Identification of sulfur components enhancing the anti-Candida effect of Lactobacillus rhamnosus Lcr35

GYNOPHILUS (Lcr REGENERANS) is a live biotherapeutic product (LBP) aimed at restoring the vaginal microbiome and contains the live biotherapeutic microorganism Lactobacillus rhamnosus Lcr35. In this study, the LBP formulation and manufacturing process significantly enhanced the anti-Candida activity of L. rhamnosus Lcr35, with a complete loss of viability of the yeast after 48 h of coincubation. Sodium thiosulfate (STS), one excipient of the product, was used as a potentiator of the anti-Candida spp. activity of Lactobacilli. This contact-independent phenomenon induced fungal cell disturbances, as observed by electron microscopy observations. Nonverbal sensory experiments showed clear odor dissimilarities between cocultures of L. rhamnosus Lcr35 and C. albicans in the presence and absence of STS, suggesting an impact of odor-active metabolites. A volatolomic approach allowed the identification of six odor-active compounds, including one sulfur compound that was identified as S-methyl thioacetate (MTA). MTA was associated with the antifungal effect of Lcr35, and its functional link was established in vitro. We show for the first time that the LBP GYNOPHILUS, which is a highly active product in the reduction of vulvovaginal candidiasis, requires the presence of a sulfur compound to fully achieve its antifungal effect.

In addition to the bacterial strains, the galenic forms contain nutrients, cryoprotectants and flow agents that are necessary to bacterial stability along the manufacturing process and for viability conservation. LBP formulation and production parameters have already been shown to promote the LBM properties, notably the anti-Candida effect of the Lcr35 in GYNOPHILUS 22 . Among the different molecules added during the manufacturing process, sulfur compounds have been widely used and sodium thiosulfate (STS) is part of the excipients of GYNOPHILUS. The aim of this study was to assess the impact of STS on the anti-Candida properties of the LBM Lcr35. Determination of the viability of C. albicans in co-incubation assays with Lcr35 indicated that STS enhanced its antifungal activities. This effect was associated with the presence of volatile organic compounds (VOCs) and we were able to identify one sulfur containing active component, S-methyl thioacetate, using a volatolomic approach.

Results
Antifungal of Lcr35 potentiated by STS. C. albicans ATCC 10231 viability was reduced by 2.7 log 10 (CFU/ml) after 48 h as compared to the non-treated control ( Fig. 1) without any impact onto Lcr35 viability (Fig. 1a). With STS supplementation, Lcr35 induced a 5.2 log 10 (CFU/ml) decrease of C. albicans viability after 24 h and a complete loss of viability after 48 h of co-incubation (Fig. 1b). A kinetic study performed by plate counting and cytometry combined with propidium iodide showed that C. albicans viability was affected as early as after 6 h of co-incubation (Fig. S1a,b). However, STS alone did not affect C. albicans growth compared to the control condition (Fig. 1a,b). Similarly, STS supplementation at 1 g/l did not affect Lcr35 viability in control . Viability of C. albicans white bar and Lcr35 grey bar determined at strain contacting (T0), after 24 h (T24) and 48 h (T48) of co-incubation with Lcr35 (a), and with STS (1 g/l) (b). Controls were incubated in same conditions. N = 3, *p < 0.05, **p < 0.01, ***p < 0.001 (one-way ANOVA test and Bonferroni correction).
Lcr35 alters the morphological structure of C. albicans in the presence of STS. Scanning Electron Microscopy (SEM) observations showed that C. albicans cells lost their homogeneous and regular surface after 48 h of co-incubation with Lcr35 and STS, without affecting the Lcr35 apparent morphology (Fig. 3). Similar effect was observed when C. albicans was incubated with Lcr35 in the absence of STS (data not shown). Neither Lcr35 nor C. albicans cellular integrity were impacted by STS supplementation in pure culture (Fig. 3a,b). Transmission Electron Microscopy (TEM) observations showed important alterations of yeast intracellular integrity after incubation in presence of Lcr35 and STS, contrary to control observations performed in the absence of Lcr35 and STS (Fig. 3b).

Contact-independent anti-Candida activity of Lcr35.
To further characterize anti-Candida mechanisms, contact-dependent inhibition assays were performed using cell culture insert. An equivalent antifungal activity of Lcr35 and STS onto C. albicans was observed when cells and bacteria were directly in contact or separated by a cell culture insert (Fig. 4a). However, the culture cell free supernatant induced only a loss of 2 log 10 of Candida albicans viability (Fig. 4a).
Additionally, protease and heat (100 °C for 15 min) treatments had no effect on the fungistatic activity of the Lcr35 cell free supernatant. Using ultrafiltration, only the fraction including molecular weight smaller than 10 kDa had an inhibition effect of C. albicans growth (data not shown).

Identification of the volatile compounds involved in anti-Candida activity of Lcr35. The 25
panelists non-verbal sensory experiment had determined the odor dissimilarities between seven conditions analyzed after 12 h of incubation (Fig. 5a). The map (2,3) showed a clear difference between the odors of C. albicans culture, C. albicans and Lcr35 co-culture and C. albicans co-incubated with Lcr35 and STS, suggesting a different modulating impact of Lcr35, Lcr35 with STS and C. albicans on odor-active metabolites produced. In addition, the 3D representation confirmed that STS modulates the odor of pure cultures of C. albicans and Lcr35. The odor of this binary combinations (C. albicans and STS, Lcr35 and STS) differed from the odor of the control medium inoculated with STS alone (STS) confirming that the odor of STS excipients may not explain the odor differences. The volatolomic analysis of seven conditions of C. albicans and Lcr35 cultures revealed different volatile profiles. Figure 5b presents the average aromagram obtained by compiling the aromagrams of the four samples points out 24 odor zones. Out of them, 15 odor zones were classified in one of the dairy (1-5, 6, 10), fruity floral (7,9,11,15,18), grass-vegetable (13) or mushroom (16,21) odor poles whereas the 9 remaining ones were less consensually described by the panelists. Cross analyses of these odor zones with significant MS data at corresponding retention time enabled the tentative identification of most odor-active compounds (Fig. S5). Six out of these tentatively identified odor-active compounds were found to discriminate the four cultures according to their odor-activity, confirming the existence odor-active markers of the four cultures investigated.
Among them, one sulfur compound was identified the 3-methylthiopropanal. Then to extend the marker discovery to VOC, a focus on sulfur VOCs determined four discriminate sulfur compounds including S-methyl thioacetate, dimethyldisulfide, 3-methylthiopropanal and 2-acetylthiazole. Only the S-methyl thioacetate (MTA, CAS 1534-08-3) was discriminating between the conditions analyzed ( Fig. 5c). Production of MTA by C. albicans was evaluated in reaction mixtures containing or not STS. Whereas the STS addition did not change MTA www.nature.com/scientificreports/ abundance in C. albicans culture. Conversely, the STS supplementation increased the MTA abundance by more than sixfold in Lcr35 culture. A significant increase in MTA abundance was also pointed out when STS was added to a co-culture between C. albicans and Lcr35. The low abundance of MTA was near to MS noise level in the control (culture medium supplemented with STS) confirming that the increase in MTA abundance cannot be due to its presence as an impurity in STS preparation.
S-methyl thioacetate potentiates the anti-Candida effect of Lcr35. The potentiating effect of MTA was evaluated using a concentration range 11.2 mM to 1.12 µM (Fig. 6). C. albicans ATCC 10231 was susceptible to Lcr35 with MTA. In contrast, MTA alone had no antimicrobial effect against C. albicans. Concentration 11.2 mM of MTA induced the most significant anti-Candida effect of Lcr35, similarly to supplementation with STS (1 g/l). The anti-Candida effect of Lcr35 was not potentiated with concentrations of MTA lower than 11.2 µM.

Discussion
The beneficial properties of microorganisms are usually evaluated with native culture strains. However, LBM are often administered as complex manufactured formulated products and positive impacts of the manufacturing process have been previously demonstrated in several studies [22][23][24] . In this study, we demonstrated that the vaginal formulation of Lactobacillus rhamnosus Lcr35, GYNOPHILUS, was able to kill C. albicans ATCC 10231 after 48 h, whereas the native strain alone only reduced the pathogen viability by 3 log 10 (CFU/ml) over the same length of time 22 . This vaginal LBP is composed of the LBM Lcr35 and of the excipient sodium thiosulfate. We showed that this component potentiates the anti-Candida albicans activity of the 8 Lactobacillus strains tested in this work, suggesting a Lactobacillus genus impact of STS. In addition, an inhibitory effect of Lcr35 in the   (Fig. 3a). However, we demonstrated that physical interaction was not essential for Lcr35 anti-Candida activity. Moreover, cell free supernatant of Lcr35 inhibits Candida growth and this effect is potentiated in the presence of STS. These results suggest the presence of anti-microbial effectors in the extra-cellular medium of Lcr35, the production, release and/or stability and efficacy of which could be enhanced by STS. Some lactic acid bacteria, including Lactobacilli species, have anti-Candida activity, probably by direct inhibition, through competition for adhesion sites, biofilms formation, filamentation or interference on virulence genes 25 . Some underlying mechanisms involved the produ ction of exopolysaccharides or chitinase by L rhamnosus GG able to interfere with hyphal formation of Candida 26,27 . In this study, we looked for the loss of viability of C. albicans induced by Lcr35. The production of secondary metabolites (hydrogen peroxide or lactic acid) able to lyse fungal cells has been previously described 28,29 , but these hypotheses did not apply to Lcr35 since its antifungal activity was not modified using buffered and catalase-enriched media (data not shown).
Analysis of the Lcr35 cell free supernatant showed that the potential extra-cellular components involved in C. albicans ATCC 10231 growth inhibition were present in the fraction of less than 10 kDa. Among the small molecules secreted by Lactobacilli, some secondary metabolites have been reported like antimicrobial peptides or VOCs 30,31 . Since the investigation of antimicrobial peptides was unsuccessful (data not shown), we focused on the characterization of the VOCs produced by Lcr35 in the presence of STS. VOCs are typically small compounds with low molecular mass, odorous, high vapor pressure, low boiling point and a lipophilic character 30,32 . In addition, STS is composed of two sulfur atoms, and this element is known for its odoriferous properties when it is embedded in some molecules. It has been shown already that VOCs produced by the bacteria Paraburkholderia graminis were able to inhibit hyphal growth of the fungi Rhizoctonia solani 33 . Moreover, in the context of soil, prominent mediator molecules involved in keeping this complex ecosystem in balance, are inorganic and organic Viability of C. albicans determined after 24 h of co-incubation with Lcr35 white bar, with Lcr35 and STS (1 g/l) white bar and Lcr35 with MTA grey bar. As control, C. albicans was incubated in MRS, MRS with 1 g/l STS and MRS with 11.2 mM MTA. N = 3, *p < 0.05, **p < 0.01, ***p < 0.001, significant difference between the condition and its control ### p < 0.001 (one-way ANOVA test and Bonferroni correction). www.nature.com/scientificreports/ microbial volatile compounds (mVOCs) 34 . Among VOCs, sulfurous VOCs such as dimethyl di-and trisulfide have been previously shown to have an effect on growth of several plant pathogenic fungi [35][36][37] . In our study, a non-verbal sensory experiment showed that Lcr35, with or without STS, develops a particular and perceptible odor in the presence of C. albicans. GC-8O/MS analyses identified six odor-active compounds, including one sulfur compound that was identified as S-methyl thioacetate (MTA). The bacterial synthesis of this thioester can occur through two pathways, the alcoholysis of acyl-CoA via an acyl alcohol transferase or the chemical esterification of an organic acid with an alcohol 38 . Genes potentially involved in these two pathways, i.e. encoding putative acetyl-CoA acetyltransferase and thioesterases, respectively, as well as homologous of genes encoding a two-component system involved in the regulation of the biosynthesis of MTA in Pseudomonas 39 are present in the genome of Lcr35 ; their analysis will contribute to characterizing MTA production in this bacterium. In our study, MTA when combined with Lcr35 presented an anti-Candida effect, suggesting that this thioester is an active metabolic intermediate. In addition, we showed that STS potentiates the anti-fungal action of several Lactobacillus species; it is thus likely that MTA fulfills a similar universal function. Indeed, MTA has already been reported as an antifungal VOC in bacterial soil 39 but its mechanism of action remains to be elucidated.
In conclusion, ours results show that STS is an active excipient that contributes to the anti-Candida properties of the Lcr35 containing LBP GYNOPHILUS via an unknown mechanism involving MTA. Further postgenomic investigations of the interaction LBM-Candida will be necessary to elucidate the molecular mechanism driving the enhanced anti-Candida properties of Lcr35 in presence of STS. Nevertheless, our data demonstrate that STS could be widely used as a potentiator of LBM capacities and be included as excipient in LBP formulation to promote the beneficial effect.
STS (PANREAC, Barcelona, Spain) was added at 1 g/l in the growth media i.e. either MRS or Sabouraud dextrose broth. S-Methyl thioacetate (MTA, CAS number 1534-08-3, Sigma-Aldrich, St-Quentin-Fallavier, France) was added at different concentration (11.2 mM to 1.12 µM by 1:10 dilutions) in the growth media of Lcr35 to determine the dose effect of anti-Candida potential.
For quantification of cultivable cells, the number of viable cells was determined by a plate count method on MRS agar (bioMerieux, Marcy l'Etoile, France) for detection of Lactobacillus spp. and Sabouraud agar for Candida strains. MRS plates were incubated at 37 °C for 72 h and Sabouraud petri dishes at 25 °C for 120 h 22,50 . The results were expressed as colony forming units per milliliter (CFU/ml) and the limit of detection was 10 CFU/ml. Anti-Candida effect. After 48 h of preculture, a mixed culture broth was made with 50% of Lactobacilli subculture containing or not STS (from 0.5 to 5 g/l) or MTA (11.2 mM to 1.12 µM) and 50% of Candida sub- Similar experiments were performed using cell culture inserts (TRANSWELL, Dutsher, Brumath, France). The membrane cell culture inserts, with 0.4 µm-pore-size polycarbonate membrane inserts, were placed in 12-well plates containing a final volume of 2 ml. Then, Lcr35 subculture with or without STS (1 g/l) was added into six-well cell culture plates and C. albicans ATCC 10231 subculture was added into the upper chambers and vice versa. The culture plates were incubated at 37 °C for 24 h. Controls were performed included pure cultures of C. albicans or Lcr35, with and without STS (1 g/l) and a co-culture without cell culture inserts. The viability of microorganisms was determined. Each assay was performed in triplicate.
The cell viability was tested using flow cytometry. The abundance of dead C. albicans cells was measured during co-incubation; one milliliter of the suspension was centrifuged (11,000×g, 10 min, at room temperature), and re-suspended in PBS (pH 7; Lonza, Levallois-Perret, France). Propidium Iodide (PI; Sigma-Aldrich, St-Quentin-Fallavier, France) was added (1 µg/ml) before incubation in the dark for 40 min at room temperature. One hundred microliters of each sample were seeded in 96 well plates 51 . The flow cytometry analysis was performed on a BD Accuri™ C6 Cytometer (BD Biosciences, Le Pont de Claix, France) and data were analyzed using the CFlow PLUS version 1.0 software (BD Biosciences). At least 2.10 5 elements per sample were analyzed at a slow flow rate (14 µl/ml). Yeast and bacteria were discriminated in two different areas (P1 and P2). Two controls, one with live Lcr35 or C. albicans, and another with heat shock (100 °C, 15 min) killed Lcr35 or C. albicans were used to calibrate the gate.
Inhibition of pathogens growth using cell free supernatant. The cell-free supernatants were collected after 48 h of L. rhamnosus Lcr35 subculture with or without STS and then centrifuged (10,000×g, 15 min) and filtered through a 0.2 µm filter (Millex; Merck, Darmstadt, Germany). The supernatants were diluted to obtain a proportion range of 10%, 30% and 50% in Sabouraud dextrose broth. C. albicans ATCC 10231 was inoculated at 10 6 CFU/ml in medium containing the supernatant dilutions and incubated at 37 °C during 24 h. The C. albicans growth was followed by reading OD 600 nm every 15 min after shaking with automate Tecan microplate readers (TECAN, Genius, Männedorf, Switzerland). The controls were performed with a proportion range of MRS broth with 1 g/l STS and proportion range of sterile water.
Similar experiments were performed with cell free supernatants boiled (15 min at 100 °C) or treated with 0.2 mg/ml pronase E (Merck KGaA, Darmstadt, Allemagne.) and 0.2 mg/ml proteinase K 0.2 mg/ml (Sigma-Aldrich, St-Quentin-Fallavier, France) during 2 h at 37 °C. The enzymes were inactivated by thermal shock at 80 °C during 20 min.
The supernatant ultrafiltration was performed using vivacell100 (Sartorius, Dourdan, France) (20 min at 2000×g) using a 100 kDa 10 kDa and cut-off membrane. The residue was subsequently solubilized with an equivalent volume of H 2 O. The residual anti-C. albicans activities of each supernatant fraction was evaluated by reading OD 600 nm using proportion range of 50%. Untreated cell free supernatant was included in each experiment as control.
Electron microscopy analysis. Samples were treated as previously described 52 . Details of the technical process were done into the supplemental materials and methods section.
Sample preparation for Purge and trap gas chromatography-mass spectrometry analysis. The filtered microbial cultures were aliquoted in 4 ml screwed caps vials and stored at − 20 °C. Analyses were done on the culture media corresponding to each of the 7 treatments (C. albicans; MRS broth + STS; C. albicans + STS; Lcr35 + C. albicans; Lcr35; Lcr35 + STS; Lcr35 + C. albicans + STS) with three replicates. The vials were left to thaw overnight at 4 °C, the sample was homogenized by 5 s vortexing just before extraction.

Assessment of odor dissimilarity by non-verbal sensory analysis. A panel of 25 untrained judges
was recruited from the research teams of the University Clermont-Auvergne. To avoid altering their olfactory sense, the judges were asked the day of the session not to perfume themselves, not to consume anything after their breakfast except water and to avoid strong or predominant atmospheres (solvents, cigarettes, ferment-ers…). Three sessions (one per culture time studied: T = 0 h, T = 12 h and T = 24 h) of about 30 min each were organized at 10h45 am over 15 days. The analyses were carried out in a room equipped with individual boxes. Six to seven vials closed with screw caps containing the different culture media numbered using 3-digit numbers and randomized were placed on a sheet of paper (40 × 60 cm 2 ) in front of each judge. The judges were instructed to smell the vials several times to compare them. After smelling them, they had to position them on the sheet of paper using the entire surface so that two vials were as closer as their odors were similar and were as distant as their odors appeared different. Once the positioning was complete, the vial numbers were recorded on the paper where the vials were placed. The x and y coordinate points of each vial were then recorded. For each napping session, a distance matrix was calculated from the vial coordinates: (x 2 − x 1 ) 2 + y 2 − y 1 2 , x and y representing the coordinates of the points.

Volatile compounds analysis by dynamic headspace-gas chromatography-mass spectrometry (DH-GC-MS) and
Olfactometry. DH-GC-MS analyses were performed on three replicates of seven conditions corresponding to culture media with or without STS inoculated or not with C. albicans, Lcr35 or both, as described in supplemental materials and methods section. www.nature.com/scientificreports/ The odorous compounds of 4 selected conditions (Lcr35; Lcr35 with STS; Lcr35 and C. albicans; Lcr35 with STS and C. albicans,) were analyzed by eight-way olfactometry coupled to gas chromatrography-mass spectrometry (DH-GC-MS/8O) 53,54 . The analyses consisted in the simultaneous acquisition of the GC-MS signal of the sample in one side and the GC-8O signal consisting in the odor profile detected by the panel of eight judge in the other side. Details of the procedure were mentioned on the supplemental materials and methods section. Statistical analysis. Statistical analyses were performed using a one-way ANOVA test followed by Bonferroni correction for growth experiments. Statistical analyses were carried out using GraphPad Prism 5 software (www.graph pad.com/prism ) (*p < 0.05; **p < 0.01; ***p < 0.001).
Data treatment of sensory analyses were processed and analyzed using STATISTICA version 13 software (Statsoft, Maisons-Alfort, France). For assessment of odor dissimilarity, a panel distance matrix was constructed by averaging the individual distance matrices of judges (n = 25). The resulting distance matrix was processed by Multi-Dimensional Scaling (MDS). The number of dimension of the model was set based on 1/ Stress value, 2/ Alienation value and 3/ Shepard diagram following the recommendation of Ding, 2018.
For the determination of distinct odor-active compounds by DH-GC-MS/8O, a Kruskal-Wallis nonparametric test (p < 0.05) was processed on the intensity given by each judge for each odor zone. The relevant odor-active compounds related to each significant odor zones were considered as distinct markers of the four cultures.
For the screening of sulfur compounds as markers of culture discrimination, one-way ANOVA (model: peak area = culture, p < 0.05) were performed on triplicates of DH-GC-MS peak areas. Tukey multiple mean comparison tests (p < 0.05) were processed on significant sulfur compounds to investigate the significance of inter-culture differences.
Statistical analysis of gas chromatography-mass spectrometry were performed using Statistica Software. One-way ANOVAs were performed on peak areas of the specific ion (m/z) of each volatile compound to identify significant effects of the treatments (abundance of compound specific ion: treatment, p < 0.05). Newman-Keuls multiple comparison tests were performed to determine significant differences between treatments (p < 0.05).