Binding of Inositol Hexakisphosphate (IP6) to Ku but Not to DNA-PKcs *

The nonhomologous DNA end joining (NHEJ) pathway is responsible for repairing a major fraction of double strand DNA breaks in somatic cells of all multicellular eukaryotes. As an indispensable protein in the NHEJ pathway, Ku has been hypothesized to be the first protein to bind at the DNA ends generated at a double strand break being repaired by this pathway. When bound to a DNA end, Ku improves the affinity of another DNA end-binding protein, DNA-PKcs, to that end. The Ku·DNA-PKcs complex is often termed the DNA-PK holoenzyme. It was recently shown that myo-inositol hexakisphosphate (IP6) stimulates the joining of complementary DNA ends in a cell free system. Moreover, the binding data suggested that IP6 bound to DNA-PKcs (not to Ku). Here we clearly show that, in fact, IP6 associates not with DNA-PKcs, but rather with Ku. Furthermore, the binding of DNA ends and IP6 to Ku are independent of each other. The possible relationship between inositol phosphate metabolism and DNA repair is discussed in light of these findings.

Double strand DNA breaks are among the most lethal DNA lesions, and they arise in somatic cells of multicellular eukaryotes spontaneously in the absence of external factors (1,2). Recently, we demonstrated that oxidative metabolism is the cause of at least a substantial fraction of these breaks (3).
There are two pathways for repairing chromosome breaks (4). Homologous recombination repairs breaks that arise during late S and G 2 of the cell cycle in vertebrate cells (5). Nonhomologous DNA end joining (NHEJ) 1 repairs double strand DNA breaks that arise during G 0 , G 1 , and early S phases of the vertebrate cell cycle (5).
The NHEJ repair process is thought to begin with the binding of a protein called Ku, which consists of Ku70 (70 kDa) and Ku86 (83 kDa) subunits, to DNA ends (6). Ku improves the DNA end binding affinity of the 469-kDa DNA-dependent protein kinase (DNA-PK cs ) (7). DNA-PK cs is a serine/threonine protein kinase that is only active when bound to DNA ends (8). When Ku and DNA-PK cs are both bound at the same DNA end, the complex is referred to as the DNA-PK holoenzyme, whereas the 469-kDa DNA-PK cs is termed the catalytic subunit (hence the subscript designation, cs) (9). The physiologic phosphorylation target of the 469-kDa DNA-PK cs has not been identified (10). After the DNA ends are trimmed into a form that is ligatable, a complex of XRCC4 and DNA ligase IV is responsible for carrying out the ligation specifically of double-strand breaks (11)(12)(13)(14)(15)(16)(17).
Evidence for the importance of Ku in eukaryotic NHEJ has been documented by a large body of genetic evidence (reviewed in Refs. 6, 18, and 19). Cells genetically deficient for either subunit of Ku are ionizing radiation sensitive. In addition, they are unable to carry out the rejoining phase of V(D)J recombination, which is a physiologic DNA double strand breakage and rejoining process restricted to lymphoid cells. The same phenotypic characteristics apply to other NHEJ components. In vitro, Ku has a high affinity for DNA ends (K D ϭ 10 Ϫ9 ) (7,20) and binds other single to double strand DNA transitions (21).
Inositol phosphate metabolism is a known form of intracellular signaling, primarily in the cytoplasm (22). There are only two lines of work describing effects of inositol phosphates on nuclear processes. One set of studies describes the roles of inositol phosphates on mRNA transport and transcriptional regulation (23,24). The second nuclear role of inositol phosphates was in NHEJ in nuclear extracts in which compatible DNA ends were studied for joining (25). In this latter study, it was reported that inositol hexakisphosphate stimulates the joining of complementary DNA ends in a cell-free system and associates with DNA-PK based on the co-elution of IP 6 and DNA-PK kinase activity from a gel filtration column. It was not determined in this study whether IP 6 was binding to the 469-kDa DNA-PK cs or to the Ku that was in the DNA-PK preparation. However, the homology of DNA-PK cs to proteins related to phosphatidylinositol metabolism was taken as an implicit indication that the IP 6 binds to the DNA-PK cs rather than to Ku. Here we demonstrate that, in fact, the IP 6 binds to Ku but not to DNA-PK cs , and the binding of DNA ends and IP 6 to Ku seem to be independent of each other.
Protein Purification-Native DNA-PK cs was purified as described previously (26) except that HeLa cells were used as the source for purification. C-terminal His-tagged Ku70 and nontagged Ku86 were co-expressed in the baculovirus system and purified as described previously (27). Nontagged Ku was expressed and purified as described previously (28) (a gift of Dr. J. Goldberg). The concentration of purified proteins was estimated by comparing to bovine serum albumin (BSA) standards on a Coomassie Blue stained SDS-PAGE gel.
In Vitro Immuno Pull-down of Inositol Phosphates-Immuno pulldown of IP 6 was performed in 50-l reactions that contain 10 mM Tris, pH 7.5, 50 mM NaCl, 10 mM MgCl 2 , 1 mM EDTA, 1 mM dithiothreitol, * This work was supported by National Institutes of Health grants (to M. R. L.) and by a P01 grant from the NIA. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
10% glycerol, and 0.1 mg/ml BSA. Anti-Ku70 monoclonal antibody (clone N3H10, NeoMarkers, Fremont, CA), anti-Ku86 monoclonal antibody (clone 111, NeoMarkers), and anti-DNA-PK cs monoclonal antibodies (clone 42-27 and 25-4) were used as indicated to reach a final total antibody amount of 20 g in each reaction. The rest of the reagents were added to each 50-l reaction as indicated: 1 pmol of [ 3 H]IP 6 or [ 3 H]IP 3 , 25-50 pmol of 35-bp DNA YM-8/YM-9 (blunt ds DNA, YM-8 5Ј-AGG CTG TGT TAA GTA TCT GCG CTC GCC CTC AGA GG-3Ј), 6.2-10 pmol of Ku, and 2.5 pmol of DNA-PK cs . Then, 100 l of 50% slurry of protein G-Sepharose (Amersham Biosciences) and 350 l of binding buffer were added to make the final volume 500 l. The pulldown reactions were allowed to proceed for 1 h at 4°C with constant mixing. After washing in the same binding buffer (without BSA) for 4 times (0.8 -1 ml of buffer for each time), the beads were resuspended in 200 l (2 ϫ 100 l) of binding buffer (without BSA) then mixed into 2 ml of ScintiVerse II scintillation fluid (Fisher Scientific). The amounts of bead-associated 3 H were measured in a liquid scintillation analyzer (model Tri-Carb 2100TR, Packard BioScience, Meriden, CT).
In Vitro Pull-down of Inositol Phosphates by DNA Beads-In each 50-l reaction (contains the same buffer components as the immuno pull-down reactions), 50 pmol of 5Ј-biotinylated 35-bp DNA YM-8/YM-9, 6.2 pmol of Ku, and 1 pmol of inositol phosphate were used. 100 l of 50% streptavidin agarose (Sigma) were then added into each reaction. The binding, washing, and scintillation counting steps were carried out as described above.
Electrophoretic Mobility Shift Assay (EMSA)-The gel shift assay was performed as described previously (20), and 0.5 nM labeled YM-6/ YM-7 (20) and 1 nM Ku were used for all the reactions. IP 6 and IP 3 were diluted in H 2 O first and then added to the reaction mixtures to the indicated final concentrations.
Surface Plasmon Resonance (SPR)-The SPR experiment was done as described previously (7,20). A 5Ј-biotinylated 35-bp DNA (sequence is the same as YM-8/YM-9) was immobilized on the streptavidin-coated surface of the sensor chip (Sensor chip SA, Biacore, San Diego, CA). Ku was diluted in the running buffer and injected at a flow rate of 5 l/min for 4 min. The injection of Ku was repeated to ensure that the surfaceimmobilized DNA on the sensor chip was saturated with Ku. Then DNA-PK cs (to a final concentration of 6.1 nM) and additional reagents (1 mM ATP, 1 mM MgCl 2 , and 0.1 mM IP 6 or IP 3 , final concentration indicated) were mixed in the running buffer and injected at 5 l/min for 4 min. Proteins bound to the sensor chip were allowed to dissociate for 6 min before the surface was regenerated with 0.05% SDS. The resulting sensorgrams were edited using the BIAevaluation software (version 3.0).

RESULTS
Ku, but Not DNA-PK cs , Binds to IP 6 -A previous study on the association of DNA-PK and IP 6 used a DNA-PK preparation that contained Ku and, possibly, contaminating DNA fragments (25). We were interested in testing whether binding of IP 6 to DNA-PK was dependent on Ku and/or DNA ends. Native DNA-PK cs and recombinant Ku were purified as described previously. DNA-PK cs was immobilized on protein G-Sepharose beads via monoclonal antibodies against it. Tritiated-IP 6 was incubated with these Immunobeads in the absence or presence of Ku and a 35-bp DNA (under the buffer conditions specified under "Experimental Procedures"). The DNA length of 35 bp was chosen, because this permits the binding of Ku and DNA-PK cs simultaneously to the same DNA molecule (7). The radioactivity associated with the extensively washed beads was then measured using a liquid scintillation counter. Surprisingly, IP 6 showed no association with DNA-PK cs alone, above the low level of background binding to the protein G-Sepharose beads (Fig. 1A, histogram bar 2 versus bar 1). This lack of association between IP 6 and DNA-PK cs was obtained regardless of the presence of the 35-bp linear DNA (Fig. 1A, bar 3  versus bar 2). When IP 6 was added to DNA-PK cs immunobeads along with Ku, there was only a near background level of binding detcted (Fig. 1A, bar 4), in agreement with our observation that Ku does not associate with DNA-PK cs in the absence of DNA (see below).
Interestingly, when the binding of IP 6 to DNA-PK cs was tested in the presence of Ku and 35-bp DNA, a substantial amount of binding was observed (Fig. 1A, bar 5). This could be because Ku induces a conformational change in DNA-PK cs to permit its binding of IP 6 . Alternatively, IP 6 might bind at the interface between Ku and DNA-PK cs when both are present on DNA, because 35 bp is long enough for Ku and DNA-PK cs to co-localize (7). Alternatively, IP 6 may simply bind to Ku. To test this, we omitted DNA-PK cs entirely and examined the binding of IP 6 to bead-immobilized Ku via monoclonal anti-Ku antibodies. In this case, we observed a very high level of IP 6 binding (Fig. 1A, bar 6). The specificity of binding of IP 6 to Ku was tested with another inositol polyphosphate, IP 3 , and only a background level of binding was detected (Fig. 1A, bar 7).
FIG. 1. IP 6 binds to Ku but not DNA-PK cs . A, antibody-bound protein G-Sepharose beads were generated with anti-DNA-PK cs (bars 1-5 and 7) or anti-Ku (anti-Ku70 and anti-Ku86, bar 6) monoclonal antibodies. DNA-PK cs and/or His-tagged Ku was added to the Immunobeads in the presence or absence of a 35-bp duplex oligonucleotides. Then tritiated IP 6 or IP 3 was mixed into each reaction. The residual radioactivity (in counts/min (cpm)) of the extensively washed beads is shown for each reaction. "ϩ" and "Ϫ" indicate presence and absence, respectively. B, nontagged Ku (bars 1 and 2) or His-tagged Ku (bars 3 and 4) were immobilized onto protein G-Sepharose beads with anti-Ku (anti-Ku70 and anti-Ku86) monoclonal antibodies. The association of tritiated IP 6 or IP 3 to the immunobeads was examined in absence of DNA as described in the legend to A.
Hence, it appeared that IP 6 was in fact binding to Ku rather than to DNA-PK cs We wanted to rule out the possibility that the six-amino acid histidine affinity tag on Ku70 might be responsible for the association between IP 6 and recombinant Ku. To test this, we compared the binding of IP 6 to purified Ku with and without the tag, and the association of IP 6 to these different preparations of Ku was indistinguishable (Fig. 1B). Therefore, the binding of IP 6 to Ku is completely unaffected by the affinity tag.
DNA Does Not Enhance or Inhibit the Binding of IP 6 to Ku-To investigate the role of DNA in the interaction between IP 6 and Ku, we used a monoclonal antibody against Ku70 to immobilize the Ku heterodimer. The level of binding was quite high, regardless of whether the 35-bp ds DNA was present ( Fig.  2A, bars 1 and 2). Immobilization with an anti-Ku86 monoclonal antibody gave indistinguishable results ( Fig. 2A, bars 3  and 4). As shown above, the binding of IP 6 to Ku was specific because IP 3 failed to bind ( Fig. 2A, bar 5). As was the case for anti-DNA-PK cs monoclonal antibodies (Fig. 1, bars 1-4 and 7), none of the anti-Ku antibodies bound to IP 6 (data not shown) nor to IP 3 ( Fig. 2A, bar 5). These data clearly indicate that IP 6 binds to Ku rather than to DNA-PK cs , and this binding is independent of the presence of DNA.
Further evidence for the binding of IP 6 to Ku can be observed when the experiments are configured in a manner where a linear DNA fragment was immobilized on streptavidin-agarose beads via a 5Ј-biotin linkage, and Ku was added to these DNA beads. IP 6 associated with the Ku⅐DNA beads, while IP 3 did not (Fig. 2B, bars 1 and 2). The streptavidin-agarose beads coated with biotinylated DNA could pull down the Ku⅐IP 6 complex as efficiently as protein G-Sepharose coated with anti-Ku antibodies (Fig. 2B, bar 1 versus bars 3 and 5). This is additional conclusive evidence that IP 6 binds to Ku, and this interaction is unaffected by DNA.
IP 6 Does Not Alter the Binding Properties of Ku to DNA Ends-We and others have previously done detailed studies of the binding of Ku to DNA ends. Recently, we showed that on linear DNA long enough to bind two Ku molecules (two-site linear DNA), the second Ku molecule loads with a 14-fold higher equilibrium constant (20). Namely, the two Ku molecules bind cooperatively to a two-site DNA molecule. We were interested in testing whether the binding properties of the Ku⅐IP 6 complex loading onto this two-site DNA would be altered compared with the loading of Ku alone onto the same DNA. To test this, concentrations of Ku and two-site DNA were chosen such that about 50% of Ku would bind to DNA; under these conditions, each 45-bp DNA molecule contains either one or two Ku molecules. The concentration of IP 6 was then varied over a 100,000-fold range that surrounded the physiologic concentration range of IP 6 in eukaryotic cells. Neither the noncooperative first Ku binding nor the cooperative second Ku binding was altered by IP 6 (Fig. 3A). Hence, it does not appear that IP 6 alters the association of Ku with DNA ends.
Ku has been designated "the DNA-binding subunit" of DNA-PK holoenzyme due to the fact that it enhances the binding of DNA-PK cs to a Ku-bound DNA end and stimulates the kinase activity of DNA-PK cs (9). Therefore, the binding of IP 6 to Ku might affect the fraction of DNA-PK cs that is in a complex with Ku and DNA. The kinase activity of DNA-PK on a synthetic substrate is not altered by IP 6 (25). To investigate the possibility that IP 6 might interfere or enhance the binding of DNA-PK cs to a Ku-bound DNA end, an SPR experiment was designed. The SPR technology allows the real-time monitoring of the changes in the mass of macromolecules associated with a surface. In our experiment, a biotinylated 35-bp ds DNA was immobilized on a streptavidin-coated surface. Ku was then allowed to bind to the free end of the DNA molecules on the surface to saturation or near saturation (achieved by two consecutive injections of the same Ku solution). The binding of DNA-PK cs was tested in the presence of IP 6 or IP 3 . Neither the association phase (the ascending phase starting with the injection of DNA-PK cs ) nor the dissociation phase (the descending phase starting with the termination of injection) was altered by IP 6 . This suggests that IP 6 does not affect DNA-PK cs in the association with the Ku⅐DNA complex.

DISCUSSION
Ku Binds to IP 6 -These results demonstrate that IP 6 , but not IP 3 , binds to Ku, and neither IP 6 nor IP 3 associate with DNA-PK cs . Previous work indicated that IP 6 enhanced the efficiency of DNA end joining in a cell extract system that was sensitive to inhibitors of DNA-PK cs (25). In that study when the mixture of IP 6 and DNA-PK was fractionated by a gel filtration column, one peak of IP 6 (detected by scintillation counting of each fraction) and the peak of DNA-PK (determined by assaying DNA-PK activity of each fraction) co-migrated. In light of our data, which indicate that Ku actually associates with IP 6 , one plausible explanation for the earlier observation could be that there was some contaminating DNA in the commercial DNA-PK preparation that allowed the association of Ku and DNA-PK cs and, therefore, the association of IP 6 and DNA-PK.
We have previously shown that DNA-PK cs and Ku do not associate in the absence of DNA ends (27), and this observation has been confirmed by our laboratory (7). In the current study, the failure to co-immunoprecipitate the Ku⅐IP 6 complex by DNA-PK cs immunobeads (Fig. 1, bar 4) further supported this point. Trace levels of Ku⅐DNA-PK cs associations seen by some laborartories may actually be due to low levels of contaminating DNA that can co-purify with DNA-PK cs , if specific purification steps are not included to remove the DNA. The contaminating DNA can then serve to bind both Ku and DNA-PK cs on the same fragments, given that Ku⅐DNA complexes bind DNA-PK cs 100-fold more efficiently than DNA alone can bind to DNA-PK cs .
The Potential Significance of Ku-IP 6 Interaction-IP 6 was reported to enhance the efficiency of joining of compatible DNA ends by crude cell extracts (25), and our study suggests that this might occur through association with Ku. One possible significance of this interaction could be that IP 6 alters some unidentified function(s) of Ku. A second possibility is that the Ku⅐IP 6 complex might interact with other factors in a way such that the overall NHEJ efficiency could be stimulated. Ku has been reported to physically and functionally interact with DNA-PK cs (only in presence of DNA ends) (9), DNA ligase IV⅐XRCC4 complex (29,30), and the Werners helicase, WRN (31,32). The alteration of the activities of any of these enzymes might result in a profound effect on the outcome of NHEJ. The elucidation of the function of Ku-IP 6 interaction awaits further studies.