Limited nucleosome migration can completely randomize DNA repair patches in intact human cells

Following irradiation of human cells with ultraviolet light, DNA repair patches are initially inserted near the 5' and 3' ends of nucleosome core DNA leaving a "gap" in repair synthesis (of approximately 50 bases) near the center of the core DNA. With time, however, these same repair patches become randomized, apparently by nucleosome migration. We have developed both an analytical expression and a computer algorithm (which simulates nucleosome migration along DNA) to determine the average distance nucleosomes must migrate to change the initial, non-uniform distribution of repair patches in nucleosomes to a random distribution. Both of these methods yielded the same result: nucleosomes must migrate an average of about 50 base-pairs in either direction to produce the randomization observed.     

A nonuniform distribution of excision repair synthesis in nucleosome core DNA

We have investigated the distribution in nucleosome core DNA of nucleotides incorporated by excision repair synthesis occurring immediately after UV irradiation in human cells. We show that the differences previously observed for whole nuclei between the DNase I digestion profiles of repaired DNA (following its refolding into a nucleosome structure) and bulk DNA are obtained for isolated nucleosome core particles. Analysis of the differences obtained indicates that they could reflect a significant difference in the level of repair-incorporated nucleotides at different sites within the core DNA region. To test this possibility directly, we have used exonuclease III digestion of very homogeneous sized core particle DNA to "map" the distribution of repair synthesis in these regions. Our results indicate that in a significant fraction of the nucleosomes the 5' and 3' ends of the core DNA are markedly enhanced in repair-incorporated nucleotides relative to the central region of the core particle. A best fit analysis indicates that a good approximation of the data is obtained for a distribution where the core DNA is uniformly labeled from the 5' end to position 62 and from position 114 to the 3' end, with the 52-base central region being devoid of repair-incorporated nucleotides. This distribution accounts for all of the quantitative differences observed previously between repaired DNA and bulk DNA following the rapid phase of nucleosome rearrangement when it is assumed that linker DNA and the core DNA ends are repaired with equal efficiency and the nucleosome structure of newly repaired DNA is identical with that of bulk chromatin. Furthermore, the 52-base central region that is devoid of repair synthesis contains the lowest frequency cutting sites for DNase I in vitro, as well as the only "internal" locations where two (rather than one) histones interact with a 10-base segment of each DNA strand.     

Stability of nucleosome placement in newly repaired regions of DNA

Rearrangements of chromatin structure during excision repair of UV-damaged DNA appear to involve unfolding of nucleosomal DNA while repair is taking place, followed by refolding of this DNA into a native nucleosome structure. Recently, we found that repair patches are not distributed uniformly along the DNA in nucleosome core particles immediately following their refolding into nucleosomes (Lan, S. Y., and Smerdon, M. J. (1985) Biochemistry, 24,7771). Therefore, the distribution of repair patches in nucleosome core DNA was used to monitor the stability of nucleosome placement in these regions. Our results indicate that in nondividing human cells undergoing excision repair there is a slow change in the positioning of nucleosomes in newly repaired regions of chromatin, resulting in the eventual randomization of repair patches in nucleosome core DNA. Furthermore, the nonrandom placement of nucleosomes observed just after the refolding event is not re-established during DNA replication. Possible mechanisms for this change in nucleosome placement along the DNA are discussed.     

Enhanced DNA repair synthesis in hyperacetylated nucleosomes

We have investigated the level of "early" DNA repair synthesis in nucleosome subpopulations, varying in histone acetylation, from normal human fibroblasts treated with sodium butyrate. We find that repair synthesis occurring during the first 30 min after UV irradiation is significantly enhanced in hyperacetylated mononucleosomes. Nucleosomes with an average of 2.3 acetyl residues/H4 molecule contained approximately 1.8-fold more repair synthesis than nucleosomes with an average of 1.5 or 1.0 acetyl residues/H4 molecule. Fractionation of highly acetylated nucleosomes by two-dimensional gel electrophoresis yielded an additional 2.0-fold enrichment of repair synthesis in nucleosomes containing 2.7 acetyl residues/H4 molecule as compared to nucleosomes containing 1.9 acetyl residues/H4 molecule. This enhanced repair synthesis is associated primarily with nucleosome core regions and does not appear to result from increased UV damage in hyperacetylated chromatin. In addition, the distribution of repair synthesis within nucleosome core DNA from hyperacetylated chromatin is nonrandom, showing a bias toward the 5' end which is similar to that obtained for bulk (unfractionated) chromatin. These results provide strong evidence that enhanced repair occurs within nucleosome cores of hyperacetylated chromatin in butyrate-treated human cells. Finally, pulse-chase experiments demonstrate that the association of enhanced repair synthesis with hyperacetylated nucleosomes is transient, lasting only about 12 h after UV damage.     

Nucleosome rearrangement in human cells following short patch repair of DNA damaged by bleomycin.

We have examined the structure of newly repaired regions of chromatin in intact and permeabilized human cells following exposure to bleomycin (BLM). The average repair patch size (in permeabilized cells) was six to nine bases, following doses of 1-25 micrograms/mL BLM, and greater than 80% of the total repair synthesis was resistant to aphidicolin. In both intact and permeabilized cells, nascent repair patches were initially very sensitive to staphylococcal nuclease, analogous to repair induced by "long patch" agents, and are nearly absent from isolated nucleosome core DNA. Unlike long patch repair, however, the loss of nuclease sensitivity during subsequent chase periods was very slow in intact cells, or in permeabilized cells treated with a low dose of BLM (1 microgram/mL), and was abolished by treatment with hydroxyurea (HU) or aphidicolin (APC). The rate of repair patch ligation did not correlate with this slow rate of chromatin rearrangement since greater than 95% of the patches were ligated within 6 h after incorporation (even in the presence of HU or APC). In permeabilized cells, repair patches induced by either 5 or 25 micrograms/mL BLM, where significant levels of strand breaks occur in compact regions of chromatin, lost the enhanced nuclease sensitivity at a rate similar to that observed following long patch repair. This rapid rate of rearrangement was not affected by APC. These results indicate that short patch repair in linker regions of nucleosomes, and/or "open" regions of chromatin, involves much less nucleosome rearrangement than long patch repair or short patch repair in condensed chromatin domains.     

Site-specific DNA repair at the nucleosome level in a yeast minichromosome

The rate of excision repair of UV-induced pyrimidine dimers (PDs) was measured at specific sites in each strand of a yeast minichromosome containing an active gene (URA3), a replication origin (ARS1), and positioned nucleosomes. All six PD sites analyzed in the transcribed URA3 strand were repaired more rapidly (greater than 5-fold on average) than any of the nine PD sites analyzed in the nontranscribed strand. Efficient repair also occurred in both strands of a disrupted TRP1 gene (ten PD sites), containing four unstable nucleosomes, and in a nucleosome gap at the 5' end of URA3 (two PD sites). Conversely, slow repair occurred in both strands immediately downstream of the URA3 gene (12 of 14 PD sites). This region contains the ARS1 consensus sequence, a nucleosome gap, and two stable nucleosomes. Thus, modulation of DNA repair occurs in a simple yeast minichromosome and correlates with gene expression, nucleosome stability, and (possibly) control of replication.     

Completion of excision repair in human cells. Relationship between ligation and nucleosome formation

The completion of excision repair patches in human cells, following UV irradiation, was compared to the refolding of these regions into nucleosomes. Incomplete repair patches were detected by their enhanced sensitivity to exonuclease III. This enhanced sensitivity was due to the presence of gaps (or displaced parental strands) at the 3' end rather than unligated nicks, indicating that ligation occurs rapidly after repair synthesis is completed. Different rates of completion were achieved by treatment with the inhibitors hydroxyurea and sodium butyrate, as well as by using a (partially) ligase-deficient human cell strain. Hydroxyurea caused a marked decrease in both the rate of completion and the level of repair incorporation in all three cell types studied, while sodium butyrate yielded different effects in each cell type. In each case, however, a decrease in the rate of repair patch completion resulted in a concomitant decrease in the level of nucleosome formation. To determine the temporal relationship of these two events, the levels of repair-incorporated nucleotides in isolated 146-base pair nucleosome core DNA were compared on native and denaturing gels. The data indicate that little (or no) nucleosome formation occurred in the nascent DNA regions prior to ligation regardless of the cell type or treatment used. Furthermore, comparison of the fraction of unligated repair patches and the fraction of repair patches in a nonnucleosomal state indicated that in the absence of inhibitors there was a significant time lag between ligation and nucleosome formation. This lag time, however, decreased when cells were treated with hydroxyurea. Thus, the formation of nucleosomes in newly repaired regions of DNA occurred after the ligation step in all cases and these two features of the excision repair process are not "tightly coupled" events.     

Nucleosome structure and repair of N-methylpurines in the GAL1-10 genes of Saccharomyces cerevisiae

Nucleosome structure and repair of N-methylpurines were analyzed at nucleotide resolution in the divergent GAL1-10 genes of intact yeast cells, encompassing their common upstream-activating sequence. In glucose cultures where genes are repressed, nucleosomes with fixed positions exist in regions adjacent to the upstream-activating sequence, and the variability of nucleosome positioning sharply increases with increasing distance from this sequence. Galactose induction causes nucleosome disruption throughout the region analyzed, with those nucleosomes close to the upstream-activating sequence being most striking. In glucose cultures, a strong correlation between N-methylpurine repair and nucleosome positioning was seen in nucleosomes with fixed positions, where slow and fast repair occurred in nucleosome core and linker DNA, respectively. Galactose induction enhanced N-methylpurine repair in both strands of nucleosome core DNA, being most dramatic in the clearly disrupted, fixed nucleosomes. Furthermore, N-methylpurines are repaired primarily by the Mag1-initiated base excision repair pathway, and nucleotide excision repair contributes little to repair of these lesions. Finally, N-methylpurine repair is significantly affected by nearest-neighbor nucleotides, where fast and slow repair occurred in sites between pyrimidines and purines, respectively. These results indicate that nucleosome positioning and DNA sequence significantly modulate Mag1-initiated base excision repair in intact yeast cells.     

Distribution of DNA damage in chromatin and its relation to repair in human cells treated with 7-bromomethylbenz(a) anthracene.

We have examined the relationship between the distribution of DNA damage and repair in chromatin from confluent human fibroblasts treated with the carcinogen 7-bromomethylbenz (a) anthracene. Analysis of staphylococcal nuclease (SN)4 digestion kinetics and gel electrophoresis revealed that more total damage occurs in nucleosome core DNA (approximately 80-85% of chromatin DNA) than in SN sensitive DNA (APPROXIMATELY15-20%). Furthermore, over a 24 hr period, damage is removed at about the same rate from these two regions. In contrast, virtually all of the nucleotides incorporated during repair synthesis are initially SN sensitive even when measured at 12 hr after damage. With time many repair-incorporated nucleotides become SN resistant and coelectrophorese with nucleosome core DNA. To explain these data we propose a model whereby excision repair occurs in both linker and core DNA; however, in core DNA the repair process induces conformational changes resulting in temporarily increased SN sensitivity; subsequently, rearrangement occurs and results in the re-establishment of native or near-native nucleosome conformation and SN resistance.     

Rearrangements of chromatin structure in newly repaired regions of deoxyribonucleic acid in human cells treated with sodium butyrate or hydroxyurea

The rate and extent of redistribution of repair-incorporated nucleotides within chromatin during very early times (10-45 min) after ultraviolet irradiation were examined in normal human fibroblasts treated with 20 mM sodium butyrate, or 2-10 mM hydroxyurea, and compared to results for untreated cells. Under these conditions, DNA replicative synthesis is reduced to very low levels in each case. However, DNA repair synthesis is stimulated by sodium butyrate and partially inhibited by hydroxyurea. Furthermore, in the sodium butyrate treated cells, the core histones are maximally hyperacetylated. Using methods previously described by us, it was found that treatment with sodium butyrate had little or no effect on either the rate or the extent of redistribution of repair-incorporated nucleotides during this early time interval. On the other hand, there was a 1.7-2.5-fold decrease in the rate of redistribution of these nucleotides in cells treated with hydroxyurea; the extent of redistribution was unchanged in these cells. Since hydroxyurea has been shown to decrease the rate of completion of "repair patches" in mammalian cells, these results indicate that nucleosome rearrangement in newly repaired regions of DNA does not occur until after the final stages of the excision repair process are completed. Furthermore, hyperacetylation of the core histones in a large fraction of the total chromatin prior to DNA damage and repair synthesis does not appear to alter the rate or extent of nucleosome core formation in newly repaired regions of DNA.     

Preferential repair of nuclear matrix associated DNA in xeroderma pigmentosum complementation group C

The distribution of ultraviolet-induced DNA repair patches in the genome of xeroderma pigmentosum cells of complementation group C was investigated by determining the molecular weight distribution of repair labeled DNA and prelabeled DNA in alkaline sucrose gradients after treatment with the dimerspecific endonuclease V of bacteriophage T4. The results were consistent with the data reported by Mansbridge and Hanawalt (1983) and suggest that DNA-repair synthesis in xeroderma pigmentosum cells of complementation group C occurs in localized regions of the genome. Analysis of the spatial distribution of ultraviolet-induced repair patches in DNA loops attached to the nuclear matrix revealed that in xeroderma pigmentosum cells of complementation group C repair patches are preferentially situated near the attachment sites of DNA loops at the nuclear matrix. In normal human fibroblasts we observed no enrichment of repair-labeled DNA at the nuclear matrix and repair patches appeared to be distributed randomly along the DNA loops. The enrichment of repair-labeled DNA at the nuclear matrix in xeroderma pigmentosum cells of complementation group C may indicate that the residual DNA-repair synthesis in these cells occurs preferentially in transcribing regions of the genome.     

Nucleosomal DNA is digested to repeats of 10 bases by exonuclease III

Nucleosomes were treated with increasing concentrations of exonuclease III (Exo III) from E. coli. At low levels of Exo III, the heterogeneous distribution of monomers (with associated DNA fragments ranging in size between 140 and 170 bp) is "trimmed" down to a discrete core of 140 bp. The "trimming" of monomers to 140 bp results from a 3' exonucleolytic digestion accompanied by a 5' clipping activity which is specific for the conformation of internucleosomal DNA. At higher concentrations of Exo III, the enzyme digests the 140 bp "trimmed" nucleosome core from both 3' ends without associated 5' nuclease activity. Most striking is the observation that the fragments produced during such a digestion display discrete single-stranded lengths that are integer multiples of 10 bases. For some dimer nucleosomes, Exo III can digest as many as 200 bases from at least one 3' end and produce a 10 base interval ladder from about 400 bases down to 180 bases. This suggests that the enzyme can traverse the length of an entire nucleosome without destroying whatever structural features are necessary to produce a 10 base DNA ladder.     

Escherichia coli uracil- and ethenocytosine-initiated base excision DNA repair: rate-limiting step and patch size distribution

The rate, extent, and DNA synthesis patch size of base excision repair (BER) were measured using Escherichia coli GM31 cell-free extracts and a pGEM (form I) DNA substrate containing a site-specific uracil or ethenocytosine target. The rate of complete BER was stimulated (approximately 3-fold) by adding exogenous E. coli DNA ligase to the cell-free extract, whereas addition of E. coli Ung, Nfo, Fpg, or Pol I did not stimulate BER. Hence, DNA ligation was identified as the rate-limiting step in the E. coli BER pathway. The addition of exogenous DNA polymerase I caused modest inhibition of BER, which was overcome by concomitant addition of DNA ligase. Repair patch size determinations were performed to assess the distribution of DNA synthesis associated with both uracil- and ethenocytosine-initiated BER. During the early phase (0-5 min) of the BER reaction, the large majority of repair events resulted from short patch (1-nucleotide) DNA synthesis. However, during the late phase (>10 min) both short and long (2-20 nucleotide) patches were observed, with long patch BER progressively dominating the repair process. In addition, the patch size distribution was influenced by the ratio of DNA polymerase I to DNA ligase activity in the reaction. A novel mode of BER was identified that involved DNA synthesis tracts of >205 nucleotides in length and termed very-long patch BER. This BER process was dependent upon DNA polymerase I since very-long patch BER was inhibited by DNA polymerase I antibody and addition of excess DNA polymerase I reversed this inhibition.     

Native HMGB1 protein inhibits repair of cisplatin-damaged nucleosomes in vitro

The high mobility group box (HMGB) 1 protein, one of the most abundant nuclear non-histone proteins has been known for its inhibitory effect on repair of DNA damaged by the antitumor drug cisplatin. Here, we report the first results that link HMGB1 to repair of cisplatin-treated DNA at nucleosome level. Experiments were carried out with three types of reconstituted nucleosomes strongly positioned on the damaged DNA: linker DNA containing nucleosomes (centrally and end-positioned) and core particles. The highest repair synthesis was registered with end-positioned nucleosomes, two and three times more efficient than that with centrally positioned nucleosomes and core particles, respectively. HMGB1 inhibited repair of linker DNA containing nucleosomes more efficiently than that of core particles. Just the opposite was the effect of the in vivo acetylated HMGB1: stronger repair inhibition was obtained with core particles. No inhibition was observed with HMGB1 lacking the acidic tail. Binding of HMGB1 proteins to different nucleosomes was also analysed. HMGB1 bound preferentially to damage nucleosomes containing linker DNA, while the binding of the acetylated protein was linker independent. We show that both the repair of cisplatin-damaged nucleosomes and its inhibition by HMGB1 are nucleosome position-dependent events which are accomplished via the acidic tail and modulated by acetylation.     

Chromatin dynamics. Fast and slow modes of nucleosome movement revealed through psoralen binding and repair

Psoralen adducts, when formed in DNA at low frequencies that permit extensive survival of normal and repair-deficient cells, are found in both linker and core regions of nucleosomes, but are slightly enriched in the linker sites. The relative frequencies of adducts obtained with 5-methylisopsoralen and angelicin, which form only monoadducts, and 8-methoxypsoralen and trimethylpsoralen, which form monoadducts and cross-links, represent an enrichment in linker DNA that is approx. 2-3-fold higher per nucleotide than in core DNA. 5-Methylisopsoralen monoadducts, which are initially in linker DNA, become randomized during 12 h of growth. This suggests a slow lateral movement of nucleosomes with respect to DNA and implies that linker and core regions of DNA are not permanent assignments. Randomization of 5-methylisopsoralen adducts is independent of the synthesis of DNA, RNA, protein, or poly(ADP-ribose) and is also independent of DNA repair. Excision repair of these adducts, in contrast, causes rapid local changes in nucleosome conformation and an initial increase in staphylococcal nuclease sensitivity that reverts to the sensitivity of bulk chromatin in less than 1 h. Chromatin, therefore, can undergo at least two distinct dynamic changes under physiological conditions: a slow randomization of the nucleosomes with respect to DNA, and a rapid but transient local rearrangement to facilitate repair.     

Deoxyribonucleic acid excision repair in chromatin after ultraviolet irradiation of human fibroblasts in culture.

We have exposed confluent normal human fibroblasts to ultraviolet (UV) fluences of 5, 14, or 40 J/m2 and monitored the specific activity of post-UV repair synthesis in chromatin with [3H]thymidine pulses. We have shown that under conditions where no semiconservative deoxyribonucleic acid (DNA) synthesis is detectable, the specific activity of repair label in micrococcal nuclease resistant (core particle) DNA is about one-fifth that in bulk DNA at all three UV fluences. On the other hand, the distribution of thymine-containing pyrimidine dimers in bulk and nuclease-resistant regions measured either immediately after irradiation or at later times showed no significant differences; preferential labeling of linker (nuclease-sensitive) DNA during repair synthesis is thus apparently not due to a predominance of UV-induced photoproducts in linker relative to core particle DNA in the nucleosome. Pulse and pulse--chase experiments at 14 or 40 J/m2 with normal human or repair-deficient xeroderma pigmentosum (XP) cells showed that at most 30% of repair label in all these cells shifts from nuclease-sensitive (linker) DNA to nuclease-resistant (core particle) DNA.     

Photofootprint of nucleosome core DNA in intact chromatin having different structural states

Recently, we reported that the distribution of ultraviolet light (u.v.) induced pyrimidine dimers in nucleosome core DNA has a striking 10.3(+/- 0.1) base periodicity and the regions of enhanced quantum yield map to positions where DNA strands are farthest from the core histone surface. Improvement of the mapping procedure has allowed us to analyze this distribution in more detail, and compare the distribution pattern for nucleosome cores from intact chromatin having different higher-order structures (from the 10 nm filament to the 30 nm fiber). At all levels of chromatin compaction, we observed the following. (1) The average periodicity in pyrimidine dimer yield is 10.3 bases. (2) The peak-to-peak spacing in this distribution is significantly different from 10.3 bases in the region covering three helix turns immediately 5' of the dyad axis. (3) There is a suppression of photoproduct formation in the region of the dyad axis, especially at position 84 from the 5' end. (4) The approximately 10 base ensembles have alternating peak intensities throughout core DNA. Furthermore, peak deconvolution analysis of the pyrimidine dimer pattern yielded a striking similarity in photoproduct yield for the different levels of chromatin compaction. Irradiation of isolated core DNA yields a much more random distribution of photoproducts, although a weak modulation pattern is observed (indicating that there is a non-random alignment of adjacent pyrimidines in our core DNA preparations). This pattern includes a depression in photoproduct yield near position 95, suggesting that the sequence in this region plays a role in nucleosome positioning. These results show that the u.v. photofootprint is a sensitive, diagnostic probe of core histone-DNA interactions in intact chromatin, and these interactions are not significantly altered by changes in the structural state of the chromatin fiber.