De novo missense variants in the PP2A regulatory subunit PPP2R2B in a neurodevelopmental syndrome: potential links to mitochondrial dynamics and spinocerebellar ataxias

Abstract

The heterotrimeric protein phosphatase 2A (PP2A) complex catalyzes about half of Ser/Thr dephosphorylations in eukaryotic cells. A CAG repeat expansion in the neuron-specific protein PP2A regulatory subunit PPP2R2B gene causes spinocerebellar ataxia type 12 (SCA12). We established five monoallelic missense variants in PPP2R2B (four confirmed as de novo) as a cause of intellectual disability with developmental delay (R149P, T246K, N310K, E37K, I427T). In addition to moderate to severe intellectual disability and developmental delay, affected individuals presented with seizures, microcephaly, aggression, hypotonia, as well as broad-based or stiff gait. We used biochemical and cellular assays, including a novel luciferase complementation assay to interrogate PP2A holoenzyme assembly and activity, as well as deregulated mitochondrial dynamics as possible pathogenic mechanisms. Cell-based assays documented impaired ability of PPP2R2B missense variants to incorporate into the PP2A holoenzyme, localize to mitochondria, induce fission of neuronal mitochondria, and dephosphorylate the mitochondrial fission enzyme dynamin-related protein 1. AlphaMissense-based pathogenicity prediction suggested that an additional seven unreported missense variants may be pathogenic. In conclusion, our studies identify loss-of-function at the PPP2R2B locus as the basis for syndromic intellectual disability with developmental delay. They also extend PPP2R2B-related pathologies from neurodegenerative (SCA12) to neurodevelopmental disorders and suggests that altered mitochondrial dynamics may contribute to mechanisms.

Keywords: cerebellar ataxias; dynamin-related protein 1; mitochondrial dynamics; neurodevelopmental disorders; protein phosphatase 2A.

Figure 1

Effect of PPP2R2B variants on PP2A holoenzyme assembly. (A) Facial photograph of patient 4 showing dysmorphic features. (B and C) A novel NanoBiT luciferase complementation assay (principle in (B)) shows significantly weakened interactions of all phenotypically characterized PPP2R2B/Bβ2 variants with the catalytic Cα subunit in transfected COS-1 cells; RR168EE is a mutation at the interface to the A subunit known to disrupt the PP2A complex. Data points are means of four independent experiments with 6 technical replicates each. (D and E) Co-immunoprecipitation of FLAG-GFP tagged Bβ2 variants from COS-1 cells shows that all variants associate with less endogenous A and C subunits compared to wild-type (WT) Bβ2. (E) Data points are normalized to C and A subunit levels in the co-IP from 6 independent experiments. Molecular mass markers are shown in kDa on the right of the representative immunoblot in (D). Error bars show means ±95% CI. Asterisks denote significant differences compared to WT based on one-way ANOVA analysis with Dunnett’s multiple comparisons post-hoc test. ^*P < 0.05; ^**P < 0.01; ^***P < 0.005; ^****P < 0.0001.

Figure 2

The R149P variant promotes PPP2R2B/Bβ2 degradation. (A) Representative blots of a Bβ2-HaloTag protein turn-over time course. COS-1 cells were transiently transfected with Bβ2 WT or variants that are C-terminally-fused with FLAG and HaloTag subunits. Cells were then pulse-labeled with TMR fluorescent ligand (30 min, 50 nM) followed by chase with media containing 7-bromoheptanol (10 μM) for the indicated times. Following SDS-PAGE of total lysates, a decay in TMR fluorescence was detected by infrared imaging and FLAG was detected for total protein by immunoblotting of the same blot. Molecular weight markers (in kDa) are indicated on the right of the blots. (B–D) Bβ2 degradation over time was quantified as the ratio of TMR to total Bβ2-FLAG-HaloTag signal and further normalized to time zero (no chase). Data points in (B) are means ± SEM from 6–7 independent experiments. (C) Half-lifes (t_1/2) were calculated from fitting the normalized data points to a one-phase decay curve. (D) Area-under-the curve was calculated from the same curves. Data points in (C and D) represent individual experiments and error bars represent means ±95% CI. ^*P < 0.05; ^**P < 0.01 according to one-way ANOVA with Dunnett’s post hoc test.

Figure 3

PPP2R2B/Bβ2 variant effect on mitochondrial localization and fission. (A and B) Bβ2-GFP expressed HeLa cells (green, representative images in (A)) were colocalized with mitochondria (COX2 immunofluorescence, red) and colocalization was quantified as Pearsons coefficient (B). Bβ2 SS21DD and SSS20AAA are previously described mutants with decreased and increased mitochondrial localization, respectively [8]. Data points are the means of 7-9 independent experiments (~30–40 cells per transfection). Error bars are 95% CI. (C–E) Primary rat hippocampal neurons were cotransfected with Bβ2-GFP and or outer-mitochondrial GFP (mGFP, green) together with Mito-V5-mRFP (red) at 4:1 mass ratios and fixed at 12 days-in-vitro followed by immunofluorescence imaging (representative images in (C)). (D) Data points are mean mitochondrial length (μm) of 43–67 neurons from four separate transfections and two primary cultures (error bars are means ±95% CI). (C) An XY-plot correlates mitochondrial length and form factor, two measures without common parameters, from the same set of neurons (means ±95% CI). ^*P < 0.05; ^**P < 0.01; ^***P < 0.001; ^****P < 0.0001 by one-way ANOVA followed by Dunnett’s post hoc test compared to WT (B) or Tukey’s post hoc test (C).

Figure 4

Effects of PPP2R2B/Bβ2 variants on neuronal survival and Drp1 dephosphorylation. (A and B) Primary hippocampal neurons were transfected with Bβ2-GFP WT/variant or outer-mitochondrial (Mito)-GFP negative control at 9 days-in-vitro. 72 h post-infection, cultures were fixed and transfected (GFP-positive) neurons were scored as dead or alive based on nuclear size/morphology (Hoechst stain), and integrity and MAP2B staining of GFP-positive neurites. Representative images are shown in (A); the arrow points to a condensed, apoptotic nucleus. (B) Quantification of neuronal death (dead/(dead + live) GFP-positive neurons; 200–300 neurons per experiment) in 4 independently prepared cultures normalized to the Mito-GFP negative control. (C) Model of regulation of Drp1 activity and mitochondrial fission by phosphorylation/dephosphorylation of the inhibitory Ser637 site. PP2A/Bβ2 and PKA/AKAP1 activity antagonize each other in the control of mitochondrial fission. (D and E) Cell-based Drp1 dephosphorylation assay utilizing HEK293 cells in which the majority of PP2A holoenzymes is replaced by a defined PP2A/Bβ2 holoenzyme [17]. Drp1 is largely dephosphorylated under basal (b) conditions; the asterisk indicates a non-specific band. However, upon forskolin/rolipram treatment (F/R, 20/2 μM for 20 min), pSer637-Drp1 is detectable reflecting the balance of PKA and PP2A/Bβ2 WT/variant activities (D). (E) Quantification of pSer637 to Drp1 ratios normalized to WT from 4–6 independent experiments. ^*P < 0.05; ^**P < 0.01; ^***P < 0.001; either by one-way ANOVA followed by Dunnett’s post hoc test (B) or by Kruskal-Wallis test with Dunn’s post hoc test (E).