A Dominant Factor for Structural Classification of Protein Crystals

• Published in: Journal of chemical information and modeling 2015-08, Vol.55 (8), p.1673-1685
• Main Authors: Qi, Fei, Fudo, Satoshi, Neya, Saburo, Hoshino, Tyuji
• Format: Article
• Language: English
• Subjects: Amino acids; Animals; Chickens; Cluster Analysis; Crystal structure; Crystallization; Crystallography, X-Ray; Databases, Protein; Hemoglobin; Hemoglobins - chemistry; HIV; HIV Protease - chemistry; HIV-1 - chemistry; Human immunodeficiency virus; Humans; Mice; Models, Molecular; Muramidase - chemistry; Mutation; Myoglobin - chemistry; Protein Conformation; Proteins; Rabbits; Serum Albumin - chemistry; Sperm Whale; Swine
• ISSN: 1549-9596; 1549-960X
• DOI: 10.1021/acs.jcim.5b00052

With the increasing number of solved protein crystal structures, much information on protein shape and atom geometry has become available. It is of great interest to know the structural diversity for a single kind of protein. Our preliminary study suggested that multiple crystal structures of a single kind of protein can be classified into several groups from the viewpoint of structural similarity. In order to broadly examine this finding, cluster analysis was applied to the crystal structures of hemoglobin (Hb), myoglobin (Mb), human serum albumin (HSA), hen egg-white lysozyme (HEWL), and human immunodeficiency virus type 1 protease (HIV-1 PR), downloaded from the Protein Data Bank (PDB). As a result of classification by cluster analysis, 146 crystal structures of Hb were separated into five groups. The crystal structures of Mb (n = 284), HEWL (n = 336), HSA (n = 63), and HIV-1 PR (n = 488) were separated into six, five, three, and six groups, respectively. It was found that a major factor causing these structural separations is the space group of crystals and that crystallizing agents have an influence on the crystal structures. Amino acid mutation is a minor factor for the separation because no obvious point mutation making a specific cluster group was observed for the five kinds of proteins. In the classification of Hb and Mb, the species of protein source such as humans, rabbits, and mice is another significant factor. When the difference in amino sequence is large among species, the species of protein source is the primary factor causing cluster separation in the classification of crystal structures.