CdaA regulation

Even though the presence of c-di-AMP has been described to be essential for some bacteria, an overproduction of c-di-AMP was demonstrated to be harmful for the cell. Hence, c-di-AMP was named the “essential poison” which requires a tight control of its synthesis (Gundlach et al. 2015a; Huynh and Woodward 2016; Blötz et al. 2017).

CdaA is present in a conserved gene cluster which encodes besides cdaA the regulatory protein CdaR and the glucosamine mutase GlmM (Mehne et al. 2013; Rismondo et al. 2016; Zhu et al.

2016). Therefore, a functional relation between these three proteins was suggested. Indeed, a direct interaction between CdaA and CdaR as well as modulation of the CdaA activity were reported previously (Mehne et al. 2013; Gundlach et al. 2015a; Rismondo et al. 2016). In silico experiments suggested a positioning of CdaR extracellular (Corrigan Rebecca M and

Gründling 2013). This was confirmed in Chapter 3. Here it was also shown that CdaR interacts through the transmembrane domain with CdaA suggesting that the CdaA TM importantly con-tributes to the activity of the DAC domain and also the transfer of a sensed signal by CdaR.

The importance of the membrane domain was also demonstrated by mutations resulting in a decrease of CdaA activity (Zhu et al. 2016). It is suggested that CdaR is a signal receptor sens-ing an external signal. So far, a signal sensed by CdaR and the signal transduction through the interacting TM domains of CdaR and CdaA is not known. In order to shed light on this un-solved problem it is of great interest to identify the signal which is transduced to the DAC domain. Crystallographic and biochemical experiments might also help to further understand the mode of interaction between CdaR and CdaA and how the activity of DAC domain is mod-ulated.

The second protein encoded in the conserved gene cluster is GlmM. A direct interaction be-tween CdaA and GlmM has been reported in B. subtilis, L. lactis and S. aureus (Gundlach et al. 2015a; Zhu et al. 2016; Tosi et al. 2019). In Chapter 3 an interaction between GlmM and CdaA was also confirmed in vivo and in vitro for L. monocytogenes. Furthermore, it was demonstrated that GlmM inhibits the c-di-AMP synthesis of CdaA under hyperosmotic stress preventing the cell from uncontrolled water loss through e.g. carnitine and betaine uptake (Zhu et al. 2016, Chapter 3). Zhu and colleagues discovered in L. lactis under hyperosmotic stress conditions a point mutation in GlmM at position 154 (I¹⁵⁴ to F) leading to an osmoresistent strain. This strain showed a reduced c-di-AMP level in comparison to the wild type GlmM strain, suggesting an interaction between CdaA and GlmM. Sequence alignments of GlmM from related bacteria unveiled a phenylalanine in S. aureus, L. monocytogenes and B. subtilis at the position of an isoleucine (I¹⁵⁴) in L. lactis wild type. As the L. lactis mutant strain carries also a Phe, it was suggested that this amino acid might play an important role in modulating CdaA and therefore might be involved in CdaA and GlmM interaction. Indeed, mutations of the phenylalanine to an isoleucine showed a similar effect in L. monocytogenes as the wildtype GlmM in L. lactis. Interestingly, the phenylalanine (Phe¹⁵⁵) in the available S. aureus GlmM (PDB: 6GYZ) structure is exposed on the protein surface and could indeed be crucial for pro-tein-protein interactions (Tosi et al. 2019).

How could an interaction of GlmM with CdaA result in altering cyclase activity?

The crystal structure of GlmM reveals, as suggested from biochemical data, a homodimer as a biologically active form. This oligomerization of two GlmM monomers results in a V-shaped structure, right above the dimeric interface is the previously discussed Phe¹⁵⁵ exposed to the surface (Fig. 9). So far, no crystallographic data of a CdaA and GlmM complex are available.

However, Tosi et al. performed SAXS experiments in which the ab initio SAXS molecular envelope suggested that, at least in solution, CdaA sits on top of the GlmM homodimer inter-face. It was suggested that GlmM blocks the formation of higher CdaA oligomers which might be important for activity (described in 6.2). Nonetheless, these results seem to be inconsistent with the fact that only a single mutation (in L. monocytogenes Phe¹⁵⁴ to Ile) leads to an increase

in the intracellular c-di-AMP level. The question to address is whether a single amino acid mutation in GlmM could have such a strong impact on the GlmM-CdaA complex formation and is able to perturbate the inhibition of the suggested higher CdaA oligomers. So far, no data are available describing the influence of the single mutation on the GlmM-CdaA binding af-finity. In order to argue about the mode of GlmM-CdaA interaction additional experimental data would be required. Even though we know that GlmM inhibits CdaA as a reaction of hy-perosmotic stress, the signal leading to a complex formation and thereby inhibition of c-di-AMP synthesis still needs to be elucidated.

Structural characterization of the GlmM-CdaA complex would be of great interest in order to understand CdaA inhibition on a structural level which in turn could be useful for a CdaA drug discovery campaign.

Figure 9: Crystal structure of the Glucosamine mutase GlmM from S. aureus. (A) GlmM homodimer (mon-omer 1: dark green, mon(mon-omer 2: light green) with a V-shaped dimeric interface in which the two Phe¹⁵⁵ are ex-posed on the protein interface. The protein is depicted in a cartoon mode in light and dark green. The two phenyl-alanine are shown in stick mode (carbon in light and dark green, oxygen in red and nitrogen in blue). These two Phe¹⁵⁵ were shown to negatively influence CdaA activity and its position is indicated by the two red stars. (B) A detailed view of the two Phe¹⁵⁵ exposed on the protein surface (PDB code: 6GYZ) (Tosi et al. 2019).