cocci J. Bacteriol. 86:702-707. 1963.-A survey has been made of bacteriocine production by a wide variety of well-characterized strains of group D streptococci.
SURVEY OF THE BACTERIOCINES OF ENTEROCOCCI THOMAS D. BROCK,1 BARBARA PEACHER, AND DEBORAH PIERSON2 Department of Bacteriology, Indiana University, Bloomington, Indiana Received for publication 29 April 1963
ABSTRACT BROCK, THOMAS D. (Indiana University, Bloomington), BARBARA PEACHER, AND DEBORAH PIERSON. A survey of the bacteriocines of enterococci J. Bacteriol. 86:702-707. 1963.-A survey has been made of bacteriocine production by a wide variety of well-characterized strains of group D streptococci. On the basis of spectrums and sensitivity to chloroform, heat, and proteolytic enzymes, five distinct bacteriocines can be defined. Type 1 is produced by all Streptococcus zymogenes (S. faecali.8 var. zymogenes) strains, is active against a wide variety of gram-positive bacteria, and is also a hemolysin. Type 2 is produced by some S. liquefaciens (S. faecalis var. liquefaciens) strains, and acts on many enterococci as well as on certain other lactic acid bacteria. Type 3 is produced by certain strains of both S. faecalis and S. faecium, and inhibits a wide variety of group D streptococci, but is inactive against all other lactic acid bacteria tested except Leuconostoc citrovorum. Type 4 is produced by certain S. faecium strains and resembles in certain ways the type 3 activity, but differs from it in other ways. Type 5 has been found to be produced by only one proteolytic strain of S. zymogenes, and this bacteriocine has a very narrow spectrum. The strain that produces this bacteriocine also produces type 1 activity. No strain is sensitive to a bacteriocine of the type it produces. In surveys of bacterial strains for lysogeny, a by-product of the research is frequently the discovery of antibiotic substances (or bacteriocines) produced by certain strains which are active against other strains. These substances differ from classical antibiotics in (i) the narrowness of their spectrums, (ii) their action on strains IU.S. Public Health Service Research Career Development Awardee, grant AI-K3-18,403. 2National Science Foundation Undergraduate Research Participant.
closely related to the producer, and (iii) their chemical complexity. To date, antibiotics of this sort have been detected in a wide variety of bacteria (reviewed by Fredericq, 1957; Ivanovics, 1962; Hamon and Peron, 1963). Preliminary detection of bacteriocines in group D streptococci was made by Kjems (1955) and Pohonek (1961), and Brandis and Brandis (1962) made a detailed characterization of one of the bacteriocines detected by Kjems. Further, DeKlerk and Coetzee (1961) observed bacteriocines in lactobacilli that seem to be related to those of the enterococci. The present work represents a survey of bacteriocine production by a wide variety of wellcharacterized strains. It is shown that at least five different bacteriocines are produced by various strains of enterococci, and the association of bacteriocine type with other taxonomic characters is considered. It is also shown that certain of these bacteriocines are able to act on other lactic acid bacteria as well as on some bacteria more distantly related. In an accompanying paper, a detailed study of the bacteriocine produced by the hemolytic enterococci is presented, and the probable identity of this bacteriocine with the hemolysin will be demonstrated. An underlying question, unanswered by these studies, is what role bacteriocines play in the life
of bacteria. MATERALS AND METHODS Organisms. A wide variety of enterococci obtained from various culture collections were used. Not all of the strains examined will be mentioned explicitly; the ones used most frequently are listed in Table 1. The nomenclature of these strains follows Skadhauge (1950), Barnes (1956), and Shattock (1955), and the kety diagnostic characteristics associated with the species epithets are given in Table 2. The organisms were all grown in Todd-Hewitt broth or agar (Difco). Bacteriocine test procedure. Overnight broth
702
86, 1963
VOL.
BACTERIOCINES OF ENTEROCOCCI
703
cultures of the strains were inoculated by loop TABLE 2. Key diagnostic characters for speciation in the group D streptococci onto the surface of agar plates, five or six strains per plate, and the plates were incubated 18 to Tellu- Tetra- Proteolytic HemoSpecies name oeytc rite zohlum lytc 24 hr at 37 C. At this time, 2 ml of 0.75% agar, to which 0.1 ml of an indicator strain in the late S. faecium ........ S exponential phase of growth had been added, were S. durans ......... S + poured on the surface of the plate; the plate was S.faecalis ........ R + again incubated at 37 C for 6 to 8 hr, at which S. liquefacienst... R + + + ortime the indicator had fully grown and the zones S. zymogenes 4 R + + of inhibition could be seen. The distance from * For the procedures used in these tests, see the outer edge of the colony of the bacteriocine Skadhauge (1950) and Barnes (1956). Symbols: producer to the edge of the inhibition zone was S = sensitive; R = resistant; + = positive; ...
- = negative.
TABLE 1. Streptococcus strains used in this study Code
Species
Source
~
_
X3 X4 X12 X13 X14
X18 X21 X22
X33
S. faecium S. liquefaciens S. faecium S. faecium S. zymogenes S. faecium S. liquefaciens S. zymogenes S. liquefa-
ciens
T. D. Brock ATCC 13398
ND* ND
ATCC 9790 ATCC 8043t I. U. stock
ND ND 1
NCTC 4725t NCTC 8175
3
NCTC 8176
1 2
2
Isolated by E. Buecher 4 Isolated by E. Buecher ND ATCC 6056 ND ATCC 11700 NCIB 2702 3 HF8AG from W. R. 4 Chesbro From Paul H. Kop- 2 pent 3 ATCC 349t 5 Charles Rogers, strain 2 3 Charles Rogers, strain J 7 2 NIRD 588
X38
S. faecium
X45 X46 X47 X51
S. durans S. faecalis S. faecalis S. faecium
X52
X75
S. liquefaciens S. faecium S. zymogenes S. faecium
X82
S. liquefa-
X98
S. zymoge- P. M. F. Shattock nes H69D6 S. zymoge- M. D. Appleman 15A nes
X56 X74
Bacteriocine type
ciens X99 *
None detected.
t This strain was received as S. faecalis. t This strain is also proteolytic.
1 1
t S. faecalis var. liquefaciens. I S. faecalis var. zymogenes. measured in millimeters. It was preferable to measure the zones after only 6 to 8 hr to avoid the confusion resulting from the additional growth of colonies of the bacteriocine producer that arose from cells spread around the plate during the pouring of the soft agar layer. However, in many cases the zones of inhibition could still be read after overnight incubation. The use of chloroform was avoided (except where noted below), because one of the bacteriocines was inactivated by chloroform vapors. Rarely one strain would be lysogenic against another, and this would result in the formation of plaques, but these plaques were always restricted to the area where the producer was growing and never interfered with the observations of bacteriocine inhibition zones which extended some distance from the producer. Further, the plaques produced by the group D temperate phages were always small and indistinct, and only appeared after 18 hr of incubation. Inactivation of the bacteriocines. Sensitivity to chloroform was determined by exposing a plate containing the growth of six bacteriocine producers to chloroform vapors for 15 min at room temperature, removing the chloroform by 30 min of aeration at 37 C, and then pouring on the soft agar layer containing indicator. Inactivation by heat was determined by placing the plate in an oven at 80 C for 10 to 20 min, cooling, and adding indicator in soft agar. Inactivation by Streptococcus liquefaciens (S. faecaltis var. liquefaciens) protease was determined by streaking strain X52 on the outer chord of a plate and cross-streaking the bacteriocine pro-
704
BROCK, PEACHER, AND PIERSON
J . BACTERIOL .
ducers against it. After overnight incubation, (iv) Bacteriocines of S. zymogenes (S. faecalis soft agar containing indicator was poured on the var. zymogenes) and S. liquefaciens (S. faecalis plate, and the patterns of inhibition were ob- var. liquefaciens) types are resistant to protease served 6 hr later. If the X52 protease had action, whereas those of S. faecium and S. inactivated the bacteriocine, the zone around the faecalis types are sensitive. producer was absent in the area where the pro(v) All strains of S. zymogenes produce a ducer was adjacent to the X52 streak. An bacteriocine which is a hemolysin (see accomanalogous procedure was used in which crystalline panying paper) and which is chloroform-sensitrypsin was deposited on the outer chord and tive. This bacteriocine is active against all the producers were streaked up to this chord. enterococci except S. zymogenes. (vi) Strain X74, which is hemolytic and proRESULTS duces a type 1 bacteriocine, also produces an A total of 99 strains were tested for bacterio- additional bacteriocine which is active against cine production against 19 indicators. The indi- other S. zymogenes strains and which differs cators were selected as representative of all the from all the other bacteriocines. The two bacspecies listed in Table 2, and included both teriocines of X74 could be distinguished because bacteriocine producers and strains that appeared the type 1 bacteriocine is active against X13 but not to produce bacteriocines. The bacteriocines not against X14 and is chloroform-sensitive, were then characterized further by determining whereas the type 5 bacteriocine is active against their sensitivity to chloroform, heat, and proteo- X14 but not against X13 and is chloroformlytic enzymes. From these results, the bacterio- resistant. When the type 1 activity is destroyed cines could be divided into five types (Table 3). by chloroform, all activity against strain X13 A detailed study of Table 3 reveals a number disappears, but the activity against X14 is of interesting points. unaffected. The type 5 bacteriocine resembles (i) A strain is always resistant to its own that of the type 2 (liquefaciens) in being chlorobacteriocine and to that produced by others of form-resistant, protease-resistant, and heatits type. sensitive. It differs in that it acts on strains that (ii) A bacteriocine producer may be sensitive are resistant to the liquefaciens type, such as S. liquefaciens and S. zymogenes strains. Note to bacteriocines of other types. (iii) All bacteriocines of the same type show that strain X74 is proteolytic as well as hemolytic, the same behavior to heat, chloroform, and whereas the other hemolytic strains are not protease. proteolytic.
TABLE 3. Production of bacteriocines by enterococci*
Sensitivityt to
Bacteri-
ocine type tye CHC13
Heat
~~~Pro-
Producers of each type
Sensitive strains
Resistant strains
Range of
zone sizes
tease mm
1
S
R
R
X14,X22,X98,X99, X13,18,21,33,38,47, X14,22,74 X74 51, 52, 56, 75, 3, 4,
2
R
S
R
X21, X33, X52, X82
3
R
R
S
X18,X47,X56,X75
4
R
R
S
X38, X51
X13,18,38,47,51,56, X14,21,33,52, 75, 3, 12 74, 4, 46 X13,14,21,33,38,51, X18,47,56,75 52, 74, 46, 3, 4, 12 X14,21,33,52,56,74, X12,13,18,38,
X74
46, 3, 4 X14, 21, 33, 52, 46
2 to 8
46, 12
5
R
S
R
47, 51 X74, 4, 13, 38, 51, 3, 12, 18
3 to 5
7 to 10 4 to 5 4 to 5
* For a strain to be considered positive, the zone of inhibition had to be greater than 2 mm. There were only occasional equivocal strains, and these are not included above. t S = sensitive; R = resistant.
VOL. 86, 1963
705
BACTERIOCINES 4JIF ENTEROCOCCI
(vii) Although all S. liquefaciens strains that produce a bacteriocine produce one of the same type (viz., X21, 33, 52, 82), not all S. liquefaciens strains produce a bacteriocine (viz., X4). However, all S. liquefaciens strains (as well as S. zymogenes strains) are resistant to this type 2 bacteriocine, whether or not they produce it. This bacteriocine is active against all S. faecium strains tested and some, but not all, S. faecalis strains. (viii) Type 3 bacteriocine is produced by some strains of S. faecium and S. faecalis, but only a very few (4 of 40 strains tested), and this bacteriocine is active against all S. zymogenes and S. liquefaciens strains as well as against some S. faecium and S. faecalis strains. It inhibits more group D strains than any of the other bacteriocines, and gives the largest zones. In its properties, it seems to resemble the bacteriocine studied by Brandis and Brandis (1962). (ix) Type 4 bacteriocine is produced by only two S. faecium strains (of about 15 strains tested). This bacteriocine is active against some S. faecium and S. faecalis strains (viz., X3, X46) and not others (X13, X18, X47), but is active against all S. zymogenes and S. liquefaciens strains. Note that type 4 producers and type 3 producers do not always exhibit reciprocal cross resistance, even though certain strains in each group are cross-resistant with those of the other group. Thus, type 3 strains X18 and X47 are resistant to type 4, but X56 (also type 3) is sensitive. The bacteriocines studied by Kjems (1955) appear to resemble those listed here as types 3 and 4. Spectrum against other bacteria. A variety of gram-positive bacteria were tested by the agarlayer method for sensitivity to the five bacteriocines. As can be seen in Table 4, S. zymogenes (type 1) bacteriocine was active against a wide variety of gram-positive bacteria, and only the gram-negative bacteria and Bacillus polymyxa were completely resistant. This relatively broad spectrum separates this bacteriocine from the others, and, since this bacteriocine is also a hemolysin (see accompanying paper), it may be suggested that most gram-positive bacteria may have some structural component similar to one found on red cells. The type 2 bacteriocine inhibits the other lactic acid bacteria tested and Corynebacterium hoagii, but no other bacteria. The type 3 and 4 bacteriocines have little effect on bacteria other
TABLE 4. Action of bacteriocines on other bacteria Bacteriocine type* Bacteria 1
Leuconostoc citrovorum Lactobacillus fermenti L. plantarum Streptococcus lactis S. mastitidis Bacillus subtilis B. cereuts B. cereus var. mycoides B. polymyxa B. megaterium Clostridium sporogenes Staphylococcus aureus S. albus (S. epi-
-___ 2
3
4
5
(X14)
(X21)
(X18)
(X38) (X74)
S
S
S
S
R
S
S
R
R
R
S S
S
WS
R R
R R
R R
S
WS
WS S S
R R R
R R R R
R R R R
R R R R
R S S
R R R
R R R
R R R
R R R
S
R
R
R
R
S
R
R
R
R
S
R
R
R
R
S S S
R R
WS
WS WS
Ws
R R
R
R R R
S
R
WS
R
R
R R
R R
R R
R R
R R
dermidis) Micrococcus roseus M. lysodeikticus
Sarcina lutea Corynebacterium hoagiiC. pseudodiphtheriticum Escherichia coli Proteus vulgaris
* The code of the species which produces each type is given in parentheses. Symbols: R = resistant; S = sensitive; WS = weakly sensitive.
than group D streptococci, except for Leuconostoc citrovorum. And, finally, the type 5 bacteriocine affects none of these bacteria, and its action seems to be restricted to group D streptococci only. Note, however, that the type 3 bacteriocine is active against most strains of group D streptococci. This is, then, the bacteriocine most likely to be of use in taxonomic studies with group D streptococci, and its action against L. citrovorum would encourage speculation on a close relationship between this organism and the group D streptococci. We thus see that the group D streptococcus bacteriocines present an extremely diverse picture when viewed from their action on gram-positive bacteria.
706
BROCK, PEACHER, AND PIERSON
DISCUSSION The overall impression the reader will gain from the above work is that the situation is confused. Unfortunately, there is no reason to think that future work will make things simpler, because it is likely that a detailed study of the bacteriocines in any one type will reveal that each type is really a mixture of subtypes. Thus, the results presented here may inhibit rather than encourage further work. When it is realized that of the almost 100 strains tested over 50% produced one bacteriocine or another, it is clear that these bacteria are endowed with a mysterious but widespread property. It is safe to assume that these bacteria do not produce these substances accidentally, but that a bacteriocine plays some role in the life of the organism. The only group of strains which exhibit any uniformity are the hemolytic S. zymogenes strains, but, since the bacteriocine and hemolvsin activities of these organisms are probably due to the same substance, this uniformity is understandable; we are categorizing strains as S. zymogenes by their hemolytic character, and this automatically (if the identity of the hemolysin and bacteriocine is accepted) makes them uniform in regard to the bacteriocine. Note, however, that S. durans strains, which are hemolytic, do not produce a bacteriocine; this observation further emphasizes the differences between S. zymogenes and S. durans. Hamon and Peron (1963) considered briefly the function of bacteriocines, and concluded that they probably serve as fertility factors. Although it is true that cell-to-cell contact is mediated in some enteric bacteria by bacteriocines, there is no prior reason why this should be so in all kinds of bacteria. If this idea is correct, however, it is clear that we have at hand a simple method of determining sexual compatibilities between strains. Our current efforts will therefore be to see whether we can obtain any evidence of genetic recombination between some of the strains listed in Table 3. Results to date with strain X14 have been negative. Another role for bacteriocines is in invasion of a host. Hamon and Peron (1963) rejected this idea, but Pohonek (1961) postulated that streptococci which produce bacteriocines that attack vaginal lactobacilli may be able, because of this property, to invade the vagina and replace the
J. BACTERIOL.
resident bacteria. A similar idea is discussed in the accompanying paper for the S. zymogenes strains. The enterococcus taxonomist, looking at Table 3, will lose no time in concluding that bacteriocinogeny will be of little value in delineating the grand design of the enterococcus group. Since bacteriophage typing is also of little value in this regard (Brock, unpublished data), and since serology (Sharpe and Shattock, 1952) has been similarly disappointing, we may be left with the notion that various strains of enterococci in reality have very little in common, and represent a group of organisms so diverse that no grand design exists. The production of antibiotics of the nisin type by strains of S. lactis may be an analogous phenomenon to that reported here. Like nisin, the enterococcus bacteriocines act on certain other gram-positive bacteria, and the types 3 and 4 described here may be structurally similar to nisin (see, for instance, Brandis and Brandis, 1962). However, the type 1, type 2, and type 5 bacteriocines, which are insensitive to proteolytic action, and especially type 1, which is chloroform-sensitive, differ markedly from the nisin type. Indeed, type 1 bacteriocine may represent a new type of antibacterial agent.
AcKNOWLEDGMENT This study was supported by grant AM04630-03 of the U.S. Public Health Service and grant DA-CML-18-064-61-G13 from the U.S. Army Chemical Corps Biological Laboratories. LITERATURE CITED BARNES, E. M. 1956. Tetrazolium reduction as a means of differentiating Streptococcus faecalis from Streptococcus faecium. J. Gen. Microbiol. 14:57-68. BRANDIS, H., AND U. BRANDIS. 1962. tUber einen bacteriocinartigen Stoff aus Enterokokken. Pathol. Microbiol. 25:632-640. DEKLERK, H. C., AND J. N. COJTZEE. 1961. Antibiosis among lactobacilli. Nature 192:340-341. FREDERICQ, P. 1957. Colicins. Ann. Rev. Microbiol. 11:7-22. HAMON, Y., AND Y. PERON. 1963. etude du pouvoir bacteriocinogene dans le genre Listeria. II. Individualit6 et classification des bac-
teriocines en cause. Ann. Inst. Pasteur 104:5565.
VOL. 86, 1963
BACTERIOCINES OF ENTEROCOCCI
IvANOVICS, G. 1962. Bacteriocins and bacteriocinlike substances. Bacteriol. Rev. 26:108-118. KJEnMS, E. 1955. Studies on streptococcal bacteriophages. I. Techniques of isolating phageproducing strains. Acta Pathol. Microbiol. Scand. 36:433-440. POHONEK, M. 1961. Streptococci antagonizing the vaginal Lactobacillus. J. Hlyg. Epidemiol. Microbiol. Immunol. (Prague) 3:267-270.
707
SHARPE, M. E., AND P. M. F. SHATTOCK. 1952. The serological typing of group D streptococci associated with outbreaks of neonatal diarrhoea. J. Gen. Microbiol. 6:150-165. SHATTOCK, P. M. F. 1955. The identification and classification of Streptococcus faecalis and some associated streptococci. Ann. Inst. Pasteur Lille 7:95-100. SKADHAUGE, K. 1950. Studies on enterococci. Einer Munksgaards Forlag, Copenhagen.
