Is Cryptococcus neoformans a pleomorphic fungus?

Abstract

Improved understanding of the human fungal pathogen Cryptococcus neoformans, classically described as a basidiomycete budding yeast, has revealed new infection-relevant single cell morphologies in vivo and in vitro. Here, we ask whether these morphologies constitute true morphotypes, requiring updated classification of C. neoformans as a pleomorphic fungus. We profile recent discoveries of C. neoformans seed cells and titan cells and provide a framework for determining whether these and other recently described single-cell morphologies constitute true morphotypes. We demonstrate that multiple C. neoformans single-cell morphologies are transcriptionally distinct, stable, heritable, and associated with active growth and therefore should be considered true morphotypes in line with the classification in other well-studied fungi. We conclude that C. neoformans is a pleomorphic fungus with an important capacity for morphotype switching that underpins pathogenesis.

Introduction

The fungal kingdom is diverse and beautiful, encompassing everything from macrostructures (mushroom fruiting bodies) to microscopic single-celled organisms. Changes in morphotype are a common strategy employed by human fungal pathogens, which tend to be single celled, upon infection. The temperature-dependent dimorphic switch from hyphal to yeast form, for example, enables immune evasion and dissemination [1]. For Cryptococcus species, Candida glabrata, and other budding yeasts that lack a stable infectious hyphal morphotype, the budding yeast form has historically been considered the sole morphotype. However, the term ‘budding yeast’ is inexact, and recent work suggests that this form contains many more morphological variants and stable subpopulations than previously thought. For example, different variants within the budding yeast population of a single species can be transcriptionally distinct and self-propagating [2–4]. Here, we argue that, for Cryptococcus neoformans, these variants represent distinct subpopulations worthy of the ‘morphotype’ designation. We conclude that impressive variation in morphology exists and that this directly underpins pathogenesis.

The human fungal pathogen Cryptococcus neoformans is readily identified in patient samples by its distinctive and unique morphology: a budding yeast surrounded by a dramatic polysaccharide capsule that easily distinguishes it from other fungi (Figure 1a). However, this classical presentation belies a complex and fascinating biology that we are only beginning to fully appreciate. As a basidiomycete, Cryptococcus yeast-phase growth is distinct in key ways from ascomycete budding yeasts, including cell wall organization, molecular regulation of polarity, bud site positioning and emergence, nuclear dynamics and cellular division, and genome organization [5–11]. As sexual dimorphs, Cryptococcus spp. can switch to filamentous growth in the context of nitrogen limitation or temperature (monokaryotic fruiting) or upon cell–cell fusion (dikaryotic hyphae), which terminate in basidia that produce dispersible basidiospore (Figure 1a) thought to be important for infection of the mammalian lung [12–14].

Figure 1.

Cryptococcus Morphotypes. (a) Recognized Cryptococcus morphotypes: (left) Yeast-phase cells are identified as Cryptococcus through the production of a robust capsule revealed by india ink. (Middle) Hypha with clamp cells terminate in two basidia, with spores visible emerging from the top. Hyphae resulting from KN99 MatA x KN99 MatAlpha mating on Murashige-Skoog medium for 7 days were stained with calcofluor white for chitin. (Right) A budding yeast imaged by TEM reveals the layers of cell wall resulting from progressive rounds of budding. (b) Newly identified morphotypes: (top) seed cell culture (Denham et al., 2022) generates encapsulated seed cells (4–6 μm), comparable in size and shape to budding yeast (4–6 μm). (bottom) Titan culture (Dambuza et al., 2018) generates a heterogenous population, including titanides (2–3 μm oval cells with minimal capsule) and titan cells (> 10 μm). TEM, Transmission Electron Microscopy.

In the context of host infection, Cryptococcus haploid cells can undergo a variety of morphological transitions 15]. The majority of cells during infection are encapsulated haploid yeast, approximately 5–7 μm in size [16,17]. Very rarely, there have been reports of pseudohyphae during infection [18–22]. Most dramatically, yeast can engage an inducible change in body plan to become large, polyploid titan cells that propagate by budding, and these are observed both in vivo and in vitro in response to a variety of signals [23–30]. Yet small cells are also relevant: The production of very small (1 μm), round, thick-walled ‘microcells’ has been documented in vivo and in clinical isolates [31,32]. Thin-walled oval ‘titanides’, 2–3 μm, have been observed in vitro and in vivo [27,33]. Finally, infection-relevant ‘seed cells’, 4–6 μm, comparable in size to in vitro grown yeast, were recently identified in vivo as being critical for dissemination, and these are linked to growth in phosphate-rich bird guano, a major environmental niche [17]. Together, these diverse, environmentally relevant morphologies suggest Cryptococcus is in fact pleomorphic and that these morphological changes underpin environmental stress resistance and pathogenesis. In this review, we lay out the evidence for and against such an expanded understanding of C. neoformans cell biology.

How do we define morphotypes?

In the dimorphic fungi, yeast and hyphae both meet several important criteria that establish them as ‘forms’ and not just variants within a population of high phenotypic divergence. (1) Each morphotype has a transcriptional signature [34]. (2) Each morphotype is either directly reproduced in the daughter or the mother dictates a distinct morphotype of the daughter. The daughter then replicates faithfully in the new form. This ability is therefore passed down across generations, rather than being a transient, phenotypic response to a stimulus. (3) Morphotypes are distinct from terminal sexual structures, which have distinct, developmentally relevant morphologies and definitions.

The three criteria are easily met for the yeast and hyphae forms of thermal dimorphs, such as Histoplasma capsulatum and Blastomyces dermatitidis [1]. Coccidioides species, while classified as dimorphs, also have multiple, nonsexual hyphae and spherule morphotypes [35]. Candida species are often dimorphic, with both yeast and hyphal phases, and recent work has identified multiple morphotypes within the yeast phase of Candida albicans [36]. The classic example is white versus opaque cells, which are transcriptionally and phenotypically distinct and reasonably stable due to self-propagating transcriptional networks [37]. Two recently discovered cell states, gray [3] and gastrointestinal-induced transition cells [4,38], can occur during infection [39]. C. albicans also produce ‘Goliath cells’, which exhibit distinct adhesive and stress resistance phenotypes [40]. This diversity of morphotypes, beyond sexual and thermal dimorphism, has led to the concept of Candida albicans as a pleomorphic fungus.

Applying these definitions, we can distinguish single-cell cryptococcal morphotypes (Figure 1) from changes that occur over less than one generation, such as capsule and melanin production [41,42], or cell wall changes that occur in response to stress or aging [5,43,44]. We additionally recognize single-cell morphotypes as distinct from morphogenesis during the mating phase (Figure 1a), which involves cell types that cannot directly propagate (i.e. the zygote must polarize to form a hypha and spores must germinate to yeast) [45,46]. Please note that we do not intend to offer proscriptive functional definitions of these morphotypes, as complete descriptions can lag initial discovery. Instead, we hope to highlight the existence of new, distinct single-cell morphotypes with the view that these may have different potentials for proliferation in the environment or for mediating pathogenesis.

Morphotypes versus developmental states

The existence of Cryptococcus mating hyphae have classically not been sufficient to classify Cryptococcus species as di- or pleomorphic [1]. Instead, hyphae are a developmental state associated with meiotic processes (the so-called ‘perfect state’) (Figure 1a). In the case of dikaryotic mating, following a/alpha cell fusion, the transient dikaryotic zygote fully polarizes to form septate hyphae. Clamp cells mediate independent mitosis of the a and alpha nuclei and subsequent translocation of parent-derived nuclei to the basidium, where they fuse and undergo meiosis to produce spores [47]. Morphogenesis and progression to meiosis are tightly co-ordinated, with the different developmental forms dictated by cell cycle events, including heterokaryon dynamics (clamp cells), nuclear fusion (basidium), and meiosis (spore emergence) [13,48] (reviewed in Ref. [49]).

Spores and conidia, the asexual propagule of filamentous fungi, are important infectious propagules that fulfill many of the phenotypic requirements of a morphotype: they are transcriptionally distinct and their gross morphology is unique [12,50–54]. However, to propagate they must germinate, entering a transcriptionally distinct state, and form either yeast cells, in the case of Cryptococcus, or hyphae, in the case of molds [12,55]. Spores and conidia are therefore transcriptionally distinct but not stable. By our definition, we consider them developmental structures rather than morphotypes.

Cryptococcus species can also undergo monokaryotic fruiting, primarily studied in the model species C. deneoformans [56,57]. C. deneoformans can undergo same sex mating [58], and there is some evidence that this also occurs in C. neoformans environmental isolates [59]. C. deneoformans isolates can also switch to monokaryotic filaments in response to a temperature-induced G2 arrest [14]. The terminal basidium of these cells is a site of meiosis and genome reduction, and so we consider these to be developmental structures.

Does Cryptococcus exhibit multiple morphotypes?

Given these definitions, we argue that C. neoformans has at least three budding cellular states that meet our definition of morphotypes — yeast, seed cells, and titan cells — and could well have several more, including the morphologically distinct titanide and microcells, as well as others relevant to pathogenesis.

Yeast cells

Yeast phase cells are the most commonly studied form of C. neoformans and have historically been considered phenotypically homogenous. However, a growing body of evidence demonstrates that yeast cells are far more varied than previously thought, especially when grown in environments that differ from rich yeast media.

Yeast cells grown in standard lab media are uniformly round, 5–7 μm, haploid, and produce clonal daughters through repeated budding. Importantly, C. neoformans are molecularly and phenotypically distinct from other well-studied budding yeasts. The C. neoformans cell wall is composed of chitin as well as deacetylated chitosan, along with beta-glucan and mannan, typically 200 nm thick [5,27,60,61]. Unlike S. cerevisiae and C. albicans, which place new daughter cells adjacent to the mother, generating a series of bud scars as cells age [44], C. neoformans (and other basidiomycetes) repeatedly bud from the same site, causing a thickening of the cell wall at the bud scar [9] (Figure 1a). This phenotypic difference is underpinned by molecular rewiring of polarity compared to S. cerevisiae [6–8,10,11]. C. neoformans has a well-defined cell cycle, yet the transcriptional programs that regulate transitions between cell cycle phases are distinct from those of S. cerevisiae [62]. Moreover, nuclear dynamics during mitosis are unique, with the nucleus moving into the daughter cell and dividing across the mother/daughter neck via semiopen mitosis [10]. Under host-relevant conditions, Cryptococcus cells produce melanin and an elaborate GXM/GalXM capsule [41,63] and also secrete factors associated with virulence, including urease and the cysteine-rich secreted factor Cpl1 [64,65]. Finally, specific transcriptional programs mounted in response to stress and environmental conditions in ascomycetes [66,67] underpin pathogenesis in Cryptococcus [68–71].

Seed cells

Seed cells (Figure 1b) are a smaller (~5 μm cell body, ~7 μm cell body + capsule) morphotype that appear in the lungs and brain around the time that C. neoformans cells start to disseminate in extrapulmonary organs [17] and correlate with brain fungal burden across different strains and species of Cryptococcus. Compared with encapsulated or yeast medium-grown unencapsulated yeast, seed cells have different cell surfaces, including increased conconavalin A binding, which suggests increased exposure of mannoproteins or other mannose moieties. This mannose exposure results in greater fungal cell uptake into the liver. Seed cells are also better able to enter the brain compared with encapsulated or yeast medium-grown cells. They are inducible and have statistically distinct transcriptomes from encapsulated yeast cells, identifiable by principle component analysis, and appear to have haploid genomes. In vitro, phosphate is sufficient to cause encapsulated yeast cells to produce seed cell daughters (Figure 1b, and personal communication, Jessica Brown). Pigeon guano, one of the environmental niches of C. neoformans cells, is also sufficient to induce seed cell formation [17]. While much about the specific biology remains to be discovered, seed cells fulfill our definition of a morphotype: transcriptionally distinct, stable with induction, and associated with active growth rather than a mating or developmental phase.

Titan cells

Titan cells (Figure 1b) are large (> 10 μm cell body, > 15 μm including capsule) in vivo–relevant cells that were first described in patient samples [23,24]. These were originally considered a ‘dead-end’ or ‘stress-induced’ host-specific morphology; however, early work demonstrated they can be reproducibly induced in vivo and propagate asymmetrically [25,26,31]. Specifically, Nielsen and Zaragoza defined these as cells with a diameter greater than 10 μm, ploidy greater than or equal to 4C, a more densely crosslinked capsule, and a thickened cell wall [27,72]. Work by Mukaramera et al. provided refined understanding of the cell wall composition, highlighting altered chitin/chitosan ratios in large versus small cells recovered from the host lung [73]. Subsequent work by Probert et al. identified differences in capsule epitope localisation during in vitro culture [74]. Okagaki et al. also identified that titan cells are associated with reduced phagocytosis of both small and large cells, consistent with the production of a titan cell–associated secreted factor with inhibitory activity against phagocytes [75].

Titan cells emerge when yeast cells are incubated at low density in the presence of specific environmental triggers, including the host environment, nutrient limitation, bacteria, activators of cAMP signaling, and mitochondrial stress [25–30]. However, under in vitro and in vivo conditions, titan cells comprise a minority of the total cell population (Figure 1b), complicating transcriptional characterisation. Despite this, RNAseq of heterogenous titan cultures suggest they have distinct transcriptional profiles relative to yeast-phase growth, and there are clear indicators of both positive and negative regulation of the morphological switch [28]. CO₂ is a major activator of cAMP signaling [76], and loss of cAMP signaling inhibits titan cell formation [27,28,77]. In addition, the CO₂-responsive transcription factor Usv101 negatively regulates titan cell formation [27,78]. Extracellular pH is a key determinant of whether Cryptococcus proliferates as yeast or switches morphotype [17,79], and the pH-responsive transcription factor Rim101 is required for titan cell formation and cell wall integrity [77,80]. High cell density represses titan cell formation [27]. The transcription factor Pdr802 regulates quorum sensing components OPT1, PQP1, and LIV3. and pdr802Δ mutants are hypertitanizing, consistent with a failure to sense cell density [81]. However, the exact nature of the quorum sensing molecule that mediates this remains unclear: The addition of exogenous synthetic quorum sensing peptide Qsp1 reduces, but does not completely block, the formation of titan cells, and deletion of CQS1, the gene encoding the Qsp1 peptide, does not make cells insensitive to high density, suggesting the existence of additional density-dependent regulators [28].

Do titan cells fit the criteria of a true morphotype?

Overall, this suggests titan cells are an inducible, transcriptionally distinct cell type; however, the identity of titan cells as a true morphotype is more subtle. Further examination of the events driving titan cell formation may suggest they are more analogous to fruiting bodies than they are to yeast.

The capacity to switch to titan growth appears to be specific to the Cryptococcus lineage [82]. Titan cells do not divide symmetrically [72], but titan cells morphology is progressive across generations, with daughter characteristics dictated by mother: Haploid yeast cells endoreduplicate to form uninucleate polyploid titan cells, and these then divide asymmetrically to produce haploid, aneuploid, or diploid daughter cells that in turn have the capacity to form titan cells [25,27,83]. Given their capacity for inducible changes in ploidy, it follows that the formation of titan cells is inextricably linked to cell cycle–dependent events. Titan cells initially form when cells fail to progress through G2 arrest, instead endoreduplicating their genomes, either through exogenous DNA damage or through altered cell cycle regulation [13,84]. Loss of the cyclin Cln1, required for exit from G2, results in elevated titan cell formation under inducing conditions [84]. In addition, the negative regulator Usv101 sits downstream of Swi6 (CNAG_01438), a component of the mitotic SBF/MBF regulatory complex that also regulates capsule and melanin [27,41,78]. Genes linked to the meiotic cell cycle are also implicated in titan cell formation. For example, MATa cells in the KN99 lineage produce titan cells at elevated rates (> 20%), and this is dependent on the alpha pheromone receptor Ste3a [26]. The receptor Gpr5 also appears to be required for in vivo and in vitro titan cell formation [27,77], likely acting via Gpa1 [77]. Gpa1 localises to the plasma membrane in response to phosphorylation by the Cyclin-dependent Kinase-related kinase Crk1, also involved in meiosis [85,86]. Finally, meiotic genes mediate ploidy reduction in response to genotoxic stress [13]. It has been proposed that the capacity to alter ploidy is a driver of diversification of otherwise clonal populations and may underpin drug resistance [83]. Overall, this raises the possibility that titan cells may be a developmental morphotype with implications for disease outcome.

Additional potential morphotypes and discovery of new morphotypes

Images of C. neoformans in human or murine lungs are impressive in their heterogeneity [24,87,88], and this variance could well indicate additional morphotypes. Several promising candidates have been reported but more data are needed to assess whether these are additional true morphotypes. These include a variety of distinct ‘small cells’ (< 4 μm).

Titanides

Titanides (Figure 1b) appear in conditions that also induce titan cells, yet their origin remains unclear [27]. Distinct from round yeast and seed cells (both 4–6 μm), titanides are distinctly oval in shape, 2–3 μm on their long axis, and have significantly thinner cell walls than yeast [27]. They have been observed in multiple different in vitro assay conditions as well as in vivo and are produced by a variety of VNI isolates at varying rates [27,33,89,90].

Microcells

Microcells were originally defined as < 1 μm with thick cell walls [31]. Their round shape, very small size, thick cell wall and robust capsule easily distinguish them from seed and titanide cells [17,27,32]. They appear to be limited to C. neoformans lineages and are associated with the increased virulence of this species relative to other Cryptococcus species [32].

Viable but nonculturable and dormant cells

Another potential morphotype observed in vivo are viable but nonculturable (VBNC) cells [91]. Although poorly understood, these may be analogous to bacterial persister cells [92], with intact membranes and low metabolic activity [91]. While whole transcriptome data are lacking, VBNCs are transcriptionally distinct from other, higher metabolism cells for limited set of 37 test genes. They also have relatively high calcofluor white staining, indicative of elevated cell wall chitin levels [91]. When isolated from mice, VBNCs can resume growth in the presence of serum [91] or Sabouraud dextrose agar [93]. Whether VBNCs are resistant to antifungal drugs, as bacterial persister cells are resistant to antibacterial agents, has not been established. However, VBNCs exhibit signs of cellular stress responses [93], which can confer stress adaptation that is important for virulence [94]. Moreover, recent work identified increased antifungal resistance in metabolically dormant Cryptococcus subpopulations [95]. Overall, we think that VBNCs could well represent a new morphotype, but the challenges of determining viability and whether different groups are working on the same subpopulation of cells suggest that more evidence is needed.

Conclusions

Based on the established definition for morphotypes, we conclude that so-called ‘yeast’ phases can be complex, comprising multiple stable, independent morphotypes and that Cryptococcus exhibits multiple, initially subtle morphotypes that are biologically important both to fungal proliferation and to pathogenesis. We argue that Cryptococcus should therefore be termed ‘pleomorphic’. We additionally highlight that titan cells may extend the current working definition of morphotype by co-opting meiotic processes normally reserved for developmental forms. Finally, a common thread across the morphotypes is changes in cell wall, antigen exposure, and secreted factors that have important implications for host response [17,27,51,60,64,65,73–75,91,96]. Cell wall changes can also occur during aging [44] or as transient responses to stress [5], independent of morphotype. Overall, it is clear that, while these changes alone do not underpin morphotype definitions, investigations of fungal morphotypes should include a robust analysis of the fungal cell wall and secretome.

Data Availability

No data were used for the research described in the article.