Did you begin by purifying the membrane fraction to minimize contamination with cytoplasmic proteins?

Dear Luiz,

Adam is right, membrane fraction purification should be the solution. Moreover I found this:

Based on this phase separation method and the earlier observation that detergents with low cmc values (such as Triton X-100 and Triton X-114) are relatively inefficient in solubilizing the GPI-anchored proteins, Hooper and Bashir developed a technique of differential solubilization and temperature-induced phase separation in Triton X-114 to distinguish between the GPI-anchored proteins and those anchored by a simple membrane-spanning polypeptide.

maybe it would be of help,

best regards,

Cedric

Might I suggest a recent paper that describes the separation of the membrane proteins from the GPI-anchored proteins using phase-separation as Cedric has mentioned (see attached).  You may also consider collecting the GPI-anchored proteins by activatin phospholipase C,  collecting the supernatant (therefore removing the cell bodies and contaminants) and concentrating the supernatant (A salt-ethanol wash of your cells and then freeze-drying the collected supernatant for analysis).  Hopefully this is helpful.

Article Proteomic Analysis of the Ciliary membrane of Paramecium tetraurelia

As Megan mentioned, GPI anchored proteins are cleaved from plasma membrane by phospholipase C. The cleaved GPI-anchored proteins are partitioned into aqueous phase by Triton X-114  phase separation. I do not know whether commercial anti-CRD (cross-reacting determinant) antibody is available now. This antibody recognizes a part of remained GPI anchor.  If you get, you can purify cleaved GPI-anchored proteins by immunoprecipitation.  

It might help.

Thank you all for the attention. I'll read the article sent by Megan and follow the tips that you sent me.