The maintenance of genome integrity and fidelity is essential for the proper functioning and survival of all organisms. The research in our laboratory is centered on defining the molecular mechanisms underlying the DNA repair processes in human cells. A long-term goal of our work is to exploit this knowledge for the development of new therapeutic strategies for the treatment of cancer, which are based on targeting specific DNA-repair pathways with small molecule inhibitors.


Sensing and signaling of DNA damage

DNA damage triggers a complex network of pathways that sense and signal problems in the DNA, arrest cell-cycle progression and activate appropriate DNA-repair mechanisms. The key regulators of the mammalian DNA damage response are the ATM (ataxia telangiectasia mutated) and ATR (ATM and Rad3-related) protein kinases. ATM is the primary mediator of cellular response to DNA double-strand breaks (DSBs), while ATR has a crucial role in stabilizing the genome during DNA replication. ATR-mediated suppression of dormant replication origins shields active forks against irreversible breakage via preventing exhaustion of nuclear RPA that contributes to fork stabilization. We aim to gain a thorough understanding of the molecular mechanisms underlying the activation of ATR kinase in response to replication stress. Recent studied in our laboratory have identified the mismatch-binding protein MutSβ as a new DNA damage sensor in the process of ATR activation by replication-associated DSBs. Our work has demonstrated that MutSβ binds to hairpin loops persisting in RPA-coated single-stranded DNA (ssDNA) at sites of DNA damage and mediates the recruitment of the ATR-ATRIP complex, a prerequisite for ATR activation by TOPBP1. Based on these findings, we have proposed a model wherein the formation of hairpin loops in ssDNA generated at sites of DNA damage could signal lack of free RPA in the cell and serve as a trigger for ATR activation in a process mediated by MutSβ (Figure 1). Our current studies aim to gain further insight into the molecular mechanism of MutSβ-dependent ATR activation. Given synthetic lethal interaction between ATR and the ATM-p53 tumor suppressor pathway, we also aim to explore the possibility of using ATM inhibitors for specific killing of colorectal cancers associated with loss of MutSβ. 

Figure 1

Figure 1. Model for the role of MutSβ in ATR activation. MutSβ (MSH2–MSH3) binds to hairpin loop structures formed in RPA-coated ssDNA at sites of DNA damage and recruits the ATR-ATRIP complex. Subsequently, ATR-ATRIP binds directly to RPA-ssDNA leading to ATR activation by TOPBP1. ATP binding by MutSβ triggers its dissociation from ssDNA and the ATR-ATRIP complex.


Repair of DNA double-strand breaks by homologous recombination

We have a long-standing interest in understanding the biochemistry of DSB repair by homologous recombination (HR). HR, which is only operative during S/G2 phases of the cell cycle, is initiated by nucleolytic resection of broken DNA ends to generate 3’-ssDNA tails. One of these ssDNA tails is utilized for RAD51-mediated homology search on the undamaged sister chromatid, and after invasion and pairing with a homologous region, it primes DNA synthesis to restore the DNA continuity at the break site. Majority of HR events in mitotic cells proceed via the so-called synthesis-dependent strand annealing (SDSA) pathway, where the extended joined DNA molecule (D-loop) is disrupted by specialized DNA helicases and the newly synthesized DNA is annealed to the ssDNA tail of the other part of the broken chromosome, followed by DNA flap removal and ligation. However, HR has a second branch where the ssDNA tail of the non-invading DNA end is annealed to the D-loop to form a double Holliday junction (DHJ), whose resolution can lead to exchanges of the flanking DNA sequences between the donor and acceptor DNA molecules, the so-called crossovers. Work in our laboratory provided insights into the molecular mechanism underlying the preferential use of the SDSA pathway. Specifically, we have identified RECQ5 DNA helicase as a factor that acts during the post-synaptic phase of SDSA to prevent formation of aberrant RAD51 filaments on the extended invading strand, thus limiting its channeling into the potentially hazardous crossover pathway. In addition, we have investigated the biochemical mechanism of DSB-end resection in human cells. Studies in yeast have demonstrated that DNA-end resection during HR is mediated by either of the two nucleases: Exonuclease 1 (Exo1) and Dna2. Dna2 acts as a ssDNA-specific endonuclease and hence requires a DNA helicase to open the DNA duplex. Work conducted in our laboratory has revealed that human DNA2 catalyzes long-range DNA-end resection in conjunction with either WRN or BLM, both of which belong to the RecQ family of DNA helicases (Figure 2). Moreover, we have demonstrated that BLM mediates DNA-end resection as part of the BLM-TOPOIIIα-RMI1-RMI2 complex that is also involved in DHJ dissolution (Figure 2).

figure 2


Figure 2. Scheme of homologous recombination pathway for repairing DNA double-strand breaks.Long-range DNA-end resection is mediated by EXO1 or DNA2 nucleases. DNA2 acts in conjunction with WRN or the BLM-TOPOIIIα-RMI1-RMI2 (BTRR) complex.


Processing of stalled replication forks

The progression of replication forks is frequently impaired by various physical obstacles such as unrepaired DNA lesions, active transcription complexes or secondary DNA structures, and slows down globally upon activation of oncogenes that deregulate the replication process. Replication fork stalling can have serious implications for genome stability and cell survival. Unreplicated DNA regions hamper proper chromosome segregation and can cause DNA damage in mitosis, which can lead to chromosomal rearrangements in the following G1 phase. Some genomic regions are particularly difficult to replicate and are unstable under conditions of replication stress. Most prominent amongst these are the so-called common fragile sites (CFSs), late-replicating regions that are frequently damaged in precancerous lesions and coincide with sites of recurrent chromosomal translocations found in cancers. One protein that plays a key role in the processing of stalled replication forks is the MUS81-EME1 endonuclease. Interestingly, we have found that MUS81 forms a stable complex with RECQ5 DNA helicase in human cells. Moreover, our work has revealed that RECQ5 binds directly to MUS81 and stimulates cleavage of forked DNA structures by MUS81-EME1 in vitro. More recently, we have obtained several lines of evidence suggesting that RECQ5 cooperates with MUS81-EME1 in the processing of late replication intermediates at CFSs during early mitosis to facilitate faithful chromosome segregation (Figure 3). Our ongoing studies aim to explore the role of RECQ5 in MUS81-mediated restart of stalled replication forks during S-phase.


Figure 3


Figure 3. RECQ5 DNA helicase associates with common fragile sites and promotes their processing by MUS81 endonuclease in early mitosis. (A) Specific binding of RECQ5 and MUS81 to common fragile sites in response to replication stress. U2OS cells were treated with 0.2 μM aphidicolin (Aph) or DMSO for 16 h. Chromatin immunoprecipitation analysis coupled with quantitative PCR was performed using primers for three CFSs (FRA3B, FRA16D and FRA7H) and one control locus (GAPDH). (B) Examples of intact and broken metaphase chromosomes of U2OS cells treated with 0.2 μM Aph for 16 h. Arrows denote chromatid breaks. (C) Western blot analysis of extracts from U2OS cells transfected for 3 days with indicated siRNAs. siCtrl, control siRNA. (D) Quantification of fragile site expression in U2OS cells transfected with indicated siRNAs. Aph (or DMSO) was added 2 days after siRNA transfection. (E) Examples of anaphase cells with PICH-positive bridges. U2OS cells synchronized at G2/M transition by RO-3306 treatment (16 h) were released into mitosis for 1.5 hours before fixation and immunostaining with anti-PICH antibody. (F) Frequency of PICH-positive anaphase bridges in cells transfected with indicated siRNAs.


Molecular basis of transcription-associated genomic instability

DNA can be damaged during transcription if the nascent transcript pairs with the template DNA strand behind the transcription complex forming a three-stranded structure called an R-loop. In this aberrant structure, the non-transcribed DNA strand is left exposed as an extended ssDNA, which makes DNA bases in this strand more prone to chemical modifications and formation of secondary DNA structures that can compromise the progression of replication machinery. Formation of R-loops is facilitated by negative supercoiling generated behind the transcription complex and is favored in the transcriptional units containing runs of Gs in the non-transcribed strand. Recent studies provided evidence that R-loops form as a consequence of collisions between transcription and replication complexes, particularly if the progression of replication forks is perturbed. In addition, it has been hypothesized that formation of R-loops is the major source of genomic instability caused by oncogene-induced replication stress. We aim to identify the loci that are prone to R-loop formation upon replication stress. As a tool for the detection of R-loops, we generated a cell line conditionally expressing a catalytically-inactive form of RNase H1 fused to green fluorescent protein [RNH1(D210N)-GFP]. This RNaseH1 mutant stably binds RNA:DNA duplexes in R-loops, but does not cleave them. Our preliminary experiments have shown that RNH1(D210N)-GFP forms discrete nuclear foci in cells exposed to replication stress,  indicative of R-loop formation (Figure 4). Ongoing work in the laboratory uses this tagged RNase H1 mutant to isolate by affinity chromatography R-loops from fragmented chromatin of cells exposed to replication stress for analysis of their DNA sequence by Illumina HiSeq sequencing. It is our keen interest to identify the proteins associated with R-loops by mass spectrometry and explore their roles in the maintenance of genomic stability.

Figure 4


Figure 4. Formation of nuclear foci of GFP-tagged RNase H1 (D210N) in response to replication stress. Expression of RNH1(D210N)-GFP in U2OS-T-REx cells was induced by doxycycline for 24 hours. For the last 6 hours, cells were treated with 10 mM hydroxyurea or, for the last 16 hours, with 0.2 µM aphidicolin. Control cells were treated with DMSO. After pre-extraction with PBS supplemented with 0.5% Triton X-100, cells were fixed with formaldehyde, immunostained for γ-H2AX and analyzed by fluorescence microscopy.