You didn't say that the protein didn't elute from the column. You said you lost the activity. Did the protein come off the column, but in an inactive state? Try running fractions on SDS-PAGE or doing a protein assay on them to see if the protein came off.

If the enzyme does not elute, try including some salt (150 mM) in the elution buffer to keep the protein from sticking to the column. If your enzyme binds Sepharose, which is made from agarose, then including a high concentration of the D-galactose might compete with binding.

Thanks Adam and Alim,

my enzyme does not elute. I will try with both suggestions. Thanks a lot.

Please see the link

http://www.gelifesciences.co.jp/catalog/pdf/18117288.pdf

Although not so highly glycosylated, Sephadex G-200 has given the good result for membrane glycoprotein of brain lipoamidase (with 0.1% Nonidet P-40 (a trademark of Shell Chemicals or of Royal Dutch Shell) or octylphenoxypolyethoxyethanol; Please see Fig.2 in file. Purify brain lipoamidase).   High-speed SEC may be achievable by using HPLC-Soap-SEC (Please see file; SEC column 300A silica) using following eluent; i.e., 0.1 M sodium phosphate, 1.5 M sodium chloride, glycerol 40% (v/v), 2-propanol 10% (v/v), and 1% (v/v) of PGC (polyethylene-glycol 1000 monocetyl ether; Brij-58), and finally filtrate by PTFE membrane (Please see file; SEC fucoidan determination).   It is noteworthy that PGC (polyethylene-glycol 1000 monocetyl ether) should be used of Wako (Osaka, Japan) instead of Sigma (St. Louis, MO, USA), since the range of molecular weight (Mr) distribution of the PGC seems narrower in Wako's one (as declared in the article of "SEC fucoidan determination").   Please use glycerol of  special grade and 2-propanol of HPLC grade.   

Enzyme activity may be recovered by HPLC-Soap-SEC (please see file; HPLC-SEC BIN NP-40).

By the way, I do not understand why the term "Brij-58" (trademark of Croda International Public Limited Company; Snaith, East Riding of Yorkshire, UK) is inhibited to use in the scientific research article.   I must have to use the another term "PGC" (polyethylene-glycol 1000 monocetyl ether) in the research paper instead of the term "Brij-58" (see file; SEC fucoidan determination). 

. 

Thanks Kou

Our enzymes bind specifically to dextran, which is a component of both Sephacryl and Sephadex. To avoid this, we use polyacrylamide gel filtration media, such as the  BioGel P-series. Sephadex can be useful for us as an affinity purification step, in which case we elute the enzymes with soluble dextran solutions. Of course, then the enzyme also contains dextran, which can be an issue.

Thanks a lot