Characterization of caveolin-rich membrane domains isolated from an endothelial-rich source: implications for human disease.

Caveolae are 50-100-nm membrane microdomains that represent a subcompartment of the plasma membrane. Previous morphological studies have implicated caveolae in (a) the transcytosis of macromolecules (including LDL and modified LDLs) across capillary endothelial cells, (b) the uptake of small molecules via a process termed potocytosis involving GPI-linked receptor molecules and an unknown anion transport protein, (c) interactions with the actin-based cytoskeleton, and (d) the compartmentalization of certain signaling molecules, including G-protein coupled receptors. Caveolin, a 22-kD integral membrane protein, is an important structural component of caveolae that was first identified as a major v-Src substrate in Rous sarcoma virus transformed cells. This finding initially suggested a relationship between caveolin, transmembrane signaling, and cellular transformation. We have recently developed a procedure for isolating caveolin-rich membrane domains from cultured cells. To facilitate biochemical manipulations, we have applied this procedure to lung tissue--an endothelial and caveolin-rich source-allowing large scale preparation of these complexes. These membrane domains retain approximately 85% of caveolin and approximately 55% of a GPI-linked marker protein, while they exclude > or = 98% of integral plasma membrane protein markers and > or = 99.6% of other organelle-specific membrane markers tested. Characterization of these complexes by micro-sequencing and immuno-blotting reveals known receptors for modified forms of LDL (scavenger receptors: CD 36 and RAGE), multiple GPI-linked proteins, an anion transporter (plasma membrane porin), cytoskeletal elements, and cytoplasmic signaling molecules--including Src-like kinases, hetero-trimeric G-proteins, and three members of the Rap family of small GTPases (Rap 1--the Ras tumor suppressor protein, Rap 2, and TC21). At least a fraction of the actin in these complexes appeared monomeric (G-actin), suggesting that these domains could represent membrane bound sites for microfilament nucleation/assembly during signaling. Given that the majority of these proteins are known molecules, our current studies provide a systematic basis for evaluating these interactions in vivo.