Activation of Protein Kinase Tel1 through Recognition of Protein-Bound DNA Ends ^▿

Strains and plasmids.

The FLAG-tagged TOP1 construct has been described previously (42). Other tagged constructs have been described previously (26, 46, 47). The mre11-3 strain was constructed as follows. The mre11-3 mutant, containing the H125L and D126V substitutions (7), was cloned into a LEU2-marked YIp plasmid, creating the YIpL-mre11-3 plasmid. Cells were then transformed with linearized YIp-mre11-3. The tel1-KN or xrs2-11 mutation has been described previously (46, 47). Deletion mutations were obtained by PCR-based methods (26, 46). YCpT-Rad9-HA was constructed by transferring the RAD9-HA construct from YIp-RAD9-HA (45). The pGAL-FLAG-TEL1-KN plasmid was created from pGAL-FLAG-TEL1 by replacing the NheI-SalI fragment with that from YEp-TEL1-KN-HA (47). The pGAL-EcoRI plasmid (YCp50 carrying GAL-EcoRI) was obtained from M. Resnick (34). Other plasmids have been described elsewhere (26, 46).

Protein purification.

The Mre11, Rad50, and Xrs2 proteins were individually purified and assembled into the heterotrimer MRX complex as described previously (8). The Flag-Tel1 or Flag-Tel1-KN protein was purified from xrs2Δ mutants carrying pGAL-FLAG-TEL1 or pGAL-FLAG-TEL1-KN, respectively (26). GST-Rad53 proteins were expressed in Escherichia coli and were purified from extracts as described previously (47).

DNA substrates.

The unmodified 150-bp DNA fragment (N) was prepared by PCR with the KSX001-KSX002 primer set using pUC19 as a template. The 5′-biotinylated 150-bp DNA fragment (5′) was generated with the KS1926-KS2058 primer set (each of these primers is biotinylated at the 5′ end) using pUC19 as a template. The 3′-biotinylated 150-bp DNA fragment (3′) was prepared as follows. First, a DNA fragment was amplified by PCR with the KSX001-KSX053 primer set using pUC19 as a template. The resulting fragment, containing the BamHI cleavage sequence at the both ends, was first cleaved with BamHI and then treated with Klenow fragment in a buffer containing 0.2 mM dATP, dGTP, dTTP, and biotin-dCTP (Perkin-Elmer). The 150-bp DNA fragment with biotin embedded internally (I) was prepared by PCR using the KS2351-KS2352 primer set in a reaction mixture containing 0.2 mM dATP, dGTP, dTTP, and biotin-dCTP. The KS2351 and KS2352 oligonucleotides anneal each other, generating two primer-template junctions. The ssDNA regions of the annealed oligonucleotides are filled in the first cycle of PCR, and the resulting DNA fragments act as a template for PCR. Each of the PCRs incorporates a unique C into both strands of the DNA fragment. The nonbiotinylated 150-bp DNA fragment (N1) was generated using primers KS2351 and KS2352 in a reaction mixture containing deoxynucleoside triphosphates (dNTP) without biotin-dCTP. All PCR products were purified with the QIAquick PCR purification kit (Qiagen). For the DNA mobility shift or nuclease assay, DNA fragments were ³²P labeled during PCR in a reaction mixture containing [α-³²P]dATP. To prepare Fab-tethered DNA fragments, biotinylated DNA fragments were incubated with monovalent Fab fragments of affinity purified goat antibiotin antibodies (Rockland) in TE (10 mM Tris-Cl, 1 mM EDTA) for 20 min on ice. The unmodified 800-bp DNA fragment containing the GAL1-GAL10 promoter sequence was generated by PCR with the KSX001-KSX002 primer set using YCplac33 carrying the GAL1-GAL10 promoter sequence as a template. The sequences of the oligonucleotides described here are given in Table 1.

Tel1 kinase assay.

HA-tagged Tel1 protein was immunoprecipitated as described previously (26, 47). Cells were disrupted in a lysis buffer lacking Mg²⁺ or Mn²⁺ (26). Kinase reactions were initiated by the addition of a substrate (2 μg of glutathione S-transferase [GST]-Rad53 or 1 μg of PHAS-1) to immunoprecipitates in 80 μl of the reaction buffer (20 mM HEPES-KOH [pH 7.5], 4 mM MgCl₂, 4 mM MnCl₂, 3 μM ATP) containing 3 μCi [γ-³²P]ATP (3,000 Ci/mmol; Perkin-Elmer). Kinase reactions were also reconstituted in vitro with 2 pmol of purified Tel1 or Tel1-KN proteins, MRX complexes, or DNA fragments in 80 μl of the reaction buffer containing 3 μCi [γ-³²P]ATP. After 10 min of incubation at 30°C, the reaction was terminated by the addition of 20 μl of 5× sodium dodecyl sulfate (SDS)-polyacrylamide gel loading dye. The reaction mixtures were separated on SDS-polyacrylamide gels, and phosphorylation was quantified with a phosphorimager system (Typhoon 8600; GE Healthcare).

Nuclease assay.

Nuclease assays of the MRX complex were monitored with some modifications according to the method described previously (69). Two picomoles of ³²P-labeled DNA fragments was incubated with 2 pmol of the MRX complex in 80 μl of the reaction buffer at 30°C. After treatment with proteinase K, the reaction mixtures were subjected to gel electrophoresis under nondenaturing conditions and were analyzed by a phosphorimager system.

DNA binding of the MRX complex.

Two picomoles of MRX was incubated with 2 pmol of DNA fragments in 80 μl of the binding buffer (20 mM HEPES-KOH [pH 7.5], 4 mM MgCl₂, 3 μM ATP, 0.01% Triton X-100) for 10 min at 30°C. After incubation, MRX complexes and DNA fragments were cross-linked with 1% formaldehyde and were subjected to immunoprecipitation with anti-Xrs2 antibodies (a gift from J. Petrini). Immunoprecipitated DNA fragments were quantified by real-time PCR. Relative enrichment was determined by normalizing signals from DNA fragments immunoprecipitated by anti-Xrs2 antibodies to control signals from DNA fragments treated with mock antibodies.

Detection of DNA-cross-linked histone and Top1.

DNA-protein complexes were detected by the following method after some modifications of previously published procedures (24, 29). Cells were suspended in a denaturing lysis buffer (8 M guanidine hydrochloride, 30 mM Tris base, 10 mM EDTA, 1% Sarkosyl [pH 7.5]) and were disrupted mechanically with glass beads. Extracts were loaded on a CsCl gradient consisting of four layers (1.82, 1.72, 1.50, and 1.45 g/ml). The gradients were centrifuged using a SW60 rotor at 40,000 rpm for 24 h and were fractionated from the top. The fractions containing chromosomal DNA were precipitated by the addition of trichloroacetic acid (TCA). Precipitates were subjected to immunoblotting with anti-FLAG antibodies, anti-H2B antibodies (Millipore), or anti-H3 antibodies (a gift from A. Verreault). Precipitates were also examined by real-time PCR for DNA quantification.

Other methods.

Immunoblotting and the chromatin immunoprecipitation (ChIP) assay were performed as described previously (26, 47). The antibody against hemagglutinin (HA) was purchased from Covance. The anti-Mre11 antibody was a gift from L. Symington. Pulsed-field gel electrophoresis and Southern blotting were performed as described previously (67). Chromosomes were separated on a contour-clamped homogeneous electric field (CHEF) apparatus (Bio-Rad) at 14°C for 22 h (run conditions were as follows: a 25-s pulse, 6 V/cm, a 120° angle). ³²P-labeled DNA fragments containing RDN5-2 within the ribosomal DNA were used as a probe to detect chromosome XII. Smaller degraded fragments are well separated from the large chromosome XII fragment.

Protein kinase Tel1 is activated after DSB induction.

We investigated whether Tel1 catalytic activity is increased after DSB induction (Fig. 1). Activation of the Tel1 pathway is controlled in a cell cycle-dependent manner (47). We therefore monitored the activity of Tel1 kinase in G₂/M-arrested cells. Cells expressing HA-tagged Tel1 protein were arrested at G₂/M with nocodazole and were subsequently treated with phleomycin. HA-Tel1 proteins were immunoprecipitated from cell extracts with anti-HA antibodies, and immunoprecipitates were subjected to an in vitro kinase assay. We used the GST fusion construct with the Rad53 C terminus (GST-Rad53) as a substrate for Tel1 kinase (47) (Fig. 1A). Mec1/Tel1-dependent phosphorylation at the C terminus of Rad53 is implicated in Rad53 activation (68). The phosphorylation of GST-Rad53 was detected with immunoprecipitates from untreated cells, but the level increased after phleomycin treatment. No phosphorylation was associated with the kinase-negative version of Tel1 protein (Tel1-KN) (Fig. 1A). Similar results were obtained with PHAS-1, a substrate commonly used for ATM/ATR family proteins (data not shown). Thus, Tel1 is activated in response to DSB induction. Mre11 was detected in the Tel1 immunoprecipitates, supporting the view that Tel1 interacts with the MRX complex through the C terminus of Xrs2 (26, 46). Notably, similar Mre11-Tel1 interactions were observed before and after phleomycin treatment. Likewise, chromosomal DNA was detected in the immunoprecipitates at similar levels before and after phleomycin treatment. Tel1 was activated 2-fold after exposure to 20 μg/ml phleomycin (Fig. 1A), although stronger activation was not observed at higher concentrations of phleomycin (Fig. 1B).

Tel1 localizes to DSB ends in a manner dependent on the C terminus of Xrs2 (46). The xrs2-11 mutation, which truncates the C terminus, confers defects in Tel1 recruitment to DSB ends but does not affect MRX localization (46). We next examined whether Tel1 kinase activation requires DSB localization (Fig. 1C). Although Tel1 from the xrs2-11 mutant exhibited basal kinase activity, no increase in kinase activity was detected after phleomycin treatment. Thus, activation of Tel1 kinase depends on its localization to DSB ends. Increased kinase activity might promote the accumulation of Tel1 at DSB ends. We then examined the effect of a kinase-negative tel1 (tel1-KN) mutation on the localization of Tel1 to DSB ends by a chromatin immunoprecipitation assay. Tel1-KN mutant proteins were found to associate with DSBs as efficiently as wild-type Tel1 proteins (Fig. 1D). Thus, protein kinase activity of Tel1 is not required for the localization of Tel1 to DSB ends. These results support the idea that Tel1 is activated at DNA ends by interacting with the MRX complex.

Effect of sae2Δ mutation on Tel1 activation after phleomycin treatment.

Deletions of SAE2 enhance activation of the Tel1 checkpoint pathway (71), but it is not known whether the sae2Δ mutation targets Tel1 kinase. We next examined the effect of the sae2Δ mutation on Tel1 kinase activation after DSB induction (Fig. 2A). The sae2Δ mutation did not affect Tel1 catalytic activity before phleomycin treatment but enhanced Tel1 activation after phleomycin treatment. Thus, Tel1 is more significantly activated after phleomycin treatment in sae2Δ mutants than in wild-type cells. The sae2Δ mutation did not affect the amount of Mre11 or chromosomal DNA in the immunoprecipitates (data not shown).

To trigger checkpoint signaling, Mec1 and Tel1 phosphorylate Rad9 in response to DNA damage (15, 74). Phosphorylated Rad9 interacts with the downstream kinase Rad53, thereby leading to Rad53 activation (53). Protein kinase activity of Rad53 is well correlated with its phosphorylation status (53). To better understand how Tel1 activates checkpoint signaling, we compared Rad9 and Rad53 phosphorylation in wild-type and sae2Δ mutant cells after phleomycin treatment (Fig. 2B). To detect Tel1-specific signaling, phosphorylation of Rad9 and Rad53 was monitored in a mec1Δ background. Rad9 phosphorylation was detected in mec1Δ mutants, although it was much weaker than that in wild-type cells. In contrast, Rad53 phosphorylation was undetectable in mec1Δ mutants. However, the introduction of a sae2Δ mutation dramatically enhanced both Rad9 and Rad53 phosphorylation in mec1Δ mutants after phleomycin treatment (71). These results support the idea that although Tel1 is activated after DSB induction, Tel1-mediated Rad9 phosphorylation is insufficient to trigger Rad53 activation in the presence of Sae2 protein.

Increased covalent attachment of histone to DNA in sae2Δ mutants after phleomycin treatment.

Sae2 exhibits a structure-specific nuclease activity in vitro (33) and plays a critical role in removing covalently bound Spo11 from DNA ends during meiosis (29, 49). We therefore considered the possibility that modifications at DNA ends result in enhanced Tel1 activation in sae2Δ mutants. Like phleomycin treatment, expression of the EcoRI endonuclease induces multiple DSBs in yeast (34). EcoRI is expected to generate clean DSB ends; EcoRI cleaves DNA at the recognition sequence, generating a cohesive DNA end with a 3′ OH and a 5′ phosphate (25). When phleomycin binds to DNA, it produces superoxide and hydroxide free radicals (64). Phleomycin thus may generate dirty DNA ends with various DNA modifications. We next investigated whether the sae2Δ mutation stimulates Tel1 catalytic activity after EcoRI expression (Fig. 2C). Tel1 kinase was activated after EcoRI expression, as found for phleomycin treatment. However, the sae2Δ mutation did not enhance Tel1 activation after EcoRI expression. Tel1 activation did not reach a plateau after EcoRI expression, since subsequent phleomycin treatment enhanced Tel1 activation in sae2Δ mutants (Fig. 2C). It seems unlikely that weak Tel1 activation after EcoRI expression resulted from the generation of few DSB lesions, since similar levels of chromosome degradation were induced after EcoRI expression and phleomycin treatment (Fig. 2D). These observations raise the possibility that phleomycin, but not EcoRI, generates DNA modifications at DSB ends in the absence of Sae2, thereby stimulating Tel1 catalytic activity.

Histones and other proteins are cross-linked to DNA upon exposure of cells to genotoxic agents (4, 5). We examined whether histone is covalently attached to chromosomal DNA in wild-type cells or sae2Δ mutants (Fig. 2E). Cells were arrested at G₂/M with nocodazole and were either treated with phleomycin or mock treated as described above. Chromosomal DNAs were then purified through isopycnic centrifugation and were subjected to immunoblot analysis with an antibody against histone H2B or H3. Histones H2B and H3 were detected in the DNA fraction from sae2Δ mutants after phleomycin treatment, whereas no association was observed in wild-type cells. No histone accumulation was detected in sae2Δ mutants after EcoRI expression (Fig. 2F).

The MRX complex is also required for the removal of covalently bound Spo11 from DNA ends during meiosis (29, 49). We further addressed whether a Mre11 nuclease deficiency increases histone cross-linking and stimulates Tel1 activation after phleomycin treatment. The mre11-3 mutation carries substitutions in the phosphoesterase signature motif (7) and eliminates the nuclease activity of the MRX complex (33, 51). No covalent DNA-histone H3 interaction was detected in mre11-3 mutants after phleomycin treatment (Fig. 2E). Moreover, the mre11-3 mutation did not affect Tel1 activation after phleomycin treatment; immunoprecipitated Tel1 proteins from mre11-3 cells exhibited a level of kinase activity similar to that from wild-type cells (Fig. 2G). These results suggest that covalent protein attachment to the chromosome influences Tel1 activation.

Activation of protein kinase Tel1 after CPT treatment.

Although histone is covalently attached to DNA after DNA damage, the chemical nature of the histone-DNA complex is not known. We therefore used camptothecin (CPT), which generates DSBs by inhibiting topoisomerase I (Top1). Top1 acts to remove the superhelical tension that could accumulate during DNA replication by transiently cleaving and religating a single strand of DNA (75, 76). This reaction involves the formation of a covalent bond between a tyrosine residue on Top1 and the 3′-phosphoryl end of the single-strand break. CPT stabilizes reversible Top1-DNA covalent complexes (35, 54). Since DSB induction after CPT treatment depends on DNA replication, it has been proposed that Top1-DNA intermediates at single-strand breaks can be converted to Top1-bonded DSBs if replication fork movement is blocked (2, 27, 66, 70).

Cells carrying a sae2Δ mutation are sensitive to CPT treatment (12) (Fig. 3A). We first examined whether Sae2 is required for the removal of covalently bound Top1 from the chromosome (Fig. 3B). Cells expressing FLAG-tagged Top1 were either treated with CPT or mock treated in S phase, and chromosomal DNA was analyzed as described above to detect covalently attached Top1. Top1 was detected in the DNA fraction specifically after CPT treatment, and the amount of Top1 was greater in sae2Δ mutants than in wild-type cells. We next compared Tel1 catalytic activities in wild-type and sae2Δ mutant cells after exposure to CPT in S phase (Fig. 3C and D). Cells expressing HA-Tel1 were treated with various concentrations of CPT and were subjected to the in vitro kinase assay. CPT treatment activated Tel1 in sae2Δ mutants more strongly than in wild-type cells. Repair of CPT-mediated DNA damage also depends on Mre11 nuclease activity (12). Cells carrying the mre11-3 mutation were, like sae2Δ mutants, sensitive to CPT treatment (Fig. 3A). The mre11-3 mutation was also found to confer defects in Top1 removal (Fig. 3B) and to enhance Tel1 activation after CPT treatment (Fig. 3C). Similar Mre11-Tel1 interactions were observed in wild-type cells, sae2Δ mutants, and mre11-3 mutants before and after CPT treatment (Fig. 3C). Likewise, chromosomal DNA was detected at similar levels in those cells.

We further investigated the effect of the mre11-3 or sae2Δ mutation on the activation of the Tel1 checkpoint pathway after CPT treatment. To this end, we monitored Rad53 phosphorylation in mec1Δ sae2Δ or mec1Δ mre11-3 mutants after CPT treatment as described above (Fig. 3D; see also Fig. 2B). The introduction of the sae2Δ or mre11-3 mutation increased Rad53 phosphorylation in mec1Δ mutants after CPT treatment, although the sae2Δ mutation exhibited a stronger effect than the mre11-3 mutation. Thus, increased Tel1 activation in mre11-3 or sae2Δ mutants after CPT treatment correlates with accumulation of Top1-DNA conjugates on chromosomes.

CPT-induced DNA damage repair involves several other proteins, including Tdp1 and Rad1. The Tdp1 pathway has been established as one mechanism for the removal of Top1 from DNA (56), but Tdp1 plays a role highly redundant with those of other pathways, for example, the Rad1 pathway (12, 40, 73). Notably, Tdp1 and Rad1 act differently from Sae2; the rad1Δ sae2Δ tdp1Δ triple mutant exhibits stronger sensitivity to CPT than either the rad1Δ tdp1Δ double mutant or the sae2Δ single mutant (12). We thus examined the effect of the rad1Δ tdp1Δ double mutation on the accumulation of DNA-Top1 intermediates and the activation of protein kinase Tel1 after CPT treatment. Like the sae2Δ mutation, the rad1Δ tdp1Δ double mutation increased the amount of chromosome-bound Top1 (Fig. 3E). Tel1 activation was more effective in rad1Δ tdp1Δ mutants than in wild-type cells at lower doses of CPT but not as robust as in sae2Δ mutants (Fig. 3F). Thus, the accumulation of DNA-Top1 intermediates on chromosomes does not necessarily increase Tel1 catalytic activity.

Activation of protein kinase Tel1 by incubation of MRX and DNA fragments in vitro.

Because MRX and Sae2 are involved in DNA end processing, it is possible that DNA-Top1 intermediates accumulate at DSB ends and stimulate Tel1 activation in mre11-3 or sae2Δ mutants. However, DNA-Top1 intermediates could remain unprocessed at the 3′ ends of single-strand breaks on chromosomes in these mutants. To test the hypothesis that protein binding near DNA ends enhances Tel1 activation, we reconstituted MRX-mediated Tel1 activation in vitro. We first set up an in vitro kinase assay with purified Tel1 and MRX (Fig. 4). We tested whether incubation of MRX or DNA activates protein kinase Tel1 by using GST-Rad53 as a substrate. Phosphorylation of GST-Rad53 was detected with Tel1 in the absence of MRX or DNA fragments. However, phosphorylation was increased when Tel1 was incubated with both the MRX complex and DNA fragments (Fig. 4A). Phosphorylation was dependent on Tel1 catalytic activity, as evidenced by the fact that no phosphorylation was observed with Tel1-KN. DNA fragments were not fully degraded under the reaction conditions, although the MRX complex possesses nuclease activities (data not shown; see below). We also addressed whether MRX-mediated Tel1 activation depends on the presence of DNA ends. Tel1 activation was observed by incubation with SmaI-treated, but not undigested, pUC18 plasmids (Fig. 4B). Thus, incubation with MRX and DNA fragments increases Tel1 catalytic activity, consistent with the view that Tel1 is activated at DNA ends through interaction with the MRX complex.

Effects of DNA end modification on protein kinase activity of Tel1 in vitro.

We next examined whether protein binding at DNA ends increases Tel1 catalytic activity (Fig. 5). To modify DNA ends, we tethered anti-biotin Fab fragments at the 5′- or 3′-biotinylated DNA ends (Fig. 5A). For a control experiment, we also prepared DNA fragments with Fab fragments embedded in the middle (Fig. 5A). Binding of Fab fragments to biotinylated DNA fragments was confirmed by a DNA mobility shift assay (data not shown). Covalent protein attachment to the chromosome has been considered to inhibit DNA repair processes (4). We first examined whether Fab tethering attenuates MRX-dependent DNA degradation. MRX degraded unmodified DNA fragments, consistent with the previous observation that Mre11 possesses nuclease activities (69). Fab tethering at either the 5′ or the 3′ end decreased DNA degradation, whereas the placement of Fab fragments at the middle of the DNA fragment did not (Fig. 5B and C). Biotinylation by itself did not affect DNA degradation (data not shown). It was possible that Fab fragments at DNA ends interfered with DNA binding of the MRX complex, since Fab tethering decreased MRX-dependent DNA degradation. To determine whether Fab tethering alters the binding of MRX to DNA fragments, we monitored the DNA binding of Xrs2 by immunoprecipitation and subsequent PCR amplification (Fig. 5D). Xrs2 was detected at similar levels on modified and nonmodified DNA fragments. Thus, Fab fragments at either end attenuate MRX-dependent DNA degradation without interfering with DNA binding of the MRX complex. We next examined the effect of Fab fragments at the 5′ or 3′ end on Tel1 catalytic activity (Fig. 5E). Tel1 was incubated with MRX proteins and modified DNA fragments and was then subjected to the in vitro kinase assay as described above. Fab binding at either the 5′ or the 3′ end increased the phosphorylation of GST-Rad53, whereas Fab tethering at the middle of a DNA fragment did not. Enhanced phosphorylation was fully dependent on the MRX complex, as evidenced by the fact that Fab-dependent Tel1 activation was not observed without incubation with the MRX complex. No phosphorylation was detected with kinase-negative Tel1-KN mutants. These results indicate that Fab tethering at DNA ends enhances MRX-mediated Tel1 kinase activation.

As MRX degrades DNA fragments, the DNA fragments become shorter. It was therefore possible that longer DNA fragments activated protein kinase Tel1 more efficiently. We compared Tel1 catalytic activities after incubation of MRX with a 150-bp or an 800-bp DNA fragment (Fig. 6A and B). Even though MRX degraded both DNA fragments, the reaction mixture with the 800-bp DNA fragment contained longer DNA fragments than that with the 150-bp fragment (Fig. 6A). Under this reaction condition, however, Tel1 was activated similarly by incubation with these fragments (Fig. 6B), suggesting that the length of DNA fragments does not significantly affect Tel1 catalytic activity. This is consistent with the observation that Tel1 activation depends fully on the presence of DNA ends (Fig. 4B). Since Fab tethering at DNA ends decreases DNA degradation, defective DNA end processing might increase the activation of Tel1 kinase. To test this possibility, we examined the effects of the nuclease-defective mre11-3 mutation on Tel1 activation using unmodified DNA fragments. The MRX complex containing Mre11-3 mutant protein (MRX^mre11-3) was defective in degrading unmodified DNA fragments (Fig. 6C). However, the MRX^mre11-3 complex behaved like the wild-type MRX complex with regard to Tel1 activation; incubation with the MRX^mre11-3 complex did not stimulate Tel1 catalytic activity at unmodified DNA ends (Fig. 6D). Moreover, incubation with wild-type MRX or the mutant MRX^mre11-3 complex similarly enhanced Tel1 activation at Fab-tethered DNA ends (Fig. 6D). These results indicate that the nuclease activity of the MRX complex is dispensable for Tel1 activation. It thus seems unlikely that a stalled MRX complex at DNA ends by itself enhances the activation of protein kinase Tel1.

Effect of Fab removal and DNA degradation on Tel1 kinase activity in vitro.

Since MRX can, albeit slowly, degrade Fab-tethered DNA fragments (Fig. 5B), we addressed whether DNA end processing affects the activation status of Tel1 (Fig. 7). To monitor Tel1 catalytic activity in the course of DNA degradation, we performed a modified two-step in vitro assay using Fab-tethered or nontethered DNA fragments. In this assay, Tel1 was first incubated with MRX and DNA fragments, and GST-Rad53 was added later. In parallel, we also examined the DNA status during the in vitro reaction. DNA degradation of nontethered DNA fragments was readily detected at 10 min after incubation with MRX and Tel1, and no DNA fragments of the original size were observed at 30 min (Fig. 7A). In contrast, degradation of Fab-DNA fragments was detected at 30 min. Activation of protein kinase Tel1 with nontethered DNA fragments was observed at early time points, but kinase activity gradually dropped to the basal level (Fig. 7B). As shown above, Fab tethering enhanced Tel1 activation. However, enhanced activation was observed only at early time points (up to 30 min). Once DNA degradation occurred at 90 min, kinase activity dropped to the basal level. Decreased kinase activity does not result from denaturation of Tel1 protein, since kinase activation remained if the reaction mixture contained the nuclease-defective MRX^mre11-3 complex. No DNA degradation was observed with the MRX^mre11-3 complex in the course of this experiment (Fig. 7A). These results suggest that Tel1 exhibits enhanced kinase activity only when a DNA-protein complex exists at DNA ends and also that Tel1 becomes inactivated immediately after the completion of DNA end processing. Together, our findings support the model in which the MRX complex recognizes protein-linked DNA ends and stimulates Tel1 activation.