Redistribution and differential extraction of soluble proteins in permeabilized cultured cells. Implications for immunofluorescence microscopy

Immunofluorescence microscopy is widely used to characterize the cellular distribution of both soluble and structural proteins. Control experiments generally address only the specificity of the antibodies used. The permeabilization/fixation conditions used to prepare cells for antibody application are assumed to preserve faithfully the in vivo distributions of the protein(s) being examined. We systematically tested the extent to which soluble proteins are redistributed into inappropriate locations and are differentially extracted from native locations during the permeabilization and fixation of the cells before antibody application. We separately introduce six soluble FITC-conjugated proteins of different net charges and sizes into living cultured cells. The labeled proteins do not adhere to the external surfaces of living cells and are evenly distributed throughout the cytoplasm with the larger proteins being excluded from the nucleus. The cells are then prepared as if for immunofluorescence using several conditions that encompass many of the methods commonly used for this purpose. Cells permeabilized with 0.1-0.2% Triton X-100 before fixation with 3.7% paraformaldehyde show a striking localization of all but one of the test proteins to the nucleus and/or nucleoli of 60–80% of labeled cells. Punctate cytoplasmic labeling and cytoskeletal-like arrays of labeled protein are also observed. Extraction with 1% detergent prior to fixation removes most but not always all of the exogenous proteins from the cell remnants. Permeabilization of cells with 0.1% detergent after paraformaldehyde fixation leaves a reticular, uneven cytoplasmic distribution of the labeled proteins, and some of the larger proteins are redistributed to the nuclei. Direct fixation/permeabilization with -20 degrees C methanol largely preserves the in vivo distributions of fluorescent proteins with some preferential localization of these proteins to nuclei, nucleoli and the perinuclear region. These results show that misleading apparent localizations of soluble proteins can result from their redistribution and/or differential extraction during the preparation of cells for primary antibody application.