Medium-chain fatty acids (MCFAs) are carboxylic acids with 6–12 carbons. MCFAs have demonstrated anti-pathogenic and immunomodulatory health effects, and are known to play a role in the regulation of body fat metabolism. Accordingly, MCFAs such as caprylic acid (C8:0), capric acid (C10:0), lauric acid (C12:0), or their combination with other compounds have been shown to inhibit pathogens like Candida albicans (CA) (Murzyn et al., 2010b; Lee et al., 2021), Clostridium difficile (Yang et al., 2017), Listeria monocytogenes (Hao et al., 2020), and Salmonella (Moschonas et al., 2012). In Candida spp, MCFAs may mimic the quorum sensing molecule farnesol that interferes with communications among fungal populations, leading to its potency against Candida (Lee et al., 2021). With regards to inflammation, C10:0 has been found to induce the upregulation of GPR84 and PPARγ and the downregulation of HIF-1α (Sam et al., 2021). In another study, C8:0 has been found to enhance inflammatory IL-8 secretion by >35% in human fetal intestinal epithelial cells (Andoh et al., 2000). Given MCFAs’ potency against pathogens and their roles in immunomodulation, MCFAs, their derivatives, and probiotics producing MCFAs have been widely used as anti-pathogenic agents and health supplements.

Saccharomyces cerevisiae var. boulardii (also named as Saccharomyces boulardii) is a probiotic yeast with versatile health effects. Not only are they anti-pathogenic and involved in immunomodulation, but they also promote the growth of beneficial gut microbiota—especially those associated with MCFAs produced and secreted by certain boulardii strains. For instance, Murzyn et al. found that a boulardii strain inhibits CA’s adhesion to intestinal cell lines and its extract reduces cytokine-induced inflammatory responses in Caco-2 cells as revealed by suppressed IL-8 expression (Murzyn et al., 2010a). Boulardii strains have also been found to diminish filamentation, biofilm formation, and CA translocation (Berg et al., 1993; Krasowska et al., 2009). Meanwhile, Tomicic and others reported how a boulardii strain significantly reduced the adherence ability of Candida glabrata ZIM2344 and ZIM2369 in a dose-dependent manner (Tomicic et al., 2016). The administration of the boulardii strains CNCM I-745, CNCM-I-1079, or CNCM-I-1079 + Lactobacillus rhamnosus GG was also found to expedite the reestablishment of a healthy microbiota after diarrheic dysbiosis (More and Swidsinski, 2015). In a piglet model, supplementation with a boulardii strain mafic-1701 was observed to improve feed conversion efficiency (Zhang et al., 2020). In some cases, boulardii strains can natively produce and secrete MCFAs represented by C10:0, which inhibits CA’s filamentous growth, adhesion, and biofilm formation (Murzyn et al., 2010b; Kunyeit et al., 2020; Duysburgh et al., 2021). Given these beneficial effects, boulardii strains have been widely used as a probiotic yeast to promote human health by inhibiting pathogens like Candida and modulating the gut microbiota. To date, several boulardii strains have been commercialized to promote human health.

C. albicans is a major opportunistic fungal pathogen. Clinically, current therapeutics for Candida infections mainly rely on the administration of anti-fungal drugs such as fluconazole, which has been implicated in multi-drug resistance (MDR). Notably, MDR associated with Candida has become a key clinical challenge in fighting fungal infections and a threat to public health (Berman and Krysan, 2020; Kunyeit et al., 2020). Consequently, alternative solutions are required to fight Candida infections more effectively. Given the proven anti-Candida effects of probiotic yeast and MCFAs, MCFA-producing boulardii strains could potentially offer an effective therapy against Candida infections. Nevertheless, MCFA production and secretion by wild type boulardii strains are strain- and condition-dependent, which could make it difficult to achieve desired anti-Candida effects. Therefore, a boulardii strain that stably produces and secretes MCFAs in a controllable manner would be ideal, and could be realized by expressing a heterologous MCFA biosynthetic pathway through synthetic biology-driven metabolic engineering strategies (Liu et al., 2016; Yu et al., 2016; Chen et al., 2018; Xia et al., 2019; Chen et al., 2020; Hossain et al., 2020).

In this study, we aimed to engineer a commercial probiotic yeast strain (S. cerevisiae var boulardii) to produce MCFAs with antagonistic effects against the opportunistic pathogen C. albicans. Briefly, we identified the ploidy of the boulardii strain, then constructed and expressed a heterologous MCFA biosynthetic pathway into Δura3. Next, we examined MCFA production and secretion, then confirmed MCFAs’ effects on anti-biofilm and anti-hyphal formations in C. albicans SC5314 cells and its immunomodulatory effects in the human epithelial cell line Caco-2. This represents the first study describing the metabolic engineering of a commercial probiotic yeast strain to produce MCFAs with anti-Candida effects. Our study provides useful insights into developing live biotherapeutics against Candida infections through metabolic pathway engineering.

Strains, growth conditions, plasmids, and chemicals. Strains and plasmids used are listed in Supplementary Table S1. S. cerevisiae S288C, S. cerevisiae var. boulardii, and C. albicans were grown in YPD (yeast extract 10 g/L, peptone 20 g/L, dextrose 20 g/L), YGD (yeast nitrogen base 6.7 g/L, dextrose 20 g/L, and yeast synthetic dropout medium supplements without uracil 1.92 g/L), or YRGD (yeast nitrogen base 6.7 g/L, D-raffinose 10 g/L, yeast synthetic dropout medium supplements without uracil 1.92 g/L, and D-(+)-galactose 20 g/L) at 30°C. Escherichia coli was grown in Luria–Bertani medium at 37°C. Hygromycin (200 mg/L) or ampicillin (100 mg/L) were added when applicable. DNA preparation kits were purchased from Qiagen (Germany), PCR reagents from Biorad (United States), restriction enzymes and DNA ligase from New England Biolabs (United Kingdom), and chemicals from Sigma-Aldrich (United States) unless specified.

Ploidy characterization by PCR. We isolated S. cerevisiae var. boulardii, designated as CLPY01, from Jarrow Formulas^® yeast probiotics (Saccharomyces Boulardii MOS). To verify the boulardii strain, we performed PCR using genomic DNA as a template along with intron splice site primers EI1 and LA1 (Fietto et al., 2004). We then compared the amplicon’s profile with that of S. cerevisiae S288C, boulardii U28, and M2 (Hamedi et al., 2013). To elucidate the boulardii’s ploidy after strain verification, we performed PCR using genomic DNA as a template by primers that pair with the MATa locus (MatF) or MATα locus (MatαF) and the reverse primer MatR (Supplementary Table S1).

Gene deletion. We amplified the URA3 gene deletion cassette from pUG75, with this cassette containing a 45-bp upstream and downstream region flanking URA3’s coding sequence. The obtained gene deletion cassette was then transformed into CLPY01. To enable effective transformation, we used the following method for preparing competent cells: We prepared and diluted an overnight culture to OD₆₀₀ 0.3 in fresh YPD medium, after which we grew it at 30°C and 225 rpm until the culture reached OD₆₀₀ 1.6. Cells were collected and washed once with deionized H₂O and then with an ice-chilled buffer (buffer 1: 1 M sorbitol + 1 mM CaCl₂). Cells were suspended in 5 mL buffer (buffer 2: 0.1 M LiAc + 10 mM DTT) and incubated at 30°C and 225 rpm for 30 min. We then washed the cells once using 5 mL buffer 1 and then once using 1 mL 0.1 M LiAc. The cell pellet was then transformed using a LiAc/PEG-based method (Jung et al., 2021) followed by heat shock at 37°C for 1 h. Cells were recovered in YPD medium at 37°C for 1 h and spread on selective plates. We then screened colonies based on the presence of resistance to hygromycin and 5-fluoroorotic acid (5-FOA). The colonies showing resistance to both hygromycin and 5-FOA were further validated using YPD plates with and without uracil. Next, marker rescue was performed in a colony that grew only on a YPD plate (Ling et al., 2015). The resulting strain Δura3 was named to CLPY02.

Gene cloning and overexpression. Genes including hSFP, mhFAS, and rTEII (sequences are shown in Supplementary Table S2) were synthesized by Life Technologies and cloned into pESC-URA standard protocols (Shao et al., 2009), resulting in the plasmid pESC-HR with hSFP under P_GAL10 and mhFAS-rTEII under P_GAL1. The recombinant plasmid pESC-HR was confirmed by sequencing and transformed into CLPY02 for MCFA production under D-galactose induction. Next, we replaced the promoters P_GAL10 and P_GAL1 to P_TDH3 (for hSFP expression) and P_TEF1 (for mhFAS-rTEII expression) in the opposite direction (sequences are shown in Supplementary Table S2), resulting in the plasmid pESC-gHR as confirmed by DNA sequencing. The used primers are listed in Supplementary Table S2.

Gene integration. We employed the DNA assembler method (Shao et al., 2009) to integrate the MCFA-biosynthetic pathway in pESC-gHR into CLPY02. Given the large size of the fragment, we divided the pathway into three fragments, including the terminator of CYC1 and ADH1 (HR1: 2965bp, HR2: 3016 bp, HR3: 4050 bp). HR3 was fused to a cassette expressing the hygromycin resistance gene (HygR) of pUG75 (HR3-HyrR) to facilitate colony screening with chromosomal integration. HR1, HR2, and HR3 contained 500-bp regions that overlapped with each other. HR1 and HR3-HyrR contained a 500-bp upstream or downstream region flanking the URA3 locus. The agar plate with both hygromycin and uracil added was used to select colonies integrated with the gene cassette, and marker rescue was performed prior to the second round of chromosomal integration. To integrate another copy of the MCFA biosynthetic pathway, HR3 was fused to a cassette with corresponding overlapping regions that express the URA3 of pUG72 (HR3-URA3) for chromosomal integration. HR1 and HR3-URA3 both contain a 500-bp upstream or downstream region flanking a URA3 locus. The agar plate excluding uracil was used to select colonies integrated with the gene cassette. The resulting strain carrying two copies of the MCFA biosynthetic pathway was named CLPY04 and used for subsequent experiments.

Fatty acid extraction and detection. To measure fatty acids produced in engineered strains, single colonies were pre-cultured in appropriate media overnight. Cells were then inoculated into a fresh medium and incubated at 30°C and 225 rpm. Five milliliters of cells were harvested by centrifugation, and cell pellets were washed using deionized water. We added an internal standard heptadecanoic acid and lysed cells by adding 1.5 mL hydrochloride (HCl) and incubating for 1 h at 70°C. To derivatize fatty acids, we added 1.5 mL 10% HCl–methanol (v/v) into the lysate, followed by vortexing and incubation for 3 h at 62°C. After cooling down, fatty acid methyl esters (FAMEs) were extracted by adding 2 mL hexane and vortexing. Lastly, we centrifuged the sample and collected the upper hexane layer for GC analysis. For the detection of extracellular fatty acids, we added the internal standard and 10% HCl–methanol (v/v) into the supernatant and proceeded with fatty acid derivatization. GC analysis was performed following a previously reported method using an HP 7890 B GC system with an Agilent 5977 A MSD equipped with a HP-5MS column (Yu et al., 2016; Ng et al., 2020). GC peaks were identified by comparing the retention times and mass spectra of FAME standards. Data analysis was performed using the Agilent Enhanced Data Analysis software.

Biofilm and hyphal formation assays. C. albicans SC5314 was inoculated for overnight cultivation. The overnight culture was diluted to OD₆₀₀ 0.1 in a microplate well containing supernatant (30%, v/v) + YPD (70%, v/v). The medium was replaced every 24 h. After incubation at 75 rpm and 37°C for 4 d, the biofilm was stained by crystal violet and measured at 590 nm. Under a microscope, hyphal formation was observed after 6 h incubation in supernatant (30%, v/v) + DMEM medium (DMEM1, without sodium bicarbonate and HEPES) (Sigma-Aldrich) (70%, v/v) at the appropriate pH. YRGD or YGD medium with the same volume as the supernatant was added as negative controls.

Co-culturing of Caco-2 and C. albicans SC5314. Caco-2 cells were seeded in a 6-well plate containing Gibco™ DMEM (DMEM2, with sodium bicarbonate and HEPES) + 10% fetal bovine serum with 5% CO₂ supply at 37°C. The medium was changed every 24 h. On day 4, Caco-2 cells were pre-treated by DMEM1 for 1.5 h. Next, SC5314 cells (∼4 × 10⁷ CFU/mL) and MCFA-containing supernatant (30%, v/v) were added and incubated for 4 h. Caco-2 cells were collected and used for total RNA extraction as described below. YGD medium with the same volume as the supernatant was added as a negative control.

RNA extraction and real-time PCR (RT-PCR) analysis. The overnight culture of SC5314 was washed twice using customized DMEM and diluted in the appropriate medium to OD₆₀₀ 2 in a 6-well plate. Cells were collected after incubation at 75 rpm and 37°C at the appropriate timepoints. Total RNA was extracted using an RNeasy Kit (Qiagen) following the provided manual. Complementary DNA synthesis and RT-PCR were performed to analyze the transcription levels of BCR1, EFG1, HGC1, HWP1, and UME6 based on a reported method (Ling et al., 2015) using the 2^−ΔΔCT method. Gene expression was normalized to ACT1.

The total RNA of the co-cultured Caco-2 cells was extracted using a Trizol reagent (Hwang et al., 2021). RT-PCR was then performed to analyze the transcription levels of IL-6, IL-8, LL-37, SA1009, CCL20, and hBD-3 using the 2^−ΔΔCT method. Gene expression was normalized to the ACTB gene.

In this study, we characterized and metabolically engineered a commercial probiotic yeast S. cerevisiae var. boulardii strain for antagonistic effects against the opportunistic pathogen C. albicans. The engineered boulardii strain was shown to stably produce and secrete an MCFA mixture comprising C6:0, C8:0, and C10:0. We found that the secreted MCFAs showed antagonistic effects to the opportunistic pathogen C. albicans, as demonstrated by reduced biofilm and hyphal formations. Through RT-PCR analysis, we confirmed that the secreted MCFAs downregulated the expression of virulence-related genes in C. albicans. Furthermore, we found that MCFAs induced the upregulation of immune response genes in Caco-2 cells co-cultured with C. albicans, indicating MCFAs’ roles in immunomodulation (Figure 5B). To our knowledge, our study represents the first study tackling the metabolic engineering of a commercial probiotic yeast for anti-Candida applications. Overall, our findings provide the basis for potentially developing a live biotherapeutics probiotic yeast against fungal infections such as Candida strains.

The engineered boulardii strain was proven to produce and secrete MCFAs showing antagonistic effects against C. albicans. We confirmed the antagonistic effects through in vitro assays by supplying the MCFA-containing supernatant from the engineered boulardii strain into the co-culturing system. In the future, we could further investigate the attribution of individual MCFAs, as well as specific combinations of different MCFAs at varying doses to the downregulation of virulence-related genes in Candida strains. Future efforts could also consider further optimizing the engineered strains and validating their antagonistic effects. For instance, native fatty acid synthesis could be tuned and blocked, while other competing pathways could be removed to increase the metabolic fluxes towards MCFA biosynthesis. We observed a slower growth and lower cell density of CLPY04 compared to the control CLPY02 (Supplementary Figure S3), suggesting a need to further improve cell growth. Moreover, it would be interesting to determine CLPY04’s genetic stability. Given the boulardii strains’ relatively low engineering efficiency, strain engineering could be facilitated by using synthetic biology toolkits such as CRISPR-based genome editing (Liu et al., 2016) and an optimal method with higher transformation efficiency. The engineered MCFA-producing strain could be directly co-cultured to validate its antagonistic effects against Candida. In addition, the efficacy of the engineered boulardii strain could be validated using an in vivo model.