eGFP vs mCherry CRISPR-tagged proteins

We have recently generated an endogenous reporter of an ubiquitously expressed protein using CRISPR/Cas9 to introduce either eGFP (from pJJR82 plasmid) or mCherry (from pJJR83 plasmid) in the same gene (C-terminal).

We observe very quick photobleaching of the mCherry signal in contrast to a more stable GFP signal. Moreover, we see accumulation of mCherry signal in the germline (and probably also in the intestine) and intranuclear mCherry foci that are not observed using the eGFP reporter for the same protein.

Has someone observed something like that? Any explanation for that?


Double check whether your eGFP is the monomer variant (A206K), if not, you might be getting a dimer with the GFP, but not with the mCherry, and the dimerized protein may have different stability and localization properties.

Here’s a blog page with information about GFP dimerization-

Per this report, mCherry is among the least likely FPs (of those tested) to not oligomerize. People often see it in puncta and assume it’s oligomerization, but in the gold standard assay, mCherry does not oligomerize…much.Even in that report they suggest there may be conditions where mCherry does aggregate, but it is unclear whether that is due to the protein being tagged, or the environment.

However, this is just one potential hypothesis for why you’re seeing distinctly different localization and stability. I would suggest mNeonGreen as a third choice, I understand that to be brighter than GFP and monomeric.

My experience is the opposite: GFP is nearly always better, non-punctate, and bright. The same construct as an mCherry fusion tends to photobleach rapidly and form punctate structures. I think some of the newer RFP fusions are better, TagRFP seems good and mScarlet, mRuby, mKate (the newest variants) might be even better. There are a few papers that do pair-wise comparisons on brightness, photo stability, etc. I think that mCherry loses every time.

My experience is a long time ago now, but I also found that mCherry was extremely prone to forming aggregates. mStrawberry seemed less so.

I would recommend checking if your tagged proteins are actually functional with a rescue assay (expressing a transgene in a null mutant, for example, and check if it rescues phenotypes). These tags may affect the folding or stability of the protein.

Furthermore, GFP and mcherry are often silenced in the germline due to endogenous small rnas. You can try to cross the strain with a mut-7 mutant and then back to wild type background and check for the accumulation of GFP in the germline. There are also point mutations in the gfp sequence that can prevent piRNA silencing doi: 10.1093/nar/gky277

I’ve found that using different linkers can affect brightness and stability of my CRISPR tagged proteins.

I would also like to add that I have had great success with mNeonGreen and in every case I’ve looked at so far it has been brighter than the over-expressed eGFP extra chromosomal arrays we used before.

Another anecdote: Pmyo-2::gfp and Pmyo-2::mCherry are tolerated much differently by animals. With GFP, it can get very bright before the animals are noticeably starved looking. With mCherry, this also happens at lower intensity (and lower injection concentration). We also see a lot of mCherry aggregates in those brighter pharynges. So that doesn’t prove that mCherry aggregation is the cause of the starved appearance, but it makes me nervous. We still use it, but are very careful about concentrations. I wouldn’t use mCherry for CRISPR.