https://www.nature.com/articles/s42003-026-11034-w
An enigmatic phosphatase domain in a deadly fungal pathogen
Cryptococcal meningitis remains a major therapeutic challenge in part because Cryptococcus neoformans, like its human host, is a eukaryote. Many processes required for fungal growth and survival therefore have counterparts in human cells, leaving fewer pathogen-specific vulnerabilities that can be targeted selectively. Developing treatments that eliminate the fungus while minimizing damage to the host is consequently far more difficult than simply identifying processes essential for fungal survival.
C. neoformans is an environmental yeast that can be inhaled into the lungs and, in people with weakened immune systems, disseminate through the bloodstream to the brain. The US Centers for Disease Control and Prevention estimates that cryptococcal meningitis causes around 112,000 deaths each year among people living with HIV, and the World Health Organization has placed C. neoformans in its critical-priority group of fungal pathogens. Features specific to the fungus, or sufficiently distinct from human counterparts, are therefore valuable for understanding cryptococcal pathogenesis and identifying opportunities for selective therapeutic targeting.
One such feature led us to Cac1, the adenylyl cyclase of C. neoformans. Cac1 produces cyclic AMP, or cAMP, a signalling molecule that helps the fungus respond to environmental changes through a pathway regulating melanin and capsule production, sexual differentiation, and titan cell formation. Loss of Cac1 disrupts these processes and eliminates virulence in experimental infection models.
The unusual part of Cac1 was not the region that makes cAMP. Cac1 also contains a unique protein phosphatase 2C, or PP2C, domain adjacent to its cAMP-producing region, whereas human adenylyl cyclases lack this domain. Phosphatases regulate signalling by removing phosphate groups from proteins, but whether the PP2C domain within fungal adenylyl cyclases retains phosphatase activity had remained unknown. Its presence in fungal but not human adenylyl cyclases made its function particularly intriguing to investigate.
An earlier study of predicted phosphatases in C. neoformans had included Cac1 because its sequence contained a PP2C-like region. Sequence annotation, however, is only a prediction, and no biochemical evidence had established that this domain could remove phosphate groups or explained its contribution to Cac1 function. Two questions therefore shaped this study. Did the domain truly function as a phosphatase, and what did it contribute to Cac1 function?
From predicted structure to confirmed activity
Structural modelling complicated the first question. Several residues normally found at the active site of canonical PP2C phosphatases were replaced by different amino acids in Cac1. Although the overall fold resembled a PP2C enzyme, activity could not be inferred from structure alone.
The biochemical experiment presented a more practical obstacle. Much of the isolated PP2C protein was insoluble during preparation, making direct activity measurements difficult. Adding manganese proved to be a turning point, allowing us to obtain enough soluble protein for testing. The purified domain removed phosphate groups from phosphorylated serine and threonine substrates, but not phosphorylated tyrosine. Despite its unconventional active site, the Cac1 PP2C domain therefore retained serine/threonine phosphatase activity.
The biochemical result answered one question but raised another. What role does this phosphatase activity play in C. neoformans?
Removing one domain at a time
To separate the two activities within Cac1, we generated strains lacking either the PP2C domain or the cAMP-producing domain while retaining the remainder of the protein. Because instability could have complicated interpretation of the phenotypes, we confirmed comparable expression of the domain-deletion variants at both transcript and protein levels.
The two deletions produced different outcomes. Loss of the cAMP-producing region largely resembled complete loss of Cac1, causing broad defects in melanin, capsule, and developmental processes. By contrast, loss of PP2C generally produced intermediate phenotypes, suggesting that this domain contributes to full Cac1 function without being solely responsible for it. The extent of these defects also varied with the conditions tested, indicating a context-dependent contribution.
Titan cell formation provided one of the clearest distinctions. Cells lacking the PP2C domain failed to form titan cells under our laboratory induction conditions, yet cAMP supplementation restored the phenotype. The rescue suggested that PP2C contributes to the cAMP signalling required for this increase in cell size. Changes in the cell wall and capsule further indicated that PP2C loss affected the fungal surface encountered by the host.
Transcriptomic and phosphoproteomic profiling added another layer. The two deletions affected overlapping but distinct signalling outputs. PP2C loss caused extensive changes in protein phosphorylation despite more limited changes in gene expression, suggesting that much of its regulatory effect occurs at the post-translational level. Although phosphoproteomic analysis identified several PP2C-dependent candidate targets, the proteins directly dephosphorylated by PP2C remain unknown.
Separating mortality from tissue damage
The infection experiments produced a more complex picture. Mice infected with the PP2C-deficient strain survived, consistent with the reduced virulence-associated phenotypes observed in vitro, whereas mice infected with strains expressing intact Cac1 did not.
The unexpected result emerged in the lungs. At 16 and 26 days after infection, fungal burdens in mice infected with the PP2C-deficient strain remained comparable to those in mice infected with the strains carrying intact Cac1. Despite surviving, these mice developed a type 2-skewed immune response together with progressive, emphysema-like pulmonary damage. By 111 days, very few fungal cells could be recovered, yet signs of lung injury persisted.
Loss of the PP2C domain therefore separated mortality from another important outcome of infection, tissue damage. The infection was no longer fatal, but it was far from harmless. Alterations in the fungal surface may contribute to the unusual host response, although the relationship remains to be determined.
The study began with an uncharacterized domain embedded in a familiar signalling enzyme and ultimately linked it to phosphatase activity, fungal morphogenesis, cell-surface organization, signalling, and disease outcome. An important next step will be to identify the direct substrates of PP2C and determine how its activity is integrated with cAMP signalling.
The feature that first drew our attention to this domain remains relevant. Fungal adenylyl cyclases carry an integrated PP2C domain that human adenylyl cyclases lack. That distinction does not yet make PP2C a validated drug target, but it could prove valuable for selective antifungal strategies.
The link of this article: https://www.nature.com/articles/s42003-026-11034-w