How Bacterial Cell Division Might Cheat Turgor Pressure – a Uniﬁed Mechanism of Septal Division in Gram-Positive and Gram-Negative Bacteria

Insights & Perspectives Hypotheses

How bacterial cell division might cheat turgor pressure – a uniﬁed mechanism of septal division in Gram-positive and Gram-negative bacteria

Harold P. Erickson

An important question for bacterial cell division is how the invaginating septum can periplasm, which is a major focus of the overcome the turgor force generated by the high osmolarity of the cytoplasm. I present analysis. To achieve division all suggest that it may not need to. Several studies in Gram-negative bacteria have layers of the cell envelope must invaginate into the division site, although not shown that the periplasm is isoosmolar with the cytoplasm. Indirect evidence at the same time, and eventually fully suggests that this is also true for Gram-positive bacteria. In this case the invagination cover the daughter cells. of the septum takes place within the uniformly high osmotic pressure environment, In 2008, Osawa et al. [6] demon- and does not have to fight turgor pressure. A related question is how the V-shaped strated that FtsZ-mts could assemble Z constriction of Gram-negative bacteria relates to the plate-like septum of Gram- rings when reconstituted in tubular liposomes. The mts (membrane-target- positive bacteria. I collected evidence that Gram-negative bacteria have a latent ing sequence) is an amphipathic helix capability of forming plate-like septa, and present a model in which septal division is that can tether FtsZ directly to mem- the basic mechanism in both Gram-positive and Gram-negative bacteria. brane. These reconstituted Z rings could constrict the thick walls of Keywords: multilamellar liposomes. Both Z-ring .bacteria; cell division; cytokinesis; FtsZ; peptidoglycan; tubulin; turgor assembly and constriction were achieved by FtsZ-mts alone no other division protein was needed. This Introduction cycle. Near the end of the cell cycle the Z discovery elevated FtsZ as the prime ring, including all the downstream candidate for generating the constric- In bacterial cytokinesis FtsZ provides proteins, constricts over a period of tion force in bacterial cell division. the cytoskeletal framework for the Z minutes to pinch the cell in two. The How could FtsZ constrict mem- ring, which serves as the docking site for mechanisms of FtsZ have been investi- branes? Previous work had demon- up to two dozen other division proteins. gated intensively since the discovery strated that FtsZ protofilaments can The Z ring assembles early in the cell 25 years ago that FtsZ is a homolog of switch conformation from straight to cycle and remains in place without any eukaryotic tubulin, and that it is local- curved, and it was suggested that the obvious constriction for most of the ized at the site of constriction [1–5]. curved protofilaments could generate a Bacteria are surrounded by a cell bending force on the membrane and DOI 10.1002/bies.201700045 envelope. In Gram-positive bacteria the pull it inward [7, 8]. Additional experi- cell envelope consists of an inner lipid ments with “inside-out Z rings” sup- Department of Cell Biology, Duke University membrane (IM) and a rigid cell wall ported this mechanism [9–12]. Medical School, Durham, NC 27710, USA composed primarily of peptidoglycan Recently this Z-centric hypothesis (the abbreviation PG will refer to the cell has been questioned. Coltharp and Corresponding author: wall, implicitly including all other Xiao [13, 14] suggested that the force Harold P. Erickson E-mail: [email protected] components). Gram-negative bacteria from bending FtsZ protofilaments have an additional outer lipid mem- would be insufficient to constrict the Abbreviations: brane (OM), which is closely attached to IM against the turgor pressure gener- EM, electron microscopy; IM, inner membrane; MDO, membrane-derived oligosaccharide; OM, the PG. The PG layer is separated from ated by the high osmolarity of the outer membrane; PG, peptidoglycan. the IM by a variable space termed the bacterial cytoplasm. They suggested

Bioessays 39: 1700045, ß 2017 WILEY Periodicals, Inc. www.bioessays-journal.com 1700045 (1 of 10) H. P. Erickson Insights & Perspectives.....

that the major force for constriction is and relies on the natural contrast of the likely generated by inward growth of the membranes and protein structures with PG wall, pushing the IM from the surrounding water. Most of the results outside. In this mechanism the role of discussed below (Figs. 1–3) were FtsZ is primarily to serve as a docking obtained from high pressure freezing site for the PG synthetic enzymes, and to followed by cryosectioning. An alter- regulate assembly of the PG. native to cryosectioning is cryoEM Here, I am not arguing for one tomography, where the whole bacte- constriction mechanism or the other. I rium is imaged at multiple tilts and a 3- present the hypothesis that the essen- D image is reconstructed (Fig. 4). The tial steps of cell division, constricting thin samples for tomography are typi- Hypotheses the IM, and building new PG to form a cally prepared by plunge freezing Figure 1. CryoEM (high pressure freezing E. septum, likely take place completely without the need for high pressure. and cryosection) of the cell envelope of coli K-12. The black arrows point to the IM within the high osmolarity environ- Matias, Beveridge, and colleagues and OM. The PG is the lightly stained ment, and never have to resist the applied the technique of high pressure narrow band indicated by the white arrow. turgor pressure. This would mean freezing and cryosectioning to bacteria The periplasm is the lighter stained zone that both potential mechanisms, FtsZ in a series of seminal papers from 2003 between the PG and IM. Bar is 50 nm. bending and cell wall ingrowth, would to 2008. Figure 1 shows the cell enve- Reprinted with permission from [16]. need only modest forces to achieve lope of E. coli K12 [16]. The IM and OM division. are each seen as a single, thin, dark staining layer. The PG appears as a line medium density “outer wall zone” that of slightly enhanced density close to the can be identified as the PG layer. The cell envelope of OM. The periplasm is the lighter stain- Several other Gram-positive bacteria Gram-positive and Gram- ing space between the IM and the PG. have been imaged bythe same technique, Matias et al. [16] noted two important anddimensionsof theircellenvelopes are negative bacteria features of the periplasmic space: its given in Table 1. For most of these the visualized by cryoEM width varied by 10% over non- periplasm is 20 nm thick versus 10 nm compressed sections of the cell enve- for E. coli, and the PG is much thicker, The structural relationships of the mem- lope; and it could be significantly 20–30 nm versus 6.4 nm for E. coli.An branes and PG wall have been determined compressed by the sectioning knife. exception is Mycobacterium bovis, whose by thin section electron microscopy (EM). They suggested that the periplasmic periplasm and PG are closer to the In the early days of EM, bacteria were space has inherent compressibility or dimensions of E. coli. Interestingly, M. chemically fixed, dehydrated and infil- flexibility that allows the IM to move bovis and related Corynebacterineae were trated with plastic, and thin sections were closer or farther away from the PG. It is shown to have an outer lipid bilayer, cut. Aldehydes and OsO4 were the most not a rigid gel. chemically different but structurally sim- popular fixation agents. There was, how- The term “periplasm” has had two ilar to the OM of E. coli [19]. ever, always concern about changes in the different meanings in the literature. A structure that might occur during the time number of authors use the term to refer to for the fixation to work and during the entire space between the IM and OM, The periplasm of Gram- subsequent processing. with the PG layer being contained within negative bacteria is Rapid freezing offered the promise this periplasm. However, we will use the of instantaneous fixation, but ice terms periplasm or periplasmic space to isoosmotic with the crystal formation disrupted tissue refer specifically to thespacebetweenthe cytoplasm structures more than a few mmfrom PG and IM. This designation was used in the freezing surface. Cryoprotectants the original study of osmolarity of the An important concern for bacterial divi- could reduce ice crystals, but might compartments [17], and it is also consis- sionishowitdealswiththeturgorforceof alter the cellular structures. A major tent with the definition of periplasm in the bacterial cytoplasm. To understand advance was the development of high Gram-positive bacteria. this we need to know which elements of pressure freezing; see McDonald for a Early studies had suggested that in the cell envelope generate and which review of the history and applica- Gram-positive bacteria the IM was elements contain the turgor force. tions [15]. Briefly, the sample is sub- pressed in tight apposition to the In Gram-negative bacteria the OM is jected to 2,000 bar pressure relatively thick PG cell wall [18]. In tightly coupled to the PG by the very simultaneously with rapid freezing, 2005, Matias and Beveridge [18] over- abundant lipoprotein Lpp (other attach- both being achieved in 30 micro- turned this model when they applied ments exist but Lpp is the most seconds. The high pressure prevents cryosectioning to image the envelope of abundant). Lpp is a triple helical rod, formation of ice crystals, and the rapid Bacillus subtilis. Figure 2 reproduces 8.3 nm long with three fatty acid chains freezing preserves the organic struc- their definitive images. The cell enve- on its N-terminus that insert into the tures. The frozen specimen can then be lope comprises the IM, resolved here as OM [20]. About 1/3 of the chains are sectioned with a cryo microtome. This bilayer; an “intermediate wall zone” covalently attached to PG at the C generally involves no chemical stain identified here as the periplasm; and a terminus, so most of the trimers should

1700045 (2 of 10) Bioessays 39: 1700045, ß 2017 WILEY Periodicals, Inc...... Insights & Perspectives H. P. Erickson

through the OM-PG wall into the is flexible and unable to support a periplasm. Importantly, the IM is im- pressure gradient ... Under all condi- permeable to sucrose. This means that tions the periplasm and cytoplasm Hypotheses when sucrose is added to the outside remained isoosmotic” [17]. medium it will diffuse through the OM- The osmolarity of the Gram-negative PG layer into the periplasm, but will not periplasm is thought to be achieved by move across the IM into the cytoplasm. membrane-derived oligosaccharides The cytoplasm has a high concen- (MDOs). MDOs are anionic glucose tration of charged macromolecules and oligomers, 2,300 Da with an average salts to neutralize them, creating a high 3 charge, that are too large to pass osmolarity and a turgor pressure. Over through the small pores of the OM-PG the years there have been two models layer [24–26]. They create a Donnan for how this turgor pressure is sup- equilibrium with an osmolarity that ported. In one model the turgor pressure equals that of the cytoplasm. Upon is generated by the IM. The IM is not plasmolysis the expanded spaces per- structurally rigid, so it would have to be sist for 30 min or more, but if cells are supported by the rigid PG layer. Since maintained in a high osmolar medium EM shows that the IM is separated from they down-regulate synthesis of the PG wall by a 10 nm periplasmic MDOs [24–26], the cytoplasm swells space, this would require that the and the periplasm shrinks. periplasm be a rigid, inelastic gel that Cayley et al. [26] expanded the study can support the turgor pressure on the to measure effects of osmotic shock on IM. In the second model the turgor cells growing in low to high osmolar pressure is generated by the OM-PG medium. This is now the definitive study layer, and the cell maintains the of osmolarity of the periplasm. They periplasm in a high osmolar state to concluded that “the periplasm and match that of the cytoplasm. The IM cytoplasm are isoosmotic, and that E. then experiences no turgor pressure, coli maintains turgor pressure across and floats freely between the cytoplasm the cell wall and not across the cyto- Figure 2. CryoEM of the cell envelope of B. subtilis. A: cross section through the middle and periplasm. This second model is plasmic membrane” [26]. of a whole cell. B: higher magnification view. now widely accepted, based on work IM is designated PM (plasma membrane), discussed next. periplasm is designated IWZ and the PG layer Stock et al. [17] did the pioneering The periplasm of Gram- is OWZ. Bars are 200 nm (A) and 50 nm (B). study of the periplasm of E. coli and positive bacteria is Reprinted with permission from [18]. Salmonella typhimurium using radio- labeled probes to determine the fraction probably isoosmotic with form a tight bridge from the OM to PG. of water in the cytoplasm and peri- the cytoplasm There are 2.4 105 Lpp trimers per plasm. They found that the periplasmic cell [20], so they should be closely volume increased rapidly in response to Since Gram-positive bacteria have no spaced on the OM, with 30 nm2 per an abrupt increase in sucrose in the OM, the outer semipermeable barrier is Lpp trimer. In view of this tight external medium. As noted above, the PG wall itself. The PG layer in Gram- coupling, the OM-PG will be considered sucrose crosses the OM-PG barrier and positive bacteria is much thicker than in to be a unit that creates the turgor rapidly increases the osmolarity of the Gram-negative, and probably operates as pressure and provides the mechanical periplasm, but cannot cross the IM. This a molecular sieve. The question is, what support to contain it. Note that this is causes water to flow from the cyto- size molecules can pass through this the first filtration barrier an external plasm, increasing its osmotic pressure sieve and what size are blocked? osmolyte would encounter. If it passes until it matches the newly increased Scherrer and Gerhardt [27] measured this barrier it will enter the periplasm, level of the periplasm. The flow of water the ability of dextrans and polyglycols where it will then encounter the more from the cytoplasm reduces cytoplasmic to permeate into living Bacillus mega- selective barrier of the IM. volume and pulls the IM inward, terium. That paper is sometimes misun- The filtration barrier of the OM-PG substantially increasing the periplasmic derstood as demonstrating that the layer is based primarily on the size of volume. The IM does not retract uni- “number average molecular weight” the osmolyte. Decad and Nikaido [21] formly, but forms a small number of exclusion limit of the Bacillus cell wall assayed its permeability to sugars and “plasmolysis spaces,” often at the cell is 70,000–120,000 Da. That was indeed concluded that “the cell wall acts as a poles and division sites. Schwarz and their initial finding based on tests of molecular sieve, with an exclusion limit Koch [22] imaged these expanded pock- unfractionated polymers. But the im- near 550–650 Daltons for saccharides.” ets of periplasm by EM, and Sochacki portant point of their study was to The limit of 550–650 Da is probably et al. [23] used periplasmic GFP to image recognize that the dextrans and poly- imposed by the porins in the OM. them in living cells. Stock et al. con- glycols are polydisperse, and that the Sucrose, at 342 Da, easily passes cluded that the “cytoplasmic membrane apparent passage is highly tilted toward

Bioessays 39: 1700045, ß 2017 WILEY Periodicals, Inc. 1700045 (3 of 10) H. P. Erickson Insights & Perspectives.....

It is remarkable that 40 years after the deﬁnitive demonstration of isoosmolarity for Gram-negative bacteria [17], there is no comparable measurement for Gram-positive bacteria. We really need a deﬁnitive measure of the pore size of the PG wall in living Gram-positive bacteria, and a measure of the osmolarity of the periplasm. Another important question is the nature of the osmolytes in the Gram-positive periplasm. MDOs have not Hypotheses been reported for Gram-positive bacteria, but teichoic acids are a candidate [31, 32].

Gram-positive bacteria divide by septation

Figure 3. Septal division of S. aureus. A: The bacterium at lower left has a completed Several species of Gram-positive bacte- septum but has not separated. The bacterium in the center has a septum partially ria show clear evidence for division by a invaginated. B: A newly forming septum in cross-section. White arrows indicate the PG wall septation mechanism, in which the of the mother cell and the two forming daughter PG layers. (C) A scanning EM of a cell just beginning to pop, and one that has just popped. (A and B) are from [33], (C) is from [34]. invaginating IM forms a thin plate. Bars are (A) 250 nm, (B) 50 nm, (C) 1,000 nm. The PG wall of the mother cell remains in place around the outside, and new PG wall for the daughter cells grows into the smaller molecules in the sample. not passive transport through fixed the septum. Matias and Beveridge [33] Based on fractionated polyglycols they pores [30]. obtained superb images of septa in concluded that “... monodisperse mol- Thebestevidenceforanisoosmo- Staphylococcus aureus using high pres- ecules can penetrate the cell wall only if lar periplasm are the cryoEM images sure freezing and cryosectioning. equivalently smaller than a glycol of showing that the IM of Gram-positive Figure 3A shows bacteria with forming Mn ¼ 1,200 Da” [27]. This is about twice bacteria is not plastered against the PG and completed septa and Fig. 3B shows the size that can penetrate the OM-PG of cell wall, but is separated by a 20 nm a high magnification cross section of a Gram-negative bacteria. periplasmic space (Fig. 2) [18]. This newly forming septum. The IM is Demchick and Koch [28] reported a means that, to support the cytoplasmic resolved as a bilayer, defining the edges much larger permeability barrier of turgor pressure, the periplasm must of the septal plate, which has a uniform 30–40 kDa for isolated B. subtilis PG either be a rigid gel or be isoosmotic width of 75 nm. Cell wall material is sacculi. They used dextrans that were with the cytoplasm. Evidence against a identified as two “high-density zones,” size fractionated, so their measures rigid gel was obtained by plasmolysis which are the future outer wall PG of the should not be distorted by polydisper- experiments [18]. Sudden treatment daughter cells. They are apparently sity. However, the sacculi they assayed with20%glycerolplus5%NaCl fused to the outer wall PG, although were prepared by treatment with boiling caused the cytoplasm to shrink and the EM images do not inform on the SDS and then with trypsin, which would the periplasmic space to expand. Some molecular linkage. The septum was held remove all wall proteins. This may have expanded periplasmic spaces con- together primarily by its attachment to increased the pore size. tained small vesicles that could easily the outer cell wall PG layer. The PG layer Anothersuggestionoflargerporesize distort the space. Matias and Beveridge was 19 nm thick over most of the is that Gram-positive bacteria are sus- concluded that “Whatever is in the bacterium, but increased to 40 nm at ceptible to antimicrobial peptides IWZ [periplasm], it is quite compact- the septum, the “outer wall bridge.” and proteins. The antimicrobial peptides able ... and can be deformed by the This bridge presumably holds the sep- are 10–50 amino acids (1,100–5,500 Da), action of the vesicles” [18]. They tum together until it has completed and theymust traverse thepeptidoglycan suggested that there must be a sub- invagination, and it is then disrupted to cell wall to reach the plasma membrane stance in the periplasm that can permit the daughter cells to separate. target [29]. Antimicrobial proteins, such balance the turgor pressure of the Recent work has shown that in S. as the 14 kDa phospholipase A2 are cytoplasm, even if it is not sufficiently aureus this bridge is not digested gradu- even larger [30]. However, in this case concentrated to generate visible den- ally. It fails abruptly in a dramatic antimicrobial activity depended on cell sityintheEM.Thesestructuralobser- “popping” event lasting 1 msec [34]. growth and autolysins, suggesting vations suggest that the periplasm is The failure is mechanical, beginning as a that the large protein traversed the isoosmotic with the cytoplasm in crack at one point on the ring and PG wall by active and perhaps tran- Gram-positive bacteria, as it is in propagating around, leaving the daugh- sient openings during PG remodeling, Gram-negative bacteria. ter cells connected at a hinge (Fig. 3C).

1700045 (4 of 10) Bioessays 39: 1700045, ß 2017 WILEY Periodicals, Inc...... Insights & Perspectives H. P. Erickson Hypotheses

Figure 4. A–C: Division in C. crescentus imaged by cryoEM tomography (reprinted with distinct density for the PG layer, but if C. permission from [37]). The shallow V shape was interpreted to indicate a constrictive mode crescentus is similar to E. coli, the PG of division. A and B are slices through a 3D tomogram, while (C) is a projection through the layer is tightly attached to the OM. In E. coli whole bacterium. D: Division of imaged by cryoEM tomography, showing a slice this case, as the IM invaginates ahead of through the 3D tomogram (reprinted with permission from [39]. The white arrow points to the OM-PG it expands the periplasmic small dots identified as FtsZ filaments in cross-section. This cell has a sharp, V-shaped constriction, typical of constrictive division. The periplasm of this cell is considerably space. Eventually the IM fuses to form expanded, especially at lower right. E: A dividing E. coli cell prepared by high pressure around each daughter cell, well before freezing, freeze substitution (reprinted with permission from [40]. The division shows a V- the OM-PG has finished constricting shaped constriction and a clearly extended invagination of the IM, forming a septum. (Fig. 4C). It should be noted that Judd et al. did not consider their images Note that the new cell wall retains the flat consensus opinion that “Gram-negative representative of Gram-negative bacte- shape it had in the septum, and is only bacteria in general, ... seem to show a ria in general. They were aware of the gradually remodeled to the ovoid shape. process of constriction involving all work of Burdett and Murray [38] show- Overall the picture seems very clear for layers at once as if a noose were being ing a septation mechanism for E. coli S. aureus. Zuber et al. [35] reported a very tightened around the equator of the (discussed below) and they concluded similar septal structure in Enterococcus cell” [36]. This results in a V-shaped that the constrictive division they found gallinarum, also using high pressure constriction of all layers (Figs. 4A, D and was specific for C. crescentus [37]. freezing and cryosectioning. Division by 5A), very different from the septal plate However, a recent cryoEM study invagination of a thin, plate-like septum established for Gram-positive bacteria. provided evidence for a constrictive seems to be the general mechanism for Recent cryoEM provides support for division in E. coli (Fig. 4D) [39]. The main Gram-positive bacteria. constrictive division. In 2005, Judd focus of that study was to image FtsZ et al. [37] undertook a study of Caulo- filaments, and this was its major success; bacter crescentus division using cryoEM note the rows of dots under the IM in Constrictive division in tomography. Their images showed Fig. 4D, which are identified as ribbons of Gram-negative bacteria broad constrictions that initially main- FtsZ protofilaments in cross-section. The tained the 30 nm spacing of the IM and authors did not address the question of may be a modified OM. This was consistent with the constrictive vs. septal division, but septation mechanism classical constriction mechanism. As Fig. 4D shows a classic V-shaped con- constriction progressed, however, the striction. The PG layer is a thin, lightly Based on early EM studies, Steed and IM was seen to move ahead of the OM contrasted line close to the OM, as seen in Murray in 1966 summarized the (Fig. 4B). Their images did not show a Fig. 1. In this image the periplasmic space

Bioessays 39: 1700045, ß 2017 WILEY Periodicals, Inc. 1700045 (5 of 10) H. P. Erickson Insights & Perspectives.....

Table 1. Dimensions of the cell envelope of various bacteria determined from cryoEM of cryosections following high pressure freezing.

Thickness (nm) Bacterial strain IM periplasm PG Ref. E. coli K12 5.8 9.3 6.4 [16] E. coli K12 6.3 11.5 6.4 [19] P. aeruginosa PAO1 6.0 15 2.4 [16] B. subtilis 168 6.6 22 33 [18] S. aureus D2C 5.4 16 19 [49] B. subtilis 168 20 30 [35]

Hypotheses B. subtilis W23 20 33 [35] S. gordonii 16 26 [35] M. bovis 6.3 14 6.3 [19]

seems to have been greatly expanded, that the conventional fixatives did not to determine their time of exit from the Z suggesting possible plasmolysis. Also, in preserve the short septum, allowing it to ring at the end of division. FtsZ departed the upper left constriction the IM at the be pulled apart into an expanded V- first, followed by FtsA and ZipA. FtsI tip of the V has advanced ahead of the shaped cleft [38]. When the septum was and FtsN remained at the division site OM, similar to Fig. 4B. preserved it was always at the bottom of longer. Their images also provided In contrast, one recent image of the V-shaped cleft, in contrast to the surprising new information on the dividing E. coli prepared by high pressure septa in Gram-positive bacteria, which radial localization of the division pro- freezing shows an apparent septum extend all the way to the cylindrical cell teins. FtsZ, FtsA and ZipA co-localized (Fig. 4E) [40]. This is a clear example of wall. We will refer to this mechanism as in a ring throughout division, consistent what we will term “septation-constric- “septation-constriction.” with the direct binding of FtsA and ZipA tion.” The OM-PG wall has formed a V- Burdette and Murray [38] also to the C-terminal peptide of FtsZ. FtsN, shaped constriction, but at the bottom of obtained beautiful images of septa in FtsI, FtsL, and FtsQ were mostly co- the V the IM has invaginated much the E. coli mutant CRT 97. In this strain, localized with each other. Surprisingly, further, forming a thin septum. We will which may be defective in one or more the ring of FtsZ moved ahead of the ring nextpresent additional evidence that this murein hydrolases, septa were easily of FtsN in constricting cells. The struc- structure may be the underlying mecha- preserved by almost any fixative (Fig. 5C, tured illumination microscopy (SIM) nism of division in E. coli. D). Note that these septa also extend from used for these images has a resolution In 1966, Steed and Murray presented the base of a V-shaped constriction, of about 100 nm, but the centroid of the evidence that Gram-negative bacteria similar to the septation-constriction divi- green and red labeled proteins could be actually did have septa at the division sion seen in wild type E. coli fixed by localized much more precisely, to site,atleastundercertainconditions[36]. acrolein-glutaraldehyde. A later study of about 10 nm. As shown in Fig. 6a, the Specifically,theyfoundseptain50% ofE. Heidrich et al. [41] created mutants FtsZ ring was significantly inside the coli cells grown at 45˚C, using conven- defective in various combinations of FtsN ring. The rings differed in diameter tional fixation. Burdett and Murray [38] the 18 known murein hydrolases. Some by 105 nm over repeated measurements. made a major advance when they ofthesegrewaschainsofcells,withsepta Figure 6f showed that FtsI almost co- explored a wide range of fixatives. spanning the full diameter, that is they localized with FtsN; repeated measure- Fixation with OsO4 or glutaraldehyde had no V-shaped constriction. E. coli thus ments gave a 26 nm smaller diameter for alone, the favored fixatives at that time, has the capability of forming clear plate- FtsI. Combining these two measure- showed dividing cells with the likesepta,suggestingthatthisisthebasic ments and converting to radius, the characteristic V shape of constrictive mechanism for its division. FtsZ ring is about 40 nm closer to the division (Fig. 5A). However, a mixture Overall, these early studies of Mur- cell center than FtsI. of 5% acrolein plus 0.25% glutaralde- ray and colleagues set the stage for The images of Soderstrom€ et al. [42] hyde fixative preserved a large fraction of understanding constrictive division as a can be further interpreted to provide dividing cells with a clear septum modification of septation. We will evidence for a septal mechanism of (Fig. 5B). Importantly, the cells still had pursue this idea below. division, based in particular on the the V shape characteristic of constrictive width of the FtsI and FtsZ rings. These division, but at the bottom of the V a appear to be diffraction limited (100 nm septum continued toward the interior of Super resolution light in the SIM images) in both radial and the cell. The septum was characterized as microscopy provides axial directions. If the division were an annular plate, with the IM on the purely constrictive, with sides of outside defining its 50 nm width, and evidence for septal the V at 45 degrees, the FtsN would two layers of developing PG cell wall in division in E. coli have to expand axially as it moved the middle (Fig. 5B). The OM remained on radially outward from FtsZ. The 50 nm the outside of the septum at the top of Soderstrom€ et al. [42] recently used GFP- radial separation would convert the the V. Burdette and Murray suggested and mCherry-labeled division proteins FtsN ring axially to a double ring

1700045 (6 of 10) Bioessays 39: 1700045, ß 2017 WILEY Periodicals, Inc...... Insights & Perspectives H. P. Erickson

CRT 97 fixed with conventional glu- have completed their fusion for the taraldehyde plus OsO4.Theearlier daughter cells. The E. coli hydrolase study of Steed and Murray [36] found mutants of Heidrich et al. [41], which Hypotheses septa in E. coli preserved by conven- formed septa spanning the full diam- tional fixation in certain growth eter of the mother cell, suggest that E. conditions. A more recent study by coli has the potential for similar high-pressure freezing imaged a clear extreme septation. (B) An intermedi- septum in E. coli [40]. These studies ate mechanism, which we term sep- clearly show that E. coli has the ability tation-constriction, has a short to divide by formation of septa. septum extending inward, but it In most cases the septa of E. coli splits into a V-shaped constriction invaginate from the bottom of a on the outside. This could be formed V-shaped constriction. In some cases, if hydrolysis of the outer layers of PG especially early in division, begins while the septum is still the V-shaped constriction domi- advancing. This would permit the nated [37, 39]. forming septum to separate into a These structures can be resolved V-shaped constriction above the site by postulating septation as the uni- of hydrolysis, while the septum versal division mechanism, but rec- advances below. This is apparently ognizing three variations in how the the state captured in the images in PG wall is hydrolyzed. These are Figs.5B–D,4E.(C)Finally,atthe diagrammed in Fig. 7. (A) At one other extreme, are images showing a extreme are Gram-positive bacteria, V-shaped constriction, with no obvi- which divide by a full septation ous septum (Fig. 4A–D). We suggest mechanism. Here the bridge PG of these are produced when hydrolysis the mother cell is maintained in place follows so closely after the advancing Figure 5. E. coli division imaged by conven- until the septum has completed its septum that it virtually obliterates the tional embedding and sectioning following invagination. The bridge PG is hydro- septal structure. Importantly how- different fixation methods. A: Fixation with lyzed and/or mechanically broken ever, even these images retained a OsO4, the standard method in the 1970’s, only after the IM and the PG plates hint of septation. The IM appeared to gives the conventional picture of constrictive division. B: Acrolein plus glutaraldehyde, a novel fixative introduced by Burdett and Murray, preserved clear septa. The two dark lines within the septa are the newly forming PG layers. C, D: E. coli CRT 97 is a chain- forming strain apparently deficient in murein hydrolases. Septa in these strains can be preserved by any fixation, OsO4 being shown here. Reprinted with permission from [38].

100 nm apart. This would not be resolved by SIM, but it should produce an obvious axial broadening. This was not seen. The images are consistent with a septal width of 50 nm or less for both FtsZ and FtsN.

Hypothesis constrictive division in Gram-negative bacteria is a variation of septation

The study of Burdett and Murray [38] Figure 6. Localization of FtsZ, FtsN and FtsI by SIM superresolution microscopy. Dashed provided convincing images of septa white lines indicate the cell boundaries. Panel a shows that FtsZ (green) is signiﬁcantly in wild type E. coli ﬁxed with acrolein- toward the center of the cell relative to FtsN (red). Panel f shows that FtsI is almost co- glutaraldehyde, and in the mutant localized with FtsN, but slightly toward the center. Reprinted with permission from [42].

Bioessays 39: 1700045, ß 2017 WILEY Periodicals, Inc. 1700045 (7 of 10) H. P. Erickson Insights & Perspectives..... Hypotheses

Figure 7. Three variations on a general model of bacterial division by septation. A: The Gram-negative division is the timing of Gram-positive septum. The components are: FtsZ, a 5 nm diameter sphere with a 10 nm murein hydrolases digesting the old PG peptide linker that connects to FtsA, a 5 nm sphere with an amphipathic helix that inserts wall. In Gram-positive division this into the IM. The Z ring shown here in cross-section is a ribbon of four protofilaments, occurs only after the septum is com- consistent with recent cryoEM [39]; see also [50] for arguments supporting this narrow width. The IM, a 4 nm thick lipid bilayer that defines the two sides of the septum 70 nm apart; PG plete, while in Gram-negative it occurs parental, a 30 nm wide sheet represented as a blue to green gradient inside to outside; PG while the septum is still advancing. bridge, a thickening of the parental PG at the septum [33]; periplasm, the 20 nm space The structure of the V raises the between the IM and PG; FtsI (PBP3 in E. coli and PBP2B in B. subtilis), a transmembrane question, why doesn’t the turgor pres- protein with an elongated periplasmic domain [51] that can bridge the 10 nm wide sure collapse it outward? Probably periplasmic space in the septum. FtsI is shown here spaced 40 nm above the FtsZ ring [42]). because the PG is rigid not just to This septum remains connected to the outer wall PG of the mother cell until invagination is complete. B: The Gram-negative “septation-constriction” mechanism. Most components are stretching, but also to compression and the same as in the Gram-positive septum with different dimensions: The periplasm is 10 nm bending. An example of this rigidity is and the PG is 6 nm thick. In addition the Gram-negative bacteria have an OM, a 4 nm thick the structure of the cell walls of S. bilayer (ignoring here the oligosaccharide projections) covalently attached to the PG by the aureus immediately following the pop- abundant Lpp lipoprotein rods [20]. In this septation-constriction mechanism the outer layers ping separation (Fig. 3C). Here they are of the septum have split to generate the V-shaped constriction. (C) In the (apparent) abruptly exposed to turgor pressure, but constriction-only mechanism, the splitting of the septum on the outside follows very closely its invagination toward the center, so that the septum is hardly visible. they maintain their flat shape. Rigidity of the PG layer is also suggested by the advance ahead of the OM-PG, leading the sides of the V. This bridge must also observations that E. coli cells forced to to an extended periplasmic space and somehow be attached to the daughter grow into odd shapes retain their a sharper V for the IM (Fig. 4B, D). PG in the septum. Its growth on the shapes when released, and only reform An essential part of the constriction- inside and hydrolysis on the outside after one or more cell cycles [43, 44]. septation mechanism is the PG bridge, would need to be coupled to avoid a Another point in favor of a common at the point of the V, which must rupture. In this scenario the major septal mechanism is based on the maintain a seal with the OM-PG on difference between Gram-positive and argument of Gupta that the earliest

1700045 (8 of 10) Bioessays 39: 1700045, ß 2017 WILEY Periodicals, Inc...... Insights & Perspectives H. P. Erickson

bacteria were the simpler Gram-posi- Acknowledgments 17. Stock JB, Rauch B, Roseman S. 1977. tive “monoderms,” and that the Gram- I thank Dr. Masaki Osawa for many Periplasmic space in Salmonella typhimurium negative “diderms,” with the extra informative andchallenging discussions. and Escherichia coli. JBiolChem252: 7850–61. 18. Matias VR, Beveridge TJ. 2005. Cryo- Hypotheses outer membrane, evolved from Supported by NIH grant GM066014. I electron microscopy reveals native polymeric them [45]. In this case the septal declare no conflicts of interest. cell wall structure in Bacillus subtilis 168 and mechanism would have been inherited The authors declare no conflict of the existence of a periplasmic space. Mol Microbiol 56: 240–51. by the earliest Gram-negative bacteria. interest. 19. Zuber B, Chami M, Houssin C, Dubochet J Consistent with this, Deinococcus spe- et al. 2008. Direct visualization of the outer cies, which appear to be a primitive membrane of mycobacteria and corynebac- teria in their native state. J Bacteriol 190: Gram-negative form, divide by a dis- References 5672–80. tinctly septal mechanism [46]. The V- 20. Shu W, Liu J, Ji H, Lu M. 2000. Core structure shaped constrictive division would 1. Bi E, Lutkenhaus J. 1991. FtsZ ring structure of the outer membrane lipoprotein from then be a later modification on the associated with division in Escherichia coli. Escherichia coli at 1.9 A resolution. J Mol Nature 354: 161–4. Biol 299: 1101–12. septal mechanism. This is supported by 2. RayChaudhuri D, Park JT. 1992. Escherichia 21. Decad GM, Nikaido H. 1976. Outer mem- the observation that E. coli still retains coli cell-division gene ftsZ encodes a novel brane of gram-negative bacteria. XII. Molec- the possibility of dividing by a septal GTP-binding protein. Nature 359: 251–4. ular-sieving function of cell wall. J Bacteriol 128: 325–36. mechanism (Fig. 5B–D). 3. de Boer P, Crossley R, Rothfield L. 1992. The essential bacterial cell division protein 22. Schwarz H, Koch AL. 1995. Phase and This overall hypothesis is not really FtsZ is a GTPase. Nature 359: 254–6. electron microscopic observations of osmoti- new. Steed and Murray [36] concluded 4. Mukherjee A, Dai K, Lutkenhaus J. 1993. cally induced wrinkling and the role of endocy- in 1966: “These electron microscope Escherichia coli cell division protein FtsZ is a totic vesicles in the plasmolysis of the Gram- guanine nucleotide binding protein. Proc Natl negative cell wall. Microbiology 141: 3161–70. observations indicate that it is possible Acad Sci USA 90: 1053–7. 23. Sochacki KA, Shkel IA, Record MT, Weis- to regularly demonstrate an incontro- 5. Erickson HP. 2017. The discovery of the shaar JC. 2011. Protein diffusion in the vertible septum in these organisms prokaryotic cytoskeleton: 25th anniversary. periplasm of E. coli under osmotic stress. Biophys J 100: 22–31. (Gram-negative), which show only con- Mol Biol Cell 28: 357–8. 6. Osawa M, Anderson DE, Erickson HP. 24. Kennedy EP. 1982. Osmotic regulation and strictive divisions (when grown and 2008. Reconstitution of contractile FtsZ rings the biosynthesis of membrane-derived oligo- fixed under certain conditions). There in liposomes. Science 320: 792–4. saccharides in Escherichia coli. Proc Natl Acad Sci USA 79: 1092–5. is no escaping the suspicion that 7. Erickson HP, Taylor DW, Taylor KA, Bram- 25. Miller KJ, Kennedy EP, Reinhold VN. 1986. septum formation is the normal prelude hill D. 1996. Bacterial cell division protein FtsZ assembles into protofilament sheets and Osmotic adaptation by gram-negative bacte- to cell division in E. coli.” minirings, structural homologs of tubulin poly- ria: possible role for periplasmic oligosac- mers. Proc Natl Acad Sci USA 93: 519–23. charides. Science 231: 48–51. Record MT, Jr. 8. Lu C, Reedy M, Erickson HP. 2000. Straight 26. Cayley DS, Guttman HJ, . and curved conformations of FtsZ are regu- 2000 Biophysical characterization of Conclusions and outlook lated by GTP hydrolysis. J Bacteriol 182: changes in amounts and activity of Escher- 164–70. ichia coli cell and compartment water and turgor pressure in response to osmotic stress. what generates the 9. Osawa M, Anderson DE, Erickson HP. Biophys J 78: 1748–64. 2009. Curved FtsZ protofilaments generate constriction force? 27. Scherrer R, Gerhardt P. 1971. Molecular bending forces on liposome membranes. sieving by the Bacillus megaterium cell wall EMBO J 28: 3476–84. Coltharp et al. [13] suggested that FtsZ pf and protoplast. J Bacteriol 107: 718–35. 10. Erickson HP, Anderson DE, Osawa M. 28. Demchick P, Koch AL. 1996. The perme- bending would not be able to generate a 2010. FtsZ in bacterial cytokinesis: cytoskel- ability of the wall fabric of Escherichia coli and force sufficient to overcome the turgor eton and force generator all in one. Microbiol Bacillus subtilis. J Bacteriol 178: 768–73. Mol Biol Rev 74: 504–28. pressure, and proposed PG wall growth 29. Pasupuleti M, Schmidtchen A, Malmsten 11. Osawa M, Erickson HP. 2011. Inside-out Z M. 2012. Antimicrobial peptides: key compo- as the primary constriction force. How- rings constriction with and without GTP nents of the innate immune system. Crit Rev ever, if septation takes place within the hydrolysis. Mol Microbiol 81: 571–9. Biotechnol 32: 143–71. high osmolar environment of the cyto- 12. Housman M, Milam SL, Moore DA, Osawa 30. Foreman-Wykert AK, Weinrauch Y, Els- M et al. 2016. FtsZ protofilament curvature is plasm-periplasm, weaker forces would bach P, Weiss J. 1999. Cell-wall determi- the opposite of tubulin rings. Biochemistry 55: nants of the bactericidal action of group IIA suffice. It is possible that both FtsZ pf 4085–91. phospholipase A2 against Gram-positive bending and PG ingrowth contribute. 13. Coltharp C, Buss J, Plumer TM, Xiao J. bacteria. J Clin Invest 103: 715–21. Two recent studies have shown that FtsI 2016. Defining the rate-limiting processes of 31. Matias VR, Beveridge TJ. 2008. Lipoteichoic bacterial cytokinesis. Proc Natl Acad Sci USA acid is a major component of the Bacillus moves in circular paths around the Z 113: E1044–53. subtilis periplasm. J Bacteriol 190: 7414–8. ring, driven by patches of treadmilling 14. Coltharp C, Xiao J. 2017. Beyond force 32. Bhavsar AP, Brown ED. 2006. Cell wall FtsZ [47, 48]. Bisson-Filho et al. [48] generation: why is a dynamic ring of FtsZ assembly in Bacillus subtilis: how spirals and polymers essential for bacterial cytokinesis? spaces challenge paradigms. Mol Microbiol suggested that FtsZ filaments in the BioEssays 39: 1–1. 60: 1077–90. treadmilling patch may deform the 15. McDonald KL. 2014. Out with the old and in 33. Matias VR, Beveridge TJ. 2007. Cryo- membrane, and PG synthesis could with the new: rapid specimen preparation electron microscopy of cell division in Staph- reinforce the deformation on the other procedures for electron microscopy of sec- ylococcus aureus reveals a mid-zone be- tioned biological material. Protoplasma 251: tween nascent cross walls. Mol Microbiol 64: side. “Thus multiple sites of local 429–48. 195–206. deformation and coupled reinforcing 16. Matias VR, Al-Amoudi A, Dubochet J, 34. Zhou X, Halladin DK, Rojas ER, Koslover synthesis moving around the division Beveridge TJ. 2003. Cryo-transmission elec- EF et al. 2015. Bacterial division. Mechanical tron microscopy of frozen-hydrated sections crack propagation drives millisecond daugh- site would iteratively build the invagi- of Escherichia coli and Pseudomonas aeru- ter cell separation in Staphylococcus aureus. nating septum.” ginosa. J Bacteriol 185: 6112–8. Science 348: 574–8.

Bioessays 39: 1700045, ß 2017 WILEY Periodicals, Inc. 1700045 (9 of 10) H. P. Erickson Insights & Perspectives.....

35. Zuber B, Haenni M, Ribeiro T, Minnig K topology and functional periplasmic residues 46. Murray RG, Hall M, Thompson BG. 1983. et al. 2006. Granular layer in the periplasmic in the Escherichia coli cell division protein Cell division in Deinococcus radiodurans and space of gram-positive bacteria and fine FtsK. J Biol Chem 289: 23287–301. a method for displaying septa. Can J Micro- structures of Enterococcus gallinarum and 41. Heidrich C, Ursinus A, Berger J, Schwarz H biol 29: 1412–23. Streptococcus gordonii septa revealed by et al. 2002. Effects of multiple deletions of 47. Yang X, Lyu Z, Miguel A, McQuillen R et al. cryo-electron microscopy of vitreous sec- murein hydrolases on viability, septum cleav- 2017. GTPase activity-coupled treadmilling of tions. J Bacteriol 188: 6652–60. age, and sensitivity to large toxic molecules in the bacterial tubulin FtsZ organizes septal cell 36. Steed P, Murray RG. 1966. The cell wall and Escherichia coli. J Bacteriol 184: 6093–9. wall synthesis. Science 355: 744–7. cell division of gram-negative bacteria. Can J 42. Soderstrom B, Mirzadeh K, Toddo S, 48. Bisson-Filho AW, Hsu YP, Squyres GR, Microbiol 12: 263–70. von Heijne G et al. 2016. Coordinated Kuru E et al. 2017. Treadmilling by FtsZ 37. Judd EM, Comolli LR, Chen JC, Downing disassembly of the divisome complex in filaments drives peptidoglycan synthesis and KH et al. 2005. Distinct constrictive pro- Escherichia coli. Mol Microbiol 101: bacterial cell division. Science 355: 739–43. cesses, separated in time and space, divide 425–38. 49. Matias VR, Beveridge TJ. 2006. Native cell caulobacter inner and outer membranes. 43. Takeuchi S, DiLuzio WR, Weibel DB, wall organization shown by cryo-electron Hypotheses J Bacteriol 187: 6874–82. Whitesides GM. 2005. Controlling the shape microscopy confirms the existence of a 38. Burdett ID, Murray RG. 1974. Septum forma- of filamentous cells of Escherichia coli. Nano periplasmic space in Staphylococcus aureus. tion in Escherichia coli: characterization of Lett 5: 1819–23. J Bacteriol 188: 1011–21. septal structure and theeffects of antibioticson 44. Mannik J, Driessen R, Galajda P, Keymer 50. Erickson HP, Osawa M. 2017. FtsZ constric- cell division. J Bacteriol 119: 303–24. JE et al. 2009. Bacterial growth and motility in tion force curved protofilaments bending 39. Szwedziak P, Wang Q, Bharat TA, Tsim M sub-micron constrictions. Proc Natl Acad Sci membranes. Subcell Biochem 80: 139–160. et al. 2014. Architecture of the ring formed by USA 106: 14861–6. 51. Han S, Zaniewski RP, Marr ES, Lacey BM the tubulin homologue FtsZ in bacterial cell 45. Gupta RS. 2011. Origin of diderm (Gram- et al. 2010. Structural basis for effectiveness division. eLife 3: e04601. negative) bacteria: antibiotic selection pressure of siderophore-conjugated monocarbams 40. Berezuk AM, Goodyear M, Khursigara CM. rather than endosymbiosis likely led to the against clinically relevant strains of Pseudo- 2014. Site-directed fluorescence labeling evolution of bacterial cells with two membranes. monas aeruginosa. Proc Natl Acad Sci USA reveals a revised N-terminal membrane Antonie Van Leeuwenhoek 100:171–82. 107: 22002–7.

1700045 (10 of 10) Bioessays 39: 1700045, ß 2017 WILEY Periodicals, Inc.