Measurement of histone-DNA interaction free energy in nucleosomes.

Nucleosome positioning DNA sequences are of increasing interest because of their proposed roles in gene regulation and other chromosome functions in vivo, and because they have revealed new insights into the sequence-dependent structures and mechanics of DNA itself. Here, we describe methods to quantify the relative affinities of histone-DNA interactions in nucleosomes, i.e., the nucleosome positioning power of differing DNA sequences. We review methods developed by others and then discuss in detail our own approach to measurement of histone-DNA interaction free energies. Compared to earlier methods, our dialysis-based approach reduces the possibility that non-equilibrium or irreproducible results could be obtained. It facilitates a direct comparison of free energies for many sequences at the same time and it allows analysis of DNAs having a wide range of relative affinities.

Effects of histone acetylation on the equilibrium accessibility of nucleosomal DNA target sites.

Posttranslational acetylation of the conserved core histone N-terminal tail domains is linked to gene activation, but the molecular mechanisms involved are not known. In an earlier study we showed that removing the tail domains altogether by trypsin proteolysis (which leaves nucleosomes nevertheless intact) leads to 1.5 to 14-fold increases in the dynamic equilibrium accessibility of nucleosomal DNA target sites. These observations suggested that, by modestly increasing the equilibrium accessibility of buried DNA target sites, histone acetylation could result in an increased occupancy by regulatory proteins, ultimately increasing the probability of transcription initiation. Here, we extend these observations to a more natural system involving intact but hyperacetylated nucleosomes. We find that histone hyperacetylation leads to 1.1 to 1.8-fold increases in position-dependent equilibrium constants for exposure of nucleosomal DNA target sites, with an average increase of 1.4(+/-0.1)-fold. The mechanistic and biological implications of these results are discussed.

Effects of histone tail domains on the rate of transcriptional elongation through a nucleosome.

The N-terminal tail domains of the core histones play important roles in gene regulation, but the exact mechanisms through which they act are not known. Recent studies suggest that the tail domains may influence the ability of RNA polymerase to elongate through the nucleosomal DNA and, thus, that posttranslational modification of the tail domains may provide a control point for gene regulation through effects on the elongation rate. We take advantage of an experimental system that uses bacteriophage T7 RNA polymerase as a probe for aspects of nucleosome transcription that are dominated by the properties of nucleosomes themselves. With this system, experiments can analyze the synchronous, real-time, single-passage transcription on the nucleosomal template. Here, we use this system to directly test the hypothesis that the tail domains may influence the "elongatability" of nucleosomal DNA and to identify which of the tail domains may contribute to this. The results show that the tail domains strongly influence the rate of elongation and suggest that the effect is dominated by the N-terminal domains of the (H3-H4)(2) tetramer. They further imply that tail-mediated octamer transfer is not essential for elongation through the nucleosome. Acetylation of the tail domains leads to effects on elongation that are similar to those arising from complete removal of the tail domains.

Effects of core histone tail domains on the equilibrium constants for dynamic DNA site accessibility in nucleosomes.

The N and C-terminal tail domains of the core histones play important roles in gene regulation, but the mechanisms through which they act are not known. These tail domains are highly positively charged and are the sites of numerous post-translational modifications, including many sites for lysine acetylation. Nucleosomes in which these tail domains have been removed by trypsin remain otherwise intact, and are used by many laboratories as a model system for highly acetylated nucleosomes. Here, we test the hypothesis that one role of the tail domains is to directly regulate the accessibility of nucleosomal DNA to other DNA-binding proteins. Three assays are used: equilibrium binding by a site-specific, DNA-binding protein, and dynamic accessibility to restriction enzymes or to a non-specific exonuclease. The effects of removal of the tail domains as monitored by each of these assays can be understood within the framework of the site exposure model for the dynamic equilibrium accessibility of target sites located within the nucleosomal DNA. Removal of the tail domains leads to a 1.5 to 14-fold increase in position-dependent equilibrium constants for site exposure. The smallness of the effect weighs against models for gene activation in which histone acetylation is a mandatory initial event, required to facilitate subsequent access of regulatory proteins to nucleosomal DNA target sites. Alternative roles for histone acetylation in gene regulation are discussed.

Sequence motifs and free energies of selected natural and non-natural nucleosome positioning DNA sequences.

Our laboratories recently completed SELEX experiments to isolate DNA sequences that most-strongly favor or disfavor nucleosome formation and positioning, from the entire mouse genome or from even more diverse pools of chemically synthetic random sequence DNA. Here we directly compare these selected natural and non-natural sequences. We find that the strongest natural positioning sequences have affinities for histone binding and nucleosome formation that are sixfold or more lower than those possessed by many of the selected non-natural sequences. We conclude that even the highest-affinity sequence regions of eukaryotic genomes are not evolved for the highest affinity or nucleosome positioning power. Fourier transform calculations on the selected natural sequences reveal a special significance for nucleosome positioning of a motif consisting of approximately 10 bp periodic placement of TA dinucleotide steps. Contributions to histone binding and nucleosome formation from periodic TA steps are more significant than those from other periodic steps such as AA (=TT), CC (=GG) and more important than those from the other YR steps (CA (=TG) and CG), which are reported to have greater conformational flexibility in protein-DNA complexes even than TA. We report the development of improved procedures for measuring the free energies of even stronger positioning sequences that may be isolated in the future, and show that when the favorable free energy of histone-DNA interactions becomes sufficiently large, measurements based on the widely used exchange method become unreliable.

New DNA sequence rules for high affinity binding to histone octamer and sequence-directed nucleosome positioning.

DNA sequences that position nucleosomes are of increasing interest because of their relationship to gene regulation in vivo and because of their utility in studies of nucleosome structure and function in vitro. However, at present our understanding of the rules for DNA sequence-directed nucleosome positioning is fragmentary, and existing positioning sequences have many limitations. We carried out a SELEX experiment starting with a large pool of chemically synthetic random. DNA molecules to identify those individuals having the highest affinity for histone octamer. A set of highest-affinity molecules were selected, cloned, and sequenced, their affinities (free energies) for histone octamer in nucleosome reconstitution measured, and their ability to position nucleosomes in vitro assessed by native gel electrophoresis. The selected sequences have higher affinity than previously known natural or non-natural sequences, and have a correspondingly strong nucleosome positioning ability. A variety of analyses including Fourier transform, real-space correlation, and direct counting computations were carried out to assess non-random features in the selected sequences. The results reveal sequence rules that were already identified in earlier studies of natural nucleosomal DNA, together with a large set of new rules having even stronger statistical significance. Possible physical origins of the selected molecules' high affinities are discussed. The sequences isolated in this study should prove valuable for studies of chromatin structure and function in vitro and, potentially, for studies in vivo.

Nucleosome packaging and nucleosome positioning of genomic DNA.

The goals of this study were to assess the extent to which bulk genomic DNA sequences contribute to their own packaging in nucleosomes and to reveal the relationship between nucleosome packaging and positioning. Using a competitive nucleosome reconstitution assay, we found that at least 95% of bulk DNA sequences have an affinity for histone octamer in nucleosomes that is similar to that of randomly synthesized DNA; they contribute little to their own packaging at the level of individual nucleosomes. An equation was developed that relates the measured free energy to the fractional occupancy of specific nucleosome positions. Evidently, the bulk of eukaryotic genomic DNA is also not evolved or constrained for significant sequence-directed nucleosome positioning at the level of individual nucleosomes. Implications for gene regulation in vivo are discussed.

Twist constraints on linker DNA in the 30-nm chromatin fiber: implications for nucleosome phasing.

Previous work has shown that nucleosome repeat lengths, and hence linker DNA lengths, are preferentially quantized to a set of values differing by integral multiples of the helical twist of DNA. An explanation was proposed in which this preferential quantitation is due to twist constraints on linker DNA arising from nucleosome-nucleosome interactions in folded chromatin. Here we report the results of a study, using ethidium intercalation, designed to test whether twist constraints do indeed exist. Electron microscopy reveals that ethidium intercalation causes decondensation of dinucleosomes. Direct measurement of the free energy of intercalation by fluorescence spectroscopy reveals competition between chromatin folding and ethidium intercalation. Results from other laboratories establish that these effects of ethidium are due to ethidium-induced changes in the twist of linker DNA, and not to a variety of other possible effects. We conclude that twist constraints on linker DNA do exist. These may explain the observation of preferentially quantized linker DNA lengths. Implications of these results for mechanisms of nucleosome phasing and the mechanisms of drug action are discussed.

Co-operative binding of the globular domain of histone H5 to DNA.

The globular domain of histone H5 (GH5) was prepared by trypsin digestion of H5 that was extracted from chicken erythrocyte nuclei with NaCl. Electron microscopy, sucrose gradient centrifugation, native agarose gel electrophoresis and equilibrium density gradient ultracentrifugation show that GH5 binds co-operatively to double-stranded DNA. The electron microscopic images suggest that the GH5-DNA complexes are very similar in structure to co-operative complexes of intact histone H1 (or its variants) with double-stranded DNA, studied previously, which have been proposed to consist of two parallel DNA double helices sandwiching a polymer of the protein. For complexes with GH5 or with intact H1, naked DNA co-sediments with the protein-DNA complexes through sucrose gradients, and DNA also appears to protrude from the ends and sides of the complexes; measurements of the protein-DNA stoichiometry in fractionated samples may not reflect the stoichiometry in the complexes. An estimate of the stoichiometry obtained from the buoyant density of fixed GH5-DNA complexes in CsCl suggests that sufficient GH5 is present in the complexes for the GH5s to be in direct contact, as required by a simple molecular mechanism for the co-operative binding. Chemical crosslinking demonstrates that GH5s are in close proximity in the complexes. In the absence of DNA, GH5-GH5 interactions are weak or non-existent.

Linker DNA bending induced by the core histones of chromatin.

We have previously reported that ionic conditions that stabilize the folding of long chromatin into 30-nm filaments cause linker DNA to bend, bringing the two nucleosomes of a dinucleosome into contact [Yao, J., Lowary, P. T., & Widom, J. (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 7603-7607]. Dinucleosomes are studied because they allow the unambiguous detection of linker DNA bending through measurement of their nucleosome-nucleosome distance. Because of the large resistance of DNA to bending, the observed compaction must be facilitated by the histones. We have now tested the role of histone H1 (and its variant, H5) in this process. We find that dinucleosomes from which the H1 and H5 have been removed are able to compact to the same extent as native dinucleosomes; the transition is shifted to higher salt concentrations. We conclude that histone H1 is not essential for compacting the chromatin filament. However, H1 contributes to the free energy of compaction, and so it may select a single, ordered, compact state (the 30-nm filament, in long chromatin) from a family of compact states which are possible in its absence.

Direct detection of linker DNA bending in defined-length oligomers of chromatin.

Linker DNA, which connects between nucleosomes in chromatin, is short and, therefore, may be essentially straight and inflexible. We have carried out hydrodynamic and electron microscopic studies of dinucleosomes--fragments of chromatin containing just two nucleosomes--to test the ability of linker DNA to bend. We find that ionic conditions that stabilize the folding of long chromatin cause linker DNA in dinucleosomes to bend, bringing the two nucleosomes into contact. The results uphold a key prediction of the solenoid model of chromosome folding and suggest a mechanism by which proteins that are separated along the DNA can interact by direct contact.

Higher-order structure of Saccharomyces cerevisiae chromatin.

We have developed a method for partially purifying chromatin from Saccharomyces cerevisiae (baker's yeast) to a level suitable for studies of its higher-order folding. This has required the use of yeast strains that are free of the ubiquitous yeast "killer" virus. Results from dynamic light scattering, electron microscopy, and x-ray diffraction show that the yeast chromatin undergoes a cation-dependent folding into 30-nm filaments that resemble those characteristic of higher-cell chromatin; moreover, the packing of nucleosomes within the yeast 30-nm filaments is similar to that of higher cells. These results imply that yeast has a protein or protein domain that serves the role of the histone H 1 found in higher cells; physical and genetic studies of the yeast activity could help elucidate the structure and function of H 1. Images of the yeast 30-nm filaments can be used to test crossed-linker models for 30-nm filament structure.

An RNA mutation that increases the affinity of an RNA-protein interaction.

The introduction of a cytidine in place of one of the two single stranded uridines in the R17 replicase translational operator results in a much tighter binding to R17 coat protein. The complex containing the variant RNA is stable to gel electrophoresis and has a binding constant about 50 times greater than the one with wild type RNA. The nearly thirty percent increase in the free energy of binding for the variant RNA is primarily due to a more favorable enthalpy of interaction. A possible explanation for this surprising result is that the U to C change leads to a greater extent of formation of a transient covalent complex between the protein and the RNA.

Interaction of R17 coat protein with its RNA binding site for translational repression.

The interaction between bacteriophage R17 coat protein and its RNA binding site for translational repression was studied as an example of a sequence-specific RNA-protein interaction. A nitrocellulose filter retention assay is used to demonstrate equimolar binding between the coat protein and a synthetic 21 nucleotide RNA fragment. The Kd at 2 degrees C in a buffer containing 0.19 M salt is about 1 nM. The relatively weak ionic strength dependence of Ka and a delta H = -19 kcal/mole indicates that most of the binding free energy is due to non-electrostatic interactions. Since a variety of RNAs failed to compete with the 21 nucleotide fragment for coat protein binding, the interaction appears highly sequence specific. We have synthesized more than 30 different variants of the binding site sequence in order to identify the portions of the RNA molecule which are important for protein binding. Out of the five single stranded residues examined, four were essential for protein binding whereas the fifth could be replaced by any nucleotide. One variant was found to bind better than the wild type sequence. Substitution of nucleotides which disrupted the secondary structure of the binding fragment resulted in very poor binding to the protein. These data indicated that there are several points of contact between the RNA and the protein and the correct hairpin secondary structure of the RNA is essential for protein binding.

Interaction of R17 coat protein with synthetic variants of its ribonucleic acid binding site.

The specificity of the interaction between R17 coat protein and its site of translational repression on R17 RNA was studied by enzymatically synthesizing 23 sequence variants of the RNA binding site and measuring their affinity to the coat protein by a nitrocellulose filter binding assay. Experiments using oligomers truncated on the 3' and 5' termini allowed precise determination of the edges of the binding domain. Several oligomers which disrupted one or more of the base pairs in the binding site failed to bind coat protein, establishing the importance of RNA secondary structure for the interaction. Substitution at two single-stranded positions with each of the common bases affected Ka very differently. In one case, Ka was reduced substantially no matter which base was substituted for an adenine. At the other position, when a uracil was substituted with a purine, Ka decreased 10-100-fold, whereas when it was substituted by a cytosine, Ka increased about 5-fold. These studies indicate that the protein and the RNA hairpin loop interact over an extensive area and that several different types of contacts form to stabilize the complex.

Binding of yeast tRNAPhe anticodon arm to Escherichia coli 30 S ribosomes.

A 15-nucleotide fragment of RNA having the sequence of the anticodon arm of yeast tRNAPhe was constructed using T4 RNA ligase. The stoichiometry and binding constant of this oligomer to poly(U)-programmed 30 S ribosomes was found to be identical to that of deacylated tRNAPhe. The anticodon arm and tRNAPhe also compete for the same binding site on the ribosome. These data indicate that the interaction of tRNAPhe with poly(U)-programmed 30 S ribosomes is primarily a result of contacts in the anticodon arm region and not with other parts of the transfer RNA. Since similar oligomers which cannot form a stable helical stem do not bind ribosomes, a clear requirement for the entire anticodon arm structure is demonstrated.

Role of the constant uridine in binding of yeast tRNAPhe anticodon arm to 30S ribosomes.

Twenty-two anticodon arm analogues were prepared by joining different tetra, penta, and hexaribonucleotides to a nine nucleotide fragment of yeast tRNAPhe with T4 RNA ligase. The oligomer with the same sequence as the anticodon arm of tRNAPhe bind poly U programmed 30S ribosomes with affinity similar to intact tRNAPhe. Analogues with an additional nucleotide in the loop bind ribosomes with a weaker affinity whereas analogues with one less nucleotide in the loop do not bind ribosomes at all. Reasonably tight binding of anticodon arms with different nucleotides on the 5' side of the anticodon suggest that positions 32 and 33 in the tRNAPhe sequence are not essential for ribosome binding. However, differences in the binding constants for anticodon arms containing modified uridine residues in the "constant uridine" position suggest that both of the internal "U turn" hydrogen bonds predicted by the X-ray crystal structure are necessary for maximal ribosome binding.