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

The possible anchoring to the cytoskeleton of the four antigens was assayed by extraction of live cells with 0.5 % TX-100 in extraction buffer (EB; 100 mM KC1, 4 mM EGTA, 2 mM MgCI2, 1 mM PMSE 60 mM Pipes, pH 6.9), which maximizes preservation of the MDCK suhmembrane cytoskeleton. As described by Fey et al. (25) , the extracted cells The antibody type was determined by agar immunodiffusion assay against specific antibodies. Mr was determined by immunoblots from SDS-PAGE of MDCK cell postnuclear supernatants. Some supernatants were treated with 50 mM sodium periodate and then assayed by immunoblot. Triton X-114 extractability: MDCK cell monolayers were extracted at 0°C in 0.5% Triton X-114, the detergent phase was spun at 37"C, and resuspended/spun twice. The proteins in the detergent phase were assayed by SDS-PAGE and immunoblot. The specific bands were cut and counted. The 100% was calculated as the cpm in the specific band from an SDS extract obtained from the same number of cells. pH 11 extraction: MDCK cell posmuclear supernatants were incubated in icecold 0.1 M Na2CO3, pH 1 i .0, or in PBS, pH 7.5, (control) for 30 rain (26) . The membranes were spun for 1 h at 100,000 g, and each membrane protein was determined by SDS-PAGE and immunoblot. The specific bands were cut and counted (100%: untreated membranes). kept to a large extent the nonextracted cell shape at the phase contrast and EM levels. In agreement with previous observations (18, 44) , an actin-rich submembrane cytoskeleton was observed associated with both apical and basolateral surfaces of MDCK cells by rhodamine phalloidin fluorescence on semi-thin frozen sections; larger amounts of actin were detected on the apical surface due to the presence of microvilli ( Fig. 1 A) . As shown by other groups (25, 31, 43) , TX-100 in EB failed to extract F-actin ( Fig. 1 B) and preserved the ultrastructural features of the cortical cytoskeleton including microvilli ( Fig. 1 C) . The time course of the extraction by TX-100-EB of the four MDCK antigens was studied in monolayers plated for 24 h by RIA with monoclonal antibodies (Fig. 2) . The bind-ing of both B1 and B2 monoclonals increased during the initial 30 s, presumably due to increased antibody accessibility to the basolateral domain, followed by a 50% decrease, for B1, or a slight 5% decrease, for B2, during the next 10 min (Fig. 2 A) . At this time a plateau was reached. The apical protein A1 displayed similar biphasic extraction kinetics, presumably due to some intracellular and basolateral localization (Fig. 2 B ; see also Fig. 3 C) ; the maximum level of extraction, 40% of the peak value, was reached after 10 min. On the other hand, A2, which is exclusively distributed on the apical plasma membrane in confluent monolayers with normal cell-cell contacts (82, 83) , was continuously extracted to <10% of the initial signal (Fig. 2 B) . Using the same protocol the total cellular extractability in TX-100 was assayed in 4-d confluent MDCK monolayers. The TX-100-extractable fractions in steady-state were similar to those shown in Fig. 2 : A1, 24%; A2, 90%; B1, 36%; and B2, 10%.