It is possible to image GFP and autofluorescence does not need to need a limiting factor, particularly if you use time-domain optical imaging (McCormack E et al. Mol. Imaging 2007) or spectral unmixing whereby autofluorescence can be gated out. The major limitation with GFP is due to absorption of both excitation and emission spectra by haemoglobin or deoxyhaemoglobin. Thus, one should choose a reporter with ideally excitation and emission in the NIR. The main flaw with these far-red fluorescent proteins is that they generally are not very bright in comparison to GFP for in vivo imaging. But of course it depends on what your going to be looking at in vivo. If we are talking about superficial organs you might be able to use a fluorescent protein with high enough gene expression. If you want to go below 2-3 mm I would use a luciferase or nitroreductase reporter.

For in vivo imagery using CCD camera, it is better to use lumniscence as compared to fluorescence. Among fluorescent reporters mCherry is superior to GFP. GFP frequencey has issues due to body's autofluorescence. I had once tested plasmids expressing GFP, mCherry, and firefly luciferase. Firefly luciferase is best followed by mCherry. You will get a lot of background fluorescence when using GFP, due autofluorescence.

@ Dear Zaheer : Thanks for your information.. so then i need to modify my experiments...

Depends on your experimental design, and the optical imaging equipment (any of the IVIS?), but the golden rule for in vivo (animal imaging) is - never use GFP.

Yes you can! if you use balb/c mice, there will be lesser auto fluorescence compared to Black 6 mice. If your animal choice itself different better you try with FAR RED FPs

I have tested in both Balb/C and nude mice with IVIS machine. GFP is simply no good. I had injected plasmids into muscle. If you are working on a tumor or cancer model and your cells are located deeper, Luciferase may be a better option. However, as substrate has to be injected I/P, it may add to some variability which will have to be controlled.

Muscular injection need you focus in one muscel.

It is possible to image GFP and autofluorescence does not need to need a limiting factor, particularly if you use time-domain optical imaging (McCormack E et al. Mol. Imaging 2007) or spectral unmixing whereby autofluorescence can be gated out. The major limitation with GFP is due to absorption of both excitation and emission spectra by haemoglobin or deoxyhaemoglobin. Thus, one should choose a reporter with ideally excitation and emission in the NIR. The main flaw with these far-red fluorescent proteins is that they generally are not very bright in comparison to GFP for in vivo imaging. But of course it depends on what your going to be looking at in vivo. If we are talking about superficial organs you might be able to use a fluorescent protein with high enough gene expression. If you want to go below 2-3 mm I would use a luciferase or nitroreductase reporter.

If you wish to continue with your current protocol, I would suggest investigating RFP or the newer IFP - both are generally better for in vivo use than GFP. However, depending on the depth of your target, it is possible to devise an effective spectral unmixing protocol that will minimize autofluorescent effects...but not eliminate them completely. Look at the attached image - you can see the issues you have to deal with when using GFP...

One additional to Emmet McCormack's point about the NIR dyes not being as bright - oftentimes this is as much a factor of the detector than it is of the dye itself, as many CCDs have a dramatic dropoff in Quantum Efficiency starting around 650nm...so it is also important to match your dye with a) your camera's peak efficiency, and b) an appropriate filter ex/em set.

Dear Scott,

Please read the post again, I never said NIR dyes are not as bright - I said far red fluorescent proteins are not as bright. I have tried this both with your system (CCD based) and the Optix, and while the PMT based Optix is better for NIR fluorescence in comparison to CCD, theres no disputing the fact that I can achieve higher expression of GFP in my cells than RFP with our vectors so as to affect my imaging results. (Please note that Im using "i" and "my"...so Im not generalizing for anyone other than myself). However, our group have used the Nitroreductase NIR imaging platform (using a quenched NIR dye) and find this to be superior to fluorescent protein and luciferase imaging in our models of disseminated and metastatic cancer (McCormack et al Cancer Res 2013).

Ah - understood, and agreed. And your nitroreductase platform sounds quite interesting...I just downloaded the relevant papers from your page.

Have you been able to try IFP yet? I'd be interested in knowing more...

No we havent tried it yet. We do lots of Heme and metastatic models so I need something that works in these models and Ive only seen good SC imaging with IFP so far really. Its great that its NIR but probably needs a few more rounds of mutation to be optimized for in vivo imaging of mets.