Phenotypic diversity in E. coli populations evolving in glucose-limited chemostats. (a) The sympatric divergence of at least 39 isolates randomly picked from each population grown at slow-dilution rate (0.1 h-1; populations (Pop) 1 to 4) and fast-dilution rate (0.6 h-1; populations 5, 8, 10 and 11) were based on three characteristics: iodine staining of colonies, sensitivity to methyl α-glucoside (α-MG) and malG-lacZ expression level (fold- change in Miller units relative to ancestral strain). The individual properties of each test are listed in Table S1 in Additional file 1 and the criteria used for defining the phenotypic classes (PCs) are shown in Table S2 in Additional file 1. Each of 11 identified PCs 1 to 11 are marked with different colors. Numbers in the colored boxes indicate the frequency (percentage) of the isolate in the population. Boxes without colors indicate that the corresponding PCs are absent from the population. (b) Relationships of the 11 PCs were obtained by the neighbor-joining method rooted to the ancestral strain, BW2952 (black box white type), as described in PAST . The dendrograms were based on the three characteristics described in (a). Each terminal branch represents a phenotypic class (PC) and letters in parentheses indicate the individual phenotypic property belonging to the PC (+, wild-type iodine staining; -, no staining; P, partial staining; S, sensitive to α-MG; R, resistance to α-MG) and numbers 1 to 10 in parentheses indicate malG-lacZ expression level (fold changes in Miller units relative to ancestral strain). The bootstrap values at the branches are percentages based on 1,000 replicates. (c) The number of phenotypic classes in slow-dilution rate or 0.1 h-1 and fast-dilution rate or 0.6 h-1 populations. The phenotypic classes were based on (a), in which each colored box represents one PC. (d) The level of phenotypic diversity in 0.1 h-1 and 0.6 h-1 populations presented as the Simpson's diversity index (Di) of four replicate evolving populations from each dilution rate. The difference in Di was significant (two-tailed t-test, P = 0.033). The error bars are the standard deviation based on replicate populations.
Maharjan et al. Genome Biology 2012 13:R41 doi:10.1186/gb-2012-13-6-r41