RESEARCH LETTER

Plasmid-associated bacteriocin production by Lactobacillus

LMG21688 suppresses Listeria monocytogenes growth rebound
in a food system
Privat Kouakou1, Carine Dortu1, Robin Dubois-Dauphin1, Micheline Vandenbol2 & Philippe Thonart1
1
Faculté Universitaire des Sciences Agronomiques de Gembloux, Gembloux, Belgium; and 2Unité de biologie animale et microbienne,
Gembloux, Belgium

Downloaded from https://academic.oup.com/femsle/article-abstract/306/1/37/479127 by guest on 30 August 2019

Correspondence: Privat Kouakou, Faculté Abstract
Universitaire des Sciences Agronomiques de
Gembloux, 2 Passage des Déportés, B-5030
Bacteriocin produced by Lactobacillus curvatus CWBI-B28wt is not completely
Gembloux, Belgium. Tel.: 1328 162 2305; effective against Listeria monocytogenes in food models. There is evidence suggest-
fax: 1328 161 4222; e-mail: ing that bacteriocin-degrading proteolytic enzymes produced by the CWBI-B28wt
bioindus@fsagx.ac.be strain and/or present in the food matrix contribute to this rebound of Listeria
growth. To limit this problem, we have partially characterized an approximately
Received 12 December 2009; revised 1 10-kb plasmid responsible for bacteriocin production in L. curvatus CWBI-B28wt.
February 2010; accepted 8 February 2010. This plasmid was transferred by high-voltage electroporation into a less proteoly-
Final version published online 8 March 2010.
tic, but technologically competent Lactobacillus strain. When the transformed
strain was used as a starter culture in a model food system, a high bacteriocin level
DOI:10.1111/j.1574-6968.2010.01932.x
was maintained for a longer time than with CWBI-B28wt, and Listeria growth
Editor: Wolfgang Kneifel
rebound was delayed by 2 weeks (occurring after 3 weeks of apparently total
inhibition, instead of one).
Keywords
plasmid curing; PCR amplification;
electrotransformation; Lactobacillus curvatus;
MICROBIOLOGY LETTERS

bacteriocins; Listeria monocytogenes.

Introduction
Lactic acid bacteria (LAB) are important industrially, mainly
in food fermentation processes (Gilliland, 1985; Chassy,
1987; McKay & Baldwin, 1990). In addition to causing rapid
acidification of the raw material through the production of
organic acids (mainly lactic acid), they produce a number
of compounds, such as acetic acid, ethanol, aroma com-
pounds, bacteriocins, exopolysaccharides, and enzymes, that
increase the shelf-life and microbial safety of the end
product, improve its texture, or contribute to a pleasant
sensory profile. Direct addition of selected starter cultures to
raw materials has been a breakthrough in the processing of
fermented foods, allowing end-product standardization and
a high degree of control over the fermentation process
(Oberman & Libudzisz, 1998).
Among the metabolites synthesized by LAB, bacteriocins
are important for their antibacterial action. These riboso-
mally synthesized proteinaceous compounds typically inhi-
bit the growth of strains closely related to the producer
strain (Tagg et al., 1976), but they can also affect more
distantly related species such as Listeria monocytogenes, a
foodborne pathogen that has received considerable atten-
tion (Klaenhammer, 1988). Bacteriocin, however, is sensitive
to proteolytic enzymes present in the food matrix and/or
synthesized by the producer strain (Schillinger et al., 1991;
Kouakou et al., 2008). Evidence suggests that the proteolytic
degradation of bacteriocin may contribute to the ‘rebound’
of listerial growth observed after its initial inhibition in
bacteriocin-containing systems (Kouakou et al., 2008). The
present work was an attempt to limit this problem. Our
initial focus was on Lactobacillus curvatus CWBI-B28wt
(henceforth called wt), a strain isolated by Benkerroum
et al. (2002) and known to produce a bacteriocin, probably
from a plasmid-borne gene. Dortu et al. (2008) have shown
that this bacteriocin is a sakacin P, and Kouakou et al. (2008)
have demonstrated the (limited) antilisterial action of wt
added to a model meat system. Here, the aim was to confirm
the plasmid location of this strain’s sakacin P gene and, if
successful, to transfer the bacteriocin-encoding plasmid into

FEMS Microbiol Lett 306 (2010) 37–44 

c 2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
38
a nonbacteriocinogenic, but technologically competent Lac-
tobacillus strain with low proteolytic activity. The transfer
method chosen was high-voltage electroporation, used
successfully on various Lactobacillus species (e.g. Chassy &
Flickinger, 1987; Badii et al., 1989; Josson et al., 1989). Our
work has led to the creation of a strain whose ability to
maintain a high level of bacteriocin for a prolonged period
in a model food system delays Listeria growth rebound.
Materials and methods
Bacterial strains, strain storage conditions, and
broth/plate cultures
Lactobacillus curvatus CWBI-B28wt (wt), described by Ben-
kerroum et al. (2002), is an antilisterial bacteriocin-produ-
cing strain. Lactobacillus curvatus LMG 21688 (Diop et al.,
2008) is a bacteriocin-nonproducing strain (noted LMG
hereafter) purchased from Gent University and included in
the THT company’s laboratory culture collection (Gem-
bloux, Belgium). Listeria monocytogenes M, originally iso-
lated from bacon, was obtained from the collection of the
Centre Wallon des Bio-Industries (Gembloux, Belgium). It
is sensitive to the bacteriocin produced by wt and was used
as an indicator and to artificially contaminate meat samples.
It was spread regularly over Palcam agar (Oxoid, Beauvais,
France) plates and activated in tryptone soy broth (Biokar,
Beauvais, France) at the time of its use.
Strains mt (obtained by curing wt of its plasmids) and
LMGel (obtained by electroporation of LMG with a wt-
derived plasmid) are described in the present work. All
strains were grown in the meat system described below (see
Meat system and meat sampling) or on DeMan, Rogosa, and
Sharpe medium (MRS, Biokar) (broth or with 1.5% agar, as
specified). To avoid plasmid loss by the LMGel strain, MRS
medium was rendered selective for plasmid-containing cells
(see Results) by addition of streptomycin (50 mg mL1) or by
replacing 2% glucose with either 2% D-celobiose, 2%
gentiobiose, or 1% of each of these sugars. The correspond-
ing media are henceforth, respectively, called MRSStr, MRSC,
MRSG, and MRSCG.
All strains were stored at  80 1C in their respective
media with added 40% glycerol (v/v). Once the antibiotic
sensitivity profile conferred by the identified plasmid was
obtained (see Results), selection for its presence was carried
out on a medium containing streptomycin (50 mg mL1).
Meat system and meat sampling
The model food system used was as described by Kouakou
et al. (2009), except that the meat was first rubbed with
D-celobiose and gentiobiose (each at 1%) to favour plasmid
stability in LMGel. Then, briefly, 50-g blocks of raw pork
meat (listed characteristics: 60% moisture; 15% protein;

c 2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
P. Kouakou et al.
13% fat; 5% minerals; and 7% carbohydrates, pH 5.65,
at 24 h) were transferred to sterile Stomacher bags, homo-
genized in deionized water, transferred to sterile bottles,
and coinoculated with L. monocytogenes and the specified
L. curvatus strain (103 CFU of each bacterium g1). A con-
trol with only L. monocytogenes (initial concentration:
103 CFU g1) was included. The treated homogenates were
then incubated for 4 weeks at 4 1C. Meat samples (20 g
crushed meat) were taken at the start of the experiment (day
0, before storage) and on days 7, 14, and 28. Samples were
diluted with 10 mL of sterile saline (0.85% NaCl) and
Downloaded from https://academic.oup.com/femsle/article-abstract/306/1/37/479127 by guest on 30 August 2019
homogenized in a Stomacher bag.
Plasmid curing procedure
The method used to cure wt of its plasmid(s) combined
heating, as described by Sonstein & Baldwin (1972), with
sodium dodecyl sulphate (SDS) treatment according to
Collins & Harvey (1962). A single colony of wt was picked
from an MRS agar plate, inoculated into 5.0 mL MRS broth,
and grown overnight at 37 1C. Then 5.0 mL fresh medium
containing SDS (1%) was seeded with 0.1 mL of culture and
incubated overnight at 42 1C. This culture was centrifuged at
4424 g for 5 min. The pellet was resuspended in 5.0 mL
saline, diluted 105-fold, spread on MRS agar plates, and
incubated overnight at 37 1C. The cells were then overlaid
with 3 mL soft (0.75%) agar medium inoculated with 30 mL
culture of the indicator strain (at c. l06 CFU mL1). The
plates were then incubated again overnight at 37 1C and
colonies with no clear zone surrounding them were ran-
domly selected, isolated, purified, and tested for the pre-
sence of plasmids.
DNA isolation and plasmid detection
Total DNA was extracted using the Wizard Genomic DNA
Purification Kit (Promega, Madison) and plasmid DNA
using the ‘Qiagen plasmid kit’ (Diagen, Dusseldorf, Ger-
many) according to the manufacturers’ instructions (the
latter preparation is henceforth called ‘crude plasmid
DNA’). When specified, a single plasmid band was excised
and cleaned according to instructions of the Silica Bead
DNA gel extraction kit (Fermentas, Dusseldorf, Germany)
and used for further study (such preparations are henceforth
called ‘gel-purified plasmid DNA’).
Locating the sakacin P gene (SppA) of the wt
strain
First, a digoxigenin (DIG)-labelled probe corresponding to
part of the SppA gene of strain wt was generated using the
PCR DIG Probe Synthesis kit (Roche, Basel, Switzerland)
according to the manufacturer’s instructions. Total wt DNA
was used as a template and the primers were SakPfw
FEMS Microbiol Lett 306 (2010) 37–44
Suppression of Listeria growth rebound in a food system
(5 0 -GAA (T/A)T(A/G)(C/A)(C/A)A NCA ATT A(C/T)
(A/C) GGT GG-3 0 ) and SakPrev (5 0 -GGC CCA GTT TGC
AGC TGC AT-3 0 ), based on the SppA sequence deposited in
the GenBank database.
Southern blotting was then performed on restriction
fragments of gel-purified plasmid from wt (the products of
separate HindIII, CfoI, and EcoRI digestions were run in
parallel). Restriction mixtures were electrophoresed through
a 0.8% agarose gel, along with the 500-bp marker (Biorad)
and the Big Dye Marker (Roche, Penzberg, Germany). The
resulting bands were blotted onto a Hybond N1nylon
membrane and allowed to hybridize with the DIG-labelled
probe for 20 h at 65 1C. Chemiluminescence detection was
performed using the DIG-DNA kit (Boehringer, Mannheim,
Germany) according to the manufacturer’s instructions.
The plasmid location of the SppA gene was confirmed by
subjecting gel-purified plasmid DNA to PCR with Taq DNA
polymerase (Applied Biosystems, Milan, Italy). The reaction
mixture (final volume: 25 mL, placed in a 0.2-mL Eppendorf
tube) contained 10  Taq Buffer, 1.5 mM MgCl2, 0.2 mM
dNTP, each primer at 0.5 mM, 5 U mL1 Taq DNA polymer-
ase (Applied Biosystems), and 25 mg mL1 DNA in sterile
milli-Q water. Amplification was carried out in a Mastercy-
cler Personal thermocycler (Eppendorf, Pecq, France). The
heating/cooling program was as follows: a first cycle at 94 1C
for 2 min, 55 1C for 1 min, 72 1C for 1 min, followed by 32
cycles of 94 1C for 1 min, 50 1C for 45 s, and 72 1C for 1 min.
The target amplicon was detected with the DIG-labelled
probe.
Electrotransformation of the LMG strain
To electroporate LMG with the wt-derived plasmid, the
method described by Kim et al. (1992) was used with a
slight modification in the pulse (electric field strength 2 kV;
capacitance 25 mF; resistance 800 O).
Bacteriocin detection: bacteriocin and
proteolytic activity assays
Bacteriocin was detected qualitatively as follows: the super-
natant of a wt, mt, LMG, or LMGel culture was adjusted to
pH 6.5 with 10 N NaOH and heated at 62 1C for 30 min. A 5-
mL portion was spotted onto the surface of an agar plate
containing 102 CFU freshly prepared L. monocytogenes M
and allowed to diffuse for 4 h. A clear zone around a spot
was indicative of the presence of bacteriocin.
Bacteriocin activity was assayed by means of the agar well
diffusion assay of Parente & Hill (1992), as further described
by Kouakou et al. (2008). It is expressed in arbitrary units
(AU) defined as the reciprocal of the highest dilution
showing a definite inhibition zone around the well.
Extracellular proteolytic activity was measured by the
spectrophotometric assay described by Church et al.
FEMS Microbiol Lett 306 (2010) 37–44
39
(1983), using O-phthaldialdehyde. The substrate used was
lyophilized cell-adsorbed bacteriocin prepared as described
by Kouakou et al. (2008). All results are means of duplicate
assays. Activities are expressed in U mL1 extract, 1 U being
defined as the activity corresponding to an absorbance
increase of 0.001 min1 in the assay.
Antibiotic resistance and carbohydrate
fermentation profiling
The wt, LMG, and LMGel strains were tested for resistance
Downloaded from https://academic.oup.com/femsle/article-abstract/306/1/37/479127 by guest on 30 August 2019
to chloramphenicol, ampicillin, streptomycin, vancomycin,
erythromycin, and tetracycline according to a slightly mod-
ified version of the macrodilution broth method developed
by Jones et al. (1985). Each antibiotic was tested individually
in the concentration range 4–1000 mg mL1. All experiments
were conducted three times, with determinations in tripli-
cate. The carbohydrate fermentation profiles of the wt, mt,
and LMG strains were determined using the API 50 CHL kit
(BioMérieux, Marcy-l’Etoile, France) according to the man-
ufacturer’s instructions.
Microbiological analysis and plasmid
stability test
Listeria monocytogenes CFU were counted in meat samples
after sample homogenization in peptone water as described
by Katla et al. (2001) and plating on Palcam agar. Plates were
incubated at 37 1C for 48–72 h.
The growth of L. curvatus strains in MRS or modified
MRS was monitored in 100-mL cultures inoculated with
106 CFU mL1 in 100 mL. At specific time intervals, 1-mL
samples were mixed with 9 mL peptone water. A decimal
dilution series was prepared from each sample and
L. curvatus CFU were counted after plating on MRS and
incubation at 37 1C for 24–48 h.
The stability of the wt-derived plasmid in strain LMGel
was tested by serially subculturing the cells in modified MRS
broth or nonselective MRS broth. The initial cultures were
seeded at 103 CFU mL1 from an overnight culture on
MRSStr. This was followed by six rounds of subculturing
(seeding at 103 CFU mL1, growth for 15 h, and storage at
4 1C for 9 h). After each 15-h growth period, aliquots were
diluted as above and plated on MRS agar, MRSStr agar,
MRSG agar, and MRSC agar. Colonies were counted after
incubation at 37 1C for 24–48 h.
Statistical analysis
Each trial was performed twice and each determination was
carried out in triplicate. Data are presented as means of two
independent experiments with SD. Statistical analysis (ANOVA
a = 0.05 and Student’s t-test) was performed using EXCEL
software.

c2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
40 P. Kouakou et al.

Results along with its plasmid, the ability to ferment D-celobiose,

gentiobiose, and N-acetylglucosamine. This suggests that
The sakacin P gene of the wt is plasmid-borne the plasmid present in wt bears determinants of the ability to
ferment these compounds.
Benkerroum et al. (2002) previously used ethidium bromide
treatment to cure the wt strain of any plasmids it might
contain. As they obtained bacteriocin-nonproducing
mutants in this way, they assumed, but did not confirm, Size (bp)
Mq L1 L2 L3 L4
a plasmid location of the strain’s bacteriocin gene. As
these mutants did not seem to be completely plasmid-cured
(data not shown), we first sought to use a more radical 10 000 bp

Downloaded from https://academic.oup.com/femsle/article-abstract/306/1/37/479127 by guest on 30 August 2019

10 000
curing procedure to obtain similar bacteriocin-nonprodu- 9500 bp
4000
cing mutants. As neither SDS treatment nor heat treatment
alone proved satisfactory, we combined the two. Several 2000
isolates showing no antilisterial action in the screening test 1000
were thus obtained. Figure 1 shows the plasmid profiles of 500
wt and one of the isolated mutants, mt (lane 1 vs. lane 4).
Whereas wt showed two plasmid bands near 10 kb, mt
appeared to be totally cured. This result confirms the
correlation between plasmid curing and loss of antilisterial
action. Fig. 1. Plasmid DNA profile on a 0.7% w/v agarose gel. Lane Mq, shows
Each of the plasmid bands revealed in wt was isolated the high-range Promega DNA ladder (molecular weight markers); lane 1,
by excision from the gel. Parallel restrictions were per- wt; lane 2, LMG; lane 3, LMGel; lane 4, mt, a completely cured derivative
formed with HindIII, CfoI, and EcoRI, and the resulting of the wt strain obtained after SDS treatment at 42 1C.
fragments were analysed by gel electrophoresis and Southern
blotting. In each restriction mixture, one fragment was 1 2 3 4 5 6
recognized by the sakacin-specific probe (Fig. 2). This Size (bp)
confirms the plasmid location of the wt strain’s SppA gene.
The material in each plasmid band displayed the same 23 130
restriction pattern and probe-binding profile, which sug- 9416
gests the presence of a single plasmid in two different coiling 6557
states.
4361

Further characterization of the SppA-bearing

plasmid 2322
2027
Before using the plasmid of wt to electrotransform
strain LMG, we examined whether it might bear selectable
or otherwise useful markers. Antibiotic resistance profiling
of the wt and mt strains was carried out with chloramphe- 564
nicol, ampicillin, streptomycin, vancomycin, erythromycin,
and tetracycline, chosen because the corresponding
resistance genes are frequently plasmid-borne (Axelsson
et al., 1988; Danielsen, 2002; Gevers et al., 2003; Gfeller
et al., 2003). One difference between the two strains was
observed: the MIC for streptomycin was 5 mg mL1 for
mt and above 50 mg mL1 for wt, suggesting the presence
of a plasmid-encoded determinant of streptomycin
resistance in wt. As LMG and mt showed similar pro- Fig. 2. Southern blot analysis of the restriction fragments obtained in
parallel digestions with HindIII (lanes 1 and 2), EcoR1 (lanes 3 and 4), and
files, streptomycin resistance was used later on as a
Cfo1 (lanes 5 and 6) of gel-purified plasmid from wt. The blot was
positive selection marker for bacteriocinogenic LMG probed with the DIG-labelled sakacin-specific probe. Lanes 1, 3, 5: lower
electrotransformants. plasmid band (see Fig. 1); Lanes 2, 4, 6: upper plasmid band. The lane on
The carbohydrate fermentation profiles of wt and mt the far left shows the DIG-labelled molecular weight markers [HindIII
were also compared. The mt strain was found to have lost, molecular size (bp) standard].


c 2010 Federation of European Microbiological Societies FEMS Microbiol Lett 306 (2010) 37–44
Published by Blackwell Publishing Ltd. All rights reserved
Suppression of Listeria growth rebound in a food system 41

Construction of a new strain with Plasmid stability in LMGel

plasmid-associated bacteriocin production
Our next step was to introduce the wt plasmid by electro-
poration into the nonbacteriocinogenic strain LMG.
Electrotransformants were selected for streptomycin resis-
tance. The number of streptomycin-resistant colonies
obtained was 4–7 mg1 DNA. Figure 1 confirms the absence
of the 10-kb plasmid in LMG and its presence in a
tested electrotransformant, LMGel (lane 2 vs. lane 3).
Figure 3 shows the results obtained when total DNA from
Table 1 shows the results of a subculturing experiment
performed to test the stability of the wt-derived plasmid in
LMGel (see Microbiological analysis and plasmid stability
test). In nonselective MRS broth, the percentage of cells
expressing plasmid-linked traits was found to decrease
rapidly. By the end of the second subculture (i.e. after about
21 generations, see Antilisterial effects in MRS broth and
in a meat system), only 1% of the cell population still
displayed streptomycin resistance (for the tested fermenta-

Downloaded from https://academic.oup.com/femsle/article-abstract/306/1/37/479127 by guest on 30 August 2019

LMG and gel-purified plasmid DNA from wt and LMGel
were subjected to PCR amplification with sakacin-specific
primers. The results, which were the same for the upper and
the lower plasmid bands, confirm the absence of a sakacin
P-encoding plasmid in LMG and its presence in both wt
and LMGel.
Size (bp)
3000
500
200
190 bp
100
Fig. 3. Agarose gel electrophoresis [1% agarose (w/v)] of PCR frag-
ments generated with primers SakP-F and SakP-R from gel-purified
plasmid DNA from wt and LMGel and total DNA from LMG. Lane Mq,
molecular weight markers; lane 1, wt; lane 2, LMG; lane 3, LMGel.
tion traits, the percentages were similar). The bacteriocin
activity measured at the end of a growth round was also
found to decrease from one round to the next, from
522 AU mL1 at the end of the first culture period to 0 at
the end of the third subculture (not shown). When the sole
carbon source was D-celobiose or gentiobiose (sugars whose
fermentation requires the presence of plasmid-borne genes,
but that do not kill plasmid-free cells), 49% of the cell
population was found to have retained the streptomycin
resistance marker, and a similar proportion to have retained
each tested fermentation trait, after seven rounds of growth
(about 49 generations). The bacteriocin activity measured
was 2133 AU mL1 at the end of each growth period (not
shown). The plasmid derived from wt is thus unstable in
LMGel, but its presence in an LMGel population can be
maintained for a longer time if plasmid-bearing cells have a
selective advantage.
Antilisterial effects in MRS broth and in a meat
system
The wt, mt, LMG, and LMGel strains were then cocultured
with L. monocytogenes in MRS broth (modified in the case of
LMGel) and in a meat matrix (see Meat system and meat
sampling). In both systems (Fig. 4), mt and LMG were
found to exert only a minor negative effect on Listeria
growth. In broth cultures (Fig. 4a), the L. monocytogenes

Table 1. Plasmid stability in LMGel during successive rounds of seeding (at approximately 103 CFU mL1), growth (for 15 h), followed by sampling, and
storage at 4 1C (for 9 h)
Culturing/subculturing on MRS followed by plating on: Culturing/subculturing on MRSCG, followed by plating on:

Passage no. MRS MRSStr,w MRSC MRSG MRS  MRSStr ,w MRSC MRSG
0 257  51 52  6 (20%) 45  8 30  7 215  52 132  7 (61%) 128  15 145  10
1 215  53 11  2 (5%) 72 52 182  35 115  6 (63%) 98  10 90  15
2 233  56 3  1 (1%) 41 21 213  50 122  9 (57%) 85  8 65  7
3 356  82 0 0 0 257  72 135  8 (53%) 110  21 85  10
4 488  35 0 0 0 225  64 127  9 (56%) 95  12 107  20
5 342  50 0 0 0 208  52 93  7 (45%) 115  15 80  10
6 294  72 0 0 0 233  74 115  8 (49%) 134  25 130  22

The initial culture (passage no. 0) was seeded with an appropriate amount of overnight culture on MRSStr.
CFU concentrations in the culture/subculture, expressed in 105 CFU mL1.
w
In the MRSStr column, the percentage of the population expressing streptomycin resistance is also given.

FEMS Microbiol Lett 306 (2010) 37–44 

c2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
42 P. Kouakou et al.

(a) 1.E+11 2500

1.E+10
1.E+09 2000
1.E+08
CFU mL–1

1.E+07

AU mL–1
1500
1.E+06
1.E+05
1.E+04 1000
1.E+03
1.E+02 500
1.E+01
1.E+00 0
0 10 20 30 40 50 60 70
Time (h) at 37°C

Downloaded from https://academic.oup.com/femsle/article-abstract/306/1/37/479127 by guest on 30 August 2019

(b) 1.E+08 2500
1.E+07
2000
1.E+06
1.E+05
CFU g–1

1500

AU g–1
1.E+04
1.E+03 1000 Fig. 4. Evolution of the Listeria monocytogenes
CFU count in (a) MRS broth (modified for LMGel)
1.E+02
500 and (b) a meat matrix. The broth or the meat
1.E+01 matrix was seeded with Listeria alone (^) or
1.E+00 0 coinoculated with wt (&), mt (n), LMG (m), or
0 1 2 3 4 5 6 LMGel (’). Evolution of bacteriocin activity in the
Time (weeks) at 4°C 
cultures containing wt (0) or LMGel ( ).

count decreased quickly in the presence of wt or LMGel,
declining below the detection limit within 24 h. Growth
rebound occurred in both cases, albeit 24 h later in the
LMGel-Listeria than in the wt-Listeria coculture.
After 1 week of storage at 4 1C, the L. monocytogenes
count in both wt- and LMGel-treated raw pork was found to
have declined below the detection limit (Fig. 4b). Growth
rebound occurred after another 2 weeks in the wt-treated
system vs. 4 weeks in the LMGel-treated system.
Bacteriocin activity levels were also measured in these
cultures. As expected, no bacteriocin was detected in sam-
ples containing mt or LMG. In samples containing wt or
LMGel, the lowest Listeria CFU count coincided with peak
bacteriocin activity: 2133 AU mL1 after 24 h in MRS broth
and 2133 AU g1 after 2 weeks in raw meat. In both systems,
the bacteriocin activity decreased more quickly in the
presence of wt than in the presence of LMGel, i.e. 24 h
sooner in the broth culture and 1 week sooner in the meat
system.
In broth, strain LMG grew fastest (mmax = 0.52), followed
by LMGel (0.32) and finally wt (0.22). As the bacteriocin
level increased initially at the same rate and reached the
same peak value in the LMGel and wt cultures, it would
seem that LMGel is a somewhat less efficient bacteriocin
producer than wt, possibly because of plasmid instability.
As expected, bacteriocin-degrading proteolytic activity
remained low (at about 7.5 U mL1) in broth cultures seeded
with LMG or LMGel. In cultures seeded with wt or mt, it
was about twice as high at the start of the culture and seven
times as high at the end.
Discussion
Here, we demonstrate that sakacin P production in
L. curvatus is encoded by a plasmid. This is clear from the
nonbacteriocinogenic phenotype of our plasmid-cured mu-
tants, from binding of a sakacin-specific probe to plasmid
restriction fragments on Southern blots, and from binding
of the same probe to the amplicon produced by PCR from
gel-purified plasmid DNA. This approximately 10-kb plas-
mid, furthermore, appears to exist in two different forms
and additionally to confer streptomycin resistance and the
ability to ferment D-celobiose, gentiobiose, and N-acetylglu-
cosamine. These extra features facilitate selection of plas-
mid-containing cells (with streptomycin) or offer a means of
enhancing plasmid stability (using carbohydrates that can
be fermented only if the plasmid is present).
Plasmids associated with bacteriocin production, anti-
biotic resistance, and/or metabolic traits are common in
lactobacilli (reviewed in Wang & Lee, 1997). Sometimes, the
bacteriocin-encoding plasmid also carries a gene conferring
immunity to the bacteriocin concerned, and plasmid loss
results in sensitivity to that bacteriocin (Møretrø et al.,
2005). Here, our plasmid-cured mt isolates remained resis-
tant to sakacin P (plate assays and high-titre exposure, data


c 2010 Federation of European Microbiological Societies FEMS Microbiol Lett 306 (2010) 37–44
Published by Blackwell Publishing Ltd. All rights reserved
Suppression of Listeria growth rebound in a food system
not shown). This contrasts with the situation described by
Møretrø et al. (2005).
Naturally occurring LAB have long been used in food
technology, and the use of genetically engineered LAB with
improved features is envisaged for many applications. In the
European Union, regulation EC1829/2003 prohibits the use
of genetically modified organisms in human food unless the
proposed use has been approved on the basis of evidence
that it is safe for human health, animal health, and the
environment and that it neither misleads the consumer nor
diminishes the nutritional quality of the food. This regula-
tion is applicable to strain LMGel. Regarding the safety of
this strain, it is worth stressing that this is genetic engineer-
ing performed between two LAB strains of the same genus
that are suitable for use in food. Moreover, the transferred
plasmid is a naturally occurring plasmid of one of them.
This should be quite safe, but the plasmid should never-
theless be sequenced to ensure that it contains no known
toxin genes. Furthermore, the instability of the plasmid in
LMGel could be a problem in the framework of industrial
applications. In a review on the genetics of lactobacilli in
industrial fermentations, Vogel & Ehrmann (1996) mention
the poor segregational and structural stability of certain
plasmids transferred between L. curvatus species. It might be
worth investigating the cause of the instability observed in
the present case.
In our meat system enriched with D-celobiose and
gentiobiose, these sugars are not the sole carbon sources,
but the plasmid appears to be maintained in a sufficient
proportion of the LMGel cells to allow a substantial level
of bacteriocin production and to delay Listeria growth
rebound.
In conclusion, LMGel requires further study and im-
provement before it, or the plasmid it contains, can be used
industrially to prevent Listeria growth in meat fermenta-
tions. Yet, the ability of this strain to delay Listeria growth
rebound in a model meat system seems very promising.
References
Axelsson LT, Ahrné SE, Andersson MC & Stahl SR (1988)
Identification and cloning of a plasmid-encoded erythromycin
resistance determinant from Lactobacillus reuteri. Plasmid 20:
171–174.
Badii R, Jones S & Warner PJ (1989) Sphaeroplast and
electroporation-mediated transformation of Lactobacillus
plantarum. Lett Appl Microbiol 9: 41–44.
Benkerroum N, Ghouati Y, Ghalfi H, Elmejdoub T, Roblain D,
Jacques P & Thonart P (2002) Biocontrol of Listeria
monocytogenes in a model cultured milk (lben) by in situ
bacteriocin production from Lactococcus lactis ssp. lactis.
Int J Dairy Technol 55: 145–151.
FEMS Microbiol Lett 306 (2010) 37–44
43
Chassy BM (1987) Prospects for the genetic manipulation of
Lactobacilli. FEMS Microbiol Rev 46: 297–312.
Chassy BM & Flickinger JL (1987) Transformation of
Lactobacillus casei by electroporation. FEMS Microbiol Lett 44:
173–177.
Church FC, Swaisgood HE, Porter DH & Catignani GL (1983)
Pectrophotometric assay using o-phthaldialdehyde for
determination of proteolysis in milk and isolated milk
proteins. J Dairy Sci 66: 1219–1227.
Collins EB & Harvey RJ (1962) Failure in the production of
citrate permease by Streptococcus diacetilactis. J Dairy Sci 45:
Downloaded from https://academic.oup.com/femsle/article-abstract/306/1/37/479127 by guest on 30 August 2019
32–35.
Danielsen M (2002) Characterization of the tetracycline
resistance plasmid pMD5057 from Lactobacillus plantarum
5057 reveals a composite structure. Plasmid 48: 98–103.
Diop MB, Dubois-dauphin R, Dortu C, Destain J, Tine E &
Thonart P (2008) In vitro detection and characterization of
bacteriocinlike inhibitory activity of lactic acid bacteria (LAB)
isolated from Senegalese local food products. Afr J Microbiol
Res 2: 206–216.
Dortu C, Huch M, Holzapfel WH, Franz CMAP & Thonart P
(2008) Anti-listeria activity of bacteriocin-producing
Lactobacillus curvatus CWBI-B28 and Lactobacillus sakei
CWBI-B1365 on raw beef and poultry meat. Lett Appl
Microbiol 47: 581–586.
Gevers D, Danielsen M, Huys G & Swings J (2003) Molecular
characterization of tet(M) genes in Lactobacillus isolates from
different types of fermented dry sausages. Appl Environ Microb
69: 1270–1275.
Gfeller KY, Roth M, Meile L & Teuber M (2003) Sequence and
genetic organization of the 19.3-kb erythromycin- and
dalfopristin-resistance plasmid pLME300 from Lactobacillus
fermentum ROT1. Plasmid 50: 190–201.
Gilliland SE (1985) Bacterial Starter Cultures for Foods. CRC Press
Inc., Boca Raton, FL.
Jones RN, Barry AL, Gavan TL & Washington JA (1985)
Susceptibility tests: microdilution and macrodilution broth
procedures. Manual of Clinical Microbiology (Lennette EH,
Balows A, Hausler WJ & Shadomy HJ, eds), pp. 972–977.
American Society for Microbiology, Washington, DC.
Josson K, Scheirlinck T, Michiels F, Platteeuw C, Stansens P, Joos
H, Dhaese P, Zabeau M & Mahillon J (1989) Characterization
of a Gram-positive broad-host-range plasmid isolated from
Lactobacillus hilgardii. Plasmid 21: 9–20.
Katla T, Moretro T, Aasen IM, Holck A, Axelsson L & Naterstad K
(2001) Inhibition of Listeria monocytogenes in cold smoked
salmon by addition of sakacin P and/or live Lactobacillus sakei
cultures. Food Microbiol 18: 431–439.
Kim WJ, Ray B & Johnson MC (1992) Plasmid transfers by
conjugation and electroporation in Pediococcus acidilactici.
J Appl Bacteriol 72: 201–207.
Klaenhammer TR (1988) Bacteriocins of lactic acid bacteria.
Biochimie 70: 337–349.
Kouakou P, Ghalfi H, Destin J, Dubois-Dauphin R, Evrard P &
Thonart P (2008) Enhancing the antilisterial effect of

c 2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
44
Lactobacillus curvatus CWBI-B28 in pork meat and cocultures
by limiting bacteriocin degradation. Meat Sci 80: 640–648.
Kouakou P, Ghalfi H, Destin J, Dubois-Dauphin R, Evrard P &
Thonart P (2009) Effects of curing sodium nitrite additive and
natural meat fat on growth control of Listeria monocytogenes
by the bacteriocin-producing Lactobacillus curvatus strain
CWBI-B28. Food Microbiol 26: 623–628.
McKay LL & Baldwin KA (1990) Applications for biotechnology:
present and future improvements in lactic acid bacteria. FEMS
Microbiol Rev 87: 3–14.
Møretrø T, Kristine N, Ellen W, Inga M, Aasen B, Stephane C,
Monique Z & Lars A (2005) Sakacin P non-producing
Lactobacillus sakei strains contain homologues of the sakacin P
gene cluster. Res Microbiol 156: 949–960.
Oberman H & Libudzisz Z (1998) Fermented milks. Microbiology
of Fermented Foods, Vol. 1 (Wood BJB, ed), pp. 308–350.
Blackie Academic & Professional, London.

c 2010 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved
P. Kouakou et al.
Parente E & Hill C (1992) A comparison of factors affecting the
production of two bacteriocins from lactic acid bacteria. J Appl
Bacteriol 73: 290–298.
Schillinger U, Kaya M & Lücke FK (1991) Behavior of Listeria
monocytogenes in meat and its control by a bacteriocin-
producing strain of Lactobacillus sake. J Appl Bacteriol 70:
473–478.
Sonstein SA & Baldwin JN (1972) Loss of the penicillinase
plasmid after treatment of Staphylococcus aureus with sodium
dodecyl sulfate. J Bacteriol 109: 262–265.
Tagg JR, Dajani AS & Wannamaker LW (1976) Bacteriocins
Downloaded from https://academic.oup.com/femsle/article-abstract/306/1/37/479127 by guest on 30 August 2019
of gram-positive bacteria. Bacteriol Rev 40:
722–756.
Vogel RF & Ehrmann M (1996) Genetics of lactobacilli in food
fermentations. Biotechnol Ann Rev 2: 123–127.
Wang TT & Lee BH (1997) Plasmids in Lactobacillus. Crit Rev
Biotechnol 17: 227–272.
FEMS Microbiol Lett 306 (2010) 37–44