In our preliminary subcellular localization experiment we demonstrated that annexin II co-localized with submembranous actin in subpopulations of both cultured fibroblasts and keratinocytes. To investigate the physical interaction between annexin II and actin at the cell periphery, in vitro reconstitution experiments were carried out with keratins used as a control. Annexin II, isolated by immunoaffinity column chromatography, was found to exist as globular structures measuring 10 to 25 nm in diameter by rotary shadowing, similar to a previous report. We believe that these structures represent its polymeric forms. By negative staining, monomeric annexin II was detectable as tapered rods, measuring 6 nm in length and 1 to 2 nm in diameter. When annexin II was mixed with actin in 3 mM piperazine-N, N-bis-2-ethanesulfonic acid (PIPES) buffer with 10 mM NaCl2, 2 mM MgCl2 and 0.1 mM CaCl2, thick twisting actin bundles formed, confirming previous reports. This bundling was much reduced when calcium was removed. In the presence of 5 mM ethylenediamine tetra-acetic acid (EDTA) in 5 mM tris, pH 7.2, keratins were found to form a network of filaments, which began to disassemble when the chelator was removed and became fragmented when 0.1 mM CaCl2 was added. Keratins under the same conditions did not fragment when annexin II was present. These results suggest that annexin II, in conjunction with Ca2+, may be involved in a flexible system accommodating changes in the membrane cytoskeletal framework at the cell periphery in keratinocytes.

Cultured human epithelial cells stained with antibody to desmosomal proteins by indirect immunofluorescence showed linear arrays of desmosomes en face between stratified cells. To confirm that an extensive linear pattern existed on the cell surface, subconfluent cultures were viewed using scanning electron microscopy. Aligned arrays of blunt protrusions lying parallel to each other and extending in the direction of the long axis of the cell were observed on the surface of groups of superficial cells in intact cultures. That this pattern was indeed related to desmosomal distribution was verified by transmission microscopy of thin sections cut in a plane between the upper and lower surfaces of flattened stratified cells to view desmosomes directly. A similar arrangement of desmosomes was seen in intact tissue, using epidermal sheets separated from newborn foreskin. The same pattern found in flattened cells was sometimes apparent in more rounded basal cells where the cytoplasm was beginning to extend. Since desmosomal plaques are associated with keratin filaments, the alignment of desmosomes must occur in association with cytoskeletal changes as cells become flattened toward the distal epithelial surface. The primary initiation of desmosomal alignment remains to be investigated. However, the present findings demonstrate an increasingly regular membrane-cytoskeletal spatial interaction as stratified epithelial cells of skin mature.