Microorganisms | Free Full-Text | 16S-rRNA-Based Metagenomic Profiling of the Bacterial Communities in Traditional Bulgarian Sourdoughs

3.1. Physicochemical and Microbiological Characterization of Bulgarian Sourdoughs

In the present study, we explored five samples of traditional sourdough collected from three geographical locations: sample D5 from Smolyan town (Rhodope mountains, South Bulgaria), samples D8 and D9 from Bansko town (at the foot of Pirin mountain, Southwest Bulgaria), and samples D11 and D12 from Ruse (Danube Plain, North Bulgaria) (Figure 1). These areas were chosen because they have various geographical and climatic qualities, and the selected local bakeries also have a solid track record of producing traditional Bulgarian sourdough.
All samples were characterized with regards to their dry matter content, pH, total titratable acidity, and lactic acid and acetic acid content. Total viable counts of the two major microbial groups—LAB and yeasts—were also estimated. The results of the sourdough characterizations are presented in Table 1.
The dry matter content ranged from 41.95% (sample D8) to 56.57% (D5), with variations resulting from the raw materials used, sourdough formulation, and manufacturing technology. Dry matter content was positively correlated to the resulting pH of the sourdoughs. The higher water content ensured more free water available and easier flour hydrolysis, therefore a higher amount of fermenting sugars and an easier access of the growing microbiota to the nutrient ingredients, in particular the LAB that produce organic acids. As the dry matter increased, a tendency toward elevated pH was observed: the lowest pH value estimated in sample D8 correlated with the lowest dry matter content, and the highest pH value with the highest dry matter content in sample D5. Vera et al. [31] studied the fermentation dynamics of natural French sourdough, reporting dry matter content of 44.41% at the end of the fermentation, a pH of 3.70, and TTA of 22.4 mL NaOH. While the pH value correlated with dry matter content in a similar pattern as shown in our study, the reported TTA value was much higher than that of the Bulgarian sourdoughs.
According to a recently published review by Calvert et al. [32], sourdough starters should have a pH of 3.9–4.1 and TTA of 14–16 [4]. These reference values vary a little in the literature: a pH of 3.5–4.0 is typically accepted as optimal for sourdough breads [33,34]. Most of our analyzed sourdoughs showed values within this pH range, except for sample D5 (pH 4.28). However, the pH values we observed were similar to the pH intervals of 3.74–4.28 and 3.41–3.70 reported for Italian and French sourdoughs, respectively [4,31].

Regarding TTA, only sample D8 is within the recommended values, with TTA of 15.6. The other four samples had lower TTA values, which could be primarily attributed to the higher dry matter content (that, in turn, influences the dominant microorganisms) but also to other factors such as the number of backsloppings and other parameters employed in the preparation method.

The development of the sourdough fermentation is often evaluated through the fermentation quotient (FQ, i.e., lactic-to-acetic-acid molar ratio). The FQ is a parameter usually used to evaluate the sourness of sourdough bread [4]. In our study, sourdoughs D8 and D9, prepared from wholegrain wheat flour, showed higher FQ values than the samples from white flour—6.31 ± 0.72 and 5.09 ± 0.91, respectively. All samples from white flour had an FQ within the optimal range.

The LAB viable counts of the analyzed sourdoughs were between 9.92 ± 0.96 (D11) and 11.68 ± 0.58 log CFU/g (D5), and the yeast viable counts in the sourdoughs varied from 7.72 ± 0.34 (D11) to 9.59 ± 0.68 log CFU/g (D9). The high TTA could not be linked to the LAB viable counts alone. It may be affected by the homo-/heterofermentative character of the predominant LAB strains in sourdoughs as a major factor for the type and amount of organic acids produced during fermentation, as well as other related endogenous and exogenous baking parameters, such as the fermentation time and temperature, type of cereal and its extraction rate, dough yield, fermentation quotient, number of backsloppings, etc. It was interesting to find that sample D8, which had the lowest DM content, also harbored the highest ratio between LAB and yeast viable counts (2680). This could also be correlated with the lowest pH and highest TTA values recorded for this sample. Samples D5 and D12, which had similar DM values, showed similar LAB-to-yeast ratios, suggesting a link between DM content and the high prevalence of lactic acid bacteria over yeast.

Other studies on sourdough microbiota reported LAB to yeast ratios of 10:1 to 100:1, with 100:1 considered ideal [35,36]. Only one of our studied samples (D9) showed a ratio within this interval—59:1. The other four samples had higher, widely varying ratios, which again could be attributed to the differences in the raw materials and preparation technology.
Organic acids are major metabolites of sourdough fermentation with effects on the gluten network structure and dough elasticity, which also significantly contribute to the specific organoleptic properties of sourdough products compared to other commercially leavened baked products [37,38]. The quantity and type of the organic acids produced during sourdough fermentation depend on various factors such as the microbiota composition and processing parameters (e.g., flour type, dry matter content, dough yield, fermentation time and temperature, and NaCl concentration) [32,39,40].
Many authors have reported higher levels of lactic acid compared to acetic acid in traditional Type-I sourdough starters, which is attributed to the higher number of homofermentative LAB in the sourdough and the lower viable number of acetic acid bacteria (AAB) [41]. The same trend was observed in our study, where LA levels were significantly higher than AA levels in all analyzed samples, from 2.11 times for sample D5, which had the highest DM content (56.57 ± 0.96%), to 6.31 times for sample D8, which had the lowest DM content (41.95 ± 0.47%) (Table 1). It is interesting to note that samples D5 and D12, which had similar DM content, generated the highest AA contents of 37.8 ± 0.9 mM and 29.8 ± 0.8 mM, respectively. These observations also confirm that, generally, the DM of sourdoughs plays a significant role both for the total organic acid production and the ratio between the main organic acids.
The variations in lactic acid and acetic acid in our samples are similar to other reported values of these compounds in sourdough bread [41,42]. The estimated levels of lactic acid in our SD samples varied from 70.9 ± 1.8 to 97.2 ± 1.5 mM (6.38–8.75 g/kg). Debonne et al., (2020) [43] reported a lower LA content of 53 ± 2 mM/kg in Type-C sourdough, while Komatsuzaki et al., (2019) [44] estimated much higher LA values for sourdoughs prepared from wheat and rye flour, reaching approximately 15 mg/g at 8 °C and 30 mg/g at 28 °C. However, in this case, the fermentation was performed with two specially selected natural LAB and yeast strains, which confirmed that the composition of the fermenting microbiota had a major effect on LA formation. Therefore, starter culture selection is essential for the production of sourdough with the desired organic acid levels and profiles.
Acetic acid in our samples ranged from 15.1 ± 1.0 mM (D11) to 37.8 ± 0.9 mM (D5). Debonne et al., (2020) [43] reported similar AA concentrations of 39 ± 1 mM/kg in 1 commercial sourdough, but in 2 other commercial sourdough samples, the AA levels were more than 2 times higher at 89–99 mM/kg sourdough, which was attributed to the presence of Fructolactobacillus sanfranciscensis.
Some researchers attribute increased acetic acid levels to yeast activity since yeast releases fructose, which is transformed into acetic acid by heterofermentative LAB [45,46]. Indeed, acetic acid production can be increased by employing obligate heterofermentative LAB and fructose and citrate as alternative electron acceptors [33,34]. However, it was also found that acetic acid inhibits yeast in sourdough more than lactic acid [47,48,49]. In our study, no correlation between yeast viable counts and AA production was observed, which indicates that organic acid production and tolerance are more dependent on LAB, AAB, and yeast species and strains, as reported by other authors [45,50,51]. No correlation between the geographical origin of the samples and their physicochemical characteristics and the estimated LAB and yeast counts was observed. Clearly, the differences found were mainly related to other factors, such as the type of flour used, the recipe, and the preparation method.

3.4. Bacterial Communities in the Studied Sourdough Samples

The bacterial sequences from 16S rRNA genes assigned to bacterial phyla and their relative abundance (Figure 2A) varied slightly between the samples. As expected, all mature sourdough samples were dominated by Firmicutes phyla (69.2% to 95.4%), with a lower abundance of Proteobacteria (3.24% to 20.05%).
V3-V4 regions have been mostly used for identifying 16S rRNA gene sequences [52,53]. An increased presence of Cyanobacteria spp. was found in samples D11 and D12. Since these two sourdoughs originated from the same bakery, Cyanobacteria could have been introduced through the use of contaminated water (water source, pipelines, or containers). Several studies on cyanotoxin contamination of food have found evidence of Cyanobacteria in aquatic products, grains, fresh produce, dietary supplements, and maize [54,55,56], with the usual contamination route of irrigation with contaminated water or the use of contaminated water during food processing. It has been demonstrated that Cyanobacteria contamination can quickly occur throughout the entire food chain [57,58].
Most of the OTUs were classified at the genus level and are shown in Figure 2B. Lactobacillus was the main genus found in all sourdough samples (D5—68.77%, D8—67.24%, D9—51.02%, D11—78.92%, and D12—59.73%). The genus Pediococcus was present at very low levels (0.2%), with the exception of sample D12 (7.53%). Menezes et al. (2020) [23] also reported a low relative abundance of Pediococcus pentosaceus (0.02%) in Brazilian sourdoughs, whereas Pediococcus was the dominant genus in western China sourdoughs [59]. Similarly, the genus Pediococcus was reported as the most abundant genus in wheat-based sourdoughs in Iran [60].
Interestingly, Weissella was the second most predominant genus of LAB found in the wholegrain samples D8 and D9 (from the same bakery in Bansko), with 16.98% and 42.84%, respectively. In comparison, it was found in low abundance (<1.5%) in the rest of the samples made of white flour. In this case, the origin of the same bakery is related to the presence of Weissella representatives. At the same time, the two samples differed in the relative amount of Weissella, which may be related to the different wheat species (T. aestivum and T. monococcum) used for the wholegrain sourdough. Weissella confusa has been isolated from various habitats, including the skin, milk, and feces of animals; human saliva, breast milk, and feces; and, in particular, in starchy or cereal-based foods [61,62]. Based on these habitats, its presence in traditional Asian and African fermented foods and in European sourdoughs is quite common, as reported by several authors [10,63,64]. Weissella spp. are attracting much interest among researchers since various strains have been found to exert probiotic potential, and others can produce valuable oligosaccharides with prebiotic properties and polysaccharides with biotechnological applications. However, some representatives of this species (W. cibaria and W. confusa) have been recognized as opportunistic pathogens [62]. Therefore, identification at the species level of Weissella spp. found in samples D8 and D9 would be necessary in order to assess whether the present strains would have a beneficial or detrimental effect.
The genus Serratia was found in high abundance (11.73%) in 1 of the samples (D5), while in all other samples the amount was below 1%. Since this sample is from a different location to the other four samples, the presence of Serratia sp. may be related to the microflora specific to the producing bakery. Serratia spp. are widely distributed in nature and have been isolated from the soil [65]. Therefore, their presence in a hospital environment and different foods—dairy, meat, and fish products [66,67]—is associated with breaches in hygienic practice.
Herbaspirillum spp. were found in all the samples, with higher abundance levels of between 2.5% and 3.8% in samples D5, D11, and D12 and lower levels in the wholegrain samples D8 and D9. The presence of Herbaspirillum species is closely associated with plants, both endophytically and epiphytically. These bacteria were isolated from different cereals, such as rice, maize, and sorghum, which might indicate a route for their presence in sourdoughs [68]. In addition, studies by Barton et al., (2006) and Salter et al., (2014) [69,70] found that Herbaspirillum spp. were among the contaminating species in PCR reagents and DNA extraction kits, which could also be the reason for their detection in sourdoughs.
Other taxonomic units that fell under the top 12 identified genera (Figure 2) with low abundance in all samples were Bacteroides (most dominant with 1.12% in D12), Leuconostoc (most dominant with 1.12% in D11), Bacillus (<1%), and Sphingomonas (0.1–0.5%).
Regarding the profile of LAB, which tend to be the most critical group for sourdough matrices, fewer reads were classified to the species level (Figure 3).
The diversity of LAB in sourdoughs has been found to depend on several factors, such as the raw materials, geographical origin [60,71], preparation technology, microbiota at the bakery (production environment), human microbiota, and methods applied for sampling and identification. Moreover, the results from next-generation sequencing can be affected by the choice and selectivity of the primers used in targeted approaches, sequencing depth, read length, and the algorithms and databases available and utilized for post-sequencing data analysis [72].

The samples in our study were not representative enough to make conclusions regarding all of these factors. Yet, the results from different samples originating from the same bakery show that the production location (environment) is related to the presence of some common species. For example, Lacticaseibacillus casei (Lactobacillus casei) and Loigolactobacillus coryniformis (Lactobacillus coryniformis) were mainly presented in samples D8 and D9 taken from one bakery (Bansko town, Southwest Bulgaria), and Companilactobacillus farciminis (Lactobacillus farciminis) and Weissella cibaria were found in both sourdoughs from Ruse town, North Bulgaria (at the Danube River). Sourdough specificity related to location is an important feature that could be used for sourdough metagenomic authentication. However, we observed that samples from different bakeries prepared from white wheat flour shared common species. For example, Lactobacillus delbrueckii was in almost equal proportions in samples D5 (Smolyan, South Bulgaria) and D8 (Ruse, North Bulgaria). Ligilactobacillus agilis (Lactobacillus agilis), Limosilactobacillus fermentum (Lactobacillus fermentum), and Leuconostoc mesenteroides were the most common in these two sourdoughs. However, some specific species were found only in specific samples, independent of the bakery, such as Pediococcus parvulus in sample D12, Schleiferilactobacillus perolens (Lactobacillus perolens) in sample D5, and Ligilactobacillus salivarius (Lactobacillus salivarius) in sample D11, which could be attributed to other factors.

Other authors [23,73] have found Companilactobacillus farciminis in Italian and Brazilian wheat sourdoughs. It was also isolated from Croatian organic flours and applied as a starter culture in sourdough fermentation to reduce bread spoilage [74]. These diverse locations show that this species is related to the raw materials rather than the geographical factor. Another interesting species, Lig. agilis, was identified in traditional Iranian wheat sourdough [75] and in the fermented soy-based product tempeh, which indicates the ability of this species to ferment raw materials of different plant origin. Along the grain-feed chain, a strain of Lig. agilis was isolated from pig manure and showed promising probiotic potential [76].
The former genus Lactobacillus was predominant in the studied sourdoughs and had the highest diversity with 13 identified species. Leuconostoc mesenteroides (D11 and D5, from different bakeries, but both from white flour) and W. cibaria (D9 and D8 from the same bakery, both from wholegrain flour) were the only identified species of the respective genera. Other researchers [77,78] found that W. cibaria was related to the fermentation of sourdoughs from alternative grains such as quinoa, amaranth, and sorghum, which indicates a broader distribution of this species during grain fermentation.
Pediococcus spp. were mainly represented by Pediococcus parvulus (sample D12), and Pediococcus pentosaceus was detected at very low levels in samples D8 and D9 from the same bakery. The occurrence of these two species is not very common in sourdoughs, but it has been reported by other authors [79]. In a review by Landis et al., (2021) [72], P. parvulus was detected in old sourdoughs. In other fermented products, such as wine, Pediococcus spp. produce various enzymes that generate desirable aroma compounds [80]. Immerstrand et al., (2010) [81] explored this species’ potential probiotic and culture-protective characteristics, which indicates the potential for other beneficial applications of strains originating from sourdoughs.

Read more here: Source link