Fig 2: Perphenazine, iHAP1 and DT-061 do not activate PP2A complexes in vitro ATitration experiment of different PP2A holoenzyme preparations to determine appropriate concentration to use in enzymatic assays. Experiment was performed twice using the concentrations of enzyme stated in the figure. Shown is one representative experiment.B–E In vitro dephosphorylation assays using model phosphopeptides and purified PP2A holoenzymes. The phosphopeptides were incubated with specific PP2A holoenzymes together with PPZ, iHAP1 or DT-061 at the indicated concentrations. The reaction was stopped after 15 min, and the amount of phosphate released was measured. In (E) substrates are BRCA2 WT: P(pS)QKAEITELSTILEESGSQW and BRCA2 2A: P(pS)QKAEITEASTALEESGSQW. Mean and SD from three technical replicates of one out of three independent experiments is shown. Experiment in (E) is representative of two independent experiments.FInjection of PPP2R1A (300 µM) into the ITC buffer (5% DMSO) used for the DT-061, iHAP experiments. The experiments were performed at 25°C. No significant effects due to dilution of protein (control experiment for Fig 1B)

Fig 3: Perphenazine, DT-061 and iHAP1 have no direct effect on PP2A holoenzymes Chemical structure of perphenazine, DT-061 and iHAP1.Thermogram (top panel) for the calorimetric titration of perphenazine (400 µM, in syringe) in PPP2R1A (30 µM, in calorimetric cell) or PPP2R1A (300 µM, in syringe) in DT-061 (20 µM, in calorimetric cell) or iHAP1 (20 µM, in calorimetric cell) at 25°C and corresponding ligand-normalized integrated heats (bottom panel). Experiments were performed in 5% DMSO. At least two independent experiments were performed. Control injection of PPP2R1A shown in Fig EV1F. In vitro dephosphorylation assay using a basofilic phosphopeptide and purified PP2A holoenzymes. The peptide was incubated with specific PP2A holoenzymes together with iHAP1 or DT-061 as indicated. The reaction was stopped after 15 min, and the amount of phosphate released directly measured. Mean and SD from three technical replicates of one out of three independent experiments are shown.MP mass distribution of the PP2A-B56a holoenzyme at a final concentration of 50 nM in absence (blue) and in presence (green) of 2.5 µM DT-061 in PBS. The mass photometry measurement of the holoenzyme shows a similar distribution of species in presence or absence of DT-061. A representative experiment of two independent experiments is shown.

Fig 4: In vivo and structural analysis of drug activity on cells CRISPR-Cas9-inducible HeLa cell lines for PP2A regulatory subunits. Western blotting shows the removal of the indicated B56 subunits 3 and 12 days after addition of doxycycline, respectively. Actin was used as loading control. At least three biological replicates of the experiment were performed.Incucyte growth assay using a HeLa cell line stably expressing inducible YFP-B56a. Induced cells were treated with Doxycycline for 12 h before imaging. Cell was treated with 20 µM DT-061 or DMSO (Ctrl). Experiment was repeated in three times independently of each other with each experiment performed in technical triplicate. A representative independent experiment is shown with mean and SD.A total cell extract was prepared from HeLa cells stably expressing doxycycline-inducible YFP-B56a. Cells were either untreated or expression induced for 24 h with Doxycycline and the complexes in the cell extracts separated on a Superdex200 column. The migration of molecular weight markers or recombinant untagged B56a is indicated on top. The fractions were analyzed by quantitative Licor western blot for PP2A/A, B56a and PP2A/C. The ratio of B56a or Venus-B56a to PP2A/A in the peak fractions is shown below. PP2AC levels co-migrating with PPP2R1A increased 2–3-fold in the sample where YFP-B56a was induced. Experiment was performed once.(i) The superposition of the crystallographic electron densities maps of PDB 2IAE chain a,b,c (green) and PDB 3FGA (light pink) with cryo-EM map EMD-0510 (gray) at a map threshold of 0.25 shows that in all three maps, density is protruding into the center of the horseshoe-shaped PP2A holoenzyme at the same position (closeup, right). The fits of all three PP2A structures—PDB 6NTS (ii), PDB 3FGA (iii) and PDB 2IAE, (chain a,b,c) (iv) —into their respective densities show that in all three cases, this area of density was attributed to the C-terminal tail of the PP2A catalytic subunit. In the case of PDB 6NTS, the overall map quality and local resolution in that area reveals less easily interpretable features than in the case of PDBs 3FGA/2IAE. Only the last three C-terminal tail residues (307-9) were built, and the DT061 ligand (red) was modeled right next to them. This leaves most of the visible extra density (asterisk) uninterpreted (ii), in a position where in the higher resolution models of PP2A/C (3FGA/2IAE) further residues of the PP2A/C C-terminus have been placed.