Bacillus subtilis is a bacterium with many talents. Some cells devote themselves to exploration, abandoning their relatives to live as swimming ‘motile’ cells. Others instead choose a cooperative, community-oriented lifestyle by settling down with their neighbors to form connected ‘chains’ of cells. In a previous study1, we adapted a device2 that allowed us to follow cells over hundreds of generations of growth, so that we could learn about how cells split their time between the two states. We found that the motile cells decide to switch lifestyles in a completely random fashion—some ‘family trees’ would remain motile for only a few generations, while others remained motile for hundreds. In contrast, cells were very picky about how long they grew as chains, and precisely timed out almost exactly eight generations before becoming motile again. This is a remarkable trick for a bacterium to execute; motile cells and chain cells share all of the same genes, yet have completely different properties and can retain their states for many generations of growth. In our new study, we set out to understand the genetic mechanism that allows cells to choose between the two lifestyles.
By building mathematical models that describe the key genes involved in the process, we stumbled on an unexpected explanation: the cells decide between states by playing a molecular version of Scrabble. In our model, two proteins— which we think of as vowels and consonants—are produced at steady rates, and are destroyed sticking together to form complexes (words). When the proteins rapidly form complexes, this leads to an interesting effect: whichever protein happens to be more abundant at a given time will consume the entire supply of the other. This effect explains a frustration familiar to any serious scrabble player: at one point or another, your hand consists entirely of vowels or consonants. The bacteria use this phenomenon to make lifestyle decisions. When ‘vowels’ outnumber ‘consonants’, the bacteria is motile, and when ‘consonants’ outnumber ‘vowels’ it is in a chain.
This model turned out to have substantial explanatory power. Not only did it capture the existence of two distinct states, but it also predicted the precise timing behavior of the lifestyles—random switching for the motile state, and clock-like timing of the chain state. To prove that this mechanism could actually work, we re-built the ‘Scrabble’ system in a distantly-related bacterium that lacks the two lifestyles. This rebuilt system generated the key features of the Bacillus switch with remarkable fidelity.
Understanding how genetically identical cells select and retain different lifestyles is a long-standing challenge in developmental biology. This ‘cell fate’ selection process is generally thought to require intricately-wired genetic circuits which often defy mechanistic understanding. The mechanism we propose is remarkable because of its simplicity. The major requirement for it to work is two proteins that enthusiastically stick to one another, a surpassingly common feature of many genetic circuits. We are excited to see if this mechanism similarly explains other cases of cell fate selection.
2 Wang, P. et al. Robust growth of Escherichia coli. Current biology 20, 1099-1103 (2010).