Non specific yeast transformations (or none at all) - (Oct/07/2009 )

I seem to have hit a rough patch with my yeast transformations of late....Basically, I have been trying to do two types of transformations:

1) I am trying to tag EGFP to the C-terminal region of a gene in yeast cells.
I use a plasmid which contains EGFP and a selectable marker as template (pKT127, pKT128), and using appropriate forward and reverse primers (a 40 bp overhang for the gene of interest and 18-20 bp recommended region of overlap with the plasmid DNA); I do a PCR. I use the PCR product to transform yeast cells using the Lithium Acetate method, following the Gietz protocol pretty closely. (I have also tried running the PCR product on a gel, where it shows the correct size, then extracting the DNA from the appropriate band using gel extraction and using this as the DNA for transformation).

I obtain a few (1-5 or so) colonies per transformation in the G418 marker case, and none with the His selectable marker (I am puzzled by the latter). But invariably, none of the G418 strains show correct EGFP localization (I can see EGFP in the cells, so it is transformed, but it is not localized to the expected region for the specific gene of interest; which suggests non-specific integration). I have tried this several times.

2) I am trying to delete a gene in yeast cells, using say a URA3 selectable marker
I use a strain in which a gene was deleted with the URA3 marker, and using appropriate primers and the genomic DNA as the template, I do a PCR. Again, I checked the PCR product and it is of the right size. But when I use this product to do the transformation (same Gietz method), I again obtain a lot of spurious transformation results. I must note that I have obtained a successful transformation once or twice, but the frequency of this is very low (5-10%?) (when integration is successful, I am able to confirm it with a second PCR).

I am a little puzzled, because I used to have better success in the past (a couple of times when I had done it, starter's luck??). I was under the impression that the homologous recombination should be relatively robust (I would be happy with a 50% efficiency too, but it's frustrating to do this several times with no success, and only false positives...)

I have tried switching PCR enzymes and other materials, preparing new transformation mixtures etc... I also try (as much as I can given gene sequence constraints) not to have too many repetitive regions of A's, T's, G's or C's during the primer design

Any suggestions? It would be truly hepful!!!

I would suggest increasing the length of sequence homologous to target sequence on your targeting construct. That would increase targeting efficiency and specificity, 40bp is the very minimum you can use for gene targeting. Yeast tend to be unhappy under His selection, so expected some rough water.