Dear Shaung Liang and other buddies,
I know about a couple of examples of problems similar to what you
describe, both from other people's experience and from my own experiments.
You are right, the GAL promoter IS leaky. It depends on the gene
hooked to it, whether the basal transcription level is high enough
to disturb your plans; in your case it clearly is. This is probably
a general problem of any conditional system based on transcriptional
regulation since we encountered something similar also with the MET3
promoter which should be shut down in the presence of methionine.
As a rule, transcriptionally regulated conditional mutants are likely
to work only if the protein of interest is unstable or if the cell
really needs a lot of it.
So, what you can do is:
1. Try to destabilize your protein e.g. by making a ubiquitin fusion.
But I wouldn't do this, I would rather
2. Give up the GAL promoter and try to make a thermosensitive allele
by mutagenic PCR. It works quite fine - just reduce the concentration
of one of the nucleotides by a factor of five and add MnCl2 (final
conc. 0.5 mM) and mercaptoethanol (final conc. 10 mM) to the reaction.
IMPORTANT: use a native Taq polymerase, not AmpliTaq - that wouldn't
make any mutations.