Periodicity of deoxyribonuclease I digestion of chromatin.

Two methods have been used to measure the single-strand lengths of the DNA fragments produced by deoxyribonuclease I digestion of chromatin. The average lengths obtained are muliples of about 10.4 bases, significantly different from the value of 10 previously reported. This periodicity in fragment lengths is closely related to the periodicity of the DNA double helix in chromatin, but the two values need not be exactly the same.

Chromatin reconstitution on small DNA rings. V. DNA thermal flexibility of single nucleosomes.

The thermal flexibility of DNA minicircles reconstituted with single nucleosomes was measured relative to the naked minicircles. The measurement used a new method based on the electrophoretic properties of these molecules, whose mobility strongly depended on the DNA writhe, either of the whole minicircle, when naked, or of the extranucleosomal loop, when reconstituted. The experiment was as follows. The DNA length was first increased by one base-pair (bp), and the correlative shift in mobility resulting from the altered DNA writhe was recorded. Second, the gel temperature was increased so that the former mobility was restored. Under these conditions, the untwisting of the thermally flexible DNA due to the temperature shift exactly compensates for the increase in the DNA mean twist number resulting from the one bp addition. The relative thermal flexibility was then calculated as the ratio between the increases in temperature measured for the naked and the reconstituted DNAs, respectively. The figure, 0.69 (+/- 0.07), was used to derive the length of DNA in interaction with the histones, 109 (+/- 25) bp. Such length was in good agreement with the mean value of 115 bp we have previously obtained from the distribution of the angles between DNAs at the entrance and exit of similar nucleosomes measured from high resolution electron microscopy. This consistency further reinforces our previous conclusion that minicircle-reconstituted nucleosomes, with 1.3(109/83) to 1.4(115/83) turns of superhelical DNA, show no crossing of entering and exiting DNAs when the loop is in its most probable configuration, and therefore, that these nucleosomes behave topologically as "single-turn" particles. The present data are also within the range of values, 50 to 100 bp of thermally rigid DNA per nucleosome, obtained by others for yeast plasmid chromatin, suggesting that the "single-turn" particle notion may be extended to this particular case of naturally-occurring H1-free chromatin. However, these data are quite different from the 230 bp figure derived from thermal measurements of reconstituted H1-free minichromosomes. It is proposed that nucleosome interactions occurring in this chromatin, but not in yeast chromatin, may be partly responsible for the discrepancy.

Nucleosome dynamics V. Ethidium bromide versus histone tails in modulating ethidium bromide-driven tetrasome chiral transition. A fluorescence study of tetrasomes on DNA minicircles.

Protein and DNA contributions in the chiral transition of DNA minicircle-reconstituted tetrasomes (the particles made of DNA wrapped around the histone (H3-H4)(2) tetramer) to a right-handed conformation have been investigated in a recent article from this laboratory. As the evidence for a protein contribution, a sterical hindrance introduced at the H3/H3 interface of the two constituent H3-H4 dimers by oxidation of H3 cysteine 110 blocked the tetramer in a half-left-handed or semi-right-handed conformation, depending on the SH-reagent used. The DNA contributed at the level of the dyad region, which appeared to act through its sequence-dependent deformability in modulating both the loop threshold positive constraint required to trigger the transition, and the tetrasome lateral opening. This opening, which electron microscopic visualizations directly showed to be associated with the transition, is expected to help remove the clash between the entering and exiting DNAs. In this work, the transition mechanism was further investigated by applying a positive constraint in the loop through ethidium bromide (EtBr) intercalation. This technique, including the determination of binding isotherms, has first been used with mononucleosomes on DNA minicircles, and has revealed that these particles could tolerate large positive supercoilings without disruption, owing to the loop ability to cross positively in a histone tail-dependent manner. The transition of 359 bp tetrasomes was found to go to completion in lower salt (10 mM), but not in higher salt (100 mM), whereas the transition of 256 bp tetrasomes was already hindered in lower salt. Histone acetylation relieved that lower salt hindrance but enhanced the higher salt hindrances. These data again pointed to the DNA in the dyad region as a regulator of the transition. The block was indeed expected to originate from a local EtBr intercalation in that DNA, which opposed its overtwisting during the transition. The occurrence of the block, or its relief, then depended on the outcome of the competition between the tails and EtBr for binding to that region, that is, on whether the tails could prevent EtBr intercalation before the ongoing transition hampered both bindings. Destabilization of the tails in the course of the transition is documented in an accompanying article through a relaxation study of a 351-366 bp tetrasome series.

Properties of supercoiled DNA in gel electrophoresis. The V-like dependence of mobility on topological constraint. DNA-matrix interactions.

The dependence of the electrophoretic mobility of small DNA rings on topological constraint was investigated in acrylamide or agarose gels as a function of DNA size (from approximately 350 to 1400 base-pairs), gel concentration and nucleotide sequence. Under appropriate adjustment between the size of the DNA and the gel concentration, this dependence was found to be V-shaped in a limited interval around constraint O, the minimum mobility at the apex of the V being obtained for relaxed DNA. Analysis of the DNA size dependence of the V suggests that it is the result of a modulated compaction of the DNA rings by the gel matrix. Compaction appears to be maximum upon relaxation, and to decrease with increase in supercoiling. Consistent with this interpretation, gels were found to oppose structural departures from the B helix, such as Z transition and cruciform extrusion, which tend to relax the DNA molecule and make it more expanded. In contrast, when DNA size or gel concentration are large enough relative to one another, U shapes are observed instead of Vs, as a consequence of an increase in the mobility of the rings closer to relaxation. The relevance of these results to the situation of superhelical DNA in vivo is discussed. Application of the V to the measurement of the DNA helical twist is mentioned.

Chromatin reconstitution on small DNA rings. I.

Chromatin was reconstituted using the four core histones on 359 base-pair nicked and closed rings by salt dialysis and/or at physiological ionic strength by means of polyglutamic acid. The products, which consisted of mono- and dinucleosomes, were characterized by gel electrophoresis, sedimentation in sucrose gradients and high-resolution electron microscopy. The results were as follows. (1) The efficiency of the reconstitution was found first to increase with the negative linking difference of the closed rings relative to their relaxed configuration to reach a maximum for -2 turns, and then to decrease for the largest difference of -3 turns. Discrepancies between topoisomers were also observed with regard to differential formation of mono- and dinucleosomes. Topoisomer -1 reconstituted monomers easily but reconstituted dimers with difficulty, whilst this discrimination was virtually absent in the case of topoisomers -2 and -3. Moreover, mononucleosomes on the nicked ring were, with respect to their electrophoretic mobility, similar to mononucleosomes formed on topoisomer -1 but not to those on the other topoisomers, whose mobilities were greater. These features were interpreted in terms of the linking number change associated with the formation of a nucleosome monomer and dimer, approximately -1 and -2 turns, respectively. (2) Two dinucleosome subtypes were found to form in a sequential manner. Their different electrophoretic mobilities and sedimentation coefficients suggested that the early subtype is lighter, probably because of an incomplete histone complement in the second nucleosome of that subtype as a result of an impaired co-operativity in octamer assembly due to the small ring size. (3) An electron microscopic examination of the chromatin reconstituted on topoisomer -2 revealed that both mono- and dinucleosomes adopt two different, salt-dependent, morphologies each: in type I, entering and exiting DNAs do not cross, whilst they do in type II. Type I configuration is favoured in lower salt, whereas type II is favoured in higher salt. Such behaviour explains why nucleosomes in dimers were found to be always diametrically opposed on the rings rather than sometimes apposed, as would have been expected from a random deposition of the histone cores.

Nucleosome dynamics. Protein and DNA contributions in the chiral transition of the tetrasome, the histone (H3-H4)2 tetramer-DNA particle.

Our laboratory has previously reported the chiral transition of DNA minicircle-reconstituted tetrasomes (the particles made of DNA wrapped around the histone (H3-H4)2tetramer). This transition was induced by DNA positive torsional constraint, generated either by initial supercoiling of the loop or by its thermal fluctuations during topoisomerase relaxation. Taking into account the wrapping of the DNA around the histones into less than a turn, and its negative crossing at the entry-exit, the transition was proposed to involve a 360 degrees rotation of the loop around the particle dyad axis, and the formation of a positive crossing. The tetramer horseshoe-shaped conformation within the octamer further suggested that this process could be mediated by a reorientation of the two sector-like H3-H4 dimers about their H3/H3 interface, which would switch the overall handedness of the proteinaceous superhelix from left to right-handed. We now provide additional evidence for such a contribution of the protein by showing, through gel electrophoresis, topoisomerase relaxation and electron microscopy, that a sterical hindrance at the H3/H3 interface, introduced by covalent linking of bulky adducts through thiol oxidation of H3 cysteine 110, interferes with the transition. Such interference varies, depending on the particular SH-reagent used; but the most remarkable effect was obtained with 5, 5'-dithiobis (2-nitrobenzoic acid) (DTNB), which displaces the preferred conformation of the tetrasomes from left-handed to semi-right-handed, and at the same time preserves a significant degree of chiral flexibility. DNA contribution was evidenced by a specific fractionation of circular tetrasomes in gel electrophoresis which, together with a different positioning of control and DTNB tetrasomes on linear DNA, pointed to an interdependence between tetrasome conformation and positions. Moreover, linear tetrasomes fluctuate between crossed and uncrossed conformations in a salt-dependent equilibrium which appears to vary with their positions on the DNA. These data suggest a modulatable role of the DNA around the dyad in the transition, depending primarily on its sequence-dependent deformability. This role is played at both levels of H3-H4 dimer reorientation and lateral opening, a mechanism by which the particle may relieve the clash between its entering and exiting DNAs. These properties make the tetrasome an attractive potential intermediate in nucleosome dynamics in vivo, in particular duringX transcriptional activation and elongation.

Chromatin reconstitution on small DNA rings. II. DNA supercoiling on the nucleosome.

DNA supercoiling on the nucleosome was investigated by relaxing with topoisomerase I mono- and dinucleosomes reconstituted on small DNA rings. Besides 359 base-pair (bp) rings whose linking differences were integers, two additional series of rings with fractional differences, 341 and 354 bp in size, were used. Mononucleosomes reconstituted on 359 bp rings were found to relax into a single mononucleosome form. In contrast, 341 and 354 bp mononucleosomes relaxed into a mixture of two forms, corresponding to two adjacent topoisomers. The observation that the ratio between these two forms was, within each ring series, virtually independent of the initial linking number of the topoisomer used for the reconstitution suggested that each partition reflected an equilibrium. Comparison with the equilibria observed for the same rings in the absence of histones showed that the formation of a single nucleosome is associated with a linking number change of -1.1(+/-0.1) turn. Dinucleosomes, in contrast, were not relaxed to completion and do not reach equilibria. The corresponding linking number change per nucleosome was, however, estimated to be similar to the above figure, in agreement with previous data from the literature obtained with circular chromatins containing larger numbers of nucleosomes. DNA structure in mononucleosomes was subsequently investigated by means of high-resolution electron microscopy and gel electrophoresis. It was found that the above linking number reduction could be ascribed to a particle with a large open extranucleosomal DNA loop and with no more than 1.5 turns of a superhelix around the histone core. A theoretical model of a nucleosome on a small ring was constructed in which one part of the DNA was wrapped around a cylinder and the other part was free to vary both in torsion and flexion. The linking number reduction predicted was found to be most consistent with experimental data when the twist of the DNA in the superhelix was between 10.5 and 10.65 pb per turn, suggesting that wrapping on the nucleosome does not alter the twist of the DNA significantly. A lower estimate of the linking number reduction associated with a two-turn nucleosome was also derived, based on an analysis of recent data obtained upon treatment of reconstituted minichromosomes with gyrase. The value, 1.6 turns, set a lower limit of 10.44 bp per turn for the twist of nucleosomal DNA, in agreement with the above estimate.(ABSTRACT TRUNCATED AT 400 WORDS)

Nucleosome dynamics. II. High flexibility of nucleosome entering and exiting DNAs to positive crossing. An ethidium bromide fluorescence study of mononucleosomes on DNA minicircles.

H2A-H2B exchange with the intranuclear histone pool upon chromatin transcription in vivo is generally viewed as being triggered by the DNA positive supercoiling wave pushed by the elongating polymerase. This notion was tested here by investigating a potential release of H2A-H2B by ethidium bromide-induced positive supercoiling in the loop of mononucleosomes reconstituted on DNA minicircles. The results of gel electrophoresis, fluorescence titration and electron microscopy showed that such a positive supercoiling was not able to release H2A-H2B, nor to unfold the nucleosome to any detectable extent. The reason appeared to be the ease with which the loop could undergo a positive crossing, a surprising observation in view of the DNA left-handed wrapping around the octamer. Moreover, the influence of histone acetylation suggested that such loop flexibility to positive crossing is mediated by histone N-terminal tails which, by interacting with entering and exiting DNAs, reduce their electrostatic repulsion. These conclusions are confirmed and extended in the accompanying article through relaxation with topoisomerase I.

Nucleosome dynamics. III. Histone tail-dependent fluctuation of nucleosomes between open and closed DNA conformations. Implications for chromatin dynamics and the linking number paradox. A relaxation study of mononucleosomes on DNA minicircles.

The mean linking number () of the topoisomer equilibrium distribution obtained upon relaxation of DNA minicircles with topoisomerase I did not increase linearly, but rather in a step wise fashion, with DNA size between 351 and 366 bp. As a consequence, the corresponding linking number difference () did not remain equal to 0, but rather oscillated between +/-0.3 with the periodicity of the double helix. This oscillation, not observed with plasmid-size DNA, is an expected consequence of the stiffness of short DNA. When minicircles were reconstituted with a nucleosome, the associated oscillated between approximately -1.4 +/-0. 2. This oscillation appears to result from the combined effects of DNA stiffness, and nucleosome ability to thermally fluctuate between three distinct DNA conformational states. Two of these states, a closed approximately 1.75-turn DNA conformation with negatively crossed entering and exiting DNAs, and an open approximately 1.4-turn conformation with uncrossed DNAs, are well known, whereas the third state, with a closed DNA conformation and DNAs tending to cross positively rather than negatively, is less familiar. Access to both closed "negative" and "positive" states appears to be mediated by histone N-terminal tails, as shown by specific alterations to the oscillation caused by histone acetylation and phosphate ions, a potent tail destabilizator. These results extend previous observations of ethidium bromide fluorescence titration in the accompanying article, which have pointed to an histone tail-dependent flexibility of entering and exiting DNAs to positive crossing. They also show that DNA wrapping around the histones occurred without twist alteration compared to the DNA free in solution, and reveal an intriguing new facet of the "linking-number-paradox" problem: the possibility for linkers in chromatin to adopt different crossing status within an overall dynamic equilibrium which may be regulated by histone acetylation.

Linker histone-dependent DNA structure in linear mononucleosomes.

We have examined the binding of the linker histone H5 (LH) to mononucleosomes. Mononucleosomes reconstituted on short DNA fragments display a series of discrete bands on a gel corresponding to various nucleosome positions along the DNA. When a series of engineered H5s with differing extents of the C-terminal tail are bound to these mononucleosomes, the electrophoretic mobilities of the resulting complexes are altered. Not only is there a general increase in mobility upon complex formation, but there is a reduction in the differences in mobility of the most distal nucleosomes. The complexes were also visualized by electronmicroscopy. From these two complementary studies, we conclude the following. (1) Entering and exiting DNAs are uncrossed in the LH-free particles, despite a DNA wrapping of 1.65 to 1.7 turns around the histone core. This results from a bending of the entering and exiting DNA away from each other and the histone surface, presumably as a consequence of electrostatic repulsion. This confirms and extends conclusions derived from our recent examination of the same particles in 3D through cryo-electron microscopy. (2) Binding of the globular domain of H5 increases DNA wrapping to 1.8 to 1.9 turns, but fails to induce a crossing due to an accentuation of the bends. (3) The C-terminal tail of H5 bridges entering and exiting DNAs together into a four-stranded stem over a distance of about 30 bp. The occurrence of such a stem may introduce constraints on models of the 30 nm chromatin fiber.

Nucleosome dynamics. VI. Histone tail regulation of tetrasome chiral transition. A relaxation study of tetrasomes on DNA minicircles.

We have recently described the relaxation of mononucleosomes on an homologous series of 351-366 bp DNA minicircles, as a tool to study nucleosome structure and dynamics in vitro. Nucleosomes were found to have a tail-regulated access to three distinct DNA conformations, depending on the crossing between the entering and exiting DNAs, and its polarity. This approach was now used to explore tetrasome chiral transition, and the influence of the histone tails. The data confirmed the existence of two states, with linking number differences DeltaLk(t)=-0.74(+/-0.01) and +0.51(+/-0.06). As expected, the particle free energy is higher in the right-handed state (DeltaG(t)=1.9(+/-0.I) kT), but it decreased (to 1.3(+/-0.1) kT) upon histone acetylation and the addition of phosphate, a potent tail destabilizer. Removal of the tails with trypsin further decreased DeltaG(t) (to 0.6 kT), and also induced a loss of supercoiling in both states, to DeltaLk(t)=-0.64(+/-0.03) and +0. 35(+/-0.05). The loop end-conditions, and hence the parameters of the DNA superhelix, were then calculated for both states using the explicit solutions to the equations of the mechanical equilibrium in the theory of elastic rod model for DNA. Whereas the pitch of the DNA superhelix may be approximately equal and opposite in the two conformations, its radius (r) was 20% larger in the right-handed conformation, confirming previous observations by electron microscopy of a tetrasome lateral opening in that conformation. The above supercoiling losses were found to reflect a further 3 % increase in r (to 23 %) upon removal of the tails in the right-handed conformation, and a 14 % increase in the left-handed conformation. The use of composite tetramers with one histone tail intact and the other removed showed these effects to be essentially due to the H3 tails. Altogether, these results show that the H3 tails oppose the tetrasome opening which is expected to be required to relieve the clash between the entering and exiting DNAs in the course of the transition, but which also appears to be intrinsic to the protein reorientation mechanism. We propose that the block against opening results from the H3 tails intercalating into the small groove of the double helix at +/-10 bp from the dyad, and acting as wedges against local DNA straightening. The tails (especially H3) may therefore regulate tetrasome chiral transition in vivo.

Chromatin reconstitution on small DNA rings. III. Histone H5 dependence of DNA supercoiling in the nucleosome.

Mononucleosomes were reconstituted on small DNA rings in the presence of histone H5 and relaxed to an equilibrium using calf thymus topoisomerase I. DNA products, when compared to the equilibria observed with the same minicircles in the absence of histones, showed that a linking number reduction of 1.6 to 1.7 was associated with this reconstitution, in contrast with the 1.1 to 1.2 figure reported in our recent study of the H5-free nucleosome. Gel electrophoretic properties and electron microscopic visualization of the nucleosomes suggest a correlation between this increase and a further wrapping of the DNA around the histone core from less than 1.5 turns of the superhelix in the absence of H5, to close to two turns in its presence. Implications for DNA topology in chromatin are discussed.

Chromatin reconstitution on small DNA rings. IV. DNA supercoiling and nucleosome sequence preference.

Nucleosome formation on inverted repeats or on some alternations of purines and pyrimidines can be inhibited in vitro by DNA supercoiling through their supercoiling-induced structural transitions to cruciforms or Z-form DNA, respectively. We report here, as a result of study of single nucleosome reconstitutions on a DNA minicircle, that a physiological level of DNA supercoiling can also enhance nucleosome sequence preference. The 357 base-pair minicircle was composed of a promoter of phage SP6 RNA polymerase joined to a 256 base-pair fragment containing a sea urchin 5 S RNA gene. Nucleosome formation on the promoter was found to be enhanced on a topoisomer with in vivo superhelix density when compared to topoisomers of lower or higher superhelical densities, to the nicked circle, or to the linear DNA. In contrast, nucleosomes at other positions appeared to be insensitive to supercoiling. This observation relied on a novel procedure for the investigation of nucleosome positioning. The reconstituted circular chromatin was first linearized using a restriction endonuclease, and the linear chromatin so obtained was electrophoresed as nucleoprotein in a polyacrylamide gel. The gel showed well-fractionated bands whose mobilities were a V-like function of nucleosome positions, with the nucleosome near the middle migrating less. This behavior is similar to that previously observed for complexes of sequence-specific DNA-bending proteins with circularly permuted DNA fragments, and presumably reflects the change in the direction of the DNA axis between the entrance and the exit of the particle. Possible mechanisms for such supercoiling-induced modulation of nucleosome formation are discussed in the light of the supercoiling-dependent susceptibility to cleavage of the naked minicircle with S1 and Bal31 nucleases; and a comparison between DNase I cleavage patterns of the modulated nucleosome and of another, non-modulated, overlapping nucleosome.