Selective PP2A Enhancement through Biased Heterotrimer Stabilization

Abstract

Impairment of protein phosphatases, including the family of serine/threonine phosphatases designated PP2A, is essential for the pathogenesis of many diseases, including cancer. The ability of PP2A to dephosphorylate hundreds of proteins is regulated by over 40 specificity-determining regulatory "B" subunits that compete for assembly and activation of heterogeneous PP2A heterotrimers. Here, we reveal how a small molecule, DT-061, specifically stabilizes the B56α-PP2A holoenzyme in a fully assembled, active state to dephosphorylate selective substrates, such as its well-known oncogenic target, c-Myc. Our 3.6 Å structure identifies molecular interactions between DT-061 and all three PP2A subunits that prevent dissociation of the active enzyme and highlight inherent mechanisms of PP2A complex assembly. Thus, our findings provide fundamental insights into PP2A complex assembly and regulation, identify a unique interfacial stabilizing mode of action for therapeutic targeting, and aid in the development of phosphatase-based therapeutics tailored against disease specific phospho-protein targets.

Keywords: B56α; PP2A; SMAP; c-Myc; cancer; cryo-EM; phosphatase; small molecule.

Fig. 1.. DT-061 increases the population of B56α containing heterotrimers.

(A) PP2A-A split bioluminescence tagged (NanoBit) H358 cells treated with DMSO or 30 μM DT-061 for 15 min followed by washout and analysis at 15 min (blue), 30 min (red), 1 hr (green), 2 hr (magenta), 6 hr (orange), or 12 hrs (black), demonstrate an increase in signal 1–2 hours after DT-061 treatment followed by a subsequent return to baseline. Two-way ANOVA with Holm-Sidak multiple comparison test, *** = p < 0.001, **** = p < 0.0001 (B) Co-IP of V5-tagged PP2A-A from xenograft tumors treated for 29 days with DMSO or DT-061 followed by mass spectrometry and SAINT analysis of co-immunoprecipitated proteins, expressed as dot-plots using ProHits-viz, indicates an increase in B56α holoenzymes with DT-061 treatment. Color intensity of dot represents raw spectral count, size of dot represents normalized intensity, intensity of dot border represents statistical confidence. (C-D) Cell culture Co-IP and whole lysate (input) from H358 cells stably expressing V5-tagged PP2A-A treated with 3 μM DT-061, probed for B55α, B56α, PR72, V5, or vinculin demonstrates selective increase in B56α holoenzymes with DT-061 treatment. (E) PP2A-A and B56α split bioluminescence tagged (NanoBRET) H358 cells treated with 20 μM or 30 μM DT-061 confirms enhanced population of B56α containing holoenzymes. ANOVA with Dunnett multiple comparison test, *** = p < 0.001, **** = p < 0.0001. (See also Figure S1).

Fig. 2.. DT-061 directly stabilizes AB56αC heterotrimers.

(A) Size-exclusion chromatography (SEC) profiles of PP2A holoenzyme complex components, including C monomer, AC dimer and AB56αC trimer complex at 6 μM concentration (black lines). Purified AB56αC trimer starts to dissociate at 100 nM concentration during SEC analysis (blue line), while the complex remains stable (red line) in the presence of 2 μM DT-061 in SEC running buffer. The peak height is UV absorbance at 280 nm normalized to the peak of each plot. (B) 6-FAM labeled recombinant B56α incubated with increasing concentrations of AC analyzed by fluorescence polarization to determine the affinity of B56α for AC in the absence (red line, KD ~130 ± nM) and presence (blue line, KD ~ 53.0 ± 2 nM) of DT-061 (2 μM). (C) Surface immobilized B56α subject to SPR with increasing concentrations of AC dimer (colors) in the absence (top) or presence (bottom) of DT-061 (2 μM). SPR data of B56α in the presence of DT-061 demonstrates a similar enhanced overall affinity for AC. (D) Surface immobilized B56γ subject to SPR with increasing concentrations of AC dimer (colors) in the absence (top) or presence (bottom) of DT-061 (2 μM) illustrates no effect of DT-061 on B56γ binding kinetics for AC. (See also Figure S2).

Fig. 3.. The binding pocket of DT-061 resides at the trimeric interface of AB56αC.

(A) A 3.6 Å cryo-EM density map of AB56αC holoenzyme in complex with DT-061 (green density) bound PP2A trimer. The individual subunits, A, B56α and, C, are colored in red, blue and orange respectively. (B) Close-up view of DT-061, shown in grey mesh density with carbons represented in green. Part of DT-061 is buried inside a cavity between the A (red) and B (blue) subunits, positioned underneath the C-tail (F307-L309) of the C-subunit (orange). The phenoxazine moiety of DT-061 stacks on a solvent exposure interface area between the C-tail of the C-subunit and the A-subunit. (C) A close-up view of DT-061 (green) partly buried in a cavity between A (red) and B (blue) subunits directly adjacent to the C-tail (F307-L309) of the C subunit (orange). Cryo-EM density of DT-061 and C-tail illustrated by respectively colored meshes. L309 methylation is highlighted in blue ball and stick model. (D) Atomic details of DT-061 binding pocket. Side-chain of residues involved in formation of the binding pocket are colored in red, blue and orange for A, B and C subunits respectively. (E) Representative micrographs and 2D class averages, with (F) schematics, of AB56αC holoenzyme complex without DT-061 (top, blue box) and with DT-061 (bottom, red box). The black scale bar is 200 Å. (See also Figure S3 and Table S1).

Fig. 4.. Structure based design of inactive DT-061 chemical derivatives confirms binding pocket, biochemical stabilization, and apoptotic effects of DT-061.

(A) DT-061 chemical derivative structures with one methyl group (-CH3; NZ-035), two methyl groups (NZ-043) or a benzene (Bn; NZ-044) group added. (B) Fluorescence polarization determined affinities of wildtype B56α with AC in the presence of DT-061 or the derivative compounds demonstrates that all derivative compounds lose the ability to stabilize B56α to the AC dimer. (C) Cell viability dose response curves of H358 lung adenocarcinoma cells treated with DT-061 (red line), NZ-035 (orange line), NZ-043 (green line), or NZ-044 (magenta line) at 24 hours. In line with the inability to stabilize B56α heterotrimers, all derivative compounds demonstrate a complete loss of apoptotic potential. (See also Figure S4).

Fig. 5.. DT-061 dependent B subunit selection is class and isoform specific.

(A) The DT-061 (EM density in mesh) binding region in the AB56αC holoenzyme as compared to that of (B) the previously determined AB56γC holoenzyme. The T281 residue in AB56αC is represented by a lysine (yellow highlight) at the analogous position (K246, yellow highlight in B) in AB56γC, which forms a salt bridge with E100 of the A subunit (dashed line). (C) Sequence alignment of the 15 different B56 subunits demonstrating the divergence of the analogous T281 position. (D) A single point mutation, T281K, of B56α, mimicking B56γ at this position, forms a complex with AC dimer at 6 μM concentration (black line) that dissociates at 100 nM concentration (cyan line) that is not rescued by addition of DT-061 (100 nM, pink line). (E) Fluorescence Polarization and (F) SPR of the T281K B56α mutant with AC dimer determined in the absence (solid line) and presence of DT-061 (dashed line) illustrates that the binding kinetics and overall affinity of the mutant complex is unaffected by DT-061. (See also Figure S4 and S5).

Fig. 6.. Methylation of the PP2A catalytic subunit correlates with DT-061 activity in vivo.

(A) Immunohistochemical analysis of xenograft treated tumors probed for mL309 (left) or tPP2A-C (right) demonstrates an increase in mL309 at early time points following DT-061 treatment. Representative images from each time point (n=4–6) are shown at 100x magnification, black scale bar is 10 μM. Densitometric quantification of western blots probed for (B) methyl-L309 PP2A-C (mL309), (C) total-PP2A-C (tPP2A-C) and vinculin from tumors treated with a single dose of DT-061 (5mg/kg) for the designated times (n=5–13 tumors per group) mirrors the IHC data whereby DT-061 treatment enhances mL309 between 1–3 hours after treatment followed by a return to baseline. Individual data points represent ratios, mL309/tPP2A-C or tPP2A-C/Vinculin. Box-whiskers represent average ± s.d. one-way parametric ANOVA with Dunnett multiple comparison test presented comparing treated groups to vehicle control, *** = p < 0.001, **** = p < 0.0001. (D) Immunohistochemical c-MYC staining of single dose vehicle or DT-061 treated xenograft tumors demonstrate PP2A activation by DT-061 inversely correlates with changes in c-MYC detected in vivo. Images are recorded at 100x magnification, black scale bar is 10 μM. (E) Representative fluorescent microscopy images of treated and control H358 xenograft tumors resected at 1, 2, 3, 6, 12, or 24 hours after DT-061 treatment, stained for TUNEL and counterstained with DAPI, white scale bar is 100 μM. (F) Quantification represents percentage of TUNEL-positive cells per treatment group. Bar graphs represent average ± s.d. one-way parametric ANOVA with Dunnett multiple comparison test comparing treated groups to vehicle control, * = p < 0.05 ** = p < 0.01 *** = p < 0.001. (See also Figure S6)

Fig. 7.. Methylation and interactome deficiencies drive DT-061 resistance in a recurrent cancer derived PP2A-A mutation, R183W, in vivo.

(A) Western blot of co-IP and lysate samples from H358 cells expressing V5 tagged EGFP, WT scaffold, or R183W mutant scaffold subunit. Lysate probed for methylated L309 (mL309), total PP2A-C (tPP2A-C), V5, and total PP2A-A (tPP2A-A), demonstrate marked reduction in mL309 in R183W holoenzymes. (B) Percent of L309 methylation of the PP2A catalytic subunit bound to WT-A or R183W-A holoenzymes in cell culture as determined by mass spectrometry. (C) Co-IP of V5-tagged WT versus R183W PP2A-A from H358 tumors followed by mass spectrometry and SAINT analysis of co-immunoprecipitated proteins, expressed as dot-plots using ProHits-viz, indicates R183W mutant has dramatic deficiencies in B-subunit binding, particularly B56α. Color intensity of dot represents raw spectral count, size of dot represents normalized intensity, intensity of dot border represents statistical confidence. (D) Subcutaneous xenograft tumors expressing V5 tagged EGFP (left), WT scaffold (middle), or R183W scaffold subunit (right), treated with DMA control (blue line), dual kinase inhibitor combination (AZD6244/MK2206, 24mg/kg:6mg/kg, green line), or DT-061 (5mg/kg, red line). (E) Co-IP of WT scaffold and R183W scaffold subunit from xenograft tumors treated with DMA or DT-061 for 29 days demonstrates that WT holoenzymes have ~10% more L309 methylation after DT-061 treatment while R183W holoenzymes have a reduction in L309 methylation as determined by mass spectrometry. (F) Co-IP, mass spectrometry, SAINT analysis, and ProHits dot-blot visualization of R183W interactome in tumors treated with either DMA (left) or DT-061 (right) demonstrates no significant increase in B56α containing mutant holoenzymes with DT-061 therapy. (See also Figure S7).