Abstract

   The fungal pathogen Cryptococcus neoformans is well adapted to its host
   environment. It has several defence mechanisms to evade oxidative and
   nitrosative agents released by phagocytic host cells during infection.
   Among them, melanin production is linked to both fungal virulence and
   defence against harmful free radicals that facilitate host innate
   immunity. How C. neoformans manipulates its redox environment to
   facilitate melanin formation and virulence is unclear. Here we show
   that the antioxidant glutathione is inextricably linked to redox-active
   processes that facilitate melanin and titan cell production, as well as
   survival in macrophages and virulence in a murine model of
   cryptococcosis. Comparative metabolomics revealed that disruption of
   glutathione biosynthesis leads to accumulation of reducing and acidic
   compounds in the extracellular environment of mutant cells. Overall,
   these findings highlight the importance of redox homeostasis and
   metabolic compensation in pathogen adaptation to the host environment
   and suggest new avenues for antifungal drug development.
     __________________________________________________________________

   C. neoformans has emerged as a valuable fungal model for studying
   microbial virulence and host–pathogen interactions^[47]1–[48]4.
   Cryptococcal infections begin in the lungs and subsequently disseminate
   to other organs, including the brain, which can result in
   life-threatening meningoencephalitis^[49]4–[50]6. The fungus evades
   host immune mechanisms via several virulence factors which may have
   evolved in response to environmental predation^[51]5,[52]6. Notably,
   melanin is critical for the stress response of C. neoformans during
   infection as it can scavenge harmful free radicals and protect cells
   from oxidative bursts during phagocytosis^[53]3,[54]7. Furthermore,
   cryptococcal meningoencephalitis is associated with fungal-mediated
   conversion of catecholamines (for example, dopamine) from the central
   nervous system (CNS) into melanin^[55]2,[56]7–[57]11. Thus, mechanisms
   underlying the redox-mediated processes for melanin production are
   crucial for understanding virulence and defence strategies of
   pathogenic fungi, which are becoming more prevalent due to limited
   diagnostic and treatment capabilities and a growing at-risk
   population^[58]12,[59]13.

   Redox-active thiols, such as glutathione (GSH), are key for the
   oxidative stress response of fungi and constitute systems that depend
   on enzymatic and non-enzymatic mechanisms to evade host immunity and
   cause disease^[60]14–[61]16. Several GSH-dependent enzymes, including
   glutathione peroxidases (Gpx) and glutathione reductase (GR), are
   important for C. neoformans stress response pathways and virulence;
   others (for example, glutaredoxins) execute antioxidant functions that
   require GSH for oxidoreductase activity^[62]17–[63]19. GSH is also a
   signalling molecule for pathogens such as Listeria monocytogenes, which
   requires bacterial and host-derived GSH to activate the master
   virulence regulator PrfA^[64]14,[65]20. The importance of GSH is
   therefore not limited to its redox capabilities. This is clearly
   demonstrated by the inability of γ-Glu-Cys—a potent antioxidant and
   direct precursor of GSH—to compensate for non-antioxidant functions in
   Saccharomyces cerevisiae mutants lacking Gsh2, the terminal enzyme in
   the GSH biosynthetic pathway^[66]21. In fact, GSH is ubiquitous across
   all kingdoms of life and serves in a vast array of cellular processes;
   it is arguably the most important intracellular antioxidant for
   maintaining redox homeostasis^[67]15,[68]22. Perhaps unsurprisingly,
   GSH has also been linked to melanogenesis in animals—a process that
   relies heavily on the cellular redox environment^[69]23–[70]27.

   The involvement of GSH in the growth, survival and virulence of C.
   neoformans has not been investigated. In addition to maintaining redox
   homeostasis during infection, GSH might support the cellular conditions
   needed for fungal melanin production—a phenomenon that has not yet been
   examined in a human pathogen. We employed genetic and metabolomic
   approaches to demonstrate that GSH biosynthesis supports the
   proliferation of C. neoformans, and that mutants lacking GSH2 have
   attenuated virulence in a murine model of cryptococcosis. Loss of Gsh2
   results in an enhanced reducing state that perturbs the specific
   ‘Goldilocks’ redox conditions required for melanogenesis. Furthermore,
   we demonstrate that an altered redox environment influences titan cell
   formation, urease production and interactions with macrophages. These
   results indicate a critical role for GSH biosynthesis in regulating
   metabolic and redox-dependent factors that contribute to cryptococcal
   pathogenesis.

Results

Mutants lacking GSH2 have attenuated virulence

   First, we examined the impact of GSH2 deletion (and thus the lack of
   GSH) on virulence-related phenotypes using two independent gsh2Δ
   deletion mutants and a gsh2Δ::GSH2 complemented strain, and found that
   gsh2Δ mutants did not show defects in capsule production or
   thermotolerance ([71]Extended Data Fig. 1a–[72]c). However, the gsh2Δ
   mutants were unable to melanize on catecholamine or phenolic substrates
   ([73]Fig. 1a,[74]b, [75]Extended Data Fig. 1d). Accordingly, we
   predicted that the gsh2Δ mutant would have impaired virulence in mice
   because melanin is linked to the dissemination and neurotropism of C.
   neoformans. We found that mice inoculated with the gsh2Δ mutant had
   reduced disease symptoms and prolonged survival compared to the
   wild-type (WT) and complemented strains ([76]Fig. 1c,[77]d). Mice
   infected with the mutant also had a lower fungal burden in the blood
   and affected organs at the experimental endpoint ([78]Fig. 1d).
   Interestingly, fungal burden was particularly diminished in the brain
   tissue of mutant-infected mice, which is a source of catecholamine
   precursors used for melanin production during infection^[79]7. Despite
   multiple attempts, we were unable to isolate melanized cells from
   tissue homogenates, but confirmed that mutants retrieved from lung
   homogenates could not melanize ex vivo ([80]Extended Data Fig. 1e).

Fig. 1 |. Loss of GSH2 impairs melanin and titan cell formation and
attenuates virulence.

   Fig. 1 |
   [81]Open in a new tab

   a, Melanin from cell suspensions incubated for 48 h in l-DOPA medium.
   Absorbance (OD[490]) was measured before (left) and after (right)
   digestion ([82]Methods) and normalized to 10^7 cells per ml. CWB, cell
   wall-bound. Bars represent the mean ± s.d. of n = 3 biological
   replicates. Significance relative to WT was calculated using one-way
   ANOVA with Dunnett’s correction for multiple comparisons. b, Melanin
   production of 10^6 cells on solid l-DOPA medium after 48 h incubation
   at 30 °C. The image is representative of three independent experiments.
   c,d, Survival (c) and fungal burden (d) at time of euthanasia in groups
   of 10 mice after intranasal infection with WT (red), gsh2Δ (blue) and
   gsh2Δ::GSH2 (blue segmented or dark blue) strains. Statistical analysis
   was performed using log-rank (Mantel–Cox) and Kruskal–Wallis (with
   Dunn’s correction for multiple comparisons) tests, respectively (n = 10
   mice per group). e, Visualization of cells retrieved from murine lungs
   (left, in vivo) or cell culture (right, in vitro) with DIC microscopy.
   In vivo images represent cells from eight murine lungs per strain at
   each time point, and three independent experiments per strain (n = 50)
   for in vitro images (scale bars, 10 μm). f, Cell body diameters of
   cells from e (left, in vivo; right, in vitro). Lines represent mean
   values of n = 60 cells from eight lungs per strain per time point (in
   vivo) or n = 50 cells per strain from three independent experiments (in
   vitro). Significance relative to the WT for in vitro and in vivo
   measurements was calculated using one-way or two-way ANOVA tests,
   respectively, with Dunnett’s correction for multiple comparisons. The
   dashed line indicates the diameter threshold for cryptococcal cells
   defined as titan cells. g, Phagocytosis, PI and IPR of control or
   opsonized (+18B7) strains co-incubated with macrophages for 2–24 h.
   Bars represent mean ± s.d. of n = 6 or n = 12 independent experiments
   for phagocytosis, or PI and IPR, respectively. Significance of
   phagocytosis relative to WT was calculated using two-way ANOVA with
   Šidák’s correction for multiple comparisons. Significance of PI and IPR
   was calculated using unpaired, two-tailed Student’s t-tests.

   The onset and dissemination of fungal infection was also markedly
   delayed for gsh2Δ mutants, which may be partly attributed to an
   observed inability to form titan cells in vitro and in vivo—as titan
   cells are strongly linked to cryptococcal persistence and virulence
   ([83]Fig. 1e,[84]f and [85]Extended Data Fig. 1f,[86]g)^[87]4.
   Additionally, and compared with the WT, gsh2Δ mutants had reduced
   phagocytosis, increased intracellular proliferation (IPR) in
   bone-marrow-derived macrophages (BMDMs), and severely diminished
   progression to the brain, which was colonized in less than half of
   mutant-infected mice ([88]Fig. 1g and [89]Extended Data Fig. 1f). By
   contrast, fungal burden increased in the lungs, liver, spleen and
   kidneys of mice infected with gsh2Δ mutants 26 days post-infection,
   suggesting that gsh2Δ-infected mice likely succumbed to cryptococcal
   pneumonia and/or visceral organ damage rather than brain infection
   ([90]Extended Data Figs. 1f and [91]2a). This hypothesis is supported
   by increased lung pro-inflammatory cytokines (IFN-γ, TNF-α, IL-6)
   observed 26 days post-infection in mutant-infected mice, which were
   absent 14 days post-infection ([92]Extended Data Fig. 2b). Overall, GSH
   contributes to key aspects of cryptococcal disease including
   melanogenesis, titan cell formation, interactions with phagocytes and
   dissemination to the brain.

GSH is critical for growth upon nutrient deprivation

   Next, we characterized the impact of GSH deficiency on growth upon
   nutrient depletion, which mimics the nutrient-sequestered host
   environment. We found that gsh2Δ mutants had markedly impaired growth
   in minimal medium but no growth defect in nutrient-rich medium
   ([93]Fig. 2a,[94]b). Exogenous GSH is also important for combating
   external stressors in bacterial pathogens, and the secretion and
   extracellular accumulation of GSH in S. cerevisiae is crucial for
   survival and replication under high temperature stress^[95]28,[96]29.
   We therefore investigated whether C. neoformans secreted GSH
   extracellularly and if loss of this extracellular GSH pool impacted
   gsh2Δ mutant function. First, we quantified extracellular total GSH
   (reduced plus oxidized GSH) for strains grown in minimal medium. We
   found that the WT and gsh2Δ::GSH2 strains secreted and accumulated GSH
   extracellularly ([97]Fig. 2c). By contrast, gsh2Δ mutants had no
   detectable extracellular GSH, consistent with the inability of these
   mutants to synthesize GSH endogenously ([98]Fig. 2c).

Fig. 2 |. Gsh2 is required for growth upon nutrient depletion and WT cells
secrete extracellular GSH.

   Fig. 2 |
   [99]Open in a new tab

   a, Growth curve analysis of WT, gsh2Δ and gsh2Δ::GSH2 strains grown in
   minimal medium (YNB with amino acids plus 2% glucose) with and without
   GSH supplementation. Data points indicate mean OD[600] values ± s.d. of
   n = 3 biological replicates at each time point. The initial inoculum
   for each strain was 2 × 10^4 cells ml^−1 in minimal medium and growth
   was monitored for 72 h, with OD[600] values measured every 24 h. b,
   Spot assays of growth on solid agar medium starting at 10^6 cells with
   10-fold serial dilutions with or without GSH or GSSG supplementation at
   the indicated concentrations. Images are representative of three
   biological replicates. c, Quantification of extracellular total GSH
   (reduced and oxidized GSH) for the indicated strains normalized to
   OD[600]. Bars represent mean ± s.d. of n = 3 independent experiments.
   Significance was calculated relative to WT using one-way ANOVA with
   Dunnett’s correction for multiple comparisons. d, Schematic for the
   setup of growth curves in SM. e, Growth curve analysis of strains
   growth in SM. SM was isolated from WT, gsh2Δ and gsh2Δ::GSH2 log-phase
   cells grown in minimal medium for 16 h and was used to monitor growth
   of fresh cells for each of the strains. Points indicate mean OD[600]
   values ± s.d. of n = 3 biological replicates at each time point.

   Because WT C. neoformans secreted and accumulated GSH, we posited that
   transferring gsh2Δ mutants to WT-conditioned medium could help
   elucidate the function of extracellular GSH. Although mutants
   transferred from rich to minimal medium no longer grew ([100]Fig. 2a),
   transferring mutants into spent minimal medium (SM), conditioned only
   with WT cells, fully rescued gsh2Δ mutant growth ([101]Fig. 2d,[102]e).
   This outcome suggests that an extracellular factor (likely GSH)
   supports mutant growth in nutrient-limited conditions. Growth of gsh2Δ
   mutants was only partially restored in SM conditioned with the
   gsh2Δ::GSH2 strain ([103]Fig. 2e), which may be due to partial
   complementation and/or lower accumulation of GSH in this strain. By
   contrast, gsh2Δ mutants grown in SM conditioned with the gsh2Δ mutant
   did not recover, confirming that mutants are unable to secrete and
   accumulate extracellular GSH ([104]Fig. 2e).

   To eliminate the possibility that other components in the WT SM
   influenced mutant growth, we tested the impact of exogenous GSH on the
   gsh2Δ mutant. Indeed, mutant growth in minimal medium supplemented with
   GSH was restored to WT levels ([105]Fig. 2a,[106]b). Liquid-grown
   cultures were particularly sensitive to GSH supplementation and were
   fully rescued with concentrations as low as 50 μM GSH—a concentration
   similar to that found in WT supernatant ([107]Fig. 2c and [108]Extended
   Data Fig. 3a). Addition of the oxidized disulfide GSSG also rescued
   gsh2Δ mutant growth, though to a lesser extent than GSH ([109]Fig.
   2a,[110]b). This effect may be attributed to the delay in regenerating
   GSH via reduction of GSSG by GR^[111]30, resulting in the initial lag
   in growth for mutants supplemented with GSSG. Given the overall
   positive effect of GSH on gsh2Δ mutant growth compared to other
   compounds tested (cysteine, methionine and ascorbic acid; [112]Fig. 2
   and [113]Extended Data Fig. 3b,[114]c), we conclude that GSH is
   critical for the growth of C. neoformans and could impact survival and
   proliferation in vertebrate hosts.

Loss of GSH2 enhances antioxidant defense

   S. cerevisiae gsh2Δ mutants accumulate γ-Glu-Cys^[115]21, an
   antioxidant and direct precursor of GSH, and we hypothesized that a
   similar accumulation of upstream GSH constituents would occur in C.
   neoformans gsh2Δ mutants. Disruption of GSH biosynthesis and loss of
   extracellular GSH could also impact the external redox environment and
   response to stress. Consistent with observations in S. cerevisiae, we
   found that mutants treated with H[2]O[2] had lower ROS accumulation
   than the WT and complemented strains^[116]21 ([117]Fig. 3a,[118]b).
   These observed differences led us to investigate changes in other
   antioxidant functions that could potentially compensate for the loss of
   GSH in mutants. Indeed, gsh2Δ mutants had enhanced activity of
   superoxide dismutase and catalase antioxidant enzymes, which are
   crucial for the fungal oxidative stress response ([119]Fig. 3c). We
   further speculated that loss of the extracellular GSH pool could affect
   extracellular redox homeostasis. Such changes were of particular
   interest because melanin assembly occurs at the cell wall and is highly
   dependent on redox-active processes that drive radical
   polymerization^[120]7,[121]10. Furthermore, accumulation of l-DOPA
   (l-3,4-dihydroxyphenylalanine) in the cell supernatant can cause
   toxicity and impaired growth^[122]31—a trend observed in gsh2Δ mutants
   grown in l-DOPA medium, which was rescued with ≥50 μM GSH
   supplementation ([123]Extended Data Fig. 3d,[124]e). We therefore
   examined the antioxidant/reducing power of the mutant supernatant using
   a radical-scavenging assay with
   2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid (ABTS), which
   positively correlates reducing capacity with decolorization of a
   blue/green ABTS radical chromophore. The supernatant isolated from
   gsh2Δ mutants grown in l-DOPA medium fully quenched the ABTS radical
   suggesting a highly reductive extracellular environment consistent with
   the reduced accumulation of ROS in this strain ([125]Fig. 3d,[126]e).
   However, because l-DOPA is known to scavenge ABTS radicals, we also
   tested the scavenging potential of cells grown in minimal medium
   lacking l-DOPA. Both WT and mutant supernatant from cells grown in this
   medium were unable to quench the ABTS radical, suggesting that
   accumulated l-DOPA in the gsh2Δ mutant supernatant contributed to the
   ABTS radical scavenging effect ([127]Extended Data Fig. 3f). Because
   the gsh2Δ mutants had less ROS accumulation than the WT and higher IPR
   in BMDMs ([128]Figs. 1g and [129]3a,[130]b), we suspected that other
   mechanisms contributed to the observed changes in redox homeostasis. We
   postulated that loss of the extracellular GSH pool altered the
   composition of the mutant supernatant, likely due to accumulation
   and/or secretion of metabolites upstream of GSH—some of which (for
   example, γ-Glu-Cys and cysteine) could compensate for the antioxidant
   and/or reducing abilities of GSH. We did find modestly elevated thiol
   content in gsh2Δ mutant supernatant relative to the WT or complemented
   strains ([131]Fig. 3f), and thiols are known to interfere with
   oxidation of l-DOPA^[132]25–[133]27. Thus, GSH deficiency may induce
   dysregulation of the cellular redox environment and impair formation of
   virulence-related traits (for example, melanin and titan cell
   formation) of gsh2Δ mutants.

Fig. 3 |. Deletion of GSH2 reduces susceptibility to H[2]O[2] stress and
alters non-GSH antioxidant functions.

   Fig. 3 |
   [134]Open in a new tab

   a, 2D density plots of the indicated strains stained with 16 μM DCFDA
   (untreated, +H[2]O[2]) or without (control), and analysed via flow
   cytometry. Each measurement represents 30,000 gated single cells with
   or without 1 mM H[2]O[2] treatment 1 h prior to harvesting from minimal
   medium. Graphs represent results from three independent experiments. b,
   The percentage of DCFDA-stained (DCFDA^+) cells after H[2]O[2]
   treatment relative to untreated cells was measured for each strain from
   a. Statistical significance relative to untreated cells was calculated
   using a two-way ANOVA with Dunnett’s correction for multiple
   comparisons. Bars represent mean percentage of DCFDA^+ cells ± s.d. NS,
   not significant. c, Superoxide dismutase (Sod) and catalase (Cat)
   enzyme activity (U mg^−1 protein) of WT (red) and gsh2Δ mutant (blue)
   cell lysate normalized by protein concentration. Bars represent the
   mean ± s.d. and significance was calculated relative to WT using
   unpaired, two-tailed Student’s t-tests. d,e, ABTS antioxidant assay for
   the proportion of ABTS radical (ABTS^•, blue colouration in e) quenched
   by supernatant isolated from the strains indicated after 72 h
   incubation in l-DOPA medium. Measurements were taken every 24 h for 72
   h and were quantified as a percentage of total quenched ABTS radical
   (d). Decreased pigmentation (light blue and clear in e) indicates
   radical scavenging activity. Data points in d represent the mean ± s.d.
   from three biological replicates for each strain, and images in e
   represent three independent experiments. f, Fluorescence quantification
   of supernatant thiol content after 48 h growth in l-DOPA medium for the
   indicated strains (OD[535]). Bars represent the mean thiol
   concentration ± s.d. per 10^7 cells per ml^−1 and significance was
   calculated relative to WT using a one-way ANOVA with Dunnett’s
   correction for multiple comparisons. Data are representative of three
   biological replicates for each experiment.

Mutants lacking Gsh2 have a dysregulated metabolome

   To explain the metabolic changes incited by GSH2 deletion, including
   changes to regulatory machinery for redox control, we next compared the
   metabolomes of WT and gsh2Δ mutant cells grown under melanizing
   conditions, as melanin synthesis depends on oxidative processes to
   facilitate polymerization. In particular, we searched for changes in
   the relative abundance of reducing compounds that could explain the
   altered extracellular redox potential of gsh2Δ mutants. This included
   quantification of thiol and/or antioxidant compounds directly upstream
   of Gsh2 (including sulfur-containing compounds and other reductants)
   with reducing power. Analysis of WT and gsh2Δ mutant supernatants and
   cellular extracts via liquid chromatography–high-resolution tandem mass
   spectrometry (LC-HRMS/MS) untargeted metabolomics revealed peak areas
   of more than 1,500 distinct deconvoluted molecular features (positive
   ionization mode), of which 439 extracellular and 229 intracellular
   metabolites were notably different (fold change (FC) > 1.5 or < 0.667,
   P < 0.05) between strains ([135]Extended Data Fig. 4a). Multivariate
   heat map and principal component analysis (PCA) revealed a high degree
   of similarity between WT and mutant cell extracts ([136]Extended Data
   Fig. 4b,[137]c), but substantial differences in extracellular
   metabolites between the two strains ([138]Extended Data Fig.
   4b,[139]c). Of the 439 differentially abundant metabolites in the
   supernatant, 294 had a more-than 1.5-fold increased abundance in the
   gsh2Δ mutant relative to the WT and 80 had more-than a 50-fold
   increase. However, features with a maximum FC ≥ 50 were interpreted
   with caution, as excessively high FC values could be attributed to
   abundances below the limit of quantification in the WT strain.

   To evaluate the biological relevance of these metabolic changes, we
   performed mummichog pathway enrichment analysis (based on KEGG (Kyoto
   Encyclopedia of Genes and Genomes) pathway data) and identified notable
   differences in several key metabolic pathways including amino acid and
   secondary metabolite metabolism, as well as carbohydrate/energy
   metabolism ([140]Fig. 4a). We further annotated specific metabolites
   using a robust scoring technique ([141]Methods) and confidently
   identified more than 100 features that contributed to the observed
   pathway changes. We used these annotated metabolites for further
   analysis^[142]32, and found substantial extracellular accumulation of
   several aromatic amino acids and weak acids (for example, dicarboxylic
   acids, ketones/keto-acids and hydroxy fatty acids) in the gsh2Δ mutant
   and a depleted intracellular amino acid content ([143]Fig. 4b,[144]c).
   The mutant also had substantially depleted intracellular levels of key
   energy pathway intermediates including adenosine and adenosine
   monophosphate (AMP), l-carnitine and propionylcarnitine, pantothenate,
   α-ketoglutaric acid and β-nicotinamide adenine dinucleotide (NAD) and
   nicotinamide, as well as the urea cycle intermediates ornithine and
   argininosuccinic acid ([145]Supplementary Table 1). Consistent with the
   melanin defect and enhanced ABTS radical scavenging, gsh2Δ mutants had
   an increased relative abundance of the melanin precursor l-DOPA
   ([146]Extended Data Fig. 5), suggesting an inability to utilize l-DOPA
   for melanin production. Finally, the mutant accumulated substantial
   amounts of extracellular glucosinolates (1,4-dimethoxyglucobrassicin
   and indolylmethyl-desulfoglucosinolate), sulfur-containing compounds
   and cysteine derivatives, and other antioxidants that could account for
   some of the characterized redox changes ([147]Supplementary Table 1).
   For instance, the antioxidants caffeic acid, salvianolic acid and
   pyrogallol-2-O-glucuronide were more abundant in the gsh2Δ mutant than
   the WT ([148]Supplementary Table 1). Anti-melanogenic compounds were
   also found, including myo-inositol and hydroxytyrosol, the latter of
   which was highly abundant in the gsh2Δ mutant supernatant
   ([149]Extended Data Fig. 5). We also detected an approximately 10-fold
   increase of the cysteine derivative S-(5-histidyl)cysteine sulfoxide, a
   precursor to the potent antioxidant ovothiol which has Gpx-like
   activity but has not yet been described in C. neoformans
   ([150]Supplementary Table 1)^[151]33. Of note, the WT strain had a high
   intracellular concentration of the antioxidant ergothioneine (EGT), an
   integral redox buffer in several non-yeast fungi, cyanobacteria and
   certain Gram-positive bacteria including Mycobacterium tuberculosis
   ([152]Supplementary Table 1)^[153]34. Although EGT has not been
   described in C. neoformans, our findings further support the proposed
   interdependency of EGT synthesis on GSH and/or the enzymes involved in
   GSH biosynthesis^[154]34.

Fig. 4 |. Dysregulation of GSH biosynthesis affects cellular metabolism.

   Fig. 4 |
   [155]Open in a new tab

   a, Peak list profile of enriched metabolic pathways using mummichog
   analysis (v.2) from MetaboAnalyst 5.0. Metabolic pathways are grouped
   by S. cerevisiae KEGG pathway module and class datasets. b,
   Classification of differentially abundant metabolites (P < 0.05,
   measured using unpaired, two-tailed Student’s t-tests with
   Benjamini–Hochberg FDR corrections for multiple comparisons) between
   gsh2Δ mutant (FC < 0.667) and WT (H99; FC > 1.5) strains using
   ClassyFire batch compound classification (v.2.0,
   [156]https://cfb.fiehnlab.ucdavis.edu/). Only metabolites with high
   confidence annotations (Progenesis QI score ≥40) from either positive
   or negative ionization modes were included in the classification
   analysis. c, Relative abundances of amino acids and derivatives
   identified via LC-HRMS/MS in both the cell extract (top) and
   supernatant (bottom) fractions. The horizontal axis represents
   directional intensity of the metabolite peak abundance FC, and the
   vertical axis represents statistical significance. Dashed lines
   indicate significance thresholds for −log[10] FDR-corrected P values
   (FDR < 0.05, horizontal) and log[2] FC (FC > 1.5 or < 0.667, vertical)
   were used to determine significance, which was calculated using an
   unpaired, two-tailed Student’s t-test with Benjamini–Hochberg FDR
   corrections for multiple comparisons in MetaboAnalyst (a, c). Blue dots
   indicate metabolites with higher abundance in gsh2Δ mutants relative to
   WT; red dots indicate metabolites with higher abundance in WT relative
   to gsh2Δ mutants.

Acidification and metabolite changes impair melanization

   We further characterized melanin formation as a readily assayable
   phenotype for examining the impact of metabolite differences between WT
   and mutant cells. Disparities in the abundance of extracellular acids
   predicted pH differences, and we found that the supernatant of the
   gsh2Δ mutant culture was more acidic than that of the WT ([157]Fig.
   5a). This was true of mutant cells grown in either l-DOPA or
   l-asparagine minimal media, which demonstrates that accumulated l-DOPA
   in the mutant supernatant was not responsible for acidification
   ([158]Extended Data Fig. 6a). We also investigated urease activity
   because melanin formation is influenced by extracellular pH via the
   urease-dependent production of ammonia^[159]35, and mutant growth in
   l-DOPA medium resulted in a low extracellular pH. Consistent with an
   acidic extracellular pH, gsh2Δ mutants had impaired urease activity
   ([160]Fig. 5b). Cell wall laccase activity was also diminished in
   mutants relative to the WT, suggesting that reduced expression and/or
   improper localization of the enzyme contributed to the melanin defect,
   potentially due to pH-dependent regulation of laccase activity^[161]36
   ([162]Fig. 5c). Buffering the gsh2Δ culture medium with
   3-(N-morpholino)propanesulfonic acid (MOPS) restored melanin
   production, validating an inhibitory role for extracellular
   acidification ([163]Fig. 5d,[164]e).

Fig. 5 |. Extracellular acidity and reducing pontential inhibits melanin
formation.

   Fig. 5 |
   [165]Open in a new tab

   a, pH values of supernatant isolates from cells grown in l-DOPA medium
   and normalized to 10^7 cells ml^−1. b, Spot assays of 10^6 cells ml^−1
   on solid agar urea broth following incubation at 30 °C for 24 h (top).
   The diffusion radius of ammonia into the medium (pink colouration,
   measured in the bottom panel) indicates urease activity of cells. For a
   and b (bottom), statistical significance was calculated using unpaired,
   two-tailed Student’s t-tests and bars show mean ± s.d. c, Laccase
   activity of the cell wall fraction of cell lysate measured by
   conversion of the ABTS substrate, which can be detected at OD[734].
   Bars represent mean OD[734] values ± s.d. of n = 3 biological
   replicates for each strain. d, Melanin formation of cells grown in
   liquid l-DOPA medium with or without 1 M MOPS pH-buffering. e,
   Extracellular (left) and CWB (right) melanin were quantified by
   measuring absorbance of cell supernatant and digests at OD[490]
   ([166]Methods) and normalizing by CFU to 10^7 cells ml^−1. Significance
   in e was calculated using two-way ANOVA with Tukey’s correction for
   multiple comparisons. Bars represent mean OD[490] values ± s.d. of n =
   3 biological replicates. f,g, Melanin formation of cultures grown in
   liquid (f) or on solid agar (g) l-DOPA medium for 72 h with or without
   supplementation as indicated. f, Extracellular (top panel) and CWB
   (bottom) melanin of cells grown with or without 5 mM GSH, 2.5 mM GSSG,
   or 5 mM γ-Glu-Cys supplementation and quantified by absorbance of cell
   supernatant and digests at OD[490] ([167]Methods). Bars represent mean
   OD[490] values ± s.d. of n = 3 biological replicates for each strain
   and condition. Significance was calculated relative to the l-DOPA
   condition for each strain using two-way ANOVA with Dunnett’s correction
   for multiple comparisons. g, Spot assays of 10^6 cells plated on solid
   l-DOPA medium with or without supplementation (GSH (5 mM), GSSG (2.5
   mM), ascorbic acid (AA, 10 mM), and 5 mM or 10 mM cysteine (+Cys or
   ++Cys, respectively)). Images representative of three independent
   experiments.

   Metabolomic analyses also predicted changes in antioxidants, which
   could inhibit redox-dependent radical polymerization for
   melanization^[168]7. Such changes corroborated the potent reducing
   power of the gsh2Δ mutant supernatant ([169]Fig. 3d). We tested other
   GSH pathway intermediates and ascorbic acid, a well-documented
   antioxidant, to determine whether the melanin defect of gsh2Δ mutants
   was specific to loss of GSH. GSH, cysteine and ascorbic acid had
   comparable radical scavenging activities to the gsh2Δ mutant
   supernatant ([170]Extended Data Fig. 6b,[171]c), and high
   concentrations of these compounds blocked melanogenesis of all strains
   including the WT and complement ([172]Fig. 5f,[173]g). We also found
   that the GSH pathway intermediate γ-Glu-Cys inhibited melanin
   production in all strains ([174]Fig. 5f), commensurate with the ability
   of this compound to compensate for antioxidant functions in gsh2Δ
   mutants of S. cerevisiae^[175]21. This intermediate also fully quenched
   the ABTS radical chromophore at high concentrations but was less potent
   than GSH ([176]Extended Data Fig. 6c). We note that γ-Glu-Cys was not
   detected via LC-HRMS/MS, perhaps due to chemical conjugation or use in
   alternate biochemical pathways. For instance, γ-Glu-Cys can directly
   interact with and neutralize ROS and serves as a cofactor for
   antioxidant enzymes^[177]37. By contrast, GSSG fully restored melanin
   production in gsh2Δ mutants and even enhanced melanin production of the
   WT and complemented strains ([178]Fig. 5f,[179]g). Because gsh2Δ
   mutants retain GR activity, we questioned whether trace amounts of GSH
   derived from GSSG were responsible for restoring melanin production in
   GSSG-treated mutants. We performed a GSH titration of mutant cells
   grown in l-DOPA medium, and found that low levels of GSH (50–500 μM)
   restored extracellular melanin in gsh2Δ mutants and further enhanced
   melanin in the WT ([180]Fig. 6a,[181]b). Notably, the GSH
   concentrations required to restore melanin production in gsh2Δ mutants
   were similar to the concentrations in the WT supernatant ([182]Fig.
   2c). More intriguingly, extracellular melanin production was blocked in
   both the mutant and WT when treated with GSH concentrations of 750 μM
   or higher ([183]Fig. 6a,[184]b), suggesting a ‘tipping point’ for
   redox-dependent melanogenesis. GSH at all concentrations impaired
   melanin deposition at the cell wall in the WT and at high
   concentrations in the gsh2Δ mutant ([185]Fig. 6a). However, GSSG
   addition (which is oxidizing) and MOPS buffering restored gsh2Δ mutant
   cell wall-bound melanin—an effect not seen with low-concentration GSH
   treatment ([186]Figs. 5d–[187]g and [188]6a).

Fig. 6 |. GSH modulates redox homeostasis to influence melanin production.

   Fig. 6 |
   [189]Open in a new tab

   a, Melanin formation of cells grown in liquid l-DOPA medium with or
   without GSH supplementation at the indicated concentrations.
   Extracellular (a, left) and CWB (a, right) melanin were quantified by
   measuring absorbance of cell supernatant and digests at OD[490] (see
   [190]Methods) and normalizing by CFU to 10^7 cells ml^−1. Significance
   was calculated relative to cells grown without GSH supplementation
   using one-way ANOVA with Dunnett’s correction for multiple comparisons.
   Bars represent mean OD[490] values ± s.d. from n = 3 biological
   replicates per strain for each condition. b, Images representing three
   independent experiments of extracellular melanin from WT and gsh2Δ
   mutant cells grown in l-DOPA medium with or without GSH
   supplementation. c, Schematic for the setup of melanization assays in
   l-DOPA SM. d, WT and gsh2Δ mutant cells melanized for 72 h on
   agar-supplemented with l-DOPA or l-DOPA SM isolated from log-phase
   cells grown in l-DOPA medium. Images from the upper panel show
   individual wells of a 48-well plate and the lower panel shows
   enhanced-zoom images of spots from the upper panel and represent three
   independent experiments. e, Extracellular and CWB melanin of WT, gsh2Δ
   and gsh2Δ::GSH2 cells grown for 48 h in SM isolated from log-phase
   cells grown in l-DOPA medium. Melanin content was quantified after
   normalizing by CFU to 10^7 cells ml^−1 ([191]Methods). Bars indicate
   mean OD[490] values ± s.d. of n = 3 biological replicates for each
   condition. Significance for each condition was calculated using two-way
   ANOVA with a two-stage setup method of Benjamini, Krieger and Yekutieli
   FDR correction for multiple comparisons.

   Since supernatant isolated from the WT had a higher pH, contained GSH
   at low concentrations, and restored gsh2Δ mutant growth, we reasoned
   that l-DOPA medium conditioned with WT cells could restore gsh2Δ
   melanin production. We tested this hypothesis and found that mutants
   transferred to WT SM had increased extracellular melanin and modestly
   elevated cell wall melanin ([192]Fig. 6d,[193]e). By contrast, WT cells
   transferred to gsh2Δ SM had less released melanin (P = 0.077) and less
   cell-wall-bound melanin ([194]Fig. 6d,[195]e). Overall, we conclude
   that changes to extracellular pH and antioxidant composition in gsh2Δ
   mutants lead to a redox environment that precludes melanin formation.

Discussion

   GSH is a potent antioxidant and a key player in redox homeostasis for
   many organisms^[196]1,[197]38. For C. neoformans. a pathogen of global
   health importance^[198]12,[199]13, we found that manipulation of GSH
   metabolism by deletion of the GSH2 gene attenuated both proliferation
   in host tissue and dissemination to the brain in a murine model of
   cryptococcosis. The diminished virulence of gsh2Δ mutants, which are
   non-melanized and have reduced uptake by phagocytes, affirms the
   critical role for melanin in dissemination via survival in macrophages
   and traversal of the blood–brain barrier^[200]39. Melanin is well
   established as an antioxidant that protects cells against oxidative
   bursts in the phagosome of phagocytic cells and promotes dissemination
   to the brain^[201]3,[202]7. The gsh2Δ mutant showed enhanced
   proliferation in phagocytes, consistent with greater extracellular
   antioxidant activity observed in our analysis. This phenotype may not
   be sufficient to overcome impaired dissemination to the brain, but may
   support persistence in the lung that contributes to the delayed onset
   of disease. The gsh2Δ mutant also failed to produce titan cells in
   vitro and during infection, and this phenotype has a complex influence
   on dissemination^[203]40. A previous study found that higher
   proportions of titan cells in the mouse lung were associated with
   greater dissemination^[204]41. In contrast, Reuwsaat et al. found that
   a regulatory mutant that overproduced titan cells had poor
   proliferation and reduced dissemination^[205]39. Titan cells can also
   induce Th2 immunity, which is non-protective and skews adaptive
   immunity to favour cryptococcal dissemination to the CNS, thereby
   enhancing disease^[206]4,[207]40. We hypothesize that this contribution
   of titan cells is particularly relevant for the gsh2Δ mutant because we
   observed an increased abundance of the Th2 cytokine IL-4 in the lungs
   of WT-infected mice at the time of death, but not in mutant-infected
   mice. Conversely, the surge of Th1-associated cytokines (IFN-γ and
   TNF-α) in gsh2Δ-infected mice at the experimental endpoint suggests
   enhanced inflammation that may damage lungs and exacerbate
   disease^[208]42. These findings suggest that changes to redox
   homeostasis due to loss of GSH2 impacted virulence both through
   influencing interactions with phagocytic cells and host adaptive
   immunity. Consistent with our study, GSH influences virulence in
   several fungal and bacterial species and plays a critical role in
   mitigating the impact of host-derived ROS released during
   infection^[209]15,[210]16,[211]20,[212]28,[213]41.

   Our global metabolomic analysis revealed that loss of GSH dramatically
   altered the metabolome and resulted in dysregulation of the cellular
   redox environment—a key determinant of survival in the host^[214]43.
   For example, the mutant supernatant had elevated thiol content and
   substantial quantities of glucosinolates and phenolic glycosides with
   antioxidant activity^[215]44–[216]46. Of note, we found a twofold
   increase in pyroglutamic acid, a γ-glutamyl cycle intermediate formed
   from the GSH precursor γ-Glu-Cys. This finding is intriguing because
   glutathione synthetase deficiency, a heritable disorder in humans, also
   leads to hyperaccumulation of pyroglutamic acid resulting in high anion
   gap metabolic acidosis—a form of acidosis defined by acid accumulation
   and low pH of the blood^[217]47,[218]48. Consistent with this
   phenotype, our mutants accumulated and secreted high quantities of
   amino acids and other weak acids (for example, lactic acid,
   dicarboxylic acids and keto acids) leading to lower extracellular pH
   and urease activity compared to the WT. Mutant supernatant fractions
   also contained several amino acid derivatives with known antioxidant
   and anti-melanogenic properties including hydroxytyrosol^[219]49. We
   propose that acidification of mutant supernatant (along with
   accumulation of antioxidants) precludes the chemical oxidation of the
   catecholamine precursor l-DOPA needed for melanization^[220]2,[221]9,
   as this process is supported by an alkaline pH^[222]35,[223]50,[224]51
   and was reversible by MOPS-buffering of gsh2Δ mutant media. Consistent
   with this mechanism, treatment with exogenous reducing compounds (for
   example, cysteine, GSH, ascorbic acid and γ-Glu-Cys) also inhibited
   melanization.

   In summary, we have shown that perturbation of the cellular redox
   environment via loss of GSH upsets the ‘Goldilocks’ conditions needed
   for melanin formation, as well as other virulence traits including
   titan cell formation, urease activity and dissemination to the brain.
   These findings are consistent with the roles of GSH across fungal and
   bacterial species^[225]20,[226]52 and provide a framework for
   understanding how microorganisms not only mitigate harmful radicals
   released by host phagocytic cells but also manipulate the extracellular
   metabolite environment to condition virulence. These findings
   collectively support GSH metabolism as a potential therapeutic target
   for cryptococcosis, and we note that glutathione metabolism has been
   targeted in human diseases and cancer^[227]53,[228]54 but not for
   antifungal drug discovery.

Methods

Ethics statement

   All animal experiments performed with mice in this study were conducted
   in accordance with the guidelines of the Canadian Council on Animal
   Care and approved by the University of British Columbia’s Committee on
   Animal Care (protocol A21–0105).

Strains and growth media

   A detailed list of strains and oligonucleotides used in this study are
   provided in [229]Supplementary Tables 2 and [230]3, respectively. C.
   neoformans var. grubii (serotype A) strain H99 was used for generation
   of gsh2Δ deletion mutants and as the WT strain ([231]Supplementary
   Table 2). WT and mutant strains were maintained on yeast peptone
   dextrose medium (YPD, Difco; 1% yeast extract, 2% peptone, 2% dextrose,
   2% agar). Cells for growth and phenotypic assays were grown overnight
   at 30 °C (with shaking at 220 r.p.m.) in liquid YPD and were then
   either harvested or transferred to experiment-specific media for
   further experimentation. Yeast nitrogen base minimal medium (YNB,
   Difco; with amino acids, supplemented with 2% glucose, pH 5.6), which
   lacks GSH, was used as minimal medium. Solid YPD medium containing 200
   μg ml^−1 hygromycin was used for selection of gsh2Δ deletion mutants,
   and medium containing 100 μg ml^−1 nourseothricin was used to select
   for the gsh2Δ::GSH2 reconstituted strain in the gsh2Δ mutant
   background.

Strain construction

   The gsh2Δ mutants were constructed with deletion cassette prepared via
   three-step overlapping PCR using primers listed in [232]Supplementary
   Table 3 with WT genomic DNA and the pJAF15 plasmid as templates. The
   resulting construct was biolistically transformed into the WT strain as
   described previously^[233]55, and positive transformants were selected
   on hygromycin (200 μg ml^−1) and confirmed via PCR and Southern blot
   analysis. The gsh2Δ::GSH2 complemented strain was generated by
   insertion of a gene encoding a C-terminal Gsh2-HA fusion protein at the
   native locus. GSH2-HA was constructed using three-step overlapping PCR
   with WT and Dnj1-HA genomic DNA, as described previously^[234]56. The
   resulting construct was biolistically transformed into the gsh2Δ mutant
   background and positive transformants were selected on nourseothricin
   (100 μg ml^−1). Primers used for the complement construct are listed in
   [235]Supplementary Table 3.

Murine infection and virulence assays

   Virulence of WT, gsh2Δ mutant and gsh2Δ::GSH2 strains was tested in a
   murine inhalation model of cryptococcosis using 4–6 week-old female
   BALB/c mice (Mus musculus) from Charles River Laboratories. Mice were
   housed in a facility with 14 h light/10 h dark cycles, 40–60% humidity
   and an ambient temperature range of 21–23 °C. Feed (PicoLab Rodent Diet
   20, LabDiet, 0007688) and water were provided ad libitum for the
   duration of the experiments. Briefly, fungal cells were grown in YPD at
   30 °C overnight, washed twice with phosphate-buffered saline (PBS) and
   resuspended in PBS. For survival and endpoint fungal burden, groups of
   10 BALB/c mice were intranasally inoculated with a suspension of 2 ×
   10^5 cells in 50 μl for each strain and mice reaching the humane
   endpoint were euthanized by CO[2] anoxia. For time-course measurements
   of fungal burden, groups of eight BALB/c mice were intranasally
   inoculated with a suspension of 2 × 10^5 cells in 30 μl for each strain
   and timepoint. Mice infected for 7, 14, 21 or 26 days were euthanized
   by CO[2] anoxia. Mouse organs were harvested by dissection at humane or
   pre-determined endpoints, weighed and homogenized in sterile PBS and
   plated on YPD agar containing chloramphenicol (100 μg ml^−1) to
   quantify colony-forming units (CFUs) for each organ. No statistical
   methods were used to pre-determine sample sizes, but sample sizes were
   similar to those reported in previous
   publications^[236]17–[237]19,[238]35. To enhance experimental rigour
   and minimize bias for data collection, mice from different litters were
   mixed and randomly assigned to one of three experimental groups
   (WT-infected, gsh2Δ-infected or gsh2Δ::GSH2-infected). Cages from each
   group were chosen at random for euthanasia at pre-determined
   experimental endpoints. The health status of mice for all experiments
   was monitored daily post-inoculation.

Microscopy of C. neoformans and lung tissue sections

   For in vitro titan cell assays, cells were grown on Sabouraud dextrose
   agar (SDA; 4% glucose, 1% peptone, 1.5% agar) for 2–5 days at room
   temperature (~20–22 °C), as described previously^[239]57. After
   incubation, 10^7 cells were resuspended in a T25 flask with 10 ml YPD
   and cultured for 22 h at 30 °C with shaking at 150 r.p.m. until
   stationary phase (approximately 2 × 10^8 cells ml^−1). Cells were
   washed twice with minimal medium, adjusted to 10^6 cells ml^−1, and
   incubated at 30 °C for 48 h with shaking at 800 r.p.m. In vivo titan
   cell formation was evaluated using WT and gsh2Δ cells isolated from
   murine lung homogenate and washed twice with minimal medium. For
   polysaccharide capsule images, cells were incubated for 48 h in
   capsule-inducing medium (CIM) prepared as previously
   described^[240]19,[241]58. All cells were stained with India ink prior
   to imaging. Titan cell and polysaccharide capsule formation were
   evaluated by differential interference contrast (DIC) microscopy using
   a Zeiss Axiopolan 2 microscope equipped with a Plan-Apochromat
   100×/1.46 objective lens and an ORCA-Flash4.0 LT CMOS camera (Hamamatsu
   Photonics). Zeiss Zen 2 Blue edition (v.2.3) and ImageJ^[242]59
   (v.2.14.0) were used for collection and analysis of microscopy images,
   respectively. Cells with a body size >10 μm were considered titan
   cells, as described previously^[243]57.

   Histological samples were obtained from lung tissue of four infected
   mice for each strain and timepoint. Lung tissue was stored in 10%
   formalin and submitted to Wax-It Histology Services for processing.
   Resin-embedded tissues were sectioned, mounted and stained with
   hematoxylin and eosin prior to imaging.

In vitro macrophage infections with C. neoformans

   BMDMs were isolated from BALB/c mice and differentiated at 37 °C with
   5% CO[2] for 7 days in media consisting of DMEM, 10% FBS, 1%
   nonessential amino acids, 1% penicillin-streptomycin, 1% GlutaMAX
   (Gibco, 35050061), 1% HEPES buffer, 20% L-929 cell-conditioned
   supernatant and 0.1% 2-mercaptoethanol^[244]35. Differentiated cells
   were detached from culture dishes using a cell scraper (Falcon, 353085)
   and washed into DMEM. Cells were seeded at a density of 2 × 10^5 cells
   per well in 24-well plates and incubated at 37 °C in 5% CO[2] for 24 h
   in DMEM supplemented with 100 U ml^−1 of IFN-γ (Gibco, PMC4031) to
   activate. Overnight cultures of WT and gsh2Δ C. neoformans grown in YPD
   were washed twice with PBS, counted and resuspended in DMEM at a
   density of 2 × 10^7 cells ml^−1. Fungal cells were opsonized by
   incubation with 5 μg ml^−1 18B7 monoclonal antibody (mAb) for 1 h at
   room temperature. BMDMs were then infected with 2 × 10^6 cells ml^−1 of
   opsonized or non-opsonized (−18B7, control) C. neoformans and incubated
   for 2–24 h at 37 °C. After infection, BMDMs were washed twice with PBS
   to remove extracellular yeast cells and lysed with 1 ml sterile water.
   Serial dilutions of BMDM lysate were plated on YPD and grown for 2–3
   days at 30 °C. Colonies were then counted to determine CFUs.

   Phagocytosis of control and opsonized cells was calculated as the
   number of BMDMs containing cryptococcal cells after 2 h incubation
   divided by the total number of BMDMs. Phagocytic index (PI) was
   determined by dividing the CFUs from lysed cells 2 h post-infection
   (hpi) by the total number of BMDMs. IPR was calculated by dividing the
   CFUs at 24 hpi by the CFUs at 2 hpi.

Lung cytokine isolation and analysis

   Lung cytokines were quantified from the supernatant of thawed organ
   homogenates (stored at −80 °C) from each experimental timepoint using
   the BD Mouse Th1/Th2/Th17 Cytometric Bead Array (CBA) Kit (BD
   Biosciences, 560485). Briefly, supernatants were incubated for 2 h at
   room temperature in equal proportions with anti-cytokine mAb-coated
   beads and a cytokine phycoerythrin (PE) detection reagent. After
   incubation, samples were washed once and measured on a Northern Lights
   full spectrum flow cytometer (Cytek Biosciences) with SpectroFlo
   software (v.3.2.1). Data were analysed using FlowJo (BD Biosciences,
   v.10.10) with the CBA Plug-in (v.5.2.2).

Serial dilution spot assays and growth curves

   In a typical growth experiment, single colonies of the WT, gsh2Δ
   mutants or gsh2Δ::GSH2 strains were selected from an agar plate and
   incubated at 30 °C in YPD medium overnight with shaking at 220 r.p.m.
   Cells were harvested and washed twice with sterile water. For spot
   assays, 10-fold serial dilutions were prepared starting at 10^6 cells
   and 5 μl of cell suspensions were spotted onto solid agar medium.
   Plates were then incubated at 30 °C, 37 °C or 39 °C for 2–4 days and
   imaged. For growth curves, overnight cultures were first counted to
   determine CFUs. Cells were then diluted in the appropriate medium and
   grown in either a 96-well plate or 5 ml liquid culture at 30 °C for 72
   h with shaking at 220 r.p.m. For YPD and YNB medium growth curves,
   absorbance at an optical density of 600 nm (OD[600]) was measured every
   24 h using a Tecan Infinite M200 PRO microplate reader with Tecan
   i-control (v.1.7.1.12). For growth in l-DOPA medium, cells were
   inoculated in chemically defined l-DOPA medium containing 0.1%
   l-asparagine, 0.1% dextrose, 3 mg ml^−1 KH[2]PO[4], 0.25 mg ml^−1
   MgSO[4]·7H[2]O, 1 μg ml^−1 thiamine, 5 ng ml^−1 biotin and 0.2 mg ml^−1
   l-DOPA (Sigma-Aldrich, D9628) and counted every 24 h to determine CFUs.

Quantification of melanin production in liquid culture

   Cells from overnight cultures were harvested and counted to determine
   CFUs. Each strain was then diluted and transferred at 1 × 10^6 cells
   ml^−1 to l-DOPA medium with or without supplementation at the
   concentrations indicated. Cultures were grown for 48 h at 30 °C with
   shaking at 220 r.p.m., then harvested and counted to determine CFUs.
   Absorbance of the supernatant was measured at OD[490] using a Tecan
   Infinite M200 PRO microplate reader to determine extracellular melanin
   content. The remaining cell pellet was washed twice with water, diluted
   to 10^7 cells ml^−1, and digested in 100 μl of 1 M NaOH with 10%
   dimethyl sulfoxide (DMSO) for 1 h at 95 °C. Cell digests were
   centrifuged and the OD[490] was measured to determine cell wall melanin
   content. Extracellular melanin content was normalized to growth as
   determined by CFUs for each strain.

Media transfer and quantification of extracellular GSH

   Following overnight incubation, WT, gsh2Δ mutant and gsh2Δ::GSH2 cells
   were washed twice with sterile water, resuspended and transferred into
   25 ml YPD broth at an OD[600] of 0.3 for an 8 h growth period.
   Log-phase cells were then harvested, counted to determine CFUs and
   diluted in 5 ml of either YNB or l-DOPA medium at concentrations of 3 ×
   10^4 or 1 × 10^6 cells ml^−1, respectively. After 16 h of growth, cells
   were harvested via centrifugation and the resulting supernatant was
   harvested and sterilized using 0.22 μm syringe filter sterilizing units
   (VWR, 76479–016) to remove residual cells. For YNB medium transfer
   experiments, fresh overnight cells of each strain were then inoculated
   into the filtered spent medium (SM) and aliquoted into a 96-well plate
   at an OD[600] of 0.0001 and grown for 72 h at 30 °C with shaking at 220
   r.p.m. and OD[600] measurements taken every 24 h. For liquid l-DOPA
   medium transfer experiments, fresh cells were aliquoted into 3 ml spent
   l-DOPA medium at an initial density of 1 × 10^6 cells ml^−1 and grown
   for 48 h (with shaking at 220 r.p.m.) and then harvested and counted to
   determine CFUs. Melanin content of liquid cultures was then quantified
   (as described above). In parallel, fresh cells of each strain were
   grown in YNB or l-DOPA media to compare cell growth in fresh versus
   filtered SM. For solid medium experiments, spent l-DOPA medium was
   supplemented with 2% sterile agar and dispensed into a 48-well plate.
   After the medium solidified, fresh cells were spotted onto solid SM at
   1 × 10^6 cells ml^−1 and incubated for 48 h at 30 °C prior to imaging.

   Extracellular GSH was quantified by centrifugation and filter
   sterilization (VWR, 76479–016) of culture supernatants to remove cells.
   The concentration of GSH was measured using a glutathione assay kit
   according to the manufacturer’s instructions (Cayman Chemical, 703002).
   Absorbance at OD[415] was measured using a Tecan Infinite M200 PRO
   microplate reader and GSH concentration was determined by constructing
   a standard curve using GSH solutions provided by the manufacturer.

Measurement of ROS content using flow cytometry

   Cells were grown for 24 h in liquid YPD medium, washed twice with water
   and the absorbance at OD[600] was measured. Resuspended cells were
   inoculated in fresh YNB (pH 5.6) at an OD[600] of 0.3 and grown for 16
   h with shaking at 30 °C. Following a 16 h incubation period, cells were
   harvested via centrifugation, washed twice using sterile PBS and
   resuspended in PBS at a cell density of OD[600] = 1.0. Cells exposed to
   oxidative stress were treated with 1 mM H[2]O[2] for 1 h with shaking
   at 30 °C. After treatment with H[2]O[2], the treated cells were
   centrifuged, washed twice and resuspended in PBS. Cells were then
   treated with 16 μM of the ROS-sensitive fluorogenic probe
   2′,7′-dichlorodihydrofluorescein diacetate (DCFDA, Sigma-Aldrich) for 1
   h in the dark at 30 °C. Stained cells were then washed once with PBS
   and ROS levels were analysed using a CytoFLEX S flow cytometer (Beckman
   Coulter) with lasers at wavelengths of 405 nm (violet), 488 nm (blue),
   561 nm (yellow) and 633 nm (red). Results were gated to single yeast
   cells and PBS was used as a blank control. Fluorescence of DCFDA was
   measured using the FITC-GFP channel, and data were acquired and
   analysed using the CytExpert cytometry analysis software (Beckman
   Coulter).

ABTS free-radical scavenging assays

   Preparation of the radical ABTS solution was performed as previously
   described^[245]60. To examine the radical scavenging activity in WT and
   gsh2Δ mutant culture supernatants, cells were first grown at 1 × 10^6
   cells ml^−1 in l-DOPA medium, centrifuged at 20,627g and 1 ml of the
   supernatant was collected and dried overnight at room temperature via
   rotary evaporation. Samples were then resuspended in 200 μl of nanopure
   water and 10 μl of each sample was transferred to a 96-well plate. For
   testing specific chemicals, 10 μl of each compound was serially diluted
   to the indicated concentrations. Then 190 μl of radical ABTS solution
   was added to each sample, as described previously^[246]60. Following a
   5 min incubation in the dark, absorbance at OD[734] was measured using
   a Tecan Infinite M200 PRO microplate reader to quantify decolourization
   and values were normalized to 10^7 CFUs per strain.

Thiol quantification and antioxidant enzyme activity

   Thiol concentration of culture supernatants was measured using a
   fluorometric thiol assay kit according to the manufacturer’s
   instructions (Sigma-Aldrich, MAK151). Fluorescence was measured using a
   BioTek Synergy H1 microplate reader (Agilent) with BioTek Gen5 software
   (v.3.15.15) at an excitation wavelength of 490 nm and an emission
   wavelength of 525 nm. Thiol concentration was determined by
   constructing a standard curve using serial dilutions of a GSH standard
   provided by the manufacturer and normalized by CFU.

   The activities of superoxide dismutase and catalase were measured using
   enzyme-specific colorimetric assay kits according to the manufacturer’s
   instructions (Cayman Chemical, 706002 and 707002, respectively).
   Absorbance was measured at 440 nm (superoxide dismutase) or 540 nm
   (catalase) using a Tecan Infinite M200 PRO microplate reader, and
   enzyme activity was normalized to protein concentration as determined
   by Pierce BCA protein quantification (Thermo Fisher Scientific, 23225)
   relative to a BSA standard.

Measurement of urease and cell wall laccase activity

   Overnight cultures were harvested and washed twice with PBS at 4,000g
   for 4 min. For urease assays, cells were diluted 1:40 in PBS and 25 μl
   of diluted culture was spotted onto urease detection agar medium (2%
   urea, 1.5% agar, 0.2% monopotassium phosphate, 0.1% peptone, 0.1%
   dextrose, 0.5% sodium chloride, 0.0012% pH 6.8 phenol red). Plates were
   incubated for 24 h at 30 °C and imaged. Pink halos indicative of urease
   activity were measured using ImageJ software^[247]59. For cell wall
   laccase activity, washed cultures were added 1:100 in 100 ml of
   l-asparagine medium (0.1% l-asparagine, 0.1% dextrose, 3 mg ml^−1
   KH[2]PO[4], 0.25 mg ml^−1 MgSO[4]·7H[2]O, 1 mg ml^−1 thiamine, 5 ng
   ml^−1 biotin) and grown for 48 h, shaking at 120 r.p.m. Following
   incubation, cultures were harvested and washed twice in PBS at 14,000g
   for 15 min. Cells were then resuspended in 10 ml of cold PBS and
   incubated with cOmplete mini EDTA-free protease inhibitor (Roche,
   11836170001) for 30 min at 4 °C. Cells were kept on ice and lysed using
   a French pressure cell press (Glen Mills) with five passes at 25,000
   psi. The resulting lysate was then centrifuged at 3,000g for 10 min and
   the supernatant was removed. The resulting cell wall pellet was washed
   twice in PBS at 3,000g for 10 min and resuspended in 10 ml PBS. Then,
   180 μl aliquots of resuspended cell pellet were added to a 96-well
   plate in triplicate and 20 μl of 10 mM ABTS was added to each sample.
   Plates were incubated for 2 h at 30 °C with constant shaking.
   Absorbance at 734 nm was measured after 2 h using a SpectraMax iD5
   spectrophotometer with SoftMax Pro software (v.7.1).

Cell preparation for metabolome profiling

   Samples for metabolite analysis were incubated for 48 h in l-DOPA
   medium starting at 1 × 10^6 cells ml^−1. WT and gsh2Δ mutant cells were
   then harvested via centrifugation at 15,493g for 10 min at 4 °C and
   separated into supernatant and cell pellet fractions. Both fractions
   were kept on ice. Next, 2.5 volumes of 100% MeOH + 0.05% butylated
   hydroxytoluene (BHT; Cayman Chemical) were added to the supernatant
   fraction to limit oxidation and/or isomerization of reactive
   compounds^[248]61. The resulting mixture was incubated for 1 h at −20
   °C to precipitate residual salts from the chemically defined medium.
   Supernatant fractions were then centrifuged (20,627g for 10 min, 4 °C)
   and a 1 ml aliquot was removed and stored on ice. Two-phase extraction
   was then performed as described previously, with slight modifications,
   to isolate intracellular contents of fungal cells^[249]62. First, the
   remaining cell pellets were washed twice with pre-cooled nanopure water
   and 150 μl 100% MeOH + 0.05% BHT was added. The resulting solution was
   mixed by vortexing and then shaken gently for 2 min at 4 °C. Pre-cooled
   nanopure water (200 μl) was added with ~150 μl glass beads. Samples
   were homogenized using a Mini-BeadBeater-8 (BioSpec) for three rounds
   of 20 s each, with 30 s resting on ice between each round. A 500 μl
   aliquot of methyl-tert-butyl-ether (MTBE) was then added to each sample
   followed by vortexing to mix and gentle shaking for 10 min at 4 °C.
   After centrifugation (20,627g for 10 min, 4 °C) to induce phase
   separation, the aqueous layer was extracted and placed in a separate
   microfuge tube. Both the aqueous cell extract and cell supernatant were
   then dried using a rotary evaporator overnight at room temperature.
   Fractions were stored at −80 °C until LC-HRMS/MS analysis.

Liquid chromatography–mass spectrometry

   For untargeted metabolomics analysis, samples were analysed using an
   Impact II high-resolution mass spectrometer (Bruker Daltonics) coupled
   with an Elute UHPLC system (Bruker). Separation of compounds was
   achieved using a multigradient method on an Inertsil Ph-3 UHPLC column
   (2 μm, 150 × 2.1 mm; GL Sciences) equipped with a Ph-3 guard column (2
   μm, 2.1 × 10 mm). The mobile phase consisted of water (A) with 0.1% v/v
   formic acid and methanol (B) with 0.1% v/v formic acid. The separation
   was conducted using a multi-step gradient ranging from 5% to 99% mobile
   phase B over 18 min as follows: 0 min 5% B; 0–1 min 5% B; 1–8 min 35%
   B; 8–10.5 min 99% B; 10.5–14 min 99% B; 14–14.5 min 5% B; 14.5–18 min
   5% B. The column temperature was set to 55 °C, while the autosampler
   was maintained at 4 °C, and the flow rate was 0.3 ml min^−1.

   Aliquots of 100 μl from each sample were pooled to generate a quality
   control sample (QC) used for evaluating instrument performance. A QC
   sample was injected every six samples. Data-dependent acquisitions were
   conducted in positive (ESI+) and negative (ESI−) electrospray
   ionization modes to obtain precursor and fragment ion information for
   annotating compounds. For ESI+, the mass spectrometer settings were as
   follows: capillary voltage of 4,500 V, nebulizer gas pressure of 2.0
   bar, dry gas flow rate of 10 l min^−1, dry gas temperature of 220 °C,
   mass scan range of 50–1300 m/z, spectra acquisition rate of 10 Hz and
   cycle time of 0.4 s. Collision energy of 20 V was ramped through each
   MS/MS scan from 100 to 250%. For ESI−, the capillary voltage was set at
   −3,500 V. The mass spectrometer was calibrated with sodium formate at
   the beginning of each run to ensure accuracy. Average mass error in
   annotation was below 2.0 ppm (ref. [250]32).

Data processing and annotation of metabolites

   Raw data processing was performed using Progenesis QI software
   (v.3.0.7600.27622) with METLIN plugin v.1.0.7642.33805 (NonLinear
   Dynamics) and entailed peak picking, alignment, normalization, and
   database searching^[251]63. Annotations were performed as previously
   described^[252]63,[253]64. However, to increase confidence in
   annotations, only experimental MS/MS data were used for querying and
   matching against the in-house library (Mass Spectrometry Metabolite
   Library of Standards, MSMLS, supplied by IROA Technologies), METLIN and
   MassBank of North America^[254]65,[255]66. A Progenesis QI score ≥40
   was considered to select a candidate for annotation in accordance with
   reporting criteria for chemical analysis suggested by the Metabolomics
   Standards Initiative (MSI)^[256]67,[257]68. Ions generated from QC
   samples were retained for annotation and included in the dataset if the
   coefficient of variation (CV) for abundance did not exceed 25%. When
   compounds were detected in both ion modes, the one with the highest
   signal-to-noise ratio was retained. Relative quantities of metabolites
   were determined by calculating their corresponding peak areas. Finally,
   putative metabolites with high confidence annotations were categorized
   using the ontology-based ClassyFire Batch Compound Classification web
   platform (v.2.0, [258]https://cfb.fiehnlab.ucdavis.edu/) and visualized
   with the ggplot2 package (v.3.5.0) in R.

   Data was normalized using a robust Progenesis QI built-in approach
   designed explicitly for untargeted
   metabolomics^[259]63,[260]64,[261]69. Progenesis QI employed median and
   means absolute deviation analysis based on all detected abundances to
   reduce bias and noise in the data^[262]70. A unique gain factor is then
   calculated for each sample (represented by a scalar multiplier αk) and
   compound ion abundances were adjusted to a normalization reference
   run^[263]69. In brief, all compounds were included to normalize the
   data, resulting in an aggregate matrix with measurements for every
   compound ion in each run. This aggregate matrix was used to generate a
   ratio for comparison of compound ion abundance in a particular run with
   the corresponding value in the normalization reference (typically a
   QC). Ratios were log transformed to create a normal distribution for
   all ratio data within each run for all samples. Finally, scalar
   estimations were employed to align log distributions with that of the
   normalization reference. The tight clustering exhibited by QC samples
   at the center of the PCA analysis indicated excellent system
   performance and stability ([264]Fig. 4c). The resulting dataset and
   parameters used in statistical analyses were made publicly
   available^[265]32.

Statistical analysis

   Statistical analyses were performed using GraphPad Prism 8 (v.8.2.1),
   unless otherwise specified. Statistical significance between two groups
   was calculated using unpaired, two-tailed Student’s t-tests;
   significance of three or more groups was calculated using one- or
   two-way analysis of variance (ANOVA) and corrected for multiple
   comparisons using the methods indicated. The threshold for statistical
   significance was defined as α = 0.05 or q = 0.05 for false discovery
   rate (FDR)-adjusted P values. Exact P values are not provided when P <
   0.0001, but information on statistical analyses can be found in figure
   legends and relevant Source Data files. For parametric tests (t-tests,
   ANOVA), data distribution was assumed to be normal, but this was not
   formally tested (individual data points are shown in figures). Sample
   sizes were determined based on those reported in previous
   publications^[266]17,[267]19,[268]35. No data were excluded from the
   analyses. The investigators were not blinded to allocation during the
   experiments and outcome assessment.

   Metabolomics data were analysed using the web-based platform
   Metaboanalyst 5.0 ([269]https://www.metaboanalyst.ca/faces/home.xhtml).
   Univariate analysis of peak intensity values was performed by volcano
   plot analysis wherein P values were determined with a two-sided
   Wilcoxon rank-sum test and then adjusted for multiple testing by FDR
   correction using an adjusted P value threshold of q = 0.05. Adjusted P
   values were plotted on a negative log[10] scale. Volcano plot FC values
   were calculated as the ratio of peak intensity means between samples
   for each feature and plotted on a log[2] scale with cut-off values of
   FC > 1.5 or < 0.667.

   Multivariate analysis was performed to identify interactions and
   changes in the overall metabolome profile of WT and gsh2Δ mutant
   strains, as well as supernatant and cellular extract fractions.
   Metabo-Analyst was used to conduct a PCA and a heatmap analysis to
   visualize distribution of samples and relative intensity of LC-HRMS/MS
   features, respectively. Metabolic pathway enrichment analysis was
   performed using MetaboAnalyst peak list profiling which utilizes the
   mummichog v.2 algorithm, based on KEGG pathway data, to quantify
   enrichment of putatively annotated peaks at the network level, as
   described previously^[270]71. The S. cerevisiae KEGG pathway data
   (accessed 9 November 2022) was used for enrichment analysis as this
   dataset most closely resembles C. neoformans pathway information of
   datasets available on the MetaboAnalyst platform. Enrichment data were
   visualized using the ggplot2 package (v.3.5.0) in R.

Extended Data

Extended Data Fig. 1 |. Loss of GSH2 does not affect capsule size but delays
virulence and impairs dissemination to the brain.

   Extended Data Fig. 1 |
   [271]Open in a new tab

   a, Polysaccharide capsule (left) and cell body diameter (right) of
   cells grown for 48 h. Measurements represent mean values ± s.d. of n =
   50 cells from three independent experiments per strain. Significance
   was calculated relative to the WT using one-way ANOVA with Dunnett’s
   correction for multiple comparisons. b, Visualization of polysaccharide
   capsule with DIC microscopy. Images represent three independent
   experiments (scale bars = 10 μm). c, Spot assay of growth on solid YPD
   medium starting at 10^6 cells ml^−1 with 10-fold serial dilutions.
   Plates were incubated at 30 °C, 37 °C, 39 °C. d, Melanin production of
   10^6 cells on epinephrine (Epi, 0.1 g L^−1), dopamine (Dm, 0.1 g L^−1),
   and Niger seed (NS)-containing minimal media at 30 °C. NS medium was
   prepared from 70 g 0.1 g L^−1 seed extract supplemented with 0.1 g L^−1
   glucose, 20 g L^−1 Bacto Agar, and 0.05% Tween 20. e, Melanin
   production of 10^6 cells per strain retrieved from murine lungs and
   spotted on solid l-DOPA medium. Agar plates in c, d, and e were grown
   for 48 h before imaging, and images are representative of three
   biological replicates. f, Time-course of fungal burden in mice
   intranasally infected with WT (red), gsh2Δ (light blue), and
   gsh2Δ::GSH2 (dark blue) strains. Significance was calculated using
   two-way ANOVA with Dunnett’s correction for multiple comparisons. Solid
   bars indicate mean fungal burden (n = 8 mice per group) and segmented
   bars represent interquartile range. g, Visualization of fungal cells
   retrieved from murine lungs with DIC microscopy (top). Images represent
   fungal cells retrieved from 8 murine lungs per strain (n = 60 cells per
   sample) per timepoint (bars = 10 μm). Cell body diameter of gsh2Δ
   mutant cells retrieved from murine lungs (bottom). Lines represent mean
   values of n = 60 cells from 8 lungs per strain per time point.

Extended Data Fig. 2 |. Mutants lacking GSH2 have reduced lung proliferation
and cell size in vivo.

   Extended Data Fig. 2 |
   [272]Open in a new tab

   a, Representative images of H & E stained lung tissue sections from
   four mice infected with WT, gsh2Δ, or gsh2Δ::GSH2 strains at each of
   the time points indicated (arrows = cryptococcal cells; bars = 10 μm).
   Segmented squares indicate enhanced zoom sections shown in the
   lower-right corner of each image. b, Cytokine levels in murine tissue
   homogenate of lungs infected with the WT (red) or gsh2Δ mutant (blue)
   strains at the indicated timepoints (n = 8 lungs for each strain at
   each time point). Uninfected mice (Naïve, gray, inoculated with PBS)
   were used as a control. Significance was calculated relative to
   treatment naïve mice using two-way ANOVA with either Tukey’s (7 & 14
   dpi) or Šidák’s (21 & 26 dpi) correction for multiple comparisons.
   Solid bars indicate mean cytokine level and segmented bars represent
   interquartile range.

Extended Data Fig. 3 |. GSH is required for growth in minimal medium and is
not restored with ascorbate or sulfur-containing amino acids.

   Extended Data Fig. 3 |
   [273]Open in a new tab

   a, Growth curve analysis of WT (light blue circles), gsh2Δ (blue
   squares or circles), and gsh2Δ::GSH2 (light blue squares) strains grown
   in rich or minimal medium with and without methionine (Met), cysteine
   (Cys), or ascorbic acid (AA) supplementation at the indicated
   concentrations. b, Spot assays on minimal media. Each strain was
   spotted starting at 10^6 cells with 10-fold serial dilutions on agar
   minimal medium with or without Met, Cys, or AA supplementation at the
   indicated concentrations. Images are representative of three biological
   replicates. c, Growth of WT, gsh2Δ, and gsh2Δ::GSH2 strains in minimal
   medium with or without GSH supplementation at the indicated
   concentrations. Data points for a and c indicate mean OD[600] values ±
   s.d., and the initial inoculum for each strain was 2 × 10^4 cells
   ml^−1. Growth was monitored for 72 h with OD[600] values measured every
   24 h. d–e, Growth of WT, gsh2Δ, and gsh2Δ::GSH2 cells in 5 ml liquid
   cultures of l-DOPA medium (with or without GSH supplementation in e).
   Initial inoculum for each strain was 10^6 cells ml^−1 and CFUs were
   counted every 24 h for 72 h (d) or at 48 h (e). Points (d) or bars (e)
   represent mean CFUs ml^−1 ± s.d. at each time point or GSH
   concentration, respectively. f, ABTS antioxidant assay (see
   [274]Methods) for the proportion of ABTS radical quenched by
   supernatant isolated from the indicated strains after 72 h incubation
   in l-asparagine minimal medium (lacking l-DOPA). Pigmentation (blue
   colouration) indicates presence of the ABTS radical. Images represent
   three independent experiments. Data are representative of three
   biological replicates for each experiment.

Extended Data Fig. 4 |. Metabolic profiling of the gsh2Δ mutant relative to
WT reveals differences in extracellular and intracellular fractions.

   Extended Data Fig. 4 |
   [275]Open in a new tab

   a, Volcano plots showing QC-normalized LC-HRMS/MS feature data and
   differentially abundant metabolites between WT and gsh2Δ mutant cells
   in supernatant (left) and cell extract (right) fractions. The
   horizontal axis represents the directional intensity of the metabolite
   peak abundance fold change (FC) and the vertical axis represents
   statistical significance. A P-value threshold of P < 0.05 and FC
   threshold of > 1.5 or < 0.667 (segmented lines) were used to identify
   differences between the WT and mutant strains, which were determined
   using an unpaired, two-tailed Student’s t-test with Benjamini-Hochberg
   FDR correction for multiple comparisons in MetaboAnalyst. b, Heatmap
   comparing relative intensity of metabolite abundances in the
   supernatant (sup) and cell extract (ext) of WT and gsh2Δ mutants (KO);
   n = 3 biological replicates were analyzed for each fragment. Higher and
   lower intensity values are coloured red and blue, respectively. c, PCA
   score plot of the first two principal components from WT and gsh2Δ
   mutant supernatant and cellular extract fractions with QC samples for
   data normalization. Each data set represents n = 3 biological
   replicates. Purple = WT; red = gsh2Δ mutant; green = QC.

Extended Data Fig. 5 |. Mutants lacking GSH2 show distinct changes in
relative abundance of key energy metabolites, antioxidants, and extracellular
acids.

   Extended Data Fig. 5 |
   [276]Open in a new tab

   Relative abundance of select compounds identified via LC-HRMS/MS
   between WT (blue) and gsh2Δ mutant (red) strains. Both cell extract
   (ext) and supernatant (sup) fractions were analyzed for relative peak
   intensity. Bars represent mean relative abundances ± s.d. for n = 3
   biological replicates per strain. Significance was calculated relative
   to the WT supernatant (sup) fraction using multiple unpaired,
   two-tailed Student’s t-tests with Benjamini-Hochberg FDR corrections
   for multiple comparisons.

Extended Data Fig. 6 |. Extracellular acidification of gsh2Δ mutants is
independent of l-DOPA and GSH pathway metabolites influence melanin
formation.

   Extended Data Fig. 6 |
   [277]Open in a new tab

   a, pH values of supernatant isolates from WT and gsh2Δ mutant cells
   grown in l-asparagine minimal medium and normalized to 10^7 cells
   ml^−1. Statistical significance was calculated using an unpaired,
   two-tailed Student’s t-test. Bars represent mean pH values ± s.d. b–c,
   ABTS antioxidant assay indicates the reducing power of compounds tested
   at the indicated concentrations. Decreased pigmentation (light blue &
   clear in image) in c indicates increased ABTS^• radical scavenging
   activity. GSH = glutathione; AA = ascorbic acid; Cys = cysteine.

Supplementary Material

   Supplementary Figures
   [278]NIHMS2063451-supplement-Supplementary_Figures.pdf^ (1.1MB, pdf)

Acknowledgements