The term "non-genetic individuality" is applied to organisms from a genetically identical population that display differences in phenotype from individual to individual. Among many instances of this phenomenon in both prokaryotic and eukaryotic organisms is the individuality in the swimming behaviour of E. coli, which was first observed a quarter of a century ago [Spudich and Koshland (1976) Nature 262:467-71]. One explanation for this property is that it could arise from permanent differences in the number of protein molecules from cell to cell due to stochastic gene expression. BCT provides the means for using this proposed mechanism to simulate individuality in swimming behaviour. The population in each simulation consisted of 10,000 individuals, in each of which the concentrations of the chemotaxis proteins were chosen from independent normal distributions with constant standard deviations relative to the mean wild-type concentration. Figures reproduced from Levin et al. (1998).

 

CheYp distributions

CheYp is the signalling molecule involved in switching events at the flagellar motor. As the relative standard deviation of protein concention increases, the CheYp distribution becomes broader and more skewed. The intracellular CheYp concentration can be related to the motor bias through the Hill equation

    bias = 1 - [CheYp]5.5 / ((17/3)*Set_Yp5.5 + [CheYp]5.5)

where Set_Yp is the mean concentration of CheYp in a wild-type population, so that when [CheYp] = Set_Yp, bias = 0.85.

 

Bias distributions

The bias distributions are more skewed than the CheYp distributions from which they were derived. Note that the distributions have different modal biases but very similar mean biases:

Standard deviation of protein concentration (% of mean) Bias distribution
Mean Standard deviation
5 0.85 0.07
10 0.84 0.14
15 0.83 0.20
20 0.82 0.24

 

Comparison with experimental data

Bob Bourret and Carl Morton-Firth measured the biases of a population of 500 wild-type E. coli bacteria by tethering their shorn flagella to glass coverslips. They recorded the movement of each cell on videotape and manually logged the changes in direction of rotation on playback. The bias was calculated as the fraction of time the cell rotated in an anti-clockwise direction. The simulated bias distribution with a standard deviation of protein concentration of 10% of the mean is in good agreement with the experimental distribution:

Using the above Hill equation we can deduce the CheYp concentration in a cell from its measured bias, and again compare the simulated and the experimental distributions: