ANTIBIOTIC. the proportion of streptomycin resistant cells steadily decreased until it reached that normally

Transcription

1 THE CO-KILLING OF PENICILLIN SENSITIVE AND PENICILLIN RESISTANT BACTERIA AT LOW CONCENTRATIONS OF THE ANTIBIOTIC ARTHUR K. SAZ AND HARRY EAGLE Section on Experimental Therapeutics, Laboratory of Infectious Diseases, National Microbiological Institute, National Institutes of Health,' Bethesda, Maryland Received for publication March 9, 1953 There are a number of reports that bacteria which have been made resistant to antibiotics may be rendered normally susceptible by contact with sensitive organisms, with extracts of such organisms, or with culture filtrates. Some of these reports are mutually contradictory in important respects, and the results have not been reproducible regularly by other workers. Voureka (1948) found that when penicillin resistant staphylococci or streptococci were grown in mixed culture with penicillin sensitive pneumococci, staphylococci, streptococci, or Corynebacterium diphtheriae and then reisolated, a large proportion of the colonies were sensitive to penicillin. Surprisingly, organisms naturally resistant to penicillin, such as Salmonella typhosa, also were capable of "sensitizing" the penicillin resistant staphylococci. Further, it was not necessary that the two bacterial strains actually grow together since exposure of resistant staphylococci to the sensitizing strains for as little as five minutes at incubator, room, or refrigerator temperatures often sufficed. Lysates produced by bacteriophage or chemicals, and autolysates of various organisms also induced penicillin sensitivity as did exposure to chloramphenicol and to antisera (Voureka, 1951), or exposure to sublethal amounts of various toxic agents including penicillin (Voureka, 1952). Although some of the early observations of Voureka were confirmied by Winner (1948) and George and Pandalai (1949), they could not be repeated by Bennison and Schwabacher (1948). Barber (1948), in commenting on Voureka's findings, pointed out that single colonies derived from a presumably resistant culture may be much less resistant than the parent culture and emphasized the necessity for careful controls on the 1 U. S. Public Health Service, Dept. of Health, Education & Welfare. spontaneous reversion of resistant cultures toward decreased resistance. Welsch (1949) found that when a mixture of streptomycin sensitive and streptomycin resistant staphylococci were grown together in broth, the proportion of streptomycin resistant cells steadily decreased until it reached that normally present in a sensitive culture. Linz and co-worker (1951, 1951a, 1951b) made similar observations with Escherichia coli, staphylococci, and Pseudomonas aeruginosa. Contrary to the findings of Voureka with penicillin and staphylococci, the phenomenon was strain-specific, i.e., there was no decrease in the proportion of streptomycin resistant cells when they were mixed with sensitive cells of another strain; and culture filtrates of streptomycin sensitive cells were ineffective. Linz (1951) considered and discarded the possibility that the apparent reversal from resistance to sensitivity in the Welsch experiment reflected a more rapid multiplication of the sensitive cells. The opposite change, in the direction of increased resistance, also has been reported. Extracts and lysates of streptomycin resistant Enterobaderiaceae (Roland and Stuart, 1951), salmonellae (Zinder and Lederberg, 1952), and Hemophilus influenzae (Alexander and Leidy, 1953) have induced increased streptomycin resistance in the homologous sensitive organism; and Hotchkiss (1951) similarly has increased the penicillin resistance of pneumococci by growing the sensitive organisms in the presence of desoxyribosenucleic acid isolated from penicillin resistant pneumococci. In the course of studies in this laboratory on the development of resistance to antibiotics it was noted that when large numbers of sensitive organisms were plated in agar in the presence of threshold concentrations of penicillin fewer colonies grew out than when smaller inocula were 347

2 348 ARTHUR K. SAZ AND HARRY EAGLE [VOL. 66 used. It was found subsequently that when a few highly resistant organisms were added to large numbers of penicillin sensitive organisms of the same species, concentrations of antibiotic which would normally have permitted growth of the resistant organisms now effected sterilization of the entire bacterial population, sensitive and resistant cells alike. As will be discussed in the present paper, this phenomenon differs in important respects from those outlined above, and its explanation is not yet clear. MATERIALS AND METHODS The bacterial strains used were grown for 6 hours at 37 C in nutrient broth containing 0.2 per cent glucose; 2 per cent defibrinated horse blood was added for the cultivation of Diplococcus pneumoniae, type III, and Streptococcus pyogenes (C-203 strain). The cultures were diluted in nutrient broth to contain approximately 109, 108, 107, and 10, organisms per milliliter, as determined by dark field microscopic count; and one ml of these dilutions was plated into 14 ml of nutrient agar containing 0.2 per cent glucose and penicillin at the appropriate concentration. The agar medium for D. pneumoniae and S. pyogenes was enriched with 5 per cent defibrinated horse blood. After incubation at 37 C for 4 to 6 days, the colonies were counted with the aid of a Quebec colony counter. In critical plates showing no apparent growth, the agar was examined microscopically for micro-colonies; in addition, areas 1 cm square were macerated in a Potter-Elvehjem grinder in the appropriate fluid medium and plated out to determine the viable bacterial count in the clear areas. EXPERIMENTAL RESULTS The paradoxical decrease in the number of penicillin resistant colonies which grotw out from inocula of increasing size. The number of colonies of Micrococcus pyogenes var. aureus which grew out in a typical experiment when varying numbers of organisms were plated in jg of penicillin per ml is shown in the first row of table 1. In this experiment 39 colonies appeared on the plate inoculated with 106 organi. Plates inoculated with 107, 108, and 109 bacteria therefore would have been expected to contain 390, 3,900, and 39,000 colonies. Instead, 236 and 374 colonies grew out on plates inoculated with 107 and 108 organisms, and there were no visible colonies in the plate inoculated with 10' organisms. No micro-colonies were seen on the latter on microscopic examination, and three agar sectors each 1 cm square gave no growth when emulsified and subcultured in pour plates. Illustrative experiments with S. pyogenes and D. pneumoniae, with qualitatively similar results, are shown in table 1. The effect with Streptococcus faecalis (table 1) was much less pronounced than with either the micrococcus, the,-hemolytic streptococcus, or the pneumococcus, and at times the number of survivors was roughly proportional to the number of cells plated; while with E. coli (K-12 strain) the number of survivors has been related regularly to the number of organisms inoculated, and the "sensitization" phenomenon has never been observed. In all the experiments the sensitization phenomenon was demonstrable only in a relatively narrow zone of penicillin concentrations. When the concentration of penicillin was too low, large numbers of organisms survived and the sensitization effect was not demonstrable. On the other hand, if the concentration of penicillin was too high, then no resistant colonies grew out in any of the plates. Particularly with Micrococcus pyogenes var. aureus, but with the other species as well, occasional experiments failed to show the sensitization phenomenon just described. The reason for this variability in results is not yet clear. The "senwization" of highly resistant cells by nomal organisms of the same strain. In the foregoing experiments, the sensitization to penicillin was evidenced by a paradoxical decrease in the number of resistant colonies which grew out of a normal culture inoculum at threshold concentrations of penicillin as the size of that inoculum was increased. It was found subsequently that cells made highly resistant to penicillin by serial passage in media containing increasing concentrations of penicillin were similarly "sensitized" by large numbers of normal bacteria. An illustrative experiment is shown in table 2. Normal staphylococci in varying numbers (106, 107, 108, and 109) were plated into 15 ml of nutrient agar containing penicillin at 0.033,ug per ml. After solidification, this was overlaid with 5 ml of agar inoculated with 200 to 2,000 staphylococci resistant to 80 ug of penicillin per ml, or 1,400 times

3 1953] CO-KILLING OF PENICILLIN BACTERIA the concentration actually used in the experi- not S. faecalis or E. coli; normal pneumococci ment. (These resistant cells had been derived sensitized only pneumococci; and normal strepfrom the same strain by serial passage in broth tococci, only streptococci. TABLE 1 The paradoxical inverse relationship between the size of the inoculum and the number of colonies growing out at threshold concentrations of penicillin 349 PENICILLI NUMER OP VISIBLE COLONIES APPEARING ON ORGANISM CONCZNTRA- PLS NOCTDW TIONS IN AGAR 10G 10' 10' 10. CONCLUSION pg/ml Micrococcus pyogenes var No colonies develop aureus with sufficiently large inocula, and Diplococcus pneumoniae, , the inoculum is Type III , dead Streptococcus pyogenes, strain C Streptococcus faecalis Slight sensitizing ef fect of large in ocula Escherichia coli, strain K-12' ,820 overgrown No sensitizing effect overgrown by large inocula or agar containing increasing amounts of penicillin.) The agar overlay also contained ug of penicillin per ml. The colonies growing out in the agar overlay were counted after 4 to 6 days' incubation at 37 C. As shown in table 2, when no sensitive cells were in the underlay, there was esentially 100 per cent survival of the resistant cells in the overlay. However, as the number of sensitive cells was increased from 107 to 10, to 10', the number of resistant cells which grew out to form colonies fell progressively from the original 1,900 to 534, 80, and 0, respectively. The total population of highly resistant organisms had been rendered nonviable in the presence of large numbers of sensitive cells dying under the impact of penicillin. In order to determine the specificity of this sensitization, experiments were carried out with various bacterial species in the underlay and overlay, using the empirically determined optimal concentrations of penicillin in each case. As shown in table 3, in every instance the sensitization was strictly species-specific: normal staphylococci sensitized resistant staphylococci to penicillin, but TABLE 2 The "sensitization" of penicillin resistant staphylococci by sensitive staphylococci in the presence of penicillin NUMBER OP RESISTANf ORGMS WHICH GREW OUT IN PENICILIN-AGAR PLATES CONTAINNG SI OP RESIS- INDICATED NUMBER OP SENSITIVE TANT INOCULUMO STAPHYLOCOCCI* ' 2 X X 10' 1,900k - 1, * Normal organisms were plated in 15 ml agar containing penicillin at micrograms per ml; after solidification, this was overlaid with 5 ml penicillin-agar containing the indicated number of penicillin resistant staphylococci derived from the same strain. A major difficulty in the study of the phenomenon has been the nonreproducibility of the experiments. Even with normal and resistant strains known to produce the phenomenon under consideration, the results have often been irregular. The reason for this variability is not yet clear.

4 350 ARTHUR K. SAZ AND HARRY EAGLE [vol. 66 Further, with M. pyogenes var. aureus used in the present experiments, not every resistant subculture proved susceptible to the sensitizing action of normal bacteria. Thus, in one experiment, three strains were tested simultaneously, all deriving from the same parent Smith strain: two were resistant to 3.5 and 1,ug per ml, and one was resistant to 0.7 ug per ml. When 103 resistant organisms were mixed with 109 normal organisms in a 15 ml agar plate containing jig per ml, the first two strains grew out quantitatively, while the latter gave only 29 colonies, instead of account for the results obtained (e.g., the killing at jug penicillin per ml of bacteria originally resistant to 80 jg per ml). Conceivably also, the introduction of such large numbers of bacteria could have led to the rapid exhaustion of an essential metabolite or to the production of toxic catabolic products, thereby limiting growth. The following experiments indicated that this was not the case. Plates contauiing threshold concentrations of penicillin and an inoculum of 10' normal staphylococci, on which no colonies had appeared after TABLE 3 The species 8pecificty of co-killing phenomenon Approximately 109 penicillin sensitive organisms were overlaid or mixed with small numbers of resistant organisms in 15 ml agar containing threshold concentrations of penicillin. O - resistant organisms fail to grow when mixed in agar plate with large numbers of sensitive bacteria and threshold concentrations of penicillin. + - no "sensitization": resistant organisms grew out. - - not done, because the minimum concentration of penicillin necessary to kill the "sensitizing" organism was also bactericidal for the heterologous species (vertical column). SdSS=NSmZINC; S11= a T FTC N= S WIPIAX'=8 PYOWCM 8%ZPC=AS ZSCMRU tyi A.A. - Concentration of penicillin, e/ml Relatively resistant* species tested for "sensitization" Streptococcus pyogenes Diplococcus pneumoniae, type III Micrococcus pyogenes var. aureus Streptococcus faecalis Escherichia coli * Rendered resistant where necessary by serial passage through penicillin. 1,210, and the clear areas on the plate were sterile. DISCUSSION There are a number of possible explanations for the observed "sensitization" of penicillin resistant organisms by large numbers of sensitive organisms dying under the impact of the drug. The addition of 1 to 1.5 ml (109 cells) of a fully grown glucose broth culture to 14 ml of agar might conceivably have decreased the ph in the agar, thus increasing the bactericidal activity of the penicillin (Eagle et al., 1952). However, the phenomenon was observed when the normal organisms were centrifuged and resuspended in agar containing no added glucose. The minor change in ph observed under these conditions could not 4 to 6 days' incubation, were flooded with penicillinas to inactivate residual penicillin. The surface of the treated plates was treated then with serial dilutions of the homologous organism. As a control, a previously uninoculated agar plate was treated similarly. Growth in the two series was quantitatively the same, measured either by the number of colonies or by their size. Similarly, if penicillin-containing-agar plates, in which a small number of resistant pneumococci or staphylococci had failed to grow out in the presence of large numbers of normal bacteria, were autoclaved and then reinoculated with the same organisms, either normal or resistant, the number of emergent colonies and the rate of growth of these colonies were the same as in normal media. One may conclude reasonably

5 1953] CO-KILLING OF PENICILLIN BACTERIA 351 that the death of penicillin resistant cells in penicillin-agar containing large numbers of susceptible bacteria was not due to the exhaustion of the medium. Either of two general mechanisms could be responsible for the phenomenon described: (a) a release from penicillin killed sensitive bacteria of a species-specific toxic factor which kills the resistant cells directly, without the further intervention of penicillin, or (b) the release of a species-specific transforming agent which converts penicillin resistant cells to penicillin sensitive cells which then are killed by the antibiotic present in the medium. In keeping with the first possibility, the organisms might produce a specific lytic substance similar to that reported for E. coli by Zamenhof (1945). However, culture filtrates and cell-free extracts2 of penicillin sensitive staphylococci did not kill resistant cells; and similar filtrates and extracts of penicillin treated sensitive staphylococci were also without effect. Such negative experiments, however, do not rule out the possibility that under the impact of penicillin the organisms may produce or release a labile toxic factor responsible for the results here reported. Similarly, the possibility that a bacteriophage is released from lysogenic penicillin sensitive cells, which then attacks the resistant cells, merits consideration despite the fact that extracts of the experimental mixtures have been uniformly ineffective. It is known that penicillin may cause the lysis of sensitive cells. Conceivably, desoxyribonucleic acid present in such lysates could have induced a transformation from penicillin resistance to penicillin sensitivity. Transformation or transduction in the reverse direction, from penicillin or streptomycin sensitivity to resistance, by extracts of resistant cells, has been reported by several workers (Roland and Stuart, 1951; Hotchkiss, 1951; Zinder and Lederberg, 1952; Alexander and Leidy, 1953). It should be noted that the proportion of cells so transformed is extremely small and generally less than 1 per cent; while in the present experiments more than 99 per cent of the resistant organisms were often rendered nonviable. Further, the addition of desoxyribonuclease to the 2Produced either by rapid vibration with micro glass beads ("ballotini") in the Mickle tissue disintegrator, or by vibration in the 9 kc Raytheon sonic oscillator. medium had no demonstrable effect on the cokilling phenomenon described. Attempts to demonstrate a putative toxic or transforming principle in the agar medium under the conditions of the present experiments to date have been uniformly unsuccessful. It is of interest in this connection that the co-killing phenomenon described to date has not been observed in fluid media, and the results obtained with pneumococci and streptococci in agar medium were more consistent if the reaction mixture was incubated under anaerobic conditions. SUMMARY When varying numbers of Micrococcus pyogenes var. aureus, Streptococcus pyogene8, or Diplococcu8 pneumwniae, type III, were plated into agar containing threshold concentrations of penicillin, the number of emergent colonies paradoxically decreased as the size of the inoculum was incrased. The weakly resistant organisms which would normally grow out became nonviable in the presence of large numbers of bacteria dying under the impact of the antibiotic. This "sensitization" phenomenon occurred only irregularly with Streptococcu faecalis and was not observed with Escherichia coli, strain K-12. Highly penicillin resistant organisms (M. pyogenes, Streptococcus pyogenes, or Diplococcus pneumoniae), derived from an initially sensitive strain by serial transfer through penicillin, were killed similarly when plated together with large numbers of the parent sensitive strain and low concentrations of penicillin. This phenomenon was strictly species specific. Possible mechanisms have been discussed in the text. REFERENCES ALEXANDER, H. E., AND LEIDY, G Induction of streptomycin resistance in sensitive Hemophilus influenzae by extracts containing desoxyribonucleic acid from resistant Hemophilus in.fuenzae. J. Exptl. Med., 97, BARBER, M Sensitisation of penicillinresistant staphylococci. Lancet, May 8. BENNISON, W. H., AND SCHWABACHER, H Sensitisation of penicillin-resistant bacteria. Lancet, I, 885. EAGLE, H., LEVY, M., AND FLEISCHMAN, R Effect of the ph of the medium on the antibacterial action of penicillin, streptomycin, chloramphenicol, terramycin, and bacitracin. Antibiotics & Chemotherapy, 2,

6 352 ARTHUR K. SAZ AND HARRY EAGLE [VOL. 66 GEORGE, M., AND PANDALAI, K. M Sensitisation of penicillin-resistant pathogens. Lancet, 1, HOTCHKI5S, R. D Transfer of penicillin resistance in pneumococci by the desoxyribonucleate derived from resistant cultures. Cold Spring Harbor Symposia Quant. Biol., 16, LINZ, R Sur le mdcanisme du ph6nomene de Welsch. Compt. rend. soc. biol., 145, 146. LINz, R., AND LECOCQ, E. 1951a Evolution des melanges de Bact6ries sensibles et r6sistantes a la streptomycine (ph6nom6ne de Welsch). Compt. rend. soc. biol., 145, 143. LINZ, R., AND LECOCQ, E. 1951b Evolution des m6langes artificiels de cellules de Mycobacterium tuberculosis sensibles et r4sistantes a la streptomycine. Compt. rend. soc. biol., 145, 149. ROLAND, F., AND STUART, F "Directed mutation" toward streptomycin resistance in Salmonella typhi. Antibiotics Chemotherapy, 1, VOuREKA, A Sensitisation of penicillinresistant bacteria. Lancet I, 62. VOuREKA, A Production of bacterial variants in vitro with chloramphenical and specific antiserum. Lancet, I, 1, VOURZKA, A Induced variations in a penicillin-resistant staphylococcus. J. Gen. Microbiol., 6, WELSCH, M Quelques aspects de la r6sistance du staphylocoque & la streptomycine. Compt. rend. soc. biol., 158, WINNER, H. I Quantitative sensitisation of a penicillin-resistant staphylococcus. Lancet, I, 674. ZAMENHOF, S A specific lytic substance in Escherichia coli. J. Bact., 49, 413. ZINDER, N. D., AND LEDERBERG, J Genetic exchange in Salmonella. J. Bact., 64, Downloaded from on May 11, 2016 by PENN STATE UNIV