Polypeptide chain configurations in crystalline proteins

Abstract

Astbury’s studies of α-keratin, and X-ray studies of crystalline haemoglobin and myoglobin by Perutz and Kendrew, agree in indicating some form of folded polypeptide chain which has a repeat distance of about 5·1 Å, with three amino-acid residues per repeat. In this paper a systematic survey has been made of chain models which conform to established bond lengths and angles, and which are held in a folded form by N—H—O bonds. After excluding the models which depart widely from the observed repeat distance and number of residues per repeat, an attempt is made to reduce the number of possibilities still further by comparing vector diagrams of the models with Patterson projections based on the X-ray data. When this comparison is made for two-dimensional Patterson projections on a plane at right angles to the chain, the evidence favours chains of the general type proposed for a-keratin by Astbury. These chains have a dyad axis with six residues in a repeat distance of 10·2 Å, and are composed of approximately coplanar folds. As a further test, these chains are placed in the myoglobin structure, and a comparison is made between calculated and observed F values for a zone parallel to the chains; the agreement is remarkably close taking into account the omission from the calculations of the unknown effect of the side-chains. On the other hand, a study of the three-dimensional Patterson of haemoglobin shows how cautious one must be in accepting this agreement as significant. Successive portions of the rod of high vector density which has been supposed to represent the chains give widely different projections and show no evidence of a dyad axis. The evidence is still too slender for definite conclusions to be drawn, but it indicates that a further intensive study of these proteins, and in particular of myoglobin which has promising features of simplicity, may lead to a determination of the chain structure.