Pre-colonization with the fungus Candida glabrata exacerbates infection by the bacterial pathogen Clostridioides difficile in a murine model

ABSTRACT The contributions of commensal fungi to human health and disease are not well understood. Candida species such as C. albicans and C. glabrata are opportunistic pathogenic fungi and common colonizers of the human intestinal tract. They have been shown to affect the host immune system and interact with the gut microbiome and pathogenic microorganisms. Therefore, Candida species could be expected to play important ecological roles in the host gastrointestinal tract. Previously, our group demonstrated that pre-colonization of mice with C. albicans protected them against lethal C. difficile infection (CDI). Here, we show that mice pre-colonized with C. glabrata succumbed to CDI more rapidly than mice that were not pre-colonized suggesting an enhancement in C. difficile pathogenesis. Further, when C. difficile was added to pre-formed C. glabrata biofilms, an increase in matrix and overall biomass was observed. These effects were also shown with C. glabrata clinical isolates. Interestingly, the presence of C. difficile increased C. glabrata biofilm susceptibility to caspofungin, indicating potential effects on the fungal cell wall. Defining this intricate and intimate relationship will lead to an understanding of the role of Candida species in the context of CDI and novel aspects of Candida biology. IMPORTANCE Most microbiome studies have only considered the bacterial populations while ignoring other members of the microbiome such as fungi, other eukaryotic microorganisms, and viruses. Therefore, the role of fungi in human health and disease has been significantly understudied compared to their bacterial counterparts. This has generated a significant gap in knowledge that has negatively impacted disease diagnosis, understanding, and the development of therapeutics. With the development of novel technologies, we now have an understanding of mycobiome composition, but we do not understand the roles of fungi in the host. Here, we present findings showing that Candida glabrata, an opportunistic pathogenic yeast that colonizes the mammalian gastrointestinal tract, can impact the severity and outcome of a Clostridioides difficile infection (CDI) in a murine model. These findings bring attention to fungal colonizers during CDI, a bacterial infection of the gastrointestinal tract.

T he contributions of commensal fungi to human health and disease are not fully understood. Candida species such as C. albicans are opportunistic pathogenic fungi and common colonizers of the human intestinal tract (1-3). They have been shown to affect the host immune system (4)(5)(6)(7)(8), and interact with the gut microbiome and pathogenic microorganisms (9)(10)(11)(12)(13)(14). Therefore, Candida species could be expected to play important ecological roles in the host gastrointestinal tract. Our group's demonstration C57BL/6 mice were co-housed and provided cefoperazone in their drinking water. At the end of 10 days, mice were switched to water without cefoperazone and the day after this transfer, given PBS or C. glabrata BG2 via oral inoculation. Mice were administered clindamycin intraperitoneally and fed C. difficile UK1 spores (3, 5 × 10 5 spores per mouse) on the next day. The presence of C. glabrata exacerbated CDI leading to earlier disease onset, increased severity, and mortality (Fig. 1B). Mice pre-colonized with C. glabrata and challenged with C. difficile began succumbing 2 days earlier compared to the C. difficile only group (P = 0.0018; log rank test) (Fig. 1B). Importantly, mice pre-colonized only with C. glabrata did not exhibit any visible disease.
A separate group of C. glabrata only mice were followed for 28 days. These mice continued to be colonized with C. glabrata (10 7 cfu/gm in fecal pellets) and gained weight normally (Fig. S1). These findings show that C. glabrata can colonize mice for extended periods of time without apparent detrimental effects to the host.
Mice pre-colonized with C. glabrata contained higher cecal C. difficile CFU 1 day after C. difficile inoculation compared to C. difficile only mice (Fig. 1C). Moreover, all mice (n = 12) pre-colonized with C. glabrata contained detectable toxin in their ceca 1 day after C. difficile infection compared to only four mice (n=12) with detectable toxin from the C. difficile only group (Fig. 1D). These findings show that C. glabrata colonization leads to CDI exacerbation, which could be due in part to earlier toxin production. Importantly, no correlation was observed between CFU, and amount of toxin detected at day 1 or 2 after infection, similar to previous reports (17,53). Further, no significant differences in cecum tissue histology were detected between the groups at either of the two timepoints (Fig.  S2). Thus, in contrast to C. albicans (15), C. glabrata pre-colonization produces a distinct effect, negatively impacting CDI outcome. The exact mechanisms of this exacerbation are most likely multifactorial. in (A). The C57BL/6 mice were co-housed in a large cage and received antibiotics as described in Materials and Methods (cefoperazone 10 days). Mice were then transferred to standard size cages, some mice were fed C. glabrata, and all mice were switched to water without antibiotics for 2 days. Mice received clindamycin intraperitoneally followed by C. difficile spores (3, 5 × 10 5 spores of strain UK1, a NAP1/027/BI human epidemic strain) by oral inoculation on the next day. (B) Survival of the mice was monitored for 6 days post-inoculation with spores (C. glabrata, n = 12; C. difficile, n = 12; C. glabrata and C. difficile, n = 12 log rank test P = 0.0018). (C) C. difficile bacteria were enumerated by plating the homogenized cecum contents collected from mice sacrificed on day 1 or 2 post-inoculation. Homogenates were plated on a TCCFA medium. Bar indicates the geometric mean. Mann-Whitney P-values are displayed above each set. *, P = 0.0112. (D) C. difficile toxin activity in the cecum contents from the mice sacrificed on day 1 or 2 post-inoculation was measured using a cell-rounding assay. The inverse of the greatest dilution that yielded 100% cell rounding is plotted. Mann-Whitney P-values are displayed above each set. **, P = 0.007 combined results of three experimental trials.

Research Article mSphere
Microbiota composition of the host gastrointestinal tract plays a significant role in preventing CDI by a variety of mechanisms (e.g., limiting access to resources and secretion of inhibitory metabolic products). Indeed, CDI is highly dependent on the host microbiota composition and several bacterial genera have been implicated in antago nistic roles toward C. difficile (41,42). Candida species have been shown to influence microbiota reassembly and composition after antibiotic treatment and immunosuppres sion (22,(54)(55)(56)(57). Therefore, we characterized the cecal microbiota composition in our murine model to understand its role in disease exacerbation. Mice were divided into uninfected, C. glabrata only, C. difficile only, and C. glabrata + C. difficile cohorts as described above (Fig. 1A). Two days post-inoculation with C. difficile spores, mice were sacrificed, and the distal cecum tip was harvested, placed in a collection tube and frozen immediately. Samples were processed as described in Materials and Methods. The composition of the cecal bacterial community was characterized by sequencing the V4 region of the 16S rRNA gene and analyzed using QIIME 2 (2018.8) (58) (Fig. 2 and 3).
Alpha diversity analyses highlighted a significant increase in diversity of the population between the C. difficile only and C. glabrata + C. difficile groups ( Fig. 2A and B), although no significant differences were observed using the Chao1 and Faith's alpha diversity analyses (Fig. S3). The weighted UniFrac distance between bacterial communi ties is represented in a principal coordinate analysis (PCoA) (Fig. 3A). The PERMANOVA analysis of these results identified significant differences between C. difficile and C. glabrata + C. difficile (P = 0.016).
To identify possible differences in genera between the groups, the relative abundan ces of bacterial genera were compared (Fig. 3B). All bacterial genera with a median fraction greater than zero in at least one of the two groups were included in this analysis (Fig. 3B). Due to antibiotic treatment in preparation for the C. difficile infection, mice were colonized with relatively few genera. We observed differences in the relative abundance of certain groups. An increase in relative abundance of members of the Peptostreptococcaceae family (the family that includes C. difficile) and a decrease in the genus Enterococcus (Fig. 3B) were observed. Importantly, these results showed compara ble trends to findings previously reported in patients (59). Lastly, similar changes were observed in phyla (Fig. S4).
Total levels of bacteria were measured using qPCR and universal 16S rRNA primers as previously described (60). Relative abundance was multiplied by the total level of bacteria per milligram of cecum sample (Fig. S5) to determine absolute abundance in arbitrary units per mg of cecum contents. Absolute abundances of genera in mice with C. difficile and either with or without C. glabrata were very similar, but a statistically Research Article mSphere significant increase in the Peptostreptococcaceae family (P = 0.0057) was observed in mice with C. difficile and C. glabrata (Fig. S5). These findings suggest that bacterial microbiota composition is not the sole factor impacting disease progression and outcome in this model.

C. glabrata and C. difficile form robust polymicrobial biofilms
Both Candida species and C. difficile have been shown to interact with members of the microbiome (9,41,61,62). Therefore, an additional factor impacting disease exacerba tion could be polymicrobial interactions between C. difficile and C. glabrata. We detected an interaction between C. difficile and C. glabrata using a co-culturing system focused on biofilms.
A highly robust and replicable fungal biofilm model was adapted to study the effects of C. difficile on pre-formed fungal biofilms (Fig. 4). Pre-formed fungal biofilms were used because presumably, C. glabrata would be present and established in the human gastrointestinal tract before C. difficile was present. C. glabrata biofilms were grown in tissue culture media (DMEM, 10% FBS, and 1% NEAA) at 37°C under anaerobic conditions for 24 h before the addition of C. difficile as described in Materials and Methods. After the addition of C. difficile, biofilms were further incubated for 24 h. The addition of C. difficile to a C. glabrata biofilm led to a significant enhancement in overall biofilm biomass and matrix compared to either C. difficile or C. glabrata mono-microbial biofilms ( Fig. 4A and B). Importantly, biomass from the co-culture was greater than the biomass predicted due to an additive effect. This observation is significant as both organisms formed weak biofilms independently under the same conditions at 24 h (63). Furthermore, no significant change in the metabolic activity of C. glabrata was observed when C. difficile was present (Fig. 4C). Metabolic activity was measured using the XTT assay as previously described (64,65). To distinguish between the metabolic activity of both organisms, plates were taken out of the anaerobic chamber and washed multiple times with aerobic PBS to reduce C. difficile viability. C. difficile mono-microbial biofilms treated the same way did not reduce XTT (not shown). Additionally, plating of C. glabrata cells from the biofilms on YPD showed that there was no significant difference in CFU between C. glabrata only and C. glabrata with C. difficile ( Fig. 4D; Fig. S6). These results suggest that the presence of C. difficile leads to an enhancement in biofilm formation independent of C. glabrata cell number. Moreover, these results show that interaction with C. difficile was not cytotoxic to C. glabrata. Similar experiments were conducted to assess C. difficile viability except BHIC-TA was used as the plating medium and the plates were incubated anaerobically. There were no significant differences in CFU from C. difficile only biofilms or polymicrobial biofilms and as with C. glabrata, the enhancement in biofilm biomass was independent of C. difficile CFU (Fig. 4E). A similar phenotype of biofilm enhancement was observed when introducing C. difficile to biofilms of two distinct oral clinical isolates of C. glabrata ( Fig. S7) (66). Lastly, C. difficile toxin production was assessed in mono or polymicrobial biofilms (Fig. S8). Toxin titers were similar between biofilms, and thus, in vitro, the presence of C. glabrata did not enhance toxin production. These findings show that these two organisms form a highly robust polymicrobial biofilm in vitro without impacting each other's viability.

C. difficile sensitizes C. glabrata to the cell wall targeting agent caspofungin and impacts cell wall related processes
Polymicrobial biofilms are highly resistant to antimicrobial treatments (67). We sought to characterize the antifungal sensitivity profile of C. glabrata in polymicrobial biofilms with C. difficile using the model described above. The XTT assay was used as a readout for C. glabrata metabolic activity as described in Materials and Methods. The presence of C. difficile significantly sensitized C. glabrata to the cell wall targeting antifungal caspofungin at most concentrations tested (Fig. 5A). This increase in susceptibility was not observed with other antifungals tested (Fig. S9) suggesting that C. difficile impacts C. glabrata cell wall biology.
Additionally, a dose response assay was conducted to measure susceptibility to the antibiotic vancomycin. The PrestoBlue reagent was utilized to distinguish C. difficile metabolic activity from that of C. glabrata. The presence of C. glabrata increased C. difficile metabolic activity (no vancomycin, P = 0.0004) suggesting an enhancement in biofilm formation (Fig. S10). Moreover, the presence of C. glabrata appeared to decrease C. difficile's susceptibility to vancomycin (Fig. S10).
To better understand the impact of C. difficile on C. glabrata cell wall biology, transcription of genes whose products are involved in a variety of cell wall and biofilm related processes was measured (Fig. 5B). We focused on ERG11, which enco des Lanosterol 14-alpha-demethylase, a key enzyme in ergosterol biosynthesis (68), KNH1 encoding a product important in β1,6-glucan synthesis (69), MNN2 encoding an α-1,2-mannosyltransferase important for the glycosylation of cell wall components (70), and GAS2 encoding a glycosylphosphatidylinositol (GPI)-anchored cell surface protein that has been shown to be a regulator of β1,3-gulcan and important in drug treated biofilms of C. glabrata (71)(72)(73). Further, TEC1 encodes a product important for biofilm formation and adaptation to low pH stress in C. glabrata (74) and UGP1 encodes a Putative UDP-glucose pyrophosphorylase involved in the β1,6-gulcan biosynthetic process (75,76). We also measured the expression of EPA6, a gene whose product is one of the key adhesins in C. glabrata and is important for biofilm formation (77)(78)(79). While ERG11 expression was upregulated (8-fold), the rest of these cell wall and biofilm related genes were significantly downregulated in the presence of C. difficile. These findings support the model that C. difficile impacts fungal cell wall biology. These interactions could have significant implications for CDI progression in the gastrointestinal tract.

DISCUSSION
While the contributions of commensal fungi during bacterial infections are not well understood, fungi have been implicated in a variety of polymicrobial interactions that could significantly impact host health (9, 55, 80-83) including those involving CDI (15,24,25,28,29,60,(84)(85)(86). More recently, Santus and co-workers showed that Salmo nella enterica can utilize commensal and dietary fungal siderophores which promote Salmonella colonization (87). This study highlights the complexity of trans kingdom polymicrobial interactions and underscores the critical need to characterize the role of the mycobiome in host health and disease.
Here, we describe the use of a murine model of CDI coupled with a fungal coloni zation model to characterize the impact of the opportunistic pathogenic fungus, C. glabrata during CDI. We show that pre-colonization of mice with C. glabrata exacerbates CDI leading to more rapid mortality. Moreover, the co-colonized mice harbored higher levels of C. difficile (detected by plating for CFU and sequencing) and more detectable toxin compared to mice that only carried C. difficile at an early time point. These two factors probably contribute to disease exacerbation, but exacerbation is likely multifac torial as toxin production does not correlate with disease outcome in CDI (17,53). Moreover, in our model, CFU did not correlate with toxin production at either day 1 or 2 (not shown), further highlighting the complexity of the system. Previously, Jawhara and co-workers utilized a dextran sulfate sodium (DSS)-induced colitis murine model to characterize the contribution of C. glabrata cell wall components during gastrointestinal colonization (88). In this study, C. glabrata was able to colonize the murine gastrointes tinal tract without causing visible disease, which was also seen in our studies. When DSS was fed to mice in their drinking water, 25% of the C. glabrata colonized animals died, suggesting that C. glabrata can exploit a colitis environment. This study attributed (A) Biofilms were grown anaerobically and exposed to the antifungal caspofungin anaerobically in a dose response assay and the impact on the biofilm was quantified using the metabolic XTT assay. (B) RNA was isolated from the polymicrobial biofilms and expression of C. glabrata genes in the presence of C. difficile relative to C. glabrata alone was measured using qRT PCR (3 experimental trials).
Research Article mSphere disease to enhanced colonization and the ability of C. glabrata to cause damage in colitis tissues and did not assess dissemination of C. glabrata to other tissues. Toxins produced by C. difficile induce colitis in the gastrointestinal tract (89-94); therefore, it is possible that C. glabrata exploits this environment during CDI. While we did not observe an increase in inflammation in the C. glabrata + C. difficile compared to the C. difficile only group (Fig. 3), β-glucans and chitin in the fungal cell wall have been shown to have antiinflammatory effects (54,95) potentially obscuring any additional inflammatory effects in the C. glabrata + C. difficile groups. C. difficile infection is highly dependent on the status of the host gastrointestinal microbiome. While we did not identify large changes in overall bacterial microbiota composition in co-colonized mice compared to the other groups, we did identify changes in relative abundance of some taxa including a significantly higher relative abundance of the Peptostreptococcaceae family. Moreover, the trends observed in our murine models were similar to those reported in patients previously (59). Interestingly, in the co-colonized mice, we identified a lower relative abundance of the genus Entero coccus, a group of organisms that is often enriched in C. difficile infections (37,96). Importantly, Enteroccoci have been shown to be antagonistic to C. difficile growth (97), or in more recent studies, enhance C. difficile virulence via a nutrient restriction and cross-feeding mechanism (98). The presence of C. glabrata may enhance C. difficile fitness, allowing it to outcompete other members of the microbiome. The findings from our murine studies suggest that the presence of C. glabrata fuels CDI by enhancing early C. difficile growth and toxin production.
Utilizing an in vitro co-culturing system, we further show that C. difficile and C. glabrata form robust polymicrobial biofilms compared to their monomicrobial coun terparts. Interestingly, while they do not affect each other's viability, the presence of C. difficile impacts the expression of C. glabrata genes whose products are involved in cell wall biogenesis and maintenance and sensitizes C. glabrata to caspofungin, a cell wall targeting antifungal. These findings suggest that C. difficile can directly or indirectly impact fungal cell wall biology. We propose that these findings translate to the gastrointestinal tract via three potential mechanisms. First, Bacteriodetes have been shown to utilize yeast mannans, a mechanism that fuels growth in the gastrointes tinal tract (99). C. difficile could be utilizing a similar strategy. Second, Candida species colonizing the gastrointestinal tract mask glucans present in their cell walls to evade immune detection (4, 100-102). It is possible that C. difficile leads to dysregulation of the fungal cell wall modifying the glucan masking mechanism and eliciting an immune response from the host. These effects may lead to enhanced mortality in our model. Third, C. difficile has been shown to form biofilms with other members of the micro biome, which could enhance its survival and proliferation in the host gastrointestinal tract (98, 103). Polymicrobial biofilms with members of the mycobiome such as C. glabrata could play a significant role in CDI.
Currently, knowledge of mucosal biofilms in the gastrointestinal tract is limited, but reports suggest that they contribute to colonization resistance, community stability and resilience, food digestion, and drug chemical modifications (104). The role of biofilms in CDI is currently not known. C. difficile biofilms were first reported clogging a biliary stent (105) and have begun to receive interest in recent years (103,(106)(107)(108)(109)(110)(111)(112)(113)(114)(115)(116). Importantly, biofilms have been suggested to serve as reservoirs that could play a role in recurrence of CDI (115), but this has yet to be demonstrated.
It is currently unknown whether this ability to form a polymicrobial biofilm plays a significant role in the mammalian host. Interestingly, biofilms formed by members of the host microbiome, including the fungus C. parapsilosis, have recently been shown to harbor C. difficile in a gut chemostat model (103). This ability to form polymicrobial biofilms with members of the host microbiome, including fungi, could serve to harbor C. difficile and serve as a reservoir for recurrent infections.
Our in vitro model used biofilms to study interactions between C. difficile and C. glabrata and yielded highly interesting findings suggesting a mechanism involving the fungal cell wall. Interestingly, previous studies have reported an increase in virulence of cells involved in a biofilm (117). While fungi and bacteria in the gut may not form true biofilms, it is possible that the cells are in close proximity to each other, potentially allowing them to interact in ways that resemble the interactions detected in biofilms. These possibilities will require further investigation in the future.
Utilizing the murine model presented here will allow the characterization of the role of individual members of the mycobiome and will lead to the development of novel therapeutic interventions for CDI.

Strains and growth conditions
C. glabrata strain BG2 (118, 119), a clinical isolate from a vaginitis case that did not respond to fluconazole, was used for all studies (except for Fig. S7). Clinical isolates AE2 and D1 were kindly provided by Dr. Célia F. Rodrigues (52) and incubated at 37°C for 2 days in an anaerobic chamber. Samples were heated at 60°C for 10 min, followed by plating on pre-reduced TCCFA plates for spore enumeration.

Murine model of gastrointestinal colonization and CDI
All the experiments using animals were done in compliance with the NIH Guide for the Care and Use of Laboratory Animals and Tufts University IACUC guidelines. Animal experimentation was approved by the Tufts Institutional Animal Care & Use Committee, 10/08/2018, protocol number B2018-84. The murine model was recently described in detail (52). Briefly, 5-week-old female C57BL/6 mice (Jackson Laboratory, Bar Harbor, Maine, USA) were co-housed in a large cage (24″ × 17″) for 20 days total. After 10 days, mice were treated with the antibiotic cefoperazone (Sigma-Aldrich, St. Louis, Missouri, USA cat. C4292, 0.5 gm/L) in drinking water for the remaining 10 days of co-housing. At the end of the cefoperazone treatment, all mice were tested and shown to be negative for cultivable fungi on YPD-SA agar medium [YPD agar plus 100 µg/mL streptomycin (Sigma-Aldrich, St. Louis, Missouri, USA cat. S6501) and 50 µg/mL ampicillin (Sigma-Aldrich, St. Louis, Missouri, USA cat. A9518)] incubated for 2 days at 37°C. After 20 days of co-housing, all mice were transferred from the large cage to standard sized cages, housing four mice per cage and switched to water without cefoperazone. Mice were then given C. glabrata orally (5 × 10 8 cells in 25 µL) or PBS. For C. glabrata colonization experiments (n = 8), mice were allowed to rest for 28 days. C. glabrata colonization was measured as described previously (52). Briefly, 3-4 fecal pellets were collected from each animal into a pre-weighed tube containing PBS. The tube was weighed after collection. Pellets were then homogenized using a microtube vortex, serially diluted, and plated for CFU on YPD-SA. For the C. difficile infection experiments (n = 12), mice were treated as above with cefoperazone and orally inoculated with C. glabrata or PBS, then allowed to rest for 2 days. All the mice used in these experiments were shown to be negative for C. difficile colonization prior to inoculation with spores by collecting fecal pellets and plating on pre-reduced TCCFA. After the 2 days of rest, they were administered clindamycin intraperitoneally (Sigma-Aldrich, St. Louis, Missouri, USA cat. C5269) (10 mg/ kg). The following day, mice were orally inoculated with C. difficile spores (3, 5 × 10 5 spores per mouse). Mice were then monitored for survival or sacrificed at 1 or 2 days after C. difficile inoculation to assess disease progression. The mice were weighed daily at the same time of day and sacrificed when moribund. Mice were considered moribund if they exhibited severe signs of illness (extreme inactivity, hunched posture, ruffled fur). If their weight loss exceeded 20%, they were also sacrificed. To ascertain if C. glabrata was able to disseminate during C. difficile infection, murine liver, kidneys, and tongues were collected. The tissues were macerated, diluted, and plated on YPD-SA for C. glabrata and BHI for aerobic bacteria. The survival data show the combined results of the mice from multiple experimental trials. The relative weights were compared using the t test. Survival was compared using the log rank test.

C. difficile toxin titer assay
African green monkey kidney epithelial cells (Vero cells) (ATCC CCL-81) were seeded in 96-well microtiter plates at a concentration of 4.3 × 10 3 cells per well in DMEM (Corning Cell Gro, Corning, New York, USA cat. MT10-013CV) with 10% heat-inactivated fetal bovine serum (Atlanta Biologicals S11150) and 1% MEM non-essential amino acids (ThermoFisher Scientific, Waltham, MA, USA cat. 11140076) and allowed to adhere for 24 h at 37°C in 5% CO 2 (52,60). Cecum contents from mice were weighed and diluted 1:10 with DMEM with 10% FBS and 1% MEM non-essential amino acids. Serial threefold dilutions in DMEM with 10% FBS and 1% MEM non-essential amino acids, were added to adhered Vero cells and incubated for 24 h at 37°C in 5% CO 2 . Cells were scored visually at 10× magnification for cell rounding at 24 h. The toxin titer is defined as the inverse of the greatest dilution that resulted in 100% cell rounding.

Histology
One or 2 days post-inoculation with C. difficile, cecum tissue was collected from mice and fixed in buffered formalin and processed for staining with Hematoxylin and Eosin (H&E) by the Tufts Comparative Pathology Core facility. Scoring was conducted at 10× magnification on 30-40 fields for each tissue of view by a blinded investigator. The scores were calculated by dividing the number of fields showing infiltration/inflammation by the total number of fields examined for that tissue. Each point on the graph represents an individual mouse. A non-parametric t-test was used to compare the groups.

Microbiota analyses
Two days post C. difficile inoculation, mice were sacrificed, and the distal tip of the cecum was immediately frozen on dry ice. Microbial DNA was extracted using the QIAaMP PowerFecalPro DNA Kit (Qiagen, Hilden, Germany cat. 12830-50) according to the manufacturer's protocols. The recommended additional steps to ensure optimal yield of DNA from Gram-positive bacterial was also followed. Libraries were prepared and sequenced as previously described (122). Briefly, the V4 region of the 16S rRNA was amplified via PCR with 515F and 806R primers including adapters for Illumina sequenc ing and 12-mer Golay barcodes for multiplexing. Two hundred fifty base pair paired-end sequencing was performed using an Illumina MiSeq following manufacturer's protocols. CASAVA 1.8 was used to perform base calling. The resulting fastq files were used for analysis with QIIME2 (2018.8) (58). Raw sequences were demultiplexed and filtered with q2-deumix plugin followed by denoising with DADA2 (q2-dada2) (123). Amplicons were aligned with mafft (q2-phylogeny) (124). The de novo phylogenetic tree from fasttree2 (q2-phylogeny) (125) was used to construct phylogeny trees. The q2-diversity was used to estimate Alpha (Shannon, Chao1, Simpson, and Faith) and Beta-diversity [weighted UniFrac (126)] metrics. To summarize the weighted UniFrac distance matrix, we used a principal coordinate analysis (PCoA). PERMANOVA analysis was performed using QIIME2. The operational taxonomic units (OTUs) were determined by aligning reads to the Greengenes Database (version 13_8) at 99% identity (127). Feature tables describing the relative abundance of bacterial taxa were used for analysis of diversity within each sample. The total levels of bacteria per cecal tip sample were measured by qPCR using eubacterial primers as previously described (15,60). The absolute abundance of bacterial genera was calculated by multiplying the fraction of total reads for a genus by the total level of bacteria per mg of cecum sample (in arbitrary units).

Biofilm assays
We used a modified version of the 96-well flat bottom microtiter plate model of Candida biofilm formation previously described (65,128). Briefly, C. glabrata was grown overnight in YPD broth at 30°C. Cells were then centrifuged, washed twice in PBS, moved into an anaerobic chamber, and diluted to 1 × 10 6 cells/mL in pre-reduced DMEM (Corning Cell Gro, Corning, NY, USA cat. MT10-013CV) with 10% heat-inactivated fetal bovine serum (Atlanta Biologicals S11150) and 1% MEM non-essential amino acids (ThermoFisher Scientific, Waltham, MA, USA cat. 11140076). The cell suspension was used to seed wells of a 96-well flat bottom microtiter with 100 µL. Appropriate positive (absence of C. difficile to measure C. glabrata biofilm formation only) and negative (no cells, to monitor contamination and to be able to calculate percent inhibition) controls were included in each experiment. Plates were incubated anaerobically for 24 h at 37°C to allow for biofilm formation. After 24 h, the media was removed and replaced with fresh media containing C. difficile at a concentration of 1 × 10 5 cells/mL and incubated for a further 24 h. Each experiment was conducted three separate times (biological replicates), with multiple technical replicates per plate. The 2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazo lium-5-carboxanilide (XTT) assay was specifically used to measure C. glabrata viability, and in turn, biofilm formation, in polymicrobial biofilms. After the incubation period following the addition of C. difficile or fresh media without C. difficile, plates were taken out of the anaerobic chamber, washed twice with 200 µL/well of aerobic PBS to reduce C. difficile viability. Biofilm inhibition was measured using the XTT colorimetric assay as previously described (65,128,129).
The PrestoBlue cell viability reagent (Invitrogen, Carlsbad, CA, USA cat. A13261) was used to quantify C. difficile viability, and in turn, biofilm formation, in polymicro bial biofilms as previously described (130). Briefly, after incubation with or without C. glabrata, the supernatant was removed, and samples were washed twice with 200 µL of prereduced PBS. The PBS was removed and 200 µL of 1:10 vol/vol PrestoBlue reagent in DMEM supplemented with 10% FBS and 1% NEAA were added. Plates were then sealed with parafilm, wrapped in foil, and taken out of the anaerobic chamber to incubate for 1 h in a 5% CO 2 incubator at 37°C. Finally, 100 µL from each well was transferred into a new 96-well plate for fluorescent readings in a plate reader at 560/25 nm excitation and 590/35 emission.
To measure overall biofilm biomass, biofilms were stained with crystal violet as previously described (131). After incubation, plates were washed twice with PBS and 100 µL of methanol was added to each well for 20 min to fix the biofilms. Methanol was removed and plates allowed to dry. Biofilms were then stained for 10 min with 150 µL of 3% (wt/vol) crystal violet. Crystal violet was removed, and plates were allowed to dry. This was followed by three washes with 200 µL of distilled water. Plates were allowed to dry. Each well received 100 µL of 33% glacial acetic acid and was left to de-stain for 5 min while shaking slowly. A total of 85 µL was transferred to a new microtiter plate for OD 550 measurement. For microscopy, stained samples were directly observed on the 96-well plate using a 20× objective in an inverted system microscope equipped for photography.

Antifungal and antibacterial susceptibility assays
For antifungal dose response assays, polymicrobial biofilms were formed as described above. After the 24 h incubation with C. difficile, media was removed and replaced with media containing twofold dilutions either 5-Flucytosine (Sigma-Aldrich, St. Louis, MO, USA cat. 1272000-200MG), amphotericin B (ThermoFisher Scientific, Waltham, MA, USA cat. MT30003CF), caspofungin (Merck & Co., Inc., Whitehouse Station, NJ), or vancomycin (ThermoFisher Scientific, Waltham, MA, USA cat. BP29581) starting at 200 µg/mL, 1 µg/mL, 16 µg/mL, and 64 µg/mL, respectively. Caspofungin was obtained as a powder and was stored at 4°C; a stock solution was prepared in PBS at 2 mg/mL the same day before its addition to well plates. After anaerobic incubation, the plates were washed twice with 200 µL/well of PBS, and biofilm viability was measured using the XTT and Presto Blue colorimetric assays described above.

Real-time quantitative PCR analysis of C. glabrata gene expression in polymicrobial biofilms
C. glabrata RNA was extracted using the Ambion Purelink mini kit (Invitrogen, Carlsbad, CA, USA cat. 12183555). Briefly, media was removed, and biofilms were washed twice in cold PBS. Biofilms were then resuspended in 800 µL of lysis buffer (from kit) contain ing 1% BME. The resuspended biofilm was then transferred to 2 mL gasketed tubes containing 1.3 mL of zirconium/glass beads. The tubes containing the zirconium/glass beads were stored at −20°C for 30 min before use. Bead beating was conducted thrice in a Biospec Bead Beater 24 for 1 min with a 5 min ice incubation step in between. This was followed by purification with the Ambion Purelink Mini kit. cDNA was synthesized using SuperScript III (Invitrogen, Carlsbad, CA, USA cat. 18080093) as per the manufacturer instructions. qPCR reactions were performed using SYBR Green Master Mix (Applied Biosystems) and a LightCycler 480 II (Roche) instrument. Gene specific primers were previously described (73,78).

Statistical analysis
Statistical analyses were performed using GraphPad Prism (San Diego, CA, USA) and Excel.

ACKNOWLEDGMENTS
The authors thank Drs. Linc Sonenshein and Aimee Shen for helpful discussion and critical expertise. We also thank Anne Kane for assistance with the analysis of microbiota composition. Lastly, we would like to thank Drs. David Stewart and Regina Lamedella for helpful discussions.

DATA AVAILABILITY
The raw sequencing data were submitted in FASTQ format to the NCBI Sequence Read Archive. The BioProject ID is PRJNA918677 and it can be found on https:// www.ncbi.nlm.nih.gov/sra.