Replication-fork stalling and processing at a single psoralen interstrand crosslink in Xenopus egg extracts

Abstract

Interstrand crosslink (ICL)-inducing agents block the separation of the two DNA strands. They prevent transcription and replication and are used in clinics for the treatment of cancer and skin diseases. Here, we have introduced a single psoralen ICL at a specific site in plasmid DNA using a triplex-forming-oligonucleotide (TFO)-psoralen conjugate and studied its repair in Xenopus egg extracts that support nuclear assembly and replication of plasmid DNA. Replication forks arriving from either side stalled at the psoralen ICL. In contrast to previous observations with other ICL-inducing agents, the leading strands advanced up to the lesion without any prior pausing. Subsequently, incisions were introduced on one parental strand on both sides of the ICL. These incisions could be detected whether one or both forks reached the ICL. Using small molecule inhibitors, we found that the ATR-Chk1 pathway, but not the ATM-Chk2 pathway, stimulated both the incision step and the subsequent processing of the broken replication intermediates. Our results highlight both similarities and differences in fork stalling and repair induced by psoralen and by other ICL-forming agents.

Figure 1. TFO and purification of monomodified plasmid.

(A) Structure of the TFO linked in 3′ to 4, 5′, 8-trimethylpsoralen and in 5′ to biotin. (B) Localization of the TFO binding site and position of the psoralen ICL on pTUC plasmid. P and A are schematic representations of primers for q-PCR used in (D). P primers surround the psoralen ICL site and amplify a region of 113 nt, the A primers amplify an undamaged regions of 129 nt. (C) Analysis of the crosslinked plasmid after UV irradiation and before (input) or after (purified) DTT elution from a strepatvidin column. DNA was digested with BamHI + EcoRI and radioactively labelled before electrophoresis on a 10% polyacrylamide denaturing gel. Interpretative diagrams of the relevant molecular species are shown on lane sides. (D) Input and purified plasmids were subjected to q-PCR with primers P and A. Mean values of 3 q-PCRs for the input and 12 q-PCRs for the purified plasmid are presented. For calculation details see Material and Methods. (E) After purification pTUC-PSO was analyzed on a 0.8% agarose gel TBE 1x alongside with control plasmid and a molecular weight ladder.

Figure 2. Two-dimensional gel electrophoretic analysis of pTUC and pTUC-PSO plasmids replicating in Xenopus egg extracts.

Plasmids pTUC (A) and pTUC-PSO (B) were incubated 95 min in Xenopus egg extract in the continuous presence of [α-³²P]-dATP. Plasmids were purified, linearized by AflIII digestion and analyzed by 2D gel electrophoresis. Phosphorimager images of the dried gels and interpretative diagrams are shown. (C) Interpretation of the pTUC-PSO specific arcs. (D) Relative intensity of total signal, 1x spot and replication intermediates (RIs) from 3 independent experiments (Exp1, Exp2, Exp3).

Figure 3. Replication fork leading strands progress up to the psoralen crosslink.

(A) Sequence of the plasmid around the psoralen ICL with restriction sites used in B and C. Double arrows indicate the size of the replicated strands (lagging or leading) spanning the ICL site for the control or repaired plasmid. Single arrows indicate the progression of the leading strand to the ICL for the crosslinked plasmid. The strand size is shown above each product. (B) Mapping of leading strand progression for pTUC and pTUC-PSO after 65 min of incubation in Xenopus egg extracts in the continuous presence of [α-³²P]-dATP. Plasmids were digested by the indicated enzymes and analyzed on 10% polyacrylamide denaturing gel. Faint bands visible in lanes 3 and 4 likely result from a star activity of Sac II. The two 28 nt size indicators are at two different positions due to gel smiling. (C) Mapping of leading strand progression for pTUC and pTUC-PSO at 25, 35, 50, 65, 85, 95, 120 and 180 min. Plasmids were digested with EcoRI and BamHI and subjected to migration on a 10% polyacrylamide denaturing gel. 28 nt and 46 nt fragments correspond to the DNA size expected for the leading strand arriving at the psoralen from the EcoRI side or BamHI side, respectively.

Figure 4. Effect of ATR and Chk1 inhibition on processing of stalled replication forks at the psoralen ICL.

(A) pTUC-PSO was incubated for 95 min in a Xenopus egg extract in absence (NT; not treated) or presence of 5 mM caffeine or 10 µM UCN-01 and [α-³²P]-dATP and analyzed on 2D gels. Total signal and spots 1, 2 and 3 (as labeled on the interpretative diagram) were quantified in the presence of caffeine (B), UCN-01 (C), KU55933 (D) or C3742 (E). Results were normalized to the signals obtained with the non-treated plasmid. Exp1, Exp2, Exp3, Exp4, Exp5 and Exp6 are six experiments in six independently prepared egg extracts.

Figure 5. Model for psoralen ICL repair in Xenopus egg extract.

During replication of plasmid DNA containing a single psoralen ICL, replication forks stall at the ICL advancing up to 1 nt from the lesion. Coordinated incisions of one parental strand around the lesion can occur whether one (right) or both (left) forks have stalled, resulting in unhooking of the lesion. Next, translesion synthesis occurs across the unhooked psoralen adduct and other events implying HR and/or NER regenerate two repaired duplexes as proposed elsewhere . Both incision events and subsequent processing of the broken replication intermediates are stimulated by the ATR-Chk1 pathway.