The quality of the crystals is fundamental for preparation of heavy-atom derivatives, X-ray diffraction measurements and phasing of protein structure. Proteins have often domain structure (domains are connected by flexible linkers). Flexibility of domains in a proteins may generate heterogenity in structure. Removal of a flexible linker region between two domains can help in crystallization by minimizing effects resulting from this microheterogenity. Another types of heterogenity (disordering amino- and carboxy- terminal fragments) may be removed by engineering of terminal sequences or by synthesis of more compact domains. Side chains of surface residues are involved in crystal contacts. When crystals are of poor quality, residues on the protein surface may be mutated to remove bad contacts or some residues can be introduced to promote good surface interaction.

Human H ferritin - By protein engineering in the sequence of the intermolecular contact region have been changed and crystals isomorphous with the homologous rat L ferritin has been obtained [34]

GroEL - large protein assembly involved in the ATP-dependent folding of polypeptide chains The original native crystals diffract poorly, double point mutation Arg 13 Gly and Ala 126 Val produced crystalls with better diffraction qualities. The multimeric protein assembly were less stable, although still functional. The crystal structure at 2.8 A resolution has been determined [35]

Human Interferon gamma - five carboxy-terminal residues have been deleted by protein engineering. The C-terminus was targeted bacause it was known to be protease sensitive. Structure of a compact domain (generated by the mutations) has been determined [36]

HIV-1 integrase - Crystals of the native protein diffract poorly. A point mutation Phe 185 Lys resulted in a protein with improved solubility and native activity.The crystal structure of the catalytically active core domain (residues 50 to 212) of HIV-1 integrase was determined at 2.5 A resolution [37]