An Insight into Goat Cheese: The Tales of Artisanal and Industrial Gidotyri Microbiota

The purpose of this study was to determine for the first time the microbiota in artisanal-type and industrial-type Gidotyri cheeses and investigate the influence of the cheese-making practices on their composition using culture-independent techniques. The microbiota present in artisanal with commercial starters (Artisanal_CS, n = 15), artisanal with in-house starters (Artisanal_IHS, n = 10) and industrial (Ind., n = 9) Gidotyri cheese samples were analyzed using a targeted metagenomic approach (16S rRNA gene). The Ind. Gidotyri cheese microbiota were less complex, dominated by the Streptococcaceae family (91%) that was more abundant compared to the artisanal Gidotyri cheeses (p < 0.05). Artisanal cheeses were more diverse compositionally with specific bacterial species being prevalent to each subtype. Particularly, Loigolactobacillus coryniformis (OTU 175), Secundilactobacillus malefermentans (OTU 48), and Streptococcus parauberis (OTU 50) were more prevalent in Artisanal_IHS cheeses compared to Artisanal_CS (p ≤ 0.001) and Ind. (p < 0.01) Gidotyri cheeses. Carnobacterium maltaromaticum (OTU 23) and Enterobacter hormaechei subsp. hoffmannii (OTU 268) were more prevalent in Artisanal_CS cheeses compared to Artisanal_IHS cheeses (p < 0.05) and Ind. cheeses (p < 0.05). Hafnia alvei (OTU 13) and Acinetobacter colistiniresistens (OTU 111) tended to be more prevalent in Artisanal_CS compared to the other two cheese groups (p < 0.10). In conclusion, higher microbial diversity was observed in the artisanal-type Gidotyri cheeses, with possible bacterial markers specific to each subtype identified with potential application to traceability of the manufacturing processes’ authenticity and cheese quality.


Introduction
Cheese has constituted an important component of the human diet for millennia, leading to the industrialization of its production. Nevertheless, artisanal cheeses are gaining increasing interest from consumers. As no consensus definition exists, artisanal cheeses are generally considered to be hand-made cheeses produced using traditional cheese-making practices from cow, sheep and/or goat milk (preferably raw) on-farm or in small-scale dairies interlinked to a geographical region and culture [1][2][3]. The traditionally produced artisanal cheeses are characterized by increased microbial diversity associated with superior flavor, aroma and texture, while the standardized manufacturing of industrial cheeses leads to a less complex microbiota in an attempt to improve product safety at the expense of sensorial quality [3].
The composition of the microbial community collectively known as the cheese microbiota has a fundamental role in the production processes of curd formation and ripening, safety, and quality of the final product. Lactic acid bacteria (LABs) are the principal component of this microbial community, while pathogenic and spoilage bacteria such as Escherichia coli, Staphylococcus spp., Pseudomonas spp. are absent or present in low numbers

Collection of Gidotyri Cheese (Goat Cheese) Samples
Thirty-four Gidotyri cheese samples were collected in their original packages (1 kg net weight for industrial Gidotyri cheeses and 500 g net weight for artisanal Gidotyri cheeses) from eight dairy establishments located in the Epirus region production area ( Figure 1A) and then transported under refrigeration (4 • C) in no more than two hours, for laboratory analysis. Cheese samples were obtained from six different industrial producers (Ind.) and two artisanal dairies, one using commercial starters (Artisanal_CS) and the other using in-house starter cultures (Artisanal_IHS), with different production procedures ( Figure 1B) such as thermalization (63 • C, 15 min for artisanal cheeses) or pasteurization (72 • C, 15 s for industrial cheeses) of milk, type of starter cultures for the acidification procedure, production capacity (small-scale dairies vs industrial dairies). Samples from different production batches were collected from each dairy after 3 months of ripening (milk collection for cheese production took place in July 2022). All cheese samples were collected in November 2022.
Microorganisms 2023, 11,123 3 of 17 Thirty-four Gidotyri cheese samples were collected in their original packages (1 kg net weight for industrial Gidotyri cheeses and 500 g net weight for artisanal Gidotyri cheeses) from eight dairy establishments located in the Epirus region production area ( Figure 1A) and then transported under refrigeration (4 °C) in no more than two hours, for laboratory analysis. Cheese samples were obtained from six different industrial producers (Ind.) and two artisanal dairies, one using commercial starters (Artisanal_CS) and the other using in-house starter cultures (Artisanal_IHS), with different production procedures ( Figure 1B) such as thermalization (63 °C, 15 min for artisanal cheeses) or pasteurization (72 °C, 15 sec for industrial cheeses) of milk, type of starter cultures for the acidification procedure, production capacity (small-scale dairies vs industrial dairies). Samples from different production batches were collected from each dairy after 3 months of ripening (milk collection for cheese production took place in July 2022). All cheese samples were collected in November 2022.  For each dairy, samples were collected from different cheese-making lots at the same ripening stage (3 months). The number of Gidotyri cheese samples from each dairy type is presented in the figure. (B) The flow chart shows the main production steps of Gidotyri cheese including the differences between industrial and artisanal manufacturing processes. CS: commercial starter cultures; IHS: in-house starter cultures; Ind.: Industrial.

DNA Extraction
A 25 g aliquot was sampled from each cheese core once and homogenized in 225 mL of buffered peptone water (LAB M, Bury, Lancashire, UK), using a Stomacher (Laboratory Blender Stomacher 400; Seward, London, UK) for 2 min at 260 rpm. Ten milliliters of the filtered homogenized sample were collected in a 15-mL conical centrifuge tube and high-quality total DNA was extracted using DNeasy PowerFood Microbial kit (Qiagen, Hilden, Germany) according to the manufacturer's instruction. DNA concentrations were measured using a fluorescence spectrometer (Qubit, Life Technologies, Carlsbad, CA, USA). The samples were stored at −20 • C until analysis.

High Throughput 16S rRNA Sequencing
Aliquots of the obtained DNA of each sample proceeded for the characterization of the microbial diversity through PCR amplification, library preparation and high-throughput sequencing of the V3-V4 region of the 16s RNA gene. The targeted DNA region was amplified using the primers 341F and 806R [42]. The amplicon libraries were prepared using Nextera XT index kit (Illumina Inc., San Diego, CA, USA) and purified using the AMPure XP system (Beckmann Coulter, Krefeld, Germany) according to the manufacturers' instructions. Sequencing was conducted in a paired-end mode (PE300; only using reads of 275 each) with pooled samples containing 20% (v/v) PhiX standard library in the MiSeq Sequencing System (Illumina Inc., San Diego, CA, USA) using the MiSeq Reagent Kit v2 (300-cycles) (Illumina Inc., San Diego, CA, USA) amplifying the 465 bp fragment.

Data Analysis and Bioinformatics
The 16S rRNA gene amplicon data were analyzed and further processed using the "Integrated Microbial Next-generation sequencing" platform based on UPARSE. (IMNGS, www.imngs.org, accessed on 7 July 2022). A de-multiplexing (demultiplexer v3.pl) was performed before the sequences were trimmed by ten nucleotides. Sequences with nucleotides <200 and >600 and expected errors in paired reads >3 were excluded, and samples were screened for chimeras [43]. Operational taxonomic units (OTUs) were clustered at 97% similarity and OTUs with a relative abundance of <0.25% were removed. To generate a graphical overview of the alpha and beta diversity and the microbial composition, taxonomic binning was performed by Rhea using the set of R-scripts described by [44]. For all given results, p-values were corrected for multiple comparisons using the Wilcoxon rank-sum and/or Kruskal−Wallis Rank Sum statistical tests, unless stated otherwise. Significant OTUs were then identified at species level by EzBioCloud's 16S rRNA gene-based ID (www.ezbiocloud.net, accessed on 16 September 2022). Data were visualized using Illustrator CS6 Version 16.0.0 (Adobe Inc., San José, CA, USA).

DNA Sequencing Analysis and Alpha Diversity
A total of 1,399,249 raw paired-end reads were sequenced from the thirty-four Gidotyri cheese samples. After merging quality filtering, chimera removal and normalization, a total of 1,116,188 high-quality sequences were obtained, with an average of 32,829 reads per sample (range 10,425 to 55,131). In total, 222 OTUs were observed.
Alpha diversity metrics, namely Shannon and Simpson diversity, were calculated; however, we considered effective richness as a more accurate measurement of bacterial diversity between the samples, as this index is not affected by sequence depth or normalization steps and takes into account bacterial taxa with a relative abundance over 0.25% for each sample [45]. In this study, both Artisanal_CS and Artisanal_IHS Gidotyri cheeses had increased effective richness compared to Ind. Gidotyri cheeses (p < 0.05, Figure 2). Increased microbial diversity has also been observed in previous studies comparing the microbiota of artisanal-and Ind.-type cheeses [27,40,46]. This finding further confirms the assumption that traditional cheese-making practices are directly linked to a more diverse cheese microbial profile.
Increased microbial diversity has also been observed in previous studies comparing the microbiota of artisanal-and Ind.-type cheeses [27,40,46]. This finding further confirms the assumption that traditional cheese-making practices are directly linked to a more diverse cheese microbial profile.

Beta Diversity
We calculated beta diversity to evaluate the similarity of the microbial profiles between the three Gidotyri cheese groups. Different microbial communities were revealed between all three cheese groups (Permanova p < 0.05) demonstrated by the three separated clusters ( Figure 3). Similarly, classification of industrial and homemade Feta cheese samples to separate clusters has previously been observed [27]. It is worth noting that, both in the current and the previously mentioned studies, the artisanal cluster was characterized by a higher dispersal of the samples, indicative of the higher variation among the microbial profile of the respective samples. Regarding the increased variability that was evident in the Artisanal_CS cluster, this can be explained by the fact that milk used for each cheese sample was supplied by a different goat farm. Contrarily, Artisanal_IHS cheese samples made from milk produced by a single goat farm formed a tighter cluster. The influence of the raw milk microbiota on the cheese microbial profile has already been extensively reviewed [3,47,48], supporting our observations.

Beta Diversity
We calculated beta diversity to evaluate the similarity of the microbial profiles between the three Gidotyri cheese groups. Different microbial communities were revealed between all three cheese groups (Permanova p < 0.05) demonstrated by the three separated clusters ( Figure 3). Similarly, classification of industrial and homemade Feta cheese samples to separate clusters has previously been observed [27]. It is worth noting that, both in the current and the previously mentioned studies, the artisanal cluster was characterized by a higher dispersal of the samples, indicative of the higher variation among the microbial profile of the respective samples. Regarding the increased variability that was evident in the Artisanal_CS cluster, this can be explained by the fact that milk used for each cheese sample was supplied by a different goat farm. Contrarily, Artisanal_IHS cheese samples made from milk produced by a single goat farm formed a tighter cluster. The influence of the raw milk microbiota on the cheese microbial profile has already been extensively reviewed [3,47,48], supporting our observations.

Microbiota Diversity in Gidotyri Cheese Samples
This is the first study that utilized high throughput sequencing to achieve a more indepth characterization of the microbiota in Artisanal and Industrial Gidotyri cheeses. Four phyla were present in all the samples of the three cheese groups with Firmicutes being predominant. In particular, higher abundance was observed in Ind. Gidotyri cheese samples (99.51%) followed by Artisanal_IHS (99.03%) and Artisanal_CS (88.10%) Gidotyri cheese samples with all being significantly different from each other (p < 0.05). Proteobac-

Microbiota Diversity in Gidotyri Cheese Samples
This is the first study that utilized high throughput sequencing to achieve a more in-depth characterization of the microbiota in Artisanal and Industrial Gidotyri cheeses. Four phyla were present in all the samples of the three cheese groups with Firmicutes being predominant. In particular, higher abundance was observed in Ind. Gidotyri cheese samples (99.51%) followed by Artisanal_IHS (99.03%) and Artisanal_CS (88.10%) Gidotyri cheese samples with all being significantly different from each other (p < 0.05). Proteobacteria had a higher abundance in Artisanal_CS Gidotyri cheese samples (11.74%) compared to Artisanal_IHS (0.66%) and Ind. Gidotyri (0.43%) cheese samples (p > 0.05), while Bacteroidota and Actinobacteriota were present at <0.3% in all three Gidotyri cheese groups. The observed phyla with the predominance of Firmicutes has been observed in a variety of cheese types [4,23,38,46,49,50].
In the present study, we focus on the families with differences in the relative abundance between the three cheese groups; however, the complete list of families identified are presented in Figure S1 of the Supplementary File. Within the Firmicutes phylum, three families were significantly different between the three Gidotyri cheese groups ( Figure 4A). Streptococcaceae had a higher abundance in Ind. Gidotyri cheeses (90.96%) compared to the other two cheese groups (73.71% for Artisanal_IHS and 63.85% for Artisanal_CS) (p < 0.05). Lactobacillaceae had a higher abundance in Artisanal_IHS (25.25%) compared to the Ind. Gidotyri cheeses (7.63%) (p < 0.05). The relative abundance of Lactobacillaceae in Artisanal_CS cheeses (20.48%) was numerically higher than Ind. Gidotyri cheeses (p > 0.05) and closer to the Artisanal_IHS. Based on our finding, the microbiota of Ind. Gidotyri cheese is less complex with the Streptococcaceae family representing >90% of the bacterial taxa. Contrarily, Artisanal Gidotyri cheeses include a significant percentage of the Lactobacillaceae family as well. Our findings resemble the ones reported by Samelis and Kakouri [40] that also observed a dominance of members of the Streptococcaceae in industrial galotyri cheeses, while Lactobacillaceae members were more prevalent in artisanal galotyri cheeses. The dominance of Streptococaceae followed by Lactobacillaceae in artisanal goat cheese was also reported in a recent study [51]. Carnobacteriaceae were solely present in Artisanal_CS despite being identified in a single Ind. Gidotyri cheese sample (p < 0.05). This family has been considered among the families involved in the acidification of milk during cheese production [3] and has been associated with anti-listerial activity in smear-ripened cheeses [52]. Within the Proteobacteria phylum, two families, namely Enterobacteriaceae and Moraxellaceae, were solely present in the Artisanal_CS despite the former being observed in a single Ind. Gidotyri cheese sample (p < 0.05, Figure 4B). These families are commonly found in raw milk and are considered markers of the hygiene conditions during the cheese production process [37,38,53]. There was also a tendency for increased prevalence of Hafniaceae in the Artisanal_CS compared to the other two cheese groups (p < 0.10, Figure 4B).
The thirty-five most abundant genera in all three Gidotyri cheese groups are presented in Figure 5, while the complete list of genera identified are given in Figure S2 of the Supplementary File. Lactococcus and Streptococcus represented the two major genera in all three cheese groups with relative abundance 43.4% and 30.6%, respectively, followed by Lactiplantibacillus (8.9%), Secundilactobacillus (4.4%) and Lactobacillus (2.9%), with the remaining genera being present at <1.6% in accordance with observations from previous studies on different cheese types with variations in the abundances of these genera [23,27,39,46,49,51]. Five genera were found to be significantly different in terms of prevalence (Fisher's test) between the three cheese groups. Acinetobacter, Carnobacterium and Enterobacter had higher prevalence in Artisanal_CS cheeses (7 out of 15; 10 out of 15; 7 out of 15 samples) compared to Artisanal_IHS cheeses (0 out of 10 for all genera, p < 0.05) and Ind. cheeses (0 out of 9; 1 out of 9; 0 out of 9 samples, p < 0.05). Loigolactobacillus and Secundilactobacillus had higher prevalence in Artisanal_IHS cheeses (9 out of 10; 10 out of 10 samples) compared to Artisanal_CS cheeses (2 out of 15; 3 out of 15, p < 0.001) and Ind. cheeses (0 out of 9; 2 out of 9 samples, p < 0.001). In the present study, the representative species of each genus that dominated in the respective Gidotyri cheese group was identified and will be discussed in the subsequent section.
families, namely Enterobacteriaceae and Moraxellaceae, were solely present in the Arti-sanal_CS despite the former being observed in a single Ind. Gidotyri cheese sample (p < 0.05, Figure 4B). These families are commonly found in raw milk and are considered markers of the hygiene conditions during the cheese production process [37,38,53]. There was also a tendency for increased prevalence of Hafniaceae in the Artisanal_CS compared to the other two cheese groups (p < 0.10, Figure 4B). and Industrial (Ind.) Gidotyri cheeses. The red bold lines represent the median, while symbolGidotyri* indicates statistically significant differences in abundance between cheese groups (Mann-Whitney U statistical test) with the number of stars representing the level of significance (< 0.05, * < 0.01, ** p< 0.01). Symbol (+) indicates statistically significant differences in prevalence between cheese groups (Fisher's statistical test) with the number of crosses representing the level of significance (+ < 0.05, + + < 0.01).
The thirty-five most abundant genera in all three Gidotyri cheese groups are presented in Figure 5, while the complete list of genera identified are given in Figure S2 of the Supplementary File. Lactococcus and Streptococcus represented the two major genera in all three cheese groups with relative abundance 43.4% and 30.6%, respectively, followed by Lactiplantibacillus (8.9%), Secundilactobacillus (4.4%) and Lactobacillus (2.9%), with the remaining genera being present at <1.6% in accordance with observations from previous studies on different cheese types with variations in the abundances of these genera [23,27,39,46,49,51]. Five genera were found to be significantly different in terms of prevalence (Fisher's test) between the three cheese groups. Acinetobacter, Carnobacterium and Enterobacter had higher prevalence in Artisanal_CS cheeses (7 out of 15; 10 out of 15; 7 out of 15 samples) compared to Artisanal_IHS cheeses (0 out of 10 for all genera, p < 0.05) and . Differential abundance or prevalence of selected bacterial families belonging to Lactobacillales (A) and Enterobacterales (B) between the Artisanal_CS (Art. CS), Artisanal_IHS (Art. IHS) and Industrial (Ind.) Gidotyri cheeses. The red bold lines represent the median, while symbolGidotyri* indicates statistically significant differences in abundance between cheese groups (Mann-Whitney U statistical test) with the number of stars representing the level of significance (<0.05, * <0.01, ** p < 0.01). Symbol (+) indicates statistically significant differences in prevalence between cheese groups (Fisher's statistical test) with the number of crosses representing the level of significance (+ < 0.05, + + < 0.01).

The Most Prevalent Bacterial Species in Artisanal-and Industrial-Type Gidotyri Cheese
The relative abundance of the bacterial species present in the three Gidotyri cheeses was additionally determined with the 29 most abundant based on their average relative abundance across all cheese samples from all cheese groups being presented in Table 1, along with their respective roles in cheese production and quality. Eight OTUs corresponding to eight different bacterial species were found to be significantly different in terms of prevalence (Fisher's test) between the three cheese groups. Loigolactobacillus coryniformis (OTU 175), Secundilactobacillus malefermentans (OTU 48), and Streptococcus parauberis (OTU 50) had higher prevalence in Artisanal_IHS cheeses (9 out of 10; 10 out of 10; 9 out of 10 samples) compared to Artisanal_CS cheeses (2 out of 15; 3 out of 15; 2 out of 15 samples, p ≤ 0.001) and Ind. cheeses (0 out of 9; 2 out of 9; 1 out of 9 samples, p < 0.01). L. coryniformis has exhibited antibacterial activity against pathogens and spoilage microorganisms associated with acid, H 2 O 2 and bacteriocin production [54,55], while S. malefermentans, a bacterium exclusively fermenting carbohydrates at low temperature, was recently identified as a core member of sauerkraut (fermented food product) carrying genes encoding for enzymes with significant contribution to the aroma development of the final product [56,57]. Str. parauberis is associated with small ruminant mastitis and has been identified as a minor component of cheeses produced with their milk [38,58,59]. Other mastitis-causing pathogens can be found in cheese via contaminated milk [60][61][62][63]. Carnobacterium maltaromaticum (OTU 23) and Enterobacter hormaechei subsp. hoffmannii (OTU 268) had higher prevalence in Artisanal_CS cheeses (10 out of 15; 7 out of 15 samples) compared to Artisanal_IHS cheeses (0 out of 10; 0 out of 10 samples, p < 0.05) and Ind. cheeses (1 out of 9; 0 out of 9 samples, p < 0.05). C. maltaromaticum comprises an important member of the cheese ripening microflora due to its contribution to aroma development, control of spoilage bacteria and anti-listerial bacteriocin production [64][65][66]. The role of E. hormaechei, an isolate from several artisanal sheep cheeses, is a bit controversial as it is considered both a poor hygiene indicator of cheese production and a potential contributing bacterium to cheese flavor [61,67,68]. A tendency for higher prevalence of Hafnia alvei (OTU 13) and Acinetobacter colistiniresistens (OTU 111) was additionally observed in Artisanal_CS compared to the other two cheese groups (p < 0.10). H. alvei is a frequent member of the microbiota in traditional cheeses possibly related to distinct organoleptic properties while also displaying antibacterial activity against foodborne pathogens [69]. A. colistiniresistens, a bacterial species with intrinsic resistance to polymyxins, has not been previously isolated in cheese or other food products [70]. Concerning the two Artisanal cheeses, Streptococcus thermophilus (OTU 3) was more prevalent in the Artisanal_IHS (10 out of 10 samples) cheese samples compared to Artisanal_CS (7 out of 15 samples, p < 0.05). This bacterial species is a widely used starter culture in cheese and other dairy products with well-known technological properties [71]. Ind. cheeses (0 out of 9; 1 out of 9; 0 out of 9 samples, p < 0.05). Loigolactobacillus and Secundilactobacillus had higher prevalence in Artisanal_IHS cheeses (9 out of 10; 10 out of 10 samples) compared to Artisanal_CS cheeses (2 out of 15; 3 out of 15, p < 0.001) and Ind. cheeses (0 out of 9; 2 out of 9 samples, p < 0.001). In the present study, the representative species of each genus that dominated in the respective Gidotyri cheese group was identified and will be discussed in the subsequent section.

The Most Prevalent Bacterial Species in Artisanal-and Industrial-Type Gidotyri Cheese
The relative abundance of the bacterial species present in the three Gidotyri cheeses was additionally determined with the 29 most abundant based on their average relative abundance across all cheese samples from all cheese groups being presented in Table 1, along with their respective roles in cheese production and quality. Eight OTUs corre- As a final point of this study, it is worth mentioning that the type of heat treatment implemented on milk influences the microbial composition of the produced cheese [48,72,73]. Therefore, the more diverse microbiota in artisanal Gidotyri cheeses can probably be attributed to the thermization of the raw milk (63 • C for 15 min) instead of the industrial practice of pasteurization (73 • C for 15 s). We assume that the higher prevalence of bacterial populations used as poor hygiene indicators in Artisanal_CS is likely associated with milk being obtained from multiple sources, namely 15 different goat farms, in contrast with Arti-sanal_IHS which was solely made from milk of a single goat farm. Based on our findings, specific bacterial species related to artisanal Gidotyri cheeses were detected, indicating the probability of linking traditional cheese-making practices to distinct microbial markers that could be used as traceability models to ensure their authenticity.

Conclusions
In this study, the microbiota of Gidotyri, a traditional Greek goat cheese, was determined for the first time using HTS methodology. Furthermore, the impact of the cheese-making practices, industrial-type versus artisanal-type, was examined. It was clearly demonstrated that artisanal Gidotyri cheeses were characterized by a more complex microbiota with specific families, genera and species linked to each subtype, namely Ar-tisanal_CS and Artisanal_IHS. Contrarily, Ind. Gidotyri cheeses were dominated by the Streptococcaceae family. This initial screening study provides evidence on the presence of possible microbial markers which could serve as authenticity signatures of the manufacturing processes while also possessing distinct technological and beneficial properties that merit further investigation.  Data Availability Statement: Primary sequencing data were uploaded to ENA public repository with the accession number PRJEB58122.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.