Native Cell Wall Organization Shown by Cryo-Electron Microscopy Confirms the Existence of a Periplasmic Space in Staphylococcus aureus

The current perception of the ultrastructure of gram-positivecell envelopes relies mainly on electron microscopy of thinsections and on sample preparation. Freezing of cells into amatrix of amorphous ice (i.e., vitrification) results in optimalspecimen preservation and allows the observation of cell envelopeboundary layers in their (frozen) hydrated state. In this report,cryo-transmission electron microscopy of frozen-hydrated sectionsof Staphylococcus aureus D₂C was used to examine cell envelopeorganization. A bipartite wall was positioned above the plasmamembrane and consisted of a 16-nm low-density inner wall zone(IWZ), followed by a 19-nm high-density outer wall zone (OWZ).Observation of plasmolyzed cells, which were used to artificiallyseparate the membrane from the wall, showed membrane vesicleswithin the space associated with the IWZ in native cells anda large gap between the membrane and OWZ, suggesting that theIWZ was devoid of a cross-linked polymeric cell wall network.Isolated wall fragments possessed only one zone of high density,with a constant level of density throughout their thickness,as was previously seen with the OWZs of intact cells. Theseresults strongly indicate that the IWZ represents a periplasmicspace, composed mostly of soluble low-density constituents confinedbetween the plasma membrane and OWZ, and that the OWZ representsthe peptidoglycan-teichoic acid cell wall network with its associatedproteins. Cell wall differentiation was also seen at the septumof dividing cells. Here, two high-density zones were sandwichedbetween three low-density zones. It appeared that the septumconsisted of an extension of the IWZ and OWZ from the outsideperipheral wall, plus a low-density middle zone that separatedadjacent septal cross walls, which could contribute to cellseparation during division.

INTRODUCTION

Top

Introduction

Staphylococcus aureus is a gram-positive pathogen often involvedin nosocomial infections and food-borne diseases, and thereis now growing concern about the spread of multidrug-resistantstrains of methicillin-resistant S. aureus (4, 25, 26, 31, 39,45). The staphylococcal cell wall plays an important role forthis organism's success, as it withstands tremendous turgorpressures throughout the cell cycle (ca. 20 to 30 atm), stronglyinteracts with the cell's external environment, particularlyin infection processes, and is intimately involved in cell division(1-3, 15, 21, 22, 33). Over the past two decades much knowledgehas been accumulated on the primary structure of staphylococcalcell wall components, i.e., peptidoglycan (3, 15, 45), (lipo)teichoicacids (34, 47, 48), and surface proteins (33, 43), but thereis still a lack of structural information at high resolutionon the spatial organization of this complex cell wall (4, 9,17, 46, 49).

As with other gram-positive bacteria, the envelope of S. aureusis commonly seen in conventional thin sections by electron microscopyas composed of a plasma membrane tightly bound by a thick andundifferentiated wall (Fig. 1). This simple architecture hasbeen attributed to the harsh procedures required to produceresin embeddings (4, 5). Chemical fixation, dehydration, heavymetal staining, and plastic embedding introduce many artifacts,which are aggravated by long treatment times. Rapid freezingused as a fast physical means of fixation, in combination withthe simultaneous dehydration and chemical fixation at low temperaturesin the freeze-substitution method, has provided better preservationof staphylococci for transmission electron microscopy (TEM)of thin sections (4, 44, 47, 49). Such thin sections of S. aureusprovide more structural detail than conventional sections doand show a tripartite cell wall surrounding the membrane. Ahighly stained thin inner zone (region 1) is enclosed by a low-contrastthicker region (region 2) and is followed by a more fibrouswall (region 3) (Fig. 2). This is similar to the layering seenin the freeze-substituted cell wall of Bacillus subtilis (16,17). However, this arrangement must be attributed to dissimilarwall architectures, since in B. subtilis the three-zoned wallreflects the way the cell wall turns over, whereas in S. aureuswall growth proceeds differently, as explained below.

View larger version (162K):
[in this window]
[in a new window]
FIG. 1. Thin section of the cell envelope of a conventionally embedded Staphylococcus aureus D₂C. The plasma membrane (PM) appears tightly bound by a thick and amorphous cell wall (CW). Bar, 50 nm.

View larger version (240K):
[in this window]
[in a new window]
FIG. 2. Thin section of the cell envelope of a freeze-substituted cell. A tripartite wall is seen enclosing the plasma membrane: 1, heavily stained inner thin zone; 2, intermediate translucent region; 3, outer highly stained fibrous surface. Region 1 is so heavily stained that visualization of the plasma membrane is difficult. The plasma membrane resides immediately below this heavily stained layer of the wall (17). Bar, 50 nm.

Cell elongation in the rod-shaped bacterium B. subtilis proceedswith the insertion of newer wall material on the inner faceof the cylindrical wall, which is pushed outwards and stretchedbecause of cellular turgor pressure (and becomes the middlezone). Older material on the outer surface is fragmented intofibers, due to the action of autolysins, and is shed (3, 20).New wall material progresses through the wall from inside tooutside as the wall matures and is finally excised. The translucentand fibrous walls seen in freeze-substituted B. subtilis areconsistent with the reactive sites for heavy metal binding withfully intact and partially degraded walls, respectively (16,17). In contrast, the staphylococcal wall grows predominantlywith the insertion of new wall material at division sites inthe middle of the cell, forming a cross wall or growing septum,which eventually fuses to develop a complete septum (1, 2, 15,21). In S. aureus, autolysins have hitherto been localized onlyat the septum, where they have been implicated in splittingthe cross wall into two leaflets, each leaflet forming one hemisphereof the newly generated daughter cells (15, 53). Since wall synthesisand (eventually) autolytic activity are associated with theseptum, new wall materials do not seem to be incorporated intothe cell wall of S. aureus by an inside-to-outside mechanism.

The tripartite wall displayed in freeze-substituted S. aureushas been interpreted as defining the segregation of componentswithin the wall. Freeze-substitution of isolated cell walls,in combination with immunogold labeling, has indicated thatthe fuzzy coat of S. aureus wall consists mostly of teichoicacids (47). The existence of a periplasmic space has been proposedto explain the heavily stained thin zone adjacent to the membrane(44, 47). Although structural preservation is better achievedby freeze-substitution, we believe specimen dehydration andheavy metal staining can still cause perturbations of the nativewall conformation.

In this report, cryo-TEM of frozen-hydrated sections is usedto further evaluate staphylococcal cell wall organization. Here,immobilization of structures at the molecular level is achievedby vitrifying samples within a matrix of amorphous ice, resultingin optimal structural preservation. Chemical fixatives and heavymetal stains are not used, and after vitrification, samplesare kept in their hydrated state for viewing. Vitrified specimensare cryo-sectioned, and cells in frozen sections are observedat low temperatures by cryo-TEM with the aid of phase contrast(12, 29). Production of frozen-hydrated sections is the mostdifficult step of the technique, but advances in the sectioningapparatus and high-pressure freezing have improved results (11,29, 37, 42, 54). This approach has recently resulted in optimalstructural preservation of bacteria and has allowed the observationof a periplasmic space in B. subtilis as a model gram-positivebacterium (30). Motivated by this result, we have investigatedthe polymeric network of the wall and the possible existenceof a periplasmic space in S. aureus as an example of an importantpathogen, especially since this gram-positive bacterium's modeof wall expansion and cell growth differs substantially fromthat of B. subtilis.

MATERIALS AND METHODS

Top

Materials and Methods

Bacterial strains and growth conditions. S. aureus Newman D₂C and B. subtilis 168 were grown at 37°Cto a mid-exponential growth phase (optical density at 470 nm,0.5 to 0.8) in either Trypticase soy broth (TSB) or TSB containing10% (wt/wt) glycerol (used as a cryoprotectant). Harvested cells(centrifuged at 6,000 x g for 5 min) were washed three timesin 50 mM HEPES (pH 7.0) or in 10 mM HEPES containing 10% (wt/wt)glycerol. The pellet of cells grown in TSB with glycerol andwashed in buffered glycerol was directly used for freezing.

Isolation of cell wall fragments. S. aureus grown in TSB was resuspended in 50 mM HEPES (pH 7.0)containing 50 mg/ml DNase and 50 mg/ml RNase, and 0.5-ml aliquotsof the pellet were added to 1.5 ml of 0.1-mm zirconia/silicabeads in 2-ml tubes. Bacteria were mechanically broken by two1-min agitations at 4,600 rpm in a Bead Beater (Biospec), withcooling of the tubes on ice between agitations. The supernatantwas centrifuged to remove intact cells and residual beads (3,000x g). The remaining supernatant, containing the cell wall fragments,was boiled in 4% (wt/vol) sodium dodecyl sulfate (SDS) for 2h (41). For the extraction of teichoic acids (TA), water-washedcell wall fragments were treated with 10% (wt/vol) trichloroaceticacid for 2 h at 60°C (19). Further, water-washed TA-extractedcell walls were incubated in a protease solution (500 µg/ml ${alpha}$ -chymotrypsin in 50 mM HEPES [pH 7.2] containing 10 mM CaCl₂)for 2 h at 30°C to remove any remaining protein associatedwith the wall. Protease-treated walls were then boiled in 4%SDS for 1 h to remove any remaining protease. All wall fragmentswere washed five more times in deionized water and three timesin 17.5% (wt/wt) glycerol in 10 mM HEPES (pH 7) for freezing.

Plasmolysis experiment. In order to plasmolyze D₂C cells, 1 ml of the pellet was resuspendedin 40 ml of either 10, 20, 25% (wt/vol) NaCl, or 10% NaCl containing10% (wt/vol) glucose, all prepared in 10 mM HEPES (pH 7.0).In all cases, the concentration of the salt was sufficient toprovide cryoprotection. Pellets obtained were used for freezing.

Freezing and sectioning of bacteria. Bacteria and cell wall fragments were frozen and sectioned aspreviously described (30). Briefly, pellets of samples for freezingwere drawn into a disposable plastic syringe, and bacteria orwalls were injected into copper tubes from the syringe and immediatelyvitrified using a Leica EM PACT high-pressure freezer (for detailson high-pressure freezing, see references 32 and 37). Frozensamples were sectioned in a Leica cryo-ultramicrotome and mountedon carbon-coated 1500-mesh copper grids.

Conventional embedding and freeze-substitution. Cells for conventional embedding were prepared as describedby Beveridge et al. (5). For freeze-substitution, copper tubeswith frozen cells were cut, under liquid nitrogen, into roughly2-mm-long pieces. These pieces were transferred to small vialscontaining 0.5 ml of freeze-substitution medium (2% osmium tetroxideand 2% uranyl acetate in anhydrous ethanol), and vials weretransferred to a Leica automatic freeze-substitution apparatus.Freeze-substitution was carried out at –90°C for 18h, at –60°C for 15 h, and at –30°C for 12h, and then the temperature was slowly ramped up to room temperatureover an 18-h period. Cells were washed in ethanol, embeddedin LR White (London Resins), thin sectioned at room temperature,and sections were poststained with uranyl acetate and lead citrate(17).

Cryo-TEM and chemical analyses. Grids containing the frozen-hydrated thin sections were mountedinto a Gatan cryo-holder for direct observation at –170°Cin a LEO 912AB energy-filtered cryo-TEM operating at 120 kV.Energy filtering improves image contrast by eliminating inelasticallyscattered electrons, which causes a blurring effect on micrographs.Zero-loss energy-filtered images were taken using a slow-scancharge-coupled device camera (Proscan) (1,024 x 1,024 pixels).Freeze-substituted cells were observed at room temperature,under the same operating conditions. Images were stored andanalyzed using analySIS software (SIS, Munster, Germany). Lengthmeasurements were done on the least deformed regions of frozen-hydratedcells (29, 30).

Phosphorous analysis of cell walls was performed using a VarianVista-pro radial induction-coupled plasma spectrometer witha Cetac ultrasonic nebulizer. Muramic acid, aspartic acid, andthreonine analyses were carried out by high-performance cation-exchangechromatography using a Beckman system Gold amino acid analyzer(7). Phosphorus, muramic acid, and the amino acids threonineand aspartic acid were used as markers for teichoic acid, peptidoglycan,and protein, respectively.

RESULTS

Top

Results

Freezing, sectioning, and cutting artifacts. S. aureus grown in the presence of 10% (wt/wt) glycerol wasconsistently vitrified by high-pressure freezing. Growth inthe presence of the cryoprotectant was chosen so as to minimizeosmotic effects on cells, and compared to growth without glycerol,there was little change in the growth rate (doubling time of35 min for glycerol versus doubling time of 30 min without glycerol).

Low-magnification micrographs of S. aureus showed the cuttingartifacts commonly associated with frozen-hydrated sections(Fig. 3). Unlike resin-embedded samples, which use the surfacetension of water to stretch and float sections during sectioning,frozen-hydrated sections are produced on a dry knife. Hence,tensional stresses created during sectioning remain on the sections,leaving artifacts on them. Cutting artifacts can neverthelessbe systematically taken into account, as they are all relatedto the cutting direction. These artifacts include knife marks,crevasses, compression along the cutting direction, and icecrystal contamination (for a detailed discussion on cuttingartifacts, see references 12, 29, and 30). Previous work onrod-shaped bacteria also faced an additional problem. Freezingof concentrated cell pellets in capillary copper tubes causedthe alignment of cells along the length of the tubes, preventingthe observation of longitudinal sections. Here, due to the sphericalshape of S. aureus, cells were found randomly aligned on thesections, and frequent observation of cell division was thereforepossible (Fig. 3).

View larger version (233K):
[in this window]
[in a new window]
FIG. 3. Energy-filtered image of a frozen-hydrated section of S. aureus D₂C at low magnification. Long arrows point to knife marks, white arrowheads to crevasses, double black arrowheads to compression in the cutting direction (this is why cells look more oblong than circular), and a short white arrow to a crack in the supporting film. White asterisks indicate cells possessing a septum, and short black arrows point to the septa. Bar, 500 nm.

Frozen-hydrated S. aureus and its cell wall organization. Frozen-hydrated sections of S. aureus showed cytoplasm filledwith evenly dispersed ribosomes without visible aggregationof DNA (Fig. 4). The cell wall was well preserved at the leastcompressed regions of cells (30), where the plasma membraneappeared bounded by a two-zoned wall at nonseptal regions (Fig.5A). A 16-nm inner wall zone (IWZ) showing low contrast precededa high-contrast 19-nm outer wall zone (OWZ) (Table 1 and Fig.5A). Since in frozen-hydrated samples, contrast is directlyproportional to density, this bipartite view of the staphylococcalwall represents regions of different distinct densities (12).Although a bipartite staphylococcal wall differs from the tripartitewall seen in freeze-substituted specimens, it resembles theB. subtilis wall seen in frozen-hydrated sections (30, 44, 47).

View larger version (244K):
[in this window]
[in a new window]
FIG. 4. Frozen-hydrated section of S. aureus D₂C at intermediate magnification. Ribosomes (seen as dense 20-nm particles) are evenly disbursed throughout the cytoplasm, which appears without visible aggregation of DNA. The cell at the middle of the figure is dividing and has a septum. Bar, 200 nm.

View larger version (115K):
[in this window]
[in a new window]
FIG. 5. Frozen-hydrated section at high magnification showing the S. aureus cell envelope. (A) At nonseptal regions, the plasma membrane (PM) is bound by a bipartite wall; a low-density inner wall zone (IWZ) precedes a high-density outer wall zone (OWZ). (B) At the septum, five different zones of alternating low (white arrows) and high (black arrows) densities are distinguished between the two membranes of the septum (arrowheads). Both images are shown at the same magnification. Bar, 50 nm.

View this table:
[in this window]
[in a new window]
TABLE 1. Measurements of structures and compartments of S. aureus D₂C^a

At the septum, five zones of alternating densities were observedbetween the membranes of the two daughter cells, with two high-densityzones appearing to be sandwiched by three low-density zones(Fig. 5B). This septal profile is markedly different comparedto both conventional embeddings and freeze-substituted resultsof S. aureus, which show a heavily stained midline at the crosswall (4, 15, 44, 46). Nevertheless, the two zones next to themembrane at the septum seemed to be an extension of the bipartitewall of the cell wall envelope (cf. Fig. 5A and B). This willbe discussed in more detail below.

Plasmolysis experiments. In order to help distinguish the constitution of each wall zoneseen in the S. aureus envelope, cells were osmotically shockedto artificially separate the membrane from the walls. For thispurpose, cells grown without a cryoprotectant were subjectedto a high-osmolarity solution. Solutions of 20 and 25% NaClgave the best plasmolysis results, but the high concentrationof salts made it difficult to obtain sections without pronouncedcutting artifacts, necessary for higher image resolution (notshown). Plasmolysis was still achieved after incubation in asolution of 10% NaCl containing 10% glucose, and the presenceof glucose seemed to improve section quality. It is possiblethat the glucose acted as both an osmolyte and cryoprotectant.The high osmolarities necessary to plasmolyze S. aureus reflectthe high osmotolerance of this organism, which is able to growover a wide range of osmotic conditions (from low osmolaritiesup to a NaCl concentration of 3.5 M [36, 38]). Even with theseharsher plasmolysis conditions, S. aureus did not undergo plasmolysisas readily as seen previously in B. subtilis (30). PlasmolyzedS. aureus produced an OWZ that was separated further from theprotoplast, which shrank as water was removed from its cytoplasm(Fig. 6A and Table 1). At higher magnification, the IWZ wasshown to increase in thickness particularly in certain areasof cells, where membrane vesicles were found between the plasmamembrane and OWZ (Fig. 6B). Since plasmolysis causes a considerablereduction of the total area of the membrane associated withthe contraction of the protoplast, part of the membrane seemsto have been excised, generating vesicles outside the protoplast(these resemble the mesosome bodies found after plasmolysisof Bacillus megaterium [52]). The observation of vesicles withinthe IWZ after plasmolysis is similar to the observations inprevious B. subtilis experiments (30) and implies that thiswall zone is composed mostly of soluble low-density constituents(i.e., it possesses soft materials that allow vesicles to expandwithin it). In addition, this easily achieved, wide separationof the OWZ from the protoplast is consistent with a deformablespace, which would expand upon reduction of turgor pressureduring plasmolysis (Fig. 6C). Since many membrane-associatedlipidated components interact strongly with the wall matrixin gram-positive bacteria (3, 33), it is possible that suchmembrane-associated enzymes (27), lipoteichoic acids (34), andlipoproteins (43) could aid in removing portions of membranefrom the protoplast during shrinking, resulting in the formationof vesicles. Interestingly, in areas lacking vesicles, the thicknessof the IWZ remained similar to the thickness in unplasmolyzedcells. This could be related to a tighter bonding between theOWZ and plasma membrane by larger amounts of lipidated wallcomponents (i.e., membrane-bound wall enzymes, lipoteichoicacids, and lipoproteins; Fig. 6D). The thickness of the OWZincreased compared to that of unshocked cells, as expected,since with lower turgor pressures, wall networks expand dueto reduction of stress (Fig. 6C and D and Table 1).

View larger version (226K):
[in this window]
[in a new window]
FIG. 6. Plasmolyzed S. aureus D₂C cells. (A) Cells show variable separations between the plasma membrane and OWZ. (B) Separation between the OWZ and protoplast is larger at certain regions of the envelope, where membrane vesicles are seen (arrows). (C) Vesicles retaining high-density materials are confined between the membrane and OWZ, which appears thicker than in intact cells (arrow). (D) In areas of the envelope without vesicles, the thickness of the IWZ is similar to that of the IWZs in unplasmolyzed cells, while the OWZ is thicker. Bars, 500 nm (A), 150 nm (B), and 50 nm (C and D).

Cell wall fragments. To corroborate the plasmolysis data on the constitution of theIWZ and OWZ, crude cell wall fragments were isolated from mechanicallybroken cells using a Bead Beater. These fragments were boiledin SDS to remove cytoplasmic contaminants and noncovalentlybound wall proteins (WP), and they possessed the major wallconstituents (i.e., peptidoglycan, teichoic acids, and covalentlybound wall proteins; Table 2). As previously seen in B. subtilis,sections of cell wall fragments revealed them to be monopartitein contrast to the bipartite view seen on cells (Fig. 7A and B)(30). S. aureus wall fragments consisted of a single 33-nm zoneshowing contrast similar to that of the OWZs of cell envelopeson intact cells (Fig. 7B and Table 1). As these fragments consistedof isolated walls (Fig. 7B and Table 2), this provides strongevidence that the OWZ seen on cells corresponds to the cellwall polymeric network. The IWZ must be composed mostly of solublecomponents that were washed away during cell wall isolationand SDS boiling. The increase in thickness of wall fragmentscompared to the thickness of walls of intact cells is consistentwith the removal of turgor pressure after mechanical breakageof cells (Table 1), as this causes the relaxation of the peptidoglycanfabric in the absence of stress. This relaxation also increasesthe density of the (now) unstretched polymers, since it resultsin smaller spacings between the polymers. As indicated by theplasmolysis experiment, these data strongly suggest that theIWZ is composed of highly deformable and less substantial wallconstituents, implying the presence of a periplasmic space confinedby both the plasma membrane and OWZ. As a consequence, the outerwall must represent the peptidoglycan-teichoic acid-proteincell wall network.

View this table:
[in this window]
[in a new window]
TABLE 2. Chemical analyses of S. aureus D₂C cell wall fragments

View larger version (156K):
[in this window]
[in a new window]
FIG.7. S. aureus cell wall fragments. (A) Cell wall fragments obtained from mechanically broken and SDS-boiled fragments, containing all major wall components (i.e., peptidoglycan [PG], TA, and WP), only partially retain the shape of cells. The white arrowhead points to ice crystal contamination. (B) Same specimen as in panel A but at a higher magnification shows only one wall zone similar to the OWZs of intact cells. Black arrowheads point to cracks in the supporting film. (C) TA-extracted walls (possessing PG and WP) have even more pliable forms than cell wall fragments do. Long black arrows point to knife marks. (D) At high magnification, TA-extracted walls also show only one wall zone and lower contrast than cell wall fragments; the white arrow points to a septum coming off an outside wall. (E) WP-and-TA-extracted wall fragments (comprising predominantly PG) lack a defined shape. (F) At high magnification, wall fragments stripped of WP and TA show the least contrast and are the most pliable of all wall fragments. Bars, 500 nm (A, C, and E) and 50 nm (B, D, and F).

Sections of isolated walls also showed that fragments retainedonly partially the shape of cells (Fig. 7A). Cell walls typicallyprovide a rigid exoskeleton responsible for cell shape in bacteria,but this requirement may not be so strict in these cocci, asa spherical geometry would represent the most energeticallyfavored shape. Moreover, a more flexible wall could aid S. aureusto better respond to variations in environmental osmolarities,a feature very important for this highly osmotolerant pathogen(22, 50).

Removal of teichoic acids and proteins from wall fragments. Wall fragments were further treated so as to provide informationon cell wall organization. Teichoic acid was extracted fromwall fragments using trichloroacetic acid, which removed 94%of TA on the basis of phosphorous analysis (Table 2). TA isresponsible for most of the wall's anionic charge due to itsabundance and the presence of phosphate groups at both the linkageand repeat units of the polymer (10, 13, 34). TA-extracted wallswere composed of predominantly peptidoglycan and wall-boundproteins (Table 2) and, similar to untreated cell wall fragments,had a monopartite appearance (Fig. 7D). Yet, they were morepliable and little cellular shape was left (cf. Fig. 7A and C).

TA-extracted walls were subsequently incubated with a proteaseto remove most of the covalently bound wall proteins, and thesewalls were essentially composed of only peptidoglycan (Table2). WP can be abundant components of walls, particularly ingram-positive cocci. Protein A alone accounts for 7% of thewall content (dry weight) in S. aureus (3, 40). Reduction inthe concentrations of aspartic acid and threonine, used as proteinmarkers in the cell wall, showed that 90% of wall-bound proteinswere removed from these wall fragments (Table 2). Removal ofproteins from TA-extracted walls had a significant influenceon the shape and rigidity of walls, and WP-and-TA-extractedwall fragments were still monopartite, with very pliable forms(Fig. 7E and F).

The progressive loss of shape and rigidity after extractionof TA and WP point to an important role for both on wall rigidityin staphylococci (Fig. 7A, B, and C). It is possible that ionicand hydrophobic interactions among peptidoglycan, TA, and WPare especially important in the direction perpendicular to thewall's thickness and increase the wall's stiffness. Since secondarywall components interdigitate with peptidoglycan polymers asthey span the wall, their removal could result in wall expansionand loss of rigidity. Interestingly, wall thickness was thesame in all three fractions of wall fragments (Table 1). Althoughthe presence of TA and proteins seemed to aid in restrictingwall bending, they both had negligible influence on the expansionof the wall thickness. This indicates that secondary polymerbonding is not important in maintaining a constant wall thicknessin cells and that the bonding between peptidoglycan polymersis more important. Yet, large forces due to turgor pressurecan extend peptidoglycan components laterally, thereby causingcontraction of wall thickness. Moreover, the continuous monopartiteview seen with all wall fragments provides evidence that TAand WP are not segregated to a particular region of the wall.

Density tracings of the cell envelope and cell walls. Densitometry plots provided information on the distributionof mass through the thickness of walls. Density tracing of thecell envelope showed that the OWZ possessed a relatively constantdensity along the wall thickness, with similar densities atboth the wall's inner and outer surfaces (Fig. 8A). The OWZalso appeared very similar in the three types of wall fragments(Fig. 8B to D), providing more evidence that the OWZ is thepeptidoglycan-TA-WP cell wall matrix and that TA and WP areevenly distributed throughout the wall thickness.

View larger version (51K):
[in this window]
[in a new window]
FIG. 8. High-magnification images of S. aureus D₂C with the corresponding density tracings. (A) Cell envelope. (B) Cell wall fragments. (C) Teichoic acid-extracted walls. (D) Protease-digested and teichoic acid-extracted walls. Bar, 50 nm.

The constant density across the wall thickness of this gram-positivebacterium seems to be inconsistent with the concept that peptidoglycanfibers are more stretched at the inner zone because of turgorpressure and their closer proximity to the membrane (17). TheS. aureus sacculus is composed of very short peptides (six disaccharideunits long [6]) and has one of the highest degrees of cross-linkingamong bacterial peptidoglycans, on the order of 80 to 90% (14).Since this peptidoglycan is a highly cross-linked network ofsmall building blocks, this organism has pliable walls (observedin frozen sections of isolated walls; Fig. 7). Turgor is highin this organism (about 25 to 30 atm) and should stretch thepeptidoglycan fibers throughout the thickness of the cell wall.Stretching must affect "relaxed" polymer density by decreasingit. If the cell wall had different in situ densities or if certainregions of the wall were stretched differently by turgor pressure,our technique would distinguish it. No different densities wereseen as indicated by the constant levels of density within thewall observed in the density tracing of the cell envelope (Fig.8A). If the inner region of the wall was the one bearing mostof the stress caused by turgor pressure, one would expect agradual increase in density from inside to outside of the wall,which is not observed. Therefore, the hypothesis that the IWZrepresents stretched wall is not valid.

The densitometry plots of the OWZ and wall fragments also showedsimilar density values on both sides of the wall (Fig. 8A to D),which adds to the notion that turnover of walls in S. aureusis mainly localized to the septum. Here, new wall polymers arelaid down as the septum grows, and once the septum is completedand daughter cell separation commences, autolytic action splitsthe septum to release the two new daughter cells (Fig. 8A to D).In other systems (such as in the preexisting cylindrical wallsof B. subtilis) that require the input of new polymers and theshedding of old polymers by turnover, there exists an inside-to-outsidemode of wall growth, which is reflected by a decay of wall densityfrom inside to outside (Fig. 9) (30).

View larger version (137K):
[in this window]
[in a new window]
FIG. 9. High-magnification image of the B. subtilis envelope with corresponding density tracing. Progressive decrease in the wall density from inside (arrow) to outside (arrowhead) is consistent with the concept of wall turnover. This was not seem in the S. aureus cell wall (Fig. 8A). Bar, 50 nm.

Septation. At the S. aureus septum, where most of the wall synthesis takesplace, density tracings showed five alternating zones of lowand high densities (Fig. 10). The structure of the two zonesadjacent to the membrane at the septum compared well to thewall organization of the cell envelope (cf. Fig. 5A and 10).In both cases, a low-density zone was seen above the membranewith a higher density zone above it. The low-density zone nextto the membrane appeared to be an extension of the IWZ (Fig.5B). The high-density zones appeared to be similar to the OWZ,showing a nearly constant level of density throughout the wallthickness (Fig. 10).

View larger version (170K):
[in this window]
[in a new window]
FIG. 10. Density tracing of an S. aureus septum at the center of a cell. Two high-density zones are seen sandwiched between three zones of low density. Bar, 50 nm.

In sections of conventionally embedded and freeze-substitutedS. aureus, a highly stained midline is seen at the septum (15,44). This midline has been attributed to a splitting system,presumably involved in cell separation and consisting of concentricallyarranged rings of wall substance (15). Our current cryo-TEMresults showed that the middle zone of the cross wall possesseslow density and has no discernible structure inside. The low-densitymiddle zone of the cross wall resembled the other two low-densityregions (IWZs) and could hold similar components, such as penicillin-bindingproteins (PBPs) and other enzymes involved in the wall synthesisand hydrolysis.

DISCUSSION

Top

Discussion

In this paper we present the cell wall organization of S. aureusas revealed by cryo-TEM of frozen-hydrated sections. The plasmamembrane was found bound by a bipartite wall, consisting ofa low-density 16-nm IWZ, followed by a 19-nm OWZ of higher density.This arrangement was found around cells and extended into theseptum, where another low-density zone appeared between thetwo high-density zones. Our data strongly suggest that, as withB. subtilis (30), the IWZ is a periplasmic space, whereas theOWZ consists of the peptidoglycan-teichoic acid cell wall networkwith its associated proteins. Although suggested before (44,47), this is the first time strong evidence supports the existenceof a periplasmic space in S. aureus. This report has shown thefollowing. (i) The IWZ possessed lower density than the OWZ.(ii) Membrane blebs were found in plasmolyzed cells betweenthe protoplast and the OWZ. (iii) Cell wall fragments possessedonly one zone of high density similar to that of the OWZ. (iv)The density profile of OWZ was very similar to the profilesof all isolated wall fragments. (v) Wall fragments were thickerthan the OWZ after reduction and removal of turgor in plasmolyzedcells and cell wall fragments.

By analogy with gram-negative bacteria, the periplasmic spacerepresents an extraprotoplasmic compartment confined betweenthe plasma membrane and an outer structure (outer membrane versusa peptidoglycan-teichoic acid-protein network), and its constituentperiplasm is composed mostly of highly deformable, low-density,soluble components (18, 30). Such a compartment in gram-positivebacteria would represent an advantageous strategy used by thesebacteria to provide sufficient space for proper functioningand folding of membrane-bound enzymes involved in the synthesisand transport of macromolecules, such as wall components, exoenzymes,and secreted proteins, situated away from the highly anioniccell wall matrix (27, 28). It is intriguing how gram-positivebacteria maintain a separation between the membrane and thewall fabric, since turgor is of the order of 20 to 30 atm inthese bacteria (3, 22). Although the exact mechanism for maintainingtheir periplasmic space has yet to be elucidated, our plasmolysisresults showed stronger bonding between the membrane and wallfabric in certain areas of the envelope, implying that membrane-boundpolymers, which interact with the wall (i.e., PBPs, lipoteichoicacids, lipoproteins [3, 27, 34, 43]), could form a scaffoldingthat keeps the wall away from the protoplast.

The frozen-hydrated profile of the S. aureus cell envelope appearsmarkedly different from the tripartite wall seen by freeze-substitution,but one should remember that contrast is achieved in differentways by each technique. Freeze-substitution uses heavy metalstains to make visible structures with metal binding ability,whereas in frozen-hydrated sections, contrast is proportionalto the mass of the structural component. Accordingly, frozen-hydratedsections show a low-density periplasmic space in the cell envelope,which appears heavily stained in freeze-substituted specimensbecause of the high heavy metal affinity of the periplasm, similarto what is observed in gram-negative bacteria (29). The thicknessof the frozen-hydrated cell wall matrix compared reasonablywell with the translucent wall revealed by freeze-substitution(region 2 [Fig. 2]), and these walls show constant mass andconstant metal binding throughout their thickness. However,preservation of the plasma membrane and periplasm seems remarkablybetter achieved in frozen-hydrated sections, which shows bilayeredmembranes and a much wider periplasmic space. Admittedly, itis still possible that the cryoprotectant used in this studyand required to vitrify samples could have had subtle effectson the sizes of wall structures. The only feature not seen infrozen-hydrated S. aureus walls was the highly stained surfacefringe of freeze-substituted cells which has been suggestedto consist mostly of teichoic acids (47). This identificationis based on the results of immunolabeling experiments that labeledmost of the wall outer surface using nanogold particles, whichgive limited resolution due to particle size (5 to 10 nm [47]).Our results give no indication of segregation of teichoic acidsin the wall fabric, as wall fragments possessed the same thicknessbefore and after extraction of teichoic acids and density tracingsof the OWZ and all wall fragments showed an almost constantdensity over the thickness of walls. It is probable that whateveris on the surface of the wall possesses high metal binding abilityand lower local density than the outer medium (10% glycerolin our frozen-hydrated sections) and does not provide enoughcontrast for cryo-TEM (which uses differential mass for imaging).

Despite the optimal structural preservation assured by vitrification,the molecular organization of the wall components within thewall fabric could not be discerned in our frozen-hydrated sections.Their visualization is certainly a very challenging task becauseof the gel-like and heterogeneous nature of the wall, attestedto by many high-resolution techniques (e.g., X-ray scattering,TEM of isolated walls, and atomic force microscopy [8, 24, 46]).Accordingly, many models have been proposed to describe thetertiary structure of peptidoglycan. In the most traditionalview, growing glycan strands (bearing stem peptides and orientedparallel to the membrane) are incorporated into the wall, forminga multilayered peptidoglycan thick enough to withstand turgorpressures of the order of 25 atm (23, 24, 51). A recent "scaffold"model of murein architecture, which depicts glycan strands asoriented perpendicularly to the plasma membrane, has been proposedfor the S. aureus peptidoglycan tertiary structure, and thedegree of cross-linking seems to reflect experimental data better(9). The frozen-hydrated data on S. aureus cannot distinguishbetween the "parallel to membrane"' (horizontal) and "scaffolding"(vertical) models.

Division of S. aureus could be observed in frozen-hydrated sections.S. aureus septa possessed five distinct zones of different densities,but the two zones adjacent to the membrane seemed to be an extensionof the envelope. The middle zone of the cross wall had low densitywith no distinguishing structures within it and could, likethe periplasmic space, consist of soluble substances. Presumably,the middle zone has a high heavy metal binding ability, whichwould result in a heavily stained midline in both conventionaland freeze-substituted sections. The use of a heavy metal labelspecific for PBPs has in fact increased the thickness of thismidline, suggesting the presence of PBPs at this region of theseptum (35), even though most PBPs must be associated with themembrane. Further work is necessary to unravel the nature ofthe middle zone of the septum and the complex mechanisms involvedin cross wall formation of S. aureus.

A view of the intriguing complexity of the staphylococcal cellwall has emerged with the use of cryo-TEM of frozen-hydratedsections. The existence of a periplasmic space and of a complexarchitecture at the septum in this pathogen points to elaboratecellular mechanisms of division and cell wall synthesis. Certainlymore studies will be necessary to reveal the structural basisinvolved in the maintenance of a periplasmic space and in crosswall synthesis in gram-positive bacteria, but cryo-TEM of frozen-hydratedsections will be essential to help find these answers.

ACKNOWLEDGMENTS

We thank A. Saxena of our laboratory for technical assistance.

This work was supported by a Natural Science and EngineeringResearch Council of Canada (NSERC) Discovery grant to T.J.B.V.R.F.M. was the recipient of a Ph.D. scholarship from CNPq/Brazilduring part of this study. Microscopy was performed in the NSERCGuelph Regional Integrated Imaging Facility (GRIIF), which ispartially funded by an NSERC Major Facility Access grant toT.J.B.

FOOTNOTES

* Corresponding author. Mailing address: Department of Molecular and Cellular Biology, College of Biological Science, University of Guelph, Guelph, Ontario, Canada N1G 2W1. Phone: (519) 824-4120, ext. 58904. Fax: (519) 837-1802. E-mail: vmatias{at}uoguelph.ca.

REFERENCES

Top