Rapid proliferation due to better metabolic adaptation results in full virulence of a filament-deficient Candida albicans strain

The ability of the fungal pathogen Candida albicans to undergo a yeast-to-hypha transition is believed to be a key virulence factor, as filaments mediate tissue damage. Here, we show that virulence is not necessarily reduced in filament-deficient strains, and the results depend on the infection model used. We generate a filament-deficient strain by deletion or repression of EED1 (known to be required for maintenance of hyphal growth). Consistent with previous studies, the strain is attenuated in damaging epithelial cells and macrophages in vitro and in a mouse model of intraperitoneal infection. However, in a mouse model of systemic infection, the strain is as virulent as the wild type when mice are challenged with intermediate infectious doses, and even more virulent when using low infectious doses. Retained virulence is associated with rapid yeast proliferation, likely the result of metabolic adaptation and improved fitness, leading to high organ fungal loads. Analyses of cytokine responses in vitro and in vivo, as well as systemic infections in immunosuppressed mice, suggest that differences in immunopathology contribute to some extent to retained virulence of the filament-deficient mutant. Our findings challenge the long-standing hypothesis that hyphae are essential for pathogenesis of systemic candidiasis by C. albicans.

T he polymorphic yeast Candida albicans colonizes up to 70% of the human population as a commensal on mucosal surfaces 1,2 . As an opportunistic fungal pathogen, however, C. albicans is able to cause superficial as well as life-threatening systemic infections promoted by disturbances in the microbiota and impaired host defenses 3,4 . Despite antifungal therapy, disseminated candidiasis is associated with high mortality rates of up to 50% 5,6 .
The reversible transition of spherical budding yeast to pseudohyphal or hyphal filaments is promoted in vivo by body temperature, serum, physiological pH and elevated CO 2 concentration 7,8 . Nevertheless, both morphologies are present in tissues during systemic infection 9 . While filaments facilitate tissue invasion, damage and escape from host cells 10,11 , yeast cells are believed to be important for mucosal colonization, dissemination through the bloodstream, adherence to endothelial cells and biofilm formation 12 . Furthermore, yeast and hypha are recognized differentially by immune cells 13 . Saville et al. dissected the relative contribution of C. albicans filamentation to virulence during systemic infection by using a regulable expression system, placing one copy of the negative regulator of filamentation NRG1 under the control of a tetracycline-regulable promotor 14 . Mice challenged intravenously with the tet-NRG1 strain succumbed to infection when hyphal growth was permitted but survived when fungal cells were enforced to grow in the yeast form. Similarly, other C. albicans yeast-locked mutants, such as the cph1Δ/Δ efg1Δ/Δ double and hgc1Δ/Δ mutant have been shown to be avirulent or are strongly attenuated in virulence 15,16 . On the other hand, mutants locked in the filamentous form like tup1Δ/Δ or nrg1Δ/Δ are less virulent as well [17][18][19] , implying that morphological plasticity is essential for virulence in murine disseminated candidiasis.
One of the factors required for maintenance of hyphal growth, invasion and damage of C. albicans is encoded by EED1 (Epithelial Escape and Dissemination 1 7,20 ). A homozygous eed1Δ/Δ deletion mutant is still able to initiate germ tube formation, but fails to elongate these into hypha and eventually switches back to yeast cell growth 7,20,21 . Here we show that the eed1Δ/Δ mutant is fully virulent in a murine model of systemic candidiasis and confirmed this finding using a tetracycline-regulable expression system to induce or repress EED1 in C. albicans in vivo. Virulence of these filament-deficient mutants is associated with high yeast proliferation rates in vivo and enhanced growth in vitro on nutrient sources likely encountered in the host, suggesting metabolic adaptation as the underlying mechanism for retained virulence in the absence of filamentation. Analysis of host responses furthermore supports a contribution of altered immunopathology to C. albicans virulence in the absence of filamentation.

Results
EED1 is required for virulence in a murine intraperitoneal infection model. Hypha formation has been shown to be essential for tissue invasion and damage in vitro and in vivo 11,14,22 . Consistent with this, the eed1Δ/Δ mutant, unable to maintain hyphal growth, was not able to damage renal, hepatic and oral epithelial cells in vitro within 24 h of co-incubation (Fig. 1a). To assess the impact of EED1 for tissue invasion and damage in vivo, we employed an intraperitoneal infection model. In this model, C. albicans filamentation facilitates invasion from the peritoneal cavity into intraperitoneal organs such as liver, spleen and pancreas 23 . In line with the in vitro results, deletion of EED1 led to reduced damage of liver and pancreatic tissue indicated by lower serum levels of tissue-specific enzymes and lower clinical scores compared to mice infected with the wild type strain (WT; Supplementary Fig. 1). This attenuated virulence phenotype was confirmed using a conditional knock out mutant (t-EED1), in which one EED1 allele was deleted and the other was placed under the control of a tet-OFF promoter 14 . In the presence of doxycycline (t-EED1 + ), repressed EED1 expression leads to yeast cell growth. In absence of doxycycline (t-EED1−), the gene is constitutively expressed resulting in increased filamentation on solid media (Supplementary Fig. 2a) and filamentation comparable to the WT in liquid media ( Supplementary Fig. 2b). In the presence of doxycycline, mice developed significantly less clinical symptoms after intraperitoneal infection with the t-EED1+ yeast compared to the WT and the t-EED1− filamentous strain 24 h post infection (p.i.) (Fig. 1b). Histological analysis showed t-EED1+ yeast that remained in the upper tissue layers below the liver capsule whereas WT and t-EED1− hyphae invaded deeply into liver parenchyma (Fig. 1c). Furthermore, infection with WT and t-EED1− induced significantly increased serum levels of alanine aminotransaminase (ALT) and pancreatic amylase indicating liver and pancreatic injury, respectively (Fig. 1d). In contrast, reduced tissue invasion of the t-EED1+ yeast resulted in serum enzyme levels of ALT and pancreatic amylase comparable to the uninfected control. Interestingly, a significantly higher fungal burden was recovered from liver tissue of mice infected with either the t-EED1+ or eed1Δ/Δ mutant compared to the respective WT ( Fig. 1e and Supplementary Fig. 1c). Thus, tissue damage and clinical symptoms correlated with the ability to form hyphae in the intraperitoneal infection model, and deletion or repression of EED1 resulted in attenuation of virulence despite a higher liver fungal burden.
Deletion of EED1 affects dissemination from the gastrointestinal tract in immunosuppressed mice. It is generally believed that the gastrointestinal tract is the main reservoir for C. albicans in humans 24,25 . In order to disseminate systemically, the fungus needs to translocate across the mucosal barrier 24,25 . Consistent with the in vitro results for renal, hepatic and oral epithelial cells, deletion of EED1 significantly reduced the ability of C. albicans to damage enterocytes under normoxic and hypoxic conditions (Fig. 2a), resulting in maintained barrier function and reduced translocation in vitro (Fig. 2b, c). Colonization of the intestinal tract was moderately increased for eed1Δ/Δ compared to the WT in antibiotic-treated immunocompetent mice, but not in mice treated with cyclophosphamide ( Fig. 2d, e). Dissemination of the eed1Δ/Δ mutant to internal organs was, however, only observed upon cyclophosphamide treatment (Fig. 2f), where a significantly higher number of colony forming units (CFU) in kidneys were observed for eed1Δ/Δ compared to the WT. Of note, while cyclophosphamide reduces the numbers of immune cells and also impacts the intestinal barrier function 26,27 , we did not observe any symptoms of disseminated disease in the mice. This suggests that the treatment was not sufficient to induce the level of immunosuppression and/ or intestinal damage required for translocation of a sufficiently high number of fungal cells to cause disease 27 . Nonetheless, these results support the concept that growth in the yeast form is beneficial for intestinal colonization 28,29 (at least in immunocompetent mice) and dissemination 12 , but that filamentation is required for active translocation 22 .
EED1 is not essential for virulence in a systemic infection model. Based on the in vitro data and results from the intraperitoneal infection model, we expected the EED1 mutants also to be attenuated in a murine systemic (intravenous) infection model that mimics catheter-associated disseminated candidiasis in humans. Surprisingly, the eed1Δ/Δ mutant showed enhanced virulence compared to the WT when low infectious doses were used (Fig. 3a) whereas with a high infectious dose mortality was delayed for t-EED1 + yeast (Fig. 3b). Intermediate infectious doses led to virulence and disease progression indistinguishable to the WT when t-EED1 was either forced to grow as yeast or as hyphae (Fig. 3c, Supplementary Fig. 3). Histology confirmed that t-EED1 + grew as yeast in vivo (Fig. 3d). Consistent with previous reports, experiments conducted in parallel with the tet-NRG1 strain 14 resulted in clinical disease only when filamentation was facilitated by the presence of doxycycline ( Supplementary  Fig. 4a-c 14 ). We also noted that mice infected with the WT (THE1-CIp10) reached humane endpoints earlier when receiving doxycycline; this is likely a consequence of lower water uptake in these groups (see method section for details). Determination of CFU in mice infected with the intermediate dose showed similar initial fungal burden 6 h p.i. in kidney, liver, brain and spleen (Fig. 3e). However, already 24 h p.i. we observed significantly increased (~25-fold higher) fungal burden in the kidneys of mice infected with t-EED1 + yeast compared to the filamentous strains. Likewise, fungal burden increased for tet-NRG1-yeast, reaching significantly higher levels 72 h p.i. compared to the WT-, but in comparison to t-EED1 + yeast declined until the end of the experiment (Supplementary Fig. 4d). Furthermore, fungal burden of t-EED1 + yeast in the brain continued to increase until mice became moribund, whereas after infection with filamentous strains cell numbers peaked 24 h p.i. and were stable or declined thereafter. After an initial decline, the CFU of the t-EED1 + yeast stabilized and increased in liver and spleen while they continuously declined after infection with the WT, t-EED1-or tet-NRG1 yeast or hypha ( Fig. 3e; Supplementary Fig. 4d). To test the possibility that differences in fungal load were due to the underestimation of CFU plated from filamentous C. albicans strains, we performed quantitative PCR targeting the C. albicans 18 S rRNA gene RDN18 in complex DNA isolated from infected kidney homogenates 30 . A strong correlation (R 2 = 0.7076) between CFU count and DNA content irrespective of the fungal morphology ( Supplementary Fig. 5) supports the CFU results.
During hematogenously disseminated candidiasis EED1 deficiency leads to delayed renal cytokine production but increased immune cell infiltration at later time points. We hypothesized that the increased fungal burden observed for t-EED1 + yeast could be the result of an altered early innate immune response leading to reduced fungal killing. Therefore, we quantified immune cells in the kidney over the course of infection. Leukocytes started to infiltrate kidneys 24 h p.i. and no difference was observed in the overall number or subpopulations of leukocytes 24 and 48 h p.i. between WT + and t-EED1 + yeast (Supplementary Figs. 6-7). However, morphology-dependent significant differences between WT + and t-EED1+ were observed 72 h p.i.: Total leukocyte numbers were approximately 2-fold higher in response to t-EED1+ yeast (Fig. 4a), with increased numbers of monocytes (4.5-fold), macrophages (3.2-fold) and DCs and NK cells (2-fold). This coincided with 190-fold higher yeast fungal burden compared to WT+ (Fig. 3e). Interestingly, significantly less leukocytes accumulated in kidneys after infection with the t-EED1− filamentous strain compared to WT− ( Supplementary  Fig. 8), despite comparable CFU 72 h p.i. (Fig. 3e). While renal cytokine levels induced by the filamentous t-EED1− were comparable to the WT− throughout the course of infection (Supplementary Figs. 9-10), less pro-inflammatory cytokines were induced by t-EED1+ yeast than by WT+ 24 h p.i. (Fig. 4b), despite higher fungal burden of t-EED1+ yeast. In contrast, cytokine levels were comparable 48 h p.i., and a tendency towards increased cytokine release, with significant higher amounts of IL-18, IP-10, RANTES and IFN-ɣ in response to t-EED1+ yeast, was observed 72 h p.i. (Fig. 4b). In moribund mice no morphologydependent difference in cytokine levels was observed (Supplementary Fig. 10). To investigate whether the local differences in renal cytokine production were reflected by differences in systemic inflammation, we determined serum levels of the sepsis markers soluble triggering receptor expressed on myeloid cells (sTREM-1) and neutrophil gelatinase-associated lipocalin (NGAL). NGAL and sTREM-1 levels increased rapidly and throughout the course of infection (Fig. 4c) irrespective of the fungal morphology.
Lower cytokine levels induced by t-EED1+ yeast in the presence of similar leukocyte numbers 24 h p.i. suggested differences in the interaction with leukocytes and/or epithelial cells and their interaction with fungal cells was therefore analyzed in vitro. Murine bone marrow (BM) neutrophils and BM-derived macrophages (BMDMs) were capable of phagocytosing and killing C. albicans WT and eed1Δ/Δ mutant to the same extent in vitro after 1 h and 2 h, respectively (Fig. 5a, b; Supplementary  Fig. 11a). At that time point both strains had formed germ tubes (Supplementary Fig. 11b and Supplementary Fig. 12). After 6 h, the WT maintained filamentation while the eed1Δ/Δ mutant had switched back to yeast cell growth ( Fig. 5c; Supplementary  Fig. 12). Analysis at 6 h showed that a significantly larger proportion of C. albicans eed1Δ/Δ was killed by BMDMs compared to the WT. However, we noted that the relative increase in CFU of the eed1Δ/Δ mutant from 0 h to 6 h was considerably higher than for the WT both in the presence and absence of macrophages (Fig. 5d). This resulted in significantly higher numbers of eed1Δ/Δ than WT cells in the presence of BMDMs. In addition, the eed1Δ/Δ mutant showed a reduced capacity to damage BMDMs (Fig. 5e) and to stimulate TNF-α release by BMDMs 24 h p.i. (Fig. 5f). At that time wells that contained the C. albicans WT were completely covered by hyphae whereas wells with eed1Δ/Δ showed yeasts only and macrophages were visible from time to time ( Supplementary Fig. 12). In contrast, activation of murine PMNs determined by the release of IL-10, IL-6 and TNF-α ( Supplementary Fig. 11c) and production of reactive oxygen species (ROS; Supplementary Fig. 11d) was comparable between the strains. Infection of renal, hepatic and oral epithelial cells with the eed1Δ/Δ mutant induced no increase of pro-inflammatory cytokines compared to uninfected controls (Fig. 5g), likely due to the lack of damage caused by the mutant in this model (Fig. 1a). For oral epithelial cells, no significant increase was observed for the WT either, although a tendency of higher cytokine production was observed. Thus, the initial lower cytokine release in response to t-EED1+ yeast in vivo is likely due to reduced damage of and lower cytokine production by renal and hepatic cells and macrophages, which is balanced by the response of recruited immune cells at later time points. eed1Δ/Δ yeast display enhanced growth on physiologically relevant carbon sources. EED1 depleted yeast were able to increase their cell number approximately 100-fold from 6 h to 24 h p.i. in murine kidneys (which equals a generation time of 2.6 h Fig. 1 Hyphal elongation of C. albicans is required for tissue damage in vitro and virulence in an intraperitoneal infection model. a Damage of oral (TR146), hepatic (HepaRG) and renal (A498) epithelial cells was quantified by measuring the release of lactate dehydrogenase (LDH) into the supernatant after co-incubation for 24 h with WT (SC5314) or eed1Δ/Δ mutant. Uninfected cells served as negative control. LDH released by WT was set to 100%. Data are presented as mean ± SD from four biologically independent experiments. Within each experiment, supernatants from three wells infected by the same strain were pooled and LDH was measured. Data was analyzed by two-tailed students t-test. b-e t-EED1 + yeasts are attenuated in invasion, damage and virulence potential despite higher organ fungal loads in the intraperitoneal infection model 24 h post infection. Mice were infected intraperitoneally with 1 × 10 8 cells of WT (THE1-CIp10) or t-EED1 in the presence (+) or absence (−) of doxycycline supplied via the drinking water. Data derived from two independent experiments, WT− and WT+ n = 10, t-EED1− and t-EED1+ n = 9 animals per group. b Repression of EED1 by doxycycline led to significantly reduced clinical symptoms in mice. The semiquantitative clinical score was determined by assessing fur, coat and posture, behavior and lethargy, fibrin exudation and other symptoms like diarrhea. The score ranges from 0 (no symptoms) to 10 (severe illness). c Representative images of periodic acid-Schiff stained histological sections of liver tissue 24 h p.i., fungal cells are stained purple. Arrows point towards the liver surface (capsule, partially destroyed). The scale bar represents 100 µm and applies to all images. d Damage of liver and pancreas was quantified by measuring serum levels of alanine aminotransaminase (ALT) and pancreatic amylase, respectively. Enzyme levels of uninfected mice (n = 9) served as negative control. e Fungal burden in the liver. b, d, e Median and interquartile range are shown, two-sided Mann-Whitney test. Asterisks above bars represent significant differences compared to the uninfected control *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. Source data are provided as a Source Data file. compared to 7.1 h for the WT). Since we found no evidence for reduced recognition or killing by immune cells, this observation cannot be explained by increased immune evasion resulting in less killing. We thus hypothesized that EED1-deficient yeasts are metabolically better adapted to acquire locally available nutrients resulting in increased proliferation. Therefore, we analyzed the growth of C. albicans WT and the eed1Δ/Δ mutant in the presence of different fermentable and non-fermentable carbon sources at 37°C (Fig. 6). Growth of both strains was similar with 2% glucose in complex (YPD) and defined media (SD). However, in the presence of physiological glucose concentrations of 0.1% in SD medium, the mutant reached lower cell densities at stationary phase. In YPD with 0.1% glucose, in contrast, the eed1Δ/Δ mutant continued to grow steadily after depletion of glucose reaching higher ODs than the WT. Importantly, the eed1Δ/Δ mutant showed enhanced growth in the presence of the physiological relevant alternative carbon sources lactate, acetate, citrate, amino acids and the amino sugar N-acetyl-glucosamine (GlcNAc). Additionally, the eed1Δ/Δ mutant showed faster onset of growth in YCB-BSA compared to the WT indicating that this strain is highly proteolytically active. Furthermore, eed1Δ/Δ showed better growth than the WT in murine kidney homogenates. Taken together, these data suggest that the absence of EED1 results in better metabolic adaptation in vivo, facilitating faster proliferation resulting in increased fungal burden.
To gain information on possible mechanisms mediating the enhanced proliferation of the eed1Δ/Δ mutant on alternative carbon sources, we performed RNAseq analysis of cells grown in SD with glucose at 30°C for 10 h (designated as 0 h time point), followed by a shift to citrate or casamino acids as sole carbon source at 37°C (2 h, 6 h, and 12 h; Supplementary Fig. 13a). Growth in SD with glucose resulted in yeast morphology for both strains ( Supplementary Fig. 13b); on casamino acids, both strains formed germ tubes within 2 h, but only the WT formed hyphae at later time points (Supplementary Fig. 13b). With citrate as sole carbon source, filamentation was observed only for the WT ( Supplementary Fig. 13b). In both, citrate and amino acid media, the eed1Δ/Δ mutant showed enhanced growth as observed by optical density as well as by determination of dry weight ( Supplementary Fig. 14a). Principal component analysis (PCA) of the RNAseq data showed that biological replicates clustered together and that the transcriptome of both strains varied over time ( Supplementary Fig. 14b). Comparison of eed1Δ/Δ mutant and WT for each condition and time point (with a log 2 fold change of 2.0 as cut off) revealed that only one yet uncharacterized gene (orf19.2962) was differentially expressed during yeast growth at the 0 h time point (2.1-fold down-regulated). After the shift to media containing either citrate or casamino acids as sole carbon sources, relatively few genes (<320) were differentially expressed in the mutant compared to the WT at any given time point, and the majority of differentially expressed genes (DEGs) were down-regulated in eed1Δ/Δ compared to the WT (Supplementary Fig. 15). Consistent with differences in morphology, down-regulated genes in both conditions included genes associated with filamentous growth, biofilm formation, and adhesion (e.g. ECE1, HGC1, BRG1, UME6, HYR1, SAP5, HWP1 and ALS1; Supplementary Fig 16a). The expression of secreted aspartyl proteinases likewise coincided with morphology, with the hypha-associated genes SAP4-6 31 showing reduced transcription while transcription of the yeast-associated genes SAP1 and SAP3 31 was enhanced in the eed1Δ/Δ mutant during growth on casamino acids ( Supplementary Fig. 16b). The highest number of DEGs was observed at the 12 h time point (Fig. 7a), when the eed1Δ/Δ mutant grew as yeast but also the filamentous WT switched back to yeast cell growth to some extent (Fig. 7b). Genes that were up-regulated in eed1Δ/Δ mutant compared to WT in both media included the ammonium permease MEP1, the putative 2-isopropylmalate synthase LEU4, the secreted yeast wall protein YWP1, the transcription factor MSS11, and 5 yet uncharacterized open reading frames. Most of the DEGs showed a medium-specific regulation. After 12 h of growth with citrate as sole carbon source Gene Ontology (GO) term analysis identified enrichment of genes associated with organic acid, (long-chain) fatty acid and small molecule biosynthetic and metabolic processes within the up-regulated DEGs in the eed1Δ/Δ mutant (Fig. 7c), which might explain the better growth of the mutant on citrate. Genes that were up-regulated exclusively in citrate included e.g., the dicarboxylic acid transporter JEN2 and a key enzyme of gluconeogenesis, PCK1. Of note, the only significantly enriched GO term after 12 h of growth on casamino acids in the genes up-regulated in eed1Δ/Δ was found to be carbohydrate transport. However, individual genes that were significantly upregulated included the broad specificity amino acid permease GAP2, the predicted amino acid transmembrane transporter UGA5, the GATA-type transcription factor GAT1 that regulates nitrogen utilization, and the glyoxylate cycle enzyme isocitrate lyase ICL1 (Supplementary Data 1), which might contribute to enhanced growth. In addition to GAT1, three other transcription factors (TFs) were up-regulated and 13 down-regulated in the eed1Δ/Δ mutant 12 h after shift to either citrate of casamino acids ( Supplementary Fig. 16c). Several down-regulated TFs were associated with filamentous growth and thus, differential expression likely reflects differences in morphology (e.g., UME6, SFL2, BRG1, TEC1, OFI1 32 , ACE2 33 ). Counterintuitively, though, transcription of MSS11, encoding a factor interacting with Flo8 to activate transcription of hypha-specific genes 34 , was higher in the mutant in both citrate and amino acid media. Similarly, expression of RFX2 was reduced in the eed1Δ/Δ mutantdeletion of this gene, however, results in hyper-filamentation and increased expression of hypha-associated genes 35 . Interestingly, both WOR1 and WOR3 were significantly down-regulated after prolonged growth in both media, possibly indicating increased commitment of the eed1Δ/Δ mutant to grow as white cells. White cells have been shown to have a metabolic advantage over opaque cells with various nutrients at 37°C 36 , lead to higher renal fungal burden, and are more virulent in systemic infection models 37,38 .

High numbers of yeast cells result in renal damage in vivo.
To determine if immune cell recruitment together with high fungal burden during t-EED1 + infection was sufficient to cause renal damage in the absence of invading hyphae, we measured KIM-1, a biomarker for renal proximal tubule injury in the urine of infected mice 39,40 . KIM-1 levels increased after infection with all strains without significant differences (Fig. 8a). Additionally, we Fig. 2 The eed1Δ/Δ mutant shows a reduced translocation capacity in vitro and in a gut dissemination model in immunocompetent but not in immunocompromised mice. a-c C2BBe1 cells were infected with C. albicans WT (SC5314) or the eed1Δ/Δ mutant and incubated under normoxic (21% O 2 ) or hypoxic (1% O 2 ) conditions for 24 h. a Damage of C2BBe1 cells was quantified by measurement of LDH and is shown in % of high control (cells lysed with Triton X-100). Uninfected cells served as negative control. b Integrity of epithelial barrier was quantified by TEER measurement. Data are expressed as change of TEER 24 h post infection compared to TEER values prior to infection in percent. c Translocation across the epithelial barrier was assessed by CFU plating. a-c Graphs show the mean ± SD of three independent biological replicates, two-tailed students t-test. p-values are shown in the graph. d Fungal burden in feces of antibiotic-treated immunocompetent and mice rendered immunocompromised by cyclophosphamide treatment. Immunocompetent, day 2 and 4 n = 8; day 7 and 14 n = 4 mice per group, one experiment. Immunocompromised WT infected n = 3; eed1Δ/Δ n = 8, one experiment. e Schematic overview of the timeline for the gastrointestinal dissemination model. f Fungal burden in liver (left) and kidneys (right) of immunocompetent or cyclophosphamide-treated immunosuppressed (+Cyclo) mice colonized with either C. albicans WT or eed1Δ/Δ mutant 14 d post infection. WT n = 3, one experiment; WT + Cyclo and eed1Δ/Δ n = 4, one experiment; eed1Δ/Δ + Cyclo n = 8 mice, data from two independent experiments. Dashed lines indicate limit of detection. d, f Shown is the median with interquartile range, two-sided Mann-Whitney test, p-values are shown in the graph. Source data are provided as a Source Data file. Fig. 3 C. albicans retains its virulence potential in a mouse model of systemic candidiasis even in the absence of hyphal elongation due to repression or deletion of EED1 accompanied by higher organ fungal loads. Mice were intravenously infected with the following doses: a Low doses of 1 × 10 3 or 1 × 10 4 CFU/g body weight of WT (SC5314) or eed1Δ/Δ mutant. One experiment, n = 8 mice per group (except for WT 1 × 10 3 , n = 7). b A high dose of 1.25 × 10 5 CFU/g body weight of WT (THE1-CIp10) or t-EED1 in the presence of doxycycline. One experiment, n = 10 mice per group. c An intermediate dose of 2.5 × 10 4 CFU/g body weight with WT (THE1-CIp10) or t-EED1 in the presence (+) or absence (−) of doxycycline. One experiment, n = 10 mice per group. a-c Survival was monitored over a course of 20 or 21 days and is shown as Kaplan-Meyer curve. Survival curves were compared using the two-sided Log-rank (Mantel-Cox) test, p-values are shown in the graph. d Representative images of PAS stained histological cross sections of kidneys from moribund mice infected with 2.5 × 10 4 CFU/g body weight with WT (THE1-CIp10) or t-EED1 in the presence (+) or absence (−) of doxycycline. Fungal cells are stained purple. Scale bar represents 20 µm and applies to all images. e Organ fungal loads of mice infected with 2.5 × 10 4 CFU/g body weight with WT (THE1-CIp10) or t-EED1 in the presence (+) or absence (−) of doxycycline. Shown is the number of CFU per gram organ from mice sacrificed 6, 24, 48 and 72 h post infection and from moribund mice (mor). Data are shown as median with interquartile range. Dashed gray lines indicate limit of detection. Moribund: one experiment, n = 10; Kidney: two independent experiments n = 10. Liver, spleen, brain: three independent experiments, n = 15. Two-sided Mann-Whitney test, asterisks indicate significant changes of t-EED1+ compared to WT+ groups at the indicated time points. *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24095-8 ARTICLE NATURE COMMUNICATIONS | (2021) 12:3899 | https://doi.org/10.1038/s41467-021-24095-8 | www.nature.com/naturecommunications evaluated kidney function by measuring serum creatinine and blood urea nitrogen (BUN) levels. Both increased at comparable rates, reaching highly elevated levels indicating severe renal dysfunction 41,42 in moribund mice only (Fig. 8b, c). Finally, the same extent of apoptosis was observed in the kidneys of moribund mice infected with either WT or t-EED1 (Fig. 8d, e). Therefore, enhanced fungal proliferation can compensate for the reduced capacity of hypha-mediated damage in C. albicans lacking expression of EED1.
Rapid yeast proliferation in combination with immunopathology contributes to mortality. To investigate whether mortality caused by the yeast form is the result of high organ fungal loads interfering with organ function or whether immunopathology is the driving force in the pathogenic process, mice successfully depleted of peripheral neutrophils and monocytes ( Supplementary Fig. 17) were intravenously infected with low infectious doses (1 × 10 2 and 1 × 10 3 CFU/g body weight). Although the eed1Δ/Δ mutant led to 100% mortality in immunosuppressed mice, this was significantly delayed in comparison to WT strain infections (Fig. 9a). However, in immunocompetent mice challenged with 1 × 10 3 CFU/g body weight, the obverse effect was observed: while 6/7 mice survived in the WT, only 3/8 mice survived in the eed1Δ/Δ group until the end of the experiment (Fig. 3a). In addition, histological analysis of renal tissue showed more pronounced immune cell infiltrations in response to eed1Δ/Δ yeast in immunocompetent mice (Fig. 9b), suggesting some contribution of immunopathology to yeast driven mortality. In the absence of neutrophils and monocytes, fungal burden of WT and eed1Δ/Δ mutant progressively increased during the course of infection. However, organ fungal load of the eed1Δ/Δ mutant exceeded that of the WT in all organs tested, with significant differences detectable as early as 12 h post infections in liver and kidneys (Fig. 9c). At the humane endpoint, the fungal burden of the eed1Δ/Δ mutant exceeded those of the WT in all organs tested ( Fig. 9c and Supplementary Fig. 18a) with a 150fold and 280-fold increase in kidney and spleen fungal burden, respectively. Furthermore, the urinary KIM-1 to creatinine ratio steadily increased after infection with C. albicans eed1Δ/Δ, indicating that filaments are not essential for induction of renal injury (Fig. 9d). However, kidney function as determined by measurement of BUN in serum of immunosuppressed mice was not affected at any time after infection with WT or eed1Δ/Δ mutant ( Supplementary Fig. 18b), possibly because the time to the humane endpoint was too short to allow accumulation of waste products by impaired renal clearance.

Discussion
C. albicans hypha formation has been linked to invasion and damage, and C. albicans strains locked in either the yeast or hyphal morphology were repeatedly found to be less virulent in systemic infection models [15][16][17][18] . Consequently, morphological plasticity is considered as an important virulence trait of C. albicans 43,44 . Consistent with the role of filamentation for invasion and damage, repression or absence of EED1 in C. albicans, interfering with hyphal elongation and leading to yeast growth, resulted in significantly reduced capacity to damage epithelial cells in vitro, and reduced virulence in a murine intraperitoneal infection model in vivo although the mutant was present in higher numbers than the WT in the liver. The observation that following intraperitoneal injection EED1-deficient cells superficially invaded the liver parenchyma but remained below the liver capsule, without causing detectable damage, resembles the described behavior of the eed1Δ/Δ mutant in a reconstituted human epithelium model 20 . Thus, our results support previous research demonstrating that hyphae are important for tissue infiltration. In the gut, however, C. albicans can be found in both yeast and hyphal form 28,45,46 . Some studies found higher colonization rates for strains locked in the yeast morphology (e.g., the cph1Δ/Δ efg1Δ/Δ double mutant and the hgc1Δ/Δ mutant 28 ), whereas enforced filamentous growth resulted in lower intestinal colonization (e.g. C. albicans nrg1Δ/Δ and tup1Δ/Δ mutants 27,28 , repression of TUP1 29 , and constitutive expression of UME6 28,45 ). We likewise observed higher colonization rates for the eed1Δ/Δ mutant in immunocompetent mice, supporting the concept that yeast growth favors intestinal colonization. How C. albicans is translocating across the intestinal barrier remains unknown. In vitro, translocation requires filamentation 22 , but low level translocation has been observed in immunocompetent mice independent of fungal morphology 28 . This has been suggested to be mediated by transport via host cells that sample the intestinal content 28 . We observed reduced basal dissemination of the eed1Δ/Δ mutant to the liver, which might be a consequence of altered interaction of the mutant with dendritic cells or M cells mediating entry into the bloodstream by lumen sampling or transcytosis 28,47 in the absence of hyphal formation, but this aspect was not investigated further in this study. In contrast, increased fungal burden in liver and kidney was observed in C. albicans eed1Δ/Δ colonized mice treated with cyclophosphamide, and in these mice kidney CFU were significantly higher for the eed1Δ/Δ compared to the WT. However, this data does not allow any conclusion on the translocation process, as the higher fungal load could also be a consequence of increased proliferation of the mutant within this organ. It should furthermore be noted that the fungal burden in kidneys following translocation from the gut was substantially lower than that observed after intravenous infection, which explains why colonized mice did not develop signs of systemic candidiasis upon cyclophosphamide treatment.
Surprisingly, the filamentation defect caused by EED1 deficiency did not impair virulence in a systemic infection model when intermediate infectious doses were injected directly into the bloodstream. Remarkably, infectious doses lower than 10 4 cells per gram body weight resulted in increased virulence of the eed1Δ/Δ mutant compared to the WT, whereas with high infectious doses or in absence of neutrophils and monocytes mortality was delayed but still reached 100%. In the murine model of hematogenously disseminated candidiasis, infection is established via the lateral tail vein mimicking disseminated C. albicans infections in humans 48 . While virtually all organs can get affected during infection the kidney is not able to control fungal growth 49 . The fungus reaches the kidneys via the renal artery and the afferent arterioles. In order to invade into tubules and the renal parenchyma, the fungus needs to adhere to endothelial cells and pass through into the renal cortex. This step probably does not require fungal invasion by hypha formation, as tet-NRG1-yeast and a hgc1Δ/Δ mutant defective in hyphal formation were shown to traverse from the bloodstream to the renal endothelium 14,16 . Although attenuated in its damage capacity, the eed1Δ/Δ mutant Asterisks above bars represent significant differences compared to the uninfected control. Two independent experiments, n = 10, except for uninfected+, n = 9. MIP-1α, TNF-α for uninfected+, n = 5. IL-10 data derived from one experiment, n = 5. c Markers for systemic inflammation sTREM-1 and NGAL were quantified by ELISA in serum of mice after 6, 24, 48, 72 h and when mice become moribund (mor). Dashed lines indicate median serum protein levels of uninfected controls in the presence (dark gray) or absence of doxycycline (light gray). Two independent experiments. sTREM-1 n = 10, except for 6 h n = 6 and mor n = 5; for WT-24 h n = 8 and 48 h n = 9. NGAL n = 10, except for WT− 24 h n = 9, t-EED1− 24 h n = 8 and 72 h n = 9. a-c Shown is the median and interquartile range, two-sided Mann-Whitney test. a, b p-values are shown in the graph and b asterisks indicate significant differences in comparison to the uninfected control (*p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001). Source data are provided as a Source Data file. is able to adhere and invade into epithelial cells in vitro 20,50 , which might be sufficient to mediate entry into renal tissue in vivo. Thus, while filamentation is required for active tissue penetration in the intraperitoneal infection model, it might be dispensable if defensive barriers are breached by direct injection into the blood stream.
Following establishment of the fungus in the kidney in immunocompetent mice, pathogenesis is thought to be driven by both direct fungus-mediated damage and immunopathology. Enhanced immunopathology could compensate for the lack of hypha-mediated fungal damage during pathogenesis; however, in the early time points following infection with an intermediate infectious dose of t-EED1+ yeast we rather observed a reduced induction of pro-inflammatory cytokines, likely as a result of the reduced capacity of the yeast to directly damage epithelial cells and macrophages and stimulate cytokine release. Macrophages have been shown to be less prone to killing by hypha-deficient strains 51 and to release less TNF-α after stimulation with C. albicans strains unable to germinate; 52 likewise, murine macrophages produced less TNF-α in response to the eed1Δ/Δ mutant. However, the amount of TNF-α in infected kidneys was not affected by absence of EED1-driven filamentation and the interaction with murine neutrophils in vitro was comparable for WT and eed1Δ/Δ mutant. Interestingly, while killing of the mutant by BMDMs was increased after 6 h of co-incubation, the relative increase in CFU was higher for the eed1Δ/Δ mutant than for the WT, suggesting that the increased proliferation is sufficient to compensate for the higher killing rate mediating survival and allowing for continuous organ colonization. It thus appears unlikely that the higher renal fungal burden of the yeast is solely due to reduced fungal clearance as a result of impaired recognition by and/or activation of immune cells.
We hypothesized that increased fitness mediated by better metabolic adaptation to the locally available nutrients facilitated rapid proliferation of EED1 deficient yeast in vivo in the systemic infection model, and that the higher fungal burden in the liver after intraperitoneal infection might likewise reflect increased growth of the mutant. Upon infection C. albicans faces a hostile environment with varying nutritional compositions. Preferred carbon sources such as glucose are present only in low concentrations in the bloodstream (0.06-0.1% 53 ,) and can become scarce in microenvironments or deprived by phagocytes upon ingestion 10 . Consequently, C. albicans relies on the assimilation of alternative carbon sources in order to survive and proliferate within the host 54 . As Crabtree-negative yeast, C. albicans is able to assimilate glucose and alternative carbon sources such as amino acids, fatty acids, and carboxylic acids, at the same time 55 . This metabolic flexibility is known to increase colonization, resistance to phagocytic recognition 56 and killing, and enhances pathogenicity 57,58 . The eed1Δ/Δ mutant indeed showed enhanced growth in kidney homogenates and in the presence of alternative carbon sources that are available in vivo, such as acetate, lactate, amino acids, citrate, and the amino sugar N-acetyl-glucosamine. Additionally, the eed1Δ/Δ mutant displayed an enhanced proteolytic activity, possibly supporting utilization of host proteins and immune evasion by degradation of complement proteins and antimicrobial peptides in vivo [59][60][61] . Whereas hypha-associated SAPs appeared to be down-regulated during growth on casamino acids, enhanced expression of SAP1, SAP3, SAP7, SAP9 and SAP10 was observed. Since Sap1-3, Sap4-6 and Sap9-10 have different pH optima for activity, pH 3-5, pH 5-7 and pH 5-8, respectively, and in addition Saps differ in their substrate specificity 31,62 , this compensatory gene expression might lead to the earlier onset of BSA utilization observed for the eed1Δ/Δ mutant. The increased growth on citrate is especially interesting as citrate is present in blood in a range of 0.05 to 0.3 mM, freely filtered in the glomerulus, and extensively reabsorbed in the nephrons of the kidney 63 . Therefore, citrate levels are c Network analysis of Gene Ontology (GO) term enrichments of significantly up-regulated genes (+ log 2 2 and adjusted p-value <0.05) in eed1Δ/Δ mutant compared to WT after 12 h of growth on citrate. Font size represents the p-value (p < 0.1) of the GO term: big letterssmall p value, ranging from p < 2.3 × 10 −5 to low p-value p < 0.09 indicated by small letters. Color is representing the uniqueness, indicating how particular each term is with respect to the set of terms being evaluated (original data analyzed are available in Supplementary Data 1). approximately 3-4-fold higher in the renal cortex than in plasma. Furthermore, mice develop metabolic alkalosis early after systemic infection 48 which is known to further increase citrate concentration in the cortex, and in addition enhances renal citrate excretion 20-fold compared to the steady state 63 . Therefore, citrate could be an abundant source of carbon available for C. albicans in the renal cortex and tubules during the early phase of systemic infection. Thus, enhanced growth on citrate might contribute to the higher renal fungal burden in the absence of EED1 expression. Transcriptional profiling during growth of C. albicans on citrate as sole carbon source revealed only few genes that were up-regulated in the eed1Δ/Δ mutant compared to the WT after 12 h. These, however, included genes involved in metabolism of carboxylic acids and hence could aid in the metabolism of citrate. Similarly, upregulation of GAT1 might be promoting growth on casamino acids. However, many metabolic enzymes are regulated by posttranscriptional modifications to respond quickly to changing environmental conditions 64,65 , and it is therefore likely that the transcriptional changes described here reflect only in part the Infection with t-EED1+ yeast resulted in higher organ fungal burden, significantly more leukocytes infiltrating the kidneys and increased renal pro-inflammatory cytokine production at later time points contributing to local immunopathology 49,66 . Markers of systemic inflammation, sTREM-1 and NGAL, increased rapidly in serum of infected mice irrespective of fungal morphology. NGAL thresholds defining sepsis in humans 67 and sTREM-1 thresholds in mice 68 were reached within 24 h after infection, while damage of proximal tubular cells measured by KIM-1 became detectable slightly later (between 24 and 48 h p.i.). Impaired kidney function determined by increased blood urea nitrogen level and serum creatinine was only evident in moribund mice, consistent with findings by Spellberg et al. 48 , again without significant differences between hypha-forming and filamentdeficient strains. Histologically, t-EED1+ yeast were found in large numbers mainly within tubules and in the renal pelvis; it appears possible that accumulation of yeast led to obstruction of tubules and the collecting duct system 69 , increasing the intrarenal pressure and thereby, in addition to immunopathology, causes renal tissue damage. This is supported by our data showing that the area of apoptotic tissue in kidneys did not differ between mice infected with WT and t-EED1+ and suggests that the lack of hyphae-driven direct damage can be compensated by enhanced yeast cell growth. The exact molecular mechanisms by which absence of EED1 influences both morphogenesis and growth on alternative carbon sources remains unknown; however, it should be noted that retention of some level of virulence in the absence of hyphal elongation is not a unique feature of EED1 deficient strains. Homozygous deletion of UME6, resulting in a similar filamentation defect as observed for the eed1Δ/Δ mutant, leads to significantly attenuated virulence, but still causes lethal infections in mice 70 . Virulence of a sfl2Δ/Δ mutant was comparable to the WT despite a filamentation defect 71 . In contrast to the eed1Δ/Δ mutant, systemic infection with the sfl2Δ/Δ mutant, however, did not result in increased fungal burden, and some filamentation occurs in the absence of SFL2 71,72 . The role of SFL2 for interaction with immune cells has not been investigated so far, and it thus remains unclear if differences in immunopathology contribute to the virulence of the sfl2Δ/Δ mutant. Furthermore, a screen of a C. albicans deletion library identified mutants with defects in filamentation but with unaltered infectivity; 73 a detailed analysis of these mutants might reveal additional strains for which virulence in murine systemic candidiasis models does not depend on filamentation.
Of note, significantly higher CFUs in kidneys were also observed for tet-NRG1-yeast on day 3 after systemic infection compared to both the parental strain ( Supplementary Fig. 4d) and the same strain with doxycycline (p = 0.001), consistent with observations by Saville et al. 14 . However, the fungal burden of tet-NRG1-yeast was approximately 1-log lower than for t-EED1+ yeast at this time point (mean 3.19 × 10 5 ± 2.5 × 10 5 compared to 3.45 × 10 6 ± 1.15 × 10 5 ), and declined over time ( Supplementary  Fig. 4d). This indicates that while increased proliferation in vivo might be a feature shared by different filament-deficient strains, it quantitatively differs; the lower fungal burden resulting from infection with tet-NRG1 yeast compared to t-EED1 yeast is likely one factor contributing to the difference in virulence between these two strains. In addition to metabolic fitness, resistance to host defense mechanisms determines the extent to which C. albicans can proliferate in vivo. While C. albicans locked in the yeast form by constitutive expression of NRG1 (tet-NRG1) were avirulent in immunocompetent mice in this and previous studies 14,74 , the same strain was capable of inducing lethal infection in mice rendered leukopenic by combined treatment with cyclophosphamide and cortisone acetate 75 . In these mice, the yeast-locked strain reached a significantly higher fungal burden than the corresponding filamentous strain 75 , which was also higher than observed in immunocompetent animals in this study and by others 14 . Furthermore, in the same study the fungal load observed in mice with different types of immunosuppression and infected with the filamentous strain was comparable at the humane endpoint, even though this was reached at different time points after infection 75 . Similarly, MacCallum and Odds 76 observed comparable kidney burden in mice at the humane endpoint even if animals were challenged with different infectious doses and survived for a different duration. Together, this suggests that a certain number of fungal cells can be tolerated within the kidney, and that this threshold is higher for yeast than for hyphae.
Comparison of tet-NRG1 and EED1-deficient strains furthermore indicates that, despite a shared morphology, different yeast strains display substantial differences in the ability to proliferate in vivo in the presence of functional innate immunity.
To our knowledge, this is the first study showing that increased metabolic fitness of C. albicans not only contributes to virulence in hematogenously disseminated candidiasis, but that enhanced proliferation of yeast cells can result in pathogenesis and mortality indistinguishable from infection with hyphae-forming WT and t-EED1− strains. Of note, the eed1Δ/Δ mutant caused 100% mortality in a systemic infection model in immunosuppressed mice, although delayed compared to the WT. In the absence of hypha-mediated damage and overt immunopathology, this indicates that rapid proliferation resulting in high organ fungal loads might be sufficient to drive pathogenesis. Previous studies have reported similar results: tet-NRG1 yeast, which are avirulent in immunocompetent mice ( Supplementary Fig. 4 14 ), can cause lethal infection in immunosuppressed mice 75 . Whether this was due to increased fungal burden was not determined. The relevance of fungal load for pathogenesis is furthermore supported by the fact that the course of disease and rate of mortality in mice with systemic candidiasis is highly dependent on the initial infection dose 76,77 . Furthermore, by adjusting the infectious dose of different strains Odds et al. achieved a comparable mean survival time of mice infected either with C. albicans SC5314 or an isolate (RV4688) that displayed less filamentation in the murine kidney. Interestingly, less filamentation of RV4688 also coincided with a higher fungal burden compared to SC5314 in this study 77 .
Considering that immunosuppression is a major risk factor for the development of candidemia 78 the findings in this study might provide some explanation for the virulence of non-albicans Candida species, such as C. glabrata and C. auris, that do not form true hyphae 79,80 . Interestingly, the yeast form is the virulent cell type in most pathogenic dimorphic fungi, such as Histoplasma spp. and Blastomyces spp., that grow as mycelia in the environment but switch to yeast upon entering the host 81 . While the pathogenicity factors and pathogenesis mechanisms employed by these yeast cells differ significantly from C. albicans yeast, this underscores the general concept that yeast cells are not per se less virulent than hyphae. Fig. 9 Systemic infection with the C. albicans eed1Δ/Δ mutant leads to delayed mortality of immunosuppressed mice despite higher fungal burden accompanied by increased kidney injury compared to mice infected with the WT. Mice that were rendered immunosuppressed by depletion of neutrophils and monocytes using the RB6-8C5 antibody were intravenously infected with 1 × 10 2 , 1 × 10 3 or 1 × 10 4 CFU/g body weight of C. albicans WT (SC5314) or eed1Δ/Δ mutant. a Survival of mice was monitored after infection with 1 × 10 2 CFU/g or 1 × 10 3 CFU/g body weight for 21 days. Survival of mice infected with 1 × 10 2 CFU/g (n = 6 per group) or 1 × 10 3 CFU/g body weight (n = 8 per group) is shown as Kaplan-Meyer curve and curves were compared using the two-sided Log-rank (Mantel-Cox) test. p-values are shown in the graph. b Representative images of PAS stained histological cross sections of kidneys from moribund immunocompetent or immunosuppressed mice infected with 10 4 CFU/g body weight of WT or eed1Δ/Δ mutant. Immunocompetent and immunosuppressed uninfected control mice were sacrificed 7 d after mock infection. Black arrows point towards purple stained C. albicans hyphae (WT) or yeast (eed1Δ/Δ mutant), white arrows towards immune cells infiltrating the renal tissue. Scale bar represents 20 µm and applies to all images. c Organ fungal burden of immunosuppressed mice infected with 1 × 10 2 CFU/g body weight 12 h and 18 h post infection and when moribund (mor). d Quantification of kidney injury by measuring and normalizing urinary KIM-1 to urinary creatinine level. c, d Two independent experiments, n = 6, except for eed1Δ/Δ mutant 12 h, n = 4; controls, n = 4. Data are shown as median with interquartile range and were compared using the two-sided Mann-Whitney test. p-values are shown in the graph. d p-values above bars represent significant changes in comparison to the uninfected control. Source data are provided as a Source Data file.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-24095-8 ARTICLE Methods C. albicans strains, strain construction and growth conditions. The t-EED1 strain was generated in the THE1 82 background. Therefore, the URA3-tetracyclineregulable (TR) promoter region was amplified from p99CAU1 82 using primers Eed1-TET-F and Eed1-TET-R (Supplementary Table 1). Fragments were used to replace the endogenous promotor of one allele of EED1. The second allele of EED1 was deleted using the SAT1 flipping method 83 with plasmids already containing the EED1-flanking regions used to generate the homozygous eed1Δ/Δ deletion mutant M1315 7 . Transformants were selected on YPD with 200 mg/ml nourseothricin 83 and were verified by PCR and Southern Blot analysis.
The C. albicans clinical isolate SC5314 84 , the isogenic eed1Δ/Δ mutant 7 , t-EED1 and the respective parental strain THE1-CIp10 18 were maintained as glycerol stocks and grown on YPD (1% yeast extract, 2% peptone, 2% glucose) agar plates. Single colonies were inoculated into liquid YPD and grown overnight at 30°C with horizontal shaking at 180 rpm. When needed, C. albicans cells were grown to exponential phase by diluting liquid cultures to an optical density at 600 nm (OD 600 ) of 0.2, followed by incubation at 30°C and 180 rpm for 3-4 h. Cultures were washed twice with phosphate buffered saline (PBS) prior to experiments.
For infection experiments, fresh C. albicans colonies were inoculated into liquid YPD and grown to late exponential phase (14-16 h) at 30°C with horizontal shaking at 180 rpm. Doxycycline hyclate (50 µg/ml; Sigma Aldrich) was added to t-EED1 cultures to prevent hypha formation. Cells were washed twice with sterile PBS and resuspended in Dulbecco's Phosphate Buffered Saline (DPBS, Gibco), counted using a hemocytometer and adjusted to the desired concentrations in DPBS. Infectious doses were confirmed by serial dilutions and plating on YPD agar plates.
Morphology of tetracycline-regulable strains in vitro. To investigate fungal morphology of the tet-regulable t-EED1 strain in vitro, THE1-CIp10 and t-EED1 were streaked on YPD plates in the absence or presence of 50 µg/ml doxycycline. Plates were incubated under non-hypha-inducing conditions at 25°C for 2 days before pictures were taken from single colonies with an inverse microscope (Axio Vert.A1; Zeiss). To test for morphology under hypha-inducing conditions, 5 × 10 4 cells of the strains were seeded in RPMI1640 in the absence or presence of 50 µg/ml doxycycline per well in 12-well plates and incubated at 37°C and 5% CO 2 . Pictures were taken after various time points by inverse microscopy.
Growth curves. To evaluate growth in the presence of different carbon sources SC5314 and the eed1Δ/Δ mutant were diluted to OD 600 of 0.1 in YPD medium with 2 or 0.1% glucose or in SD minimal medium (0.67% yeast nitrogen base; BD Biosciences) in the presence of 2% glucose, 0.1% glucose, 2% N-acetyl-glucosamine (GlcNAc; Sigma-Aldrich), 2% sodium-DL-lactate (Sigma Aldrich), 2% potassium acetate (Merck), 2% citric acid monohydrate (Roth) or 2% casamino acids (BD Biosciences). For testing of proteolytic activity, C. albicans strains were grown in SD medium overnight, washed twice with PBS and OD 600 was set to 0.1 in YCB-BSA (1.17% Yeast Carbon Base (BD Biosciences), 1% glucose, 0.5% BSA (Serva)), pH 4.0. For growth in kidney homogenates, kidneys of uninfected mice were removed aseptically, homogenized and diluted to 25 mg/ml in DPBS and filtered through 70 µm and 40 µm cell strainers before filter sterilization. Growth was recorded by measuring OD 600 in a microplate reader at 37°C. Measurements were performed in 30 min intervals over a course of 60 h. Blank values were subtracted from all measurements.
Sample preparation for RNA isolation, RNA sequencing and analysis of data. For RNAseq, overnight cultures of C. albicans WT SC5314 and eed1Δ/Δ mutant were grown in SD with 2% glucose at 30°C and 180 rpm (Supplementary Fig. 13a). To synchronize cultures cells were inoculated at an OD 600 of 0.1 in SD with 2% glucose and grown at 30°C and horizontal shaking at 180 rpm. After 10 h samples for the 0 h time point were taken and cells were transferred to SD medium containing 2% of citrate (Roth) or 2% of casamino acids (CAA; BD Bacto) as sole carbon source at an OD 600 of 0.2 in individual flasks for each time point. Cells were grown at 37°C and 180 rpm. After 2 h, 6 h and 12 h samples were removed from the cultures and cell pellets for RNA isolation were obtained by centrifugation at 20,000 × g for 3 min and were immediately frozen in liquid nitrogen. Additional samples were taken at these and intermediate time points to determine the optical density at 600 nm, dry mass and morphology of fungal cells. For the determination of dry mass, nylon Whatman® membrane filters with a pore size of 0.2 µm were dried at 55°C for 24 h in a hybridization oven, weight and placed on a bottle top vacuum filter. Cells were loaded on the membrane by filtration and washed with ddH 2 O. After drying for additional 24 h, membranes were weighed again and dry mass was calculated. Experiments were conducted in triplicates and RNA isolation was performed as previously described 7 . In brief, pellets were resuspended in 400 µl AE-buffer (50 mM sodium acetate, 10 mM EDTA) and 40 µl 10% SDS. An equal volume of phenol/chloroforme/isoamylalcohol was added followed by incubation at 65°C for 5 min. Homogenous solutions were frozen at −80°C for 10 min, transferred to 65°C for 5 min. Freezing and thawing was repeated once. Solutions were centrifuged for 10 min at 20,800 × g and the upper phase was transferred into a new reaction tube. 10% volume 3 M sodium acetate (pH 5.3) and 1 volume 2propanol were added. Precipitation of RNA was carried out for 30 min at −20°C.
After centrifugation for 10 min at 12,000g the supernatant was discarded and RNA pellets were washed twice with 70% ethanol. RNA was solved in RNase free water. RNA quantity was determined with a Nanodrop ND1000 (Peqlab) and quality was assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies). Library preparation and RNA sequencing was carried out at Novogene (UK) Company Limited. Sequencing was performed using an Illumina NovaSeq 6000 system to obtain 150 bp paired-end reads.
Mapping of the fastq files delivered by the company (raw data are accessible at NCBI under BioProject accession number PRJNA714826, https://www.ncbi.nlm. nih.gov/bioproject/PRJNA714826) and counting of the gene transcription reads was performed using the European Galaxy server 85 that is providing an environment and sets of tools for the following analysis steps: First, FastQC was used to assess the quality of the sequences and Cutadapt was applied. Sequences were mapped to the genome of C. albicans WT (SC5314) using the RNA-Star tool, the "length of the genomic sequence around annotated junctions" parameter was set to 149. From that analysis, bam files were constructed that were used to quantify gene counts with the FeatureCount function. The reverse stranded bam files were processed allowing for fragment counts but not multimapping. The minimum mapping quality per read was set to 10. Final count files were analyzed in R, using the Deseq2 package 86 that allows searching for differentially expressed genes (DEGs) by comparing the count tables of different conditions. For this purpose, transcription profiles of the mutant were compared to the WT at the 0 h time point, as well as for each time point in medium containing citrate or casamino acids. The principal component analysis was done using the PCA function from the Deseq2 package in R. All genes were used for the calculation. Gene Ontology (GO) term enrichment analysis was performed from significantly up-or downregulated genes (+ or -log 2 2 and adjusted p-value <0.05) in eed1Δ/Δ mutant compared to WT using the CGD GO Term Finder 87 . Based on the analysis of GO terms with REVIGO 88 graphs were created using cytoscape 89 . Venn diagrams were created using Venny2.1 (https://bioinfogp.cnb.csic.es/tools/venny/).
Intraperitoneal and systemic infection model. For intraperitoneal infection and survival analyses after intravenous infection with THE1-CIp10, t-EED1 and tet-NRG1, mice received drinking water containing 5% sucrose without (−) or with (+) 2 mg/ml doxycycline starting 3 days prior to infection. Water was replaced every two days. The low acceptance of doxycycline containing water resulted in a loss of body weight in the doxycycline group only, indicating possible dehydration. This could have aggravated the consequences of impaired renal function caused by systemic infection and likely explains while mice infected with THE1-CIp10 reached the humane endpoints earlier if they received doxycycline (Fig. 2c). Therefore, in further experiments mice received a diet containing 625 mg/kg doxycycline (Envigo Teklad, catalog no. TD.120769). Food was replaced every 5 days; acceptance was high, resulting in body weights comparable to the nondoxycycline groups. For the intraperitoneal infection model, mice were infected intraperitoneally with 1 × 10 8 CFU in 500 µl DPBS and mice were humanely sacrificed 24 h post infection. For hematogenously disseminated candidiasis mice were infected with 1 × 10 2 to 1.25 × 10 5 CFU/g body weight in 100 µl DPBS via the lateral tail vein at day 0. For survival experiments, mice were euthanized when showing signs of severe illness (details described below), and these animals are referred to as "moribund". Groups of mice (n = 5) were sacrificed 6, 24, 48 and 72 h post infection with 2.5 × 10 4 CFU/g body weight for analysis of fungal burden, immune cell infiltration, renal cytokines, serum and urinary marker protein progression during the acute phase of infection. Uninfected control mice received (ctr + ) or did not receive doxycycline (ctr−) containing food. Time point experiments were repeated two to three times.
Murine model of gastrointestinal colonization and dissemination. To avoid environmental contamination, cages, bedding, bottles and drinking water were sterilized prior to use and mice were handled exclusively in laminar flow hoods. Mice received sucrose-containing drinking water and sucrose concentration was increased from 2.5% starting 7 d prior to infection for 2 d to 5% for 1 d. From day −4 until the end of the experiment mice received antibiotics to reduce the intestinal bacterial flora: 1500 U/ml penicillin and 2 mg/ml streptomycin were added to 7.5% sucrose-containing drinking water that was replaced daily; Mice were fed with chow containing 625 mg/kg doxycycline sterilized by irradiation. On day 0, mice were inoculated by gavage with 100 µl DPBS containing 5 × 10 7 C. albicans WT (SC5314) or eed1Δ/Δ. Mice were divided in two groups: one group was only colonized whereas in the other group dissemination was induced by injecting 200 mg/kg body weight cyclophosphamide (Endoxan, Baxter) intraperitoneally on day 7 and 10 post infection. Successful depletion of immune cells was confirmed by white blood cell differential count using the hematology analyzer. Feces were collected from individual mice on day 2, 4, 7 and 14, weighed and plated on YPD agar with or without 80 µg/ml chloramphenicol for determination of fungal and bacterial CFUs, respectively. On the end of the experiment (14 days p.i.) mice were humanely sacrificed and fungal burden were determined in liver and kidney.
Determination of serum and urinary biomarkers. Blood was collected by cardiac puncture (intraperitoneal infection and survival experiment) or via the vena cava inferior from mice euthanized at defined time points. Serum enzyme levels of pancreatic amylase and alanine aminotransaminase (ALT) were measured using the EuroLyser CCA 180 Vet system (QinLAB Diagnostik) according to standard methods recommended by the International Federation of Clinical Chemistry. The Mouse TREM-1 ELISA Kit (RayBiotech), DetectX® Urea Nitrogen (BUN) Detection Kit (Arbor Assays), Mouse Lipocalin-2 (NGAL) ELISA Kit (RayBiotech) and DetectX® Serum Creatinine Kit (Arbor Assays) were used to measure the respective parameters in serum of mice. Urine of mice was collected from mice euthanized 24, 48 and 72 h p.i. from moribund and uninfected mice. Either spontaneous urine was collected or gentle trans-abdominal pressure was applied onto the bladder and urine was collected using untreated glass capillary tubes. Urinary KIM-1 and creatinine levels were measured using the Mouse TIM-1 ELISA Kit (RayBiotech) and the Creatinine Parameter Assay Kit (R&D Systems), respectively. KIM-1 levels were normalized to urinary creatinine to account for differences in urinary concentration.
Quantification of immune cells by flow cytometry. To evaluate immune cell infiltration during the course of infection, organs were perfused with normal saline after withdrawal of blood. Organs were removed and weighed. One half of each kidney was cut into small pieces and digested in the presence of collagenase D (30 µg/ml; Sigma Aldrich) and DNase I (0.7 mg/ml; Sigma Aldrich) in RPMI (RPMI 1640; Gibco) supplemented with 10% fetal bovine serum (FBS; Bio&SELL), Pen Strep (100 U/ml Penicillin and 100 µg/ml Streptomycin; Life Technologies), and 1 mM sodium pyruvate (Gibco) for 30 min at 37°C with moderate horizontal shaking (70 rpm). Single cells were obtained by passing the digested tissue through a 70 µm cell strainer. Cells were washed and erythrocytes were lysed by addition of red blood cell lysis buffer (0.15 mM NH 4 Cl, 10 mM KHCO 3 , 1 mM Na 2 EDTA, pH 7.2). Remaining cells were washed, resuspended in 70% Percoll (GE Healthcare) and layered under 30% Percoll. Leukocytes were enriched by density gradient centrifugation (400g, 20 min, room temperature (RT), acceleration 1, deceleration 0). Leukocytes were collected from the interphase, washed with PBS and volumes were determined. Cells were transferred in a 96-well plate. Leukocytes were stained for flow cytometric analysis and acquired on a FACSVerse (BD Biosciences).  145-2C11, eBioscience). Fc receptors were blocked by addition of anti-mouse CD16/32 (93; BioLegend) 1:50 to the staining mixture. Dead cells were excluded from analysis using the Fixable Viability Dye eFluor® 506 (eBioscience) prior to specific antibody staining. A detailed description of the gating strategy is provided in Supplementary Fig. 19. Data were analyzed using FlowJo V.10.0.8 software.
Histopathology and immunohistochemistry. For histology, longitudinal sections of kidneys were fixed with buffered formalin and embedded in paraffin, cut into 3-4 µm slices, and stained with periodic acid-Schiff (PAS) staining according to standard protocols. Apoptotic cells in the kidney were detected by immunohistochemistry using the ApopTag in situ apoptosis detection kit (EMD Millipore) following the manufacturer's directions. Briefly, paraffin-embedded sections were rehydrated in Histo-Clear II (National Diagnostics) and alcohols followed by washing with phosphate-buffered saline (PBS). The sections were pre-treated with 20 μg/ml Proteinase K (Ambion) in PBS for 15 min at RT. Endogenous peroxidases were blocked by incubating slides for 15 min in 3% hydrogen peroxide. Sections were incubated with equilibration buffer (EMD Millipore) for 30 s at RT, followed by terminal deoxynucleotidyl transferase (TdT; EMD Millipore) incubation at 37°C for 1 h. Sections were further exposed to anti-Digoxigenin (EMD Millipore, Cat Number S7100) for 30 min at RT, and the positive reaction was visualized with DAB 3, 3-diaminobenzidine (DAB) substrate (Thermo Scientific). After counterstaining the specimens with 0.5% methyl green (Sigma), they were imaged by bright field microscopy. For quantification, apoptotic areas were quantified using PROGRES GRYPHAX ® software (Jenoptik).
Neutrophil isolation from bone marrow and differentiation of bone marrowderived macrophages. Bone marrow was obtained from 8-20 week old female BALB/c mice as described previously 91 . Briefly, mice were euthanized by cervical dislocation and femora, tibiae and humeri were removed and placed in RPMI supplemented with Pen Strep. Bone marrow was flushed with supplemented RPMI and single cell suspensions were obtained by continuous pipetting. Bone residues were removed by filtration through a 40 µm pore-size filter. Cells were pelleted and erythrocytes were lysed by addition of RBC lysis buffer. Cells were resuspended in Hanks' balanced saline solution without Ca and Mg (HBSS -; Lonza). Mature neutrophils were purified using a discontinuous Percoll gradient consisting of 52%, 69%, and 78% Percoll in HBSS -. Mature neutrophils were recovered from the 69%/ 78% interphase after centrifugation (1500 g, 4°C, 30 min, acceleration 2, deceleration 2), washed and resuspended in HBSS -. Neutrophils were counted using the hematology analyzer. Purity of neutrophils was confirmed by flow cytometry to be between 89 -95%. For differentiation into macrophages, bone marrow cells were seeded at a density of 5 × 10 6 cells in 175 cm 2 cell culture flasks in RPMI containing 10% heat-inactivated (h.i.; 30 min at 56°C) FBS, Pen Strep and 40 ng/ml recombinant murine M-CSF (ImmunoTools). Cells were incubated in a humidified incubator at 37°C with 5% CO 2 and medium was exchanged every 2-3 days. After 7 days, adherent cells were detached in RPMI + FBS by scrapping. Viable cells were counted using trypan blue exclusion and diluted to desired concentrations. For phagocytosis assays 5 × 10 5 neutrophils or macrophages were allowed to adhere to sterile coverslips in a 24-well plate for 1-2 h at 37°C, 5% CO 2 in a humidified incubator. To increase the adherence of neutrophils, coverslips were pre-treated with 0.1% gelatin and incubated at 4°C overnight. Wells were washed twice with PBS before seeding. For cytokine measurement, survival and damage assays 8 × 10 4 neutrophils or macrophages were seeded in 96-well plates in RPMI supplemented with 1% mouse serum. Cells were allowed to adhere to the substrate by culturing them for 1-2 h at 37°C, 5% CO 2 in a humidified incubator prior to infection.
Phagocytosis, survival and damage assays. To quantify phagocytosis, cells were infected with C. albicans at a multiplicity of infection (MOI) of 1 in the presence of 1% murine serum in a total volume of 500 µl. After 1 h of co-incubation at 37°C with 5% CO 2 , cells were fixed with 2% paraformaldehyde. Extracellular C. albicans cells were stained with Alexa Fluor 647-conjugated Concanavalin A (Thermo Fisher Scientific) for 30 min, intra-and extracellular fungal cells were stained with Calcofluor White (Sigma-Aldrich) after permeabilization of immune cells with 0.5% Triton X-100. Coverslips were mounted with ProLong Gold antifade reagent (Thermo Fisher Scientific) and fluorescence images were recorded using the Axio Observer.Z1 (Carl Zeiss Microscopy). The phagocytic index was determined by counting the numbers of C. albicans cells phagocytosed by 100 immune cells.
Fungal survival in the presence of immune cells was determined by infecting macrophages or neutrophils with C. albicans (MOI1) in the presence of 1% murine serum in a total volume of 150 µl. After 2 or 6 h, immune cells were lysed by addition of 50 µl 5% Triton X-100. Fungal cells were resuspended by rigorously pipetting and lysates were diluted and plated onto YPD plates and incubated for 48 h at 37°C. Survival rates were calculated by normalization from control wells containing no immune cells and the increase in fungal CFU was calculated by normalization to the starting inoculum for cells in the presence or absence of BMDMs. Fungal morphology was recorded by inverse microscopy using an Axio Vert.A1 microscope (Zeiss) after various time points. To quantify damage and TNF-α, BMDMs were co-incubated with C. albicans (MOI1) for 24 h. For total LDH release (high control), BMDMs were lysed by addition of 20 µl 5% Triton X-100, incubated for 10 min at 37°C. Supernatants were obtained by centrifugation at 300 x g for 10 min. LDH was quantified using the Cytotoxicity Detection Kit (Roche) and TNF-α was quantified by ELISA (Ready-SET-Go, eBioscience) according to manufacturer's instructions.
Cytokine and ROS production. Macrophages and neutrophils were infected with living or heat-killed (HK; 70°C, 10 min) C. albicans WT cells (MOI1) in a total volume of 200 µl. Unstimulated immune cells and cells treated with 100 nM phorbol 12-myristate 13-acetate (PMA; Sigma Aldrich) or 100 ng/ml lipopolysaccharide (LPS; Sigma Aldrich) served as negative and positive controls, respectively. After co-incubation for 24 h at 37°C with 5% CO 2 supernatants were recovered after centrifugation (1500 × g, 4°C, 15 min) and TNF-α, IL-6 and IL-10 were determined by commercially available ELISA kits (Invitrogen) according to manufacturer's instructions.
Total ROS accumulation by neutrophils was quantified by luminol-enhanced chemiluminescence. Therefore, 5 × 10 4 freshly isolated neutrophils were seeded into white clear-bottom 96-well plates (Corning) in RPMI without phenol red (Gibco). Cells were allowed to attach for 30 min at 37°C and 5% CO 2 prior to infection. Neutrophils were infected with C. albicans (MOI1) left untreated or were stimulated with PMA as positive control. Immediately after stimulation, 50 µl of RPMI without phenol red containing 200 mM luminol (Fluka) and 16 U horseradish peroxidase (Sigma Aldrich) were added. Luminescence was recorded every 2.5 min for 190 min at 37°C in a Tecan Infinite microplate reader. The area under the curve was calculated with GraphPad Prism 7.
Epithelial cell infection. The following human epithelial cell lines were used in this study: Hepatic epithelial cells (HepaRG; Gibco) were maintained in William's Medium E with GlutaMAX and HepaRG Thaw, Plate & General Purpose Medium Supplement; renal epithelial cells (A498; DSMZ) were cultivated in Minimum Essential Medium with L-glutamine supplemented with 10% h.i. FBS; oral epithelial cells (TR146; Episkin) were cultivated in Dulbecco's Modified Eagle Medium (DMEM) with high glucose supplemented with 10% h.i. FBS; intestinal epithelial cells (Caco-2 clone type C2BBe1; ATCC®CRL-2102 TM ) were maintained in DMEM supplemented with 10% h.i. FBS and 10 µg/ml human holotransferrin (Merck Millipore). Cells were cultured in a humidified incubator at 37°C with 5% CO 2 under normoxic conditions (21% O 2 ) if not stated otherwise. In addition, C2BBe1 cells were cultivated under hypoxic conditions (1% O 2 ) in a temperature controlled Hypoxystation (H35, Don Whitley Scientific) to mimic physiological intestinal O 2 concentrations. For damage assays and the quantification of cytokines, cells were detached and 2 × 10 4 cells (TR146, A498) or 4 × 10 4 cells (HepaRG) were seeded in 96-well plates 2 d prior to infection. Cells were washed and infected with exponentially grown C. albicans strains at a MOI of 1 in a volume of 200 µl. Medium without fungal cells served as mock control. After 24 h of coincubation, supernatants were recovered after centrifugation (200 × g, 5 min). To measure epithelial integrity and translocation, Corning® Transwell® polycarbonate membrane inserts with 5 µm pore size and 6.5 mm in diameter were coated with 100 µl of 10 µg/ml collagen I for 2 h at RT before they were washed twice and placed in a 24-well plate filled with 600 µl supplemented DMEM. 2 × 10 4 C2BBe1 cells were seeded in 200 µl supplemented DMEM in inserts and cultivated for 14 d at 37°C, 5% CO 2 with 21% O 2 or 1% O 2 . Medium was replaced on day 5 and every second day thereafter. Epithelial cells were infected by adding 1 × 10 5 C. albicans cells to the upper compartment and incubated for 24 h at the conditions mentioned above. To measure epithelial integrity, the trans-epithelial electrical resistance (TEER) was quantified using a chopstick electrode connected to the Epithelial Voltohmmeter EVOM2 (WPI) before and 24 h post infection. TEER measurements from inserts containing medium only served as blank values and were subtracted from all measurements. Supernatants of the upper compartment were kept for measurement of lactate dehydrogenase (LDH). To quantify the potential of the different C. albicans strains to translocate through the C2BBe1 cell layer, 24 h after infection the lower compartment was treated with 20 U/l zymolyase (Amsbio) for 2 h at 37°C and 5% CO 2 22 . Detached fungal cells were plated on YPD agar and CFUs were counted. Epithelial cell damage was quantified by measurement of LDH in supernatants using the Cytotoxicity Detection Kit (Roche). Uninfected cells served as negative control. For total cells lysis (high control) 10 µl of 5% Triton X-100 were added. Human IL-6 and IL-8 were quantified by ELISA (Invitrogen) according to the manufacturer's instructions.
Quantitative PCR. DNA was isolated from kidneys infected with either the WT (THE1-CIp10) or t-EED1 in the presence of doxycycline. Kidneys were homogenized and centrifuged for 15 min at 1500 × g. DNA was extracted from pellets using the Yeast DNA Extraction Kit (Thermo Scientific) following manufacturer's instructions. For amplification of the C. albicans 18 S rRNA gene RDN18 the following primers were used: sense amplification primer, 5'-GGACCCAGCCGAGCCTT-3' and antisense amplification primer, 5'-AAGTAAAAGTCCTGGTTCGCCA-3' 30 . Quantitative PCR was conducted using 1 µl of template DNA and the QPCR Mix EvaGreen (Bio&SELL) on a CFX 96 Real time System (BioRad). The following condition were used for product amplification: 95°C for 15 min, 40 cycles of each 95°C for 15 s, 59°C for 15 s and 72°C for 15 s. To confirm PCR product specificity, a melting curve was generated. The resulting Ct values were plotted against the CFU determined from the homogenized tissue.
Statistical analysis. GraphPad Prism 7 was used to analyze all data sets. Shown are either the mean and standard deviation (SD) or the median and the interquartile range as indicated in the figure legends. The two-tailed student's t-test or the Mann-Whitney test was used to test for statistical significances. p-values ≤ 0.05 were considered significant, *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. Survival curves were compared using the Log-rank (Mantel-Cox) test.