Structure-function relationships in miscoding by Sulfolobus solfataricus DNA polymerase Dpo4: guanine N2,N2-dimethyl substitution produces inactive and miscoding polymerase complexes

Abstract

Previous work has shown that Y-family DNA polymerases tolerate large DNA adducts, but a substantial decrease in catalytic efficiency and fidelity occurs during bypass of N2,N2-dimethyl (Me2)-substituted guanine (N2,N2-Me2G), in contrast to a single methyl substitution. Therefore, it is unclear why the addition of two methyl groups is so disruptive. The presence of N2,N2-Me2G lowered the catalytic efficiency of the model enzyme Sulfolobus solfataricus Dpo4 16,000-fold. Dpo4 inserted dNTPs almost at random during bypass of N2,N2-Me2G, and much of the enzyme was kinetically trapped by an inactive ternary complex when N2,N2-Me2G was present, as judged by a reduced burst amplitude (5% of total enzyme) and kinetic modeling. One crystal structure of Dpo4 with a primer having a 3'-terminal dideoxycytosine (Cdd) opposite template N2,N2-Me2G in a post-insertion position showed Cdd folded back into the minor groove, as a catalytically incompetent complex. A second crystal had two unique orientations for the primer terminal Cdd as follows: (i) flipped into the minor groove and (ii) a long pairing with N2,N2-Me2G in which one hydrogen bond exists between the O-2 atom of Cdd and the N-1 atom of N2,N2-Me2G, with a second water-mediated hydrogen bond between the N-3 atom of Cdd and the O-6 atom of N2,N2-Me2G. A crystal structure of Dpo4 with dTTP opposite template N2,N2-Me2G revealed a wobble orientation. Collectively, these results explain, in a detailed manner, the basis for the reduced efficiency and fidelity of Dpo4-catalyzed bypass of N2,N2-Me2G compared with mono-substituted N2-alkyl G adducts.

FIGURE 1.

N²-MeG and N²,N²-Me₂G.

FIGURE 2.

Extension of ³²P-labeled primer opposite N²,N²-Me₂G by Dpo4 in the presence of all four dNTPs. A 24-mer primer/36-mer N²,N²-Me₂G template complex (100 nm) was extended using Dpo4 (0, 0.5, 2, and 10 nm).

FIGURE 3.

Crystal structures of Dpo4 bound to N²,N²-Me₂G-modified DNA. A, superimpositions of DMG-1 (red), DMG-2 (molecule A, cyan), and DMG-3 (molecule A, green) reveal overall similarity in Dpo4 structure. B, representative electron density near the active site of Dpo4 in the DMG-1 structure. The 3F_o− 2F_c map (gray mesh) is shown contoured to the 1σ level. The F_o− F_c difference maps are shown contoured to 3σ and −3σ for positive (red mesh) and negative (green mesh) density, respectively. The terminal C_dd residue is flipped out of base-stacking orientation. but the incoming dGTP forms a Watson-Crick pair with the cytosine to the 5′-side of N²,N²-Me₂G (DMG).

FIGURE 4.

Structural examination of wobble pairing with N²,N²-Me₂G-modified DNA in Dpo4. A, overall structure of DNA and corresponding electron density observed in the active site of DMG-2 are shown. The orientation of bases observed in molecule A (B) and molecule B (C) of the DMG-2 structure is shown. D, overall structure of DNA and corresponding electron density observed in the active site of DMG-2 are shown. The orientation of bases observed in molecule A (E) and molecule B (F) of the DMG-2 structure is shown. In all panels, the 3F_o− 2F_c map (gray mesh) for DNA bound in the DMG-2 structure is shown contoured to 1σ level (gray mesh) with the F_o− F_c difference maps shown contoured to 3σ and −3σ for positive (red mesh) and negative (green mesh) density, respectively.

FIGURE 5.

Comparison between catalytically inhibited Dpo4 N²,N²-Me₂G wobble pairs and the highly efficient 8-oxoG:dCTP structure (68, 69). A, orientation of wobble base pairs observed in DMG-2 (cyan carbons) and DMG-3 (green carbons) is shown with the active-site residues and Mg²⁺ ions labeled accordingly. B, rotated view of A illustrating the widened C-1′–C-1′ distance observed for the N²,N²-Me₂G:C_dd wobble relative to N²,N²-Me₂G:dTTP. C, superimposing molecules A (base pair and metal ions in yellow, everything else in green) and B (base pair and metal ions in red, everything else in green) from the DMG-3 structure with a ternary structure of Dpo4 inserting dCTP opposite 8-oxoG (base pair and metal ions in blue, everything else in gray; PDB code 2c2e) shows that the catalytic metal ion is shifted in N²,N²-Me₂G:dTTP pairing events. D, rotated view of C shows how the wobble pairing mode shifts the incoming dNTP away from the active site of Dpo4.

FIGURE 6.

Incorporation of dCTP, (S_p)-dCTPαS, and dTTP opposite G and N²,N²-Me₂G by Dpo4. A, pre-steady-state incorporation of dCTP (1 mm, ■) or (S_p)-dCTPαS (1 mm, ●) into a 24-mer primer opposite a 36-mer containing N²,N²-Me₂G (120 nm) by Dpo4 (70 nm). The burst rates (k_p) were estimated to be 1.5 (± 0.2) s⁻¹ for dCTP and 1.1 (± 0.1) s⁻¹ for dCTPαS. B, incorporation of dTTP (1 mm) opposite 36-mer at the G (■) and N²,N²-Me₂G (●) sites by Dpo4 (70 nm). No bursts were observed, and the data points were fit to a linear equation, with rates estimated as 0.46 (± 0.01) s⁻¹ for the G oligonucleotide and 0.026 (± 0.004) s⁻¹ for the N²,N²-Me₂G oligonucleotide.

FIGURE 7.

Estimation of k_pol and K_d,_dCTP for by pre-steady-state burst kinetic analysis. Single turnover experiments were done with Dpo4 (200 nm) and 100 nm 24-mer primer/36-mer template complex containing G (A) or N²,N²-Me₂G (B) and varying dCTP concentrations (2–1800 μm). Burst rates (k_obs, fit to Equation 2) versus [dCTP] were fit to a hyperbolic equation (Equation 3). (Data from the experiments with N²-MeG are not shown.) The burst amplitudes for G, N²-MeG, and N²,N²-Me₂G were 42, 71, and 3 nm, respectively. Estimated values of k_pol for G, N²-MeG, and N²,N²-Me₂G were 3.0, 2.8, and 3.4 s⁻¹, respectively. Estimated K_d,_dCTP values for G, N²-MeG, and N²,N²-Me₂G were 10, 9, and 220 μm, respectively.

FIGURE 8.

Kinetic simulations of bursts. A, general DNA polymerase mechanism. Individual steps are numbered. E, polymerase; D_n:DNA substrate; E*, conformationally modified polymerase; E^#, inactive polymerase conformation; D_n+1, DNA extended by one base, and PP_i, pyrophosphate. Forward and reverse rate constants for each step are presented from simulations of pre-steady-state nucleotide incorporation reactions with unmodified substrate. See Table 6 for adjustments of k₋₂, k₄, k₈, and k₋₈. All other rate constants remain the same in B–D. B, simulated fit to G data points. C, simulated fit to N²-MeG data points. D, simulated fit to N²,N²-Me₂G data points.