As mentioned before, potassium homeostasis is an essential function of c-di-AMP inB. subtilis. Cells lacking the second messenger are not able to inhibit potassium import. Consequently, subjection to high (5 mm) instead of low (0.1 mm) potassium concentrations leads to toxic accumulation of the ion and cell lysis, presumably caused by concomitant water influx (Gundlachet al., 2017b). In the aforementioned studies (see Section 3.8), we also analyzed growth of the B. subtilis DAC (GP2222) deletion mutant in a range of different media, including MSSM medium with ammonium and elevated potassium concentrations (5 mm and higher). Surprisingly, the c-di-AMP-free ∆DAC strain was able to grow with very high (250 mm) amounts of potassium in MSSM medium with ammonium as the nitrogen source
but not with 5 mm.
To verify and refine our initial observation, we investigated which potassium concentrations inhibit growth of the ∆DAC strain on MSSM medium with ammonium as the nitrogen source in a drop dilution experiment. The ∆DAC strain (GP2222) was first grown in MSSM medium with ammonium as the nitrogen source and low (0.1 mm) amounts of potassium and the wild type (168) was carried along as a control. Cells were washed extensively and samples of a serial dilution were spotted on MSSM medium agar plates with ammonium as the nitrogen source and the indicated potassium concentration (0.1 to 500 mm). Growth was documented after 48 h at 37◦C. Although elevated amounts of potassium were thought to be toxic for a c-di-AMP-free B. subtilis strain, the results shown in Figure 3.16 refine this postulation.
Surprisingly, increasing the potassium concentration to 100 mmand more stabilized growth of the ∆DAC strain. This is particularly interesting since the results are counterintuitive at first glance. Subjection of the ∆DAC strain to 5 mmKCl already resulted in toxic accumulation of K+ and cell lysis, as documented before (Gundlachet al., 2017b), but a 20- to 50-fold increase allows for growth again.
KCl [mM]
WT ∆DAC
0.1 5 50 100 150 250 500
Figure 3.16: Elevated potassium amounts stabilize growth of a c-di-AMP-free strain again.
Growth ofB. subtiliswild type (WT, 168) and of a c-di-AMP-free deletion mutant devoid ofcdaA,disAand cdaS (∆DAC, GP2222) on MSSM medium agar plates with ammonium as the nitrogen source and indicated KCl concentration after 48 h at 37◦C. Numbers on top indicate the spotted dilution of an OD600 of 1.0.
Growth at 500 mm KCl was slightly worse than growth at 250 mmfor both the wild type and the ∆DAC strain, most likely because the hyperosmotic condition leads to dehydration of the cells (Hoffmann and Bremer, 2017). It should be noted that unstressedB. subtilis cells accumulate potassium to around 300 mm (Whatmoreet al., 1990). The results indicate that, although c-di-AMP cannot inhibit K+ uptake anymore, increasingly abundant potassium amounts prevent the build-up of a huge K+gradient which would then lead to increased water influx and cell lysis in the ∆DAC strain.
The second messenger c-di-AMP is essential for many bacteria that produce it and becomes dispensable only under very specific growth conditions or by accumulation of suppressor mutations (Commichauet al., 2018b). DarA is a prominent c-di-AMP receptor inB. subtilis and is highly conserved among Gram-positive, c-di-AMP-producing firmicutes. Similar to other c-di-AMP targets, DarA is not essential (Gundlach et al., 2015a). In the past, our and other groups tried to elucidate the function of DarA in B. subtilis and its homologue PstA inL. monocytogenes and S. aureus. Despite extensive research the function of DarA has remained enigmatic. In this thesis we showed that DarA is acting on a cytosolic target.
Specific phenotypes of adarAdeletion mutant were shown, several hypotheses could be ruled out and DarA was linked to glutamate homeostasis. This is especially interesting since the homeostases of c-di-AMP, potassium and glutamate seem to be intricately linked to each other (Gundlachet al., 2018). So far, no target of c-di-AMP has been reported to be involved in the homeostasis of glutamate.
4.1 No apparent functional link within the darA operon
DarA is conserved among Gram-positive, c-di-AMP-producing firmicutes with only few exceptions (Gundlachet al., 2015a). The genomic arrangement of DarA with the essential genes tmk (upstream, coding for the thymidylate kinase) and holB (downstream, coding for theδ0 subunit of the DNA polymerase III) is also highly conserved among bacteria that express darA hinting to a functional link (Dandekar et al., 1998; Nicolas et al., 2012). In addition, in almost allBacillus species a gene of unknown function (yaaR) is located directly betweendarAandholB. However, the presence ofyaaRis not conserved among the firmicutes, suggesting that there is no functional link between DarA and YaaR (Dandekaret al., 1998;
Nicolaset al., 2012).
Conserved gene pairs or operons can often be attributed to ribosomal genes or genes involved in other fundamental cellular processes like cell division or cell wall synthesis and those likely originate from a common ancestral genome. In addition, many proteins encoded by conserved gene pairs appear to physically interact with each other. This allows for the prediction of functional linkages and protein–protein interactions (Dandekar et al., 1998;
Thomaset al., 2000; Watanabe et al., 1997). In fact, there are also examples of conserved genomic linkages and interactions of the gene products among PII proteins. For instance, the genes for the PIIprotein GlnK and the ammonium importer AmtB are encoded in a conserved operon in almost all bacteria that produce them. The physical interaction of GlnK with AmtB has been structurally resolved for E. coli and verified experimentally in different bacteria includingB. subtilis (Conroy et al., 2007; Coutts et al., 2002; Detsch and Stülke, 2003).
Stimulated by the genetic linkage oftmk,darA andholB, earlier studies tried to prove physical interactions between the three. This has been especially interesting as DarA itself is not an essential target of the essential second messenger c-di-AMP but of course its interaction partner might as well be (Gundlach et al., 2015b). Intriguingly, no interaction of DarA with HolB or Tmk could be verified. This work is in agreement with previous studies and accounted for c-di-AMP-bound and apo-DarA which was not always done before (Hach, 2015; Jäger, 2015). No specific interaction of DarA with Tmk or HolB could be detected in a variety of different experimental setups (see Section 3.1). These results, together with the preceding work, strongly indicate that there is no interaction present.
This raises the question: Why is there such a conserved genomic arrangement at all?
One reason might be that this conservation does not reflect a physical interaction but rather an indirect functional linkage of the gene products as in the case of GlnB, the founding member of the PIIprotein family. GlnB indirectly regulates the glutamine synthetase (GS) GlnA, thus nitrogen assimilation, on a transcriptional and post-transcriptional level in many proteobacteria, andglnAandglnBare genetically linked in many diazotrophicα-proteobacteria.
It should be noted that this is not the case in the γ-proteobacterial model organism E. coli, although the GS is still regulated on a transcriptional and post-transcriptional level by GlnB (Arcondéguy et al., 2001; Huergo et al., 2013; Leigh and Dodsworth, 2007). The genomic conservation of thedarAoperon might point to a functional relation of the encoded proteins without physical interaction. However, there are no experimental indications for this and our results associate DarA with glutamate metabolism.
There is one report of PII proteins from the α-proteobacterial Rhodobacter capsulatus where a yeast two-hybrid experiment showed an interaction of GlnB and GlnK with PcrA (helicase superfamily 1) and Era (Ras-like GTPase), respectively. This is quite unusual and would be the first time that a PII protein regulates the unwinding of the DNA (PcrA). The other protein, Era, is involved in several processes, among them ribosome assembly, energy metabolism and cell cycle regulation, and would be an unusual target for a PII protein as well (Pawlowskiet al., 2003). However, yeast two-hybrid analysis is prone to false positive indications and validation by other approaches is still missing (Brückneret al., 2009).