A comparison of six DNA bending models

The predictions of six DNA bending models were compared with experimental relative mobility data. The study showed that all the models are reasonably accurate in predicting bending in synthetic sequences and in a natural sequence. The least accurate of these models is the Calladine-Dickerson model. The most consistent model is the ApA Wedge, possibly because it distributes the bends into base-roll and base-tilt components.     

DNA mismatch binding activities in Chlorella pyrenoidosa extracts and affinity isolation of G-T mismatch binding proteins.

DNA mismatch recognition proteins contained in the extracts of unicellular alga Chlorella pyrenoidosa were isolated by affinity adsorption and 2-D gel electrophoresis. Incubation of the algal extracts with a 38-mer duplex oligonucleotide carrying a single DNA simple mispair generated a few gel retardation complexes. G-T mispair was recognized significantly better than C-T, G-G, G-A, and C-C mispairs by the algal extracts and these extracts bound very weakly to G-A and C-C mispairs, displaying a universal trend of mismatch binding efficiency. The levels of mismatch recognition complexes were slightly increased in the presence of 1 mM ATP. Two 13-kDa G-T binding polypeptides possessing pIs of 5.3 and 5.5 were isolated after resolving affinity-captured proteins by 2-D gel electrophoresis and the two factors were found to bind 5.5- and 2.8-fold stronger to heteroduplex than to homoduplex DNA, respectively. No proteins significantly homologous to the two algal G-T binding proteins were found by peptide mass fingerprinting (PMF). The sequence of a peptide generated from trypsin-cleavage of one G-T binding factor (pI 5.5) could be aligned with the amino acid sequences that form the C-terminal active sites of human and mouse mismatch-specific uracil/thymine-DNA glycosylases, suggesting the possibility of this factor as an algae- or a Chlorella-specific DNA mismatch glycosylase.     

A comparison of the different DNA binding specificities of the bZip proteins C/EBP and GCN4.

The bZip proteins GCN4 and C/EBP differ in their DNA binding specificities: GCN4 binds well to the pseudopalindromic AP1 site 5'-A4T3G2A1C0T1C2'A3'T4'-3' and to the palindromic ATF/CREB sequence 5'-A4T3G2A1-C0*G0'T1'C2'A3'T4'-3'; C/EBP preferentially recognizes the palindromic sequence 5'-A4T3T2G1C0*G0'C1'A2'-A3'T4'-3'. According to the X-ray structures of GCN4-DNA complexes, five residues of the basic region of GCN4 are involved in specific base contacts: asparagine -18, alanine -15, alanine -14, serine -11 and arginine -10 (numbered relative to the start point of the leucine zipper, which we define as +1). In the basic region of C/EBP position -14 is occupied by valine instead of alanine, the other four residues being identical. Here we analyse the role of valine -14 in C/EBP-DNA complex formation. Starting from a C/EBP-GCN4 chimeric bZip peptide which displays C/EBP specificity, we systematically mutated position -14 of its basic region and characterized the DNA binding specificities of the 20 possible different peptides by gel mobility shift assays with various target sites. We present evidence that valine -14 of C/EBP interacts more strongly with thymine 2 than with cytosine 1' of the C/EBP binding site, unlike the corresponding alanine -14 of GCN4, which exclusively contacts thymine 1' of the GCN4 binding sites.     

Sequence comparison of single-stranded DNA binding proteins and its structural implications

The primary sequences were compared among several proteins: gene product 5 protein (GP5) from phage M13; PIKE from phage Ike; gene product 32 protein (GP32) from phage T4; RecA, SSB and SSF from Escherichia coli. These proteins bind strongly and cooperatively to single-stranded DNA with no sequence specificity. GP5 is the smallest in this group and its three-dimensional structure is well-characterized. Using the entire sequence of GP5 as a template we searched for the regions in other single-stranded DNA binding proteins yielding the best alignment of aromatic and basic residues. The identified domains show alignment of five aromatic and four charged residues in these proteins. The domains in PIKE, GP32 and RecA exhibit statistically significant sequence homology with GP5. These observations strongly favor the hypothesis that the protein-single-stranded DNA complex in this class of proteins is stabilized by the stacking interaction of the aromatic residues with the bases of the DNA, and by the electrostatic interaction of the basic residues with the phosphate groups of the DNA. We also find that the DNA binding domains of these proteins have similar secondary structural preferences, mainly beta structures. The triple-stranded beta-sheet may be a common motif in the DNA binding domains of these proteins.     

Selective DNA bending by a variety of bZIP proteins.

We have investigated DNA bending by bZIP family proteins that can bind to the AP-1 site. DNA bending is widespread, although not universal, among members of this family. Different bZIP protein dimers induced distinct DNA bends. The DNA bend angles ranged from virtually 0 to greater than 40 degrees as measured by phasing analysis and were oriented toward both the major and the minor grooves at the center of the AP-1 site. The DNA bends induced by the various heterodimeric complexes suggested that each component of the complex induced an independent DNA bend as previously shown for Fos and Jun. The Fos-related proteins Fra1 and Fra2 bent DNA in the same orientation as Fos but induced smaller DNA bend angles. ATF2 also bent DNA toward the minor groove in heterodimers formed with Fos, Fra2, and Jun. CREB and ATF1, which favor binding to the CRE site, did not induce significant DNA bending. Zta, which is a divergent member of the bZIP family, bent DNA toward the major groove. A variety of DNA structures can therefore be induced at the AP-1 site through combinatorial interactions between different bZIP family proteins. This diversity of DNA structures may contribute to regulatory specificity among the plethora of proteins that can bind to the AP-1 site.     

Topological comparison of two helix destabilizing proteins: ribonuclease A and the gene 5 DNA binding protein

A topological comparison of the two helix destabilizing proteins, pancreatic ribonuclease A and the gene 5 DNA binding protein of bacteriophage fd has been completed utilizing the available high resolution tertiary structures of each protein. The results indicate these two proteins are structurally if not also evolutionarily related. Regions of closet topological equivalence occur between beta loops directly involved in nucleotide binding or are required for the maintenance of their respective oligonucleotide binding channels. In addition, there is a similar placement of critical amino acid side chains about the binding site. Further evidence for this structural relationship is obtained by comparison of structural data for the mode of complexation of polynucleotides to each protein. The results of topological comparison suggest the essential property shared by helix destabilizing proteins, whether specialized DNA binding proteins such as G5BP or proteins with other primary functional roles, like ribonuclease A, is the presence of an elongated oligonucleotide binding channel. Although ribonuclease A and G5BP are structurally related, it seems likely any protein with this structural feature will exhibit a helix destabilizing capacity. This conclusion is supported by the diversity of molecular characteristics shown by other proteins having this activity.     

DNA binding activities of three murine Jun proteins: stimulation by Fos

Three members of the Jun/AP-1 family have been identified in mouse cDNA libraries: c-Jun, Jun-B, and Jun-D. We have compared the DNA binding properties of the Jun proteins by using in vitro translation products in gel retardation assays. Each protein was able to bind to the consensus AP-1 site (TGACTCA) and, with lower affinity, to related sequences, including the cyclic AMP response element TGACGTCA. The relative binding to the oligonucleotides tested was similar for the different proteins. The Jun proteins formed homodimers and heterodimers with other members of the family, and they were bound to the AP-1 site as dimers. When Fos translation product was present, DNA binding by Jun increased markedly, and the DNA complex contained Fos. The C-terminal homology region of Jun was sufficient for DNA binding, dimer formation, and interaction with Fos. Our general conclusion is that c-Jun, Jun-B, and Jun-D are similar in their DNA binding properties and in their interaction with Fos. If there are functional differences between them, they are likely to involve other activities of the Jun proteins.     

Two hypervariable minisatellite DNA binding proteins.

Hypervariable minisatellite DNA sequences are short, tandemly repeated sequences present at numerous loci in eukaryotes. They stimulate intermolecular homologous recombination up to 13-fold in human cells in culture and may be specific sites for the initiation of recombination in the eukaryotic genome (Wahls, W.P., Wallace, L.J., & Moore, P.D. (1990) Cell 60, 95-103). Reported here is the detection and partial purification of two hypervariable minisatellite DNA binding proteins, called Msbp-2 and Msbp-3, present in the nuclear extracts of human HeLa cells. The proteins elute from a gel filtration column with a native mass of 200-250 kDa and have sizes of 77 kDa and 115 kDa respectively.     

DNA binding and bending to initiate packaging of phage lambda DNA

Physical and genetic studies verify that the DNA binding domain of protein gpNu1 (which initiates packaging of phage lambda DNA) is a winged helix-turn-helix (w HTH) and that gpNu1 dimers bind sites that are brought close through DNA bending.     

Ion concentration and temperature dependence of DNA binding: comparison of PurR and LacI repressor proteins.

Purine repressor (PurR) binding to specific DNA is enhanced by complexing with purines, whereas lactose repressor (LacI) binding is diminished by interaction with inducer sugars despite 30% identity in their protein sequences and highly homologous tertiary structures. Nonetheless, in switching from low- to high-affinity DNA binding, these proteins undergo a similar structural change in which the hinge region connecting the DNA and effector binding domains folds into an alpha-helix and contacts the DNA minor groove. The differences in response to effector for these proteins should be manifest in the polyelectrolyte effect which arises from cations displaced from DNA by interaction with positively charged side chains on a protein and is quantitated by measurement of DNA binding affinity as a function of ion concentration. Consistent with structural data for these proteins, high-affinity operator DNA binding by the PurR-purine complex involved approximately 15 ion pairs, a value significantly greater than that for the corresponding state of LacI (approximately 6 ion pairs). For both proteins, however, conversion to the low-affinity state results in a decrease of approximately 2-fold in the number of cations released per dimeric DNA binding site. Heat capacity changes (DeltaC(p)) that accompany DNA binding, derived from buried apolar surface area, coupled folding, and restriction of motional freedom of polar groups in the interface, also reflect the differences between these homologous repressor proteins. DNA binding of the PurR-guanine complex is accompanied by a DeltaC(p) (-2.8 kcal mol(-1) K(-1)) more negative than that observed previously for LacI (-0.9 to -1.5 kcal mol(-1) K(-1)), suggesting that more extensive protein folding and/or enhanced structural rigidity may occur upon DNA binding for PurR compared to DNA binding for LacI. The differences between these proteins illustrate plasticity of function despite high-level sequence and structural homology and undermine efforts to predict protein behavior on the basis of such similarities.     

DNA binding and bending properties of the post-meiotically expressed Sry-related protein Sox-5.

Sox-5 is one of a family of genes which show homology to the HMG box region of the testis determining gene SRY. We have used indirect immunofluorescence to show that Sox-5 protein is localized to the nucleus of post-meiotic round spermatids in the mouse testis. In vitro footprinting and gel retardation assays demonstrate that Sox-5 binds specifically to the sequence AACAAT with moderately high affinity (Kd of approximately 10(-9) M). Moreover, interaction of Sox-5 with its target DNA induces a significant bend in the DNA, characteristic of HMG box proteins. Circular dichroism spectroscopy of the Sox-5 HMG box and its specific complex with DNA shows an alteration in the DNA spectrum, perhaps as a consequence of DNA bending, but none in the protein spectrum on complex formation. The dependence of the change in the CD spectrum with protein to DNA ratio demonstrates the formation of a 1:1 complex. Analysis of the structure of the Sox-5 HMG box by 2D NMR suggests that both the location of helical secondary structure as well as the tertiary structure is similar to that of HMG1 box 2.