An Intrinsically Disordered APLF Links Ku, DNA-PKcs, and XRCC4-DNA Ligase IV in an Extended Flexible Non-homologous End Joining Complex

Abstract

DNA double-strand break (DSB) repair by non-homologous end joining (NHEJ) in human cells is initiated by Ku heterodimer binding to a DSB, followed by recruitment of core NHEJ factors including DNA-dependent protein kinase catalytic subunit (DNA-PKcs), XRCC4-like factor (XLF), and XRCC4 (X4)-DNA ligase IV (L4). Ku also interacts with accessory factors such as aprataxin and polynucleotide kinase/phosphatase-like factor (APLF). Yet, how these factors interact to tether, process, and ligate DSB ends while allowing regulation and chromatin interactions remains enigmatic. Here, small angle X-ray scattering (SAXS) and mutational analyses show APLF is largely an intrinsically disordered protein that binds Ku, Ku/DNA-PKcs (DNA-PK), and X4L4 within an extended flexible NHEJ core complex. X4L4 assembles with Ku heterodimers linked to DNA-PKcs via flexible Ku80 C-terminal regions (Ku80CTR) in a complex stabilized through APLF interactions with Ku, DNA-PK, and X4L4. Collective results unveil the solution architecture of the six-protein complex and suggest cooperative assembly of an extended flexible NHEJ core complex that supports APLF accessibility while possibly providing flexible attachment of the core complex to chromatin. The resulting dynamic tethering furthermore, provides geometric access of L4 catalytic domains to the DNA ends during ligation and of DNA-PKcs for targeted phosphorylation of other NHEJ proteins as well as trans-phosphorylation of DNA-PKcs on the opposing DSB without disrupting the core ligation complex. Overall the results shed light on evolutionary conservation of Ku, X4, and L4 activities, while explaining the observation that Ku80CTR and DNA-PKcs only occur in a subset of higher eukaryotes.

Keywords: APLF; DNA ligase IV; DNA repair; DNA-dependent serine/threonine protein kinase (DNA-PK); Ku; XRCC4; intrinsically disordered protein; non-homologous end joining; protein complex; small-angle X-ray scattering (SAXS).

FIGURE 1.

Purification of proteins. A, schematic of APLF showing the N-terminal FHA domain (residues 21–102), the ATM-dependent phosphorylation site Ser¹¹⁶, the Ku binding motif (KBM) (Arg¹⁸²-Arg¹⁸⁴ and Trp¹⁸⁹) and the PAR (poly(ADP-ribose)) binding domain (Cys³⁷⁹-His⁴⁴⁰). Also shown are representations of CK2-phosphorylated XRCC4 and/or XRCC1 that interact with the FHA domain of APLF. Below is a prediction of the unfolded nature of APLF from FoldINdex (70). B, upper panel: DNA-PKcs and Ku70/80 were purified from HeLa cells, human APLF was purified as a GST fusion protein from bacteria and the GST tag cleaved off with PreScission protease. 1 μg of each protein was run on an SDS-PAGE gel and stained with Coomassie Blue. Molecular mass markers are shown on the left-hand side in kDa. The predicted molecular mass of APLF (accession number BC041144.1, 511 amino acids) is 56,956 Da. Bacterially expressed APLF runs higher than the predicted molecular mass on SDS-PAGE, at ∼80 kDa. Full-length human XRCC4-ligase IV (X4L4) complex was purified from baculovirus-infected insect cells as described under “Experimental Procedures.” Approximately 0.2 μg of purified protein was analyzed on SDS-PAGE and stained with Coomassie Blue. C, MALDI-TOF MS spectrum of purified APLF. GST-APLF protein was purified from an E. coli expression system, the GST tag was removed by PreScission protease and the sample analyzed by mass spectrometry as described under “Experimental Procedures.” D, SDS-PAGE gel of APLF (lane 2) in comparison to APLF treated with glutaraldehyde (lane 3). Note that treated APLF predominantly ran at the molecular mass of the monomer ∼57 kDa.

FIGURE 2.

SAXS analyses of Ku·DNA-PKcs·APLF complexes. A, dimensionless Kratky plots for APLF (purple), Ku (cyan), Ku/20bpDNA (blue), Ku/APLF(green), Ku/20bpDNA/APLF (pink, batch mode; black, collected in SEC-SAXS mode) indicate the level of disorder. B, normalized P(r)s calculated for the experimental data shown in panel C. The area of the P(r) is normalized relative to the M_r estimated by SAXS (Mr_SAXS) and is listed in Table 1. The highlighted areas under the APLF, P(r), and between Ku/APLF and Ku P(r)s indicates 1:1 stoichiometry of the complex. C, SAXS curves for APLF, Ku, Ku·20bpDNA, Ku·APLF, and Ku·20bpDNA·APLF colored according to the panel A. Green curves indicate theoretical SAXS profiles for corresponding ensemble models shown in panels E–H. Inset, Guinier plots for the SAXS curves. D–H, ensemble models of APLF, Ku·APLF, Ku·20bpDNA, and Ku·20bpDNA·APLF. The determined percentage in the ensemble and R_g value of each conformer is indicated.

FIGURE 3.

Integrity of NHEJ complexes on SEC. A and B, SEC profiles of APLF, Ku, DNA-PKcs, Ku·APLF, Ku·DNA-PKcs, Ku·DNA-PKcs·APLF, and Ku·X4L4 in the presence of 20bpDNA are colored as indicated. SEC profiles are normalized at the maxima of the main peak. B, zoom in of the SEC profiles of the complexes shown in panel A. C, SDS-PAGE of APLF SEC-peak fractions together with stock solution of APLF (lane 1) and Ku heterodimer (lane 2). Note that APLF and Ku chain Ku70 run at the same level. D, SDS-PAGE of Ku/APLF SEC-peak fractions in the presence of 20bpDNA shows the presence of Ku70, Ku80, and APLF as indicated. E, Ku·20bpDNA·APLF SEC fractions 1 and 2 (lanes 6 and 7) from panel D together with positive and negative controls as indicated were boiled in SDS sample buffer, loaded onto SDS-PAGE gels, and immunoblotted with antibodies to APLF. F–H, SDS-PAGE of DNA-PKcs·Ku, DNA-PKcs·Ku·APLF, and Ku·X4L4·APLF SEC, peak fractions in the presence of 20bpDNA shows the presence of the complex components as indicated. I, the most concentrated fraction, 1, of Ku·20bpDNA·X4L4·APLF from panel H (lane 2) together with positive control were boiled in SDS sample buffer, loaded onto SDS-PAGE gels, and immunoblotted with antibodies to APLF.

FIGURE 4.

APLF interacts with the Ku·DNA-PKcs·DNA complex. A, His-APLF was immobilized on nitrilotriacetic acid beads and incubated with HeLa whole cell extracts. Beads were washed either in the absence (−) or presence (+) of ethidium bromide (EtBr, 50 μg/ml), then boiled in SDS sample buffer, loaded onto SDS-PAGE gels, and immunoblotted with antibodies to His (for His-APLF), DNA-PKcs, and Ku80 as indicated. B, GST (lane 2) or GST-APLF (lanes 3–6) were immobilized on glutathione-Sepharose 4B beads and incubated with whole cell extracts from HeLa cells that had been either unirradiated (−) or irradiated (10 gray IR) and allowed to recover for 1 h. Beads were washed either in the absence (−) or presence (+) of EtBr (50 μg/ml), then boiled in SDS sample buffer, loaded onto SDS-PAGE gels, and immunoblotted with antibodies to GST (for GST-APLF), DNA-PKcs, and Ku80 as indicated. The lower panel represents a longer exposure of the Ku80 blot to show a signal in the input lanes. Lane 1 contained 50 μg of extract from unirradiated cells as a positive control. C, HeLa cells were transiently transfected with FLAG-tagged APLF (lanes 3 and 4) or empty vector (lane 2), then extracts were immunoprecipitated with anti-FLAG antibody, run on SDS-PAGE, and immunoblotted with antibodies to FLAG (for FLAG-APLF), DNA-PKcs and Ku as indicated. Where indicated, ethidium bromide (50 μg/ml) was added to immunoprecipitation wash buffers. Note: a duplicated sample lane has been removed between lanes 2 and 3. All blots were from the same exposure of the same gels. D, purified DNA-PKcs and/or Ku were incubated with GST-APLF immobilized on glutathione-Sepharose 4B beads in either the absence (−) or presence (+) of CT-DNA (10 μg/ml). Samples were run on SDS-PAGE and immunoblotted with antibodies to GST (for GST-APLF), DNA-PKcs and Ku as indicated. E, purified DNA-PKcs and Ku were incubated with GST-APLF (lanes 3–8) or GST (lane 2) immobilized on glutathione-Sepharose 4B beads in the presence of different lengths of DNA (10 μg/ml) and then immunoblotted with antibodies as indicated. In lanes 2 and 4, proteins were incubated in the presence of 10 μg/ml of CT-DNA, lane 5 contained 40 base ssDNA; lane 6, 40-bp dsDNA; lane 7, 100 base ssDNA; and lane 8, 100-bp dsDNA. Lane 1 contained 100 ng each DNA-PKcs and Ku. Lane 3 contained no DNA. F, purified DNA-PKcs and Ku were incubated with either GST alone (lanes 2 and 3), GST-APLF (lanes 4 and 5), or GST-APLF residues 1–120 (lanes 6 and 7), 110–360 (lanes 8 and 9), or 360–511 (lanes 9 and 10) that had been bound to glutathione-Sepharose 4B beads either in the absence (−) or presence (+) of CT-DNA (80 μg/ml). Samples were washed, run on SDS-PAGE, and immunoblotted. Lane 1 contains 100 ng each DNA-PKcs and Ku. The upper panel is a Ponceau Red-stained membrane, whereas the lower panels show immunoblots for DNA-PKcs and Ku80, respectively. Positions of molecular mass markers (in kDa) are shown on the left-hand side on the Ponceau-stained blot. G, GST alone, GST-APLF, or GST-APLF with mutations of R182E/K183E/R184E or W189G were bound to glutathione-Sepharose 4B beads and incubated with purified DNA-PKcs and Ku in the absence (−) or presence (+) of CT-DNA as above then immunoblotted with antibodies to GST (for GST-APLF), DNA-PKcs, and Ku80 as indicated.

FIGURE 5.

EMSA of Ku·DNA-PKcs·APLF complexes. A, purified Ku, DNA-PKcs, or APLF were incubated with 6 pmol of 3′-FAM-labeled 40-bp dsDNA and analyzed by EMSA as described under ”Experimental Procedures.“ Lane 1 contained DNA alone. Lanes 2 and 3 contained 6 pmol of BSA or APLF, respectively. Samples in lanes 4–12 contained purified Ku70/80 heterodimer (6 pmol). DNA-PKcs was present at 6 pmol in lanes 6–12. APLF was added at 6 (lanes 5 and 7), 12 (lanes 8), or 24 pmol (lane 9). Lanes 10–12, contained 6, 12, or 24 pmol of BSA, respectively. B, purified proteins were incubated with 6 pmol of 3′-FAM labeled 25-bp dsDNA and analyzed as above. Lane 1 contained DNA alone. Lanes 2 and 3, contained DNA-PKcs (6 pmol) or APLF (24 pmol), respectively. Lane 4 contained DNA-PKcs and APLF. Samples in lanes 5-8 contained purified Ku70/80 heterodimer (6 pmol). APLF was added at 24 pmol in lanes 6 and 8 and DNA-PKcs was added (6 pmol) in lanes 7 and 8.

FIGURE 6.

Solution structure reconstructions of Ku·20bpDNA·DNA-PKcs·APLF complexes. A, SEC-MALS chromatographs for Ku (cyan), DNA-PKcs (green), APLF (purple), and Ku, Ku·DNA-PKcs or Ku·DNA-PKcs·APLF (blue, yellow, red, respectively) in the presence of 20bpDNA at molar ratios of 1:1:1:1 at a final concentration of 5.3 μm. Solid lines represent the light scattering signal (Rayleigh ratio in arbitrary units), the symbols represent molecular mass versus elution time. B, SEC-SAXS profiles for DNA-PKcs (green), Ku·20bpDNA·DNA-PKcs (yellow) and Ku·20bpDNA·DNA-PKcs·APLF (red) showing I(0) (lines) and R_g (symbols) values are shown for each collected frame across the SEC peak. The black box indicates the frame, which gives the SAXS profile (shown in the panel C)) used in the SAXS analysis. C, SAXS curves for DNA-PKcs (green, taken from Ref. ; cyan, curve obtained from SEC-SAXS), Ku·20bpDNA·DNA-PKcs (yellow), and Ku·20bpDNA·DNA-PKcs·/APLF (red) complexes. Inset: Guinier plot for the SAXS curves. The black line shows the model fit for the single phase envelope (Ku·20bpDNA·DNA-PKcs·APLF) calculated by DAMMMIF or multiphase envelope with the individual phases DNA-PKcs, Ku·20bpDNA·DNA-PKcs calculated by MONSA. D, P(r) calculated for the experimental SAXS of DNA-PKcs (green), Ku·20bpDNA·DNA-PKcs (yellow), and Ku·20bpDNA·DNA-PKcs·APLF (red) shown in panel C. The area of the P(r) is normalized relative to the Mr_SAXS listed in Table 1. Inset, dimensionless Kratky plots for DNA-PKcs (green), Ku·20bpDNA·DNA-PKcs (yellow), and Ku·20bpDNA·DNA-PKcs·APLF (red) indicate the level of disorder as indicated. E and F, four representative SAXS envelopes of Ku·20bpDNA·DNA-PKcs and Ku·20bpDNA·DNA-PKcs·APLF complexes superimposed with the atomistic model of Ku·20bpDNA and DNA-PKcs crystal structure (30). The right panel shows envelopes in ∼70% of their volume to highlight the hollow feature of the DNA-PKcs central region. E, far left panel: representative single and average multiphase envelopes of Ku·20bpDNA·DNA-PKcs. Phases for DNA-PKcs and Ku·DNA are colored as indicated. Model fit for the individual phases and complexes are shown in the panel C. Models in all panels are to the same scale.

FIGURE 7.

APLF interacts with the Ku·X4L4 complex on dsDNA. A, human XRCC4, XRCC4-T233A, and GST-CK2 were expressed and purified from E. coli as described previously (33). X4L4 was purified from baculovirus-infected insect cells as described in the text. 1 μg of XRCC4 or XRCC4-T233A or 2 μg of X4L4 protein was either untreated, or phosphorylated by CK2 in vitro and/or treated with λ phosphatase as described previously (33). Samples were analyzed by SDS-PAGE and either stained by Coomassie Blue (top panel) or probed with a phosphospecific antibody to XRCC4-pT233 (33), followed with an antibody to total XRCC4 protein as indicated on the right. Top panel, molecular mass markers (in kDa) are shown on the left-hand side. B, Ku, X4L4, and APLF at a ratio of 1:1:1 and a final concentration of 14 μm were incubated for 10 min in the presence or absence of DNA (20bpDNA or 20bpDNA-10nt) as indicated. Loading buffer was added and samples were analyzed on 4–12% acrylamide non-denaturing gels, followed by ethidium bromide staining. C, double strand ligation by X4L4 and effects of Ku, DNA-PKcs, APLF, and XLF. Ligation reactions were carried out with 0.25 pmol of DNA substrate and indicated proteins (0.25 pmol each) as described under “Experimental Procedures.” Their products were resolved by denaturing polyacrylamide gel electrophoresis, and their mobilities in the gel are indicated. M1 and M2 are size markers for single strand ligation and double strand ligation, respectively. Formation of the doublets as the double strand ligation products is due to partial annealing of the DNA during denaturing gel electrophoresis. The structure of the DNA substrate is schematically diagramed in the left side of the gel.

FIGURE 8.

SEC-MALS and SEC-SAXS of the Ku·X4L4·APLF complex. A, SEC-MALS chromatographs for Ku (blue), X4L4 (green), Ku·X4L4 (yellow), and Ku·X4L4·APLF (red) in the presence of one 20bpDNA or two 20bpDNA-10nt with complementary end groups at molar ratios of 1:1:1:1 (final concentrations 14 μm). Solid lines represent the light scattering signal (Rayleigh ratio in arbitrary units), the symbols represent molecular mass versus elution time. Bottom panels show EMSA gels for the SEC-MALS peak fraction for Ku·20bpDNA·X4L4·APLF in the presence of one 20bpDNA (red) or two 20bpDNA-10nt with complementary end groups (violet). B, SEC-SAXS profiles for Ku·X4L4·APLF in the presence of stem-loop DNA molecules with either one (20bpDNA, red) or two complementary ends (20bpDNA-10nt, violet). I(0) (lines) and R_g (symbols) values are shown for each collected frame across the SEC peak. The black box indicates the frame that gives the SAXS curves (shown in Fig. 9E) used in further SAXS analysis. C, P(r) of X4L4 (gray) and Ku·X4L4·APLF in the presence of one 20bpDNA (red) or two 20bpDNA-10nt (violet) as indicated in panel A. P(r) have been calculated for the SAXS curves shown in Fig. 9E. The area of the P(r) is normalized relative to the Mr_SAXS listed in the Table 1.

FIGURE 9.

Solution structure reconstruction of the Ku·X4L4·APLF complex. A, average and four representative SAXS envelopes of one-site Ku·20bpDNA·X4L4·APLF complexes reconstructed with P1 symmetry operator. B, average and four representative single multiphase envelopes of one-site Ku·20bpDNA·X4L4·APLF complex with three phases for Ku/DNA, APLF, and X4L4 colored as indicated. Bottom, representative fits of the individual phases (Ku·DNA, blue; APLF, violet; X4L4, yellow) and complexes (Ku·20bpDNA·X4L4·APLF, red) for the multiphase models calculated by MONSA. C, average and four representative SAXS envelopes of two-site Ku·20bpDNA-10nt·X4L4·APLF complexes reconstructed with P2 symmetry. Bottom, the average envelope reconstructed with P1 symmetry. The putative locations of L4 catalytic domain, X4, and Ku are indicated. Envelopes are at the same scale. D, ensemble models of X4L4 according Ref. is shown in schematic representation and colored as indicated. Average SAXS envelope of a one-site Ku·20bpDNA·X4L4·APLF complex is superimposed with X4L4 and Ku·20bpDNA·APLF atomistic model. Two orthogonal views of the average SAXS envelopes of a two-site Ku·20bpDNA-10nt·X4L4·APLF complex. Two one-site complexes were superimposed with SAXS envelope. These atomistic models were used to match the experimental SAXS curves as shown in the panel E. Models are to the same scale. E, SAXS curves for Ku·X4L4·APLF in the presence of 20bpDNA (red) or two complementary DNAs (20bpDNA-10nt, purple) in comparison with the SAXS curve of X4L4 in the absence of DNA (gray). Theoretical SAXS profiles (green) of the ensemble models from panel D in the fit to the corresponding experimental curves are shown together with the fit-residuals for single (red) and ensemble models (green). Determined weights in the ensemble of two and corresponding χ values are indicated.

FIGURE 10.

The architecture of the NHEJ core complex based on combined crystallographic and SAXS structures explains the synergy of Ku/X4L4/XLF interactions for ligating DSBs. A–C, three orthogonal views of the atomistic model of aligned DNA ends stabilized by the Ku·X4L4·XLF·APLF complex are shown in molecular surface representation. The APLF-FHA domain interacts with phosphorylated X4 (229–235 region) (71). At the same time, APLF residues Arg¹⁸²-Lys¹⁸³-Arg¹⁸⁴ and Trp¹⁸⁹ in the central region (residues 110–360) bridge Ku80 and DNA-PKcs. The Ku-nucleated X4/XLF filament appears suitable to maintain DNA end alignment via its grooved DNA binding surface proposed in Ref. and tested as described in Ref. . The parallel X4/XLF filaments are shown as seen in the crystal structure (49) and tested by SAXS (48, 49). The crystal structure of DNA-PKcs from Ref. is shown tethered by Ku (45) through interactions with the extended Ku80CTR (25) as visualized through the SAXS models presented in this study. The central X4 tethered to L4 as shown in the crystal structures of the X4L4 complex (72, 73), the SAXS model of X4L4 (38, 39), and presented here as a SAXS model of the Ku·X4L4 complex. Distant DNA-PKcs and L4 catalytic domains are proposed to allow DNA-PKcs activation of NHEJ partners and providing steric access to L4 catalytic domains with the DNA. NHEJ partners are colored according to the legend. D, SAXS envelopes (gray) of one- and two-site Ku·DNA·X4L4·APLF complexes are superimposed with the corresponding region of the models. E, schematic representations of DNA pairing by the Ku-nucleated X4L4/XLF filament groove. The role of APLF is to scaffold X4L4 with the further located Ku/DNA-PKcs assembly, which allowed DNA pairing and final alignment. Models are distinguished by having DNA moved distal to the Ku undisturbed by L4 catalytic domains or DNA-PKcs. The schematic of aligned DNA shows DNA ends in an end-to-end configuration in which the ends are compatible for ligation.