Insights into Substrate Binding and Catalytic Mechanism of Human Tyrosyl-DNA Phosphodiesterase (Tdp1) from Vanadate and Tungstate-inhibited Structures

Abstract

Tyrosyl-DNA phosphodiesterase (Tdp1) is a DNA repair enzyme that catalyzes the hydrolysis of a phosphodiester bond between a tyrosine residue and a DNA 3′-phosphate. The only known example of such a linkage in eukaryotic cells occurs normally as a transient link between a type IB topoisomerase and DNA. Thus human Tdp1 is thought to be responsible for repairing lesions that occur when topoisomerase I becomes stalled on the DNA in the cell. Tdp1 has also been shown to remove glycolate from single-stranded DNA containing a 3′-phosphoglycolate, suggesting a role for Tdp1 in repair of free-radical mediated DNA double-strand breaks. We report the three-dimensional structures of human Tdp1 bound to the phosphate transition state analogs vanadate and tungstate. Each structure shows the inhibitor covalently bound to His263, confirming that this residue is the nucleophile in the first step of the catalytic reaction. Vanadate in the Tdp1-vanadate structure has a trigonal bipyramidal geometry that mimics the transition state for hydrolysis of a phosphodiester bond, while Tdp1-tungstate displays unusual octahedral coordination. The presence of low-occupancy tungstate molecules along the narrow groove of the substrate binding cleft is suggestive evidence that this groove binds ssDNA. In both cases, glycerol from the cryoprotectant solution became liganded to the vanadate or tungstate inhibitor molecules in a bidentate 1,2-diol fashion. These structural models allow predictions to be made regarding the specific binding mode of the substrate and the mechanism of catalysis.

Introduction

Tyrosyl-DNA phosphodiesterase (Tdp1) catalyzes the hydrolysis of a phosphodiester bond between a tyrosine residue and a DNA 3′-phosphate.¹ The only known occurrence of such a linkage in eukaryotic cells is in the transient covalent enzyme–DNA intermediate formed when a type IB DNA topoisomerase nicks double-stranded DNA. It has therefore been suggested that Tdp1 functions as a DNA repair enzyme to remove such complexes when they become stalled on the DNA in the cell.¹ Stalled topoisomerase I–DNA complexes can be induced by various forms of DNA damage such as nicks, gaps, or abasic sites; by the incorporation of nucleoside analogs such as Ara-C; and by treatment with anti-cancer drugs such as camptothecin.2., 3., 4.. The cytocidal effect of camptothecin appears to require DNA synthesis,5., 6., 7., 8. suggesting that it is the collision of a replication fork with the stalled topoisomerase I that leads to cell death.9., 10. The development of inhibitors of Tdp1 may be useful in combined drug therapy with camptothecin for treatment of cancers, since Tdp1 theoretically counteracts the effects of camptothecin and its derivatives.

Recent biochemical studies have shown that both human and Saccharomyces cerevisiae Tdp1 catalyze removal of glycolate from 3′-phosphoglycolate DNA, although the reaction is less efficient than the removal of tyrosine from 3′-phosphotyrosyl DNA.¹¹ 3′-phosphoglycolate is a common by-product of DNA double-strand breaks caused by oxidative fragmentation of DNA sugars, which can occur as a result of ionizing radiation. The human apurinic/apyrimidinic endonuclease Ape1 can remove 3′-phosphoglycolate from internal breaks in double-stranded DNA, but phosphoglycolates on 3′ overhangs are completely refractory to Ape1.11., 12., 13. Inamdar et al.¹¹ concluded that Tdp1, in addition to repairing stalled topoisomerase I covalent complexes, is also responsible for the repair of free radical-mediated DNA double-strand breaks bearing terminally blocked 3′ overhangs. Thus inhibitors of human Tdp1 may also function as potent radiosensitization agents, since a decrease in human Tdp1 activity could potentially inhibit the repair of radiation-induced double-strand breaks.

Sequence comparisons, mutational analyses,¹⁴ and structure determination¹⁵ have shown that human Tdp1 is a member of the phospholipase D (PLD) superfamily. Enzymes in the PLD superfamily include phospholipases D, bacterial phosphatidyl serine and cardiolipin synthetases, a bacterial nuclease from Salmonella typhimurium (referred to as Nuc), a bacterial toxin and a number of possibly catalytically inactive poxvirus envelope proteins.16., 17., 18. The catalytic domain of the type II restriction endonuclease Bfi I also appears to be a member of the PLD superfamily.¹⁹ All of the enzymes in the PLD superfamily catalyze phosphodiester bond cleavage through a phosphoryl transfer mechanism where the acceptor is either an alcohol or water. The similar chemistry for the members of the family appears to derive from a pair of motifs (referred to as HKD motifs) most of which contain a highly conserved histidine, lysine, and aspartate residue in the sequence HxK(x)₄D.¹⁸ Structural studies of a Streptomyces PLD and Nuc20., 21. have shown that the aspartate residue of the HKD motif is not located near the active site of the enzyme and is probably not involved in catalysis. Most likely, this aspartate residue is critical for tertiary structure stabilization of certain members of this protein family,²¹ and indeed is not conserved in human Tdp1 or any of its orthologs,¹⁴ nor is it conserved in the Bfi I restriction endonuclease.¹⁹ Analyses of reaction intermediates and products have led to the suggestion that catalysis in the PLD superfamily proceeds with the formation of a phosphoenzyme intermediate that is subsequently cleaved to form the final product.22., 23., 24. Starting from substrate, the presence of a phosphoenzyme intermediate has been demonstrated for Streptomyces PLD²⁵ and for human Tdp1.¹⁴ In two cases, a histidine in one of the HKD motifs has been directly implicated as the nucleophile in the reaction leading to the covalent intermediate.26., 27. The crystal structure of Streptomyces PLD reveals that the conserved histidine and lysine residues of both HKD motifs are clustered to form a single active site.²⁰ Although Nuc contains only a single HKD motif, the active enzyme is dimeric, with a single active site located at the dimer interface.²¹

The three-dimensional structure of a catalytically active N-terminal truncation of human Tdp1 (Δ1–148) was recently solved by X-ray crystallography.¹⁵ The tertiary structure of human Tdp1 can be described as an α–β–α–β–α sandwich composed of two α–β–α domains that are related by a pseudo-twofold axis of symmetry (Figure 1(a)). The N-terminal domain extends from Gly149 to Thr350, and the C-terminal domain runs from Asn351 to Ser608. Each domain has in common a seven-stranded mixed parallel–antiparallel beta sheet and two conserved alpha helices. A single catalytic active site is located along the pseudo-twofold axis of symmetry.

The active site region containing the pairs of highly conserved histidine and lysine residues of the HKD motifs obeys pseudo-twofold symmetry about an axis between the two domains, at a point where the convex sides of the large beta sheets from each domain come in closest contact with one another. The key histidine and lysine residues of the active site of human Tdp1 have been shown by mutagenesis studies to be important for catalysis.¹⁴ His263 is the most important of these four residues, as an H263A mutant is catalytically inactive, whereas H493A, H493N, K495S, and K265S mutants retain limited enzymatic activity.¹⁴ Four additional residues that are conserved in all Tdp1 orthologs, Asn283, Gln294, Asn516, and Glu538, are also located near the active site.

The exact substrate for Tdp1 in the repair of topoisomerase I–DNA covalent complexes in vivo is uncertain, but a complex containing the native form of human topoisomerase I is cleaved poorly if at all by human Tdp1 (H. Interthal & J. Champoux, unpublished results). Similarly, S. cerevisiae Tdp1 is unable to efficiently cleave intact topoisomerase I–DNA complexes.²⁸ This result is understandable in view of the structure of the topoisomerase I–DNA covalent complex, where the tyrosyl-DNA phosphodiester bond is completely buried within the complex.²⁹ Interestingly, heating a trapped human topoisomerase I–DNA complex for 10 minutes at 65 °C transforms the complex into a substrate for human Tdp1 (H. Interthal & J. J. Champoux, unpublished results). Similarly, bacteriophage λ Int protein–DNA complex (which also contains a 3′-DNA phosphotyrosine linkage) is poorly cleaved by yeast Tdp1; however, prior heating of the complex to 65 °C also greatly increased cleavage by the enzyme.¹ These observations suggest that the initial topoisomerase I–DNA covalent complex must be denatured or otherwise processed in some way to make the tyrosine-DNA phosphodiester bond accessible to Tdp1.¹

The human Tdp1 active site is centrally located near the middle of an irregularly shaped cleft that extends approximately 40 Å across an entire surface of the enzyme, nearly perpendicular to the boundary between domains (Figure 1(a)). The two halves of this potential substrate binding cleft on either side of the active site are distinct in terms of both shape and charge distribution. On one side of the active site, the cleft is relatively narrow (approximately 8 Å) and lined with predominantly positively charged amino acid residues. On the other side of the active site, the cleft has a more mixed charge distribution and flares out into a bowl-shaped basin, extending from 8 Å near the active site to as wide as 20 Å. On the basis of the shape and surface charge distribution of the substrate binding cleft we previously proposed that the narrow, positively charged side may act as a DNA-binding region, while the more open side with the more negative charge distribution is better suited to binding the peptide moiety of the substrate.¹⁵ The narrow, positive groove is only large enough to accommodate a single strand of DNA. The larger half of the cleft, although too small to accommodate an intact molecule of topoisomerase I, could easily accommodate a moiety larger than a single tyrosine residue or a short peptide. The putative peptide binding pocket may be sufficient to bind the intact 6.3 kDa C-terminal domain of topoisomerase I that contains the catalytic tyrosine residue.¹⁵

In order to gain additional insight regarding the catalytic mechanism and substrate binding, we solved two crystal structures of Tdp1 bound to the phosphate transition state analogs, vanadate and tungstate. In each case, these inhibitors were found to be covalently bound to His263, implicating this residue as the nucleophile in the first step of the catalytic mechanism. Both structures also reveal a glycerol molecule contributed by the cryoprotectant solution bound to the vanadate or tungstate moieties. To our knowledge, this is the first time that vanadate–glycerol or tungstate–glycerol adducts have been observed in the context of a biological molecule. These structures allow the formulation of a model describing not only the substrate binding mode of Tdp1, but also the catalytic mechanism of the enzyme.