A selectively driven molecular clock

Abstract

The problem of whether mutations that increase a species' fitness are incorporated into the genome as a function of generation interval or as a function of absolute time is of some interest because of the clock-like nature of amino acid substitutions known to have occurred in the evolution of certain well studied proteins. The molecular clock features a rate of amino acid substitution that is relatively constant when averaged over a sufficiently long time, and the constant rate of substitution is independent of the generation interval of the species involved1,2. Furthermore, the variance in the rate of substitution is roughly twice the mean2. Some authors have argued that the very existence of a molecular clock implies that most amino acid substitutions must be selectively neutral or nearly neutral3. On the other hand, there is no compelling reason to assume that selectively driven gene substitutions would not also exhibit clock-like behaviour, in which case the rate of increase in fitness of a population should be a function of absolute time rather than generation interval. In this report, we have investigated the issue of rate of increase in fitness using initially isogeneic strains of Escherichia coli that had been allowed to evolve independently in glucose-limited chemostats4 for 500 h. The nutrient supply was adjusted in such a manner that some of the strains maintained a doubling time of 2.5 h and others a doubling time of 5.0 h, and the relative fitnesses of the evolving strains were measured periodically. Our main conclusions are that (1) selectively driven gene substitutions detectable as increases in fitness do exhibit clocklike behaviour, (2) when averaged over a sufficiently long time, the rate of increase in fitness is a function of absolute time and independent of generation interval, and (3) except for an initial period related to adjustment to chemostat conditions, the variance in the rate of increase in fitness is very nearly twice the mean.