Scavenger Receptor SCARA5 Acts as an HMGB1 Recognition Molecule Negatively Involved in HMGB1-Mediated Inflammation in Fish Models

Scavenger receptors (SRs) represent one group of the most important pattern recognition receptor family members; this group consists of a large subfamily with eight classes (A–H) according to their similarities in multidomain protein structure (1). SRs were initially found to recognize acetylated low-density lipoproteins, participate in lipid metabolism, and contribute to various lipid metabolism–relevant disorders, such as atherosclerosis and Alzheimer’s disease (2). Consequently, SRs have become increasingly attractive for their critical functions in innate immunity, among which class A members have received particular attention (2). Class A members are composed of the macrophage SR, macrophage receptors with long collagenous structure, SR with C-type lectin domain, and SR class A member 5 (SCARA5). These receptors can bind with an array of pathogen-associated molecular patterns (PAMPs), such as LPS, peptidoglycan, lipoteichoic acid, and unmethylated CpG-DNA, which stimulate a number of host innate inflammatory responses (3–5). Naturally, the host innate inflammatory responses are essential for host defense against pathogen invasions. However, excessive inflammatory responses may result in severe inflammatory disorders (6). Therefore, the innate inflammatory reactions present a double-edged mechanism, which must be tightly regulated to maintain immune homeostasis and thus prevent inflammatory disorders. Apart from the role of PAMPs in the augmentation of inflammation, damage-associated molecule patterns (DAMPs), such as the recently identified high-mobility group box 1 (HMGB1) protein, also endorse key functions to enhance inflammation (7, 8). SRs play crucial anti-inflammatory roles by eliminating PAMPs; however, whether they participate in DAMP-mediated inflammation remains poorly understood.

HMGB1 was first identified as a nonhistone nuclear protein bound to the minor groove of DNA without sequence specificity that stabilizes DNA structure and modulates transcriptional activity (9). Recently, this protein has been defined as a classical proinflammatory mediator in the pathogenesis of both sterile and infectious inflammatory processes outside the cell (10, 11). HMGB1 promotes inflammation through various mechanisms. For example, it acts as a late mediator of endotoxemia to enhance endotoxin lethality and associates with IL-1β, LPS, Pam3CSK4, and CpG-oligodeoxynucleotides (ODNs) to induce the burst of inflammatory cytokines (12–14). During bacterial or viral infection, HMGB1 can be secreted by the activated monocytes and macrophages to act as a molecular chaperone for the transplant of LPS and CpG-ODNs to their receptors TLR4 and TLR9 to elicit inflammatory responses (14–16). To date, RAGE is identified to be the major potent receptor for HMGB1, although TLR2 and TLR4 are also closely involved in HMGB1 signaling (17–19). However, other receptors, specifically the negative regulatory receptor for HMGB1, have never been described.

Our previous study identified a SCARA5 homolog from pufferfish (Tetraodon nigroviridis) model and showed that SCARA5 is a LPS recognition receptor, which negatively regulates NF-κB activation by competing with TNFR-associated factor 2 recruitment into the TNF-α signaling pathway (4). In the present study, we report SCARA5 acts as an HMGB1 receptor that negatively regulates CpG-ODN–triggered inflammation, in which HMGB1 is an essential partner for CpG-ODN transplant to TLR9 in the endosomes (20). To our knowledge, this report is the first to show that SCARA5 is an HMGB1 receptor inhibiting the synergic effect of CpG-ODN with HMGB1 on TLR9-mediated inflammatory response; thus, a new receptor for HMGB1 is added, and new insights into the function of SR family members in host innate inflammatory responses are provided. The present findings also contribute to enhance our understanding about the biology of HMGB1.

Wild-type spotted green pufferfish (T. nigroviridis) and AB zebrafish (Danio rerio) were bred and maintained in a circulating water bath at 26–28°C under standard conditions as previously described (21). The fish were held in the laboratory for at least 2 wk before the experiment to evaluate their overall health. Only healthy fish, as determined by their general appearance and level of activity, were used in the study. Zebrafish embryos were collected at different stages of embryonic development in accordance with previously established protocols (22). Imprinting control region mice (4–8 wk) were obtained from the Laboratory Animal Unit of Zhejiang Academy of Medical Sciences (Hangzhou, China) and housed under standard conditions. All animal experiments were performed in accordance with legal regulations and were approved by a local Ethics Committee.

cDNAs of Tetraodon HMGB1 (TnHMGB1) and TnTLR9 were amplified by RT-PCR by using the homologous sequences predicted from the University of California Santa Cruz and National Center for Biotechnology Information genome databases. Full-length encoding sequence of Tetraodon SCARA5 (TnSCARA5) was generated on the basis of our previously published sequence in the National Center for Biotechnology Information genome database (no. AFX69239.1). The primers used in cloning are listed in Supplemental Table I. PCR products were purified using a gel extraction kit (Qiagen), inserted into the pGEM-T easy vector (Promega), and transformed into competent Escherichia coli TOP10 cells (Invitrogen). Plasmid DNA was purified using a plasmid Miniprep kit (Qiagen) (23). Sequence alignments of TnHMGB1 and TnTLR9 were analyzed using the ClustalW program (version 1.81) (24, 25). Phylogenetic trees were constructed with MEGA 4.0 by using the neighbor-joining method.

Encoding sequences for wild-type and mutant TnHMGB1 proteins with deficiency of A, B, or T boxes were inserted into pET28a to construct prokaryotic expression vectors, namely, pET28a-TnHMGB1, pET28a-TnHMGB1-mutA, pET28a-TnHMGB1-mutB, and pET28a-TnHMGB1-mutT. The encoding sequences of TnHMGB1 and TnTLR9 were subcloned into pDsRED-C1 and pcDNA6/myc-His B (Invitrogen) to obtain the recombinant fusion proteins of TnHMGB1 with enhanced red fluorescent protein and TnTLR9 with Myc-tag at the C-terminals, respectively. The plasmids encoding recombinant fusion proteins of TnSCARA5 with enhanced GFP and Myc/His-tag (p-enhanced GFP [EGFP]-TnSCARA5 and pcDNA6/myc-His-TnSCARA5) were constructed previously in our laboratory (4). NF-κB luciferase and pRL-TK (Renilla luciferase reporter) constructs were purchased from Clontech (Palo Alto, CA) and Promega. The primers used for construct generation are listed in Supplemental Table I. Plasmids for transfection and microinjection were prepared free of endotoxin by using an EZNA plasmid mini kit (Omega Bio-Tek).

Total RNA from leukocytes or selected tissues, including liver, intestine, spleen, kidney, skin, gill, muscle, and brain, in normal or CpG-Singapore grouper iridovirus (SGIV)-ODN–stimulated fish was isolated and transcribed into first-strand cDNA with oligo(dT) (Takara Bio) as previously described (23). The expression levels of inflammatory cytokines IL-1β, IL-6, TNF-α, and IFN-γ were quantified via real-time PCR by using a SYBR Premix Ex Taq kit (Takara Bio) in accordance with the manufacturer’s instructions. The total amount of mRNA was normalized to endogenous β-actin mRNA. Real-time PCR was performed as previously described (26). In all cases, each PCR trial was performed with triplicate samples and repeated at least three times. The primers used in the analysis are listed in Supplemental Table I.

Subcellular localization and induced secretion of TnHMGB1 were examined in human embryonic kidney 293T (HEK293T) cells. For subcellular localization, cells were seeded into multiwell plates (Corning) and cultured in DMEM supplemented with 10% FBS at 37°C in 5% CO₂ to allow growth into 70–90% confluence. Subsequently, cells were transiently transfected with pDsRED-C1-TnHMGB1 plasmid by using FuGENE HD (Promega) transfection reagent in accordance with the manufacturer’s protocol (23). At 24 h posttransfection, cells were fixed in 4% paraformaldehyde for 10 min and stained with DAPI at 37°C for 5 min. Red fluorescence images were obtained using a laser scanning confocal microscope (LSM 510; Carl Zeiss) with a ×63 lens. For induced secretion, Tetraodon leukocytes were freshly isolated from the head kidney, and spleen tissues by Lympholyte-H and -M (vol ratio, 3:1; density, 1.081) (Sango), and the interlayer of cell suspensions was collected after centrifugation for 25 min at 2500 rpm (27). Afterward, the separated leukocytes were stimulated with 1 μM CpG-SGIV-ODN6 (Table I) for 8 h, and culture supernatants were collected after ultrafiltration for Western blot analysis.

Ab against TnHMGB1 was prepared using a novel Ag epitope-based protocol (27). Briefly, the antigenic epitopes of TnHMGB1 were predicted through online software ABCPred (http://www.imtech.res.in/raghava/abcpred/) and BepiPred (http://www.cbs.dtu.dk/services/BepiPred/). The tertiary structure was modeled by the SWISS-MODEL program to detect whether the predicted epitopes were exposed on the surface of the protein, and two typical epitopes (MGKDPKKPRGKMSC and TAADDKQPYEKKAC) were used as alternative Ags to prepare the specific Abs against the TnHMGB1 protein. The epitope peptides were synthesized by InvivoGen. New Zealand White rabbits, aged 4–6 wk old and weighing 2.0–2.5 kg, were immunized with the synthetic peptides (28). Antiserum was collected after the last immunization, and Ab was affinity purified into IgG isotype by a protein A–agarose column (Qiagen) (29). Western blot analysis was performed to evaluate the specificity of the Ab. The Ab against TnSCARA5 was prepared in our previous research with purified TnexSCARA5 protein (4).

Small interfering RNA (siRNA)–encoding lentiviruses (LVs) against TnHMGB1 and TnSCARA5 were produced using a previously described protocol (23, 27). Briefly, DNA oligonucleotides encoding for the selected short hairpin RNAs containing siRNAs or the nonrelated scrambled siRNA were synthesized and constructed into the pSUPER vector downstream of the H1 promoter (30, 31). The generated constructs or control pSUPER, along with pCMV-Tag2B-TnHMGB1 or pcDNA6-TnSCARA5, were cotransfected into HEK293T cells to examine the efficacy of siRNAs via real-time PCR and Western blot analysis. The U6 promoter cassette in lentiviral vector pLB was replaced by the H1-siRNA cassette excised from the siRNA constructs with the highest efficacy to produce pLB-TnHMGB1 and pLB-TnSCARA5 lentiviral vectors. These two vectors were cotransfected with pCMV-dR8.2 and pCMV-VSVG packaging vectors into HEK293T cells by using FuGENE HD (Promega). Lentiviral supernatant was harvested three times within 48–96 h posttransfection and concentrated via ultracentrifugation at 25,000 rpm for 90 min. Viral titers were determined by transduction and FACS analysis of GFP expression in HEK293T cells. After injecting Tetraodon with LVs (5 × 10⁵ transducing units [TU]/fish) once every 24 h for three times, the silencing efficacy of the constructed LVs was examined in PBLs and head kidney leukocytes (HKLs) through real-time PCR.

The prokaryotic expression vectors for wild-type and mutant TnHMGB1 proteins were separately transformed into E. coli Rosetta (DE3) cells. The positive colonies were inoculated into 100 ml of Luria–Bertani medium containing chloramphenicol (100 μg/ml) and kanamycin (50 μg/ml). The inoculated Luria–Bertani medium was incubated in a shaking incubator at 37°C and 250 rpm until the OD₆₀₀ value reached 0.6. Subsequently, isopropyl β-d-thiogalactoside was added to a final concentration of 1 mM and shaken at 100 rpm and 20°C overnight. The proteins were expressed in soluble forms and determined by 10% SDS-PAGE. The recombinant TnHMGB1 proteins were purified by Ni-NTA agarose affinity chromatography in accordance with the QIAexpressionist manual (Qiagen). The concentration of the purified proteins was measured using the Bradford method (29). To prepare oxidized TnHMGB1 or reduced TnHMGB1, wild-type TnHMGB1 was exposed to 100 mM H₂O₂ for 1 h or 5 mM DTT for 1 h before adding it into cells for interaction analysis (57).

NF-κB activation was examined by luciferase reporter assays in HEK293T cells and zebrafish embryos as previously described (26). For HEK293T cells, plasmids in different combinations (pcDNA6, pcDNA6-TnTLR9, and pcDNA6-TnTLR9 plus pcDNA6-TnSCARA5), as well as NF-κB luciferase reporter construct (200 ng/ml; Clontech) and pGL-TK Renilla luciferase internal control (20 ng/ml; Clontech), were transfected into cells. After 24 h transfection, CpG-SGIV-ODN6 (5 μg/ml) was added into the supernatant for stimulation for 5 h to evaluate the involvement of TnSCARA5 in CpG-ODN/TnHMGB1–mediated inflammation. For zebrafish embryos, different concentrations (0.25, 0.5, 0.75, and 1.0 μg/μl) of CpG-SGIV-ODN6 or pcDNA6-TnSCARA5 plasmids (50, 75, 100, and 125 μg/μl) under 1.0 μg/μl CpG-SGIV-ODN6 stimulation added with NF-κB and pGL-TK were microinjected into one-cell stage embryos. After 24 h stimulation or postmicroinjection, the firefly and Renilla luciferase activities of the HEK293T cells and embryos per group were measured using a Dual-Luciferase reporter assay system (Promega). The relative luciferase activity unit was determined as the ratio of firefly luciferase activity divided by the Renilla luciferase activity. Each trial was repeated at least three times.

The association of TnSCARA5 with TnHMGB1 was examined by quantitative ELISA (qELISA) and coimmunoprecipitation (CoIP). For ELISA, various concentrations (8–256 nM) of recombinant TnHMGB1 proteins were coated onto microtiter wells overnight at 4°C. After blocking with 2% BSA for 2 h and washing three times, the wells were incubated with different concentrations (36 and 72 nM) of TnSCARA5 for 2 h at 37°C, followed by incubation with rabbit anti-TnSCARA5 Ab for 2 h at 37°C and HRP-conjugated goat anti-rabbit IgG for 30 min. Color was developed by tetramethylbenzidine and H₂O₂ as substrates, and OD₄₅₀ was read by a Synergy H1 hybrid reader (BioTek) (29). For CoIP, HEK293T cells were transfected with pcDNA6-TnSCARA5 fused with myc and His tag. After 24 h transfection, the cell supernatant was changed to serum-free buffer, and recombinant wild-type, mutant TnHMGB1 proteins or TnHMGB1 with different redox forms were separately added for incubation for 20 min. Afterward, cells were treated with 450 μl of lysis buffer (1% Triton X-100, 150 mM NaCl, 1 mM EDTA, 0.1% SDS, 1% sodium deoxycholate, 50 mM Tris-HCl [pH 7.4]) with protease inhibitor mixture (Roche) for 30 min at 4°C. After centrifugation for 10 min at 14,000 rpm, the supernatant was collected and incubated at 4°C overnight with mouse anti-myc Ab (Abcam). Up to 50 μl of protein G–agarose bead (Sigma-Aldrich) was then added with shaking for 4 h at 4°C and washed three times with cold PBS. The precipitants were denatured in loading buffer for detection on 12% SDS-PAGE and analysis by Western blot.

The association of CpG-SGIV-ODN6 or the CpG-SGIV-ODN6-TnHMGB1 complex to TnSCARA5, as well as the association of CpG-SGIV-ODN6 to TnHMGB1, was estimated by a streptavidin pulldown assay. Biotin-labeled CpG-SGIV-ODN6 (5 μg) with a phosphorothioate or phosphodiester backbone was immobilized onto 10 μl of streptavidin Sepharose beads (Cell Signaling Technology) and then incubated with the recombinant TnHMGB1 protein or/and TnSCARA5 lysate from HEK293 cells transfected with the TnSCARA5 plasmid pcDNA6-TnSCARA5 for 16 h at 4°C. After washing three times with lysis buffer, the precipitates were analyzed by Western blot.

TnSCARA5-mediated internalization of TnHMGB1 was determined in Tetraodon leukocytes by FACS and ELISA and evaluated by the induced trafficking of TnSCARA5 under TnHMGB1 stimulation in HEK293 cells. For FACS assay, leukocytes (10⁶ cells/ml) were preincubated with anti-TnSCARA5 Ab and normal rabbit IgG as control for 2 h at 28°C. After dealing with cytochalasin B (80 μg/ml; Sigma-Aldrich) for 1 h, the cells were treated with various concentrations (20, 30, and 40 μg) of FITC-conjugated TnHMGB1 protein for 30 min. After washing five times with ice-cold PBS to remove the unbound protein, the internalization signals were determined by a FACScan flow cytometer (BD Biosciences) with emission at 488 nm. For each sample, at least 10,000 individual cells were recorded using a dot plot combination of low-angle forward-scattered and right-angle side-scattered laser light (25). For ELISA, leukocytes were blocked with or without anti-TnSCARA5 Ab for 2 h. After incubating with the recombinant TnHMGB1 protein, the supernatant was collected at different time points and then coated onto microtiter wells (1:10 dilution) overnight at 4°C. OD₄₅₀ was read as previously described (28). For trafficking analysis, the HEK293 cells transfected with pcDNA6-TnSCARA5 were stimulated with TnHMGB1 for 30 or 60 min and detected via immunofluorescence staining as previously described (22). Briefly, cells were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100 for 15 min, and blocked with 5% goat serum for 1 h at room temperature. After incubating with rabbit anti-TnSCARA5 Ab (1:400 dilution) at 4°C overnight and washing three times with PBS, the cells were incubated with DyLight 549–conjugated goat anti-rabbit IgG for 1 h and stained with 0.1% DAPI (Invitrogen) for 3–5 min at room temperature. Fluorescence signals were observed under a confocal laser scanning microscope (LSM 510; Carl Zeiss) with a ×63 lens.

Colocalization of TnHMGB1 and TnSCARA5 with lysosome was determined in Tetraodon leukocytes and HEK293 cells. Tetraodon leukocytes were incubated with FITC-conjugated TnHMGB1 (10 μg/ml) at 28°C for 1 h and then stained with DMEM containing 75 nM LysoTracker Red (Beyotime) at 37°C for 1 h. After being incubated with 0.1% DAPI (Invitrogen) for 3–5 min at room temperature, the leukocytes were photomicrographed under a confocal laser scanning microscope. HEK293 cells transfected with pEGFP-TnSCARA5 were fixed with 4% paraformaldehyde at room temperature for 15 min, incubated with LysoTracker Red and DAPI, and then photomicrographed as described above.

A functional blockade assay was performed on mouse leukocytes to evaluate whether SCARA5 also plays an inhibitory role in CpG-ODN–induced inflammation in mammals. Leukocytes were freshly isolated from the spleen by Lympholyte-M (Sango), and the interlayer of cell suspensions was collected after centrifugation at 2200 rpm for 25 min. After blocking with different concentrations (5, 10, 15, and 20 μg/ml) of anti-mouse SCARA5 Ab and stimulation with CpG-SGIV-ODN6 (3 μg/ml), the mRNA levels of inflammatory cytokines (IL-1β, IL-6, IFN-γ, and TNF-α) in leukocytes were quantified by real-time PCR as previously described.

All experiments were replicated at least three times. All data are expressed as means ± SD. Statistical evaluations of the differences between the values of experimental groups were conducted using multiple Student t tests. A p value <0.05 was considered statistically significant.

With human HMGB1 and TLR9 sequences as queries, two homologous genes (TnHMGB1 and TnTLR9) were cloned from a model pufferfish (T. nigroviridis) through computational searching of the Tetraodon genome/EST databases. TnHMGB1 and TnTLR9 genes share overall conserved gene organization and genome synteny to their mammalian counterparts. TnHMGB1 cDNA consists of 649 bp with a 600-bp open reading frame that encodes 199 aa (Supplemental Fig. 1A; GenBank accession no. KU214880). TnHMGB1 protein is predicted to contain three conserved cysteines (Cys²², Cys⁴⁴, and Cys¹⁰⁶), an aspartic acid and glutamic acid–rich T box, and two DNA-binding domains (A and B boxes), all of which are functionally important in mammalian HMGB1 molecules. Multiple alignment analysis showed that TnHMGB1 shares high amino acid identities (74%) with its mammalian homologs (Supplemental Fig. 1A). TnTLR9 cDNA consists of 3175 bp with a 3123-bp open reading frame that encodes 1040 aa. TnTLR9 protein contains typical TLR structural hallmarks completely conserved throughout different species, including several leucine-rich repeat structures, a transmembrane domain, and a Toll/IL-1–resistance intracellular region (Supplemental Fig. 2A; GenBank accession no. KU214881). Phylogenetic analysis showed that TnHMGB1 and TnTLR9 are independently clustered with their homologs with a high bootstrap probability (Supplemental Figs. 1B, 2B). All of these observations suggest that TnHMGB1 and TnTLR9 are highly conserved molecules from fish to mammals throughout the vertebrate evolution. Additionally, TnSCARA5 was also identified as a highly conserved molecule with typical structural hallmarks of the SCARA5 family in our previous study (GenBank accession no. AFX69239.1) (4). Thus, fish is an attractive model for the functional studies of HMGB1, SCARA5, and TLR9 as a whole.

As an endogenous inflammatory factor, HMGB1 can be translocated from nucleus to cytoplasm and secrete from the cells under stimulation with PAMPs, such as LPS and CpG-ODNs in mammals. To investigate whether TnHMGB1 has such a feature, an induced secretion assay was performed by stimulating Tetraodon leukocytes with CpG-ODNs. For this purpose, we synthesized six viral genomic DNA fragments with typical CpG stimuli motifs derived from the SGIV genome (CpG-SGIV-ODN1–6) (Table I). The separated Tetraodon leukocytes were stimulated with these CpG-SGIV-ODNs, and the levels of inflammatory cytokines (IL-1β and IFN-γ) were detected by real-time PCR. Among the six CpG-SGIV-ODNs examined, CpG-SGIV-ODN6 has the strongest stimulatory effect (Fig. 1A). This CpG-SGIV-ODN6 fragment was then used in the subsequent experiments. HEK293T cells were transfected with pDsRed-TnHMGB1, and the subcellular localization of TnHMGB1 was detected by the DsRed fusion protein with red fluorescence. Results showed that TnHMGB1 is distributed in the cell nucleus without stimulation. After stimulation with CpG-SGIV-ODN6 for 8 h, TnHMGB1 migrated into the cytoplasm gradually (Fig. 1D) and secreted into the cultural supernatant as detected by the Western blot analysis (Fig. 1B, 1C).

The induced secretion of TnHMGB1 by CpG-SGIV-ODN6 implies its involvement in CpG-ODN–induced inflammation. Emerging studies have suggested that HMGB1 is an inflammatory accessory molecule that confers stimulatory function associated with various proinflammatory stimulants, such as LPS, IL-1β, Pam3CSK4, and CpG-ODN (14). HMGB1 has also been found as a CpG-ODN-binding protein that initiates TLR9 signaling pathway (20). In the present study, we explored whether TnHMGB1 can act in synergy with CpG-ODN to augment the inflammatory responses. We initially demonstrated the strong ability of TnHMGB1 protein to interact with CpG-SGIV-ODN6 through a streptavidin pulldown assay (see Fig. 9). Leukocytes isolated from Tetraodon spleen and kidney tissues were treated with TnHMGB1 and/or CpG-SGIV-ODN6 in different combinations (TnHMGB1 or CpG-SGIV-ODN6 alone, and TnHMGB1 plus CpG-SGIV-ODN6), and inflammatory cytokines (IL-1β, IL-6, TNF-α, and IFN-γ) were examined after stimulation for 4 h. The expression levels of inflammatory cytokines were significantly (p < 0.01) increased (>5-fold increase for IFN-γ and TNF-α, and >10-fold increase for IL-1β and IL-6) in the TnHMGB1 and CpG-SGIV-ODN6 combined-stimulating group compared with that in the TnHMGB1 or CpG-SGIV-ODN6 mono-stimulating group (Fig. 2). Thus, TnHMGB1 dramatically augmented the CpG-ODN–induced inflammatory responses.

For further confirmation, an in vivo TnHMGB1 knockdown assay was performed by using an LV-based siRNA delivery system. Three siRNAs targeting different TnHMGB1 regions were selected from an online design program. Of the three generated constructs, pSUPER-TnHMGB1–3 is the most effective (80%) in inducing TnHMGB1 mRNA and protein degradation in HKE293T cells (Fig. 3B). This construct was used to produce the recombinant LV. The generated lentivirus (LV-TnHMGB1) reached a titer of ∼1 × 10⁴ TU/μl by flow cytometry examination (Fig. 3A) and showed a highly infectious efficacy as determined by most GFP⁺ HEK293T cells (Fig. 3C). To detect the in vivo interference efficiency for TnHMGB1, the LV was administered into Tetraodon for 3 d, and real-time PCR was performed to test the silencing of TnHMGB1 in PBLs and HKLs. The expression of TnSCARA5 was detected to exclude cross-reaction. The results showed that TnHMGB1 expression can be sharply downregulated by 68 and 75% in PBLs and HKLs, respectively (Fig. 3D). Subsequently, CpG-SGIV-ODN6 was administrated into the TnHMGB1-depleted fish, and inflammatory cytokines (IL-1β and IFN-γ) were examined in leukocytes, as well as in spleen and kidney tissues. As expected, TnHMGB1 depletion dramatically (p < 0.01) inhibited the expression of inflammatory cytokines in response to CpG-SGIV-ODN6 stimulation (Fig. 3E), which demonstrates the synergy of TnHMGB1 in CpG-ODN–induced inflammation.

To evaluate whether TnSCARA5 is involved in TnHMGB1/CpG-ODN–mediated inflammatory responses, an induced expression assay of TnSCARA5 in response to TnHMGB1 stimulation was initially performed. Experimental fish were injected with TnHMGB1 protein for 18 h, and the expression of TnSCARA5 was detected by real-time PCR. The results showed that the transcripts of TnSCARA5 were significantly (p < 0.01) upregulated in most tissues examined, with an 8- to 12-fold increase in intestine, kidney, brain, and skin, as well as a 2- to 3-fold increase in gill and muscle (Fig. 4A). Additionally, the expression of TnSCARA5 was also increased in kidney and spleen leukocytes in response to TnHMGB1 stimulation in a dose-dependent manner (Fig. 4B). These observations provide initial insights into the involvement of TnSCARA5 in TnHMGB1/CpG-ODN–mediated inflammation.

The involvement of TnSCARA5 in TnHMGB1/CpG-ODN–induced inflammation was determined in HEK293T cells and zebrafish embryos. No significant NF-κB activation was observed in HEK293T cells in response to CpG-SGIV-ODN6 stimulation because of the low TLR9 expression in this cell line. With the exotic expression of TnTLR9 in HEK293T cells, the NF-κB activation was significantly (p < 0.05) upregulated in cells under stimulation. However, with the co-overexpression of TnTLR9 and TnSCARA5 for 24 h, the NF-κB activation was significantly (p < 0.05) reduced to the basal level after stimulation for 5 h (Fig. 5A). The zebrafish embryo model showed that the administration of different concentrations (0.25, 0.5, 0.75, and 1 μg/μl) of CpG-SGIV-ODN6 induced NF-κB luciferase activation in a dose-dependent manner (>4- to 10-fold increase) (Fig. 5B). Nevertheless, coinjection of CpG-SGIV-ODN6 (1 μg/μl) with different concentrations (50, 75, 100, and 125 ng/μl) of pCDNA6-TnSCARA5 significantly (p < 0.01) inhibited the NF-κB activation in a dose-dependent manner (>3- to 5-fold decrease) (Fig. 5C). Correspondingly, the expression of inflammatory cytokines (IL-1β, IL-6, IFN-γ, and TNF-α) was significantly (p < 0.05) increased (>4-fold increase) in embryos stimulated with CpG-SGIV-ODN6 (0.5 and 1 μg/μl) (Fig. 5D–G); this increase was dramatically (p < 0.05) inhibited in pCDNA6-TnSCARA5 (50, 75, and 100 ng/μl)–administrated embryos in a dose-dependent manner under stimulation with CpG-SGIV-ODN6 (1 μg/μl) for 16 h (Fig. 5H, 5I).

Furthermore, the function of TnSCARA5 in TnHMGB1/CpG-ODN–mediated inflammation was examined by an in vivo knockdown assay. Among the five constructs (pSUPER-TnSCARA5-1–5) generated with five siRNAs targeting different regions of TnSCARA5, pSUPER-TnSCARA5-2 has the most effective silencing (63%) for TnSCARA5 (Fig. 6B). This construct was further used to produce the recombinant LV (LV-TnSCARA5) with a titer of ∼10⁴ TU/μl (Fig. 6A) and high infectious efficacy in HEK293T cells (Fig. 6C). An in vivo interference assay showed that the TnSCARA5 expression declined by 64 and 60% in PBLs and HKLs, respectively, after the fish received the LV-TnSCARA5 for 3 d (Fig. 6D). Accordingly, the expression of inflammatory cytokines (IL-1β and IFN-γ) in TnSCARA5-depleted leukocytes, kidney, and spleen tissues was overall increased by 38–64% compared with that of the control group under stimulation of CpG-SGIV-ODN6 (Fig. 6E). These findings again demonstrated the inhibitory role of TnSCARA5 in HMGB1/CpG-ODN–elicited inflammation.

To determine whether TnSCARA5 acts as a recognition receptor for TnHMGB1, the association of TnSCARA5 with TnHMGB1 was examined by qELISA and CoIP. For qELISA, recombinant TnSCARA5 and TnHMGB1 proteins were purified by Ni-NTA agarose affinity chromatography (Fig. 7A, 7B), and various concentrations of these two proteins were used in different combinations. As expected, increased binding activities between TnSCARA5 and TnHMGB1 were clearly detected with the increase of TnHMGB1 (from 8 to 32 nM, or from 8 to 64 nM) to a designated concentration of TnSCARA5 at 36 or 72 nM, respectively. In contrast, no association was detected in the control groups, in which TnHMGB1 was replaced by various concentrations of nonrelated BSA proteins (Fig. 7D). These results indicated that TnSCARA5 interacts with TnHMGB1 in a dose-dependent manner and at a ratio of ∼1:1, as estimated by the saturated concentration of TnHMGB1 for the interaction with TnSCARA5 in the binding curves. For CoIP, the pcDNA6-TnSCARA5 construct was transfected into HEK293T cells, and two different concentrations (150 and 300 nM) of TnHMGB1 protein were incubated with the cells for 20 min at 24 h posttransfection. The associated TnSCARA5 and TnHMGB1 proteins were precipitated with Ab against Myc-tag fused to TnSCARA5 and detected by Western blot analysis with Ab against His-tag fused to TnHMGB1. Correspondingly, empty pcDNA6.0 construct–transfected cells incubated with the same amounts of TnHMGB1 protein were used as negative control. As expected, the interaction signals from TnSCARA5 with TnHMGB1 were clearly detected in the experimental groups but not in the controls (Fig. 7E). In agreement with the qELISA observations, the association of TnSCARA5 with TnHMGB1 was significantly enhanced with the increase of the administrated TnHMGB1 proteins (Fig. 7F).

To further determine which domain of TnHMGB1 is involved in the association of TnSCARA5 with TnHMGB1, three mutant TnHMGB1 proteins with deletion of A, B, and T boxes (named TnHMGB1-mutA, TnHMGB1-mutB, and TnHMGB1-mutT, respectively) were constructed (Fig. 7C), and the interactions of TnSCARA5 with these mutants were examined by CoIP. Minimal interaction of TnSCARA5 with TnHMGB1-mutA or TnHMGB1-mutB was detected compared with that of TnSCARA5 with wild-type TnHMGB1. Notably, the absence of T box in TnHMGB1 did not reduce but instead promoted the association of TnHMGB1 with TnSCARA5 (Fig. 7G). To confirm this observation, a competitive binding assay was performed to detect the decline of the increased binding of TnHMGB1-mutT to TnSCARA5 by replacing 50% of the TnHMGB1-mutT with wild-type TnHMGB1. As expected, the enhanced interaction in the TnHMGB1-mutT plus TnSCARA5 group was dramatically decreased in the TnHMGB1 and TnHMGB1-mutT (1:1) plus TnSCARA5 group (Fig. 7H, 7I). This result indicated that the absence of T box promoted the binding capacity of TnHMGB1 to TnSCARA5. Thus, T box in TnHMGB1 may play a negative regulatory role in the association of TnSCARA5 with TnHMGB1, and both A and B boxes may be necessary for the association.

Three functionally important cysteines exist in the A (Cys²² and Cys⁴⁴) and B boxes (Cys106). Previous studies have shown that HMGB1 can switch its redox states (from a fully reduced form to a disulfide-bonded form) by forming a reversible intracellular disulfide bond between Cys²² and Cys⁴⁴, and different redox forms exercise different functions. The three cysteines in HMGB1 can also be terminally oxidized to sulfonates (fully oxidized form) in an irreversible manner by treating the protein with H₂O₂. The disulfide-bonded form has a more compact folding polypeptide chain than does the fully reduced or fully oxidized form; hence, this form would have an increased electrophoretic mobility under nonreducing conditions (32). To evaluate which redox form was possessed by TnHMGB1, the recombinant TnHMGB1 protein pretreated with or without DTT (5 mM for 1 h) was detected in the presence or absence of 2-ME (0.5 mM for 10 min). The results of Coomassie brilliant blue staining (Fig. 8A) and Western blot analysis (Fig. 8B) clearly showed that the TnHMGB1 protein had an increased electrophoretic mobility under nonreducing conditions. This result suggests the presence of a disulfide bond in the recombinant TnHMGB1. To explore whether the interaction of TnHMGB1 with TnSCARA5 is associated with the redox state (disulfide bonded, fully reduced, or fully oxidized) of TnHMGB1, CoIP assays were carried out to determine the binding efficiency of the three forms of TnHMGB1 to TnSCARA5. The fully reduced and fully oxidized forms of TnHMGB1 were prepared by pretreating the recombinant TnHMGB1 protein (disulfide-bonded form) with DTT and H₂O₂ as previously described. Results showed that the association signal in the fully oxidized TnHMGB1-administered CoIP sample was significantly weaker (p < 0.05) than that of the two other samples administered with fully reduced TnHMGB1 and untreated TnHMGB1 with a disulfide bond. This result suggests that the fully oxidized TnHMGB1 had a significantly weaker combination with TnSCARA5 than did the two other forms of TnHMGB1. However, no significant difference in association signals was found between the fully reduced and disulfide-bonded TnHMGB1-administered samples. This result suggests similar binding capacities of the two TnHMGB1 forms to TnSCARA5 (Fig. 8C, 8D). The specific redox forms of TnHMGB1 that were used in the binding assays were detected from the CoIP complexes by their electrophoretic patterns under reducing and nonreducing conditions through Western blot analysis. As expected, TnHMGB1 pretreated with DTT (fully reduced) or H₂O₂ (fully oxidized) migrated as a major band with a lower electrophoretic mobility both under reducing and nonreducing conditions. In contrast, disulfide-bonded TnHMGB1 migrated under nonreducing conditions as a single band with an increased electrophoretic mobility (Fig. 8E) and shifted under reducing conditions to the same position as fully reduced or fully oxidized TnHMGB1 did (Fig. 8C). These results suggest that the different redox forms of TnHMGB1 produced remained specific during the binding assays.

Finally, we explored whether TnSCARA5 possesses the ability to associate with the TnHMGB1–CpG-ODN complex in addition to its interaction with the TnHMGB1 protein. For this purpose, TnSCARA5 was overexpressed in HEK293T cells and then incubated with the TnHMGB1–CpG-SGIV-ODN6 complex (preincubated TnHMGB1 protein with biotin-labeled CpG-SGIV-ODN6 at 4°C for 16 h), biotin-labeled CpG-SGIV-ODN6, and His-tagged TnHMGB1 protein (as positive controls). The TnSCARA5 was then precipitated with streptavidin beads or Abs against the Myc-tag fused to the TnSCARA5 protein. Western blot analysis showed that no detectable TnSCARA5 was precipitated in the TnHMGB1–CpG-SGIV-ODN6 pull-down assay. In contrast, obvious TnSCARA5 precipitates were detected in both of the two control assays (Supplemental Fig. 3A). Considering that the CpG-ODNs with a phosphorothioate backbone might have unspecific binding, we synthesized CpG-SGIV-ODN6 with a phosphodiester backbone and verified its binding abilities to TnHMGB1 and TnSCARA5 proteins. Similar to CpG-SGIV-ODN6 with a phosphorothioate backbone, CpG-SGIV-ODN6 with a phosphodiester backbone also showed strong binding abilities to TnHMGB1 and TnSCARA5 proteins. Meanwhile, as with the TnHMGB1–CpG-SGIV-ODN6 (with a phosphorothioate backbone) complex, the TnHMGB1–CpG-SGIV-ODN6 (with a phosphodiester backbone) complex showed no association to TnSCARA5 protein (Fig. 9). The above-mentioned observations suggest that TnSCARA5 associates with TnHMGB1 protein or CpG-ODN but does not do so with the TnHMGB1–CpG-ODN complex.

The aforementioned experiments showed that TnHMGB1 exerts a synergistic effect on CpG-ODN–induced inflammatory responses, and TnSCARA5 serves as a TnHMGB1 receptor with an inhibitory role in TnHMGB1/CpG-ODN–mediated inflammation. Therefore, we supposed that TnSCARA5 exerts its function by the clearance of TnHMGB1 through internalizing exogenous TnHMGB1 protein. To confirm this hypothesis, an internalization assay was performed on Tetraodon leukocytes. For the assay, the recombinant TnHMGB1 protein was conjuncted with FITC in advance. Internalization of FITC-TnHMGB1 was determined by FACS analysis. The performance of the TnSCARA5-mediated internalization of TnHMGB1 was examined by blocking the receptor with anti-TnSCARA5 Ab. Results show that the administration of FITC-TnHMGB1 led to a dose-dependent increase in percentages of leukocytes with fluorescence (from 4.09 to 13.18%) (Fig. 10A–C); in contrast, pretreatment with anti-TnSCARA5 Ab significantly reduced (p < 0.05) the percentages of the fluorescent cells from 13.18 to 8.25% (Fig. 10D). To exclude the adsorption, pretreatment with cytochalasin B (an actin inhibitor that can suppress receptor-mediated endocytosis) was set as control. As expected, the number of fluorescent leukocytes rarely increased in the presence of cytochalasin B (Fig. 10E, 10F). For further confirmation, TnSCARA5-mediated internalization of exogenous TnHMGB1 was determined by the decreased level of extracellular TnHMGB1 at different time points. ELISA results showed that the amount of the recombinant TnHMGB1 protein in the supernatant of the cultural leukocytes decreased after 30 min; however, the TnHMGB1 protein in the control group pretreated with anti-TnSCARA5 Ab remained unchanged (Fig. 11A).

In an attempt to explore whether the inhibitory function of TnSCARA5 is associated with the clearance of TnHMGB1, colocalization of TnHMGB1 with lysosome was detected in Tetraodon leukocytes. As shown in Fig. 11B, FITC-TnHMGB1 was taken up by the leukocytes and colocalized with lysosome in the cytoplasm. Meanwhile, the kinetic subcellular distribution of TnSCARA5 was tracked after TnHMGB1 stimulation. In this step, a Myc-tagged TnSCARA5 construct (pcDNA6-TnSCARA5) was transfected into HEK293 cells. After stimulation with TnHMGB1 (300 nM), the localization signal of TnSCARA5 was detected by immunofluorescence staining with secondary Ab against primary anti-TnSCARA5 Ab with red fluorescence. The result clearly demonstrated that TnSCARA5 was transferred to the surface membrane after stimulation for 30 min and returned to the cytoplasm after 1 h stimulation (Fig. 12A). Moreover, the EGFP-labeled TnSCARA5 in HEK293 cells (transfected with pEGFP-TnSCARA5) showed that the TnSCARA5 can be colocalized with the lysosome tracker (LysoTracker Red) after returning to the cytoplasm (Fig. 12B). These findings showed a kinetic change in the TnSCARA5 subcellular localization from the cytoplasm to the cell surface membrane and return to the lysosomes in the cytoplasm. Hence, TnSCARA5 can be recruited onto the cell membrane upon TnHMGB1 stimulation, which contributes to the clearance of TnHMGB1 by the receptor-mediated internalization of this protein into lysosome.

SCARA5 is an evolutionarily conservative molecule that is present in mammals, birds, and fish. Thus, whether SCARA5 also plays an inhibitory role in CpG-ODN–induced inflammation in mammals was determined. For this, nonactivated splenic leukocytes were freshly isolated from mouse, stimulated with CpG-SGIV-ODN6 (3 μg/ml), and collected at 2, 4, 6, 8, and 10 h after stimulation. Results showed that the expression of inflammatory cytokines (IL-1β, IL-6, TNF-α, and IFN-γ) in the stimulated splenic leukocytes peaked at 6 h (Supplemental Fig. 3B). Then, the expression of these cytokines in response to the stimulation of different concentrations of CpG-SGIV-ODN6 (1–5 μg/ml) was determined at 6 h. CpG-SGIV-ODN6 induced the expression of the inflammatory cytokines in a dose-dependent manner (Supplemental Fig. 3C). Ab blockade assays showed that the expression of IL-1β, IL-6, TNF-α, and IFN-γ in splenic leukocytes treated with different concentrations of anti-mouse SCARA5 Ab (5–20 μg/ml) increased by 26–41% after CpG-SGIV-ODN6 stimulation compared with that of the control cells, which received no Ab treatment (Fig. 13A). Similarly, the expression of inflammatory cytokines increased by 20–70% after splenic leukocytes were treated with anti-mouse SCARA5 Ab (15 μg/ml) and stimulated by different concentrations of CpG-SGIV-ODN6 (1–4 μg/ml) (Fig. 13B). These results suggest that SCARA5 also plays an inhibitory role in CpG-ODN–induced inflammation in mammals. However, the detailed mechanisms underlying SCARA5-mediated regulation need further investigation.

The activation of innate immunity initiated by exogenous PAMPs or endogenous DAMPs through pattern recognition receptors is essential for the establishment of host defense against infections. However, excessive activation results in both infectious and sterile inflammatory diseases and autoimmune disorders, such as sepsis, systemic inflammatory response syndrome (SIRS), systemic lupus erythematosus, ulcerative colitis, and rheumatoid arthritis (29, 33, 34). The LPS and unmethylated CpG-ODNs derived from a range of bacterial and viral genomes and even mitochondrial DNA are the two major stimulators of the occurrence of sepsis, SIRS, and autoimmune diseases, such as systemic lupus erythematosus and rheumatoid arthritis (35, 36). LPS induces septic shock through TLR4 in mammals by associating with CD14, LPS-binding protein, and myeloid differentiation protein 2, because TLR4-deficient (TLR4^−/−) mice were found to be tolerant of endotoxin lethality (37). However, the CpG-ODNs induce SIRS by the TLR9 located in the endosomes of macrophages, monocytes, and dendritic cells through the MyD88-dependent NF-κB signaling pathway (38). Mice challenged with excessive amounts of bacterial CpG-ODN were shown to die within 3 d because of the release of abundant proinflammatory cytokines, such as IL-1β, IL-6, and TNF-α (35). Thus, the burst of these cytokines has been used as an indicator for SIRS, which is an inflammatory reaction closely associated with pathogen infections or intense tissue damage (39). Moreover, mitochondrial DNA with unmethylated CpG motifs also activates the TLR9 signaling pathway and contributes to posttraumatic SIRS (36, 40, 41).

Emerging evidence has shown that HMGB1 endorses critical functions in the pathogenesis of both sterile and infectious inflammation outside the cells. In sterile injury, damage to cellular integrity results in a nearly instantaneous, passive release of HMGB1. For instance, HMGB1 release was initiated in ischemia–reperfusion injury within 1 h and remained elevated for 24 h. The protein served as an early mediator that in turn activated TNF-α and other cytokines (42). However, HMGB1 acts as a potent late mediator of infectious inflammation, such as sepsis, SIRS, and other disease models, in which HMGB1 can be secreted by activated macrophages and monocytes exposed to pathogens (12). The role of HMGB1 in sepsis has been extensively investigated. The results showed that HMGB1 was released by macrophages for >8 h and maintained at peak levels from 16 to 32 h after LPS exposure in mouse model, which is different from the TNF-α and IL-1β releases within minutes. The administration of anti–TNF-α Abs did not improve survival, but the administration of neutralizing anti-HMGB1 Abs or recombinant HMGB1 A box protein significantly attenuates endotoxin lethality in mice (12, 43). The recombinant A box protein is a competitive antagonist of HMGB1 that suppresses HMGB1-promoted cellular activation, which further highlights the importance of HMGB1 in sepsis (44). In contrast, the role of HMGB1 in SIRS induced by CpG-ODNs is limited, although HMGB1 has been shown to exhibit great importance in the CpG-ODN-TLR9 signaling pathway (45). Under CpG-ODN stimulation, HMGB1 shuttles in and out of immune cells and regulates inflammatory responses to CpG-DNA. Extracellular HMGB1 executes different roles depending on its redox form. Reduced cysteines render HMGB1 a chemoattractant that promotes the recruitment of inflammatory cells through CXCR4, whereas a disulfide bond renders it a proinflammatory cytokine that accelerates the delivery of CpG-ODNs to its receptor TLR9 through RAGE. Whether using the CpG-ODN-HMGB1 complex or CpG-ODN uptake, RAGE is important in CpG/HMGB1-induced inflammatory responses (46, 47). Intracellular HMGB1 specifically associates with TLR9 in the endoplasmic reticulum–Golgi intermediate compartment before and during the association of CpG-ODN to TLR9. Lack of the intracellular TLR9-associated HMGB1 can be compensated by extracellular HMGB1. Additionally, some earlier investigations suggested that HMGB1 can directly stimulate monocytes to release TNF-α and other proinflammatory cytokines in the late period of inflammation (12, 14). However, recent studies have shown that highly purified HMGB1 cannot induce detectable proinflammatory cytokines on its own. Instead, HMGB1 associates with proinflammatory factors to intensify the inflammatory responses (14, 15). Therefore, HMGB1 may act as a major chaperone mediator in synergy with LPS and CpG-ODNs to promote inflammation in sepsis, SIRS, and various autoimmune disorders.

Animal models are powerful tools for understanding the major immunological events and disease mechanisms difficult to address in humans. Aside from the classical mouse model, fish models, such as zebrafish and pufferfish, have become increasingly attractive in studies of comparative immunology and in modeling human diseases, such as T cell acute lymphoblastic leukemia, diabetes mellitus, Alagille syndrome, alcoholic liver disorder, and hepatocellular carcinoma; fish are suitable models because they exhibit precise and conservative innate immunity during vertebrate evolution (48–50). In the present study, we again showed the advantages of fish models in uncovering the involvement of HMGB1 in CpG-ODN–induced inflammation and the regulatory mechanisms underlying the inflammatory reactions. Structurally, the HMGB1 identified from the pufferfish T. nigroviridis (TnHMGB1) is highly homologous to its human counterpart, as evidenced by a high amino acid identity (>74%) and conserved functional domains, including the A box, B box, T box, and Cys residues (10). TnHMGB1 also distributes in the nucleus of the resting cells and can be secreted out of the cells under stimulation of the proinflammatory stimulant CpG-ODN. By stimulating Tetraodon leukocytes with TnHMGB1 and CpG-ODN in different combinations, TnHMGB1 was demonstrated to exhibit a strong synergistic effect on CpG-ODN–induced inflammation. The expression of inflammatory cytokines was dramatically augmented in TnHMGB1 and CpG-ODN combined groups. Furthermore, TnHMGB1-deficient leukocytes exhibited impaired responsiveness to CpG-ODN stimulation, which support the synergy of TnHMGB1 with CpG-ODN stimulant. Additionally, we also found that the inflammatory effect of TnHMGB1 with CpG-ODN is dependent on TLR9, because no detectable inflammatory responses can be induced in HEK293 cells with low expression of TLR9. However, such behavior can be significantly improved by the exogenous expression of TnTLR9. Thus, TnHMGB1 clearly acts as an inflammatory accessory mediator that confers stimulatory function associated with CpG-ODN through TLR9 signaling pathway, which should be highly conserved between fish and mammals. These attributes make fish an attractive model to explore the CpG-ODN/HMGB1–induced inflammatory reactions and relevant disorders, such as SIRS.

As shown by the data, the major recognition receptors for extracellular HMGB1 were identified to be RAGE, TLR2, and TLR4. These receptors are physically potent in activating cellular growth, differentiation, migration, and inflammation in innate immunity through MyD88-dependent NF-κB and MAPK signaling pathways (22, 51). However, the negative regulatory pathways and mechanisms for HMGB1-mediated biological reactions are poorly understood. In the present study, we demonstrated that SCARA5 is an HMGB1 receptor that is negatively involved in HMGB1/CpG-ODN–elicited inflammation. This finding adds a new member to the HMGB1 receptor family and provides new insights into the regulatory role of SCARA5 in innate immunity. A number of experimental lines support this conclusion. Functionally, the overexpression of TnSCARA5 in HEK293T cells and in zebrafish embryos significantly inhibited the NF-κB activation and inflammatory cytokine expressions in response to TnHMGB1/CpG-ODN stimulation. However, the depletion of TnSCARA5 in leukocytes, kidney, and spleen tissues extremely enhanced the inflammatory responses. By qELISA and CoIP assays, TnSCARA5 was demonstrated to exhibit a strong ability to associate with TnHMGB1 in a dose-dependent manner. By constructing mutant TnHMGB1 proteins with deletions of A, B, and T boxes, we showed that both A and B boxes were essential for the interaction of TnHMGB1 with TnSCARA5. However, the T box negatively regulated the association, because the ablation of T box promoted the interaction of the two molecules. To our knowledge, this result is the first observation showing the regulatory function of HMGB1 T box in molecular interactions. The biological significance and mechanism underlying such a process must be further clarified. Alternatively, TnSCARA5 was demonstrated to preferentially interact with TnHMGB1 but not with the TnHMGB1–CpG-ODN complex. From previous studies, we found that A and B boxes are two tandem DNA-binding domains linked by a short, flexible linker that lies in the minor groove and HMG boxes; these domains contribute to DNA bending by interacting with DNA through hydrophobic residues (52). Thus, the A and B boxes are common functional domains for the interactions of HMGB1 with DNA/CpG-ODN and with SCARA5. Hence, explaining why TnSCARA5 preferentially interacts with TnHMGB1 rather than with the TnHMGB1–CpG-ODN complex seems reasonable. In the latter interaction, the A and B boxes might have been spatially occupied by the CpG-ODN, which competitively excludes the association of TnHMGB1 with TnSCARA5 through the A and B boxes simultaneously. However, further investigation is needed to clarify the details of this competitive interaction. The preferential binding of SCARA5 to HMGB1 instead of the HMGB1–CpG-ODN complex may be of great biological significance. As a double-edged sword, HMGB1 released earlier forms complexed with CpG-ODN as a proinflammatory stimulant for TLR9 to initiate inflammation, which is essential for the activation of host innate immunity. However, HMGB1-released later forms complex with CpG-ODN and other stimulants to boost excessive inflammation, which frequently result in inflammatory disorders. Thus, the preference of SCARA5 for HMGB1 might be a later event in the inflammatory reactions that occur to prevent a second inflammatory burst by the clearance of HMGB1 chaperone molecule without disturbing the early activation of innate immunity by HMGB1–CpG-ODN stimulation. This hypothesis was supported by the observation that the expression of TnSCARA5 was significantly induced under TnHMGB1 stimulation. The finding also suggests the existence of a precise feedback regulatory mechanism between SCARA5 and HMGB1. Furthermore, the T box of HMGB1 might act as a switch to turn off the interaction of SCARA5 with HMGB1–CpG-ODN. Hence, a more extended T box might be formed with the binding of CpG-ODN to HMGB1. This binding might make TnSCARA5 more preferable to interact with TnHMGB1 than the TnHMGB1–CpG-ODN complex. The exact mechanisms underlying this process definitely need elucidation.

Structurally, two receptor internalization–related motifs of YXXØ (Ø represents hydrophobic amino acids) exist in the intracellular part close to the transmembrane domain of TnSCARA5. These motifs include Y-A-I-V (residues 61–64) and Y-I-L-V (residues 67–70). The YXXØ motif, characteristic of internalized receptors, is the binding site of clathrin adaptor protein 2 (53). Thus, TnSCARA5 might mediate clathrin-dependent endocytosis by interacting with clathrin/adaptor protein 2. In accordance with this notion, we found that the TnSCARA5 expressed in HEK293T cells can be recruited to the cell membrane and return to the cytoplasm in response to TnHMGB1 stimulation. This finding is consistent with that in our last study, in which LPS drives the recycling process of TnSCARA5 (4). By a further internalization assay, we demonstrated that exogenous TnHMGB1 can be internalized by Tetraodon leukocytes through TnSCARA5-mediated endocytosis in a dose-dependent manner. This observation was indicated by the dose-dependent increase of the fluorescent cells incubated with FITC-TnHMGB1 and the decreased phagocytosis of leukocytes pretreated with anti-TnSCARA5 Ab or cytochalasin B, an inhibitor for actin polymerization essential in receptor-mediated endocytosis (54, 55). After internalization with TnHMGB1 into the cytoplasm, TnSCARA5 and TnHMGB1 were clearly demonstrated to be localized in the lysosomes, strongly suggesting that the inflammatory inhibition of TnSCARA5 may be largely ascribed to its involvement in the clearance of the TnHMGB1 chaperone mediator. Alternatively, the lysosomal acidic environment is well accepted to be strictly maintained by the ATP-dependent proton pump in the lysosome membrane. With CpG motif stimulation, the acidification of CpG-DNA in lysosomes is coupled with the rapid generation of intracellular reactive oxygen species (ROS) (56). Three cysteine residues are present in HMGB1; two residues in the A box form a reversible disulfide bond under mildly oxidizing conditions and can be further oxidized into sulfonates by the increase in ROS (57–60). Functional diversity arises depending on the redox of HMGB1, as shown by the facts that reduced cysteines make HMGB1 a chemoattractant, whereas a disulfide bond makes HMGB1 a proinflammatory cytokine (57). These observations imply that a high level of ROS can be generated with the transport of CpG-ODN and HMGB1 into lysosomes by TnSCARA5. This process results in the terminal oxidization of cysteine residues into sulfonates. Such redox of TnHMGB1 might lead to the dissociation of TnHMGB1 from TnSCARA5. To explore this hypothesis, we examined the binding capacity of TnHMGB1 to TnSCARA5 under H₂O₂ treatment. As expected, the association of TnHMGB1 with TnSCARA5 was significantly decreased after the exposure of TnHMGB1 to H₂O₂. Thus, HMGB1 potentially disassociates from SCARA5 in the lysosome because of the acidic and oxidative environment under CpG-ODN stimulation. Such conditions may enable TnSCARA5 recycling from the lysosome to the surface membrane. Such notions may explain the ability of TnSCARA5 to interact with CpG-ODN besides its binding to HMGB1, which proves that ROS production by CpG-ODN stimulant is essential for the disassociation of HMGB1 from SCARA5.

In summary, the present study presents pioneering investigations of the role of SCARA5 as an HMGB1 recognition receptor crucial for the inhibition of CpG-ODN/HMGB1–induced inflammation in a pufferfish model. We previously found that TnSCARA5 is an LPS receptor negatively involved in LPS-induced inflammation. Besides this information, we demonstrated that SCARA5 is a multifunctional receptor with broad biological activities in innate immunity. Moreover, we showed that the protein is greatly important in preventing inflammation during the second inflammatory burst. This study will not only enrich current knowledge on the regulatory function of SCARA5 in innate immunity, but also serve useful in guiding the clinical treatment of diseases mediated by TnHMGB1.

The authors have no financial conflicts of interest.