lable at ScienceDirect

Food Microbiology 49 (2015) 23e32

Contents lists avai

Food Microbiology

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

The microbial diversity of an industrially produced lambic beer sharesmembers of a traditionally produced one and reveals a coremicrobiota for lambic beer fermentation

Freek Spitaels a, Anneleen D. Wieme a, b, Maarten Janssens c, Maarten Aerts a,Anita Van Landschoot b, Luc De Vuyst c, Peter Vandamme a, *

a Laboratory of Microbiology, Faculty of Sciences, Ghent University, K.L. Ledeganckstraat 35, B-9000 Ghent, Belgiumb Laboratory of Biochemistry and Brewing, Faculty of Bioscience Engineering, Ghent University, Valentin Vaerwyckweg 1, B-9000 Ghent, Belgiumc Research Group of Industrial Microbiology and Food Biotechnology (IMDO), Faculty of Sciences and Bioengineering Sciences, Vrije Universiteit Brussel,Pleinlaan 2, B-1050 Brussels, Belgium

a r t i c l e i n f o

Article history:Received 8 October 2014Received in revised form9 January 2015Accepted 27 January 2015Available online 7 February 2015

Keywords:Lambic beerAABLABYeastsSpontaneous fermentationMALDI-TOF MS

* Corresponding author. Tel.: þ32 9 264 51 13; fax:E-mail address: Peter.Vandamme@UGent.be (P. Va

http://dx.doi.org/10.1016/j.fm.2015.01.0080740-0020/© 2015 Elsevier Ltd. All rights reserved.

a b s t r a c t

The microbiota involved in lambic beer fermentations in an industrial brewery in West-Flanders,Belgium, was determined through culture-dependent and culture-independent techniques. More than1300 bacterial and yeast isolates from 13 samples collected during a one-year fermentation process wereidentified using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry followedby sequence analysis of rRNA and various protein-encoding genes. The bacterial and yeast communitiesof the same samples were further analyzed using denaturing gradient gel electrophoresis of PCR-amplified V3 regions of the 16S rRNA genes and D1/D2 regions of the 26S rRNA genes, respectively. Incontrast to traditional lambic beer fermentations, there was no Enterobacteriaceae phase and a largervariety of acetic acid bacteria were found in industrial lambic beer fermentations. Like in traditionallambic beer fermentations, Saccharomyces cerevisiae, Saccharomyces pastorianus, Dekkera bruxellensis andPediococcus damnosus were the microorganisms responsible for the main fermentation and maturationphases. These microorganisms originated most probably from the wood of the casks and were consid-ered as the core microbiota of lambic beer fermentations.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Lambic sour beers are among the oldest types of beers stillbrewed. They are the weakly carbonated products of a spontaneousfermentation process that lasts for one to three years beforebottling (De Keersmaecker, 1996). The sour character of the beeroriginates from the metabolic activities of lactic acid bacteria (LAB),acetic acid bacteria (AAB) and various yeasts (Spitaels et al., 2014c;Van Oevelen et al., 1977; Verachtert and Iserentant, 1995). Thesebeers can be drunk as such or are used to produce gueuze or fruitlambic beers. Except for an American coolship ales study based on16S rRNA gene sequence analysis (Bokulich et al., 2012), previousmicrobial studies on lambic beers only used phenotypic identifi-cation techniques (Van Oevelen et al., 1977; Verachtert and

þ32 9 264 50 92.ndamme).

Iserentant, 1995). A recent in-depth analysis of lambic brew sam-ples of the most traditional lambic brewery of Belgium revealed acharacteristic microbial succession of Enterobacteriaceae in the firstmonth (representing the first phase of lambic beer fermentation),Pediococcus damnosus, Saccharomyces cerevisiae and Saccharomycespastorianus after two months (reflecting the main fermentationphase), and Dekkera bruxellensis after six months (characteristic forthe maturation phase) (Spitaels et al., 2014c).

Although lambic beers were originally only brewed in the Senneriver valley and southeast of Brussels, they are now also brewedelsewhere in Belgium. InWest-Flanders, themost western provinceof Belgium and thus outside the Senne river valley, two indepen-dent breweries produce lambic beers. In the past, both breweriesobtained the necessary lambic wort and lambic beer from brew-eries located in the Senne river valley. One brewery filled theircasks with purchased wort to be able to produce their own lambicbeer. The other brewery purchased finished lambic beers to blendthem and produce gueuze beers. Because of the growing interest in

F. Spitaels et al. / Food Microbiology 49 (2015) 23e3224

beers of spontaneous fermentation, both breweries stopped buyinglambic wort or beer from traditional lambic breweries and startedto brew a lambic-type of beer from fresh wort, conform to thelambic beer production process.

The production activities of American craft breweries, includingAmerican coolship ales and other types of beers, resemble those ofthe industrial lambic breweries (Bokulich et al., 2012). The latterbreweries do not only produce lambic beers and products derivedthereof, but also the more typical ales and lager beer brands. In-dustrial lambic breweries mostly filter, pasteurize and carbonatetheir spontaneously fermented beers, which are sometimes alsosweetened (Van Oevelen et al., 1976). Moreover, they can brewlambic-type beers continuously, because they have the capacity toprechill the wort before its transfer into the open cooling tun andhence do not need the cold winter months to properly cool theirwort in one night as traditional lambic breweries do. Also, indus-trial brewers generally do not use old, small wine or cognac casksfor fermentation (2e6 hL). Instead, their wooden casks are usuallylarger and custom-made on-site (about 170e200 hL).

The present study aimed to characterize the microbial succes-sion in an industrial lambic beer fermentation process during oneyear, and to compare this succession of microorganismswith that ofa lambic beer fermentation in a traditional lambic brewery.

2. Materials and methods

2.1. Brewery

The selected brewery was an industrial lambic brewery locatedinWest-Flanders, approximately 70 km to the west of Brussels. Thisbrewery started to produce lambic beers in 1981. Before 1981, thisbrewery produced gueuze based on the blending of lambic beerspurchased from traditional lambic breweries.

2.2. Brewing process and sampling to study the succession of themicrobiota

Mash was prepared and boiled for 1.5 h in the brewery ac-cording to the brewer's recipe, which included acidification of thewort to pH 4 through the addition of lactic acid. This acidification iscommonly performed in all industrial lambic beer breweries. Afterthe acidification, the wort was prechilled to 40 �C and centrifugedto remove the hot break. The prechilled wort was then transferredinto a cleaned open cooling tun and a 500-mL sample was takenaseptically. A second 500-mL sample was taken from the wort inthe cooling tun after overnight cooling at the start of the worttransfer into the 170 hL cask. The transfer process required about8 h. The cooling tun was sampled a third time shortly before thewort was completely transferred into the wooden cask. Sampleswere taken from this cask after the transfer of the cooled wort andafter 1, 2 and 3 weeks and 1, 2, 3, 6, 9 and 12 months. Two lambicbatches were analyzed. Batch Awas brewed on January 4, 2011 andwas sampled at all time points mentioned above. The wort tem-perature of batch A after overnight cooling was about 22 �C. Batch Bwas brewed on July 27, 2010 and was sampled at the same timepoints for threemonths only. The wort temperature of batch B afterovernight cooling was about 29 �C. Two weeks after transfer of thebatch A wort into the cask there was no apparent production offoam, indicating that the fermentation process did not start, hencethe brewer decided to mix batch A (which is further referred to asthe acceptor batch A) with a 3-months old fermenting lambic wortfrom another batch (further referred to as the donor batch A) toinitiate the fermentation. Mixing occurred through the bottomapertures of the casks andwas performed in a ratio of 5 hL to 165 hL(±3%, vol/vol). Both the donor and acceptor batches Awere sampled

at the time of mixing, which is referred to below as the mixingpoint. The acceptor batch A was sampled prior to and 15 min aftermixing, enabling debris to settle.

All casks were already used several times for the production oflambic beers and were located in a single, separate building of thebrewery at ambient temperature and contained three apertures: amanhole at the top, closed with a loose panel, a valve at the bottomto fill and empty the cask, and a sampling tap located at about 1/3 ofthe total height of the cask. Before sampling, the tap was cleanedwith 70% (vol/vol) ethanol and approximately 100 mL of ferment-ing wort were discarded. Samples (500 mL) were collected in asterile bottle and transported on ice to the laboratory to be pro-cessed on the same day.

2.3. Denaturing gradient gel electrophoresis (DGGE) analysis

Crude brew samples (50 mL) were centrifuged at 8000 � g for10 min (4 �C) at the day of sampling and cell pellets were storedat �20 �C until further processing. DNA was prepared from thepellets as described by Camu et al. (2007). The DNA concentration,purity, and integrity were determined using 1% (wt/vol) agarosegels stained with ethidium bromide and by optical density (OD)measurements at 234, 260, and 280 nm. The quality of the DNAwas assessed as good, when absorbance ratios were OD260/OD280 > 1.8 and OD234/OD260 > 0.5. Total DNA solutions werediluted to an OD260 of 1. Amplification of about 200 bp of the V3region of the 16S rRNA genes with the F357 (with a GC clamp) andR518 primers, followed by denaturing gradient gel electrophoresis(DGGE) analysis, and processing of the resulting fingerprints wasperformed as described previously (Duytschaever et al., 2011),except that DGGE gels were run for 960 min instead of 990 min.For the amplification of about 200 bp of the D1 region of the LSU-rRNA genes of eukaryotic microorganisms, NL1 (with GC clamp)and LS2 primers were used, as previously reported by Cocolinet al. (2000).

All DNA bands were assigned to band classes using the Bio-Numerics 5.1 software (Applied Maths, Sint-Martens-Latem,Belgium). Dense DNA bands and/or bands that were present inmultiple fingerprints were excised from the polyacrylamide gels byinserting a pipette tip into the bands and subsequently transferredinto 40 mL 1� TE buffer (10 mM TriseHCl, 5 mM EDTA, pH 8) at 4 �Cfor the overnight elution of the DNA from the gel slices. The posi-tion of each extracted DNA band was confirmed by repeat DGGEexperiments using the excised DNA as template. The extracted DNAwas subsequently re-amplified and sequenced using the sameprotocol and primers (but without GC clamp). EzBioCloud andBLAST (Altschul et al., 1997; Kim et al., 2012) analyses were per-formed to determine the most similar sequences in the NCBIsequence databases.

2.4. Culture media, enumeration and isolation

Samples were serially diluted in 0.9% (wt/vol) saline and 50 mL ofeach dilution was plated in triplicate on multiple isolation mediasolidified with 1.5% agar. Bacterial isolationmedia [deMan-Rogosa-Sharpe (MRS) agar (Oxoid, Erembodegem, Belgium) (De Man et al.,1960), violet red bile glucose (VRBG) agar (Mossel et al., 1978, 1962)and acetic acid medium (AAM) agar (Lisdiyanti et al., 2003)] weresupplemented with 5 ppm amphotericin B (SigmaeAldrich, Bor-nem, Belgium) and 200 ppm cycloheximide (SigmaeAldrich) toinhibit fungal growth and were selected as described earlier(Spitaels et al., 2014c). Inoculated MRS agar plates were incubatedat 28 �C aerobically and at 20 �C anaerobically for the isolation ofLAB; inoculated VRBG agar plates were incubated at 28 �C aerobi-cally for the isolation of Enterobacteriaceae; and inoculated AAM

F. Spitaels et al. / Food Microbiology 49 (2015) 23e32 25

agar plates were incubated at 28 �C aerobically for the isolation ofAAB.

Yeast isolation media were supplemented with 100 ppmchloramphenicol (SigmaeAldrich) to inhibit bacterial growth andwere incubated aerobically at 28 �C. DYPAI agar (2.0% glucose, 0.5%yeast extract, 1.0% peptone and 1.5% agar; wt/vol) was used as ageneral yeast agar isolation medium. To favor growth of the slow-growing Dekkera/Brettanomyces, DYPAI was supplemented withan additional 50 ppm cycloheximide (DYPAIX) (Abbott et al., 2005;Licker et al., 1998; Su�arez et al., 2007). Furthermore, universal beeragar (Oxoid) was supplemented with 25% (vol/vol) commercialgueuze (Belle-Vue; AB Inbev, Anderlecht, Belgium) as recom-mended by the manufacturer and was used as an additional uni-versal yeast agar isolation medium (UBAGI).

Colonies on plates comprising 25e250 colony forming units(CFU) were counted after 3e10 days of incubation and for each ofthe seven isolation conditions about 20e25 colonies, or all coloniesif the counts were lower, were randomly picked up.

2.5. Matrix-assisted laser desorption/ionization time-of-flight massspectrometry (MALDI-TOF MS)-based dereplication andidentification

Isolates were subcultured twice using the respective isolationconditions and MALDI-TOF MS was performed using the thirdgeneration of pure cultures by means of a 4800 Plus MALDI TOF/TOF™ Analyzer (AB SCIEX, Framingham, MA, USA), as describedpreviously (Wieme et al., 2012). In short, Data Explorer 4.0 software(AB SCIEX) was used to convert the mass spectra into txt-files toimport them into a BioNumerics 5.1 (Applied Maths) database. Thespectral profiles were compared using the Pearson productemo-ment correlation coefficient (PPMCC) and a dendrogram was builtusing the unweighted pair group method with arithmetic mean(UPGMA) cluster algorithm. Homogeneous clusters consisting ofisolates with visually identical or virtually identical mass spectrawere delineated. From most clusters, isolates were randomlyselected for further identification through sequence analysis of 16SrRNA genes and other molecular markers. Sequence analyses of thepheS gene were performed to identify LAB (De Bruyne et al., 2008;De Bruyne et al., 2007; Naser et al., 2007, 2005) and of dnaK, groELand rpoB genes to identify AAB (Cleenwerck et al., 2010). Yeastisolates were identified through sequence analysis of the D1/D2region of the 26S rRNA gene (Kurtzman and Robnett, 1998) and,whenever needed, also by determination of ACT1 gene sequences(Daniel and Meyer, 2003). Some isolates of the present studygrouped in clusters of lambic isolates that were examined in aprevious study using the same polyphasic approach and weretherefore considered identified (Spitaels et al., 2014c).

All PCR assays were performed as described by Snauwaert et al.(2013). Bacterial DNA was obtained via the protocol described byNiemann et al. (1997), whereas yeast DNA was obtained using theprotocol of Harju et al. (2004).

2.6. Analysis of the microbiota in the brewery environment

To analyze the microbiota of the brewery environment, twosamples were taken from the cooling tun, the ceiling above thecooling tun, the walls and ceiling of the cellar and outside of thecasks, by swabbing a surface of about 100 cm2, using a moist swab.Furthermore, the inside of a cask that was not monitored duringthis study was swabbed as well. A first swab was streaked on eachof the agar isolation media; a second sample was transferred into5mL of saline and filtered through a 0.45-mm filter that was cut intopieces and subsequently transferred into 30 mL of MRS, VRBG,AAM, DYPAI and DYPAIX broth each, and incubated as described

above. Enrichment cultures that showed growth after 3e10 days ofincubation were subcultured on their respective agar media andmorphologically distinct colonies were selected for further analysis.Isolates were identified as described above. Air samples were takenusing a MAS-100 air sampler (Merck, Darmstadt, Germany) with aflow rate of 0.1 m3/min placed about 1 m above the floor, for 1 or10 min using yeast and bacterial agar isolation media, respectively.

2.7. Nucleotide sequence accession numbers

The Genbank accession numbers for the sequences generated inthis study are KJ541146eKJ541153.

3. Results

For both batches A and B, neither the cooling tun samples northe first wort sample obtained from the cask yielded DNA for DGGEanalysis. Bacterial and yeast DNA was successfully extracted fromall subsequent samples and DGGE amplicons were generated fromsamples taken from one week onwards (Fig. 1). Table 1 provides anoverview the enumeration analyses. For both batches, the freshlyboiled wort sample did not yield growth and there was no apparentinoculation of the cooling tun samples taken after overnight cool-ing. Moreover, neither the early samples of both batches nor thoseof the remainder of the fermentation process yielded growth ofEnterobacteriaceae and these bacteria were not detected in theDGGE community profiles.

3.1. Microbiota succession in batch A

3.1.1. First week of fermentationThe bacterial community profiles of the batch A sample taken

after one week revealed a single dense band in the low % G þ Cregion (Fig. 1A, band class 1). Sequence analysis (SupplementaryFig. S1 and Supplementary Table S1) demonstrated that thisdense band originated from a yeast, Hanseniaspora sp., whichconfirmed that the V3 primers also amplify some eukaryotic DNA(Scheirlinck et al., 2008; Spitaels et al., 2014c; Van der Meulen et al.,2007). Two dense DNA bands, originating from Hanseniaspora(Fig. 1B, band class 6) and Candida/Pichia (Fig. 1B, band class 7)strains were present in the corresponding yeast community pro-files. An overview of the microbiota identified through DGGE bandsequencing is provided in Supplementary Table S2. An overview ofthe identification results of isolates (and the numbers of isolatesinvestigated) per MALDI-TOF MS cluster is presented inSupplementary Table S3. Cultivation experiments of the one-weekold sample yielded primarily yeasts while bacterial counts werelow to zero (Table 1). Yeast isolates were identified as Hansenias-pora uvarum and Pichia fermentans (Fig. 2), which confirmed theDGGE results.

3.1.2. Mixing pointH. uvarum remained the most isolated species in the 2-weeks

old sample prior to the mixing point, along with Pi. fermentansand D. bruxellensis (Fig. 2). All MRS agar isolates were identified asGluconobacter cerinus, whereas AAM agar isolates were identifiedas Acetobacter lambici (Fig. 3) (Spitaels et al., 2014a). The bacterialand yeast community profiles of the donor batch A samplecomprised a single dense band each (Fig. 1A, band class 2 andFig. 1B, band class 8, respectively), which both originated fromSaccharomyces strains (Supplementary Table S2).

Immediately after themixing point, the acceptor batch A sampleyielded several additional DNA bands that originated from AAB(Fig. 1A, bands grouped in band class box 4) and one faint band thatoriginated from Pediococcus/Lactobacillus (Fig. 1A, band class 3). In

Fig. 1. DGGE community profiles of both bacteria and yeast communities of batch A and batch B. DGGE banding patterns of the bacterial and yeast communities of batch A (A and B,respectively) and batch B (C and D, respectively); w, week(s); m, month(s). Band classes 1e20 are indicated with numbers and some are grouped in a band class box. Samples onlyyielded DNA and PCR amplicons after one week of fermentation. 1. Hanseniaspora sp.; 2. Saccharomyces sp.; 3. Pediococcus/Lactobacillus; 4. AAB; 5. Pediococcus/Lactobacillus; 6.Hanseniaspora sp.; 7. Candida/Pichia; 8. Saccharomyces sp.; 9. Saccharomyces sp.; 10. Saccharomyces sp.; 11. Dekkera sp.; 12. Hanseniaspora sp.; 13. Kregervanrija sp.; 14. Dekkera sp.;15. Saccharomyces sp.; 16. Pediococcus/Lactobacillus; 17. AAB; 18. Pediococcus/Lactobacillus; 19. Saccharomyces sp.; 20. Dekkera sp. *These bands did not yield PCR amplicons afterband excision and subsequent DNA amplification. The 35e70 % denaturing gradient is represented from left to right on the gels.

F. Spitaels et al. / Food Microbiology 49 (2015) 23e3226

the yeast community profiles, two new Saccharomyces bands (bandclasses 8 and 9) were present (Fig. 1B). Bacterial counts of the donorbatch Awere generally equal to or lower than those of the acceptorbatch A, except for counts on anaerobically incubated MRS whichwere higher (Table 1). Most donor batch sample isolates fromaerobically incubated MRS agar were identified as P. damnosus(Fig. 3), while this was the only species isolated from anaerobicallyincubated MRS agar. AAM agar isolates belonged to Acetobacterorientalis and Acetobacter fabarum (Fig. 3). Donor batch A containedmore cycloheximide-sensitive yeasts than the acceptor batch A [asrevealed by the difference in colony counts on DYPAI and UBAGIversus DYPAIX agar media (Table 1)]. DYPAI and UBAGI agar isolatesfrom the donor batch A sample yielded the cycloheximide-sensitiveS. cerevisiae and S. pastorianus (Fig. 2), while DYPAIX agar isolateswere identified as D. bruxellensis, Hanseniaspora meyeri andH. uvarum (Supplementary Fig. S2). Immediately after the mixingpoint, bacterial counts on aerobically incubated MRS and AAM agarmedia increased, which is likely explained by the presence ofsettled microbiota in either the donor or acceptor batch A that was

at least partially resuspended, due to the physical mixing of thecask contents (Table 1). H. uvarum and Pi. fermentans remained theonly yeasts isolated from the acceptor batch A (Fig. 2). Glucono-bacter cerevisiae (Spitaels et al., 2014b) was the only isolated speciesfrom AAM agar, although it was not isolated from previous samplesnor from the donor batch A (Fig. 3). Aerobically incubated MRS agaryielded G. cerinus isolates only; anaerobically incubated MRS agaryielded P. damnosus isolates only (Fig. 3).

3.1.3. Three weeks of fermentationAlthough foam was produced and the fermentation therefore

had started one week after the mixing point, the enumeration re-sults showed no profound changes (Table 1). Not unexpectedlybecause of its presence in the donor cask A, S. cerevisiae was iso-lated in addition toH. uvarum and Pi. fermentans from the yeast agarisolation media (Fig. 2). A diverse group of AAB species was iso-lated: A. lambici, G. cerevisiae, G. cerinus and A. fabarum (Fig. 3). Thefew colonies picked from anaerobically incubatedMRS (n¼ 4) wereidentified as P. damnosus.

Table 1Plate counts on different agar isolation media under different incubation conditions. MRS agar was used for the growth of LAB, AAM agar for the growth of AAB, DYPAI andUBAGI agars were used as general yeast growth media and DYPAIX agar was used to favor the growth of Dekkera species. VRBG agar, used for the growth of Enterobacteriaceae,did not yield any growth. Values represent log CFU/mL ± standard deviation. ULD: under limit of detection (<20 CFU/mL); ULQ: under limit of quantification (the estimatedCFU/mL is provided between brackets).

MRS 28 �C MRS 20 �C AN AAM 28 �C DYPAI 28 �C UBAGI 28 �C DYPAIX 28 �C

Batch AFreshly boiled wort ULD ULD ULD ULD ULD ULD1 night cooling tun ULD ULD ULD ULD ULD ULD1 night cask ULQ (20) ULD ULD ULD ULD ULD1 week ULQ (20) ULD ULD 6.85 ± 0.08 6.88 ± 0.02 6.83 ± 0.052 weeks (before mixing point) 3.13 ± 0.06 ULD 3.03 ± 0.13 6.51 ± 0.08 6.42 ± 0.09 6.63 ± 0.13Donor batch 3.61 ± 0.05 3.80 ± 0.06 ULQ (440) 5.57 ± 0.05 5.62 ± 0.01 3.90 ± 0.032 weeks (after mixing point) 4.95 ± 0.12 ULQ (40) 5.12 ± 0.02 6.72 ± 0.10 6.87 ± 0.13 6.46 ± 0.083 weeks 5.37 ± 0.05 ULQ (50) 5.35 ± 0.08 6.62 ± 0.05 6.66 ± 0.13 6.48 ± 0.041 month 5.31 ± 0.06 ULQ (200) 5.92 ± 0.08 6.06 ± 0.01 5.97 ± 0.08 5.50 ± 0.062 months 4.35 ± 0.09 3.38 ± 0.10 5.05 ± 0.07 5.98 ± 0.05 5.95 ± 0.05 3.25 ± 0.063 months 6.98 ± 0.12 6.78 ± 0.06 3.98 ± 0.01 6.37 ± 0.03 6.44 ± 0.02 ULQ (300)6 months 6.59 ± 0.04 6.53 ± 0.04 ULQ (20) 4.76 ± 0.06 4.75 ± 0.03 4.94 ± 0.149 months ULQ (480) 2.83 ± 0.06 ULD 3.84 ± 0.17 3.83 ± 0.02 3.73 ± 0.1012 months 4.38 ± 0.05 4.33 ± 0.05 ULD 3.90 ± 0.17 3.30 ± 0.09 2.88 ± 0.06Batch BFreshly boiled wort ULD ULD ULD ULD ULD ULD1 night cooling tun ULD ULD ULD ULD ULD ULD1 night cask ULQ (70) ULD ULD 3.00 ± 0.08 3.00 ± 0.09 ULQ (100)1 week ULD ULD ULQ (20) 5.92 ± 0.08 5.92 ± 0.03 ULQ (20)2 weeks ULQ (390) ULD ULQ (410) 6.24 ± 0.05 6.33 ± 0.06 ULD3 weeks 5.37 ± 0.02 ULQ (200) 5.28 ± 0.01 6.05 ± 0.09 6.13 ± 0.03 ULD1 month 5.56 ± 0.04 5.49 ± 0.05 5.46 ± 0.03 5.95 ± 0.04 5.93 ± 0.05 ULQ (100)2 months 7.56 ± 0.05 6.63 ± 0.05 4.81 ± 0.02 5.04 ± 0.03 4.51 ± 0.05 5.02 ± 0.063 months 7.40 ± 0.03 7.44 ± 0.04 ULQ (20) 4.62 ± 0.08 4.62 ± 0.08 4.64 ± 0.06

F. Spitaels et al. / Food Microbiology 49 (2015) 23e32 27

3.1.4. One month of fermentationOne month after brewing, the composition of the cultivable

microbiota of the samples changed remarkably (Fig. 1). While Pi.fermentans, S. cerevisiae, S. pastorianus and H. uvarumwere isolatedfrom DYPAI and UBAGI agar, H. uvarum remained the only yeastspecies isolated from DYPAIX agar (Fig. 2 and SupplementaryFig. S2). Aerobically incubated MRS and AAM agars yieldedlargely the samemicroorganisms as those identified oneweek after

Fig. 2. Identification of random isolates from DYPAI and UBAGI agars of ba

the mixing point (Fig. 3) and P. damnosus remained the onlymicroorganism isolated from anaerobically incubated MRS agar.

3.1.5. Second and third month of fermentationThe number of colonies on anaerobically incubated MRS agar

gradually increased to 106 CFU/mL (Table 1). In contrast, colonycounts on aerobically incubated MRS agar showed a decrease in thesecondmonth, followed by an increase in the thirdmonth (Table 1).

tch A cask samples. The number of isolates is given between brackets.

Fig. 3. Identification of random isolates from MRS and AAM agars of batch A cask samples. The identification of anaerobically incubated MRS agar isolates is not shown, as allisolates were identified as Pediococcus damnosus. The number of isolates is given between brackets.

F. Spitaels et al. / Food Microbiology 49 (2015) 23e3228

At the 2-months sampling point, primarily P. damnosuswas isolatedfrom both aerobically and anaerobically incubated MRS agar; AABincluded A. lambici, A. fabarum G. cerinus, and G. cerevisiae (Fig. 3).The decrease and subsequent increase in colony counts on aero-bically incubated MRS agar are likely caused by the decrease of AABand the subsequent increase of LAB, as shown by the colony countson anaerobically incubated MRS (Table 1).

Changes in the microbial communities led to the presence ofnew bands in the bacterial community profiles of batch A after 3months of fermentation (Fig. 1A, bands grouped in band class box5), which originated from Pediococcus/Lactobacillus(Supplementary Table S2). In the yeast community profiles, Han-seniaspora (Fig. 1B, band class 6) and two SaccharomycesDNA bands(Fig.1B, band classes 8 and 9) were present, although band classes 6and 9 were no longer detected at 3 months. In the 2-months oldsample, multiple new Saccharomyces DNA bands were present(Fig. 1B, bands grouped in banding box 10). At 2 months, Pi. fer-mentans and S. pastorianus were isolated from DYPAI and UBAGIagars (Fig. 2), whereas H. uvarum remained the only species iso-lated from DYPAIX agar (Supplementary Fig. S2). The former twospecies were the main yeast species isolated at 3 months, in addi-tion to a small number of S. cerevisiae isolates (Fig. 2). BothD. bruxellensis and H. uvarum were isolated from DYPAIX agar(Supplementary Fig. S2). Band class 11 (Fig. 1B) originated fromDekkera strains and was first detected in the yeast communityprofiles at 3 months.

3.1.6. Six months to one year of fermentationAfter the 3-months sampling point, Pediococcus/Lactobacillus

remained the single dominant bacterium in the fermentation pro-cess, while AAB virtually disappeared (Fig. 1A and Fig. 3). Bacterialcounts on MRS agar remained high (106 CFU/mL) until 6 months offermentation, but then started to decrease (Table 1). Yeast countsstarted to decrease from 3 months onwards. In the 6-months oldand subsequent samples, D. bruxellensiswas the main yeast species

(Fig. 2). Band class 8 (Saccharomyces) (Fig. 1B) was the mainremaining band at 6 months, together with two faint bands (Fig. 1B,bands grouped in band class box 12) assigned toHanseniaspora. Theyeast community profiles of the samples at 9 and 12 months weresimilar: band class 8 was no longer present, band class 11 reap-peared and two new bands, originating from Kregervanrija (bandclass 13) andDekkera (band class 14) strains were present. After oneyear, D. bruxellensis, Wickerhamomyces anomalus and Yarrowia lip-olytica were isolated from DYPAIX agar (Supplementary Fig. S2).

3.2. Microbiota succession in batch B

The early bacterial and yeast DGGE community profiles of batchB contained only a single dense band each (Fig. 1C, band class 15;Fig. 1D, band class 19), which were both assigned to Saccharomycesstrains. Only DYPAI and UBAGI agars yielded substantial growthimmediately after the transfer of the wort into the cask (103 CFU/mL; Table 1). Isolates from these samples were identified as Pichiakudriavzevii (n ¼ 8 isolates examined) and Debaryomyces hansenii(n ¼ 1). Isolates from MRS agar were all identified as A. orientalis(n ¼ 7).

After one week, yeast counts increased to 105 CFU/mL (Table 1).Isolates from DYPAI and UBAGI agar media were identified asS. cerevisiae (n ¼ 38) and Pi. kudriavzevii (n ¼ 2), whereas only oneisolate, identified as D. bruxellensis, was obtained fromDYPAIX agar.Bacterial counts were low (Table 1) and only one bacterial isolate,identified as A. orientalis, was obtained from MRS agar.

Band class 16 (originating from Pediococcus/Lactobacillus) waspresent in the bacterial DGGE community profile of the 2-weeksold sample and increased in intensity in later samples until itvirtually disappeared after one month (Fig. 1C). In the 3-weeks and1-month samples, one band originating from AAB could be detec-ted in the high % G þ C region (Fig. 1C, band class 17). Although thebacterial community profiles changed during the second month ofthe fermentation, all DNA bands (Fig. 1C, bands grouped in band

F. Spitaels et al. / Food Microbiology 49 (2015) 23e32 29

class box 18) were assigned to Pediococcus/Lactobacillus. The bac-terial community profiles after 3 months of fermentation werehighly similar to the bacterial community profile obtained from thesample of the 6-months fermented batch A. Yeast communityprofiles were nearly identical and DNA bands originated fromSaccharomyces (Fig. 1D, band class 19) and Dekkera (Fig. 1D, bandclass 20) strains.

Colony counts on all yeast agar isolationmediawere comparableafter three months (about 104 CFU/mL; Table 1). S. cerevisiae(n ¼ 46) and Pi. kudriavzevii (n ¼ 17) were the main yeast speciesisolated during the first month. D. bruxellensis (n ¼ 5) wasincreasingly recovered from one month of fermentation onwardsand was the only isolated yeast species (n ¼ 36) in the samplestaken after 2 and 3months of fermentation. From 2weeks onwards,A. fabarum was the sole AAB species that could be isolated fromboth AAM agar (n ¼ 23) and aerobically incubated MRS agar(n ¼ 48). This species was isolated up to 3 months of fermentationand counts on AAM agar reached a maximum of 105 CFU/mL after 3weeks, but decreased below the level of quantification at 3 months(Table 1). From 3 weeks onwards, P. damnosus was isolated fromMRS agar and it was the most isolated bacterial species during theremainder of the fermentation (n ¼ 57).

3.3. Microbiota of the brewery environment

No yeasts or bacteria could be recovered from samples of thebrewery ceilings, walls and cooling tun surface; in addition, themicroorganisms that were isolated from air samples were notfound in the lambic beer fermentation process, such as Klebsiellaoxytoca, Bacillus spp. and Staphylococcus spp. (Table 2). In contrast,swab samples taken from both inside and outside of the caskssurfaces yielded several species found in the fermenting lambic. P.damnosus, D. bruxellensis and Dekkera anomala (Table 2), microor-ganisms isolated frequently from 6 months of fermentation on-wards, were isolated from the inside of a cleaned cask and weretherefore readily present when the wort entered the cask. S. cer-evisiae and S. pastorianus were not isolated from air samples norfrom the casks surfaces.

4. Discussion

S. cerevisiae, S. pastorianus, D. bruxellensis and P. damnosus weredominating the lambic beer fermentation processes in the presentstudy. In addition, a considerable diversity of AAB species waspresent, unlike in traditional lambic beer fermentations (Spitaelset al., 2014c). In the latter lambic beer fermentations, the micro-bial inoculation occurs during the overnight cooling of the wort inthe cooling tun and the acidification of the wort is mediatedthrough the action of acid-producing microorganisms. In contrast,the industrial lambic beer fermentation process of the presentstudy was steered or started spontaneously as soon as the acidifiedchilled wort received microorganisms from the surroundings whenit was transferred into the cask. The acidification of the wort afterboiling prevented the growth of Enterobacteriaceae (Priest andStewart, 2006), which contrasted with both traditional lambicbeer as well as American coolship ale fermentations, where thisgroup of bacteria is dominantly present in the cooled wort samplefrom the cooling tun until the end of the first month of fermenta-tion (Bokulich et al., 2012; Spitaels et al., 2014c; Van Oevelen et al.,1977; Verachtert and Iserentant, 1995). A sluggishly startingfermentation, such as batch A in the present study, can be steeredby adding fermenting wort from another lambic batch in the samefermentation phase. The identification of H. uvarum in the sluggishbatch A confirms earlier reports [in which its asexual name,Kloeckera apiculata (Meyer et al., 1978), was used (Van Oevelen

et al., 1977; Verachtert and Iserentant, 1995)]. H. uvarum was notdetected in batch B nor in a recent study of the lambic beerfermentation in a traditional brewery (Spitaels et al., 2014c) and,therefore, its presence is most probably not necessary for lambicbeer fermentation. This species has a low fermentative capacity andis commonly found during the spontaneous fermentation of winesand cider, where its contribution to flavor complexity is increas-ingly appreciated (Bezerra-Bussoli et al., 2013; de Arruda MouraPietrowski et al., 2012; Valles et al., 2007).

Since the Enterobacteriaceae phase was absent, the first phase ofthe industrial lambic beer fermentation processes studied was themain fermentation phase. The latter was characterized by thepresence of Saccharomyces spp., which were isolated in batch Auntil the third month. The ratio of S. pastorianus to S. cerevisiaeincreased for a reason that is not known but may be related tofermentation temperature adaptation (Cousseau et al., 2013;Vidgren et al., 2010). The latter phenomenon was also seen dur-ing a traditional lambic beer fermentation studied previously(Spitaels et al., 2014c). The slow onset and pace of the mainfermentation phase in batch A caused a late proliferation ofD. bruxellensis and hence of the maturation phase in batch A(Spitaels et al., 2014c; Van Oevelen et al., 1977; Verachtert andIserentant, 1995). In batch B, S. cerevisiae was the most isolatedyeast species during the main fermentation phase, untilD. bruxellensis became predominant after one month.

The different start of the two lambic beer batches may beexplained by differences in the temperature of the chilled wortafter overnight cooling and by differences in the environmentaltemperatures. All casks in the industrial brewery were similar inheight, were exclusively used for lambic beer fermentation, andwere stored at ambient temperature at the same location. There-fore, these environmental factors were more uniform than those ofa traditional lambic brewery, where casks are smaller, have a pre-vious use in wine or cognac production, and are often located indifferent rooms (Spitaels et al., 2014c). However, the wort tem-peratures after overnight cooling differed by 7 �C (22 �C versus29 �C) and the ambient temperatures of batch A that started inJanuary (winter) and batch B that started in July (summer) werevery different. Therefore, the present paper hypothesizes that bothtemperature factors likely influenced the successful initiation andpace of the main fermentation phase as well as the start of thematuration phase. Since thewort had an initial pH of 4 after boiling,the pH drop caused by the production of lactic acid by LAB was lesspronounced compared to a non-acidified wort. In agreement withthis, the start of the maturation phase, which is characterized bythe presence of D. bruxellensis, occurred after one month offermentation in batch B. Simultaneously, P. damnosus was the onlyisolated bacterium. This effect was not apparent in batch A, wherethe maturation phase occurred from 6 months onwards, but whichwas most probably masked by the delayed start of the fermentationaltogether. Similarly, recent studies of the microbiology of sponta-neous beer fermentation processes did not reveal an extendedacidification phase (Bokulich et al., 2012; Spitaels et al., 2014c).Hence, the acidification and maturation phases seemed to proceedsimultaneously and it may be more appropriate to consider thispart of the lambic beer fermentation as a single long maturationphase, characterized by the presence of D. bruxellensis andP. damnosus.

Opportunistic contaminants such as Candida parapsilosis,D. anomala,W. anomalus, and Y. lipolyticawere occasionally isolatedfrom fermenting wort samples. C. parapsilosis was previously re-ported as a wild yeast isolated from lager beers, but failed to growin wort or beer, and was therefore regarded as a contaminant (Vander Aa Kühle and Jespersen, 1998). The yeast species W. anomalusand Y. lipolytica were also isolated at the end of a lambic beer

Table 2Overview of the microorganisms isolated from the brewery environment and their isolation sources.

Accessionnumber

Accession numberclosest hit

Similarity(%)

Present infermentation

Air atticafter cooling

Air attic beforecooling

Aircellar

Caskexterior

Caskinterior

Bacteriaa

Aerococcus urinaeequi D87677 100 þBacillus aerophilus AJ831844 100 þBacillus aryabhattai EF114313 100 þBacillus licheniformis AE017333 100 þBacillus simplex AB363738 100 þBacillus subtilis AMXN01000021 100 þKlebsiella oxytocad AB004754 100 þKocuria kristinae X80749 100 þLactococcus lactis AE005176 100 þLeuconostoc citreum KJ541152 AF111948 99 þLysinibacillus macroides AJ628749 100 þPediococcus damnosus þ þPediococcus pentosaceusb AM749815 100 þPropionibacterium

cyclohexanicumKJ541151 D82046 99 þ

Propionibacterium thoenii KJ541153 AJ704572 98 þPseudomonas azotoformans D84009 100 þRumeliibacillus pycnus AB271739 100 þStaphylococcus aureus D83355 100 þStaphylococcus caprae AB009935 100 þ þStaphylococcus epidermidis L37605 100 þStaphylococcus haemolyticus L37600 100 þStaphylococcus hominis X6601 100 þ þStaphylococcus petrasii AY953148 100 þStaphylococcus saprophyticus AP008934 100 þStaphylococcus succinus AF004220 100 þStaphylococcus warneri L37603 100 þStreptococcus parauberis NR_043001 100 þYeastsc

Blastobotrys arbuscula DQ442689 100 þCryptococcus carnescens AB035054 100 þDebaryomyces hanseniid JQ689041 100 þDekkera anomala þ þDekkera bruxellensis þ þTrichosporon domesticum JN939449 100 þa Identification was based on the 16S rRNA gene sequence.b Identification was based on the pheS gene sequence.c Identification was based on the D1/D2 26S rRNA gene sequence.d Identification was confirmed by MALDI-TOF MS, clustering together with isolates obtained during a previous study (Spitaels et al., 2014c).

F. Spitaels et al. / Food Microbiology 49 (2015) 23e3230

fermentation in a traditional lambic brewery (Spitaels et al., 2014c).Y. lipolytica primarily occurs in dairy and meat products, but also insoil and wastewaters (Knutsen et al., 2007). The typical presence ofthis yeast at a late stage of lambic beer fermentation may suggest aspecific role and deserves further attention.

As mentioned above, traditional lambic beers are assumed tobe spontaneously inoculated by the air microbiota of the Senneriver valley during the overnight cooling in the cooling tun(Martens et al., 1991; Verachtert and Iserentant, 1995). However,whereas air samples of the industrial lambic brewery harboredmicroorganisms not relevant for lambic beer fermentation, noneof the cooling tun samples studied yielded DNA or microbialgrowth. This indicated that the cooling tun samples were sterileor that very low numbers of microorganisms were present.Therefore, the microbiota must have been inoculated when thewort entered the cask, and therefore it must originate from thecask wood or from residues of the previous fermentation batches.Casks are indeed cleaned superficially using only a pressurewasher to remove yeast and bacterial clumps from the ceiling,sides and bottom of the casks. Unlike in traditional breweries, noefforts are made to kill the residual microbiota using, for instance,steam or other sanitizing agents. Again contrasting with tradi-tional lambic breweries, the industrial brewery uses anti-fungalpaint on all walls and ceilings in the brewery. S. cerevisiae andS. pastorianus are responsible for the main fermentation phase butwere not isolated at the end of the fermentation process. Yet,

these yeasts and bacteria may penetrate into the wood andeffectively form a biofilm in the wood of the casks (Swaffield andScott, 1995; Swaffield et al., 1997). Therefore, the present paperhypothesizes that Saccharomyces yeasts may remain present inthe cask wood and thus may survive the maturation phase to re-emerge when fresh wort enters the cask. Likewise, AAB maysurvive in the cask wood. As P. damnosus, D. bruxellensis andD. anomala were isolated from the inside of the casks, thesespecies could enter the wort directly after the transfer from thecooling tun into the casks.

5. Conclusion

The present study demonstrated that industrial and traditionallambic beer fermentations involve the same main actors, includingS. cerevisiae, S. pastorianus, P. damnosus and D. bruxellensis, whichconfirms and extends previous observations (Spitaels et al., 2014c;Van Oevelen et al., 1977; Verachtert and Iserentant, 1995). Thesespecies could therefore be regarded as the core microbiota oflambic beer fermentation. Although these main actors were thesame in both breweries, the present data showed a differentfermentation profile in the industrial brewery compared to thetraditional brewery studied. This was mainly due to the absence ofthe Enterobacteriaceae phase, as the industrially producedwort wasacidified after wort boiling. Furthermore, P. damnosus was the soleLAB species detected, indicating a unique adaptation of this species

F. Spitaels et al. / Food Microbiology 49 (2015) 23e32 31

to grow under the harsh conditions of the lambic beer fermentationprocess.

Acknowledgments

The authors thank the brewery and brewers involved in thisstudy for their generous contributions of lambic wort and beersamples.

This research was funded by a Ph.D. grant of the Agency forInnovation by Science and Technology (IWT) and by the ResearchFoundation Flanders (FWO-Vlaanderen). The authors furtheracknowledge their finances from the research fund of the Univer-sity College Ghent (ADW), the Vrije Universiteit Brussel (HOA, SRP,IRP, and IOF projects; MJ and LDV), and the Hercules Foundation.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.fm.2015.01.008.

References

Abbott, D.A., Hynes, S.H., Ingledew, W.M., 2005. Growth rates of Dekkera/Bret-tanomyces yeasts hinder their ability to compete with Saccharomyces cer-evisiae in batch corn mash fermentations. Appl. Microbiol. Biotechnol. 66,641e647.

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 databasesearch programs. Nucleic Acids Res. 25, 3389e3402.

Bezerra-Bussoli, C., Baffi, M.A., Gomes, E., Da-Silva, R., 2013. Yeast diversity isolatedfrom grape musts during spontaneous fermentation from a Brazilian winery.Curr. Microbiol. 67, 356e361.

Bokulich, N.A., Bamforth, C.W., Mills, D.A., 2012. Brewhouse-resident microbiota areresponsible for multi-stage fermentation of American coolship ale. Plos One 7,e35507.

Camu, N., De Winter, T., Verbrugghe, K., Cleenwerck, I., Vandamme, P.,Takrama, J.S., Vancanneyt, M., De Vuyst, L., 2007. Dynamics and biodiversity ofpopulations of lactic acid bacteria and acetic acid bacteria involved in spon-taneous heap fermentation of cocoa beans in Ghana. Appl. Environ. Microbiol.73, 1809e1824.

Cleenwerck, I., De Vos, P., De Vuyst, L., 2010. Phylogeny and differentiation ofspecies of the genus Gluconacetobacter and related taxa based on multilocussequence analyses of housekeeping genes and reclassification of Acetobacterxylinus subsp. sucrofermentans as Gluconacetobacter sucrofermentans (Toyo-saki et al. 1996) sp. nov., comb. nov. Int. J. Syst. Evol. Microbiol. 60,2277e2283.

Cocolin, L., Bisson, L.F., Mills, D.A., 2000. Direct profiling of the yeast dynamics inwine fermentations. FEMS Microbiol. Lett. 189, 81e87.

Cousseau, F.E., Alves Jr., S.L., Trichez, D., Stambuk, B.U., 2013. Characterization ofmaltotriose transporters from the Saccharomyces eubayanus subgenome of thehybrid Saccharomyces pastorianus lager brewing yeast strain Weihenstephan34/70. Lett. Appl. Microbiol. 56, 21e29.

Daniel, H.M., Meyer, W., 2003. Evaluation of ribosomal RNA and actin gene se-quences for the identification of ascomycetous yeasts. Int. J. Food Microbiol. 86,61e78.

de Arruda Moura Pietrowski, G., dos Santos, C.M., Sauer, E., Wosiacki, G.,Nogueira, A., 2012. Influence of fermentation with Hanseniaspora sp. yeaston the volatile profile of fermented apple. J. Agric. Food Chem. 60,9815e9821.

De Bruyne, K., Franz, C.M., Vancanneyt, M., Schillinger, U., Mozzi, F., de Valdez, G.F.,De Vuyst, L., Vandamme, P., 2008. Pediococcus argentinicus sp. nov. fromArgentinean fermented wheat flour and identification of Pediococcus species bypheS, rpoA and atpA sequence analysis. Int. J. Syst. Evol. Microbiol. 58,2909e2916.

De Bruyne, K., Schillinger, U., Caroline, L., Boehringer, B., Cleenwerck, I.,Vancanneyt, M., De Vuyst, L., Franz, C.M., Vandamme, P., 2007. Leuconostocholzapfelii sp. nov., isolated from Ethiopian coffee fermentation and assessmentof sequence analysis of housekeeping genes for delineation of Leuconostocspecies. Int. J. Syst. Evol. Microbiol. 57, 2952e2959.

De Keersmaecker, J., 1996. The mystery of lambic beer. Sci. Am. 275, 74e81.De Man, J., Rogosa, M., Sharpe, M.E., 1960. A medium for the cultivation of lacto-

bacilli. J. Appl. Microbiol. 23, 130e135.Duytschaever, G., Huys, G., Bekaert, M., Boulanger, L., De Boeck, K., Vandamme, P.,

2011. Cross-sectional and longitudinal comparisons of the predominant fecalmicrobiota compositions of a group of pediatric patients with cystic fibrosis andtheir healthy siblings. Appl. Environ. Microbiol. 77, 8015e8024.

Harju, S., Fedosyuk, H., Peterson, K.R., 2004. Rapid isolation of yeast genomic DNA:Bust n' Grab. BMC Biotechnol. 4, 8.

Kim, O.S., Cho, Y.J., Lee, K., Yoon, S.H., Kim, M., Na, H., Park, S.C., Jeon, Y.S., Lee, J.H.,Yi, H., Won, S., Chun, J., 2012. Introducing EzTaxon-e: a prokaryotic 16S rRNAgene sequence database with phylotypes that represent uncultured species. Int.J. Syst. Evol. Microbiol. 62, 716e721.

Knutsen, A.K., Robert, V., Poot, G.A., Epping, W., Figge, M., Holst-Jensen, A., Skaar, I.,Smith, M.T., 2007. Polyphasic re-examination of Yarrowia lipolytica strains andthe description of three novel Candida species: Candida oslonensis sp. nov.,Candida alimentaria sp. nov. and Candida hollandica sp. nov. Int. J. Syst. Evol.Microbiol. 57, 2426e2435.

Kurtzman, C.P., Robnett, C.J., 1998. Identification and phylogeny of ascomycetousyeasts from analysis of nuclear large subunit (26S) ribosomal DNA partial se-quences. Antonie Leeuwenhoek 73, 331e371.

Licker, J., Acree, T., Henick-Kling, T., 1998. What is “brett” (Brettanomyces) flavor?: apreliminary investigation. In: Waterhouse, A., Ebeler, S. (Eds.), Chemistry ofWine Flavor. ACS Publications, Washington, DC, USA, pp. 96e115.

Lisdiyanti, P., Katsura, K., Potacharoen, W., Navarro, R.R., Yamada, Y., Uchimura, T.,Komagata, K., 2003. Diversity of acetic acid bacteria in Indonesia, Thailand, andthe Philippines. Microbiol. Cult. Collect. 19, 91e98.

Martens, H., Dawoud, E., Verachtert, H., 1991. Wort enterobacteria and othermicrobial-populations involved during the 1st month of lambic fermentation.J. Inst. Brew. 97, 435e439.

Meyer, S.A., Smith, M.T., Simione Jr., F.P., 1978. Systematics of Hanseniaspora zikesand Kloeckera janke. Antonie Leeuwenhoek 44, 79e96.

Mossel, D., Elederink, I., Koopmans, M., Van Rossem, F., 1978. Optimalisation of aMacConkey-type medium for the enumeration of Enterobacteriaceae. Lab. Pract.27, 1049e1050.

Mossel, D., Mengerink, W., Scholts, H., 1962. Use of a modified MacConkey agarmedium for the selective growth and enumeration of Enterobacteriaceae.J. Bacteriol. 84, 381.

Naser, S.M., Dawyndt, P., Hoste, B., Gevers, D., Vandemeulebroecke, K.,Cleenwerck, I., Vancanneyt, M., Swings, J., 2007. Identification of lactobacilli bypheS and rpoA gene sequence analyses. Int. J. Syst. Evol. Microbiol. 57,2777e2789.

Naser, S.M., Thompson, F.L., Hoste, B., Gevers, D., Dawyndt, P., Vancanneyt, M.,Swings, J., 2005. Application of multilocus sequence analysis (MLSA) for rapididentification of Enterococcus species based on rpoA and pheS genes. Microbi-ology 151, 2141e2150.

Niemann, S., Puhler, A., Tichy, H.V., Simon, R., Selbitschka, W., 1997. Evaluation ofthe resolving power of three different DNA fingerprinting methods todiscriminate among isolates of a natural Rhizobium meliloti population. J. Appl.Microbiol. 82, 477e484.

Priest, F.G., Stewart, G.G., 2006. Microbiology and Microbiological Control in theBrewery, Handbook of Brewing, Second ed. CRC Press, Boca Raton, FL, USA,pp. 607e629.

Scheirlinck, I., Van der Meulen, R., Van Schoor, A., Vancanneyt, M., De Vuyst, L.,Vandamme, P., Huys, G., 2008. Taxonomic structure and stability of the bacterialcommunity in Belgian sourdough ecosystems as assessed by culture and pop-ulation fingerprinting. Appl. Environ. Microbiol. 74, 2414e2423.

Snauwaert, I., Papalexandratou, Z., De Vuyst, L., Vandamme, P., 2013. Character-ization of strains of Weissella fabalis sp. nov. and Fructobacillus tropaeoli fromspontaneous cocoa bean fermentations. Int. J. Syst. Evol. Microbiol. 63,1709e1716.

Spitaels, F., Li, L., Wieme, A., Balzarini, T., Cleenwerck, I., Van Landschoot, A., DeVuyst, L., Vandamme, P., 2014a. Acetobacter lambici sp. nov. isolated from fer-menting lambic beer. Int. J. Syst. Evol. Microbiol. 64, 1083e1089.

Spitaels, F., Wieme, A.D., Balzarini, T., Cleenwerck, I., Van Landschoot, A., DeVuyst, L., Vandamme, P., 2014b. Gluconobacter cerevisiae sp. nov.isolated from the brewery environment. Int. J. Syst. Evol. Microbiol. 64,1134e1141.

Spitaels, F., Wieme, A.D., Janssens, M., Aerts, M., Daniel, H.M., Van Landschoot, A., DeVuyst, L., Vandamme, P., 2014c. The microbial diversity of traditional sponta-neously fermented lambic beer. Plos One 9, e95384.

Su�arez, R., Su�arez-Lepe, J., Morata, A., Calder�on, F., 2007. The production of ethyl-phenols in wine by yeasts of the genera Brettanomyces and Dekkera: a review.Food Chem. 102, 10e21.

Swaffield, C.H., Scott, J.A., 1995. Existence and development of natural microbialpopulations in wooden storage vats used for alcoholic cider maturation. J. Am.Soc. Brew. Chem. 53, 117e120.

Swaffield, C.H., Scott, J.A., Jarvis, B., 1997. Observations on the microbial ecology oftraditional alcoholic cider storage vats. Food Microbiol. 14, 353e361.

Valles, B.S., Bedrinana, R.P., Tascon, N.F., Simon, A.Q., Madrera, R.R., 2007. Yeastspecies associated with the spontaneous fermentation of cider. Food Microbiol.24, 25e31.

Van der Aa Kühle, A., Jespersen, L., 1998. Detection and identification of wild yeastsin lager breweries. Int. J. Food Microbiol. 43, 205e213.

Van der Meulen, R., Scheirlinck, I., Van Schoor, A., Huys, G., Vancanneyt, M.,Vandamme, P., De Vuyst, L., 2007. Population dynamics andmetabolite target analysis of lactic acid bacteria during laboratoryfermentations of wheat and spelt sourdoughs. Appl. Environ. Microbiol.73, 4741e4750.

Van Oevelen, D., L'Escaille, F., Verachtert, H., 1976. Synthesis of aroma componentsduring the spontaneous fermentation of lambic and gueuze. J. Inst. Brew. 82,322e326.

F. Spitaels et al. / Food Microbiology 49 (2015) 23e3232

Van Oevelen, D., Spaepen, M., Timmermans, P., Verachtert, H., 1977. Microbiologicalaspects of spontaneous wort fermentation in the production of lambic andgueuze. J. Inst. Brew. 83, 356e360.

Verachtert, H., Iserentant, D., 1995. Properties of Belgian acid beers and theirmicroflora. Part I. The production of gueuze and related refreshing acid beers.Cerevisia, Belg. J. Brew. Biotechnol. 20, 37e41.

Vidgren, V., Multanen, J.-P., Ruohonen, L., Londesborough, J., 2010. The temperaturedependence of maltose transport in ale and lager strains of brewer's yeast.FEMS Yeast Res. 10, 402e411.

Wieme, A., Cleenwerck, I., Van Landschoot, A., Vandamme, P., 2012. Pediococcus loliiDSM 19927T and JCM 15055T are strains of Pediococcus acidilactici. Int. J. Syst.Evol. Microbiol. 62, 3105e3108.