Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2009 Mar 1.
Published in final edited form as: Structure. 2008 Mar;16(3):468–477. doi: 10.1016/j.str.2007.12.014

Cryoelectron Microscopy Structure of the DNA-dependent Protein Kinase Catalytic Subunit (DNA-PKcs) at Subnanometer Resolution Reveals α-Helices and Insight into DNA Binding

Dewight R Williams 1, Kyung-Jong Lee 2, Jian Shi 1, David J Chen 2, Phoebe L Stewart 1,*
PMCID: PMC2323513  NIHMSID: NIHMS42929  PMID: 18334221

SUMMARY

The DNA-dependent protein kinase catalytic subunit (DNA-PKcs) regulates the non-homologous end joining pathway for repair of double-stranded DNA breaks. Here we present a 7Å resolution structure of DNA-PKcs determined by cryoEM single particle reconstruction. This structure is composed of density rods throughout the molecule indicative of α-helices and reveals new structural features not observed in lower resolution EM structures. Docking of homology models into the DNA-PKcs structure demonstrates that up to 8 helical HEAT repeat motifs fit well within the density. Surprisingly, models for the kinase domain can be docked into either the crown or base of the molecule at this resolution, although real space refinement suggests that the base location is the best fit. We propose a model for the interaction of DNA with DNA-PKcs in which one turn of dsDNA enters the central channel and interacts with a newly resolved α-helical protrusion.

INTRODUCTION

DNA double strand breaks (DSBs) are a common occurrence arising from both endogenous mechanisms, such as V-D-J recombination or attack by reactive oxygen species, as well as from exogenous factors, such as ionizing radiation or DNA replication inhibitors. DNA DSBs are considered particularly deleterious because they can be a precursor to chromosome translocations and deletions. When DNA DSBs are irreparable, cell death is the most common fate. DSBs can also lead to chromosome aberrations including gene fusions that are associated with initial steps of oncogenesis (Burma et al., 2006; Mitelman et al., 2007). The two main pathways for repair of DNA DSBs are homologous recombination (HR) and non-homologous end joining (NHEJ). NHEJ is the predominant repair pathway for DSBs in mammalian cells and does not require the presence of a homologous template. There are four distinct steps to NHEJ: DNA end recognition, bridging of the DNA ends, DNA end processing, and ligating the broken strands. In addition to DSB repair, NHEJ components are essential for V-D-J recombination which generates antibody and T cell receptor diversity required for lymphocyte maturation (Gellert, 2002). Molecular defects in the components of the NHEJ pathway result in severe combined immune deficiency (SCID) in mammals as well as sensitivity to ionizing radiation (radiosensitive or RS-SCID). These phenotypes have helped to genetically define the NHEJ components (Schwarz et al., 2003).

The DNA-dependent protein kinase (DNA-PK) plays a central role in regulating NHEJ and is a heterotrimer composed of Ku70, Ku80, and the large (469 kDa) DNA-dependent kinase catalytic subunit (DNA-PKcs). NHEJ is initiated when two Ku heterodimers recognize and stably bind broken DNA ends. Once bound to DNA, Ku recruits DNA-PKcs to the damage site and the kinase function of DNA-PKcs is activated. Recruitment of DNA-PKcs results in a translocation of Ku70/80 about one helical turn inward from the DNA end (Yoo and Dynan, 1999). The DNA-PK holoenzyme may promote formation of a bridging complex that keeps the DNA ends tethered (DeFazio et al., 2002; Weterings and van Gent, 2004; Weterings et al., 2003). DNA-PKcs is a serine/threonine kinase that undergoes phosphorylation at a cluster of six sites between Thr2609 and Thr2647 (Douglas et al., 2002), as well as additional sites including Ser2056 (Chen et al., 2005; Cui et al., 2005). The phosphorylation status of DNA-PKcs may serve to regulate DNA end processing (Cui et al., 2005; Ding et al., 2003). Impairing the kinase activity of DNA-PKcs or mutating the 7 major phosphorylation sites does not block localization of DNA-PKcs to DSB sites, but lowers the rate of exchange between DNA-bound and free DNA-PKcs. This observation led to the proposal that autophosphorylation is required to destabilize the initial protein–DNA complex that in turn facilitates additional repair steps (Uematsu et al., 2007).

DNA-PKcs has also been shown to phosphorylate several of the NHEJ components, including Ku70, Ku80, Artemis, and XRCC4 (Meek et al., 2004). Artemis is a 5’-3’ exonuclease that possesses hairpin opening and DNA overhang endonucleolytic activities when bound to and phosphorylated by DNA-PKcs (Ma et al., 2005). Truncation of the C-terminus of Artemis, which is the site of DNA-PKcs phosphorylation, relieves inhibition of the endonucleolytic and hairpin opening activities similar to DNA-PKcs activation (Niewolik et al., 2006). Because this activity is heightened in the presence of DNA-PKcs and ATP, it has been proposed that DNA-PKcs may configure 5’ and 3’ overhang into the appropriate substrates for Artemis' endonucleolytic activity (Niewolik et al., 2006). In a cell-free system, DNA-PKcs kinase activity is required for both DNA end processing and for ligation (Budman et al., 2007). XRCC4 and DNA ligase IV function as a complex, regulated by XLF/Cernunnos (Ahnesorg et al., 2006; Buck et al., 2006). Both DNA-PKcs and Ku are necessary for targeting of the ligase complex to the DSB site.

DNA-PKcs has been classified on the basis of sequence analysis as a member of phosphatidylinositol-3 (PI-3) kinase-like kinase (PIKK) superfamily, which contains the human proteins ataxia telangiectasia mutated (ATM), ATM-Rad3 related (ATR), FKBP-rapamycin associated protein (FRAP) or mammalian target of rapamycin (mTOR), SMG-1, and transformation/transcription domain associated protein (TRRAP). The PIKKs regulate divergent processes including genome surveillance, response to cellular stress, and nonsense mediated mRNA decay (Abraham, 2004; Bakkenist and Kastan, 2004). The N-terminal region of PIKK family members is not well conserved, and is predicted to be mostly α-helical and composed of multiple α-helical HEAT (huntingtin-elongation-A-subunit-TOR), TPR (tetratricopeptide repeats), or PFT (protein farnesyl transferase) repeats (Brewerton et al., 2004; Perry and Kleckner, 2003). The C-terminal region has a PI3K-like kinase domain flanked by a FAT (FRAP, ATM, TRRAP) domain and a FAT-C domain.

In the absence of an atomic resolution structure, DNA-PKcs has been the focus of several three-dimensional electron microscopy (EM) studies. Three independently derived EM structures have been determined, two by single particle reconstruction (Chiu et al., 1998; Rivera-Calzada et al., 2005) and one by electron crystallography (Leuther et al., 1999). The resolution of the two single particle reconstructions is similar, 20 or 21 Å, by the standard 0.5 Fourier shell correlation criterion. The electron crystallography structure included data out to resolution of 22 Å. These structures have three main features in common: a flat base region, a more globular crown or head region, and a large central channel between the base and crown. However the low resolution nature of these structures has made direct comparison difficult and molecular docking with these structures has generated conflicting theories for domain assignments. To address these differences, we have undertaken a subnanometer (<10 Å) resolution structure determination of DNA-PKcs applying recent advances in data acquisition and image processing.

Here we present a 7 Å resolution cryo-electron microscopy (cryoEM) structure of DNA-PKcs that reveals α-helices throughout the molecule and new features that were not resolved in previous structures. Homology models of up to 8 HEAT repeats can be docked reasonably well into the cryoEM density. The cryoEM density is consistent with short regions of HEAT repeats interspersed with non-helical regions or with other types of helical repeats as predicted by Brewerton et al. (Brewerton et al., 2004). Our results with docking of kinase domain homology models indicate that the fold of the DNA-PKcs kinase domain may differ significantly from the closest available homology models. Possible locations for the kinase domain are identified in both the crown and the base of the molecule. Real space refinement of these docked models shows that the best fit is in the base of DNA-PKcs. This site for the kinase domain is also consistent with a previous EM study of the DNA-PKcs/Ku70/Ku80/DNA the complex (Spagnolo et al., 2006) and biochemical data on the interaction between DNA-PKcs and Ku70/80. Our structure reveals an α-helical region that was not observed at lower resolution and that we term the central protrusion. This protrusion is formed by a stacked layer of density rods that projects into the large central channel of the molecule. Two groups have proposed that double-stranded DNA (dsDNA) binds in the central channel of DNA-PKcs (Boskovic et al., 2003; Leuther et al., 1999). A 12 base pair DNA fragment has been shown to bind DNA-PKcs, although it is not long enough to activate kinase activity (Leuther et al., 1999). Photocross-linking studies have indicated that a ∼10 base pair region proximal to the free DNA end makes contact with DNA-PKcs (Yoo and Dynan, 1999). We propose that the protrusion in the central channel, which is located ∼34Å from the edge of the DNA-PKcs molecule, is well positioned to be the main interaction site with dsDNA.

RESULTS

CryoEM Structure of DNA-PKcs at 7 Å Resolution

To determine a subnanometer resolution structure of DNA-PKcs we collected a dataset of 356,135 cryoEM particle images on an FEI Polara microscope (300kV, FEG). The refined structure shows the characteristic crown, base, and central channel observed at lower resolution, as well as new α-helical features. The overall dimensions of the structure are 120 Å in width and 150 Å in height as oriented in Figure 1A. Two well resolved connections are observed on either side of the central channel. Just off center in the channel is a large protrusion that projects down toward the base of the molecule. The resolution of the structure is 6.7 Å as assessed by the 0.5 Fourier shell correlation criterion (Fig. 1B). A collection of −5 μm defocused particle images chosen from the entire dataset for their high image contrast is shown with matching projections of the DNA-PKcs structure (Fig. 1C). These matched pairs show excellent agreement and this can be taken as evidence for the consistency of the final reconstruction with the aligned particle images. The presence of high resolution information in the dataset can be assessed by plotting the radially averaged power spectrum for particles with the same or similar defocus values. The power spectrum of a selected subset of 132 particle images (0.04% of the dataset) with a narrow range of defocus values confirms that the cryoEM particle images contain information beyond 10 Å resolution (Fig. 1D).

Figure 1. CryoEM structure of human DNA-PKcs at 7 Å resolution.

Figure 1

Open in a new tab

(A) Radial height color coded depiction of DNA-PKcs with the isosurface level set to enclose 469 kDa (100% of the expected volume). Scale bar, 50 Å. (B) Plot of the Fourier shell correlation calculated from two half datasets indicating 6.7 Å resolution at the 0.5 threshold. (C) Raw images (top image in each pair) with large (−5 μm) defocus values selected from the 356,135 particle dataset. The refined orientational parameters of these particle images were used to project the DNA-PKcs structure in matching orientations (bottom image in each pair). (D) Plot of the radially averaged power spectrum (dashed line) for a selected subset of 132 particle images compared to the empirically fit CTF curve generated with CTFIT (solid line). The particle images were selected to have a narrow range of defocus values (2.35 to 2.37 μm) for both the defocus1 and defocus2 parameters after FREALIGN refinement. The information in this subset of particles extends beyond 10 Å resolution.

The DNA-PKcs cryoEM structure of Rivera-Calzada et al. filtered to 30 Å resolution shows the same basic structural features (crown, base, channel) as the 7 Å resolution cryoEM structure, however the higher resolution features are not consistent (Rivera-Calzada et al., 2005). This may indicate either an overestimation by Rivera-Calzada et al. of the resolution or possibly a conformational difference captured by the two structures. The resolution reported for the earlier structure is ∼13Å based on the FSC 0.14 and 3σ criteria, and 19.78 Å at the FSC 0.5 criterion. When our structure is displayed at the 100% isosurface level the base of the molecule appears to be a solid region of density. At the 30% isosurface level the base appears to have two struts of density (Supplementary Fig. 1), which are similar to the description of the proximal claws in the Rivera-Calzada structure.

Our DNA-PKcs structure displays a distinct handedness. The two side connections are clearly different and the central protrusion is off center. Since cryoEM single particle reconstruction from untilted projection images can not determine the absolute hand of a structure, we also collected a small dataset of tilt pair projections. Applying a computational method for absolute hand determination (Rosenthal and Henderson, 2003) gave a clear indication of the correct hand of the DNA-PKcs structure (Supplementary Fig. 2). Leuther et al. also determined the absolute hand of DNA-PKcs by electron crystallography with negatively stained 2D crystals (Leuther et al., 1999). The electron crystallographic structure is flattened and is at a significantly lower resolution (∼22 Å) than the cryoEM structure. Nevertheless the hand appears consistent with our results.

Structural Regions of DNA-PKcs

The top of the DNA-PKcs molecule has been referred to as the crown or head due to its distinctive shape. The crown forms a bi-lobed cap that extends down the back of the molecule (Fig. 2A,B). The brow of the DNA-PKcs structure is a ∼30 Å thick cylinder that wraps half way around the molecule. The crown and brow enclose a cavity ∼30 Å across. This internal cavity was previously observed in the lower resolution DNA-PKcs structures, although it was best resolved in the electron crystallography structure (Leuther et al., 1999). Our structure supports the earlier interpretation by Leuther et al. that this chamber is accessible by ∼10 Å openings. A new feature resolved in our structure is the region we refer to as the central protrusion. The central protrusion comes within 10 Å of the base as displayed in Fig. 2 and connects to the base at >100% isosurface levels, indicating that conformational changes could be relayed from the protrusion to the base. The crown and brow are connected to the base by two distinct side connections. The side-1 connection is broader and has weaker density than the rest of the structure suggesting conformational flexibility (Fig. 2C). In contrast the side-2 connection appears to have a distinct stacking pattern with a repeat distance of ∼12−13 Å, which is similar to that between α-helical repeats in HEAT and PFT motifs (Fig. 2D).

Figure 2. Structural regions of DNA-PKcs.

Figure 2

Open in a new tab

(A) Full and cropped views of the structure are shown with the regions colored as follows: crown (red), brow (purple), central protrusion (green), side connections (yellow), and base (blue). This color coding is not meant to imply assignment of structural regions to particular amino acids or protein domains. The upper part of the molecule has a cavity that is almost completely enclosed. The lower part of the molecule has the large central channel, which is ∼45 Å wide at the narrowest point. Scale bar, 50 Å. (B) Four views of the DNA-PKcs structure related by 90° rotations. Scale bar, 50 Å. (C) Enlarged view of the side-1 connection. (D) Enlarged view of the side-2 connection. The bracket indicates 12 Å, which is the spacing observed between density rods in this region.

Resolution of α-Helices within the DNA-PKcs Structure

Secondary structure prediction algorithms assign between 110 and 160 α-helices of 8 or more residues to the full length sequence of human DNA-PKcs (4128aa). We only counted predicted α-helices of 8 or more residues, since secondary prediction algorithms while quite good at predicting the locations of helices are often incorrect by 1−2 residues in length. We know from experience that α-helices of 10 or more residues are resolved in cryoEM density maps at 6−7Å resolution (Saban et al., 2006). When we raise the isocontour level for our DNA-PKcs structure, density rods of the appropriate dimensions for α-helices are found throughout the molecule. These density rods can be accentuated by applying a small negative temperature factor (B = −150 Å2) to the density map, which serves to restore contrast at high resolution. A slab of density through the temperature-factor corrected map reveals numerous density rods in the crown, brow, side-2 connection, and base of the molecule (Fig. 3A). The central protrusion also shows particularly well resolved density cylinders with the correct spacing (∼12 Å) for stacked α-helices (Fig. 3B). The CoLoRes quantitative docking tool in the SITUS package (Chacon and Wriggers, 2002) was used to position α-helices with 4, 5 or 6 turns within the cryoEM density. The resulting top 300 α-helical fits were visually inspected to ensure that the model α-helix fit well within a continuous rod of cryoEM density and were positioned uniquely in the density. Approximately 70 unique positions were found throughout DNA-PKcs (Supplementary Fig. 3). Enlarged views of three α-helical fits within density rods are shown in Figure 3C.

Figure 3. Alpha-helices resolved in the DNA-PKcs cryoEM structure.

Figure 3

Open in a new tab

(A) A slab of density ∼35 Å thick through the DNA-PKcs cryoEM structure with an applied temperature factor (B = −150 Å2) showing numerous density rods. Scale bar, 10 Å. (B) The central protrusion, shown extracted from the DNA-PKcs structure, has a layer of stacked density rods. In parts A and B the isosurface level is set to most clearly reveal the density rods and encloses ∼75% of the expected volume. Scale bar, 10 Å. (C) Three examples of α-helical fits within the cryoEM density. The numbers (1−3) in part A indicate their locations within the structure. The isosurface levels range from 90−65%.

At ∼10Å resolution identification of α-helices as density rods begins to be possible in cryoEM density maps (Jiang and Ludtke, 2005; Saban et al., 2006), although 6−7Å resolution provides better α-helical resolution (Saban et al., 2005; Saban et al., 2006). Visualization of α-helices allows more accurate fitting of atomic models into cryoEM density than is possible at lower resolutions. Confirmation that α-helical density rods align in three dimensions with α-helices in an atomic model is a significantly more discriminating test than aligning atomic models by overall shape.

Molecular Docking Analysis of HEAT Repeat Motifs

The HHpred homology detection server (Soding et al., 2005) was used to retrieve the best homology models for three overlapping portions of DNA-PKcs (aa 1−1,800; aa 1165−2964; and aa 2329−4128). The models with the highest probabilities were for the N-terminal and C-terminal regions of DNA-PKcs. Cand1 (aa 133−1189, PDB 1U6G) is predicted as a model for a large segment of the N-terminal region (aa 64−1072) with a probability of 99.5%, E-value of 2.2E-05, and score of 58.2. The percent sequence identity is 14% with 883 residues aligned. This region of Cand1 contains 24 α-helical HEAT repeats, which have a U-shaped fold in the crystal structure of the Cand1-Cul1-Roc1 complex (Goldenberg et al., 2004). CoLoRes was used to dock the coordinates for the predicted Cand1 segment into the DNA-PKcs cryoEM density. The best fit has a modest Laplacian correlation score (5.0E-02) and is oriented with the N-terminal region of the Cand1 model in the base of DNA-PKcs, HEAT repeats extending through the side-2 connection, and the C-terminal region in the central protrusion (Supplementary Fig. 4). While approximately half of the Cand1 homology model in the top scoring position fits the cryoEM density, the other half strays outside of the density. This leads us to conclude that if DNA-PKcs contains the same number of continuous HEAT repeats they do not follow the exact superhelical arc found in the Cand1 structure.

Therefore we truncated the Cand1 homology model into segments of 2, 4 or 8 HEAT repeats and used CoLoRes to dock these shorter elements into the DNA-PKcs structure. Since the superhelical twist of Cand1 varies from a right-handed structure in the N-terminal region to a left-handed structure in the C-terminal region, we selected short elements from the beginning, middle and end of the Cand1 homology region. All of the short HEAT models, with the exception of the left-handed 8-repeat model of Cand1, gave higher Laplacian correlation scores than the 24-repeat model. The best scoring model (Laplacian correlation 7.4E-02) was found for the 4-repeat model from the N-terminal region of Cand1 (aa 133−282) in the base of DNA-PKcs. In this position the 4-repeat model fits well within this region of cryoEM density, which has a repeat pattern with the correct spacing for HEAT motifs (Fig. 4A,B). Close inspection of this density region reveals resolved density rods for 4 of the 8 α-helices within the homology model. The additional 4 α-helices in the model do fit within cryoEM density, but this density doesn't clearly resolve into separated rods. The top scoring 4-repeat model positioned by CoLoRes was refined and evaluated with the real space refinement program RSRef2000 (Korostelev et al., 2002). Real space refinement resulted in a correlation coefficient for this model with the cryoEM density of 0.74. This correlation is as high as expected for an intermediate resolution cryoEM density map (Gao and Frank, 2005).

Figure 4. Fit of HEAT repeat models in the cryoEM density.

Figure 4

Open in a new tab

(A) The top scoring CoLoRes position for a 4-repeat model within the base of DNA-PKcs shown with a 75% isosurface level. (B) Perpendicular and stereo view of the same fit with a 10% isosurface level. (C) The top scoring CoLoRes position for an 8-repeat model within the base and side-2 connection shown with a 75% isosurface level. (D) Enlarged stereo view of this fit. The DNA-PKcs cryoEM structure is shown with an applied temperature factor.

The best scoring 8-repeat model (Laplacian correlation 7.0E-02) was found by CoLoRes for Cand1 (aa 133−508) in the base and side-2 connection of DNA-PKcs (Fig. 4C,D). This model correctly follows the curve of the base into the side-2 connection, as well as the density spacing in the side-2 connection, but not every model α-helix is within a cryoEM density rod. Real space refinement for this model resulted in a moderate correlation coefficient of 0.51 with the cryoEM density. Based on our docking results with various length HEAT repeat models, we conclude that the CoLoRes fitting result indicating 70 unique α-helix positions is most likely an underestimate of the true number of α-helices in the DNA-PKcs structure.

Molecular Docking Analysis of Kinase Domain Models

The highest scoring HHPred homology model for the kinase domain of DNA-PKcs is porcine PI3Kγ (probability of 100%; E-value of 2E-44; score of 472.5), with DNA-PKcs residues 3654−4113 aligning with 592−936 of PI3Kγ (PDB 1E7U) (Walker et al., 2000). HHPred indicates a sequence identity of 26% for 328 aligned residues. This region of the porcine PI3Kγ includes most of the helical domain, which has a fold similar to HEAT repeats, the N-terminal lobe of the catalytic domain, and about one quarter of the catalytic domain C-terminal lobe (Fig. 5A). There are three large insertions of 25 to 52 residues in the DNA-PKcs sequence relative to PI3Kγ as aligned by HHPred. These insertions are in the N- and C-lobes of the catalytic domain of PI3Kγ. Therefore there may be significant differences in the folds of the PI3Kγ and the DNA-PKcs catalytic domains.

Figure 5. Fit of kinase domain model in the cryoEM density.

Figure 5

Open in a new tab

(A) The HHPred predicted model for the DNA-PKcs kinase domain is a region of porcine PI3Kγ (aa592−936, PDB 1E7U) that includes part of the helical domain (green), the N-terminal lobe (red), and part of the C-terminal lobe of the catalytic domain (yellow). HHPred alignment indicates three large insertions in the DNA-PKcs sequence relative to PI3Kγ as indicated (blue). (B) A cropped view of the DNA-PKcs structure (100% isosurface level) shown with the top scoring crown and base positions. (C) Stereo view of the crown and base positions shown within a transparent mesh (75% isosurface level). (D) Perpendicular view from the bottom of the molecule. The DNA-PKcs cryoEM structure is shown with an applied temperature factor.

CoLoRes docking of the porcine PI3Kγ homology model into the DNA-PKcs structure gave only modest Laplacian correlation scores in the range of 3.7−4.5E-02. Moreover the top ten results were divided between locations in the crown and base of the molecule. Visual inspection of the highest scoring crown and base positions (Laplacian scores 4.5E-02 and 3.9E-02, respectively) indicated that the model in the base location fit within the overall density envelope better (Fig. 5B,C,D). The model in the crown location protrudes somewhat into the cavity in the upper half of the molecule. In either position only about 2 of the 10 α-helices of 10 more residues in the PI3Kγ homology model align with rods of EM density, indicating structural divergence between the PI3Kγ and DNA-PKcs catalytic domains. We also built models for the DNA-PKcs kinase domain with MODELLER (Sali and Blundell, 1993), and PHYRE, the successor of 3D-PSSM (Kelley et al., 2000), but after CoLoRes docking neither model showed any better alignment of α-helices with density rods. Real space refinement with RSRef2000 of the highest scoring crown and base positions for the unmodified PI3Kγ (aa 592−936) model yielded a significantly higher correlation coefficient for the position in the base (0.53), than for the crown (0.11). The real space refinement results together with our visualize assessment indicate that the base position is more likely, although docking with currently available homology models is not definitive.

Potential DNA Binding Sites within DNA-PKcs

Our structure reveals two distinct side connections between the crown/brow region and the base of the molecule that were not clearly observed in any previous structure. The two side connections create an enclosed passageway through the central channel that is the proposed site of DNA binding (Boskovic et al., 2003; Leuther et al., 1999). The new structure reveals a protrusion composed of a stacked layer of α-helical like density rods that is positioned ∼34 Å from the entrance of the central channel along the distance of closest approach. DNA-PKcs is activated upon binding to dsDNA ends with the strongest activation occurring when the DNA ends are unpaired (Jovanovic and Dynan, 2006). Additionally a minimum of 12bp has been shown to be required to activate the kinase in the absence of Ku 70/80 and when a DNA-PKcs/Ku70/80/DNA complex is formed at least 10bp are required to activate DNA-PKcs (West et al., 1998; Yoo and Dynan, 1999). One of the most common DNA binding motifs is the helix-loop-helix motif of transcription factors, which bind within the major groove of the DNA. The protrusion in the DNA-PKcs structure is well placed to bind the free end of dsDNA upon insertion into the central channel (Fig. 6A). The modeled configuration shows that approximately 1 turn of dsDNA would need to enter the central channel in order to have the major groove of DNA interact with α-helices in the central protrusion (Fig. 6B). Sitting above the central channel is a smaller cavity that is separated from the central channel by protein density. This smaller cavity is only large enough to accommodate single-stranded DNA. We postulate that this smaller cavity, which is directly above the central protrusion, may be involved in binding ssDNA during end processing reactions. Several studies have shown that the DNA-PKcs molecule is capable of binding both ssDNA and dsDNA in non-competing sites (Gottlieb and Jackson, 1993; Leuther et al., 1999). However, only dsDNA is capable of activating the kinase. We have not pursued DNA bound structures in this high resolution study, but previous studies at low resolution have shown DNA binds near or in the central channel (Boskovic et al., 2003), supporting the configuration modeled in Figure 6.

Figure 6. Modeling of potential DNA binding sites within DNA-PKcs.

Figure 6

Open in a new tab

(A) The central channel in DNA-PKcs is large enough to accommodate dsDNA. Scale bar, 50 Å. (B) A cropped view shows that the smaller upper cavity is only large enough for ssDNA. Approximately one helical turn of dsDNA spanning 34 Å would be required to reach the highly α-helical central protrusion if DNA enters from the side of the molecule shown in front in part A. Small pores in the roof of the central channel are present that might allow ssDNA to pass into the upper cavity. Such a configuration could facilitate end processing or end discrimination for DSB repair pathway selection. The view in B is rotated 90° from that in part A and the DNA-PKcs density is shown cropped in half.

DISCUSSION

CryoEM combined with image reconstruction has resulted in the determination of several icosahedral virus structures at subnanometer (<10 Å) resolution (Zhou and Chiu, 2003) and a handful of non-icosahedral assembly structures including the ribosome at subnanometer resolution (Jiang and Ludtke, 2005). In this study, we present one of the highest resolution cryoEM single particle reconstructions of an asymmetric particle. At better than 10 Å resolution, secondary structural elements, particularly α-helices, can be directly visualized. Visualization of α-helices in turn facilitates accurate docking of atomic resolution structures of component molecules or domains. We observe α-helical density rods throughout the 7Å resolution DNA-PKcs structure.

Docking with various length HEAT, TPR, and PFT repeat motifs indicates that for the most part the shorter segments fit best within the DNA-PKcs cryoEM density. Given the large variation in superhelical twist of the HEAT, TPR, and PFT repeat proteins in the Protein Data Bank, it is not surprising that the full extent of these proteins does not exactly match the α-helical density rods of DNA-PKcs. We have found that 4 and 8 HEAT repeat models from the N-terminal region of Cand1 follow density rods in the base and side-2 connection of DNA-PKcs reasonably well. Real space refinement of these models gave good correlation coefficients with the cryoEM density. Our docking results indicate that short regions of α-helical repeats are likely to be present in other regions of the molecule including the central protrusion, crown, and base. Fold prediction studies on the PIKK superfamily of proteins reveals a potential for HEAT like helical domains, but also suggested that the likelihood of an extended superhelical region was limited and that the HEAT like motifs would probably be dispersed by regions of non-HEAT α-helical folds (Brewerton et al., 2004; Perry and Kleckner, 2003). This may best describe our results with CoLoRes fitting of various HEAT models into the DNA-PKcs structure with the best fits found for dispersed regions of 4−8 HEAT repeats.

The lack of an atomic structure for any of the PIKK family members has necessitated molecular docking with the kinase domain from the closest homologous structure, the PI3Kγ. Other groups have also attempted to dock homology models for the kinase domain into lower resolution EM structures of DNA-PKcs. One group reported the best fit within the crown (Rivera-Calzada et al., 2005) and the other reported the best fit within the base (Brewerton et al., 2004) of DNA-PKcs. We find that PI3Kγ kinase domain homology models can fit reasonably well within either the crown or the base based purely on volumetric assessment with CoLoRes. Real space refinement of the top scoring crown and base locations suggested that the base location is the best fit with the cryoEM density. However we also find that the precise three-dimensional distribution of α-helices within the homology models does not match the observed distribution of density rods in either the crown or base of DNA-PKcs. The lack of a good fit for the helices is not surprising given that there are several significant gaps of 25−52 amino acids in the HHPred alignment of the PI3γ and DNA-PKcs kinase domains. Presumably an atomic resolution structure of the DNA-PKcs kinase domain, or a closer homology model, would more readily be docked into a unique position within the cryoEM density.

A previous EM study of DNA-PKcs extended their kinase molecular fitting experiments with antibody localization using a commercially available antibody (sc-5282) directed against a large C-terminal region (aa 2965−4128) (Rivera-Calzada et al., 2005). This study showed sc-5282 antibody density between the crown and the base of the molecule. We also have performed antibody localization with a monoclonal antibody, DNA-PKcs Ab-2 (Clone 25−4), with a similarly broad epitope (3198−4127). Our reconstruction of the DNA-PKcs/Ab-2(clone 25−4) complex reveals antibody density in the central region of the molecule near the side-2 connection (Supplementary Fig. 5). The problem with this approach is that the precise epitopes of these antibodies have not been determined and they could be located 450 to 680aa away from the beginning of the HHPred predicted kinase domain. A stretch of HEAT repeats formed by this many residues can span a distance of 100−160Å. Given that the maximum dimension of DNA-PKcs is ∼150 Å, localization of the kinase domain with this approach lacks the precision needed to discriminate between the possible crown and base locations.

The negative stain EM structure of the DNA-PKcs/Ku70/Ku80 complex (Spagnolo et al., 2006), which resembles our cryoEM structure filtered to 25Å resolution shows extra density assigned to Ku70/80 at the base of the DNA-PKcs. Weaver and coworkers (Jin et al., 1997) found that two DNA-PKcs protein fragments (aa 3002−3450 and 3414−3850) bind Ku independently. The more C-terminal fragment includes the region of DNA-PKcs that is homologous to the FKBP12-rapamycin binding (FRB) domain of FKBP12-rapamycin associated protein (FRAP) (aa 3582−3666). Although the Ku association regions are N-terminal to the DNA-PKcs residues expected to be involved in the catalysis of phosphotransfer (aa 3917−3922 and 3936−3938) (Hartley et al., 1995; Milne et al., 1996), the C-terminal Ku-binding fragment overlaps with the PI3Kγ homology region found by HHpred (aa 3654−4113) by ∼200 residues. Given this overlap and assuming Ku binds to the base of DNA-PKcs as indicated by the previous negative stain EM study, we infer that the PI3Kγ homology region is likely to reside in the base of DNA-PKcs. We note that the base of DNA-PKcs represents ∼33% of the volume of the molecule and is easily large enough to contain the kinase domain, as well as the flanking FAT and FAT-C domains.

An outstanding question for DNA-PKcs is how its kinase activity is DNA dependent. This structural study advances the notion that the kinase domain is likely within the base of the molecule and dsDNA binds to the protrusion within the central channel. Interaction of dsDNA with the protrusion could cause a conformational change that directly affects the base of the molecule leading to activation of the kinase domain. Such a mechanism of activation does not require gross molecular rearrangement of the molecule and places the kinase in a region where DNA-dependent regulation is easy to envision. High resolution structures of individual DNA-PKcs domains, as well as complexes with DNA and other components of the NHEJ pathway will be required for a more detailed understanding of this important component in genome surveillance.

EXPERIMENTAL PROCEDURES

Protein Purification

HeLa nuclear extracts were prepared as described previously (Chan et al., 2002). Nuclear extracts were sequentially bound to phosphocellulose (PC11) resin, then a diethylaminoethyl-sepharose (DEAE) column, and then heparin-agarose; equilibrated in column buffer (CB) [50mM Tris-HCl pH 7.9, 1 mM ethylenediamine tetraacetic acid (EDTA), 2% glycerol, 1mM dithiotreitol (DTT), and 0.1M KCl, with freshly added protease inhibitors]. Proteins were eluted with a 100mM step gradient from 0.1M to 1 M KCl. Peak DNA-PK containing fractions were pooled and equilibrated to CB before progressing to the next round chromatography. DNA-PKcs containing fractions were further purified by two final rounds of chromatography: a single-stranded DNA cellulose matrix followed by a MonoQ (Amersham Biosciences) anion exchange, each eluted with a linear salt gradient from 0.1 to 1M KCl. Purified DNA-PKcs was flash frozen and stored at −80°C.

Cryoelectron Microscopy

DNA-PKcs samples with at 20−50 μg/ml in 20 mM HEPES 8.0, 100 mM KCl, 0.5% glycerol were applied to freshly prepared EM grids with homemade holey carbon film. The excess liquid from a 2−3 μl droplet was blotted away with filter paper, and the sample grid was immediately plunged into liquid ethane cooled by liquid nitrogen. All data incorporated into the final structure of DNA-PKcs was collected on an FEI Tecnai Polara equipped with a field emission gun (FEG) operating at 300 kV in nanoprobe mode. The sample grids were maintained at liquid nitrogen temperature during data acquisition. Images were digitally recorded on a Gatan UltraScan 4000 (4k × 4k) CCD camera. Micrographs were collected with a defocus range of −0.9 to −7 μm and with a magnification of either 253,654× or 397,878×. The absolute magnification values were refined previously to within +/−0.25% (Saban et al., 2006). An initial small dataset of 14,989 particle images was collected at 253,654× with a defocus range of −3 to −6 μm and an average electron dose of 50 e−/Å2 to maximize image contrast and generate an initial reconstruction. The remainder of the data was collected with an average electron dose of 20 e−/Å2.

A total of 17,787 digital micrographs were saved with the aid of a Tecnai and Gatan scripting interface package that facilitated semi-automated targeting, acquisition, and image quality assessment (Shi, Williams, Stewart, In preparation). Micrographs were evaluated by image binning and Fourier transformation before saving to disk and typically 30−50% were rejected for noticeable drift, astigmatism or charging effects.

Image Processing

The micrographs were binned by various factors, ranging from 3 to 8, during the course of image processing to produce images with pixels size of 1.5 to 4.7 Å on the molecular scale. The coarsest images were used in the initial processing steps. Particle images were selected manually for the initial small dataset using the BOXER routine in EMAN (Ludtke et al., 1999). The remainder of the particle images were automatically selected using the Batchboxer routine in EMAN with characteristic projections of the initial DNA-PKcs reconstruction used as references for autoboxing. A total of 356,135 particle images were selected for image processing.

An initial reconstruction was calculated for the small data dataset of 14,989 particle images. This structure was at ∼15 Å resolution (FSC 0.5 threshold) and contained features consistent with the first cryoEM structure of DNA-PKcs (Chiu et al., 1998). This initial structure served as the input 3D map for the first round of refinement in FREALIGN with the full dataset. The final reconstruction is generated from a subset of 349,719 particles with the best FREALIGN phase residual values. The resolution of the final reconstruction indicated by the FREALIGN FSC 0.5 threshold is 6.7 Å. The RMEASURE resolution assessment program (Sousa and Grigorieff, 2007) indicated a resolution of 7.7 Å at the FSC 0.5 threshold and 5.6 Å at the FSC 0.143 threshold. For the purpose of accentuating α-helical density rods a map was also generated with a small negative temperature factor (B = −150 Å2) using BFACTOR (Niko Grigorieff, http://emlab.rose2.brandeis.edu/grigorieff/). See Supplementary Methods for more details.

Hand Determination for the DNA-PKcs Structure

A small set of cryoelectron micrographs were collected at a magnification of 253,654× as tilt pairs, with one untilted (0°) micrograph and one tilted (+15° compustage rotation) micrograph of the same field of DNA-PKcs particles. These micrographs yielded 191 selected particle images with two projection views (0° and +15°) for each particle. The physical axis systems of the compustage and the microscope room were related to that of two micrograph display packages (Digital Micrograph and EMAN). A simulated tilt pair dataset was created for a test molecule, Cand1, which has a distinctive overall handedness (PDB 1U6G)(Goldenberg et al., 2004). Particle images from the simulated micrographs were put a computational procedure to determine hand (using FREALIGN and ANGPLOT) developed by Rosenthal and Henderson (Rosenthal and Henderson, 2003). A clear minimum in the average phase residual was observed for RotX=0°, RotY=15°. The actual cryoEM tilt pair dataset was put through the same computational procedure and a clear minimum was found for RotX=0°, RotY=−15° (Supplementary Fig. 2). This indicated that hand of the current reconstruction needed to be flipped. The correct hand of the final DNA-PKcs reconstruction is shown in all figures and was used for the molecular fitting experiments. See Supplementary Methods for more details.

Molecular Fitting

HHpred (Soding et al., 2005) was used to find homology models for regions of DNA-PKcs. Homology models were also built with MODELLER (Sali and Blundell, 1993) and PHYRE (Kelley et al., 2000). The CoLoRes routine in SITUS (Chacon and Wriggers, 2002) was used to perform exhaustive 6 dimensional searches of the homology models within the final reconstruction of DNA-PKcs filtered to 6 Å resolution. The results were evaluated visually with UCSF Chimera (Pettersen et al., 2004). The program RSRef2000 (Korostelev et al., 2002) was used to refine and quantitatively assess the molecular docking results. The parameters used for RSRef2000 were resolution range (12−7 Å); atomsize and refsize (12 Å and 6 Å); weight for electron density component (30,000,000); refinement scheme (slowcool); molecular dynamics type (Cartesian); starting temperature (5000); and cool rate (500). Refinement of the 4 and 8 HEAT repeat models resulted in correlation coefficients of 0.74 and 0.51, real space R factors of 0.16 and 0.18, and density residuals of 0.94E+08 and 0.12E+09, respectively. The PI3Kγ kinase domain homology model in the top scoring CoLoRes crown and base locations refined to give final correlation coefficients of 0.11 and 0.53, real space R factors of 0.19 and 0.17, and density residuals of 0.47E+09 and 0.32E+09, respectively.

In addition to the HEAT repeat and kinase domains models described in the Results section, homology models for other regions of DNA-PKcs were identified, docked with CoLoRes, and evaluated for their fit within the cryoEM density. See Supplementary Methods for more details.

ACKNOWLEDGEMENTS

We would like to thank Dr. Richard Henderson for the ANGPLOT program and the Vanderbilt Advanced Computing Center for Research & Education (ACCRE) for their assistance with computational issues. Support from the National Institutes of Health (F32 GM71276 to DRW and CA50519 and CA92584 to DJC) is gratefully acknowledged.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Supplementary Material

01

REFERENCES

  1. Abraham RT. PI 3-kinase related kinases: ‘big’ players in stress-induced signaling pathways. DNA repair. 2004;3:883–887. doi: 10.1016/j.dnarep.2004.04.002. [DOI] [PubMed] [Google Scholar]
  2. Ahnesorg P, Smith P, Jackson SP. XLF interacts with the XRCC4-DNA ligase IV complex to promote DNA nonhomologous end-joining. Cell. 2006;124:301–313. doi: 10.1016/j.cell.2005.12.031. [DOI] [PubMed] [Google Scholar]
  3. Bakkenist CJ, Kastan MB. Initiating cellular stress responses. Cell. 2004;118:9–17. doi: 10.1016/j.cell.2004.06.023. [DOI] [PubMed] [Google Scholar]
  4. Boskovic J, Rivera-Calzada A, Maman JD, Chacon P, Willison KR, Pearl LH, Llorca O. Visualization of DNA-induced conformational changes in the DNA repair kinase DNA-PKcs. EMBO J. 2003;22:5875–5882. doi: 10.1093/emboj/cdg555. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brewerton SC, Dore AS, Drake AC, Leuther KK, Blundell TL. Structural analysis of DNA-PKcs: modelling of the repeat units and insights into the detailed molecular architecture. J. Struct. Biol. 2004;145:295–306. doi: 10.1016/j.jsb.2003.11.024. [DOI] [PubMed] [Google Scholar]
  6. Buck D, Malivert L, de Chasseval R, Barraud A, Fondaneche MC, Sanal O, Plebani A, Stephan JL, Hufnagel M, le Deist F, et al. Cernunnos, a novel nonhomologous end-joining factor, is mutated in human immunodeficiency with microcephaly. Cell. 2006;124:287–299. doi: 10.1016/j.cell.2005.12.030. [DOI] [PubMed] [Google Scholar]
  7. Budman J, Kim SA, Chu G. Processing of DNA for nonhomologous end-joining is controlled by kinase activity and XRCC4/ligase IV. J. Biol. Chem. 2007;282:11950–11959. doi: 10.1074/jbc.M610058200. [DOI] [PubMed] [Google Scholar]
  8. Burma S, Chen BP, Chen DJ. Role of non-homologous end joining (NHEJ) in maintaining genomic integrity. DNA repair. 2006;5:1042–1048. doi: 10.1016/j.dnarep.2006.05.026. [DOI] [PubMed] [Google Scholar]
  9. Chacon P, Wriggers W. Multi-resolution contour-based fitting of macromolecular structures. J. Mol. Biol. 2002;317:375–384. doi: 10.1006/jmbi.2002.5438. [DOI] [PubMed] [Google Scholar]
  10. Chan DW, Chen BP, Prithivirajsingh S, Kurimasa A, Story MD, Qin J, Chen DJ. Autophosphorylation of the DNA-dependent protein kinase catalytic subunit is required for rejoining of DNA double-strand breaks. Genes Dev. 2002;16:2333–2338. doi: 10.1101/gad.1015202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Chen BP, Chan DW, Kobayashi J, Burma S, Asaithamby A, Morotomi-Yano K, Botvinick E, Qin J, Chen DJ. Cell cycle dependence of DNA-dependent protein kinase phosphorylation in response to DNA double strand breaks. J. Biol. Chem. 2005;280:14709–14715. doi: 10.1074/jbc.M408827200. [DOI] [PubMed] [Google Scholar]
  12. Chiu CY, Cary RB, Chen DJ, Peterson SR, Stewart PL. Cryo-EM imaging of the catalytic subunit of the DNA-dependent protein kinase. J. Mol. Biol. 1998;284:1075–1081. doi: 10.1006/jmbi.1998.2212. [DOI] [PubMed] [Google Scholar]
  13. Cui X, Yu Y, Gupta S, Cho YM, Lees-Miller SP, Meek K. Autophosphorylation of DNA-dependent protein kinase regulates DNA end processing and may also alter double-strand break repair pathway choice. Mol. Cell. Biol. 2005;25:10842–10852. doi: 10.1128/MCB.25.24.10842-10852.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. DeFazio LG, Stansel RM, Griffith JD, Chu G. Synapsis of DNA ends by DNA-dependent protein kinase. EMBO J. 2002;21:3192–3200. doi: 10.1093/emboj/cdf299. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ding Q, Reddy YV, Wang W, Woods T, Douglas P, Ramsden DA, Lees-Miller SP, Meek K. Autophosphorylation of the catalytic subunit of the DNA-dependent protein kinase is required for efficient end processing during DNA double-strand break repair. Mol. Cell. Biol. 2003;23:5836–5848. doi: 10.1128/MCB.23.16.5836-5848.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Douglas P, Sapkota GP, Morrice N, Yu Y, Goodarzi AA, Merkle D, Meek K, Alessi DR, Lees-Miller SP. Identification of in vitro and in vivo phosphorylation sites in the catalytic subunit of the DNA-dependent protein kinase. Biochem. J. 2002;368:243–251. doi: 10.1042/BJ20020973. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Gao H, Frank J. Molding atomic structures into intermediate-resolution cryo-EM density maps of ribosomal complexes using real-space refinement. Structure. 2005;13:401–406. doi: 10.1016/j.str.2005.01.004. [DOI] [PubMed] [Google Scholar]
  18. Gellert M. V(D)J recombination: RAG proteins, repair factors, and regulation. Annu. Rev. Biochem. 2002;71:101–132. doi: 10.1146/annurev.biochem.71.090501.150203. [DOI] [PubMed] [Google Scholar]
  19. Goldenberg SJ, Cascio TC, Shumway SD, Garbutt KC, Liu J, Xiong Y, Zheng N. Structure of the Cand1-Cul1-Roc1 complex reveals regulatory mechanisms for the assembly of the multisubunit cullin-dependent ubiquitin ligases. Cell. 2004;119:517–528. doi: 10.1016/j.cell.2004.10.019. [DOI] [PubMed] [Google Scholar]
  20. Gottlieb TM, Jackson SP. The DNA-dependent protein kinase: requirement for DNA ends and association with Ku antigen. Cell. 1993;72:131–142. doi: 10.1016/0092-8674(93)90057-w. [DOI] [PubMed] [Google Scholar]
  21. Hartley KO, Gell D, Smith GC, Zhang H, Divecha N, Connelly MA, Admon A, Lees-Miller SP, Anderson CW, Jackson SP. DNA-dependent protein kinase catalytic subunit: a relative of phosphatidylinositol 3-kinase and the ataxia telangiectasia gene product. Cell. 1995;82:849–856. doi: 10.1016/0092-8674(95)90482-4. [DOI] [PubMed] [Google Scholar]
  22. Jiang W, Ludtke SJ. Electron cryomicroscopy of single particles at subnanometer resolution. Curr. Opin. Struct. Biol. 2005;15:571–577. doi: 10.1016/j.sbi.2005.08.004. [DOI] [PubMed] [Google Scholar]
  23. Jin S, Kharbanda S, Mayer B, Kufe D, Weaver DT. Binding of Ku and c-Abl at the kinase homology region of DNA-dependent protein kinase catalytic subunit. J. Biol. Chem. 1997;272:24763–24766. doi: 10.1074/jbc.272.40.24763. [DOI] [PubMed] [Google Scholar]
  24. Jovanovic M, Dynan WS. Terminal DNA structure and ATP influence binding parameters of the DNA-dependent protein kinase at an early step prior to DNA synapsis. Nucleic Acids Res. 2006;34:1112–1120. doi: 10.1093/nar/gkj504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kelley LA, MacCallum RM, Sternberg MJ. Enhanced genome annotation using structural profiles in the program 3D-PSSM. J. Mol. Biol. 2000;299:499–520. doi: 10.1006/jmbi.2000.3741. [DOI] [PubMed] [Google Scholar]
  26. Korostelev A, Bertram R, Chapman MS. Simulated-annealing real-space refinement as a tool in model building. Acta Crystallogr. D Biol. Crystallogr. 2002;58:761–767. doi: 10.1107/s0907444902003402. [DOI] [PubMed] [Google Scholar]
  27. Leuther KK, Hammarsten O, Kornberg RD, Chu G. Structure of DNA-dependent protein kinase: implications for its regulation by DNA. EMBO J. 1999;18:1114–1123. doi: 10.1093/emboj/18.5.1114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Ludtke SJ, Baldwin PR, Chiu W. EMAN: semiautomated software for high-resolution single-particle reconstructions. J. Struct. Biol. 1999;128:82–97. doi: 10.1006/jsbi.1999.4174. [DOI] [PubMed] [Google Scholar]
  29. Ma Y, Pannicke U, Lu H, Niewolik D, Schwarz K, Lieber MR. The DNA-dependent protein kinase catalytic subunit phosphorylation sites in human Artemis. J. Biol. Chem. 2005;280:33839–33846. doi: 10.1074/jbc.M507113200. [DOI] [PubMed] [Google Scholar]
  30. Meek K, Gupta S, Ramsden DA, Lees-Miller SP. The DNA-dependent protein kinase: the director at the end. Immunol. Rev. 2004;200:132–141. doi: 10.1111/j.0105-2896.2004.00162.x. [DOI] [PubMed] [Google Scholar]
  31. Milne GT, Jin S, Shannon KB, Weaver DT. Mutations in two Ku homologs define a DNA end-joining repair pathway in Saccharomyces cerevisiae. Mol. Cell. Biol. 1996;16:4189–4198. doi: 10.1128/mcb.16.8.4189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mitelman F, Johansson B, Mertens F. The impact of translocations and gene fusions on cancer causation. Nat. Rev. Cancer. 2007;7:233–245. doi: 10.1038/nrc2091. [DOI] [PubMed] [Google Scholar]
  33. Niewolik D, Pannicke U, Lu H, Ma Y, Wang LC, Kulesza P, Zandi E, Lieber MR, Schwarz K. DNA-PKcs dependence of Artemis endonucleolytic activity, differences between hairpins and 5' or 3' overhangs. J. Biol. Chem. 2006;281:33900–33909. doi: 10.1074/jbc.M606023200. [DOI] [PubMed] [Google Scholar]
  34. Perry J, Kleckner N. The ATRs, ATMs, and TORs are giant HEAT repeat proteins. Cell. 2003;112:151–155. doi: 10.1016/s0092-8674(03)00033-3. [DOI] [PubMed] [Google Scholar]
  35. Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC, Ferrin TE. UCSF Chimera--a visualization system for exploratory research and analysis. J. Comput. Chem. 2004;25:1605–1612. doi: 10.1002/jcc.20084. [DOI] [PubMed] [Google Scholar]
  36. Rivera-Calzada A, Maman JD, Spagnolo L, Pearl LH, Llorca O. Three-dimensional structure and regulation of the DNA-dependent protein kinase catalytic subunit (DNA-PKcs). Structure. 2005;13:243–255. doi: 10.1016/j.str.2004.12.006. [DOI] [PubMed] [Google Scholar]
  37. Rosenthal PB, Henderson R. Optimal determination of particle orientation, absolute hand, and contrast loss in single-particle electron cryomicroscopy. J. Mol. Biol. 2003;333:721–745. doi: 10.1016/j.jmb.2003.07.013. [DOI] [PubMed] [Google Scholar]
  38. Saban SD, Nepomuceno RR, Gritton LD, Nemerow GR, Stewart PL. CryoEM structure at 9A resolution of an adenovirus vector targeted to hematopoietic cells. J. Mol. Biol. 2005;349:526–537. doi: 10.1016/j.jmb.2005.04.034. [DOI] [PubMed] [Google Scholar]
  39. Saban SD, Silvestry M, Nemerow GR, Stewart PL. Visualization of alpha-helices in a 6-angstrom resolution cryoelectron microscopy structure of adenovirus allows refinement of capsid protein assignments. J. Virol. 2006;80:12049–12059. doi: 10.1128/JVI.01652-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sali A, Blundell TL. Comparative protein modelling by satisfaction of spatial restraints. J. Mol. Biol. 1993;234:779–815. doi: 10.1006/jmbi.1993.1626. [DOI] [PubMed] [Google Scholar]
  41. Schwarz K, Ma Y, Pannicke U, Lieber MR. Human severe combined immune deficiency and DNA repair. Bioessays. 2003;25:1061–1070. doi: 10.1002/bies.10344. [DOI] [PubMed] [Google Scholar]
  42. Soding J, Biegert A, Lupas AN. The HHpred interactive server for protein homology detection and structure prediction. Nucleic Acids Res. 2005;33:W244–248. doi: 10.1093/nar/gki408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Sousa D, Grigorieff N. Ab initio resolution measurement for single particle structures. J. Struct. Biol. 2007;157:201–210. doi: 10.1016/j.jsb.2006.08.003. [DOI] [PubMed] [Google Scholar]
  44. 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;22:511–519. doi: 10.1016/j.molcel.2006.04.013. [DOI] [PubMed] [Google Scholar]
  45. Uematsu N, Weterings E, Yano K, Morotomi-Yano K, Jakob B, Taucher-Scholz G, Mari PO, van Gent DC, Chen BP, Chen DJ. Autophosphorylation of DNA-PKCS regulates its dynamics at DNA double-strand breaks. J. Cell Biol. 2007;177:219–229. doi: 10.1083/jcb.200608077. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Walker EH, Pacold ME, Perisic O, Stephens L, Hawkins PT, Wymann MP, Williams RL. Structural determinants of phosphoinositide 3-kinase inhibition by wortmannin, LY294002, quercetin, myricetin, and staurosporine. Mol. Cell. 2000;6:909–919. doi: 10.1016/s1097-2765(05)00089-4. [DOI] [PubMed] [Google Scholar]
  47. West RB, Yaneva M, Lieber MR. Productive and nonproductive complexes of Ku and DNA-dependent protein kinase at DNA termini. Mol. Cell. Biol. 1998;18:5908–5920. doi: 10.1128/mcb.18.10.5908. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Weterings E, van Gent DC. The mechanism of non-homologous end-joining: a synopsis of synapsis. DNA repair. 2004;3:1425–1435. doi: 10.1016/j.dnarep.2004.06.003. [DOI] [PubMed] [Google Scholar]
  49. Weterings E, Verkaik NS, Bruggenwirth HT, Hoeijmakers JH, van Gent DC. The role of DNA dependent protein kinase in synapsis of DNA ends. Nucleic Acids Res. 2003;31:7238–7246. doi: 10.1093/nar/gkg889. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Yoo S, Dynan WS. Geometry of a complex formed by double strand break repair proteins at a single DNA end: recruitment of DNA-PKcs induces inward translocation of Ku protein. Nucleic Acids Res. 1999;27:4679–4686. doi: 10.1093/nar/27.24.4679. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Zhou ZH, Chiu W. Determination of icosahedral virus structures by electron cryomicroscopy at subnanometer resolution. Adv. Protein Chem. 2003;64:93–124. doi: 10.1016/s0065-3233(03)01003-9. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

01
NIHMS42929-supplement-01.pdf (376.4KB, pdf)

RESOURCES