Unraveling the Microbiota of Natural Black Cv

microorganisms

Article Unraveling the Microbiota of Natural Black cv. Kalamata Fermented Olives through 16S and ITS Metataxonomic Analysis

Maria Kazou 1,*, Aikaterini Tzamourani 2, Efstathios Z. Panagou 2,* and Eﬃe Tsakalidou 1 1 Laboratory of Dairy Research, Department of Food Science and Human Nutrition, School of Food and Nutritional Sciences, Agricultural University of Athens, Iera Odos 75, 11855 Athens, Greece; [email protected] 2 Laboratory of Food Microbiology and Biotechnology, Department of Food Science and Human Nutrition, School of Food and Nutritional Sciences, Agricultural University of Athens, Iera Odos 75, 11855 Athens, Greece; [email protected] * Correspondence: [email protected] (M.K.); [email protected] (E.Z.P.); Tel.: +30-210-529-4628 (M.K.); +30-210-529-4693 (E.Z.P.) Received: 17 April 2020; Accepted: 4 May 2020; Published: 6 May 2020

Abstract: Kalamata natural black olives are one of the most economically important Greek varieties. The microbial ecology of table olives is highly influenced by the co-existence of bacteria and yeasts/fungi, as well as the physicochemical parameters throughout the fermentation. Therefore, the aim of this study was the identification of bacterial and yeast/fungal microbiota of both olives and brines obtained from 29 cv. Kalamata olive samples industrially fermented in the two main producing geographical regions of Greece, namely Aitoloakarnania and Messinia/Lakonia. The potential microbial biogeography association between certain taxa and geographical area was also assessed. The dominant bacterial family identified in olive and brine samples from both regions was Lactobacillaceae, presenting, however, higher average abundances in the samples from Aitoloakarnania compared to Messinia/Lakonia. At the genus level, Lactobacillus, Celerinatantimonas, Propionibacterium and Pseudomonas were the most abundant. In addition, the yeasts/fungal communities were less diverse compared to those of bacteria, with Pichiaceae being the dominant family and Pichia, Ogataea, and Saccharomyces being the most abundant genera. To the best of our knowledge, this is the first report on the microbiota of both olives and brines of cv. Kalamata black olives fermented on an industrial scale between two geographical regions of Greece using metagenomics analysis.

Keywords: table olives; Kalamata olives; Greek style fermentation; microbiological analysis; 16S metagenomics analysis; ITS metagenomics analysis; microbiota; high-throughput sequencing

1. Introduction Table olives are one of the oldest and most popular fermented vegetables in the Mediterranean basin that has a signiﬁcant contribution in the world production of fermented olives, with Spain, Greece, Italy and Portugal being among the main EU table olive producers (ca. 35% of world production). The renewed interest of consumers for high-quality food products with enhanced nutritional properties, such as bioactive compounds, dietary ﬁbers, fatty acids and antioxidants, resulted in an increasing trend in the production of processed olives in recent years, which is expected to reach 2.9 million tons in the 2019/2020 season [1]. Table olive processing in Greece is of paramount socio-economic importance for the country. According to a recent report by the Greek Interprofessional Association for Table Olives [2], more than 64,000 farmers and 100 businesses are directly involved in the primary and secondary production and processing sectors, respectively. The annual amount of the end (processed) product has increased

Microorganisms 2020, 8, 672; doi:10.3390/microorganisms8050672 www.mdpi.com/journal/microorganisms Microorganisms 2020, 8, 672 2 of 24 to 215,000 tons, from which 85% are exported, rendering the country as the second largest world exporter of table olives, right after Spain. The value of the market size of Greek table olive exports exceeds 450 million euros representing 9.2% of the overall exports of Greek agricultural commodities. Several trade preparations of Greek table olives are renowned brand names in the international market, including Halkidiki green olives, as well as Kalamata and Conservolea (or Amfissis) natural black olives. Although all trade preparations of table olives (i.e., treated, natural, dehydrated and/or shriveled olives, and olives darkened by oxidation) [3] are produced in the country, Greece has a long tradition in the production of natural black olives, which are internationally known as “Greek style olives in brine”, using olives from the two most economically important varieties, namely Conservolea and Kalamata [4]. In particular, the latter variety is highly esteemed for its exceptional organoleptic characteristics, the dark color and crisp texture of the end product increasing, therefore, the demand and the exports for Kalamata black olives. Nowadays, this variety is cultivated primarily in the prefectures of Aitoloakarnania (Western Greece), Messinia and Lakonia (Southern Peloponnese) and Fthiotida (Mainland Greece), but new olive trees are being planted every year in other areas as well. It is expected that, within the next 10 years, the production of Kalamata natural black olives will exceed 100,000 tons [2]. During Greek-style processing, olives are initially subjected to quality inspection to remove defective drupes. Subsequently, olives are immersed directly in a brine solution of about 6–10% (w/v) salt concentration where they undergo spontaneous fermentation for 8–12 months. The microbiota of table olives fermentation is a complex set of dynamics, in which the co-existence of lactic acid bacteria (LAB) and yeasts is of fundamental importance to obtain high quality products [5,6]. Until recently, the microbiota of table olive fermentation has been described mostly by culture-dependent approaches. However, it should be stressed out that almost 90% (or 99% according to other researchers) of microorganisms present on natural environments cannot be cultivated in synthetic microbiological media [7–10]. To overcome these limitations, culture-independent analyses and, most importantly, next generation sequencing (NGS) intensely changed the study of the microbial ecology of fermented foods [11–13]. Nowadays, high throughput sequencing (HTS) coupled with omics approaches are used for the characterization of food microbiota. Among them, amplicon-based metagenomics analysis is the most widely used. It should be noted that the majority of culture-independent approaches for the assessment of the microbial diversity in table olive fermentations have been applied on green olive fermentations, either by the Spanish method or directly brined [14–25]. However, limited information can be found in the literature for the microbial diversity of natural black olive fermentations using culture-independent approaches, especially for Greek table olive varieties [26]. The aim of this study was to unravel the microbial diversity of fermented natural black olives from the Kalamata variety using the state-of-the-art approach of amplicon-based metagenomics analysis. The samples were taken directly from fermentation vessels of processing installations located in different locations from two geographical regions in Greece, namely Aitoloakarnania and Messinia/Lakonia, where the main volume of this variety is cultivated and processed. To the best of our knowledge, this is the first study to elucidate the microbial community diversity of cv. Kalamata olives processed on industrial scale according to the traditional Greek-style method using the metagenomic approach.

2. Materials and Methods

2.1. Origin of the Samples Twenty-nine (29) samples of fermented cv. Kalamata (Olea europea var. ceraticarpa) natural black olives were obtained during the 2018–2019 season from two geographical areas in Greece, namely Aitoloakarnania in western Greece and Messinia/Lakonia in southern Peloponnese (Figure1). The samples (3 kg each) were taken in coincidence with the ﬁnal stage of black olive fermentation as traditionally implemented in the two regions. Overall, 14 samples were taken from the area of MicroorganismsMicroorganisms2020 20,208,, 6728, x FOR PEER REVIEW 3 of3 24 of 24

93 samples (3 kg each) were taken in coincidence with the final stage of black olive fermentation as / 94 Aitoloakarnaniatraditionally implemented and 15 samples in the from two the regions. area of Overall, Messinia 14 samplesLakonia were (Table taken1). Fermentation from the area was of 95 undertakenAitoloakarnania in 5000-6000-L and 15 sampl capacityes from polyester the area vessels of Messinia/Lakonia containing 60–70% (Table olives 1). Fermentation and 30–40% was brine. 96 Olivesundertaken were processed in 5000-6000 according-L capacity to traditionalpolyester vessels Greek-style containing anaerobic 60–70% processing. olives and 30 Speciﬁcally,–40% brine. cv. 97 KalamataOlives were black processed olive processing according included to traditional harvesting Greek at-style the anaerobic right stage processing. of ripening Specifically, (i.e., 3/4of cv. the 98 mesocarpKalamata has black attained olive blackprocessing color), included transportation harvesting to at the the processing right stage installations, of ripening (i.e., quality 3/4 of control the 99 inspectionmesocarp for has the attained removal black of defected color), transportation drupes and to ﬁnally the processing brining in installations, 5.0–6.0% (w quality/v) NaCl, control where 100 theinspection olives were for subjectedthe removal to of spontaneous defected drupes fermentation and finally by brining the autochthonous in 5.0–6.0% (w/v bacterial) NaCl, where and fungal the 101 microbiota.olives were During subjected fermentation, to spontaneous salt level fermentation was adjusted by the to autochthonous the initial concentration bacterial and by periodicfungal 102 additionsmicrobiota. of coarse During salt fermentation, at the top of eachsalt level fermentation was adjusted vessel to to the facilitate initial theconcentration dominance by of periodic LAB over 103 yeasts.additions In late of Maycoarse/early salt at June, the top depending of each fermentation on ambient vessel temperature, to facilitate salt the was dominance added to of the LAB vessels over to 104 adjustyeasts. salinity In late to May/early 7.0–8.0% toJune, avoid depending spoilage on due ambient to the upcomingtemperature, high salt summer was added temperatures to the vessels [27 to]. 105 adjust salinity to 7.0–8.0% to avoid spoilage due to the upcoming high summer temperatures [27].

106 107 Figure 1. Geographical origin of fermented cv. Kalamata natural black olives from (A) Figure 1. Geographical origin of fermented cv. Kalamata natural black olives from (A) Aitoloakarnania 108 Aitoloakarnania (western Greece) and (B) Messinia/Lakonia (southern Peloponnese). (western Greece) and (B) Messinia/Lakonia (southern Peloponnese).

109 Table 1. Sample code and geographical origin of the fermented cv. Kalamata natural black olive 110 samples.

Sample Code Sample Origin Sample Code Sample Origin 1 Aetofolia/Messinia 16 Messolongi/Aitoloakarnania

Microorganisms 2020, 8, 672 4 of 24

Table1. Sample code and geographical origin of the fermented cv. Kalamata natural black olive samples.

Sample Code Sample Origin Sample Code Sample Origin 1 Aetofolia/Messinia 16 Messolongi/Aitoloakarnania 2 Geraki/Lakonia 17 Aitoliko/Aitoloakarnania 3 Manesis/Messinia 18 Stamna/Aitoloakarnania 4 Velika/Messinia 19 Kefalovriso/Aitoloakarnania 5 Geraki/Lakonia 20 Chrysovergi/Aitoloakarnania 6 Messene/Messinia 21 Kainourgio/Aitoloakarnania 7 Trikorfo/Messinia 22 Aitoliko/Aitoloakarnania 8 Aitoliko/Aitoloakarnania 23 Mastro/Aitoloakarnania 9 Neochori/Aitoloakarnania 24 Neochori/Aitoloakarnania 10 Androusa/Messinia 25 Neochori/Aitoloakarnania 11 Klima/Messinia 26 Stamna/Aitoloakarnania 12 Meligalas/Messinia 27 Chrisafa/Lakonia 13 Meropi/Messinia 28 Geraki/Lakonia 14 Klada/Messinia 29 Trikorfo/Messinia 15 Messolongi/Aitoloakarnania

2.2. Microbiological Analysis Classical microbiological analysis was performed in both olive and brine samples to enumerate the main microbial groups implicated in table olive fermentation, i.e., LAB, yeasts and enterobacteria [28]. For this purpose, olive (25 g) or brine (1 mL) samples were aseptically transferred into 225 mL or 9 mL sterile Ringer’s solution (0.9% (w/v) NaCl), respectively. Olive samples were homogenized in a stomacher device (LabBlender, Seward Medical, London, UK) for 60 s at room temperature. The resulting suspensions were serially diluted in the same diluent and duplicate 1 or 0.1 mL of the appropriate dilutions were poured or spread on the following agar media to enumerate: (i) LAB on de Man–Rogosa–Sharpe (MRS) medium (Biolife, Milan, Italy, pH adjusted to 5.7) containing 0.05% ◦ (w/v) cycloheximide (AppliChem GmbH, Darmstadt, Germany) and incubated at 25 C for 72 h under anaerobic conditions (double agar layer); (ii) yeasts and molds on Rose Bengal Chloramphenicol ◦ agar (RBC; supplemented with selective supplement X009, Bury, UK), incubated for 48 h at 25 C; and ﬁnally (iii) Enterobacteriaceae on Violet Red Bile Glucose agar (VRBGA; Biolife, Milan, Italy), ◦ incubated under anaerobic conditions (double agar layer) for 24 h at 37 C. Thereafter, agar media containing 30–300 colonies were used for the enumeration of microbial counts and 20% of colonies present in the plates were randomly selected and examined for typical morphological traits related with each medium. The results were log transformed and expressed as log CFU/g or mL of olives or brine, respectively.

2.3. pH and Salt Measurement After sampling, the pH was measured in both brine and olive samples using a digital pHmeter (model RL150, Russel Inc., Boston, MA, USA). The pH of the brine was recorded by immersing the electrode of the instrument directly in the brine sample, whereas the pH of olives was measured in a ﬂuid slurry that was prepared by homogenizing 50 g of depitted olives in 100 mL of distilled water using an Ultra Turrax T25 blender (IKA, Staufen, Germany) [29]. Salt (sodium chloride) determinations in the brines were carried out by titration as described elsewhere [30]. The results are expressed as a percentage (w/v) of NaCl.

2.4. Total DNA Extraction Microbial DNA from olive samples was extracted based on the protocol of Medina et al., slightly modified [22]. In detail, 12 g of olive sample were homogenized with 30 mL of sterile Ringer’s solution ◦ in a stomacher apparatus for 60 s at room temperature. After centrifugation (10,000 g/10 min/20 C), the pellet was washed twice with 30 mL of phosphate-buffered saline (PBS) pH 7.4, and the DNA was Microorganisms 2020, 8, 672 5 of 24 extracted using the DNeasy PowerSoil Kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s ◦ instructions. DNA was eluted in 30 µL of preheated (70 C) DNA-free PCR-grade water and stored ◦ at −20 C until further analysis. In addition, microbial DNA from brine samples was extracted according to a combined protocol of Pitcher et al. [31] and Kopsahelis et al. [32]. In brief, 10 mL ◦ of brine sample was centrifuged (10,000 rpm/10 min/20 C) and the pellet was washed twice with ◦ 1 mL of PBS. After centrifugation (10,000 g/10 min/20 C), the pellet was resuspended in 1 mL of ◦ PBS and incubated at 65 C for 10 min to decrease the content of PCR inhibitors. The sample was centrifuged under the same conditions, the supernatant was removed, and then 500 µL of freshly prepared lysozyme (Sigma-Aldrich Chemie Gmbh, Munich, Germany) (50 mg/mL) in Tris-EDTA (TE) buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0), 100 µL RNAse A (Sigma-Aldrich) (10 mg/mL), 40 µL mutanolysin (Sigma-Aldrich) (5 U/µL), 500 µL of 1 M sorbitol, 0.1 M EDTA (pH 7.5) and 200 µL lyticase (Sigma-Aldrich) (1 U/µL) were added to the pellet and vortexed briefly until it was completely ◦ dissolved. The suspension was incubated at 37 C for 3 h, and, subsequently, 20 µL proteinase K ◦ (Sigma-Aldrich) (25 mg/mL) was added and an incubation at 55 C for 1 h followed. The lysate was ◦ centrifuged at 10,000 g/10 min/20 C, the pellet was resuspended in 0.5 mL of 50 mM Tris-Cl (pH 7.4), ◦ 20 mM EDTA and 50 µL of 20% v/v SDS and the mixture was incubated at 65 C for 30 min. Afterwards, 250 µL cold ammonium acetate (7.5 mol/L) was added and the tube was held on ice for 1 h. After ◦ centrifugation (10,000 g/10 min/4 C), 1 mL of the supernatant was transferred to a new Eppendorf ◦ tube, 1 volume of cold isopropanol was added and the tube was kept overnight at -20 C. The next ◦ day, the fibrous DNA was pelleted by centrifugation (10,000 g/20 min/4 C), washed twice with 700 µL ice-cold ethanol (70% v/v) and, after the complete removal of ethanol, the pellet was resuspended in ◦ 30–50 µL of TE buffer (pH 8.0) and stored at −20 C until use. DNA concentration was measured using a Quawell Q5000 Read First photometer (Quawell Technology Inc, San Jose, CA, USA), and DNA quality was checked according to Papademas et al. [33].

2.5. Amplicon-Based Metagenomics Analysis ® Amplicon sequencing (bTEFAP ) was performed on the Illumina MiSeq at Molecular Research DNA (MR DNA, Shallowater, Texas) in order to evaluate the bacterial and yeast/fungal microbiota. 0 0 0 Therefore, primers 27F (5 -AGR GTT TGA TCM TGG CTC AG-3 ) and 519R (5 -GTN TTA CNG CGG 0 CKG CTG-3 ) were used for the amplification of the V1-V3 hypervariable region of the bacterial 16S 0 0 0 rRNA gene, as well as primers ITS1F (5 -CTT GGT CAT TTA GAG GAA GTA A-3 ) and ITS2R (5 -GCT 0 GCG TTC TTC ATC GAT GC-3 ), to amplify the yeast/fungal internal transcribed spacer (ITS) DNA region, namely ITS1-ITS2. The PCR conditions and purification of amplicon products were performed according to Papademas et al. [33]. Operational taxonomic units (OTUs) were defined after the removal of singleton sequences, clustered at 3% divergence (97% similarity) and taxonomically assigned using the Nucleotide Basic Local Alignment Search Tool (BLASTn) against a curated National Center for Biotechnology Information (NCBI) deriving database [34]. Subsequently, clustered OTUs were used to construct the rarefaction curves in order to assess species richness and estimate the sequencing depth. Finally, alpha-diversity indices, namely observed species, Shannon and inverse Simpson were used to evaluate microbial diversity based on richness and evenness [35]. Raw sequencing data are deposited at the European Nucleotide Archive (ENA) under the study ID PRJEB37064.

2.6. Statistics and Exploratory Data Analysis One-way analysis of variance (ANOVA) followed by post hoc comparisons with Tukey’s test was undertaken for the determination of differences in microbial enumerations and pH between fermented table olives from the two regions. In addition, differences between salt concentration in the brine samples from the two regions were determined using an unpaired t-test. In all cases, p < 0.05 was considered to be statistically significant. Data analysis was performed using the GraphPad Prism ver. 5.0 software (GraphPad Software, San Diego, CA, USA). In addition, Hierarchical cluster analysis (HCA) was undertaken for the unsupervised discrimination of the characterized microbiota (bacteria Microorganisms 2020, 8, 672 6 of 24

and yeasts) from the brine and olive environment based on the two geographical sampling regions (i.e., Aitoloakarnania and Messinia/Lakonia). As input matrix, the bacterial or fungal OTU table with a Microorganisms 2020, 8, x FOR PEER REVIEW 6 of 24 relative abundances higher than 1% was used, for olive or brine samples. HCA was performed based 197 ongraphically Euclidean illustrated distance and in Ward’sthe form linkage of heatmaps. as similarity Heatmaps measure are andcommonly clustering used algorithm, in bioinformatics respectively. as a 198 Datatwo- weredimensional transformed visualization by autoscaling technique, before where analysis data and are the arranged results in were rows graphically and columns illustrated so that 199 insimilar the form columns of heatmaps. are grouped Heatmaps together are commonly with their used similarity in bioinformatics presented by as a two-dimensional dendrogram [36]. 200 visualizationFurthermore, technique, a supervised where classification data are arranged technique, in rows namely and columns Partial so Least that Squares similar columns Discriminant are 201 groupedAnalysis together (PLS-DA) with, was their also similarity used in this presented work. byPLS a-DA dendrogram establishes [36 a]. linear Furthermore, regression a supervisedbetween the 202 classiﬁcationX matrix of technique,independent namely variables Partial (in Least our case Squares OTUs Discriminant derived from Analysis NGS (PLS-DA),analysis) and was matrix also used Y of 203 independent this work. variables PLS-DA establishes(classes). HCA a linear and regressionPLS-DA were between conduc theted X using matrix MetaboAnalyst of independent 4.0 variables [37]. (in our case OTUs derived from NGS analysis) and matrix Y of dependent variables (classes). HCA and 204 PLS-DA3. Results were conducted using MetaboAnalyst 4.0 [37].

205 3.3.1. Results Differences in Microbial Population, pH Values and Salt Concentration According to the Region of 206 Origin 3.1. Differences in Microbial Population, pH Values and Salt Concentration According to the Region of Origin 207 Out of the 29 samples of fermented cv. Kalamata natural black table olives examined, 14 samples Out of the 29 samples of fermented cv. Kalamata natural black table olives examined, 14 samples 208 originated from the area of Aitoloakarnania in western Greece and 15 samples were from the area of originated from the area of Aitoloakarnania in western Greece and 15 samples were from the area of 209 Messinia/Lakonia in southern Peloponnese. Differences in LAB populations were found according to Messinia/Lakonia in southern Peloponnese. Differences in LAB populations were found according to the 210 the geographical origin of the samples. Specifically, the average value of LAB counts of olives geographical origin of the samples. Specifically, the average value of LAB counts of olives originating 211 originating from Messinia/Lakonia was 3.27  0.64 log CFU/g that was significantly lower (p < 0.05) from Messinia/Lakonia was 3.27 0.64 log CFU/g that was significantly lower (p < 0.05) compared 212 compared to LAB counts of olives from Aitoloakarnania, which equaled to 5.98  0.43 log CFU/g to LAB counts of olives from Aitoloakarnania, which equaled to 5.98 0.43 log CFU/g (Figure2A). 213 (Figure 2A). A similar trend was obtained for LAB enumerated in the brine samples. In this case, the A similar trend was obtained for LAB enumerated in the brine samples. In this case, the average 214 average LAB counts from brine samples from Messinia/Lakonia, amounting to 3.85  0.66 log LAB counts from brine samples from Messinia/Lakonia, amounting to 3.85 0.66 log CFU/mL, 215 CFU/mL, was significantly lower (p < 0.05) than those of Aitoloakarnania (6.77  0.36 log CFU/mL). was significantly lower (p < 0.05) than those of Aitoloakarnania (6.77 0.36 log CFU/mL). It should 216 It should be noted though that, in four samples from Messinia/Lakonia and one sample from be noted though that, in four samples from Messinia/Lakonia and one sample from Aitoloakarnania, 217 Aitoloakarnania, LAB could not be enumerated in the brine and/or olives (Figures S1 and S2). LAB could not be enumerated in the brine and/or olives (Figures S1 and S2).

218 Figure 2. Boxplots of lactic acid bacteria (LAB) (A) and yeast (B) counts as well as pH values (C) from 219 Figure 2. Boxplots of lactic acid bacteria (LAB) (A) and yeast (B) counts as well as pH values (C) from the olives and brines, along with salt concentration (D) from the brines of natural black cv. Kalamata 220 the olives and brines, along with salt concentration (D) from the brines of natural black cv. Kalamata olives from the two geographical regions. Different letters indicate statistically significant differences 221 olives from the two geographical regions. Different letters indicate statistically significant differences (p < 0.05). 222 (p < 0.05).

223 On the other hand, fewer differences were found for yeasts in Kalamata natural black olives. 224 Specifically, the average populations of yeasts for olive and brine samples originating from the area 225 of Messinia/Lakonia were 3.52  0.50 log CFU/g and 4.51 log CFU/mL, respectively, whereas lower 226 populations were found in the samples from Aitoloakarnania, i.e., 2.92  0.33 log CFU/g and 3.94 

Microorganisms 2020, 8, 672 7 of 24

On the other hand, fewer differences were found for yeasts in Kalamata natural black olives. Specifically, the average populations of yeasts for olive and brine samples originating from the area of Messinia/Lakonia were 3.52 0.50 log CFU/g and 4.51 log CFU/mL, respectively, whereas lower populations were found in the samples from Aitoloakarnania, i.e., 2.92 0.33 log CFU/g and 3.94 0.25 log CFU/mL. No statistical differences could be established among the olive and brine samples between the two areas, with the exception of olive samples from Aitoloakarnania that presented on average lower counts compared to brine samples from Messinia/Lakonia (Figure2B). It needs to be noted that, in two samples from Aitoloakarnania, no yeasts were detected on the olives (Figure S2). Finally, no enterobacteria were present in any of the analyzed olives and brine samples from both regions. Furthermore, the average values of pH in both brines and olives indicated successful lactic acid fermentation processes in the two regions. Brine samples from Aitoloakarnania and Messinia/Lakonia presented average pH values of 3.92 0.06 and 3.85 0.07, respectively. Slightly higher pH values were measured in olive samples from both areas, i.e., 4.14 0.06 (Aitoloakarnania) and 4.09 0.07 (Messinia/Lakonia) (Figure2C). The average pH values did not significantly di ffer between the two regions, with the exception of the average pH value of the brine samples from Messinia/Lakonia that was significantly lower (p < 0.05) from the average pH value of olive samples from Aitoloakarnania. The pH values in the brines and olives of the individual samples analyzed are illustrated in Figures S3 and S4. Given that the pH values in the brine of natural black olives should not exceed 4.30 according to the “Trade standard applying to table olives” of the IOC to ensure the safety of the final product, it can be concluded that no deviations were noticed. Moreover, the average salt concentration in the brines was 7.7% 0.28% and 7.5% 0.43% for the areas of Messinia/Lakonia and Aitoloakarnania, respectively, whereas no statistical differences could be noticed between the two geographical regions (Figure2D). The salt concentration of the individual samples of brine analyzed is shown in Figure S5. It must be mentioned that in 21 brine samples (13 from Messinia/Lakonia and eight from Aitoloakarnania), salt concentration was in the range of 7–8% or above, which is considered as an adequate salt concentration to prevent the spoilage of olives, especially in the summer period when high temperatures prevail.

3.2. Amplicon-Based Metagenomics Analysis A total of 3,188,642 bacterial raw sequences were obtained from the 58 samples (29 olives and 29 brine samples). After the removal of low-quality reads, 1,783,446 sequences (an average of 29,724 1530 sequences per sample) were used for metagenomic analysis. In total, 2133 bacterial OTUs were assigned among the samples, with an average of 317 131 and 271 58 OTUs per olive and brine sample, respectively. On the other hand, the number of yeasts/fungal raw sequences obtained from the 58 samples was lower, i.e., 2,742,360. However, the number of sequences used for metagenomic analysis after quality control was similar to that of bacterial sequences, i.e., 1,769,686 (an average of 29,495 1914 sequences per sample). Among the 58 samples, 1900 yeast/fungal OTUs were identiﬁed with an average of 303 125 and 348 204 OTUs per olive and brine sample, respectively. The microbial complexity was estimated on the basis of alpha-diversity indices, namely observed, Shannon and inverse Simpson. The richness estimation according to the Observed species indicated that the bacterial and yeast/fungal microbiota of olives from Messinia/Lakonia was signiﬁcantly higher (p < 0.05), and, according to Shannon and inverse Simpson indexes, more abundant (p < 0.05) than that of samples from the Aitoloakarnania region (Figure3A,C). This was also the case for yeast /fungal diversity (both richness and evenness) of brine samples from Messinia/Lakonia and Aitoloakarnania regions (Figure3D). On the contrary, observed species, Shannon and inverse Simpson index values for the bacterial microbiota of brine samples were relatively similar for both the Messinia/Lakonia and Aitoloakarnania regions (Figure3B). Microorganisms 2020, 8, 672 8 of 24 Microorganisms 2020, 8, x FOR PEER REVIEW 8 of 24

270 271 FigureFigure 3.3. BoxplotsBoxplots of alpha-diversityalpha-diversity indices, indices, namely namely o observed,bserved, Shannon Shannon and and inverse inverse Simpson Simpson for for 272 bacterialbacterial communitiescommunities in olive ( A)) and brine brine ( (BB)) samples, samples, as as well well as as for for yeast/fungal yeast/fungal microbiota microbiota in in 273 olivesolives ((CC)) andand brines brines ( D(D).). Samples Samples are are colored colored according according to to the the two two geographical geographical regions, regions, i.e., i.e. pink, pink for 274 Aitoloakarnaniafor Aitoloakarnania and and light light blue blue for Messiniafor Messinia/Lakonia/Lakonia. .

275 InIn addition,addition, thethe rarefaction curves revealed revealed the the differences differences of of bacterial bacterial and and yeasts/fungal yeasts/fungal species species 276 richnessrichness amongamong thethe samplessamples analyzed (Figure S6 S6).). In In details, details, rarefaction rarefaction curves curves of of both both 16S 16S and and ITS ITS 277 datadata ofof oliveolive andand brinebrine samples attained the saturatio saturationn plateau, plateau, indicating indicating that that the the sequencing sequencing depth depth 278 waswas susufficient.fficient. AsAs alsoalso revealedrevealed by the alpha-diversityalpha-diversity indices, indices, the the species species richness richness of of yeasts/fungal yeasts/fungal 279 microbiotamicrobiota of of both both olive olive and and brine brine samples samples (Figure (Figure S6C,D), S6C and,D), the and bacterial the bacterial microbiota microbiota of olive of samples olive 280 (Figuresamples S6A) (Figure was higherS6A) was in the higher samples in fromthe samples Messinia from/Lakonia Messinia/Lakonia compared to the compared Aitoloakarnania to the 281 region,Aitoloakarnania while the region,species whilerichness the of species bacteria richness communities of bacteria of brine communities samples was of brine similar samples between was the 282 twosimilar geographical between the regions two geographical (Figure S6B). regions (Figure S6B). 283 TheThe bacterialbacterial microbiota of of all all the the olive olive and and brine brine samples samples analyzed analyzed was was covered covered by 24 by phyla, 24 phyla, in 284 inwhich which Firmicutes, Firmicutes, Proteobacteria, Proteobacteria, Bacteroidetes Bacteroidetes and and Actinobacteria Actinobacteria were were the the predominant. predominant. Among Among 285 Firmicutes,Firmicutes, Lactobacillaceae was was the the main main bacterial bacterial family family identified, identified, reaching reaching approxim approximatelyately an 286 anaverage average abundance abundance of 35.14 of 35.14%% ± 24.37%24.37% and 45.06 and% 45.06%± 43.16% among43.16% the among olive theand olivebrine andsamples, brine 287 samples,respectively, respectively, from Messinia/Lakonia, from Messinia /whereasLakonia, average whereas abundances average abundances were higher were in olives higher and in brines olives 288 andfrombrines Aitoloakarnania, from Aitoloakarnania, i.e., 63.53% ± 26.79 i.e.,%63.53% and 70.71%26.79% ± 23.39%, and respectively 70.71% (Figure23.39%, 4). This respectively was in 289 (Figureaccordance4). This with was the in classical accordance microbiological with the classical analysis, microbiological since LAB counts analysis, of olives since LAB and countsbrines 290 oforiginating olives and from brines Messinia/Lakonia originating from were Messinia significantly/Lakonia lower were compared significantly to these lower from compared Aitoloakarnania to these 291 fromregion Aitoloakarnania (Figure 2A). Apart region from (Figure Lactobacillaceae,2A). Apart fromCelerinatantimonadaceae, Lactobacillaceae, Celerinatantimonadaceae, Leuconostocaceae and 292 LeuconostocaceaeAcetobacteraceae were and Acetobacteraceaemore abundant families weremore in olive abundant and brine families samples in olivefrom andMessinia/Lakonia brine samples 293 than Aitoloakarnania as well. Interestingly, several families, including Propionibacteriaceae, from Messinia/Lakonia than Aitoloakarnania as well. Interestingly, several families, including 294 Staphylococcaceae and Chitinophagaceae, were mainly identified in olives, while Propionibacteriaceae, Staphylococcaceae and Chitinophagaceae, were mainly identified in olives, 295 Pseudomonadaceae, Cardiobacteriaceae and Acetobacteraceae were identified in brine samples while Pseudomonadaceae, Cardiobacteriaceae and Acetobacteraceae were identified in brine samples 296 (Table S1A,B). At the genus level, Lactobacillus, Celerinatantimonas, Propionibacterium, Staphylococcus (Table S1A,B). At the genus level, Lactobacillus, Celerinatantimonas, Propionibacterium, Staphylococcus and 297 and Leuconostoc dominated the olive samples with varying abundances depending on the Leuconostoc dominated the olive samples with varying abundances depending on the geographical 298 geographical region (Figure 5A and Table S2A). According to the Venn diagram constructed on the region (Figure5A and Table S2A). According to the Venn diagram constructed on the basis of the average 299 basis of the average bacterial microbiota of olive samples from Messinia/Lakonia and olive samples bacterial microbiota of olive samples from Messinia/Lakonia and olive samples from Aitoloakarnania, 300 from Aitoloakarnania, the majority of the identified genera (25 out of 44) were shared among the the majority of the identified genera (25 out of 44) were shared among the olives from the two regions 301 olives from the two regions (Figure 5C). Among the core genera, Lactobacillus, Celerinatantimonas, 302 Propionibacterium, Staphylococcus, Sediminibacterium, Leuconostoc, Sphingomonas and Enterobacter were

Microorganisms 2020, 8, 672 9 of 24

Microorganisms 2020, 8, x FOR PEER REVIEW 9 of 24 (Figure5C). Among the core genera, Lactobacillus, Celerinatantimonas, Propionibacterium, Staphylococcus, 303 Sediminibacterium,the most abundant. Leuconostoc, However, Sphingomonas 10 and 9 bacterialand Enterobacter genera werewere unique the most in olives abundant. from Messinia/Lakonia However, 10 and / 304 9and bacterial Aitoloakarnania, genera were respect uniqueively, in olives all below from Messinia1.0% abundance,Lakonia exce andpt Aitoloakarnania, of Serratia (1.75% respectively, ± 6.77%) in / 305 allolives below from 1.0% Messinia/Lakonia abundance, except and of AcetobacterSerratia (1.75% (1.34 ± 6.77%)5.00%) inin olivesolives fromfrom MessiniaAitoloakarnaniaLakonia (Table and 306 AcetobacterS2A). Furthermore,(1.34 5.00%) the bacterial in olives frommicrobiota Aitoloakarnania of brine samples (Table S2A). from Furthermore, both geographical the bacterial regions 307 microbiotaconsisted of of brine 44 dominant samples from genera, both geographicalincluding Lactobacillus, regions consisted Pseudomonas, of 44 dominant Sphingomonas, genera, including Suttonella, 308 Lactobacillus,Celerinatantimonas Pseudomonas, and Leuconostoc Sphingomonas,, reaching Suttonella, approximately Celerinatantimonas 80 % of the andbacterialLeuconostoc sequences, reaching (Figure 309 approximately5B and Table S2B). 80% ofBased the bacterialon the Venn sequences diagram, (Figure only5 14B andgenera Table were S2B). shared Based among on the the Venn brine diagram, samples / 310 onlyfrom 14 Messinia/Lakonia genera were shared and among Aitoloakarnania, the brine samples compared from to Messiniathe 25 coreLakonia genera and of olives Aitoloakarnania, (Figure 5D). / 311 comparedBrine samples to the from 25 core the genera Messinia/Lakonia of olives (Figure region5D). had Brine the samplesmajority fromof unique the Messinia genera, i.e.Lakonia, 18; however, region 312 hadonly the three majority had abundance of unique genera,above 1.0 i.e., %, 18; namely however, Bacteroides only three (1.83 had% ± abundance2.81%), Celerinatantimonas above 1.0%, namely (4.71% 313 Bacteroides± 12.22%)( 1.83% and Weissella2.81% ),(1.05Celerinatantimonas% ± 4.08%). On(4.71% the contrary,12.22%) among and Weissella the 12 (1.05% unique bacterial4.08%). On genera the 314 contrary,identified among in brine the 12samples unique from bacterial Aitoloakarnania, genera identiﬁed five were in brine the samples most abundant, from Aitoloakarnania, i.e., Swaminathania ﬁve 315 were(1.83 the% ± most 6.37%), abundant, Acetobacter i.e., (1.24Swaminathania% ± 3.82%),(1.83% Salinicola6.37%), (1.13%Acetobacter ± 2.16%), Empedobacter(1.24% 3.82%), (1.04%Salinicola ± 3.89%) 316 (1.13%and Novosphingobium2.16%), Empedobacter (1.03% (1.04%± 3.84%) (Table3.89%) S2B). and Novosphingobium (1.03% 3.84%) (Table S2B).

317 318 FigureFigure 4. Average4. Average relative relative abundance abundance (%) of dominant (%) of dominantbacterial families bacterial obtained families by 16S obtained metagenomics by 16S / 319 analysismetagenomics of olive (A analysis) and brine of (oliveB) samples (A) and from brine the Messinia (B) samplesLakonia from and the Aitoloakarnania Messinia/Lakonia regions. and 320 Aitoloakarnania regions.

Microorganisms 2020, 8, 672 10 of 24 Microorganisms 2020, 8, x FOR PEER REVIEW 10 of 24

321 Figure 5. Average relative abundance (%) of dominant bacterial genera obtained by 16S metagenomics 322 Figure 5. Average relative abundance (%) of dominant bacterial genera obtained by 16S analysis of olive (A) and brine (B) samples from the Messinia/Lakonia and Aitoloakarnania regions. 323 metagenomics analysis of olive (A) and brine (B) samples from the Messinia/Lakonia and Venn diagram showing the number of unique and shared bacterial genera among olives (C) and brines 324 Aitoloakarnania regions. Venn diagram showing the number of unique and shared bacterial genera (D) from the two geographical regions. 325 among olives (C) and brines (D) from the two geographical regions. The yeasts/fungal community of olive and brine samples from both geographical regions was 326 The yeasts/fungal community of olive and brine samples from both geographical regions was characterized by a high level of Ascomycota phylum (89.58%). Compared to bacteria, yeasts/fungal 327 characterized by a high level of Ascomycota phylum (89.58%). Compared to bacteria, yeasts/fungal microbiota was less diverse, with Pichiaceae, Aspergillaceae, Saccharomycetaceae and Debaryomycetaceae 328 microbiota was less diverse, with Pichiaceae, Aspergillaceae, Saccharomycetaceae and being the dominant among the 27 families identified in all samples (Figure6 and Table S1C,D). 329 Debaryomycetaceae being the dominant among the 27 families identified in all samples (Figure 6 and The average yeasts/fungal microbiota from olives originating from Messinia/Lakonia was similar 330 Table S1C,D). The average yeasts/fungal microbiota from olives originating from Messinia/Lakonia ff 331 towas that similar of Aitoloakarnania. to that of Aitoloakarnania. However, a However, few di erences a few weredifferences observed were between observed the between two microbial the two 332 fingerprints,microbial fingerprin such asts, the such relatively as the relatively high abundance high abundance of the Cladosporiaceae of the Cladosporiaceae (3.68% (3.686.42%)% ± 6.42 and%) / ff 333 Ascomycotaand Ascomycota (3.30% (3.307.17%)% ± 7.17 families%) families in Messinia in Messinia/LakoniaLakonia and the and Pha theomycetaceae Phaffomycetaceae (2.97% (2.977.10%)% ± / 334 family7.10%) in family Aitoloakarnania in Aitoloakarnania microbiota microbiota (Figure6A and(Figure Table 6A S1C). and Interestingly, Table S1C). the Interestin yeastsgly,fungal the 335 communityyeasts/fungal of brinecommunity samples of brine was less samples diverse was than less thatdiverse of olives,than that including of olives, 12 including families 12 (Figure families6B 336 and(Figure Table 6 S1D).B and TheTable average S1D). The abundance average abundance of Pichiaceae of Pichiaceae family was family relatively was relatively higher in higher brine samplesin brine / 337 fromsamples Messinia from LakoniaMessinia/Lakonia (74.77 25.10%)(74.77 ± compared25.10%) compared to Aitoloakarnania to Aitoloakarnania (42.27% (42.2736.32%),% ± 36.32 while,%), 338 onwhile, the contrary, on the Saccharomycetaceae contrary, Saccharomycetaceae (25.27% 34.48% (25.27 vs.% 14.60%± 34.4823.87%),% vs. Debaryomycetaceae14.60% ± 23.87%), ff 339 (15.73%Debaryomycetaceae32.51% vs. 5.12%(15.73% ±10.05%), 32.51% Phavs. 5.12omycetaceae% ± 10.05%), (8.99% Phaffomycetaceae20.85% vs. 0.11%(8.99% ± 0.41%) 20.85%and vs. 340 Cystofilobasidiaceae0.11% ± 0.41%) and Cystofilobasidiaceae (5.05 15.41 vs. 0.70 (5.05 ±1.91%) 15.41 vs. families 0.70 ± 1.91 were%) mainlyfamilies identifiedwere mainly in identified the brine 341 samplesin the brine from Aitoloakarnaniasamples from Aitoloakarnania region. At the region. genus level,At thePichia, genus Penicillium, level, Pichia, Ogataea, Penicillium, Saccharomyces, Ogataea, 342 MalasseziaSaccharomyces,and Millerozyma Malasseziawere and predominantMillerozyma inwere olive predominant samples, with in varying olive samples, abundances with depending varying 343 onabundances the geographical depending region on (Figure the geographical7A and Table region S2C). Among (Figure the 7A 39 and yeasts Table/fungal S2C). genera Among identified the 39 344 inyeasts/fungal olive samples, genera approximately identified 43.6% in olive were samples, shared among approximately the samples, 43.6% compared were shared to 56.81% among of the 345 bacterialsamples, core compared genera. Interestingly,to 56.81 % of fromthe bacterial the total core of 39 genera. genera Interestingly, identified, 15 from and seventhe total were of unique39 genera for 346 olivesidentified, originating 15 and from seven Messinia were/Lakonia unique and for Aitoloakarnania, olives originating respectively from Messinia/Lakonia (Figure7C and Table and 347 S2C).Aitoloakarnania, It should be noted respectively though, that(Figure the abundances7C and Table of the S2C). majority It should of the be unique noted genera though, were that below the 348 1.0%,abundances except of ofEngyodontium the majority (3.30%of the unique 7.17%) genera and Brettanomyces were below 1.0(1.93% %, except 5.54%) of Engyodontium for Messinia /(3.30Lakonia% ± 349 and7.17Wickerhamomyces%) and Brettanomyces(2.96% (1.93%7.10%) ± 5.54%) for for Aitoloakarnania Messinia/Lakonia regions. and Wickerhamomyces As already mentioned (2.96% ± 7.10 above,%) 350 thefor richness Aitoloakarnania of the yeasts regions./fungal As microbiota already of mentioned brine samples above, was the lower richness compared of the to yeasts/f that ofungal olive 351 microbiota of brine samples was lower compared to that of olive samples, which is also shown in the 352 Venn diagram (Figure 7B,D). The majority of the yeast/fungal genera identified (62.5%) were shared

Microorganisms 2020, 8, 672 11 of 24

Microorganisms 2020, 8, x FOR PEER REVIEW 11 of 24 samples, which is also shown in the Venn diagram (Figure7B,D). The majority of the yeast /fungal 353 generaamong identiﬁedthe brine samples (62.5%) werefrom the shared two amongregions, the with brine Pichia, samples Ogataea, from Saccharomyces the two regions, and Millerozyma with Pichia, 354 Ogataea,being the Saccharomyces dominant onesand (TableMillerozyma S2D). beingHowever, the dominantfive and four ones genera (Table S2D).were However,unique for ﬁve brines and from four 355 generaMessinia/Lakonia were unique andfor brines Aitoloakarnan from Messiniaia, / respectively,Lakonia and Aitoloakarnania, all below 1.0% respectively, abundance, all below except 1.0% of 356 abundance,Zygosaccharomyces except (7.72 of Zygosaccharomyces% ± 20.70%) for (7.72% Messinia/Lakonia 20.70%) for and Messinia Kluyveromyces/Lakonia (1.49 and%Kluyveromyces ± 3.96%) for 357 (1.49%Aitoloakarnania 3.96%) forregions Aitoloakarnania (Table S2D). regions (Table S2D).

358 Figure 6. Average relative abundance (%) of dominant yeast/fungal families obtained by internal 359 Figure 6. Average relative abundance (%) of dominant yeast/fungal families obtained by internal transcribed spacer (ITS) metagenomics analysis of olive (A) and brine (B) samples from the Messinia/ 360 transcribed spacer (ITS) metagenomics analysis of olive (A) and brine (B) samples from the Lakonia and Aitoloakarnania regions. 361 Messinia/Lakonia and Aitoloakarnania regions.

Microorganisms 2020, 8, 672 12 of 24 Microorganisms 2020, 8, x FOR PEER REVIEW 12 of 24

362 Figure 7. Average relative abundance (%) of dominant yeast/fungal genera obtained by ITS 363 Figure 7. Average relative abundance (%) of dominant yeast/fungal genera obtained by ITS metagenomics analysis of olive (A) and brine (B) samples from Messinia/Lakonia and Aitoloakarnania 364 metagenomics analysis of olive (A) and brine (B) samples from Messinia/Lakonia and regions. Venn diagram showing the number of unique and shared yeast/fungal genera among olives 365 Aitoloakarnania regions. Venn diagram showing the number of unique and shared yeast/fungal (C) and brines (D) from the two geographical regions. 366 genera among olives (C) and brines (D) from the two geographical regions. 3.3. Differences in Microbial Community Structure By Multivariate Analysis 367 3.3. Differences in Microbial Community Structure By Multivariate Analysis A two-dimensional graphical representation (heatmap) was employed for data illustration, 368 in whichA two each-dimensional microbial graphical species isrepresentation characterized (heatmap) by a single was row employed and each for data column illustration, represents in 369 anwhich individual each microbial olive or species brine sample is characterized analyzed by from a single the two row geographical and each column regions. represents Specifically, an 370 Figureindividual8 shows olive the or heatmapbrine sample of the analyzed bacterial from community the two forgeographical Kalamata naturalregions. black Specifically, olives clustered Figure 8 371 byshows Euclidean the heatmap distance. of The the bacterial clusters indicated community higher for Kalamata bacterial diversitynatural black in both olives olive clustered and brine by 372 samplesEuclidean originating distance. The from clusters Messinia indicated/Lakonia higher compared bacterial to diversity the Aitoloakarnania in both olive region. and brine Regarding samples 373 originating from Messinia/Lakonia compared to the Aitoloakarnania region. Regarding the bacterial the bacterial community structure of brines, five samples, namely B1,B2,B4,B27, and B28, from the 374 Messiniacommunity/ Lakonia structure area of brines, presented five highersamples abundances, namely B1, B in2, Pseudomonas,B4, B27, and B28 Microbacterium,, from the Messinia/Lakonia Caulobacter, 375 Sphingomonas,area presented Sediminibacterium, higher abundances Burkholderia, in Pseudomonas, Comamonas, Microbacterium,and Bacteroides Caulobacspeciester, Sphingomonas, (Figure8A). 376 InSediminibacterium, addition, Lactobacillus Burkholderia,spp. and Comamonas,Leuconostoc spp.and were Bacteroides abundant species in several (Figure brine samples8A). In from addition, both 377 geographicalLactobacillus areas,spp. and whereas, Leuconostoc on the otherspp. hand, wereLactococcus abundantspp. in was several abundant brine in samples three samples from from both 378 geographical areas, whereas, on the other hand, Lactococcus spp. was abundant in three samples from Aitoloakarnania, namely B18,B22 and B23. 379 Aitoloakarnania, namely B18, B22 and B23.

Microorganisms 2020, 8, 672 13 of 24 Microorganisms 2020, 8, x FOR PEER REVIEW 13 of 24

380

381 FigureFigure 8. 8Hierarchically. Hierarchically clustered clustered heatmap heatmap of theof the bacterial bacterial community community from from cv. Kalamatacv. Kalamata natural natural black 382 black olives from the brine (A) and olives (B). Classes A and M correspond to the samples from the olives from the brine (A) and olives (B). Classes A and M correspond to the samples from the region of 383 region of Aitoloakarnania (A) and Messinia/Lakonia (M), respectively. The sample codes are Aitoloakarnania (A) and Messinia/Lakonia (M), respectively. The sample codes are indicated in Table1. 384 indicated in Table 1. The bacterial community of olives from the Kalamata variety from both geographical areas 385 is presentedThe bacterial in Figure community8B, with ofLactobacillus olives fromspp. the Kalamata and Leuconostoc variety spp.from both being geographical abundant in areas several is 386 presented in Figure 8B, with Lactobacillus spp. and Leuconostoc spp. being abundant in several samples samples from both areas. Two samples from Aitoloakarnania, namely O2 and O16, presented a high 387 from both areas. Two samples from Aitoloakarnania, namely O2 and O16, presented a high abundance abundance in enterobacteria (Pantoea spp., Serratia spp., Klebsiella spp. and Enterobacter spp.), whereas 388 in enterobacteria (Pantoea spp., Serratia spp., Klebsiella spp. and Enterobacter spp.), whereas Corynebacterium spp., Propionibacterium spp. and Staphylococcus spp. were mostly abundant in olive 389 Corynebacterium spp., Propionibacterium spp. and Staphylococcus spp. were mostly abundant in olive samples from Messinia/Lakonia. It needs to be noted that Staphylococcus spp. was highly correlated 390 samples from Messinia/Lakonia. It needs to be noted/ that Staphylococcus spp. was highly correlated with four samples (O1,O4,O6 and O12) from Messinia Lakonia, a fact that should raise concern about 391 with four samples (O1, O4, O6 and O12) from Messinia/Lakonia, a fact that should raise concern about the hygiene conditions in these fermentations. 392 the hygiene conditions in these fermentations. The structure of the yeasts/fungal community of olive and brine samples from both regions is 393 The structure of the yeasts/fungal community of olive and brine samples from both regions is illustrated in Figure9. Higher fungal diversity can be observed in fermented olives compared to brines 394 illustrated in Figure 9. Higher fungal diversity can be observed in fermented olives compared to 395 regardlessbrines regardless of the geographical of the geographical region of region origin. of Theorigin. most The represented most represented genus in genus the brines in the of brines samples of / 396 fromsamples Messinia fromLakonia Messinia/Lakonia was Pichia followed was Pichia by Candida,followed Zygosaccharomyces, by Candida, Zygosaccharomyces, Zygoascus, Zygotorulaspora, Zygoascus, 397 Brettanomyces,Zygotorulaspora,and Schanniomyces. Brettanomyces, Onand the Schanniomyces. contrary, brine samplesOn the from contrary, Aitoloakarnania brine samples were abundant from 398 inAitoloakarnaniaSaccharomyces spp., wereWickerhamomyces abundant in Saccharomycesspp., and Ogataea spp., Wickerhamomycesspp. (Figure9A). spp., A less and clear Ogataea clustering spp. / 399 pattern(Figure between 9A). A the less two clear regions clustering could be pattern obtained between for the the yeasts twofungal regions community could be of obtained fermented for olives. the / / 400 Theyeasts/fungal most abundant community yeasts fungi of fermented from the olives. area of The Messinia most abundantLakonia wereyeasts/fungiPichia spp., fromCandida the areaspp., of 401 AureobacidiumMessinia/Lakoniaspp., wereAlternaria Pichia spp.,spp., CandidaPenicillium spp.,spp., AureobacidiumCladosporium spp.,spp., Alternaria and Phyllactinia spp., Penicilliumspp., spp., while 402 SaccharomycesCladosporium spp.spp., and andOgataea Phyllactiniaspp. spp., were highlywhile Sacc relatedharomyces to fermented spp. and olives Ogataea from spp. Aitoloakarnania were highly 403 (Figurerelated9B). to fermented olives from Aitoloakarnania (Figure 9B).

Microorganisms 2020, 8, 672 14 of 24

Microorganisms 2020, 8, x FOR PEER REVIEW 14 of 24

404

405 FigureFigure 9. 9.Hierarchically Hierarchically clustered clustered heatmapheatmap of the yeasts/fungal yeasts/fungal community community from from cv. cv. Kalamata Kalamata natural natural 406 blackblack olives olives from from the the brine brine (A(A)) andand olivesolives ((BB).). ClassClass A and M corresponds to to the the samples samples from from the the 407 regionregion of of Aitoloakarnania Aitoloakarnania (A) (A) and and Messinia Messinia/Lakonia/Lakonia (M), respectively. (M), respectively. The sample The s codesample are codes indicated are 408 inindicated Table1. in Table 1.

409 PLS-DAPLS-DA analysis analysis using using species-level species-level OTUs OTUs revealed revealed several di severalfferences differences in microbial in communities microbial 410 amongcommuni theties olive among and brine the olive samples and brine analyzed. samples The analyzed. PLS-DA scoresThe PLS plot-DA for scores the bacterial plot for microbiotathe bacterial in 411 themicrobiota brines showed in the discriminationbrines showed betweendiscrimination the two between geographical the two regions, geographical although regions, an overlapping although wasan 412 evidentoverlapping among was a number evident of among samples a (Figurenumber 10 ofA). samples In addition, (Figure the 10 loadingsA). In ad plotdition, showed the loadings the correlation plot 413 ofshowed each bacterial the correlation species toof theeach brine bacterial samples species from to the the two brine regions samples (Figure from 10theB). two It can regions be concluded (Figure 414 10B). It can be concluded that brine samples from Aitoloakarnania (A) were highly correlated with that brine samples from Aitoloakarnania (A) were highly correlated with Lactobacillus spp., whereas 415 Lactobacillus spp., whereas brine samples from Messinia/Lakonia (M) were associated with brine samples from Messinia/Lakonia (M) were associated with Pseudomonas spp., Microbacterium spp., 416 Pseudomonas spp., Microbacterium spp., Caulobacter spp., Burkholderia spp. and Sphingomonas spp. Caulobacter spp., Burkholderia spp. and Sphingomonas spp. Furthermore, the influence of bacterial 417 Furthermore, the influence of bacterial species (X-variable) on the geographical regions (Y-response) species (X-variable) on the geographical regions (Y-response) can be also expressed with the Variable 418 can be also expressed with the Variable Importance in Projection (VIP) coefficient which, in our case, Importance in Projection (VIP) coefficient which, in our case, indicates the bacterial species having 419 indicates the bacterial species having the highest importance in explaining the Y-variance (i.e., the the highest importance in explaining the Y-variance (i.e., the origin of the samples). A VIP value 420 origin of the samples). A VIP value of 1.0 has generally been accepted as a cut-off limit in variable of 1.0 has generally been accepted as a cut-off limit in variable selection; thus, variables exceeding 421 selection; thus, variables exceeding this limit can be considered to be highly influential. Based on the this limit can be considered to be highly influential. Based on the VIP values, Microbacterium spp., 422 VIP values, Microbacterium spp., Sphingomonas spp., Burkholderia spp., Bacterioides spp., Pseudomonas 423 Sphingomonasspp., Caulobacterspp., spp.,Burkholderia Pelomonasspp., spp.,Bacterioides Sediminibacterspp., Pseudomonas spp., and Celerinatantimonasspp., Caulobacter spp., spp.Pelomonas were highlyspp., 424 Sediminibacterassociated withspp., the and brineCelerinatantimonas samples from thespp. area were of highlyMessinia/Lakonia associated with (Figure the brine11). On samples the contrary, from the / 425 areaLactobacillus of Messinia spp.,Lakonia Salinicola (Figure spp., 11 ).Lactococcus On the contrary, spp., IdiomarinaLactobacillus spp.,spp., andSalinicola Martellelaspp., spp.Lactococcus were highlyspp., 426 Idiomarinacorrelatedspp., with andthe brineMartellela samplesspp. from were the highlyarea of correlatedAitoloakarnania. with theFurthermore, brine samples the discrimination from the area 427 ofprofile Aitoloakarnania. according to Furthermore, the PLS-DA analysis the discrimination for the bacterial profile communities according to in the the PLS-DA olive samples analysis was for 428 thesimilar bacterial to that communities of brines; however, in the olive it should samples be noted was similarthat the tooverlapping that of brines; observed however, in brine it should samples be 429 notedwas absent that the in overlappingolives (Figures observed S7). Interestingly, in brine samples a different was pattern absent of in influential olives (Figure bacterial S7). Interestingly,species was 430 aobserved different patternon the olive of influential samples compared bacterial speciesto the brines, was observed based on on the the VIP olive scores. samples Specifically, compared olive to 431 thesamples brines, from based Messinia/Lakonia on the VIP scores. were Specifically, highly associated olive samples with Staphylococcus from Messinia spp.,/Lakonia Propionibacterium were highly 432 associatedspp., Micrococcus with Staphylococcus spp., Corynebacteriumspp., Propionibacterium spp., Celerinatantimonasspp., Micrococcus spp., Arcicellaspp., spp.,Corynebacterium Streptococcus spp.,spp., 433 CelerinatantimonasRubrobacter spp., andspp., PrevotellaArcicella spp.,spp., whileStreptococcus, on the contrary,spp., Rubrobacter olive samplesspp., from and AitoloakarnaniaPrevotella spp., were while, 434 onhighly the contrary, correlated olive with samples Lactobacillus from spp., Aitoloakarnania Lactococcus spp., were and highly Leuconostoc correlated spp. with(FigureLactobacillus S8). spp., Lactococcus spp., and Leuconostoc spp. (Figure S8).

MicroorganismsMicroorganisms 20202020, 8, ,x8 FOR, 672 PEER REVIEW 15 of15 24 of 24 Microorganisms 2020, 8, x FOR PEER REVIEW 15 of 24

435 435 436 Figure 10. Plot of scores (A) and loadings (B) of the first two latent variables of the PLS-DA model of 436 FigureFigure 10. 10.Plot Plot of of scoresscores ((AA)) andand loadings ( (BB)) of of the the first ﬁrst two two latent latent variables variables of ofthe the PLS PLS-DA-DA model model of 437 the bacterial community build on the brine samples from the region of Aitoloakarnania (A) and 437 ofthe the bacterial bacterial community community build build on on the the brine brine samples samples from from the the region region of of Aitoloakarnania Aitoloakarnania (A) (A) and and 438 Messinia/Lakonia (M). 438 MessiniaMessinia/Lakonia/Lakonia (M). (M).

439 439 440 Figure 11. Most influential bacterial genera of the brine samples from the region of Aitoloakarnania 440 FigureFigure 11.11. MostMost influential influential bacterial bacterial genera genera of the of the brine brine samples samples from from the the region region of ofAitoloakarnan Aitoloakarnaniaia 441 (A) and Messinia/Lakonia/ (M) based on the Variable Importance in Projection (VIP) scores from the 441 (A)(A) and and Messinia/Lakonia Messinia Lakonia (M) (M) based based on onthe the Variable Variable Importance Importance in Projection in Projection (VIP (VIP)) scores scores from from the the 442 Partial Least Squares Discriminant Analysis (PLS-DA) analysis. The color bars indicate the intensity 442 PartialPartial Least Least Squares Squares Discriminant Discriminant Analysis Analysis (PLS (PLS-DA)-DA) analysis. analysis. The The color color bars bars indicate indicate the the intensity intensity of 443 of thevariable in the respective group. 443 of thethevariablevariable in the respective group. 444 OnOn the the other other hand, hand, the PLS-DA the PLS analysis-DA analysis based on based the yeasts on the/fungal yeasts/fungal community community structure provided structure a 444 On the other hand, the PLS-DA analysis based on the yeasts/fungal community structure 445 clearprovided discrimination a clear discrimination among the brine among samples the frombrine the samples two geographical from the two areas geographical (Figure 12A). areas The (Figure absence 445 provided a clear discrimination among the brine samples from the two geographical areas (Figure 446 of12 overlappingA). The absence among the of brineoverlapping samples among indicates the that brine the yeasts samples/fungal indicates community that could the yeasts/funga be used morel 446 12A). The absence of overlapping among the brine samples indicates that the yeasts/fungal 447 effcommunityectively in the could discrimination be used more of theeffectively sample’s in originthe discrimination compared to theof the bacteria sample’s community. origin compared In addition, to 447 community could be used more effectively in the discrimination of the sample’s origin compared to 448 thethe plot bacteria of loadings community. revealed In addition, that the brinethe plot samples of loadings from therevealed area of that Messinia the brine/Lakonia samples were from highly the 448 the bacteria community. In addition, the plot of loadings revealed that the brine samples from the 449 associatedarea of Messinia/Lakonia with Pichia spp., wereCandida highlyspp., associatedCryptococcus withspp., PichiaCystofilobasidium spp., Candida spp.,spp., Cryptococcus and Brettanomyces spp., 449 area of Messinia/Lakonia were highly associated with Pichia spp., Candida spp., Cryptococcus spp., 450 spp.,Cystofilobasidium whereas Wickerhamomyces spp., and Brettanomycesspp., Debaryomyces spp., whereasspp., KluyveromycesWickerhamomycesspp. spp., and DebaryomycesSaccharomyces spp.,spp. 450 Cystofilobasidium spp., and Brettanomyces spp., whereas Wickerhamomyces spp., Debaryomyces spp., 451 wereKluyveromyces correlated spp. with and the Saccharomyces brine samples spp. from were the correlated area of Aitoloakarnania with the brine samples (Figure from12B). the The area highly of 451 452Kluyveromyces spp. and Saccharomyces spp. were correlated with the brine samples from the area of influentialAitoloakarnania yeasts/fungal (Figure species 12B). wereThe highly further influential identified yeasts/fungal by the VIP scores species (cut were off value further of 1) identified [38], in which by 452 453Aitoloakarnania the VIP scores (Figure (cut off 12 valueB). The of highly 1) [38] influential, in which Pichiayeasts/fungal spp. was species also found were tofurther be highly identified correlated by Pichia spp. was also found to be highly correlated with brine samples from Messinia/Lakonia, together 453 454the VIPwith scores brine (cut samples off value from of 1) Messinia/Lakonia, [38], in which Pichia together spp. was with also Brettanomyc found toes be spp., highly Candida correlated spp., 454 with brine samples from Messinia/Lakonia, together with Brettanomyces spp., Candida spp.,

Microorganisms 2020, 8, x FOR PEER REVIEW 16 of 24 MicroorganismsMicroorganisms2020 2020, 8,, 8 672, x FOR PEER REVIEW 1616 of of 24 24 455 Zygosaccharomuces spp., and Zygotorulaspora spp. On the contrary, for brine samples from 456455 Zygosaccharomuces spp., and Zygotorulaspora spp. On the contrary, for brine samples from Aitoloakarnania,with Brettanomyces highspp., correlationsCandida spp., wereZygosaccharomuces observed for Wickerhamomycesspp., and Zygotorulaspora spp. andspp. Kluyveromyces On the contrary, spp. 457456 (FigureAitoloakarnania, 13). high correlations were observed for Wickerhamomyces spp. and Kluyveromyces spp. 457 for(Figure brine 13 samples). from Aitoloakarnania, high correlations were observed for Wickerhamomyces spp. and Kluyveromyces spp. (Figure 13).

458458 459 Figure 12. Plot of scores (A) and loadings (B) of the first two latent variables of the PLS-DA model 459 FigureFigure 12 12.. PlotPlot of of scores scores ( A) and loadingsloadings ( B(B) of) of the the ﬁrst first two two latent latent variables variables of the of PLS-DA the PLS model-DA model of 460460 of ofthe thethe yeasts/fungal yeastsyeasts/fungal/fungal community community build build build on on theon the the brine brine brine samples samples samples from from from the regionthe the region region of Aitoloakarnania of of Aitoloakarnania Aitoloakarnania (A) and (A)(A) 461461 Messinia/Lakonia (M). andand Messinia/Lakonia Messinia/Lakonia (M). (M).

462462 463463 FigureFigureFigure 13.13 13. MostMost. Most inﬂuential influential influential yeasts yeasts/fungal yeasts/fungal/fungal genera generagenera of the brine ofof the samplesthe brine brine from samples samples the region from from of Aitoloakarnania the the region region of of 464464 Aitoloakarnania(A)Aitoloakarnania and Messinia (A)/ Lakonia(A) and and Messinia/Lakonia Messinia/Lakonia (M) based on the (M) (M) VIP based based scores on on fromthe the VIP VIP the scores PLS-DAscores from from analysis. the the PLS PLS The--DADA color analysis.analysis. bars 465465 TheindicateThe color color thebars bars intensity indicate indicate of the thethe intensity variableintensity of in of the the the variable respective variable in in the group. the respective respective group. group. / 466466 Moreover,Moreover,Moreover, the the the score score score plot plot plot of of of the the the PLS PLS-DA PLS-DA-DA ana analysis analysislysis forfor for the thethe yeasts/fungal yeastsyeasts/fungalfungal communities communities,communities,, althoughalthough itit it 467467 discriminateddiscriminateddiscriminated the the the olive olive olive samples samples samples from from from the the the two two two geographical geographicalgeographical regions, regions,regions, revealed revealedrevealed an an overlap overlap among among somesome some 468468 ofof ofthem them them (Figure (Figure S9).S 9S).9 According).According According to to theto the the VIP VIP VIP scores, scores, scores,Pichia Pichia Pichiaspp., spp., spp.,Candida Candida Candidaspp., spp., spp., and andBrettanomycesand BrettanomycesBrettanomycesspp. spp. werespp. / 469469 werealsowere highlyalso also highly correlatedhighly correlated correlated with thewith with olive the the samplesolive olive samples samples from Messinia from from Messinia/Lakonia, Messinia/Lakonia,Lakonia, as in the asas case inin thethe of brinecasecase of samplesof brinebrine

Microorganisms 2020, 8, 672 17 of 24 from the same area Cladosporium spp., Alternaria spp., and Engyodontium spp. Similar to the brines, Wickerhamomyces spp. and Kluyveromyces spp. were found to be highly correlated with the olives from Aitoloakarnania, along with Ogatae spp. (Figure S10).

4. Discussion In this study, the bacterial and yeast/fungal microbiota of the olives and brine samples of cv. Kalamata natural black olives were elucidated using amplicon-based metagenomic analysis. Samples were industrially fermented according to the traditional Greek-style method and obtained from the two main producing regions of this cultivar, namely Messinia/Lakonia in southern Peloponnese and Aitoloakarnania in western Greece. Due to the highly diverse terrain and climate conditions between the two regions, the potential microbial biogeography association between certain taxa and geographical area was also assessed. A critical factor that affects the fermentation process and determines the survival of LAB during the fermentation is the salt concentration in the brine. It has been reported that high salt levels (>8%) could favor the dominance of yeasts and inhibit the growth of LAB [39], rendering a final product with less preservation characteristics. Nowadays, the Greek table olive industry has reduced the salt level in the brine to 6–7% during the period of active fermentation to ensure the dominance of LAB and thus improve the preservation and sensory attributes of the final product. According to our analysis, the average salt concentration in the brine samples from both regions was less than 8% (Figure2D), although great variations were observed in the individual samples (Figure S5), indicating that further measures must be undertaken by the industry to standardize the fermentation process so as to improve the quality of the final product. As already mentioned, the co-existence of bacterial and yeast/fungal communities is of fundamental importance during the fermentation process. These communities could be considered as a natural resource associated with a specific agroecosystem, shaping the microbial “terroir” of table olives [40]. The term “terroir”, although initially employed in oenology, has been extended, nowadays, to other crops and food products to link geographical origin and environmental conditions to quality aspects of agricultural commodities [41]. According to the alpha-diversity analysis, the diversity (both richness and evenness) of the yeast/fungal microbiota of both olive and brine samples (Figure3C,D), and the bacterial microbiota of olives (Figure3A), was higher in the samples from Messinia /Lakonia compared to the Aitoloakarnania region, while bacteria communities of brine samples (Figure3B) were similar between the two geographical regions. Interestingly, based on the identified OTUs, the yeasts/fungal microbiota of both olives and brines was less diverse compared to bacteria. This could be explained since the olives were spontaneously fermented by the indigenous microbiota, which is subjected to seasonal variations [42], as well as to a variety of contaminating microorganisms from fermentation vessels, pipelines, pumps and other equipment in contact with olives and brine, that could be considered as a persistent microbiota in the factory facilities [18,22,43]. According to the results of the 16S metagenomics analysis, Lactobacillaceae was the dominant family identified in olive and brine samples from both regions (Figure4). The high average abundance of the family Lactobacillaceae in the olive samples from both regions confirmed the ability of LAB to colonize the surface of olives [44,45]. Among the family Lactobacillaceae, Lactobacillus and Leuconostoc were the most abundant genera identified, indicating that these were all lactic acid fermentations, as also verified by the classical microbiological analysis (Figure2). Both genera are commonly found in the microbiota of fermented green and black olives using both classical microbiological and metagenomics analyses [6,17,20,22–24]. This observation updates the findings of other researchers who reported that LAB could not be detected throughout the spontaneous fermentation of Kalamata variety or at least they could be detected only in the initial stage of the process [46,47]. In a recent work, the bacterial and yeast/fungal microbiota of brines from three Greek olive varieties, namely Kalamata, Conservolea and Kerasoelia, originating from different geographical regions of Greece, was assessed using an amplicon-based metagenomic technique [26]. The authors reported the dominance of the genus Lactobacillus (74%) in the Kerasoelia variety, whereas in the Kalamata and Conservolea varieties, Microorganisms 2020, 8, 672 18 of 24 the genus Cellulosimicrobium prevailed, with abundances ranging from 39% to 90%. In our samples, the genus Cellulosimicrobium was found in only one brine sample, namely B27, from the region of Messinia/Lakonia with relatively low abundance, i.e., 1.26% (Table S2B). The almost absence of the genus Cellulosimicrobium in our samples could be attributed to the different geographical regions, as the olive samples analyzed in the work of Papadimitriou et al. [26] originated from the island of Evoia in the eastern part of mainland Greece, thus strengthening the hypothesis that the conditions of different agroecosystems could have a decisive role in the diversity of microbial communities [48]. Moreover, it is noteworthy that the genus Staphylococcus was identified in four olive samples from the area of Messinia/Lakonia, with abundances ranging from 8% to 17% (Table S1A). As some species of this bacterial genus have been characterized as opportunistic pathogens [49], special attention should be given to the hygiene conditions in these specific fermentation vessels to avoid the adaptation of this species to the processing environment of olives. Consequently, intense cleaning would be advisable in these fermenters to avoid persistence of undesirable bacteria, although some researchers report that intense hygiene conditions may also negatively affect the establishment of a favorable autochthonous microbiota in the processing environment [18]. However, our results concerning the identification of opportunistic pathogens should be evaluated down to species or even strain level. The presence of the genus Staphylococcus in low abundances has been also confirmed by other researchers in directly brined green cracked Aloreña de Málaga olive fermentations [22], as well as in Spanish-style green olive fermentation of Manzanilla and Gordal varieties on an industrial scale [18,40,50]. No other food-related bacterial pathogens, such as Listeria, Clostridium, Salmonella and Escherichia were detected in any of the olive and brine samples analyzed from both regions. This finding is of paramount importance regarding the safety of table olives as the outbreaks due to table olive consumption are scarce. Our results are in accordance with other studies that employed metagenomics approaches to elucidate the bacterial community structure in Italian and Spanish olive varieties and reported the absence or low abundance of bacterial potential pathogens [17,20,22]. Furthermore, the family Enterobacteriaceae was found in several olives and brines from both regions, with Enterobacter being the most abundant genus. However, the majority of olive and brine samples analyzed presented relatively low abundances of this family, with the exception of two olive / samples, namely O2 and O16, from Messinia Lakonia and Aitoloakarnania, respectively. The presence of this family in the fermentation of table olives is rather habitual, with a well-known negative contribution in the quality of the final product [42]. However, it should be noted that, according to the classical microbiological analysis, no enterobacteria were counted in any of the analyzed samples. The absence of culturable enterobacteria can be attributed to the low pH values of brines and olives that ensure the safety of the final product. Thus, the identification of the family Enterobacteriaceae in several samples highlights the disadvantage of DNA-based metagenomics techniques, which is the amplification of DNA that may arise from dead or compromised cells [12]. Luckily, Proteobacteria, including Enterobacteriaceae, are inhibited in pH values below 4.5, and thus olives are protected from spoilage (gas pockets) due to the outgrowth of this microbial group during the fermentation [27,51]. Among the Gram-negative bacteria, the genera Pseudomonas and Sphingomonas were detected in all brine samples from both regions. The presence of the genus Pseudomonas has been initially detected on the surface of olives and the brines during black olive fermentation using culture dependent methods [39,52,53]. Recently, the genus Pseudomonas was also found in directly brined green olives of the Aloreña de Málaga variety using culture-independent techniques [14,22]. The presence of this proteolytic genus in table olives could have a negative effect on the overall stability and safety of the final product, either directly through the reduction of acidity in the brines and swelling of cans [54], or indirectly through the development of biogenic amines [14]. Moreover, the presence of the genus Sphingomonas has been reported for the first time in directly brined Alorena green olives fermented on an industrial scale in southern Spain [14], as well as in the Greek-type fermentation of Bella di Cerignola green olives [20]. The genus Sphingomonas is strictly aerobic, widely distributed in soil and water, as well as in plant root systems due to its ability to survive in low concentration of nutrients [55]. In addition, Microorganisms 2020, 8, 672 19 of 24 the genus Propionobacterium was also found in our olive samples, with varying abundances depending on the geographical region. This genus is associated with undesirable secondary fermentations, where the lactic acid produced during the main fermentation process is assimilated by propionic acid bacteria producing propionic acid, acetic acid and CO2, resulting in a gradual increase in pH and volatile acidity values, with the concurrent development of off-odors and undesirable sensory perceptions in table olives (zapateria) [27]. The control of salt concentration and pH values is thus necessary to eliminate the activity of these microorganisms, especially during the summer months [30]. The yeast/fungal microbiota of olive and brine samples from both geographical regions was less diverse compared to bacteria. Among the filamentous fungi identified, the genus Penicillium presented the highest abundances in the olive and brine samples from both regions (Table S2). This result is in agreement with a recent work [56] that investigated the fungal community associated with fermented black olives of Italian and Greek olive varieties, including Kalamata natural black olives. The authors isolated and identified 60 strains as Penicillium crustosum, Penicillium roqueforti, Penicillium paneum, Penicillium expansum, Penicillium polonicum and Penicillium commune. It has been documented that the presence of filamentous fungi on olives is associated with various mycotoxins (ochratoxin, aflatoxin B, and citrinin) in green and black olives [57–59], but their concentration is too low to pose a risk to human health [60]. With regard to the yeast community, Pichia, Saccharomyces, Ogatea and Millerozyma were the dominant genera identified in both brine and olive samples (Table S2C,D). The presence of the former two yeast genera is well-documented in the fermentation of natural black olives of Kalamata and Conservolea varieties [26,46,61,62]. The role of yeasts in table olive production could enhance or deteriorate the quality of the final product. On the positive side, they are present throughout the fermentation process and produce compounds with important sensory attributes determining the quality of fermented olives. On the negative side, they can also be spoilage microorganisms during olive fermentation and subsequent storage and packing causing gas pockets, swollen containers, cloudy brines and off-flavors and odors [63]. The presence of the genus Pichia in the fermentation may have a positive impact, as several strains have antioxidant activity that protects olives from oxidation and peroxide formation [64]. In addition, several strains belonging to this genus present strain-specific killer activity against spoilage yeasts and could thus influence fermentation by affecting the yeast association during the process [65,66]. Regarding the genus Saccharomyces, it has been reported that some strains may have lipase activity and increase the free fatty acid content in olives during fermentation [67], whereas other strains may synthesize bioactive compounds, such as carotenoids, tocopherols, citric acid and glutathione, with interesting antioxidant potential [68]. Additionally, we used the bacterial and yeast/fungal microbiota of olives and brines as “microbial fingerprints” to discriminate the samples according to their geographical origin. The outcome of the PLS-DA analysis illustrated that a satisfactory discrimination could be obtained between the two regions, although some degree of overlapping could not be avoided. The best discrimination was obtained by the ITS data of the brine samples (Figure 12). Based on the VIP values derived from PLS-DA analysis, the most discriminative genus among the yeasts/fungal communities was Pichia for both brine and olive samples that were highly associated (VIP > 2.5 and 2.0 for brine and olive samples, respectively) with the region of Messinia/Lakonia (Figure 12 and Figure S10). Regarding bacterial communities, the genus Lactobacillus was highly influential (VIP > 2.2) for the differentiation of olive samples from the area of Aitoloakarnania (Figure S8). These observations highlight the influence of the microclimate conditions on the microbiota of, not only fermented olives [40], but also of other food commodities as well [69,70].

5. Conclusions The results obtained from the present study reveal the complex microbial community structure of bacterial and yeast/fungal microbiota of 29 olives and brines of cv. Kalamata olives industrially fermented in the Messinia/Lakonia and Aitoloakarnania regions. To the best of our knowledge, this is the first systematic study on the microbial biogeography of Kalamata variety natural black olives Microorganisms 2020, 8, 672 20 of 24 from the two main producing regions of Greece, using the state-of-the-art approach of metagenomics. However, further studies on the occurrence of these taxa should be undertaken in other table olive industries in the same geographical regions to enhance our knowledge on the microbial communities present during Kalamata olives fermentation and their role on the quality and safety of the final product. Finally, our better understanding on the microbial composition of Kalamata table olives would result in the development of specific starter cultures to promote unique organoleptic characteristics in the final product of each region.

Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/2076-2607/8/5/672/s1. Figure S1: Microbiological counts of LAB and yeasts from the olives (O) and brine (B) of fermented cv. Kalamata natural black olives in the area of Messinia/Lakonia. Sample codes are indicated in Table1. Figure S2: Microbiological counts of LAB and yeasts from the olives (O) and brine (B) of fermented cv. Kalamata natural black olives in the area of Aitoloakarnania. Sample codes are indicated in Table1. Figure S3: pH values in olives and brines of fermented cv. Kalamata natural black olives in the area of Messinia/Lakonia. Sample codes are indicated in Table1. Figure S4: pH values in olives and brine of fermented cv. Kalamata natural black olives in the area of Aitoloakarnania. Sample codes are indicated in Table1. Figure S5: Salt concentration (%, w/v) in the brines of fermented cv. Kalamata natural black olives from the areas of Messinia/Lakonia (A) and Aitoloakarnania (B). Sample codes are indicated in Table1. Figure S6: Rarefaction curves of species richness for bacteria communities in olive (A) and brine (B) samples, as well as for yeast/fungal microbiota in olives (C) and brines (D). Samples are colored according to the two geographical regions, i.e., pink for Aitoloakarnania and light blue for Messinia/Lakonia. Figure S7: Plot of scores (A) and loadings (B) of the first two latent variables of the PLS-DA model of the bacterial community built on the olive samples of fermented cv. Kalamata olive fermentations in the region of Aitoloakarnania (A) and Messinia/Lakonia (M). Figure S8: Most influential bacterial species from olive samples of fermented cv. Kalamata olives based on VIP scores from PLS-DA analysis from the region of Aitoloakarnania (A) and Messinia/Lakonia (M). Color bars indicate intensity of variable in the respective group., Figure S9: Plot of scores (A) and loadings (B) of the first two latent variables of the PLS-DA model of the fungal community built on the olive samples of fermented cv. Kalamata olive fermentations in the region of Aitoloakarnania (A) and Messinia/Lakonia (M). Figure S10: Most influential fungal species from olive samples of fermented cv. Kalamata olives based on VIP scores from PLS-DA analysis from the region of Aitoloakarnania (A) and Messinia/Lakonia (M). Color bars indicate intensity of variable in the respective group. Table S1A: Abundance of dominant bacterial families (>1% abundance) identified in olive samples from Messinia/Lakonia and Aitoloakarnania regions. Table S1B: Abundance of dominant bacterial families (>1% abundance) identified in brine samples from Messinia/Lakonia and Aitoloakarnania regions. Table S1C: Abundance of dominant yeast/fungal families (>1% abundance) identified in olive samples from Messinia/Lakonia and Aitoloakarnania regions, Table S1D: Abundance of dominant yeast/fungal families (>1% abundance) identified in brine samples from Messinia/Lakonia and Aitoloakarnania regions. Table S2A: Abundance of dominant bacterial genera (>1% abundance) identified in olive samples from Messinia/Lakonia and Aitoloakarnania regions. Table S2B: Abundance of dominant bacterial genera (>1% abundance) identified in brine samples from Messinia/Lakonia and Aitoloakarnania regions. Table S2C: Abundance of dominant yeast/fungal genera (>1% abundance) identified in olive samples from Messinia/Lakonia and Aitoloakarnania regions. Table S2D: Abundance of dominant yeast/fungal genera (>1% abundance) identified in brine samples from Messinia/Lakonia and Aitoloakarnania regions. Author Contributions: Conceptualization, E.Z.P. and E.T.; Data curation, M.K. and E.Z.P.; Formal analysis, A.T.; Funding acquisition, E.Z.P. and E.T.; Methodology, M.K. and E.Z.P.; Project administration, E.Z.P.; Software, M.K. and E.Z.P.; Supervision, E.Z.P. and E.T.; Validation, M.K. and E.Z.P.; Visualization, M.K. and E.Z.P.; Writing—original draft, M.K. and E.Z.P.; Writing—review & editing, M.K., E.Z.P. and E.T. All authors have read and agreed to the published version of the manuscript. Funding: This research has been financed by Greek national funds through the Public Investments Program (PIP) of General Secretariat for Research & Technology (GSRT), under the Emblematic Action “The Olive Road” (project code: 2018ΣE01300000). Acknowledgments: The authors would like to thank the companies AGROVIM S.A. and Mani Foods in the area of Messinia as well as the company GAIA S.A. and the Table Olive Quality Centre in the area of Aitoloakarnania for providing the samples of cv. Kalamata natural black olives. 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. Microorganisms 2020, 8, 672 21 of 24

References

1. IOC, International Olive Council. World Table Olive Figures. 2020. Available online: https://www.internatio naloliveoil.org/wp-content/uploads/2020/01/OT-W901-29-11-2019-P.pdf (accessed on 3 March 2020). 2. DOEPEL, Interprofessional Association for table Olives. 2020. Available online: https://olivetreeroute.gr/wp -content/uploads/Studies_Publications_017a.pdf (accessed on 3 March 2020). 3. IOC, International Olive Council. Trade Standard Applying to Table Olives; International Olive Council: Madrid, Spain, 2004. 4. Grounta, A.; Tassou, C.C.; Panagou, E.Z. Greek-style table olives and their functional value. In Olives and Olive Oil as Functional Foods; Kiritsakis, A., Shahidi, F., Eds.; Wiley: Oxford, UK, 2017; pp. 325–342. 5. Botta, C.; Cocolin, L. Microbial dynamics and biodiversity in table olive fermentation: Culture-dependent and –independent approaches. Front. Microbiol. 2012, 3, 245. [CrossRef][PubMed] 6. Perpetuini, G.; Prete, R.; García-González, N.; Khairul Alam, M.; Corsetti, A. Table olives. More than a fermented food. Foods 2020, 9, 178. [CrossRef] 7. Giraffa, G. Studying the dynamics of microbial populations during food fermentation. FEMS Microbiol. 2004, 28, 251–260. [CrossRef][PubMed] 8. Riesenfeld, C.S.; Schloss, P.D.; Handelsman, J. Metagenomics: Genomic analysis of microbial communities. Ann. Rev. Genet. 2004, 38, 525–552. [CrossRef][PubMed] 9. Justé, A.; Thomma, B.P.H.J.; Lievens, B. Recent advances in molecular techniques to study microbial communities in food-associated matrices and processes. Food Microbiol. 2008, 25, 745–761. [CrossRef] 10. Kergourlay, G.; Taminiau, B.; Daube, G.; Champomier Verges, M.C. Metagenomic insights into the dynamics of microbial communities. Int. J. Food Microbiol. 2015, 213, 31–39. [CrossRef] 11. Ercolini, D. High-throughput sequencing and metagenomics: Moving forward in the culture-independent analysis of food microbial ecology. Appl. Environ. Microbiol. 2013, 79, 3148–3155. [CrossRef] 12. Cocolin, L.; Ercolini, D. Zooming into food-associated microbial consortia: A ‘cultural’ evolution. Curr. Opin. Food Sci. 2015, 2, 43–50. [CrossRef] 13. Ferrocino, I.; Cocolin, L. Current perspectives in food-based studies exploiting multi-omics approaches. Curr. Opin. Food Sci. 2017, 13, 10–15. [CrossRef] 14. Abriouel, H.; Benomar, N.; Lucas, R.; Gálvez, A. Culture-independent study of the diversity of microbial populations in brines during fermentation of naturally fermented Aloreña green table olives. Int. J. Food Microbiol. 2011, 144, 487–496. [CrossRef] 15. Muccilli, S.; Caggia, C.; Randazzo, C.L.; Restuccia, C. Yeast dynamics during the fermentation of brined green olives treated in the field with kaolin and Bordeaux mixture to control the olive fruit fly. Int. J. Food Microbiol. 2011, 148, 15–22. [CrossRef][PubMed] 16. Randazzo, C.L.; Ribbera, A.; Pitino, I.; Romeo, F.V.; Caggia, C. Diversity of bacterial population of table olives assessed by PCR-DGGE analysis. Food Microbiol. 2012, 32, 87–96. [CrossRef][PubMed] 17. Cocolin, L.; Alessandria, V.; Botta, C.; Gorra, R.; De Filippis, F.; Ercolini, D.; Rantsiou, K. NaOH-debittering induces changes in bacterial ecology during table olives fermentation. PLoS ONE 2013, 8, e69074. [CrossRef] [PubMed] 18. Lucena-Padrós, H.; Caballero-Guerrero, B.; Maldonado-Barragán, A.; Ruiz-Barba, J.L. Genetic diversity and dynamics of bacterial and yeast strains associated to Spanish-style green table olive fermentations in large manufacturing companies. Int. J. Food Microbiol. 2014, 190, 72–78. [CrossRef][PubMed] 19. Lucena-Padrós, H.; Jiménez, E.; Maldonado-Barragán, A.; Rodríguez, J.M.; Ruiz-Barba, J.L. PCR-DGGE assessment of the bacterial diversity in Spanish-style green table olive fermentations. Int. J. Food Microbiol. 2015, 205, 47–53. [CrossRef] 20. De Angelis, M.; Campanella, D.; Cosmai, L.; Summo, C.; Rizzello, C.G.; Caponio, F. Microbiota and metabolome of un-started and started Greek-type fermentation of Bella di Cerignola table olives. Food Microbiol. 2015, 52, 18–30. [CrossRef] 21. Arroyo-López, F.N.; Medina, E.; Ruiz-Bellido, M.A.; Romero-Gil, V.; Montes-Borrego, M.; Landa, B.B. Enhancement of the knowledge on fungal communities in directly brined Aloreña de Málaga green olive fermentations by metabarcoding analysis. PLoS ONE 2016, 11, e0163135. [CrossRef] Microorganisms 2020, 8, 672 22 of 24

22. Medina, E.; Ruiz-Bellido, M.A.; Romero-Gil, V.; Rodríguez-Gómez, F.; Montes-Borrego, M.; Landa, B.B.; Arroyo-López, F.N. Assessment of the bacterial community in directly brined Aloreña de Málaga table olive fermentations by metagenetic analysis. Int. J. Food Microbiol. 2016, 236, 47–55. [CrossRef] 23. Randazzo, C.L.; Todaro, A.; Pino, A.; Pitino, I.; Corona, O.; Caggia, C. Microbiota and metabolome during controlled and spontaneous fermentation of Nocellara Etnea table olives. Food Microbiol. 2017, 65, 136–148. [CrossRef] 24. Rodríguez-Gómez, F.; Ruiz-Bellido, M.Á.; Romero-Gil, V.; Benítez-Cabello, A.; Garrido-Fernández, A.; Arroyo-López, F.N. Microbiological and physicochemical changes in natural green heat-shocked Aloreña de Málaga table olives. Front. Microbiol. 2017, 8, 2209. [CrossRef] 25. De Castro, A.; Sánchez, A.H.; López-López, A.; Cortés-Delgado, A.; Medina, E.; Montaño, A. Microbiota and Metabolite Profiling of Spoiled Spanish-Style Green Table Olives. Metabolites 2018, 8, 73. [CrossRef] [PubMed] 26. Papadimitriou, K.; Kazou, M.; Alexandraki, V.; Chavella, G.; Koutsoumpou, M.; Paximadaki, A.; Anastasiou, R.; Georgalaki, M.; Tsakalidou, E. Metagenomics analysis of naturally fermented Greek table olives. In Proceedings of the 12th Symposium on Lactic Acid Bacteria, Egmond aan Zee, The Netherlands, 27–31 August 2017. 27. Lanza, B. Abnormal fermentations in table-olive processing: Microbial origin and sensory evaluation. Front. Microbiol. 2013, 4, 91. [CrossRef][PubMed] 28. Heperkan, D. Microbiota of table olive fermentations and criteria of selection for their use as starters. Front. Microbiol. 2013, 4, 143. [CrossRef][PubMed] 29. Panagou, E.Z. Effect of different packing treatments on the microbiological and physicochemical characteristics of untreated green olives of the Conservolea cultivar. J. Sci. Food Agric. 2004, 84, 757–764. [CrossRef] 30. Garrido-Fernández, A.; Fernández-Díez, M.J.; Adams, R.M. Table Olives: Production and Processing, 1st ed.; Chapman and Hall: London, UK, 1997. 31. Pitcher, D.G.; Saunders, N.A.; Owen, R.J. Rapid extraction of bacterial genomic DNA with guanidium thiocyanate. Lett. Appl. Microbiol. 1989, 8, 151–156. [CrossRef] 32. Kopsahelis, N.; Nisiotou, A.; Kourkoutas, Y.; Panas, P.; Nychas, G.J.E.; Kanellaki, M. Molecular characterization and molasses fermentation performance of a wild yeast strain operating in an extremely wide temperature range. Bioresour. Technol. 2009, 100, 4854–4862. [CrossRef][PubMed] 33. Papademas, P.; Aspri, M.; Mariou, M.; Dowd, S.E.; Kazou, M.; Tsakalidou, E. Conventional and omics approaches shed light on Halitzia cheese, a long-forgotten white-brined cheese from Cyprus. Int. Dairy J. 2019, 98, 72–83. [CrossRef] 34. Dowd, S.E.; Callaway, T.R.; Wolcott, R.D.; Sun, Y.; McKeehan, T.; Hagevoort, R.G.; Edrington, T.S. Evaluation of the bacterial diversity in the feces of cattle using 16S rDNA bacterial tag-encoded FLX amplicon pyrosequencing (bTEFAP). BMC Microbiol. 2008, 8, 1–8. [CrossRef] 35. Thukral, A.K. A review on measurement of alpha diversity in biology. Agric. Res. J. 2017, 54, 1–10. [CrossRef] 36. Moon, J.Y.; Jung, H.J.; Chung, B.C.; Choi, M.H. Heat-map visualization of gas chromatography-mass spectrometry based quantitative signatures on steroid metabolism. J. Am. Soc. Mass Spectr. 2009, 20, 1626–1637. [CrossRef] 37. Chong, J.; Wishart, D.S.; Xia, J. Using MetaboAnalyst 4.0 for comprehensive and interactive metabolomics data analysis. Curr. Protoc. Bioinf. 2019, 68, e86. [CrossRef][PubMed] 38. Afanador, N.L.; Tran, T.N.; Buydens, L.M.C. An assessment of the jackknife and boostrap procedures on uncertainty estimation in the variable importance in the projection metrics. Chemometr. Intell. Lab. Syst. 2014, 137, 162–172. [CrossRef] 39. Tassou, C.C.; Panagou, E.Z.; Katsaboxakis, K.Z. Microbiological and physicochemical changes of naturally black olives fermented at different temperatures and NaCl levels in the brines. Food Microbiol. 2002, 19, 605–615. [CrossRef] 40. Lucena-Pardós, H.; Ruiz-Barba, J.L. Microbial biogeography of Spanish-style green olive fermentations in the province of Seville, Spain. Food Microbiol. 2019, 82, 259–268. [CrossRef] 41. Lucini, L.; Rocchetti, G.; Trevisan, M. Extending the concept of terroir from grapes to other agricultural commodities: An overview. Curr. Opin. Food Sci. 2020.[CrossRef] 42. Sánchez-Gómez, A.H.; García-García, P.; Rejano Navarro, L. Elaboration of table olives. Grasas Y Aceites 2006, 57, 86–94. [CrossRef] Microorganisms 2020, 8, 672 23 of 24

43. Grounta, A.; Doulgeraki, A.I.; Nychas, G.J.E.; Panagou, E.Z. Biofilm formation on Conservolea natural black olives during single and combined inoculation with a functional Lactobacillus pentosus starter culture. Food Microbiol. 2016, 56, 35–44. [CrossRef] 44. Domínguez-Manzano, J.; Olmo-Ruiz, C.; Bautista-Gallego, J.; Arroyo-Lopez, F.N.; Garrido-Fernández, A.; Jiménez-Díaz, R. Biofilm formation on abiotic and biotic surfaces during Spanish-style green table olive fermentation. Int. J. Food Microbiol. 2012, 157, 230–238. [CrossRef] 45. Sisto, A.; Lavermicocca, P. Suitability of a Lactobacillus paracasei strain as a starter culture in olive fermentation and development of the innovative patented product “probiotic table olives”. Front. Microbiol. 2012, 3, 1–5. [CrossRef] 46. Bleve, G.; Tufariello, M.; Durante, M.; Grieco, F.; Ramires, F.A.; Mita, G.; Tasioula-Margari, M.; Logrieco, A.F. Physico-chemical characterization of natural fermentation process of Conservolea and Kalamata table olives and development of a protocol for the pre-selection of fermentation starters. Food Microbiol. 2015, 46, 368–382. [CrossRef] 47. Chytiri, A.; Tasioula-Margari, M.; Bleve, G.; Kontogianni, V.G.; Kallimanis, A.; Kontominas, M.G. Effect of different inoculation strategies of selected yeast and LAB cultures on Conservolea and Kalamata table olives considering phenol content, texture, and sensory attributes. J. Sci. Food Agric. 2020, 100, 926–935. [CrossRef] 48. Martiny, J.B.H.; Bohannan, B.J.M.; Brown, J.H.; Colwell, R.K.; Fuhrman, J.A.; Green, J.L.; Horner-Devine, M.C.; Kane, M.; Krumins, J.A.; Kuske, C.R.; et al. Microbial biogeography: Putting microorganisms on the map. Nat. Rev. Microbiol. 2006, 4, 102–112. [CrossRef][PubMed] 49. Doulgeraki, A.; Di Ciccio, P.; Lanieri, A.; Nychas, G.J.E. Methicillin-resistant food-related Staphylococcus aureus: A review of current knowledge and biofilm formation for future studies and applications. Res. Microbiol. 2017, 168, 1–15. [CrossRef] 50. Benítez-Cabello, A.; Bautista-Gallego, J.; Garrido-Fernández, A.; Rantsiou, K.; Cocolin, L.; Jiménez-Díaz, R.; Arroyo-López, F.N. RT-PCR-DGGE analysis to elucidate the dominant bacterial species of industrial Spanish-style green table olive fermentations. Front. Microbiol. 2016, 7, 1291. [CrossRef][PubMed] 51. Perricone, M.; Bevilacqua, A.; Corbo, M.R.; Sinigaglia, M. Use of Lactobacillus plantarum and glucose to control the fermentation of “Bella di Cerignola” table olives, a traditional variety of Apulian region (Southern Italy). J. Food Sci. 2010, 75, M430–M436. [CrossRef][PubMed] 52. Nychas, G.J.E.; Panagou, E.Z.; Parker, M.L.; Waldron, K.W.; Tassou, C.C. Microbial colonization of naturally black olives during fermentation and associated biochemical activities in the cover brines. Lett. Appl. Microbiol. 2002, 34, 173–177. [CrossRef][PubMed] 53. Ercolini, D.; Villani, F.; Aponte, M.; Mauriello, G. Fluoresence in situ hybridization detection of Lactobacillus plantarum group to be used in natural fermentations. Int. J. Food Microbiol. 2006, 112, 291–296. [CrossRef] 54. Harmon, S.M.; Kautter, D.A.; McKee, C. Spoilage of anchovy-stuffed olives by heterofermentative lactobacilli. J. Food Saf. 1987, 8, 205–210. [CrossRef] 55. Steinberg, J.P.; Burd, E.M. Other Gram-negative and Gram-variable bacilli. Principles and Practices of Infectious Diseases; Bennett, J.E., Dolin, R., Blaser, M.J., Eds.; Elsevier: Philadelphia, PA, USA, 2015; pp. 2667–2683. 56. Bavaro, S.L.; Susca, A.; Frisvad, J.C.; Tufariello, M.; Chytiri, A.; Perrone, G.; Mita, G.; Logrieco, A.F.; Bleve, G. Isolation, characterization, and selection of molds associated to fermented black table olives. Front. Microbiol. 2017, 8, 1356. [CrossRef]

57. Ghitakou, S.; Koutras, K.; Kanellou, E.; Markaki, P. Study of aflatoxin B1 and ochratoxin A production by natural microflora and Aspergillus parasiticus in black and green olives of Greek origin. Food Microbiol. 2006, 23, 612–621. [CrossRef] 58. Heperkan, D.; Meric, B.E.; Sismanoglu, C.; Dalkiliç, C.; Güler, F.K. Mycobiota, mycotoxigenic fungi, and citrinin production in black olives. Adv. Exp. Med. Biol. 2006, 571, 203–210. [PubMed] 59. Franzetti, L.; Scarpellini, M.; Vecchio, A.; Planeta, D. Microbiological and safety evaluation of green table olives marketed in Italy. Ann. Microbiol. 2011, 61, 843–851. [CrossRef] 60. Medina-Pradas, E.; Arroyo-Lóopez, F.N. Presence of toxic microbial metabolites in table olives. Front. Microbiol. 2015, 6, 873. [CrossRef][PubMed] 61. Nisiotou, A.A.; Chorianopoulos, N.; Nychas, G.-J.E.; Panagou, E.Z. Yeast heterogeneity during spontaneous fermentation of black Conservolea olivesin different brine solutions. J. Appl. Microbiol. 2010, 108, 396–405. [CrossRef][PubMed] Microorganisms 2020, 8, 672 24 of 24

62. Bonatsou, M.; Paramithiotis, S.; Panagou, E.Z. Evolution of yeast consortia during the fermentation of Kalamata natural black olives upon two initial acidification treatments. Front. Microbiol. 2018, 8, 2673. [CrossRef][PubMed] 63. Arroyo-López, F.N.; Querol, A.; Bautista-Gallego, J.; Garrido-Fernández, A. Role of yeasts in table olive production. Int. J. Food Microbiol. 2008, 128, 189–196. [CrossRef] 64. Hernández, A.; Martín, A.; Aranda, E.; Pérez-Nevado, F.; Górdoba, M.G. Identification and characterization of yeasts isolated from the elaboration of seasoned green table olives. Food Microbiol. 2007, 24, 346–351. [CrossRef] 65. Psani, M.; Kotzekidou, P. Technological characteristics of yeasts strains and their potential as starter adjuncts in Greek-style black olive fermentation. World J. Microbiol. Biotechnol. 2006, 22, 1329–1336. [CrossRef] 66. Hernández, A.; Martín, A.; Górdoba, M.G.; Benito, M.J.; Aranda, E.; Pérez-Nevado, F. Determination of killer activity in yeasts isolated from the elaboration of seasoned green table olives. Int. J. Food Microbiol. 2008, 121, 178–188. [CrossRef] 67. Ciafardini, G.; Zullo, B.A.; Cioccia, G.; Iride, A. Lipolytic activity of Williopsis californica and Saccharomyces cerevisiae in extra virgin olive oil. Int. J. Food Microbiol. 2006, 107, 27–32. [CrossRef] 68. Abbas, C.A. Production of antioxidants, aromas, colours, flavours, and vitamins, by yeasts. In Yeasts in Foods and Beverages; Querol, A., Fleet, H., Eds.; Springer: Berlin, Germany, 2006; pp. 285–334. 69. Ohshima, C.; Takahashia, H.; Insang, S.; Phraephaisarn, C.; Techaruvichit, P.; Khumthong, R.; Haraguchi, H.; Lopetcharat, K.; Keeratipibul, S. Next generation sequencing reveals predominant bacterial communities during fermentation of Thai fish sauce in large manufacturing plants. LWT-Food Sci. Technol. 2019, 114, 108375. [CrossRef] 70. Phewpan, A.; Phuwaprisirisan, P.; Takahashi, H.; Ohshima, C.; Lopetcharat, K.; Techaruvichit, P.; Keeratipibul, S. Microbial diversity during processing of Thai traditional fermented shrimp paste, determined by next generation sequencing. LWT-Food Sci. Technol. 2020, 122, 108989. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).