Topological effects of the TATA box binding protein on minicircle DNA and a possible thermodynamic linkage to chromatin remodeling.

DNA ring closure experiments on short restriction fragments ( approximately 160 bp) bound by the TATA box binding protein (TBP) have demonstrated the formation of negative topoisomers, consistent with crystallographically observed TBP-induced DNA untwisting but in contrast to most previous results on topological effects in plasmid DNA. The difference may be due to the high free energy cost of substantial writhe in minicircles. A speculative mechanism for the loss of TBP-induced writhe suggests that TBP is capable of inducing DeltaTw between 0 and -0.3 in minicircles, via loss of out-of-plane bending upon retraction of intercalating Phe stirrups, and that TBP can thus act as a "supercoil shock absorber". The proposed biological relevance of these observations is that they may model the behavior of DNA in constrained chromatin environments. Irrespective of the detailed mechanism of TBP-induced supercoiling, its existence suggests that chromatin remodeling and enhanced TBP binding are thermodynamically linked. Remodeling ATPases or histone acetylases release some of the negative supercoiling previously restrained by the nucleosome. When TBP takes up the supercoiling, its binding should be enhanced transiently until the unrestrained supercoiling is removed by diffusion or topoisomerases. The effect is predicted to be independent of local remodeling-induced changes in TATA box accessibility.

Designed hyperstable Lac repressor.DNA loop topologies suggest alternative loop geometries.

Lac repressor (LacI) forms DNA loops which are critical for efficient operator binding and transcriptional repression. Designed DNA loops formed on three constructs with lac operators bracketing phased A-tract bends were characterized by mobility shift, footprinting, and DNA cyclization and topology. Operator dyad axes point either in or out relative to the sequence-induced curvature. Possible conformations suggested from X-ray structures of LacI and LacI.DNA include "wrapping away" (WA), "simple loop" (SL), and "wrapping toward" (WT) models. The WA loop should be preferentially stabilized by the outward operators, the SL and WT loops by the inward operators. Competition experiments demonstrated increased loop stability for all the bent constructs, with the SL/WT construct supporting hyperstable loops (t1/2 of days). This offers a general approach to stabilizing multi-protein DNA complexes on short DNA. DNA cyclization of loops gave minicircle products with altered topologies. WA constructs afforded relaxed and positive topoisomers, and cyclization kinetics indicated slow interconversion of precursors to the two topoisomers. The SL/WT construct gave a relaxed topoisomer, with a small amount of negative supercoil. These results suggest that while it is possible to force the WA loop to form (as in a model proposed from the LacI.DNA structure), the most stable loop geometry is different, probably a U-shape around an extended LacI tetramer. The topological results show how a protein-induced positive supercoil can be reconciled with LacI's preference for binding negatively supercoiled DNA. We suggest that looping proteins can affect the assembly of subsequent proteins by controlling loop topology.

TATA box DNA deformation with and without the TATA box-binding protein.

DNA ring closure methods have been applied to TATA box DNA and its complex with the TATA box-binding protein (TBP). The J factors for cyclization (effective concentrations of one DNA end about the other) have been measured using cyclization kinetics, with and without bound TBP, for 18 DNA constructs containing the adenovirus major late promoter TATA box (TATAAAAG) separated by a variable helical phasing adapter from sequence-induced A-tract DNA bends. Six phasing lengths were used at three overall DNA lengths each. Cyclization kinetics were also measured in the absence of protein for the same set of molecules bearing a mutant TATA box (TACAAAAG). The results suggest that the TATA box DNA itself is strongly bent and anisotropically flexible, in a direction opposite to the bend induced by TBP, and that the mutant TACA box is much less bent/flexible. The bending and flexibility of the free DNA may govern the energetics of recognition of different DNA sequences by TBP, and the intrinsic bend may act to repress transcription complex assembly in the absence of TBP. The cyclization kinetics of TBP-DNA complexes in solution predict a geometry generally consistent with crystal structures, which show dramatic bending and unwinding. The novel observation of TBP-induced topoisomers suggests that this minicircle approach is able to distinguish TBP-induced unwinding from writhe (these cancel out in larger DNA), and this in turn suggests that changes in supercoiling in small topological domains can control TBP binding.

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.

Measurement of the DNA bend angle induced by the catabolite activator protein using Monte Carlo simulation of cyclization kinetics.

A Monte Carlo simulation method for studying DNA cyclization (or ring-closure) has been extended to the case of protein-induced bending, and its application to experimental data has been demonstrated. Estimates for the geometric parameters describing the DNA bend induced by the catabolite activator protein (CAP or CRP) were obtained which correctly predict experimental DNA cyclization probabilities (J factors), determined for a set of 11 150 to 166 bp DNA restriction fragments bearing A tracts phased against CAP binding sites. We find that simulation of out-of-phase molecules is difficult and time consuming, requiring the geometric parameters to be optimized individually rather than globally. A wedge angle model for DNA bending was found to make reasonable predictions for the free DNA. The bend angle in the CAP-DNA complex is estimated to be 85 to 90 degrees, in agreement with estimates from gel electrophoresis and X-ray co-crystal structures. Since the DNA is found to have a pre-existing bend of 15 degrees, the change in bend angle induced by CAP is 70 to 75 degrees, in a agreement with an estimate from topological measurements. We find evidence for slight (approximately 10 degrees) unwinding by CAP. The persistence length and helical repeat of the unbound portion of the DNA are in accord with literature-cited values, but the best-fit DNA torsional modulus C is found to be 1.7 (+/- 0.2) x 10(-19) erg. cm, versus literature estimates and best-fit values for the free DNA of 2.0 x 10(-19) to 3.4 x 10(-19) erg.com. Simulations using this low value of C predict that cyclization of molecules with out-of-phase bends proceeds via undertwisting or overtwisting of the DNA between the bends, so as to align the bends, rather than through conformations with substantial writhe. We present experiments on the topoisomers formed by cyclization with CAP which support this conclusion, and thereby rationalize the surprising result that cyclization can actually be enhanced by out-of-phase bends if the twist required to align the bends improves the torsional alignment of the ends. The relationship between the present work and previous studies on DNA bending by CAP is discussed, and recommendations are given for the efficient application of the cyclization/simulation approach to DNA bending.

Identification and characterization of genomic nucleosome-positioning sequences.

Positioned nucleosomes are believed to play important roles in transcriptional regulation and for the organization of chromatin in cell nuclei. Here, we have isolated the DNA segments in the mouse genome that form the most stable nucleosomes yet characterized. In separate molecules we find phased runs of three to four adenine nucleotides, extensive CA repeats, and in a few cases phased TATA tetranucleotides. The latter forms the most stable nucleosome yet characterized. One sequence with CAG repeats was also found. By fluorescence in situ hydridization the selected sequences are shown to be localized at the centromeric regions of mouse metaphase chromosomes.

Sequence elements responsible for DNA curvature.

Intrinsic DNA bending or curvature is a phenomenon that has been shown to play an important role in a variety of DNA transactions. Large curvature occurs when short homopolymeric (dA.dT)4-6 runs (A-tracts) are repeated in phase with the helical screw. We have used electrophoretic mobility modulation to examine how bending depends on the nature of the 5 bp DNA sequence between the A tracts in molecules of the form (A5-6N5)n. We show that A-tract-induced DNA curvature can indeed be affected by other sequence elements, although by only about +/- 10%. The small observed curvature modulation implies that the overall helix axis deflection contributed by 5-bp B-DNA segments between A-tracts varies little from one sequence to another. This result validates, to first order, the assumption that DNA curvature results from inserting A-tracts at integral turn phasing into generic B-DNA. Therefore, if, as has been proposed, A-tracts have zero roll between the base-pairs and all curvature results from positive roll in the B-DNA segments, then this must be a general property of approximately 5 bp B-DNA sequences, not just special cases. This interpretation would require that the canonical structure of B-DNA be revised to include systematic roll between the base-pairs of about 6 degrees. Alternatively, the data are also consistent with zero average roll in the B-DNA sequences, and wedge angles dominated by negative roll in the A-tracts, or with an appropriate mixture of the two models. It is not possible to resolve this ambiguity using comparative electrophoresis or existing structural data. We show that published wedge angle parameters successfully predict the measured direction and, with appropriate rescaling, the magnitude of curvature due to a non-A-tract sequence containing the protein-free lac operator CAP protein binding site.

Detection of localized DNA flexibility.

The bending and flexibility of DNA are important in packaging, recombination and transcription. Bending decreases electrophoretic mobility in a manner depending on bend position within a fragment (circular permutation) and on the distance between bends (phasing analysis). Bending can also affect DNA ring closure (cyclization). The lack of a complete theory for the mechanism of gel retardation hampers measurement of bend magnitudes by electrophoresis, whereas cyclization is done entirely in solution and is well understood theoretically. Disagreements between bend angles estimated by the two electrophoretic assays have been ascribed to DNA flexibility. Here we test this interpretation using an internal loop as a model flexible locus. Whereas the circular permutation and helical phasing experiments are only subtly affected by the loop, DNA cyclization kinetics detects and quantifies substantial increases in torsional and bending flexibility. Furthermore, the results support a functional role for the stress of DNA bending in inducing base-pair opening.

Sequence and structure of cmp, the replication enhancer of the Staphylococcus aureus plasmid pT181.

The Staphylococcus aureus plasmid pT181 possesses a DNA replication enhancer element, called cmp, that is required in cis for optimal utilization of the initiator protein by the origin of replication. The minimal nucleotide sequence required for cmp activity was defined by testing progressively smaller DNA fragments for their ability to restore cmp activity in a plasmid mutant deleted for cmp. These experiments indicate that cmp is a sequence of 100 base pairs (bp) characterized by a loosely repeated sequence motif and phased oligo(dT) tracts. Intrinsic DNA bending at cmp was detected by a circular permutation assay of the locus using polyacrylamide-gel electrophoresis and by computer modeling. The cmp element was found to contain two loci of intrinsically bent DNA that confer an overall bent conformation to this replication enhancer.

Protein-induced bending and DNA cyclization.

We have applied T4 ligase-mediated DNA cyclization kinetics to protein-induced bending in DNA. The presence and direction of a static bend can be inferred from J factors for cyclization of 150- to 160-base-pair minicircles, which include a catabolite activator protein binding site phased against a sequence-directed bend. We demonstrate a quasi-thermodynamic linkage between cyclization and protein binding; we find that properly phased DNAs bind catabolite activator protein approximately 200-fold more tightly as circles than as linear molecules. The results unambiguously distinguish DNA bends from isotropically flexible sites and can explain cooperative binding by proteins that need not contact each other.

RNA folding during transcription by Escherichia coli RNA polymerase analyzed by RNA self-cleavage.

We have used a self-cleaving RNA molecule related to a subsequence of plant viroids (a "hammerhead") to study the length-dependent folding of RNA produced during transcription by Escherichia coli RNA polymerase. Transcript elongation is arrested at defined positions using chain-terminating ribonucleoside triphosphate analogues (3'-deoxyNTP's or 3'-O-methylNTP's). When the transcript can form the "hammerhead" structure it self-cleaves to give a truncated product. The experiment yields an RNA sequencing ladder which terminates at the length at which cleavage becomes possible; the sequencing ladder is compared to those generated by using a noncleaving transcript or by using [alpha-thio]ATP in place of ATP. We have shown that 15-18 nucleotides (nt) of RNA past the cleavage point must be synthesized before the transcript can self-cleave within a ternary complex, whereas RNA freed from the complex by heating can cleave with only 3 or more nt present beyond the cleavage point. There are sequence-dependent as well as length-dependent effects. The results suggest that 12 +/- 1 nt are sequestered within the ternary complex and are consistent with the presence of a DNA-RNA hybrid within the transcription bubble, as proposed by others. The results indicate that the "hammerhead" structure does not disrupt the hybrid. It appears that the RNA beyond the hybrid is not restrained by interactions with the enzyme, since the last stem of the self-cleaving structure forms as soon as the RNA composing it emerges from the DNA-RNA hybrid. Self-cleaving of the transcript offers a simple structural probe for studying less well-characterized transcription complexes. The relevance of the results to models for transcription termination is discussed.

Reversibility of nucleotide incorporation by Escherichia coli RNA polymerase, and its effect on fidelity.

During transcription, Escherichia coli RNA polymerase is capable of removing the nucleotide that it has just added to a growing RNA chain, and this removal depends on the presence of small concentrations of pyrophosphate. Chemically, the removal reaction is simply the reversal of the incorporation reaction, and we have observed the generation of free triphosphate as a result. After the removal the enzyme can continue synthesis. To test whether this reaction can provide an error correction mechanism, misincorporation rates were measured at a single position in an RNA transcript by withholding the correct nucleotide for that position, measuring the amount of readthrough transcript, and analyzing the readthrough transcripts with nearest-neighbor analysis and enzymatic RNA sequencing. The removal of pyrophosphate increases the rate of misincorporation. We present a theory that explains how reversible incorporation can increase the available discrimination free energy between correct and incorrect nucleotides and therefore may increase the fidelity of transcription. The formation of a covalent phosphodiester bond allows discrimination on the basis of helical structure as well as base-pairing. We propose that the important discrimination step is the translocation of the enzyme from one site on the DNA template to the next, and that reversible incorporation is necessary in order to take full advantage of the maximum discrimination free energy.