SHIP1 modulates antimalarial immunity by bridging the crosstalk between type I IFN signaling and autophagy

ABSTRACT Stringent control of the type I interferon (IFN-I) signaling is critical for host immune defense against infectious diseases, yet the molecular mechanisms that regulate this pathway remain elusive. Here, we show that Src homology 2 containing inositol phosphatase 1 (SHIP1) suppresses IFN-I signaling by promoting IRF3 degradation during malaria infection. Genetic ablation of Ship1 in mice leads to high levels of IFN-I and confers resistance to Plasmodium yoelii nigeriensis (P.y.) N67 infection. Mechanistically, SHIP1 promotes the selective autophagic degradation of IRF3 by enhancing K63-linked ubiquitination of IRF3 at lysine 313, which serves as a recognition signal for NDP52-mediated selective autophagic degradation. In addition, SHIP1 is downregulated by IFN-I-induced miR-155-5p upon P.y. N67 infection and severs as a feedback loop of the signaling crosstalk. This study reveals a regulatory mechanism between IFN-I signaling and autophagy, and verifies SHIP1 can be a potential target for therapeutic intervention against malaria and other infectious diseases. IMPORTANCE Malaria remains a serious disease affecting millions of people worldwide. Malaria parasite infection triggers tightly controlled type I interferon (IFN-I) signaling that plays a critical role in host innate immunity; however, the molecular mechanisms underlying the immune responses are still elusive. Here, we discover a host gene [Src homology 2-containing inositol phosphatase 1 (SHIP1)] that can regulate IFN-I signaling by modulating NDP52-mediated selective autophagic degradation of IRF3 and significantly affect parasitemia and resistance of Plasmodium-infected mice. This study identifies SHIP1 as a potential target for immunotherapies in malaria and highlights the crosstalk between IFN-I signaling and autophagy in preventing related infectious diseases. SHIP1 functions as a negative regulator during malaria infection by targeting IRF3 for autophagic degradation.

M alaria parasite infection activates complex and delicate immune responses, including the innate immunity which severs as the first line of host defense against invading Plasmodium (1,2). Parasite DNA, RNA, protein-DNA complex, glycosyl phosphatidylinositol, and hemozoin are recognized by the pattern recognition receptors of host cells, such as Toll-like receptors (TLRs), Retinoic-acid-inducible gene I (RIG-I)-like receptors, NOD-like receptors (NLRs), and nucleic acid sensors, to activate nuclear factor kappa B (NF-κB), IRF-3/7, and inflammasome signaling to produce type I interferon (IFN-I) and pro-inflammatory cytokines (3)(4)(5). IFN-I signaling plays an important role in inhibiting the blood stages of multiple Plasmodium strains at early infection (5), and its production must be tightly regulated to maintain immune balance. Recent studies have unveiled the complexity of regulation in IFN-I responses by various regulators during Plasmodium infections, including suppressor of cytokine signaling 1 (6), FOS-like antigen-1 (7), membrane-associated ring-CH-type finger 1 (8), and receptor transporter protein 4 (RTP4) (5). However, the complex regulatory network and the dynamics of the IFN-I pathway in response to malaria parasite infections remain unresolved (9).
SHIP1 is a member of the inositol polyphosphate-5-phosphatase (inpp5) family. SHIP1 has an N-terminal SH2 domain, an inositol phosphatase domain, and two C-terminal protein interaction domains (10) and is an important regulator of immune cell activation (11). SHIP1 dampens B cell receptor signaling through its receptor-dependent function and functions as an intrinsic brake on activation signaling (12). Moreover, substantial evidence indicates that SHIP1 controls NF-κB, mitogen-activated protein kinase, and phosphatidylinositol 3 kinases (PI3K)/serine/threonine kinase signaling via phosphatase activity. Besides, SHIP1 is also critical for hematopoietic cell development through its nonenzymatic functions (13). Given the importance of SHIP1 in regulating immune cell function and PI3K signaling, therapeutic targeting of SHIP1 has been tested in treatments of leukemia, lymphoma, and inflammatory and autoimmune disease (14). The transcrip tion level of inositol polyphosphate 5-phosphatase D (Inpp5d or Ship1) has been reported to be downregulated in Plasmodium falciparum (P.f.)-infected patients (15). This suggests that it may be involved in malaria infection. However, it is unknown how SHIP1 functions in antimalarial immunity.
Autophagy is a highly conserved eukaryotic degradation process involving the cellular recycling of multiple cytoplasmic components during physiological conditions as well as in response to various stresses (16). Accumulated evidence emphasizes that autophagy is highly selective that delivers substrates such as protein aggregates and damaged or superfluous organelles to autophagosomes for lysosome degradation via several cargo receptors (17). Indeed, cargo receptors such as sequestosome 1 (SQSTM1/ p62), CALCOCO2/NDP52, optineurin (OPTN), and neighbor of BRCA1 (NBR1), harbor both ubiquitin-binding domains and LC3-interacting regions, and direct ubiquitin-linked substrates for autophagic degradation (17,18). Recent reports support that autophagy modulates immune responses in infectious diseases (19), but the molecular mechanism underlying the multilayer regulation of selective autophagy in antimalarial immune responses remains largely elusive (20).
In this study, we report that SHIP1 is a key negative regulator of IFN-I signaling by promoting autophagic degradation of IRF3 during malaria infection. After malaria parasite Plasmodium yoelii nigeriensis N67 (P.y. N67) infection, Ship1-deficient mice produce higher levels of IFN-I and have significantly lower parasitemia than wild-type (WT) mice. The association between IRF3 and cargo protein NDP52 is enhanced by SHIP1 that promotes the K63-linked ubiquitination of IRF3 at K313 and severs as a signal for NDP52-dependent selective autophagic degradation. Moreover, SHIP1 is downregu lated by IFN-I-induced miR-155-5p after malaria parasite infection and functions as a feedback loop between IFN-I signaling and autophagy. This study reveals an important role of SHIP1 in mediating the crosstalk between IFN-I response and autophagy during antimalarial immune responses.

SHIP1 impairs host antimalarial immune responses
To study the potential role of SHIP1 in host antimalarial immune response, we genetically ablated the Ship1 gene in WT bone marrow (BM) cells using the CRISPR/Cas9 system and generated Ship1-chimeric mice to investigate the function of SHIP1 during malaria infection (Fig. 1A). We injected intraperitoneally (i.p.) 2 × 10 5 of N67 infected red blood cells (iRBCs) into WT and Ship1-chimeric mice and evaluated SHIP1 expression in the BM and the spleen. Significantly reduced SHIP1 expressions were observed in the BM and spleen of the infected Ship1-chimeric mice, respectively ( Fig. 1B and C). Parasitemia in WT mice infected with N67 rose rapidly for several days then reduced to low levels at day 5, before increasing again (3). The infected Ship1-chimeric mice showed significantly reduced parasitemia after N67 infection (Fig. 1D), and survived significantly longer than N67-infected WT mice which typically died on days 25-30 (3) (Fig. 1E), indicating that SHIP1 is detrimental for host antimalarial immunity.
IFN-I has been reported to be critical for stimulating stronger immune responses against N67 infection (3,5,21,22), thus, we detected whether SHIP1 deficiency affects the production of IFN-I and found that N67C-infected Ship1-chimeric mice produced significantly higher levels of IFN-α and IFN-β in the serum than the infected WT mice ( Fig.  1F and G). Meanwhile, the mRNA extracted from BMs of mice on days 1 and 6 post-N67 infection showed higher Ifn-a/b expression in Ship1-chimeric mice than the WT mice ( Fig.  S1A and B). Considering the negative role of Ship1 in the NF-κB pathway (23), we also measured the production of pro-inflammatory cytokines such as TNF-α and IL-6, but detected no significant difference between Ship1-chimeric mice and WT mice after N67 infection ( Fig. S1C and D). To investigate the effect of Ship1 in other murine malaria models, we also generated Ship1-chimeric mice and infected them with Plasmodium berghei (P.b.) ANKA. The infected Ship1-chimeric mice showed reduced parasitemia and prolonged survival after ANKA infection (Fig. 1H). Also, we detected higher levels of IFN-β in the serum (Fig. 1I) and higher Ifn-a/b expression in the spleen ( Fig. S1E and F) of Ship1chimeric mice than the WT mice. We also determined whether Ship1 mediates antimalar ial immunity in non-virulent malaria, and observed lower parasitemia in Ship1-chimeric mice after the non-lethal strain P.y. 17XNL infection (Fig. S1G). These results suggest that SHIP1 inhibits IFN-I signaling during lethal or non-lethal malaria infection and genetic deletion of Ship1 leads to a stronger IFN-I response and antimalarial immunity during malaria infection.
To investigate whether the resistance of Ship1-chimeric mice to N67 infection is dependent on the increased production of IFN-I, we infected WT and Ifnar −/− mice with N67 after the tail vein injection (IOCV) of scramble small interfering RNA (siRNA) and Ship1 siRNA, respectively; then, we detected lower parasitemia and longer survived in WT mice which handled with Ship1 siRNA (Fig. 1J and K), and higher production of IFNα/β in the serum and higher Ifn-a/b expression in the spleen than the infected WT mice which handled with scramble siRNA (Fig. 1L and M). Moreover, higher parasitemia and shorter survival time were observed in Ifnar −/− mice after N67 infection, and the inhibition of Ship1 on host antimalarial immunity was abolished in Ifnar −/− mice ( Fig. 1J and K). These results suggested that SHIP1-mediated inhibition on host antimalarial immunity is IFN-I-IFNAR dependent.

SHIP1 negatively regulates malaria-induced IFN-I signaling
Malaria infection results in host immune dysfunction, and a strong IFN-I response triggered by RNA polymerase III and melanoma differentiation-associated protein 5 (MDA5), binding of parasite DNA/RNA contributed to a decline of parasitemia in N67infected mice (3,24). However, the mechanism of IFN-I regulation during N67 infection remains unclear. To further investigate the role of SHIP1 in regulating malaria infectioninduced IFN-I signaling, we purified genomic DNA (gDNA) (21,25) and RNA from N67 parasites and performed luciferase reporter assay to measure IFN-I response in 293T cells after parasite gDNA or RNA stimulation. Ectopic expression of SHIP1 substantially inhibited ISRE-luc and IFN-β-luc activities triggered by P.y. gDNA or RNA ( Fig. 2A to D).
We next genetically disrupted the Ship1 gene in RAW 264.7 cells using CRISPR/Cas9 gene-editing technology and stimulated WT and SHIP1 knockout (KO) cells with P.y. gDNA or RNA for the indicated times. After 6 or 9 h stimulation with P.y. gDNA or RNA, significantly higher mRNA transcripts for Ifnb and ISGs, such as Ifit3, Ifit2, and Isg56, were detected in the SHIP1 KO cells than in WT cells ( Fig. 2E and G; Fig. S2A and B). In addition, phosphorylated IRF3, which indicates activation of the IFN-I pathway, was more abun dant in the SHIP1 KO cells after P.y. gDNA or RNA stimulation ( Fig. 2F and H). Consistently, ELISA analysis revealed that SHIP1 deficiency promoted the production of IFN-β cytokine in the supernatants after P.y. gDNA or RNA stimulation( Fig. S2C and D). Taken together, these results demonstrate that SHIP1 deficiency significantly increases the expression of IFN-β and ISGs following malaria parasite gDNA/RNA stimulations in vitro, supporting that SHIP1 is a negative regulator of IFN-I signaling.
We next generated a Flag-tagged SHIP1 doxycycline (doxy)-inducible RAW 264.7 cells to examine the effects of SHIP1 overexpression on IFN-I response and found that doxymediated expression of SHIP1 resulted in lower mRNA levels of Ifnb, Ifit3, Ifit2, and Isg56 after P.y. gDNA or RNA stimulation ( Fig. 2I and J; Fig. S2E and F). Consistent with the results of lower mRNA levels, doxy-induced SHIP1 potently blocked the phosphorylation of endogenous IRF3 after P.y. gDNA or RNA stimulation ( Fig. 2K and L). In addition, IFN-β protein was also decreased in the supernatants of SHIP1 doxy-induced cell culture compared with those of WT cells after P.y. gDNA or RNA stimulations( Fig. S2G and H). Collectively, these results support that SHIP1 negatively regulates malaria-induced IFN-I signaling.
We previously showed that MyD88-mediated IFN-I signaling in plasmacytoid dendritic cells (pDCs) played an important role in P.y. YM model (21). To verify whether Ship1 inhibited the MyD88-mediated IFN in pDCs after N67 infection, we cultured pDCs by using FMS-like tyrosine kinase 3 ligand (FLT3L) to stimulate murine BM cells, and stimulated pDCs with parasite N67 gDNA for the indicated times after the handle of siRNA to knockdown the expression of SHIP1 (Fig. S2I). We observed the production of IFN-β in the supernatants (Fig. S2J) and the expression of Ifn-a/b ( Fig. S2K and L), and found that there is no difference between scramble siRNA and Ship1 siRNA treated cells. These results suggested that SHIP1 did not affect the MyD88-mediated IFN in pDCs after N67 gDNA stimulation.

SHIP1 interacts with the IRF association domain of IRF3
To investigate the molecular target of SHIP1 in IFN-I signaling pathways, we performed a luciferase reporter assay (8) and found that SHIP1 decreased the luciferase reporter activity induced by cyclic GMP-AMP synthase (cGAS)/stimulator of interferon genes (STING), RIG-I, MDA5, mitochondrial antiviral-signaling protein (MAVS), TANK-binding kinase 1 (TBK1), or IRF3-5D (a constitutively active mutant of IRF3) ( Fig. 3A and B). Since SHIP1 can inhibit the IFN-I response by reducing IRF3 phosphorylation, we investigated whether SHIP1 could directly interact with IRF3 and other signaling molecules in the IFN-I pathway. Co-immunoprecipitation (co-IP) and immunoblot analysis showed that HAtagged SHIP1 interacted with IRF3 and IRF3-5D ( Fig. 3C; Fig. S3A), but not with cGAS, STING, RIG-I, MDA5, MAVS, or TBK1 (Fig. S3A). To investigate the interaction between SHIP1 and IRF3 under physiological conditions, we stimulated RAW 264.7 cells, bone marrow-derived macrophages (BMDMs), and THP-1 cells with P.y. gDNA and collected the cell lysates at indicated times. We found that the endogenous association between SHIP1 and IRF3 increased upon P.y. gDNA stimulation ( Fig. 3D; Fig. S3B and C). Further more, we investigated the co-localization of SHIP1 and IRF3 using confocal microscopy and showed slight co-localization between SHIP1 and IRF3 in BMDMs, and stronger colocalization after stimulation of parasite gDNA (Fig. 3E). These data indicate direct interaction between SHIP1 and IRF3.
IRF3 comprises the DNA binding domain (DBD), proline-rich region (PRO), IRF association domain (IAD), and reaction domain (RD) (17). To map the essential domains of IRF3 that interact with SHIP1, we generated several deletion constructs of IRF3 (Fig. 3F)  Research Article mBio and found FL (full length), IR (IAD+RD), DPI (DBD+PRO+IAD), ΔC (DBD+PRO+IAD), and ΔDBD (PRO+IAD+RD) of IRF3 interacted with the full-length SHIP1 protein, while IRF3 DP (DBD+PRO) and DBD did not associated with SHIP1 ( Fig. 3H), indicating that the IAD domain of IRF3 is important for the IRF3-SHIP1 interaction. To further determine which domain of SHIP1 is responsible for its interaction with IRF3, we also generated several deletion constructs of SHIP1 (Fig. 3G). Co-IP and immunoblot analyses revealed that only the phosphatase domain of SHIP1 interacted with IRF3 (Fig. 3I). Consistently, the phosphatase domain of SHIP1 could interact with the IR (IAD+RD) domain of IRF3 (Fig.  3I). We also observed that the phosphatase domain of SHIP1 maintained the SHIP1mediated inhibition of ISRE activity (Fig. S3D), indicating that the phosphatase domain of SHIP1 is critical for the suppression of IFN-I signaling. Taken together, these results suggest that SHIP1 inhibits IFN-I signaling by interacting with IRF3 through the phospha tase domain and IAD domains, respectively.

SHIP1 promotes autophagic degradation of IRF3
SHIP1 is known as a phosphatase that could decrease the protein phosphorylation level (12), and our results indicated that the phosphatase domain of SHIP1 is essential for its association with IRF3; thus, we hypothesized that the phosphatase activity of SHIP1 is required for its inhibition on IFN-I pathway. To test this possibility, we generated a construct encoding catalytically inactive SHIP1 with T1936C substitution, a mutant reported in 30% acute myeloid leukemia (AML) patients (26). 293T cells were transfected with plasmids encoding Flag-IRF3 together with increasing doses of WT SHIP1 or the inactive SHIP1 form (T1936C). Interestingly, the levels of IRF3 and p-IRF3 were decreased and the IRF3 protein levels were negatively correlated with that of SHIP1 protein no matter the cells were co-transfected with WT SHIP1 or the T1936C mutant (Fig. 4A). Moreover, the mRNA level of Irf3 remained unchanged with increasing SHIP1, suggesting that SHIP1 promotes IRF3 protein degradation (Fig. 4B). Consistently, the phosphatase domain of SHIP1 alone could cause the degradation of IRF3 (Fig. S4A), suggesting SHIP1 mediated the activity of IRF3 in a catalytic activity-independent manner, but through its phosphatase domain. To exclude other proteins in the IFN-I pathway, we performed similar experiments and found that SHIP1 specifically degraded IRF3, but not RIG-I, MDA5, cGAS, STING, MAVS, or TBK1 ( Fig. S4B to F). Next, we ectopically expressed an increasing amount of Flag-SHIP1 with doxy induction and found that endogenous IRF3 was downregulated upon increasing expression of SHIP1 (Fig. 4C). Consistently, we observed that IRF3 turnover rates were reduced in SHIP1 KO cells in the presence of translational inhibitor cycloheximide (CHX) (27) (Fig. 4D). To assess whether SHIP1-mediated IRF3 degradation could be potentiated by malaria parasite DNA, we compared the degradation rates of IRF3 between WT and SHIP1 KO cells upon P.y. gDNA stimulation and found that KO of SHIP1 in RAW 264.7 cells decelerated endogenous IRF3 degradation elicited by P.y. gDNA stimulation (Fig. 4E). Conversely, doxy-induced expression of SHIP1 further accelerated the downregulation of endogenous IRF3 after P.y. gDNA stimulation (Fig. 4F). Furthermore, we observed much more protein amounts and phosphorylation of IRF3 in the BM of Ship1-chimeric mice than in WT mice (Fig. 4G). These results indicate that SHIP1 mediates IRF3 degradation upon Plasmodium DNA stimulation or infection.
The proteasome, lysosome, and autolysosome pathways are three major systems that eukaryotic cells used to control protein degradation (17). We next investigated which degradation system mediated the degradation of IRF3 by SHIP1 and found the degrada tion of IRF3 could be blocked by the autolysosome inhibitor chloroquine (CQ) (28) and autophagy inhibitor 3-methyladenine (3-MA) (17) but not by the proteasome inhibitor MG132 (Fig. 4H), suggesting that SHIP1 promoted autophagic degradation of IRF3. As previously reported (17), we also observed that IRF3 turnover rates were reduced in autophagy-related gene 5 (ATG5) and Beclin-1 (BECN1) KO cells (17) in the presence of CHX, in which the autophagy is significantly impaired (Fig. S4G and H). In addition, the degradation of IRF3 triggered by SHIP1 was almost abolished in ATG5 and BECN1 KO cells  Fig. S4I and J). Collectively, these results suggest that SHIP1 mediates IRF3 degradation through autophagy. Given that IRF3 forms dimers after phosphorylation and translocates to the nucleus to activate downstream signaling (29), we next investigated which form (active or inactive) of IRF3 is regulated by SHIP1. We co-transfected 293T cells with Flag-IRF3-5D (a constitutively active mutant of IRF3) (30) or Flag-IRF3-5A (the substitution of these five S/T residues with A greatly impairs the function of IRF3) and HA-SHIP1 and observed that SHIP1 could promote the degradation of WT IRF3 and IRF3-5D, but not IRF3-5A (Fig.  4K), suggesting that SHIP1 mediated degradation of active IRF3 only. In addition, we performed a two-step IP assay and pulled down the phosphorylated and nonphosphory lated IRF3 separately after P.y. gDNA stimulation. The IP assay revealed the endoge nous association between p-IRF3 and SHIP1 increased after P.y. gDNA stimulation, and there is no interaction between non-phosphorylated IRF3 and SHIP1 (Fig. 4L). Taken together, these data indicate that SHIP1 could interact with activated IRF3 and induce its autophagic degradation.

SHIP1 interacts with cargo receptor NDP52 and facilitates the association between NDP52 and IRF3
Accumulating evidence indicates that cargo receptors play essential roles in deliver ing substrates to the autophagosome for selective degradation (31). Thus, we next attempted to identify the potential cargo receptor responsible for the autophagic degradation of IRF3. Since SHIP1 is not a cargo receptor, we hypothesized that SHIP1 might bridge IRF3 to the cargo receptors for autophagic degradation. The co-IP assay showed that SHIP1 interacted with NDP52 but not other cargo receptors, such as p62, Nix, OPTN, NBR1, and Tollip (Fig. 5A), through its phosphatase domain (Fig. S5A). Therefore, SHIP1 interacted with both IRF3 and NDP52 through the phosphatase domain.
We next investigated whether NDP52 is involved in SHIP1-mediated autophagic degradation of IRF3. Indeed, SHIP1 could mediate IRF3 degradation in p62 KO cells, but not in NDP52 KO cells ( Fig. 5B and C). Additionally, the inhibition of IFN-I activation by SHIP1 was abolished in NDP52 KO cells (Fig. 5D). Likewise, CHX-chase assay results showed that the degradation of IRF3 was inhibited in NDP52 KO 293T cells, compared with WT 293T cells ( Fig. S5B and C). Taken together, these data suggest that SHIP1 contributes to IRF3 degradation through NDP52-mediated autophagy.
To further validate the role of SHIP1 in the association between IRF3 and NDP52. We performed confocal microscopy analysis and showed reduced co-localization puncta of IRF3-NDP52 when SHIP1 was KO in RAW 264.7 cells (Fig. 5E and G), and their co-localiza tion was elevated upon doxy-induced SHIP1 overexpression ( Fig. 5F and G). Furthermore, we observed that SHIP1 KO dramatically disrupted the interaction of endogenous IRF3 and NDP52 (Fig. 5H). Conversely, doxy-inducible of SHIP1 in RAW 264.7 cells promoted the association between IRF3 and NDP52 (Fig. 5I). Consistently, the interaction of SHIP1 and IRF3 was reduced in NDP52 KO 293T cells than in WT 293T cells (Fig. 5J). Altogether, these results indicate that SHIP1 facilitates the interaction between IRF3 and NDP52 for selective autophagic degradation.

SHIP1 increases the K63-linked ubiquitination of IRF3 at the K313 site
It is well known that cargo receptors recognize ubiquitin chains linked to substrates and target for autophagic degradation (17), and IRF3 has been reported to be degraded by inducing its connection with different kinds of ubiquitin chains, including K11, K27, K33, K48, and K63 ubiquitination after infections (17). We next investigated whether SHIP1 could affect the ubiquitination of IRF3 and found that SHIP1 depletion remarkably decreased the poly-ubiquitination of endogenous IRF3 (Fig. 6A; Fig. S6A). Meanwhile, doxy-inducible of SHIP1 in RAW 264.7 cells increased the ubiquitination of IRF3 ( Fig. 6B;  Fig. S6B). To further investigate how SHIP1 regulates IRF3 ubiquitination, we transfected 293T cells with Flag-IRF3, cMyc-SHIP1, and HA-tagged K11-linked ubiquitin (K11-Ub), K27-Ub, K33-Ub, K48-Ub, or K63-Ub and found that SHIP1 specifically increased K27-and K63-linked ubiquitin chains on IRF3, but not others ( Fig. 6C; Fig. S6C). Although mouse NDP52 and human NDP52 differ in the structures, with mouse NDP52 lacking the LIM-L domain (Fig. S6D) (32), both mouse and human NDP52 could bind to the ubiquitin molecules in 293T cells ( Fig. S6E and F). We also found that the endogenous association between NDP52 and Ub in mouse BMDMs (Fig. S6G). Moreover, SHIP1 mediated stronger association between NDP52 and IRF3 was abolished when ubiquitination of IRF3 was blocked by ubiquitin inhibitor TAK-243 (MLN7243) (33) (Fig. S6H), indicating that SHIP1 promotes the ubiquitination of IRF3, thus facilitating the interaction between IRF3 and NDP52. Several ubiquitination sites (K5, K39, K313, K360, and K366) of IRF3 have been reported previously (17). We next investigated the possible ubiquitination sites on IRF3 mediated by SHIP1. We transfected 293T cells with HA-SHIP1 and Flag-IRF3 or its mutants and observed that SHIP1-mediated degradation of IRF3 was abolished only when the IRF3 K313 site was mutated ( Fig. 6D and E; Fig. S6I). Moreover, IRF3 K313R displayed a reduction in K63-linked ubiquitination but not K27-linked ubiquitination after cotransfected with HA-SHIP1 ( Fig. 6F; Fig. S6J). In summary, these results reveal that SHIP1 induces the K63-linked ubiquitination of IRF3 at K313 and disrupts the stabilization of IRF3.

SHIP1 is downregulated by IFN-I-induced miR-155-5p during malaria infection
Finally, we investigated whether SHIP1 expression could also be regulated in response to malaria infection. Real-time PCR and western blot analyses revealed strong downregula tion of SHIP1 at both the mRNA and protein levels after P.y. N67 infection ( Fig. 7A and B). To further confirm the expression levels of SHIP1, we also stimulated RAW 264.7 cells or BMDMs with P.y. gDNA or RNA and observed similar trends of reduction of SHIP1 mRNA in RAW 264.7 cells or BMDMs after P.y. gDNA or RNA stimulation (Fig. S7A to D), which indicated that the expression of SHIP1 is restrained upon malaria infection. Since human malaria models may differ from the murine malaria models, we also used human P.f. 3D7 gDNA and RNA to stimulate THP-1 cells, and found a similar trend of downregulation in the mRNA levels of Ship1 after P.f. 3D7 gDNA or RNA stimulation ( Fig. S7E and F), as in the murine models. microRNAs (miRNAs) are small endogenous RNAs that are critical regulators of gene expression and promising candidates for biomarker development (34). About 100 miRNAs are altered during malaria infection (35). We detected several miRNAs that are most correlated with malaria (36), including miR-25-3p, miR-146a-5p, and miR-155-5p, for the decreased trends of Ship1, and found that miR-155-5p has the most corresponding trends with Ship1 (Fig. S7G to M). Since SHIP1 was reported to be a direct and function ally significant target of miR-155-5p in the context of viral and bacterial infections (37), we next sought to determine whether miR-155-5p is involved in the downregulation of SHIP1 during malaria infection. Indeed, the luciferase reporter assay showed that miR-155-5p mimics blocked the relative activity of the reporter luciferase of Ship1-3′UTR (Fig. S7N), confirming the direct interaction between miR-155-5p and Ship1 gene. In addition, the expression of miR-155-5p mimics decreased the expression of SHIP1 protein in 293T cells, while miR-155-5p inhibitors enhanced the protein level of SHIP1 ( Fig. S7O and P). Next, we asked whether miR-155-5p can inhibit SHIP1 during malaria parasite infection. A significant increase in the miR-155-5p expression levels during P.y. N67 infection (Fig. 7C) and nucleic acid components stimulation ( Fig. 7D; Fig. S7G to I) was detected, and there was an inverse correlation between the miR-155-5p and the Ship1 mRNA levels during malaria infection (Fig. 7E), indicating that SHIP1 could be downregu lated by miR-155-5p after P.y. N67 infection. Furthermore, we observed that the induction of miR-155-5p and the downregulation of SHIP1 (Fig. 7F and G) were abolished in Ifnar −/− BMDM after P.y. gDNA stimulation, compared with WT BMDM (Fig. 7H and I), suggested that the change of miR-155-5p and SHIP1 was controlled by downstream of IFN-I. Moreover, the decrease in SHIP1 triggered by IFN-β stimulation was restored when the cells were co-transfected with miR-155-5p inhibitors (Fig. 7J). Altogether, these results suggest that SHIP1 is downregulated by IFN-I-induced miR-155-5p during malaria infection and serves as a negative feedback mechanism between IFN-I signaling and autophagy.

DISCUSSION
IFN-I must be tightly controlled to maintain a balanced immune response between defense and homeostasis in infectious diseases (38,39). Mounting evidence has shown that multiple regulators are involved in the regulation of IFN-I production during infection, and regulators' expressions are also modulated by the activation of IFN-I signaling as well (40,41). Many negative regulators, such as Tetherin, NLRC5, NLRP11, and RTP4, are induced after infection, to feedback and inhibit the production of IFNs (5,32,42,43). The induction of negative regulator is important for organisms to maintain immune homeostasis, which also plays key roles in the escape mechanism of pathogens (44). However, other negative molecules are downregulated to relieve their inhibition on IFN-I signaling and enhance host immune responses against infection (45,46). Here, we discover another IFN-I response regulator SHIP1 and show that SHIP1 is downregulated by IFN-I-mediated miR-155-5p. Downregulation of SHIP1 increases phosphorylation and stability of IRF3 and IFN-I production, leading to improved host antimalarial immun ity after N67 parasite infection. Although reduction of SHIP1 after P.y. N67 infection could ameliorate degradation of IRF3 and enhance IFN-I in WT mice, which still cannot generate sufficient immune responses against malaria infection. Our data showed that only disruption of Ship1 at an early time of infection could enhance IFN-I response timely and protect host against P.y. N67 parasites, which is consistent with our previous study that only the-early-stage-IFNs can protect the organism (25,47). The timing of production of IFNs is critical for antimalarial immune responses (3), and that is why we observed more resistance in Ship1-chimeric mice in comparison to WT mice during malaria infection. During the N67 infection, the production of IFN-I peaks at day 1 after infection, which is consistent with previous publication (3,22), and decreases to a basal line at the late stage of infection. Although the amounts of IFN-α/β are low on day 6, Ship1-chimeric mice still generate high levels of IFN-α/β than the WT mice, which further suggests the inhibition of SHIP1 on the production of IFN-I throughout the course of malaria infection.
Decades of studies have revealed that SHIP1 is a multifunction protein that carries out its tasks by distinct means (48). Early studies mainly focused on SHIP1's function in bone biology, cancer, and mucosal inflammation (49). SHIP1 promotes the development of bone through its nonenzymatic functions (50); meanwhile, the inhibition of enzymatic  The interactions of SHIP1, IRF3, and NDP52 enhance activated IRF3 autophagic degradation and turns down IFN-I responses against malaria parasite infection. Meanwhile, IRF3-dependent IFN-I activated by parasite gDNA and RNA induce miR-155-5p to downregulate SHIP1, which serves as a negative feedback loop between IFN-I signaling and autophagy. Data in (A, C, D, F, G, H and I) are means ± SEM of at least three independent experiments, *** P < 0.001; NS, not significant (two-tailed Student's t test).
Research Article mBio activity of SHIP1 enhances the survival in cancer cells (26). The loss of SHIP1 function causes severe inflammation in human Crohn's disease (51). Recently, accumulating evidence has shown that SHIP1 affects cellular behaviors in innate and adaptive immune cells by dephosphorylating the PI3K (48). A study in AML patients showed that different SHIP1 mutants influence its phosphatase activity, and the AML mice increased their lifespan after lentiviral-mediated overexpression of SHIP1 (26). In autoimmune diseases, such as systemic lupus erythematosus and rheumatoid arthritis, SHIP1 functions as a critical negative regulator of TLR3-mediated IFN-β production through affecting the localization and activity of TBK1 (52). However, different from the previous studies, here we demonstrated that SHIP1 mainly targets on IRF3, but not TBK1, after P.y. N67 infection. Although SHIP1 may dephosphorylate IRF3 by its phosphatase activity, it can also regulate the stability of IRF3 through its nonenzymatic activity and promotes selective autophagic degradation of IRF3 to inhibit IFN-I production during malaria infection. Although gDNA stimulation lowers the expression of Ship1, its association with IRF3 increased after stimulation, which indicated that more SHIP1 interacted with activated IRF3 in the cytoplasm, even though total SHIP1 may be downregulated by miR-155 or undergo degradation upon malaria infection. Besides IFN-I signaling, NF-κB pathway is also critical during malaria infection. Parasite protein circumsporozoite protein was reported to interfere with the NF-κB pathway by outcompeting NF-κB nuclear import (53). In addition, the activation of heat shock protein 70 (Hsp70) and NF-κB could be suppressed by Hemozoin during severe malarial anemia (54). Thus, Artemisia annua L. 's active ingredient, artemisinin, has been widely used for malaria therapy by its numerous bioactivity including inhibition of the NF-κB pathway (55). SHIP1 is also well known for its regulation on NF-κB signaling by targeting MyD88 (48). However, SHIP1 did not affect NF-κB pathway and production of pro-inflammatory cytokines during malaria parasite N67 infection in our study. Our data suggested that the main role of SHIP1 in antimalarial immunity was dependent on its regulation of IFN-I signaling, which indicates that multiple innate immune signaling pathways contribute synergistically to host antimalarial immunity. Malaria is typically transmitted to humans by the bite of female Anopheles mosquito, which carriers the Plasmodium parasites (56). The transcription of Ship1 has been reported to be downregulated in P.f. infected patients (15). Due to the limitations in obtaining human samples, we used the murine malaria model for this study. We observed Ship1-chimeric mice can obtain more protection in both the nonlethal murine malaria stain (P.y. N67 and 17XNL) and the lethal strain (P.b. ANKA) models. Significantly higher levels of IFN-I responses, lower parasitemia, and better survival rates were observed in Ship1-chimeric mice than in WT mice. Altogether it verified SHIP1-mediated IFN-I immunity is not only specific for P.y. nigeriensis but also applicable to the other virulent and nonvirulent malaria.
Phosphorylation of IRF3 is a critical step for IFN-I signaling activation. IRF3 locates in the cytoplasm in the unstimulated state and is phosphorylated at specific serine residues which allow its dimerization and nuclear translocation to initiate IFN-I signaling upon infection. IFN-I needs to be tightly controlled to maintain host immune balance; thus, the post-translational modification of IRF3 plays a critical role in determining the immune state against infection. We previously reported that the phosphorylated IRF3 could undergo K27-linked ubiquitination, which led to the NDP52-mediated autophagic degradation and the immune suppression of IFN-I signaling after Sendai virus infection (17). Different from the previous study, here we demonstrate that SHIP1 also modulates the stability of activated IRF3 through K63-linked ubiquitination of IRF3 at K313, which is also a recognition signal for NDP52-dependent selective autophagic degradation resulting in inhibition of antimalarial immunity. These results indicate that different types of ubiquitination control IRF3 activity and stability through autophagic degradation and lead to distinct outcomes during different microbe infections. However, since SHIP1 is not an E3 ligase, it may regulate certain E3 ligase of IRF3 and then participate in the regulation of IRF3's ubiquitination. Whether this K63-linked ubiquitination of IRF3 mediated by SHIP1 also occurs in other infections requires further investigations.
Autophagy is a conserved degradation pathway that disassembles and recycles dysfunctional cellular components (57). CQ, as a traditional antimalarial drug, could directly target host's autophagy system in malaria, which also has additional functions, including antivirus, antibacteria, antiprotozoan, antiautoimmunity, and anticancer effects (58). However, the mechanism of CQ in malaria and other diseases is not understood well (58). Accumulating evidence has revealed the crosstalk between autophagy and IFN-I signaling in infectious diseases. During viral infection, coiled-coil domain-contain ing protein 50 (CCDC50) functions as autophagy receptor and negatively regulates the IFN-I by promoting the autophagic degradation of RIG-I/MDA5 (59). An Epstein-Barr virus encoding BCL2 homolog (BHRF1) inhibits IFN-β induction by targeting the mitochondria and blocking the nuclear translocation of IRF3 (60). Recent study showed that MGF-505-7R promotes the expression of the autophagy related protein ULK1 to degrade STING in African swine fever virus infection model (61). However, autophagy is a double-edged sword in infectious disease, and the relationship between autophagy and IFN-I is still not clarified. Autophagy can induce the production of IFN-I or degrade some important signal molecules in a direct or indirect way, thus dissecting the interactions between autophagy and IFN-I signaling in infectious diseases may help us better understand the underlying mechanism (19). Here, we demonstrate that SHIP1 enhances the association between cargo protein NDP52 and IRF3 and promotes the autophagic degradation of IRF3, leading to inhibition of IFN-I production and suppression of immune responses against N67 infection. Our study suggests targeting autophagy could be an effective interference for modulating antimalarial immunity.
Noncoding RNAs, including miRNAs and long noncoding RNAs, have emerged as key regulators of gene expression in immune cell development and function. Their expression is altered in different physiological and disease conditions, hence making them attractive targets for the understanding of disease etiology and developing complementary control strategies (36). The previous study summarized the miRNAs in malaria patients and highlighted several miRNAs that are involved in different stages of malaria development, such as miR-146a-5p, miR-25-3p, miR-155-5p, and others (62). miR-155-5p was reported to contribute to the pathogenesis of cerebral malaria by regulating endothelial activation, microvascular leak, and blood-brain barrier, and then aggravate the malaria process (63). In our study, we revealed that SHIP1 was targeted and downregulated by miR-155-5p, which was consistent with a previous study (62). However, we still do not know whether the proteasome or the autolysosome system is used to control SHIP1 degradation. Since 3-MA was reported as an potential antago nist of SHIP1 (63), the degradation of SHIP1 during malaria parasite infection may be influenced by autophagy in some way.
In summary, we identify a novel role for SHIP1 in negative regulation of antimalar ial immunity. Based on our findings, we proposed a working model to illustrate how SHIP1 suppresses IFN-I production during malaria infection. After Plasmodium infection, IRF3-dependent IFN-I is activated by parasite gDNA and RNA (3). However, SHIP1 binds to phosphorylated IRF3 to promote K63-linked poly-ubiquitination on IRF3 at K313, which stabilizes the interaction of IRF3 and its selective cargo receptor NDP52, and targets the activated IRF3 for autophagic degradation. Degradation of IRF3 thus turns off IFN-I responses against malaria infection. Meanwhile, SHIP1 is also regulated by IFN-I-induced miR-155-5p, which serves as a negative feedback loop controlling the crosstalk between IFN-I and autophagy (Fig. 7K). Inhibition of SHIP1 activity will stabilize pIRF3, enhance early IFN-I response, and suppress parasitemia. SHIP1 could be a potential therapeutic target for the treatment of malaria and other infectious diseases.

Generation of SHIP1 stable expression and knockout cell lines
The retroviral vectors were co-transfected with an expression plasmid for the vesicular stomatitis virus G protein into the HEK293T cell line. The medium was changed the following day, and the viral containing supernatant was collected 48 h after trans fection, filtered through a 0.45-μm filter and subsequently used to infect cells with polybrene (10 μg/mL). For SHIP1 ectopic expression, lentiviral particles were produced by transfecting HEK293T cells with FG-EH-DEST-SHIP1, VSVG, and Δ8.9. RAW 264.7 cells were infected by incubation with retrovirus-containing supernatant for 48 h. For ATG5 KO, BECN1 KO, NDP52 KO, and p62 KO cells, target sequences were cloned into pLenti CRISPRv2 by cutting with BsmBI. Transduced cells were purified by puromycin selection. SHIP1 KO RAW 264.7 cells and BM cells were generated using the CRISPR/Cas9 system, and the sequences of target sgRNA are as follows: Ship1-sgRNA (CACCGACGCAGAGTG CGTAGGCCCG).

Extraction of malarial gDNA and RNA
Parasite-infected mice blood was collected in saline solution and filtered to deplete white blood cells. Parasites were spun down after RBC lysis buffer treatment, and lysate was incubated with buffer A (150 mM NaCl, 25 mM EDTA, 10% SDS, and protein kinase) overnight. gDNAs were isolated using phenol/chloroform, and RNAs were isolated using TRIzol reagent (Invitrogen, CA, USA).

Animal and chimeric model
The packaged virus of sg-v2 or sg-m-ship1 was produced by 8.9/VSVG lentivirus expression system and infected BM cells of 6-to 10-week-old donor female mice. Recipient mice aged 8-10 weeks (female, WT C57BL/6) were irradiated with 900 cGy and randomly divided into two groups. They received BM cells (5 × 10 6 ) infected with sg-v2 or sg-m-ship1, respectively. The chimeric mice were housed in the SPF animal room and the BMs or spleens from two groups were collected to detect the level of SHIP1 after 6-8 weeks.

Malaria parasite infection procedures
The parasites P.y. N67, 17XNL, P.f. 3D7, and P.b. ANKA were initially obtained from the Malaria Research and Reference Reagent Resource Center (MR4, https://www.beiresour ces.org/About/MR4.aspx). The parasites P.y. N67 and P.b. ANKA have been previously described (24). For N67 infection, 2 × 10 5 iRBCs suspended in 200 μL PBS from the donor mice were injected i.p. into experimental mice (8-week-old female C57BL/6 mice were acquired from the Experimental Animal Centre of Southern Medical University). For ANKA infection, 2 × 10 5 iRBCs suspended in 200 μL PBS from the donor mice were injected i.p. into experimental mice. For P.y. 17XNL infection, 1 × 10 5 iRBCs suspended in 200 μL PBS from the donor mice were injected i.p. into experimental mice. All mouserelated procedures were performed according to experimental protocols authorized by the Southern Medical University Animal Care and Use Committee (SMUL2019243). Detailed information could be found in the Supporting Information.

Statistical analysis
The data of all quantitative experiments are presented as mean±SEM of at least three independent experiments. Curve data were assessed by GraphPad Prism 8.0 (San Diego, California, USA). And comparisons between groups for statistically significant differences were analyzed with a two-tailed Student's t test. The statistical significance was defined as P < 0.05.

ETHICS APPROVAL
All animal experiments were approved by the Southern Medical University Animal Care and Use Committee (SMUL2019243).