The crystal structure of insulin I. The investigation of air-dried insulin crystals.

Abstract

The crystal structure of insulin is of interest not only in itself but as part of the wider problem of protein structure considered as a whole. The early X-ray work on protein fibres gave strong support to the original chemical theories of the existence of polypeptide chains as the essential backbone of the protein molecule. The comparatively small spacings of 3-4, 4∙5 and 10-11 A which are most prominent on a large section of X-ray photographs of the scleroproteins are easily interpretable in terms of a fundamental chain structure (Astbury and Woods 1933; Astbury and Sisson 1935). It is on the other hand difficult to reconcile a chain structure with certain of the physical properties of the globular proteins, particularly their crystalline form and their behaviour in the ultracentrifuge as approximately spherical masses of molecular weight some multiple of 35,000 (Svedberg 1937 a ). Yet there is much evidence, physical as well as chemical, that the two types of protein structure are closely interrelated. Certain of the natural protein fibres show X-ray reflexions corresponding to very long spacings superimposed on the simple β keratin type of fibre pattern. These long spacings—up to 300 A in feather keratin (Astbury and Marwick 1932), tendon and collagen (Wyckolf and Corey 1936)—are more of the order of magnitude of the spacings which define the crystal structure, i. e. intermolecular distances, in crystals of the soluble proteins. And further fibre forms can be obtained from the crystalline proteins, particularly relevant observations being those on edestin and excelsin (Astbury, Dickinson and Bailey 1935). Several attempts have been made on the theoretical side to reconcile these facts, and two views of protein structure in particular have been stated in some detail. Astbury has advocated a variety of straight chain structures for the different fibre proteins which may be coiled into spiral forms or folded in layers in the globular proteins (Astbury and others 1935), as also discussed by Pauling and Mirsky (1936), while D. M. Wrinch (1937 a ) has shown that the occurrence of an internal cyclizing process might lead to the development from polypeptide chains of beautifully symmetrical “cyclol” networks both open and space enclosing. But on the experimental side many more observations are necessary before even these quite different hypotheses can be distinguished, and one possible approach seems to be through a more complete X-ray study of the crystalline globular proteins.