Transcriptome analysis of mRNAs localized to synapses by high-throughput sequencing

• Main Authors: Yu-Ling Huang, 黃于玲
• Other Authors: Hwan-Ching Tai
• Format: Others
• Language: en_US
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/fn6kg9

碩士 === 國立臺灣大學 === 化學研究所 === 107 === Synaptic dysfunction underlies many neurological disorders such as dementia, autism, schizophrenia, mental retardation, etc. In Alzheimer’s disease (AD), the most common form of senile dementia, synaptic loss precedes neuronal death in the prolonged process of neurodegeneration. The loss of synapse shows a very strong correlation with cognitive decline in AD, stronger than the levels of senile plaques, neurofibrillary tangles, beta-amyloid (Aβ) oligomers, or hyperphosphorylated tau. There are three important proteins involved in the pathogenesis of AD—Aβ, tau, and apolipoprotein E, all of which are synaptic proteins. While much attention has been paid to protein abnormalities at synaptic terminals, it should also be noted that nerve endings and synaptic terminals also contain numerous RNA molecules. These include mRNAs, microRNAs, and long non-coding RNAs. The synapse where proteins are translated often contains polyribosomes, and the newly synthesized proteins play critical roles in synaptic plasticity. However, the typical excitatory synapses in mammalian brains are very small, only about 500 nm in diameter, and they contain a very low copy number of mRNAs. It is therefore extremely challenging to identify mRNAs localized to synapses by conventional fluorescence in situ hybridization (FISH) techniques. To optimize FISH assays for thousands of genes to probe their synaptic localization is largely impractical. We devised a strategy to identify the mRNA composition of synaptic terminals that takes advantage of the sensitivity and high throughput of next generation sequencing (NGS). We prepared synaptically enriched biochemical fractions by simple sedimentation (crude preparation) and by sucrose gradient centrifugation (sucrose preparation). Starting with very small amounts of mouse cortical synaptic terminals, we utilized highly sensitive transcriptome amplification kits designed for single cells to convert their mRNAs into cDNA libraries. Altogether, we identified about 500 transcripts in these synaptic preparations by blasting against mouse genome. Several bioinformatics methods were used to analyze the relationship of these genes, including Gene Ontology and Clusters of Orthologous Groups. They are mostly associated with synapses, extracellular vesicles, ribonucleoproteins, and mitochondria. In total, we obtained about 1.6 billion reads from sequencing runs with 9.5% alignment rate after BLAST. The sucrose preparation shows better synaptic enrichment than the crude preparation but both still contain various contaminating organelles. The low alignment rate suggests that mouse protein databases are insufficient for the full analysis for the mouse transcriptome, which may express unusual gene isoforms and alternative splicing variants. In the future, we will acquire high-purity synaptic terminals by fluorescence activated cell sorting and carry out deeper transcriptome analysis by building custom gene databases.