A genome-wide screening uncovers the role of CCAR2 as an antagonist of DNA end resection

Introduction

DNA double-strand breaks (DSBs) are the most dangerous form of DNA damage. Unrepaired breaks lead to cell death, while improperly repaired breaks cause an increase in genomic instability and, in humans, diseases such as cancer and premature aging1,2. There are two major pathways to repair DSBs: non-homologous end-joining (NHEJ) and homologous recombination (HR)3,4,5. NHEJ consists of a ligation of two DNA ends without using homology5. In HR, a homologous sequence is used as an information donor in a highly regulated mechanism3. Several recombination subpathways have been described, each one with distinct outcomes and consequences3. The choice between these two repair mechanisms is highly regulated, and changes in the ratio between them can increase genomic instability4. So far, the best-known regulated step of the DSB repair pathway choice is the so-called DNA end resection4. Here strands are degraded 5′–3′ at each DNA end, giving rise to ssDNA tails that are immediately coated by the replication protein A (RPA) complex for protection4. RPA-coated ssDNA is an obligatory substrate of HR and hampers NHEJ4. A major player in the choice between NHEJ and HR is CtIP (CtBP interacting protein), which licenses HR by activating DNA end resection6. Multiple signals converge on CtIP to initiate DNA end resection at those breaks that will be repaired by HR4,6,7. In order to find and characterize new factors involved in this crucial choice, we took advantage of the SeeSaw Reporter (SSR), a system designed to assess the balance between NHEJ and HR8. Using a genome-wide human esiRNA library, we found that downregulation of 1.35% of the genes shifts the NHEJ:HR ratio towards NHEJ, while depletion of a further 0.71% has the opposite effect. We focused on CCAR2, which we found to cause hyper-recombination when depleted. We show that it acts as an inhibitor of recombination. Specifically, we found that CCAR2 inhibits initiation and limits the extent of DNA end resection through its functional interaction with CtIP. This regulation of DNA end processing modulates the choice between NHEJ and HR.

Results

A genome-wide screening for regulators of the NHEJ:HR ratio

The SSR2.0 system (Fig. 1a) was designed to calculate the balance between NHEJ and HR as the ratio of green fluorescent protein (GFP)-positive versus red fluorescent protein (RFP)-positive cells of a lone DSB induced by the meganuclease I-SceI (ref. 8). Note that, in this reporter, mainly a specific subtype of HR termed single-strand annealing (SSA) is measured, which is Rad51-independent and does not require strand invasion3. SSA is very sensitive to DNA end resection but does not require additional steps; thus, our screening focused on the early steps shared by the various HR subpathways. We measured the ratio of green versus red cells using a high-throughput microscope after individually downregulating human genes using a genome-wide esiRNA library (Fig. 1b). We used 96-well plates and included esiRNA against luciferase in each plate as a control. We discarded the results of any plate in which the green versus red cell ratio of the luciferase control varied more than 10% relative to the average value from all luciferase controls. The ratio of green versus red cells was calculated for each esiRNA and normalized with the value of the internal esiRNA against luciferase. The experiment was repeated independently three times (Supplementary Data 1). Genes were ranked accordingly to average GFP:RFP cell ratio normalized with luciferase and represented graphically (Fig. 1c). We observed three categories of genes with respect to the shape of the curve. Downregulation of the majority of the genes showed a NHEJ:HR ratio similar to the control (for example, normalized GFP:RFP ratio close to 1; dashed black rectangle, Fig. 1c). Depletion of 0.71% of the genes skewed the balance towards an increase in HR (for example, normalized GFP:RFP ratio below 0.5; red ellipse, Fig. 1c). As downregulation of those genes increased HR, we categorize them as genes that naturally favour NHEJ. An additional 1.35% of the genes favoured HR, that is, NHEJ increased when downregulated (for example, normalized GFP:RFP ratio above 3; green ellipse, Fig. 1c). The thresholds of 0.5 and 3 were established with respect to the inflection points of the curve. Data analyses revealed false-positive signals for some genes because of a single experiment with extreme values. To eliminate those, we established the following criteria (Supplementary Data 2): genes for which depletion caused an average normalized GFP:RFP ratio below 0.5, with an individual GFP:RFP normalized ratio below 0.75 for all three replicas, were included in the category of genes that favour NHEJ. In contrast, genes for which depletion caused an average normalized GFP:RFP ratio above 3, with an individual GFP:RFP normalized ratio above 2 for all three replicas, were included in the category of genes that favour HR.

Figure 1: A genome-wide screening for factors that control the balance between NHEJ and HR.

(a) Schematic representation of the SSR. An I-SceI-induced DSB can be repaired by NHEJ, thus reconstructing an active GFP gene, or by homologous recombination using RFP fragments, thus creating a functional RFP gene. (b) Workflow of the screening method. Individual esiRNAs deposited in a 96-well plate format were used to reverse-transfect U2OS cells harbouring a single copy of the SSR system. After 36 h, lentiviral particles bearing I-SceI and the BFP genes were transduced into cells. After 2 days, cells were imaged using a high-throughput microscope for green, red and blue fluorescence. The number of green and red cells was then established for each individual esiRNA. Scale bar, 100 μm (c). The ratio between green and red cells was established for each individual esiRNA and normalized to an internal control (esiRNA targeted against luciferase) that was included in each plate. The individual NHEJ:HR ratio was then ordered and plotted. Cells with a normalized ratio close to one occupy the central dashed rectangle. Genes for which depletion reduced the ratio (that is, HR was increased) are marked in the red ellipse and correspond to proteins that naturally favour NHEJ. In contrast, genes that encode pro-recombination proteins (that is, NHEJ increased when they were downregulated) are included inside the green ellipse. (d) Functional categories enriched among candidates. The list of candidates that alter the NHEJ:HR ratio was analysed using the IPA Software and categorized into functional groups. Those categories in which genes were statistically significantly over-represented in the set of genes that favour NHEJ (left) or HR (right) are plotted. (e) Network of genes involved incell cycle, DNA repair, replication and recombination and cancer, which appear in the set of candidates isolated in the screening. Genes that favour HR are labelled in green, and those that facilitate NHEJ, in red.