Immunoblot detection of antigens in immunoprecipitates.

Co-immunoprecipitation of antigens from cell or tissue extracts is used in a wide variety of experimental systems to corroborate the existence of authentic protein-protein interactions in vivo. Typically, stable protein-protein associations are detected by immunoprecipitation with antibodies specific for one protein, followed by immunoblot analysis with antibodies against potential interacting proteins. However, in many cases, the only immunoreagents available for blotting are antibodies that have been generated in the same animal species as the antibodies used for immunoprecipitation. This can give rise to extremely high levels of background staining on blots because secondary antibody conjugates bind directly to the electrophoresed immunoglobulins in the immunoprecipitate. We have overcome this problem by probing immunoblots of immunoprecipitates with biotinylated antibodies, followed by an avidin-peroxidase conjugate. We used rabbit antibodies against human ezrin (2) to demonstrate the difference between a conventional method and our improved method for detecting antigens in immunoprecipitates (Figure 1). Ezrin is an 81-kDa membrane-cytoskeletal linking protein (Figure 1, arrowhead) that is highly enriched in microvilli, which line the human placental trophoblast epithelium (1). Immunoprecipitates of ezrin were prepared from soluble extracts of isolated placental microvilli, separated by SDS-PAGE, transferred to membranes, and then stained using either unconjugated (Figure 1A) or biotinylated (Figure 1B) ezrin antibodies, followed by goat antirabbit IgG-peroxidase or avidin-peroxidase secondary reagents, respectively. In the conventional method, a major band at 81 kDa, corresponding to the mobility of ezrin, was easily observed in immunoprecipitates under reducing conditions (Figure 1A, lane 6). However, this band was not seen under non-reducing conditions (Figure 1A, lane 8) because the high levels of background staining on the blot occlude it. This background staining, which was particularly abundant in non-reduced samples, is due to the direct binding of the secondary goat anti-rabbit IgG to the rabbit immunoglobulins present in the Laemmli (3) sample of the immunoprecipitate (compare Figure 1A, lanes 5–8, and Figure 1A, lanes 5′–8′). In contrast, Benchmarks