Immuno(histo)fluorescence: what's the best method?

Tommy the Terrorist u58563 at uic.edu
Thu Mar 7 22:08:10 EST 1996


I'd like to do some immunofluorescence on slides of skin and intestine to
visualize low-abundance proteins with publishable results.  There are
lots of new products in the catalogs, but I'm not sure what the best
method is.  Protocols are out of date, companies are obviously biased,
and worse, I get the feeling that due to patents they are effectively
censored from telling us what *could* be done with the products they sell
("not for that use").  So far, I've considered the following:

1) Molecular Probes sells a new "ELF" technique, the significant feature
of which is that it uses alkaline phosphatase to convert many molecules
of a fluorogenic compound into the fluorescent signal molecule.  The
problem?  We're worried about endogenous alkaline phosphatase, especially
in intestine (that won't be affected by levamisole).  On the other hand,
I've seen sources that say not to worry so much about it in
normally-fixed tissues.  Then there's the other question:  Does It
Work...?

2) Vector has a nice range of avidin-biotin based methods.  Stain with
primary then biotinylated secondary antibody, then with labeled avidin;
this allows their background-blocking kit (first avidin then biotin, then
wash) to be used.  Or, stain with primary, biotinylated secondary, then
avidin, then biotinylated anti-avidin antibody, then more labeled avidin.

3) Alternatively, how about multiple enzyme-linked substrates with ABC
kits, in brightfield?  Apparently you can get away with this, but I'm not
really sure how.  And Pierce has this nifty DAB substrate-enhancing thing
that they show a pretty impressive improvement with in their catalog... 
it's enough to make a person consider leaving out the "fluorescence"
part, although fluorescence *must* give the best localization (right?)

Then there's the question of multiple labeling:  which of these methods
would be most compatible with what other souped-up visualization
technique?

I don't expect a reply to answer all of these questions; if anyone can
just say "I tried this and it works" or "I tried this and it *didn't*
work" it'll be a good beginning!



More information about the Methods mailing list