Panning was used to select co-transfected cells expressing plasmid-encoded ion channels. Adherent cells were cotransfected by the CaPO4 method with a plasmid encoding a cell surface marker (CD8) along with another plasmid encoding an ion channel. At 1-3 days post-transfection, the cells were suspended, treated with a biotinylated CD8-specific antibody and placed into streptavidin-coated bacterial petri dishes. After 2 h, these dishes were washed with a saline solution to remove cells that did not adhere to the streptavidin-coated dishes. By using molar ratios of > or = 8:1 of the ion channel encoding plasmid to the CD8 plasmid, we found that > or = 50% of the panned cells that adhered to coated dishes were positive for expression of the co-selected gene. Cells expressing plasmid-encoded channels (voltage-dependent sodium channels or cystic fibrosis transregulator chloride channels) were assayed using whole-cell recording, perforated-patch recording and single-channel recording. The method was tested on tsA201 and NIH3T3 cells, the latter of which transfected very poorly (usually < 4% efficiency) with our standard protocols. When the co-selected plasmid encoded the bacterial beta-galactosidase gene, it was possible to determine by histological assay the percentage of positively transfected cells (with and without panning). Panning in some cases increased the percentage of positively cotransfected cells by more than 20-fold. This technique is particularly useful when selecting co-transfected cells for electro-physiological recordings on individual cells.