β-Arrestins Negatively Regulate the Toll Pathway in Shrimp by Preventing Dorsal Translocation and Inhibiting Dorsal Transcriptional Activity*

The Toll signaling pathway plays an important role in the innate immunity of Drosophila melanogaster and mammals. The activation and termination of Toll signaling are finely regulated in these animals. Although the primary components of the Toll pathway were identified in shrimp, the functions and regulation of the pathway are seldom studied. We first demonstrated that the Toll signaling pathway plays a central role in host defense against Staphylococcus aureus by regulating expression of antimicrobial peptides in shrimp. We then found that β-arrestins negatively regulate Toll signaling in two different ways. β-Arrestins interact with the C-terminal PEST domain of Cactus through the arrestin-N domain, and Cactus interacts with the RHD domain of Dorsal via the ankyrin repeats domain, forming a heterotrimeric complex of β-arrestin·Cactus·Dorsal, with Cactus as the bridge. This complex prevents Cactus phosphorylation and degradation, as well as Dorsal translocation into the nucleus, thus inhibiting activation of the Toll signaling pathway. β-Arrestins also interact with non-phosphorylated ERK (extracellular signal-regulated protein kinase) through the arrestin-C domain to inhibit ERK phosphorylation, which affects Dorsal translocation into the nucleus and phosphorylation of Dorsal at Ser276 that impairs Dorsal transcriptional activity. Our study suggests that β-arrestins negatively regulate the Toll signaling pathway by preventing Dorsal translocation and inhibiting Dorsal phosphorylation and transcriptional activity.

The Toll signaling pathway plays an important role in the innate immunity of Drosophila melanogaster and mammals. The activation and termination of Toll signaling are finely regulated in these animals. Although the primary components of the Toll pathway were identified in shrimp, the functions and regulation of the pathway are seldom studied. We first demonstrated that the Toll signaling pathway plays a central role in host defense against Staphylococcus aureus by regulating expression of antimicrobial peptides in shrimp. We then found that ␤-arrestins negatively regulate Toll signaling in two different ways. ␤-Arrestins interact with the C-terminal PEST domain of Cactus through the arrestin-N domain, and Cactus interacts with the RHD domain of Dorsal via the ankyrin repeats domain, forming a heterotrimeric complex of ␤-arrestin⅐Cactus⅐ Dorsal, with Cactus as the bridge. This complex prevents Cactus phosphorylation and degradation, as well as Dorsal translocation into the nucleus, thus inhibiting activation of the Toll signaling pathway. ␤-Arrestins also interact with non-phosphorylated ERK (extracellular signal-regulated protein kinase) through the arrestin-C domain to inhibit ERK phosphorylation, which affects Dorsal translocation into the nucleus and phosphorylation of Dorsal at Ser 276 that impairs Dorsal transcriptional activity. Our study suggests that ␤-arrestins negatively regulate the Toll signaling pathway by preventing Dorsal translocation and inhibiting Dorsal phosphorylation and transcriptional activity.
The immune system is composed of humoral and cellular immunity. Toll and Toll-like receptor pathways play important roles in humoral and cellular immune responses in Drosophila (1, 2). Subsequent studies have revealed the central roles of mammalian Toll-like receptors in innate immunity (TLRs) 2 (3).
Gram-positive bacterial or fungal infection activates the Toll pathway, which leads to production of several antimicrobial peptides (AMPs) that kill infective pathogens in Drosophila melanogaster (4). The Toll pathway is also involved in the hematopoiesis, encapsulation, and killing of parasites (5). The core components of the Toll pathway in Drosophila include the cytokine-like ligand Spätzle, the receptor Toll, the intracellular adaptor MyD88, the kinases Tube and Pelle, and transcription factors, such as Dorsal and Dorsal-related immunity factor (Dif). After activation of the Toll pathway, Dorsal or Dif translocates into the nucleus, leading to activation of several AMP genes (2,6,7).
The Toll pathway is regulated by multiple factors at different levels. For example, Pellino, a Pelle/IL-1R-associated kinase (IRAK) interacting protein, acts as a positive regulator of the Toll pathway (8). Drosophila Pellino mutants have impaired drosomycin expression and reduced survival against Grampositive bacteria. As all Pellino proteins contain a RING domain, it is speculated that Drosophila Pellino may ubiquitinate Pelle in a similar fashion to mammalian Pellinos' polyubiquitination of IRAK1 (9). G protein-coupled receptor kinase 2 (Gprk2) was identified as a regulator of the Toll pathway (10). Gprk2 interacts with Cactus in S2 cells but is not involved in Cactus degradation; the detailed mechanism remains to be investigated. A recent study showed that specific calcineurin isoforms are also involved in Drosophila Toll immune signaling as positive regulators (11).
The strength and duration of the activation of the Toll signaling pathway must be tightly controlled, because overactivation of Toll or TLRs can be dangerous to the host. The Toll signaling pathway is negatively regulated by different molecules that target each of the key molecules in Toll signaling through various mechanisms to prevent or terminate excessive immune responses. In Drosophila, the Cactus, a homolog of IB in mammals, is the main cytoplasmic inhibitor of Dorsal (a homolog of mammalian NF-B) (12). The activation of the Toll pathway is also negatively regulated by serpins (13). Kurtz, a ␤-arrestin in Drosophila, negatively controls Toll signaling and systemic inflammation at the level of sumoylation (14). A very recent report demonstrates that Drosophila Pellino functions as a negative regulator by targeting MyD88 for ubiquitination and degradation in Toll-mediated signaling (15).
The arrestin family in mammals includes four members as follows: two visual ␤-arrestins (␤arrs), which are expressed in the rod and cone photoreceptor cells of the retina, respectively, and two ubiquitously expressed ␤-arrestins (␤arrs 1 and 2) (16). As adaptor proteins, ␤arrs are critical for mediating endocytosis of G protein-coupled receptors (GPCRs). In addition, ␤arrs function in the desensitization and endocytosis of different cell surface receptors (17). ␤arrs are also scaffold proteins, linking GPCRs to other signaling proteins, such as the Src family kinases and members of the mitogen-activated protein kinase (MAPK) cascade. ␤arrs are involved in the regulation of multiple signaling pathways (18). Previous studies suggested that the role of ␤arrs in cell signaling is much broader and that ␤arrs regulate signaling molecules by modulating phosphorylation, ubiquitination, and/or subcellular distribution of their binding partners. ␤arrs appear to interact with TRAF6 and IB␣ in the TLR signaling pathway and inhibit NF-B activity (19 -21). ␤arr1 functions as a positive regulator of CD4 ϩ T cell survival and autoimmunity (22). ␤arrs also inhibit cell apoptosis by inhibiting pro-apoptotic extracellular signal-regulated protein kinases (ERK1/2) and p38 MAPKs and anti-apoptotic Akt signaling pathways in mouse embryonic fibroblasts (23). In Drosophila, Kurtz inhibits MAPK and Toll signaling during development (24). Another study connected Kurtz activity to sumoylation and found that Kurtz negatively controls Toll signaling and systemic inflammation at the level of sumoylation in Drosophila, although the mechanism was not clear (14). Therefore, mammalian ␤arrs and Drosophila Kurtz both downregulate NF-B signaling, but in different ways. It would be interesting to determine whether ␤arrs in other invertebrates regulate Toll/NF-B signaling pathways like mammals or Drosophila.
In shrimp, the key molecules of the Toll pathway have been identified, such as Spätzle, Toll receptors, MyD88, Pelle, Cactus, and Dorsal (25)(26)(27)(28)(29)(30)(31), and reviewed briefly for the signal pathway by Li and Xiang (32). Toll-interacting protein (Tollip) from Litopenaeus vannamei was reported to negatively regulate the shrimp antimicrobial peptide gene, penaeidin-4 (PEN4) (33). ERK was identified and involved in defense against White spot syndrome virus invasion in Fenneropenaeus chinensis (34). However, the function and regulation of the Toll signaling are seldom studied. In our work, we found that the bacterial clearance, shrimp survival rate, and expression of antimicrobial peptides in kuruma shrimp Marsupenaeus japonicus were declined significantly after RNA interference (RNAi) of Dorsal, the transcription factor of the pathway, and subsequently infected with Staphylococcus aureus. These results demonstrated that the Toll pathway plays a central role in host defense against S. aureus by regulating AMPs expression. Then the regulation of the signaling was analyzed. We found that two ␤arrs (designated as Mj-␤arr1 and Mj-␤arr2) interacted with Mj-Cactus and Mj-ERK. Knockdown of ␤arrs enhanced Dorsal translocation into the nucleus to induce expression of AMPs, although knockdown of ERK showed the opposite results. The possible mechanism was further studied. ␤arrs interacted with Cactus to form a ␤arr⅐Cactus⅐Dorsal complex, which retained Dorsal in the cytoplasm of cells. ␤arrs also interacted with non-phosphor-ylated ERK to inhibit its phosphorylation and affected Dorsal phosphorylation at Ser 276 . Therefore, ␤arrs negatively regulate the Toll signaling pathway by forming a ␤arr⅐Cactus⅐Dorsal complex to prevent Dorsal translocation and impair Dorsal phosphorylation by inhibiting ERK phosphorylation. To the best of our knowledge, this is the first work to demonstrate that Toll signaling plays a central role in shrimp against Gram-positive bacteria, and ␤arrs regulate the Toll pathway in two different ways.

Experimental Procedures
Bacterial Challenge and Sample Collection-Kuruma shrimp M. japonicus (6 -8 g each) were purchased from a fish market in Jinan, Shandong Province, China, and cultured for 1 day in laboratory aquarium tanks with aerated seawater at 22°C for acclimation to the new environment. The 2 ϫ 10 7 cfu of S. aureus (Shandong University Organism Culture Collection) was injected into the abdomen of shrimp. The control group was injected with PBS (140 mM NaCl, 10 mM sodium phosphate, pH 7.4). Hemocytes, heart, hepatopancreas, gill, stomach, and intestine were collected from at least three shrimp. For the hemocytes collection, the hemolymph was extracted with a syringe preloaded with 1 ml of anticoagulant buffer (0.45 M NaCl, 10 mM KCl, 10 mM EDTA, and 10 mM HEPES, pH 7.45) and immediately centrifuged at 800 ϫ g for 8 min at 4°C, and the hemocytes were suspended in PBS. The hemocytes and other tissues were used for RNA or protein extraction. Total RNA was extracted using the TRIzol reagent (Cwbio, Beijing, China).
cDNA Cloning and Sequence Analysis-The full-length cDNA sequences of Mj-Dorsal, Mj-␤arr1, Mj-␤arr2, and Mj-ERK were obtained from hemocyte and intestine transcriptome sequencing of M. japonicus. The open reading frame of Mj-Cactus was also obtained by transcriptome sequencing. Similarity analysis was conducted using BLASTx (www.ncbi.nlm. nih.gov). The corresponding cDNA was conceptually translated, and the deduced proteins were predicted using ExPASy. The domain architecture prediction of the proteins was performed using SMART. MEGA5 was used for phylogenetic analysis.
Tissue Distribution and Expression Pattern Analysis by Reverse Transcription-PCR (RT-PCR) and Western Blotting-RT-PCR was used to assess the tissue distribution of Mj-␤arr1, Mj-␤arr2, Mj-Dorsal, Mj-ERK, and Mj-Cactus using the primers shown in Table 1. Protein samples obtained from shrimp organs were separated by 10% SDS-PAGE and then transferred onto a nitrocellulose membrane. The membrane was blocked for 1-2 h with 3% nonfat milk in Tris-buffered saline (10 mM Tris-HCl, pH 8.0, 150 mM NaCl) and incubated with 1:200 diluted antiserum against the proteins of interest (Mj-␤arrs, Mj-ERK, or Mj-Dorsal) in TBS with 3% nonfat milk for 3 h. After washing three times in TBS, alkaline phosphatase-conjugated goat anti-rabbit IgG (1:10,000 diluted in TBS) was added. After incubation with the membrane for 3 h, unbound IgG was washed away. The membrane dipped into the reaction system was visualized by 4-chloro-1-naphthol oxidation in the dark for 5 min. Mj-␤arr1, Mj-␤arr2, Mj-ERK, and Mj-Cactus antisera were prepared in our laboratory using recombinant proteins.
Antibodies recognizing the phosphorylated forms of ERK were purchased from Abcam, and NF-B p65 (Ser 276 ) antibody was purchased from ABGENT (San Diego, CA).
Quantitative Real Time PCR-The cDNA templates were diluted 50-fold in nuclease-free water for qRT-PCR analysis. SYBR Premix Ex Taq (TaKaRa, Dalian, China) was used in a real time thermal cycler (Bio-Rad) with a total volume of 10 l containing 5 l of 2ϫ Premix Ex Taq, 1 l of the 1:100 diluted cDNA, and 2 l (1 M) each of the forward primer and the reverse primer. qRT-PCR was performed with the following conditions: 94°C for 3 min, 40 cycles of 94°C for 10 s, and 60°C for 1 min, and a final dissociation protocol from 65 to 95°C. Three parallel experiments were conducted to increase the credibility of this study. We used the 2 Ϫ⌬⌬Ct method to calculate the mRNA relative expression (36). The obtained data were subjected to statistical analysis and unpaired sample t test.
RNA Interference Assay-The cDNA fragments amplified by primers Fi and Ri (Table 1) linked to the T7 promoter (Table 1) were used as templates for the synthesis of dsRNA. The cDNA fragment of GFP used for dsGFP synthesis was amplified using primers GFP-Fi and GFP-Ri (Table 1).
The assay for dsRNA synthesis was performed in accordance with previous reports (35). The dsRNA (40 g) of Mj-MEK, Mj-␤arr1, Mj-␤arr2, Mj-ERK, or Mj-Dorsal was injected into the abdominal segment of each shrimp. To enhance the RNAi effect, a second injection was performed 12 h after the first injection. The dsGFP was used as the control. The intestine was collected from the shrimp 24 h after the second injection, and total RNA was extracted and detected by RT-PCR using primers RT-F and RT-R (Table 1) to check the efficiency of RNAi. The experiments were repeated three times.
After setting up the RNAi assay, S. aureus was injected into the gene-silenced shrimp. The expression levels of AMPs and the tumor necrosis factor (TNF) superfamily (TNFSF) gene regulated by the Toll pathway (37) in hemocytes and intestine and gill, and the bacterial clearance and the survival rate (see the following method) of the shrimp were analyzed for Mj-Dorsal RNAi shrimp. The translocation of Mj-Dorsal in hemocytes and intestine, the AMPs expression in hemocytes, intestine, or gill, the bacterial clearance, and the survival rate were detected for Mj-␤arr1 and Mj-ERK RNAi shrimp. The AMPs expression was analyzed by quantitative real time reverse transcription PCR (qRT-PCR) with the AMPs primers (Table 1). ␤-Actin was used as the control. Data show the mean Ϯ S.D. from three independent repeats. The p value was calculated by Student's t test for paired samples, and significant or most significant differences were accepted at p Ͻ 0.05 or p Ͻ 0.01.
Nuclear and Cytoplasmic Protein Extraction-To confirm Mj-Dorsal and Mj-ERK translocation into the nucleus, nuclear and cytoplasmic protein extraction assays were performed. Shrimp intestine and gill were cut into pieces with small scissors, washed with PBS, and then centrifuged at 1000 ϫ g for 7 min. The precipitate was suspended in 100 l of buffer A (10 mM HEPES, pH 7.5, 10 mM KCl, 0.2 mM EDTA, 3 mM MgCl 2 , 1 mM DTT, 1 mM PMSF, 1% inhibitor mixture (Merck, Germany), 1% Nonidet P-40) and incubated on ice for 10 min. The tissue solution was then centrifuged at 1000 ϫ g for 10 min. The obtained supernatant contained cytoplasmic proteins, and the precipitate was resuspended in the buffer A, incubated for 10 min on ice to wash off the remaining cytoplasm, and then centrifuged at 1000 ϫ g for 10 min again. Buffer B (100 l) (20 mM HEPES, 400 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 1 mM PMSF, 1% mixture) was added to the precipitate and vortexed for 30 min at 4°C. The solution was centrifuged at 14,000 ϫ g for 10 min to obtain the nuclear proteins. Nuclear and cytoplasmic proteins were then subjected to Western blotting.
Bacteria Clearance and Survival Rate Assays-The shrimp (6 -8 g each) were divided into four groups as following: three groups for RNAi of target genes and one group for GFP-RNAi control. After knockdown of Mj-Dorsal by RNAi, shrimp were injected with S. aureus (2 ϫ 10 7 CFU) at 24 h after the second dsRNA injection. At 6 h after S. aureus injection, S. aureus (2 ϫ 10 7 CFU) was injected into shrimp again, and the hemolymph was extracted at 1 h after the second S. aureus injection. The hemolymph was then diluted and cultured on LB agar plates overnight at 37°C. The number of bacterial colonies was counted. The same methods were performed for Mj-␤arr1 and Mj-ERK RNAi and bacterial clearance.
Survival Rate Assay-After Mj-Dorsal was knocked down by RNAi, the shrimp were injected with S. aureus (2 ϫ 10 7 cfu). The dead shrimp were monitored and counted every day. Data show the mean Ϯ S.D. from three independent repeats. The p value was calculated by Student's t test for paired samples, and significant or most significant differences were accepted at p Ͻ 0.05 or p Ͻ 0.01, respectively.
Pulldown Assay-The recombinant proteins (30 g) were added to 20 l of charged nickel-nitrilotriacetic acid beads (for His-tagged proteins) or glutathione resin (for GST-tagged proteins) and incubated at room temperature for 2 h with slight rotation. The mixture (resin and binding proteins) was washed three times by centrifugation at 500 ϫ g for 3 min to remove the unbound proteins. The test protein, without a His tag or GST tag, was added into the mixture containing the nickel-nitrilotriacetic acid beads or glutathione resin and the tagged protein and gently rotated at room temperature for 2 h. After washing three times, the mixture was analyzed by SDS-PAGE. We also used recombinant proteins to pull down the natural proteins from shrimp. Shrimp gills were homogenized with lysis buffer (150 mM NaCl, 1.0% Nonidet P-40, 0.1% SDS, 50 mM Tris-containing protease inhibitor mixture (Abcam)) and then centrifuged at 14,000 ϫ g for 12 min. The supernatant (1000 l) was added into the GST resin with recombinant proteins and incubated at room temperature for 2 h. The resin was washed three times by centrifugation at 500 ϫ g for 3 min and then analyzed using SDS-PAGE. The proteins in the gel were transferred onto a nitrocellulose membrane, followed by blocking with 3% nonfat milk dissolved in Tris-buffered saline (10 mM Tris-HCl, 150 mM NaCl, pH 7.5) and incubated with 1:100 diluted antiserum in TBS with 3% nonfat milk for 2 h. Horseradish peroxidase goat anti-rabbit IgG (1:10,000 diluted in TBS) was then added. After incubation with the membrane for 2 h, unbound IgG was washed off. The membrane was dipped into the reaction system (9 ml of TBS, with 1 ml of 4-chlorine naphthol and 6 l of H 2 O 2 ) in the dark for 5 min to visualize the target protein.
Immunocytochemical Assays-The hemolymph was collected from three shrimp using a syringe preloaded with 1 ml of anticoagulant and then fixed by adding 4% paraformaldehyde. The hemocytes were isolated by centrifugation (700 ϫ g for 4 min at 4°C), washed with PBS, and incubated in 0.2% Triton X-100 at 37°C (5 min) twice. After blocking with 3% bovine serum albumin (BSA) for 30 min at 37°C, hemocytes were incubated overnight with anti-Mj-Dorsal serum (1:100 in blocking ␤-Arrestins Negatively Regulate the Toll Pathway in Shrimp buffer) at 4°C. After washing with PBS, the hemocytes were then incubated with 3% BSA for 10 min, after which the second antibody, goat anti-rabbit Alexa Fluor 488 (1:1,000 dilution in 3% BSA), was added. The reaction was kept in the dark for 1 h at 37°C and then washed with PBS. Hemocytes were stained with 4Ј,6-diamidino-2-phenylindole dihydrochloride (DAPI, AnaSpec Inc., San Jose, CA) for 10 min at room temperature and washed again. Hemocytes were observed under a fluorescence microscope (Olympus BX51).
Co-immunoprecipitation (Co-IP)-Sixty milligrams of CNBractivated Sepharose 4B (Amersham Biosciences) was swollen in 500 l of HCl (1 mM) and 100 l of medium, and then washed twice with 1 mM HCl. The antibodies of interested were incubated with CNBr-activated Sepharose 4B with rotation at 25°C for 1 h. The resin was washed five times with coupling buffer and then equilibrated in 0.1 M Tris-HCl, pH 8.0, at 25°C for 2 h. Then, after four rounds of alternate washing with acetic acid buffer (0.1 M sodium acetate, 0.5 M NaCl, pH 4.0) and Tris buffer (0.1 M Tris-HCl, 0.5 M NaCl, pH 8.0), the resin was equilibrated in Tris buffer. The supernatant (1000 l) from normal and challenged shrimp hemocytes, intestine, or gill was prepared and then mixed with 100 l of resin and gently rotated overnight at 4°C. After five washes with Tris buffer, the native protein was eluted using elution buffer (0.1 M glycine, pH 2.5) and neutralized with neutralization buffer (1 M Tris-HCl, pH 8.0). The purified proteins were confirmed by Western blotting.
PD98059 Inhibitor Assay in Shrimp-PD98059, a mitogenactivated protein kinase kinase MEK1/2 inhibitor, was purchased from Cell Signaling Technology (Danvers, MA). PD98059 (4 M) was injected into shrimp (8 -10 g), and the same volume of DMSO injection was used as the control. S. aureus was injected after 1 h of inhibitor injection. Subsequently, shrimp intestines were collected 1 h after bacterial injection for Western blotting. Antibodies against Mj-ERK and Mj-Dorsal were used in the analysis. The total RNAs from intestine were extracted from shrimp 6 h after S. aureus injection to detect AMPs expression.

Results
Toll Pathway Is Activated by S. aureus Challenge and Plays a Central Role in Defense against S. aureus in Shrimp-To study the function and regulation of the Toll pathway in shrimp during bacterial infection, we first ascertained the activation of the Toll pathway by translocation of Mj-Dorsal (GenBank TM accession number KU160503) and expression of AMPs after bacterial infection. The Gram-positive bacterium S. aureus was used to activate the Toll pathway. After bacterial activation of the Toll path-way in shrimp, the protein level of Mj-Dorsal increased in hemocytes, intestine, and gill (Fig. 1, A-C). Mj-Dorsal was translocated from the cytoplasm into the nucleus in hemocytes, intestine, and gill at 1 h after S. aureus challenge (Fig. 1, D and  F). The expression of AMPs, ALF-C2 (GenBank TM accession number KU160498), CruI-1 (GenBank TM accession number KU160502), and another effector TNFSF (GenBank TM accession number KU160505) but not ALF-D1 (GenBank TM accession number KU160499) regulated by the Toll pathway also, was increased significantly (Fig. 1, G, J and M). After knockdown of Mj-Dorsal by RNA interference and the subsequent challenge by S. aureus, the expression of the effectors ALF-C2, CruI-1, and TNFSF in hemocytes (Fig. 1H), intestine (Fig. 1K), and gill (Fig. 1N) was significantly decreased compared with the control (Fig. 1, I, L, and O). These results indicated that expression and translocation of Mj-Dorsal and expression of the Toll pathway effectors (AMP genes) were significantly increased in hemocytes, intestine, and gill after S. aureus infection and suggest that systemic immune responses occurred also in the intestine and gill in addition to hemocytes. Therefore, we mainly used intestine or gill in the following studies for the convenience of sample collection.
To evaluate whether the Toll pathway plays a central role in host defense against S. aureus, the bacterial clearance and survival of shrimp were conducted after shrimp were injected with dsRNA of Mj-Dorsal and subsequently infected with S. aureus. dsGFP injection followed by S. aureus infection was used as the control. The results showed that after RNAi of Mj-Dorsal, S. aureus clearance was significantly reduced compared with the control group (Fig. 1P), and the survival rate of Mj-Dorsal-RNAi shrimp infected with S. aureus was significantly lower than that of the control from day 3 to day 5 post-infection (Fig.  1Q). These data indicated that the Toll signaling may play a central role in defense against Gram-positive bacteria in shrimp.
Mj-␤arr1, Mj-␤arr2, and Mj-ERK Participate in Regulation of the Toll Pathway-The full-length cDNA sequences of Mj-␤arr1 (GenBank TM accession number KU160500), Mj-␤arr2 (GenBank TM accession number KU160501), and Mj-ERK (GenBank TM accession number KU160504) were obtained by transcriptome sequencing of M. japonicus. To analyze whether Mj-␤arr1, Mj-␤arr2, and Mj-ERK participate in regulation of the Toll signaling pathway, translocation of Mj-Dorsal and expression of the Toll pathway effectors were detected after RNAi knockdown of ␤arrs and ERK, followed by bacterial challenge. The results showed that in the ␤arr-si-

␤-Arrestins Negatively Regulate the Toll Pathway in Shrimp
APRIL 1, 2016 • VOLUME 291 • NUMBER 14 lenced shrimp, most of Mj-Dorsal was detected in the nucleus of intestine cells by Western blotting (Fig. 2, A and A1), and the expression of effector transcripts, such as ALF-C2, CruI-1, and TNFSF mRNAs, was significantly up-regulated (Fig. 2A2). The Mj-Dorsal translocation was also detected in hemocytes in ␤arr-silenced shrimp following bacterial challenge (Fig. 2, A3  and A4). The results showed that in the normal (untreated) shrimp, Mj-Dorsal could be observed in the cytoplasm and ␤-Arrestins Negatively Regulate the Toll Pathway in Shrimp nucleus of hemocytes, but most of the Mj-Dorsal was located in the cytoplasm (Fig. 2A4). In the dsGFP-injected shrimp, most of Mj-Dorsal was translocated into the nucleus, and only a small amount of Mj-Dorsal could be detected in the cytoplasm after S. aureus challenge (Fig. 2A4). After ␤arr knockdown and bacterial challenge, Mj-Dorsal was only detected in the nuclei of hemocytes (Fig. 2A4). These results suggested that ␤arrs may inhibit the expression of the Toll pathway effector genes by preventing Mj-Dorsal from translocation into the nucleus. On the contrary, most Mj-Dorsal was detected in the cytoplasm but not in the nuclei of intestine cells (Fig. 2B1) after Mj-ERK was knocked down (Fig. 2B) followed by S. aureus challenge. This result indicated that Mj-Dorsal could not be translocated into the nucleus after bacterial challenge in the Mj-ERK knockdown shrimp, and the expression of effectors was also significantly decreased (Fig. 2, B1 and B2) at 6 h after bacterial injection. In the hemocyte immunocytochemical assay, almost no Mj-Dorsal could be detected in the nucleus after Mj-ERK knockdown (Fig. 2, B3 and B4). These results suggested that Mj-ERK promoted the nuclear translocation of Mj-Dorsal and increased the expression of the AMPs. Taken together, our results suggested that Mj-␤arrs negatively regulate the Toll pathway, and Mj-ERK positively regulates the pathway.
To further confirm that Mj-␤arrs and Mj-ERK participate in regulating the Toll pathway, the bacteria clearance and survival rate of Mj-␤arr1-RNAi or Mj-ERK-RNAi shrimp were also analyzed. In the Mj-␤arr1-RNAi shrimp injected with S. aureus, bacteria clearance was enhanced, although in the Mj-ERK-RNAi shrimp bacteria clearance was reduced compared with the control group (Fig. 2, A5 and B5). The survival rate of the Mj-␤arr1-RNAi shrimp injected with S. aureus was about 55% at day 5 post-infection, which was significantly higher than that (20%) of the GFP-RNAi control group (Fig. 2A6). However, the survival rate of Mj-ERK-RNAi shrimp injected with S. aureus was about 40% at day 3 post-infection, which was significantly lower than that (75%) of dsGFP shrimp (Fig. 2B6). These data indicated that Mj-␤arr1 and Mj-ERK play important roles in the antibacterial immunity of shrimp.
Mj-␤arrs Interact with Mj-Cactus to Form a ␤arr⅐Cactus⅐ Dorsal Complex-To understand the regulatory mechanism of Mj-␤arrs in the Toll pathway, the interactions of Mj-␤arrs with Mj-Cactus and Mj-Dorsal were analyzed by Co-IP assays. We first determined the amount of Mj-Cactus after S. aureus challenge, and the results showed that the amount of Mj-Cactus was decreased from 10 to 60 min in hemocytes, intestine, and gill of the shrimp, and only small amounts of Mj-Cactus were detected in hemocytes at the 60-min post-bacterial challenge (Fig. 3A). These results suggested that Mj-Cactus was degraded after activation of Toll pathway. The antibody against Mj-␤arr1 was coupled to CNBr-activated Sepharose 4B for binding of natural Mj-Cactus and Mj-Dorsal from unchallenged and bacterially challenged shrimp. The results showed that in the unchallenged shrimp, Mj-Cactus and Mj-Dorsal could be coimmunoprecipitated with Mj-␤arr1, but in the bacterially challenged shrimp, no Mj-Cactus nor Mj-Dorsal was co-immunoprecipitated (Fig. 3, B and B1). It is known that after activation of the Toll pathway by S. aureus, Cactus is degraded to release Dorsal, which then translocates into the nucleus. Therefore, the above results suggest that Mj-␤arrs may interact with Mj-Cactus and/or Mj-Dorsal. When Mj-Cactus antibody was used for the Co-IP assay, Mj-␤arr1, Mj-␤arr2, and Mj-Dorsal were coprecipitated in the unchallenged shrimp, but only a little Mj-Dorsal was co-precipitated in the bacterially challenged shrimp, and almost no Mj-␤arrs were detected in gill (Fig. 3, C  and C1). The results suggested that Mj-Cactus was not completely degraded in gill of shrimp challenged by bacteria. From  Fig. 3A, we can see that most of Mj-Cactus was degraded at 60 min post-bacterial challenge; we used hemocytes to do the coimmunoprecipitation assay, and the results showed that Mj-␤arr1, Mj-␤arr2, and Mj-Dorsal were not co-immunoprecipitated in the bacterially challenged shrimp (Fig. 3, D and D1). These results suggested that Mj-Cactus interacted with Mj-Dorsal and Mj-␤arrs, and almost no interaction of Mj-Cactus and Mj-␤arrs was detected at the early stage of bacterial infection in S. aureus challenged shrimp, probably because the upstream signal pathway was activated, and the Mj-Cactus would be degraded to lead to the dissociation of Mj-␤arrs with Mj-Cactus. The Mj-Dorsal antibody was also used for Co-IP assays, and the results showed that Mj-␤arrs and Mj-Cactus were detected in the unchallenged shrimp, and no Mj-␤arrs and Mj-Cactus were detected in the challenged shrimp (Fig. 3, E and  E1). Taken together, our results suggested that Mj-␤arrs, Mj-Cactus, and Mj-Dorsal could form a heterotrimeric complex, and the complex would disassociate followed by Mj-Cactus degradation.
To further confirm the possibility of forming a heterotrimeric complex in vivo, the tissue distribution of Mj-␤arr1, Mj-␤arr2, Mj-Dorsal, and Mj-Cactus was detected at the mRNA and protein levels. The results showed that the four mRNAs and proteins were distributed in all tissues tested, including hemocytes, heart, hepatopancreas, gill, stomach, and intestine (Fig. 3F), suggesting that heterotrimeric complex of Mj-␤arr-Cactus-Dorsal could be formed in shrimp.
N-terminal Domain of Mj-␤arrs Interacts with the PEST Domain of Mj-Cactus-Pulldown assays were performed to confirm the above-mentioned interactions using recombinant proteins of Mj-␤arrs and Mj-Cactus. Mj-␤arr contains arrestin-N and arrestin-C domains, whereas Mj-Cactus contains an N-terminal domain (amino acid residues 1-168), ANK repeats domain (amino acid residues 169 -395), and a PEST domain (amino acid residues 396 -452) at the C terminus (Fig. 4A). The full-length ␤arrs and their two individual domains were expressed and purified. The full-length Mj-Cactus protein, the N-terminal domain-ANK repeats domain (amino acid residues 1-395), N-terminal domain (amino acid residues 1-168), ANK repeats domain (amino acid residues 169 -395), and PEST domain (amino acid residues 396 -452) were also expressed and purified. Two different pulldown assays with GST-binding resin and His-binding resin were performed.
The recombinant Mj-␤arrs and Mj-Cactus were used for GST and His pulldown assays. As shown in Fig. 4, B and C, Mj-␤arr1 or Mj-␤arr2 interacted with Mj-Cactus. To under-FIGURE 3. Mj-␤arrs interacts with Mj-Cactus and Mj-Dorsal, forming a ␤arr⅐Cactus⅐Dorsal complex. A, shrimp was injected with S. aureus, and the proteins from different tissues (hemocytes, intestine, and gill) were extracted at 10, 30, and 60 min post-injection and analyzed by Western blotting using Mj-Cactus antibody. IB, immunoblot. B, interaction of Mj-␤arr1 with Mj-Cactus and Mj-Dorsal in gill was analyzed by a co-immunoprecipitation assay using an anti-Mj-␤arr1 antibody. The antibody was coupled with CNBr-activated Sepharose 4B. The gill tissues of untreated shrimp or S. aureus-challenged shrimp were collected at 1 h after bacterial injection. The tissue extraction solution was incubated with CNBr-activated Sepharose 4B to bind the target proteins. The eluate was assayed by Western blotting. The Mj-Cactus and Mj-Dorsal could be detected in the unchallenged shrimp gill tissues, but not in the bacterially challenged shrimp. B1, interaction of Mj-␤arr2 with Mj-Cactus and Mj-Dorsal in gill was analyzed by co-immunoprecipitation using an anti-Mj-␤arr2 antibody. The antibodies were coupled with CNBr-activated Sepharose 4B to bind natural Mj-Cactus and Mj-Dorsal. C and C1, interaction of Mj-Cactus with Mj-␤arrs and Mj-Dorsal in gill was analyzed by co-immunoprecipitation using an Mj-Cactus antibody. The antibodies were coupled with CNBr-activated Sepharose 4B to bind natural Mj-␤arrs and Mj-Dorsal. D and D1, interaction of Mj-Dorsal with Mj-␤arrs and Mj-Cactus was analyzed by co-immunoprecipitation using an anti-Mj-Dorsal antibody. The antibody was coupled with CNBr-activated Sepharose 4B to bind natural Mj-Cactus and Mj-␤arr1 or Mj-␤arr2. E and E1, interaction of Mj-Cactus with Mj-␤arrs and Mj-Dorsal in hemocytes was analyzed by co-immunoprecipitation using an Mj-Cactus antibody. The antibody was coupled with CNBr-activated Sepharose 4B to bind natural Mj-␤arrs and Mj-Dorsal. F, distribution of Mj-␤arr1, Mj-␤arr2, Mj-ERK, Mj-Cactus, and Mj-Dorsal in mRNA and proteins was examined by semi-quantitative RT-PCR and Western blotting. The experiments were repeated three times.

␤-Arrestins Negatively Regulate the Toll Pathway in Shrimp
stand which domain of Mj-␤arrs was responsible for the interaction with Mj-Cactus, the N-and C-terminal domains of ␤arrs (His-Mj-␤arr1-N, His-Mj-␤arr1-C and His-Mj-␤arr2-N, His-Mj-␤arr2-C) were used in GST and His pulldown assays. The results showed that Mj-␤arrs-N, but not Mj-␤arrs-C, interacted with Mj-Cactus (Fig. 4, D-G) to further study which domain of Mj-Cactus interacted with Mj-␤arrs-N. The GSTtagged N-ANK and PEST domains of Mj-Cactus were used for pulldown assays. The results showed that The Mj-Cactus-N-ANK(1-395) could not interact with Mj -␤arrs-N (Fig. 4, H and  I); however, Mj-Cactus-PEST (Fig. 4, F and G) interacted with Mj-␤arrs-N (Fig. 4, J and K) 5B) were used for pulldown assays. As shown in Fig. 5C, Mj-Cactus could interact with Mj-Dorsal, and further study showed that Mj-Cactus interacted with the RHD domain but not the IPT domain of Mj-Dorsal (Fig. 5D). Recombinant Mj-Cactus-N-ANK, Mj-Cactus-N, Mj-Cactus-ANK, and Mj-Cactus-PEST were used for pulldown assays to analyze the interaction with RHD domain of Mj-Dorsal (Fig. 5E). The results showed that only the ANK repeats domain of Mj-Cactus interacted with the RHD domain of Mj-Dorsal.
In summary, the arrestin-N domain of Mj-␤arrs interacts with the PEST domain of Mj-Cactus. The ANK repeat domain of Mj-Cactus interacts with RHD domain of Mj-Dorsal. The ␤arr⅐Cactus⅐Dorsal complex is formed using Mj-Cactus as a bridge. To confirm the result, we used recombinant Mj-␤arr1 or Mj-␤arr2, Mj-Cactus, and Mj-Dorsal in the GST-pulldown assay, and the result showed that Mj-␤arrs, Mj-Cactus, and Mj-Dorsal could form a complex (Fig. 5F) but Mj-␤arrs could not interact with Mj-Dorsal (Fig. 5H). Fig. 2, B and B4, we noticed that ERK positively regulated Dorsal translocation into the nucleus and expression of AMPs. The possible mechanism was further studied. The interactions of Mj-␤arrs with Mj-ERK were first studied by pull down (Fig. 6A) and Co-IP (Fig. 6B) assays. Here, we used recombinant proteins to pull down the native proteins from shrimp. The results showed that Mj-ERK interacted with native Mj-␤arr1 or Mj-␤arr2 (Fig. 6A) and that Mj-␤arrs interacted with native non-phosphorylated Mj-ERK, but little native phosphorylated Mj-ERK was co-precipitated (Fig. 6B). To further confirm whether native phosphorylated Mj-ERK could interact with Mj-␤arr1 or Mj-␤arr2, phosphorylated Mj-ERK antibody was used to do the Co-IP assay. The results showed almost no native Mj-␤arr1 or Mj-␤arr2 was coprecipitated (Fig. 6C). Therefore, the phosphorylated Mj-ERK could not interact with Mj-␤arr1 or Mj-␤arr2.

Mj-␤arrs Interacts with Non-phosphorylated Mj-ERK to Decrease Phosphorylation of Mj-ERK-In
After knockdown of Mj-␤arr1 or Mj-␤arr2 (Fig. 6D), the phosphorylation level of Mj-ERK was detected by Western blotting. The results showed that the phosphorylation level of Mj-ERK was increased in Mj-␤arr1 and Mj-␤arr2 knockdown shrimp challenged with S. aureus for 1 h (Fig. 6E). These results suggested that Mj-␤arrs inhibited ERK phosphorylation by interacting with non-phosphorylated Mj-ERK.
Mj-ERK Affected the Phosphorylation of Mj-Dorsal in the Nucleus-To further study the possible mechanism of ERK in regulation of the Toll pathway, we analyzed the phosphorylation of Mj-ERK and Mj-Dorsal using antibodies to phosphorylated ERK (Abcam) or phosphorylated Mj-Dorsal. To determine whether commercial phosphorylated Dorsal antibody could recognize the phosphorylated Mj-Dorsal from shrimp, Mj-Dorsal was knocked down in shrimp and then challenged with S. aureus, and Western blot was performed to detect the Dorsal. As shown in Fig. 8A, phosphorylated Mj-Dorsal could be detected by the antibody to phosphorylated Dorsal. Then phosphorylation of Mj-Dorsal and Mj-ERK was analyzed by Western blotting. The results showed that S. aureus challenge enhanced phosphorylation of Mj-ERK and Mj-Dorsal. The phosphorylation of Mj-ERK increased rapidly at 15 min and phosphorylation at Ser 276 of Mj-Dorsal increased at 30 -60 min (Fig. 8B). After knockdown of Mj-ERK and challenge with S. aureus in shrimp, the phosphorylation of Mj-Dorsal obviously decreased (Fig. 8C). To confirm the result, Mj-MEK, which

␤-Arrestins Negatively Regulate the Toll Pathway in Shrimp
affects the Mj-ERK activity, was knocked down, and the phosphorylation of Mj-Dorsal and Mj-ERK was detected. The results showed phosphorylation of Mj-Dorsal and Mj-ERK was decreased in Mj-MEK-silenced shrimp (Fig. 8E). PD98059, a MEK1/2 inhibitor, was also used to inhibit ERK phosphorylation, and a similar result was obtained (Fig. 8F). We further detected the subcellular distribution of phosphorylated Mj-Dorsal by Western blotting. The distribution and phosphor- ylation levels of Mj-Dorsal in the nucleus decreased after PD98059 injection and challenge with S. aureus for 1 h (Fig.  8G). The expressions of effectors (MjALF-C2, CruI-1, and TNFSF) regulated by the Toll pathway in the intestine were significantly decreased after PD98059 injection and challenge with S. aureus at 6 h (Fig. 8H). Taken together, the results showed that phosphorylation of Mj-ERK induced Mj-Dorsal phosphorylation and the expression of effectors regulated by the Toll pathway. However, the interaction of Mj-␤arr with Mj-ERK inhibited the phosphorylation of Mj-ERK (Fig. 6) and subsequently reduced Mj-Dorsal phosphorylation, which inhibited Toll signaling and reduced the expression of the effectors.

Discussion
The Toll signaling pathway plays an important role in innate immunity in invertebrates (38,39). Dorsal is a transcription factor that modulates the transcriptional activity of the downstream targets. In Drosophila, the Toll signal pathway is activated by fungi and Gram-positive bacteria to regulate expression of AMPs to defend against pathogen invasion (40,41). In shrimp, the Toll pathway plays an important role in regulating AMP expression (42,43). In this report, we find that the bacterial clearance and survival rate of Mj-Dorsal-silenced shrimp infected with S. aureus had declined significantly, suggesting that the Toll pathway plays a central role in host defense against S. aureus in shrimp. Toll signaling is activated by S. aureus, resulting in Mj-Dorsal nuclear translocation and an increase in its phosphorylation level. The expression of the effectors (MjALF-C2, CruI-1, and TNFSF) is regulated by the Toll pathway. Mj-␤arrs interacted with Mj-Cactus and Mj-ERK to negatively regulate the Toll pathway.
The systemic immune response in Drosophila induced the synthesis of several families of AMPs by cells in the fat body, but AMP induction in the epidermis and epithelia of gut is regulated by the immune deficiency pathway for local immune challenges (44). In our study, we found that S. aureus challenge could induce the expression of the same kinds of AMPs in hemocytes, intestine and gill. AMP expression was not induced in these tissues of Mj-Dorsal RNAi shrimp. These results indicated that S. aureus challenge could induce the systemic immune responses in intestine and gill.
␤-Arrestins were initially identified as mediators of GPCR desensitization and endocytosis (45). Nowadays, they are acknowledged to have signaling functions in a wide variety of signaling pathways and modes of regulation, including the Hedgehog, Wnt, Notch, and TGF␤ pathways (46). As negative regulators, ␤-arrestins regulate the TLR-IL-1R signal pathway in mammals by interacting with TRAF6 and IKB␣ (19,20). at 1 h. A, GST pulldown assay demonstrated the interaction of GST-Mj-ERK with native Mj-␤arrs. The extract solution of gill tissues from normal shrimp was incubated with GST-resin bound to Mj-ERK. The eluate was assayed by Western blotting. The recombinant GST protein was used as control. IB, immunoblot. B, co-immunoprecipitation assay using anti-Mj-␤arr1 or Mj-␤arr2 antibodies. Shrimp intestine tissues were homogenized and centrifuged. The Mj-␤arr1 or Mj-␤arr2 antibodies were incubated with the supernatant, and then isolated using CNBr-activated Sepharose 4B. C, co-immunoprecipitation assay using Mj-ERK or Mj-ERK-P antibodies. Shrimp intestine was homogenized and centrifuged. These antibodies were coupled with CNBr-activated Sepharose 4B to bind natural Mj-␤arrs. D, Mj-␤arr1 and Mj-␤arr2 expression at the mRNA level in intestine after injection of dsMj-␤arr1 or dsMj-␤arr2. E, distribution of Mj-ERK in the nucleus and cytoplasm was detected in intestine after knockdown of Mj-␤arr1 or Mj-␤arr2 and then challenged by S. aureus for 1 h. Bound proteins were eluted and detected by Western blotting with the antibody of interest. F, left panels are the statistical analysis of Mj-ERK-P expression in shrimp for the right panels. The experiments were repeated three times. Asterisks indicate significant differences (*, p Ͻ 0.05; **, p Ͻ 0.01).

FIGURE 6. Mj-␤arrs interacts with non-phosphorylated Mj-ERK and phosphorylation of Mj-ERK was increased in the Mj-␤arrs-knockdown shrimp challenged by S. aureus
␤-Arrestins also regulate the MAPK signaling pathway by interacting with inactive ERK to inhibit its phosphorylation (47,48). In Drosophila, Kurtz inhibits MAPK and Toll signaling during development (24). Anjum et al. (14) demonstrated that Kurtz and SUMO protease Ulp1 work synergistically to regulate the Toll immune pathway in Drosophila. In shrimp, we found that the arrestin-N domain of Mj-␤arrs interacted with the C-terminal PEST domain of Mj-Cactus; the ANK repeats domain of Mj-Cactus interacted with the RHD domain of Mj-Dorsal, and the ␤arr⅐Cactus⅐Dorsal complex was formed using Mj-Cactus as a bridge. The complex prevented phosphorylation and degradation of Mj-Cactus and Mj-Dorsal translocation from the cytoplasm to the nucleus. The C-terminal domain of Mj-␤arrs also interacted with non-phosphorylated Mj-ERK and inhibited its phosphorylation. Phosphorylation of Mj-ERK could promote Mj-Dorsal translocation into the nucleus and the phosphorylation of Ser 276 of Mj-Dorsal (47). The interaction of Mj-␤arrs and Mj-ERK inhibited Mj-ERK phosphorylation and prevented Mj-Dorsal translocation and phosphorylation.
MAPK signaling has been implicated in stress and immunity in evolutionarily diverse species (45). The relationship between MAPK and Toll signaling still needs to be studied in invertebrates. ERK, as one of the three MAPKs, plays a role in regulat-ing transcription activity of p65 in mammals (47,48). In our study, Mj-␤arrs bound to non-phosphorylated Mj-ERK and inhibited its phosphorylation by MEK. Mj-ERK phosphorylation regulated Mj-Dorsal nuclear translocation and phosphorylation of Ser 276 of Mj-Dorsal and then regulated Mj-Dorsal transcriptional activity and DNA binding activity. Therefore, the interaction of Mj-␤arrs with Mj-ERK indirectly inhibited Mj-Dorsal translocation and phosphorylation and limited Toll signaling activity. Uncontrolled immune responses have detrimental outcomes in organisms; therefore, the intensity and duration of immune signaling must be finely regulated. As described above, activation of the Toll pathway is under multiple layers of control, which would limit the damage caused an inappropriate immune response.
Antimicrobial responses in insects mainly consisted of a twostage process as follows: constitutive defense, such as engulfment and melanization, acts immediately, and more than 99% of bacteria in the hemocoel can be cleared in less than an hour. The second stage is inducible defense, i.e. is the production of antimicrobial peptides, which eliminates or suppresses the remaining microbes (49). Similar immune responses occurred in shrimp. However, some pre-synthetized AMPs, such as penaeidins, stored in hemocytes could be released upon immune stimulation (50). Therefore, we performed an initial challenge and then made a second challenge against S. aureus in the bacterial clearance assay (see under "Experimental Procedures").
In mammals, ␤arrs appear to interact with TRAF6 and IB␣ in the TLR signaling pathway and inhibit NF-B activity (19 -21). In Drosophila, Kurtz negatively controls Toll signaling and systemic inflammation at the level of sumoylation (14). In our study, we found that ␤-arrestins could interact with Cactus and inhibit Toll signaling activity. The function of ␤arrs in shrimp is similar to that of mammals.
In summary, our study revealed that in shrimp the bacterial challenge could activate the Toll signaling pathway, and the pathway played a central role in antibacterial immunity. We also showed that that Mj-␤arrs could block or limit Toll signaling by forming a heterotrimeric complex of ␤arr-Cactus-Dorsal to prevent Mj-Cactus phosphorylation and degradation. Mj-␤arrs also inhibited phosphorylation of Mj-ERK, which affected Mj-Dorsal translocation and phosphorylation. Therefore, Mj-␤arrs negatively regulate the Toll pathway by limiting the translocation and phosphorylation of Mj-Dorsal (Fig. 9). These findings for the first time demonstrate that Mj-␤arrs negatively regulate the Toll signaling pathway in two different ways in shrimp.