Try refolding at different pH (say 5, 6, 7, 8 and 9) to see if that helps. Often that can make a huge difference in yield.  (remember arginine is a buffer so you'll want to pH the final solutions before refolding) You can also try  slowing your dialysis (going down to 0 M denaturant in 0.5 M steps, for example, at 24 hours each step at 4 C or 25 C). Since your protein precipitates when the his tag is removed, can you simply leave the tag attached?  You also might want to vary your final conditions a bit as well.  Try 0 M and 400 mM KCl during your tag cleavage and also some glycerol (say 20 % or removing the glycerol)

My personal experience with refolding is that each kind of unnatural tag, especially the his-tag is making refolding tasks even more complicated. Sometimes it depends on the location (N- or C-terminal) but generally I would try to work without his-tags in refolding procedures and I would express the target protein in its native form without any tag. You might ask why? There are many reasons, I will pick only the most important ones.

- Your protein is expressed into inclusion bodies. Isolation of these inclusion bodies is a very efficient initial purification step, which is often as powerful as his-tag based IMAC, so you don't gain additional purification advantages with a his-tag.

- The his-tag is a nonphysiologic charge accumulation at one side of your protein, which can have tremendously negative effects on the natural folding behavior of your protein.

- The strong charge of a his-tag can attract other proteins to become part of the inclusion bodies, thus weakening the purifying effect of inclusion body isolation.

- You are cleaving the his-tag, so you introduce an additional experimental step, you contaminate your liquid chromatography columns and pumps with a protease (some of your colleagues might not be that happy about this) and, as you are cleaving the tag, there seems to be no special need for it. Why are you still using it?

I would recommend the following:

Express your target protein without any tag. Then isolate your inclusion bodies and perform a screening in small volumes for both, optimal resolubilizing conditions and optimal refolding conditions. Follow Mathews hints regarding pH, temperature and ionic strength screening and add some additional compounds to your refolding and resolubilization solutions.

What is the native localization of your protein? The background of this question is difference of cytosolic, extracellular etc. conditions. E.g. for the reconstitution of disulphide bonds it is often required to add reducing agents such as DTT, TCEP or beta-mercaptoethanol. A defined redox environment can be set up using appropriate amounts of reduced and oxidized glutathione to facilitate the formation of the correct disulphide bonds.

Adding appropriate amounts of urea can often be helpful, as it supports the formation of correct secondary structures.

Sometimes it is also important to add some detergent like tween or triton-x to resolubilize the hydrophobic parts of your inclusion bodies.

Slow changes of chemical conditions can be achieved via dialysis and can help your protein to walk along the energetic road of folding enthalpies to the correctly folded energy minimum. This can be supported sometimes by thermal gradients from higher to lower temperatures. Keep in mind that your protein is optimized for a certain temperature, so it might be possible that a human protein from a 37C body temperature cannot be refolded efficiently at 4C.

Directly after refolding I would perform ion exchange chromatography, e.g. on a Q-sepharose, as misfolded proteins often have many hydrophobic residues on their surface, so that only the correctly folded molecules will interact with the stationary phase. If you are working with large volumes, you can also directly put the sepharose beads into your refolding solution and after some mixing directly pack the column with the preloaded material.

Hope that helps....

Konrad

I actually forgot to mention another important aspect. In order to find out the optimal refolding protocol for your target protein you need to monitor the refolding progress. If your protein is an enzyme than you are lucky, you can quantify the correctly folded protein with an enzymatic assay (by the way, adding cofactors to refolding buffers of enzymes is often required for correct folding). If you don't have an enzyme, monitoring is a little harder and typically achieved by spectroscopic methods, such as circular dichroism or IR-spectra measurements. For those spectroscopic methods it is important to have a good positive control of your correctly folded protein and a negative control as the initially misfolded protein. Collecting spectroscopic data from positive control, negative control and your refolding screening batches will give you the chance to estimate the degree and success of the respective refolding condition.

I agree with the suggestions above...they are very good. And if you don't do ion exchange first,  you should be careful to concentrate your protein as little as possible before SEC. You don't want to force your folded protein up against the unfolded portion because they will stick together and you will lose refolded protein .  Also, the cost of NDSB-201 has come down significantly over the last couple years. Try your standard refolding, but replace the Arginine with an equivalent molarity of NDSB-201 from Santa Cruz biotechnology (cat number sc-202237A) or some other vendor.  This almost always increases the yield compared to arginine refolding.

I agree with the suggestions. However, appreciate the risks involved in pursuing refolding proteins.

There is no universal solution for refolding proteins. It has to be determined empirically for each one. It could mean 10s or 100s of conditions to be tested. Some proteins can never be functionally refolded. These aspects are under emphasized by PIs and initially not appreciated by students and postdocs. 

Good luck with your work

I would suggest trying an on-column method for refolding. I have used the attached method with great success - it was adapted from Oganesyan, N., S. H. Kim and R. Kim (2005). "On-column protein refolding for crystallization." J Struct Funct Genomics 6(2-3): 177-182.

I did once refolding of protein on NiNTA column and it was quite successful ... It was quite a long time ago so I can describe you in general, how it went:

1. resuspended and sonicated the cells in tris buffer (pH 8.0) excessively to ensure complete cell disruption, centrifuged and collected the pellet

2. repeating step (1) but in tris buffer containing small amount of detergent (0.5% triton X-100)

3. wash cell pellet with tris buffer pH 8.0 (vortexed)

4. resuspended the pellet (50 mg/ml) in tris buffer pH 8.0 containing 6 M GdmHCl (or 6 M Urea), 5 mM beta-mercaptoethanol (or 1 mM DTT), 300-500 mM NaCl. I did this solubilization in a 30 ml nalgene high-centrifugation tube, the volume was about 5 ml, put in a small stirring bar. I put the tube vertical on a magnetic stirrer, run at mid speed for overnight (minimum 6 hours) at room temperature

5. centrifuged the tube for 30 minutes at 15,000 g, 4 deg. carefully took the supernatant (don't touch the black/dark precipitate). Here I have the solubilized inclusion bodies (IB). important: always start with as pure as possible material.

6. after equilibrated the Ni-NTA column with solubilizing solution (see above), I loaded the solubilized IB onto the column at low flow rate (0.5 ml/min), then eluted the column with solubilization solution (about 3 CV).

7. started gradient elution with the same buffer but low GdmHCl concentration (0.25 mM) and containing GSSG:GSH redox system for about 5 CV, even slower flow rate of 0.2 ml/min. 

8. eluted with tris buffer pH 8 for about 5 CV, then elute the protein out with usual elution of imidazole in tris buffer. centrifuge the collected protein solution at 15,000 g for 30 mins.

I successfully recovered the protein, which behaved as a monomer upon migration in gel filtration column and displayed monodispersity upon dynamic light scattering experiment. Unfortunately the protein (enzyme, actually) showed much less activity than expected, which I believed from difficulty to install the metal cofactor in the active site (it's a metallo-enzyme). The activity measured was true, though, because it still displayed similar inhibition profile and was far above the biomimetic activity. Further, SAXS experiment indicated that the amount of metal was less than expected. Thus most of my refolded protein was indeed uncharged.

Mind you, all the experiment was conducted in the cold room (about 4-7 deg). For this reason also, beware of tris buffer because its pH changes according to the temperature (and mess up the refolding). The refolding conditions must also be experimentally probed, unfortunately. But you can try some basic conditions and start from there (use of arginine, glycerol, redox system and ratio, etc.).

I did also successful refolding on size exclusion column or using direct dilution (different proteins) but for the aforementioned explanation, may be tricky if you have HisTag.

Good luck.