The DNA binding and 3'-end preferential activity of human tyrosyl-DNA phosphodiesterase

Abstract

Human tyrosyl-DNA phosphodiesterase (Tdp1) processes 3'-blocking lesions, predominantly 3'-phosphotyrosyl bonds resulting from the trapping of topoisomerase I (Top1) cleavage complexes. The controversial ability of yeast Tdp1 to hydrolyze 5'-phosphotyrosyl linkage between topoisomerase II (Top2) and DNA raises the question whether human Tdp1 possesses 5'-end processing activity. Here we characterize the end-binding and cleavage preference of human Tdp1 using single-stranded 5'- and 3'-fluorescein-labeled oligonucleotides. We establish 3'-fluorescein as an efficient surrogate substrate for human Tdp1, provided it is attached to the DNA by a phosphodiester (but not a phosphorothioate) linkage. We demonstrate that human Tdp1 lacks the ability to hydrolyze a phosphodiester linked 5'-fluorescein. Using both fluorescence anisotropy and time-resolved fluorescence quenching techniques, we also show the preferential binding of human Tdp1 to the 3'-end. However, DNA binding competition experiments indicate that human Tdp1 binding is dependent on DNA length rather than number of DNA ends. Lastly, using surface plasmon resonance, we show that human Tdp1 selectively binds the 3'-end of DNA. Together, our results suggest human Tdp1 may act using a scanning mechanism, in which Tdp1 bind non-specifically upstream of a 3'-blocking lesion and is preferentially stabilized at 3'-DNA ends corresponding to its site of action.

Figure 1.

Processing of fluorescein-labeled oligonucleotides by recombinant human Tdp1. (A) Representative denaturing PAGE gel showing Tdp1-mediated conversion of 5′-end radiolabeled 14-mer oligonucleotides containing a 3′-tyrosine (p14Y) or 3′-fluorescein (p14F) to a 3′-phosphate product (p14p). The chemical structures of the 3′-tyrosine and 3′-6-FAM are shown above the gel. (B) Representative denaturing PAGE gel showing the generation of an unknown product (indicated by question mark) from Tdp1 cleavage of a 3′-end radiolabeled 14-mer oligonucleotide containing a 5′-fluorescien (F14pA). Tdp1 concentrations in (A) and (B) are as follows: 0.01 nM (lanes 2 and 8), 0.1 nM (lanes 3 and 9), 1 nM (lanes 4 and 10), 10 nM (lanes 5 and 11) and 100 nM (lanes 6 and 12). (C) 3′-end processing of the F14pA oligonucleotide [from (B)] by the combinations of Tdp1 (100 nM) and wild-type T4 PNK [PNK (1 unit), lane 5] or Tdp1 and 3′-phosphatase-minus mutant T4 PNK [PNK⁻ (1 unit), lane 7]. (D) 3′-end processing of the different substrates. (E) Schematic illustrating the substrates and products of the sequential action of Tdp1 and PNK on the F14pA oligonucleotide. In all instances the asterisk indicates the ³²P-labeled phosphate.

Figure 2.

Equilibrium binding of recombinant human Tdp1 to fluorescein-labeled oligonucleotides using fluorescence anisotropy. (A) Schematic diagram showing the principle of fluorescence anisotropy. (B) Fluorescence anisotropic signal for 14F (filled squares), F14 (filled circles) and F14p (open circles) as a function of Tdp1 concentration. Each point of the graph represents the mean (n = 3) ±95% confidence interval (error bars). (C) Analysis of the competition between oligonucleotides in solution using fluorescence anisotropy. Representative plots of Tdp1 binding competition between F14p and unlabeled 14p (open circles) or 50p (filled diamonds).

Figure 3.

Binding and processing of phosphorothioate-linked fluorescein containing oligonucleotides by recombinant human Tdp1. (A) Simplified representation of the quenched and unquenched states of the terminal fluorescein in the absence and presence of Tdp1. (B) Representative denaturing PAGE gel showing the hydrolyzability and non-hydrolyzability of a 3′-fluorescein attached via a phosphodiester linkage (15F) and phosphorothioate linkage [15(pS)F] by Tdp1, respectively. (C) Graphical representation of the relative change in the fluorescence lifetime of the unquenched state for F(pS)15 (filled circles) and 15(pS)F (open circles) as a function of Tdp1 concentration. (D) Fluorescence anisotropic signal for 15F (filled squares), 15(pS)F (open squares), F15 (filled circles) and F(pS)15 (open circles) as a function of Tdp1 concentration. Each point on the graph represents the mean (n = 3) ±95% confidence interval (error bars).

Figure 4.

Differential binding of recombinant human Tdp1 to 3′- and 5′-phosphate terminated oligonucleotides using SPR. SPR sensorgrams showing the interaction of Tdp1 with the (A) B14p oligonucleotide and (B) p14B oligonucleotide (see sequences above). Nine successive 2-fold dilutions of Tdp1 from 500 nM were tested for each oligonucleotide. The inset in (A) shows the equilibrium binding of Tdp1 to B14p fitted to the 1: 1 Langmuir binding model.