International Journal of Food Microbiology 163 (2013) 71–79

Contents lists available at SciVerse ScienceDirect

International Journal of Food Microbiology

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

The lactic acid bacteria and yeast microbiota of eighteen used for the manufacture of traditional Italian sweet leavened baked goods

Anna Lattanzi a, Fabio Minervini a,⁎, Raffaella Di Cagno a, Annamaria Diviccaro a, Livio Antonielli b, Gianluigi Cardinali b, Stefan Cappelle c, Maria De Angelis a, Marco Gobbetti a a Department of Scienze del Suolo, della Pianta e degli Alimenti, University of Bari Aldo Moro, via Amendola 165/a., Bari, 70126, b Department of Applied Biology, Microbiology Division, University of Perugia, via Borgo 20 Giugno, Perugia, 74-06121, Italy c Beldem S.A., Rue Borrie 12, Andenne, Belgium article info abstract

Article history: The lactic acid bacteria and yeast microbiota of eighteen sourdoughs used for the manufacture of some traditional Received 27 September 2012 Italian sweet leavened baked goods were studied through culture-dependent method and pyrosequencing. Received in revised form 7 February 2013 Flours used for back slopping and sourdoughs were also biochemically characterized. Principal component Accepted 22 February 2013 analysis was applied to explore eventual correlations between process parameters applied during back slopping, Available online 1 March 2013 some flour nutrients, profile of microbiota, and biochemical characteristics of sourdoughs. The median values of Keywords: the cell density of lactic acid bacteria and yeasts were 8.05 and 7.03 log CFU/g, respectively. As shown by fi culture-dependent method, Lactobacillus sanfranciscensis was identi ed in all the sourdoughs, except for Lactic acid bacteria Panaredda and Torcolo di San Costanzo. For eleven sourdoughs, all the lactic acid bacteria isolates were allotted Yeasts to this species. For Buccellato di Lucca, Mbriagotto, , and Nadalin sourdoughs, at least 80% of the isolates was allotted to this species. Other lactic acid bacteria isolated with a relatively high frequence were Lactobacillus plantarum and Leuconostoc citreum.Pyrosequencingconfirmed and complemented the culture-dependent approach, detecting L. sanfranciscensis also in Panaredda and Torcolo di San Costanzo sourdoughs. Saccharomyces cerevisiae was identified in all the sourdoughs, except for Mbriagotto, Ciambella di Mosto and Pandolce Genovese. These latter sourdoughs harbored strains of Candida humilis,whereasfive sourdoughs combined the presence of both yeast species. Positive correlations were found between time of back slopping and cell density and main metabolites of lactic acid bacteria. Percentage of sourdough used as inoculum was mainly correlated with the cell density of yeasts and the concentration of ethanol. This study provided a comprehensive and comparative approach to highlight the dominant microbiota of Italian sourdoughs, which could be exploited further to guarantee a highly reproducible quality of the Italian sweet goods studied, while preserving their traditional trait. © 2013 Elsevier B.V. All rights reserved.

1. Introduction (Gobbetti, 1998; Zannini et al., 2010). Sourdough is a preparation of flour and water, containing high numbers of metabolically active lactic The annual production of Italian sweet leavened baked goods is acid bacteria and yeasts (Gobbetti, 1998; Hammes and Gaenzle, estimated to be ca. 31033 t (ISTAT, 2011). Several sweet leavened 1998). According to traditional procedure, dominant microorganisms baked goods are consumed mainly in the occasion of some religious are kept metabolically active by back slopping (De Vuyst and Neysens, feasts (e.g., Panettone, Pandoro and Nadalin, and di Pasqua, 2005). More than 50 species of lactic acid bacteria, especially belonging Panaredda and Colomba for Christmas and Easter, respectively). to the genus Lactobacillus, and more than 20 species of yeasts, mostly Despite their geographical origin (e.g., Lumbardy for Panettone), belonging to the genera Saccharomyces and Candida, were found during sweet leavened baked goods have, nowadays, a national and interna- sourdough propagation for making traditional/typical leavened baked tional diffusion (ISTAT, 2011). goods (Corsetti et al., 2001; Minervini et al., 2012a; Palomba et al., Traditional sweet leavened baked goods share the use of a sour- 2011; Venturi et al., 2012). Whereas the sourdough microbiota of tradi- called “madre” (mother sponge) as the basis for their production tional breads was largely studied (Catzeddu et al., 2006; Minervini et al., 2012a; Reale et al., 2011; Valmorri et al., 2010; Zotta et al., 2008), that of ⁎ Corresponding author. Tel.: +39 080 5442948; fax: +39 080 5442911. sweet leavened baked goods received little attention. Although most E-mail address: [email protected] (F. Minervini). studies involved a limited number of products, a certain degree of

0168-1605/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ijfoodmicro.2013.02.010 72 A. Lattanzi et al. / International Journal of Food Microbiology 163 (2013) 71–79 microbial diversity was found in the sourdoughs used for manufacturing 2. Material and methods sweet leavened baked goods (Foschino et al., 2004; Garofalo et al., 2008; Palomba et al., 2011; Venturi et al., 2012). 2.1. Sourdoughs The microbial composition of sourdough is influenced by several parameters. Specific technological process parameters include pH and Eighteen sourdoughs used for the manufacture of some traditional redox potential, dough hydration and yield, number of sourdough Italian sweet leavened baked goods, representative of several Italian back slopping steps, fermentation time between back slopping, leave- regions (Fig. S1), were studied. Except for sourdough used for making ning and storage temperature, the use of starters and/or baker's yeast, Cornetto di semola (T. durum flour), all sourdoughs were made with and the percentage of starter cultures, or sourdough, or baker's yeast T. aestivum flour. The only ingredients used for preparing the sour- used as inoculum. Compared to the above parameters, microbiological, were: flour, sourdough coming from a previous fermentation, chemical and enzymatic composition of the flour and the house and tap water. The sourdoughs mainly differed in terms of the following microbiota may be regarded as parameters which are not fully under process parameters (Table 1): percentage (expressed on total dough the control whoever manages the sourdough (De Vuyst et al., 2009). weight) of sourdough used as inoculum; number of back slopping It has been recently reported that the type of flour (Triticum durum or steps necessary to get the “mature” sourdough; time of back slopping; Triticum aestivum), as well as the concentration of some flour nutrients and temperature of back slopping. Three batches of each sourdough (fermentable carbohydrates and free amino acids) required by the were collected in three consecutive days at local bakeries (Weckx microorganisms dominating the sourdough ecosystem, may have a et al., 2010). All samples were taken immediately at the end of the last key role for selecting the population of lactic acid bacteria (Minervini back slopping (mature sourdough), i.e. before being used as leavening et al., 2012a). Consequently, the microbiota structure distinguished agent of doughs containing sugar, eggs, and butter. Mature sourdoughs almost all the sourdoughs used for the manufacture of nineteen were stored at 4 °C for 12 h before analyses. Eighteen sourdoughs were traditional/typical Italian breads (Minervini et al., 2012a). Overall, the analyzed in three different days in duplicate. sourdoughs used for manufacturing sweet products mainly differ from those used for manufacturing breads in terms of: (i) higher inoculum ratio; (ii) shorter fermentation time; and (iii) higher number of back 2.2. Measurement of pH and determination of carbohydrates, organic slopping steps. acids, ethanol and free amino acids This study was aimed at describing the composition of the lactic acid bacteria and yeast microbiota of eighteen sourdoughs, which The values of pH were measured by a pH-meter. The water- are used for the manufacture of traditional Italian sweet leavened soluble extracts were obtained as previously described (Minervini baked goods. Dominating lactic acid bacteria and yeasts were et al., 2012a) and were used for determining the concentrations of monitored by culture-dependent and -independent methods. Possible carbohydrates, organic acids and ethanol through HPLC (Zeppa et al., correlations between some flour nutrients (soluble carbohydrates 2001) and the concentrations of individual free amino acids and free amino acids), process parameters and composition of the (FAA) through a Biochrom 30 Amino Acid Analyser (Biochrom LTD, sourdough microbiota, and between composition of the microbiota Cambridge Science Park, UK) (De Angelis et al., 2007). The quotient of and the biochemical characteristics of sourdoughs, were explored fermentation (QF) was calculated as the molar ratio between D,L-lactic through principal component analysis. and acetic acids.

Table 1 Process parameters of sourdoughs used for the manufacture of traditional Italian sweet leavened baked goods.

Sweet leavened Code Sourdough No. of back slopping Time (h) and temp (°C) of back sloppingb baked good (% on total dough weight)a steps before use

Buondì SA 40 3 4.5, 26 Panettone (Puglia) SB 29 (1st back slopping); 2 4.5, 28 38 (2nd back slopping) Buccellato di Lucca SC 50 2 6, 25 Cornetto di semola SD 40 1 9, 27 Pizza di Pasqua SE 40 2 5, 25 Panaredda SG 42 4 4, 25 Torcolo di San Costanzo SH 40 1 3.5, 25 Mbriagotto SI 40 4 8, 22 (1st back slopping); 3, 25 (all the other back slopping steps) Pandoro SL 40 3 4, 28 Nadalin SM 40 5 3, 28 (1st and 2nd back slopping steps); 16, 4 (3rd back slopping); 2, 28 (4th and 5th back slopping steps) Panettone () SN 40 3 4, 28 Resta di Como SO 32 2 4, 28 Panettone basso SP 25 (1st back slopping); 2 4, 25 (1st back slopping); 6, 25 (2nd back slopping) 31 (2nd back slopping) Biscotti del Lagaccio SQ 45 (1st back slopping); 23,28 41 (2nd back slopping) Pandolce Genovese SR 40 4 3.5, 30 (1st, 2nd, and 3rd backslopping steps); 10, 12 (4th backslopping) Colomba SS 41 3 3, 30 Veneziana ST 40 3 2.5, 30 Ciambella di Mosto SU 40 1 4, 23

a Percentage of sourdough used during back slopping. b First number indicates the length of back slopping (h), and the second number indicates the temperature (°C) of incubation. Mature sourdough, after the last back slopping, was subjected to analyses. A. Lattanzi et al. / International Journal of Food Microbiology 163 (2013) 71–79 73

2.3. Enumeration and isolation of lactic acid bacteria and yeasts were purified as previously described (Minervini et al., 2012a)and sequencedbyABIPRISM3730xlwithprimersITS1(5′-TCCGTAGGTG Cell densities of lactic acid bacteria were determined by plating AACCTGCGG) and NL4 (5′-GGTCCGTGTTTCAAGACGG) (O'Donnell, serially diluted sourdoughs (Minervini et al., 2012a) on Sourdough 1993). Sequencing electropherogram data were processed with Bacteria (SDB) and MRS5 agar media at 30 °C for 48 h. Each liter of Geneious (http://www.geneious.com). MRS5 medium contained 10 g of tryptone, 5 g of meat extract, 5 g of yeast extract, 10 g of maltose, 5 g of fructose, 5 g of glucose, 5 g of C2H3NaO2 ∙ 3H2O, 3 g of NH4Cl, 2.6 g of K2HPO4 ∙ 3H2O, 4 g of 2.6. Pyrosequencing analysis KH2PO4, 0.1 g of MgSO4 ∙ 7H2O, 0.05 g of MnSO4∙4H2O, 0.5 g of cysteine-HCl, cobalamin, folic acid, nicotinic acid amide, pantothenic Ninety milliliters of 50 mM potassium phosphate buffer, pH 7.0, acid, pyridoxal phosphate, and thiamine (2 × 10−4 g each) and 1 ml were added to 10 g of sourdough and homogenized for 5 min. DNA of Tween 80 (pH 5.8) (Meroth et al., 2003). Both SDB and MRS5 were extraction from sourdoughs was carried out as described by Minervini supplemented with cycloheximide (0.1 g/l). Cell densities of yeasts et al. (2010). DNA was dried through a vacuum centrifuge (SPD121P were estimated by plating serially diluted sourdoughs on Sabouraud SpeedVac Concentrator, Thermo Scientific, NC) and used as template Dextrose agar (SDA) (Oxoid) medium, supplemented with chloram- for bacterial 16S tag-encoded FLX Titanium amplicon pyrosequencing phenicol (0.1 g/l), at 30 °C for 48 h. analysis. Pyrosequencing was carried out using primers Gray28F For each medium (SDB and MRS5), five to fifteen colonies were (5-TTTGATCNTGGCTCAG) and Gray519r (5-GTNTTACNGCGGCKGCTG) randomly selected from the plates containing the two highest sample to amplify a 520-bp fragment of the V1–V3 region of the 16S rRNA dilutions. Ten to thirty colonies of presumptive yeasts were randomly gene (Andreotti et al., 2011). Tag-encoded FLX amplicon pyrosequencing selected from the highest SDA plate dilutions. Pure cultures of analyses utilized a Roche 454 FLX instrument with Titanium reagents, presumptive lactic acid bacteria and yeasts were isolated and stocked and Titanium procedures were carried out at the Research and Testing as previously described (Minervini et al., 2012a). Yeast isolates were Laboratory (RTL, Lubbock, TX) based on RTL protocols. Following assayed for the capacity to ferment galactose, sucrose, maltose, raffinose sequencing, all failed sequence reads, low-quality sequence ends, tags, and trehalose, and for the capacity to grow at 37 °C (Barnett et al., and primers were removed, and sequence collections were depleted 2000). of any nonbacterial ribosome sequences and chimeras using B2C2 (Gontcharova et al., 2010), as described previously by Callaway et al. 2.4. Genotypic characterization by RAPD–PCR analysis (2010). To determine the identities of bacteria in the remaining sequences, the sequences were denoised, assembled into clusters, and Genomic DNA of lactic acid bacteria was extracted according to used to query a database of high-quality 16S bacterial sequences derived De Los Reyes-Gavilán et al. (1992). Lactic acid bacteria isolates were from NCBI with a distributed BLASTn algorithm (Dowd et al., 2005). biotyped by means of three oligonucleotides: P4 5′-CCGCAGCGTT-3′, Database sequences were characterized as high quality based on similar P7 5′-AGCAGCGTGG-3′ (Corsetti et al., 2003), and M13 5′-GAGGG criteria used by RDP, version 9. By using a .NET and C# analysis pipeline, TGGCGGTTCT-3′ (Zapparoli et al., 1998). RAPD–PCR profiles were the resulting BLASTn outputs were compiled, validated using taxonomic acquired and their similarity was evaluated as previously described distance methods, and subjected to data reduction analysis, as described (Minervini et al., 2012a). The reproducibility of RAPD fingerprints by Andreotti et al. (2011). Sequences with identity scores greater than was assessed by comparing the PCR products obtained from three 97% (3% divergence) for known or well-characterized 16S rRNA gene separate cultures of the same strain. Based on the reproducibility sequences were resolved at the genus or, when possible, at the species of the RAPD–PCR analysis, isolates showing significantly (P b 0.10) level. The identities of all hits were greater than 98%. These parameters different profiles were subjected to genotypic identification. were previously evaluated to enable reliable identification at least at Genomic DNA from yeast isolates was extracted as described the genus level (Dowd et al., 2008). The percentage of each bacterial by Cardinali et al. (2001). RAPD–PCR analysis with Inter Simple species was analyzed individually for each sample, providing a relatively

Sequence Repeats (ISSR) primer (GACA)4 was used to discriminate abundant information among the samples based on the relative numbers yeasts at the level of strain. PCR protocol was carried out according of reads within each (Andreotti et al., 2011). to Andrade et al. (2006). (GACA)4–ISSR profiles were converted in a binary matrix and cluster analysis was carried out with ADE4 package (Dray and Dufour, 2007). Hierarchical clustering was calculated 2.7. Statistical analysis by means of DIANA algorithm (Kaufman and Rousseeuw, 1990) and then a dendrogram was built up. Data (at least three biological replicates) of maltose, glucose, fructose, FAA, pH, organic acids and ethanol, and cell density of 2.5. Genotypic identification of lactic acid bacteria and yeasts lactic acid bacteria and yeasts were subjected to one-way ANOVA, and pair-comparison of treatment means was achieved by Tukey's Identification of lactic acid bacteria strains was carried out using procedure at P b 0.05, using the statistical software Statistica 7.0 two primer pairs (Invitrogen Life Technologies, Milan, Italy), LacbF/ for Windows. Principal component analysis (PCA), using a correla- LacbR and LpCoF/LpCoR, to amplify 16S rRNA gene fragment of tion matrix, was carried out to find the effect of different variables lactic acid bacteria (De Angelis et al., 2006). Primers designed on the distribution of the sourdoughs (Scheirlinck et al., 2007). on recA gene were also used to distinguish Lactobacillus plantarum, Data from the concentration of some flour nutrients (maltose, Lactobacillus pentosus and Lactobacillus paraplantarum species using glucose, fructose, total and individual FAA), parameters used multiplex PCR (Torriani et al., 2001). Primers pheS (Naser et al., 2005) during back slopping (percentage of sourdough used as inoculum, were used to identify at the level of species within the genus Leuconostoc time, temperature and number of back slopping), biochemical (Ouadghiri et al., 2009). Primers casei/para were used to discriminate characteristics (pH, concentrations of lactic and acetic acids, among the species Lactobacillus casei, Lactobacillus paracasei and ethanol and total FAA, QF, residual concentrations of carbohy- Lactobacillus rhamnosus (Ward and Timmins, 1999). drates) of sourdoughs and microbial community (cell density of Genomic DNA of yeasts was extracted as described above. lactic acid bacteria and yeasts, number of strains, percentage of The ITS1, 5.8S, ITS2 and D1/D2 domains of the 26S rDNA were amplified obligately homofermentative, and obligately and facultatively using primers ITS5 (5′-GGAAGTAAAAGTCGTAACAAGG) and LR6 (5′-CG heterofermentative lactic acid bacteria) were used as variables CCAGTTCTGCTTACC) as indicated by White et al (1990). Amplicons for PCA analyses. 74 A. Lattanzi et al. / International Journal of Food Microbiology 163 (2013) 71–79

3. Results and discussion The concentration of maltose, glucose and fructose of the flours showed median values of ca. 0.26, 0.32 and 0.44% (Table S1), respectively. 3.1. Process parameters, biochemical and microbiological characteristics After sourdough fermentation, the median values were ca. 0.56, 0.34 and of sourdoughs 0.35% (Table S2) for maltose, glucose and fructose, respectively. Based on these values, it may be issued that soluble carbohydrates did not repre- Because the approach used in this study was similar to that previously sent a limiting factor for microbial growth during sourdough fermen- used to characterize the microbiota of nineteen sourdoughs used for the tation. By considering that carbohydrates are continuously liberated manufacture of traditional/typical Italian breads (Minervini et al., 2012a), during fermentation by especially flour endogenous enzymes, it was an initial comparison, based on the process parameters applied for not possible to estimate their consumption at the end of fermentation. getting the mature sourdough, is noteworthy. In general, the percentage The concentration of total free amino acids (FAA) of the flours of inoculum and the temperature of back slopping of sweet baked good ranged from ca. 904 (Panaredda sourdough) to ca. 1385 mg/kg sourdoughs was higher (median value ca. 39.39% and ca. 25 °C, (Panettone basso sourdough) (Table S1). The median value was ca. respectively) than those found for sourdoughs used for the manufacture 1101 mg/kg. The eighteen sourdoughs used for making traditional of traditional/typical Italian breads (median value ca. 23.75% and 22 °C, Italian sweet leavened baked goods contained variable concentra- respectively) (Minervini et al., 2012a). On the contrary, the time of back tions of FAA. They ranged from ca. 368 (Biscotti del Lagaccio sour- slopping was lower than that used for sourdough breads (median value dough) to ca. 2576 (Buondì sourdough) mg/kg (Tables 2 and S2). of ca. 4.72 vs. 10.20 h, respectively). Such differences in the fermentation The median value was ca. 1169 mg/kg, higher than that found for schemes are probably due to the need of the pastry-makers of favoring sourdoughs used for the manufacture of traditional/typical Italian yeasts over lactic acid bacteria activity.Indeed,comparedtobread,the breads (median value ca. 591 mg/kg) (Minervini et al., 2012a). leavening of a dough containing “rich” ingredients (mainly sugar, butter Although glutamine is by far the most abundant amino acid in wheat and eggs) is more challenging for the dominant microbiota of sourdough proteins, glutamic acid was present at the highest concentration (Venturi et al., 2012). Furthermore, for most of sweet goods considered among FAA (Table S2). Similar results were previously reported for in this study, one or two final long (6–12 h) fermentation steps are sourdoughs used for manufacturing traditional/typical Italian breads, required. This increases the risk of getting sourness, which is highly such as Pane di Altamura PDO, Pane di Laterza, Pane di Matera PGI, undesirable in sweet goods (personal communication from Luigi Biasetto Pane di Cappelli, Moddizzosu, and Pane Carasau (Minervini et al., and Vincenzo Benvenuto). The only way for decreasing such a risk is to 2012a). This could be attributed to the conversion of glutamine into limit the activity of lactic acid bacteria starting from the first steps glutamate by lactic acid bacteria, contributing to the tolerance of these (from one to five back slopping steps using just flour and water) of the microorganisms to the acid conditions of sourdough (Vermeulen et al., protocol of production. Consequently, the median value of pH of 2007). FAA, other than glutamic acid, present at high concentrations the sourdoughs used for manufacturing sweet goods was higher in most (eleven of eighteen) of the sourdoughs considered in this than that found for the sourdoughs used for bread production (4.38 study were leucine, aspartic acid and arginine (Table S2). vs. 3.97) (Minervini et al., 2012a). In details, the sourdoughs consid- Cell densities of lactic acid bacteria (estimated on SDB agar) and ered in this study showed values of pH, which ranged from ca. 3.90 yeasts in the eighteen sourdoughs studied are shown in Table 3.The (Pizza di Pasqua sourdough) to ca. 5.01 (Panaredda sourdough) cell density of lactic acid bacteria varied from ca. 6.3 (Panettone basso (Table 2). The concentration of D,L-lactic acid was in agreement sourdough) to ca. 9.2 (Buondì sourdough) log CFU/g (median value with the values of pH and it ranged from ca. 42 mM (Panaredda sour- of 8.05 log CFU/g). Except for sourdoughs used for manufacturing dough) to ca. 85 mM (Buccellato di Lucca sourdough) (median value Cornetto di semola and Resta di Como, which were characterized by a of ca. 67 mM). The median value for the concentration of acetic acid low cell density of yeast, this parameter ranged from ca. 6.5 (Pandoro was ca. 11 mM, with variations from ca. 5 (Panaredda sourdough) to sourdough) to ca. 7.6 (Panaredda sourdough) log CFU/g (median ca. 27 mM (Pizza di Pasqua sourdough). Quotient of fermentation (QF) value of 7.03 log CFU/g). The sourdough used for making Panaredda rangedfromca.3.0toca.9.3(medianvalueofca.5.2).Ethanolranged was atypical. Indeed, besides showing a relatively high value of pH from ca. 0.09 (Resta di Como sourdough) to ca. 35 M (Panaredda sour- (5.01), it was characterized by an unusual dominance of yeasts versus dough) (median value of ca. 0.16 M). lactic acid bacteria. The use of two different culture media to count

Table 2 Mean values (and standard deviation) of pH, concentration of organic acids (mM), quotient of fermentation, concentration of ethanol (M) and total free amino acids (mg/kg) of sourdoughs used for the manufacture of traditional Italian sweet leavened baked goods.

Sourdoughsa pH Lactic acid (mM) Acetic acid (mM) Quotient of fermentation Ethanol (M) Free amino acids (mg/kg)

Buondì 4.03 (0.04)j 76 (2)b 19 (1)b 4.0 (0.1)h 0.32 (0.03)a 2576 (27)a Panettone (Puglia) 4.04 (0.02)j 70 (4)cd 19 (2)b 3.7 (0.2)h 0.20 (0.01)de 1689 (19)d Buccellato di Lucca 3.93 (0.04)k 85 (4)a 21 (2)b 4.0 (0.2)h 0.27 (0.01)b 2149 (24)b Cornetto di semola 4.45 (0.01)de 74 (1)b 20 (3)b 3.7 (0.5)h 0.10 (0.02)ij 1533 (6)ef Pizza di Pasqua 3.90 (0.02)k 82 (1)a 27 (2)a 3.0 (0.1)i 0.26 (0.01)bc 1529 (11)ef Panaredda 5.01 (0.05)a 42 (2)i 5 (1)f 8.4 (0.1)bc 0.35 (0.02)a 1278 (22)g Torcolo di San Costanzo 4.40 (0.04)ef 62 (3)ef 8 (3)ef 7.7 (1.8)cd 0.16 (0.02)fg 452 (25)l Mbriagotto 4.20 (0.03)h 74 (2)b 11 (2)de 6.7 (0.8)d 0.17 (0.03)ef 686 (15)k Pandoro 4.15 (0.01)i 75 (4)b 15 (3)c 5.0 (0.6)fg 0.12 (0.01)hi 790 (27)i Nadalin 4.62 (0.03)b 60 (4)f 11 (1)de 5.4 (0.1)f 0.16 (0.02)fg 754 (20)j Panettone (Lombardia) 4.36 (0.04)fg 60 (2)f 7 (2)f 8.6 (1.7)ab 0.17 (0.01)ef 1051 (18)h Resta di Como 4.50 (0.02)cd 55 (3)g 10 (2)def 5.5 (0.7)ef 0.09 (0.01)j 810 (25)i Panettone basso 4.57 (0.02)bc 50 (1)h 11 (3)de 4.5 (0.9)gh 0.12 (0.02)hi 818 (17)i Biscotti del Lagaccio 4.30 (0.01)g 68 (1)cd 11 (3)de 6.2 (1.3)de 0.16 (0.02)fg 368 (30)m Pandolce Genovese 4.40 (0.03)ef 73 (2)bc 11 (1)de 6.6 (0.3)d 0.13 (0.02)ghi 1561 (16)e Colomba 4.45 (0.04)de 65 (4)de 7 (1)f 9.3 (0.7)a 0.15 (0.01)fgh 2082 (20)c Veneziana 4.50 (0.05)cd 62 (2)ef 13 (2)cd 4.8 (0.5)fg 0.22 (0.01)cd 1060 (21)h Ciambella di Mosto 4.37 (0.03)fg 66 (1)de 13 (2)cd 5.1 (0.6)fg 0.15 (0.03)fgh 1516 (14)f a–mValues within a column with different superscript letters are significantly different (P b 0.05). a Sourdoughs are indicated with the names of the related sweet baked goods. A. Lattanzi et al. / International Journal of Food Microbiology 163 (2013) 71–79 75

Table 3 fingerprints was assessed by comparing the PCR products obtained Mean values (and standard deviation) of cell densities (log CFU/g) of lactic acid bacteria with primers P4, P7, and M13 and DNA extracted from three separate (estimated on SDB agar) and yeasts (estimated on SDA) in sourdoughs used for the cultures of the same strain. For this purpose, 10 strains were studied, manufacture of traditional Italian sweet leavened baked goods. and the patterns for the same strain were similar at a level of ca. 90% a Sourdoughs Lactic acid bacteria Yeasts (data not shown), as estimated by UPGMA analysis. Cluster analysis of Buondì 9.2 (0.07)a 7.5 (0.05)ab the RAPD profiles revealed that diversity among isolates ranged from ca. Panettone (Puglia) 8.9 (0.10)bc 7.3 (0.08)bc 1.5 to 25.5%. At the 15.2% of diversity, ca. 93% of the strains grouped into ab bc Buccellato di Lucca 9.1 (0.06) 7.3 (0.06) twenty-four clusters (I–XXIV) (Fig. S2). Overall, clustering of the strains Cornetto di semola 8.4 (0.08)d 5.2 (0.07)h Pizza di Pasqua 9.1 (0.09)ab 7.2 (0.05)cd was not related to sourdoughs they were isolated from. The only Panaredda 6.7 (0.10)j 7.6 (0.06)a exceptions were strains isolated from Pizza di Pasqua (cluster XXI) Torcolo di San Costanzo 8.7 (0.08)c 7.1 (0.05)cde and Buccellato di Lucca (cluster XXIV) sourdoughs (Fig. S2). e cde Mbriagotto 8.1 (0.09) 7.1 (0.06) Strains were identified by partial sequence analysis of 16S rRNA, Pandoro 8.0 (0.07)ef 6.5 (0.09)g recA,andpheS genes. Except for Panaredda and Torcolo di San Costanzo, Nadalin 7.7 (0.09)g 7.0 (0.06)de fi Panettone (Lombardia) 7.1 (0.06)i 7.0 (0.06)de Lactobacillus sanfranciscensis was identi ed in all the sourdoughs Resta di Como 7.2 (0.08)hi 4.9 (0.10)i (Table 4). For eleven out of eighteen sourdoughs, all the lactic acid Panettone basso 6.3 (0.07)k 6.6 (0.09)fg bacteria isolates were allotted to this species. For Buccellato di Lucca, fg ef Biscotti del Lagaccio 7.8 (0.08) 6.8 (0.11) Mbriagotto, Pandoro, and Nadalin sourdoughs, at least 80% of the Pandolce Genovese 8.7 (0.11)c 7.0 (0.09)de Colomba 8.1 (0.06)e 6.9 (0.06)e isolates was allotted to this species. L. sanfranciscensis is the key Veneziana 7.4 (0.05)h 6.9 (0.05)e lactic acid bacterium in most of type I sourdoughs. The dominance of Ciambella di Mosto 7.4 (0.08)h 7.2 (0.1)cd L. sanfranciscensis could be ascribed to suitable process parameters a–kValues within a column with different superscript letters are significantly different (e.g., back slopping practices, leavening and storage temperature and/ (P b 0.05). or pH of the dough) and to the intrinsic capacity of adaptation to the a Sourdoughs are indicated with the names of the related sweet baked goods. sourdough environment (Gobbetti and Corsetti, 1997). The genome analysis of L. sanfranciscensis TMW 1.1304 showed that this bacterium presumptive lactic acid bacteria showed significant differences for responds quickly to favorable conditions of its most favorable environ- six sourdoughs only: Pandoro, Nadalin, Panettone (Lombardia), Resta ment (sourdough), and suddenly initiates fermentative metabolism di Como, Panettone basso (cell density estimated on MRS5 higher and fast growth (Vogel et al., 2011). than that estimated on SDB), and Pandolce Genovese (cell density Other lactic acid bacteria identified in this study were: L. plantarum estimated on MRS5 lower than that estimated on SDB). Furthermore, (Panaredda, Torcolo di San Costanzo, Mbriagotto, Pandoro, and Nadalin no presumptive lactic acid bacteria were detectable by plating 1 g sourdoughs), Leuconostoc citreum (Cornetto di semola sourdough), of Buccellato di Lucca sourdough on MRS5 agar medium (data not Weissella cibaria (Buccellato di Lucca sourdough), Lactobacillus brevis shown). Such differences could be attributed to qualitative and quan- (Torcolo di San Costanzo sourdough), Lactobacillus curvatus (Cornetto titative differences among SDB and MRS5 agar media in terms of di semola and Panaredda sourdoughs), L. casei (Torcolo di San Costanzo nutrients and to different metabolic capacities among strains harbored sourdough), Lactococcus lactis (Panaredda and Mbriagotto sourdoughs), in each sourdough (Minervini et al., 2012a). Leuconostoc mesenteroides and Lactobacillus fermentum (Panaredda sourdough), and Pediococcus acidilactici (Torcolo di San Costanzo sour- dough) (Table 4). Previous studies on sourdoughs used for the manufac- 3.2. Dominant lactic acid bacteria and yeasts isolated from sourdoughs ture of Italian sweet baked goods also found L. sanfranciscensis as the dominant species (Galli et al., 1988; Ottogalli et al., 1996; Venturi et al., Gram-positive, catalase-negative, nonmotile, cocci and rods able 2012). Other species found, even in codominance with L. sanfranciscensis, to acidify SDB or MRS5 broth (three hundred and six isolates) were were L. brevis (Garofalo et al., 2008) or belonged to the genera subjected to RAPD–PCR analysis. The reproducibility of the RAPD Lactococcus and Leuconostoc (Palomba et al., 2011).

Table 4 Distribution of dominant lactic acid bacterium and yeast species in sourdoughs used for the manufacture of traditional/typical Italian sweet leavened baked goods.

Sourdoughsa L. L. L. L. L. L. Lc. Ln. Ln. P. W. C. S. brevis casei curvatus fermentum plantarum sanfranciscensis lactis mesenteroides citreum acidilactici cibaria humilis cerevisiae

Buondì 24/24 4/8 4/8 Panettone (Puglia) 18/18 8/8 Buccellato di Lucca 10/15 5/15 2/8 6/8 Cornetto di semola 1/28 1/28 26/28 8/10 2/10 Pizza di Pasqua 18/18 10/10 Panaredda 3/16 2/16 6/16 2/16 3/16 6/6 Torcolo di San Costanzo 4/30 3/30 22/30 1/30 5/5 Mbriagotto 1/15 13/15 1/15 7/7 Pandoro 3/12 9/12 9/9 Nadalin 3/15 12/15 6/6 Panettone (Lombardia) 9/9 5/6 1/6 Resta di Como 16/16 6/6 Panettone basso 15/15 5/5 Biscotti del Lagaccio 9/9 7/7 Pandolce Genovese 14/14 6/6 Colomba 20/20 5/5 Veneziana 19/19 5/7 2/7 Ciambella di mosto 13/13 5/5

The first number indicates the number of isolates allotted to a species, whereas the second number indicates the total amount of isolates. L., Lactobacillus; Lc., Lactococcus; Ln., Leuconostoc; P., Pediococcus; W., Weissella; C., Candida; S., Saccharomyces. a Sourdoughs are indicated with the names of the related sweet baked goods. 76 A. Lattanzi et al. / International Journal of Food Microbiology 163 (2013) 71–79

Preliminarily, one hundred eighty-eight isolates of yeasts were we are not able to state that the association between L. sanfranciscensis, assayed for the fermentation of galactose, sucrose, maltose, raffinose L. plantarum and S. cerevisiae found in the Pandoro sourdough occurs in and trehalose, as well as for the capacity of growing at 37 °C. At all the sourdoughs used for manufacturing Pandoro. Indeed, even when least five isolates for each sourdough, including all isolates showing the same sourdough is propagated in two different environments by different fermentation profiles (data not shown), were genetically using the same flour and applying identical technological process typed and identified. Cluster analysis of RAPD–PCR profiles revealed parameters, taxonomic differences may be observed, because of that the level of diversity among isolates ranged from ca. 2.5 to the influence of the different house microbiota (Vrancken et al., 2010; 42.5% (Fig. S3). At the diversity level of 15%, twenty-five strains Minervini et al., 2012b). were grouped into nine clusters (I–IX), whereas nineteen were unclustered. Overall, clustering of the strains was not related to sour- 3.3. Pyrosequencing analysis of the bacterial microbiota doughs they were isolated from. Compared to lactic acid bacteria (Fig. S2), the number of yeast strains found in a given sourdough Compared to yeasts, lactic acid bacteria showed higher diversity. was lower, ranging from one to four. The only exceptions were Therefore, pyrosequencing analysis was carried out to characterize represented by Cornetto di semola (8 strains), Buccellato di Lucca the domain Bacteria. Sourdoughs were subjected to extraction of (5), Panettone (Puglia) (6), and Pizza di Pasqua (7) (Fig. S3). The DNA for pyrosequencing. The number of sequence reads per sample low diversity of yeasts should be related to environmental contamina- ranged from 5478 to 5805. A total of 48 taxonomic units were tion by very competitive strains (Vrancken et al., 2010) and/or to the found. For graphical representation, only sequences with a minimal presence of stable associations between lactic acid bacteria and yeasts incidence (1%) were considered (Fig. 1). L. sanfranciscensis had an (Gobbetti, 1998). incidence above 99% in sourdoughs used for manufacturing Buondì, Yeast strains were identified by partial sequence analysis of the 26S Panettone (Puglia), Buccellato di Lucca, Pizza di Pasqua, Panettone rRNA gene (Table 4). Except for Mbriagotto, Ciambella di Mosto (Lombardia), Resta di Como, Panettone basso, Biscotti del Lagaccio, and Pandolce Genovese, Saccharomyces cerevisiae was identified in Pandolce Genovese, Colomba, Veneziana, and Ciambella di Mosto. all the sourdoughs. These above three sourdoughs harbored only one Mbriagotto sourdough harbored L. sanfranciscensis as the most abun- (Mbriagotto and Ciambella di Mosto) or two (Pandolce Genovese) dant bacterial species (ca. 79%), followed by Lc. lactis (ca. 12%), and strains belonging to the species Candida humilis. Buondì, Buccellato di L. plantarum (ca. 8%). L. sanfranciscensis was also the most abundant Lucca, Cornetto di semola, Panettone (Lombardia) and Veneziana sour- bacterial species in sourdoughs used for manufacturing Pandoro and doughs combined the presence of C. humilis together with S. cerevisiae Nadalin, followed by L. plantarum. Torcolo di San Costanzo sourdough (Table 4). Previously (Foschino et al., 2004; Garofalo et al., 2008; harbored the highest percentage of L. plantarum, followed by L. brevis, Palomba et al., 2011; Venturi et al., 2012; Vernocchi et al., 2004), L. casei, and L. sanfranciscensis. Sphingobium spp. had an incidence S. cerevisiae and/or C. humilis were found as the dominant yeast species of ca. 65% in Panaredda sourdough. Methylobacterium sp. (24%), in sourdoughs used for making Panettone, Cornetto and Colomba. L. sanfranciscensis (4.74%), and Buchnera aphidicola (1.10%) were In this study, four sourdoughs (Buondì, Buccellato di Lucca, Panettone also detected in this sourdough. Thus, the results of pyrosequencing Lombardia and Veneziana) were characterized by the coexistence of were in agreement with the results of the culture-dependent method, S. cerevisiae, C. humilis,andL. sanfranciscensis (Table 4). This multiple except for Cornetto di semola, Torcolo di San Costanzo, and Panaredda association occurred in Italian type I sourdoughs, and the stability of sourdoughs. Indeed, according to pyrosequencing, L. sanfranciscensis this association seemed to be promoted by quite similar times of gene- was the most abundant in Cornetto di semola sourdough, followed by ration and the steady presence of soluble carbohydrates (Venturi et al., Ln. citreum (Fig. 1), whereas the opposite was found through the 2012). It has to be highlighted that the microbial associations found in culture-dependent method (Table 4). Furthermore, L. sanfranciscensis the various sourdoughs considered in this study have validity as far as was detected, although with quite a low incidence, in Torcolo di one refers to the sourdough sampled in a given bakery. For instance, San Costanzo and Panaredda sourdoughs (Fig. 1), which it had not

Fig. 1. Relative abundance, based on bacterial 16S tag-encoded FLX Titanium amplicon pyrosequencing analysis, of bacterial taxonomical units (species or genus) of sourdoughs used for the manufacture of traditional Italian sweet leavened baked goods. A. Lattanzi et al. / International Journal of Food Microbiology 163 (2013) 71–79 77 been isolated from (Table 4). The sourdough used for making Panaredda order to explore possible correlations between process parameters, was the only one, which was characterized by high relative abundance microbial composition and biochemical characteristics of sourdoughs of two proteobacteria, Sphingobium sp. and Methylobacterium sp. These (Minervini et al., 2012a; Scheirlinck et al., 2007, 2008). Concentrations bacteria were obviously not detected by culture-dependent method, of soluble carbohydrates and FAA in the flour were not correlated because of the use of selective media for lactic acid bacteria. It is note- with any other variable (data not shown). The time of back slopping worthy that these two bacterial genera, both Gram-negative rods, was positively correlated (r = 0.30) with the cell density of lactic acid aerobic, chemorganotrophs, and ubiquitous (often found in soils), have bacteria, as well as with the concentration of lactic (r = 0.29) and acetic never been associated with food ecosystems. Similar observations were (r = 0.54) acids and total FAA (r = 0.36).Indetails,sourdoughsused reported in a recent paper for different bacterial genera or species for the manufacture of Cornetto di semola, Buccellato di Lucca, Pizza di detected in beef through pyrosequencing (Ercolini et al., 2011). Although Pasqua, Buondì and Panettone (Puglia) were grouped into quadrants I to understand the actual role of these two bacteria in the sourdough and II because of the highest time of back slopping and cell density of ecosystem could deserve further experimentation, it can be hypothesized lactic acid bacteria (Table 1 and Fig. 2). Cornetto di semola sourdough that they could have contaminated kernels allowing the DNA to be was located separately from the others due to the low cell density of detected, but their viability in this food ecosystem remains questionable yeasts. Although the time of back slopping was 10 h, Pandolce Genovese (Humblot and Guyot, 2009). sourdough was located in the III quadrant. Compared to the previous sourdoughs, this also showed the lowest temperature of back slopping 3.4. Correlations between process parameters, microbial community (12 °C) (Table 1 and Fig. 2). Except for sourdough used for the manu- profile and biochemical characteristics of sourdoughs facture of Cornetto di semola, the percentage of sourdough used for back slopping was positively correlated (r = 0.48) with the cell density PCA was carried out using the following variables: concentration of of yeasts, as well as with the concentration of ethanol (r = 0.49) maltose, glucose, fructose, and FAA in the flours, process parameters (e.g., Buccellato di Lucca and Panaredda sourdoughs) (I and IV quadrants used during back slopping (percentage of sourdough used as inoculum, of Fig. 2). and time, temperature and number of back slopping), profiles of the sourdough microbiota (cell density of lactic acid bacteria and yeasts, 4. Conclusions number of strains, and percentage of obligately homofermentative, and obligately and facultatively heterofermentative lactic acid bacteria) First, this study gave a comparative view of the lactic acid bacteria and biochemical characteristics of the sourdoughs (pH, concentration of and yeast microbiota of eighteen sourdoughs used for the manufacture lactic and acetic acids, ethanol, maltose, glucose, fructose and total FAA, of traditional Italian sweet baked goods. It was highlighted that bio- and QF) (Fig. 2). A similar analysis had been previously performed in chemical characteristics (e.g., pH, organic acids, FAA and carbohydrates)

4 I IV SG_1 SG_3 SG_2 3 SA_1 SA_3 Ethanol Yeasts SM_3 SA_2 % sourdough SM_2 SC_2SC_1 SM_1 SC_3 2 Glucose-S BS Number Hof Obl Fructose-S LAB FAA-S Hef Fa SI_2SI_1 1 SE_1 LAB strainsSL_2 SI_3 SE_2 Lactic acid SL_1 SH_2 SE_3 SH_1 SL_3 Acetic acid SH_3 Maltose-S QF SB_3SB_1 0 YeastSB_2 strains SR_1SR_3 SS_2 BS Time SS_3 SR_2 SS_1 BS TempST_1ST_2 SN_2 SU_1SU_3 ST_3SQ_1 SN_1SN_3 SQ_3SQ_2

Factor 2: 25,59% -1 SU_2 Hef Obl SD_1 SD_3 -2 SD_2

SP_1 SP_3SP_2 -3

SO_1 SO_2 SO_3 II III -4 -5 -4 -3 -2 -1 0 1 2 3 4 Factor 1: 42,35%

Fig. 2. Score and loading plots of first and second principal components after principal component analysis based on data of process parameters used during back slopping (percentage of sourdough used as inoculum, and time, temperature and number of back slopping), profiles of microbial community (cell densities of lactic acid bacteria and yeasts, number of strains, and percentage of obligately homofermentative and obligately and facultatively heterofermentative lactic acid bacteria) and biochemical characteristics (concen- trations of lactic and acetic acids, ethanol, total free amino acids, maltose, glucose and fructose, and quotient of fermentation) from the sourdoughs used for the manufacture of traditional Italian sweet leavened baked goods. Sourdoughs are indicated with the following alphanumeric codes: Buondì, SA_1, SA_2, SA_3; Panettone (Puglia), SB_1, SB_2, SB_3; Buccellato di Lucca, SC_1, SC_2, SC_3; Cornetto di semola, SD_1, SD_2, SD_3; Pizza di Pasqua, SE_1, SE_2, SE_3; Panaredda, SG_1, SG_2, SG_3; Torcolo di San Costanzo, SH_1, SH_2, SH_3; Mbriagotto, SI_1, SI_2, SI_3; Pandoro, SL_1, SL_2, SL_3; Nadalin, SM_1, SM_2, SM_3; Panettone (Lombardia), SN_1, SN_2, SN_3; Resta di Como, SO_1, SO_2, SO_3; Panettone basso, SP_1, SP_2, SP_3; Biscotti del Lagaccio, SQ_1, SQ_2, SQ_3; Pandolce Genovese, SR_1, SR_2, SR_3; Colomba, SS_1, SS_2, SS_3; Veneziana, ST_1, ST_2, ST_3; Ciambella di Mosto, SU_1, SU_2, SU_3. % sourdough, percentage of sourdough used as inoculum; BS Time, time of back slopping; BS Temp, temperature of back slopping; BS Number, number of back slopping; LAB, cell density of lactic acid bacteria; Yeasts, cell density of yeasts; LAB strains, number of lactic acid bacteria strains; Yeast strains, number of yeast strains; Hof Obl, obligately homofermentative lactic acid bacteria; Hef Fa, facultatively heterofermentative lactic acid bacteria; Hef Obl, obligately heterofermentative lactic acid bac- teria FAA-S, total free amino acids; QF, quotient of fermentation. 78 A. Lattanzi et al. / International Journal of Food Microbiology 163 (2013) 71–79 of these sourdoughs are different from those of sourdoughs used for polymorphisms in this species. Applied and Environmental Microbiology 58, 3429–3432. the manufacture of typical/traditional Italian breads (Minervini De Vuyst, L., Neysens, P., 2005. Biodiversity of sourdough lactic acid bacteria. Trends et al., 2012a). Further, compared to bread sourdoughs (Minervini in Food Science and Technology 16, 43–56. et al., 2012a), a higher dominance of L. sanfranciscensis was found De Vuyst, L., Vrancken, G., Ravyts, F., Rimaux, T., Weckx, S., 2009. Biodiversity, ecological determinants, and metabolic exploitation of sourdough microbiota. Food Microbiology for sweet baked sourdough goods (ca. 28% and 72% of the total lactic 26, 666–675. acid bacteria isolates, respectively). Secondly, culture-dependent Dowd, S.E., Zaragoza, J., Rodriguez, J.R., Oliver, M.J., Payton, P.R., 2005. Windows. method and pyrosequencing analysis were complementarily used NET network distributed basic local alignment search tool-kit (W.ND-BLAST). to describe the diversity of the sourdough microbiota. Finally, BMC Bioinformatics 6, 93. Dowd, S.E., Sun, Y., Secor, P.R., Rhoads, D.D., Wolcott, B.M., James, G.A., Wolcott, R.D., the present work constitutes a base for exploiting further, upon 2008. Survey of bacterial diversity in chronic wounds using pyrosequencing, adequate characterization, lactic acid bacteria and yeasts strains DGGE, and full ribosome shotgun sequencing. BMC Bioinformatics 8, 43. as tools to guarantee a highly reproducible quality of the Italian Dray, S., Dufour, A.B., 2007. The ade4 package: implementing the duality diagram for ecologists. Journal of Statistical Software 22, 1–20. sweet goods studied, while preserving their traditional trait. Ercolini, D., Ferrocino, I., Nasi, A., Ndagijimana, M., Vernocchi, P., La Storia, A., Laghi, L., Mauriello, G., Guerzoni, M.E., Villani, F., 2011. Monitoring of microbial metabolites and bacterial diversity in beef stored under different packaging conditions. Applied Acknowledgments and Environmental Microbiology 77, 7372–7381. Foschino, R., Gallina, S., Andrighetto, C., Rossetti, L., Galli, A., 2004. Comparison of cultural methods for the identification and molecular investigation of yeasts This work was supported by Beldem S.A. (Andenne, Belgium). We from sourdough for Italian sweet baked products. FEMS Yeast Research 4, 609–618. thank the following for their kind collaboration: Antica Forneria Galli, A., Franzetti, L., Fortina, M.G., 1988. Isolation and identification of sour dough mi- dei Sassi di Cosimo Martino (Matera), Atelier Biasetto (Selvazzano croflora. Microbiologie - Aliments - Nutrition 6, 345–351. Dentro, Padova), Vincenzo Benvenuto and Cipriano Berloco (Caffetteria Garofalo, C., Silvestri, G., Aquilanti, L., Clementi, F., 2008. PCR-DGGE analysis of lactic acid bacteria and yeast dynamics during the production processes of three varieties del Viale, Altamura, Bari), Buccellato Taddeucci (Lucca), Fabrizio Del of Panettone. Journal of Applied Microbiology 105, 243–254. Papa (180° gradi, Fermo), Angela Gargani (Preti Dolciaria, Sant'Olcese, Gobbetti, M., 1998. Interactions between lactic acid bacteria and yeasts in sourdoughs. – Genova), Panetteria Pasticceria Caffetteria Molinelli (Rovescala, Pavia), Trends in Food Science & Technology 9, 267 274. Gobbetti, M., Corsetti, A., 1997. Lactobacillus sanfrancisco a key sourdough lactic Vincenzo Paolillo (Il Fornaio dei Vecchi Mulini, Barletta), Pasticceria acid bacterium: a review. Food Microbiology 14, 175–187. Paolo Caridi (Reggio Calabria), Pasticceria Fuin (Como), Pasticceria Gontcharova, V.Y., Wolcott, E.R.D., Hollister, E.B., Gentry, T.J., Dowd, S.E., 2010. Black Gertosio (Torino), Pasticceria Perbellini Ernesto (Bovolone, Verona), box chimera check (B2C2): a Windows-based software for batch depletion of fi chimeras from bacterial 16S rRNA gene datasets. Open Microbiology Journal 4, 6. Pasticceria Sandri (Perugia), Pasticceria Taglia co (Genova), Giovanni Hammes, W.P., Gaenzle, M.G., 1998. Sourdough breads and related products, In: Wood, B.J.B. Pellegrino (Dolci Sapori, Valenzano, Bari), Carlo Pozza (Pasticceria (Ed.), Microbiology of Fermented Foods, 2nd ed. Blackie Academic and Professional, Davenicio, Arzignano, Vicenza), and Mauro Scaglia (Pasticceria Mascale, an imprint of Thomson Science, London, pp. 199–216. Humblot, C., Guyot, J.P., 2009. Pyrosequencing of tagged 16S rRNA gene amplicons Roncadelle, Brescia). for rapid deciphering of the microbiomes of fermented foods such as pearl millet slurries. Applied and Environmental Microbiology 75, 4354–4361. ISTAT, 2011. (www3.istat.it/…/catalogo/20110728_00/Tavole_fuori_testo/primo% Appendix A. Supplementary data 20volume/3.1.5.1/v1_p3_c1_5_UE_19.xls). Kaufman, L., Rousseeuw, P.J., 1990. Finding Groups in Data. Wiley, New York. Supplementary data to this article can be found online at http:// Meroth, C.B., Walter, J., Hertel, C., Brandt, M.J., Hammes, W.P., 2003. Monitoring the bacterial population dynamics in sourdough fermentation processes by using PCR- dx.doi.org/10.1016/j.ijfoodmicro.2013.02.010. denaturing gradient gel electrophoresis. Applied and Environmental Microbiology 69, 475–482. Minervini, F., De Angelis, M., Di Cagno, R., Pinto, D., Siragusa, S., Rizzello, C.G., Gobbetti, M., References 2010. Robustness of Lactobacillus plantarum starters during daily propagation of wheat flour sourdough type I. Food Microbiology 27, 897–908. Andrade,M.J.,Rodrìguez,M.,Sànchez,B.,Aranda,E.,Cordoba,J.J.,2006.DNAtyping Minervini, F., Di Cagno, R., Lattanzi, A., De Angelis, M., Antonielli, L., Cardinali, G., methods for differentiation of yeasts related to dry-cured meat products. International Cappelle, S., Gobbetti, M., 2012a. Lactic acid bacterium and yeast microbiotas of Journal of Food Microbiology 107, 48–58. 19 sourdoughs used for traditional/typical Italian breads: interactions between Andreotti, R., Perez de Leon, A.A., Dowd, S.E., Guerrero, F.D., Bendele, K.G., Scoles, G.A., ingredients and microbial species diversity. Applied and Environmental Microbiology 2011. Assessment of bacterial diversity in the cattle tick Rhipicephalus (Boophilus) 78, 1251–1264. microplus through tag-encoded pyrosequencing. BMC Microbiology 11, 6. Minervini, F., Lattanzi, A., De Angelis, M., Di Cagno, R., Gobbetti, M., 2012b. Influence of Barnett, J.A., Payne, R.W., Yarrow, D., 2000. Yeast: Characteristics and Identification. artisan bakery- or laboratory-propagated sourdoughs on the diversity of lactic acid Cambridge University Press, Cambridge. bacterium and yeast microbiotas. Applied and Environmental Microbiology 78, Callaway, T.R., Dowd, S.E., Edrington, T.S.E., Anderson, R.C., Krueger, N., Bauer, N., 5328–5340. Kononoff, P.J., Nisbet, D.J., 2010. Evaluation of bacterial diversity in the rumen Naser, S.M., Thompson, F.L., Hoste, B., Gevers, D., Dawyndt, P., Vancanneyt, M., Swings, and feces of cattle fed different levels of dried distillers grains plus solubles using J., 2005. Application of multilocus sequence analysis (MLSA) for rapid identifica- bacterial tag-encoded FLX amplicon pyrosequencing. Journal of Animal Science tion of Enterococcus species based on rpoA and pheS genes. Microbiology 151, 88, 3977–3983. 2141–2150. Cardinali, G., Bolano, A., Martini, A., 2001. A DNA extraction and purification method O'Donnell, K., 1993. Fusarium and its near relatives. In: Reynolds, D.R., Taylor, J.W. for several yeast genera. Annals of Microbiology 51, 121–130. (Eds.), The Fungal Holomorph: Mitotic, Meiotic and Pleomorphic Speciation in Catzeddu, P., Mura, E., Parente, E., Sanna, M., Farris, G.A., 2006. Molecular characterization Fungal Systematics. CAB International, Wallingford, pp. 225–233. of lactic acid bacteria from sourdough breads produced in Sardinia (Italy) and Ottogalli, G., Galli, A., Foschino, R., 1996. Italian bakery products obtained with sour multivariate statistical analyses of results. Systematic and Applied Microbiology 29, dough: Characterization of the typical microfl ora. Advances in Food Sciences 18, 138–144. 131–144. Corsetti, A., Lavermicocca, P., Morea, M., Baruzzi, F., Tosti, N., Gobbetti, M., 2001. Phenotypic Ouadghiri, M., Vancanneyt, M., Vandamme, P., Naser, S., Gevers, D., Lefebvre, K., Swings, J., and molecular identification and clustering of lactic acid bacteria and yeasts from Amar, M., 2009. Identification of lactic acid bacteria in Moroccan raw milk and wheat (species Triticum durum and Triticum aestivum) sourdoughs of Southern Italy. traditionally fermented skimmed milk ‘lben’. Journal of Applied Microbiology 106, International Journal of Food Microbiology 64, 95–104. 486–495. Corsetti, A., De Angelis, M., Dellaglio, F., Paparella, A., Fox, P.F., Settanni, L., Gobbetti, M., Palomba, S., Blaiotta, G., Ventorino, V., Saccone, A., Pepe, O., 2011. Microbial character- 2003. Characterization of sourdough lactic acid bacteria based on genotypic and ization of sourdough for sweet baked products in the Campania region (southern cell-wall protein analysis. Journal of Applied Microbiology 94, 641–654. Italy) by a polyphasic approach. Annals of Microbiology 61, 307–314. De Angelis, M., Siragusa, S., Berloco, M., Caputo, L., Settanni, L., Alfonsi, G., Amerio, M., Reale, A., Di Renzo, T., Succi, M., Tremonte, P., Coppola, R., Sorrentino, E., 2011. Grandi, A., Ragni, A., Gobbetti, M., 2006. Selection of potential probiotic lactobacilli Identification of lactobacilli isolated in traditional ripe wheat sourdoughs by from pig feces to be used as additives in pelleted feeding. Research in Microbiology using molecular methods. World Journal of Microbiology and Biotechnology 27, 157, 792–801. 237–244. De Angelis, M., Di Cagno, R., Gallo, G., Curci, M., Siragusa, S., Crecchio, C., Parente, E., Scheirlinck, I., Van der Meulen, R., Van Schoor, A., Vancanneyt, M., De Vuyst, L., Gobbetti, M., 2007. Molecular and functional characterization of Lactobacillus Vandamme, P., Huys, G., 2007. Influence of geographical origin and flour type sanfranciscensis strains isolated from sourdoughs. International Journal of Food on diversity of lactic acid bacteria in traditional Belgian sourdoughs. Applied and Microbiology 114, 69–82. Environmental Microbiology 73, 6262–6269. De Los Reyes-Gavilán, C.G., Limsowtin, G.K.Y., Tailliez, P., Séchaud, L., Accolas, J.P., 1992. Scheirlinck, I., Van der Meulen, R., Van Schoor, A., Vancanneyt, M., De Vuyst, L., A Lactobacillus helveticus-specific DNA probe detects restriction fragment length Vandamme, P., Huys, G., 2008. Taxonomic structure and stability of the bacterial A. Lattanzi et al. / International Journal of Food Microbiology 163 (2013) 71–79 79

community in Belgian sourdough ecosystems assessed by culturing and population Ward, L.J.H., Timmins, M.J., 1999. Differentiation of Lactobacillus casei, Lactobacillus fingerprinting. Applied and Environmental Microbiology 74, 2414–2423. paracasei and Lactobacillus rhamnosus by polymerase chain reaction. Letters in Torriani, S., Felis, G.E., Dellaglio, F., 2001. Differentiation of Lactobacillus plantarum, Applied Microbiology 29, 90–92. L. pentosus, and L. paraplantarum by recA gene sequence analysis and multiplex Weckx, S., Van der Meulen, R., Maes, D., Scheirlinck, I., Huys, G., Vandamme, P., PCR assay with recA gene-derived primers. Applied and Environmental Microbiology De Vuyst, L., 2010. Lactic acid bacteria community dynamics and metabolite 67, 3450–3454. production of rye sourdough fermentations share characteristics of wheat and spelt Valmorri, S., Tofalo, R., Settanni, L., Corsetti, A., Suzzi, G., 2010. Yeast microbiota associated sourdough fermentations. Food Microbiology 27, 1000–1008. with spontaneous sourdough fermentations in the production of traditional wheat White, T.J., Bruns, T., Lee, S., Taylor, J.W., 1990. Amplification and direct sequencing sourdough breads of the region (Italy). Antonie Van Leeuwenhoek 97, of fungal ribosomal RNA genes for phylogenetics. In: Innis, M.A., Gelfand, D.H., 119–129. Sninsky, J.J., White, T.J. (Eds.), PCR Protocols: A Guide to Methods and Applications. Venturi, M., Guerrini, S., Vincenzini, M., 2012. Stable and non-competitive association Academic Press, New York, pp. 315–322. of Saccharomyces cerevisiae, Candida milleri and Lactobacillus sanfranciscensis Zannini, E., Garofalo, C., Clementi, F., 2010. I prodotti dolciari lievitati italiani. In: during manufacture of two traditional sourdough baked goods. Food Microbiology Gobbetti, M., Corsetti, A. (Eds.), Biotecnologia dei prodotti lievitati da forno. 31, 107–115. Casa Editrice Ambrosiana, Milano, pp. 243–262. Vermeulen, N., Gaenzle, M.G., Vogel, R.F., 2007. Glutamine deamidation by cereal- Zapparoli, G., Torriani, S., Dellaglio, F., 1998. Differentiation of Lactobacillus sanfranciscensis associated lactic acid bacteria. Journal of Applied Microbiology 103, 1197–1205. strainsbyrandomlyamplified polymorphic DNA and pulsed-field gel electrophoresis. Vernocchi, P., Valmorri, S., Gatto, V., Torriani, S., Gianotti, A., Suzzi, G., Guerzoni, M.E., FEMS Microbiology Letters 166, 324–332. Gardini, F., 2004. A survey on yeast microbiota associated with an Italian traditional Zeppa, G., Conterna, L., Gerbi, V., 2001. Determination of organic acids, sugars, diacetyl, sweet-leavened baked good fermentation. Food Research International 37, 469–476. and acetoin in cheese by high-performance liquid chromatography. Journal of Vogel, R.F., Pavlovic, M., Ehrmann, M.A., Wiezer, A., Liesegang, H., Offschanka, S., Voget, S., Agricultural and Food Chemistry 49, 2722–2726. Angelov, A., Böcker, G., Liebl, W., 2011. Genomic analysis reveals Lactobacillus Zotta, T., Piraino, P., Parente, E., Salzano, G., Ricciardi, A., 2008. Characterization of lactic sanfranciscensis as stable element in traditional sourdoughs. Microbial Cell acid bacteria isolated from sourdoughs for cornetto, a traditional bread produced Factories 10, S6. in Basilicata (southern Italy). World Journal of Microbiology and Biotechnology Vrancken, G., De Vuyst, L., Van der Meulen, R., Huys, G., Vandamme, P., Daniel, H.M., 24, 1785–1795. 2010. Yeast species composition differs between artisan bakery and spontaneous laboratory sourdoughs. FEMS Yeast Research 10, 471–481.