Activity assay

To be able to detect both low and high activities of the NUDIX hydrolases, we performed the assay at two enzyme concentrations, low (5 nM) and high (200 nM) with 25 or 50 µM substrate, respectively (when available) (Supplementary Fig. ^{2b, c}). NUDIX protein activity with a panel of possible substrates was assessed in technical triplicates, in reaction buffer (100 mM Tris Acetate pH 7.5, 40 mM NaCl, 10 mM MgAc, and 1 mM dithiothreitol) at 22 °C. 50 µM of the respective substrate was incubated together with 0, 5, or 200 nM NUDIX protein diluted in reaction buffer, either without coupling enzyme, or with E. coli pyrophosphatase (PPase) (0.2 U ml⁻¹) or alkaline phosphatase from bovine intestinal mucosa (10 U ml⁻¹) (Sigma Aldrich), in order to detect inorganic phosphate (Pi) produced from the reaction products pyrophosphate and sugar-5-phosphates, respectively (Supplementary Fig. ^2b). After 30 min incubation under shaking conditions, the generated Pi was detected by addition of a malachite green reagent³⁶. After 15 min incubation, the reaction was stopped by addition of 10 µl 0.4 M sodium citrate to the 40 µl reaction mixture per well in a 384-well plate and the absorbance was read at 630 nm in an EnVision plate reader (Perkin Elmer). We normalized the absorbance signal to the control in the absence of coupling enzymes. In some cases, such as MTH1 activity toward 8-oxo-dGTP, the low concentration of enzyme was sufficient to completely convert the substrate into product, providing a maximum signal (Supplementary Fig. ^2c).