Chromatin factor siRNA screen in epidermal stem cells
Primary human epidermal stem cells (keratinocytes) were transfected with a siRNA library targeting 332 known and putative chromatin factors. These screens were performed in 5 conditions in triplicate. Results are depicted as activity scores and Z-scores, respectively. ..more
Primary human epidermal stem cells (keratinocytes) were transfected with a siRNA library targeting 332 known and putative chromatin factors. These screens were performed in 5 conditions in triplicate. Results are depicted as activity scores and Z-scores, respectively.
Cell culture and differentiation:
Primary normal human keratinocytes (oral lka, foreskin kc and km strains), obtained with appropriate ethical consent, were cultured on feeders as described42. Prior to induction of differentiation, cells were grown, feeder-free, in Keratinocyte Serum Free Medium (KSFM containing 30 mug/ml Bovine Pituitary Extract and 0.2 ng/ml EGF; Gibco) for 2-3 days. At ~70% confluency cells were incubated in KSFM containing 10 muM AG1478 (Calbiochem), 200 ng/ml recombinant human BMP2/7 (R&D systems), both, or 10% foetal bovine serum (PAA).
siRNA nucleofections were performed with the Amaxa 96-well shuttle system (Lonza). Keratinocytes were grown in KSFM to ~70% confluency, harvested and resuspended in cell line buffer SF. 2x105 cells were used for each 20 mul transfection (program FF-113) with 1-2 muM siRNA duplexes. This is equivalent to 5-10 nM siRNA in conventional liposome-based transfections. Transfected cells were incubated at ambient temperature for 5-10 minutes and subsequently resuspended in pre-warmed KSFM. Silencer Select siRNAs were used (Ambion/Applied Biosystems).
We designed a custom library (Silencer Select product, purchased from Ambion) targeting a comprehensive set of human genes encoding known or putative chromatin-factors. We included factors containing any of the following domains or functions: PHD, BROMO, CHROMO, PWWP, tandem BRCT, TUDOR, BAH, MBT, SET (including DOT1L), JMJC, JMJN, PRMT, HAT, HDAC, SIRT, DNMT, MBD, and SNF2 ATP-dependent remodelers. After manual curation for redundant entries a final list of 332 chromatin-associated factors was obtained.
siRNA screening and data processing:
We used passage 2 lka keratinocytes for the siRNA screens. Our custom library of 332 siRNA pools (3 duplexes/pool) was plated in four 96-well plates. Following transfection, keratinocytes were manually dispensed into twenty 96-well plates (8,000 cells/well) containing pre-warmed KSFM. This allowed analysis of quadruplicate plates for each of the five treatment groups (vehicle, AG1478, BMP2/7, AG1478+BMP2/7 and 10% serum). Medium was refreshed the next day. 72 hours after transfection cells were differentiated for 48 hours. Cells were fixed in 4% paraformaldehyde (10 min, RT), washed and permeabilised in PBS+0.2% Triton X-100 (10 min, RT). Following blocking (PBS+10%serum, 30 min, RT), cells were stained using Transglutaminase I specific antibodies (1:2000 mouse monoclonal BC.1 in blocking buffer) for at least 1 hour (RT). After three washes, cells were stained with IR800 anti-mouse secondary antibodies (1:2000, LiCor) and a DNA stain, DRAQ5 (1:2000, Biostatus Ltd), in blocking buffer for 1 hour at RT. Cells were washed 3 times and a final volume of 100 mul TBS was added. Plates were scanned and analysed using the LiCor Odessey system and software with consistent settings throughout the whole screen.
As a transfection efficiency dependent background control we included two independent siRNAs targeting TG1. After scanning and quantification, this background was subtracted and TG1 levels for each individual well were normalised to DRAQ5 signal to give a measure of differentiation/cell for each population of siRNA transfected cells. High data quality was ensured by confirming high pearson-correlation coefficients (Pearson-correlation>0.95) of each replicate versus the mean of the quadruplicates.
A Z-score was subsequently calculated for each population.
Where Chi is the background corrected normalised intensity of a specific well, Alpha and delta are the mean and standard deviation of background corrected normalised intensities of all test siRNAs on the plate, respectively. This calculation effectively transforms the data (without affecting its distribution) to an average of zero and a standard deviation of one. This standardised format allowed us to compile the results of all conditions and experiments into one dataset of ~6,600 quantitative measurements. The three best replicates were used for hierarchical clustering and network analysis.
† RNAi Target.
Data Table (Concise)