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

The size of the detergent-extractable population for each membrane protein in each domain was calculated by comparing the size of fluorescence peaks in digitized images of frozen sections from detergent-extracted and control monolayers. Since antibodies penetrate the full thickness of 0.5-~tm frozen sections, we determined the variability in section thickness and found that it was <15 %. This variability was statistically normalized by collecting images of several different sections from control and detergent-extracted cells. The linear correlation between antibody concentration and fluorescence intensity was also assayed in parallel controls (see Materials and Methods). The size of the fluorescence peaks was calculated by subtracting the average extracellular fluorescence background (typically 50-90 pixel units; i.e., see Fig. 4 , E and F, small white arrows). The extracellular fluorescence was not affected by detergent extraction. The resulting values of extractability in detergent are labeled E in Table II . Since this method cannot completely discriminate the contribution of cytoplasmic fluorescence, it was measured within the cell (Fig. 4, E and F, central white arrows, c) . Cytoplasmic fluorescence was 6-40 pixel units higher than the extracellular background. This relatively small fluorescence was very extractable in TX-100 (Table II, Cytoplasm). For A1, B1, and B2, the fluorescence on the plasma membrane was higher than the cytoplasmic fluorescence ("peaks"), even in the case of the smaller peaks, namely those in the incorrect domain from detergent extracted cells. Therefore, even these smaller peaks could be measured with no uncertainty. This was not the case for B2-m. For this marker, only the lateral domain showed clear peaks above the cytoplasmic fluorescence. The fluorescence profile did not show any apical or basal peak in nearly 60% of the cells. In those cases, the values were taken from the pixels on the corresponding domain as judged by parallel phase images. These values (apical and basal for B2-m) may contain a larger error due to the cytoplasmic fluorescence. The corresponding extractabilities are shown in Table II for comparison with the much lower extractability in the lateral domain. To determine the possible contribution of cytoplasmic fluorescence, a parallel set of extractability values was obtained for all the markers by subtracting cytoplasmic fluorescence from peak values (Table II, C) . Since extracellular background was not decreased by detergent extraction we considered E values as unbiased and will use them hereafter.