MD-2 Homologue Recognizes the White Spot Syndrome Virus Lipid Component and Induces Antiviral Molecule Expression in Shrimp

Home , White spot syndrome

MD-2 Homologue Recognizes the White Spot Syndrome Virus Lipid Component and Induces Antiviral Molecule Expression in Shrimp This information is current as of October 1, 2021. Jie Gao, Jin-Xing Wang and Xian-Wei Wang J Immunol 2019; 203:1131-1141; Prepublished online 22 July 2019; doi: 10.4049/jimmunol.1900268

http://www.jimmunol.org/content/203/5/1131 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2019/07/20/jimmunol.190026 Material 8.DCSupplemental http://www.jimmunol.org/ References This article cites 38 articles, 18 of which you can access for free at: http://www.jimmunol.org/content/203/5/1131.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

by guest on October 1, 2021 • No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2019 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

MD-2 Homologue Recognizes the White Spot Syndrome Virus Lipid Component and Induces Antiviral Molecule Expression in Shrimp

Jie Gao,* Jin-Xing Wang,*,†,‡ and Xian-Wei Wang*,†,‡

The myeloid differentiation factor 2 (MD-2)–related lipid-recognition (ML) domain is found in multiple proteins, including MD-2, MD-1, Niemann–Pick disease type C2, and mite major allergen proteins. The significance of ML proteins in antibacterial signal transduction and in lipid metabolism has been well studied. However, their function in host–virus interaction remains poorly understood. In the current study, we found that the ML protein family is involved in resistance against white spot syndrome virus in kuruma shrimp, Marsupenaeus japonicus. One member, which showed a high similarity to mammalian MD-2/MD-1 and was designated as ML1, participated in the antiviral response by recognizing cholesta-3,5-diene (CD), a lipid component of the white spot syndrome virus envelope. After recognizing CD, ML1 induced the translocation of Rel family NF-kB transcription factor Downloaded from Dorsal into the nucleus, resulting in the expression of Vago, an IFN-like antiviral cytokine in arthropods. Overall, this study revealed the significance of an MD-2 homologue as an immune recognition protein for virus lipids. The identification and characterization of CD–ML1–Dorsal–Vago signaling provided new insights into invertebrate antiviral immunity. The Journal of Immunology, 2019, 203: 1131–1141.

yeloid differentiation factor 2 (MD-2)–related lipid- with MD-2 and Niemann–Pick disease type C2 (NPC2), respec- http://www.jimmunol.org/ recognition (ML) family proteins are characterized tively, as representatives for the two typical roles. M by the presence of an ML domain. Comprising ∼150 MD-2 was identified as an accessory receptor accompanying residues, the ML domain shows an overall structure of two anti- TLR4 in LPS sensing (2). MD-2 is coexpressed with TLR4, and parallel b sheets that are built from multiple b strands. The two the TLR4/MD-2 heterodimer forms before LPS binding (3). The sheets enclose a cavity to accommodate lipids or lipid-like mol- hydrophobic cavity of MD-2 provides a site for five of six acyl ecules. The hydrophobic residues in the binding pocket are mainly chains of LPS. The remaining exposed acyl chain of LPS interacts responsible for interacting with the side chains of lipids. A high with the conserved hydrophobic phenylalanines located in the C degree of sequence variation in the ligand-binding site, especially terminus of the extracellular domain of a second TLR4, leading by guest on October 1, 2021 the hydrophobic residues, allows the ML domain to recognize a to dimerization of two TLR4–MD-2 complexes (4, 5). The close variety of lipids (1). Through interacting with specific lipids origi- proximity of the TLR4 intracellular domains results in the re- nated from nonself or self via the ML domain, the ML family is cruitment of downstream adaptors, the initiation of an immune mainly involved in antibacterial immunity and lipid metabolism, signaling cascade, and finally, the expression of inflammatory cytokines (6). MD-1, an ortholog of MD-2 sharing ∼20% sequence homology with MD-2, is also involved in LPS signaling *Shandong Provincial Key Laboratory of Animal Cells and Developmental Biology, together with radioprotective 105, which shares 30% sequence School of Life Sciences, Shandong University, Qingdao, Shandong 266237, China; †Laboratory for Marine Biology and Biotechnology, Qingdao National Labora- identity with TLR4 (7, 8). The radioprotective 105/MD-1 com- tory for Marine Science and Technology, Qingdao, Shandong 266237, China; and plex promotes LPS-induced cell growth and Ab production in ‡ State Key Laboratory of Microbial Technology, Shandong University, Qingdao, B cells and suppresses LPS-induced immune signaling in den- Shandong 266237, China dritic cells and macrophages by interacting with the TLR4/MD-2 Received for publication March 4, 2019. Accepted for publication June 22, 2019. complex (9, 10). This work was supported by the National Science Foundation of China (Grants NPC2 was named for its involvement in the NPC2, which is a 31622058 and 31873043), the National Key Research and Development Program of China (2018YFD0900505), and the Young Scholars Program of Shandong Uni- neurodegenerative lysosomal lipid storage disorder (11). Mutation of versity (Grant 2015WLJH26) (to X.-W.W.). the NPC2 gene would lead to the impairment of egress of choles- The sequences presented in this article have been submitted to GenBank under terol from lysosomes. The NPC2 protein is secreted, recaptured accession numbers MK993577, MK993578, MK993579, MK993580, MK993581, from extracellular sites, and transported into lysosomes. By bind- and MK993582. ing and transferring cholesterol in late endosomes or lysosomes to Address correspondence and reprint requests to Dr. Xian-Wei Wang, School of Life Sciences, Shandong University, Qingdao, Shandong 266237, China. E-mail address: NPC1, another important participator in cholesterol trafficking, [email protected] NPC2 plays an important role in regulating intracellular cholesterol The online version of this article contains supplemental material. homeostasis (12). The structural basis of NPC2-cholesterol bind- Abbreviations used in this article: CD, cholesta-3,5-diene; ChIP, chromatin immu- ing has been revealed. Unliganded bovine NPC2 adopts an overall noprecipitation; F, fusion; GM2A, GM2 activator; ITC, isothermal titration calorim- Ig-like b-sandwich structure, similar to other ML proteins. The etry; MD-2, myeloid differentiation factor 2; ML, MD-2–related lipid-recognition; NPC2, Niemann–Pick disease type C2; qRT-PCR, quantitative real-time RT-PCR; hydrophobic interior core of bovine NPC2 provides a site to RNAi, RNA interference; RSV, respiratory syncytial virus; siRNA, small interfering accommodate cholesterol (13). Similar to NPC2, GM2 activator RNA; SPR, surface plasmon resonance; WNV, West Nile virus; WSSV, white spot (GM2A) is also a small protein involved in lipid metabolism (14). syndrome virus. Through its lipid-binding ability, GM2A functions as a cofactor for Copyright Ó 2019 by The American Association of Immunologists, Inc. 0022-1767/19/$37.50 b-hexosaminidase A, which converts ganglioside GM2 to GM3. www.jimmunol.org/cgi/doi/10.4049/jimmunol.1900268 1132 MD-2 HOMOLOGUE RECOGNIZES VIRUS LIPID IN SHRIMP

Deficiency or functional impairment of GM2A leads to GM2 ac- Expression profile analysis cumulation and neuronal disease (15). Semiquantitative RT-PCR was performed to study the tissue distribution of In addition to LPS signaling and lipid metabolism, studies have MjML mRNAs using the gene-specific primers listed in Table I, following suggested the importance of the ML family in host–virus inter- a conventional procedure: 94˚C for 3 min; 30 cycles of 94˚C for 15 s, 54˚C action. For example, mouse mammary tumor virus enhances its for 30 s and 72˚C for 30 s; and a final 72˚C for 10 min, using the EasyTaq infectivity by acquiring host LPS binding factors, including CD14, PCR SuperMix (TransGen Biotech, Beijing, China). PCR products were analyzed using 1.5% agarose gel electrophoresis. Quantitative real-time TLR4, and MD-2, to exploit LPS from commensal bacteria (16). RT-PCR (qRT-PCR) was performed to determine the expression profiles Human MD-2 interacts with the respiratory syncytial virus (RSV) of MjML mRNAs upon WSSV challenge using the iQ SYBR Green fusion (F) protein to activate TLR4–NF-kB–mediated cytokine Supermix (Bio-Rad Laboratories, Hercules, CA) and the CFX96 Real- expression (17). Loss of host NPC2 greatly increased the cho- Time System (Bio-Rad Laboratories), using the same primers used for RT-PCR. The PCR procedure was as follows: 94˚C for 5 min, followed lesterol level in late endosomes/lysosomes and enhanced HIV by 40 cycles of 94˚C for 10 s and 60˚C for 1 min, and a final melting from infectivity (18). The involvement of ML proteins in host–virus 65˚C to 95˚C. PCR data were analyzed using the 22OOCt method. b-Actin interaction was also observed in invertebrates. Silencing of a was used as the internal reference for both semiquantitative RT-PCR and dengue virus–inducible ML protein in Aedes aegypti, AaegML33, qRT-PCR. The expression level was normalized to the control group at resulted in a significantly lower dengue virus titer in the mosquito each time point. Three independent experiments were performed, and the results represented the mean 6 SD. midgut. Further study showed that AaegML33 likely facilitated viral infection as a potential negative regulator for the JAK/STAT Bioinformatics analysis pathway and the immune deficiency pathway (19, 20). In a study The human MD-2 sequence was used as a template to perform local basic to identify the proteins responsible for white spot syndrome

local alignment search tool searching using BioEdit software to find pos- Downloaded from virus (WSSV) resistance in Pacific white shrimp (Litopenaeus sible homologues from the transcriptome data obtained from our previous vannamei), an ML family member was found to be more abundant transcriptome sequencing of healthy shrimp, which was performed by BGI in virus-resistant shrimp than in virus-susceptible shrimp, suggest- Group (Shenzhen, China). The domain architecture was predicted using SMART (http://smart.embl-heidelberg.de/). A neighbor-joining phyloge- ing its possible role in WSSV resistance (21). We also observed the netic tree was built using MEGA 6.0 with 1000 bootstraps. Multiple frequent appearance of ML family proteins after WSSV infection alignment of protein sequences was performed using the online tool in kuruma shrimp (Marsupenaeus japonicus) when attempting ClustalW2 (https://www.ebi.ac.uk/Tools/msa/clustalw2/). The second- to identify virus-inducible genes using transcriptomic analysis ary structure was predicted using the online tool PredictProtein (https:// http://www.jimmunol.org/ (J. Gao, J.X. Wang, and X.W. Wang, unpublished observations). www.predictprotein.org/). The promoter sequence was analyzed using Promoter Scan online (https://www-bimas.cit.nih.gov/molbio/proscan/). All above information indicated that the ML family plays important roles in the host–virus interaction. However, the specific RNA interference mechanism remains largely elusive. In the current study, we in- A partial DNA fragment was amplified using primers linked to a T7 vestigated the participation of the kuruma shrimp ML family in promoter (Table I). The products were used as templates to synthesize WSSV infection. Specifically, the function and mechanism of dsRNA using an in vitro T7 Transcription Kit (Takara Bio, Dalian, China). an MD-2 homologue was studied. The finding that the shrimp dsGFP was synthesized as a control. The oligonucleotides containing the MD-2 homologue recognizes a WSSV envelope lipid compo- T7 promoter and the small interfering RNA (siRNA) sequence (Table I) were commercially synthesized and used as templates to synthesize siRNA, using by guest on October 1, 2021 nent and regulates the expression of antiviral molecules pro- the in vitro T7 Transcription Kit (for siRNA synthesis) (Takara Bio). For vides new insights into the significance of the ML family in each gene, two specific siRNAs were synthesized and mixed for use. The host–virus interaction. siRNA specific for GFP sequence was synthesized as a control. Specific dsRNAs or siRNAs (30 mg) were injected into the shrimp hemocoel at the abdominal segment, and the control group was injected with an equal Materials and Methods amount of control dsRNA or siRNA. The RNA interference (RNAi) effi- Animal cultivation, virus challenge, and sample preparation ciency was determined using qRT-PCR at 48 h after dsRNA injection or 24 h after siRNA injection. Healthy kuruma shrimp (M. japonicus; 3–5 g) were purchased from a To test the function of MjMLs in virus infection, WSSV infection market in Jinan, Shandong, China and cultivated in air-pumped artificial (5 3 105 copies per shrimp) was performed at 48 h after dsRNA or 24 h seawater (25˚C) for at least 1 wk before the experiments. The WSSV strain after siRNA injection. The expression level of WSSV VP28 was then ana- used in this study was collected from infected red swamp crayfish (Pro- lyzed using qRT-PCR with the primers listed in Table I. The experiments cambarus clarkii), and the original inoculum was gifted from the East were performed independently three times. China Sea Fisheries Research Institute (Shanghai, China). Shrimp were artificially infected via i.m. injection with the original inoculum, and Generation of recombinant protein 2 moribund shrimp were collected and stored at 80˚C before use for The sequence encoding the MjML1 mature peptide was amplified using the WSSV propagation. To prepare successive inoculums, shrimp gills (1 g) specific primers listed in Table I and ligated into the pET32a(+) plasmid. were homogenized in 10 ml of PBS (140 mM NaCl, 2.7 mM KCl, 10 mM The recombinant vector was transformed into Escherichia coli Rosetta Na HPO , and 1.8 mM KH PO [pH 7.4]). The homogenate was frozen 2 4 2 4 (DE3) strain for expression under induction with 0.5 mM isopropyl-b-D- 3 g and thawed twice and then centrifuged at 3000 rpm for 10 min at 4˚C. thiogalactopyranoside. Inclusion bodies were extracted, washed, and The supernatant was filtered through a 0.45-mm filter. The filtrate was used m dissolved in buffer (0.1 mM Tris-HCl [pH 8], 10 mM DDT, and 8 M urea) as the WSSV inoculum. Viral DNA was extracted from 100 l of the fil- and renatured by dialysis in PBS with 5% glycerol. Recombinant proteins trate using MagExtractor Genome (TOYOBO, Shanghai, China) to de- were then purified using affinity chromatography with ProteinIso Ni-NTA termine the viral titer according to a previously described method (22). The 2 Resin (TransGen Biotech). Endotoxins were removed by thorough washing rest of the filtrate was stored at 80˚C and diluted to the appropriate titer with cold 0.1% Triton X-114 before the final elution of the protein from with PBS before use. For WSSV challenge, each shrimp was injected with 3 5 the column (23). Purified proteins were then dialyzed in PBS and stored WSSV inoculum at a dose of 5 10 virus copies. PBS was injected as the at 280˚C before use. A tag expressed by the empty vector was prepared control. At different times postinfection, shrimp hemolymph was collected simultaneously. into cold anticoagulant (0.45 M NaCl, 10 mM KCl, 10 mM EDTA, and 10 mM HEPES [pH 7.45]) and centrifuged at 800 rpm 3 g for 7 min at Application of the recombinant protein in vivo 4˚C to obtain the hemocytes pellet. Other tissues were collected simultaneously. Each sample originated from at least five shrimp. Total RNA was Shrimp were divided into four groups (10 individuals in each group). extracted using TRIzol (Invitrogen, Waltham, MA) from ∼100 mg of tissue Three groups were injected with different amounts of rMjML1 (10, 3, or 2 3 107 cells, and the first stranded cDNA was synthesized using a and 1 mg/shrimp) and WSSV (5 3 105 copies per shrimp), and the ReverTra Ace qPCR RT Master Mix with gDNA Remover (TOYOBO) other group was injected with the control tag protein (10 mg/shrimp) according to the manufacturers’ instructions. and WSSV (5 3 105 copies per shrimp). Thereafter, total RNA was The Journal of Immunology 1133

Table I. Primers used for this study

Primer Sequence (59-39) (q)RT-PCR MjML1RTF 59-GCTTGGAAATCCTGGGTGG-39 MjML1RTR 59-CGGGGTATCGTCTCCTCATC-39 MjML2RTF 59-GTAACGGCAGTGAAGGCA-39 MjML2RTR 59-CTGATTGGAAGCGAAGATGT-39 MjML3RTF 59-CTGCGAACTGATTGTGGAGG-39 MjML3RTR 59-CAAGACGAATAGCGAAATGGAT-39 MjML4RTF 59-ATGGTCGTAGCGGTTGG-39 MjML4RTR 59-AAAGATGGCTCGTAGATTGTG-39 MjML5RTF 59-CGTCGCCTCAACAACAATAG-39 MjML5RTR 59-AATCTTCTCGGTTCCATCCA-39 MjML6RTF 59-GTCTGGACACGGACGGCT-39 MjML6RTR 59-CTACAAACATCTTGGCTACACGA-39 b-ActinRTF 59-CAGCCTTCCTTCCTGGGTATGG-39 b-ActinRTR 59-GAGGGAGCGAGGGCAGTGATT-39 VP28RTF 59-AGCTCCAACACCTCCTCCTTCA-39 VP28RTR 59-TTACTCGGTCTCAGTGCCAGA-39 MjVago1RTF 59-GCCTGTTTCCCTTCTGTGG-39 MjVago1RTR 59-GCTTCTCAATCTCTGCCTGG-39 MjVago3RTF 59-CGTGTCCTGTTGTTCCTCGC-39 Downloaded from MjVago3RTR 59-CCGCACCCTGTCTCCTCATA-39 MjVago4RTF 59-CACTCAAGTTCCGCTCGTTT-39 MjVago4RTR 59-TCGCCTCGGTCTTCTCTCA-39 MjVago5RTF 59-GGCGGAGGCAAAAGCATC-39 MjVago5RTR 59-GTGGCGAGTGTCACCATAAGC-39 MjVago6RTF 59-TCCTGTTAGAGATGGCGGTT-39 9 GACACACAATCAAGGCAGAGTTA 9 MjVago6RTR 5 - -3 http://www.jimmunol.org/ RNAi MjML1RNAiF 59-GCGTAATACGACTCACTATAGGCACCAGCGGGCACACTTA-39 MjML1RNAiR 59-GCGTAATACGACTCACTATAGGACGGGGTATCGTCTCCTCAT-39 MjML2RNAiF 59-GCGTAATACGACTCACTATAGGCAGTGAAGGCAAAAGTAA-39 MjML2RNAiR 59-GCGTAATACGACTCACTATAGGTTTCTGAATGATGGACAA-39 MjML3Si1-1 59-GATCACTAATACGACTCACTATAGGGCCAAGATGACCGGGATCAATT-39 MjML3Si1-2 59-AATTGATCCCGGTCATCTTGGCCCTATAGTGAGTCGTATTAGTGATC-39 MjML3Si2-1 59-AACCAAGATGACCGGGATCAACCCTATAGTGAGTCGTATTAGTGATC-39 MjML3Si2-2 59-GATCACTAATACGACTCACTATAGGGTTGATCCCGGTCATCTTGGTT-39 MjML4RNAiF 59-GCGTAATACGACTCACTATAGGAGGTCAGGATAGGGAGCA-39 MjML4RNAiR 59-GCGTAATACGACTCACTATAGGGGTGTCTACATCTAAACGGG-39 by guest on October 1, 2021 MjML5Si1-1 59-GATCACTAATACGACTCACTATAGGGGGACATAGTCGTCGAGATATT-39 MjML5Si1-2 59-AATATCTCGACGACTATGTCCCCCTATAGTGAGTCGTATTAGTGATC-39 MjML5Si2-1 59-AAGGACATAGTCGTCGAGATACCCTATAGTGAGTCGTATTAGTGATC-39 MjML5Si2-2 59-GATCACTAATACGACTCACTATAGGGTATCTCGACGACTATGTCCTT-39 MjML6Si1-1 59-GATCACTAATACGACTCACTATAGGGGCTCGTGTAGCCAAGATGTTT-39 MjML6Si1-2 59-AAACATCTTGGCTACACGAGCCCCTATAGTGAGTCGTATTAGTGATC-39 MjML6Si2-1 59-AAGCTCGTGTAGCCAAGATGTCCCTATAGTGAGTCGTATTAGTGATC-39 MjML6Si2-2 59-GATCACTAATACGACTCACTATAGGGACATCTTGGCTACACGAGCTT-39 MjVago5Si1-1 59-GATCACTAATACGACTCACTATAGGGGCCCTACCTCTGTATGCCCTAAATATT-39 MjVago5Si1-2 59-AATATTTAGGGCATACAGAGGTAGGGCCCCTATAGTGAGTCGTATTAGTGATC-39 MjVago5Si2-1 59-GATCACTAATACGACTCACTATAGGGGCGCCTTTGCAAATGACCCAGTGATTT-39 MjVago5Si2-2 59-AAATCACTGGGTCATTTGCAAAGGCGCCCCTATAGTGAGTCGTATTAGTGATC-39 GFPRNAiF 59-GCGTAATACGACTCACTATAGGTGGTCCCAATTCTCGTGGAAC-39 GFPRNAiR 59-GCGTAATACGACTCACTATAGGCTTGAAGTTGACCTTGATGCC-39 GFPSi1-1 59-GATCACTAATACGACTCACTATAGGGGGAGTTGTCCCAATTCTTGTT-39 GFPSi1-2 59-AACAAGAATTGGGACAACTCCCCCTATAGTGAGTCGTATTAGTGATC-39 GFPSi2-1 59-AAGGAGTTGTCCCAATTCTTGCCCTATAGTGAGTCGTATTAGTGATC-39 GFPSi2-2 59-GATCACTAATACGACTCACTATAGGGCAAGAATTGGGACAACTCCTT-39 Recombinant expression MjML1EF 59-CGCGGATCCGCAGAGGTCTACGAGATCCC-39 MjML1ER 59-CCGCTCGAGTTACATGATTTTAATACTGA-39 ChIP MjVago5chipF 59-TACAAACACCGACAATGAGAAA-39 MjVago5chipR 59-GCGTACCGTAAACACTGAAGAT-39

extracted to determine the transcription level of WSSV VP28 using Western blotting qRT-PCR. Protein samples were prepared to determine the translation level of VP28 using Western blotting with anti-VP28 Abs. Shrimp gills were thoroughly homogenized in PBS, and the homogenate 3 Genomic DNA was also extracted to determine the virus copy was centrifuged at 12,000 rpm g for 20 min to collect the supernatant. The number in shrimp tissues, following a previously described method protein concentration was determined using the Bradford Protein Assay Kit (24). For the survival analysis, 30 shrimp were injected with WSSV (Sangon Biotech, Shanghai, China). The supernatant was added to Protein (5 3 106 copies per shrimp) together with 5 mg of rMjML1 or Loading Dye (Sangon Biotech), boiled at 100˚C for 5 min, and centri- control tag. The survival rates were recorded for both groups every fuged at 8000 rpm 3 g for 3 min. The protein samples were separated by 12 h for 72 h. 10% SDS-PAGE with 100 mg protein loaded per lane. The proteins were 1134 MD-2 HOMOLOGUE RECOGNIZES VIRUS LIPID IN SHRIMP transferred onto a nitrocellulose membrane using a conventional binding model. The equilibrium dissociation constant was presented semidry transfer protocol with a constant voltage of 9 V for 50 min as the ratio dissociation rates/association rates. The SPR assay was using a Jim-X Semi-Dry Blotter (Jim-X, Dalian, China). After blocking performed independently three times. with 3% nonfat milk in TBS for 1 h, the membranes were incubated with specific primary Abs for 3 h at 25˚C. After washing three Immunocytochemistry assay times with TBST, the membrane was then incubated with HRP-labeled Shrimp were injected with WSSV (5 3 105 copies), CD (1 mg), or rMjML1 secondary Ab (1:10,000 diluted in nonfat milk) for 2 h. Unbound IgG (5 mg) with PBS, DMSO, or tag as the control. At specific time points after was washed away using TBST. Bands were visualized using the High- injection, the hemocytes were collected as described above. Fresh anti- sig ECL Western Blotting Substrate (Tanon Science & Technology, coagulant with 4% paraformaldehyde was used to wash and to incubate Shanghai, China). The chemiluminescent signal was detected using the hemocytes for 10 min. The hemocytes were then collected and a 5200 Chemiluminescence Imaging System (Tanon Science & Tech- b resuspended in PBS and smeared onto poly-L-lysine–coated glass slides. nology). The anti-VP28, anti-Dorsal, and anti– -actin Abs were made in One hour later, 0.2% Triton X-100 in PBS was added onto the slides and our laboratory by immunizing New Zealand rabbits with the recombinant incubated for 10 min. After washing three times with PBS (for 6 min each proteins (24–26). The anti-histone 3 Abs were purchased from Proteintech time), the glass slides were blocked by 3% BSA in PBS at 37˚C for 30 min. Group (Wuhan, China); all secondary Abs were purchased from Zhongshan The hemocytes on the slides were then incubated overnight with anti- (Beijing, China). Dorsal Abs (1:500 in blocking buffer) at 4˚C. After washing with PBS five times, the hemocytes were incubated with 3% BSA for 10 min. Then, Separation of nuclear and cytoplasmic proteins goat anti-rabbit Alexa Fluor 488 (1:1000 diluted in 3% BSA) was added The separation of nuclear and cytoplasmic proteins was performed using a for 1 h in the dark. After washing with PBS five times, the hemocytes were 9 Nuclear Protein Extraction Kit (Solarbio, Beijing, China) according to the stained with 4 ,6-diamidino-2-phenylindole dihydrochloride (AnaSpec, San manufacturer’s instructions. Briefly, shrimp gills were washed with PBS Jose, CA) for 10 min at 25˚C and then washed six times with PBS. Finally, three times and homogenized with 1 ml of cytoplasmic protein extraction the slides were observed under an Olympus DP71 fluorescent microscope reagent containing 1 mM PMSF. The homogenate was processed by five (Olympus, Tokyo, Japan). The colocalization percentage of Dorsal- and Downloaded from rounds of vortex shaking for 15 s and incubation in an ice bath for 3 min. The homogenate was centrifuged at 15,000 rpm 3 g for 25 min at 4˚C, and the supernatant was collected as the pool of cytoplasmic proteins. The sediment was washed with PBS three times and then resuspended in nucleoprotein extraction reagent containing 1 mM PMSF. The suspension was processed by five rounds of vortex shaking for 15 s and incubation in an ice bath for 3 min and then centrifuged at 15,000 rpm 3 g for 25 min at 4˚C. The supernatant protein was collected as the pool of nuclear proteins. The http://www.jimmunol.org/ protein concentration was determined as described above.

Lipid-binding assay The characterization of the binding of lipid cholesta-3,5-diene (CD) (Sigma-Aldrich, St. Louis, MO) to MjML1 was performed in three ways. An ELISA was performed as a preliminary binding assay. The 96-well plates were coated with 50 ml of CD or cholesterol solution (5 mg), incubated at 4˚C until air dried for 3 d in a frost-free refrigerator, and

blocked with 3 mg/ml of BSA solution. Purified rMjML1 was added to by guest on October 1, 2021 each well to a final concentration ranging from 5 to 100 nM. The plate was incubated at 25˚C for 3 h and then washed with TBS three times. Mouse anti-His Abs (1:1000 diluted in 0.1 mg/ml of BSA solution) were added and incubated for 4 h at 25˚C. Alkaline phosphatase–conjugated horse anti-mouse IgG (1:10,000 diluted in 0.1 mg/ml of BSA solution) was added after washing with TBS three times. After incubation for 3 h at 25˚C, the wells were washed four times with TBS and added with p-nitro-phenyl phosphate (1 mg/ml in 10 mM diethanolamine with 0.5 mM MgCl2). The absorbance at 405 nm was then measured after incubation for 30 min at 25˚C. Data are presented as the mean 6 SD derived from three independent repeats. The lipid-binding site of MjML1 was predicted by aligning MjML1 with human MD-2 and chicken MD-1. This short peptide was commercially synthesized by GenScript (Nanjing, China) at .95% purity. Using the synthesized peptide, isothermal titration calorimetry (ITC) and surface plasmon resonance (SPR) assays were performed to characterize the binding between MjML1 and CD in detail. The ITC assay was performed using MicroCal PEAQ-ITC (Malvern Panalytical, Malvern, U.K.). CD and the peptide were dissolved in 15% DMSO to a final concentration of 2 and 0.2 mM, respectively. According to the manufacturer’s instructions, CD was pumped in using a syringe, and the peptide was injected into the ITC FIGURE 1. Neighbor-joining phylogenetic analysis of the ML family cell. Injection of 2 ml of CD solution over a period of 150 s at a stirring from kuruma shrimp (M. japonicus) with typical vertebrate MLs. The tree speed of 750 rpm 3 g was performed. For the blank control test, CD was built using MEGA 6.0 with bootstrapping of 1000. The accession solution was pumped into the sample cell of dissolvent. Before the anal- numbers of the sequences used in the phylogenetic analysis are as follows: ysis, the data were subtracted with that from the blank control test. A Homo sapiens NPC2, AAH02532; Mus musculus Npc2, AAH07190; Gallus negative control test was performed as described above using an unrelated gallus Npc2, NP_001026374; Xenopus laevis Npc2, AAH60392; Danio peptide comprising a partial peptide of Vibrio anguillarum flagellin A that rerio Npc2, AAH45895; Ciona intestinalis Npc2, XP_002127695; was synthesized in the same way as the MjML1 peptide. This assay was M. japonicus ML5, MK993581; M. japonicus ML3, MK993579; M. japo- repeated three times independently. nicus ML6, MK993582; M. japonicus ML2, MK993578; G. gallus GM2A, The SPR assay was performed using a BIACORE T200 (GE Healthcare Life Sciences, Chicago, IL). CD was coated onto the CM5 sensor chip XP_003642107; D. rerio Gm2a, XP_005173176; H. sapiens GM2A, AAD25741; using an Amine Coupling Kit (GE Healthcare Life Sciences). The syn- M. musculus GM2A, CAJ18497; Rattus norvegicus GM2A, AAH72474; thesized peptide was dissolved to a series of final concentrations ranging M. japonicus ML1, MK993577; M. japonicus ML4, MK993580; H. sapiens from 40.625 to 1300 mM and injected to flow through the chip at a flow MD-1, AAC98152; M. musculus MD-1, BAA32399; G. gallus MD-1, rate of 30 ml/min. Biacore Evaluation Software was used to calculate the NP_001004399; H. sapiens MD-2, BAA78717; Bos taurus MD-2, BAC67682; association rates and dissociation rates using the one-to-one Langmuir M. musculus MD-2, BAA93619; and Cricetulus griseus MD-2, AAK57984. The Journal of Immunology 1135

DAPI-stained nuclei was analyzed using the Wright Cell Imaging Facility separated into two clusters. MjML2, MjML3, MjML5, and MjML6 ImageJ software. were clustered together with vertebrate NPC2 proteins, whereas Chromatin immunoprecipitation assay MjML1 and MjML4 were clustered together with vertebrate MD-2 and MD-1 proteins. The relatively close relationship suggested that The chromatin immunoprecipitation (ChIP) assay was performed using a MjML 1 and MjML4 might play similar functions to vertebrate MD- ChIP Assay Kit (Beyotime Biotechnology, Wuhan, China) according to the manufacturer’s instructions. Shrimp were injected with CD (1 mg) or 2 and MD-1, which have been proven as important participants in DMSO. The hemocytes were collected 6 h after injection and used as the innate immunity by recognizing LPS or other foreign lipid-like pool for ChIP. The primers used for the MjVago5 promoter are listed in objects. Table I. The experiment was performed independently twice. Statistical analysis The shrimp ML family responded to WSSV infection and played antiviral roles Most results in this study were analyzed using Student t test, and significance was accepted with p , 0.05. For the expression profiles, data were analyzed The tissue distributions of MjMLs transcripts were first analyzed using one-way ANOVA, followed by Tukey multiple comparison test using (Table I). As shown in Fig. 2A, MjML1, MjML2, and MjML4 were GraphPad Prism software. For the survival assay, the results were analyzed widely expressed in all tested tissues, whereas MjML3 and MjML5 using the log-rank (Mantel–Cox) test using GraphPad Prism software. were expressed only in the hepatopancreas. To confirm the involvement of the ML family in WSSV infection, the expression Results profiles of MjMLs after virus infection were studied. The results Identification of the ML family from kuruma shrimp showed a significant induction of both MjML1 and MjML3 in Using the human MD-2 as a template, six proteins were identified the hepatopancreas. MjML1 expression was induced more than Downloaded from as putative ML family members using local basic local alignment 20-fold at the early stage of infection (6 h), whereas MjML3 ex- search tool searching from the data set of several previous tran- pression gradually increased, reaching a peak at the mid and late scriptomic analyses (J. Gao, J.X. Wang, and X.W. Wang, unpublished stage of infection (24–72 h) (Fig. 2B). RNAi was then performed observations). Each protein contained a signal peptide and a typical to knock down MjMLs expression to determine their specific roles ML domain, and they were named MjML1–MjML6 (GenBank ac- during WSSV infection. As shown in Fig. 3A, MjML expression cession numbers: MK993577, MK993578, MK993579, MK993580, was successfully inhibited by dsRNA or siRNA injection. Knock- MK993581, and MK993582). Multiple alignment of the shrimp down of the expression of any MjML family member led to en- http://www.jimmunol.org/ ML family and mammalian ML proteins suggested that the cysteine hanced WSSV gene expression, with the highest enhancement residues responsible for the formation of disulfide bonds to stabilize occurring after knockdown of MjML1 or MjML3 (Fig. 3B). Taken the overall structure were conserved among all sequences together, these results suggested that MjML1 and MjML3 play (Supplemental Fig. 1). To reveal the possible relationship important roles in restricting WSSV infection. between the shrimp ML family and other typical vertebrate ML We decided to focus on MjML1. The enhanced WSSV in- proteins, phylogenetic analysis was performed. As shown in Fig. 1, fection after MjML1 knockdown was confirmed by observing three subgroups of vertebrate ML proteins (MD-2 and MD-1, NPC2, higher levels of the of WSSV VP28 protein in the knock- and GM2A) were distinctly clustered, and the MjML family was down group compared with that in the control group (Fig. 4A). by guest on October 1, 2021

FIGURE 2. Expression profiles of MjML genes. (A) Tissue distribution of MjML mRNAs. Total RNAs were extracted from healthy shrimp. Expression was studied using RT-PCR with b-actin as the internal reference. Each sample comprised at least five shrimp. The data are representative of two independent experiments. (B) Expression profiles of MjML mRNAs in the hepatopancreas after WSSV infection. qRT-PCR was performed to check MjML mRNA expression with b-actin as the reference. The expression level was normalized to the control test for each time point. The results are shown as the mean 6 SD. Different characters indicate a significant difference, analyzed using one-way ANOVA, followed by Tukey multiple comparison test in GraphPad Prism. Repeats were performed in triplicate with at least five shrimp for each sample. 1136 MD-2 HOMOLOGUE RECOGNIZES VIRUS LIPID IN SHRIMP

FIGURE 3. The function of MjML in WSSV infection, as analyzed using RNAi. (A) RNAi efﬁciency of MjML. dsRNA or siRNA (30 mg) was injected into shrimp hemocoel. RNAi efﬁciency was studied 48 h after dsRNA injection or 24 h after siRNA injection using qRT-PCR. (B) Expression of WSSV VP28 after MjML knockdown. WSSV infection (5 3 105 copies per shrimp) was performed 48 h after dsRNA or 24 h after siRNA injection. VP28 ex-

pression was checked another 24 h later using qRT-PCR. Three repeated experiments were performed independently. Each sample comprised at least ﬁve Downloaded from shrimp. Data show the mean 6 SD. *p , 0.05, *p , 0.01, ***p , 0.001, as calculated using Student t test.

Recombinant MjML1 was then expressed, puriﬁed, and applied data suggested that MjML1 plays an important role in the antiviral in vivo to further study its function in inhibiting virus infection response. (Fig. 4B). The results showed that rMjML1 application could MjML1 recognized WSSV envelope lipid component CD suppress VP28 transcription (Fig. 4C) and VP28 translation http://www.jimmunol.org/ (Fig. 4D) and reduced the viral titer (Fig. 4E) in shrimp tissue in a The mechanism by which MjML1 exerts its antiviral role was dose-dependent manner (Fig. 4C). Moreover, after shrimp were investigated. The close phylogenetic relationship between MjML1 injected with rMjML1 or the control tag combined with WSSV and mammalian MD-2, which functions in the host response by inoculums, the survival rate of the rMjML1 group was always recognizing nonself targets, prompted us to speculate whether higher than that of the control group (Fig. 4F). Collectively, these MjML1 could bind any lipid components of WSSV, especially by guest on October 1, 2021

FIGURE 4. Antiviral function of MjML1. (A) Expression level of WSSV VP28 protein after MjML1 knockdown. Protein samples were extracted 24 h postinfection in MjML1 preknockdown shrimp. VP28 protein expression was studied using Western blotting with anti-VP28 Abs. Data are representative of two independent repeats. (B) Recombinant expression and purification of MjML1 and control tag. rMjML1 and tag were expressed with recombinant and empty pET32a(+) vector in E. coli Rosseta (DE3) cells and purified using affinity chromatography. (C) Expression of WSSV genes after rMjML1 application in vivo. Shrimp were injected with different amounts of rMjML1 (10, 3, and 1 mg) and WSSV (5 3 105 copies). The control group was injected with tag (10 mg) and WSSV. WSSV VP28 expression was detected using qRT-PCR 24 h later. (D) Expression of the WSSV VP28 protein after rMjML1 application in vivo. Protein samples were extracted at 24 and 48 h postinfection and analyzed using Western blotting with anti-VP28 Abs. Data are representative of two independent experiments. (E) WSSV copy number quantification after rMjML1 application in vivo. Genomic DNA was extracted at 48 h after WSSV infection and rMjML1 application and was used to determine the WSSV copy number using qRT-PCR. At least five shrimp were used to prepare a sample. Three independent repeats were performed. Data are shown as the mean 6 SD. **p , 0.01, ***p , 0.001, as analyzed using Student t test. (F) Survival rate after rMjML1 application. Shrimp were injected with rMjML1 or tag (5 mg) and WSSV (1 3 106 copies). Each group consisted of 30 shrimp. The survival rate was recorded every 12 h for 72 h. The results were analyzed using the Log-rank (Mantel–Cox) test in GraphPad Prism. The Journal of Immunology 1137 those of the WSSV surface envelope. A previous study found that (Fig. 5E) assays confirmed the CD-binding ability of MjML1, with although the WSSV envelope was generated from the host nu- similar binding properties (dissociation constant) obtained (101 6 cleoplasm, the cholesterol and hydrocarbons in WSSV were much 23.1 mM in the ITC assay, 130.4 6 29.86 mMintheSPRassay). more complicated than in the host hemocytes. Among the cho- These data suggested that MjML1 might exert its antiviral role by lesterol and hydrocarbon-like molecules present in the WSSV recognizing the lipid component of the WSSV envelope. envelope and absent in host nuclei, CD, the dehydroxyl derivative of cholesterol (Fig. 5A), accounted for the highest proportion other MjML1 regulated Vago5 expression after recognizing CD than cholesterol (27). Thus, we detected whether MjML1 could We next attempted to identify the downstream effector molecule after bind CD. As shown in Fig. 5B, rMjML1 preferentially bound to MjML1 recognition of CD. We specifically focused on the Vago CD rather than cholesterol when coated onto the plates, and the family, which was reported as the functional analogues of IFN in binding increased as the dose of rMjML1 increased. insect antiviral immunity (28). The Vago family is also expressed in ITC and SPR assays were then performed to study the binding L. vannamei and is important for shrimp antiviral immunity (29). properties of MjML1. The lipid-binding region of MjML1 was first Analysis of the transcriptomic data of kuruma shrimp and the identified by alignment of MjML1 with human MD-2 and chicken sequences deposited in GenBank identified five Vago proteins in MD-1, whose lipid-binding properties had been revealed based on kuruma shrimp with the accession numbers BAW35380, BAW78900, their crystal structures (Fig. 5C). The peptide encompassing the BAW78901, BAW78902, and BAW78903. As shown in Fig. 6A, lipid-binding region of MjML1, located between b strands G and H, MjVago5 expression could be significantly induced by WSSV in- was commercially synthesized to meet the demand of high purity of fection, suggesting its possible involvement in the antiviral response. protein in the ITC and SPR assays. Both the ITC (Fig. 5D) and SPR Knockdown of MjVago5 expression using siRNA (Fig. 6B) led to a Downloaded from http://www.jimmunol.org/ by guest on October 1, 2021

FIGURE 5. Binding of CD by MjML1. (A) Chemical structures of cholesterol and CD. (B) Binding of CD by rMjML1, analyzed using ELISA. CD or cholesterol (5 mg) was coated into 96-well plates and air dried. After blocking with 3 mg/ml of BSA, proteins (5 to 100 nM) were added into the wells for 3 h at 25˚C. Bound proteins were detected using a conventional ELISA. Data show the mean 6 SD. *p , 0.05, Student t test. (C) Identiﬁcation of the lipid- binding site of MjML1. Alignment was performed using online ClustalW2 analysis. The b strands are labeled with arrows. The lipid-binding sites are in red. (D) Binding of CD by MjML1 lipid-binding peptide, analyzed using ITC. The ITC assay was performed with MicroCal PEAQ-ITC (Malvern Pan- alytical). CD (2 mM) was pumped in using a syringe, and peptide (0.2 mM) was injected into the ITC cell. Injection of 2 ml of CD solution over a period of 150 s at a stirring speed of 750 rpm 3 g was performed. Data shown are representative of three independent repeats. (E) Binding of CD by MjML1 lipid- binding peptide analyzed using SPR. SPR was performed with BIACORE T200 (GE Healthcare Life Sciences). CD was coated onto the CM5 sensor chip. Peptide solution (40.625 to 1300 mM) was injected to ﬂow through the chip with a rate of 30 ml/min. This assay was performed independently in triplicate, and a representative result is shown. 1138 MD-2 HOMOLOGUE RECOGNIZES VIRUS LIPID IN SHRIMP Downloaded from http://www.jimmunol.org/

FIGURE 6. Regulation of MjVago5 expression by CD-MjML1. (A) Induction of MjVago expression by WSSV. Shrimp were infected with WSSV (5 3 105 copies). Expression of MjVago was detected by qRT-PCR 24 h later. The fold induction was expressed as the ratio between the WSSV infection and the control group. The results show the mean 6 SD from three independent repeats. (B) RNAi efficiency for MjVago5. siRNA (30 mg) was injected into shrimp. The expression of MjVago5 was detected 24 h later. (C) WSSV VP28 protein expression after MjVago5 knockdown. WSSV was injected into shrimp 24 h after MjVago5 siRNA injection. Protein samples were extracted at 24 and 48 h postinfection and analyzed using Western blotting with anti- VP28 Abs. Data are representative of two independent experiments. (D) Expression of MjVago5 after MjML1 knockdown or rMjML1 application. WSSV (5 3 105 copies) was injected 24 h after MjML1 siRNA injection (30 mg) or together with rMjML1 (5 mg). The expression of MjVago5 was detected 24 h later. Data are shown as the mean 6 SD from three independent repeats. (E) The expression of MjML1 and MjVago5 after CD stimulation. Shrimp were by guest on October 1, 2021 injected with CD (1 mg) with DMSO as a control. The expression of MjML1 and MjVago5 was detected using qRT-PCR. (F) The effect of MjML1 on the induction of MjVago5 by CD stimulation. Shrimp were injected with CD (1 mg) 24 h after MjML1 siRNA injection (30 mg) or together with rMjML1 (5 mg). MjVago5 expression was detected 24 h later. The results are shown as the mean 6 SD from three independent experiments. The results were analyzed using Student t test. *p , 0.05, **p , 0.001, ***p , 0.001. higher expression of WSSV VP28 than in the control group (Fig. 6C), Toll pathway. By separating and analyzing the cytoplasmic and confirming the significance of MjVago5 in the host–virus interaction. nucleoproteins after WSSV infection or CD stimulation, obvious The regulation of MjVago5 by MjML1 was then characterized. translocation of Dorsal from the cytoplasm to the nucleus was As shown in Fig. 6D, MjML1 knockdown obviously downregu- observed (Fig. 7A). This suggested that WSSV could induce the lated MjVago5 expression, regardless of the absence or presence of activation of the shrimp Toll pathway and that CD was indeed WSSV infection. Conversely, rMjML1 application increased the a molecular pattern for WSSV and was sufficient to imitate WSSV expression of MjVago5, further increasing the induction of MjVago5 infection. As expected, rMjML1 application also led to Dorsal by WSSV infection. Besides, CD significantly induced the expres- translocation, suggesting that MjML1 might regulate Dorsal sion of both MjML1 and MjVago5 (Fig. 6E). These data suggested function after recognizing the WSSV lipid component (Fig. 7A). that MjML1 transmitted the WSSV/CD signal to induce MjVago5 To confirm the translocation of Dorsal after the above treatments, expression. To test this hypothesis, CD stimulation was performed immunocytochemical analysis was performed to determine the in the shrimp in which MjML1 expression was presilenced. The distribution of Dorsal in shrimp hemocytes. As shown in Fig. 7B results showed that CD-mediated induction of MjVago5 was sup- and 7C, WSSV infection, CD stimulation, and rMjML1 appli- pressed by MjML1 silencing. Moreover, rMjML1 application en- cation could indeed induce the nuclear translocation of Dorsal. hanced CD-mediated induction of MjVago5 expression (Fig. 6F). These data suggested that MjML1 regulates MjVago5 expression This supported the view that MjML1 is a primary determinant for through Dorsal after immune recognition of WSSV/CD. transferring the signal from CD to downstream MjVago5. CD stimulation enhanced the transcription of MjVago5 MjML1 regulated the subcellular location of the NF-kB by Dorsal transcription factor Dorsal The above results suggested an immune signaling transmission Mammalian MD-2 participates in LPS signaling by forming a from WSSV/CD to MjML1, Dorsal, and finally, MjVago5. However, receptor complex with TLR4 and activating NF-kB transcription whether MjVago5 expression was directly or indirectly regulated by factor–mediated inflammatory gene expression; therefore, we in- Dorsal remained uncertain. The promoter region of MjVago5 was vestigated whether MjML1 could function through regulating cloned, and two NF-kB sites were identified in the promoter se- Dorsal, the principal NF-kB transcription factor of the arthropod quence (Fig. 8A). Next, a ChIP assay was performed to check The Journal of Immunology 1139

FIGURE 7. Induction of Dorsal translocation into nucleus by WSSV, CD, and rMjML1. (A) Distribution of the Dorsal protein after treatment, analyzed by Western blotting. Shrimp were injected with WSSV (5 3 105 copies), rMjML1 (5 mg), or CD (1 mg). Hemocytes were collected at specific times, and the cytoplasmic and nuclear proteins were separated. b-Actin and Histone3 were detected as references for the cytoplasmic and nuclear proteins, respectively. The data shown are the representative of two independent repeats. (B) Dorsal translocation analyzed using immunocytochemistry. Hemocytes were collected after specific treatments and fixed. The slides were analyzed using immunohis- tochemistry with anti-Dorsal Abs. Data are Downloaded from representative of two independent repeats. Scale bar, 10 mm. (C) Digitization of the result in (B). The colocalization percentage of Dorsal and nuclei was analyzed using WCIF ImageJ software. ***p , 0.001, as analyzed using Student t test. http://www.jimmunol.org/

whether Dorsal could bind the DNA fragment containing the of Npc2a in sterol homeostasis and ecdysteroid biosynthesis NF-kB site. As shown in Fig. 8B, a positive signal was de- suggested that it was indeed the ortholog of mammalian NPC2. tected after immunoprecipitation using anti-Dorsal Abs with CD- However, together with other ML proteins in D. melanogaster, by guest on October 1, 2021 stimulated hemocytes for the pool of ChIP. This suggested that Npc2a also plays a role in innate immunity by binding bacterial Dorsal could directly transcriptionally regulate MjVago5 expres- polysaccharides and modulating the immune deficiency pathway, sion and that CD-stimulated translocation of Dorsal to the nucleus suggesting the complexity and significance of the ML family in was essential for MjVago5 transcription. Thus, the above data invertebrate physiological processes (32). In the current study, supported the view that direct signal transduction from MjML1’s we identified six ML proteins in kuruma shrimp M. japonicus. recognition of CD subsequently induced Dorsal translocation in- Through a phylogenetic analysis of the vertebrate ML family to the nucleus to regulate the expression of the antiviral effector members, we preliminarily classified these six proteins as MD-2– and MjVago5 (Fig. 9). MD-1–like or NPC-2–like, based on the sequence information.

Discussion Vertebrate ML proteins can be clearly divided into three groups based on their sequence similarity, with the following proteins as the representative of each group: MD-2, NPC2, and GM2A. Ac- cordingly, these three groups exert different functions in innate immune recognition and lipid metabolism (1). Compared with that in vertebrates, the ML family exhibited signiﬁcant expansion in arthropods. Taking insects as examples, the A. aegypti genome encodes 24 ML proteins, the Anopheles gambiae genome encodes 13 ML proteins, and the Drosophila melanogaster genome encodes 8 ML family members (30). However, because of the high sequence variation of these proteins, it is difﬁcult to separate the majority of them into distinct clusters like those in vertebrates. Moreover, most invertebrate ML proteins have not been func- FIGURE 8. Regulation of MjVago5 transcription by Dorsal. (A) Anal- 9 tionally characterized, and thus, it might not be appropriate to ysis of the MjVago5 promoter. The 5 untranslated region of MjVago5 was cloned and analyzed using the online Promoter Scan tool. (B) A ChIP assay simply name a certain ML protein only after its sequence homo- was used to check the binding of Dorsal to the MjVago5 promoter. Shrimp logue in mammals. Until now, only a few invertebrate ML proteins were stimulated with CD (1 mg) or DMSO. The hemocytes were collected have been studied functionally. For example, among the eight ML 6 h later as the pool for ChIP using a ChIP Assay Kit (Beyotime Bio- proteins in D. melanogaster, Npc2a showed the highest sequence technology). Immunoprecipitates were used for RT-PCR with the primers similarity with mammalian NPC2, and mutation of Npc2a led designed against the MjVago5 promoter. The data are representative of two to abnormal cellular sterol distribution (31). The participation independent experiments. 1140 MD-2 HOMOLOGUE RECOGNIZES VIRUS LIPID IN SHRIMP

WSSV lipid component and regulates antiviral molecule expression sheds new light on the mechanism of MD-2–like proteins in host–virus interaction. Among the limited reports of the relevance of MD-2–like proteins to virus infection, only one study found that human MD-2 could act as a receptor for the RSV F protein. By interacting with the RSV F protein in the hydrophobic pocket, MD-2 is essential for F protein–activated TLR4–NF-kB signaling (17). Our finding that MjML1 could directly bind WSSV lipids expanded the binding spectrum of MD-2–like proteins when acting as recognition proteins for viruses. To the best of our knowledge, this is the first report that the ML family directly recognizes the lipid components of viral pathogens. Arthropod Vago proteins are regarded as functionally similar to vertebrate IFNs. Vago proteins contain a single von Willebrand factor type C domain and are secreted. The Vagos from D. melanogaster and Culex quinquefasciatus are induced by Drosophila C virus and West Nile virus (WNV), respectively, and control viral loads postinfection (28, 38). Moreover, C. quinquefasciatus Vago re- FIGURE 9. Model of the MjML1-mediated antiviral mechanism. MjML1 stricts viral infection by activating the JAK/STAT pathway to recognizes WSSV by binding to its envelope lipid component, CD. MjML1 Downloaded from upregulate STAT-dependent virus-inducible molecules in neigh- then transduces the viral signal to Toll4, leading to canonical TLR pathway activation and nuclear translocation of the NF-kB transcription factor, Dorsal. boring cells. This IFN-like function was also proved in the white Dorsal transcriptionally regulates the expression and secretion of the antiviral shrimp, L. vannamei (29). The results of the current study sup- molecule Vago, which activates the JAK/STAT pathway in neighboring cells. ported the importance of Vago in the antiviral response in kuruma shrimp. In addition, the finding that MjML1 induces shrimp Vago5 expression through the NF-kB transcription factor Dorsal This classification provided a clue for the subsequent functional increased our knowledge of the regulation of Vago expression. http://www.jimmunol.org/ analysis, which could in turn help to validate the classification. A previous study reported that induction of Culex Vago by WNV Previous studies (21) and J. Gao, J.X. Wang, and X.W. Wang is dependent on Dicer-2, which is phylogenetically related to (unpublished data) supported the view that the ML family is im- vertebrate RIG-I–like receptors and might sense viral nucleotides. portant in WSSV infection. ML proteins could bind lipid ligands; Recognition of WNV by Dicer-2 leads to the activation of TRAF therefore, two hypotheses of how shrimp ML proteins participate and the cleavage and nuclear translocation of Rel2, another NF-kB in WSSV infection could be proposed, referring to the data ob- ortholog. Rel2 then binds the NF-kB site in the promoter of Vago, tained for the mammalian ML family and for WSSV pathogenesis. and this binding is essential for Vago induction (39). In the current First, the WSSV envelope originates from host cell nuclei; how- study, kuruma shrimp Vago5 transcription was also mediated by by guest on October 1, 2021 ever, the lipid compositions and contents between them are NF-kB. However, the signal transduction upstream of NF-kB somewhat different. In particular, there are more species of cho- might be different in mosquito and shrimp. Both MD-2 and NF-kB lesterol and hydrocarbons in the WSSV envelope than in that of factor are essential members of TLR pathway; therefore, it could be the host nuclei (27). Thus, the lipid material present in WSSV but hypothesized that the recognition of viral lipids by MjML1 leads to absent in the host membrane might be recognized as a nonself the activation of the shrimp TLR pathway. A recent study showed target by ML proteins, which play an immune recognition func- that Toll4 in L. vannamei is critical for the defense against WSSV tion in this case. Second, WSSV enters host cells in a cholesterol- by inducing the nuclear translocation and phosphorylation of Dor- dependent manner (33, 34). The cholesterol-enriched lipid raft sal, which then transcriptionally regulates the expression of several might provide a suitable microenvironment for virus attachment antiviral and antimicrobial peptides (40). However, the recognition and endocytosis (35). Thus, ML proteins might regulate WSSV events upstream of Toll4 remain unclear. Based on this information, infection by influencing the transport of cholesterol. In the current a possible model could be proposed. MjML1 acts as a cofactor of study, we proved that MjML1 functioned in the antiviral response Toll4 in kuruma shrimp. By recognizing the lipid component of the by recognizing the lipid component of WSSV. This MD-2–like WSSV envelope, MjML1 senses viral infection and transduces the immune recognition function was matched with the classification viral signal to Toll4. Through the canonical signaling of the TLR based on the sequence information. In addition, another WSSV- pathway, Dorsal is activated to transcriptionally regulate the ex- inducible ML protein, MjML3, also exerted an antiviral function. pression of the antiviral molecule Vago5 (Fig. 9). The results of the MjML3 is an NPC2-like molecule; therefore, further study should current study revealed the significance of the ML family in antiviral be performed to check whether MjML3 influences the WSSV response in kuruma shrimp. The identification of the CD–MjML1– infection by regulating lipid metabolism similarly to NPC2 in Dorsal–Vago signaling pathway provided new insights into inver- mammals. tebrate antiviral immunity. Sequence analysis suggested that MjML1 adopted an overall canonical MD-2–like structure, with the conservation of mul- Disclosures b tiple strands. Several studies of the structures of MD-2 and The authors have no financial conflicts of interest. MD-1 revealed that the lipid-binding site was located between b strands G and H (36, 37). We identified the lipid-binding site of MjML1 based on sequence alignment and successfully References proved its interaction with lipids. Among MD-2–like proteins, 1. Inohara, N., and G. Nun˜ez. 2002. ML -- a conserved domain involved in innate there is a pronounced sequence variation of the lipid-binding immunity and lipid metabolism. Trends Biochem. Sci. 27: 219–221. 2. Shimazu, R., S. Akashi, H. Ogata, Y. Nagai, K. Fukudome, K. Miyake, and sites, suggesting structural plasticity of the binding pocket M. Kimoto. 1999. MD-2, a molecule that confers lipopolysaccharide respon- to host lipid ligands. The discovery that MjML1 recognizes a siveness on Toll-like receptor 4. J. Exp. Med. 189: 1777–1782. The Journal of Immunology 1141

3. Kim, H. M., B. S. Park, J. I. Kim, S. E. Kim, J. Lee, S. C. Oh, P. Enkhbayar, 22. Wang, S., X. F. Zhao, and J. X. Wang. 2009. Molecular cloning and character- N. Matsushima, H. Lee, O. J. Yoo, and J. O. Lee. 2007. Crystal structure of the ization of the translationally controlled tumor protein from Fenneropenaeus TLR4-MD-2 complex with bound endotoxin antagonist Eritoran. Cell 130: 906–917. chinensis. Mol. Biol. Rep. 36: 1683–1693. 4. Ohto, U., K. Fukase, K. Miyake, and T. Shimizu. 2012. Structural basis of 23. Reichelt, P., C. Schwarz, and M. Donzeau. 2006. Single step protocol to purify species-specific endotoxin sensing by innate immune receptor TLR4/MD-2. recombinant proteins with low endotoxin contents. Protein Expr. Purif. 46: 483– Proc. Natl. Acad. Sci. USA 109: 7421–7426. 488. 5. Park, B. S., D. H. Song, H. M. Kim, B. S. Choi, H. Lee, and J. O. Lee. 2009. The 24. Wang, X. W., Y. H. Xu, J. D. Xu, X. F. Zhao, and J. X. Wang. 2014. Collabo- structural basis of lipopolysaccharide recognition by the TLR4-MD-2 complex. ration between a soluble C-type lectin and calreticulin facilitates white spot Nature 458: 1191–1195. syndrome virus infection in shrimp. J. Immunol. 193: 2106–2117. 6. Lu, Y. C., W. C. Yeh, and P. S. Ohashi. 2008. LPS/TLR4 signal transduction 25. Sun, J. J., J. F. Lan, X. Z. Shi, M. C. Yang, G. J. Niu, D. Ding, X. F. Zhao, pathway. Cytokine 42: 145–151. X. Q. Yu, and J. X. Wang. 2016. b-Arrestins negatively regulate the toll pathway 7. Miura, Y., R. Shimazu, K. Miyake, S. Akashi, H. Ogata, Y. Yamashita, in shrimp by preventing dorsal translocation and inhibiting dorsal transcriptional Y. Narisawa, and M. Kimoto. 1998. RP105 is associated with MD-1 and activity. J. Biol. Chem. 291: 7488–7504. transmits an activation signal in human B cells. Blood 92: 2815–2822. 26. Wang, X. W., W. T. Xu, X. W. Zhang, X. F. Zhao, X. Q. Yu, and J. X. Wang. 8. Miyake, K., R. Shimazu, J. Kondo, T. Niki, S. Akashi, H. Ogata, Y. Yamashita, 2009. A C-type lectin is involved in the innate immune response of Chinese Y. Miura, and M. Kimoto. 1998. Mouse MD-1, a molecule that is physically white shrimp. Fish Shellfish Immunol. 27: 556–562. associated with RP105 and positively regulates its expression. J. Immunol. 161: 27. Zhou, Q., H. Li, Y. P. Qi, and F. Yang. 2008. Lipid of white-spot syndrome virus 1348–1353. originating from host-cell nuclei. J. Gen. Virol. 89: 2909–2914. 9. Nagai, Y., T. Kobayashi, Y. Motoi, K. Ishiguro, S. Akashi, S. Saitoh, 28. Deddouche, S., N. Matt, A. Budd, S. Mueller, C. Kemp, D. Galiana-Arnoux, Y. Kusumoto, T. Kaisho, S. Akira, M. Matsumoto, et al. 2005. The radiopro- C. Dostert, C. Antoniewski, J. A. Hoffmann, and J. L. Imler. 2008. The DExD/H- tective 105/MD-1 complex links TLR2 and TLR4/MD-2 in antibody response to box helicase Dicer-2 mediates the induction of antiviral activity in drosophila. microbial membranes. J. Immunol. 174: 7043–7049. Nat. Immunol. 9: 1425–1432. 10. Yazawa, N., M. Fujimoto, S. Sato, K. Miyake, N. Asano, Y. Nagai, O. Takeuchi, 29. Li, C., H. Li, Y. Chen, Y. Chen, S. Wang, S. P. Weng, X. Xu, and J. He. 2015. K. Takeda, H. Okochi, S. Akira, et al. 2003. CD19 regulates innate immunity by Activation of Vago by interferon regulatory factor (IRF) suggests an interferon the toll-like receptor RP105 signaling in B lymphocytes. Blood 102: 1374–1380. system-like antiviral mechanism in shrimp. Sci. Rep. 5: 15078. 11. Naureckiene, S., D. E. Sleat, H. Lackland, A. Fensom, M. T. Vanier, R. Wattiaux, 30. Waterhouse, R. M., E. V. Kriventseva, S. Meister, Z. Xi, K. S. Alvarez, Downloaded from M. Jadot, and P. Lobel. 2000. Identification of HE1 as the second gene of L. C. Bartholomay, C. Barillas-Mury, G. Bian, S. Blandin, B. M. Christensen, Niemann-Pick C disease. Science 290: 2298–2301. et al. 2007. Evolutionary dynamics of immune-related genes and pathways in 12. Storch, J., and Z. Xu. 2009. Niemann-Pick C2 (NPC2) and intracellular cho- disease-vector mosquitoes. Science 316: 1738–1743. lesterol trafficking. Biochim. Biophys. Acta 1791: 671–678. 31. Huang, X., J. T. Warren, J. Buchanan, L. I. Gilbert, and M. P. Scott. 2007. 13. Friedland, N., H. L. Liou, P. Lobel, and A. M. Stock. 2003. Structure of a Drosophila Niemann-Pick type C-2 genes control sterol homeostasis and steroid cholesterol-binding protein deficient in Niemann-Pick type C2 disease. Proc. biosynthesis: a model of human neurodegenerative disease. Development 134: Natl. Acad. Sci. USA 100: 2512–2517. 3733–3742. 14. Liu, Y., A. Hoffmann, A. Grinberg, H. Westphal, M. P. McDonald, K. M. Miller, 32. Shi, X. Z., X. Zhong, and X. Q. Yu. 2012. Drosophila melanogaster NPC2

J. N. Crawley, K. Sandhoff, K. Suzuki, and R. L. Proia. 1997. Mouse model of proteins bind bacterial cell wall components and may function in immune signal http://www.jimmunol.org/ GM2 activator deﬁciency manifests cerebellar pathology and motor impairment. pathways. Insect Biochem. Mol. Biol. 42: 545–556. Proc. Natl. Acad. Sci. USA 94: 8138–8143. 33. Huang, J., F. Li, J. Wu, and F. Yang. 2015. White spot syndrome virus enters 15. Wright, C. S., Q. Zhao, and F. Rastinejad. 2003. Structural analysis of lipid crayﬁsh hematopoietic tissue cells via clathrin-mediated endocytosis. Virology complexes of GM2-activator protein. J. Mol. Biol. 331: 951–964. 486: 35–43. 16.Wilks,J.,E.Lien,A.N.Jacobson,M.A.Fischbach,N.Qureshi,A.V.Chervonsky, 34. Chen, R. Y., K. L. Shen, Z. Chen, W. W. Fan, X. L. Xie, C. Meng, X. J. Chang, and T. V. Golovkina. 2015. Mammalian lipopolysaccharide receptors incorporated into L. B. Zheng, J. Jeswin, C. H. Li, et al. 2016. White spot syndrome virus entry is the retroviral envelope augment virus transmission. Cell Host Microbe 18: 456–462. dependent on multiple endocytic routes and strongly facilitated by Cq- 17. Rallabhandi, P., R. L. Phillips, M. S. Boukhvalova, L. M. Pletneva, K. A. Shirey, GABARAP in a CME-dependent manner. Sci. Rep. 6: 28694. T. L. Gioannini, J. P. Weiss, J. C. Chow, L. D. Hawkins, S. N. Vogel, and 35. Cossart, P., and A. Helenius. 2014. Endocytosis of viruses and bacteria. Cold J. C. Blanco. 2012. Respiratory syncytial virus fusion protein-induced toll-like Spring Harb. Perspect. Biol. 6: a016972. receptor 4 (TLR4) signaling is inhibited by the TLR4 antagonists Rhodobacter 36. Yoon, S. I., M. Hong, G. W. Han, and I. A. Wilson. 2010. Crystal structure of

sphaeroides lipopolysaccharide and eritoran (E5564) and requires direct inter- soluble MD-1 and its interaction with lipid IVa. Proc. Natl. Acad. Sci. USA 107: by guest on October 1, 2021 action with MD-2. MBio 3: e00218-12. 10990–10995. 18. Coleman, E. M., T. N. Walker, and J. E. Hildreth. 2012. Loss of Niemann Pick type 37. Resman, N., A. Oblak, T. L. Gioannini, J. P. Weiss, and R. Jerala. 2014. Tet- C proteins 1 and 2 greatly enhances HIV infectivity and is associated with accu- raacylated lipid A and paclitaxel-selective activation of TLR4/MD-2 conferred mulation of HIV Gag and cholesterol in late endosomes/lysosomes. Virol. J. 9: 31. through hydrophobic interactions. J. Immunol. 192: 1887–1895. 19. Jupatanakul, N., S. Sim, and G. Dimopoulos. 2014. Aedes aegypti ML and 38.Paradkar,P.N.,L.Trinidad,R.Voysey,J.B.Duchemin,andP.J.Walker. Niemann-Pick type C family members are agonists of dengue virus infection. 2012. Secreted Vago restricts West Nile virus infection in Culex mosquito Dev. Comp. Immunol. 43: 1–9. cells by activating the Jak-STAT pathway. Proc.Natl.Acad.Sci.USA109: 20. Sim, S., J. L. Ramirez, and G. Dimopoulos. 2012. Dengue virus infection 18915–18920. of the Aedes aegypti salivary gland and chemosensory apparatus induces genes 39. Paradkar, P. N., J. B. Duchemin, R. Voysey, and P. J. Walker. 2014. Dicer-2- that modulate infection and blood-feeding behavior. PLoS Pathog. 8: dependent activation of Culex Vago occurs via the TRAF-Rel2 signaling path- e1002631. way. PLoS Negl. Trop. Dis. 8: e2823. 21. Zhao, Z. Y., Z. X. Yin, S. P. Weng, H. J. Guan, S. D. Li, K. Xing, S. M. Chan, and 40. Li, H., B. Yin, S. Wang, Q. Fu, B. Xiao, K. Lǚ, J. He, and C. Li. 2018. RNAi J. G. He. 2007. Profiling of differentially expressed genes in hepatopancreas of screening identifies a new Toll from shrimp Litopenaeus vannamei that restricts white spot syndrome virus-resistant shrimp (Litopenaeus vannamei) by sup- WSSV infection through activating Dorsal to induce antimicrobial peptides. pression subtractive hybridisation. Fish Shellfish Immunol. 22: 520–534. PLoS Pathog. 14: e1007109. Fig. S1

Fig. S1 Multiple alignment of kuruma shrimp ML family and mammalian (human and mouse) MD-2, MD-1 and Npc2. The alignment was performed by online Clustal Omega and presented by GENEDOC software. The conserved residues were shown in black and in grey.