Structure–function studies of a plant tyrosyl-DNA phosphodiesterase provide novel insights into DNA repair mechanisms of Arabidopsis thaliana

Abstract

TDP1 (tyrosyl-DNA phosphodiesterase 1), a member of the PLD (phospholipase D) superfamily, catalyses the hydrolysis of a phosphodiester bond between a tyrosine residue and the 3′-phosphate of DNA. We have previously identified and characterized the AtTDP gene in Arabidopsis thaliana, an orthologue of yeast and human TDP1 genes. Sequence alignment of TDP1 orthologues revealed that AtTDP has both a conserved C-terminal TDP domain and, uniquely, an N-terminal SMAD/FHA (forkhead-associated) domain. To help understand the function of this novel enzyme, we analysed the substrate saturation kinetics of full-length AtTDP compared with a truncated AtTDP mutant lacking the N-terminal FHA domain. The recombinant AtTDP protein hydrolysed a single-stranded DNA substrate with K_m and k_cat/K_m values of 703±137 nM and (1.5±0.04)×10⁹M⁻¹·min⁻¹ respectively. The AtTDP-(Δ1–122) protein (TDP domain) showed kinetic parameters that were equivalent to those of the full-length AtTDP protein. A basic amino acid sequence (RKKVKP) within the AtTDP-(Δ123–605) protein (FHA domain) was necessary for nuclear localization of AtTDP. Analysis of active-site mutations showed that a histidine and a lysine residue in each of the HKD motifs were critical for enzyme activity. Vanadates, inhibitors of phosphoryl transfer reactions, inhibited AtTDP enzymatic activity and retarded the growth of an Arabidopsis tdp mutant. Finally, we showed that expression of the AtTDP gene could complement a yeast tdp1Δrad1Δ mutant, rescuing the growth inhibitory effects of vanadate analogues and CPT (camptothecin). Taken together, the results of the present study demonstrate the structure-based function of AtTDP through which AtTDP can repair DNA strand breaks in plants.

INTRODUCTION

TDP1 (tyrosyl-DNA phosphodiesterase 1) specifically hydrolyses the phosphodiester bond between a tyrosine residue and the 3′-phosphate of DNA in the DNA–TOP1 (topoisomerase I) complex [1]. The tyrosine residue of TOP1 cleaves one strand of DNA and forms a 3′-phosphotyrosine intermediate, changing DNA topology by altering the linkage of the DNA strand [2]. TDP1 was first identified in Saccharomyces cerevisiae [3], where its protein was shown to remove TOP1-linked DNA breaks [4]. Yeast tdp1 mutant cells are hypersensitive to CPT (camptothecin), an anticancer drug that stabilizes the transient TOP1–DNA complex [3,4].

On the basis of sequence comparison, the TDP1 protein has been shown to belong to the PLD (phospholipase D) superfamily, which comprises proteins that are ubiquitous in bacteria, yeast, plants and mammals [5]. Members of the PLD superfamily catalyse the hydrolysis of a phosphodiester bond in glycerophospholipids, such as phosphatidylcholine, generating phosphatidic acid and free choline [6]. They contain an N-terminal region that varies in both size and amino acid identity. The function of this variable N-terminal region is unknown. Indeed, deletion of the N-terminal amino acids of human TDP1 was not found to affect enzyme activity in vitro [6]. A universal structure in the PLD superfamily is a repeated catalytic HKD motif [HXK(X)₄D(X)₆GSXN], which arose as a result of a gene duplication event [3,6]. These HKD motifs have highly conserved histidine (H) and lysine (K) residues [6–10]. Crystal structure analysis revealed that a histidine and a lysine residue in both HKD motifs are clustered together at the active centre of the enzymes [11].

A point mutation (H493R) in the human TDP1 gene is physiologically important, as, in the homozygous state, it is responsible for SCAN1 (spinocerebellar ataxia with axonal neuropathy), an autosomal recessive neurodegenerative syndrome [12]. This recessive mutation reduces enzyme activity and, importantly, causes the accumulation of a TOP1–DNA complex lesion. TDP1-mutant SCAN1 cells are defective in the repair of DNA single-stranded breaks induced by CPT [13–15]. Moreover, human TDP1 was found to repair single-stranded DNA breaks caused by ionizing radiation [16]. TDP1 has DNA and RNA exonuclease activity and can hydrolyse various substrates blocking 3′-termini at strand breaks, including 3′-phosphoglycolates, 3′-phosphoamide and terminal abasic sites [9,17,18].

We have previously identified AtTDP, an orthologue of human TDP1 in Arabidopsis [19]. Unlike yeast and human TDP1 proteins, AtTDP consists of two domains: a SMAD/FHA (forkhead-associated) domain in the N-terminal region and a TDP domain in the C-terminal region. AtTDP, like other TDP1 orthologues, can remove the tyrosyl group from tyrosyl-DNA substrate in vitro. Moreover, AtTDP protein showed substrate specificity for single-stranded, blunted, tailed and nicked duplex DNA. The knockout tdp mutant exhibited a dwarf phenotype due to reduced cell numbers, which was caused by the accumulation of DNA damage and progressive cell death during Arabidopsis development [19].

Until now, little was known about the enzymatic function of TDP1 in plants. In the present study, we describe the possible roles of the TDP and FHA domains of TDP in Arabidopsis. Specifically, we demonstrate both that the TDP domain alone is sufficient to hydrolyse a 3′-phosphate DNA–tyrosine complex, and that the FHA domain is necessary for targeting to the nucleus and contains a putative novel nuclear localization signal sequence (RKKVKP). Taken together, the results of the present study suggest that the distinctive structure of Arabidopsis TDP1 protein causes it to have a molecular function unlike that of its yeast and human orthologues.

EXPERIMENTAL

Production of recombinant AtTDP-(Δ1–122) (TDP domain) protein

The region of the AtTDP gene encoding the TDP domain was amplified from the leaves of 2-week-old plants by RT (reverse transcription)–PCR (TaKaRa). RT–PCR was performed using the specific primers 5′-AGAGGATCCATGGCTGAAGACGATGTAGAG-3′ and 5′-TTGAGCTCCTATCTTGGCCAGACTTGTCCA-3′. The TDP domain sequence was PCR-amplified, and the PCR products were cloned into the pGEM-T Easy vector (Promega), then subcloned into a pET30a (+) vector and transferred to Escherichia coli BL21(DE3) cells. Recombinant TDP domain protein expression was induced by 1 mM IPTG (isopropyl β-D-thiogalactopyranoside) at 28°C and the protein was purified using His-Bind Resin and the His-Bind Kit (Novagen), according to the manufacturer's protocol. The purity of the enzyme was verified by SDS/PAGE (12% gel).

Enzyme assays and kinetic determination

The ability of AtTDP and the TDP domain to function as phosphodiesterases and cleave a tyrosyl residue were examined as described previously [19]. The 18-Y (5′-TCCGTTGAAGCCTGCTTT-Tyr-3′) oligonucleotide, an artificial substrate, was used for all enzyme assays. AtTDP or the TDP domain protein was incubated with 18-Y substrate in an appropriate buffer [50 mM Tris/HCl (pH 8.0), 80 mM KCl, 2 mM EDTA, 1 mM dithiothreitol, 40 μg/ml BSA and 5% glycerol] at 28°C for 5 min. Reactions (20 μl) were stopped by the addition of 10 μl of formamide quenching buffer. All of the reaction solutions were subjected to electrophoresis on a 20% acrylamide sequencing gel (acrylamide/bisacrylamide=29:1) and stained by GelRed (Sigma) in TBE buffer (1×TBE=45 mM Tris/borate and 1 mM EDTA) for 30 min. To determine the kinetic parameters of the AtTDP and TDP domain proteins, each protein was incubated with various concentrations of 18-Y substrate. The amount of substrate converted into the product, 3′-phosphate oligonucleotide, was measured by densitometry analysis of the gel image. Initial velocities were determined by plotting the amount of 3′-phosphate oligonucleotide as a function of time. All lines extrapolated to zero product at the start of the reaction, and at least three time points were used to determine the velocity. The concentrations of enzymes were determined using the Bradford protein assay [20]. The specific activity was determined as μmol of product formed/min per mg of enzyme. The kinetic parameters, K_m, k_cat and k_cat/K_m, were determined by the Michaelis–Menten equation V=(V_max[S])/(K_m+[S]) and Lineweaver–Burk plot 1/V=(K_m/V_max·1/[S])+1/V_max and k_cat=V_max/[E₀].

Subcellular localization

The AtTDP-(Δ123–605) (FHA domain) and AtTDP-(Δ1–122) (TDP domain) GFP (green fluorescent protein)-fusion proteins were generated using the same methods as described previously for AtTDP [19], by amplifying from cDNA using FHA domain primers (5′-GGATCCATGTTAAAGGAAGACAATTCATCTC-3′/5′-CATGGATCCTCTAGAGTCCCCCGT-3′) and TDP domain primers (5′-TGGATCCAAGGAGATATACCAATG-3′/5′-GGA TCCATAGCTTGAAGAAATGGTGACCAG-3′) containing a BamHI site respectively, and subcloned into the smGFP (soluble-modified GFP) vector. Rosette leaves of wild-type plants grown for 2 weeks were used for the isolation and transformation of protoplasts. Plasmid DNA (10 μg) containing the FHA domain–GFP or TDP domain–GFP constructs were transfected into Arabidopsis mesophyll protoplasts using the PEG [poly(ethylene glycol)] method [21]. Protoplasts were then incubated in the dark at 24°C for 24 h. Images were obtained using a confocal microscope (Radiance 2000/MP, Bio-Rad Laboratories).

For generation of a GFP-fusion construct containing the basic amino acid sequence from the FHA domain, a duplex oligonucleotide (5′-pGATCCATGAGAAAGAAAGTGAAACCTAG-3′/5′-pGATCCTAGGTTTCACTTTCTTTCTCATG-3′) was synthesized by Biomedic. The basic amino acid sequence was subcloned into pUC19 containing a 35S::smGFP construct, and transfected into the protoplasts of Arabidopsis thaliana by PEG-calcium transfection [22]. After incubation of transfected protoplasts at room temperature (22°C) for 16 h, DAPI (4′,6-diamidino-2-phenylindole, 10 μg/ml) staining was used to visualize protoplast nuclei as a control. Images of smGFP-fusion protein expression in transfected protoplasts stained with DAPI were obtained using a fluorescence microscope (DE/Axio Imager A1, Carl Zeiss) and a multi-photon CLSM (confocal laser-scanning microscope; DE/LSM 510 NLO, Carl Zeiss). The excitation wavelength was 488 nm and emission was detected between 500 and 530 nm.

In vitro mutagenesis

Mutations of various residues in AtTDP to alanine (H236A, K238A, H466A and K468A) were produced by Mutagenex and confirmed by sequencing. The PCR-amplified products were treated with PNK (polynucleotide kinase) and DpnI, and then purified from an agarose gel. The purified PCR products were self-ligated and transformed. To generate each mutant plasmid, we amplified the whole vector using two mutagenic primers. The primer sequences used were as follows: H236A-F2, 5′-gcTTCGAAGGCTATATTTCTTG-3′; H236A-R2, 5′-GTGTGTCCCGAATGAAATAG-3′; K238A-F3, 5′-gcGGCTATATTTCTTGTCTATC-3′; K238A-R3, 5′-CGAATGGTGTGTCCCGAATG-3′; H466A-F, 5′-gcTATAAAGACGTTCACACGTTAC-3′; H466AR, 5′-TGGCATGGCACGACCGC-3′; K468A-F, 5′-gcGACGTTCACACGTTACAACG-3′; and K468A-R, 5′-TATATGTGGCATGGCACG-3′. The mutant plasmids were introduced into E. coli BL21 (DE3) cells using the pET expression system (Novagen). The recombinant protein products were purified using His-Bind Resin and the His-Bind Kit (Novagen), according to the manufacturer's instructions.

Vanadate analogue treatment and measurement of total chlorophyll content in Arabidopsis tdp and wild-type plants

We evaluated the sensitivity of wild-type and tdp plants to vanadate analogues. All Arabidopsis plants used were ecotype Columbia. The loss-of-function AtTDP mutation was isolated from an Arabidopsis activation T-DNA (transferred DNA) tagging mutant pool as described by Lee et al. [19]. Homozygous tdp mutant and wild-type plants were germinated on plates containing half-strength MS (Murashige and Skoog) medium, 1% (w/v) sucrose and 0.7% agar. Seedlings (5-day-old) were transferred on to MS plates supplemented with various concentrations of the vanadate analogues. Seedlings in vanadate-treated medium were allowed to grow for approximately 2 weeks. They were grown at 21°C under constant light or 22°C with a regime of 16 h of light/8 h of dark. The vanadate-treated wild-type and tdp mutant seedlings were freshly weighed and ground in 25 mM Hepes (pH 7.5) and 80% acetone solution. The extracts of Arabidopsis seedlings were centrifuged and the supernatant was used in this experiment. The total chlorophyll content was measured as described by Porra [23]. Briefly, the absorbance of the supernatants at 646.6 nm and 663.6 nm was measured using a spectrophotometer (GE Healthcare), and chlorophyll content was calculated using the following equation: Chls_a+b=17.76A_646.6+7.34A_663.6.

Yeast strains and measurement of drug sensitivity

Yeast strains YW465 (MATa), YW812 (YW465 tdp1Δ::kanMX4rad1Δ::HIS3), YW463 (YW465 rad1Δ::HIS3) and YW625 (YW465 tdp1Δ::URA3) were used in the present study. Full-length AtTDP and the TDP domain sequence were subcloned into the p424 ADH vector (generating the p424 ADH–AtTDP and p424 ADH–TDP domain plasmids) [24]. To test sensitivity to CPT and vanadate analogues, yeast strains were grown at 30°C in YPD [1% (w/v) yeast extract/2% (w/v) peptone/2% (w/v) glucose] or selective medium containing 2% (w/v) dextrose-tryptophan, with or without 20 μM CPT or 1 mM vanadate analogues. Plates were incubated at 30°C for 2 days, and then photographed.

RESULTS

Substrate saturation kinetics of AtTDP protein

We previously determined that AtTDP catalyses the hydrolysis of the phosphodiester bonds between a tyrosine residue and the 3′-phosphate of DNA [19]. To further examine the kinetic parameters of the AtTDP protein, we carried out enzyme reactions with the synthesized single-stranded 18-Y oligonucleotide, which is an 18-mer containing a tyrosine residue at the 3′-terminus, as the substrate (Supplementary Figure S1 at http://www.BiochemJ.org/bj/443/bj4430049add.htm). The recombinant AtTDP has a molecular mass of 70 kDa. As shown in Figure 1 (inset), AtTDP protein efficiently hydrolysed the phosphodiester bond between tyrosine and the 3′-end of the DNA strand, and the amount of product formed was proportional to the assay time. We further examined enzyme activity using serial dilutions of 18-Y substrate. The Michaelis–Menten plot of AtTDP protein activity produced a typical hyperbolic curve (Supplementary Figure S2A at http://www.BiochemJ.org/bj/443/bj4430049add.htm). The V_max of AtTDP was 7.9±0.5 μmol of product formed/min per mg of protein and the turnover number of the enzyme (k_cat) was 1077±242 min⁻¹ (Figure 2 and Table 1).

Figure 1. Time-dependence of AtTDP activity.

The reaction was performed by incubating 0.1 ng of enzyme with 1.5 μM 18-Y substrate at 28°C. The reaction velocity plot was generated by measuring the amount of 18-P product. The inset shows a gel image of a cleavage reaction used to calculate a reaction velocity. Reaction products before the addition of AtTDP protein (lane 1), and 1, 2, 3, 4 and 5 min after the addition of enzyme (lanes 2–6) are shown. All experiments were performed three times. Products were measured by densitometry after separation on a 20% polyacrylamide gel.

Figure 2. Determination of the kinetic parameters K_m and k_cat for AtTDP protein.

Lineweaver–Burk plot of AtTDP activity. The reactions were performed by incubating 1 ng of enzyme and a variety of concentrations of 18-Y substrate at 28°C. Reactions were terminated after 5 min by the addition of formamide quenching solution. All experiments were performed three times. Products were measured by densitometry after separation on a 20% polyacrylamide gel.

Table 1. Summary of kinetic properties of recombinant AtTDP and TDP domain proteins.

Values are means±S.D. (n=3)

Enzyme	Km (nM)	kcat (min−1)	kcat/Km (M−1·min−1)
AtTDP	703±137	1077±242	1.5 × 109±0.04 × 109
TDP domain	770±93	950±115	1.2 × 109±0.02 × 109

Substrate saturation kinetics of Arabidopsis AtTDP-(Δ1–122) (TDP domain) protein

AtTDP consists of a SMAD/FHA domain and a TDP domain [19]. The TDP domain is the most highly conserved domain of AtTDP, as in yeast and human TDP1, which is consistent with its role as a member of the PLD superfamily. To determine whether the TDP domain is responsible for the enzymatic activity of the protein, we generated a recombinant TDP domain polypeptide consisting of amino acids 123–605, using the pET expression system (Novagen). The recombinant TDP domain protein has a molecular mass of 57 kDa. The TDP domain efficiently hydrolysed the phosphodiester bond between tyrosine and the 3′-end of DNA. The amount of product formed was proportional to the assay time (Figure 3). On the basis of the reciprocal Lineweaver–Burk plot (correlation coefficient r²=0.9989, Figure 4), the recombinant TDP domain has a V_max of 8.0±0.9 μmol of product formed/min per mg of enzyme and a k_cat of 950±115 min⁻¹ (Table 1 and Figure 4, and Supplementary Figure S2B).

Figure 3. Time-dependence of AtTDP-(Δ1–122) (TDP domain) activity.

The reaction was performed by incubating 0.1 ng of enzyme with 1.5 μM 18-Y substrate at 28°C. The reaction velocity plot was generated by measuring the amount of 18-P product. The inset shows a gel image of a cleavage reaction used to calculate a reaction velocity. Reaction products before the addition of the TDP domain protein (lane 1), and 1, 2, 3, 4 and 5 min after the addition of enzyme (lanes 2–6) are shown. All experiments were performed three times. Products were measured by densitometry after separation on a 20% polyacrylamide gel.

Figure 4. Determination of the kinetic parameters K_m and k_cat for AtTDP-(Δ1–122) (TDP domain) protein.

Lineweaver–Burk plot of the TDP domain protein activity. The reactions were performed by incubating 1 ng of enzyme and a variety of concentrations of 18-Y substrate at 28°C. Reactions were terminated after 5 min by the addition of formamide quenching solution. All experiments were performed three times. Products were measured by densitometry after separation on a 20% polyacrylamide gel.

The FHA domain basic amino acid sequence, RKKVKP, is a nuclear localization signal

AtTDP has a distinctive additional FHA domain in the N-terminal region, from amino acids 16 to 109. In sequence comparisons between Arabidopsis and yeast or human TDP1, the N-terminal region exhibited 15 and 25% identity respectively (Supplementary Figure S3 at http://www.BiochemJ.org/bj/443/bj4430049add.htm). Among plants, TDP1 proteins showed approximately 35–45% identity with Arabidopsis AtTDP (Supplementary Figure S4 at http://www.BiochemJ.org/bj/443/bj4430049add.htm). Among them, the Medicago truncatula Tdp1 proteins contained an FHA domain, in common with AtTDP, but its function has not been determined. To investigate the role of the FHA domain, we separated the FHA domain [AtTDP-(Δ123–605)] and the TDP domain [AtTDP-(Δ1–122)] and fused each to smGFP (Figures 5A and 5B). As a control, the RFP (red fluorescent protein) gene fused to TFL2 (terminal flower 2) was expressed under the control of the CaMV (cauliflower mosaic virus) 35S promoter. The results show that the FHA domain–GFP fusion protein was targeted to the nucleus (Figures 5C–5E), whereas the TDP domain–GFP fusion protein was expressed in the cytosol (Figures 5F–5H). Thus we further examined whether a specific amino acid sequence within the FHA domain acts as a nuclear localization signal. We chose a short basic amino acid sequence, RKKVKP, from the FHA domain and performed in vivo targeting experiments using a construct composed of this basic sequence fused with smGFP (Figure 5I). The construct showed efficient nuclear localization (Figures 5J–5M). Specifically, more than 80% of RKKVKP–GFP was localized to the nucleus (Supplementary Figure S5 at http://www.BiochemJ.org/bj/443/bj4430049add.htm).

Figure 5. Subcellular localization of the isolated TDP and FHA domains from AtTDP.

Schematic diagrams of the (A) FHA domain–smGFP and (B) TDP domain–smGFP constructs. (C) Light microscope image, (D) GFP fluorescence and (E) RFP fluorescence of a protoplast transfected with FHA domain–GFP and TFL2–RFP. TFL2–RFL expression indicates the nucleus. (F) Light microscope image, (G) GFP fluorescence and (H) RFP fluorescence of a protoplast transfected with TDP domain–GFP and TFL2–RFP. (I) The basic amino acid sequence within the FHA domain–smGFP construct. A transparent protoplast transfected with the basic amino acid sequence–GFP construct is shown (J) under a light microscope, (K) by DAPI fluorescence, (L) by GFP fluorescence and (M) with images from (J), (K) and (L) merged.

Mutation of the most conserved amino acid residues in the HKD motifs drastically reduces the enzymatic activity of TDP protein

Being members of the PLD superfamily, TDP1 orthologues have highly conserved histidine and lysine residues in both of their C-terminal HKD motifs (Supplementary Figure S6A at http://www.BiochemJ.org/bj/443/bj4430049add.htm). Four conserved histidine and lysine residues (His²⁶³, His⁴⁹³, Lys²⁶⁵ and Lys⁴⁹⁵) within the HKD motifs are necessary for the catalytic mechanism of human TDP1 [6]. To confirm whether the conserved amino acid residues of the HKD motifs in Arabidopsis also contribute to the active site, we generated several missense mutations of these residues (Table 2 and Supplementary Figure S6B). Mutation of His²³⁶ to an alanine (H236A) completely disrupted the enzymatic activity of AtTDP. As shown by the appearance of the product 18-P, mutation of Lys²³⁸ to an alanine (K238A) was approximately 10-fold less active than AtTDP (Table 2 and Supplementary Figure S7 at http://www.BiochemJ.org/bj/443/bj4430049add.htm). Mutation of His⁴⁶⁶ (H466A) or Lys⁴⁶⁸ (K468A) reduced enzymatic activities by ~100-fold respectively (Table 2 and Supplementary Figure S7).

Table 2. The enzymatic activities of AtTDP proteins with mutated amino acid residues in the HKD motifs.

Values are means±S.D. (n=3). ND, the H236A mutation completely suppressed the enzyme activity.

Enzyme/mutant	Specific activity (μmol of substrate converted into product/min per mg of enzyme)	Fold reduction
AtTDP	1.5±0.5	1
H236A	ND	–
K238A	0.16±0.001	9.4
H466A	0.015±0.008	100
K468A	0.016±0.001	94

Vanadate analogues inhibit AtTDP enzyme activity

Vanadate analogues (transition metal oxoanions) directly bind to tyrosine [25] and are known to be inhibitors of a variety of enzymes that are involved in phosphoryl transfer reactions. Their inhibitory effects may be based on their ability to mimic phosphates or act as transition stage analogues [26]. In addition to TDP1, some other members of the PLD superfamily are also inhibited by vanadate [27]. The human TDP1–vanadate complex structure has a trigonal bipyramidal geometry that mimics the transition state of hydrolysis of a phosphodiester bond [8].

To determine which vanadate analogues were more effective inhibitors of AtTDP activity, we assayed the enzyme activities of AtTDP and the TDP domain proteins in the presence of various vanadates. The enzymatic activities of AtTDP and the TDP domain proteins were completely inhibited by treatment with 20 and 40 mM sodium orthovanadate respectively. Furthermore, AtTDP and the TDP domain proteins were partially inhibited by treatment with sodium metavanadate and ammonium metavanadate (Figure 6 and Supplementary Figure S8 at http://www.BiochemJ.org/bj/443/bj4430049add.htm).

Figure 6. Inhibitory effects of vanadate analogues.

(A) AtTDP protein and (B) TDP domain protein. Enzyme reactions were performed by incubating 0.1 ng of enzyme with 0.5 μM 18-Y substrate at 28°C. A dilution series of the vanadate analogues was used. The reactions were terminated after 5 min by adding formamide quenching solution. The reaction velocity plots were generated by measuring the amount of 18-P products. Values are means±S.D. (three independent replicates). Products were measured by densitometry after separation on a 20% polyacrylamide gel. Na₃VO₄, sodium orthovanadate; NaVO₃, sodium metavanadate; NH₄VO₃, ammonium metavanadate.

We further examined in vivo whether vanadate analogues affect the growth of wild-type or tdp mutant plants. The seedlings of wild-type and tdp mutant plants were grown in the presence of various concentrations of vanadate analogues. The tdp mutant plants were significantly hypersensitive to vanadate analogues (Figure 7). At a concentration of 0.05 mM, vanadate analogues caused severe growth defects in AtTDP loss-of-function mutant plants relative to wild-type (Figure 7).

Figure 7. Inhibition of growth by vanadate analogues in wild-type and tdp mutant plants.

Wild-type and tdp plants (20 days old) were treated with various concentrations of (A) sodium orthovanadate (Na₃VO₄) (left-hand panel), and the chlorophyll content of these plants was analysed (right-hand panel); (B) sodium metavanadate (NaVO₃) (left-hand panel), and the chlorophyll content of these plants was analysed (right-hand panel); and (C) ammonium metavanadate (NH₄VO₃) (left-hand panel), and the chlorophyll content of these plants was analysed (right-hand panel). Values are means±S.D. (three independent replicates).

Test for complementation of tdp1Δrad1Δ mutant yeast by AtTDP and the TDP domain

The budding yeast (Saccharomyces cerevisiae) tdp1Δrad1Δ strain exhibits a TOP1-dependent growth delay and shows significant sensitivity to CPT [28]. CPT stabilizes the TOP1–DNA complex and is an effective chemotherapeutic agent [25]. To test whether the yeast tdp1Δrad1Δ strain could be complemented by the introduction of plant AtTDP, we examined the CPT and vanadate analogue sensitivity of tdp1Δrad1Δ and of tdp1Δrad1Δ strains expressing full-length AtTDP or the TDP domain. When the tdp1Δrad1Δ strain was exposed to CPT, it showed a significant sensitivity, in contrast with wild-type, tdp1Δ, or rad1Δ strains (Supplementary Figure S9 at http://www.BiochemJ.org/bj/443/bj4430049add.htm). However, tdp1Δrad1Δ expressing full-length AtTDP or the TDP domain did not show significant sensitivity to CPT (Figure 8, left-hand panel). None of the strains examined (tdp1Δrad1Δ, tdp1Δrad1Δ expressing full-length AtTDP or tdp1Δrad1Δ expressing the TDP domain) were sensitive to vanadate analogues (Figure 8, upper panels). However, tdp1Δrad1Δ treated with vanadate analogues displayed significant sensitivity to CPT, and expression of full-length AtTDP or the TDP domain partially rescued this sensitivity (Figure 8).

Figure 8. Complementation test in a tdp1Δrad1Δ yeast strain.

Serial dilutions of the tdp1Δrad1Δ strain, tdp1Δrad1Δ expressing full-length AtTDP or tdp1Δrad1Δ expressing the TDP domain were spotted on to selective medium containing 2% (w/v) dextrose, with or without CPT and vanadate analogues.

DISCUSSION

The TDP1 protein was originally found, in yeast cells, to be responsible for repair of the TOP1–DNA complex in vivo, catalysing the hydrolysis of topoisomerases from the 3′-end of DNA during double-strand break repair [3,4,29]. Similarly, in humans, TDP1 was found to participate in the repair of TOP1-induced lesions during single-strand break repair, and a recessive mutation of the TDP1 gene is responsible for the hereditary disorder SCAN1 [12]. We previously identified and characterized a novel plant TDP1, the Arabidopsis TDP1 protein AtTDP. A loss-of-function mutation of AtTDP resulted in developmental defects and a dwarf phenotype in Arabidopsis. Recombinant AtTDP protein can hydrolyse the phosphodiester bond between a tyrosine residue and the 3′-terminus of single-stranded DNA, or blunted, tailed or nicked double-stranded DNA [19]. Recently, the expression profiles of the MtTdp1α and MtTdp1β genes in Medicago truncatula in response to oxidative stress have been reported [30].

On the basis of sequence alignments, TDP1 proteins have a highly conserved TDP domain at the C-terminus and poorly conserved sequence identity at the N-terminus. As members of the PLD superfamily, TDP1 orthologues, including AtTDP, have two highly conserved HXK(X)₄D(X)₆GSXN sequences, known as HKD motifs, in the C-terminal region. To understand the structure-based function of the AtTDP protein, we created a TDP domain polypeptide through deletion of the N-terminal region, and compared the enzymatic kinetic parameters of the TDP domain with those of the full-length AtTDP protein. The turnover number of AtTDP and truncated TDP domain proteins for the single-stranded substrate seem to be identical as compared with those of human TDP1 (~1×10⁸ M⁻¹·s⁻¹) in the gel-based analysis of human TDP1 activity [31,32]. The results of the present study demonstrate that the TDP domain is itself sufficient to confer the full phosphodiesterase activity of the AtTDP enzyme.

Sequence analysis of AtTDP showed it to contain a unique SMAD/FHA domain in the N-terminal region, unlike yeast and human TDP1s. SMAD/FHA domains are found as a component of forkhead transcription factor complexes and in various eukaryotic and prokaryotic signalling proteins involved in similar protein–protein interactions [33,34]. Generally, the FHA-domain-containing proteins in Arabidopsis show the conserved residues glycine, arginine, serine, histidine and asparagine within the FHA domain [35]. Even though the FHA domain of AtTDP only contains glycine, arginine, serine and histidine residues, it is still categorized as an FHA-domain-containing protein in TAIR (The Arabidopsis Information Resource, http://www.arabidopsis.org) and the Pfam-A matches (http://pfam.sanger.ac.uk).

In the case of AtTDP, although the FHA domain did not contain a known nuclear localization sequence, the FHA domain–GFP protein construct showed nuclear localization. We previously showed that full-length AtTDP protein is localized to the nucleus, whereas a TDP domain–GFP protein construct without the FHA domain is localized to the cytosol [19]. Therefore we focused on a possible nuclear-targeting amino acid sequence (RKKVKP) within the FHA domain of AtTDP, and generated a GFP construct to examine its nuclear localization. The results of the present study demonstrate that this basic amino acid sequence within the FHA domain possesses the ability to confer nuclear localization to AtTDP, which is consistent with its demonstrated enzyme activity. Human TDP1 contains two putative NLSs at residues 56–74 of the non-essential N-terminal region and residues 216–223 within its TDP domain, and must be efficiently localized and expressed in the nucleus for repair mechanisms [36,37].

In humans, a mutation in TDP1 was found to be responsible for the human genetic disorder SCAN1, an autosomal recessive disorder that arises as a result of a point mutation (H493R) in the second HKD motif [12]. Crystallographic studies of human TDP1 revealed that two conserved residues (histidine and lysine) in each HKD motif form an active site that is necessary for enzyme activity [7,8]. Various mutations (H263A, H493A, K265A or K495A) of the HKD motifs in human TDP1 reduce or abolish enzymatic activity [6,31]. To test whether plant TDP1 possesses the same active sites within the HKD motifs, we generated several point mutations in these motifs. The TDP1 reaction is a nucleophilic attack by His²³⁶ on the tyrosyl-DNA 3′ phosphate bond and this releases the DNA from this covalent intermediate [14]. Our results show that the H236A mutation (equivalent to H263A in human TDP1) within the first HKD motif completely obliterated the phosphodiesterase activity. Mutations of a lysine residue (Lys²³⁸) in the first HKD motif, or the histidine (His⁴⁶⁶) or lysine (Lys⁴⁶⁸) residue of the second HKD motif also partially reduced the phosphodiesterase activity. Therefore we concluded that the catalytic mechanism of the plant enzyme is similar to that of other PLD family members.

In structural studies of human TDP1–vanadate complexes, the apparent trigonal bipyramidal geometry mimics the transition state of hydrolysis of a phosphodiester bond, with vanadate covalently bound to His²⁶³ [7,8]. Vanadates can also inhibit the enzyme activity of human TDP1 through a phosphate-mimicking function [38]. Using biochemical assays, we found that the enzyme activities of AtTDP and the TDP domain protein were completely or partially inhibited by sodium orthovanadate, sodium metavanadate and ammonium metavanadate. In addition, Arabidopsis tdp mutant plants were highly sensitive to very low concentrations of sodium orthovanadate and sodium metavanadate in in vivo experiments. These results may be explained by the possibility that somehow tdp mutant plants have plasma membranes that are more permeable to vanadates and/or have an increased affinity of membrane ‘vanadate transporter proteins’ for vanadates. However, future studies are needed determine the biochemical basis for this interesting observation.

Moreover, we assessed the in vivo ability of AtTDP and the TDP domain to rescue the sensitivity of a budding yeast tdp1Δrad1Δ strain to CPT and vanadate analogues, since the treatment of both compounds should be explained by the synergy effect in complementation. CPT is an antitumour agent that acts as a topoisomerase inhibitor by binding to the covalent TOPI–DNA complex and preventing the DNA religation step [3]. The finding that full-length AtTDP or the TDP domain can rescue the sensitivity of a tdp1Δrad1Δ strain to CPT and vanadate analogues suggests a role for AtTDP in the repair of damage induced by CPT and vanadate analogues.

In conclusion, the TDP domain alone is sufficient for enzyme activity, and the FHA domain of Arabidopsis thaliana TDP1 is necessary for its nuclear localization. A basic amino acid sequence within the FHA domain is a novel nuclear localization signal. Furthermore, inhibitor analysis increases our understanding of the dynamics of structure-based enzyme inhibition in plants.

AUTHOR CONTRIBUTION

Hoyeun Kim and Sang-Gu Kim designed the experiments and wrote the paper. Hoyeun Kim, Je-Chang Woo and Sang-Gu Kim analysed and interpreted the data. Hoyeun Kim, Sang Hyeon Na, Young-Min Jeong, So-Young Lee, Hyun-Ju Hwang and Jae Young Hur carried out the experiments. Sang-Hyun Park optimized the yeast experiments.

ACKNOWLEDGEMENTS

We thank Professor Donald J. Armstrong (Oregon State University, Corvallis, OR, U.S.A.) for his critical review of this paper prior to submission, Professor Thomas E. Wilson (University of Michigan Medical School, Ann Arbor, MI, U.S.A.) for providing yeast strains and Won-Ki Huh (Seoul National University, Seoul, Korea) for providing the p424 ADH vector.

FUNDING

This work was supported by a Korea Research Foundation Grant funded by the Korean Government (MOEHRD) [grant numbers KRF-2007-314-C00275, KRF-2008-313-C00841]; and the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology [grant number 2011-0024730].