Have you looked into the ScaleA2 method? It has worked well for me in the past, and can be achieved with common (and inexpensive) lab chemicals :

http://www.nature.com/neuro/journal/v14/n11/fig_tab/nn.2928_F1.html

I'd say that in our tests, Clarity is the best for preserving endogenous fluorescence of GFP or Tomato. Cubic works on slices, but the clearing is less potent than Clarity, and the GFP preservation a bit less good. If you don't see the endogenous GFP signal with Clarity, you might not see it with any other method.

However, I'd strongly recommend immunostaining for many reasons. Unless you're using a super strong promoter, GFP expression often needs a boost. Also, immunolabeling allows you to use far-red fluorophores that yields better optical properties for volume imaging. (thanks Andreas for citing the iDISCO method!).

Thanks everyone for the feedback so far.  For Immunostaining, should that be done before clearance, as in this recent 3DISCO paper from Belle and Chedotal Nov 2014?

I have heard that doing immuno after clearance, especially with CLARITY, is difficult b/c the structure of the brain is very hard to penetrate.

http://www.sciencedirect.com/science/article/pii/S2211124714009061

Yes, that would be the idea. The paper from Belle and Chedotal and ours are based on the same principle: immunostain and then clear with 3DISCO. The iDISCO protocol is optimised for adult brains (http://www.ncbi.nlm.nih.gov/pubmed/25417164), and enables full antibody penetration in adult tissues.

Thank you Nicolas

Dear Nicolas - when clearing endogenous GFP samples with CLARITY, what do you use for the last step? FocusClear or something else? I'm asking because we had some problems with fluorescence loss after passive CLARITY lately.

Dear Monika, we use RIMS which does a decent job, but this is just for testing purposes, as we don't use Clarity in our lab for "real" experiments. In some specific cases (viral tracings for instance), immunostaining can recover the lost GFP.

Thank you for quick answers! Perhaps we should try RIMS too. Another person in my lab does immunostaining of slices after CUBIC, but we aim at thicker samples that are probably too thick for immuno.

Mario: we've used glycerol already, but not the way you describe it, thanks for the link!

I just ran across your question from last year and that hope that your experiments went well. In case you are looking to clear a mouse brain in the future, I have a procedure that uses ethanol and BABB. It preserves emission from fluorescent proteins. I have also used it the look at unlabeled cleared mouse legs using second harmonic generation. 

The procedure was published a few years ago in Journal of Visualized Experiments (JoVE):

SG Parra, SS Vesuna, TA Murray, MJ Levene (2012) Multiphoton microscopy of cleared mouse brain expressing YFP. J. Vis. Exp., pub. 9/24/2012, e3848 10.3791/3848, DOI: 10.3791/3848. Available online at: http://www.jove.com/video/3848/multiphoton-microscopy-of-cleared-mouse-brain-expressing-yfp.