Substantial evidence has shown that overexpression of the inhibitor of apoptosis protein (IAP) survivin in human tumors correlates significantly with treatment resistance and poor patient prognosis. Survivin serves as a radiation resistance factor that impacts the DNA damage response by interacting with DNA-dependent protein kinase (DNA-PKcs). However, the complexity, molecular determinants, and functional consequences of this interrelationship remain largely unknown. By applying coimmunoprecipitation and flow cytometry-based Förster resonance energy transfer assays, we demonstrated a direct involvement of the survivin baculovirus IAP repeat domain in the regulation of radiation survival and DNA repair. This survivin-mediated activity required an interaction of residues S20 and W67 with the phosphoinositide 3-kinase (PI3K) domain of DNA-PKcs. In silico molecular docking and dynamics simulation analyses, in vitro kinase assays, and large-scale mass spectrometry suggested a heterotetrameric survivin–DNA-PKcs complex that results in a conformational change within the DNA-PKcs PI3K domain. Overexpression of survivin resulted in enhanced PI3K enzymatic activity and detection of differentially abundant phosphopeptides and proteins implicated in the DNA damage response. The survivin–DNA-PKcs interaction altered the S/T-hydrophobic motif substrate specificity of DNA-PKcs with a predominant usage of S/T-P phosphorylation sites and an increase of DNA-PKcs substrates including Foxo3. These data demonstrate that survivin differentially regulates DNA-PKcs-dependent radiation survival and DNA double-strand break repair via formation of a survivin–DNA-PKcs heterotetrameric complex.

Significance:

These findings provide insight into survivin-mediated regulation of DNA-PKcs kinase and broaden our knowledge of the impact of survivin in modulating the cellular radiation response.

See related commentary by Iliakis, p. 2270

The activation of signal transduction pathways by ionizing radiation to induce DNA damage repair and to regulate cell-cycle progression/arrest and cell death is a main feature of the cellular radiation response (1). There is substantial evidence for the involvement of survivin, the smallest and structurally unique member of the mammalian inhibitor of apoptosis protein (IAP) family (2), in all of these pathways. However, the complexity and molecular basis of survivin’s impact on radiation response, especially on its interaction with the key repair factor DNA-PKcs, remains largely unknown.

Survivin encompasses a single IAP family characteristic baculovirus IAP repeat (BIR) motif, which modulates protein–protein interactions and an amphipathic α-helical coiled-coil domain at the C-terminus, common in microtubule-associated proteins (3) and is implicated in multiple tumor signaling pathways (4). Survivin binds to a huge number of protein partners, including tubulin, nuclear and heat shock proteins, caspases, and diverse members of the IAP family such as the X chromosome-linked IAP (XIAP) and cellular (c)IAP1 (4). In addition, survivin shuttles to multiple cellular compartments, including mitochondria, cytoplasm, and nucleus (5). Nuclear localization is essential for chromosomal segregation and cytokinesis as survivin is an indispensable member of the chromosomal passenger complex (CPC) along with the mitotic kinase Aurora B, Borealin/DasraB, and Inner centromere protein (INCENP; ref. 6). Clinically, survivin is overexpressed in the majority of tumors and significantly correlates with tumorigenesis, treatment resistance, and patients’ poor prognosis (4, 7). In line with that, survivin is a radiation resistance factor and knockdown of the protein sensitizes tumor cells to radiation therapy by caspase-dependent and -independent mechanisms (8–11). In that context, a nuclear accumulation of survivin, localization at the site of DNA damage, and interaction with components of the DNA double-strand break (DSB) repair apparatus by forming a complex with DNA-PKcs was reported by our group and others (11, 12).

Non-homologous end joining (NHEJ), the predominant DNA DSB repair pathway in mammalian cells, requires DNA-PK, which assembles when the Ku70 and Ku80 heterodimer binds to DNA, encircles its end, translocates inwards, and recruits the DNA-PKcs kinase (13). Using electron microscopy, dimeric complexes were reported to contain two DNA-PK holoenzymes interacting to promote the synapsis of broken DNA ends (14). DNA-PKcs is a member of the phosphatidylinositol 3-kinase (PI3K) related kinase family that mediates DNA damage repair by phosphorylating multiple factors including histone H2AX and members of the NHEJ machinery, such as Artemis, Ligase IV, and X-ray repair cross-complementing protein 4 (XRCC4; ref. 15). In addition, DNA-PKcs is reported to impact on diverse cellular processes including telomere maintenance, apoptosis regulation, modulation of chromatin structure, and transcriptional regulation (16).

In this study, we show that S20 and W67 residues located within the BIR domain of survivin are essential for association with the catalytic PI3K domain of DNA-PKcs, resulting in a novel heterotetramer complex with increased kinase activity and repair capacity. Large-scale (phospho)proteomics studies further indicate differentially abundant candidate phosphosites and proteins implicated in DNA damage repair and response mediated by the survivin–DNA-PKcs interaction.

Cell culture and irradiation procedure

The colorectal carcinoma cell lines SW480, DLD-1, HCT-15, and human embryonic kidney cell line HEK293T were purchased from the ATCC (LGC Promochem) and maintained according to their respective guidelines (maximum 10 passages from thawing). All cell lines were regularly tested for Mycoplasm contamination by using Venor Gem Mycoplasma Detection Kit (Biochrom, WVGM-050; latest test date: July 11, 2020). Irradiation of cells was performed using a linear accelerator (ELEKTA) with a 6 MV photon energy, a 100 cm focus-surface distance, and a dose rate of 6 Gy/min.

Constructs, siRNA, and transfection

Primer and siRNA sequences used are provided in Supplementary Table S1. Site-directed mutagenesis was performed using a QuikChange II Mutagenesis Kit (Agilent Technologies) following the manufacturer’s instructions. Survivin cDNA was amplified from a plasmid kindly provided by Roland H. Stauber (University Hospital of Mainz, Mainz, Germany) and was inserted into KpnI/BamHI sites of the pEGFP-N1 or pEYFP-N1 expression vector (Clontech). For HEAT1, FATC and PI3K domain constructs of DNA-PKcs and PI3K domain of ATM, total RNA was isolated from EA.hy926 cells and cDNA was synthesized by using M-MLV Reverse Transcriptase (Promega) and random primer hexamers and amplified fragments were inserted into pECFP-N1 expression vector. Flag-tagged survivin and PI3K domain constructs of DNA-PKcs were generated by insertion of fragments into EcoRI/KpnI sites of the p3xFlag-CMV10/14 expression vectors (Sigma-Aldrich). In-frame deletions and mutated sites were confirmed by sequencing (Eurofins Genomics). Stable transfections and siRNA-mediated knockdown at a concentration of 20 nmol/L siRNA were performed using Roti-Fect PLUS transfection reagent (Carl Roth) according to the manufacturer’s recommendations.

NanoBiT complementation assay

To perform luciferase-based protein interaction analysis, a NanoBiT PPI Starter Systems Kit was used (Promega). Vectors including recombinant luminescence complementation tags LgBiT and SmBiT were fused to survivin and DNA-PKcs PI3K domain sequences and transfected into HEK293T cells using the Effectene transfection protocol (Qiagen). At 1 hour after a 4 Gy irradiation, Nano-Glo live-cell reagent was added and luminescence was measured with an Infinite M200 Pro Elisa Reader (TECAN).

Flow cytometer-based Förster resonance energy transfer analyses

Förster resonance energy transfer (FRET) is based upon an energy transfer from an excited donor fluorophore to a near-by (≤10 nm) suitable acceptor fluorophore, resulting in enhanced fluorescence emission of the acceptor and detection of protein–protein interactions in living cells. Genes of interest were cloned into N- or C-terminal tagged EYFP/ECFP vectors, cells were transiently transfected using jetPRIME (Polyplus), and 1 hour after a 4 Gy irradiation FRET signals were determined using a 529/24 filter after excitation with the 405 nm laser (ECFP) and with a 488 nm laser (EYFP) using a CytoFLEX S cytometer (Beckman Coulter). An ECFP-EYFP fusion protein, kindly provided by Michael Schindler (University of Tübingen, Tübingen, Germany), served as a positive control (17).

Cell-cycle analysis

Analysis of cell-cycle distribution was performed by staining with 40 mg/mL propidium iodide (Merck) in the presence of 40 mg/mL RNase A (Qiagen) and measured with a CytoFLEX S flow cytometer. The subsequent analysis was performed by applying the CytExpert Software (Beckman Coulter) as described previously (18).

3D colony formation assay

Three dimensional (3D) cell survival assays were performed as reported (18) and survival curves were fitted according to the linear quadratic equation (SF = exp [−α × Dβ × D2] with D = dose using EXCEL software (Microsoft).

Measurement of caspase-3/7 assay

For quantification of caspase-3/7 activity a CASPASE GLO-assay (Promega) was used according to the manufacturer’s protocol.

Coimmunoprecipitation and Western immunoblotting

SW480 cells were transfected with survivin/PI3K expression vectors for 48 hours and cell lysates were prepared at 1 hour after a 4 Gy irradiation. Equal amounts of lysates were added to coupled Dynabeads Protein G (Thermo Fisher Scientific) and antibody complexes: Anti-GFP (#ab290; Abcam, RRID:AB_303395), anti-Flag (#2368S; Cell Signaling Technology), anti-DNA-PKcs (#MS-369-P1; Thermo Fisher Scientific) or IgG (Mouse #SC-2025, Rabbit #SC-2027; Santa Cruz Biotechnology) and incubated in a rotating mixer at 5 rpm overnight at 4°C. Next, beads were washed with PBS and proteins were eluted by boiling, subjected to Western immunoblotting and detected with primary antibodies: anti-survivin (AF886; R&D Systems), anti-GFP (ab290; Abcam), anti-DNA-PKcs (#MS-369-P1; Thermo Fisher Scientific), anti-Flag-HRP (#ab49763; Abcam), anti-β-actin (A5441-.2ML; Sigma-Aldrich), anti-Foxo3 (#2497S), or anti-Foxo3 pS253 (#9466S; Cell Signaling Technology). For detection, horseradish peroxidase-conjugated goat anti-rabbit and goat anti-mouse secondary antibodies (Santa Cruz Biotechnology) and LI-COR WesternSure Premium Chemiluminescent Substrate (LI-COR) were used.

Immunofluorescence imaging

Residual DNA DSBs were quantified using foci assays as described previously (19) with detection of histone γH2AX (clone JBW301; Millipore) and 53PB1 (#NB-100-304; Novus Biologicals) in at least 50 nuclei per single experiment using an Axio Imager Z.1 microscope (Zeiss). Acute DSBs per Gy are approximately 35 to 40 foci per cell and increase linearly with the irradiation dose (20). Thus, to ascertain microscopic assessment/quantification of distinct and individual foci cells were irradiated with a dose of 2 Gy.

In vitro DNA-PKcs kinase assay

DNA-PKcs kinase activity was quantified with a SignaTECT DNA-Dependent Protein Kinase Assay System (Promega) using purified DNA-PKcs (30 units, #V5811; Promega) and Surv-EGFP [survivin wild type (wt) inserted into pEGFP-N1] immunoprecipitated at 1 hour after 4 Gy irradiation. One μmol/L DNA-PK inhibitor (KU 0060648; Tocris) treated preparations were used as a negative control. (γ-32P)ATP (Perkin Elmer) was quantified by a Fujifilm BAS-1500 Imager (GE Healthcare Life Sciences) and evaluated by using TINA Image Analysis Environment software (OSMIA Project, EU IST Program).

Molecular docking analysis

For initial in silico protein docking analysis, DNA-PKcs and survivin structures were pre-processed using the Protein Preparation Wizard (21) in the Schrödinger Release 2018-1 to refine the missing/truncated parts followed by a restrained minimization. Refined structures were docked by using PIPER (22) and PatchDOCK (RRID:SCR_017589; ref. 23) and post-processed by FireDOCK software (24). Interaction residues were determined according to their binding energies by using PRIME MM-GBSA (25) and distances were calculated by Find Clashes/Contacts tool of Chimera 1.13.1 (26).

For large-scale molecular docking (MD) analyses, the DNA-PKcs head domain (residues 2802-4128) model was refined using the MODELLER software (RRID:SCR_008395; ref. 27). Large-scale MD analysis was performed using the head domain of DNA-PKcs and survivin by the global docking protocol of Rosetta (28). The 98884 poses generated were next used as an input for the local docking protocol, using Rosetta all-atom energy function (29). The subsequent analysis regarding the proximity of BIR domain residues and the PI3K domains was done using custom python scripts and the bioinformatics library Biotite (30). We calculated the minimum distance between the docked survivin and the whole PI3K region by the following formula:

formula

|{R_{{\rm{PI}}3{\rm{K}}}}$| and |{R_{{\rm{BIR}}}}$| indicate atom coordinates. The docking quality is described by increasingly negative Rosetta interface scores. These scores are defined as the differences in Rosetta score for the structure with two docking poses together and separated by 100 Å, eventually mimicking no interaction (31). Calculations are described in more detail in the Supplementary Materials and Methods.

Molecular dynamics simulation and calculation of the molecular in silico attraction

To analyze the stability and dynamics of the heterotetramer structure, 200 ns molecular dynamics simulations were performed by employing Gromacs 2019.4 (RRID:SCR_014565; ref. 32) with the CHARMM36 force field (33). Structural changes were measured by root mean square deviation (RMSD) and radius of gyration (Rg; Supplementary Fig. S3A).

formula
formula

Finally, the measurement of the active site cavity dynamics was performed by a unique and simple quantitative measurement strategy. Briefly, the base of a cylindrical reference volume was fitted to the outer surface of PI3K active site and the atoms that were inside the cylindrical reference volume during the molecular dynamics simulations were counted (Supplementary Fig. S3B). A more detailed description of the method is provided in the Supplementary Materials and Methods.

Liquid chromatography-mass spectroscopy

Wild-type (wt), EGFP, or Surv-EGFP stably transfected SW480 cells were subjected to siRNA-mediated knockdown by control siRNA or survivin siRNA or treated with DNA-PK inhibitor (1 μmol/L). At 1 hour after irradiation, cells were harvested with ice-cold PBS, centrifuged (100 × g, 5 min, 4°C), and pellets were lysed as described previously (34). Lysates were incubated with TCEP-chloroacetamide buffer and were precipitated by using a methanol–chloroform mixture. Proteins were resuspended in Urea-EPPS buffer and digested with LysC and trypsin. Digests were acidified using trifluoroacetic acid (TFA) and peptides were purified using SepPak C18 columns according to the manufacturer’s instructions. Phosphopeptides, enriched by using a High-Select Fe-NTA Phosphopeptide Enrichment Kit, were subsequently purified using C18 resin material and qual amounts of peptides and phosphopeptides were labeled with tandem mass tag (TMT) reagents. Samples were dried and resuspended in formic acid buffer and separated on an easy-nLC 1200 nano HPLC system using a nanoFlex ion source (Thermo Fisher Scientific). Peptide fractions were eluted by a nonlinear gradient on a self-made 32 cm long, 75 μm inner diameter (ID) fused-silica column and sprayed into an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific). The more exact settings and detailed materials and methods are given in the Supplementary Materials and Methods.

Liquid chromatography-mass spectroscopy data analysis

Log2-fold changes were calculated by log2 transformation of the ratio between the treated sample versus the control sample. Hierarchical clustering analysis was performed using the Perseus software package (version 1.6.10.45, RRID:SCR_015753) with default settings after centering and scaling of data (Z scores). Graphs were created by using Prism 8 software (RRID:SCR_002798). Principal component analysis (PCA) plots were generated by the calculation ratios using Perseus software package (version 1.6.10.45) with default settings. PCA is a reduction method, which reduces the dimensions of a dataset by transformation while preserving the main information from the original data.

Statistical analysis

Mean SDs were calculated with reference to controls defined in a 1.0 scale. To test statistical significance, a two-sided unpaired Student t test was applied using EXCEL software. Results were considered statistically significant with a P value of less than 0.05.

Availability of supporting data

The mass spectrometry (phospho)proteomics data are available via the PRIDE repository (https://www.ebi.ac.uk/pride/; accession code: PXD020489). Animations of molecular dynamics simulations, scripts, and raw data of large-scale molecular docking and molecular dynamics simulations are available at https://github.com/entropybit/survivinpkcs and http://www.cbs.tu-darmstadt.de/SurvivinDnaPkcs. Pathway analysis was performed using Pathway Commons (https://www.pathwaycommons.org) and consensus motif analysis was performed by Seq2Logo (http://www.cbs.dtu.dk/biotools/Seq2Logo/).

The survivin BIR domain is important for clonogenic radiation survival and repair of DNA DSBs

To unravel the structural determinants of survivin’s impact in radiation response, we stably expressed survivin–EGFP fusion proteins harboring a deletion of the functional BIR (ΔBIR) or microtubule binding (ΔMicTub) domain in endogenous survivin-depleted SW480 colorectal cancer cells (Fig. 1A). We confirmed similar expression of exogenous constructs and attenuation of endogenous survivin levels by more than 90% by Western immunoblotting (Fig. 1B). As survivin is well described to impact on apoptosis regulation (4), apoptosis induction was used as experimental endpoint to assess functional expression of the recombinant proteins. Caspase-3/7 activity assays revealed a significantly elevated apoptosis induction upon knockdown of endogenous survivin in EGFP-control, survivin ΔBIR, or ΔMicTub mutant transfected cells, whereas recombinant survivin wt (Surv. wt) rescued the knockdown (Fig. 1C). SW480 cells were next subjected to a three-dimensional colony formation assay to further investigate long-term cellular responses upon BIR or MicTub domain deletion. Survivin knockdown resulted in a significantly decreased clonogenic survival in EGFP-control and ΔBIR mutant cells, whereas overexpression of Surv. wt or ΔMicTub rescued survival in SW480 cells (Fig. 1D; Supplementary Fig. S2B). We confirmed these findings in a second colorectal cancer cell line, DLD-1 (Supplementary Fig. S1A). Further, radiation-induced residual (24 hours) γH2AX/53BP1 foci were significantly increased upon knockdown of endogenous survivin and expression of EGFP and ΔBIR, but not in SW480 or DLD-1 cells expressing Surv. wt or ΔMicTub (Fig. 1E; Supplementary Figs. S1B and S2C). Thus, we regarded the BIR domain (residues 18–88) as pivotal for impacting clonogenic radiation survival and residual DNA damage response, whereas the MicTub domain seemed not to affect these processes.

Figure 1.

Survivin BIR deletion mutant fails to rescue 3D radiation survival and radiation-induced DNA damage repair. A, Schematic presentation of enhanced green fluorescent protein (EGFP) tagged survivin vectors: Surv. wt, survivin wild type; ΔBIR, survivin BIR domain deletion mutant; ΔMicTub, survivin microtubules binding site deletion mutant. B, Expression of recombinant survivin–EGFP fusion proteins and knockdown of endogenous (endog.) survivin was confirmed by Western immunoblotting. β-Actin served as a loading control. siCtrl, control siRNA; siSurv, survivin siRNA. C, Caspase-3/7 activity of indicated stable lines transfected with siCtrl or siSurv (20 nmol/L) was measured 48 hours after irradiation with 0 or 6 Gy and calculated relative to mock-irradiated siCtrl-transfected cells (n = 4). D, Clonogenic radiation survival of the indicated SW480 3D cell cultures was analyzed after irradiation with 0, 2, 4, or 6 Gy (single dose). Results represent means ± SD (n ≥ 4). E, SW480 cells, stably expressing indicated survivin mutants, were subjected to siRNA transfection (siCtrl, siSurv, 20 nmol/L) and subsequently irradiated with a dose of 2 Gy. At 24 hours after irradiation (residual damage), cells were fixed and stained for γH2AX/53BP1, whereas nuclei were counterstained with DAPI. Nuclear γH2AX/53BP1 foci were microscopically counted (50 nuclei per experiment). Results represent mean foci per cell + SD (n = 3). *, P < 0.05; **, P < 0.01; ***, P < 0.001; t test.

Figure 1.

Survivin BIR deletion mutant fails to rescue 3D radiation survival and radiation-induced DNA damage repair. A, Schematic presentation of enhanced green fluorescent protein (EGFP) tagged survivin vectors: Surv. wt, survivin wild type; ΔBIR, survivin BIR domain deletion mutant; ΔMicTub, survivin microtubules binding site deletion mutant. B, Expression of recombinant survivin–EGFP fusion proteins and knockdown of endogenous (endog.) survivin was confirmed by Western immunoblotting. β-Actin served as a loading control. siCtrl, control siRNA; siSurv, survivin siRNA. C, Caspase-3/7 activity of indicated stable lines transfected with siCtrl or siSurv (20 nmol/L) was measured 48 hours after irradiation with 0 or 6 Gy and calculated relative to mock-irradiated siCtrl-transfected cells (n = 4). D, Clonogenic radiation survival of the indicated SW480 3D cell cultures was analyzed after irradiation with 0, 2, 4, or 6 Gy (single dose). Results represent means ± SD (n ≥ 4). E, SW480 cells, stably expressing indicated survivin mutants, were subjected to siRNA transfection (siCtrl, siSurv, 20 nmol/L) and subsequently irradiated with a dose of 2 Gy. At 24 hours after irradiation (residual damage), cells were fixed and stained for γH2AX/53BP1, whereas nuclei were counterstained with DAPI. Nuclear γH2AX/53BP1 foci were microscopically counted (50 nuclei per experiment). Results represent mean foci per cell + SD (n = 3). *, P < 0.05; **, P < 0.01; ***, P < 0.001; t test.

Close modal

The catalytic PI3K domain of DNA-PKcs interacts with the BIR domain of survivin

As survivin localizes at the site of radiation-induced DNA–DSBs and interacts with DNA-PKcs (11, 12), we next aimed to characterize putative molecular interaction sites. We cloned recombinant fusion fluorescence proteins covering DNA-PKcs-PI3K, -HEAT1, -FATC, Surv. wt, and ΔBIR constructs. The PI3K domain of ataxia–telangiectasia mutated (ATM) served as a control (Fig. 2A). SW480 and embryonic kidney HEK293T cells transiently transfected with these constructs were subjected to live-cell protein–protein interaction assays using FACS–FRET (Supplementary Fig. S1C; ref. 17). Combined survivin and DNA-PKcs PI3K domains but not ΔBIR mutants showed an enhanced FRET signal, whereas additional DNA-PKcs or ATM PI3K domains led to less pronounced interaction (Fig. 2B). Notably, kinetic analysis of the survivin–DNA-PKcs PI3K complex revealed the highest abundance 1 hour after irradiation followed by a gradual dissociation over time (Supplementary Fig. S4). We further performed a luciferase-based complementation assay (NanoBiT) to ascertain a close interrelationship between survivin and the PI3K domain. We detected a high luminescence signal when the two luciferase subunits (LgBiT and SmBiT) were fused to survivin and the DNA-PKcs PI3K domain (Fig. 2C). Immunoprecipitation analyses further revealed DNA-PKcs binding to EYFP-tagged Surv. wt (and vice versa), but not to the ΔBIR mutant, confirming the involvement of the BIR domain in the interaction with the DNA-PKcs PI3K domain (Fig. 2D).

Figure 2.

BIR domain of survivin interacts with the PI3K domain of DNA-PKcs. A, Schematic presentation of the domains of survivin, DNA-PKcs, and ATM, which were subjected to FACS-FRET analyses, with vectors expressing the enhanced yellow or cyan fluorescent protein (EYFP, ECFP). Surv. wt, survivin wild type; ΔBIR, survivin BIR domain deletion mutant; PI3K, DNA-PKcs PI3K domain; HEAT1, DNA-PKcs Huntingtin-elongation factor 3-PP2A-TOR1 (HEAT1) domain; FATC, FRAP-ATM-TRRAP C-terminal (FATC) domain-ECFP; PI3K-ATM, ATM PI3K domain. B, FACS-FRET analysis in SW480 or HEK293T cells performed at 1 hour after 4 Gy irradiation to measure the interaction between survivin and the HEAT1-repeat, PI3K and FATC domains of DNA-PKcs and PI3K domain of ATM. Results represent means + SD (n = 3; ***, P < 0.001; t test). C, NanoBiT protein interaction analysis performed by generating survivin and PI3K domain constructs fused with LgBiT (large subunit of luciferase) and SmBiT (small subunit of luciferase). HEK293T cells were cotransfected with the corresponding constructs and luciferase signal was measured with an ELISA reader. Results represent means + SD (n = 3; **, P < 0.01; t test). D, Whole cell lysates of SW480 cells, transiently transfected with EYFP, Surv. wt, or ΔBIR expression constructs, were isolated one hour after irradiation with 4 Gy. EYFP fusion proteins were immunoprecipitated (IP) using anti-GFP antibody and DNA-PKcs with an anti-DNA-PKcs antibody. Mouse and rabbit mAb isotype controls (IgG) served as controls. Subsequently, EYFP constructs and DNA-PKcs were detected by Western immunoblotting. Data display one representative out of two independent experiments. E, Schematic representation of the interaction domains of survivin and DNA-PKcs. F,In silico molecular docking analysis. Binding-free energy (ΔG) and proximity analysis of potential interaction amino acids of survivin and DNA-PKcs are shown. G, Representative docked poses of survivin (cyan) with DNA-PKcs-PI3K domain (red).

Figure 2.

BIR domain of survivin interacts with the PI3K domain of DNA-PKcs. A, Schematic presentation of the domains of survivin, DNA-PKcs, and ATM, which were subjected to FACS-FRET analyses, with vectors expressing the enhanced yellow or cyan fluorescent protein (EYFP, ECFP). Surv. wt, survivin wild type; ΔBIR, survivin BIR domain deletion mutant; PI3K, DNA-PKcs PI3K domain; HEAT1, DNA-PKcs Huntingtin-elongation factor 3-PP2A-TOR1 (HEAT1) domain; FATC, FRAP-ATM-TRRAP C-terminal (FATC) domain-ECFP; PI3K-ATM, ATM PI3K domain. B, FACS-FRET analysis in SW480 or HEK293T cells performed at 1 hour after 4 Gy irradiation to measure the interaction between survivin and the HEAT1-repeat, PI3K and FATC domains of DNA-PKcs and PI3K domain of ATM. Results represent means + SD (n = 3; ***, P < 0.001; t test). C, NanoBiT protein interaction analysis performed by generating survivin and PI3K domain constructs fused with LgBiT (large subunit of luciferase) and SmBiT (small subunit of luciferase). HEK293T cells were cotransfected with the corresponding constructs and luciferase signal was measured with an ELISA reader. Results represent means + SD (n = 3; **, P < 0.01; t test). D, Whole cell lysates of SW480 cells, transiently transfected with EYFP, Surv. wt, or ΔBIR expression constructs, were isolated one hour after irradiation with 4 Gy. EYFP fusion proteins were immunoprecipitated (IP) using anti-GFP antibody and DNA-PKcs with an anti-DNA-PKcs antibody. Mouse and rabbit mAb isotype controls (IgG) served as controls. Subsequently, EYFP constructs and DNA-PKcs were detected by Western immunoblotting. Data display one representative out of two independent experiments. E, Schematic representation of the interaction domains of survivin and DNA-PKcs. F,In silico molecular docking analysis. Binding-free energy (ΔG) and proximity analysis of potential interaction amino acids of survivin and DNA-PKcs are shown. G, Representative docked poses of survivin (cyan) with DNA-PKcs-PI3K domain (red).

Close modal

Specific residues located in the BIR domain of survivin are vital for the interaction with the PI3K domain of DNA-PKcs

We next aimed to further dissect the molecular basis of this interrelationship and investigated the interaction interfaces of the DNA-PKcs PI3K domain by performing in silico molecular docking analyses. Post-docking free energy and bond-forming proximity analyses revealed specific residues, such as S20, F27, C31, D53, and W67, to display low-binding free energies and high affinity to the PI3K domain (Fig. 2E). On the basis of the lowest binding free energy, S20 and W67 residues (scheme depicted in Fig. 2F) were next subjected to in silico mutagenesis. A phospho-mimicking mutant of S20 (S20D) or alanine substitution of W67 (W67A) showed increased side-chain bond forming atom distances and less attraction compared with wt conditions (Fig. 2G), indicating a prominent role of both residues for the interaction. To further corroborate these findings, site-directed mutagenesis of S20D, F27A, E29A (served as a negative control), C31A, D53A, and W67A was performed and recombinant proteins were transiently expressed in SW480 cells. FACS-FRET measurement revealed a decreased interaction of survivin and the DNA-PKcs PI3K domain for all residues mutated, with the highest degree of reduction (∼70%) observed for the S20D-W67A double mutant construct (Fig. 3A). This finding was confirmed in DLD-1 and HCT-15 colorectal cancer cell lines (Fig. 3B). Moreover, coimmunoprecipitation experiments verified a gradual decrease in survivin and DNA-PKcs PI3K domain interaction in the case of S20D, W67A, or S20D-W67A double mutant (Fig. 3C). On a functional level, cells expressing W67A or S20D-W67A mutants (expression profiles in Supplementary Fig. S2A) failed to rescue 3D clonogenic radiation survival upon knockdown of endogenous survivin in SW480 (Fig. 3D; Supplementary Fig. S2B) and in DLD-1 cells (Supplementary Fig. S1A) as opposed to Surv. wt cells. Finally, residual γH2AX/53PB1 foci in SW480 (Fig. 3E; Supplementary Fig. S2C) or DLD-1 (Supplementary Fig. S1B) cultures were significantly increased upon knockdown of endogenous survivin in W67A and S20D-W67A mutants, whereas Surv. wt overexpression rescued the phenotype. In conclusion, these findings favored a model in which the S20 and W67 residues mediated survivin-modulated radiation survival and DNA DSB repair.

Figure 3.

Specific amino acids located in the BIR domain of survivin are essential for the interaction with the PI3K domain of DNA-PKcs. A, FACS–FRET interaction analysis of EYFP-tagged BIR domain deletion (ΔBIR) and alanine and/or aspartic acid substitution forms of survivin interaction sites in SW480 cells were measured 1 hour after 4 Gy irradiation. Results on single (left) and double/triple mutations (right) are given as means + SD (n = 3; *, P < 0.05; **, P < 0.01; ***, P < 0.001; t test). B, Assessment of the interaction potentials of survivin S20D, W67A, and S20D-W67A double mutant was performed 1 hour after 4 Gy irradiation by FACS–FRET analysis in DLD-1 and HCT-15 colorectal cells. Results are given as means + SD (n = 3; *, P < 0.05; **, P < 0.01; ***, P < 0.001; t test). C, SW480 cells, transiently transfected with PI3K-Flag and EYFP-tagged Surv. wt, survivin ΔBIR, S20D, W67A, or survivin S20D+W67A expression constructs; whole cell lysates were isolated 1 hour after 4 Gy irradiation. Mock-transfected cells served as a control. EYFP fusion proteins were immunoprecipitated using an anti-GFP antibody and PI3K domain of DNA-PKcs was precipitated with an anti-Flag antibody. Mouse and rabbit mAb isotype controls (IgG) served as controls. Subsequently, EYFP constructs and PI3K domain of DNA-PKcs were detected by Western blotting. β-Actin served as a loading control. One representative out of two independent experiments is shown. D, Clonogenic radiation survival was analyzed after irradiation with 0, 2, 4, or 6 Gy (single dose). Results represent means ± SD (n ≥ 3; *, P < 0.05; **, P < 0.01; ***, P < 0.001; t test). E, SW480 cells, stably expressing indicated survivin mutants were subjected to siRNA-mediated survivin attenuation and irradiated with 2 Gy. At 24 hours after irradiation, cells were stained for γH2AX/53BP1 and DAPI. Nuclear γH2AX/53BP1 foci were microscopically counted (50 nuclei per experiment). Results represent mean foci per cell + SD (n ≥ 3; ***, P < 0.001; t test).

Figure 3.

Specific amino acids located in the BIR domain of survivin are essential for the interaction with the PI3K domain of DNA-PKcs. A, FACS–FRET interaction analysis of EYFP-tagged BIR domain deletion (ΔBIR) and alanine and/or aspartic acid substitution forms of survivin interaction sites in SW480 cells were measured 1 hour after 4 Gy irradiation. Results on single (left) and double/triple mutations (right) are given as means + SD (n = 3; *, P < 0.05; **, P < 0.01; ***, P < 0.001; t test). B, Assessment of the interaction potentials of survivin S20D, W67A, and S20D-W67A double mutant was performed 1 hour after 4 Gy irradiation by FACS–FRET analysis in DLD-1 and HCT-15 colorectal cells. Results are given as means + SD (n = 3; *, P < 0.05; **, P < 0.01; ***, P < 0.001; t test). C, SW480 cells, transiently transfected with PI3K-Flag and EYFP-tagged Surv. wt, survivin ΔBIR, S20D, W67A, or survivin S20D+W67A expression constructs; whole cell lysates were isolated 1 hour after 4 Gy irradiation. Mock-transfected cells served as a control. EYFP fusion proteins were immunoprecipitated using an anti-GFP antibody and PI3K domain of DNA-PKcs was precipitated with an anti-Flag antibody. Mouse and rabbit mAb isotype controls (IgG) served as controls. Subsequently, EYFP constructs and PI3K domain of DNA-PKcs were detected by Western blotting. β-Actin served as a loading control. One representative out of two independent experiments is shown. D, Clonogenic radiation survival was analyzed after irradiation with 0, 2, 4, or 6 Gy (single dose). Results represent means ± SD (n ≥ 3; *, P < 0.05; **, P < 0.01; ***, P < 0.001; t test). E, SW480 cells, stably expressing indicated survivin mutants were subjected to siRNA-mediated survivin attenuation and irradiated with 2 Gy. At 24 hours after irradiation, cells were stained for γH2AX/53BP1 and DAPI. Nuclear γH2AX/53BP1 foci were microscopically counted (50 nuclei per experiment). Results represent mean foci per cell + SD (n ≥ 3; ***, P < 0.001; t test).

Close modal

Analyzing the attraction between the BIR and PI3K domains in a heterotetramer complex model

Given that the previous docking approach (∼150 poses, Fig. 2E and G) indicated two prominent poses, we next aimed to extend the interaction analyses in a broader scale approach (∼100k poses). We calculated the minimum distance between any docked survivin BIR domain and the entire PI3K region. The two-dimensional histogram of |{d_{{\rm{BIR}}}}$| and the Rosetta interface score |{I_{{\rm{sc}}}}$| confirmed the existence of two clusters, with the BIR domain atoms showing 20 to 40 Å distance to the PI3K domain in the first cluster (Fig. 4A, left) and 1 to 2Å in the second cluster (Fig. 4A, right). According to the first set of molecular docking analysis (Fig. 2E and G) and experimental findings depicted in Fig. 3, large-scale global docking data were next filtered for spatial proximity of S20 and W67 residues to the DNA-PKcs PI3K domain. Two differing states were computed for S20 and W67, showing contradictory configurations and orientations towards the PI3K region. For geometrical constraints it was not possible to fulfil both poses at the same time. To address this discrepancy, we hypothesized two distinct spatial binding conformations: chain B of survivin via S20 in close proximity to chain A of the head-domain (residues 2802–4128) of DNA-PKcs and chain C of survivin in close proximity via W67 to chain D of the head-domain of DNA-PKcs. These analyses resulted in a heterotetramer model, in which survivin binds to the surface of a DNA-PKcs dimer with spatial constraints and an energetic minimum independently fulfilled. The model also indicated an interface between the FKBP-rapamycin-binding (FRB) domain and the FAT region of the two DNA-PKcs head domains (Fig. 4B). To experimentally validate live-cell occurrence of the heterotetramer complex, we performed FACS-FRET assays with recombinant ECFP/EYFP PI3K domain constructs confirming a dimerization in mock- and control siRNA-treated SW480 cells (Fig. 4C). Importantly, although dimerization was not affected by siRNA knockdown of survivin, Flag-survivin overexpression significantly increased the dimerization by more than 40% (Fig. 4C). These results suggested that the heterotetramer structure may transform to an established DNA-PKcs head-dimer complex by survivin binding. To further establish an energetic superiority of the molecular structure, three individual molecular dynamics simulations were performed on the heterotetramer, a head-dimer interacting via the FRB and FAT regions and a reference model via in silico docking of two identical FRB domain sites. For the heterotetramer model, convergence in RMSD was observed for both kinases (chain A and D) after ∼50ns, although chain A stabilized significantly faster (Fig. 4D). The RMSD and Rg results of the head-dimer were essentially identical to the heterotetramer for both single head domains (chain A and D). Only for chain A, a lower value in Rg was observed. In contrast, the reference model showed differing behavior in both RMSD and Rg for chain D (Fig. 4D). The RMSD was stable for survivin chain B where S20 was in proximity to the PI3K domain (Fig. 4E). In particular, survivin chain C, in which W67 was in proximity to the PI3K domain, showed deviations between 30 ns and 80 ns for both RMSD and Rg, suggesting structural changes (Fig. 4E). The head dimer displayed slightly fluctuating values over the simulation time and showed an opening of the active site for both chains only after ∼50 ns. Notably, the chains of the heterotetramer structure exhibited different behavior with a far greater surface accessibility for chain D (Fig. 4F). Particle density analysis further revealed that the active site regions of the heterotetramer structure were diverging, particularly for chain D (green line) in proximity with W67. This conformation resulted in a more accessible active site region at the end of the simulation (Fig. 4G, bottom) compared with the beginning (Fig. 4G, top).

Figure 4.

In silico molecular docking revealed a DNA-PKcs–survivin–DNA-PKcs heterotetramer with increased accessibility for the active site of PI3K domain. A, Distribution of distances between PI3K–BIR domain residues (dBIR [Å]) versus interface scores (Isc). Right, closer proximity (dBIR ≤20 Å) distribution. B, Representation of the structure of heterotetramer complex. Chain-A head-domain (green), chain-B survivin (cyan), chain-C survivin (purple), chain-D head-domain (yellow), PI3K domains of both head-domains (white), and active sites (red). C, SW480 cells were cotransfected with ECFP-PI3K and EYFP-PI3K and mock/control/survivin siRNA or survivin-Flag construct and subjected to FACS-FRET analyses at 1 hour after a 4 Gy irradiation. D, RMSD (top) and Rg (bottom) analyses during MD simulations of heterotetramer (red), head-dimer (blue), and reference model (orange) structures. E, RMSD (top) and Rg (bottom) analyses during MD simulations of survivin dimer for both chains, B (red) and C (blue). F, Particle density analysis of the active site region of PI3K domain for both chains, A (orange) and D (green). G, Visualization of the active site of energy minimized and equilibrated heterotetramer at the start (top) and end (bottom) of MD simulation. Chain-A head-domain (green), chain-B survivin (cyan), chain-C survivin (purple), chain-D head-domain (yellow), PI3K domains of both head-domains (white), and active sites (red). H, Effect of survivin on the kinase activity of DNA-PKcs. EYFP or Surv. wt-EYFP–expressing SW480 cells were subjected to survivin siRNA transfection and irradiated with 4 Gy. One hour after irradiation, cells were lysed and immunoprecipitated with anti-YFP or IgG antibodies. Equal amounts of immunoprecipitated proteins were used for kinase reaction and quantified by phosphoimaging. Results represent means + SD (n = 5; *, P < 0.05; **, P < 0.01; ***, P < 0.001; t test vs. EGFP). Pos. control, DNA-PK + Substrate-peptide; IgG control, Pos. control + IgG pull-down; EGFP, Pos. control + EGFP pull-down; Surv. wt, Pos. control + Surv. wt pull-down; DNA-PK inh., Pos. control + DNA-PK inhibitor; Neg. control, DNA-PK.

Figure 4.

In silico molecular docking revealed a DNA-PKcs–survivin–DNA-PKcs heterotetramer with increased accessibility for the active site of PI3K domain. A, Distribution of distances between PI3K–BIR domain residues (dBIR [Å]) versus interface scores (Isc). Right, closer proximity (dBIR ≤20 Å) distribution. B, Representation of the structure of heterotetramer complex. Chain-A head-domain (green), chain-B survivin (cyan), chain-C survivin (purple), chain-D head-domain (yellow), PI3K domains of both head-domains (white), and active sites (red). C, SW480 cells were cotransfected with ECFP-PI3K and EYFP-PI3K and mock/control/survivin siRNA or survivin-Flag construct and subjected to FACS-FRET analyses at 1 hour after a 4 Gy irradiation. D, RMSD (top) and Rg (bottom) analyses during MD simulations of heterotetramer (red), head-dimer (blue), and reference model (orange) structures. E, RMSD (top) and Rg (bottom) analyses during MD simulations of survivin dimer for both chains, B (red) and C (blue). F, Particle density analysis of the active site region of PI3K domain for both chains, A (orange) and D (green). G, Visualization of the active site of energy minimized and equilibrated heterotetramer at the start (top) and end (bottom) of MD simulation. Chain-A head-domain (green), chain-B survivin (cyan), chain-C survivin (purple), chain-D head-domain (yellow), PI3K domains of both head-domains (white), and active sites (red). H, Effect of survivin on the kinase activity of DNA-PKcs. EYFP or Surv. wt-EYFP–expressing SW480 cells were subjected to survivin siRNA transfection and irradiated with 4 Gy. One hour after irradiation, cells were lysed and immunoprecipitated with anti-YFP or IgG antibodies. Equal amounts of immunoprecipitated proteins were used for kinase reaction and quantified by phosphoimaging. Results represent means + SD (n = 5; *, P < 0.05; **, P < 0.01; ***, P < 0.001; t test vs. EGFP). Pos. control, DNA-PK + Substrate-peptide; IgG control, Pos. control + IgG pull-down; EGFP, Pos. control + EGFP pull-down; Surv. wt, Pos. control + Surv. wt pull-down; DNA-PK inh., Pos. control + DNA-PK inhibitor; Neg. control, DNA-PK.

Close modal

In line with a report that survivin binding increases the enzymatic activity of Aurora kinase B (35), we investigated whether survivin binding to DNA-PKcs PI3K domain may confer comparable effects. Thus, Surv-EGFP was immunoprecipitated from SW480 cells 1 hour after a irradiation with 4 Gy and subjected to an in vitro DNA-PKcs kinase activity assay. We observed a significant 45% increase in kinase activity in the presence of survivin (Fig. 4H). These findings were in favor of an increased accessibility by ATP and the substrates of DNA-PKcs, presumably resulting in enhanced kinase activity.

Survivin affects phosphorylation of DNA-PKcs substrates by regulating its kinase activity

On the basis of our findings of an increased DNA-PKcs kinase activity, we next investigated the functional consequences and aimed to uncover factors implicated in survivin/DNA-PKcs associated regulation of DNA DSB repair. We analyzed changes in the global proteome and phosphoproteome of SW480 cells treated with survivin siRNA, control siRNA, or DNA PKcs inhibitor, in wt cells or cells stably expressing EGFP or survivin-EGFP 1 hour after a 4 Gy exposure (Fig. 5A). Replicate treatments clustered together closely, as assessed by Euclidean distance (Fig. 5B) and PCA, while showing distinctive differences across conditions (Fig. 5B; Supplementary Fig. S5A). Differences across experimental conditions were not driven by changes in the cell cycle (Supplementary Fig. S6).

Figure 5.

Enhanced kinase activity of DNA-PKcs upon binding of survivin differentially increases the phosphorylation of DNA damage response proteins. A, Schematic representation of the experimental setup for LC-MS2/3 evaluation. B, LC-MS2 quantification of TMT-labeled, phospho-enriched and fractionated phosphopeptides revealed a similarly clustered differential phosphorylation profile. Datasets were combined, Z scores calculated, and hierarchical clustering performed using Euclidean distance between the samples. C, Analysis of threshold of interest candidates by setting log2-fold change cut-offs [≥0.25] for Surv-EGFP vs. EGFP, [≤−0.25] for DNA-PK inh. vs. siCtrl and siSurv vs. siCtrl conditions by plot analyses (n = 2). Qualified phosphosites from C (top graph) were next used as an input for further qualification considering the effect of DNA-PK inhibitor (bottom graph). D, The final set of phosphosites (orange in C bottom) normalized with proteome data represented by two heatmap graphs; first with high log2-fold change cut-offs [≤−0.5], [≤−0.5], and [≥0.5], second with normal log2-fold change cut-offs [≤−0.25], [≤−0.25], and [≥0.25] for DNA-PK inh. vs. siCtrl, siSurv vs. siCtrl, and Surv-EGFP vs. EGFP. Asterisks (*) indicate DNA damage/repair related proteins. E, Western blot validation of candidate Foxo3 S253 residue phosphorylation by conditions given in A. F, The abundance of phosphosite candidates (D) subjected to pathway analysis by using Pathway Commons and individual literatures. G, Consensus motif analysis of phosphosites revealed a highly conserved S/T-P motif. Bits on y-axis state the information content units as the weighted prevalence frequency of indicated motif generated by Seq2Logo. The input motifs were used from the hits in D. The generation of the logo based on probability weighted Kullback–Leibler logo type and Hobohm algorithm for sequence weighting type.

Figure 5.

Enhanced kinase activity of DNA-PKcs upon binding of survivin differentially increases the phosphorylation of DNA damage response proteins. A, Schematic representation of the experimental setup for LC-MS2/3 evaluation. B, LC-MS2 quantification of TMT-labeled, phospho-enriched and fractionated phosphopeptides revealed a similarly clustered differential phosphorylation profile. Datasets were combined, Z scores calculated, and hierarchical clustering performed using Euclidean distance between the samples. C, Analysis of threshold of interest candidates by setting log2-fold change cut-offs [≥0.25] for Surv-EGFP vs. EGFP, [≤−0.25] for DNA-PK inh. vs. siCtrl and siSurv vs. siCtrl conditions by plot analyses (n = 2). Qualified phosphosites from C (top graph) were next used as an input for further qualification considering the effect of DNA-PK inhibitor (bottom graph). D, The final set of phosphosites (orange in C bottom) normalized with proteome data represented by two heatmap graphs; first with high log2-fold change cut-offs [≤−0.5], [≤−0.5], and [≥0.5], second with normal log2-fold change cut-offs [≤−0.25], [≤−0.25], and [≥0.25] for DNA-PK inh. vs. siCtrl, siSurv vs. siCtrl, and Surv-EGFP vs. EGFP. Asterisks (*) indicate DNA damage/repair related proteins. E, Western blot validation of candidate Foxo3 S253 residue phosphorylation by conditions given in A. F, The abundance of phosphosite candidates (D) subjected to pathway analysis by using Pathway Commons and individual literatures. G, Consensus motif analysis of phosphosites revealed a highly conserved S/T-P motif. Bits on y-axis state the information content units as the weighted prevalence frequency of indicated motif generated by Seq2Logo. The input motifs were used from the hits in D. The generation of the logo based on probability weighted Kullback–Leibler logo type and Hobohm algorithm for sequence weighting type.

Close modal

To determine the modulatory effects of the survivin–DNA-PKcs interaction on kinase activity, we selected phosphopeptides increased by survivin overexpression and decreased by survivin knockdown (Fig. 5C, top) and further filtered for phosphorylation decrease by the DNA-PK inhibitor (Fig. 5C, bottom). Qualified phosphosites were normalized for changes in protein levels (subtraction from proteome changes) and presented in two separate heatmaps (Fig. 5D). To validate our results, we monitored forkhead-box-protein O3 (Foxo3) phosphorylation and found that S253 phosphorylation was decreased in response to DNA-PK inhibitor and increased upon survivin overexpression (Fig. 5E). The list of differentially regulated phosphosites in survivin overexpressing cells further revealed a vast number of DNA damage/repair-related proteins [marked with asterisks (*) in Fig. 5D] including substrates of DNA-PKcs, in line with markers implicated in chromosome organization, cell cycle, transcriptional regulation, and apoptosis (Fig. 5F; Supplementary File S1). Importantly, by analyzing the sequence motifs that increased in phosphorylation after survivin overexpression using amino acid sequence profile alignment and weighted logo generator Seq2Logo 2.0 (http://www.cbs.dtu.dk/biotools/Seq2Logo/), we discovered a highly-enriched conserved S/T-P consensus phosphorylation motif (Fig. 5G).

Evaluation of full-proteome data revealed a differential protein expression profile and similar clustering (Fig. 6A; Supplementary Fig. S5B). To qualify proteins regulated by survivin–DNA-PKcs interrelationship, directly and inversely impacted proteins were extracted across the data sets (Fig. 6B). Candidates that showed differential protein expression in a survivin- and DNA-PKcs-dependent manner (Fig. 6C and D) dominantly cover DNA repair-related proteins [marked with asterisks (*)]. In addition, by employing a pairwise intersection matrix representation using different cut-off log2 values, we addressed differentially regulated proteins with a sole dependence on survivin overexpression or knockdown. Directly regulated proteins covered known interactors like INCENP, CDCA8/Borealin, and Aurora kinase B (Fig. 6E) whereas inversely regulated proteins contained markers of transcriptional repression, nucleolar reorganization, immune response, and motility/invasiveness (Fig. 6F). In sum, these findings indicate that a modulation of DNA–DSB repair is mainly associated with a survivin–DNA-PKcs interaction.

Figure 6.

Proteomic landscape of differentially expressed proteins. A, LC-MS3 quantification of TMT-labeled and fractionated peptides revealed a similarly clustered differential expression profile of proteins. B, Analysis of directly (orange) and inversely (purple) regulated thresholds of interest by setting log2-fold change cut-offs: orange threshold of interest, [≥0.25] for Surv-EGFP vs. EGFP, [≤−0.25] for DNA-PK inh. vs. siCtrl and siSurv vs. siCtrl; purple threshold of interest [≤−0.25] for Surv-EGFP vs. EGFP, [≥0.25] for DNA-PK inh. vs. siCtrl and siSurv vs. siCtrl conditions by comparative distribution representations of the status of expressions as plot graphics. Results represent mean log2-fold change value (n = 2). Qualified proteins (orange and purple in B, top) were used as an input for further qualification to consider also the effect of DNA-PK inhibitor. C, Final set of directly regulated proteins (orange in B, bottom) represented as heatmap graphs by setting log2-fold change cut-offs, [≥0.25] for Surv-EGFP vs. EGFP, [≤−0.25] for DNA-PK inh. vs. siCtrl and siSurv vs. siCtrl. Asterisks (*) indicate proteins related to DNA damage/repair. D, Final set of inversely regulated proteins (purple in Fig. 6B, bottom) represented as heatmap graphs by setting log2-fold change cut-offs, [≤−0.25] for Surv-EGFP vs. EGFP, [≥0.25] for DNA-PK inh. vs. siCtrl, and siSurv vs. siCtrl conditions. Asterisks (*) indicate proteins related to DNA damage/repair. Pairwise intersection matrix classifies directly regulated (E) and inversely regulated (F) proteins by log2-fold changes as low [±0.25], high [±0.5], and highest [± 1.0] differentially expressed in response to survivin overexpression and survivin knockdown by representing the differential expression levels as heatmap graphs. DNA-PK inh., control siRNA + DNA-PK inhibitor-treated (1 μmol/L); siCtrl, control siRNA; siSurv, survivin siRNA; EGFP, survivin siRNA + EGFP overexpression. Surv-EGFP, survivin siRNA + survivin-EGFP overexpression.

Figure 6.

Proteomic landscape of differentially expressed proteins. A, LC-MS3 quantification of TMT-labeled and fractionated peptides revealed a similarly clustered differential expression profile of proteins. B, Analysis of directly (orange) and inversely (purple) regulated thresholds of interest by setting log2-fold change cut-offs: orange threshold of interest, [≥0.25] for Surv-EGFP vs. EGFP, [≤−0.25] for DNA-PK inh. vs. siCtrl and siSurv vs. siCtrl; purple threshold of interest [≤−0.25] for Surv-EGFP vs. EGFP, [≥0.25] for DNA-PK inh. vs. siCtrl and siSurv vs. siCtrl conditions by comparative distribution representations of the status of expressions as plot graphics. Results represent mean log2-fold change value (n = 2). Qualified proteins (orange and purple in B, top) were used as an input for further qualification to consider also the effect of DNA-PK inhibitor. C, Final set of directly regulated proteins (orange in B, bottom) represented as heatmap graphs by setting log2-fold change cut-offs, [≥0.25] for Surv-EGFP vs. EGFP, [≤−0.25] for DNA-PK inh. vs. siCtrl and siSurv vs. siCtrl. Asterisks (*) indicate proteins related to DNA damage/repair. D, Final set of inversely regulated proteins (purple in Fig. 6B, bottom) represented as heatmap graphs by setting log2-fold change cut-offs, [≤−0.25] for Surv-EGFP vs. EGFP, [≥0.25] for DNA-PK inh. vs. siCtrl, and siSurv vs. siCtrl conditions. Asterisks (*) indicate proteins related to DNA damage/repair. Pairwise intersection matrix classifies directly regulated (E) and inversely regulated (F) proteins by log2-fold changes as low [±0.25], high [±0.5], and highest [± 1.0] differentially expressed in response to survivin overexpression and survivin knockdown by representing the differential expression levels as heatmap graphs. DNA-PK inh., control siRNA + DNA-PK inhibitor-treated (1 μmol/L); siCtrl, control siRNA; siSurv, survivin siRNA; EGFP, survivin siRNA + EGFP overexpression. Surv-EGFP, survivin siRNA + survivin-EGFP overexpression.

Close modal

Multiple lines of evidences indicate that survivin is a radiation resistance factor and its knockdown or therapeutical attenuation radiosensitizes tumor cells in vitro and in vivo (36). However, the molecular determinants for impacting on the complex mechanisms of radiation response far exceed a simple inhibition of caspase-dependent apoptosis and autophagy, instead also suggesting effects on DSB processing and DNA damage repair (11, 12). Here, we demonstrate that the BIR domain of survivin is implicated in the modulation of a long-term radiation response (Fig. 1D) and short-term DNA damage repair as exemplified by a reduced clonogenic survival (6–9 days) and increased residual (24 hours) γH2AX/53BP1 foci detection (Fig. 1D and E), whereas both, the BIR and MicTub domains, are implicated in apoptosis regulation (Fig. 1C). This may indicate a selective requirement of a specific domain of survivin in the modulation of the DNA damage repair. Further, our corroborative findings by initial global blind molecular docking analysis (Fig. 2E and G), FACS-FRET, and biochemical analyses (Fig. 3) revealed an important functional role of residues S20 and W67, both located in the BIR domain (Figs. 2F and 3). As survivin represents a promising oncotherapeutic target, these findings may pave the way to develop novel DNA repair centered therapeutic strategies based on inhibition of S20- and W67-mediated interaction, for example, by developing small molecule inhibitors.

DNA-PKcs acts as a key kinase in the NHEJ repair pathway by its catalytic function, which is activated by a Ku70/80 holoenzyme complex formation in the presence of DNA (37). NHEJ rejoins broken DNA ends by final ligation, but for that process the ends need to be fixed in close proximity. The mechanism of bridging remains controversial (38) and recent models indicate a dimerization of DNA-PKcs that guides the formation of a synaptic complex (39). Cryoelectron microscopy and biochemical studies indeed demonstrate a dimerization of the DNA-PKcs subunit or the entire holoenzyme (14, 40, 41), enabling subsequent steps for DNA break processing and ligation (42). In addition, a recent study suggested the FAT/kinase domain regions to generate a self-heterodimer to autophosphorylate the PQR cluster (43). These reports are in close correlation with our model for the survivin–DNA-PKcs complex. In detail, in silico molecular docking analyses and experimental verifications (Fig. 4) resulted in a heterotetramer model where a survivin homodimer (3) binds to the surface of the PI3K domains of a DNA-PKcs dimer with spatially and energetically favored conformations. Notably, FACS-FRET measurements (Fig. 4C) indicated that the dimerization of DNA-PKcs PI3K domains was not affected by a siRNA-mediated knockdown of survivin and that binding of survivin may not be mandatory for primary DNA-PKcs dimerization. In addition, X-ray crystallography and CryoEM approaches revealed that DNA-PKcs dimerization seems to be a highly dynamic process displaying different spatial dimerization conformations (14, 40–44). Accordingly, it is likely that the survivin–DNA-PKcs complex may cover only a fraction of the DNA-PKcs protein dimer configurations. In favor of this assumption, complex formation time kinetics revealed a gradual decline of complex abundance following a peak 1 hour after irradiation (Supplementary Fig. S4). This observation is in line with our previous findings that a nuclear accumulation of survivin peaks at 1 hour after irradiation (12). Moreover, results from molecular simulations indicated a change in kinase active site accessibility upon binding of survivin with enhanced accessibility of potential substrates. These findings were analogous to a report that association of survivin results in an activation of the catalytic subunit of Aurora kinase B (35) and localization to its substrate histone H3. Although the exact molecular mechanism(s) of this activation remains elusive, recognition of phosphorylated (T3) histone H3 by a binding pocket in the BIR domain of survivin is involved in this process (45).

DNA-PKcs preferentially phosphorylates a S/T-Q motif present at a multitude of substrates including Artemis S516/S645, DNA Ligase IV S672, and H2AX S139 (39), whereas efficient phosphorylation of a S/T-hydr (hydrophobic residues: G, A, V, L, I, P, F, M, W) motif by DNA-PKcs was reported as well (46). Remarkably, analysis of phosphosites in our findings revealed an enrichment (60.5%) of S/T-hydr. motifs with a high abundance of S/T-P motifs (33.3%), whereas the S/T-Q motif was conserved only by 1.2% (Fig. 5G), suggesting a shift of the substrate preference of DNA-PKcs upon interaction with survivin. Indeed, our phosphoproteomics analyses did not indicate enrichment of S/T-Q DNA-PKcs substrates like Artemis, XRCC4, and DNA Ligase IV (15), but demonstrated markers including forkhead family transcription factor 3 (Foxo3) S253 and Filamin-A T2336 phosphorylation with S/T-hydr motifs. With the dominant changes on substrate specificity, we propose survivin to change DNA-PKcs activity towards previously less-known downstream substrates. However, the question arises whether survivin binding causes additional conceptional impact on modulation of the cellular radiation response.

Foxo3 is a member of forkhead family of transcription factors, which mainly facilitates the transcriptional regulation of apoptosis genes such as Bim, Noxa, Puma, FasL, and TRAIL (47). T32 and S253 phosphorylation of Foxo3 targeted to 14-3-3 protein results in deactivation or ubiquitin-mediated degradation (47). More recently, Foxo3 was shown to be phosphorylated at S413 by DNA-PKcs under nutrient starvation (48). Our findings indicate Foxo3 S253 to be an additional phosphorylation site regulated by DNA-PKcs that suggests an additional mechanism to prevent apoptosis induction in relationship with the survivin–DNA-PKcs complex. Previous findings further indicate that the multi-complex interacting scaffold Filamin-A is required to stabilize the DNA-PK holoenzyme complex in a breast cancer type 1 susceptibility protein (BRCA1) dependent manner (49). In addition, Filamin-A knockout cells are deficient in DNA repair (50). In conclusion, our findings support the idea of a multi-effector dynamic DNA-PK holoenzyme complex with expanded functional capacity by the recruitment and binding of survivin and possible other proteins such as Filamin-A.

Among the differentially expressed proteins inversely regulated by the survivin–DNA-PKcs heterotetramer, we uncovered high-mobility group (HMGA, HMGN) family proteins implicated in the regulation of transcription and chromosome condensation via altering the chromatin architecture (51). Markedly, via altering the steady-state complex of the DNA-PK holoenzyme and delaying the release of DNA-PKcs from the DNA DSB ends, HMGA2 overexpression revealed an inhibitory effect on DNA-PKcs during the NHEJ process (52). Thus, downregulation of the expression of HMGA2 by the survivin–DNA-PKcs complex may preserve the DNA binding of DNA-PKcs and increase efficacy and resolution of the NHEJ process.

In sum, our study advances our conceptual understanding of the regulation of DNA-PKcs kinase activity by interaction with the IAP survivin towards an altered substrate specificity and may broaden the knowledge of the impact of survivin on DNA–DSB repair and radiation responsiveness. In addition, the data add a novel facet in the diversity of mechanisms implicated in the radiation protective effect of survivin and are translationally relevant for devising novel survivin-tailored strategies and to develop new avenues of drug discovery including targeting residue S20- and W67-mediated survivin–DNA-PKcs interaction.

Ö. Güllülü reports grants from Deutsche Forschungsgemeinschaft (DFG) during the conduct of the study. B.E. Mayer reports grants, personal fees, and non-financial support from DFG during the conduct of the study. C. Petraki reports grants from Deutsche Forschungsgemeinschaft and International Graduate College (GRK) 1657 during the conduct of the study. M. Hoffmann reports grants from Deutsche Forschungsgemeinschaft (DFG) during the conduct of the study. M.J. Dombrowsky reports grants from DFG (GRK 1657) during the conduct of the study. P. Kunzmann reports other support from Hesse state during the conduct of the study. K. Hamacher reports grants from Deutsche Forschungsgemeinschaft during the conduct of the study and grants from Bundesministerium für Bildung und Forschung outside the submitted work. C. Münch reports grants from DFG during the conduct of the study. F. Rödel reports grants from German Research Foundation (DFG) during the conduct of the study. No disclosures were reported by the other authors.

Ö. Güllülü: Conceptualization, validation, investigation, methodology, writing–original draft, writing–review and editing. S. Hehlgans: Conceptualization, validation, investigation, methodology, writing–original draft, writing–review and editing. B.E. Mayer: Software, formal analysis, investigation, methodology, writing–original draft. I. Gößner: Formal analysis, investigation, methodology. C. Petraki: Validation, investigation, methodology, writing–original draft. M. Hoffmann: Validation, investigation, writing–original draft. M.J. Dombrowsky: Software, investigation, methodology, writing–original draft. P. Kunzmann: Software, investigation, methodology, writing–original draft. K. Hamacher: Conceptualization, software, formal analysis, validation, investigation, writing–original draft. K. Strebhardt: Supervision, validation, investigation, methodology, writing–original draft, writing–review and editing. E. Fokas: Supervision, validation, methodology, writing–original draft, writing–review and editing. C. Rödel: Resources, supervision, validation, investigation, methodology, writing–original draft, writing–review and editing. C. Münch: Conceptualization, formal analysis, validation, investigation, methodology, writing–original draft, writing–review and editing. F. Rödel: Conceptualization, formal analysis, supervision, funding acquisition, validation, investigation, methodology, writing–original draft, project administration, writing–review and editing.

This study was supported by a grant of the German Research Foundation (DFG; Graduate school GRK 1657, projects 2B and 1A). C. Münch was supported by the German Research Foundation Emmy Noether Programm (DFG, MU 4216/1-1). The authors thank Professor R.H. Stauber (University of Mainz, Germany) and Professor Michael Schindler (University of Tübingen) for kindly providing survivin cDNA and ECFP-EYFP fusion vector and gratefully acknowledge Mr. Julius Oppermann’s excellent technical assistance. Calculations for this research were conducted on the Lichtenberg high performance computer of the TU Darmstadt. The Graphical abstract was created with BioRender.com.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Roos
WP
,
Thomas
AD
,
Kaina
B
. 
DNA damage and the balance between survival and death in cancer biology
.
Nat Rev Cancer
2016
;
16
:
20
33
.
2.
Ambrosini
G
,
Adida
C
,
Altieri
DC
. 
A novel anti-apoptosis gene, survivin, expressed in cancer and lymphoma
.
Nat Med
1997
;
3
:
917
21
.
3.
Verdecia
MA
,
Huang
H
,
Dutil
E
,
Kaiser
DA
,
Hunter
T
,
Noel
JP
. 
Structure of the human anti-apoptotic protein survivin reveals a dimeric arrangement
.
Nat Struct Biol
2000
;
7
:
602
8
.
4.
Altieri
DC
. 
Survivin, cancer networks and pathway-directed drug discovery
.
Nat Rev Cancer
2008
;
8
:
61
70
.
5.
Colnaghi
R
,
Connell
CM
,
Barrett
RM
,
Wheatley
SP
. 
Separating the anti-apoptotic and mitotic roles of survivin
.
J Biol Chem
2006
;
281
:
33450
6
.
6.
Honda
R
,
Korner
R
,
Nigg
EA
. 
Exploring the functional interactions between Aurora B, INCENP, and survivin in mitosis
.
Mol Biol Cell
2003
;
14
:
3325
41
.
7.
Kanwar
JR
,
Kamalapuram
SK
,
Kanwar
RK
. 
Targeting survivin in cancer: the cell-signalling perspective
.
Drug Discov Today
2011
;
16
:
485
94
.
8.
Chakravarti
A
,
Zhai
GG
,
Zhang
M
,
Malhotra
R
,
Latham
DE
,
Delaney
MA
, et al
Survivin enhances radiation resistance in primary human glioblastoma cells via caspase-independent mechanisms
.
Oncogene
2004
;
23
:
7494
506
.
9.
Iwasa
T
,
Okamoto
I
,
Suzuki
M
,
Nakahara
T
,
Yamanaka
K
,
Hatashita
E
, et al
Radiosensitizing effect of YM155, a novel small-molecule survivin suppressant, in non-small cell lung cancer cell lines
.
Clin Cancer Res
2008
;
14
:
6496
504
.
10.
Rodel
F
,
Hoffmann
J
,
Distel
L
,
Herrmann
M
,
Noisternig
T
,
Papadopoulos
T
, et al
Survivin as a radioresistance factor, and prognostic and therapeutic target for radiotherapy in rectal cancer
.
Cancer Res
2005
;
65
:
4881
7
.
11.
Wang
X
,
Beitler
JJ
,
Huang
W
,
Chen
G
,
Qian
G
,
Magliocca
K
, et al
Honokiol radiosensitizes squamous cell carcinoma of the head and neck by downregulation of survivin
.
Clin Cancer Res
2018
;
24
:
858
69
.
12.
Capalbo
G
,
Dittmann
K
,
Weiss
C
,
Reichert
S
,
Hausmann
E
,
Rodel
C
, et al
Radiation-induced survivin nuclear accumulation is linked to DNA damage repair
.
Int J Radiat Oncol Biol Phys
2010
;
77
:
226
34
.
13.
Hill
R
,
Lee
PW
. 
The DNA-dependent protein kinase (DNA-PK): More than just a case of making ends meet?
Cell Cycle
2010
;
9
:
3460
9
.
14.
Spagnolo
L
,
Rivera-Calzada
A
,
Pearl
LH
,
Llorca
O
. 
Three-dimensional structure of the human DNA-PKcs/Ku70/Ku80 complex assembled on DNA and its implications for DNA DSB repair
.
Mol Cell
2006
;
22
:
511
9
.
15.
Jette
N
,
Lees-Miller
SP
. 
The DNA-dependent protein kinase: A multifunctional protein kinase with roles in DNA double strand break repair and mitosis
.
Prog Biophys Mol Biol
2015
;
117
:
194
205
.
16.
Goodwin
JF
,
Knudsen
KE
. 
Beyond DNA repair: DNA-PK function in cancer
.
Cancer Discov
2014
;
4
:
1126
39
.
17.
Banning
C
,
Votteler
J
,
Hoffmann
D
,
Koppensteiner
H
,
Warmer
M
,
Reimer
R
, et al
A flow cytometry-based FRET assay to identify and analyse protein-protein interactions in living cells
.
PLoS One
2010
;
5
:
e9344
.
18.
Hehlgans
S
,
Petraki
C
,
Reichert
S
,
Cordes
N
,
Rodel
C
,
Rodel
F
. 
Double targeting of Survivin and XIAP radiosensitizes 3D grown human colorectal tumor cells and decreases migration
.
Radiother Oncol
2013
;
108
:
32
9
.
19.
Hehlgans
S
,
Oppermann
J
,
Reichert
S
,
Fulda
S
,
Rodel
C
,
Rodel
F
. 
The SMAC mimetic BV6 sensitizes colorectal cancer cells to ionizing radiation by interfering with DNA repair processes and enhancing apoptosis
.
Radiat Oncol
2015
;
10
:
198
.
20.
Rothkamm
K
,
Lobrich
M
. 
Evidence for a lack of DNA double-strand break repair in human cells exposed to very low x-ray doses
.
Proc Natl Acad Sci U S A
2003
;
100
:
5057
62
.
21.
Sastry
GM
,
Adzhigirey
M
,
Day
T
,
Annabhimoju
R
,
Sherman
W
. 
Protein and ligand preparation: parameters, protocols, and influence on virtual screening enrichments
.
J Comput Aided Mol Des
2013
;
27
:
221
34
.
22.
Kozakov
D
,
Brenke
R
,
Comeau
SR
,
Vajda
S
. 
PIPER: an FFT-based protein docking program with pairwise potentials
.
Proteins
2006
;
65
:
392
406
.
23.
Schneidman-Duhovny
D
,
Inbar
Y
,
Nussinov
R
,
Wolfson
HJ
. 
PatchDock and SymmDock: servers for rigid and symmetric docking
.
Nucleic Acids Res
2005
;
33
:
W363
W7
.
24.
Mashiach
E
,
Schneidman-Duhovny
D
,
Andrusier
N
,
Nussinov
R
,
Wolfson
HJ
. 
FireDock: a web server for fast interaction refinement in molecular docking
.
Nucleic Acids Res
2008
;
36
:
W229
32
.
25.
Jacobson
MP
,
Pincus
DL
,
Rapp
CS
,
Day
TJ
,
Honig
B
,
Shaw
DE
, et al
A hierarchical approach to all-atom protein loop prediction
.
Proteins
2004
;
55
:
351
67
.
26.
Pettersen
EF
,
Goddard
TD
,
Huang
CC
,
Couch
GS
,
Greenblatt
DM
,
Meng
EC
, et al
UCSF Chimera—a visualization system for exploratory research and analysis
.
J Comput Chem
2004
;
25
:
1605
12
.
27.
Webb
B
,
Sali
A
. 
Comparative protein structure modeling using MODELLER
.
Curr Protoc Bioinformatics
2016
;
54
:
5 6 1–5 6 37
.
28.
Leaver-Fay
A
,
Tyka
M
,
Lewis
SM
,
Lange
OF
,
Thompson
J
,
Jacak
R
, et al
ROSETTA3: an object-oriented software suite for the simulation and design of macromolecules
.
Methods Enzymol
2011
;
487
:
545
74
.
29.
Alford
RF
,
Leaver-Fay
A
,
Jeliazkov
JR
,
O’Meara
MJ
,
DiMaio
FP
,
Park
H
, et al
The rosetta all-atom energy function for macromolecular modeling and design
.
J Chem Theory Comput
2017
;
13
:
3031
48
.
30.
Kunzmann
P
,
Hamacher
K
. 
Biotite: a unifying open source computational biology framework in Python
.
BMC Bioinformatics
2018
;
19
:
346
.
31.
Gray
JJ
,
Moughon
S
,
Wang
C
,
Schueler-Furman
O
,
Kuhlman
B
,
Rohl
CA
, et al
Protein-protein docking with simultaneous optimization of rigid-body displacement and side-chain conformations
.
J Mol Biol
2003
;
331
:
281
99
.
32.
Abraham
MJ
,
Murtola
T
,
Schulz
R
,
Páll
S
,
Smith
JC
,
Hess
B
, et al
GROMACS: high performance molecular simulations through multi-level parallelism from laptops to supercomputers
.
SoftwareX
2015
;
1–2
:
19
25
.
33.
Vanommeslaeghe
K
,
Hatcher
E
,
Acharya
C
,
Kundu
S
,
Zhong
S
,
Shim
J
, et al
CHARMM general force field: a force field for drug-like molecules compatible with the CHARMM all-atom additive biological force fields
.
J Comput Chem
2010
;
31
:
671
90
.
34.
Potel
CM
,
Lin
MH
,
Heck
AJR
,
Lemeer
S
. 
Defeating major contaminants in Fe(3+)- immobilized metal ion affinity chromatography (IMAC) phosphopeptide enrichment
.
Mol Cell Proteomics
2018
;
17
:
1028
34
.
35.
Chen
J
,
Jin
S
,
Tahir
SK
,
Zhang
H
,
Liu
X
,
Sarthy
AV
, et al
Survivin enhances Aurora-B kinase activity and localizes Aurora-B in human cells
.
J Biol Chem
2003
;
278
:
486
90
.
36.
Rodel
F
,
Sprenger
T
,
Kaina
B
,
Liersch
T
,
Rodel
C
,
Fulda
S
, et al
Survivin as a prognostic/predictive marker and molecular target in cancer therapy
.
Curr Med Chem
2012
;
19
:
3679
88
.
37.
Chang
HHY
,
Pannunzio
NR
,
Adachi
N
,
Lieber
MR
. 
Non-homologous DNA end joining and alternative pathways to double-strand break repair
.
Nat Rev Mol Cell Biol
2017
;
18
:
495
506
.
38.
Graham
TG
,
Walter
JC
,
Loparo
JJ
. 
Two-stage synapsis of DNA ends during non-homologous end joining
.
Mol Cell
2016
;
61
:
850
8
.
39.
Lees-Miller
SP
,
Anderson
CW
. 
The human double-stranded DNA-activated protein kinase phosphorylates the 90-kDa heat-shock protein, hsp90 alpha at two NH2-terminal threonine residues
.
J Biol Chem
1989
;
264
:
17275
80
.
40.
Sibanda
BL
,
Chirgadze
DY
,
Ascher
DB
,
Blundell
TL
. 
DNA-PKcs structure suggests an allosteric mechanism modulating DNA double-strand break repair
.
Science
2017
;
355
:
520
4
.
41.
Hammel
M
,
Yu
Y
,
Mahaney
BL
,
Cai
B
,
Ye
R
,
Phipps
BM
, et al
Ku and DNA-dependent protein kinase dynamic conformations and assembly regulate DNA binding and the initial non-homologous end joining complex
.
J Biol Chem
2010
;
285
:
1414
23
.
42.
DeFazio
LG
,
Stansel
RM
,
Griffith
JD
,
Chu
G
. 
Synapsis of DNA ends by DNA-dependent protein kinase
.
EMBO J
2002
;
21
:
3192
200
.
43.
Baretic
D
,
Maia de Oliveira
T
,
Niess
M
,
Wan
P
,
Pollard
H
,
Johnson
CM
, et al
Structural insights into the critical DNA damage sensors DNA-PKcs, ATM and ATR
.
Prog Biophys Mol Biol
2019
;
147
:
4
16
.
44.
Chaplin
AK
,
Hardwick
SW
,
Liang
S
,
Kefala Stavridi
A
,
Hnizda
A
,
Cooper
LR
, et al
Dimers of DNA-PK create a stage for DNA double-strand break repair
.
Nat Struct Mol Biol
2020
.
45.
Kelly
AE
,
Ghenoiu
C
,
Xue
JZ
,
Zierhut
C
,
Kimura
H
,
Funabiki
H
. 
Survivin reads phosphorylated histone H3 threonine 3 to activate the mitotic kinase Aurora B
.
Science
2010
;
330
:
235
9
.
46.
Yu
Y
,
Wang
W
,
Ding
Q
,
Ye
R
,
Chen
D
,
Merkle
D
, et al
DNA-PK phosphorylation sites in XRCC4 are not required for survival after radiation or for V(D)J recombination
.
DNA Repair
2003
;
2
:
1239
52
.
47.
Liu
Y
,
Ao
X
,
Ding
W
,
Ponnusamy
M
,
Wu
W
,
Hao
X
, et al
Critical role of FOXO3a in carcinogenesis
.
Mol Cancer
2018
;
17
:
104
.
48.
Shiga
S
,
Murata
Y
,
Hashimoto
T
,
Urushihara
Y
,
Fujishima
Y
,
Kudo
K
, et al
DNA-PKcs is activated under nutrient starvation and activates Akt, MST1, FoxO3a, and NDR1
.
Biochem Biophys Res Commun
2020
;
521
:
668
73
.
49.
Yue
J
,
Huhn
S
,
Shen
Z
. 
Complex roles of filamin-A mediated cytoskeleton network in cancer progression
.
Cell Biosci
2013
;
3
:
7
.
50.
Velkova
A
,
Carvalho
MA
,
Johnson
JO
,
Tavtigian
SV
,
Monteiro
AN
. 
Identification of Filamin A as a BRCA1-interacting protein required for efficient DNA repair
.
Cell Cycle
2010
;
9
:
1421
33
.
51.
Zhang
S
,
Mo
Q
,
Wang
X
. 
Oncological role of HMGA2 (Review)
.
Int J Oncol
2019
;
55
:
775
88
.
52.
Li
AY
,
Boo
LM
,
Wang
SY
,
Lin
HH
,
Wang
CC
,
Yen
Y
, et al
Suppression of nonhomologous end joining repair by overexpression of HMGA2
.
Cancer Res
2009
;
69
:
5699
706
.