Available via license: CC BY-NC-ND 4.0
Content may be subject to copyright.
610
Journal of Food Protection, Vol. 66, No. 4, 2003, Pag es 610–617
Copyright q, In ternation al Associa tion fo r Fo od Pr otection
Effect of Ethanol on the Growth of Clostridium botulinum
DAPHNE PHILLIPS DAIFAS,1JAMES P. SMITH,1*BURKE BLANCHFIELD,2BRIGITTE CADIEUX, 2
GREG SANDERS,2A ND JOHN W. AUSTIN2
1Department of Foo d Science and Agricultural Chemi stry, M cGill University, Macdonald Camp us, 21,111 Lakeshore Roa d, Ste. Anne de Bellevue,
Quebec, Canada H9X 3V9; a nd 2Bureau of Microbial Hazards, Food Directorate, Health Products and Food Branch, Health Canada, Ottawa,
Ontario, Can ada K1A 0L2
MS 02 -161: Received 21 May 200 2/Accepted 20 October 2002
ABSTRACT
Model broth studies were carried out to investigate the effect of ethanol on the growth of proteolytic (group I) strains of
Clostridium botulinum. Ethanol extended the pathogen’s lag phase, decreased its exponential growth rate, and decreased its
nal level of growth in the stationary phase. In all cases, botulinum neurotoxin production was associated with growth.
Micrographs of C. botulinum cultures grown at 378C in trypticase peptone glucose yeast extract (TPGY) broths containing 2
and 4% ethanol showed elongation of vegetative cells and interference with cell division. The inhibition of growth and toxin
productionat the ethanol level predicted (5.5%, wt/wt) was conrmed by microscopy and by the mouse bioassay.A subsequent
study was carried out to determine the combined effect of ethanol (0 to 8% [wt/wt]), water activity (aw; 0.953 to 0.997), and
pH (6.2 to 8.2) on the probability of the growth of and neurotoxin production by proteolytic strains of C. botulinum (103
spores per ml). Growth and neurotoxin production occurred in 1 to 3 days in TPGY broths without ethanol (0%) and in 2 to
4 days in broths containing 2% ethanol regardless of the awor pH levels (P,0.005). Growth and neurotoxin production
were delayed by an ethanol concentration of 4% ethanol and completely inhibited by a concentration of 6%. At an ethanol
concentration of 4%, the probability of growth and toxin production over 365 days (Pt) was in uenced by awand pH. After
365 days, the maximum probability of growth and toxin production (Pmax ) was 1 for all but one combination. However, t,
the time it took for 50% of all eventually positive replicates for any given combination of barriers to show growth and/or
turbidity, ranged from ,3 to 229 days. All tubes of TPGY broths that showed no growth after 365 days were subcultured in
fresh TPGY broths. In all cases, growth and toxin production occurred within 24 h at 378C, indicating the reversible(sporostatic
and/or bacte riostatic) effect of e thanol on C. botulinum.
Modi ed atmosphere packaging (MAP) involving vac-
uum packaging, gas ushing, and oxygen absorbent tech-
nology has been used to extend the shelf lives of high-
moisture bakery products (20). However, concern about the
safety of these products has been expressed, particularly
with respect to Clostridium botulinum, a spore-forming an-
aerobe that produces a potent neurotoxin. In minimally pro-
cessed foods such as bakery products, proteolytic spores, if
present in the raw ingredients, will readily sur vive baking.
Since many high-moisture bakery products have a water
activity (aw) of .0.95 and a pH of .6(20), these products
have the potential to support the growth of and toxin pro-
duction by this pathogen (9). It is therefore important to
ensure that the potential growth of C. botulinum in mini-
mally processed high-moisture MAP bakery products is
prevented, since the consumption of food in which this
pathogen has grown and produced neurotoxin often results
in fatal botulism.
To extend product shelf life and increase safety, the
food industr y commonly uses a multibarrier approach. This
approach relies on several factors, or hurdles, that act in
conjunction with each other additively or synergistically to
inhibit or prevent microbial growth (12). While appropriate
combinations of low awand low pH may be effective
* Author for corresp ondence. Tel: (514 ) 398-7923; Fax: (514) 398-7977;
E-mail: jsmith@macdo nald.mcgill.ca.
against the growth of C. botulinum in bakery products,
many products cannot be reformulated to such levels with-
out a loss of sensor y acce ptability. Howev er, o ptimum com -
binations of inhibitory factors involving lower levels of
each factor can often result in the enhancement of the safety
of products.
One additional barrier that, in conjunction with reduced
awand/or reduced pH, may prove effective in controlling
the growth of C. botulinum is ethanol. Recently, ethanol
has been shown to inhibit the growth of and toxin produc-
tion by proteolytic C. botulinum in English-style crumpets,
a high-moisture bakery product that was found to be toxic
with this pathogen at ambient temperature within 1 week
in challenge studies (6). Although ethanol delayed the
growth of and toxin production by C. botulinum in inocu-
lated crumpets, the shelf life of this product was limited
because of the product absorbed ethanol from the package
headspace (7). Nevertheless, these studies demonstrated the
antibotulinal effects of ethanol. However, little is known
about the minimum levels of ethanol that could be used to
enhance the safety of minimally processed bakery products
without compromising product quality. Therefore, the ob-
jective of this study was to model the effects of ethanol,
alone and in combination with awand pH, on the growth
of and toxin production by C. botulinum in model broth
studies.
Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/4/610/1675368/0362-028x-66_4_610.pdf by guest on 12 January 2023
J. Foo d Prot., Vol. 66, No. 4 EFFECT OF ETHANOL ON C. BOTULINUM 611
MATERIALS AND METHODS
Growth curve: cultures and growth conditions. A total of
six group I (i.e., proteolytic) strains of C. botulinum (17A, 62A,
CK2A, MRB, 13983IIB, and 368B) were used in this study. Each
C. botulinum strain was grown from frozen stock culture on
McClung Toab e ag ar (Difco , Bect on-Dickin son, Sparks, M d.) a nd
incubated anaerobically in an atmosphere of 10% H2, 80% N2,
and 10% CO2at 358C. Overnight broth cultures were prepared in
tryp ticase pept one gluco se yeast ex tract (TPGY) br oth (Difco ) and
incubated anaerobically to provide nal inoculum levels of ap-
proximately 108CFU/m l. Ino culum lev els were veri ed by plating
appropriate dilutions in duplicate on McClung Toabe agar. Flasks
of TPGY broth (500 ml) preheated to 378C and containing 0, 1,
2, 3, or 4% (wt/wt) 95% ethanol were inoculated with 0.3 ml of
the composite mixture of C. botulinum to provide a nal inoculum
level of 2 3104CFU/ml. All asks were tightly covered with
sterile aluminum foil and incubated in an atmosphere of 10% H2,
80% N2, and 10% CO2in an anaerobic chamber (Coy Laborato-
ries, Detroit, Mich.) at 378C. Growth was monitored by plating
(in the anaerobic chamber) appropriate decimal dilutions of in-
oculated TPGY on TPGY agar and incubating the agar anaero-
bically at 378C at ;90-min intervals for 28 h. Viable cell counts
were determined and growth was simultaneously monitored
through the measurement of the optical density at 600 nm (OD600)
at 90-min intervals with a Bausch and Lomb Spec 20 spectro-
photom eter ( Fish er Scie nti c, Ottawa, Ontario, Canad a) un til t he
stationary phase was reached. At the end of the growth period (28
h), 4-ml portions of culture from each ask were lter sterilized
with a 0.22-mm lter (Acrodisc, Gelman Sciences, Ann Arbor,
Mich.). Filtrates were held at 48C until they were assayed for
botulinum neurotoxin by the mouse bioassay.
Growth curve: data treatment. Growth cur ve data were
tted to sigmoidal curves with a four-parameter logistic equation
by using Prism 2.01 (GraphPad Software, Inc., San Diego, Calif.),
and the rst (]CFU/ml/]t) and second derivatives of the sigmoidal
curves were determined with Mathcad 5.0 (MathSoft Inc., Cam-
bridge, Mass.). The times at which the absolute growth rates were
highest were determined as the maxima of the cur ves of the rst
derivatives plotted versus time. The ends of the lag and exponen-
tial growth phases were analyzed as the maxima and minima,
respectively, of the second derivativesplotted versus time accord-
ing to the method of Buchanan and Cygnarowicz (3). The ethanol
level predicted to result in the complete inhibition of proteolytic
strains of C. botulinum was obtained through extrapolation of a
linear regression analysis of the maximal absolute growth rates
versus the percentage (wt/wt) of ethanol.
Combined effects of ethanol, aw, and pH: preparation of
spore inoculum and inoculation of TPGY broths. Equal num-
bers of spores of each strain were combined to form a single
suspension of approximately 105spores per ml. Inoculum levels
were veri ed prior to the inoculation of tubes of TPGY broth as
described above. The spore mixture was heat shocked at 758C for
20 min prior to inoculation. Flasks of TPGY broth were prepared
at appropriate strengths to ensure that all broths would contain
the same amount of medium after ethanol was added. Flasks of
TPGY broth were then adjusted with appropriate volumes of 0.1
N HCl or 0.1 N NaOH to provide a pH range of 6.2 to 8.2. The
pH was checked with a previously calibrated Fisher Accumet pH
meter ( Fisher Scie nti c). Th e pH-ad justed TPGY b roths were sub -
sequently adjusted with appropriate volumes of glycerol to pro-
vide an awrange of 0.953 to 0.997, taking into account that eth-
anol would have a depressant effect on the awof the broth (15).
Water activity was determined with an Aqu alab water activity me-
ter (Decagon Devices, Inc., Pullman, Wash.). After autoclaving
and cooling to 258C, portions of TPGY broth were aseptically
dispensed into sterile tubes. Ethanol (0 to 8% [wt/wt]) was added
to tubes of adjusted sterile TPGY broth immediately prior to in-
oculation with 0.1 ml of spore suspension (105spores per ml) to
provide a nal inoculum level of 103spores per ml. Tubes were
covere d with appr oximately 3 ml of sterile Vaspar (45% p araf n,
55% white petroleum) and incubated at 258C. The combination
of factors (ethanol, aw, and pH) and the number of replicatetubes
for each treatment condition were based on the primary model
previously used by Whiting and Call (23) and are shown in Table
1. Tubes were checked twice daily for turbidity and/or gas pro-
ductio n. Tube s showi ng eviden ce of gro wth were a ssayed for neu-
rotoxin by the mouse bioassay.
Data treatment. To determine the probability of growth for
a given set of conditions (percentage [wt/wt] of ethanol, aw, and
pH) as a function of time, data were tted to a logistic equation
with Prism 2.01 (GraphPad). This cumulative probability distri-
bution function has previously been used by Whiting and Call
(23) and by Whiting and Oriente (24) to model the time to tur-
bidity for proteolytic and nonproteolyticC. botulinum, respective-
ly. As outlined by these authors, the parameters of this primary
model describe the n umber of positive s amples accumulati ng with
increasing storage time and provide information about the time to
turbidity, the rate at which samples show evidence of turbidity,
and the proportion of samples that do not become turbid. This
model is as follows:
Pmax
P5
tk(t2t)
[1 1e]
where Ptis the cumulative probability of growth at time t(i.e.,
the probability that growth occurred between [0, t]), tis the time
in days, Pmax is the maximum probability of growth after 365
days (i.e., the number of tubes under a given set of conditions
showing evidence of growth [turbidity and/or gas production] di-
vided by the total number of tubes under that set of conditions),
kis the rate of increase in the number of positive (i.e., turbid)
tubes (per day), and tis the time of the midpoint of the function
in days (i.e., the time taken for 50% of the replicates that will
eventually become toxic to show turbidity and/or gas production).
TEM. Preparation for transmission electron microscopy
(TEM) was carried out essentially according to the method of
Austin et al. (1). Cells were xed in 0.2 M cacodylate buffer (pH
7.4) containing 2.5% (vol/vol) glutaraldehyde and post- xed in
0.2 M cacodylate buffer (pH 7.4) containing 1% osmium tetrox-
ide. Cells were enrobed in 1% Nob le ag ar and dehydra ted through
a graded series of ethanol concentrations (15 min each in 50, 70,
and 90% ethanol, followed by three 20-min incubations in 100%
ethanol). Samples were then in ltrated and embedded in Taab 812
resin ( Marivac, Halif ax, Nova Scotia, Ca nada). Thi n sect ions were
cut on a Re ichert-Ju ng Ultracut E ultramicro tome (C. Reiche rt Ag,
Vienna, Austria) and stained with uranyl acetate and lead citrate
(14). Thin sections were examined with a Zeiss EM902 transmis-
sion electron microscope (Carl Zeiss, Thornwood, N.Y.) operating
at 80 kV with the energy loss spectrometer in place.
Nature of the effect of ethanol. Portions (0.1 ml) of inoc-
ulated TPGY broths from tubes containing 0 to 8% (wt/wt) eth-
anol that did not show any evidence of growth (i.e., turbidity or
gas production) or toxin production after 1 year at 258C were
transferred to fresh TPGY broth to dilute the ethanol to ,0.08%
(wt/wt), incubated anaerobically at 378C, and observed for evi-
Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/4/610/1675368/0362-028x-66_4_610.pdf by guest on 12 January 2023
J. Food Prot., Vol. 66, No. 4612 DAIFAS ET AL.
TABLE 1. Effects of various combinations of ethanol, water activity (aw), and pH on probability of growth of, and toxin production
by, Clostridium botulinum (103spores per ml) in trypticase peptone glucose yeast broth at 258C
Conditio n
% (wt/wt)
ethanol awpH
No. of
replicate
tubes
Day of
growtha
Value for model parametersb
Pmax t k
1
2
3
4
5
6
7
8
0
0
0
0
0
0
0
0
0.956
0.960
0.966
0.976
0.980
0.997
0.997
0.997
7.7
7.2
6.2
8.2
6.7
6.2
6.7
7.2
10
10
10
10
10
5
5
5
3
3
3
3
2
1
1
1
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
9
10
11
12
13
14
15
0
0
2
2
2
2
2
0.997
0.997
0.954
0.959
0.965
0.970
0.972
7.7
8.2
7.7
7.2
6.2
8.2
7.2
8
5
20
20
15
15
15
1
3
3
4
3
3
3
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
16
17
18
19
20d
20e
21
2
2
4
4
4
4
4
0.989
0.989
0.956
0.959
0.962
0.962
0.965
6.7
7.2
7.7
7.2
8.2
8.2
6.2
15
10
25
25
25
25
25
2
2
49–.365c
60–143c
14–104c
192–263c
23–26c
NF (NF)
NF (NF)
0.18 (0.01)
1 (0.13)
0.22 (0.04)
1 (0.06)
1 (0.1)
NF (NF)
NF (NF)
210.2 (0.4)
96.66 (1.5)
59.4 (0)
232.9 (0.5)
24.49 (0.2)
NF (NF)
NF (NF)
0.92 (0.04)
0.298 (0.07)
0.93 (0.08)
0.026 (0)
1.33 (0.2)
22
23
24
25
26
27
28
4
4
4
6
6
6
6
0.967
0.981
0.981
0.953
0.954
0.973
0.973
7.2
7.2
7.7
7.2
7.7
7.2
7.7
25
25
25
15
15
15
15
6–9c
4
3
NG
NG
NG
NG
1 (0.06)
NF (NF)
NF (NF)
0 (NF)
0 (NF)
0 (NF)
0 (NF)
8.97 (0.2)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
4.8 (4.7)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
29
30
8
8
0.964
0.964
7.2
7.7
5
5
NG
NG
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
NF (NF)
aDay on which growth was observed simultaneously in all replicates. NG, no growth over 365 days.
bPmax is the maximum probability of growth after 365 days (i.e. the number of tubes under a given set of conditions showing evidence
of growth (turbidity and/or gas production) divided by the total number of tubes under that set of conditions), kis the rate of increase
in the number of positive tubes (per day), and tis the time of the midpoint of the function (days) (i.e., the time taken for 50% of the
replicates that will become toxic to show turbidity and/or gas production). Standard errors are given in parentheses. NF, not t.
cRange of days when growth was rst and last observed in replicate tubes.
dData for days 1 to 190.
eData for days 191 to 365.
dence of growth. Supernatantsfrom all subcultured tubes showing
evidence of growth were subsequently assayed for toxin.
Detection of toxin. Botulinum neurotoxin was detected by
the mouse bioassay. Vaspar (used in the second study only) was
removed with a sterile Pasteur pipet prior to the ltering of vor-
texed cultures through a 0.22-mm lte r (Gelman Scien ti c). A
portio n (0.5 ml) o f ltr ate was inject ed intrap eritone ally into each
of two 20- to 28-g mice (Charles River, Quebec, Canada). Mice
were observed for up to 72 h for typical signs of botulism, in-
cludin g ruf e d fu r, pinch ed waist, labored b reathing, l imb pa resis,
and general paralysis. Mice showing severe distress were euthan-
atized by asphyxiation with CO2according to Health Canada An-
imal Care Committee gu idelines . Toxin n eutraliz ation was carried
out fo r rando mly selecte d rep resentat ive posit ive sa mples by usi ng
antisera (Connaught Laboratories, North York, Ontario, Canada)
to botulinum neurotoxinsas previously described(6, 7) to con rm
that toxicity was due to botulinum neurotoxin.
RESULTS AND DISCUSSION
Effect of ethanol on the growth of C. botulinum. The
growth pro les of the composite mixture of six proteolytic
strains of C. botulinum grown in TPGY broth containing 0
to 4% (wt/wt) 95% ethanol as monitored by viable cell
density and OD60 0 are shown in Figure 1A and 1B, re-
spectively. It is evident from Figure 1A and 1B that there
was no linear relationship between the log of viable cell
counts and OD60 0. It has been shown that a limitation of
the monitoring of growth by optical density is that cell
numbers cannot be directly determined from optical den-
sity, since the number of cells is generally proportional to
Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/4/610/1675368/0362-028x-66_4_610.pdf by guest on 12 January 2023
J. Foo d Prot., Vol. 66, No. 4 EFFECT OF ETHANOL ON C. BOTULINUM 613
FIGURE 1. Effects of (m) 0, (v) 1, ( ) 2, (l) 3, and (.) 4%
(wt/wt) ethanol on growth as measured by (A) viable cell density
(CFU/ml) and (B) OD600 of a composite proteolytic C. botulinum
mixture in TPGY broth at 378C.
optical density over only a very narrow range. Accordingly,
the estimation of microbial growth parameters is best
achieved with the use of viable cell counts (2, 8). Therefore,
viable-count growth curves (Fig. 1A) were used to estimate
growth parameters. The growth of C. botulinum was rapid
in control asks of TPGY broth (0% ethanol). However,
growth (as measured by the log of viable cell counts) was
inhibited as the concentration of ethanol increased.
Increasing levels of ethanol resulted in a concentration-
dependent extension of the lag phase when C. botulinum
was grown in TPGY broth (Fig. 2B). The end of lag phase
growth is represented as the maxima of the curves shown
in Figure 2B. While the end of the lag phase occurred after
4.6 h in control broths (0% ethanol), it was delayed to 5.2,
5.3, 9, and 9.5 h, respectively, in broths containing 1, 2, 3,
and 4% ethanol (Fig. 2B).
Although ethanol (at 1 to 4% [wt/wt]) delayed the lag
phase, the growth of C. botulinum was not prevented.
Growth, once initiated, was rapid; however, the rate of
growth (]CFU/ml/]t) during the logarithmic phase was in-
uenced by the concentration of ethanol in the medium.
The times at which the maximum C. botulinum growth
rates occurred in TPGY containing ethanol (at 0 to 4% [wt/
wt]) are shown in Figure 2A. These times ranged from 6.4
h for cultures grown without ethanol (0%) to 14.9 h for
cultures grown in 4% (wt/wt) ethanol. Linear regression (r2
50.879) and extrapolation of these data predicted that a
level of 5.5% (wt/wt) ethanol would result in complete in-
hibition of the growth of C. botulinum in TPGY broth at
378C (Fig. 3).
The maximum growth reached at the stationary phase
was also decreased in the presence of ethanol. The maxi-
mum log of the viable cell count for C. botulinum grown
in TPGY broth without ethanol was approximately 8.3,
compared with 7.6 to 7.3 for cultures grown in broth con-
taining 1 to 4% (wt/wt) ethanol (Fig. 1A) (i.e., ethanol re-
sulted in a ,1-log reduction in counts at the stationary
phase regardless of the concentration of ethanol).
In summar y, the effects of ethanol on the growth of
the strains of C. botulinum used in this study were (i) an
increase in the lag phase, (ii) a decrease in the rate of ex-
ponential growth, and (iii) a ;1-log reduction in counts at
the stationary phase. The effects of ethanol on the growth
of C. botulinum observed in the present study are consistent
with those found by Yammamoto et al. (25), who reported
that ethanol exerted a concentration-dependent inhibitionof
gram-positive food spoilage and pathogenic bacteria.
Toxin production. While the growth of C. botulinum
was inhibited by 1 to 4% (wt/wt) ethanol, toxin was still
detected in all cultures after 24 h of growth at 378C re-
gardless of the ethanol concentration. Although it is pos-
sible that culture supernatants of TPGY with higher levels
of ethanol contained less toxin after 24 h of growth, this
possibility was not tested.
Combined effects of ethanol, aw, and pH. On the
basis of the extrapolation of the reduction of the C. botu-
linum growth rate by ethanol in the initial growth study,
5.5% (wt/wt) ethanol was predicted to be the level that
would completely inhibit the growth of the composite C.
botulinum mixture of in TPGY broth at 378C (optimal
growth conditions). However, this level exceeds the maxi-
mum level (2%, wt/wt) permitted for use to extend the shelf
lives of certain bakery products, speci cally pizza crusts.
Therefore, a second study was designed to determine the
combined effect of ethanol, aw, and pH on the probability
of the growth of and toxin production by a composite mix-
ture of spores of the same six strains of proteolytic C. bot-
ulinum used in the initial growth study. Tubes of TPGY
broth were visually examined for turbidity and gas produc-
tion twice daily. For all broths containing 0% ethanol (con-
ditions 1 through 10, Table 1), growth and toxin production
occurred in 1 to 3 days regardless of the awand pH values
(P,0.005). Since growth was observed simultaneously in
all replicate tubes of TPGY broths without ethanol, it was
impossible to t the parameters of the logistic distribution
(Pt) to the data for 0% ethanol. However, since 100% of
the tubes were positive within 3 days, the maximum prob-
ability of growth and toxin production for conditions in-
volving 0% ethanol at $3 days (P3) was 1.
In broths containing 2% ethanol, growth and toxin pro-
duction occurred in 2 to 4 days (conditions 11 through 17,
Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/4/610/1675368/0362-028x-66_4_610.pdf by guest on 12 January 2023
J. Food Prot., Vol. 66, No. 4614 DAIFAS ET AL.
FIGURE 2. First (A) and second (B) derivative plots of growth curves (shown in Fig. 1). Maximum absolute growth rates (maxima of
curves in panel A) and the end of the lag phases (maxima of curves in panel B) for a composite C. botulinum mixture grown in 0%
(solid line), 1% (small dashes), 2% (large dashes), 3% (medium dashes), and 4% (dotted line) (wt/wt) ethanol.
FIGURE 3. Ethanol le vel (%, wt/wt ) (inter cept of xaxis) predicted
to completely inhibit the growth of a proteolytic C. botulinum
mixture in TPGY broth at 378C.
Table 1), again regardless of the awand pH values (P,
0.005). Since growth and toxin production were observed
simultaneously in all replicate tubes of TPGY broths con-
taining 2% ethanol, it was impossible to t the parameters
of the logistic distribution to data for this level of ethanol.
Nevertheless, P451.
All combinations of 4% (wt/wt) ethanol, aw, and pH
(conditions 18 through 24, Table 1) also showed growth
and toxin production. However, the rates at which growth
occurred in replicate tubes of broths containing 4% (wt/wt)
ethanol was in uenced by the awand/or pH of the medium.
Complete inhibition of growth was observed at ethanol
concentrations of 6 and 8%. Furthermore, no toxin was de-
tected for any combination of factors (ethanol, aw, pH) at
these ethanol levels (conditions 25 through 30, Table 1)
after 1 year (365 days) at 258C. These results are consistent
with those of the preliminary growth study resulting in the
prediction of the inhibition of C. botulinum by 5.5% etha-
nol for overnight cultures grown in TPGY at 378C. Yam-
mamoto et al. (25) reported the inhibition of gram-positive
food spoilage and pathogenic bacteria with 9 to 11% eth-
anol, while Cook and Pierson (5) reported the inhibition of
C. botulinum type A with 10% ethanol.
The C. botulinum growth inhibition observed in this
study was concentration dependent. For example, in broths
with an awof ;0.954, a pH of 7.7, and 0, 2, 4, or 6% (wt/
wt) ethanol, growth was rst observed on 3, 3, 49, and
.365 days into the study, respectively (Table 1). The ad-
dition of ethanol to a liquid medium results in the depres-
sion of its aw. TPGY broths containing 6 and 8% (wt/wt)
ethanol would have estimated awvalues of 0.971 and 0.964,
respectively, according to the Ross equation (15). However,
since growth was observed within 3 days in TPGY broths
containing 0% ethanol and adjusted to an awof ,0.964
(Table 1), the inhibitory effect of ethanol on C. botulinum
is attributable to more than simply a depression of aw. Sim-
ilar conclusions regarding the antimicrobial effects of eth-
anol have been reached by other authors (17, 21).
As previously described, the rate of the growth of and
toxin production by C. botulinum in replicate tubes of
TPGY broths containing 4% ethanol (conditions 18 through
24, Table 1) was in uenced by the awand pH of the me-
dium. These data were tted to a mathematical model that
Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/4/610/1675368/0362-028x-66_4_610.pdf by guest on 12 January 2023
J. Foo d Prot., Vol. 66, No. 4 EFFECT OF ETHANOL ON C. BOTULINUM 615
FIGURE 4. Proportions of replicate tubes
showing growth and toxin production by
103C. botulinum spores per ml for select-
ed co mbination s of 4% ( wt/wt) e thanol, aw,
and pH ( , aw0.965, pH 6.2; V, aw0.962,
pH 8.2; m, aw0.959, pH 7.2; l, aw0.956,
pH 7.7) in TPGY broth at 258C.
describes the rate at which replicate tubes become toxic as
a function of time. The ts of the model to each combi-
nation of 4% ethanol, aw, and pH, along with the standard
errors for each parameter, are shown in Table 1. Such mod-
els, which estimate the probability of growth and toxin pro-
duction for a given set of conditions, are generally consid-
ered valid for the evaluation of the botulism hazard asso-
ciated with the growth of C. botulinum in food. There is
evidence that considerable growth is required before mea-
surable toxin is produced and that most toxin is produced
after cell lysis (18, 19). Daifas et al. (6) observed that the
time at which toxin was rst detected in crumpets inocu-
lated with 5 3102spores of proteolytic C. botulinum per
g and stored at 258C corresponded to counts of approxi-
mately 105CFU/g. However, the growth of this pathogen,
once initiated, is rapid and is always associated with a neu-
rotoxin of extreme potency. Therefore, any growth of C.
botulinum is considered unacceptable (23).
Selected conditions involving 4% ethanol that inhibited
the growth of C. botulinum are shown in Figure 4. Describ-
ing the effect of selected levels of barriers such as ethanol,
aw, and pH on the growth of C. botulinum is a practical
approach to food safety, since the safety of bakery products
with extended shelf lives can be predicted with this ap-
proach. Furthermore, the results of this study indicate that
while some combinations of barriers could be used to en-
hance the safety of a bakery product with a short shelf life,
other combinations would be more effective for products
with longer shelf lives. For example, for an ethanol con-
centration of 4%, an awof 0.959, and a pH of 7.2 (condition
19, Table 1), the probability of toxin production by day 60
is 0.003. However, the probability of growth and toxin pro-
duction increases rapidly with time. By day 97, 50% of all
replicates were toxic (P9 7 50.5), and by day 125, all rep-
licates were toxic (Pmax 51) (Fig. 4). Therefore, this com-
bination of barriers may be appropriate for the enhancement
of the safety of a bakery product with a shelf life of ,30
days. However, although safety may be enhanced by this
combination, sensor y quality may be adversely affected.
Foods with intermediate water activity packaged with eth-
anol vapor have been shown to absorb less ethanol than
foods with high water activity packaged similarly (16, 21).
Therefore, packaging with this level of ethanol may be
more suitable for intermediate-moisture bakery products.
Furthermore, some studies have shown that alcohol aromas
may dissipate once packages are opened and/or products
are heated (16). Although many bakery products are con-
sumed without heating, par-baked breads and rolls and
crumpets are examples of products that are further heated
prior to consumption and may be appropriate for this type
of packaging.
The tvalue for broths containing 4% ethanol ranged
from ,3 days for broth with an awof 0.981 and a pH of
7.7 to 229 days for broth with an awof 0.962 and a pH of
8.2. With the exceptio n of one combination of b arriers (con-
dition 18, Table 1), the Pm ax values for all broths containing
4% ethanol were ca. 1. For the most inhibitory combination
of barriers (4% ethanol, an awof 0.956, and a pH of 7.7;
condition 18, Table 1), Pmax 50.22 and t 5 207 days (Fig.
4).
Data for only one combination of barriers (4% ethanol,
an awof 0.962, and a pH of 8.2; condition 20, Table 1) did
not t the probability function used. The pattern of growth
observed for these barrier levels was different from all other
growth patterns. At the onset of the study, several replicate
tubes for this condition became positive, and then no fur-
ther growth was observed (Fig. 4). Data for this condition
(condition 20, Table 1) tted the model well for the rst
200 days of growth. Indeed, if this study had been termi-
nated before 200 days, the Pm ax for this condition would
have been assumed to be 0.25 (Fig. 4). However, after 200
days, tubes again showed evidence of growth, and this trend
continued until all tubes were positive (by day 263). Al-
though the pattern of overall growth observed for these
barrier levels did not t the model well, there was a good
t for data for both periods of growth (Fig. 4). While it is
possible that this combination of barriers inhibited the ger-
mination or outgrowth of spores of C. botulinum initially,
Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/4/610/1675368/0362-028x-66_4_610.pdf by guest on 12 January 2023
J. Food Prot., Vol. 66, No. 4616 DAIFAS ET AL.
FIGURE 5. Effects of (a) 0% (bar represents 0.5 mm), (b) 2%
(bar represents 1 mm), and (c and d) 4% (bars represent 2 and
1mm, respectively) (wt/wt) ethanol on vegetative cells of C. bot-
ulinum in TPGY broth at 378C.
this inhibition was overcome in time, resulting in visible
signs of growth and toxicity for more tubes. McMeekin et
al. (13) proposed a point (the growth/no growth interface)
at which interacting hurdles restrict growth and emphasized
the importance of considering the physiological mecha-
nisms that occur in pathogens at this interface. It is possible
that this combination of barriers (4% ethanol, an awof
0.962, and a pH of 8.2) was close to this growth/no growth
interface.
Effects of ethanol on cells. These two studies con-
rmed the inhibitory effect of ethanol on C. botulinum. In
order to in vestigate t he effec t of ethanol on veg etative cells,
overnight cultures of C. botulinum 62A were grown in
TPGY broth containing 0, 2, 4, 5, and 6% ethanol at 378C.
No growth was visible in cultures grown in 5 or 6% etha-
nol, again con rming the predicted inhibitory ethanol level
of 5.5%. Micrographs of these cultures grown in TPGY
broth with 0, 2, and 4% ethanol are shown in Figure 5a
through 5d. Growth and cell division are evident in cultures
grown with 0% ethanol (Fig. 5a). The effect of 2 and 4%
ethanol on cells is shown in Figure 5b through 5d. The
interference of growth, as demonstrated by cell elongation
and septation at an ethanol concentration of 2% (Fig. 5b),
is even more pronounced at an ethanol concentration of 4%
(Fig. 5d). Reversible interference with cell division and the
elongation of Escherichia coli cells grown with ethanol
have been reported by Fried and Novick (10), although
these authors also isolated mutants for which ethanol stim-
ulated cell division. Ethanol has been shown to act primar-
ily on the cell membrane, either directly on the membrane
and/or membrane-associated enzymes or indirectly through
the impairment of biosynthesis of membrane components
(16). Empty, ‘‘ghostlike’’ cells, i.e., cells that were col-
lapsed and devoid of intracellular contents, were observed
in cultures grown with either 2 or 4% ethanol but not in
cultures grown with 0% ethanol. This nding may be due
to the leakage of solutes across the membrane or to cell
lysis following decreased pe ptidogly can cross-linki ng in th e
growing cell wall (11).
Nature of the effect of ethanol. In the study on the
combined effect of ethanol, aw, and pH, some tubes of
TPGY broths (including all broths containing 6 or 8% eth-
anol [conditions 25 through 30, Table 1] and 20 of 25 rep-
licates containing 4% ethanol [condition 18, Table 1])
showed no evidence of growth (turbidity and/or gas pro-
duction) even after 1 year at 258C. To determine whether
ethanol had a lethal or a static effect at these levels, all
tubes of TPGY broths showing neither turbidity nor gas
production after 1 year were subcultured into fresh TPGY
broth and incubated anaerobically at 378C for 24 h. All of
these subcultured tubes showed turbidity and gas produc-
tion within 24 h and subsequently tested positive for toxin
by the mouse bioassay. Although it is not clear how the
growth of inoculated spores was inhibited for 365 days at
258C, the inhibitory effect of ethanol was clearly reversible.
Reversible inhibition of the germination of spores of Ba-
cillus subtilis var. niger and Bacillus pumilus by low levels
of ethanol has been reported by Trujillo and Laible (22),
who attributed this reversible inhibition to an enzymatic
mechanism. Possible mechanisms involve the direct inhi-
bition of lytic enzymes, which degrade peptidoglycan in the
spore cortex, by ethanol or hydrophobical interaction be-
tween ethanol and the spore coat structure that in turn af-
fects germination enzymes. Chaibi et al. (4) reported that
ethanol at concentrations of #3% did not inhibit the spore
germination, outgrowth, or cell multiplication of C. botu-
linum 62A. Cook and Pierson (5) reported the inhibition of
the germination of C. botulinum type A spores by 10%
ethanol, but it is not known whether this inhibition was
reversible.
In conclusion, these studies have con rmed that etha-
nol could be used to delay the growth of C. botulinum.
Furthermore, modeling studies have shown that the proba-
bility of growth and toxin production is in uenced by the
concentration of ethanol in the medium involved and the
awand pH of the medium. However, model broth studies
have certain limitations with respect to the safety of bakery
products: (i) they involve pure cultures of C. botulinum
spores and do not consider the background micro ora of
products, and (ii) they do not consider the sensory aspect
of a product to determine whether toxigenesis preceded
spoilage or vice versa. While caution must be exercised in
extrapolating results from model broth systems to food
products, these studies have nevertheless shown that etha-
nol has the potential to be a viable part of a multibarrier
food safety system for bakery products. However, a limi-
tation of ethanol in liquid form is that it would necessitate
the spraying of high levels ($4%, wt/wt) of ethanol onto
Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/4/610/1675368/0362-028x-66_4_610.pdf by guest on 12 January 2023
J. Foo d Prot., Vol. 66, No. 4 EFFECT OF ETHANOL ON C. BOTULINUM 617
the product surface to prevent the growth of and toxin pro-
duction by C. botulinum. To overcome this limitation, the
inclusion of ethanol vapor–generating sachets in a product’s
package might reduce the amount of ethanol required, since
ethanol has been shown to be most effective in the vapor
state (21). Therefore, the application of lower levels of eth-
anol in the vapor phase in conjunction with awand/or pH
reduction may result in inhibition similar to or more exten-
sive than that achieved through the direct application of
liquid ethanol to the surfaces of bakery products.
ACKNOWLEDGMENT
The au thors acknowledge the Natural Sciences and Engineering Re-
search Council for it s nan cial support o f th is study.
REFERENCES
1. Austin, J. W., M. Stewart, and R. G. Murray. 1990. S tructural and
chemical characterization of the S layer o f a Pseudomo nas-like bac-
terium. J. Bact eriol. 172:808–817.
2. Begot, C., I. Desnier, J. D. Daudin, J. C. Labadie, and A. Leb ert.
1996. Recommend ations fo r calculating growth parameters b y opti-
cal density measur ements. J. Microbiol. Methods 25:225–232.
3. Buchanan, R. L., and M. L. Cygnarowicz. 1990. A math ematical
approach to de ning and calculating the du ration of the lag phase.
Food Micriobiol. 7:237–240.
4. Chaibi, A., L. H. Ababouch, K. Belasri, S. Boucetta, and F. F. Busta.
1997. In hibition of germination an d veg etative g rowth of Bacillus
cereus T and Clostridium botu linum 62A spores by essential oi ls.
Food Microbiol. 14:1 61–174.
5. Cook, F. K., and M. D. Pierson. 1 983. In hibition of bacterial spores
by antimicro bials. Food Technol. 18:115–126.
6. Daifas, D. P., J. P. Smith, B. Blanch eld, and J. W. Austin . 1999.
Growth and toxin produ ction by Clostridium bo tulinum in English
style crumpets packaged under mod i ed atmosph eres. J. Food Prot.
62:349 –355.
7. Daifas, D. P., J. P. Smith, B. Blanch eld, and J. W. Austin . 2000.
Effect of ethanol vapor on growth and toxin pro duction by Clostrid-
ium botulinum in a high moisture bakery product. J. Food S ci. 20:
113–127.
8. Dalgaard, P., and K. Koutsoumanis. 2001. Comparison of maximum
speci c growth rates and lag times estimated from absorban ce and
viable coun t data by different math ematical models. J. Microbiol.
Methods 43:183–196.
9. Dodds, K. L., and J. W. Austin . 1997. Clostridium botulinum, p.
288–304. In M . P. Doyle, L. R. Beuchat, and T. J. Montville (ed.),
Food microbiology. Fundamentals and frontiers. ASM Press, Wash-
ington, D. C.
10. Fried, V. A., and A. Novick. 1 973. Organic solvents as prob es for
the structure and function of the bacterial membrane: effects of eth-
anol on the wil d type of an et hanol-resistant mutant of Escherichia
coli K-12. J. Bacteriol. 114:239–248.
11. Ingram, L . O., and T. M. Buttke. 1984. E ffects of alcohols on micro-
organisms. Adv. Microbiol. Physiol. 25:253–300.
12. Leistner, L. 2 000. Basic aspects of food preservation by h urdle tech-
nology. Int . J. F ood Micro biol. 55:181–186.
13. McMeekin, T. A., K. Presser, D. Ratkowsky, T. Ross, M. Salter, and
S. Tienu ngoon. 2000. Quantifying th e hu rdle concept by modelling
the bacterial gro wth/no growth interface. Int. J. Food Microbiol. 55:
93–98.
14. Reynolds, E. S. 1963. The use of lead citrate as high pH as an
electron-opaqu e stain in electron microscop y. J. Cell Biol. 17:2 08–
212.
15. Ross, K. D. 1975. Estimation of water activity i n intermed iate mois-
ture foods. Food Technol . 29:26–34.
16. Seiler, D. A. L., and N. J. Russell. 1991. Ethanol as a food preser-
vative, p. 15 3–171. In N. J. Russell and G. W. Gould ( ed.), Food
preservativ es. Blackie, London.
17. Shapero, M., D. A. Nelso n, and T. P. Labuza. 197 8. Ethanol inhi-
bition of Staphylococcu s aureus at li mited water activ ity. J. Food
Sci. 43:14 67–1469.
18. Siegel, L. S., and J. F. Metzger. 1979. Toxin production by Clostrid-
ium botulinum type A under various fermentation conditions. Appl.
Environ. Microbiol. 38:606–611.
19. Siegel, L. S., and J. F. Metzger. 1980. Effect of fermentation con-
ditions o n toxin prod uction by Clostrid ium botulinum t ype B. App l.
Environ. Microbiol. 40:1023–1026.
20. Smith, J. P. 1992. Bakery products, p. 134–169. In R. T. Parry (ed.),
Principles and applications of modi ed atmosphere packaging of
food. Blackie Academic and Professional, London.
21. Smith, J. P., B. Ooraikul, W. J. Koersen, F. R. van de Voort, E. D.
Jackson, and R. A. Lawren ce. 1987. Shelf life extension of a bakery
product u sing ethanol vap or. Food Microbiol. 4:329–337.
22. Trujillo, R., and N. Laible. 1970. Reversible inhibition of spore ger-
mination by alcohols. Appl. Microbiol. 20:620–62 3.
23. Whiting, R. C., and J. E. Call. 1993. Time o f growth model for
proteolyti c Clostridium botulinum. Food Microbiol. 10:295–301.
24. Whiting, R. C., and J. C. Oriente. 1997. Time-to-turbidity model for
non-pro teolytic type B Clostridium botulinum. Int. J. Food Micro-
biol. 35:4 9–60.
25. Yammamoto, Y., K. Higashi, and H. Yoshi. 19 84. Inhib itory activity
of ethanol on food spoilage bacteria. (Studies on g rowth inhibition
of food spo ilage microorgani sms for low salt foo ds. Part II). Nippo n
Shokuhin Ko gyo Gakkaishi 31:531–535.
Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/4/610/1675368/0362-028x-66_4_610.pdf by guest on 12 January 2023