Introduction

Catastrophic declines of hundreds of amphibian species across six continents have been causally linked to the chytrid fungi, Batrachochytrium dendrobatidis (Bd) and Batrachochytrium salamandrivorans (Bsal)^{1, 2}. Motile Bd zoospores readily colonize keratinized skin of adult amphibians and keratinized mouthparts of tadpoles³. Bd skin infections culminate in chytridiomycosis, ultimately disrupting the function of this respiratory and barrier tissue^{3, 4, 5}. Effectively combatting chytrid infections requires a holistic understanding of amphibian cutaneous immune defenses. Research efforts up to this point have focused on antifungal capacities of amphibian skin-produced antimicrobial peptides^{6, 7}; commensal antifungal products^{7, 8, 9}; antifungal properties of mucus⁹, alkaloids¹⁰, and lysozymes^{11, 12} as well as the roles of antibodies^4,13. Many of these studies suggest that Bd exposure can elicit some immune protection¹⁴. However, the contribution of skin-resident immune cells to amphibian anti-Bd defenses remains almost entirely unexplored.

Mammalian mast cells serve as sentinels of mucosal and connective tissues, concentrated in barrier tissues like skin, where they maintain homeostasis and regulate immune responses¹⁵. Other granulocyte-lineage cells such as neutrophils are generally not found in healthy tissues and only extravasate into sites of inflammation¹⁶. Consequently, mast cells are among the first immune cells to recognize and respond to skin-infiltrating pathogens. When activated, mast cells release pre-formed and de novo-generated immunomodulatory compounds that may serve to elicit, exacerbate, or ameliorate inflammatory responses¹⁵. Cells bearing hallmark mast cell cytology have been reported across a range of non-mammalian species^{17, 18}, including amphibians¹⁹. Notably, the principal mast cell growth factor, stem cell factor (SCF, KIT ligand) required for mast cell differentiation and survival²⁰ is expressed by all vertebrates examined to-date.

Here, we combine comprehensive in vitro and in vivo approaches to define the roles of amphibian (Xenopus laevis) mast cells during Bd infections. Our results provide compelling evidence that skin-resident immune cells contribute to anti-Bd defenses.

Results

Frog mast cells possess archetypal mast cell cytology and transcriptional profiles

We produced X. laevis recombinant (r)SCF, and used this reagent to generate mast cell cultures from bone marrow-derived myeloid precursors²¹. Mast cells were compared to bone marrow-derived neutrophilic granulocytes (hereafter referred to as ‘neutrophils’), differentiated using a recombinant X. laevis colony-stimulating factor-3²¹ (rCSF3, i.e., granulocyte colony-stimulating factor; GCSF). While the neutrophil cultures were comprised of cells with hyper-segmented nuclei and neutral-staining cytoplasms (Fig. 1A), the mast cell cultures consisted predominantly of mononuclear cells with basophilic cytoplasm (Fig. 1B). We confirmed the granulocyte-lineage of X. laevis mast cells using specific esterase (SE) staining (Fig. 1D). As expected, X. laevis neutrophils were also SE-positive (Fig. 1C). Mast cell and neutrophil morphology was further explored with electron microscopy (Fig. 1E-H). SEM imaging demonstrated that X. laevis mast cells possess extensive folding of their plasma membranes (Fig. 1F). This mast cell-characteristic membrane ruffling appeared as projections resembling pseudopods via TEM, which further revealed electron-dense heterogenous granules, few mitochondria, and round to elongated nuclei (Fig. 1H) typical of mammalian mast cells²². X. laevis neutrophils also exhibited pronounced membrane ruffling (Fig. 1E) but strikingly distinct intracellular appearance including multilobed nuclei (Fig. 1G).

Figure 1.

Frog mast cells and neutrophils also exhibited distinct transcriptional profiles of immune- related genes including those encoding lineage-specific transcription factors, immune receptors, downstream signaling components and adhesion molecules, as well as non-immune genes (Fig. 1I). Frog mast cells and neutrophils each expressed greater levels of lineage-specific transcription factors associated with mammalian mast cell (gata1, gata2, and mitf)²³ and neutrophil (cebp family members)²⁴ counterparts, respectively (Fig. 2A). Notably, mast cells expressed greater levels of enzyme and cytokine genes associated with tissue remodeling (carboxypeptidase; cpa²⁵), immune suppression (indoleamine 2,3 dioxygenase-1; ido1²⁶) and amelioration of cutaneous inflammation (leukemia inhibitory factor, lif²⁷; Fig. 2B). Conversely, neutrophils expressed predominantly proinflammatory enzymes and cytokine genes such as leukotriene 4 hydrolase (lta4h; Fig. 2B) and tumor necrosis factor alpha (tnfα, Fig. 2B). In addition, mast cells and neutrophils each had greater expression of genes encoding their respective growth factor receptors, kit and csf3r (Fig. 2B).

Figure 2.

Enriching mast cells in frog skin offers protection against Bd

Although all granulocyte-lineage cells possess SE activity, mast cells are the predominant mononuclear granulocytes to reside in vertebrate tissues²⁸. Therefore, we selectively enriched mast cells in X. laevis skin via subcutaneous rSCF administration (note SE-stained cells indicated by arrows in r-ctrl-injected skins, Fig. 3A, versus r-SCF-injected skins, Fig 3B). We confirmed SE- positive cells in rSCF-treated skins also possessed round-oval nuclei (Fig. 3C). Maximum mast cell enrichment was observed 12 hours post injection (hpi) of rSCF (Fig. 3D). When we challenged X. laevis with Bd, we did not detect differences in skin fungal loads between control (r-ctrl- injected) and mast cell-enriched groups at 7 days post infection (dpi; Fig. 3E). However, fungal loads were almost two-fold lower in mast cell- compared to control-enriched animals at 21 dpi, allowing for approximately four cycles of zoospore to mature zoosporangia development (Fig. 3E).

Figure. 3.

Mast cells protect frogs from Bd-elicited inflammation-associated pathology

To explore potential mechanisms of mast cell-mediated protection against Bd, we compared the gene expression profiles of r-ctrl- and rSCF-administered Bd-infected frog skins at 21 dpi. Among the top differentially expressed genes, we noted mast cell-enriched, Bd-infected skins possessed greater transcripts for genes associated with cutaneous strength and integrity (lamc2), epidermal cell maturation, lymphocyte recruitment (b3gnt3.1)^{29, 30, 31}, as well as ion and nutrient flow (gjb3l). (Fig. 4A). Moreover, mast cell-enriched Bd-challenged skin exhibited greater expression of genes associated with protection of the mucosa and epithelial healing (ttf3.6s)³², pathogen nutrient deprivation in mucus (slc1a5s)³³, and mucus production (duoxa1.s³⁴; Fig. 4A). In striking contrast, skins from control Bd-infected frogs revealed greater expression of genes associated with leukocyte infiltration and inflammation (e.g., ccl19, cxcl16, adamt13, csf3r; Fig. 4A). These transcriptional profiles were supported by our histological observations wherein control Bd- infected skins exhibited hyperkeratosis, epidermal hyperplasia, jagged stratum corneum, and extensive leukocyte infiltration (Fig. 4B), while mast cell-enriched Bd-infected tissues appeared largely devoid of these pathologies (Fig. 4C).

Figure. 4.

Cutaneous neutrophil enrichment results in increased Bd fungal loads

Neutrophils are one of the first leukocytes to infiltrate infected tissues, typically amplifying inflammation³⁵. All vertebrate neutrophils depend on CSF3 for their differentiation and function³⁶, and our previous work has demonstrated the gene encoding the CSF3 receptor (csf3r) is a marker of X. laevis neutrophils^{37, 38}. Because csf3r expression was markedly elevated in control over mast cell-enriched skins of infected frogs (Fig. 4A), we next assessed the outcomes of enriching skin for neutrophils via subcutaneous rCSF3 administration. We confirmed neutrophil enrichment peaked 12 hp rCSF3 injection (Fig. 4F) and resulted in a thickened epidermis in comparison with r-ctrl injected skins of otherwise healthy animals (i.e., no Bd; Fig. 4D, E). When challenged with Bd, frogs with neutrophil-enriched skin possessed significantly greater Bd loads than control frogs (Fig. 4G). This suggests inflammation may be exacerbating Bd infections.

In vitro analyses of frog mast cells responses to Bd

Amphibians rely heavily on skin-produced antimicrobial peptides (AMPs) for antifungal protection³⁹, and mast cells produce antimicrobial AMPs⁴⁰. Thus, we examined whether mast cells could be sources of such AMPs during Bd infections. As anticipated, mast cells, but not neutrophils challenged with Bd, upregulated their expression of the AMP-encoding genes, pgla and magainin (Fig. S1A).

Mammalian mast cells are recognized as potent producers of the pleotropic anti- inflammatory cytokine, interleukin-4 (IL4)⁴¹. Notably, X. laevis mast cells challenged in vitro with Bd significantly upregulated their il4 gene expression (Fig. S1A), whereas almost no il4 expression was detected from either unstimulated or Bd-challenged neutrophils. To examine whether the frog IL4 might offer anti-Bd protection, we generated the X. laevis IL4 in recombinant form (rIL4) and confirmed subcutaneous injection of this moiety augmented expression of genes typically activated by the mammalian IL4⁴² (cd36, metalloproteinase inhibitor 3-timp3, and monoamine oxidase A- maoa; Fig. S1B). However, subcutaneous administration of rIL4 in Bd-infected frogs did not alter their fungal loads (Fig. S1C), suggesting IL4 is probably not a major mechanism responsible for the observed mast cell-mediated anti-Bd protection.

Enrichment of frog skin mast cells alters mucus gland composition

Because mast cell-enriched frog skin had greater expression of genes associated with mucosal tissue integrity and mucus production (Fig. 4A), we next investigated whether the anti-Bd protection identified in mast cell-enriched skins could be due, at least in part, to differences in mucus production. Interestingly, while mucus glands in mock- and Bd-infected control skin sections contained normal levels of mucus, those from both mock- and Bd-infected frogs that had been enriched for cutaneous mast cells were significantly more filled with higher levels of mucin (Fig. 5A-C).

Figure. 5.

Enrichment of frog skin mast cells alters skin microbial composition

Amphibian skin-produced mucus may offer antimicrobial protection and serve as a selective substratum for commensal microbes, many of which are antifungal⁹. We found no significant differences in direct Bd-killing capacities of mucus isolated from mock- or Bd-challenged control or mast cell-enriched frogs (Fig. S2). By contrast, we observed substantial differences in skin microbiomes, including changes in bacterial composition and richness as well as relative abundances of Bd-inhibitory bacteria (Fig. 5D-G, Fig. S3). A total of 1645 bacterial amplicon sequence variants (ASVs) were identified from 20 bacterial phyla, seven of which were predominant (Fig. 5D). Of these, Verrucomicrobiota were only present on uninfected animals, whereas Acidobacteriota was only seen after 21 dpi on both control and mast cell-enriched, infected animals (Fig. 5D).

At 10 dpi, mast cell-enrichment resulted in a nominal shift in community composition compared to control frogs (Fig. 5E). Notably, while control, Bd-infected animals exhibited a drastic shift in community composition, mast cell-enriched animals possessed substantially less deviated community composition (Fig. 5E), suggesting that these cells are somehow counteracting the adverse effects of Bd on microbiome structure. These mast cell-mediated effects persisted to 21 dpi (Fig. S3A).

At 10 dpi, mast cell-enriched and mock-infected frogs possessed significantly greater abundance of Bd-inhibitory bacteria such as Chyrseobacterium sp., compared to control, mock infected animals (Fig. 5F)⁴³. This suggests that mast cells may promote skin flora composition that is more antifungal. Control (non-enriched) Bd-infected frogs possessed significantly greater abundance of Bd-inhibitory bacteria than all other treatment groups (Fig. 5F). While mast cell- enriched, Bd-infected frogs had lower abundance of Bd-inhibitory bacteria than control infected frogs, they possessed higher abundance of inhibitory taxa than uninfected control animals (Fig. 5F). The Bd-inhibitory bacteria seen in greater abundance on mast cell-enriched, Bd-infected animals, included Roseateles sp., Flavobacterium sp., and Kaistia sp. We did not see significant differences in Bd-inhibitory bacteria across the treatment groups at 21 dpi (Fig. S3B).

Mast cell-enriched uninfected frogs exhibited increased frog skin bacterial richness at 10 dpi (Fig. 5G). While control Bd-infected animals exhibited significantly reduced skin microbial richness, mast cell-enriched Bd-infected frogs did not exhibit such a reduction in bacterial richness (Fig. 5G), supporting the idea that mast cells may be counteracting the adverse effects of Bd on skin microbiome composition.

Discussion

Amphibian extinction rates far outpace those of any other vertebrate class⁴⁴. It is now well- established that chytrid fungi are major contributors to these declines, and strikingly, are considered the greatest infectious disease threat to biodivsersity⁴⁵. The development of effective mitigation strategies though, is hindered by incomplete understanding of amphibian immune defenses and skin integrity. In this respect, while mast cells are recognized as key immune sentinels of tissues such as skin⁴⁶, relatively little is known about this cell lineage outside of mammals. Our findings provide the first in-depth functional analyses of these cells in amphibians and explore their protective mechanisms during chytrid infections. This work presents a unique perspective on the evolution of mast cell functionality and will serve as a new avenue to explore ways to counteract the amphibian declines.

Mammalian mast cells directly interact with neurons innervating the skin⁴⁷, and mucus production in mammals is tightly controlled by neuronal signaling⁴⁸. In turn, mucus serves as a critical substratum for microbiota at mucosal barriers⁴⁹. Here, we demonstrate that frog skin mast cells appear to confer significant protection against Bd by counteracting Bd-elicited inflammation, maintaining skin integrity, promoting mucus production, and preventing the disassembly of established skin microbiomes. Amphibian skin is much thinner and more permeable than that of mammals⁷, and as such, arguably represents a more penetrable barrier to pathogens. Because mammalian skin is relatively impermeable, mast cells are absent from mammalian epidermis and are instead found exclusively in their dermal layers⁵⁰. In contrast, we have shown here and have observed across several classes of amphibians¹⁶ that mast cells are found in both epidermal and dermal layers. This localization has presumably evolved to support the more intimate contact between amphibian skin and their environments. Considering the importance of amphibian cutaneous integrity to their physiology⁵¹, it is likely that skin-resident mast cells co-evolved to support skin immunity, physiology, and symbiotic microbiota. Ongoing research continues to reveal functional differences between mammalian connective tissue and mucosa-resident mast cells⁵², and we suspect that there is similarly much to learn about the distinct physiological and immune roles of amphibian epidermal and dermal mast cells.

Bd infections caused major reductions in bacterial taxa richness, changes in composition and substantial increases in the relative abundance of Bd-inhibitory bacteria early in the infection. Similar changes to microbiome structure occur during experimental Bd infections of red-backed salamanders and mountain yellow-legged frogs^{8, 53}. In turn, progressing Bd infections corresponded with a return to baseline levels of Bd-inhibitory bacteria abundance and rebounding microbial richness, albeit with dissimilar communities to those seen in control animals. These temporal changes indicate that amphibian microbiomes are dynamic, as are the effects of Bd infections on them. Indeed, Bd infections may have long-lasting impacts on amphibian microbiomes⁸. While Bd infections manifested in these considerable changes to frog skin microbiome structure, mast cell enrichment appeared to counteract these deleterious effects to their microbial composition. Presumably, the greater skin mucosal integrity and mucus production observed after mast cell enrichment served to stabilize the cutaneous environment during Bd infections, thereby ameliorating the Bd-mediated microbiome changes. While this work explored the changes in established antifungal flora, we anticipate the mast cell-mediated inhibition of Bd may be due to additional, yet unidentified bacterial or fungal taxa. Intriguingly, while mammalian skin mast cell functionality depends on microbiome elicited SCF production by keratinocytes⁵⁴, our results indicate that frog skin mast cells in turn impact skin microbiome structure and likely their function. It will be interesting to further explore the interdependent nature of amphibian skin microbiomes and resident mast cells.

In contract to the mast cell effects, enrichment of neutrophils seemed to result in greater Bd skin burdens. This suggests that neutrophils and the cytokines they release may result in a greater level of inflammation, which is not protective. This is consistent with other studies that suggest that a strong immune response in the skin compartment of some species may be detrimental^{55, 56}

It has become apparent that amphibian host-Bd interactions are highly complex and multifaceted while different amphibian species exhibit marked differences in their susceptibilities to this devastating pathogen^{2, 4}. The findings described here emphasize the importance of skin- resident mast cells for successful anti-Bd defenses and demonstrate that these immune sentinels are intimately linked to many aspects of frog skin physiology. Our results indicate that when mast cells are enriched, the ensuing changes in the skin allow for greater resistance to infection by developing zoospores. Presumably, distinct amphibian species have evolved disparate interconnections between their skin mast cells and their cutaneous defenses, as dictated by their respective physiological and environmental pressures. In turn, these species-specific differences likely dictate whether and to what extent the skin-resident mast cells of a given amphibian species recognize and appropriately respond to Bd infections. We postulate that such differences may contribute to the differences in susceptibilities of amphibian species to pathogens like chytrid fungi.

Methods

Animals

Outbred 1 year-old (1.5-2”), mixed sex X. laevis were purchased from Xenopus 1 (Dexter, MI). All animals were housed and handled under strict laboratory regulations as per GWU IACUC (Approval number 15-024).

Recombinant cytokines and bone marrow granulocyte cultures

The X. laevis rSCF, rCSF3 and rIL4 were generated as previously described for rCSF3⁵⁷ and as detailed in the supplemental materials.

Bone marrow isolation, culture conditions, and establishment of neutrophil cultures have been previously described⁵⁸. Mast cell cultures were generated according to protocols adapted from Koubourli et al. (2018) and Meurer et al. (2016) ^{59, 60}. Isolated bone marrow cells were treated with 250 ng/µl of rSCF on Day 0, Day 4, Day 7, and collected for further analysis on Day 9. Cell cultures were maintained at 27°C with 5% CO2 in amphibian medium supplemented with 10% fetal bovine serum and 0.25% X. laevis serum. Neutrophil-like granulocytes were generated as above but with 250 ng/µl of rCSF3 on Day 0, Day 3, and collected for further analysis on Day 5. Cell cultures were maintained at 27°C with 5% CO2 in amphibian serum-free medium supplemented with 10% fetal bovine serum, 0.25% X. laevis serum, 10 μg/mL gentamicin (Thermo Fisher Scientific, Waltham, Massachusetts, USA), 100 U/mL penicillin, and 100 μg/mL streptomycin (Gibco, Thermo Fisher Scientific).

Enrichment of skin granulocyte subsets

Animals were subcutaneously injected between the skin and muscle layers with 5 µg/animal of rSCF, rCSF3, or r-ctrl in 5 µL of saline using finely pulled glass needles. Optimal time course and dose for in vivo mast cell and neutrophil enrichment were determined during preliminary experiments.

Recombinant interleukin-4 treatment

The capacity of the recombinant interleukin-4 (rIL4) to induce expression of genes associated with mammalian IL4 responses were assessed by injecting frogs (N=6 per treatment group subcutaneously with rIL4 (5 µg/animal) or r-ctrl in 5 µl of saline. After 6hrs, animals were sacrificed, and skins were isolated for gene expression analyses.

To examine the effect of rIL4 on Bd loads, frogs were infected with Bd by water bath (10⁵ zoospores/mL) as described below and 1 day later injected subcutaneously, dorsally with rIL4 (5 µg/animal) or r-ctrl in 5 µl of saline. After an additional 9 days of infection, animals were sacrificed and their dorsal skin Bd loads examined.

Bd stocks and fungal challenge

Bd isolate JEL 197 was grown in 1% tryptone broth or on 1% tryptone agar plates (Difco Laboratories, Detroit, MI) supplemented with 100 U/mL penicillin and 100 μg/mL streptomycin (Gibco) at 19°C.

In vitro Bd killing was evaluated by incubating live Bd (maturing zoosporangia) with mast cells or neutrophils at ratios of 5:1 or 1:1 Bd cells per granulocyte. Cells were incubated at 27°C for three days before fungal loads were analyzed by absolute qPCR. Experimental groups were compared to pathogen DNA amounts derived from equal quantities of live Bd plated alone in otherwise identical conditions.

For in vivo infection studies, zoospores were harvested by flooding confluent tryptone agar plates with 2 mL sterile A-PBS for 10 minutes. Twelve hours post rSCF, rCSF3, rIL4, or r-ctrl injection, animals were infected with 10⁷ zoospores or mock-infected in 100 mL of water (10⁵ zoospores/mL). After 3 hrs, 400 mL of water was added to each tank. Skins were collected for histology and gene expression analysis on 1, 7, and 21 dpi.

Histology

X. laevis leukocyte cytology and cutaneous SE staining has been described⁵⁷ and is detailed in the supplemental materials. An Alcian Blue/PAS staining kit (Newcomer Supply, Middleton, WI) was used to quantify mucin content from in vivo experiments. Histological analyses are detailed in the supplemental materials.

Electron Microscopy

Processing and imaging of cells for transmission and scanning electron microscopy (TEM and SEM) was conducted at the GWU Nanofabrication and Imaging Center (GWNIC). For transmission electron microscopy, cells were fixed as monolayers on six-well plates with 2.5% glutaraldehyde and 1% paraformaldehyde in 0.1 M sodium cacodylate buffer for one hour. Cells were treated with 1% osmium tetroxide in 0.1 M sodium cacodylate buffer for 1 hr. Following washes, cells were en bloc stained with 1% uranyl acetate in water overnight at 4°C. Samples were dehydrated through an ethanol series and embedded in epoxy resin using LX112. Inverted Beem capsules were placed into each tissue culture well to create on face blockfaces for sectioning. Resin was cured for 48 hrs at 60°C. The 95 nm sections were post-stained with 1% aqueous uranyl acetate and Reynold’s lead citrate. All imaging was performed at 80 KV in a Talos 200X transmission electron microscope (Thermo Fisher Scientific, Hillsboro, OR).

For SEM, cells were fixed with 2.5% glutaraldehyde / 1% paraformaldehyde in Sodium cacodylate buffer, followed by 1% OsO4, then dehydrated through an ethyl alcohol series. Coverslips were critical point dried and coated with 2 nm iridium. Cells were imaged using a Teneo Scanning Electron Microscope (Thermo Fisher Scientific).

Analyses of immune gene expression and Bd skin loads

These analyses have been described⁵⁷ and are detailed in supplemental materials.

RNA sequencing

For transcriptomic profiling, bone marrow-derived neutrophil and mast cell cultures were generated as described above and FACS-sorted according to preestablished size and internal complexity parameters to isolate the respective subsets for further analyses. Sorted cells were immediately processed to extract and purify RNA. Flash frozen samples were sent to Azenta Life Sciences for all library preparation, RNA sequencing, and analyses. In short, polyadenylated RNA was isolated using Oligo dT beads. Enriched mRNAs were then fragmented for first and second strand cDNA synthesis. cDNA fragments were end repaired, 5’ phosphorylated, and dA-tailed.

Fragments were then ligated to universal adaptors and PCR-amplified. 150-bp paired-end sequencing was performed on an Illumina HiSeq platform.

FastQC was used to evaluate raw data quality. Adaptors sequences and poor-quality nucleotides were removed from reads using Trimmomatic v.0.36. The STAR aligner v.2.55.2b was used to map these reads to the Xenopus_laevis_9_2 reference genome from ENSEMBL. To determine differential gene expression, featureCount (Subread package v.1.5.2) was first used to count unique gene hits, which were then used with DESeq2 to calculate absolute log2-fold change.

Skin microbiome analyses

Towards microbiome studies, frogs were housed individually (N=6/treatment group). At indicated times, frogs were gently rinsed with sterile deionized water to remove transient microbes and gently swabbed 20 times, dorsally. Genomic DNA was extracted from swabs using a PowerSoil Pro kit on a Qiacube HT (QIAGEN, MD). One-step PCR library prep and dual-index paired-end Illumina sequencing was used to sequence the skin microbiome of individual frogs. A ∼380 base pair region in the V3-V5 region of the 16S rRNA gene using the universal primers 515F-Y (GTGYCAGCMGCCGCGGTAA) and 939R (CTTGTGCGGGCCCCCGTCAATTC) was used for amplification. Negative and positive controls (ZymoBIOMICS D6300 & D6305, Zymo, CA) were included in each round of extraction and PCR. Reactions were done in duplicate for each sample, pooled, cleaned with in-house Speed-beads (in a PEG/NaCl buffer), quantified with a Qubit4 (Invitrogen, MA) and pooled into a final library in equimolar proportion. The pooled library was sequenced on two Illumina MiSeq runs (v3 chemistry: 2x300 bp kit) at the Center for Conservation Genomics, Smithsonian National Zoo & Conservation Biology Institute.

Analyses were performed in the R environment version 4.0.3 (R Core Team, 2020) using methods detailed in the supplemental materials.

Statistical analyses

Differences in transcript expression were calculated with one-way or multi-way ANOVAs followed by Tukey post-hoc tests. Student’s t-tests were used to determine differences in Bd loads between treatments only. Statistical differences in mucin content and mucosome Bd-killing were assessed with the two-way ANOVA calculator available online through Statistics Kingdom. For RNA sequencing, p-values were calculated with the Wald test and were adjusted using the Benjamini-Hochberge procedure.

Acknowledgements

We thank Dr. Cynthia S. Dowd and Ben England (Chemistry Department, GWU) for their help with X. laevis mucus lyophilization. We thank Dr. Christine Brantner (GWU Nanofabrication and imaging center) for assistance with TEM and SEM imaging. KAH, MRHH, and AY were supported by Wilbur V. Harlan summer graduate research fellowships. KAH, MRHH, NR, NK, AY, MZ, and LG thank the GWU Biology Department.

Funding

National Science Foundation grant 2131061 (LG, CRMW) National Science Foundation grant 2147466 (LG, LAR-S) National Science Foundation grant 2147467 (LAR-S)

Author contributions

Conceptualization: KAH, LG

Methodology: KAH, MRHH, LKG, CNG, NR, NK, AY, MZ, EJJ, AND, LAR-S, CRMW, LG

Investigation: KAH, MRHH, LKG, CNG, NR, NK, AY, MZ, EJJ, AND, LAR-S, CRMW, LG

Visualization: KAH, MRHH, EJJ, AND, LG

Funding acquisition: LG, CRMW, LAR-S

Project administration: LG

Supervision: LG, CRMW, LAR-S

Writing – original draft: KAH, LAR-S, CRMW, LG

Writing – review & editing: KAH, LAR-S, CRMW, LG

Competing interests

The authors declare no competing interests.

Data and materials availability

All data are available in the main text or the supplementary materials.

Supplementary materials

Production of recombinant cytokines

The X. laevis SCF and CSF3 sequences representing the signal peptide-cleaved transcripts were ligated into the pMIB/V5 His A insect expression vectors (Invitrogen). SCF-ligated, CSF3-ligated, or empty vectors were transfected into Sf9 insect cells (cellfectin II, Invitrogen). Recombinant proteins contain a V5 epitope, and western blot with an anti-V5-HRP antibody (Sigma) confirmed their presence. Positive transfectants were selected using 10 μg/mL blasticidin (Gibco). Expression cultures were scaled up to 500 mL liquid cultures, grown for 5 days, pelleted by centrifugation, and the supernatants collected. Supernatants were dialyzed overnight at 4°C against 150 mM sodium phosphate, concentrated against polyethylene glycol flakes (8 kDa) at 4°C, dialyzed overnight at 4°C against 150 mM sodium phosphate, and passed through Ni-NTA agarose columns (Qiagen). Columns were washed with 2 × 10 volumes of high stringency wash buffer (0.5% Tween 20, 50 mM Sodium Phosphate, 500 mM Sodium Chloride, 100 mM Imidazole) and 5 x 10 volumes of low stringency wash buffer (as above but with 40 mM Imidazole). Recombinant proteins were eluted with 250 mM imidazole. After recombinant protein purification, a halt protease inhibitor cocktail (containing AEBSF, aprotinin, bestatin, E-64, leupeptin and pepstatin A; Thermo Scientific) was added. Intact recombinant protein presence was confirmed again by western blot and the protein concentrations quantified by Bradford protein assays (BioRad). Protein aliquots were stored at -20°C until use.

Bone marrow granulocyte cultures

Bone marrow cells were isolated as previously described ⁵⁸. Briefly, adult X. laevis (approximately one year old) were euthanized in 5% tricaine mesylate followed by cervical dislocation. Femurs were removed and washed in ice-cold Amphibian-PBS (A-PBS) in sterile conditions. Each femur was flushed with 5 mL of A-PBS. Red blood cells were removed from culture via a differential gradient generated with 51% Percoll. Bone marrow cell counts were generated using trypan blue exclusion and cells were seeded at a density of 10⁴ cells/well for gene expression experiments, 5x10⁴ cells/well for histology analyses, and 10⁵ cells/well for electron microscopy analyses.

Histology

Paraffin-embedded tissue sections (5 μm) were deparaffinized, rehydrated through A- PBS, and stained with Naphthol AS-D Chloroacetate (specific esterase; Sigma) or Alcian Blue/PAS (Newcomer, Middleton, WI) according to the manufacturers’ instructions and optimized for Xenopus skin tissues. Cells collected from in vitro cultures were cytocentrifuged onto glass microscope slides (VWR). Cells were stained immediately with Giemsa (Sigma) for 7 minutes or fixed with 10% neutral buffered formalin for 30 minutes and stained with specific esterase according to the manufacturers’ instructions. Slides stained with Alcian Blue/PAS (Newcomer) were used to quantify mucin content from in vivo experiments. Images were taken using identical microscope settings under 20x magnification. Images were converted to 8-bit in Fiji by ImageJ and threshold adjusted such that positive staining for mucus was captured within the mucus glands (threshold held constant across images). The percentage of each mucus gland positively stained and the average percent-positive per field of view were subsequently calculated. Positive staining of both acidic and neutral mucins was included in analyses. All slides were imaged with a Leica DMi8 Inverted Fluorescent Microscope with all mucus glands assessed for each respective frog skin section (Leica Microsystems, Davie, FL).

Analyses of immune gene expression and Bd skin loads

Cells and tissues were homogenized in Trizol reagent, flash frozen on dry ice, and stored at -80°C until RNA and DNA isolation. RNA isolation was performed using Trizol according to the manufacturer’s directions. RNA-Seq is described in detail below. For qRT-PCR gene expression analysis, RNA (500 ng/sample) was reverse transcribed into cDNA using cDNA qscript supermix (Quantabio, Beverly, MA). Following RNA extraction, back extraction buffer (4 M guanidinethiocyanate, 50 mM sodium citrate, 1 M Tris pH 8.0) was mixed with the remaining Trizol layer and centrifuged to isolate the DNA-containing aqueous phase. DNA was precipitated overnight with isopropanol, pelleted by centrifugation, washed with 80% ethanol, and resuspended in TE buffer (10 mM Tris pH 8.0, 1 mM EDTA). DNA was purified by phenol:chloroform extraction and resuspended in molecular grade water (VWR).

Quantitative gene expression analyses for both Bd and X. laevis cells and tissues were performed using the CFX96 Real-Time System (Bio-Rad Laboratories, Hercules, CA) and iTaq Universal SYBR Green Supermix (Bio-Rad Laboratories). The Bio-Rad CFX Manager software (SDS) was employed for all expression analysis. All expression analyses were conducted using the ΔΔCt method relative to the gapdh endogenous control gene for X. laevis. Fungal load quantification was assessed by absolute qPCR. Isolated Bd DNA (JEL 197 isolate) was serially diluted and used as the standard curve. Primers were designed and validated against the Bd ribosomal RNA internal transcribed spacer 1 (ITS1). The primers used are listed in Table S1.

Analyses of mucus Bd-killing capacities

Mucosomes were collected from mast cell- or vector- enriched X. laevis that were either mock- or Bd-infected for 10 or 21 days. To this end, individual X. laevis were soaked in a 5 mL water bath for 1 hour. Each water sample was then lyophilized, reconstituted with 500 µl of molecular grade water, and passed through a sterile cell strainer to remove large debris.

Bd was seeded in opaque white 96-well plates (20,000 zoospores in 50 µl of tryptone broth / well). Next, 50 µl of mucosome solution was added to each well (100 µl total well volume) in three replicate wells per individual X. laevis mucosome. Mucosomes, tryptone broth, and water were each plated alone as controls. Plates were sealed with parafilm and incubated at 19°C for 16 hrs with gentle mixing (20 rpm).

Zoospore viability was determined with the CellTiter-Glo 2.0 Cell Viability assay kit (Promega) according to the manufacturer’s instructions and using a SpectraMax plate reader (Molecular Devices, San Jose, CA). Luminescence readings were fitted to a standard curve (descending proportions of heat-killed zoospores to viable zoospores) to calculate the number of viable zoospores in each well. Zoospores were heated-killed at 65°C for 15 mins.

Skin microbiome analyses

All analyses were performed in the R environment version 4.0.3 (R Core Team, 2020). Demultiplexed reads were imported from Basespace into R environment for sequence processing. Package “dada2”⁶¹ was used to perform quality filtering using their standard filtering parameters (i.e., maxEE = 2), which collapsed high quality reads into amplicon sequence variant (ASV) and removed chimeras. Bacterial taxonomy was assigned using Silva version 138.1. The R package “phyloseq”⁶² was used to import and merge the final ASV table, taxonomy table, and metadata to create a phyloseq object to perform further analyses. Sequences classified as cyanobacteria/chloroplast and those unclassified at kingdom were removed. Singletons were filtered out (i.e., ASVs with only one sequence read in one individual). The R package “decontam”⁶³ was used to remove potential contaminants using the method “combined.” The ZymoBIOMICS microbial community standards (positive controls) were analyzed, and we found genera in similar relative abundances as described by Zymo.

To determine how Bd and mast cell treatments impacted skin microbiomes, the microbiome structure was examined. The components of microbiome structure were: ASV richness (measured as Bd-inhibitory ASV richness and total ASV richness), microbial composition (measured by Jaccard and Bray-Curtis distances), and sequence abundance of Bd-inhibitory ASVs (measured as individual Bd-inhibitory ASV sequence counts and total relative abundance of Bd-inhibitory ASVs). To characterize variation in microbiome structure, Mast Cell (mast cell normal and mast cell+), Bd (Bd- and Bd+) and their interaction as explanatory variables at two-time points Day 10 and Day 21 post Bd infection were included. For this characterization, log-transformed ASV richness in ANOVAs, microbial composition measures in PERMANOVAs and log-transformed raw sequence counts in ANOVAs (with post-hoc corrections for multiple comparisons) were used. For identification of Bd-inhibitory ASVs, methods as described in by Jimenez et al., 2022⁶⁴ were followed.