Spatial centrosome proteome of human neural cells uncovers novel interactors and disease hubs

Despite the crucial importance of the centrosome in brain development and disease, its comprehensive proteome remains uncharacterized in neural cells. Here, we used spatial proteomics to elucidate protein interaction networks at the centrosome of iPSC-derived human neural stem cells (NSCs) and neurons. Centrosome-associated proteins were largely cell-type specific, with protein hubs involved in RNA-dynamics. Analysis of neurodevelopmental disease cohorts identified a significant overrepresentation of NSC centrosome proteins with variants in patients with periventricular heterotopia (PH). Expressing the PH-associated mutant splicing factor PRPF6 reproduced the periventricular misplacement in the developing mouse brain, highlighting mis-splicing of transcripts of the MAP-kinase SAD-A at centrosomal location as essential for the phenotype. Collectively, cell-type specific centrosome interactomes explain how genetic variants in ubiquitous proteins may convey brain-specific phenotypes. Overall design: RNA from FACS sorted electroporated (GFP+) cells was isolated in Extraction Buffer (Arcturus), heated to 42 and stored at -80 cDNA was synthesized from 300 pg of total RNA using SMART-Seq v4 Ultra Low Input RNA Kit for Sequencing (Clontech), according to the manufacturer's instructions. Prior to generating the final library for Illumina sequencing, the Covaris AFA system was used to perform the cDNA shearing, resulting in 200- to 500-bp-long cDNA fragments. The quality and concentration of the sheared cDNA were assessed on Agilent 2100 Bioanalyzer before proceeding to library preparation using MicroPlex Library Preparation kit v2 (Diagenode). Final libraries were evaluated and quantified using an Agilent 2100 Bioanalyzer, and the concentration was measured additionally with Quant-iT PicoGreen dsDNA Assay Kit (InvitrogenTM) before sequencing. The uniquely barcoded libraries were multiplexed onto one lane and 150-bp paired-end deep sequencing was carried out on HiSeq 4000 (Illumina) that generated ~ 30 million reads per sample.