This document outlines detailed experimental protocols for molecular cloning, adeno-associated virus (AAV) production, in vivo animal studies, in vitro primary neuron culture, and various biochemical and bioinformatic analyses.
Molecular Cloning and AAV Production
The Ribo-STAMP construct, initially a Tet-On hRPS2–APOBEC lentiviral expression vector, was meticulously cloned into pLIX_403 derivatives. This process incorporated P2A-mRuby fluorescent protein sequences. Further variants, including Rpl10a–APOBEC and Rpl22–APOBEC fusions, were generated as synthetic gBlocks and subsequently cloned into a parental backbone.
Tet-On Ribo-STAMP constructs were then cloned into mammalian gene expression AAV vectors (VectorBuilder), which included WPRE and BGH polyA signals, utilizing Gibson assembly for efficient integration. A constitutive RPS2-STAMP construct, designed specifically for in vivo experiments, was also assembled into an AAV vector. Finally, AAV virions were prepared for packaging using the PHP.eB capsid.
Animal Protocols
C57BL/6J mice were utilized for these studies and maintained under group-housed conditions. All experimental protocols strictly adhered to guidelines set by the National Institutes of Health (NIH) and received official approval from the Institutional Animal Care and Use Committee at Scripps Research Institute.
Primary Neuron Culture and Treatments
Cortices were carefully dissected from E18 C57BL/6J embryos, subsequently dissociated, and then plated. Neurons were maintained in neurobasal medium to support their growth. To limit the proliferation of dividing cells, Floxuridine was introduced at DIV 3.
At DIV 6, neurons were infected with inducible Ribo-STAMP AAV virions. Ribo-STAMP expression was induced with Dox for a duration of 48 hours before cell collection at DIV 14. For various experimental conditions, BDNF, Anisomycin, and Puromycin were applied for specific durations immediately prior to cell collection. Cell viability throughout these experiments was routinely assessed using the Cell Counting Kit Assay-8.
Biochemical Assays
Proximity Ligation Assay (PLA)
PLA was performed using the Duolink kit following puromycin treatment. The protocol involved several critical steps: initial fixation, permeabilization, blocking, primary and PLA probe incubations, ligation, and amplification. This was followed by subsequent immunocytochemistry for visualization.
Immunocytochemistry (ICC)
PFA-fixed cells were prepared for ICC by permeabilization and blocking. They were then incubated with primary and secondary antibodies, mounted, and subsequently imaged using a Nikon A1 confocal microscope.
FUNCAT and Immunohistochemistry (IHC)
P25 C57BL/6 male mice were co-injected intraperitoneally with AHA and either PTZ or PBS. Brains were then perfused, postfixed, sliced, and subjected to click labeling. Following this, slices were prepared for IHC, which included blocking, primary and secondary antibody incubations, mounting, and final imaging.
Ribosome Structure Analysis
Mouse ribosome structures were accessed from the Protein Data Bank (PDB entries 7CPU, 6Y0G, 7LS2). PyMol software was utilized for subsequent structure superimposition and the generation of illustrative figures.
RNA Sequencing and Bioinformatic Analyses
eCLIP Sequencing
Anti-RPS2 antibody was employed for eCLIP sequencing. This was followed by total RNA isolation, ribosomal RNA depletion, and single-end sequencing on an Illumina NovaSeq platform. Reads were then aligned to the mm10 genome using STAR, and eCLIP-based translational efficiency was calculated.
Ribo-STAMP Bulk RNA Sequencing
RNA was isolated, and Illumina TruSeq Stranded mRNA libraries were prepared. Sequencing was performed on an Illumina NovaSeq. FASTQ files underwent trimming, alignment with STAR, and processing using the SAILOR pipeline and feature counts to derive EPKM (edited reads per kilobase million) and EPR (edited reads per total reads) values.
Differential Transcription vs. Translation (BDNF)
DESeq2 was applied to calculate RNA log2 FC and P-values. A two-sided t-test was used to compare BDNF-treated and no-treatment EPR values, with Benjamini–Hochberg correction for multiple comparisons.
Gene Ontology Analyses
Gene Ontology analyses were performed using clusterProfiler in R studio and SynGo for the identification of enriched synaptic terms.
Single-Cell Library Preparation and Analysis (scRNA-seq)
Single-cell suspensions were prepared from the hippocampi of EF1a-Ribo-STAMP virus-injected mice. Libraries were constructed using 10x Genomics kits and subsequently sequenced. Cell Ranger, Scrublet, and Scanpy were employed for alignment, doublet removal, quality control, batch correction, and cell assignments.
A Python package was developed for rapid, memory-efficient identification of RNA edits at single-cell and isoform-level resolution, benchmarked against existing tools.
MARINE Tool Development
A Python package named MARINE was developed for the rapid and memory-efficient identification of RNA edits at both single-cell and isoform-level resolution. This tool was rigorously benchmarked against existing methodologies.
Short-Read EPR Pipeline
This pipeline involved several steps: filtering edit sites, calculating EPR values, normalizing to Stamp expression using linear regression, and applying cell-specific normalization.
Clustering Cells by EPR
To overcome edit sparsity, pseudocells were created by grouping cells from the same cell-type assignment. This enabled UMAP and Leiden clustering based on EPR to validate cell-type specificity.
Differential EPR Analysis
Scanpy’s rank_gene_groups method was utilized to identify EPR marker genes for each annotated cell type.
EPR Relationship to RNA
Pseudobulk EPR and RNA counts were calculated per cell type and sample. This was followed by filtering and Spearman correlation analysis to understand their relationship.
Translational Cell State Analysis
Gaussian mixture modeling was used to categorize CA3 cells into high/low translation groups based on their editing rate. Differential gene expression analysis using Scanpy was then performed for EPR.
Differential EPR vs. RNA (CA3 vs. CA1)
EPR differences between CA3 and CA1 neurons were assessed across replicates, with vectorized bootstrap resampling used for statistical significance. Correlations to existing RiboTag, RIBOmap/STARmap, and human hippocampus mass spectrometry datasets were also performed.
Single-Cell Long-Read Library Preparation and Analysis (MAS-seq)
10x Genomics 3′ cDNA was processed using MAS-seq. HiFi reads were processed with SKERA, the PacBio workflow, and IsoQuant for isoform assignment. MARINE identified C>T edit loci, aggregated them by isoform, and calculated EditsC (ratio of edited cytosines to total cytosines).
Long-Read Pseudobulk EditsC and RNA
Pseudobulk EditsC and RNA values were calculated by summing edits/reads across cells per sample and then averaging across samples.
Cell-Type-Specific EditsC and RNA Analysis
Pseudobulk datasets were filtered for transcripts possessing multiple isoforms per gene. Heatmaps were then generated to visualize cell-type-specific patterns.
EditsC and RNA Isoform Correlations
Pairwise Spearman correlations were computed between pseudobulk EditsC and RNA counts for genes exhibiting two or more isoforms.
Transcript Feature Analysis
This analysis involved obtaining 5′ UTR and 3′ UTR coordinates to calculate GC content, perform motif analysis (MEME, TomTom), and assess miRNA binding (TargetScanMouse) and ELAV binding sites.
Genome Browser Tracks
BED files containing SNP-filtered edit information were generated and subsequently visualized using IGV (Integrative Genomics Viewer).
Differential Transcript Analysis
Pairwise comparisons between cell types for pseudobulk EditsC and RNA expression were performed, aiming to identify discordant translated isoforms.
Statistics and Reporting
Experiments used at least three biological replicates, with mice randomly assigned to groups.
Experiments consistently utilized at least three biological replicates, with mice randomly assigned to experimental groups. While efforts were made to ensure robust data collection, blinding of experimenters was not always feasible. EPR and EditsC analyses were conducted without specific adjustments unless explicitly stated within the protocol.