Introduction

The dynamic organization of chromatin in the eukaryotic genome not only allows efficient packing of genomic information, but is also essential for the regulation of DNA metabolism. These dynamics can operate at the level of the basic chromatin subunit, a nucleosome, through the regulation of DNA accessibility by assembly and disassembly events (Kornberg and Lorch, 1992; Felsenfeld, 1996; Krude and Elgin, 1996; Wolffe and Pruss, 1996). Nucleosome disruption can be promoted by specific molecular machines including SWI/SNF, NURF, brahma and RSC (Cairns et al., 1996; Steger and Workman, 1996). Access to nucleosomal DNA may be facilitated by histone chaperones such as nucleoplasmin (Chen et al., 1994) as well as post-translational modification of histones including acetylation (Turner and O'Neill, 1995; Roth and Allis, 1996) and poly ADP-ribosylation (Althaus, 1992).

The resetting of nucleosomal structures after a disassembly event is equally important. In the case of nucleotide excision repair (NER), one of the major DNA repair pathways, nucleosomal organization is assumed to be transiently disrupted so that repair enzymes have access to lesions and can remove them from DNA (Smerdon, 1989). Unfolding during repair could involve chromatin remodelling mechanisms similar to those used for transcription (Friedberg, 1996). A proper nucleosomal organization must then be re-established after repair. Recently, we found that de novo chromatin assembly on DNA occurs concomitantly with UV damage-dependent repair synthesis in extracts from Xenopus eggs and human cells, suggesting a mechanistic coupling between the two processes. Human chromatin assembly factor 1 (CAF1) was necessary for this repair-associated chromatin formation (Gaillard et al., 1996). CAF1 was first identified as a nucleosome assembly factor associated with DNA replication in human cell-free systems (Stillman, 1986; Smith and Stillman, 1989; Krude and Knippers, 1993; Kaufman et al., 1995; Verreault et al., 1996).

The disruption and reassembly of chromatin during NER is likely to have some parallels with the corresponding sequence of events during DNA replication. However, the precise mechanisms coupling chromatin assembly with DNA repair may be significantly different from those used for replication. Previous studies on chromatin assembly at the replication fork have revealed that nucleosomes are deposited onto newly replicated DNA when a sufficient length of DNA has been synthesized (Annunziato, 1990; Gruss and Sogo, 1992; Kaufman and Botchan, 1994; Krude, 1995). In this case, the chromatin assembly machinery may function in a sequential, stoichiometric fashion. In contrast, during NER, a very different situation arises. Although the extent of disruption of nucleosomal organization at a site of repair has been estimated to reach up to 2 kb (Althaus, 1992), the amount of DNA synthesized is limited to a short patch of 30 nucleotides (nt) (Edenberg and Hanawalt, 1972; Hansson et al., 1989; Cleaver et al., 1991; Svoboda et al., 1993) and the mechanism of nucleosome reformation after repair has not been defined. It is critical to understand how a process of DNA synthesis limited to such a defined region can be sufficient to ensure the resetting of an extended nucleosomal array. A mechanism ensuring a proper regulation of assembly during NER is essential because of the importance of chromatin organization for DNA metabolism.

DNA molecules containing single, defined sites of damage have proven to be valuable tools for analysis of the NER process (Sancar, 1996; Wood, 1996). Their use in conjunction with a model system in which both NER and chromatin assembly can be followed in parallel represents a powerful way to examine chromatin assembly at a unique site of repair. Drosophila preblastoderm embryo extracts in which chromatin assembly is supported by the endogenous pool of histones (Becker and Wu, 1992) are used here for this purpose. We have defined conditions under which it is possible to observe selectively the chromatin assembly pathway coupled to efficient NER, enabling a precise analysis of nucleosome formation at a defined site of repair. The results support a mechanism in which NER initiates nucleosomal formation from the repair site, followed by a bidirectional propagation of chromatin assembly that leads to regular arrays of nucleosomes. Chromatin assembly is still effective when the repair synthesis step in the NER process is inhibited. This novel chromatin assembly mechanism linked to a site of NER may be significant in the response to genotoxic stress.

Results

Selection of a chromatin assembly pathway specifically coupled to repair in a Drosophila cell-free system

Extracts from preblastoderm Drosophila embryos have a high capacity to assemble chromatin in vitro (Becker and Wu, 1992). These extracts contain pools of histones and factors which facilitate the efficient assembly of nucleosomes on DNA (Foe and Alberts, 1983). Early embryos replicate their DNA at high rates (Blumenthal et al., 1974) and at this time of development are expected to rely on efficient repair of damaged DNA. We explored whether such Drosophila extracts were capable of simultaneous DNA repair and chromatin assembly. A simple assay, initially set up with Xenopus egg extracts, consists of incubating UV-irradiated circular plasmids with a cell-free extract in the presence of radiolabelled dNTPs and analysing both radiolabel incorporation and supercoiling of the purified DNA (Gaillard et al., 1996). The labelling represents DNA repair synthesis after incision of UV photoproducts (Wood et al., 1988) whereas the supercoiling is indicative of chromatin formation on the DNA (Germond et al., 1979).

We analysed the Drosophila extract in this way, using our Xenopus egg extract as a reference. Preferential incorporation of dNTPs in UV-treated plasmids compared with untreated DNA (Figure 1, top panel, compare lanes + and -) indicated that the Drosophila extract was competent for UV damage-dependent DNA synthesis. The influence of salt concentration was tested, as it is important for NER reactions in vitro (Wood et al., 1988). The contribution from the extracts (in equivalent KCl) was measured by conductimetry and the final concentration was obtained by adding KCl. Comparison of Drosophila and Xenopus extracts provided two sets of information (Figure 1, 'Labeled DNA'). Firstly, throughout the range of salt conditions tested, the damage-dependent radiolabel incorporation in UV-treated plasmids remained constant with Drosophila extract. Secondly, for all salt conditions tested in the Drosophila extract, a constant fraction of labelled DNA was supercoiled, suggesting its assembly into chromatin. Limited labelling in the control plasmid indicated the presence of a few nicks in the DNA substrates. In contrast, both the extent of supercoiling and repair synthesis were decreased at higher salt concentrations in the Xenopus extract.

When examining the total population of DNA molecules by ethidium bromide staining (Figure 1, 'Total DNA'), formation of supercoiled molecules was also detected, revealing the existence of a repair-independent chromatin assembly pathway in both Xenopus and Drosophila extracts. This pathway was salt-sensitive, as assessed on unirradiated molecules in which the labelled fraction was very limited (Figure 1, bottom panel, lanes 11, 13, 15, 17 and 19). A time-course analysis using Drosophila extract revealed that under high salt conditions only limited supercoiling was observed even after 6 h, as opposed to low salt conditions where significant levels of control DNA were supercoiled after 30 min (Figure 2A, bottom panel). Although the supercoiling assay is a convenient test for chromatin assembly (Germond et al., 1979) it was necessary to confirm that authentic chromatin assembly was taking place. Micrococcal nuclease (MNase) digestion analysis was used to verify that the supercoiled plasmids contained regularly spaced nucleosomes (Figure 2B). At low salt concentrations, nucleosomal ladders were observed for the control and the UV-irradiated plasmid DNA for both labelled and total DNA. In contrast, at higher salt concentration within the physiological range, defined nucleosomal ladders were detected only for the labelled plasmid DNA. This shows that elevated salt concentration suppresses repair-independent chromatin assembly by Drosophila embryo extracts. These conditions were used in subsequent experiments for the selective analysis of chromatin assembly tightly linked to DNA repair.

Efficient nucleotide excision repair in Drosophila embryo extracts

The DNA repair activity of Drosophila preblastoderm embryo extracts has not been reported before, and it was important to analyse further and quantify the repair properties. During a 2 h reaction, radiolabel incorporation increased with the UV-C dose applied to the plasmid DNA (Figure 3A, left). An estimate of the percentage of lesions repaired can be made by assuming that, as in other eukaryotes, UV-induced DNA synthesis is limited to short patches of 30 nt characteristic of NER. Up to 45% of lesions were repaired in a plasmid irradiated with 50 J/m² (Figure 3A), a dose that induces 1.5 pyrimidine dimers per pBS plasmid (Wood et al., 1995). Thus, the Drosophila cell-free system exhibits a high repair efficiency. The time-course of repair was followed for a plasmid irradiated at 500 J/m² (Figure 3A, right). At 15 min, 5% of the lesions were already repaired and up to 38% and 63% of lesions were repaired after 2 and 8 h, respectively. We also tested UV-irradiated DNA from which pyrimidine hydrates had been removed by treatment with Nth protein (endonuclease III) (Wood et al., 1995) and found similar results (data not shown), strongly suggesting that NER is responsible for most of the UV-induced DNA synthesis. This is consistent with a previous report of NER by extracts from a cultured embryonic Drosophila cell line (Shimamoto et al., 1995).

For a high-resolution analysis of repair at the site of damage, the Drosophila system was further analysed using DNA containing a single 1,3-intrastrand d(GpTpG)–cisplatin crosslink, a lesion that is avidly repaired by NER in human cell extracts (Moggs et al., 1996). In mammalian cells, dual incision of the damaged DNA strand proceeds via an open intermediate (Evans et al., 1997) with cleavage of the 3' and 5' single-strand/double-strand DNA junctions by the structure-specific endonucleases XPG and XPF-ERCC1 respectively (Sijbers et al., 1996). With Drosophila extract, oligonucleotides 24–34 nt long, predominated by 29- and 30-mers, were detected in an incision assay (Figure 3B, lanes 2–4). The size of the excised fragments is characteristic of authentic NER, but it is interesting that the pattern with Drosophila embryo extracts was different in detail from that observed using human cell extracts, where the predominant excised oligonucleotides are 26 nt long (Figure 3B, lane 5; Moggs et al., 1996). Species-specific differences in the size and shape of the repair proteins or their developmental modification may alter the positions of incisions. We also examined the extent of radiolabel incorporation at the lesion site and, as expected in a typical NER reaction, repair synthesis was limited to a short patch found in the 33, 38 and 68 bp restriction fragments around the site of the cisplatin adduct (Figure 3C). The efficiency of repair of the cisplatin DNA adduct was 40% after 2 h as measured by radiolabel incorporation into the 33 bp fragment containing the lesion site, consistent with the repair efficiency calculated above for a low dose of UV irradiation. The combination of this high repair efficiency and the ability to select a repair-linked chromatin assembly pathway provides a powerful system for investigating chromatin assembly linked to NER.

Nucleosome formation initiated at a target site for NER

The coupling mechanism between NER and chromatin assembly was investigated under high salt conditions using the DNA containing a single lesion. Formation of nucleosomal arrays was analysed by MNase digestion, including similarly constructed undamaged DNA as a control. As expected, significant radiolabel incorporation was detected only on the platinated DNA (Figure 4A, upper panel, compare lanes - and +). Repair was already associated with mononucleosomal DNA after 40 min. After 2 h the DNA that had undergone repair synthesis was assembled into regular nucleosomal arrays, and progressively larger nucleosomal DNA fragments appeared after 6 h. Analysis of the total population of DNA molecules by ethidium bromide staining (Figure 4A, lower panel) paralleled these observations, with nucleosomal ladders being detectable only for platinated DNA after 2 h. This is consistent with the high repair efficiency and selection of repair-linked chromatin assembly at high salt concentrations in this system. However, in contrast to the labelled DNA, a significant proportion of the total DNA formed a background smear, which is likely to represent the unrepaired fraction (60%). This excludes that the simple presence of a DNA lesion can promote chromatin assembly and rather favours a chromatin assembly mechanism linked to the repair process. Furthermore, proteins such as HMGs which might promote chromatin assembly by binding to DNA distortions (Travers et al., 1994) do not interact with a 1,3-intrastrand cisplatin–DNA crosslink (Pil and Lippard, 1992). For undamaged DNA, MNase-generated fragments migrated essentially at the position of mononucleosomal DNA after 6 h, reflecting a slow, non-specific assembly of bulk DNA into chromatin under these conditions.

We then verified that the labelled DNA organized into chromatin resulted from repair at the cisplatin lesion site (Figure 4B). Undamaged and platinated substrates were incubated in Drosophila embryo extracts for 45 min or 6 h, and chromatin was digested to mononucleosomes by extensive MNase treatment. DNA purified after protein extraction was then challenged with two restriction enzymes, HindIII and XhoI, each recognizing a unique cleavage site near the cisplatin adduct. Mononucleosomal DNA labelled by a specific NER reaction at the lesion site should be cleavable by at least one of the two enzymes. In contrast, labelled mononucleosomal DNA resulting from random non-specific DNA synthesis elsewhere on the molecule was expected to be resistant to cleavage by both enzymes (Figure 4B, upper panel). Labelled mononucleosomes from reactions performed with platinated DNA were sensitive to both enzymes, reflecting the presence of a nucleosome spanning the repaired site (Figure 4B, lower panel). These results demonstrate that a nucleosome is assembled at the repaired site and that subsequent assembly events lead to extended regular arrays of nucleosomes.

Bidirectional propagation of nucleosomal arrays from a target site for NER

To analyse how NER at a single site, limited to a short, 30 nt region, can lead to nucleosome formation, the quality of the chromatin was examined at various distances from the lesion site. The DNA blot from Figure 4A was probed with specific oligonucleotides (Figure 5A) complementary to either the lesion site (Pt site), to regions located 800 nt 5' and 3' to the lesion (-800, +800) and to a region opposite the cisplatin lesion, 3500 nt away (distal). Using the probe at the Pt site, nucleosomal ladders were observed similar to those in Figure 4A, for the platinated substrate after 2 h, with larger nucleosomal fragments appearing after 6 h (Figure 5B). However, probing the membrane with the 'distal' oligonucleotide gave only a smear of fragments on both the platinated and the undamaged DNA after a 2 h reaction (Figure 5B). This indicates that at a time where regular nucleosomal arrays were detectable at the repair site, little assembly had occurred 3500 nt from the lesion. After a 6 h reaction the signal detected with the distal probe did not correspond to a regular and extended array of nucleosomes.

Considering the 5'3' directionality of the repair synthesis step, we wanted to test whether the propagation of nucleosome assembly initiated at the repair site had a corresponding unidirectionality. The membrane was probed successively with the '+800' and '-800' probes (Figure 5A). Nucleosomal ladders were detected with both probes for the platinated substrate only (Figure 5B). Interestingly, the formation of nucleosome arrays proceeded with comparable kinetics 5' and 3' to the lesion, with a nucleosomal ladder detectable after 2 h and longer arrays detectable after 6 h (as observed with the 'Pt site' probe, Figure 5B). Densitometric scanning was carried out to examine more closely the profiles of the nucleosomal ladders at the different sites (Figure 6A). Profiles obtained for the labelled fragments had a peak distribution similar to the profile with the 'Pt site' probe. In addition, the background smear representing the unrepaired fraction of single lesion DNA was also detected by the probe. Profiles obtained for the +800 or -800 probes had maximum peaks for the low molecular weight nucleosomal fragments, consistent with a less extensive assembly than at the site of the lesion. The smeared profile obtained with the distal probe confirmed that only poor assembly occurred far away (3500 bp) from the lesion site. Interpretation of these data enabled us to propose the scheme in Figure 6B, where chromatin assembly initiates from the NER repair site, leading to the bidirectional propagation of regular nucleosome arrays over a limited distance around the DNA molecule.

DNA synthesis is not required for chromatin assembly from a target site for NER

The unidirectionality (5'3') of the repair synthesis step in the NER process contrasts with the bidirectional propagation of chromatin assembly. This suggests that the DNA repair synthesis step itself might not be essential for the coupling of NER and chromatin assembly. To test this hypothesis, the formation of nucleosomal arrays was analysed on undamaged and platinated substrates, in the presence of aphidicolin, a potent inhibitor of DNA synthesis. Aphidicolin specifically inhibits DNA repair synthesis without affecting any of the preceding steps of the NER process, including recognition of the lesion and dual incision of the damaged strand (Coverley et al., 1992; Calsou and Salles, 1994; Moggs et al., 1996). We confirmed that dual incisions were formed by Drosophila embryo extract in the presence of aphidicolin (data not shown). When aphidicolin was added to the reaction, all label incorporation was eliminated, consistent with an efficient inhibition of DNA synthesis (Figure 7, compare lanes - and +, Labeled DNA). Despite this inhibition, a nucleosomal ladder was detected on the platinated substrate with an oligonucleotide probe (Figure 7). This ladder was comparable with that observed on the platinated substrate without aphidicolin (Figure 7, compare - and +). In contrast, no ladder was detected on the undamaged substrate with or without aphidicolin. These results demonstrate that the complete repair synthesis step is not required for chromatin assembly during NER. Earlier steps in the NER process involving either a specific DNA structure or a component of the NER machinery could serve as a signal of recognition and recruitment.

Discussion

Chromatin assembly and NER in Drosophila embryo extracts

Drosophila preblastoderm embryo extracts are shown here to perform NER efficiently and to promote a chromatin assembly pathway linked to a specific site of DNA repair. Previous studies using either preblastoderm (Becker and Wu, 1992) or postblastoderm (Kamakaka et al., 1993) embryo extracts have identified a number of chromatin assembly factors. These include nucleosome assembly factor-1 (NAP-1; Ishimi et al., 1984; Ito et al., 1996), decondensation factor 31 (DF31; Crevel and Cotterill, 1995) and high mobility group protein (HMGP; Ner and Travers, 1994). Additional chromatin factors have been found that can mediate disrupting events associated with transcriptional activation, such as the nucleosome rearrangement factor (NURF; Tsukiyama and Wu, 1995), or that ensure nucleosome mobility (Varga-Weisz et al., 1995). These factors may participate in our specific repair-coupled chromatin assembly reaction. However, the only factor shown to be associated with a specific DNA transaction is CAF1. Human CAF1 was initially identified through the analysis of an SV40 replication-coupled assembly pathway (Stillman, 1986). Analogous activities have also been found in yeast, Xenopus and Drosophila (Gaillard et al., 1996; Kamakaka et al., 1996; Enomoto et al., 1997; Kaufman et al., 1997). In addition, human CAF1 is required for chromatin assembly coupled to NER in a human cell-free system (Gaillard et al., 1996). Notably, deletion of any of the three genes encoding CAF1 subunits in Saccharomyces cerevisiae increases sensitivity to UV irradiation, without sensitizing cells to -rays (Kaufman et al., 1997). This is consistent with an involvement of CAF1 in NER in vivo. It is not yet known whether CAF1 directly interacts with components of either DNA replication or repair.

The Drosophila system defined here provides a new powerful approach to follow the biochemical mechanism of chromatin assembly coupled to NER. It also offers the possibility of developmental and genetic investigations, taking advantage of known UV-sensitive mutants (Dusenbery and Smith, 1996). This can be expected to open up new avenues for investigation of both chromatin assembly and NER and the interrelations between these mechanisms.

Initiation and propagation of nucleosome formation from a defined site for NER

The use of a uniquely placed lesion facilitates the study of the chromatin assembly mechanism by focusing on a single site of NER. At early times in the reaction, mononucleosomes were formed at the repair site and contained DNA labelled during repair synthesis. The appearance of labelled polynucleosomal DNA at later times indicates that repair at a single site subsequently leads to formation of extended nucleosomal arrays (Figure 4). This demonstrates that chromatin assembly specifically operates from the repair site to generate arrays of nucleosomes. A mechanism of propagation for the assembly of nucleosomes appears particularly appropriate for the resetting of relatively large destabilized structures. Propagation may only be possible over a limited region of the genome because incomplete assembly was observed 3.5 kb away from the lesion (Figures 5B and 6). It will now be important to examine more closely the extent of nucleosomal disruption for various damaging agents in vivo.

Changes in the nuclease sensitivity of DNA during repair of UV lesions in mammalian cells are known to occur (Smerdon, 1989). DNA repair patches in the bulk of chromatin show an initial, transient sensitivity to nucleases. This may be accounted for if chromatin unfolds during NER (Friedberg et al., 1995). The mechanism by which a native nucleosomal organization is re-established after NER is not well understood. Repair patch ligation precedes the loss of the transient nuclease sensitivity associated with unfolded or altered nucleosomal structures on DNA undergoing NER (Hunting et al., 1985; Smerdon, 1986). Furthermore, most repair patches associated with isolated nucleosomes were observed to contain ligated repair patches. These results suggest that both DNA synthesis and ligation occur before the formation of a complete nucleosome structure at the repair site. However, since disruption may not be limited to a single nucleosome during NER it is also important to examine nucleosomal organization at various distances from the repair site.

We conclude that nucleosomal arrays are formed by a bidirectional propagation mechanism from the repair site, even when repair synthesis is inhibited. In the presence of aphidicolin, the region of repair was particularly sensitive to MNase (data not shown), reflecting either the presence of a sensitive intermediate in NER such as an unligated gap, or the absence of a completely formed nucleosome. Under conditions permitting the complete repair reaction to occur, labelled mononucleosomes were found as soon as synthesis was detected (Figure 4A, top panel). Thus, although DNA synthesis may be needed in order to form a stable canonical mononucleosome at the repair site in agreement with previous reports (Smerdon, 1989), it does not appear to be necessary for entry of a presumed chromatin assembly machinery during NER. Instead, we propose that the link between NER and extensive de novo chromatin assembly is at an earlier step in the NER process.

Comparison of repair- and replication-coupled chromatin assembly mechanisms

Although DNA synthesis is a common event in both DNA repair and DNA replication, in each of these processes it occurs in a distinct manner and it may not be surprising that the common link involves a reaction step distinct from DNA synthesis. The incision/excision step is of particular interest since it seems possible that the simple presence of nicks can promote repair-coupled chromatin assembly (Figure 2A and B, high KCl, non-irradiated labelled DNA). Identification of proteins specifically interacting with these intermediates may reveal a factor that recruits a presumed chromatin assembly machinery.

The assembly mechanism presented here implies that a single hit initiates nucleosome assembly from one point and then further assembly proceeds in a concerted manner. Two situations can be considered. In the first scenario, the chromatin assembly machinery is recruited at the repair site and remains there to ensure that subsequent nucleosomes form at the same site, with the newly formed nucleosome being pushed toward either side to leave the repair site free for the formation of the next one. Thus, older nucleosomes would be further away from the repair site, and activities ensuring nucleosome mobility (Varga-Weisz et al., 1995) would be critical in this process. A second possibility is that the chromatin assembly machinery recruits the first nucleosome over the repaired site, followed by recruitment events on both sides.

Physiological significance of nucleosome spreading

The importance of the nucleosome spreading phenomenon should be emphasized, as it allows the generation of regular arrays of nucleosomes. Although our experiments were carried out on naked DNA templates as simple models to follow an assembly process, they may also represent an open chromatin organization that may be encountered over limited regions of the genome undergoing active transcription, repair and other DNA transactions. In response to genotoxic stress, a mechanism of assembly propagating from a single site of repair in these regions would be useful to establish rapidly a nucleosomal array that could repress new DNA transactions. Such a mechanism could be used as a rapid response to genotoxic stress which arrests DNA metabolism, thus helping to establish cell cycle checkpoints.

Materials and methods

DNA templates

UV-irradiated pBluescript KS+ (pBS) plasmid DNA was prepared using a 254 nm (peak) germicidal UV lamp (Philips). The UV flux rate was measured with a Latarjet dosimeter (France). A flux of 100 J/m² induces 1 pyrimidine dimer photoproduct in 1000 bp [0.75 cyclobutane pyrimidine dimer and 0.25 (6-4) photoproduct]. In control experiments we treated UV-irradiated plasmid DNA with Escherichia coli Nth (endonuclease III) protein to cleave DNA at sites of pyrimidine hydrates (Wood et al., 1995). The single lesion substrate (Pt-GTGx), containing a 1,3-intrastrand d(GpTpG)–cisplatin crosslink, and the corresponding control substrate (Con-GTGx), were constructed as described (Moggs et al., 1996).

Preparation of Xenopus egg, Drosophila embryo and human cell extracts

Xenopus egg high-speed extracts were prepared as described (Almouzni et al., 1991). Drosophila preblastoderm embryo extracts were obtained as described in Becker et al. (1994). Protein concentration, determined using a Bradford protein assay kit (Bio-Rad), was typically 40 mg/ml for both types of extract. The human whole cell extract used in Figure 3B was derived from SW48 cells.

Repair and supercoiling assay

Xenopus egg extracts and Drosophila embryo extracts were used in 25 l reactions at 23°C containing 400 g protein extract, 300 ng of control or UV-C irradiated pBS plasmid, 5 mM MgCl₂, 40 mM HEPES–KOH, pH 7.8, 0.5 mM DTT, 4 mM ATP, 5 Ci [-³²P]dCTP (3000 Ci/mmol), 40 mM phosphocreatine, 2 g creatine phosphokinase (Type I, Sigma). The KCl equivalent in each protein extract was measured using a conductance meter (Hanna Instrument). Final equivalent KCl concentrations were obtained by the addition of KCl. Reactions were terminated by the addition of EDTA to 15 mM, SDS to 0.35% and RNase A to 200 g/ml. After 30 min at 37°C, proteinase K was added to 1 mg/ml and the reaction further incubated for 30 min before phenol/chloroform extraction. DNA was precipitated with 2.5 M ammonium acetate and 2 volumes of ethanol, and subjected directly to electrophoresis in a 1.2% agarose gel in TAE buffer. The total DNA loaded on the gel was visualized after ethidium bromide staining under near-UV transillumination with a CCD camera linked to an image analyser unit (Vilbert-Lourmat, France). The gel was then dried and exposed to an X-ray film to visualize the migration pattern of the plasmids labelled during the repair assay.

Quantitative analysis of repair synthesis during NER

Reactions were carried out as described above with 400 g of Drosophila embryo extract and 300 ng of either non-treated pBS, UV-C irradiated pBS, Pt-GTGx or Con-GTGx DNA substrates. pBS plasmids were linearized with BamHI before electrophoresis in a 1% agarose gel. The Pt-GTGx and Con-GTGx substrates were digested by BstNI, and the digestion products were resolved in a non-denaturing 12% polyacrylamide gel. For quantitation, the bands corresponding to the DNA substrates were excised from the dried gel, and the incorporated radioactivity was directly measured by Cerenkov counting. In the case of the single lesion substrate only the 33 bp restriction fragment was excised. The incorporation of dCMP during the repair synthesis reaction was calculated taking into account the specific activity of the radiolabel and the background incorporation in the control DNA substrate. The dNTP contribution of the Drosophila embryo extract was determined to be 60 M by isotopic dilution as described (Blow and Laskey, 1986). The repair efficiency was estimated based on the number of lesions per plasmid assuming a repair patch size of 30 nt. In the case of the single lesion substrate the number of fmol incorporated was normalized to the base composition of the damaged DNA strand.

Analysis of the dual incisions formed during NER

For dual incision reactions, 300 ng Pt-GTGx or Con-GTGx DNA substrates were incubated with Drosophila embryo extract at 23°C under the same conditions as the repair synthesis assay except that radiolabelled dNTPs were omitted from the reaction buffer. The analysis of dual and uncoupled 3' incisions was performed as described previously (Sijbers et al., 1996). Briefly, DNA purified from incision reactions was cleaved with HindIII and XhoI before separation in a denaturing 12% polyacrylamide gel. DNA was then transferred to a nylon membrane (Hybond-N+, Amersham) by capillary transfer for 90 min. Fixed membranes were hybridized with a radiolabelled 27-mer oligonucleotide complementary to the sequence of the damaged DNA strand flanking the cisplatin–DNA adduct. The sizes of platinated oligonucleotides were determined by comparison with a 5'-phosphorylated 24-mer oligonucleotide 5'-TCTTCTTCTGTGCACTCTTCTTCT-3' having a single 1,3-intrastrand d(GpTpG)–cisplatin crosslink bridging bases 10 and 12.

Analysis of micrococcal nuclease digestion products

Reactions were scaled up to a final volume of 75 l but using 5-fold less specific activity. When aphidicolin was added at a final concentration of 400 g/ml, reactions contained 6% dimethyl sulfoxide (v/v). Digestion reactions were carried out with 45 units of micrococcal nuclease (MNase) at 23°C and aliquots (90 l) were removed at various times and processed as described (Becker et al., 1994). MNase digestion products were separated in a 1.3% agarose gel in TBE buffer. Total DNA was visualized after ethidium bromide staining of the gel as described above. For reactions carried out with pBS substrates, the gel was dried and exposed to an X-ray film to visualize the DNA labelled during the repair assay. For reactions carried out with the single-lesion substrates, the gel was processed for vacuum transfer of the DNA onto a nylon membrane (Hybond-N+, Amersham) and exposed to an X-ray film before hybridization with radiolabelled probes. The autoradiograph revealed, as for a dried gel, the migration pattern of the DNA labelled during the repair assay. Oligonucleotides (Eurogentec) specific for the damaged site ('Pt site' 5'-GAAGAGTGCACAGAAGAAGAGGCCTGG-3'), for a region diametrically opposed to the damaged site ('distal' 5'-CACCAGTAGCACCATTACCATTAGCA-3'), or specific for regions 800 bp away from either side of the lesion ('+800' 5'-CTCAGAGCATAAAGCTAAATCGGT-3'; '-800' 5'-TTAATTTGCGTGATGGACAGACTCT-3'), were end-labelled with [-³²P]ATP (>5000 Ci/mmol) using T4 DNA polynucleotide kinase (New England Biolabs). Radiolabelled probes were then used successively on the same membrane to analyse the pattern of nucleosome packaging at various locations on the single-lesion substrate. Southern blot analysis was performed as described (Becker et al., 1994). After hybridization with each probe the membrane was exposed to X-ray film before being stripped. The membrane was then probed with a different oligonucleotide after checking that no significant residual label remained. Because of the high specific activity of the probes (10⁷ c.p.m./pmol of oligo) the signal due to repair synthesis can be neglected.

Acknowledgements

We are grateful to John Diffley, Manuel Buchwald and Ethel Moustacchi for critical reading of the manuscript. We thank Sylvie Milley for help and our colleagues for numerous discussions. This work was supported by the ARC, la Ligue National Contre le Cancer, la FRM, an ATIPE from the CNRS, the EEC and the Imperial Cancer Research Fund. P.-H.L.G. is an MENESR fellow. J.G.M. is supported by a long-term EMBO fellowship.

References

Almouzni G, Méchali M and Wolffe AP (1991) Transcription complex disruption caused by a transition in chromatin structure. Mol Cell Biol, 11, 655–665. | PubMed | ChemPort |

Althaus FR (1992) Poly ADP-ribosylation – a histone shuttle mechanism in DNA excision repair. J Cell Sci, 102, 663–670. | PubMed | ChemPort |

Annunziato AT (1990) Chromatin Replication and Nucleosome Assembly, Strauss-Wilson, Vol. 2, pp. 687–712. | ChemPort |

Becker PB and Wu C (1992) Cell-free system for assembly of transcriptionally repressed chromatin from Drosophila embryos. Mol Cell Biol, 12, 2241–2249. | PubMed | ISI | ChemPort |

Becker PB, Tsukiyama T and Wu C (1994) Chromatin assembly extracts from Drosophila embryos. Methods Cell Biol, 44, 207–223. | PubMed | ISI | ChemPort |

Blow JJ and Laskey RA (1986) Initiation of DNA replication in nuclei and purified DNA by cell-free extracts of Xenopus eggs. Cell, 47, 577–587. | PubMed | ISI | ChemPort |

Blumenthal AB, Kriegstein HJ and Hogness DS (1974) The units of DNA replication in Drosophila melanogaster chromosomes. Cold Spring Harbor Symp Quant Biol, 38, 205–223. | PubMed | ChemPort |

Cairns BR et al. (1996) RSC, an essential abundant chromatin-remodelling complex. Cell, 87, 1249–1260. | PubMed | ISI | ChemPort |

Calsou P and Salles B (1994) Properties of damage-dependent DNA incision by nucleotide excision repair in human cell-free extracts. Nucleic Acids Res, 22, 4937–4942. | PubMed | ChemPort |

Chen H, Li B and Workman JL (1994) A histone-binding protein, nucleoplasmin, stimulates transcription factor binding to nucleosomes and factor-induced nucleosome disassembly. EMBO J, 13, 380–390. | PubMed | ChemPort |

Cleaver JE, Jen J, Charles WC and Mitchell DL (1991) Cyclobutane dimers and (6-4) photoproducts in human cells are mended with the same patch sizes. Photochem Photobiol, 54, 393–402. | PubMed | ChemPort |

Coverley D, Kenny MK, Lane DP and Wood RD (1992) A role for the human single-stranded DNA binding protein HSSB/RPA in an early stage of nucleotide excision repair. Nucleic Acids Res, 20, 3873–3880. | PubMed | ISI | ChemPort |

Crevel G and Cotterill S (1995) DF 31, a sperm decondensation factor from Drosophila melanogaster: purification and characterization. EMBO J, 14, 1711–1717. | PubMed | ISI | ChemPort |

Dusenbery RL and Smith PD (1996) Cellular responses to DNA damage in Drosophila melanogaster. Mutat Res, 364, 133–145. | Article | PubMed | ChemPort |

Edenberg H and Hanawalt P (1972) Size of repair patches in the DNA of ultraviolet-irradiated HeLa cells. Biochim Biophys Acta, 272, 361–372. | Article | PubMed | ChemPort |

Enomoto S, McCune-Zierath P, Geraminejad M, Sanders M and Berman J (1997) Rfl2, a subunit of yeast chromatin assembly factor-I, is required for telomeric chromatin function in vivo. Genes Dev, 11, 358–363. | PubMed | ISI | ChemPort |

Evans E, Fellows J, Coffer A and Wood RD (1997) Open complex formation around a lesion during nucleotide excision repair provides a structure for cleavage by human XPG protein. EMBO J, 16, 625–638. | Article | PubMed | ChemPort |

Felsenfeld G (1996) Chromatin unfolds. Cell, 86, 13–19. | PubMed | ISI | ChemPort |

Foe VE and Alberts BM (1983) Studies of nuclear and cytoplasmic behaviour during the five mitotic cycles that precede gastrulation in Drosophila embryogenesis. J Cell Sci, 61, 31–70. | PubMed | ISI | ChemPort |

Friedberg EC (1996) Cockayne syndrome – a primary defect in DNA repair, transcription, both or neither. BioEssays, 18, 731–738. | PubMed | ISI | ChemPort |

Friedberg EC, Walker GC and Siede W (1995) DNA Repair and Mutagenesis, ASM Press, Washington, DC, pp. 291–294.

Gaillard P-HL, Martini EMD, Kaufman PD, Stillman B, Moustacchi E and Almouzni G (1996) Chromatin assembly coupled to DNA repair – a new role for chromatin assembly factor-I. Cell, 86, 887–896. | PubMed | ISI | ChemPort |

Germond JE, Rouvière-Yaniv J, Yaniv M and Brutlag D (1979) Nicking-closing enzyme assembles nucleosome-like structures in vitro. Proc Natl Acad Sci USA, 76, 3779–3783. | PubMed | ChemPort |

Gruss C and Sogo JM (1992) Chromatin replication. BioEssays, 14, 1–8. | PubMed | ChemPort |

Hansson J, Munn M, Rupp WD, Kahn R and Wood RD (1989) Localization of DNA repair synthesis by human cell extracts to a short region at the site of a lesion. J Biol Chem, 264, 21788–21792. | PubMed | ChemPort |

Hunting DJ, Dresler SL and Lieberman MW (1985) Multiple conformational states of repair in chromatin during DNA excision repair. Biochemistry, 24, 3219–3225. | PubMed | ChemPort |

Ishimi Y, Hirosumi J, Sato W, Sugasawa K, Yokota S, Hanaoka F and Yamada M (1984) Purification and initial characterization of a protein which facilitates assembly of nucleosome-like structure from mammalian cells. Eur J Biochem, 142, 431–439. | PubMed | ISI | ChemPort |

Ito T, Bulger M, Kobayashi R and Kadonaga JT (1996) Drosophila NAP-1 is a core histone chaperone that functions in ATP-facilitated assembly of regularly spaced nucleosomal arrays. Mol Cell Biol, 16, 3112–3124. | PubMed | ISI | ChemPort |

Kamakaka RT, Bulger M and Kadonaga JT (1993) Potentiation of RNA polymerase II transcription by Gal4-VP16 during but not after DNA replication and chromatin assembly. Genes Dev, 7, 1779–1795. | PubMed | ISI | ChemPort |

Kamakaka RT, Bulger M, Kaufman PD, Stillman B and Kadonaga JT (1996) Postreplicative chromatin assembly by Drosophila and human chromatin assembly factor I. Mol Cell Biol, 16, 810–817. | PubMed | ChemPort |

Kaufman PD and Botchan MR (1994) Assembly of nucleosomes: do multiple assembly factors mean multiple mechanisms? Curr Opin Genet Dev, 4, 229–235. | PubMed | ChemPort |

Kaufman PD, Kobayashi R, Kessler N and Stillman B (1995) The p150 and p60 subunits of chromatin assembly factor I – a molecular link between newly synthesized histones and DNA replication. Cell, 81, 1105–1114. | PubMed | ISI | ChemPort |

Kaufman PD, Kobayashi R and Stillman B (1997) Ultraviolet radiation sensitivity and reduction of telomeric silencing in Saccharomyces cerevisiae cells lacking chromatin assembly factor I. Genes Dev, 11, 345–357. | Article | PubMed | ISI | ChemPort |

Kornberg R and Lorch Y (1992) Chromatin structure and transcription. Annu Rev Cell Biol, 8, 563–587. | Article | PubMed | ISI | ChemPort |

Krude T (1995) Nucleosome assembly during DNA replication. Curr Biol, 5, 1232–1234. | PubMed | ChemPort |

Krude T and Elgin SCR (1996) Chromatin: pushing nucleosomes around. Curr Biol, 6, 511–515. | PubMed | ChemPort |

Krude T and Knippers R (1993) Nucleosome assembly during complementary DNA strand synthesis in extracts from mammalian cells. J Biol Chem, 268, 14432–14442. | PubMed | ChemPort |

Moggs JG, Yarema KJ, Essigmann JM and Wood RD (1996) Analysis of incision sites produced by human cell extracts and purified proteins during nucleotide excision repair of a 1,3-intrastrand d(GpTpG)–cisplatin adduct. J Biol Chem, 271, 7177–7186. | Article | PubMed | ISI | ChemPort |

Ner SS and Travers AA (1994) HMG-D, the Drosophila melanogaster homolog of HMG1 protein, is associated with early embryonic chromatin in the absence of histone H1. EMBO J, 13, 1817–1822. | PubMed | ChemPort |

Pil PM and Lippard SJ (1992) Specific binding of chromosomal protein-HMG1 to DNA damaged by the anticancer drug cisplatin. Science, 256, 234–237. | PubMed | ISI | ChemPort |

Roth SY and Allis CD (1996) Histone acetylation and chromatin assembly: a single escort, multiple dances? Cell, 87, 5–8. | PubMed | ISI | ChemPort |

Sancar A (1996) DNA excision repair. Annu Rev Biochem, 65, 43–81. | Article | PubMed | ISI | ChemPort |

Shimamoto T et al. (1995) Expression and functional analyses of the DXPA gene, the Drosophila homolog of the human excision repair gene. XPA J Biol Chem, 270, 22452–22459. | Article | ChemPort |

Sijbers AM et al. (1996) Xeroderma pigmentosum group F caused by a defect in a structure-specific DNA repair endonuclease. Cell, 86, 811–822. | PubMed | ChemPort |

Smerdon MJ (1986) Completion of excision repair in human cells: relationship between ligation and nucleosome formation. J Biol Chem, 261, 244–252. | PubMed | ChemPort |

Smerdon MJ (1989) DNA excision repair at the nucleosome level of chromatin. In Lambert,M.W. and Laval,J. (eds), DNA Repair Mechanisms and their Biological Implications in Mammalian Cells. Plenum Press, New York, pp. 271–294. | ChemPort |

Smith S and Stillman B (1989) Purification and characterization of CAF-I, a human cell factor required for chromatin assembly during DNA replication in vitro. Cell, 58, 15–25. | PubMed | ISI | ChemPort |

Steger DJ and Workman JL (1996) Remodeling chromatin structures for transcription: what happens to the histones? BioEssays, 18, 875–884. | PubMed | ChemPort |

Stillman B (1986) Chromatin assembly during SV40 DNA replication in vitro. Cell, 45, 555–565. | PubMed | ISI | ChemPort |

Svoboda DL, Taylor JS, Hearst JE and Sancar A (1993) DNA repair by eukaryotic nucleotide excision nuclease: removal of thymine dimer and psoralen monoadduct by HeLa cell-free extract and of thymine dimer by Xenopus laevis oocytes. J Biol Chem, 268, 1931–1936. | PubMed | ChemPort |

Travers AA, Ner SS and Churchill MEA (1994) DNA chaperones: a solution to a persistence problem? Cell, 77, 167–169. | PubMed | ISI | ChemPort |

Tsukiyama T and Wu C (1995) Purification and properties of an ATP-dependent nucleosome remodeling factor. Cell, 83, 1011–1020. | PubMed | ISI | ChemPort |

Turner BM and O'Neill LP (1995) Histone acetylation in chromatin and chromosomes. Semin Cell Biol, 6, 229–236. | Article | PubMed | ISI | ChemPort |

Varga-Weisz PD, Blank TA and Becker PB (1995) Energy-dependent chromatin accessibility and nucleosome mobility in a cell-free system. EMBO J, 14, 2209–2216. | PubMed | ChemPort |

Verreault A, Kaufman PD, Kobayashi R and Stillman B (1996) Nucleosome assembly by a complex of CAF-1 and acetylated histones H3/H4. Cell, 87, 95–104. | PubMed | ISI | ChemPort |

Wolffe AP and Pruss D (1996) Targeting chromatin disruption: transcription regulators that acetylate histones. Cell, 84, 817–819. | PubMed | ISI | ChemPort |

Wood RD (1996) DNA repair in eukaryotes. Annu Rev Biochem, 65, 135–167. | Article | PubMed | ISI | ChemPort |

Wood RD, Robins P and Lindahl T (1988) Complementation of the Xeroderma pigmentosum DNA repair defect in cell-free extracts. Cell, 53, 97–106. | PubMed | ChemPort |

Wood RD, Biggerstaff M and Shivji MKK (1995) Detection and measurement of nucleotide excision repair synthesis by mammalian cell extracts in vitro. Methods: A Companion to Methods in Enzymology, 7, 163–175. | Article | ChemPort |