Problems with purifying and concentrating a GST-fusion protein

[Lori Buetow]

Hi,
I am currently trying to purify a protein from a bacterial expression of a GST-fusion construct of the protein.  When my protein is eluted from a glutathione column, it co-elutes with a band of approximately 60 kDa in size -- I think the co-eluted protein is the chaperone GroEL.  My protein is a homolog of another protein with 97% sequence identity that has been expressed and purified from the same vector/bacterial system and the authors identified their co-eluted protein as GroEL.  

When my GST-fusion protein is eluted from a glutathione column, it has a 260/280 ratio of approximately 1.  Fractions with high concentrations of protein have high absorbance readings at 320nm but there is no visible evidence of precipitate in solution.  In addition, when I concentrate the GST-fusion protein, the 320 increases as though the protein is precipitating, but there is no visible evidence of precipitate in solution.  After cleaving the GST tag while dialyzing, the absorbance readings are comparable to pre-cleavage, suggesting that small molecules are not contributing factors to the strange absorbance readings.  Is this characteristic of a soluble aggregate or is a characteristic of a chaperone/protein interaction?  

I have tried the prep 3 times.  The first two times, I was able to obtain purified protein but the yields were too low for sufficient experimental work.  The third time I tried the prep I had better yields initially (GST-fusion protein) but concentration of the protein was very slow and dramatically decreased the yield.  For the third prep, I had optimized expression levels (about a 10-fold improvement over the 1st two) and I had used more pellet. 

Has anyone seen this type of behavior with a GST-fusion protein before and does anyone have any advice on how to increase the soluble/functioning yields from this prep?

Thanks,
Laurie