A memory of eS25 loss drives resistance phenotypes

In order to maintain cellular protein homeostasis, ribosomes are safeguarded against dysregulation by myriad processes. Remarkably, many cell types can withstand genetic lesions of certain ribosomal protein genes, some of which are linked to diverse cellular phenotypes and human disease. Yet the direct and indirect consequences from these lesions are poorly understood. To address this knowledge gap, we studied in vitro and cellular consequences that follow genetic knockout of the ribosomal proteins RPS25 or RACK1 in a human cell line, as both proteins are implicated in direct translational control. Prompted by the unexpected detection of an off-target ribosome alteration in the RPS25 knockout, we closely interrogated cellular phenotypes. We found that multiple RPS25 knockout clones display viral- and toxin-resistance phenotypes that cannot be rescued by functional cDNA expression, suggesting that RPS25 loss elicits a cell state transition. We characterized this state and found that it underlies pleiotropic phenotypes and has a common rewiring of gene expression. Rescuing RPS25 expression by genomic locus repair failed to correct for the phenotypic and expression hysteresis. Our findings illustrate how the elasticity of cells to a ribosome perturbation can drive specific phenotypic outcomes that are indirectly linked to translation and suggests caution in the interpretation of ribosomal protein gene mutation data. RNA-seq of polyadenylated mRNAs as the first step to acquire a snapshot of the transcriptional state of wild-type, RPS25 KOs, and eS25-HA AB HAP1 cells HAP1 cell lines were seeded into the wells of 6-well plates with equal live cell counts (250,000 cells/well), grown under standard conditions, and harvested by trypsinization, pelleting and washing cells with PBS 48 hrs post-seeding. To normalize growth conditions prior to harvest, media was removed 6 hrs prior to harvest and replenished with fresh media. RNA was purified from cells pellets using the PureLink RNA mini kit (ThermoFisher cat.#12183025). Prior to library preparation, RNA concentration and sample quality by RNA integrity number (RIN) were checked using Tapestation RNA ScreenTape reagents (Agilent cat.# 5067-5576). cDNA libraries were prepared with the SureSelect Strand-Specific RNA Library Prep Kit (Agilent cat.#G9691B) on an Agilent Bravo Automated Liquid Handling Platform accordingly to the protocol (Version E0, March 2017, G9691-90030). Library concentration and integrity were checked using Tapestation D1000 ScreenTape reagents (Agilent cat.#5067-5582) and Qubit dsDNA BR Assay Kit (Invitrogen cat.#Q32850). Sequencing was performed on an Illumina HiSeq 4000 with 2x101 base pair reads and Illumina Single Index. Reads were aligned to the hg38 reference genome using STAR v2.5.3a (Dobin et al., 2013) and differential expression between samples was computed using R v3.4.0 and the DESeq2 package (Love et al., 2014) (detailed pipeline and options available on https://github.com/emc2cube/Bioinformatics/). In the rare case that fold change P-values were too small for R to calculate and estimated as 0, these were imputed at 10-300 to facilitate logarithmic conversion.