Analysis of noise mechanisms in cell size control

Abstract

At the single-cell level, noise features in multiple ways through the inherent stochasticity of biomolecular processes, random partitioning of resources at division, and fluctuations in cellular growth rates. How these diverse noise mechanisms combine to drive variations in cell size within an isoclonal population is not well understood. To address this problem, we systematically investigate the contributions of different noise sources in well-known paradigms of cell-size control, such as the adder (division occurs after adding a fixed size from birth) and the sizer (division occurs upon reaching a size threshold). Analysis reveals that variance in cell size is most sensitive to errors in partitioning of volume among daughter cells, and not surprisingly, this process is well regulated among microbes. Moreover, depending on the dominant noise mechanism, different size control strategies (or a combination of them) provide efficient buffering of intercellular size variations. We further explore mixer models of size control, where a timer phase precedes/follows an adder, as has been proposed in Caulobacter crescentus. While mixing a timer with an adder can sometimes attenuate size variations, it invariably leads to higher-order moments growing unboundedly over time. This results in the cell size following a power-law distribution with an exponent that is inversely dependent on the noise in the timer phase. Consistent with theory, we find evidence of power-law statistics in the tail of C. crescentus cell-size distribution, but there is a huge discrepancy in the power-law exponent as estimated from data and theory. However, the discrepancy is removed after data reveals that the size added by individual newborns from birth to division itself exhibits power-law statistics. Taken together, this study provides key insights into the role of noise mechanisms in size homeostasis, and suggests an inextricable link between timer-based models of size control and heavy-tailed cell size distributions.

Mycobacteria Modify Their Cell Size Control under Sub-Optimal Carbon Sources

“…This is coupled with the observation that growth rate is dependent on birth length (Figure 2D), with smaller cells growing faster than larger cells at birth (for cells grown in glycerol but not for cells grown in acetate or pyruvate; Figures 5B, 6B); a dependence that is not observed in E. coli . How these correlations affect the properties of cell size control is not clear and should be the subject of further modeling studies (Modi et al, 2016), though we note that the striking negative dependence of growth rate on birth length has recently been observed in mammalian cells (Ginzberg et al, 2017). …”

Mitochondrial heterogeneity, metabolic scaling and cell death

“…2C), adder dynamics imposes a characteristic cell size associated with the mode of the distribution, which is often referred to as an 'optimal' cell size. The width of the final distribution is set by the variability in partitioning of volume between daughters at division (although other sources of variability may be envisaged such as noise in the growth rate or the amount of volume added per cell cycle [54], which are neglected here). Under these simulation dynamics, the modal cell volume and the maximum value of rDC are essentially independent ( Fig.…”

Stochasticity in bacterial division control: Preliminary consequences for protein concentration

“…Cell division is suspected to explain part of this fluctuations in gene expression (Amir 2014, Modi, et al 2017. In rod shaped organisms, physiological implications can potentially affect DNA concentration, surface transport and biosynthesis rates, as well as proteome composition (Willis and Huang 2017).…”