First Evidence of Metabolically Active Intracellular Bacteria in Saccharomyces cerevisiae

2.1. Chemicals and Reagents

AHL standards (C8-HSL, 3-oxo-C6-HSL, and 3-oxo-C10-HSL), mono- and polysaccharides, ethyl acetate, acetonitrile, methanol, acetic acid, 2-propanol, formic acid, 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal), gentamycin solution (30 μg mL^–1), chloramphenicol, and cycloheximide were purchased from Merk Life Science S.r.l (Milan, Italy). Reverse-phase silica gel and TLC silica gel plates were purchased from VWR International (Milan, Italy). Water for LC-MS analysis was obtained by a Milli-Q apparatus (Millipore, Milan, Italy).

2.2. Yeast Strains

Saccharomyces cerevisiae OS3 (DBVPG 6765; NCYC 3264) and S. cerevisiae V5 (M6–3) were the strains used in this study.

Strains were stored in Malt Extract Broth (Oxoid) containing 20% glycerol (Sigma) at −20 °C. Routine cultures were maintained in WL Nutrient Agar (Oxoid) slants at 4–6 °C. Working cultures were grown in YPD (10 g/L yeast extract, 20 g/L peptone, 20 g/L dextrose). Purity was tested by streaking on WL Nutrient Agar and simultaneously observing under a microscope.

2.3. Growth of S. cerevisiae OS3 and V5 in Chemically Defined Medium

Growth was monitored in 0.67% Yeast Nitrogen Base without amino acids (YNBwAA) (Sigma, Y0626) containing 20 g/L dextrose in triplicate. Water and dextrose solutions were sterilized by autoclave treatment; 10 × YNBwAA solutions were sterilized by filtration. Sterile solutions were mixed in sterile conditions to obtain the final medium (3 × 2.8 L total volume in a 5.0 L sterile Duram bottle containing a stirring bar). The strain OS3 was pre-grown in YPD broth, containing 100 mg/L chloramphenicol, at 30 °C until the OD600 nm (Honda spectrophotometer V10-Plus) reached 1.6–1.8, and then was inoculated in YNBwAA at a rate of 1%. The growth was monitored for 72 h at 30 °C under static conditions. Uninoculated medium (CN), inoculated medium at the beginning of fermentation (t₀ h), and at different times during the growth (t = 6, 11, 24, 48, and 72 h) were sampled. Before each sampling, the cultures were stirred at 200 rpm for 5 min. The growth was monitored by evaluating the OD and by total viable yeast counts on WL Nutrient Agar containing 100 mg/L chloramphenicol at 30 °C for 5 days. Bacterial contamination was evaluated by plating 0.1 mL on PCA (Oxoid) and on MRS agar (Oxoid) containing 40 mg/L of cycloheximide to inhibit Saccharomyces, and by simultaneously observing under a microscope (Nikon Eclipse E400). During fermentation, the pH was monitored by using a pH 60 VioLab (XS instruments), while residual glucose and produced ethanol, succinic and acetic acids, and glycerol were determined by high-performance liquid chromatography (HPLC) analyses as previously described. Samples for chemical analyses (200 mL each) were collected at t = 0, 1, 2, 3, 4, 5, 6, 12, 24, 48, and 72 h and immediately centrifuged 6000 rpm (4800×g) for 10 min (Rotor A8–50, Centrifuge NEYA8) to separate cells from fermented broth, and both were stored at −80 °C.

S. cerevisiae V5 strain was cultured as described for the OS3 strain (see above), but samples for chemical analyses were collected at the key time points (t = 0, 6, and 24h) for the determination of tyrosol content, the detection and identification of AHLs, and enzymatic activities.

2.4. Microscope Observation

Two different approaches were used for the microscope observations. Yeast cultures, stored at −20 °C, were grown in YPD medium containing 100 mg/L chloramphenicol and then inoculated on different media: WL Nutrient Agar (WLNA, Oxoid) supplemented with 40 mg/L of cycloheximide; Nutrient Agar (NA, Oxoid) supplemented with 40 mg/L of cycloheximide; Baird Parker Agar Base supplemented with Egg’s Yolk Tellurite Steril Emulsion (BP-EYT) (Schalau); Pseudomonas agar Base with CFC Selective supplement (PA-CFC) (Schalau); Yeast Pre-Sporulation Medium (YPSM: yeast extract 10 g/L; dextrose, 100 g/L; potassium acetate, 20 g/L; agar, 20 g/L); Yeast Sporulation Medium (YSM: yeast extract 10 g/L; dextrose, 0.1 g/L; potassium acetate, 10 g/L; agar, 20 g/L). All cultures were incubated at 30 °C for 5–10 days. All grown cultures were analyzed by light microscopy, and images and video were captured by VisiCam 10.0 (VWR International).

Six μL of the cell suspension was also placed on a microscope slide, coverslips were placed on the slides, and the samples were observed under an optical microscope (Axio Imager D2Carl Zeiss Microscopy, Jena, Germany) fitted with a 63× and 100× objective lens and an attached camera (Axiocam MRmCarl Zeiss Microscopy, Jena, Germany). Fiji was used to generate the scale bar. Moreover, with the attempt to verify the contamination by anaerobic bacteria, yeast colonies grown on BP and YSM, in which Moving Bacteria-Like bodies (MBLBs) were present, were inoculated in (a) YCFAGSC medium supplemented with 40 mg/L of cycloheximide; (b) BBA-HVK, Brucella Blood Agar with Hemin and Vitamin K1 (Becton Dickinson GmbH) with 40 mg/L of cycloheximide; (c) CMM, Chopped Meat Medium (DSMZ no. 78) supplemented with 40 mg/L of cycloheximide. All media were incubated anaerobically at 37 °C.

2.5. Starvation Protocol

The used protocol was as described by Yue et al. Briefly, cells were first pulled out from glycerol stocks in solid YPD (23 °C, overnight); subsequently, they were streaked to confirm that there was no bacterial contamination. A single colony was cultivated in liquid YPD (2% dextrose, 1% yeast extract, 2% peptone) for 24 h, at 23 °C. Afterward, cells were diluted 1:50 in 10 mL of presporulation medium (YPA: 2% potassium acetate, 1% yeast extract, 2% peptone) and incubated for 48 h at 23 °C (shaking 220 rpm). Presporulated cells were finally resuspended in 25 mL of Sporulation medium (2% KAc) in 250 mL flasks. The flasks were kept at 23 and 30 °C and shaken at 220 rpm for 96 h. The samples were observed under an optical microscope in each of the three steps of the protocol.

2.6. MBLBs Visualization by DNA Dyes

To localize the MBLBs inside the yeast cells, a single colony of each strain (OS3 and V5), was inoculated in 10 mL of YPD. The tubes were incubated for up to 6 days at 23 °C under shaking conditions (220 rpm). The samples were stained with the LIVE/DEAD BacLight Bacterial Viability Kit (Cat. No. L7012) according to the manufacturer’s instructions. Briefly, the cells were washed with 0.85% NaCl and supplemented with 3 μL of the dye mixture for each milliliter of the yeast suspension. The cell suspension was mixed thoroughly and incubated in the dark for 15 min before the microscopy analysis. A fluorescent microscope (Axio Imager D2) fitted with a 63× and 100× objective lens and an attached camera (Axiocam MRm) was used for visualization and image capture. Photographs were taken at different time intervals to reveal the movement of the BLBs. Fiji was used to process the images.

2.7. Extraction of Cell-Free Medium

The spent medium (200 mL) from S. cerevisiae OS3 and V5 strains cultures at different growth stages (t = 0, 6, 12, 24, 48, and 72 h) was centrifuged as previously reported. The supernatants were then extracted with ethyl acetate (1:1 v/v; three times) and subsequently dried under vacuum at T < 40 °C. For each growth time, three media extracts were collected and investigated by MS spectrometry and NMR spectroscopy. The extracts were dissolved in methanol to a final concentration of 0.6 mg/mL and directly analyzed by UPLC-MS/MS. Additionally, the extracts were tested for the detection of QS signal molecules by means of the TLC-overlay assay.

2.8. QSMs Identification and Characterization

2.9. Enzymatic Investigation

Cell pellets of OS3 and V5 strains isolated from the culture medium whole were suspended in PBS (Phosphate Buffered Saline) (1×, pH 7.4), with a ratio of 1:7 (mg of humid cells/μL of used buffer) and disrupted as previously described by Tramice et al. Cells were disrupted by sonication for 3 min and homogenized by ULTRA-TURRAX for a further 2 min in an ice–water bath. Cell debris was removed by centrifugation at 5524×g at 4 °C for 45 min.

The cell-free supernatants from different growth conditions of S. cerevisiae OS3 and V5 strains were tested for their protein content (Table S1), stored at −80 °C, and used as crude enzyme preparations in enzymatic digestions. Protein concentration was routinely estimated using the Bio-Rad Protein System, and bovine serum albumin was used as standard.

The crude protein enzymatic extracts were tested for the presence of α- and β-glucanase activities by using several substrates such as starch, amylose, amylopectin, pullulan, glycogen, β-glucan from barley, curdlan, laminarin, laminaribiose, and laminaripentaose. Solutions of 5 mg/mL of each substrate in 100 mM sodium acetate buffer at pH 5 were incubated at 40 °C under magnetic stirring in the presence of a fixed protein amount of the four extracellular enzymatic solutions obtained from different growth conditions.

All reactions were carried out using 60 μg of total protein per milligram of reagent. Reactions were monitored over time (0–48 h) by TLC analysis. The TLC solvent system used was EtOAc:AcOH:2-Propanol:HCOOH:H₂O, 25:10:5:1:15 by vol. Compounds on TLC plates were visualized under UV light or charring with α-naphthol reagent.

The enzymatic digestions of pullulan, amylopectin, and laminarin were stopped in an ice-bath at 48 h for 5 min, and 0.100 mL of each sample was assayed for the glucose equivalent production (DNS assay). The tests were conducted in triplicate, and the calibration curve of glucose was elaborated in a concentration interval of 0.05:1 mg mL^–1. Moreover, fasta36 was used to search for pullulanase and laminarase sequences in the strains data set at EVOMICS, and in the pangenome.

2.10. DNA Isolation and Meta-Taxonomic Analysis

The yeast culture showing MBLBs was grown on YSM and BP-YET for 5 days. Cells were harvested from agar plates and washed with sterile water, and DNA was isolated using a Nucleo Spin Food kit (Macherey-Nagel) following the supplier's instructions. Bacterial communities were assessed by HTS of the amplified V3–V4 regions within the 16S rRNA gene (∼460 bp). PCR was carried out with primers (S-D-Bact-0341-b-S-17/S-D-Bact0785-a-A-21) connecting with barcodes as previously described by Aponte et al. PCR products with proper sizes were selected by 2% agarose gel electrophoresis. The same amount of PCR products from each sample was pooled, end-repaired, A-tailed, and further ligated with Illumina adapters. The library was checked with Qubit and real-time PCR for quantification and a bioanalyzer for size distribution detection. Quantified libraries were pooled and sequenced on a paired-end Illumina platform to generate 250 bp paired-end raw reads. Paired-end reads were joined using FLASH. The DADA2 method was used for noise reduction. ASVs (Amplicon Sequence Variants) were further filtered using QIIME2 software (Version QIIME2–202202) and identified using the Silva Database 138.1.

2.11. Alignment of Lux Genes against S. cerevisiae Assemblies

To exclude the presence of Lux genes as HGT integrated inside yeast genomes, we put together all the long-reads assemblies two studies , among which is also available OS3 (as DBVPG6765), and we used both Blastn and Tblastx (BLAST suite v. 2.13.0) to search for the presence of LuxR (sequence searched: NC_011186.1:1162619-1163371) and acyl-homoserine-lactone synthase (sequences searched: NZ_SILH01000001.1:2962926-2963591, NZ_MRDH01000005.1:140513-141178, NZ_CP098803.1:3577150-3577815, NZ_JAAXQT010000001.1:c446238-445573, NZ_CP071612.1:3025255-3025920, NZ_CP071454.1:424262-424927, NZ_FYAH01000010.1:25839-26489, NZ_CP031495.1:286275-287471, NZ_CP009467.1:2502877-2504076, NZ_AP025466.1:c1546624-1545446). Reads for DBVPG6765 were also mapped with the Burrow-Wheeler Aligner (BWA) software against a multi-FASTA, containing Saccharomyces representatives and the previously mentioned LuxR and acyl-homoserine-lactone synthase.

2.12. Statistical Analysis

A One-way repeated measures ANOVA test was conducted to compare the amount of tyrosol, C8-AHL, and C10-AHL expressed in S. cerevisiae samples collected at different time points during the fermentation process. Once the normal distribution of the data set, Mauchly’s test of sphericity was applied to the data, and Greenhouse-Geisser (ε < 0.75) or Huynh-Feldt (ε near or above 0.75) corrections were used for the within-subjects effect in case of sphericity violation. Multiple comparisons were assessed with the Bonferroni correction method (Table S2). Statistical tests were elaborated with the OriginPro 9.1 software package (OriginLab Corporation, Northampton, USA) and R software.

3.1. Growth Monitoring of S. cerevisiae OS3 and V5 in Chemically Defined Medium

S. cerevisiae OS3 (DBVPG 6765; NCYC 3264) strain was originally isolated from lychee fruit (Indonesia). The monitoring of its growth in Yeast Nitrogen Base without amino acids (YNBwAA) medium was performed. Figure shows the growth kinetics of the OS3 strain and the production of the main metabolites during 72 h of fermentation at 30 °C. In the first 12 h, the strain reached the stationary phase. In fact, viable counts rapidly increased from 5.71 ± 0.15 log CFU/mL at 0 h to 7.44 ± 0.24 log CFU/mL after 12 h. The last level was maintained constantly until 48 h to slightly decrease at 72 h, reaching 6.97 ± 0.15 log CFU/mL (Figure A). An opposite behavior was observed for the pH: a rapid decrease in the first 12 h (from 5.13 to 2.86 units) was recorded with a stationary value for the rest of the monitoring time (Figure A). The glucose of the medium was depleted in the first 24 h, and consequently, ethanol reached its maximum level (about 8.0 g/L) (Figure B). In this phase, glycerol and succinic acid also reached their maximum production levels, i.e., =1.3 and 0.6 g/L, respectively. The strain V5, isolated from Passito wine (Italy), monitored for 24 h in the same medium, showed similar growth performances reaching an optical density of 1.764 OD600, a viable count of 7.35 ± 0.05 log CFU/mL, and a pH of 2.83 (data not shown) after 24 h of fermentation.

Contamination from aerobic bacteria during the fermentation process was carefully avoided and checked by analyzing 0.1 mL of each replica batch on PCA and MRS agar, both containing cycloheximide. Moreover, a microscope examination of the fermentation broth did not reveal the presence of any possible bacteria.

3.2. Extraction and Identification of Yeast and Bacterial Signaling Molecules

The cell-free supernatant extracts of S. cerevisiae OS3 and V5 strains at different growth time points (0, 6, 12, 24, 48, and 72 h for the OS3 strain and 0, 6, and 24 h for the V5 strain) were analyzed by NMR spectroscopy to detect secondary metabolites involved in inter- and intraspecies communication. A comparison with NMR spectra of pure tyrosol, farnesol, 2-phenylethanol, 2-phenoxyethanol, and tryptophol secured the exclusive presence of tyrosol as the fungal QS metabolite recovered in the ethyl acetate extract spectra of the spent medium. In the ¹H NMR spectra, the aromatic proton signals of tyrosol were easily detected at 7.02 ppm (doublet, J 8.89 Hz) and 6.69 ppm (doublet, J 8.89 Hz); the alchilic proton signals (2.71, triplet, J 7.31 Hz) and 3.67 (triplet, J 7.31 Hz) ppm overlapped with other signals and were assigned by homonuclear and heteronuclear 2D experiments and by comparison with spectrum of tyrosol standard (Figures S1–S5). Significant levels of tyrosol were detected in the OS3 strain starting after 6 h of growth (49.53 ± 14.29 nM), which increased over time to 230.58 ± 60.61 nM at 48 h and 309.14 ± 15.45 nM at 72 h (Figure ). Differently, the tyrosol concentration evaluated in V5 was negligible at 6 h, and of 144.30 ± 15.30 nM at 24 h of growth.

To evaluate possible interference of fungal QSMs with QS system of bacteria, the cell-free medium extracts of S. cerevisiae OS3 and V5 strains were analyzed at different time (0, 6, 12, 24, 48, and 72 h for OS3 strain and 0, 6, and 24 h for V5 strain) using the A. tumefaciens NTL4 (pZLR4) AHL bioreporter in the TLC-overlay assay. Several extracts were found to activate the AHL bioreporter starting from t = 12 h (Figure A,B). The first hypothesis was the presence of AHL-mimicking compounds. This behavior has already been described in the literature, where biosensors for AHL detection were activated by diketopiperazines, as shown in the study of Tommonaro et al. The samples exhibiting positive results in the overlay-assay TLC were analyzed by UPLC-mass spectrometry and, surprisingly, two AHLs were identified, N-octanoyl l-homoserine lactone (C8-HSL) and N-decanoyl l-homoserine lactone (C10-HSL) (Figures and S6–S8), two typical QSMs produced by bacteria. C8-HSL was the most accumulated molecule; however, the production of both signal molecules increased over time, following the same trend. At t = 48 h, there was the maximum production of compounds with concentrations of 0.835 ± 0.287 nM for C8-AHL and 0.174 ± 0.059 nM for C10-AHL (Figure ).

3.3. Whole-Genome Analysis, Microscope Observations, and Metataxonomic Analysis

AHL-dependent QS are typically synthesized by a LuxI-type enzyme and detected by LuxR-type transcriptional regulators. We therefore searched for the presence of LuxR (sequence searched: NC_011186.1:1162619-1163371) and acyl-homoserine-lactone synthase in the S. cerevisiae OS3 genome with the aim of detecting AHL synthase gene (luxI) and a transcriptional regulator gene (luxR). For this, we used the Basic Local Alignment Search Tool suite (BLASTN and TBLASTX routines, v. 2.13.0). The alignment of lux genes against the S. cerevisiae genome did not match any sequence. Therefore, considering the strictly applied experimental conditions of growth to rule out possible external contamination, the production of AHLs supports the hypothesis of the coexistence of bacteria in S. cerevisiae OS3 and V5 strains.

Suspensions of cells cultivated in liquid YPD and in the presporulation media YPA, YPSM, YSM, and BP were analyzed by an optical microscope. The analysis showed MBLBs inside the cells of S. cerevisiae OS3 and V5 strains grown on YSM medium, as the bodies occupy different positions in the A and B pictures of Figure (1000× magnification), taken a few seconds apart (red arrows). This is more evident in the Supporting Information, Video S1. By using presporulation YPA medium (stress conditions), the presence of MBLBs became greater and more evident (Figure , pictures C and D).

After 48 h of growth in YPA medium, as stress conditions increased, MBLBs exited from the OS3 cells (black dots in Figure C), leaving them empty. In potassium acetate, after 96 h at 30 °C, under severe stress conditions caused by the absence of nutrients, we observed that the OS3 yeast cells were surrounded by MBLBs, with the occurrence of dark dots (indicated by the red arrow) within the cell. Dark dots surrounding the vacuole do not move and might consist of stress granules. Moreover, the microscopic analysis of YPA-induced S. cerevisiae cells of some strains from our collection (including V5 and EC1118) showed, in many cases, the presence of MBLBs inside the vacuoles (data not shown).

Fluorescent microscopy (LIVE/DEAD BacLight) showed the presence of green-stained, small, round, moving BLBs within the yeast’s vacuole (white arrows) of strains OS3 and V5 (Figure S9).

The Sanger sequencing results of amplified 16S rDNA, using DNA isolated from OS3 yeast presenting MBLBs as template, showed confused chromatograms and an overlapping peak, suggesting a mix of 16S rDNA amplicons. This suggested the presence of more than one bacterial species inside the yeast strain, which was further investigated by the Metataxonomic approach (amplicon sequencing). The meta-taxonomic analysis was performed on yeast cultures showing MBLBs, and bacterial communities were assessed by HTS of the amplified V3–V4 regions within the 16S rRNA gene. The main results of two replicas of the strain OS3 are shown in supplementary Figures S10–S13. The core bacterial microbiota retrieved was represented by Firmicutes, Bacteroidota, and Proteobacteria (Figure S10). Bacteroidaceae, Lachnospiraceae, and Ruminococcaceae covered about 75% of the population (Figure S11). Particularly, the genus Bacteroides with 8 different species was the most occurring (Figures S12 and S13). We also performed several attempts to isolate obligate anaerobic bacteria of the Bacteroides group on YCFAGSC, BBA-HVK, and CMM media, but they were resulted unsuccessful.

3.4. Enzymatic Analysis

The hydrolytic potentials of the S. cerevisiae yeast OS3 and V5 strains were qualitatively investigated. Enzymatic digestion products were qualitatively investigated by TLC (Figure ), aiming to separate and identify the components of the reaction mixtures, and for each process, the depolymerization degree was evaluated by considering the percentage of the produced glucose equivalent (Table S3, DNS assay).

These results indicated that in OS3 cells from different growths, among the α-glucans, amylopectin and pullulan were better hydrolyzed (hydrolysis percentage of ∼5:10% by DNS assay) than glycogen; these results suggested the presence in the OS3 cellular system of a pullulanase activity. Pullulanases (EC 3.2.1.41) are well-known starch-debranching enzymes widely used to hydrolyze α-1,6-glucosidic linkages in starch, pullulan, amylopectin, and other oligosaccharides, with application potentials in food, brewing, and pharmaceutical industries. They were essentially recovered from various bacteria such as Bacillus sp., Thermococcus sp., Klebsiella sp., Geobacillus sp. Differently, the typical isoamylase activity of yeasts was in OS3 and V5 growth conditions, scarcely present (Figure , Table S3).

Among β-1,3-glucans and β-1,3-oligosaccharides, the curdlan, β-glucan from barley, laminarin, and laminaribiose and laminaripentaose were selected for the digestions with the OS3 and V5 crude enzymatic extracts. If curdlan, [a linear polymer consisting of β-(1,3)-linked glucose] and β-glucan from barley, [a linear glucose homopolymer with mixed β(1→3):β(1→4) interglycosidic linkages] were not degraded at all, laminarin, [a glucose polymer made up of β(1→3)-glucan with β(1→6)-branches in the β(1→3):β(1→6) ratio of 3:1] was hydrolyzed by crude enzymatic extracts of OS3 and V5 cells with a percentage of hydrolysis of 9 and 30% (DNS assay), respectively. Oligosaccharides with a degree of polymerization (DP) higher than 5 were produced, and glucose was scarcely detected (Figure ). It is worth noting that for the OS3 strain under sporulation medium conditions, with increased production of moving bacteria-like bodies, the laminarin digestion was favored. Laminariobiose and laminaripentaose were partially consumed, but glucose was not produced; after 48 h of reactions, oligosaccharides with a DP of 3 or higher were detected (data not shown).

Our results suggest the presence of an endo-1,3-β-glucanase activity and in particular, of a possible laminarinase (EC 3.2.1.39) activity in both OS3 and V5 strains. , The characterization of the genetic system governing 1,3-β-glucanase synthesis in yeasts has been reported, even though the exo-1,3-β-glucanase-encoding genes were preferentially detected and investigated with respect to the endo-1,3-β-glucanase-encoding genes. ,

Since the genomes of OS3 and V5 are known, we searched a sequence of pullulanase (X52181.1, 4091 bp long) and laminarase (AB179717.1, 2410 bp long), in the whole set of assemblies from long read sequences from EVOMICS, and in the described pangenome. No hits were found in any strain, either with an alignment longer than 200bp, regardless of the identity, or with an alignment length over 25bp and identity over 90%. The absence of genes encoding these enzymes could suggest the bacterial origin of these activities, and these results could represent a further indication of the possible presence of bacteria in the OS3 and V5 yeast cell systems.