Extracellular ATP released from Candida albicans activates non-peptidergic neurons to augment host defense

Intestinal microbes release ATP to modulate local immune responses. Herein we demonstrates that Candida albicans, an opportunistic commensal fungus, also modulates immune responses via secretion of ATP. We found that ATP secretion from C. albicans varied between standard laboratory strains. A survey of eighty-nine clinical isolates revealed heterogeneity in ATP secretion, independent of growth kinetics and intracellular ATP levels. Isolates from blood released less ATP than commensals, suggesting that ATP secretion assists with commensalism. To confirm this, cohorts of mice were infected with strains matched for origin, and intracellular ATP concentration, but high or low extracellular ATP. In all cases fungal burden was inversely correlated with ATP secretion. Mice lacking P2RX7, the key ATP receptor expressed by immune cells in the skin, showed no alteration in fungal burden. Rather, treatments with a P2RX2/3 antagonist result in increased fungal burden. P2RX2/3 is expressed by non-peptidergic neurons that terminate in the epidermis. Cultured sensory neurons flux Ca2+ when exposed to supernatant from heat-killed C. albicans (HKCA), and these non-peptidergic fibers are the dominant subset that respond to HKCA. Ca2+ flux, but not CGRP-release, can be abrogated by pretreatment of HKCA supernatant with apyrase. To determine whether non-peptidergic neurons participate in host defense, we generated MRGPRD-DTR mice. Infection in these mice resulted in increased CFU only for those C. albicans strains with high ATP secretion. Taken together, our findings indicate that C. albicans releases ATP, which is recognized by non-peptidergic nerves in the skin resulting in augmented anti-Candida immune responses. Author Summary Bacterial release of ATP has been shown to modulate immune responses. Candida albicans displays heterogeneity in ATP release among laboratory strains and commensal clinical isolates release more ATP than invasive isolates. C. albicans strains with high ATP secretion show lower fungal burden following epicutaneous infection. Mice lacking P2RX7, the key ATP receptor expressed by immune cells, showed no alteration in fungal burden. In contrast, treatment with P2RX2/3 antagonists resulted in increased fungal burden. P2RX3 is expressed by a subset of non-peptidergic neurons that terminate in the epidermis. These non-peptidergic fibers are the predominant responders when cultured sensory neurons are exposed to heat-killed C. albicans in vitro. Mice lacking non-peptidergic neurons have increased infection when exposed to high but not low ATP-secreting isolates of C. albicans. Taken together, our findings indicate that C. albicans releases ATP which is recognized by non-peptidergic nerves in the skin resulting in augmented anti-Candida immune responses. Bullet points ATP released from heat killed C. albicans activates non-peptidergic sensory neurons Live C. albicans clinical isolates release variable amounts of ATP Elevated levels of ATP released by C. albicans correlates with reduced infectivity in vivo MRGPRD-expressing cutaneous neurons are required for defense against ATP-secreting C. albicans


Introduction
Skin is a barrier tissue that is exposed to both commensal microorganisms and pathogens. Candida albicans is a dimorphic fungus that typically grows as a commensal at barrier surfaces, but it can also become pathogenic. Pathenogenic C.
albicans infection can manifest as chronic mucocutaneous candidiasis or disseminated candidiasis in immunocompromised individuals, leading to fungal sepsis in severe cases [1]. In healthy individuals, type-17 antifungal immunity limits C. albicans growth at barrier tissues [2]. Patients with genetic mutations in this pathway can suffer chronic mucocutaneous candidiasis [3]. In most mice strains, C. albicans is a foreign pathogen that is commonly used as a model to interrogate mechanisms of fungal immunity and host defense. As in humans, type-17 immunity mediates anti-candida responses at barrier tissues [2,4,5].
Neurons of the somatosensory nervous system, particularly unmyelinated free nerve endings that project extensively throughout barrier tissues, are ideally located to detect danger signals. Unmyelinated C-fibers can be divided into multiple categories based on unique gene expression patterns [9,10]. Broadly speaking, these can be divided into peptidergic neurons that express TRPV1 and CGRP, and non-peptidergic neurons, many of which express the ATP receptor P2RX3 [11,12]. In skin, TRPV1-expressing neurons terminate primarily within the dermis and in the stratum spinosum, while terminals of nonpeptidergic neurons extend into the stratum granulosum [12]. During epicutaneous C. albicans infection, activation of TRPV1-expressing neurons is both necessary and sufficient to trigger the development of protective type-17 host defense through a mechanism that depends on CGRP [13,14].
Neurons directly recognize both bacteria and C. albicans. Staphylococcus aureus triggers neuron activation as measured by Ca 2+ flux in sensory neurons [15].
Cutaneous neurons also detect C. albicans, but the mechanisms involved are less clear. C. albicans is sufficient to induce Ca 2+ flux in both TRPV1-expressing peptidergic neurons and non-TRPV1 expressing neurons [13]. Dectin-1 and TRP channels are required for optimal CGRP release [17] and exposure of cultured dorsal root ganglia (DRG) neurons to soluble β-glucan is sufficient to evoke CGRP release [18]. Despite strong evidence that TRPV1-expressing neurons recognize pathogens and modulate host defense, little is known about the role of TRVP1-negative sensory neurons in host defense.
Most non-TRPV1 expressing neurons flux Ca 2+ in response to C. albicans through an unknown mechanism [13]. Interestingly, ATP modulates β-glucan induced mechanical allodynia during C. albicans infection [18]. In this model, extracellular ATP (eATP) released from keratinocytes was suggested as the source. Notably, eATP is released by E. coli in the intestine as well as some other bacteria species and functions to modulate the development of local immune responses [19,20]. This raises the possibility that pathogen-derived eATP could trigger neuronal recognition of pathogens in skin.
Herein we show that C. albicans release eATP. The amount of eATP is highly variable between strains and positively correlates with the effectiveness of host defense that requires P2RX2/3 receptors. We also show that the major subset of non-peptidergic Cfiber neurons known to express P2RX3 are required for optimal host defense only against strains of C. albicans that express high levels of eATP. These data reveal an important unexpected host defense function for non-peptidergic neurons that relies on recognition of a novel C. albicans-derived PAMP.

ATP from heat-killed C. albicans induces Ca 2+ flux in cultured DRG neurons
In an effort to determine C. albicans-associated PAMPS responsible for activating cutaneous sensory neurons, we examined Ca 2+ flux in Fura2-AM labeled cultured DRGs. C. albicans strain SC5314 was harvested at mid-log phase and washed into neuron recording buffer (NRB) prior to heat killing at 100° for 30-60 minutes. Heat killed C. albicans (HKCA) cells were separated into washed yeast cell bodies and effluence by centrifugation. Live neurons, identified by a response to 50mM potassium chloride (KCl), fluxed Ca 2+ in response to HKCA effluence but not yeast cell bodies (Fig 1A-B).
Approximately 45% of the 154 neurons examined responded to HKCA effluence. Ca 2+ flux was evident within 1 second following application of effluence and responses were absent in Ca 2+ -free buffer (unpublished observation). Both capsaicin responsive (TRPV1+) and non-responsive (TRPV1-) neurons fluxed Ca 2+ in the presence of HKCA effluence (Fig 1A). This is consistent with our prior observations and suggests that multiple populations of sensory afferent neurons respond to C. albicans. β(1-3)-glucan is a well-defined component of the C. albicans yeast cell wall that drives recognition of C. albicans by cells of the innate immune system through interaction with the c-type lectin receptor Dectin-1 [21]. Although mRNA encoding for Dectin-1 (i.e. Clec7a mRNA) is not abundant in DRG single cell RNAseq datasets [9,22], incubation of cultured DRGs with C. albicans-derived soluble β-glucan is sufficient to trigger release of the neuropeptide CGRPα in a Dectin-1 dependent manner [18].
Unexpectedly, we observed that equal numbers of sensory neurons isolated from wild-type (WT) or Clec7a -/mice fluxed Ca 2+ in response to HKCA effluence (Fig 1C).
However, neurons isolated from Clec7a -/mice released significantly less CGRPα compared to neurons from WT mice (Fig 1D). These data support earlier reports that CGRPα release is at least partially dependent upon Dectin-1 [18], but also indicate that sensory neurons can be activated by a component in HKCA effluence other than βglucan that is not dependant on Dectin-1.
The immediate Ca 2+ flux observed in response to HKCA effluence (Fig 1A) suggests activation of a ligand gated ion channel. As many epidermal nonpeptidergic neurons express the P2RX3 receptor we examined the possible involvement of ATP [12]. To determine whether ATP could be responsible for Dectin-1 independent DRG activation, we first identified HKCA effluence-responsive cells based on Ca 2+ flux (Fig 1E).
Neurons were then tested for responsiveness to ATP by exposing them to 10μM ATP. PPADS, a selective P2 purinergic receptor antagonist that blocks activity of ligand-gated P2RX1-3 and P2RX5 receptors [23,24], was added to cultures prior to exposure to HKCA effluence. The presence of PPADS decreased the number total of neurons responding to HKCA effluence by 44% (Fig 1E-F). Notably, at least two kinds of HKCA effluence responsive neurons could be identified. Some neurons fluxed Ca 2+ following exposure to ATP and their response to HKCA effluence could be inhibited by PPADS.
Other neurons were unresponsive to ATP and their response to HKCA was not inhibited by PPADS.
To confirm that eATP in HKCA effluence was responsible for Ca 2+ flux, we first determined that HKCA effluence contains approximately 10µM ATP and it could be efficiently reduced by treatment with apyrase, an ATP-diphosphohydrolase that catalyzes the hydrolysis of ATP to AMP (S1A Fig). Exposure of cultured DRGs to apyrase treated HKCA effluence reduced the number of responding neurons by 68% compared with HKCA effluence (Figs 1G-H). Notably, release of CGRPα was unaffected by treatment with apyrase (Fig 1I). Moreover, αβme-ATP, a selective P2RX1, P2RX3, and P2RX2/3 agonist [25], was not sufficient to elicit CGRPα release ( Fig 1I). Taken together, these data support a model in which at least 2 subsets of neurons can respond to HKCA effluence. One releases CGRPα and is partially dependent on Dectin-1. This is likely the well-studied TRPV1 + expressing subset of cutaneous afferent neurons [13,14,18]. The other subset expresses a P2X receptor and is not linked to CGRPα release. These data also raise the exciting possibility that C. albicans secretes eATP, which could represent an unrecognized fungal PAMP that may participate in eliciting host defense.

C. albicans strains secrete varying amounts of ATP
To determine whether live C. albicans releases eATP, mid-log-phase yeasts of C. albicans strain SC5314 were washed and suspended into PBS, and either incubated at 30°C or heat killed at 100°C for 1 hr. The quantity of eATP in the centrifuged supernatant of live C. albicans showed a significant amount of ATP that approached the level obtained by heat killing, thereby demonstrating that live C. albicans can release eATP (Fig 2A). To evaluate whether eATP resulted from cell death, we performed a kinetic analysis of eATP during growth in YPD broth. SC5314 yeast were inoculated into YPD and levels of eATP in the broth were analyzed over the course of a 12 hour 30° shaking incubation. eATP in the broth was first detectable after 2 hours and peaked at 6 hours during mid-log phase. The same samples were analyzed for cell viability by Propidium iodide staining and flow cytometry. Control HKCA were >99% PI+ while samples grown in YPD showed less than 0.5% staining at all time points tested (Fig 2B,   S2 Fig). Thus, C. albicans SC5314 releases eATP during culture in both YPD and PBS that is unrelated to cell death, suggesting that eATP is actively secreted by viable cells.
To determine whether the SC5314 strain is unique in its secretion of ATP, we compared growth kinetics and eATP secretion in a collection of common prototrophic SC5314derived laboratory strains. Notably, SC5314 and CAI4 released eATP while RM1000, SN87HL and BWP17 released very little eATP (Fig 2C-D). We next compared the capacity of CAI4 and RM1000, two strains with similar growth kinetics, to infect mouse skin using an epicutaneous skin infection model [13,26]. On day 3 following infection, CAI4, which secretes eATP, resulted in lower colony forming units (CFU) when compared to RM1000, a genetically related strain, that does not secrete eATP (Fig 2E).
These data suggest that eATP enhanced C. albicans host defense though there are many genetic differences between CAI4 and RM1000 that could also potentially explain these results.
We next sought to identify a single gene mutant that results in reduced eATP. Our prediction was that these mutants should result in higher levels of infection in vivo. We focused exclusively on mutants with reduced eATP since the predicted increased levels of infection would be unlikely to result from the effects of genetic mutation that affect other aspects of C. albicans biology. A library screen of S. cerevisiae identified several genes required for release of eATP [27]. The majority of available C. albicans orthologous mutant strains, however, were on genetic background that did not release eATP (e.g. BWP17). We were able to identify one orthologous mutant, hda1 Δ/Δ , that showed normal growth kinetics compared with its parental strain (J4-2.1) but did not release eATP (Fig 2F-G). As predicted, on day +3 following epicutaneous infection, CFUs of the hda1 Δ/Δ mutant were significantly higher than those from control J4-2.1 cells (Fig 2H). Finally, we screened a series of dox-inducible OE strains based on our prediction from the data in Peters et al. (2016) that they would be affected in eATP secretion [28,29]We found two clones that secreted appreciable amounts of eATP and showed similar growth kinetics. Incubation with 50 μg/ml of doxycycline reduced eATP release from P TET -ORF19.2898 but not from P TET -O-ORF19.185 (Fig 2I-J). Both strains were grown in YPD augmented with doxycycline or vehicle and used to epicutaneously infect mice. On day +3 following epicutaneous infection only P TET -ORF19.2898 grown in the presence of doxycycline showed enhanced CFU (Fig 2K). From these data using laboratory strains of C. albicans, we conclude that reduced levels of released eATP resulted in enhanced epicutaneous skin infection in vivo.

C. albicans skin infectivity negatively correlates with ATP secretion
To extend our findings beyond laboratory strains of C. albicans, we screened for release of eATP across 89 well characterized clinical isolates [30]. Isolates were categorized into 3 groups: commensal, if they were obtained from asymptomatic barrier sites (i.e. mouth, GI, or vagina); superficial if they were obtained from symptomatic barrier sites (e.g. oral thrush, urine); or invasive if they were obtained from blood. Isolates were grown in YPD for 6 hrs and analyzed for OD 600 nm, levels of extracellular ATP in the YPD broth, and intracellular ATP levels from washed cells. Growth appeared relatively similar in most isolates (Fig 3A) but levels of eATP released into the broth varied across 3.5 logs and was independent of levels of intracellular ATP (Fig 3B-C). Notably, levels of eATP released by invasive blood isolates were approximately 10-fold lower than commensal and superficial isolates (Fig 3B). Release of eATP varied somewhat by genetic clusters (Fig S3A) suggesting that genetic differences determine levels of eATP. These data are consistent with our observations in laboratory strains that lower eATP is associated with greater infectivity in vivo.
To determine whether secretion of ATP correlates with skin infection, we selected paired high eATP-and low eATP commensal isolates from the mouth (D10:CEC3626, B7:CEC3530), GI tract (C4:CEC3548, C5:CEC3553) and vagina (D2:CEC3606, C6:CEC3556). These isolates transitioned normally from yeast to hypha when cultured in serum at 37° (S3B Fig), had similar growth kinetics (Fig 3D), but only C4, D10, and D2 released eATP (Fig 3E). Mice epicutaneously infected with high eATP strains from the mouth (D10), GI (C4), and vagina (D2) all showed lower CFU on day 3 following infection than their matched low eATP counterparts (Fig 3F). Moreover, the amount of eATP normalized to levels of intracellular ATP for every isolate tested has a very strong negative correlation with CFU (Fig 3G). Taken together, these data demonstrate that C. albicans releases eATP that varies greatly between strains. Moreover, low levels of eATP strongly correlate with increased infection efficiency suggesting that eATP released by C. albicans is recognized by the host and augments host defense.

Recognition of C. albicans-derived ATP is mediated by P2RX2/3
Most immune cells in the skin express the purinergic receptors P2RX7 and/or P2RX4 (Fig 4A; reviewed in [31,32]). Of particular interest, is P2RX7, which is known to participate in the development of cutaneous Type-17 immune responses [33]. To determine whether host defense to C. albicans requires P2RX7, we epicutaneously infected wild type (WT) and P2rx7 -/mice with a high ATP secreting strain (C4:CEC3548). On day 3 following infection, C. albicans CFU was equivalent in both groups indicating that P2RX7 is redundant for C. albicans host defense (Fig 4B). To test the requirement for P2RX4, WT mice were injected with 1mM 5-BDBD (a selective P2RX4 inhibitor) or vehicle intradermally (i.d.) starting on the day of C. albicans infection and twice daily continuing until harvest on day 3 (Fig 4C). Inhibiting P2RX4 did not affect C. albicans CFU.
In the skin, P2RX3 is primarily expressed by nonpeptidergic neurons particularly the subset identified based on expression of the Mas-related G-protein coupled receptor D (MRGPRD) [9,12,34]. To test the requirement for P2RX3, WT mice infected with C.
Based on our in vitro work (Fig 1) showing that eATP is not involved in release of CGRP, it is unlikely that CGRP-expressing TRPV1+ peptidergic neurons are the neuron subset responding to eATP. Rather, we hypothesize that MRGPRD+ nonpeptidergic neurons are the key cell type responding to C. albicans-derived eATP.

Augmented host defense to C. albicans-derived ATP requires MRGPRD+ neurons
MRGPRD-expressing non-peptidergic neurons are a distinct subset of c-fibers that terminate within the epidermis and are characterized by binding IB4 and high expression of GFRa2, and P2RX3 [12,37,38]. We bred MRGPRD-Cre [39] with ROSA26.TdTomato (TdT) reporter mice to confirm appropriate expression of Cre in cutaneous neurons. As expected, TdT was highly expressed by cutaneous neurons terminating in the epidermis (Fig 5A). Immunofluorescent microscopy of DRGs also revealed that TdT was expressed primarily by cells binding IB4 and less by cells expressing CGRPα or TRPV1 (Fig 5B and C). We next examined Ca 2+ flux following exposure to HKCA effluence in DRG cultured neurons from ROSA26.Tdt mice. Four distinct populations of cells could be identified (Fig 5C). The majority of effluence responding cells expressed TdT. Most did not respond to capsaicin indicating that these are MRGPRD+, TRPV1-nonpeptidergic neurons, but an overlapping capsaicinresponsive TdT+ population was evident. A small population of TdT-negative cells were capsaicin responsive indicating that these are likely TRPV1+ CGRP+ peptidergic neurons. A capsaicin-unresponsive TdT-population was also present.
To ablate MRGPRD neurons, we bred MRGPRD-Cre to ROSA26.DTR [39]. These mice allow for the inducible ablation of MRGPRD-expressing neurons following administration of Diphtheria toxin (DT). Immunofluorescent microscopy of epidermal whole mounts and transverse sections showed efficient ablation of epidermal GFRα2expressing neurons (Fig 6A, 6B). We were able to confirm the absence of cre expression in tissues outside the nervous system, such as in the aorta (not shown) and To determine whether MRGPRD+ neurons participate in host defense against C.
albicans, vehicle or DT treated MRGPRD-DTR mice were epicutaneously infected with either a high eATP strain (C4:CEC3548) or a low eATP strain (B7:CEC3530). On day 3 following infection, cutaneous C. albicans CFU was determined (Fig 5G). DT treated mice in which MRGPRD+ neurons were ablated showed a significantly higher CFU compared with vehicle treated mice when infected with the high eATP strain. In contrast, DT and vehicle treated mice showed equivalent CFU when infected with the low ATP-secreting strain. Notably, CFU in DT treated mice infected with the high eATP strain was similar to mice infected with the low eATP strain. From these data we conclude that the augmented host defense that occurs as a result of C. albicans-derived eATP requires epidermal MRGPRD+ neurons. Moreover, the fact that CFU was independent of the presence of MRGPRD+ neurons during infection with a low eATPsecreting strain indicates that MRGPRD+ nerves do not participate in host defense in response to host-derived factors including keratinocyte-derived eATP or to C. albicansderived factors other than eATP.

Discussion
By initially examining the components of C. albicans responsible for activating cutaneous sensory afferent neurons, we found that neurons were not activated by yeast cell bodies but that a large number of non-peptidergic, TRPV1-negative neurons were activated by eATP. Some laboratory C. albicans strains released eATP and genetic alterations that reduced eATP secretion resulted in increased epicutaneous infectivity.
Clinical isolates of C. albicans all released eATP with isolates from systemic infection showing greatly reduced levels of eATP. Infectivity across multiple isolates correlated well with increased eATP and decreased CFU recovered from infected tissue. Host recognition of C. albicans-derived eATP required P2RX3 which is expressed by epidermal non-peptidergic neurons. Ablation of MRGPRD non-peptidergic neurons resulted in greater infectivity for C. albicans that release eATP but had no effect on C. albicans releasing little, or no eATP. It is known that MRGPRD-expressing neurons respond to ATP [41] and these neurons have been implicated in modulating mechanosensation [42,43], histamine-independent itch [44] and extreme heat and cold sensation [45]. While these neurons are known to be involved in hyperalgesia following infection or injury [18,41], they have not previously been shown to play a role in the immune response to infection. Taken together, these data demonstrated that nonpeptidergic cutaneous sensory afferents sense C. albicans through a novel, unappreciated mechanism that mediates host defense.
The release of eATP by C. albicans was unexpected. Although it is undetectable in some commonly used laboratory strains, all of the 89 clinical isolates we examined released amounts of eATP that were easily detectable. Furthermore, eATP levels were independent of levels of intracellular ATP, which is well known to be essential for growth, filament formation, and successful invasion [22]. ATP is released by S. cerevisiae and a mutational analysis revealed that many genes related to the secretory pathways were required for ATP efflux [27]. This suggests that release of eATP occurs through a defined mechanism and is not unique to C. albicans. Several bacterial species are also known to secrete eATP [46][47][48][49]. Notably, eATP from gut commensal bacteria is biologically active and is well documented to modulate the host immune response [19,20,50]. Interestingly, levels of eATP released by clinical isolates of C.
albicans were highly variable. Invasive isolates identified in blood released less eATP and were over-represented in C. albicans clade 4 suggesting genetic control of eATP release. The observation that multiple microorganisms actively release ATP suggests that eATP may be beneficial to the yeast. In the case of C. albicans, we speculate that eATP may be related to a C. albicans-intrinsic process such as environmental sensing antagonists, PPADS and A-317491, had increased CFU in a skin infection model. In the skin, P2RX2/3 expression is largely restricted to neurons. P2RX3 is expressed in many non-peptidergic fibers [12] and very highly expressed in the MRGPRD subset [9].
Approximately 92% of MRGPRD-expressing neurons flux Ca 2+ in response to 50µM ATP [41]. P2RX2/3 is not highly expressed in immune cells based on publicly available datasets (Immgen) or in other cells in the skin [51], nor is it highly expressed in glia [10].
We thus focused on the MRGPRD-expressing subset of nonpeptidergic sensory afferents.
In mouse, TRPV1-expressing and MRGPRD-expressing neurons are largely nonoverlapping subsets. Only 5% of MRGPRD neurons respond to 1µM Capsaicin, and TRPV1 is expressed by only 9% of MRGPRD neurons [12,41]. We observed that The ablation of MRGPRD-expressing neurons resulted in increased CFU of high eATP releasing, but not low eATP releasing, C. albicans strains. Thus, this neuron subset, which has dense free nerve endings in the epidermis, plays an important role in recognizing eATP and mediating the host response. This represents a novel function for this subset of sensory afferents. The absence of a phenotype with low eATP releasing C. albicans indicates that ATP from other potential sources during infection, such as keratinocytes, either does not activate these neurons or has no role in host defense. Possible mediators expressed by MRGPRD + neurons that could modulate host defense are unclear, but a number of biologically active soluble mediators are expressed by these neurons [9,51].
In summary, we have identified that two major subsets of cutaneous sensory neurons recognize distinct Candida-associated PAMPS and together provide synergistic host defense. The mechanistic details of this synergy as well as the benefit to C. albicans derived from eATP release and its genetic control all remains as exciting future areas of exploration.

Animals
Mice used in this study are described in supplemental methods. Male and female mice between the ages of 7-13 weeks were used. Mice were housed in microisolator cages and fed irradiated food and acidified water.

Candida strains and culturing
C. albicans SC5314, lab strains created from SC5314 [53], and lab strains created from clinical isolate SZ306, were all included in this study (see Supplemental Methods Table   1). Additional clinical isolates of different origins were previously described [30]. To measure growth, C. albicans was cultured in liquid YPD at 30°C/ 250 rpm overnight ~16 hrs. These cultures were then diluted 1:50 in fresh YPD and incubated as above.
Optical density was determined at the indicated time points. P TET -driven over-expression strains were cultured in YPD containing 50 μg/mL doxycycline and in the same medium without doxycycline as a control [28].For induction of hyphae in liquid medium, C. albicans was sub-cultured for 3 hrs at 37°C in RPMI 1640 medium plus 10% (vol/vol) heat-inactivated FBS [54].
To stimulate neurons in culture, SC5314 was grown, as above, to ~OD1.5, and washed three times in neuron recording buffer (NRB) to remove residual YPD. Candida was then resuspended in 10 mL NRB and heat-killed in a water bath for 30min-1hr at 100°C (assay dependent). HKCA was centrifuged at 3000RPM/5min and the NRB in which they were heat-killed -what we call effluence, was collected. HKCA were washed again and resuspended in NRB to a concentration of 1x10 7 /mL. HKCA and effluence were stored at -80°C until use. A similar method was used to create HKCA in PBS for ATP measurement.

Neuron cultures
The caudal 6 thoracic DRGs and all 6 lumbar DRGs were dissected out and cultured for Ca 2+ imaging, while all thoracic and lumbar DRGs were used for CGRP release assays.
Dissection and methods for culture are based on those described in [55]. Following euthanasia and perfusion with ice cold Ca 2+ /Mg 2+ -free HBSS, isolated ganglia were placed into cold HBSS. In a two-step enzymatic digestion, DRGs were transferred to 3mL of filter-sterilized 60U papain/ 1mg L-cys/ HBSS buffered with NaHCO 3 at 37°C for 10 minutes, then to filter sterilized collagenase II (4 mg/mL)/dispase II (4.67 mg/mL) in HBSS for 20 minutes at 37°C. Enzymes were neutralized in Advanced DMEM/F12 media with 10% FBS and 1% penicillin/streptomycin. Neurons were dissociated in media with trituration through a series of increasingly smaller-bored fire-polished glass pipettes. Cells for Ca 2+ imaging were plated overnight in Advanced DMEM/F12 (as above) on poly-d-lysine coated 12mm round coverslips (Corning, Corning, NY). Cells for CGRP release assays were plated at a concentration of 20,000 cells/well in a 12 well plate in media supplemented with 10 ng/mL NGF 2.5S (Harlan, Indianapolis, IN) or 7S (ThermoFisher, Pittsburgh, PA) and cultured for ~48 hr.

Calcium imaging
Coverslips coated with neurons were transferred to a dish containing 2 mL 0.5% BSA in

CGRP release assay
The CGRP release assay was performed as previously described [13], but with slight modifications. After 48 hours in culture, 20,000 neurons/well were washed with 1xHBSS for 5 minutes and then incubated in NRB for 10 minutes to measure basal release.

Determination of ATP level of C. albicans
Aliquots of supernatant were collected from C. albicans cultures during growth and centrifuged at 3000 rpm for 5 mins at 4°C to remove yeast. ATP level in supernatant was determined using ATPlite luminescence assay system (PerkinElmer, Waltham, MA), as previously described [48]. To determine intracellular ATP level, the C. albicans pellet at each indicated time was washed in PBS and incubated for 30 min with ATPlite lysis buffer at room temperature. The extract was then spun down at 3000 rpm for 5 mins and the supernatant was transferred to a fresh tube for the ATP assay.
Measurements were made with SOFTmax Pro for Lmax 1.1 software and a Lmax Microplate Luminometer (Molecular Devices, San Jose, CA).
One hour treatments were determined to be sufficient for efficient ATP removal and thus were used in neuron stimulation assays.

Skin Infection Model
Skin infection was performed as previously described [26]. Mice were anesthetized with ketamine and xylazine in sterile PBS (100/10 mg/kg body weight), their backs shaved with an electric clipper, and then chemically depilated with Neet sensitive skin hair remover cream (Reckitt Benckiser, Slough, England, UK). The stratum corneum was removed with 15-17 strokes using 220 grit sandpaper (3M, St Paul, MN). 2x10 8 C.
albicans in 50 μl of sterile PBS was then applied to the upper back.
For ATP receptor antagonist treatments, mice were injected intradermally across 5 sites on the back with a total of 100µL antagonist (500µM PPADS, 500µM A317491 in PBS or 1mM 5BDBD in DMSO/PBS) or PBS (with DMSO for 5-BDBD control) after sandpapering, but before application of Candida. Treatments were additionally given twice-daily at 12 hour intervals on d+1 and d+2. Skin was harvested 3-days postinfection, homogenized, and serially diluted onto YPAD plates. Plates were incubated at 30°C for 24-48 hours to determine colony forming units (CFU)/cm 2 .

Immunofluorescent microscopy
For sectioned tissue, dissected mouse skin or DRGs were fixed in 4% paraformaldehyde (PFA) at RT for 1hr, washed twice in PBS, cryoprotected in 12% then

Flow cytometry
Single cell suspensions were obtained from uninfected skin and lymph nodes as previously described [13]. Intracellular CD207 staining was performed using BD

Statistical Analysis:
Statistical analyses were performed using GraphPad Prism software and is represented graphically as mean ± standard error. Chi squared analysis was used for comparisons of groups for Ca 2+ imaging data, while Mann-Whitney and Student's t test were used for comparisons between two groups. Kruskal-Wallace one-way analysis of variance (ANOVA) was used to compare corresponding groups as indicated.

Ethics Statement
All animal experiments were approved by The University of Pittsburgh institutional care and use committee (IACUC #19126552). Mice were anesthetized with ketamine and xylazine in sterile PBS (100/10 mg/kg body weight). Euthanasia was performed using CO2.