Pole age affects cell size and the timing of cell division in Methylobacterium extorquens AM1

Abstract

A number of recent experiments at the single-cell level have shown that genetically identical bacteria that live in homogeneous environments often show a substantial degree of phenotypic variation between cells. Often, this variation is attributed to stochastic aspects of biology-the fact that many biological processes involve small numbers of molecules and are thus inherently variable. However, not all variation between cells needs to be stochastic in nature; one deterministic process that could be important for cell variability in some bacterial species is the age of the cell poles. Working with the alphaproteobacterium Methylobacterium extorquens, we monitored individuals in clonally growing populations over several divisions and determined the pole age, cell size, and interdivision intervals of individual cells. We observed the high levels of variation in cell size and the timing of cell division that have been reported before. A substantial fraction of this variation could be explained by each cell's pole age and the pole age of its mother: cell size increased with increasing pole age, and the interval between cell divisions decreased. A theoretical model predicted that populations governed by such processes will quickly reach a stable distribution of different age and size classes. These results show that the pole age distribution in bacterial populations can contribute substantially to cellular individuality. In addition, they raise questions about functional differences between cells of different ages and the coupling of cell division to cell size.

Fig. 1.

Representative lineage trees of M. extorquens (top) and E. coli (bottom). Each lineage tree is based on a time-lapse microscopy experiment. A single cell (represented at the bottom of each lineage tree) was placed on a pad of nutrient medium under the microscope and grown into a clonal population. Images were recorded every 20 min for M. extorquens and every 4 min for E. coli, and the lineage tree of the clonal population was reconstructed from these images. The lineage trees are ordered according to pole age. At each cell division (represented by branching events in the lineage tree), the branch that represents the cell with the new cell pole extends to the left, and the branch that represents the cell with the older pole extends to the right. (For the first cell division, at the bottom of each lineage tree, the old and new poles cannot be distinguished; these branches are thus dashed.) (A) Old- and new-pole cells of M. extorquens differ in the interdivision intervals. Branches that extend to the left tend to be longer than branches that extend to the right, indicating that new-pole cells take more time than old-pole cells until they initiate the next division. (B) In order to illustrate the pole age nomenclature, a section (section a1, framed by the dashed line) in the M. extorquens lineage tree is magnified. Each color in the magnified section denotes one individual: A, A′, and A″ are the same individual at increasing ages; B is the daughter of A. (See the main text for our definitions of “individual” and “daughter.”) Individuals at different ages are represented by rounded rectangles: the number on the right side in each rounded rectangle denotes the age of the older (age-defining) pole of this individual, and the number on the left side (in gray) denotes the age of the other pole, which is zero for all cells. (C) Magnification of section a2 illustrates cell pole age during cell division. Cells are labeled as in panels A and B. Individual B is flipped horizontally compared to panel B, in order to show the orientation of the cell poles as a cell emerges from division. (D) In E. coli, no systematic differences between old- and new-pole cells manifest in the lineage tree.

Fig. 2.

Cell size and the timing of cell division as a function of the age of a cell's pole. Information from eight independent lineage trees for M. extorquens (left; based on eight independent microcolonies with a total of 1,619 cell divisions) and three independent lineage trees for E. coli (right; based on three independent microcolonies with a total of 842 cell divisions) was separated according to the pole age of each cell. Each line depicts information for cells with poles of a particular age. The left end of each line depicts the moment when cells of a particular pole age are formed by division. The y-position of this end is the cell size after division, and the x-position is the time since division, which is zero. The right end depicts the moment just before a cell divides again. The y-position is the size prior to division, and the x-position is the time since the last division (i.e., the interdivision interval). The numbered arrows mark the following events: arrow 1, birth of a cell with a cell pole of age 1; arrow 2, division of a cell with a cell pole of age 1; arrow 3, formation of a cell with a cell pole of age 2. Error bars are standard errors (SE) of the mean, with lineage trees as units of replication. In M. extorquens, an individual's pole age affects its cell size and the interval between divisions. In E. coli, no such effects are discernible.

Fig. 3.

Detailed analysis of the sizes after and before division and the interdivisional intervals for M. extorquens. To account for possible (pole-age-unrelated) changes of cell size and interdivison intervals during the growth of a microcolony, we calculated for each cell how much these traits differ from all other cells that were present at the same stage of microcolony growth. These differences are “residuals,” and the corrected traits are “residual cell size after division,” “residual cell size before next division,” and “residual time to division.” The patterns that we observed in Fig. 2 are robust: cell size after division (A) and just before the next division (B) increases with increasing age of a cell's pole, and the interdivision interval decreases (C). These patterns are statistically significant: the letters above each column label homogeneous subsets, based on a Duncan post hoc test, with lineage trees as units of replication. Columns with different letters are statistically significantly different. Panels D to F focus on cells with a new cell pole (of age 1), and ask whether the cell pole ages of these cell's mothers have an effect on their properties. Cells that are produced by mothers with young poles are born smaller (D), and they take longer until they divide again (F); there are no consistent effects on their size prior to the next division (E).

Fig. 4.

Stable age distribution in M. extorquens populations, as predicted by a theoretical model. A population of 10,000 individuals of M. extorquens was simulated in an individual-based model. Within a time span corresponding to about 50 h, the fraction of individuals in different age classes reaches equilibrium. (A) Stable distribution of individuals in different classes, where each class is characterized by the pole age of cells (first horizontal axis) and by the pole age of the mothers of these cells, at the moment the cells were produced (second horizontal axis). Most individuals have a pole age of 1 and a mother of pole age 1. (B) Distribution of cell size after division in a population in stable age equilibrium. According to the results of the theoretical model, the fraction of cells in different size classes remains constant once a population reaches the age equilibrium.