ARTICLE IN PRESS

FOODMICROBIOLOGY

0740-0020/$ - se

doi:10.1016/j.fm

�CorrespondE-mail addr

Food Microbiology 25 (2008) 278–287

www.elsevier.com/locate/fm

Succession of dominant and antagonistic lactic acid bacteria infermented cucumber: Insights from a PCR-based approach

Atul Kumar Singh, Aiyagari Ramesh�

Department of Biotechnology, Indian Institute of Technology Guwahati, Guwahati 781 039, Assam, India

Received 13 August 2007; received in revised form 12 October 2007; accepted 14 October 2007

Available online 22 October 2007

Abstract

The goal of the investigation was to study the succession of major groups of lactic acid bacteria (LAB) and their antagonism in salt-

fermented cucumber using PCR. In a direct detection method as well as a short enrichment process, PCR enabled detection of

Leuconostoc and Lactobacillus during early hours of fermentation. Subsequently, Lactobacillus and Pediococcus emerged as the dominant

genera. Nucleic acid sequence of culture-independent clones confirmed the detection of Pediococcus as a dominant genera emerging

during late stages of fermentation. PCR also revealed time-dependent emergence of mesentericin, pediocin and plantaricin A producers

and accounted for the LAB succession in the fermenting samples. A total of 328 LAB isolates were obtained collectively from 30

cucumber samples, of which PCR could identify an overwhelming 186 Lactobacillus isolates followed by 113 Pediococcus and 29

Leuconostoc isolates, respectively. Based on antimicrobial assay against target strain Leuconostoc mesenteroides NRRL B640, 28% of the

LAB were bacteriocin producers, of which pediocin producers were substantial, followed by plantaricin A and mesentericin producers.

The bacteriocins elaborated by the isolates were active against a large number of Gram-positive target LAB strains and pathogenic

bacteria including Bacillus cereus, Enterobacter aerogenes, Enterococcus faecalis, Escherichia coli, Listeria monocytogenes and

Staphylococcus aureus.

r 2007 Elsevier Ltd. All rights reserved.

Keywords: Microbial succession; Diversity; Lactic acid bacteria; Fermented cucumber; PCR; Bacteriocin

1. Introduction

Lactic acid bacteria (LAB) comprise a versatile group ofmicroorganisms enjoying a generally regarded as safestatus and have profound applications in food fermenta-tion. Apart from their beneficial attributes such as acidproduction, proteolytic activity, and aroma formation,which have direct relevance in food fermentation processes,they have an enormous potential to inhibit microorganismsthrough the production of bacteriocins (Caplice andFitzgerald, 1999; Leroy and De Vuyst, 2004; Drideret al., 2006). Bacteriocins are mostly proteinaceous innature and majority of the bacteriocins produced by LABbelong to the Class IIa category which are essentially smallheat-stable peptides with considerable sequence homologyand strong antimicrobial activity against closely related

e front matter r 2007 Elsevier Ltd. All rights reserved.

.2007.10.010

ing author. Tel.: +91361 2582205; fax: +91 361 2582249.

ess: aramesh@iitg.ernet.in (A. Ramesh).

Gram-positive bacteria (Ennahar et al., 2000; Drider et al.,2006). Class IIa bacteriocins like pediocins are particularlyknown for their strong antilisterial activity, making themattractive candidates for application in food systems(Jagannath et al., 2001; Rodrı́guez et al., 2002). Owing toan impending threat from foodborne pathogens, thediscovery of new food processes and the consumer demandfor minimally processed food products, LAB have come tothe forefront as biopreservative agents and substantialresearch has been focused in recent times to isolatebacteriocin-producing LAB from natural sources andexploit them for production and preservation of fermentedproduct.Microbial succession is a universal phenomenon ob-

served in natural fermentation processes and is often areflection of ensuing microbial interaction, competition forintrinsic growth factors such as nutrients and resistance togrowth inhibitory environmental perturbations like highacidity and production of antimicrobials. Consequently,

ARTICLE IN PRESS

Table 1

Reference strains of lactic acid bacteria (LAB) and pathogenic bacteria

used in the present investigation

Organism Strain

Lactobacillus acidophilus MTCC 447

Lactobacillus casei subsp. casei NRRL B1922

Lactobacillus delbrueckii subsp. bulgaricus NRRL B548

Lactobacillus delbrueckii subsp. lactis MTCC 911

Lactobacillus fermentum MTCC 903

Lactobacillus johnsonii NRRL B2178

Lactobacillus plantarum MTCC 1325

Lactobacillus rhamnosus MTCC 1408

Lactobacillus sakei NRRL B1917

Lactococcus lactis subsp. cremoris MTCC 1484

Lactococcus lactis subsp. lactis MTCC 3041

Leuconostoc mesenteroides NRRL B640

Leuconostoc mesenteroides MTCC 107

subsp. mesenteroides

Pediococcus acidilactici NRRL B14009, CFR K7a

Escherichia coli MTCC 443

Bacillus cereus MTCC 1305

Listeria monocytogenes Scott A

Staphylococcus aureus MTCC 96

Enterobacter aerogenes MTCC 2822

Enterococcus faecalis MTCC 439

NRRL: Northern Regional Research Laboratory, Peoria, IL, USA;

MTCC: Microbial Type Culture Collection, Institute of Microbial

Technology (IMTECH), Chandigarh, India.aCulture provided by Dr. Prakash Halami, Central Food Technological

Research Institute (CFTRI), Mysore, India.

A.K. Singh, A. Ramesh / Food Microbiology 25 (2008) 278–287 279

robust microbial populations that enjoy a selectiveadvantage owing to either acid resistance or antagonistictraits such as the ability to produce bacteriocins emerge asthe dominant population at the end of fermentation.Knowledge of such microbial interaction in a naturalfermentation process not only provides a fundamentalunderstanding of the process parameters, but can alsoform the basis of designing a strategy of isolatingantagonistic cultures from a natural source. Convention-ally, isolation of bacteriocin-producing LAB is accom-plished using microbiological methods by obtainingcolonies from fermented samples and performing amicrobial assay. However, this exercise mostly results inthe isolation of a dominant population and owing to thetarget specificity of bacteriocins, the process demandsinclusion of a large number of target strains and is thuscumbersome.

Investigations concerning microbial dynamics in foodsystem have adopted a PCR-based approach to studydevelopment of community structure and the fate ofinteracting microorganisms in defined fermented products(Ercolini, 2004; Randazzo et al., 2006; Rantsiou andCocolin, 2006; Delbe’s et al., 2007). However, these studieslargely reported the succession of a dominant populationand do not reveal the antagonistic attributes, whichcontribute to their persistence. A holistic view of theensuing microbial succession in a natural fermentationprocess can be rationally obtained from an experimentaldesign, which is able to detect all the major groups ofinteracting microorganisms, and relate the observedsuccession to their antagonistic properties. The presentinvestigation is primarily based on the aforesaid conceptand is aimed to analyze the succession of LAB in salt-fermented cucumber and concomitantly detects the emer-gence of bacteriocin producers as fermentation proceeds.We have selected cucumber as a model system based on aprevious investigation, which demonstrated that cucumbercan be a promising source for isolating pediocin-likebacteriocin producers (Halami et al., 2005). A molecularapproach based on PCR was employed to ascertain thesuccession of LAB existing as inherent microflora infermented cucumber samples and correlate the successionwith the emergence of bacteriocin-producing strains. Tothe best of our knowledge, this is the first report on acomprehensive time-dependent analysis using culture-independent and -dependent methods to determinethe emergence of dominant LAB genera in fermentedcucumber and simultaneously monitor antagonism bydetecting bacteriocin producers at various time periods offermentation.

2. Materials and methods

2.1. Bacterial strains and culture conditions

The reference LAB and pathogenic bacterial strains usedin the present investigation are shown in Table 1. The

standard and natural isolates of LAB were maintained asfrozen stocks in milk and glycerol (10% each) at �20 1Cand subcultured regularly in de Man, Rogosa and Sharpe(MRS) broth (HiMedia, Mumbai, India) at 37 1C in a staticincubator every fortnight. Strains of Escherichia coli,Listeria monocytogenes, Staphylococcus aureus and Enter-

ococcus faecalis were propagated in brain–heart infusionbroth (HiMedia, India) whereas Bacillus cereus andEnterobacter aerogenes were grown in nutrient broth(HiMedia, India).

2.2. Molecular detection of LAB succession in fermented

cucumber

A total of 30 cucumber samples randomly collected fromvarious vegetable markets of Guwahati city were chosenfor this study. Cucumber samples were fermented in 4%saline solution in static mode at 35 1C. Samples wereperiodically taken at intervals of 6, 12, 18, 24, 36, 48 and72 h after the onset of fermentation. In the culture-independent method, 1.0ml aliquot of the samples werecentrifuged briefly at 8000 rpm for 5min at 4 1C to pelletcells and washed thrice with phosphate-buffered saline(PBS) followed by single wash with autoclaved ultrafilteredwater. Washed cells were resuspended in 100 ml of sterilewater and 400 ml of 1% Triton X-100. For DNAextraction, resuspended cell were incubated in boiling

ARTICLE IN PRESSA.K. Singh, A. Ramesh / Food Microbiology 25 (2008) 278–287280

water for 5min and snap cooled on ice. In the culture-dependent method, 1.0ml aliquots of the sample weresubjected to short enrichment in MRS broth for 6 h at37 1C. Cells were separated from a 1.0ml aliquot ofenriched culture broth by centrifugation, washedthrice with PBS and DNA was extracted using 1% TritonX-100 as mentioned earlier. A 2.0 ml aliquot of extractedDNA from both the methods was used in PCR inconjunction with genus-specific and bacteriocin-specificprimers.

2.3. Genus specific and bacteriocin primers

Nucleotide sequence of the 16S rRNA gene of LABcomprising of strains belonging to genera Lactobacilli,Pediococci, Lactococci and Leuconostoc (28 sequences) andselect non-LAB were retrieved from the GenBank database(http://www.ncbi.nlm.nih.gov). The sequences were alignedusing the web-based multiple alignment program (http://www.genomatix.de/cgi-bin/dialign/dialign.pl). Primersspecific to the genera Lactobacilli, Pediococci, Lactococci

and Leuconostoc were designed as shown in Table 2.The sequences of the primers were subjected to BLASTanalysis to determine their specificity. The specificity of thegenus specific primers was also validated by PCR usingreference strains of LAB. Primers were also designed forthe following bacteriocins—pediocin, mesentericin, enter-ocin A, nisin and plantaricin A using the program Pri-mer 3.0 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). The sequences of bacteriocin primers areindicated in Table 2.

Table 2

16S rRNA and bacteriocin gene-specific primers used in the present investigat

Primer

name

Nucleotide sequence (50–30)

1F AGAAGAGGACAGTGGAAC

1R TTACAAACTCTCATGGTGTG

2F TAAAGCGAGCGCAGGTGG

2R GGTTACCTTGTTACGACTT

3F CTGAATGAGATTTTAACACG

3R GGTTTTAAGAGATTAGCT

4F AGAGATGGATCCGCGGTGCA

4R TTACAAACTCCCATGGTGTG

IF TGGCCAATATCATTGGTGGT

IR GCATTTATGATTACCTTGATGTCC

IIF ACCAAAATCCATTTCCACCA

IIR TCTGTGGAAGCATATCAGCAA

IIIF CAGACACAACTTATCTATGGGGGTA

IIIR CCTGGAATTGCTCCACCTAA

IVF CTTGGATTTGGTATCTGTTTCG

IVR TTGCTTACGTGAATACTACAATGACA

VF GAAAATTCAAATTAAAGGTATGAAGCA

VR ATTGCAGTTGCCCCCATCT

2.4. PCR conditions

Amplification of bacterial DNA was performed in a totalreaction volume of 25 ml, which contained 2 ml of DNAsamples comprising either genomic DNA from referencestrains and natural isolates or DNA extracted fromfermented cucumber samples. Separate PCR reactionswere performed to detect a particular genus of LAB or abacteriocin gene. The reaction mixture consisted of 10times diluted PCR buffer, 200 mM of each deoxynu-cleoside triphosphate (dNTP), 1U of Taq DNA polymer-ase (Sibenzyme, Novosibirsk, Russia) and 50 pmoleach of forward and reverse primer. Template DNAwas initially denatured at 94 1C for 2min. Subsequently,a total of 35 amplification cycles were carried out in aprogrammable thermal cycler (Gene Amp Gold PCRSystem, Applied Biosystems, USA). Each cycle con-sisted of denaturation for 1min at 94 1C, primer annea-ling for 1min at 55 1C and extension for 1min at 72 1C.The last cycle was followed by a final extension at 72 1Cfor 10min. Amplification of bacteriocin genes was carriedout under similar conditions except for an annealingtemperature of 58 1C. PCR products were analyzed byagarose (1%) gel electrophoresis (Sambrook and Russell,2001).

2.5. Enumeration and detection of antagonistic LAB in salt-

fermented cucumbers

Samples were collected periodically and analyzed fortotal LAB counts by pour plating in MRS-agar media

ion

Primer annealing

position (50–30)

Reference gene

accession no.

PCR target

677–694 AB104855 Lactobacillus

1424–1443

572–589 AF515226 Lactococcus

1492–1510

95–114 AF515229 Pediococcus

1307–1324

232–251 M23035 Leuconostoc

1416–1435

37–56 AY083244 Pediocin

165–188

383–402 AY286003 Mesentericin

531–551

40–64 AF240561 Enterocin A

178–197

21–42 AY526091 Nisin

145–170

553–579 X75323 Plantaricin A

639–657

ARTICLE IN PRESS

Fig. 1. Amplicons obtained from reference strains of lactic acid bacteria

using genus-specific primers designed for 16S rRNA gene. Lanes: M: lDNA EcoRI/HindIII double digest size marker; (1) Lactobacillus casei

subsp. casei NRRL B1922; (2) Lactobacillus plantarum MTCC 1325; (3

and 4) Pediococcus acidilactici CFR K7 and NRRL B14009; (5)

Lactococcus lactis subsp. cremoris MTCC 1484; (6) Lactococcus lactis

subsp. lactis MTCC 3041; (7) Leuconostoc mesenteroides subsp. mesenter-

oides MTCC 107; (8) Leuconostoc mesenteroides NRRL B640.

A.K. Singh, A. Ramesh / Food Microbiology 25 (2008) 278–287 281

(HiMedia, India). Aliquots of fermented cucumber sampleswere centrifuged briefly and the pH of the supernatant wasmeasured as an index of acid production resulting from thegrowth of LAB. The number of bacteriocin-producingcolonies (Bac+) at each time period was determined byoverlaying LAB colonies obtained on MRS agar plateswith 106 cells of freshly grown target strain of Leuconostoc

mesenteroides NRRL B640 seeded in MRS soft agar(0.8% agar) and observing zone of growth inhibitionaround the colonies. Select colonies of LAB consis-ting of bacteriocin producers (Bac+) and non-producers(Bac�) obtained collectively from 30 samples were main-tained as frozen stocks in milk and glycerol (10% each)at �20 1C. The isolates were propagated overnight inMRS medium at 37 1C under static condition, prior tosubsequent experiments. The genomic DNA from theisolates was extracted using the Triton-X 100 method asmentioned before and the genera of the LAB isolates andthe class of bacteriocin produced were identified by PCRusing genus specific 16S rRNA and bacteriocin primers(Table 2).

2.4. Antimicrobial spectrum of bacteriocinogenic LAB

isolates

The bacteriocinogenic isolates obtained from cucumberwere grown in MRS broth at 37 1C for 18 h. Cells wereseparated by centrifugation at 10,000 rpm for 10min at4 1C and the pH of the culture filtrate (CF) was adjustedto 7.0 with 2M NaOH, filtered through a 0.2 mmmembrane filter (Millipore, India) and stored at �20 1Ctill further use. Prior to assay, the samples were incu-bated in boiling water for 20min, cooled to roomtemperature and then used against indicator strainsconsisting of LAB strains or pathogenic bacteria(Table 1). The antimicrobial activity of the samples wasascertained by the well diffusion assay. Briefly, 100 mlaliquot of an overnight culture of the test organism grownin the respective medium was used to seed 20ml of MRSsoft agar (0.8%w/v). The wells were loaded with 50 ml oftest sample and the plate was incubated at 4 1C for aminimum of 3 h to allow diffusion before overnightincubation at 37 1C. The antimicrobial activity of thesamples was determined by the zone of inhibition producedaround the wells.

2.5. Nucleic acid sequence

Select amplicons obtained using genus-specific primersfor Pediococcus were directly sequenced in the NationalFacility for Automated DNA Sequencing, Department ofBiochemistry, Delhi University, South Campus. Thesequences obtained were subjected to BLAST analysisand deposited in GenBank database with accessionnumbers EF636660, EF636661 and EF636662 and thesizes of the sequenced amplicons were 656, 689 and 696 bp,respectively.

3. Results and discussion

3.1. Genus-specific primers

Multiple alignment of the 16S rRNA gene sequence ofLAB enabled us to select candidate regions for designinggenus specific primers. The expected size of amplicons forLactobacilli, Lactococci, Pediococci and Leuconostoc wereapproximately 800, 960, 1200 and 1100 bp, respectively.The primers designed for specific detection of LAB generawere validated by performing PCR with DNA isolatedfrom standard LAB strains indicated in Table 1. Ampli-cons obtained from two strains each belonging to the fourmajor genera of LAB is indicated in Fig. 1. It is evidentfrom the figure that the size of the amplicons coincidedwith the expected size. The specificity of the designedprimers was further confirmed by the lack of cross-reactivity of the primers amongst the LAB genera andthe non-LAB pathogens listed in Table 1. This was criticalin the context of the present study, which involvedsimultaneous detection of mixed LAB population likelyto be present as a natural microflora in cucumber. Previousstudies have focused on the application of 16S rRNAprimers to detect a specific LAB genus or species usingconventional and multiplex PCR approach (Dubernetet al., 2002; Settanni et al., 2005). Substantial researchefforts have also been directed on the design of 16S rRNAprimers and its application to monitor LAB dynamics andcommunity development in food system (Randazzo et al.,2006; Delbe’s et al., 2007). In the present investigation, ourefforts were concerted towards the design 16S rRNA genespecific primers that enabled us to simultaneously detectand predict interactions between the major LAB groups ina natural fermentation process.

ARTICLE IN PRESSA.K. Singh, A. Ramesh / Food Microbiology 25 (2008) 278–287282

3.2. PCR-based detection of dominant LAB and bacteriocin

producers in fermented cucumber

PCR was employed to salt-fermented cucumber samplesto detect the succession of major LAB groups existing asinherent microflora and correlate the succession with theemergence of bacteriocin-producing LAB. The cucumbersamples were collected from various vegetables markets inand around Guwahati city to account for any samplevariations. The method of analysis comprised time-dependent sampling during fermentation with an aim tounderstand microbial dynamics, and included a culture-independent and -dependent approach to decipher any biasintroduced due to selection pressure exerted by the growthmedium. Table 3 indicates the results obtained from 30samples using a culture-independent and an enrichmenttechnique. In the direct detection method, PCR enabled thedetection of Lactobacillus and Leuconostoc as early as 12 hof fermentation and revealed their coexistence upto 18 h, in11 samples, whereas in the remaining 19 samples onlyLactobacillus could be detected. The trend for the enrichedsamples was similar wherein Lactobacillus and Leuconostoc

was observed in 5 samples and Lactobacillus alone in theremaining 25 samples after 18 h of fermentation. FromTable 3 it is also evident that as the fermentationprogressed, Leuconostoc could be detected in only 6samples by 36 h in the direct detection method. Concomi-tantly, the emergence of Pediococcus was observed in thesamples. The persistence of Lactobacillus was evident inthe samples during this time period, as it co-evolved eitherwith Pediococcus and Leuconostoc (6 samples) or with

Table 3

Molecular detection of LAB succession and bacteriocin producers in salt-ferm

Fermentation

time (h)

Direct detection

Genusa Bacteriocinb

6 Nil Nil

12 Lb., Leu. (11) Pediocin (08)

Lb. (19) Mesentericin (04)

18 Lb., Leu. (11) Pediocin (14)

Lb. (19) Mesentericin (04)

24 Lb., Leu. (08) Pediocin (20)

Lb. (22) Mesentericin (02)

36 Lb., Ped, Leu. (06) Pediocin (24)

Lb.,Ped. (24) Mesentericin (02)

Plantaricin A (04)

48 Lb., Ped. (30) Pediocin (26)

Plantaricin A (10)

72 Lb., Ped. (30) Pediocin (27)

Plantaricin A (11)

aGenus-specific 16S rRNA primers were used for Lactobacillus: Lb., Pedioc

imply the number of samples in which the indicated LAB genera were obtainbBacteriocin gene-specific primers were used for Pediocin, Mesentericin, En

number of samples in which bacteriocin-encoding genes were detected at spec

Pediococcus alone (24 samples). In contrast, Leuconostoc

and Pediococcus could not be detected in the enrichedsamples by 36 h as all the 30 samples revealed the existenceof only Lactobacillus. In the latter part of fermentation(48 and 72 h), the persistence of Lactobacillus andPediococcus was clearly established as they were detectedin all 30 samples in the direct detection as well asenrichment method. The genus Lactococcus could not bedetected in any of the samples by the PCR method. Thiscould possibly be attributed to the presence of extremelylow numbers of Lactococcus in cucumber samples incomparison to other dominant genus. Consequently, inpresence of overwhelming competing DNA, ampliconsspecific for the genus Lactococcus could not be detectedby PCR.In the present investigation, submersion of sliced

cucumber samples in salt solution and their tight packingin screw-cap bottles resulted in an anaerobic environment.Consequently, the growth of aerobic non-LAB wassuppressed and lactic acid fermentation was initiated byfacultative anaerobes. Besides, salt favors the developmentof anaerobic conditions in the fermentation vessel andexerts a selective effect on LAB naturally present oncucumber. Although the population of LAB present inplant material is presumably small, nonetheless it isassumed that plants are a natural habitat for many LABspecies (Daeschel et al., 1987). Leuconostocs are facultativeanaerobes, present on plant materials, and they are knownto initiate fermentation in vegetable products. The initialloads of population, growth rate, salt and acid tolerance, aswell as bacteriocin-producing attributes are apparently

ented cucumber

Enrichment

Genusa Bacteriocinb

Nil Nil

Lb., Leu. (08) Pediocin (05)

Lb. (22) Mesentericin (02)

Lb., Leu. (05) Pediocin (07)

Lb. (25) Mesentericin (02)

Lb., Leu.(03) Pediocin (11)

Lb. (27)

Lb. (30) Pediocin (19)

Lb., Ped. (30) Pediocin (22)

Plantaricin A (06)

Lb., Ped. (30) Pediocin (23)

Plantaricin A (06)

occus: Ped., Lactococcus: Lc., Leuconostoc: Leu. Numbers in parenthesis

ed.

terocin A, Nisin and Plantaricin A. Numbers in parenthesis indicate the

ific time periods of fermentation.

ARTICLE IN PRESSA.K. Singh, A. Ramesh / Food Microbiology 25 (2008) 278–287 283

involved in the succession of various LAB in vegetablefermentations. Pediococcus pentosaceus and Lactobacillus

plantarum are more acid tolerant and are generally presentduring the later stages of fermentation as the dominantspecies (Daeschel et al., 1987; Sandhu and Shukla, 1996;Harris, 1998). The initial load of LAB in cucumber samplesis known to be low at the onset of fermentation and theinability of conventional microbiological techniques toenumerate and identify specific LAB genera is a bottleneckto study the succession of LAB in the samples. However, asimple detergent-based DNA extraction process providingPCR-compatible templates, in conjunction with the de-signed genus-specific primers adopted in the present studycould overcome this limitation and detect Lactobacillus andLeuconostoc in the samples during early hours of fermenta-tion. The inability of the PCR method to detect Lacto-

coccus and Pediococcus during initial stages offermentation is possibly a reflection of the limit ofdetection of the PCR method, owing to the presenceof copious amount of competing DNA. Select ampliconsobtained in the culture-independent method using genus-specific primers for Pediococcus were subjected to nucleicacid sequencing. The sequences were deposited in GenBankwith accession numbers EF636660, EF636661 andEF636662 and the sizes of the amplicons were 656, 689and 696 bp. Homology search of the sequence of theculture-independent clones by BLAST revealed that theclones exhibited very high similarity (99%) to strains ofP. pentosaceus. Thus, the specificity of the primers used aswell as the identity of the dominant strain evolving duringthe course of cucumber fermentation was validated.

PCR was also used to detect the succession ofbacteriocin producers. Pediocin and mesentericin produ-cers could be detected in the samples by 12 h offermentation. After 24 h of fermentation, pediocin produ-cers could be detected in majority of the samples by thedirect detection method (20 samples) whereas the numberof mesentericin producers dropped drastically as they weredetected in only 2 samples. In contrast, only pediocinproducers could be detected in the enriched samples(11 nos.) by 24 h of fermentation. The ability to producebacteriocin is an inherent attribute in LAB. The signifi-cance of bacteriocin-producing LAB is even more in thecontext of microbial interaction observed in a naturalfermentation process, wherein antagonistic cultures areexpected to enjoy a distinct selective advantage asfermentation proceeds. In the present study, it wasobserved that Leuconostocs thrive during the early stagesof fermentation. This could possibly account for thedetection of mesentericin producers in those samples. Itwas also observed that Lactobacillus could be detected inthe initial stages of fermentation. It is quite likely that somestrains of Lactobacillus were presumably able to producepediocin. Heterologous production of pediocin by Lacto-

bacillus has been reported earlier (Ennahar et al., 1996). Inan earlier study a pediocin-specific probe was used incolony hybridization experiments for isolating pediocin

producers from fermented cucumber (Halami et al., 2005).However, the study was not conducted on a time-dependent basis to evaluate the succession of LAB strainsin the samples.As fermentation progressed (36 h), the number of

pediocin producers enhanced and they could be detectedin 24 samples by the direct detection method. However, thenumber of mesentericin producers still remained very low(only 2 samples) whereas plantaricin A producers started toemerge and were detected in 4 samples. For the enrichedsamples, only pediocin producers could be detected in 19samples after 36 h of fermentation. The elimination ofmesentericin producers and the steady increase in thenumber of pediocin producers in the samples corroborateswell with our earlier results which reveal the emergence ofLactobacillus and Pediococcus as the dominant genera after36 h of fermentation. Subsequently, pediocin and plantar-icin A producers persisted and could be detected after 48and 72 h of fermentation in both the direct and enrichmentmethod. The detection of plantaricin A producers can beaccounted by the emergence of Lactobacillus as thedominant genus during the later stages of fermentation.It is evident from Table 3 that there is a qualitativedifference observed in the number of pediocin producers inthe direct and enrichment method. It is also observed thatalthough there is a general increase of Lactobacillus due toenrichment, the proportion of samples indicating thepresence of pediocin producers were relatively low com-pared to the direct detection method. The population ofLactobacillus obtained from fermented cucumber samplesprobably consist of pediocin producers and pediocinresistant population of Lactobacillus which may not bepediocin producers themselves. Presumably the enrichmentprocess favors the selective growth of the pediocin resistantpopulation of Lactobacillus resulting in the observationthat a steady increase in the detection of Lactobacillus isnot accompanied by a proportionate increase in thenumber of samples having pediocin producers. Enterocinand nisin producers could not be detected in any of thesamples by the PCR method.Considerable research has been carried out to demon-

strate LAB dynamics in fermented food samples. Acommon strategy in majority of the investigations is theuse of molecular techniques such as PCR-DGGE toascertain microbial dynamics and community developmentin food systems (Fontana et al., 2005; Randazzo et al.,2006). Some studies have also been conducted on commu-nity structure and detection of native LAB in traditionalfermented vegetable products (Kim and Chun, 2005;Tamang et al., 2005). These methods essentially identifythe dominant LAB species, but do not reveal theirantagonistic attributes that are consequential in conferringselective advantage to the dominant population. In thepresent study, for the first time, a combined PCR-basedholistic approach to detect the major LAB genera as well ascharacteristic Class IIa bacteriocin producers addresses theaforementioned phenomenon and provides a pragmatic

ARTICLE IN PRESSA.K. Singh, A. Ramesh / Food Microbiology 25 (2008) 278–287284

picture of LAB succession in fermented cucumber samples.The presence of different LAB genera and Class IIabacteriocin producers in cucumber samples provides thepossibility of rapid large scale screening of samples throughthe application of multiplex PCR which is amicable toautomation. The use of automated PCR techniques toanalyze LAB succession and dynamics has been reportedearlier (Lazzi et al., 2004; Brusetti et al., 2006).

3.3. Growth profile of LAB, isolation and classification of

bacteriocin-producing LAB in fermented cucumber

The growth profile of LAB and acid production in salt-fermented cucumber is shown in Fig. 2 as an average of 30samples. Analyses of samples at the initial stages offermentation revealed the presence of approximately5.5 log10 cfu/ml. However, from Table 3 it can be seen thatPCR failed to detect any LAB genera and bacteriocinproducers at 6 h of fermentation. The population of LABobtained at 6 h of fermentation indicates the total LABcount and comprises of all the genera. The proportion ofindividual LAB genus present in early stages of fermenta-tion is presumably low. Thus the sensitivity of PCR-baseddetection is compromised due to the presence of minisculeamount of target DNA in comparison to competingDNA from other genera. It is evident that there is a rapidincrease in growth in the first 18 h of fermentation whereinthe average cell number amounted to approximately8.0 log10 cfu/ml. Subsequently, the LAB population fol-lowed a steady decline to reach approximately 7.0 log10 cfu/ml after 72 h of fermentation. The increase in LAB cellnumber was accompanied by a concomitant decline of pHfrom 6.7 to 3.25, indicating growth-associated acidproduction. Since fermentation mainly occurs in the liquidphase, rapid release of nutrients from sliced cucumber

Fig. 2. Kinetics of growth of LAB ( ) and acidification (&) in

fermented cucumber samples. Results are demonstrated as average of 30

samples.

samples results in accelerated growth and high acidproduction. The rapid drop in pH observed in the presentstudy is a combined effect of high acid production andinadequate buffering capacity of the fermenting cucumbersamples. The variations observed for growth kinetics andacid production in the samples analyzed is essentially afunction of the type, availability and concentration offermenting substrate, buffering capacity, competing micro-organisms and antagonists inherently present in cucumber.A total of 328 LAB isolates obtained from 30 cucumber

samples at different time periods of fermentation weresubjected to PCR-based genus identification. It is clearfrom Table 4 that Lactobacillus comprised a significantproportion of the population, amounting to nearly 56%(186 nos.) of the isolates followed by Pediococcus, whichaccounted for approximately 34% (118 nos.) of theisolates. A small fraction of the isolates belong to genusLeuconostoc (29 nos.). The overwhelming presence ofLactobacillus and Pediococcus agree well with our earlierresults shown in Table 3 wherein they could be detected inlarge number of samples in both the direct and enrichedmethod. Low numbers of Leuconostoc are expectedconsidering that they were obtained from samples takenduring early hours of fermentation. Subsequently, due toacid production and copious numbers of bacteriocinproducers the Leuconostoc population was eliminated.Bacteriocin-producing attribute of the LAB isolates was

ascertained by overlaying the colonies with a lawn of thesensitive strain L. mesenteroides NRRL B640. A represen-tative assay plate indicating the presence of bacteriocin-producing colonies of LAB is shown in Fig. 3. It is clearfrom the figure that bacteriocin-producing colonies wereclearly discernible from non-producers as they produce alarge zone of inhibition on a lawn of the indicator strain.The total number of bacteriocin-producing LAB coloniesobtained collectively from 30 samples was 92, whichcorrespond to approximately 28% of the total LABcolonies obtained. The class of bacteriocin produced bythe antagonistic isolates was assessed by PCR usingprimers specific for the bacteriocins (Table 2). Thequantitative comparison of the class of bacteriocinproducers obtained from the cucumber samples is depictedin Table 4. A total number of 7 mesentericin producerswere obtained after 12–18 h of fermentation. The numberof pediocin producers encompassed a vast majority of thebacteriocin producers (57 nos.). Presumably, strainsbelonging to genus Lactobacillus and Pediococcus togetheraccount for the large number of pediocin producersobserved. Out of a total of 57 pediocin producers, thehighest numbers (21 nos.) were obtained after 72 h offermentation. Likewise, a total of 16 plantaricin Aproducers emerged late following 48 and 72 h of fermenta-tion. Thus different classes of bacteriocin producers evolveat defined time-periods and this can form a strategic basisfor isolating potent bacteriocin-producing strains from anatural source. Amongst the bacteriocin producers, 12additional LAB isolates were established as bacteriocin

ARTICLE IN PRESS

Fig. 3. Colonies of lactic acid bacteria obtained from salt fermented

cucumber. Arrow indicates a bacteriocin-producing colony with zone of

inhibition produced on a lawn of the target strain Leuconostoc

mesenteroides NRRL B640.

Table 4

PCR-based identification of LAB and bacteriocin producers obtained at different time periods from fermented cucumber samplesa

Fermentation time period Total count

6 h 12 h 18 h 24 h 36 h 48 h 72 h

LAB generab

Lactobacillus 03 10 17 23 31 48 54 186

Pediococcus 02 07 15 19 20 22 28 113

Leuconostoc 03 14 08 04 Nil Nil Nil 29

Lactococcus Nil Nil Nil Nil Nil Nil Nil Nil

Total LAB number 328

Bacteriocin producersc

Pediocin Nil 04 04 06 09 13 21 57

Plantaricin A Nil Nil Nil Nil Nil 07 09 16

Mesentericin Nil 04 03 Nil Nil Nil Nil 07

Enterocin A Nil Nil Nil Nil Nil Nil Nil Nil

Nisin Nil Nil Nil Nil Nil Nil Nil Nil

Unidentified bacteriocin producers 12

Total no. of bacteriocin producers 92

aData shown as aggregate of 30 cucumber samples.bLAB genera was determined by PCR using genus-specific primers for 16S rRNA gene.cBacteriocin-producing colonies were selected by colony overlay assay using Leuconostoc mesenteroides NRRL B640 as target strain. Bacteriocin

encoding genes were detected in the isolates by PCR using gene-specific primers for pediocin, plantaricin A, mesentericin, enterocin A and nisin.

A.K. Singh, A. Ramesh / Food Microbiology 25 (2008) 278–287 285

producers by colony overlay assay. However, no specificamplicon could be detected using the bacteriocin gene-specific primers. Evidently, the bacteriocin produced bythese isolates are different from the ones tested by the PCRmethod in the present investigation and hence could not beidentified.

3.4. Antimicrobial spectrum of bacteriocin-producing LAB

The culture filtrate of all the 92 bacteriocin-producingLAB was tested against a range of target LAB strains andpathogenic bacteria indicated in Table 5. It is evident fromthe table that a large number of bacteriocin producersexhibited activity against the chosen LAB target strains.The highest numbers of antagonistic isolates were obtainedagainst L. mesenteroides NRRL B640, which was originallyused as a target strain to isolate the bacteriocin producersfrom fermented cucumber. The culture filtrate of asignificant number of isolates (87 nos.) also demonstratedinhibitory activity against Lactobacillus casei subsp. casei

NRRL B1922, Lactobacillus fermentum MTCC 903,Lactobacillus sakei NRRL B1917 and Pediococcus acid-

ilactici NRRL B14009. Amongst the LAB used as targetstrain, Lactobacillus johnsonii NRRL B2178 and Lactoba-

cillus rhamnosus MTCC 1408 were resistant. Majority ofthe bacteriocins detected in the natural isolates (mesenter-icin, pediocin and plantaricin A) belong to the Class IIacategory, which is known to act on closely related LABstrains (Ennahar et al., 2000). It is also evident from thetable that the culture filtrate of the isolates was inhibitoryto pathogenic bacterial strains. A large number of isolates(48 nos.) could readily inhibit the growth of the foodbornepathogen Listeria monocytogenes Scott A. Class IIabacteriocin-like pediocins are known to have strongantilisterial activity (Jagannath et al., 2001; Rodrı́guezet al., 2002). The antimicrobial activity of the culturefiltrates of selected isolates on a lawn of L. monocytogenes

Scott A is also depicted in Fig. 4 wherein zone of growthinhibition around the well is conspicuous. Some isolates

ARTICLE IN PRESS

Table 5

Antimicrobial spectrum of bacteriocinogenic LAB isolates (Bac+)

obtained from salt-fermented cucumber samples against target LAB and

pathogenic bacterial strains

Target bacterial strain No. of Bac+

LAB isolates

Lactobacillus acidophilus MTCC 447 48

Lactobacillus casei subsp. casei NRRL B1922 87

Lactobacillus delbrueckii subsp. bulgaricus NRRL B

548

33

Lactobacillus delbrueckii subsp. lactis MTCC 911 28

Lactobacillus fermentum MTCC 903 72

Lactobacillus johnsonii NRRL B2178 Nil

Lactobacillus plantarum MTCC 1325 14

Lactobacillus rhamnosus MTCC 1408 Nil

Lactobacillus sakei NRRL B1917 68

Lactococcus lactis subsp. cremoris MTCC 1484 42

Lactococcus lactis subsp. lactis MTCC 3041 11

Leuconostoc mesenteroides NRRL B640 92

Leuconostoc mesenteroides subsp. mesenteroides

MTCC 107

51

Pediococcus acidilactici NRRL B14009 65

Escherichia coli MTCC 443 05

Bacillus cereus MTCC 1305 17

Listeria monocytogenes Scott A 48

Staphylococcus aureus MTCC 96 14

Enterobacter aerogenes MTCC 2822 04

Enterococcus faecalis MTCC 439 06

NRRL: Northern Regional Research Laboratory, Peoria, IL, USA;

MTCC: Microbial Type Culture Collection, Institute of Microbial

Technology (IMTECH), Chandigarh, India.

Fig. 4. Antimicrobial activity of the culture filtrate of select LAB isolates

obtained from fermented cucumber against pathogenic bacteria. (A)

Escherichia coli MTCC 433; (B) Bacillus cereus MTCC 1305; (C)

Staphylococcus aureus MTCC 96 and (D) Listeria monocytogenes ScottA.

Bacteriocin activity is revealed as a zone of growth inhibition around the

well.

A.K. Singh, A. Ramesh / Food Microbiology 25 (2008) 278–287286

were antagonistic to Gram-positive pathogens such asBacillus cereus MTCC 1305 and S. aureus MTCC 96(Table 5 and Fig. 4). Interestingly, bacteriocin present inthe culture filtrate of few isolates could also hinder thegrowth of Gram-negative bacteria such as E. coli MTCC443. Bacteriocins working against Gram-positive patho-genic bacteria have been reported earlier (Albano et al.,2007). In our work we could obtain bacteriocin producers,

which were antagonistic to even Gram-negative pathogenlike E. coli. The broad spectrum bacteriocin-producingLAB cultures obtained in our investigation can find usefulapplications in food fermentation and preservation pro-cesses. Currently, biochemical characterization of thesebacteriocins is underway.

4. Conclusion

The present investigation clearly demonstrates thestrength of the PCR-based approach in simultaneouslydetecting major LAB groups in fermenting cucumber. Thesalient achievements of the investigation include: (a)Judicious design of genus-specific and bacteriocin-specificprimers in conjunction with a simple DNA extractionmethod enabled detection of LAB genera as well asmultiple bacteriocin producers in fermented samples.(b) Time-dependent succession of LAB genera in thefermenting samples were monitored and the succession ofthe LAB population could be rationally linked to theemergence of bacteriocin-producing strains. (c) The meth-od was highly sensitive and specially proved useful todetect small numbers of the initial LAB population presentin cucumber and also demonstrate their bacteriocin-producing ability. (d) Qualitative differences could beobserved in the culture-independent and -dependentmethods thus accounting for the biased selection usuallyassociated with enrichment techniques. The PCR-basedapproach not only provided a comprehensive snapshot ofLAB succession in fermenting cucumber samples but alsorevealed the possibility of designing strategic methods ofisolating bacteriocin-producing LAB from natural fermen-tation processes. Interestingly, a large number of antag-onistic LAB were obtained with potent activity againstpathogenic bacteria. On further characterization, thesenatural isolates can either be used as starter cultures ortheir antagonistic attributes can be genetically transferredto construct competitive LAB starters and achieve superiorshelf life of fermented food products.

Acknowledgments

A.R. thanks Ministry of Human Resource Development(MHRD), Govt. of India for financial support in TATproject. A.K.S. thanks Indian Institute of TechnologyGuwahati for a doctoral fellowship.

References

Albano, H., Todorov, S.D., van Reenen, C.A., Hogg, T., Dicks, L.M.T.,

Teixeira, P., 2007. Characterization of two bacteriocins produced by

Pediococcus acidilactici isolated from ‘‘Alheira,’’ a fermented sausage

traditionally produced in Portugal. Int. J. Food Microbiol. 116,

239–247.

Brusetti, L., Borin, S., Mora, D., Rizzi, A., Raddadi, N., Sorlini, C.,

Daffonchio, D., 2006. Usefulness of length heterogeneity-PCR for

monitoring lactic acid bacteria succession during maize ensiling.

FEMS Microbiol. Ecol. 56, 154–164.

ARTICLE IN PRESSA.K. Singh, A. Ramesh / Food Microbiology 25 (2008) 278–287 287

Caplice, E., Fitzgerald, G.F., 1999. Food fermentations: role of

microorganisms in food production and preservation. Int. J. Food

Microbiol. 50, 131–149.

Daeschel, M.A., Anderson, R.E., Fleming, H.P., 1987. Microbial ecology

of fermenting plant materials. FEMS Microbiol. Rev. 46, 357–367.

Delbe’s, C., Ali-Mandjee, L., Montel, M.C., 2007. Monitoring bacterial

communities in raw milk and cheese by culture-dependent and

-independent 16S rRNA gene-based analyses. Appl. Environ. Micro-

biol. 73, 1882–1891.

Drider, D., Fimland, G., Hechard, Y., Mcmullen, L.M., Prevost, H., 2006.

The continuing story of Class IIa bacteriocins. Microbiol. Mol. Biol.

Rev. 70, 564–582.

Dubernet, S., Desmasures, N., Gueguen, M., 2002. A PCR-based method

for identification of lactobacilli at the genus level. FEMS Microbiol.

Lett. 214, 271–275.

Ennahar, S., Aoude-Werner, D., Sorokine, O., van Dorsselaer, A.,

Bringel, F., Hubert, J.C., Hasselmann, C., 1996. Production of

pediocin AcH by Lactobacillus plantarum WHE92 isolated from

cheese. Appl. Environ. Microbiol. 62, 4381–4387.

Ennahar, S., Sashihara, T., Sonomoto, K., Ishizaki, A., 2000. Class IIa

bacteriocins: biosynthesis, structure and activity. FEMS Microbiol.

Rev. 24, 85–106.

Ercolini, D., 2004. PCR-DGGE fingerprinting: novel strategies for

detection of microbes in food. J. Microbiol. Methods 56, 297–314.

Fontana, C., Vignolo, G.T., Cocconcelli, P.S., 2005. PCR-DGGE analysis

for the identification of microbial populations from Argentinean dry

fermented sausages. J. Microbiol. Methods 63, 254–263.

Halami, P.M., Ramesh, A., Chandrashekar, A., 2005. Fermenting

cucumber, a potential source for the isolation of pediocin like

bacteriocin producers. World J. Microbiol. Biotechnol. 21, 1351–1358.

Harris, L.J., 1998. The microbiology of vegetable fermentations. In:

Wood, B.J.E. (Ed.), Microbiology of Fermented Foods, vol. 1, second

ed. Blackie Academic & Professional, London, pp. 45–72.

Jagannath, A., Ramesh, A., Ramesh, M.N., Chandrashekar, A.,

Varadaraj, M.C., 2001. Predictive model for the behavior of Listeria

monocytogenes ScottA in Shrikhand, prepared with a biopreservative

pediocin K7. Food Microbiol. 18, 335–343.

Kim, M., Chun, J.T., 2005. Bacterial community structure in kimchi, a

Korean fermented vegetable food, as revealed by 16S rRNA gene

analysis. Int. J. Food Microbiol. 103, 91–96.

Lazzi, C., Rossetti, L., Zago, M., Neviani, E., Giraffa, G., 2004.

Evaluation of bacterial communities belonging to natural whey

starters for Grana Padano cheese by length heterogeneity-PCR.

J. Appl. Microbiol. 96, 481–490.

Leroy, F., De Vuyst, L., 2004. Lactic acid bacteria as functional starter

cultures for the food fermentation industry. Trends Food Sci. Technol.

15, 67–78.

Randazzo, C.L., Vaughan, E.E., Caggia, C., 2006. Artisanal and

experimental Pecorino Siciliano cheese: microbial dynamics during

manufacture assessed by culturing and PCR-DGGE analyses. Int. J.

Food Microbiol. 109, 1–8.

Rantsiou, K., Cocolin, L., 2006. New developments in the study of the

microbiota of naturally fermented sausages as determined by

molecular methods: a review. Int. J. Food Microbiol. 108, 255–267.

Rodrı́guez, J.M., Martı́nez, M.I., Kok, J., 2002. Pediocin PA-1, a wide-

spectrum bacteriocin from lactic acid bacteria. Crit. Rev. Food Sci.

Nutr. 42, 91–121.

Sambrook, J., Russell, D.W., 2001. Molecular Cloning: A Laboratory

Manual, third ed. Cold Spring Harbor, Cold Spring Harbor

Laboratory, New York.

Sandhu, K.S., Shukla, F.C., 1996. Methods for pickling of cucumber

(Cucumis sativus)—a critical appraisal. J. Food Technol. 33,

455–473.

Settanni, L., van Sinderen, D., Rossi, J., Corsetti, Aldo., 2005. Rapid

differentiation and in situ detection of 16 sourdough Lactobacillus

species by multiplex PCR. Appl. Environ. Microbiol. 71, 3049–3059.

Tamang, J.P., Tamang, B., Schillinger, U., Franz, C.M.A.P., Gores, M.,

Holzapfel, W.H., 2005. Identification of predominant lactic acid

bacteria isolated from traditionally fermented vegetable products of

the Eastern Himalayas. Int. J. Food Microbiol. 105, 347–356.