Many proteins can be damaged by lyophilization. Use ultrafiltration to concentrate the protein. Store it by flash-freezing and keeping at -80 or in liquid nitrogen.

thanks a lot, Dr. Shapiro, I'm wondering what kind of damage exactly may happen?  is adding some glycerol helpful? I lost a lot protein by using centrifugal filter unit (Amicon Ultra),  even my protein is several times bigger than the cutoff if the filter. not sure why..

Lyophilization causes the protein to aggregate and precipitate and depending on how well your protein resoluabilizes may not be a huge problem. For example, lysozyme can be lyophilized, resuspended and crystalized. Some disorder is necessary as you need some nucleation to form crystals. Most people are not working with proteins as stable as lysozyme, and so most prefer not to lyophilize. It is helpful to avoid batch effects during crystalization trials by mass producing your protein of interest and then snap freezing many aliquots in liquid nitrogen. No cryoprotection necessary. If you are having a hard time concentrating your protein, you could try to dry it down gently with Ammonium sulfate. Spectrum labs has very small dialysis units that you can use with ammonium sulfate concentration. BTW if you want to test protein 'happiness' you can do it more analytically with dynamic light scattering. If the resuspended protein is mono modal and mono disperse than it is most likely well soluble. Good luck!! 

Cryoprotectants, especially sugars like trehalose, are used to help stabilize proteins during lyophilization. Glycerol is also useful as a stabilizer, but it can't be used for lyophilization because it is a non-volatile liquid. It can be used as a stabilizer when freezing the protein for storage, but you will probably have to remove it before setting up your crystallization trials. The same goes for trehalose or other cryoprotectants.

If your protein is sticking to the concentrator, try using one with a different membrane chemistry. Another way to concentrate a protein sample is to put it in a dialysis bag and immerse the bag in a concentrated PEG solution, or bury it in a solid water-absorbent substance like carboxymethylcellulose. Still another method is to bind it to an ion-exchange resin, then elute it in a small volume of a high-salt buffer. Obviously, this leaves you with a salty sample. If the protein is His-tagged, you can do the same thing with nickel resin, eluting with a high concentration of imidazole.

BTW: I would avoid freezing of protein samples. Some proteins are already denatured or aggregated by freezing/thawing alone.

Personally, I like Amicon stirred cells (now sold by Millipore) for concentrating. The water, salts and other small molecules are pressed through a semipermeable membrane with nitrogen at moderate pressure, large molecules like proteins are retained by the filter. Because the content of the cell is stirred, protein aggregation on the membrane is minimised.

There are so-called "ordered water molecules" in many protein structures, and this set of water molecules are considered integral part of the protein structure.  If you remove them by lyophilization or spray drying in vacuo, you defeat the purpose of  solving the structure of your protein, because the structure will have been destroyed by drying.

Thanks a lot all of your helpful insights.I really appreciate.

I'm using lyophilizer just to reduce the volume of the sample, not waiting until all water are removed. I'll check with LC/MS again to make sure no protein modification resulting from this step.