In the following publication they show the mobilities of standard phosphoproteins (sucha as casein, pepsin, and ovalbumin) in different buffer systems during Phos-tag SDS-PAGE. Offcourse the exact shift will depend on the specific method and tags ( Zn2+, Mn2+, ...) you are using.

http://onlinelibrary.wiley.com/doi/10.1002/pmic.201100524/full

Thanks Giel Detienne! I will read this paper carefully to perform my experiment .

Hi, Giel and Yifeng. This is Eiji, who is one of authors of the paper, which was introduced by Giel. If you have any questions, please feel free to contact me. Many thanks!

Hi,Eiji. Thanks for your kindly help. I will.

Use isoelectric focusing-- alone, if you've already purified the protein; or in 2D electrophoresis if not.  Two references:

http://www.pnas.org/content/103/44/16153.full

http://www.ncbi.nlm.nih.gov/pubmed/18466305

Three years ago I followed this paper exactly, same protein and antibody, but failed to detect phosph-protein. Still don't know what is wrong. :(

http://www.ncbi.nlm.nih.gov/pubmed/20661282

That is a complicated paper, Xianfeng--  and it focuses on the ER, so are you sure the extraction procedure you used was correct for your own protein?  There were other worrisome parts of the procedure, such as phosphatase reversal-- not enzymes I'd wish to have floating around the lab.  Perhaps this earlier Kinoshita paper is useful, if you're married to the idea of using molecular tags-- http://www.ncbi.nlm.nih.gov/pubmed/16340016/

HRP-labeled tags always require wisdom in analysis, the enzyme can have unwelcome chemical consequences (I prefer immunogold). An isoelectric focusing method would exploit the change in pI as a consequence of phosphorylation-- for a review of many types of post-translational modifications that exhibit a pI shift see: http://nar.oxfordjournals.org/content/32/suppl_2/W638.full