To date, a selection strategy that imparts a growth advantage specifically to cells that express beta-gal is not available. However, beta-gal can be detected in live cells using fluorescent substrates, and fluorescent cells can be efficiently separated from mixed populations rapidly and efficiently using flow cytometry without affecting their viability. This assay is relatively simple and can be performed in less than half an hour when the number of samples to be analyzed is limited. Furthermore, it is quantitative and we use it routinely to follow the kinetics of a given interaction in response to specific signals. The traditional substrate used for this assay, fluorescein di-b-D-galactopyranoside (FdG) is not cell permeable but it can be introduced into the cells by hypotonic shock. Cleavage by beta-gal results in the production of free fluorescein, which is also unable to cross the plasma membrane and is trapped inside the beta-gal positive cells. The cells to be analyzed are trypsinized, resuspended in phosphate buffered saline (PBS) containing 5% fetal bovine serum (PBS/FBS) and pelleted in 5ml polystyrene round-bottom tubes (Falcon 2058). The cells are then resuspended in 100 ml of PBS/FBS and an equal volume of bi-distilled water containing the substrate at a concentration of 1 mM is added. After three minutes at room temperature, the hypotonic conditions are quenched by adding ten volumes of ice-cold PBS/FBS containing 1mg/ml Propidium Iodine (PI). PI is a red fluorescent compound that is actively excluded from living cells but accumulates in dead cells, allowing their exclusion from the analysis or the sorting. After quenching, the cells are pelleted again, resuspended in approximately 200 ml of ice cold PBS/FBS and analyzed on a Becton-Dickinson FACScan or sorted on a Becton-Dickinson FACStar flow cytometer. In most clonal populations, the cells respond to signals inducing the test interaction with an homogenous increase in beta-gal activity, and the mean fluorescence in each sample can be used as a reliable measure of the interaction. In polyclonal populations however, we have often noticed that a subset of the cells are not responding to the inducing signal, and the mean fluorescence is less useful. However, the main advantage of the FACS based assay is that it provides a simple means of selecting for cells in which a given interaction takes place, opening the path to screening cDNA libraries for novel interaction partners of a given "bait" protein in mammalian cells. This rapid method of selection is in our opinion superior to growth-based selection protocols, as many interactions are likely to be transient under physiological conditions and could possibly affect the growth characteristics of the cells.