HI Michael,

Other method you can use to follow your refolding is by HPLC. Also you can do MS analysis of your folded protein, and compare it to the reduced protein mass (folded mass M-12, for 6 disulfide bond and one Cys that is not oxidized.)

You can also follow the folding by alkylation analysis with iodoacetamide or iodoacetic acid, and check by MS.

Usually for the folding especially in the case of disulfide-rich proteins one need to have a redox buffer, mixture of oxidized and reduced glutathione (GSH and GSSG) are usually used.

Good luck,

Norman.

Wow, that's very cysteine rich. I assume the major difference between your original scale of purification and the problem you have now is that the concentration of your protein is higher. In your original preparation the protein concentration was low enough to prevent intermolecular disuphides forming quicker than the correct intra molecular ones. You could simply use a dilution of your new prep to achieve the same results but the volumes are likely to be prohibitive. The main approach used to deal with this is column refolding. I don't know what chromatography steps you used but any affinity column can be used. The protein is loaded in a denatured and reduced state and allowed to slowly refold and oxidise by buffer exchange.

You can determine how many free thiols are in your protein by reaction with DTNB. This will give a qantitative measure which in conjunction with an accurate protein concentration will allow you to calculate moles of thiols per mole of protein.

I think it would be worth optimising your screen further. When solublising definately use reducing agent (as much as you think you can get away with), then with and without during refolding. Eventually your aggregate will be retained in the gel filtration as you expect - after which you'll need somekind of reference before running an activity assay, CD, or what have you...

In addition you could look at alignments of this protein and its structure/ predicted structure/ homologs. I'm guessing you don't have the actual structure ;-). Looking here could indicate important cys's and potentially ones you could afford to lose.

Remember cysteins could also be involved in structural motifs other than disulfides.

Dear Michael,

you could include low concentration of reductant in the buffer to prevent aberant disulfide bond between chains. If the correct intra-chain disulfides are buried in the structure then they may not be acessible for reduction. You can check that you have inter-chain disulfides by running SDS-PAGE under non-reduction conditions.

If you want to see if you have the correct number of disulfide bonds formed for native structure then use DTNB (Ellman’s Reagent). This reacts with free cysteines. Use under native and denaturing conditions (e.g. 6M Guandinium Chloride). Remove any unattached dye with a desalting column. Comparing the change in absorbance of the dye relative to the total protein concentration will give the number of free cysteines.

Your binding activity would be the best way to see how much of your protein was native. If have a quantitative binding assay, then using protein concentrations higher than the dissociation constant will give a tritration where the apparent binding constant is at half the total number of sites. Similarly if you perform ITC then the stoichometry will tell you what fraction of protein is folded.

Also one can measure the number of thiols present in the refolded solution with Ellman's reagent. A similar measurement is made with papain since it has an odd thiol in its oxidized state (Brocklehurst, K., Kirestan, M., Little, G. The reaction of papain with Ellman's reagent, Biochem. J. 1972 128 811-816 discusses the various number of free thiols detected by DTNB). If the concentration of your protein solution is fairly dilute, it should decrease the aggregation problem. I would recommend a ratio of 10 GSH/1GSSG for the redox buffer. It might be anticipated that even under these conditions that dimerization would be possible. Your project sounds very interesting and difficult. Best of luck.

Using circular dichroism (CD), you can check the protein secondary structure (if its native structure is known) after refolding. Reg disulfide bonds; Number of thiols can be measured by DTNB assay. If you have enough sample volume, Raman spectroscopy can help determining conformation of disulphide bonds in 'refolded' protein (Please refer Tu AT (1982) Raman Spectroscopy in Biology: Principles and Applications. New York: Wiley. 448 p.)

A number of possible answers have already been suggested, but I will confine myself confine myself to what I think is a different possible approach. (For the record my lab has had extensive experience folding a 10,000 MW protein (neurophysin) that also has seven disulfides , but no "extra" cysteine. If I interpret your report correctly,you say that folding the protein in the presence of its natural ligand is successful, but that large scale attempts at folding are not successful. If I assume that the large scale attempts are not carried out in the presence of its natural ligand, that may be the problem, and, if possible, you might try adding the ligand to the large scale folding attempt ( or to an intermediate scale attempt if there is not enough ligand.) The presence of its ligand during folding should favor formation of the correct disulfides (because of the favorable free energy of binding) and allow them to form before unproductive disulfide formation and aggegation can can occur.

Peptide mapping by reverse phase HPLC or MS after proteolytic digestion of the refolded protein should help you to compare the disulfide bonding with that of the natural protein.

I think that the free cysteine is giving  you problems (unbonding cysteine is leading to unwanted agg) at large scale especially if it sits on the surface. This problem would also occur at small scale but at lower concentrations the free cysteine can form a mixed disulphide so therefore does not react. Maybe worth checking the pH of your buffers and amount of redox used was the same in both cases.

In my experience with similar molecules the odd number of cysteines always gave mixed populations (and depended on the cysteine vicinity) and the way we addressed it was by mutating the free cysteine to a serine or alanine. The mutation gave a correct mass by SDSPAGE (nr) and was found to be homogeneous.