Extracellular vesicles from diverse fungal pathogens induce species-specific and endocytosis-dependent immunomodulation

Abstract

Microbial pathogens generate extracellular vesicles (EVs) for intercellular communication and quorum sensing. Microbial EVs also induce inflammatory pathways within host innate immune cells. We previously demonstrated that EVs secreted by Candida albicans trigger type I interferon signaling in host cells specifically via the cGAS-STING innate immune signaling pathway. Here, we show that despite sharing similar properties of morphology and internal DNA content, the interactions between EVs and the innate immune system differ according to the parental fungal species. EVs secreted by C. albicans, Saccharomyces cerevisiae, Cryptococcus neoformans, and Aspergillus fumigatus are differentially endocytosed by murine macrophages triggering varied cytokine responses, innate immune signaling, and subsequent immune cell recruitment. Notably, polysaccharide and hydrophobic protein structures on the outer layers of C. neoformans and A. fumigatus EVs inhibit efficient internalization by macrophages and dampen innate immune activation. Our data uncover the functional consequences of the internalization of diverse fungal EVs by immune cells and reveal novel insights into the early innate immune response to distinct clinically significant fungal pathogens.

Author summary

Similar to bacteria and viruses, fungi can infect humans and cause devastating consequences, most significantly in immunocompromised individuals. Currently, significant gaps exist in our understanding of the important interactions between fungal pathogens and host immune systems. Cells, including microorganisms and human cells, release small particles termed “extracellular vesicles”, that can carry proteins and genetic materials to communicate between cells. In our study, we investigated how host macrophages, one of the early immune responders to infection, respond to extracellular vesicles produced by different species of fungi that cause human disease. We found that macrophages engulf fungal vesicles that have significant structures on their surfaces, like sugars or proteins, to a lesser extent than vesicles without these surface structures. Without this vital first step, downstream inflammatory pathways are not as robustly activated, which could lead to less activity from our immune systems to fight off disease. This study provides novel insight into fungal vesicle properties that affect interactions with host immune systems.

Introduction

Invasive fungal infections have become more prevalent in recent years, largely impacting the growing population of immunocompromised individuals [1]. Eukaryotic in nature, fungal infections are difficult to treat due to increasing antimicrobial resistance and the limited availability of antifungal therapeutics compared to their bacterial counterparts. The observed increase in mortality from fungal infection necessitates fundamental research on the molecular interactions between fungal pathogens and host organisms [2]. Among others, the type I interferon (IFN) response has gained traction as a host defense against several common invasive fungal pathogens [3–5]. In particular, the stimulator of IFN genes (STING) pathway is an orchestrated signaling cascade triggered by microbial pathogens that regulates IFN and interferon-stimulated genes (ISGs) during the early stages of infection. STING activation induces innate immune responses against several fungal organisms, including Candida albicans and Aspergillus fumigatus [6,7]. Major IFN-producing pathways, including the STING pathway, are well studied in the context of viral and bacterial infections [8–13], however, their role during fungal infections remains understudied and represents a significant gap in knowledge. We recently identified the molecular mechanism behind C. albicans activation of the pathway. C. albicans-derived extracellular vesicles (EVs) trigger the induction of both IFNβ, a cytokine released upon STING pathway activation, and viperin, an ISG product, in a cGAS- and STING-dependent manner. C. albicans EVs and the DNA carried within them elicit an extensive proinflammatory response and phosphorylation of essential pathway components including IFN regulatory factor 3 (IRF3) and TANK-binding kinase1 (TBK1) [14]. As whole fungal cells are typically sequestered into phagolysosomes upon host cell internalization, the fusion of C. albicans EVs (Ca EVs) to the plasma membrane provide a mechanism for how fungal nucleic acids access and activate the cytosolic DNA sensor cGAS. These findings led us to explore the interactions of additional species of fungal EVs with innate immune cells and their subsequent effects within host cells.

As described in the MISEV 2023 guidelines [15], EVs are membrane-bound vesicles released from nearly all cell types, including both host and pathogen cells. They carry proteins, lipids, and nucleic acids and were first identified as messengers capable of shuttling information and cargo between the same cell types [16]. It is now known that host cells can take up EVs released by pathogens and exploit the use of these microbial products to elicit robust and diverse pro-inflammatory or anti-inflammatory immune signaling in the host [17].

In our study, we identified four different species of EV-generating fungi with diverse stimulatory, pathogenicity, and morphological properties to compare their early interactions with the innate immune system and any immunomodulatory impacts. Specifically, we examined the impact of significant surface structures on EVs as virulence factors, selecting two species lacking surface components (C. albicans and Saccharomyces cerevisiae) and two species possessing them (Cryptococcus neoformans and Aspergillus fumigatus).

S. cerevisiae is a low-virulent yeast that generates EVs (Sc EVs) that do not have notable structures on their surfaces. Sc EVs have been studied in the context of cell wall remodeling and immune cell maturation [18,19]. However, the mechanism of their internalization by macrophages and the extent to which Sc EVs trigger a type I IFN response is unknown [17,18].

The encapsulated yeast C. neoformans is an opportunistic fungal pathogen that can cause meningoencephalitis in immunocompromised individuals [20]. Not only is this pathogenic yeast in a completely different phylum than S. cerevisiae and C. albicans, but it also possesses virulence factors and extracellular capsule that make it difficult to be phagocytosed. The presence of its surrounding capsule, comprised mainly of a robust sugar molecule glucuronoxylomannan (GXM), is a potent virulence factor of C. neoformans [21]. GXM is present on C. neoformans EVs (Cn EVs), but the immunostimulatory properties of these EVs are understudied [22].

Additionally, we explored fungal EV interactions with innate immune cells not only by yeasts, but also by a filamentous mold previously linked to type I IFN signaling, A. fumigatus [7]. Although A. fumigatus conidia are not encapsulated by sugars like C. neoformans, they are surrounded by a hydrophobic layer of rodlet proteins, which may also be involved in immune evasion [23]. Furthermore, A. fumigatus-derived EVs (Af EVs) prime murine macrophages for increased fungal clearance capacity and increased survival in the wax moth model Galleria mellonella after a fungal challenge [24,25]. Thus, our study aimed to examine early host responses to EVs secreted from all four of these pathologically distinct fungal organisms.

Our findings reveal novel patterns of macrophage ability to endocytose fungal EVs according to fungal species and this is further dependent upon the cell surface layers that populate these EVs. Ca EVs and Sc EVs are more significantly endocytosed by macrophages compared to Cn EVs and Af EVs. In addition to our previous findings on Ca EVs, we identified how EVs from diverse fungal organisms differentially activate the STING pathway. Interestingly, the degree of STING pathway activation is EV species-dependent, revealing that EVs lacking structural outer layers are more efficiently endocytosed and can more robustly activate the STING pathway. As EVs may play a role in priming the host for infection and activating important proinflammatory mediators, these data provide new insights into the early host immune response to fungal pathogens and underscore the potential for fungal EVs to modulate innate immunity. By elucidating these interactions, this study generates the potential for new therapeutic targets and strategies in combatting fungal infections.

Results

EV interactions with immune cells are species-specific

Since different species of fungi have unique impacts on host organisms, we inquired how interactions between fungal EVs and immune cells would compare between fungal species. We isolated EVs from C. albicans, S. cerevisiae, C. neoformans, and A. fumigatus cultures through a series of centrifugation steps (isolation methods summarized in S1A Fig) for co-culture with macrophages (immortalized from C57BL/6 murine bone marrow cells) to compare EV internalization by innate immune cells. Interestingly, by staining the EVs with DiI lipophilic stain and using flow cytometry, we observed that macrophages had different capacities to uptake EVs depending on the species of EV. After a 3h co-incubation, over 50% of the macrophage population internalized Ca EVs and Sc EVs, compared to less than 10% of macrophages having internalized Cn EVs and Af EVs (Figs 1A and S2A). To determine the mechanism of macrophage uptake of EVs, we inhibited dynamin-dependent endocytosis in wildtype (WT) macrophages using dynasore [14,26]. We further assessed through flow cytometry that dynasore treatment inhibited macrophages from endocytosing Ca EVs and Sc EVs by 80.4% and 88.9%, respectively (Figs 1B and S2B). Entry into dynasore-treated WT macrophages by Cn EVs and Af EVs was also significantly blocked, further confirming endocytosis as the mechanism of EV entry into macrophages (S2C and S2D Fig).

Fig 1. Murine macrophage internalization, cytokine production, and neutrophil recruitment differ in responses to different fungal EVs.

A, Percent internalization of Ca EVs, Sc EVs, Cn EVs, and Af EVs by murine macrophages. Significance assessed by an ordinary one-way ANOVA and Dunnet’s multiple comparisons test, ***p = 0.0002 vs PBS treated macrophages. (5x10¹⁰ EVs/mL added per stimulation), n = 3. B, Percent endocytosis of Ca EVs and Sc EVs with murine macrophages treated with DMSO or 100µM dynasore. Significance assessed by a two-way ANOVA and Tukey’s multiple comparisons test, ****p= < 0.0001 vs respective DMSO controls. (5x10¹⁰ EVs/mL added per stimulation), n = 2. C, Induction of MCP-1 (in pg/mL) in WT and Sting^-/- macrophages stimulated with PBS, 2.5µg cGAMP, Ca EVs, Sc EVs, Cn EVs, and Af EVs (1x10¹⁰ EVs/mL added per stimulation). Significance assessed by an unpaired t-test, *p ≤ 0.05, **p ≤ 0.009 vs PBS treated WT macrophages, n = 2. D, Number of neutrophils that transmigrated towards supernatants from WT macrophage stimulated by PBS, 2.5µg cGAMP, Ca EVs, Sc EVs, Cn EVs, and Af EVs (5x10¹⁰ EVs added per stimulation), and 0.1µM fMLP as a positive control. Significance assessed by an ordinary one-way ANOVA and Dunnet’s multiple comparisons test, *p ≤ 0.0203, **p ≤ 0.0073, ****p < 0.0001 vs PBS-treated macrophages, n = 3.

Given the differences in macrophage EV endocytosis, we hypothesized that innate immune responses would vary depending on the EV fungal species. To address this hypothesis, we used a mouse inflammation panel that measures the induction of 13 main cytokines produced by innate immune cells to assess altered immune activity by macrophages exposed to fungal EVs. As our previous findings revealed that Ca EVs activate the STING pathway in macrophages [14], we assessed levels of these 13 cytokines in both WT and Sting^-/- macrophages stimulated with our fungal EVs, PBS, or cGAMP (positive control) with the mouse inflammation panel kit. Only three of the 13 cytokines showed changes compared to unstimulated macrophages: TNFα, IFNβ, and MCP-1 (Figs 1C, S3A and S3B). WT macrophages secreted higher levels of TNFα when stimulated by cGAMP, Ca EVs, Sc EVs, and Cn EVs (S3A Fig, top panel). Similarly to our prior results in mouse macrophages and human peripheral blood mononuclear cells [14], we observed an increased level of TNFα secretion in Sting^-/- macrophages stimulated with Ca EVs compared to PBS-treated macrophages. This pattern was consistent with cGAMP, Sc EVs, and Cn EVs. We did not observe an increase in TNFα production in either Af EV-treated WT or Sting^-/- macrophages. IFNβ was induced by cGAMP, Ca EVs, and Sc EVs in WT but not Sting^-/- macrophages, or in macrophages stimulated with Cn or Af EVs (S3A Fig, bottom panel). Increased MCP-1 secretion was induced by cGAMP and all EVs in WT, however, only cGAMP and Sc EVs induced moderate MCP-1 secretion in Sting^-/- macrophages (Fig 1C).

As MCP-1 plays a role in immune cell recruitment, we next assessed the recruitment of murine neutrophils (immortalized cell line WT Cas9-ER-HoxB8-GMP) in response to a co-culture of macrophage and fungal EVs. As a positive control for neutrophil recruitment, we used N-Formylmethionine-leucyl-phenylalanine (fMLP), a potent neutrophil chemoattractant [27]. We revealed significant neutrophil recruitment in response to supernatants from macrophages co-cultured with all EVs (compared to untreated macrophages) (Fig 1D), which complimented MCP-1 induction in macrophages stimulated by Ca EVs, Sc EVs, and Cn EVs (Fig 1C). Interestingly, we note that macrophages exposed to Af EVs also caused significant neutrophil recruitment despite having minimal MCP-1 secretion, indicating that MCP-1 is not solely responsible for this neutrophil chemotaxis. Altogether, these data indicate that EVs derived from both yeast and filamentous fungi result in both similar and differential responses in macrophage host defenses.

Differential capacity of fungal EVs from diverse sources on STING pathway activation

Since we observed IFNβ to be induced by several fungal EVs in a STING-dependent manner (S3A Fig, bottom panel) and our previous work confirmed the induction of IFNβ in macrophages by Ca EVs as STING-dependent [14], we next assessed STING pathway activation in macrophages stimulated by diverse pathogen-derived EVs. We assessed the expression of viperin, p-IRF3, IRF3, p-TBK1, and TBK1 by immunoblot as readouts of STING pathway activation. Viperin is an ISG product, and the phosphorylation of IRF3 and TBK1 is essential for the downstream transcriptional action of type I IFNs and pro-inflammatory cytokines, which are key components of the host immune response. Similar to IFNβ production, we observed a significant increase in viperin induction in WT macrophages stimulated with Ca EVs and Sc EVs (Fig 2A and 2B). Moderate viperin induction was observed in WT macrophages stimulated with Cn EVs, whereas Af EVs failed to induce viperin (Fig 2A and 2B). As predicted, viperin expression in Sting^-/- macrophages was absent upon EV stimulation, showing the necessity of this pathway (S3C and S3D Fig). Furthermore, phosphorylation of the key downstream components of the STING pathway (IRF3 and TBK1) were induced to varying levels in response to fungal EVs in WT macrophages. Notably, we saw greater IRF3 and TBK1 phosphorylation from Ca EVs, Sc EVs, and Cn EVs than from Af EVs (Fig 2A and 2B).

Fig 2. Fungal EVs differentially activate the STING pathway.

A, Representative immunoblots of viperin, phosphorylated IRF3, total IRF3, phosphorylated TBK1, total TBK1, and actin in response to PBS, transfected cGAMP (2.5µg), Ca EVs, Sc EVs, Cn EVs, and Af EVs (5x10¹⁰ EVs/mL added per stimulation) in WT macrophages. B, Densitometry of viperin, p-IRF3, and p-TBK1 relative to appropriate loading controls. Significance assessed by a two-way ANOVA and Dunnet’s multiple comparisons test, *p ≤ 0.0497, **p = 0.0068, ***p = 0.0007, ****p < 0.0001 vs PBS treated macrophages in each respective group, n = 3. C, Immunoblot of viperin in WT macrophages treated with either DMSO (vehicle control) or 100µM dynasore and stimulated with Ca EVs, Sc EVs, Cn EVs, and Af EVs (5x10¹⁰ EVs/mL added per stimulation).

To assess whether this STING pathway activation relies on the endocytosis of EVs into macrophages, we next used flow cytometry to examine how preventing endocytosis with dynasore treatment would impact viperin production in macrophages stimulated with EVs. Compared to the vehicle control (DMSO) which revealed a strong viperin induction in response to Ca EVs and Sc EVs, macrophages treated with dynasore showed no induction of viperin, regardless of EV species (Fig 2C). This further confirms the requirement of endocytosis for STING pathway activation and subsequent viperin induction by stimulating EVs.

Translocalization of cGAS in response to fungal EVs

The localization of the DNA sensor cGAS is dynamic, as it remains tethered to the nuclear membrane in non-stimulatory conditions to prevent autoreactivity, and subsequently translocates to the cytosol upon stimulation in the presence of cytosolic DNA [28–30]. To confirm previously published finding, we stained cells expressing cGAS-GFP with DAPI (4’,6-diamidino-2-phenylindole) and observed a co-localization of cGAS with the nucleus in an unstimulated state (S4A Fig) [31]. Stimulation of cGAS-GFP macrophages with Ca EVs, however, increased non-nuclear localization of cGAS [14]. Whether cGAS localization is impacted similarly across EVs from other fungal species remains unknown. Thus, we used confocal microscopy to investigate how EVs from other fungal organisms impact cGAS localization. The lipid membranes of the EVs were fluorescently labeled with DiI, cultured with cGAS-GFP macrophages, and then imaged to assess the cGAS-localization in macrophages that had internalized EVs. Macrophages that endocytosed fungal EVs had greater non-nuclear localization of cGAS compared to the untreated macrophages without internalized EVs. However, the total percentage of non-nuclear cGAS localization varied among fungal species. The three yeast EVs (i.e., Ca EVs, Sc EVs, Cn EVs), induced noticeably more non-nuclear cGAS localization compared to Af EVs (Fig 3A and 3B). When cGAS was observed as non-nuclear, it was either localized solely cytoplasmic or a mix of cytoplasm and nucleus. While still inducing non-nuclear cGAS localization, the Ca EVs isolated from whole fungal organisms resulted in less localization of cGAS to the cytoplasm than we previously reported in stimulated macrophages with EVs isolated from C. albicans 48h grown biofilms [14]. The mature nature of the 48h grown C. albicans biofilm EVs may explain the increased cGAS translocation, as a recent publication has noted changes in EV cargo content and membrane fluidics as a direct function of C. albicans biofilm growth time [32]. Regardless, our results demonstrate the ability of fungal EVs to alter the cellular localization of the cGAS-sensing molecule and these subcellular patterns reflect the ability of these EVs to induce downstream cGAS-STING signaling, viperin induction, and secretion of IFNβ.

Fig 3. Fungal EVs induce translocation of cGAS from the nuclear membrane to the cytosol.

A, Representative images of the localization of cGAS (green) when cGAS-GFP expressing macrophages are stimulated for 3h with DiI-labeled (red) PBS, Ca EVs, Sc EVs, Cn EVs, or Af EVs (8x10⁹ EVs added per stimulation). Scale bar = 10µm. B, Semi-quantitative analysis of the percentage of cGAS localization that is either nuclear or non-nuclear. cGAS localization analyses were performed on approximately 100 cGAS-GFP expressing macrophages that successfully endocytosed DiI-labeled EVs.

EVs share similar characteristics across diverse fungal species

EVs can vary in morphology and cargo depending on the cellular origin. Thus, to compare differences across our fungal species EVs, we analyzed several EV characteristics, including size, gross morphology, and due to the observed changes in STING pathway activation, internal DNA concentration and DNA GC content. Standard preparations of EVs from all species of interest yielded similar EV concentrations, typically between 1x10¹⁰ – 1x10¹² EVs/mL. Although we noticed some variability, the differences were not statistically significant (Fig 4A). Secreted EVs are known to be heterogenous in size, even when released by the same cell [33,34]. We investigated the heterogeneity amongst our preparations using Nanoparticle Tracking Analysis to measure the sizes of EVs isolated from our fungal organisms. Sizes ranged from 50-400 nm in diameter, with both the mode and median sizes of EVs from all four species falling within the 100–200 nm diameter range (Fig 4B and 4C). We next examined the gross morphology of our isolated EVs by transmission electron microscopy (TEM) to explore any significant structural differences between species (Fig 4D). All EVs appeared spherical with thick edges and noticeable depressions in the centers, a finding consistent with previous TEM imaging of EVs [35]. There are no significant differences between species visible by TEM, indicating that fungal EV morphology is broadly conserved regardless of species.

Fig 4. Physical and DNA characteristics of Ca EVs, Sc EVs, Cn EVs, and Af EVs.

A, The -log₁₀ values of EV concentration from standard isolation preps. Significance assessed using an ordinary one-way ANOVA and Tukey’s multiple comparisons test with no significance determined. n ≥ 5. B, Quantification of EV concentration according to diameter for Ca EVs, Sc EVs, Cn EVs, and Af EVs. n ≥ 5. C, The median and mode diameters of Ca EVs, Sc EVs, Cn EVs, and Af EVs. n ≥ 5. D, Representative images from TEM of Ca EVs, Sc EVs, Cn EVs, and Af EVs. E, The average DNA concentration (pg per 5x10¹⁰ EVs) of Ca EVs, Sc EVs, Cn EVs, and Af EVs. Significance assessed using an ordinary one-way ANOVA and Tukey’s multiple comparisons test with no significance determined. n ≥ 3. F, Percent GC content of EV DNA reads that mapped to the reference fungal genome as a truncated violin box plot. G, Percent GC content of the whole fungal genome compared to the EV DNA.

We previously unveiled that the DNA in Ca EVs triggers immune responses in macrophages through the STING pathway by binding to the cGAS DNA sensor. Given that the interaction between DNA and cGAS is needed for viperin induction, we hypothesized that Ca and Sc EVs might carry more DNA, as they induced more viperin in a STING-dependent manner, compared to Cn and Af EVs. Thus, we extracted and quantified EV DNA from all four species, and instead found similar concentrations across fungal species. The mean EV DNA concentration ranged from 54-1415 pg per 5x10¹⁰ EVs and there was no statistically significant difference between species (Fig 4E).

Finally, we more closely examined the DNA carried within these diverse EVs. First, we treated the EVs with a DNase (benzonase) to rid any extracellular DNA associated with the membranes and found no difference in viperin induction compared to untreated EVs (S3E Fig). Thus, we confirmed that DNA packaged within EVs is responsible for activating the DNA-dependent STING pathway, as opposed to external DNA. Although there was no difference in the quantity of DNA packaged in the different EVs (Fig 4E), we did find that Ca and Sc EV DNA had lower guanine-cytosine (GC) contents than Cn and Af EVs (Fig 4F). Low GC composition in DNA has been linked to DNA structure (i.e., nucleosome formation) and stability in some cells, potentially explaining the more stimulatory nature of the DNA carried in Ca and Sc EVs [36–39]. However, we also noted that whole C. albicans and S. cerevisiae genomes have lower GC contents than C. neoformans and A. fumigatus genomes [40–42] (Fig 4G). Therefore, this altered GC content might simply be a reflection of the DNA found in the parent fungal genome and not directly affect DNA recognition by cGAS or other immune sensors. Future studies examining the accessibility of the DNA carried in these diverse fungal EVs will elucidate the relevance of these findings.

Outer structural layers on EVs impact endocytosis by macrophages

As we observed similarities in EV size and morphology from four different fungi, we concluded that other EV attributes may account for the observed differences in endocytosis patterns by macrophages and subsequent degrees of STING pathway activation. One potential contributor may be physical differences in the EV membrane characteristics. To explore specific physical properties that might drive differences between EV internalization by immune cells and subsequent cGAS-STING pathway activation, we assessed the most external outer structural layers present on fungal EVs that exhibited reduced internalization by macrophages, specifically the GXM polysaccharide capsule that surrounds C. neoformans yeast and the rodlet layer surrounding A. fumigatus conidia. We added C. neoformans and Cn EVs to WT macrophages and used confocal microscopy to identify the presence of GXM on both WT yeast cells and their EVs (Fig 5A) as previously described [43,44]. We also used cap59∆ yeast, a C. neoformans strain lacking a GXM layer [45,46]. We confirmed the absence of GXM on whole cap59∆ yeasts and their EVs by confocal microscopy (Fig 5A). To assess the effect of the rodlet layer on EV endocytosis, we used a mutant strain of A. fumigatus that produces conidia that lack this outermost rodlet layer, ΔrodA [23]. With a RodA polyclonal antibody (Epigentek A55804-050), we confirmed decreased RodA protein levels in the ΔrodA mutant EVs (Fig 5B lane 3). We note that the observed RodA protein with WT Af EVs (Fig 5B lane 2) does not definitively prove the presence of rodlet layer on the surface, but rather a general presence of the RodA protein that is notably absent in the ΔrodA mutant EVs. To investigate whether the RodA protein populates the surface of Af EVs (similarly to Af conidia), future studies will need to examine closely the outer structures of these EVs. To rule out differences other than the extracellular structures on Cn EVs and Af EVs compared to cap59∆ EVs and ΔrodA EVs, we compared important characteristics between EVs derived from the WT and mutant Cn and Af strains. We compared the concentrations of EVs isolated from standard preps, EV size, and EV internal DNA content, and found no statistical differences among these characteristics (S5A–C Fig). We then stimulated macrophages with DiI stained-EVs and used flow cytometry to assess the abilities of these mutant EVs to be endocytosed by macrophages. We note that although the mutant EVs have different membranes from their WT counterparts, we confirmed successful lipophilic staining by flow cytometry. Furthermore, we found that those without protective surface layers, cap59∆ EVs and ΔrodA EVs, exhibited significantly increased endocytosis compared to their respective WT EVs (Fig 5C and 5D). Furthermore, incubation of the cap59∆ EVs or ΔrodA EVs with WT macrophages yielded increased viperin expression compared to their respective WT EV counterparts (Fig 5E and 5F). Therefore, not only have we identified a species-specificity to the activation of the cGAS-STING pathway and downstream innate immune activation, but we have also uncovered a role for protective surface layers on Cn and Af EVs as suppressor agents that prevent EV entry and subsequently dampen STING pathway activation. These findings further elucidate the mechanism by which EVs released by pathogenic fungi rewire the innate immune response.

Fig 5. Outer structural layers on EVs inhibit endocytosis and STING pathway activation.

A, Representative confocal microscopy showing GXM layer present on WT C. neoformans (H99) yeast and EVs but not on cap59∆ yeast and EVs. 1x10⁷ WT yeasts, 1x10⁷ cap59∆ yeasts, and 5x10¹⁰ WT and cap59∆ EVs were added to the respective stimulations. Scale bar = 10µm. B, Representative immunoblot of RodA (28.3 kDa) on Af EVs compared to mutant ΔrodA EVs. C, Percent endocytosis of Cn EVs and cap59∆ EVs by WT macrophages. (5x10¹⁰ EVs/mL added per stimulation), n = 2. D, Percent endocytosis of Af EVs and ΔrodA EVs by the total number of WT macrophages (5x10¹⁰ EVs/mL added per stimulation), n = 2. E, Immunoblot of viperin and actin in WT macrophages stimulated by PBS, Cn EVs, cap59∆ EVs, and the positive control for yeast, Ca EVs (5x10¹⁰ EVs/mL were added per stimulation). F, Immunoblot of viperin and actin in WT macrophages stimulated by PBS, Af EVs, and ΔrodA EVs (5x10¹⁰ EVs/mL were added per stimulation), and positive control for Aspergillus: 2.5µg cGAMP.

Discussion

Here, we uncovered that specific surface components of EVs disrupt their endocytosis by macrophages, subsequently disrupting EV-mediated cGAS-STING pathway activation. EVs from C. neoformans and A. fumigatus have protective outer layers composed of polysaccharides and hydrophobins, respectively. These external structures inhibit endocytosis and subsequently inhibit cGAS-STING activation. While the immunomodulatory properties of surface structures have been studied for whole fungal organisms [47–54], our findings provide novel insights connecting surface coatings and immune evasion in the context of fungal EVs. GXM and galactoxylolomannan (GXMGal), major components of the outer capsule of C. neoformans, modulate both the innate and adaptive immune systems, including apoptosis induction of important inflammatory cells. For example, GXM and GXMGal from C. neoformans induce macrophage apoptosis both in vitro and in vivo, a phenomenon mediated by Fas/FasL interactions [47]. GXMGal also interacts with glycoreceptors, triggering apoptosis of T cells via caspase-8 [48,49]. Beyond influencing immune cell apoptosis, these capsular polysaccharides from various Cryptococcus species disrupt phagocyte migration, hinder phagocytosis, and inhibit NET production by human neutrophils, all contributing to immune evasion abilities of the encapsulated fungus [50–52]. Similarly, surface proteins that form a hydrophobic layer over the surface of Aspergillus conidia play a key role in immune evasion tactics. WT Aspergillus conidia do not activate human dendritic cells and murine alveolar macrophages in vitro, whereas ΔrodA mutant conidia lacking the outer hydrophobic layer are highly stimulatory [53]. The rodlet layer is also responsible for the dampening of pro-inflammatory signaling such as NF-κB and cytokine production via Dectin-1 and Dectin-2 [54]. Additionally, in a murine infection model, mice infected with WT A. fumigatus resulted in a greater fungal burden compared to mice infected with ΔrodA [54]. Furthermore, curvature-sensing peptides are used in EV detection procedures, where affinities of these peptides to bacterial EVs depend on the presence or absence of capsular polysaccharides on the EVs [55]. Further studies elucidating extracellular structures of microbial EVs and how they impact immunity will be important for gaining insight into host-pathogen relationships and developing tools for pathogen detection techniques.

Building on our previously published results demonstrating that dynamin-dependent endocytosis is necessary for the induction of viperin in macrophages exposed to C. albicans EVs [14], this study revealed a species-dependent EV endocytosis efficiency that directly correlates with activation of the cGAS-STING pathway. These observations are likely due to the protective and shielding nature of surface components of poorly stimulating fungal EVs. While this is the first time diverse fungal EV internalization by macrophages has been studied in the context of endocytosis and innate immune activation, the necessity for endocytosis of pathogens to elicit immune signaling has been studied extensively [56–58]. Dynamin-dependent endocytosis is necessary for inflammatory cytokine (IFNγ, IL-6, and CCL5) induction in response to LPS [59]. Dynamin-dependent endocytosis is also important for LPS-induced TLR4 internalization and initiation of downstream signaling [60]. Additionally, dynasore-treated phagocytes have reduced induction of type III IFNs and decreased cell surface expression of CD86 and HLA class II molecules in Streptococcus thermophilus-stimulated cells [59]. It is important to note that dynasore can work to disrupt dynamin-dependent endocytosis through inhibition of the GTPase activity of dynamin, blocking early constriction and fission, or by reducing plasma membrane cholesterol, which inhibits membrane mobilization [61]. Therefore, it could be argued that the effects our study and others observed in dynasore-treated cells could be due to either altered early invagination or changes in macropinocytosis. Future studies examining the specific endocytic properties affected by dynasore treated macrophages will provide further insights into fungal EV internalization. However, these nuances do not diminish our overall finding that dynamin-dependent endocytosis is needed for efficient uptake of fungal EVs and downstream innate immune signaling and that this efficiency is affected by the outer surface layers of fungal EVs.

Importantly, initiation of cGAS-STING signaling is a critical priming step in the immune response to viral or foreign DNA [62]. Recent evidence shows that cGAS binds to DNA via a “dual signal” model in which the subcellular localization of this interaction and the downstream immune activity initiated is dependent on the type of DNA (i.e., foreign or self). Specifically, endocytosis of viral or other foreign DNA, an initial priming step occurs where spleen tyrosine kinase (SYK) and cGAS translocate to endosomes, where cGAS binds to the DNA and initiates a robust immune response. Alternatively, in response to the presence of self-DNA, cGAS localizes to the plasma membrane, which initiates a mild IFN response. Our studies did not investigate whether fungal EVs and EV DNA initiate this dual signal process and future experiments assessing the role of SYK in our model will elucidate whether this first priming step is essential following endocytosis of fungal EVs [63]. Furthermore, future in vivo studies would be beneficial for determining the biological relevance of these EVs in an infection setting. Since there are no studies that confirm physiological concentrations of EVs during fungal infection, in this study we normalized the number of EVs added to each macrophage stimulation. By adding the same number of EVs across macrophage wells, we created a controlled experiment. However, future studies investigating physiological levels of EVs in infection would enhance further investigations into the immunomodulatory properties of EVs.

Finally, our studies further elucidate the differing abilities of fungal EVs to activate the STING pathway and type I IFN signaling. The STING pathway has historically been studied in the context of viral infections, bacterial infections, and autoimmunity [64]. However, there are increasing reports of interactions between this pathway and fungal organisms. For example, C. albicans and Ca EVs activate the STING pathway and STING-deficient mice have improved survival following a C. albicans infection compared to their WT counterparts [14,6]. A. fumigatus also elicits inflammatory responses in the host responsible for fungal keratitis via cGAS and STING [7]. As EVs gain more attention as important mediators of pathogenic infection, investigating their immunostimulatory properties alongside the whole organism will be informative. Expanding our research questions to pathogenic fungi beyond those in our current study would contribute to a more comprehensive understanding of fungal properties that impact host-pathogen interactions. Our novel findings on the in vitro immunomodulatory abilities of Ca EVs, Sc EVs, Cn EVs, and Af EVs inform early interactions with the innate immune system and provide details on infection mechanisms utilized by these clinically significant fungal pathogens.

Methods

Cell culture and macrophage cell line generation

All macrophage lines were immortalized from C57BL/6 murine bone marrow cells. All immortalized bone marrow-derived macrophages (BMDM) were cultured in complete Dulbecco’s modified eagle medium (cDMEM) with added 10% FBS, penicillin and streptomycin (Pen+Strep), L-glutamine and 1M hydroxyethelpiperazine ethane sulfonic acid at 37°C and 5% CO₂. Macrophages were washed with PBS and lifted by 0.05% trypsin in a 1:10 dilution. The WT immortalized BMDM were gifted by Douglas Golenbock (University of Massachusetts Medical School), and the cGAS^-/- and Sting^-/- macrophages were generated as previously described [65], as were the cGAS-GFP expressing macrophages.

Fungal strains

C. albicans is strain SC5314. S. cerevisiae is strain S228C from ATCC 204508. C. neoformans is strain H99 from John Perfect (Duke University). A. fumigatus is strain Af293 from Robert Cramer (Dartmouth College). cap59∆ mutant is derived from parent background H99. ΔrodA mutant is derived from parent background G10 and from Jean Paul-Latge (Institut Pasteur, Paris, France). We note the difference in strain background between the WT Af293 and the ΔrodA G10, however there does not yet exist an Af293 strain with the RodA gene deleted. Future studies will assess EVs from all strains of A. fumigatus and their associated rodlet layers in this process.

EV isolation and quantification

EVs from C. albicans (SC5314), S. cerevisiae (S228C), and C. neoformans (H99) were isolated and purified from YPD agar lawns. Each species was inoculated from a frozen stock in 5mL of YPD broth and incubated overnight at 30°C on a rotating wheel. Cells were washed in PBS, counted, and plated at 3.5x10⁷ cells/mL in 300µL per YPD agar plate, then spread by glass beads to form a lawn. A standard prep included 20 plates (STEMCELL Technologies 100mm x 15mm culture dish, non-treated) for each species. Plates were incubated at 30°C for 48h. The cell lawns were carefully scraped from the plates and transferred to conical tubes with 30mL of PBS. The conicals were centrifuged at 1,500g for 10 minutes to pellet the cellular debris. The EV-rich supernatant was filtered through 0.45µm filter. The filtered supernatants were collected (in 17mL Beckman Coulter centrifuge tubes Ref 337986) and ultracentrifuged at 100,000g for 1h at 4°C to pellet the EVs. The supernatants were aspirated, and the pellets were resuspended in 300µL each. EV sizes and concentrations were quantified using NanoSight LM10 & Nanoparticle Tracking Analysis (NTA) 3.4 analytical software (NanoSight Ltd., Minton Park, Amesbury, Wiltshire SP4 7RT, UK) and stored at -20°C for further use.

EVs from A. fumigatus (Af293) were isolated from liquid SBD broth. Af293 from a frozen stock was inoculated on GMM slants and incubated at 37°C for 72h and the conidia were harvested. Harvested conidia were inoculated at 2x10⁸ conidia/L into one liter of liquid SBD broth and incubated at 37°C for 96h at constant shaking at 150 rpm (New Brunswick Scientific I2500 Incubator Shaker). The culture was filtered through a miracloth-lined funnel and then through a 0.22µm filter overnight at 4°C. The filtered supernatant was concentrated by the Vivaflow 50 Concentrator with a 30kDa cut-off membrane. The concentrated supernatant was ultracentrifuged at 100,000g (in 17mL Beckman Coulter centrifuge tubes Ref 337986) for 1 hour at 4°C to pellet the EVs. The supernatants were aspirated and the pellets were resuspended in 300µL each. EV sizes and concentrations were quantified using NanoSight LM10 & Nanoparticle Tracking Analysis (NTA) 3.4 analytical software (NanoSight Ltd.,Minton Park, Amesbury, Wiltshire SP4 7RT, UK) and stored at -20°C for further use.

Endocytosis analysis

WT immortalized macrophages were seeded at 1.5x10⁵ cells/mL in 1mL in a 12-well tissue culture treated plate and incubated overnight at 37°C and 5% CO₂. We fluorescently labeled the lipid membranes of 5x10¹⁰ EVs from C. albicans, S. cerevisiae, C. neoformans, and A. fumigatus with DiI (DiI lipophilic stain (Invitrogen D282). 2μL of 1:100 dilutions of DiI in 100% ethanol were added to the EVs, followed by a 30 min incubation at RT. The EVs were then centrifuged at 100,000xG for 1h at 4°C (in 1.5mL Beckman Coulter microcentrifuge tubes REF357448). This allowed for the EVs to pellet, and for any excess stain in the supernatant to be discarded. The supernatant was aspirated, and then the pellet was resuspended in 100μL of PBS to be co-incubated with the WT macrophages. A PBS-treated well was used for negative control. The stimulated macrophages were then incubated at 37°C and 5% CO₂ for 3 h. After incubation, we aspirated old media, washed with 1mL of PBS, aspirated and added 0.5mL of 0.05% trypsin to allow cells to detach from the well. We then centrifuged the samples to pellet the cells (1200rpm for 5 minutes), aspirated to remove the supernatant, and resuspended the pellet in 300μL PBS. Samples were filtered into FACS tubes to analyze the percent endocytosed EVs located in macrophages using a Cytek Aurora Spectral flow cytometer (Cytek Biosciences, California, USA). A total of 50,000 cells was collected per sample, recorded using the Spectroflo (version 3.0.3, Cytek Biosciences). Finally, endocytosis was determined using negative unloaded macrophages and appropriate positive controls in duplicate. Data were gated and analyzed using FlowJo (version 10.10.0, BD Biosciences, New Jersey, USA).

LEGENDplex multiplex analysis

To assess cytokine induction by macrophages co-cultured with EVs, we used a LEGENDplex 13-plex Mouse Inflammation Panel kit (cat. 740150). Macrophages were seeded at 1.5x10⁵ cells/mL in DMEM in a 12-well TC-treated plate overnight at 37°C and 5% CO₂. Macrophages were stimulated with PBS, transfected cGAMP (2.5µg cGAMP), or pathogen-derived EVs at 1x10¹⁰ EVs/mL. Transfection for cGAMP was prepared with Lipofectamine 3000 reagents (Invitrogen L3000015). Supernatants from the co-incubations were collected after a 6h incubation at 37°C and 5% CO₂, and the LEGENDplex Mouse Inflammation Panel protocol was followed accordingly to the kit instructions. The concentration of cytokines in each sample was determined using a BD FACSCelest Flow Cytometer. This multiplex ELISA was completed in duplicate, and data were analyzed using the LEGENDplex Data Analysis Software Suite and PRISM10 software version 10.2.3 (GraphPad Software).

Neutrophil Transmigration Assays

WT immortalized macrophages were seeded at 1.5x10⁵ cells/mL in 1mL in a 12-well tissue culture treated plate and incubated overnight at 37°C and 5% CO₂. Macrophages were stimulated with 5x10¹⁰ EVs/mL or with cGAMP as a positive control. After 5 h of co-incubation at 37°C and 5% CO₂, 500µL of the supernatant was transferred to the receiving plate of 24-well. N-Formylmethionine-leucyl-phenylalanine (fMLP) was added to one well of the receiving plate as a positive control. 3μm transwell inserts (Corning Product No. 3415) were placed over the wells and 1x10⁶ murine neutrophils (immortalized cell line WT Cas9-ER-HoxB8-GMP), isolated as previously described [66], were added to the transwells in cRPMI media. Neutrophils were allowed to migrate for 2 h at 37°C and 5% CO₂. The cells in the receiver plate were then lysed with Triton X while shaking for 20 min, and then treated with citrate buffer. 100µL of the samples were added in triplicate to a 96-well plate, and then treated with ABTS solution made from Di water, 2,2’-Azino-bis(3-ethylbenzo-thiazoline-6-sulfonic acid) diammonium salt (Sigma Aldrich) and citrate buffer. Neutrophil standards were also subjected to the same conditions. Absorbance was detected with an i3X spectrophotometer (Molecular Devices, LLC) at 405. Data were analyzed using PRISM10 software version 10.2.3 (GraphPad Software).

Immunoblot analysis

For pathogen-derived EV and EV DNA stimulation experiments, immortalized macrophages were seeded at 1.5x10⁵ cells/mL in a 12-well or at 1.2x10⁵ cells/mL in a 48-well TC-treated plate overnight for cell adhesion at 37°C and 5% CO₂. Pathogen-derived EVs were added to wells at 5x10¹⁰ EVs/mL or 1x10¹⁰ EVs/mL, and EV DNA was transfected at 300ng/mL. PBS was used as a negative control and cGAMP as a positive control. cGAMP and EV DNA were transfected using Lipofectamine 3000 reagents (Invitrogen L3000015). After a 6h incubation at 37°C and 5% CO₂, lysates were collected by aspirating the media, washing cells with PBS, and then lysing cells with mammalian protein extraction reagent lysis buffer (Thermo Scientific, no. 78501) with sodium orthovanadate and protease inhibitors and shaking for 5 min. Lysates were collected and centrifuged at 14,000g for 5 min at 4°C. Lysates were transferred to new Eppendorf tubes with 4x NuPage lithium dodecyl sulfate loading buffer and 10x NuPage reducing agent. Protein expression was identified by western blot with 4–12% NuPage gel with 2-[N-morpholino]ethanesulfonic acid running buffer (NuPage gels, Thermo Fisher Scientific) at 180V for 1h, and transferred to methanol-activated polyvinylidene difluoride membrane (Perkin Elmer, Waltham, MA) with transfer buffer (0.025 M Tris, 0.192 M glycine and 20% methanol) and electrophoretic transfer at 100V for 1h. Membranes were blocked in 5% milk in 1X tris-buffered saline with Tween 20 (TBST) shaking for 1h at room temperature. For viperin detection, membranes were incubated rotating for 1 hour at room temperature in 1% BSA and a 1:5,000 dilution of primary antibody (Sigma MABF106) in 1X TBST. For TBK1 (Cell Signaling 3504) detection, membranes were incubated rotating for 1h at room temperature in 1% BSA and a 1:1,000 dilution of primary antibody in 1X TBST. For IRF3 (Cell Signaling 4302), membranes were incubated rotating for 1h at room temperature in 1% BSA and a 1:1,000 dilution of primary antibody in 1X TBST. For phospho-IRF3 (Cell Signaling E6F7Q) and phospho-TBK-1 (Cell Signaling D52C2) detection, membranes were incubated rotating overnight at 4°C in 5% BSA and 1:1,000 dilution of primary antibody in 1X TBST. After incubation of the primary antibody, all membranes were washed 3x with 1X TBST before incubation with secondary antibody swine anti-rabbit horseradish peroxidase-conjugated antibody at 1:2,000 (Agilent DAKO, P0399) or secondary peroxidase AffiniPure goat anti-mouse IgG (H and L) (Jackson ImmunoResearch) at 1:2,000 dilution in 1% milk in 1X TBST for 1h at room temperature. Membranes were then washed 3x with 1X TBST and then prepared with Western Lightning Plus ECL chemiluminescent substrate (Perkin Elmer) on Kodak BioMax XAR film (MilliporeSigma) for signal detection. Film sheets were scanned for electronic upload. Any contrast adjustments were applied evenly to the entire image and adhered to standards set forth by the scientific community. All reported Western blots/immunoblots were repeated in at least biological duplicate. Quantification of protein expression from western blots was performed with FIJI software version 1.53.

EV DNA isolation

DNA was extracted from the isolated pathogen-derived EVs using the MasterPure Yeast DNA Purification Kit (Lucigen). Approximately 5x10¹⁰ EVs were added to 300 μL of yeast cell lysis solution and incubated at 65°C for 15 min. The samples were placed on ice for 5 min, and then 150 μL of milk protein concentrate protein precipitation reagent was added, vortex mixed for 10 seconds, and centrifuged in a table-top microcentrifuge for 10 min. The supernatants were transferred to new Eppendorf tubes, followed by the addition of 500 μL of isopropanol and mixed thoroughly by inversion. DNA was collected by centrifugation in a table-top microcentrifuge for 10 min at the maximum speed. The pellets containing the environmental DNA were washed with 500 μl of 70% ethanol, dried briefly at 42°C, and added to 50 μL of 1X TE buffer. DNA concentration was quantified by the Quant-iT PicoGreen dsDNA Assay Kit (Invitrogen) with fluorescent readings performed on an i3X spectrophotometer (Molecular Devices, LLC).

EV DNA sequencing

DNA from fungal EVs was extracted using Masterpure genomic extraction kit (above). DNA was submitted to the NextGen sequencing core at MGH and sequencing was done on a NextSeq Flowcell (P3 Single End 50 Run per million QC for Lanes after the First Lane). Reads that mapped to the reference genome were extracted with samtools view -h -F 4 and exported to fastq format with samtools bam2fq. These DNA sequences are publicly available on the SRA (sequence read archive) repository (Submission ID: SUB15319812, BioProject ID: PRJNA1262535, https://dataview.ncbi.nlm.nih.gov/object/PRJNA1262535). The GC content analysis was done only the reads that mapped to the expected fungal reference genome. GC content was calculated from these reads by taking the sum of the counts of all G and C residues divided by the total number of bases.

Confocal microscopy

For the cGAS localization experiment, cGAS-GFP expressing macrophages were seeded at 1x10⁵ cells/mL in Nunc lab-tek II chambered coverglass (Thermo Scientific) in 500µL of cDMEM and incubated overnight at 37°C and 5% CO₂. Macrophages were stimulated with fungal EVs at a concentration of 1.6x10¹⁰ cells/mL or stimulated with PBS. Before stimulation, 5μM DiI lipophilic stain (Invitrogen D282) was added to the EVs for 30 min at room temperature. The EVs were then ultracentrifuged at 100,000g for 1h at 4°C. A PBS sample also underwent the same staining and ultracentrifuge procedure. The macrophages were washed with PBS and refreshed with 500µL DMEM without FBS to rid of any serum-derived mammalian EVs. Finally, the macrophages were stimulated with the DiI labeled fungal EVs or the DiI-stained PBS and allowed to incubate for 3h at 37°C and 5% CO₂. After incubation and phagocytosis of EVs, the cells were imaged on a Nikon Inverted Microscope Eclipse Ti-E with a CSU-X1 confocal spinning disk head (Yokogawa), and a 6-laser Cairn Multiline launch equipped with 405nm, 445nm, 488nm, 515nm, 561nm, and 647nm solid-state lasers was used to excite the sample. The objective was a 100x high-numerical aperture objective (Nikon, 1003, 1.49 numerical aperture, oil immersion) and an EMCCD camera (Hamamatsu Photonics). For image acquisition, we used MetaMorph software version 7.10.5.476 (Molecular Devices). The photos were cropped in FIJI software and organized in Adobe Illustrator, version 28.5 (Adobe Systems). The brightness/contrast values were set to the recommended “auto” levels for each individual image. As these cells were cGAS-GFP expressing macrophages, green fluorescence indicated the localization of cGAS. The localization of cGAS in the macrophages was quantified as “nuclear” if cGAS was only present in the nucleus, “cytosolic” if cGAS was only present in the cytoplasm, or “mixed” if cGAS was only present in both the nucleus and cytoplasm.

To identify macrophage nuclei, cGAS-GFP expressing macrophages were seeded at 1x10⁵ cells/mL in Nunc lab-tek II chambered coverglass (Thermo Scientific) in 500µL of cDMEM and incubated overnight at 37°C and 5% CO₂. The next morning, we added 150uL of a 1:5,000 dilution of DAPI (4’,6-diamidino-2-phenylindole) stock solution (Invitrogen D3571) at 5mg/mL to the appropriate wells, followed by a three-minute incubation at room temperature. Images were then captured by confocal microscopy and analyzed as previously described.

For the GXM staining, WT macrophages were seeded at 1x10⁵ cell/mL in 500µL of cDMEM in Nunc lab-tek II chambered coverglass (Thermo Scientific) and incubated overnight at 37°C and 5% CO₂. 10uL of antiGXM antibody (Millipore Sigma Cat. No. MABF2069) was added to 1x10⁷ H99 cells and 1x10⁶ cap59∆ yeast cells. 10uL of antiGXM antibody (Millipore Sigma Cat. No. MABF2069) was also added to 5x10¹⁰ H99 EVs and 5x10¹⁰ cap59∆ EVs, followed by a 1-hour incubation of all four of these samples at 37°C while shaking. 5uL of secondary anti-mouse AF488 (ThermoFisher Scientific Ref A32723 Lot WF323912) antibody was then added to the samples, followed by another 1-hour incubation at 37°C while shaking. Samples were ultracentrifuged for 1 hour at 4°C at 100,000g. Supernatant was removed and the pellets were resuspended in 50uL of PBS which was then added to the plated macrophages and co-incubated for 3 hours. Images were then captured by confocal microscopy and analyzed as previously described.

RodA detection

We performed an immunoblot analysis using an unconjugated Polyclonal Rabbit anti-Aspergillus fumigatus rodA/Hydrophobin Antibody to probe for RodA on Aspergillus EVs. We loaded either PBS alone (negative control), 1 x10¹⁰ Af293 EVs, and 1 x10¹⁰ ΔrodA EVs (boiled and combined with 4x NuPage lithium dodecyl sulfate loading buffer and 10x NuPage reducing agent) onto a 4–12% NuPage gel and transferred/blocked as above. For RodA detection we used the rodA Polyclonal Antibody, FITC Conjugated, Raised in Rabbit, from Epigentek A55804-050 at a 1:1000 dilution in 5% BSA rotating overnight at 4°C. Membranes were washed and developed as above. RodA ~ 28 kDa. No loading control was performed as EVs alone do not contain sufficient amounts of beta actin.

TEM imaging (Negative staining)

10-20µL of EV samples at approximately 1x10¹¹ EVs/mL were added onto 200 mesh Formvar/carbon coated nickel grids and allowed to adsorb 15min. Grids were blotted to remove excess suspension, rinsed briefly with filtered distilled deionized water, and contrast-stained for 10min in a tylose/uranyl acetate solution on ice. Grids were blotted again and allowed to air dry prior to analysis. Examination of preparations was done using a JEOL JEM 1011 transmission electron microscope at 80kV. Images were collected using an AMT digital imaging system with proprietary image capture software (Advanced Microscopy Techniques, Danvers, Massachusetts).

Supporting information

S1 Fig. Isolation of fungal EVs.

A, Isolation procedures of EVs from C. albicans, S. cerevisiae, and C. neoformans. B, Isolation procedures of EVs from A. fumigatus cultures. Created in BioRender. Brown, H. (2025) https://BioRender.com/g11i747.

(TIF)

S2 Fig. Endocytosis as EV internalization mechanism by macrophages.

A, Flow cytometry gating strategy for internalization of DiI-labeled fungal EVs by WT macrophages. B, Flow cytometry gating strategy for endocytosis of DiI-labeled fungal EVs by WT macrophages treated with DMSO or 100µM Dynasore. C, Percent endocytosis of Cn EVs with murine macrophages treated with DMSO or 100µM dynasore. Significance assessed by a two-way ANOVA and uncorrected Fisher’s LSD test, ***p = 0.0002 vs respective DMSO controls, n = 2. (5x10¹⁰ EVs/mL added per stimulation). D, Percent endocytosis of Af EVs with murine macrophages treated with DMSO or 100µM dynasore. Significance assessed by a two-way ANOVA and uncorrected Fisher’s LSD test, ****p=<0.0001 vs respective DMSO controls, n = 3. (5x10¹⁰ EVs/mL added per stimulation).

(PDF)

S3 Fig. Immunomodulatory properties of fungal EVs.

A, Heat maps of TNFα and IFNβ secreted by WT and Sting^-/- macrophages when stimulated by PBS, cGAMP, Ca EVs, Sc EVs, Cn EVs, and Af EVs (EVs added at 1x10¹⁰ EVs/mL). B, Heat map of 10 additional cytokines secreted by WT and Sting^-/- macrophages when stimulated by PBS, cGAMP, Ca EVs, Sc EVs, Cn EVs, and Af EVs (EVs added at 1x10¹⁰ EVs/mL). C, Immunoblots of viperin and actin in WT and Sting^-/- macrophages stimulated by PBS, 2.5µg cGAMP, Ca EVs, or Sc EVs (EVs added at 1x10¹⁰ EVs/mL). D, Immunoblot of viperin and actin in WT and Sting^-/- macrophages stimulated by PBS, 2.5µg cGAMP, Cn EVs, or Af EVs (EVs added at 1x10¹⁰ EVs/mL). E, Immunoblot of viperin and actin in WT macrophages stimulated by PBS, benzonase-treated EVs, and untreated EVs (EVs added at 5x10¹⁰ EVs/mL).

(TIF)

S4 Fig. cGAS localization in unstimulated macrophages.

A, Localization of the nuclei (panel 2, blue) and cGAS (panel 3, green) of cGAS-GFP expressing macrophages in an unstimulated state. Macrophages were stained with DAPI for visualization of the nucleus.

(TIF)

S5 Fig. Mutant EV characteristics compared to WT EVs.

A, The -log₁₀ values of EV concentration from standard isolation preps of cap59∆ EVs and ΔrodA EVs compared to Cn EVs and Af EVs. Significance assessed using an ordinary one-way ANOVA and Tukey’s multiple comparisons test with no significance determined, n ≥ 2. B, The median and mode diameters of cap59∆ EVs and ΔrodA EVs. Significance assessed using an ordinary one-way ANOVA and Tukey’s multiple comparisons test with no significance determined, n ≥ 2. C, The average DNA concentration (pg per 5x10¹⁰ EVs) of cap59∆ EVs and ΔrodA EVs compared to Cn EVs and Af EVs. Significance assessed using an ordinary one-way ANOVA and Tukey’s multiple comparisons test with no significance determined, n ≥ 3.

(TIF)

Data Availability

This is a confirmation that all data (including data points used to generate graphs, standard error bars, and statistical significance) are displayed clearly in each figure in the manuscript. We have also made the fungal EV DNA sequences discussed in this study publicly available on the SRA (sequence read archive) database (Submission ID: SUB15319812, BioProject ID: PRJNA1262535).

Funding Statement

The work in the Vyas Lab is supported by grants from the NIH: R01AI150181, R01AI136529, R21AI152499 (JMV). JN and DZM are supported by R01AI171093 and R21AI156104. HBH is supported by BroadIgnite Fellowship. Electron microscopy was performed in the Massachusetts General Hospital (MGH) Microscopy Core of the Program in Membrane Biology, which is partially supported by an Inflammatory Bowel Disease Grant P30DK043351 and a Boston Area Diabetes and Endocrinology Research Center (BADERC) Award P30DK135043. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.