gfp reports higher expression but rt-pcr data do not show change


I have recently made a strain that contains a gene tagged with gfp in a mutant background. After confirming homozygosity by genotyping the worms seem to show higher gfp expression (confirmed by fluorescence measurement by confocal microscopy) than the wild-type strain. When I did rt-pcr with primers specific to that gene, there seems to be no difference in expression between wt and the mutant for that gene. In this case, should one trust the rt-pcr data more than the gfp fluorescence? Or try to do other experiments to test if two pieces of data will agree with one another?

Thank you for your insights.

One obvious possibility: when you say “a gene tagged with gfp”, do you mean that the product should be a translational fusion of the protein and GFP? Because if so, there’s one obvious way you could have different levels of GFP protein without having different levels of transgene mRNA …

The transgenic reporter strain used is a transcriptional fusion reporter, not a translational fusion…

Just to be clear. You have transcriptional reporter for gene X, and you’re comparing the expression of that transgene in wild type or mutant gene Y. Fluorescence says it’s brighter, but RT-PCR says the same, correct? Is this whole worm fluorescence, a particular cell or tissue?

Yes, that’s exactly it, Kevinem. It is whole worm fluorescence, observed at 10X and 20X magnification.

The hypodermal cells in particular are brightened up in the gene Y mutants compared to wt worms, which do not show that expression…

Hi, when you used the confocal microscopy to observe the expression pattern of your mutant strain and WT, did you use the exactly the same setting to do comparison? Also, is it same while you were measuring fluorescence?

Yes, the settings were all the same. I have also done the rt-pcr several times, and there is no difference in gene expression between the wt and the mutant. I have also done the cross a second time to remove the possibility of a mutation happening during creation of the strain, and this time I got the same difference in expression pattern and fluorescence…

It may be because the increased expression is a tissue-specific phenomenon (mostly in the hypodermis from what you said), which is detectable by GFP but gets diluted out when you isolate RNA from the whole animal. When you observe this increase, is it at all life stages? And when you isolate RNA, are you doing it in a mixed population? If so, you’d also miss it.

Perhaps if you could clarify all of your experimental parameters and observations it would make it easier to help troubleshoot.


Hi Steve, the increase is seen only in L4 and adult stage worms. For RNA isolation, I wash off plates of a mixed population of worms with M9, so you probably have a point. I guess I will manually pick a fixed number of gravid worms and then do the qpcr. And if the same thing happens even then I guess I will probably have to do other experiments to confirm and to explain how this might take place in the cell.

Thank you for your helpful tips!

Is it RT-PCR of the GFP transgene or of the endogenous gene to which the promoter fusion was generated? It seems like you’d want data for the endogenous gene. But yeah, since it is an age/tissue-specific result, it might be too weak to detect. You could use arrested L1’s to synchronize and do it only on L3’s vs. L4’s/adults.

You could also try a hypodermal-specific rescue of the gene to see if the GFP goes away. You’d want another promoter to make sure you don’t get sequelching, and co-express RFP from the same promoter. That way you could do some sort of mosaic analysis. Red(-)cells should be GFP(+), Red(+) cell should be GFP(-).


I am slightly confused by your question. You said you made a transcriptional reporter for that gene , that means you just have the promoter fusion to GFP right? Then, how do you expect to see any change in the gene transcript level by rt-pcr when your construct does not have the full gene?
Please correct me if I am wrong in understanding this.

Okay, I will try to clarify. I have transcriptional reporter for gene X, and I’m comparing the expression of that transgene in wild type or a mutant strain which has a mutation in gene Y. To do this, I first made the strain gene X::gfp;gene Y(-). I compared gfp fluorescence of these worms to that of the wt gfp strain, gene X::gfp. gfp reports much greater expression of gene X in the gene Y(-) background, but that difference, observed in the whole worm, is seen only in L4/adult worms.

Next I wanted to see mRNA transcript levels of gene X in the wt and in the gene Y mutants. So I do not use the gene X::gfp;gene Y mutants for RT-PCR, but N2 and gene Y mutants (and not the gfp strain). But RT-PCR says the expression is the same.

Hi senor_abditis,

There might be miRNA regulation involved. In your transgene I guess you use the unc-54 3’UTR and not the native one.


or gene y could be involved in transgene silencing at the posttranscriptional level.

Some thoughts:

  1. You take your images on a confocal microscope, but was that necessary to see the difference? If the GFP is very faintly expressed in the hypodermis and you need to use a confocal to see the difference, I wouldn’t be surprised that you don’t see a difference in qRT-PCR. As it was stated before, the qRT-PCR isn’t sensitive enough. (GFP is correct, qRT-PCR is a false negative)

  2. You are looking at protein levels (via GFP fluorescence) and comparing that to protein message levels. Protein levels do not always correlate with mRNA levels. So you are somewhat comparing apples to oranges. As others have suggested (microRNAs and 3’UTR choice), post-transcriptional regulation of expression may be responsible for differences. (Both GFP and qRT-PCR is correct)

  3. GFP is very stable and you are over-expressing gene X, which makes it possible that you have an over-expression artifact. (qRT-PCR is correct and GFP is a false positive).

It will take some effort to decipher what is correct :-
Good luck!