Mammalian FMRP S499 Is Phosphorylated by CK2 and Promotes Secondary Phosphorylation of FMRP

FMRP S499 is phosphorylated by multiple kinases in vitro. A, FMRP S499 phosphorylation is not altered by 3-h staurosporine treatment. N2a cells were treated for 3 h with increasing doses of staurosporine. Staurosporine decreased rpS6 S240/244 phosphorylation without affecting FMRP S499 phosphorylation. B, Quantification of rpS6 S240/244 (pS6/S6, blue) and FMRP S499 phosphorylation after staurosporine treatment. There was no significant effect of staurosporine treatment on FMRP phosphorylation (Kruskal–Wallis one-way ANOVA [H(5) = 2.32, p = 0.8034]), but there was a significant effect on rpS6 S240/244 phosphorylation (Kruskal–Wallis one-way ANOVA [H(5) = 16.04, p = 0.0067]. Post hoc Dunn’s test for multiple comparisons showed p < 0.05 for 50- and 150-µm treatment groups compared with 0 µm. n = 3, error bars = SEM. C, Initial Kinexus kinase assay. rFRMP S500 was incubated with recombinant kinase listed above the SDS-PAGE on 10% Tris-glycine gels. Samples were run on two independent blots. The leftmost lane is the ladder. Arrows indicate where intervening lanes have been removed for clarity. D, Repeat kinase assay including GRK2. The two panels are from the same membrane. After transfer, the membrane was probed with anti-pFMRP S499, stripped, and reprobed with tFMRP.

CK2 phosphorylates mammalian FMRP S499. A, Immunoblots for 3-h treatment of N2a cells with vehicle (DMSO), D4476 (25 µm), IC261 (20 µm), PHA-767491 (5 µm), DRB (50 µm), TBB (25 µm), or βARK (200 µm) followed by Western blot for protein indicated on the left. B, Quantification of pFMRP/tFMRP signal in A. ns, not significant (Kruskal–Wallis one-way ANOVA analysis [H(8) = 4.56, p = 0.8034], n = 4). C, Immunoblot for 24-h treatment of N2a cells with the same agents listed in A. D, Quantification of pFMRP/tFMRP signal in C (Kruskal–Wallis one-way ANOVA [H(8) = 7.239, p = 0.5111], n = 4, error bars = SEM). E, Baseline, untreated N2a cells were collected at time 0, and the remainder of the cells were treated with either DMSO or CX-4945 (5 or 1 µm) for 24 h. tFMRP and pFMRP S499 signals increased in DMSO-treated samples; however, only tFMRP increased in CX-treated samples, thereby causing a significant reduction in relative FMRP S499 phosphorylation. All immunoblot signals are from the same membrane; however, intervening lanes have been removed for clarity. F, Quantification of pFMRP S499 to tFMRP ratio from D; one-way ANOVA, n = 4, error bars = SEM, *p < 0.05. G, HEK293 cells were collected at baseline (time 0) or treated for 24 h with DMSO or CX-4945. H, CX-4945 significantly reduced FMRP S499 phosphorylation compared with DMSO (one-tailed Mann–Whitney test, p = 0.0143). I, J, Mouse cortical neurons at 7 d in vitro treated with 1 µm CX-4945 for 24 h exhibited a significant reduction in FMRP S499 phosphorylation compared with DMSO-treated neurons (one-tailed Mann–Whitney test, p = 0.05).

CK2 phosphorylates mammalian FMRP S499 in vitro. rFMRP S500 or S500D was incubated with or without recombinant CK2 for 30 min. Samples were resolved by SDS-PAGE and probed with pFMRP S499, tFMRP, or CK2a1 antibodies. Only rFMRP S499 incubated with CK2 showed a positive phosphosignal.

FMRP S499 rephosphorylation kinetics after CX-4945 treatment and washout mirrors a known CK2 target. A, N2a cells were collected either at baseline or 24 h after DMSO or CX-4945 treatment. Additional samples were collected after 24-h treatment and washout of DMSO or CX-4945 for different time periods. B, Quantification of [p/t]FMRP or [p/t]AKT compared with baseline. Changes in ratios were quantified by two-way ANOVA followed by Dunnett’s multiple comparisons post hoc test, n = 4 per data point, error bars = SEM. ****p < 0.0001; ***p < 0.001; *p < 0.05.

CK2 phosphorylates primarily newly translated FMRP. A, Baseline N2a cells were collected without treatment. The remainder of N2a cells were treated with DMSO or DMSO with cycloheximide for 1 h. After 1 h, the medium was replenished with fresh medium containing DMSO or DMSO with cycloheximide, and cell lysates were collected 6 h later. B, Bar graphs of the relative tFMRP and pFMRP signals at baseline, 6 h after media change + DMSO, and 6 h after media change + DMSO + cycloheximide. Total FMRP was significantly increased 6 h after media change (one-way ANOVA [H(4) = 6.736, p = 0.0194], n = 4), and this increase was blocked by treatment with cycloheximide. The relative pFMRP levels were not significantly different in any of the three conditions (one-way ANOVA [H(4) = 4.344, p = 0.1131], n = 4). C, Baseline cells were collected without treatment, and the remainder of cells were treated with above agents for 1 h. After 1 h, cells were replenished with fresh medium, and cell lysates were collected 6 h later. D, Bar graphs of pFMRP/tFMRP under the different conditions. There was a statistically significant difference of pFMRP/tFMRP between treatment groups and compared with the DMSO control group (one-way ANOVA [H(3) = 11.22, p = 0.0009]). Dunn’s post hoc multiple comparisons test was not significant between any group and the DMSO control group. However, given the a priori Bayesian hypothesis that CX-4945 decreases pFMRP S499, we performed a nonparametric t test between DMSO and CX-4945 groups and found a significant decrease in pFMRP S499 due to 6-h CX-4945 treatment after media change (p = 0.0143, n = 4).

FMRP S499 phosphorylation promotes secondary FMRP phosphorylation on serine/threonine residues. A, Immunoblots (IB) for anti-pSer (left), anti-pThr (middle), and anti-pTyr (right) after immunoprecipitation with anti-GST for the different FMRP constructs (S499, S499A, and S499D). pRes, phosphoresidue. B, Quantification of phosphoresidue over GST blots. **p < 0.01 compared with S499 pRes by one-way ANOVA, n = 4 per data point, error bars = SEM. C, Post-IP eluates were separated by SDS-PAGE and Coomassie stained overnight. The GST-FMRP and heavy antibody chain bands are indicated by the arrows on the right. The ∼120-kD band corresponding to the size of GST-FMRP was excised for phosphoproteomic analysis. D, The rectangle above represents FMRP and indicates the location of the identified phosphopeptides after phosphoproteomics. The graph below shows the relative phosphopeptide to total peptide ratio normalized to GST-FMRP S499 and indicates that the proportion of phospho-to-total peptide ratio is selectively decreased for peptide S483–521 in the GST-FMRP A499 condition. The bars for the peptides corresponding to aa 58–70 have been omitted given that their low abundance led to widely spread results that precluded easy representation on the bar graph. The relative abundances for the peptide corresponding to aa 58–70 were GST-FMRP S499 = 1, GST-FMRP A499 = 6.96, and GST-FMRP D499 = 3.03. Note that the overall abundance of the aa 58–70 phosphopeptide relative to all FMRP peptides was <0.04%.

FMRP S499 influences the phosphorylation of other FMRP residues, possibly downstream of PP2A and mGluR-I. A, N2a cells were transfected with GST-FMRP S499 and treated with DMSO, CX-4945, okadaic acid, or calyculin for 6 h. After treatment, FMRP was isolated via GST immunoprecipitation, and samples were probed for phosphothreonine. The same membranes were then stripped and probed for total GST. CX-4945 dramatically decreased the phosphothreonine signal. The PP2A inhibitor okadaic acid dramatically increased the phosphothreonine signal, whereas the PP1 inhibitor had no detectable effect. B, N2a cells were transfected with GST-FMRP S499, D499, and A499 constructs. Cells were treated with okadaic acid for 1 h or DHPG for 2 min before lysis. FMRP was isolated by GST immunoprecipitation and probed for phosphothreonine. Okadaic acid and DHPG increased phosphothreonine signals for S499 and D499 compared with DMSO control, whereas A499 phosphorylation was unresponsive to either treatment.