Gary Felsenfeld

Controlling the double helix

Chromatin is the complex of DNA and proteins in which the genetic material is packaged inside the cells of organisms with nuclei. Chromatin structure is dynamic and exerts profound control over gene expression and other fundamental cellular processes. Changes in its structure can be inherited by the next generation, independent of the DNA sequence itself.

Insulators: exploiting transcriptional and epigenetic mechanisms

Insulators are DNA sequence elements that prevent inappropriate interactions between adjacent chromatin domains. One type of insulator establishes domains that separate enhancers and promoters to block their interaction, whereas a second type creates a barrier against the spread of heterochromatin. Recent studies have provided important advances in our understanding of the modes of action of both types of insulator. These new insights also suggest that the mechanisms of action of both enhancer blockers and barriers might not be unique to these types of element, but instead are adaptations of other gene-regulatory mechanisms.

Silencing of transgene transcription precedes methylation of promoter DNA and histone H3 lysine 9

Transgenes stably integrated into cells or animals in many cases are silenced rapidly, probably under the influence of surrounding endogenous condensed chromatin. This gene silencing correlates with repressed chromatin structure marked by histone hypoacetylation, loss of methylation at H3 lysine 4, increase of histone H3 lysine 9 methylation as well as CpG DNA methylation at the promoter. However, the order and the timing of these modifications and their impact on transcription inactivation are less well understood. To determine the temporal order of these events, we examined a model system consisting of a transgenic cassette stably integrated in chicken erythroid cells. We found that histone H3 and H4 hypoacetylation and loss of methylation at H3 lysine 4 all occurred during the same window of time as transgene inactivation in both multicopy and low‐copy‐number lines. These results indicate that these histone modifications were the primary events in gene silencing. We show that the kinetics of silencing exclude histone H3 K9 and promoter DNA methylation as the primary causative events in our transgene system.

The Organization of Histones and DNA in Chromatin: Evidence for an Arginine-Rich Histone Kernel

We have examined the role played by various histones in the organization of the DNA of the nucleosome, using staphylococcal nuclease as a probe of DNA conformation. When this enzyme attacks chromatin, a series of fragments evenly spaced at 10 base pair intervals is generated, reflecting the histone-DNA interactions within the nucleosome structure. To determine what contribution the various histones make to DNA organization, we have studied the staphylococcal nuclease digestion patterns of complexes of DNA with purified histones. Virtually all possible combinations of homogeneous histones were reconstituted onto DNA. Exhaustive digestion of a complex containing the four histones H2A, H2B,H3, and H4 yields a DNA fragment pattern very similar to that of whole chromatin. The only other combinations of histones capable of inducing chromatin-like DNA organization are H2A/H2B/H4 and those mixtures containing both H3 and H4. From an examination of the kinetics of digestion of H3/H4 reconstitutes, we conclude that although the other histones have a role in DNA organization within the nucleosome, the arginine-rich histone pair, H3/H4, can organize DNA segments the length of the nucleosome core in the absence of all other histones.

Higher order structure of chromatin: Orientation of nucleosomes within the 30 nm chromatin solenoid is independent of species and spacer length

We have used electric dichroism to study the arrangement of nucleosomes in 30 nm chromatin solenoidal fibers prepared from a variety of sources (CHO cells, HeLa cells, rat liver, chicken erythrocytes, and sea urchin sperm) in which the nucleosome spacer length varies from approximately 10 to approximately 80 bp. Field-free relaxation times are consistent only with structures containing 6 +/- 1 nucleosomes for every 11 nm of solenoidal length. With very few assumptions about the arrangement of the spacer DNA, our dichroism data are consistent with the same orientation of the chromatosomes for every chromatin sample examined. This orientation, which maintains the faces of the radially arranged chromatosomes inclined at an angle between 20 degrees-33 degrees to the solenoid axis, thus appears to be a general structural feature of the higher order chromatin fiber.

Conformation of polyribouridylic acid in solution

Abstract The conformation of polyribouridylic acid in aqueous solution has been studied under ideal solvent conditions, using light scattering, viscometry and sedimentation velocity and equilibrium measurements. The unperturbed dimensions of the poly rU molecule do not vary greatly with temperature, as expected for a polynucleotide with little or no base stacking. Despite the absence of stacking, the characteristic ratio 2 > nl 2 is 17.6, corresponding to a quite highly extended random coil. Only slightly smaller values are found by extrapolation of the data for poly rA to the completely unstacked form, suggesting that the degree of extension of an unstacked polynucleotide does not depend upon the nature of the base. This degree of extension can only be accounted for if rotation about all of the bonds of the polynucleotide backbone is highly restricted, as has been suggested recently by Sundaralingam (1969) on the basis of his survey of the crystallographic literature. The restrictions he proposes appear to be sufficient to account for our result.

Tissue-specific factors additively increase the probability of the all-or-none formation of a hypersensitive site.

DNase I‐hypersensitive sites lack a canonical nucleosome and have binding sites for various transcription factors. To understand how the hypersensitivity is generated and maintained, we studied the chicken erythroid‐specific beta(A)/epsilon globin gene enhancer, a region where both tissue‐specific and ubiquitous transcription factors can bind. Constructions containing mutations of this enhancer were stably introduced into a chicken erythroid cell line. We found that the hypersensitivity was determined primarily by the erythroid factors and that their binding additively increased the accessibility. The fraction of accessible sites in clonal cell lines was quantitated using restriction endonucleases; these data implied that the formation of each hypersensitive site was an all‐or‐none phenomenon. Use of DNase I and micrococcal nuclease probes further indicated that the size of the hypersensitive site was influenced by the binding of transcription factors which then determined the length of the nucleosome‐free gap. Our data are consistent with a model in which hypersensitive sites are generated stochastically: mutations that reduce the number of bound factors reduce the probability that these factors will prevail over a nucleosome; thus, the fraction of sites in the population that are accessible is also diminished.

The dispersion of the hyperchromic effect in thermally induced transitions of nucleic acids

The increase in optical absorbance in the ultraviolet light band which occurs when nucleic acids are heated has been shown to have a different functional dependence upon temperature at each wavelength, because there is a correlation between the denaturation temperature of a denaturing unit and its absorption spectrum. The dependence of the median temperature (Tm) of absorbance increase upon wavelength is plotted to give an optical denaturation–dispersion curve characteristic of the nucleic acid. It is shown that the magnitude of the dispersion in native calf thymus DNA is much greater than in native T4 phage DNA, either before or after shearing of the phage DNA. This result gives information about the composition of denaturation “nuclei” in the phage DNA molecule. Multiple wavelength data are also analysed to give separate curves showing the degree of denaturation of adenine–thymine (A–T) base pairs and guanine–cytosine (G–C) base pairs as a function of temperature. It is shown that in native calf thymus DNA the values of Tm for A–T and G–C pairs are within a few degrees of one another, but that in reheated denatured calf thymus DNA the values of Tm are widely separated. The shape of the A–T and G–C curves in the denatured DNA is shown to be interpretable in terms.of structural models previously proposed by other investigators. A similar analysis of S-RNA denaturation shows that the structure of S-RNA is different from that of denatured calf thymus DNA. It is concluded that the longest regions of ordered structure in S-RNA consist of some sequences relatively pure in G–C base pairs, and other sequences relatively pure in A–U base pairs.

Chromatin structure as probed by nucleases and proteases: Evidence for the central role of hitones H3 and H4

We have examined the role by each histone in forming the structure of the nu-body. When DNAase I, DNAase II, trypsin and chymotrypsin attack chromatin, characteristic discrete DNA and protein digest fragments are produced. Using this restriction of accessibility as diagnostic for chromatin structure, we have examined complexes of DNA with virtually all possible combinations of histones. The results strongly support our previous conclusion (Camerini-Otero, Sollner-Webb, and Felsenfeld, 1976) that the arginine-rich histones are unique in their ability to create, with DNA a structure with many features of native chromatin. Acting together, slightly lysine-rich histones then modify this complex into one very similar to native chromatin. An analysis of the rate constants of staphylococcal nuclease digestion also confirms that the complex of H3, H4, and DNA is crucial to the structure of the nu-body.

Transcription of chromatin in vitro

Abstract The interaction of Escherichia coli RNA polymerase with calf thymus DNA and chromatin has been studied under conditions which permit direct measurement of binding sites capable of supporting chain initiation. The number of such sites was determined by counting the number of growing RNA chains, using either sucrose gradient ultracentrifugation or incorporation of [γ-32P]ATP and [γ-32P]GTP. The number of sites was also determined by titrating template with polymerase and counting the number of molecules bound at the end-point. Both measurements show that there is one site per 1000 to 1400 nucleotide pairs on DNA and one tenth that number on chromatin. The rate of chain elongation on chromatin is about one-third that on DNA, under the assay conditions used. Measurements were made under conditions in which chromatin was soluble. The results show that the number of binding sites for polymerase on chromatin is much smaller than on DNA, but that all bound enzyme molecules are capable of chain elongation. They also show that chromatin solubility is not a factor in template restriction. Studies of chromatin template activity in vitro have generally made use of the rate of RNA synthesis as a measure of the amount of template present. The methods presented here make it possible to distinguish between the number of growing chains and the rate of chain elongation and therefore provide an unambiguous determination of template concentration.