Identification of the main barriers to Ku accumulation in chromatin

Abstract

Repair of DNA double strand breaks by the non-homologous end-joining pathway is initiated by the binding of Ku to DNA ends. Given its high affinity for ends, multiple Ku proteins load onto linear DNAs in vitro. However, in cells, Ku loading is limited to ~1-2 molecules per DNA end. The mechanisms enforcing this limit are currently unknown. Here we show that the catalytic subunit of the DNA-dependent protein kinase (DNA-PKcs), but not its protein kinase activity, is required to prevent excessive Ku entry into chromatin. Ku accumulation is further restricted by two mechanisms: a neddylation/FBXL12-dependent process which actively removes loaded Ku molecules throughout the cell cycle and a CtIP/ATM-dependent mechanism which operates in S-phase. Finally, we demonstrate that the misregulation of Ku loading leads to impaired transcription in the vicinity of DNA ends. Together our data shed light on the multiple layers of coordinated mechanisms operating to prevent Ku from invading chromatin and interfering with other DNA transactions.

Keywords: ATM; CtIP; DNA end resection; DNA repair; DNA-PKcs; FBXL12; Ku; MRN; NHEJ; Xenopus; neddylation.

Figure 1:. DNA-PKcs presence but not activity limits Ku accumulation at DSBs.

A. Immunoblot of U2OS cells wild-type (WT) or knocked-out for the indicated genes. See Fig. S1A for the IR sensitivity analysis of each of these cells. B. U2OS WT, DNA-PKcs KO (PKcs-KO), DNA ligase IV KO (LIG4-KO) or DNA-PKcs and Lig4 KO (LIG4/PKcs-KO) received 5 Gy of IR before being post-incubated 5 or 60 min and processed for Ku foci imaging. Representative pictures are shown on the left panel. Ku foci average intensity was measured and normalized to the Ku foci intensity measured after 5 Gy of IR in WT U2OS to compute the fold change in Ku foci intensity in each condition, depicted on the graph on the right panel. C. WT or PKcs-KO U2OS cells received 5 Gy of IR and were post-incubated for 60 min before being processed for Ku foci imaging. Cells were pre-treated with 250 nM nedisertib (PKi) for 1 h before treatment when indicated. Fold change in Ku foci intensity in each condition is displayed. D. PKcs-KO or LIG4-KO U2OS cells received 5 Gy of IR and were post-incubated for the indicated time with or without NEDi before being processed for Ku foci imaging. Representative pictures are shown on the left, while the ratio between the changes of Ku foci intensity with NEDi versus without NEDi (DMSO) are plotted on the graph on right. The corresponding graph depicting the fold change in Ku foci intensity in each condition corresponds to Fig. S1B. E. WT, PKcs-KO, LIG4-KO or LIG4/PKcs-KO U2OS cells received 5 Gy of IR and were post-incubated 5 or 60 min with or without NEDi before being processed for Ku foci imaging. Representative pictures are shown on the left panel, while the ratio between the change of Ku foci intensity with NEDi versus without NEDi (DMSO) is plotted on the graph on right. The corresponding graph depicting the fold change in Ku foci intensity in each condition corresponds to Fig. S1C. Fig. S1D,E and Fig. S1F,G depict similar analyses in HeLa cells and in complemented PKcs-KO U2OS, respectively, while Fig. S2 reports the use of STORM imaging to monitor the size and shape of Ku foci in absence of DNA-PKcs +/−NEDi. F. WT, PKcs-KO or LIG4-KO U2OS incubated with 3 μM of NU7441 (PKi) were treated with 3 nM Cali for 1 h with or without NEDi, before being collected and processed to separate a chromatin fraction from a soluble fraction which were both analyzed by immunoblotting. SAF-A and nucleolin were used as loading controls for the chromatin and the soluble fraction, respectively. G. Control (IgG) or anti-Ku immunoprecipitation were performed from the soluble fractions to monitor Ku ubiquitination in response to Cali with or without NEDi.

Figure 2:. DNA-PKcs prevents Ku overloading onto a single DNA end.

A. Scheme depicting the single molecule assay used to quantify Ku loading on DNA in presence or in absence of DNA-PKcs. A Cy3-labeled 100 bp DNA substrate was attached to a glass surface, incubated for 60 min with Cy5-labelled xenopus Halo-Ku80:Ku70 alone or in non-cycling Xenopus eggs extracts. After incubation and wash-out of the oxygen scavenger, the photobleaching of individual Ku molecules was monitored under continuous illumination and the number of bleaching steps was used as a readout of the number of Ku molecules on DNA. B. Representative trajectories highlighting photobleaching events for Halo-Ku80:Ku70 bound to DNA in buffer. Representative trajectories for each condition are shown in Fig. S3J. C–H: On each panel the normalized histograms depicting fractional occupancy of Ku70/80 on DNA ends, constructed from mean fractions per occupancy bin calculated in 3 independent experiments. The number of events and total number of molecules observed for each experiment are reported in Table 1. C. The number of Ku molecules was monitored as described in A. using purified Cy5-labeled Ku incubated in ELB wash buffer. D. The number of Ku molecules was monitored as described in A. using Xenopus eggs extracts containing Cy5-labeled Ku. E. The number of Ku molecules was monitored as described in A. using Xenopus eggs extracts immunodepleted for DNA-PKcs (ΔDNA-PKcs) and containing purified Cy5-labeled Ku. F. The number of Ku molecules was monitored as described in A. using Xenopus eggs extracts immunodepleted for XLF and XRCC4 (ΔXLF/ΔXRCC4) and containing purified Cy5-labeled Ku. G. The number of Ku molecules was monitored as described in A. using purified Cy5-labeled Ku mixed with 60 nM Lig4-XRCC4, 60 nM XLF, and incubated in ELB wash buffer. H. The number of Ku molecules was monitored as described in A. using purified Cy5-labeled Ku mixed with 60 nM DNA-PKcs and incubated in ELB wash buffer.

Figure 3:. Different mechanisms limit Ku loading during the cell cycle.

A. PKcs-KO U2OS were pre-treated with NEDi, ATMi or both and received 5 Gy of IR before being post-incubated for the indicated time before being processed for immunofluorescence. A PCNA staining was used to identify the cells in S-phase. Representative pictures are shown on the left panel, with insets at the bottom to illustrate the PCNA staining. Ku foci average intensity was measured and normalized to the Ku foci intensity in U2OS WT measured 5 min after 5 Gy of IR to compute the fold change in Ku foci intensity in each condition, displayed on the graph on the right panel. B. PKcs-KO U2OS were transfected by siRNA control or against FBXL12 before being treated and processed as described in A. Representative pictures are shown on the left panel, with insets at the bottom to illustrate the PCNA staining. An immunoblot showing the depletion of FBXL12 is shown in Fig. S5A, while the graph on the right panel shows the fold change in Ku foci intensity computed as in A. C. PKcs-KO U2OS were transfected by siRNA control or against CtIP before being treated and processed as described in A. Representative pictures are shown on the left panel, with insets at the bottom to illustrate the PCNA staining. An immunoblot showing the depletion of CtIP is shown in Fig. S5E, while the graph on the right panel shows the fold change in Ku foci intensity computed as in A.

Figure 4:. DNA-PKcs deficiency impacts on transcription at the DNA end vicinity.

A. Scheme depicting the linear substrate used to monitor Ku interference with transcription. Cells are co-transfected with a circular plasmid coding for mCherry and with a linear PCR product with a minimal transcription unit coding for GFP with 5’DNA ends protected against exonuclease by five phosphorothioate linkages or with the same transcription unit inserted in a circular plasmid. Upon transfection, Ku is expected to bind to the linear substrate DNA ends and thread in, physically impeding its transcription. B. WT, LIG4-KO or LIG4/PKcs-KO U2OS cells were co-transfected with a mCherry-coding circular plasmid together with the circular (left panel) or linear (right panel) GFP-coding substrates for 24 h before being analyzed by flow cytometry. The graphs represent the percentage of GFP positive cells among the cells successfully transfected, as identified using the mCherry fluorescence. C. WT or LIG4/PKcs-KO U2OS cells, treated with 250 nM nedisertib (PKi) when indicated, were co-transfected with a mCherry-coding circular plasmid together with the circular or linear GFP-coding substrates for 24 h before being analyzed by flow cytometry. The graphs represent the percentage of GFP positive cells among the cells successfully transfected, as identified using the mCherry fluorescence.

Figure 5:. Model summarizing the different barriers to Ku overloading.

Upon DSB formation, Ku quickly load on DNA ends. The formation of a Ku-DNA-PKcs complex physically restrains Ku entry into chromatin, enforcing the 1:1 Ku-DNA end stoichiometry (A). In absence of DNA-PKcs, multiple Ku proteins can load and slide from the DNA end into chromatin. The progressive accumulation of larger amounts of Ku into chromatin is actively restricted by two mechanisms (B). In all cell cycle phases (bottom left), Ku is evicted via its ubiquitination by a FBXL12-containing SCF complex whose activity relies on its neddylation. In S-phase (bottom right), an ATM/CtIP-dependent mechanisms overcomes Ku accumulation through DNA end resection.