Dear colleagues,
I have been using siRNA interventions for some time now to study the impact of my proteins of interest on gene transcription.
For these investigations I have been transfecting epithelial cells (BEAS-2B) using individual siRNA's (Qiagen) delivered by RNAiMAX lipofectamine (1 ul per ml) at a final concentration of 5 - 10 nM.
For each target, I have been testing 4 individual siRNA's, which resulted 48-hour post transfection in clear knock-down of the proteins of interest on protein level (knockdown of 80-90%) compared to scrambled controls and lipofectamine.
However when examining the gene induction after my chosen stimulation, I observe clear "off-target" effects for particular siRNA's. As I observe an enhancement/reduction of gene induction after single siRNAs that are not shared among the other siRNAs targetting the same protein.
Unfortunately these off-target effects are quite common, so I wonder how I can reduce these off-target effects to make interpretation of the data more straightforward.
I am quite sure more people experience similar kind of problems, so I hope you can share with me any tips/suggestions around this topic!
Take a look at PMID 24875475. Maybe won't solve your problem in the short term, but will definitely draw your attention.
Hello Martijn,
a simple method is to make a BlastN with your siRNA sequence in the NR or transcriptome of the desired organism. You will find then if your siRNA is complementary to other transcrips and causing maybe a knock down.
You should also consider that the knock down of your gene of interest can have severe effects on the other interaction partners in the pathway. You can also see the dereguation of the other genes as prove that your gene is required for the signalling. On my own I have determined 2 fold and even higher changes of transcription factor expression after knocking down a chaperone indicating that they are highly cross-linked and regulated confirming that the determined proliferation inhibition of the cancer cells is due to the loss of this particular gene.
Think over this possibility and Alejandro´s advice with siPools is also good attemt but if it is not making sense in your setup you can proceed with the better siRNAs doing no weired stuff because several journals only want that you show 2 tested siRNAs.
Regards
Hi Martijn,
In a study, only a small fraction of the experimentally validated off-targets were identified by in silico methods based on alignment and sequence comparison tools, suggesting that overall sequence identity was a poor predictor of the number and identity of off-targeted genes. A report from Dharmacon suggests certain approaches to reduce off target effects. You may go through this report, I found it useful.
http://dharmacon.gelifesciences.com/uploadedFiles/Resources/off-target-tech-review-technote.pdf
Another Nature Reviews Drug Discovery article addresses some of these points:
http://www.nature.com/nrd/journal/v9/n1/full/nrd3010.html
Regards
Hi Martijn,
Beside siRNA pool effect, your observation can also be related to the transfection reagent itself, you can refer to the following publication that compared up to 26 reagents for HTS siRNA screening
https://www.ncbi.nlm.nih.gov/pubmed/24661213
Best regards
Olivier
Thank you all for your replies. I found them very helpful!
In regards of Olivier's comment, I can be quite clear that is not transfection-reagent related in my situation. Because the induction of my genes of interest in untransfected and lipofectamine treated cells are similar. And also my scrambled- and Lamin siRNa (non-target) sRNA behave quite similar to the lipofectamine-treated cells.
I like the idea of using pooled siRNAs! I do however wonder how you can exclude the possibility that one of the siRNAs in your pool is not contributing to the phenotype you observe in your assay?
The chances on these off-target effects are probably even lower when using the by Alejandro mentioned siPool of 60 siRNA's compared to the more standard pools of 4 siRNAs. However I did notice these off-target effects by single siRNA administration also at relatively low concentrations of siRNA (< 1 nM).
Have people used this siPool for publication? And if so what was the reviewers response on validation?
If you are seeing 'off-target' effects at 1 nM, I think this is a 'specific sequence effect'. I've had siRNAs with 5 mismatches against a target mRNA that still functioned. A paper in Nucleic Acids Research (2005) demonstrated that as little as 7 nt complementation with the target mRNA was enough to induce RNAi.
Off-target effects usually arise from (1) the sense strand, (2) the antisense strand, or (3) competition with endogenous miRNAs (Nat Biotech 2009). You can't do anything about #3. You can eliminate the sense strand from being incorporated into RISC by sequence design (more GC in position 1 and 2, more AU in position 16-19) or chemical modification (2-O-methyl on position 1 and 2). To prevent the antisense strand from being used as a miRNA, one study showed 2-O-methyl modification on position 2 was helpful. I've had better luck with position 3 on the antisense strand.
I'm not a fan of pools. As I feel that most siRNA have some off-target effects, why add more siRNAs that result in more off-targets?
The best and only way to show siRNA specificity is rescue experiments (cotransfect plasmid expressing target mRNA of interest with silent mutations to prevent siRNA targeting). If the off target effects are reduced/eliminated, the effects are due to knockdown of your target. If the effects remain, they are 'off-target effects' but they are sequence specific.
Hi John,
It's nice explanation about the off-target site. I have a question regarding to selection of antisense oligo nucleotide. Suppose this is the designed antisense oligo for Nicotiana tabacum (NtVI A3.6 722L20 5’CTATCTTCTTCTGCACTTCG 3’). I want to check on target and is their any off-target site.when i am doing homology search of this antisense with N.tabacum database then how i can determine that this designed antisense oligo have no any off-target site. what are the range of e-value that will be acceptable.
Hi Martijn, our company developed siPOOLs and they have been used for several years now with good feedback from customers. There is also a growing list of publications found here: https://www.sitoolsbiotech.com/publications.php that demonstrate siPOOLs are well-recognized by the research community as a method of validating loss-of-function phenotypes.
For more info about siPOOLs, you can refer to this pdf: https://www.sitoolsbiotech.com/pdf/170926_siPOOLsProductBrochure2017_webPDF.pdf
We also have a blog that dishes out several interesting articles a month on functional genomic screening: https://blog.sitoolsbiotech.com
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology