Fig 1: Copper is required for PERK kinase activity in vitro(A, B, and F) Representative western blot images, probed as indicated, of kinase assays with PERK-FLAG (A) and (B) or HRI-FLAG (E) variants, as labeled, carried out at 37°C for the indicated time points (A), (B), or 1 h (F).(C) Quantification n = 3 replicates of the experiment in (B) plotted as enzyme progress curves. Data were fit to an exponential plateau model by least-squares regression with all parameters constrained to greater than zero, and YM constrained to be shared. There was a difference of 13.64 in AICc for one curve for all datasets vs. a different curve for at least one dataset (WT vs. PCM1 vs. PCM2), conferring a 99.9% probability that the curve is different for at least one dataset. The separate dataset models are plotted with asymmetric 95% confidence interval error bars. R-squared and best-fit values for curve parameters are as follows: WT-0.80, Y0 = 0.10, k = 0.17; PCM1–0.67, Y0 = 0.04, k = 0.16; and PCM2–0.19, Y0 = 0.02, k = 0.00. Shared YM was calculated to be 1.23.(D) Plot of enzyme progress curve (p-eIF2α accumulation as monitored by absorbance at 450 nm, over time) for n = 3 replicates of ELISA analyzed kinase reactions with different PERK variants, as indicated. Data were fit to an exponential plateau model by least-squares regression with all parameters constrained to greater than zero, and YM constrained to be shared between 0 and 4 (the detection limit of the spectrometer). There was a difference of 131.1 in AICc for one curve for all datasets vs. a different curve for at least one dataset, conferring a >99.99% probability that the curve is different for at least one dataset. The separate dataset models are plotted with asymmetric 95% confidence intervals. R-squared and best-fit values for curve parameters are as follows: WT-0.92, Y0 = 0.58, k = 0.07; PCM1–0.98, Y0 = 0.54, k = 0.07; and PCM2–0.46, YM = 0.37, Y0 = 0.21, k = 0.00. Shared YM was calculated to be 3.34.(E) Quantification of p-eIF2α levels from n = 6 PERK-FLAG protein kinase reactions carried out at 37°C for approximately 5 min as monitored by ELISA absorbance at 450 nm. One-sample t test for mean equal to 100 (WT), ****p < 0.0001.(G) Quantification of n = 3 replicates of the experiment shown in F. One-sample t test for mean equal to 1 (HRI), p = 0.09.(H) Western blot images, probed as indicated of PERK-FLAG kinase assay reactions carried out at 37°C for 10 min in the presence of increasing titrations of copper chelators TTM (5–50 μM) or BCS (5–500 μM).(I and J) Quantification of n = 3 variable replicates of the experiment shown in H. Data were fit to a four-parameter (variable slope) [inhibitor] vs. response model, by least-squares regression, with all parameters except the Hill slope constrained to greater than zero. Models are plotted with asymmetric 95% confidence intervals. IC50, Hill slope, and R-squared of the fit curves indicated. All data points represent individual experimental replicates, with replicate indicated by data point shape. Bars represent mean ± SD. Also see Figure S2.