SUMMARY MRE11 within the MRE11-RAD50-NBS1 (MRN) complex acts in DNA double-strand break repair (DSBR), detection and signaling; yet, how its endo- and exonuclease activities regulate DSB repair by non-homologous end-joining (NHEJ) versus homologous recombination (HR) remains enigmatic. Here we employed structure-based design with a focused chemical library to discover specific MRE11 endo- or exonuclease inhibitors. With these inhibitors we examined repair pathway choice at DSBs generated in G2 following radiation exposure. Whilst endo- or exonuclease inhibition impairs radiation-induced RPA chromatin binding, suggesting diminished resection, the inhibitors surprisingly direct different repair outcomes. Endonuclease inhibition promotes NHEJ in lieu of HR, whilst exonuclease inhibition confers a repair defect. Collectively, the results describe nuclease-specific MRE11 inhibitors, define distinct nuclease roles in DSB repair, and support a mechanism whereby MRE11 endonuclease initiates resection, thereby licensing HR followed by MRE11 exo and EXO1/BLM bidirectional resection towards and away from the DNA end, which commits to HR.

No abstract

Nucleotide excision repair removes damages from the DNA by incising the damaged strand on the 3 and 5 sides of the lesion. In Escherichia coli, the two incisions are made by the UvrC protein, which consists of two functional halves. The N-terminal half contains the catalytic site for 3 incision and the C-terminal half contains the residues involved in 5 incision. The genome of E. coli contains an SOS-inducible gene (ydjQ) encoding a protein that is homologous to the N-terminal half of UvrC. In this paper we show that this protein, which we refer to as Cho (UvrC homologue), can incise the DNA at the 3 side of a lesion during nucleotide excision repair. The incision site of Cho is located 4 nt further away from the damage compared with the 3 incision site of UvrC. Cho and UvrC bind to different domains of UvrB, which is probably the reason of the shift in incision position. Some damaged substrates that are poorly incised by UvrC are very efficiently incised by Cho. We propose that E. coli uses Cho for repair of such damages in vivo. Initially, most of the lesions in the cell will be repaired by the action of UvrC alone. Remaining damages, that for structural reasons obstruct the 3 incision by UvrC, will be repaired by the combined action of Cho (for 3 incision) and UvrC (for 5 incision). The UvrC protein is the endonuclease responsible for incisions at the 3Ј and 5Ј sides of a damaged site during nucleotide excision repair in Escherichia coli (1, 2). The protein binds to the UvrB-DNA preincision complex, which is formed by stable binding of the UvrB protein to a damaged site, initially as part of the UvrA 2 B complex, but from which the UvrA subunits have subsequently dissociated (for reviews see refs. 3 and 4). The two incisions occur in a defined order, first at the 3Ј side and then at the 5Ј side of the damage. The UvrC protein consists of two functional parts. The N-terminal half contains the catalytic domain for 3Ј incision (2), which is homologous to the catalytic domain of the GIY-YIG family of intron-encoded homing endonucleases ( Fig. 1; ref. 5). In addition, the N-terminal half of UvrC contains a region that interacts with a homologous domain of the UvrB protein in the UvrBC-DNA incision complex (6, 7). Mutational analyses have shown that interaction by means of these homologous domains is needed for the 3Ј incision but not for the 5Ј incision (8, 9). A truncated UvrC protein lacking the N-terminal half is capable of 5Ј incision on a substrate that is prenicked at the 3Ј incision position (2), showing that the C-terminal half of UvrC contains all of the elements for the second incision event.The complete genome sequence of E. coli (10) revealed the presence of a gene, ydjQ, encoding a protein of 295 aa (molecular mass, 33.7 kDa) that shares significant homology with the N-terminal half of UvrC (Fig. 1). With Northern analysis it was shown that ydjQ is part of the SOS regulon, controlled by the LexA repressor (11). Interestingly the uvrA and uvrB genes are also under control of the SOS response, whereas u...

Essential genome transactions, such as homologous recombination, are achieved by concerted and dynamic interactions of multiple protein components with DNA. Which proteins do what and how, will be reflected in their relative arrangements. However, obtaining high-resolution structural information on the variable arrangements of these complex assemblies is a challenge. Here we demonstrate the versatility of a combined total internal reflection fluorescence and scanning force microscope (TIRF-SFM) to pinpoint fluorescently labeled human homologous recombination protein RAD54 interacting with presynaptic (ssDNA) and postsynaptic (dsDNA) human recombinase RAD51 nucleoprotein filaments. Labeled proteins were localized by superresolution imaging on complex structures in the SFM image with high spatial accuracy. We observed some RAD54 at RAD51 filament ends, as expected. More commonly, RAD54 interspersed along RAD51-DNA filaments. RAD54 promotes RAD51-mediated DNA strand exchange and has been described to both stabilize and destabilize RAD51-DNA filaments. The different architectural arrangements we observe for RAD54 with RAD51-DNA filaments may reflect the diverse roles of this protein in homologous recombination.DNA break repair | genome stability | DNA-protein interaction | image registration | single-molecule microscopy T o understand how proteins cooperate to perform elaborate genome transactions, we need to know how they are arranged relative to each other in functional complexes. Determining the structure of such complex and often variable assemblies is a challenge. The scanning force microscope (SFM) is ideal for visualizing DNA-protein assemblies under native conditions at a resolution limited by the radius of curvature of the scanning tip (usually 5-10 nm) (1-3). This method is limited because identification of specific proteins in heterogeneous assemblies depends on distinct structural features, whereas most proteins have a similar globular shape. Molecular recognition in high-resolution SFM images can be achieved by combining SFM with a fluorescence microscope capable of single-fluorescence detection (4-6). By labeling individual proteins or single DNA molecules, it is possible to distinguish the position of different components in a highly resolved complex.The multiprotein complexes of homologous recombination (HR) are an ideal example to establish methods for hybrid microscopy that can be validated on the basis of known structures and used to describe unknown arrangements. HR is a mesoscale DNA rearrangement process achieved by the coordinated action of several proteins and used for DNA repair (7,8). In the core reaction of HR, human recombinase RAD51-DNA filaments search and invade homologous undamaged dsDNA. After strand exchange between the invading ssDNA and its complement, the homologous strand of the duplex is used as template for repair synthesis. Genetic and biochemical evidence show that several mediator proteins, such as RAD54, are required to regulate RAD51 nucleoprotein filament function and pr...

Caffeine is a widely used inhibitor of the protein kinases that play a central role in the DNA damage response. We used chemical inhibitors and genetically deficient mouse embryonic stem cell lines to study the role of DNA damage response in stable integration of the transfected DNA and found that caffeine rapidly, efficiently and reversibly inhibited homologous integration of the transfected DNA as measured by several homologous recombination-mediated gene-targeting assays. Biochemical and structural biology experiments revealed that caffeine interfered with a pivotal step in homologous recombination, homologous joint molecule formation, through increasing interactions of the RAD51 nucleoprotein filament with non-homologous DNA. Our results suggest that recombination pathways dependent on extensive homology search are caffeine-sensitive and stress the importance of considering direct checkpoint-independent mechanisms in the interpretation of the effects of caffeine on DNA repair.