Separation on a SEC system is directly linked to the hydrodynamic radius of the protein and not the molecular weight. The hydrodynamic radius can be similar for an elongated dimer and a much more globular multimer.

Do you try another column such as superdex 75 or 200 to improve the separation in this range of size?

Dear Richelle

to provide an opinion i need to know some more details about your protein sequence as isoelettric point, number cisteines ? due you tried to check if addiction of reducing agent or edta affect the aggregation for example?

best regards

Manuele

To follow up on Sebastian Schmitt's comment.

I think your protein or proteins are associating or - in worst case - aggregating.

Did you change anything in your buffers? Composition? Salts or other things added? Do you have Ca2+ or other divalent ions in your buffer? Especially Ca2+, Mg2+, but also Cu2+ nd Zn2+ can affect protein association.

Thanks for your responses.

Sébastien Brûlé, yes I have tried G75 and S200 as well and I got similar profiles from both. I'm thinking that maybe if the columns would be longer, the resolution would be better? In the previous batch, I used an XK50 column with 91 cm height (~2L CV) since that was only a smaller scale. But this time I used a BPG100 and a BPX200 (6L and 28L CV, respectively). Relatively comparing the columns, the length of the bigger columns are less than that of XK50 in relation to the column diameters.

Sebastian Schmitt

Manuele Martinelli, my POI has a pI of 5.77 and 9 Cys residues. I haven't checked its profile on chromatography in reducing conditions, but on SDS-PAGE with b-mercaptoethanol, the multimers do break down to monomers and few dimers.

Achim Recktenwald, yes, I am quite convinced that's the case. The buffers I am using are the exact same ones I used previously (1st step:(A)50 mM Sodium phosphate pH 7.2, (B)A + 500 mM Imidazole; 2nd step: 20 mM Sodium acetate, 150 mM NaCl pH 5.5). I didn't change any buffer composition at all.

I also inquired if there was a difference in the production process of this batch compared with the previous small scale, but there was no difference (as I was told). The only different thing I observed with the starting material is the amount of the POI, which is a bit lesser this time.

Dear Richelle

if in sds page multimer brake down into monomer and dimer when you are reducing agent, it is probably because multimer is formed by aspecific interaction of some S-S and i think that to obtain an homogeneus and reproducible preparation you have to solve this issue.

theoretically you can try to run the sec with reducing agent but practically of some of the cys are involved in good intramolecular s-s you may unfold your protein. you can try to use sta page as a test to tune the concentration of reducing agent and see if you cam find a concentration able to broke multimer and not dimer and perform sec in this condiction but you need a lot of lucky.

is the structure of the protein or some homologues known? do you know whih are the cys involved in s-s? Because out of 9 cysteines, at least 1 is Free and it seems to be able to induce multimerization and probably the best way to solve your problem is, if you find a way to identify it is mutate it to serine?

one more question? how do you express this protein,? whith e.coli or mammalian cells? Because with so many cys, mammalian eg hek293 or expi293 are preferable.

good luck

Manuele

Did you try gel filtration or is gel filtration your second step in your purification?

Do you have thiols in your protein, even free ones? If yes, they could lead to cross-linking. If the conditions for cross-linking are borderline, it can sometimes happens and sometimes not.

Add some DTT, ~1 to 3mM, or ~10 to 15mM (stinky) mercpatoethanol to your buffer. With the latter, make sure that it is not much older than 6 months or so, as it easily oxidizes in air; it becomes yellow.

If I were in your shoes, i would use the original method S100 since you know that worked, with part of your sample. If you get good results, stick with it and do the several batches. Look at all the time it’s taking to try all these other things. And yes, a high column length to diameter ratio will always give better separation for SEC.

In the distant past my lab set up a recycling column so the peak of primary interest was sent back into the top of the column. A few passes did the trick. Is it clear that the formation and dissociation of dimers is slow? The co-elution of the two forms suggests a fairly fast equilibrium between the two.

Hi Robert C Woodworth, what you mean by "recycling column" is that you were using just a single column having the output tubing set-up to make the eluents reloaded? Or several columns of the same type were actually used? I am curious about how is this actually done.

I am not sure yet regarding the rate of formation and dissociation of the two forms, but just recently I found that by just lowering the pH a little bit, the multimers also decreased. I have yet to check again its resolution on gel filtration.

Hello Richelle, is the formation of the multimers perhaps dependent on a component in your buffer?

For instance, a lot of proteins bind to other proteins, form dimers and multimers in the presence of divalent ions. Especially Ca++ and Mg++, but Zn++ and others can play this role, as well. These ions do not need to be in the buffer you are using for the run, but can be a carry-over from previous steps. The concentrations needed are often very tiny.

Try adding 10mM to 20mM EDTA in one run. And if you know that it is Ca++, add EGTA, but EDTA will work, too. If it helps, you can then test reducing the EDTA to more normal concentrations in further runs.