OUP user menu

Comparison of cultural methods for the identification and molecular investigation of yeasts from sourdoughs for Italian sweet baked products

Roberto Foschino, Silvia Gallina, Christian Andrighetto, Lia Rossetti, Antonietta Galli
DOI: http://dx.doi.org/10.1016/j.femsyr.2003.12.006 609-618 First published online: 1 March 2004


Twenty-five yeast strains isolated from sourdough samples for Panettone, Pandoro and Cornetto brioche manufactured by eight different bakeries in northern Italy were characterised. Classification was performed by the simplified identification method (SIM), Kurtzman and Fell's identification protocol, the API system from bioMérieux (France) and the MicroLogTM system from Biolog (USA). Genetic diversity was investigated by randomly amplified polymorphic DNA fingerprinting and mitochondrial-DNA restriction enzyme analysis. Sequences of the internal transcribed spacers between 18S and 26S rDNA genes were analysed. Candida humilis was the predominant species (56% of isolates), whereas the remaining strains (44%) were related to the Saccharomyces cerevisiae sensu stricto group. Identification systems based on phenotypic analysis proved to be unreliable to identify yeasts from sourdough. Either RAPD-PCR or mtDNA restriction analysis showed to be suitable for the identification of species, but could not be used to differentiate among the isolates at the strain level. Sequencing of the ITS region permitted a consistent classification of the sourdough yeasts.

  • Yeast
  • Sourdough
  • Candida humilis
  • Saccharomyces cerevisiae
  • Panettone
  • Pandoro

1 Introduction

Sourdough is a spontaneously fermented mixture of flour and water or, more often, it is inoculated with a wild microbial starter called “mother”, which is a sourdough constantly renewed in a cycling way, using strict conditions of recipe and ripening [13]. The sourdough microflora consists of lactic acid bacteria (108–109 CFU g−1), responsible for acidification and weak proteolysis, and yeasts (106–107 CFU g−1) that mainly carry out the leavening of the dough. Metabolic activities of these microorganisms also affect the sensorial characteristics and prolong the shelf life of the products [47]. The use of natural starters containing mixed cultures, without the addition of baker's yeast at the beginning of the process, is still applied in bakeries that make some Italian sweet goods such as Panettone, Colomba, Pandoro, Cornetto brioche and other local cakes and buns [3,8,9].

The primary source of the sourdough mycoflora has not yet been elucidated clearly. Yeasts are rare in wheat flour since they contaminate the outside parts of caryopsis, which are removed during cleaning, sieving and milling operations [10,11]. The predominant microorganisms in sourdoughs probably originate from the traditional practice to add vegetable matters, which are naturally rich in yeasts and lactic acid bacteria, such as grape must, raisins, apples, figs, lemon or orange peels, bran, hay or horse dung, to the starting dough to prepare the mother culture. Then, stringent conditions of ripening of the following doughs (long times and low temperature of incubation, anaerobic state by binding, and final low pH value) enable the establishment of a typical microflora. The phenomenon is not easy to reproduce. Some “mothers” for Panettone and Pandoro are more than sixty years old and are preserved according to one's private custom carried on from generation to generation. This information was collected by interviewing the manufacturers during the sampling and from a previous work [8].

Most reviews about sourdough microflora have dealt with the characterisation of lactobacilli and their role during fermentation [1216], while a smaller number of papers have been devoted to investigate the mycological component [1720], presumably because baker's yeast is widely used in sourdough bread technology. Regarding traditional sweet baked products, Galli and Ottogalli [21] first isolated Torulopsis holmii from sourdoughs for Panettone. Later, the same group of researchers [22] found yeasts ascribed to Saccharomyces exiguus, S. cerevisiae and Candida stellata in sourdough samples used for Panettone and Cornetto brioche. Recently, Foschino et al. [8] characterised the microbial composition of sourdoughs for sweet baked products from ten different bakeries in Lombardy and isolated strains belonging to Candida holmii, the asexual form of S. exiguus, and S. cerevisiae. However, in all these papers the classification of yeasts was obtained only by phenotypic analysis. Moreover, the genetic diversity within species and strains of sourdough yeasts has not yet been well explored. The aim of this investigation was to characterise yeasts isolated from sourdoughs for sweet baked products in different industrial bakeries by phenotypic analysis and molecular methods. Particularly, RAPD-PCR technique and mitochondrial-DNA restriction analysis, that have been usefully applied to strain typing and identification of yeasts isolated from wines [2325] and cheeses [2628], were used. The sequence of the internal transcribed spacers (ITS) occurring between the 18S and 28S rDNA genes was investigated in order to confirm the identification at species level [29,30].

2 Materials and methods

2.1 Microbial counts and pH determination of sourdough samples

Samples of a ripened mother culture used to prepare sweet baked products were obtained from eight industrial bakeries located in northern Italy (Table 1). They were aseptically collected and quickly refrigerated until analysis. Approximately 10 g of sample was diluted in 90 ml of sterile peptone water and homogenised in a Stomacher 400 Circulator for 5 min at 260 rpm (Seward, Thetford, Norfolk, United Kingdom). Lactobacilli were counted by plating on Sanfrancisco Medium [15], modified in this work, and on de Man Rogosa Sharpe (MRS) agar [31] with cycloheximide (100 mg l−1). Incubation was done under anaerobic conditions (GasPak System, BBL, Syracuse, NY, USA) at 30 °C for 3–5 days. In the recipe of Sanfrancisco Medium, rye flour and fresh baker's yeast were replaced by fresh yeast extract (150 ml l−1) prepared according to Kline and Sugihara [32]. Yeasts counts were determined by plating on yeast glucose chloramphenicol agar (YGC) [33] and malt extract agar (MEA) [34] after incubation at 25 °C for 5 days. The pH values were checked using undiluted samples by a combined glass–calomel electrode connected to a pH meter (Gibertini DP 100-NE, Milano, Italy).

View this table:
Table 1

Microbial counts, pH measurements and some information about sourdoughs for sweet baked products collected in northern Italy

FirmCityYeast counts (CFU g−1)Lactobacilli counts (CFU g−1)pHBaked product
AMilano3.9 × 1074.8 × 1083.80Cornetto
BMilano5.4 × 1062.9 × 1084.05Panettone
1.4 × 1074.4 × 1083.95
1.4 × 1061.7 × 1093.75
1.0 × 1079.5 × 1074.20
2.0 × 1062.0 × 1084.10
1.7 × 1072.1 × 1094.05
1.3 × 1074.2 × 1084.15
CMilano1.1 × 1071.9 × 1083.95Panettone
DMilano6.2 × 1073.4 × 1083.90Cornetto
1.2 × 1073.3 × 1083.95
EMilano6.6 × 1062.2 × 1083.85Panettone
FRavenna1.3 × 1074.1 × 1083.95Cornetto
2.7 × 1061.6 × 1084.05
9.6 × 1061.9 × 1084.00
1.9 × 1071.5 × 1084.00
GVarese7.6 × 1068.1 × 1074.10Cornetto
HVerona2.1 × 1073.4 × 1083.90Pandoro
1.7 × 1075.5 × 1083.85
1.2 × 1071.9 × 1083.95
1.1 × 1072.8 × 1083.95
9.8 × 1064.5 × 1083.85
1.2 × 1075.5 × 1083.90
1.3 × 1071.1 × 1093.90
1.0 × 1071.8 × 1084.00

2.2 Isolation and classification of yeast strains based on phenotypic analysis

Colonies were randomly selected from high dilutions on YGC plates, which proved more useful than MEA to recover yeasts and discriminate the morphology of the colonies. Purity was checked by streaking on yeast extract peptone dextrose agar (YEPD) [23] and pure cultures were kept on slants of the same medium at 4 °C. The morphological aspect of the cells was observed microscopically. Before executing phenotypic tests the isolates were subcultured twice on YEPD plates. Physiological characteristics of each isolate were investigated according to the simplified identification method (SIM) [34] and Kurtzman and Fell's protocol [35]. The assimilation capabilities on different carbon sources were also examined by API ID 32C (BioMérieux, Marcy l'Étoile, France) and by Microplate YT (Biolog Inc., Hayward, CA, USA) according to the suppliers’ recommendations. In the latter case a MicroLogTM MicroStationTM with Biolog Microlog3 4.20 software version was used.

2.3 DNA isolation

Yeast cells were grown overnight in 5 ml of YEPD broth under agitation at 200 rpm and collected by centrifugation at 6000g for 10 min at 4 °C. After washing the cells were resuspended in TE buffer (10 mM Tris–HCl, pH 8.0; 1 mM EDTA). DNA was extracted according to the protocol of Querol et al. [23], modified by using 25 mg ml−1 of lytic enzyme from Rhizoctonia solani (in 1 M sorbitol, 0.1 M EDTA, pH 7.5), with incubation at 45 °C for 2 h, in order to digest the cell wall.

The following reference strains were used to compare the results of genotypic analyses: C. holmii CBS 135, C. humilis CBS 6897, C. stellata CBS 157, Pichia anomala CBS 5759, S. barnettii CBS 5648, S. bayanus CBS 380, S. cerevisiae ATCC 9763, S. exiguus CBS 379, S. pastorianus CBS 1538, S. unisporus CBS 398.

2.4 Randomly amplified polymorphic DNA fingerprinting

PCR amplifications were performed in a 25-μl reaction volume containing 10 mM Tris–HCl (pH 9), 50 mM KCl, 1.5 mM MgCl2, 200 μM each of dATP, dCTP, dGTP and dTTP, 80 ng of DNA and 2U of Taq-DNA Polymerase (Applied Biosystems, Foster City, CA, USA) using a T Gradient Biometra Thermocycler (Biometra, Göttingen, Germany). For primer M13 (5 GAGGGTGGCGGTTCT 3) [36], the temperature profile was: denaturation step at 94 °C for 1 min, annealing step at 40 °C for 20 s, ramp to 72 °C at 0.6 °C s−1 and extension step at 72 °C for 2 min and this was repeated for 35 cycles. For primer RF2 (5 CGGCCCCTGT 3) [37], the temperature profile was: denaturation step at 95 °C for 30 s, annealing step at 36 °C for 1 min and extension step at 72 °C for 1 min and 30 s and was repeated for 35 cycles.

Amplification products were resolved by electrophoresis in 1.8% (w/v) agarose gels in TAE buffer (40 mM Tris–acetate, pH 8.2; 1 mM EDTA) at 100 V for 3 h, stained with ethidium bromide and photographed under UV illumination.

2.5 Mitochondrial DNA restriction analysis

Extracted DNA was digested with restriction enzymes HinfI and RsaI (Amersham Pharmacia Biotech, Uppsala, Sweden) at 37 °C for 4 h. The mtDNA restriction fragments were separated by electrophoresis on 0.7% (w/v) agarose gels in TAE buffer (40 mM Tris–acetate, pH 8.2; 1 mM EDTA) at 110 V for 2 h, stained with ethidium bromide and photographed under UV illumination.

2.6 Determination of the ITS rDNA gene sequences

The internal transcribed spacers present between the 18S and 26S rDNA genes (ITS1-5.8S-ITS2) were amplified directly from genomic DNA as described by Montrocher et al. [29]. The amplification was performed in a 100-μl reaction mixture containing 10 mM Tris–HCl (pH 9), 50 mM KCl, 1.5 mM MgCl2, 200 μM each of dATP, dCTP, dGTP and dTTP, 80 ng of DNA, 0.1 μM of primer 2234C (5 GTTTCCGTAGGTGAACCTGC 3), 0.1 μM of primer 3126T (5 ATATGCTTAAGTTCAGCGGGT 3) and 1 U of Taq-DNA Polymerase (Applied Biosystems, USA). A T Gradient Biometra Thermocycler (Biometra, Germany) was used. The temperature profile was: denaturation step at 95 °C for 1 min, annealing step at 50 °C for 1 min and extension step at 72 °C for 1 min and this was repeated for 35 cycles. The amplified products were purified by QIAquick spin columns (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions and sequenced with the ABI Prism® BigDyeTM Terminators v. 3.0 Cycle Sequencing Kit (Applied Biosystems, USA) in an ABI Prism® 310 Genetic Analyser. The BLAST program was used to search the GenBank database for homologous sequences (http://www.ncbi.nlm.nih.gov/).

The sequences of the ITS1-5.8S-ITS2 region of C. humilis CBS 6897 and C. stellata CBS 157 type strains were determined and assigned to DDBJ/EMBL/GenBank Accession Nos. AY188851 and AY188852, respectively.

2.7 Cluster analysis

Polymerase chain reaction fingerprinting and mtDNA restriction profiles were analysed with the pattern analysis software package Gel Compare version 4.0 (Applied Maths, Kortrijk, Belgium). Calculation of similarities of band profiles was based on the Pearson product–moment correlation for the PCR assay and on the Dice coefficient for the mtDNA analysis. Dendrograms were obtained by means of the unweighed pair group method using an arithmetic average (UPGMA) clustering algorithm [38].

A phylogenetic tree based on the ITS rDNA genes sequences was obtained by the neighbour-joining method [39] using a MicroSeq Analysis Software v. 1.39 (Applied Biosystems, USA). Relevant sequences of some reference strains were downloaded from the National Center for Biotechnology Information database.

3 Results

3.1 Microbial counts and pH determination of sourdough samples

Yeast counts in sourdough samples ranged from 1.4 × 106 to 6.2 × 107 CFU g−1 with a mean value of 1.1 × 107 CFU g−1, whereas lactobacilli counts in SF-modified medium varied from 8.1 × 107 to 1.7 × 109 with a mean value of 3.4 × 108 CFU g−1 (Table 1). Values of lactic acid bacteria detected in MRS agar ranged from 1.0 × 103 to 3.7 × 104 CFU g−1 with a mean value of 4.1 × 103 CFU g−1. The failure of predominant microorganisms to grow on MRS medium let us suppose that most lactobacilli belonged to Lactobacillus sanfranciscensis, which is unable to form colonies on MRS medium. This supposition was confirmed by the identification of some isolates randomly picked from SF-modified medium plates at high dilutions, using a species-specific PCR with 16S rRNA-targeted primers [40] (data not shown). In all sourdough samples a high degree of acidification was observed with pH values varying between 3.75 and 4.20, with a mean value of 3.95.

3.2 Phenotypic characteristics of yeast strains

After isolation each yeast strain was microscopically observed and its morphology recorded. Results obtained by different methods of identification based on phenotypic analysis are presented in Table 2. The identification of the species by four procedures was consistent only for one isolate (LB). Identifications with the SIM method and the API 32 ID system were frequently concordant. This was the case for the isolates 110, 112 and 221, which were ascribed to S. cerevisiae, and isolates 2PL3, GMAB, MM2, MM5, RM2, RM3, RM4, RM5, 18R, 20R and 21R, which were attributed to S. exiguus or its imperfect state C. holmii. The isolates VA, VB, LS and 1R were identified as S. bayanus either by the SIM method or Kurtzman and Fell's protocol. Microbial characterisation by the Biolog system attributed the isolates RM4, 110, 112 and 3B to S. boulardii. This latter species is a synonym of S. cerevisiae. Besides, the Biolog system failed to identify some maltose-negative strains from bakery B (strains MM2, MM5, RM2, RM3, and RM5), which were identified as Kluyveromyces lodderae.

View this table:
Table 2

Identification of yeast strains performed with different systems based on phenotypic analysis and molecular techniques

FirmStrainSIM methodKurtzman and Fell's methodAPI ID32C (BioMérieux)Microplate YT (Biolog)Molecular techniques
A2PL3S. exiguusS. bayanusC. holmiiaC. humilis
BGMABS. exiguusS. bayanusC. holmiiS. cerevisiaeC. humilis
MM2S. exiguusC. stellataC. holmiiKl. lodderaea
MM5S. exiguusC. stellataC. holmiiKl. lodderaea
RM2S. exiguusNot identifiedC. holmiiKl. lodderaeC. humilis
RM3S. exiguusC. stellataC. holmiiKl. lodderaea
RM4S. exiguusS. bayanusC. holmiiS. boulardiia
RM5S. exiguusC. stellataC. holmiiKl. lodderaea
CLBS. cerevisiaeS.cerevisiaeS. cerevisiaeS. cerevisiaea
DVAS. bayanusS. bayanusS. cerevisiaeS. cerevisiaeS. pastorianus
VBS. bayanusS. bayanusC. holmiiS. cerevisiaea
ELGATaaC. holmiiaC. humilis
F110S. cerevisiaeS. bayanusS. cerevisiaeS. boulardiiS. cerevisiae
111S. pastorianusS. cerevisiaeC. holmiiP. jadiniiS. cerevisiae
112S. cerevisiaeS. bayanusS. cerevisiaeS. boulardiia
221S. cerevisiaeNot identifiedS. cerevisiaeS. cerevisiaeS. cerevisiae
GLSS. bayanusS. bayanusS. cerevisiaeS. cerevisiaea
H3BS. pastorianusS. bayanusS. cerevisiaeS. boulardiiS. cerevisiae
18BS. pastorianusS. bayanusS. cerevisiaeS. cerevisiaeS. cerevisiae
22BS. pastorianusS. bayanusS. cerevisiaeS. cerevisiaeS. cerevisiae
1RS. bayanusS. bayanusC. holmiiS. cerevisiaea
18RS. exiguusS. bayanusC. holmiiaa
19RS. pastorianusNot identifiedC. holmiiaC. humilis
20RS. exiguusNot identifiedC. holmiiaC. humilis
21RS. exiguusS. cerevisiaeC. holmiiS. cerevisiaea
  • aNot done.

For the other isolates the identification results obtained by the different methods were never in agreement.

3.3 Polymerase chain reaction fingerprinting

All the isolates and reference strains were subjected to RAPD-PCR analysis with primers M13 and RF2. The dendrogram based on the analysis of the random amplified DNA fragments revealed that yeasts isolated from sourdoughs grouped in five clusters at a level of similarity of approximately 45% (Fig. 1). The first (cluster A) contained the patterns of isolates LB, LS, 3B, 22B, 18B, and 1R, which shared a similarity of 89%, as well as those of isolates 111 and 112, which had a similarity of 94%. The second and third clusters (B and C) were formed by only one strain, namely the isolates 110 and 221. All the above-mentioned isolates proved to be remotely related with each of the reference strains used in this work and, unexpectedly, they were distinct from S. cerevisiae ATCC 9763. The fourth cluster (D) was composed of isolates VA and VB showing a correlation of 48% with S. pastorianus CBS 1538 type strain. The fifth cluster (E) showed a similarity of more than 60% and grouped the isolates 2PL3, GMAB, MM2, MM5, RM2, RM3, RM4, RM5, LGAT, 18R, 19R, 20R and 21R. This cluster demonstrated a similarity of approximately 47% with C. humilis CBS 6897 type strain. Reference strains of other species gave different profiles and these were not included in any cluster containing yeasts isolated from sourdoughs.

Figure 1

Cluster analysis of RAPD-PCR patterns from sourdough yeast strains obtained with primers M13 (a) and RF2 (b). Letters from A to E indicate cluster names. Distance values between branches in the dendrogram are reported as percentage of similarity (0–100%).

3.4 Mitochondrial DNA restriction analysis

The results of the mtDNA restriction analysis almost agreed with those of the grouping obtained with RAPD-PCR (Fig. 2). Actually, the restriction profiles of isolates LS, 22B, 1R, 3B and LB showed a similarity of 89%, whereas isolates 221 and 111 correlated at a lower similarity level (cluster α). Contrary to the results obtained by RAPD-PCR analysis, the isolate 18B appeared to be different from other S. cerevisiae isolates since it produced a digestion pattern, which shared only 47% similarity with this cluster. Isolate 110 was confirmed as another divergent strain of the same species revealing a low level of similarity (35%) with the S. cerevisiae group. Regarding the cluster of C. humilis (cluster β) the restriction profiles of isolates RM5, 19R, 2PL3, 21R, MM2 and 18R grouped at a similarity value higher than 90%, but they matched with that of the isolate LGAT at only 52%. Moreover, the digestion pattern of isolate VB, ascribed presumptively to S. pastorianus, had only 30% similarity with cluster β.

Figure 2

Cluster analysis of mtDNA patterns from sourdough yeast strains obtained with the HinfI and RsaI restriction endonucleases. Letters α and β indicate cluster names. Distance values between branches in the dendrogram are reported as percentage of similarity (0–100%).

3.5 Sequence analysis of ITS1-5.8S-ITS2 region

A single ITS1-5.8S-ITS2 DNA fragment was obtained from all the isolates used in this work with a length ranging approximately from 670 to 820 bp. Forward and reverse sequencing was carried out on thirteen isolates belonging to the five clusters identified by the RAPD-PCR analysis. Nucleotides from the beginning of ITS1 to the end of ITS2, including the 5.8S gene, were compared to available on-line sequences of reference strains using BLAST software. The amplicons from isolates 3B, 18B, 22B, 110, 111 and 221 had a homology ranging from 97% to 99% with S. cerevisiae, whereas that from isolate VA aligned with S. pastorianus at 99%. The amplicons from isolates LGAT, RM2, 2PL3, GMAB, 19R and 20R shared high homology values (97–100%) with C. humilis (Table 3). The sequence alignments were used to construct a phylogenetic tree by the neighbour-joining method (Fig. 3). This comparative analysis validated the distribution of sourdough strains into two main groups. The dendrogram showed that the strains ascribed to C. humilis, which were collected in different bakeries, strictly grouped in the major cluster. In contrast, strains attributed to S. cerevisiae were divided into two subgroups, which were each linked to a single bakery. Isolates 110, 111 and 221 derived from bakery F, while isolates 3B, 18B and 22B originated from bakery H. The only sourdough strain belonging to S. pastorianus (isolate VA from bakery D) was placed into the appropriate cluster.

View this table:
Table 3

Comparison between ITS sequences of representative strains isolated in this work and reference strains found in the NCBI data base

FirmStrainReference strainPercentage identityIdentities (bp)Gaps (bp)
A2PL3C. humilis CBS 6897100556/5560/556
GMABC. humilis CBS 6897100558/5580/558
RM2C. humilis CBS 689798553/5637/563
DVAS. pastorianus IFO 116799739/7411/741
ELGATC. humilis CBS 689797549/5630/563
F110S. cerevisiae IFO 1021798699/7121/712
111S. cerevisiae CBS 490398717/7295/729
221S. cerevisiae IFO 1021797697/7222/722
H3BS. cerevisiae CBS 490399732/7342/734
18BS. cerevisiae CBS 490399736/7393/739
22BS. cerevisiae CBS 490399735/7394/739
19RC. humilis CBS 6897100563/5630/563
20RC. humilis CBS 6897100563/5630/563
Figure 3

Unrooted phylogenetic tree of sourdough yeast strains and related species based on the ITS1-5.8S-ITS2 sequences. The bar indicates the percentage of sequence divergence.

4 Discussion

4.1 Phenotypic identification of yeast strains

Identification systems of yeasts based on physiological tests are time-consuming and sometimes hard to interpret due to ambiguous responses by some isolates in the media tested. Besides, phenotypic characters can be influenced by strain diversity and cultivation conditions. Therefore, we compared the laboratory protocols provided for each identification method [34,35]. This investigation demonstrated the presence of maltose-negative yeasts that could not be identified in a consistent way by any of the procedures of identification used in this work. Concerning the performance of the SIM method [34], the critical tests for identifying the species were those performed in liquid medium, such as the 0.01%-cycloheximide and vitamin-free growth tests, since no reproducible results were obtained. In the same way, using the key of Kurtzman and Fell [35] the vitamin-free growth test proved to be crucial. When this was positive the isolates could be attributed to S. bayanus, although they were maltose-negative, and when it was negative the isolates were identified as C. stellata, even if they grew with gluconate and 0.01% cycloheximide. Therefore, both identification methods cannot be used to identify C. humilis and similar maltose-negative species found in sourdoughs. The commercial identification kits, API 32 ID and Microplate YT Biolog, were unable to discriminate among sourdough yeasts, probably because they were designed to meet the needs of clinical diagnostics. Praphailong et al. [41] evaluated the potential use of the Biolog system for the identification of 72 yeast strains of 21 species isolated from food and wine, and considered some factors affecting the test, such as the culture medium, inoculum density and culture age. While 68% of the strains were correctly identified at the species level, the remaining strains, including S. bayanus, were not. Sand and Rennie [42] pointed out that misidentifications of species belonging to the genus Candida frequently occur using the Biolog and API systems. Furthermore, Mäntynen et al. [20] reported that the Biolog system had attributed wrongly some collection strains of C. milleri and most maltose-negative yeast isolates from Finnish sourdough to Kl. lodderae. In the same study other baker's isolates, which were ascribed to S. boulardii by the Biolog system, clustered together with S. cerevisiae reference strains by molecular methods. These observations are very similar to those presented in this work. All together these data indicate that rapid commercial systems are not sufficiently sensitive for an accurate identification of sourdough yeasts, though they are reliable tools for the phenotypic characterisation of strains, including assimilation patterns.

4.2 Genotypic identification and typing of yeast strains

Candida humilis is the main yeast species involved in sourdoughs used to prepare Italian sweet baked products. Other authors have also evidenced a meaningful presence of this species in sourdoughs. Gullo et al. [43] collected isolates of C. humilis from samples for the production of durum wheat bread at different intervals of time, ascertaining that this species was dominant and steady. On the basis of 18S rDNA and EF-3 PCR-RFLP pattern analyses, Mäntynen et al. [20] found strains closely related to C. milleri, a synonym of C. humilis, among yeasts isolated from Finnish industrial sourdoughs used for the production of rye bread. The C. humilis isolates collected in this work revealed a high degree of intraspecific homogeneity as shown by the results obtained by RAPD fingerprinting. This PCR technique demonstrated to be efficient in identifying this species without differentiating at the strain level. The same analytical method, when applied to strains from different bakeries belonging to the S. cerevisiae group, highlighted a high similarity between them and a certain distance from the reference strains. This observation allows us to suppose a genetic difference correlating with the adaptation to the habitat of recovery. This variation can be a useful tool to recognise S. cerevisiae originating from sourdough. Moreover, isolates from samples of the same bakeries often showed amplification patterns correlating over 80% similarity as is the case of isolates MM5, RM5 and RM2 (bakery B), VA and VB (bakery D), 111 and 112 (bakery F), and 18R, 19R and 21R (bakery H).

Mt-DNA restriction analysis did not recognise a significant polymorphism among the strains. This assumption was supported by our results obtained after pattern analysis. Nevertheless, strains LGAT and 110, isolated from sourdoughs of the bakeries E and F, respectively, showed a significant diversity from the relevant cluster, which was supported by some differences detected in the ITS sequence. This is especially true for isolate LGAT, which was derived from the oldest mother culture for Panettone (since 1920) investigated in this sampling. However, comparable works on the characterisation of yeast strains from food have reported dissimilar comments about the capability of this molecular technique to provide a good discrimination at the strain level [25,26,28,44].

The ITS sequences turned out to be a reliable tool for the identification of strains at the species level, since the groups obtained using the neighbour-joining method placed the strains in coherent clusters together with the reference strains. The discriminative potential of the ITS1-5.8S-ITS2 DNA region to investigate species boundaries was successfully used by some authors using RFLP analysis and sequencing [4548]. For some yeasts, the sequence variation within this part of the genome distinguishes species and subspecies better than more conserved portions of the ribosomal operon. This was particularly the case for the differentiation among representatives of the Saccharomyces sensu stricto group [48].

In this study no significant relationship was observed between phenotypic or genotypic characteristics of strains and geographical areas of isolation or type of the final products. Even if S. cerevisiae strains were isolated in sourdough samples for Cornetto and Pandoro, while C. humilis strains were collected only in those for Pandoro and Panettone, this difference regarding yeast species distribution was not substantial since most manufacturers used the same mother culture for the preparation of different final products throughout the year. Our work demonstrated the presence of dominant maltose-positive S. cerevisiae strains in Italian sourdoughs from different bakeries that formed a tight cluster among them, but not with the reference strain. Besides, it is interesting to note that in the same sourdough samples from bakeries C and H, C. humilis was detected at a lower concentration (data not shown). The occurrence of these two species in the same mother cultures gives a new outlook on the role of these yeasts and the relationship between them and with lactic acid bacteria in sourdough fermentation.


The authors greatly thank Dr. A. Lombardi, Prof. G. Soncini and Dr. Carlo Trivisano for their profitable advice, precious collaboration and technical support.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
View Abstract