There can be variations within the same polymorph or a completely different polymorph. A well known example for this is lysozyme. Changing the precipitant leads to growth of a different polymorph.

Have no proper facts to explain above phenomenon but changing condition refers different environment that protein is facing during crystallization. Protein shapes i.e. folding highly influenced by their environment. With this perspective of-course conditions may force a protein to adopt varying shapes.

Different conditions may produce crystals in different spacegroups, but these may or may not contain different confations of the protein. If the portion of the protein which exhibits different conformations is involved in crystal contacts then this would be indicative that the crystal packing has 'frozen out' the conformation observed. It may be that this region is flexible in solution. It is rarely (but sometimes) the case that a protein can exhibit two completely different folds, for example http://www.ebi.ac.uk/pdbe/quips?story=Transformer

For single domain proteins, the structures tend to be highly similar when solved under different conditions. However the introduction of flexible regions, multiple domains. and functionally-important conformational states (as present in most human proteins) means that a crystal structure may catch a number of different conformational and assembly states. The crystal lattice energy is not enough to change a fold fold, but it is enough to change a flexible region or trap one of several conformational states. Small angle x-ray scattering (SAXS) provides a complementary approach that provides a complete description of the solution ensemble under near physiological conditions. For SAXS analyses you can do mail-in SAXS of your sample and a matching buffer to get mass (see Nature 496, 477-81, 2013) and http://bl1231.als.lbl.gov.

Adding to the previous answers, but most directly John's, the energy of crystallization is not usually sufficient to change a fold, but the conditions preceding crystallization could have stabilized alternate conformations. So the answer is "yes, there is a possibility that different crystallization conditions give different conformations." But this would usually indicate that the different conditions reflect alternate biochemical milieus (each conformation having functional capacities in each milieu.)

Different shapes? Yes: What could happen is that N- and C-terminal extensions of a globular protein (e.g. some helical coil) may extend into solution in the presence of high salt in the crystallization condition

In my experience, globular compact proteins tends to have the same conformation in different crystallization conditions and sapce groups. Flexibles multidomains protiens could presents different conformations in different crystallization conditions and space groups.

Dear Santhosh,

Even though many participants are correct in their assertions they do not provide the key info. Proteins exist as an ensemble of different conformers. So speaking about a unique shape is a bit misnomer. Which of these conformers dominate in which condition is not well understood. Crystallization by definition is different (usually) from the native conditions, therefore it selects a dominant or a few dominating conformers. For more quantitative description of divergence in ensembles I refer you to two papers:

Global distribution of conformational states derived from redundant models in the PDB points to non-uniqueness of the protein structure

PV Burra, Y Zhang, A Godzik, B Stec

Proceedings of the National Academy of Sciences 106 (2009), 10505-10510

Sampling of the native conformational ensemble of myoglobin via structures in different crystalline environments

DA Kondrashov, W Zhang, R Aranda, B Stec, GN Phillips

Proteins: Structure, Function, and Bioinformatics 70 (2008), 353-362

In essence the same sequence can be found in highly divergent conformations no matter whether the protein is small or large, whether it has a single or multiple domains. It strongly depends on the design principle whether its function demands it to be, more, or less energetically stable.

However, be very careful when using people's formulations about the "crystal energy" or "energy of crystallization". These are rather misused and mystical terms. To learn more about what drives crystallization and how the energy is assigned to crystals please read a bit works by Peter Vekilov who concluded that in most systems the crystallization is driven by entropic gain not by enthalpic contributions. Crystal is formed from a subpopulation of similar matching conformers, therefore so called crystal cannot, by definition, distort anything.