Three-dimensional structure analysis of the crystalline moiety of A-starch

Abstract

Despite the fact the first first X-ray diffraction diagrams of starch were reported in the 1930s (I) , unambiguous description of the native crystalline arrangement is not yet complete. Kainuma and French (2) were the first to suggest a double-stranded helix for the B polymorph (commonly found in tubers). This origianl proposal was extensively considered by Wu and Sarko (3,4) who arrived at the conclusion that both B and A types (found in cereals) can be described as right-handed, parallel-stranded double helices, packed in an antiparallel fashion. However, A and B polymorphs differ considerably in the crystalline packing of the duplexes. Current models derived from biochemical data (5,6) rely only on parallel arrangements of molecules. Similar arguments about chain polarity also arise from biosynthetic considerations; it would be more reasonable to expect parallel packing of parallel-stranded double helices since all the chains would have the same polarity. There is therefore a discrepancy between the currently accepted models (7). The present paper is concerned with the threedimensional structure analysis of the A-type polymorph . New crystallographic data were obtained by using two complementary methods: electron diffraction on single, micron-sized crystals (8) and deconvolution of X-ray powder patterns (9) into individual peaks. Wide-angle X-ray diffraction diagrams recorded were similar to those displayed by native cereal starch granules and those corresponding to fibrallar samples. They all can be indexed using an orthorhombic-like unit cell having a volume of -2.2 nrrr'. According to the measured density of 1.48, 12 glucose residues, together with a few water molecules, can be accommodated in the unit cell. Computer modelling of IX (1-4)-linked glucose units indicated that the observed repeat of c = 1.069 nm can either correspond to an extended conformation of a maltotriose unit or to a shallow conformation of a maltoheaxose unit. Therefore, three structural hypotheses had to be tested against their packing constraints: (i) four single three-fold helices (n = 3, h = 0.355 nm); (ii) two six-fold helices (n == 6, h = 0.178 nm); and (iii) two double six-fold helices repeating in 2c = 2.139 nm [in order to restore the observed repeat of 1.069 nm, each double helix has to be parallel stranded, the single strands (n = 6, h = 0.355 nm) being associated to its counterpart through a two-fold axis]. All cases were thoroughly examined and only the latter was Shown to be a valid candidate. Further molecular modelling showed that the conformation at the glycosidic linkage corresponds to the intersectionof n = 6 and h -0.355 nm (left-handedchirality) (Figure I). Such a situation corresponds to one of the stable energy minima (10) and is reminiscent of the observed conformation of a crystalline amylose oligomer (11). Assembling two such parallel single strands through a two-fold symmetry operation permits forma-