Vol. 174, No. 7JOURNAL OF BACTERIOLOGY, Apr. 1992, p. 2241-22520021-9193/92/072241-12$02.00/0Copyright © 1992, American Society for Microbiology

Fan-Shaped Ejections of Regularly Arranged Murosomes Involvedin Penicillin-Induced Death of Staphylococci

PETER GIESBRECHT,* THOMAS KERSTEN, AND JORG WECKERobert Koch-Institute of the Federal Health Office, Nordufer 20, D-1000 Berlin 65, Germany

Received 23 October 1991/Accepted 22 January 1992

Electron microscopic research into the murosomes of staphylococci has shown that the number of murosomesinvolved in penicillin-induced death varies depending on the experimental conditions employed. With 0.1 ,ugof penicillin G per ml, only 1 of a total of about 20 murosomes, the "killing murosome," completely perforatedthe pressure-stabilized peripheral cell wall during a three-step process. This strictly localized event was mainlyattributed to a mechanical effect being comparable to the process of aneurysm formation. Wall perforation wasalso considered to mark the very moment of penicillin-induced death ("nonlytic killing event"), whilebacteriolysis started only postmortem. By varying the osmolarity of the growth medium, the number ofmurosomes involved in penicillin-induced killing increased considerably, which resulted in the ejection of afan-shaped row of murosomes at the second division plane. These data are compatible with the finding that, inuntreated or chloramphenicol-treated staphylococci, the activation of the murosomes resulted in (i) theformation of regularly arranged "blebs" on the cell surface, containing traces of disintegrated wall material,and (ii) the subsequent liberation of the murosomes lying underneath, leaving behind their former sites in theperipheral wall as a row of regularly arranged "pores" in every division plane. The number, distribution, andpositioning of these blebs corresponded with those of the pores and the original murosomes. The significanceof wall autolysins liberated from the first division plane for penicillin-induced wall perforation at the seconddivision plane is discussed.

The essential components of the mechanism of penicillin-induced death of staphylococci have been elucidated inprevious reports (6, 15). After the reaction of the ,-lactamantibiotic with the penicillin-binding proteins (2, 20), it is notthe traditional alteration in the activity of wall autolysins orin cell wall synthesis (22, 24) which initiates penicillin-induced death (19) but rather a special morphogenetic errorincorporated into the cell wall during cross-wall formation.This killing process involves novel extraplasmatic wall or-ganelles, the murosomes, which in untreated staphylococciregularly punch minute "pores" into the peripheral cell wallabove the completed cross wall, thereby initiating cell sep-aration. In the presence of penicillin, however, this separa-tion process is no longer preceded by the assembly ofsufficient cross-wall material in the second division plane sothat, at the onset of the pore-punching process at this now"unprotected" site of the peripheral wall, the integrity of thewall is destroyed and the cell dies (6, 15). Another strictlylocalized lytic wall process was recently reported to be thevery reason for penicillin-induced death in Escherichia coli(12).Murosomes of staphylococci are not only detected in the

peripheral wall above the nascent cross walls but also,although in rare cases, in the peripheral wall between twoconsecutive division planes. This finding led to the specula-tion that murosomes might be transported in some way fromone division plane to the site of the next cross-wall initiation(6). Furthermore, since only a single murosome-induced wallperforation could be observed in the second division plane inthe presence of penicillin, it seemed that only one murosomewas involved in the killing process in staphylococci (6, 15).

Since we consider a detailed knowledge of the mechanism

* Corresponding author.

of penicillin-induced death to be an important prerequisitefor further improving ,B-lactam antibiotic therapy, we ana-lyzed in more detail the sequence of events taking place inthe presence of penicillin, especially at the very momentwhen murosome-induced perforation of the peripheral cellwall initiated the killing of staphylococci. The new datapresented here has led to a better mechanistic view of thesequence of events leading to penicillin-induced death.

MATERIALS AND METHODS

Bacterial strains and growth conditions. Staphylococcusaureus strains SG 511 Berlin and SG 511 Pelzer (from theculture collection of the Robert Koch-Institute, Berlin) wereemployed. The basal growth medium was peptone watercontaining 2.5% (wt/vol) Bacto Peptone (Difco) and 0.5%(wt/vol) NaCl, pH 7.2. In parallel experiments, this basalmedium was altered by increasing the NaCl concentration to3.0% (wt/vol). Cells grown for 16 h in peptone water at 37°Cwere diluted 1:15 into fresh medium. Following growth for 2h at 37°C in a rotary shaker (first exponential-phase culture),an aliquot of culture was diluted into fresh medium as beforeand grown for an additional 2.5 h (second exponential-phaseculture). The cell densities of both cultures were adjustedwith fresh medium to an optical density at 578 nm of 0.2 to0.3. The second cultures were transferred to flasks contain-ing the appropriate amounts of NaCl (0.5 or 3.0% [wt/vol]final concentration). Immediately after the sodium chloridewas dissolved, penicillin G (potassium salt; Sigma) wasadded to a final concentration of 0.1 ,ug/ml.

Referring to earlier findings that chloramphenicol inter-feres with the activity of the autolytic wall enzymes (10, 25),exponential-phase staphylococci were cultivated in the pres-ence of 20 to 40 ,ug of chloramphenicol per ml (Boehringer,Mannheim, Germany) for 8 to 14 h. After this treatment, the

2241

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

2242 GIESBRECHT ET AL. J. BACTERIOL.

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

PENICILLIN-INDUCED DEATH 2243

bacteria were treated for 4 h with 1 to 1,000 ,ug of egg whitelysozyme per ml (Serva, Heidelberg, Germany), dissolved in0.1 M sodium acetate buffer (pH 5) in order to induce areactivation of autolytic wall enzymes (28).

Preparation of samples for transmission and scanning elec-tron microscopy. Staphylococci were fixed overnight at 4°Cwith 2.5% (wt/vol) glutaraldehyde (Serva) in 0.1 M cacody-late buffer, pH 7.2. After being washed with cacodylatebuffer, the cells were postfixed for 1 h at room temperaturewith 1.5% (wt/vol) osmium tetroxide (Serva) and 1.65%(wt/vol) potassium bichromate in 0.1 M cacodylate buffer.The fixed samples were washed again and poststained foranother hour with 0.5% (wt/vol) uranyl acetate. The fixedand stained bacteria were dispersed in 2% (wt/vol) agar(Noble agar; Difco). Small blocks of agar were dehydrated inethanol and finally embedded in LR White plastic resin(Science Services, Munich, Germany). Thin sections werecut with a Reichert OM U3 ultramicrotome and viewed withan Elmiskop I (Siemens) or a Philips 400 electron micro-scope.For freeze-etching, unfixed bacteria were quickly frozen

without cryoprotectants in propane which was supercooledby liquid nitrogen. In order to get more freeze sections thanfreeze fractures, we dehydrated, in some cases (see Fig. 1B),unfixed staphylococci by a brief exposure to 3 M sucrose justbefore they were frozen (9). Replicas were prepared in aBalzers BA 300 freeze-etching device with a turbomolecularpump and electron beam evaporator. Replicas were cleanedin 40% (wt/vol) chromic acid.For scanning electron microscopy, a drop of the staphy-

lococcal culture containing 0.5 or 3.0% (wt/vol) NaCl wasput on a glass slide which had been pretreated with poly-L-lysine (Sero-Med, Berlin, Germany) to ensure a firm adher-ence of the bacteria and the released murosomes. The cellswere allowed to sediment for 20 min and were then fixedovernight at 4°C with 2.5% (wt/vol) glutaraldehyde in 0.1 Mcacodylate buffer, pH 7.2. Samples were washed, dehy-drated successively in ethanol and acetone, dried at thecritical point (model CPD 030; Balzers, Balzers, Liechten-stein), and sputtered with an 18-nm gold layer (SEM coatingunit E 5100; Polaron Equipment Ltd., Watford, UnitedKingdom). Examinations were carried out with a Hitachi H5010 electron microscope equipped with a scanning device.

RESULTS

In untreated staphylococci, cell separation of daughtercells was indicated after cross-wall completion by the ap-pearance of spherical structures or pores in the peripheralcell wall directly adjacent to the splitting system of the crosswall (7) (Fig. 1A). They were especially distinct in therelatively thick cross walls of staphylococci in their station-ary phase of growth. We have named these structuresmurosomes and have described them in detail in previousreports (3, 6). In thin sections, murosomes could never be

observed unless the cross wall was completed. However,after application of our special freeze-etching procedure (9),we were able to show in longitudinal fractures through thenascent cross wall that such structures already existed at thebeginning of cross-wall formation; they were disk shapedand regularly arranged in a row (Fig. 1B).Murosomes or pores appeared very rapidly at the onset of

cell separation. In order to retard this process, we firstdeactivated the autolytic wall system of staphylococci bychloramphenicol treatment (10, 25) and slowly reactivated itafterwards by the addition of a cationic protein (lysozyme atpH 5) which is known to act in staphylococci mainly as anactivator of autolytic wall enzymes rather than as a wall-lyticenzyme (28, 31). By this procedure, the row of pores in theperipheral cell wall could be better demonstrated (Fig. 1C);it was comparable to the row of murosomes in the peripheralcell wall of untreated staphylococci (Fig. 1B). However,after chloramphenicol treatment, these pores often proved tobe arranged in a zigzag line or sometimes even in two distinctparallel rows (Fig. 1C). Apparently, the process of poreformation started with the development of regularly ar-ranged "blebs" on the cell surface of the next division plane(Fig. 2A), the number of blebs corresponding to the numberof pores in untreated or chloramphenicol-treated staphylo-cocci. The result of such an activation process was therelease of the murosomes, 30 to 40 nm in diameter (Fig. 2B),leaving behind the characteristic spherical pores in theextremely thick peripheral cell wall. Such pores still resem-bled the murosomes in shape but exhibited distinctly largerdiameters, 50 to 60 nm. A prolonged treatment with lyso-zyme at pH 5 resulted in an overall disintegration of theperipheral wall.As already described in a previous report (6), the forma-

tion of the first cross wall after the addition of penicillin tookplace in a manner similar to what is seen with untreated orchloramphenicol-treated staphylococci. However, most ofthe cells were no longer able to assemble the splittingsystem, and often only an incomplete and deformed crosswall consisting of rather "loose" wall material was built upin the first division plane (Fig. 3A). Nevertheless, suchstaphylococci started cell separation normally, i.e., via theformation of lytic murosomal sites in their peripheral cellwalls, despite the fact that their cross walls were incomplete.Longitudinal sections through the cross wall revealed aboutas many lytic sites (Fig. 3B) as there were pores on thesurface of untreated or chloramphenicol-treated staphylo-cocci.

In the second division plane, cross-wall formation wasagain initiated by the deposition of murosomes, but theassembly of the newly formed cross-wall material continuedmainly in the first division plane (6). Consequently, in thesecond division plane, only a rather thin basal cross-walllayer was formed beneath the murosome (Fig. 4A) or noteven traces of cross-wall material could be detected here butonly "naked" murosomes (Fig. 4B). Even then, murosome-

FIG. 1. Thin section and freeze-etched surfaces of S. aureus SG511 Pelzer. (A) Thin section of an untreated cell at the final step ofcross-wall formation (18 h of growth), showing the splitting system of the cross wall (small arrows) and a pore (large arrow) in the peripheralcell wall directly above the closed cross wall. Bar, 200 nm. (B) Freeze-etching of an untreated cell at the very beginning of cross-wallformation, showing disk-shaped structures (arrows) in the peripheral cell wall. (This cell is fractured at an angle of 900 to the cell of panel A.)Bar, 200 nm. (C) Surfaces of freeze-etched cells in which the autolytic wall system was deactivated by a 14-h treatment with 20 ,ug ofchloramphenicol per ml. The autolytic wall system was subsequently reactivated by a 4-h treatment with 10 pLg of lysozyme per ml in 0.1 Macetate buffer at pH 5. Different stages of activation of regularly arranged pores at the cell surface in the first division plane are shown.Sometimes, the pores appear to be pushed apart and consequently are arranged in a zigzag line (black arrows) or even in two distinct parallelrows (white arrows). Bar, 1,000 nm.

VOL. 174, 1992

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

FIG. 2. Freeze-etched surface and thin section of S. aureus SG511 Pelzer. (A) Freeze-etched cell in which the autolytic wall system wasdeactivated by an 8-h treatment with 40 ,ug of chloramphenicol per ml. The autolytic wall system was subsequently reactivated by a 4-htreatment with 1,000 ,ug of lysozyme per ml in 0.1 M acetate buffer at pH 5. Regularly arranged blebs (arrows) on the cell surface in the seconddivision plane are shown. 1.1, first division plane. Bar, 200 nm. (B) Thin section of a cell in which the autolytic wall system was deactivatedby a 14-h treatment with 20 ,ug of chloramphenicol per ml. The autolytic wall system was subsequently reactivated by a 4-h treatment with1 ,ug of lysozyme per ml in 0.1 M acetate buffer at pH 5. Release of a murosome (Mu), leaving behind a characteristic pore, is shown. Thepores (asterisks) are no longer located directly above the cross wall (cf. Fig. 1A) but are pushed apart. Bar, 200 nm.

2244

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

FIG. 3. Thin sections of penicillin-treated cells of S. aureus SG511 Berlin. (A) Thin section of a cell treated for 4 h with 0.1 ,ug of penicillinG per ml. An incomplete, deformed cross wall consisting of loose wall material which lacks the splitting system has been formed (cf. Fig. 1A).Lytic murosomal sites (arrows) above the cross wall of the first division plane can be seen. Bar, 200 nm. (B) Longitudinal section throughthe cross wall of a cell treated for 6 h with 0.1 ,ug of penicillin G per ml. The arrangement of lytic murosomal sites (arrows) and supposedmurosomal sites (arrowheads) in the first division plane can be seen. Bar, 200 nm.

2245

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

FIG. 4. Thin sections of penicillin-treated cells of S. aureus SG511 Berlin. (A) Thin section of a cell treated for 4 h with 0.1 p.g of penicillinG per ml. Large amounts of cross-wall material have been deposited further on in the first division plane (1.....1), while in the second divisionplane only a little cross-wall material was deposited beneath the murosome (Mu). Bar, 200 nm. (B) Thin section of a cell treated for 4 h with0.1 ,ug of penicillin G per ml. At the second division plane, only a naked murosome (Mu) was deposited without any trace of cross-wallmaterial underneath. However, large amounts of cross-wall material have been added to the incomplete, deformed cross wall in the firstdivision plane (1..... 1). Bar, 200 nm.

2246

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

PENICILLIN-INDUCED DEATH 2247

induced cell separation started normally in the second divi-sion plane and, because of the lack of sufficient protectivecross-wall material, this separation resulted in the perfora-tion of the peripheral cell wall (Fig. SA). Near the site ofwall perforation, a characteristic ribosome-free transparentzone in the cytoplasm could often be observed. High-resolution pictures of the site of wall perforation revealedthat part of the cytoplasm, still covered with the cytoplas-mic membrane, was often bulged out like an aneurysm (Fig.SB). Furthermore, at the perforation site, characteristicY-shaped broken edges of the peripheral cell wall couldbe demonstrated. Very important is the observation thatin those cases in which several murosomes were found inthis area, only one of them perforated the cell wall com-

pletely, causing the release of cytoplasmic constituents,while the adjacent murosomes (Mu in Fig. SB), althoughactivated as well, perforated only some outer layers, leav-ing the wall layers lying underneath apparently intact andresistant to the internal pressure. Consequently, we havenever observed the release of cytoplasmic componentsfrom such adjacent murosomal sites. In some cases, adjacentmurosomes even remained trapped within a slit in thewall (Fig. SB) that resembled two of those Y-shaped partsof the cell wall, still connected with each other. More-over, such trapped murosomes revealed that their mem-

branes could not be distinguished from the cytoplasmicmembrane.Our earlier findings that, in penicillin-treated staphylo-

cocci, murosomal sites in the peripheral wall appeared notonly at the nascent cross walls of the first and seconddivision planes but also anywhere in between (6) promptedus to analyze a large number of thin sections of penicillin-treated cells. In this way, it was possible to demonstrate thatthere were about the same number of murosomal sites in thesecond (Fig. 6) and in the first division plane (Fig. 3B). This,however, seemed to contradict our finding that there existedonly one single killing site in the second division plane ofpenicillin-treated staphylococci. Consequently, we had toconsider two possibilities: (i) the single killing site merelymarked the location of the most active of all peripheralmurosomes, or (ii) we had simply been unable to detect theother extrusions.To clear up this question, we cultivated the staphylococci

in the presence of penicillin plus 3% (wt/vol) sodium chlo-ride, which established an external osmotic pressure almostequalling the pressure inside the staphylococcal cells (17).After replacement of this high-pressure growth medium witha normal medium, sufficient fixation fluid was added topreserve the induced effect. In this way, we were able toshow by scanning electron microscopy that a varying num-

ber of murosomes was ejected simultaneously from thesurface of a staphylococcus. Three types of cells were

observed (Fig. 7A): (i) staphylococci from which a row ofseveral murosomes was ejected (Fig. 7A, upper cell), (ii)staphylococci from which only a few murosomes were

ejected (Fig. 7A, left and right cells), and (iii) staphylococcifrom which no murosomes at all were ejected (Fig. 7A, lowercell). In a penicillin-treated population, there were foundonly a few cells of group i and a significant number of cells ofgroup ii, while the majority of cells belonged to group iii.However, we cannot exclude the possibility that we were

losing some of the ejected murosomes during the prepara-

tion.In some cells, a fan-shaped row of several murosomes was

ejected in the second division plane of one daughter cell (Fig.7B), while the second daughter cell exhibited only one single

ejection site. Finally, it should be noted that the number ofejections seen by scanning electron microscopy (Fig. 7B, leftdaughter cell) corresponded to the number of lytic muroso-mal sites found in the second division plane in thin sections(Fig. 6) and also to the number of blebs revealed by freezefracture analysis (Fig. 2A). Murosome ejection was some-times followed by the release of part of the cytoplasm (Fig.7B, left daughter cell), while some other cells seemed toeject only murosomes (Fig. 7A, upper cell).

DISCUSSION

The data presented here confirm the existence of minutestaphylococcal wall organelles, the murosomes (3, 6), whichin untreated cells function to initiate cell separation, and alsoconfirm that they are involved in the killing process inducedby lytic concentrations of penicillin G (0.1 pLg/ml). Whilemost of the murosomes were surrounded by wall material(Fig. 4A and SB; cf. Fig. 1A, 1B, and 2B), there were othersthat appeared in the absence of cross-wall material in thesecond division plane as naked murosomes lying betweenthe cytoplasmic membrane and the peripheral cell wall at thepresumptive site of the cross wall (Fig. 4B). These nakedmurosomes point to the possibility that murosome formationprecedes cross-wall formation. Such a conclusion would bein line with the evidence that murosomes, although diskshaped at that time, were already detected at the verybeginning of cross-wall formation in untreated staphylococci(Fig. 1B).

Confirming previous studies (6), the lethal event after theaddition of lytic concentrations of penicillin (0.1 ,ug/ml) tostaphylococcal cultures was shown to result from a 3-lac-tam-mediated morphogenetic mistake; this occurred at theonset of cell separation in the second division plane, becausethe formation of murosome-derived pores in the peripheralwall was no longer preceded by the underlaying of sufficient"protective" cross-wall material (Fig. 4). Consequently,unprotected pores appeared in this plane and, during extru-sion of the "killing murosome" (Fig. 5), the integrity of thewall, functioning as a pressure-stabilized exoskeleton (11,21, 30), was destroyed. Consequently, the cell lost part of itscytoplasm (Fig. 5A and 7B), the internal pressure droppedsuddenly, and the cell eventually died. The penicillin-in-duced death must, therefore, mainly be attributed to astrictly localized mechanical effect resembling the well-known process of aneurysm formation (Fig. SB). Which ofthe approximately 20 murosomes finally turns out to be thekilling murosome probably depends on the actual amount ofcross-wall material deposited beneath each individual muro-some in the second division plane. Assuming this is the case,it is considerably more likely that a naked murosome (Fig.4B and SA) functions as the killing murosome than that amurosome under which at least some cross-wall materialwas deposited does so (Fig. 4A).

During the rapid wall perforation process, part of thecytoplasm was ejected along with the murosome (Fig. 7B),resulting also in a certain loss of ribosomes, especially fromcytoplasmic areas close to the perforation site (Fig. SA). Inan earlier study (13), we had presented evidence that at theonset of the killing process, as deduced from a sudden dropin the number of viable cells, the release of radiolabelledRNA is rapidly enhanced. The bulk of this released RNAproved to be of high molecular weight and was precipitableby trichloroacetic acid (5). These data are compatible withearlier reports of a considerable loss of RNA during penicil-

VOL. 174, 1992

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

2248 GIESBRECHT ET AL.

lin treatment (16, 18), which could also be documented inthin sections of Proteus mirabilis (14) and S. aureus (5).To elucidate the mechanism of ,-lactam action, it ap-

peared necessary to study also the murosome-induced poreformation process in the cell walls of chloramphenicol-treated staphylococci. Under the experimental conditionsemployed, pore formation introducing the separation of suchcells proved to be a two-step process. It started with thelocal liberation of minute amounts of peripheral wall materialfrom the surface of the murosomes, giving rise to theformation of about 20 regularly arranged blebs (Fig. 2A).Since these blebs probably contained only traces of disinte-grated wall material, this process could hitherto not bedetected by monitoring the release of labelled wall material(19). The bleb formation was followed by the release ofabout 20 murosomes (Fig. 2B; cf. Fig. 7B), leaving behindabout 20 pores in the peripheral wall (Fig. 1C). However, inthe presence of penicillin, pore formation during attemptedcell separation at the unprotected second division plane wasfollowed by wall perforation, leaving behind characteristicY-shaped broken edges in the peripheral cell wall (Fig. 5B).These broken edges must, therefore, be the result of athree-step process comprising (i) bleb formation, (ii) theformation of pores, and (iii) the destruction of the basal walllayers beneath these pores. Only in cases of naked muro-somes (Fig. 4B and 5A) did the centrifugally directed lyticactivity of these organelles apparently result in direct wallperforations.By salt-mediated compensation of the internal osmotic

pressure of the bacteria, we were able, for the first time, toincrease considerably the number of murosomes involved inthe killing process (Fig. 7A, upper cell). It could be demon-strated (Fig. 7B, left daughter cell) that the number ofejected murosomes corresponded to the number of blebs(Fig. 2A) and also to the number of pores (Fig. 1C and 3B).The row of murosomes, ejected like a fan-shaped "muro-some fountain" (Fig. 7B), ruled out the possibility that onlyone single killing murosome might be located in the seconddivision plane. This fact, together with the findings from thinsectioning that the numbers of murosome-induced lytic sitesin the first and second division planes of penicillin-treatedstaphylococci were about the same (Fig. 3B and 6), led to theconclusion that an earlier-considered transport of muro-somes within the peripheral wall (6) is unlikely.

Interestingly, those penicillin-treated staphylococci fromwhich the ejection of activated murosomes had been pre-vented by a high external osmotic pressure released differentnumbers of murosomes while being transferred to normalgrowth medium (Fig. 7A). This observation pointed to astepwise expulsion of activated murosomes during cell sep-aration, which is in line with earlier findings obtained withuntreated cells; in these staphylococci, cell separationstarted at one single "starting pore" of the peripheral wall

and proceeded from there via sequential formation of adja-cent pores (7) along the splitting system (1, 8, 10). Hence, thebest interpretation of the different numbers of murosomesejected from the cells in Fig. 7A is that different stages ofmurosome activation have been reached within the individ-ual cell separation processes. Another striking observationwas that, under the same experimental conditions, evennonseparated daughter cells ejected different numbers ofmurosomes (Fig. 7B), which indicated that the activationof the murosomes in the two daughter cells had beeninitiated asynchronously. This indication is compatible withour earlier findings that, under the influence of penicillinor chloramphenicol, the cross walls within daughter cellsof staphylococci were not formed simultaneously; in manycases, one of two nonseparated daughter cells had al-ready completed its cross wall when the other one hadonly initiated its formation. Sometimes, we even observedthat one daughter cell had formed a complete cross wall,while in the other one cross-wall formation was blocked (10,26).One of the most important problems in connection with

penicillin-induced death was the question why, during theearly period of the action of this drug, i.e., before cell wallsynthesis was completely inhibited (19), the new cross-wallmaterial continued to be assembled in the first rather than inthe second division plane (Fig. 4 and SA). Since this mor-phogenetic mistake was the very reason for the subsequentdeath, it has to be considered in some detail. The cross-wallmaterial of the first division plane appeared to be lesscompact (6) than in control cells (Fig. 3A, 4, and SA; cf. Fig.1A), but we can presently only speculate that this mightreflect not only the well-known reduced cross-linking (23)but also the liberation of autolytic wall enzymes from thecontainers of the splitting system of this cross wall whichwere missing or defective in the presence of penicillin (Fig.3A; cf. Fig. 1A). Such liberated wall enzymes could havecut numerous free ends into the cross-wall material, whichmight then have served as additional acceptors for newlyformed strings of wall material, probably by far exceed-ing those supposed to be created normally in a newlyinitiated division plane. Earlier experiments, employingcombinations of penicillin and anionic polyelectrolytes, sup-ported such considerations (27). If this is true, "drifting"wall autolysins from the splitting system of the first divi-sion plane could be one of the most important prerequisitesfor penicillin-induced death. Without this defect, the newlyformed cross-wall material, bound for the second divi-sion plane, would probably not be detoured to the firstdivision plane; consequently, the murosomes of the seconddivision plane would become protected by sufficient cross-wall material and no wall perforation would take place at thissite.

It is important to mention that thin sections of penicillin-

FIG. 5. Thin sections of penicillin-treated cells of S. aureus SG511 Berlin. (A) Thin section of a cell treated for 4 h with 0.1 ,ug of penicillinG per ml. Perforation of the peripheral cell wall in the second division plane has resulted in the ejection of the naked murosome (Mu) locatedat this site. Large amounts of cross-wall material have been added to the incomplete, deformed cross wall in the first division plane (1..... 1).Asterisk, ribosome-free, rather transparent zone of the cytoplasm. Bar, 200 nm. (B) Thin section of a cell treated for 2 h with 0.1 ,ug ofpenicillin G per ml. In the second division plane, the area of wall perforation (killing site), showing a bulged part of the cytoplasm still coveredwith the cytoplasmic membrane (asterisk), can.be detected. Characteristic Y-shaped broken edges of the perforated peripheral cell wall (Y),with their outsides reaching over the cell surface like "volcanos" (Vo), are to be seen. At the upper side, a murosome (Mu) is still trappedwithin the wall material. This murosome has perforated only some outer layers of the cell wall, leaving some inner wall layers intact. Themurosomal membrane (MuM) cannot be distinguished from the cytoplasmic membrane (CM). Bar, 200 nm.

J. BACTERIOL.

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

PENICILLIN-INDUCED DEATH 2249VOL. 174, 1992

F

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

2250 GIESBRECHT ET AL.

treated staphylococci did not reveal any morphologicaldifferences between dead and living cells (apart from theminute wall perforations and the loss of some ribosomes;Fig. 5A). The same observation was made after penicillin-induced killing in the presence of an anionic polyelectrolytethat completely blocked bacteriolysis (5, 29). Also, theobservation that bacteriolysis, as detected by the drop inoptical density, normally started more than one genera-tion time after the penicillin-induced death (13) indicatedthat the lethal wall perforation process can no longer beidentified with the process of bacteriolysis. Since inhibi-

tion of bacteriolysis did not interfere substantially with thekilling rate, we conclude that the penicillin-induced deathby lytic penicillin concentrations (0.1 ,ug/ml) represents anonlytic killing process, while bacteriolysis is assumed to benothing but a postmortem side effect (5). We should, there-fore, no longer synonymize "bacteriolysis" with "lyticdeath" (6, 15, 22).

Finally, we wish to emphasize that the possibility that thenumber of murosomes functioning as killing murosomes isconsiderably increased by varying the osmolarity shouldstimulate attempts to improve ,-lactam therapy by applying

FIG. 6. Thin section of S. aureus SG511 Berlin treated for 4 h with 0.1 ,ug of penicillin G per ml. Arrangement of lytic murosomal sites(arrows) and supposed murosomal sites (arrowheads) in the second division plane, comparable to the distribution of the lytic murosomal sitesin the first division plane (cf. Fig. 3B). 1..... 1, first division plane. Bar, 200 nm.

FIG. 7. Scanning electron micrographs of S. aureus SG511 Berlin, treated for 2 h with 0.1 ,ug of penicillin G per ml in a growth mediumcontaining 3% NaCl, after transfer into normal growth medium. (A) A characteristic group of cells from which a varying number ofmurosomes was ejected; either a row of several murosomes (upper cell; large arrows) or only a few murosomes (left and right cells; smallarrows) or no murosome at all (lower cell) are to be seen. No detectable amounts of cytoplasm are ejected along with the murosomes. Bar,200 nm. (B) Fan-shaped ejection of several murosomes (arrows) along with part of the cytoplasm (asterisks) only in the second division planeof the left daughter cell. Supposed ejection sites are marked with arrowheads. 1..... 1, first division plane. In the right daughter cell, only onesingle killing murosome has been ejected (large arrow) in the second division plane. Bar, 200 nm.

J. BACTERIOL.

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

VOL. 174, 1992 PENICILLIN-INDUCED DEATH 2251

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

2252 GIESBRECHT ET AL.

ext,rnal factors that enlarge the number of lytically activevesicular structures (4) or increase the activation of suchstructures (27).

ACKNOWLEDGMENTS

We thank Gerta Tochtermann, Bernd Nurnberg, Chang Han,

Georg Kniffke, and Ellen Albrecht for their excellent technicalassistance.

REFERENCt:S1. Amako, K., A. Umeda, and K. Murata. 1982. Arrangement of

peptidoglycan in the cell wall of Staphylococcus spp. J. Bacte-riol. 150:844-850.

2. Beise, F., H. Labischinski, and P. Giesbrecht. 1988. Selectiveinhibition of penicillin-binding proteins and its effects on growthand architecture of Staphylococcus aureus. FEMS Microbiol.Lett. 55:195-202.

3. Giesbrecht, P. 1984. Novel bacterial wall organelles ("muro-somes") in staphylococci: their involvement in wall assembly,p. 177-186. In C. Nombela (ed.), Microbial cell wall synthesisand autolysis. Elsevier Science Publishers, Amsterdam.

4. Giesbrecht, P. 1991. Perturbation of the membrane-wall-in-terlayer by volume reductions of staphylococci results inwall disintegration and bacteriolysis after formation of mesoso-mal vesicles, being no fixation artifacts but rather multiplied<<exosomes>>, p. 173-177. In J. Jeljaszewicz and J. Ci-borowski (ed.), The staphylococci. Gustav Fischer Verlag,Stuttgart.

5. Giesbrecht, P., M. Franz, D. Krfiger, H. Labischinski, and J.Wecke. Bacteriolysis of staphylococci is only a side-effect ofpenicillin-induced death. In H. Kleinkauf and H. von Dohren(ed.), 50 years of penicillin application-history and trends, inpress. Walter de Gruyter, Berlin.

6. Giesbrecht, P., H. Labischinski, and J. Wecke. 1985. A specialmorphogenetic wall defect and the subsequent activity of"murosomes" as the very reason for penicillin-induced bacte-riolysis in staphylococci. Arch. Microbiol. 141:315-324.

7. Giesbrecht, P., and J. Wecke. 1971. Zur Morphogenese derZellwand von Staphylokokken. I. Querwandbildung und Zell-trennung. Cytobiologie 4:349-368.

8. Giesbrecht, P., and J. Wecke. 1980. On the structure andfunction of autolytic wall systems in gram-positive bacteria, p.

446-453. In P. Brederoo and W. de Priester (ed.), Proceedingsof the 7th European Congress on Electron Microscopy, vol. 2.Seventh European Congress on Electron Microscopy Founda-tion, Leiden.

9. Giesbrecht, P., and J. Wecke. 1986. A new method in bacterialcell wall research: controlled transition from freeze-fractioningto freeze sectioning in latent living staphylococci. J. ElectronMicrosc. 35(Suppl.):2057-2060.

10. Giesbrecht, P., J. Wecke, and B. Reinicke. 1976. On the mor-

phogenesis of the cell wall of staphylococci. Int. Rev. Cytol.44:225-317.

11. Holtje, J.-V., and U. Schwarz. 1985. Biosynthesis and growth ofthe murein sacculus, p. 77-119. In N. Nanninga (ed.), Molecularcytology of Escherichia coli. Academic Press, London.

12. Kohlrausch, U., and J.-V. Holtje. 1991. Analysis of murein andmurein precursors during antibiotic-induced lysis ofEscherichiacoli. J. Bacteriol. 173:3425-3431.

13. Labischinski, H., H. Maidhof, M. Franz, D. Kriiger, T. Sidow,and P. Giesbrecht. 1988. Biochemical and biophysical investiga-tions into the cause of penicillin-induced lytic death of staph-ylococci: checking predictions of the murosome model, p.

242-257. In P. Actor, L. Daneo-Moore, M. L. Higgins, M. R. J.Salton, and G. D. Shockman (ed.), Antibiotic inhibition ofbacterial cell surface assembly and function. American Societyfor Microbiology, Washington, D.C.

14. Lorian, V., L. D. Sabath, and M. Simionescu. 1975. Decrease inribosomal density of Proteus mirabilis exposed to subinhibitoryconcentrations of ampicillin or cephalothin (38888). Proc. Soc.Exp. Biol. Med. 149:731-735.

15. Maidhof, H., L. Johannsen, H. Labischinski, and P. Giesbrecht.1989. Onset of penicillin-induced bacteriolysis in staphylococciis cell cycle dependent. J. Bacteriol. 171:2252-2257.

16. McDowell, T. D., and K. E. Reed. 1988. Hydrolysis of RNAduring penicillin-induced nonlytic death in a group A strepto-coccus, p. 231-241. In P. Actor, L. Daneo-Moore, M. L.Higgins, M. R. J. Salton, and G. D. Shockman (ed.), Antibioticinhibition of bacterial cell surface assembly and function. Amer-ican Society for Microbiology, Washington, D.C.

17. Mitchell, P., and J. Moyle. 1956. Osmotic function and structurein bacteria, p. 150-180. In E. T. C. Spooner and B. A. D.Stocker (ed.), Bacterial anatomy. Sixth Symposium of theSociety for General Microbiology held at the Royal Institution,London. Cambridge University Press, Cambridge.

18. Prestige, L. S., and A. B. Pardee. i957. Induction of bacteriallysis by penicillin. J. Bacteriol. 74:48-59.

19. Reinicke, B., P. Bluimel, H. Labischinski, and P. Giesbrecht.1985. Neither an enhancement of autolytic wall degradation noran inhibition of the incorporation of cell wall material arepre-requisites for penicillin-induced bacteriolysis in staphylo-cocci. Arch. Microbiol. 141:309-314.

20. Reynolds, P. E. 1988. The essential nature of staphylococcalpenicillin-binding proteins, p. 343-351. In P. Actor, L. Daneo-Moore, M. L. Higgins, M. R. J. Salton, and G. D. Shockman(ed.), Antibiotic inhibition of bacterial cell surface assembly andfunction. American Society for Microbiology, Washington,D.C.

21. Rogers, H. J., H. R. Perkins, and J. B. Ward. 1980. Microbialcell walls and membranes. Chapman and Hall, London.

22. Shockman, C. D., L. Daneo-Moore, T. D. McDowell, and W.Wong. 1981. Function and structure of the cell wall-its impor-tance in the life and death of bacteria, p. 31-65. In M. Salton andG. D. Shockman (ed.), P-Lactam antibiotics. Academic Press,New York.

23. Sidow, T., L. Johannsen, and H. Labischinski. 1990. Penicillin-induced changes in the cell wall composition of Staphylococcusaureus before the onset of bacteriolysis. Arch. Microbiol.154:73-81.

24. Tomasz, A. 1979. The mechanism of the irreversible antimicro-bial effects of penicillin: how beta-lactam antibiotics kill andlyse bacteria. Annu. Rev. Microbiol. 33:113-137.

25. Tomasz, A., and S. Waks. 1975. Mechanism of action ofpenicillin: triggering of the pneumococcal autolytic enzyme byinhibitors of cell wall synthesis. Proc. Natl. Acad. Sci. USA72:4162-4166.

26. Wecke, J., M. Franz, and P. Giesbrecht. 1990. Inhibition of thebacteriolytic effect of ,B-lactam antibiotics on Staphylococcusaureus by the polyanionic drugs suramin and Evans Blue*.APMIS 98:71-81.

27. Wecke, J., T. Kersten, and P. Giesbrecht. The modulation of thebacteriolytic effect of P-lactam antibiotics by non-antibiotics.APMIS, in press.

28. Wecke, J., M. Lahav, I. Ginsburg, and P. Giesbrecht. 1982. Cellwall degradation of Staphylococcus aureus by lysozyme. Arch.Microbiol. 131:116-123.

29. Wecke, J., M. Lahav, I. Ginsburg, E. Kwa, and P. Giesbrecht.1986. Inhibition of wall autolysis of staphylococci by sodiumpolyanethole sulfonate "liquoid". Arch. Microbiol. 144:110-115.

30. Weidel, W., and H. Pelzer. 1964. Bagshaped macromolecules: anew outlook on bacterial cell walls. Adv. Enzymol. 26:193-232.

31. Westmacott, D., and H. R. Perkins. 1979. Effects of lysozyme onBacillus cereus 569: rupture of chains of bacteria and enhance-ment of sensitivity to autolysins. J. Gen. Microbiol. 115:1-11.

J. BACTERIOL.

on April 7, 2020 by guest

http://jb.asm.org/

Dow

nloaded from