Membrane-cytoskeleton associations during myogenesis deviate from traditional definitions

Plasma membrane-cytoskeleton associations involving four membrane proteins (A5, H58, H36, and I20) were studied in developing L8E63 rat skeletal muscle cells using immunofluorescence microscopy and photometry on the basis of three criteria: Triton-insolubility, colocalization with cytoskeletal components, and sensitivity to cytoskeleton-directed drugs. The results presented demonstrate that there are developmental stage-specific associations between membrane proteins and the cytoskeleton during skeletal myogenesis. Several inconsistencies were found with traditional expectations of membrane-cytoskeleton associations. For example, although A5 is Triton-insoluble and sensitive to cytochalasin, its distribution generally does not correspond with any known cytoskeletal structure. Furthermore, the topography of A5 is dependent on the integrity of the plasma membrane. H36 and I20 are completely soluble in Triton and therefore by accepted definitions would not be expected to be associated with any cytoskeletal component. Yet H36 and actin codisrupt in the presence of cytochalasin, while I20, whose distribution does not correspond with microtubules, is uniquely sensitive to their disruption. These results demonstrate that (i) neither Triton-solubility nor colocalization alone predicts all membrane-cytoskeleton associations; some associations between the membrane and cytoskeleton are unstable in nonionic detergent; (ii) the native distribution of proteins in the membrane may not reflect their cytoskeletal associations; and (iii) the topography of some membrane proteins with no apparent association with the cytoskeleton may be greatly influenced by the cell cytoskeleton.