Mutation in PHACTR1 associated with multifocal epilepsy with infantile spasms and hypsarrhythmia

Abstract A young boy with multifocal epilepsy with infantile spasms and hypsarrhythmia with minimal organic lesions of brain structures underwent DNA diagnosis using whole‐exome sequencing. A heterozygous amino‐acid substitution p.L519R in a PHACTR1 gene was identified. PHACTR1 belongs to a protein family of G‐actin binding protein phosphatase 1 (PP1) cofactors and was not previously associated with a human disease. The missense single nucleotide variant in the proband was shown to occur de novo in the paternal allele. The mutation was shown in vitro to reduce the affinity of PHACTR1 for G‐actin, and to increase its propensity to form complexes with the catalytic subunit of PP1. These properties are associated with altered subcellular localization of PHACTR1 and increased ability to induce cytoskeletal rearrangements. Although the molecular role of the PHACTR1 in neuronal excitability and differentiation remains to be defined, PHACTR1 has been previously shown to be involved in Slack channelopathy pathogenesis, consistent with our findings. We conclude that this activating mutation in PHACTR1 causes a severe type of sporadic multifocal epilepsy in the patient.


| INTRODUCTION
Epilepsy is a diverse group of disorders characterized by recurrent seizures. The idiopathic form of epilepsy is thought to be genetically determined, 1 and over 500 genes have been associated with epilepsy in human and mouse. 2 Most cases are associated with mutations in genes encoding subunits of ion channels, neurotransmitter receptors or synaptic vesicle cycle proteins that play essential roles in neuronal activity. 3 Nevertheless, these findings do not explain all cases of congenital forms of epilepsy, and new genes continued to be uncovered. 4 Protein phosphatases and protein kinases have a crucial role in the functioning of the mammalian brain. The coordinated activity underlies two major forms of synaptic plasticity known as long-term potentiation and long-term depression (LTD). 5,6 LTD-inducing stimuli promote distribution of PP1 to dendritic spines, 7 where its association with the actin-rich postsynaptic density (PSD) is critical for effective dephosphorylation of its substrates and down-regulation of synaptic function. [8][9][10] Numerous pathogenic conditions involve altered function of PP1 in the brain. [11][12][13][14] Epilepsy can also be regarded as exhibiting synaptic plasticity with elevated excitatory synapse activity. 15 Modulation of kinase/phosphatase balance has been shown to induce status epilepticus and/or seizure activity in rats. 16,17 Two members of the Phactr family of G-actin binding PP1 cofactors, PHACTR1 and PHACTR3, are highly expressed in the brain. 18,19 Interaction with PP1 appears critical for the function of PHACTR proteins. 20,21 The presence of PHACTR1 in the PSD indicates that it is likely to play a role in neuronal function, 18 while PHACTR3 was shown to inhibit axon elongation in primary rat cortical neurons via regulation of actin dynamics. 19 Moreover, Phactr3 was recently shown to be necessary for regulation of neuronal morphology by increasing the percentage of mature dendritic spines as well as up-regulating dendritic complexity. 22 Phactr family proteins interact with G-actin through four conserved RPEL motifs (named after ultraconservative amino-acids building up the motif). 18,[23][24][25][26] Studies in non-neuronal cells have shown that PHACTR1/G-actin interaction controls its subcellular localization and inhibits its interaction with PP1. 23,24 A recent study has demonstrated that PHACTR1 specifically interacts with Slack, a Na + -activated K + channel (K Na 1.1) which is encoded by KCNT1 gene, and dissociates from it upon channel activation. 27 Mutations in Slack are associated with malignant migrating partial seizures of infancy (MMPSI), characterized by infantile seizures and intellectual disability. 28 Interestingly, MMPSI mutations in Slack both constitutively activate the channel, and affect its interaction with PHACTR1, raising the possibility that the MMPSI phenotype may in part reflect altered PHACTR1 activity. Indeed, recently, it was demonstrated that de novo mutations in PHACTR1 gene are associated with West syndrome characterized by infantile spasms, hypsarrhythmia, and mental retardation. 29 Although authors do provide evidence for the pathogenic role of these PHACTR1 mutations in a rodent model how this relates to pathogenesis of West syndrome remained unclear.
Here we used whole-exome sequencing (WES) to identify a novel mutation in PHACTR1 that occurred de novo in a proband with multifocal epilepsy with infantile spasms and hypsarrhythmia. We show that this mutation, L519R, reduces the affinity of PHACTR1 for G-actin and increases its propensity to form complexes with the catalytic subunit of PP1 in vitro. These properties are associated with altered subcellular localization of PHACTR1 and increased ability to induce cytoskeletal rearrangements in vivo. We propose that L519R acts as an activating mutation of PHACTR1. Overall, our study provides a clue to the molecular pathogenesis of PHACTR1-associated epileptic encephalopathy of infancy.
2 | MATERIALS AND METHODS

| Source of DNA
Peripheral blood samples were collected from affected individual and his parents as well as from 120 unrelated healthy control individuals of the Central Russian origin. Genomic DNA from the samples was extracted using standard protocols.

| Cell lines, transfection, and immunofluorescence microscopy
Mouse NIH3T3 fibroblasts were maintained in DMEM with 10% FCS. Immunofluorescence microscopy was performed as described earlier. 24,31 Phactr1 mutants were detected by anti-FLAG antibody (Sigma F7425). F-actin was visualized using FITC-phalloidin (Invitrogen), and nuclei were detected with DAPI.

| Proteins and peptides
Actin was isolated from rabbit skeletal muscle essentially as described 32,33 and saturated with Latrunculin B (Calbiochem) as described previously. 34 N-terminally fluorescein isothiocyanate (FITC)-conjugated peptides were synthesized and HPLC-purified by the Crick Peptide Chemistry

| Fluorescence anisotropy
The fluorescence anisotropy assay was performed as described earlier. 34 Dissociation constants (K d ) were calculated by nonlinear regression in Prism software using equation: where X is protein concentration, Y is total anisotropy, A b is anisotropy from bound ligand and A f is anisotropy from free ligand. 36 K d values were derived from three independent experiments each in duplicate.

| Co-immunoprecipitation
Cells transfected with Flag-Phactr1 and HA-PP1 were lysed in lysis buffer (0.5% Nonidet P-40, 1 mM EDTA, 50 mM Tris pH 8.0, 120 mM NaCl, 0.1 mM sodium vanadate, protease inhibitors). In 1 mg of total protein was used to immunoprecipitate with anti-HA-agarose beads (A2095 SIGMA). After three washes in lysis buffer, proteins were separated by SDS-PAGE and immunoblotted using anti-FLAG (Sigma

| The L519R mutation reduces G-actin binding affinity, thereby increasing PP1 binding
The activity and intracellular location of Phactr1 are regulated by G-actin, which inhibits importin α-β binding to nuclear import signals associated with the Phactr1 N-terminal RPEL motif and Cterminal RPEL domain, and PP1 binding to its conserved Cterminal sequences. 23,24 The L519R mutation lies at a conserved position within RPEL3 (Figure 2(A)). 24 While there is no obvious steric clash between the more bulky arginine side chain and actin, its charged character would be expected to weaken the interaction with the hydrophobic ledge, potentially reducing G-actin binding affinity (Figure 2(B)). To test whether the L519R mutation affects G-actin binding affinity we used the fluorescence polarization anisotropy assay, with FITC-labeled RPEL3 peptides. The  starved and serum-stimulated conditions (Figure 3(C)). These results are consistent with a model in which the decreased affinity of PHACTR1 L519R for G-actin decreases its ability to inhibit PP1 binding, thereby potentiating formation of the PHACTR1/PP1 complex (see Discussion).

| The Phactr1 L519R mutant exhibits aberrant subcellular localization and function
We previously showed that in NIH3T3 fibroblasts, G-actin binding is critical for regulation of PHACTR1 subcellular localisation and for PP1-dependent induction of cytoskeletal rearrangements. 24 We therefore next examined the effect of L519R mutation on PHACTR1 localisation. In NIH3T3 fibroblasts, serum stimulation induces PHACTR1 nuclear accumulation, and the integrity of the C-terminal RPEL motifs, including RPEL3, is required to maintain its cytoplasmic localisation in resting cells. 24  The PHACTR proteins each contain four RPEL motifs: one at the N-terminus, and three within the C-terminal RPEL domain. The most C-terminal motif, RPEL3, overlaps the sequences required for interaction with PP1. Our in vitro G-actin binding experiments indicate that the L519R mutation reduces the RPEL3 affinity some 3-fold. In contrast, mutation of the core RPEL3 arginine, R507A, reduces affinity 18-fold. Nevertheless, in co-immunoprecipitation experiments we found that PHACTR1 L519R exhibited greatly increased association with PP1, to an extent comparable to the R507A mutant. We suggest that G-actin concentrations in the expression assays are such that the relatively small decrease in binding affinity is sufficient to tip the balance of interaction in favor the formation of PHACTR1/PP1 complex.
In fibroblasts, over-expression of PHACTR1 RPEL mutants that exhibit defective actin binding induces assembly of thick F-actin fibers and F-actin foci, and this requires PP1 binding. 24 We found that the L519R mutant also exhibited elevated levels of F-actin remodeling, comparable to those seen with the R507A mutant. This result is consistent with the increased ability of L519R to interact with PP1.
In addition to its effects on PP1 interaction and cytoskeletal rearrangement, G-actin binding also affects the subcellular localization of PHACTR1. In transfected fibroblasts, PHACTR1 is predominantly localized in the cytoplasm, because binding of importin αβ to NLS associated with RPEL motifs, is prevented by G-actin binding. 24 Analysis of core arginine mutations at individual RPEL motifs showed that RPEL3, R507A, had the biggest effect on PHACTR1 subcellular localization. 24 In line with the effects seen in the PP1 binding analysis, Given the effects of the L519R mutation on PHACTR1 subcellular localization as well as its interaction with PP1, it is therefore possible that its phenotype reflects redistribution of PP1.
The PHACTR1 L519R mutation acts positively to increase formation of the PHACTR1PP1 complex, and therefore stands in contrast to the mouse humdy mutant, which carries a mutation that prevents formation of the Phactr4/PP1 complex. 20 (Figure 2(A)). While these data emphasize the detrimental effect of changes at PHACTR1 amino-acid 521, further work is needed to establish the molecular basis of the connection between the p.R521C mutation and epilepsy.
In conclusion, we have identified an activating mutation, L519R, in the PHACTR1 gene in a sporadic case of multifocal epilepsy. By altering PHACTR1's ability to bind G-actin, L519R effectively activates the PHACTR1/PP1 complex, leading to enhanced dephosphorylation of target substrates. G-actin binding also controls PHACTR1 subcellular localization, and the L519R mutation could also exert its effects through changes in PHACTR1/PP1 substrate targeting. Our findings suggest that actin treadmilling and PHACTR1 is involved in regulation of the phosphorylation/dephosphorylation balance in synapses and that activation of PHACTR1/PP1 is involved in epileptogenesis. This could be used for etiotropic therapy opportunity as well as for genetic counseling of the affected family. Our study adds PHACTR1 to the list of epilepsy-related genes and contributes to the diverse molecular mechanisms of the disease.