Tyrosyl-DNA Phosphodiesterase I Catalytic Mutants Reveal an Alternative Nucleophile That Can Catalyze Substrate Cleavage*

Background: Tdp1 repair of adducted DNA involves a covalent enzyme-DNA intermediate, formed and resolved by Hisnuc and Hisgab residues. Stabilized Hisgab mutant-DNA complexes are cytotoxic. Results: DNA adduct cleavage by an adjacent His residue in HisnucAla mutants impair enzyme-DNA intermediate resolution. Conclusion: Alterations in active site geometry enhance the stability of cytotoxic Tdp1-DNA intermediates. Significance: These findings provide the rationale for developing chemotherapeutics that poison Tdp1. Tyrosyl-DNA phosphodiesterase I (Tdp1) catalyzes the repair of 3′-DNA adducts, such as the 3′-phosphotyrosyl linkage of DNA topoisomerase I to DNA. Tdp1 contains two conserved catalytic histidines: a nucleophilic His (Hisnuc) that attacks DNA adducts to form a covalent 3′-phosphohistidyl intermediate and a general acid/base His (Hisgab), which resolves the Tdp1-DNA linkage. A Hisnuc to Ala mutant protein is reportedly inactive, whereas the autosomal recessive neurodegenerative disease SCAN1 has been attributed to the enhanced stability of the Tdp1-DNA intermediate induced by mutation of Hisgab to Arg. However, here we report that expression of the yeast HisnucAla (H182A) mutant actually induced topoisomerase I-dependent cytotoxicity and further enhanced the cytotoxicity of Tdp1 Hisgab mutants, including H432N and the SCAN1-related H432R. Moreover, the HisnucAla mutant was catalytically active in vitro, albeit at levels 85-fold less than that observed with wild type Tdp1. In contrast, the HisnucPhe mutant was catalytically inactive and suppressed Hisgab mutant-induced toxicity. These data suggest that the activity of another nucleophile when Hisnuc is replaced with residues containing a small side chain (Ala, Asn, and Gln), but not with a bulky side chain. Indeed, genetic, biochemical, and mass spectrometry analyses show that a highly conserved His, immediately N-terminal to Hisnuc, can act as a nucleophile to catalyze the formation of a covalent Tdp1-DNA intermediate. These findings suggest that the flexibility of Tdp1 active site residues may impair the resolution of mutant Tdp1 covalent phosphohistidyl intermediates and provide the rationale for developing chemotherapeutics that stabilize the covalent Tdp1-DNA intermediate.


Tyrosyl-DNA phosphodiesterase I (Tdp1) catalyzes the repair of 3-DNA adducts, such as the 3-phosphotyrosyl linkage of DNA topoisomerase I to DNA. Tdp1 contains two conserved catalytic histidines: a nucleophilic His (His nuc ) that attacks DNA adducts to form a covalent 3-phosphohistidyl intermediate and a general acid/base His (His gab ), which resolves the Tdp1-DNA linkage. A His nuc to Ala mutant protein is reportedly inactive,
whereas the autosomal recessive neurodegenerative disease SCAN1 has been attributed to the enhanced stability of the Tdp1-DNA intermediate induced by mutation of His gab to Arg. However, here we report that expression of the yeast His nuc Ala (H182A) mutant actually induced topoisomerase I-dependent cytotoxicity and further enhanced the cytotoxicity of Tdp1 His gab mutants, including H432N and the SCAN1-related H432R. Moreover, the His nuc Ala mutant was catalytically active in vitro, albeit at levels 85-fold less than that observed with wild type Tdp1. In contrast, the His nuc Phe mutant was catalytically inactive and suppressed His gab mutant-induced toxicity. These data suggest that the activity of another nucleophile when His nuc is replaced with residues containing a small side chain (Ala, Asn, and Gln), but not with a bulky side chain. Indeed, genetic, biochemical, and mass spectrometry analyses show that a highly conserved His, immediately N-terminal to His nuc , can act as a nucleophile to catalyze the formation of a covalent Tdp1-DNA intermediate. These findings suggest that the flexibility of Tdp1 active site residues may impair the resolution of mutant Tdp1 covalent phosphohistidyl intermediates and provide the ration-ale for developing chemotherapeutics that stabilize the covalent Tdp1-DNA intermediate.
Tyrosyl-DNA phosphodiesterase I (Tdp1) 3 is a eukaryotic DNA repair enzyme that can remove a wide variety of 3Ј-phospho-DNA adducts. The 3Ј-phosphotyrosyl linkage of DNA topoisomerase I (Top1) to a single DNA end was first identified as a Tdp1 substrate (1). The covalent enzyme-DNA complex is a requisite intermediate in the catalytic cycle of Top1 and constitutes the sole cellular target of the camptothecin (CPT) class of cancer therapeutics. CPTs, which include the Food and Drug Administration-approved analogs topotecan and irinotecan, reversibly stabilize the covalent Top1-DNA complex to induce S phase-specific lethal DNA lesions (2)(3)(4)(5)(6). More recently, Tdp1 has been shown to process a wide range of other DNA adducts, including 3Ј-modified nucleotides, such as phosphoglycolates and chain-terminating nucleotides, and a 3Ј-phosphohistidyl intermediate formed during Tdp1 catalysis (7)(8)(9)(10)(11)(12)(13).
As a member of the phospholipase D superfamily, Tdp1 contains conserved histidine-lysine motifs (HXK) that provide the two catalytic histidine residues for hydrolysis of its substrates (see Fig. 1A) (14 -17). The more N-terminal catalytic histidine (herein referred to as His nuc , which is His 182 in yeast and His 263 human Tdp1) functions in the nucleophilic attack on the 3Ј-phosphoryl-adduct linkage to form a 3Ј-phosphohistidyl bond between Tdp1 and DNA. A more C-terminal catalytic histidine (referred to as His gab , which in yeast is His 432 and in human His 493 ) then functions as a general acid/base to activate a water molecule to hydrolyze the Tdp1-DNA linkage, dissociating Tdp1 from the DNA (see Fig. 1B). Tdp1 leaves behind a DNA nick with 3Ј-phosphoryl and 5Ј-OH ends, which must be processed by polynucleotide kinase/phosphatase for DNA ligation (18 -21). The catalytic mechanism of Tdp1 and catalytic pocket architecture are highly conserved between yeast and human Tdp1 (see Fig. 1, A and B) (7, 12, 15-17, 22, 23).
A catalytic His gab to Arg substitution in human Tdp1 (H493R) was identified in patients with the rare autosomal recessive neurodegenerative disease spinocerebellar ataxia with axonal neuropathy (SCAN1) (13). Replacing His gab with Arg impairs the resolution of the Tdp1-DNA phosphohistidyl linkage, resulting in a longer-lived covalent Tdp1-DNA reaction intermediate in vitro (10,23). Consequently, SCAN1 lymphoblastoid cells or yeast cells expressing the analogous yeast His gab to Arg (H432R) mutant are more sensitive to CPT-induced DNA lesions or to oxidative DNA damage induced by H 2 O 2 , ␥-irradiation, or bleomycin (7,(23)(24)(25). However, in patients, the SCAN1 phenotype is not evident until the second decade of life, is restricted primarily to the neurodegeneration of cerebellar neurons, and is not associated with an increased risk of cancer or immune deficiency (13). Thus, the defect in Tdp1 catalysis imparted by the H432R mutation does not suffice to induce acute cellular toxicity; rather it is likely that the cumulative effects of chronic Tdp1-DNA stabilization induce the neuropathology of SCAN1.
In structural studies of the analogous SCAN1-like mutation in yeast Tdp1 (H432R), crystallographic data revealed a more shallow catalytic pocket with increased positive charge, in comparison with wild type Tdp1 (23). Such alterations in electrostatic potential appear to enhance the noncovalent binding of Tdp1 to DNA. Yet this increment in substrate affinity fails to explain the defect in (H432R) SCAN1 mutant catalysis. For example, a similar change in electrostatic potential and DNA binding was induced in the H432K mutant; however, this substitution did not impair the resolution of the covalent Tdp1-DNA linkage. Indeed, the catalytic activity of this mutant is similar to that of wild type Tdp1 (23). In contrast, substitution of His gab with residues that contain smaller polar or aliphatic side chains (including Asn, Glu, Leu, Ala, Ser, or Thr) induces an even more severe lethal phenotype, effectively transforming Tdp1 into a potent Top1-dependent cellular toxin, even in the absence of CPT (7,23).
To investigate how seemingly subtle alterations in His gab have profoundly distinct effects on cell viability, we focused on the coordinated function of His gab with His nuc during mutant Tdp1 catalysis (see Fig. 1A). For example, mutation of the His nuc to Ala in human and yeast Tdp1 has been reported to inhibit Tdp1 catalytic activity in in vitro assays (12,16,26,27). We therefore hypothesized that introducing the His nuc Ala mutation into the His gab mutant proteins would suppress the enhanced CPT sensitivity of the SCAN1 H432R mutant and the Top1-dependent lethality of the more toxic H432N mutant (7,23,24). Surprisingly, however, we report here that the His nuc Ala substitution (H182A) actually potentiated the cytotoxicity of the His gab mutants and induced a Top1-independent lethal phenotype in combination with the H432N mutant. In contrast, the introduction of a bulky Phe residue at residue 182 completely suppressed the catalytic activity and the toxicity of the His gab mutants. Our findings suggest that in the H182A mutant, a conserved His 181 residue provides an alternative nucleophile for substrate hydrolysis, an activity that is constrained by the active site geometry of wild type Tdp1 or the H182F mutant. Moreover, the enhanced flexibility of H182A active site residues appears to impair the resolution of mutant His gab covalent phosphohistidyl intermediates. These results are discussed in terms of the catalytic mechanism of Tdp1 and provide the rationale for developing chemotherapeutics that stabilize rather than inhibit the formation of the covalent Tdp1-DNA intermediate.
For cell viability studies, the tdp1 alleles were expressed from the galactose-inducible promoter GAL1 in YCpGAL1TDP1⅐L vectors as described in Gajewski et al. (23). tdp1 mutant alleles were created using QuikChange Site-Directed Mutagenesis kit (Stratagene). yTop1 was constitutively expressed at elevated levels from the GPD promoter YCpGDP-TOP1⅐U (28). The yTop1Y727F catalytic inactive mutant was expressed from the galactose-inducible GAL1 promoter of the YCpGAL1-Top1Y727F⅐U plasmids (7,31). For protein purification from yeast, we first introduced GAL4 coding sequences downstream of the bidirectional 10GAL1 promoter in YEpGAL4 -10GAL1⅐L to enhance galactose-induced expression from the GAL1 promoter. N-terminal Flag-tagged tdp1 constructs, produced by PCR amplification, were subsequently cloned into this vector to yield YEpGAL4 -10GAL1-tdp1⅐L vectors. All TDP1 alleles were confirmed by DNA sequencing. PCR primer sequences are available upon request.
Immunoblotting of Yeast Cell Extracts-Anti-yeast Tdp1 antibodies were previously described (7) and anti-tubulin antibodies (MCA77G) were from AbD Serotec. Briefly, as described (7), cultures expressing Tdp1 from YCpGAL1 constructs were induced for 6 h with 2% galactose, corrected to the same A 595 and lysed by vortexing with acid-washed glass beads at 4°C in 50 mM Tris, pH 8.0, 2 mM EDTA, 2 mM EGTA, 10% glycerol, and Complete EDTA-free protease inhibitor (Roche). Proteins were resolved on a 12% SDS-PAGE, blotted onto PVDF, immunostained with anti-Tdp1 and anti-tubulin antibodies, and visualized by chemiluminescence.
Yeast Cell Viability Assays-Semi-quantitative colony formation assays were performed as described in He et al. (7). Briefly, cultures of yeast cells, co-transformed with indicated vectors, were diluted to A 595 ϭ 0.3 in TE buffer (50 mM Tris, pH 8.0, 5 mM EDTA) and serially diluted 10-fold. Aliquots (5 l) were then spotted onto selective media plates containing 2% galactose, with or without the indicated concentration of CPT, 0.0025% DMSO, and 25 mM HEPES, pH 7.2, and incubated for 4 days at 30°C. At least three independent experiments were performed.
Quantitative colony formation assays were also performed, as described in Ref. 23. Briefly, exponential cultures of cells transformed with the indicated vectors were corrected to A 595 ϭ 0.3, 10-fold serially diluted, and plated on selective medium containing 2% galactose. The number of viable cells forming colonies were counted following incubation at 30°C. All experiments (n Ն 3) were performed in duplicate.
Tdp1 in Vitro Activity Assays-Tdp1 activity was assessed as described (23). Briefly, a 14-mer (5Ј-GATCTAAAAGACTT-3Ј) oligonucleotide with a 3Ј-phosphotyrosine was used as substrate, and an identical oligonucleotide with a 3Ј-phosphate served as substrate control (Midland Certified Reagent Company, Inc.). 5Ј-32 P end-labeled oligonucleotide, 16.7 nM under unsaturated conditions, and 1000 nM for saturated conditions was incubated in reaction buffer (50 mM Tris, pH 7.5, 2 mM EDTA, 100 mM KCl, 2 mM DTT) with the indicated amounts of Tdp1. After 10 min at 30°C, half the sample was combined with USB stop buffer and 8 M urea for analysis in 20% polyacrylamide/urea sequencing gels, whereas the other half was subjected to 12% SDS-PAGE. To determine enzyme kinetics of wild type Tdp1 and the H182A, H432R, and H432N mutant enzymes, we used the same reaction conditions as described above but at one enzyme concentration (that converts 40% or less substrate to product) and increasing concentrations of substrate. Reaction products were visualized by phosphorimaging analysis and quantified with ImageQuant TL 1D version 7.0 (GE Healthcare). Product concentrations and initial velocity were calculated from the ratio of the integrated intensities of substrate and product (which is equal to enzyme-DNA intermediate ϩ 3Ј-phosphoryl product) bands. Four independent reactions were analyzed using nonlinear regression fitting of the Michealis-Menten equation to calculate the K m and V max values with the VisualEnzymics software (Soft Zymics, Inc., Charlottesville, VA) in the IGOR PRO 6 program 2010 (WaveMetrics, Portland, OR).
nLC-ESI-MS(MS)-In vitro reactions, prepared with 2 g of Tdp1H182A with or without oligonucleotide (as above), were incubated with 3000 units of micrococcal nuclease at 37°C for 30 min, then resolved by 4 -12% SDS Bis-Tris PAGE, and stained with colloidal blue (Invitrogen). ␣-Casein (5 g) was included as both a carrier protein and as an internal quality control. For in gel digestion, bands corresponding to ϳ70 kDa (Tdp1) and 19 -25 kDa (␣-casein) were excised, equilibrated in 100 mM ammonium bicarbonate, reduced with 10 mM DTT, carbidomethylated with 55 mM iodoacetamide, dehydrated, and digested with Trypsin Gold (Promega). The digested peptides were concentrated under vacuum, resolubilized in 0.1% formic acid, and injected onto a Surveyor HPLC plus (Thermo Scientific) using a split flow configuration on the back end of a 100-micron-inner diameter ϫ 13-cm pulled tip C-18 column (Jupiter C-18 300 Å, 5 microns; Phenomenex). Peptide fractions were then directly sprayed into a Thermo Orbitrap Velos Pro hybrid mass spectrometer equipped with a nano-electrospray source over a 1-h gradient set to increase from 0 -30% acetonitrile in deionized H 2 O containing 0.1% formic acid with a flow rate of 0.3 l/min. Following each parent ion scan, fragmentation data were collected on the top most intense 15 ions, in collision-induced dissociation mode, with the following instrument settings: spray voltage, 1.9 kV; capillary temperature, 170°C; and 1 microscan, with the scan window set at 300 -2,00 m/z, and maximum inject times of 500 and 100 ms for the parent ion and fragmentation scans, respectively. The parent ion scans were obtained at 60 K resolution in the Oritrap with a minimum signal threshold of 2000 counts. The activation settings were: charge state, ϩ2; isolation width, 2.0; m/z, normalized collision energy, 35.0; activation Q, 0.250; and activation time, 25 ms. For data-dependent settings, monoisotopic precursor selection was enabled, in addition to charge state screening with rejection of 1ϩ ions, and dynamic exclusion with repeat count of 2, repeat duration of 30 s, exclusion list size of 500, and exclusion duration of 90 s.
The XCalibur RAW files collected in profile mode were centroided and converted to MzXML using ReADW v. 3.5.1. The .mgf files were created using MzXML2Search for all scans with a precursor mass between 300 and 2,000 Da. The data were searched using SEQUEST set for three maximum missed cleavages, a precursor mass window of 20 ppm, trypsin digestion, and variable modification C at 57.0293 and M at 15.9949. For the fragment-ion mass tolerance, 0.0 Da was used. Searches were performed with the SGD database with protein sequences specific to these experiments included, and peptide IDs were filtered using Scaffold (Proteome software), with cutoff values set with peptide length (Ͼ5 amino acids) and excluding peptides with a MHϩ1 charge state. Peptide probabilities were set to Ͼ90% C.I., with the number of peptides/protein set at 2 or more and protein probabilities set to Ͼ97% confidence interval, resulting in protein IDs with Ͼ99% confidence. Two-dimensional time and mass plots were generated in Refiner MS (Genedata, Expressionist).

Expression of Tdp1H182A Induces Top1-dependent Toxicity-
To study the cellular consequences of Tdp1 catalytic dysregulation, we used the yeast S. cerevisiae in which the TOP1 and TDP1 genes were deleted (top1⌬, tdp1⌬) in an otherwise DNA damage response and repair-proficient background (7,23). Mutant tdp1 alleles were then expressed from plasmid-borne, galactose-inducible (GAL1) promoted constructs to prevent adaptation to cytotoxic mutant enzymes. These studies were also facilitated by elevated Top1 expression to increase cellular levels of the Tdp1 substrate, covalent Top1-DNA covalent complexes.
To investigate the mechanism by which His gab substitutions cause increased Top1-dependent toxicity (23), we focused on the role of the His nuc within the first HXK motif (Fig. 1A). In yeast Tdp1, His nuc is His 182 , and His gab is His 432 . Biochemical studies previously reported that mutation of His nuc to Ala resulted in a catalytically inactive enzyme (16,26). This outcome is logical, because removing the nucleophile that initiates substrate cleavage should suppress Tdp1 catalysis. Thus, we reasoned that combining the H182A mutation with the His gab mutants would suppress the cytotoxic phenotypes induced by defects in the resolution of the phosphohistidyl intermediate. Surprisingly, however, expression of the H182A mutant, either alone or in combination with the SCAN1-like H432R mutant, actually potentiated the Tdp1-induced toxicity associated with elevated Top1 levels (1Top1) and enhanced cell sensitivity to CPT ( Fig. 2A and Table 1). In these experiments, Tdp1 mutants were expressed in a wild type TOP1 strain, with or without subtoxic CPT concentrations, to verify that the lethal phenotypes were a consequence of increased covalent Top1-DNA complexes and not elevated Top1 protein levels per se. In cells deleted for wild type Tdp1 (tdp1⌬ strains), galactose-induced expression of Tdp1H182A suppressed the growth of cells with increased levels of Top1 protein (1Top1) ( Fig. 2A and Table 1) and enhanced the CPT sensitivity of cells expressing endogenous levels of Top1 (in Fig. 2A, compare TDP1 and tdp1H182A in the TOP1/CPT panel). In the context of the His gab mutants, introducing the H182A substitution into the H432R allele induced an even more pronounced Top1-dependent lethal phenotype (Table 1 and Fig. 2A, TOP1/CPT and 1Top1 panels), not evident with the single H432R mutant. The results obtained with the H432N mutant were even more surprising. Although expression of the H432N mutant alone induces Top1-dependent toxicity ( Fig. 2A, TOP1/CPT and 1Top1 panels; Table 1; and Refs. 7 and 23), expression of the double H182A,H432N mutant was toxic even in the absence of Top1 ( Fig. 2A, top1⌬ panel). These data suggest that the inappropriate processing of endogenous Tdp1 substrates, other than Top1-DNA intermediates, also contributes to the Tdp1 mutant-induced toxicity.
To determine whether Tdp1 mutant-induced lethality required a catalytically active Top1, as was previously shown for H432N A, overlay of the crystal structure of yeast and human Tdp1 active site residues that comprise the HXKX n N motifs. Human Tdp1 residues are labeled in green (Protein Data Bank code 1NOP) (22), and yeast Tdp1 is labeled in cyan (Protein Data Bank code 1Q32) (7). His 263/182 functions as a nucleophile (His nuc ), whereas His 493/432 acts as a general acid-base (His gab ). B, the two-step Tdp1 catalytic cycle. In step 1, Tdp1 hydrolysis of the Top1-DNA covalent complex by nucleophilic attack of His nuc on the 3Ј-phosphotyrosyl linkage, forming a 3Ј-phosphohistidyl bond (Tdp1His nuc -DNA covalent complex). In step 2, water is activated by the general acid/base histidine, His gab , to hydrolyze the Tdp1-DNA linkage, dissociating Tdp1 from the nicked DNA strand. The singlestrand break ends need subsequent processing by polynucleotide kinase/ phosphatase and DNA ligase to produce intact DNA. FIGURE 2. Tdp1H182A induces a substrate-dependent toxicity. A, top1⌬,tdp1⌬ cells were co-transformed with vector control (top1⌬) or YCpGPD-Top1⅐U ( 1Top1) and control vector (tdp1⌬) or the indicated YCpGAL1-Tdp1⅐L, whereas tdp1⌬ cells were co-transformed with vector control (TOP1) and control vector (tdp1⌬) or the indicated YCpGAL1-Tdp1⅐L. Exponential growing cultures were diluted to A 595 ϭ 0.3, then 10-fold serially diluted, and spotted onto 2% galactose selective media plates supplemented with or without 1 g/ml CPT (only for tdp1⌬-transformants, shown between brackets). B, top1⌬,tdp1⌬ cells were transformed with YCpGAL1-Top1Y727F⅐U ( 1Top1Y727F) and indicated YCpGAL1-Tdp1⅐L or control vector (tdp1⌬), diluted, and spotted (as in A) onto selective media plates containing 2% galactose, with or without 5 g/ml CPT. In addition, top1⌬,TDP1 cells were transformed with vector control (top1⌬) or YCpGPD-Top1⅐U ( 1Top1) and the indicated YCpGAL1-Tdp1⅐L or vector control (tdp1⌬), diluted, and spotted (as in A) onto selective media plates containing 2% galactose. All plates were incubated for 4 days at 30°C. C, total cell extracts collected from 6-h galactose-induced top1⌬,tdp1⌬ cells co-transformed with vector control (top1⌬), and the indicated YCpGAL1-Tdp1vL or control vector (tdp1⌬) used in A were resolved on 12% SDS-PAGE and immunoblotted with anti-yeast Tdp1 and anti-␣-tubulin. (23), the single H182A and double H182A,H432R mutants were expressed in cells with elevated levels of catalytically inactive Top1Y727F (31,32). Indeed, as seen in Fig. 2B, cell viability was maintained, either with or without CPT (compare 1Top1Y727F and 1Top1Y727F/CPT panels). Moreover, these Top1-dependent lethal phenotypes were not simply a result of mutant Tdp1 overexpression. Lower levels of the H182A,H432R mutant protein produced a more severe phenotype than that induced by higher levels of the single H182A mutant protein (Fig. 2C). Together, these data support the concept that the degree of Tdp1-induced toxicity is dependent upon the levels of the Top1-DNA substrate and their conversion to Tdp1-DNA intermediates. To address this possibility, we also expressed these mutants in a wild type Tdp1 strain (top1⌬, TDP1).

mutant-induced toxicity
Tdp1 is capable of resolving the 3Ј-phosphohistidyl bond of the Tdp1-DNA intermediate. Thus, yeast and human Tdp1 His gab mutants (including SCAN1) display a recessive phenotype (7,13,23,24,33). Indeed, expression of endogenous wild type Tdp1 suppressed the Top1-dependent lethality induced by the H182A or H182A,H432R mutant proteins (Fig. 2, A and B, compare 1Top1 panel). These data suggest that the cytotoxic lesion(s) induced by the H182A and H182A,H432R mutant proteins is repaired by wild type Tdp1 and likely involves the formation of a covalent Tdp1-DNA complex. Similar results were obtained with the double H182A,H432N mutant, although in this case, the Top1-independent lethality observed in a top1⌬ strain was only partially suppressed by wild type Tdp1 (data not shown).
Tdp1H182A Is Catalytically Active in Vitro-Our findings suggest that rather than suppressing Tdp1 catalysis, the H182A mutation actually alters the resolution of the Tdp1-DNA intermediate, possibly in a manner similar to the gain of function imparted by the SCAN1 mutant. To directly address the catalytic activity of the H182A mutant, we purified N-terminal, Flag epitope-tagged Tdp1 proteins from a protease-deficient (top1⌬, tdp1⌬, pep4⌬, prb1⌬) yeast strain, as described previously (23). As diagrammed in Fig. 3A, the proteins were then incubated in reaction mixtures using a 32 P 5Ј-end-labeled single-stranded 14-mer oligonucleotide modified with a 3Ј-phosphotyrosyl as substrate (23). Each reaction was then split into two samples and analyzed either for conversion of the 3Ј-phosphotyrosyl linkage to a 3Ј-phosphoryl-oligonucleotide product by denaturing 8 M urea, 20% PAGE or for the formation of covalent Tdp1-DNA intermediates by 12% SDS-PAGE as described (23). Consistent with previous results under unsaturated substrate conditions, ϳ1 nM of wild type Tdp1 protein was sufficient to convert 40% of the 16.7 nM substrate to product, whereas the H432R protein exhibited a ϳ40-fold reduction in activity compared with wild type (Fig. 3, B and C). The more toxic mutant, H432N, showed a reduction of ϳ167-fold (ϳ167 nM enzyme) in catalytic activity versus wild type (Fig. 3, B and  C). Yet at the highest enzyme amount (416.7 nM), we generally did not observe full conversion of substrate (Fig. 3C). No product was detected in "mock" purified extracts (no Tdp1) (data not shown). The H432R reactions also produced covalent protein-DNA intermediates, resolved as a radiolabeled band near the top of the 8 M urea, 20% polyacrylamide gel, or as a ϳ70-kDa radiolabeled Tdp1 protein band by SDS-PAGE (Fig. 3, B and D,  asterisks). In stark contradiction to earlier reports, but consistent with our yeast data, the H182A protein was also catalytically active, albeit with a specific activity ϳ85-fold lower than that of wild type Tdp1. However, we failed to detect covalent Tdp1H182A-DNA intermediates at protein concentrations sufficient for quantitative conversion of the substrate to product (the absence of a high molecular weight radiolabeled band with 138.9 nM of H182A in Fig. 3, B and D). These findings contrast with the results obtained with H432R, where covalent reaction intermediates accumulated prior to reaction products (at 13.9 nM in Fig. 3B). Instead, 10 -25-fold higher concentrations of protein were required to detect low levels of covalent H182A-DNA complexes by denaturing PAGE and SDS-PAGE (Fig. 3E). In similar experiments, a 20-fold molar excess of wild type Tdp1 (over substrate) failed to produce detectable enzyme-DNA intermediates (data not shown).
Mass spectrometry and SDS-PAGE analyses of our H182A protein preparations confirmed that over 90% of the protein was Tdp1H182A. Minor contaminants, including the translation elongation factor EFA-1␣ (ϳ50 kDa), do not covalently bind DNA and are significantly smaller than the ϳ70-kDa band detected in Fig. 3E. Thus, we think it unlikely that the low levels of protein-DNA intermediates detected with H182A are due to minor protein contaminants. Rather, our data demonstrate the catalytic activity of the H182A mutant protein in vitro and suggest that the formation of covalent H182A-DNA intermediates may contribute to the Top1-dependent toxicity observed in yeast.
To define the kinetics of wild type and mutant Tdp1 enzyme activity, an enzyme concentration that converts less than 40% of substrate to product was incubated with increasing concentrations of substrate. The data produced a typical hyperbolic curve in a Michaelis-Menten plot (Fig. 3F). Using nonlinear regression fit with the Michaelis-Menten equation, we determined the Michealis-Menten constant (K m ) and V max and calculated k cat and the second order rate constants (K m /k cat ) ( Table 2). For all reactions, enzyme-DNA intermediates and 3Ј-phosphoryl reaction products were defined as products, the concentration of which was then divided by enzyme amount (ng) to allow comparison of the different Tdp1 enzymes. Comparison of K m values revealed that the single H182A, H432N, and H432R mutant enzymes have a 2-5-fold lower affinity for substrate than wild type Tdp1 (depicted in Fig. 3F). These dif- ferences may reflect increased substrate mobility within the larger catalytic pocket of the mutant enzymes, prior to the interactions with active site residues Lys 184 and Lys 434 , which allow for His nuc attack of the 3Ј-phosphotyrosyl linkage. This model is supported by changes in the three-dimensional topology of the catalytic pockets of the mutants versus wild type, as revealed in the apo-protein crystal structures (23). However, the relative specific activities of the mutant enzymes (wild type k cat divided by mutant k cat ; Table 2) is similar to those obtained with single-turnover experiments for the H432R and H182A mutants, but not the H432N mutant. Even if we compare the second order rate constants (k cat /K m ) of the mutant enzymes with wild type, the H432N mutant still shows the highest relative activity ( Table 2). The k cat /K m value represents how rapidly enzyme turns over under low substrate concentrations. The H432N mutant revealed a higher relative activity (k cat values) under higher substrate conditions (ϳ25-fold less then wild type) than under single-turnover conditions (ϳ170-fold less active then wild type), where substrate is rate-limiting. Yet these relative activities were obtained with a similar enzyme: substrate ratio of ϳ10 (10-fold more enzyme). These findings suggest that only a fraction of the H432N protein is catalytically active under our near physiological conditions of pH 7.5. Indeed, our modeling of pK a values of catalytic residues in different Tdp1 mutant enzymes suggested that the His nuc within the H432N mutant catalytic pocket has a high probability to be protonated and therefore catalytically inactive (23). The Rad9 DNA Damage Checkpoint Suppresses the Top1-dependent Toxicity of Mutant Tdp1-Given the discrepancies between the relatively higher levels of DNA-linked H432R versus H182A complexes observed in vitro, with the more severe   (26, 34 -36). Here we asked whether rad9⌬ would also exacerbate the Top1-dependent lethal phenotypes induced by H182A or the His gab mutants in the absence of drug or in the presence of subtoxic concentrations of CPT. As shown in Fig. 4, in wild type TOP1 cells, rad9⌬ had no effect on the viability of cells expressing the single H182A, H432R, or H432N mutants. In the presence of low concentrations of CPT, rad9⌬ had no impact on the viability of cells expressing wild type Tdp1, no Tdp1 (tdp1⌬), or the H432R mutant. In contrast, rad9⌬ cells expressing H432N or H182A (and Top1) exhibited a 3-4 log increase in CPT sensitivity, relative to isogenic RAD9 cells. These results implicate a selective role for the Rad9 DNA checkpoint in mediating the repair of H182A-or H432N-induced DNA lesions in the presence of CPT stabilized Top1-DNA complexes, but not in response to increased stabilization of the covalent H432R-DNA intermediate. The Top1-dependent lethality of the double H182A,H432N mutant was also enhanced in rad9⌬ cells, even in the absence of CPT (Fig. 4). However, rad9⌬ had no impact on the Top1independent toxicity induced by H182A,H432N expression in top1⌬ cells. The extent of H182A,H432N mutant-induced lethality was indistinguishable between the top1⌬,tdp1⌬ strain shown in Fig. 2A and the isogenic top1⌬, tdp1⌬, rad9⌬ strain shown in Fig. 4. Thus, whereas the Top1-dependent toxicity elicited by H182A and/or H432N-induced alterations in Tdp1 catalysis appears to involve Rad9-mediated DNA repair, the resolution of Top1-independent lesions induced by the H182A,H432N mutant does not require the Rad9 DNA checkpoint.
The Tdp1H182F Protein Is Catalytically Inactive-The Top1-dependent cytotoxicity of the H182A mutant was surprising, in that this phenotype necessitates the function of a nucleophile, other than His 182 , in the formation of a Tdp1-DNA intermediate. Given the coordinated activity of the His nuc and His gab residues, and the spatial geometry of these residues in the crystal structures of Tdp1, we next asked whether the smaller side chain of the Ala residue at position 182 was a critical determinant of mutant-induced toxicity, and which residue within the catalytic pocket serves as a nucleophile in the absence of His 182 . To address these questions, we first generated additional substitutions of His 182 . Among the mutants generated, we included the polar residues Asn and Gln, because neither can function as a nucleophile yet serve to address whether the size of the side chain at residue 182 is a critical determinant of toxicity. We also mutated His 182 to Phe to retain the ring structure of the side chain, but without nucleophilic capability. Expression of either the H182N or H182Q mutant induced a similar Top1-dependent toxicity as that observed with the H182A mutant (Fig. 5A). However, expression of H182F had no adverse effect on cell viability. Moreover, unlike the H182A,H432R mutant ( Fig. 2A), combining H182F with the H432R mutant failed to induce a Top1-dependent lethal phenotype (Fig. 5A). These data suggest that in contrast to the single H182A, H182N, and H182Q mutants, substitution of His 182 with Phe results in a catalytically inactive Tdp1 protein.
To directly address this possibility, we purified the H182F mutant protein and assayed enzyme activity using the substrate described in Fig. 3A. In reactions containing 5000 fmol of H182F, which represents a 20-fold molar excess of protein over substrate, no conversion of the 3Ј-phosphotyrosine substrate to a 3Ј-phosphoryl product was detected (Fig. 5B). In contrast, H182A activity was evident at 1250 fmol. These data demonstrate that substitution of the His nuc with Phe suppresses both Tdp1 catalysis and Top1-dependent cellular toxicity.
H182A-induced Toxicity Requires an Adjacent, Conserved His Residue-The observation that H182A possesses some catalytic activity, whereas the H182F mutant is inactive, suggests that the phenolic side chain of Phe prevents the nucleophilic attack on Tdp1 substrates that occurs with a smaller residue at the His nuc position. To ascertain alternative nucleophiles within the Tdp1 active site, we considered the conservation of active site residues in Tdp1 orthologs. As shown in a representative list in Fig. 6A, a conserved histidine residue (His 181 in yeast and His 262 in human) is immediately N-terminal to the His nuc (Fig. 1A) in all but four of 19 known orthologs of Tdp1. To determine whether this conserved adjacent His functions in the formation of a covalent H182A-DNA intermediate, we mutated His 181 to Ala. Indeed, as seen in Fig. 6B, the H181A mutation abolished the Top1-dependent lethality induced by the single H182A and double H182A,H432R and H182A,H432N mutant enzymes ( Fig. 6B and data not shown). This is probably because these His 181 FIGURE 4. Rad9 DNA damage checkpoint suppresses the Top1-dependent toxicity of mutant Tdp1. top1⌬,tdp1⌬,rad9⌬, TOP1,tdp1⌬, and TOP1,tdp1⌬,rad9⌬ cells were transformed with the indicated YCpGAL1-Tdp1⅐L or control vector (tdp1⌬) and then 10-fold serially diluted and spotted (as in Fig. 2A) onto 2% galactose selective media plates supplemented with or without 0.5 g/ml CPT. The plates were incubated for 4 days at 30°C. FIGURE 5. Tdp1H182F protein is catalytically inactive. A, top1⌬,tdp1⌬ cells were transformed with vector control (top1⌬) or YCpGPD-Top1⅐U ( 1Top1) and the indicated YCpGAL1-Tdp1⅐L or control vector (tdp1⌬), 10-fold serially diluted and spotted (as in Fig. 2A) onto selective media plates with 2% galactose and incubated for 4 days at 30°C. B, in vitro Tdp1 catalytic activity assay (as described in Fig. 3A), with various amounts (fmol) of indicated full-length Tdp1 enzymes resolved in 20% 8 M urea denaturing polyacrylamide gel to detect conversion of 3Ј-phosphotyrosine (S) to 3Ј-phosphate (P). FIGURE 6. Tdp1H182A toxicity depends on a conserved adjacent histidine residue. A, alignment of a selection of Tdp1 orthologs N-terminal HHXKX n N motif amino acid sequence. The His nuc is italicized, and the adjacent histidine is in bold. AA#, amino acid number. UniProt K D /Swiss-Prot protein numbers were as follows: Homo sapiens (Q9NUW8), S. cerevisiae (P38319), Schizosaccharomyces pombe (Q9USG9), Caenorhabditis elegans (Q9TXV7), Leishmania major (Q4Q3N1), Trypanosoma brucei (Q586D1), and Xenopus laevis (Q6DFE8). B, top1⌬,tdp1⌬ cells were transformed with vector control (top1⌬) or YCpGPD-Top1⅐U ( 1Top1) and the indicated YCpGAL1-Tdp1⅐L or control vector (tdp1⌬), diluted, and spotted (as in Fig. 2A) onto selective media plates containing 2% galactose and incubated for 4 days at 30°C. C, conversion of 3Ј-phosphotyrosine (S) to 3Ј-phosphate (P) resolved in 20% 8 M urea denaturing polyacrylamide gel as described for Fig. 3 (A and B) using 16.7 nM substrate. The mean and standard deviation were determined (graph not shown) as described for Fig. 3C, revealing a reduction of relative activity (versus wild type enzyme; Fig. 3, B and C) of ϳ20-fold for H181A and ϳ210-fold for H182A,H432N. Relative activity of H182A,H432R could not be determined because it did not convert 50% of substrate. D, two-dimensional plot of mass and time tags (557-560 m/z, 33-40 min.). The non-cross-linked His 181 -Ala 182 -containing tryptic fragment 170 LIEITMPPFASHATK 184 (Tdp1 peptide, rectangle) is illustrated with a mass and time tag of 557.96 m/z at 38.3 min (3ϩ charge state), with good signal intensity in the oligonucleotide-free sample (Tdp1) and yet not observed in the oligonucleotide reacted sample (Tdp1-Oligo). Serving as a control peptide 65 IIDLTNQEQDLSER 78 (control peptide, circle) presented with a similar mass and time tag of 558.62 m/z @ 33.7 min (3ϩ charge state) and was observed at nearly equal intensities in both samples Ϯ oligonucleotide. Neither fragment was observed in the negative control sample (Neg. Control).
double and triple mutant proteins are not catalytically active. Indeed, the double H181A,H182A mutant protein did not show any catalytic activity (Fig. 6C), suggesting that the triple mutants are also catalytically inactive. Conversely, the single H181A mutant protein is still catalytically active and only ϳ21-fold less active then wild type Tdp1 (Fig. 6C and data not shown). Moreover, the double H182A,H432R and H182A,H432N mutant proteins reveal a dramatic reduction in catalytic activity (Fig. 6C). As such, the H182A,H432R enzyme was not able to convert 50% of the substrate at the highest concentration tested (675 nM), whereas the H182A,H432N enzyme revealed a reduction in catalytic activity of ϳ210-fold (ϳ270 nM enzyme) compared with wild type (Fig.  6C and data not shown). This is similar to the single H432N mutant, under conditions where ϳ200 nM of enzyme converted 50% of the 16.7 nM substrate to product (Figs. 3B and 6C, compare  H432N). This further supports the concept that the loss of a general acid-base residue at position 432 (H432N) increases the probability that the His nuc is protonated to render the enzyme inactive ( Fig. 3B and Table 2) (23).
To address the ability of His 181 to function as a nucleophile in a H182A background, we next assessed the formation of a phosphohistidyl bond between His 181 and the Tdp1 substrate shown in Fig. 3A. In vitro reactions were performed with 2 g of purified H182A protein, with or without a nonradioactive substrate, to be analyzed by mass spectrometry. The reaction mixture was first incubated with micrococcal nuclease to remove the bulk of the oligonucleotide substrate, and then the H182A protein was resolved by SDS-PAGE. Protein within the molecular weight range of Tdp1 was recovered from the gel and trypsinized. The resultant peptides were then analyzed by nLC-ESI-MS(MS), with the goal of defining the presence of the His 181 -Ala 182 -containing sequence 170 LIEITMPPFASHATK 184 bound to mononucleotide 3Ј-T/TCAG. Mass spectrometry analysis revealed similar sequence coverage of Tdp1 in both samples (with or without oligonucleotide). However, whereas the non-crosslinked His 181 -Ala 182 -containing fragment (mass and time tag; 557.96 m/z @ 38.3 min, 3ϩ charge state) was characterized in the oligonucleotide-free sample with good signal intensity (Fig.  6D, rectangle), the same tryptic peptide was not observed in the oligonucleotide reacted sample. In contrast, the Tdp1-derived peptide 65 IIDLTNQEQDLSER 78 (mass and time tag; 558.62 m/z @ 33.7 min, 3ϩ charge state) was observed at nearly equal intensities in both samples with and without oligonucleotide (Fig. 6D, circle). Despite several attempts, we were unable to positively characterize the nucleotide-linked amino acid residue by nLC-ESI-MS(MS) analysis, presumably because the mononucleotide modification likely interferes with ionization and fragmentation efficiency. Nevertheless, the comparative analyses of samples obtained with or without oligonucleotide are consistent with the model in which the His 181 -Ala 182 -containing tryptic peptide is covalently attached to the mononucleotide via His 181 .

DISCUSSION
Tdp1 catalyzes the repair of 3Ј-phospho-DNA adducts and, to a limited extent, 5Ј-DNA adducts. DNA topoisomerases, on the other hand, catalyze changes in the linkage of DNA strands or helices, whereas site-specific recombinases catalyze the reciprocal exchanges of double-stranded DNA molecules (reviewed in Ref. 37). Yet, despite the biological diversity of these activities, these enzymes share a reaction mechanism that involves the formation of a covalent enzyme-DNA intermediate. In each case, the chemistry of coupled transesterification reactions obviates the need for ATP hydrolysis to effect substrate cleavage or covalent complex resolution. With DNA topoisomerases and site-specific recombinases, a nucleophilic tyrosine or serine initiates the coupled DNA cleavage and religation reactions required to effect changes in DNA topology or DNA recombination. With Tdp1, the coordinated activity of two histidines function in a two-step cycle to remove DNAphospho-adducts. Moreover, in contrast to the intact DNA strands produced by DNA topoisomerases and recombinases, Tdp1 processing of adducted DNA ends yields a nicked DNA substrate that is subsequently repaired by the concerted action of PNK/DNA ligase (Fig. 1B).
Among DNA repair proteins, Tdp1 appears to be unique in the formation of a phosphohistidyl intermediate with a DNA substrate. Aprataxin also relies on a two-step mechanism of catalytic histidines to remove AMP from 5Ј-adenylated DNA lesions. However, in this case, the nucleophilic histidine forms a transient covalent bond with the AMP adduct, rather than the adducted DNA, to liberate a free 5Ј-phosphoryl DNA end (38). It has also been reported that the multifunctional enzyme Tdp2 (also known as TTRAP or EAPII) is more efficient in catalyzing the resolution of 5Ј-phosphotyrosyl adducts than Tdp1 (39). However, Tdp2 is a member of the exonuclease-endonucleasephosphatase (or EEP) nuclease family that uses single Mg 2ϩ -dependent catalysis to cleave the 5-phosphotyrosine bond and functions in the nonhomologous end joining pathway (40,41). Thus, Tdp1 and Tdp2 constitute complementary DNA repair activities, with distinct reaction mechanisms (27,42,43).
Although the chemistry of DNA transactions catalyzed by topoisomerases and Tdp1 does not require ATP hydrolysis, the downside to the formation of a transient covalent enzyme-DNA intermediate is the potential formation of toxic DNA lesions. This concept has been extensively exploited in the development of microbial and cancer chemotherapeutics, which induce toxicity via the stabilization of enzyme-DNA intermediates (including ciprofloxacin for bacterial DNA gyrase, etoposide for Top2, and CPTs for Top1) (2,44,45). In addition, the accumulation of enzyme-DNA intermediates can also result from mutation-induced alterations in DNA cleavage/religation; e.g. the self-poisoning Top1 mutants Top1T722A and Top1N726H (46,47). The best example of a naturally occurring mutation, the Tdp1H493R mutant identified in human SCAN1 patients, is associated with a progressive cerebellar atrophy that presents in the second decade of life (13). Substitution of Arg for His gab enhances the stability of the Tdp1-DNA intermediate in vitro and produces a mild toxicity that is recessive to wild type Tdp1 (7,13,23,33).
We previously reported that substituting smaller polar or aliphatic side chains (such as Asn, Glu, Leu, Ala, Ser, or Thr) for the His gab induced more severe cytotoxic phenotypes in yeast (7,23). Yet, in contrast to results obtained with the yeast SCAN1-like H432R mutant, the use of conventional oligonucleotide-tyrosine substrates did not reveal increased covalent An Alternative Nucleophile Poison Tdp1 MARCH 6, 2015 • VOLUME 290 • NUMBER 10 JOURNAL OF BIOLOGICAL CHEMISTRY 6211 intermediates in vitro. Nevertheless, the Top1-dependent depletion of Tdp1H432N protein in cells was consistent with enhanced H432N-DNA complexes in cells.
The current study extends this concept to demonstrate that alterations in His nuc catalysis can also induce Top1-dependent cytotoxic DNA adducts and that the combination of His nuc and His gab substituents induced an even more severe Top1-independent toxic phenotype. Liu et al. (26) previously reported that expression of the His nuc (H182A) mutant in a tdp1⌬ yeast strain generated a toxic Top1-dependent phenotype and increased cell sensitivity to CPT. However, these studies did not address the catalytic mechanism of this mutant protein. In contrast, our observations suggest that the temporal and spatial coordination of His nuc (His 182 ) and His gab (His 432 ) activities are critical for Tdp1 catalyzed resolution of DNA substrates (Fig.  1B). First, expression of H182A (and Tdp1 mutants with other small side chain change substituents of His nuc ) induced cell lethality, which was not evident with the introduction of Phe at position 182 ( Figs. 2A and 5A). Our mass spectrometry and mutational analyses of His 181 (Fig. 6) support the formation of a phosphohistidyl intermediate involving an adjacent His 181 residue as the probable cause of Tdp1H182A-induced cell lethality ( Figs. 2A and 3E). These data suggest flexibility within the enzyme active site that allows for the formation of His 181 -DNA intermediates when the His nuc residue is replaced with residues containing a small side chain. Indeed, a His residue immediately N-terminal to His nuc is conserved in the Tdp1 enzyme family. However, in the crystal structures of yeast and human Tdp1, including the apo structure of yeast H182A, His 181 is facing away from the active site and forms a hydrogen bond with Tyr 244 (Fig. 7A) (7,22,23). Rotation of His 181 , for nucleophilic attack on a DNA adduct, would be easier in the H182A mutant (Fig. 7B), because more space is available within the active site, whereas the bulky side chain of Phe at residue 182 would sterically hinder such rotation. Docking or interaction of the Top1-DNA substrate with Tdp1H182A might also induce the rotation of His 181 and is supported by the predicted ionization state changes of Tdp1 catalytic residues with substrate docking (23). In this context, it is worth noting that lower levels of His 181 and His 182 mutant proteins (single or combined with other substitutions) were detected in immunoblots of cell lysates, with a corresponding reduction in the yield of affinitypurified proteins (Fig. 2C and data not shown). Because these catalytic residues are located within structurally defined loop domains (Fig. 1A), the His 181 residue may act to stabilize the structure of the protein, thereby facilitating the positioning of the His nuc for efficient catalysis. Indeed, a hydrogen bond between yeast His 181 and Tyr 244 (Fig. 7), or the orthologous human His 262 and Tyr 330 , suggested in crystal structures of yeast and human Tdp1 (7,15,22), may contribute to Tdp1 protein stability in cells.
Another consideration is that the increased distance between the resultant His 181 -DNA linkage and the His gab residue could preclude effective resolution of the covalent intermediate. Indeed, increased covalent H182A-DNA complexes were observed using oligonucleotide-based substrates in vitro, although they were only observed at much higher enzyme concentrations than that observed with the SCAN1-like H432R mutant (Fig. 3, B, D, and E). The low turnover rate (k cat ) and 280-fold reduction in second order reaction constant (k cat /K m ) of the H182A mutant, relative to wild type Tdp1, can also be explained by the rotation of His 181 .
Several lines of evidence also suggest the stabilization of covalent H182A-DNA intermediates in cells. First, the Top1dependent lethality induced by H182A required the expression of a catalytically active topoisomerase I capable of forming Tdp1 substrates (Fig. 2B) and was enhanced by CPT targeting of topoisomerase I. Second, the lethal phenotype induced by H182A expression was recessive to wild type Tdp1, consistent with the resolution of adducted DNA by the wild type enzyme. Together, these data support the concept that the degree of Tdp1-induced toxicity is dependent upon the levels of the Top1-DNA substrate and their conversion to Tdp1-DNA intermediates. Similar results were obtained with H432N (7); however, only the double H182A,H432N mutant exhibited a Topoisomerase I-independent lethality, which was only partially suppressed by wild type Tdp1. These findings suggest that other, spontaneous DNA adducts also serve as endogenous substrates for Tdp1-mediated repair.
Our biochemical characterizations and studies in yeast strains deleted for the Rad9 DNA damage checkpoint (rad9⌬ cells) further suggest that the DNA adducts induced by Tdp1H182A and H432N are distinct from those induced by the SCAN1-like H432R mutant enzyme. Notably, rad9⌬ induced a Ͼ3-log drop in the viability of CPT-treated cells expressing either the H182A or H432N mutant enzyme, but not the H432R mutant (Fig. 4). These data indicate that alterations in Tdp1H182A or H432N catalysis suffice to induce DNA lesions  (7)) and His 181 , Ala 182 , and His 432 residues within the Tdp1H182A crystal structure in blue (Protein Data Bank code 3SQ3 (23)). In gray is shown the ␣-carbon backbone trace of Tdp1 protein and surrounding residues that might influence path of rotation. B, H182A catalytic pocket focusing on the proposed rotation (orange arrow, with the rectangle symbolizing end point after rotation) of His 181 upon docking of the substrate to face the substrate within the active site, which would allow His 181 to perform a nucleophilic attack. Note the increase distance from His 432 to His 182 in A and to potential location of His 181 (after rotation, orange rectangle) in B.
that activate the Rad9 DNA damage checkpoint. Given the recessive nature of the Top1-dependent lethality of these mutant enzymes, we posit that such adducted DNA likely involves covalent mutant Tdp1-DNA complexes. If we consider the kinetics of Tdp1-DNA intermediate formation using the oligonucleotide substrate, no intermediates are detected with wild type Tdp1. With the H432R mutant enzyme, reaction intermediates are detected; however, they accumulate prior to detectable product, suggesting a delay in covalent complex resolution. If the reaction intermediate is buried within the active site of Tdp1H432R, possibly because of the tighter DNA binding suggested by the surface charge topography of the apo crystal structure (23), then the sequestration of this DNA adduct might prevent activation of the Rad9 DNA damage checkpoint, even in the presence of CPT.
In contrast, Tdp1H182A-DNA intermediates are detected at 10 -25-fold higher concentrations than those needed for substrate conversion to product, which suggests the inability of H182A or H432N mutant enzymes to resolve the covalent intermediate. Thus, the increased flexibility or spatial distortion of the Tdp1 active site, which impairs the resolution of the reaction intermediates, may also trigger alternative Rad9-dependent DNA repair pathways to induce even more toxic DNA lesions. These results are reminiscent of CPT poisoning of Top1-DNA complexes or of the self-poisoning mutants Top1N722H and Top1T718A, which accumulate Top1-DNA complexes via distinct defects in catalysis (46,47). Top1N722H exhibits an enhanced rate of DNA cleavage (formation of the covalent Top1-DNA complex), and Top1T718A displays a reduced religation rate (slower resolution of the covalent intermediate), whereas the double mutant is even more toxic than either single mutant enzyme. The same attributes may be extended to the H182A and H432N Tdp1 mutants, where double mutants are more toxic than the single mutants (Figs. 2A and 4) and induces a unique Top1-independent lethality that is not suppressed by the Rad9 checkpoint or wild type Tdp1 (Figs. 2 and 4). Taken together, our findings demonstrate distinct mechanisms of cytotoxicity are induced by mutation of the catalytic His nuc or His gab residues in Tdp1. The conservation of Tdp1 enzyme-induced lethality is supported by the similar dysregulation of human and yeast SCAN1-associated His gab to Arg mutants, including the increased half-life of the covalent reaction intermediate in vitro and enhanced cell sensitivity to CPT (7,16,23,33). Moreover, preliminary studies of mammalian cells expressing human hTdp1H493N and hTdp1H263A mutants (analogous to yeast Tdp1H432N and Tdp1H182A, respectively, (23)) indicate the induction of a similar cytotoxic phenotype. 4 Thus, the mechanisms of yeast Tdp1 enzyme dysregulation, described in these studies, also appear to be a feature of human Tdp1. Moreover, these observations support the concept that the targeting of Tdp1 to develop small molecules that poison the enzyme by increasing the stability of Tdp1-DNA intermediates represents a novel chemotherapeutic strategy.