ANTAGONISMS AMONG ENTERIC PATHOGENS AND COLIFORM BACTERIA

ANTAGONISMS AMONG ENTERIC PATHOGENS AND COLIFORM BACTERIA MAX LEVINE AND RALPH H. TANIMOTO Bureau of Laboratories, Territorial Department of Health, Honolulu, Hawaii Received for publication October 26, 1953 That some strains of the coliform group of strains but also against several salmonellae, bacteria exert inhibitory effects on other microorganisms, when grown in association in liquid Salmonella newport, two strains of Salmonella including a strain of Salmonella tennessee and media, was reported first by Nissle in 1916. enteritidis, and four of Salmonella schottmuelleri Because of the multiplicity of factors involved (paratyphoid B). Cook, Blachard, Robbins, and when organisms are growing in heterogeneous Parr (1953) observed that some "colicins" were mixtures in liquid media, the nature of the inhibiting agent is difficult to ascertain. types, SalmoneUa paratyphi, and a number of effective against several SalmoneUa typhi phage Gratia of Belgium (1925) observed a strain gram positive orgams. of Escherichia which, when grown on a solid In general the "colicins" have been reported medium, produced a diffusible substance that to exert a high degree of specificity, a particular was antibiotic against another coliform strain coliform culture being antagonistic to a few and ShigeUa dysenteriae. It was demonstrated other strains of the coliform and Shigella groups, later by Gratia and Fredericq (1946) that antibiotic coliform strains are not at all infrequent SalmoneUa. and occasionally against members of the genus and that the antibiotic substance produced, This brief report is concerned with observations on antibiosis by Salmonella, Shigella, and which they designated "colicin", was peculiarly distinct and characteristic for the particular Alkalescens-Dispar types, but especially two coliform strain producing it. They observed that coliform strains which (in contrast to previous some strains produced several "colicins" with reports of limited antibiotic activity by coliform characteristically different antibiotic spectra bacteria) were found to be antibiotic against and that antibiotic strains themselves might be almost all Salmonella and ShigelL types against susceptible to "colicins" produced by other coliform organisms. which they were tested. Heatley and Florey (1946) and Halbert and METHODS Magnuson (1948) presented data on the nature The technique employed to detect antibiotic of these "colicins" which were found to be acetone precipitable and thermostable. which is briefly as follows: strain was that of Gratia and Fredericq (1946) Halbert (1948) detected antibiotic coliform A small amount (about 1 mm loop) of a broth strains in the stools of 25 per cent of 147 individuals in Texas and 18.3 per cent of 1,243 cul- of an antibiotic is stabbed into previously culture of a strain to be examined for production tures isolated from this group; and in another poured and dried tryptone glucose extract agar. series of 287 individuals in an institution for (Generally, four to eight cultures were examined feeble minded in New York State, 30 per cent on a single agar plate.) The medium is incubated were found to be harboring antibiotic coliform then for 48 hours to provide an opportunity for strains, which comprised 11.7 per cent of 2,458 each inoculum to develop into a colony and for cultures examined, when testedagainst a strain of the products of growth to diffuse out into the ShigeUa flezr III (Z). Many of these coliform strains were also antibiotic against ShigeUa sonnei I, but Halbert did not observe evidence of antibiosis against salmonellae or a variety of other gram negative and gram positive bacteria. Fredericq and Levine (1947) reported a culture of Escherichia which was markedly antibiotic not only against coliform and Shigella agar. At the end of this time the organisms are killed by exposure, for about two hours, to chloroform which is then permitted to evaporate into the air. Then the entire surface of the agar is inoculated with a broth culture of an organism being examined for susceptibility. This is performed by placing a piece of sterile filter paper on the surface of the agar, saturating it with the

538 MAX LEVINE AND RALPH H. TANIMOTO [VOL. 67 TABLE 1 Incidence of antibiotic enteric bacteria GENUS OR GROUP NER ANBIOTIC EXAM- INED No. Per cent Salmonella types 52 2* 3.9 Shigella types 18 3t 16.7 Alkalescens-Dispar group 5 4$ 80.0 types Escherichia strains Fecal and water 174 33 19.0 Infants' throats 169 117 69.2 * Salmonella carrau; Salmonella schleissheim. t Shigella flexneri VI (64); Shigella boydii I and III. t Alkalescens I; Dispar I, IIa, IIb. broth culture being examined for susceptibility, then removing the filter paper and placing the reinoculated petri dish in the incubator overnight. If an antibiotic had been produced, growth of a sensitive strain was inhibited in a circular zone around the colonies of the chloroform killed coliform strains. Inhibition zones of 10 to over 50 mm in diameter were observed with various antibiotic coliform strains against many Shigella and Salmonella types. RESULTS A. Incidence of antibiotic strains. In table 1 are shown the results obtained with 52 Salmonella, 18 Shigella, 5 Alkalescens-Dispar types, and 343 strains of the coliform group. Of the latter, 174 were obtained from stools and water sources and 169 from throats of infants (2 to 5 days old). A strain was considered to be antibiotic if it produced a growth inhibition zone at least 10 mm in diameter against one or more enteric bacteria. It is evident that antibiotic strains were rarely encountered among the Salmonella types (3.9 per cent), more frequently among the Shigella types (15.7 per cent) and coliform strains obtained from water and stools (19.0 per cent), and especially frequently among the few types (only 5 types examined) of the Alkalescens-Dispar group (80 per cent) and the coliform strains obtained from the throats of infants of which 69.2 per cent were found to exhibit antibiosis. With reference to the latter, it is interesting to note that coliform strains were found in the throats of 40 (17.2 per cent) of 232 infants examined and that 27 (11.6 per cent) of these infants were harboring Escherichia in their throats. It is particularly worthy to note that of 117 coliform strains (obtained from 15 of these infants) which were examined for antibiosis, 99 or 84.6 per cent were antagonistic to growth of the pathogenic Escherichia serotype 0-55 B5, and 64 (54.7 per cent) were also effective against Escherichia serotype 0-111 B4. Furthermore, 30 of these strains (2 obtained from each of the 15 children) were antibiotic against Shigella sonnei I; 13 (obtained from 8 infants) were antibiotic against SalmoneUa grumpenesi; and 2 (obtained from one child) inhibited the growth of Salmonella 8chottmuellri in addition to Shigella eonnei I, and the pathogenic coliform types previously referred to. B. Frequency of antibioeis among strains of enteric bacteria. Each of 84 strains of enteric bacteria-52 Salmonella types, 18 Shigella types, 5 types of Alkalescens-Dispar group, and 9 strains of Escherichia-was examined for antibiosis against and susceptibility to each other. The results are shown in table 2. Of 7,056 observations (or trials) constituting this series, evidence of growth inhibition was noted in 253 or 3.7 per cent of all trials. Antibiosis was observed in only 6 (0.1 per cent) of 4,368 trials with Salmonella; in 14 (0.9 per cent) of 1,512 trials with Shigella; in 57 (13 per cent) of 420 trials with the Alkalescens-Dispar group; and in 186 (or 24.6 per cent) of 756 trials with Escherichia strains. It will be noted that none of the 52 Salmonella or 18 Shigella types was antibiotic against any of the Salmonella or Alkalescens-Dispar group types. Salmonella types showed antibiosis in 0.9 per cent of the trials against Escherichia and in only 0.2 per cent with Shigella. The Shigella types were only occasionally antibiotic against Escherichia strains (3.1 per cent of trials) and Shigella types (2.8 per cent of trials). The organisms of the Alkalescens-Dispar group employed were only occasionally antibiotic against Salmonella (2.3 per cent of trials) or the Alkalescens-Dispar group types (4.0 per cent of trials), but frequently against Eacherichia strains (33.3 per cent of trials) and various Shigella types (38.8 per cent of trials). It may be seen from table 2 that the Es-

19541 ENTERIC PATHOGENS AND COLIFORM BACTERIA TABLE 2 Frequency of antibiosis among enteric bacteria 539 SllIAINS OBSEBVZD FORt GRtOWTH INHIBITION STRAINS TESTED FOR EVIDENCE OF ANTBIOSIS Salmonella Shigella Alkalescens- Escherichia All strains Dispar 52* 18* 5* 9t 84 Per cent of trials showing growth inhibition Salmonella 52* 0.0 0.2 0.0 0.9 0.1 (4,368)1 Shigella 18* 0.0 2.8 0.0 3.1 0.9 (1,512) Alkalescens-Dispar 5* 2.3 38.8 4.0 33.3 13.0 (420) Escherichia 9t 21.4 34.0 17.8 28.4 24.6 (756) All strains 84 2.4 6.1 2.1 6.2 3.7 (7,056) (4,368) (1,512) (420) (765) t Number of strains. $ Figures in ( ) indicate number of trials-product of number of strains employed for evidence of production of an antibiotic and number of strains tested for susceptibility. cherichia strains employed showed by far the widest antibiotic spectra against other gram negative bacteria and, in comparison with the other enteric organisms, were particularly effective against Salmonella types. Thus, evidence of growth inhibition was observed in 21.4 per cent of trials against Salmonella, 34 per cent against Shigella, 17.8 per cent with the Alkalescens-Dispar group, and 21.4 per cent when tested against various strains of the genus Escherichia. It is interesting to note that on the basis of antibiotic properties the Alkalescens- Dispar group appears to be allied more closely to the genus Escherichia than to the genus Shigella to which it was formerly allocated-a relationship which, to some extent, finds a counterpart in antigenic components which are common to some Escherichia and the Alkalescens- Dispar strains. C. Antibiotic spectra of individual enteric bacterial types or strains. Of the 52 Salmonella types studied, only two-salmonella carrau isolated from a stool submitted from Canton Island and a stock culture of Salmonella schleissheim obtained from the Communicable Disease Center-were antibiotic; and, of the 18 Shigella types examined, antibiosis was observed with three-shigella flexnei VI (64); Shigella boydii I, and Shigella boydii III. In addition to these, Alkalescens I, Dispar I, IIa, and IIb, and many strains of the genus Escherichia were found to exhibit antibiosis against various enteric bacteria. In table 3 are indicated the spectra of the two antibiotic Salmonella and three Shigella types, and the specific enteric bacteria which were susceptible are indicated in the footnotes. Thus, Salmonella carrau was antibiotic only against one Shigella strain (Shigella flexneri VI (64)) and one coliform strain; Salmonella schleissheim inhibited the growth of a strain of Shigella sonnei II and three coliform strains. TABLE 3 Antibiosis by specific Salmonella and Shigella types against enteric bacteria Alka- Sal- Shi- les- Esche- 0_111 0-55 ZXA0N=FORGROCMmo- glacensn- EXAINEFO GRWTHnella gladis- chia B4 BS INHIBITION par S2* 18* 5* 9t 1 1i ANTIIOTIC.UR1EIBIOTIC TYPE 1 Number of types (or strains) showing ~growt inhibition S. carrau 0 1t 0 1 0 0 S. schleissheim 0 1 0 3 0 0 S. fiexneri VI (64) 0 311 0 0 0 0 S. boydii I 0 31 0 2 0 0 S. boydii III 0 31 0 3 0 0 t Number of strains. $ Shigella Jlexneri VI (64). Shigella sonnei II. 11 Shigella dysenteriae IV (Q1167) and VI (Q454); Shigella boydii IV. 1 Shigella flexneri VI (1389) and "Y" variant; Shigella dysenteriae VI (Q454).

540 MAX LEVINE AND RALPH H. TANIMOTO [vol. 67 TABLE Antibiosis by 8pecific Alkalescens-Dispar and Escherichia types (or strains) against enteric bacteria AULk- Sal- Sh-lea- Esche- 011, 0-55 EXAWMIN OX GROWTH nea gella Di cj B4 B5 INHIBTION par 52* 18 55 9 10 1* ANTIOTIC TY1 Number of types (or strains) (ol STHAN) showing growth inhibition Alkalescens I 4t 15 1 6 1 1 Dispar I 0 10 0 5 0 1 Dispar IIa 2t 511 0 3 0 0 Dispar lib 0 31 0 1 0 0 Escherichia H-10** 48 18 5 8 1 1 Escherichia H-23** 50 17 3 8 1 1 Escherichia 0-111 1 0 0 3 0 1 B4 Escherichia0-55 B5 0 0 0 0 0 0 t S. gallinarum; S. newington; S. typhi Bs and C. t S. muenster; S. worthington. I S. flezneri VI (64); S. boydii III and VI. 1j S. flexneri VI (64); S. boydii III and VI; S. sonnei I and II. S. glostrup. ** These strains (H-10 and H-23) were also antibiotic against stock cultures of 22 S. typhi phage types; 10 freshly isolated S. typhi cultures; and 3 Shigella types which were not tested for susceptibility to the Salmonella, Shigella and Alkalescens- Dispar types employed in this study. Shigeeneri VI (64), was antibiotic against three other Shigella types, namely: Shigella dywnteriwe IV (Q1167), ShigeUa dyentriae VI (Q454), and ShigeUa boydii IV. Shiea boydii I and III were each antibiotic against ShigeUa fleneri VI (1389), ShigeUa nezneri "Y" variant, and Shigella boidii VI (Q454). It will be noted that Alkalescens I was antibiotic against 4 Salmonella types-salmoneua gauinarum, SalmoneUa newington, and Salmonela typhi phage types Bs and C-and a large proportion (15 out of 18) of the ShigelLa types. Dispar IIa was effective against Salmonela muenster and Salmonella worthington and 5 Shigella types, whereas Dispar I and Dispar IIb were not antibiotic against any of the Salmonella, but the Dispar I was frequently and the Dispar IIb only occasionlly antibiotic against Shigella (10 and 3 Shigella types, respectively). The four strains of Escherichia selected include two (H-10 and H-23) isolated from stools in Hawaii because they were strikingly effective against a large variety of enteric organisms, the pathogenic Escherichia serotypes 0-111 B4 and 0-55 B5 because of their relationship to diarrhea in infants. In conjunction with the tables 3 and 4 it should be pointed out that Escherichia serotypes 0-111 B4 and 0-55 B5 were included among the 9 Escherichia strains but are indicated separately in the table because of their possible special interest. In table 4 are shown the results obtained with antibiotic types of the Alkalescens-Dispar group and four selected strains of Escherichia-cultures H-10 and H-23 isolated from stool specimens in Hawaii, and Escherichia serotypes 0-111 B4 and 0-55 B5 obtained from the Communicable Disease Center at Atlanta, Georgia. Escherichia 0-55 B5 was not antibiotic against any of a large number of coliform and other enteric types, whereas coliform serotype 0-111 B4 was observed occasionally to show growth inhibition of some Escherichia strains, including Escherichia serotypes 0-55 B5 and one strain of Salmonella (SalmoneUa glostrup). It should be noted particularly that Escherichia H-10 was effective against 48 and Escherichia H-23 against 50 of the 52 types of Salmonella observed in this study. Escherichia H-10 was antibiotic against all and Escherichia H-23 against 17 of the 18 Shigella types, as well as all of the coliform strains shown in the table, including Escherichia serotypes 0-111 B4 and 0-55 B5. SUMMARY Enteric bacteria with antibiotic properties were encountered most frequently among the Escherichia and Alkalescens-Dispar groups, occasionally among the Shigella, and only rarely among the Salmonella types. Thus, 33 (19 per cent) of 124 strains of Escherichia isolated from feces and water, 117 (69.2 per cent) of 169 Escherichia strains obtained from infants' throats, 4 out of 5 Alkalescens-Dispar types, 3 (16.7 per cent) of 18 Shigella types, and only 2 (3.9 per cent) of 52 Salmonella types were found to exhibit antibiotic properties against one or more of the gram negative enteric bacteria.

1954] ENTERIC PATHOGENS AND COLIFORM BACTERIA Of 52 Salmonella and 18 Shigella types studied, evidence of antibiosis was not observed against any of the Salmonella or Alkalescens-Dispar types, and they were only occasionally antagonistic to growth of ShigeLla types. Enteric bacteria with markedly antibiotic properties against many Salmonella and Shigella types were particularly prone to be found among strains of the genus Eecherichia isolated from feces, streams, and sea water and especially from infantsw throats. The pathogenic coliform serotype 0-55 B5 did not exhibit antibiosis, but serotype 0-111 B4 was antibiotic against coliform serotype 0-55 B5 and Salmonella glostp, five of the ShigeUafleneri, two ShigeUa boydii types, and SMgea sonnei I and II. Sver EWchriChia coli strains were found to be antibiotic aginst 80 of 84 Salmonella and 22 of 26 Shigella and Alkalesens-Dispar types as well as many coliform bacteria, including the pathogenic serotypes 0-55 B5 and 0-111 B4. REFERENCES Coox, M. K., BLANCARD, V. L., ROBBINS, M. L., AND PAR, L. W. 1953 An investigation of the antibacterial spectrum of colicines. Antibiotics & Chemotherapy, 3, 195-202. FEEDERXCQ, P., AND LEVINE, MAX 1947 Antibiotic interrelationships among the enteric group of bacteria. J. Bact., 54, 785-792. GRATIA, A. 1925 Sur un remarquable exemple d'antagonisms entre deux souches de colibacilli. Compt. rend. soc. biol., 93, 1040-1043. GRATIA, A., AND FREDERICQ, P. 1946 Diversitd des souches antibiotiques de B. coli et etendue variable de leur champ d'action. Comp. rend. soc. biol., 10, 1032-1033. HALBEET, S. P. 1948 The relation of antagonistic coliform organisms to Shigella infections. I. Survey observations. J. Immunol., 60, 23-3; II. Observations in acute infections. J. Immunol., 60, 359-381. HALBzRT, S. P., AND MAGNUSON, H. J. 1948 Studies with antibiotic-producing strains of Escherichia coli. J. Immunol., 58, 397415. HUATLEY, N. G., AND FLowy, H. W. 1948 An antibiotic from Bacterium coli. Brit. J. Exptl. Pathol., 27, 378-390. NIssiE, A. 1916 Ueber die grundlagen einer neuen ursaechlichen bekaempfung der pathologenicher darmfiora. U, 1181-1184. 541 Deut. med. Wochschr., Downloaded from http://jb.asm.org/ on April 22, 2019 by guest