lable at ScienceDirect

Food Microbiology 26 (2009) 504–513

Contents lists avai

Food Microbiology

journal homepage: www.elsevier .com/locate/ fm

Evaluation of a facile method of template DNA preparation for PCR-baseddetection and typing of lactic acid bacteria

Atul Kumar Singh, Aiyagari Ramesh*

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

a r t i c l e i n f o

Article history:Received 20 October 2008Received in revised form28 February 2009Accepted 10 March 2009Available online 21 March 2009

Keywords:Lactic acid bacteriaTemplate DNA isolationUrea–SDS–NaOHPolymerase chain reactionFood matrixTyping

* Corresponding author. Tel.: þ91 361 258 2205; faE-mail address: aramesh@iitg.ernet.in (A. Ramesh

0740-0020/$ – see front matter � 2009 Elsevier Ltd.doi:10.1016/j.fm.2009.03.006

a b s t r a c t

The objective of our investigation was to develop a convenient and reliable method of generatingtemplate DNA for routine PCR-based detection and typing of lactic acid bacteria (LAB). Template DNAextracted from Lactobacillus, Lactococcus, Pediococcus and Leuconostoc using a combination of urea, SDSand NaOH yielded amplicons of expected size in PCR with genus-specific primers. Apart from LAB, theproposed method could also be adopted to generate PCR-compatible template DNA from a number ofGram-positive and Gram-negative bacterial strains. DNA template prepared by the proposed methodfrom various standard strains of Lactobacillus sp. also generated discriminating fingerprints with BOXA1Rprimer in rep-PCR. A significant finding of the investigation was that a comparable banding profile of LABstrains was obtained in rep-PCR using template DNA prepared by urea–SDS–NaOH method anda commercially available DNA isolation kit. This was further evidenced by high dice coefficient valuesobtained in the range of 81.8–96.7 when cluster analysis was performed by UPGAMA method. Theapplication potential of this DNA extraction method for PCR-based direct detection of LAB in fermentedfood samples such as dahi, idli batter and salt-fermented cucumber was validated by detecting specificamplicons of LAB genera in the fermented samples. The applicability of the proposed template DNAextraction method was further substantiated when 29 bacteriocinogenic LAB strains (Bacþ) previouslydetected in salt-fermented cucumber by PCR [Singh, A.K., Ramesh, A., 2008. Succession of dominant andantagonistic lactic acid bacteria in fermented cucumber: Insights from a PCR-based approach. Food.Microbiol. 25, 278–287] generated differentiating fingerprints in BOX element based rep-PCR and formedclusters with reference LAB strains.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Lactic acid bacteria (LAB) belong to the low GþC group ofmicroaerophilic Gram-positive bacteria and assume a vital role infood fermentation. The starter culture and flavor forming traits ofLAB have been recognized as cardinal features, which have signif-icantly contributed to their widespread use in food fermentationprocesses (Coolbear et al., 2008;Di Cagno et al., 2008; Leroy and DeVuyst, 2004; van Hylckama Vlieg and Hugenholtz, 2007). LAB arealso known for probiotic attributes and can influence the activityand composition of the intestinal microbiota (Ouwehand et al.,2002; Saito, 2004). Apart from the starter culture and probioticfeatures, the capacity of LAB to produce antimicrobial agents suchas bacteriocin and inhibit foodborne pathogens has reiterated theirpromise and strengthened their application potential in foodfermentation processes and biopreservation (Castellano et al.,

x: þ91 361 258 2249.).

All rights reserved.

2008; Drider et al., 2006; Jagannath et al., 2001; Jones et al., 2008;Settani and Corsetti, 2008).

The choice of an appropriate LAB strain for food fermentationdemands critical attention to strain identification. Besides, the needto characterize LAB strains is justified since estimating the geneticdiversity of strains and appraising the fate of LAB starter culturesare vital in the context of defining technological parameters forfood fermentation processes. The classical approaches based onmorphological and biochemical identification of LAB strains arearduous and error-prone. Alternative DNA based techniquesinclude 16S rRNA probes, plasmid profiles, and pulsed-field gelelectrophoresis (Cusick and O’Sullivan, 2000; Heilig et al., 2002;Ventura and Zink, 2002). However, these methods are cumbersomefor routine large-scale application. Polymerase chain reaction(PCR)-based methods offer rapid and convenient options for LABstrain identification, monitoring of starter cultures, detection ofantagonistic LAB and molecular typing (Coudeyras et al., 2008;Nielsen et al., 2007; Settani et al., 2005; Singh and Ramesh, 2008;Yost and Nattress, 2002).

Table 1Reference bacterial strains used in the present investigation.

Bacteria Strain

Lactobacillus acidophilus MTCC 447Lactobacillus brevis NRRL B4527, NCIM 2090Lactobacillus casei NCIM 2151, MTCC 1423Lactobacillus casei subsp. casei NRRL B1922Lactobacillus delbrueckii subsp. delbrueckii NCIM 2025Lactobacillus delbrueckii subsp. bulgaricus NRRL B548Lactobacillus delbrueckii subsp. lactis MTCC 911Lactobacillus fermentum MTCC 903Lactobacillus gasseri NRRL B4240, NRRL B14168Lactobacillus helveticus NCIM 2126Lactobacillus johnsonii NRRL B2178Lactobacillus lactis NCIM 2368Lactobacillus paracasei subsp. paracasei NRRL B4564Lactobacillus plantarum MTCC 1325, MTCC 1746,

MTCC 1407, NCIM 2083, NCIM 2592Lactobacillus reuteri NRRL B14171Lactobacillus rhamnosus MTCC 1408, NRRL B442Lactobacillus sakei NRRL B1917Lactobacillus salivarius subsp. salivarius NRRL B1949Lactococcus lactis subsp. chacetylactis MTCC 3042Lactococcus lactis subsp. lactis MTCC 440, MTCC 3041, MTCC 3038Leuconostoc mesenteroides NRRL B640Leuconostoc mesenteroidessubsp. mesenteroides

MTCC 107

Leuconostoc oenos NCIM 2219Pediococcus acidilactici CFR K7a NRRL B14009, NRRL B14958,

NRRL B1153Pediococcus pentosaceus NCIM 2295Bacillus cereus MTCC 1305Bacillus subtilis MTCC 441Enterococcus faecalis MTCC 439Listeria monocytogenes Scott AMicrococcus luteus NCIM 2704Staphylococcus aureus MTCC 96Enterobacter aerogenes MTCC 2822Escherichia coli MTCC 118, MTCC 443Pseudomonas aeruginosa MTCC 2488Yersinia enterocolitica MTCC 859

NRRL: Northern Regional Research Laboratory, Peoria, IL, USA; MTCC: MicrobialType Culture Collection, Institute of Microbial Technology (IMTECH), Chandigarh,India; NCIM: National Collection of Industrial Microorganism, National ChemicalLaboratory (NCL), Pune, India.

a Culture provided by Dr. Prakash Halami, Central Food Technological ResearchInstitute (CFTRI), Mysore, India.

A.K. Singh, A. Ramesh / Food Microbiology 26 (2009) 504–513 505

The successful application of PCR-based techniques for LABdetection and typing essentially depends on the quality of thetemplate DNA. Hence, efficient extraction of template DNA fromLAB strains is a decisive factor to confer reliability to PCR-basedanalysis. The crux of the problem is that LAB being Gram positivepossess a recalcitrant cell wall, necessitating the use of cell lyticenzymes like lysozyme (Chagnaud et al., 2001). In other instances,protocols for efficient DNA isolation from LAB report the use ofchelex resins, detergents like CTAB, nuclease inhibitors likeproteinase K and organic solvents like phenol–chloroform(Dubernet et al., 2002; Giraffa et al., 2000). However, the afore-mentioned interventions make the entire process laborious andcost intensive. Besides, the presence of even trace amounts ofthese reagents as contaminants in the template DNA sample is ofgreat concern and could significantly hamper subsequent PCR-based analysis. Another cardinal issue is the fact that LAB strainsare known to be diverse and it is quite likely that a variety of nativeLAB isolates with different cell wall characteristics could beencountered in food samples. Hence, the DNA extraction methodshould be able to address this issue and readily provide PCR-compatible templates from a wide range of native LAB strains. It isalso envisaged that the DNA extraction method should have wide-scale adaptability in detecting LAB present in food samples bygenerating inhibitor-free PCR-compatible templates. Lastly,exploitation of the DNA extraction method to generate templatesthat facilitate PCR-based fingerprinting of native LAB strains willalso be an added advantage in estimating the genetic diversity ofthe strains.

In the present investigation our aim was to develop a robustmethod of DNA extraction which generates PCR-compatibletemplates from various LAB strains, with minimal sample manip-ulation steps. A combination of urea, SDS and NaOH was able torapidly generate PCR-compatible templates and allowed us todetect and type various LAB strains. Interestingly, the proposedmethod of DNA preparation exhibited broad-spectrum utility as wewere able to generate PCR-compatible template DNA from a widerange of Gram-positive and Gram-negative bacteria. The reliabilityof the method was established as the fingerprinting patternsobtained from the templates from select LAB strains were compa-rable to that obtained using a commercially available DNA isolationkit. We have also demonstrated the utility and application potentialof this method in PCR-based detection of LAB strains from fer-mented food samples and in estimating the genetic diversity ofbacteriocin producing (Bacþ) native LAB isolates obtained fromsalt-fermented cucumber.

2. Materials and methods

2.1. Bacterial strains and culture conditions

The reference strains of LAB and other Gram-positive andGram-negative bacteria used in the present investigation areshown in Table 1. All LAB strains were maintained as frozen stocksin milk and glycerol (10% each) at �20 �C. The strains were sub-cultured every fortnight by adding 0.1 ml of the stock culture to5.0 ml of de Man, Rogosa and Sharpe (MRS) broth (HiMedia,Mumbai, India) and growing the cells at 37 �C in a static incubator.Strains of Enterococcus faecalis, Listeria monocytogenes, Staphylo-coccus aureus and Yersinia enterocolitica were propagated aerobi-cally at 37 �C in brain–heart infusion (BHI) broth (HiMedia, India)whereas Bacillus cereus, Bacillus subtilis, Enterobacter aerogenes,Escherichia coli, Micrococcus luteus and Pseudomonas aeruginosawere grown aerobically at 37 �C in nutrient broth (NB) medium(HiMedia, India).

2.2. Template DNA preparation from bacterial strains

DNA was isolated from LAB and other bacterial strains listed inTable 1. The bacterial strains were grown for 6 h under the specifiedgrowth conditions mentioned before, prior to isolation of templateDNA. The template DNA isolation protocol was essentially based onan earlier method (Ramesh et al., 2002) with some modifications,and the scheme is outlined in Fig. 1. The purity of the isolatedtemplate DNA was ascertained for select strains of LAB bymeasuring A260/A280 (Sambrook and Russell, 2001).

2.3. Evaluation of urea–SDS–NaOH method of templateDNA isolation

The ability of the urea–SDS–NaOH method of DNA isolation togenerate PCR-compatible templates from select strains of LAB wasevaluated by using 2.0 ml of the extracted DNA sample in conven-tional PCR along with genus-specific primers for LAB. The univer-sality of the DNA isolation method was ascertained by extractingDNA from LAB strains as well as various Gram-positive and Gram-negative bacterial strains (Table 1) and performing PCR usinguniversal primers for 16S rRNA gene. Template DNA extracted from25 standard strains of Lactobacillus was also used in fingerprintinganalysis by performing rep-PCR with BOX element specific primer.

Fig. 1. Scheme of template DNA isolation.

A.K. Singh, A. Ramesh / Food Microbiology 26 (2009) 504–513506

In another set of experiments, DNA was isolated from select LABstrains using the urea–SDS–NaOH method as well as a GenElutebacterial genomic DNA isolation kit (Sigma, USA). A conventionalPCR using universal primers for 16S rRNA gene and rep-PCR withBOX element primer was performed with the templates generatedby both methods. The quality of amplicons and fingerprintingprofiles obtained from LAB strains were compared. In the case ofrep-PCR, reproducibility of the fingerprint profiles was also evalu-ated by extracting DNA separately from triplicate samples andperforming PCR.

Table 2Primers used in the present investigation.

Primer Nucleotide sequence (50–30) PC

U1F AGAGTTTGATCCTGGCTCAG UnU1R GGTTACCTTGTTACGACTT1F AGAAGAGGACAGTGGAAC Lac1R TTACAAACTCTCATGGTGTG2F TAAAGCGAGCGCAGGTGG Lac2R GGTTACCTTGTTACGACTT3F CTGAATGAGATTTTAACACG Pe3R GGTTTTAAGAGATTAGCT4F AGAGATGGATCCGCGGTGCA Leu4R TTACAAACTCCCATGGTGTGBOXA1R CTACGGCAAGGCGACGCTGACG Int

2.4. PCR conditions and fingerprinting profiles

PCR amplification was performed in a total reaction volume of25.0 ml. The proportions of PCR components and cycle parameterswere essentially same as mentioned before (Singh and Ramesh,2008). The sequence of universal primers for bacteria, genus-specific primers for Lactobacilli, Lactococci, Pediococci and Leuco-nostoc and conserved BOX element specific primers are indicated inTable 2. A total of 35 amplification cycles were performed usinga programmable thermal cycler (Gene Amp Gold PCR System,

R target Reference

iversal-16S rRNA Weisburg et al. (1991)

tobacillus-16S rRNA Singh and Ramesh (2008)

tococcus-16S rRNA Singh and Ramesh (2008)

diococcus-16S rRNA Singh and Ramesh (2008)

conostoc-16S rRNA Singh and Ramesh (2008)

erspersed repetitive DNA element Versalovic et al. (1994)

Fig. 2. Agarose gel electrophoresis of PCR amplicons obtained with genus-specificprimers targeting 16S rRNA gene of LAB strains. Lanes: (M) l DNA EcoRI/HindIII doubledigest size marker; (1) Lactobacillus acidophilus MTCC 447; (2) Lactobacillus casei NCIM2151; (3) Lactobacillus helveticus NCIM 2126; (4) Lactobacillus gasseri NRRL B4240; (5, 6and 7) Lactococcus lactis subsp. lactis MTCC 440, MTCC 3041 and MTCC 3038; (8)Lactococcus lactis subsp. chacetylactis MTCC 3042; (9, 10 and 11) Pediococcus acidilacticiCFR K7, NRRL B1153 and NRRL B14958; (12) Pediococcus pentosaceus NCIM 2295; (13)Leuconostoc oenos NCIM 2219; (14) Leuconostoc mesenteroides NRRL B640; (15) Leu-conostoc mesenteroides subsp. mesenteroides MTCC 107.

A.K. Singh, A. Ramesh / Food Microbiology 26 (2009) 504–513 507

Applied Biosystems, USA). Primer annealing temperatures were setat 55 �C for 1 min for genus-specific and universal primers and50 �C for 1 min in case of BOX element based primer. The PCRproducts were analyzed by agarose (0.8%) gel electrophoresis(Sambrook and Russell, 2001). Analysis of the fingerprintingprofiles obtained in rep-PCR was accomplished by using QuantityOne software (Bio-Rad, USA). Similarity of the profiles was evalu-ated by obtaining dice coefficient values and the phylogenetic treewas drawn on the basis of unweighted pair group method usingarithmetic averages (UPGAMA).

2.5. Sensitivity of detection

Template DNA was extracted by urea–SDS–NaOH method froma varying cell number (106–101 cfu/ml) of standard LAB strainsbelonging to genus Lactobacillus, Pediococcus, Lactococcus andLeuconostoc. PCR was then performed in conjunction with genus-specific 16S rRNA primers. PCR reaction and agarose gel electro-phoresis were carried out according to the method mentioned inSection 2.4. The amplicon intensity was quantified by a plot profileanalysis with NIH ImageJ software (http://rsb.info.nih.gov/ij).

2.6. LAB detection in fermented samples

Five samples each of dahi (an indigenous lactic cultured milkproduct), idli batter (fermented batter produced from rice powderand dehulled black gram, which is steamed to prepare idli) andcucumber samples fermented in 4% saline solution for 72 h (Singhand Ramesh, 2008) were chosen as model system for theseexperiments. The following steps were adopted for direct extrac-tion of DNA from the samples. Aliquots of each sample (0.3 ml ofdahi and salt-fermented cucumber and 0.3 g in case of idli batter)were taken in separate tubes and mixed with 0.7 ml of phosphatebuffered saline containing 0.5% Tween 20 (PBS-T). The sampleswere thoroughly vortexed for 10 min and centrifuged at low speedof 450 � g for 2 min at room temperature to separate the foodmatrix. The supernatant containing bacterial cells was collectedand centrifuged at 8000 � g for 5 min at room temperature and thecell pellet was processed for template DNA extraction using themethod outlined in Fig. 1. The extracted template DNA was sub-jected to PCR with Lactobacillus, Lactococcus, Pediococcus and Leu-conostoc genus-specific primers.

2.7. Template DNA extraction for molecular fingerprinting ofbacteriocinogenic LAB

The application potential of the template DNA extractionmethod for molecular typing analysis was also tested for 29bacteriocin producing (Bacþ) LAB strains obtained earlier from salt-fermented cucumber (Singh and Ramesh, 2008). DNA templatesfrom Bacþ LAB were prepared by urea–SDS–NaOH method and rep-PCR was performed with BOXA1R element primer. Standard strainsof Lactobacillus and Pediococcus were included as reference strainsto compare the fingerprinting profiles. Cluster analysis andconstruction of phylogenetic tree was accomplished as statedbefore.

3. Results and discussion

3.1. Template DNA preparation

The scheme of urea–SDS–NaOH method of template DNApreparation used in the present investigation is shown in Fig. 1.DNA was isolated from 1.0 ml of 6 h grown culture (w106 cfu) ofvarious LAB strains as well as other Gram-positive and Gram-

negative bacteria (Table 1). Our aim was to establish a convenientDNA extraction method using readily available and cost-effectivereagents. We also wanted to ensure that the method is rapid,involving minimal sample manipulation steps and applicable tolarge number of bacterial strains with varying cell wall composi-tion. A major concern was to obtain PCR-compatible DNA templatesthat facilitate LAB identification and typing analysis. The primarychallenge in bacterial DNA extraction is effective cell lysis. Based onan earlier study where urea was employed for DNA extraction frommilk samples (Ramesh et al., 2002) we selected urea as a chaotropicagent and induced cell lysis. It is known that urea can promotedenaturation of proteins by destabilizing protein–water interac-tions. Hydrophobic effects are the main determinants of urea-induced protein denaturation and urea can preferentially solvatenon-polar and aromatic amino acid residues, as well as the peptidebackbone in proteins (Stumpe and Grubmuller, 2007). The overallweakening of the protein–water interactions by urea may possiblyrender denaturation of bacterial membrane proteins by exposure ofthe hydrophobic core and dissociation of secondary structureelements of membrane proteins. The destabilization of the bacterialcell membrane was further propagated by inclusion of a stronganionic detergent such as SDS. As a consequence of exposure tourea and SDS, and the subsequent boiling step, the bacterial cellmembrane was weakened considerably. The fragile cells were thenreadily lysed on exposure to 0.2 N NaOH resulting in effectiverelease of DNA from the bacterial strain. The denatured proteinsand other cellular debris were removed by centrifugation. Theextracted DNA was then precipitated in presence of absolutealcohol. A260/280 measurements for the extracted template DNAfrom select LAB were in the range of 1.87–1.92, indicating a lack ofprotein contamination in the samples.

3.2. PCR-based detection of LAB and other bacterial strains

Template DNA was isolated from reference LAB strains using theproposed method and PCR-compatibility of the extracted templateDNA was assessed by using genus-specific primers for Lactobacillus,Lactococcus, Pediococcus and Leuconostoc. Fig. 2 depicts the ampli-cons obtained from PCR reactions. It is evident that distinct

A.K. Singh, A. Ramesh / Food Microbiology 26 (2009) 504–513508

amplicons were obtained from the LAB strains. The sizes of theamplicons were around 800, 960, 1200 and 1100 bp, which coin-cided with the expected amplicon size of the LAB strains asreported in an earlier investigation (Singh and Ramesh, 2008). Apositive PCR reaction demonstrated that the urea–SDS–NaOHmethod was effective in generating PCR-compatible templates.Previous reports on PCR-based LAB detection describe the appli-cation of specialized resins to extract bacterial DNA such as chelexresin (Giraffa et al., 2000) or the use of strong detergents such asCTAB (Dubernet et al., 2002; Elegado et al., 2004), zirconium beads(Delbes et al., 2007) and cell lytic enzymes such as lysozyme(Chagnaud et al., 2001; Dubernet et al., 2002). The use of nucleaseinhibitors like proteinase K and deproteinizing agents such asphenol–chloroform have also been reported (Dubernet et al., 2002;Sharma and Singh, 2005). In contrast, our method of DNA extrac-tion from LAB is relatively simple and avoids the use of strongorganic solvents such as phenol–chloroform. Organic solvents arepotential inhibitors of the PCR reaction and even trace amounts aresufficient to completely preclude the amplification process. In ourmethod, we have eliminated this possibility by completely avoidingthe use of organic solvents. Additionally, a distinct advantage of thismethod is that the reagents used are readily available and inex-pensive. Further, the method is quite rapid considering the rela-tively short time taken to prepare the template DNA. In the presentstudy, the observation that amplification could be achieved fromDNA templates prepared from LAB strains belonging to various

Fig. 3. UPGAMA based cluster analysis for standard strains of Lactobacillus on the basis ofbased primer.

genera indicated that the DNA isolation method was not hamperedby the choice of the strains.

The versatility of the template DNA preparation method wassubstantiated when positive amplification was obtained fromselected LAB and non-LAB bacterial strains with universal primersfor 16S rRNA gene. Amplicons of expected size (w1485 bp) wereobtained from various LAB as well as non-LAB strains (gel picturenot shown). It may be mentioned here that the selection of thesebacterial strains encompassed a combination of Gram-positive andGram-negative bacteria, with evidently varying cell wall composi-tion. The rationale of the selection of non-LAB strains was todemonstrate the potential of the DNA extraction method inproviding PCR-compatible templates from these strains. Foodbornepathogenic bacterial strains such as Bacillus cereus, Staphylococcusaureus, Listeria monocytogenes, Enterococcus faecalis, and Yersiniaenterocolitica are often present in fermented food products andhence need to be detected to ensure food safety and quality control.In this context, our method of DNA extraction is indeed handy inreadily providing PCR-compatible templates and is thus a prom-ising tool in molecular detection of LAB as well as foodbornepathogenic strains. Interestingly, PCR-compatible template DNAcould be readily extracted even from bacterial strains such asStaphylococcus aureus MTCC 96, which is known to possessa recalcitrant cell wall and thus cell lytic enzymes such as lysos-taphin, are often used in the DNA isolation protocol (Hein et al.,2005; Padmapriya et al., 2003).

dice coefficient values of banding pattern generated in rep-PCR with BOXA1R element

Table 3Dice coefficient based similarity analysis of fingerprint profilesa from DNA templatesof LAB strains.

LAB strain No. of bandsfrom template Ab

No. of bandsfrom template Bc

Dice coefficientvalued

Leuconostoc oenosNCIM 2219

6.0 5.0 81.8

Leuconostoc mesenteroidesNRRL B640

4.0 4.0 95.0

Lactobacillus casei subsp.casei NRRL B1922

5.0 5.0 94.0

Lactobacillus gasseriNRRL B4240

7.0 7.0 96.5

Lactobacillus helveticusNCIM 2126

6.0 7.0 88.4

Lactobacillus plantarumMTCC 1325

7.0 7.0 85.2

Lactobacillus reuteriNRRL B14171

6.0 5.0 85.7

Lactobacillus salivarius subsp.salivarius NRRL B1949

4.0 4.0 90.2

Lactococcus lactis subspchacetylactis MTCC 3042

5.0 5.0 96.7

Pediococcus acidilactici CFR K7 4.0 4.0 92.3Pediococcus acidilactici

NRRL B149583.0 3.0 79.4

a Fingerprinting profile obtained by rep-PCR with BOXA1R primer.b Template A prepared by GenElute bacterial genomic DNA isolation kit (Sigma,

USA).c Template B prepared by urea–SDS–NaOH method.d Dice coefficient values for matched banding patterns obtained with template A

and B from LAB strain.

A.K. Singh, A. Ramesh / Food Microbiology 26 (2009) 504–513 509

We have also ascertained the PCR compatibility of the templateDNA prepared by the proposed urea–SDS–NaOH methodvis-a-vis DNA extraction using a commercially available bacterialgenomic DNA isolation kit. Template DNA prepared from selectstandard strains of Lactobacillus and Pediococcus using the urea–SDS–NaOH method as well as the kit were subjected to PCR inconjunction with universal primers for bacterial 16S rRNA gene.Amplicons of equivalent size (w1485 bp) and comparable intensitywere obtained when template DNA prepared by the proposedmethod and the kit were used in PCR (gel picture not shown). Theseresults indicated that the quality of template DNA prepared byurea–SDS–NaOH method was comparable to that obtained usingthe kit. This renders the proposed DNA method potentially appli-cable for routine PCR-based screening of LAB isolates circum-venting the need of expensive kit-based DNA isolation.

3.3. Fingerprinting profiles of Lactobacillus strains

Template DNA was extracted from 25 reference strains ofLactobacillus using the urea–SDS–NaOH method and subjected torep-PCR with primers for BOX elements. The fingerprinting profilesof LAB strains consisted of approximately 3–7 DNA bands as visu-alized in Fig. 3. For example, the number of prominent bandsobtained was 3 in the case of L. plantarum MTCC 1325, L. brevisNCIM 2090, L. acidophilus MTCC 447 and L. helveticus NCIM 2126.However, for some strains such as L. sakei NRRL B1917 and L. lactisNCIM 2368 the number of bands obtained was 5 and 7, respectively.The banding pattern also revealed that the approximate size of theobtained fragments was in the range of 0.6–5.0 kb. It is also evidentfrom Fig. 3 that discrete clusters based on dice coefficient valuescould be obtained for various standard Lactobacillus strains. Ataround 0.77 dice coefficient value, various species of Lactobacilluscould be discriminated. The similarity index for strains of L. gasseri(NRRL B14168 and NRRL B4240) was considerably high based ondice coefficient values (0.85) whereas for strains of L. plantarum(MTCC 1746, NCIM 2592, MTCC 1325 and NCIM 2083), dice coeffi-cient values were even greater (0.92). Consequently, they wereclustered in close vicinity. rep-PCR-based fingerprinting of LABstrains using BOXA1R primer has been reported earlier (De Angeliset al., 2007; Terzic-Vidojevic et al., 2007).

In our investigation, the overall clustering of reference strains ofLactobacillus is in conformity with the expected pattern based ongenetic relatedness amongst the various species of Lactobacillus. Forinstance, it is known that the L. casei group comprises of strainsbelonging to L. casei, L. paracasei and L. rhamnosus (Ryu et al., 2001;Song et al., 2000). As evident from Fig. 3, L. casei NRRL B1922 andNCIM 2151 share a reasonably high similarity as evident from a dicecoefficient value of 0.62, whereas L. rhamnosus NRRL B442 andL. casei MTCC 1423 exhibit a dice coefficient value of 0.56 and areclustered in close proximity. Further, L. johnsonii NRRL B2178revealed a high dice coefficient value of 0.75 with strains ofL. gasseri and clustered in close vicinity. It can be mentioned thatL. johnsonii and L. gasseri are genetically close and known to belongto the L. acidophilus group (Ryu et al., 2001). Similarly, strains ofL. brevis were clustered along with strains of L. plantarum which isin agreement with the taxonomic proximity of these strains on thebasis of their genetic relatedness (Makarova and Koonin, 2007). Inour study the unusual grouping observed in some cases whereinstrains belonging to a given species did not cluster together (forexample some strains of L. plantarum) indicates heterogeneityamongst these strains. Similar findings have also been reportedearlier (De Angelis et al., 2001).

The reproducibility of the banding pattern obtained with rep-PCR using BOX primer was also assessed for select Lactobacillusstrains. Triplicate samples of template DNA isolated separately from

each LAB strain was subjected to rep-PCR analysis. There was noqualitative difference in banding patterns and the subsequentcluster analysis was not affected (gel picture not shown). Thus thepresent method of DNA extraction was consistent in generatinga PCR-compatible template that facilitated reproducible moleculartyping analysis of LAB.

Template DNA extraction from various LAB strains was alsoaccomplished using the proposed method and a bacterial genomicDNA isolation kit as mentioned before. rep-PCR fingerprintinganalysis with BOXA1R primer was performed to compare thebanding patterns produced from the templates and ascertain theauthenticity of fingerprinting profiles. DNA fingerprints obtainedfrom select LAB strains were compared based on dice coefficientvalues. It is quite clear from Table 3 that nine standard LAB strainsout of a total of 11 clustered together with high dice coefficientvalues (>85.0). The highest dice coefficient value of 96.7 wasobtained for L. lactis subsp. chacetylactis MTCC 3042 followed byL. gasseri NRRL B4240 for which the dice coefficient value was 96.5.For some strains such as L. oenos NCIM 2219, L. helveticus NCIM 2126and L. reuteri NRRL B14171, the number of bands obtained from thetwo templates differed. However, the fingerprinting profiles forthese strains were still comparable as evident from dice coefficientvalues which were considerably high in the range of 81.8–88.4,whereas P. acidilactici NRRL B14958 revealed the lowest dice coef-ficient value of 79.4. Overall, the results depicted in Table 3 reflectsignificantly high dice coefficient values and thus strengthen thepossibility of using the urea–SDS–NaOH method of template DNApreparation for consistent and reliable PCR-based fingerprinting ofLAB.

3.4. Sensitivity of detection

Template DNA extracted by urea–SDS–NaOH method fromvarying cell number of LAB strains belonging to genus Lactobacillus,Pediococcus, Lactococcus and Leuconostoc was used in PCR toestablish the sensitivity of detection. A representative agarose gel

Fig. 4. Sensitivity of PCR-based detection of LAB. (A) Amplicons obtained from Lactobacillus casei subsp. casei NRRL B1922 (lanes 1–6: 106–101 cfu/ml) and Pediococcus acidilacticiNRRL B14009 (lanes 7–12: 106–101 cfu/ml) (B) Quantification of amplicons with NIH ImageJ software.

A.K. Singh, A. Ramesh / Food Microbiology 26 (2009) 504–513510

picture shown in Fig. 4A depicts the amplicons obtained from 106–101 cfu/ml of Lactobacillus casei subsp. casei NRRL B1922 and Ped-iococcus acidilactici NRRL B14009. It is quite evident from the figurethat there was a decrease in the amplicon intensity which wasproportional to the cell numbers. Quantitative analysis of theamplicons in Fig 4B reflects higher band intensity for Lactobacilluscasei subsp. casei NRRL B1922 in comparison to Pediococcus acid-ilactici NRRL B14009. The limit of detection for all the strains testedwas 102 cfu/ml. The achieved level of detection has significantimplications for routine PCR-based detection of LAB in fermentedfood samples.

Fig. 5. Agarose gel electrophoresis of amplicons obtained using template DNAextracted directly from fermented samples by urea–SDS–NaOH method. Lactobacillus(Lb), Lactococcus (Lc), Pediococcus (Ped) and Leuconostoc (Leu) genus-specific primersfor 16S rRNA gene was used in PCR. Lane M indicates l DNA EcoRI/HindIII double digestsize marker.

3.5. Application of template DNA extraction method

3.5.1. Direct detection of LAB in fermented samplesBased on the results obtained for pure cultures of various

standard LAB strains, we were encouraged to apply the DNAextraction method to fermented samples and ascertain the appli-cability of the method in facilitating PCR-based direct detection ofLAB in such samples. Template DNA was extracted from dahi, idlibatter and salt-fermented cucumber samples as mentioned inSection 2.6. Fig. 5 is a representative agarose gel picture whichshows the amplicons obtained in PCR using template DNA extrac-ted from fermented samples. It is evident that Lactobacillus andLactococcus could be detected in dahi samples whereas in idli batterLactobacillus, Pediococcus and Leuconostoc could be detected. Dahi isan indigenous lactic cultured milk product and the presence ofLactobacillus and Lactococcus has been reported earlier in dahisamples (Varadaraj et al., 1993). The prevalence of Lactobacillus,Pediococcus and Leuconostoc in idli batter is in agreement with theinherent LAB profile reported earlier for such samples (Blandinoet al., 2003). In the case of fermented cucumber samples, directDNA extraction followed by PCR enabled the detection of Lactoba-cillus and Pediococcus. This finding corresponds to our earlier studywherein PCR-based detection of Lactobacillus and Pediococcus wasdemonstrated for samples of salt-fermented cucumber (Singh andRamesh, 2008). Similar trend of LAB genus detection was observedfor all five samples of dahi, idli batter and fermented cucumber. Thelevel of detection in the fermented samples corresponds to a cellnumber which is either equivalent or greater than the limit ofdetection (102 cfu/ml), which was established in an earlier experi-ment (Fig. 4). This level of detection is comparable with an earlierinvestigation wherein LAB could be quantified by PCR in fermentedmilk product (Furet et al., 2004). The absence of certain LAB genusin the fermented samples could be attributed to either a very low

initial load of the cells or an overwhelming presence of competingDNA obtained from bacterial cells of other genus. Sequestration ofbacterial cells from food matrix is a crucial prerequisite that caninfluence the outcome of PCR-based culture-independent detec-tion. We have included a short pre-treatment of the fermentedsamples with PBS-T in which the presence of Tween-20 promotesefficient cell separation from the food matrix. PCR results obtainedin our experiments also point out that the template DNA extracteddirectly from fermented samples was either free from PCR inhibi-tors or the concentration of inhibitors, if any, was lower than thelevels which can inhibit the amplification reaction. Significantly,the template DNA prepared from fermented samples was also

Fig. 6. UPGAMA based cluster analysis of bacteriocin producing (Bacþ) Lactobacillus sp. and Pediococcus sp. isolates obtained from salt-fermented cucumber. Clusters were generatedon the basis of dice coefficient values of banding pattern obtained in rep-PCR with BOXA1R element based primer. Genus identity of Bacþ native LAB isolates is indicated inparentheses (Lactobacillus : Lb; Pediococcus: Ped).

A.K. Singh, A. Ramesh / Food Microbiology 26 (2009) 504–513 511

amicable for multiple LAB genus detection in PCR, emphasizing theanalytical value of the urea–SDS–NaOH method for multiplex LABdetection in food samples.

Further, we were also interested in investigating whether thePCR results obtained with template DNA isolated by the proposedurea–SDS–NaOH method was comparable to the results obtainedwhen alternate DNA isolation protocols were adopted. Suspen-sions of food samples (dahi, idli batter and cucumber) wereprepared in PBS-T as mentioned before and DNA was extractedfrom the samples using the methods reported by Chagnaud et al.(2001) and Dubernet et al. (2002). PCR with these templatesfailed to yield any amplicons with genus-specific primers. It maybe emphasized here that the aforementioned DNA extractionmethods were essentially applied on pure cultures of LAB by theauthors. The efficiency of these methods in providing inhibitor-free template DNA directly from food samples needs to be furtherinvestigated.

3.5.2. Molecular typing analysis of Bacþ LABIt is evident from the results obtained in the previous section

that the application potential of the proposed template DNAextraction method was validated by PCR-based detection ofmultiple LAB strains in liquid broth and fermented samples.Further, the same templates could also generate reliable finger-printing profiles for LAB strains. As an extension of this application,our next aim was to assess the utility of the DNA extraction methodin generating PCR-compatible templates from select bacteriocinproducing (Bacþ) LAB strains and determine their genetic diversityby rep-PCR. This study assumes significance in the context ofassigning taxonomic affiliation and identifying signature patternsfor various native Bacþ LAB strains.

The antagonistic LAB strains were previously isolated from salt-fermented cucumber and identified as Lactobacillus sp. and Ped-iococcus sp. by PCR (Singh and Ramesh, 2008). DNA was extractedfrom 29 Bacþ isolates as well as reference LAB strains of

A.K. Singh, A. Ramesh / Food Microbiology 26 (2009) 504–513512

Lactobacillus and Pediococcus genus by urea–SDS–NaOH method.PCR amplification was carried out with BOXA1R primer. The DNAfingerprints of native Bacþ isolates were compared with the refer-ence strains to calculate the dice coefficient values and generate theclusters (Fig. 6). It is evident that the bacteriocin producing LABstrains yielded profiles which resulted in the generation of discreteclusters. The number of bands obtained from the Bacþ LAB isolatesand standard LAB strains varied from 3 to 7. At 0.40 dice coefficientvalue eight distinct clusters could be delineated. The nativeantagonistic isolates obtained from fermented cucumber weredistributed in clusters II–VII. Clusters II, III and V solely consisted ofBacþ isolates which were identified as Lactobacillus sp. using genus-specific primers for 16S rRNA gene in our previous investigation(Singh and Ramesh, 2008). Cluster IV consisted of 7 LAB strains ofwhich 3 were Bacþ isolates. These isolates clustered with standardLactobacillus strains (Fig. 6). In this cluster isolate CUAK 13 revealedhigh similarity with L. paracasei subsp. paracasei NRRL B4564 (dicecoefficient value of 0.64). Cluster VI consisted of 11 LAB strains ofwhich 4 included Bacþ isolates (CUAK 6, 8, 7 and 55). It can beobserved from Fig. 6 that isolate CUAK 6 revealed high similaritywith L. reuteri NRRL B14171 (dice coefficient value of 0.64) whereasisolate CUAK 55 clustered with P. acidilactici CFR K7 (dice coefficientvalue of 0.83). Bacþ LAB strains which formed cluster VII consistedof Pediococcus sp. (9 isolates). One of the isolates CUAK 19 revealedhigh similarity with P. pentosaceus NCIM 2295 (dice coefficientvalue of 0.70). Cluster VIII consisted of 14 Lactobacillus standard LABstrains belonging to the Lactobacillus acidophilus group. The clustersobserved for standard LAB strains did not exactly match with theearlier profile shown in Fig. 3. The plausible explanation for thisdeviation is that a greater genotypic discrimination is likely to beobserved due to a higher number of strains used for profilecomparison. This observation has also been reported earlier (DeAngelis et al., 2001). The application of BOXA1R primers in rep-PCRfor typing of LAB strains has been demonstrated earlier in case ofdifferentiation of L. sanfranciscensis isolates (De Angelis et al., 2007)and for molecular typing of probiotic L. rhamnosus 35 (Coudeyraset al., 2008). The overall clustering observed for the native Bacþ LABisolates in our investigation conformed to their expected taxonomicunits. Thus the utility of the template DNA extraction method inestimating genetic diversity of native Bacþ strains was clearlydemonstrated as adequate discriminating fingerprinting patternswere obtained for the strains in rep-PCR.

4. Conclusion

In our investigation we have been able to demonstrate a simpleand effective method of generating PCR-compatible template DNAfrom various LAB as well as other bacterial strains. The method israpid, cost-effective and obviates the need of commonly used celllytic enzymes and strong organic solvents. Interestingly, the reli-ability of the proposed method was validated as templatesprepared from LAB strains by the urea–SDS–NaOH method anda commercially available kit yielded comparable results inconventional PCR as well as rep-PCR-based fingerprinting analysis.A major highlight of the investigation was the successful imple-mentation of the template DNA extraction method in facilitatingPCR-based direct detection of LAB in fermented samples. Anotherimportant application of the template DNA extraction method wasemphasized when we could extract PCR-compatible DNA frombacteriocin producing native LAB strains and assess their geneticdiversity by rep-PCR. The results obtained in this investigation arerewarding, and the DNA extraction method outlined in the presentinvestigation can be adopted on a routine basis in food microbi-ology laboratories for PCR-based detection of LAB in variousindigenous fermented foods as well as in studies that aim to

establish the genetic diversity of native LAB strains by PCR-basedfingerprinting.

Acknowledgements

A.R. thanks Ministry of Human Resource Development (MHRD),Govt. of India for financial support in the TAT project. A.K.S. thanksCouncil of Scientific and Industrial Research (CSIR), New Delhi,India for a Senior Research Fellowship.

References

Blandino, A., Al-Aseeri, M.E., Pandiella, S.S., Cantero, D., Webb, C., 2003. Cereal-based fermented foods and beverages. Food Res. Int 36, 527–543.

Castellano, P., Belfiore, C., Fadda, S., Vignolo, G., 2008. A review of bacteriocinogeniclactic acid bacteria used as bioprotective cultures in fresh meat produced inArgentina. Meat Sci. 79, 483–499.

Chagnaud, P., Machinis, K., Coutte, L.A., Marecat, A., Mercenier, A., 2001. Rapid PCR-based procedure to identify lactic acid bacteria: application to six commonLactobacillus species. J. Microbiol. Methods 44, 139–148.

Coolbear, T., Crow, V., Harnett, J., Harvey, S., Holland, R., Martley, F., 2008. Devel-opments in cheese microbiology in New Zealand-Use of starter and non-starterlactic acid bacteria and their enzymes in determining flavour. Int. Dairy J 18,705–713.

Coudeyras, S., Macrchandin, H., Fajon, C., Forestier, C., 2008. Taxonomic and strain-specific identification of probiotic strain Lactobacillus rhamnosus 35 within theLactobacillus casei Group. Appl. Environ. Microbiol. 74, 2679–2689.

Cusick, S.M., O’Sullivan, D.J., 2000. Use of a single, triplicate arbitrarily primed-PCRprocedure for molecular fingerprinting of lactic acid bacteria. Appl. Environ.Microbiol. 66, 2227–2231.

De Angelis, M., Corsetti, A., Tosti, N., Rossi, J., Corbo, M.R., Gobbetti, M., 2001.Characterization of non-starter lactic acid bacteria from Italian ewe cheesesbased on phenotypic, genotypic, and cell wall protein analyses. Appl. Environ.Microbiol. 67, 2011–2020.

De Angelis, M., Di Cagno, R., Gallo, G., Curci, M., Siragusa, S., Crecchio, C., Parente, E.,Gobbetti, M., 2007. Molecular and functional characterization of Lactobacillussanfranciscensis strains isolated from sourdoughs. Int. J. Food Microbiol. 114,69–82.

Delbes, C., Ali-Mandjee, L., Montel, M., 2007. Monitoring bacterial communities inraw milk and cheese by culture-dependent and -independent 16S rRNA gene-based analyses. Appl. Environ. Microbiol. 73, 1882–1891.

Di Cagno, R., Surico, R.F., Siragusa, S., De Angelis, M., Paradiso, A., Minervini, F., DeGara, L., Gobbetti, M., 2008. Selection and use of autochthonous mixed starterfor lactic acid fermentation of carrots, french beans or marrows. Int. FoodMicrobiol. 127, 220–228.

Drider, D., Fimland, G., Hechard, Y., Mcmullen, L.M., Prevost, H., 2006. Thecontinuing story of Class IIa bacteriocins. Microbiol. Mol. Biol Rev. 70, 564–582.

Dubernet, S., Desmasures, N., Gueguen, M., 2002. A PCR-based method for identi-fication of lactobacilli at the genus level. FEMS Microbiol. Lett. 214, 271–275.

Elegado, F.B., Guerra, M.A.R.V., Macayan, R.A., Mendoza, H.A., Lirazan, M.B., 2004.Spectrum of bacteriocin activity of Lactobacillus plantarum BS and finger-printing of RAPD-PCR. Int. J. Food Microbiol. 95, 11–18.

Furet, J., Quenee, P., Tailliez, P., 2004. Molecular quantification of lactic acid bacteriain fermented milk products using real-time quantitative PCR. Int. J. FoodMicrobiol. 97, 197–207.

Giraffa, G., Rossetti, L., Neviani, 2000. An evaluation of chelex-based DNA purifi-cation protocols for the typing of lactic acid bacteria. J. Microbiol. Methods 42,175–184.

Heilig, H.G.H.J., Zoetendal, E.G., Vaughan, E.E., Marteau, P., Akkermans, A.D.L., deVos, W.M., 2002. Molecular diversity of Lactobacillus spp. and other lactic acidbacteria in the human intestine as determined by specific amplification of 16Sribosomal DNA. Appl. Environ. Microbiol. 68, 114–123.

Hein, I., Jørgensen, H.J., Loncarevic, Semir, Wagner, M., 2005. Quantification ofStaphylococcus aureus in unpasteurised bovine and caprine milk by real-timePCR. Res. Microbiol. 156, 554–563.

Jagannath, A., Ramesh, A., Ramesh, M.N., Chandrashekar, A., Varadaraj, M.C., 2001.Predictive model for the behavior of Listeria monocytogenes ScottA in Shrik-hand, prepared with a biopreservative pediocin K7. Food Microbiol. 18,335–343.

Jones, R.J., Hussein, H.M., Zagorec, M., Brightwell, G., Tagg, J.R., 2008. Isolation oflactic acid bacteria with inhibitory activity against pathogens and spoilageorganisms associated with fresh meat. Food Microbiol. 25, 228–234.

Leroy, F., De Vuyst, L., 2004. Lactic acid bacteria as functional starter cultures for thefood fermentation industry. Trends Food Sci. Technol.15, 67–78.

Makarova, K.S., Koonin, E.V., 2007. Evolutionary genomics of lactic acid bacteria. J.Bacteriol 189, 1199–1208.

Nielsen, D.S., Teniola, O.D., Ban-Koffi, L., Owusu, M., Andersson, T.S., Holzapfel, W.H.,2007. The microbiology of Ghanaian cocoa fermentations analysed usingculture-dependent and culture-independent methods. Int. J. Food Microbiol.114, 168–186.

A.K. Singh, A. Ramesh / Food Microbiology 26 (2009) 504–513 513

Ouwehand, A.C., Salminen, S., Solauri, E., 2002. Probiotics: An overview of beneficialeffects. Antonie Van Leeuwenhoek 82, 279–289.

Padmapriya, B.P., Ramesh, A., Chandrashekar, A., Varadaraj, M.C., 2003. Staphylo-coccal accessory gene regulator (sar) as a signature gene to detect enter-otoxigenic staphylococci. J. Appl. Microbiol. 95, 974–981.

Ramesh, A., Padmapriya, B.P., Chandrashekar, A., Varadaraj, M.C., 2002. Applicationof a convenient DNA extraction method and multiplex PCR for the directdetection of Staphylococcus aureus and Yersinia enterocolitica in milk samples.Mol. Cell. Probes 16, 307–314.

Ryu, C.S., Czajka, J.W., Sakamoto, M., Benno, Y., 2001. Characterization of theLactobacillus casei group and the Lactobacillus acidophilus group by automatedribotyping. Microbiol. Immunol 45, 271–275.

Saito, T., 2004. Selection of useful probiotic lactic acid bacteria from the Lactoba-cillus acidophilus group and their applications to functional foods. Animal. Sci. J75, 1–13.

Sambrook, J., Russell, D.W., 2001. Molecular Cloning: A Laboratory Manual, third ed.Cold Spring Harbor, Cold Spring Harbor Laboratory, New York.

Settani, L., Corsetti, A., 2008. Application of bacteriocin in vegetable food bio-preservation. Int. J. Food Microbiol. 121, 123–138.

Settani, L., Massitti, O., Van Sinderen, D., Corsetti, A., 2005. In situ activity ofa bacteriocin-producing Lactococcus lactis strain. Influence on the interactionsbetween lactic acid bacteria during sourdough fermentation. J. Appl. Microbiol.99, 670–681.

Sharma, A.D., Singh, J., 2005. A nonenzymatic method to isolate genomic DNA frombacteria and actinomycete. Anal. Biochem 337, 354–356.

Singh, A.K., Ramesh, A., 2008. Succession of dominant and antagonistic lactic acidbacteria in fermented cucumber: Insights from a PCR-based approach. FoodMicrobiol. 25, 278–287.

Song, Y., Kato, N., Liu, C., Matsumiya, Y., Kato, H., Watanabe, K., 2000. Rapid iden-tification of 11 human intestinal Lactobacillus species by multiplex PCR assaysusing group- and species-specific primers derived from the 16S-23S rRNAintergenic spacer region and its flanking 23S rRNA. FEMS Microbiol. Lett. 187,163–173.

Stumpe, M.C., Grubmuller, H., 2007. Interaction of urea with amino acids. Impli-cation for urea-induced protein denaturation. J. Am. Chem. Soc. 129,16126–16131.

Terzic-Vidojevic, A., Vukasinovic, M., Veljovic, K., Ostojic, M., Topisirovic, L., 2007.Characterization of microflora in homemade semi-hard white zlatar cheese. Int.J. Food Microbiol. 114, 36–42.

Varadaraj, M.C., Devi, N., Keshava, N., Manjrekar, S.P., 1993. Antimicrobial activityof neutralized extracellular culture filtrates of lactic acid bacteria isolatedfrom a cultured Indian milk Product (dahi). Int. J. Food Microbiol. 20,259–267.

Ventura, M., Zink, R., 2002. Specific identification and molecular typing analysis ofLactobacillus johnsonii by using PCR-based methods and pulse-field gel elec-trophoresis. FEMS Microbiol. Lett. 217, 141–154.

Versalovic, J., Schneider, M., de Bruijn, F.J., Lupski, J.R., 1994. Genomic fingerprintingof bacteria using repetitive sequence based PCR (rep-PCR). Methods Cell. Mol.Biol. 5, 25–40.

van Hylckama Vlieg, J.E.T., Hugenholtz, J., 2007. Mining natural diversity of lacticacid bacteria for flavour and health benefits. Int. Dairy J 17, 1290–1297.

Weisburg, W.G., Barns, S.M., Pelletier, D.A., Lane, D.J., 1991. 16S ribosomal DNAamplification for phylogenetic study. J. Bacteriol 173, 697–703.

Yost, C.K., Nattress, F.M., 2002. Molecular typing techniques to characterize thedevelopment of a lactic acid bacteria community on vacuum-packaged beef. Int.J. Food Microbiol. 72, 97–105.