2 ideas I can think of in this instance.

1) concentrator membrane chemistry - there are concentrators which use Cellulose-based membranes, some use PES, and I think (from memory) some use PTFE.

Which membrane are you using?  If your protein binds to one type, may it be compatible with a different type?

2) Can you try using another additive such as Arginine in the buffer to reduce unwanted binding, then dialyse this away once the protein is sufficiently concentrated?

Good luck!

The two most common centrifugal concentrating devices are made of

(1) "regenerated cellulose" and

(2) "polyether sulfone (PES).  

The concentraters made of regenerated cellulose are known to interact with many carbohydrate binding proteins as regenerated cellulose has glucose monomers as building block.

The PES is known to interact less with proteins.

If your protein is known to bind to carbohydrates you might want to try PES concentraters.

There are other methods of concentrating that does not involve centrifugation which might be an option as well.

Changing salt Conc. or some  other buffer components might help. 

Good Luck.

Venu.

If your protein can tolerate a little bit of surfactant, I would put in something benign like 0.1% NP-40 or Tween-20. For any other purpose besides crystallization, BSA would ideally reduce non-specific binding of extra sticky proteins. Do you follow the manufacturer's protocol to wet the membrane with the buffer your protein is in before concentrating?

When you describe "adhering to the membrane" is that a conclusion based on loss of total product and/or loss of flux?  Is it certain that the proteins are not aggregating in the polarized concentrate layer during the concentration?  At which stage of purification are you concentrating?  Usually the downstream steps in a purification train will employ PES membrane chemistry for concentration/buffer exchange due to the lack of membranous impurities while upstream UF/DF unit operations will utilize cellulose-based membranes because of less fouling by hydrophobic solutes.  These operations are done in a TFF mode which can also be done on a very small scale so that the membrane doesn't collect or build up protein.   What pH range can the proteins tolerate before re-adjustment and is there any irreversible inactivation by mild chaotrope concentrations (urea, dimethyl urea, crystallography solvents, etc.) that can be exchanged out later with little volume change?  

 "adhering to the membrane" is based on the loss of 50% total protein. I also think this could be low solubility of this protein. This is a GST fusion protein and concentrating is done after the cleavage from the beads. This protein is a RNA helicase has the ability to bind Mg2+ and the pI is around 6.8. The GST fusion purification protocol I used was an old standard protocol and the lysis buffer was just PBS with DTT which has a pH of 7.4. However, in other labs they use different lysis buffer such Tris/HEPES with high concentration of NaCl and reducing agent. do you think I should come up with a new buffer instead of using PBS? 

I think a few things I could try: 

1. Add some Mg2+ before culture induction, since it is a Mg2+ binding protein it may help stabilize the protein. 

2. Lowering the induction temperature and extending the over expression time.

3. Change buffer, find a new buffer that could make the protein more stable. 

4. Add charged Arginine amino acid in the buffer  to make the protein more soluble and stable.

5. Try different concentrator. 

Thank you for all the help you guys provided, and I really appreciate it :) 

Is this a bacterial helicase or mammalian DEAD-box?  What ever is left in solution, is it active or needs to be active?  A 60 kDa loss left from a 90kDa enzyme  seems quite significant If you're expecting the 30kDa fragment to correctly fold.  If cysteines are still present in the mutant sequences then do they require correct pairing to oxidize and form disulfides for correct folding?  If this is a bacterial helicase then the cysteines may need to be kept reduced.  Solubility is certainly dependent upon correct folding, especially at high concentrations.  If Mg is important then it should have been in the previous protocols, right?  Does the Mg associate with the protein in order to bind ATP but not necessarily for folding?  Is the DTT present to keep the system reduced because of the reduced glutathione elution but the helicase needs to disulfide bridges?

Jindou Dong ,

Hi, Could you give me the update how you have solved your problem? I am facing same problem, I am losing proteins during concentration. I am using cellulose membrane concentrator. Have you tried PES one that others are suggesting here? does it work ?