Relative activity factors (RAFs) and immunoquantified levels of cytochrome P450 (CYP) isoforms both have been proposed as scaling factors for the prediction of hepatic drug metabolism from studies using cDNA-expressed CYPs. However, a systematic comparison of the two approaches, including possible mechanisms underlying differences, is not available. In this study, RAFs determined for CYPs 1A2, 2B6, 2C19, 2D6, and 3A4 in 12 human livers using lymphoblast-expressed enzymes were compared to immunoquantified protein levels. 2C19, 2D6, and 3A4 RAFs were similar to immunoquantified enzyme levels. In contrast, 1A2 RAFs were 5- to 20-fold higher than CYP1A2 content, and the RAF:content ratio was positively correlated with the molar ratio of NADPH:CYP oxidoreductase (OR) to CYP1A2. The OR:CYP1A2 ratio in lymphoblast microsomes was 92-fold lower than in human liver microsomes. Reconstitution experiments demonstrated a 10- to 20-fold lower activity at OR:CYP1A2 ratios similar to those in lymphoblasts, compared with those in human livers. CYP2B6-containing lymphoblast microsomes had 29- and 13-fold lower OR:CYP and cytochrome b(5):CYP ratios, respectively, than did liver microsomes and yielded RAFs that were 6-fold higher than CYP2B6 content. Use of metabolic rates from cDNA-expressed CYPs containing nonphysiologic concentrations of electron-transfer proteins (relative to human liver microsomes) in conjunction with hepatic CYP contents may lead to incorrect predictions of liver microsomal rates and relative contributions of individual isoforms. Scaling factors used in bridging the gap between expression systems and liver microsomes should not only incorporate relative hepatic abundance of individual CYPs but also account for differences in activity per unit enzyme in the two systems.