Direct Observation of Staphylococcus aureus Cell Wall Digestion by Lysostaphin

The advent of Staphylococcus aureus strains that are resistantto virtually all antibiotics has increased the need for newantistaphylococcal agents. An example of such a potential therapeuticis lysostaphin, an enzyme that specifically cleaves the S. aureuspeptidoglycan, thereby lysing the bacteria. Here we trackedover time the structural and physical dynamics of single S.aureus cells exposed to lysostaphin, using atomic force microscopy.Topographic images of native cells revealed a smooth surfacemorphology decorated with concentric rings attributed to newlyformed peptidoglycan. Time-lapse images collected followingaddition of lysostaphin revealed major structural changes inthe form of cell swelling, splitting of the septum, and creationof nanoscale perforations. Notably, treatment of the cells withlysostaphin was also found to decrease the bacterial springconstant and the cell wall stiffness, demonstrating that structuralchanges were correlated with major differences in cell wallnanomechanical properties. We interpret these modificationsas resulting from the digestion of peptidoglycan by lysostaphin,eventually leading to the formation of osmotically fragile cells.This study provides new insight into the lytic activity of lysostaphinand offers promising prospects for the study of new antistaphylococcalagents.

INTRODUCTION

Top

INTRODUCTION

The gram-positive bacterium Staphylococcus aureus is a versatilepathogen capable of causing a wide range of human diseases (23,28). Peptidoglycan, the major cell wall components of S. aureus,is known to be the target of important antibiotics (22, 39).Among these, β-lactams (e.g., penicillins) exert theirantimicrobial activity by interfering with the enzymes (penicillinbinding proteins) specifically involved in peptidoglycan synthesis,while glycopeptides (e.g., vancomycin) bind with high affinityand specificity to peptidoglycan precursors, thereby preventingtheir incorporation into the bacterial cell wall (38). Currently,there is a constant need for new antistaphylococcal drugs owingto the development of antibiotic-resistant strains (4, 38, 39).An example of such a novel drug is lysostaphin, an enzyme secretedby Staphylococcus simulans which specifically cleaves the peptidoglycancross-linking pentaglycine bridges of S. aureus, thereby hydrolyzingthe cell wall and lysing the bacteria (7, 27, 33). Due to itsunique specificity, lysostaphin could have high potential inthe treatment of antibiotic-resistant staphylococcal infections(27). For instance, lysostaphin, given alone or in combinationwith vancomycin, was shown to be more effective in the treatmentof experimental methicillin-resistant S. aureus aortic valveendocarditis than vancomycin alone (9). Lysostaphin was alsoreported to be effective in the treatment of eye infectionscaused by S. aureus (12).

Knowledge of the molecular mechanisms underlying peptidoglycandigestion by lysostaphin is a key toward its efficient use inantistaphylococcal therapies. Early structural investigationsusing electron microscopy revealed that S. aureus cells exposedto lysostaphin showed perforations and more extensive damage,including the separation of walls from the plasma membranesand the disintegration of large sections of the walls (34).As a consequence, most cells exposed to the enzyme were renderedosmotically fragile. Among these, cells that retain the capacityto revert to normal staphylococci were designated spheroplasts,as opposed to protoplasts, in which the whole cell wall wasdigested away. Although powerful, electron microscopy techniquesare not suited for tracking dynamic, time-dependent processeson single cells, and they cannot probe the cell wall mechanicalproperties.

Recently, atomic force microscopy (AFM) (20, 30) has providednew opportunities for studying microbial cell walls at the single-celland single-molecule levels (14, 15). The technique is particularlywell-suited for visualizing the surfaces of single live cellswhile they grow (11, 31, 37), yielding information on cell wallassembly and dynamics that cannot be obtained with traditionalmicroscopies. For instance, Touhami et al. (37) were able tomonitor cell growth and division events in S. aureus using AFMcombined with thin-section transmission electron microscopy.Nanoscale holes were seen around the septal annulus at the onsetof division and were attributed to cell wall structures possessinghigh autolytic activity. After cell separation, concentric ringswere observed on the surface of the new cell wall and were suggestedto reflect newly formed peptidoglycan. AFM can also be usedto study the effect of drugs on microbial cell walls. In onesuch study, mycobacteria were observed prior to and after incubationwith isoniazid, ethionamide, ethambutol, or streptomycin (1).Upon drug treatment, major structural alterations were observedin the form of layered structures, striations, and porous morphologies,suggesting that they reflect the inhibition of the synthesisof three major cell wall constituents, i.e., mycolic acids,arabinans, and proteins.

Here, we used AFM to track the structural and physical dynamicsof single S. aureus cells exposed to lysostaphin. AFM imagesshow that the enzyme causes substantial swelling of the cells,favors splitting of the septum, and induces the creation ofnanoscale perforations. Consistent with this, nanomechanicalmeasurements demonstrate that these structural changes correlatewith a major decrease of bacterial turgor pressure and cellwall stiffness, reflecting the formation of osmotically fragilecells.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Bacterial strain and growth conditions. Staphylococcus aureus (methicillin-susceptible S. aureus strainATCC 25923; American Type Culture Collection [ATCC], Manassas,VA) was grown in Mueller-Hinter broth (Becton Dickinson, France)to mid-exponential growth phase (optical density at 600 nm =0.6) at 37°C with continuous stirring. Cells were centrifuged(7,000 rpm for 10 min) and washed twice by resuspension in phosphate-bufferedsaline (PBS) and centrifugation.

AFM measurements. AFM images and force-distance curves were recorded in PBS solution(10 mM PBS, 150 mM NaCl, pH 7.4) at room temperature, usinga Nanoscope IV multimode AFM (Veeco Metrology Group, Santa Barbara,CA). Cells were immobilized by mechanical trapping into porouspolycarbonate membranes (Millipore) with a pore size similarto the bacterial cell size (13, 26). After a concentrated cellsuspension was filtered, the filter was gently rinsed with PBS,carefully cut (1 cm by 1 cm), and attached to a steel samplepuck (Veeco Metrology Group) using a small piece of double-faceadhesive tape, and the mounted sample was transferred into theAFM liquid cell while avoiding dewetting. The spring constantsof the cantilevers were measured using the thermal noise method(Picoforce; Veeco Metrology Group), yielding a mean value of0.0255 ± 0.001 N/m. For lysostaphin experiments, PBSsolutions containing 16 µg/ml lysostaphin from S. simulans(Sigma, Belgium) were injected into the AFM liquid cell.

Mechanical properties were measured by recording arrays of 32-by-32force curves, using a maximum applied force of 1.75 to 2 nNto avoid sample damage. The bacterial spring constant, k_b, wasdetermined from the slope of the linear portions of the curvesfor raw deflection (nm) versus piezo displacement (nm) (5):k_b = (k_C·s)/(1 – s), where k_C is the spring constantof the cantilever and s the observed slope. The bacterial Youngmodulus was obtained by fitting the curves for force versusindentation using the Hertz model with a conical (tip)-plane(cell surface) geometry (24): F = {2Etan ${alpha}$ /[ ${pi}$ (1 – ${nu}$ ²)]}· ${delta}$ ²,where F is the force, ${delta}$ is the indentation depth, E is the Youngmodulus, ${nu}$ is the Poisson coefficient, and ${alpha}$ is the semi-top angleof the tip. This simple model is valid for elastic surfacesand does not take into account tip-surface adhesion. The latterassumption is reasonable since adhesion forces were negligiblein the present study. The mathematical analysis was performedwith an automatic Fortran C++ algorithm as described elsewhere(16, 17).

RESULTS AND DISCUSSION

Top

RESULTS AND DISCUSSION

High-resolution imaging of S. aureus. Cells immobilized in porous membranes were imaged in buffersolution. A representative low-resolution deflection image ofa single cell is shown in Fig. 1A. Due to the large curvatureof the specimen, height images had fairly poor resolution whiledeflection images were much more sensitive to the surface relief.Using small imaging forces ( $~$ 100 pN), images of the same areawere obtained repeatedly without detaching the cell or significantlyaltering the surface morphology. The cell, located at the centerof the image, was surrounded by artifactual features resultingfrom the contact between the AFM tip and the pore edges. Thecell surface was smooth and showed circular or concentric ringsenclosing a central depression. The high-resolution height imagerecorded on top of the cell, shown in Fig. 1B, reveals thatthe rings were separated by 20 to 50 nm. These features aresimilar to those reported earlier by Touhami et al. (37), whoobserved concentric rings on the newly formed cell walls ofS. aureus cells but not on older regions of the cell walls.The ring structures are also reminiscent of the circular patternsseen by several groups using electron microscopy techniques(2, 3, 6, 21, 22). We therefore conclude that the rings reflectstructural features of the peptidoglycan strands found on thesurface of the new wall after daughter cells have separated.As pointed out by Touhami et al. (37), the concentricity ofthe rings implies that the peptidoglycan strands are themselvesoriented in a similar fashion and are added or removed in anorderly manner. Note that a number of cells did not show anyring structures, presumably because their surfaces consistedof older cell wall material (Fig. 2, upper panels).

View larger version (53K):
[in this window]
[in a new window]

FIG. 1. Imaging of single S. aureus cells. (A) Low-resolution deflection image recorded in PBS, showing a single cell trapped in a pore of the polycarbonate membrane. (B) High-resolution height image recorded in the square region highlighted in panel A, together with a vertical cross-section taken along the white line.

View larger version (49K):
[in this window]
[in a new window]

FIG. 2. Imaging of single S. aureus cells following incubation with lysostaphin. A series of height (left) and deflection (right) images recorded in real time for a single cell prior to and after incubation with 16 µg/ml lysostaphin in PBS for 80, 210, and 260 min is shown. The lower panel compares vertical cross-sections taken at 0 min (continuous line) and 260 min (dashed line). White arrows show nanoscale perforations enlarging with time, while black arrows show splitting of the septum. Similar data were obtained with different cells from four independent cultures.

Lysostaphin induces major structural alterations. Peptidoglycan is an important component of the S. aureus cellwall, conferring strength and rigidity to the cell, allowinggrowth and division, maintaining cell shape, and protectingagainst osmotic lysis (8, 21). In S. aureus and other gram-positivebacteria, where the external membrane is absent, the multilayeredpeptidoglycan represents the limit of the cell and thereforethe site for exchanges and interactions with the outside environment(10). In particular, peptidoglycan is known to be an importanttarget for antimicrobial compounds (35), including lysostaphin,which hydrolyzes the pentapeptide cross-linkages (27).

With this in mind, we used in situ AFM imaging to observe structuralchanges on single S. aureus cells exposed to lysostaphin. Figure2 shows height and deflection images recorded for the same cellfollowing incubation with 16 µg/ml lysostaphin for 0,80, 210, and 260 min. As can be seen in the vertical cross-sections,a significant increase in cell size was noted over the courseof the experiment, with the cell height increasing by $~$ 100 nmafter 260 min. In agreement with early viability studies (25,34), we interpret such cell swelling as evidence for the formationof lysostaphin-induced osmotically fragile cells, resultingfrom peptidoglycan hydrolysis.

Moreover, progressive alteration of the cell surface structurewas clearly observed after lysostaphin addition. After 80 min,nanoscale perforations were seen, which were about 50 to 100nm in diameter and 25 to 75 nm in depth (Fig. 2). With time(e.g., after 260 min), these holes enlarged until they mergedtogether to form larger perforations. These structures are reminiscentof the small depressions seen at the onset of division (37)and may reflect so-called murosomes (22), i.e., regions of thecell wall having high autolytic activity. Interestingly, splittingof the septum was observed after 210 min (Fig. 2). We note thatthese features, i.e., perforations and splitting of the septum,were never observed when cells were imaged in PBS in the absenceof lysostaphin, even for prolonged periods of time (up to 240min), confirming that they result from the lytic activity ofthe enzyme. It is also worth mentioning that protoplasts inwhich the cell wall had been completely digested could neverbe observed by AFM, most likely because they were immediatelyruptured by the scanning tip. Besides localized surface modifications,we also found that lysostaphin increased the cell surface roughness(Fig. 3). Power spectral density analysis of the fast Fouriertransform of height images revealed that the root mean squareroughness for native and treated cells increased with the lengthscale to reach a plateau after 200 nm. Notably, the roughnesson 500-nm by 500-nm height images increased from 1.6 ±0.3 nm for untreated cells to 5.4 ± 1.2 nm for cellstreated for 200 min (Fig. 3). These results are qualitativelyconsistent with earlier electron microscopy investigations (34)showing progressive disintegration of the cell wall and separationfrom the plasma membrane after exposure to lysostaphin.

View larger version (41K):
[in this window]
[in a new window]

FIG. 3. Imaging of lysostaphin-treated cells at high resolution. A series of deflection images recorded on top of a single cell after incubation with 16 µg/ml lysostaphin for 20, 40, 80, 170, 210, and 260 min is shown. The lower panel shows the temporal evolution of the root mean square roughness measured on height images.

Taken together, our AFM images demonstrate that incubation ofS. aureus with lysostaphin causes substantial swelling of thecell body, favors splitting of the septum, induces nanoscaleperforations, and increases the cell surface roughness. Presumably,these time-dependent cell wall modifications result from peptidoglycandigestion, eventually leading to the formation of osmoticallyfragile spheroplasts.

Lysostaphin alters the cell wall mechanical properties. In view of the role of peptidoglycan in providing rigidity andprotection against osmotic lysis, we then addressed the pertinentquestion as to whether the observed structural changes werecorrelated with differences in cell wall mechanical properties.To this end, cells incubated with lysostaphin were probed usingnanoindentation measurements (Fig. 4 and 5). Figure 4A showstypical force-versus-piezo displacement curves obtained forthe polymer support and for a single cell at increasing incubationtimes. Consistent with the work of Gaboriaud et al. (18), thecurves recorded on the cell surface showed two domains, i.e.,a nonlinear domain at low loading forces followed by a linearone at high loading forces. From the shape of these curves,it can be seen that the cell wall was already substantiallysofter than the support and that lysostaphin caused a progressivesoftening of the cell wall.

View larger version (27K):
[in this window]
[in a new window]

FIG. 4. Mechanical properties of native and lysostaphin-treated cells. (A) Representative force-versus-displacement curves recorded on the polymer support (dashed line, S) and on a single cell prior to (0 min) and after (20, 60, 100, and 300 min) incubation with lysostaphin at 16 µg/ml. Open symbols are the raw data, while the solid lines show the theoretical fit (Hooke's law) used to extract the bacterial spring constant. (B) Evolution of the bacterial spring constant as a function of incubation time. Each data point represents the mean and standard deviation of 1,024 measurements obtained at different locations in the same cell. (C) Force-indentation curves obtained from the curves shown in panel A. Open symbols show the data, while the solid lines show the theoretical fits (Hertz model) used to extract the Young modulus values. (D) Evolution of the Young modulus as a function of incubation time. Each data point represents the mean and standard deviation of 1,024 measurements obtained at different locations in the same cell. Similar behaviors were observed for six different cells using two different tips.

View larger version (57K):
[in this window]
[in a new window]

FIG. 5. Mapping of cell surface elasticity. (A and D) High-resolution deflection images recorded for an S. aureus cell prior to and after incubation with lysostaphin for 80 min. (B and E) Elasticity maps (z-range = 3,000 kPa). (C and F) Distribution of elasticity values (n = 1,024 force curves) corresponding to the elasticity maps.

To provide quantitative information on the cell wall mechanicalproperties, two parameters were extracted from the force curves,i.e., the bacterial spring constant, which is related to theinner turgor pressure of the cell (5, 18, 19, 40), and the Youngmodulus, which reflects cell surface elasticity (17, 18, 29,36). Bacterial spring constant values, k_b, were determined fromthe slopes of the linear portions of the force-versus-piezodisplacement curves (Fig. 4A). The k_b value of native cellswas found to be 0.013 ± 0.001 N/m (Fig. 4B), which agreesreasonably well with values reported earlier for six Lactobacillusstrains (0.016 to 0.053 N/m) (32). Our slightly lower valuescould be due to the higher ionic strength that we used or todifferences in the cell wall organization. Notably, incubationwith lysostaphin caused a progressive decrease of k_b, from 0.013± 0.001 N/m to 0.006 ± 0.001 N/m after 100 minof exposure. This change may be attributed to a decrease ofthe inner turgor pressure of the cell (5, 18, 19, 40), in agreementwith the observed cell swelling (Fig. 2) and with the notionthat lysostaphin generates osmotically fragile cells.

Next, we estimated Young modulus values by converting the forcecurves into force-indentation curves and analyzing them usingthe Hertz theory (Fig. 4C). Figure 4D shows that within 100min, the Young modulus of the cells decreased from 1,764 ±218 kPa to 189 ± 49 kPa, demonstrating that lysostaphininduces a dramatic reduction of cell wall stiffness. This findingis fully consistent with the mode of action of the enzyme, whichcleaves the peptidoglycan cross-linking pentaglycine bridges.Also of note is the comparison between the level of indentationand cell swelling: for a constant force of 1 nN, the differencein indentation depth (for native cells versus cells treatedfor 300 min) was $~$ 100 nm, which is on the order of the heightincrease observed in the images (Fig. 2). Finally, arrays of32-by-32 force curves were recorded across the cell surfaceto map local variations of elasticity. As shown in Fig. 5, elasticitymaps on native and treated cells showed rather uniform contrasts,reflecting homogeneous distributions of elasticity. The homogeneouslydark contrast observed after 80 min confirmed that the entirecell wall was softened upon treatment with the enzyme.

In conclusion, we have observed, for the first time in situand on a nanoscale, the digestion of the S. aureus cell wallby lysostaphin, revealing that the enzyme causes substantialswelling of the cells and major alterations of their surfacestructure (septum splitting, nanoscale perforations, and increasedroughness). Using force measurements, we also showed that thesestructural changes correlate with major differences in mechanicalproperties, i.e., with a decrease of bacterial turgor pressureand of cell wall stiffness. This study shows that AFM is a promisingtool for exploring the organization and assembly of peptidoglycanand for investigating its interactions with enzymes and drugs.

ACKNOWLEDGMENTS

This work was supported by the National Foundation for ScientificResearch (FNRS); the Université Catholique de Louvain(Fonds Spéciaux de Recherche); the Federal Office forScientific, Technical and Cultural Affairs (InteruniversityPoles of Attraction Programme); and the Research Departmentof the Communauté Française de Belgique (ConcertedResearch Action). Y.F.D. is a research associate at the FNRS.

FOOTNOTES

* Corresponding author. Mailing address: Unité de Chimie des Interfaces, Université Catholique de Louvain, Croix du Sud 2/18, B-1348 Louvain-la-Neuve, Belgium. Phone: (32) 10 47 36 00. Fax: (32) 10 47 20 05. E-mail: yves.dufrene{at}uclouvain.be

Published ahead of print on 3 October 2008.

REFERENCES

Top