(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



Cryptococcus neoformans adapts to host CO2 concentrations via metabolic and stress-response remodeling [1]

['Laura C. Ristow', 'Department Of Pediatrics', 'Carver College Of Medicine', 'University Of Iowa', 'Iowa City', 'Lowa', 'United States Of America', 'Emma E. Blackburn', 'Department Of Microbiology', 'University Of Georgia']

Date: 2026-05 

Cryptococcus neoformans is an environmental pathogen that remodels its cellular physiology to survive within mammals and, in susceptible hosts, cause life-threatening meningoencephalitis. Of the many distinctions between the external environment and mammalian tissues, CO 2 concentration in the host is two orders of magnitude higher than in the environment and represents a critical stress for C. neoformans. C. neoformans strains that do not replicate at host CO 2 concentrations are less virulent in mouse models of infection, further supporting CO 2 tolerance as a virulence trait. To further understand the genetic determinants of C. neoformans CO 2 tolerance, we performed a near genome-wide screen for deletion mutants with altered CO 2 fitness using a competitive growth assay. A total of 301 of 4,692 deletion mutants showed altered CO 2 tolerance (245 reduced fitness; 56 increased fitness) demonstrating the global effect of host CO 2 on C. neoformans physiology. Based on this data set as well as a metabolomic analysis of C. neoformans adaptation to host CO 2 , we show that remodeling of central carbon metabolism, oxidative stress buffering, and membrane homeostasis represent an integrated response to CO 2 stress that is mediated in part by the TOR-Ypk1 signaling axis. We propose that CO 2 -induced capsule formation leads to reduced cellular glucose which, in turn, triggers remodeling of central carbon metabolism toward utilization of alternative carbon sources and increased mitochondrial respiration/reactive oxygen generation. Thus, these data provide a near genome-wide profile of the genetic determinants of C. neoformans CO 2 tolerance as well as a model for how this important environmental human fungal pathogen alters its physiology to proliferate in the host.

The results of this screen have provided new insights into the breadth of cellular systems affected by host CO 2 as well as the multiple stress response pathways that contribute to the ability of the CO 2 -tolerant reference strain H99 to compensate to elevated, host-relevant concentrations of CO 2 . We focused our follow-up studies on the metabolic and oxidative stress pathways and processes required for CO 2 tolerance. The results of these experiments suggest that metabolic remodeling, buffering of ROS species, and membrane homeostasis represent an interconnected and integrated response to high CO 2 concentrations that is critical for C. neoformans to adapt to this host-relevant stressor.

We have been interested in identifying the genetic and physiologic determinants of C. neoformans CO 2 tolerance. In previous work, we showed that the target of rapamycin (TOR) pathway is a critical mediator of CO 2 tolerance and that, in part, this was due to the role of TOR in remodeling of phospholipid asymmetry of the plasma membrane in response to host CO 2 [ 12 ]. In addition, we found that multiple genetic loci contribute to the CO 2 fitness of tolerant strains through quantitative genetic trait analysis of backcrossed strains [ 11 ]. To gain a more complete profile of the genes that affect CO 2 tolerance, we carried out a large-scale competitive fitness screen using the C. neoformans deletion mutant libraries constructed by the Madhani lab [ 13 ].

In addition to elevated body temperature, the mammalian host environment presents additional stressors that could also represent significant fitness bottlenecks to C. neoformans; low levels of nutrients and reactive oxygen species (ROS) are two commonly cited examples [ 6 , 7 ]. Our group found that the elevated concentrations of CO 2 present in host tissues (5%) relative to the external environment (~0.04%) represent a significant stressor for C. neoformans [ 8 , 9 ]. Overall, environmental strains have reduced fitness in host concentrations of CO 2 relative to clinical isolates. Importantly, environmental strains that are CO 2 -tolerant are also able to cause disease in mouse models [ 10 ]. In addition, CO 2 -sensitive environmental strains recovered from infected mice have increased CO 2 -fitness as well as increased fitness in the host [ 11 ]. Taken together, these data strongly support the concept that tolerance of host CO 2 concentrations is a virulence trait for C. neoformans [ 9 ].

Cryptococcus neoformans is an environmental fungus that causes life-threatening infections of the central nervous system in humans [ 1 ]. People with altered immune function such as organ transplant patients and those living with HIV/AIDS are at particular risk for invasive cryptococcal infection [ 2 ]. C. neoformans has been cultured from a wide range of environmental niches throughout the world [ 3 ]. The majority of environmental C. neoformans isolates are non-pathogenic in mammalian models of infection whereas C. neoformans isolated from patients are generally more virulent [ 4 ]. One well-established reason for this distinction is that most environmental strains of C. neoformans are unable to grow at mammalian body temperature (~37 °C, [ 5 ]). Growth at human body temperature, polysaccharide capsule formation, and cell wall melanization represent three of the most widely studied C. neoformans virulence traits [ 5 ]. Despite the prominence of the “big-three” virulence traits [ 6 ], other aspects of C. neoformans biology contribute to its ability to cause disease and the identification of traits that can better distinguish pathogenic and non-pathogenic strains remains an area of strong research interest in fungal pathogenesis [ 7 ].

Results

The transition to host CO 2 concentrations affects a wide range of cellular processes in C. neoformans To identify genes with altered fitness at host concentrations of CO 2 (5%), a moderate throughput competitive fitness assay that we previously reported [12] was used to screen 4,692 strains from the Madhani C. neoformans deletion library; all deletion mutants were in the CO 2 -tolerant KN99α strain background (S1 Table). Each mutant was inoculated into the well of a 96-well plate with an approximately equal cell concentration of mNeonGreen-labeled reference strain H99. The strains were incubated in RPMI buffered with MOPS in ambient air or 5% CO 2 at 37 °C for 24 h. The ratios of the two strains were quantified using flow cytometry, and the competitive fitness calculated as described in Materials and methods. Mutants with statistically significant (three replicates, P < 0.05 Student t test) competitive fitness changes of ±0.2 were considered to have altered fitness. A total of 301 genes met these criteria with 245 mutants showing reduced fitness and 56 mutants displaying increased fitness (Fig 1A and S1 Table). We previously identified multiple kinase mutants in the Bahn Library that have altered fitness in 5% CO 2 ; importantly, a subset of these were also in the Madhani Library and showed the same fitness phenotypes [12], including mutants of MPK1, BCK1, MKK2, PKA1, ARK1, PKA1, CBK1, FBP26, YPK1, and CEX1. In addition, recently, it has been reported that plating errors in the distribution of the Madhani deletion set have occurred, leading to obvious problems with reproducibility. Our group has previously reported screening a subset of the Madhani library (obtained independently) under different conditions (yeast peptone 2% dextrose (YPD), 10% CO 2 on solid agar plates, [13]); a total of 25 mutants identified by Chadwick and colleagues were also identified in this work [13]. Finally, we have also independently constructed mutants for further study and confirmed those phenotypes. A summary of these corroboration results is provided in S2 Table. While large-scale screening experiments have inherent false positive and negative rates, our follow-up data indicate that the screening results are reasonably robust within limitations expected for a genome-wide mutant screen. PNG larger image

TIFF original image Download: Fig 1. Genome-wide deletion collection screening revealed a wide range of cellular processes involved in response to CO 2 . A. Volcano plot of the average competitive fitness score and significance of three biological replicates for each mutant in 5% CO 2 conditions relative to ambient air in the Madhani deletion collection. Competitive fitness was determined after combining a 1:1 ratio of mNeonGreen-labeled H99 and unlabeled mutant cells in RPMI 1,640 medium with 165 mM MOPS, pH 7, and incubating at 37 °C in ambient air or 5% CO 2 for 24 h. Cell populations were characterized by flow cytometry, and the percentage of mNeonGreen-negative cells in 5% CO 2 was divided by those in ambient conditions to calculate competitive fitness. Colored data points indicate values that were below 0.8 (red) or above 1.2 (blue) and statistically significant by Student t test (P < 0.05). Biological-process GO term analysis for the genes identified in (A) at 24 h are depicted in B (significantly reduce CO 2 fitness) and D (significantly increase CO 2 fitness). C. Cell lysates from H99 cells exposed to ambient air or 5% CO 2 for 2 h were analyzed by western blot for Hog1 expression and phosphorylation status. Results are representative of three biological replicates. E. Venn diagram comparing mutants with reduced CO 2 fitness (identified in A) to mutants with reduced fitness in a mouse model of cryptococcosis reported in [14]. E. Venn diagram comparing mutants with reduced fitness at 37 °C to mutants with reduced fitness in a mouse model of cryptococcosis reported in [14]. The data underlying this Figure can be found in S1 Table and S1 Data. https://doi.org/10.1371/journal.pbio.3003561.g001 Biological process GO term analysis of the set with reduced fitness provided some insight into the overall functions involved in CO 2 tolerance (Fig 1B). Response to the detergent sodium dodecyl sulfate (SDS) and salt were the two terms with the lowest adjusted p values (Benjamini-Hochberg). SDS interferes with membrane function and is consistent with our previous findings that CO 2 leads to remodeling of the plasma membrane and that drugs targeting sterols (fluconazole) and sphingolipid biosynthesis (myriocin) are more active against C. neoformans at host CO 2 concentration than in ambient air [8]. The response to salt appears to be driven by a large set of protein kinases with pleomorphic phenotypes [14] that include sensitivity to salt (CCK1, IPK1, CBK1, YPK1, SNF1, TPS2). If the host CO 2 concentrations directly triggered a stress response that overlaps with salt stress, then it may increase activity of the Hyperosmolar Glycerol Pathway (HOG) pathway, an important part of the cellular response to elevated salt [15]. Our previous screen of a protein kinase library [12], however, did not find that mutants of the HOG1 pathway were less fit in CO 2 (in the current screen the hog1∆ mutant showed modestly reduced fitness 0.78, S1 Table). Furthermore, transition to high CO 2 did not change the phosphorylation status of Hog1 (Fig 1C). Therefore, it seems sensitivity to salt is an indirect correlation with CO 2 sensitivity. Finally, genes involved in response to oxidative stress were also enriched in those required for CO 2 fitness. The set of gene mutations that increased CO 2 fitness is also enriched for response to salt along with response to pH and multivesicular body (MVB) formation (Fig 1D). The presence of the latter two groups of genes is likely related to the fact that the Rim101 pathway is maladaptive during CO 2 stress [12] and that some genes that contribute to the MVB are involved in Rim101 processing [16]. Interestingly, the set of genes required for full CO 2 fitness was enriched for those that are also annotated to affect the “big three” virulence traits [5,6] of capsule formation, melanin metabolism, and response to heat (Fig 1B). However, the set of genes whose loss of function increases fitness in 5% CO 2 was also enriched for those with effects on capsule, melaninization, and salt stress (Fig 1D). The connection between genes that affect CO 2 fitness and the “big three” virulence traits is likely due to the fact that all of these phenotypes are indirectly affected by many physiologic processes. Of these, capsule formation is the trait most likely to share physiologically direct connections to CO 2 fitness because it is a well-characterized inducer of capsule formation in vivo [17]. Finally, we compared the set of mutants with reduced CO 2 fitness to the set of mutants with reduced fitness in a mouse model of cryptococcosis as recently reported by Boucher and colleagues [14]. The Venn diagram in Fig 1E shows that 138/245 (56%) mutants with reduced CO 2 fitness in vitro also have reduced fitness during infection. We do not know what level of CO 2 fitness under these in vitro conditions is required to maintain in vivo fitness. However, we took advantage of the large-scale phenotyping data from Boucher and colleagues [14] to compare the CO 2 /in vivo fitness correlation with that between fitness at body temperature (37 °C) and in vivo fitness. The ability to grow at host body temperature is one of the three well-established C. neoformans virulence traits. Interestingly, the overlap of mutants with reduced fitness at 37 °C with in vivo fitness was remarkably similar (152/272, 56%) to that for CO 2 fitness (Fig 1F). These comparisons suggest that both CO 2 and host body temperature tolerance are virulence traits that display variable penetrance rather than virulence factors that show a binary, present/absent, distribution. Further, these traits are likely to interact with each other as well as with additional virulence traits to determine the overall in vivo fitness of a given mutant or strain.

The RIM101 and PKA pathways negatively regulate CO 2 tolerance in an independent manner The set of mutants associated with increased CO 2 fitness was enriched for those involved in response to alkaline pH (Fig 1D). In addition to re-identifying the rim101∆ mutant, deletion mutants of components of the pathway required for proteolytic processing (Rim13 and Rim20), and hence activation, of Rim101 were also found to have increased fitness in 5% CO 2 . Activation of the Rim101 pathway has been linked to Endosomal Sorting Complex Required for Transport (ESCRT) complex function in model yeasts [18]. Furthermore, the Kronstad [19] and Alspaugh [16] labs have shown that the ESCRT complex is involved in Rim101 activation in C. neoformans as well. In keeping with these observations, deletion mutants of five genes with homology to ESCRT complex components (VPS22/23/25/28&36) showed increased fitness in CO 2 . In principle, the increased fitness of mutants related to tolerance of elevated pH could be due to pH-related functions or due to other functions with pH response as a common but distinct function. Therefore, we compared the CO 2 phenotypes of the ESCRT complex mutants to non-ESCRT complex VPS genes (Fig 2A) and re-tested them on YPD plates (S1 Fig); the CO 2 phenotypes were consistent between the two conditions for all strains. All five ESCRT complex mutants were more fit in CO 2 than H99 (VPS22/23/25/28&36) while non-ESCRT VPS mutants either had reduced CO 2 fitness (VPS1 and VPS31) or no phenotype (VPS29/35/&41); the increased fitness of the vps29∆ mutant is statistically significant but not likely to be large enough to be biologically relevant. Regardless of their CO 2 tolerance phenotype, all VPS mutants except for the vps29∆ mutant showed reduced growth on YPD adjusted to pH 8 (Fig 2B). Since both CO 2 -tolerant and -intolerant VPS mutants are generally sensitive to pH 8, it appears that the CO 2 fitness phenotypes are not directly related to an impaired response to elevated pH but rather due to other functions of the ESCRT/Rim101 pathway or VPS genes. PNG larger image

TIFF original image Download: Fig 2. The RIM101 and PKA pathways negatively regulate CO 2 tolerance in an independent manner. A. Competitive fitness was determined for the indicated vacuolar protein sorting (VPS) mutants after combining a 1:1 ratio of mNeonGreen-labeled H99 and unlabeled mutant cells in RPMI 1,640 medium with 165 mM MOPS at pH 7 and incubating at 37 °C in ambient air or 5% CO 2 for 24 h. Cell populations were characterized by flow cytometry, and the percentage of mNeonGreen-negative cells in 5% CO 2 was divided by those in ambient conditions to calculate competitive fitness. Bars represent the average of three biological replicates with error bars indicating SD. Asterisks indicate a significant difference relative to H99 as determined by Student t test (P < 0.05). B. Overnight cultures of rim101∆ and vps mutants were washed, and concentration determined by OD 600 reading. Strains were diluted to a starting OD 600 :1, and serial 10-fold dilutions were performed. Three µL per replicate were spotted onto standard YPD agar plates, or YPD buffered with 150 mM HEPES and adjusted to pH 8. Plates were incubated at 37 °C in ambient air for 48 h before images were acquired. Images are representative of three biological replicates. Competitive fitness of pkr1∆ and pka1∆ mutants (C) and rim101 mutants (D) in 5% CO 2 as described in (A). The data underlying this Figure can be found in S1 Table and S1 Data. https://doi.org/10.1371/journal.pbio.3003561.g002 O’Meara and colleagues have demonstrated that some functions of the Rim101 pathway are regulated by the Protein Kinase A (PKA) pathway [20]. Previously, we showed that, like loss of function mutants in the RIM101 pathway, deletion of PKA1 increases CO 2 fitness [12]; an observation we re-confirmed in this screen (Fig 2C). The PKA pathway is negatively regulated by Pkr1 and, consistent with that relationship, the pkr1∆ mutant is less fit in CO 2 (Fig 2C). Therefore, we tested whether the PKA pathway directly regulates Rim101 during CO 2 stress. To do so, we took advantage of a strain constructed by O’Meara and colleagues in which the only source of RIM101 is an allele in which the PKA phosphosite (Ser773) was mutated to alanine (rim101Ser773Ala, [20]). The rim101Ser773Ala mutant did not show a significant difference from WT in fitness under 5% CO 2 (Fig 2D). Therefore, both the PKA and RIM101 pathways negatively affect CO 2 tolerance, but they do so in a manner independent of each other.

Regulation of membrane homeostasis through both the TOR and UPR pathways is required for CO 2 tolerance The TOR pathway is also a key positive regulator of CO 2 tolerance, and prior work has shown that it regulates remodeling of the plasma membrane asymmetry in response to elevated CO 2 concentrations [12]. Further validating the importance of the TOR pathway to CO 2 tolerance, two putative components of the TOR complex, SIN1 and a protein homologous to Avo2 (CNAG_02727), as well as SIT4, a phosphatase that appears to function downstream of TOR in S. cerevisiae [21], were among the mutants with reduced fitness (S1 Table). A key function of the Tor-Ypk1 axis in model yeast is the regulation of sphingolipid biosynthesis [22]. In S. cerevisiae, Ypk1 phosphorylates and activates the serine-palmityl transferase (SPT, [23]) which regulates the first committed step of sphingolipid biosynthesis (Fig 3A). As previously reported, the potency of the SPT inhibitor, myriocin, against C. neoformans is increased in 5% CO 2 , indicating sphingolipid synthesis is required for CO 2 tolerance [8,12]. In addition, Lee and colleagues have previously demonstrated that Ypk1 is required for sphingolipid biosynthesis in C. neoformans [23]. PNG larger image

TIFF original image Download: Fig 3. Regulation of membrane homeostasis through both the TOR and UPR pathways is required for CO 2 tolerance. A. Schematic representation of the early steps in C. neoformans sphingolipid biosynthesis pathway. Genes required for CO 2 fitness are indicated in red. Created in BioRender. Ristow, L. (2026) https://BioRender.com/jgg870j. B. Competitive fitness was determined for glucosylceramide synthase mutant gcs1∆ after combining a 1:1 ratio of mNeonGreen-labeled H99 and unlabeled mutant cells in RPMI 1,640 medium with 165 mM MOPS at pH 7 and incubating at 37 °C in ambient air or 5% CO 2 for 24 h. Cell populations were characterized by flow cytometry, and the percentage of mNeonGreen-negative cells in 5% CO 2 was divided by those in ambient conditions to calculate competitive fitness. Bars represent the average of three biological replicates with error bars indicating SD. Asterisks indicate a significant difference relative to H99 as determined by Student t test (P < 0.05). C. Cells from overnight cultures of H99 were spread on solid RPMI agar plates buffered with 165 mM MOPS to pH 7. Sterile disks were placed on plates, and 100 µg of aureobasidin A was added to each disk. Plates were incubated at 37 °C in ambient air or 5% CO 2 for 48 h before images were acquired. Images are representative of three biological replicates. D. Competitive fitness of the unfolded protein response (UPR) mutants ire1∆ and hxl1∆ in 5% CO 2 as described in (B). The data underlying this Figure can be found in S1 Table. https://doi.org/10.1371/journal.pbio.3003561.g003 Downstream of SPT (Fig 3A), the sphingolipid biosynthesis pathway has two branches: glucosylceramide (GC) and inositol-phosphoryl-ceramide (IPC). We identified the gcs1∆ mutant as having reduced CO 2 fitness, demonstrating the importance of the GC branch in CO 2 tolerance (Fig 3B). IPC synthase (Ipc1) is an essential gene [24] and, therefore, to determine if the IPC branch is also required for CO 2 tolerance, we compared the potency of the specific IPC inhibitor, aureobasidin A in ambient and 5% CO 2 [25]. Indeed, the zone of clearance induced by aureobasidin A is increased at 5% CO 2 relative to ambient air (Fig 3C). These data firmly establish that an essential function of the Tor-Ypk1 axis during CO 2 stress is to maintain sphingolipid homeostasis and that both major sphingolipid classes contribute to CO 2 tolerance. Membrane homeostasis is also critically dependent on the function of the endoplasmic reticulum (ER) for the synthesis of lipids such as sterols, for the delivery to membrane proteins, and for the generation of the cell wall which buffers the cells against the effects of membrane-directed stressors. The unfolded-protein response (UPR) pathway functions to maintain ER homeostasis in the presence of a variety of stressors that affect its function [26]. The two key proteins in this pathway are the kinase IRE1 and the downstream transcription factor HXL1 (aka BZP1). Mutants lacking either component of the UPR pathway have reduced fitness in 5% CO 2 (Fig 3D). These data contribute to the growing body of evidence that host CO 2 exerts a significant stress on the membranes of C. neoformans and, consequently, key mediators of membrane biosynthesis and membrane stress response pathways are critical for the ability of C. neoformans to replicate under those conditions.

Host CO 2 concentrations remodel C. neoformans carbon metabolism Through its conversion to bicarbonate, CO 2 is an essential substrate for many key reactions in central carbon metabolism [27]. At the same time, it is generated as a byproduct in other reactions of central carbon metabolism such as those in the tricarboxylic acid cycle in the mitochondria. Twenty-five mutants involving central carbon metabolic processes were identified as being required for CO 2 fitness (Table 1). Sixteen of these genes are associated with mitochondrial functions. In addition, the deletion mutants of two non-essential genes involved in glycolysis (FBP26 and PFK1) have reduced fitness as do five genes associated with the synthesis of acetyl-CoA or CoASH (YHM2, CAB3, LSC1, POT1, and ALD4). Lastly, multiple genes in the kynurenine pathway, which generates nicotinamide, were also less fit in 5% CO 2 . PNG larger image

TIFF original image Download: Table 1. Carbon metabolism genes required for fitness in host concentrations of CO 2 . https://doi.org/10.1371/journal.pbio.3003561.t001 To further explore the effect of host CO 2 concentrations on the metabolic state of C. neoformans, we compared the non-lipid metabolomic profile of C. neoformans incubated in buffered RPMI at ambient CO 2 to that from cells incubated in 5% CO 2 after 24 h at 37 °C. The abundance of 90 molecules differed between 5% CO 2 and ambient air (± 1 log 2 , adjust P < 0.05, S3 Table). Strikingly, glucose; two glycolysis intermediates, phosphoenopyruvate (PEP) and dihydroxyacetone-3-phosphate (DHAP); and fructose (Fig 4A) are reduced, suggesting that exposure to 5% CO 2 depletes glucose and glycolytic intermediates. PNG larger image

TIFF original image Download: Fig 4. Host CO 2 concentrations remodel C. neoformans carbon metabolism. The normalized area of each metabolite was determined by tandem mass spectrometry as described in materials and methods for H99 incubated in buffered RPMI in ambient air or 5% CO 2 . Asterisks indicate statistically significant differences in abundance at 5% CO 2 relative to ambient air (Welch corrected, two-tailed t test, P < 0.05). A. Relative amounts of glycolytic intermediates. B. Relative amounts of UDP-Glucuronic Acid (UDP-GluA). C. Competitive fitness of the uxs1∆ and uge1∆ mutants relative to H99 was determined for indicated strains after combining a 1:1 ratio of mNeonGreen-labeled H99 and unlabeled mutant cells in RPMI 1,640 medium with 165 mM MOPS at pH 7 and incubating at 37 °C in ambient air or 5% CO 2 for 24 h. Cell populations were characterized by flow cytometry and the percentage of mNeonGreen-negative cells in 5% CO 2 was divided by those in ambient conditions to calculate competitive fitness. Bars represent the average of three biological replicates with error bars indicating SD. Asterisks indicate a significant difference relative to H99 as determined by Students t test (P < 0.05). D. H99 was incubated in buffered RPMI medium with standard glucose (0.2%) or the indicated amount of glucose supplementation at 37 °C at ambient or 5% CO 2 . The optical density of the cultures was measured at 48 h. Growth was normalized to that of standard RPMI for each CO 2 concentration. The bars indicate mean of two independent replicates, and error bars indicate standard deviation. The differences between conditions were analyzed by One-way ANOVA with post hoc correction for multiple comparisons. The data underlying this Figure can be found in S1 and S3 Tables and S1 Data. https://doi.org/10.1371/journal.pbio.3003561.g004 Host CO 2 is a strong inducer of the polysaccharide capsule in C. neoformans [17], and we have previously shown that genes involved in capsule biosynthesis are enriched in the set of genes upregulated by exposure of H99 to host levels of CO 2 [12]. The carbohydrates that make up the capsule are derived from glucose, suggesting that host CO 2 may divert glucose from glycolysis to capsule biosynthesis [28]. Indeed, the levels of a glucose-derived capsule biosynthesis intermediate, UDP-glucuronate (UDP-GlnA), are reduced 4-fold in 5% CO 2 relative to ambient CO 2 levels (Fig 4B), supporting the idea that flux through this pathway is increased. In the next step of capsule biosynthesis, UDP-GlnA is converted to UDP-xylose by UDP-xylose synthase (UXS1, [29]). Consistent with this function, the uxs1∆ mutant has reduced capsule thickness [30]. Deletion of UXS1 increases CO 2 fitness (Fig 4C), suggesting that interrupting capsule biosynthesis early in the pathway provides a growth advantage. In contrast, deletion of UGE1, the enzyme that epimerizes UPD-glucose to UDP-galactose, increases capsule thickness [31] and reduces CO 2 fitness (Fig 4C). If this model is operative, we hypothesized that increasing glucose concentrations would improve growth of H99 in 5%CO 2 . RPMI medium contains 0.2% glucose, and we previously reported that increasing glucose concentrations to 2% did not affect the growth of H99 in 5% CO 2 [9]; therefore, we supplemented the medium to 2%, 5%, and 10% and incubated H99 at 37 °C in either ambient or 5% CO 2 for 48 h. Cultures of H99 containing increased glucose that were incubated at ambient CO 2 reached a higher final cell density relative to standard RPMI; however, there was no apparent dose effect between 2%–10%, indicating that at glucose concentrations above 2% other nutrients were growth limiting (Fig 4D). In 5% CO 2 , the effect of increasing glucose concentrations on the final cell density is blunted relative to ambient CO 2 (Fig 4D). A statistically significant effect on final cell density is not observed until the concentration of glucose is 10%. These data support the model that at host CO 2 concentrations, glucose is being diverted to processes other than replication and that high glucose concentrations can saturate the non-replicative pathway which, in turn, leads to increased final culture density. In contrast, at ambient CO 2 , 2% glucose is sufficient to render glucose a non-limiting nutrient. Taken together with our genetic and metabolic data, we propose that CO 2 -induced capsule formation diverts glucose from glycolysis and other energy-generating processes, leading to the use of alternative carbon sources at host CO 2 concentration.

Protein catabolism and urea cycle activity is increased in response to host CO 2 concentrations Metabolomic studies in other yeast have shown that depletion of glucose increases protein catabolism in order to mobilize amino acids as carbon sources [32]. As shown in Fig 5A, nine di- and tripeptides are increased at 5% CO 2 , a strong indication that C. neoformans protein catabolism is operative under these conditions. In addition, the amounts of six amino acids (Leu, Phe, Tyr, Trp, His, and Met) are increased by ≥ 2-fold in cells grown in 5% CO 2 (S3 Table). Of the nine di- and tri-peptides identified, all contain ketogenic amino acids, and 7/9 peptides contain Leu and Lys which are exclusively ketogenic amino acids. Ketogenic amino acids can be further catabolized in the mitochondria to acetyl-CoA which then enters the TCA cycle as an alternative to acetyl-CoA derived from glycolysis-derived pyruvate [33]. Thus, it appears that C. neoformans mobilizes alternative carbon nutrients by protein catabolism to compensate for reduced amounts of glucose after shift to host levels of CO 2 . PNG larger image

TIFF original image Download: Fig 5. Host CO 2 concentrations induce increased protein catabolism and urea cycle activity. The normalized area of each metabolite was determined by tandem mass spectrometry as described in materials and methods for H99 incubated in buffered RPMI in ambient air or 5% CO 2 . Asterisks indicate statistically significant differences in abundance at 5% CO 2 relative to ambient air (Welch corrected, two-tailed t test, P < 0.05). A. Relative abundance of the indicated di- and tripeptides in ambient and 5% CO 2 . B. Competitive fitness of mutants involved in vacuolar ATPase function. Competitive fitness of the indicated VMA mutants relative to H99 was determined for indicated strains after combining a 1:1 ratio of mNeonGreen-labeled H99 and unlabeled mutant cells in RPMI 1,640 medium with 165 mM MOPS at pH 7 and incubating at 37 °C in ambient air or 5% CO 2 for 24 h. Cell populations were characterized by flow cytometry, and the percentage of mNeonGreen-negative cells in 5% CO 2 was divided by those in ambient conditions to calculate competitive fitness. Bars represent the average of three biological replicates with error bars indicating SD. Asterisks indicate a significant difference relative to H99 as determined by Students t test (P < 0.05). C. Schematic of the initial steps of the urea cycle in C. neoformans. Created in BioRender.com. Ristow, L. (2026) https://BioRender.com/3jm5mu3. Relative abundance of the urea cycle intermediates citrulline (D) and argininosuccinate (E) in ambient and 5% CO 2 . The data underlying this Figure can be found in S1 and S3 Tables. https://doi.org/10.1371/journal.pbio.3003561.g005 Protein catabolism occurs in the vacuole and is dependent upon vacuolar ATPase (V-ATPase) to generate the acidic environment necessary for the activation of vacuolar peptidases [34]. Multiple strains lacking genes coding for V-ATPase components or proteins involved in the assembly of the V-ATPase complex show reduced CO 2 fitness (Fig 5B), indicating that maintenance of vacuolar pH is critical for tolerance of high CO 2 . Although V-ATPase mutants have pleomorphic effects on cells, the metabolomic data clearly indicate protein catabolism occurs during the initial phases of CO 2 stress. Accordingly, loss of V-ATPase function would likely interfere with efficient generation of ketogenic amino acids necessary to maintain mitochondrial respiration. CO 2 /HCO 3 - is essential for the cell, but high levels of CO 2 are stressful and can ultimately be toxic. As such, CO 2 is both generated and consumed by metabolic reactions in the cell. The urea cycle is one of the CO 2 -consuming processes utilized by eukaryotic cells (Fig 5C). Ura2 is a carbamoyl phosphate synthase that catalyzes the rate-limiting condensation of CO 2 with ammonia which feeds into the urea cycle. Previous work by our group found that deletion of URA2 reduces CO 2 fitness [11] and our screen confirmed this result (Table 1). Carbamoyl-phosphate enters the urea cycle by reacting with ornithine to generate citrulline which, in turn, is converted to argininosuccinic acid (Fig 5C). Both the citrulline and argininosuccinic acid levels are elevated at 5% CO 2 relative to ambient levels (Fig 5D and 5E), suggesting increased conversion of CO 2 to urea cycle intermediates relative to ambient CO 2 . Supporting this conclusion, in S. cerevisiae, urea cycle activity is enhanced under increased CO 2 concentrations to deplete excess CO 2 /HCO 3 [35]. Ura2 is also required for the synthesis of uridine/uracil and deletion mutants are auxotrophic for those nutrients. However, Chadwick and colleagues found that the reduced fitness of the ura2∆ mutant in 5% CO 2 was not rescued by uracil or uridine supplementation while the auxotrophic phenotypes were reversed [11]. Therefore, Ura2 appears to buffer the cell against high levels of CO 2 by converting it into intermediates of the urea cycle. Together, these data indicate that host levels of CO 2 lead to substantial alteration of central carbon metabolism through multiple mechanisms.

Host concentrations of CO 2 alter C. neoformans cellular redox balance In S. cerevisiae, acute depletion of glucose leads to an increase in mitochondrial respiration. Banerjee and colleagues have also demonstrated that glucose depletion in C. neoformans leads to increased oxidative stress which is a byproduct of mitochondrial activity [36]. Based on these precedents as well as the large set of mitochondrial genes required for CO 2 fitness, we predicted that host CO 2 concentrations would increase mitochondrial respiration and membrane potential. To test this hypothesis, we characterized the membrane potential of mitochondria in ambient and 5% CO 2 using JC-1 staining as previously reported by the Kronstad lab [37]. In this assay, increased mitochondrial membrane potential leads to a red shift in the JC-1 dye. As shown in Fig 6A, growth in 5% CO 2 leads to an increase in the proportion of red-shifted cells relative to ambient CO 2 . This increase in mitochondrial membrane potential is consistent with an increase in respiration at host concentrations of CO 2 [38]. The membrane potential is driven by the TCA cycle and electron transport chain enzymes. Accordingly, deletion mutants of the TCA cycle enzyme succinyl CoA synthetase (LSC1) as well as three complex I/ubiquinone NADH dehydrogenase components CNAG_05267, CNAG_02403, and CNAG_02833 have reduced fitness in host CO 2 concentrations (Table 1). PFK1 and FBP26 reduce the ability of yeast to transition from glycolysis to respiration [39]. We, therefore, suggest this function may contribute to the reduced fitness of the pfk1∆ and fbp26∆ mutants (Table 1) at 5% CO 2 . PNG larger image

TIFF original image Download: Fig 6. Host concentrations of CO 2 alter C. neoformans cellular redox balance and engage multiple acetyl-CoA generating pathways for tolerance. A. The mitochondrial membrane potential of H99 in ambient or 5% CO 2 conditions for 24 h was assessed by JC-1 staining. JC-1 staining was analyzed by dividing the MFI resulting from excitation with the blue (488 nm) laser in the BL2 emission filter (574/26) by the MFI in the BL1 emission filter (530/30), i.e., red/green ratio of single yeast cells. Bars represent the average red/green ratio of a minimum of 10,000 cells per condition and error bars represent the SEM. Asterisks indicate a significant difference relative to H99 as determined by Student t test (P < 0.05). Cellular stores of NAD/NADH (B) and NADP+/NADPH (C) were determined for H99 after incubation in buffered RPMI medium at ambient or 5% CO 2 conditions for 24 h. Data are presented as ratios of the oxidized or reduced forms. Bars represent the average of three biological replicates and error bars represent SD. Asterisks indicate a significant difference relative to ambient conditions as determined by an unpaired t test (P < 0.05). D. The metabolic profile of H99 grown for 24 h in ambient air or 5% CO 2 was analyzed by tandem mass spectrometry. Data were processed and are presented for 6-P-Gluconate as normalized area. Asterisk indicates P < 0.05 for the comparison of ambient to 5% CO 2 using Welch corrected, two-tailed t test. E. Competitive fitness was determined for indicated strains after combining a 1:1 ratio of mNeonGreen-labeled H99 and unlabeled mutant cells in RPMI 1,640 medium with 165 mM MOPS at pH 7 and incubating at 37 °C in ambient air or 5% CO 2 for 24 h. Cell populations were characterized by flow cytometry, and the percentage of mNeonGreen-negative cells in 5% CO 2 was divided by those in ambient conditions to calculate competitive fitness. Bars represent the average of three biological replicates with error bars indicating SD. Asterisks indicate a significant difference relative to H99 as determined by one-way ANOVA with Tukey’s multiple comparisons test (P < 0.05). The competitive fitness of H99 and indicated strains (acl1∆ or yhm2∆) was determined in vivo after two inoculation routes in CD-1 mice. F. For intranasal inoculation (IN), wild-type and mutant cells were mixed in a 1:1 ratio prior to inoculation with 5 × 104 CFU/animal in a 50 µL volume. Organs were harvested at day 10 post-inoculation. G. For intravenous inoculation (IV), wild-type and mutant cells were mixed in a 1:1 ratio prior to inoculation with 1 × 105 CFU/animal in 100 µL volume. Organs were harvested at day 4 post-inoculation. Tissues were homogenized and serially diluted on YNB or YNB with 400 µg/mL hygromycin and incubated at 30 °C for two days before counting CFUs. The competitive fitness was determined by dividing the CFU of mutant cells (HYG resistant) by the CFU of H99 cells in lung (F) from the IN model or brain (G) from the IV model. Data are represented with the average CFU/mouse and SD. Asterisks indicate a significant difference relative to H99 as determined by unpaired t test (P < 0.05). The data underlying this Figure can be found in S3 Table and S1 Data. https://doi.org/10.1371/journal.pbio.3003561.g006 To further characterize the effect of CO 2 on the redox state of C. neoformans, we measured the ratios of NAD+/NADH and NADP+/NADPH under ambient and host CO 2 concentrations. As shown in Fig 6B, CO 2 increases the NAD + /NADH ratio by ~2.5 fold. The increased NAD+/NADH ratio is consistent with the higher mitochondrial membrane potential observed in Fig 6A and further supports the conclusion that 5% CO 2 increases respiration and TCA cycle activity relative to ambient CO 2 concentrations. One of the consequences of increased respiration is increased oxidative stress in the cell. NADPH is a critical component of the biochemical processes that detoxify ROS [38]. Interestingly, the ratio of NADP+/NADPH is unchanged at 5% CO 2 compared to ambient CO 2 (Fig 6C). This suggests that CO 2 -tolerant C. neoformans can compensate for the increase in respiration and oxidative state of the cell at 5% CO 2 without depleting NADPH stores significantly. The pentose-phosphate pathway generates two molecules of NADPH through the oxidation of glucose first to 6-phospho-gluconate and then, through an oxidative decarboxylation step, to ribulose-5-phosphate [40]. The levels of 6-phosphogluconate are reduced in 5% CO 2 relative to ambient CO 2 conditions which is consistent with increased pentose phosphate pathway activity (Fig 6D), possibly to maintain NADPH levels and compensate for increased ROS resulting from increased mitochondrial respiration.

Multiple acetyl-CoA-generating pathways are required for C. neoformans tolerance of host CO 2 concentrations Acetyl-CoA is not only required for mitochondrial respiration through the TCA cycle but is also a fundamental substrate for lipid biosynthesis as well as for histone acetyltransferases involved in the regulation of gene expression [41]. Acetyl-CoA cannot diffuse across organelle membranes and, therefore, distinct pools of acetyl-CoA are present in the mitochondria, cytosol, and nucleus [41]. In the presence of glucose, glycolysis delivers pyruvate-derived acetyl-CoA to the mitochondria and TCA cycle. Our metabolomics data suggests that glucose and glycolysis are reduced at 5% CO 2 and that ketogenic amino acids are mobilized to provide acetyl-CoA to support the increased mitochondrial respiration driven by the acute reduction in glucose. Lipid catabolism through β-oxidation contributes to the acetyl-CoA pool and, therefore, could also contribute to acetyl-CoA homeostasis during CO 2 stress [32]. Supporting this hypothesis, we identified three mutants in the β-oxidation pathway that have reduced CO 2 fitness: the enoyl CoA hydratase CNAG_04531, the multi-functional protein MFE2, and the 3-oxo-acyl CoA thiolase POT1 (Table 1). Enoyl-CoA hydratases catalyze the second step of fatty acid β-oxidation while Pot1 directly generates acetyl-CoA. MFE2 is a peroxisomal enzyme that catalyzes the first two steps of β-oxidation [38]. MFE2 is subject to glucose repression in C. neoformans, and POT1 has been shown to be negatively regulated by glucose in S. cerevisiae [42]. The expression of both MFE2 (1.8-fold, FDR = 9e-15) and POT1 (3-fold, FDR 1e-68) is increased in 5% CO 2 , further supporting its likely role in generating acetyl-CoA under these conditions [12]. In addition, β-oxidation is positively regulated by the AMP-activated protein kinase Snf1 in other systems [43,44], and this kinase is required for full fitness at host CO 2 concentrations (competitive index: 0.67; p = 0.012, S1 Table). Finally, POT1, MFE2, and enoyl CoA hydratase CNAG_04531 are required for full fitness during infection [14]. Thus, the in vivo fitness defects of genes related to β-oxidation may be due, at least in part, to their importance in the adaptation of C. neoformans to host CO 2 . Two additional enzymes could also contribute to the generation of acetyl-CoA: acetyl-CoA synthetase (ACS1) and aceto-acetyl-CoA synthetase (KBC1, [45]). Acs1 converts acetate to acetyl-CoA. Kbc1 condenses the ketone body acetoacetate with CoA to give aceto-acetyl-CoA which, in turn, is a substrate for Pot1, leading to two molecules of acetyl-CoA. We, therefore, tested the CO 2 fitness of acs1∆, kbc1∆, and the corresponding double deletion mutants using the competition assay. The kbc1∆ acs1∆ double mutant showed a modest reduction in CO 2 fitness whereas neither single mutant differed from WT (Fig 6E), suggesting that these enzymes play a minor role in maintaining acetyl-CoA homeostasis during CO 2 stress. Amino acid catabolism and β-oxidation to acetyl-CoA occur in the mitochondria and directly provide acetyl-CoA for use in the TCA cycle. However, acetyl-CoA is also required in other cellular components to support biosynthesis and gene expression. An important source of cytosolic and nuclear acetyl-CoA in C. neoformans and other cells is, in fact, the mitochondria. Indeed, previous work by the Kronstad lab showed that the majority of cytosolic and nuclear acetyl-CoA appears to be derived from citrate generated in the mitochondria [46]. Citrate is exported to the cytosol, where it is converted to acetyl-CoA and oxaloacetate by ATP-citrate lyase (ACL1). In prior work, we found that the acl1∆ mutant has a severe growth defect in 5% CO 2 [45]. This suggests that nuclear and cytosolic acetyl-CoA stores remain primarily dependent on Acl1. The deletion mutant of a tricarboxylic acid transporter, CNAG_06096, has a strong fitness defect in 5% CO 2 (Table 1). CNAG_06096 is homologous to S. cerevisiae Yhm2 which is a mitochondrial citrate-oxoglutarate antiporter [47]. In other fungi with ATP-citrate lyase enzymes such as Aspergillus spp., deletion of the citrate-oxoglutarate transporter leads to reduced acetyl-CoA production [48]. Presumably, this is due to reduced cytosolic citrate in the absence of the transporter. If Yhm2 plays this role during mammalian C. neoformans infection, then, like the acl1∆ mutant, it should have a profound fitness defect in vivo. To test this hypothesis, we infected mice with a 1:1 mixture of H99 and either the acl1∆ or yhm2∆ mutant via inhalation (Fig 6F) and intravenous injection (Fig 6G). As expected, based on the previous report of its reduced virulence, the acl1∆ had dramatically reduced fitness in both lung and brain tissue. The yhm2∆ mutant showed a similarly profound reduction in fitness relative to WT. These data indicate that mitochondrial citrate-derived acetyl-CoA is required for C. neoformans tolerance of CO 2 and infection of mammals. Taken together, these data indicate that host concentrations of CO 2 establish a relatively low glucose state in C. neoformans. In response, oxidation of fatty acids and amino acid catabolism generates mitochondrial acetyl-CoA which then supports increased respiratory metabolism. Cytosolic and nuclear pools of acetyl-CoA remain largely dependent on the export of mitochondrial citrate through the transporter Yhm2; citrate is then converted to acetyl-CoA by Acl1 in the nucleus and cytosol.

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3003561

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/