MutT homologue 1 (MTH1) catalyzes the hydrolysis of mutagenic O6-methyl-dGTP

Abstract

Nucleotides in the free pool are more susceptible to nonenzymatic methylation than those protected in the DNA double helix. Methylated nucleotides like O6-methyl-dGTP can be mutagenic and toxic if incorporated into DNA. Removal of methylated nucleotides from the nucleotide pool may therefore be important to maintain genome integrity. We show that MutT homologue 1 (MTH1) efficiently catalyzes the hydrolysis of O6-methyl-dGTP with a catalytic efficiency similar to that for 8-oxo-dGTP. O6-methyl-dGTP activity is exclusive to MTH1 among human NUDIX proteins and conserved through evolution but not found in bacterial MutT. We present a high resolution crystal structure of human and zebrafish MTH1 in complex with O6-methyl-dGMP. By microinjecting fertilized zebrafish eggs with O6-methyl-dGTP and inhibiting MTH1 we demonstrate that survival is dependent on active MTH1 in vivo. O6-methyl-dG levels are higher in DNA extracted from zebrafish embryos microinjected with O6-methyl-dGTP and inhibition of O6-methylguanine-DNA methyl transferase (MGMT) increases the toxicity of O6-methyl-dGTP demonstrating that O6-methyl-dGTP is incorporated into DNA. MTH1 deficiency sensitizes human cells to the alkylating agent Temozolomide, a sensitization that is more pronounced upon MGMT inhibition. These results expand the cellular MTH1 function and suggests MTH1 also is important for removal of methylated nucleotides from the nucleotide pool.

Figure 1.

MTH1 is an efficient catalyst of O6-methyl-dGTP hydrolysis. (A) Activity assessment of MTH1 with O6-methyl-dGTP and N2-methyl-dGTP in comparison to dGTP and 8-oxo-dGTP. Activity of 5 nM MTH1 was tested with 50 μM substrate in MTH1 reaction buffer with PPase (0.2 U/ml). Formed Pi was detected using the malachite green reagent. Activity differences between samples in quadruplicate were found to be statistically significant by multiple comparisons using One way Anova in the GraphPad Prism 6.0 software. (B) Activity of MTH1 (45 nM) with 50 μM O6-methyl-GTP and GTP was measured as in (A) with samples in quadruplicate. Statistic significance was analysed using Paired Two-tailed T-test using the GraphPad Prism 6.0 software. (C) Saturation curve for MTH1 with O6-methyl-dGTP were produced by determining initial rates using 1.25 nM MTH1 and O6-methyl-dGTP ranging in concentration between 0 and 40 μM. (D) Saturation curve for MTH1 with O6-methyl-GTP. 50 nM MTH1 and O6-methyl-GTP ranging from 0 to 400 μM were used. Shown are representative saturation curves out of two independent experiments for each substrate with data points recorded in duplicate. P ≤ 0.05 are considered to be statistically significant and are indicated by *, P ≤ 0.01 are indicated by **, P ≤ 0.001 are indicated by *** and P ≤ 0.0001 are indicated by ****.

Figure 2.

Hydrolysis activity of MTH1 with O6-methyl-dGTP is exclusive among NUDIX hydrolases. Activity screen of human NUDIX proteins with 50 μM O6-methyl-dGTP was tested using 0, 5, or 200 nM NUDIX enzyme in presence of an excess of PPase (A) monitoring formation of Pi and PPi, or without coupled enzyme (B) detecting formation of Pi. Pi was detected using malachite green reagent and measurement of absorbance at 630 nm. Data points were recorded in triplicate. Statistically significant differences in activity compared to the mock control was assessed using multiple comparison and two-way ANOVA in GraphPad Prism 6.0. P ≤ 0.05 are considered to be statistically significant and are indicated by *, P ≤ 0.01 are indicated by **, P ≤ 0.001 are indicated by *** and P ≤ 0.0001 are indicated by ****.

Figure 3.

Activity of MTH1, NUDT17 and NUDT18 with methylated and nonmethylated nucleotide. (A) MTH1 (5 nM) and (B) NUDT17 (200 nM) and NUDT18 (200 nM) were screened for activity with a panel of methylated and nonmethylated deoxyribonucleotides and ribonucleotides by incubation with 50 μM nucleotide at 22°C in MTH1 reaction buffer for 30 min. Formed Pi was detected using the Malachite green assay. Graphs show average and SEM of [Pi] (μM) per min per enzyme concentration ([E_tot]) (μM) from two independent experiments with data points in triplicate. Statistically significant differences between activity with nonmethylated and methylated nucleotides were determined using two-way ANOVA.

Figure 4.

Activity of MTH1 with O6-methyl-dGTP has been conserved through evolution. (A) Comparison of activities of MTH1 (NUDT1) (1.5 nM) from different species with 75 μM 8-oxo-dGTP and 75 μM O6-methyl-dGTP (hsNUDT1, human NUDT1; mmNUDT1, mouse NUDT1; rnNUDT1, rat NUDT1; ssNUDT1, pig NUDT1; clNUDT1, dog NUDT1; atNUDX1, Arabidopsis thaliana NUDT1; zfNUDT1, zebrafish MTH1; MutT, E. coli MutT). Reaction was performed in MTH1 reaction buffer and reaction time was 15 min. Formed PPi was detected using PPiLight Inorganic Pyrophosphate Assay kit from Lonza. Data points were recorded in triplicate. (B) Ratio between activities with O6-methyl-dGTP and 8-oxo-dGTP for MTH1 from different species.

Figure 5.

Crystal structures of hMTH1 and zfMTH1 in complex with O6-methyl-dGMP. (A) Electron density of O6-methyl-dGMP bound to human MTH1: 2Fo-Fc map at 2.0 σ. hMTH1 is shown in off-white and O6-methyl-dGMP in green. Important binding residues and residues of the hydrophobic pocket are shown as sticks and labelled. Hydrogen bonding distances in Angstroms are shown. (B) Electron density of O6-methyl-dGMP bound to zebrafish MTH1: 2Fo-Fc map at 2.0 σ. zfMTH1 is shown in purple and O6-methyl-dGMP in cyan. Important binding residues and residues of the hydrophobic pocket are shown as sticks and are labelled and hydrogen bonding distances in Angstroms are indicated. (C) Comparison of O6-methyl-dGMP and 8-oxo-dGMP bound to human MTH1. Human MTH1 is shown in off-white for the O6-methyl-dGMP bound structure and in pink for the 8-oxo-dGMP bound structure. O6-methyl-dGMP is shown in green and 8-oxo-dGMP is shown in orange. Important binding residues and residues of the hydrophobic pocket are shown as sticks and labelled. The E52, E56 and E100 binding the magnesium critical for the catalytic activity are also labelled and shown as sticks.

Figure 6.

Active zfMTH1 is crucial for zebrafish embryo survival after O6-methyl-dGTP exposure. (A) The zfMTH1 enzyme catalyzes the hydrolysis of O6-methyl-dGTP efficiently. 100 μM dGTP or O6-methyl-dGTP was incubated with 5 nM zfMTH1 or hMTH1 for 20 min. Formed PPi was converted to Pi by using an excess of E. coli PPase and Pi was detected using malachite green reagent. Statistical significance was determined using multiple comparison and Two way Anova using the GraphPad Prism 6.0 software. (B) O6-methyl-dGTP (150 μM) was injected into fertilized zebra fish eggs followed by treatment with DMSO, TH588 (1.5 μM) or TH1579 (1.5 μM). Picture shows zebrafish embryos from a representative experiment. (C) Quantification of zebrafish survival. Inhibition of zfMTH1 in combination with microinjecting O6-methyl-dGTP in zebrafish is clearly toxic to fish embryos. Graph shows average and standard deviations from three independent experiments. (D) Levels of O6-methyl-dG per million dN in DNA as measured by LC–MS/MS. DNA was extracted from DMSO or TH588 treated zebrafish embryos, zebrafish embryos microinjected with O6-methyl-dGTP or from O6-methyl-dGTP microinjected and TH588 treated zebrafish embryos. Graph shows mean and SEM from two independent experiments. Statistic significance in C and D was tested using multiple comparisons and One way Anova, P ≤ 0.05 are indicated by *. (E) Percentage dead embryos after microinjection of O6-methyl-dGTP and inhibition of zfMTH1 and MGMT through treatment with TH588 (1.5 μM) and Lomeguatrib (10 μM), alone and in combination, compared to untreated zebrafish embryos. Graph shows average ± SD from three independent experiments. (F) Co-treatment of O6-methyl-dGTP injected zebrafish eggs with TH588 and Lomeguatrib significantly decreases the survival of zebrafish embryos compared to the effects of the combined individual treatments.

Figure 7.

MTH1 deficiency sensitizes cells to Temozolomide. Cell viability of U251 (MTH1 proficient) and U251-MTH1 (MTH1 deficient) was measured using resazurin after 96 h of treatment with Temozolomide and Lomeguatrib. (A) EC₅₀ values for Temozolomide was determined to 62±4 μM for MTH1 proficient and 17±4 μM for MTH1 deficient cells with active MGMT, respectively. (B) EC₅₀ values for Temozolomide were 79±2 μM for MTH1 proficient and 11±2 μM for MTH1 deficient cells when MGMT is inhibited with 12.5 μM Lomeguatrib. MTH1 deficiency causes a 3.6-fold sensitization with active MGMT, and increases to 6.6-fold upon MGMT inhibition. Figures show representative experiments out of two independent experiments per treatment with data points in duplicate. Caspase 3/7 activity, detecting early apoptosis, was assayed using U251 and U251-MTH1 cells after treatment with Temozolomide ranging from 0 to 100 μM for 48 h (C) or with both Temozolomide and Lomeguatrib (12.5 μM) (D). MTH1 deficient cells show increased caspase 3/7 activity at 50 and 100 μM Temozolomide when MGMT is inhibited with 12.5 μM Lomeguatrib. Shown are representative experiments out of three independent experiments with data points in duplicate. Data are presented as average ± SEM. Statistical significant differences were determined using multiple comparison and Two way Anova in GraphPad Prism 6.0, P ≤ 0.01 is indicated by**.