Anaerobe 18 (2012) 405e413

Contents lists available at SciVerse ScienceDirect

Anaerobe
journal homepage: www.elsevier.com/locate/anaerobe

Clinical microbiology

In vitro evaluation of single- and multi-strain probiotics: Inter-species inhibition

between probiotic strains, and inhibition of pathogens
C.M.C. Chapman*, G.R. Gibson, I. Rowland
Department of Food and Nutritional Sciences, University of Reading, P.O. Box 226, Whiteknights, Reading RG6 6AP, United Kingdom

a r t i c l e i n f o a b s t r a c t

Article history: Many studies comparing the effects of single- and multi-strain probiotics on pathogen inhibition
Received 27 September 2011 compare treatments with different concentrations. They also do not examine the possibility of inhibition
Received in revised form between probiotic strains with a mixture. We tested the ability of 14 single-species probiotics to inhibit
24 April 2012
each other using a cross-streak assay, and agar spot test. We then tested the ability of 15 single-species
Accepted 27 May 2012
Available online 4 June 2012
probiotics and 5 probiotic mixtures to inhibit Clostridium difficile, Escherichia coli and S. typhimurium,
using the agar spot test. Testing was done with mixtures created in two ways: one group contained
component species incubated together, the other group of mixtures was made using component species
Keywords:
Probiotic
which had been incubated separately, equalised to equal optical density, and then mixed in equal
Multi-species volumes. Inhibition was observed for all combinations of probiotics, suggesting that when used as such
Pathogen there may be inhibition between probiotics, potentially reducing efficacy of the mixture. Significant
Optical density inter-species variation was seen against each pathogen. When single species were tested against
Inhibition mixtures, the multi-species preparations displayed significantly (p < 0.05 or less) greater inhibition of
pathogens in 12 out of 24 cases. Despite evidence that probiotic species will inhibit each other when
incubated together in vitro, in many cases a probiotic mixture was more effective at inhibiting pathogens
than its component species when tested at approximately equal concentrations of biomass. This suggests
that using a probiotic mixture might be more effective at reducing gastrointestinal infections, and that
creating a mixture using species with different effects against different pathogens may have a broader
spectrum of action that a single provided by a single strain.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction information as to whether mixing of strains results in synergistic or

Probiotics are defined as live organisms which, when adminis-
tered in sufficient amounts, can have a beneficial effect on host
health [1]. They have been demonstrated to be effective in a wide
range of conditions including travellers’ diarrhoea, antibiotic-
associated diarrhoea, upper respiratory tract infections, atopic
eczema, and some inflammatory conditions although there are
species- and strain-specific aspects to their activities [2,3]. Potential
mechanisms of action include modulation of the intestinal immune
system, and displacement of potential pathogens via competitive
exclusion or production of antimicrobial agents including acids [3].
In a recent review [4], we found evidence that probiotic
mixtures also have beneficial effects against a wide range of
disorders, although the evidence that mixtures are more effective
than their component strains is more limited. There is little
* Corresponding author. Tel.: þ44 (0) 118 378 8700; fax: þ44 (0) 118 931 0080.
E-mail addresses: cb010202@reading.ac.uk, c.m.c.chapman@pgr.reading.ac.uk
(C.M.C. Chapman).
even additive effects in terms of bioactivity or in reduced efficacy
due to mutual inhibition by the component strains. Of 16 studies - 7
in vivo, 9 in vitro - in which a direct comparison was made between
a probiotic mixture and one or more of its component strains
against a variety of endpoints, 12 showed the mixture to be more
effective than the single strains [4]. For example, the use of pro-
biotic bacteria in combination can cause greater inhibition of
growth of three pathogens - Escherichia coli, Vibrio cholerae and
Salmonella enteritidis - than that produced by a single Lactobacillus
strain [5], and that combinations of lactobacilli were more effective
than single strains at preventing adhesion of Enterobacter sakazakii
to intestinal mucus suggesting a reduced risk of bacterial infection
[6]. However, Barone et al. [7] showed that a mixture containing 8
probiotic species had an effect which was not statistically different
from that of a single Lactobacillus species.
In many studies, the comparison was not strictly valid because in
each study the dose of multi-strain probiotic differed from that of the
single strain. In order to assess effectiveness of multi-strain probiotics,
research is needed using identical doses. It is of interest that 8 of the 12

1075-9964/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved.
10.1016/j.anaerobe.2012.05.004
406 C.M.C. Chapman et al. / Anaerobe 18 (2012) 405e413

studies showing greater effectiveness of mixtures used Lactobacillus Table 1

spp. while in all 4 studies where the mixture was less effective, Single probiotic strains used.

a Bifidobacterium sp. in the mixture was used. Conclusions that can be Name Collection number Probiotics
drawn from this are firstly that the presence of a Lactobacillus sp. may International
enhance the effectiveness of a mixture, while bifidobacteria may be reference number.

inhibited in the presence of other probiotic genera. Our review [4] also Lactobacillus acidophilus NCIMB 30184 PXN 35
Lactobacillus delbrueckii NCIMB 30186 PXN 39
concluded that the presence of a greater variety of probiotic genera
subsp. bulgaricus
within a mixture may reduce its effectiveness through mutual inhi- Lactobacillus casei NICMB 30185 PXN 37
bition by the different species, possibly by production of antagonistic Lactobacillus plantarum NCIMB 30187 PXN 47
agents, or by competition for either nutrients or binding sites within Lactobacillus rhamnosus NCIMB 30188 PXN 54
the gastrointestinal tract. Lactobacillus salivarius NCIMB 30225 PXN 57
ssp. Salivarius
A further potential advantage of multi-strain probiotics, in addition Lactobacillus fermentum NCIMB 30226 PXN 44
to exerting additive or synergistic effects, is that strain-specific effects Lactobacillus helveticus NCIMB 30224 PXN 45
of individual probiotic components could together exert a broader Bifidobacterium bifidum NCIMB 30179 PXN 23
spectrum of activity for example inhibition of a wider variety patho- Bifidobacterium breve NCIMB 30180 PXN 25
Bifidobacterium infantis NCIMB 30181 PXN 27
genic bacteria. Currently evidence for this is lacking.
Bifidobacterium longum NCIMB 30182 PXN 30
This study had three main aims: the first was to provide Streptococcus thermophilus NCIMB 30189 PXN 66
evidence for which probiotic strains may inhibit the growth of Lactococcus lactis NCIMB 30222 PXN 63
others in vitro, using firstly the cross-streak assay, then the agar Bacillus subtilis NCIMB 30223 PXN 21
spot test. The second aim was to determine whether multi-strain
probiotics would show greater of inhibition of a variety of patho-
gens than their constituent strains, using the agar spot test. Many Basingstoke, United Kingdom), species being allowed to grow
studies compare single- and multi-strain probiotics at different to different concentrations. The mixtures for this method were
concentrations, which confounds comparisons of efficacy [4]. Since created by using equal volumes of the different species, at
optical density (OD) is often used to provide rapid measurement of different concentrations. Thus single strains and mixtures were
biomass in bacterial cultures [8] e the correlation being a complex used at different concentrations of species. A commercial
function of particle size, shape and refractive index [9] e in this mixture of strains e ProtexinÔ (Probiotics International Ltd,
study we used the method of equalising the ODs of cultures to Stoke-sub-Hamdon, Somerset, UK) was prepared using
provide a method of testing these probiotics at approximately equal a different method. One capsule of the commercially-available
concentrations. The third aim was to compare a series of probiotic lyophilised bacteria was incubated overnight anaerobically in
mixtures to ascertain whether the component strains of a mixture MRS broth, and used in each assay.
had a significant influence upon the ability of that mixture to 2. Overnight cultures of single-strain probiotics in MRS broth
inhibit pathogens. were adjusted to the same OD at 650 nm (254 Colorimeter,
Sherwood Scientific, Cambridge, UK) in order to represent
2. Materials and methods approximately equal biomass [8]. From these cultures, various
probiotic mixtures were created using equal proportions of
2.1. Organisms used their constituent strains. This was not possible with the
commercial ProtexinÔ mixture.
Table 1 lists the probiotic species tested. Table 2 lists the
compositions of the mixtures used. All were obtained from the Thus 6 testing scenarios were created; three pathogens, each
collection of Probiotics International Ltd, Stoke-sub-Hamdon, tested using the adjusted and non-adjusted OD methods.
Somerset, UK. Pathogens tested were Clostridium difficile ATCC The method used by Barbosa et al. [11] was then adopted; i.e.
43594, E. coli NCFB 1989 and Salmonella typhimurium LT2. overnight cultures of probiotics were spotted onto MRS agar plates
in 5 ml aliquots e these spots represented the tester strain; after
2.2. Cross-streak assay

The method of Yoshida et al. [10] was adapted. Briefly, each Table 2
strain was streaked in 3 parallel lines onto de Man Rogosa Sharpe Constituent strains of mixtures used.
(MRS) agar (Oxoid, Basingstoke, United Kingdom) using a 1 ml loop.
Once these lines had dried (approximately 15 min), test strains
were streaked perpendicular to these initial strains in the same
fashion, giving 3 possible zones of inhibition for each combination
of strains. Using each of 14 species against each other gave 196
possible combinations, and each plate was done in triplicate. For
each combination, the species at the top of Table 3 was the (first-
streaked) tester strain. It was assumed that where there was inhi-
bition, it was caused by the tester strain hindering the growth of the
second-streaked (indicator) strain.
2.3. Agar spot test
This procedure was carried out using two methods.
1. From streak plates, individual probiotic species were incubated
anaerobically together overnight in MRS broth (Oxoid,
Mixture name Constituent strains
15 mix L. acidophilus, L. delbrueckii subsp.
Bulgaricus, L. casei, L. plantarum,
L, rhamnosus, L. salivarius ssp. Salivarius,
L. fermentum, L. helveticus, B. bifidum,
B. breve, B. infantis, B. longum,
Str. thermophilus, Lac. Lactis, B. subtilis
Bif mix B. bifidum, B. breve, B. infantis, B. longum
CC mix B. bifidum, B. breve, B. infantis, B. longum,
L. helveticus
LB mix L. acidophilus, L. delbrueckii subsp. Bulgaricus,
L. casei, L. plantarum, L, rhamnosus, L. salivarius
ssp. Salivarius, L. fermentum, L. helveticus
ProtexinÔ All strains at 5% of total unless otherwise stated:
L. acidophilus, L. delbrueckii subsp. Bulgaricus,
L. casei, L. plantarum (15%), L, rhamnosus (10%),
L. salivarius ssp. Salivarius, L. helveticus,
B. bifidum, B. breve, B. infantis, B. longum,
Str. thermophilus, Lac. Lactis, B. subtilis (20%)
 C.M.C. Chapman et al. / Anaerobe 18 (2012) 405e413 407

24 h incubation at 37  C in anaerobic conditions (DW Scientific,

was caused by the tester strain hindering the growth of the second-streaked (indicator) strain. Inhibition is rated from - to þþþ, - indicating no inhibition at all, þþþ indicating total inhibition, or no growth at all in the zone
Inhibition of probiotic strains by other probiotic strains using cross-streak method. For each combination, the species at the top of the Table was the (first-streaked) tester strain. It was assumed that where there was inhibition, it

L. helveticus
Shipley, UK) these plates were overlaid with 0.7%(w/v) MRS agar
inoculated with 5 ml of the indicator strain or pathogen. After 24 h
of further anaerobic incubation at 37  C, the zones of inhibition

þþ

þþ
þþ
þþ
þþ

þþ
þ

þ
þ
e

e
e
were measured with a clear ring around the spotted test strain was
L. fermentum
taken to indicate a zone of inhibition. Mean inhibition scores for
each probiotic test strain were taken to give an indication of that
strain’s inhibitory nature relative to the other strains used. OD for
þþþ

þþþ
þþþ
þþþ
þþþ

þþ
each probiotic treatment was measured in order to ascertain
þ

þ
þ

þ
e
e

e
a relationship between biomass of probiotic treatment and inhib-
L. salivarius

itory effect.
þþþ

þþþ

þþþ

þþþ

þþþ
2.4. Statistical analysis
þþ

þþ

þþ
þþ
þ
þ

þ
e

A Pearson Chi-square test was calculated for the cross-streak

L. rhamnosus

assay data to determine statistical significance between genera.

For the agar spot test, a one-way analysis of variance was per-
þþþ

þþþ
þþþ
þþþ
þþþ

þþþ
formed with post-hoc tests for Least Significant Difference to
þþ
þ
þ

þ
þ
e

determine the effect on the size of inhibition zone of probiotic

genus, and of single- or multi-strain treatments. Further Analysis of
L. plantarum

Variance with post-hoc tests for Least Significant Difference was

performed to rank performances of each probiotic mixture against
þþþ

þþþ
þþþ
þþþ

þþþ
þþ

þþ

þþ

all pathogens in both equal and non-equal proportions. p values of

þ

þ
e

<0.05 were considered significant. Calculations were performed

L. casei

using the SPSS software v.18 (Chicago, ILL, USA).

þþþ

þþþ
þþþ
þþþ
þþþ
þþ

þ
þ
e

e
e
e

3. Results
L. delbrueckii

3.1. Mutual inhibition of probiotic strains tested by cross-streak

assay
þþþ

þþþ
þþþ
þþþ
þþþ

þþþ
þ

þ
þ
þ
e
e

Table 4 shows the overall inhibition for each pair of strains.

L. acidophilus

Although all strains inhibited some others to a degree, a genus-

specific effect was observed (p < 0.01) indicating that lactobacilli
appeared more effective in inhibiting probiotic strains, while bifi-
þþþ

þþþ
þþþ
þþþ
þþ

þþ

dobacteria, with the exception of Bifidobacterium breve, had less

þ
e

e
e
e
e

effect in inhibiting any other species. Pearson chi-squared test

L. lactis

analysis of the inhibition scores between strains of different genera

þþþ

þþþ
þþþ

(data not shown) indicated that lactobacilli appeared more effec-

þþ

þþ
þ
e

e
e

e
e
e
e

tive at inhibiting strains of other genera, bifidobacteria far less so.

Streptococcus and Lactococcus species showed little ability to inhibit
S. Salivarius

strains from other genera. However, intra-genus inhibition data

(not shown) suggests that bifidobacteria were more effective at
þþþ

þþþ
þþ

þþ

inhibiting a strain of its own genus, while lactobacilli were far less
þ
e

e
e
e
e
e
e
e

inhibitory towards their own genus.

B. longum

3.2. Inhibition of probiotic strains tested by agar spot test

þþ

þ
e

e
e
e
e
e
e
e
e
e
e

Fig. 1 shows mean inhibition zone measurements for agar spot

B. infantis

test against other probiotic species (0.2e5.2 mm). All species

showed some inhibitory activity towards other probiotics. Within
the species tested, ANOVA analysis reveals a significant (p < 0.001)
þ
þ

þ
þ
þ
e
e
e
e

e
e
e

effect for genus, lactobacilli showing greater inhibition towards all

B. breve

probiotic strains (mean 4.2 mm). Bifidobacteria were less inhibitory

þþþ

þþþ
þþþ
þþþ
þþþ

(mean 2.0 mm) with the exception of B. breve which showed

þ
þ
þ

þ
þ
e
e

inhibitory capacity equivalent to that of lactobacilli. As in the cross-

where the streaks converged.

streak assay, Bacillus, Lactococcus and Streptococcus species tested

B. bifidum

showed little ability to inhibit probiotics of any genus.

þþ

þþ
þþ

þþ

þþ

þþ

When tested against strains of their own genus, bifidobacteria

þ

þ
e
e

tolerated each other well (range 1.2e2.6 mm, mean 1.7 mm), while
lactobacilli showed greater mutual inhibition (2.0e4.8 mm, mean
L. acidophilus

L. fermentum
L. rhamnosus
L. delbrueckii

L. plantarum

L. helveticus
S. Salivarius

L. salivarius

4.1 mm (data not shown).

B. bifidum

B. longum
B. infantis
B. breve

L. lactis

L. casei

When tested against strains of other genera, lactobacilli

Table 3

(1.9e5.5 mm, mean 4.5 mm) showed greater inhibition (p < 0.05)
than did bifidobacteria (0.3e4.4 mm, mean 2.1 mm), Streptococcus,
408 C.M.C. Chapman et al. / Anaerobe 18 (2012) 405e413

Table 4
Summary results of statistical analysis (one-way ANOVA) comparing inhibitory ability of mixtures vs. single-strain probiotics.

Pathogen tested

C. difficile E. coli S. typhimurium

Equal OD Non-equal OD Equal OD Non-equal OD Equal OD Non-equal OD

Multiple LB No difference Mix > single LB > mix No difference No difference LB > mix (p < 0.05)
mix vs. single LB (p < 0.001) (p < 0.01)
Multiple Bif Mix > single Mix > single Mix > single Mix > single No difference Mix > single (p < 0.05)
mix vs. single Bif (p < 0.001) (p < 0.05) (p < 0.05) (p < 0.01)
Multi-genus LB > mix No difference No difference LB > mix Mix > single Mix > single (p < 0.05)
mix vs. single LB (p < 0.001) (p < 0.01) (p < 0.001)
Multi-genus Mix > single No difference No difference Mix > single Mix > single Mix > single (p < 0.01)
mix vs. single Bif (p < 0.001) (p < 0.001) (p < 0.001)

Lactococcus or Bacillus sp. (0.2e3.3 mm, mean 1.7 mm). Lactoba-
cillus fermentum showed less ability to inhibit strains of other
genera as with strains of its own genus. B. breve (4.4 mm) showed
the greatest ability of any Bifidobacterium strain to inhibit other
genera (data not shown).
Using these data, and the data below showing the inhibitory
ability of each strain against the pathogens, it was suggested that
the optimum mixture might be created using the inhibitory ability
of a Lactobacillus sp. alongside the ability of bifidobacteria to co-
exist with less mutual inhibition. To this end, a mixture of all 4
bifidobacteria and Lactobacillus helveticus was created, indicated
here as “CC mix.”
3.3. Inhibition of Clostridium difficile tested by agar spot test
3.3.1. Equal optical densities
Mean measurements for inhibition using equal levels of biomass
(Fig. 2) ranged from 0.0 to 9.0 mm. Lactobacillus was the most
inhibitory single genus (mean 7.2 mm) and the mixture of
lactobacilli was the most inhibitory multi-strain treatment with
a mean zone of 7.3 mm. Zone diameters for bifidobacteria ranged
0.5e4.3 mm, while the mixture of all 4 bifidobacteria had a mean
zone of 6.2 mm. The 3 other genera showed little inhibitory ability
when used as single-strain treatments. Multi-strain preparations
showed a range of zone diameters from 4.9 to 7.3 mm. ANOVA
analysis reveals that under these testing conditions, Lactobacillus
spp. tested singly had a greater inhibitory effect than the multi-
genus preparations (p < 0.001), while the probiotic mixtures
were significantly more effective (p < 0.001) than the individual
bifidobacteria and other genera used singly.
3.3.2. Different optical densities
Mean measurements for inhibition against C. difficile were
obtained using the agar spot test for all 15 single-strains and all 5
multi-strain probiotics (Fig. 3). For single-strain treatments, lacto-
bacilli were the most inhibitory (mean 12.5 mm) ahead of bifido-
bacteria (mean 5.0 mm) and other species (0.1e5.9 mm, mean
3.8 mm). The mixture of all 15 strains scored higher (17.1 mm) than

Fig. 1. Agar spot test inhibition zones testing probiotic strains against each other. Results shown are mean values of inhibition zones produced by each named strain when tested
against all others. Bars indicate standard error (n ¼ 3 independent experiments).
 C.M.C. Chapman et al. / Anaerobe 18 (2012) 405e413 409

Fig. 2. Agar spot test inhibition zone sizes for inhibition of C. difficile using single- and multi-strain probiotics at equal optical densities (OD). Values shown are means (bar ¼ SE) of 3
independent experiments. ANOVA was used to assess differences between the different mixtures (p < 0.05) Lactobacillus spp. (p < 0.001) Bifidobacterium spp. (p < 0.001).

any single strain (0.1e16.0 mm). The mixture of lactobacilli showed
greater inhibition than any single-strain Lactobacillus sp, 20.1 mm
compared with 8.5e16 mm. The 4-species bifidobacteria mixture
showed a mean inhibition zone of 12.1 mm, higher than any single
species of the same genus with the exception of B. breve (14.1 mm).
ANOVA analysis shows that for treatments of different OD,
mixtures of lactobacilli and bifidobacteria were more inhibitory
than their component strains (p < 0.001, p < 0.05), although multi-
genus mixtures were no more or less effective than single strains
Fig. 4.
OD at 650 nm was measured for each treatment to observe any
relationship between biomass of probiotic and its ability to inhibit

Fig. 3. Agar spot test inhibition zone sizes for inhibition of C. difficile using single- and multi-strain probiotics at different optical densities (OD). Values shown are means (bar ¼ SE)
of 3 independent experiments. ANOVA was used to assess differences between the different mixtures (p < 0.001) Lactobacillus spp. (p < 0.01) Bifidobacterium spp. (p < 0.001).
410 C.M.C. Chapman et al. / Anaerobe 18 (2012) 405e413
Fig. 4. Agar spot test inhibition zone sizes for inhibition of E. coli using single- and multi-strain probiotics at equal optical densities (OD). Values shown are means (bar ¼ SE) of 3
independent experiments. ANOVA was used to assess differences between the different mixtures (no significant difference) Lactobacillus spp. (p < 0.001) Bifidobacterium spp.
(p < 0.001).
the pathogen (data not shown). An R2 value of 0.7492 was observed,
indicating a strong positive correlation between probiotic cell
density and inhibition.
3.4. Inhibition of Escherichia coli tested by agar spot test
3.4.1. Equal optical densities
Of the single-strain treatments, lactobacilli (mean 11.9 mm)
were more effective than bifidobacteria (mean 9.5 mm). Multi-
strain treatments created a mean inhibition zone diameter of
11.4 mm. The mixture of all 8 lactobacilli (9.7 mm) showed less
inhibition than any single strain of that genus, while the mixture of
4 bifidobacteria (12.2 mm) showed a relatively high level of inhi-
bition. In this experiment, ANOVA analysis indicates that multi-
genus probiotics had no significantly different inhibitory effect
than single Lactobacillus, Lactococcus, Streptococcus or Bacillus
species. The bifidobacteria mixture was more effective than Bifi-
dobacterium species tested singly (p < 0.05), while individual lac-
tobacilli showed greater inhibition than a mixture of this genus
(p < 0.01).
3.4.2. Different optical densities
Overall, the study obtained mean inhibition zone scores against
E. coli for 15 single-strain probiotics and 5 multi-strain mixtures
with a wide range of values (Fig. 5). Mean measurements for multi-
strain preparations compared favourably with single-strains: the
mixture of 8 lactobacilli had a mean score of 14.5 mm, at the upper
end of the range of scores for single-species lactobacilli. The
mixture of 4 bifidobacteria had a mean score of 12.7 mm, higher
than the range of scores from single strain. ANOVA analysis
suggests that an individual Bifidobacterium sp. has a lesser inhibi-
tory effect than mixtures of either bifidobacteria or multiple genera
(p < 0.01, p < 0.001), while an individual Lactobacillus is more
inhibitory than a multi-genus mixture (p < 0.01) and no different
from a multi-genus mixture.
OD at 650 nm was measured for each treatment to observe any
relationship between biomass of probiotic and its ability to inhibit
the pathogen. The R2 value of 0.6974 was observed (data not
shown).
3.5. Inhibition of Salmonella typhimurium tested by agar spot test
3.5.1. Equal optical densities
Zones of inhibition against S. typhimurium were measured
(Fig. 6). Lactobacillus was the most inhibitory single genus (9.7 mm)
while bifidobacteria generated a mean inhibition zone of 8.1 mm.
Little inhibition was shown by other single species (mean 6.0 mm).
A mean inhibition zone of 12.2 mm was observed for probiotic
mixtures. Under these testing conditions, ANOVA suggests that
single lactobacilli and bifidobacteria are no more inhibitory than
using a mixture of these genera, while a multi-genus mixture is
more effective than either a single Lactobacillus or Bifidobacterium
sp (p < 0.001).
3.5.2. Different optical densities
Inhibition zones were obtained for all 15 single strains, and all 5
mixtures (Fig. 7). Lactobacillus was the most inhibitory single
species (mean 9.7 mm). Bifidobacterium spp (7.1e11.2 mm, mean
10.0 mm) showed lower levels of inhibition, although the mixture
of 4 bifidobacteria showed greater inhibition than any of its single
components. Other single species showed lower degrees of inhi-
bition (mean 4.6 mm). Multi-strain treatments gave inhibition
zones with a mean value of 12.2 mm. ANOVA analysis of this assay
indicates that when tested against a single Lactobacillus or Bifido-
bacterium, a multi-genus probiotic mixture has significantly greater
effect (p < 0.05, p < 0.01). Further, a mixture of bifidobacteria
shows greater inhibition than a single species (p < 0.05), although
a single Lactobacillus sp. shows greater inhibition than a mixture of
that genus (p < 0.05).
 C.M.C. Chapman et al. / Anaerobe 18 (2012) 405e413 411

Fig. 5. Agar spot test inhibition zone sizes for inhibition of E. coli using single- and multi-strain probiotics at different optical densities (OD).). Values shown are means (bar ¼ SE) of
3 independent experiments. ANOVA was used to assess differences between the different mixtures (no significant difference) Lactobacillus spp. (p < 0.001) Bifidobacterium spp.
(p < 0.05).

Optical densities at 650 nm were measured for each probiotic
treatment to observe any relationship between biomass of pro-
biotic and its ability to inhibit the pathogen. The R2 value of 0.2565
was observed (data not shown), suggesting a weak correlation
between probiotic cell density and inhibition of this pathogen.
3.6. Analysis of single- vs. multi-strain preparations
Table 4 shows the results of Analysis of Variance of the inhibi-
tion of all pathogens, with post-hoc tests to compare the efficacy of
mixtures with that of lactobacilli, bifidobacteria and other single

Fig. 6. Agar spot test inhibition zone sizes for inhibition of S. typhimurium using single- and multi-strain probiotics at equal optical densities (OD).). Values shown are means
(bar ¼ SE) of 3 independent experiments. ANOVA was used to assess differences between the different mixtures (p < 0.001) Lactobacillus spp. (p < 0.001) Bifidobacterium spp.
(p < 0.001).
412 C.M.C. Chapman et al. / Anaerobe 18 (2012) 405e413

Fig. 7. Agar spot test inhibition zone sizes for inhibition of S. typhimurium using single- and multi-strain probiotics at different optical densities (OD). Values shown are means
(bar ¼ SE) of 3 independent experiments. ANOVA was used to assess differences between the different mixtures (p < 0.001) Lactobacillus spp. (p < 0.001) Bifidobacterium spp. (no
significant difference).

strains. In 50% (12/24) tests, there was a statistically greater level of used in a multi-strain preparation. This is consistent with the
inhibition when a multi-strain treatment was used. In 8 cases, there findings of Be’er et al. [12] who observed inhibition between
was no statistical difference when single- or multi-strain treatment similar, or closely-related, strains of bacteria. This could potentially
was used, while in 4 cases using a single-strain Lactobacillus was reduce the efficacy of a mixture when compared with using single
more effective than using a probiotic mixture. strains as reported by Grandy et al. [13] and Myllyluoma et al. [14].
In both our assays, lactobacilli showed the greatest inhibition of
3.7. Overall ranking of probiotic mixtures other probiotic species, followed by the bifidobacteria, with the
other species showing little or no inhibition of other probiotic
Initial analysis (Table 5) of the mean inhibition zone sizes strains. Three explanations are plausible here. Lactobacilli may
created by each multi-strain treatment suggests that the mixtures produce a greater quantity of antimicrobial substances, or these
can be ranked in order of effectiveness. This was confirmed by post- substances have a broader spectrum of activity. A further possibility
hoc tests for mixture from a one-way Analysis of Variance; is that lactobacilli compete better for nutrients than other strains.
however, the differences between mixtures were non-significant. Results from both assays indicate that bifidobacteria showed
good tolerance to strains of the same genus. This suggests that such
4. Discussion inhibitory properties may be due to the production of acids,
bacteriocins or other metabolites to which other bifidobacteria are
The first aim of the study was to assess the ability of probiotic tolerant. Different species of bifidobacteria may produce identical
strains to inhibit the growth of others in vitro, in order to inform or near-identical metabolites, although data on metabolic activities
choices of which strains to use in multi-species probiotic treatment. of this genus are limited [15].
Results from both the cross-streak assay and agar spot test showed The aims of the agar spot test against pathogens were to
that all the probiotic strains were able to inhibit at least some other determine whether single strains of probiotics were more or less
probiotic strains suggesting that they may inhibit each other when effective than mixtures and to compare the relative efficacy of 5
multi-strain probiotic treatments in the inhibition of a range of
pathogens. These assays reveal an overall genus-specific effect in
Table 5 that lactobacilli were more inhibitory towards pathogens than
Overall mean inhibition zone diameter (mm) for each mixture taking into account bifidobacteria and strains from other genera (p < 0.05). The results
all pathogens tested and both testing methods (equal and non-equal OD). also show species-specific effects within the genera. The most
Non-equal Equal OD Combined notable of such effects was that lactobacilli and bifidobacteria at
OD mixtures mixtures methods data equal OD, showed species-specific variation in inhibition of all
15 mix 14.2 8.9 12.8 pathogens (p < 0.001). These data lend weight to the hypothesis
LB mix 13.2 9.1 12.1 that different probiotics have different levels of inhibition towards
Bif mix 13.4 11.4 11.1 different pathogens. In this assay we observed a similar genus-
ProtexinÔ 10.4 11.5 11.0
CC mix 10.3 9.9 10.1
specific effect of inhibition against two Gram negative, facultative
anaerobes (S. Typhimurium and E. coli) and a Gram positive, strictly
 C.M.C. Chapman et al. / Anaerobe 18 (2012) 405e413
anaerobic pathogen (C. difficile). This suggests that antimicrobial
effects of probiotics may occur irrespective of oxygen tolerance or
the Gram positive/negative nature of the pathogen.
It is noteworthy that in 12 out of 24 scenarios against all path-
ogens with probiotics at equal and non-equal OD (Table 4), a pro-
biotic mixture was statistically more effective than single strains
suggests a broad spectrum of inhibition, as observed by Lin et al.
[16]. Greater inhibition by multi-strain treatments observed here
also support data from Apella et al. [17] and Drago et al. [5] which
suggest a multi-species probiotic can have a greater inhibitory
effect than single strains included in that mixture.
In most previous studies of mixtures versus single strains, the
probiotic cultures were used at different concentrations, as described
by Köll et al. [18] and Hutt et al. [19]. Under these conditions, our
results show that using a probiotic mixture was more effective in 7 of
12 situations, as observed by Ridwan et al. [20]. Mixtures always
showed greater or equal inhibition against C. difficile while against
E. coli a single Lactobacillus strain was more effective than a multi-
genus mixture and, against S. typhimurium, a mixture of lactobacilli
was less effective than a single Lactobacillus species. By analysing the
relationship between OD and the inhibitory effects of the various
strains and mixtures, a correlation could be established between the
concentration of the probiotic and its ability to inhibit pathogen
growth. This suggests that the greater effectiveness of a multi-strain
over a single-strain preparation seen above and in other studies
[5,16,17] may be due simply to a greater concentration of probiotics in
the probiotic mixture, and emphasises the importance of making
comparisons on the basis of equal biomass.
In the assays using equal OD, in 5 of 12 cases, the mixture
showed significantly more inhibition, suggesting a synergistic
effect of strains, supported by the range of pathogens inhibited
here. Since a mixture was more inhibitory than a single strain in at
least one assay against each pathogen, these results lend greater
weight to the hypothesis that a probiotic mixture may have both
a greater effect and a wider range of prevention of gastrointestinal
pathogens, since these assays represent a method of testing
a multi-species probiotic against its constituent strains at approx-
imately equal treatment concentration.
However, in testing E. coli against probiotics of equal OD, indi-
vidual Lactobacillus spp. proved more inhibitory than a mixture of
lactobacilli, indicating that in this scenario using a mixture of strains
may not be the most effective at preventing gastrointestinal infection.
The fact that with each method there were instances where the
mixture proved equally or less inhibitory than a single strain may point
to one drawback of the OD method e this method cannot differentiate
between living and dead cells. In the cases where the mixture was no
more effective, it may be that in these cases there were fewer viable
probiotic cells, leading to reduced efficacy of the mixtures.
Comparative data ranking the efficacy of the 5 mixtures failed to
establish a correlation between the constituent strains and their
inhibitory ability, since there was no statistical difference in the
effectiveness of the mixtures. The CC mix was formulated as a result
of data indicating lower inhibition between its component strains.
However, this mix was statistically no more effective than the other
mixtures. This indicates that the selection criteria for component
strains in a mixture are dependent upon more than inhibition
betweens strains, and that the strains used in this mixture may not
have shared the synergy shown by other mixtures.
We can draw several conclusions from this study. First, within
a multi-strain probiotic preparation there can be inhibition
between the component strains (lactobacilli showing greater
inhibition than other genera) leading to the possibility of reduced
efficacy within a mixture. However, data from the agar spot test
supports our hypothesis that despite mutual inhibition, mixtures
can generally maintain inhibition of a range of pathogens as
413
effective as that of their component strains in vitro, although there
is some variation. This occurs even when single- and multi-strain
preparations are tested at an approximately equal amount of
biomass, something which has seldom been tested in previous
studies. Against S. typhimurium, a multi-genus mixture appeared
more effective than a single Lactobacillus or Bifidobacterium sp.,
regardless of biomass. However, biomass density appears to be
important, as our data show that a correlation can be made
between increased biomass and increased inhibition of pathogens.
Acknowledgement
This work forms part of a PhD project funded by Probiotics
International, Somerset, UK. Probiotics International had no role in
the preparation of the article, study design, data collection, analysis,
interpretation, or report writing.
References
[1] Joint FAO/WHO Expert Consultation on Evaluation of health and Nutritional
properties of probiotics in Food including Powder Milk with live Lactic acid
Bacteria (2001) Health and Nutritional properties of probiotics in Food
including Powder Milk with live Lactic acid bacteria. Food and Agriculture
Organization of the United Nations, World Health Organization.
[2] Rowland I, Capurso L, Collins K, Cummings J, Delzenne N, Goulet O, Guarner F,
Marteau P, Meier R. Current level of consensus on probiotic scienceereport of an
expert meetingeLondon, 23 November 2009. Gut Microbe 2010;1(6):436e9.
[3] Gibson GR, McCartney AL, Rastall RA. Prebiotics and resistance to gastroin-
testinal infections. Br J Nutr 2005;93(Suppl 1):S31e4.
[4] Chapman CM, Gibson GR, Rowland I. Health benefits of probiotics: are
mixtures more effective than single strains? Eur J Nutr 2011;50(1):1e17.
[5] Drago L, Gismondo MR, Lombardi A, de Haen C, Gozzini L. Inhibition of in vitro
growth of enteropathogens by new Lactobacillus isolates of human intestinal
origin. FEMS Microbiol Lett 1997;153(2):455e63.
[6] Collado MC, Isolauri E, Salminen S. Specific probiotic strains and their
combinations counteract adhesion of Enterobacter sakazakii to intestinal
mucus. FEMS Microbiol Lett 2008;285(1):58e64.
[7] Barone C, Pettinato R, Avola E, Alberti A, Greco D, Failla P, et al. Comparison of
three probiotics in the treatment of acute diarrhea in mentally retarded
children. Minerva Pediatr 2000;52(3):161e5.
[8] Shuler M, Kargi F. Bioprocess engineering: basic concepts. 2nd ed. Singapore:
Pearson Education; 2005.
[9] Eaton A, Cleresci L, EW R, AB G, editors. Standard methods for the examination
of water and wastewater. Washington, D.C: American Public Health Associa-
tion; 2005.
[10] Yoshida K, Nasu Y, Shitami N, Toyoda H, Takemura H, Oomori K. A novel
convenient method for high bacteriophage titer assay. Nucleic Acids Symp Ser
(Oxf) 2009;53:315e6.
[11] Barbosa TM, Serra CR, La Ragione RM, Woodward MJ, Henriques AO.
Screening for bacillus isolates in the broiler gastrointestinal tract. Appl
Environ Microbiol 2005;71(2):968e78.
[12] Be’er A, Zhang HP, Florin EL, Payne SM, Ben-Jacob E, Swinney HL. Deadly
competition between sibling bacterial colonies. Proc Natl Acad Sci U S A 2009;
106(2):428e33.
[13] Grandy G, Medina M, Soria R, Teran CG, Araya M. Probiotics in the treatment
of acute rotavirus diarrhoea. A randomized, double-blind, controlled trial
using two different probiotic preparations in Bolivian children. BMC Infect Dis
2010;10(1):253.
[14] Myllyluoma E, Ahonen AM, Korpela R, Vapaatalo H, Kankuri E. Effects of
multispecies probiotic combination on helicobacter pylori infection in vitro.
Clin Vaccine Immunol 2008;15(9):1472e82.
[15] WIl ed.. In: Versalovic J, Wilson M, editors. Therapeutic Microbiology: pro-
biotics and related strategies. Washington DC: ASM Press; 2008
[16] Lin JS, Chiu YH, Lin NT, Chu CH, Huang KC, Liao KW, et al. Different effects of
probiotic species/strains on infections in preschool children: a double-blind,
randomized, controlled study. Vaccine 2009;27(7):1073e9.
[17] Apella MC, Gonzalez SN, Nader de Macias ME, Romero N, Oliver G. In vitro
studies on the growth of Shigella sonnei by Lactobacillus casei and Lact.
acidophilus. J Appl Bacteriol 1992;73(6):480e3.
[18] Koll P, Mandar R, Marcotte H, Leibur E, Mikelsaar M, Hammarstrom L. Char-
acterization of oral lactobacilli as potential probiotics for oral health. Oral
Microbiol Immunol 2008;23(2):139e47.
[19] Hutt P, Shchepetova J, Loivukene K, Kullisaar T, Mikelsaar M. Antagonistic
activity of probiotic lactobacilli and bifidobacteria against entero- and uro-
pathogens. J Appl Microbiol 2006;100(6):1324e32.
[20] Ridwan BU, Koning CJ, Besselink MG, Timmerman HM, Brouwer EC, Verhoef J,
et al. Antimicrobial activity of a multispecies probiotic (Ecologic 641) against
pathogens isolated from infected pancreatic necrosis. Lett Appl Microbiol
2008;46(1):61e7.