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. 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 unclear. 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 that actively removes loaded Ku molecules throughout the cell cycle and a CtIP/ATM-dependent mechanism that 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 mechanisms operating to prevent Ku from invading chromatin and interfering with other DNA transactions.

Keywords: ATM; CP: Molecular biology; CtIP; DNA end resection; DNA repair; DNA-PK; 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 that are wild type (WT) or knocked out for the indicated genes. See Figure S1A for the IR sensitivity analysis of each of these cells. (B) WT, PKcs-KO, LIG4-KO, or 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. Ku foci intensity was measured and normalized to the average 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. (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 nedisertib (PKi) for 1 h before treatment where 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 in Ku foci intensity with NEDi versus without NEDi (DMSO) are shown in the graph on the right. The graph depicting the fold change in Ku foci intensity in each condition corresponds to Figure S1C. (E) WT, PKcs-KO, LIG4-KO, or LIG4/PKcs-KO U2OS cells received 5 Gy of IR and were post-incubated for 5 or 60 min with or without NEDi before being processed for Ku foci imaging. Representative pictures are shown on the left, while the ratio between the change in Ku foci intensity with NEDi versus without NEDi (DMSO) is plotted in the graph on the right. The graph in Figure S1D depicts the fold change in Ku foci intensity in each condition. (F) WT, PKcs-KO, or LIG4-KO U2OS cells incubated with 3 μM NU7441 (PKi) were treated with 3 nM Cali for 1 h with or without NEDi before being collected and processed to separate the chromatin fraction from the 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 (immunoglobulin G [IgG]) or anti-Ku immunoprecipitation was performed from the soluble fractions of PKcs-KO U2OS to monitor Ku ubiquitination in response to Cali with or without NEDi. For all panels, error bars represent SD from the means of n ≥ 3 independent experiments. p values are as follow: ns p > 0.05, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. Scale bars, 5 μm.

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 the presence or the absence of DNA-PKcs. A Cy3-labeled 100 bp DNA substrate was attached to a glass surface and incubated for 60 min with Cy5-labeled Xenopus Halo-Ku80:Ku70 alone or in non-cycling Xenopus eggs extracts. After incubation and washout 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 Figure S3J. (C–H) Normalized histograms depicting the 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 egg lysis buffer (ELB) wash buffer. (D) The number of Ku molecules was monitored as described in (A) using Xenopus egg extracts containing Cy5-labeled Ku. (E) The number of Ku molecules was monitored as described in (A) using Xenopus egg 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 egg 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 and 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. For all panels, error bars represent SD from the means of n ≥ 3 independent experiments.

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

(A) PKcs-KO U2OS cells were pre-treated with NEDi, ATMi, or both and received 5 Gy of IR before being post-incubated for the indicated time and then processed for immunofluorescence. PCNA staining was used to identify the cells in S phase. Representative pictures are shown on the left, with insets at the bottom to illustrate the PCNA staining. Ku foci intensity was measured and normalized to the average 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 in the graph on the right. (B) PKcs-KO U2OS cells were transfected with siRNA control or against FBXL12 before being treated and processed as described in (A). Representative pictures are shown on the left, with insets at the bottom to illustrate the PCNA staining. An immunoblot showing the depletion of FBXL12 is shown in Figure S5A, while the graph on the right shows the fold change in Ku foci intensity computed as in (A). (C) PKcs-KO U2OS cells were transfected with siRNA control or against CtIP before being treated and processed as described in (A). Representative pictures are shown on the left, with insets at the bottom to illustrate the PCNA staining. An immunoblot showing the depletion of CtIP is shown in Figure S5E, while the graph on the right shows the fold change in Ku foci intensity computed as in (A). For all panels, error bars represent SD from the means of n ≥ 3 independent experiments. p values are as follow: ns p > 0.05, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p < 0.0001. Scale bars, 5 μm.

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

(A) Scheme depicting the linear substrate used to monitor Ku interference with transcription. The linear PCR product bears a minimal transcription unit coding for GFP with 5′ DNA ends protected against exonuclease by five phosphorothioate linkages. Upon transfection, Ku is expected to bind to the linear substrate DNA ends and by threading in to physically impede GFP transcription. (B) WT, LIG4-KO, or LIG4/PKcs-KO U2OS cells were co-transfected with an mCherry-coding circular plasmid together with the circular (GFP unit inserted in a plasmid) 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, identified using the mCherry fluorescence. (C) WT or LIG4/PKcs-KO U2OS cells, treated with PKi where indicated, were transfected and analyzed by flow cytometry as described in (B). (D) LIG4-KO or LIG4/PKcs-KO U2OS cells were transfected with biotinylated (Biot.) or non-biotinylated (Non-biot.) linear substrates for 4 h before being lysed. The substrates with the associated proteins were recovered from the extracts using streptavidin pull-down and analyzed by immunoblotting. A 20th of the extracts used for pull-down was used as input control. The graph on the right corresponds to the quantification of the Ku70 signal, normalized to the input, from 3 independent experiments. (E) WT or PKcs-KO U2OS mAID-Ku70 cells were treated for 8 h with IAA before being lysed and analyzed by immunoblotting. (F) WT or PKcs-KO U2OS mAID-Ku70 cells were transfected and analyzed by flow cytometry as described in (B). Where indicated, cells were treated with PKi, and Ku degradation was induced with IAA, added 19 h before transfection. The graphs correspond to four independent experiments. For all panels, error bars represent SD from the means of n ≥ 3 independent experiments. p values are as follow: ns p > 0.05, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p < 0.0001.

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

Upon DSB formation, Ku quickly loads on DNA ends. (A) The formation of a Ku-DNA-PKcs complex physically restrains Ku entry into chromatin, enforcing the 1:1 Ku:DNA end stoichiometry. (B) In the 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. In all cell-cycle phases (bottom left), Ku is evicted via its ubiquitination by an FBXL12-containing SCF complex whose activity relies on its neddylation. In S phase (bottom right), an ATM/CtIP-dependent mechanism overcomes Ku accumulation through DNA end resection. Ku ubiquitination also contributes to Ku removal in S phase.