Immune mediation of HMG-like DSP1 via Toll-Spätzle pathway and its specific inhibition by salicylic acid analogs

Abstract

Xenorhabdus hominickii, an entomopathogenic bacterium, inhibits eicosanoid biosynthesis of target insects to suppress their immune responses by inhibiting phospholipase A2 (PLA2) through binding to a damage-associated molecular pattern (DAMP) molecule called dorsal switch protein 1 (DSP1) from Spodoptera exigua, a lepidopteran insect. However, the signalling pathway between DSP1 and PLA2 remains unknown. The objective of this study was to determine whether DSP1 could activate Toll immune signalling pathway to activate PLA2 activation and whether X. hominickii metabolites could inhibit DSP1 to shutdown eicosanoid biosynthesis. Toll-Spätzle (Spz) signalling pathway includes two Spz (SeSpz1 and SeSpz2) and 10 Toll receptors (SeToll1-10) in S. exigua. Loss-of-function approach using RNA interference showed that SeSpz1 and SeToll9 played crucial roles in connecting DSP1 mediation to activate PLA2. Furthermore, a deletion mutant against SeToll9 using CRISPR/Cas9 abolished DSP1 mediation and induced significant immunosuppression. Organic extracts of X. hominickii culture broth could bind to DSP1 at a low micromolar range. Subsequent sequential fractionations along with binding assays led to the identification of seven potent compounds including 3-ethoxy-4-methoxyphenol (EMP). EMP could bind to DSP1 and prevent its translocation to plasma in response to bacterial challenge and suppress the up-regulation of PLA2 activity. These results suggest that X. hominickii inhibits DSP1 and prevents its DAMP role in activating Toll immune signalling pathway including PLA2 activation, leading to significant immunosuppression of target insects.

Fig 1. Secretion of Se-DSP1 from nuclei to plasma upon infection to Gram-positive bacterium, E. mundtii (‘Em’, 4 × 10⁵ cells/larva), in S. exigua.

(A) An immunofluorescence assay of Se-DSP1 in fat body at 6 h after bacterial injection. F-actin and nucleus were stained with phalloidin and DAPI, respectively. Se-DSP1 was detected with its polyclonal antibody. (B) Western blotting analysis of Se-DSP1 in the plasma of naïve or Em-challenged larvae. Plasma samples were collected at 6 h after bacterial challenge. Each lane was loaded with 5 μL of plasma. (C) Exosome analysis for secreted Se-DSP1 in the plasma using western blotting against CD9 (an exosome-specific protein) and Se-DSP1. (Left panel) Exosomes isolated from plasma of naïve or Em-challenged larvae. (Right panel) Plasma samples of Em-challenged larvae before and after exosome extraction. Each lane was loaded with 20 μg proteins. Coomassie-staining bands against a larval storage protein (‘LSP’) indicated the same amount of protein loading for plasma samples. A cytoskeletal protein, α-tubulin, was detected by western blotting to indicate the same amount of protein loading in exosome analysis.

Fig 2. Immune mediation of Se-DSP1 in S. exigua.

rSe-DSP1 (‘DSP1’) was injected into L5 larvae at a dose of 0.8 μg/larva. Inactivation of DSP1 used heat treatment at 95°C for 10 min. Immune challenge used an injection of E. mundtii (‘Em’) to L5 larvae at a dose of 4 × 10⁵ cells/larva. At 8 h PI, hemolymph and fat body were collected. Hemolymph was used for analyzing activities of phenoloxidase (‘PO’) and sPLA₂. Fat body was used for analyzing cPLA₂ activity. For antimicrobial peptide (‘AMP’) analysis, fat body from treated larvae was collected at 12 h PI. (A) Up-regulation of PO activity by Se-DSP1. RNAi used injection of gene-specific dsRNA against Se-DSP1 (dsDSP1) at a dose of 1 μg/larva. At 24 h PI, Em was used for treatment. Control dsRNA (dsCON) used dsRNA specific to a viral gene, CpBV302. (B) Up-regulation of PLA₂ activity by Se-DSP1. (C) Up-regulation of AMP gene expression. Expression levels of AMP were presented as fold changes in comparison with those in naïve larvae. Expression level of a ribosomal gene, RL32, was used as reference to normalize expression levels of target genes. Each treatment was replicated three times. Different letters above standard deviation bars denote significant difference among means at Type I error = 0.05 (LSD test).

Fig 3. Toll receptors (Se-Tolls) of S. exigua.

(A) Gene map of 10 Se-Tolls predicted in this study on chromosome(s), showing three linkage groups (LGs). (B) Their domain and phylogeny analyses. Predicted domains included SP (signal peptide), LRR (leucine-rich-repeat), LRR-CT (leucine-rich-repeat C terminal), LRR-NT (leucine-rich-repeat N terminal), Q-rich region (glutamine rich region), transmembrane region, and TIR (Toll-interleukin receptor). Neighbor-joining method was applied for constructing a phylogenetic tree. Bootstrap values on nodes are obtained from 1,000 repetitions. (C) Expression profile of 10 Toll genes (‘T1-T10’) in naive fat body tissues of S. exigua. Expression level of a ribosomal gene, RL32, was used as reference to normalize expression levels of target genes. Each treatment was replicated three times. Different letters above standard deviation bars denote significant difference among means at Type I error = 0.05 (LSD test).

Fig 4. Functional assay of 10 Se-Tolls (‘T1-T10’) for immune mediation of Se-DSP1 in S. exigua by individual knocking-down of gene expression using RNAi.

(A) RNAi efficiencies of 10 Se-Toll genes by injecting gene-specific dsRNAs (‘dsToll1-dsToll10’, 1 μg/larva) to L5 larvae. A viral gene, CpBV302, was used to prepare control dsRNA (‘dsCON’). Expression level of a ribosomal gene, RL32, was used as reference to normalize expression levels of target genes. (B-D) Changes in immune responses after individual RNAi treatments. At 24 h PI dsRNA, rSe-DSP1 (‘DSP1’) was injected to L5 larvae at a dose of 0.8 μg/larva. At 8 h PI, hemolymph and fat body were collected. Hemolymph was used for analyzing activities of phenoloxidase (‘PO’) and sPLA₂. Fat body was used for analyzing cPLA₂ activity. For analysis of expression levels antimicrobial peptide (gallerimycin (‘Gal’), gloverin (‘Glv’), and lysozyme (‘Lyz’)) genes, fat bodies were collected from treated larvae at 12 h PI. Each treatment was replicated three times. Different letters above standard deviation bars denote significant difference among means at Type I error = 0.05 (LSD test).

Fig 5. Two Spätzles (Se-Spz) of S. exigua.

(A) Functional domains of two proSpätzles (‘Se-ProSpz1’ and ‘Se-Spz2’), including a SP (signal peptide), a D/ERR (aspartic/glutamic acid rich region), a TRR (threonine rich region), and a Toll binding domain. Scissors indicate cleavage sites during post-translational modification from Se-ProSpz1/2 to active Se-Spz1/2. Disulfide bonds are indicated by linking cysteine residues. (B) Phylogeny analysis of Se-ProSpz1/2 and those of other insects. The analysis was performed using MEGA6 program with a Neighbor-joining method. Bootstrapping values were obtained with 1,000 repetitions to support branching and clustering. Amino acid sequences were retrieved from GenBank with accession numbers shown in S2 Table. (C) Expression profile of Se-Spz1/2 in hemocyte (‘HC’) and fat body (‘FB’) of L5 larvae of S. exigua. rSe-DSP1 (‘DSP1’) was injected into L5 larvae at a dose of 0.8 μg/larva. Immune challenge was performed by injecting E. mundtii (‘Em’) to L5 larvae at a dose of 4 × 10⁵ cells/larva. At 8 h PI, hemolymph and fat body were collected for RT-qPCR analysis. Control insects were injected with PBS. Expression levels of Se-Spz1/2 were presented as fold changes in comparison with those in control larvae. Expression level of a ribosomal gene, RL32, was used as reference to normalize expression levels of target genes. Each treatment was replicated three times. Different letters above standard deviation bars denote significant difference among means at Type I error = 0.05 (LSD test).

Fig 6. Functional assay of two Spätzles (‘Se-Spz1 and Se-Spz2’) for immune mediation of Se-DSP1 in S. exigua by individual knocking-down of gene expression using RNAi.

(A) RNAi efficiencies of two Se-Spz genes by injecting gene-specific dsRNAs (‘dsSpz1 and dsSpz2’, 1 μg/larva) to L5 larvae. A viral gene, CpBV302, was used to prepare control dsRNA (‘dsCON’). Expression level of a ribosomal gene, RL32, was used as reference to normalize expression levels of target genes. (B-D) Changes in immune responses after individual RNAi treatment. At 24 h PI dsRNA, rSe-DSP1 (‘DSP1’) was injected to L5 larvae at a dose of 0.8 μg/larva. For bacterial immune challenge, Enterococcus mundtii (Em, 4 × 10⁵ cells/larva) was injected into larvae. Naïve larvae were injected with PBS. At 8 h PI, hemolymph and fat body were collected. Hemolymph was used for analyzing activities of phenoloxidase (‘PO’) and sPLA₂. Fat body was used for analyzing cPLA₂ activity. For analyzing expression levels of antimicrobial peptide (gallerimycin (‘Gal’), gloverin (‘Glv’), and lysozyme (‘Lyz’)) genes, fat bodies were collected from treated larvae at 12 h PI. Each treatment was replicated three times. Different letters above standard deviation bars denote significant difference among means at Type I error = 0.05 (LSD test).

Fig 7. Deletion mutants of SeToll9 using CRISPR/Cas9 and their insensitivity to Se-DSP1 to express immune responses in S. exigua.

(A) Construction of SeToll9-deletion mutants (ΔToll9) with two single-stranded guide RNAs (sgRNAs), with protospacer adjacent motif (PAM) denoted in red color. Twelve different types of ΔToll9 (‘M1-M12’) are confirmed by sequence analysis of 808 bp around both sgRNA-specific deletion sites by comparing with the corresponding sequence of wild type (‘Wild’). Deletion or insertion sizes by CRISPR/Cas9 are denoted by ‘+’ or ‘-’ in parentheses. (B) Insensitivity of mutants to Se-DSP1 in activation of PLA₂. rSe-DSP1 (‘DSP1’) was injected into L5 larvae at a dose of 0.8 μg/larva. Control (‘CON’) larvae were injected with PBS. At 8 h PI, hemolymph and fat bodies were collected. Hemolymph was used for sPLA₂ activity analysis. Fat body was used for cPLA₂ activity analysis. (C) Insensitivity of mutants to Se-DSP1 in activating expression of antimicrobial peptide (gallerimycin (‘Gal’), gloverin (‘Glv’), and lysozyme (‘Lyz’)) genes. Fat bodies were collected from treated larvae at 12 h PI of Se-DSP1. Expression level of a ribosomal gene, RL32, was used as reference to normalize expression levels of target genes. Each treatment was replicated three times. Different letters above standard deviation bars denote significant difference among means at Type I error = 0.05 (LSD test).

Fig 8. Comparative analysis of susceptibility of S. exigua larvae to an entomopathogenic bacterium, Bacillus thuringiensis (‘Bt’), after individual RNAi treatment targeting each of 10 SeTolls (‘T1-T10’) (A) and two Spätzle (‘SeSpz1 and SeSpz2’) (B) genes.

dsRNA-specific to individual genes were injected to L5 larvae (1 μg/larva). Control larvae were injected with dsRNA (‘CON’) specific to a viral gene CpBV302. At 24 h PI, 500 ppm of Bt was orally fed to larvae with a leaf-dipping method. Mortality was recorded at 3 days after treatment (‘DAT’). Each treatment was replicated three times and each replication used 10 larvae. Different letters above standard deviation bars denote significant difference among means at Type I error = 0.05 (LSD test).

Fig 9. Screening bacterial metabolites of X. hominickii (Xh) for their binding affinities for Se-DSP1.

(A) Thermal shift assay for screening binding affinities using protein denaturation curve occurring with increasing ambient temperature. Shifting of the maximal value at higher temperature presumes that Se-DSP1 is completely denatured and maximally bound to a fluorescence dye. Dissociation constant (Kd) is estimated based on the relation between the maximal dissociation temperature and test compound concentration. (B) Binding assays for four organic extracts of Xh culture broth after 48 h of grown in TSB (S2 Fig). Extracts used included hexane (HEX), ethyl acetate (EAX), chloroform (CX), and butanol (BX) extracts. (C) Binding assay of 15 fractions (‘F1-F15’) of BX. (D) Binding assays for nine subfractions isolated from F2 fraction of BX (‘BX-F2’). (E) Binding assays for 11 subfractions isolated from F4 fraction of BX (‘BX-F4’) (F) Binding assays for nine subfractions isolated from F6 of BX (‘BX-F6’). Each measurement was replicated three times (three independent samples for each replicate). Different letters above standard deviation bars indicate significant differences among means at Type I error = 0.05 (LSD test).

Fig 10. Identification of compounds binding to Se-DSP1 from bacterial culture broth of X. hominickii.

(A) Prediction of Se-DSP1-binding compounds from purified butanol extract (‘Xh-BX’) of X. hominickii culture broth using GC-MS analysis. For example, ‘F2-2’ stands for subfraction #2 from ‘BX-F2’ fraction in Fig 9. See GC-MS chromatograms of compounds in S3 Fig (B) Binding affinity (Kd) estimations of seven bacterial metabolites and salicylic acid to rSe-DSP1 assessed by thermal shift assay. Each treatment was replicated three times with individual samples. Different letters following standard deviation (‘SD’) indicate significant difference among means at Type I error = 0.05 (LSD test).

Fig 11. Inhibitory activity of 3-ethoxy-4-methoxyphenol (‘EMP’) on Se-DSP1 secretion to plasma in response of S. exigua to immune challenge.

L5 larvae were injected with E. mundtii (‘Em’, 4 × 10⁵ cells/larva) and EMP (1 μg/larva). At 6 h PI, hemolymph and fat body were collected. Immunofluorescence assays of Se-DSP1 in hemocytes (A) and fat body (B) are shown. F-actin and nucleus were stained with phalloidin and DAPI, respectively. Se-DSP1 was detected with its polyclonal antibody. (C) Western blotting analysis of Se-DSP1 in the plasma of naïve or treated larvae. Each lane was loaded with 5 μL of plasma. Coomassie-staining bands against a larval storage protein (‘LSP’) indicated that the same amount of proteins in plasma samples was loaded.

Fig 12. Inhibitory effect of 3-ethoxy-4-methoxyphenol (‘EMP’) or salicylic acid (‘SA’) on immune responses mediated by Se-DSP1 in S. exigua.

L5 larvae were injected with rSe-DSP1 (‘DSP1’, 0.8 μg/larva) and EMP or SA (1 μg/larva). At 8 h PI, hemolymph and fat body were collected. Hemolymph was used for analyzing activities of phenoloxidase (‘PO’) and sPLA₂. Fat body was used for analysis of cPLA₂ activity. For antimicrobial peptide (‘AMP’) analysis, fat bodies were collected from treated larvae at 12 h PI. (A) Inhibition of up-regulated PO activity by EMP or SA. (B) Inhibition of up-regulated PLA₂ activities by EMP or SA. (C) Inhibition of up-regulation of antimicrobial peptide (gallerimycin (‘Gal’), gloverin (‘Glv’), and lysozyme (‘Lyz’)) genes by EMP or SA. Expression levels of AMP were presented as fold changes in comparison with those in naïve larvae. Expression level of a ribosomal gene, RL32, was used as reference to normalize expression levels of targeted genes. Each treatment was replicated three times. Different letters above standard deviation bars denote significant difference among means at Type I error = 0.05 (LSD test). (D) Enhanced susceptibility of S. exigua larvae to Bacillus thuringiensis (‘Bt’), an entomopathogenic bacterium, after treatment with RNAi specific to Se-DSP1 or EMP. RNAi was performed by injecting gene-specific dsRNA against Se-DSP1 (dsDSP1) at a dose of 1 μg/larva. At 24 h PI, Bt was used for treatnet. Control dsRNA (dsCON) used dsRNA specific to a viral gene, CpBV302. EMP treatment was performed by mixing with Bt suspension. Each treatment was replicated three times. Different letters above standard deviation bars denote significant difference among means at Type I error = 0.05 (LSD test).

Fig 13. A working hypothesis of Se-DSP1 for mediating immune responses via Toll-Spätzle (‘Spz’) signalling pathway and its inhibition by a bacterial metabolite, 3-ethoxy-4-methoxyphenol (‘EMP’), in S. exigua.

Upon challenge with bacteria including an entomopathogenic bacterium, X. hominickii, Se-DSP1 is secreted to the plasma to activate serine protease (SP) cascade for activating phenoloxidase (‘PO’) and Spz. Activated PO can catalyze melanin formation to suppress the growth of pathogenic bacteria. Activated Spz can bind to SeToll receptor to activate PLA₂ and the expression of antimicrobial peptide (gallerimycin (‘Gal’), gloverin (‘Glv’), and lysozyme (‘Lyz’)) genes. Activated PLA₂ can catalyze eicosanoid biosynthesis to mediate cellular immune responses to defend bacterial infection along with AMPs. To overcome tmmune responses mediated by Se-DSP1, X. hominickii produces and secretes secondary metabolites including EMP to inhibit the secretion of Se-DSP1 and prevent its immune-mediating activity.