Localized Production of Cell Wall Precursors May Be Critical for Regulating the Mycobacterial Cell Wall

ABSTRACT

The paper “Cell wall damage reveals spatial flexibility in peptidoglycan synthesis and a nonredundant role for RodA in mycobacteria” by Melzer et al. (E. S. Melzer, T. Kado, A. Garcia-Heredia, K. R. Gupta, et al., J Bacteriol 204:e00540-21, 2022, https://doi.org/10.1128/JB.00540-21) presents several new observations about the localization and function of cell wall enzymes in Mycobacterium smegmatis and their responses to stress. This work illustrates some important aspects of cell wall physiology in mycobacteria and also points to a new model for how peptidoglycan synthesis may be organized in pole-growing bacteria.

TEXT

In their article, Melzer et al. (1) show that the peptidoglycan transglycosylase RodA localizes evenly around the Mycobacterium smegmatis cell, as does the essential class A penicillin-binding protein (PBP) PonA1 (2, 3). Despite their localization, both enzymes contribute similarly to polar peptidoglycan insertion in log phase, though RodA seems to have increased importance under stress. In all the well-studied lateral growers (Bacillus subtilis, Caulobacter crescentus, Escherichia coli), cell wall elongation is mediated by a large protein complex that spans the inner membrane and includes cytoplasmic regulators such as MreB and the peptidoglycan synthesis enzymes RodA and class B PBP2 (bPBP2). The localization of RodA and PBP2 to the Rod complex is dependent on MreB (4, 5). Until the work by Melzer et al, several papers had indicated that a similar model of localization of transglycosylases to the peptidoglycan insertion point was likely to prevail in the pole growers as well. In Corynebacterium glutamicum, RodA’s localization is restricted to the pole through an interaction with the pole-associated DivIVA protein (6, 7). The Rhizobiales lack RodA and use PBP1a as the key elongative transglycosylase (8); this protein also localizes to the growing pole (9). Now we can see that localization of transglycosylases to the site of cell wall expansion is not essential for polar growth. How, then, is peptidoglycan insertion restricted to the poles when the enzymes that catalyze it are everywhere?

Previous work has shown that M. smegmatis has a biochemically distinct domain of the inner membrane, called the IMD, that localizes largely near the cell poles and is the site where many cell wall precursors enzymes are localized (10–13), including MurG (3, 14). These studies, in light of the new data showing that the RodA and PBP1 are delocalized, indicate that localized production of cell wall substrates may be the critical factor for restricting cell wall synthesis to the poles. To determine if this is true, we would need to be able to observe regulation of this system. Importantly, Melzer et al. (1) show us that cell wall damage causes peptidoglycan synthesis to relocalize from the poles to the lateral walls, and, supporting the substrate regulation model, lipid II synthesis enzyme MurG relocalizes similarly. This substrate model of regulation is also supported by work showing that other cell wall precursor synthesis/IMD proteins relocalize along the lateral walls during starvation and d-cycloserine treatment (11).

The idea that lipid II is a critical regulatory node in cell wall synthesis is not novel. In E. coli and C. crescentus, MurG also associates with the Rod complex (15, 16). Thus, lipid II is produced in the same protein complex in which it is consumed by RodA. B. subtilis has distinct lipid domains that are arranged in circumferential arcs like the Rod complex and which presumably include lipid II, as they are dependent on MurG expression (17). Lipid II, in turn, is required for the proper membrane localization of MreB (18) and PBPs in several Gram-positive organisms (19, 20, 21). Thus, in many species, lipid II is spatially linked to the localization of the transglycosylases, either through the interaction of MurG (E. coli and C. crescentus) or lipid II itself (B. subtilis) with the Rod complex. The work of Melzer et al. indicates that MurG and lipid II are likely to be critical regulators in mycobacteria as well, but they appear to control the activity of the transglycosylases without controlling their localization.

Melzer et al. provide a new framework to approach the question of how rod shape is maintained in pole-growing bacteria. This study illustrates that the localization of the peptidoglycan transglycosylases is perhaps not so important for maintaining rod shape. Studies in Bacillus subtilis (22) and E. coli (23) indicate that perhaps the fundamental problem to solve, for the maintenance of rod shape, is the orientation of the movement of peptidoglycan transglycosylases, not their localization. It appears increasingly clear that the glycan chains of peptidoglycan must be arranged circumferentially around the cell in order to withstand osmotic pressure and push cell wall expansion to occur such that cells elongate and do not widen (24–27). There is also evidence of circumferential cell wall construction in mycobacteria (28, 29). In the laterally growing model rods, the transglycosylases are directed to move circumferentially by the dynamic cytoskeletal protein MreB, which moves with the RodA synthesis complex (27). Pole growers lack MreB. While the polar protein DivIVA/Wag31 is often invoked as a functional analog of MreB, there is no evidence that it helps orient transglycosylases circumferentially. If DivIVA does perform this function, it would have to do so without being mobile, like MreB is; DivIVA in Actinobacteria has been shown to be largely immobile (30).

This leaves us with three possibilities that I can currently think of: (i) DivIVA orders circumferential peptidoglycan synthesis while remaining stationary itself, (ii) there is an unknown mobile functional homolog of MreB that directs circumferential peptidoglycan synthesis, and (iii) circumferential peptidoglycan synthesis is enforced through the maintenance of a thin ring of substrate availability around the cell pole.

With respect to the first possibility, if DivIVA proteins function to orient peptidoglycan synthesis while remaining stationary, the DivIVA homo-oligomeric network would need to have directional orientation and be arranged circumferentially around the cell. Perhaps a transmembrane protein that interacts with the peptidoglycan synthesis complex could slide along an organized track of the DivIVA network to keep synthesis oriented. Rhizobiales, which lack DivIVA, have a protein, GPR or RgsE, which forms a multimeric ring along the inner membrane that circles the cell near the growth pole (31, 32). GPR, like DivIVA in mycobacteria (33), is required to restrict peptidoglycan synthesis near the poles (34), and its positioning might allow it to serve as a track for circumferential peptidoglycan synthases to move along.

As there are no data to support or exclude the possibility of an unknown functional homolog of MreB, I will not discuss that model further.

The third possibility comes out of the work done by Melzer et al. (1) and other studies showing that cell wall precursors are synthesized primarily near the pole in the IMD (3, 10, 11). Because of the higher concentration of lipid II near the poles, peptidoglycan synthesis would most frequently begin near that region. If there are no cytoskeletal factors to direct peptidoglycan synthase complexes to move circumferentially, perhaps they can initiate synthesis in any direction. If synthesis initiates along the short axis of the cell, then substrate would always be abundant, allowing rapid circumferential synthesis. If synthesis occurs along the major cell axis, the synthase complex would soon leave the polar region and stall due to lack of substrates. This would cause either strand termination or a pause in synthesis during which the synthase complex could, while still tethered to the synthesized glycan, diffuse back toward the IMD, where it would be more likely to encounter more lipid II to continue synthesis; this could reorient synthesis toward the short axis. In this model, the transglycosylases might need to be highly processive to allow synthetic complexes to diffuse back toward the IMD without terminating the strand. Alternatively, if lack of substrate causes strand termination, this model would result in shorter strands that are oriented along the long axis of the cell, and longer strands would be oriented circumferentially. This model is indifferent to the transglycosylase: all enzymes are stalled by lack of substrate. This model may also be viable in the Rhizobiales: no IMD-like lipid domain has been described as of yet, but the GPR/RsgE protein contains several apolipoprotein domains, which have been shown to organize lipids in other species (34). Methods to manipulate the localization of MurG/J and time-lapse microscopy of single peptidoglycan synthesis complexes will be required to test this model.

In summary, this work by Melzer et al. indicates that the location of cell wall precursors is likely to be highly regulated in mycobacteria and to be more important than the localization of the peptidoglycan synthesis enzymes for organized polar cell wall expansion. The distributed localization of the key peptidoglycan transglycosylases seems to prepare the cell for lateral reinforcement of the wall during cell wall damage.

The views expressed in this article do not necessarily reflect the views of the journal or of ASM.