DNA folding in the nucleosome

Digestion of chromatin with a number of nucleases shows that the DNA is regularly folded in the nucleosome. Particularly cleavage by pancreatic DNase (DNase I) in the 140 base-pair nucleosome has been examined. This nuclease nicks the DNA every ten bases on each strand as demonstrated by labeling the 5′-ends of the 140 base-pair nucleosome. Cleavage sites on opposite strands are staggered by two bases. This proves that the DNA is arranged on the outside of the histone core in a regular way. The probability distribution of nicking might indicate a 2-fold symmetry of the 140 base-pair nucleosome. In particular it is shown that the predominant band of 80 bases is derived from several regions within the 140 base-pairs and suggested to reflect the pitch of the DNA superhelix surrounding the histone core of the nucleosome. Its possible significance with respect to chromatin structure is discussed.     

Location of the primary sites of micrococcal nuclease cleavage on the nucleosome core

The positions and relative frequencies of the primary cleavages made by micrococcal nuclease on the DNA of nucleosome core particles have been found by fractionating the double-stranded products of digestion and examining their single-stranded compositions. This approach overcomes the problems caused by secondary events such as the exonucleolytic and pseudo-double-stranded actions of the nuclease and, combined with the use of high resolution gel electrophoresis, enables the cutting site positions to be determined with a higher precision than has been achieved hitherto. The micrococcal nuclease primary cleavage sites lie close (on average, within 0.5 nucleotide) to those previously determined by Lutter (1981) for the nucleases DNase I and DNase II. These similarities show that the accessible regions are the same for all three nucleases, the cleavage sites being dictated by the structure of the nucleosome core. The differences in the final products of the digestion are explained in terms of secondary cleavage events of micrococcal nuclease. While the strongly protected regions of the nucleosome core DNA are common to all three nucleases, there are differences in the relative degrees of cutting at the more exposed sites characteristic of the particular enzyme. In particular, micrococcal nuclease shows a marked polarity in the 3'-5' direction in the cutting rates as plotted along a single strand of the nucleosomal DNA. This is explained in terms of the three-dimensional structure of the nucleosome where, in any accessible region of the double helix, the innermost strand is shielded by the outermost strand on the one side and the histone core on the other. The final part of the paper is concerned with the preference of micrococcal nuclease to cleave at (A,T) sequences in chromatin.     

Primary organization of the nucleosome core particles sequential arrangement of histones along DNA

A high-resolution map for the arrangement of histones along DNA in the nucleosome core particles has been determined by a new sequencing procedure. The lysine groups of histones were crosslinked to partly depurinated DNA at neutral pH. One strand of DNA was split at the points of crosslinking, thus leaving the 5′-terminal DNA fragments bound to histones. The lengths of these crosslinked DNA fragments were measured to determine the position of histones on one strand of the core DNA from its 5′ end.The results demonstrate that histones are bound to regularly arranged, discrete DNA segments about six nucleotides long. These segments are separated by histone-free gaps about four nucleotides wide located at a distance of about 10n nucleotides from the 5′ end of DNA. The first 20 nucleotides from the 5′ ends of DNA seem to be free of histones. Histones appear to be arranged symmetrically and in a similar way on both DNA strands. Any one histone, being bound predominantly to discrete segments on one or other of the strands, can oscillate at the same time between the two strands across the major DNA groove. Two symmetrical models for the arrangement of two molecules of each core histone on linearized and folded DNA are proposed.     

Deoxyribonuclease I produces staggered cuts in the DNA of chromatin

The relationship of cuts made by deoxyribonuclease I (DNase I, EC. 3.1.4.5) on the two strands of DNA of chromatin has been investigated. DNA was extracted from a DNase I digest of rat liver nuclei and incubated with the large fragment of DNA polyrnerase I. Analysis of the products of this incubation indicates the cuts made by DNase I on opposite strands are staggered with respect to one another. A cut on one strand is about two bases in the 3′ direction or eight bases in the 5′ direction from the position on its own strand which is directly across from the cut on the other strand. A different result is obtained when a DNase I digest of native DNA is analyzed. Current models for the organization of DNA in the nucleosome are discussed with respect to these results.     

Precise location of DNase I cutting sites in the nucleosome core determined by high resolution gel electrophoresis.

The precise locations of the DNase I cutting sites in the nucleosome core have been determined by analysis of the DNA products of a DNase I digestion of 32P end-labelled mucleosome cores on a high resolution gel electrophoresis system. This system is capable of resolving fragments of mixed sequence DNA differing by one base into the region of 160 bases in length. The DNase I cutting sites in the core are found to be spaced at multiples of about 10.4 (i.e. clearly different from 10.0) bases along the DNA, but show significant variations about this value. In addition to the location of the sites, the stagger between individual sites on opposite strands has been determined and is found to be inconsistent with at least one proposed mechanism for nuclease cleavage of chromatin DNA. Finally, a calculated distribution of fragment lengths in a DNase I digest of nuclei has been determined from the data obtained from the nucleosome core and found to be in reasonable agreement with the observed distribution. The periodicity of 10.4 is discussed with respect to the number of base pairs per turn of chromatin DNA and the number of superhelical turns of DNA per nucleosome.     

DNase II digestion of the nucleosome core: precise locations and relative exposures of sites.

The precise locations and relative exposures of the DNase II-accessible sites in the nucleosome core DNA are determined using techniques previously employed for the enzyme DNase I. It is found that there are a number of similarities between the site exposure patterns for the two enzymes but that in general the DNase II seems to discriminate less among adjacent sites' accessibilities than does DNase I. The two enzymes attack essentially the same positions in the DNA, the average difference between the precise location of the site being less than one-half base for the two enzymes. Such close similarities in the digestion patterns of two enzymes with such different mechanisms of scission show that the patterns reflect the structure of the nucleosome core and not merely the properties of the particular enzyme used.     

Benzo(a)pyrene 7,8-dihydrodiol-9,10-oxide modification of DNA: Relation to chromatin structure and reconstitution

Purified duck reticulocyte DNA was incubated in vitro with a 7,8-dihydrodiol-9,10-oxide derivative of benzo(a)pyrene (BPDE). The carcinogen-modified DNA was somewhat more susceptible to partial digestion by the single strand specific endonuclease S₁ than unmodified DNA, suggesting slight denaturation of the helix at sites of modification. Chromatin was reconstituted in vitro utilizing this carcinogen-modified DNA and unmodified-chromatin associated proteins. This reconstituted chromatin showed the same kinetics and extent of digestion by Staphylococcal nuclease, and similar nucleosome profiles on sucrose density gradient centrifugation, as those obtained with native chromatin or chromatin reconstituted with unmodified DNA. Moreover, polyacrylamide gel electrophoresis of DNA fragments obtained from nuclease digests gel electrophoresis of DNA fragments obtained from nuclease digests of the reconstituted chromatins suggested that the chromatin containing carcinogen-modified DNA had the same subnucleosome structure as that reconstituted with unmodified DNA. In a separate set of studies intact duck reticulocyte chromatin was reacted directly with BPDE. Nuclease digestion studies indicated that 65% of the carcinogen was bound to the ‘open’ regions of chromatin, and 35% to ‘closed’ regions.These results indicate that although covalent binding of a benzo(a)pyrene (BP) derivative to DNA produces local distortions in conformation of the helix, this modification does not appear to interfere with the ability of the DNA to associate with histones to form nucleosome structures. In addition, although DNA in the open regions of chromatin is more susceptible to reaction with the BP derivative, there is appreciable reaction with the DNA associated with histones.     

Rapid Deamination of Cyclobutane Pyrimidine Dimer Photoproducts at TCG Sites in a Translationally and Rotationally Positioned Nucleosome in Vivo

Sunlight-induced C to T mutation hot spots in skin cancers occur primarily at methylated CpG sites that coincide with sites of UV-induced cyclobutane pyrimidine dimer (CPD) formation. The C and 5-methyl-C in CPDs are not stable and deaminate to U and T, respectively, which leads to the insertion of A by the DNA damage bypass polymerase η, thereby defining a probable mechanism for the origin of UV-induced C to T mutations. Deamination rates for T^mCG CPDs have been found to vary 12-fold with rotational position in a nucleosome in vitro. To determine the influence of nucleosome structure on deamination rates in vivo, we determined the deamination rates of CPDs at TCG sites in a stably positioned nucleosome within the FOS promoter in HeLa cells. A procedure for in vivo hydroxyl radical footprinting with Fe-EDTA was developed, and, together with results from a cytosine methylation protection assay, we determined the translational and rotational positions of the TCG sites. Consistent with the in vitro observations, deamination was slower for one CPD located at an intermediate rotational position compared with two other sites located at outside positions, and all were much faster than for CPDs at non-TCG sites. Photoproduct formation was also highly suppressed at one site, possibly due to its interaction with a histone tail. Thus, it was shown that CPDs of TCG sites deaminate the fastest in vivo and that nucleosomes can modulate both their formation and deamination, which could contribute to the UV mutation hot spots and cold spots.     

Adenovirus DNA replication in vitro: Origin and direction of daughter strand synthesis

We have characterized a soluble enzyme system from adenovirus-infected cells that is capable of replicating exogenously added adenovirus DNA in vitro. Maximal DNA synthesis is observed when DNA-protein complex, isolated from purified adenovirus virions, is added as template. Under these conditions DNA replication starts at or near either end of the template. Daughter strand synthesis then proceeds in the 5′ to 3′ direction displacing the parental strand of the same polarity. Thus, the r daughter strand is synthesized from right to left on the conventional map of the adenovirus genome, and the l daughter strand is synthesized from left to right. This course of events is the same as that which occurs during adenovirus DNA replication in vivo. In contrast, when deproteinized adenovirus DNA is added to the in vitro system, the limited DNA synthesis that is observed appears to be due to a repair-like reaction. In particular, synthesis can begin at many sites within the template, and the synthetic product consists largely of short DNA chains that are covalently linked to template DNA strands.     

Effect of 5-Methylcytosine on the Structure and Stability of DNA. Formation of Triple-Stranded Concatenamers by Overlapping Oligonucleotides

Abstract

A triple helix can be formed upon binding of a pyrimidine oligonucleotide to the major groove of a homopurine-homopyrimidine (R·Y) double-stranded DNA target site. Here, we report that this reaction can be influenced by base methylation. The pyrimidine strand 5′- TmCTmCTmCTmCTTmCT (mY₁₂), whose cytosine residues are methylated at C5, does not bind the duplex 5′-AGAGAGAGAAGA·3′-TCTCTCTCTTCT (R₁₂·Y₁₂) to yield a 12-triad triplex, as would be expected from these DNA sequences. Rather, a complex of overlapping oligonucleotides, which we define concatenamer, is formed. The concatenamer is clearly evidenced by Polyacrylamide gel electrophoresis (PAGE) since it migrates with a smeared band of very low mobility. The stoichiometry of the concatenamer, determined by both UV mixing curves and electrophoresis, is surprisingly found to be (R₁₂· 2mY₁₂)_n, thus showing that the unmethylated Y₁₂ strand is excluded from the complex. Denaturation experiments performed by ultraviolet absorbance (UV) and differential scanning calorimetry (DSC) show that the concatenamers melt with a single and highly cooperative transition whose Tm strongly depends on pH. Overall, the data point to the conclusion that the concatenamers are in triple helix, where the methylated mY₁₂ strand is engaged in both Watson-Crick and Hoogsteen base pairings, thus displacing the Y₁₂ strand from the R₁₂·Y₁₂ duplex. A possible mechanism of concatenamer formation is proposed. The results presented in this paper show that 5-methylcytosine brings about a strong stabilizing effect on both double and triple DNA helices, and that pyrimidine oligonucleotides containing 5-methylcytosine can displace from R·Y duplexes the analogous non-methylated strand. The advantage of using methylated oligonucleotides in antisense technology is discussed.     

Non-productive DNA damage binding by DNA glycosylase-like protein Mag2 from Schizosaccharomyces pombe

Schizosaccharomyces pombe contains two paralogous proteins, Mag1 and Mag2, related to the helix-hairpin-helix (HhH) superfamily of alkylpurine DNA glycosylases from yeast and bacteria. Phylogenetic analysis of related proteins from four Schizosaccharomyces and other fungal species shows that the Mag1/Mag2 duplication is unique to the genus Schizosaccharomyces and most likely occurred in its ancestor. Mag1 excises N3- and N7-alkylguanines and 1,N⁶-ethenoadenine from DNA, whereas Mag2 has been reported to have no detectible alkylpurine base excision activity despite high sequence and active site similarity to Mag1. To understand this discrepancy we determined the crystal structure of Mag2 bound to abasic DNA and compared it to our previously determined Mag1–DNA structure. In contrast to Mag1, Mag2 does not flip the abasic moiety into the active site or stabilize the DNA strand 5′ to the lesion, suggesting that it is incapable of forming a catalytically competent protein–DNA complex. Subtle differences in Mag1 and Mag2 interactions with the DNA duplex illustrate how Mag2 can stall at damage sites without fully engaging the lesion. We tested our structural predictions by mutational analysis of base excision and found a single amino acid responsible at least in part for Mag2's lack of activity. Substitution of Mag2 Asp56, which caps the helix at the base of the DNA intercalation loop, with the corresponding serine residue in Mag1 endows Mag2 with ?A excision activity comparable to Mag1. This work provides novel insight into the chemical and physical determinants by which the HhH glycosylases engage DNA in a catalytically productive manner.     

Crystal Structures of Nucleosome Core Particles Containing the ‘601’ Strong Positioning Sequence

Nucleosome positioning plays a key role in genomic regulation by defining histone-DNA context and by modulating access to specific sites. Moreover, the histone-DNA register influences the double-helix structure, which in turn can affect the association of small molecules and protein factors. Analysis of genomic and synthetic DNA has revealed sequence motifs that direct nucleosome positioning in vitro; thus, establishing the basis for the DNA sequence dependence of positioning would shed light on the mechanics of the double helix and its contribution to chromatin structure in vivo. However, acquisition of well-diffracting nucleosome core particle (NCP) crystals is extremely dependent on the DNA fragment used for assembly, and all previous NCP crystal structures have been based on human α-satellite sequences. Here, we describe the crystal structures of Xenopus NCPs containing one of the strongest known histone octamer binding and positioning sequences, the so-called ‘601’ DNA.Two distinct 145-bp 601 crystal forms display the same histone-DNA register, which coincides with the occurrence of DNA stretching-overtwisting in both halves of the particle around five double-helical turns from the nucleosome center, giving the DNA an ‘effective length’ of 147 bp. As we have found previously with stretching around two turns from the nucleosome center for a centromere-based sequence, the terminal stretching observed in the 601 constructs is associated with extreme kinking into the minor groove at purine-purine (pyrimidine-pyrimidine) dinucleotide steps. In other contexts, these step types display an overall nonflexible behavior, which raises the possibility that DNA stretching in the nucleosome or extreme distortions in general have unique sequence dependency characteristics. Our findings indicate that DNA stretching is an intrinsically predisposed site-specific property of the nucleosome and suggest how NCP crystal structures with diverse DNA sequences can be obtained.     

Differential DNase I Sensitivity of the Two Complementary Nucleosomal DNA Strands in Cycloheximide-Treated Ehrlich Ascites Tumor Cells

Abstract

The accessibility of the two complementary DNA strands in newly replicated chromatin of Ehrlich ascites tumor (EAT) cells grown under conditions of cycloheximide-inhibrted protein synthesis was studied by analysis of the DNase I digestion of isolated nuclei. Bulk DNA was labeled with ¹⁴C-thymidine and the newly synthesized strands - with bromodeoxyu ridine and ³H-thymidine. The DNase I digests were fractionated in two successive CsCl density gradient centrifugations to obtain a dense fraction containing 15–20% newly replica ted DNA Analysis of the distribution of ¹⁴C-labeled parental DNA fragments complementary to the ³H-nascent strand has shown that the ¹⁴C-labeled fragments prevail in the region of 30–50 nucleotides. Simulation experiments using the rate constants for DNase I attack show that this result may be explained by an enhanced accessibility at the nucleosomal 5′-end region of the parental strands, where the H2a-H2b dimer interacts with DNA. This asymmetry seems tobe induced by interactions in the chromatin.     

The (gt)n(ga)m containing intron 2 of HLA-DRB alleles binds a zinc-dependent protein and forms non B-DNA structures

We studied protein binding and structural features of perfect and imperfect composite (gt)_n(ga)_m blocks from different HLA-DRB1 alleles in their original genomic and artificial environments. The major retarded protein/DNA complex of the genomic (gt)_n(ga)_m fragments comprises a zinc-dependent protein present in nuclear extracts from different cell types. The protein binding is characterized by moderate affinities independent of the polymorphic form of the physiological microsatellite allele. The binding affinity depends on the 5′ and 3′ adjacent single copy parts. DNase I footprinting of genome-derived fragments revealed that the 5′ adjacent sequence and the (gt)_n repeat are preferentially protected on the (gt)_n(ga)_m strand. Comparing three alleles, a regular pattern of footprints was not detectable in the (gt)_n part, indicating that the zinc-dependent protein recognizes structural rather than sequence-specific features in this region. Chemical probing resulted in a pattern characteristic for Z-DNA in the (gt)_n tract of the fragments. However, EMSA experiments using the Z-DNA specific monoclonal antibody mABZ-22 did not prove the presence of Z-DNA. As demonstrated by chemical modifications of the different (ga)_m targets, only one of three (gt)_n(ga)_m fragments formed intramolecular triplexes of the type H-y3 and H-y5. DNase I footprinting revealed only weak protection, if any, in the homopurine tract. Rather, the (tc)_m strands are hypersensitive for DNase I. This is probably due to structural conversions into intramolecular *H-triplexes after binding of HIZP.     

Mechanical properties of symmetric and asymmetric DNA A-tracts: implications for looping and nucleosome positioning

A-tracts are functionally important DNA sequences which induce helix bending and have peculiar structural properties. While A-tract structure has been qualitatively well characterized, their mechanical properties remain controversial. A-tracts appear structurally rigid and resist nucleosome formation, but seem flexible in DNA looping. In this work, we investigate mechanical properties of symmetric A_nT_n and asymmetric A_2n tracts for n = 3, 4, 5 using two types of coarse-grained models. The first model represents DNA as an ensemble of interacting rigid bases with non-local quadratic deformation energy, the second one treats DNA as an anisotropically bendable and twistable elastic rod. Parameters for both models are inferred from microsecond long, atomic-resolution molecular dynamics simulations. We find that asymmetric A-tracts are more rigid than the control G/C-rich sequence in localized distortions relevant for nucleosome formation, but are more flexible in global bending and twisting relevant for looping. The symmetric tracts, in contrast, are more rigid than asymmetric tracts and the control, both locally and globally. Our results can reconcile the contradictory stiffness data on A-tracts and suggest symmetric A-tracts to be more efficient in nucleosome exclusion than the asymmetric ones. This would open a new possibility of gene expression manipulation using A-tracts.     

Interaction of single-stranded DNA with the second DNA-binding site of the RecA nucleoprotein filament

RecA first forms a filament on single-stranded DNA (ssDNA), thereby forming the first site for ssDNA binding and, simultaneously, the second site for binding double-stranded DNA (dsDNA). Then, the nucleoprotein filament interacts with dsDNA, although it can bind ssDNA as well. The resulting complex searches for homology sites and performs strand exchange between homologous DNA molecules. The interaction of various ssDNAs with the second DNA-recognizing site of RecA was studied by gradually increasing the structural complexity of the DNA ligand. Recognizing ssDNA with the second site, the protein interacts with each nucleotide of the ligand, forming contacts with both internucleotide phosphate groups and nitrogen bases. Pyrimidine oligonucleotides d(pC) _n and d(pT) _n interacted with the second site of the RecA filament more efficiently than d(pA) _n did. This was due to a more efficient interaction of the RecA filament with the 5′-terminal nucleotide of pyrimidinic DNA and to the difference in specific conformational changes of the nucleoprotein filament in the presence of purinic and pyrimidinic DNAs. A comparison of thermodynamic characteristics of DNA recognition at the first and second DNA-binding sites of the filament showed that, at n > 10, d(pC) _n and d(pN) _n were bound at the second site less tightly than at the first site. At n > 20, the second site bound d(pA) _n more efficiently than the first site. The difference in d(pN) _n affinity for the first and second sites increased monotonically with increasing n. Possible mechanisms of a RecA-dependent search for homology and DNA strand exchange are discussed.     

The influence of linker chain length on the sequence specificity of DNA damage by n-bromoalkylphenanthridinium bromides in plasmid DNA and in intact human cells

The sequence specificity of DNA damage of n-bromoalkylphenanthridinium bromides, with linker chain lengths (n) of 4,6,8 and 10 methylene groups, was investigated in the plasmid pUC8 and in intact human cells. A linear amplification assay was used to elucidate the DNA sequence specificity of the alkylating agents. In this assay Taq DNA polymerase extends from an oligonucleotide primer up to the damage site and the products run on a DNA sequencing gel to reveal the precise sites of DNA damage. For both the plasmid and cellular experiments, the compound that caused the most damage to DNA was the n = 6 compound, followed by (in decreasing order) the n = 4, n = 8, and n = 10 compounds. There were significant differences in the sequence specificity of DNA damage between n-bromoalkylphenanthridinium bromides of different linker chain length: (1) the main sites of damage were at guanines for the n = 4,6 and 8 compounds but at guanines and adenines for the n = 10 compound; (2) a consensus sequence of 5′-c(a/t)Ggg-3′ was obtained for the n = 4,6 and 8 compounds but 5′-c(a/c)(G/A)(g/a)-3′ for the n = 10 compound; (3) runs of consecutive Gs were the major site of damage for the n = 4,6 and 8 compounds, but consecutive Gs or consecutive As for the n = 10 compound; (4) for damage at single isolated guanines, the most damaged sequences were at 5′-Ga-3′ for the n = 4 compound but at 5′-Gt-3′ for the n = 6,8 and 10 compounds. The tandemly repeated alpha RI DNA sequence was the DNA target in intact human K562 cells. In intact human cells, the compounds produced damage with similar DNA sequence selectivity to that found in plasmid DNA. The n = 4 and 6 compounds possess marginal anti-tumour activity and these compounds produced the most damage in intact human cells. The n = 8 and 10 compounds do not demonstrate significant anti-tumour activity and these compounds resulted in the least damage in cells.