Shieldin and CST co-orchestrate DNA polymerase-dependent tailed-end joining reactions independently of 53BP1-governed repair pathway choice

Abstract

Tumor suppressor p53-binding protein 1 (53BP1) regulates DNA end joining in lymphocytes, diversifying immune antigen receptors. This involves nucleosome-bound 53BP1 at DNA double-stranded breaks (DSBs) recruiting Rap1-interacting factor 1 homolog (RIF1) and shieldin, a poorly understood DNA-binding complex. The 53BP1-RIF1-shieldin axis is pathological in BRCA1-mutated cancers, blocking homologous recombination (HR) and driving illegitimate nonhomologous end joining (NHEJ). However, how this axis regulates DNA end joining and HR suppression remains unresolved. We investigated shieldin and its interplay with the Ctc1-Stn1-Ten1 (CST) complex, which was recently implicated downstream of 53BP1. Immunophenotypically, mice lacking shieldin or CST are equivalent, with class-switch recombination coreliant on both complexes. Ataxia-telangiectasia mutated kinase-dependent DNA damage signaling underpins this cooperation, inducing physical interactions between these complexes that reveal shieldin as a DSB-responsive CST adaptor. Furthermore, DNA polymerase ζ functions downstream of shieldin, establishing DNA fill-in synthesis as the physiological function of shieldin-CST. Lastly, we demonstrate that 53BP1 suppresses HR and promotes NHEJ in BRCA1-deficient mice and cells independently of shieldin. These findings showcase the versatility of the 53BP1 pathway, achieved through the collaboration of chromatin-bound 53BP1 complexes and DNA end-processing effector proteins.

Fig. 1. Equivalent immunodeficiencies in mice lacking shieldin and CST.

a, Schematic representation of the Shld2 and Shld3 (knockout) and Ctc1 (conditional) alleles generated in this study. TSS, transcription start site. b, Absolute numbers of B220⁺ B cells in the bone marrow (one femur and one tibia) and spleen (n = 4–8 mice per genotype, where each data point is a single mouse). Significance was determined by an unpaired two tailed t-test (mean ± s.e.m.). NS, not significant. c, Absolute numbers of B cell precursors (Hardy fraction A, B220⁺CD43⁺BP1⁻CD24⁻; Hardy fraction B, B220⁺CD43⁺BP1⁻CD24⁺; Hardy fraction C, B220⁺CD43⁺BP1⁺CD24⁺; Hardy fraction D, B220⁺CD43⁻IgM⁻IgD⁻; Hardy fraction E, B220⁺CD43⁻IgM⁺IgD⁻; Hardy fraction F, B220⁺CD43⁻IgM⁺IgD⁺) in the bone marrow (one femur and one tibia) (n = 4–8 mice per genotype, where each data point is a single mouse). Significance was determined by an unpaired two tailed t-test (mean ± s.e.m.). d, Splenic B cells cultured with the indicated stimuli (96 h) and stained for surface IgG1, IgE, IgG2b or IgG3 (n = 4–7 mice per genotype, where each data point is a single mouse). CSR 100%, mean immunoglobulin isotype switch frequency of two control animals in each experiment. Significance was determined by a two-way analysis of variance (ANOVA) with Tukey’s correction (mean ± s.e.m.). e, CTV-labeled splenic B cells were stimulated as indicated and stained for surface IgG1 after 96 h. Representative data, n > 6 mice. f, IgG1⁺ B cells as a proportion of total B cells (%) for each cell generation as determined by CTV staining and proliferation-associated dye dilution (n = 4–6 mice per genotype, where each data point is a single mouse; mean ± s.e.m.). g, CTV dilution in purified B cells cultured in the presence of LPS and IL-4 for 96 h. Representative data, n > 6 mice. h, NP-specific serum IgM (left) and IgG1 (right) at indicated times after NP-CGG immunization. Representative data, n = 2 independent experiments, each with four mice. Significance was determined by an unpaired two tailed t-test (mean ± s.e.m.). Gating strategies for the above flow cytometry data are provided in Supplementary Fig. 1. Source data

Fig. 2. DNA damage induces interactions between shieldin and CST.

a, Schematic depicting shieldin purification and proteomic elucidation strategy. Briefly, Shld1/3-knockout CH12-F3 cell lines were lentivirally transduced with TwinStrep–HA–Shld1/3 espression transgenes. First, 300-ml suspension cultures were either mock-treated or irradiated (10 Gy) and lysates were prepared following a 1-h recovery. Shieldin complexes captured from lysates on MagStrep-XT resin were eluted in biotin, captured and tryptic-digested on S-Trap columns, with resulting peptides analyzed by LC–MS/MS. The schematic was generated using BioRender.com. b, B cell SHLD3 versus control (beads only) interactomes as defined by LC–MS/MS and LFQ. The scatter plot depicts the log₂ fold enrichment of indicated MagStrep-XT purified complexes across two independent experiments. c, Comparison of mock and irradiated B cell SHLD3 interactomes as defined by LC–MS/MS and LFQ. The scatter plot depicts the log₂ fold enrichment of indicated MagStrep-XT purified complexes across two independent experiments. d, Immunoblot analysis of CH12-F3 cell extracts from SHLD3 pulldown experiments. Representative of n = 2 independent experiments. Source data

Fig. 3. Pol ζ cooperation with shieldin supports NHEJ in CSR.

a, Schematic representation of the Rev3l conditional allele. b, Absolute numbers of B220⁺ B cells in the bone marrow (one femur and one tibia) and spleen (n = 4–6 mice per genotype, where each data point is a single mouse). Significance was determined by an unpaired two tailed t-test (mean ± s.e.m.). c, CTV dilution in purified B cells cultured in the presence of LPS and IL-4 for 96 h. Representative data, n > 6 mice. d, IgG1⁺ B cells as a proportion of total B cells (%) for each cell generation as determined by CTV staining and proliferation-associated dye dilution (n = 4–6 mice per genotype, where each data point is a single mouse; mean ± s.e.m.). e, Splenic B cells cultured with the indicated stimuli (96 h) and stained for surface IgG1, IgE, IgG2b or IgG3 (n = 4–7 mice per genotype, where each data point is a single mouse). CSR 100%, mean immunoglobulin isotype switch frequency of two control animals in each experiment. Significance was determined by a two-way ANOVA with Tukey’s correction (mean ± s.e.m.). f, CTV-labeled splenic B cells were stimulated as indicated and stained for surface IgG1 after 96 h. Representative data, n > 6 mice. g, Splenic B cells cultured with the indicated stimuli (96 h) and stained for surface IgG1, IgE, IgG2b or IgG3 (n = 4–7 mice per genotype, where each data point is a single mouse). CSR 100%, mean immunoglobulin isotype switch frequency of two control animals in each experiment. Significance was determined by a two-way ANOVA with Tukey’s correction (mean ± s.e.m.). h, Splenic B cells cultured with the indicated stimuli (96 h) and stained for surface IgG1, IgE, IgG2b or IgG3 (n = 4–7 mice per genotype, where each data point is a single mouse). CSR 100%, mean immunoglobulin isotype switch frequency of two control animals in each experiment. Significance was determined by a two-way ANOVA with Tukey’s correction (mean ± s.e.m.). Source data

Fig. 4. REV7-independent and NHEJ-independent functions of Pol ζ support B cell development.

a, Absolute numbers of B cell precursors (Hardy fraction A, B220⁺CD43⁺BP1⁻CD24⁻; Hardy fraction B, B220⁺CD43⁺BP1⁻CD24⁺; Hardy fraction C, B220⁺CD43⁺BP1⁺CD24⁺; Hardy fraction D, B220⁺CD43⁻IgM⁻IgD⁻; Hardy fraction E, B220⁺CD43⁻IgM⁺IgD⁻; Hardy fraction F, B220⁺CD43⁻IgM⁺IgD⁺) in the bone marrow (one femur and one tibia) (n = 4–6 mice per genotype, where each data point is a single mouse). Significance was determined by an unpaired two tailed t-test (mean ± s.e.m.). b, Absolute numbers of B220⁺ B cells in the bone marrow (one femur and one tibia) and spleen (n = 6–12 mice per genotype, where each data point is a single mouse). Significance was determined by an unpaired two tailed t-test (mean ± s.e.m.). c, Absolute numbers of B cell precursors in the bone marrow (one femur and one tibia) (n = 6–12 mice per genotype, where each data point is a single mouse). Significance was determined by an unpaired two tailed t-test (mean ± s.e.m.). d, Absolute numbers of mature IgM⁺IgD⁺ B cells in the spleen (n = 4 mice per genotype, where each data point is a single mouse). Significance was determined by an unpaired two tailed t-test (mean ± s.e.m.). Source data

Fig. 5. 53BP1-dependent HR suppression in BRCA1-deficient cells is shieldin independent.

a, Representative image of 53bp1^−/−Brca1^−/− mice and relevant controls at ~12 weeks. Genotypes are indicated. b, Representative image of a Shld2^−/−Brca1^+/− control embryo and Shld2^−/−Brca1^−/− double-knockout embryo at E10.5. c, Survival of the indicated MEF cell lines grown for 7 days in the presence of the indicated doses of olaparib (n = 3 biological experiments; mean ± s.d.). d, Immunofluorescence microscopy of Rad51 IRIF in indicated MEF cell lines. Cells were irradiated (5 Gy) and fixed 2 h later. Representative images of n = 3 biological experiments. e, Quantification of RAD51 IRIF in Edu⁺ cells from d. Integrated intensity and foci quantifications were made using CellProfiler. Boxes indicate the 25th–75th percentiles with the median denoted and whiskers indicate the 10th–90th percentiles. Significance was determined by a two-sided Kruskal–Wallis H test with Dunn’s correction for multiple comparisons. ****P < 0.0001 and **P < 0.005 (n = 3 biological experiments). f, Survival of the indicated BARD1^AID/AID HCT-116 cell lines grown for 7 days in the presence or absence of IAA (1 mM), doxycycline (2 mg ml⁻¹) and the indicated doses of olaparib (n = 3 biological experiments; mean ± s.d.). g, Immunofluorescence microscopy of Rad51 IRIF in indicated BARD1^AID/AID cell lines. Cultures were supplemented with doxycycline (2 mg ml⁻¹ for 24 h) before the addition of IAA (1 mM). After 24 h, cells were irradiated (5 Gy) and fixed 2 h later. Representative images of n = 3 biological experiments. h, Quantification of RAD51 IRIF in Edu⁺ cells from g. Integrated intensity and foci quantifications were made using CellProfiler. Boxes indicate the 25th–75th percentiles with the median denoted and whiskers calculated by the Tukey method. Significance was determined by a two-sided Kruskal–Wallis H test with Dunn’s correction for multiple comparisons (n = 4 biological experiments). Source data

Fig. 6. Independent and synergistic contributions of 53BP1 and shieldin to genomic instability in BRCA1-deficient cells.

a, Representative metaphase images from the indicated BARD1^AID/AID cell lines. Cultures were supplemented with doxycycline (2 mg ml⁻¹for 24 h) before the addition of IAA (1 mM). After a further 24 h, cells were treated with 500 nM olaparib and fixed 24 h later. Representative images of n = 3 biological experiments. b, Quantification of 40–65 metaphases per genotype per experiment. Significance was determined by a two-sided Kruskal–Wallis H test with Dunn’s correction for multiple comparisons. Horizontal bars indicate the mean values ± 95% confidence interval. ****P < 0.0001, ***P < 0.0005, **P < 0.005 and *P < 0.05 (n = 3 biological experiments). c, Model for 53BP1-dependent DSB repair. Chromatin-associated 53BP1 complexes suppress HR and promote V(D)J recombination. DNA damage-dependent signaling by RIF1 is conferred to shieldin through interactions with SHLD3. DSB signaling by ATM stimulates the assembly of shieldin–CST complexes on ssDNA, promoting the recruitment of Pol α–primase to initiate fill-in synthesis. Pol ζ REV3L increases the processivity of Pol α–primase-catalyzed fill-in synthesis. Source data

Extended Data Fig. 1. Equivalent immunodeficiencies in mice lacking shieldin and CST.

a) Representative image of PCR amplicons from genomic DNA obtained by ear biopsies from mice of the indicated genotype. Bands of different size correspond to the Shld2^tm1a allele (274 bp), Shld2^tm1c allele (554 bp), Shld2^tm1d allele (240 bp) and WT allele (391 bp). n = >20 mice per genotype. b) Representative image of PCR amplicons from genomic DNA obtained by ear biopsies from mice of the indicated genotype. Bands of different size correspond to the Shld3^EM1-B6 allele (left) or Shld3^EM2-B6 allele (right) (~312 bp) and WT allele (459 bp). n = >20 mice per genotype. c) Representative image of PCR amplicons from genomic DNA obtained by ear biopsies (left), splenic b cells or purified splenic b cells (right) from mice of the indicated genotype. Bands of different size correspond to the Ctc1^fl allele (413 bp), Ctc1^−/− allele (658 bp) and WT allele (210 bp). n = >20 mice per genotype except for purified B cells; n = 2 independent experiments each with 2-3 mice. d) Peripheral blood stained as indicated. Reticulocytes (RET), micronucleated reticulocytes (MN-RET), normochromatic erythrocytes (NCE) and micronucleated normochromatic erythrocytes (MN-NCE). Representative data, n ³ 5 mice of each sex per genotype. e) MN-NCE as a percentage of total NCEs from tail vein blood of mice at 8-12 weeks. n = 5-12 mice of each sex per genotype, where each data point is a single mouse. Significance was determined by unpaired two tailed t-test, Mean ± SEM. f) Flow cytometric sub-classification of nature splenic b cell fractions. n = 4-8 mice per genotype, where each data point is a single mouse. Significance was determined by unpaired two tailed t-test, Mean±SEM. g) Absolute numbers of B220⁺ B cells in the bone marrow (one femur and one tibia) and spleen and live lymphocytes in the thymus. n = 10-18 mice per genotype, where each data point is a single mouse. Significance was determined by unpaired two tailed t-test, Mean ± SEM. h) Absolute numbers of B cell precursors in the bone marrow (one femur and one tibia). n = 15-19 mice per genotype, where each data point is a single mouse. Significance was determined by unpaired two tailed t-test, Mean±SEM. i) Flow cytometric sub-classification of mature splenic b cell fractions. n = 6-17 mice per genotype, where each data point is a single mouse. Significance was determined by unpaired two tailed t-test, Mean ± SEM. j) Absolute numbers of thymic T-cells, n = 9-17 mice per genotype, where each data point is a single mouse. Mean ± SEM. k) Absolute numbers of T cell precursors in double negative (CD4- CD8-) in the thymus, n = 9-17 mice per genotype, where each data point is a single mouse. Mean ± SEM. l) CTV-labelled purified B cells were stimulated as indicated and stained for surface IgG/IgE on day 4. Representative of n > 6 experiments. m) Splenic B cells cultured with the indicated stimuli (96 h) and stained for surface IgG1, IgE, IgG2b or IgG3. n = 11-14 mice per genotype. CSR 100%, mean immunoglobulin isotype switch frequency of 2 control animals in each experiment. Significance was determined by two-way ANOVA with Tukey’s correction. Mean ± SEM. n) NP-specific serum IgM (left) and IgG1 (right) at indicated times after NP-CGG immunization. AU, arbitrary units. Representative data, n = 2 independent experiments, each with 2-3 mice. Significance was determined by unpaired two tailed t-test. Mean ± SEM. Source data

Extended Data Fig. 2. DNA damage induces interactions between shieldin and CST.

a) IgM-to-IgA CSR frequencies in Shld3 knockout CH12-F3 cell lines, n = 3 independent experiments, where each data point is the average of 3 technical replicates from a single experiment. Mean ± SEM. Representative flow cytometry plots for IgM-to-IgA CSR. Representative data, n > 3 independent experiments. b) IgM-to-IgA CSR frequencies in Shld1 knockout CH12-F3 cell lines, n = 3 independent experiments, where each data point is the average of 3 technical replicates from a single experiment. Mean ± SEM. Representative flow cytometry plots for IgM-to-IgA CSR. Representative data, n > 3 independent experiments. c) B cell Shld1 interactome as defined by LC–MS/MS and LFQ. Scatter plot depicts log₂ fold-enrichment of indicated TwinStep–immunocomplexes across 2 independent experiments. d) B cell Shld1 IR-dependent interactome as defined by LC–MS/MS and LFQ. Scatter plot depicts log₂ fold-enrichment of indicated TwinStep–immunocomplexes across 2 independent experiments. e) Western blot analysis of CH12-F3 cell extracts from pulldown experiments. Cells were pre-treated with ATM, ATR inhibitor or both for 1 hr prior to irradiation. Representative of n = 2 independent experiments. f) Western blot analysis of CH12-F3 cell extracts from pulldown experiments. Representative of n = 2 independent experiments. g) Splenic B cells cultured with the indicated stimuli (96 h) and stained for surface IgG1, IgE, IgG2b or IgG3. n = 3-5 mice per genotype. CSR 100%, mean immunoglobulin isotype switch frequency of 2 control animals in each experiment. Significance was determined by two-way ANOVA with Tukey’s correction. Mean ± SEM. h) CTV-labelled purified B cells were stimulated as indicated and stained for surface IgG on day 4. Representative of n = 2 experiments. Source data

Extended Data Fig. 3. Pol ζ cooperation with shieldin supports NHEJ in CSR.

a) Representative image of PCR amplicons from genomic DNA obtained by ear biopsies (left), splenic b cells or purified splenic b cells (right) from mice of the indicated genotype. Bands of different size correspond to the Rev3l^fl allele (458 bp), Rev3l^−/− allele (300 bp) and WT allele (384 bp). n = >20 mice per genotype except for purified B cells; n = 2 independent experiments each with 2-3 mice. b) CTV-labelled purified B cells were stimulated as indicated and stained for surface IgG/IgE on day 4. Representative of n > 5 experiments.

Extended Data Fig. 4. REV7-independent and NHEJ-independent functions of Pol ζ support B cell development.

a) Flow cytometric sub-classification of nature splenic b cell fractions. n = 4-6 mice per genotype, where each data point is a single mouse. Significance was determined by unpaired two tailed t-test, Mean±SEM. b) Flow cytometry analysis of B cell development in the bone marrow of Mb1^+/Cre, Rev7^f/lf Mb1^{+/Cre or} Rev3l^fl/fl Mb1^+/Cre mice; gating on B220⁺CD43⁺ (top, Hardy fractions A, B and C) and on B220⁺CD43⁻ (bottom, Hardy fractions D, E and F). Representative data; n = 3 experiments. c) Flow cytometry analysis of mature b cell populations in the spleen of Mb1^+/Cre, Rev7^f/lf Mb1^{+/Cre or} Rev3l^fl/fl Mb1^+/Cre mice; gating on B220⁺CD19⁺. Representative data; n = 3 experiments. d) Representative image of PCR amplicons from genomic DNA obtained by ear biopsies (left), splenic b cells or purified splenic b cells (right) from mice of the indicated genotype. Bands of different size correspond to the p53^fl allele (353 bp), p53^−/− allele (450 bp) and WT allele (200 bp). n = >20 mice per genotype except for purified B cells; n = 2 independent experiments each with 2-3 mice. e) Flow cytometric sub-classification of nature splenic b cell fractions. n =6-12 mice per genotype, where each data point is a single mouse. Significance was determined by unpaired two tailed t-test, Mean±SEM. f) Flow cytometry analysis of B cell development in the bone marrow of Rev^+/+ p53^fl/fl Mb1^{+/Cre or} Rev3l^fl/fl p53^fl/fl Mb1^+/Cre mice; gating on B220⁺CD43⁺ (top, Hardy fractions A, B and C) and on B220⁺CD43⁻ (bottom, Hardy fractions D, E and F). Representative data; n = 3 experiments. g) Flow cytometry analysis of mature b cell populations in the spleen of Rev^+/+ p53^fl/fl Mb1^{+/Cre or} Rev3l^fl/fl p53^fl/fl Mb1^+/Cre mice; gating on B220⁺CD19⁺. Representative data; n = 3 experiments. h) Flow cytometry analysis of B cell development in the bone marrow of MD4^+/Tg, Rev3^fl/fl Mb1^+/Cre MD4^+/Tg, 53BP1^−/− MD4^+/TG mice; gating on B220⁺. Representative data; n = 3 experiments. Source data

Extended Data Fig. 5. 53BP1-dependent HR suppression in BRCA1-deficient cells is shieldin independent.

a) Immunofluorescent microscopy of Rad51 IRIF in indicated MEF cell lines using Rabbit serum gifted by R. Kanaar. Cells were irradiated (5 Gy) and fixed 2 h later. Representative images of n = 3 biological experiments. b) Quantification of RAD51 IRIF in Edu⁺ cells from D. Integrated intensity and foci quantifications were made using CellProfiler. Boxes indicate the 25th–75th percentiles with the median denoted, and whiskers indicate the 10^th-90^th percentile. Significance was determined by two-sided Kruskal–Wallis H test with Dunn’s correction for multiple comparisons. p values; ****p < 0.0001, **p < 0.0005, **p < 0.005, *p < 0.05. n = 3 biological experiments. c) Survival of the indicated BARD^AID/AID cell lines grown for 7 days in the presence of doxycycline (2 μg ml⁻¹) and the indicated doses of olaparib, n = 3 biological experiments, Mean±SEM. d) Survival of the indicated BARD1^AID/AID HCT-116 cell-lines grown for 7 days in the presence or absence of IAA (1mM), doxycycline (2 mg ml⁻¹) and the indicated doses of Olaparib, n = 3 biological experiments, Mean±SD. e) Survival of the indicated BARD1^AID/AID HCT-116 cell-lines grown for 7 days in the presence or absence of IAA (1mM), doxycycline (2 mg ml⁻¹) and the indicated doses of Olaparib, n = 3 biological experiments, Mean±SD. f) Immunofluorescent microscopy of Rad51 IRIF in indicated BARD1^AID/AID cell lines. Cultures were supplemented with doxycycline (2 mg ml⁻¹ for 24 h) before addition of IAA (1 mM). After 24 h cells were irradiated (5 Gy) and fixed 2 h later. Representative images of n = 3 biological experiments. g) Quantification of RAD51 IRIF in Edu⁺ cells from F. Integrated intensity and foci quantifications were made using CellProfiler. Boxes indicate the 25th–75th percentiles with the median denoted, and whiskers calculated by the Turkey method. Significance was determined by two-sided Kruskal–Wallis H test with Dunn’s correction for multiple comparisons. p values; ****p < 0.0001, ***p < 0.0005, **p < 0.005, *p < 0.05. h) Immunofluorescent microscopy of Rad51 IRIF in indicated KB1P-G3 mouse mammary tumour cell lines. Cells were irradiated (5 Gy) and fixed 2 h later. Representative images of n = 3 biological experiments. i) Quantification of RAD51 IRIF in Edu⁺ cells from H. Integrated intensity and foci quantifications were made using CellProfiler. Boxes indicate the 25th–75th percentiles with the median denoted, and whiskers calculated by the Turkey method. Significance was determined by two-sided Kruskal–Wallis H test with Dunn’s correction for multiple comparisons. p values; ****p < 0.0001, ***p < 0.0005, **p < 0.005, *p < 0.05. Source data

Extended Data Fig. 6. Independent and synergistic contributions of 53BP1 and shieldin to genomic instability in BRCA1-deficient cells.

a) Quantification of between 40 and 65 metaphases analysed per genotype/experiment. Cultures were supplemented with doxycycline (2 μg ml⁻¹ for 24 h) before addition of IAA (1 mM). After a further 24 h cells were treated with DMSO and fixed 24 h later. Significance was determined by two-sided Kruskal–Wallis H test with Dunn’s correction for multiple comparisons. p values; ****p < 0.0001, ***p < 0.0005, **p < 0.005, *p < 0.05. All comparisons not indicated are not significant. n = 3 biological experiments. Source data