I'd guess you are seeing cellular autofluorescence due to flavins, cytochromes, and so forth.

At first glance, it looks like fluorescence bleedthrough to me, but it definitely shouldn't happen between mCherry and GFP, unless the mCherry signal is insanely strong (which doesn't seem to be the case here).

What is weird is that I can see some cells that are green, but not (or almost not) red. If it is a viral injection, could it be that you had a contamination in your syringe/pipette, if you re-use it? For example, with remnants of another virus expressing GFP? Or a contamination in your virus aliquot?

I can suggest a quick test: incubate your slices with (up to) 3% H2O2 diluted in PBS for 30 min. It should quench the endogenous mCherry fluorescence. Then check under the microscope if the green signal is still present. If it is, then it means that it's definitely not a fluorescence bleedthrough.

Well flavins can be excited by 488nm and emit at 550nm. So that could be one answer.

Adam B Shapiro, thank you for your suggestion. I searched a bit for some examples of autofluorescence. It seems to be the case. One of the first articles I found (10.1016/j.jneumeth.2011.01.029). The images look really similar to mine

Vincent Simon , thank you for your answer. I do not think that contamination is the reason. I used a new pipet every time and only one type of virus (in this particular experiment). Aliquotation was performed with great care as well. So I believe and hope it is not the case. You also mentioned that strong mCherry signals may lead to fluorescent bleedthrough. Does it mean that, close to the injection site, mCherry-positive neurons will very likely have a strong green fluorescence? I think I see it near the injection site. Thank you for the suggestion; I will check it.

David M Jameson , thank you for the answer. After a little searching, I think flavins are really the reason in my case.

The green signal outside the mCherry-expressing cells is most probably from flavins. Its intensity looks a bit higher than I would expect to see considering similar power densities for all fluorescence channels. But we do not know exactly what laser sources were used. If the green laser was an Andor HLE (2 W), then its power density at "50 ms/20%" would be more than 10-fold higher than that from an ILE 637 nm laser.

Concerning the green signal co-localized with the red one from mCherry, I am pretty unsure that it could be attributed to flavins. I'd rather suppose that you observed fluorescence of the mCherry proteins with incomplete chromophore maturation. It is a widespread phenomenon among RFPs (see, for instance, here 10.1007/s43630-021-00060-8) to show some minor green and/or blue populations. Moreover, fixation procedures can likely chemically modify the red chromophore and thus lead to appearance of green fluorescence.

Alexey M Bogdanov ,thank you for your reply. Thank you for the notice about the laser model. I will check what model we use exactly and read the specifications.

Photoconversion is a new phenomenon for me. I have never heard about it. Thank you for the reference. Do you have any ideas on how I can confirm the mChery photoconversion?

What I meant by "fluorescence bleedthrough" is that if your mCherry signal is extremely strong, and if it's the only source of fluorescence in the tissue (except from autofluorescence due to endogenous molecules or formaldehyde), then you might detect this fluorescence using filters that are normally not adapted to mCherry. Even suboptimal excitation wavelength and detection window may be able to detect it.

What I find surprising in your picture is that this green fluorescence is present only in the cells that have clearly been infected by the virus (or around the injection site at least). If it were flavins, I would expect other cells in the brain to have this kind of fluorescence, but when you look away from the injection site, the tissue seems pretty clean. Did you look at other parts of the brain to see if you have a similar staining pattern in unrelated regions? One explanation would be that the infected cells have an increased amount of flavins, but I don't know if it's the case. I don't remember having such problem when using mCherry-expressing viruses...

Related to what Alexey M Bogdanov suggested, about the chemical modification of red chromophores due to fixation (which I was unaware of): what fixation protocol did you use?

Regarding the "fluorescence bleedthrough" suggestion proposed by Vincent Simon , it is of a great relevance if the laser power used for excitation in the green channel was much higher than that in the red channel. According to https://www.fpbase.org/protein/mcherry/, 488 nm source can still excite around 8% of mCherry molecules but only a negligible fraction (well under 5%) of emitted photons could pass through the 512-555 nm band-pass filter. Therefore, one would need to use a 200-300-fold power advantage in the green channel to see a well-pronounced bleedtrough (of course, if I did not miss any additional details when calculating).

Are your tissues fixed? Do you use standard type fixatives? If so, aldehyde fixatives are a good source of background autoflorescence. I switched to rhodamine type fluorochromes and never again worried about endogenous fluorescence. Also, tissues have a certain degree of autoflorescence, so depending on your tissue type and fixative you can get a persistent low yellow-green background glow.

Thank you for the ideas, Vincent Simon! I've got it. According to the spectrum, I think mCherry can unlikely bleed through the green filter. Unless its real spectrum is different from the theoretical.

The very weak fluorescence can be observed in other regions far from the injection site. I think it could be flavins. Their Ex/Em spectrums are fit.

Sheldon R Gordon , that's right, the tissue is fixed with 4% formaldehyde. It is the standard protocol that is used in my lab. I hope the real signals from green fluorophores will be much higher than those from autofluorescent neurons. Thank you for your thoughts!