International Journal of Food Microbiology 128 (2008) 268–273

Contents lists available at ScienceDirect

International Journal of Food Microbiology

j ourna l homepage: www.e lsev ie r.com/ locate / i j foodmicro

Microbial diversity of traditional Vietnamese alcohol fermentation starters(banh men) as determined by PCR-mediated DGGE

Vu Nguyen Thanh ⁎, Le Thuy Mai, Duong Anh TuanDepartment of Microbiology, Food Industries Research Institute, 301-Nguyen Trai, Thanh Xuan, Hanoi, Vietnam

⁎ Corresponding author. Tel.: +84 4 8583450; fax: +8E-mail address: thanh@firi.ac.vn (V.N. Thanh).

0168-1605/$ – see front matter © 2008 Elsevier B.V. Aldoi:10.1016/j.ijfoodmicro.2008.08.020

a b s t r a c t

a r t i c l e i n f o

Article history:

The diversity of fungi and Received 21 February 2008Received in revised form 3 July 2008Accepted 28 August 2008

Keywords:VietnamTraditional starterAlcohol fermentationBanh menDGGEMicrobial community

bacteria associated with traditional Vietnamese alcohol fermentation starters(banh men) was investigated by PCR-mediated DGGE. From 52 starter samples, 13 species of fungi (includingyeasts) and 23 species of bacteria were identified. The fungal composition of the starters was consistent withlittle variation among samples. It consisted of amylase producers (Rhizopus oryzae, R. microsporus, Absidiacorymbifera, Amylomyces sp., Saccharomycopsis fibuligera), ethanol producers (Saccharomyces cerevisiae,Issatchenkia sp., Pichia anomala, Candida tropicalis, P. ranongensis, Clavispora lusitaniae), and (opportunistic)contaminants (Xeromyces bisporus, Botryobasidium subcoronatum). The bacterial microflora of starters washighly variable in species composition and dominated by lactic acid bacteria (LAB). The most frequent LAB werePediococcus pentosaceus, Lactobacillus plantarum, L. brevis,Weissella confusa, andW. paramesenteroides. Species ofamylase-producing Bacillus (Bacillus subtilis, B. circulans, B. amyloliquefaciens, B. sporothermodurans), acetic acidbacteria (Acetobacter orientalis, A. pasteurianus), and plant pathogens/environment contaminants (Burkholderiaubonensis, Ralstonia solanacearum, Pelomonas puraquae) were also detected. Fungal DGGEwas found to be usefulfor evaluating starter type and starterquality.Moreover, inviewof thehighbiological diversity of these substrates,bacterial DGGE may be useful in determining the identity of a starter. The constant occurrence of opportunisticcontaminants highlights the need for careful examination of the role of individual components in starters.

© 2008 Elsevier B.V. All rights reserved.

1. Introduction

Like other Asian countries, Vietnam has centuries of history in thetraditional production of alcohol from rice (Haard et al., 1999). Alcoholicfermentation of steamed rice is carried out by banh men, the speciallyprepared and preserved starter. Banh men is made from an uncookeddough of rice, water, and a variety of herbs and spices. The dough isinoculatedwith starter from a previous batch, shaped into small balls orflattened tablets, and incubated at around30 °C for about aweek.Duringthis period a microbial community develops and the tablets dehydrate.The tablets canbestoredat ambient temperaturewithno significant lossof viability for at least 6 months (Aidoo et al., 2006).

For alcohol production, banh men is pulverized and inoculated intocooked rice. The mash is fermented with aeration for 1–2 days. Wateris then added and an anaerobic fermentation is carried out for aboutone week. Once the fermentation is complete, an alcoholic drink isobtained by means of filtration or distillation. Although rice is themost popular raw material, maize and cassava are also used. Earlytermination of the anaerobic fermentation allows banh men to be used

4 4 8584554.

l rights reserved.

also for making ruou nep, a traditional snack with mild alcoholic andsweet tastes.

A previous study (Lee and Fujio 1999) of 12 banh men samplesindicated that the starter is acidic with a mean pH of 5.67 and has amicrobial community dominated by molds and yeasts at cell densitiesof 1.3×106 and 4.3×106 cfu.g−1 fresh sample, respectively. The mouldsisolated from banh men were identified as Rhizopus oryzae, Mucorindicus, M. circinelloides, and Amylomyces rouxii. The yeasts were Sac-charomyces cerevisiae, Hyphopichia burtonii, Saccharomycopsis fibuligera,Pichia anomala, and Candida sp. The authors concluded that banh menis similar to ragi and other Asian fermentation starters. Until now,no information has been available on the bacteria associated with banhmen.

The microbial communities associated with banh men are the resultof long term selection and are regarded as part of the cultural andtechnological heritage of Vietnam.Banhmen is often considered as a richsource for the isolation and selection of microorganisms that can beutilized in the food industry. Strains of A. rouxii and S. cerevisiae havebeen selected for the production of defined granulated starters for theproduction of high-quality Vietnamese rice wine (Dung, 2004; Dunget al., 2005). The success of industrial alcohol production as well as thecommercialization of factory produced starters in Vietnam in the lastdecade has had a major impact on the diversity of traditionalfermentations. It appears that a decline in traditional fermentations isinevitable, resulting in a loss of biological diversity. In order to preserve

269V.N. Thanh et al. / International Journal of Food Microbiology 128 (2008) 268–273

that biodiversity, we have established a starter bank based on freeze-dried banh men samples collected from around the country. This studydeals with the microbiological analysis of these banh men samples.

Although there are no reports of non-culturable microorganismsassociated with amylolytic starters, a complete picture of thebiodiversity present in a specific type of starter is often elusive as itrequires expertise with a wide range of microorganisms includingyeasts, filamentous fungi, and bacteria. This task involves the isolationand identification of a large number of strains and the use of numerousselective isolation media. For this reason, in most reports on thebiodiversity of traditional alcohol fermentation starters, the range ofmicroorganisms covered is predicated on the quality of the identifica-tion technique. In this report we describe the microbial diversityassociated with banh men using PCR-mediated DGGE, a technique thatallows the detection of a broad range of microorganisms without theneed to culture them. The technique has proven to be a powerful tool ininvestigating different types of traditional fermentations (Haruta et al.,2006; Meroth et al., 2003; Prakitchaiwattana et al., 2004; Rantsiouet al., 2005). Various applications of DGGE in food microbiology werecomprehensively reviewed by Ercolini (2004).

2. Materials and methods

2.1. Sample collection and starter bank

Collections of banh men were conducted during the years 2004,2005, and 2006. The collection sites spanned across northern, middle,and southern parts of Vietnam. Samples were collected at farmers'houses and local markets. After collection, samples were asepticallycrushedmanually in plastic bags to reduce the particles size to less than1mm. Approximately 1 g of samplewas placed in a sterile glass ampule(inner diameter=10 mm) and plugged with cotton wool. The startersamples were dried overnight in a Lioalfa-6, Telstar (Spain) freeze dryer.After primarydrying, the ampuleswere connected to amanifold forfinaldrying.When the pressure inside the drying chamber reached 1×10−2 –5×10−2 mbar (condenser temperature=−80 °C) the ampules weresealed with a propane–oxygen burner. The ampules were labeled andstored in plastic bags at 4–10 °C.

2.2. DNA extraction from starters

To ensure adequate sampling, 52 representative banh men sampleswere analyzed. For DNA extraction, ca. 0.1 g of freeze-dried sampleswere incubated with 1 ml of×2 SSC (0.3 M NaCl, 30 mM sodiumcitrate) (Sambrook et al., 1989) in 1.5 ml micro-centrifuge tubes at99 °C for 10 min in a dry block heater (Thermomixer Comfort,Eppendorf, Germany). The tubes were centrifuged at 10,000 g for5 min and the upper layer was discarded. Sterile distilled water(200 µl), 200 µl phenol-chloroform, and 300 µl glass beads (0.2–0.5 mm diameter, Roth, Germany) were added. Microbial cells weredisrupted using a mini-bead beater (BioSpec, USA) for 4 min withchilling on ice after each 1min. The tubes were centrifuged at 10,000 gfor 10 min. The upper layer was transferred into new tubes and DNAwas precipitated with 2.5 volumes of ethanol. The tubes were storedat −20 °C for 30 min and centrifuged at 10,000 g for 10 min. Thesupernatant was discarded and the pellet was washed twice with icecold 70% ethanol. The pellet was dried and re-dissolved in 50 µl water.Residual starch present in the samples interfered with PCR reactions.For this reason, the DNAs were therefore subjected to furtherpurification by adsorption to a silica matrix using a Fermentas (USA)DNA extraction kit according to the manufacturer's instructions.

2.3. DNA amplification and DGGE

The microbial composition of banh men was investigated byamplifying partial sequences of ribosome encoding genes from total

DNA extracts. For bacteria, fragments of ca. 470 bases were amplifiedby using primers 16F945 (5′-GGGCCCGCACAAGCGGTGG-3′) (Wagner-Döbler et al., 1998) and 16R1401 (5′-CGGTGTGTACAAGACCC-3′)(Nübel et al., 1996) corresponding to positions 927 to 945 and 1385to 1401 of the Escherichia coli 16 S rDNA. Similarly, ca. 320 bases in theD2 domain of the fungal large subunit rDNAs were amplified usingprimers NL3A (5′-GAGACCGATAGCGAACAAG-3′) and NL4 (Kurtzmanand Robnett, 1997). The use of the entire D1/D2 domain (ca. 610 bases)using primers NL1 (5′-GCATATCAATAAGCGGAGGAAAAG-3′) and NL4(5′-GGTCCGTGTTTCAAGACGG-3′) (Kurtzman and Robnett, 1997) wasalso explored. An improved 40-nucleotide GC clamp (5′-CGCCCGCC-GCGCGCGGCGGGCGGGGCGGGGGCACGGGGGG-3′) (Davies et al.,2004) was attached to the 5′ end of the reverse primers (indicatedas 16R1401GC and NL4GC). All primers were prepared by MetabionGmbH (Germany). Prokaryotic and eukaryotic rDNAs were amplifiedseparately using a GeneAmp 9700 thermocycler (PE — AppliedBiosystem, USA) with a PCR program consisting of an initialdenaturation step of 95 °C for 2 min followed by 35 cycles of 94 °Cfor 30 s, 60 °C for 40 s, and 72 °C for 1min. The final extension stepwas72 °C for 7 min.

The PCR products were separated on a 9% polyacrylamide (37.5:1acrylamide: bisacrylamide) 20 cm×20 cm gel with a denaturationgradient from 20% to 60%. The 100% denaturation gel contained 7 Murea and 40% (v/v) formamide in 1×TAE (40 mM Tris, 20 mM aceticacid, 1 mM EDTA). The electrophoresis was run at 60 °C and 70 V for16 h in a 6-liter DGGE tank (Scie-Plas, UK). The gel was then rinsed indeionized water for 30 min to remove denaturant (urea andformamide), stained in 0.5×TAE containing 0.5 µg/ml ethidiumbromide, and viewed on a Benchtop Ultraviolet Transilluminator(VWR Scientific, USA). Images were captured on a Camedia 4000Zdigital camera (Olympus, Japan).

2.4. Identification of bands

Representative bands were excised from denaturing gels underlow UV exposure. The excised bands were placed in micro-centrifugetubes and washed with 1.0 ml 0.5×TAE and 1.0 ml deionized water,and then destained with 1.0 ml of isopropanol. The isopropanol wasdiscarded and the gels excerpts were dried in air. Glass beads (ca.70 µl) and 50 µl deionized water were added, following which the gelpieces were ground with disposable plastic pestles. Deionized water(100 µl) was added and the gels were homogenized in a mini-beadbeater (BioSpec, USA) for 4 min with interval chilling on ice after each1 min. The tubes were then centrifuged at 10,000 g for 5 min and thesupernatants were used as templates for PCR. The excised bands werere-amplified using the same primer pairs and tested by DGGE again toconfirm homogeneity. The PCR products were purified with QIAEX IIagarose gel extraction kits (Qiagen, USA) and sequenced by thedideoxy chain termination method, using non-clamped primers (i.e.16F945 for bacteria and NL3A for yeast and fungi). A model 377automated DNA sequencer (Applied Biosystems, USA) was used. Foridentification, the resulting sequences were compared with GenBankdatabase using the web-based nucleotide–nucleotide BLAST searchengine hosted by the National Center for Biotechnology Information,Bethesda, USA (http://www.ncbi.nlm.nih.gov) (Altschul et al., 1997).

3. Results and discussion

3.1. Sample collection and starter bank

A total of 116 banh men samples were collected. The sampling sitescovered the whole country but were concentrated around the denselypopulated regions of the Red river delta and to a lesser extent theMekong basin. During collection trips it became evident that extensivetrading of starters takes place. As starters are made from otherstarters, the mixing of different types is inevitable with the result that

Table 1List of NL3A-NL4GC DGGE bands and identification by DNA sequencing

Band Putative species Related GenBanksequence

Identity

F-110.1 Rhizopus oryzae AY213624 363/363 (100.0%)F-166.1 Rhizopus oryzae AY213624 362/363 (99.7%)F-166.5 Rhizopus microsporus AB363776 369/369 (100.0%)F-34.2 Rhizopus microsporus AB363776 79/79 (100.0%)F-34.7 Absidia corymbifera AF113446 331/331 (100.0%)F-22.4 Amylomyces sp. DQ466612 105/113 (92.9%)F-166.3 Saccharomycopsis fibuligera U09238 293/293 (100.0%)F-5.5 Saccharomycopsis fibuligera U09238 275/275 (100.0%)F-110.3 Saccharomycopsis fibuligera U09238 280/280 (100.0%)F-34.3 Saccharomyces cerevisiae EU268656 273/273 (100.0%)F-166.2 Saccharomyces cerevisiae EU268656 300/300 (100.0%)F-110.2 Saccharomyces cerevisiae EU268656 298/298 (100.0%)F-34.6 Issatchenkia sp. EU268659 273/283 (96.5%)F-29.1 Pichia anomala DQ318803 284/285 (99.6%)F-89.2 Candida tropicalis AB281286 229/232 (98.7%)F-5.4 Pichia ranongensis AB334112 243/243 (100.0%)F-110.5 Clavispora lusitaniae EF694616 239/239 (100.0%)F-5.6 Botryobasidium subcoronatum AY647212 267/269 (99.3%)F-110.4 Xeromyces bisporus AF365020 245/246 (99.6%)

270 V.N. Thanh et al. / International Journal of Food Microbiology 128 (2008) 268–273

boundaries between the communities are blurred and the diversity isreduced. From the 116 banh men samples collected, only 45 wereproduced locally at the site of collection and the remaining 71 wereintroduced from elsewhere. Among the 116 collected samples, 4 weremen la (a special type of banh men produced by the people of an ethnicminority),10 were produced by small scale factories, 8 were in powderform, 4 were introduced from China, and 4 were used for makingsweet fermenting rice (ruou nep) rather than spirit.

One of our objectives was to preserve the genetic diversity ofmicroorganisms associated with traditional starters. One would assumethat this is best accomplished by isolating all microorganisms from thecollected samples and preserving them in a culture collection. However,this approach is hampered by the difficulty of retrieving all existingmicroorganisms. Another difficulty has to do with the reconstruction ofstarters basedon the individual strains. The stability of the starters is owedto the symbiotic relationship among the microorganisms, which isaffected by the proportion of each species as well as the abiotic factorsunique to eachmicrobial community. Until now, the reconstruction of theseveral types of traditional starters has not been possible. An example ofsuch a complex system is the starter used for kefir fermentation (Lopitz-Otsoa et al., 2006). For this reason, we opted for a different approach,whereby thewhole community, andnot individual strains, is preservedbyfreeze-drying. Lyophilized banh men starters are expected to be stable for20–30 years. Preliminary work has shown that freeze-drying does notaffect plate count numbers or microbial community composition. Thismakes sense since traditional starters are often supplied in dry form andcan be stored routinely for up to six months (Aidoo et al., 2006).

3.2. DGGE analysis

Fungi and bacteria were assessed separately with specific DGGEprimers. The 3′ end of 16 S rDNA and the D2 domain of the LSU rDNAwere chosen because of their relatively high divergence and the widecoverage of the GenBank database for these regions. In the case offungi, the primer pair NL3A-NL4GC reproducibly differentiates speciesbased on four or more nucleotide differences (Thanh and Manh,2005). In this study we employed slightly longer fragments (ca. 320bases for fungi and ca. 470 bases for bacteria) than those used by otherauthors (Haruta et al., 2006; Yoshida et al., 2005), to ensure a moreprecise identification of isolated bands. However, the use of largerfragments was not successful, as it results in a major loss ofelectrophoretic resolution (data not shown).

Fig. 1. Yeast and filamentous fungal community associated with traditional alcoholfermentation starters as revealed by DGGE. Lanes corresponding to different startersamples are indicated by numbers at the top. Numbers on the gel show bands that wereexcised for identification (see Table 1).

3.3. Yeast and fungal community

The DGGE patterns of amplicons generatedwith primer pair NL3A-NL4GC showed high reproducibility. Different batches of DNAextractions from the same starter sample and repeated PCR from thesame DNA extract gave nearly identical results. The DGGE profile ofthe fungal community associated with banh men was not overlycomplex. Major bands were recovered in many samples and a fewminor bands differed from sample to sample. Individual samples mostoften produced less than 10 bands. A typical DGGE profile of fungalcommunity is presented in Fig. 1. The list of identified bands ispresented in Table 1. The identification was extrapolated to otherbands based on similarity in mobility and thus the microbiologicalstructure of all banh men samples could be evaluated. The rate ofoccurrence of individual fungal (and yeast) species is presented inTable 2.

Four species of filamentous fungi in the family Mucoraceae (R. oryzae,R. microsporus, Absidia corymbifera, and Amylomyces sp.), were commonlyfound in banh men. These species are strong amylase producers and arefrequently found in amylolytic Asian fermentation starters (Haard et al.,1999; Hesseltine, 1983; Hesseltine et al., 1988). R. oryzae was the most

Table 2Occurrence of yeast and fungal species in traditional Vietnamese alcohol fermentationstarters as revealed by DGGE

Species Number of bandsidentified

Occurrence(per 52 samples)

Amylase producer 9 49Rhizopus oryzae 2 47Rhizopus microsporus 2 9Absidia corymbifera 1 12Amylomyces sp. 1 8Saccharomycopsis fibuligera 3 32Ethanol producer 8 44Saccharomyces cerevisiae 3 41Issatchenkia sp. 1 18Pichia anomala 1 7Clavispora lusitaniae 1 3Candida tropicalis 1 2Pichia ranongensis 1 2Substrate inhabitant/contaminant 2 35Xeromyces bisporus 1 35Botryobasidium subcoronatum 1 2

Note: Species were grouped on the basis of putative functions or origins (in bold text).The presence of a group in a given sample was indicated by the presence of at least onemember of the group in the sample.

Fig. 2. Bacterial community associated with traditional alcohol fermentation starters asrevealed by DGGE. Lanes corresponding to different starter samples are indicated bynumbers at the top. Numbers on the gel show bands that were excised for identification(see Table 3).

271V.N. Thanh et al. / International Journal of Food Microbiology 128 (2008) 268–273

common fungus of any kind in the community associated with banhmen, having been identified in 47 out of 52 tested samples. The secondmost abundant amylase producer was S. fibuligera.

In a previous study of amylolytic microorganisms involved intraditional alcohol fermentations, Mucoraceae were not predominantand S. fibuligera was the principal species (Thanh et al., 2005).S. fibuligera was also the main amylase producer in Loog-pang, atraditional Thai alcohol starter similar to banh men (Limtong et al.,2002). The strong banding and the highest frequency of occurrence ofR. oryzae in the DGGE profiles obtained in this study are at variancewith those observations. One explanation is that the cells or hyphaewalls of Mucoraceae are relatively larger or more fragile than yeastcells and thus facilitate DNA extraction and PCR. Moreover, theprevalence of Mucoraceae in the starter does not necessarily ensurethat other species will not prevail during the actual fermentationprocess, where the substrate and environment conditions aredifferent. This does not rule out the possibility that species ofMucoraceae play a role in ethanol production. In fact, a pure cultureof A. rouxii was used successfully for the production of wine from redglutinous rice (Dung, 2004; Dung et al., 2005).

S. cerevisiae, Issatchenkia sp., P. anomala, Candida tropicalis, Pichiaranongensis, and Clavispora lusitaniae were the main ethanol producersidentified. Among these, S. cerevisiaewas the most frequent, having beendetected in41outof52examined samples.Ahigh frequencyofoccurrencewas confirmed by the culture method (Thanh et al., 2005). S. cerevisiae isthe most effective ethanol producer known so far (Vaughan-Martini andMartini, 1995). The next frequent species in starters was a species of Is-satchenkia with one band (F-34.6) identified. The yeast was closest toIssatchenkia orientalis but a precise assignment was hindered by thepresenceof ambiguousbases. Theyeast occurred in18 samples. I. orientaliswas found to be common in the Thai starter Loog-pang (Limtong et al.,2002) and the Indonesian brem (Sujaya et al., 2001). I. orientalis is a strongfermenter that can produce up to 6% (w/v) of ethanol in broth culture(Limtong et al., 2002).P. anomalawas detected in 7 samples byDGGE. Thisascosporic yeast was also found in banh men by the microbiologicalisolation technique (Thanh et al., 2005). P. anomala (syn. Hansenulaanomala) is a regular component in several types of Asia-Pacific alcoholfermentation starters (Haard et al.,1999; Limtong et al., 2002; Sujaya et al.,2001). The yeast can accumulate up to 5% (w/v) ethanol in broth culture(Limtong et al., 2002). Less frequent were C. lusitaniae, C. tropicalis andPichia ranongensis, which occurred in 2–3 samples (Table 2). C. lusitaniaeand C. tropicalis are cosmopolitan species that can be found in a vastvariety of substrates including fruits, soil, water, and humans (Lachanceand Phaff, 1998; Meyer et al., 1998). P. ranongensis is a yet undescribednovel ascomycetous yeast isolated from water in mangrove forest inThailand, represented in GenBank as sequence AB334112.

Botryobasidium subcoronatum was among the species withunknown functions found in the starters. B. subcoronatum is abasidiomycetous species in the order Agaricomycetes. It has a cosmo-politan distribution with a broad spectrum of substrates. Thesaprotrophic species occurs on wood of both gymnosperms andangiosperms (http://www.uni-tuebingen.de/uni/bbm/ mycology/botryoba/subcoron.htm). In banh men, B. subcoronatum was found intwo of 52 tested samples and thus we assume that it is a contaminantthat may be associated with plant materials use in the preparation ofthe starter. The most unexpected was Xeromyces bisporus, found in 35of the 52 samples examined. X. bisporus is an extreme xerophylicascomycete with optimum growth at aw of ca. 0.92. The fungus ceasesto grow at aw values above 0.97 (equal to 24% w/w glucose solution;Gock et al., 2003; Pitt and Christian, 1968; Rosso and Robinson, 2001).The biology of X. bisporus allows us to predict that this fungusdevelops on the starter during the drying and storage processes. Thespecies cannot grow during mash fermentation due to the high wateractivity. The frequent occurrence of X. bisporus in starters and its lackof functionality in the fermentation process raise the question ofwhether it is appropriate to assign a role to individual species based

strictly on frequency of occurrence and enzymatic abilities withoutdocumenting its growth in the fermentation mash. We believe thatthe main difference between traditional and industrial starters is thattraditional starters have a self-regulated microbiota capable ofreproducing in relatively non-sterile condition. Only a fraction ofthe species found in traditional starters serves a function in thesubsequent fermentation process. The rest either play a role inmaintaining the stability of the starter or are just contaminants.Thus, traditional starters are far more complex than industrial startersand establishing the role of each component will require carefulexamination.

3.4. Bacterial community

The diversity of bacteria associated with banh men was analyzedfor 48 starter samples. The bacterial microbiota was much morecomplex than that of fungi (Figs. 1 and 2). In total, 45 bands were cutand identified (Table 3). Due to the high complexity of the system,extrapolation of identified bands to others was liberal. The frequenciesof occurrence of individual species are listed in Table 4.

Despite the high complexity of the system, the bacterial speciesassociated with banh men could be categorized into 5 clear anddistinctive ecological and functional groups. A first group comprisesspecies of the genus Bacillus renown for fast aerobic growth andstrong amylase activity, namely B. subtilis, B. circulans, B. amylolique-faciens, and B. sporothermodurans. Among them, B. subtilis was mostprevalent. All Bacillus species detected in this study are common instarch substrates and have found various applications in foodproduction and biotechnology (Pirttijärvi et al., 2001).

The second group consisted of two species of acetic acid producingbacteria, Acetobacter orientalis and A. pasteurianus. The genus Aceto-bacter is characterized by the ability to convert ethanol to acetic acid inthe presence of oxygen. A. orientalis strains have been isolated fromcanna flower, starfruit, coconut, curd of tofu, and tempe. They canproduce acid on starch substrates and tolerate 7% ethanol (Lisdiyantiet al., 2001). A. pasteurianus is common in various traditionalfermentation products and considered one of the major acetic acidproducers in these materials (Gandjar, 1999; Gullo et al., 2006). Inbanh men, Acetobacter is not very prevalent as they occurred only in3 samples. The presence of Acetobacter is not desirable due to theoxidation of ethanol and the development of a vinegar flavour in thefinal product.

Table 4Occurrence of bacterial species in traditional Vietnamese alcohol fermentation startersas revealed by DGGE

Species Number of bandsidentified

Occurrence(per 48 samples)

Starch degrader 6 11Bacillus amyloliquefaciens 1 1Bacillus circulans 3 3Bacillus sporothermodurans 1 1Bacillus subtilis 1 8Acetic acid producer 3 3Acetobacter orientalis 1 2Acetobacter pasteurianus 2 2Lactic acid producer 32 44Enterococcus faecium 1 1Lactobacillus agilis 1 5Lactobacillus brevis 2 19Lactobacillus fermentum 2 5Lactobacillus manihotivorans 2 1Lactobacillus plantarum 2 22Lactobacillus sp. 1 5Lactococcus lactis 3 14Leuconostoc citreum 1 1Leuconostoc garlicum 1 8Leuconostoc mesenteroides 1 3Pediococcus pentosaceus 3 29Weissella confusa 7 17Weissella paramesenteroides 5 15Plant pathogen/contaminant 4 15Burkholderia ubonensis 1 8Ralstonia solanacearum 1 6Pelomonas puraquae 2 3

Note: Species were grouped on the basis of putative functions or origins (in bold text).The presence of a group in a given sample was indicated by the presence of at least onemember of the group in the sample.

Table 3List of PCR-mediated 16 S rDNA DGGE bands and identification of bacterial species

Band Putative species Related GenBanksequence

Identity

B-89.6.1 Acetobacter orientalis AB180930 167/167 (100.0%)B-89.6.2 Acetobacter pasteurianus AB299165 446/448 (99.6%)B-89.7 Acetobacter pasteurianus AB299165 232/233 (99.6%)B-95.5 Bacillus amyloliquefaciens EF473131 412/416 (99.0%)B-155.9 Bacillus circulans DQ870737 450/451 (99.8%)B-95.8 Bacillus circulans DQ870737 452/453 (99.8%)B-34.2 Bacillus circulans DQ870737 176/176 (100.0%)B-37.9 Bacillus sporothermodurans AB190039 97/99 (98.0%)B-95.6 Bacillus subtilis EU271854 453/454 (99.8%)B-22.9 Burkholderia ubonensis EU024179 159/159 (100.0%)B-41.4 Enterococcus faecium AB362603 408/408 (100.0%)B-89.2 Lactobacillus agilis M58803 323/326 (99.1%)B-29.8 Lactobacillus brevis EF120367 456/457 (99.8%)B-37.8 Lactobacillus brevis EF120367 376/378 (99.5%)B-155.10 Lactobacillus fermentum EU287483 456/457 (99.8%)B-95.7 Lactobacillus fermentum DQ399352 449/450 (99.8%)B-42.7 Lactobacillus manihotivorans AF000162 446/447 (99.8%)B-42.8 Lactobacillus manihotivorans AF000162 455/456 (99.8%)B-42.4 Lactobacillus plantarum EU273819 163/165 (99.0%)B-29.7 Lactobacillus plantarum EU273819 456/457 (99.8%)B-89.4 Lactobacillus sp. AY445129 175/181 (96.7%)B-155.8 Lactococcus lactis EU104368 449/451 (99.6%)B-42.5 Lactococcus lactis EU104368 151/152 (99.3%)B-166.5 Lactococcus lactis EU104368 420/422 (99.5%)B-42.3 Leuconostoc citreum AB362723 431/439 (98.2%)B-48.4 Leuconostoc garlicum AB362722 110/110 (100.0%)B-48.3 Leuconostoc mesenteroides AB023247 226/228 (99.1%)B-48.5 Pediococcus pentosaceus AB362605 445/446 (99.8%)B-89.3 Pediococcus pentosaceus AB362605 170/171 (99.4%)B-91.2 Pediococcus pentosaceus AB362605 456/457 (99.8%)B-9.3 Pelomonas puraquae AM501441 445/448 (99.3%)B-95.4 Pelomonas puraquae AM501441 386/391 (98.7%)B-20.2 Ralstonia solanacearum DQ924957 134/137 (97.8%)B-42.9 Weissella confusa AJ874741 455/456 (99.8%)B-166.7 Weissella confusa AJ874741 349/351 (99.4%)B-166.8 Weissella confusa AJ874741 456/456 (100.0%)B-42.10 Weissella confusa AJ874741 449/450 (99.8%)B-166.6 Weissella confusa AJ874741 196/198 (99.0%)B-91.4 Weissella confusa AJ874741 457/457 (100.0%)B-47.4 Weissella confusa AJ874741 444/445 (99.8%)B-89.5 Weissella paramesenteroides AB023238 449/449 (100.0%)B-111.5 Weissella paramesenteroides AB023238 434/436 (99.5%)B-20.1 Weissella paramesenteroides AB023238 62/63 (98.4%)B-91.3 Weissella paramesenteroides AB023238 302/304 (99.3%)B-42.6 Weissella paramesenteroides AB023238 84/85 (98.8%)

272 V.N. Thanh et al. / International Journal of Food Microbiology 128 (2008) 268–273

The third group of prokaryotes, themost abundant, consisted of thelactic acid bacteria, accounting for 32 of the 45 bands identified anddetected in 44 out of 48 tested samples (Table 4). This group was alsothe most diverse with 6 genera and 14 species detected. Among LAB,the following species were most frequent: Pediococcus pentosaceus,L. plantarum, L. brevis, W. confusa, W. paramesenteroides (Table 4).Although not verified empirically, the role of lactic acid bacteria intraditional alcohol fermentation starter is often thought to be anacidification that favours the growth of yeasts and filamentous fungiwhile suppressing the growth of food spoilage bacteria (Gandjar, 1999,Haard et al., 1999). Excess growth of lactic acid bacteria is howeverundesirable because it causes sugar losses and off-flavors in the finalproduct. Ethanol production and lactic acid production are competingprocesses in traditional alcohol fermentations. The balance of the twoprocesses will affect production efficiency and product quality.

Using DGGE we detected two groups of bacteria with no obviousfunctionality in banh men. They included the plant pathogens(Burkholderia ubonensis and Ralstonia solanacearum) as well as speciesthat are most likely to be environmental contaminants (Pelomonaspuraquae). B. ubonensis belongs to the Burkholderia cepacia complex,which contains versatile organisms that occupy a wide range ofecological niches. B. ubonensis was described only recently based onsoil isolates (Zhang and Xie, 2007). R. solanacearum is a Gram-

negative, plant pathogenic, soil bacterium. It colonizes the phloem,causing a bacterial wilt in a very wide range of potential host plants(Yao and Allen, 2006). We presume that plant pathogens wereintroduced in starters as contaminants of plant tissue and have nofunction or activity in the fermentation process. P. puraquae is arecently described species isolated from industrial water in Sweden(Gomila et al., 2007). The genus Pelomonas was created recently forPseudomonas saccharophila based on a phylogenetic analysis of 16 SrDNA (Xie and Yokota, 2005). There is no report thus far of thepresence of Pelomonas in traditional fermentation products and therole or possible contamination source of the bacterium remainsunknown. Amongst 48 tested samples the bacterium was found inonly three (Table 4).

The bacterial microbiota of banh men is highly similar to that ofIndonesian tape. Sujaya et al. (2001) have found P. pentosaceus, Entero-coccus faecium, Lactobacillus curvatus, Weissella confusa, and W. para-mesenteroides as major LAB in ragi tape. Similarly species of the generaBacillus and Acetobacter have been found in tape ketan (Ardhana andFleet, 1989).

One of the major differences between the fungal and bacterialmicrobiota of banh men is consistency. Whereas fungal microflorashowed little variation between samples, the bacterial communityexhibited a rather “spontaneous” species composition. This might bedue to different selectionpressures exercises on these two communitiesby the various processes. It is widely accepted that fungi rather thanbacteria, play major roles in alcohol fermentations (Haard et al., 1999)such that alterations in fungal species composition would largely affectfermentation efficiency and product quality. Some deviations fromstandard DGGE profiles were detected and correlated with the type ofstarter. For example, an industrially produced starter imported fromChina for alcohol production from rice showed only one band ofS. cerevisiae in fungal DGGE (Fig. 1, lane 185). This was confirmed withthe culture method. The source of amylase in such commercialpreparations is unspecified. Banh men intended for the production of

273V.N. Thanh et al. / International Journal of Food Microbiology 128 (2008) 268–273

sweet fermenting rice (ruou nep) was also notable for the lack of S.cerevisiae (Fig. 1, lane 174). Strong ethanol production in such a type offermentation is undesirable. Regarding bacteria, some alteration wasalso detected in a factory produced starter. In sample BM94 (Fig. 2, lane94),we have detected only a single bandofB. subtilis and thismost likelyindicated the incorporation of pure culture to the process.

The fungal composition of banh men as revealed by PCR-mediatedDGGE showed considerable agreement with our previous knowledgederived from the culture method (Lee and Fujio, 1999; Thanh et al.,2005) except for the presence of the extreme osmophilic fungusX. bisporus, which would not grow in standard cultivation media. Thebacterial community associated with banh men was similar to that ofIndonesial ragi tape. Fungal DGGE profiles could be useful in evaluatingstarter type and quality, meanwhile, due to high variation, bacterialDGGE could be utilized in determining the identity of a starter.

Although DGGE electrophoresis does provide a broad overview of amicrobial community, interpretation of DGGE data should be accom-panied with caution. The technique cannot differentiate dead and livingcells, which would preclude the application of DGGE in microbialsuccession studies, especiallywhen change is rapid. Band intensity is notalways correlated with population density, as DNA extraction efficien-cies vary between microorganisms (Ercolini, 2004; Prakitchaiwattanaet al., 2004). In DGGE studies, identified bands are often extrapolated toothers based on similarity in mobility. We found this approach veryusefulwhen similar communities are being compared. Thatwas the casefor the fungal microbiota of banh men. However, with a number of banhmen samples, extrapolation of bacterial bands was error-prone.

Acknowledgements

We thank Nguyen Thi Quynh Anh and Nguyen Thi Thu Huong fortheir valuable assistance, Prof. M.A. Lachance at the University ofWestern Ontario for the helpful suggestions and comments. This studywas supported by the International Foundation for Science (Sweden)through research grant E/3696-1.

References

Aidoo, K.E., Nout, M.J., Sarkar, P.K., 2006. Occurrence and function of yeasts in Asianindigenous fermented foods. FEMS Yeast Research 6, 30–39.

Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J.,1997. Gapped BLAST and PSI-BLAST: a new generation of protein database searchprograms. Nucleic Acids Research 25, 3389–3402.

Ardhana, M.M., Fleet, G.H., 1989. The microbial ecology of tape ketan fermentation.International Journal of Food Microbiology 9, 157–165.

Davies, C.E., Hill, K.E., Wilson, M.J., Stephens, P., Hill, C.M., Harding, K.G., Thomas, D.W.,2004. Use of 16 S ribosomal DNA PCR and denaturing gradient gel electrophoresisfor analysis of the microfloras of healing and nonhealing chronic venous leg ulcers.Journal of Clinical Microbiology, 42, 3549–3557.

Dung, N.T.P., 2004. Defined fungal starter granules for purple glutinous rice wine, Ph.D.Thesis. Wageningen University, Wageningen, The Netherlands, p. 110.

Dung, N.T.P., Rombouts, F.M., Nout, M.J.R., 2005. Development of defined mixed-culturefungal fermentation starter granulate for controlled production of rice wine.Innovative Food Science and Emerging Technologies 6, 429–441.

Ercolini, D., 2004. PCR-DGGE fingerprinting: novel strategies for detection of microbesin food. Journal of Microbiological Methods 56, 297–314.

Gandjar, I., 1999. Fermented foods — fermentations of the Far East. In: Robinson, R.K.,Batt, C.A., Patel, P.D. (Eds.), Encyclopedia of Food Microbiology. Academic Press,London, pp. 767–773.

Gock, M.A., Hocking, A.D., Pitt, J.I., Poulos, P.G., 2003. Influence of temperature, wateractivity and pH on growth of some xerophilic fungi. International Journal of FoodMicrobiology 81, 11–19.

Gomila, M., Bowien, B., Falsen, E., Moore, E.R., Lalucat, J., 2007. Description of Pelomonasaquatica sp. nov. and Pelomonas puraquae sp. nov., isolated from industrial andhaemodialysis water. International Journal of Systematic and EvolutionaryMicrobiology. 57, 2629–2635.

Gullo, M., Caggia, C., De Vero, L., Giudici, P., 2006. Characterization of acetic acid bacteria in“traditional balsamic vinegar”. International Journal of Food Microbiology 106, 209–212.

Haard, N.F., Odunfa, S.A., Lee, C.H., Quintero-Ramírez, R., Lorence-Quinones, A., Wacher-Radarte, C., 1999. Fermented cereals a global perspective. FAO Agricultural ServicesBulletin 138, 97.

Haruta, S., Ueno, S., Egawa, I., Hashiguchi, K., Fujii, A., Nagano, M., Ishii, M., Igarashi, Y.,2006. Succession of bacterial and fungal communities during a traditional potfermentation of rice vinegar assessed by PCR-mediated denaturing gradient gelelectrophoresis. International Journal of Food Microbiology 109, 79–87.

Hesseltine, C.W., 1983. Microbiology of oriental fermented foods. Annual Review ofMicrobiology 37, 575–601.

Hesseltine, C.W., Rogers, R., Winarno, F.G., 1988. Microbiological studies on amylolyticoriental fermentation starters. Mycopathologia 101, 141–155.

Kurtzman, C.P., Robnett, C.J., 1997. Identification of clinically important ascomycetousyeasts based on nucleotide divergence in the 5′ end of the large-subunit (26 S)ribosomal DNA gene. Journal of Clinical Microbiology 35, 1216–1223.

Lachance, M.A., Phaff, H.J., 1998. Clavispora Rodrigues deMiranda, In: Kurtzman, C.P., Fell, J.W.(Eds.), The Yeasts, a Taxonomic Study, 4th edn. Elsevier, Amsterdam, pp. 148–152.

Lee, A.C., Fujio, Y., 1999. Microflora of banh men, a fermentation starter from Vietnam.World Journal of Microbiology and Biotechnology 15, 57–62.

Limtong, S., Sintara, S., Suwannarit, P., Lotong, N., 2002. Yeast diversity in Thaitraditional alcoholic starter. Kasetsart Journal (Natural sciences) 36, 149–158.

Lisdiyanti, P., Kawasaki, H., Seki, T., Yamada, Y., Uchimura, T., Komagata, K., 2001.Identification of Acetobacter strains isolated from Indonesian sources, and proposalsof Acetobacter syzygii sp. nov., Acetobacter cibinongensis sp. nov., and Acetobacterorientalis sp. nov. Journal of General and Applied Microbiology 47, 119–131.

Lopitz-Otsoa, F., Rementeria, A., Elguezabal, N., Garaizar, J., 2006. Kefir: a symbioticyeasts–bacteria community with alleged healthy capabilities. Revista iberoamer-icana de micología 23, 67–74.

Meroth, C.B., Hammes, W.P., Hertel, C., 2003. Identification and population dynamics ofyeasts in sourdough fermentation processes by PCR-denaturing gradient gelelectrophoresis. Applied and Environmental Microbiology 69, 7453–7461.

Meyer, S.A., Payne, R.W., Yarrow, D., 1998. Candida Berkhout, In: Kurtzman, C.P., Fell, J.W.(Eds.), The Yeasts, a Taxonomic Study, 4th edn. Elsevier, Amsterdam, pp. 454–573.

Nübel, U., Engelen, B., Felske, A., Snaidr, J., Wieshuber, A., Amann, R.I., Ludwig, W.,Backhaus, H., 1996. Sequence heterogeneities of genes encoding 16 S rRNAs inPaenibacillus polymyxa detected by temperature gradient gel electrophoresis.Journal of Bacteriology 178, 5636–5643.

Pirttijärvi, T.S., Wahlström, G., Rainey, F.A., Saris, P.E., Salkinoja-Salonen, M.S., 2001.Inhibition of bacilli in industrial starches by nisin. Journal of Industrial Microbiologyand Biotechnology 26, 107–114.

Pitt, J.I., Christian, J.H., 1968. Water relations of xerophilic fungi isolated from prunes.Applied Microbiology 16, 1853–1858.

Prakitchaiwattana, C.J., Fleet, G.H., Heard, G.M., 2004. Application and evaluation ofdenaturing gradient gel electrophoresis to analyse the yeast ecology of wine grapes.FEMS Yeast Research 4, 865–877.

Rantsiou, K., Urso, R., Lacumin, L., Cantoni, C., Cattaneo, P., Comi, G., Cocolin, L., 2005. Culture-dependent and — independent methods to investigate the microbial ecology of Italianfermented sausages. Applied and Environmental Microbiology 71, 1977–1986.

Rosso, L., Robinson, T.P., 2001. A cardinalmodel to describe the effect of water activity onthe growth of moulds. International Journal of Food Microbiology 63, 265–273.

Sambrook, J., Fritsch, E.F., Mainiatis, T., 1989. Molecular cloning: A Laboratory Manual.Cold Spring Harbor Laboratory Press, New York.

Sujaya, I.N., Amachi, S., Yokota, A., Asano, K., Tomita, F., 2001. Identification andcharacterization of lactic acid bacteria in ragi tape. World Journal of Microbiologyand Biotechnology 17, 349–357.

Thanh, V.N., Manh, L.D., 2005. Resolution power of DGGE technique using 26 S rDNA(D1/D2) in biodiversity assessment. Proceedings of National Conference 2005 —

“Biotechnology in Basic Research”, Hanoi, Vietnam, December 6, 2005, pp. 115–117.Thanh, V.N., Thu Huong, N.T., Quynh Anh, N.T., 2005. Conservation of the biodiversity of

Vietnam traditional alcohol fermentation starters. MSA Golden Jubilee InternationalScience Congress (ISC) 2005, August 3–6, 2005, Kuala Lumpur, Malaysia, pp. 160–161.

Vaughan-Martini, A., Martini, A., 1995. Facts, myths and legends on the prime industrialmicroorganism. Journal of Industrial Microbiology 14, 514–522.

Wagner-Döbler, I., Bennasar, A., Vancanneyt, M., Strömpl, C., Brümmer, I., Eichner, C.,Grammel, I., Moore, E.R., 1998. Microcosm enrichment of biphenyl-degradingmicrobial communities from soils and sediments. Applied and EnvironmentalMicrobiology 64, 3014–3022.

Xie, C.H., Yokota, A., 2005. Reclassification of Alcaligenes latus strains IAM 12599 T andIAM 12664 and Pseudomonas saccharophila as Azohydromonas lata gen. nov., comb.nov., Azohydromonas australica sp. nov. and Pelomonas saccharophila gen. nov.,comb. nov., respectively. International Journal of Systematic and EvolutionaryMicrobiology 55, 2419–2425.

Yao, J., Allen, C., 2006. Chemotaxis is required for virulence and competitive fitness ofthe bacterial wilt pathogen Ralstonia solanacearum. Journal of Bacteriology 188,3697–3708.

Yoshida, N., Takahashi, N., Hiraishi, A., 2005. Phylogenetic characterization of apolychlorinated-dioxin-dechlorinating microbial community by use of microcosmstudies. Applied and Environmental Microbiology 71, 4325–4334.

Zhang, L., Xie, G., 2007. Diversity and distribution of Burkholderia cepacia complex in therhizosphere of rice and maize. FEMS Microbiology Letters 266, 231–235.