A dominant role for DNA secondary structure in forming hypersensitive structures in chromatin

Chromatin structure of genes during the cell cycle

K562 cells were separated by centrifugal elutriation into 5 fractions each of which had a unique distribution of cells at different phases of the cell cycle. A study of DNAseI sensitivity of the epsilon- and gamma-globin genes, which are expressed in these cells, and the lambda light chain immunoglobulin gene, which is not expressed, in nuclei from each of the 5 fractions was undertaken. There was no detectable difference in the arrangement of DNAseI-hypersensitive sites (DHS) in chromatin of each of the three genes in the different fractions.

Unwinding of supercoiled DNA by cis- and trans-diamminedichloroplatinum(II): influence of the torsional strain on DNA unwinding

The effective unwinding angle, phi, for cis-diamminedichloroplatinum(II) (cis-DDP) and trans-DDP was determined by utilizing high resolution gel electrophoresis and supercoiled phi X174 RF DNA as a substrate. The effective unwinding angle was calculated by equating the reduction in mobility of the DDP-modified DNA to the removal of a number of superhelical turns. The value of the effective unwinding angle for both DDP isomers was greatest at the low levels of DDP bound and decreased with increasing amounts of unwinding agent. The cis-isomer is a better unwinding agent than is the trans-isomer, being nearly twice as effective in unwinding the supercoiled DNA at the DDP levels investigated. A comparison of the magnitude of phi below rb values of 0.005 and those at high levels of binding reveals that the extent of torsional strain in the supercoiled DNA influences the magnitude of the unwinding of the DNA by these complexes. When this method is used in the analysis of the unwinding angle for a covalently bound species on supercoiled DNA, it may provide a more reliable estimate of the magnitude of phi at high degrees of supercoiling and at low levels of modification.

Enhanced gene expression by the poly(dT-dG).poly(dC-dA) sequence

The sequence poly(dT-dG).poly(dC-dA) (TG-element) is a ubiquitous component of eucaryotic genomes and has the potential to adopt a left-handed DNA conformation (Z-DNA). In this report, we have tested the hypothesis that the TG-element can modulate gene expression. Human genomic DNA fragments (1 to 1.5 kilobases) containing a (dT-dG)n.(dC-dA)n tract (30, 40, or 50 base pairs) or chemically synthesized (dT-dG)n.(dC-dA)n fragments (50 to 130 base pairs) were inserted in the pSV2-cat (simian virus 40 enhancer plus) or pA10-cat (enhancer minus) expression vector plasmid. These constructs were transfected into CV-1 cells or HeLa cells, and their transcription was monitored by assaying chloramphenicol acetyltransferase activity. The results showed that pSV2-cat with the TG-element and pA10-cat with the TG-element synthesized more chloramphenicol acetyltransferase activity (2 to 10 times, depending on the location of the TG-element) than did parental pSV2-cat and pA10-cat DNAs, respectively. Furthermore, the TG-element appeared to have characteristics similar to those of viral enhancers: (i) the TG-element enhanced transcription from a distance, (ii) its closer location to the promoter was more effective, and (iii) its orientation was not crucial. However, its enhancer-like activity was much weaker than that of the simian virus 40 enhancer, and, unlike many viral enhancers, it was equally active in monkey and in human cells. These results suggest that the TG-element may influence the expression of cellular genes.

Characterization of genomic poly(dT-dG).poly(dC-dA) sequences: structure, organization, and conformation

Hybridization studies suggest the abundant presence of poly(dT-dG).poly(dC-dA) (TG-element), a potential Z-DNA sequence, in eucaryotic genomes. We have isolated and characterized TG-elements from different locations in the human genome: from randomly isolated clones, associated with the actin gene family, and linked to another repeated element. The results indicate that the following features are typical of these TG-elements: the elements consist of 20 to 60 base pairs of (dT-dG)n.(dC-dA)n, the sequences characterized in our study were not flanked by direct or inverted repeats, the sequences are interspersed rather than in satellite blocks, the elements are not usually associated with other repeated elements, and some of the elements are found near coding sequences or in introns. Studies on the conformation of a genomic TG-element in a supercoiled plasmid indicate several distinct properties of the TG-element: it is in the Z-form only at low ionic strength, S1 nuclease recognizes its Z-form with a marked preference for one of the B-Z junctions, and the sensitive region extends for 20 base pairs near the B-Z junction. In contrast to the result with the supercoiled plasmid, S1 nuclease failed to recognize the TG-element in minichromosomes.

Quantitation of eukaryotic topoisomerase I reactivity with DNA. Preferential cleavage of supercoiled DNA

A method has been used to quantitate the reaction between eukaryotic type I DNA topoisomerase and topological forms of DNA. This procedure (Trask, D.K., DiDonato, J.D. and Muller, M.T. (1984) Eur. Mol. Biol. Organ. J. 3, 671-676) measures the efficiency of DNA cleavage and concurrent formation of a covalent enzyme/DNA complex. Eukaryotic type I topoisomerases react preferentially by 5-10-fold with supercoiled DNA. The effect of supercoiling is clearly evident in that both the initial rate and final extent of the reaction is elevated. Because the dissociation rate is much lower than the association rate, it is possible to isolate native topoisomerase/DNA complexes. These complexes are comprised of enzyme molecules which are catalytically active when challenged with a second supercoiled DNA substrate. Collectively, the data support the conclusion that a functional intermediate in the reaction sequence is being detected and that the avian topoisomerase I preferentially cleaves supercoiled DNA.

Effects of exposing Drosophila melanogaster parents to ethanol on expression of vestigial in their progeny

When parental Drosophila melanogaster were chronically exposed at 28 degrees C or 24 degrees C to ethanol during their larval and pupal stages of development, their progeny, produced when parents were 5-16-day-old adults, showed modified expression of vestigial alleles in heterozygous and homozygous combinations. Parental alcohol effects were dependent on parental rearing temperature. We conclude that parental environment (alcohol, temperature) causes heritable but transitory changes in progeny phenotype that are elicited by exposure of germ cells to alcohol.

Analysis of inactive X chromosome structure by in situ nick translation

Nick translation assays of fixed interphase female fibroblasts with tritiated nucleotides demonstrated a characteristic absence of label over sex chromatin. The chromatin bodies were nearly always peripheral in location and a ribbon of nick translatable DNA was detected between the sex chromatin and the nuclear envelope. High voltage electron microscopy indicated the possibility of a special nuclear envelope attachment region. The apparent resistance of sex chromatin to nick translation did not appear to be due to resistance to DNase I attack.

The chromatin structure of the human epsilon globin gene: nuclease hypersensitive sites correlate with multiple initiation sites of transcription

We have mapped sites in chromatin flanking the epsilon-globin gene in the K562 cell which are hypersensitive to digestion with DNAseI, micrococcal nuclease and S1 nuclease. Many of those in the 5' flanking region correspond to minor upstream transcriptional starts. However, one prominent site occurs upstream of the boundary of transcription; it maps to a region with an unusual DNA sequence. In baby hamster kidney cells stably transformed with recombinant DNA containing the human epsilon-globin gene and in Cos 7 cells transiently transfected with DNA containing the epsilon-globin gene, hypersensitive sites can be demonstrated.

Human alpha- and beta-globin gene transcription in mouse erythroleukaemia cells

Human beta-globin genes introduced into mouse erythroleukaemia (MEL) cells by DNA co-transformation are correctly regulated when erythroid cell differentiation is induced by dimethylsulphoxide (DMSO). In contrast, cloned human alpha-globin genes are efficiently transcribed in MEL cells before induction, and no increase in the level alpha-globin mRNA is observed when the cells differentiate. These observations suggest that the mechanisms by which alpha- and beta-globin genes are activated during erythroid cell differentiation are fundamentally different. Analysis of the transcription of hybrid human alpha-beta-globin genes in MEL cells revealed that the sequences responsible for differences in transcription of the intact alpha- and beta-globin genes are located on the 3' side of the mRNA capping site of the two genes, suggesting that cis-acting regulatory sequences are located within the structural genes.

Molecular analysis of erythropoiesis. A current appraisal

This article considers recent evidence concerning the molecular mechanisms involved in the coordinate regulation of gene expression during red blood cell (RBC) differentiation. Contrary to popular belief, recent evidence shows that only a few of the characteristic RBC proteins are restricted to the erythroid lineage: apart from the globins, an RBC lipoxygenase and (possibly) glycophorin are the only examples for which there is reasonably good evidence. In contrast, the proteins forming the RBC cytoskeleton (spectrin, ankyrin, band 4.1, actin and possibly the major anion exchange transmembrane protein by which the cytoskeleton is attached to the plasma membrane) have closely-related variants in other cell types. Yet two beta-spectrin variants are found exclusively in certain terminally differentiated cells, often only in certain specific regions of the cell membrane. Certain RBC isozymes (e.g. for pyruvate kinase and carbonic anhydrase) and an RBC 19 kD protein (ep19) are also expressed only in a subset of other cell types. This illustrates the importance of gene families which are differentially regulated in certain subsets of cell types during differentiation and development. The expression of the globin genes seems to be regulated mainly at the transcriptional level, although transport of these transcripts to the cytoplasm may be controlled by interactions with other RNAs: stabilisation of globin mRNAs by ribonucleoprotein complexes in the cytoplasm may also be important. In fact, the expression of the globin genes involves two distinct phases: first, structural changes occur in the chromatin surrounding the genes (as determined by sensitivity to digestion by nucleases) and these can be maintained independently of any subsequent transcription. In many cases, these nuclease-sensitive sites in the chromatin correspond to low-level transcription initiation sites and to DNA sequences with regulatory functions when the isolated genes are assayed for transcription in vivo after transfection into cells. How the unlinked alpha- and beta-globin genes are coordinately regulated is not yet understood. Indeed, the alpha- and beta-gene promoters have quite different properties as judged by their responses to DNA replication and to factors known to affect viral gene function (e.g. the cis-acting SV40 enhancer elements and the trans-acting adenovirus regulatory protein, Ela). Other evidence shows that a nuclear protein present only in erythroid cells is able to bind to the beta-globin gene precisely in the region that is hypersensitive to nuclease digestion in chromatin from erythroid cells.(ABSTRACT TRUNCATED AT 400 WORDS)

Enhanced gene expression by the poly(dT-dG).poly(dC-dA) sequence

The sequence poly(dT-dG).poly(dC-dA) (TG-element) is a ubiquitous component of eucaryotic genomes and has the potential to adopt a left-handed DNA conformation (Z-DNA). In this report, we have tested the hypothesis that the TG-element can modulate gene expression. Human genomic DNA fragments (1 to 1.5 kilobases) containing a (dT-dG)n.(dC-dA)n tract (30, 40, or 50 base pairs) or chemically synthesized (dT-dG)n.(dC-dA)n fragments (50 to 130 base pairs) were inserted in the pSV2-cat (simian virus 40 enhancer plus) or pA10-cat (enhancer minus) expression vector plasmid. These constructs were transfected into CV-1 cells or HeLa cells, and their transcription was monitored by assaying chloramphenicol acetyltransferase activity. The results showed that pSV2-cat with the TG-element and pA10-cat with the TG-element synthesized more chloramphenicol acetyltransferase activity (2 to 10 times, depending on the location of the TG-element) than did parental pSV2-cat and pA10-cat DNAs, respectively. Furthermore, the TG-element appeared to have characteristics similar to those of viral enhancers: (i) the TG-element enhanced transcription from a distance, (ii) its closer location to the promoter was more effective, and (iii) its orientation was not crucial. However, its enhancer-like activity was much weaker than that of the simian virus 40 enhancer, and, unlike many viral enhancers, it was equally active in monkey and in human cells. These results suggest that the TG-element may influence the expression of cellular genes.

Characterization of genomic poly(dT-dG).poly(dC-dA) sequences: structure, organization, and conformation

Hybridization studies suggest the abundant presence of poly(dT-dG).poly(dC-dA) (TG-element), a potential Z-DNA sequence, in eucaryotic genomes. We have isolated and characterized TG-elements from different locations in the human genome: from randomly isolated clones, associated with the actin gene family, and linked to another repeated element. The results indicate that the following features are typical of these TG-elements: the elements consist of 20 to 60 base pairs of (dT-dG)n.(dC-dA)n, the sequences characterized in our study were not flanked by direct or inverted repeats, the sequences are interspersed rather than in satellite blocks, the elements are not usually associated with other repeated elements, and some of the elements are found near coding sequences or in introns. Studies on the conformation of a genomic TG-element in a supercoiled plasmid indicate several distinct properties of the TG-element: it is in the Z-form only at low ionic strength, S1 nuclease recognizes its Z-form with a marked preference for one of the B-Z junctions, and the sensitive region extends for 20 base pairs near the B-Z junction. In contrast to the result with the supercoiled plasmid, S1 nuclease failed to recognize the TG-element in minichromosomes.

Human U1 RNA genes contain an unusually sensitive nuclease S1 cleavage site within the conserved 3' flanking region

We find that the cloned DNAs of human U1 small nuclear RNA genes contain two nuclease S1-sensitive sites, one about 1.8 kilobases downstream of the U1 RNA coding region and the other around 0.3 kilobase upstream. The downstream site is unusually sensitive to the nuclease, being cleaved in both linear and negatively supercoiled DNAs. The extent of cleavage at this site is enhanced at lower pH and reduced concentrations of NaCl; the effects of salt are more apparent on linear than supercoiled DNAs. The nuclease S1 sensitivity of this downstream site is dependent on the presence of the sequence (dC-dT)n X (dA-dG)n, where n = 15-25. (One gene with n = 5 is resistant to nuclease S1 cleavage in this region.) In contrast, the nuclease S1 site upstream of the coding region is cleaved only when the DNA is supercoiled. This site also has a homopyrimidine X homopurine bias in the DNA strands, but the sequence is less regular. In the course of these studies, we detected several discrepancies between our restriction maps of some U1 RNA genes and those published by others. Our maps demonstrate that all seven cloned human U1 RNA genes are very similar in sequence for as much as 2.3 kilobases downstream of the U1 RNA coding region.

Thermodynamics of the ColE1 cruciform. Comparisons between probing and topological experiments using single topoisomers

The sensitivity of the ColE1 cruciform to four enzyme and chemical probes of secondary structure has been studied as a function of plasmid topology. Purified topoisomers of pColIR515 have been probed with S1 nuclease, Bal31 nuclease, phage T4 endonuclease VII or osmium tetroxide, and site-specific reaction quantified. Closely similar profiles of reactivity as a function of linking difference were obtained for each probe. Electrophoresis of the pure topoisomers on polyacrylamide/agarose gels revealed a discontinuity in migration as a function of linking difference. Above a threshold linking difference, topoisomers exhibit pronounced reduction in mobility. The linking difference at which this band shift is found correlates precisely with that required for site-specific reaction with the four probes. We conclude that both probing and topological methods are valuable in the study of cruciform structure in supercoiled DNA. The band shift has been measured with accuracy to allow the calculation of the twist change that accompanies the transition, corresponding to delta Tw = -3.2 +/- 0.1. Using this value together with the critical linking difference we calculate a free energy of formation for this structure delta G = 18.4 +/- 0.5 kcal mol-1 (1 kcal = 4.184 kJ).

Internal promoter elements of transfer RNA genes are preferentially exposed in chromatin

We have examined the chromatin organization of a cluster of transfer RNA genes at the cytogenic locus 90BC on the right arm of the third Drosophila melanogaster chromosome. We find that the internal promoter sequences are preferentially exposed to micrococcal nuclease and DNase I in chromatin. Moreover, these exposed sequences have an unusual single strand nuclease-sensitive conformation.

S1-hypersensitive sites in eukaryotic promoter regions

We have examined by fine mapping the S1 nuclease-hypersensitivity of the 5' flanking regions of the human beta-globin and rat preproinsulin II genes and of the SV40 origin/enhancer region. In all cases S1-hypersensitive sites are located in known or presumed promoter/regulatory regions. Though a consensus DNA sequence is not evident, all of these sites reside in predominantly homopurine-homopyrimidine stretches. The alternate (non-B) DNA structure which is revealed by the enzymatic probe is a sequence-dependent feature of a short stretch of DNA, which is retained upon transplantation into a foreign environment. The alternate structure exhibits S1-nicking patterns uniquely different from those associated with the presence of Z-DNA.

Possible structures of homopurine-homopyrimidine S1-hypersensitive sites

S1 nuclease hypersensitive sites in the 5' flanking regions of eukaryotic genes, present in small artificial supercoiled DNA circles, reside in homopurine-homopyrimidine stretches. The hierarchical behavior which these sites exhibit is consistent with the notion that they act as sinks of torsional free energy. By employing DMS as a single-strand-specific reagent, we show that these sites (despite their S1 sensitivity) are regions of duplex DNA. A simple thermodynamic treatment indicates that the high torsional stress in the small DNA circles is almost certain to be relieved by the formation of alternate DNA structures. The same treatment places some constraints on the types and sizes of the regions with alternate conformation. While no definitive structural conclusions can be drawn, left-handed helices seem most consistent with the extent and the pattern of sensitivity to S1 nuclease.

Transcriptionally active chromatin is sensitive to Neurospora crassa and S1 nucleases

We have examined the distribution of Neurospora crassa and S1 nuclease cleavage products in the chromatin of the 87A7 heat shock locus of Drosophila melanogaster. Both of these nucleases generate single and double-strand breaks in chromatin at specific sites in the 87A7 locus. Before heat induction, we find that the 5' ends of the two 87A7 hsp 70 genes contain N. crassa and S1 nuclease hypersensitive sites, while there are only a few cleavage products from elsewhere in the locus. With N. crassa nuclease, we observe one major 5' fragment, and this is derived from cleavage in a DNA segment mapping about 90 to 115 base-pairs from the beginning of the transcription unit. With S1 nuclease, we find two 5' cleavage products. The first maps about 120 to 130 base-pairs from the beginning of the gene. Interestingly, this site is also sensitive to S1 nuclease in supercoiled but not linear naked DNA. The other fragment maps very close to the transcription start site (approximately 0 to -15 base-pairs). After heat induction, there is a transition in the chromatin architecture of 87A7. First, there is a marked reduction in the yield of the prominent 5' N. crassa and S1 nuclease fragments. Second, the entire hsp 70 gene, as well as the spacer DNA just downstream from the 3' end of the gene, becomes highly sensitive to both of these nucleases.

Drosophila DNA topoisomerase I is associated with transcriptionally active regions of the genome

The distribution of DNA topoisomerase I within Drosophila polytene chromosomes was observed by immunofluorescent staining with affinity-purified antibodies. The enzyme is preferentially associated with active loci, as shown by prominent staining of puffs. The heat shock loci 87A-87C are stained after, but not before, heat shock induction. A detailed comparison of the distribution of topoisomerase I with that of RNA polymerase II reveals a similar, although not identical, pattern of association. Topoisomerase I is also found in association with the nucleolus, the site of transcription by RNA polymerase I.

Expression of recombinant genes containing herpes simplex virus delayed-early and immediate-early regulatory regions and trans activation by herpesvirus infection

The promoter-regulatory regions from the herpes simplex virus type 1 (HSV-1) gene for the immediate-early, 175,000-molecular-weight (175K) protein and the HSV-2 delayed-early gene for a 38K protein were linked to the readily assayable bacterial gene for the enzyme chloramphenicol acetyltransferase (CAT). Unexpectedly, in measurements of the constitutive expression of the recombinant genes 40 to 50 h after transfection of Vero cells, enzyme levels expressed from the delayed-early 38K-promoter-CAT construct (p38KCAT) were at least as high as those from the immediate-early 175K-promoter-CAT construct (p175KCAT). In contrast, enzyme levels expressed after transfection of a similar recombinant gene containing a second delayed-early promoter region, that of the HSV-1 thymidine kinase gene, were ca. 20-fold lower. The amounts of enzyme expressed from both p38KCAT and p175KCAT could be increased by up to 20- to 40-fold after infection of the transfected cells with HSV. In comparison, virus infection had no significant effect on enzyme levels expressed from recombinant CAT genes containing the simian virus 40 early promoter region, with or without the 72-base-pair enhancer element. Experiments with the temperature-sensitive mutants HSV-1 tsB7 and HSV-1 tsK indicate that induction of expression from p175KCAT was mediated by components of the infecting virus particle, whereas that from p38KCAT required de novo expression of virus immediate-early proteins. In addition, we show that functions required to induce expression from both p175KCAT and p38KCAT could also be provided by infection with pseudorabies virus and cytomegalovirus.

Mung bean nuclease cleavage of a dA + dT-rich sequence or an inverted repeat sequence in supercoiled PM2 DNA depends on ionic environment

We have determined the nucleotide sequences around two alternative sites cleaved in supercoiled PM2 DNA by single-strand-specific mung bean nuclease in different ionic environments. In 10 mM Tris-HC1 (pH 7.0, 37 degrees C), the major site is a dA+dT-rich sequence which maps with a known early denaturation region at 0.75 map units. About 30 cleavages occurred in a 135 bp region. Cleavages were largely excluded at (dA)n . (dT)n (n = 3-7) sequences. Cleavage patterns of this type have not been previously observed in dA+dT-rich sequences. With the addition of 0.1 M NaC1 the major alternative site occurred in a hyphenated inverted repeat sequence 500 bp away (0.70 map units) and did not map to an early denaturation region. One major and 4 minor cleavages occurred in the region between the repeats, suggesting that a hairpin containing at most a 12 bp stem and 10 base loop is recognized. The basis for nuclease recognition of the dA+dT-rich sequence is not clear. The differences in the sequences and cleavage patterns at the alternative sites indicate that their secondary structures differ.

Transcriptionally active chromatin

Eukaryotic chromatin has a dynamic, complex hierarchical structure. Active gene transcription takes place on only a small proportion of it at a time. While many workers have tried to characterize active chromatin, we are still far from understanding all the biochemical, morphological and compositional features that distinguish it from inactive nuclear material. Active genes are apparently packaged in an altered nucleosome structure and are associated with domains of chromatin that are less condensed or more open than inactive domains. Active genes are more sensitive to nuclease digestions and probably contain specific nonhistone proteins which may establish and/or maintain the active state. Variant or modified histones as well as altered configurations or modifications of the DNA itself may likewise be involved. Practically nothing is known about the mechanisms that control these nuclear characteristics. However, controlled accessibility to regions of chromatin and specific sequences of DNA may be one of the primary regulatory mechanisms by which higher cells establish potentially active chromatin domains. Another control mechanism may be compartmentalization of active chromatin to certain regions within the nucleus, perhaps to the nuclear matrix. Topological constraints and DNA supercoiling may influence the active regions of chromatin and be involved in eukaryotic genomic functions. Further, the chromatin structure of various DNA regulatory sequences, such as promoters, terminators and enhancers, appears to partially regulate transcriptional activity.

Supercoil-dependent features of DNA structure at Drosophila locus 67B1

We have analyzed the pattern of supercoil-dependent, single strand-specific nuclease cleavage sites across 11.6 kb (11.6 X 10(3) base-pairs) of cloned Drosophila melanogaster DNA from locus 67B1. This region contains coding sequences for the heat shock proteins hsp23, hsp26 and hsp28 as well as for a 1.6 kb developmentally regulated transcript (R). Two major sites are detected on digestion with S1 nuclease or mung bean nuclease. The most prominent site maps 100 base-pairs upstream of hsp26 in a very pyrimidine-rich region adjacent to a known region of chromatin DNAase I hypersensitivity. The other site is located approximately 800 base-pairs upstream of hsp28 in an area devoid of such chromatin-specific features. BAL31 nuclease produces a different array, with three to six strong cleavages located in the spacer DNA approx. 0.1 to 1.0 kb upstream of the DNAase I hypersensitive sites of hsp28, hsp23 and R. Thus, for each gene in the cluster a localized sequence sensitive to the winding state of the DNA is observed 5' to the gene. However, there is no precise coincidence of any of the major sites sensitive to BAL31 nuclease in the supercoiled plasmid with the sequences sensitive to DNAase I in chromatin. While all of the enzymes utilized in this study have prominent single strand-specific endonucleolytic activity, it is clear that they recognize different variants in the DNA structure induced by supercoiling. At least two classes of DNA perturbation have been detected.

Differential response of multiple epsilon-globin cap sites to cis- and trans-acting controls

The human epsilon-globin gene has a number of alternative transcription-initiation sites located upstream of the canonical mRNA cap site. In three nonerythroid cell lines, "leaky" epsilon-globin transcription occurs exclusively from one of these upstream sites, the -200 cap site. Using a transient expression assay, we have shown that transcription initiation from the -200 cap site and the major cap site can be independently regulated in response to plasmid replication, SV40 enhancer sequences in cis, and the adenovirus E1A gene in trans. The -200 cap site is located within a region of S1 hypersensitivity in the supercoiled plasmid, and in the absence of viral enhancer sequences it is the main initiation site following transfection into a number of cell lines. We suggest that the -200 cap site acts as a polymerase entry site by virtue of its accessible chromatin structure. The efficiency of polymerase binding at this site may be altered by trans-acting regulatory molecules.

DNAase I hypersensitive sites may be correlated with genomic regions of large structural variation

Helical-twist, roll and torsion-angle variations calculated by the Calladine (1982)-Dickerson (1983) rules were scanned along several nucleotide sequences for which DNAase I cleavage data are available. It has been shown that for short synthetic oligomers DNAase I cuts preferentially at positions of high helical twist (Dickerson & Drew, 1981; Lomonossoff et al., 1981). Our calculations indicate that DNAase I sensitive and hypersensitive sites in chromatin are correlated with regions of successive, large, helical-twist angle variations from regular B-DNA. In many cases these regions exhibit large variations in base-pair roll and backbone torsion angles as well. It has been suggested that DNAase I cuts in the vicinity of cruciforms. However, it was recently demonstrated by Courey & Wang (1983) and Gellert et al. (1983) that such cruciform formation in a negatively supercoiled DNA is kinetically forbidden under physiological conditions. We thus propose that clustering of large twist-angle (and/or roll and backbone torsion angle) variations may be among the conformational features recognized by the enzyme. Specific cuts can then preferentially occur at base-pair steps with high helical twists.

Poly(pyrimidine) . poly(purine) synthetic DNAs containing 5-methylcytosine form stable triplexes at neutral pH

Poly(pyrimidine) . poly(purine) tracts have been discovered in the 5'-flanking regions of many eucaryotic genes. They may be involved in the regulation of expression since they can be mapped to the nuclease-sensitive sites of active chromatin. We have found that poly(pyrimidine) . poly(purine) DNAs which contain 5-methylcytosine (e.g. poly[d(Tm5C)] . poly[d(GA)]) will form a triplex at a pH below 8. In contrast, the unmethylated analogue, poly[d(TC)] . poly[d(GA)] only forms a triplex at pHs below 6. Synthetic DNAs containing repeating trinucleotides and poly[d(Um5C)] . poly[d(GA)] behave in a similar manner. Thus the stability of a triplex can be controlled by methylation of cytosine. This suggests a model for the regulation of expression based upon specific triplex formation on the 5'-side of eucaryotic genes.

The histone H5 gene is flanked by S1 hypersensitive structures

The potential of the cloned histone H5 gene to form altered DNA structures has been examined by S1 nuclease digestion of supercoiled recombinant plasmids containing up to 8.8 kbp of chicken DNa. The three main nicking sites map at the upstream and downstream sequences flanking the structural gene. The cleavage sites share sequence homology, strand specificity, and do not seem to be single-stranded. The sequence of the S1-sensitive sites does not suggest that the fragments can adopt any of the known DNA secondary structures.

Reversible and persistent changes in chromatin structure accompany activation of a glucocorticoid-dependent enhancer element

A derivative of mouse mammary tumor virus (MTV) DNA, LTL, was constructed in vitro and introduced into the genome of mouse L cells. Transcription of LTL was stimulated by dexamethasone, a glucocorticoid hormone. Two features of LTL chromatin structure are altered upon hormone treatment. First, "moderate" DNAase I sensitivity of the entire LTL element increases following addition of dexamethasone; this alteration persists after hormone withdrawal, when LTL transcription is shut off. Second, a discrete DNAase I-hypersensitive region is induced with a time course that closely parallels the rate of increasing transcription from the MTV promoter; this structure disappears upon hormone removal. The induced hypersensitive region coincides with a segment of the MTV long terminal repeat sequence that specifically binds purified glucocorticoid receptor in vitro and functions as a hormone-dependent enhancer element in vivo. We suggest that specific glucocorticoid receptor-DNA interactions may alter the configuration of DNA or chromatin in the vicinity of the binding sites, thereby creating an active transcriptional enhancer.

Differences in human alpha- and beta-globin gene expression in mouse erythroleukemia cells: the role of intragenic sequences

Human beta-globin genes introduced into mouse erythroleukemia (MEL) cells by DNA cotransformation are correctly regulated when erythroid cell differentiation is induced by dimethylsulfoxide (DMSO). In contrast, cloned human alpha-globin genes are efficiently transcribed in MEL cells prior to induction, and no increase in the level of alpha-globin mRNA is observed when the cells differentiate. These observations suggest that the mechanisms by which alpha- and beta-globin genes are activated during erythroid cell differentiation are fundamentally different. Analysis of the transcription of hybrid human alpha/beta-globin genes in MEL cells revealed that the sequences responsible for differences in transcription of the intact alpha- and beta-globin genes are located on the 3' side of the mRNA capping site of the two genes, suggesting that cis-acting regulatory sequences are located within the structural genes.

Histone-H1-dependent chromatin superstructures and the suppression of gene activity

I have identified a chromatin particle containing DNA as large as 20-40 kb that migrates as a discrete entity on agarose gels. With increasing nuclease digestion, the particle becomes cleaved in the linker regions between nucleosomes, but remains intact, probably held together by the outer histones, H1 and H5. By hybridization analysis, inactive genes are found in these particles. Active genes (and their flanking sequences) are also found in particles containing H1 and H5, but in contrast to inactive supranucleosome particles, active polynucleosome particles are not held together after cleavage of linker DNA. This suggests that H1 cross-links adjacent nucleosomes in inactive regions and that H1 is bound differently in expressed regions. The results raise the possibility that the marked degree of suppression of repressed, tissue-specific genes may be determined, in part, by their assembly into these inactive supranucleosome structures.

Simian virus 40 illegitimate recombination occurs near short direct repeats

We have analysed nucleotide sequences at the junction between simian virus 40 (SV40) and cellular DNA in the Fisher rat transformed line tsA30-N2. This line contains a single insertion of one complete SV40 genome with a terminal duplication of 267 nucleotides, the recombination sites being located at nucleotides 439 and 705 in the late region of SV40. These two positions are located within short direct repeats in the virus genome. In order to test the significance of such repeats with respect to illegitimate recombination events, we analysed two series of published sequences of SV40 recombination sites: the first one consists of eight SV40 insertion endpoints derived from four SV40-transformed cell lines; the second one consists of 18 junction points from SV40 evolutionary variants. Our analysis demonstrates that in both cases, recombination preferentially takes place near short direct repeats in the virus genome. A model involving a "slipped mispairing" mechanism is proposed in order to account for this finding.

DNA strand breakage by wheat germ type 1 topoisomerase

Properties of strand breakage in duplex and single-stranded DNA by the wheat germ type 1 DNA topoisomerase were investigated. Strand breakage in duplex DNA is dependent upon the use of denaturing conditions to inactivate the enzyme and terminate the reaction, whereas breakage of single-stranded DNA occurs under the normal reaction conditions and is not dependent upon denaturation. Breakage generates a free 5' hydroxyl group and enzyme bound to the 3' side of the break, presumably via the 3' phosphate group. The location of sites of breakage with both duplex and single-stranded DNA is not random. In all these respects the wheat germ enzyme closely resembles the rat liver type 1 topoisomerase. A comparison of the locations of the sites of breakage in duplex DNA generated by the topoisomerases from wheat germ and rat liver indicates a number of common sites, although the patterns of breakage are not identical.

Structural specificities of five commonly used DNA nucleases

Five commonly used nucleases were surveyed for their ability to distinguish among several different DNA backbone configurations. The digestion data suggest that: (1) DNAase I binds across the minor groove; whereas (2) nuclease S1 and (3) micrococcal nuclease bind to an exposed single strand; (4) copper/phenanthroline seeks a base-pair step; and (5) DNAase II requires just a stacked single strand of limited exposure. Only micrococcal nuclease is demonstrably base-specific, with a strong preference for T, A over C, G in any structural context.

Barriers to nuclease Bal31 digestion across specific sites in simian virus 40 chromatin

A portion of the nucleoprotein containing viral DNA extracted from cells infected by simian virus (SV40) is preferentially cleaved by endonucleases in a region of the genome encompassing the origin of replication and early and late promoters. To explore this nuclease-sensitive structure, we cleaved SV40 chromatin molecules with restriction enzymes and digested the exposed termini with nuclease Bal31. Digestion proceeded only a short distance in the late direction from the MspI site, but some molecules were degraded 400 to 500 base pairs in the early direction. By comparison, BglI-cleaved chromatin was digested for only a short distance in the early direction, but some molecules were degraded 400 to 450 base pairs in the late direction. These barriers to Bal31 digestion (bracketing the BglI and the MspI sites) define the borders of the same open region in SV40 chromatin that is preferentially digested by DNase I and other endonucleases. In a portion of the SV40 chromatin, Bal31 could not digest through the nuclease-sensitive region and reached barriers after digesting only 50 to 100 base pairs from one end or the other. Chromatin molecules that contain barriers in the BglI to MspI region are physically distinct from molecules that are open in this region as evidenced by partial separation of the two populations on sucrose density gradients.

Nucleosome core particle stability and conformational change. Effect of temperature, particle and NaCl concentrations, and crosslinking of histone H3 sulfhydryl groups

We have studied the reversible dissociation of core size DNA from chicken erythrocyte nucleosome core particles in solutions containing 0 X 1 M to 0 X 6 M-NaCl. Dissociation increases with increasing NaCl concentration, increasing temperature and decreasing particle concentration. At high particle concentrations, no free DNA is observed below 0 X 3 M-NaCl, whereas above 0 X 3 M-NaCl a lower limit of dissociation is reached. A theoretical analysis based on the migrating-octamer mechanism of Stein is in disagreement with his conclusions concerning dependence of core particle dissociation on particle concentration, but provides a good explanation for our observations, and those of others, using salt concentrations up to 1 M-NaCl. It appears that the core particle is not stabilized primarily by electrostatic interactions. DNA length is not critical for core particle stabilization. The conformation of remaining intact nucleosome core particles changes only moderately within the range of NaCl concentrations studied. Crosslinking by copper phenanthroline of the Cys110 histone H3 single sulfhydryl groups in the intact nucleosome core particle leads to a decrease in stability, yet essentially unchanged hydrodynamic properties are maintained at 0 X 6 M-NaCl, confirming conclusions derived from the behavior of the native core particles. Values for density increments of nucleosome core particles over a range of NaCl concentrations are also given. A method is described for studying binding of histones to nucleosome core particles in the ultracentrifuge by scanning at 230 and 260 nm.

How damaged is the biologically active subpopulation of transfected DNA?

Relatively little is known about the damage suffered by transfected DNA molecules during their journey from outside the cell into the nucleus. To follow selectively the minor subpopulation that completes this journey, we devised a genetic approach using simian virus 40 DNA transfected with DEAE-dextran. We investigated this active subpopulation in three ways: (i) by assaying reciprocal pairs of mutant linear dimers which differed only in the arrangement of two mutant genomes; (ii) by assaying a series of wild-type oligomers which ranged from 1.1 to 2.0 simian virus 40 genomes in length; and (iii) by assaying linear monomers of simian virus 40 which were cleaved within a nonessential region to leave either sticky, blunt, or mismatched ends. We conclude from these studies that transfected DNA molecules in the active subpopulation are moderately damaged by fragmentation and modification of ends. As a whole, the active subpopulation suffers about one break per 5 to 15 kilobases, and about 15 to 20% of the molecules have one or both ends modified. Our analysis of fragmentation is consistent with the random introduction of double-strand breaks, whose cause and exact nature are unknown. Our analysis of end modification indicated that the most prevalent form of damage involved deletion or addition of less than 25 base pairs. In addition we demonstrated directly that the efficiencies of joining sticky, blunt, or mismatched ends are identical, verifying the apparent ability of cells to join nearly any two DNA ends and suggesting that the efficiency of joining approaches 100%. The design of these experiments ensured that the detected damage preceded viral replication and thus should be common to all DNAs transfected with DEAE-dextran and not specific for viral DNA. These measurements of damage within transfected DNA have important consequences for studies of homologous and nonhomologous recombination in somatic cells as is discussed.

Chromatin assembly in Xenopus oocytes: in vitro studies

We describe and characterize a complex reaction that catalyzes DNA supercoiling and chromatin assembly in vitro. A Xenopus oocyte extract supplemented with ATP and Mg++ converts DNA circles into minichromosomes that display a native, 200 bp periodicity. When supercoiled DNA is added to this extract it undergoes a time-dependent series of topological changes, which precisely mimic those found when the DNA is microinjected into oocytes. As judged by the conformation of the subsequently deproteinized DNA, the supercoiled DNA is first relaxed, in a reaction that takes 4 min, and then it is resupercoiled in a slower process that takes 4 hr. The relaxation is partially inhibited by EDTA, to an extent that suggests that that it is catalyzed by a type I DNA topoisomerase. The resupercoiling , on the other hand, requires ATP and Mg++, is completely inhibited by EDTA, and is inhibited by novobiocin in a manner that suggests it is catalyzed by a type II DNA topoisomerase. These findings, and the ones reported in the preceding paper ( Ryoji and Worcel , 1984), lead us to propose that chromatin assembly is an active, ATP-driven process.

Chromatin assembly in Xenopus oocytes: in vivo studies

We have followed the time course of chromatin assembly, DNA supercoiling, and transcription on a Xenopus 5S RNA gene clone injected into germinal vesicles of Xenopus oocytes. During the first 2 hr after DNA injection, there is a gradual enhancement in transcription that correlates with the increase in superhelical density of the DNA template; on the other hand, nucleosome assembly is already completed by 10-30 min after DNA injection. To probe further the DNA structure in the assembled minichromosomes, we injected enzymes and chemicals into the germinal vesicle. DNAase I and topoisomerase I injections reveal that the circular DNA has been assembled into two discrete and equally abundant types of chromatin: one type, which we call "dynamic" chromatin, is torsionally strained and is thus fully relaxed by those two enzymes. The other type, which we call "static" chromatin, still yields supercoiled DNA molecules after deproteinization. The dynamic chromatin is also relaxed by injection of novobiocin, and simultaneously, 5S RNA transcription is turned off. The results of our in vivo experiments suggest that the dynamic chromatin is the one that is transcriptionally active. We discuss the biological relevance of these findings.

DNase I sensitivity of the alpha 2(I) collagen gene: correlation with its expression but not with its methylation pattern

The chromatin structure of the chick alpha 2(I) collagen gene was probed with DNase I. Because our previous work strongly suggested that the 5' end of this gene is not methylated whereas the rest of the gene is methylated whether or not the gene is expressed, we compared the relative DNase I sensitivity of the methylated and unmethylated segments. Both regions demonstrate similar relative DNase I sensitivities within a given tissue. In chromatin of chick embryo fibroblasts, we find a DNase I hypersensitive site which maps between 100 and 300 bp preceding the start of transcription. This site is not found in brain chromatin but is present in chick embryo fibroblasts transformed by Rous Sarcoma virus although the rate of transcription of the alpha 2(I) collagen gene is greatly reduced in these cells. Hence, the mechanism responsible for the large decrease in alpha 2(I) collagen gene expression in RSV transformed cells is different from the mechanism that is responsible for the presence of a DNase I hypersensitive site in the promoter. Furthermore, changes in the DNase I sensitivity of the chromatin of the alpha 2(I) collagen promoter occur without changes in the methylation pattern of the gene.