I want to separate the plasma membrane fraction from the cultured muscle cells upon sample treatment in order to analyze Glut4 translocation into the plasma membrane. I usually extract the cells with 0.1% NP-40 buffer followed by a homogenization step using a handheld pestle homogenizer. The homogenization is carried out for 30 s followed by 10s interval, which is repeated 5 times. Then the cell homogenate is centrifuged and the supernatant is collected as the cytosolic protein fraction. The precipitate is retreated with 1% NP-40 buffer after incubating on ice for 1h. Then the plasma membrane fraction is collected as the supernatant. Glut4 translocation is determined by western blotting. However, I often find the reference protein for plasma membrane (IR-beta) in the cytosolic protein fraction. Is this due to the homogenization step? If so, what is the optimal conditions for homogenization? Prolong or just one or two strokes? Or else, how could I optimize the protocol to ensure a complete extraction of plasma membrane proteins?