tures resulting from the action of this reagent con- sisted of impure ... termed solution L: MgSO4-7H20, 0.10 (g per liter); citric acid ... shown by the following experiment. Cells were ..... dicts that lysis would also be caused by reagents which can ...
LYSIS OF ESCHERICHIA COLI BY SULFHYDRYL-BINDING REAGENTS M. SCHAECHTER' AND KATHERINE A. SANTOMASSINO Department of Microbiology, College of Medicine, University of Florida, Gainesville, Florida Received for publication February 23, 1962
ABSTRACT SCHAECHTER, M. (College of Medicine, University of Florida, Gainesville) AND K. SANTOMASSINO. Lysis of Escherichia coli by sulfhydryl-binding reagents. J. Bacteriol. 84:318325. 1962-Washed suspensions of gram-negative rods were lysed by low concentrations of some sulfhydryl-binding and oxidizing reagents, but not by reducing agents. Some kinetic aspects of this phenomenon were studied with p-chloromercuribenzoate and Escherichia coli B/r. Structures resulting from the action of this reagent consisted of impure cell walls. These could be purified by treatment with trypsin. Cell walls prepared mechanically and cell membranes obtained by lysing protoplasts were not overtly affected by this chemical.
chemicals under investigation were prepared at the same time. Bacteria were grown in Trypticase Soy Broth (BBL). After washing, they were resuspended and tested either in 0.04 M phosphate buffer (pH 7.0) or in the following nitrogen-free medium, termed solution L: MgSO4-7H20, 0.10 (g per liter); citric acid, 1.00; Na2HPO4.2H20, 5.00; KH2PO4, 1.03; KCl, 0.74; glucose, 2.00; pH 7.0. Microscopy. A Wild phase-contrast microscope was used for routine observation. Electron micrographs were taken with a 100A Phillips instrument. Preparations were shadowed with chromium. Bacterial strains. Cultures of this department's stock collection were employed. With the exceptions shown in Table 2, the strains had no special designation.
In the course of physiological investigations, it was noted that low concentrations of some mercurial compounds lysed washed suspensions of Escherichia coli. A search of the literature did not uncover previous descriptions of the phenomenon, despite the long use of such compounds in microbiology. We have undertaken to investigate the susceptibility of various bacterial genera, the activity of selected chemicals, and some kinetic aspects of this mode of lysis.
RESULTS
MATERIALS AND METHODS
Test systems. Bacterial suspensions were prepared as follows. Exponentially growing cultures containing about 100 ,ug of bacteria (dry wt) per ml were rapidly collected on 150-mm membrane filters (Schleicher and Schuell, coarse grade). They were washed and resuspended to about the same density in prewarmed test medium. The suspensions were incubated at 37 C without agitation. Their optical density was measured at 450 m,A in a Zeiss PMQ II spectrophotometer. Control suspensions in solutions which did not contain the ' Present address: Department of Microbiology, Tufts University School of Medicine, Boston, Mass. 318
Effect of p-chloromercuribenzoate (p-CMB) on suspensions of E. coli B/r. The addition of p-CMB to suspensions of E. coli B/r in solution L brought about a marked decrease in optical density (Fig. 1). Lysis of the cells took place in a reproducible manner at any given p-CMB concentration. The addition of 10-5 M p-CMB resulted in the following events. No changes were discernible for 6 to 10 min. Thereafter, turbidity decreased at an accelerated rate over a period of 15 to 20 mi. This was followed by lysis at a maximal, linear rate, where the optical density decreased about 20% of the original per 10-min interval. The rate of lysis then gradually diminished. Various concentrations of p-CMB affected the various components of this lysis curve differently (Fig. 1). The onset of lysis occurred sooner at higher p-CMB concentrations, up to 2 X 10-4 M. The maximal rate of lysis was the same at pCMB concentrations above 5 X 10-6 M. At lower concentrations, lysis proceeded at a slower rate. The total extent of lysis was the same with concentrations of p-CMB between 10-3 and 2 x 10-4 M. However, between 2 X 10-4 and 5 X 10-6 M, slightly more lysis was consistently observed at lower p-CMB concentrations.
VOL.
84, 1962
LYSIS OF E. COLI
319
00
30
60
90 MINUTES
180
150
120
FIG. 1. Effect of different concentrations of p-CMB on suspensions of Escherichia coli B/r. The ordinate expresses the optical density in per cent of that at the time of addition of p-CMB.
The lysis curve obtained with 10-5 M p-CMB was strongly affected by temperature. All its components were considerably slower at 20 C than at 37 C. In this range the length of the initial period during which no changes were perceptible was 2.5 times shorter per 10 C rise in temperature. The rate of maximal lysis increased 4.5 times per 10 C rise. The reaction of p-CMB with the cells apparently took place in a rapid manner. This was shown by the following experiment. Cells were washed 1, 3, 5, and 10 min after the addition of 2 X 10-5 M p-CMB and resuspended in solution L without p-CMB. Cells washed after 1 or 3 min lysed at a slower rate, but to the same extent as unwashed cells. Cells washed 5 or 10 min after pCMB addition behaved like unwashed cells. Reversal of the effect of p-CMB. The addition of 0.3 M sucrose protected E. coli B/r suspended in solution L from p-CMB-induced lysis (Fig. 2). However, under these conditions p-CMB produced latent damage to the cells. Upon threefold dilution with solution L, the optical density (corrected for dilution) dropped precipitously to that of the suspension treated with p-CMB in the absence of sucrose. The addition of a tenfold molar excess of cysteine or sodium hydrosulfide inhibited lysis if added soon after p-CMB. Lysis was greatly reduced if cysteine was added during the first 10 min (Fig. 3). This effect was significant, but much
4-
-
so
200 80
Control
~ ~ ~ ~ ~ ~Sucrose*p-CMB
0
6
j
40 -.
p-CMB Control
20 Not Treated Wth
30
Cysttine 60 90 MINUTES
Trooted
Wih Cystoine
120
FIG. 2. Effect of p-CMB on suspensions of Escherichia coli B/r in solution L containing 0.8 M sucrose and cysteine. To samples treated wvith 105 M p-CMB, 104 M cysteine was added at the times indicated. After 5 min, these and similar samples not treated with cysteine were diluted 1: 3 with solution L. The upper curve in each case represents the sample treated with cysteine, the lower curve the sample not so treated.
smaller, when cysteine was added after 20 min. At this time the optical density had already dropped by about 17%. Cysteine added 40 min after p-CMB, when the optical density had decreased to about one-half of the original, had an insignificant effect. This reversal of the p-CMB effect by cysteine
320
SCHAECHTER AND SANTOMASSINO
30
60 90 MINUTES
J. BACTERIOL..
120
FIG. 3. Effect of cysteine on treatment of suspensions of Escherichia coli B/r with p-CMB. Cysteine (10-4 M) was added at different times after addition of 10-5 M p-CMB to suspensions in solution L.
was also studied in treated cells whose lysis was continued to divide. This continuance of cell inhibited by the presence of sucrose. Cysteine was division upon transfer of growing cells to a meadded at different times after p-CMB to cells sus- dium which does not support growth has been pended in sucrose-containing solution L. After 5 previously described (e.g., Schaechter, 1961). min of contact with cysteine, the suspensions were Cells suspended in solution L containing 2 X 10-5 diluted 1:3 with solution L. As seen in Fig. 2, M p-CMB and no sucrose rapidly lost viability. treatment with cysteine 20 min after p-CMB re- In the presence of sucrose and p-CMB, the loss sulted, upon dilution, in a considerable decrease of viability was considerably slower. After 18 in lysis as compared with the sample not treated min, about 25% of the cells were still viable, with cysteine. This effect was smaller, but still whereas at this time the number of viable cells in significant, when cysteine was added after 40 min. the suspension lacking sucrose was reduced about It was very small, but measurable, when cysteine 104-fold. was added after 60 min. The optical density of Physiological state and sensitivity to p-CMB. untreated cells suspended in 0.3 M sucrose was The importance of the physiological state of E. nearly unaffected by a similar dilution. coli B/r and the composition of the solution in Effect of p-CMB on viability of E. coli B/r. which they were suspended were also investiTable 1 shows the effect of p-CMB on viability of gated. The addition of p-CMB at several consuspensions of E. coli B/r. As in the previous ex- centrations to cells growing in glucose-salts miniperiments, p-CMB was added at zero time to sus- mal medium (or washed out of this medium and pensions in solution L with and without 0.3 M placed in solution L) resulted in about half the sucrose. Samples were taken at different times, amount of lysis seen with broth-grown cells. diluted in 100-fold steps in saline blanks contain- Overnight broth cultures were highly refractory ing '4 volumes of broth and 0.3 M sucrose, and to lysis upon addition of 10-3 to 10-5 M p-CMB. plated on Trypticase Soy Agar plates containing The maximal decrease in optical density was 0.3 M sucrose. Similar results were obtained when about 20% of that before the addition of p-CMB. 10-3 M sodium hydrosulfide was added to the dilu- Broth-grown cells became increasingly more retion blanks and to the agar. The plates were fractory to lysis when placed in solution L. When scored after 24 hr at 37 C. The following results 2 X 10-5 M p-CMB was added to such cells were obtained. In control suspensions, without suspended for 1 hr in solution L, lysis proceeded p-CMB, no loss of viability was seen and the cells at about half the rate seen with freshly washed
321
LYSIS OF E. COLI
VOL. 84, 1962
TABLE 1. Effect of p-CMB on the viability of Escherichia coli B/r X 10-s p-CMB, sucrose no (2
Control Time of sampling min
1 14 28 47
Viable cells per ml
1.54X 2.04 X 2.24 X 2.72 X
10; 107 107 107
X 10-5 p-CMB sucrose(2 (0.3 m) m)
a)
Time of sampling min
Viable cells per ml
3 15 30 48
6.86X 106 1.60 X 103 Less than 10 Less than 10
Time of sampling
Viable cells per ml
min
5 18 33 49
2.05X 5.20 X 2.46 X 6.60 X
107 106 103 102
cell membranes (Fig. 8). Upon p-CMB treatment of this preparation, no changes in optical density or morphology were detected (Fig. 9). Spheroplasts of Salmonella typhimurium obtained by the proceaure of Landman, Altenbern, and Ginosa (1958) and suspended in 0.3 M sucrose showed no effect upon the addition of various concentrations of p-CMB. Sensitivity of various bacterial species to p-CMB. Table 2 shows the results of exposure of various bacterial suspensions to p-CMB for 2 hr. For each species, concentrations ranging from 10-3 to 10-5 M were used. Since the amount of lysis in the absence of p-CMB varied, the results could not be directly compared between species. For this reason, the difference in the lysis seen with and without p-CMB was only a relative measure of sensitivity to this substance. Nonetheless, gramnegative rods, and particularly enteric bacteria, were distinctly more sensitive than were grampositive bacteria. With the exception of Staphylococcus epidermidis, Streptococcus faecalis, and Lactobacillus arabinosus, which showed a small amount of lysis, the gram-positive organisms and the Neisseria catarrhalis tested could be considered resistant to lysis by p-CMB. The shape of the lysis curve of sensitive organisms resembled that of E. coli B/r, except in the case of Proteus vulgaris. In the presence of 10-4 or (Fig. 6). Purified cell walls prepared in the Mickle dis- 10-5 M p-CMB, the optical density of suspensions integrator and subsequently treated with trypsin of this organism began to decrease after about 1 were exposed to 2 X 10-5 M p-CMB for 2 hr. The min. The morphology of the structures resulting optical density remained constant and no morphological change was apparent under the from lysis of different sensitive species was analoelectron microscope. A preparation of impure cell gous. When the extent of lysis was not great, the membranes was made by distilled-water treat- cell envelopes contained larger amounts of cytoment of spheroplasts (obtained after 3 hr of con- plasmic material. Gram-positive organisms did tact of a broth culture in 0.5 M sucrose with 100 not show significant morphological changes even units/ml of penicillin). This preparation cer- when, as with L. arabinosus, the optical density of tainly contained cell-wall material, but was as- the suspension had decreased appreciably. Effect of other chemicals on suspensions of E. coli sumed also to be composed of sizable amounts of
cells. The addition of 0.1% ammonium sulfate, or the withdrawal of glucose from solution L, had no effect on broth-grown cells. Lysis with various concentrations of p-CMB took place more slowly if the cells were suspended in 0.04 M phosphate buffer (pH 7.0). Morphological changes associated with lysis by p-C.M1B. Phase-contrast microscopy of suspensions of E. coli B/r after extensive lysis due to pCMB revealed the presence of membranous structures of the same size and shape as the original bacteria. The number of these structures was the same as that of the cells originally present (when counted with the Coulter Counter, an electronic particle analyzer). These particles were quite homogenous in appearance and no whole cells were seen. Thus, lysis affected all the cells in the suspension to a similar degree. Under the electron microscope, the residual structures were seento consist of cell envelopes containing considerable amounts of shrunken, irregularly shaped cytoplasmic residues (Fig. 4). Amorphous blobs of material could be seen either attached to these structures or free on the supporting film. When suspensions of this material were treated for 2 hr at 37 C with trypsin (1 mg/ml), they became morphologically indistinguishable (Fig. 7) from cell walls of the same organism prepared with the Mickle disintegrator
FIG. 4-9. Electron micrographs of shadowed preparations. The bar in each picture represents 1 ,u. (Fig. 4) Residual structures after treatment of Escherichia coli Blr for 2 hr with 1O- M p-CMB. (Fig. 5) Residual structures after treatment of E. coli B/r for 2 hr with 1-4 M phenylhydrazine. (Fig. 6) Cell walls of E. coli B/r obtained by disintegration in the Mickle shaker with Ballotini beads and treatment with trypsin. (Fig. 7) Cell walls of E. coli B/r obtained by trypsin treatment of the material in Fig. 5. (Fig. 8) Impure cell membranes obtained by osmotic shock of E. coli B/r spheroplasts. (Fig. 9) As in Fig. 8, after treatment with p-CMB (105 M) for 2 hr. 322
TABLE 2. Effect of p-CMB on various bacteria*
PretActual
Organism
Per centPe cent p-CMB ilysis of s lysis - (by control wlCM differ-
PMBence)
Escherichia coli B/r E. coli 15TSalmonella typhimurium LT-2 Aerobacter aerogenes Alkaligenes faecalis Serratia marcescens Proteus vulgaris P. mirabilis Pseudomonas aeruginosa P. fluorescens Neisseria catarrhalis Sarcina lutea Staphylococcus epidermidis Gaffkya tetragena Streptococcus faecalis Lactobacillus arabinosus Bacillus megaterium KM B. subtilis
Corynebacterium hoffmanii
323
LYSIS OF E. COLI
VOL. 84, 1962
10 13 0
15 11 22
15 20 24 7 23 16 7 0 5 20 17 19 9
79 78 70 75 50 67 74 58 58 30
69 65 70 60 39 45 59 38 34 23
22 17
-1 -1
30 18 35
23 18 30 24 -2
44 15 23 20
4 11
Suspensions of the various strains in solution L were treated for 2 hr at 37 C with 2 X 10-5 M pCMB. The results are expressed as per cent decrease in optical density from the time of addition of p-CMB. *
activity. Other compounds listed were largely ineffective under the conditions of the test. The morphological appearance of cells treated with 10-4 M phenylhydrazine was different from that obtained with p-CMB (Fig. 5). The cytoplasmic residues contained within the cell envelopes were homogeneous in appearance and little amorphous material was seen on the outside of the cells. Phenylhydrazine or periodic acid at concentrations of 10-3 to 10-5 M did not lyse nongrowing, overnight cultures. It was of interest to note that most of the agents which were effective in causing lysis reduced the optical density of suspensions to a lesser extent at higher concentrations. This was particularly striking with mercuric chloride which, at 10-3 M, caused a 25% increase in the optical density of the suspension. It is probable that this effect was caused by coagulation of the cytoplasm. This could prevent its leaking out, even if the cell envelopes were greatly damaged. This interesting response was not so extensive with other reagents. However, it was highly reproducible. Despite small variations in the final optical densities of suspensions treated on different days, the inverse concentration dependence was always found with the chemicals indicated.
Blr. In an attempt to determine the specificity
DISCUSSION
of the action of p-CMB, a number of other chemicals were tested. Several other mercurials, as well as other well-known sulfhydryl-binding compounds such as N-ethylmaleimide and iodosobenzoic acid, effectively caused lysis of E. coli (Table 3). These compounds are regarded as highly reactive with sulfhydryl groups, but are not entirely specific since under certain conditions they are known to react with other groups. Other compounds known to react with sulfhydryls, iodoacetic acid and arsenite, had no effect. Lysis was induced by a number of substances which, under the conditions of the test, were thought to be oxidizing. The lysis curves were similar to those obtained with p-CMB. Among them, sodium hypochlorite had two different effects at various concentrations. At high concentrations (10- and 10-3 M), cells were lysed immediately and completely. At lower concentrations (10-5 M), a lysis curve similar to that induced by p-CMB was obtained. One compound, copper sulfate, had intermediate and equivocal
It may be inferentially assumed that the chemically induced lysis here reported is the consequence of damage to cell envelopes. On the basis of the present experiments it is not possible to determine whether structural changes are effected directly (as in the case of lysozyme) or indirectly (as with penicillin). The addition of pCMB to isolated suspensions of cell walls and cell membranes did not result in discernible morphological changes. However, it is possible that these structures are impaired in a functional sense by virtue of damage more subtle than is apparent by electron microscopy. The sensitivity to osmotic shock of p-CMBtreated cells resembles that of protoplasts. Both types of structures are protected from lysis by high osmotic pressure. However, this may not be taken as an indication that p-CMB primarily affects the cell wall, since it is conceivable that cells may withstand lysis at high osmotic pressures when the cell membrane, rather than the cell wall, is damaged.
324
J. BACTERIOL.
SCHAECHTER AND SANTOMASSINO
TABLE 3. Effect of various substances on the optical density of suspensions of Escherichia coli B/r* Substance added
Molar concn
Effect on
integrity
° cells el 1ll 14 10-6 of Io2ESe2
p-Chloromercuribenzoate (1, 2) Phenylmercuriacetate (1, 2) Merthiolate (1) Mercuric chloride (1) N-ethylmaleimide (1) Iodosobenzoic acid (1, 2)
72 73 -
Phenylhydrazine (2) Hydroxylamine (2) Periodic acid (2) Potassium permanganate (2) Sodium hypochlorite (2) Iodine (2)
42 60 49
65 -
+25 65 71
70 55 75 +3 29 62
78 75 61 -
0 43 -
+ + + + + +
65 16 33 70 62 68
-
+ + + + + +
100
69 66 67 0 100
-
-
71 54 60 23 21 33
Copper sulfate (2)
-
14
24
55
-
i
Iodoacetic acid (1, 2) Sodium arsenite (2) Zinc sulfate (2) Perchloric acid (2) Trichloroacetic acid (2) Nitric acid (2) Sodium hydrosulfide (2, 3) Potassium borohydride (1, 2) Sodium bisulfite (2) Sodium cyanide (1) Hydrogen sulfide (1)t
43
20 7
18 16 13 14 14
14 16 13
-
-
14
-
13 15 12 14 11 21
-
-
-
-
-
12 -
18 56 19 12
-
-
13 21 11 15 10
-
15 5 17 14
6 -
-
-
-
12 12 6
-
-
Suspensions of E. coli B/r were treated for 2 hr at 37 C with the compounds listed. The minus sign indicates that the test was not performed with that particular concentration. The compounds marked (1) were tested in solution L, those marked (2) in 0.04 M phosphate buffer (pH 7.0), and those marked (3) in distilled water. The results are expressed as per cent decrease in optical density from the time of addition. They are the average of two to five experiments. When the chemical was tested in two different media, the results of the first listed are given. The action of each chemical tested was similar in both media used. Mercuric chloride (10-3 and 10-4 M) produce and increase in optical density. Compounds which induced lysis, at concentrations of 10-3 M or lower, were considered effective (+), if the optical density was reduced by at least 60%. A 30 to 59%70 drop was considered equivocal (d). If the reduction in optical density was less than 30%, the compounds were considered ineffective (-). t Hydrogen sulfide was bubbled through bacterial suspensions, and gave a result of 7% lysis. *
Several aspects of the kinetics of lysis by p- siderably more than two. A study of the binding CMB are of interest. The interaction of this sub- of p-CMB to various cell components may help stance with cells apparently takes place within a elucidate the phenomenon. few minutes of its addition. For the first 10 min The effectiveness of sulfhydryl-binding comits effects on cell integrity and viability can be pounds in inducing lysis led us to hypothesize partially reversed by an excess of cysteine. The that disulfide bonds in structural proteins play a reversibility of the action of p-CMB is rapidly role in maintaining bacterial-cell integrity. Comlost after this initial period. There is not an abrupt pounds like p-CMB do not affect disulfide bridges transition from a reversible to a nonreversible except under special conditions. However, since condition. Rather, this change is gradual. Thus, if growing cells are sensitive to lysis but nonparticular events leading either to lysis or loss of growing cells are not, it is possible that the formaviability are concatenated, there must be con- tion of disulfide bridges could be impaired by
VOL.
84, 19692
LYSIS OF E. COLI
binding of sulfhydryl groups. This scheme predicts that lysis would also be caused by reagents which can break already formed disulfides, either by oxidation or reduction. However, reagents which specifically carry out such reactions are not available. A number of oxidizing agents tested proved effective. They are highly reactive and could be expected to have other activities on structural components and elicit lysis for other reasons. Indeed, most of them react with carbohydrates. The reducing agents tested were ineffective. We do not know whether these substances actually reduce disulfides under the conditions employed. In view of the chemical complexity of the structural components involved, the hypothesis is difficult to prove. A direct objection to this hypothesis is the fact that nongrowing cultures were not lysed by oxidizing agents. It would be expected that such agents would affect nongrowing cells, whose disulfide bridges may be stable, as well as growing cells. However, it was not determined whether this indeed was the case or whether disulfides of nongrowing cells were unavailable to the reagents. Although we are unable to adduce definitive proof of the role of disulfide bridges in maintaining cell integrity, this hypothesis seems quite reasonable. Impairment of these bonds can be expected to cause changes in surface properties of sufficient magnitude to account for the observed
effects. Systematic differences between gram-positive and gram-negative organisms cannot be invoked to explain differences in sensitivity to chemicals
325
tested. Although gram-negative bacteria are thought to contain a greater complement of proteins in their cell walls, a comparison of the composition of their cell membranes is not possible at the present for lack of adequate data. Lysis by some of the chemical agents used in this work, followed by trypsin treatment, may provide a useful method for the quantitative recovery of purified cell walls. Further work is needed to demonstrate probable chemical changes in the composition of cell walls thus obtained. ACKNOWLEDGMENTS
W. R. Roderick suggested several experiments and interpretations. M. Fried, J. J. Cebra, and R. E. Ecker stimulated the authors through interesting discussions. J. W. Carlisle performed the electron microscopy. The help of these persons is very gratefully acknowledged. This investigation was supported by research grant E-2744 from the National Institute of Allergy and Infectious Diseases, U.S. Public Health Service. The first author holds a Senior Research Fellowship, GSF-4333, from the U.S. Public Health Service. LITERATURE CITED LANDMAN, 0. E., R. A. ALTENBERN, AND H. S. GINOSA. 1958. Quantitative conversion of cells and protoplasts of Proteus mirabilis and Escherichia coli to the L-form. J. Bacteriol. 75:567-576. SCHAECHTER, M. 1961. Patterns of cellular control during unbalanced growth. Cold Spring Harbor Symposia Quant. Biol. 26:53-62.
