DNA-PKcs has KU-dependent function in rRNA processing and haematopoiesis

Abstract

The DNA-dependent protein kinase (DNA-PK), which comprises the KU heterodimer and a catalytic subunit (DNA-PKcs), is a classical non-homologous end-joining (cNHEJ) factor¹. KU binds to DNA ends, initiates cNHEJ, and recruits and activates DNA-PKcs. KU also binds to RNA, but the relevance of this interaction in mammals is unclear. Here we use mouse models to show that DNA-PK has an unexpected role in the biogenesis of ribosomal RNA (rRNA) and in haematopoiesis. The expression of kinase-dead DNA-PKcs abrogates cNHEJ². However, most mice that both expressed kinase-dead DNA-PKcs and lacked the tumour suppressor TP53 developed myeloid disease, whereas all other previously characterized mice deficient in both cNHEJ and TP53 expression succumbed to pro-B cell lymphoma³. DNA-PK autophosphorylates DNA-PKcs, which is its best characterized substrate. Blocking the phosphorylation of DNA-PKcs at the T2609 cluster, but not the S2056 cluster, led to KU-dependent defects in 18S rRNA processing, compromised global protein synthesis in haematopoietic cells and caused bone marrow failure in mice. KU drives the assembly of DNA-PKcs on a wide range of cellular RNAs, including the U3 small nucleolar RNA, which is essential for processing of 18S rRNA⁴. U3 activates purified DNA-PK and triggers phosphorylation of DNA-PKcs at T2609. DNA-PK, but not other cNHEJ factors, resides in nucleoli in an rRNA-dependent manner and is co-purified with the small subunit processome. Together our data show that DNA-PK has RNA-dependent, cNHEJ-independent functions during ribosome biogenesis that require the kinase activity of DNA-PKcs and its phosphorylation at the T2609 cluster.

Extended Data Fig. 1 |. Myeloid proliferative disease with MDS features in DNA-PKcs^KD/KDTp53^−/− mice.

a, Representative fluorescence-activated cell sorting (FACS) analyses of pro-B cell lymphomas from DNA-PKcs^KD/KDTp53^−/− mice. b, The frequency of megakaryocytes with hyperchromatic or pyknotic nuclei and total megakaryocytes per 400× field in the spleen. Eight fields from three independent DNA-PKcs^KD/KDTp53^−/− mice were quantified and are presented here. Mean ± s.e.m.; two-sided unpaired student’s t-test, **P < 0.01, ***P < 0.001. c, Representative FACS analyses of myeloid proliferative disease in DNA-PKcs^KD/KDTp53^−/− mice. The CD19-negative bone marrow from DNA-PKcs^KD/KDTp53^−/− mice shows a higher percentage of myeloid cells (CD11b⁺) at the expense of RBCs (TER119⁺). CD19⁺ B cells were excluded from the wildtype bone marrow, as DNA-PKcs^KD/KDTp53^−/− mice have no pre-B and mature B cells (CD19⁺) owing to cNHEJ defects. There is also an increase in the percentage of c-KIT⁺ cells among the myeloid cells (CD11b⁺GR1⁺) from DNA-PKcs^KD/KDTp53^−/− mice. Mid-sized reticulocytes and RBC progenitors are reduced in the DNA-PKcs^KD/KDTp53^−/− bone marrow. Exact P values and defined sample sizes (n) are provided in Supplementary Data 1.

Extended Data Fig. 2 |. Generation and characterization of the DNA-PKcs^PQR/PQR and DNA-PKcs^5A/5A mouse models.

a, Diagram of the protein sequence changes in the DNA-PKcs^PQR allele. b, c, Diagram of the protein sequence changes in the DNA-PKcs^5A allele and the targeting scheme. The TQ sites are shown in upper case for ease of identification. Asterisks denote sites mutated in the previously published DNA-PKcs^3A allele. The schematic diagram (not to scale) represents the mouse Prkdc (DNA-PKcs) locus (top), targeting vector (second row), targeted allele (DNA-PKcs^5AN, third row), and the neo-deleted mutant allele (DNA-PKcs^5A, bottom). The red exon contains the mutations. Boxes, exons; triangles, FRT sites. d, Southern blot analyses of SpeI-digested DNA from DNA-PKcs^+/+ and DNA-PKcs^+/5AN (with NeoR) ES cells. e, f, The concentration of RBCs and platelets (PLT) in the peripheral blood of 5-month-old DNA-PKcs^PQR/PQR mice and DNA-PKcs^+/+ controls. g, The platelet concentration in the peripheral blood of 2-week-old DNA-PKcs^+/+, DNA-PKcs^5A/5A and DNA-PKcs^−/− mice. h, i, Representative images show pale kidney (h) and liver (i) from 2-week-old DNA-PKcs^5A/5A mice, indicating extreme anaemia. e–g, Mean ± s.e.m.; two-sided unpaired Student’s t-test, ***P < 0.001; n.s., not significant, P > 0.05). Exact P values and defined sample sizes (n) are provided in Supplementary Data 1.

Extended Data Fig. 3 |. Haematopoiesis and translation defects in DNA-PKcs^5A/5A mice.

a, b, Flow cytometry analyses (a) and quantification (b) of 2-week bone marrow progenitor cell populations. The same gating strategy was used throughout to define progenitor populations. Lin⁻SCA1⁺c-KIT⁺ haematopoietic stem and progenitor cells (LSK, HSPC) and LK cells (Lin⁻SCA1⁻c-KIT⁺) were further divided into CMPs, granulocyte-monocyte progenitors (GMPs) and MEPs. c, d, Flow cytometry analyses (c) and quantification (d) of E14.5 fetal liver HSPC (Lin⁻SCA1⁺c-KIT⁺) frequency from DNA-PKcs^KD/KD and DNA-PKcs^5A/5A embryos. e, Flow cytometry gating strategy for analysing protein translation in Lin⁺ cells. The gating strategy for progenitors is shown in a. This strategy (a, e) was used to determine global translation for each population in Extended Data Fig. 7d. Mean ± s.e.m.; two-sided unpaired Student’s t-test; ***P < 0.001, *P < 0.05, n.s. P > 0.05. Exact P values and defined sample sizes (n) are provided in Supplementary Data 1.

Extended Data Fig. 4 |. Normal V(D)J recombination in DNA-PKcs^5A/5A mice and cells.

a, Flow cytometry analyses of lymphocyte development in DNAPKcs^5A/5ATp53^−/− and control mice. b, Frequency of B220⁺IgM⁺ naive mature B cells from spleen. c, Quantification of peripheral RBCs in DNA-PKcs^5A/5A mice with and without TP53 deficiency. Note that wild-type, DNA-PKcs^5A/5A and DNA-PKcs^−/− data from Fig. 2b are included here for comparison. d, e, Absolute counts of peripheral lymphocytes (d) and neutrophils (e) in DNA-PKcs^5A/5A mice with or without Tp53 or Ku70 deficiency. TP53-deficiency rescued both the lymphocytopaenia and the neutropaenia in DNA-PKcs^5A/5A mice. Notably, KU70 deficiency caused severe lymphocytopaenia by itself or with the DNA-PKcs^5A/5A mutation, but rescued the neutropaenia in DNA-PKcs^5A/5A mice. f–h, Chromosomal V(D)J recombination measured in v-ABL kinase-transformed B cells. f, Diagram of pMX-INV chromosomal V(D)J recombination substrate. Empty and filled triangles represent recombination signal sequences. RV, EcoRV, N, NcoI. Dashed line marks the C4 probe used for Southern blotting. g, Southern blotting reveals the accumulation of normal coding joins (CJ) and signal joins (SJ) and suggests successful V(D)J recombination in DNA-PKcs^5A/5A B cells. h, Successful V(D)J recombination places Gfp in the same orientation as the promoter and leads to GFP expression. Plots show the frequency of GFP⁺ cells after 4 days of STI571 (3 μM) treatment. i, Quantification of DN2 and DN3 cells out of all double-negative cells (left, reflecting TCRβ V(D)J recombination) and percentage of CD4⁺ or CD8⁺ single-positive (SP) cells (right, reflecting TCRα V(D)J recombination). b–e, i, Mean ± s.e.m.; two-sided unpaired Student’s t-test. ***P < 0.001, **P < 0.01, *P < 0.05, n.s. P > 0.05. Exact P values and defined sample sizes (n) are provided in Supplementary Data 1.

Extended Data Fig. 5 |. DNA-PKcs^5A/5A MEFs and ES cells are not hypersensitive to DNA-damaging agents.

a, b, DNA-PKcs^5A/5A MEFs are not hypersensitive to ionizing radiation (IR) or mitomycin C (MMC). Fancd2^−/− cells are included as positive controls. c–e, Representative colony formation assays (c) and quantification of sensitivity to MMC (d) and IR (e) in DNA-PKcs^5A/5A and control ES cells. Mean ± s.e.m. of a representative experiment in triplicate. Two-sided unpaired Student’s t-test; P > 0.05 (n.s). f, Normal proliferation curve of DNA-PKcs^5A/5A MEFs. g, DNA-PKcs^5A/5A MEFs have normal S phase frequency. The accumulation of aneuploid cells is unique to Tp53^−/− cells. a, d–f, Mean ± s.e.m. Exact P values and defined sample sizes (n) are provided in Supplementary Data 1.

Extended Data Fig. 6 |. DNA-PKcs^5A/5A ES cells have KU-dependent proliferation defects that cannot be fully explained by cNHEJ or telomere defects.

a, Representative colony formation assay for DNA-PKcs^5A/5A ES cells. Note the frequent accumulation of small colonies in DNA-PKcs^5A/5A cells, which is rescued by deletion of KU80. b, Quantification of colony size (A.U., arbitrary units) from DNA-PKcs^5A/5A and control ES cells. Two independently derived ES cells were assayed. c, Cell cycle analyses of two independently derived DNA-PKcs^5A/5A and control ES cell lines. The percentage of BrdU⁺ cells is shown. d, Frequency of metaphase with telomere abnormalities (see below for definition). e, Representative telomere fluorescence in situ hybridization (FISH) images of a normal mouse chromosome (top, with four telomere dots), a chromosome with a chromatid break (middle, showing loss of one telomere signal among the two sister chromatids), or a chromatid fusion without telomere signal (bottom). f, Quantitative analyses of telomere instability and chromosomal breaks in metaphase. Telomere FISH analyses of MEFs were performed with the telomere-specific PNA probe as previously described. Normal mouse chromosomes have four discrete telomere signals (e, top). Telomere instability or breaks considered include: i) telomere instability (indicated by more than one telomere signal per chromatid), ii) telomere/chromosome fusion (e, bottom; with telomere at the fusion junction (telomere fusion) or without telomere signal at the fusion junction (non-telomere fusion)), iii) chromosome breaks (S.B.; loss of both telomere signals on the paired sister chromatids) and iv) chromatid breaks (T.B.; loss of one of the two chromatids (e, middle)). The number of metaphases with at least one telomere instability, break or fusion is shown in d as a percentage of metaphases with abnormalities. Data derived from four independent MEF lines of each genotype. g, Representative flow cytometry analyses of erythroblasts from age-matched (2 weeks) DNA-PKcs^+/+ and DNA-PKcs^−/− mice. h, Representative protein translation analyses of S1 (CD71⁺Ter119⁻), S2 (CD71⁺Ter119^mid) and S3 erythroblasts (CD71⁺Ter119^high) from 2-week-old DNA-PKcs^+/+, DNA-PKcs^−/− and Ku70^−/− mice. Quantification is shown in Fig. 2i. b, d, Mean ± s.e.m.; two-sided unpaired Student’s t t-test, **P < 0.01, *P < 0.05, n.s. P > 0.05. Exact P values and defined sample sizes (n) are provided in Supplementary Data 1.

Extended Data Fig. 7 |. Global translation defects in DNA-PKcs^5A/5A haematopoietic cells.

a, MCV of RBCs from 2-week-old DNA-PKcs^5A/5A mice with or without TP53 deficiency. Wild-type and 5A/5A data from Fig. 2e included for comparison. b, Relative percentage of S1, S2 and S3 cells among all erythroblasts in fetal liver from DNA-PKcs^5A/5A or DNA-PKcs^+/+ E14.5 embryos. S4 and S5 populations have not yet evolved significantly in the E14.5 fetal liver. c, The relative frequencies of S5, the most mature erythroblast, in P14 bone marrow from DNA-PKcs^5A/5A or DNA-PKcs^+/+ mice. d, Relative OP-puro levels in P14 wild-type bone marrow haematopoietic cells. The average of total bone marrow was set to 1. The erythroblasts (S1, S2 and S3^high) clearly have the highest OP-Puro levels. As shown in Fig. 2h, S3 erythroblasts have an OP-puro^high and an OP-puro^low population. The mean OP-puro levels of both populations are included. e, Representative global protein translation in S1, S2 and S3 erythroblasts from E14.5 DNA-PKcs^+/+ and DNA-PKcs^5A/5A fetal liver. f, Frequency of OP-Puro^low among S1, S2, and S3 erythroblasts in E14.5 fetal liver. OP-Puro labelling of fetal liver cells was performed for 30 min. g, Quantification of the frequency of OP-Puro^low among S1and S2 erythroblasts from 2-week-old DNA-PKcs^5A/5ATp53^+/− mice. h, Relative OP-puro levels (normalized to the levels in untreated DNA-PKcs^+/+ B cells) of DNA-PKcs^KD/KD and DNA-PKcs^3A/3A B cells. Unpaired Student’s t-test, ***P < 0.001, *P < 0.05. i, Relative OP-puro levels of DNA-PKcs^+/+ B cells treated with ATM kinase inhibitor (KU55933, 15 μM for 17 h). Data represent three independent biological experiments (P = 0.38 for unpaired Student’s t-test). a–d, f–i, Two-sided unpaired Student’s t-test, ***P < 0.001, **P < 0.01, *P < 0.05, n.s. P > 0.05. All graphs show mean ± s.e.m. Exact P values and defined sample sizes (n) are provided in Supplementary Data 1.

Extended Data Fig. 8 |. rRNA synthesis-dependent localization of DNA-PK in nucleoli of human and mouse cells, and U3 ChIRP-MS of SSU processome.

a, b, Immunofluorescence staining of endogenous KU86 (a) and DNA-PKcs (b) in U2OS cells. DDX21 RNA helicase is used as a positive control for nucleoli. The CSK buffer contains Triton X-100 for pre-extraction before fixation (see Methods). When indicated, the cells were treated with 50 nM ActD for 1 h before pre-extraction, fixation and staining. c, Localization of ectopically expressed GFP-tagged KU70 in mouse ES cells. a–c, n = 3 biologically independent experiments. d, U3 ChIRP-qRT–PCR analysis from HeLa cells. Enrichment levels, relative to input samples, of the U3, 7SK, 18S, and RMRP RNAs were assessed from experimental (−RNase A) or control (+RNase A) ChIRP samples. Data are from two independent biological replicates. e, DNA-PK was also recovered from U3 ChIRP-MS in IMR90 cells. Peptide spectral match (PSM) counts for control (RNase A) and experimental (U3) samples are shown. n = 2 biological replicates.

Extended Data Fig. 9 |. KU associates with SSU processome components.

a, KU86 IP-MS and U3 ChIRP-MS overlap. Zhou et al. (n = 2 independent MS assays; for each assay, n = 2 technical replicates) identified 292 proteins enriched with KU86 protein. U3 ChIRP-MS identified 483 proteins enriched more than twofold with the U3 snoRNA (Supplementary Table 1). These identified factors were intersected, resulting in 153 proteins in common between the two affinity purification strategies, which is highly significant (hypergeometric P < 6.650 × 10⁻¹⁶⁶). To understand what types of protein were enriched only with KU86, only with U3, or together with both factors, we isolated the enrichment values (−log₁₀(Benjamini)) for GO biological process terms for each of these sets and compared them. Factors commonly bound were biased for rRNA processing, ribosomal terms and SSU biogenesis. U3-specific factors had additional enrichment in these categories and KU86 had a set of unique terms that were not well represented in U3. b, Independent repeat of northern blot analyses of 18S rRNA maturation in v-ABL kinase-transformed B cells from noted genotypes. The probe covers the sequence just after the 18S rRNA (red line). This experiment was repeated independently four times. Another repeat is shown in Fig. 3c. c, d, MS characterization of commercial DNA-PK holoenzyme (Promega) used in EMSA and kinase reactions. A detailed description of proteins and their quality in this mixture has not been published. We subjected the DNA-PK enzyme mix as provided to SDS–PAGE separation followed by LC–MS identification of proteins from mass ranges between 65 and 600 kDa. KU70 and KU86 were clearly present in the gel and via MS. For masses above 130 kDa, DNA-PKcs was the major protein identified. For each of the five slices analysed (coloured regions) we tabulated the starting positions of peptides matching the DNA-PKcs polypeptide and mapped them to the position within DNA-PKcs. As expected, in the highest-molecular-weight slices, we identified peptides across the majority of the length of DNA-PKcs. DNA-PKcs peptides were present in lower slices, but poorer overall coverage was evident, suggesting that these are degradation products. As this was a confirmatory experiment of a validated and commercially available product (see Methods), it was conducted only once.

Extended Data Fig. 10 |. DNA-PKcs and KU86 bind nuclear and nucleolar RNAs in living cells.

a, b, irCLIP of DNA-PKcs (a) and KU86 (b) in HeLa cells. RNA (IR800) and protein blots show specific enrichment of RNA UV-crosslinked to DNA-PKcs and KU86. The increased signal in the RNaseA-treated samples corresponds to the accumulation of RNaseA-protected fragments that are directly associated with either KU86 or DNA-PKcs. Results shown are representative of three biologically independent experiments. Vertical line marks the RNA extracted for sequencing. c, e, Fractional distribution of DNA-PKcs (c) and KU86 (e) irCLIP RT stops on major genomic features annotated with HOMER. d, f, Detailed HOMER annotation of DNA-PKcs (d) and KU86 (f) RT stops categorized as intronic and intergenic from c and e. g, h, Genome browser tracks of DNA-PKcs and KU86 irCLIP on the Neat1 (g) and Terc (h) RNA. RT stop values are normalized to 10⁷ mapped reads. i, j, Histograms of DNA-PKcs (i) and KU86 (j) irCLIP RT stops mapping to the 5′ETS, normalized for total reads mapped per experiment. Data from DMSO- (top) or ActD-treated (bottom) HeLa cells are shown. Grey highlighted region is as in Fig. 4a.

Extended Data Fig. 11 |. DNA-PK interacts with structured RNAs that can activate its kinase.

a, b, Correlation analysis of total RT stops mapping to non-repeat snoRNA transcripts from DDX21 compared to DNA-PKcs (a) or KU86 (b) irCLIP experiments from DMSO-treated HeLa cells. Correlation analysis was performed using Pearson’s correlation coefficient. n denotes number of snoRNA transcripts bound by each protein. c, EMSA of purified human DNA-PK and in vitro transcribed U3-SL1. Lane 1 contains only Cy7-labelled U3-SL1. Lanes 2–4 show that KU assembles with U3-SL1 at a 1:3 molar ratio, while DNA-PK holoenzyme assembly occurs at a 1:25 molar ratio. Lanes 5 and 6 show that unlabelled U3-SL1 RNA competes away bound labelled U3-SL1 in a dose-dependent manner. d, Supershift EMSA of DNA-PK and U3-SL1 RNA with KU86 antibody. The addition of anti-KU86 confirms the identity of the KU–U3-SL1 band and also shifts up the complex to higher molecular weights. e, A structural mutant of U3-SL1 was generated by introducing point mutations predicted to disrupt the stem-loop structure. This mutant was unable to compete away wildtype U3-SL1 for binding to the KU complex, while unlabelled wild-type U3-SL1 competed efficiently. f, DNA-PK in vitro kinase phosphorylation assay in the presence of increasing amounts of U3-SL1 or DNA. Western blot was performed with an antibody recognizing DNA-PKcs phosphorylated at the T2609 cluster. Asterisks denote cross-reactive fragments that probably include phosphorylated DNA-PKcs fragments, on the basis of MS analyses of the DNA-PK complex (Extended Data Fig. 9c, d). g, As in f, but using an antibody recognizing DNA-PKcs phosphorylated at the S2056 cluster. h–j, As in f with the following changes. h, dsDNA was used to activate DNA-PK, NU7441 was included to inhibit specific DNA-PK activity, and western blot analysis monitored the total DNA-PK (total DNA-PKcs) or phosphorylated DNA-PK (DNA-PKcs phoT2609). i, U3-SL1 RNA was used to activate DNA-PK in the absence or presence of the DNA-PK inhibitor NU7441. j, U3-SL1 RNA was used to activate DNA-PK, hydrolysable (ATP) or non-hydrolysable (AppCp) ATP was provided, and western blot analysis monitored KU86 (loading control) or phosphorylated DNA-PK (phoT2609). k, Baculovirus-purified human DNA-PK in vitro kinase phosphorylation assay in the presence of increasing amounts of U3-SL1 or DNA. Western blot was performed with antibodies recognizing DNA-PKcs phosphorylated at the S2056 cluster (top), total DNA-PK (middle) and KU86 (bottom). All EMSA and western blots presented here are representative of three biologically independent experiments.

Fig. 1 |. Kinase dead or phosphorylation deficient DNA-PKcs causes haematopoietic defects independent of cNHEJ.

a, Kaplan–Meier survival curve of DNA-PKcs^KD/KDTp53^−/− and control (DNA-PKcs^KD/KDTp53^+/−, DNA-PKcs^−/−Tp53^−/−, Xrcc4^−/−Tp53^−/− and Tp53^−/−) mice. Log-rank (Mantel–Cox) test; ***P < 0.001, NS (not significant) P > 0.05. b, Cause of death for tumour cohorts by genotype. c, The life expectancy for mice of different genotypes plotted by tumour type. Pro-B cell lymphomas in DNA-PKcs^KD/KDTp53^−/− mice have similiar latency as in DNA-PKcs^−/−Tp53^−/− and Xrcc4^−/−Tp53^−/− mice (two-sided unpaired Student’s t-test; ***P < 0.001, NS P > 0.05). Data shown as mean ± s.e.m. d, Representative histological analyses of bone marrow and spleen from 3-week-old DNA-PKcs^KD/KDTp53^−/− mice. The samples were stained for the myeloid marker myeloperoxidase before being counterstained with haematoxylin (for nuclei). Arrows indicate megakaryocytes. Scale bars, 100 μm. e, Representative histology analyses of bone marrow from 2-week-old DNA-PKcs^+/+ and DNA-PKcs^5A/5A mice with or without Tp53 deficiency. Scale bars, 100 μm. Exact P values and defined sample sizes (n) are provided in Supplementary Data 1.

Fig. 2 |. Mutations in DNA-PKcs cause KU-dependent haematopoietic failure and translation defects.

a, Kaplan–Meier survival curve of DNA-PKcs^5A/5A, DNA-PKcs^5A/5ATp53^+/−, DNA-PKcs^5A/5ATp53^−/−, DNA-PKcs^5A/5AKu70^−/− and Ku70^−/− mice. b, Ku70 deficiency rescued the peripheral RBC counts of 2-week-old DNAPKcs^5A/5AKu70^−/− mice. c, d, Representative flow cytometry analyses (c) and the absolute number of haematopoietic stem and progenitor cells (Lin⁻SCA1⁺c-KIT⁺) (d) from 2-week-old DNA-PKcs^5A/5A and DNA-PKcs^5A/5AKu70^−/− mice. e, MCV of RBCs from 2-week-old DNA-PKcs^5A/5A and DNA-PKcs^5A/5AKu70^−/− mice. f, Representative flow cytometry analyses of bone marrow RBCs from 2-week-old DNA-PKcs^5A/5A and DNA-PKcs^5A/5AKu70^−/− mice. CD71 and TER119 staining determines the stage of RBC differentiation. S1, CD71⁺TER119^low; S2, CD71⁺TER119^mid; S3, CD71⁺TER119^high; S4, CD71^midTER119^high; S5, CD71⁻TER119^high. Forward scatter of TER119⁺ RBCs is shown (bottom). g, Relative frequencies of S2, S3 and S4 among all immature erythroblasts (S1–S4). The S5 percentage is much higher in DNA-PKcs^5A/5A mice (Extended Data Fig. 7c). h, Measurements of translation in S1, S2 and S3 erythroblasts from P14 DNA-PKcs^+/+ and DNA-PKcs^5A/5A mice bone marrow. i, Quantification of the frequency of OP-Puro^low among S1, S2 and S3 erythroblasts by OP-Puro labeling of bone marrow cells for 1 h. j, Inhibition of DNA-PKcs with NU7441 leads to a DNA-PKcs- and Ku-dependent reduction in global translation. Graphs show fold change in OP-Puro fluorescence (mean ± s.d.) normalized to DNA-PKcs^+/+ cells (set to 100%). Wild-type with or without NU7441 (44% reduction with NU7441); DNA-PKcs^−/− with or without NU7441 (~1% reduction with NU7441), and Ku70^−/− with or without NU7441 (~10% reduction with NU7441). b, d, e, g, i, Mean ± s.e.m. a, Two-sided log-rank (Mantel–Cox) test, ***P < 0.001. j, Two-sided paired Student’s t-test; all other panels, two-sided unpaired Student’s t-test, ***P < 0.001; **P < 0.01; *P < 0.05; NS, P > 0.05. Exact P values and defined sample sizes (n) are provided in Supplementary Data 1.

Fig. 3 |. DNA-PK, but not other cNHEJ factors, co-purifies with the U3 snoRNA and regulates rRNA processing.

a, A protein–protein interaction (grey lines) network of U3 ChIRP-MS from HeLa cells of nucleolar and ribosomal biogenesis factors. Colours map to the enrichment log₂[ChIRP/control] value of each protein. Known complexes are physically grouped and labelled. Complete hit list detailed in Supplementary Table 1. b, Cellular compartment term analysis of the U3 ChIRP-MS from HeLa cells using the DAVID tool. False discovery rate (FDR) was estimated (Benjamini corrected P) by comparing the 483 enriched proteins to the human proteome; the top ten most enriched terms are shown. c, d, Northern blot analyses of 18S and 5.8S rRNA maturation in v-ABL kinase transformed B cells from noted genotypes. The probe covers the 3′ edge of 18S rRNA and 5.8S rRNA, respectively (blue line). The normalized relative intensities of 21S/34S, 18S/21S and 12S/32S are marked below the gels. n = 4 (c), n = 3 (d) biologically independent experiments.

Fig. 4 |. U3 snoRNA drives assembly, activation, and auto-phosphorylation of DNA-PK at the T2609 cluster.

a, DNA-PKcs and KU86 RT stops that map to the four pre-rRNA introns; 5′ETS, ITS1, ITS2, and 3′ETS. irCLIP was performed in biological duplicate. b, Transcript normalized histogram of DNA-PKcs (orange) and KU86 (blue) irCLIP RT stops from DMSO-treated HeLa cells mapping to U3. c, In silico (mFold)-predicted secondary structure of U3. The top three and two peaks of DNA-PKcs and KU86, respectively, are annotated on the secondary structure. The length of the 5′end stem-loop is shown. d, Baculovirus-purified human DNA-PK in vitro kinase phosphorylation assay in the presence of increasing amounts of U3-SL1. Western blots were performed with antibodies against DNA-PKcs phosphorylated at the T2609 cluster (top), total DNA-PK (middle), and KU86 (bottom). e, Baculovirus-purified human DNA-PK in vitro kinase phosphorylation assay against purified human TP53. Activated DNA-PK phosphorylates TP53 on serine 15. An in vitro ATM kinase assay was used as a control. Western blots were performed with antibodies against TP53 phosphorylated on serine 15 (top), total TP53 (middle) and KU86 (bottom). d, e, n = 3 biologically independent experiments.