Summary
A key component of insect immunity is melanin encapsulation of microbes. Melanization is also a part of an immune process known as nodulation, which occurs when insect hemocytes surround microbes and produce melanin. Insect nodules are analogous to mammalian immune granulomas. Melanin is believed to kill microbes through the production of toxic intermediates and oxidative damage. However, it is unclear to what extent immune melanin is directly fungicidal during infections of insect hosts. We reported previously that C. neoformans cells are encapsulated with host-derived melanin within hemocyte nodules. Here we report an association between melanin-based immune responses by Galleria mellonella wax moth larvae and fungal cell death of C. neoformans during infection. To monitor melanization in situ, we applied a tissue-clearing technique to G. mellonella larvae, revealing that nodulation occurs throughout the organism. Further, we developed a protocol for time-lapse microscopy of extracted hemolymph following exposure to fungal cells, which allowed us to visualize and quantify the kinetics of the melanin-based immune response. Using this technique, we found evidence that cryptococcal melanins and laccase enhance immune melanization in hemolymph. We used these techniques to also study the fungal pathogen Candida albicans infections of G. mellonella. We find that the yeast form of C. albicans was the primary targets of host melanization, while filamentous structures were melanin-evasive. Approximately 23% of melanin-encapsulated C. albicans yeast survive and break through the encapsulation. Overall, our results provide direct evidence that the melanization reaction functions as a direct antifungal mechanism in insect hosts.
Introduction
Insects occupy essential niches in global ecosystems, including many that directly affect human health and survival 1. In addition, insects serve as powerful model systems for infectious disease research, and help to reduce reliance on vertebrates recommended by “3R” -Replace, Reduce, and Refine – programs 2. Insects are also targeted by environmental pathogens and have evolved complex immune mechanisms that partially overlap with mammalian innate immunity. Understanding the dynamics of insect-pathogen interactions and the factors involved is vital to both ensure ecosystem stability and establish invertebrate immunological models in research.
Fungi are an important class of pathogens for insects, and emerging fungal pathogens are predicted to become bigger threats to human health and agriculture in the coming years 3,4. Consequently, studying host-fungal interactions using insect models is important and timely. Although insects do not produce antibodies or other mammal-like adaptive immune responses, the antifungal immune defenses of insects involve cell-mediated and humoral innate immune processes 5. Hemocytes, the immune cells of invertebrates which circulate in the hemolymph, have roles comparable to macrophages and neutrophils in mammals. Hemocytes are responsible for clearance of fungi via phagocytosis, release of extracellular damaging reactive oxygen species (ROS) and inflammatory molecules, and the creation of granuloma-like structures through a process called nodulation 6. During nodulation, hemocytes surround the microbe and form an aggregate of insect cells, within which, clotting factors, immune enzymes, and immune complexes are released and activated 6–8. These structures immobilize the fungus and lead to its destruction. Also, during infection, the production of prostaglandins by the plasmatocyte subset of hemocytes in Lepidopteran species cause the lysis of other hemocytes called oenocytoids. The lysis of oenocytoid cells results in the release of antimicrobial peptides, signaling molecules, and enzymes important to immune function 9–11. One class of host enzymes that are often released and activated during oenocytoid lysis and nodulation are phenoloxidases (PO) 9,11. POs are enzymes responsible for converting catecholamines in the hemolymph into melanin 12. Melanin is a the black-brown pigment that is an important component of insect immune defense and wound repair 13. Melanization produces oxidative species and cytotoxic intermediates that are hypothesized to result in the death of the microbe 12,14. Additionally, melanin may act as a physical barrier, restricting gas exchange and nutrient uptake, and thus prevent fungal replication and dissemination to other tissues 15. At this time, in vitro evidence strongly links PO activity and resulting melanin intermediates with killing of fungi, bacteria, and viruses 16–18, but comparable direct evidence for the microbicidal effect of POs and their toxic intermediates in vivo during insect infections is challenging to measure directly. Consequently, obtaining direct evidence that the process of melanization is fungicidal in vivo is important for establishing insect melanin as an important mechanism for clearing fungal infections.
Larvae of the wax moth Galleria mellonella are commonly used as a model organism for studying fungal pathogenesis 5. G. mellonella larvae are readily available in large numbers at low cost. Their larger size (2-3 cm) relative to other model insects such as Drosophila melanogaster makes them amenable to research approaches requiring larger volumes of hemolymph, insect hemocytes, and soluble immune factors. The study of G. mellonella hemolymph can prove valuable for understanding the insect’s immune response to infection and stress. G. mellonella are also commonly used as a model for studying mammalian pathogens, including human pathogenic fungi Cryptococcus neoformans and Candida albicans 19–21. While G. mellonella is a model for mammalian fungal infections because of similarities between the G. mellonella immune responses and the mammalian innate immune responses 5, a more thorough understanding of the insect immune response is needed to fully benefit from studying host-microbe interactions in G. mellonella.
The differences between mammalian versus insect hosts also provide important new insights into host-microbe interactions and mechanisms of fungal virulence factors 5,19,20,22,23. For example, laccase, a fungal enzyme that oxidizes mammalian and insect catecholamines, is an important virulence factor in both hosts but, seemingly by distinct and diverse mechanisms 19,24,25. In insects, fungal laccase appears to oxidize and deplete host catecholamines required for encapsulating the fungus in melanin, thus weakening the host immune response. Fungal laccases also help detoxify reactive oxygen species that form during insect immune processes 24. In contrast, during mammalian infection, fungal laccase enhances production of fungal melanin to evade key mammalian immune defenses 26. Thus, fungal melanins increase virulence in mammals, but decrease virulence in G. mellonella 27–29. These seemingly different and opposite roles in which fungal melanins interact with mammalian and insect hosts is unexplored and as of now unexplained in literature.
In this study, we describe the first direct evidence that the melanin-based immune response in vivo is fungicidal against C. neoformans. Our data show a direct link between melanin encapsulation during infection within the G. mellonella larvae and fungal death by visualizing death using an endogenously expressed GFP-based fungal viability assay. We then used a series of in situ, in vivo, and in vitro methods to study the melanin-based immune response of G. mellonella larvae. For in situ experiments, we modified a previously published tissue-clearing protocol to visualize melanized nodules and their tissue specificity, or lack thereof. We have also developed time-lapse microscopy method for visualizing the melanin-based immune system. We applied this method to quantify the melanization kinetics during in vitro fungal infection, which improved our understanding of how fungal components, such as laccase and melanin, interact with insect melanization. We gained insight into how Candida albicans activates and evades the melanin-based immune response through morphological switching. Overall, our findings strongly suggest that melanization has direct antimicrobial activity in vivo in the insect immune system, and we subsequently explore methods to further study the melanization immune response.
Results
Galleria mellonella kill C. neoformans through melanin encapsulation in nodules
Previously, we found that C. neoformans is encapsulated inside immune system-produced melanins during infection of G. mellonella, providing evidence that the melanin-based immune response is activated against C. neoformans in G. mellonella 30 (Figure 1A). To evaluate whether insect melanin encapsulation kills C. neoformans, we assessed viability using a GFP-expressing strain of C. neoformans, which expresses GFP under an actin promotor. The GFP-expressing strain as a reporter for fungal viability C. neoformans was validated using the standard dead cell stain propidium iodide. Propidium iodide staining was nearly mutually exclusive with GFP fluorescence in untreated cells, and GFP fluorescence was extinguished when heat killed (Figure 1B, Supplementary Figure S1A and S1B). Using GFP fluorescence as a proxy for cell viability, we found fewer GFP-positive fungal cells in association with nodules located in the hemolymph compared to non-melanin encapsulated fungal cells at both room temperature and at 30°C at 24 and 72 h post-infection (Figure 1C and D). Melanin produced by C. neoformans in culture did not quench or obscure the GFP fluorescence, as determined by imaging melanized versus non-melanized C. neoformans H99-GFP (Supplementary Figure S1C-D). Hence, this result suggests that the immune melanization reaction itself is associated with fewer GFP-positive cells, consistent with death of C. neoformans in vivo during infections of G. mellonella.
Within nodule-encapsulated GFP-positive C. neoformans, we measured the degree of immune melanin intensity and GFP fluorescence intensity. We found that the yeast with the weakest GFP signal tended to be encapsulated with more melanin surrounding them, compared to the population of brightly fluorescent cells (Figure 1E). The result was an inverse correlation between melanization and fluorescence within GFP-positive cells. Most GFP-positive cells had little to no melanization, with a mean gray value around 105, which is the background gray value intensity. The occurrence of faint signal in some cells suggests that these emanate from cells in the process of dying or having been recently killed.
We attempted to use PI as an additional technique to study fungal viability within the nodules and to show that the GFP-negative cells are indeed dead. Surprisingly, PI staining did not result in the expected fluorescence in nodule-encapsulated yeast cells, but there was staining in some of the hemocytes that surrounded the yeast cells, and the external periphery of the nodules (Figure 1F). Given that PI staining was extracellular to the fungal cell, this staining could reflect released fungal or hemocyte DNA. The absence of PI staining for fungal cells in nodules suggests that PI was unable to reach the center of the nodules where the fungi are found and shows that some of the hemocytes involved in surrounding the fungus may undergo cell death in the process. (Figure 1F). The permeability or access issues that may arise when using added dyes to measure microbial viability within the nodule show the usefulness of using a live-dead indicator that is endogenously present within the fungus, such as the constitutively expressed GFP.
To confirm that insect melanin killed C. neoformans, we performed complementary experiments in vitro by assessing the ability G. mellonella melanization to inhibit the growth of C. neoformans. We incubated C. neoformans cells with extracted hemolymph from G. mellonella in a 96-well plate. Using various concentrations of a phenoloxidase-specific competitive inhibitor, phenylthiourea (PTU), we were able to generate a range of melanin inhibition conditions in the hemolymph-fungal mixture. After 24 h, we removed a small aliquot of the mixture and plated it on nutrient rich agar to allow fungal growth. The number of CFUs following a 24 h incubation with the hemocytes and PTU melanin inhibitor was directly proportional to the concentration of PTU (Figure 1G), and thus inversely proportional to the degree of melanization (Figure 1G, inset). Melanization, as measured by mean gray value, correlated with low CFUs (Figure 1H). This result strongly suggests that immune melanization, inhibits the growth of C. neoformans in vitro.
Time-lapse microscopy of hemocyte-fungal interactions and the melanization response
To record the kinetics of the hemolymph melanization response, we developed a protocol to extract hemocytes and watch their interactions with fungi (Figure 2A). Using this time-lapse microscopy, we were able to measure the rate and magnitude of the anti-cryptococcal immune melanization response following treatment with different species of fungi, mutants, or isolated virulence factors (Supplementary Video 1-2). Using particle analysis, we could quantify the area covered by melanization in minute intervals, and we can record the rate of hemolymph melanization (Figure 2B-D). By analyzing the melanization kinetics between different mutant and wildtype strains, we can determine how the mutant gene of interest affects fungal interactions with the melanization immune response. We observed variation in the extent of melanization between different experiments, likely due to variability in the fungi in the field of view, and biological variability from non-isogenic G. mellonella larvae. To overcome the risk of interpreting variability-derived artifacts as results, each experiment was performed with a corresponding control (i.e., parental strain and mutant experiments were performed at same time, using the same stock of hemolymph and the same pool of extracted hemocytes).
When we evaluated the rate and magnitude of hemolymph melanization in response to the lac1Δ mutant, we found that there was a dramatically reduced rate and magnitude of hemolymph melanization compared to the wildtype parental strain (Figure 2E). While the overall magnitude of melanization varied between replicates, the ratio of insect melanization that occurred in the lac1Δ versus H99 (WT) was statistically significantly lower than 1, and consistently around 0.35 (Figure 2F).
Many fungi, including most of those that infect people, produce melanin in their cell walls, which enables them to persist within mammalian hosts and avoid destruction from oxidative stress and antimicrobial agents 26. However, literature shows that melanized fungi are less virulent in G. mellonella compared to their non-melanized or albino mutant counterparts 27,28,31. We hypothesized that the fungal melanin could act as a pathogen or damage-associated molecular pattern (PAMP/DAMP), resulting in enhanced activation of the melanization immune reaction. We found that isolated C. neoformans melanin activated immune melanization, both with (Figure 2G) and without hemocytes present compared to heat killed C. neoformans (Supplementary Video 3-5). Since melanin ghosts contain trace amounts of fungal cell wall components that could theoretically activate the melanization immune response, we used heat-killed non-melanized cells as a control for the cell wall components that would be present. This indicates that there is a mechanism by which fungal melanin is specifically recognized and activates the phenoloxidase cascade. Further, the isolated melanin ghosts are aggregated by the hemocytes throughout the course of the time-lapse microscopy, even when immune melanization is not activated (Supplementary video 6).
We used this technique to compare the immune melanization between C. neoformans and C. albicans, the latter of which is known to trigger robust melanization of the hemolymph. In the time-lapse microscopy, we saw that C. albicans activated the melanization response faster (beginning as early as 15 minutes) and to a significantly greater extent than did C. neoformans (Figure 2H). This corresponds to the levels of melanization previously reported that occurs during G. mellonella infection with C. albicans versus C. neoformans and validates that our system, at least in part, corresponds to what occurs during actual infection.
Evaluating the melanin-based immune response of G. mellonella using tissue clearing
Tissue clearing is a technique that allows for visualization of structures deep within an organism or tissue sample, without significant disruption of the native tissue anatomy. We adapted a previously reported protocol 32 to visualize the anatomical localization of the anti-cryptococcal melanization response in G. mellonella (Figure 3A,B). We found that using this technique, we could visualize melanized nodules in situ that are formed only during infection with C. neoformans and not in uninfected controls (Figure 3C,D). These in situ melanized nodules (Figure 3D-F) appeared very similar to those that are collected from extracted hemolymph, which represent an in vivo method of visualizing the nodules (Figure 1A,3G). The visual similarities between Figure 3E,F and Figure 3G clarified that the nodules observed in extracted hemolymph are generally representative of the entirety of nodules in the organism. Both the in vivo and in situ techniques could be quantified to determine the average melanized nodule area and degree of melanization (Figure 3H). However, the cleared tissue had some opacities or normally darker tissues (i.e. digestive tract contents, legs, prolegs, spiracles, cuticle pigmentation, etc.), which can result in the detection of dark particles even in the uninfected controls, albeit at a much lower frequency (Figure 3H). Further, while there are some large C. neoformans nodules in situ that appeared aggregated together (Figure 3I, arrows), there was no clear anatomical tropism for nodule formation and the nodules are found throughout the larvae, implying that the infection was disseminated throughout the body of larvae, possibly through the insect’s open circulatory system. The large, aggregated nodules can be imaged along the Z-axis, which allowed 3D reconstruction of the nodule for a better understanding of the native nodule structure compared to the in vivo preparations compressed under a slide (Supplementary Video 7). However, compared to the in vivo experiments, the resolution of the melanin-encapsulated C. neoformans in situ is limited, and variations in opacity and tissue thickness could interfere with measurements.
Candida albicans with the Melanin-based Immune Response
C. albicans is a fungus known to elicit a strong melanization reaction in hemolymph of infected G. mellonella larvae33. We thus employed the in vitro, in vivo, and in situ techniques described above to gain insight into the host-microbe interactions of C. albicans with the G. mellonella melanin-based immune response.
Using the in vivo technique of extracting infected hemolymph to analyze melanized nodules, we observed melanin-encapsulated C. albicans cells within nodule structures (Figure 4A). These melanized nodules are like those observed during G. mellonella’s infection with C. neoformans, however, the borders of the melanin itself appeared less distinct, blurry, and smudged. An additional difference from cryptococcal infection was the presence of filamentous C. albicans structures within the nodules. These hyphae or pseudohyphae were melanin-encapsulated, but seemingly to a lesser extent than the yeast morphology.
Analysis of the tissue of larvae infected with C. albicans for 24 h using the in situ tissue clarification technique (Figure 4B) revealed groupings of the melanized nodules, often in long string-like patterns (Figure 4B), that did not appear particularly associated with any organs or tissues (Figure 4C-D, supplemental Figure 2A-F). Under higher magnification, we observed lightly pigmented hyphae and dark spherical melanized particles within the cleared larvae after 24 h of infection (Figure 4C-F). These images validate that filamentation occurred within G. mellonella, which was previously observed with histology. The melanin-encapsulated fungi form large aggregates (Figure 4C,D), which we initially thought could be indicative of a tropism for a specific tissue such as the chitinous trachea. However, upon dissection of uncleared infected larvae, there did not appear to be an association of these clusters with any specific tissues (Supplementary Figure 2 A-F). Differences in pigmentation between the two C. albicans morphologies, particularly as seen in the Z-projection in (Figure 4F), indicated that the hyphae were encapsulated with less melanin during infection compared with the yeast form of the fungus. One potential bias in interpreting this data is that since we are only looking at melanin pigmentation, we are likely missing any non-melanin encapsulated fungi which would blend in with the insect tissue.
We used the in vitro time-lapse microscopy to observe the melanization dynamics of C. albicans in hemolymph. As seen earlier (Figure 2H), C. albicans triggered a more robust melanization response than C. neoformans (Figure 5A). In time-lapse microscopy performed without the addition of insect hemocytes, we observed that the C. albicans began to grow in filamentous forms (Figure 5B2), consistent with the importance of filamentation in the pathogenesis of C. albicans within G. mellonella and the in situ data (Figure 3)23. In mammalian hosts, filamentation is triggered by serum, neutral pH, and temperature 34,35. However, in the G. mellonella system, filamentation in vitro does not occur when the C. albicans is only incubated with hemocytes without hemolymph, indicating that a component of the hemolymph is necessary for the morphological switch. Interestingly, as the time-lapse movie progressed, we observed that the hyphae did not get encapsulated by melanin in comparison to the yeast form of C. albicans (Figure 5B2). In mammalian hosts, filamentation by C. albicans is used to evade immune detection in part due to changes in cell wall structure and expression that prevent binding of C-type Lectins. After about 12 hours of filamentous growth, we observed the formation of blastoconidium (yeast) along the hyphae (Figure 5B3). The formation of these yeast cells then corresponded with a subsequent “bloom” of melanization (Figure 5A, B4, Supplementary Video 8). A similar temporal progression of C. albicans morphology and melanin-encapsulation is seen in C. albicans infected larvae dissected at various timepoints post-infection (Supplementary Figure 2G). The average time of this melanin bloom was about 840 minutes, with a 95% confidence interval between approximately 720 minutes and 960 minutes (Figure 5C).
We observed that some of the melanin-encapsulated yeast survived the immune reaction and then underwent hyphal and/or pseudohyphal growth (Figure 5D, Supplementary Video 9). This occurred in about 23% of melanin-encapsulated yeast, with 8% of single yeast and 38% of budded/pairs of yeast being able to escape (Figure 5E) The time until hyphal or pseudohyphal growth was significantly delayed in the melanin-encapsulated cells; the median time for a non-melanin encapsulated C. albicans cells to begin filamentation is 97 minutes, while the melanin-encapsulated counterparts take 230 minutes, with some taking as long as 520 minutes (Figure 5F). This delay could be reflective of physical barriers as the fungus breaks through the melanin layer and/or delays in initiating cellular growth because of cell damage caused by the immune response.
Altogether, these data demonstrate three phases of the C. albicans-immune melanization interactions: 1) yeast become encapsulated with melanin, with nearly 25% surviving and breaking through the pigment, 2) cells undergo a yeast-to-hyphal transition, with the hyphal and pseudohyphal cells evading the melanization immune response; and 3) filamentous C. albicans begins to produce more yeast cells (referred to as blastoconidia or blastospores), which then causes a second bloom of melanization to occur (Summarized in Figure 6). While host melanization and fungal filamentation have been well-reported during the course of C. albicans infection23,36–39, this is the first indication that the hyphae and pseudohyphae are melanin-evasive. Although the presence of blastoconidia has been reported in G. mellonella infected with C. albicans 36, these data indicate for the first time that lateral blastoconidia growing from hyphae induce a strong “second wave” melanization response.
Discussion
Melanin has been appreciated as a key part of the insect immune defense against microbes and parasites for the greater part of the past century 13,40. Immune melanization has been implicated as a major process in neutralizing entomopathogenic fungi upon infection41. The insect phenoloxidase and the melanization cascade produce toxic intermediates such as dihydroxyindole (DHI) and high levels of oxidative stress that can overwhelm and kill the fungus or microbe in vitro 13,17. However, there have been no in vivo studies showing that melanization directly kills fungi during these immune reactions within the insect. In this paper, we fill that gap by showing that melanization within nodules is associated with the death of C. neoformans using a GFP viability reporter assay and provide additional in vitro data for a fungicidal role in immune defense.
During cryptococcal infection of G. mellonella, melanin encapsulation of the fungus melanin encapsulation within nodules was associated with diminished or lost fluorescence signal in these GFP-expressing C. neoformans strains. Additionally, the melanin-encapsulated fungi that remained GFP positive had weaker signals and the intensity of the GFP signal was more intense for the non-melanin encapsulated fungi within the nodules. The expression of GFP in these cells is under the control of an actin promotor, and while actin is generally presumed to be constitutively expressed in cells, growth conditions have been shown to lead to some alterations in cryptococcal actin expression 42,43. If the environmental conditions within the nodule abolished actin expression in some cells without killing the fungus, we would expect that condition to equally affect the melanin and non-melanin encapsulated fungi, and as a result, see similar GFP-negative: GFP-positive ratios between the melanin-encapsulated and not melanin-encapsulated cells. The association between melanin encapsulation and disappearance in GFP fluorescence provides strong evidence for the notion that the melanization reaction kills fungal cells during infection. This is the first evidence that G. mellonella immune melanization directly and effectively neutralizes C. neoformans during infection and the first demonstration that melanin encapsulation results in fungal death within the insect. Previously, the death of microbes, specifically bacteria, was attributed to the enzymatic activity of the melanin-producing phenoloxidase (PO) in an in vitro reaction 16. In addition to our association of melanin encapsulation and fungal death in vivo, we sought to reproduce these results in vitro using extracted hemolymph in buffer. We used the PO-specific inhibitor, phenylthiourea (PTU), to inhibit PO activity and melanization and found that PO-inhibited wells of hemolymph had higher recoverable CFUs of C. neoformans compared. The inverse correlation of melanization with CFUs further supports the claim that melanin plays a role in neutralizing C. neoformans. Since we only assayed CFUs from these in vitro experiments, we cannot determine whether the melanization in the in vitro experiments directly killed the fungus or just inhibited fungal growth.
In addition to studying the extracted hemolymph, we used a developed in vitro time-lapse microscopy assay. We investigated the impact of fungal melanins on the insect immune response. We found that isolated fungal melanins, termed “melanin ghosts,” activated the melanin-based immune response whereas the heat-killed C. neoformans did not. This suggests that fungal melanins can activate the immune system which could help promote fungal clearance. This is interesting in the context of naturally-occurring fungal pathogens of insects, which tend to have a white color and/or do not naturally produce melanin pigment, such as Beauveria bassiana and Metarhizium anisophilae 24,25. This may be because of evolutionary pressure that selects for entomopathogenic fungi that produce less fungal melanin and thus are better at evading the insect’s melanin-based immune response. Since melanin is a component of the insect wound response, it is possible that these exogenous melanins are recognized by the insect as a damage associated molecular pattern (DAMP) and launches an inappropriate wound repair response.
We found that the lac1Δ mutant, which is unable to produce the enzyme laccase, causes less melanization in the hemolymph. This implies that some of the trigger for melanization comes from laccase-catalyzed initiation of melanin formation using host-derived catecholamines in the hemolymph. This is consistent with the observation that for B. bassiana infection of G. mellonella, that laccases play a role in virulence by oxidizing the hemolymph catecholamines and preventing them from producing anti-fungal melanization and reducing the oxidative burden on the fungus 24. It is also worth noting that the lac1Δ mutant is less virulent in G. mellonella infections compared to the parental 19. Together, these observations paint a nuanced picture of the role that laccase and fungal melanin play during fungal pathogenesis in G. mellonella – both fungal melanin and fungal laccases activate the melanin-based immune response, while fungal melanins are associated with decreased virulence, fungal laccases enhance virulence. We note that laccase is secreted by C. neoformans and is found in extracellular vesicles, which could transport laccase away from the fungal cell and reduce the antifungal damage from its effects on triggering insect immune melanization. We were also able to compare the amount of melanization that C. neoformans triggers with the amount triggered by other fungal species such as C. albicans. The differences in hemolymph-induced melanization during exposure to C. albicans and C. neoformans were previously described 33, and our results confirm those findings.
The second method used to evaluate the melanization response to C. neoformans was tissue clarification, which enabled us to visualize melanized nodules in situ deep within the larvae. We modified a recently developed protocol 32 to view the nodules that formed during infections, and saw the native structures of the nodules and their anatomical location in the larvae. This offers an advantage over dissection of uncleared larvae, because during the dissection process: 1) the tissue organization is disrupted, 2) some organs such as the nerve cord and cardiac system might be disrupted, and 3) the geography of infection patterns may not be apparent. Additionally, melanized nodules may not be visible within or behind opaque tissues and organs. Tissue dissection of opaque larvae was helpful when evaluating tissue tropism since tissue boundaries may not be fully visible in clarified larvae. A bias involved in studying fungal infections using both tissue clearing and dissection is that the non-melanized nodules or fungi may be missed, as unpigmented fungi will likely blend in with surrounding tissue. However, in the clearing method, we viewed the nodules throughout the entire depth of the larvae at a low to moderate (4x to 40x) magnification using light microscopy. However, the objective and microscope limitations only permitted imaging the superficial melanized nodules at 100x magnification, which provided a lower resolution of the nodules compared to the imaging of the extracted hemolymph. While in the case of C. neoformans, the nodules within the hemolymph appeared congruent to those viewed in situ, that might not always be the case. Nodules in extracted hemolymph during other fungal infection may not be entirely representative of those found throughout the entire larvae, so only viewing the hemolymph nodules may give a biased understanding of the fungal infection.
We also examined the melanization response to Candida albicans infection. C. albicans is known to trigger large scale systemic melanization in G. mellonella larvae 33,44. Similar to C. neoformans, we found melanized nodules in the hemolymph from larvae infected with C. albicans. Interestingly, the center of these nodules had melanized and smoothened areas that seemed more amorphous than those seen with C. neoformans, and additionally, we saw hyphal structures appeared less melanized than the spherical yeast-like structures. Using the tissue clarification method, we noted that the melanin-encapsulated C. albicans formed large rope-like aggregates without tissue tropism, with yeast being preferentially melanized over hyphal cells. Using in vitro time-lapse microscopy, we found that rapid melanization occurred, even in the absence of hemocytes. Additionally, after the melanization plateaus, the surviving fungus can break free from the melanin encapsulation and undergo melanin-evasive filamentation. This is followed by production of laterally-budding blastoconidium and a bloom in melanization around these newly formed yeast cells. Similar fungal morphologies and timelines were observed in dissected infected larvae, although the temporal kinetics were less resolved and identification of blastoconidium was less clear. Together, these data paint an interesting picture and allow insight into the pathogenesis of C. albicans within G. mellonella host. Hence, it appears that the melanin encapsulation can clear most of the yeast upon infection, however, cells that survive can then filament and evade subsequent melanin-mediated killing. The hyphae are known to penetrate and infect organs within the insect 23. The hyphae then produce yeast, which again triggers a burst of melanization that would likely cause damage to the surrounding tissue and eventually death of the organism.
In summary, we found evidence that the G. mellonella wax moth directly kills C. neoformans by encapsulating it with melanin in vivo using a GFP-expressing strains where fluorescence indicates viability. This association between melanin encapsulation and reduced viability provides the first direct evidence for fungal killing via melanin encapsulation in vivo. We also describe three different methodological approaches for studying the melanization response to fungi in G. mellonella and employ these techniques to study C. neoformans and C. albicans interactions with the melanin-based immune response. With C. neoformans, we show that both fungal melanins and fungal laccases can activate the insect’s melanization immune response, furthering our understanding of how these fungal components interact with insect immunity and alter the fungus’ pathogenesis. In C. albicans, we are able to observe how some melanin-encapsulated yeast are able to break through the melanization, and form melanin-evasive hyphae and pseudohyphae during infection. The direct association of insect melanization with antifungal defense further heighten concern that pesticides that inhibit the melanin reaction 30 could have untoward and unpredictable effects on insect populations.
Author contributions
Conceptualization – DFQS, QD, MK, JMH, AC; Methodology – DFQS, AC, QD, MK; Software – QD; Validation – DFQS; Formal analysis – DFQS, MK; Investigation – DFQS, MK; Resources – AC, JMH; Data curation – DFQS; Writing (Original Draft) – DFQS, AC; Writing (Review and Editing) – DFQS, QD, MK, JMH, AC; Visualization – DFQS, MK; Supervision – DFQS, JMH, AC; Project Administration – DFQS, AC; Funding Acquisition – JMH, AC
Declaration of Interests
Authors have no interests to declare
Figure Legends
Materials and Methods
Biological materials
G. mellonella last-instar larvae were obtained through Vanderhorst Wholesale, St. Marys, Ohio, USA. C. neoformans strain H99 (serotype A), C. neoformans strain H99-GFP 45, C. neoformans lac1Δ mutant, and Candida albicans strain 90028 were kept frozen in 20% glycerol stocks and sub-cultured into yeast peptone dextrose (YPD) broth for 48 h at 30°C prior to each experiment. For H99-GFP infections, frozen stock was streaked out first onto YPD agar, and green colonies were inoculated into YPD broth for 48 h at 30°C prior to each experiment. The yeast cells were washed twice with PBS, counted using a hemocytometer (Corning, New York, USA), and adjusted to 107 cells/ml for an injection inoculum of 1 × 105 cells/larva. C. albicans infections were performed at 5 × 105 cells/larva.
Extraction of hemolymph from fungal-infected Galleria mellonella larvae
Infection of G. mellonella larvae was performed as previously described 30. Briefly, washed C. neoformans or C. albicans cultures, resuspended to 107 cells/ml were injected in the right rear proleg of larvae ranging from 175 to 225 mg. Infected larvae were then incubated at 30°C. Three days following infection, larvae were removed from incubator, and hemolymph was extracted by puncturing the right rear proleg with an 18 G needle. Removed hemolymph from 3 larvae was collected directly into 1 ml anticoagulation buffer at room temperature 46. Hemolymph was centrifuged for 5 minutes at 4,000 x g and resuspended in 200 µl insect physiological saline (IPS) (150 mM sodium chloride, 5 mM potassium chloride, 7.21 mM calcium chloride, 1 mM sodium bicarbonate, pH 6.90 – adapted from 47–49). Samples were placed on slides and nodules were imaged using Olympus AX70 microscope with a 100x oil immersion objective.
C. neoformans GFP Viability Assay
H99-GFP strain was streaked from frozen stock on YPD agar and incubated at 30°C. 2 ml YPD was inoculated with H99-GFP and incubated for ∼18 h at 30°C with rotation. Culture was diluted to OD 0.5 and 100 µL was incubated at 70°C for 1 h using a thermocycler. 100 µL of untreated and heated samples were stained with 10 µg/ml propidium iodide (Invitrogen). 10 μL of stained samples were loaded onto a hemocytometer and imaged using a 10X objective and Zeiss AxioImager M2 (60x Olympus objective) equipped with a Hamamatsu Orca R2 camera and Volocity Software (Perkin Elmer). Images were analyzed using ImageJ/FIJI software. Fluorescence channel images were processed by adjusting the minimum pixel value to 10 and maximum to 90. Number of fluorescent cells for each channel were counted using Measure Particles. Number of double fluorescence positive and double fluorescence negative cells were enumerated manually.
C. neoformans GFP Fungal Survival Assay in vivo
G. mellonella larvae were infected as previously described using H99-GFP. Larvae were incubated for 3 days at 30°C, and hemolymph was extracted. Melanized nodules were visualized using an Olympus AX70 microscope with 488 excitation/520 nm emission fluorescence microscopy to visualize the GFP signal. Images were taken at 100x magnification with the same exposure, and manually marked as positive or negative for GFP fluorescence, and melanin-encapsulated or unencapsulated. For fluorescence and melanin intensity measurements, images were analyzed using the Measure tool in FIJI 50, and the 8-bit mean gray value of each cell was measured in both channels. The region selected for the melanin measurements extended the edge of the fungal capsule, and the GFP intensity measurements were from selections limited to the fluorescent cell’s body.
Phenoloxidase Inhibition and Fungal Survival Assay in vitro
Serial dilutions of phenylthiourea (PTU) were performed in 100 µl IPS buffer, to which 5 µl of 106 cells of C. neoformans was added. G. mellonella hemolymph was extracted as previously described into insect physiological saline, and 100 µl of the mixture was added to each well. The mixture was incubated at room temperature for 24 hours protected from light. Following the incubation, the contents of each well were resuspended and diluted 1:16 in PBS. From the dilution, 5 µl was spotted on a Sabouroud agar plate. The plate was incubated at 30°C for 24 h and colonies were enumerated under a dissection microscope.
Tissue Clearing of Galleria mellonella following fungal infection
G. mellonella larvae were infected with C. neoformans or C. albicans as described above. Five days following infection, groups of three larvae were removed from incubator and injected with 10 µl of 1 M ascorbic acid to inhibit new melanization and oxidation of endogenous catecholamines during the tissue clearing process. Ten minutes following the ascorbic acid injection, larvae were placed at -20°C for fifteen minutes to euthanize them, then injected with an additional 10 µl of 1 M ascorbic acid. Larvae were immediately placed in 40 mL of 4% paraformaldehyde. Larvae were fixed, permeabilized, and cleared in Benzyl Alcohol and Benzyl Benzoate (BABB) solution as previously described 32. Following 5 to 7 days of tissue clearing, larvae were removed from the BABB solution and pressed between two glass microscope slides. Once flattened, a coverslip was placed on top of the larvae and parafilmed into place. Larvae were imaged using Olympus AX70 microscope with 4x, 20x, and 100x objectives.
Imaging Galleria mellonella hemocytes in vitro
To collect and isolate hemocytes G. mellonella larvae were surface sterilized in two sequential baths of 70% ethanol, followed by 10% bleach, then dried on sterile paper towels. Five to 10 drops of hemolymph were extracted as described above into room temperature anticoagulation buffer and inverted 3 times. Hemolymph was centrifuged at 400 x g for 4 minutes, the supernatant was removed, and the hemocytes were resuspended in 1 ml anticoagulation buffer and centrifuged. The supernatant was completely removed and hemocytes were resuspended in 200 µl of insect physiological saline (IPS). The 200 µl suspension of hemocytes were added to the coverslip of a MatTek dish and allowed to settle for 10 minutes. Following the 10 minutes, the buffer and unsettled hemocytes were removed, and the coverslip was washed 4 times with 1 ml of IPS. The hemocytes are seeded into the coverslip at a cell density of 1.5 × 106 cells/ml and the resulting hemocyte density after washing is approximately 2-3 × 103 cells/mm2.
While the hemocytes were being isolated, cell-free hemolymph was being prepared. Approximately 10 drops of hemolymph were removed from G. mellonella larvae and collected directly into 1 ml IPS. To remove hemocytes, the mixture was filtered using a 0.22 µm syringe-driven PVDF filter. Cell-free hemolymph was stored up to a week at -80C. Penicillin-Streptomycin (Gibco, Thermo Fisher) antibiotic was added at 1x concentration to the cell-free hemolymph. For experiments looking at the interaction of hemocytes with fungi or a virulence factor, the cells or component are added at this stage.
Following the hemocyte washes, 1 ml of cell-free hemolymph was added to the entirety of the MatTek dish, followed by an addition 1 ml of IPS. The MatTek dish was covered and imaged using the OpenFlexure microscope and software and time-lapse microscopy was performed every minute for 16-24 hours 51. This protocol is summarized in Supplementary Figure 3.
All timelapse data was analyzing using FIJI 50 and particle measurements were made by converting the image sequence to 8-bit, setting a threshold of 0-50 gray value, and analyzing any particle over the size of 4 pixels2. Measuring the time until germ tube formation was done manually by recording the frame in which the first of the germ tube was visible.
Melanin Ghost Isolation
C. neoformans cultures were grown in minimal media with 1 mM L-DOPA for 7 days at 30°C. Cells were collected and mixed 1:1 with 12 N hydrochloric acid (HCl), for a final concentration of 6 N HCl. Cells were heated for 1 hour at 85°C under constant shaking at 350 RPM. Control cells were heat killed cells were incubated for 1 hour at 85°C in PBS. Cells were washed twice in PBS and subsequently used in the time-lapse microscopy.
Multimedia Files
Supplementary Video 1_C. neoformans timelapse
Supplementary Video 2_C. neoformans timelapse
Supplementary Video 3_Melanin Ghost vs heat killed
Supplementary Video 4 _Melanin ghost without hemocytes
Supplementary Video 5_Melanin ghost timelapse
Supplementary Video 6_Hemocyte-ghost interactions
Supplementary Video 7_In situ nodule projection
Supplementary Video 8_Melanin Bloom Candida
Supplementary Video 9_Candida albicans escape
Supplementary Video 10_C. neoformans Anticoagulation Buffer
Supplementary Video 11_No fungus timelapse
Acknowledgements
We would like to thank the entire Casadevall Lab for their contributions during lab meetings and other discussions of this project. We would like to thank Maryann Smith, Thomas Hitzelberger, and Kathy Spinnato for placing the years of weekly G. mellonella orders. Figure 6 was created using Biorender.com. D.F.Q.S., Q.D., and A.C. are funded by National Institute of Allergy and Infection Disease R01 AI052733. D.F.Q.S. is funded by National Institutes of Health 5T32GM008752-18 and 1T32AI138953-01A1. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The salaries of D.F.Q.S., Q.D., and A.C. are in part funded by the National Institute of Allergy and Infection Disease. The salaries of D.F.Q.S. and Q.D., are in part funded by the National Institutes of Health.