Candida albicans triggers NADPH oxidase-independent neutrophil extracellular traps through dectin-2

Candida albicans is one of the top leading causes of healthcare-associated bloodstream infection. Neutrophil extracellular traps (NET) are known to capture and kill pathogens. It is reported that opsonized C. albicans-triggered NETosis is NADPH oxidase-dependent. We discovered a NADPH oxidase-independent NETosis pathway in neutrophil response to unopsonized C. albicans. While CR3 engagement with opsonized C. albicans triggered NET, dectin-2 recognized unopsonized C. albicans and mediated NET formation. Engagement of dectin-2 activated the downstream Syk-Ca2+-PKCδ-protein arginine deiminase 4 (PAD4) signaling pathway which modulated nuclear translocation of neutrophil elastase (NE), histone citrullination and NETosis. In a C. albicans peritonitis model we observed Ki67+Ly6G+ NETotic cells in the peritoneal exudate and mesenteric tissues within 3 h of infection. Treatment with PAD4 inhibitor GSK484 or dectin-2 deficiency reduced % Ki67+Ly6G+ cells and the intensity of Ki67 in peritoneal neutrophils. Employing DNA digestion enzyme micrococcal nuclease, GSK484 as well as dectin-2-deficient mice, we further showed that dectin-2-mediated PAD4-dependent NET formation in vivo restrained the spread of C. albicans from the peritoneal cavity to kidney. Taken together, this study reveals that unopsonized C. albicans evokes NADPH oxidase-independent NETosis through dectin-2 and its downstream signaling pathway and dectin-2-mediated NET helps restrain fungal dissemination.

Introduction Candida albicans is a commensal in the mucosa surface and skin in most humans. Environmental changes in temperature, nutrition, or the presence of serum induce its transformation from yeast to hyphae. C. albicans infection is one of the top leading causes of overall healthcare-associated bloodstream infection in medical centers as well as regional hospitals. Invasive candidiasis affects more than 250,000 people worldwide each year and leads to more than 50,000 deaths. Mortality among patients with invasive candidiasis is as high as 40% even after receiving antifungal therapy [1][2][3]. Patients with neutropenia or genetic deficiency in NADPH oxidase are susceptible to invasive candidiasis [4,5], showing that neutrophils and NADPH oxidase activation are indispensable for host defense against C. albicans infection.
NADPH oxidase activation requires the assembly of its regulatory subunits, p40 phox , p47 phox , and p67 phox with its core proteins gp91 phox and p22 phox , resulting in ROS production [6,7]. In addition to generating ROS, NADPH oxidase activation also induces neutrophil release of nuclear DNA to form a sticky web-like structure named neutrophil extracellular trap (NET) that binds histones, granular proteins and antimicrobial peptides [8]. In vitro studies showed that pathogens trapped by NET are in contact with and killed by concentrated antimicrobial factors [9,10]. NET is known to capture and kill C. albicans through a NADPH oxidase-dependent mechanism [11]. However, it is also reported that human neutrophils are capable of killing unopsonized C. albicans through a ROS-independent mechanism [12]. Whether C. albicans can induce NET through a NADPH oxidase-independent mechanism is a question to be addressed.
The process of NET formation is called NETosis. Neutrophils undergoing NETosis are characterized by disintegrated nuclear envelop and release of decondensed chromatin into the cytoplasm [8]. Recent study uncovers that cell cycle pathway controls NETosis. NETotic neutrophils have phosphorylated retinoblastoma protein and lamins and express cell cycle marker Ki67 [13]. Chromatin decondensation is the result of protein arginine deiminase 4 (PAD4)dependent histone citrullination and neutrophil elastase (NE)-mediated histone degradation [14,15]. NADPH oxidase facilitates both nuclear translocation of NE and PAD4 activation through stimulating myeloperoxidase activation [16]. It has been shown that opsonized C. albicans induces NET through autophagy, ROS, and NE, but not PAD4, apoptosis nor necroptosis [17]. Unopsonized C. albicans is also known to induce NET formation [11]. Since neutrophils use different receptors to recognize serum-opsonized and unopsonized C. albicans [12], it is important to investigate the receptor and the molecular mechanism by which unopsonized C. albicans uses to evoke NET.
Multiple receptors participate in modulating neutrophil anti-C. albicans functions. Fcγ receptor mediates human neutrophil killing of antibody-opsonized fungus through the Syk and PKC signaling pathways, whereas complement receptor 3 (CR3) are involved in killing of unopsonized C. albicans [12]. Mouse neutrophils utilize CR3 for recognition and killing of opsonized C. albicans [18]. Dectin-2 is marginally involved in opsonized C. albicans-induced neutrophil ROS production [19]. Dectin-1 as a phagocytosis receptor for C. albicans yeasts negatively regulates NETosis through interfering with nuclear translocation of granule NE [20]. Although CR3 as a receptor recognizing fibronectin-coated matrix is responsible for C. albicans-induced NETosis [21], which receptor(s) mediates NET formation in response to unopsonized C. albicans alone awaits to be determined.
Here we sought to study the receptor and signaling pathway that mediate unopsonized C. albicans-induced NET formation. Our study revealed the role of dectin-2 and its downstream Syk-Ca 2+ -PKCδ-PAD4/NE pathway in inducing NETosis in a NADPH oxidase-independent manner. Dectin-2 functions to restrain C. albicans spread from peritoneal cavity to kidney through modulating NET.

Both opsonized and unopsonized C. albicans induce NET formation
It has been reported that opsonized C. albicans induces NETosis [17,22,23]. Our results revealed that neutrophils released web-like extracellular DNA fibers in response to opsonized as well as unopsonized C. albicans ( Fig 1A). NETosis is characterized by disintegration of the nuclear envelope [8]. While stimulation by opsonized C. albicans resulted in nuclear membrane disintegration (Fig 1B), we also observed nuclear envelope breakdown and cytoplasmic membrane rupture following stimulation by unopsonized C. albicans (Fig 1B). Fluorescence images at high magnification clearly demonstrated that similar to opsonized C. albicans stimulation, unopsonized C. albicans hyphae were entangled with histone H3-containing web-like extracellular DNA structure (Fig 1C), although the percentage of NETotic cells in response to unopsonized C. albicans (6.8%) was lower than that to opsonized organisms (13.9%) (Fig 1D). Since opsonin-containing fresh serum also facilitates the germination of C. albicans, we allowed C. albicans yeasts to germinate before adding them to wells containing neutrophils. PicoGreen dsDNA assay showed that the hyphal form but not yeast-locked C. albicans (strain HLC 54) induced NETosis whether it was opsonized or not (Fig 1E), signifying the importance of hyphal formation in triggering NETosis. Our data indicate that not only opsonized but also unopsonized C. albicans in its hyphal form induces NETosis.

Unopsonized C. albicans-induced NET formation is independent of NCF-1
We further explored the requirement of ROS in unopsonized C. albicans-induced NET formation. Results showed that NCF-1 (NADPH oxidase subunit p47 phox )-deficient neutrophils formed histone H3-containing web-like NET structure as readily as NCF-1-sufficient cells upon stimulation by unopsonized C. albicans (Fig 2A and 2B). Time-lapse live cell imaging showed that neutrophils underwent robust NETosis after encountering opsonized C. albicans (Fig 2C and S1 Video) whereas NET formation induced by unopsonized organism was less so (Fig 2C and S2 Video). Importantly, similar to stimulation of NCF-1-sufficient neutrophils with opsonized or unopsonized C. albicans (Fig 2C, S1 and S2 Videos), stimulation of NCF-1-deficient neutrophils by unopsonized C. albicans resulted in loss of lobular shape in the nucleus and chromatin decondensation (Fig 2C and S3 Video). These data demonstrate that NCF-1-deficient neutrophils underwent NETosis after C. albicans challenge. Consistently, Ncf-1 -/and Ncf-1 +/+ neutrophils releases comparable levels of dsDNA in response to stimulation by unopsonized C. albicans, whereas the response was greatly reduced in Ncf-1 -/neutrophils upon challenge with opsonized organisms (Fig 2D). Additionally, treatment with MitoTEMPO (mitochondrial ROS inhibitor) did not affect NET formation in neutrophils Yellow arrows point to disintegrated nuclear envelop, whereas the red arrow points to disrupted cell membrane where cytosolic materials are being released. (C) Cells were stained with anti-histone H3 antibody (green), cell-impermeable DNA dye SYTOX Orange (red), and cell-permeable DNA dye stimulated by unopsonized C. albicans (Fig 2E). Our results indicate that unopsonized C. albicans-induced NET formation is independent of NADPH oxidase and mitochondrial ROS.

Dectin-2 mediates NET formation through Syk-Ca 2+ -PKCδ signaling pathway in response to unopsonized C. albicans
We used pharmacological inhibitors to inhibit activation of signaling molecules and found that inhibition of Syk, Ca 2+ influx, and PKCs significantly diminished NET formation (Fig 4A,  4B and 4C). While different isoforms of PKC family have their unique roles in modulating NETosis [24], our results showed that inhibition of PKCδ dose-dependently, but not PKCα and PKCβ, reduced the level of NETosis (Fig 4D). These results together indicate that Syk, Ca 2 + influx, and PKCδ are involved in unopsonized C. albicans-induced NET. Furthermore, cells treated with Syk inhibitor had lower levels of Ca 2+ influx, less Ca 2+ -positive cells (S1 Fig and  Fig 4E) and lowered the level of phosphorylated PKCδ (Fig 4F). Cells treated with Ca 2+ chelator had lower levels of phosphorylated PKCδ but not that of phosphorylated Syk (Fig 4G). In line with the observation that dectin-2 deficiency reduced the levels of phosphorylated Syk and PKCδ after stimulation (Fig 4H), our results clearly demonstrate that unopsonized C. albicans induces NETosis through dectin-2 and its downstream Syk-Ca 2+ -PKCδ pathway.
Hoechst 33342 (blue). Immunofluorescence images were viewed under fluorescence microscope at 3 h of stimulation. DIC, differential interference contrast image. Images in the boxed areas are enlarged and shown on the right. (D) % NETotic cells = the number of cells that had NET morphology (SYTOX Orange + , web-like structure) after stimulation with opsonized (Ops) or unopsonized (Unops.) C. albicans divided by the total number of cells (blue) counted in images as prepared in (C). (E) C. albicans strains HLC 54 (yeast-locked) and SC 5314 (germination-competent) were incubated in RPMI medium for 4 h to allow competent cells to germinate. Neutrophils were then stimulated with opsonized and unopsonized HLC 54 (Yeast) and SC 5314 (Hyphae). Ctrl for the unops. group was cells incubated in HBSS only. Ctrl for the ops. group was cells incubated in HBSS containing 5% mouse serum. Extracellular DNA was quantified by Quant-iT PicoGreen dsDNA assay (n = 3). � , p < 0.05; ��� , p < 0.005; n.s., not significant, as analyzed by Student's t test. https://doi.org/10.1371/journal.ppat.1008096.g001

NE nuclear translocation is involved in NCF-1-independent NETosis through Syk-Ca 2+ -PKCδ
Neutrophils treated with neutrophil elastase inhibitor sivelestat had reduced NET formation upon stimulation by unopsonized C. albicans (Fig 5A). Immunofluorescence images showed that NE was distributed in the cytoplasm and was separated from the nuclear region before stimulation ( Fig 5B). Responding to unopsonized C. albicans challenge, NE aggregated into larger puncta and started translocating to the nucleus, especially to the decondensed area by 1 h of stimulation (Fig 5B). At 2 h after stimulation, the granules containing NE began to disintegrate, and NE was localized in the nucleus of cells that was ready for NETosis (2 h, Fig 5B). By 3 h after stimulation when NETotic structure began to form, NE was released to the extracellular space and bound to extracellular DNA fibers (3 h, Fig 5B). While nuclear translocation of NE occurred after stimulation with unopsonized C. albicans (Ctrl. in Fig 5C), neutrophils treated with pharmacological inhibitor to Syk, Ca 2+ influx, PKCδ or NE exhibited condensed chromatin structure and their NE remained in the perimeter of the nucleus (Fig 5C), suggesting that Syk, Ca 2+ influx, PKCδ and NE activity regulate chromatin decondensation as well as NE nuclear translocation. Together these data indicate that dectin-2 downstream signaling pathway mediates NE translocation. Importantly, NCF-1 deficiency did not affect NE nuclear translocation upon stimulation with unopsonized C. albicans

Unopsonized C. albicans-induced NET formation is dependent on PAD4 enzymatic activity
Treatment with inhibitor to PAD1-4 or PAD4 reduced unopsonized C. albicans-induced NET formation ( Fig 6A). PAD4 inhibitor also impeded NE nuclear translocation and chromatin decondensation ( Fig 6B) as well as the formation of histone H3-containing web-like NET structure ( Fig 6C). Thus, PAD4 activity is required for unopsonized C. albicans-induced nucleus decondensation and NE nuclear translocation. Interestingly, inhibition of PKCδ reduced the level of citrullinated H3 (Fig 6D). Since histone H3 citrullination is catalyzed by PAD4 [14], these results show that recognition of unopsonized C. albicans by dectin-2 leads PAD4-dependent NET formation.

NET formation in peritoneal cavity after C. albicans infection
A C. albicans peritonitis model was established to study the role of NET formation in vivo.
Since there are few neutrophils in the peritoneal cavity of normal mice, we gave mice two intraperitoneal injections of casein 18 h apart to enrich the neutrophil population. Mice were given C. albicans yeasts intraperitoneally 4 h after the second injection of casein when cells were stained with cell-permeable DNA dye Draq5 (blue) and cell-impermeable DNA dye SYTOX Orange (red) before stimulation with GFPexpressing C. albicans strain OG1 (green). NETosis in response to opsonized and unopsonized pre-germinated C. albicans was observed over 180 min after addition of C. albicans. Zeiss LSM 780 confocal microscope was employed for time-lapse imaging (Images were obtained from S1, S2 and S3 Videos separately). (D) Cells were stimulated with opsonized and unopsonized C. albicans for 3 h. Cell supernatants were collected for Quant-iT PicoGreen dsDNA assay. (n = 3) (E) Cells were pretreated with 5, 10, 20 μM of mitochondria ROS inhibitor, MitoTEMPO, for 30 min before stimulation with unopsonized C. albicans. Cell supernatants were collected for Quant-iT PicoGreen dsDNA assay. (n = 3). All experiments were performed three times. Data from one representative experiment are presented as mean ± standard deviation (SD). �� , p < 0.01; n.s., not significant, as analyzed by Student's t test.
https://doi.org/10.1371/journal.ppat.1008096.g002 Data from one representative experiment are shown and presented as mean ± SD. � , p < 0.05; �� , p < 0.01; n.s., not significant, as analyzed by Student's t test. (E) Neutrophils were stimulated with pre-germinated (Unops. Hyphae) or ungerminated (Unops. Yeast) GFP-expressing C. albicans OG1 (green). Cells were cytospun, permeabilized and stained for dectin-2 (red) and nucleus (blue). Arrows point to where dectin-2 contacts either the yeast or hyphal form of the fungus. (F) Clec4n +/+ and clec4n -/neutrophils were stimulated with unopsonized C. albicans at MOI of 2 for 3 h and stained with anti-histone H3 antibody (green), cell-impermeable DNA dye SYTOX Orange (red), cell-peritoneal neutrophils constituted about 78.8% of the whole peritoneal cell population ( Fig  7A). In vivo imaging system (IVIS) spectrum images showed that C. albicans infection induced release of extracellular DNA into peritoneal cavity as early as 1.5 h after infection and remained at relatively the same level until 3 h later ( Fig 7B). Web-like DNA structures that were positive for Ki67 (a novel marker for mature neutrophils that undergo NETosis [13]), histone H3, and Ly6G cells were observed in peritoneal exudates from mice given C. albicans ( Fig  7C). Cells on the mesenteric tissues collected from infected mice also stained positive for Ki67 and Ly6G ( Fig 7D). In the peritonitis candidiasis model, we observed NET formation in the peritoneal cavity and NETotic cells on the mesenteric tissues.

NCF-1-independent NETosis restrains C. albicans spread from peritoneal cavity to kidney
To monitor C. albicans spread, we infected mice with dTomato-expressing C. albicans when neutrophils were enriched in the peritoneal cavity. IVIS images showed that the intensity of fluorescence in the peritoneal cavity remained at relatively the same level at 1 and 2 h after infection and decreased thereafter ( Fig 8A). Coinciding with decrease in intensity of dTomato, fungal burden in the peritoneal cavity also decreased by 3 h after infection and in the meantime, it was increased in the kidney ( Fig 8B). We then treated mice with NET digestion enzyme micrococcal nuclease (MNase) or its heat-inactivated form (h.i. MNase) intraperitoneally and discovered that treatment with MNase compared to h.i. MNase decreased fungal burden in the peritoneal cavity [from (3.4 ± 2.3) × 10 5 to (1.6 ± 0.7) × 10 5 CFU] and increased that in the kidney [from (2.1 ± 1.0) × 10 4 to (3.6 ± 0.7) × 10 4 CFU/kidney) ( Fig 8C). These results demonstrate that NET functions to restrain C. albicans in the peritoneal cavity and keep it from spread to the kidney. In the meanwhile, Ncf-1 -/mice were infected by C. albicans intraperitoneally to assess whether NCF-1 participates in inducing NETosis in vivo. Data in S3A and S3B Fig showed that NCF-1 deficiency did not affect NET formation in peritoneal exudate nor did it affect the Ki67 + Ly6G + population in mesenteric tissues, indicating that an NCF-1-independent NETosis response to C. albicans occurs in vivo. Interestingly, however, Ncf-1 -/mice had significantly higher fungal burden in the peritoneal cavity but a comparable level in the kidney compared to Ncf-1 +/+ mice at 3 h after infection (S3C Fig). Since ROS is important to phagocytic cell clearance of C. albicans [18], these results indicate that unopsonized C. albicansinduced NCF-1-independent NETosis that restricts fungal spread does occur in vivo, but fungal clearance involves more than just NETosis.

PAD4 is important to prevent fungal spread
Our data in Fig 6 showed that PAD4 is involved in NE nuclear translocation and inducing NET formation in vitro. To explore the role of PAD4 in C. albicans infection, we treated mice with PAD4 inhibitor GSK484. Mice were given GSK484 before intraperitoneal injection of C.
permeable DNA dye Hoechst 33258 (blue). Immunofluorescence images were viewed under fluorescence microscope. DIC, differential interference contrast image. Yellow arrows point to cells that undergo NETosis. (G) WT, Itgam -/-, and clec4n -/neutrophils were incubated with unopsonized C. albicans at MOI of 2 in HBSS supplemented with 10 U/ml of MNase or heat-inactivated MNase (h.i. MNase). Wells containing C. albicans only without neutrophils were used as control. Controls were incubated in medium containing h.i. MNase or MNase. Three hours after incubation, medium was collected and cold H 2 O (pH = 11) was added to lyse cells. C. albicans was detached by mini cell scraper and vigorous pipetting. The number of viable fungi was determined by plating the supernatant on yeast-peptone-dextrose agar plate. Colony counts (CFU) were enumerated 2-3 days later. The ability of neutrophils to kill Candida is presented as % killing of C. albicans which was calculated by dividing the difference of CFU counts between the control group (without neutrophils) and neutrophiladded groups with MNase or h.i. MNase treatment by the counts of respective control. WT, n = 9; Itgam -/and clec4n -/-, n = 5 each. Each n represents neutrophils collected from one mouse. Data were pooled from 3 independent experiments and presented as mean ± SD. � , p < 0.05; n.s., not significant, as analyzed by Student's t test by comparing the 2 groups linked by a bracket.
https://doi.org/10.1371/journal.ppat.1008096.g003 After treatment, cells were stimulated (DMSO and Syki) or not (Ctrl) with unopsonized C. albicans at MOI of 4. Intracellular Ca 2+ content was analyzed by flow cytometry from before stimulation until 450 sec after. Arrow points to the time when C. albicans was added. Maximal % of Ca 2+ -positive cells in unopsonized C. albicans-stimulated group was taken as 100% intensity (Max). The Ca 2+ response of other time points was normalized against the maximal response and is shown as relative Ca 2+ -positive cells (%) (calculated under the kinetic mode of FlowJo software). Line graphs showing the kinetics of Ca 2+ responses in the three groups were analyzed and overlaid by FlowJo software. Bar graph on the right shows % Ca 2+ -positive cells at maximal response (20-50 sec, gating strategy is shown in S1 Fig). (F, G) Cells were pre-treated with Syk inhibitor SykI (F) or Ca 2+ chelator BAPTA-AM (G) before stimulation by pre-germinated unopsonized C. albicans at MOI of 2. At 30 min of stimulation, cell lysates were collected and subject to Western blot analysis for phosphorylated-Syk and -PKCδ. β-actin was used as loading control. Relative intensities of p-Syk and p-PKCδ are quantified by ImageJ and shown as bar graphs next to the blot. (H) WT and clec4n -/neutrophils were stimulated with pre-germinated unopsonized C. albicans at MOI of 2 for 15 and 30 min. Cell lysates were collected and subject to Western blot analysis for p-Syk and p-PKCδ as described in (G). Relative intensities of p-Syk and p-albicans. Results showed that inhibition of PAD4 reduced the formation of web-like structure and Ki67 expression in peritoneal Ly6G + cells and Ki67 + Ly6G + cell population in mesenteric tissues (Fig 9A and 9B). Flow cytometric analysis also revealed that GSK484 treatment reduced the percentage and the level of Ki67 in peritoneal infiltrating neutrophils (Fig 9C). In addition, GSK484 treatment decreased fungal CFU in the peritoneal cavity and increased that in the kidney by 3 h after infection (Fig 9D). Results of our in vitro (Fig 6) and in vivo studies together demonstrated that PAD4 regulates NETosis response to C. albicans.
Thus, it appears that CR3 recognition of opsonized C. albicans sends signals to activate NADPH oxidase-dependent NET formation through Syk-PKC-ROS cascade, but PAD4 and Ca 2+ do not take part in NET formation. Our study employing confocal microscopy, fluorescence microscopy, transmission electron microscopy, live cell imaging and PicoGreen dsDNA assay shows that unopsonized C. albicans triggers neutrophils to undergo a NADPH oxidaseindependent NETosis. Distinct from opsonized organisms, unopsonized C. albicans-induced NET is through ligation of dectin-2 that drives Syk-Ca 2+ -PKCδ-NE/PAD4 signaling pathway. Our study together with those of others reveal that opsonized and unopsonized C. albicans utilize different receptors and different signaling pathways to trigger NETosis.
Employing PAD4 -/and WT neutrophils Guiducci et al. showed in their recent publication [23] that in response to opsonized C. albicans, PAD4 deficiency reduced histone H3 citrullination (confocal microscopy of neutrophils in vitro). However, they found that PAD4 was not PKCδ in stimulated cells were normalized against their respective unstimulated controls and shown as relative fold induction. All Western blot experiments were performed three times. Data from one representative experiment are shown. Data are presented as mean ± standard error of the mean (SEM). � , p < 0.05; �� , p < 0.01, as analyzed by Student's t test comparing the 2 groups linked by a bracket. https://doi.org/10.1371/journal.ppat.1008096.g004 Candida albicans triggers NADPH oxidase-independent NET through dectin-2 . Immunofluorescence images were viewed under confocal microscope. DIC, differential interference contrast image. Ctrl., cells incubated in HBSS containing 0.1% DMSO. (C) Neutrophils pretreated or not (Ctrl) with PAD4 inhibitor (10 μM of GSK484, PAD4i) were stimulated with unopsonized C. albicans at MOI of 2 for 3 h. Cells were stained with anti-histone H3 antibody (green), cell-impermeable DNA dye SYTOX Orange (red), cell-permeable DNA dye Hoechst 33258 (blue). Immunofluorescence images were viewed under fluorescence microscope. DIC, differential interference contrast image. Arrows point to H3-containing web-like structure. (D) Cells were pre-treated with PKCδ inhibitor (PKCδI, 20 μM of Rottlerin) before stimulation with C. albicans. At 30 min of stimulation, cell lysates were collected and subject to Western blot analysis for citrullinated-histone H3 (cit.H3). GAPDH was used as a loading control. The experiment was performed 3 times. Data from one representative experiment are shown. Relative intensities of cit.H3 against GAPDH are shown below the blot. Data are presented as mean ± standard error of the mean (SEM). � , p < 0.05; �� , p < 0.01, as analyzed by Student's t test comparing the two groups treated with and without inhibitor (A, D).
https://doi.org/10.1371/journal.ppat.1008096.g006 WT mice were given peritoneal injections of casein as described above. At 4 h after the second casein injection, mice received 1 × 10 8 of C. albicans intraperitoneally. (B) Mice were injected with SYTOX Orange intraperitoneally at the same time when C. albicans strain SC5314 was administered. Mice were imaged on the side by IVIS (Ex/Em = 570/620) to record SYTOX Orange signals for 3 h starting at the time when C. albicans was administered. Photons in user-specified region of interest (ROI, gated area) was measured by Living Image 3.2 software. Relative intensity of total photons in ROI at each time point was required for opsonized C. albicans-induced NETosis (quantified by SytoxGreen assay, immunofluorescence staining for confocal microscopic imaging and electron microscopic imaging). We showed that GSK484 treatment of neutrophils challenged with unopsonized C. albicans reduced NE translocation (immunofluorescence staining for confocal microscopic imaging), Sytox Organge + histone H3 + web-like structure formation (immunofluorescence staining for fluorescence microscopic imaging) and NETosis (quantified by Quant-iT PicoGreen dsDNA assay). Therefore, it appears that unlike challenge with opsonized C. albicans, PAD4 activation results in NETosis when neutrophils are challenged with unopsonized organisms in vitro. Furthermore, it is shown that PAD4 deficiency increased fungal CFU in kidney on day 3 and 7 after intravenous infection, decreased that in the tongue on day 1 after sublingual infection but no other time points [23]. We treated mice with PAD4 inhibitor GSK484 before intraperitoneally infected them with C. albicans. Inhibition of PAD4 reduced Ki67 expression and web-like structure formation in Ly6G + cells in peritoneal exudate and Ki67 + Ly6G + cell population in mesenteric tissues (intracellular Ki67 straining followed by flow cytometric analysis and immunofluorescence staining for fluorescence microscopic imaging). In addition, GSK484 treatment decreased fungal CFU in the peritoneal cavity and increased that in the kidney at 3 h after infection. These data clearly demonstrated that PAD4 is important to C. albicans-induced NETosis in vivo. Regarding the role of PAD4 in host defense, our work with peritonitis infection together with that reported by Guiducci et al. with sublingual infection show that PAD4 functions to restrain fungal spread from the inoculation site to distal site during early phase of C. albicans infection [23]. We speculate that since NETosis-mediated restrain of fungal spread [16,26] does not affect eventual fungal clearance, PAD4 is not required for control of fungal infection.
C-type lectin receptor engagement elicits proinflammatory cytokine response to stimulation by fungal ligand through Syk-mediated PKCδ activation [27]. PKCδ activity modulates CARD9/Malt1/Bcl10 signalosome formation to facilitate downstream NFκB translocation and subsequent cytokine production [27]. Deletion of Prkcd, but not Prkca nor Prkcb genes, abolishes TNF, IL-6 and IL-1β production by dendritic cells upon zymosan, curdlan or C. albicans stimulation [27], indicating the unique role of PKCδ in fungal challenge. We use pharmacological inhibitors for different PKC isoforms and uncover the importance of dectin-2 downstream PKCδ in histone H3 citrullination and NETosis in response to unopsonized C. albicans. Conventional PKCs are known to mediate NADPH oxidase-dependent ROS-mediated NETosis [16]. Our results reveal a PKCδ (novel PKC isoform)-mediated signaling pathway that is involved in unopsonized C. albicans-induced NADPH oxidase-independent NETosis. Our finding also suggests that CARD9/Malt1/Bcl10 signalosome which is downstream of PKCδ may function to mediate NE translocation and PAD4 activation, histone H3 citrullination and trigger NETosis.
C-type lectin dectin-1 has been reported to interfere with C. albicans-induced NET formation in human neutrophils through promoting phagocytosis [20]. NE is normally associated with granule membrane [28]. Upon phagocytosis, it is delivered to phagosome and sequestered within C. albicans-containing phagosome [20]. After which, its access to decondensed chromatin is blocked [20]. Metzler et al. observed in human neutrophils that NE is dissociated from granule membrane via ROS production to gain access to decondensed chromatin [28]. C. calculated based on the intensity at 1 h after infection. n = 10. Data are presented as mean ± SD. ��� , p < 0.005, as analyzed by Student's t test comparing the intensity at each time point to that at 1 h after infection. (C) Three hours after C. albicans infection, peritoneal exudates were collected and seeded on coverslips. Cells on the coverslips were permeabilized and stained for Ki67 (orange), histone H3 (red), Ly6G (green) and nucleus (blue) and viewed under fluorescence microscope. Arrows point to Ly6G + Ki67 + cells. (D) Three hours after C. albicans infection, mesenteric tissues were collected and embedded in O. C.T. Cryosections were stained for Ki67 (red), Ly6G (green) and nucleus (blue) and viewed under confocal microscope. Arrows point to Ly6G + Ki67 + cells.
https://doi.org/10.1371/journal.ppat.1008096.g007   albicans yeast is known to be taken up and cleared by human neutrophils and the neutrophils remain intact [29]. Mouse neutrophils, however, allow engulfed C. albicans to germinate, resulting in membrane rupture and eventual cell death [29]. Our data show that dectin-1 and NADPH oxidase are not involved in NETotic response to unopsonized C. albicans. We also found that inhibiting NE reduces unopsonized C. albicans-induced NET formation, and that NE translocates to the nucleus is independent of NCF-1. We further demonstrated that Syk-Ca 2+ -PKCδ-PAD4 pathway modulates NE nuclear translocation and its access to decondensed chromatin without the involvement of ROS. Thus, it appears that chromatin decondensation and NE translocation as a result of dectin-2 downstream Syk-Ca 2+ -PKCδ-PAD4 signaling pathway are important to NADPH oxidase-independent NETosis.
Urban et al. employed pulmonary and subcutaneous C. albicans infection models to investigate NET formation [30]. In vivo NET formation in infected tissues was identified as structures that stained positive for cell-impermeable DNA dye (pre-injected before animals were killed), myeloperoxidase and histone. NET formation in subcutaneous tissue was observed on 6 days after subcutaneous infection and in the lungs at 24 h after intranasal infection [30]. NADPH oxidase deficiency greatly reduces NET formation in pulmonary A. fumigatus infection [31]. It appears that NADPH oxidase is involved in host NETosis response to pulmonary fungal infection in vivo. Ki67 is recently established as a NETosis marker [13]. Neutrophils undergoing NETosis in the lungs of C. albicans-infected mice as well as that in the brain of fungus-infected humans are Ki67-positive [13]. We enriched neutrophils in the peritoneal cavity of a mouse by injecting casein twice before C. albicans infection. Abundant peritoneal neutrophils allow a robust NET response to C. albicans infection. IVIS imaging showed extracellular DNA release as early as 1.5 h after infection and histone H3 + Ki67 + Ly6G + neutrophils with web-like structures were observed in peritoneal exudates. The % of Ly6G + neutrophils undergoing NETosis was quantified by flow cytometric analysis of intracellular Ki67. This C. albicans peritonitis model where neutrophils could be easily obtained without tissue homogenization and enzyme digestion is useful in quantifying NET and studying their function.
The role of dectin-2 in host defense against C. albicans is well documented. Dendritic cells utilize dectin-2 to recognize C. albicans for IL-6, IL-1β, and IL-23 production [32]. Dectin-2 signaling induces cytokine production through Syk-CARD9-NF-κB pathway [32]. While dectin-2-deficient dendritic cells have greatly reduced ability to produce cytokines in response to fungal α-mannan, dectin-2 deficiency also dampens host cytokine response to C. albicans [32]. Systemic C. albicans infection results in higher mortality in dectin-2-deficient mice than WT mice [32]. Neutrophils are major effector cells that kill C. albicans [5]. Whether dectin-2 plays a role in neutrophil response to C. albicans is an interesting question. It is reported that the role of dectin-2 in neutrophil ROS response to opsonized C. albicans is only marginal [19]. Our results show that dectin-2 recognizes unopsonized C. albicans for NETotic response, and dectin-2 deficient mice have reduced ability to restrain fungal spread from peritoneal cavity to (blue) and viewed under fluorescence microscope. DIC, differential interference contrast image. (C) Peritoneal exudates were stained with anti-CD11b, -Ly6G and -Ki67 antibodies and subject to flow cytometric analysis. The percentages of Ki67 + cells among total neutrophils (CD11b + Ly6G + ) population are shown as % Ki67 + of neutrophils. The mean fluorescence intensity (MFI) of Ki67 represents the level of Ki67 expression in Ki67 + cells. Data were pooled from 2 independent experiments. Gating strategy for Ki67 was illustrated in S4B Fig. Naïve mice are uninfected mice receiving casein injection only. (D) Fungal counts in total peritoneal fluid and kidney homogenates were determined by plating. Fungal colonies were counted 2-3 days later. (Clec4n +/+ n = 13; Clec4n -/n = 9). Data were pooled from four independent experiments. ��� , p < 0.005, as analyzed by Student's t test. (E) Clec4n -/mice were given two injections of 9% casein as described above. Four hours after the second injection, mice were injected with 100 U of MNase or heat-inactivated MNase (h.i. MNase, in otherwise equivalent amount) intraperitoneally at the time when 1×10 8 of C. albicans was administered. At 3 h after infection, peritoneal exudates and kidneys were harvest. Fungal counts in total peritoneal fluid and kidney homogenates were determined by plating. Fungal colonies were counted 2-3 days later. (n = 7-8). Data were pooled from 2 independent experiments. n.s., not significant, as analyzed by Student's t test.
https://doi.org/10.1371/journal.ppat.1008096.g010 kidney. These results suggest that C. albicans injected to the peritoneal cavity remain unopsonized at least for a short period of time to be recognized by dectin-2 and reveal a new role for dectin-2 in neutrophil anti-C. albicans functions.
In summary, this study showed that recognition of unopsonized C. albicans by dectin-2 triggers NET formation through a NADPH oxidase-independent pathway. Signaling pathway leading to NETosis involves Syk-Ca 2+ -PKCδ-NE/PAD4. Dectin-2-mediated NET as revealed in the C. albicans peritonitis model functions to control fungal spread from peritoneal cavity to kidney. Our work provides a better understanding of the molecular mechanism involved in NADPH oxidase-independent NET formation and sheds light on the role of dectin-2 in neutrophil anti-C. albicans function.

Fungus and infection
C. albicans strain SC5314 (ATCC MYA-2876), its isogenic mutant strain HLC54 (yeast-locked strain, efg1/efg1 cph1/cph1), GFP-expressing strain OG1 [33], and dTomato-expressing strain CFA2-dTomato [34] were used in this study. All strains were cultured on yeast-peptone-dextrose (YPD) agar (DIFCO) plate at 30˚C. Mice were injected intraperitoneally with C. albicans yeasts prepared in HBSS buffer. Unopsonized C. albicans was prepared in phenol red free HBSS buffer for experiments. To opsonize, C. albicans yeasts were added to phenol red free HBSS containing 10% fresh mouse serum and let stand at room temperature for 30 min. To induce hyphal formation, C. albicans yeasts were incubated in RPMI 1640 medium at 37˚C for 4 h before use.

Mice
Wild-type (C57BL/6), Itgam -/-, Ncf-1 -/-(originally purchased from the Jackson Laboratories, Bar Harbor, ME, USA), Clec7a -/-(from Dr. Gordon Brown, University of Cape Town, Cape Town, South Africa) [35], Clec4n -/- [19] and MyD88 -/-(from Dr. Tsung-Hsien Chuang, National Health Research Institutes, Taiwan) mice were bred and maintained at the Laboratory Animal Center of National Taiwan University College of Medicine. All mice used in this study were maintained under specific pathogen-free conditions. Mice at 6-12 weeks of age were used in all of the experiments.

Bone marrow neutrophils
Bone marrow cells were harvested from the femurs and suspended in dPBS buffer before overlaid on discontinuous percoll gradients (55%, 62%, and 81% in the order from top to bottom) (GE healthcare). After centrifugation at 1,400 × g for 30 min, cells at the interface between 62% and 81% gradients were harvested and washed. Flow cytometric analysis showed that 90-95% of cells were CD11b + Ly6G + .

NET induction and quantification
Two hundred thousand neutrophils suspended in HBSS were seeded in 96 well-plate before addition of 4 × 10 5 opsonized or unopsonized C. albicans. The plate was centrifuged at 800 × g for 3 min to spin down cells. Wells were treated with 0.5 U of micrococcal nuclease (MNase, NEB) 3 h later and incubated at 37˚C for 10 min to partially digest NET. Supernatants were collected and cell free dsDNA were quantified by Quant-iT PicoGreen dsDNA assay kit (Life technology) according to manufacturer's instruction.

Flow cytometric analysis of Ca 2+ influx in neutrophils
One million neutrophils were suspended in 200 μl of phenol red-free HBSS buffer. Loading dye for intracellular Ca 2+ staining was prepared by adding Calcium Indicator to Signal Enhancer at the ratio of 1:1000 according to the manufacturer's recommendation (BD Biosciences Calcium Assay Kit, 640176). Two hundred μl of loading dye was added to cells and the mixture was incubated at 37˚C for 45 min. After resting in room temperature for 20 min, cells were placed in a 5 ml FACS tube and analyzed by flow cytometry to set the basal level of intracellular Ca 2+ intensity (30 sec). Unopsonized pre-germinated C. albicans prepared in 10 μl of HBSS was then added to the tube for continuous flow cytometric analysis for additional 300-420 sec. Data were analyzed by the kinetic mode of FlowJo software. All procedures including sample acquisition and data analysis followed that of modified Bio-protocol published by S. Lee [36]. Original FACS contour plot for measurement of Ca 2+ intensity is shown in S1 Fig.

Immunofluorescence staining
Five hundred thousand neutrophils mixed with C. albicans at a ratio of 1:2 were plated on coverslips and incubated at 37˚C for 3 h. Coverslips were fixed in 10% formaldehyde for 15 min and permeabilized with 0.5% Triton X-100. After thorough wash, coverslips were blocked (10% FBS in PBS) and stained with anti-neutrophil elastase (abcam, 1:50) or anti-histone H3 antibody (Cell Signaling, 1:100) at 4˚C overnight. Coverslips were stained with cell-permeable DNA dye Hoechst 33258 (1 μg/ml, Invitrogen) or cell-impermeable DNA dye SYTOX Orange (1 μM, Life technology) diluted in blocking buffer and left on ice for 15 min. Coverslips were then mounted by mounting gel and subject to fluorescence microscopic or confocal microscopic analysis.

NET fungicidal activity assay
Twenty thousand neutrophils suspended in HBSS were seeded in 96-well plate and allowed to adhere for 30 min before addition of 4 x 10 5 unopsonized C. albicans yeasts. Wells containing C. albicans yeasts without neutrophils were used as control. Plate was centrifuged at 800 × g for 3 min to spin down yeasts. HBSS containing micrococcal nuclease or heat-inactivated MNase (h.i. MNase) was added to the final concentration of 10 U/ml. Buffer was collected and cold H 2 O (pH = 11) was added 3 h later to lyse cells. C. albicans was detached by mini cell scraper and vigorously pipetting. The number of viable fungi was determined by plating the supernatant on yeast-peptone-dextrose agar plate. Colony counts (CFU) were enumerated 2 days later. The ability of neutrophils to kill C. albicans is presented as % killing of C. albicans which was calculated by dividing the difference of CFU counts between the control group (without neutrophils) and neutrophil-added groups by the counts of the control.

Peritoneal C. albicans infection in casein-peritonitis mice
Mice were injected with 1 ml of 9% casein (9 g of casein sodium salt dissolved in 100 ml of hot PBS) intraperitoneally twice with an interval of 18 h. Four hours after the second casein injection, mice were given 1×10 8 of C. albicans intraperitoneally. For some experiments, HBSS containing 100 U of micrococcal nuclease or heat-inactivated MNase (h.i. MNase) was administered intraperitoneally at the time of infection. At indicated time after infection, peritoneal exudates, mesentery tissues and kidneys were collected and subject to following experiments.

In vivo imaging and quantification of fluorescence
Before imaging, mice were injected with 1 ml of 9% casein intraperitoneally twice with an interval of 18 h. Four hours later, mice were infected by 5 × 10 8 of C. albicans (SC 5314 or CAF2-dTomato) in 200 μl HBSS intraperitoneally. To monitor NET release, mice were given an additional intraperitoneal injection of 100 μl SYTOX Orange (5 μM) at the time of infection. A Xenogen IVIS Imaging System 200 series (PerkinElemer Inc.) was used to quantify fluorescent C. albicans (CAF2-dTomato) or the release of extracellular DNA over time. Photons in user-specified region of interest (ROI, gated area) during 10 sec exposure were measured by Living Image 3.2 software. Relative intensity of total photons in ROI was calculated based on the intensity at 1 h after infection.

Tissue and peritoneal exudate cell immunofluorescence staining
Mesentery were embedded in O.C.T. and allowed to freeze at -80˚C overnight. Five μm thick tissue sections were cut and mounted on gelatin-coated slides. Peritoneal exudates were diluted 1:5 and seeded on coverslips for 1 h. Slides and coverslips were fixed with 4% paraformaldehyde, left on ice for 30 min, permeabilized with 0.5% Triton X-100 and let sit at room temperature for 5 min. After wash, samples were blocked (PBS containing 5% FBS) and stained with anti-histone H3 (Cell Signaling), -Ki67 and -Ly6G (Biolegend) antibodies at 4˚C overnight. Nucleus was stained with Hoechst 33258 (1 μg/ml, Invitrogen) for 15 min. Samples were then mounted by mounting gel and subject to confocal microscopic analysis.

Quantification of fungal load
Kidneys were collected and homogenized in a tissue grinder with 1 ml of RPMI 1640 medium. Peritoneal exudate was collected from mice after intraperitoneal injection of 5 ml of HBSS. Homogenates and peritoneal fluids were treated with 0.5 U of micrococcal nuclease and incubated at 37˚C for 10 min to digest DNA entangled with fungus. One hundred microliter of kidney homogenates and peritoneal exudates were plated on YPD agar. Colonies were counted after incubation at 30˚C for 2-3 days.

Ki67 staining
Peritoneal exudates from naïve and infected mice were collected, treated with 0.5 U/ml MNase and subject to surface staining for neutrophil marker CD11b and Ly6G. After fixation in 4% paraformaldehyde and permeabilization with 1% saponin, cells were stained with anti-Ki67 antibody prepared in staining buffer (0.5% saponin) overnight. Cells were fixed in 1% PFA and subject to flow cytometric analysis.

Statistics
Student t test was used to compare the difference between two groups. Statistical significance was defined as P < 0.05.

Ethics statement
Mouse study was carried out in strict accordance with the recommendations in the Guidebook for the Care and Use of Laboratory Animals, The Third Edition  Ncf-1 +/+ and Ncf-1 -/mice were injected with two doses of 9% casein intraperitoneally. At 4 h after second injection, mice were given C. albicans (1 × 10 8 ) intraperitoneally. At 3 h after infection, peritoneal exudates, mesenteric tissues and kidneys were collected. (A) Peritoneal exudates were seeded on coverslips and incubated for 1 h. Cells were then permeabilized and stained for Ki67 (orange), histone H3 (red), Ly6G (green) and nucleus (blue) and viewed under fluorescence microscope. DIC, differential interference contrast image. Arrows point to Ki67 + cells. (B) Mesenteric tissues were collected and embedded in O.C.T. Cryosections were stained for Ki67 (red), Ly6G (green) and nucleus (blue) and viewed under fluorescence microscope. (C) Fungal counts in total peritoneal fluid and kidney homogenates were determined by plating. Fungal colonies were counted 2-3 days later. ��� , p < 0.005, as analyzed by Student's t test.
Research Center at Academia Sinica, Taiwan) for providing the equipment and technical support for IVIS imaging system.