Selective anchoring in the specific plasma membrane domain: a role in epithelial cell polarity

We studied by FRAP the diffusional parameters of antigens localized in the apical and basolateral membrane. The probes were added differentially to both domains in cells grown on collagen-coated nylon nets, similar to those originally used by Cereijido et al. (12) , except that in our case the collagen gels were native to allow diffusion of antibodies. We could use only three monovalent probes in FRAP measurements: B1 Fab, B2-m, and A2 Fab. A2 is a highly polarized antigen (polarity ratio >15:1; reference 82), so no signal could be detected on the basolateral domain. B1 and B2-m appear in the apical domain in large amounts when the ceils are prevented from forming cell-cell contacts (either in sparse cells or in cells incubated in low calcium medium). Upon establishment of a confluent monolayer, B1 (82) and B2 (unpublished observations) become polarized over a 4-d period. Significant amounts of B1 and B2 are still present in the apical (incorrect) domain 24 h after plating and are amenable for FRAP measurements, while A2 is already polarized and cannot be detected in the basolateral (incorrect) domain. In the 4-d stage, on the other hand, only tracer amounts of B1 and B2 are present in the apical domain. These amounts could be determined as small peaks above background in the fluorescence profiles on semi-thin frozen sections, but we could not discriminate them from cellular autofluorescence in intact cells. Therefore FRAP could be measured in the incorrect domain only when significant amounts of the antigen were present. Under these technical restrictions, we found small differences in the diffusion coefficients and no differences in the recovery fractions between apical and basolateral membranes for both basolateral markers. The values of these parameters reported here for B1 and B2-m are in good agreement with similar values reported by Jesaitis and Yguerabide (34) for (Na÷-K +) ATPase on the free surface of subconfluent MDCK cells and in scraped confluent monolayers. The latter condition has been extensively used to load macromolecules in various cell types and is likely to increase cytoplasmic calcium levels that may affect the cytoskeletal functions. This basolateral marker of MDCK cells shows similar kinetics of polarization and similar dependence on intercellular contacts as our B1 and B2 markers (48, 74) .