In this study, metabolically radiolabeled Escherichia coli cell extracts were used to systematically evaluate protein recoveries at each step of two-dimensional (2-D) electrophoresis and using different sample application methods. Sample application using sample cups resulted in better protein recovery compared with sample loading by rehydration when the Multiphor system was used. At least 50% or more of an E. coli extract was lost when high protein amounts (500 mu g) were loaded by rehydration using this system, which employs separate holders for rehydration and isoelectric focusing (IEF). In contrast, when the IPGphor system was used, rehydration sample loading consistently yielded the highest overall protein recoveries. These improved protein recoveries were due to integration of rehydration and electrophoretic separation in a single unit. Even at high protein loads (500 mu g), less than 15-20% of the proteins were lost when proteins were loaded by rehydration using sample buffer containing 2% carrier ampholytes in the ceramic immobilized pH gradient (IPG) strip holders used for both rehydration and IEF. Regardless of the loading conditions used, carrier ampholytes in the sample buffer increased protein recoveries. Use of thiourea did not significantly affect protein recoveries but did improve protein resolution in 2-D gels as expected. in summary, these results show the best protein recoveries are obtained for ail protein loads when samples are applied to IPG strips during rehydration using a single device for both rehydration and IEF. In contrast, the poorest recoveries are obtained when rehydration and IEF are performed in separate devices, and losses increase dramatically with increasing protein loads using this approach.