Remodeling of the Candida glabrata cell wall in the gastrointestinal tract affects the gut microbiota and the immune response

The gastrointestinal (GI) microbiota acts a natural barrier to the proliferation of opportunistic pathogens. Candida glabrata is an opportunistic yeast pathogen that has adapted to colonize all segments of the human GI tract. We observed an increase in Escherichia coli, Enterococcus faecalis, and Bacteroides vulgatus populations, and a decrease in Lactobacillus johnsonii, Bacteroides thetaiotaomicron, and Bifidobacterium animalis in mice with DSS-induced colitis. This reduction was more pronounced for L. johnsonii during C. glabrata overgrowth. In addition, C. glabrata overgrowth increased mouse mortality and inflammatory parameters, and modulated the expression of intestinal receptors and signaling pathways. The C. glabrata cell wall underwent various changes during the course of C. glabrata colonization, and showed a significant increase in chitin. C. glabrata deficient in chitin synthase-3 induced fewer inflammatory parameters than the parental strain during intestinal inflammation. Oral administration of chitin attenuated the impact of colitis, and reduced the number of aerobic bacteria and C. glabrata overgrowth, while chitinase-3-like protein-1 increased. This study provides evidence that inflammation of the gut alters the microbial balance and leads to C. glabrata cell wall remodeling through an increase in chitin, which is involved in promoting persistence of C. glabrata in the gut.

The fungal cell wall is the predominant site of interaction between the fungus and its host. This cell wall consists of a complex structure of polysaccharides, proteins, and lipids 7 , but its composition is dynamic, responding to changes in the local environment 7,10 . Expansion of the fungal wall during growth involves permanent remodeling of the cell wall polysaccharide network, which is comprised of three major types of polysaccharide: mannans, β-glucans, and chitin. Chitin is a homopolymer of β1,4-N-acetylglucosamine (GlcNAc) and is essential for biological functions in fungi, including cell division 11 , forming the primary septum of all septa, hyphal growth 12 , and virulence 13,14 . Chitin synthesis in C. glabrata is carried out by chitin synthases 15 . Deregulation of chitin biosynthesis is a potential mechanism of virulence and resistance to antifungal treatments.
Chitin has been reported to have anti-ulcer 16 , anti-tumor 17 , and anti-inflammatory 18 properties. Chitin is recognized by different receptors, triggering an immune response. Previous investigations have shown that NOD-2 and TLR-9 recognize chitin and act together to mediate an anti-inflammatory response via secretion of the cytokine, interleukin (IL)-10 19 .
In the present study, we investigated the impact of C. glabrata colonization on the gut microbiota diversity in a DSS-induced colitis model, and assessed how the C. glabrata cell wall is remodeled in order to persist in the gut environment. We also analyzed the effect of chitin deficiency on C. glabrata-host interactions in the DSS mouse model.

Results
Measurement of inflammatory parameters. Mice were administered an oral dose of C. glabrata WT and were exposed to DSS treatment for 2 weeks in order to induce acute colitis. Mice given C. glabrata WT only showed no signs of inflammation or mortality (Supplementary data Fig. 1 and Table 1). In contrast, mice given DSS or C. glabrata WT-DSS showed, from day 6, a decrease in body weight and mortality rates of 10% and 20%, respectively. The clinical score for inflammation was significantly higher in C. glabrata WT-DSS mice than in the DSS group, starting gradually from day 6, in which the first clinical symptoms of inflammation appeared, including diarrhea and bloody stools (Supplementary data Fig. 1B). The histologic score, which was based on the degree of inflammatory cell infiltration and the amount of tissue damage, was significantly higher in C. glabrata WT-DSS mice than in DSS-treated mice (Supplementary data Fig. 1C). Epithelial damage was observed along all parts of the colon mucosa, and leukocyte infiltrates, crypt abscesses and mucosal edema were more common in the colons from C. glabrata WT-DSS mice than in those from mice treated with DSS only (Supplementary data Fig. 1D).

Impact of C. glabrata colonization on the gut microbiota.
The number of C. glabrata CFUs and the changes in microbiota diversity were determined in freshly collected stool samples from tagged mice, using traditional culture methods based on agar plates (Fig. 1). A high number of C. glabrata CFUs was observed in both C. glabrata WT and C. glabrata WT-DSS groups on day 1 (Fig. 1A). In the absence of DSS, C. glabrata was eliminated from the mice from day 2. In contrast, the number of C. glabrata CFUs increased from day 2 in the C. glabrata WT-DSS group, and remained significantly higher than that in the C. glabrata WT group up to day 14 (P < 0.05).
To assess the extent of C. glabrata colonization in the GI tract, we assessed the number of yeasts adhering to the stomach, caecum, and colon (Fig. 1B). Significantly higher numbers of viable C. glabrata were detected in the stomach, caecum, and colon of C. glabrata WT-DSS mice than in mice colonized with C. glabrata WT (P < 0.001).
DSS-induced colitis promoted C. glabrata dissemination to the liver and kidneys (Fig. 1C). Regarding the microbiota diversity, cultures confirmed that the number of E. coli and E. faecalis colonies increased significantly from day 6 to day 14 in both the DSS and C. glabrata WT-DSS groups when compared to the C. glabrata WT or control groups, suggesting that irrespective of C. glabrata colonization, DSS-induced colitis promotes an increase in E. coli and E. faecalis populations in mice ( Fig. 2A,C). In contrast to Bacteroides vulgatus populations, the number of B. thetaiotaomicron and Bacteroides sp. TP5 decreased significantly in both DSS and C. glabrata WT-DSS mice (Fig. 2B,D and E). In terms of anaerobic bacteria, the number of Bifidobacterium spp., in particular Bifidobacterium animalis, and L. johnsonii colonies was significantly reduced in both DSS and C. glabrata WT-DSS mice, but the reduction in the L. johnsonii population was significantly more pronounced in the C. glabrata WT-DSS group, suggesting that overgrowth of C. glabrata impacts on the L. johnsonii population during the development of colitis ( Fig. 2F and G). In contrast, the Lactobacillus reuteri population revealed unpredictable changes and significant fluctuations over the course of DSS-induced colitis suggesting no association between C. glabrata-induced intestinal inflammation and L. reuteri levels (Fig. 2H).
Analysis of cytokines, signaling pathways, and receptor expression. To understand the mechanism by which changes in the microbiota modulate the inflammatory mediator responses during colitis and colonization with C. glabrata, pro-inflammatory cytokine expression (including IL-1β, and IL-6) was measured in the mice colons. Expression of IL-1β and IL-6 was significantly higher in the colons of C. glabrata WT-DSS mice than in DSS mice ( Fig. 3A-D). Conversely, expression of these cytokines was significantly lower in the colons of C. glabrata WT mice and control groups. In terms of PPARγ and Myd88 expression, there were no significant differences between DSS and C. glabrata WT-DSS groups (Fig. 3E,G). To determine the activation/expression of receptors in response to C. glabrata sensing and colitis, the expression levels of TLR-4, TLR-9, and MBL-C were examined. DSS-induced colitis significantly increased TLR-4, TLR-9, and MBL-C expression in colon tissues. TLR-9 and MBL-C expression increased significantly in response to both C. glabrata and colitis, when compared to mice treated with DSS only (Fig. 3H-J), although TLR-4 expression decreased.
Remodeling of the C. glabrata cell wall after passage through the GI tract. To assess whether C. glabrata undergoes cell wall remodeling during fungal colonization in the DSS-induced colitis model, we analyzed the cell wall fitness in terms of α-mannans (α-mans), β-mans, and chitin in the stools of tagged mice over a In the absence of DSS, C. glabrata was rapidly eliminated from the mouse gut; the fitness of the fungal cell wall was therefore only analyzed over 2 days. A significant increase in α-mans was observed, while the chitin level did not change significantly between day 0 and day 1 (Fig. 4).
In DSS-induced colitis, a significant increase in chitin and β-man levels was observed, while α-mans decreased significantly (Fig. 4C,D,F). Remodeling of the fungal cell wall was also assessed by confocal microscopy and was consistent with the flow cytometry analysis, indicating that both β-mans and chitin are involved in promoting persistence of C. glabrata in the gut. Impact of chitin deficiency on C. glabrata virulence in the mouse colitis model. To assess the impact of chitin deficiency on persistence of C. glabrata in the gut, C. glabrata strains ΔChs1 (Cg ΔChs1), and Δchs3 (Cg ΔChs3) were used in the DSS-induced colitis model and were compared with their parental strain, HTL (Fig. 5). Phenotypically, no growth defect was found in any strain when grown on YPD medium (data not shown). Flow cytometry revealed a significant decrease in chitin levels in C. glabrata ΔChs3 (80-90%), while chitin levels were increased in C. glabrata ΔChs1 when compared to C. glabrata HTL (Supplementary data Fig. 2). Additionally, β-man levels were increased in C. glabrata ΔChs1 when compared to C. glabrata ΔChs3 or C. glabrata HTL (Supplementary data Fig. 2). These data were consistent with confocal microscopy, which revealed that C. glabrata ΔChs3 stained weakly with WGA, while C. glabrata ΔChs1 and C. glabrata HTL stained strongly with WGA, and this staining was particularly strong in C. glabrata ΔChs1.
Mice were administered a unique inoculum of these C. glabrata strains orally. In the absence of DSS, neither mouse mortality, nor clinical inflammation was observed in groups receiving C. glabrata (Fig. 5).
In mice with DSS-induced colitis, mortality rates were recorded as 20% in DSS mice, as 20% in C. glabrata HTL-DSS, and as 62.5% in C. glabrata ΔChs1 mice. This suggests that C. glabrata ΔChs1 is more virulent than the other strains (Fig. 5A).
Significantly higher clinical and histologic scores for inflammation were observed in C. glabrata ΔChs1-DSS mice than in the C. glabrata ΔChs3-DSS and C. glabrata HTL-DSS groups (Fig. 5C,D). In addition, histologic sections of colons from C. glabrata ΔChs1-DSS mice revealed high levels of leukocyte infiltration, epithelial damage, and edema when compared to colons from C. glabrata ΔChs3-DSS or C. glabrata HTL-DSS mice. In terms of fungal colonization, all C. glabrata strains were eliminated within 2 days after challenge from mice without colitis. In mice with DSS-induced colitis, there was a significant increase in number of C. glabrata colonies detected in stools when the first clinical signs of inflammation appeared. Although C. glabrata ΔChs1 is highly virulent, the number of CFUs of this strain was lower than that of the C. glabrata HTL parental strain (Supplementary data Fig. 3).
To assess the impact of chitin deficiency on C. glabrata colonization of the gut, we measured the number of C. glabrata CFUs adhering to the stomach, caecum, and colons in the different groups of mice (Supplementary data Fig. 3B).
The number of C. glabrata ΔChs1 colonies was significantly higher in the stomachs of mice treated with DSS than in C. glabrata ΔChs3 and C. glabrata HTL-DSS treated mice. In terms of fungal dissemination from the gut to the organs, higher numbers of C. glabrata ΔChs1 colonies were observed while the C. glabrata ΔChs3 strain was unable to disseminate. A trend towards a large number of C. glabrata ΔChs1 colonies was observed in the kidneys of DSS-treated mice (Supplementary data Fig. 3C).
The impact of C. glabrata colonization on the gut microbiota was also investigated. The change in E. coli numbers was determined in freshly collected stool samples (Supplementary data Fig. 4). In contrast to mice not treated with DSS, where the population of E. coli remained relatively stable, the E. coli population increased significantly in the DSS, C. glabrata ΔChs1, and C. glabrata HTL groups, as inflammation progressed, when compared to C. glabrata ΔChs3 mice.   Effect of oral administration of chitin on intestinal inflammation, C. glabrata overgrowth and the gut microbiota. Before conducting our experiments in mice, we determined the effect of chitin on transepithelial electrical resistance and permeability of Caco-2 cells and whether chitin can reduce the adherence of C. glabrata to Caco-2 cells. We used different concentrations of chitin (1, 3, 5, and 10 mg). We found that chitin concentrations greater than 1 mg reduced C. glabrata adherence to Caco-2 cells (Supplementary data Fig. 6). Furthermore, the addition of chitin to Caco-2 cells before adding C. glabrata promoted intestinal barrier function as measured by a significant increase in transepithelial electrical resistance (data not shown). To determine the effect of chitin on modulation of intestinal inflammation and microbiota diversity including C. glabrata, chitin purified from C. glabrata was administered orally (3 mg/dose) to mice for 5 days after C. glabrata challenge. No mortality was observed in mice that received chitin during DSS-induced inflammation and colonization with C. glabrata. Administration of chitin also decreased the clinical and histologic scores when compared to untreated mice (Supplementary data Fig. 7). Mice treated with chitin showed decreased levels of C. glabrata in the stools as well as in the stomach and colon (Supplementary data Fig. 8). In terms of bacterial biodiversity, chitin treatment re-established the anaerobic bacteria including L. reuteri, L. johnsonii, Bifidobacterium and Bacteroides spp. and reduced aerobic bacteria such as E. coli and E. faecalis ( Fig. 6 and Supplementary data Fig. 9). In addition, TLR-8, dectin-1, NOD-2, and PPARg receptors were significantly activated as a result of chitin administration while the expression of TLR-4, TLR-5, TLR-7, and TLR-9 was decreased (Fig. 7).
When mRNA and protein expression of chitinase-3-like protein-1 was assessed after chitin treatment, a significant increase in chitinase-3-like protein-1 was observed, correlated with a decrease in C. glabrata CFUs in the gut indicating that both mRNA and protein expression of chitinase-3 increased after chitin treatment (Fig. 7).

Discussion
Dysbiosis is a change in the normal gut microbiome with a breakdown of host-microbial crosstalk 20 . This change in the gut microbiome promotes the overgrowth of opportunistic pathogens that contribute to intestinal mucosal inflammation 21,22 . Patients with CD are more frequently colonized with Candida species than healthy subjects 23 . Experimental studies show that colonization with Candida species exacerbates intestinal inflammation in the DSS-induced colitis model 3,4 . These clinical and experimental observations reveal the major role of opportunistic yeasts in modulation of the host immune-inflammatory response. In the present study, we assessed how the opportunistic yeast C. glabrata changes its cell wall composition in order to persist in the gut, and how overgrowth of this fungus together with intestinal inflammation affects the gut microbiome through modulation of both cytokine production and pathogen-recognition receptor (PRR) stimulation. The study also explored the effect of the fungal cell wall component chitin on modulation of the gut microbiome and its biological activities that confer a benefit to the host in terms of reducing inflammation. We also investigated the changes in C. glabrata virulence related to chitin gene deletions in the DSS-induced colitis model. In this DSS model, which is used to experimentally mimic IBD, mice received a single inoculum of C. glabrata by oral gavage and C. glabrata colonization increased gradually when DSS-induced intestinal inflammation was maintained in the mice. Erosions and ulceration of the mucosal surface of the gut promoted overgrowth of C. glabrata, which in turn exacerbated inflammation. In contrast, C. glabrata was rapidly eliminated from the gut in the absence of DSS-induced colitis, around 2 days after fungal challenge.
In the present study, we observed that regardless of C. glabrata colonization, DSS-induced colitis triggered changes in the gut microbiome. We focused in particular on cultivable bacteria belonging to the phyla Firmicutes, Bacteriodetes, Proteabacteria, or Actinobacteria by using selective bacterial media since these bacteria are known to be involved in CD (i.e. E. coli and E. faecalis) 24 . We observed a decrease in Firmicutes and Bacteroidetes while Proteobacteria and Enterobacteria increased in mice that developed colitis. These data are consistent with the clinical study, which showed a decrease in Firmicutes and Bacteroidetes, and an increase in Proteobacteria and fungal load, in particular C. glabrata, in CD patients 25 . Kim et al. showed that the dual association of E. coli and E. faecalis, both commensal organisms, rapidly induced severe pancolitis with dysplasia 26 . In addition, E. coli numbers increased during colitis and had a high pro-inflammatory potential to trigger inflammation via TLR-4 27,28 . E. faecalis is a facultative anaerobic bacterium that likely benefits from potentially increased oxygen availability in the inflamed intestine, in a manner similar to E. coli, which is strictly aerobic. Furthermore, E. coli can even exploit the environment in an inflamed gut to obtain a growth advantage when compared to anaerobic bacteria (Lactobacillus or Bifidobacterium). In response to epithelial cell death and tissue damage during intestinal inflammation, the dead cells could provide extra nutrients, such as ethanolamine, to support E. coli overgrowth 29,30 .
Aerobic culture of stomach and colon samples showed an increase in E. coli and E. faecalis populations in DSS-treated mice when compared to control mice. These data are consistent with clinical studies, which showed that aerobic cultures of biopsies obtained by colonoscopy from control colons were often sterile, whereas colons from patients with CD contained increased bacterial numbers in the sub-mucosa, a relatively well oxygenated site, more than half of which were E. coli 31 . In parallel, we observed that the increase in E. coli population in the gut was correlated with C. glabrata overgrowth. Centeno et al. observed that piliated E. coli strains can enhance Candida attachment to epithelial cells 32 . E. coli also exhibits a synergistic effect with Candida by inducing high mouse mortality during experimental microbial peritonitis 33 . In biofilm studies, Hoarau et al. showed that the interaction of E. coli with C. tropicalis combined with Serratia marcescens enhanced fungal filamentation and biofilm maturation 34 .
In terms of anaerobic bacteria, the population of Bifidobacterium spp., B. thetaiotaomicron and other Bacteroides spp. decreased in response to intestinal inflammation and to C. glabrata overgrowth, while B. vulgatus was not affected by either DSS-induced colitis or C. glabrata overgrowth. Interestingly, C. glabrata overgrowth significantly decreased the L. johnsonii population during the development of colitis. This observation is consistent with previous reports, which indicate that Lactobacillus growth can antagonize colonization by Candida 35 .
Receptors and signaling pathways were selected according to our previous studies or other studies showing the involvement of these receptors in fungal recognition, in particular TLRs, MBL, or in the stimulation of signaling pathways in response to C. glabrata sensing 7,36,37 . In this study, C. glabrata WT was found to increase MBL-C and TLR-9 expression. This finding supports the findings of Choteau et al. who showed that the activation of MBL-C by Candida sensing in intestinal epithelial cells promoted the rapid elimination of C. glabrata from the gut 37 . Conversely, the C. glabrata WT strain decreased TLR-4 expression in the mouse colons as colitis developed when compared to colons of mice treated with DSS only, suggesting that stimulation of TLR-4 is driven by overgrowth of E. coli and E. faecalis, while the increase in MBL-C expression is related to Candida sensing 35 . Pro-inflammatory cytokine mRNA and protein expression (IL-6 and IL-1β) increased in the colons of DSS-treated mice, and a higher level of expression of these cytokines was observed in response to Candida sensing. This is consistent with different reports, which showed that DSS-induced colitis alone or combined with fungal colonization promotes overexpression of pro-inflammatory mediators that amplify the inflammatory cascade through NF-κB and Myd88 expression.
Previous studies have shown how the local environment affects morphogenesis, virulence gene expression, and stress resistance, but very little is known about how the inflammatory gut environment impacts on the fungal cell wall composition. In the present study, we found that C. glabrata undergoes cell wall remodeling during fungal colonization in the DSS-induced colitis model. Modification of the composition of the C. glabrata cell wall is related to intestinal inflammation and not to bacteria biodiversity changes, since the C. glabrata cell wall modification appears in the days corresponding to the onset of inflammation and not while the bacterial population changes. In the absence of colitis development, the level of α-mans increased in the C. glabrata cell wall on days 1 and 2 after fungal challenge and then this yeast was eliminated rapidly from the gut. The increase in C. glabrata α-man level enables C. glabrata to behave like S. cerevisiae transiting through the mouse gut. In contrast, in DSS-induced colitis, C. glabrata appears to alter towards a pathogenic form close to that of C. albicans, resulting in an increase in chitin and β-man, and a decrease in α-man levels. These data reveal that the inflammatory gut environment impacts on the C. glabrata cell wall leading to adaptation of the fungal cell within the host, which allows C. glabrata to persist in the gut environment.
In a previous study, we showed that a deficiency in β-mans reduced C. glabrata adherence to intestinal epithelial cells, favoring fungal elimination from the mouse gut, indicating that β-mans contribute to C. glabrata virulence. However, the role of chitin in modulation of intestinal inflammation, the gut microbiome and fungal colonization has not yet been thoroughly investigated. In the present study, orally administered chitin purified from C. glabrata decreased intestinal inflammation and C. glabrata overgrowth. Several studies have shown that chitin enhances the immune response and increases the clearance of pathogenic bacteria in animal models; this supports our observations that chitin from C. glabrata has beneficial activities for C. glabrata.
To assess the impact of chitin deficiency on C. glabrata in the inflammatory gut environment, we selected the two genes that encode chitin synthase involved in chitin biosynthesis, Chs1 and Chs3 38 . Flow cytometry and confocal microscopy of the fungal cell wall revealed a significant reduction in chitin in C. glabrata Δchs3 while chitin levels in C. glabrata Δchs1 increased, indicating that in contrast to Chs1, chitin synthase Chs3 is a crucial enzyme for cell wall chitin synthesis (Supplementary data Fig. 8). This observation is consistent with other reports that show that chitin synthase Chs3 is involved in generating 80-90% of the total fungal chitin, while Chs1 is responsible for repairing the septa and the weakened cell walls of daughter cells after their separation from mother cells 39 . In the absence of DSS, mutant strains did not induce inflammation in the mice. In DSS-induced colitis, C. glabrata Δchs1, rich in chitin and β-glucan, is highly virulent in terms of colonization, fungal dissemination to the organs and inflammatory parameters. In contrast, a deficiency of chitin synthase Chs3 reduced the pathogenicity of C. glabrata in the inflammatory gut environment. Furthermore, C. glabrata Δ chs1, which has a high level of chitin, does not induce the expression of chitinase-3-like protein-1. Thus, this promotes C. glabrata overgrowth and persistence in the gut. In addition to the increase in chitin level in the C. glabrata Δchs1 cell wall, the β-man level, which is expressed in the outer fungal cell wall layer, was also increased. These data are consistent with other reports, which show that C. glabrata strains deficient in β-mans are less virulent than the WT strain suggesting the involvement of β-mans in the virulence and resistance of C. glabrata in the intestinal tract 5 .
Interestingly, the virulent C. glabrata Δchs1 strain reduced the expression of MBL-C, TLR-2, and TLR-9, while C. glabrata Δchs3 increased the expression of these receptors as colitis developed, indicating that the cell wall of C. glabrata Δchs3 is rich in α-mans, which are potential ligands for MBL-C and TLR4. In contrast, C. glabrata Δchs1, which has high levels of expression of β-man epitopes on its cell wall surface, is capable of reducing these intestinal epithelial cell receptors as an escape mechanism from the host defense.
Chitin acts as an anti-inflammatory agent and is involved in the process of wound healing 40 . Oral administration of chitin reduced all of the inflammatory parameters, and led to overgrowth of C. glabrata and reestablishment of the biodiversity of the gut microbiota. Chitin treatment increased chitinase-3-like protein-1, promoting chitin breakdown and the generation of small sized chitin particles that induce IL-10 production via PPARg, NOD-2, and TLR-8 sensing. These results are consistent with those of Wagner et al. who showed that chitin oligosaccharides have the potential to induce IL-10 secretion, via NOD-2 and TLR-9 signaling, promoting the attenuation of inflammation responses 19 . Surprisingly, the expression levels of TLR-4, TLR-5, TLR-6, TLR-7, TLR-9, and MBL-C were not upregulated by chitin treatment.
In conclusion, inflammation in the gut increased the aerobic bacteria population, in particular E. coli and E. faecalis, but decreased the population of anaerobic bacteria such as L. johnsonii, B. thetaiotaomicron, and Bifidobacterium spp. DSS-induced colitis led to cell wall remodeling through an increase in chitin production, which was involved in promoting C. glabrata overgrowth and persistence in the gut. C. glabrata colonization modulated the intestinal epithelial receptors, in particular MBL-C, TLR-4, and TLR-9, as well as expression of the signaling pathways, NF-κB and Myd88. In terms of fungal cell wall components, oral administration of chitin to mice reduced the overgrowth of aerobic bacteria and C. glabrata as well the production of inflammatory parameters through stimulation of intestinal receptors. Chitin deficiency affected the C. glabrata cell wall composition. This deficiency was compensated for by an increase in α-man levels in C. glabrata chs3, which induced fewer inflammatory parameters than the parental strain. Chitin treatment increased chitinase-3-like protein-1, enabling chitin digestion and the generation of small sized chitin particles that induced IL-10 production via PPARg, NOD-2, and TLR-8 sensing, promoting the attenuation of colitis and C. glabrata elimination. Finally, this study has increased our understanding of the nature of yeast molecular components that differentially affect inflammation and/or C. glabrata clearance.

Methods
Animals. Eight-to-10-week-old, female C57BL/6 mice were purchased from Charles River Laboratories, France. Mice were allocated to six experimental groups (Supplementary data, Table 1) and eight control groups, including assessment of the effect of DSS alone. Four complete experimental series were performed independently.
All experimental procedures were approved by the subcommittee for Research Animal Care, Regional Hospital Center, Lille, France (00550.05), and in accordance with institutional (86/609/CEE) and European guidelines for the care and use of laboratory animals.  38 . All C. glabrata strains were maintained at 4 °C in yeast peptone dextrose broth (YPD; 1% yeast extract, 2% peptone, 2% dextrose).
Extraction of chitin from C. glabrata. C. glabrata cell pellets were washed twice in phosphate-buffered saline (PBS). Chitin was extracted from C. glabrata yeast cells as described previously 41 . Briefly, the cell pellet was incubated twice in 20 ml of 10% KOH and autoclaved at 120 °C for 2 h. After washing several times with distilled water, the supernatant was removed and the pellet was oxidized with 50% hydrogen peroxide and 50% acetic acid, and then autoclaved at 120 °C for 2 h. After washing several times with distilled water, the chitin fraction was lyophilized. The nature of the chitin was confirmed by nuclear magnetic resonance (NMR) analysis. Intact chitin purified from C. glabrata was suspended in deuterated hexafluroisopropanol (Euriso-Top) at 70 °C until dissolved. All experiments were performed using a Bruker Avance 600 MHz (13.1 T) spectrometer with Bruker standard pulse programs. The chitin concentration was determined using a BiCinchoninic acid assay. The standard range for the N-acetyl-glucosamine (GlcNAc) control was 0.1-5 mg/mL.

Measurement of cell wall chitin content.
The chitin content was measured as described previously 42 .
C. glabrata HTL, C. glabrata ΔChs1, and C. glabrata ΔChs3 strains were grown in YPD broth. Yeast cells were suspended in 10 mL of PBS to a final concentration of 10 9 cells/mL and were disrupted with glass beads. The cells were then washed several times with 1 M NaCl and extracted in SDS-MerOH buffer (50 mM Tris, 2% sodium dodecyl sulfate (SDS), 0.3 M β-mercaptoethanol, 1 mM EDTA; pH 8.0) at 100 °C for 10 min, and then washed in distilled water. The chitin concentration was determined using a BiCinchoninic acid assay. The standard range for the N-acetyl-glucosamine (GlcNAc) control was 0.1-5 mg/mL.

Inoculum preparation and induction of colitis.
The mice were inoculated on day 1 by oral gavage with 200 µL PBS containing 5 × 10 7 live C. glabrata cells. From day 1 to day 14, mice were also administered 2% DSS (36−50 kDa; MP Biomedicals, LLC, Germany) in drinking water in order to induce intestinal inflammation. For chitin treatment, mice were administered with chitin purified from C. glabrata (3 mg per mouse) orally and daily for 5 days, starting on day 1. The presence of C. glabrata in the intestinal tract was monitored daily by measuring the number of colony-forming units (CFUs) in feces (approximately 0.1 g/sample) collected from each animal 4 . Fecal samples were suspended in 1 mL saline, homogenized in a glass tissue homogenizer, and samples were then cultured on Candi-Select medium (Bio-Rad Laboratories, Marnes la Coquette, France) 43 . The number of C. glabrata colonies was counted after incubation of the plates at 37 °C for 48 h. The results are expressed as CFU/µg of feces. In order to determine the degree of C. glabrata colonization of the gut, the mice were sacrificed and the GI tract was removed and separated into the stomach, ileum, and colon. These portions of the GI tract were cut longitudinally and the intestinal contents were removed. The tissue samples were then washed several times in PBS in order to minimize contamination from yeasts present in the lumen 44 . Serial dilutions of the homogenates were prepared and plated onto Candi-Select medium. The number of colonies was noted after 48 h incubation at 37 °C and expressed as C. glabrata CFU/mg of tissue.
For the isolation of bacteria, we performed serial dilutions of the gut contents (stomach, colon, and caecum) or fecal samples collected from the mice. The samples were cultured on non-selective bacterial media (AC agar) focusing on the most representative cultivable anaerobic and aerobic bacteria that can undergo changes during intestinal inflammation 45 .
For the isolation of aerobic bacteria, the fecal samples and tissues were plated onto MacConkey agar (Sigma-Aldrich), bile esculin azide agar (BEA; Sigma-Aldrich), and MRS agar (Sigma-Aldrich) plates. Serial dilutions of the samples were prepared. Bacteroides bile esculin (BBE; Sigma-Aldrich), Columbia agar (Sigma-Aldrich), and Bifidus selective medium agar (BSM; Sigma-Aldrich) were used for the isolation of anaerobic bacteria. These plates were incubated in anaerobic jars containing an anaerobic atmosphere generator pack (AnaeroGen TM 2.5 L; Sigma-Aldrich) at 35 °C. All aerobic and anaerobic media contained 60 mg/L fluconazole (Fresenius Kabi) to inhibit the growth of fungal cells.
All plates were incubated at 37 °C and examined after 24 h and 48 h. To identify any bacteria on the plates, the colonies were mixed with 1.5 μl of matrix solution (α-cyano-4-hydroxycinnamic acid; Bruker Daltonics) dissolved in 50% acetonitrile, 47.5% water, and 2.5% trifluoroacetic acid, and analyzed by MALDI-TOF MS (Microflex-Bruker Daltonics).

Analysis of C. glabrata cell wall remodeling after passage through the GI tract. Fresh fungal cells
obtained from fecal samples, and collected daily from each tagged animal, were diluted in PBS. 10 7 C. glabrata cells were prepared and washed several times with PBS. Washed yeast cells were incubated with PBS containing 2% fetal calf serum for 20 min, and were then incubated with fluorescein isothiocyanate-wheat germ agglutinin (FITC-WGA), rhodamine-labeled concanavalin A (Con A), or 5B2 antibody (FITC anti-rat IgM secondary antibody and the control isotype rat IgM). Analysis of the expression of glycan epitopes on the C. glabrata cell wall was performed by flow cytometry (Accuri ® Sampler ™ ). Mean fluorescence intensity (MFI) of each histogram was calculated as: labeled strain-unlabeled strain/mean fluorescence of labeled strain according to the number of days.
Assessment of clinical and histological scores. Body weight and mortality of the mice were recorded daily. The data were expressed as mean percent change from initial body weight. Clinical scores ranging from 0 to 12 and histologic scores ranging from 0 (no changes) to 6 (extensive cell infiltration and tissue damage) were calculated as described previously 4,46 . Real-time mRNA quantification of pro-inflammatory cytokines and innate immune receptors.
Total RNA was extracted from mouse colons using a Nucleospin RNA kit (Macherey-Nagel). RNA was quantified Scientific RePoRTS | (2018) 8:3316 | DOI:10.1038/s41598-018-21422-w by spectrophotometry (Nanodrop; Nyxor Biotech, France). Reverse transcription of mRNA was carried out from 1 µg total RNA using a high capacity cDNA RT kit (Applied Biosystems) in a final volume of 20 µL. cDNA was amplified by PCR in the one-step system (Applied Biosystems) using Fast SYBR green (Applied Biosystems). The intensity of SYBR green dye was assessed using one-step software. The reference gene, POLR2A 47 , was used to normalize the results.

Quantification of cytokines and chitinase 3-like 1 by ELISA. Representative pro-inflammatory (IL-
1β, IL-6) and anti-inflammatory (IL-10) cytokine profiles were selected in this study. Cytokine concentrations (IL-1β, IL-6, and IL-10) in the colons were measured using a commercial ELISA kit according to the manufacturer's instructions (eBioscience, San Diego, CA) whereas the detection of murine chitinase 3-like 1 levels was done using an ELISA kit from R and D systems. The data are expressed as pg/mL. Statistical analysis. All data are expressed as the mean ± standard deviation (SD) for each experimental group. Pairs of groups were compared using the Mann-Whitney U test. Differences were considered to be statistically significant when the P value was as follows: p < 0.05; p < 0.01; p < 0.001.
All statistical analyses were carried out using GraphPad Prism 4.0 and XLSTAT.