1. Introduction
Sourdough technology, based on the use of dough fermented by lactic acid bacteria (LAB) and yeasts which coexist and establish stable interactions, has attracted interest since sourdough positively affects the nutritional, textural, and sensorial characteristics of cereal-based products [
1,
2,
3,
4].
It is well known that phenols, bioactive compounds widely present in wine, olive oil, and plant-based foods [
5,
6,
7], including cereals and their products [
8,
9,
10], are antioxidants and help prevent cardiovascular diseases, chronic inflammation, cancer, and diabetes [
11,
12,
13,
14,
15].
Microorganisms determine changes in chemical compounds present in the raw material [
16,
17,
18,
19] affecting also the bioactive compounds within cereal products [
20,
21], particularly in sourdough derived products rather than in yeast- or chemically-leavened bakery products [
4,
22]. Sourdough fermentation also affects the rheological properties of dough [
23] as well as the textural and sensorial characteristics of bread [
2,
24,
25,
26,
27,
28,
29]. In different geographical regions, research has been carried out aimed at improving the level of bioactive compounds and/or sensorial characteristics of bread, using starters or fortification of wheat flour or a combination of the two [
30,
31,
32,
33,
34,
35,
36,
37,
38]. However, sourdough exhibits a wide biodiversity, and the present research aims to investigate autochthonous Calabrian sourdoughs and strains.
In the southern Italian region of Calabria, the production of wheat sourdough bread, usually called either Pane di grano, Pane tradizionale, or Pane casereccio, is well established in many cities and villages. The mother dough, prepared either using soft or durum wheat flour, is stored and continuously propagated using traditional procedures handed down from one generation to another.
The biodiversity of the different artisanal bakery productions is poorly studied [
39,
40], and there have been studies neither on the antioxidant and sensorial properties of Calabrian sourdoughs and bread, nor on the modification of existing artisanal preparations using starters and technological procedures with an expected impact on the bread’s characteristics.
The need for research is highlighted both by the lack of information about existing traditional products, and by the current consumer demand for food which is healthier and more functional [
41] as well as appealing to the senses [
42,
43]. Research into traditional products would not only reveal the characteristics of traditional artisanal sourdoughs, but would also indicate a starting point for new formulations. Consumer demand for healthier bread justifies the creation of a multistrain starter able to outgrow the original microbiota of artisanal sourdoughs. For these reasons, autochthonous wheat-related Calabrian strains, selected based on useful technological properties, were chosen (see Material and Methods section for details) to make the starter culture used in this study.
The suitability of a starter is firstly determined by its stability over time, thus maintaining a certain reproducibility of a loaf’s characteristics, and secondly, by verifying its role and impact on the products. Therefore, the main aims of this work were to verify the persistence in sourdough of the multistrain starter culture—Lactobacillus sanfranciscensis, Leuconostoc citreum, and Candida milleri—throughout all stages of production (from laboratory to bakery plant) and to evaluate its role in modifying the content of polyphenols, the antioxidant capacity, and the textural properties of novel produced sourdoughs and bread.
There are two novel aspects of this work: firstly, it sheds light on artisanal sourdough biodiversity, and secondly, it combines artisanal knowledge of sourdough bread-making with the scientific basis for producing new bread.
2. Materials and Methods
Starter strains: Two strains of LAB,
L. sanfranciscensis B450 and
L. citreum B435, and a yeast strain,
C. milleri L999, were used as multispecies starter culture to produce a sourdough—in this study named SD. These strains were previously isolated from artisanal sourdoughs and selected for their properties useful in bread-making [
40]: the production of CO
2, a fast acidifying activity, the production of exopolysaccharides, and the exhibition of proteolytic activity. The yeast
C. milleri has a known stable association with
L. sanfranciscensis due to maltose metabolism. In particular the strain L999 was chosen for its resistance to high salt concentration, tolerance to low pH, and growth in the presence of acetic acid.
Strains B450 and B435 were propagated overnight at 30 °C in Sour Dough Bacteria (SDB) broth [
44] and in de Man–Rogosa–Sharpe (MRS) broth (VWR International s.r.l., Milan, Italy), respectively. Strain L999 was propagated overnight at 30 °C in Yeast Peptone Dextrose (YPD) Broth (Amresco, Milan, Italy).
Sourdough and bread production: Each overnight culture was harvested by centrifugation (5000 rpm for 10 min), washed once in 0.9% NaCl solution, resuspended to OD600 of 1.0 in the same solution, and used to prepare SD sourdough. Inoculated (SD) and noninoculated control (SDC) sourdoughs (500 g) were prepared in duplicate using durum wheat remilled semolina (Industria molitoria Minnini s.r.l, Italy) and sterile still mineral water with a dough yield (weight of the dough/weight of the flour × 100) of 170 that gave the best kneading performance. The raw ingredients for SD were inoculated with the prepared multispecies starter culture at 1% of the dough’s total volume. Both SD and SDC sourdoughs were kneaded by gloved hands in sterile stainless-steel trays, covered with plastic wrap, and incubated at 25 °C for 21 h to allow fermentation. Afterwards, sourdough propagations were carried out every two days and, after leavening, the sourdoughs were stored at 4 °C. After two months, SD and SDC sourdoughs were kneaded by a spiral kneading machine (CHEF Planetary mixer, Sigma s.r.l., Torbole Casaglia (BS), Italy) for 5 min and left to ferment for 4 h at 30 °C before baking 12 loaves × 500 g (6 for each type of sourdough) in an oven (BE-1 System Oven, Angelo Po Grandi Cucine s.p.a., Carpi (MO), Italy) at 200 °C for 15 min followed by a further 5 min at 180 °C.
In a bakery plant experiment, the SD sourdough propagated in laboratory (mother dough-SD m) was mixed both with artisanal Calabrian sourdoughs PF7 and PF9 [
40], here used as mother doughs, in a ratio of 1:1 to produce novel sourdoughs named PF7 M (SD m + PF7) and PF9 M (SD m + PF9). The PF7 and the PF9 are made by two different bakeries using their traditional mother dough. In detail, PF7 M and PF9 M were composed of soft wheat flour type 00, 20% mother dough, 2% salt, and 60% tap water. The PF7 M and PF9 M sourdoughs were kneaded by a spiral kneading machine (Mecnosud s.r.l., Flumeri (AV), Italy) for 15 min and left to ferment for 4 h at 25 °C before baking at 250 °C for 30 min 16 loaves × 500 g for each type of sourdough. The same procedural scheme was followed for the bakery sourdoughs PF7 F and PF9 F to produce traditional bakery loaves (
Figure 1).
Microbiological analysis: Plate counting was performed in triplicate for SD and SDC throughout 0–1140 h (2 months), for the SD m, PF7, PF9, PF7 F, PF9 F, for sourdoughs PF7 M (SD m + PF7), and PF9 M (SD m + PF9), and for PFC (baker’s yeast dough) as a control dough. After this, the sourdoughs were homogenized in 0.9% NaCl solution (1:10) and then diluted ten-fold. LAB were plated onto MRS agar and SDB agar supplemented with 100 mg/L cycloheximide (Oxoid, Milan, Italy), while yeasts were plated onto YPD agar (Amresco, Milan, Italy) supplemented with 100 mg/L chloramphenicol (Liofilchem Diagnostici, Roseto degli Abruzzi, Italy). Plates were incubated at 30 °C for 48 h anaerobically and aerobically for LAB and yeasts, respectively.
Monitoring of the added multistrain starter: Colonies from SD and SDC (0–21 h; 2 months), PF7 M and PF9 M (before baking) were randomly isolated based on their appearance and purified by restreaking on the above reported growth media. The presumptive LAB were tested for catalase and for Gram by the KOH method [
45]. All the purified isolates were stored at −80 °C. To verify the persistence of the added strains to the experimental sourdoughs, each pure culture was subjected to DNA extraction by InstaGene Matrix (Bio-Rad Laboratories, Milan, Italy) according to the manufacturer’s instructions, and then analyzed for Random Amplified Polymorphic DNA (RAPD)—PCR technique. PCRs were performed in a MasterCycler Nexux GX2 (Eppendorf, Milan, Italy) using the primer M13 in a 25 µL reaction as reported by [
46]. PCR amplicons were electrophoretically separated on 1.5% agarose gel using the GeneRuler 100 bp Plus (Thermo Fisher Scientific, Monza, Italy) as a ladder. The gels were stained with RealSafe Nucleic Acid Staining Solution (Real, Paterna, Valencia, Spain), checked under UV transillumination, and documented by the MicroDoc system (Cleaver Scientific, Warwickshire, UK).
The polymorphic profiles obtained from the pure cultures were compared with profiles of the axenic strains B450, B435, and L999 inoculated as starters.
PCR-Denaturant Gradient Gel Electrophoresis analysis: Sourdough SD and SDC were subjected to PCR-DGGE of a portion of the 16S and 26S rRNA [
47]. LAB reference strains were:
Lactobacillus plantarum subsp.
plantarum LMG 06907
T,
Lactobacillus pentosus LMG 10755
T,
Lactobacillus sanfranciscensis LMG 16002
T,
Lactobacillus brevis LMG 07944
T,
Lactobacillus buchneri LMG 06892
T,
Lactobacillus fructivorans LMG 09201
T,
Lactobacillus reuterii LMG 09213
T,
Pediococcus pentosaceus LMG 11488
T,
Lactobacillus pontis LMG 14187
T,
Lactococcus lactis subsp.
lactis LMG 06890
T, and
Pediococcus acidilactici LMG 11384
T together with
L. sanfranciscensis B450 and
L. citreum B435, previously isolated [
40]. Yeasts reference strains were:
C. milleri CBS 6897
T,
Kluyveromyces marxianus CBS 834
T,
Saccharomyces cerevisiae CBS 1171
T,
Wickerhamomyces anomalus CBS 5759
T together with
C. milleri L999, previously isolated [
40]. PCR products were analyzed by the D-code apparatus (Bio-Rad Laboratories) loading them onto 8% (
w/
v) polyacrylamide gels (acrylamide/bis-acrylamide, 37.5:1) in 1× TAE buffer containing 30% to 60% urea-formamide linear denaturating gradient increasing in the direction of electrophoresis. The electrophoresis parameters were 100 V with a running temperature of 60 °C for 7 h. After staining, DNA from bands of interest was reamplified; then, the PCR products were purified by Illustra™ GFX™ PCR DNA and Gel Band Purification Kit (GE Healthcare, Freiburg, Germany) and sequenced (Eurofins Genomics, Ebersberg, Germany). The sequences obtained were compared with those available at NCBI by the Blast search tool [
48] and submitted to GenBank for accession numbers (see Results section).
Chemical and rheological analyses: These analyses were carried out on sourdoughs and/or breads made in the laboratory and at the bakery. Total phenolic content, DPPH, and texture parameters were also performed on ten (PF1–PF10) artisanal Calabrian sourdoughs. The PFC dough produced with baker’s yeast supplied by a local bakery was used as a control. After baking, the loaves were left to cool and then analyzed, sampling both crust and crumb.
In detail, pH and total titratable acidity (TTA) were carried out in triplicate on SD; SDC sourdoughs (0–2–4–6–8–21 h); PFC (laboratory experiments); and on SD m, PF7, PF9, PF7 F, PF9 F, PF7 M, and PF9 M sourdoughs (bakery experiments). The pH was determined by a spin electrode pH meter (HI99161, Hanna Instruments, Ronchi di Villafranca Padovana, Italy), and the total titratable acidity (TTA) was determined by titration using 0.1 N NaOH on 10 g of each sample and expressed as mL of NaOH.
The organic acids were detected for SD; SDC sourdoughs/breads (laboratory experiments: 0-8-21 h and 2 months); PFC dough; the mother doughs for the bakery experiment (SD m, PF7, and PF9); PF7 F, PF9 F, PF7 M, and PF9 M sourdoughs/breads (bakery experiments). The organic acid extraction was carried out in triplicate according to Martorana et al. [
40]. In brief, sourdough homogenates were centrifuged at 5000×
g for 15 min, and the supernatant was filtered with 0.45 μm PTFE filter (Supelco Sigma-Aldrich, Milan, Italy). Then, the obtained water-/salt-soluble extract was analyzed by HPLC equipped with an Acclaim OA 5 μm (4 × 250 mm) at 30 °C and with the UV detector operating at 210 nm, with a flow rate of 0.6 mL/min The isocratic mobile phase was 100 mM Na
2SO
4 acidified with methanesulfonic acid to a pH of 2.65. External standard method was used to quantify the compounds detected. Data were expressed as mg/g. The quotient of fermentation (QF) was calculated as the molar ratio between
d, l-lactic and acetic acids.
To determine total phenolic content and antioxidant activity, five grams of each sourdough/bread was added to 50 mL of 80% methanol, mixed for 30 min, and centrifuged at 6000×
g for 20 min. The obtained extracts were analyzed for phenolic content by the Folin–Ciocalteu method [
49] and by the 2,2-diphenyl-1-picrylhydrazyl radical (DPPH) free radical scavenging method [
50].
Rheological analyses were performed on sourdoughs and breads by a TA-XT Plus Texture Analyzer (Stable Micro Systems Ltd., Godalming, UK). Data acquisition and curve integration were carried out by Exponent software 6.1.4.0 (Stable Micro Systems Ltd., Godalming, UK).
The PF1-PF10 sourdoughs; mature SD and SDC sourdoughs; PFC; the mother sourdoughs SD m, PF7, PF9; the traditional fermented sourdoughs PF7 F and PF9 F; and the novel sourdoughs PF7 M and PF9 M were subjected to the Stickiness, Penetration, Warburtons, and Kieffer tests.
The Stickiness test was performed using a Chen and Hoseney probe (A/DSC) (Stable Micro Systems Ltd., Godalming, UK) on 20 g of sample. The probe in contact with the sample measured the forces of insertion and withdrawal from the dough. To evaluate this attribute, the following parameters were used: pretest speed: 0.50 mm/s; test speed: 0.50 mm/s; post-test speed: 10.00 mm/s; distance: 4.0 mm; trigger force: 5.0 g; data acquisition rate: 400 pps. For each sample, ten replicates were carried out. In order to assess the hardness of the dough, the Penetration test was performed on 110 g of sample using a 6 mm cylindrical probe (P/6) (Stable Micro Systems Ltd., Godalming, UK). The set parameters were: pretest speed: 2.00 mm/s; test speed: 3.00 mm/s; post-test speed: 10.00 mm/s; distance: 20.0 mm; trigger force: 5.0 g; data acquisition rate: 200 pps. For each sample, six replicates were carried out. The Warburtons test was performed using the Warburtons dough stickiness system (A/WDS500) (Stable Micro Systems Ltd., Godalming, UK) on 500 g of sample. The sample, placed into a box, is slightly compressed by a retaining plate with a slot in the middle. Then, a blade is driven through the slot to obtain indications of its consistency. The test was carried out applying the following parameters: pretest speed: 5.00 mm/s; test speed: 2.00 mm/s; post-test speed: 10.00 mm/s; distance: 40.0 mm; trigger force: 10.0 g; data acquisition rate: 500 pps. For each sample, five replicates were carried out. The Kieffer test was performed using a Kieffer probe (A/KIE) (Stable Micro Systems Ltd., Godalming, UK) on 50 g of sample. In this test, the dough is stretched by the probe. The following parameters were used: pretest speed: 2.00 mm/s; test speed: 3.30 mm/s; post-test speed: 10.00 mm/s; distance: 75.0 mm; trigger force: 5.0 g; data acquisition rate: 400 pps. For each sample, ten replicates were carried out.
Concerning the bread, the PF7 F, PF9 F, PF7 M, and PF9 M breads were analyzed for the Penetration test, Cut slice test, and Texture Profile Analysis (TPA). Three replicate measurements were made on bread samples.
A single whole loaf was used in the Penetration test to assess its hardness and fracturability. The test was performed using a 3 mm cylindrical probe (P/3) (Stable Micro Systems Ltd., Godalming, UK) that was driven into the bread with the following parameters: pretest speed: 1.00 mm/s; test speed: 0.50 mm/s; post-test speed: 10.00 mm/s; distance: 10.0 mm; trigger force: 5.0 g; data acquisition rate: 400 pps. For each sample, three replicates were carried out. Moreover, the hardness was evaluated on slices of bread. In detail, 2 cm thick slices were analyzed by the Cut Slice test using a Blade Set with Knife probe (HDP/BSK) (Stable Micro Systems Ltd., Godalming, UK). The set parameters were: pretest speed: 1.50 mm/s; test speed: 2.00 mm/s; post-test speed: 10.00 mm/s; distance: 30.0 mm; trigger force: 25.0 g; data acquisition rate: 400 pps. For each sample, three replicates were carried out.
The TPA test was performed using a 100 mm compression platen (P/100) probe (Stable Micro Systems Ltd., Godalming, UK) on a whole loaf sample with the following parameters: pretest speed: 1.00 mm/s; test speed: 5.00 mm/s; post-test speed: 5.00 mm/s; distance: 20.0 mm; trigger force: 5.0 g; data acquisition rate: 400 pps. For each sample, three replicates were carried out. From test results, the chewiness, springiness, and resilience parameters were taken into consideration.
Statistical analyses: Excel 2010 software (Microsoft, Milan, Italy) was used to calculate the means and standard deviations on three replicates. The means were analyzed by one-way ANOVA and a Tukey’s test, at 5% probability, using the SPSS 17.0 software (SPSS Inc., Chicago, IL, USA).
4. Discussion
In this study, a multistrain starter culture was developed and used to manufacture novel wheat sourdoughs and breads exploiting its effect on antioxidant and rheological properties.
The results regarding microbial loads are consistent with those reported for traditional and experimental sourdoughs [
30,
35,
37,
40], and the pH and TTA evolution is comparable as well [
40,
51].
The concentration of lactic and acetic acids in our sourdough samples was similar to values reported by Settanni et al. [
51]. To present an interpretation of the role of these organic acids, the QF is commonly reported [
35,
63,
64]. The sourdough made in the laboratory and the novel sourdoughs can positively influence the aroma profile and the structure of the final products, having a QF in the range 1.5–4 that is considered positive in sourdough bread-making [
65].
Different studies have reported the stability of the microbial population during continuous propagation of sourdough [
30,
35,
66,
67]. The mature sourdough SD made in the laboratory and the novel sourdoughs made at the bakery were dominated by
L. sanfranciscensis and
C. milleri confirming the key role of this LAB [
68] and its strict association with
C. milleri [
63,
69].
The PCR-DGGE analysis of the experimental sourdoughs revealed their microbial ecology throughout the propagation time. The size of the bands is in accordance with Gatto et al. [
47]. The bands identified as
T. aestivum and
C. turicensis with a sequence identity lower than 97%, were poorly related to those deposited in the GenBank database. Nevertheless, the detection of
T. aestivum is consistent with the primers used that also amplify plant material as reported by Gatto et al. [
47]. Concerning the detection of
C. turicensis, various authors have previously reported its presence in flour, cereal-based products, and bread [
70,
71,
72,
73]. Moreover, DGGE band sequencing demonstrated the presence of members of
Enterobacteriaceae in the naturally fermented sourdough. Other authors have detected enterobacteria in sourdoughs [
47,
74]. Some members of enterobacteria such as
Enterobacter cowanii and
P. agglomerans have been isolated from nonsterilized flour [
75]; therefore, it is highly probable that they derived from the flour used to prepare the sourdoughs SD and SDC. It is interesting to highlight the role of the multistrain starter added in SD in outgrowing the above reported harmful bacteria that instead dominated the SDC bacteria population up to 2 months.
The phenolic content is usually related to different factors other than the microbial composition. Although cereal grains are a good source of bioactive compounds, their level and activity are affected by cereal species and varieties, the nature of the antioxidants, the milling process producing flours with different degrees of refining, and food processing [
1,
76,
77,
78,
79,
80,
81,
82,
83,
84,
85]. In detail, durum wheat has higher phenol and antioxidant activity than soft wheat [
76,
86]; similarly, whole flour and bread more than the refined [
87,
88,
89,
90], since phenolic compounds are mostly present in the outer layers of the grain that are usually removed by the milling industries [
91].
Consistent with the above reported literature, lower values characterize the majority of PFs (see
Table S1) made with soft wheat and the PFC (lowest value), which is, in addition, made with baker’s yeast. The phenolic content increase observed in the sourdough made in the laboratory (see
Table 4) could be due to the progress of lactic acid fermentation that determines a pH decrease. The acidification, in fact, can enhance the phenol extractability, and LAB can determine the hydrolysis of complex phenolic forms shaping the phenolic profile of a matrix and its radical scavenging activity [
1,
20,
92,
93]. The higher values of SD compared to SDC could be attributable to the lower pH and higher LAB loads than the SDC values. These results are in agreement with the reported correlation between the antioxidant properties and the level of inoculum and pH value [
94].
The difference in DPPH inhibition capacity observed between the novel sourdoughs could be due to the different LAB biodiversity revealed by the RAPD profiles, higher in PF9 M than in PF7 M; in fact, different microbial fermentation could lead to a variable effect due to the activation of inactive compounds but also deactivation of bioactive compounds [
94].
Correlating phenolic content and antioxidant activity, our results could be attributable to the type of phenolics present in fermented doughs that possess different antioxidant activity [
77]; therefore, the phenols associated with sourdoughs exhibiting lower phenolic content exert the same or higher antioxidant activity per mass unit compared to others. Similar results have been reported for barley and oat grains [
95,
96].
To the best of our knowledge, no studies have investigated phenols and antioxidant properties of artisanal wheat sourdoughs. The availability of various methods of detection and the different ways to express results do not make it easy to directly compare our results with those of others in the same field. However, our results are consistent with the trend reported by other authors, demonstrating that selected starters for bread-making increase the content of phenolic compounds, and the effect on antioxidant activity is not always positively correlated with the content of phenols [
21,
97,
98,
99].
Our results on breads are consistent with findings of several authors who reported the effects of baking on phenols and antioxidant capacity of different types of cereal-based products [
86,
100,
101,
102,
103,
104] highlighting the role of cereals, bread formulations, baking parameters, as well as extraction and analytical methods. Our results on breads—decrease in phenolic content and increase in DPPH inhibition—could be attributable to Maillard reaction products such as melanoidins that possess antioxidant properties [
105,
106] but are also able to bind polyphenols to their protein backbones [
107], thus determining a decrease in phenol content. Moreover, our findings can be explained by many factors acting synergistically such as the presence of antioxidant compounds apart from phenols [
108], the formation of thermally induced degradation products [
109], and polyphenol-polysaccharide complexes [
110] that increase the antioxidant scavenging activity. The addition of the multistrain starter mother sourdough positively affected the breads produced, inducing either an increase or a restrained decrease in antioxidant activity and phenolic content. The anticipated health benefits are linked to the consumption of a portion of bread (100 g), which will deliver bioactive (antioxidant) compounds to the human body with the potential to prevent diseases [
111]. However, in order to estimate the degree of such health benefits, further studies considering the bioavailability of these compounds are necessary.
Wheat dough texture parameters are important for dough handling properties, dough machinability, and as predictors of bread quality [
112]. The texture parameters of the artisanal Calabrian sourdoughs reflect the interesting variability in bread-making; each component of the dough formulation such as type of flour, water, yeast or mother dough, and kneading method are responsible for the final characteristics of the bread [
59,
113]. The sourdough made in the laboratory (SD) is stickier, less hard, and less extensible compared to the majority of the sourdoughs made at the bakery. Most of the rheological parameters of the novel sourdoughs differentiate them from the traditional ones; this could be explained by considering the low pH due probably to a synergistic effect of traditional sourdough microbiota and the multistrain starter added (PF7 + SD m; PF9 + SD m) [
103]. The TPA parameters of the breads produced at the bakery are comparable with data reported by Casado et al. [
114]. The addition of the multistrain starter mother sourdough affected the rheological properties of the novel formulated breads. In particular, the PF9 M bread, compared to the PF9 F bread was less hard and was characterized by less chewiness (rubbery texture during mastication), maintaining its springiness (size and shape recovery after compression); features reported for the acceptability of bread [
115].