BACTERIOLYSIS OF ENTEROBACTERIACEAEII. PRE- AND CO-LYTIC TREATMENTS POTENTIATING THE ACTION OF LYSOZYME1' 2

ERIC C. NOLLER3 AND S. E. HARTSELLLaboratories of Bacteriology, Department of Biological Sciences, Purdue University, Lafayette, Indiana

Received for publication September 19, 1960

Most gram-negative bacteria are refractoryto the lytic action of lysozyme unless the cellshave been conditioned for lysis by certain pre-lytic or co-lytic treatments. There are sugges-tions that the resistance of gram-negative bac-teria is conferred by protein which protects themucopolysaccharide lysozyme substrate (Beckerand Hartsell, 1954, 1955; Salton, 1958; Weideland Primosigh, 1958; Kellenberger and Ryter,1958) from lysozyme action. It is also possiblethat the substrate exists in a state of interdis-persion with lipoprotein and that it is the lipo-protein-mucopolysaccharide complex that mustbe dissociated before lysozyme can reach itssubstrate. The function of pre- and co-lytictreatments which potentiate lysozyme actionappears to be one of releasing the lysozyme sub-strate by degradation of the protective lipo-protein moiety of the cell wall.The mucopolysaccharide is presently consid-

ered to be the wall component that lends rigidityto the gram-negative cell wall (Salton, 1955,1958). However, Grula and Hartsell (1957)observed that lysozymic degradation of gram-negative bacteria by Nakamura's technique(Nakamura, 1923) resulted in ghosts possessinga general rodlike appearance that would shrinkand swell after alternate addition of acid andalkali. Since lysozyme substrate was presumablyabsent, it appeared likely that some other typeof reticulum (protein?) may also be importantin conferring shape upon gram-negative cells.This possibility has also been suggested by Tuttleand Gest (1959).

1 This study was supported in part by the Pur-due Research Foundation (XR2042).

2 This work was taken from a dissertationsubmitted in partial fulfillment of the require-ments for the Ph.D. degree at Purdue University,Lafayette, Indiana.

3Present address: Department of Bacteriology,Oklahoma State University, Stillwater, Okla-homa.

During a comparative study of the lytic re-sponse of Enterobacteriaceae to four lytic systemsutilizing lysozyme (Noller and Hartsell, 1961),there appeared to be a similarity between themode of action of the pre- and co-lytic treatmentsthat potentiated lysozyme action. The presentreport is concerned with a comparison of theeffects of these treatments and offers evidencefor the mechanism by which heat, low pH,butanol, and Versene potentiate lysozyme ac-tion. Additional evidence to that of Grula andHartsell (1957) will also be presented indicatingthat some component of the cell wall other thanmucopolysaccharide is also important in lendingrigidity to the gram-negative cell wall.

MATERIALS AND METHODS

Bacterial strains. The following bacterial strainswere used in these investigations: Escherichia colistrains 19, U-2, and 4157; Aerobacter aerogenesstrain PU-2.

Comparison of heat and acid pretreatments.One suspension of E. coli strain 19 in water atpH 7 was heated for 15 min at 70 C and thencooled to 45 C. A second cell suspension wasincubated at 45 C at pH 3.5 for 1 hr. A controlsuspension was incubated in water at pH 7 for1 hr at 45 C. After these pretreatments, each ofthe cell suspensions was tested for its sensitivityto 10 Aug/ml lysozyme, 20 ,ug/ml trypsin, and10 ,ug/ml lysozyme plus 20 ,ug/ml trypsin (allat 45 C) in 0.0067 M phosphate buffer (pH 7).The extent of lytic sensitivity of each cell sus-pension was determined turbidimetrically (Cole-man model 14 Universal spectrophotometer).

In a second experiment, a suspension of E.coli strain 19 cells was incubated at 45 C for 30min at pH 3.5. Control suspensions were incu-bated under the same conditions except at pH 7.Each suspension was then adjusted to pH 10with 0.1 N NaOH. The extent of lysis was deter-mined before and after addition of lysozyme,circulin, Hyamine 1622, or salmine to a concen-

492

on January 16, 2021 by guesthttp://jb.asm

.org/D

ownloaded from

BACTERIOLYSIS OF ENTEROBACTERIACEAE. II

tration of 10 jig/ml. The suspensions that werenot acid pretreated were adjusted to pH 3.5with HCl for 5 sec and readjusted back to pH10 with alkali. Lysis was noted and followed byaddition of lysozyme (10 ,ug/ml) to those tubescontaining circulin, Hyamine, or salmine, andfinal observations for the degree of lysis weremade.

Butanol pretreatment. A suspension of E. colistrain U-2 cells was treated with n-butanol asindicated in Fig. 3. The extent of lysis was notedafter alkali addition to pH 10 and after additionof lysozyme at pH 10.

Versene cotreatment. A. aerogenes strain PU-2and E. coli strain U-2 were used as test speciesand were exposed in 0.03 M tris [tris(hydroxy-methyl)aminomethanel buffer (pH 8) to 133 ,ug/ml Versene, 20 ,ug/ml lysozyme, or 10 ,ug/mltrypsin as single and as combined reagents. Theextent of lysis was determined spectrophoto-metrically during incubation of the system at45 C.

Butanol, heat, or acid treatment. To evaluatethe cell wall disorganizing effects of butanol,heat, or acid treatments by some means otherthan macroscopic lysis, hydrogen ion uptakestudies on control and treated cells were madeby a method similar to that described by Gilbyand Few (1958). Washed cells of E. coli strainsU-2 and 19 and A. aerogenes strain PU-2 weresuspended in distilled water at a concentrationof 5 mg cells per ml (dry weight basis). Ten-milliliter portions were heated at 70 C for 15min and immediately cooled to room tempera-ture, treated with 5% v/v n-butanol, or leftuntreated. These 10-ml portions were titratedpotentiometrically at 26 C with 0.5 N HCl.Acid was added to the constantly stirred samplesin 0.01-ml aliquots with an Agla micrometersyringe and pH of the suspension was measured3 min after each addition. Milliequivalents ofhydrogen ion bound per gram dry weight ofcells were calculated from pH values for distilledwater and cell suspensions resulting from eachaddition of acid.

Cell structures remaining after lysis. The lysisof acid or butanol pretreated cells by alkalinelysozyme was observed with the phase andelectron microscope. E. coli strain U-2 suspen-sions were pretreated by incubation at pH 3.5for 30 min at 45 C or in the presence of 5%v/v n-butanol at pH 7 for 30 min at 45 C. Cellsfrom these suspensions were dried on cover

glasses and observed by phase microscopy duringsequential treatment with pH 10.5 alkali, 10,ug/ml lysozyme at pH 10.5, and pH 3.5 acid.Cells were also air-dried to collodion-coveredgrids and then treated on the grids by the samereagents, followed by observation of the prep-arations with the electron microscope.

In a second series of experiments, the effectof trypsin digestion of the ghosts remainingfollowing alkaline-lysozyme digestion of acidpretreated cells of E. coli strain 4157 was deter-mined. Residues remaining after treatment ofE. coli strain 4157 cells with lysozyme-trypsin-butanol, lysozyme-Versene, or lysozyme-circulinwere also subjected to alternate treatment withacid and alkali to determine whether shrinkingor swelling of these ghosts occurred.

RESULTS

Comparison of heat and acid pretreatments.Cells incubated at 45 C at neutral pH were notlysed by lysozyme or trypsin individually orcombined. Heated cells, however, exhibited tryp-sin sensitivity (Fig. 1) and synergistic lysis bythe combination of lysozyme and trypsin. Thisverifies reports of Becker and Hartsell (1954,1955). Acid pretreated cells were also lysedmarkedly by lysozyme or trypsin and were moresensitive to the action of these enzymes thanwere heated cells. No synergistic action with

Cells pre-treated Cells pre- treated15min,70C,pH7 Ihr,45C, pH 3.5

60 0 15

Minutes at 45 C

Fig. 1. Acid and heat pretreatments as po-tentiators of lysozyme and trypsin activity.Escherichia coli strain 19. L = 10 ,ug/ml lysozyme>T = 20 ,g/ml trypsin.

49319611l

on January 16, 2021 by guesthttp://jb.asm

.org/D

ownloaded from

NOLLER AND HARTSELL

Cells Pre-treatedL30min, 45 C, pH3.5|

NaOH H 10

30% Clearing

+ Lysozym + Hymine, Circulin,90% Mlel S40%ineCleor Water

90% Clearing 40% Clearing

E. coli

0,O.

0c

00.

Cells Pre-treated30min, 45 C pH7

|NaOHtopl1

15% Clearing

|+ Lysozyme | +H yamine, Circulin,|Smie,or Water

40% Clearing 40 % ClearingI I

+HCI to pH 3.5 for 5sec,NaOH to pH 1

I 185% Clearing 50% Clearing

+ Lysozyme

80%Clearing

Fig. 2. Lysis of Escherichia coli strain 19 bylysozyme at pH 10.

A. aerogenes

0 15 30 60 0 15 30 60

Minutes at 45 C

Fig. 4. Lysis of Escherichia coli strain U-2and Aerobacter aerogenes strain PU-2 by lysozyme,

trypsin, and Versene. L = 20 ,ug/ml lysozyme;T = 10 /g/ml trypsin; V = 133 tug/ml Versene.

I0-)

06'- 40

a.

0

xD 2a)

a.

A.aerogenes 26 C

Pre- treatments:A- 70 C, 15 minB- 5% n-BuOHC- None

7 6 5 4 3pH

Fig. 5. Hydrogen ion uptakeheat treated cells.

45 C, pH 7,5tn-oButano

Centrifuged ResidueRe-suspended in Water;

NaOH to pH10O

40%/ Clearing

+ Lysoizyme

90%/ Clearing

Fig. 3. Lvsozyme action at pH ].0 after pre-treatment of Escherichia coli strain U-2 cellswith 5% n-butanol.

lysozyme plus trypsin is apparent with acidtreated cells. This is undoubtedly due to thehigh activity of the individual enzymes. Thesedata suggest that neutral heat treatment (70 C)

by butanol and

and pH 3.5 treatment (45 C) accomplish similardisorganization of lipoprotein.From the results presented in Fig 2, it is ap-

parent that the presence of lysozyme at pH 3.5is not necessary for good lysing (clearing) byalkali. Addition of lysozyme after raising acidtreated cells to pH 10 with alkali resulted inimmediate clearing comparable in extent tocells treated by Nakamura's technique. Lysozymeappears quite capable of substrate degradationat a pH as high as 10. This lytic effect of lyso-zyme at pH 10 appears specific since other basicpeptides such as salmine or circulin do not pro-

duce the effect.Butanol pretreatment. Pretreatment of cells

with 5% v/v n-butanol at pH 7 appeared to have

494 [VOL. 81

on January 16, 2021 by guesthttp://jb.asm

.org/D

ownloaded from

Fig. 6. Effect of alkali plus lysozyme and acid on pH 3.5 pretreated Escherichia coli strain U-2. a) pH3.5 pretreated cells (30 min at 45 0); b and c) same cells at pH 10.5 plus lysozyme for 5 min; d and e) fol-lowed by pH 3.5 treatment; f and g) followed by pH 10.5 treatment; h and i) by pH 3.5treatment.

495

on January 16, 2021 by guesthttp://jb.asm

.org/D

ownloaded from

NOLLER AND HARTSELL

Fig. 7. Effect of alkali plus lysozyme and acid on n-butanol pretreated E8cherichia coli strain U-2.a) n-Butanol pretreated cells (30 min, pH 7, 45 C); b) same cells at pH 10.5 plus lysozyme for 5 min; c)followed by pH 3.5 treatment; d) followed by pH 10.5 treatment; e) followed by pH 3.5 treatment.

a comparable effect with acid pretreatment to acid effects and perhaps to those of heat.(Fig. 3). Addition of lysozyme following adjust- Similarity between the effects of butanol andment of the suspension of pretreated cells to pH heat pretreatments is supported by the fact that10 resulted in immediate and extensive lysis lysis curves after heat treatment (Fig. 1) are(clearing). Hence, butanol effects appear similar comparable to those obtained with butanol

496 [VOL. 8]

on January 16, 2021 by guesthttp://jb.asm

.org/D

ownloaded from

BACTERIOLYSIS OF ENTEROBACTERIACEAE. II

..ti C 8 dFig. 8. Prevention of shrinking and swelling by trypsin treatment. a) Escherichia coli strain 4157, pH

3.5 pretreated (30 min at 45 C); b) followed by pH 10.5 plus lysozyme for 5 min; c) followed by trypsintreatment at pH 8 for 5 min, then pH 3.5; d) followed by pH 10.5 treatment.

cotreated cells observed in an earlier report(Noller and Hartsell, 1961).

Versene cotreatment. The presence of Versenewas found to increase both trypsin and lysozymeactivity (Fig. 4). While E. coli strain U-2 wasmore sensitive than A. aerogenes strain PU-2to the action of Versene and trypsin or Verseneand lysozyme, both organisms were comparablylysed by a combination of all three compounds.The lysis curves again appear similar to thoseobtained with lysozyme, trypsin, and butanol(Noller and Hartsell, 1961). While metal chela-tion may influence lytic sensitivity, it appearsthat the major role of Versene in lysozyme ortrypsin potentiation could be due to lipoproteindissociation.

Butanol, heat, or acid treatment. The hydrogenion uptake curves presented in Fig. 5 were ob-tained using A. aerogenes strain PU-2 cells and

are typical of the results obtained with all threeorganisms. Control cells were observed to bindhydrogen ion slowly until a critical acidity wasreached in the vicinity of pH 3. At lower pH,hydrogen ion was bound rapidly with each addi-tion of acid. The low initial uptake representstitration of cell wall groups prior to completedisorganization of the wall and plasma mem-brane. At pH 3 or less, wall and membrane dam-age allows penetration of the cytoplasm byprotons. Heat treated cells and cells titratedin the presence of n-butanol exhibited markedproton uptake from the beginning of acid titra-tion. The most logical explanation of this resultis that heat and butanol cause prior disorganiza-tion of the wall and membrane allowing protonuptake by both wall and cytoplasm.

Cell structures remaining after lysis. The alter-nate addition of alkali and acid to acid or butanol

19611 497

on January 16, 2021 by guesthttp://jb.asm

.org/D

ownloaded from

NOLLER AND HARTSELL

pretreated cells had no effect on cell morphology.When lysozyme (10 Ag/ml) was added to acidpretrea.ted (Fig. 6) or n-butanol pretreated(Fig. 7) cells under alkaline conditions (pH10.5), marked swelling occurred with a concomi-tant loss of refractility. Moreover, the swellingand refractility loss could be reversed by loweringthe pH to 3.5 with hydrochloric acid. Alternateaddition of alkali and acid caused alternatereversal of the cell residues from a swollen to ashrunken state. This effect was first noted byGrula and Hartsell (1957) with acid pretreatedcells. The same effect was observed with butanolpretreated cells. The series of electron photo-micrographs (Fig. 6) indicates that acid treat-ment confers definite rigidity to the cells sinceshadows are cast by cells only under acid con-ditions. Residues under alkaline conditions driedflat to the grid and no shadow was cast.When the swollen residues from alkaline-

lysozyme treatment of acid pretreated cells(Fig. 8) were further treated with trypsin atpH 8, the shrinking and swelling reaction nolonger occurred. This effect provides furtherevidence that the residual structures first notedby Grula and Hartsell (1957) that undergoshrinking and swelling are truly proteinaceous(lipoprotein?).

Cell residues remaining after lysozyme-trypsin-butanol or lysozyme-Versene action also remainlargely unchanged by acid or alkali treatment.Residues from lysozyme-circulin action shrinkand swell somewhat during alternate changesfrom acid to alkaline conditions but not to theextent of acid or n-butanol pretreated cells.Tuttle and Gest (1959) observed that Rhodo-spirillum rubrum loses its spiral form followingpolymyxin treatment. They suggested thatlipoprotein was conferring structural rigidityin the untreated cell. Degradation of lipoproteinby circulin should likewise preclude the shrink-swell effect. Residues from lysozyme and circulin,acid, or n-butanol treatments were of such lowrefractility that photomicrography by phasemicroscopy was impossible.

DISCUSSION

All of the pre- and co-lytic treatments testedappear to be capable of degrading the protectivelipoprotein complex to an extent that permitslysozyme penetration. These treatments alsopotentiate the action of trypsin on the proteincomponents of the cell.

Comparison of the effects of acid and butanol

pretreatment indicated that exposure to lowpH is not the only cell treatment that permitsgood lysing of cells by alkali after lysozyme ex-posure. The prerequisite for alkali lysing is notacid pretreatment alone, but probably any treat-ment that disorganizes the lipoprotein componentof the cell wall. Lysozyme then degrades theunbound substrate and alkali aids in solubiliza-tion of the partially degraded cell structures.

It is also apparent that lysozyme is enzymati-cally active over a wide pH range. Lysozymicaction at pH 10 was apparent with acid or bu-tanol pretreated cells. Optimal lysing of MIicro-coccus lysodeikticus cells by lysozyme appearsto occur optimally at pH 6.2 (Smolelis and Hart-sell, 1952) and some activity on cell walls ofM. lysodeikticus has been observed at pH 3.5(Grula and Hartsell, 1954). Hence, lysozymeappears capable of activity over a range of atleast pH 3.5 to 10. A relatively broad pH rangefor lysozyme action has also been reported bySalton (1956). Since lysis of gram-negative cellsat pH 10 to 10.5 is as rapid when lysozyme isadded after rendering the system alkaline aswhen lysozyme was present during acid pre-treatment, it is possible that lysozyme incubatedwith cells at pH 3.5 does not have activity untilthe pH is raised by addition of alkali.The effect of Versene in lysozyme-potentiation

is quite comparable to that of butanol, and,hence, would appear to unmask lysozyme sub-strate through lipoprotein dissociation ratherthan through chelation of metals. Repaske(1958) stated that other chelating agents (8-hy-droxyquinoline, o-phenanthroline, or other Ver-senes) were ineffective as co-lytic agents withlysozyme. This observation would appear tosupport some action of Versene other than thatof metal chelation at the cell surface.The use of hydrogen ion uptake studies on

butanol and heat treated cells lends furthersupport to the cell wall-dissociating effects ofthese treatmentg. It has been frequently sug-gested that such disorganization of cell structurecould conceivably unmask the lysozyme sub-strate and allow lysozyme action. While walldisorganization by acid alone does not occuruntil approximately pH 3 at 26 C, the pH valuesat which disorganization occurs would probablybe higher at 45 C.None of the lytic treatments studied resulted

in spheroplasts when carried out in the presenceof sucrose as a stabilizer. It may be presumedthat the lytic systems caused marked dissociation

498 [VOL. 81

on January 16, 2021 by guesthttp://jb.asm

.org/D

ownloaded from

BACTERIOLYSIS OF ENTEROBACTERIACEAE. II

of the plasma membrane. Since the plasma mem-brane is believed to consist largely of lipoprotein(McQuillen, 1958), it could hardly be expectedto withstand the action of pH extremes, lipidsolvents, heat, or a highly active surfactantsuch as Hyamine 1622 or circulin. If lysozymehas degraded the mucopolysaccharide in thecell wall during treatment in the presence ofalkali, butanol and trypsin, circulin, or Versene,then it is extremely interesting to conjecture asto the structure remaining that continues toretain the cell in its usual rodlike morphology.If the mucopolysaccharide has not been de-graded to the extent that it no longer confersrigidity to the cell wall, then swelling of the resi-dues would be impossible under alkaline con-ditions. The implications of the shrinking andswelling phenomenon are important since theyindicate that some component of the cell wallother than mucopolysaccharide is also conferringshape to the gram-negative cell. Since shrinkingand swelling does not occur following tryptichydrolysis of the residue, it is apparent that theprotein (or lipoprotein) moiety of the cell wall(along with the muco-complex) also contributesto maintenance of cell morphology.

SUMMARY

Pretreatment of gram-negative bacteria byheat, low pH, or n-butanol, or cotreatment withbutanol or Versene, sensitizes the cells to thelytic action of lysozyme. These treatments alsoincrease subsequent tryptic activity on protein-aceous components of the cell. Turbidimetricand microscopic observation of lytic patternsafter such pre- or cotreatments indicate thatthe mechanism by which. sensitization occurs isthrough dissociation of the lipoprotein compon-ents of the cell wall which.protect the lysozymesubstrate. This conclusion has been further sup-ported by hydrogen ion uptake studies on heator butanol treated cells. The evidence suggeststhat Versene potentiates lysozyme and trypsinactivity through lipoprotein dissociation al-though chelation may also be important.

Studies on residual cell structures remainingafter lysis of gram-negative bacteria indicatethat some cell wall or cytoplasmic component(probably lipoprotein) also confers rigidity tothe cell along with mucopolysaccharide.

REFERENCESBECKER, M. E., AND S. E. HARTSELL 1954 Fac-

tors affecting bacteriolysis using lysozyme

in dual enzyme systems. Arch. Biochem.Biophys., 53, 402-410.

BECKER, M. E., AND S. E. HARTSELL 1955 Thesynergistic action of lysozyme and trypsinin bacteriolysis. Arch. Biochem. Biophys.,55, 257-269.

GILBY, A. R., AND A. V. FEW 1958 The perme-ability of Micrococcus lysodeikticus to hydro-chloric acid. Biochim. et Biophys. Acta,30, 421-422.

GRULA, E. A., AND S. E. HARTSELL 1954 Lyso-zyme action and its relation to the Nakamuraeffect. J. Bacteriol., 68, 302-306.

GRULA, E. A., AND S. E. HARTSELL 1957 Lyso-zyme in the bacteriolysis of gram-negativebacteria. I. Morphological changes duringuse of Nakamura's technique. Can. J.Microbiol., 3, 13-21.

KELLENBERGER, E., AND A. RYTER 1958 Cellwall and cytoplasmic membrane of Escherichiacoli. J. Biophys. Biochem. Cytol., 4, 323-325.

MCQUILLEN, K. 1958 Lysis resulting frommetabolic disturbance. J. Gen. Microbiol.,18, 498-512.

NAKAMURA, 0. 1923 Ueber Lysozymwirkungen.Z. Immunitatsforsch., 38, 425-449.

NOLLER, E. C., AND S. E. HARTSELL 1961 Bac-teriolysis of Enterobacteriaceae. I. Lysis byfour lytic systems utilizing lysozyme. J.Bacteriol., 81, 482-491.

REPASKE, R. 1958 Lysis of gram-negativeorganisms and the role of Versene. Biochim.et Biophys. Acta, 30, 225-232.

SALTON, M. R. J. 1955 Cellular structure andenzymic bacteriolysis. Proc. Intern. Congr.Biochem., 3rd Congr. (Brussels), 1955, 406-410.

SALTON, M. R. J. 1956 Studies of the bacterialcell wall. V. The action of lysozyme on cellwalls of some lysozyme-sensitive bacteria.Biochim. et Biophys. Acta, 22, 495-506.

SALTON, M. R. J. 1958 The lysis of micro-organisms by lysozyme and related enzymes.J. Gen. Microbiol., 18, 481-490.

SMOLELIS, A. N., AND S. E. HARTSELL 1952Factors affecting the lytic activity of lyso-zyme. J. Bacteriol., 63, 665-674.

TUTTLE, A. L., AND H. GEST 1959 Subcellularparticulate systems and the photochemicalapparatus of Rhodospirillum rubrum. Natl.Acad. Sci. U. S., 45, 1261-1269.

WEIDEL, W., AND J. PRIMOSIGH 1958 Bio-chemical parallels between lysis by virulentphage and lysis by penicillin. J. Gen. Micro-biol., 18, 513-517.

19611 499

on January 16, 2021 by guesthttp://jb.asm

.org/D

ownloaded from