MRE11 Function in Response to Topoisomerase Poisons Is Independent of its Function in Double-Strand Break Repair in Saccharomyces cerevisiae

Camptothecin (CPT) and etoposide (ETP) trap topoisomerase-DNA covalent intermediates, resulting in formation of DNA damage that can be cytotoxic if unrepaired. CPT and ETP are prototypes for molecules widely used in chemotherapy of cancer, so defining the mechanisms for repair of damage induced by treatment with these compounds is of great interest. In S. cerevisiae, deficiency in MRE11, which encodes a highly conserved factor, greatly enhances sensitivity to treatment with CPT or ETP. This has been thought to reflect the importance of double-strand break (DSB) repair pathways in the response to these to agents. Here we report that an S. cerevisiae strain expressing the mre11-H59A allele, mutant at a conserved active site histidine, is sensitive to hydroxyurea and also to ionizing radiation, which induces DSBs, but not to CPT or ETP. We show that TDP1, which encodes a tyrosyl-DNA phosphodiesterase activity able to release both 5′- and 3′-covalent topoisomerase-DNA complexes in vitro, contributes to ETP-resistance but not CPT-resistance in the mre11-H59A background. We further show that CPT- and ETP-resistance mediated by MRE11 is independent of SAE2, and thus independent of the coordinated functions of MRE11 and SAE2 in homology-directed repair and removal of Spo11 from DNA ends in meiosis. These results identify a function for MRE11 in the response to topoisomerase poisons that is distinct from its functions in DSB repair or meiotic DNA processing. They also establish that cellular proficiency in repair of DSBs may not correlate with resistance to topoisomerase poisons, a finding with potential implications for stratification of tumors with specific DNA repair deficiencies for treatment with these compounds.


Introduction
Topoisomerase poisons are potent drugs for treatment of cancer. Two naturally occurring topoisomerase inhibitors, camptothecin (CPT) and etoposide (ETP), are prototypes for this class of chemotherapeutics, which target Topoisomerase I (Topo I) and Topoisomerase II (Topo II) respectively. Topo I regulates DNA superhelicity ahead of transcription and replication forks by inserting itself into one strand of the DNA backbone, forming a 39covalent bond between a tyrosine residue on the enzyme and the nicked DNA strand and enabling rotation of the intact strand around the nick. CPT binds at the single-strand break 39 of the Topo I-DNA complex, stabilizing the covalent Topo I-DNA intermediate and inhibiting religation of the DNA nick. Topo II promotes decatenation of DNA following replication, inserting into both strands of the duplex to form a double-strand break (DSB), with 59-DNA ends tethered to tyrosine residues on Topo II subunits, enabling intact DNA to pass through the DSB. ETP intercalates at Topo II insertion sites, stabilizing the DSB intermediate. Defining the mechanisms of repair of damage induced by CPT and ETP is of considerable practical importance, because both are potent cytotoxic agents and their derivatives are commonly used in cancer chemotherapy.
Mutants of S. cerevisiae deficient in MRE11 (mre11D) are very sensitive to CPT [1,2,3] and to ETP [4]. MRE11 encodes a multifunctional nuclease, active as a 39-59 dsDNA exonuclease, a single-strand DNA endonuclease and an AP lyase in vitro, and shown to function in DNA repair, meiotic recombination, telomere maintenance and immunoglobulin gene diversification [5,6,7,8,9,10,11]. Its mechanism of function in response to CPT or ETP is not understood, but has been thought to correlate with activity in DSB repair, which is dependent upon its 39-59 exonuclease activity.
Here we describe a new S. cerevisiae mutant allele, mre11-H59A. We show that an S. cerevisiae mre11-H59A strain is as sensitive to HU and IR as the well-characterized DSB repair-deficient mre11-H125N strain [3,10,12,13]. However, in contrast to the mre11-H125N strain, the mre11-H59A strain is resistant to CPT. CPTresistance does not depend upon TDP1, which encodes a factor that releases covalent 39-or 59-tyrosyl DNA bonds in vitro [14,15,16,17]. The mre11-H59A and mre11-H125N strains are both resistant to ETP, which is toxic to the mre11D strain; but deficiency in TDP1 (tdp1D) causes these strains to become ETPsensitive. Neither CPT-nor ETP-resistance of the mre11-H59A strain depends upon SAE2, which is required for MRE11dependent removal of Spo11 from the ends of meiotic DNA [18] and regulates functions of MRE11 in homologous recombi-nation [19]. Thus MRE11 has distinct functions in repair of damage induced by topoisomerase poisons and repair of mitotic and meiotic DSBs.

S. cerevisiae mre11-H59A Is Sensitive to Hydroxyurea but Not CPT
To distinguish functions of MRE11 in DSB repair and response to DNA damage by topoisomerase poisons, two previous findings drew our attention to the conserved active site histidine at residue 59 (H59) in phosphodiesterase motif II of S. cerevisiae Mre11 (Fig. 1A). Biochemical analysis had shown that purified recombinant Pyrococcus furiosus Mre11 with a mutation at this site (H52S) is deficient in exonuclease activity; while genetic analysis had shown that mutation at the corresponding position of the Schizosaccharomyces pombe Mre11 homologue (rad32-H68S) did not render cells sensitive to CPT, HU or IR [8].
In order to study the effect of mutation of H59 in S. cerevisiae and compare this allele to well-characterized mutant alleles, we generated a panel of strains bearing mutations in conserved residues in the Mre11 active site phosphodiesterase motifs (Fig. 1A). S. cerevisiae Jel1 mre11D (LSY1706, MATa leu2 trp1 ura3-52 prb1-1122 pep4-3 his3::GAL10-GAL4 mre11::His3MX6; [12]) was stably transformed with pYES 2 m vectors expressing C-terminal TAPtagged Mre11 or mutant Mre11-H59A, Mre11-D56A, Mre11-D56N, Mre11-H125N or Mre11-H213Y under control of the GAL1 promoter, and gene and protein expression were confirmed by RT-PCR (not shown) and western blotting (Fig. S1). Mutations at D56 alter a conserved glutamate in close proximity to H59 in phosphodiesterase motif II (Fig. 1A), known to be important for resistance to HU and IR [12]. Mutation of H125 alters a residue in phosphodiesterase motif III that is invariant among species and necessary for stabilizing the transition state during nucleolysis [8,20,21]. S. cerevisiae mre11-H125N mutants have previously been shown to be sensitive to DNA damaging agents including CPT and IR [3,10,12,13]; and a corresponding S. pombe active site mutant (rad32-H134S) is associated with both CPT-sensitivity and IR-sensitivity [8]. The mre11-H213Y allele, one of the first MRE11 mutants studied, alters an invariant histidine in motif IV. It causes deficiency in both meiotic and mitotic DSB repair [22], and the corresponding mutant Mre11 protein fails to form a stable complex with Rad50 [12,23,24].
Sensitivity to HU and CPT was determined by spot dilution assays. Serial 10-fold dilutions of an overnight culture of each strain were spotted onto YPD plates containing various concentrations of HU or CPT. The Jel1 mre11D strain and its derivative expressing mre11-H213Y were extremely sensitive to HU; the derivatives expressing mre11-H59A and mre11-H125N were no more sensitive than the derivative expressing MRE11; and the derivatives expressing mre11-D56A and mre11-D56N were relatively resistant (Fig. 1B, left). Spot tests also showed that Jel1 mre11D and its derivative expressing mre11-H213Y were extremely sensitive to CPT, the derivatives expressing mre11-D56A, mre11-D56N and mre11-H125N slightly less sensitive, and the derivative expressing mre11-H59A relatively resistant, although slightly less so than the derivative expressing MRE11 (Fig. 1B, right). Identical results were obtained using the low copy number CEN plasmid p416ADH expressing MRE11, mre11-H59A or mre11-H213Y alleles (data not shown). Thus, the mre11-H59A mutation caused sensitivity to HU but not to CPT, as measured by spot dilution assays.
The contrasting effects of the mre11-H59A mutation on HU and CPT sensitivity were confirmed by clonogenic survival assays. These assays showed that the strain expressing mre11-H59A was sensitive to HU, less so than the extremely HU-sensitive Jel1 mre11D strain or its derivative expressing mre11-H213Y, but more sensitive than the parental Jel1 strain or the Jel1 mre11D derivatives expressing MRE11, mre11-D56A, mre11-D56N or mre11-H125N (Fig. 1C, left). Nonetheless, the Jel1 mre11D derivative expressing mre11-H59A was as resistant to CPT as the Jel1 parental line or the Jel1 mre11D derivative expressing MRE11 (Fig. 1C, right).
The S. cerevisiae mre11-H59A Strain Exhibits Intermediate Sensitivity to IR To test sensitivity to IR, we focused on a panel of six strains, exposed cells in liquid culture to radioactive cesium, plated aliquots in triplicate, determined colony number after 3-4 days growth and normalized to the unexposed samples. The mre11-H59A strain was of intermediate sensitivity to IR, less resistant than the Jel1 parental strain or the Jel1 mre11D derivative expressing MRE11, but comparable to the derivatives expressing mre11-D56N, mre11-D56A or mre11-H125N, and more resistant than Jel1 mre11D or its derivative expressing mre11-H213Y (Fig. 2).

TDP1 Does Not Contribute to CPT-Resistance of MRE11-Deficient Strains
The enzyme Tyrosyl-DNA-phosphodiesterase 1 (Tdp1) was identified as an activity in S. cerevisiae extracts capable of removing a trapped Topo I-DNA complex in vitro [14]. However, S. cerevisiae tdp1D mutants exhibit only minor increases in sensitivity to CPT, either in strains expressing MRE11 or in the MRE11-deficient strains tested thus far, including mre11-H125N [1,3,17,25]. To compare function of TDP1 and MRE11 in CPT resistance, we created Jel1 tdp1D and Jel1 mre11D tdp1D derivatives, and transformed the latter with 2 m plasmids expressing MRE11 or mre11-H59A, mre11-H125N and mre11-H213Y mutant alleles. In spot dilution assays, the Jel1 mre11D strain was much more severely CPT-sensitive than the Jel1 tdp1D strain, and sensitivity was not enhanced in the Jel1 mre11D tdp1D double mutant (Fig. 3). Moreover, the Jel1 tdp1D mre11D derivatives expressing MRE11, mre11-H59A, mre11-H125N or mre11-H213Y exhibited CPTsensitivity essentially indistinguishable from the corresponding Jel1 mre11D derivatives (Fig. 3, compare lower and upper). Thus, MRE11 was more critical to CPT-resistance than TDP1, and deficiency in TDP1 did not affect CPT sensitivity of MRE11deficient strains.
TDP1 Contributes to ETP-Resistance of mre11-H59A and mre11-H125N Mutants In S. cerevisiae, MRE11-deficiency causes sensitivity to ETP, which traps covalent complexes formed by Topo II with DNA 59ends [4]. We therefore tested ETP sensitivity of a panel of MRE11deficient strains in a spot dilution assay. We found that the Jel1 mre11D strain and its derivative expressing mre11-H213Y were extremely sensitive to ETP, while derivatives expressing mre11-H59A or mre11-H125N were as resistant as Jel1 or the Jel1 mre11D derivative expressing MRE11 (Fig. 4, above). Thus, expression of mre11-H59A conferred resistant to ETP and CPT, despite marked sensitivity to HU and IR; and expression of mre11-H125N strain conferred resistant to ETP, despite sensitivity to CPT, HU and IR.
Tdp1 can remove not only Topo I-39-DNA covalent complexes in vitro, but also Topo II-59-DNA covalent complexes [17]. We therefore asked how ETP-resistance was affected by deficiencies in TDP1 and MRE11. In spot dilution assays, Jel1 and Jel1 tdp1D were comparably ETP-sensitive, while the Jel1 mre11D strain was much more severely ETP-sensitive than the Jel1 tdp1D strain (Fig. 4, compare lower and upper). Thus, deficiency in MRE11 caused much more severe ETP-sensitivity than deficiency in TDP1. Moreover, while deficiency in TDP1 did not affect ETPsensitivity of the Jel1 mre11D derivative expressing MRE11, it did cause a modest increase in ETP-sensitivity of the Jel1 mre11D derivative expressing mre11-H59A, and a more severe increase in sensitivity of the derivative expressing mre11-H125N (Fig. 4, lower).

SAE2 Makes a Modest Contribution to CPT-or ETP-Resistance
SAE2 encodes a repair endonuclease that is essential for Mre11 function in removal of Spo11 from 59-ends of DNA in meiosis, and that regulates activities of Mre11 in homologous recombination [18,19,26]. Deficiency in SAE2 has been reported not to affect CPT-or ETP-sensitivity in S. cerevisiae [3,18], although the SAE2 ortholog, CtIP, is essential for CPT-resistance in S. pombe [27]. Because of the role of SAE2 in meiosis, it seemed important to test the effect of deficiency in SAE2 on CPT-and ETP-sensitivity of the Jel1 mre11D derivative expressing mre11-H59A. We therefore generated Jel1 sae2D and Jel1 sae2D mre11D strains, and Jel1 sae2D mre11D derivatives expressing MRE11, mre11-H59A or mre11-H213Y alleles (the latter was included as a control due to its severe meiotic defect). Consistent with the role of SAE2 in DSB repair, strains deficient in SAE2, including Jel1 sae2D, Jel1 sae2D mre11D, and derivatives of the latter expressing MRE11, mre11-H59A, or mre11-H213Y, all exhibited greater HU sensitivity than the parental SAE2 strains (Fig. 5A). However, deficiency in SAE2 caused only a very modest increase in CPT-sensitivity (Fig. 5B) or ETP-sensitivity (Fig. 5C). Thus, SAE2 is not critical to MRE11dependent repair of damage induced by CPT or ETP.

Discussion
We have shown that the S. cerevisiae mre11-H59A strain is resistant to topoiosomerase poisons CPT and ETP, which create protein-DNA covalent complexes; but sensitive to HU and sensitive to IR, which induces DSBs. These results establish that the MRE11 function in DSB repair can be separated from its function in repair of CPT-or ETP-induced DNA damage. They also establish that a cell need not be proficient in repair of DSBs to resist topoisomerase poisons.
Our results identify several clear distinctions between S. cerevisiae and S. pombe in the response to topoisomerase poisons. One  difference is in functions associated with specific residues of the highly conserved active site of the Mre11 polypeptide. The S. cerevisiae mre11-H59A strain that we have studied carries a mutation at a conserved active site histidine in phosphodiesterase motif II of the active site and is resistant to CPT but sensitive to HU and IR. An S. pombe strain carrying a mutation at the corresponding position, rad32(mre11)-H68S, is resistant to CPT and also to HU and IR [8].
The repair pathways for CPT-and ETP-induced damage in S. cerevisiae and S. pombe also exhibit distinct dependence upon other factors. In S. cerevisiae, TDP1-deficiency has little effect on CPTsensitivity, either in strains expressing wild-type MRE11 [1,25] or in the four mre11 mutant strains we examined. This contrasts with S. pombe, where TDP1-deficiency has a pronounced effect on CPTsensitivity [28]. In S. cerevisiae, SAE2 makes only a very modest contribution to CPT-or ETP-resistance. This also contrasts with S. pombe, where deficiency in the SAE2 ortholog, CtIP, impairs the ETP response, but surprisingly promotes the CPT response [27].
How might MRE11 function to prevent toxicity by topoisomerase poisons independent of DSB repair? A critical step that distinguishes the response to topoisomerase poisons from other cytotoxic treatments is removal of the covalent protein-DNA complexes that accumulate in cells treated with these compounds. Our results raise the possibility that MRE11 may promote cleavage of tyrosyl-DNA bonds, either directly or indirectly. Several factors possess this activity, including the conserved enzyme Tdp1, which can cleave both 39-and 59-covalent tyrosyl-DNA bonds in vitro [14,15,16,17]; the structure-specific nucleases Rad1/Rad10 and Mms4/Mus81, which appear to function redundantly with Tdp1 in release of 39-covalent protein DNA complexes in vivo [29]; and the recently discovered human factor, Tdp2, which can cleave 59tyrosyl-DNA covalent bonds in vitro and rescue CPT-sensitivity of S. cerevisiae tdp1D rad1D mutants in vivo [30]. Consistent with a role for MRE11 in promoting cleavage of tyrosyl-DNA covalent bonds, topoisomerase-DNA complexes have been shown to persist in an S. pombe rad32(mre11)-D65N strain following treatment with CPT or ETP [27].
Determining the mechanism of MRE11 function in the response to topoisomerase poisons has important implications for the clinical setting. A subset of human colorectal cancers is MRE11deficient [31,32]. If MRE11 functions to promote release of covalent topoisomerase-DNA complexes independent of its role in DSB repair, MRE11-deficient tumors may be priority candidates for treatment with CPT and ETP derivatives. This could have implications for stratification of tumors with specific DNA repair deficiencies for treatment with topoisomerase poisons.

HU, CPT, ETP and IR Sensitivity Assays
Published procedures were used to assay sensitivity to HU [12], CPT [1], IR [12], and ETP [4]. In brief, spot dilution assays were performed by plating 3 ml of serial 10-fold dilutions of a liquid culture, containing from 10 3 -10 6 cells/ml, on YPD plates with or without drug. Clonogenic survival assays for CPT-sensitivity were carried out on cells cultured in liquid medium containing 1-20 mM CPT for 6 hr (about 5-8 cell divisions). Aliquots were removed and plated on YPD in triplicate, colonies counted 3-4 days later and percent survival normalized to an untreated control. For HU clonogenic survival assays, cells from liquid culture were plated in triplicate on YPD plates containing various concentrations of HU, colonies counted 3-4 days later and percent survival normalized to an untreated control. Sensitivity to IR was measured by exposing cells in liquid culture to radioactive 137 Cs at the indicated dosage, plating on YPD and counting surviving colonies 3-4 days later. A JL-Shepherd Model 81-14 Cesium-137 Irradiator was used for all in vitro radiation treatments in this study. Cell suspensions in 1.5 ml tubes were exposed to 137 Cs c-rays at a dose rate of 3 Gy/min. Dosimetry was performed before every radiation treatment to ensure consistency in the dose rate. For survival curves, representative plots are shown for at least three independent experiments with similar results; error is calculated for triplicates within the experiment; and survival was normalized to untreated controls.

Supporting Information Legends
Figure S1 Expression of Mre11 protein. Western blot analysis of Mre11 protein levels in indicated strains (designated as described in Fig. 1), normalized to b-actin. Mre11 expressed from 2 m vectors carried a C-terminal TAP-tag, and was therefore slightly larger than the endogenous protein.