Abstract
Fragile X syndrome (FXS), caused by loss of the Fragile X Mental Retardation 1 (FMR1) gene product (FMRP), is the most common heritable cause of intellectual disability and autism spectrum disorders. It has been long hypothesized that the phosphorylation of serine 500 (S500) in human FMRP controls its function as an RNA-binding translational repressor. To test this hypothesis in vivo, we employed neuronally targeted expression of three human FMR1 transgenes, including wild-type (hFMR1), dephosphomimetic (S500A-hFMR1) and phosphomimetic (S500D-hFMR1), in the Drosophila FXS disease model to investigate phosphorylation requirements. At the molecular level, dfmr1 null mutants exhibit elevated brain protein levels due to loss of translational repressor activity. This defect is rescued for an individual target protein and across the population of brain proteins by the phosphomimetic, whereas the dephosphomimetic phenocopies the null condition. At the cellular level, dfmr1 null synapse architecture exhibits increased area, branching and bouton number. The phosphomimetic fully rescues these synaptogenesis defects, whereas the dephosphomimetic provides no rescue. The presence of Futsch-positive (microtubule-associated protein 1B) supernumerary microtubule loops is elevated in dfmr1 null synapses. The human phosphomimetic restores normal Futsch loops, whereas the dephosphomimetic provides no activity. At the behavioral level, dfmr1 null mutants exhibit strongly impaired olfactory associative learning. The human phosphomimetic targeted only to the brain-learning center restores normal learning ability, whereas the dephosphomimetic provides absolutely no rescue. We conclude that human FMRP S500 phosphorylation is necessary for its in vivo function as a neuronal translational repressor and regulator of synaptic architecture, and for the manifestation of FMRP-dependent learning behavior.