RNAiso Plus, First-Strand cDNA Synthesis Kit, in vitro Transcription T7 Kit and Taq Polymerase were purchased from TaKaRa Biotech (Dalian, China). Ultrapure LPS-EK (tlrl-peklps) were obtained from In vivoGen. Lipoteichoic acid (LTA, from Staphylococcus aureus) were obtained from Sigma (St. Louis, MO, USA). V. parahemolyticus and Vibrio harveyi identified in our laboratory as well as four standard strains Escherichia coli (8099), S. aureus (ATCC 6538), Bacillus megaterium (NBRC 15308) and Bacillus subtilis (ATCC 9372) were used in this study.

Pacific white shrimp P. vannamei (~ 12 g each) were purchased from a shrimp farm in Ganyu County (Lianyungang, Jiangsu, China) were cultured in a cement tank with aerated seawater and fed daily with a commercial diet. The animal experiments were strictly conducted following the rules of the Institutional Animal Care and Use Committee of China. Healthy shrimp were randomly selected to analyze the tissue distribution and expression profiles of PvML1. Shrimp hemolymph was harvested with a sterilized syringe preloaded with ice-cold anticoagulant buffer (0.45 M NaCl, 0.1 M glucose, 30 mM trisodium citrate, 26 mM citric acid, and 10 mM ethylenediaminetetraacetic acid; pH 4.6), and then centrifuged at 850 × g for 15 min at 4°C to isolate hemocytes. Other tissues, including gills, hepatopancreas, intestine, heart, muscle, stomach, and eyestalk, were also dissected, washed with sterile PBS, and pooled from at least five healthy shrimp. All these tissues together with hemocytes were used to isolate total RNA for investigation of tissue distribution. For immune challenge, each shrimp was injected with 100 μL of bacterial inoculum (2 × 10⁶ CFU V. parahemolyticus). The corresponding control was treated with an equal volume of sterile phosphate-buffered saline (PBS) (140 mM NaCl and 10 mM sodium phosphate; pH 7.4). At each time point post injection (0, 3, 6, 12, 24 and 48 h), the total RNA of hemocytes was extracted for investigating the temporal expression pattern of PvML1. The extracted RNA was kept in 75% ethanol at −80°C until needed. Two other batches of RNA samples isolated previously were used to eliminate the differences among batches.

RNAiso Plus reagent was used to extract the total RNA from hemocytes and other collected tissues. DNase I (Promega, USA) was added into the extracted RNA to remove contaminating genomic DNA. The cDNA was synthesized using the total RNA according to the manufacturer’s instructions of First-Strand cDNA Synthesis Kit.

The original cDNA sequence encoding the putative PvML1 was harvested through high-throughput transcriptome sequencing with an RNA mixture extracted from the hemocytes and hepatopancreas of shrimp. This cDNA sequence was further verified by polymerase chain reaction (PCR) with a pair of gene-specific primers (PvML1F and PvML1R, Table 1 ). The PCR was performed under the following parameters: 95°C for 3 min; 35 cycles of 94°C for 30 s, 54°C for 30 s, and 72°C for 30 s; and a final extension for 10 min at 72°C. The targeted DNA fragment was purified, cloned into a pMD-19T vector, and finally sequenced by Sangon Company (Shanghai, China).

The similarities of PvML1 with other ML family proteins were analyzed using the online Basic Local Alignment Search Tool Program (BLASTP) (http://blast.ncbi.nlm.nih.gov/Blast.cgi). The deduced protein sequences were translated and predicted on http://web.expasy.org/translate/. The putative domain was predicted using Simple Modular Architecture Research Tool (SMART) (http://smart.embl-heidelberg.de). Multiple alignment was conducted with the ClustalX 2.0 program (http://www.ebi.ac.uk/tools/clustalw2) and GENEDOC software. The theoretical molecular weight (Mw) and isoelectric point (pI) were calculated on http://web.expasy.org/compute_pi/. Signal peptide was searched with SignalP (34). A neighbor-joining phylogenetic tree was generated with MEGA 7.0 and 1000 bootstraps were used to assess reliability (35). Three-dimensional (3D) model of PvML1-lipid A complex was predicted by docking with BSP-SLIM ONLINE software (https://zhanglab.ccmb.med.umich.edu/BSP-SLIM/) and displayed by PyMOL program. The receptor protein PvML1 was modelled after the crystal structure of human MD2 (PDB ID: 2E59), and (heptosyl)2-Kdo2-lipid A was used as the ligand.

qRT-PCR was carried out to analyze the mRNA expression levels of PvML1 and antimicrobial peptide (AMP) genes in a real-time thermal cycler Quantstudio 6 Flex (ABI, USA) following the protocol in a previous study (36). The gene-specific primers for PvML1 and AMP genes ( Table 1 ) were designed to produce their respective amplicons and analyze their mRNA amounts. The primers for the internal reference gene PvEF1α (elongation factor 1-alpha, Table 1 ) were also synthesized and used to analyze the relative expression levels of PvML1 and AMP genes (37). qRT-PCR was performed in a 20-μL reaction mixture (10 μL of 2 × SYBR Premix Ex Taq, 2 μL of cDNA, and 4 μL of each primer). The reaction procedure was as follows: an initial denaturation step at 95°C for 3 min; 40 cycles at 95°C for 10 s, and 60°C for 40 s; and melting from 60°C to 95°C. The relative expression levels of PvML1 in different tissues as well as AMP genes was calculated with the method of 2^−ΔCT. The algorithm of 2^−ΔΔCT was applied to investigating the time-course profiles of PvML1 (38). All treatments were carried out thrice with individual templates, and the obtained data were subjected to the statistical analysis. Significant differences were assessed by unpaired t-test (*P < 0.05; **P < 0.01).

Recombinant PvML1 as well as the extracellular regions of two PvTolls (PvToll1, ABK58729; PvToll2, AEK86516) was overexpressed with E. coli expression system. Based on PvML1 cDNA sequence, a pair of gene-specific primers (PvML1EF and PvML1ER, Table 1 ) were designed to amplify the DNA fragment (402 bp) encoding PvML1 mature peptide. After digestion with enzymes (EcoR I and Xho I), the fragment was ligated into a pET32a vector to construct recombinant plasmid pET32a-PvML1. Similarly, the DNA sequences encoding the extracellular regions of two Tolls were produced by PCR with two pairs of primers ( Table 1 ); each fragment was digested by restricted enzymes and finally ligated into pGEX-6P-1 vectors. All these plasmids were respectively transformed into E. coli competent cells for over-expressions with isopropyl-β-d-thiogalactoside (IPTG, 0.1 mM). The recombinant PvML1 containing His tag was purified with Ni-NTA His Bind Resin, while the recombinant extracellular regions of Tolls (PvToll1ER and PvToll2ER) with GST tag were purified with glutathione sepharose 4B chromatography (Novagen, USA). Cold 0.1% Triton X-114 was used to remove contaminating endotoxins before collecting the final elution of the proteins from the column. Besides, the empty vectors pET32a and pGEX-6P-1 were also overexpressed in E. coli, and the corresponding vector proteins with thioredoxin (TRX) or GST tag were harvested and used as the negative controls.

Microorganisms, including Gram-negative bacteria (V. parahemolyticus, V. harveyi and E. coli) and Gram-positive bacteria (S. aureus, B. megaterium and B. subtilis), were applied to investigating the microorganism-binding activity of PvML1 using Western blot. The procedure was performed following our earlier study (39). Briefly, microorganisms were cultured in Luria–Bertani (LB) medium for 6 h at 37°C, and then were pelleted by centrifugation. After the pellets were washed thrice with 1 mL of TBS (50 mM Tris–HCl and 150 mM NaCl; pH 7.5), the microorganisms (1 × 10⁸ CFU) were incubated in 200 μL of rPvML1 (200 μg/mL) for 1 h at room temperature. Afterwards, they were pelleted, washed thrice with TBS, and eluted with 7% SDS by mild agitation for 5–10 min. The supernatants (eluates) were collected through centrifugation and the final pellets were harvested after three more washes with TBS. Both the eluates and the final pellets were subjected to 15% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). rPvML1 was also sampled as the positive control. After separation with SDS-PAGE, the protein samples were transferred onto a nitrocellulose membrane. The membrane was blocked by 5% non-fat milk in TBS and then incubated with peroxidase-conjugated mouse monoclonal antibody against His-tag for 2 h. rPvML1 signal was visualized with an ECL Western blot detection reagent kit.

Gram-negative bacteria were chosen to investigate the agglutinating activity of PvML1. The agglutination assay was performed in accordance with the method described by Du. et al. (40). Bacteria cultured in LB broth were harvested at mid-logarithmic phase by centrifugation at 5000 × g for 5 min, washed three times with TBS, and then resuspended in TBS (2 × 10⁸ cells mL^–1). The bacterial suspensions were incubated with equal volume (30 µL) of diluted rPvML1 in TBS at the protein concentration range of 0.8-5 µM with or without 10 mM CaCl₂ at 28°C for 1 h. TRX tag protein (200 µg/mL) was used as the negative control. Agglutination was determined by observing under a light microscope. The minimal agglutinating concentration (MAC) is defined as the lowest protein concentration yielding visible microbial agglutination compared with the negative control.

ELISA was carried out to investigate the binding activities of PvML1 to microbial polysaccharides and PvTolls. Medium-binding microtiter plates (Greiner) were used to test the binding activity of PvML1 to microbial polysaccharides following a previous method (41). In brief, the plate wells were incubated with a total of 100 µL of LPS, Lipid A, or LTA (20 µg/mL) at 37°C overnight until the plate came to desiccation. Wells serving as the blank control were incubated with 100 µL of distilled water. After blocked with 200 μL of BSA (2 mg/mL) for 2 h and washed four times with TBST (0.05% Tween-20 in TBS), the wells were incubated with serially diluted recombinant PvML1 or TRX tag protein (negative control) (0.0005-1 μM in TBS containing 0.1 mg/mL BSA) at 37°C for 3 h and then rinsed five times with TBST. Each well was then incubated with 100 µL of peroxidase-conjugated mouse monoclonal anti-His antibody (1:5000 dilution in TBS with 1 mg/mL BSA). The color reaction was developed with 0.01% 3,3’,5,5’-tetramethylbenzidine (Sigma) and stopped with 2 M H₂SO₄. The absorbance was recorded at 405 nm by a microtiter plate reader (Tecan, Switzerland). In addition, high-binding microtiter plates (Greiner) were applied to investigating the binding function of PvML1 to PvTolls. The plates were pre-incubated with a total of 100 µL of rPvtoll1ER, rPvtoll2ER or GST (200 µg/mL) at 37°C for 2 h. After blocking with BSA and washing with TBST, serially diluted recombinant PvML1 or TRX tag protein (0.0005-1 μM in TBS containing 0.1 mg/mL BSA) was added to the plates. The color reaction was performed with the same procedure as the above, and the absorbance was obtained in the same way. All assays were performed in triplicate.

A partial DNA fragment of PvML1 was amplified using primers containing a T7 promoter (PvML1iF and PvML1iR, Table 1 ). The harvested PCR product was used as the template to synthesize dsPvML1 (PvML1 dsRNA) with an in vitro Transcription T7 Kit. The dsEGFP (EGFP dsRNA) was also synthesized as negative control with primers listed in Table 1 . The healthy shrimp (~ 8 g each) were randomly divided into two groups (six shrimp in each group). Each shrimp was intramuscularly injected with 8 μg of dsPvML1 or dsEGFP into the fourth abdominal segment. A second dsRNA injection was conducted 24 h later in the same manner. At 48 h after the first dsRNA injection, hemocytes was collected for total RNA extraction, which was used to assess RNAi efficiency by qRT-PCR. Experiments were performed independently thrice. Significant differences were analyzed with unpaired t-test (*P < 0.05; **P < 0.01).

After validating that PvML1 expression could be silenced by injection of dsPvML1, we examined whether the knockdown of PvML1 could affect bacterial clearance. V. parahemolyticus at the mid-logarithmic growth phase was collected by centrifugation and re-suspended in PBS (2 × 10⁷ CFU/mL) after washing three times. Each shrimp was injected with 100 μL of bacterial suspension at 48 h after injection with dsPvML1 or dsEGFP. After mock injection with PBS, the shrimp were treated with an equal number of bacteria in the same way. At 40 min after bacterial injection, hemolymph (100 μL) was collected from shrimp and mixed with an equal volume of anticoagulant buffer. After serial dilution with PBS, the diluted hemolymph (50 μL) was smeared onto the LB plates. The plates were then incubated at 37°C until bacterial clones appeared. The number of residual bacteria in hemolymph was determined by counting the number of bacterial clones on the plates. In addition, to further confirm whether coating bacteria with PvML1 could facilitate bacterial clearance, V. parahemolyticus incubated with recombinant PvML1 or TRX tag protein was injected into shrimp following a method with slight modifications (42). Shrimp were randomly divided into two groups. Approximately 600 μL of rPvML1 or TRX tag protein in PBS (400 μg/mL) was mixed with an equal volume of bacterial suspension (2 × 10⁷ CFU/mL) with gentle rotation at room temperature for 15 min. TRX tag protein served as the control. After incubation, each shrimp was injected with 100 μL of mixture. The number of residual bacteria in hemolymph was calculated using the same method as described above. Unpaired student’s t-test was used to assess the significant differences. (*P < 0.05; **P < 0.01).

To investigate whether PvML1 knockdown can affect the expression of AMPs in shrimp, ten different AMPs expressed in hemocytes were selected as representatives to assess the effectiveness caused by the decrease of PvML1 expression. These AMPs are from three different AMP families: anti-lipopolysaccharide factors (ALFs), penaeidins (PENs) and crustins (Crus) (43). At 48 h after dsRNA (dsPvML1 or dsEGFP) injection, the total RNAs of hemocytes were extracted, and cDNAs were synthesized as the templates for qRT-PCR. The gene-specific primers for AMPs were listed in Table 1 . Unpaired t-test was used to analyze significant differences (*P < 0.05; **P < 0.01).

Survival rate assay was conducted to investigate the effect of PvML1 knockdown on host antibacterial immunity. Shrimp (~ 8 g each) were infected with 100 μL of V. parahemolyticus(1 × 10⁷ CFU)at 48 h after first dsRNA (dsPvML1 or dsEGFP) injection. Shrimp received two times of PBS injection served as blank control. The numbers of dead animals were recorded from 3 h to 24 h after bacterial injection, by which the survival percentage was determined. In addition, a total volume of 100 μL V. parahemolyticus (1 × 10⁷ CFU) pre-incubated with rPvML1 or with TRX tag protein in PBS was injected into shrimp to calculate the survival rates. Blank control was treated with an equal volume of PBS. A total of 30 shrimp was randomly selected for each group. The statistical analysis was conducted using Log-rank (Mantel-Cox) test.

Two PvTolls (PvToll1 and PvToll2) with higher similarities to human TLR4 were chosen to analyze the potential interactions between PvML1 and PvTolls by conducting GST pull-down assays according to a documented method with slight modifications (40). A total of 150 μL glutathione-Sepharose 4B resin (50% bead slurry) after wash three times with PBS was incubated with a mixture of a His-tagged protein (15 μg, rPvML1) and a GST-tagged protein (15 μg, rPvToll1ER, rPvToll2ER or GST) for 2 h at 4°C. The GST tag protein served as negative control. After incubation, the beads were washed thoroughly with PBS, and then proteins were eluted by adding PBS containing 10 mM reduced glutathione. The final washes and resultant eluates as well as the recombinant proteins, including rPvML1, rPvToll1ER and rPvToll2ER, were subjected to a 12.5% SDS-PAGE. The results were analyzed after the gel was stained with Coomassie blue.

Based on the cDNA sequences of PvToll1 and PvToll2, two pairs of gene-specific primers ( Table 1 ) were designed to amplify the DNA sequences of the extracellular region of these two PvTolls. Either the harvested DNA fragments or pcDNA3.1-myc-his-A (pcDNA3.1) vector were digested, and the targeted fragments were then ligated into a pcDNA3.1 vector to generate expression plasmids with the sequences of truncated PvTolls (pcDNA3.1-PvToll1ER and pcDNA3.1-PvToll2ER). Besides, the DNA sequence encoding the mature peptide of PvML1 was amplified with the specific primers in Table 1 , and then ligated into pcDNA3.1-EGFP to produce a recombinant plasmid pcDNA3.1-PvML1-EGFP. HEK 293T cells were cultured in high-glucose DMEM medium (Gibco) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 µg/ml streptomycin, in humidified 5% CO₂ and 95% air at 37 °C. For transient transfection, cells were seeded into 6-well microtiter plates and incubated overnight. When cells were ~ 70% confluent, the cells were co-transfected with 2 μg of His-tagged expression plasmid (pcDNA3.1-PvToll1ER or pcDNA3.1-PvToll2ER) and 2 μg of EGFP-tagged expression plasmid pcDNA3.1-PvML1-EGFP. At 36 h after transfection, the cells were lysed with NP40 lysate (Beyotime) and then centrifuged at 12,000 rpm for 20 min at 4°C, and the supernatants were incubated with anti-cGFP antibody (or anti-cMyc antibody) and Protein A + G Agarose beads overnight at 4°C with rotation. Normal rabbit IgG was used as the negative control. The beads were collected by centrifugation, washed three times with PBS, and then resuspended in 1 × SDS sample buffer. After boiling for 10 min, the resultant samples were separated by SDS-PAGE and then were analyzed by Western blot.

The complete cDNA sequence of PvML1 had 652 bp, including a 144-bp 5′ untranslated region, a 468-bp open reading frame for encoding a 155-amino acid (aa) polypeptide, and a 3′ noncoding region of 40 bp (GenBank Accession No. MN604018) ( Figure S1 ). A signal peptide of 22 residues at the N-terminus and a ML domain (29–152 aa) were found in deduced protein. The ML domain contained six conserved cysteines that may form three disulfide bonds to stabilize the overall structure. The domain architecture of PvML1 was schematically shown in Figure 1A . The mature peptide of PvML1 had an estimated Mw of 15.5 kDa and a theoretical pI of 7.85.

BLASTP search analysis showed that PvML1 shared the highest similarity (78.34%) with an ML domain-containing protein Penaeus japonicus PjML1 (QDX01882), but the similarity did not go beyond 40% with other ML proteins. For instance, PvML1 had 32.69% similarity with Scylla paramamosain SpMD2 (MK109797), 30.97% with PjML4 (QDX01885), 16.03% with P. vannamei PvML (ABD65303), and 15.85% with Homo sapiens MD2 (BAA78717). The similarities among ML proteins were further revealed by the alignment of representative ML protein sequences from different species. However, low similarities were observed among the ML protein sequences, except for the five highly conserved cysteine residues present in each ML protein ( Figure 1B ). Earlier studies revealed that four of the five cysteine residues located in similar positions of ML proteins could form two disulfide bonds, which were responsible for maintaining the overall structure and biological functions of the ML proteins. PvML1 and other representative ML proteins presented four cysteine residues that were involved in disulfide-bond formation, suggesting that these ML proteins may have a similar 3D structure.

The evolutionary relationship between PvML1 and other ML proteins was analyzed by constructing a phylogenetic tree. In this tree, the vertebrate ML proteins were separated into three different meaningful clusters: NPC2, GM2A, and MD2 and MD1. The crustacean ML proteins were grouped into two of the three aforementioned clusters ( Figure 1C ). PvML1, PjML1, PjML4, SpMD2, and HaML were clustered together with the vertebrate MD2 and MD1 homologs, whereas the other crustacean ML homologs were grouped together with the NPC2 homologs. PvML1 presented a much closer phylogenetic relationship with the vertebrate MD2 homologs, suggesting that it may possess similar immune function to human MD2 because the latter is an essential immune component of the human TLR4 signaling pathway.

Docking was performed with the receptor protein PvML1 and the ligand of lipid A (lipid portion of LPS) to determine whether PvML1 possessed LPS-binding potentials. The 3D model of the PvML1–lipid A complex with the highest docking score (5.981) was shown in Figures 1D, E . The PvML1 molecule contained a deep hydrophobic cavity, and lipid A was properly accommodated in it. This formation was similar to the HsMD2–lipid A complex, which attained a slightly higher docking score of 7.510 (not shown).

qRT-PCR was conducted to investigate the tissue distribution and time-course expression profile of PvML1 after bacterial challenge. As shown in Figure 2A , PvML1 was expressed in all tested tissues, and its relative expression level in hemocytes was much higher than those in the other tissues (gills, hepatopancreas, intestine, heart, muscle, stomach, and eyestalk). PvML1 was highly expressed in hemocytes, suggesting its important role in the open circulating system of shrimp. Then, the temporal expression profile of PvML1 in hemocytes after the bacterial challenge was further investigated. PvML1 was significantly increased 6–24 h after it was challenged with V. parahemolyticus and reached the highest expression level (nearly a sixfold increase) at 12 h post-injection ( Figure 2B ). This result revealed that PvML1 was an immune component involved in the antibacterial response of shrimp.

PvML1 and two truncated PvTolls were successfully expressed and purified. The recombinant PvML1 was expressed as a TRX-tagged fusion protein with a predicted Mw of 34.5 kDa (including the ~19 kDa TRX tag). Meanwhile, rPvToll1ER and rPvToll2ER were GST-tagged fusion proteins with predicted Mw values of 104.2 and 113.3, respectively (including the ~26 kDa GST tag). The position of each purified protein was roughly in agreement with the Mw of the corresponding recombinant protein ( Figure 3A ).

Western blot was performed to examine the microbial cell-binding ability of PvML1. rPvML1 was detected only in the eluate, suggesting its weak binding to the microorganisms. However, this protein was found in pellets, indicating strong binding ability. According to this standard, PvML1 exhibited a strong binding activity to Gram-negative bacteria (V. harveyi, V. parahemolyticus, and E. coli) ( Figure 3B ). PvML1 also displayed weak binding to other tested microorganisms (S. aureus, B. subtilis, and B. megaterium). The results suggest that PvML1 may act as a potential recognition protein for certain kinds of pathogens, especially Gram-negative bacteria. Then, the agglutination activities of PvML1 to microbes were investigated, especially since some immune components agglutinate pathogens via their microbial cell-binding activities. PvML1 exhibited remarkable agglutination to V. parahemolyticus cells in the presence of Ca²⁺ ( Figures 3C–E ). It also displayed agglutinating activities to E. coli and V. harveyi, and the agglutinating activity to the former is much stronger than the latter ( Table 2 ). These results further demonstrated that PvML1 could specifically interact with certain components on the surface of Gram-negative bacteria.

Considering that PvML1 could bind to the aforementioned microbes, certain components on the microbial cell surface might be recognized by PvML1. Furthermore, as most ML family members were determined as lipid-binding proteins, the common bacterial PAMPs with a lipid portion, such as LPS and its lipid portion (lipid A) and LTA, were selected and then applied to an ELISA. As shown in Figures 3F–H , rPvML1 could bind to both LPS and lipid A in a concentration-dependent manner within a certain concentration range. However, rPvML1 did not exhibit a significant binding activity to LTA. By contrast, the TRX tag protein exhibited much lower binding activities to both LPS and lipid A, although it could also interact with them. These results revealed the specific binding activity of PvML1 to LPS and lipid A, and its binding activity to LPS was largely contributed by the binding to lipid A. Thus, we speculate that LPS may be the key recognition site on the surface of Gram-negative bacteria, which can be sensed by binding to lipid A.

Survival assay was performed to investigate the in vivo function of PvML1 by using V. parahemolyticus cells that were pre-incubated with rPvML1 or TRX protein. After the bacterial cells were injected into shrimp, the rPvML1 significantly enhanced the shrimp resistance against bacterial infection. As shown in Figure 4A , the survival rate of the rPvML1-treated group was always higher than that of the control group from 6 h after bacterial injection. The survival percentage of the control group was approximately 50% at 15 h after infection, whereas more than 70% shrimp were alive in the experimental group at that time. The abovementioned results confirmed the role of PvML1 in host immunity to protect shrimp from bacterial infection. To further explore whether PvML1 could facilitate bacteria clearance. V. parahemolyticus cells pre-incubated with rPvML1 or TRX protein were injected into healthy shrimp. In contrast to the findings involving TRX treatment, the number of bacteria in the hemolymph was significantly decreased 40 min after the injection with rPvML1-incubated bacteria, demonstrating that pre-incubating bacteria with rPvML1 could facilitate bacterial clearance in vivo ( Figure 4B ). Taken together, these results indicate that the increased survival rate of shrimp may be attributed to the promoted bacterial clearance in hemolymph by PvML1.

RNAi of PvML1 and survival assays were conducted to investigate the in vivo function of PvML1. qRT-PCR analysis showed a considerable downregulation of the expression level of PvML1 in the hemocytes 40 and 48 h after the first injection of dsPvML1, whereas those of the remaining transcripts of PvML1 in the dsPvML1-injected group did not exceed 20% of those in the control group at each time point ( Figure 5A ). This result indicated that injecting dsPvML1 into shrimp could dramatically suppress PvML1 expression. After PvML1 knockdown, V. parahemolyticus was injected into shrimp, and the survival rates in different groups were calculated. In this manner, the antibacterial ability could be evaluated. As shown in Figure 5B , knockdown of PvML1 dramatically suppressed host’s immune function against bacteria. The survival percentage of dsPvML1-treated shrimp 15 h after bacterial infection did not exceed 20% in the experimental group, whereas approximately 50% of dsEGFP-treated shrimp was still alive in the control group. This result suggests that PvML1 may function as an important antibacterial component in shrimp.

Bacterial clearance assays were conducted after PvML1 knockdown to investigate the immune function of PvML1. After validating that PvML1 expression could be knocked down, V. parahemolyticus was injected into the dsRNA-treated shrimp. Then, the residual bacterial number in the hemolymph was counted to determine the bacterial clearance ability. As shown in Figure 5C , the number of residual bacteria in hemolymph significantly increased 40 min post-injection compared with that in the dsEGFP-treated group. The results showed that PvML1 knockdown remarkably suppressed bacterial clearance.

Aimed at determining whether the presence of AMPs in the hemolymph were relevant to bacterial clearance, the expression level of AMPs in the hemocytes of shrimp 48 h after injection with dsPvML1 or dsEGFP were investigated via qRT-PCR. The results showed the transcripts of PvALF2, PvALF4, PvPEN2, PvPEN3, PvPEN4, and PvCrus1 were significantly decreased in PvML1-silenced shrimps compared with those in the control group. PvALF1, PvCrus2, and PvCrus3 were significantly increased, and no significant changes were observed in the expressions of PvALF3 ( Figure 5D ). These results demonstrated that PvML1 knockdown significantly suppressed the expression of certain AMPs. The combined results suggested that the low expression of certain AMPs may be responsible for the decreased bacterial clearance ability caused by PvML1 knockdown.

To investigate whether PvML1 could interact with certain Toll homologs similar to human MD2 in the TLR4 signaling pathway, rPvML1 and rPvToll1ER and rPvToll2ER were prepared for the GST-pull down assay. The results were shown in Figures 6B–D . rPvML1 displayed apparent binding activities to both rPvToll1ER and rPvToll2ER but not to the GST tag protein. The interaction of rPvML1 with both rPvToll1ER and rPvToll2ER were further verified by ELISA ( Figures 6E–G ). The rPvML1 exhibited strong binding abilities to both rPvToll2ER and rPvToll1ER, and the binding activity of rPvML1 to rPvToll2ER was slightly stronger than that to rPvToll1ER. Furthermore, rPvML1 pre-incubated with LPS had a stronger binding activity to rPvToll2ER than that to rPvToll1ER. Besides conducting interaction assays with recombinant proteins, a co-immunoprecipitation assay was also performed by co-infecting PvML1 and the truncated PvTolls into HEK-293T cells ( Figures 6H, I ). We found that PvML1 specifically interacted with PvToll2ER as well as PvToll1ER. These results suggest that PvML1 may act as an accessory recognition protein for LPS in PvToll2 signaling pathway.