by cyclicAMP (1), is bent significantly in solution (2). The CRE siteconsists of two ATGAC half-sites arranged in an invertedrepeat and is unrelated to A-tract DNA. The magnitude ofthe intrinsic CRE bend was estimated at between 10° and13° toward the major groove (2). Thus the CRE bend issimilar in magnitude to that of a single A-tract, but the bendis in the opposite direction (8). Although the bend in theCRE site was detected initially by use of gel electrophoreticmethods (27), its existence has been confirmed subsequentlyby analyzing the rates of minicircle ligation reactions(28, 51) (M. A. Fabian and A. Schepartz, unpublished results)as well as by X-ray crystallography [see reference in Paolellaet al. (2)]. Here we employ a related gel electrophoreticmethod to analyze the extent of bending in a series ofoligonucleotides related to the CRE site. Our data indicatethat the intrinsic CRE bend (2) is the additive result of twomajor groove bends that are each found within the TGAsequence in each CRE half-site (ATGAC). These two TGAsequences are located in phase with one another in thesequence ATGACGTCAT but are not (completely) in phasein the related AP-1 sequence (ATGACTCAT). We proposethat this difference in phasing leads to the overall differencein bend as detected by gel-based methods.MATERIALS AND METHODSMaterials. Enzymes were purchased from New EnglandBiolabs and used with the buffers supplied. Adenosine 5′-[γ-32P]triphosphate was purchased from DuPont. Adenosine5′-triphosphate monomagnesium salt was purchased fromSigma. Phosphoramidites and other reagents employed insolid-phase oligonucleotide synthesis were purchased fromPerceptive Biosystems.†This work was supported by the NIH (Grant GM 52544). L.S.S.was supported by an NSF predoctoral fellowship.* Correspondence should be addressed to this author [email protected] 1998, 37, 7113-7118S0006-2960(97)02009-6 CCC: $15.00 © 1998 American Chemical SocietyPublished on Web 05/02/1998Oligonucleotide Synthesis. Oligonucleotides were syn-thesized on a 0.2 µmol scale with a Millipore Expedite DNA/RNA Synthesizer Model 8909 and standard phosphoramiditechemistry (29) and were purified by preparative denaturinggel electrophoresis. DNA-containing regions of the gel wereidentified by shadowing (30) and excised, and the oligo-nucleotides were eluted into TE buffer [90 mM Tris (pH8.0), 1 mM EDTA] at room temperature for at least 10 h.Concentrated eluents were dialyzed against 1 mM Tris (pH7.4) for 24 h at 25 °C and stored at -20 °C. Oligonucle-otides were labeled on the 5′ end with T4 polynucleotidekinase and [γ-32P]ATP (31), annealed to their complementarystrands by heating to 70 °C followed by slow cooling toroom temperature, and used directly in ligation ladderexperiments.Ligation Ladder Experiments.32P end-labeled oligonucle-otide duplexes (800 pmol) prepared as described above weretreated with T4 DNA ligase (800 units) in 70 µL of reactionbuffer [50 mM Tris‚HCl (pH 7.8), 10 mM MgCl2,10mMDTT, 4 mM ATP, 50 µg/mL BSA] for 90 min at 16 °C.The reaction mixtures were washed with phenol/chloroform(31) and precipitated with ethanol. The DNA was dried