Opposing roles for 53BP1 during homologous recombination

Although DNA non-homologous end-joining repairs most DNA double-strand breaks (DSBs) in G2 phase, late repairing DSBs undergo resection and repair by homologous recombination (HR). Based on parallels to the situation in G1 cells, previous work has suggested that DSBs that undergo repair by HR predominantly localize to regions of heterochromatin (HC). By using H3K9me3 and H4K20me3 to identify HC regions, we substantiate and extend previous evidence, suggesting that HC-DSBs undergo repair by HR. Next, we examine roles for 53BP1 and BRCA1 in this process. Previous studies have shown that 53BP1 is pro-non-homologous end-joining and anti-HR. Surprisingly, we demonstrate that in G2 phase, 53BP1 is required for HR at HC-DSBs with its role being to promote phosphorylated KAP-1 foci formation. BRCA1, in contrast, is dispensable for pKAP-1 foci formation but relieves the barrier caused by 53BP1. As 53BP1 is retained at irradiation-induced foci during HR, we propose that BRCA1 promotes displacement but retention of 53BP1 to allow resection and any necessary HC modifications to complete HR. In contrast to this role for 53BP1 in HR in G2 phase, we show that it is dispensable for HR in S phase, where HC regions are likely relaxed during replication.

As above but knockdown efficiency of KAP-1 siRNA was assessed under single, double and triple knockdown conditions. C-D) BRCA1 single and pool oligonucleotide knockdown efficiency was assessed by immunoblotting and immunostaining. For immunoblotting, MCM6 was used as a loading control. E) 53BP1 knockdown efficiency was monitored by IF following, single, double, triple and quadruple knockdown conditions. 53BP1 levels were greatly diminished in all cases although in the triple and quadruple knockdown conditions slightly more 53BP1 remained. F-I) 53BP1, KAP-1, DNAPK and KAP-1 knockdown efficiencies were assessed by western blotting, following single and multiple knockdown conditions. Ku80, α-tubulin and KAP-1 were used as loading controls.

Supplementary Figure 2.
(A) Highly resolved images of multiple focal planes through the nucleus (z-stacks) were taken using a confocal Zeiss LSM 510 microscope. Z-stacks were further processed using 3D rendering. (B) 3D model of 3D rendered z-stacks with rotation in z-axis. Threshold setting is required in order to analyze foci number. (C) 3D rendering converts the 3D model into a 2D image. The resulting 2D images were analyzed using Image J. The Analysis involves setting the threshold in each image for the heterochromatin and the RPA-foci and capturing the total number of RPA2 foci and the overlap of RPA2 foci and heterochromatin (setting a minimal size for the foci as a requirement).

Supplementary Figure 3. Controls for IP experiments.
A) HeLa cells were synchronized using a double thymidine block. 2 mM thymidine was added to adherent HeLa cells. 15h later, the cells were washed and incubated in thymidine free media for 8h. 2 mM thymidine was then re-added, and the cells were grown for a further 15h. At this point, the cells were highly synchronized in G1 phase (left panel) (no pH3 + cells).
Upon removal of thymidine and the addition of fresh media, the cells progressed through the cell cycle in synchrony and approximately 70% of cells were in G2 phase at 7 h post thymidine release, as assessed by H3S10p staining (right panel). B) Verification that α-H4K9Ac and α-H3K9Me3 detect euchromatin versus heterochromatin,, respectively. NIH3T3 cells were stained with either ChIP-grade α-H4K9Ac (red) or α-H3K9Me3 (green) antibodies and DAPI (blue). Staining of each marker is mostly mutually exclusive and α-H3K9me3 staining occurs at DAPI chromocentres.

Supplementary Figure 4. 53BP1 is dispensable for HR following siRNA XLF and inhibits HU induced SCEs.
A-B) RPA and Rad51 foci were enumerated in G2 cells (treated with the indicated siRNA) 2 h after exposure to 1.5 Gy IR. Results represent the mean and s.d. of three experiments. C) SCEs in Wild type (WT) and 53BP1 -/-MEFs following HU. In these experiments using MEFs, we observed enhanced SCEs in the untreated control compared to the untreated control in Figure 2C. We attribute this difference to the addition of aphidicolin to cells in Figure 2C but not here (Fig. S4C). With MEFs, we have observed that aphidicolin does not A-B) DSB repair by HR in G2 phase. A) Following DSB induction in G2 phase, 53BP1 localises to DSB ends and inhibits DNA end resection. At DSBs located at HC regions, BRCA1 overcomes the inhibitory barrier of 53BP1 to resection by excluding 53BP1from these regions. Concomitantly, 53BP1 is retained on chromatin and overcomes the barrier that HC poses to resection by tethering activated ATM and thus mediating HC relaxation via robust KAP-1 phosphorylation. Once the barriers posed to resection by 53BP1 and HC are overcome, DSB repair proceeds by HR. B) In the absence of 53BP1, BRCA1's function in overcoming 53BP1's barrier to resection is redundant. However, when 53BP1 is absent ATM cannot be tethered at the DSB sites and hence the barrier posed to resection by HC cannot be overcome. Consequently, resection stalls and impaired DSB repair by HR is observed. C-D) DSB repair by HR in S-phase. C) Following replication fork stalling/collapse in Sphase, 53BP1 localises to these regions and inhibits DNA end resection. BRCA1 functions to overcome the inhibitory barrier of 53BP1 to resection thus allowing fork restoration by HR to proceed. D) In the absence of 53BP1, BRCA1's function in overcoming 53BP1's barrier to resection is redundant. Moreover, 53BP1's role in chromatin relaxation is also redundant since during replication the HC superstructure is transiently dismantled and does therefore not pose a barrier to resection. Therefore unlike the situation in G2 phase where HR stalls in the absence of 53BP1, in S-phase, resection and HR can take place in the absence of 53BP1.