R964C Mutation of DNA Polymerase γ Imparts Increased Stavudine Toxicity by Decreasing Nucleoside Analog Discrimination and Impairing Polymerase Activity

Abstract

The R964C mutation of human DNA polymerase γ was recently linked to stavudine (d4T)-mediated mitochondrial toxicity. We utilized pre-steady-state kinetics to determine the effect of this mutation on incorporation of natural substrate dTTP and the active metabolite of d4T (d4TTP). The R964C polymerase γ holoenzyme demonstrated a 33% decrease in dTTP incorporation efficiency and a threefold-lower d4TTP discrimination relative to that of the wild-type polymerase γ, providing a mechanistic basis for genetic predisposition to nucleoside reverse transcriptase inhibitor toxicity.

Nucleoside reverse transcriptase inhibitors (NRTIs) are a cornerstone of highly active antiretroviral therapy in the treatment of human immunodeficiency virus infection (8), although side effects remain a prominent issue for prolonged treatment. Toxicity of NRTIs, including development of myopathies, lactic acidosis, and peripheral neuropathy (2), has been associated with inhibition of human mitochondrial DNA polymerase γ (Pol γ) (5, 9, 13, 14). Recently, a novel autosomal recessive Pol γ mutation, arginine 964 to cysteine (R964C), was hypothesized to impart a predisposition to stavudine (d4T)-induced mitochondrial toxicity (22) (Fig. 1A). This study aims to elucidate the molecular mechanism of increased mitochondrial toxicity by utilizing an in-depth kinetic approach.

FIG. 1.

(A) Structure of d4T. (B) R964C mutant Pol γ holoenzyme demonstrates a threefold decrease in d4TTP discrimination compared to that of the WT. (C) Molecular model of the Pol γ active site (7). R964 is shown in magenta; O and O1 helices and Y951 are highlighted in yellow. ddCTP is shown bound in the active site (gray) shown with Mg²⁺ ions coordinated with catalytic residues D1135 and E1136 (green). Note the position of the hydroxyl group of Y951, near where the 3′ OH of a natural dNTP would be located. Selected residues mutated in autosomal dominant progressive external ophthalmoplegia are shown in cyan (R943H, Y955C, and A957S) (reviewed in reference 3).

Initial biochemical studies with R964C Pol γ suggested that the mutation impaired steady-state polymerization of the natural substrate dTTP with no change in d4TTP inhibition in steady-state competition assays. Therefore, it was hypothesized that this slight impairment of Pol γ catalytic activity resulted in no observed clinical symptoms, requiring further challenge with d4T treatment to result in mitochondrial toxicity (22). Moreover, R964C has recently been found in trans with the A862T mutation in a patient with ataxia-neuropathy syndrome, indicating that this mutation may impair Pol γ catalysis (20). However, the biochemical experiments by Yamanaka et al. used wild-type (WT) and R964C exonuclease-proficient catalytic subunits in the absence of accessory subunits. The physiologically relevant form of Pol γ is a holoenzyme complex consisting of one catalytic subunit bound to two accessory subunits (21), an interaction essential for processive polymerization (10, 15). Furthermore, mechanistic studies of Pol γ inhibition by NRTIs typically utilize the exonuclease-deficient holoenzyme, since the exonuclease rate may complicate the kinetics of NRTI incorporation.

As such, WT and R964C Pol γ catalytic subunits (exonuclease deficient) and the accessory subunit were expressed, purified, and reconstituted as described previously (10, 15, 17). Initial determination of steady-state kinetic parameters for dTTP incorporation by the mutant and WT holoenzyme was carried out as described previously (17). Incorporation of various concentrations of [α-³²P]dTTP into a poly(rA)·oligo(dT)_12-18 primer template by the WT and R964C Pol γ holoenzyme was measured by using liquid scintillation counting of trichloroacetic acid-insoluble radioactivity. Our steady-state kinetic analyses confirmed that the R964C Pol γ holoenzyme demonstrates a fivefold decrease in steady-state incorporation efficiency of dTTP compared to that of the WT (data not shown), consistent with previous observations demonstrating a ninefold decrease in polymerase activity for the catalytic subunit alone (22). Since steady-state kinetic analysis reflects only the overall rate-limiting step in Pol γ polymerization, release of the elongated primer template (6), a more detailed approach is required to investigate how the R964C mutation may impact the ability of the enzyme to discriminate between natural nucleotide substrate and nucleotide analogs such as d4TTP.

To understand the mechanism for increased d4T toxicity, we employed a pre-steady-state kinetic approach to provide insight into the direct interaction between deoxynucleoside triphosphate (dNTP) and the active site of the WT and mutant Pol γ holoenzyme. Pre-steady-state kinetics measures rate-limiting steps prior to product release by monitoring the first enzyme turnover of substrate and yields the parameters K_d (binding affinity), k_pol (maximum rate of polymerization), and k_pol/K_d (overall incorporation efficiency) (11, 12). Single-turnover experiments, in which the enzyme is in excess over the substrate, were performed for d4TTP incorporation, whereas burst experiments, in which the substrate is in slight excess over the enzyme, were carried out for dTTP incorporation as described previously (18). Experiments were performed using a KinTek Instruments model RQF-3 rapid quench-flow apparatus to allow rapid mixing of reactants on the millisecond time scale. Incorporation of dTTP and d4TTP by the WT and R964C Pol γ holoenzyme was examined by monitoring incorporation into 5′-radiolabeled primer templates. Reaction products were subjected to 20% polyacrylamide gel electrophoresis and were quantitated using the Bio-Rad molecular imager FX. Observed rates were then plotted as a function of dNTP concentration and fit to a hyperbola using KaleidaGraph (Synergy).

Interestingly, our pre-steady-state kinetic studies revealed that the R964C Pol γ binding affinity for dTTP was similar to that of the WT and that the rate of incorporation was only slightly lower, resulting in a small decrease in dTTP incorporation efficiency (33%) compared to that of the WT (Table 1). On the other hand, the rates of d4TTP incorporation were nearly identical for the mutant and the WT, whereas the binding affinity of d4TTP was approximately twofold tighter for the mutant than for the WT (120 nM versus 270 nM). This resulted in a twofold increase in d4TTP incorporation efficiency by the R964C mutant. Importantly, this analysis reveals a loss of selectivity for the natural dTTP substrate. The R964C mutation results in a threefold decrease in discrimination (defined as Efficiency_dTTP/Efficiency_d4TTP) compared to that of the WT, implying a higher propensity to incorporate d4TTP relative to the WT polymerase (Fig. 1B).

TABLE 1.

Pre-steady-state kinetic parameters for dNTP incorporation^c

Polymerase γ	Nucleotide	kpol (s−1)	Kd (μM)	Efficiency (μM−1 s−1)a	Discriminationb
WT	dTTP	202.6 ± 8.7	1.9 ± 0.3	108.4 ± 17.9
	d4TTP	0.39 ± 0.01	0.27 ± 0.03	1.5 ± 0.2	73
R964C	dTTP	145.8 ± 8.7	2.0 ± 0.4	72.4 ± 14.5
	d4TTP	0.40 ± 0.01	0.12 ± 0.01	3.2 ± 0.3	23

^aEfficiency = k_pol/K_d.

^bDiscrimination = Efficiency_dTTP/Efficiency_d4TTP.

^cError values represent deviance of points from the curve fit as determined by KaleidaGraph.

A physical mechanism of this decrease in discrimination could be explained by the position of R964. R964 resides on the O1 helix, which is connected to the highly conserved O helix that directly interacts with bound nucleotides. Several residues implicated in the mitochondrial disorder progressive external ophthalmoplegia also reside on or near the O helix (3). Y951 on the O helix has been associated with dideoxynucleotide sensitivity of Pol γ and other family A polymerases via mimicking of the 3′ hydroxyl of a natural dNTP (1, 16, 19) (Fig. 1C). Therefore, the R964C mutation could affect the orientation of the O helix, modulating the position of Y951 to allow tighter binding to d4TTP.

This report suggests d4T-induced mitochondrial toxicity of patients exhibiting the R964C Pol γ mutation may be caused by decreased d4TTP discrimination over the natural substrate in addition to impaired polymerase function. Further study is required to identify additional mutations or polymorphisms of the mitochondrial polymerase linked to NRTI toxicity, such as the E1143G polymorphism recently linked to d4T lipodystrophy (4). We believe this represents a promising first step toward using pharmacogenetics for guiding NRTI therapy in the treatment of human immunodeficiency virus infection.