Effects of broad-spectrum antibiotics on the colonisation of probiotic yeast Saccharomyces boulardii in the murine gastrointestinal tract

Mouse models are commonly used to study the colonisation profiles of microorganisms introduced to the gastrointestinal tract. Three commonly used mouse models include conventional, germ-free, and antibiotic-treated mice. However, colonisation resistance in conventional mice and specialised equipment for germ-free mice are usually limiting factors in their applications. In this study, we sought to establish a robust colonisation model for Saccharomyces boulardii, a probiotic yeast that has caught attention in the field of probiotics and advanced microbiome therapeutics. We characterised the colonisation of S. boulardii in conventional mice and mice treated with a cocktail of broad-spectrum antibiotics, including ampicillin, kanamycin, metronidazole and vancomycin. We found colonisation levels increased up to 10,000-fold in the antibiotic-treated mice compared to nonantibiotic-treated mice. Furthermore, S. boulardii was detected continuously in more than 75% of mice for 10 days after the last administration in antibiotic-treated mice, in contrast to in nonantibiotic-treated mice where S. boulardii was undetectable in less than 2 days. Finally, we demonstrated that this antibiotic cocktail can be used in two commonly used mouse strains, C57BL/6 and ob/ob mice, both achieving ~ 108 CFU/g of S. boulardii in faeces. These findings highlight that the antibiotic cocktail used in this study is an advantageous tool to study S. boulardii based probiotic and advanced microbiome therapeutics.

www.nature.com/scientificreports/ Established genetic tools for S. cerevisiae 26,33,34 can also be applied to S. boulardii as they are closely related 35,36 , allowing S. boulardii to be used as an AMT chassis. Examples of this include engineering S. boulardii to secrete IL-10 to enhance its anti-inflammatory properties 37 and to secrete atrial natriuretic peptide to treat murine colitis 38 , thus demonstrating S. boulardii as a promising AMT platform, in addition to a probiotic itself. Regrettably, S. boulardii has not been considered an efficient coloniser in conventional rodents with an intact microbiome due to its low abundance and short residence time in the gastrointestinal tract 25,26,39,40 . For instance, in conventional mice S. boulardii was completely washed-out after two days, even though it can survive the low pH of the stomach and harsh conditions of gastrointestinal tract 41,42 . Conversely, S. boulardii administered to germ-free mice colonise the gastrointestinal tract for 30 days after a single administration 26 . However, the cost and accessibility of such mouse model are limiting factors. Studying a disease that requires a stable delivery of the therapeutic would require more consistent S. boulardii levels. However, accessibility to a germ-free murine model and poor colonisation in conventional murine models complicates the preclinical assessment of the probiotic yeast. Considering the importance of robustness in preclinical assessment, a need remains for standardised colonisation and testing of S. boulardii in murine models. To address this problem, we sought to establish an antibiotic-treated model for S. boulardii to achieve reproducible and robust colonisation in mice.

Results
Pre-treatment with an antibiotic cocktail significantly extends Saccharomyces boulardii colonisation in mice. Use of an antibiotic-treated mice model for S. boulardii should reduce the potential variability in colonisation introduced with the environment, diet, breed, and other factors, which alter the gut microbiota 43 . We hypothesised that treatment with a combination of broad-spectrum antibiotics would eliminate gut bacteria competing with S. boulardii, allowing for more robust colonisation and better reproducibility. We used an antibiotic cocktail containing ampicillin, kanamycin, metronidazole, and vancomycin enabling us to target multiple bacteria in the murine gastrointestinal tract [44][45][46] . We evaluated the in vitro growth performance of S. boulardii in the presence of each antibiotic individually and in combination ( Supplementary Fig. S1).
We evaluated the effect on colonisation and washout of S. boulardii in two cohorts of mice receiving either drinking water supplemented with antibiotics or normal drinking water (Fig. 1a). To confidently quantify colonisation levels, we orally administered ~ 10 7 cells of a genetically engineered S. boulardii expressing green fluorescent protein (GFP) and an antibiotic resistance selection marker (geneticin selectable marker, KanMX). Antibiotic supplemented drinking water enhanced the colonisation of S. boulardii with 100% of the treated mice showing detectable colonisation of S. boulardii for the first six days, with 75% remaining with detectable levels for a prolonged period up to 10 days (Fig. 1b). In contrast to the nonantibiotic-treated mice that demonstrated complete washout of S. boulardii after a short period (up to 2 days). Additionally, we observed a more than a 1,000-fold increase of viable CFU/g in faeces during the period of daily oral administrations of S. boulardii for the antibiotic-treated mice compared to nonantibiotic-treated mice ( Fig. 1c-d).
We further determined the colonisation of S. boulardii in different regions of the gastrointestinal tract. In all of the collected gut regions we observed a significant increase of viable S. boulardii CFU in the antibiotic-treated mice ( Fig. 1e-f). Furthermore, S. boulardii was able to colonise all regions of the antibiotic-treated mice, while in the nonantibiotic-treated mice it was undetectable for most of the stomach and small intestine samples (Fig. 1f).
Antibiotic cocktail opens a niche in the murine gastrointestinal tract for S. boulardii colonisation. Once the residence time was assessed we sought to investigate the effect on the total bacterial communities in antibiotic-treated mice. The mice were orally administered ~ 10 8 cells of S. boulardii and faecal samples were collected daily until sacrifice on day five (Fig. 2a). Receiving an antibiotic cocktail reduced the total bacterial abundance by four orders of magnitude (Fig. 2b). Additionally, receiving the oral administration of ~ 10 8 S. boulardii cells together with the antibiotic cocktail resulted in an abundance three orders of magnitude greater than the nonantibiotic-treated mice (Fig. 2c).

Discussion
Since the early days of microbiome research, mouse models have been used to investigate how microbes interact with their hosts. In particular, germ-free and antibiotic-treated mice have been a valuable tool to study the direct interaction of a single or a defined community of microbes in the gastrointestinal tract 45 . In this study, we characterised the effect of an antibiotic cocktail to build a robust mouse model for colonisation of S. boulardii. We started with orally administered ~ 10 7 cells (Fig. 1a) to correlate our observed colonisation with previous studies with S. boulardii 26 . The oral administered S. boulardii was later increased by tenfold to boost to colonisation (Figs. 2a, 3a). Here we demonstrated that treating mice with a cocktail of antibiotics enabled S. boulardii to www.nature.com/scientificreports/ remain in 75% of mice for 10 days and 25% of mice for 15 days (Fig. 1b) following the last oral administration. This was in stark contrast to the colonisation observed in the nonantibiotic-treated mice that demonstrated complete washout of S. boulardii after two days. This latter observation appears identical to a previous study that compared S. boulardii in mice treated with a cocktail of penicillin (1 mg/mL) and streptomycin (2 mg/mL), where S. boulardii still appeared to washout after two days 26 . This emphasised the importance of the use of the broadspectrum antibiotic cocktail in this study. In addition, we observed more than a 1,000-fold increase in viable count of S. boulardii in faeces from antibiotic-treated mice compared to nonantibiotic-treated mice (Figs. 1d, 2c). The gut microbiota of mice from the same breed has previously been shown to differ significantly between suppliers resulting in challenges to reproduce experiments [50][51][52] . In this study, we demonstrated the robustness of the antibiotic cocktail model, showing an insignificant difference in colonisation for the first four weeks between two different studies ( Fig. 3b-d). We showed that the antibiotic cocktail in mice exhibits reproducible colonisation of S. boulardii in two different mouse strains (C57BL/6 and ob/ob) under two different diets (standard chow and high fat diet). These studies were conducted independently with two different batches of S. boulardii, maintaining ~ 10 8 CFU/g in faeces for up to five weeks (Fig. 3b). Interestingly, here we observed ~ tenfold higher CFU/g counts in faeces of antibiotic-treated mice compared to previously reported numbers in germ-free mice 26 .
Comparing the viable counts in the different regions, we observed a significant increase in all regions by supplementing antibiotics in the drinking water, highlighting the potential of S. boulardii to act on multiple sections of the gastrointestinal tract (Fig. 1f). We observed the largest increase in viable cells in the jejunum and stomach, which increased by ~ 10,000-fold when antibiotics were supplemented. This was followed by the ileum and lower intestinal tract, which all increased by ~ 1,000-fold viable cells. Furthermore, the ability of S. boulardii to colonise the antibiotic-treated mice were 10-to 100-fold lower than some of the more common bacterial probiotics 53,54 , such as E. coli Nissle 1917 which can reach up to 10 9 cells/g in faeces 55 .
Treating mice with broad-spectrum antibiotics has been commonly used to deplete bacterial populations in the gastrointestinal tract. We used a cocktail of antibiotics that has previously been demonstrated to open a niche in the gastrointestinal tract [44][45][46] . In our study, using this antibiotic cocktail showed no growth burden for S. boulardii (Supplementary Fig. S1). Distinctly, the antibiotic cocktail of this study significantly impacted the total bacterial abundance, as observed by a 10,000-fold decrease of bacterial cells (Fig. 2b). Although the treatment www.nature.com/scientificreports/ cannot completely eliminate the gut microbiota 56 , it represents a rapid, inexpensive and accessible alternative to the germ-free model 45 . Further studies would be necessary to fully elucidate the long-term colonisation effects and dynamics of S. boulardii as well as other fungal species on the host. Indeed, it has been demonstrated that antibiotics have a long-term effect on the fungal community in the gut, shifting from mutualism towards competition 57 . To reduce the risk of other fungal species taking over or competing with S. boulardii, anti-fungal compounds that S. boulardii has been engineered to resist could be added to the antibiotic cocktail. Investigation of physiological effects of S. boulardii based probiotics and AMTs requires animal models that allow robust and consistent colonisation. In this study, we demonstrated that treatment with an antibiotic cocktail led to reproducible and robust colonisation of S. boulardii throughout the gastrointestinal tract. This model may be useful as a rapid, low-cost setup for testing S. boulardii AMTs and probiotics.   Table S1). All plasmids used and generated in this study are listed in Supplementary Information Table S3. Phusion high-fidelity DNA polymerase (Thermo Scientific, Waltham, MA, USA) was used for amplifying GFP. GFP was assembly with the integration plasmid pCfB2055 using Gibson Assembly 58 and digested with NotI prior to integration. Colony-PCR using OneTaq (Thermo Scientific, Waltham, MA, USA) was used for confirming the integration of pCfB2055_GFP. Primers flanking the integration was used to confirm the integration. One single amplification band on gel electrophoresis indicated a successful integration into both chromosome copies. Genomic DNA extraction was performed as described previously 59 .

Strains and media. One Shot® TOP10
To generate the uracil knockout strain, the gBlock with URA3 gRNA 34  Animal study. All animal experiments were conducted according to the Danish guidelines for experimental animal welfare and the study protocols were approved by the Danish Animal Experiment Inspectorate (license number 2020-15-0201-00405). The study was carried out in accordance with the ARRIVE guidelines 61 . Unless otherwise stated, all mice were housed on a 12-h light/dark cycle and given ad libitum access to standard chow (Safe Diets, A30) and water. For the high fat diet studies, 45% high fat diet (Research Diet, D12451 Rodent Diet with 45 kcal% fat) was fed ad libitum. Mouse strains used in this study were female NMRI mice (6-8 weeks old; Taconic Bioscience), male ob/ob mice -B6.Cg-Lep ob / J (The Jackson Laboratory) and male C57BL/6 mice (6-8 weeks old; Taconic Bioscience). All mice were randomised according to body weight and were acclimated for at least one week prior to the first oral administration of S. boulardii. The cohorts received the antibiotic cocktail (0.3 g/L ampicillin sodium salt, 0.3 g/L kanamycinsulfate, 0.3 g/L metronidazole, and 0.15 g/L vancomycin hydrochloride) supplemented in the drinking water, for up to five days prior to the first oral administration of S. boulardii. A single batch of S. boulardii was cultivated for 72 h before being harvested and frozen. The batch was used for the entire study. The impact of preservation of S. boulardii at − 80 °C in cryovials over time was assessed. Viable S. boulardii was assessed three, five and ten days after being frozen. Colony forming unit (CFUs) were count two days after plating ( Supplementary Fig. S2). Each animal study received a freshly prepared batch of S. boulardii. The researcher was unblinded for all mouse experiments.
Assessing the residence time (Fig. 1). The NRMI mice were divided into two cohorts (n = 16), with and without antibiotics supplemented to the drinking water. Both cohorts were orally administered via intragastric gavage with ~ 10 7 CFU of S. boulardii in 100 µL of 1 × PBS and 10% glycerol. The mice were orally administered S. boulardii for five consecutive days and faeces were collected daily for 20 days. On day five, half of the mice from each cohort (n = 8) were sacrificed. The gut content from six gastrointestinal tract regions was collected in 1 mL of 1 × PBS and 50% glycerol and weighed to determine the viability per gram of content.
Assessing the effect of antibiotics (Fig. 2). The NRMI mice were divided into two cohorts (n = 6), one with and one without antibiotic-treatment. Both cohorts were orally administered via intragastric gavage with ~ 10 8 CFU of S. boulardii in 100 µL of 1 × PBS and 10% glycerol. The mice were orally administered S. boulardii for five successive days and faeces were collected daily. The cohort receiving the antibiotic cocktail in drinking water also received an oral administration of antibiotics (1.0 g/L ampicillin sodium salt, 1.0 g/L kanamycinsulfate, 1.0 g/L metronidazole, and 1.0 g/L vancomycin hydrochloride).
Assessing the robustness of the antibiotic cocktail (Fig. 3). The C57BL/6 mice (n = 9) and ob/ob mice (n = 9) were orally administered via intragastric gavage with ~ 10 8 CFU of S. boulardii URA3∆ background strain in 100 µL of 1 × PBS and 10% glycerol to test the robustness of a strain without an antibiotic selection marker. The mice were housed individually to reduce the shared gut microbiota composition between each mouse 62 . In addition to receiving the antibiotic cocktail in drinking water, the mice also received an oral administration of the antibiotic cocktail (1.0 g/L ampicillin sodium salt, 1.0 g/L kanamycinsulfate, 1.0 g/L metronidazole, and 1.0 g/L vancomycin hydrochloride).
CFU assessment in the faecal samples. The faeces were collected in pre-weighted 1.5 mL Eppendorf tubes containing 1 mL of 1 × PBS and 50% glycerol and weighed again to determine the faecal weight. All sample preparation for assessing CFU numbers was kept at 4 °C and followed the same practice. The faecal samples www.nature.com/scientificreports/ were homogenised and vortexed at ~ 2400 rpm for 10 min. The samples were then spun down at 100 g for 30 s, followed by a dilution series, where 5 µL of each dilution was plated in duplicates or triplicates. The faecal samples were plated on YPD plates containing 200 mg/L G418 for selecting the S. boulardii containing the KanMX marker. For selecting the uracil knockout S. boulardii the faecal samples were plated on SC plates containing 100 mg/L ampicillin, 50 mg/L kanamycin, and 30 mg/L chloramphenicol (Sigma Aldrich-Merck Life Science) for preventing bacterial growth and 1 g/L 5-Fluoroorotic Acid (5-FOA; Nordic BioSite) for preventing other yeast species to grow. To lower the limit of detection, 100 µL of the faecal samples was plated in duplicates from day 10 (Fig. 1b). Bacterial cell counts were determined using 16S rRNA and GFP specific primers for the total bacteria and S. boulardii cell counts, respectively (Supplemented Table S1). Cell counts were calculated based on a standard curve generated using a series of OD from either E. coli or S. boulardii and yielded genomic DNA concentration. We adjusted the qPCR signal of 16S and GFP by dividing them by the mean copy numbers per genome 63 ; this was done to account for the fact that the 16S and GFP copy numbers per genome vary between species [63][64][65] , which consequently skew the readout of bacterial and S. boulardii cell counts. qPCR runs were performed on the Roche LightCycler® 480 Real-time PCR System in LightCycle® 480 Multiwell Plate 384, clear plates.

Real
Statistical testing. All statistical analysing were performed in RStudio. The normality of the data was checked with the Shapiro-Wilk test. A dependent sample t-test was used for normally distributed data and the Wilcoxon singed-rank test for non-normal distributed data. P values < 0.05 were considered statistically significant.

Data availability
All data generated or analysed during this study are included in this published article and its Supplementary Information files.