Problems with chemiluminescence

We have a problem that is driving us up the wall. Developing
Westerns with chemiluminescence (including Amersham, Pierce and NEN
products) we are frequently getting no signal at all. This happens
regardless of the origin of the sample, transfer conditions,
nitrocellulose or PVDF, primary or secondary antibody (or using
biotin-avidin amplification), or changing the exposure conditions.

We've changed the water, remade all the solutions, tried
different blocking solutions, an even not used block at all. Some,
but not all, of the primaries have azide, but the secondaries don't.

The same antibody, on the same samples will give a whopping
signal one day (film exposure less than 10 seconds for optimal
signal), and the next time an overnight exposure won't even give
background on the membrane. We always have a phosphorescent dot on
the sandwich, so we know it's not a film developing problem.

Now I've talked to a couple of colleagues in other universities
who say that they have been experiencing the same sort of thing. Is
anyone else out there seeing something similar? Anybody identified a
problem, or have ideas?

Any help appreciated.

Agonizing in Argentina
---

"Sean Patterson" <[Only registered users see links. ].edu.ar> wrote in message
news:a05010404bb6fd6dfd070@[10.0.10.84]...

I almost sure it is not, but could that be that your signal is too strong
and you are depoeting the reagents by the time you expose the film? That
would be obvious by the appearance of staining on the film in the place
where you rexpected band would be. I am asking this, because you wrote that
in successful experiments you would have whopping signal, whereas usually a
successful experiment is when you get a signal for ~30-60-sec exposure over
normal sensitivity film (like Kodak X-OMAT AR).

Also, always follow the same guidelines for protein loading. Your band of
interest should contain somewhere between 20-200 ng of protein.The only time
I have problems with Westerns is when I am lazy to do protein content
estimations (assuming antibodies are good and working OK).

Emir