Ahmad, FYI I am one of the authors of the review Lars mentioned and I am attaching a copy. I agree that it sounds like your protein is sticking to the column and thus eluting later than it should. Unless the ionic strength of your mobile phase is already very high (> ~.4) then you probably need to increase that. But you might want to try adding .1 to .2 M arginine hydrochloride rather than salt---sometimes arginine works very well at reducing the non-specific interactions (but not always). Note too that one can get the true molecular weight despite the odd elution position by using a classical light scattering detector after the column (SEC-MALS).

Dear Ahmad, what size do you see when you analyze the fractions on an SDS gel?

I was wondering the same as Wolfgang.

If you already run an SDS-PAGE gel prior to the gel filtration step that showed your protein was indeed 70 kDA before you loaded it onto the column, then it might have precipitated inside the column or even have bound to it.

If the 12 kDa peak fraction from your gel filtration runs at 70 kDa on an SDS-PAGE gel, you should calibrate the column again using the same buffer you used for the protein run (gel filtration columns should be calibrated regularly as the elution volumes change over time, particularly with heavy use and after aggressive cleaning). If this still shows that your 70 kDa protein elutes at 12 kDa, then there is something strange going on ... as I don't think a 70 kDa protein can run at 12 kDa regardless of its shape.

Finally, if your 12 kDa peak fraction runs at 12 kDa on an SDS-PAGE gel, then your protein has been degraded or it was never fully expressed ... or you are accidentally purifying something else (mix-ups happen even in the best labs. I once tried to purify a protein twice before I had the plasmid I was sent by a collaborator sequenced and found out it contained the wrong insert).

@Wolfgang Schechinger

@Antonio Ariza 

First of all i would like to apologized for the late reply as my hometown was affected by big flood.

My SDS page show protein of 70kDa and it was a native enzyme, not recombinant.

I run SDS page before and after I load into gel filtration column and it gave me the same size (70kDa).

may i know what does it means by "it might have precipitated inside the column or even have bound to it."? because as far as i know, gel filtration did not bind the protein to it.

Ahmad, looks to me that your GF column is not well calibrated or that your protein is indeed running strangely. As its elution is delayed, there might be an interaction with the GF matrix, as Antonio already suggested, or somehow it is so knurled up that it appears to be much smaller than expected. If that what you get from the column is what you want, I'd just go with it and note the protein's exceptional behaviour.

Edit: You also might check your column and the marker with some known proteins that might be available to you like BSA (approx 65kDa) and OVA (approx 25kD, IIRC).

thanks @Wolfgang Schechinger

i'll try to do that.

I think it is likely that your protein is interacting with your GF matrix as I have seen this on other occasions. The smaller apparent size of 12 kDa means it is eluting later than expected for a 70 kDa protein, consistent with interaction with the column and retardation in elution. I would not be concerned but if it is important to your studies you can try to run on different types of GF matrices. They are made from different combinations of acrylamide, agarose, and Sephadex polymers. One may yield your 70 kDa app size but even if they give different app sizes I think this would confirm this model. 

@Richard A Kahn

Thanks. i'll try to run the protein trough different type of GF matrices. May you suggest me any publication mentioning about a protein which interacting with GF matrix? Thank you so much.

I don't recall any specific references that contain documentation of this phenomenon. But I think it is well known that including some salts (e.g., 100 mM NaCl) in your buffer is common practice and the reason is to reduce ionic interactions between your protein and the matrix. If you need a reference I would look in the nucleoside diphosphate kinase (NDPK) literature. I once purified NDPK and I believe it eluted from the column about 2 column volumes later! Someone may have documented this unusually strong interaction, though unfortunately, I did not.

@Richard A Kahn

Thank you so much for the recommendation. Eluted after 2CV was far unusual than my protein. It was interesting to know that. Thanks again.

There is a 2010 review on matrix interactions in GF/SEC: T. Arakawa et al.: "The critical role of mobile phase composition in size exclusion chromatography of protein pharmaceuticals." J. Pharmaceutical Sciences vol 99 p 1674-1692: http://www.ncbi.nlm.nih.gov/pubmed/19894271

The first step should be to vary the NaCl concentration in the eluent and see if this varition affects the elution position. Increasing NaCl will tend to abolish any ionic interactions. But note that too much NaCl may cause hydrophobic interactions.

Ahmad, FYI I am one of the authors of the review Lars mentioned and I am attaching a copy. I agree that it sounds like your protein is sticking to the column and thus eluting later than it should. Unless the ionic strength of your mobile phase is already very high (> ~.4) then you probably need to increase that. But you might want to try adding .1 to .2 M arginine hydrochloride rather than salt---sometimes arginine works very well at reducing the non-specific interactions (but not always). Note too that one can get the true molecular weight despite the odd elution position by using a classical light scattering detector after the column (SEC-MALS).

that was very helpfull. thanks @Lars C Andersson and @John S Philo. i really appreciate it.