Structural insight into dynamic bypass of the major cisplatin-DNA adduct by Y-family polymerase Dpo4

Abstract

Y-family DNA polymerases bypass Pt-GG, the cisplatin-DNA double-base lesion, contributing to the cisplatin resistance in tumour cells. To reveal the mechanism, we determined three structures of the Y-family DNA polymerase, Dpo4, in complex with Pt-GG DNA. The crystallographic snapshots show three stages of lesion bypass: the nucleotide insertions opposite the 3'G (first insertion) and 5'G (second insertion) of Pt-GG, and the primer extension beyond the lesion site. We observed a dynamic process, in which the lesion was converted from an open and angular conformation at the first insertion to a depressed and nearly parallel conformation at the subsequent reaction stages to fit into the active site of Dpo4. The DNA translocation-coupled conformational change may account for additional inhibition on the second insertion reaction. The structures illustrate that Pt-GG disturbs the replicating base pair in the active site, which reduces the catalytic efficiency and fidelity. The in vivo relevance of Dpo4-mediated Pt-GG bypass was addressed by a dpo-4 knockout strain of Sulfolobus solfataricus, which exhibits enhanced sensitivity to cisplatin and proteomic alterations consistent with genomic stress.

Figure 1

The structures of the GG1 (A), GG2 (B), and GG3 (C) Dpo4–DNA–dNTP ternary complexes. The finger, palm, thumb, and little finger domains distinguished by their respective colours in (B). The Pt-GG lesioned template is represented in magenta, with the platinum atom shown as a cyan ball. The zoomed in boxes of the active site are covered with the Pt-anomalous maps (5 σ) in orange. The catalytic residues are shown in the middle panel of Figure 1C, which present invariantly in all three structures. There are two base conformers of Pt-GG in (B), with each Pt atom at 0.5 occupancy. The top views of the replicating base pairs are covered by a blue 2Fo-Fc maps contoured to 1.0 σ at 2.9, 1.9, and 2.0 Å resolution, respectively. The A-T pair in green sticks superimposed with GG1 (A) is taken from a Dpo4–DNA–dNTP complex structure (PDB: 1S0O), which depicts the regular position of an undamaged purine-pyrimidine base pair in the Dpo4 active site. The grey nucleotide in (A, middle box) is a ghost model for the 3′ primer base that is disordered in GG1. The green spheres are Ca²⁺ ions. (D) A top view comparison of the GG1 (blue), GG2 (red), and GG3 (cyan) replicating base pairs with an undamaged pyrimindine-purine (T-A) replicating base pair (2AGQ, black). (E) A top view comparison of the Dpo4 GG1 base pair with the yeast pol η GG1 (2R8J, beige) and an undamaged purine-pyrimidine (A-T) (1S0O, green) replicating base pairs.

Figure 2

Pt-GG adducts in the active sites of Dpo4 and ypolη. All panels are close-up views of the enzyme active sites where the finger domain is in cyan. The platinum (Pt) atom is shown as a cyan ball. The DNA is shown as stick-balls, with the template strand in orange and the primer strand in yellow. The sequences of the template/primer DNA in the complexes are shown on the tops of panels (A–D), cross-linked G bases are in red. The arrows indicate the incoming dNTP positions. (A) GG1, where dCTP (pink) is paired opposite the 3′G of Pt-GG (dark pink). (B, C) GG2, two alternate conformations of GG2 (translucent pink, GG2a in (B), GG2b in (C)), where dCTP (pink) is paired opposite the 5′G of Pt-GG. (D) GG3, where dATP is paired with the T base that is 5′ to the Pt-GG lesion. (E) Superposition of the GG1 (semi-transparent, left side) and GG2 (solid, right) structures of Dpo4, where DNA is shown as grey sticks with the 5′G of Pt-GG is in red with arrows, and the 3′G is in blue. The Pt-GG adduct is translocated in different positions and shows different conformations in the two structures. (F) Superposition of the same reaction stage complexes of ypolη with the same colour scheme as (E). In ypolη (F), the Pt-GG adduct remains in a similar position with the 5′G bases (red bases with an arrow) superimposed well in both reaction stages, no translocation of Pt-GG occurs.

Figure 3

Pt-GG conformations and helical DNA structures in the presence or absence of Dpo4 enzyme. (A) Alternate conformations of Pt-GG in the GG2 structure. The alternate Pt atom corresponds to discrete bulges of electron density for the 1.9 Å 2Fo-Fc map contoured at 1σ. (B–D) Comparison of the DNA helices in GG2 (B), GG3 (C), and enzyme-free Pt-GG structure (D) (PDB: 1AIO). The Pt-GG in the enzyme-free DNA induces a bend in the DNA helix.

Figure 4

Specificity of Dpo4-mediated nucleotide incorporation was tested for the (A) GG1 (first insertion), (B) GG2 (second insertion), and (C) GG3 (extension) reaction stages with a dNTP (A, C, G, or T). M represents the marker for the primer strand. In each panel, nucleotide incorporation assays were terminated after 10 min for undamaged DNA (left) and after 60 min for Pt-GG DNA (right).

Figure 5

Analysis of wild-type and dpo-4 mutant proteomes. Proteins from cell extracts prepared from exponentially growing cell lines, were fractionated by 2D SDS–PAGE and their identities were determined by tandem MS/MS sequencing of peptides derived by trypsin hydrolysis, with subsequent matching to the S. solfataricus genome sequence. Values for isoelectric point (pI) and mass (MW) are indicated across the top and sides of the figure, respectively. Proteins with increased abundance in the dpo-4 mutant are circled in the dpo-4 panel; proteins with decreased abundance are circled in the wild-type panel. Proteins are numbered in correspondence with Supplementary Table S2.