THE JOURNAL OF INFECTIOUS DISEASES. VOL. 143, NO.3.

MARCH 1981
© 1981 by The University of Chicago. 0022-1899/81/4303-0019$00.75

Bacterial Interference Between Clostridium difficile and Normal Fecal Flora

Rial D. Rolfe, Shushan Helebian, and From the Medical and Research Services, V.A. Wadsworth
Sydney M. Finegold Medical Center; and the Department of Medicine, UCLA
Center for the Health Sciences, Los Angeles, California

Clostridium difficile has been shown to be the cause of virtually all cases of
pseudomembranous colitis related to the administration of antimicrobial agents. It is
possible that some antimicrobial agents alter the normal bacterial flora of the
gastrointestinal tract so as to permit colonization and/or proliferation by C. difficile.
The inhibitory activity of representative fecal bacteria from 23 anaerobic and aerobic
genera against C. difficile was examined using two in vitro procedures. Strains of
bacteria in six of the genera inhibited the multiplication of C. difficile, with Lac-
tobacillus organisms and group D enterococci displaying the most antagonistic activity.

Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on August 22, 2015

C. difficile was examined for its ability to inhibit the multiplication of several fecal
strains of anaerobic and aerobic bacteria. All eight strains of C. difficile tested inhibited
the growth of particular strains of bacteria in the genera Bacteroides, Peptococcus, and
Peptostreptococcus.

Clostridium difficite is the etiologic agent of vir- Materials and Methods

tually all cases of antimicrobial agent-associated
Bacteria. The strains of bacteria examined in
pseudomembranous colitis and 1'\.1200/0 of cases of
this study, with the exception of C. difficile, were
nonspecific colitis or diarrhea without colitis
isolates obtained from stool specimens submitted
related to the administration of antimicrobial
by healthy individuals to the Microbial Disease
agents [1, 2]. The mechanisms by which antimicro-
Research Laboratory at the Wadsworth V.A.
bial agents induce intestinal disease associated
Medical Center, Los Angeles, Calif. The toxigenic
with C. difficile are unclear. These drugs may alter
strains of C. difficile were isolated from patients
the normal bacterial flora of the gastrointestinal
with antimicrobial agent-associated intestinal dis-
tract so as to permit colonization and!or pro-
ease (pseudomembranous colitis, nonspecific coli-
liferation of C. difficile, as well as toxin elabora-
tis, and diarrhea without colitis) as well as from
tion by the organism. C. difficile appears to
healthy individuals. The identifications of all bac-
flourish when the competing bacterial flora is sup-
teria was performed using standard procedures
pressed or before the normal bacterial flora has
[5-7] .
the opportunity to become established, as in neo-
Media. The inhibitory activity of all bacteria
nates [3, 4].
was examined using two types of growth media:
In this study, we examined 116 aerobic and 285
trypticase soy agar (TSA; BBL Microbiology
anaerobic isolates of fecal bacteria for their ability
Systems, Cockeysville, Md.) and brucella blood
to inhibit the multiplication of C. difficile in vitro.
agar supplemented with vitamin K1 and hemin
In addition, the inhibition of bacteria by strains of
(BBA; Difco Laboratories, Detroit, Mich.) [5].
C. difficile was studied.
Preliminary studies in our laboratory showed that
TSA supported the growth of most intestinal bac-
teria, was more sensitive than other media for
Received for publication June 19, 1980, and in revised form demonstrating inhibitory bacteria, and resulted in
November 13, 1980.
This study was supported in part by the V.A. Medical Center
the largest zones of inhibition. On the other hand,
Research Funds and by a grant from Upjohn Company. some fastidious microorganisms - particularly cer-
We thank Betty Harris and Sallie Young for technical tain obligately anaerobic bacteria - grew poorly
assistance and Anne Husz and Kimi Ishii for typing and on TSA, thereby necessitating the use of BBA for
manuscript preparation. inhibition testing.
Please address requests for reprints to Dr. Rial D. Rolfe,
V.A. Wadsworth Medical Center, Room 317, Building 114,
Detecting the inhibitory activity of bacteria
Wilshire and Sawtelle Boulevards, Los Angeles, California against C. difficile. Two different test proce-
90073. dures - simultaneous and deferred - were used to

470
C. difficile Bacterial Interference 471

detect bacterial inhibitory activity. In an extensive
review, Tagg et al. [8] indicated the importance of
using both test procedures when screening for
bacterial inhibitory activity.
The simultaneous (or direct) test procedure is a
modification of the one originally described by
Gratia [9]. In this procedure, the test (inhibitor)
and indicator (sensitive) strains of bacteria are
grown simultaneously on the same agar surface.
The demonstration of antagonism is dependent
upon the release of a diffusible inhibitory sub-
stance early in the growth of the test culture. Inhi-
bition is present when the indicator strain fails to
0.3(0) using reduced yeast-extract diluent and
then adding 0.1 ml of the diluted cell suspension to
4.5 ml of melted brucella soft agar (previously re-
duced inside the anaerobic chamber for 3 hr at 50
C). Zones of inhibition were measured at the end
of anaerobic incubation at 37 C for 48 hr.
Detecting the inhibitory activity of c. difficile
against fecal bacteria. The simultaneous and de-
ferred test procedures described above, with mi-
nor modifications, were used for examining the in-
hibitory activity of eight strains of C. difficile
against various isolates of fecal bacteria. The si-
multaneous test procedure was performed by inoc-

Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on August 22, 2015

multiply about the spot-inoculated test strain. In ulating 0.1 ml of an overnight culture of an indica-
the present study, bacteria examined for inhibi- tor strain in brain-heart infusion broth into 4.5 ml
tory activity were initially grown on BBA for 48 of melted brucella soft agar. This cell suspension
hr; individual colonies were then used to spot-in- was overlaid onto the surface of the growth medi-
oculate the surface of a TSA or BBA plate pre- um; after it had solidified, each C. difficile test
viously seeded with 4.0 x 105 cells of C. difficile
using a sterile, round wooden applicator stick (di-
ameter, 2.5 mm). We routinely spot-inoculated 12 Table 1. Aerobic bacteria not inhibiting the multipli-
different bacterial strains on the agar surface of a cation of Clostridium difficile.
petri dish (diameter, 100 mm). An overnight cul- Organism No. of strains
ture of C. difficile in brain-heart infusion broth,
A cinetobacter calcoaceticus 2
diluted in a reduced yeast-extract solution [5] to a
Aeromonas hydrophila 1
final concentration of 4 x 106 bacteria/ml (A at Bacillus species 5
600 nm, 0.027), was used to seed the plates by Candida albicans 3
spreading 0.1 ml of the diluted cell suspension on Citrobacter species 7
the surface of the growth medium with a sterile Enterobacter aerogenes 2
bent glass rod. All test procedures were performed Enterobacter cloacae 7
Escherichia coli 21
in an anaerobic glove box isolator (atmosphere: Group 0 enterococci 12*
N2 , 80070; H 2 , 10070; CO 2 , 10070). The inoculated Group 0 nonenterococci 1
plates were incubated anaerobically at 37 C and Klebsiella species 11
examined after 48 hr for zones of inhibition. Morganella morganii 3
Proteus mirabilis 2
In the deferred test procedure, each test organ-
Pseudomonas aeruginosa 2*
ism was spot-inoculated onto the surface of a Pseudomonas fluorescens 1
growth medium as described above and incubated Salmonella arizonae 1
anaerobically or aerobically for 48 hr at 37 C. The Sarcina lutea 1
plates were then removed from the anaerobic Serratia liquefaciens 2
chamber or conventional incubator and exposed Staphylococcus aureus 3*
Staphylococcus species t 1
to chloroform vapor for 30 min according to the Streptococcus equinus 2
procedure of Bauernfeind and Burrows [10]. After Streptococcus intermedius 1
being placed in a conventional incubator at 37 C Streptococcus mitis 1
for an additional 30 min to permit diffusion of Streptococcus salivarius 1
Streptococcus species 1
chloroform out of the agar, the plates were re-
turned to the anaerobic chamber, reduced for 3 NOTE. Bacterial antagonism was examined against four
hr, and gently overlaid with melted brucella soft strains of C. difficile by the deferred and simultaneous test
procedures after growth on trypticase soy agar and brucella
agar (brucella broth plus 0.75% agar) containing
blood agar (see Materials and Methods).
C. difficile. The overlay was prepared by diluting * Some strains of bacteria belonging to these species in-
an overnight culture of C. difficile in brain-heart hibited the multiplication of C. difficile (see table 3).
infusion broth to 4.5 x 107 cells/rnl (A at 600 nm, t Coagulase-negative.
472
Table 2. Anaerobic bacteria not inhibiting the multi-
plication of Clostridium difficile.
Organism
Bacteroides distasonis
Bacteroides fragilis
Bacteroides ovatus
Bacteroides thetaiotaomicron
Bacteroides vulgatus
Bacteroides species
Bifidobacterium adolescentis
B. adolescentis (biotype A)
B. adolescentis (biotype C)
Bifidobacterium breve
Bifidobacterium catenulatum
Bifidobacterium eriksonii
Bifidobacterium infantis
B. infantis (other)
Bifidobacterium species
Clostridium aminovalericum
Clostridium barkeri
Clostridium bifermentans
Clostridium butyricum
Clostridium innocuum
Clostridium mangenotii
Clostridium paraputrificum
Clostridium perfringens
Clostridium ramosum
Clostridium sorde//ii
Clostridium tertium
Clostridium species
Eubacterium aerofaciens
Eubacterium cy/indroides
Eubacterium lentum
Eubacterium moni/iforme
Eubacterium nitritogenes
Eubacterium rectale
Eubacterium species
Fusobacterium species
Gram-positive cocci
Lactobacillus acidophi/us
Lactobacillus brevis
Lactobacillus casei
Lactobacillus catenaforme
Lactobacillus delbrueckii
Lactobacillus fermentum
Lactobacillus lactis
Lactobacillus minutus
Lactobacillus plantarum
Lactobacillus rogosae
Lactobacillus species
Leptotrichia bucca/is
Peptostreptococcus micros
Peptostreptococcus parvulus
Peptostreptococcus productus
Peptostreptococcus species
Veil/onella parvula
Rolfe et al.
strain was spot-inoculated onto the overlay using a
sterile, round wooden applicator stick as described
No. of strains above. The plates were incubated anaerobically at
37 C and read for zones of inhibition after 48 hr.
3 The deferred test procedure was performed by
3
spot-inoculating each test strain of C. difficile
1
13
onto the surface of the growth medium and incu-
7 bating anaerobically or aerobically at 37 C for 48
38* hr. The plates were then removed from the anaer-
1* obic chamber or conventional incubator, exposed
4 to chloroform vapor, and aerated as described
1 above. After reduction inside the anaerobic cham-
1
ber for ""3 hr, the plates were overlaid with 4.5 ml
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on August 22, 2015
3
1 of melted brucella soft agar inoculated with 0.1 ml
2* of an overnight culture of an indicator strain in
2 brain-heart infusion broth. After anaerobic incu-
15 bation at 37 C for 48 hr, the plates were examined
3 for zones of inhibition.
1
3
1 Results
7
2 Inhibitory activity of fecal bacteria against C.
4 difflcile. The inhibitory activity of 116 strains of
12 aerobic bacteria (representing 14 genera) and 285
8 strains of anaerobic bacteria (representing nine
1 genera) was initially examined against four strains
1
of C. difficile, using both the simultaneous and
13
2 deferred test procedures on BBA and TSA. Bacte-
2 rial interference was defined in this part of the in-
4 vestigation as a clear zone of inhibition surround-
1 ing the spot-inoculated test strain of bacteria. All
1 test procedures were repeated a minimum of two
2
times; the results were very reproducible.
21
Bacteria that did not inhibit the multiplication
2
3 of any of the four strains of C. difficile are listed
8 in tables 1 and 2. Bacteria that did inhibit the mul-
2 tiplication of one or more of the four strains of C.
2 difflcile were then examined for inhibitory activity
4 against 16 additional strains of C. difficile (tables
1
3 and 4).
2
I
Examples of the inhibitory activity displayed by
4 group D enterococci and Pseudomonas aerugino-
4 sa against C. difficile are shown in figure 1. The
1 small clear zone of inhibition in the center of the
48*
2
NOTE. Bacterial antagonism was examined against four
I
strains of C. difficile by the deferred and simultaneous test pro-
1
cedures after growth on trypticase soy agar and brucella blood
2 agar (see Materials and Methods).
3 ... Other strains of bacteria belonging to these species in-
1 hibited the multiplication of C. difficile (see table 4).
C. difficile Bacterial Interference 473

Table 3. Aerobic bacteria inhibiting the multiplication of Clostridium difficile.

Bacterial interference test procedure

Deferred

Simultaneous Aerobic growth Anaerobic growth

Organism (no. of strains) TSA BBA TSA BBA TSA BBA

Staphylococcus aureus (1) 0 20 0 0 20 20

Pseudomonas aeruginosa (3) 0 0 20 0 NA NA
Group D enterococci (11) 20 20 20 20 20 20
Group D enterococci (2) 18 0 18 20 20 20
Group D enterococci (1) 0 0 2 0 0 0
Group D enterococci (1) 0 0 1 0 0 0
Group D enterococci (1) 0 0 0 0 12 0

Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on August 22, 2015

Group D enterococci (1) 0 0 0 0 17 0
Streptococcus mitis (1) 0 0 0 0 1 0
NOTE. Data are the number of strains of C. difficile that were inhibited. Twenty strains of C. difficile were tested in each pro-
cedure. For descriptions of the deferred (aerobic and anaerobic) and simultaneous test procedures, see Materials and Methods. TSA
= trypticase soy agar; BBA = brucella blood agar; NA = no anaerobic growth observed.

plate, not corresponding to a spot-inoculated bac-

terium, was the result of a contaminant. Group D
enterococci produced clear zones of inhibition
with and without a surrounding "halo" of partial
growth inhibition, whereas P. aeruginosa gave a
small zone of complete inhibition surrounded by a
very large zone of partial inhibition. All other in-
hibitory bacteria displayed only clear zones of in-
hibition.
There were no differences in inhibitory activity
among the test strains of C. difficile (tables 5 and
6). There were, however, differences among the
strains of bacteria inhibited by C. difficile (table 6).

Table 4. Anaerobic bacteria inhibiting the multiplica-

tion of Clostridium difficile.
Bacterial interference
test procedure
Simultaneous Deferred

Organism (no. of strains) TSA BBA TSA BBA

Bacteroides species (2) 2 0 0 0

Bifidobacterium adolescentis (3) 0 0 20 0
Bifidobacterium infantis (1) 0 0 20 0
Figure 1. Inhibition of Clostridium difficile by normal
Bifidobacterium longum (1) 0 0 20
bacterial intestinal flora as detected by the deferred an-
0
Lactobacillus species (2) 20 20 20 20
tagonism test procedure (see Materials and Methods)
with trypticase soy agar as the support medium. The
NOTE. Data are the number of strains of C. difficile that spot-inoculated bacterial strains were (A-E) group D
were inhibited. Twenty strains of C. difficile were tested in enterococci, (F) Escherichia coli, (G) Enterobacter aero-
each procedure. For descriptions of the deferred and simul- genes, (11) group D enterococci, (I) Streptococcus mitis,
taneous test procedures, see Materials and Methods. TSA = (1) Pseudomonas aeruginosa, and (K) Staphylococcus
trypticase soy agar; BBA = brucella blood agar. aureus.
474 Rolfe et al.

Table 5. Bacteria not inhibited by Clostridium difficile. Table 6. Bacteria inhibited by Clostridium difficile.
Bacteria No. of strains Bacterial interference
test procedure
Aerobic
Citrobacterfreundii 1 Simultaneous Deferred
Enterobacter cloacae 1
Organism (no. of strains) TSA BBA TSA BBA
Escherichia coli 4
Group D enterococci 3 Bacteroides distasonis (4) +
Hafnia alvei 1 Bacteroidesfragilis (2) +
Klebisella species 3 B. fragilis (2) + +
Morganella morganii 1 Bacteroides melaninogenicus
Proteus mirabilis 1 subspecies
Providencia species 1 intermedius (2) +
Staphylococcus species" 2 melaninogenicus (1) + + +
Streptococcus equinus 2 melaninogenicus (l) + +

Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on August 22, 2015

Streptococcus intermedius I melaninogenicus (2) +
Streptococcus mitis 2 Bacteroides ovatus (4) +
Streptococcus species 2 Bacteroidesputredinis (1) + +
Anaerobic Bacteroides thetaiotaomicron (5) +
Bacteroidesfragilis B. thetaiotaomicron (1) + +
Bacteroides melanlnogenicus Bacteroides vulgatus (5) + +
subspecies intermedius Peptococcus magnus (I) + +
B. melaninogenicus subspecies P. magnus (I) +
melaninogenicus It Peptostreptococcus anaerobius (I) + +
Bacteroidesputredinis 2 Peptostreptococcus micros (2) +
Bacteroides thetaiotaomicron It
Bifidobacterium adolescentis 3 NOTE. A plus sign indicates that all eight strains of C. dif-
Bifidobacterium bifidum 1 ficile inhibited the multiplication of the organism using the in-
Bifidobacterium breve 1 hibition procedure indicated; a minus sign indicates that none of
Bifidobacterium infantis 2 the eight strains of C. difficile inhibited the multiplication of
Bifidobacterium species 2 the organism using the inhibition procedure indicated. For
Clostridium butyricum 1 descriptions of the deferred and simultaneous test procedures,
Clostridium innocuum I see Materials and Methods. TSA = trypticase soy agar; BBA
Clostridium perfringens 1 = brucella blood agar.
Clostridium ramosum 1
Clostridium sphenoides 1
Clostridium tertium 1
Discussion
Eubacterium lentum 1
Eubacterium nitritogenes 2 The results of this study demonstrate the occur-
Eubacterium species 4
Lactobacillus acidophilus 1
rence of inhibitory interactions between bacterial
Lactobacillus catenaforme 2 components of the normal intestinal flora and C.
Lactobacillus species I difficile. The diameters of the zones of inhibition
Peptococcus asaccharolyticus 1 differed for the various bacteria (range, 7-45
Peptococcus prevotii 2 mm); for most bacteria examined, the diameter
Peptococcus species 1
Peptostreptococcus anaerobius It appeared to be independent of the growth rate.
Peptostreptococcus productus I The mechanism(s) of bacterial inhibitory activ-
Propionibacterium acnes 2 ity demonstrated in this study is unknown; how-
Veillonella parvula 2 ever, some general comments can be made. The
NOTE. Eight strains of C. difficile were examined for antagonism of C. difficile by Streptococcus spe-
bacterial antagonism against each indicator strain by the de- cies was probably not significantly related to the
ferred and simultaneous test procedures after growth on trypti- production of H 20 2 • One characteristic common
case soy agar and brucella blood agar (see Materials and
to three of the genera inhibiting C. difficile (Bifi-
Methods).
* Coagulase-negative. dobacterium, Lactobacillus, and Streptococcus) is
t Some strains of bacteria belonging to these species were the production of large quantities of lactic acid
inhibited by C. difficile (see table 6). during growth. The wide spectrum of microorgan-
C. difficile BacterialInterference
isms inhibiting C. difficile suggests that several
mechanisms may be involved, and it will be neces-
sary to determine which, if any, is important in the
intestinal tract.
C. difficile demonstrated inhibitory activity
against certain members of the intestinal bacterial
flora. This inhibitory activity was directed only
against three genera of anaerobic bacteria (Bacte-
roides, Peptococcus, and Peptostreptococcus),
but these genera are numerically prominent in the
normal fecal flora. The mechanism of this inhibi-
tory activity is unknown.
References
1. Bartlett, J. G. Antibiotic-associated pseudomembranous
colitis. Rev. Infect. Dis. 1:530-539, 1979.
2. George, W. L., Rolfe, R. D., Mulligan, M. E., Finegold,
S. M. Infectious diseases 1979-antimicrobial agent-in-
475
duced colitis: an update. J. Infect. Dis. 140:266-268,
1979.
3. Snyder, M. L. Normal fecal flora of infants between two
weeks and one year of age. I. Serial studies. J. Infect.
Dis. 66:1-16, 1940.
4. Hall, I. c., O'Toole, E. Intestinal flora in newborn infants.
Am. J. Dis. Child. 49:390-402, 1935.
5. Sutter, V. L., Citron, D. M., Finegold, S. M. Wadsworth
anaerobic bacteriology manual. 3rd ed. C. V. Mosby,
St. Louis, 1980. 131 p.
6. Holdeman, L. V., Cato, E. P., Moore, W. E. C. [ed.].
Anaerobic laboratory manual. 4th ed. Virginia Poly-
technic Institute, Blacksburg, 1977. 152 p.
7. Finegold, S. M., Martin, W. J., Scott, E. G. Diagnostic
microbiology. 5th ed. C. V. Mosby, S1. Louis, 1978.514 p.
Downloaded from http://jid.oxfordjournals.org/ at University of Sussex on August 22, 2015
8. Tagg, J. R., Dajani, A. S., Wannamaker, L. W. Bacterio-
cins of gram-positive bacteria. Bacteriological Reviews
40:722-756, 1976.
9. Gratia, A. Techniques selectives pour la recherche sys-
tematique des germes antibiotiques. C. R. Soc. BioI.
(Paris) 140:1053-1055, 1946.
10. Bauernfeind, A., Burrows, J. R. Suggested procedure al-
lowing use of plastic petri dishes in bacteriocin typing.
Appl. Environ. Microbiol, 35:970, 1978.