ANTIMICROBIAL AGENTS AND CHEMOrHERAPY, Aug. 1978, p. 218-223 0066-4804/78/0014-0218$02.00/0 Copyright © 1978 American Society for Microbiology
Vol, 14, No. 2 Printed in U.S.A.
Effect of a Staphylococcin on Neisseria gonorrhoeae DOROTHY M. MORRISS, JOHN W. LAWSON, AND MARVIN ROGOLSKY* Department of Biology and School of Medicine, University of Missouri-Kansas City, Kansas City, Missouri 64110 Received for publication 18 May 1978
Phage group 2 staphylococcal strain UT0002 contains a large 56S virulence plasmid with genes that code for both exfoliative toxin and a specific staphylococcin termed Bac R1. Four penicillinase-producing strains and three penicillinsusceptible strains of Neisseria gonorrhoeae were killed by Bac R1. Aftejr 30 min of growth of the penicillin-resistant TR1 strain in 62.5 arbitrary units of Bac R1 per ml, loss of viability was approximately 90%, and, after 5 h, an approxim4tely 99.99% loss of viability was observed. Lysis did not accompany cell death, and 84% of the Bac R1 added to the growth medium was adsorbed to the gonococcal cells. The extracellular supernatant fluid from a substrain of staphylococcal strain UT0002 cured of the plasmid for Bac R, production had no lethal effect on the gonococcal strains. Bac R, was also shown to have bactericidal activity against an L-form of N. meningitidis, indicating that the outer envelope of a neisserial cell is not needed for bacteriocin activity. Ten different normal human sera were unable to neutralize Bac R, activity. The bacteriocin lacks adsorption specificity. It binds to but does not kill Escherichia coli cells, indicating that the cell envelope of gram-negative organisms can provide protection against the staphylococcin.
Approximately 3 million new cases of gonorrhea occur annually, with an additional 3 to 5 million secondary infections arising from the base population of patients prior to cure (2). Gonococci have gradually become more resistant to penicillin, tetracycline, and other antibiotics in the past two decades (12, 21). With the advent of penicillinase-producing strains of Neisseria gonorrhoeae in the United States, control of this disease will become increasingly more difficult (2). The continued isolation of penicillinase-producing strains of N. gonorrhoeae which are resistant to high concentrations of penicillin dictates a re-thinking of our approach to the treatment of gonorrhea. The present investigation was initiated to assess the potential of bacteriocins as a new weapon in the arsenal against gonorrhea. Bacteriocins are proteinaceous substances produced by one bacterium with bactericidal activity against the same or a related species of bacteria. The lethal action of bacteriocins is believed to be due to their binding to surface receptors followed by their interaction with an intracellular target (23). The staphylococcins, bacteriocins produced by staphylococci, were first described by Frederique in 1946 (8). He distinguished several different staphylococcins on the basis of their inhibition spectra. Rogolsky et al. (17) demonstrated that certain
phage group 2 staphylococci contain a large 56S virulence plasmid containing genes that code for both exfoliative toxin and a specific staphylococcin. The staphylococcin is active against a wide variety of gram-positive organisms and also against group 2 staphylococcal strains that have been cured of the plasmid carrying the staphylococcin marker, but not against any gram-negative bacteria tested. This staphylococcin was termed Bac R1. The main purpose of the present studies was to investigate the bactericidal action of Bac R1 against N. gonorrhoeae and to gather information about the specific binding properties of this staphylococcin.
MATERLALS AND METHODS Bacterial strains. Cultures of N. gonorrhoeae
strains 474, 448, and 511 were provided by the Neisseria Repository, Naval Research Unit 1, University of California, Berkeley. Strains TR1, TR2, and TR3 were obtained from the Kansas City Department of Public Health. Strain KU! was obtained from the Kansas University Medical Center. Strains TR1, TR2, TR3, and KU1 were all penicillin resistant. In addition, strain TR1 was resistant to spectinomycin, erythromycin, and the tetracyclines. All strains were identified as N. gonorrhoeae by Gram stain, positive oxidase reaction, fermentation with glucose but not with maltose, sucrose, mannitol, lactose, and fructose, and agglutination with gonococcal antisera. Strains of N. meningitidis and their L-forms were obtained from the Neisseria Repository. Group D 218
VOL. 14, 1978
ACTION OF Bac R, AGAINST N. GONORRHOEAE
streptococcus, strain F-24, and its L-form, strain F-24L (ATCC 19634, 19635), were received from H. Gooder, University of North Carolina, Chapel Hill, N.C. Phage group 2 staphylococcal strain UT0002, used for the production of Bac RI, was isolated from clinical sources as described previously (17). Strain UT000219 is a substrain of UT0002 that has been cured of the .56S virulence plasmid and, therefore, is Bac RI negative. The indicator strain for Bac R1 activity, 502A, was supplied by B. F. Anthony and is a staphylococcal phage group 3 strain. Maintenance and storage of bacteria. Cultures of both neisserial species were grown in GC medium base broth (Difco) supplemented with 1% CVA (Grand Island Biological Co.) at 37°C in a G76 gyratory water bath shaker (New Brunswick Scientific Co.). Routine dilutions were stored in tryptic soy broth (Oxoid) or brain heart infusion broth with 6% (wt/vol) lactose at -70°C. Cultures were recovered by strealing on GC medium base agar with 1% CVA. The meningococcal L-form was adapted to tryptic soy broth with 10% (vol/vol) horse serum. Hence, stocks were maintained in this medium, and storage was performed as in the case of the parent organism. Group D streptococcal strain F-24 was maintained in Trypticase soy broth (BBL) and was stored in the same broth with 6% (wt/vol) lactose at -70°C. Strain F-24L was cultured in a medium consisting of 2% (wt/vol) NH4Cl, 0.05% glucose, and Trypticase soy broth minus glucose (BBL), and 1.2% purified agar (Difco) was added to this medium for colony formation. Staphylococcal strains UT0002, UT0002-19, and 502A were maintained on heart infusion agar (Difco) and were stored at -50°C in a solution consisting of 10% dimethyl sulfoxide and 0.5% glucose. Production and partial purification of Bac R1. Bac R1 was produced and partially purified according to the procedure of Rogolsky and Wiley (16). An inoculum of approximately 107 cells of strain UTOO2 per ml was shaken in 80 ml of heart infusion broth at 37°C in a G76 gyratory water bath shaker (New Brunswick Scientific Co.) covered by a Plexiglas hood that contained approximately 10% CO2. After 24 h, the supernatant fluids from at least eight flasks, each containing 80 ml of suspended cells, were combined and precipitated at a final concentration of 60% aqueous (NH4)2SO4. This suspension was then centrifiged, and the pellet was dissolved in approximately 30 ml of 0.02 M phosphate buffer. The resulting clear amber solution was then dialyzed and concentrated as described previously (16). Demonstration of bactericidal activity against the gonococcus. A lawn of the test strain grown overnight on GC medium base and diluted in GC medium base broth was spread on GC medium base agar especially prepared so that 6-mm wells punched into the agar held exactly 100 p1 of the purified bacteriocin. The plates were incubated overnight in approximately 10% C02 at 37°C. The degree of inhibition was determined from the highest dilution of a 250arbitrary units (AU)/ml amount of Bac R1 that produced a definite zone of clearing around the well. Controls were performed with the staphylococcal in-
219
dicator strain, 502A. AU of the Bac RI preparation were derived by using the agar plate assay as described previously (16). The staphylococcal 502A indicator strain was diluted 100-fold and used to prepare a lawn on a heart infusion agar plate into which 6-mm wells with 100-ed capacities were punched. One AU of Bac R1 was then defined as the highest dilution of a specific preparation producing a visible zone of sharp clearing in the bacterial lawn around an agar well holding 100 p1 of Bac Ri. Therefore, AU per milliliter is defined as the reciprocal of this highest dilution times 10. The bactericidal effect of partially purified Bac RI was also determined by treating equivalent cultures of the test strains with 62.5 AU of either Bac R1 or phosphate buffer per ml during growth in broth. The cultures were inoculated to a reading of approximately 10 in a Klett-Suimmerson photoelectric colorimeter (red filter), treated with Bac RI or phosphate buffer, and incubated in the gyratory water bath shaker at 370C at 150 rpm. Growth of the cultures was monitored by optical density (OD) readings and by viable count. To demonstrate that the bactericidal effect was due to Bac RH, other control experiments were run utilizing the supernatant fluid of strain UT0002-19, which had been cured of the 56S plasmid for Bac R1 production (17), in place of the staphylococcin. The supernatant fluid was precipitated with (NH4)2SO4 and concentrated according to the same procedure used to concentrate Bac RI preparations (16). This preparation was then added to the growing neisserial strain at a volume that would have yielded approximately 62.5 AU of Bac R1 per ml if the preparation had been isolated from a staphylococcal UTO02
strain. Cell adsorption experiments. GC medium base was inoculated to a Klett reading of approximately 6 with either a neisserial strain or its L-form. Similarly, group D streptococcus strain F-24 or its L-form was inoculated into appropriate growth media. Cultures were harvested by centrifugation when cell counts reached approximately 108 cells per ml. The pellets were washed twice with either phosphate buffer, pH 7.2, or the designated growth media and then treated with a final concentration of 62.5 AU of Bac RI per ml. After incubation for 1 h at 37°C, the treated cells were centrifuged in the cold at 6,500 rpm for 15 min. The supernatants were removed, filter sterilized, and assayed for residual Bac RI activity by the agar plate assay method (16). Similar experiments were performed on eucaryotic cells. Rat livers were isolated and teased through a fine wire mesh into either Hanks medium or phosphate-buffered saline, pH 7.2. CelLs were then filtered through sterile gauze into 50-ml conical centrifuge tubes and centrifuged in the cold at 1,600 rpm for 15 min. Pellets were washed three times with the same media and then counted, using trypan blue in a counting chamber, and diluted to a final suspension of 108 cells per ml. A 5-ml sample of the liver celLs was centrifuged, and the supernatant was removed. The cells were then resuspended in 3 ml of either Hanks media or phosphate-buffered saline. One milliliter of partially purified Bac RI from staphylococcal strain
MORRISS, LAWSON, AND ROGOLSKY
ANTIMICROB. AGENTS CHEMOTHER.
UT0002 was then added to give a final concentration of 62.5 AU/ml. Mixtures of media plus liver cells or media plus Bac RI served as controls. All mixtures were incubated for 1 h at 370C and then centrifuged. Supernatant fluids were collected, filter sterilized, and assayed for Bac R1 activity. Antibody neutralization experiments. The presence of neutralizing antibody or other substances inhibitory for Bac R1 and, therefore, potentially capable of interfering with antigonococcal activity in vivo was determined. Twofold dilutions of Bac R1 were made directly into sera, or, conversely, twofold dilutions of sera were mixed directly with Bac R1. Tubes were incubated at 370C for 2 h. Residual Bac R1 activity was determined by the agar plate method (16). RESULTS 'Susceptibility of N. gonorrhoeae to Bac
phate buffer was substituted for the bacteriocin in the control culture. Growth was determined during a 5-h period by OD readings and by viable counts of cells plated on GC medium base agar. The staphylococcin had a lethal effect on the susceptible cells without causing lysis (Fig. 1). After 1 h, loss of viability was approximately 90%. After 5 h, 99.99% of the gonococcal cells lost the ability to produce colonies. However, the OD of the culture incubated in the presence of Bac R1 remained constant, indicating that lysis was not accompanying cell death. Both the viable count and the OD of cultures treated with only phosphate buffer increased with time. Strain UT0002-19, cured of the plasmid for Bac R1 synthesis (17), was grown under optimal conditions for Bac R1 synthesis. The supernatant fluid from this culture was then concentrated after (NH4)2SO4 precipitation according to the identical procedure used in the purification of Bac R1 (16). The resulting solution had no bactericidal activity against the neisserial TR1 strain during growth in GC medium. Adsorption of Bac R1 to bacterial cells and bacterial cell fractions. Further investigation involved the study of the interaction of the staphylococcal bacteriocin with surface components of susceptible bacteria and their Lforms. Approximately 2.5 x 108 cells of either the neisseriae or the streptococcal strain, or their respective L-forms, were incubated with 62.5 AU of Bac R1 per ml, for 1 h in either the appropriate growth media or phosphate buffer, as described above. After centrifugation, Bac R1 activity was determined by the agar plate assay. Sixteen percent of the activity of Bac R1 remained in the supernatant fluid after incubation of N. gonorrhoeae TR1 in the presence of Bac R1 for 1 h at 370C in either GC medium or phosphate buffer. This indicated that 84% of the Bac R1 adsorbed to strain TR1 in 1 h and that adsorption was not dependent upon the medium. Eight percent of Bac R1 activity remained in the medium after 1 h of incubation with N. meningitidis, but no Bac RI activity remained after 1 h of incubation with the L-form of N. meningitidis. Both streptococcal strain F-24 and its L-form adsorbed all of the Bac R1 in 1 h. When bacteriocin was added alone to the above solutions or media, 100% of the activity was recovered after 1 h at 370C. Liver-cell suspensions of approximately 2.5 x 108 cells from Sprague-Dawley rats were treated with 62.5 AU of Bac R1 per ml and incubated at 370C for 1 h. Hanks medium or phosphatebuffered saline without cells was treated with bacteriocin in the same manner. After centrifugation, the supernatant from the bacteriocintreated cells was assayed for Bac R1. No Bac R1
220
R1. Four penicillinase-producing strains of N. gonorrhoeae and three penicillin-susceptible strains of N. gonorrhoeae were tested for their susceptibility to Bac R1 (Table 1) by using an agar plate assay. Test strains were appropriately diluted and spread onto heart infusion agar into which wells were punched that received 100 pl of a solution containing either 250 AU of bacteriocin per ml or 1:2 dilutions of this solution. All seven strains of N. gonorrhoeae were inhibited by a final concentration of 62.5 AU of Bac R1 per ml. Strains TR1 and KUl were the most susceptible gonococcal strains distinctly inhibited by a bacteriocin concentration of 31.2 AU/ml. Staphylococcus aureus 502A was more susceptible than any neisserial strain (Table 1). The bactericidal effect of Bac R1 against N. gonorrhoeae TR1 was monitored during growth of the organism in GC medium base to which 62.5 AU of bacteriocin per ml was added. PhosTABLE 1. Susceptibility of N. gonorrhoeae to Bac R1 activity Strain
474 448
Inhibition' at a bacteriocin concn (AU/ml) of: 250
125
62.5
31.25
+ +
+ +
i +
-
15.6 -
±
-
+ 511 + + + + + TR1 + ± TR2 + + TR3 + + + KUl + + + + S. aureus 502A + + + + + a A lawn of an appropriately diluted test strain was made on GC medium base agar into which wells were punched that received 100 pl of either a solution containing 250 AU of Bac R, per ml or 1:2 dilutions of this solution. b Inhibition was indicated by a sharp visible zone of clearing in the bacterial lawn around the wells containing bacteriocin: +, inhibition; -, no inhibition; ±, miinimal inhibition.
VoiL 14, 1978
ACTION OF Bac R1 AGAINST N. GONORRHOEAE
221
for 1 h in phosphate-buffered saline, was treated with 62.5 AU of Bac R1 per ml. After 1 additional h of incubation at 370C, the test solution was assayed for Bac R1 activity. There was no loss of activity. Bactericidal effect of Bac R1 on an N. meningitidis L-form. Since broth cultures of the N. gonorrhoeae L-forms were unavailable for testing, a stable L-form of N. meningitidis group A was used to determine the bactericidal effect of Bac R1 on an L-form. The L-form was grown in GC medium base broth containing 10% horse serum with 62.5 AU of Bac R1 per ml. As a control, phosphate buffer was added to a similarly treated culture. Growth of both cultures was determined in a water bath incubator shaker E at 370C over a 5-h period by OD and by viable counts on GC medium base agar with 10% horse serum (Fig. 2). Although the growth of the menz ingococcal L-form did not proceed as rapidly as that of the gonococcal strain, the death curve of the Bac RI-treated L-forn clearly followed that of the gonococcal strain. After 1 h, loss of viability of the L-forms was 90%. After 5 h, 99.99% of the L-form cells lost the ability to produce colonies. Neutralizing effect of fresh normal sera on Bac R1 activity. To assess the effect of normal human serum on the bactericidal activity of the bacteriocin, twofold dilutions of a Bac RI solution of 250 AU/ml were made directly into so0_o sera, or, altematively, twofold dilutions in phosphate buffer of the Bac RI solution (250 AU/ml) I z similarly were mixed with equal volumes of each 10.n sera to be tested. The tubes were incubated at :I _---370C for 1 to 2 h. A control of bacteriocin diluted 4 'U in phosphate buffer was similarly treated. Resid"E ual Bac RI activity was determined by the agar plate method. Six different sera in which the 4 5 3 1 2 Bac RI was serially diluted through a 1:20 diluTIME (HOURS) tion showed no inhibitory effect. When twofold FIG. 1. Bactericidal effect of Bac R, on N. gonor- dilutions of the bacteriocin in phosphate buffer rhoeae TRi cells. Equivalent cultures of strain TRI were mixed with equal volumes of undiluted serum, no neutralization of the Bac R, was noted were treated with either Bac R, (62.5 AU/mi) or phosphate buffer. Growth of the cultures containing in the four sera tested. In summary, 10 different Bac R1 was monitored by OD readings from a Klett- normal human sera were unable to neutralize Summerson photoelectric colorimeter (red filer) (0) Bac RI activity within the sensitivity of our and by viable colony count (A). Growth of the control assay method. .
40
__S___S___+'
.
2
culture was monitored by OD readings from a KlettSummerson photoelectric colorimeter (red filter) (0) and by viable colony counts (A).
was recovered. There was no loss of Bac R, activity from control preparations. This implies that Bac R1 adsorbs to the eucaryotic cells. Liver cells produce abundant extracellular proteases that could have accounted for loss of Bac R1 activity. Therefore, the supernatant fluid from 2.5 x 108 liver cells, which were incubated
DISCUSSION Seven strains of N. gonorrhoeae showed marked inhibition of growth on agar plates after exposure to the staphylococcin, Bac RI, which was produced by the phage group 2 staphylococcal strain UTO002. Three of the gonococcal strains were penicillin-susceptible laboratory strains, and the remaining four were penicillinase-producing strains. One strain, TR1, was also
222
MORRISS, LAWSON, AND ROGOLSKY
seria does not possess a typical gram-negative cell wall, and this is further supported by its unusual susceptibility to penicillin. Hebeler and Young (10) could not detect a lipoprotein in N. gonorrhoeae analogous to that observed in Escherichia coli which links the peptidoglycan with the outer membrane. Electron micrographs indicated that the outer membrane of the N. gonorrhoeae is, in fact, more loosely associated with peptidoglycan than is E. coli. The bactericidal effect of Bac R1 on N. gonorrhoeae TR1 in broth is comparable to the death rates reported against other genera (5, 13, 16). The extracellular supernatant fluid from strain UT0002-19, the Bac RI-negative substrain of UT0002 that has been cured of the 56S virulence plasmid, did not have a lethal effect on the TR1 strain. This indicated that the staphylococcin coded for by the 56S virulence plasmid of strain UT0002 was responsible for the death of the gonococcal cells. Exfoliative toxin that is also coded by the 56S plasmid has been previously shown not to have bacteriocin activity (16). Ten normal human sera were unable to neutralize Bac R1 activity, which indicates a lack of
8
10
7-
10
S
10
yr
-
-mdr --
,&-
-
A
I
I0
1j .
k
z
a Z
2 O
10 . z
o
t
0
ANTIMICROB. AGENTS CHEMOTHER.
3
10
specificity or nonspecific binding to serum components, no loss of stability, and the absence of antibodies in the random sera tested. Dajani (4), but not Gagliano and Hinsdill (9), was able to demonstrate neutralization of a phage group 2 2* \A staphylococcal bacteriocin by fresh human sera. 10 Dajani incubated specimens for 18 to 24 h, in TIME (HOURS) contrast to the 1 to 2 h of incubation used in FIG. 2. Bactericidal effect of Bac R1 on a N. men- these studies which is usually adequate for the ingitidis L-form. Equivalent cultures of the N. men- demonstration of antibody activity. Perhaps the ingitidis L-form were treated with either Bac R, (62.5 neutralization activity Dajani observed was due AU/ml) or phosphate buffer. Growth of the cultures to an enzyme or some nonspecific effect. On the containing Bac R1 was monitored by OD readings other hand, the limits of our assay method may from a Klett-Summerson photoelectric colorimeter have prevented detection of low levels of neu(red filter) (-) and by viable colony count (A). Growth of the control culture was monitored by OD readings tralization. Although, in many cases, the adsorption of from a Klett-Summerson photoelectric colorirneter bacteriocin has been shown to be highly specific (red filter) (0) and by viable colony counts (A). for susceptible bacteria (1, 6), other bacteriocins, such as staphylococcin 414 (9) and 1580 (14), resistant to erythromycin, spectinomycin, and lactocin LP 27 (24,25), and streptococcin B74628 tetracycline. Shtibel (19) noted inhibition of N. (22), appear to lack this adsorption specificity. gonorrhoeae by bacterial interference and sug- Bac R1 adsorbed not only to neisserial and strepgested that the inhibition might be due to bac- tococcal bacterial strains and their stable Lteriocins. forms, but also to rat liver cells and to E. coli, a Bac RI, previously described by Rogolsky and species whose growth is not inhibited by Bac R1 Wiley (16), possesses all the usual properties of (16). Loss of viability of the L-form was compaa bacteriocin. The inhibitory activity is associrable to that of the neisserial parent after treatated with protein, and Bac R1 is bactericidal to ment by Bac R1. the same and related species without being auOther investigators (1, 7, 9, 11, 15) have retoinhibitory. Although Bac R1 has a wide spec- ported susceptibility of protoplasts or L-forms trum of activity (16), the only gram-negative to bacteriocins. These studies and our own oborganisms found to be susceptible to Bac R1 servations suggest that the receptor sites of the were the Neisseria. This indicated that Neisbacteriocins are located on the inner cell mem"A
ACTION OF Bac R1 AGAINST N. GONORRHOEAE
VOL. 14, 1978
brane. Naturally resistant bacteria may have the access to their membrane receptor regions physically blocked by outer components of the cell envelope (3, 20, 23). Neisserial cell walls are thought to differ from other gram-negative cell walls in being more porous due to either a lack of lipoprotein or a typical outer membrane. This may explain the lethal effect of Bac R1 on Neisseria. This research was supported by grant 2-76-072 from the American Nurses Foundation.
LrITERATURE CITED 1. Anastasio, K. L., G. A. Soucheck, and H. Sugiyama. 1971. Boticinogeny and actions of the bacteriocin. J. Bacteriol. 107:143-149. 2. Berman, G. D. 1977. How dangerous is penicillin resistant gonorrhoeae? Hosp. Physician 3:20-30. 3. Bhattacharyya, P., L. Wnedt, E. Whitney, and S. Silver. 1970. Colicintolerant mutants of Escherichia coli: resistance of membranes to colicin El. Science 168:998-1000. 4. Dajani, A. S. 1973. Neutralization of phage type 71 staphylococcal bacteriocin by immune and non-immune sera. J. Infect. Dis. 128:494-499. 5. Dajani, A. S., E. D. Gray, and L W. Wannasmaker. 1970. Effect of bactericidal substance from Staphylococcus aureus on group A streptococci. I. Biochemical alterations. Infect. Immun. 1:485-490. 6. Dajani, A. S., and L W. Wannamaker. 1973. Kinetic studies on the interaction of bacteriophage type 71 staphylococcal bacteriocin with susceptible bacteria. J. Bacteriol. 114:738-742. 7. Ellison, J. S., and J. A. Kautter. 1970. Purification and some properties of two boticins. J. Bacteriol. 104:19-26. 8. Fredericq, P. 1946. Sur la sensibilite et l'activite antibiotique des Staphylocoques. C. R. Seances Soc. Biol. Paris 140:1167-1170. 9. Gagliano, V. J., and RI D. Hinsdill. 1970. Characterization of a Staphylococcus aureus bacteriocin. J. Bacteriol. 104:117-125. 10. Hebeler, B. H., and F. E. Young. 1976. Chemical composition and turnover of peptidoglycan in Neisseria gonorrhoeae. J. Bacteriol. 126:1180-1185. 11. Ivanovics, G., L Alfoldi, and E. Nagy. 1959. Mode of action of megacin. J. Gen. Microbiol. 21:51-4.
223
12. Jaffe, H. W., J. W. Biddle, C. Thornsberry, R. E. Johnson, R. E. Kaufnman, G. H. Reynolds, and P. J. Wiesner. 1976. National gonorrhoeae therapy monitoring study. In vitro antibiotic susceptibility and its correlation with treatment results. N. Engl. J. Med. 294:5-9. 13. Jetten, A. M., and G. D. Vogels. 1972. Mode of action of a Staphylococcus epidermidis bacteriocin. Antimicrob. Agents Chemother. 2:456-463. 14. Jetten, A. ML, and G. E. Vogels. 1972. Nature and properties of a Staphylococcus epidermnidis bacteriocin. J. Bacteriol. 112:243-250. 15. Jetten, A. M., and G. D. Vogels. 1974. Characteristics of the killing effect of a Staphylococcus epidermidis bacteriocin. Antonie van Leeuwenhoek J. Microbiol. Serol. 40:177-183. 16. Rogolsky, M., and B. B. Wiley. 1977. Production and properties of a staphylococcin genetically controlled by the staphylococcal plasmid for exfoliative toxin synthesis. Infect. Immun. 8:156-164. 17. Rogolsky, M., B. B. Wiley, and L A. Glasgow. 1976. Phage group H staphylococcal strains with chromosomal and extrachromosomal genes for exfoliative toxin production. Infect. Immun. 13:44-52. 18. Schailehn, G., and J. Kiamer. 1975. Studies on mode of action of a bacteriocin from Clostridium septicum. Can. J. Microbiol. 22:435-437. 19. Shtibel, RI 1976. Inhibition of growth of Neisseria gonorrhoeae by bacterial interference. Can. J. Microbiol. 22:1430-1436. 20. Smarha, J., and M. Havelkova. 1970. The effect of colicin G on spheroplasts of Proteus mirabilis. Fola Microbiol. (Prague) 15:122-124. 21. Sparling, P. F. 1972. Antibiotic resistance in Neisseria gonorrhoeae. Med. Clin. North Am. 56:1133-1144. 22. Tagg, J. R., A. S. Dajani, and L W. Wannamaker. 1975. Bacteriocin of a group B streptococcus: partial purification and characterization. Antimicrob. Agents Chemother. 7:764-772. 23. Tagg, J. RI, A. S. Dajani, and L W. Wannamaker. 1976. Bacteriocins of gram-positive bacteria. Bacteriol. Rev. 40:722-756. 24. Upreti, G. C., and RI D. Hinsdill. 1973. Isolation and characterization of a bacteriocin from a homofermentative LactobaciUus. Antimicrob. Agents Chemother. 4:487-494. 25. Upreti, G. C., and R. D. Hinalill. 1975. Production and mode of action of lactocin 27: bacteriocin from a homofermentative Lactobacillus. Antimicrob. Agents Chemother. 7:139-145.
Vol, 14, No. 2 Printed in U.S.A.
Effect of a Staphylococcin on Neisseria gonorrhoeae DOROTHY M. MORRISS, JOHN W. LAWSON, AND MARVIN ROGOLSKY* Department of Biology and School of Medicine, University of Missouri-Kansas City, Kansas City, Missouri 64110 Received for publication 18 May 1978
Phage group 2 staphylococcal strain UT0002 contains a large 56S virulence plasmid with genes that code for both exfoliative toxin and a specific staphylococcin termed Bac R1. Four penicillinase-producing strains and three penicillinsusceptible strains of Neisseria gonorrhoeae were killed by Bac R1. Aftejr 30 min of growth of the penicillin-resistant TR1 strain in 62.5 arbitrary units of Bac R1 per ml, loss of viability was approximately 90%, and, after 5 h, an approxim4tely 99.99% loss of viability was observed. Lysis did not accompany cell death, and 84% of the Bac R1 added to the growth medium was adsorbed to the gonococcal cells. The extracellular supernatant fluid from a substrain of staphylococcal strain UT0002 cured of the plasmid for Bac R, production had no lethal effect on the gonococcal strains. Bac R, was also shown to have bactericidal activity against an L-form of N. meningitidis, indicating that the outer envelope of a neisserial cell is not needed for bacteriocin activity. Ten different normal human sera were unable to neutralize Bac R, activity. The bacteriocin lacks adsorption specificity. It binds to but does not kill Escherichia coli cells, indicating that the cell envelope of gram-negative organisms can provide protection against the staphylococcin.
Approximately 3 million new cases of gonorrhea occur annually, with an additional 3 to 5 million secondary infections arising from the base population of patients prior to cure (2). Gonococci have gradually become more resistant to penicillin, tetracycline, and other antibiotics in the past two decades (12, 21). With the advent of penicillinase-producing strains of Neisseria gonorrhoeae in the United States, control of this disease will become increasingly more difficult (2). The continued isolation of penicillinase-producing strains of N. gonorrhoeae which are resistant to high concentrations of penicillin dictates a re-thinking of our approach to the treatment of gonorrhea. The present investigation was initiated to assess the potential of bacteriocins as a new weapon in the arsenal against gonorrhea. Bacteriocins are proteinaceous substances produced by one bacterium with bactericidal activity against the same or a related species of bacteria. The lethal action of bacteriocins is believed to be due to their binding to surface receptors followed by their interaction with an intracellular target (23). The staphylococcins, bacteriocins produced by staphylococci, were first described by Frederique in 1946 (8). He distinguished several different staphylococcins on the basis of their inhibition spectra. Rogolsky et al. (17) demonstrated that certain
phage group 2 staphylococci contain a large 56S virulence plasmid containing genes that code for both exfoliative toxin and a specific staphylococcin. The staphylococcin is active against a wide variety of gram-positive organisms and also against group 2 staphylococcal strains that have been cured of the plasmid carrying the staphylococcin marker, but not against any gram-negative bacteria tested. This staphylococcin was termed Bac R1. The main purpose of the present studies was to investigate the bactericidal action of Bac R1 against N. gonorrhoeae and to gather information about the specific binding properties of this staphylococcin.
MATERLALS AND METHODS Bacterial strains. Cultures of N. gonorrhoeae
strains 474, 448, and 511 were provided by the Neisseria Repository, Naval Research Unit 1, University of California, Berkeley. Strains TR1, TR2, and TR3 were obtained from the Kansas City Department of Public Health. Strain KU! was obtained from the Kansas University Medical Center. Strains TR1, TR2, TR3, and KU1 were all penicillin resistant. In addition, strain TR1 was resistant to spectinomycin, erythromycin, and the tetracyclines. All strains were identified as N. gonorrhoeae by Gram stain, positive oxidase reaction, fermentation with glucose but not with maltose, sucrose, mannitol, lactose, and fructose, and agglutination with gonococcal antisera. Strains of N. meningitidis and their L-forms were obtained from the Neisseria Repository. Group D 218
VOL. 14, 1978
ACTION OF Bac R, AGAINST N. GONORRHOEAE
streptococcus, strain F-24, and its L-form, strain F-24L (ATCC 19634, 19635), were received from H. Gooder, University of North Carolina, Chapel Hill, N.C. Phage group 2 staphylococcal strain UT0002, used for the production of Bac RI, was isolated from clinical sources as described previously (17). Strain UT000219 is a substrain of UT0002 that has been cured of the .56S virulence plasmid and, therefore, is Bac RI negative. The indicator strain for Bac R1 activity, 502A, was supplied by B. F. Anthony and is a staphylococcal phage group 3 strain. Maintenance and storage of bacteria. Cultures of both neisserial species were grown in GC medium base broth (Difco) supplemented with 1% CVA (Grand Island Biological Co.) at 37°C in a G76 gyratory water bath shaker (New Brunswick Scientific Co.). Routine dilutions were stored in tryptic soy broth (Oxoid) or brain heart infusion broth with 6% (wt/vol) lactose at -70°C. Cultures were recovered by strealing on GC medium base agar with 1% CVA. The meningococcal L-form was adapted to tryptic soy broth with 10% (vol/vol) horse serum. Hence, stocks were maintained in this medium, and storage was performed as in the case of the parent organism. Group D streptococcal strain F-24 was maintained in Trypticase soy broth (BBL) and was stored in the same broth with 6% (wt/vol) lactose at -70°C. Strain F-24L was cultured in a medium consisting of 2% (wt/vol) NH4Cl, 0.05% glucose, and Trypticase soy broth minus glucose (BBL), and 1.2% purified agar (Difco) was added to this medium for colony formation. Staphylococcal strains UT0002, UT0002-19, and 502A were maintained on heart infusion agar (Difco) and were stored at -50°C in a solution consisting of 10% dimethyl sulfoxide and 0.5% glucose. Production and partial purification of Bac R1. Bac R1 was produced and partially purified according to the procedure of Rogolsky and Wiley (16). An inoculum of approximately 107 cells of strain UTOO2 per ml was shaken in 80 ml of heart infusion broth at 37°C in a G76 gyratory water bath shaker (New Brunswick Scientific Co.) covered by a Plexiglas hood that contained approximately 10% CO2. After 24 h, the supernatant fluids from at least eight flasks, each containing 80 ml of suspended cells, were combined and precipitated at a final concentration of 60% aqueous (NH4)2SO4. This suspension was then centrifiged, and the pellet was dissolved in approximately 30 ml of 0.02 M phosphate buffer. The resulting clear amber solution was then dialyzed and concentrated as described previously (16). Demonstration of bactericidal activity against the gonococcus. A lawn of the test strain grown overnight on GC medium base and diluted in GC medium base broth was spread on GC medium base agar especially prepared so that 6-mm wells punched into the agar held exactly 100 p1 of the purified bacteriocin. The plates were incubated overnight in approximately 10% C02 at 37°C. The degree of inhibition was determined from the highest dilution of a 250arbitrary units (AU)/ml amount of Bac R1 that produced a definite zone of clearing around the well. Controls were performed with the staphylococcal in-
219
dicator strain, 502A. AU of the Bac RI preparation were derived by using the agar plate assay as described previously (16). The staphylococcal 502A indicator strain was diluted 100-fold and used to prepare a lawn on a heart infusion agar plate into which 6-mm wells with 100-ed capacities were punched. One AU of Bac R1 was then defined as the highest dilution of a specific preparation producing a visible zone of sharp clearing in the bacterial lawn around an agar well holding 100 p1 of Bac Ri. Therefore, AU per milliliter is defined as the reciprocal of this highest dilution times 10. The bactericidal effect of partially purified Bac RI was also determined by treating equivalent cultures of the test strains with 62.5 AU of either Bac R1 or phosphate buffer per ml during growth in broth. The cultures were inoculated to a reading of approximately 10 in a Klett-Suimmerson photoelectric colorimeter (red filter), treated with Bac RI or phosphate buffer, and incubated in the gyratory water bath shaker at 370C at 150 rpm. Growth of the cultures was monitored by optical density (OD) readings and by viable count. To demonstrate that the bactericidal effect was due to Bac RH, other control experiments were run utilizing the supernatant fluid of strain UT0002-19, which had been cured of the 56S plasmid for Bac R1 production (17), in place of the staphylococcin. The supernatant fluid was precipitated with (NH4)2SO4 and concentrated according to the same procedure used to concentrate Bac RI preparations (16). This preparation was then added to the growing neisserial strain at a volume that would have yielded approximately 62.5 AU of Bac R1 per ml if the preparation had been isolated from a staphylococcal UTO02
strain. Cell adsorption experiments. GC medium base was inoculated to a Klett reading of approximately 6 with either a neisserial strain or its L-form. Similarly, group D streptococcus strain F-24 or its L-form was inoculated into appropriate growth media. Cultures were harvested by centrifugation when cell counts reached approximately 108 cells per ml. The pellets were washed twice with either phosphate buffer, pH 7.2, or the designated growth media and then treated with a final concentration of 62.5 AU of Bac RI per ml. After incubation for 1 h at 37°C, the treated cells were centrifuged in the cold at 6,500 rpm for 15 min. The supernatants were removed, filter sterilized, and assayed for residual Bac RI activity by the agar plate assay method (16). Similar experiments were performed on eucaryotic cells. Rat livers were isolated and teased through a fine wire mesh into either Hanks medium or phosphate-buffered saline, pH 7.2. CelLs were then filtered through sterile gauze into 50-ml conical centrifuge tubes and centrifuged in the cold at 1,600 rpm for 15 min. Pellets were washed three times with the same media and then counted, using trypan blue in a counting chamber, and diluted to a final suspension of 108 cells per ml. A 5-ml sample of the liver celLs was centrifuged, and the supernatant was removed. The cells were then resuspended in 3 ml of either Hanks media or phosphate-buffered saline. One milliliter of partially purified Bac RI from staphylococcal strain
MORRISS, LAWSON, AND ROGOLSKY
ANTIMICROB. AGENTS CHEMOTHER.
UT0002 was then added to give a final concentration of 62.5 AU/ml. Mixtures of media plus liver cells or media plus Bac RI served as controls. All mixtures were incubated for 1 h at 370C and then centrifuged. Supernatant fluids were collected, filter sterilized, and assayed for Bac R1 activity. Antibody neutralization experiments. The presence of neutralizing antibody or other substances inhibitory for Bac R1 and, therefore, potentially capable of interfering with antigonococcal activity in vivo was determined. Twofold dilutions of Bac R1 were made directly into sera, or, conversely, twofold dilutions of sera were mixed directly with Bac R1. Tubes were incubated at 370C for 2 h. Residual Bac R1 activity was determined by the agar plate method (16). RESULTS 'Susceptibility of N. gonorrhoeae to Bac
phate buffer was substituted for the bacteriocin in the control culture. Growth was determined during a 5-h period by OD readings and by viable counts of cells plated on GC medium base agar. The staphylococcin had a lethal effect on the susceptible cells without causing lysis (Fig. 1). After 1 h, loss of viability was approximately 90%. After 5 h, 99.99% of the gonococcal cells lost the ability to produce colonies. However, the OD of the culture incubated in the presence of Bac R1 remained constant, indicating that lysis was not accompanying cell death. Both the viable count and the OD of cultures treated with only phosphate buffer increased with time. Strain UT0002-19, cured of the plasmid for Bac R1 synthesis (17), was grown under optimal conditions for Bac R1 synthesis. The supernatant fluid from this culture was then concentrated after (NH4)2SO4 precipitation according to the identical procedure used in the purification of Bac R1 (16). The resulting solution had no bactericidal activity against the neisserial TR1 strain during growth in GC medium. Adsorption of Bac R1 to bacterial cells and bacterial cell fractions. Further investigation involved the study of the interaction of the staphylococcal bacteriocin with surface components of susceptible bacteria and their Lforms. Approximately 2.5 x 108 cells of either the neisseriae or the streptococcal strain, or their respective L-forms, were incubated with 62.5 AU of Bac R1 per ml, for 1 h in either the appropriate growth media or phosphate buffer, as described above. After centrifugation, Bac R1 activity was determined by the agar plate assay. Sixteen percent of the activity of Bac R1 remained in the supernatant fluid after incubation of N. gonorrhoeae TR1 in the presence of Bac R1 for 1 h at 370C in either GC medium or phosphate buffer. This indicated that 84% of the Bac R1 adsorbed to strain TR1 in 1 h and that adsorption was not dependent upon the medium. Eight percent of Bac R1 activity remained in the medium after 1 h of incubation with N. meningitidis, but no Bac RI activity remained after 1 h of incubation with the L-form of N. meningitidis. Both streptococcal strain F-24 and its L-form adsorbed all of the Bac R1 in 1 h. When bacteriocin was added alone to the above solutions or media, 100% of the activity was recovered after 1 h at 370C. Liver-cell suspensions of approximately 2.5 x 108 cells from Sprague-Dawley rats were treated with 62.5 AU of Bac R1 per ml and incubated at 370C for 1 h. Hanks medium or phosphatebuffered saline without cells was treated with bacteriocin in the same manner. After centrifugation, the supernatant from the bacteriocintreated cells was assayed for Bac R1. No Bac R1
220
R1. Four penicillinase-producing strains of N. gonorrhoeae and three penicillin-susceptible strains of N. gonorrhoeae were tested for their susceptibility to Bac R1 (Table 1) by using an agar plate assay. Test strains were appropriately diluted and spread onto heart infusion agar into which wells were punched that received 100 pl of a solution containing either 250 AU of bacteriocin per ml or 1:2 dilutions of this solution. All seven strains of N. gonorrhoeae were inhibited by a final concentration of 62.5 AU of Bac R1 per ml. Strains TR1 and KUl were the most susceptible gonococcal strains distinctly inhibited by a bacteriocin concentration of 31.2 AU/ml. Staphylococcus aureus 502A was more susceptible than any neisserial strain (Table 1). The bactericidal effect of Bac R1 against N. gonorrhoeae TR1 was monitored during growth of the organism in GC medium base to which 62.5 AU of bacteriocin per ml was added. PhosTABLE 1. Susceptibility of N. gonorrhoeae to Bac R1 activity Strain
474 448
Inhibition' at a bacteriocin concn (AU/ml) of: 250
125
62.5
31.25
+ +
+ +
i +
-
15.6 -
±
-
+ 511 + + + + + TR1 + ± TR2 + + TR3 + + + KUl + + + + S. aureus 502A + + + + + a A lawn of an appropriately diluted test strain was made on GC medium base agar into which wells were punched that received 100 pl of either a solution containing 250 AU of Bac R, per ml or 1:2 dilutions of this solution. b Inhibition was indicated by a sharp visible zone of clearing in the bacterial lawn around the wells containing bacteriocin: +, inhibition; -, no inhibition; ±, miinimal inhibition.
VoiL 14, 1978
ACTION OF Bac R1 AGAINST N. GONORRHOEAE
221
for 1 h in phosphate-buffered saline, was treated with 62.5 AU of Bac R1 per ml. After 1 additional h of incubation at 370C, the test solution was assayed for Bac R1 activity. There was no loss of activity. Bactericidal effect of Bac R1 on an N. meningitidis L-form. Since broth cultures of the N. gonorrhoeae L-forms were unavailable for testing, a stable L-form of N. meningitidis group A was used to determine the bactericidal effect of Bac R1 on an L-form. The L-form was grown in GC medium base broth containing 10% horse serum with 62.5 AU of Bac R1 per ml. As a control, phosphate buffer was added to a similarly treated culture. Growth of both cultures was determined in a water bath incubator shaker E at 370C over a 5-h period by OD and by viable counts on GC medium base agar with 10% horse serum (Fig. 2). Although the growth of the menz ingococcal L-form did not proceed as rapidly as that of the gonococcal strain, the death curve of the Bac RI-treated L-forn clearly followed that of the gonococcal strain. After 1 h, loss of viability of the L-forms was 90%. After 5 h, 99.99% of the L-form cells lost the ability to produce colonies. Neutralizing effect of fresh normal sera on Bac R1 activity. To assess the effect of normal human serum on the bactericidal activity of the bacteriocin, twofold dilutions of a Bac RI solution of 250 AU/ml were made directly into so0_o sera, or, altematively, twofold dilutions in phosphate buffer of the Bac RI solution (250 AU/ml) I z similarly were mixed with equal volumes of each 10.n sera to be tested. The tubes were incubated at :I _---370C for 1 to 2 h. A control of bacteriocin diluted 4 'U in phosphate buffer was similarly treated. Resid"E ual Bac RI activity was determined by the agar plate method. Six different sera in which the 4 5 3 1 2 Bac RI was serially diluted through a 1:20 diluTIME (HOURS) tion showed no inhibitory effect. When twofold FIG. 1. Bactericidal effect of Bac R, on N. gonor- dilutions of the bacteriocin in phosphate buffer rhoeae TRi cells. Equivalent cultures of strain TRI were mixed with equal volumes of undiluted serum, no neutralization of the Bac R, was noted were treated with either Bac R, (62.5 AU/mi) or phosphate buffer. Growth of the cultures containing in the four sera tested. In summary, 10 different Bac R1 was monitored by OD readings from a Klett- normal human sera were unable to neutralize Summerson photoelectric colorimeter (red filer) (0) Bac RI activity within the sensitivity of our and by viable colony count (A). Growth of the control assay method. .
40
__S___S___+'
.
2
culture was monitored by OD readings from a KlettSummerson photoelectric colorimeter (red filter) (0) and by viable colony counts (A).
was recovered. There was no loss of Bac R, activity from control preparations. This implies that Bac R1 adsorbs to the eucaryotic cells. Liver cells produce abundant extracellular proteases that could have accounted for loss of Bac R1 activity. Therefore, the supernatant fluid from 2.5 x 108 liver cells, which were incubated
DISCUSSION Seven strains of N. gonorrhoeae showed marked inhibition of growth on agar plates after exposure to the staphylococcin, Bac RI, which was produced by the phage group 2 staphylococcal strain UTO002. Three of the gonococcal strains were penicillin-susceptible laboratory strains, and the remaining four were penicillinase-producing strains. One strain, TR1, was also
222
MORRISS, LAWSON, AND ROGOLSKY
seria does not possess a typical gram-negative cell wall, and this is further supported by its unusual susceptibility to penicillin. Hebeler and Young (10) could not detect a lipoprotein in N. gonorrhoeae analogous to that observed in Escherichia coli which links the peptidoglycan with the outer membrane. Electron micrographs indicated that the outer membrane of the N. gonorrhoeae is, in fact, more loosely associated with peptidoglycan than is E. coli. The bactericidal effect of Bac R1 on N. gonorrhoeae TR1 in broth is comparable to the death rates reported against other genera (5, 13, 16). The extracellular supernatant fluid from strain UT0002-19, the Bac RI-negative substrain of UT0002 that has been cured of the 56S virulence plasmid, did not have a lethal effect on the TR1 strain. This indicated that the staphylococcin coded for by the 56S virulence plasmid of strain UT0002 was responsible for the death of the gonococcal cells. Exfoliative toxin that is also coded by the 56S plasmid has been previously shown not to have bacteriocin activity (16). Ten normal human sera were unable to neutralize Bac R1 activity, which indicates a lack of
8
10
7-
10
S
10
yr
-
-mdr --
,&-
-
A
I
I0
1j .
k
z
a Z
2 O
10 . z
o
t
0
ANTIMICROB. AGENTS CHEMOTHER.
3
10
specificity or nonspecific binding to serum components, no loss of stability, and the absence of antibodies in the random sera tested. Dajani (4), but not Gagliano and Hinsdill (9), was able to demonstrate neutralization of a phage group 2 2* \A staphylococcal bacteriocin by fresh human sera. 10 Dajani incubated specimens for 18 to 24 h, in TIME (HOURS) contrast to the 1 to 2 h of incubation used in FIG. 2. Bactericidal effect of Bac R1 on a N. men- these studies which is usually adequate for the ingitidis L-form. Equivalent cultures of the N. men- demonstration of antibody activity. Perhaps the ingitidis L-form were treated with either Bac R, (62.5 neutralization activity Dajani observed was due AU/ml) or phosphate buffer. Growth of the cultures to an enzyme or some nonspecific effect. On the containing Bac R1 was monitored by OD readings other hand, the limits of our assay method may from a Klett-Summerson photoelectric colorimeter have prevented detection of low levels of neu(red filter) (-) and by viable colony count (A). Growth of the control culture was monitored by OD readings tralization. Although, in many cases, the adsorption of from a Klett-Summerson photoelectric colorirneter bacteriocin has been shown to be highly specific (red filter) (0) and by viable colony counts (A). for susceptible bacteria (1, 6), other bacteriocins, such as staphylococcin 414 (9) and 1580 (14), resistant to erythromycin, spectinomycin, and lactocin LP 27 (24,25), and streptococcin B74628 tetracycline. Shtibel (19) noted inhibition of N. (22), appear to lack this adsorption specificity. gonorrhoeae by bacterial interference and sug- Bac R1 adsorbed not only to neisserial and strepgested that the inhibition might be due to bac- tococcal bacterial strains and their stable Lteriocins. forms, but also to rat liver cells and to E. coli, a Bac RI, previously described by Rogolsky and species whose growth is not inhibited by Bac R1 Wiley (16), possesses all the usual properties of (16). Loss of viability of the L-form was compaa bacteriocin. The inhibitory activity is associrable to that of the neisserial parent after treatated with protein, and Bac R1 is bactericidal to ment by Bac R1. the same and related species without being auOther investigators (1, 7, 9, 11, 15) have retoinhibitory. Although Bac R1 has a wide spec- ported susceptibility of protoplasts or L-forms trum of activity (16), the only gram-negative to bacteriocins. These studies and our own oborganisms found to be susceptible to Bac R1 servations suggest that the receptor sites of the were the Neisseria. This indicated that Neisbacteriocins are located on the inner cell mem"A
ACTION OF Bac R1 AGAINST N. GONORRHOEAE
VOL. 14, 1978
brane. Naturally resistant bacteria may have the access to their membrane receptor regions physically blocked by outer components of the cell envelope (3, 20, 23). Neisserial cell walls are thought to differ from other gram-negative cell walls in being more porous due to either a lack of lipoprotein or a typical outer membrane. This may explain the lethal effect of Bac R1 on Neisseria. This research was supported by grant 2-76-072 from the American Nurses Foundation.
LrITERATURE CITED 1. Anastasio, K. L., G. A. Soucheck, and H. Sugiyama. 1971. Boticinogeny and actions of the bacteriocin. J. Bacteriol. 107:143-149. 2. Berman, G. D. 1977. How dangerous is penicillin resistant gonorrhoeae? Hosp. Physician 3:20-30. 3. Bhattacharyya, P., L. Wnedt, E. Whitney, and S. Silver. 1970. Colicintolerant mutants of Escherichia coli: resistance of membranes to colicin El. Science 168:998-1000. 4. Dajani, A. S. 1973. Neutralization of phage type 71 staphylococcal bacteriocin by immune and non-immune sera. J. Infect. Dis. 128:494-499. 5. Dajani, A. S., E. D. Gray, and L W. Wannasmaker. 1970. Effect of bactericidal substance from Staphylococcus aureus on group A streptococci. I. Biochemical alterations. Infect. Immun. 1:485-490. 6. Dajani, A. S., and L W. Wannamaker. 1973. Kinetic studies on the interaction of bacteriophage type 71 staphylococcal bacteriocin with susceptible bacteria. J. Bacteriol. 114:738-742. 7. Ellison, J. S., and J. A. Kautter. 1970. Purification and some properties of two boticins. J. Bacteriol. 104:19-26. 8. Fredericq, P. 1946. Sur la sensibilite et l'activite antibiotique des Staphylocoques. C. R. Seances Soc. Biol. Paris 140:1167-1170. 9. Gagliano, V. J., and RI D. Hinsdill. 1970. Characterization of a Staphylococcus aureus bacteriocin. J. Bacteriol. 104:117-125. 10. Hebeler, B. H., and F. E. Young. 1976. Chemical composition and turnover of peptidoglycan in Neisseria gonorrhoeae. J. Bacteriol. 126:1180-1185. 11. Ivanovics, G., L Alfoldi, and E. Nagy. 1959. Mode of action of megacin. J. Gen. Microbiol. 21:51-4.
223
12. Jaffe, H. W., J. W. Biddle, C. Thornsberry, R. E. Johnson, R. E. Kaufnman, G. H. Reynolds, and P. J. Wiesner. 1976. National gonorrhoeae therapy monitoring study. In vitro antibiotic susceptibility and its correlation with treatment results. N. Engl. J. Med. 294:5-9. 13. Jetten, A. M., and G. D. Vogels. 1972. Mode of action of a Staphylococcus epidermidis bacteriocin. Antimicrob. Agents Chemother. 2:456-463. 14. Jetten, A. ML, and G. E. Vogels. 1972. Nature and properties of a Staphylococcus epidermnidis bacteriocin. J. Bacteriol. 112:243-250. 15. Jetten, A. M., and G. D. Vogels. 1974. Characteristics of the killing effect of a Staphylococcus epidermidis bacteriocin. Antonie van Leeuwenhoek J. Microbiol. Serol. 40:177-183. 16. Rogolsky, M., and B. B. Wiley. 1977. Production and properties of a staphylococcin genetically controlled by the staphylococcal plasmid for exfoliative toxin synthesis. Infect. Immun. 8:156-164. 17. Rogolsky, M., B. B. Wiley, and L A. Glasgow. 1976. Phage group H staphylococcal strains with chromosomal and extrachromosomal genes for exfoliative toxin production. Infect. Immun. 13:44-52. 18. Schailehn, G., and J. Kiamer. 1975. Studies on mode of action of a bacteriocin from Clostridium septicum. Can. J. Microbiol. 22:435-437. 19. Shtibel, RI 1976. Inhibition of growth of Neisseria gonorrhoeae by bacterial interference. Can. J. Microbiol. 22:1430-1436. 20. Smarha, J., and M. Havelkova. 1970. The effect of colicin G on spheroplasts of Proteus mirabilis. Fola Microbiol. (Prague) 15:122-124. 21. Sparling, P. F. 1972. Antibiotic resistance in Neisseria gonorrhoeae. Med. Clin. North Am. 56:1133-1144. 22. Tagg, J. R., A. S. Dajani, and L W. Wannamaker. 1975. Bacteriocin of a group B streptococcus: partial purification and characterization. Antimicrob. Agents Chemother. 7:764-772. 23. Tagg, J. RI, A. S. Dajani, and L W. Wannamaker. 1976. Bacteriocins of gram-positive bacteria. Bacteriol. Rev. 40:722-756. 24. Upreti, G. C., and RI D. Hinsdill. 1973. Isolation and characterization of a bacteriocin from a homofermentative LactobaciUus. Antimicrob. Agents Chemother. 4:487-494. 25. Upreti, G. C., and R. D. Hinalill. 1975. Production and mode of action of lactocin 27: bacteriocin from a homofermentative Lactobacillus. Antimicrob. Agents Chemother. 7:139-145.
