Can I trust GFP-cotranfection - transfection control (May/20/2008 )
I have done experiments where I co-transfected one plasmid with a GFP-tag and another with Myc-tag. By immunofluorescent analysis I discovered that only 50% of the myc positive cells were also expressing GFP!! Besides issues with vector size differences already discussed, the individual preps of DNA can have dramatic effects on efficiency. If one prep is cleaner/better, this will be picked up by the cells much better than the other. I would highly suggest you go ahead and use a tag on your sequence of interest, if there is not an appropriate antibody for your gene product. This way you can easily identify exactly which cells are expressing your gene. Since your insert is so big I would go with Myc. It's small and there are lots of good antibodies commercially available. Otherwise you are going to have to show first that the cells are indeed expressing both GFP and your GOI. Only after you establish a transfection protocol that gives you the best co-expression conditions can you reliably use GFP expression as a marker for expression of the other plasmid. Be warned though, reprep the DNA and you have to retest the transfection conditions.
have you considered the factor that it could be because the methods of detection have different sensitivity limit?
I do agree that using direct immunostain method to prove this is indeed the case will provide most concret answer, however, if someone wants to do some experiment with live cells, that will be a little hard to do. a cotransfection with EGFP as a guide is a reasonable approach.
I would suggest thinking about what controls you will have to do in either case. For some imaging applications, you may not want the GFP concentration to be too high (could lead to phototoxicity). If you put the GFP and the gene in the same plasmid, then you can't independently vary the concentration of either. At least, with co-transfection you can maintain a consistent GFP concentration and vary the gene of interest independently. I would suggest co-transfecting a GFP-only plasmid and another plasmid that has a FLAG tag or some other reporter.
In case your transfection efficiency would be 10%, then you would get: 0,1*0,1=0,01 or only 1% of your total cells would contain both, which means only 10% (1%/10%) of your EGFP+ would also contain the other.
I think it depends on the stochasticity of the plasmid entry. For example, imagine a case where 1-10 plasmids enter the cell. That is vastly different than if 1000s of plasmids enter the cell. In the former case, one can see that there is a possibility of cells taking up one plasmid but not the other. In the case where hundreds or thousands of plasmids enter the cell (say during electroporation), it's unlikely that one cell would be transfected with one "species" of plasmid but not the other. Does anyone here have an idea about the absolute number of plasmids that typically transfected? I assume this depends on cell type, transfection method (electroporation vs. lipofectamine vs. virus, etc), plasmid size...
Hmmm...good one! I really hadn't given much thought to the detection sensitivity since both my GFP and Myc antibodies are wildly sensitive and both are primary conjugated but this certainly could explain my observations. For live cell imaging, I think the best route is constructing a stable cell line. However, as I mentioned, if you can assay and show that cells expressing the GFP are also expressing the GOI then in future experiments you can depend on the GFP as a marker for expression. I do think it is dangerous to just co-transfect and assume co-expression without checking first. Also, for live cell imaging, check out a relatively new technology by Invitrogen, FlAsH and ReAsH. It's a small molecule/fluorescent detection of proteins with a four cysteine tag. A bit pricey but I've heard amazing things and it gives you great opportunities to study expressed proteins in live cells over time.
In case your transfection efficiency would be 10%, then you would get: 0,1*0,1=0,01 or only 1% of your total cells would contain both, which means only 10% (1%/10%) of your EGFP+ would also contain the other.
I think it depends on the stochasticity of the plasmid entry. For example, imagine a case where 1-10 plasmids enter the cell. That is vastly different than if 1000s of plasmids enter the cell. In the former case, one can see that there is a possibility of cells taking up one plasmid but not the other. In the case where hundreds or thousands of plasmids enter the cell (say during electroporation), it's unlikely that one cell would be transfected with one "species" of plasmid but not the other. Does anyone here have an idea about the absolute number of plasmids that typically transfected? I assume this depends on cell type, transfection method (electroporation vs. lipofectamine vs. virus, etc), plasmid size...
I hadn't given much thought on how much plasmids would enter each individual cell, but just out of curiosity: wouldn't this also to some extent correlate to transfection efficiency? If your cells are hard to transfect, say your efficiency is 5%, wouldn't it be that the copy number per cell is likely lower than when you have 90% transfection efficiency? (as in the last case your cells seems to be more prone to take up at least one, so likely prone to take up a lot more).
Obviously, the definition of "efficiency" is very important. What do you mean when you say "efficiency is 5%" or "90% transfection efficiency"? Does that mean 5% or 90% of cells are transfected with some plasmid (i.e. at least one plasmid)? Or does it mean 5% (90%) of cells are transfected with at least some number of plasmids? For example, suppose you are transfecting cells with GFP. How many GFP plasmids does it take to actually light up a cell? 1? 10? 10,000? I don't know the answer, but clearly, it is possible that even dim cells take up some plasmid, just not enough to express strongly. Does that make sense? I'm an engineer, so I tend to think about these things very rigorously. To some extent, "efficiency" depends entirely on your readout for plasmid uptake. If you're using GFP to gauge efficiency, you may need to use a different definition than if you measure efficiency by some other means (i.e. cell phenotype, Western, immunoprecipiation, etc). Put in another way, does "efficiency" refer to the cell's ability to uptake plasmid, or does it refer to the ability of a particular plasmid to elicit a response in a particular cell type? I would argue both are important to consider.
Agreed. I checked into this recently. Very interesting potential uses.
Using GFP-expression as a marker, I would define efficiency as "the percentage off cells that yield detectable GFP-expression", no matter how much plasmids would be required for that.
Depending on your cell type, one plasmid can give you enough expression to be detected (that I know out of viral dilution series on cell lines and determination of EGFP-expression 24 hours later), the virus inocculum can be quantified and viral dilutions results (once in the linear range) in a nice 1:1 correlation of EGFP-expressing cells versus input (meaning that if dilution 1 would give you 10% EGFP-expressing cells, then a 5-fold dilution would yield 2% EGFP-expressing cells, beyond a certain point the linear relationship no longer occurs, most likely due to more dual/triplle infected cells).
Then again, this will also be dependant on your promotor used (how active is it in your specific cell type?), and if your gfp is strongly expressed in that cell line (not all variants are equally well expressed and some variants from different species, ZsGreen or so, are better than gfp in one cell, but worse in a different cell).
Using "percentage of cells yielding detectable expression of a protein", you could probably check for co-transfection by using 2 different fluorescent proteins and use flow cytometry or so, but it would still not be the same as the original experiment with different plasmids/different preps...
Another question is whether the plasmid, once inside the cell, has the same chance of uptake to the nucleus. From my understanding, the plasmid contains sequences like SV40 to promote nuclear uptake (which is apparently an active process). Am I correct about this?