Target mRNA abundance dilutes microRNA and siRNA activity MicroRNA Mike needs help to degrade all the mRNA transcripts! Aaron Arvey ISMB 2010 Target mRNA abundance dilutes microRNA and siRNA activity Quic kT ime™ and a dec ompres sor are needed to see this pic ture. QuickTime™ and a decompressor are needed to see this picture. QuickTi me™ and a decompressor are needed to see thi s pi cture. Erik Larsson Chris Sander QuickT i me™ and a decom pressor are needed to see this picture. Christina Leslie QuickTime™ and a decompressor are needed to see this picture. Debbie Marks Background: Small RNAs mediate mRNA degradation microRNA pathway – Protein complex that uses small RNA to guide degradation • microRNAs QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. • Small RNAs are 19-25nt • RNA Induced Silencing Complex (RISC) QuickTime™ and a decompressor are needed to see this picture. – Processed from non-coding genes – Transfected into the cell – Knockdown specific gene – Unintended “off-targets” are also downregulated siRNA pathway QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. • siRNAs (for our purposes) QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. Background: Transfected small RNAs induce target mRNA degradation Qui c kT im are n eede dec ompe™ and d to s ress a ee th or is pi c ture. microRNAs and siRNAs QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. • Double stranded RNA molecules are transfected into cell lines • Concentrations of small RNA are very high, ~100nM • Target mRNAs are degraded QuickTime™ and a decompressor are needed to see this picture. Background: Target Prediction • microRNAs – Transcript 3’ UTR is most likely to be targeted – microRNA 5’ “seed” region guides targeting microRNA targets and siRNA off-targets QuickTime™ and a decompressor are needed to see this picture. • siRNAs – Off-targets have similar targeting rules as microRNAs – Primary targets have nearly exact complementarity siRNA primary targets QuickTime™ and a decompressor are needed to see this picture. microRNAs induce different amounts of downregulation Big Shift Little Shift QuickTime™ and a decompressor are needed to see this picture. Concept: Small RNAs with many targets downregulate each individual target to a lesser extent Meta-analysis of high throughput studies to explore hypothesis • 178 transfection experiments in HeLa and HCT116 cell lines – 61 miRNA-mimics (Lim 2005, Grimson 2007, He 2007, Linsley 2007, Selbach 2008) – 98 siRNA (Kittler 2007, Anderson 2008, Jackson 2006, Schwarz 2006) – 19 chimeras (Lim et al, 2005, Anderson 2008) • Microarray assay pre- and post-transfection • RNA-Seq quantifies mRNA target abundance (Morin 2008) Downregulation is significantly correlated with target concentration QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. Downregulation of siRNA primary target and off-targets is significantly correlated with off-target concentration QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. Shared targets show that downregulation is determined by target abundance Measure target abundance on all targets Measure downregulation on shared targets QuickTime™ and a decompressor are needed to see this picture. Pairwise examples QuickTime™ and a decompressor are needed to see this picture. Examples of differential regulation on shared targets QuickTime™ and a decompressor are needed to see this picture. Pairwise examples QuickTime™ and a decompressor are needed to see this picture. • Smad5 downregulation – miR-155: -1.29 – miR-106: -0.1 • Target abundance – miR-155: 142 – miR-106: 315 • Differences – Downregulation: 1.19 – Abundance: 173 QuickTime™ and a decompressor are needed to see this picture. Shared targets are more downregulated by microRNAs with fewer targets QuickTime™ and a decompressor are needed to see this picture. Conclusion: Small RNAs with more targets downregulate each target to a lesser extent. Consequences: Endogenous regulation by microRNAs • Each microRNA is quantitatively unique – Definition of target should perhaps be different for different microRNAs, targets are likely to be quantitatively different • The cell as a very finely tuned system of regulation – Increase in one target mRNA detracts from downregulation of another target mRNA – microRNA regulation is always using all available degradation machinery (Khan et al2009), but is still stretched thin • Evolutionary constraints – Possibility 1: anti-targets (mRNA transcripts that ‘avoid’ being co-expressed with microRNA) enable the cell to avoid high target concentration – Possibility 2: microRNA expression increases when target mRNAs increase, dosage compensation Consequences: Target abundance limits siRNA activity • Limits knockdown of primary target – May limit drug efficacy, especially in small concentration – May limit functional genomic screens • Limits the knockdown of off-targets – Increase in off-targets may actually decrease toxicity (Anderson et al, 2008) Functional Examples Cancer: PTEN pseudogene 1 (PTENP1) regulates cell cycle by way of PTEN (Poliseno et al) Environmental Response: Non-coding RNA regulates phosphate starvation response (Franco Zorrilla et al) QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. Kinetics • Were we guaranteed to find this result? – No: Depends on dynamic range of kinetic relationship • Degradation is a function of speed, time, and concentration – So far, we have only considered downregulation with respect to concentration • Downregulation has been defined as the ratio: xT x 0 v log log x 0 of molecules x 0 degraded v • Can also consider the total number Background: RISC Kinetics • Multi-turnover enzyme – Single loaded RISC is able to degrade many mRNA transcripts (Hutvágner & Zamore, 2002) • RISC is saturated with small RNA upon transfection (Khan et al, 2009) • Degradation in lysate is very fast (Haley & Zamore, 2004) [RISC] + [target] [RISC+target] [RISC] + [product] Background: RISC Kinetics – Slope of line is velocity – Transcripts degraded at rate of 72-300nM transcript/day • Target concentration in cell is likely to be in the range 1-60nM • 72nM > 60nM – Ignores transcriptional rate – Ignores cellular context – Ignores localization Haley & Zamore (2004) QuickTime™ and a decompressor are needed to see this picture. 60nM 20nM 5nM 1nM Change in molecules (velocity nM/min) • Degradation kinetics depend on target concentration • 1nM RISC in lysate Product (nM) Kinetics in drosophila lysate Target Concentration (nM) Velocity is correlated with target abundance and follows Michaelis-Menten kinetics Velocity can be estimated by xT x 0 x0 x0 x0 Assayed by RNA- Seq (or microarray) xT / x 0 Assayed by microarray Velocity (a.u.) Velocity is correlated with target abundance and follows Michaelis-Menten kinetics QuickTime™ and a decompressor are needed to see this picture. Concentration of Predicted Targets (RPN) Quic kT ime™ and a dec ompres sor are needed to see this pic ture. Questions QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. Debbie Marks QuickT i me™ and a decom pressor are needed to see this picture. Christina Leslie QuickTi me™ and a decompressor are needed to see thi s pi cture. Erik Larsson Chris Sander We control for several alternative explanations • A+U content – Not correlated • 3’ UTR length – Correlated, controllable by shared targets • Expression of individual targets – Correlated, controllable by shared targets Individual target abundance is correlated with downregulation Caveats of shared-target analysis • False positive rate may increase sub-linearly – If false positive rate increases with number of predicted targets, becomes harder to control – The siRNA analysis completely controls for this (since there is only a single primary target!) • Length of UTR is 2x normal length in shared targets – Normal: 1167nt – Shared target: 2041nt – Longer 3’ UTR may lead to increased downregulation, though this would not give preference for a specific microRNA Methods: Target Prediction QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. Methods: Target Abundance QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. Methods: Downregulation QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. Time Course QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. Past Evidence - In Vivo QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. QuickTime™ and a decompressor are needed to see this picture. Franco-Zorrilla et al (2007) Log2 Expression Ratio of primary target Correlation between siRNA off-target abundance and primary target downregulation QuickTime™ and a decompressor are needed to see this picture. Off Target Abundance Past Evidence: Dilution In Solution QuickTime™ and a decompressor are needed to see this picture. Haley & Zamore (2004) Past Evidence - Toxicity Anderson et al (2008) Past Evidence - Dilution In Cells Ebert et al 2007 Normalization QuickTime™ and a decompressor are needed to see this picture.