siRNA woes

January 4, 2007 3:15 PM

Attention Biochemists or Anyone Else working with RNAi-expressing cell lines: Help me troubleshoot!

I am at my wit's end with this and thought the hive mind might be able to help. My sanity and PhD may also depend on this being resolved in a timely manner as well.

I have been having some trouble with a double stable cell line that is supposed (and used to) inducibly express an siRNA molecule in the presence of tet or dox.

The siRNA sequence was created by another member of my lab and works when transiently transfected into cells via oligofectamine. This has been verified by me and several other members of the lab.

The base cell line is the TRex cell line available from Invitrogen which is blasticidin resistant and expresses the tetR protein and is the inducible component of the system.

The construct the siRNA sequence was cloned into is the pSuper+neo+gfp vector. This construct is stably expressed by the TRex cell line. The resulting cell line, "16", is resistant to both G418 and blasticidin as well as expresses GFP for sorting purposes. The member of the lab, AB, who made this line has moved on but clearly showed knockdown of our gene (via western blotting) within 48 hours. She froze down only one vial (ugh) and left for greener pastures (ie dental school).

Now, my project requires the use of these cells so I woke up the only vial. I immediately froze down 10 vials before carrying the cells in the presence of drugs. The cells have been carried in the presence of 200 ug/mL G418 and 10 ug/mL blast since. Unfortunately, somewhere along the line they've lost the siRNA expression. Treatment with dox does not cause target knockdown. We don't have an antibody to the tetR protein so I don't know if it's expressing and I don't think we have a way to test for siRNA expression.

AB also made several other clones of this double stable which showed knockdown and were frozen down. We awoke all and tested for knockdown, none of which were positive.

So Hivemind: Why would these cells still be resistant to the drugs but cease to knockdown the protein in question, even fresh from the liN2? It's not the dox as we have bought fresh powder and tested it with bacteria (no growth). Invitrogen states that there is no difference between tet and dox when inducing tetR expression. I am losing my mind over this. I cannot, nor any of my colleagues, think of why this is wrong. Is there a mefites who has worked with TRex's or a similar system and can offer advice?

Hope me, mefites, you're my only hope.

posted by LunaticFringe to science & nature (15 answers total)

Can you do qPCR to detect tetR expression? How's the GFP?
posted by greatgefilte at 3:31 PM on January 4, 2007

Also, was there any documented knockdown after AB left? Any possibility that he/she might have manipulated the Westerns? In my old lab, we had someone who caused endless trouble for the grad student that succeeded him because he "mislabeled" a Southern.
posted by greatgefilte at 3:36 PM on January 4, 2007

I would check more thoroughly for target knock-down. Don't use a Western, use a more sensitive RT-PCR or even Q-PCR to verify that the target's transcript levels are indeed unchanged. Also, the target is endogenous correct? Its not transfected or infected such that dosage might affect ability to ascertain knockdown level by the siRNA?

Incidently, is the GFP inducible or always on? Either way, I would also verify that the resistant cells are also green fluorescent as designed, just to verify that the insert locus is still functioning correctly. It'll take you about 5 minutes to do this.
posted by dendrite at 3:39 PM on January 4, 2007

I've used pSuper-GFP (which has the original pSuper H1 promotor and shRNA cloning site cloned into pEGFP) to transiently transfect cells. I did try in the past to make a stable cell line, but it lost GFP expression after a few passages (under selective conditions, of course) and after that continued to just lose transient expression. So my first question is: do you still have GFP expression?

I'll try to think about it a bit more, but I have a cold and my head is all stuffy and not able to think about RNAi. (Normally I do know a lot about it, really, so feel free to e-mail me - then I can also give you more details that I might not want to put online. E-mail's in profile.)
posted by easternblot at 4:55 PM on January 4, 2007

I meant "after that continued to just USE transient expression"

(I said I wasn't thinking properly!)
posted by easternblot at 4:56 PM on January 4, 2007

Did the siRNA construct get cut into a single strand before getting transfected and selected?

Sometimes a construct will integrate into chromosomal RNA, but the section of interest (ie., the bit that expresses the siRNA) gets borked, but the antibiotic resistance loci are intact.

Second/third the recommendations to PCR to see if the siRNA is getting expressed.
posted by porpoise at 4:56 PM on January 4, 2007

...chromosomal DNA..., damnit.

... and linearizing as opposed to single-stranding, that is.

Linearizing it improves the chances that it'll integrate all of the portions of your construct of interest instead of hoping that the circular plasmid will break without damaging anything important.

It's highly unlikely that your construct integrated into the middle of something important for RNAi such as slicer or dicer.

Maybe try transiently transfecting the double-stable cell line with the siRNA to see if the RNA silencing mechanism is still intact in these cells?
posted by porpoise at 4:59 PM on January 4, 2007

As well as PCR to check siRNA expression, can you also purify DNA and submit for sequencing of the now presumably integrated) siRNA coding region and the target gene? siRNA knockdown is super sensitive to single nucleotide polymorphisms, so you should make sure that hasn't happened since AB started these experiments (although the possibility of that happening to all of AB's stable clones is very slim...)

Also, how long has it been since thawing and inducing the cells? If it's only been a short time, it's possible that complete knockdown at the protein level hasn't occurred yet (ie, leftover protein from before knockdown was induced), but that translation has been stopped effectively (measurable with RTPCR, as noted above)
posted by twoporedomain at 5:31 PM on January 4, 2007

Yes, all cell lines express GFP at pretty high levels but I didn't do this so I don't have any numbers (the tech is helping resolve this as well since it's also a lab problem, not just me).

The cells were induced approx. 2 passages after thawing.

I haven't done any PCR to test for knockdown but previous exp't done by AB showed lowered levels within 24 hrs with almost complete absence by 48-96 hrs detected by western blot. Our protein requires high levels of knockdown in order to get a phenotype so if it's not at least 90-95% then the cells are useless to us.

I have tried remaking the cell line but didn't linearize the vector. I will try that for the next attempt. I was under the impression that linearization didn't make a huge difference but I would like to be wrong. Anything as long as the cell line works out.

I still wonder why ALL of the lines would cease to express the siRNA even though they were frozen down.
posted by LunaticFringe at 6:01 PM on January 4, 2007

Thanks for the input so far and keep it coming!
posted by LunaticFringe at 6:01 PM on January 4, 2007

I just looked up the super-neo-gfp vector. So the siRNA is under the control of that H1 promoter? You mentioned that you're using dox to induce the siRNA. I also just read that in the TRex system, the dox serves to induce a chimeric CMVtetO promoter. I'm confused about how your setup is designed to work.

As an aside, I've learned to maintain some suspicion towards data that's been collected once, by one person. And dental school. Hmmmm. In my experience, the people who are good in the lab aren't usually the same people who are on a trajectory out of science. Just my experience. Nothing against dentists!
posted by shoos at 8:13 PM on January 4, 2007

shoos: I have had those very same concerns but my supervisor assures me that the pSuper system will work with TRex cells. In fact, I have brought this up twice. I may just have to bring it up again.
posted by LunaticFringe at 5:09 AM on January 5, 2007

The original pSUPER vector with the unmodified H1 promoter would NOT be tet regulated. There are 2nd generation pSUPERIOR vectors (OligoEngine, Inc.) that carry a modifed H1 promoter with a synthetic tetO inserted in the H1 promoter to allow tetR binding. Only the latter class of vectors will work in this system, otherwise the siRNA insert will be constitutively expressed.
posted by bigred1 at 8:47 AM on January 5, 2007

I'd verify that the siRNA construct has been integrated into the genome. Sequencing from purified genomic DNA, as twoporedomain suggested, would be the best way to do this. But perhaps a faster way would be to use the purified DNA as a substrate for a test PCR. Use a primer within the knockdown "coding" region and a primer on your heterologous promoter so you don't get endogenous background. Use DNA from cells without the siRNA vector for a negative control, and plasmid vector for a positive (PCR) control. Remember to use much more (1ug) genomic DNA than plasmid (1ng or much less) because of differences in copy number. I've successfully used this technique to check for integration of empty vector in control cells.

Then you can work your way out from the DNA: check for siRNA expression (RT-PCR), target RNA expression (qPCR), and target protein expression (Western). But start with something simple - ANY positive result will make you feel saner (it does for me, anyway).
posted by Jorus at 9:55 AM on January 5, 2007

Dear siRNA woes: I once had an identical situation as to what was described above. I traced it back to a a lack of clonal populations of cells. When generating these clones I isolated on the average 30 independent (or what I thought were independent) clones and tested them for doxycycline induction. I harvested 10 different clones from each of 3-10cm plates. They all were G418 and blasticidin resistant but the majority were not, or poorly, inducible by dox. Because I was working with a gene that, when knocked out, cell growth was suppressed, it was paramount that the cell population was clonal. Unfortunately, as it turned out, mine were not, so as they sat in dox the poorly inducible contaminate clones outgrew the others. I had to go back to my original frozen stocks, plate into 96 well plates (~1cell/plate), grow these up and select for ones properly induced by dox. It was a pain but worked. I hope this helps.
posted by djn at 7:07 AM on January 21, 2007