HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 4, 3020-3029, January 25, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/4/3020    most recent M106711200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mårtensson, S. Articles by Hammarsten, O. Search for Related Content PubMed PubMed Citation Articles by Mårtensson, S. Articles by Hammarsten, O. Social Bookmarking                      What's this?

DNA-dependent Protein Kinase Catalytic Subunit

STRUCTURAL REQUIREMENTS FOR KINASE ACTIVATION BY DNA ENDS^*

Susanne Mårtensson and Ola Hammarsten^§

From the  Department of Clinical Chemistry, Sahlgrenska University Hospital, Gothenburg University, SE-413 45 Gothenburg, Sweden

Received for publication, July 17, 2001, and in revised form, October 16, 2001

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

DNA-dependent protein kinase (DNA-PK) is a DNA end-activated protein kinase composed of a catalytic subunit, DNA-PKcs, and a DNA binding subunit, Ku, that is involved in repair of DNA double-stranded breaks (DSBs). We have previously shown that DNA-PKcs interacts with single-stranded DNA (ssDNA) ends with a separate ssDNA binding site to be activated for its kinase activity. Here, the properties of the ssDNA binding site were examined by using DNA fragments with modified ssDNA extensions. DNA fragments with a wide range of ssDNA modifictations activated DNA-PKcs, indicating a relaxed specificity for the chemical structure of terminal nucleotides of a DSB. Methyl substitution of the phosphate backbone impaired kinase activation but not binding, indicating that interaction with the DNA backbone was involved in kinase activation. Experiments with RNA and RNA/DNA hybrid fragments suggested that the discrimination between RNA and DNA ends resides in the double-stranded DNA binding function of DNA-PKcs. DNA fragments exposing only one ssDNA end activated DNA-PKcs poorly, suggesting that DNA-PKcs distinguishes between DSBs and ssDNA breaks by simultaneous interaction with two ssDNA ends. These properties potentially explain how DNA-PKcs can be specifically activated by DSBs but still recognize the diverse chemical structures exposed when DSBs are introduced by ionizing radiation.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The resolution and detection of DNA double-stranded breaks (DSBs)¹ is important in maintaining genomic integrity. A single DSB can lead to chromosomal fragmentation and cell death or to chromosomal translocations and cancer. DSBs are induced by ionizing radiation, DNA replication or V(D)J recombination, a site-specific recombination process that generates the molecular diversity of the immune system. In response to ionizing radiation, cells target DSBs for repair and also activate several signaling pathways that lead to cell cycle arrest or apoptosis (1). Several protein kinases involved in the DSB signaling pathway have been identified and include DNA-dependent protein kinase (DNA-PK), ATM (2), and ATR, members of the phosphatidylinositol 3-kinase superfamily (3) (4-6). Since a single DSB is sufficient to permanently halt the cell cycle in yeast cells (7-9) and likely in mammalian cells (10), the signal that is generated must be massively amplified and tightly regulated.

DNA-PK may be the first protein to bind to a newly generated DSB, as suggested by its specificity for binding to and activation by DSBs, its cellular abundance, and the fact that it constitutes the major end-binding activity in nuclear extracts from human cells (11-14). Other lines of evidence also indicate that DNA-PK is involved in the cellular response to DSBs. DNA-PK kinase activity is required for effective DSB repair, survival (15-18), and recovery of DNA replication following ionizing radiation (19). DNA-PK has also been reported to be required for induction of apoptosis in response to ionizing radiation in thymocytes (20). Thus, DNA-PK appears to signal the presence of DSBs by its kinase activity. DNA-PK activation by DSBs can therefore be used as a model for how proteins specifically recognize and signal the presence of DSB lesions.

DNA-PK is a serine-threonine protein kinase that is composed of a 465-kDa catalytic subunit, DNA-PKcs, (21) and a DNA-binding component, termed Ku (12). Ku is a heterodimer of 70 and 86 kDa that binds specifically to DSBs (22-25) as a preformed ring that encircles DNA (26). DNA-PKcs contains a separate DNA binding function that binds DNA ends with lower affinity and specificity compared with Ku (27-29).

DNA-PKcs and Ku show only weak association in the absence of DNA (30). However, when Ku binds to a DNA end, a 12-amino acid sequence from the Ku80 C terminus becomes exposed that specifically binds and recruits DNA-PKcs to the DNA end (31, 32). In addition, Ku translocates inwardly from the DNA end, permitting DNA-PKcs to interact with the terminal nucleotides (25, 28). The cooperative interactions between DNA-PKcs, the DNA end, and the C terminus of Ku80 result in stable assembly of DNA-PK and robust kinase activation. This model for DNA-PK assembly is supported by specific photocross-linking of DNA-PKcs, but not Ku, to the terminal nucleotides when DNA-PKcs and Ku assemble on DNA ends (33).

In its purified form, DNA-PK kinase activity is specifically activated by double-stranded DNA ends (12). We have further specified that DNA-PK activation involves interaction with ssDNA ends exposed at the DSB (34). In addition, we have provided support for the existence of separate single- and double-stranded DNA binding sites in DNA-PKcs (35). Despite these advances, we still have limited knowledge of how chemical modifications of the terminal nucleotides of a DSB affect kinase activation. This is relevant, since a majority of DSBs generated by ionizing radiation expose nucleotides with altered chemistry, such as thymidine glycol, 8-hydroxyguanine, and phosphoglycolate that preclude direct repair. (36). Because DNA-PK is required for efficient repair of ionizing radiation-induced DSBs, the enzyme must be able to recognize these altered chemical structures. Indeed, it has been shown that DNA-PK is activated by DSBs introduced by ionizing radiation and DNA fragments that end with 3'-phosphoglycolates (37). It is possible, however, that a subset of the DSBs produced by ionizing radiation are not recognized by DNA-PK and thus are inefficiently repaired, such as DNA ends damaged with cis-platin (38-40).

Several studies have shown that ssDNA breaks (SSBs) activate DNA-PK poorly (11, 37, 41). Based on a low resolution structure of DNA-PKcs, we have proposed that kinase activation results when a ssDNA end is inserted into an enclosed channel present in the enzyme (35). This model provides a possible explanation for the requirement of ssDNA ends for kinase activation, since double-stranded DNA would not be able to enter the enclosed channel. However, a SSB potentially exposes ssDNA ends by fraying. Therefore, the requirement of ssDNA ends does not provide an explanation for how DNA-PK discriminates between SSBs and DSBs.

Because of the potential general importance of DNA-PK activation by DSBs in DNA damage signaling, we have analyzed the requirements for DNA-PKcs kinase activation by DSBs. A series of symmetrical DNA fragments with defined modifications of the terminal nucleotides were synthesized. The DNA fragments were tested for their ability to bind and activate DNA-PKcs. We found that a wide range of modified nucleotides activated DNA-PKcs, indicating a relaxed specificity for the chemical nature of the DNA terminus. In addition, DNA-PKcs kinase activation seemed to require interaction with two single-stranded ends, which potentially explains why DSBs, but not SSBs, are capable of activating DNA-PK effectively. The data suggest how DNA-PK can be specifically activated by DSBs but still permissive for the diverse chemical structures exposed when DSBs are introduced by ionizing radiation.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

DNA-PKcs Purification-- Placenta extract was prepared as described previously (28) with modifications. After the 65% ammonium sulfate precipitation the pellet was dissolved in 100 ml of buffer E (50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 0.35 M NaCl, 5% glycerol, 0.02% sodium azide, 20 mM 2-mercaptoethanol, and protease inhibitors (28)) and the conductance adjusted by additions of solid NaCl corresponding to 2 M NaCl. Proteins that coprecipitate DNA-PKcs during dialysis were removed by polyethylene glycol 8000 precipitation (5%, final concentration) for 30 min followed by centrifugation for 30 min at 10,000 rpm. The supernatant was dialyzed with 10 volumes of buffer E for 10 h and diluted with dilution buffer (10 mM Tris-HCl, pH 7.4, 1 mM EDTA, 10% glycerol, 20 mM 2-mercaptoethanol, and protease inhibitors (28)) to a conductance corresponding to 0.1 M NaCl. The diluted extract was passed through a 5-ml oligonucleotide affinity column prepared as described previously (28). Bound proteins were step-eluted with buffer B supplemented with 1 M NaCl and precipitated with 65% ammonium sulfate. Precipitated proteins were dissolved in 700 µl of buffer E supplemented with glycerol to a final concentration of 25% and separated on a 75-ml Sephacryl S300HR (Amersham Biosciences, Inc.) gel filtration column equilibrated in the same buffer. Fractions were analyzed on 5% SDS-PAGE. Fractions containing DNA-PKcs was pooled, diluted with dilution buffer to a conductance corresponding to 0.1 M NaCl, and fractionated on a second oligonucleotide affinity column (1-ml resin volume) as described previously (28). The pooled fractions from the oligonucleotide affinity resin were diluted with dilution buffer to a conductance corresponding to 0.1 NaCl and DNA-PKcs bound to a 0.5-ml High Q column (Bio-Rad) equilibrated in buffer E with a final concentration of 0.05 M NaCl. DNA-PKcs was step eluted with buffer E and stored at 70 °C. For kinase experiments frozen aliquots of DNA-PK (0.5 mg/ml) were diluted 20-40-fold in storage buffer (34), stored at 20 °C, and used within 2 weeks.

Kinase Assay and Electromobility Shift Assay (EMSA)-- Kinase assays were performed as described previously (34). In these experiments, DNA was added prior to DNA-PKcs, since it was found that other orders of addition had a negative effect on the reproducibility of the assay. The final NaCl concentration in kinase reactions was 15 mM after addition of DNA-PKcs. The activity obtained in the absence of added DNA was defined as background and subtracted from each value obtained in the presence of DNA. The background-subtracted values were plotted as a function of DNA concentration and fitted to Michaelis-Menten equation with KaleidiaGraph software. EMSA sample mix was prepared by mixing of labeled f12-3T (see "Construction and Purification of Oligonucleotides") in EMSA buffer (10 mM Tris-HCl, 1 mM EDTA, 5% glycerol, 20 mM 2-mercaptoethanol, and 0.2 mg/ml bovine serum albumin) to a final concentration of 1.5 nM. EMSA sample mix (8 µl) was mixed with competitor DNA diluted in TE buffer before addition of 0.6 µl of DNA-PKcs (0.02 mg/ml in storage buffer). After incubation for 10 min at room temperature, samples were analyzed by EMSA as described previously (28).

Construction and Purification of Oligonucleotides-- Oligonucleotides were synthesized on ABI oligonucleotide synthesizers using standard coupling and deprotection protocols. Phosphoamidites were purchased from Glen Research. The sequence of the f12 series of oligonucleotides was identical to the f12 DNA fragment (upper strand, f12-1: 5'-GGCCGCACGCGT-3'; lower strand, f12-2: 5'-ACGCGTGCGGCC-3') except for the single-stranded extensions indicated for each DNA fragment in figures. Both the 5' and the 3' termini contained hydroxyl groups to facilitate comparison between different DNA fragments. Annealing the upper strand of f12-3T with the lower strand of f12-3A formed the 18-bp, blunt-ended DNA fragment, f18. Oligonucleotides with exclusively 3' or 5' termini were synthesized using the 5'-5' or 3'-3' linkage phosphoamidites, respectively, from Glen Research. Single-stranded oligonucleotides were annealed in 1:1 molar ratio and the double-stranded product purified on 15% PAGE. The gels were stained with ethidium bromide and DNA visualized under long wave length UV light. The band corresponding to the full-length product was cut out. DNA was extracted by crushing the gel fragment and incubating in 10 volumes of extraction buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 M NaCl) with vigorous shaking overnight at room temperature. Gel pieces were removed by centrifugation and the DNA bound to a 50-mg Sep-Pac C18 column (Waters) prewashed sequentially with acetonitrile, 30% acetonitrile, 100 mM TEAA (triethylammonium acetate), pH 7.0 (Sigma), and 25 mM TEAA, pH 7.0. Bound DNA was washed with 25 mM TEAA, pH 7.0, and eluted with 0.6 ml of 30% acetonitrile and 100 mM TEAA, pH 7.0. The collected eluate was then extracted with phenol/chloroform to remove ethidium bromide and precipitated with ethanol. DNA pellets were washed two times with 99% ethanol, dissolved in TE buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA) and dialyzed extensively with TE buffer.

A dumbbell DNA fragment ending with ssDNA loops and containing a 10-base gap in the middle of the fragment (D34-gap10) was formed by annealing-mediated folding over of a self-complementary oligonucleotide (5'-TTTGGCCGCACGCGTTTACGCGTGCGGCCTTTTTTTTTTTGACAGGCTCCGCTTGCGGAGCCTGTCTTT-3') followed by purification on 15% PAGE. Three bases of unpaired dT were added to the 3' and 5' end to facilitate exposure of the ssDNA ends flanking the gap (Fig. 9). Similarly, a hairpin DNA fragment (H12-ss3'/5') with a ssDNA loop at one end and unpaired ssDNA ends at the other end was formed by annealing-mediated folding over of a self-complementary oligonucleotide (5'-TTTGGCCGCACGCGTTTTACGCGTGCGGCCTTTTTTTTTT-3'). The sequence of H12-ss3'/5' is identical to one-half of the D34-gap10 oligonucleotide, including the gapped section. The structure of these DNA fragments are shown in Fig. 9.

The D45-ss5' DNA fragment was constructed as depicted in Fig. 8B. A phosphorylated oligonucleotide containing a symmetrical branch (CLONTECH) (f12-2branch: 5'-PCTCCTT-branch-GGAGACGCGTGCGGCG-3") was synthesized. The sequence of the branch was complementary to the sequence immediately 3' to it, such that one of the two arms of the branch looped back to form a hairpin structure while the other remained a single-stranded tail (Fig. 8B). Oligonucleotide f12-2branch was annealed with and ligated to a phosphorylated complementary oligonucleotide (Br.5: 5'-pTTCTCACGCCGCACGCGT-3'). The ligation product was purified on a 15% denaturing PAGE and subsequently ligated to a second hairpin structure (f55-NRE2, 5'-TGAGAAAGAGAAAGACGACATCCGCCTGTTTTTCAGGCGGATGTCGTCTTTCTCT-3') as shown in Fig. 8B. This ligation product was purified on a native 15% PAGE followed by separation on a 15% denaturing PAGE run at 75 °C. Under these conditions, the fully ligated product migrated more slowly than the partially ligated oligonucleotide as described previously (34). The fully ligated product was excised from the gel, extracted, and purified as described above and characterized by labeling with TdT or PNK described below (Fig. 8C).

The lariat-forming oligonucleotide, lariat, was constructed by fusing the 3' ends of two partially complementary oligonucleotides (L2 and L3) as depicted in Fig. 8C. First L2 (L2: 5'-TTTTGGACGCGTGCGGCCTTTTTGGTGTT-3', sequence complementary to L3 is underlined) was annealed to L1 (L1: 5'-AAGGAAACACCAAAAA-3') to form a partially double-stranded DNA fragment with a five-nucleotide 5'-extension. The five-nucleotide extension was complementary to a branched oligonucleotide synthesized using the branching phosphamidite from CLONTECH (Branch.1: 5'-TCCTT-branch-T-3', Fig. 8C). The structure of Branch.1 is depicted in Fig. 8C. The ligation product (TTTTGGACGCGTGCGGCCTTTTTGGTGTTTCCTT-branch-T-3') was purified on 20% denaturing PAGE, eluted as described above, and ligated to oligonucleotide L3 (L3: 5'-TTTGGCCGCACGCGTCCATTTTTGGTGTT-3', sequence complementary to L2 is underlined). Prior to ligation, L3 was annealed to L1 to form a five-nucleotide 5'-extension complementary to the branched oligonucleotide and to oligonucleotide f12-2 to block annealing between the 15 bases of complementary sequence in L2 and L3 during the ligation reaction. The ligation product was purified on 20% denaturing PAGE, which separated it from L1 and f12-2, and allowed to self-anneal in TE buffer. The lariat DNA fragment was analyzed by labeling with PNK and TdT (below) to show that it lacked 3' ends, analysis on native PAGE to show the abnormally slow migration expected for such a single-stranded loop, and cleavage with the restriction enzyme BstUI to show that the complementary sequences in the lariat DNA fragment annealed properly (Fig. 8D). The DNA fragments D32-ssL5, f42, and f32-ss3'/5' have been described previously (34).

Labeling and Enzyme Digestion of Oligonucleotides-- Oligonucleotides were labeled on the 5' end by incubating with 2U T4 polynucleotide kinase (Promega) and [-³²P]ATP (5000 µCi/mmol, Amersham Biosciences, Inc.) in 10 µl of kinase buffer (Promega) for 30 min. The 3' end of oligonucleotides was specifically labeled with 5U terminal nucleotidyltransferase in 10 µl of TdT buffer (Promega) with [-³²P]dCTP (5000 µCi/mmol, Amersham Biosciences, Inc.) at 37 °C for 30 min. Labeled oligonucleotides were separated from unincorporated nucleotides by spin column chromatography on Sephadex G25 resin (Amersham Biosciences, Inc.). The upper strand of f12RNA-3U (f12RNA-3U-1: 5'-UUUGGCCGCACGCGUUUU-3') and f12-3T (f12-3T-1) was labeled as described above and digested with 0.01 mg/ml RNase A (Sigma) or 0.05 unit of RNase-free DNase I (Promega) in 10 µl of TE buffer supplemented with 20 mM MgCl₂ for 30 min at room temperature. Reactions were stopped by addition of 3 µl of formamide before analysis on 15% denaturing PAGE.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Experimental Setup-- We and others (27-29) have shown that the catalytic subunit of DNA-PK, DNA-PKcs, binds to and is activated for its kinase activity by double-stranded DNA ends in the absence of Ku. Under low salt conditions and using short DNA fragments with unpaired ssDNA at the ends, the kinase activity of DNA-PKcs is actually higher when Ku is omitted from the reaction (28, 34). In addition, several lines of evidence suggest that DNA-PKcs interacts directly with ssDNA ends when DNA-PKcs assembles on DNA fragments alone and when it binds DNA in a complex with Ku (33-35). We have therefore continued to characterize DNA-PKcs interaction with ssDNA in the absence of Ku, since the exclusion of Ku makes the experimental system more amenable to analysis by simple Michaelis-Menten kinetics and determinations of binding constants.

To explore DNA-PKcs interaction with ssDNA ends, a series of DNA fragments with modified nucleotides were synthesized. The design of these DNA fragments was based on the observation that a 12-bp blunt-ended DNA fragment, f12, is too short to allow efficient DNA-PKcs activation and binding (Fig. 1) (35). In contrast, f12-3T, which differs from f12 by addition of three bases of unpaired dT at each end, activated DNA-PKcs 20-fold more efficiently and bound DNA-PKcs with over a hundredfold higher affinity compared with the f12 DNA fragment (Fig. 1). Thus, both binding and activation of DNA-PKcs by f12-3T was dependent on the three bases of unpaired ssDNA.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   The base in ssDNA is not required for binding and activation of DNA-PKcs. A, DNA-PKcs (2 nM) was incubated with different concentrations of the DNA fragments defined in B, and the resulting kinase activity was measured. The kinase activity was related to the maximal kinase activity obtained with the DNA fragment f12-3T (100 = 345 mol of PO4/mol of DNA-PKcs/min). The data were fitted to the Michaelis-Menten equation to obtain Km and Vmax values for activation by each DNA fragment that is listed in Table I. B, the DNA fragments used in this experiment include a 12-bp (f12) and an 18-bp (f18) blunt-ended DNA fragment. Annealing the upper strand of f12-3T with the lower strand of f12-3A formed the f18 DNA fragment. Fragments with unpaired single-stranded extensions include f12-3T, f12-3A, and f12-3AP with 3 bases of unpaired dT, dA, or abasic deoxynucleotides at the 3' and 5'ends, respectively. For each DNA fragment, the overall structure of the DNA fragment and the chemical structure of the three unpaired nucleotides of the 5' terminus are shown.

In this study, the difference in activation between f12 and f12-3T was used to explore which part of ssDNA is responsible for DNA-PKcs activation and DNA binding. For this purpose, symmetrical DNA fragments derived from f12-3T were constructed where the three bases of unpaired dT were replaced with normal or modified nucleotides. The DNA fragments were then tested for their ability to bind and activate DNA-PKcs. We postulated that if DNA-PKcs failed to interact with the modified bases, neither kinase activation nor DNA binding would improve compared with the f12 DNA fragment. In contrast, if DNA-PKcs was able to interact with the extension of modified bases, efficient binding and kinase activation would be observed.

DNA-PKcs kinase reactions were performed over a wide range of DNA concentrations. The data from kinase reactions were analyzed by best fit to the Michaelis-Menten equation to allow determination of V_max and apparent K_m (Table I). Because components of the reaction other than DNA were present in saturating amounts, the assumption was made that DNA binding was the rate-limiting step for kinase activation. Therefore, the apparent K_m was used to reflect the binding of DNA-PKcs to each DNA fragment.

Effect of Base Alterations on DNA-PKcs Kinase Activation by DNA-- To study DNA-PKcs interaction with the base in ssDNA, DNA fragments with unpaired dA, dT, or abasic nucleotides at each end were used to activate DNA-PKcs (Fig. 1). As reported previously, a blunt-ended DNA fragment with the same number of nucleotides as f12-3T (f18) activated DNA-PKcs to a level 2-fold lower than f12-3T and bound the enzyme with 3-fold lower affinity (Fig. 1, Table I). As expected, addition of three bases of dA also stimulated kinase activation and binding but to a lesser extent compared with f12-3T. To test whether DNA-PKcs interaction with the base in the single-stranded extension was important for kinase activation, abasic nucleotides were used to synthesize f12-3AP (Fig. 1). The resulting DNA fragment, with abasic single-stranded extensions, bound and activated DNA-PKcs to the same level as observed with f12-3A. Therefore, DNA-PKcs favors DNA fragments ending with unpaired dT, but the base in ssDNA was not required for efficient kinase activation or binding.

Effect of Strand Polarity on DNA-PKcs Kinase Activation by DNA-- To examine the strand polarity preference for DNA-PKcs interaction with ssDNA, DNA fragments with exclusively 3' (f12-3T/3') or 5' (f12-3T/5') termini were synthesized by introduction of 3'-3' or 5'-5' linkages, as shown in Fig. 2B. The sequence and overall structure were otherwise identical to the f12-3T DNA fragment.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   DNA fragments with exclusively 3' or 5' termini activate DNA-PKcs. A, DNA-PKcs (2 nM) was incubated with different concentrations of the DNA fragments defined in B and the resulting kinase activity measured. The kinase activity was related to the maximal kinase activity obtained with the DNA fragment f12-3T (100 = 345 mol of PO4/mol of DNA-PKcs/min). The data were fitted to the Michaelis-Menten equation to obtain Km and Vmax values for each DNA fragment that is listed in Table I. B, DNA fragments used in this experiment. Oligonucleotides with exclusively 5' or 3' termini were made by introduction of 3'-3' or 5'-5' linkages as described under "Materials and Methods." C, analysis of DNA fragments by labeling with different enzymes. The unmodified DNA fragment f12-3T and DNA fragments with exclusively 3' (f12-3T/3') or 5' (f12-3T/5') termini were labeled with T4 polynucleotide kinase (PNK, specific for 5' ends) or TdT (specific for 3' ends) as described under "Materials and Methods." Labeled DNA was separated on a 15% denaturing PAGE and analyzed by autoradiography.

To show that the fragments exposed the expected strand polarity the fragments were labeled with T4 PNK or TdT that specifically labels the 5' or the 3' termini of DNA, respectively. As shown in Fig. 2C, PNK failed to label f12-3T/3' and TdT failed to label f12-3T/5'. This demonstrated that f12-3T/3' exposed exclusively 3' termini, and f12-3T/5' exposed only 5' termini.

The fragment with only 3' ends (f12-3T/3') bound and activated DNA-PKcs as well as f12-3T (Table I). The DNA fragment with only 5' termini activated DNA-PKcs to even higher levels and bound the enzyme with almost 2-fold higher affinity than f12-3T (Fig. 2A and Table I). This indicates that DNA-PKcs interaction with ssDNA lacks stringent strand polarity preference.

Importance of the Terminal Nucleotide on DNA-PKcs Kinase Activation by DNA-- We have previously shown that ssDNA ends are important for efficient DNA-PKcs kinase activation. This finding raises the possibility that the terminal nucleotides of the DNA end are specifically recognized by DNA-PKcs. To explore how alterations of the chemical nature of the terminal nucleotide influenced DNA-PKcs activation, a set of DNA fragments was synthesized with hydrophilic spacers added to each 5' and 3' terminus (Fig. 3B).

View larger version (27K): [in this window] [in a new window]   Fig. 3.   The chemical structure of the 3' and 5' terminus is important for DNA-PKcs kinase activation. A, DNA-PKcs (2 nM) was mixed with the DNA fragments defined in B in kinase buffer. The kinase activity was measured and related to the maximal kinase activity obtained with the DNA fragment f12-3T (100 = 310 mol of PO4/mol of DNA-PKcs/min). The data were fitted to the Michaelis-Menten equation to obtain Km and Vmax values for each DNA fragment that is listed in Table I. B, DNA fragments used in this experiment. The extension on f12-sp1T is composed of a mixed carbon/ether spacer attached to a single dT nucleotide. Fragment f12-sp2T has the same spacer with two nucleotides of dT at the terminus. Fragment f12-3T/3'5'NH3 is similar to f12-3T but has an amino group attached via a carbon spacer and a phosphate to the 3' and 5' terminus. For each DNA fragment, the overall structure of the DNA fragment and the chemical structure of the unpaired single strands of the 5' termini are shown.

First, both the 3' and the 5' positions in f12-3T were extended with an amino-containing spacer (f12-3T3'5'NH₃), as indicated in Fig. 3B. This modification lowered kinase activation 3-fold compared with f12-3T but had only a minor effect on binding as measured by K_m (Table I) or EMSA competition assay (data not shown). This indicated that the chemical structure of the terminal nucleotide affected DNA-PKcs kinase activation but not DNA binding.

To assess whether exposure of a correct terminal nucleotide was sufficient for efficient kinase activation, DNA fragments were constructed with hydrophilic spacers ending with dT (Fig. 3B). The DNA fragment f12-sp1T contained a single dT linked via a hydrophilic spacer to the double-stranded portion of the DNA fragment. The spacer had an approximate length corresponding to two bases. This potentially allowed the terminal dT to reach out the same distance from the double-stranded portion of the DNA fragment as the terminal dT in f12-3T (Fig. 3). Therefore, if interaction with the terminal nucleotide were sufficient, f12-sp1T would activate DNA-PKcs. However, f12-sp1T bound and activated DNA-PKcs poorly. Even at high concentrations of f12-sp1T, kinase activation was still 5-fold lower than that observed with f12-3T. We were unable to reach V_max or calculate K_m even at increased DNA concentrations (data not shown). The addition of two dT at the end of the hydrophilic spacer (f12-sp2T) improved activation but still produced a V_max 2-fold lower than f12-3T (Table I). Although it is possible that the hydrophilic spacer in f12-sp1T and f12-sp2T simply inhibited kinase activation in a nonspecific manner, the data indicated that exposure of the correct DNA terminus was not sufficient for efficient DNA-PKcs activation and that the nature of the ssDNA backbone is important for kinase activation.

Effect of Phosphate Backbone Alterations on DNA-PKcs Activation by DNA-- The effect of DNA backbone alterations was further examined by means of a set of DNA fragments where the ribose or the phosphate of the terminal nucleotides were altered, as indicated in Fig. 4B. Single-stranded extensions of ribonucleotides (f12-3rU) or phosphothioester linkages (f12-3T/thio) activated DNA-PKcs with a K_m and V_max close to that obtained with f12-3T (Table I). In contrast, a DNA fragment with methylphosphonate linkages, where the hydroxyl group in the backbone phosphate was changed to a methyl group (f12-3T/CH₃), activated DNA-PKcs poorly.

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Activation of DNA-PKcs involves interaction with the phosphate backbone in DNA. A, DNA-PKcs (2 nM) was incubated with different concentrations of the DNA fragments defined in B, and the resulting kinase activity was measured. The kinase activity was related to the maximal kinase activity obtained with fragment f12-3T (100 = 345 mol of PO4/mol of DNA-PKcs/min). The data were fitted to the Michaelis-Menten equation to obtain Km and Vmax values for each DNA fragment that is listed in Table I. B, DNA fragments used in this experiment. Fragment f12-3rU has ribouridine extensions, f12-3T/thio has extensions with thioester linkages, and f12-3T/CH3 has extensions with methylphosphonate linkages. For each DNA fragment, the overall structure of the DNA fragment and the chemical structure of the three unpaired nucleotides of the 5' terminus are shown.

The inefficient activation obtained with f12-3T/CH₃ could be due to poor binding to DNA-PKcs. However, the low kinase activity prevented determination of K_m. Therefore, binding of f12-3T/CH₃ to DNA-PKcs was examined with an EMSA competition assay. In this assay, radioactively labeled f12-3T was mixed with different concentrations of unlabeled f12-3T, f12 or f12-3T/CH₃ before addition of DNA-PKcs. The resulting DNA-PKcs·DNA complexes were analyzed by electrophoresis and quantified by a phosphorimager (Fig. 5). F12-3T/CH₃ competed almost as efficiently as the f12-3T DNA fragment. In contrast, the f12 DNA fragment essentially failed to compete under these conditions. Thus, methylphosphonate-modified DNA ends bound to DNA-PKcs but failed to activate the kinase effectively.

View larger version (53K): [in this window] [in a new window]   Fig. 5.   DNA-PKcs binds efficiently to methylphosphonate-modified DNA. Labeled f12-3T (2 nM) was mixed with the unlabeled competitor DNA indicated in the figure in EMSA buffer. DNA-PKcs (4 nM) was added and the resulting protein-DNA complexes resolved on 4% PAGE as described under "Materials and Methods." The gels were dried and analyzed by autoradiography (upper panel) and by a phosphorimager to allow graphic presentation of the protein-DNA complexes (indicated by PKcs) in each lane (lower panel).

The fact that f12-3T/CH₃ bound but failed to activate DNA-PKcs inspired the construction of a hybrid DNA fragment between f12-3T and f12-3T/CH₃. By annealing the upper strand from f12-3T with the lower strand from f12-3T/CH₃, a DNA fragment with one strand permissive for DNA-PKcs activation and one nonpermissive at each end was produced (f12-3T/1/2CH₃). The resulting hybrid DNA fragment bound and activated DNA-PKcs to levels close to the activation obtained with f12-3T (Fig. 6 and Table I). Thus, DNA-PKcs was efficiently activated when only one of the strands of the DNA fragment was permissive for kinase activation.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   A double-stranded DNA fragment with one strand of methylphosphonate-modified DNA activates DNA-PKcs. A, DNA-PKcs (2 nM) was incubated with different concentrations of the DNA fragments shown in B, and the resulting kinase activity was measured. The kinase activity was related to the maximal kinase activity obtained with the DNA fragment f12-3T (100 = 410 mol of PO4/mol of DNA-PKcs/min) B, DNA fragments used in this experiment. Fragments f12-3T, f12, and f12-3T/CH3 are defined in Fig. 3. Fragment f12-3T/1/2CH3 was formed by annealing the upper strand from f12-3T with the lower strand from f12-3T/CH3, resulting in a DNA fragment with one normal and one methylphosphonate-modified strand at each end.

Double-stranded RNA Fails to Activate DNA-PKcs-- The finding that f12-3rU activated DNA-PKcs indicated that the ssDNA binding site in DNA-PKcs recognized ribonucleotides. This opened the possibility that DNA-PKcs could be activated by double-stranded RNA ends, although several previous reports have shown that double-stranded RNA fail to bind and activate DNA-PK (11, 12, 42). To test this possibility directly, we synthesized a double-stranded RNA fragment (f12RNA-3rU) with the exact overall structure as f12-3T and tested it in DNA-PKcs kinase assays (Fig. 7A). The data showed that DNA-PKcs was not activated at all by the double-stranded RNA fragment. In addition, f12RNA-3rU failed to compete in the EMSA competition assay, indicating that DNA-PKcs was unable to bind to double-stranded RNA (Fig. 7B). To confirm that the RNA fragment was correctly synthesized and deprotected the top strand of f12RNA-3rU and f12-3T was incubated with RNase A or DNase I and analyzed on denaturing PAGE (Fig. 7C). As expected, f12RNA-3rU-1 was degraded by RNase A and resistant to DNase I, showing that it displayed a correct RNA structure. Thus, although DNA-PKcs was activated by a hybrid DNA fragment with three ribonucleotides at the terminus (f12-3rU), double-stranded RNA with the exact same overall structure failed to bind and activate DNA-PKcs. This indicates that DNA-PKcs discriminate RNA ends from DNA ends by its interaction with the double-stranded portion of DNA ends.

View larger version (35K): [in this window] [in a new window]   Fig. 7.   DNA-PKcs interaction with a double-stranded RNA fragment. A, the double-stranded RNA fragment, f12RNA-3rU, failed to activate DNA-PKcs. DNA-PKcs (2 nM) was incubated with different concentrations of the f12RNA-3rU or f12-3T, and the resulting kinase activity was measured. The kinase activity was related to the maximal kinase activity obtained with the DNA fragment f12-3T (100 = 335 mol of PO4/mol of DNA-PKcs/min). B, labeled f12-3T DNA fragment (2 nM) was mixed with the unlabeled competitor DNA indicated in the figure in EMSA buffer. DNA-PKcs (4 nM) was added and the resulting protein-DNA complexes resolved on 4% PAGE as described under "Materials and Methods." The gels were dried and analyzed by autoradiography. C, analysis of f12RNA-3rU-1. The upper strand of f12RNA-3rU (f12RNA-3rU-1) and f12-3T (f12-3T-1) was labeled with T4 PNK and digested with RNase A or RNase-free DNase I as indicated in the figure. Reactions were separated on 20% denaturing PAGE and analyzed by autoradiography.

DNA Fragments with Only One ssDNA End Activate DNA-PKcs Inefficiently-- Several studies indicate that SSBs are poor activators of DNA-PK (11, 37, 41). At the position of a SSB, the DNA strands become flexible and capable of exposing ssDNA ends by fraying. A SSB could in theory satisfy the need for ssDNA ends for kinase activation (34). In addition, DNA-PK is potentially able to assemble on SSBs, since Ku, the DNA-targeting subunit of DNA-PK, binds SSBs (23, 43). It is therefore unclear how DNA-PK is able to distinguish SSBs from DSBs. However, a SSB is not able to simultaneously expose two ssDNA ends at the correct distance found at a DSB. Therefore, DNA-PK could differentiate between SSBs and DSBs if it was dependent on simultaneous interaction with two ssDNA ends.

To explore whether one or two ssDNA ends were required for kinase activation, a dumbbell DNA fragment exposing only one 5' ssDNA end was constructed using a branched oligonucleotide (Fig. 8B). After annealing and ligation, the fully ligated dumbbell DNA fragment was purified using a combination of native and denaturing gels (Fig. 8C) (34). The correct configuration of the 5' ssDNA end in D44-ss5' was verified by the ability of PNK, which labels only 5' ends, but not TdT, which labels only 3' ends, to label the fully ligated product (Fig. 8C). In addition, the ability of PNK to efficiently label D44-ss5' indicated that the ssDNA tail was exposed to the solution.

View larger version (55K): [in this window] [in a new window]   Fig. 8.   Construction of D44-ss5' and the lariat DNA fragment A. Graphical presentation of the construction of D44-ss5'. The sequence and structure of the branched oligonucleotide (f12-2branch) is shown. Note that the sequence of the four terminal nucleotides of the branches are complementary to the four nucleotides 3' of the branch, allowing formation of a hairpin. First, the hairpin-forming branched oligonucleotide was annealed to a complementary oligonucleotide as described under "Materials and Methods." After ligation (I) and purification of the ligation product, the branched hairpin oligonucleotide was ligated (II) to a second hairpin oligonucleotide, resulting in dumbbell DNA fragment with a single-stranded 5' terminus. For details see "Material and Methods." B, purification of D44-ss5'. First, the DNA products formed after ligation II were separated on native PAGE and stained with ethidium bromide. The DNA in the band corresponding to the ligation products (lane 1) was recovered and further purified on denaturing PAGE (lane 2) that resolve the partially ligated (lower band) and fully ligated (upper band) products. DNA recovered from the band in lane 1 was labeled with T4 PNK (lane 3) or TdT (lane 4), separated on denaturing PAGE, and analyzed by autoradiography. The fully ligated product (upper band) failed to label with TdT (specific for 3' terminus), but efficiently labeled with T4 PNK (specific for 5' terminus), as expected. C, construction of the lariat DNA fragment. The sequence of the branched linker oligonucleotide (Branch.1) is shown at the top. First, the branched linker oligonucleotide was ligated (I) to the 3' end of oligonucleotide L2. The ligation product was purified on denaturing PAGE and ligated (II) to the oligonucleotide L3 that contained a partially complementary sequence to L2. To mask the complementarity during ligation, L3 was annealed to an oligonucleotide complementary to this region, prior to ligation as shown. The ligation product was purified on denaturing PAGE and allowed to self-anneal. D, verification of the structure of the lariat DNA fragment. The lariat DNA fragment was labeled with PNK (lane 1) or TdT (lane 2), separated on 20% denaturing PAGE, and analyzed by ethidium bromide staining (left panel) and autoradiography (middle panel). The 68-base lariat DNA fragment migrated at the expected position and was not labeled with TdT (specific for 3' terminus), showing that it lacks 3' ssDNA ends. Proper annealing of the 15-bp double-stranded region was shown by cleavage of the lariat DNA fragment with the restriction enzyme BstUI and analysis on 20% denaturing PAGE (right panel: lane 3, undigested lariat; lane 4, digested lariat). On a 20% nondenaturing PAGE the 68-base lariat DNA fragment migrated slower than a 71-bp DNA fragment as expected for a looped DNA fragment. In contrast, the unligated annealing product between L2 and L3 (L2/L3, lane 6) migrated at a position reflecting the expected size of the DNA fragment (49 bases).

D44-ss5' activated DNA-PKcs poorly and to similar levels observed with a dumbbell DNA fragment without the 5' ssDNA end (D32-Lss5) (Fig. 9) (34)). The poor activation was not due to the presence of contaminating inhibitors, because restriction enzyme cleavage of D44-ss5' resulted in robust activation of the kinase to the level observed with a blunt-ended DNA fragment (f48). Because of the polarity of oligonucleotide synthesis, a corresponding DNA fragment exposing only one 3' end could not be constructed using a branched phosphoamidite. However, because DNA-PKcs activation lacked DNA strand polarity preference (Fig. 2), DNA-PK is likely to be inefficiently activated by a single 3' terminus as well.

View larger version (21K): [in this window] [in a new window]   Fig. 9.   DNA-PKcs kinase activation by DNA fragments exposing only one ssDNA end. DNA-PKcs (1.0 nM) was incubated with 60 nM of the DNA fragments shown in the figure, and the resulting kinase activity was measured. Kinase activity was related to the maximal kinase activity obtained with the DNA fragment f32-ss3'/5' (100 = 395 mol of PO4/mol of DNA-PKcs/min). To rule out the possibility that the preparation of D44-ss5'contained any inhibitory activity, D44-ss5' was cleaved with the restriction endonuclease MluI (D44-ss5'/MluI) and tested for kinase activation. Error bars represent data from three separate experiments. D44-ss5' is a dumbbell DNA fragment with loops of three dT at each end and a six-base 5' ssDNA branch extending from one end. D34-gap10 is a dumbbell DNA fragment with ssDNA loops of two dT at each end and a gap of ten bases of dT flanked by 12 bp of double-stranded DNA at each side. The ssDNA ends flanking the gap contain three bases of unpaired dT. H12-ss3'/5' is a half-dumbbell DNA fragment with a ssDNA loop of two dT at one end and three and ten bases of unpaired dT at the 5' and 3' termini, respectively. The sequence of H12-ss3'/5' is identical to one double-stranded portion including the gap of D34-gap10. Lariat is a looped DNA fragment with a 15-bp double-stranded portion and a 44-base single stranded loop connecting its 3' termini, as described in the legend to Fig. 8D. The 5' termini have 3 bases of unpaired dT. L2/L3 is a 15-bp double-stranded DNA fragment with ssDNA extensions of 17 bases at the 3' ends and three bases at the 5' ends. L2/L3 was formed by annealing of the oligonucleotides used to construct the lariat DNA fragment. D32-ssL5 is a 32-bp dumbbell DNA fragment with loops of five dT at each end. Fragment f42 is a blunt-ended 42-bp DNA fragment and f32-ss3'/5' is a 32-bp DNA fragment with five bases of unpaired dT at each end. DNA fragments are further defined in Ref. 34.

While the previous data suggest that DNA-PKcs kinase cannot be activated by a single ssDNA end, an alternative interpretation is that the hairpin loop at the end of D44-ss5' structurally interferes with activation by the 5' ssDNA end that would otherwise occur. We therefore explored DNA-PKcs activation by a similar dumbbell DNA fragment with a central gap of ten bases on the strand opposite two internal 3' and 5' ssDNA ends (H34-gap10, Fig. 9). This DNA fragment simulates the structure of a SSB with frayed DNA strands but may allow better access to the ssDNA ends due to the ten base gap between the open ends (Fig. 9). In addition, since the ssDNA ends flanking the gap have either 3' or 5' polarity, strand polarity preference for activation could potentially be explored. Similar to D44-ss5', the gapped DNA fragment activated DNA-PKcs poorly. A DNA fragment identical to one side of D34-gap10 (H12-ss3'5') efficiently activated DNA-PKcs, ruling out the possibility that D34-gap10 failed to activate DNA-PKcs due to the presence of a long ssDNA region or loop (Fig. 9).

Finally, we constructed a lariat DNA fragment by fusing the 3'-ends of two complementary oligonucleotides. The 3' ends of the complementary oligonucleotides were fused by ligation to a branched linker oligonucleotide containing two phosphorylated 5' ends (Fig. 8C). The resulting oligonucleotide self-annealed to a looped DNA fragment, with a 15-bp double-stranded region flanked by exclusively 5' ssDNA ends (Fig. 8, C and D). As shown in Fig. 9, the lariat DNA fragment failed to activate DNA-PKcs. In contrast, when the same oligonucleotides used for construction of the lariat were annealed without fusion of the 3'-termini (L2/L3), DNA-PKcs was efficiently activated, ruling out the possibility that long ssDNA regions in the lariat interfered with kinase activation.

Therefore, the inability of three distinct DNA constructs that expose only one ssDNA end (D44-ss5', D34-gap10, and the lariat DNA fragment) to efficiently activate DNA-PKcs suggests that robust activation of DNA-PKcs involves simultaneous interaction with two ssDNA ends.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

In previous reports, we have shown that ssDNA ends activate DNA-PKcs. This provided a possible explanation for how DNA-PK is specifically activated by broken DNA, since the DNA backbone must be damaged to expose the necessary ssDNA ends. The data did not clarify why SSBs activate DNA-PK poorly nor which part of the ssDNA ends is recognized by DNA-PK. The molecular details of DNA-PK kinase activation by DNA ends are important to understand how the cell is able to recognize the chemically diverse structures that occur when DSBs are introduced by ionizing radiation.

Collectively, our data indicate that DNA-PKcs has a relaxed specificity for the chemical structure of the terminal nucleotides of a DNA end. DNA fragments with abasic nucleotides bound and activated DNA-PKcs to the same extent as a DNA fragment ending with unpaired dA. This suggests that the base position in DNA is not specifically contacted by DNA-PKcs. It is therefore unlikely that the base modifications introduced by ionizing radiation will impede DNA-PK activation. DNA fragments with exclusively 3' or 5' polarity at the ends both activated DNA-PKcs, indicating a lack of strand polarity preference. No effect on DNA binding and kinase activation was evident when the terminal nucleotides were replaced by ribonucleotides or nucleotides with phosphothioester linkages. We also found that to prevent activation of DNA-PKcs a DSB must contain an incompatible chemical modification on both strands (Fig. 6). The data indicate that the ssDNA binding site in DNA-PKcs accepts a wide variety of chemical modifications, lacks strand polarity preference, and allows activation by DSBs even when one of the strands is incompatible with activation. Therefore, production of DSBs that avoid detection by DNA-PK is likely to be a rare event. This is in line with the finding that linear plasmid DNA generated by  radiation activates DNA-PK to the same extent as plasmid DNA cleaved with restriction enzymes (37). However, it is possible that more complex DSBs produced by high linear energy transfer radiation (44) and drugs such as calicheamicin or etoposide are not recognized by DNA-PK. In fact, DNA ends damaged with the anticancer drug cis-platin are unable to activate DNA-PK, and this could explain the ability of cis-platin to sensitize cells to ionizing radiation (38-40).

Three findings indicate that, in addition to the requirement for ssDNA ends, interaction with the DNA backbone is important for kinase activation. A DNA fragment extended with a hydrophilic spacer holding a dT at each of the four ssDNA ends bound and activated DNA-PKcs poorly. Although it is possible that the hydrophilic spacer simply interfered with activation, the data suggested that exposure of a correct terminal nucleotide was not sufficient for kinase activation. This indicated that efficient activation of DNA-PKcs involved interaction with the DNA backbone. This conclusion was further strengthened by the finding that methylphosphonate-modified DNA was bound by DNA-PKcs but activated its kinase poorly. The methylphosphonate modification replaces the negatively charged hydroxyl group in the phosphate backbone, resulting in an uncharged backbone. This suggest that the negative charge of the DNA backbone is required for kinase activation but not for binding. In addition we found, in agreement with others, that a double-stranded RNA fragment was unable to activate DNA-PKcs (11, 12). However, a hybrid DNA fragment with identical overall structure but with three ribonucleotides at all four single-stranded ends activated DNA-PKcs efficiently. This indicates that the ssDNA binding site in DNA-PKcs fails to distinguish RNA from DNA strands. Therefore DNA-PK must distinguish RNA from DNA by its interactions with the double-stranded portion of DNA ends. Together these three observations show that interactions with the backbone, not just the ssDNA ends, of a DSB is involved in DNA-PKcs kinase activation.

Ionizing radiation produces 10-fold more SSBs in DNA compared with DSBs (44). SSBs are known to be relatively nontoxic to cells (44) and are quickly repaired. In contrast, the few DSBs that are produced by ionizing radiation are very toxic to cells and take several hours to be fully repaired (45, 46). Thus, cells have a potent signaling system capable of distinguishing between SSBs and DSBs. Several reports indicate that DNA-PK is inefficiently activated by SSBs (11, 37, 41). It is therefore possible that DNA-PK plays a role in the discrimination among SSBs and DSBs in the cell. This possibility is supported by the finding that DNA-PK-deficient cells repair DSBs inefficiently, but have no defect in SSB repair (47, 48). One way for DNA-PK to discriminate between DSBs and SSBs is by simultaneous interaction with two ssDNA ends. A SSB is unable to expose two ssDNA ends with the configuration found in the context of a DSB. Therefore, a requirement for two ssDNA ends would render DNA-PK activation specific for DSBs. In support for this possibility, we found that three distinct DNA constructs that expose only one ssDNA end (D44-ss5', D34-gap10, and the lariat DNA fragment) activated DNA-PKcs poorly. It is possible that these DNA fragments contain structures that mask activation by a single ssDNA end. However, the combined data support the possibility that DNA-PK discriminates SSB from DSB by a requirement for interaction with two ssDNA ends for efficient kinase activation.

Based on a low resolution structure of DNA-PKcs, we have proposed that DNA-PKcs is activated when one ssDNA end enters an enclosed channel present in the enzyme (Fig. 10A). Although this model potentially explains the requirement for ssDNA ends for activation, it does not provide a good explanation why DNA-PKcs seemes to require simultaneous interaction with two single strands for activation (Fig. 9). One possibility is that activation of DNA-PKcs requires dimerization of two DNA-PKcs molecules each bound to a DNA end as we suggested previously (34) (Fig. 10B). This configuration would have the added benefit of restricting full activation of DNA-PK until synapsis of the two ends of a DSB is complete. Since the uncharged backbone of methylphosphonate-modified DNA bound DNA-PKcs but failed to activate its kinase activity, it is possible that specific contacts with the charged phosphate backbone are involved in kinase activation. According to this model, interaction with the charged phosphate could occur within the enclosed channel.

View larger version (41K): [in this window] [in a new window]   Fig. 10.   A model for DNA-PKcs activation by two ssDNA ends. A, electron crystallography structure of DNA-PKcs with superimposed cartoon of a bound DNA end (35). According to this model the double-stranded portion of a DNA end binds to the open channel in DNA-PKcs. Interaction with ssDNA occurs when the DNA is melted and inserted into an opening of the enclosed channel present in DNA-PKcs. The dashed line indicates the location of the cross-sectional slice through the molecule used to produce the cartoon shown in B. B, model for how DNA-PKcs is activated by two ssDNA ends. Using the cross-sectional slice as a model, two DNA-bound DNA-PKcs molecules first form a synapsis complex. The kinases are activated in concert when two ssDNA ends are threaded into two openings of the enclosed channel in each enzyme.

In conclusion, we have found that DNA-PKcs recognizes DNA ends containing a wide range of modified nucleotides and seems to require simultaneous interaction with two ssDNA ends for activation. These properties potentially make DNA-PK activation specific for DSBs but still capable of recognizing DNA ends with damaged terminal nucleotides. DNA-PK is therefore a suitable candidate for detection of DSBs when cells are damaged by ionizing radiation. With this knowledge it will be interesting to examine DNA-PK recognition of cross-linked DNA fragments and complex strand breaks such as those produced by cytotoxic cancer therapy and high linear energy transfer radiation.

	ACKNOWLEDGEMENTS

We thank Lisa G. DeFazio and James Lorens for advice and careful reading of the manuscript.

	FOOTNOTES

^* This work was supported by the Swedish Cancer Society, the Swedish research council, King Gustav V Jubilee Clinic Cancer Research Foundation, and Sahlgren's Hospital Research Foundation.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ To whom corresepondence should be addressed. Fax: 46-31-82-84-58; E-mail: ola.hammarsten@clinchem.gu.se.

Published, JBC Papers in Press, November 7, 2001, DOI 10.1074/jbc.M106711200

	ABBREVIATIONS

The abbreviations used are: DSB, double-stranded break; EMSA, electromobility shift assay; TdT, terminal deoxynucleotidyltransferase; PNK, T4 polynucleotid kinase; ssDNA, single-stranded DNA; SSB, single-stranded break; dT, deoxythymidine; dA, deoxyadenine; DNA-PK, DNA-dependent protein kinase; DNA-PKcs, DNA end-activated protein kinase composed of a catalytic subunit; TEAA, triethylammonium acetate.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

L. C. Shank, J. B. Kelley, D. Gioeli, C.-S. Yang, A. Spencer, L. A. Allison, and B. M. Paschal Activation of the DNA-dependent Protein Kinase Stimulates Nuclear Export of the Androgen Receptor in Vitro J. Biol. Chem., April 18, 2008; 283(16): 10568 - 10580. [Abstract] [Full Text] [PDF]