TIM-1 As a Signal Receptor Triggers Dengue Virus-Induced Autophagy

Dengue virus (DENV) infection triggers the activation of autophagy to facilitate the viral replication cycle from various aspects. Although a number of stimulators are proposed to activate autophagy, none of them appears prior to the uncoating process. Given that T-cell immunoglobulin and mucin domain 1 (TIM-1) receptor is a putative DENV receptor and promotes apoptotic body clearance by autophagy induction, it raises the possibility that TIM-1 may participate in the activation of DENV-induced autophagy. In this study, confocal images first revealed the co-localization of TIM-1 with autophagosomes in DENV-induced autophagy rather than rapamycin-induced autophagy, suggesting the co-transportation of TIM-1 with DENV during infection. The treatment of siRNA to knockdown TIM-1 expression in DENV-infected GFP-microtubule-associated protein light chain 3 (LC3)-Huh7.5 cells revealed that TIM-1 is required not only for DENV cellular internalization but also for autophagy activation. Furthermore, knockdown p85, a subunit of phosphoinositide 3-kinases (PI3Ks), which is co-localized with TIM-1 at rab5-positive endosomes caused the reduction of autophagy, indicating that TIM-1-mediated DENV-induced autophagy requires p85. Taken together, the current study uncovered TIM-1 as a novel factor for triggering autophagy in DENV infection through TIM-1-p85 axis, in addition to serving as a DENV receptor.


Introduction
Dengue virus (DENV), an RNA virus belonging to the flaviviridae family, which includes emerging and reemerging pathogens such as Zika virus (ZIKV), Japanese encephalitis virus (JEV), and West Nile virus (WNV), causes the most prevalent arthropod-born viral disease, with an estimated one hundred million symptomatic cases every year around the world [1]. DENV infection causes human diseases with a wide spectrum of clinical symptoms, ranging from asymptomatic infection or self-limited febrile illness named Dengue fever (DF) to life-threatening diseases including Dengue hemorrhagic fever (DHF) and Dengue shock syndrome (DSS) [2][3][4]. Currently, specific treatments for DENV are lacking. There is still an urgent need for anti-Dengue agents to prevent or treat DENV infections. As a result, more detail insights into DENV biology and Dengue-host interactions are necessary. DENV infection is a complicated and multifaceted process. DENV initiates infection of a permissive cell through binding

TIM-1 Required for the Activation of DENV-Induced Autophagy
To reveal the roles of TIM-1 on DENV-induced autophagy, immunofluorescence staining first confirmed the expression of TIM-1 in Huh7.5 cells, which is a DENV highly permissive cell lines derived from cellular hepatocellular carcinoma [10,49], and GFP-LC3 stably expressing Huh7.5 cells (GFP-LC3-Huh7.5). Both cell lines were stained with phycoerythrin (PE)-conjugated anti-TIM-1 antibodies and fluorescence signals were detected and analyzed by flow cytometry. The histograms of flow cytometry depicted a right-shift peak of PE signals in the anti-TIM-1 antibody staining group, compared to that in the unstaining and isotype-staining groups served as negative controls ( Figure 1A). The quantitative results of the mean fluorescence intensity (MFI) revealed at least five-fold stronger signals in the anti-TIM-1 antibody staining group than the control groups ( Figure 1B). Our results elucidated dual roles of TIM-1 of facilitating viral entry and activating autophagy during 95 DENV infection. 96

TIM-1 Required for the Activation of DENV-Induced Autophagy 98
To reveal the roles of TIM-1 on DENV-induced autophagy, immunofluorescence staining first 99 confirmed the expression of TIM-1 in Huh7.5 cells, which is a DENV highly permissive cell lines 100 derived from cellular hepatocellular carcinoma [10,49], and GFP-LC3 stably expressing Huh7.5 cells 101 (GFP-LC3-Huh7.5). Both cell lines were stained with phycoerythrin (PE)-conjugated anti-TIM-1 102 antibodies and fluorescence signals were detected and analyzed by flow cytometry. The histograms 103 of flow cytometry depicted a right-shift peak of PE signals in the anti-TIM-1 antibody staining group, 104 compared to that in the unstaining and isotype-staining groups served as negative controls ( Figure  105 1A). The quantitative results of the mean fluorescence intensity (MFI) revealed at least five-fold 106 stronger signals in the anti-TIM-1 antibody staining group than the control groups ( Figure 1B

119
These results suggested the expression of TIM-1 in both Huh7.5 and GFP-LC3-Huh7.5 cell lines. 120 Moreover, to investigate the roles of TIM-1 in the activation of DENV-induced autophagy, we also 121 generated TIM-1-knockdowned GFP-LC3-Huh7.5 cells by a siRNA (si-TIM-1) treatment. Dynamic 122 expression of TIM-1 of siRNA-treated GFP-LC3-Huh7.5 cells was examined and quantified by flow 123 cytometry. In comparison with antibody-free, isotype control, and non-target siRNA control (si-124 Control) groups, the expression of TIM-1 declined as time of treatment in si-TIM-1 groups increased 125 We confirmed by flow cytometry the reduced expression of TIM-1 after using specific siRNA (si-TIM-1). Representative experiments of the transfection of GFP-LC3-Huh7.5 with non-targeting siRNA (blue) and si-TIM1 with 20 pmol (orange) and 40 pmol (purple) were displayed, respectively. Two additional control groups, including unstained and isotype, were marked in blue and red. (D) The quantification of the histogram presented the mean fluorescence intensity of each group from (C).
These results suggested the expression of TIM-1 in both Huh7.5 and GFP-LC3-Huh7.5 cell lines. Moreover, to investigate the roles of TIM-1 in the activation of DENV-induced autophagy, we also generated TIM-1-knockdowned GFP-LC3-Huh7.5 cells by a siRNA (si-TIM-1) treatment. Dynamic expression of TIM-1 of siRNA-treated GFP-LC3-Huh7.5 cells was examined and quantified by flow cytometry. In comparison with antibody-free, isotype control, and non-target siRNA control (si-Control) groups, the expression of TIM-1 declined as time of treatment in si-TIM-1 groups increased ( Figure 1C,D), indicating the treatment of si-TIM-1 in GFP-LC3-Huh7.5 cells can specifically repress the expression of TIM-1.
To confirm TIM-1 is indeed required for DENV-induced autophagy, the numbers of GFP-LC3 punta that represents the formation of autophagosomes in TIM-1-knockdowned cells was examined. GFP-LC3-Huh7.5 cells were infected by DENV at a multiplicity of infection (MOI) of five in the absence or presence of siTIM-1. Apparently, compared to siRNA-free (mock) and si-Control groups, the treatment of si-TIM-1 inhibited the formation of GFP-LC3-positive autophagosomes in GFP-LC3-Huh7.5 cells at 30 min post-infection ( Figure 2A). The quantification results of time course experiments revealed that the numbers of GFP-LC3-positive punta cells were elevated in either mock or si-Control-treated DENV-infected cells at 15, 30 and 60 min post-infection ( Figure 2B). In contrast, the siTIM1 treatment effectively suppressed the formation of GFP-LC3-positive autophagosomes in DENV-infected cells ( Figure 2B). Furthermore, given that autophagy promotes DENV production [9,29,50], the effect of the siTIM1 treatment on viral production was examined by the plaque assay. Significantly, the titer of newly synthesized DENV particles decreased in the presence of si-TIM-1 ( Figure 2C). This shows that TIM-1 triggers autophagy in DENV infection, which would in turn affect viral production facilitation. As previous studies have shown that TIM-1 also serves as an entry receptor of DENV [40], this implies that it might in fact have a dual role in DENV infection. Accordingly, TIM-1 has been demonstrated to be an entry receptor of DENV because genetic ablation of TIM-1 inhibits DENV infection [40]. Taken together, these results indicated that TIM-1 might exert dual roles, facilitating DENV entry and triggering autophagy, in DENV infection.
Because DENV binds with TIM-1 to initiate infection and DENV particles are translocated into autophagosomes after internalization [9,40], we speculated that TIM-1 is located in autophagosomes after DENV infection. To investigate this possibility, the images of GFP-LC3 punta, which represent the formation of autophagosomes, and TIM-1 in GFP-LC3-Huh7.5 cells that were exposed to DENV at a MOI of 10 for one hour were visualized by fluorescence microscopy. The numbers of GFP-LC3 punta increased significantly by either the treatment of rapamycin or DENV in GFP-LC3-Huh7.5 cells, suggesting the activation of autophagy ( Figure 2D, Top panel). By merging the signals of GFP-LC3 and immunofluorescence images of TIM-1 ( Figure 2D, middle panel), we elucidated the colocalization of TIM-1 with autophagosomes. Apparently, TIM-1 signals overlapped with GFP-LC3 punta when cells were infected with DENV instead of treated with rapamycin ( Figure 2D, bottom panel). The quantification of TIM-1 colocalized with autophagosomes was determined by Mander's coefficient of TIM-1 signals overlapping GFP-LC3 punta. As shown in Figure 2E, the value of Mander's coefficient in the DENV group is 0.51, which is significantly higher than that in either mock or rapamycin-treated group. In summary, these results indicated that TIM-1 is specifically located within autophagosomes in DENV-induced autophagy rather than in rapamycin-induced autophagy. Figure 2C). This shows that TIM-1 triggers autophagy in DENV infection, which would in turn affect 140 viral production facilitation. As previous studies have shown that TIM-1 also serves as an entry 141 receptor of DENV [40], this implies that it might in fact have a dual role in DENV infection. 142

(
Accordingly, TIM-1 has been demonstrated to be an entry receptor of DENV because genetic ablation 143 of TIM-1 inhibits DENV infection [40]. Taken together, these results indicated that TIM-1 might exert 144 dual roles, facilitating DENV entry and triggering autophagy, in DENV infection. 145 and GFP-LC3 punta (green) in GFP-LC3-Huh7.5 cells treated either with rapamycin or DENV infection for an hour were determined by confocal microscopy (Olympus FV-10i). The insets showed more colocalization of TIM-1 with LC3 puncta in DENV-infected cells rather than in rapamycin-treated cells. Scale bar, 20 µm. (E) The colocalization of TIM-1 with GFP-LC3 punta was quantitatively analyzed by Mander's coefficient from three independent experiments (>25 cells). n.s. means non-significant; ** p < 0.01.

The Colocalization of TIM-1 with DENV Particles in Autophagosomes
Since DENV infection specifically enhanced co-localization of TIM-1 and autophagosomes ( Figure 2D,E), we would like to investigate the possibility that DENV particles and TIM-1 are co-localized in autophagosomes by using the single-virus imaging analysis. To visualize single DENV particles directly, DENV particles were first labeled with a fluorescence dye, atto647N, by conjugating atto647N-NHS ester with the amino group of the viral envelope protein and purified through a Sephadex G-25 size-exclusion column. As the diameter of a DENV particle is 50 nm, the fluorescent signals of atto647N existent from fraction 7 to fraction 11 ( Figure 3A, solid circles) indicated atto647N-labeled DENV particles (atto647N-DENV) when compared to the fractions containing the 40 nm fluorosphere ( Figure 3A, solid squares). No fluorescence signal was detected within the same range of fractions from the dye-free DENV group ( Figure 3A, open circles). These results elucidated that DENV particles were successfully conjugated with atto647N. The infectivity of atto647N-DENV particles was then determined by the plaque assay. Both dye-free DENV and atto647N-DENV presented similar virus titers, suggesting that the atto647N labeling procedure showed no significant influence on DENV infectivity ( Figure 3B). To visualize atto647N-DENV particles, Huh7.5 cells were incubated with atto647N-DENV cells at 4 • C for 30 min. The surrounding distribution pattern of atto647N signals indicated the binding of DENV on the surface of Huh7.5 cells ( Figure 3C). We further detected DENV virion by using an anti-DENV E protein antibody. Immunofluorescent staining revealed that the atto647N signals and the anti-DENV E protein antibody overlapped, suggesting atto647N spots can represent DENV virion images ( Figure 3C). These results confirmed the generation of infectious atto647N-DENV particles.
To confirm whether atto647N-DENV particles induce autophagy and colocalize with autophagosomes in the absence or presence of si-TIM-1, GFP-LC3-Huh7.5 cells, pre-treated with either si-control or si-TIM-1 for 48 h were incubated with atto647N-DENV cells at an MOI of 10 at 37 • C for 30 min. The confocal microscopy images showed that not only an increase in the number of GFP-LC3-positive vesicles was detected, which represents the formation of autophagosomes, but also the events of atto647N-DENV co-localization with autophagosomes elevated in both mock (siRNA-free) and si-control groups, as compared to those at the initial time point of infection ( Figure 3D). This was in line with our previous study [9]. In contrast, the number of internalized DENV particles (red) and the formation of GFP-LC3-positive autophagosomes (green) in si-TIM-1-treated Huh7.5 cells apparently decreased ( Figure 3D). In addition, triple-fluorescence imaging from confocal Z-stack images and cross-sectional analyses using fluorescence intensity profiling displayed spatial coincidence of GFP-LC3 punta (green), TIM-1 (red), and atto647N-DENV (blue) in GFP-LC3-Huh7.5 cells ( Figure 3E), suggesting TIM-1 is co-localized with the DENV particle in an autophagosome. To sum up, these results elucidated that TIM-1 plays dual roles of promoting viral entry and triggering autophagy during the early phase of DENV infection.  GFP-LC3-positive vesicles was detected, which represents the formation of autophagosomes, but also 225 the events of atto647N-DENV co-localization with autophagosomes elevated in both mock (siRNA-226 free) and si-control groups, as compared to those at the initial time point of infection ( Figure 3D). 227 This was in line with our previous study [9]. In contrast, the number of internalized DENV particles 228

The Colocalization of p85 with TIM-1 in rab5-Positive Endosomes
Given that TIM-1 interacts with p85, the PI3K adaptor, resulting in autophagy induction [48], we proposed that p85 may be a downstream signal molecule for TIM-1-mediated autophagy activation induced by DENV. In addition, as DENV particles are internalized with TIM-1 and located in Rab5-positive endosomes after clathrin-mediated endocytosis [7,40], it is highly possible that TIM-1 is co-localized with p85 in endosomes. To validate this possibility, Huh7.5 cells were transfected with Rab5-DsRed plasmids prior to DENV infection at a MOI of 10. The location of TIM-1, p85 and Rab5-positive endosomes were detected at 30 min post-infection. Compared to the virus-free control group (mock), DENV infection enhanced the co-localization of TIM-1, p85 in Rab5-positive endosomes ( Figure 4A). Fluorescence intensity profiling from cross sectional analyses of confocal Z-stack images depicted spatial coincidence of GFP-TIM-1 (green), Rab5-positive endosomes (red), and p85 (cyan) in Huh7.5 cells ( Figure 4B). These results supported that DENV infection enhances the association of TIM-1 with p85 at endosomes at the early phase of DENV-infection. Given that TIM-1 interacts with p85, the PI3K adaptor, resulting in autophagy induction [48], 237 we proposed that p85 may be a downstream signal molecule for TIM-1-mediated autophagy 238 activation induced by DENV. In addition, as DENV particles are internalized with TIM-1 and located 239 in Rab5-positive endosomes after clathrin-mediated endocytosis [7,40], it is highly possible that TIM-

p85 Knockdown Attenuates DENV-Induced Autophagy and Virus Production
To further elucidate the function of p85 in DENV-induced autophagy, we examined the formation of GFP-LC3-positive punta in GFP-LC3-Huh7.5 cells treated with siRNA against p85 (si-p85). The expression of p85 of siRNA-treated GFP-LC3-Huh7.5 cells was examined and quantified. Western blotting assays depicted clearly that the expression of p85 was reduced by si-p85 ( Figure 5A). Compared to those in the non-target siRNA control group (si-Control), quantification data showed more than 50% knockdown efficiency of p85 expression after 48 h post-transfection. To validate the requirement of p85 in the activation of DENV-induced autophagy, GFP-LC3-Huh7.5 cells were infected by atto647N-DENV at a MOI of 10 in the absence or presence of si-p85 and the formation of GFP-LC3-positive autophagosomes was examined. Apparently, compared to siRNA-free (mock) and si-Control groups, the treatment of si-p85 inhibited the formation of GFP-LC3-positive autophagosomes but not the internalization of DENV particles in GFP-LC3-Huh7.5 cells ( Figure 5B). Quantification of the time course results revealed that the numbers of GFP-LC3-positive autophagosomes were elevated in either siRNA-free or si-Control-treated DENV-infected cells at 30 and 60 min post-infection ( Figure 5C). In contrast, si-p85 treatment effectively suppressed the formation of GFP-LC3-positive autophagosomes in DENV-infected cells ( Figure 5C). Furthermore, the effect of the si-p85 treatment on the new synthesis of DENV particles was examined by the plaque assay. The production of DENV particles was significantly repressed in the presence of si-p85 ( Figure 5D). Taken together, these results confirmed that p85 is required for DENV-induced autophagy and contributes to DENV production. 8 showed more than 50% knockdown efficiency of p85 expression after 48 h post-transfection. To 265 validate the requirement of p85 in the activation of DENV-induced autophagy, GFP-LC3-Huh7.5 cells 266 were infected by atto647N-DENV at a MOI of 10 in the absence or presence of si-p85 and the 267 formation of GFP-LC3-positive autophagosomes was examined. Apparently, compared to siRNA-268 free (mock) and si-Control groups, the treatment of si-p85 inhibited the formation of GFP-LC3-269 positive autophagosomes but not the internalization of DENV particles in GFP-LC3-Huh7.5 cells 270 ( Figure 5B). Quantification of the time course results revealed that the numbers of GFP-LC3-positive 271 autophagosomes were elevated in either siRNA-free or si-Control-treated DENV-infected cells at 30 272 and 60 min post-infection ( Figure 5C). In contrast, si-p85 treatment effectively suppressed the 273 formation of GFP-LC3-positive autophagosomes in DENV-infected cells ( Figure 5C). Furthermore, 274 the effect of the si-p85 treatment on the new synthesis of DENV particles was examined by the plaque 275 assay. The production of DENV particles was significantly repressed in the presence of si-p85 ( Figure  276 5D). Taken together, these results confirmed that p85 is required for DENV-induced autophagy and 277 contributes to DENV production.

Discussion
The present study uncovers a novel role of TIM-1 during DENV infection in which TIM-1 specifically triggers DENV-induced autophagy through the PI3K pathway at early phase of infection. We, thus, would like to define an unprecedented triggering mechanism of autophagy by DENV infection as illustrated in Figure 6. To initiate the infection process, a DENV particle is first recognized by TIM-1 and internalized into a Rab5-positive endosome through clathrin-mediated endocytosis [7,40]. Next, p85 is recruited to TIM-1-positive endosomes, which activates autophagy to form autophagosomes through the PI3K pathway. Finally, the DENV particle is transported into an autophagic vesicle with TIM-1 for further processing in the viral life cycle to produce new viral particles. Recognizing that TIM-1 is not only a DENV receptor, but also a key cellular factor triggering the activation of DENV-induced autophagy through the PI3K pathway provides a mechanistic understanding of the dual roles of TIM-1 in facilitating DENV production.

Discussion 298
The present study uncovers a novel role of TIM-1 during DENV infection in which TIM-1 299 specifically triggers DENV-induced autophagy through the PI3K pathway at early phase of infection. 300 We, thus, would like to define an unprecedented triggering mechanism of autophagy by DENV 301 infection as illustrated in Figure 6. To initiate the infection process, a DENV particle is first recognized 302 by TIM-1 and internalized into a Rab5-positive endosome through clathrin-mediated endocytosis 303 [7,40]. Next, p85 is recruited to TIM-1-positive endosomes, which activates autophagy to form 304 autophagosomes through the PI3K pathway. Finally, the DENV particle is transported into an 305 autophagic vesicle with TIM-1 for further processing in the viral life cycle to produce new viral 306 particles. Recognizing that TIM-1 is not only a DENV receptor, but also a key cellular factor triggering 307 the activation of DENV-induced autophagy through the PI3K pathway provides a mechanistic 308 understanding of the dual roles of TIM-1 in facilitating DENV production. 309

315
Autophagy is initially activated by DENV in the early phase of infection before the viral 316 uncoating process [9]. Although several factors including DENV NS4A/NS4B, virus-induced ER 317 stress, and AMPK have been proposed to be required for autophagy [16,25,[30][31][32], none of them 318 appears prior to the viral uncoating process, suggesting an unidentified mechanism of activated 319 autophagy during the early phase of DENV infection. DENV has evolved an elegant strategy, named 320 viral apoptotic mimicry, in which it manipulates host cells for cellular entrance by the exposure of 321 Figure 6. The scheme of TIM-1-mediated DENV-induced autophagy activation. DENV binds with TIM-1 to internalize into cells. Following receptor-mediated endocytosis, DENV delivers with TIM-1 in the Rab5-endosome and recruits p85 association. TIM-1 interacts with p85 and Rab5 to enhance autophagosome formation. Furthermore, DENV-contained endosomes deliver DENV to autophagosomes. Autophagy is initially activated by DENV in the early phase of infection before the viral uncoating process [9]. Although several factors including DENV NS4A/NS4B, virus-induced ER stress, and AMPK have been proposed to be required for autophagy [16,25,[30][31][32], none of them appears prior to the viral uncoating process, suggesting an unidentified mechanism of activated autophagy during the early phase of DENV infection. DENV has evolved an elegant strategy, named viral apoptotic mimicry, in which it manipulates host cells for cellular entrance by the exposure of PtdSer on the surface [49]. DENV mimics apoptotic cells by exposing PtdSer at the viral membrane, to facilitate the binding of TIM-1 [45], which functions as an authentic DENV cellular receptor to promote DENV entrance and plays an active role in virus endocytosis [40]. As TIM-1 binds with PtdSer on apoptotic cells to promote apoptotic clearance by activating autophagy [48], we speculated that TIM-1 might mediate DENV-induced autophagy. Using siRNA to knockdown TIM-1 not only reduced DENV internalization ( Figure 3D), which is consistent with previous studies [45], but also significantly suppressed DENV-induced activation of autophagy (Figure 2A,B). Both of the effects may contribute to the reduction of DENV virions production ( Figure 2C) since DENV infection triggers autophagy to facilitate viral production [9,29,50]. This study thus not only corroborated with previous findings with regards to TIM-1 acting as an DENV entry receptor [40], but also elucidated a novel signaling role of TIM-1 as a trigger to activate autophagy in DENV infection.
In the current study, we revealed that treatments of both DENV and rapamycin can trigger autophagy, but the colocalization of TIM-1 with autophagosomes is only detected in DENV-induced autophagy ( Figure 1C), suggesting that the activation of DENV-induced autophagy may be independent of the conventional pathways in which DENV production can be enhanced by rapamycin treatments [9,29,51,52]. In contrast to TIM-1 that activates autophagy right after cellular internalization of DENV, other autophagy activators, such as DENV NS4A/B proteins, ER stress, and AMPK appear after viral translation [16,25,[30][31][32], suggesting the existence of multiple activators of autophagy during DENV infection. This implies that autophagy triggered by various activators plays different roles to facilitate DENV production. For instance, DENV infection activates AMPK for viral replication as well as lipophagy induction but not basal autophagy [32]. Moreover, one recent study reported that autophagy is required for the post-RNA replication process during DENV infection in HeLa cells [53]. As HeLa cells are TIM-1-negative cells [45], we speculated that autophagy cannot benefit from DENV replication in the early phase of infection in a TIM-1-absence cell line. In contrast, DENV presents a higher infectious ability in both Huh7.5 and A549 cells of TIM-1 positive cell lines [45]. Knockdown TIM-1 expression in Huh7.5 cells suppressed autophagy and reduced DENV production ( Figure 2). The results indicated that TIM-1 promotes DENV replication by activating autophagy in the early phase of infection. Moreover, there are non-autophagic functions of autophagy-relevant proteins, such as endocytosis, LC3-associated phagocytosis, and cellular transportation [54]. It would be worthy to investigate how DENV selects and controls these different pathways for viral production and whether there are additional host factors, such as autophagy-relevant proteins, involved in DENV replication.
In addition to TIM-1, the current study also revealed the requirement of p85, a regulatory subunit of PI3Ks, for DENV-induced autophagy ( Figure 5C). The initiation of autophagy is highly regulated by a group of phosphoinositides generated by PI3Ks [55]. p85 participated in KIM-1/TIM-1 phosphorylation-mediated signaling [56] and regulates starvation-induced autophagy through IKK [57]. Our results also depicted the co-localization of p85 with TIM-1 in DENV-infected cells ( Figure 4A). These results are consistent with the previous finding that KIM-1/TIM-1 interacts with p85 to induce autophagy for the clearance of apoptotic cells in the KIM-1/TIM-1-mediated phagocytosis [48]. Moreover, p110β, a catalytic subunit of PI3Ks, associates with Rab5 and enhances the rab5-Vps34 interaction to promote autophagy [55]. Given that DENV particles are internalized with TIM-1 and located in the rab5-positive endosome after clathrin-mediated endocytosis and TIM-1 is co-localized with p85 at endosomes ( Figure 4) [7,40], we would like to propose a novel TIM-1-p85 signaling axis model for activating DENV-induced autophagy. After cellular internalization with DENV, TIM-1 interacts with p85 in early endosomes and recruits PI3Ks to enhance the rab5-Vps34 interaction, promoting activation of autophagy. Further studies are thus needed in order to address the following issues: Whether p85 dynamically associates with DENV-containing vesicles, whether interaction of rab5 with Vps34 is required for DENV-induced autophagy, and whether there are additional host factors involved in DENV-induced autophagy.
Given that an increase in cellular autophagy facilitates the production of many viruses, a common evolutionary selection exists among various viruses to trigger or activate autophagy [16,22]. Additionally, TIM-1 is recognized as a cellular receptor of various viruses to facilitate viral infection [39][40][41][42][43][44][45][46]. In this study, we showed a novel role of TIM-1 to activate autophagy through the TIM-1-p85 signaling axis in the early phase of DENV infection, suggesting the dependence of TIM-1 on the activation of autophagy of DENV infection. Because TIM-1 is recognized as a therapeutic target in cancers, kidney injury, and immune diseases [58][59][60], we believe that developing potent TIM-1 inhibitors might be a promising possibility for successful anti-DENV and anti-viral drugs in the future.

DENV Amplification
DENV strain 16681 serotype 2 was propagated in C6/36 cells. C6/36 cells were grown until the confluent monolayer is formed and infected with DENV at MOI 0.02 in 2 mL serum free RPMI 1640 medium at 28 • C with 5% CO 2 for two hours. During DENV infection, cells were gently shaken at 15 min intervals. Afterwards, the infectious medium in C6/36 cells was replaced with a low serum RPMI 1640 medium containing 2% FBS, 1× NEAA, 1% P/S, and 1mM sodium pyruvate, and incubated at 28 • C with 5% CO 2 . The culture medium was collected at day 4, 7, 11, and 14, and cell debris were removed by centrifugation with 9,000× g for 30 min at 4 • C. The supernatant of culture medium was aliquoted and stored at −80 • C.

Plaque Assay
BHK-21 cells were seeded in six-well plates (4 × 10 5 cells/well) at 37 • C with 5% CO 2 . Cells were inoculated with DENV suspensions in 10-fold serial dilution by serum-free DMEM (Gibco, 12100, Grand Island, NY, USA). The cells were incubated at 37 • C with 5% CO 2 for two hours and were gently shaken every 15 min. After removal of virus suspension, the cells needed to be overlaid with 1:1 mixture of 2% low-melting agarose (Lonza, 50100, Rockland, ME, USA) with two-fold DMEM containing 4% FBS and 2% P/S. The infected BHK-21 cells were incubated at 37 • C with 5% CO 2 for 7 days. Cells were fixed with 10% formaldehyde for 2 h post-infection. Next, the agarose was removed and the cells were stained with crystal violet for 4 h. Finally, the titer of the virus was determined by counting the numbers of the virus plaques.

Immunofluorescence Staining and Flow Cytometry Analysis
Sample cells were collected with the same density and washed with a staining buffer (PBS with 1% FBS). After centrifugation at 112× g for 5 min, the cells were re-suspended in a 100 µL staining buffer and stained with a 10 µL PE-conjugated anti-human TIM-1 antibody (R&D systems, clone#219211, Minneapolis, MN, USA) at 4 • C for 30 min. The negative controls were the unstained group (without antibody) and the isotype group, in which the cells were stained with the PE-conjugated mouse IgG2B (R&D systems, clone#133303, Minneapolis, MN, USA). Cells were washed with a 200 µL staining buffer twice, re-suspended in a 500 µL staining buffer, and temporarily stored at 4 • C. Cells were analyzed using a flow cytometer (Beckman coulter, CytoFLEX flow cytometer, Brea, CA, USA).

Quantification of DENV-Induced Autophagy by GFP-LC3 Puncta Numbers
GFP-LC3-Huh7.5 cells were seeded in a 6-well with a density of 2 × 10 5 and were incubated at 37 • C with 5% CO 2 overnight. The culture medium was removed and the cells were washed with PBS, followed by inoculation with DENV at MOI of 5 in serum-free DMEM (Gibco, 11995) at 4 • C for 30 min to allow DENV to binding on the cell surface rather than entering the cells. Cells were gently washed with PBS, and were shifted to 37 • C with 5% CO 2 for DENV infection at different time points: 0, 15, 30, and 60 min. At each time point, the virus supernatant was removed and cells ware fixed with 4% paraformaldehyde at RT for 15 min. After being washed with PBS, the cells were stained with DAPI at RT for 10 min and the residual DAPI was cleaned with PBS twice. Cells were kept in PBS for fluorescence imaging. Fluorescence images were captured by Olympus IX70 equipped with 20× objective lens, and GFP-LC3 puncta numbers over 5 in one cell were regarded as a criterion for an autophagy positive cell. We determined the total cell numbers by recognizing the nucleus with DAPI staining and the numbers of autophagy positive cells. The percentage of autophagy positive cells was calculated by (autophagy positive cells/total cells) times 100%. The average percentages in autophagy positive cells were collected from 3 independent experiments and the fold change was determined by the average percentages of the DENV infection groups divided to that in mock group.

DENV Production Assay
GFP-LC3-Huh7.5 cells were seeded in a 6-well with a density of 2 × 10 5 at 37 • C, 5% CO 2 overnight. Pretreatment of siRNA in cells at 37 • C, 5% CO 2 for 48 h. The cells were washed with PBS, inoculated with DENV at MOI of 1 at 4 • C for 30 min and shifted to 37 • C. At 24 or 48 h post-infection, the cell culture medium was collected and the extracellular virus titers were measured by a plaque assay.

Conclusions
In conclusion, we demonstrated that TIM-1 activates DENV-induced autophagy in the early phase of infection by recruiting p85 to virus-containing Rab5-positive endosomes. Our studies not only confirm TIM-1 as a DENV entry receptor, but also uncover a novel role of TIM-1 as a signaling receptor to trigger autophagy, resulting in facilitating the process of DENV life cycle.