Kei-Ichi Takata

Defined DNA lesions are essential substrates for dissecting DNA repair mechanisms, yet existing approaches remain limited in positional flexibility, lesion diversity, efficiency, or compatibility with long and chromatinized DNA. Here, we present CGRID (CRISPR-mediated Gap-based Replacement for Insertion of DNA damage), a programmable strategy for site-specific installation of DNA lesions into long DNA and nucleosomes. CGRID uses paired Cas9 nickases to generate two adjacent nicks, followed by gap formation and replacement with a synthetic lesion-containing oligonucleotide. This design enables easy and efficient insertion of diverse DNA lesions at user-defined positions without pre-engineered nicking sites or substrate redesign. Applied to long λ-DNA, CGRID supports efficient installation of single lesions and internal double-strand breaks at predefined positions, which are challenging to obtain using existing methods. These molecules enable high-throughput single-molecule analysis of diverse repair pathways including base excision repair (BER), nonhomologous end joining, and microhomology-mediated end joining without hydrodynamic force. CGRID further allows efficient site-specific lesion incorporation into nucleosomal DNA, enabling systematic repositioning of damage within a nucleosome. This shows that nucleosome context and lesion orientation strongly influence APE1 incision efficiency. Together, CGRID establishes a broadly applicable platform for programmable DNA damage engineering, bridging biochemical and single-molecule methodologies, and facilitates mechanistic studies of chromatin-dependent genome maintenance.

HELQ is a 3′–5′ DNA helicase whose loss sensitizes cells to DNA-damaging agents, particularly DNA crosslinkers. HELQ interacts with the RAD51 paralog complex RAD51B–RAD51C–RAD51D–XRCC2 (BCDX2), a key mediator of replication fork reversal. Using DNA fiber assays, we show that HELQ and BCDX2 act epistatically to slow replication fork progression under replication stress. Because fork reversal transiently regresses nascent strands into a four-way junction and reduces net DNA synthesis, this fork slowing provides a functional readout of fork reversal. Directly supporting this model, electron microscopy reveals that reversed fork structures are reduced in HELQ-knockout cells. Consistent with a role in fork reversal, HELQ deletion suppresses nascent strand degradation when BRCA2- or FANCD2-dependent fork protection is lost. Mechanistically, biochemical reconstitution shows that HELQ is stimulated by RPA on fork substrates containing a leading strand gap, and these findings are consistent with the cell-based DNA fiber assays. Together, these results identify HELQ as a specialized regulator of replication fork remodeling that promotes fork reversal through the BCDX2 pathway.

DNA double-strand break (DSB) repair via homologous recombination is initiated by end resection. The extent of DNA end resection determines the choice of the DSB repair pathway. Nucleases for end resection have been extensively studied. However, it is still unclear how the potential DNA structures generated by the initial short resection by MRE11-RAD50-NBS1 are recognized and recruit proteins, such as EXO1, to DSB sites to facilitate long-range resection. We found that the MSH2-MSH3 mismatch repair complex is recruited to DSB sites through interaction with the chromatin remodeling protein SMARCAD1. MSH2-MSH3 facilitates the recruitment of EXO1 for long-range resection and enhances its enzymatic activity. MSH2-MSH3 also inhibits access of POLθ, which promotes polymerase theta-mediated end-joining (TMEJ). Collectively, we present a direct role of MSH2-MSH3 in the initial stages of DSB repair by promoting end resection and influencing the DSB repair pathway by favoring homologous recombination over TMEJ.

DNA polymerase θ (POLQ) is a unique DNA polymerase that is able to perform microhomology-mediated end-joining as well as translesion synthesis (TLS) across an abasic (AP) site and thymine glycol (Tg). However, the biological significance of the TLS activity is currently unknown. Herein we provide evidence that the TLS activity of POLQ plays a critical role in repairing complex DNA double-strand breaks (DSBs) induced by high linear energy transfer (LET) radiation. Radiotherapy with high LET radiation such as carbon ions leads to more deleterious biological effects than corresponding doses of low LET radiation such as X-rays. High LET-induced DSBs are considered to be complex, carrying additional DNA damage such as AP site and Tg in close proximity to the DSB sites. However, it is not clearly understood how complex DSBs are processed in mammalian cells. We demonstrated that genetic disruption of POLQ results in an increase of chromatid breaks and enhanced cellular sensitivity following treatment with high LET radiation. At the biochemical level, POLQ was able to bypass an AP site and Tg during end-joining and was able to anneal two single-stranded DNA tails when DNA lesions were located outside the microhomology. This study offers evidence that POLQ is directly involved in the repair of complex DSBs.

DNA polymerase theta (POLQ)-mediated end joining (TMEJ) is a distinct pathway for mediating DNA double-strand break (DSB) repair. TMEJ is required for the viability of BRCA-mutated cancer cells. It is crucial to identify tumors that rely on POLQ activity for DSB repair, because such tumors are defective in other DSB repair pathways and have predicted sensitivity to POLQ inhibition and to cancer therapies that produce DSBs. We define here the POLQ-associated mutation signatures in human cancers, characterized by short insertions and deletions in a specific range of microhomologies. By analyzing 82 COSMIC (Catalogue of Somatic Mutations in Cancer) signatures, we found that BRCA-mutated cancers with a higher level of POLQ expression have a greatly enhanced representation of the small insertion and deletion signature 6, as well as single base substitution signature 3. Using human cancer cells with disruptions of POLQ, we further show that TMEJ dominates end joining of two separated DSBs (distal EJ). Templated insertions with microhomology are enriched in POLQ-dependent distal EJ. The use of this signature analysis will aid in identifying tumors relying on POLQ activity.

Incorporation of the clinically active deoxycytidine analogue 2′-C-cyano-2′-deoxy-1-β-D-arabino-pentofuranosyl-cytosine (CNDAC) into DNA generates single-strand breaks that are subsequently converted to double-strand breaks (DSB). Here, we investigated the cellular manifestations of these breaks that link these mechanisms to cell death, and we further tested the relevance of DNA repair pathways in protection of cells against CNDAC damage. The present investigations demonstrate that following exposure to CNDAC and a wash into drug-free medium, chromosomal aberrations, DNA strand breaks, and multinucleate cells arose. These portended loss of viability and were dependent upon exposure time, CNDAC concentration, and passage through mitosis. Following a pulse incubation with CNDAC, live cell imaging using GFP-tagged histone H2B as a marker demonstrated a normal rate of progression to mitosis, but a concentration-dependent delay in passage to a second mitosis. Progression through mitosis was also delayed and accompanied by formation of multinucleate cells. CNDAC-treated cells lacking XPF-ERCC1 nuclease function showed a 16-fold increase in chromosome aberrations. Chromosomal damage in Rad51D-mutant cells (homologous recombination repair deficient) were even more severely affected with extensive aberrations. Rodent or human Polq (POLQ) mutant cells, defective in Pol θ–mediated alternative end joining, did not show enhanced cellular sensitivity to CNDAC. These findings are consistent with formation of DSBs in the second S-phase following exposure, resulting in chromosome aberrations, aberrant mitoses, and subsequent apoptosis.

To exploit vulnerabilities of tumors, it is urgent to identify associated defects in genome maintenance. One unsolved problem is the mechanism of regulation of DNA double‐strand break repair by REV7 in complex with 53BP1 and RIF1, and its influence on repair pathway choice between homologous recombination and non‐homologous end‐joining. We searched for REV7‐associated factors in human cells and found FAM35A, a previously unstudied protein with an unstructured N‐terminal region and a C‐terminal region harboring three OB‐fold domains similar to single‐stranded DNA‐binding protein RPA, as novel interactor of REV7/RIF1/53BP1. FAM35A re‐localized in damaged cell nuclei, and its knockdown caused sensitivity to DNA‐damaging agents. In a BRCA1‐mutant cell line, however, depletion of FAM35A increased resistance to camptothecin, suggesting that FAM35A participates in processing of DNA ends to allow more efficient DNA repair. We found FAM35A absent in one widely used BRCA1‐mutant cancer cell line (HCC1937) with anomalous resistance to PARP inhibitors. A survey of FAM35A alterations revealed that the gene is altered at the highest frequency in prostate cancers (up to 13%) and significantly less expressed in metastatic cases, revealing promise for FAM35A as a therapeutically relevant cancer marker.

In order to exploit the specific vulnerabilities of tumors, it is urgent to identify the basis of associated defects in genome maintenance. One unsolved problem is the mechanism of inhibition of processing of DNA double-strand break repair by REV7 and its influence on DNA repair pathways. We searched for REV7-associated proteins in human cells and found FAM35A, a protein of previously unknown function. By analyzing the FAM35A sequence we discovered that FAM35A has an unstructured N-terminal region and a C-terminal region harboring three OB-fold domains similar to single-stranded binding protein RPA. Knockdown of FAM35A caused sensitivity to DNA damaging agents, and FAM35A re-localized in damaged cell nuclei. In a BRCA1 mutant cell line, however, depletion of FAM35A increased resistance to camptothecin, suggesting that FAM35A participates in processing of DNA ends to allow more efficient DNA repair. Moreover, we found FAM35A absent in one widely used BRCA1-mutant cancer cell line (HCC1937) with anomalous resistance to PARP inhibitors. A survey of FAM35A alterations in cancer revealed that the gene is altered at the highest frequency in prostate cancers (up to 13%) and significantly less expressed in metastatic cases. The results reveal a new DNA repair factor with promise as a therapeutically relevant cancer marker.

DNA polymerase ν (pol ν), encoded by the POLN gene, is an A-family DNA polymerase in vertebrates and some other animal lineages. Here we report an in-depth analysis of pol ν–defective mice and human cells. POLN is very weakly expressed in most tissues, with the highest relative expression in testis. We constructed multiple mouse models for Poln disruption and detected no anatomic abnormalities, alterations in lifespan, or changed causes of mortality. Mice with inactive Poln are fertile and have normal testis morphology. However, pol ν–disrupted mice have a modestly reduced crossover frequency at a meiotic recombination hot spot harboring insertion/deletion polymorphisms. These polymorphisms are suggested to generate a looped-out primer and a hairpin structure during recombination, substrates on which pol ν can operate. Pol ν-defective mice had no alteration in DNA end-joining during immunoglobulin class-switching, in contrast to animals defective in the related DNA polymerase θ (pol θ). We examined the response to DNA crosslinking agents, as purified pol ν has some ability to bypass major groove peptide adducts and residues of DNA crosslink repair. Inactivation of Poln in mouse embryonic fibroblasts did not alter cellular sensitivity to mitomycin C, cisplatin, or aldehydes. Depletion of POLN from human cells with shRNA or siRNA did not change cellular sensitivity to mitomycin C or alter the frequency of mitomycin C-induced radial chromosomes. Our results suggest a function of pol ν in meiotic homologous recombination in processing specific substrates. The restricted and more recent evolutionary appearance of pol ν (in comparison to pol θ) supports such a specialized role. The work described here fills a current gap in the study of the 16 known DNA polymerases in vertebrate genomes. Until now, experiments with genetically disrupted mice have been reported for all but pol ν, encoded by the POLN gene. To intensively analyze the role of mammalian pol ν we generated multiple Poln-deficient murine models. We discovered that Poln is uniquely upregulated during testicular development and that it is enriched in spermatocytes. This, and phylogenetic analysis indicate a testis-specific function. We observed a modest reduction in meiotic recombination at a recombination hotspot in Poln-deficient mice. Pol ν has been suggested to function in DNA crosslink repair. However, we found no increased DNA crosslink sensitivity in Poln-deficient mice or POLN-depleted human cells. This is a major difference from some previous findings, and we support our conclusion by multiple experimental approaches, and by the very low or absent expression of functional pol ν in mammalian somatic cells. The present work represents the first description and comprehensive analysis of mice deficient in pol ν, and the first thorough phenotypic analysis in human cells.