Correct assay of complex I activity in human skin fibroblasts by timely addition of rotenone.

In a recent issue of Clinical Chemistry Janssen et al. (1) reported the development of a new spectropho-tometric assay to determine complex I activity in a mitochondrial fraction of human skin fibroblasts, which is based on measuring the reduction of 2,6-dichloroindophe-nol by electrons accepted from de-cylubiquinol. This is a potentially important finding because the determination of complex I in fibro-blasts is difficult owing to the high activity of contaminating rote-none-insensitive NADH dehydro-genases (2). In the reported method complex I was assayed by measuring the total NADH oxidase activity during a 4-min period, after which rotenone was added to measure the rotenone-insensitive NADH oxidase activity. By subtraction of the reaction rates, the complex I activity was calculated. Because it is well known that the accumulation of rotenone on its binding site is not instantaneous (3, 4), we questioned if the complex I assay might be affected by the delay in the inhibitory effect of ro-tenone. We investigated the relationship between the amount of mito-chondrial protein isolated from normal human skin fibroblasts exactly as described by Janssen et al. (1) and the time course of the NADH oxidase activity in the absence of rotenone. The duration of the first order kinetics decreased after 2–3 min, especially at the highest protein concentrations (Fig. 1). This result indicates that addition of rotenone 4 min after the start of the reaction and measurement of the changes in absor-bance during the next 4 min may lead to serious overestimation of the apparent rotenone sensitivity of complex I. To show the consequences of this finding, we performed the assay exactly according to the method described by Janssen et al. (1) by adding rotenone 4 min after the start of the reaction for the protein concentration range shown in Fig. 1. Obviously the complex I activity, if expressed as specific activity (in mU per mg protein), must be unrelated to the protein concentration. We found, however, a statistically significant linear relation between the specific complex I activity and the mito-chondrial protein concentration (R 2 ϭ 0.66, P ϭ 0.004, y ϭ 2.0x ϩ 81.9, where y ϭ mU/mg protein and x ϭ ␮g protein/mL), indicating that the complex I assay performed with the method of Janssen et al. is not reliable. Only with highly diluted fractions does the above-mentioned equation approach a constant, but in the regular assay reported by Janssen et al. the protein concentration amounts to …