Probiotic yeast and uses thereof

Abstract

The invention relates to non-virulent Saccharomyces boulardii, in which the ENA1 gene is knocked out. The non- virulent Saccharomyces boulardii, in which the ENA1 gene is knocked out may be used as a probiotic. The inven- tion also relates to a Saccharomyces boulardii strain having the accession number NCAIM (P) Y 001536 and its use as a probiotic.

Description

[0001] Probiotic yeast and uses thereof

[0002] FIELD OF THE INVENTION

[0003] The invention relates to non-virulent Saccharomyces cerevisiae ‘boulardii ’ with a knocked-out ENA1 and to non- virulent Saccharomyces cerevisiae ‘boulardii ’ with a knocked-out ENA1 comprising a transgene coding for a gene product of interest. The non-virulent Saccharomyces cerevisiae ‘boulardii ’ with a knocked-out ENA1 may be used as a probiotic. The invention also relates to the Saccharomyces cerevisiae ‘boulardii ’ strain having the accession number NCAIM (P) Y 001536 and its use as a probiotic.

[0004] BACKGROUND OF THE INVENTION

[0005] The probiotic yeast, the so-called ‘boulardii ’ subtype of Saccharomyces cerevisiae, clearly stands out in a market dominated by lactic acid-producing and other bacterial species as the most sought after yeast probiotic globally. Among others, it is used for the treatment of Clostridioides difficile-related and antibiotic-associated diarrheal and to improve the symptoms of irritable bowel syndrome (IBS). It is also being researched for probiotic -enhanced food products and drinks.

[0006] Despite the indisputable health benefits of S. ‘boulardii ’ there is a risk of fungemia among immunocompromised or severely ill patients, infants, and elderly people. In recent years, a growing number of such cases have been reported, often following the use or administration of yeast probiotic supplements (Imre, A. et al. A new, rapid multiplex PCR method identifies frequent probiotic origin among clinical Saccharomyces isolates. Microbiol. Res. 227, 126298 (2019)). The frequency of Saccharomyces probiotic fungemia cases is not known and there is a lack reporting obligations, and so far, only a single retrospective analysis was published. This study suggested that, in the hospital examined, the incidence of fungemia cases for S ‘boulardii ’ was similar to that of C. albicans (Wombwell, E., Bransteitter, B. & Gillen, L. Incidence of Saccharomyces cerevisiae Fungemia in Hospitalized Patients Administered Saccharomyces boulardii Probiotic. Mycoses 64, 1521-1526 (2021)). Scarcity of data is largely due to the lack of subtyping conducted during clinical diagnostics; thus, the number of undetected cases might greatly outnumber the published ones. Strain-level genotyping, e.g. with multiplex PCR may ameliorate this shortfall in diagnostics. In an assessment report from the European Medicines Agency (Saccharomyces boulardii- Scientific conclusions and grounds for the variation to the terms of the Marketing Authorisation(s). (2020)), a causal relationship was found between the use of S. ‘boulardii ’ and fungemia. As a result, bloodstream infections have been indicated as a side effect of rare frequency on the package leaflet of yeast probiotic products in the European Union. Recently, a hospital in Belgium even completely discontinued the use of S. ‘boulardii ’ probiotics based on safety concerns following high mortality rate in locally infected patients (Poncelet, A., Ruelle, L., Konopnicki, D., Miendje Deyi, V. Y. & Dauby, N. Saccharomyces cerevisiae fungemia: Risk factors, outcome and links with S. boulardii-containing probiotic administration. Infect. Dis. Now 51, 293-295 (2021)). In a recent study, 48% of clinical isolates from the university hospital in Debrecen, Hungary, proved to be probiotic yeasts (Harmath, A. et al. Phylogenomic diversity of clinical Saccharomyces cerevisiae and the prevalence of probiotic -derived isolates in a tertiary care centre in Hungary. Microbiol. Spectr. accepted f, (2026)).

[0007] At the same time, there is growing interest in extending the potential applications of S. ‘boulardii ’ by the heterologous expression of various genes, and the introduction and modification of biosynthetic pathways for immune modulating (Scott, B. M. et al. Self-tunable engineered yeast probiotics for the treatment of inflammatory bowel disease. Nat. Med. 2021 277 27, 1212-1222 (2021)) and antimicrobial effects (Li, R., Wan, X., Takala, T. M. & Saris, P. E. J. Heterologous Expression of the Leuconostoc Bacteriocin Leucocin C in Probiotic Yeast Saccharo- myces boulardii. Probiotics Antimicrob. Proteins 13, 229-237 (2021)) or other biotherapeutic applications. The proposed applications of such genetically engineered probiotic yeast strains would often target already vulnerable patients who may have predisposing factors for ftingal infections. Despite the obvious importance of safety concerns, there has been virtually no crosstalk between studies concerning the health benefits and genetic improvements of S. ‘boulardii’, and those focused on its opportunistic pathogenicity and virulence determinants. Thus, an integrated approach putting both the patients and the yeast in focus is needed to better understand probiotic yeast infections, to improve their safety and to extend their applicability.

[0008] On the patients’ side, knowledge has been steadily accumulating on S. ‘boulardii ’ fungemia cases which affect already hospitalized patients. A total of 108 cases of Saccharomyces fungemia were recently analyzed in a systematic review, and they found that the most important risk factors were hospitalization at intensive care units, parenteral nutrition or enteral feeding, diarrhea, diabetes mellitus, immunosuppression, gastrointestinal surgery, and catheter use (Bras, G. et al. Secreted Aspartic Proteinases: Key Factors in Candida Infections and Host-Pathogen Interactions. Int. J. Mol. Sci. 25, (2024)).

[0009] Virulence factors of pathogenic yeast species, like Candida spp. or Cryptococcus neoformans, upon infection in humans are well-known. However, the factors behind opportunistic pathogenicity concerning S. ‘boulardii ’ itself, are less clear. In most cases, only a low number of product isolates have been compared. In general, in vitro virulence factor tests failed to identify clear attributes enabling opportunistic pathogenicity in a mammalian model or in humans for S. ‘boulardii’ and in general, for S. cerevisiae (Fernandez-Pacheco, P., Monge, I. M. R., Fernandez-Gonzalez, M., Colado, J. M. P. & Arevalo-Villena, M. Safety Evaluation of Yeasts With Probiotic Potential. Front. Nutr. 8, 659328 (2021)). Nevertheless, using mouse bloodstream infection models, the opportunistic pathogenic nature of the probiotic strain has been experimentally confirmed (de Llanos, R. et al. In vivo virulence of commercial Saccharomyces cerevisiae strains with pathogenicity -associated phenotypical traits. Int. J. Food Microbiol. 144, 393-9 (2011)). Four commercial and ten clinical S. ‘boulardii ’ isolates, the largest number of such isolates so far to be compared, were tested by Imre et al. (Virulence Factors and in-Host Selection on Phenotypes in Infectious Probiotic Yeast Isolates (Saccharomyces ’boulardii’). J. fungi (Basel, Switzerland) 7, 746 (2021)) for their general phenotype, extracellular virulence factors, immune interactions, and virulence in Galleria melonella larva model. No unequivocal difference was found that would indicate a specific trait that enables S. ‘boulardii ’ clinical isolates to acquire higher pathogenicity compared to the commercial strains, but the clinical isolates showed various phenotypic differences.

[0010] Since all previous genomics studies aiming to identify virulence factors in Saccharomyces put S. cerevisiae isolates and strains in the center of attention, the number of studies focusing on the genetic background of the probiotic yeast’s virulence attributes is scarce. Imre et al. (Heme Oxygenase-1 (HMX1) Loss of Function Increases the InHost Fitness of the Saccharomyces ‘boulardii’ Probiotic Yeast in a Mouse Fungemia Model. J. Fungi 8, 522 (2022)) confirmed that the loss of lunction of heme oxygenase-1 (HMX1), a known virulence factor in Candida albicans, elevated and not diminished virulence in six closely related yet distinguishable probiotic yeast genetic backgrounds in a mouse model.

[0011] There is thus a need for Saccharomyces cerevisiae ‘boulardii ’ which can be used safely as a probiotic. SUMMARY OF THE INVENTION

[0012] Saccharonyces ‘boulardii ’ is provided, wherein the S. ‘boulardii ’ is engineered to lack a functional ENA1 gene. Saccharonyces ‘boulardii ’ with a reduced virulence in a mammalian host is provided, wherein the S. ‘boulardii ’ is engineered to lack a functional ENA1 gene.

[0013] Saccharonyces ‘boulardii ’ is provided, wherein the expression or the function of a product of the ENA1 gene in the S. ‘boulardii ’ is inhibited.

[0014] A method for decreasing the virulence of Saccharonyces ‘boulardii comprising inhibiting the activity of a gene product of the ENA1 gene.

[0015] A method for decreasing the virulence of Saccharonyces ‘boulardii comprising engineering the Saccharonyces ‘boulardii ’ to lack a or all of its functional ENA1 genes.

[0016] Preferably the Saccharonyces ‘boulardii ’ had a single copy of the ENA1 gene prior to the engineering. Preferably the single copy of the ENA1 gene was located on chromosome IV.

[0017] Preferably the virulence of the S. ‘boulardii ’ is reduced compared to the corresponding S. ‘boulardii ’ that is not engineered to lack a functional ENA1 gene.

[0018] Preferably the Saccharonyces ‘boulardii ’ is avirulent.

[0019] Preferably the Saccharonyces ‘boulardii ’ does not comprise any functional ENA1 genes.

[0020] Preferably the or all the ENA1 gene(s) is knocked-out.

[0021] Preferably the ENA1 gene comprises a loss of function mutation.

[0022] Preferably all copies of the ENA1 gene comprise a loss of function mutation or are knocked out.

[0023] Preferably the ENA1 gene is misfunctioning. Preferably the expression of the gene product of the ENA1 gene is inhibited or wherein the activity of the protein encoded by the ENA1 gene is inhibited. Preferably the S. ‘boulardii ’ lacks a functional protein encoded by the ENA1 gene.

[0024] Preferably the S. ‘boulardii ’ has a reduced virulence compared to S. ‘boulardii ’ having the nucleic acid sequence submitted to the National Library of Medicine (BioProject) and registered on 22 -Aug-2022, with the accession code PRJNA813746, ID: 813746 (https: / / www.ncbi.nlm.nih.gov / bioproiect / PRJNA813746 / ; downloaded 20-12- 2023).

[0025] Preferably the S. ‘boulardii ’ is viable in the gastrointestinal system of a mammal.

[0026] Preferably the S. ‘boulardii ’ has the same antimicrobial pattern or features as the corresponding non-engineered S. ‘boulardii ’ with a functional ENA 1 gene

[0027] Preferably the S ‘boulardii ’ inhibits the growth of Escherichia coli (preferably ATCC11775) and / or Bacillus sub- tilis (preferably ATCC6051).

[0028] Preferably the mammalian host or the mammal is a human.

[0029] Preferably the S. ‘boulardii ’ is further modified to produce a heterologous protein or peptide.

[0030] Preferably the S. ‘boulardii ’ comprises a transgene to produce a heterologous protein or peptide.

[0031] Preferably the heterologous protein or peptide is a bacteriocin. Preferably the bacteriocin is leucocin C. Preferably the transgene is located at the ENA1 locus.

[0032] Preferably the S ‘boulardii ’ is the S. ‘boulardii ’ having the accession number NCAIM (P) Y 001536, deposited with the National Collection of Agricultural and Industrial Microorganisms (Somloi lit 14-16, 1118 Budapest, Hungary) on 14 November, 2024, or a derivative S. ‘boulardii ’ thereof, wherein the derivative S. ‘boulardii ’ does not have a functional ENA1 gene. Preferably the derivative S. ‘boulardii ’ is avirulent in mice. Preferably the derivative S. ‘boulardii ’ is viable in the gastrointestinal system of mice. Preferably the derivative S. ‘boulardii ’ inhibits the growth of E. coli ATCC11775 and / or B. subtilis ATCC6051 as measured under conditions described in the Examples. Preferably the derivative S. ‘boulardii ’ is produced by culturing S. ‘boulardii ’ having the accession number NCAIM (P) Y 001536 or by engineering (e.g. genetically modifying) S. ‘boulardii ’ having the accession number NCAIM (P) Y 001536.

[0033] Preferably the S. ‘boulardii ’ or the derivative S. ‘boulardii ’ is used as a probiotic. Preferably the S. ‘boulardii ’ or the derivative S. ‘boulardii ’ is for use as a probiotic.

[0034] Preferably the S. ‘boulardii ’ or the derivative S. ‘boulardii ’ is for use in therapy, preferably in the treatment or prevention of diarrhea, preferably Clostridioides difficile-related, HIV / AIDS-associated, rotaviral or antibiotic- associated diarrhea, irritable bowel syndrome, Clostridioides difficile infection, Elelicobacter pylori infection, blastocystosis or acute gastroenteritis.

[0035] Preferably the S. ‘boulardii ’ or the derivative S. ‘boulardii ’ is used in a probiotic-enhanced food product or drink. A composition comprising the S. ‘boulardii ’ or the derivative S. ‘boulardii ’ is provided. Preferably the composition is a probiotic composition and comprises a carrier suitable for the S. ‘boulardii ’ or the derivative S. ‘boulardii Preferably the composition is a pharmaceutical composition and comprises a pharmaceutically acceptable carrier or excipient.

[0036] A method is provided for the treatment or prevention of diarrhea, preferably Clostridioides difficile-related, HIV / AIDS-associated, rotaviral or antibiotic -associated diarrhea, irritable bowel syndrome, Clostridioides difficile infection, Helicobacter pylori infection, blastocystosis or acute gastroenteritis, comprising administering the engineered S. ‘boulardii ’ or the derivative S. ‘boulardii ’ or the (pharmaceutical) composition to a subject in need of such treatment or prevention.

[0037] BRIEF DESCRIPTION OF THE FIGURES

[0038] Figure 1. Commercial S. ‘boulardii ’ isolates used in current study. References: 1: Imre, A. et al. A new, rapid multiplex PCR method identifies frequent probiotic origin among clinical Saccharomyces isolates. Microbiol. Res. 227, 126298 (2019); 2: Imre, A. et al. Virulence Factors and in-Host Selection on Phenotypes in Infectious Probiotic Yeast Isolates (Saccharomyces ’boulardii’). J. fungi (Basel, Switzerland) 7, 746 (2021); 3: Imre, A. et al. Heme Oxygenase-1 (HMX1) Loss of Function Increases the In-Host Fitness of the Saccharomyces ‘boulardii’ Probiotic Yeast in a Mouse Fungemia Model. J. Fungi 8, 522 (2022); 4: Pfliegler, W. P. et al. Commercial strain- derived clinical Saccharomyces cerevisiae can evolve new phenotypes without higher pathogenicity. Mol. Nutr. Food Res. 61, (2017)

[0039] Figure 2. Clinical S. ‘boulardii ’ isolates and patient data used in current study. References as on Figure 1. DUC: Debrecen, University Clinic, Hungary; SZUC: Szeged, University Clinic, Hungary

[0040] Figure 3. Oligonucleotides used for genome editing and for verification. For repair oligonucleotides, annealing region is shown in lowercase letters.

[0041] Figure 4. Virulence exhibited by the deletion strains in six -day-long infection experiments. Kidney burden. An individual data point represents the kidney burden [CFU / kidney weight (g)] of an individual mouse. Data from mice that died during or were killed at the end of the experiment are both plotted. Horizontal black lines represent the median of the data points. Whiskers extend to minimum and maximum values and individual data points are shown. Significant differences between isolates and respective deletion strains are shown as follows: *: p < 0.05; **: p < 0.01; ***: p < 0.001; ****: p < 0.0001.

[0042] Figure 5. Yeast CFU / g levels in fecal pellets during and after gavaging mice for two weeks. The inset shows the rapidly declining density of yeasts after gavaging was stopped (days 15 to 18). n.s.: non-significant. Box plots show individual data, median value as horizontal lines, and whiskers show minimum and maximum values. Gavaging was started on day 0, and finished on day 13.

[0043] Figure 6. Antagonism test results. (A) Inhibitory effect of the culture supernatants of the wild type PY0001 and its modified strains, PY0001 enal- Idenal- 1 nd PY0001 enal : :LecC / enal : :LecC against four bacterial species on Nutrient Broth agar. As a control, sterile YPD medium was used. (B) Specific inhibitory effect of the concentrated dialysed supernatant of PY0001 enal : :LecC / enal : :LecC against L. monocytogenes on Mueller-Hinton agar.

[0044] DETAILED DESCRIPTION OF THE INVENTION

[0045] Reports about infections of probiotic origin are on the rise in the recent years and probiotic yeast fungemia cases are no exception. Several S. ‘boulardii ’ clinical isolates show virulence in fungemia mouse model demonstrated by high lethality and kidney burden. Therefore, it is important to identify genetic traits that are responsible for the yeast becoming pathogenic through applying a mammalian model and stress phenotyping.

[0046] A new approach to uncover genetic factors enabling S. ‘boulardii ’ to cause fungemia and systemic infection is presented here. First the virulence of a probiotic yeast isolate collection was assessed by infection of immunosuppressed mouse model followed by kidney burden and lethality analysis. Founded on this data a selection of isolates was subjected to excessive spot plate stress phenotyping. Combined results of the data derived from mouse experiments and spot plate phenotyping led to the identification of stress resistance phenotypes and consequently genetic traits being under in-host selection pressure. This approach helped to target stress resistance genes which were subsequently deleted, and the modified strains were assessed anew, along with whole genome analysis of each isolate and its knockout version. With this method S. ‘boulardii ’ strains from six genetic backgrounds were engineered, that all show greatly diminished kidney burden as well as 100% survival of infected mice. Such strains may mitigate the risk of yeast probiotic treatments and may function as platform strains for development of designer probiotics with targeted therapeutic effects, combining safety and efficacy in a single yeast probiotic product. To illustrate this opportunity in next-generation yeast probiotics, we also show that the stress gene knockout probiotic yeast can be effectively modified to produce an md-Listeria peptide without diminishing its broad antibacterial effect or growth in standard media.

[0047] Immunosuppressed BALB / c mice were infected and followed for 6 days with 14 different S. ‘boulardi yeast isolates, originating from commercial and clinical sources. The commercial isolates were avirulent (87%-100% mouse survival), however some clinical isolates (465 / 2018, 2251 / 2018) had remarkable virulence shown by 28% and 42% survival of the applied mouse model, respectively. It is important to note that clinical isolates were derived from different niches of the human body (Figure 2), hence their potential genomic and phenotypic adaptations, triggered by the human host, might differ from one another, resulting in varying level of virulence in an animal model. This can be the reason why bloodstream infections of our mouse model with some clinical isolates result in low lethality, while the 465 / 2018 and 2251 / 2018 isolates show high lethality in mice. Additionally, higher lethality usually occurred in the case of isolates showing high kidney burden. In case of the avirulent isolates (PY0001, PY0002, DE27020, DE35762) the median kidney burden was in the range of ~102-103CFU / g, while isolates that caused high lethality (465 / 2018, 2251 / 2018, DE45866, DE42533) showed remarkably high median kidney burden in the range of 2-2.7 x lO5CFU / g. These results suggest that the level of kidney burden is a valid proxy to assess the virulence of yeast strains and isolates.

[0048] To uncover the phenotypic adaptations that enabled some isolates to be virulent, the stress tolerance of two commercial isolates (PY0001, PY0002), two isolates that showed high virulence in our mouse model (465 / 2018, 2251 / 2018), and the two blood isolates (DE6507, DE35762) was tested. In vivo and in vitro subclone lineages of the commercial and clinical isolates were isolated and this large number of subclones was used to assess what types of stress factors are most likely being under strong selection in the mammalian host. Plates were supplemented with amphotericin B, fluconazole, Congo Red, NaCl, and LiCl in spot-plate stress tolerance assays. This experiment revealed that subclones of isolate 465 / 2018, that caused the highest lethality among the mice, were highly sensitive to Congo red, and had marked NaCl tolerance both before and after mouse infection. Furthermore, all tested clinical isolates showed elevated LiCl tolerance, which increased even more (more subclones had tolerance score of 3 or 4 compared to the YPD subclones) after mouse infection and re-isolation. This shows that those isolates that were already exposed to the stress conditions present in the human host tended to adapt towards higher LiCl tolerance. Based on these results it was suspected that the high level of virulence might be correlated to genes and pathways also important in Li+and Na+ion tolerance. These ions are exported by the Nhalion antiporter or the Enal ion pump in yeast, respectively, and are involved in tolerance of various stresses, including alkaline pH. Their respective genes, NHA1 and ENA1 in the probiotic yeast were targeted, utilizing the isolate collection to represent different S. ‘boulardii ’ genetic backgrounds that differed mostly in heterozygosity but not in structural variations to study the effect of gene deletion. Using comparative genomics, it was established that these genetic editions generally did not result in unintended consequences in the genomes of the isolates. Merely one large segmental duplication was found in one strain (right arm of chr. XVI), and a ~30 kb sized deletion was uncovered in three modified strains (two NHA1 and one ENA1 mutant). Heterozygosity and other features of the strains remained remarkably consistent after the CRISPR / Cas9-based deletion of the two target genes.

[0049] The NHA1 and ENA1 deletion mutants obtained in this study were injected through the lateral tail vein into chemically immunosuppressed BALB / c mice to simulate fungemia, then lethality and kidney burden were monitored and compared with the results that were gained in the case of the wild type isolates in experimental infection lasting 6 days. The deletion of NHA 1 resulted in a surprising increased virulence in the case of three genetic backgrounds, while it decreased in the case of two. Notably median kidney burden followed the same pattern in the case of these yeasts. Thus, the NHA1 deletion did not unequivocally result in potentially safer probiotic strains. These results also highlight the usefulness of deleting a gene not just in one but multiple, yet closely related, genetic backgrounds. Epistasis and other phenomena may have a considerable effect on how phenotype and virulence is affected by the perturbations of the yeast genome during genetic modifications, similarly to a recent observations on the effect of HMX1 deletion in six S. ‘boulardii’ genetic backgrounds (Imre, A. et al. Heme Oxygenase-1 (HMX1) Loss of Function Increases the In-Host Fitness of the Saccharomyces ‘boulardii’ Probiotic Yeast in a Mouse Fungemia Model. J. Fungi 8, 522 (2022)).

[0050] Mouse survival in the case of the ENA1 deletion mutants, however, was invariable across the six genetic backgrounds and amounted to 100% across all strains during the 6-day long infection experiments, and median kidney burden values were under 2* 10 ' CFU / g, which is significantly lower than the values observed in the case of the highly virulent 465 / 2018 and 2251 / 2018 wild type isolates. Additionally, ENA1 deletion resulted in significant growth defect in the presence of LiCl, and NaCl stress. Thus, ENA1 contributed not only to the Li+and Na+tolerance but also to the virulence of the probiotic yeast isolates. This means that these engineered strains have a highly reduced potential to become pathogenic in various S. ‘boulardii ’ genetic backgrounds and hence, they may offer a safer way of application to patients as probiotic supplements. The modified knockout strain of PY0001 doesn’t show diminished growth capability in / on standard yeast media and at pH values up to pH 7.5. The antimicrobial activity of the knockout strain also matched that of the parent isolate. ENA1 can also be deleted from the probiotic yeast genome in a way that enables the heterologous expression of therapeutic molecules, as shown with the an- tilisterial peptide Leucocin C, also without negative effects on growth on / in standard yeast media or survival of lyophilization. Antifungal substance MIC values were variable among the PY0001 probiotic isolate and its modified strains but remained comparably low when the range of tested concentrations in the CLSI protocol are considered. The 21 -day-long survival tests with the wild-type probiotic yeasts and the knockout and LecC-substituted strains further show that strain engineering described herein produced yeasts with dramatically reduced chance to cause fungemia even during prolonged immunosuppression. These yeasts are potential probiotics and / or biotherapeutics that survive in the host GI tract and also similarly affect the host’s gut microbiome.

[0051] The term “functional gene” refers to a gene (DNA sequence) that is capable of fulfilling its function of encoding information for a functional molecule, i.e. a protein or RNA. The term “functional gene” refers to a gene (DNA sequence) that is capable of fulfilling its function of encoding a functional protein or to be transcribed to produce a functional RNA molecule. A “functional protein” and a “functional RNA” show their normal biological activity in an organism characteristic of the protein and the RNA in the organism under physiological circumstances. The meaning of “functional” may be the same as that of “non-deficient”.

[0052] The term “avirulent” may refer to a microorganism that is not capable of causing a disease. In case of S. 'boulardii' the term “avirulent” refers to a S. 'boulardii' microorganism that does not cause a symptom associated with a wild type, disease causing S. 'boulardii' microorganism or causes a symptom associated with a wild type, disease causing S. 'boulardii' that is less severe than a symptom caused by a wild type, disease causing S. 'boulardii'. The term “having / showing reduced virulence” may refer to a microorganism that causes less severe symptom(s) that a microorganism of the same taxonomic unit (e.g. species or strain). In case of S. 'boulardii' reduced virulence may refer to a S. 'boulardii' microorganism that causes a symptom associated with a wild type, disease causing S. 'boulardii' that is less severe than a symptom caused by a wild type, disease causing S. cerevisiae or S. 'boulardii' or does not cause a symptom caused by a wild type, disease causing S. cerevisiae or S. 'boulardii'.

[0053] The term “avirulent” may refer to a microorganism that is not capable of causing a disease in a mammal. In case of an engineered S. 'boulardii' without a functional ENA1 gene the term “avirulent” refers to a S. 'boulardii' microorganism that does not cause a symptom associated with the corresponding, disease causing S. 'boulardii' microorganism having a functional ENA 1 gene or causes a symptom associated with the corresponding, disease causing S. 'boulardii " microorganism having a functional ENA1 gene that is less severe than a symptom caused by the corresponding, disease causing S. 'boulardii' microorganism having a functional ENA1 gene. The corresponding, disease causing S. 'boulardii ' microorganism having a functional ENA 1 gene may be the S. 'boulardii ' whose ENA1 gene has been e.g. knocked out to produce the engineered S. 'boulardii' without a functional ENA1 gene. The corresponding, disease causing S. 'boulardii' microorganism having a functional ENA1 gene preferably belongs to the same taxonomic unit (e.g. strain) as the engineered S. 'boulardii' without a functional ENA1 gene. Preferably the corresponding, disease causing S. 'boulardii ' microorganism having a functional ENA 1 gene and the engineered s'. 'boulardii' without a functional ENA1 gene are essentially identical and the main differences are the presence or absence of the functional ENA1 gene and the level of virulence.

[0054] EXAMPLES

[0055] Selection regimen for wild type isolates, subclone lineages, and genetically modified strains for experiments Subclone lineages (single-cell derived colonies) were isolated from our wild-type isolate collection either after in vitro culturing or after re-isolation from mice. These are referred to as in vivo or in vitro selected lineages. This implies selection forces acting upon standing and / or de novo genetic variation (i.e. clonal heterogeneity) found among the cell population of a single isolate. For the below described infection experiments, all four commercial and ten clinical isolates in our collection were used. For further analysis, six isolates were selected to represent both commercial (n=2) and clinical (n=4) samples having higher and lower pathogenicity as well. Phenotyping experiments were carried out on ten in vitro selected and on 19-20 in vivo selected subclone lineages (from two different infected mice in all cases). Then the same six wild-type isolates were subjected to gene deletion and subsequent phenotyping and mouse infection. Finally, one probiotic isolate was chosen, based on its potential applicability as a biotherapeutic, to express an antibacterial peptide and to assess antibacterial activity before and after genetic modification. This probiotic isolate, and its gene deleted and antibacterial peptide producing derived strains were subjected to further tests of applicability.

[0056] Yeast isolates, strains, and patient data

[0057] Four commercial isolates of S. ‘boulardii ’ were used, originating from two batches each of two different probiotic supplements. For the ten clinical isolates used, detailed patient and isolation data were available (Imre, A. et al. A new, rapid multiplex PCR method identifies frequent probiotic origin among clinical Saccharomyces isolates. Microbiol. Res. 227, 126298 (2019), Imre, A. et al. Virulence Factors and in-Host Selection on Phenotypes in Infectious Probiotic Yeast Isolates (Saccharomyces ’boulardii’). J. fungi (Basel, Switzerland) 7, 746 (2021), Pfliegler, W. P. et al. Commercial strain-derived clinical Saccharomyces cerevisiae can evolve new phenotypes without higher pathogenicity. Mol. Nutr. Food Res. 61, (2017)), and they were collected from the university clinics of Debrecen and Szeged in Hungary (Figure 1 and 2). Patient data were handled in accordance with EU, state, and local regulations with a clinical study ethics approval from the Regional and Institutional Research Ethics Council of Debrecen (DE RKEB / IKEB 5194-2019). Gene deletions were performed using CRISPR / Cas9 genome editing in the case of two commercial (PY0001, PY0002) and four clinical (DE6507, DE35762, 465 / 2018, 2251 / 2018) yeast isolates (Figure 1 and 2). Stocks of the commercial and clinical isolates and deletion mutants were saved at -70 °C in YPD broth (VWR Chemicals, Solon, OH, USA, pH 5.8) supplemented with 30% v / v glycerol. Subculturing was minimized to prevent the geno- and phenotypic changes in the original isolates, except for the isolation of in vitro subclone lineages. In the latter case, the commercial and clinical isolates were all cultured in 5 mL YPD liquid medium for 24 hours at 30°C with 180 rpm shaking in culture tubes, and a dilution series was plated onto YPD agar plates. Individual colonies were randomly selected and given identifiers referring to original isolate and subclone lineage, these were also saved to our culture collection.

[0058] Lethality and kidney burden in immunosuppressed BALB / c mice and the isolation of in vivo subclones

[0059] For the kidney burden experiments BALB / c immunocompromised female mice (n= 7-9 per isolate; 21-23 g body weight; Charles River) were used. In case of PY0001, PY0002, 465 / 2018, 2251 / 2018, DE6507, DE35762 kidney burden data were used published by Imre et al. (Heme Oxygenase-1 (HMX1) Loss of Function Increases the In Host Fitness of the Saccharomyces ‘boulardii’ Probiotic Yeast in a Mouse Fungemia Model. J. Fungi 8, 522 (2022)). The animals were maintained in accordance with the Guidelines for the Care and Use of Laboratory Animals. The experiments were approved by the Animal Care Committee of the University of Debrecen, Debrecen, Hungary (permission no. 12 / 2014 DEM AB). Immuno -suppression was achieved by intraperitoneal administration of 150 mg / kg cyclophosphamide 4 days prior to infection, 100 mg / kg cyclophosphamide 1 day prior to infection, 100 mg / kg cyclophosphamide 2 days post-infection, and 100 mg / kg cyclophosphamide 5 days postinfection. Animals were infected intravenously through the lateral tail vein. The inoculum contained 1-1.5 x 107yeast cells in 0.2 mL physiological saline and for each isolate and mutant strain 7-9 mice were challenged. Inoculum density was confirmed by plating serial dilutions on Sabouraud dextrose agar. Survival of the mice were followed for 6 days post-infection, then surviving mice were euthanized, and kidneys were removed and homogenized aseptically. Kidney burden was determined by serial dilutions of the homogenized kidneys in physiological saline on YPD agar plates. Colonies were counted after 2 days of incubation at 37 °C to determine the number of living cells in the kidneys. The individual colonies were saved into our culture collection as described above and were given identifiers referring to inoculated wild-type isolate, mouse specimen, and individual colony. Thereby we possessed the original probiotic, clinical isolates, the subclone lineages from YPD medium cultures, and inhost evolved subclone lineages as well. Until the 6th day post-infection, survival of the inoculated mice was followed, and the data was used to generate 6-day survival curves which were compared as described in the Statistics section. Furthermore, 21 -day survival data was recorded for six isolates and their respective enal-A 0 / enal-A 0 and enal : :LecC / enal : :LecC modified strains. Based on results of shorter infection experiments, nhal-AO / nhal-AO mutants were excluded. BALB / c female mice (21-23 g; Charles River) were treated with 150 mg / kg cyclophosphamide 4 days and 1 day prior to infection, and with 100 mg / kg cyclophosphamide every 3 days until the 21stday post-infection. The inoculum contained 1-1.5 x 107yeast cells in 0.2 mL physiological saline and for each isolate 8 mice were used. Inoculation was performed as in the kidney burden experiment. Survival was recorded daily, and data was used for Kaplan-Meier analysis as described below.

[0060] Spot plate stress phenotyping

[0061] The presence of slightly different genetic lineages within a strain, clonal heterogeneity, and selection inside the mouse host acting upon these was taken into our focus in phenotyping. Ten in vitro subclones of wild type isolates were tested as controls, along with 19-20 in vivo selected subclones. Prior to spot-plate stress phenotyping yeasts were pre-cultured on YPD agar plates for 24 h. For the experiments, 0.5 MacFarland cell suspensions were prepared, and four 10 pl drops per isolate / subclone were inoculated in three replicates. The drops contained ~104, 103, 102, and 10 cells. Stress conditions, based on Strope et al. (The 100-genomes strains, an S. cerevisiae resource that illuminates its natural phenotypic and genotypic variation and emergence as an opportunistic pathogen. Genome Res. 125, 762-774 (2015)), were tested on synthetic defined (SD) agar plates (6.7 g / L yeast nitrogen base without amino acids, 2% glucose, 2% agar), supplemented with one of the following stressors: 0.125 pg / mL amphotericin B, 15.00 pg / mL fluconazole, 16.00 pg / mL Congo red, 342.23 mM (2 w / v %) NaCl, or 50 mM LiCl. Plates were incubated at 37°C for 3 days, then photographed to document the number of droplets showing growth. Tolerance scores were determined based on spots showing growth as follows. Score 0: no visible growth; 1: growth of the first spot (containing 104cells); 2: growth of two spots (containing 104and 103cells); 3: growth of three spots (containing 104, 103, and 102cells); 4: growth of all 4 spots (containing 104, 103, 102, and 10 cells).

[0062] Gene deletion constructs, Leucocin C gene integration, and mutant selection. Benchling (https: / / www.benchling.com / ) was used to design HDR templates and gRNA oligonucleotides for the gene knock-out experiments, considering the work of Akhmetov et al. ( Single-step Precision Genome Editing in Yeast Using CRISPR-Cas9. Bio-protocol 8, (2018)). Repair DNA was prepared by PCR reaction using Dream Taq Green polymerase (Thermo Fisher Scientific, Waltham, MA, USA). Each repair DNA included the same 20 bp insertion cassette and had uniquely designed homologous sequences to knock out the genes ENA1 and NHA1. Both of these are single-copy genes in S. ‘boulardii Oligonucleotide sequences of repair DNA and gRNA are presented in Tables 1 and 2.

[0063] The assembly of the plasmid coding for the Cas9 nuclease and gRNA was carried out using the MoClo Yeast Toolkit from Addgene (YTK, Addgene kit, Addgene, Watertown, MA, USA; catalog number: 1000000061). DH5a chemically ultra-competent Escherichia coli cells were used for all cloning steps. Cells were selected after transformation on lysogeny broth (LB, Miller) agar plates containing the appropriate antibiotics (0.1 mg / mL ampicillin, 0.025 mg / mL chloramphenicol, or 0.05 mg / mL kanamycin). Ultra-competent E. coli cells were prepared with the Inoue method (Inoue, H., Nojima, H. & Okayama, H. High efficiency transformation of Escherichia coli with plasmids. Gene 96, 23-28 (1990)) and transformations were carried out with the heat shock method (Froger, A. & Hall, J. E. Transformation of Plasmid DNA into E. coli Using the Heat Shock Method. J. Vis. Exp. 6, (2007)). Plasmid isolations were carried out with the GeneJet Plasmid Miniprep Kit (Thermo Fischer Scientific). Transformation of the HDR templates (1000 ng) and the Cas9 / gRNA coding plasmid (200 ng) into yeast were performed according to Lee et al. with modifications (Lee, M. E., Deloache, W. C., Cervantes, B. & Dueber, J. E. A Highly Characterized Yeast Toolkit for Modular, Multipart Assembly. (2015) doi: 10.1021 / sb500366v; Imre, A. et al. Heme Oxygenase-1 (HMX1) Loss of Function Increases the In-Host Fitness of the Saccharomyces ‘boulardii’ Probiotic Yeast in a Mouse Fungemia Model. J. Fungi 8, 522 (2022)). Selection of the transformed yeasts was done by using YPD plates (VWR Chemicals) supplemented with 0.1 mg / mL nourseothricin. Then multiple colonies were inoculated onto YPD agar medium to promote the loss of the Cas9 and gRNA coding plasmid carrying the nourseothricin acetyltransferase (NAT) antimycotic resistance gene. Verification of successful gene deletions were done by colony PCR, with GoTaq G2 Hot Start polymerase (Promega, Madison, WI, USA). Both the lack of the targeted gene’s ORF and the correct insertion of the HDR templates were checked. The oligonucleotides used for verification are listed in Tables 1 and 2. For gel electrophoresis, 1% low electroendosmosis (LE) agarose gel (Promega, Madison, WI, USA) was used, with Tris Acetate EDTA (TAE) buffer.

[0064] A version of the ENA1 knockout mutant with the integration of the Leucocin C (LecC) gene at the position of the gene was also designed. Leucocin C is a class Ila bacteriocin produced by a lactic acid bacterium effectively inhibiting the growth of Listeria monocytogenes and a synthetic version of this peptide’s gene has already been heterologously expressed in the probiotic yeasts from an expression plasmid (thus without chromosomal integration). In this study, a plasmid carrying the DNA fragment for the leucocin C secretion was synthesized by Integrated DNA Technologies (Coralwille, IO, USA). The sequence included the S. cerevisiae a-mating factor signal sequence to guide the extracellular secretion of the LecC peptide (Li, R., Wan, X., Takala, T. M. & Saris, P. E. J. Heterologous Expression of the Leuconostoc Bacteriocin Leucocin C in Probiotic Yeast Saccharomyces boulardii. Probiotics Antimicrob. Proteins 13, 229-237 (2021); Wan, X., Li, R., Saris, P. E. J. & Takala, T. M. Genetic characterisation and heterologous expression of leucocin C, a class Ila bacteriocin from Leuconostoc camosum 4010. Appl. Microbiol. Biotechnol. 97, 3509-3518 (2013)). A set of PCR primers were used to add the above- described homologous sequences upstream and downstream to the synthetic fragment enabling integration by homologous recombination to the ENA1 locus. The ENA1 gene was then targeted with the CRISPR / Cas9 method using the same guides as described above. For transformation 1000 ng LecC synthetic fragment and 200 ng CRISPR plasmid was used.

[0065] Whole genome sequencing and analysis

[0066] To compare the genomes of the modified strains and the original isolates, and to search for potential off -target effects and genome structure variations of the knockouts, we applied short -read sequencing and comparative genomics. Genomic DNA was isolated after 24 h growth of freshly revived stocks streaked onto YPD and incubated at 30°C. Library preparation was performed using tagmentation with the Nextera DNA Flex Library Prep kit (Illumina, San Diego, CA, USA) according to the manufacturer’s protocol, sequencing was performed using 150 bp paired-end reads on an Illumina NextSeq 500 system, with approximately 50-60 x coverage of the nuclear genome. Raw reads were deposited to NCBI SRA under BioProject PRJNA1165191. The original isolates have been sequenced before and those reads were used in Imre, A. et al. Heme Oxygenase-1 (HMX1) Loss of Function Increases the In-Host Fitness of the Saccharomyces ‘boulardii’ Probiotic Yeast in a Mouse Fungemia Model. J. Fungi 8, 522 (2022). Details of the genomics pipeline correspond to our previous analysis Imre, A. et al., supra. Briefly, filtered and trimmed reads were mapped to the PY0001 reference genome (ASM2473226vl) (Imre, A. et al., supra.), alleles were called to determine heterozygous positions, then they were compared to highlight differences among wild type isolates and modified strains. Allele ratios were plotted along the chromosomes. Coverage mapping with 10 kb windows sliding every 5 kb was used to identify large structural variations and for visualizations. Per-base coverage comparison was used to locate the exact loci of large deletions and segmental duplications and to visualize the CRISPR / Cas9 deleted loci’s neighboring regions. A draft assembly was also produced for PY0001 enal : :LecC / enal : :LecC to confirm correct integration of the leucocin C gene to the targeted ENA1 locus. Comparative and phylogenomic analysis

[0067] For Illumina sequencing, DNA isolation followed Hanna and Xiao (Isolation of nucleic acids. Methods Mol. Biol. 313, 15-20 (2006)). The Illumina FASTQ sequencing files were trimmed and filtered using fastp for further analysis (Chen, S., Zhou, Y., Chen, Y. & Gu, J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34, i884-i890 (2018)). Mapping to the PY0001 reference genome (accession no. ASM2473226vl) was performed using the mem option of BWA 0.7.17 (Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows- Wheeler transform. Bioinformatics 25, 1754-1760 (2009)). Sorted BAM files were obtained using Samtools 1.7. (Li, H. et al. The Sequence Alignment / Map format and SAMtools. Bioinformatics 25, 2078-2079 (2009)) and Picard-tools 2.23.8. was used to mark duplicated reads (Van der Auwera, G. A. et al. From fastQ data to high- confidence variant calls: The genome analysis toolkit best practices pipeline. Curr. Protoc. Bioinforma. 43, 11.10.1-11.10.33 (2013)). We used BEDTools 2.30.0 (Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841-842 (2010)) to calculate the median coverage of chromosomes in 10000 base windows sliding every 5000 bases. Plots generated from this data were corrected for ploidy and was used identify potential segmental duplications, deletions, or aneuploidies. Coverage was also visualized and compared on a per-base basis using IGV (Thorvaldsdottir, H., Robinson, J. T. & Mesirov, J. P. Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration. Brief. Bioinform. 14, 178-192 (2013)). Using BAM files, local realignment around indels and joint variant calling and filtering for the strains and isolates were performed with GATK 4.1.9.0. (Poplin, R. et al. Scaling accurate genetic variant discovery to tens of thousands of samples. bioRxiv 201178 (2018) doi: 10.1101 / 201178) with regions annotated in the PY0001 reference as centromeric regions, telomeric regions, orLTRs excluded. First, genomic VCF files were obtained with the Haplotype Caller, and joint genotyping of the gVCF files was applied. After joint calling, in the resulting VCF files, only SNPs or only INDELS were selected. SNPs were filtered according to the parameters (Fay, J. C. et al. A polyploid admixed origin of beer yeasts derived from European and Asian wine populations. PLOS Biol. 17, e3000147 (2019)): QD < 5.0; QUAL < 30.0; SOR > 3.0; FS > 60.0; MQ < 40.0; MQRankSum < -12.5; ReadPosRankSum < -8.0. INDELS were filtered according to the parameters QD < 5.0; QUAL < 30.0; FS > 60.0; ReadPosRankSum < -20.0. INDELS were then left-aligned. For the final VCF files, INDELS and SNPs were merged, filtered and non-variant sites were removed. Combined called VCF files were uploaded to FigShare (doi: 10.6084 / m9.figshare.27105919). Variants, their allelic depth, and called genotyes in the individual strains were selected and exported to a .csv file using the query option of SAMtools / BCFtools 1.10.2. Allele frequency plots were obtained by calculating the fraction of the two alleles’ dept at ach heterozygous site. Allele frequencies were used to verify ploidy, with the assumptions that disomic chromosomes have allele ratios of approx. 1 :0 or 1:1, trisomic of 1:0, 1:2, and 2: 1, tetrasomic of 1:0, 1:3, 1: 1, or 3:1, etc.( Large, C. R. L. et al. Genomic stability and adaptation of beer brewing yeasts during serial repitching in the brewery. bioRxiv 2020.06.26.166157 (2020) doi: 10.1101 / 2020.06.26.166157.). Genotype calls were used to determine levels of heterozygosity for each chromosome of the various sequenced genomes, these values were then compared.

[0068] For a draft assembly of the LecC-integrated PY0001 strain, we used Spades 3.12.0 (Pijibelski, A., Antipov, D., Meleshko, D., Lapidus, A. & Korobeynikov, A. Using SPAdes De Novo Assembler. Curr. Protoc. Bioinforma. 70, el02 (2020)) with default settings on the fastp-trimmed Illumina reads. The draft assembly produced 1127 scaffolds amounting to 11,655,577 bases, of these 44 larger than 100,000 bp. An NCBI BLAST search was used to identify the yeast-optimized LecC gene, the a-mating factor secretion signal sequence, and the regions around the knocked-out ENA1 gene in the assembled PY0001 genome. The assembly is deposited in GenBank, accession number SUB15769654.

[0069] Tolerance of gastrointestinal conditions, high pH, bile salts, and antifungals of modified strains To conduct an assessment on whether gene deletion and the integration of LecC diminished the potential probiotic applicability of the most promising modified strains, three samples were subjected to several growth and survival tests. The yeasts PY0001, PY0001 enal-AO / enal-AO, and PY0001 enal : :LecC / enal : :LecC were inoculated onto YPD agar buffered to various pH values with physiological relevance in the gastrointestinal system. The above-described spot plate assay was applied, growth was recorded after 2 days of incubation at 37°C. The used pH values in YPD plates were39: pH 6.1 (as in duodenum, using MES 50 mM buffer); pH 7.1 (as in middle small intestine, using Tris 50 mM buffer), pH 7.5 (as in distal small intestine, using Tris 50 m buffer), and pH 6 (as in cecum, using MES 50 mM buffer), along with a pH of 8.0 (using Tris 50 mM buffer) representing pronounced alkali stress for yeasts. Buffered media were filter-sterilized and mixed with an autoclaved agar solution (cooled to 65°C) to reach desired end concentrations of buffers and YPD components. Bile salts (bile salts suitable for microbiology, Sigma Aldrich) were also added to SD agar medium (cooled to 65°C) from a filter-sterilized stock solution to reach concentrations of 0.1, 0.05, and 0.025 w / v %, and spot plate inoculations were carried out as described above. Additionally, yeasts were subjected to the INFOGEST 2.040 in vitro simulation of gastrointestinal digestion protocol. Simulated salivary, gastric, and intestinal fluids with digestive enzymes and bile were prepared as described in Brodkorb et al. (INFOGEST static in vitro simulation of gastrointestinal food digestion. Nat. Protoc. 14, 991- 1014 (2019)). A yeast suspension of 5 MacFarland units was prepared in water, and a dilution series was plated onto plate count agars (PCA, from VWR) to calculate original cell density. From the undiluted suspension, 25 pl was mixed in wells of a 96-well plate and the suspension was mixed with 25 pl of simulated salivary fluid. After 2 min of incubation at 37°C, 50 pl of simulated gastric fluid was mixed to the samples and incubated for 2 h at 37°C, followed by the addition of 200 pl of simulated intestinal fluid and incubation for 2 h at 37°C. Samples were taken thereafter, diluted with physiological saline, and plated onto PCA plates to count colonies after 3 days of incubation at 37°C. The percentage of surviving colony forming units after the INFOGEST 2.0 simulated digestion was calculated, each yeast was subjected to the test in triplicates.

[0070] Furthermore, in order to assess whether any of the mentioned modified strains showed undesirable decreased antifungal susceptibility, we assessed susceptibility to fluconazole, amphotericin B, anidulafungin, caspoftmgin, and micafungin (all from Merck, Budapest, Hungary) in YPD (2% glucose, 2% peptone, 1% yeast extract) medium using the broth microdilution method, following the Clinical and Laboratory Standards Institute (CLSI) M27-A3 guideline (CLSI. M27: Reference method forbroth dilution antifungal susceptibility testing of yeasts. Wayne Clin. Lab. Stand. Inst. 4th ed, 19087-1898 (2017)). According to the guideline, the microdilution assay should be performed in RPMI-1640 medium, however, S. ‘boulardii ’ yeasts did not exhibit growth in this medium. Partial inhibition (>50% growth reduction vs. growth control) was used for fluconazole and echinocandins, while complete inhibition (100% growth reduction) was applied for amphotericin B.

[0071] Analyzing duplication time, biomass production, growth on standard yeast media, and liophilization tolerance for the modified strains

[0072] To conduct a preliminary assessment on whether gene deletion and the integration of LecC diminished the usability of the most promising modified strains as potential probiotic products, tests of growth and liophilization were carried out. The yeast PY0001 and its modified strains PY0001 enal-AO / enal-AO and PY0001 enal : :LecC / enal : :LecC were inoculated into small-scale liquid cultures in 250 mL Erlenmeyer flasks with 100 mL YPD medium, started with OD600=0.1 (optical density) of cells, in triplicates. The flasks were shaken at 180 rpm at 37°C for 24 h with regular OD measurements, after which cells were collected with centrifugation (6000 xg) and dried on filter paper (55°C, 8 h). Total dry cell mass was measured. From the OD measurements, duplication time in the exponential phase was calculated. Samples were also inoculated onto YPD agar (non-buffered), SD agar, and malt-extract (ME) agar using the above-described spot plate assay, growth was recorded after 2 days of incubation at 37°C. Samples from the same liquid culturing setup as above were also subjected to liophilization in a suspension of 120 g / L milk powder and 70 g / L trehalose. The suspension was frozen at -70°C for 2 h, then lyophilized for 3 h in triplicates (Nyanga, L. K., Nout, M. J. R., Smid, E. J., Boekhout, T. & Zwietering, M. H. Yeasts preservation: alternatives for lyophilisation. World J. Microbiol. Biotechnol. 28, 3239 (2012)). Subsequently, viable cell number was checked by plating dilution series to PCA along with a dilution series of the original suspension to calculate the ration of surviving cells.

[0073] Antagonism tests against bacteria

[0074] As earlier studies have shown antimicrobial activity of the probiotic yeast (Offei, B., Vandecruys, P., De Graeve, S., Foulquie-Moreno, M. R. & Thevelein, J. M. Unique genetic basis of the distinct antibiotic potency of high acetic acid production in the probiotic yeast Saccharomyces cerevisiae var. boulardii. Genome Res. 29, 1478-1494 (2019)), we tested whether this activity was affected by gene deletion using bacterial strains listed below: Bacillus subtilis ATCC6051, Escherichia coli ATCC11775, Klebsiella sp. UDeb-VGB2, Listeria monocytogenes NCAIM B.01934, Pseudomonas putida group UDeb-VGB 1 (strains UDeb-VGB 1 and 2 are unpublished and originate from our strain collection of foodstuff-derived bacteria). Agar-well diffusion assays were used, plates with 20 mL Nutrient Broth agar medium (HiMedia, Modautal, Germany) were inoculated with sterilized cotton swabs dipped into a 0.5 MacFarland suspension of bacteria in sterile water. After the inoculum dried, holes were punched into the agar using an 8-mm sterile pipette tip, and the bottom of holes were sealed with a droplet of molten medium. Each well was then fdled with ~150 pL of yeast culture supernatant (48 h cultures at 37°C in triplicates, as described above for biomass production tests). All agar-well plates were incubated at 37°C for 18 h and then photographed, inhibition zones were measured. For the L. monocytogenes inhibition tests, the supernatants were produced the same way, but they were also subjected to concentration of the produced LecC peptides. 40 mL supernatant was precipitated using saturation with ammonium sulfate (40% w / v), with agitation at 20°C for 30 min. Precipitates were centrifuged at 13000 xg for 30 min at 4°C, then dissolved in 1 mL of 200 mM sodium phosphate buffer (pH 6) according to Li et al. and references therein (Engineering of a probiotic yeast for the production and secretion of medium-chain fatty acids antagonistic to an opportunistic pathogen Candida albicans. Front. Bioeng. Biotech- nol. 11, (2023)). Samples were subjected to dialysis in 1 L of 20 mM PBS buffer (pH 7.2) for 24h using SnakeSkin (Sigma Aldrich) dialysis membranes with a cutoff of 5 kDa. An aliquot of dialysed samples was subjected to peptide degradation as a control using 30 min of incubation at 55°C with 5 units of Proteinase K (Zymo Research, Irvine, CA, USA). L. monocytogenes cultures were inoculated onto Mueller-Hinton agar plates with 5% horse blood (Clinichem, Budapest, Hungary) with cotton swabs as above, and the same agar well diffusion method was used with the dialysed samples as with simple supernatants. YPD controls were used for the diffusion assays. Antagonism tests were carried out with the yeasts PY0001, PY0001 enal-AO / enal-AO, and PY0001 enal: :LecC / enal : :LecC.

[0075] Mouse gavaging with yeasts

[0076] The gavaging experiment with the yeasts PY0001, PY0001 enal-AO / enal-AO, and PY0001 enal : :LecC / enal : :LecC was approved by the NC State University Institutional Animal Care and Use Committee (IACUC; protocol ID: 23-434). To test viability and clearance in the gut, 8-week-old female C57BL / 6J mice (n=24 total, n=8 per group) with normal microbiome (i.e. neither germ-free, nor specific-pathogen-free animals) were used. Mice were provided with standard chow and water ad libitum and maintained on an artificial cycle of 12-h light and 12-h dark. Mice were housed four per cage and their weight was recorded daily. Mice were acclimatized for one week before the experiment, then gavaged orally every day for 14 days with 108 cells of PY0001 and PY0001 enal-AO / enal-AO suspended in 200 pl filtered and sterilized water. Inoculum for one mouse was prepared by setting 5.0 McFarland cell suspension in 10 mL sterile water measured by McFarland densitometer, followed by centrifugation step (4000 rpm, 5 minutes), then suspension of cell pellet in 200 pl filtered and sterilized water. One group received 200 pl water as a control for the duration of the whole experiment. Cell number of gavaged yeast was assessed each day by plating the suspension on yeast peptone dextrose (YPD) agar. To assess CFU number, fecal pellets were collected on day 1, day 5, day 10, and day 14. Between day 15-18 pellets were also collected to determine yeast clearance from the gut. Mice were euthanized by CO2 and cervical dislocation on day 19. The gavage experiment and feces collection were performed in a biosafety cabinet with sterile instruments and sterile absorbent pads were used in the gavaging area, in case when pellets were dropped before collection. Weight of the pellets were measured, then pellets were suspended in 1 mL phosphate buffered saline (PBS) before plating for CFU and before DNA isolation for 16S metabarcoding sequencing. To determine CFU, samples were diluted lOx, then 50 pl suspension was plated on Dichloran Rose Bengal Chloramphenicol (DRBC) agar (Thermo Fisher Oxoid) and plates were incubated at 30°C for 2 days. CFU counts were determined as CFU / g fecal pellet.

[0077] Bacterial metabarcoding of mouse feces in feeding experiments

[0078] For 16S metabarcoding the undiluted fecal samples from days 0 and 14 obtained as described above were centrifuged at 12,100 ref for 5 minutes and the pellet was used for DNA isolation with QIAamp PowerFecal Pro DNA Kit (Qiagen LLC, USA) according to the manufacturer’s instructions. Isolated genomic DNA samples were sent for 16S (V3 / V4 region) metabarcoding analysis to SeqCenter LLC (Pittsburgh, PA, USA) using Illumina shortread sequencing as a payed service with at least 50,000 reads per sample.

[0079] Bacterial metabarcoding analysis

[0080] In the frames of the payed metabarcoding analysis at SeqCenter, sequences were imported to Qiime2 for analysis. Primer sequences were removed using Qiime2’s 14 cutadapt2 plugin using the following degenerate primer queries: CCTAYGGGNBGCWGCAG (forward) and GACTACNVGGGTMTCTAATCC (reverse). Sequences were then denoised using Qiime2’s dada2 plugin (Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods 13, 581-583 (2016)). Denoised sequences were placed into a feature table detailing which amplicon sequence variants (ASVs) were observed in which samples, and how many times each ASV was observed in each sample. The ASVs were identified taxonomically using the Silva 138 99% full-length sequence database with the VSEARCH16 utility in Qiime2’s feature -classifier plugin. ASVs were then collapsed to their lowest taxonomic units. Further analysis of the results provided by the company were carried out in-house using MicrobiomeAnalyst 2.017. First, a taxonomy file was created for the abundance data and unclear higher systematic status were manually edited to replace multiple identical entries like “uncultured bacterium” shared by unrelated species with unique ones. All genus-level names were changed to be unique, e.g. an uncultured bacte- rium entry in Lachnospiraceae was changed to Lachnospiraceae uncultured. An abundance table was also created with the modified species-level names. Then MicrobiomeAnalyst was used to filter data (low count filter; minimum count 4 in in at least 20% of occurrences) and to apply total sum scaling. Low -variance taxa were not filtered out. Stacked bar charts with relative frequency data were created for each sample, then all taxa’s relative frequencies in each group and at each timepoint was also averaged for an overview image. Alpha diversity was assessed using filtered data for genera, using Chaol diversity measure (total richness), with comparisons by ANOVA with posthoc pairwise comparisons. Beta diversity was also assessed for the genera, with PCoA method, Bray -Curtis distance index, using pairwise PERMANOVA statistics. The sequencing files for metabarcoding are deposited under the BioProject number PRJNA1358987.

[0081] Statistics

[0082] Statistical analysis was conducted by using the following online tools: VassarStats (http: / / vassarstats.net / in- dex.html), Astatsa (https: / / astatsa.com / ), and Statistics Kingdom (https: / / www.statskingdom.com / index.html). In the case of two datasets and normal distribution a two-tailed, two-sample t-test (equal variances) or Welch’s test (unequal variances) were used. In the case of a non-normal distribution, the Mann-Whitney test was used. To compare more than two datasets at once, ANOVA (normal distribution), followed by Tukey HSD test or Kruskal- Wallis (non-normal distribution), followed by Dunn post-test (corrected for Benjamini-Hochberg FDR) was used to determine which data sets differed significantly. The Log-rank (Mantel-Cox) test was applied for Kaplan-Meier mouse survival analysis. GraphPad Prism 10.2.3. was used for analysis and visualization.

[0083] Results Clinical S. ‘boulardii ’ isolates show significant differences in virulence in mice

[0084] Mouse infection experiments showed little to no difference (p > 0.05) in terms of virulence in case of the commercial isolates, but significant (p < 0.01) differences were found when clinical isolates were injected into the bloodstream of immunosuppressed mice. The following isolates showed remarkably high virulence: 465 / 2018 (28% mouse survival), DE45866, DE42533, and 2251 / 2018 (42% mouse survival).

[0085] Regarding kidney burden, the commercial isolates did not differ significantly (p > 0.05) from one another, while in the case of clinical isolates the CFU values of the isolate 2251 / 2018 were significantly higher compared to isolates DE27020, DE6507, and DE35762 [p < 0.01 (DE27020); p < 0.01 (DE6507); p < 0.01 (DE35762)]. The isolate 465 / 2018 also showed a remarkably high median CFU value (the highest one among the isolates).

[0086] Subclones of virulent isolates show increased LiCl and NaCl tolerance

[0087] To reveal whether virulence and increased survival is associated with differences in stress tolerance, spot plate stress phenotyping was applied for a set of S. ‘boulardii ’ wild-type isolates, that were chosen based on their virulence in mice. Hence, PY0001 and PY0002 (all inoculated mice survived) and 465 / 2018 as well as 2251 / 2018 (elevated virulence) were chosen for this test. The two patient hemoculture isolates, DE6507 and DE35762 were also included. In case of all six isolates, ten in vitro subclones were isolated from YPD broth, named as YPD- SubClones (YPD-SCs). Ten in vivo subclones each from the kidneys of two BALB / c mice (Mousel: Ml and Mouse2: M2) were isolated to account for clonal heterogeneity in a single strain, a common phenomenon especially in non-haploid microbes as the probiotic yeast. These were named as Mouse-SubClones (M-SCs). Based on the stress tolerance score, we determined whether the growth of the subclones was weak (score 1), medium (score 2 and 3) or strong (score 4). Subclones with tolerance score 0 were considered non-viable under the given stress. A marked difference in LiCl tolerance was observed between commercial and clinical subclones. While growth was generally weak or medium in the case of in vitro and mouse subclones of PY0001 and PY0002 commercial isolates, a significant number of in vitro and mouse subclones of clinical isolates showed strong tolerance against LiCl. Additionally, mouse subclones of these clinical isolates always showed even higher LiCl tolerance than the in vitro subclones of the same isolate.

[0088] In the case of PY0001, 40% of in vitro subclones showed weak growth on NaCl plates, while this ratio was 35% percent for mouse subclones. Some mouse subclones could not grow under this stress condition. All in vitro subclones and 85% of mouse subclones in case of 465 / 2018 showed medium growth, thus their NaCl tolerance was higher than the subclones of commercial isolate subclones. In contrast 40% of the in vitro subclones of 2251 / 2018 showed weak growth, and 10% of the subclones did not grow. In the case of mouse subclones of the same isolate, 50% did not grow and 15% grew weakly. Remarkably, 35% of the mouse subclones of DE6507 and 100% of the mouse subclones of DE35762 showed strong growth, which was not observed for the subclones of other isolates when NaCl stress was applied.

[0089] Amphotericin B tolerance was generally high among commercial isolates and their mouse subclones compared to clinical isolates and mouse subclones. 90%-100% of the commercial in vitro and mouse subclones showed strong growth in the presence of amphotericin B, while this ratio was lower in case of 465 / 2018 (YPD-SCs: 60%, M- SCs: 20%), 2251 / 2018 (YPD-SCs: 10%, M-SCs: 40%) and DE6507 (YPD-SCs: 70%, M-SCs: 45%). All in vitro subclones of DE35762 showed strong growth, however mouse subclones of the same isolate had much lower tolerance (nearly 50% of the subclones were barely viable on the plates). Tolerance against Congo Red was isolate-dependent. However, the subclones of 465 / 2018 were remarkably sensitive, since 90% of in vitro subclones and 95% of mouse subclones showed no or weak growth.

[0090] In case of fluconazole, majority of the commercial in vitro and mouse subclones of PY0001 showed weak (YPD- SCs: 20%, M-SCs: 5%) or medium growth (YPD-SCs: 80%, M-SCs: 70%). 25% of the mouse subclones showed strong growth. 90% of both PY0002 in vitro and mouse subclones showed medium growth. A high number of 465 / 2018 subclones grew weakly on fluconazole (YPD-SCs: 50%, M-SCs: 60%), meaning that these subclones were sensitive to fluconazole. Interestingly, all the in vitro subclones of 2251 / 2018, DE6507, and DE35762 showed medium growth, while all the mouse subclones of these isolates showed strong growth.

[0091] ENA1 and NHA1 as potential pathogenicity determinant genes

[0092] The ENA1 gene encodes a P-type ATPase, which maintains the ionic balance between the two sides of the cell membrane. In our S. ‘boulardii ’ isolate, PY 0001 it is located on chromosome IV from base 526602 to base 529877, as a single copy gene in contrast to the S288c reference genome, where it is found in three tandem copies. This membrane protein allows yeast to survive under high salinity or alkaline conditions. It is also capable of exporting Li+and K+ions at low salt concentrations. The baseline expression of the ENA1 gene is low, but it is rapidly induced by stress (osmotic, saline, alkaline). As a Na+ / H+antiporter the NHA1 gene is responsible for the sodium and potassium efflux at acidic pH and it is also a single-copy gene in PY000130. Interestingly, the ENA1 gene has been identified as a virulence gene in case of the pathogenic Basidiomycota yeast Cryptococcus neoformans. However, it is important to emphasize that the regulation of ENA1 and sensitivity to monovalent cations due to the deletion of this gene remarkably differs in C. neoformans and Saccharomyces. In the former, deletion mutants exhibited diminished tolerance of alkaline, but not salt stress.

[0093] Since the function of ENA1 and NHA1 might contribute to the high NaCl and LiCl tolerance (in Saccharomyces) that was observed in the case of our clinical isolates and their subclones, our hypothesis was that ENA1 and / or NHA1 are important genes for virulence, helping the colonization and pathogenicity of S. ‘boulardii To test this hypothesis, we decided to delete these genes in the strains PY0001, PY0002, 465 / 2018, 2251 / 2018, DE6507 and DE35762, leading to the generation of 12 knock-out strains (Figure 1 and 2).

[0094] Genomics analysis revealed no major differences between the wild-type and the gene knock-out strains

[0095] To investigate the importance of NHA1 and ENA1 in the virulence of our isolates, the genes were deleted in a marker-less fashion in six chosen isolates: two commercial isolates (PY0001, PY0002), two isolates that showed high virulence in immunosuppressed mice (465 / 2018, 2251 / 2018), and two isolates that originated from human bloodstream (DE6507, DE5762) using the CRISPR / Cas9 technology. Deletion of the genes was verified by PCR targeting the loci and also by sequencing and analyzing the whole genomes of the deletion strains. Compared to their respective wild type isolates, deletion strains did not show marked alterations in ploidy, genome structure, mitochondrial DNA, or heterozygosity. A single large, segmental duplication on the non-targeted chr. XVI and a deletion on chr. V was observed in DE35762 nhal-AO / nhal-AO. Furthermore, a 33,570 bp deletion was observed on the chr. IV of PY0001 nhal-AO / nhal-AO, DE6507 nhal-AO / nhal-AO, andDE6507 ena l-A0 / enal -A0 ina region not targeted by our gene deletion approach. The targeted ENA1 and NHA1 genes were all deleted starting and ending at the exact same loci on the respective chromosomes.

[0096] Heterozygosity was low in all genomes, ranging from 2336 to 2750 heterozygous positions (thereby not exceeding 0.025% of the whole nuclear genome). These heterozygosities were consistently distributed primarily on chr. I, II, III, the left arm of IV, X, XII, XIII, and XVI. In NHA1 deletion strains 95.4% to 98.3% of alleles, and in ENA1 deletion strains, 98.3 to 99.9% of alleles remained heterozygous compared to the original wild type. Mitochondrial DNA copy number variation was minor, overall calculated copy numbers ranged between -16.4 and 26.8 per haploid genome in all genomes, and the naturally occurring 2p plasmid of the yeast was present in all but one cases in higher copy numbers in knockout strains than in the wild types, their copy numbers ranging from 15.4 to 43.5 among the genomes.

[0097] Deletion of ENA1 but not NHA1 diminished virulence in S. ‘boulardii ’

[0098] The verified deletion strains were injected into immunosuppressed mice and we followed the survival of the animals and determined the kidney burden in a six-day-long infection experiment. Mouse survival of NHA1 deletion mutants was strain-dependent and varied between 35-100%. However, in the case of ENA1 deletion strains mouse survival was 100% regardless of whether the strains were derived from commercial or clinical yeast isolates. Similarly to the wild type isolates, these strains were also re-isolated from the kidneys. The median of the CFU values from individual mouse kidneys for the NHA1 deletion mutants increased with the virulence of the strain, thus in case of the avirulent PY0001 enal-AO / enal-AO it was 1.5* 10’’ CFU / g, in case of the moderately virulent PY0001 nhal-AO / nhal-AO, 2251 / 2018 nhal-AO / nhal-AO and DE6507 nhal-AO / nhal-AO strains it was 9.7* 10’’ CFU / g, 1.9* 10' CFU / g and 4.4* 10 ’ CFU / g, respectively. In the case of the most virulent strains, 465 / 2018 nhal-AO / nhal- AO and DE35762 nhal-AO / nhal-AO it was an order of magnitude higher, 1.4*105CFU / g and 1.0 * 105CFU / g respectively. On the contrary, median CFU values of the ENA1 deletion mutants were similar across the strains:

[0099] Statistical analysis of CFU / kidney weight (g) values of wild-type isolates and their corresponding knock-out strains showed the following significant differences among wild type and knockout strains: PY0001 enal-AO / enal-AO > PY0001 (p=0.000082). 465 / 2018 > 465 / 2018 enal-AO / enal-AO (p=0.020524). DE6507 enal-AO / enal-AO > DE6507 (p=0.002838). DE35762 nhal-AO / nhal-AO > DE35762 (p=0.001257). DE35762 enal-AO / enal-AO > DE35762 (p=0.002422).

[0100] To further assess virulence, another 21 -day-long infection experiment of immunosuppressed mice was conducted. As only ENA1 deletion mutants and not NHA1 deletion mutants showed promising results, the latter were omitted from this experiment. In this round of experiment, bloodstream infection with wild-type isolates resulted in 12- 62.5% survival while ENA1 deletion mutants showed values between 50-100%. (Table 1)

[0101] Table 1 Statistical comparison of survival curves with Log-rank (Mantel-Cox) method in the 21 -day infection experiments for each commercial or clinical isolate and for the respective knockout strains, using eight mice per test, n.s.: non-significant; *: p < 0.05; **: p < 0.01; and ***: p < 0.001.

[0102] In the case of PY0001, PY0002, and 465 / 2018 the deletion mutants showed significantly higher mouse survival than the wild type isolates. In the case of PY0001, the overall 25% survival of mice in the 21 days changed to 87.5% in the deletion mutant. In the case of PY0002, the 12.5% survival of the wild type changed to 100% in the ENA1 deletion strain. The wild-type 465 / 2018 isolate’s 25% survival was elevated to 75% upon ENA1 deletion. Survival curve differences were non-significant in the case of 2251 / 2018, DE6507, and DE35762 [pairwise Logrank (Mantel-Cox) tests].

[0103] The reassessment of stress tolerance in ENA1 deletion strains revealed variations across different genetic backgrounds

[0104] Based on the mouse survival and kidney burden data, we focused on ENA1 deletion strains as their pathogenic potential was greatly diminished. The knockout of ENA1 did not affect amphotericin B tolerance but resulted in higher tolerance in the case of PY0001, PY0002, and 465 / 2018 when fluconazole was applied on agar medium. Change in Congo red tolerance was strain dependent. In the case of PY0002, DE6507 and DE35762 tolerance did not change, in the case of PY0001 and 2251 / 2018 the tolerance decreased, and in the case of 465 / 2018 tolerance increased. All the ENA1 deletion strains were non-viable on LiCl and NaCl plates with the same stressor concentrations that were applied previously in the case of the wild type isolates.

[0105] Probiotic and biotherapeutic potential of ENA 1 deletion strains

[0106] Once the diminished pathogenicity of ENA1 deletion mutants was established in various genetic backgrounds, we assessed whether knockout version of the product isolate PY0001 can be cultured in standard media or show defective growth, potentially hindering their applicability as probiotics. Using small-scale laboratory tests, we showed that the knockout strain did not show significant difference in growth on standard yeast media, or in liquid cultures and had identical doubling times and biomass production. (Tables 2-5)

[0107] Table 3. Results of spot-plate assays with PY0001 and its modified strains on commonly used yeast media after 2 days of incubation at 37°C, illustrating that growth capabilities were not diminished.

[0108] Table 4. Doubling times in YPD medium at 37°C of the PY0001 and its two modified strains in triplicates, calculated from the OD values between 7 h and 9 h, rounded to two decimals. Duplication times calculated for PY 0001 and modified strains were calculated as follows. PY0001: 2.11 hours. PY0001 enal-AO / enal-AO'. 2.19 hours.

[0109] PY0001 enal : :LecC / enal : :LecC'. 2.16 hours.

[0110] Table 5. Dry biomass production of the wild type isolate PY0001 and its two mutants in 100 mL YPD medium at 37°C in 24 h

[0111] Based on these favorable results, ENA1 was not merely deleted but also successfully substituted in PY0001 with a synthetic Leucocin C bacteriocin-coding construct. The resulting recombinant PY0001 enal : :LecC / enal : :LecC strain displayed complete survival of infected mice in the 21 -day-long experimental infection and also showed similar growth as the wild type or the PY0001 enal-AO / enal-AO. Experiments on lyophilization survival showed PY0001 enal::LecC / enal::LecC to have the highest values, followed by PY0001, and PYOOOl enal-AO / enal-AO under small-scale laboratory conditions (Table 6).

[0112] Table 6. Survival percentage of cells of PY0001 and its two modified strains in triplicates after liophilization. *** p < 0.001; ****: p< 0.0001.

[0113] PY0001 enal: :LecC / enal: :LecC was also subjected to comparative genomic analysis as described above and was shown to have a trisomy of chr. I. A draft assembly confirmed the correct integration of the LecC gene with a 5’ a-mating factor secretion signal sequence at the locus of the target ENA1. Heterozygosity in this strain increased to 102.8% compared to the original PY0001 isolate.

[0114] Regarding the tolerance of higher pH values that characterize the human gastrointestinal system where probiotics are mostly expected to colonize, PY0001 enal-AO / enal-AO and PY0001 enal : :LecC / enal : :LecC were tested on buffered media for their growth. At pH values representing those found in the duodenum, middle and distal small intestine, and cecum (pH 6.1, 7.1, 7.5 and 6.0, respectively), both the PY0001 probiotic isolate and the two deletion strains showed high tolerance, while at a pronounced alkali stress of pH 8.0 the deletion strains showed markedly decreased tolerance (Table 7). Tolerance of bile acids was also shown to be similar in PY0001, PY0001 enal- AO / enal-AO, and PYOOOl enal : :LecC / enal : :LecC (Table 7).

[0115] Table 7. Results of spot-plate assays withPYOOOl and its modified strains on YPD buffered to various pH values, and on bile acid containing SD agar plates.

[0116] We also determined MIC values for planctonic cells in liquid medium for commonly used antimycotics for these three yeasts, as a substantially increased tolerance of antifungals would be undesirable in probiotic strains. For Amphotericin B, the MIC values was diminished in PY0001 enal::LecC / enal::LecC (1 mg / L to 0.5 mg / L), and for fluconazole, both PY0001 enal-AO / enal-AO and enal : :LecC / enal : :LecC showed diminished values (2 mg / L to 0.5 mg / L). However, the two deletion strains displayed increased MIC values for caspofungin (0.004 mg / L to 0.06 and 0.03 mg / L) and micafungin (0.5 mg / L to 1 mg / L). (Table 8)

[0117] Table 8. MIC values (mg / L) determined for PY0001 and its modified strains for five antifungal agents. For each antifungal, the applied concentrations in the antifungal test panel are given in parentheses.

[0118] To test antibacterial activity of the PY 0001 isolate and its two modified versions, we assessed the inhibitory effect of the culture supernatants of the yeasts cultured in YPD liquid medium. All three tested yeasts could inhibit the growth of the tested E. coli and B. subtilis strains in a similar manner, without significant differences. (Table 9)

[0119] Table 9. Inhibition zones of the supernatants and concentrated supernatants of PY0001 and its modified strains in agar well diffusion assays against various bacteria. Results are compared using on one-way ANOVA and Tukey

[0120] Its ability to express and secrete the bacteriocin was verified by the inhibitory effect of its concentrated and dialysed supernatant against L. monocytogenes. This effect was completely abolished by protease treatment of the dialysate, andPYOOOl andPYOOOl enal-AO / enal-AO idnot show any inhibition against this bacterium, verifying that the bacteriocin production was successfully achieved in PY0001 enal::LecC / enal::LecC. (Figure 6) Avirulent S. ‘boulardii ’ strain (PY0001 enal-AO / enal-AO) showed comparable gut viability to the original commercial S. ‘boulardii ’ (PY0001)

[0121] Since the ENA1 knock-out strain of PY0001 showed drastically diminished virulence in our mouse fungemia model, this knockout, the PY0001 enal : :LecC / enal : :LecC strain, and their parental strain (PY0001) were chosen to test and compare viability in the gastrointestinal tract. First, in vitro assessment showed that the survival under simulated gastrointestinal conditions was not changed in the knockout strains compared to PY0001. Table 10. Table 10. Survival of PY0001 and its two modified strains under simulated gastrointestinal conditions (simulated salivary fluid, gastric fluid, and intestinal fluid) in percentage of total cell number at the start of the experiment.

[0122] Based on these favorable results, PY0001 and the PY0001 enal-AO / enal-AO knockout strain were tested for their ability to colonize the GI tract in C57BL / 6J mice. We showed that all mice gavaged for 14 days survived and mouse weight was statistically not different in the two yeast-gavaged groups compared to the control group gavaged with water. Colony forming unit was determined on day 1, day 5, day 10 and day 14-18 by plating suspended fecal pellets on DRBC agar plates described above. No colonies were observed in case of the control group gavaged with water. In case of the groups that were gavaged with yeast, there was no significant difference in CFU values of the two yeast strains on days 1 (day after first gavaging), 5, 10, and 14 (day after last gavaging) when sampling was performed. Clearance between day 15 and day 18 was similar as well, since CFU values did not show significant difference between the two yeasts. The gavaged yeasts reached -9.64* 10’’ to -2.59* 10sCFU / g cell density in the case of PY0001 during the gavaging experiment (days 1 to 14, the day after the first gavaging and after the last, respectively) and 0 to ~1.25*106CFU / g in the case of PY0001 enal-AO / enal-AO in the fecal pellets. The CFU / g values varied considerably, but only in the case of a single mouse and a single timepoint was the CFU / g value below detection. Mean CFU / g values for the two yeasts were -8.49* 10 ' (S.D. -7.10* 10 ') and 2. 10* IO5(S.D. ~2.62*105) on days 1 to 14 for PYOOOl and PYOOOl enal-AO / enal-AO. There was no statistical difference between the two yeasts neither during the gavage experiment (days 1 to 14), nor after the gavaging (days 15 to 18) when the yeast CFU number rapidly declined in the case of both yeast samples, in most cases to below detection level (Figure 5).

[0123] During the gavaging experiments, fecal samples were also used to isolate DNA and to identify bacterial microbiome through 16S metabarcoding. Fecal samples from all control and yeast-gavaged mice were subjected to this analysis, samples were collected on day 0 just before the first gavaging, the day after last gavaging (14 d) and five days after last gavaging (18 d). The most abundant genus-level taxa were an unidentified Muribaculaceae, Lactobacillus, and an unidentified Lachnospiracae taxon. Changes in the bacterial microbiota were apparent among the samples. Lactobacilli in all three groups decreased in abundance from day 0 to day 14, and increased from day 14 to 18, while other taxa showed a variable picture. Analyses of alpha diversity showed that the mouse groups were similar at the start of the experiment in the case of the control group and the PY0001 -treated group, and the PY0001 enal-AO / enal-AO treated mice were more diverse than the control group. Groups at the end of gavaging did not differ. Start and end of gavaging samples in each group differed significantly, as well as end of gavaging and end of experiment samples in each group - alpha diversity increased when the yeast was gavaged, and decreased when gavaging stopped in all groups. Regarding beta diversity, the mouse groups did not differ at the start of the experiment. After 14 days, both yeast-treated mouse groups’ bacterial microbiota differed from that of the control group, but not from each other. Beta diversity at the start and end of gavaging differed significantly in all groups, as did end of gavaging vs. end of experiment in each group. Results of bacterial alpha-diversity comparisons in the fecal samples of mice gavaged with PY0001 and PY0001 enal-AO / enal-AO and control animals on day 0, day 14, and day 18 of the experiment showed the following significant differences (with FDR-adjusted p-values in brackets). Groups of mice are those used as controls, those gavaged with PY0001, and those gavaged with PY0001 enal-AO / enal-AO.

[0124] Groups at the start of the experiment Control day 0 < PY0001 enal-AO / enal-AO day 0 (p=0.030707).

[0125] Groups at the end of experiment: control day 18 < PY0001 enal-AO / enal-AO day 18 (p=0.047658).

[0126] Start vs. end of gavaging in each group: control day 0 < control day 14 (p=0.001167). PY0001 day 0 < PY0001 day 14 (p=0.001167). PY0001 enal-AO / enal-AO day 0 < PY0001 enal-AO / enal-AO day 14 (p=0.0023243).

[0127] End of gavaging vs. end of experiment in each group: control day 14 > control day 18 (p=0.0018824). PY0001 day 14 > PY0001 day 18 (p=0.0018922). PY0001 enal-AO / enal-AO day 14 > PY0001 enal-AO / enal-AO day 18 (p=0.027836).

[0128] All other comparisons of groups at the start of the experiment, groups at the end of gavaging, and groups at the end of experiment were statistically not different. All comparisons of start vs. end of experiment in each group were statistically not different. For these comparisons, filtered bacterial abundance data was subjected to alpha diversity calculation based on genera, with Chaol diversity measure, and ANOVA with Welch’s T-test post-hoc pairwise comparisons. Multi-testing adjustment is based on Benjamini-Hochberg procedure, all conducted in Mi- crobiomeAnalyst 2.0. p-value for the ANOVA test of all samples was p=3.716e-l 1.

[0129] Results of bacterial beta-diversity comparisons in the fecal samples of mice gavaged with PY0001 and PY0001 enal-AO / enal-AO and control animals on day 0, day 14, and day 18 of the experiment showed the following significant differences (with FDR-adjusted p-values in brackets). Groups of mice are those used as controls, those gavaged with PY0001, and those gavaged with PY0001 enal-AO / enal-AO.

[0130] Groups at the end of gavaging: control day 14 vs PY0001 day 14 (p=0.025714). Control day 14 vs PY0001 enal- AO / enal-AO dayl4 (p=0.024).

[0131] Groups at the end of experiment: control day 18 vs PY0001 enal-AO / enal-AO day 18 (p=0.0025714). PY0001 day 18 vs PY0001 enal-AO / enal-AO day 18 (p=0.0276).

[0132] Start vs. end of gavaging in each group: control day 0 < control day 14 (p=0.0036). PY0001 day 0 < PY0001 day 14 (p=0.0051429). PY0001 enal-AO / enal-AO day 0 < PY0001 enal-AO / enal-AO day 14 (p=0.032516).

[0133] End of gavaging vs. end of experiment in each group: control day 14 > control day 18 (p=0.0025714). PY0001 day 14 > PY0001 day 18 (p=0.0036). PY0001 enal-AO / enal-AO day 14 > PY0001 enal-AO / enal-AO day 18 (p=0.0025714).

[0134] Start vs. end of experiment in each group: control day 0 vs control day 18 (p=0.0025714). PY0001 day 0 vs PY0001 day 18 (p=0.0036). PY0001 enal-AO / enal-AO day 0 vs PY0001 enal-AO / enal-AO day 18 (p=0.0025714). All comparisons of groups at the start of the experiment, all other comparisons of groups at the end of gavaging, and groups at the end of experiment were statistically not different. For these comparisons, filtered bacterial abundance data was subjected to beta diversity calculation based on genera, using Bray -Curtis distance index and pair- wise PERMANOVA for comparisons. Multi-testing adjustment is based on Benjamini-Hochberg procedure, all conducted in MicrobiomeAnalyst 2.0. p-value for the PERMANOVA test of all samples was p=0.001.

Claims

CLAIMS1. Saccharonyces ‘boulardii ’ that is engineered to lack a functional ENA1 gene.

2. A method for producing Saccharonyces ‘boulardii ’ with a decreased virulence, comprising engineering the Saccharonyces ‘boulardii ’ to lack a functional ENA1 gene.

3. The S. ‘boulardii ’ according to claim 1 , wherein the S. ‘boulardii ’ does not comprise any functional ENA 1 genes.

4. The method according to claim 2, wherein the S. ‘boulardii ’ is engineered to not comprise any functional ENA1 genes.

5. The S. ‘boulardii ’ or the method according to any one of the preceding claims, wherein the ENA1 gene is knocked out.

6. The S ‘boulardii ’ or the method according to any one of the preceding claims, wherein the engineered S. ‘boulardii ’ is viable in the intestine of a mammal.

7. The S. ‘boulardii ’ or the method according to any one of the preceding claims, wherein the engineered S. ‘boulardii ’ further comprises a transgene or a synthetic protein- or peptide-coding construct.

8. The S. ‘boulardii ’ or the method according to claim 7, wherein the transgene or synthetic protein- or peptide- coding construct is localized at the locus of the ENA1 gene.

9. The S. ‘boulardii ’ or the method according to claim 7 or 8, wherein the transgene or synthetic protein- or peptide- coding construct codes for a bacteriocin.

10. The S. ‘boulardii ’ or the method according to claim 9, wherein the bacteriocin is leucocin C.

11. S. ‘boulardii ’ having the accession number NCAIM (P) Y 001536 or a derivative S. ‘boulardii ’ thereof, wherein the derivative S. ‘boulardii ’ does not have a functional ENA1 gene.

12. A composition comprising the S. ‘boulardii ’ according to any one of claims 1, 3, 5-11, the S. ‘boulardii ’ produced by the method according to any one of claims 2 and 4-11 or the derivative S. ‘boulardii ’ according to claim 11.

13. The composition according to claim 12, wherein the composition is a probiotic composition or a pharmaceutical composition.

14. The S. ‘boulardii ’ according to any one of claims 1, 3, 5-11, the S', ‘boulardii ’ produced by the method according to any one of claims 2 and 4-11 or the derivative S. ‘boulardii ’ according to claim 11 for use in therapy.

15. Use of the S', ‘boulardii ’ according to any one of claims 1, 3, 5-11, the S. ‘boulardii ’ produced by the method according to any one of claims 2 and 4-11 or the derivative S. ‘boulardii ’ according to claim 11 as a probiotic16. The S. ‘boulardii ’ according to any one of claims 1, 3, 5-11, the S. ‘boulardii ’ produced by the method according to any one of claims 2 and 4-11, the derivative S. ‘boulardii ’ according to claim 11 or the pharmaceutical composition according to claim 13 for use in the treatment or prevention of diarrhea, preferably Clostridioides difficile-related, HIV / AIDS -associated, rotaviral or antibiotic -associated diarrhea, irritable bowel syndrome, Clostridioides difficile infection, Helicobacter pylori infection, blastocystosis or acute gastroenteritis.

Images