ylm Has More than a (Z Anchor) Ring to It!

The division and cell wall (dcw) cluster is a highly conserved region of the bacterial genome consisting of genes that encode several cell division and cell wall synthesis factors, including the central division protein FtsZ. The region immediately downstream of ftsZ encodes the ylm genes and is conserved across diverse lineages of Gram-positive bacteria and Cyanobacteria. In some organisms, this region remains part of the dcw cluster, but in others, it appears as an independent operon.

ABSTRACT

The division and cell wall (dcw) cluster is a highly conserved region of the bacterial genome consisting of genes that encode several cell division and cell wall synthesis factors, including the central division protein FtsZ. The region immediately downstream of ftsZ encodes the ylm genes and is conserved across diverse lineages of Gram-positive bacteria and Cyanobacteria. In some organisms, this region remains part of the dcw cluster, but in others, it appears as an independent operon. A well-studied protein coded from this region is the positive FtsZ regulator SepF (YlmF), which anchors FtsZ to the membrane. Recent developments have shed light on the importance of SepF in a range of species. Additionally, new studies are highlighting the importance of the other conserved genes in this neighborhood. In this minireview, we aim to bring together the current research linking the ylm region to cell division and highlight further questions surrounding these conserved genes.

The dcw cluster.

Cell division is an essential process through which all living things seek to reproduce. In bacteria, this is primarily carried out through binary fission, a mechanism where one parental cell typically produces two identical daughter cells (1–3). To facilitate this process, a large protein complex, known as the divisome, forms at mid-cell and enables septum formation and cytokinesis (3). Central to the divisome is the tubulin homolog FtsZ. Regulation of FtsZ ring (Z-ring) formation is critical for division to occur properly (4). Much of what we know about the regulation of FtsZ and more generally cellular division is through studies of the rod-shaped Gram-negative and Gram-positive model organisms Escherichia coli and Bacillus subtilis, respectively (1, 4). However, there are still many gaps in our knowledge of this process in these classic model organisms and, importantly, in nontraditional model organisms, including pathogens such as Staphylococcus aureus (2).

A conserved region of the bacterial chromosome that has been well studied in the division process is the division and cell wall (dcw) cluster. It was first identified around 30 years ago and noted to encode many important cell wall- and divisome-associated proteins (5, 6). Additionally, the dcw cluster is found in diverse lineages of bacteria. Although the contents of the cluster vary between species, the structure of this region is almost always conserved (6, 7); at the 5′ end are the genes associated with cell wall synthesis, and located close to the 3′ end is ftsZ (6–8).

ylm operon.

While the conservation of the dcw region is consistent across both Gram-positive and Gram-negative bacteria, the arrangement of the region downstream of ftsZ, named the ylm operon, appears to be exclusively conserved among Gram-positive organisms (7, 9). The ylm operon is found immediately downstream of ftsZ in many species and about 10 kb downstream in B. subtilis. Carrying between five and seven genes, the ylm region is found as an operon in at least B. subtilis, S. aureus, and Streptococcus pneumoniae (9–11) (Fig. 1A). In some species, such as Streptomyces, this operon is transcribed from the ftsAZ promoter, while in other species such as B. subtilis, it is transcribed from its own promoter (10, 12). This operon typically consists of ylmD, ylmE, ylmF (sepF), ylmG, ylmH, and divIVA. DivIVA is a conserved multifunctional protein important in regulating Z-ring placement and chromosome segregation in a range of Gram-positive bacteria. DivIVA has been extensively studied previously; see recent review articles for more information on DivIVA (13, 14). As mentioned above, the ylm operon is highly conserved among Gram-positive bacteria, but there is some variability in the genes that are included. The most common change is the loss of either ylmD or ylmH (Fig. 1A and Table 1).

FIG 1.

Overview of the ylm locus conservation, interaction network, and SepF domain architecture. (A) Cartoon representation of the ylm region in Firmicutes and Actinobacteria. (B) Interaction network of SepF and other Ylm proteins. Highlighted in red are Ylm proteins. The central SepF and FtsZ interaction has been confirmed in all species displayed except Listeria monocytogenes (L. mono.) and Synechocystis. A full list of tested interactions and the methods utilized can be found in Table S1 in the supplemental material. (C) Domain architecture of SepF.

TABLE 1.

Conservation of ylm genes^a

Species	ylmD	ylmE	ylmF (sepF)	ylmG	ylmH	sepF essentiality	Z-ring anchor(s)	Operon structure
B. subtilis	X	X	X	X	X	Nonessential	FtsA, EzrA	X
S. aureus	X	X	X	X	X	Unclear	FtsA, EzrA	X
S. pneumoniae		X	X	X	X	Nonessential	FtsA, EzrA	X
E. faecalis		X	X	X	X	Nonessential	FtsA, EzrA	X
Mycobacterium	X	X	X	X		Essential		X
Streptomyces	X	X	X	X		Unclear	SsgA/Bb	X
C. glutamicum	X	X	X	X		Essential		X
Synechocystis	X	X	X	X	X	Essential	ZipN
S. elongatus		X	X	X	X	Essential	ZipN
Anabaena		X	X	X	X	Essential	ZipN

^aAnalysis of the conservation of ylm genes in a range of Firmicutes, Actinobacteria, and Cyanobacteria species.

^bThe SsgAB system has been shown to recruit FtsZ to the septum; however, whether it anchors FtsZ to the membrane is unclear.

Unlike bacteria that belong to the Firmicutes or Actinobacteria phylum, cell division in the members of Cyanobacteria are divergent in several respects. For example, they carry most of the ylm genes, but these loci are scattered in different regions of the chromosome instead of being housed in an operon (Table 1) (15). Second, in addition to harboring many of the conserved cell division orthologs, they also feature unique cell division proteins such as Ftn6, which is a homolog of DnaD (15, 16).

Although not much is known about the ylm operon as a whole, SepF has emerged as a key member of the divisome in diverse lineages of Gram-positive bacteria and Cyanobacteria. Since the discovery of the function of SepF in septum formation in B. subtilis, much has been revealed about this protein, including the essential role it plays in Actinobacteria cell division (10, 17–19). Importantly, SepF has been found to anchor the Z-ring (20).

In this minireview, we first bring together SepF literature spanning a decade and a half and then discuss recent developments investigating the roles of other ylm genes in a diverse range of species. Finally, we highlight the recent advances and discuss further questions relating to this conserved locus.

Discovery of SepF.

SepF (YlmF) was first identified as a gene disruption in S. pneumoniae that resulted in cell elongation and minicell formation (11). Further investigation into SepF was done in B. subtilis by two separate groups. Mutational analysis of the entire ylm locus resulted in aberrantly thick and uneven division septa, particularly among cells that were in the early stages of septum formation (10). This phenotype was complemented by the addition of inducible sepF alone in trans. Additionally, it was shown that SepF localizes to the division site in a FtsZ-dependent manner (10). Another group identified B. subtilis SepF through a screen for FtsZ-interacting proteins, showing that SepF is found in a complex containing FtsZ, FtsA, EzrA, and ZapA (17). The interaction between SepF and FtsZ has since been confirmed in a number of species through a range of methods (Fig. 1B; see also Table S1 in the supplemental material) (18, 20–24). Following identification of its role in cell division, SepF was found to assemble into large polymeric ring-like structures through electron microscopy (25). Furthermore, the presence of SepF in vitro promotes higher-order FtsZ polymerization by facilitating lateral interaction between FtsZ protofilaments (25). In vitro work discovered that SepF binds to the FtsZ C-terminal tail (CTT), similar to many other FtsZ-interacting proteins, and promotes FtsZ filament formation and polymer stabilization (19, 22, 26, 27).

To further characterize the interaction between SepF and FtsZ, structural studies of purified SepF protein from B. subtilis and Corynebacterium glutamicum were carried out. This revealed that SepF consists of an N-terminal domain, an intrinsically disordered linker region, and a C-terminal domain (Fig. 1C) (19, 20). The N-terminal domain consists of an amphipathic helix that facilitates binding of SepF to the membrane. This occurs independently of FtsZ, as mutants deficient in their ability to interact with FtsZ still localized to the membrane (20). Mutational analysis of the C-terminal domain of SepF has shown that it is critical for binding to FtsZ and for oligomerization (20, 25). A summary of known SepF mutations and their phenotypes is shown in Table S2.

SepF—a Z-ring anchor.

The dual ability of SepF to bind to the membrane and to FtsZ enables it to recruit FtsZ to the membrane during cell division and anchor the Z-ring at the division site (20). Recent work in B. subtilis has suggested that the thickness of the SepF ring determines the thickness of the septum and that this feature may be conserved across both the Firmicutes and Actinobacteria phyla (28).

While sepF is conserved in phylogenetically diverse species of bacteria, its essentiality varies between phyla; sepF is essential in Actinobacteria and nonessential in Firmicutes. Given that SepF has the ability to tether FtsZ to the membrane, it is intriguing that the essentiality of sepF correlates with the presence of alternative Z-ring anchors, such as FtsA and EzrA (Table 1). Firmicutes typically have FtsA and EzrA, both of which are able to anchor the Z-ring to the cell membrane during divisome formation (29–31). In B. subtilis when ftsA or ezrA is not present, sepF becomes essential through its ability to complement the function of FtsA (10, 17, 20). Along the same lines, the essentiality of SepF in Actinobacteria can be explained by the lack of both EzrA and FtsA in this family (18, 19, 22). Indeed, in Actinobacteria such as Mycobacterium tuberculosis, Mycobacterium smegmatis, and C. glutamicum, depletion of SepF results in the complete absence of septation, distinct from both B. subtilis and S. pneumoniae where septa are present but malformed (10, 11, 17–19). Unlike other Actinobacteria, Streptomyces species carry genes that encode three SepF homologs—SepF, SepF2 and SepF3 (23). However, whether these genes exhibit functional redundancy and whether sepF is essential in Streptomyces remains to be determined.

Cyanobacteria require the presence of SepF and encode an alternative Z-ring anchor, ZipN; however, the physiological reason for this is yet to be determined (15, 16, 24, 32–34). Additionally, there is some evidence to suggest that in S. aureus, a member of the Firmicutes phylum, sepF is essential despite also carrying both the ftsA and ezrA genes (35, 36).

SepF forms a complex with the other Z-ring anchors FtsA and EzrA in B. subtilis and S. aureus, but whether there is a direct interaction between these proteins remains unclear (17, 21). Results to investigate S. aureus cell division protein-protein interactions through a bacterial two-hybrid screen (in E. coli) suggest that there may be a direct interaction between SepF and EzrA (37). However, it must be noted that positive results in bacterial two-hybrid assay are indications but not definitive proof for interactions (38, 39). In Cyanobacteria, a direct interaction between SepF and ZipN has been found in both Synechocystis and Anabaena, but the significance of this interaction is yet to be elucidated (32, 33).

The importance of maintaining the correct level of FtsZ-anchoring proteins has previously been shown in E. coli cells overexpressing either the ftsA gene or the Gram-negative alternative Z-ring anchor gene zipA (40–42). Therefore, it is perhaps unsurprising that overproduction of SepF results in a similar lethal phenotype in M. smegmatis and B. subtilis (18, 43). In B. subtilis, overproduction of SepF causes a dysregulation of divisome recruitment. Early divisome proteins (shown by FtsZ and FtsA) still localize correctly, but there is a failure for the late divisome proteins such as the penicillin binding proteins (PBPs) and FtsW to be recruited and complete divisome formation (43). This results in filamentation, which is indicative of cell division arrest. Interestingly, elevated FtsZ level or the deletion of either ftsA or ezrA overcomes this phenotype (43). In M. smegmatis, overproduction of SepF results in elongated and branched cells, again signifying a block in cell division (18).

Other roles and interaction partners of SepF.

Although a lot is known about SepF in the model organism B. subtilis and the clinically important pathogen M. tuberculosis, much less is known about the role of SepF in other species. In the pathogenic coccoid bacterium Enterococcus faecalis, early work has suggested that deletion of sepF may moderately increase biofilm formation; however, the mechanism behind this still requires elucidation (44). A bacterial two-hybrid screen in Listeria monocytogenes indicates that SepF interacts with the DivIVA-like protein GpsB, but a similar screen in S. aureus has not shown this interaction to be conserved (37, 45).

In M. tuberculosis, which also lacks MreB in addition to FtsA and EzrA, SepF has been shown to interact with MurG (involved in peptidoglycan synthesis) potentially directing synthesis of new cell wall to the division site (22). Of the three Streptomyces SepF homologs, only SepF has been shown to directly interact with FtsZ (23). However, SepF2 is able to interact with both itself and SepF, indicating it may form part of the septal ring despite not containing an N-terminal membrane binding domain (23). During Streptomyces sporulation, it is thought that SepF and SepF2 provide a link between FtsZ and the two dynamin-like proteins, DynA and DynB, through direct interactions (23). SepF3 contains the conserved C-terminal domain and is known to interact with both SepF and SepF2, but its precise role remains to be determined (23).

In the Cyanobacteria Synechocystis, where sepF is an essential gene, depletion of SepF results in the formation of giant cells, indicative of cell division arrest in spherical bacteria, similar to depletion/deletion phenotypes seen in other species (16). Despite the presence of enlarged cells, it was revealed that Z-rings are formed in the absence of SepF; however, they are no longer recruited to midcell (16). Work in Anabaena, a heterocyst-forming cyanobacterium, has revealed an interaction between SepF and PatA, a transcriptional regulator, linking cellular division to transcriptional regulation of developmental genes (46). In addition to PatA, SepF in Anabaena has also been shown to interact with another divisome-associated protein, SepI (34). A full list of interaction partners (including negative results) of SepF and other Ylm proteins, and the methods used to test those interactions can be found in Fig. 1B and Table S1.

SepF as an antimicrobial target and its role in the formation of L-forms.

Given the essentiality of cell division to bacterial survival, there has long been interest in developing novel antimicrobial treatments against cell division targets (47). SepF is no exception (47). Rhodomyrtone, a natural antimicrobial compound, has been shown to have activity in B. subtilis against SepF as well as FtsA, resulting in mislocalization of the FtsZ ring, cell division arrest, and ultimately cell lysis (48). However, the effect of rhodomyrtone is likely indirect, as it affects the cell membrane architecture and membrane potential, causing FtsA and SepF to delocalize (48–50). On a related note, understanding the role of cell wall-less L-forms has been an important part of antimicrobial studies looking to understand chronic infections, in particular those that can evade cell wall-targeting drugs (51). Work in B. subtilis has shown that proliferation of L-forms is independent of FtsZ, indicating a noncanonical mode for cellular division (52). It was also revealed that mutations in SepF (T11M) and the WalR response regulator in B. subtilis allow propagation of L-forms in the laboratory (51, 53). The mechanism of L-form emergence appears to be through breaches at the sites of cell division. Whether these mutations are present in clinical samples of Gram-positive bacteria remains to be seen. Given that both SepF and WalR are conserved specifically in Gram-positive organisms and the latter being specific to Firmicutes (54), analogous mechanisms are likely present in Gram-negative bacteria.

What about the rest of the ylm operon?

Most of the work carried out on the ylm operon has been focused on SepF; however, initial work shows that the other members of this operon may also play a role in cell division and/or sporulation. Here, we focus on what is known about the other genes of the ylm operon in Firmicutes, Actinobacteria, and Cyanobacteria.

YlmD.

The first gene of the ylm operon is ylmD. In all the species in which it has been investigated, ylmD is nonessential (12, 33, 55). YlmD is predicted to be a homolog of the E. coli peptidoglycan editing factor PgeF (YfiH) (56). However, unlike ylmD, the E. coli pgeF locus is located far from the dcw cluster. In E. coli, PgeF helps maintain the composition of peptidoglycan, and deletion of pgeF from the E. coli chromosome results in noncanonical amino acids (l-serine or glycine) being incorporated into the cell wall, resulting in increased sensitivity toward β-lactam antibiotics (56). This is also seen in B. subtilis cells lacking ylmD (56). E. coli cells that contain a deletion in pgeF, as well as two other cell wall synthesis genes ampG (muropeptide permease) and amiD (amidase), exhibit a growth defect that can be resolved by expressing ylmD from either B. subtilis or M. tuberculosis, indicating conservation of function between the homologs (56).

In the Cyanobacteria Synechocystis, deletion of ylmD has no observable phenotype (33). It was shown in a bacterial two-hybrid screen that YlmD interacts with itself, the Z-ring anchor ZipN, FtsI (PBP3), and FtsQ (33, 57). The interaction of YlmD with the latter two hints that the role of YlmD in cell wall regulation may be conserved beyond B. subtilis and M. tuberculosis.

The involvement of YlmD in sporulation was investigated in Streptomyces. Here, deletion of ylmD results in a heterogenous population where some cells exhibit increased spore size as well as reduced spore viability (12, 55). Fluorescence and electron microscopy of spores lacking ylmD indicated reduced septal peptidoglycan synthesis and aberrant spacing between Z-rings. Additionally, YlmD was reported to display diffuse cytoplasmic localization throughout the hyphae and spores (12). In B. subtilis, the role of Ylm proteins in endospore formation is unclear, although it appears that the ylm operon is repressed by the sporulation-specific transcription factor Spo0A (58).

YlmE.

Homology modeling has suggested that YlmE from B. subtilis may be a functional homolog of the E. coli protein YggS, a pyridoxal 5′-phosphate binding protein important for coenzyme A homeostasis (59, 60). Previous work had suggested that YlmE may be a racemase; however, there is no evidence for this activity in vitro (59). In E. coli, loss of yggS results in extracellular l-valine accumulation that can be complemented by expressing ylmE from B. subtilis, indicating a conserved function (59). Disruption of ylmE in S. pneumoniae results in the production of cells containing thin division septa and a modest increase in cell diameter compared to wild-type cells (11).

In Streptomyces, deletion of ylmE leads to defective spore wall formation that is significantly thinner than that of the wild type (12). Like YlmD, YlmE is also found to be diffused throughout the cytoplasm of the hyphae and spores (12). Cells lacking ylmE produce unusually large, heat-sensitive, and reduced number of viable spores (12, 55). This phenotype is more severe than that of a ylmD deletion. However, interestingly, the deletion of ylmE in combination with ylmD restores the wild-type phenotype (12), giving rise to a speculation that YlmD and YlmE could form a toxin/antitoxin-like bipartite system.

YlmG.

ylmG is a small gene that is highly conserved across not only bacteria but also in the chloroplasts of photosynthetic organisms (61). It is predicted to encode a small protein of around 90 amino acids in B. subtilis and of similar size in other species, consisting of a transmembrane domain at the N and C termini linked by a short cytoplasmic domain in the middle (62). YlmG has been grouped into the YggT family of proteins. In E. coli, YggT is involved in osmoregulation; however, due to poor sequence homology, it is speculated that there may not be any functional similarity between YlmG and YggT (62, 63). Deletion of ylmG on its own in S. pneumoniae results in the production of thinner septa as observed by electron microscopy (11). When ylmG is deleted in combination with ylmE, cells divide along multiple planes, unlike wild-type streptococci that typically divide in the middle along the short axis of the cell (11).

As with ylmD and ylmE, the role of ylmG has been investigated in Streptomyces sporulation. Cells lacking ylmG produce defective spores that have a thinner cell wall and as a consequence are less heat resistant (62). Additionally, the nucleoids in these spores appear less condensed (62). Localization of the Z-ring positioning protein SsgB was found to be dependent on the presence of ylmG (62, 64). Because of its role in septation, YlmG has been renamed SepG in Streptomyces (62).

The role of YlmG in Cyanobacteria has been studied in Synechococcus elongatus. In this species, it appears that YlmG plays a role in nucleoid segregation as deletion results in nucleoid compaction (61). Overproduction of YlmG results in longer cells with abnormal nucleoid distribution (61). The placement of the Z-rings within these cells was found to be biased toward the pole of the cell that display increased nucleoid density (61). Interestingly, this gene is also conserved in plant chloroplasts and plays a role in chloroplast and plant embryo development (61, 65).

YlmH.

Of all the genes in the ylm operon, ylmH is the least studied. It appears to be conserved only in the Firmicutes phylum and some Cyanobacteria—however, some orthologs have been found in chloroplasts (15). Deletion in S. pneumoniae results in a population that lyses before reaching exponential phase, a phenotype possibly explained by the presence of thinner septa (11). Sequence analysis of YlmH suggests that it contains an S4 RNA binding domain, a domain often associated with translation-related processes (66).

Concluding thoughts.

To obtain a comprehensive understanding of cell division regulation in bacteria, it is imperative that the functions of all proteins of unknown function are elucidated and factors that are even distantly related to cell division are investigated. Thus, the roles of Ylm proteins other than SepF need to be revisited and investigated in detail. As recent studies have shown, any given protein could play an altered and/or essential role in different organisms. Therefore, studies conducted in model organisms alone are not sufficient. Recent investigations in nonmodel organisms have revealed new modes of cell division and new mechanisms of cell division regulation (2). With the advent of new genetic tools and advanced microscopy techniques, the time is ripe to undertake such exploratory expeditions.