Mechanism of template-independent nucleotide incorporation catalyzed by a template-dependent DNA polymerase

Abstract

Numerous template-dependent DNA polymerases are capable of catalyzing template-independent nucleotide additions onto blunt-end DNA. Such non-canonical activity has been hypothesized to increase the genomic hypermutability of retroviruses including human immunodeficiency viruses. Here, we employed pre-steady state kinetics and X-ray crystallography to establish a mechanism for blunt-end additions catalyzed by Sulfolobus solfataricus Dpo4. Our kinetic studies indicated that the first blunt-end dATP incorporation was 80-fold more efficient than the second, and among natural deoxynucleotides, dATP was the preferred substrate due to its stronger intrahelical base-stacking ability. Such base-stacking contributions are supported by the 41-fold higher ground-state binding affinity of a nucleotide analog, pyrene nucleoside 5'-triphosphate, which lacks hydrogen bonding ability but possesses four conjugated aromatic rings. A 2.05 A resolution structure of Dpo4*(blunt-end DNA)*ddATP revealed that the base and sugar of the incoming ddATP, respectively, stack against the 5'-base of the opposite strand and the 3'-base of the elongating strand. This unprecedented base-stacking pattern can be applied to subsequent blunt-end additions only if all incorporated dAMPs are extrahelical, leading to predominantly single non-templated dATP incorporation.

Figure 1

Gel mobility shift assay to determine Dpo4•B-1 equilibrium dissociation constant. (a) Reactions containing 100 nM 5′-[³²P]B-1 were incubated with increasing concentration of Dpo4 as indicated in the gel picture followed by native gel analysis to resolve binary complex from unbound B-1. (b) Binary complex formation [Dpo4•B-1] was plotted against Dpo4 concentration and the resulting data points were fit via quadratic regression (equation 3) to yield a K_d of 38±3 nM.

Figure 2

Series of polyacrylamide gel pictures showing time courses of product formation under the following conditions either onto B-1: (a) 200 μM dATP, (b) 1,300 μM dCTP, (c) 1,300 μM dGTP, (d) 400 μM dTTP, (e) 2 μM dPTP, or onto the B-1A substrate, (f) 1,300 μM dATP. Remaining primer, 21-mer in (a) – (e), 22-mer in (f), is shown at the bottom of each gel image with extended products located sequentially above it. Reaction time (minutes) is denoted below the corresponding lane.

Figure 3

Concentration dependence on the rate of dATP and dPTP incorporation onto B-1. A preincubated solution of Dpo4 (120 nM) and 5′-[³²P]B-1 (30 nM) was mixed with an increasing concentration of dNTP•Mg²⁺ for various times. The individual reactions were then quenched by 0.37 M EDTA. (a) The single exponential rates for each individual time course were plotted as a function of dATP concentration. The rate data were then fit to the hyperbolic equation (equation 2) yielding a k_p of 0.0035(±0.0003) s⁻¹ and a K_d of 571(±132) μM. (b) Likewise, the dependence of dPTP concentration on the single-turnover rate onto B-1 was plotted, as above, to yield a k_p of 0.0085(±0.0004) s⁻¹ and a K_d of 14(±2) μM.

Figure 4

Crystal structure of Dpo4•blunt-end X-1•ddATP (2.05 Å). (a) Overall ternary structure. Dpo4 was shown in grey ribbons while DNA and ddATP were shown as ball-and-stick models. The ddATP is highlighted in magenta. The Ca²⁺ ion was shown in a green sphere. (b) The zoomed in view of the active site including ddATP and the blunt-end base-pair. The residues in contact with ddATP were shown as ball-and-stick models (grey for atom C, red for atom O, yellow for atom S). Only the side chain and main chain atoms involved were shown. (c) 2F_o - F_c electron density map contoured at 1.2 σ (light-blue) of the active site. The amino acid residues, two blunt-end DNA base-pairs, and incoming ddATP were shown as ball-and-stick models.

Figure 5

Proposed mechanisms of blunt-end additions of (a) dPTP and (b) dATP. dATP and dPTP are represented by A and P in different colors, respectively. The Watson-Crick hydrogen bonds were drawn as dashed lines while the base-stacking interactions were shadowed in green. The stacking interactions between the 2′-deoxyribose (R) of an incoming nucleotide and the 5′-terminal base A are displayed in a green box. The van der Waals interactions between an incoming nucleotide and Dpo4 active site residues were not shown for clarity.

Figure 6

Comparison of the incoming nucleotide (magenta) positions in the active sites of three different Dpo4 ternary structures. The nucleotides were shown in the side (upper panel) and top (lower panel) views. (a) Type I (ddADP: T); (b) Blunt-end (ddATP:DNA); and (c) Mismatched T/G-1 (dGTP:T).

Scheme 1