I can think of three explanations: (1) the shRNA could have inserted into the chromosome in a region that is transcriptionally silenced, such that the shRNA is not expressed. I don't have a reference for you for this, but I do have experience with a transgene knock-in that got silenced, which is similar. (2) The particular RNA sequence targeted is simply less efficient due to secondary structure of the target mRNA. (3) Most likely, there is some sequence bias for the RNAs, some of which can be modeled based on studies that have looked at random siRNAs and seen which target reporters more efficiently than others. There could also be oher unpredictable idiosyncrasies that we have yet to understand for particular sequences. I read on the RNAi Consortium's website that their newest version of shRNA library will have GFP transgenes that contain each shRNA-matching sequence to measure GFP knockdown and thereby assess any sequence bias because of that particular sequence. Hope that helps!

I can think of three explanations: (1) the shRNA could have inserted into the chromosome in a region that is transcriptionally silenced, such that the shRNA is not expressed. I don't have a reference for you for this, but I do have experience with a transgene knock-in that got silenced, which is similar. (2) The particular RNA sequence targeted is simply less efficient due to secondary structure of the target mRNA. (3) Most likely, there is some sequence bias for the RNAs, some of which can be modeled based on studies that have looked at random siRNAs and seen which target reporters more efficiently than others. There could also be oher unpredictable idiosyncrasies that we have yet to understand for particular sequences. I read on the RNAi Consortium's website that their newest version of shRNA library will have GFP transgenes that contain each shRNA-matching sequence to measure GFP knockdown and thereby assess any sequence bias because of that particular sequence. Hope that helps!

shRNAs are great tools to use in the lab, however, one major flaw is the fact that many shRNAs are not that potent in yielding efficient target knockdown. When designing our own shRNAs, we often see that only ±2 out of 5 shRNAs give a target knockdown of over 60%. Sometimes this is higher, but it's rare that all shRNAs targeting a given gene are all very potent. This can be caused by many reasons, as already addressed by Beverly Dancy. Unfortunately, we still cannot design shRNAs that all work very well, the parameters for this are poorly understood. siRNAs on the other hand often always work, but for shRNAs it remains a major challenge.  

Also, some targets you simply cannot knockdown very efficiently since loss may be toxic to the cells. In these instances, you will only end up with living cells when the shRNAs are less effective.

Hence, it's not strange that your cells are selected OK by using a selection drug but don't show knockdown. Just screen more of them...

cheers

RJ

Check out the qRT-PCR and blot data for several shRNA sequences targeted against p300 and CBP in this supporting figure for my paper, as an example of getting a variety of knockdown efficiencies. This kind of result is fairly typical, and after the initial experiment, you can choose your best line(s) for further experiments (we chose p300 #04 based on the data shown).

http://onlinelibrary.wiley.com/doi/10.1002/cbic.201200381/suppinfo

Hi,

I'm agree with previous comments. If you have your cells growing on selection marker such as puromycin, it is very unprobable that only your shRNA is silenced and not your puromycin cassette. Most probably is your shRNA sequence not really efficient. From my experience I have designed 10 shRNA on pLKO.1 for same RNA target, by quantitative WB only 2 gave me 80-90% KD in compare to scramble and mock cells, some of them 50% and some of them not KD at all!!...for this reason I suggest you to test different shRNA sequence or buy commercial pLKO with shRNA sequences already validated functionally. Another isuue coul be try different MOI and also as it was suggested by Sandra double infection, although doing that you should change resistance marker of one of them to have double selection such as Puromycin/ Blasticidin. Good Luck!!

I will only add to the above the more unlikely scenario. Have you checked for how many splice isoforms your gene can produce and their relative percentage transcripts value which can change upon alteration of experimental conditions, etc…? It is possible that your shRNA is targeting an isoform which doesn’t correspond to the one you are checking by RT-qPCR. Normally, it is mandatory to perform western blot against your protein.

What is the promoter you use to drive the shRNA. Is it a polymerase III promoter, U6 for example, or is it a polymerase II promoter? The later can be silenced. The ability of the shRNA to yield knockdown depends a lot on the processing of your shRNA as well, did you do Northern to assess processing of the construct? Usually in my experience you can have high knockdown and good processing from U6 simple hairpin constructs, but this is more a betting game. While pol II driven pre-microRNA expressed constructs (especially human pre-microRNA 30) are having a quite reliable processing and furthermore RISC incorporation based on strand bias delivered by the pre-microRNA backbone.

Of course pol II promoters produce significantly lower initial transcript numbers, but with high efficiency of processing and incorporation that can offset an U6 driven shRNA that is poorly processed and has an unfavorable strand bias for example (assuming the intended processed/mature siRNA would be potent).

Another problem is the level of your target gene. The lower the expression level of a target gene, the harder it is often to knock down. I've found in primary cells that especially adaptor proteins in signalling cascades can be a huge pain. In this case I would recommend you try a combinatorial construct with 2-3 shRNAs against the same target at different sites (distances of 23-53 of the target site seem to act often synergistic) but express these definitely from a pol II promoter to avoid perturbation of the endogenous microRNA regulatory layer.