Electron microscopic techniques have been used to reveal two classes of subunits of tubulin in ordered arrays. Presumably the two classes correspond to the a and 13 polypeptide chains of tubulin that have been distinguished by chemical criteria. The two types of subunits alternate along individual protofilaments in microtubules, microtubule-precursor sheets, and extended zinc-tubulin sheets. The resolution of the two types of polypeptide chains is achieved by improved negative staining methods which produce micrographs with layer lines at 28 A-' and 84 A-1 in optical or computed transforms, in addition to the layer lines at 21 A-1 and 42 A-1 de- vector points in the same direction for all of the protofilaments of the microtubule. However, for the sheets assembled in the presence of zinc, adjacent protofilaments are staggered by about 21 A and oriented in an antiparallel arrangement with alterate protofilaments related by a 2-fold screw axis. The antiparallel alignment of the protofilaments in the zinc-tubulin sheets accounts for their planarity (no tubular structures are found in the presence of moderate concentrations of zinc), since the intrinsic curvature found with parallel alignment of protofilaments in the absence of zinc would be cancelled by the antiparallel arrangement. Tubulin, the principal component of microtubules, is a molecule with a molecular weight of 110,000 (1). It can be dissociated into a and 1 polypeptide chains which have similar molecular weights (about 55,000) but differ in electrophoretic mobility, amino acid sequence, and extent of phosphorylation (2-4). In terms of the structure of microtubules, morphological units spaced about 40 A along the microtubule axis are observed by electron microscopy and x-ray diffraction (5-9) corresponding in size to about 55,000 molecular weight or the size of a single a or 1 polypeptide chain. In principle, microtubules with 13 strands of protofilaments could be composed of (i) a and ,-chains alternating along protofilaments; (ii) some other regular pattern, such as a2 and 12 homodimers alternating along protofilaments; or (iii) a random arrangement of a and 13 polypeptide chains. Pattern i has been favored on the basis of crosslinking data, which support a heterodimer (a13) structure for tubulin (10), and observations on flagellar microtubules, which give an 80 A-i reflection in transforms suggestive of a pairing of a and 13 chains as would be expected for a heterodimer (7). However, the situation is still ambiguous, since the crosslinking data also reveal appreciable amounts of homodimers (10) and the 80 A reflection could also arise from pairings of polypeptide chains in arrangements other than heterodimers. Direct evidence for pattern i, such as the observation of distinct structures alternating along protofilaments, would settle this issue, and such evidence is presented here.Our evidence for an alternating pattern of a and 13 chains isThe publication costs of this article were defrayed in part by page charge payment. This article must the...