At 23:58 24/03/02 -0500, vous avez écrit:
>Can you really get decent EM from frozen sections embedded in plastic? 
>What sort of cryoprotection do you do?

I used the procedure described by L.-I. Larsson in Vol.1 of Methods in 
Chemical Neuroanatomy (pp. 195-196.

After fixation, the tissues are cryoprotected in a series of sucrose 
solutions (10%, 12% and 15% for 2 hours and finally overnight in 20% 
sucrose + 5% glycerol + 40 microM digitonin). All sucrose solutions are 
prepared in 100 mM phosphate buffer pH 7.4. The digitonin is supposed to 
improve membrane structure, but I have never checked whether membrane 
structure is impaired if I leave the digitonin out.

It is essential for EM work to use tissue slices that do not exceed 1 cm in 
at least one dimension. Ideally, they should be about 5 mm thick. The 
cryoprotected slices are then frozen in isopentane cooled to -60°C with 
liquid nitrogen. The -60°C seem to be important, because if frozen at 
higher or lower temperatures, after sectioning the floating sections curl 
up, and it becomes impossible to unroll them to obtain even immunolabelling 
throughout the section.

For preembedding EM work, even on vibratome sections, you will have to use 
some permeabilization agent if you want your antibodies to pentrate more 
than less than 1 micrometer. Some people use detergents, while some even 
suggest freeze-thaw cycles after cryoprotection the sections in DMSO. You 
may also use buffered ethanol (10%, 15% and 20% in 100 mM Phosphate buffer 
pH adjusted to 7,4). Permeabilize the sections for 5-10 min in each of 
these (going up and down). The ultrastructure of membranes is affected 
though less than one might expect and antibodies penetrate almost througout 
the section. This technique works best when combined with an alcaline 
formaldehyde fixation.
You may look up a paper by Eldred et.al. (1983) in JHC 31(2):285-92.

Another trick to improve EM structure with preembedding work is to postfix 
the sections for 1 hour with GLUTARALDEHYDE after immunostaining and BEFORE 
peroxidase detection. Glutaraldehyde only slightly affects peroxidase 
activity. For peroxidase tract tracing, only glutaraldehyde works better 
than glutaraldehyde/formaldehyde fixatives. This postfixation with 
glutaraldehyde stabilizes ultrastructure, which otherwise will be impaired 
by the free radicals liberated during the peroxidase reaction. This can be 
seen on heavily labelled structures in preembedding EM ICC.

The combination of the above techniques allows the production of more than 
decent EM structure in floating frozen sections. Some of the 
permeabilization techniques are also usefull in vibratome sections. 
Especially the buffered ethanol procedure if you want to do only optical 
work on vibratome sections. I never found an antigen which would not work 
after this procedure it it worked with detergent permeabilization. And 
compared with detergents, this procedure is much more efficient and allow 
complete penetration of the immunoreagents throughout the sections (with 10 
min in each ethanol).

I hope this answers some of your questions.

Paul
                                                                         -=-
                                                                        (o -) O
===============================oOo==(_)==OOo=============================
Paul Klosen, PhD
CNRS UMR 7518 Neurobiologie des Fonctions Rythmiques et Saisonnieres
Universite Louis Pasteur  12, rue de l'Universite
F-67000 Strasbourg, FRANCE
Tel. 03.90.24.05.01  Fax. 03.90.24.05.28
========================klosen@neurochem.u-strasbg.fr=========================