On Tue, 9 May 2000, Mark Schrenzel wrote:

> Does anybody have any suggestions or experiences with using 
> non-formaldehyde fixatives for better preservation of DNA and 
> RNA for pcr ...

In traditional histochemistry the best fixatives for DNA and RNA
contain alcohol and acetic acid. The simplest is Clarke's, which
has been around for nearly 150 years. Mix 1 volume of glacial
acetic acid with 3 volumes of absolute ethanol. Fix for as long as
it takes to penetrate the specimen (6 to 12 hours for a 5mm cube;
10 minutes or less for a layer of cells). Wash in alcohol, 
then clear, infiltrate with wax etc, as needed. Micro-anatomical 
preservation in paraffin is better than you get with any aqueous
formaldehyde. Also, nuclear chromatin patterns are more prominent,
but cytoplasm, as seen with oil-immersion, is rather coarsely
granular (coagulated), and organelles such as mitochondria and
alcohol-labile granules cannot be seen. Aggregates of rRNA (Nissl
bodies of neurons and "ergastoplasm" of secretory cells) are
stained crisply by cationic (basic) dyes such as toluidine blue
or neutral red.

Other major fixatives containing alcohol and acetic acid are
Carnoy's, methacarn (Carnoy with methanol), and the "modified Carnoy"
of James & Tas (1984). The last-named contains less acetic acid than
true Carnoy (5% instead of 10%) and is less prone to extract RNA if
the specimens are left in it for too long. There are also many
alcohol-acetic-formalin mixtures, all good fixatives for nucleic
acids, and the subject of recent HistoNet discussion, but you
won't want to know about these if you intend to steer clear of
formaldehyde.

The above remarks apply to traditional staining and histochemical
methods for DNA and RNA. It's the RNA that's more at risk; DNA is
well preserved, without chemical change, by nearly all fixatives.
In fact, it isn't fixed at all, but trapped among its large associated
basic protein molecules (nucleohistone), which are insolubilized by
fixation. I've never done in situ PCR, but would guess that any
fixative that can be washed out should be suitable. A substance that
binds itself strongly to the tissue and remains chemically active 
(such as mercuric chloride, chromium compounds or glutaraldehyde) is
generally contra-indicated if you are going to use enzymes as reagents. 
You wouldn't want your expensive enzymes to be "fixed" and inactivated
by the tissue before they can do their stuff.

Another of my too-long answers. It will be gratifying if someone who
has done a lot of in situ PCR sends in a one-liner that identifies
one or more acetic-alcohol mixtures as tried & true fixatives for this
family of methods!

 John A. Kiernan,
 Department of Anatomy & Cell Biology,
 The University of Western Ontario,
 LONDON,  Canada  N6A 5C1