Crystal structure of a bZIP/DNA complex at 2.2 A: determinants of DNA specific recognition

Electrostatic control of half-site spacing preferences by the cyclic AMP response element-binding protein CREB

Basic region leucine zipper (bZIP) proteins represent a class of transcription factors that bind DNA using a simple, dimeric, alpha-helical recognition motif. The cAMP response element-binding protein (CREB) is a member of the CREB/ATF subfamily of bZIP proteins. CREB discriminates effectively in vivo and in vitro between the 10 bp cAMP response element (ATGACGTCAT, CRE) and the 9 bp activating protein 1 site (ATGACTCAT, AP-1). Here we describe an alanine scanning mutagenesis study designed to identify those residues within the CREB bZIP element that control CRE/AP-1 specificity. We find that the preference of CREB for the CRE site is controlled in a positive and negative way by acidic and basic residues in the basic, spacer and zipper segments. The CRE/AP-1 specificity of CREB is increased significantly by four glutamic acid residues located at positions 24, 28, 35 and 41; glutamic acid residues at positions 10 and 48 contribute in a more modest way. Specificity is decreased significantly by two basic residues located at positions 21 and 23; basic residues at positions 14, 18, 33 and 34 and V17 contribute in a more modest way. All of the residues that influence specificity significantly are located on the solvent-exposed face of the protein-DNA complex and likely participate in interactions between and among proteins, not between protein and DNA. The finding that the CRE/AP-1 specificity of CREB is dictated by the presence or absence of charged residues has interesting implications for how transcription factors seek and selectively bind sequences within genomic DNA.

DNA-binding specificity and dimerization of the DNA-binding domain of the PEND protein in the chloroplast envelope membrane

The PEND protein is a DNA-binding protein in the inner envelope membrane of a developing chloroplast, which may anchor chloroplast nucleoids. Here we report the DNA-binding characteristics of the N-terminal basic region plus leucine zipper (bZIP)-like domain of the PEND protein that we call cbZIP domain. The basic region of the cbZIP domain diverges significantly from the basic region of known bZIP proteins that contain a bipartite nuclear localization signal. However, the cbZIP domain has the ability to dimerize in vitro. Selection of binding sites from a random sequence pool indicated that the cbZIP domain preferentially binds to a canonical sequence, TAAGAAGT. The binding site was also confirmed by gel mobility shift analysis using a representative binding site within the chloroplast DNA. These results suggest that the cbZIP domain is a unique DNA-binding domain of the chloroplast protein.

Consensus and variant cAMP-regulated enhancers have distinct CREB-binding properties

Recent determination of the cAMP response element-binding protein (CREB) basic leucine zipper (bZIP) consensus CRE crystal structure revealed key dimerization and DNA binding features that are conserved among members of the CREB/CREM/ATF-1 family of transcription factors. Dimerization appeared to be mediated by a Tyr(307)-Glu(312) interhelical hydrogen bond and a Glu(319)-Arg(314) electrostatic interaction. An unexpected hexahydrated Mg(2+) ion was centered above the CRE in the dimer cavity. In the present study, we related these features to CREB dimerization and DNA binding. A Y307F substitution reduced dimer stability and DNA binding affinity, whereas a Y307R mutation produced a stabilizing effect. Mutation of Glu(319) to Ala or Lys attenuated dimerization and DNA binding. Mg(2+) ions enhanced the binding affinity of wild-type CREB to the palindromic CRE by approximately 20-fold but did not do so for divergent CREs. Similarly, mutation of Lys(304), which mediates the CREB interaction with the hydrated Mg(2+), blocked CREB binding to the palindromic but not the variant CRE sequences. The distinct binding characteristics of the K304A mutants to the consensus and variant CRE sequences indicate that CREB binding to these elements is differentially regulated by Mg(2+) ions. We suggest that CREB binds the consensus and variant CRE sequences through fundamentally distinct mechanisms.

Stabilization of coiled-coil peptide domains by introduction of trifluoroleucine

Substitution of leucine residues by 5,5,5-trifluoroleucine at the d-positions of the leucine zipper peptide GCN4-p1d increases the thermal stability of the coiled-coil structure. The midpoint thermal unfolding temperature of the fluorinated peptide is elevated by 13 degrees C at 30 microM peptide concentration. The modified peptide is more resistant to chaotropic denaturants, and the free energy of folding of the fluorinated peptide is 0.5-1.2 kcal/mol larger than that of the hydrogenated form. A similarly fluorinated form of the DNA-binding peptide GCN4-bZip binds to target DNA sequences with affinity and specificity identical to those of the hydrogenated form, while demonstrating enhanced thermal stability. Molecular dynamics simulation on the fluorinated GCN4-p1d peptide using the Surface Generalized Born implicit solvation model revealed that the coiled-coil binding energy is 55% more favorable upon fluorination. These results suggest that fluorination of hydrophobic substructures in peptides and proteins may provide new means of increasing protein stability, enhancing protein assembly, and strengthening receptor-ligand interactions.

Specific down-regulation of an engineered human cyclin D1 promoter by a novel DNA-binding ligand in intact cells

Cyclin D1 is expressed at abnormally high levels in many cancers and has been specifically implicated in the development of breast cancer. In this report we have extensively analyzed the cyclin D1 promoter in a variety of cancer cell lines that overexpress the protein and identified two critical regulatory elements (CREs), a previously identified CRE at -52 and a novel site at -30. In vivo footprinting experiments demonstrated factors binding at both sites. We have used a novel DNA-binding ligand, GL020924, to target the site at -30 (-30-21) of the cyclin D1 promoter in MCF7 breast cancer cells. A binding site for this novel molecule was constructed by mutating 2 bp of the wild-type cyclin D1 promoter at the -30-21 site. Treatment with GL020924 specifically inhibited expression of the targeted cyclin D1 promoter construct in MCF7 cells in a concentration-dependent manner, thus validating the -30-21 site as a target for minor groove-binding ligands. In addition, this result validates our approach to regulating the expression of genes implicated in disease by targeting small DNA-binding ligands to key regulatory elements in the promoters of those genes.

Long-range electrostatic interactions influence the orientation of Fos-Jun binding at AP-1 sites

Heterodimeric transcription regulatory proteins that bind palindromic DNA sequences can potentially bind their recognition sites in two opposite orientations. The orientation of transcription factor binding can control transcriptional activity by altering interactions with proteins that bind to adjacent regulatory elements. Fos-Jun heterodimers bind to AP-1 sites with different flanking sequences in opposite orientations. A gel-based fluorescence resonance energy transfer assay, gelFRET, was used to define the mechanism whereby amino acid residues and nucleotide base-pairs outside the Fos-Jun-AP-1 contact interface determine the orientation of heterodimer binding. Exchange of three amino acid residues adjacent to the basic DNA contact regions between Fos and Jun reversed the binding orientation. The effects of these amino acid residues on the orientation of heterodimer binding depended on base-pairs flanking the core AP-1 recognition sequence. Single amino acid and base-pair substitutions had parallel effects on DNA bending by Fos-Jun-AP-1 complexes and on heterodimer orientation. The binding orientation exhibited a close correspondence with both the difference in bending propensities of opposite sides of the AP-1 site as well as the difference in bending potentials of the Fos and Jun subunits of the heterodimer. The influence of flanking DNA sequences on heterodimer orientation was attenuated in the presence of high concentrations of multivalent cations. Base substitutions up to one helical turn from the center of the AP-1 site affected the binding orientation. Modification of flanking base-pairs with positively or negatively charged functional groups had opposite effects on the orientation of heterodimer binding. These changes in DNA charge had converse effects on the orientation preferences of heterodimers in which charged amino acid residues adjacent to the basic regions were exchanged between Fos and Jun. These results indicate that the orientation of heterodimer binding is determined primarily by minimization of the electrostatic free energy of the Fos-Jun-AP-1 complex. Consequently, long-range electrostatic interactions influence the architecture of nucleoprotein complexes.

Solution structure and dynamics of GCN4 cognate DNA: NMR investigations

A 12 bp long GCN4-binding, self-complementary duplex DNA d(CATGACGTCATG)(2) has been investigated by NMR spectroscopy to study the structure and dynamics of the molecule in aqueous solution. The NMR structure of the DNA obtained using simulated annealing and iterative relaxation matrix calculations compares quite closely with the X-ray structure of ATF/CREB DNA in complex with GCN4 protein (DNA-binding domain). The DNA is also seen to be curved in the free state and this has a significant bearing on recognition by the protein. The dynamic characteristics of the molecule have been studied by (13)C relaxation measurements at natural abundance. A correlation has been observed between sequence-dependent dynamics and recognition by GCN4 protein.

Crystal structure of the CCAAT box/enhancer-binding protein beta activating transcription factor-4 basic leucine zipper heterodimer in the absence of DNA

The crystal structure of the heterodimer formed by the basic leucine zipper (bZIP) domains of activating transcription factor-4 (ATF4) and CCAAT box/enhancer-binding protein beta (C/EBP beta), from two different bZIP transcription factor families, has been determined and refined to 2.6 A. The structure shows that the heterodimer forms an asymmetric coiled-coil. Even in the absence of DNA, the basic region of ATF4 forms a continuous alpha-helix, but the basic region of C/EBP beta is disordered. Proteolysis, electrophoretic mobility shift assay, circular dichroism, and NMR analyses indicated that (i) the bZIP domain of ATF4 is a disordered monomer and forms a homodimer upon binding to the DNA target; (ii) the bZIP domain of ATF4 forms a heterodimer with the bZIP domain of C/EBP beta that binds the cAMP response element, but not CCAAT box DNA, with high affinity; and (iii) the basic region of ATF4 has a higher alpha-helical propensity than that of C/EBP beta. These results suggest that the degree of ordering of the basic region and the fork and the dimerization properties of the leucine zipper combine to distinguish the structurally similar bZIP domains of ATF4 and C/EBP beta with respect to DNA target sequence. This study provides insight into the mechanism by which dimeric bZIP transcription factors discriminate between closely related but distinct DNA targets.

The structure of a CREB bZIP.somatostatin CRE complex reveals the basis for selective dimerization and divalent cation-enhanced DNA binding

The cAMP responsive element-binding protein (CREB) is central to second messenger regulated transcription. To elucidate the structural mechanisms of DNA binding and selective dimerization of CREB, we determined to 3.0 A resolution, the structure of the CREB bZIP (residues 283-341) bound to a 21-base pair deoxynucleotide that encompasses the canonical 8-base pair somatostatin cAMP response element (SSCRE). The CREB dimer is stabilized in part by ionic interactions from Arg(314) to Glu(319') and Glu(328) to Lys(333') as well as a hydrogen bond network that links the carboxamide side chains of Gln(322')-Asn(321)-Asn(321')-Gln(322). Critical to family selective dimerization are intersubunit hydrogen bonds between basic region residue Tyr(307) and leucine zipper residue Glu(312), which are conserved in all CREB/CREM/ATF-1 family members. Strikingly, the structure reveals a hexahydrated Mg(2+) ion bound in the cavity between the basic region and SSCRE that makes a water-mediated DNA contact. DNA binding studies demonstrate that Mg(2+) ions enhance CREB bZIP:SSCRE binding by more than 25-fold and suggest a possible physiological role for this ion in somatostatin cAMP response element and potentially other CRE-mediated gene expression.

Structural and thermodynamic strategies for site-specific DNA binding proteins

BACKGROUND

Site-specific protein-DNA complexes vary greatly in structural properties and in the thermodynamic strategy for achieving an appropriate binding free energy. A better understanding of the structural and energetic engineering principles might lead to rational methods for modification or design of such proteins.

RESULTS

A novel analysis of ten site-specific protein-DNA complexes reveals a striking correspondence between the degree of imposed DNA distortion and the thermodynamic parameters of each system. For complexes with relatively undistorted DNA, favorable enthalpy change drives unfavorable entropy change, whereas for complexes with highly distorted DNA, unfavorable DeltaH degrees is driven by favorable DeltaS degrees. We show for the first time that protein-DNA associations have isothermal enthalpy-entropy compensation, distinct from temperature-dependent compensation, so DeltaH degrees and DeltaS degrees do not vary independently. All complexes have favorable DeltaH degrees from direct protein-DNA recognition interactions and favorable DeltaS degrees from water release. Systems that strongly distort the DNA nevertheless have net unfavorable DeltaH degrees as the result of molecular strain, primarily associated with the base pair destacking. These systems have little coupled protein folding and the strained interface suffers less immobilization, so DeltaS degrees is net favorable. By contrast, systems with little DNA distortion have net favorable DeltaH degrees, which must be counterbalanced by net unfavorable DeltaS degrees, derived from loss of vibrational entropy (a result of isothermal enthalpy-entropy compensation) and from coupling between DNA binding and protein folding.

CONCLUSIONS

Isothermal enthalpy-entropy compensation implies that a structurally optimal, unstrained fit is achieved only at the cost of entropically unfavorable immobilization, whereas an enthalpically weaker, strained interface entails smaller entropic penalties.

Geometric analysis and comparison of protein-DNA interfaces: why is there no simple code for recognition?

Structural studies of protein-DNA complexes have shown that there are many distinct families of DNA-binding proteins, and have shown that there is no simple "code" describing side-chain/base interactions. However, systematic analysis and comparison of protein-DNA complexes has been complicated by the diversity of observed contacts, the sheer number of complexes currently available and the absence of any consistent method of comparison that retains detailed structural information about the protein-DNA interface. To address these problems, we have developed geometric methods for characterizing the local structural environment in which particular side-chain/base interactions are observed. In particular, we develop methods for analyzing and comparing spatial relationships at the protein-DNA interface. Our method involves attaching local coordinate systems to the DNA bases and to the C(alpha) atoms of the peptide backbone (these are relatively rigid structural units). We use these tools to consider how the position and orientation of the polypeptide backbone (with respect to the DNA) helps to determine what contacts are possible at any given position in a protein-DNA complex. Here, we focus on base contacts that are made in the major groove, and we use spatial relationships in analyzing: (i) the observed patterns of side-chain/base interactions; (ii) observed helix docking orientations; (iii) family/subfamily relationships among DNA-binding proteins; and (iv) broader questions about evolution, altered specificity mutants and the limits for the design of new DNA-binding proteins. Our analysis, which highlights differences in spatial relationships in different complexes and at different positions in a complex, helps explain why there is no simple, general code for protein-DNA recognition.

Bending and adaptability to proteins of the cAMP DNA-responsive element: molecular dynamics contrasted with NMR

DNA bending is assumed to play a crucial role during recognition of the cAMP-responsive element (CRE) by transcription factors. However, diverging results have been obtained for the bending direction of the unbound double helix. The refined NMR structures present a bend directed toward the minor groove, while biochemical methods conclude that there is a bend toward the major groove. The present 10-ns molecular dynamics (MD) simulation of d(GAGATGACGTCATCTC)(2), which contains the octamer CRE in its center, was carried out with AMBER in explicit water and counterions. It shows that CRE is a flexible segment, although it is bent slightly toward the major groove (7 degrees -8 degrees ) on the average. The MD structure agrees with both the biochemical results and unrefined NMR data. The divergence with the NMR refined structures suggests an improper electrostatic parameterization in the refinement software. The malleability of the central CpG is certainly the major contribution to the curving of the whole CRE segment in both the unbound and bound states. Comparison with the crystal structure of CRE bound to GCN4 shows that the deformation induced by the protein is concentrated mainly on the CpG step, rendering the bound structure of CRE closer to the structure of the 12-0 tetradecanoylphorbol-beta-acetate-responsive element.

Sexual dimorphism in diverse metazoans is regulated by a novel class of intertwined zinc fingers

Sex determination is regulated by diverse pathways. Although upstream signals vary, a cysteine-rich DNA-binding domain (the DM motif) is conserved within downstream transcription factors ofDrosophila melanogaster (Doublesex) and Caenorhabditis elegans (MAB-3). Vertebrate DM genes have likewise been identified and, remarkably, are associated with human sex reversal (46, XY gonadal dysgenesis). Here we demonstrate that the structure of the Doublesex domain contains a novel zinc module and disordered tail. The module consists of intertwined CCHC and HCCC Zn2+-binding sites; the tail functions as a nascent recognition α-helix. Mutations in either Zn2+-binding site or tail can lead to an intersex phenotype. The motif binds in the DNA minor groove without sharp DNA bending. These molecular features, unusual among zinc fingers and zinc modules, underlie the organization of a Drosophila enhancer that integrates sex- and tissue-specific signals. The structure provides a foundation for analysis of DM mutations affecting sexual dimorphism and courtship behavior.

Assembly of a functional beta interferon enhanceosome is dependent on ATF-2-c-jun heterodimer orientation

Heterodimeric transcription factors, including the basic region-leucine zipper (bZIP) protein ATF-2-c-jun, are well-characterized components of an enhanceosome that mediates virus induction of the human beta interferon (IFN-beta) gene. Here we report that within the IFN-beta enhanceosome the ATF-2-c-jun heterodimer binds in a specific orientation, which is required for assembly of a complex between ATF-2-c-jun and interferon regulatory factor 3 (IRF-3). We demonstrate that correct orientation of the ATF-2-c-jun binding site is required for virus induction of the IFN-beta gene and for IRF-3-dependent activation of a composite ATF-2- c-jun-IRF site in the IFN-beta promoter. We also show that in vitro the DNA-bound ATF-2-c-jun heterodimer adopts a fixed orientation upon the binding of IRF-3 at an adjacent site in the IFN-beta enhancer and that the DNA-binding domain of IRF-3 is sufficient to mediate this effect. In addition, we show that the DNA-binding domain of ATF-2 is necessary and sufficient for selective protein-protein interactions with IRF-3. Strikingly, in vivo chromatin immunoprecipitation experiments with IFN-beta reporter constructs reveal that recruitment of IRF-3 to the IFN-beta promoter upon virus infection is dependent on the orientation of the ATF-2-c-jun heterodimer binding site. These observations demonstrate functional and physical cooperativity between the bZIP and IRF transcription factor families and illustrate the critical role of heterodimeric transcription factors in formation of the IFN-beta enhanceosome.

Helix bending as a factor in protein/DNA recognition

Normal vectors perpendicular to individual base pairs are a powerful tool for studying the bending behavior of B-DNA, both in the form of normal vector plots and in matrices that list angles between vectors for all possible base pair combinations. A new analysis program, FREEHELIX, has been written for this purpose, and applied to 86 examples of sequence-specific protein/DNA complexes whose coordinates are on deposit in the Nucleic Acid Data Base. Bends in this sample of 86 structures almost invariably follow from roll angles between adjacent base pairs; tilt makes no net contribution. Roll in a direction compressing the broad major groove is much more common than that which compresses the minor groove. Three distinct types of B-DNA bending are observed, each with a different molecular origin: (1) Localized kinking is produced by large roll at single steps or at two steps separated by one turn of helix. (2) Smooth, planar curvature is produced by positive and negative roll angles spaced a half-turn apart, with random side-to-side zigzag roll at intermediate points, rather than a tilt contribution that might have been expected theoretically. (3) Three-dimensional writhe results from significant roll angles at a continuous series of steps. Writhe need not change the overall direction of helix axis, if it is continued indefinitely or for an integral number of helical turns. A-DNA itself can be formally considered as possessing uniform, continuous writhe that yields no net helix bending. Smooth curvature is the most intricate deformation of the three, and is least common. Writhe is the simplest deformation and is most common; indeed, a low level of continuous writhe is the normal condition of an otherwise unbent B-DNA helix of general sequence. With one exception, every example of major kinking in this sample of 86 structures involves a pyrimidine-purine step: C-A/T-G, T-A, or C-G. Purine-purine steps, especially A-A, show the least tendency toward roll deformations.

GCN4 binds with high affinity to DNA sequences containing a single consensus half-site

bZip proteins contain a bipartite DNA-binding motif consisting of a "leucine zipper" dimerization domain and a highly charged "basic region" that directly contacts DNA. These transcription factors form dimeric complexes with each monomer recognizing half of a symmetric or nearly symmetric DNA site. We have found that the bZip protein GCN4 can also bind with high affinity to DNA sites containing only a single GCN4 consensus half-site. Because several recent lines of evidence have suggested a role for monomeric DNA binding by bZip proteins, we investigated the structure of the GCN4.half-site complex. Quantitative DNA binding and affinity cleaving studies support a model in which GCN4 binds as a dimer, with one monomer making specific contacts to the consensus half-site and the other monomer forming nonspecific contacts that are nonetheless important for binding affinity. We also examined the folding transition induced in the basic regions of this complex upon binding DNA. Circular dichroism (CD) studies demonstrate that the basic regions of both monomers are helical, suggesting that a protein folding transition may be required for both specific and nonspecific DNA binding by GCN4.

Short, hydrophobic, alanine-based proteins based on the basic region/leucine zipper protein motif: overcoming inclusion body formation and protein aggregation during overexpression, purification, and renaturation

GCN4 is a yeast transcriptional regulatory protein; its DNA-binding domain is a basic region/leucine zipper (bZIP) structure that comprises a dimer of alpha-helices capable of high-affinity, sequence-specific recognition of the DNA major groove. We are exploiting what nature has evolved by manipulating the bZIP motif as a molecular recognition scaffold; thus we reduced the elegantly minimal bZIP to an even more simplified structure by substitution with alanine residues-hence, a generic, Ala-based, helical scaffold. These Ala-based mutants are unusual proteins for expression as they are short ( approximately 100 amino acids) and hydrophobic (Ala-mutated basic regions, leucine-zipper dimerization domains). Hydrophobicity posed a major problem throughout the expression, isolation, and purification stages; inclusion body formation and protein aggregation were significant hurdles throughout protein production. We describe measures that solved these problems, including use of high concentrations of denaturant in all steps of protein isolation and purification and use of temperature-dependent renaturing techniques to obtain folded, functional protein. Despite these difficulties, we ultimately retrieved 5-10 mg/L of broth of active, correctly folded protein after the complete purification procedure. Homogeneity of the proteins was established by chromatography, electrophoresis, and mass spectrometry. Furthermore, characterization by circular dichroism and DNase footprinting analysis demonstrates that these alanine-based mutants retain the structure and function of the native GCN4 DNA-binding domain. Remarkably, the most heavily mutated protein, containing 24 alanines of 27 total amino acids in the DNA-binding basic region, still binds the AP-1 site, the target of native GCN4.

Protein-directed DNA structure II. Raman spectroscopy of a leucine zipper bZIP complex

Mechanisms of transcription may involve protein-directed changes in DNA structure and DNA-directed changes in protein structure. We have employed Raman spectroscopy to characterize vibrational signatures associated with such induced molecular fitting for two classes of transcription factors-the basic leucine-zipper (bZIP) motif and the high-mobility-group (HMG) box-each with a DNA target site. Results for bZIP are described here; findings for the HMG-box are reported in the preceding paper in this issue [Benevides, J. M., Chan, G., Lu, X.-J., Olson, W. K., Weiss, M. A., and Thomas, G. J., Jr. (2000) Biochemistry 39, 537-547]. The yeast activator GCN4 provides a well-studied example of bZIP recognition, wherein B-DNA serves essentially as a template for protein folding. Analysis of Raman spectra of the 57-residue GCN4 bZIP domain, its AP-1 binding site, and their specific complex confirms a DNA-induced increase in alpha-helicity, attributable to folding of GCN4 basic arms with virtually no change in B-DNA structure, consistent with previous X-ray and NMR structure determinations. The absence of DNA perturbations in the bZIP model contrasts sharply with the HMG box, where DNA structure perturbations predominate. The bZIP and HMG-box models represent two opposing extremes in a range of induced fits identifiable by Raman spectroscopy. Previously characterized lambda repressor/operator complexes [Benevides, J. M., Weiss, M. A., and Thomas, G. J. (1994) J. Biol. Chem. 269, 10869-10878] occupy an intermediate position within this range. A comprehensive tabulation of Raman markers proposed as diagnostic of different protein/DNA recognition motifs is presented. The results are analyzed in terms of available DNA crystal structures (Nucleic Acid Database) to identify details of DNA conformation that correlate with specific Raman recognition markers.

Interaction of calicheamicin gamma1(I) and its related carbohydrates with DNA-protein complexes

We report studies of the contribution of DNA structure, holding the sequence constant, to the affinity of calicheamicin gamma(1)(I) and its aryltetrasaccharide moiety for DNA. We used polynucleotide chains as models of known protein-binding sequences [the catabolite activator protein (CAP) consensus sequence, AP-1 and cAMP response element (CRE) sites] in their free and protein-bound forms. The proteins were selected to provide examples in which the minor-groove binding site for the carbohydrate is (CAP) or is not (GCN4) covered by the protein. Additionally, peptides related to the GCN4 and CREB families, which have different bending effects on their DNA-binding sites, were used. We observe that proteins of the CREB class, which induce a tendency to bend toward the minor groove at the center of the site, inhibit drug-cleavage sites located at the center of the free AP-1 or CRE DNA sites. In the case of GCN4, which does not induce DNA bending, there is no effect on calicheamicin cleavage of the CRE site, but we observe a GCN4-induced rearrangement of the cutting pattern in the AP-1 site. This effect may arise from either a subtle local conformational rearrangement not accompanied by bending or a localized reduction in DNA flexibility. Whereas GCN4 binding is not inhibited by the calicheamicin aryltetrasaccharide, binding of CAP to its DNA target is significantly inhibited, and calicheamicin cutting of DNA at the center of the CAP-DNA complex site is strongly reduced by protein binding. This result probably reflects steric inhibition of drug binding by the protein.

Protein-DNA interactions: A structural analysis

A detailed analysis of the DNA-binding sites of 26 proteins is presented using data from the Nucleic Acid Database (NDB) and the Protein Data Bank (PDB). Chemical and physical properties of the protein-DNA interface, such as polarity, size, shape, and packing, were analysed. The DNA-binding sites shared common features, comprising many discontinuous sequence segments forming hydrophilic surfaces capable of direct and water-mediated hydrogen bonds. These interface sites were compared to those of protein-protein binding sites, revealing them to be more polar, with many more intermolecular hydrogen bonds and buried water molecules than the protein-protein interface sites. By looking at the number and positioning of protein residue-DNA base interactions in a series of interaction footprints, three modes of DNA binding were identified (single-headed, double-headed and enveloping). Six of the eight enzymes in the data set bound in the enveloping mode, with the protein presenting a large interface area effectively wrapped around the DNA.A comparison of structural parameters of the DNA revealed that some values for the bound DNA (including twist, slide and roll) were intermediate of those observed for the unbound B-DNA and A-DNA. The distortion of bound DNA was evaluated by calculating a root-mean-square deviation on fitting to a canonical B-DNA structure. Major distortions were commonly caused by specific kinks in the DNA sequence, some resulting in the overall bending of the helix. The helix bending affected the dimensions of the grooves in the DNA, allowing the binding of protein elements that would otherwise be unable to make contact. From this structural analysis a preliminary set of rules that govern the bending of the DNA in protein-DNA complexes, are proposed.

Stability of the dimerization domain effects the cooperative DNA binding of short peptides

The basic region peptide derived from the basic leucine zipper protein GCN4 bound specifically to the native GCN4 binding sequences in a dimeric form when the beta-cyclodextrin/adamantane dimerization domain was introduced at the C-terminus of the GCN4 basic region peptide. We describe here how the structure and stability of the dimerization domain affect the cooperative formation of the peptide dimer-DNA complex. The basic region peptides with five different guest molecules were synthesized, and their equilibrium dissociation constants with a peptide possessing beta-cyclodextrin were determined. These values, ranging from 1.3 to 15 microM, were used to estimate the stability of the complexes between the dimers with various guest/cyclodextrin dimerization domains and GCN4 target sequences. An efficient cooperative formation of the dimer complexes at the GCN4 binding sequence was observed when the adamantyl group was replaced with the norbornyl or noradamantyl group, but not with the cyclohexyl group that formed a beta-cyclodextrin complex with a stability that was 1 order of magnitude lower than that of the adamantyl group. Thus, cooperative formation of the stable dimer-DNA complex appeared to be effected by the stability of the dimerization domain. For the peptides that cooperatively formed dimer-DNA complexes, there was no linear correlation between the stability of the inclusion complex and that of the dimer-DNA complex. With the beta-cyclodextrin/adamantane dimerization domain, the basic region peptide dimer preferred to bind to a palindromic 5'-ATGACGTCAT-3' sequence over the sequence lacking the central G.C base pair and that with an additional G.C base pair in the middle. Changing the adamantyl group into a norbornyl group did not alter the preferential binding of the peptide dimers to the palindromic sequence, but slightly affected the selectivity of the dimer for other nonpalindromic sequences. The helical contents of the peptides in the DNA-bound dimer with the adamantyl group were decreased by reducing the stability of the dimer-DNA complex, which was possibly caused by deformation of the helical structure proximal to the dimerization domain.

A novel feature of DNA recognition: a mutant Gcn4p bZip peptide with dual DNA binding specificities dependent of half-site spacing

Homodimeric DNA-binding proteins with relaxed half-site spacing requirements for their DNA targets have been described. As an example, the yeast transcriptional activator Gcn4p binds in vitro equally well to the AP1 site (5'A4T3G2A1C0T1'C2'A3'T4'3') and the ATF/CREB site (5'A4T3G2A1C0G0'T1'C2'A3'T4'3'), which have identical but differently spaced half-site blocks. We describe a novel feature for the bZip class of DNA-binding proteins. The N-14 mutant of a Gcn4p-derived bZip peptide shows a diametrically opposed base-pair recognition specificity depending on the half-site spacing of its DNA target: on pseudo-palindromic, AP1 site-like binding sites, guanine is required in position 2 for proper binding; in contrast, on palindromic, ATF/CREB site-like targets, position 2 must be cytosine to prevent a loss of binding. Modeling studies suggest that the different base-pair requirements on differently spaced DNA targets are due to minimal alterations of the distances between the relevant atoms of the N-14 side-chain and the corresponding target groups on the DNA. Although the N-14 peptide does not have a natural counterpart, its behavior hints at the possibility that dual binding modi dependent on half-site spacing may occur also for natural homodimeric DNA-binding proteins.

Structural features of protein-nucleic acid recognition sites

We analyzed the atomic models of 75 X-ray structures of protein-nucleic acid complexes with the aim of uncovering common properties. The interface area measured the extent of contact between the protein and nucleic acid. It was found to vary between 1120 and 5800 A2. Despite this wide variation, the interfaces in complexes of transcription factors with double-stranded DNA could be broken up into recognition modules where 12 +/- 3 nucleotides on the DNA side contact 24 +/- 6 amino acids on the protein side, with interface areas in the range 1600 +/- 400 A2. For enzymes acting on DNA, the recognition module is on average 600 A2 larger, due to the requirement of making an active site. As judged by its chemical and amino acid composition, the average protein surface in contact with the DNA is more polar than the solvent accessible surface or the typical protein-protein interface. The protein side is rich in positively charged groups from lysine and arginine side chains; on the DNA side the negative charges from phosphate groups dominate. Hydrogen bonding patterns were also analyzed, and we found one intermolecular hydrogen bond per 125 A2 of interface area in high-resolution structures. An equivalent number of polar interactions involved water molecules, which are generally abundant at protein-DNA interfaces. Calculations of Voronoi atomic volumes, performed in the presence and absence of water molecules, showed that protein atoms buried at the interface with DNA are on average as closely packed as in the protein interior. Water molecules contribute to the close packing, thereby mediating shape complementarity. Finally, conformational changes accompanying association were analyzed in 24 of the complexes for which the structure of the free protein was also available. On the DNA side the extent of deformation showed some correlation with the size of the interface area. On the protein side the type and size of the structural changes spanned a wide spectrum. Disorder-to-order transitions, domain movements, quaternary and tertiary changes were observed, and the largest changes occurred in complexes with large interfaces.

Comparison of the sequence-selective DNA binding by peptide dimers with covalent and noncovalent dimerization domains

Sequence-specific DNA binding proteins generally consist of more than two DNA-contacting regions to ensure the selectivity of recognition. The multiple DNA binding modules are connected either through the covalent linker or through the noncovalent dimerization domain. We have compared the DNA binding of peptide dimers with covalent and noncovalent dimerization domains to explore the potential advantage of each linkage on the sequence-specific DNA binding. Three sets of head-to-tail peptide dimers were synthesized by using the same basic region peptide to target the same DNA sequence; one dimer was assembled with a bridged biphenyl derivative as a covalent dimerization domain, and two other dimers were assembled with the cyclodextrin guest noncovalent dimerization domains. One of the noncovalent dimers was a heterodimer that consisted of cyclodextrin and guest peptides, while the other was a homodimer that consisted of peptides bearing both cyclodextrin and the guest molecule within the same chain. Both noncovalent dimers formed the specific DNA complexes within narrower ranges of peptide concentrations and showed higher sequence selectivity than the covalent dimer did. Among the three dimers, the noncovalent homodimer that can form an intramolecular inclusion complex showed the highest sequence selectivity. Because the noncovalent homodimer with the higher stability of the circular intramolecular inclusion complex exhibited the higher sequence selectivity, it was concluded that an equilibrium involving a conformational transition of a monomeric peptide effectively reduced the stability of its nonspecific binding complex, hence increasing the efficacy of cooperative dimer formation at the specific DNA sequence.

Structural code for DNA recognition revealed in crystal structures of papillomavirus E2-DNA targets

Transcriptional regulation in papillomaviruses depends on sequence-specific binding of the regulatory protein E2 to several sites in the viral genome. Crystal structures of bovine papillomavirus E2 DNA targets reveal a conformational variant of B-DNA characterized by a roll-induced writhe and helical repeat of 10.5 bp per turn. A comparison between the free and the protein-bound DNA demonstrates that the intrinsic structure of the DNA regions contacted directly by the protein and the deformability of the DNA region that is not contacted by the protein are critical for sequence-specific protein/DNA recognition and hence for gene-regulatory signals in the viral system. We show that the selection of dinucleotide or longer segments with appropriate conformational characteristics, when positioned at correct intervals along the DNA helix, can constitute a structural code for DNA recognition by regulatory proteins. This structural code facilitates the formation of a complementary protein-DNA interface that can be further specified by hydrogen bonds and nonpolar interactions between the protein amino acids and the DNA bases.

An extended alpha-helix and specific amino acid residues opposite the DNA-binding surface of the cAMP response element binding protein basic domain are important for human T cell lymphotropic retrovirus type I Tax binding

The human T cell lymphotropic retrovirus type I (HTLV-I) trans-activator, Tax, interacts specifically with the basic-domain/leucine-zipper (bZip) protein, cAMP response element binding protein (CREB), bound to the viral Tax-responsive element consisting of three imperfect 21-base pair repeats, each with a cAMP response element core flanked by G/C-rich sequences. Here, the minimal CREB-bZip necessary for Tax binding is shown to be composed of amino acid residues 280-341. The Tax-CREB interaction involves an uninterrupted and extended alpha-helix in CREB that spans most of its basic domain to include amino acid residues localized to the NH2 terminus of the DNA binding region. Mutational analyses indicate that three residues, Arg284, Met291, and Glu299 unique to this region of the CREB/activating transcription factor-1 subfamily of bZip proteins, constitute the contact surface for Tax. Amino acid substitutions in these positions had little impact on CREB-bZip binding to DNA but abrogated its binding to Tax. Each of the contact residues for Tax are spaced approximately two helical turns apart on the side of the bZip helix directly opposite to that of the invariant DNA-binding residues. Molecular modeling reveals the Tax-contact residues to be near the minor groove of the G/C-rich DNA in the 21-base pair repeat. They most likely position Tax for minor groove contact with the G/C-rich sequences.

Sequence determinants of the intrinsic bend in the cyclic AMP response element

The cyclic AMP response element (CRE site, ATGACGTCAT) is the DNA target for transcription factors whose activities are regulated by cyclic AMP (1). Recently, we discovered that the CRE site is bent by 10-13 degrees toward the major groove (2). Little or no bend is detected in the related AP-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 the bent structure in an attempt to understand the origins of the dramatically different conformations of the CRE and AP-1 sites. Our data indicate that the intrinsic CRE bend results from distortion within the TGA sequence found in each CRE half site (ATGAC). These two TGA sequences are located in phase with one another in the CRE sequence but are not (completely) in phase in the AP-1 sequence. This difference 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, provide insight into the structural reorganizations induced in DNA by bZIP proteins, and lead to a revision of the relationship between the structures of the free and bZIP-bound forms of the CRE and AP-1 sites.

The activation specificities of wild-type and mutant Gcn4p in vivo can be different from the DNA binding specificities of the corresponding bZip peptides in vitro

Single amino acid substitutions which previously have been shown to alter the DNA binding specificity of a Gcn4p bZip peptide in vitro were transformed to full length Gcn4p, and activation of a test promoter carrying various palindromic and pseudo-palindromic binding sites was measured. All mutations were found to have different phenotypes, and the first change-of-specificity mutants for Gcn4p in vivo are described. The comparison of plasmids encoding no protein or a particular Gcn4p mutant with broadened activation specificity in gcn4 and gcn4 acr1 genetic backgrounds revealed three new DNA targets of the yeast Acr1p repressor. Surprisingly, we found the activation specificities Gcn4p and the mutants tested in vivo to be generally different from DNA binding specificities of the corresponding bZip peptides in vitro. Especially, the proteins respond differently, in vitro and in vivo, on changes in half site spacing of the DNA binding sites. We present data which largely exclude that the differences between in vivo and in vitro-derived results are due to differences in protein structure, or to the presence of competing protein factors in the yeast cell. We conclude that the differences between in vitro and in vivo-derived results are caused by differences in the degree of flexibility of the target DNA sequences in vitro and in vivo.

DNA-binding domains of Fos and Jun do not induce DNA curvature: an investigation with solution and gel methods

We demonstrate the use of a DNA minicircle competition binding assay, together with DNA cyclization kinetics and gel-phasing methods, to show that the DNA-binding domains (dbd) of the heterodimeric leucine zipper protein Fos-Jun do not bend the AP-1 target site. Our DNA constructs contain an AP-1 site phased by 1-4 helical turns against an A-tract-directed bend. Competition binding experiments reveal that (dbd)Fos-Jun has a slight preference for binding to linear over circular AP-1 DNAs, independent of whether the site faces in or out on the circle. This result suggests that (dbd)Fos-Jun slightly stiffens rather than bends its DNA target site. A single A-tract bend replacing the AP-1 site is readily detected by its effect on cyclization kinetics, in contrast to the observations for Fos-Jun bound at the AP-1 locus. In contrast, comparative electrophoresis reveals that Fos-Jun-DNA complexes, in which the A-tract bend is positioned close (1-2 helical turns) to the AP-1 site, show phase-dependent variations in gel mobilities that are comparable with those observed when a single A-tract bend replaces the AP-1 site. Whereas gel mobility variations of Fos-Jun-DNA complexes decrease linearly with increasing Mg2+ contained in the gel, the solution binding preference of (dbd)Fos-Jun for linear over circular DNAs is independent of Mg2+ concentration. Hence, gel mobility variations of Fos-Jun-DNA complexes are not indicative of (dbd)Fos-Jun-induced DNA bending (upper limit 5 degrees) in the low salt conditions of gel electrophoresis. Instead, we propose that the gel anomalies depend on the steric relationship of the leucine zipper region with respect to a DNA bend.

Characterization of the ATF/CREB site and its complex with GCN4

We have studied DNA minicircles containing the ATF/CREB binding site for GCN4 by using a combination of cyclization kinetics experiments and Monte Carlo simulations. Cyclization rates were determined with and without GCN4 for DNA constructs containing the ATF/CREB site separated from a phased A-tract multimer bend by a variable length phasing adaptor. The cyclization results show that GCN4 binding does not significantly change the conformation of the ATF/CREB site, which is intrinsically slightly bent toward the major groove. Monte Carlo simulations quantitate the ATF/CREB site structure as an 8 degrees bend toward the major groove in a coordinate frame near the center of the site. The ATF/CREB site is underwound by 53 degrees relative to the related AP-1 site DNA. The effect of GCN4 binding can be modeled either as a decrease in the local flexibility, corresponding to an estimated 60% increase in the persistence length for the 10-bp binding site, or possibly as a small decrease (1 degrees) in intrinsic bend angle. Our results agree with recent electrophoretic and crystallographic studies and demonstrate that cyclization and simulation can characterize subtle changes in DNA structure and flexibility.

Mechanism of DNA binding enhancement by hepatitis B virus protein pX

At least three hundred million people worldwide are infected with the hepatitis B virus (HBV), and epidemiological studies show a clear correlation between chronic HBV infection and the development of hepatocellular carcinoma. HBV encodes a protein, pX, which abducts the cellular transcriptional machinery in several ways including direct interactions with bZIP transcription factors. These interactions increase the DNA affinities of target bZIP proteins in a DNA sequence-dependent manner. Here we use a series of bZIP peptide models to explore the mechanism by which pX interacts with bZIP proteins. Our results suggest that pX increases bZIP.DNA stability by increasing the stability of the bZIP dimer as well as the affinity of the dimer for DNA. Additional experiments provide evidence for a mechanism in which pX recognizes the composite structure of the peptide.DNA complex, not simply the primary peptide sequence. These experiments provide a framework for understanding how pX alters the patterns of transcription within the nucleus. The similarities between the mechanism proposed for pX and the mechanism previously proposed for the human T-cell leukemia virus protein Tax are discussed.

DNA-binding specificity of the homeodomain-leucine zipper domain

Homeodomain-leucine zipper (HD-Zip) proteins are putative transcription factors identified only in plants. The study of the DNA-binding properties of the ATHB-1 and -2 HD-Zip (HD-Zip-1 and -2) domains showed that they interact with DNA as homodimers and recognize two distinct 9 bp pseudopalindromic sequences, CAAT(A/T)ATTG (BS-1) and CAAT(G/C)ATTG (BS-2), respectively, as determined by selecting high-affinity binding sites from random-sequence DNA. Here, we report a mutational analysis of the HD-Zip-2 domain. We determined that conserved amino acid residues of helix 3, Val47 and Asn51, and Arg55 are essential for the DNA-binding activity of the HD-Zip-2 domain. We demonstrated that the preferential recognition of a G/C base-pair at the central position by the HD-Zip-2 domain is abolished either by the replacement of Arg55 with lysine or by the substitution of Glu46 and Thr56 with the corresponding residues of the HD-Zip-1 domain (alanine and tryptophan, respectively). In contrast, substitution of Arg55 with lysine in the HD-Zip-1 domain significantly reduced DNA-binding activity without changing the specificity of recognition. Finally, we determined that differences in residues outside helix 3 further contribute to the DNA-binding specificity of the HD-Zip domain. Taken together, the data strongly suggest that the preferential recognition of BS-2 and -1 by the HD-Zip-2 and -1 domains, respectively, may be attributable to a distinct orientation of the side-chain of Arg55 in these two domains.

Interaction of a bZip oligopeptide model with oligodeoxyribonucleotides modelling DNA binding sites. The effect of flanking sequences

A leucine zipper (bZip) binding peptide BP1 was constructed based on the DNA binding sequence of the GCN4 protein, slightly modified to make it more similar to the sequence of other bZip proteins (Jun) with related DNA binding specificity. Self-complementary DNA hexadecanucleotides containing ATF/CRE, AP-1 and C/EPB target sites were used to study peptide-DNA complex formation. Conformation changes in both components that occur on complex formation were studied by circular dichroism (CD) spectroscopy. The results show that the amount of alpha-helix formed in the peptide strongly depends not only on the target site present, but also on the type of the sequence flanking the ATF/CRE target site. Highest amount of the alpha-helix induced in the peptide was observed when homopurine homopyrimidine flanking sequences were present, whereas the presence of alternating sequences, especially of the CA/TG type, showed considerably lower effects. The change in DNA conformation on complex formation was generally small, but also depended on the type of the flanking sequence. It appears that the sequences flanking the target site can considerably modify the ability of the target sequence to bind specifically the bZip peptide, probably by slightly varying the overall DNA conformation.

Crystal structure of a PUT3-DNA complex reveals a novel mechanism for DNA recognition by a protein containing a Zn2Cys6 binuclear cluster

PUT3 is a member of a family of at least 79 fungal transcription factors that contain a six-cysteine, two-zinc domain called a 'Zn2Cys6 binuclear cluster'. We have determined the crystal structure of the DNA binding region from the PUT3 protein bound to its cognate DNA target. The structure reveals that the PUT3 homodimer is bound asymmetrically to the DNA site. This asymmetry orients a beta-strand from one protein subunit into the minor groove of the DNA resulting in a partial amino acid-base pair intercalation and extensive direct and water-mediated protein interactions with the minor groove of the DNA. These interactions facilitate a sequence dependent kink at the centre of the DNA site and specify the intervening base pairs separating two DNA half-sites that are contacted in the DNA major groove. A comparison with the GAL4-DNA and PPR1-DNA complexes shows how a family of related DNA binding proteins can use a diverse set of mechanisms to discriminate between the base pairs separating conserved DNA half-sites.

SKN-1 domain folding and basic region monomer stabilization upon DNA binding

The SKN-1 transcription factor specifies early embryonic cell fates in Caenorhabditis elegans. SKN-1 binds DNA at high affinity as a monomer, by means of a basic region like those of basic-leucine zipper (bZIP) proteins, which bind DNA only as dimers. We have investigated how the SKN-1 DNA-binding domain (the Skn domain) promotes stable binding of a basic region monomer to DNA. A flexible arm at the Skn domain amino terminus binds in the minor groove, but a support segment adjacent to the carboxy-terminal basic region can independently stabilize basic region-DNA binding. Off DNA, the basic region and arm are unfolded and, surprisingly, the support segment forms a molten globule of four alpha-helices. On binding DNA, the Skn domain adopts a tertiary structure in which the basic region helix extends directly from a support segment alpha-helix, which is required for binding. The remainder of the support segment anchors this uninterrupted helix on DNA, but leaves the basic region exposed in the major groove. This is similar to how the bZIP basic region extends from the leucine zipper, indicating that positioning and cooperative stability provided by helix extension are conserved mechanisms that promote binding of basic regions to DNA.

Electrostatic mechanism for DNA bending by bZIP proteins

Biology is replete with examples of protein-induced DNA bending, yet the forces responsible for bending have been neither established nor quantified. Mirzabekov and Rich proposed in 1979 that asymmetric neutralization of the anionic phosphodiester backbone by basic histone proteins could provide a thermodynamic driving force for DNA bending in the nucleosome core particle [Mirzabekov, A. D., & Rich, A. (1979) Proc. Natl. Acad. Sci. U.S.A. 76, 1118-1121]. Strauss and Maher lent support to this proposal in 1994 by demonstrating that replacement of six proximal phosphate residues with neutral methylphosphonates resulted in DNA bent spontaneously toward the neutralized face [Strauss, J. K., & Maher, L. J., III (1994) Science 266, 1829-1834; Strauss, J. K., Prakash, T. P., Roberts, C., Switzer, C., & Maher, L. J., III (1996) Chem. Biol. 3, 671-678; Strauss, J. K., Roberts, C.; Nelson, M. G.; Switzer, C., & Maher, J. L., III (1996) Proc. Natl. Acad. Sci. U.S.A. 93, 9515-9520]. Here it is shown that bZIP proteins bend DNA via a mechanism involving direct contacts between one or two basic side chains and a symmetry-related pair of unique, nonbridging phosphate oxygens. The locations of these phosphates provide direct experimental support for a protein-induced bending mechanism based on asymmetric charge neutralization. This straightforward mechanism is compatible with many DNA-recognition motifs and may represent a general strategy for the assembly of protein-DNA complexes of defined stereochemistries.

The role of a basic amino acid cluster in target site selection and non-specific binding of bZIP peptides to DNA

The ability of a transcription factor to locate and bind its cognate DNA site in the presence of closely related sites and a vast array of non-specific DNA is crucial for cell survival. The CREB/ATF family of transcription factors is an important group of basic region leucine zipper (bZIP) proteins that display high affinity for the CRE site and low affinity for the closely related AP-1 site. Members of the CREB/ATF family share in common a cluster of basic amino acids at the N-terminus of their bZIP element. This basic cluster is necessary and sufficient to cause the CRE site to bend upon binding of a CREB/ATF protein. The possibility that DNA bending and CRE/AP-1 specificity were linked in CREB/ATF proteins was investigated using chimeric peptides derived from human CRE-BP1 (a member of the CREB/ATF family) and yeast GCN4, which lacks both a basic cluster and CRE/AP-1 specificity. Gain of function and loss of function experiments demonstrated that the basic cluster was not responsible for the CRE/AP-1 specificity displayed by all characterized CREB/ATF proteins. The basic cluster was, however, responsible for inducing very high affinity for non- specific DNA. It was further shown that basic cluster-containing peptides bind non-specific DNA in a random coil conformation. We postulate that the high non- specific DNA affinities of basic cluster-containing peptides result from cooperative electrostatic interactions with the phosphate backbone that do not require peptide organization.

Skn-1: evidence for a bipartite recognition helix in DNA binding

Skn-1 is a maternally expressed transcription factor that specifies the fate of certain blastomeres early in the development of Caenorhabditis elegans. This transcription factor contains a basic region, but it binds to DNA as a monomer. Because other transcription factors containing basic regions bind as dimers, this finding implied that Skn represents a new DNA recognition motif. It has been proposed that the basic region helix of Skn is stabilized for binding by tertiary contacts to other parts of the protein. We have tested this proposal by carrying out circular dichroism (CD) and NMR experiments on the Skn domain and five truncated proteins. Our results have shown that the basic region of Skn is unstructured in solution and does not contact other parts of the protein; like other basic region peptides, it folds into a helix only upon binding specifically to DNA. However, there is a stably folded helical module in the Skn domain, and one of the helices in this module terminates immediately before the start of the basic region. This pre-organized helix contains a surface rich in basic amino acids, and we propose that this helix contacts the DNA distal to the basic region proper, providing an extra long helical recognition surface which helps to stabilize monomeric binding. Homology between the Skn domain and several basic-region leucine zipper (bZIP) domains raises the possibility that the affinity and perhaps the specificity of DNA binding by bZIP proteins can be modulated by incorporating a stably folded helical segment that contacts the DNA just below the basic region proper.

Coupled folding and site-specific binding of the GCN4-bZIP transcription factor to the AP-1 and ATF/CREB DNA sites studied by microcalorimetry

The site-specific interaction of the basic leucine zipper protein C62GCN4, which corresponds to the C-terminal sequence 220-281 of the yeast transcription factor GCN4, with the AP-1 and ATF/ CREB DNA recognition sites was analyzed by isothermal titration microcalorimetry. Free C62GCN4 is a dimer composed of a C-terminal leucine zipper and a basic, mainly unstructured DNA binding domain. Upon association with the target DNA, C62GCN4 folds to a fully alpha-helical dimer [Ellenberger et al. (1992) Cell 71, 1223-1237; König and Richmond (1993) J. Mol. Biol. 233, 139-154]. The protein-bound AP-1 site is straight, and the protein-bound ATF/CREB site is bent by 20 degrees toward the leucine zipper domain. The coupling between protein folding and DNA association resulting in two conformationally different complexes with C62GCN4 poses interesting thermodynamic problems. The association was strongly exothermic for both DNA target sites. The free energies of binding were indistinguishable in buffers of low salt concentration, and no change of the protonation state of C62GCN4 and/or the DNA target site occurred on formation of the complexes. Both complexes exhibited large and negative heat capacity changes. The empirical correlation between buried nonpolar and polar surfaces and the reduction in heat capacity concomitant to complexation did hold for the reaction with the AP-1 site at low salt concentration. However, in the case of the ATF/CREB site, the change in heat capacity was larger than could be accounted for by the burial of solvent-accessible surface. Potential sources of the extra decrement in the heat capacity could be restrictions in the vibrational modes of polar groups and of bound water molecules at the protein-DNA interface, thought to result from the bending of the ATF/CREB site. In the presence of high concentrations of glutamate and NaCl, the complex with the ATF/CREB site was significantly weaker than the complex with the AP-1 site.

DNA binding of Jun and Fos bZip domains: homodimers and heterodimers induce a DNA conformational change in solution

We constructed plasmids encoding the sequences for the bZip modules of c-Jun and c-Fos which could then be expressed as soluble proteins in Escherichia coli. The purified bZip modules were tested for their binding capacities of synthetic oligonucleotides containing either TRE or CRE recognition sites in electrophoretic mobility shift assays and circular dichroism (CD). Electrophoretic mobility shift assays showed that bZip Jun homodimers and bZip Jun/Fos heterodimers bind a collagenase-like TRE (CTGACTCAT) with dissociation constants of respectively 1.4 x 10(-7) M and 5 x 10(-8) M. As reported earlier [Patel et al. (1990) Nature 347, 572-575], DNA binding induces a marked change of the protein structure. However, we found that the DNA also undergoes a conformational change. This is most clearly seen with small oligonucleotides of 13 or 14 bp harboring respectively a TRE (TGACTCA) or a CRE (TGACGTCA) sequence. In this case, the positive DNA CD signal at 280 nm increases almost two-fold with a concomitant blue-shift of 3-4 nm. Within experimental error the same spectral changes are observed for TRE and CRE containing DNA fragments. The spectral changes observed with a non-specific DNA fragment are weaker and the signal of free DNA is recovered upon addition of much smaller salt concentrations than required for a specific DNA fragment. Surprisingly the spectral changes induced by Jun/Jun homodimers are not identical to those induced by Jun/Fos heterodimers. However, in both cases the increase of the positive CD band and the concomitant blue shift would be compatible with a B to A-transition of part of the binding site or a DNA conformation intermediate between the canonical A and B structures.

Simultaneous binding of a polyamide dimer and an oligonucleotide in the minor and major grooves of DNA

The effect of the polyamide ImPyPy-Dp (Im = N-methylimidazole-2-carboxamide, Py = N-methylpyrrole-2-carboxamide, and Dp = dimethylaminopropylamide), which binds as an antiparallel dimer in the Watson-Crick minor groove, on pyrimidine. purine.pyrimidine triple helix stability was investigated. A DNA restriction fragment was designed which contained two triple helix sites, one which overlapped a minor groove ligand site (proximal), and a control site 13 base pairs away (distal). Using quantitative DNase I footprint titration experiments the equilibrium association constant of oligonucleotide 5'-TTTTTm5CTTTm5CTTTm5CT-3' (1) to each site was measured in the absence and presence of the polyamide dimer. Our data indicate that triple helix formation is compatible with a polyamide dimer binding in the minor groove of DNA at an overlapping site. No cooperative effect of the polyamide dimer on the equilibrium association constant of oligonucleotide 1 was observed.