Unreliability of rate constants derived from a linear transformation of kinetic data, with special reference to cholesterol equilibration between phospholipid vesicles.

In the time-dependent transfer of a lipid from a donor to an acceptor vesicle population a(t) is the amount transferred to the acceptor vesicles at time t, a infinity is the equilibrium transfer value and a0 is the value at zero time. In order to plot kinetic data (a(t) as ln[(a infinity - a(t))/(a infinity - a(t))] against time and to fit these with a linear regression, it is necessary to know the equilibrium value, a infinity, or to choose one. Here it is shown that even if a very larger error is made in the choice of a infinity, the resulting plot can still be acceptably linear and the correlation coefficient of the regression acceptably high. When a infinity is overestimated the rate constant derived from the slope of such a plot is underestimated. In extreme cases a 10-fold error can occur.