Secondary siRNA screen (poly(A)-FISH and RNA StrandBrite)
Image-based genetic screen targeting 463 genes in human HeLa cells using RNA interference (siRNA). Cells stained with DAPI (DNA), poly(A) fluorescence in situ hybridisation (FISH) (mRNA), RNA Strandbrite (total RNA), succinimidyl ester-Alexa647 (total protein). Cells fixed 72h after siRNA transfection. Imaged at 20X magnification with two experimental replicates per condition. Further description available in Berry et al., 2022. Single-cell features provided as background-subtracted mean and sum intensity values for the nucleus, cytoplasm, or whole-cell, as appropriate, for each fluorescence channel. Data cleanup and correction applied as described in the manuscript. Summaries provided on a per-well basis.
Steps to reproduce
900 HeLa cells were plated per well in 384-well plates for reverse transfection onto a mixture of pooled siRNAs dispensed using an acoustic dispenser and Lipofectamine RNAiMAX (0.08µl per well in 10µL OptiMEM) according to manufacturer’s specifications. Cells were subsequently grown for 72 hours at 37°C in a final volume of 40µL growth media, to establish efficient knockdown of the targeted genes. Cells were fixed for 15 min in 4% paraformaldehyde, permeabilised in 0.2% Triton X-100 for 15 min. Cells were stained with RNA StrandBrite according to manufacturer's specifications, then washed, stained with DAPI and imaged. After round 1 of imaging, poly(A) FISH was performed according to the following protocol: Cells were washed into 2X saline sodium citrate buffer (SSC) containing 10% (v/v) formamide. Before hybridisation, samples were transferred to hybridisation buffer (2X SSC, 10% (v/v) formamide, 100mg/mL dextran sulfate), and pre-incubated for 1 hour at 37°C. Fluorescently labelled DNA oligonucleotide probes, dT(30)-Atto488 or dT(30)-Cy5, were purchased from Microsynth as labelled and HPLC-purified. These were diluted into the hybridisation buffer and applied to cells at a final concentration of 400nM. After 16 hours at 37°C in a rotating incubator, cells were washed into 2X SSC, 10% formamide, incubated again at 37°C for 1 hour, before washing into 2X SSC. For combining with subsequent immunofluorescence, poly(A) FISH was imaged before immunostaining. Images were aligned using the DAPI channel, which was included in both imaging rounds, using the computational procedure described previously (Gut et al., Science, 2018) Samples were imaged on an automated spinning-disk microscope (CellVoyager 7000, Yokogawa), which is equipped with four excitation lasers (405, 488, 568, 647nm) and two Neo sCMOS cameras (Andor). A 20X/NA0.75 objective was used. Computational image analysis was used to segment cells and nuclei from max-projected images, and intensity features calculated for all channels in the nucleus, cytoplasm and cell. Mitotic/apoptotic cells were identified and removed using iterative supervised machine learning in tissuemaps. Intensity values were background-subtracted and quadrant effects arising from pipetting inconsistencies were corrected using median normalisation. Intensity values were further median normalized per plate using negative control wells on each plate (scrambled siRNA).