I have had some moderate success in staining GFP containing tissue sections 
with FITC labelled antibodies in order to 'replace' fluoresence lost due to 
parrafin processing, so it is possible.  Since the part of the GFP protein 
that actually glows is fairly well protected in the protein tertiary 
structure, you may find that you can still see some fluorescence if you use 
gentle fixatives (fresh paraformaldehyde), and short fixation times prior to 
processing.  You'll lose weak GFP expression, but strong expression will still 
remain.

  The risky part with antibodies is that the epitopes you need may not survive 
the processing - using formaldehyde as a fixitive helps a lot (it doesn't 
denature the proteins, and apparently helps shield them from denaturing due to 
later alcohol exposure) - try an epitope recovery protocol (e.g., boil your 
slides in citrate buffer) before adding antibody.  Some people suggest using a 
polyclonal antibody so that protein recognition isn't hinging on a single 
epitope, but personally I've had problems with non-specific binding, so I take 
my chances with a monoclonal.

 -Tom Clarke-
  Division of Biology
  Kansas State University
  tec8435@ksu.edu

>At 2:30 PM -0400 15/5/02, Sarka Lhotak wrote:
>>Hello Netters,
>>	Has anybody done a peroxidase staining against GFP in paraffin
>>tissue sections? I am wondering whether the antibody can bind even to GFP
>>that is no longer fluorescent (due to alcohol processing etc.)?
>>	Thanks a lot,
>>
>>Sarka Lhotak
>>
>>Hamilton Regional Cancer Centre
>>McMaster University
>>Hamilton, Ontario, Canada