CARD9 mediates Dectin-1–induced ERK activation by linking Ras-GRF1 to H-Ras for antifungal immunity

C. albicans is coated with mannoproteins that prevent β-glucan moieties from being recognized by Dectin-1 on innate immune cells (Gantner et al., 2005; Wheeler and Fink, 2006). To study the effect of surface β-glucans of live fungi on Dectin-1–mediated innate immune responses, we generated a mutant strain of C. albicans, mnn5, in which the α-1,2-mannosyltransferanse gene MNN5 (Bai et al., 2006) was disrupted by homologous recombination, thereby causing the increased exposure of β-glucans on the cell wall of C. albicans yeasts (Fig. 1 A). Consistently, the β-glucan moiety on the surface of mnn5 yeast was more accessible to soluble GFP-tagged extracellular carbohydrate-binding domain (CRD) of Dectin-1 (sDectin-1^CRD) than that on WT C. albicans yeast cells (Fig. 1 B).

To further confirm that the exposed β-glucan on the surface of mnn5 yeast activates the signaling pathway through Dectin-1, we stimulated BMDMs from WT (Clec7A^+/+) or Dectin-1–deficient (Clec7A^−/−) mice with WT or mnn5 yeast cells, and found that mnn5 yeasts induced more potent NF-κB DNA binding than WT yeast (Fig. 1 C), together with more IκBα degradation (Fig. 1 D). Consistently, Dectin-1 deficiency impaired the NF-κB activation (Fig. 1 C) and IκBα degradation (Fig. 1 D). Moreover, mnn5 yeasts triggered stronger and sustained activation of ERK than WT yeast in BMDMs (Fig. 1 D), but Dectin-1 deficiency significantly impaired yeast-induced ERK activation in BMDMs (Fig. 1 D). In contrast, when these cells were stimulated with hyphal forms of C. albicans, NF-κB activation (Fig. 1 E) and IκBα degradation and ERK activation (Fig. 1 F) were no different between WT and Dectin-1–deficient cells stimulated with WT or mnn5 mutant. These data indicate that mnn5 mutation does not increase the exposure of β-glucan on the surface of hyphae form of C. albicans; thereby Dectin-1 deficiency does not affect mnn5 hyphal form stimulation-induced signaling.

We also found that blockade of Dectin-2 using its specific monoclonal antibodies (Zhu et al., 2013) in BMDMs had almost no effect on NF-κB activation induced by WT and mnn5 yeast stimulation (Fig. 1 G) and the production of proinflammatory cytokine TNF and IL-6 (Fig. 1 H). However, when challenged with WT and mnn5 hyphae, blocking Dectin-2 completely inhibited NF-κB activation (Fig. 1 G) and significantly reduced secretion of TNF and IL-6 (Fig. 1 H). Furthermore, deficiency of TRL2 or TLR4 had no influences on WT and mnn5 yeast stimulation-induced activation of NF-κB (Fig. 1 I) and ERK (Fig. 1, J and K). Together, these results indicate that the exposed β-glucan on yeast surface can induce potent proinflammatory responses through Dectin-1, but not Dectin-2, TLR2, or TLR4.

Previous studies suggest that β-glucans from C. albicans activate Dectin-1–mediated NF-κB activation through a CARD9-dependent pathway (Gross et al., 2006). To further investigate the role of CARD9 in Dectin-1–induced signaling pathways in response to fungal infection, BM-derived macrophages (BMDMs) or BM-derived DCs (BMDCs) from WT (Card9^+/+) or CARD9-deficient (Card9^−/−) mice were stimulated with WT or mnn5 strain of C. albicans yeast. We found that CARD9 deficiency in BMDMs and BMDCs had no influences on the yeast stimulation-induced nuclear translocation of NF-κB (p65 and p50 subunits; Fig. 2, A and B), as well as IκBα phosphorylation and degradation (Fig. 2, C and D, top). In contrast, when these cells were stimulated by C. albicans hyphae that activate Dectin-2 but not Dectin-1 signaling (Saijo et al., 2010), the nuclear translocation of NF-κB subunits was defective in CARD9-deficient BMDMs (Fig. 2 A) and BMDCs (Fig. 2 B). Unlike the yeast stimulation-induced NF-κB activation, we surprisingly found that C. albicans yeast stimulation-induced ERK phosphorylation was defective in CARD9-deficient BMDMs (Fig. 2 C, middle) and BMDCs (Fig. 2 D, middle), suggesting that CARD9 deficiency affects Dectin-1–induced ERK activation. In addition, CARD9 deficiency had no significant impact on Dectin-1–induced activation of p38 and JNK MAP kinases (Fig. 2 E). Together, these results suggest that CARD9 may be dispensable for Dectin-1–induced activation of NF-κB, p38, and JNK, but it is required for Dectin-1–induced ERK activation.

Because CARD9 forms a complex with Bcl10 (Bertin et al., 2000), we examined whether Bcl10 deficiency affected Dectin-1–induced ERK activation using BMDMs from WT or Bcl10-deficient (Bcl10^−/−) mice, and found that Bcl10 deficiency affected neither C. albicans yeast stimulation-induced IκBα phosphorylation and degradation nor ERK activation (Fig. 2 F). This result suggests that although Dectin-1–induced ERK activation is dependent on CARD9, it is independent of the CARD9–Bcl10–Malt1 (CBM) complex (Kingeter and Lin, 2012).

It has been shown that Dectin-1 activates NF-κB through CARD9-dependent pathway (Gross et al., 2006), and curdlan, a mixture of β-glucan from nonpathogenic bacteria, is one of the ligands for Dectin-1 (Palma et al., 2006). To further determine the role of CARD9 in Dectin-1 signaling, we used plate-bound curdlan or plate-bound α-mannans, which potently activated Dectin-1 or Dectin-2, respectively, leading to induction of NF-κB nuclear translocation and ERK phosphorylation (Fig. 3, A and B). Interestingly, we found that curdlan-induced ERK activation was defective in CARD9-deficient BMDMs and BMDCs (Fig. 3, C and D), whereas α-mannan–induced ERK activation was intact (Fig. 3, C and D). Consistent with our previous findings (Bi et al., 2010), curdlan-induced nuclear translocation of NF-κB was intact (Fig. 3, E and F). In contrast, α-mannan–induced nuclear translocation of NF-κB was defected in CARD9-deficient cells (Fig. 3, E and F), which is consistent with our previous studies that CARD9 is required for Dectin-2–induced NF-κB activation (Bi et al., 2010). Together, these results suggest that CARD9 is required for Dectin-1–induced ERK activation but is dispensable for NF-κB activation. In contrast, CARD9 is required for Dectin-2–induced NF-κB activation.

To determine the molecular mechanism by which CARD9 is linked to the signaling pathway leading to activation of ERK, we performed a Yeast Cyto-trap two-hybrid screening. Multiple positive clones, which encode H-Ras (a member of Ras family of small guanosine triphosphatases [GTPases]) and Ras-GRF1 (Ras-guanine-nucleotide-releasing factor 1), were identified in this screening (unpublished data), suggesting that CARD9 may associate with H-Ras and Ras-GRF1, which are upstream signaling components, leading to activation of ERK (Imamura et al., 2009; Wang et al., 2013). To examine whether CARD9 is associated with H-Ras and Ras-GRF1, we immunoprecipitated CARD9 from RAW264.7 cells expressing Flag-tagged CARD9 and found that immunoprecipitation of Flag-CARD9 effectively coprecipitated H-Ras and Ras-GRF1 after the stimulation by WT C. albicans yeast (Fig. 4 A). Consistently, mnn5 yeast stimulation induced a stronger association of CARD9 with both H-Ras and Ras-GRF1 (Fig. 4 A), suggesting that CARD9 inducibly forms a complex with H-Ras and Ras-GRF1 in Dectin-1–induced signaling cascade.

To further confirm the inducible association of CARD9 with H-Ras and Ras-GRF1, endogenous CARD9 was immunoprecipitated from BMDMs after stimulation by C. albicans yeasts. We found that CARD9 inducibly coimmunoprecipitated with both H-Ras and Ras-GRF1 after stimulation (Fig. 4 B), and this association was specific because CARD9 deficiency abolished this association (Fig. 4 B). Together, these results suggest that CARD9 associates with H-Ras and Ras-GRF1 in response to the stimulation of C. albicans yeasts.

Interestingly, CARD9 constitutively associated with H-Ras or Ras-GRF1 when they were coexpressed in HEK293T cells (Fig. 4 C). However, the association of CARD9 with H-Ras was significantly enhanced when three of them were coexpressed in 293T cells (Fig. 4 C). Similarly, although Ras-GRF1 weakly associated with H-Ras, this association was significantly enhanced upon coexpression of them with CARD9 (Fig. 4 D). Consistently, CARD9 deficiency impaired the yeast-induced association of Ras-GRF1 with H-Ras (Fig. 4 E). These results suggest that CARD9 bridges the interaction between Ras-GRF1 and H-Ras, which leads to activation of downstream signaling. Because Ras-GRF1 activates H-Ras (Zhu et al., 2007), we examined H-Ras activation and found that H-Ras could be effectively activated by C. albicans stimulation, whereas CARD9 deficiency impaired this activation (Fig. 4 F). Together, these results indicate that CARD9 mediates H-Ras activation through bridging Ras-GRF1 to H-Ras.

To determine whether the formation of CARD9-bridged Ras-GRF1 and H-Ras complex is dependent on Dectin-1-induced signaling, we examined this complex formation after C. albicans yeast stimulation in WT (Clec7A^+/+) and Dectin-1–deficient (Clec7A^−/−) BMDMs, and found that this complex was completely defective in Dectin-1–deficient cells (Fig. 5 A). Consistently, C. albicans yeast stimulation-induced H-Ras activation was completely defective in Dectin-1–deficient BMDMs (Fig. 5 B). In addition, treatment of BMDMs with Syk inhibitor Piceatannol also blocked this complex formation (Fig. 5 C), and the yeast-induced phosphorylation of Ras-GRF1 as well as ERK activation (Fig. 5 D). In contrast, CARD9 deficiency did not significantly impair the yeast-induced Ras-GRF1 phosphorylation (Fig. 5, E and F), indicating that CARD9 functions downstream of Ras-GRF1. Similarly, CARD9 deficiency significantly impaired curdlan-induced phosphorylation of ERK, but not Ras-GRF1 (Fig. 5 G), indicating that CARD9 functions downstream of Ras-GRF1. Consistently, knockdown of endogenous H-Ras in BMDMs specifically inhibited curdlan- or yeast-induced phosphorylation of ERK (Fig. 5 H) but had no effect on α-mannans– or LPS-induced ERK activation (Fig. 5 I). Together, these results demonstrate that Dectin-1 signaling activated by β-glucans induces Syk-dependent phosphorylation of Ras-GRF1, and then the phosphorylated Ras-GRF1 activates H-Ras through scaffold protein CARD9, leading to the activation of ERK pathway.

To determine the functional effect of CARD9 on Dectin-1–induced cytokine production, we first treated C. albicans yeasts with UV light that fixes C. albicans at the yeast stage and prevents them from transforming into hyphae, and then stimulated WT or CARD9-deficient BMDMs with UV-inactivated C. albicans yeast cells. We found that secretion of cytokines including TNF, IL-6, and IL-1β had a significant reduction, whereas IL-12 production was only slightly affected in CARD9-deficient BMDMs (Fig. 6 A). In contrast, when BMDMs were stimulated with C. albicans hyphae, which induces proinflammatory signaling through the Dectin-2– and Dectin-3–dependent pathways (Saijo et al., 2010; Zhu et al., 2013), CARD9 deficiency was significantly impaired in the expression of these cytokines (Fig. 6 B). Furthermore, we examined the effect of inhibiting ERK signaling on cytokine production by WT or CARD9-deficient BMDMs, which were treated with or without U0126, an inhibitor for ERK activation. We found that U0126 significantly suppressed TNF, IL-6, and IL-1β production by WT BMDMs in response to Dectin-1 agonists C. albicans yeast (Fig. 6 C) and curdlan (Fig. 6 D). However, U0126 treatment did not significantly affect the expression of these cytokines in CARD9-deficient BMDMs after the stimulation by C. albicans yeast (Fig. 6 C) or curdlan (Fig. 6 D), suggesting that the inhibited signaling by U0126 may be largely overlapped with CARD9-induced signaling. Consistent with the effect on CARD9-deficient on TNF and IL-6 production, knockdown of endogenous H-Ras in BMDMs also significantly reduced curdlan- or yeast-induced production of TNF and IL-6 (Fig. 6 E). As controls, U0126 treatment and knockdown of endogenous H-Ras in BMDMs had no effect on α-mannans– or LPS-induced secretions of TNF, IL-6, IL-1β, and IL-12 (Fig. 6, F and G). Together, these results suggest that the activation of CARD9-mediated H-Ras/ERK cascade induced by Dectin-1 agonist is critical for the production of proinflammatory cytokines.

To examine the effect of β-glucan exposed on the surface of C. albicans yeast on host immune responses, we compared the pathogenicity of WT, mnn5 mutant, and MNN5-reconstituted mnn5 strains (mnn5/MNN5) of C. albicans in a murine model of systemic candidiasis. Consistent with a previous report (Bai et al., 2006), mnn5 mutant-infected mice survived significantly longer than CAF2-1 (the parental strain of C. albicans)– or mnn5/MNN5-infected mice with a lethal dose (1 × 10⁶) of C. albicans (Fig. 7 A). The fungal burdens of kidneys in mice infected with mnn5 mutant were significantly lower than those infected with either CAF2-1 or mnn5/MNN5 strains (P < 0.01, Fig. 7 B). However, we found that all CARD9-deficient mice died within 5 d after infection with a sublethal dose (3 × 10⁵) of WT (CAF2-1) or mnn5 mutant strains of C. albicans (P > 0.05, Fig. 7 C) and, consistently, fungal burden in the kidney of CARD9-deficient mice had no significant difference when challenging with CAF2-1 or mnn5 strains (P > 0.05, Fig. 7 D).

To further confirm the effect of CARD9-mediated immune responses on the clearance ability to systemic C. albicans infections, we challenged WT or CARD9-deficient mice with UV-inactivated WT (CAF2-1) or mnn5 mutant C. albicans. Similar to the in vitro challenging of macrophages by UV-inactivated C. albicans (Fig. 6 A), CARD9-deficient mice had significantly lower levels of TNF, IL-6, and IL-1β production in their serum at day 2 after infected with mnn5 mutant, but only TNF production had a significant reduction when infecting CARD9-deficient mice with CAF2-1 strain (Fig. 7 E). To determine the role of Dectin-2–mediated host defense against systemic infections with mnn5 mutant, mice receiving anti–Dectin-2 or nonspecific control IgG were intravenously injected with a sublethal dose (6 × 10⁴) of WT (CAF2-1) or mnn5 mutant strains. Blocking Dectin-2 significantly decreased the survival rate of CAF2-1–infected mice (P < 0.01; Fig. 7 F) but had very little impact on mnn5 mutant-infected mice (Fig. 7 F). The fungal burden in kidney of these mice consistently showed similar phenotypes (Fig. 7 G). Together, these data suggest that mnn5 mutant C. albicans is virulent to CARD9-deficient mice, and that the faster clearance of mnn5 mutant of C. albicans by WT mice may be due to a stronger activation of β-glucan–induced Dectin-1–, but not Dectin-2–, and CARD9-dependent inflammatory responses.

To determine the contribution of CARD9-mediated ERK activation on antifungal immunity in vivo, we pretreated WT or CARD9-deficient mice with or without U0126, and then intravenously infected these mice with a low dose (5 × 10⁴) of mnn5 mutant. We found that administration of U0126 reduced the survival rate of WT mice (P < 0.05; Fig. 8 A). However, administration of U0126 did not have a significant effect on the survival of CARD9-defecient mice (P > 0.05; Fig. 8 A). Consistently, U0126 treatment only significantly affected the fungal load (P < 0.01; Fig. 8 B) and cytokine production (Fig. 8 C) in the kidney of WT but not CARD9-defecient mice after mnn5 infection, suggesting that the contribution of ERK activation on antifungal immunity is largely dependent on CARD9-mediated signaling.

During C. albicans infection, both yeast and hyphae forms can be found in infected tissues, and innate immune cells discriminate them through different PRRs, especially CLRs, which elicit protective immune responses against this infection. Previous studies have shown that CARD9, a scaffold protein, is required to mediate different CLR-induced NF-κB activation after fungal infection. In this study, we found that CARD9 is dispensable for Dectin-1–induced NF-κB activation but it is required for Dectin-1–induced ERK activation with stimulation of C. albicans yeasts, whereas Dectin-2–induced NF-κB activation is completely dependent on CARD9 with stimulation of C. albicans hyphae. Mechanistically, we found that CARD9 links Ras-GRF1 to H-Ras and regulates the activation of H-Ras, leading to activation of downstream ERK activation. Together, we have revealed a previously unknown mechanism by which Dectin-1 induces H-Ras activation that, in turn, induces downstream ERK activation.

In this study, we have also found that CARD9-mediated signaling had more significant contributions to Dectin-2–mediated proinflammatory responses than those mediated by Dectin-1, including induction of IL-1β and IL-6, two pivotal cytokines for priming downstream Th17 differentiation. Consistent with our observations, CARD9 has been shown to be indispensable for development of Th17 responses to C. albicans in mice, mostly through Dectin-2, and, to a lesser extent, Dectin-1 signaling (Robinson et al., 2009). Therefore, the CARD9-mediated different transcriptional activation programs are at the convergence of innate and adaptive antifungal immune responses downstream of Dectin-1 and Dectin-2.

Signaling from Dectin-1 after ligand binding is mediated through the cytoplasmic ITAM-like motif that becomes phosphorylated by Src family kinases, providing a docking site for Syk (Brown, 2006). Therefore, Syk is a pivotal kinase mediating Dectin-1–induced downstream cellular responses, such as cytokine production and induction of the respiratory burst (Brown, 2006). Although the components of the Dectin-1–Syk signaling pathway have yet to be fully elucidated, CARD9, which assembles with BCL10 and MALT1 (Bertin et al., 2000), has been identified as an essential downstream adaptor linking Syk-coupled CLRs to the canonical NF-κB pathway (Gross et al., 2006; Hara et al., 2007). However, in our study, we have found that NF-κB activation induced by Dectin-1 ligands, both curdlan and mutant C. albicans cells with highly exposed β-glucans, are intact in CARD9-deficient cells. These results indicate that CARD9 is dispensable for Dectin-1–induced NF-κB activation. One possible explanation for our observations is that these Dectin-1 ligands may be able to activate NF-κB through both CARD9-dependent and CARD9-independent pathways. In this case, these Dectin-1 ligands may activate another undefined coreceptor leading to activation of NF-κB via CARD9-independent pathway. In contrast, because this undefined coreceptor cannot induce ERK activation, ERK activation by Dectin-1 ligands is only dependent on CARD9-mediated signaling pathway.

It has been shown that Dectin-1 engagement by the cell wall extract zymosan or C. albicans yeasts in DCs triggers Syk and ERK pathway activation (Rogers et al., 2005; Dillon et al., 2006; Slack et al., 2007; Galán-Díez et al., 2010). However, it has not been shown how Dectin-1 signaling pathway leads to ERK activation. Our studies provide a molecular mechanism by which Dectin-1 activates ERK. In this case, stimulation of Dectin-1 with C. albicans yeasts induces Syk-dependent phosphorylation of Ras-GRF1, and the phosphorylated Ras-GRF1, in turn, forms a complex with CARD9 that further recruits H-Ras, leading to activation of H-Ras and subsequent activation of downstream ERK pathway.

It has been well documented that the small G protein H-Ras is a key regulator of many signal transduction pathways (Rebollo and Martínez-A, 1999). In its active form, H-Ras can activate a variety of downstream targets, including the ERK cascade (Takai et al., 2001). H-Ras is activated by GEFs (guanine nucleotide exchange factors) that promote the release of guanosine diphosphate (GDP) bound to H-Ras and exchange for the activating GTP (Shou et al., 1992). Ras-GRF1 catalyzes the GDP/GTP exchange needed for H-Ras activation (Quilliam et al., 2002). Our findings indicate that CARD9 mediates H-Ras activation through linking Ras-GRF1 to H-Ras in response to C. albicans yeast stimulation.

It is interesting to note that CARD9 regulates the activation of another small G protein Rac1 through dissociating LyGDI (lymphoid-specific GDP dissociation inhibitor) from Rac1 after Listeria monocytogenes stimulation (Wu et al., 2009). The activation of Rac1 induces reactive oxygen species (ROS) production, which contributes to the killing of intracellular bacteria L. monocytogenes. Therefore, CARD9-deficient macrophages are defective in ROS production in response to L. monocytogenes infection and are impaired in intracellular bacterial killing (Wu et al., 2009). Consistently, CARD9-deficient mice exhibit hyper-susceptibility to infection of L. monocytogenes (Hara et al., 2007; Hsu et al., 2007). Future studies should also determine whether CARD9 regulates Rac1 GEFs besides Rac1 GDI.

Up to now, there have only been limited studies about the roles of ERK cascade on host defense against fungal infections. It has been shown that the ERK cascade plays a minor role in modulating the killing of C. albicans by neutrophil (Hii et al., 1999) but plays a more important role for promoting the killing of C. albicans by macrophages (Ibata-Ombetta et al., 2001). However, inhibition of the ERK signaling pathway abolished neutrophil migration induced by C. albicans hyphae and selectively impaired their ability to kill C. albicans (Wozniok et al., 2008). Another study shows that activation of the ERK cascade by C. albicans hyphae, but not yeasts, is necessary to induce a cytokine response by oral epithelial cells (Moyes et al., 2010). In this study, our data shows that β-glucan exposed on the surface of C. albicans yeast can induce strong activation of both NF-κB and ERK pathways through Dectin-1 signaling. However, Dectin-1–induced ERK activation is completely dependent on the adaptor protein CARD9. Furthermore, our in vivo data shows that treatment with an ERK inhibitor significantly accelerates the death of C. albicans–infected mice in a murine systemic candidiasis model, and this inhibitory effect is largely dependent on the presence of CARD9. Therefore, we conclude that Dectin-1–mediated antifungal immune responses to C. albicans infection are mainly through CARD9-mediated ERK activation both in vitro and in vivo.

Dectin-1 genes are highly conserved across mammals, including human and mouse, indicating that the roles of Dectin-1 in innate immunity are conserved among different species. The impact of Dectin-1 deficiency on antifungal immunity has been shown in Dectin-1–deficient mice (Taylor et al., 2007). Moreover, the importance of Dectin-1 in human antifungal immunity has shown that human patients with Dectin-1 Y238X mutation, in which the protein translation of Dectin-1 gene is prematurely terminated at Tyrosine 238, are highly susceptible to fungal infection (Ferwerda et al., 2009; Cunha et al., 2010; Chai et al., 2011). Homozygotes of this mutation are associated with mucocutaneous infections, including vaginal candidiasis and onychomycosis, even though fungal killing and phagocytosis occur normally (Plantinga et al., 2009; van der Velden et al., 2010). The higher susceptibility of patients with Dectin-1 mutations is likely due to the defect of antifungal innate immune responses induced by Dectin-1–mediated NF-κB and MAPK activation.

Previous studies indicate that CARD9 is required for NF-κB activation induced by several CLRs including Dectin-1 and Dectin-2 in response to fungal infection (Gross et al., 2006; Robinson et al., 2009; Bi et al., 2010; Saijo et al., 2010). Our findings that CARD9 also regulates H-Ras activation and mediates Dectin-1–induced ERK activation further indicate the importance of CARD9 in antifungal immunity. Consistent with the critical role of CARD9 in NF-κB activation induced by multiple CLRs and ERK activation by Dectin-1, patients with CARD9 mutations (Glocker et al., 2009; Lanternier et al., 2013; Wang et al., 2014) are susceptible to the infection by a broad range of different fungal species and appear to have more severe fungal infection than those found in patients with Dectin-1 mutations (Ferwerda et al., 2009; Cunha et al., 2010; Chai et al., 2011). Considering the emerging information of the complexity of CLR signaling, further investigating CARD9-associated signaling cascades may provide novel therapeutic insight for designing antifungal agents.

Human Dectin-1 and Dectin-2 were amplified by PCR using full-length cDNA of human peripheral blood cells as a template. All PCR amplifying fragments, including the Flag-encoding DNA sequence, were inserted into a lentivirus vector, pRV3, between the SalI and BglII sites. To obtain GFP/Dectin-1^CRD fusion gene, the method of overlapping PCR was used as described previously (Guan et al., 2007). Four primers for PCR were designed as follows based on the GFP and the Dectin-1^CRD encoding sequence: GFP (sense), 5′-CAGGACATATGGTGAGCAAGGGCGAG-3′; GFP (anti-sense), 5′-GCTGCCACCT­CCACCGCTACC­GCCGCCTCCCT­TGTACAGCTC-3′; Dectin-1^CRD (sense), 5′-GGTGGAGGTG­GCAGCTTTTGG­CGACACAAT-3′; and Dectin-1^CRD (anti-sense), 5′-ACTGAGGGATCCCAGTTCCTTCTCA-3′. The GFP (anti-sense) and Dectin-1^CRD (sense) primers contained a (Gly₄Ser)₂ linker encoding sequence (boxed). The fusion PCR product was digested by NdeI and BamHI and was inserted in frame into pET28a expression vector (EMD Millipore).

Antibodies against phospho-IKKα/β, phospho-IκBα, phospho-ERK, phospho-p38, and p38 were purchased from Cell Signaling Technology; antibodies against p65, IKKα, IκBα, ERK, H-Ras, Ras-GRF1, α-actin, and PCNA were from Santa Cruz Biotechnology. Curdlan and α-mannans were from Wako and Sigma-Aldrich, respectively.

To construct mnn5 null mutant, the fragment containing 5′ and 3′ ends of MNN5 gene for homologous recombination and the 4-kb hisG-URA3-hisG fragment from the plasmid p5921 were subsequently cloned into plasmid pUCm-t (Sangon). The XhoI digested fragment of pUC-MNN5-URA3 was transformed into the ura3Δ/Δ mutant (CAI4) by standard methods (Sanglard et al., 1996). To construct mnn5 mutant revertant strain, MNN5 containing their ORFs and 1 kb up/downstream were amplified by PCR with Pyrobest polymerase (Takara Bio Inc.) and cloned into plasmid pBes116 to obtain plasmid pBes-MNN5. Then, the AscI digested fragment of pBes-MNN5 was transformed into mnn5 null mutant as previously (Sanglard et al., 1996).

C. albicans yeast cells were grown overnight in YPD rich medium at 30°C. For hyphae, yeast cells were counted and then resuspended in RPMI 1640 to grow at 37°C for 3 h.

Inactivation of C. albicans strains was performed as described previously (Wheeler and Fink, 2006). For UV inactivation of C. albicans, the equivalent of 2.5 × 10⁷ cells from a culture were washed and resuspended in 1 ml of PBS in a 6-well plate. The fungi were exposed to four doses of 100,000 µjoules/cm² in a CL-1000 UV-cross-linker (UVP; Upland), with agitation between each dose to treat cells evenly.

C. albicans yeast cells or hyphae were stained with anti-β-1,3-glucan primary antibody (Biosupplies Inc.) and FITC-labeled goat anti–mouse secondary. Stained cells were adhered to clear-bottom plates with Concanavalin A (Sigma-Aldrich) and scanned under an oil immersion objective at 100× magnification on a confocal microscope (TCS SP5; Leica).

Dectin-1^−/− mice were provided by G. Brown (University of Aberdeen, Aberdeen, Scotland). Primary cultures of BMDMs from WT or Decitn-1^−/−, CARD9^−/−, or Bcl10^−/−mice were prepared as previously described (Bi et al., 2010; Gorjestani et al., 2011). In brief, BM cells were harvested from the femurs and tibias of mice. Erythrocytes were removed from cells samples by subjecting the samples to hypotonic solution. Cells were cultured for 7 d in DMEM containing 20% FBS, 55 µM β-mercaptoethanol, 100 µg/ml streptomycin, 100 U/ml penicillin, and 30% conditioned media from L929 cells expressing macrophage CSF (M-CSF). Non-adherent cells were removed and cells were passed every 3 d. After 1 wk of cultures, flow cytometry analysis indicated that the harvested cell population contained 86–95% CD11b⁺ F4/80⁺cells as assessed.

BMDCs were prepared from mouse femurs and tibias by culture of BM cells for 6 d in complete RPMI 1640 (10% FBS, 100 U/ml penicillin, 100 µg/ml streptomycin, and 55 µM β-mercaptoethanol) supplemented with 25 ng/ml GM-CSF (PeproTech) and 1 ng/ml IL-4 (PeproTech). After 6 d of cultures, BMDCs were purified by magnetic selection with anti-CD11c microbeads. This routinely gave purities of >98%.

Human CARD9 coding sequence was inserted into pSOS vector and used to screen its binding partners by using CytoTrap XR Mouse Spleen cDNA Library according to the manufacture protocol (Agilent Technologies).

BMDMs were plated in a 6-well plate (1 × 10⁶ cells/well) in 1 ml DMEM containing 10% FBS. Cells at 50–80% knockdown efficiency were transfected with different concentrations of siRNA against murine H-Ras. Trans IT-TKO transfection reagent was used to transfect siRNA and nontargeting control siRNA according to the manufacturer’s protocol (Mirus). Cells were lysed 48 h after transfection, and relative intensities of H-Ras protein were determined by Western blotting.

BMDMs were serum-starved overnight, stimulated, and lysed in lysis buffer (150 mM NaCl, 50 mM Hepes, pH 7.4, 1 mM EDTA, 1% Nonidet P-40, and protease inhibitors). The cell lysates were immunoprecipitated with the indicated antibody-conjugated agarose. The resulting immunoprecipitates and lysates were subjected to SDS-PAGE and then blotted using the indicated antibodies.

After BMDMs were stimulated with yeast cells or hyphae, their nuclear extracts were prepared. 5 µg of the resulting nuclear protein was incubated with ³²P-labeled NF-κB or Oct-1 probe (Promega) for 15 min at 25°C, and then subjected to PAGE and exposed to x-ray films.

The enzyme-linked immunosorbent assay kit for TNF, IL-6, IL-12p40, and IL-1β ELISA Ready-SET-GO kits were purchased from eBioscience. All of the samples were measured in triplicate times according to the manufacturer’s protocol.

Ras activation was determined by pull-down assay using a Ras activation assay Biochem kit (Cytoskeleton, Inc.) according to the manufacturer’s instructions. In brief, left ventricular myocardium sample (∼100 mg) was homogenized in lysis buffer. Equal amounts of protein (∼200 µg) from left ventricular samples were incubated at 4°C for 1 h with Raf-RBD (Ras binding domain) beads (40 µl), which specifically recognize the GTP-bound form of Ras. Immunoassay of the total and activated Ras was performed for determining the Ras activation.

For the in vivo C. albicans infection, a group of C57B/L6 and Card9^−/− female mice were injected via lateral tail vein with 200 µl of a suspension containing a sublethal dose (3 × 10⁵) or a low dose (5 × 10⁴) of CAF2-1 or mnn5 strain of C. albicans yeast cells in sterile saline. Kaplan–Meier and Life Table analyses were used to estimate survival probabilities. All animal experiments were performed in compliance with the institutional guidelines and according to the protocol approved by the Institutional Animal Care and Use Committee of MD Anderson Cancer Center and the Animal Ethics Committee of Tongji University. For cytokine measurement, sera from infected mice or cultured media from DMEMs after fungal infections were used and detected by ELISA kits (eBioscience).

All the animal experiments were performed in compliance with the institutional guidelines and according to the protocol approved by the Institutional Animal Use and Care Committee of the University of Pittsburg, the Institutional Animal Care and Use Committee of MD Anderson Cancer Center, and the Animal Ethics Committee of Tongji University.

At least two biological replicates were performed for all experiments unless otherwise indicated. Log-rank testing was used to evaluate the equality of survival curves. Student’s t test for paired observations was used for statistical analyses of cytokine expression levels. Statistical significance was set at a p-value of <0.05, <0.01, or <0.001.

We would like to thank Dr. G. Brown for providing Dectin-1 KO mice.

This work was supported by Shanghai Basic Research Program (12JC1408401; X.-M. Jia), National Basic Research Program of China (Program 973, 2013CB967500; G.-T. Xu), and National Institutes of Health grant R01AI050848 (X. Lin), and X. Lin is a Changjiang Lecture Professor of the Ministry of Education, China.

The authors declare no competing financial interests.

Author contributions: X.-M. Jia, B. Tang, L.-L. Z.hu, Y.-H. Liu, X.-Q. Zhao, S. Gorjestani, Y.-M.S. Hsu, L. Yang, and J.-H. Guan performed experiments. G.-T. Xu, X.-M. Jia, and X. Lin directed the project and wrote the paper.