I had the privilege of studying the ~3 µm bacterium Bacillus subtilis (pictured above) during my PhD. When I started graduate school, I wasn’t excited about studying bacteria. “You can’t even see them,” I thought. But my PI (principal investigator) quickly changed my perspective by showing me just how much of modern biology has been driven by the study of microorganisms.
As it turns out, bacteria are anything but boring. I had underestimated them. Over six years, I grew more and more to appreciate their incredible sophistication.
One thing about B. subtilis that came to fascinate me — and it’s one of the reasons this organism has been so well studied — is its superpower of crowdsourcing DNA under stress. This is called competence, and today we will be looking at some of the early steps and decision making preceding this cell fate.
Waves of Gene Expression
Competence is a programmed physiological state that Bacillus subtilis enters. It allows the organism to acquire extracellular DNA and, in some cases, integrate it into its genome through homologous recombination (Maier 2020). Competence can be thought of as a form of “crowdsourcing” where the target is genetic material that might replace one’s own inadequate resources. This is needed in an environmental situation where the bacterium’s core genetic resources fall short; such situations often are environmental extremes like high concentrations of antibiotics, drastic changes in pH or temperature. Competence is a well-studied program within B. subtilis where much of the molecular signaling leading to the state is understood. In order to run this program, the bacterium must turn on many genes and turn off many others. It does this through a series of waves of gene expression that pass through its gene-regulatory network.
Bacillus identifies the need to enter this adaptive state by integrating sensory inputs (signals bringing information from the environment) via transcription factors binding in the promoter regions of its genes (Alon 2019). There are four major families of network motifs: autoregulation, feed-forward loops, single input modules (SIMs), and dense overlapping regulons (DORs) (Alon 2019). The latter three can all be thought of as combinatorial decision-making tools where multiple inputs (input functions) are integrated to determine the output of a gene (Alon 2019). DORs form the foundation of the network structure and within DORs most of the other motifs can be found. In E. coli, another microorganism, for example, there are several large DORs where each DOR is united by a common function such as metabolism, stress response, or biosynthesis. In sensory transcription networks DORs are not layered, meaning the output of a DOR does not serve as an input for a different DOR. This means the majority of the computation in the network is done at the layer of promoters within the DOR. This description of how transcription networks function can provide a generic answer for how B. subtilis identifies problems leading to initiation of bacterial competence. But now we will turn to exploring the specific details. Hold on, there’s a lot of complexity!
First Step: A New Bus Route for RNA Polymerase
A prerequisite for competence is stationary phase growth (a semi-dormant state) in B. subtilis. Entry into stationary phase growth happens after nutrient depletion when complex regulatory circuits determine that rapid growth can no longer be sustained. This is in part initiated using the modular system of sigma factors which are exchangeable subunits of RNA polymerase that are necessary for transcription of subsets of genes (Paget 2015). There are at least 17 sigma factors in B. subtilis that are involved in a variety of cellular processes: housekeeping, exponential growth, chemotaxis, and sporulation. These can be swapped out like cassettes, and depending on the cassette, RNA polymerase will stop to transcribe different parts of the genome, resulting in different proteins being present in the cell. Sigma-H is the cassette that provides many of the gene regulatory changes necessary to move B. subtilis from exponential into stationary phase. Sigma-H sets the stage for initiation of genetic competence or spore formation (one program excludes the other) (Britton et al. 2002). We can consider activation of Sigma-H as the bacterium’s identification that rapid growth can no longer be sustained. If confirmed by additional inputs, this factor will move a bacterial cell towards either sporulation or competence.
Second Step: Lifting Repression of the “Crowdsourcing” Gene
ComK is the master regulator of the competence regulon in Bacillus subtilis. Under normal growth conditions, its gene (comK) is repressed. (Note: By standard biological convention, the lowercase italicized comK refers to the gene, while uppercase non-italicized ComK refers to the protein.)
The comK gene must be activated before the cell can produce the machinery required for genetic competence. Its promoter region is the key region where we expect physiological signals for and against competence to be integrated to control this gene’s expression
Three major transcription factors (CodY, Rok, and AbrB) act as repressors that keep comK turned off under normal conditions. Once a small amount of ComK protein is produced, it works together with the priming protein DegU in a positive feedback loop to further activate its own expression.
In short, CodY, Rok, and AbrB function like roadblocks preventing comK transcription. When these roadblocks are removed or overcome by specific environmental signals, ComK can bind to the promoter and drive the competence program forward.
In the next section, I’ll examine how these repressive roadblocks are lifted.
CodY: A Sensor for Nutrient Scarcity
CodY is a global transcriptional regulator that is used by bacteria to prioritize different metabolic pathways depending on metabolism or virulence needs. This global regulator specifically monitors branched chain amino acid status (isoleucine, leucine, and valine) and GTP levels (think: energy level). In times of plenty, CodY represses comK, but when levels of those amino acids or GTP drop, CodY releases from the comK promoter region and if other repressors are also lifted or inactivated the gene can be transcribed (Brinsmade 2017).
Rok: A Chromosome-Organization and “Crowdsourcing” Sensor
Rok, a nucleoid-associated protein (NAP), is involved in B. subtilis genome organization and is a potent repressor of comK, and the sporulation genes (Spo0A regulated genes) (Serrano et al. 2021). Current data doesn’t indicate that Rok is removed from the ComK promoter region. Rather, its repression is overcome by having more activators like ComK itself be present on the promoter. This means Rok is not a switch that is suddenly on. Instead, it is a brake that the regulatory network gradually overcomes.
Rok is also involved with linking the nucleoid architecture to natural chromosomal transformation efficiency (Serrano et al. 2021). For example, it has a role in controlling integration of xenogenic genes to prevent a deleterious effect on the host. Homology-mediated RecA-dependent DNA integration is the critical system that moves foreign DNA into the bacterial chromosome. However, chromosome organization that is controlled by NAPs, such as Rok, plays a key role in whether this system can work or not. For example, through Rok’s action of chromosome compaction, during xenogeneic DNA integration, certain dsDNA regions might be exposed more than others to RecA and homology finding. This affects how and where xenogenic DNA can be integrated. You can sort of think of Rok like a chromosome architect that controls where, when, and how often incoming DNA can recombine with the bacterial genome.
AbrB: A Sensor of Cell Division, Stress, and Genome Organization
AbrB is a global transcriptional regulator that prevents inappropriate gene expression when bacterial cells are actively growing and dividing (Chumsakul et al. 2011). It plays a very important role in transitioning the cell from active growth into the more dormant stationary phase, particularly reorganizing expression of more than a hundred genes (Chumsakul et al. 2011). The comK promoter AbrB acts as a repressor. The master controller of sporulation, Spo0A antagonizes AbrB thus lifting repression in the early stages of stress (Maier 2020). Interestingly, a consensus sequence for AbrB binding has not been recognized and instead it has been proposed that the protein requires a specific 3D conformation of the DNA helix. Several inputs seem involved in lifting this repressor from the ComK promoter (even though we don’t even really know how it binds): cell division status, genome organization, and an early stage signal of stress (this is what Spo0A begins as).
DegU: A Priming Protein that Promotes “Crowdsourcing”
ComK also acts as an activator on its own promoter (positive autoregulation) which helps commit cells to the crowdsourcing or competence pathway. Interestingly, though, to do this ComK requires a priming protein DegU to help it bind to its promoter. DegU controls four behaviors that all involve social or population level behavior: inhibition of flagellar based motility, activation of degradative enzyme production, enhancing biofilm formation, and activation of competence (Verhamme et al. 2007). DegU can be phosphorylated, but it is the unphosphorylated state of DegU that activates competence (Verhamme et al. 2007). It has been proposed that as DegU moves up its phosphorylation gradient (more and more of it becomes phosphorylated) that it activates genetic competence, swarming, complex colony development, and exoproteases, in that order (Verhamme et al. 2007). This means that a very low level of DegU~P activates competence while a high level of DegU~P will inhibit competence. The amount of DegU~P is governed by a DegS-DegU two-component system that is also fine-tuned by several modulators.
The Gene Regulatory Network for Crowdsourcing
The gene regulatory network for “crowdsourcing” has at least four regulatory elements upstream of the master regulator ComK. Three are repressors, indicating that under many instances ComK should be off.
The dominant feature of this gene regulatory motif is positive feedback. However, this is combined with multiple parallel repressors and co-activation by DegU. The strong positive feedback (also called positive autoregulation since ComK activates its own promoter) is the key driver of bistability in competence development, where bistability is the reality that low basal expression can flip to high expression in a subset of cells once a threshold is crossed.
If an upstream factor (e.g., nutrient signals via CodY) causes the presence of a repressor on the comK operon while DegU/ComK tries to activate the operon, this can resemble an incoherent motif (opposing direct/indirect effects). This can generate pulses, accelerate responses, or filter noise, similar to Incoherent type 1-FFLs in other systems (Alon 2019). In short, this control system upstream of comK does a number of sophisticated things including creating stochastic induction of competence for a subset of cells and filtering out noise.
To summarize, critical input data determining whether a B. subtilis bacterium will take the next steps toward committing to competence seems to be nutrient availability, genome organization, differentiation between the need to sporulate and crowdsource DNA, cell division status, and population behavior. These inputs are computed at the comK promoter based upon its architecture. Every part of this gene regulatory system has a logical purpose and critical role in controlling the DNA crowdsourcing decision. Nothing here looks cobbled together or haphazard. Rather this small part of the gene regulatory network interfaces seamlessly with the greater network and neatly controls the master regulator of “DNA crowdsourcing” so that the bacterium only does this when it is necessary and prepared.
These are the same type of understandable and logical computation circuits that humans design. What can we learn from observing this similarity? If you are brave enough to venture into thinking about origins, you can use inference to the best explanation to consider whether a superintellect might just be the best explanation for the design of the gene regulatory motif upstream of comK.