Electron microscope study of the rod-to-coccus shape change in a temperature-sensitive rod- mutant of Bacillus subtilis.

The changes in cell morphology of Bacillus subtilis rodB during a temperature shift from 20 to 42 degrees C, in the absence of added anions, are described. At 20 degrees C the organisms grow as rods but gradually become spherical in shape when placed at 42 degrees C. The shape change is initiated by an increase in diameter at the cell equator, resulting in a bulged morphology, which is further modified to the morphology of a coccus. This change may involve a modification of the pattern of normal cylindrical extension such that incorporation of newly synthesized wall leads only to increase in diameter, perhaps from a growth zone of limited extent. The pattern of surface growth was followed by reconstructing the sequence of cross wall formation and pole construction in rods grown at 20 degrees C and in organisms incubated at 42 degrees C for 75 and 150 min. In thin section, wall forming the septum and nascent poles can be distinguished from the surface distal to the division site by the presence of raised tears, perhaps analogous to the wall bands of streptococci. By using an analog rotation technique involving the three-dimensional reconstruction of cells by mathematical rotation of axial thin sections about their longitudinal axis, it is shown that the proportion of septal wall increases during the shape change. In the coccal forms, all surface growth may arise from septal growth sites.     

Study of cycle of cell wall assembly in Streptococcus faecalis by three-dimensional reconstructions of thin sections of cells.

A new ultrastructural method was used to study rounds of envelope synthesis that occur in Streptococcus faecalis in "growth zones" found between pairs of naturally occurring surface markers. The technique consists of producing three-dimensional reconstructions of these growth zones from the mathematical rotation, about a central axis, of measurements taken from central, longitudinal thin sections of cells. A cycle of exponential-phase envelope growth was then simulated by arranging a series of these reconstructions in increasing order of the amount of peripheral wall surface area or the amount of cell volume that each was calculated to contain. Using this simulated cycle of growth, the geometry of a single growth zone during a round of synthesis was studied. Based on this analysis, a model was developed for the assembly of the cell wall of S. faecalis. The model states that new cell wall surface is synthesized by the regulated flow of essentially two channels of cell wall precursors into a single growth zone. One channel of precursors would be involved in the assembly of a bilayered cross wall that would proceed at a fairly constant rate until the cross wall closes. The second channel of precursors would be involved in the separation of the bilayered cross wall into two segments of peripheral wall. These precursors would intercalate into and thicken the separating layers of the cross wall. The flow of precursors through this channel would be progressively reduced through a cycle. These decreases, when coupled with internal hydrostatic pressure, apparently would result in the enlarging peripheral wall becoming increasingly more curved and would also promote cell division by reducing the total amount of cell wall that must be assembled in order for septation to occur.     

Three-dimensional reconstruction of whole cells of Streptococcus faecalis from thin sections of cells.

A new ultrastructural technique has been developed to study the geometry of cell wall assembly in Streptococcus faecalis, which is believed to occur between pairs of raised bands located on the organism's surface. Three-dimensional reconstructions of these new regions of envelope growth are produced from the mathematical rotation (around a central axis) of various measurements taken from central, longitudinal thin sections of cells. These reconstructions can be used to calculate the surface area and volume of the septal and peripheral walls that were supposedly present in any given cell before sectioning. In an accompanying paper, it is shown how such surface and volume estimations, coupled with other measurements of length, thickness, and curvature, can be used to characterize a cycle of envelope growth in this organism. The validity of the assumptions used to reconstruct cells by rotation and the possible sources of error in using this technique are discussed.     

Differences in the formation of poles of Enterococcus and Bacillus.

The pole of Enterococcus hirae (Streptococcus faecium) is more pointed than that of Bacillus subtilis; i.e. the pole of the former is prolate and the latter is oblate. Both species form their poles by constructing annular additions on the inside surface. In both cases, the thick septum starts to split from the outside before the septum is complete. Physiochemical considerations dictate that the peptidoglycan must be unstretched as laid down. However, it later becomes stressed and may stretch to increase its surface area or to change its shape. Our earlier analysis for B. subtilis demonstrated that, without the addition of new peptidoglycan, the nascent wall is stretched after it is externalized to 1.51 times the original area. The wall of partially formed poles that is already exteriorized continues to deform with further development. For E. hirae, Higgins & Shockman's measurements showed that the completed pole has a surface area 2.18 times larger than a completed septal disk and the wall changes shape very little after exteriorization. A model is presented here for the streptococcus in which the septal wall does not increase its surface area on exteriorization either by expansion or by murein insertion. Instead, the septal wall as it is split and exteriorized twists to become oblique, increasing the inner radius of the incomplete septum. In consequence of this rotation, extra layers of peptidoglycan are added to the inside face of the developing septum. This additional murein forms the more pointed pole shape for E. hirae. This "split-and-splay" model thus refines and extends the surface stress theory of E. hirae developed a decade ago by proposing a source of the extra wall needed for the formation of its prolate, more pointed, pole.     

Cell wall assembly in Bacillus subtilis: partial conservation of polar wall material and the effect of growth conditions on the pattern of incorporation of new material at the polar caps

The use of phage SP50 as marker for cell wall containing teichoic acid in Bacillus subtilis showed clear differences in the rates at which new wall material becomes exposed at polar and cylindrical regions of the wall, though the poles were not completely conserved. Following transition from phosphate limitation to conditions that permitted synthesis of teichoic acid, old polar caps fairly rapidly incorporated enough teichoic acid to permit phage binding. Electron microscopy suggested that the new receptor material spread towards the tip of the pole from cylindrical wall so that phages bound to an increasing proportion of the pole area until only the tip lacked receptor. Eventually, receptor was present over the whole polar surface. Direct electron microscopic staining of bacteria collected during transitions between magnesium and phosphorus limitations showed that new material was incorporated at the inner surface of polar wall and later became exposed at the outer surface by removal of overlying older wall. The apparent partial conservation of the pole reflected a slower degradation of the overlying outer wall at the pole than at the cylindrical surface, the rate being graded towards the tip of the pole. The relative proportions of the new wall material incorporated into polar and cylindrical regions differed in bacteria undergoing transitions that were accompanied by upshift or downshift in growth rate. These differences can be explained on the basis that growth rate affected the rate of synthesis of cylindrical but not septal wall.     

A Three-dimensional model reconstruction of pole assembly in Bacillus subtilis

In the rod-shaped bacterium Bacillus subtilis, new polar surfaces arise at division through the centripetal synthesis of a centrally located cross-wall. Subsequently, the cross-wall, analogous to a flat annulus, is converted into two inner layers of polar wall as the daughter cells separate. The junction of polar and cylindrical wall is marked by the presence of raised tears or wall bands formed by the splitting apart of the cross-wall at its base. New polar wall formed in this manner accounts for about 15% of the total surface area. The sequence of pole formation has been simulated by means of a generalized conic section based upon the mathematical rotation of a parabola about its longitudinal axis. Four basic measurements describe the stages of pole formation with reference to polar surface area: the equatorial diameter at the wall bands (D_max), the division furrow (D_min), the horizontal distance (h) from the centre of the cross-wall to D_max and the curvature of the nascent polar surfaces. These four parameters were found to yield a close fit to measurements of polar size and shape derived from electron micrographs of cell poles in sectioned organisms. Calculations of pole curvature suggest that both the initial separation of the cross-wall and separation of the daughter cells may occur very rapidly.     

Sites of Cellular Autolysis in Lactobacillus acidophilus

Ultrastructural changes which occur during cellular autolysis of Lactobacillus acidophilus strain 63AM Gasser in 0.05 M citrate buffer, pH 5.0, were examined. Early in the process, randomly distributed electron-dense patches were seen on the wall surface, along with an accompanying eversion of mesosomes. Later, after a loss of about 20% of the initial cellular turbidity, dissolution from the outside of nascent cross walls was seen. This observation was related to the normal process of cell separation. After this stage, short lengths of the cylindrical portion of the wall appeared to be completely removed in a random manner over the entire surface. This dissolution produced gaps in the wall which allowed the extrusion of membrane and cytoplasm. Although membrane was usually extruded through one major, polar, subpolar, or septal site, other secondary points of membrane extrusion were also frequently seen in the same cell section.     

Control of wall band splitting in Streptococcus faecalis

Computer reconstructions of 659 and 1325 whole mounted, shadowed cells, randomly chosen from cultures of Streptococcus faecalis undergoing balanced growth and doubling in mass every 83 min and 30 min, respectively, were used to analyse the cell cycle. The size limits and duration of phases of the cell cycle were estimated by applying a method previously described by the authors, details of which are given here to allow others to use the method. Deeply constricted cells whose primary septal radius, Rs, was less than or equal to 0.18 micron were considered as belonging to an E-phase ending the cell cycle. The statistical parameters of these E-phase cells were used to calculate the mean and coefficient of variation of dividing cells. These latter values, in turn, predicted the moments of the total population well enough so that the method's assumptions were judged adequately satisfied. Therefore, the method was considered applicable to other phases and sub-phases of the cell cycle of these two cultures. The E-phase cells were further classified as having either 0, 1 or 2 secondary growth zones, allowing us to calculate the percentage of newborn cells without growth zones. In the slow-growing cells, 69% of the cells arose with no growth zone. On the other hand, in more rapidly growing cells 16% of the cells or less arose with no growth zone. Our calculations showed that they could exist without a growth zone for only 2 and 0.1 min, respectively. We also classified cells as possessing a 'birth site' if the volume between the two daughter bands was greater than 0, but less than 0.06 micron3. From the statistical properties of such cells with new growth zones, the mean pole time, W, was estimated. We also estimated W from the size of cells in E-phase. The major conclusion is that the pole time is only slightly greater than the mass doubling time at both growth rates. Since DNA synthesis in S. faecalis takes longer (C = 50 to 52 min) than the mass doubling time in rich medium (30 min), a new round of chromosome replication must be initiated before the old round of synthesis is completed (dichotomous replication). Consequently, wall band splitting and initiation of chromosome replication do not occur simultaneously. It was also concluded that the cell initiates wall band splitting, resulting in pole formation and cell division, when the growth zones cannot function rapidly enough to allow the increase of surface area required to accommodate continuing production of cytoplasm.     

Reinitiation of Cell Wall Growth After Threonine Starvation of Streptococcus faecalis

Cultures of Streptococcus faecalis ATCC 9790 were starved of threonine for 10 hr and then allowed to reinitiate growth in a fresh complete medium. On regrowth, culture turbidity began to increase within 10 min, but the ability of cells to autolyze did not begin to increase until after 30 min. Ultrastructural studies of regrowth of the initially thick-walled cells showed, at about 30 min, centripetal linear extension of new thin cross wall. This was followed, at about 40 min, by a notching, splitting, and peeling apart of the base of the cross wall. After this, extension of new thin peripheral wall from the nascent cross wall appeared to push old thick wall toward the poles. After the first cell division, asymmetric cells with one initial generation thick-walled pole and one second generation thin-walled pole were seen. After two divisions, thick-walled hemispheres were still seen, suggesting conservation of old wall during this time. A small fraction of the initial cell population exhibited aberrations and difficulties in reinitiating linear wall extension and were useful in the establishment of a model for the reinitiation of linear wall extension.     

Control of cell morphogenesis in bacteria: two distinct ways to make a rod-shaped cell

Cell shape in most eubacteria is maintained by a tough external peptidoglycan cell wall. Recently, cell shape determining proteins of the MreB family were shown to form helical, actin-like cables in the cell. We used a fluorescent derivative of the antibiotic vancomycin as a probe for nascent peptidoglycan synthesis in unfixed cells of various Gram-positive bacteria. In the rod-shaped bacterium B. subtilis, synthesis of the cylindrical part of the cell wall occurs in a helical pattern governed by an MreB homolog, Mbl. However, a few rod-shaped bacteria have no MreB system. Here, a rod-like shape can be achieved by a completely different mechanism based on use of polar growth zones derived from the division machinery. These results provide insights into the diverse molecular strategies used by bacteria to control their cellular morphology, as well as suggesting ways in which these strategies may impact on growth rates and cell envelope structure.     

The role of surface stress in the morphology of microbes

The shapes of many prokaryotes can be understood by the assumption that the cell wall expands in response to tension created by the osmotically derived hydrostatic pressure. Different organisms have different shapes because wall growth takes place in different regions. A previous paper (Koch et al., 1981 a) considered the simplest case of prokaryotic growth, i.e. that of Streptococcus faecium. In the present paper, an elaboration of this theory is applied to two further cases - the more perfectly spherical cocci and the rod-shaped bacteria. These cases are more complex mathematically, because growth over a considerable fraction of the surface must be considered. Such diffuse growth cannot be treated analytically, but can be simulated on a computer or handled by geometric arguments. The spherical form of the cocci may result from either diffuse growth over their entire external surface, or from zonal growth in which the addition of new material only occurs in the immediate vicinity of the splitting septum. In the zonal model, it must be assumed that the least amount of previously laid down septal peptidoglycan consistent with wall growth is reworked in the formation of the new external wall. For Gram-positive rods, where the body of the rod is truly cylindrical, three kinds of growth zones are required: (1) the inward edge of the ingrowing septum, (2) the junction of septum and nascent pole, and (3) the cylindrical walls. Two modes for cylindrical elongation ara possible: (a) new wall is added in one or a few narrow annular zones, or (b) new wall material is added continuously all over the innermost surface and the outer layer is degraded. It is shown that the latter case applies to Bacillus subtilis. Also summarized in this paper are results, developed in more detail elsewhere, concerning the morphology of fusiform bacteria, Gram-negative rods and the hyphal tips of fungi.     

Cardiolipin domains in Bacillus subtilis marburg membranes

Recently, use of the cardiolipin (CL)-specific fluorescent dye 10-N-nonyl-acridine orange (NAO) revealed CL-rich domains in the Escherichia coli membrane (E. Mileykovskaya and W. Dowhan, J. Bacteriol. 182: 1172-1175, 2000). Staining of Bacillus subtilis cells with NAO showed that there were green fluorescence domains in the septal regions and at the poles. These fluorescence domains were scarcely detectable in exponentially growing cells of the clsA-disrupted mutant lacking detectable CL. In sporulating cells with a wild-type lipid composition, fluorescence domains were observed in the polar septa and on the engulfment and forespore membranes. Both in the clsA-disrupted mutant and in a mutant with disruptions in all three of the paralogous genes (clsA, ywjE, and ywiE) for CL synthase, these domains did not vanish but appeared later, after sporulation initiation. A red shift in the fluorescence due to stacking of two dye molecules and the lipid composition suggested that a small amount of CL was present in sporulating cells of the mutants. Mass spectrometry analyses revealed the presence of CL in these mutant cells. At a later stage during sporulation of the mutants the frequency of heat-resistant cells that could form colonies after heat treatment was lower. The frequency of sporulation of these cells at 24 h after sporulation initiation was 30 to 50% of the frequency of the wild type. These results indicate that CL-rich domains are present in the polar septal membrane and in the engulfment and forespore membranes during the sporulation phase even in a B. subtilis mutant with disruptions in all three paralogous genes, as well as in the membranes of the medial septa and at the poles during the exponential growth phase of wild-type cells. The results further suggest that the CL-rich domains in the polar septal membrane and engulfment and forespore membranes are involved in sporulation.     

Effect of macromolecular synthesis and lytic capacity on surface growth of Streptococcus faecalis.

Exposure of exponential-phase cultures of Streptococcus faecalis to any of three inhibitors of protein synthesis was accompanied by an increase in the average distance that the cross wall extended into the cytoplasm. This resulted in: (i) an increase in the average surface area of the cross wall (Sa) and (ii) septation occurring in the envelope growth sites that were much smaller than the controls. However, although at the concentrations used, all three antibiotics inhibited protein synthesis and autolytic capacity to the same extent and with the same kinetics, cells treated with these agents showed large differences in the rate at which Sa values increased above those of the untreated cells. The largest increases in Sa were observed in cells that synthesized the least amount of cytoplasmic macromolecules (deoxyribonucleic acid, plus ribonucleic acid, plus protein). The observations were interpreted in terms of a model in which a decreased lytic capacity reduces the rate of splitting of the nascent cross wall into two layers of peripheral wall, preferentially using wall precursors to close open cross walls. However, the extent to which centripetal growth occurs would be inversely related to the rate at which cytoplasmic macromolecules are synthesized. In contrast, inhibition of deoxyribonucleic acid synthesis was accompanied by decreased extension of the leading edge of the cross wall into the cytoplasm, thus antagonizing septation. These findings are discussed in relation to the normal cell division cycle of S. faecalis.     

Cell wall replication in Escherichia coli, studies by immunofluorescence and immunoelectron microscopy

Beachey, Edwin H. (National Institute of Allergy and Infectious Diseases, Bethesda, Md.), and Roger M. Cole. Cell wall replication in Escherichia coli, studied by immunofluorescence and immunoelectron microscopy. J. Bacteriol. 92:1245-1251. 1966.-Cell wall components of four different strains of Escherichia coli (B; B/r, try(-); O5; and O86:B7) were labeled with homologous fluorescent antibodies (FLG); the way the label was dispersed on further growth in media free of antibody was followed by fluorescence microscopy. Fluorescence diminished diffusely along longitudinal wall but remained bright at cell poles (or cross walls); newly formed cross walls did not fluoresce. In agreement, reverse labeling (preincubation in unlabeled antibody, followed by staining on the slide with homologous FLG) showed that stainability of longitudinal wall increased gradually and diffusely with increased time of incubation, whereas polar wall remained nonfluorescent or stained only faintly; newly formed poles (or cross walls), on the other hand, stained brightly. These observations were confirmed by electron microscopy, after immunoferritin labeling. Although the mode of cross-wall formation remained unclear, our findings refuted reported ideas of segmental or polar growth of cell wall in E. coli and supported the idea of wall replication by diffuse intercalation, as described for Salmonella.     

Cryo-electron microscopy of cell division in Staphylococcus aureus reveals a mid-zone between nascent cross walls

Cryo-electron microscopy of frozen-hydrated thin sections permits the observation of the real distribution of mass in biological specimens allowing the native structure of bacteria to be seen, including the natural orientation of their surface layers. Here, we use this approach to study the fine ultrastructure of the division site, or septum, of Staphylococcus aureus D₂C. Frozen-hydrated sections revealed a differentiated cell wall at the septum, showing two high-density regions sandwiched between three low-density zones. The two zones adjacent to the membrane appeared as an extension of the periplasmic space seen in this organism's cell envelope and showed no distinguishing structures within them. Immediately next to these were higher-density zones that corresponded to nascent cross walls of the septum. Unexpectedly, a rather broad low-density zone was seen separating cross walls in the septum. This mid-zone of low density appeared inflated and without visible structures in isolated cell walls, which showed only the high-density zones of the septum. Here, we suggest that frozen-hydrated thin sections have captured a highly fragile septal region, the mid-zone, which results from the dynamic action of autolysis and actively separates daughter cells during division. The two zones next to the membranes are periplasmic spaces. Immediately next to these are the growing cross walls composed of peptidoglycan, teichoic acid and protein.     

Promiscuous targeting of Bacillus subtilis cell division protein DivIVA to division sites in Escherichia coli and fission yeast

The Bacillus subtilis divIVA gene encodes a coiled-coil protein that shows weak similarity to eukaryotic tropomyosins. The protein is targeted to the sites of cell division and mature cell poles where, in B.subtilis, it controls the site specificity of cell division. Although clear homologues of DivIVA are present only in Gram-positive bacteria, and its role in division site selection is not conserved in the Gram-negative bacterium, Escherichia coli, a DivIVA-green fluorescent protein (GFP) fusion was targeted accurately to division sites and retained at the cell pole in this organism. Remarkably, the same fusion protein was also targeted to nascent division sites and growth zones in the fission yeast Schizosaccharomyces pombe, mimicking the localization of the endogenous tropomyosin-like cell division protein Cdc8p, and F-actin. The results show that a targeting signal for division sites is conserved across the eukaryote-prokaryote divide.     

Localization of new peptidoglycan at poles in Bacillus mycoides, a member of the Bacillus cereus group

Bacillus mycoides is a sporogenic Gram-positive soil bacillus of the B. cereus group. This bacillus, which forms hyphal colonies, is composed of cells connected in filaments that make up bundles and turn clock- or counterclockwise depending on the strain. A thick peptidoglycan wall gives the rod cells of these bacilli strength and shape. One approach used to study peptidoglycan neoformation in Gram positives exploits the binding properties of antibiotics such as vancomycin and ramoplanin to nascent peptidoglycan, whose localization in the cell is monitored by means of a fluorescent tag. When we treated B. mycoides strains with BODIPY-vancomycin, we found the expected accumulation of fluorescence at the midcell septa and localization along the cell sidewall in small foci distributed quite uniformly. Intense fluorescence was also observed at the poles of many cells, more clearly visible at the outer edges of the cell chains. The unusual abundance of peptidoglycan intermediates at the cell poles after cell separation suggests that the construction process of this structure is different from that of B. subtilis, in which the free poles are rarely reactive to vancomycin.     

DivIVA is required for polar growth in the MreB-lacking rod-shaped actinomycete Corynebacterium glutamicum

The actinomycete Corynebacterium glutamicum grows as rod-shaped cells by zonal peptidoglycan synthesis at the cell poles. In this bacterium, experimental depletion of the polar DivIVA protein (DivIVA(Cg)) resulted in the inhibition of polar growth; consequently, these cells exhibited a coccoid morphology. This result demonstrated that DivIVA is required for cell elongation and the acquisition of a rod shape. DivIVA from Streptomyces or Mycobacterium localized to the cell poles of DivIVA(Cg)-depleted C. glutamicum and restored polar peptidoglycan synthesis, in contrast to DivIVA proteins from Bacillus subtilis or Streptococcus pneumoniae, which localized at the septum of C. glutamicum. This confirmed that DivIVAs from actinomycetes are involved in polarized cell growth. DivIVA(Cg) localized at the septum after cell wall synthesis had started and the nucleoids had already segregated, suggesting that in C. glutamicum DivIVA is not involved in cell division or chromosome segregation.     

Atomic force microscopy of cell growth and division in Staphylococcus aureus

The growth and division of Staphylococcus aureus was monitored by atomic force microscopy (AFM) and thin-section transmission electron microscopy (TEM). A good correlation of the structural events of division was found using the two microscopies, and AFM was able to provide new additional information. AFM was performed under water, ensuring that all structures were in the hydrated condition. Sequential images on the same structure revealed progressive changes to surfaces, suggesting the cells were growing while images were being taken. Using AFM small depressions were seen around the septal annulus at the onset of division that could be attributed to so-called murosomes (Giesbrecht et al., Arch. Microbiol. 141:315-324, 1985). The new cell wall formed from the cross wall (i.e., completed septum) after cell separation and possessed concentric surface rings and a central depression; these structures could be correlated to a midline of reactive material in the developing septum that was seen by TEM. The older wall, that which was not derived from a newly formed cross wall, was partitioned into two different surface zones, smooth and gel-like zones, with different adhesive properties that could be attributed to cell wall turnover. The new and old wall topographies are equated to possible peptidoglycan arrangements, but no conclusion can be made regarding the planar or scaffolding models.     

Insertion and fate of the cell wall in Bacillus subtilis

Cell wall assembly was studied in autolysin-deficient and -sufficient strains of Bacillus subtilis. Two independent probes, one for peptidoglycan and the other for surface-accessible teichoic acid, were employed to monitor cell surface changes during growth. Cell walls were specifically labeled with N-acetyl-D-[3H]glucosamine, and after growth, autoradiographs were prepared for both cell types. The locations of silver grains revealed that label was progressively lost from numerous sites on the cell cylinders, whereas label was retained on the cell poles, even after several generations. In the autolysin-deficient and chain-forming strain, it was found that the distance between densely labeled poles approximately doubled after each generation of growth. In the autolysin-sufficient strain, it was found that the numbers of labeled cell poles remained nearly constant for several generations, supporting the premise that completed septa and poles are largely conserved during growth. Fluorescein-conjugated concanavalin A was also used to determine the distribution of alpha-D-glucosylated teichoic acid on the surfaces of growing cells. Strains with temperature-sensitive phosphoglucomutase were used because in these mutants, glycosylation of cell wall teichoic acids can be controlled by temperature shifts. When the bacteria were grown at 45 degrees C, which stops the glucosylation of teichoic acid, the cells gradually lost their ability to bind concanavalin A on their cylindrical surfaces, but they retained concanavalin A-reactive sites on their poles. Discrete areas on the cylinder, defined by the binding of fluorescent concanavalin A, were absent when the synthesis of glucosylated teichoic acid was inhibited during growth for several generations at the nonpermissive temperature. When the mutant was shifted from a nonpermissive to a permissive temperature, all areas of the cylinder became able to bind the labeled concanavalin A after about one-half generation. Old cell poles were able to bind the lectin after nearly one generation at the permissive temperature, showing that new wall synthesis does occur in the cell poles, although it occurs slowly. These data, based on both qualitative and quantitative experiments, support a model for cell wall assembly in B. subtilis, in which cylinders elongate by inside-to-outside growth, with degradation of the stress-bearing old wall in wild-type organisms. Loss of wall material, by turnover, from many sites on the cylinder may be necessary for intercalation of new wall and normal length extension. Poles tend to retain their wall components during division and are turned over much more slowly.