A disorder-related variant (E420K) of a PP2A-regulatory subunit (PPP2R5D) causes constitutively active AKT-mTOR signaling and uncoordinated cell growth

Abstract

Functional genomic approaches have facilitated the discovery of rare genetic disorders and improved efforts to decipher their underlying etiology. PPP2R5D-related disorder is an early childhood onset condition characterized by intellectual disability, hypotonia, autism-spectrum disorder, macrocephaly, and dysmorphic features. The disorder is caused by de novo single nucleotide changes in PPP2R5D, which generate heterozygous dominant missense variants. PPP2R5D is known to encode a B'-type (B'56δ) regulatory subunit of a PP2A-serine/threonine phosphatase. To help elucidate the molecular mechanisms altered in PPP2R5D-related disorder, we used a CRISPR-single-base editor to generate HEK-293 cells in which a single transition (c.1258G>A) was introduced into one allele, precisely recapitulating a clinically relevant E420K variant. Unbiased quantitative proteomic and phosphoproteomic analyses of endogenously expressed proteins revealed heterozygous-dominant changes in kinase/phosphatase signaling. These data combined with orthogonal validation studies revealed a previously unrecognized interaction of PPP2R5D with AKT in human cells, leading to constitutively active AKT-mTOR signaling, increased cell size, and uncoordinated cellular growth in E420K-variant cells. Rapamycin reduced cell size and dose-dependently reduced RPS6 phosphorylation in E420K-variant cells, suggesting that inhibition of mTOR1 can suppress both the observed RPS6 hyperphosphorylation and increased cell size. Together, our findings provide a deeper understanding of PPP2R5D and insight into how the E420K-variant alters signaling networks influenced by PPP2R5D. Our comprehensive approach, which combines precise genome editing, isobaric tandem mass tag labeling of peptides generated from endogenously expressed proteins, and concurrent liquid chromatography-mass spectrometry (LC-MS³), also provides a roadmap that can be used to rapidly explore the etiologies of additional genetic disorders.

Keywords: AKT-mTOR; Jordan’s syndrome; PP2A; PPP2R5D; PPP2R5D-intellectual disability; PPP2R5D-related neurodevelopmental disorder; phosphatase.

Figure 1

Structure of PP2A holoenzymes and generation of PPP2R5D E420K-variant cell lines. A, diagram of PP2A holoenzyme assembly with PPP2R5D in cyan. B, PPP2R5D homology model showing the likely location of the Glu⁴²⁰ variant in pink based on the crystal structure of PPP2R5C (PDB ID: 2IAE). PPP2R5D is colored as cyan and the scaffolding and catalytic subunits as gray. The catalytic metal ions are purple. C, illustration of targeted cytosine deamination by CRISPR/Cas9/BE4-Gam base editing. D, details of the gRNA interaction with the targeted genomic locus in exon 12 and genomic base editing. E, Sanger DNA sequencing data from WT encoding glutamic acid (GAG) in both alleles, E420K heterozygous (HET) encoding glutamic acid (GAG) in one allele and lysine (AAG) in the other, and E420K homozygous (HOMO) encoding lysine (AAG) in both alleles. F, IPs of endogenous PPP2R5D from WT, E420K HET, and E420K HOMO cells were separated by SDS-PAGE and stained with Coomassie Blue. LC-MS/MS spectra of peptides identified in E420K HET IP’s identified (G) wild type and (H) variant peptides. I, Western analysis of total cell extracts (input) and endogenous PPP2R5D IPs probed for PPP2R5D and PPP2CA. J, relative amounts of PPP2CA from WT or E420K cells normalized to total PPP2R5D in IPs and represented as the mean percent of WT (N = 5 independent experiments; mean ± SD, ANOVA Kruskal–Wallis p = 0.003 and H = 9.41, Dunn’s multiple comparison posthoc test, ∗p < 0.05 and NS: nonsignificant). K, LC-MS/MS detection of PPP2R1A, PPP2R1B, PPP2CA, and PPP2CB in PPP2R5D IPs of WT and E420K-variant cells. L, relative phosphatase activity of endogenous PPP2R5D IPs (n = 10). Endogenous PPP2R5D was immunoprecipitated from WT, HET, or HOMO cells and hydrolysis of DiFMUP was measured at Ex 360 nm and at Em 460 nm. Activity was normalized to total levels of PPP2R5D detected in IPs through immunoblotting shown in (K). WT activity was normalized at 30 min (endpoint) to 1 and variant activity is expressed relative to WT. N = 5 independent biological replicate of cells, each performed as n = 2 technical replicates.

Figure 2

Quantitative changes in the proteome and phosphoproteome in PPP2R5D E420K-variant cells. Workflow for quantitative proteomics (A) and phosphoproteomics (B). Volcano plots of proteomic (C, D) and protein-corrected phosphoproteomic (E, F) data (N = 3 independent biological replicates for WT, N = 4 independent biological replicates for E420K HET and HOMO). Volcano plots depict log₂ ratios of peptides, plotted against the negative log₁₀ of the p value of their fold change. Peptides above –log₁₀ = 1.3 (corresponding to p < 0.05) are considered statistically significant. Peptides shown in green or blue are twofold or more decreased or increased in abundance, respectively. Correlation analysis of proteomic (G) and phosphoproteomic (H) between E420K HET and HOMO datasets.

Figure 3

Peptide phosphorylation and phenotype analysis of proteins with altered phosphorylation site occupancy in PPP2R5D E420K-variant cells.A, Icelogo displaying enrichment and deselection of amino acids surrounding the phosphorylated residue (position 0) and (B) RegPhos analysis of kinase substrates enriched in significantly regulated, localized, single phosphorylation sites compared with all localized, single phosphorylation sites identified in the analysis. C, significantly regulated, localized, single phosphorylation sites in E420K-variant cells known or predicted to be phosphorylated by the indicated kinase. Node color (blue increased, green decreased, yellow both increased and decreased) indicates type of phosphorylation site regulation. Edges indicate known or predicted upstream kinase. D, overrepresentation analysis of human phenotype ontology data base. E, Gene ontology analysis of enriched biological processes. F, overrepresentation analysis of human KEGG pathways data.

Figure 4

Bioinformatic pathway analysis of proteins with altered phosphorylation site occupancy in PPP2R5D E420K-variant cells.A, proteins associated with enriched KEGG pathways identified in Figure 3F. Node size correlates with the number of pathways the protein is associated with. Node color indicates the type of phosphorylation site regulation. Edges indicate pathway association. B, detailed diagram of enriched KEGG pathways. Proteins with p-site(s) increased are shown in blue, proteins with p-site(s) decreased are green, and proteins with p-sites both increase and decrease are yellow.

Figure 5

Orthogonal validation of the LC-MS³data.A, representative immunoblot and (B) associated graph showing the levels of P-AKT1/2/3 (Ser^473/4/2) normalized to total AKT1/2. (C) Representative immunoblot and (D) associated graph showing the levels of P-GSK3α/ß (Ser^21/9) normalized to total GSK3ß. E, representative immunoblot and (F) associated graph showing the levels of P-S6 (Ser^235/236) normalized to total S6. G, representative immunoblot and (H) associated graph showing the levels of P-FOXO3 (Ser²⁵³) normalized to total FOXO3. I, representative immunoblot and (J) associated graph showing the levels of P-ß-catenin (Ser⁵⁵²) normalized to total ß-catenin and (K) associated graph showing levels of total ß-catenin normalized to ß-actin. All (A–K) were performed as n = 3 independent cell experiments, mean ± SD, unpaired t test ∗∗p < 0.01 ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. MS values indicate the fold change detected in the LC-MS dataset (Fig. S2). L, representative immunoblot and (M) associated graph showing levels of P-S6 (Ser^235/236) normalized to total ß-actin. Cells were assayed in 10% serum containing media at decreasing cell densities. “+” sign indicates 100% confluent cells incubated with fresh media containing 10% serum for 24 h prior to harvest. (n = 3 independent experiments, mean ± SD, ANOVA Kruskal–Wallis p = 0.008 and H = 19.11, Dunn’s multiple comparison posthoc test, ∗p < 0.05 and NS: nonsignificant). N, representative immunoblot and (O) associated graph showing levels of P-S6 (Ser^235/236) normalized to total ß-actin. Cells were grown in the presence/absence of 10% serum for 24 h and stimulated with/without EGF (0.5 ng/ml) and/or IGF-1 (0.5 ng/ml) for 30 min prior to harvest. (n = 4 independent experiments, mean ± SD, ANOVA Kruskal–Wallis p < 0.0001 and H = 45.3, Dunn’s multiple comparison posthoc test, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, and NS: nonsignificant).

Figure 6

Rapamycin inhibits aberrant mTOR signaling and decreases cell size in E420K-variant cells.A, representative histogram showing the distribution of mean FSC-H for HET cells (black) compared with WT cells (blue). B, graph showing greater mean relative FSC-H in HET compared with WT cells. Treatment with rapamycin (10 nM, 48 h) significantly decreased the mean relative cell sizes of both HET (C, D) and WT (E, F) cells. The absolute rapamycin-induced decrease in HET (red) was beyond that seen in WT DMSO-control treated cells (blue) G, H, FSC-H is in arbitrary units (AU) relative to each other (n = 4 independent experiments; mean ± SD, unpaired t test ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001). I, representative immunoblot showing dose-dependent inhibition of P-S6 (Ser^235/236) by rapamycin. Rapamycin did not affect P- AKT1/2/3 (Ser^473/4/2) and P-GSK3α/ß (Ser^21/9). ß-actin was probed as an internal control. J, graph showing levels of P-S6 (Ser^235/236) normalized to ß-actin (n = 3 independent experiment, shown is the representative experiment). K, representative immunoblot analysis of AKT IPs probing for PPP2CA and PPP2R5D. L, representative immunoblot analysis of PPP2CA IPs probing for PPP2R5D and AKT.

Figure 7

PPP2R5D E420K-variant signaling. Diagram of pathway regulation by PPP2R5D containing PP2A-holoenzymes altered in E420K variant HEK-293 cells based on the data presented. Phosphorylation sites shown in blue (hyperphosphorylated) and green (hypophosphorylated) differ in the E420K variant cells as compared with wild-type cells.