Serine Protease Networks Mediate Immune Responses in Extra-Embryonic Tissues of Eggs in the Tobacco Hornworm, Manduca sexta

Abstract

The melanization and Toll pathways, regulated by a network of serine proteases and noncatalytic serine protease homologs (SPHs), have been investigated mostly in adult and larval insects. However, how these innate immune reactions are regulated in insect eggs remains unclear. Here we present evidence from transcriptome and proteome analyses that extra-embryonic tissues (yolk and serosa) of early-stage Manduca sexta eggs are immune competent, with expression of immune effector genes including prophenoloxidase and antimicrobial peptides. We identified gene products of the melanization and Toll pathways in M. sexta eggs. Through in vitro reconstitution experiments, we demonstrated that constitutive and infection-induced serine protease cascade modules that stimulate immune responses exist in the extra-embryonic tissues of M. sexta eggs. The constitutive module (HP14b-SP144-GP6) may promote rapid early immune signaling by forming a cascade activating the cytokine Spätzle and regulating melanization by activating prophenoloxidase (proPO). The inducible module (HP14a-HP21-HP5) may trigger enhanced activation of Spätzle and proPO at a later phase of infection. Crosstalk between the two modules may occur in transition from the constitutive to the induced response in eggs inoculated with bacteria. Examination of data from two other well-studied insect species, Tribolium castaneum and Drosophila melanogaster, supports a role for a serosa-dependent constitutive protease cascade in protecting early embryos against invading pathogens.

Introduction

Extracellular serine protease cascades stimulate insect innate immune reactions by activating prophenoloxidase (proPO) and proSpätzle (proSpz) to trigger the melanization and Toll pathways, respectively. The proteases circulate in hemolymph (insect blood) as inactive zymogens and become active upon specific cleavage by proteases upstream in the pathway [1]. The mechanisms regulating these innate immune reactions have been most thoroughly investigated in adult and larval insects, primarily through genetic studies in the fruit fly Drosophila melanogaster and biochemical studies in the tobacco hornworm, Manduca sexta, and the yellow mealworm beetle, Tenebrio molitor [2, 3, 4, 5].

Autoactivation of an initiating protease upon binding of pattern recognition receptors (PRRs) to pathogen surfaces is a crucial step in eliciting insect immune responses [6]. The detection of bacteria is mediated by the recognition of peptidoglycans by two types of PRRs, peptidoglycan recognition proteins (PGRPs) and microbe binding proteins (MBPs, closely related to Gram-negative bacteria-binding proteins, GNBP) [7], whereas the fungal cell wall component β-1,3-glucan is detected by β-1,3-glucan recognition proteins (βGRPs) [8]. In D. melanogaster, genetic studies indicate that a serine protease cascade involving ModSP, Grass, Psh/Hayan, and SPE is responsible for activation of the Spz-Toll-Dorsal/Dif pathway, which induces antimicrobial peptide (AMP) expression [2]. This cascade likely branches at the positions of Psh/Hayan to Sp7, which activates proPO1 to combat invading bacteria. In M. sexta, the sequential activation of the serine proteases HP14a, HP21, HP5, HP6, and HP8 leads to the cleavage of proSpz and the activation of Toll signaling [3]. This cascade branches at the position of HP21 and HP6 to activate prophenoloxidase activating proteases (PAPs), which in turn activate proPO in the presence of two noncatalytic serine protease homologs (SPHs) [9, 10, 11, 12]. In T. molitor, MSP, SAE, and SPE form a cascade to activate proSpz [5]. Active SPE also proteolytically activates proPO in the presence of T. molitor SPH1 oligomers [4].

Not only larvae and adults but also insect eggs may be threatened by pathogens. Although insect eggs are protected by the egg shell and the underlying vitelline membrane, they can be infected by pathogens through transovarial transmission or penetration of the eggshell [13, 14]. Recent advances indicate that insect eggs can harbor endogenous immune defenses through the production of proteins by the serosa, an extra-embryonic tissue [15, 16]. However, the serosa was secondarily lost in a group of derived flies (the Schizophora), which includes the well-studied model insect D. melanogaster [17, 18]. Consistent with this absence, D. melanogaster eggs do not mount an immune response to a bacterial challenge until stage 15, a late embryonic stage [19]. The absence of a serosa might account for the poor immune response in Drosophila eggs.

To gain deeper insights into the innate immune reactions in insect eggs, we chose as our model M. sexta, a moth that possesses a serosa. Our transcriptome and proteome data and in vitro reconstitution experiments suggest that a protease network, consisting of a constitutive module (HP14b-SP144-GP6) and an inducible module (HP14a-HP21-HP5), together with other downstream SPs/SPHs, mediate immune responses in the extra-embryonic tissues (yolk and serosa) of M. sexta. Further phylogenetic analysis and expression profiling reveal that the serosa-dependent serine protease cascade for protecting early embryos may be a common mechanism ancestral to non-Schizophoran insects.

Materials and Methods

Microbial Elicitors

Insoluble β-1,3-glucan, curdlan (from Alcaligenes faecalis, #C7821), was purchased from Sigma. Soluble DAP-type peptidoglycan from the Gram-negative bacterium Escherichia coli K12 (Cat. tlrl-ksspgn) and insoluble Lys-type peptidoglycan from the Gram-positive bacterium Staphylococcus aureus (Cat. tlrl-pgns2) were from InvivoGen.

Egg Dissections

Insects were reared on an artificial diet at 26°C with a 16:8 light:dark cycle [20]. Eggs were collected within 1 h of oviposition, surface sterilized in 3.5% bleach (0.2% sodium hypochlorite) for 1 min, then rinsed with tap water; after air drying for 15 min, the eggs were placed in the 26°C incubator. Twenty-four hours after oviposition (mid-to-late stage 4, 20–25% development), the eggs were either pricked with a minuten pin (mock infected) or a minuten pin dipped in a pellet of bacteria (bacterial infected). (Equal volumes of formaldehyde-treated E. coli (CGSC #7636) and Micrococcus luteus (ATCC #700256), each at a concentration of 1.65 × 10⁸ bacteria/mL, were mixed together and centrifuged at 16,000 × g for 5 min; the supernatant was removed and the bacterial pellet stored at 4°C). Twenty-four to twenty-eight hours after inoculation (48–52 h post-oviposition, mid-to-late stage 8, 40–45% development), the eggs were dissected in cold phosphate-buffered saline (PBS), and the embryos and extra-embryonic tissues were separately transferred to 1.5 mL microcentrifuge tubes on ice. For proteomic analysis, samples were stored at −20°C; each sample contained tissue from 10 to 15 eggs. Tissues collected for RNAseq analysis were centrifuged at 4°C for 5 min at 16,000 × g. The supernatant was removed, and the tissues were suspended in 100 μL of RNAlater (Thermo Fisher Scientific), stored at 4°C for 2–5 days, then placed at −20°C until RNA isolation; each replicate contained tissues collected from 14–16 eggs.

Proteomic Analysis

The total volume of each sample was brought up to 50 μL with PBS. The samples were homogenized in a 1.5 mL microcentrifuge tube with a fitted pestle, then heated at 95°C for 10 min and centrifuged at 16,000 × g for 5 min. An equal volume of 2X non-reducing SDS buffer (100 mM Tris-HCl, pH 7.0, 4% SDS, 20% glycerol) was added, and the process was repeated. The soluble fraction was saved, and the protein concentration was determined using a Micro BCA Protein Assay Kit (Thermo Fisher Scientific). Ten μg of protein from each sample were separately mixed with 6X SDS sample buffer and heated at 95°C for 5 min, then loaded onto a 4–15% precast gradient polyacrylamide gel (Bio-Rad) and separated at a constant current of 30 mA for 45 min. Gel bands were cut, and in-gel trypsin digestion was performed as described [21]. Mass spectra of the resulting peptides were obtained by liquid chromatography tandem mass spectrometry (LC-MS/MS) [21]. The MS/MS data were searched against a database of M. sexta consisted of high-quality immunity-related protein sequences [6, 22, 23, 24, 25, 26] and RefSeq GCF_014839805.1 (NCBI).

RNAseq Analysis

The samples were centrifuged at 16,000 × g for 10 min at 4°C, and the supernatants were discarded. The tissues were suspended in 500 μL of TRIzol reagent (Thermo Fisher Scientific) and homogenized with a fitted pestle; total RNA was isolated following the manufacturer's suggested protocol. Libraries constructed with 3 μg total RNA from pooled aliquots (1 μg) from each of the three biological replicates were subjected to high-throughput sequencing on a NextSeq Illumina 500 platform. The TPM (transcripts per million) method was used to calculate the normalized expression data of each gene.

cDNA Synthesis and qRT-PCR Analysis

The single-strand cDNA, synthesized from the RNA using iScript Reverse Transcription Supermix (Bio-Rad), was used as a template for qRT-PCR. The qRT-PCR was performed on a CFX Connect Real-Time System (Bio-Rad) using iTaq Universal SYBR Green Supermix (Bio-Rad) according to the manufacturer's instructions. The primers and amplification efficiencies (E) are listed in Table S1 (see www.karger.com/doi/10.1159/000527974 for all online suppl. material). Relative mRNA levels were calculated as (1 + E_rpS3)^{Ct, rpS3}/(1 + E_x)^{Ct, x}.

Recombinant Protein Preparation

M. sexta PGRP1 [27], MBP [28], proHP14a [29], proHP21 [9], proHP5 [3], proHP6 [11], proPAP1 [30], proPAP3 [31], proSPH1b [12], proSPH2 [32], and proSpz1a [33] were expressed in baculovirus-infected Sf9 cells. M. sexta βGRP1 [34] and proPO [35] were isolated from hemolymph and stored at −80°C.

M. sexta HP14b, SP144 and GP6 cDNA Cloning, E. coli Expression, and Antiserum

The coding region of mature proHP14b was amplified by PCR using forward primer (J184 5′- CCATGGGAATTCAAACATCCTTGATTAGATC), reverse primer (J185 5′- CTGCAGTTACTCGAGGTCGTCAGCGCTCCA), and cDNA from M. sexta eggs as template. Similarly, proSP144 was amplified using primers J186 (5′- CCATGGGAATTCACGACGTGTGTATT) and J187 (5′- CTGCAGTTACTCGAGCCACAAGACGCTCT), and proGP6 was amplified using primers J188 (5′- CCATGGGAATTCGAAAGTGTATCATAGATTAC) and J189 (5′- CTGCAGTTACTCGAGATGATCTAAAATCCAG). The purified products were ligated with pGEM-T vector (Promega) before transformation and plasmid isolation. Following sequence validation, the NcoI-PstI fragments were subcloned into the same sites in H6pQE60. E. coli JM109 cells harboring the plasmid yielded an insoluble polypeptide with an amino terminal tag of MHHHHHHAMGI followed by the full-length proteins. The recombinant protein was solubilized in 8 M urea and purified by nickel affinity chromatography under denaturing conditions, followed by preparative SDS-PAGE [36]. The resulting protein was used as antigen to generate a rabbit antiserum (Cocalico Biologicals).

Production of M. sexta proHP14b, proSP144, and proGP6 in the Baculovirus Expression System

The plasmids proHP14b/pGEM-T, proSP144/pGEM-T, and proGP6/pGEM-T described above were digested using restriction enzymes EcoRI and XhoI. The EcoRI-XhoI fragments were subcloned into the same sites in pMFH6, which encodes the honeybee melittin signal peptide and a carboxyl-terminal hexahistidine tag [32]. The plasmid was used to produce a bacmid and a high-titer viral stock for infecting Sf9 cells. Recombinant protein was purified from 300 mL of the conditioned medium. Briefly, cells were removed by centrifugation at 5,000 × g for 20 min, and the cell-free medium containing secreted proteins was collected and diluted with an equal volume of deionized H₂O, supplemented with 1 mM benzamidine and adjusted pH to 6.4 before applied to a dextran sulfate-Sepharose CL-6B column (40 mL) equilibrated in DS buffer (10 mM KPO₄, 1 mM benzamidine, 0.001% Tween-20, pH 6.4). After washing with 200 mL DS buffer, a linear gradient of 0–1.0 M NaCl in DS buffer was employed to elute the bound proteins at 1.5 mL/min for 100 min, followed by 1.0 M NaCl in DS buffer at 1.5 mL/min for 50 min. Collected fractions were analyzed by 10% SDS-PAGE and immunoblotting using a ×6His tag antibody (GenScript). Fractions containing target proteins were pooled and adjusted to pH 7.5, then applied to a Ni²⁺-NTA agarose column (2 mL) (Qiagen) and equilibrated with Ni²⁺-NTA buffer (50 mM Tris-HCl, pH 7.5, 300 mM NaCl, 10 mM imidazole, 5% glycerol, 1 mM benzamidine). After washing with Ni²⁺-NTA buffer (10 mL), bound proteins were eluted from the column with a linear gradient of 10–100 mM imidazole in Ni²⁺-NTA buffer (40 mL). Finally, tightly bound proteins were eluted with 10 mL of Ni²⁺-NTA buffer containing 250 mM imidazole. All purification steps were carried out at 4°C. After electrophoretic analysis, fractions with purified target proteins were combined and concentrated using Amicon ultracentrifugal 30K MWCO filter devices (Millipore). Following buffer exchange to 20 mM Tris-HCl, pH 7.5, 50 mM NaCl, aliquots of the recombinant protein were stored at −80°C.

Amidase Activity Assay

The activation of proPAP1 and proPAP3 were confirmed by measuring their amidase activity. Briefly, the in vitro recombinant protein reactions were mixed with 50 μM acetyl-Ile-Glu-Ala-Arg-p-nitroanilide (IEARpNa) [37] in 150 μL reaction buffer (0.1 M Tris-HCl, 0.1 M NaCl, 5 mM CaCl₂, pH 8.0) and the change in absorbance at 405 nm was recorded every 30 s for 20 min. The reaction rate was calculated from the slope of the initial, linear portion of each curve generated. One unit of activity was defined as a change in A₄₀₅ of 0.001 per min.

proPO Activation Assay

Day 2, fifth instar larvae of M. sexta were injected with a mixture of heat-killed (121°C for 15 min) E. coli (2 × 10⁷ cells), M. luteus (20 μg) (Sigma-Aldrich), and curdlan (20 μg, insoluble β-1,3-glucan from Alcaligenes faecalis) (Sigma-Aldrich) in 30 μL H₂O. Hemolymph was collected 24 h later and centrifuged at 5,000 g for 5 min at 4°C to remove hemocytes and obtain induced (but uninfected) plasma (IP). Aliquots of 1:10 buffer A (20 mM Tris-HCl, pH 7.5, 20 mM NaCl, 5 mM CaCl₂, 0.001% Tween-20) diluted IP (5.0 μL) were incubated with various amounts (0.1 μg, 0.2 μg, 0.3 μg, 0.4 μg, 0.5 μg) of proHP14b, or proSP144, or proGP6 in 20 μL buffer A (20 mM Tris-HCl, pH 7.5, 20 mM NaCl, 5 mM CaCl₂, 0.001% Tween-20) for 45 min at 25°C prior to PO activity assay using dopamine as substrate [37].

Expression Profiling of the Immunity-Related SP and SPH Genes in T. castaneum

The orthologs of M. sexta immunity-related SPs/SPHs in T. castaneum were identified based on published phylogenetic data [38]. RPKM (reads per kilobase of template per million mapped reads) values of the SPs/SPHs in whole insects at 4 life stages were extracted from the published data [39].

Multiple Sequence Alignment and Phylogenetic Analysis

Multiple sequence alignments of the entire proteins of the 173 nondigestive M. sexta SPs and SPHs [23, 40] were performed using MUSCLE, one module of MEGA X (http://www.megasoftware.net). The following parameters were used: gap opening penalty = −2.9, gap extension penalty = 0, hydrophobicity multiplier = 1.2, maximum iterations = 100, clustering method (for iterations 1 and 2) = UPGMB, and maximum diagonal length = 24. The aligned sequences were used to construct a neighbor-joining phylogenetic tree using MEGA X with bootstrap method for the phylogeny test (1,000 replications, Poisson model, uniform rates, and partial deletion of gaps or missing data 95%).

Results

Extra-Embryonic Tissues of Early-Stage M. sexta Eggs Are Immune Competent

Toll pathway activation and melanization are conserved insect immune responses, producing AMPs and generating melanin through active PO. A previous study demonstrated that bacterial challenge stimulates melanization and AMP synthesis in extra-embryonic tissues of M. sexta eggs [41]. To better understand the defense reaction in eggs, we performed transcriptome and proteome studies of M. sexta embryos and extra-embryonic tissues.

To identify sources of the immune proteins, we examined the transcriptomes of embryos and extra-embryonic tissues (yolk and serosa) from mock-injected and bacteria-inoculated M. sexta eggs. The same treatment groups, as well as naïve control, were obtained for proteome analysis. Eggs were inoculated at 24–28 h after oviposition and dissected 24 h later (48–52 h post-oviposition) for the extraction of protein and RNA. We examined differential expression of transcripts for immune effector genes, including 86 AMPs and 2 proPO genes identified in the M. sexta genome [25], setting a threshold of a 10-fold increase in transcript level for upregulation and 10-fold decrease in transcript level for downregulation (online suppl. Fig. S1). One AMP gene (WAP3) in embryonic tissue and 30 AMP genes in extra-embryonic tissue had >10-fold higher expression in bacteria-inoculated eggs compared to the mock-injected eggs. Notably, inoculation with bacteria stimulated >100-fold higher expression of 13 AMPs (attacin1/2/4/7/8/10, cecropin2/5/6/7, gallerimycin1, gloverin, 3-tox4, and 4-tox3) in extra-embryonic tissues compared with the mock-injected group. Similarly, when comparing the two tissue sources after bacterial inoculation, only WAP3 had higher expression in embryos, while 54 AMP genes were upregulated in extra-embryonic tissues; these included attacins, cecropins, moricins, lebocins, gallerimycins, defensins, and x-tox genes. These results are consistent with the proteomics data (online suppl. Table S2), in which 28 AMPs were identified, with 3 (e.g., attacin2) present in the mock injection and greatly elevated after the immune challenge, 14 (e.g., cecropin1) only detected after bacterial inoculation, and 4 (e.g., lysozyme1/2) present at low constitutive levels and highly abundant upon immune challenge. The transcripts and proteins of the two proPO genes were at higher levels in extra-embryonic tissues than in embryos and did not change significantly in response to bacterial inoculation in either type of sample. Together, these data indicate that early-stage M. sexta eggs are immune competent and that extra-embryonic tissues may protect embryos from invading pathogens by synthesis of AMPs and activation of proPO via immune signaling pathways.

Melanization and Toll Pathway Related Proteins in M. sexta Eggs

Melanization and Toll pathways are activated by extracellular serine protease cascades in M. sexta larvae. To examine whether similar cascade pathways may occur in eggs, we identified related proteins in the proteomes from embryos and extra-embryonic tissues. PRRs including PGRP1, MBP, βGRP1, and βGRP2 were detected in extra-embryonic tissues and embryos (online suppl. Table S2), indicating that proteins capable of sensing bacterial and fungal infections are present in eggs. However, proteins for the initial three proteases of the larval immune protease cascade, HP14a, HP21, and HP5 [3], were not detected in the naïve or mock inoculated groups (online suppl. Table S2). In contrast, three proteases identified by our previous transcript profiling as specific to the egg stage [23], HP14b, SP144, and GP6, were detected in naïve and mock infected samples. Of note, the gene pairs HP14b-HP14a, SP144-HP21, and GP6-HP5 are paralogs (online suppl. Fig. S2), located closely on the chromosome or genome contig (online suppl. Table S3), suggesting that each gene pair may have arisen from gene duplication and sequence divergence. We speculated that the constitutively expressed proteases HP14b, SP144, and GP6 may have the conserved functions in eggs of their corresponding paralogs in larval hemolymph [3, 8], acting as the first 3 enzymes of a protease immune cascade. Experiments below test this hypothesis.

Upon immune challenge, the protein levels of HP14b, SP144, and GP6 decreased (online suppl. Table S2). In contrast, HP14a and HP5 proteins were not detected in naïve eggs but were present after immune challenge. Contrasting changes in the two sets of closely similar proteases led us to compare their mRNA levels by qRT-PCR analysis in eggs after bacterial inoculation. The mRNA levels of HP14a, HP21, and HP5 were 2.6-, 1.5-, and 2.8-fold higher, respectively, in extra-embryonic tissues upon immune challenge than in tissues from mock infected eggs, whereas HP14b and SP144 mRNA levels were decreased compared with the mock infected control (GP6 showed no statistical difference between the control and immune challenged tissues) (Fig. 1). The induced expression of HP14a, HP21, and HP5 was infection-dependent, suggesting that they may play a role in the immune responses in a later phase of infection after their synthesis is stimulated by detected infection. Transcripts for all three pairs of proteins were present at very low levels in embryos, indicating that their presence in the serosa or yolk may fulfill an immune function to protect the embryo.

Fig. 1.

Expression levels of 6 serine protease genes in M. sexta eggs. Gene expression was determined by qRT-PCR, using ribosomal protein S3 (rpS3) as a reference gene. Results were normalized to the value obtained with the sample EM. Data represent means ± SDs of three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.005 and n.s. indicates not significant (Student's t test). EM, embryos of mock inoculated eggs; XM, extra-embryonic tissues of mock inoculated eggs; EB, embryos of bacteria-inoculated eggs; and XB, extra-embryonic tissues of bacteria-inoculated eggs.

Thus, we hypothesized that serine protease networks downstream of PRRs mediate two phases of immune response in extra-embryonic tissues of M. sexta, with two upstream protease modules: a constitutive module of HP14b-SP144-GP6 may propagate rapid early immune signaling, and an inducible module of HP14a-HP21-HP5 may provide an enhanced response in a later phase of infection. Both modules may stimulate proteolytic activation of HP6, HP8, PAP, SPH, and Spz precursors present in eggs (online suppl. Table S2), as occurs in larval hemolymph [11, 33, 42]. We further investigated the function of the constitutive module in experiments described below. A Proteolytic Cascade Specific to the Egg Stage Mediates the Activation of the Toll Pathway in M. sexta Eggs.

HP14b has a domain architecture similar to other initiating proteases of the Toll pathway, such as ModSP in D. melanogaster, MSP in T. molitor, and HP14a in M. sexta. It includes four low-density lipoprotein receptor class A repeats (LA1-LA4), a complement control protein domain (CCP), a cysteine-rich domain (7C), and a catalytic serine protease domain (Fig. 2a). SP144 and GP6 are CLIP family serine proteases, belonging to subfamilies C and B, respectively [23]. To test the hypothesis that HP14b activates SP144, which in turn activates GP6 as a constitutive protease cascade module propagating an early signal of invasion in eggs, we produced recombinant HP14b, SP144, and GP6 zymogens with a C-terminal 6×His tag, and generated rabbit antisera to HP14b, SP144, and GP6. When purified recombinant proHP14b was analyzed by SDS-PAGE, four major bands were identified (Fig. 2b). The predominant full-length HP14b zymogen migrated as an 85 kDa band on the stained gel, recognized by both HP14b and 6× His antibodies. Two lower mass bands, C1 and C2, were also detected by the 6× His tag antibody, indicating that they included the C-terminus. C2 corresponds in size to the catalytic domain of HP14b, whereas C1 may have been generated by an unknown protease during proHP14 production. The C-terminal bands were not recognized by the HP14b antibody, likely due to poor antigenicity of the protease domain, as was observed with HP14a [8]. A 50-kDa band N, not recognized by 6× His antibody, was detected by HP14b antibody and is consistent in size with the N-terminal LA-CCP-7C region. Recombinant proSP144 migrated as a doublet with molecular masses of 50 and 52 kDa, perhaps due to differences in glycosylation. The purified proGP6 migrated as a single band at 60 kDa (Fig. 2c, d).

Fig. 2.

Structural features and production of recombinant proHP14b, proSP144, and proGP6. a Domain architecture of HP14b, SP144 and GP6. The protease zymogen is activated by specific cleavage. After the cleavage, N-terminal fragment and C-terminal catalytic SP domain remain connected via a disulfide bond. For HP14b, four low-density lipoprotein receptor class A repeats (LA1-LA4), a complement control protein domain (CCP), a cysteine-rich domain (7C), and a catalytic SP domain are indicated. For SP144 and GP6, a CLIP domain and a catalytic SP domain are indicated. b−d The 10% SDS-PAGE and immunoblot analysis of the purified proHP14b (b), proSP144 (c), and proGP6 (d) from Sf9 cells. Aliquots of the protein (2.0 μg for staining; 200 ng for immunodetection) were treated by SDS sample buffer, separated along with protein size markers, and detected by Coomassie brilliant blue (CBB) or using 1:1,000 diluted respective antiserum as indicated at the bottom. Positions and sizes of the Mr makers are indicated. For the 4 bands of recombinant proHP14b, f: full-length zymogen; N: N-terminal fragment; C1: C-terminal fragment 1; C2: C-terminal fragment 2.

To investigate proHP14b activation in the presence of β-1,3-glucan, we used βGRP1 purified from the plasma of naïve M. sexta larvae, as it was more abundant in the egg proteome than βGRP2 (online suppl. Table S2). After incubation of proHP14b and βGRP1 with insoluble β-1,3-glucan (curdlan), the 85 kDa proHP14b was completely converted to a 50 kDa band detected by HP14b antibody (Fig. 3a left, lane 5), consistent with the size of band N observed in purified proHP14b (Fig. 2b, 3a left, lane 1). ProHP14b was not cleaved when incubated with β-1,3-glucan or βGRP1 alone (Fig. 3a left, lanes 3 and 4), suggesting that the binding of βGRP1 to β-1,3-glucan is necessary for proHP14b autoactivation. After activation, the catalytic domain of HP14b migrated to a position around 33 kDa visible with 6× His tag antibody, at the same position as band C2 observed in purified proHP14b (Fig. 2b, 3a right, lane 5). Similar autoactivation was observed when proHP14b was incubated with PGRP1 and MBP and soluble DAP-type peptidoglycan from E. coli (Ec-PG) (Fig. 3b, lane 12). However, in the absence of either recognition protein or peptidoglycan, proHP14b was not activated. These results suggest that PGRP1 and MBP are both required for recognizing DAP-PG and stimulating proHP14b activation. A high level of proHP14b cleavage was not observed when Ec-PG was substituted with Lys-type peptidoglycan from S. aureus (Sa-PG) (Fig. 3b, lane 13), consistent with the preferential binding of PGRP1 to DAP-PG [43].

Fig. 3.

Autoactivation assay of M. sexta proHP14b. a Autoactivation in the presence of β-1,3-glucan and βGRP1. Purified proHP14b (200 ng) was incubated with curdlan (10 μg), βGRP1 (150 ng), and buffer A (to 25 μL) for 2 h at 37°C. The reaction mixture and controls were subjected to 10% SDS-PAGE under reducing conditions, followed by immunoblot analysis using HP14b (left) and 6 × His (right) antibodies. Triangle indicates the N-terminal fragment of HP14b produced after activation cleavage. F Full-length zymogen; C1: C-terminal fragment 1; C2: C-terminal fragment 2. b Autoactivation in the presence of E. coli or S. aureus peptidoglycan, PGRP1, and MBP. Purified proHP14b (200 ng) was incubated with E. coli peptidoglycan (1 μg) or S. aureus peptidoglycan (1 μg), PGRP1 (500 ng), MBP (500 ng), and buffer A (to 25 μL) for 2 h at 37°C. The reaction mixture and controls were separated by 10% SDS-PAGE under reducing conditions and followed by immunoblot analysis using HP14b antibody. The triangle indicates the N-terminal fragment of HP14b produced after activation cleavage. Sizes and positions of the Mr markers are indicated on the left.

We then tested whether active HP14b could process proSP144. We mixed activated HP14b with proSP144 and detected proSP144 cleavage by immunoblot analysis using SP144 antibody. ProSP144 was cleaved by HP14b activated in the presence of β-1,3-glucan and βGRP1 (Fig. 4a, lane 5). The 50 kDa proSP144 band decreased in intensity, and a band appeared at the 30 kDa position, the expected size of the activated C-terminal protease domain of SP144. We then tested whether this putatively active SP144 may process proGP6. After incubation of proGP6 with HP14b-activated SP144, all of the proGP6 was converted to a 30 kDa band detected by GP6 antibody, at the expected size of the C-terminal protease domain after cleavage activation (Fig. 2a, 4b, lane 5). HP14b, in the absence of SP144, did not cleave proGP6 (Fig. 4b, lane 3). These results indicate that SP144 activated by HP14b was responsible for the observed proGP6 cleavage. In the control mixture of proSP144 and proGP6, the decreased intensity of the 60 kDa proGP6 band and a faint band at the size of GP6 protease domain was observed, perhaps due to a trace amount of active SP144 in the proSP144 preparation (Fig. 4b, lane 4). Considering the distinctive constitutive expression pattern of these egg-specific proteases, these results are consistent with a hypothesis that the module of HP14b-SP144-GP6 may function to initiate an immune cascade immediately after infection of eggs. We next tested whether GP6 could function as a terminal SP to process proSpz. When proSpz1a was incubated with active GP6, a 15 kDa band corresponding to the activated cystine-knot domain (Spz1a-C108) [33] was detected by Spz1a-C108 antibody (Fig. 4c, lane 5). In the absence of DTT, this protein migrated to a position around 25 kDa (Fig. 4c, lane 6), consistent with a disulfide-linked dimer, which was previously identified as the active form of M. sexta Spz1a [33]. Therefore, our in vitro reconstitution experiments suggest that the constitutive module forms a three-step proteolytic cascade (HP14b-SP144-GP6) that recognizes microbial polysaccharides through specific PRRs and activates the Toll pathway via specific cleavage of proSpz1a.

Fig. 4.

Sequential activation of M. sexta proSP144, proGP6 and proSpz1a. a proSP144 activation. Purified proSP144 (200 ng) was incubated with curdlan (10 μg), βGRP1 (150 ng), proHP14b (200 ng), and buffer A (to 25 μL) for 2 h at 37°C. The reaction mixture and controls were separated by 10% SDS-PAGE under reducing condition and detected by immunoblotting using SP144 antibody. Triangle indicates the catalytic domain of SP144 produced after activation cleavage. b proGP6 activation. Purified proGP6 (200 ng) was incubated with curdlan (10 μg), βGRP1 (150 ng), proHP14b (200 ng), proSP144 (200 ng), and buffer A (to 25 μL) for 1 h at 37°C. The reaction mixture and controls were subjected to 10% SDS-PAGE and immunoblot analysis using GP6 antibody. Triangles indicate the catalytic domain of GP6 produced after activation cleavage. c proSpz1a activation. Purified proSpz1a (400 ng) was incubated with a mixture of curdlan (10 μg), βGRP1 (100 ng), proHP14b (100 ng), proSP144 (100 ng), proGP6 (200 ng), and buffer A (to 25 μL) for 1 h at 37°C. The reaction mixture and controls were resolved by 12% SDS-PAGE under the reducing condition (lane 1-5) or non-reducing condition (lane 6) and detected by immunoblotting using Spz1a-C108 antibody. Triangles indicate Spz1a cystine-knot domain (Spätzle-C108). Sizes and positions of the Mr markers are indicated on the left.

The Protease Cascade Module in Eggs Can Promote Prophenoloxidase Activation

We tested whether HP14b, SP144, or GP6 could directly activate proPO purified from the plasma of M. sexta larvae, but no proPO cleavage was detected by proPO antibody after treatment with these proteases (online suppl. Fig. S3). In M. sexta, three prophenoloxidase activating proteases (PAPs) have been identified, PAP1, PAP2, and PAP3. We detected PAP1 and PAP3 in the egg proteome samples but not PAP2, which is consistent with previous identification of PAP1 and PAP3 by immunoblot analysis in 2 h and 24 h M sexta eggs [41]. Hence, we tested whether the protease module of HP14b-SP144-GP6 can activate PAPs. HP21, the paralog of SP144, activates proPAP2 [9] and proPAP3 [10] in larval hemolymph. Therefore, we then tested whether SP144 can similarly cleave and activate proPAP3. When proPAP3 was incubated with activated SP144, the 48 kDa proPAP3 zymogen was converted to a 37 kDa band (Fig. 5a lane 5), the same size as the catalytic domain of proPAP3 generated by cleavage with HP21 [10]. Confirming the proteolytic activation of PAP3, the artificial PAP substrate, IEARpNa, was hydrolyzed in the curdlan/βGRP1-proHP14b-proSP144-proPAP3 reaction (Fig. 5b), but very little IEARpNA amidase activity was present in the control reactions. These results suggest that SP144 can activate proPAP3.

Fig. 5.

Cleavage activation of proPAP3 by M. sexta SP144. a Activation of proPAP3. Purified proPAP3 (200 ng) was incubated with a mixture of curdlan (10 μg), βGRP1 (150 ng), proHP14b (200 ng), proSP144 (200 ng), and buffer A (to 25 μL) for 1 h at 37°C. The reaction mixture and controls were resolved by 10% SDS-PAGE under the reducing condition and detected by immunoblotting using PAP3 antibody. Triangle indicates the catalytic domain of PAP3 produced after activation cleavage. Sizes and positions of the Mr markers are indicated on the left. b Amidase activity of SP144-generated PAP3. In a duplicated experiment, amidase activities in the reaction and control mixtures were determined using 150 μL of 50 μM of IEARpNA and plotted as mean ± SD (n = 3).

In M. sexta larvae, proPAP1 is activated by HP6 [11], a protease downstream of HP5 [3]. Taking into consideration that GP6 clusters phylogenetically with HP5, we tested whether GP6 can also activate proHP6. After proHP6 was incubated with active GP6, two bands of approximately 30 kDa, corresponding to the C-terminal protease domain of HP6, were detected by the HP6 antibody (Fig. 6a, lane 5). To test if this cleaved form of HP6 is proteolytically active against its natural substrate, we then incubated it with proPAP1. After incubation of proPAP1 with GP6-activated HP6, all of the proPAP1 was converted into two bands recognized by the PAP1 antibody, representing the catalytic domain (29 kDa) and clip domain (20 kDa), respectively (online suppl. Fig. S4a, lane 5). In the absence of HP6, GP6 did not cleave proPAP1 (online suppl. Fig. S4a, lane 3). In the control mixture of proHP6 and proPAP1, two faint bands corresponding to the sizes of the catalytic and clip domains of PAP1 were also observed, perhaps due to a trace amount of active HP6 in the proHP6 preparation (online suppl. Fig. S4a, lane 4). A significant increase in IEARase activity (above that of proHP6-proPAP1 reaction) was observed in the presence of GP6-activated HP6 and proPAP1 (online suppl. Fig. S4b), thus confirming the activation of proPAP1 by HP6 that had been activated by GP6.

Fig. 6.

Cleavage activation of proHP6, proSPH1b, and proSPH2 by GP6. Two hundred ng of proHP6 (a), proSPH1b (b), or proSPH2 (c) were respectively incubated with a mixture of curdlan (10 μg), βGRP1 (100 ng), proHP14b (100 ng), proSP144 (100 ng), proGP6 (200 ng), and buffer A (to 25 μL) for 1 h at 37°C. The reaction mixture and controls were resolved by 10% (b, c) or 12% (a) SDS-PAGE under reducing conditions and detected by immunoblotting using HP6, SPH1, and SPH2 antibodies, respectively. Triangles indicate HP6 catalytic domain, SPH1b, and SPH2 heavy chain. Sizes and positions of the Mr markers are indicated on the left. d Function of SPH1b and SPH2 as PAP cofactors. Active GP6 was produced by incubating proGP6 (1.5 μg) with curdlan (10 μg) and 50 ng each of βGRP1, proHP14b, SP144, and buffer A (to 30 μL) at 37°C for 1 h. Activated GP6 (100 ng) was next mixed with proSPH1b (100 ng) and proSPH2 (100 ng) and incubated at 25°C for 30 min PAP3 (5 ng) and proPO (200 ng) were then added and incubated on ice for 30 min. PO activities were measured using dopamine as substrate and plotted as mean ± SD (n = 3). Student's t test was used to identify statistically significant differences between the PO activities of the indicated groups. **p < 0.01.

Efficient proPO activation by PAPs in M. sexta requires a cofactor consisting of two serine protease homologs, SPH1, and SPH2, that lack protease activity [44, 45]. The SPH cofactors themselves must be activated by proteolytic cleavage [12]. Both SPH1 and SPH2 proteins were present in eggs (online suppl. Table S2), but no distinction could be easily made among paralogs SPH1a, SPH1b, SPH4, and SPH101, due to their high sequence identity. As SPH1b was most abundant in the egg transcriptome (online suppl. Table S4), we tested cleavage activation of proSPH1b and proSPH2. Active GP6 processed both proSPH1b (Fig. 6b, lane 5) and proSPH2 (Fig. 6c, lane 5), yielding a 37 kDa band, similar to what was observed when those same two SPHs were processed by PAP3 [12]. We then tested whether proSPH1b and proSPH2, cleaved by GP6, can promote proPO activation by PAP3. PO activity increased approximately 6-fold when PAP3 and proPO were incubated with both proSPH1b and proSPH2 (Fig. 6d); this activation was not observed when either SPH was incubated individually with PAP3 and proPO. PO activity was further increased by the addition of active GP6, presumably through the cleavage of the SPHs, as the addition of GP6 alone, or in combination with only one SPH, did not increase PO activity. Thus, both SPH1b and SPH2 are required to form a functional cofactor for PAP3, and their cleavage by GP6 further enhances proPO activation. Collectively, these results demonstrated that the egg-specific protease cascade initiating module can trigger activation of PAP and SPH proteins that lead to proPO activation and melanization in M. sexta eggs.

Crosstalk between the Constitutive and Inducible Protease Immune Modules in Challenged Eggs

As proposed above, the constitutive protease module (HP14b-SP144-GP6) forms a cascade to initiate melanization and Toll activation in M. sexta eggs. The inducible module (HP14a-HP21-HP5) mediates the same responses in larval hemolymph [3] and is present in eggs after inoculation with bacteria. We investigated whether proHP5 and proGP6 may be activated by SP144 (constitutive) and HP21 (induced), respectively. Both proHP5 and proGP6 were completely activated by HP14b-activated SP144 (Fig. 7a, lane 5) and by HP14a-activated HP21 (Fig. 7b, lane 5). The conserved proteolytic activities of these two sets of proteases indicate that a crosstalk may occur between the constitutive and inducible protease modules during the transition to the induced phase of the immune response in eggs. The functional conservation led us to test the ability of the HP14b-SP144-GP6 module to activate proPO in larval plasma. When we incubated recombinant proHP14b, proSP144, and proGP6 with plasma from bacteria-induced larvae, PO activity increased in a manner dependent on the concentrations of the added recombinant proteases (Fig. 8). These results further illustrated the functional conservation and potential crosstalk between the two protease modules in eggs.

Fig. 7.

Crosstalk between the constitutive module (HP14b-SP144-GP6) and the inducible module (HP14a-HP21-HP5). a Proteolytic activation of proHP5 by SP144. Purified proHP5 (100 ng) was incubated with curdlan (10 μg), βGRP1 (150 ng), proHP14b (200 ng), proSP144 (200 ng), and buffer A (to 25 μL) for 1 h at 37°C. The reaction mixture and controls were resolved by 10% SDS-PAGE under reducing conditions and detected by immunoblotting using HP5 antibody. Triangle indicates the catalytic domain of HP5. b Proteolytic activation of proGP6 by HP21. Purified proGP6 (200 ng) was incubated with E. coli peptidoglycan (1 μg), PGRP1 (500 ng), MBP (500 ng), proHP14a (200 ng), proHP21 (100 ng), and buffer A (to 25 μL) for 1 h at 37°C. The reaction mixture and controls were subjected to 10% SDS-PAGE under reducing conditions and immunoblot analysis using GP6 antibody. Triangle indicates the catalytic domain of GP6. Sizes and positions of the Mr markers are indicated on the left. c Proteolytic activation of proSpz1a by HP5. Purified proSpz1a (400 ng) was incubated with E. coli peptidoglycan (1 μg), PGRP1 (500 ng), MBP (500 ng), proHP14a (200 ng), proHP21 (100 ng), proHP5 (100 ng), and buffer A (to 25 μL) for 1 h at 37°C. The reaction mixture and controls were subjected to 12% SDS-PAGE under reducing (lanes 1-5) or non-reducing conditions (lane 6) and immunoblot analysis using Spz1a-C108 antibody. Triangles indicate the cystine-knot domain (Spätzle-C108) of Spz1a. Sizes and positions of the Mr markers are indicated on the left.

Fig. 8.

Concentration-dependent proPO activation caused by proHP14b, proSP144, and proGP6 in the absence of a microbial elicitor. Aliquots of 1:10 diluted induced plasma (IP, 5.0 μL) were incubated with various amounts of recombinant serine protease zymogen for 45 min at 25°C. PO activities were measured using dopamine as substrate and plotted as mean ± SD (n = 3). “+” indicates the addition of IP.

Our previous studies have shown that HP5 activates proHP6, HP6 activates proHP8 (closely related to Spӓtzle processing enzymes (SPEs) from other insects), which in turn activates proSpz1a [3, 33]. HP5 can activate Toll signaling through the HP5-HP6-HP8-Spz1a cascade [3]. We tested whether HP5 may also directly activate proSpz1a, similar to the activity of the HP5 paralog GP6 (Fig. 4c). Active HP5 produced from the pathway of Ec-PG/PGRP1/MBP-HP14a-HP21-HP5 efficiently cleaved proSpz1a to yield the 15 kDa band consistent with the active Spz1a (Fig. 7d, lane 5); in the absence of DTT, the Spz1a cystine-knot domain migrated as a disulfide-linked dimer of about 25 kDa (Fig. 7d, lane 6). The finding that HP5 and GP6 (related to Grass in D. melanogaster [38] can process proSpz1 expands our knowledge on the direct activation of the Toll pathway. Altogether, the in vitro reconstitution and PO activation assays suggest a conserved role of the constitutive and inducible modules in an initial and enhanced extended immune response in eggs.

Serosa-Dependent SP Cascade May Be a Common Mechanism for Non-Schizophoran Insects

Several studies indicate that the extra-embryonic serosal epithelium provides insect eggs with immune defenses in several insects, including M. sexta, Tribolium castaneum, Oncopeltus fasciatus, and Locusta migratoria [15, 16, 41]. In contrast, the innate immune response of early-stage Drosophila eggs is poor [19], perhaps due to the absence of the serosa, a trait limited in insects to flies in the Schizophora group [17, 18]. We examined whether orthologs of the immune cascade proteases found in Manduca eggs are present in a beetle, T. castaneum, and D. melanogaster, two species with high-quality serine protease annotation and reliable expression profiles.

Several egg-specific clip serine proteases were identified in T. castaneum, including cSP61 (ortholog of HP21/SP144), cSP33 and cSP93/cSP94 (orthologs of HP5/GP6) (Fig. 9). Of two orthologs of the modular protease HP14a/HP14b in T. castaneum, mSP3 and mSP13, only mSP13 was detected in eggs (Fig. 9). We speculate that mSP13, cSP61, and cSP33/cSP93/cSP94 may function in a cascade in the beetle eggs, equivalent to HP14b-SP144-GP6 in M. sexta. In contrast, we did not find any immunity-related serine proteases with egg-specific expression in transcriptomes or proteomes of D. melanogaster [38]. These results are consistent with potentially conserved protease cascade immune pathways in insects that produce a serosa and lacking in a Schizophoran fly that does not produce a serosa during development of the egg.

Fig. 9.

Transcript profiles of the T. castaneum immunity-related SP/SPH genes during different developmental stages. The SP/SPH genes in T. castaneum were selected based on the phylogenetic relationship with the immunity-related SP/SPH genes in M. sexta [38]. The putative orthologs of T. castaneum SP/SPH genes in M. sexta are indicated on the right. The mRNA levels [39], as represented by log₂(RPKM+1) values, are shown in the gradient heat map from blue (0) to maroon (≥10). The values are labeled as 0–0.49 (0), 0.50–1.49 (1), 1.50–2.49 (2), 2.50–3.49 (3), 3.50–4.49 (4), 4.50–5.49 (5), 5.50–6.49 (6), 6.50–7.49 (7), 7.50–8.49 (8), 8.50–9.49 (9), and 9.50–10.49(A). No distinction could be made between cSP90/cSP91, cSP93/cSP94, cSP7/cSP8, and cSP136/cSP137 because they were assembled into the same transcript in the database used for data analysis [39].

Discussion

To protect against infections, insect eggs rely on (a) physical barriers such as the eggshell and hardened secretions covering the eggs [46], (b) endogenous egg defenses triggering expression of immunity-related genes [47], and (c) maternal care, including selection of oviposition sites [48]. Trans-generational immune priming may provide defensive structures or chemicals into eggs [49]. The trans-generational transmission of bacteria from mothers to the egg yolk of their offspring was observed in M. sexta [50], which for immune priming would require that the egg itself has the ability to respond to the transmitted bacteria or fragments thereof. In fact, the serosa, an extra-embryonic epithelium around the yolk and embryo, provides endogenous immunity to insect eggs, which has been demonstrated in holometabolous insects M. sexta and T. castaneum and in hemimetabolous insects O. fasciatus and L. migratoria [15, 16, 41]. Transcriptome and proteome data described here indicate that bacterial infection elicits substantial expression of multiple AMPs in extra-embryonic tissues, the yolk, and the serosa. In contrast, the embryo itself showed little immune response. We analyzed transcripts and proteins from embryos at developmental stages 8–9 or 40–45% development time [51], before there was substantial differentiation of hemocytes [52] or the fat body [51], the primary immune cell types of the larval and adult stages. Therefore, the serosa and yolk appear to provide a source of antimicrobial proteins that protect the developing embryo. However, it has been unclear how the immune response is regulated in insect eggs.

In this work, we provide evidence for serine protease cascade-mediated immune responses in an insect egg. In the M. sexta model (Fig. 10), bacterial and fungal infections in eggs are sensed by the same pattern recognition proteins that occur in larval hemolymph. This recognition triggers the activation of a module of constitutively expressed proteases, HP14b, SP144, and GP6. The sequential activation of these three proteases leads to the proteolytic activation of the cytokine Spz1a, which activates Toll signaling [33]. This constitutive pathway can branch at the position of SP144 to activate PAP3, which in turn activates proPO. GP6 cleaves and likely activates the serine protease homologs SPH1b and SPH2, as well as clip protease HP6. SPH1b and SPH2 are PAP cofactors, which are required for efficient generation of highly active PO [12]. HP6 can activate HP8, which can process the cytokine substrate Spz1a [33] or regulate melanization by activating PAP1 [11]. At a later phase of infection, an inducible module composed of proteases HP14a, HP21, and HP5 may further stimulate the Toll and melanization pathways, leading to an enhanced and prolonged immune response. The functional conservation of these two sets of proteases allows crosstalk during their transition, especially as SP144 and HP21 process HP5 and GP6, respectively. It appears that a redundant function for these proteases in Spz and proPO activation has evolved and may have some fitness benefit, which needs to be evaluated in embryo survival tests using HP14b, SP144, and GP6 knockout mutants. Functional redundancy was also found in D. melanogaster; Hayan and Psh merge signals from PRR and activate a common extracellular pathway upstream of Toll [2]. The phenomenon of HP5 or GP6 both directly (HP5/GP6→Spz) and indirectly (HP5/GP6→HP6→HP8→Spz) activating Spz is intriguing. Similarly, proPO can also be activated by a short (HP21/SP144→PAP3→PO) and a long (HP21/SP144→HP5/GP6→HP6→ PAP1→PO) cascades. Functional redundancy of serine protease molecules or serine protease cascades may be a strategy evolved in insects to defend against pathogen infection.

Fig. 10.

A simplified model of the serine protease network in M. sexta eggs. The blue arrows are elucidated in the current study. The system outputs (colored red) include PO and Spz1a.

Phylogenetic analysis and expression profiling to identify potential counterparts of the constitutive protease module in a beetle, T. castaneum, suggests that egg-specific proteases mSP13-cSP61-cSP33/cSP93/cSP94 might have the same types of immune function as M. sexta HP14b, SP144, and GP6. Previous research revealed that the serosa provides T. castaneum eggs with immune responses, with 44 immunity-related genes exhibiting serosa-dependent expression [15], including cSP61, cSP93, and cSP94. In light of this study, failure of experimental serosa-less T. castaneum eggs to restrict bacterial growth [15] could be caused in part by the lack of a key extracellular protease cascade. In D. melanogaster, which lacks a serosa, the innate immune response of the egg is limited [19]. While induced CecA1-lacZ expression was detected in embryonic yolk after LPS or E. cloacae injection [53], attacin, diptericin, and drosocin production was unaffected. In comparison, robust induction of AMP synthesis was observed in M. sexta extra-embryonic tissues (online suppl. Fig. S1; Table S2). We also examined expression profiles of the extracellular SP cascade members, but did not find ModSP, Grass, Psh, Hayan, or SPE expressed higher in early-stage eggs or ovaries of mated adult females than other tissues [38, 54, 55]. In contrast, Nudel, Gd, Snake, and Easter mRNA levels in the ovary sample were 4–8-fold higher than in testes, Malpighian tubules, and ovaries of virgin females, supporting a maternal transfer of the factors for establishing the dorsal-ventral axis. Similar transfer of the immunity-related proteases is less likely. It is possible insects like higher Diptera that have very fast embryonic development do not need to depend as much on an immune response in eggs. Thus, the serosa-mediated immune responses, through a serosa-dependent SP cascade, may be a common mechanism for immune protection of the embryo in non-Schizophoran insects. More evidence is needed to reveal whether serosa-dependent serine protease cascades are an ancestral feature for non-Schizophoran insects. Identification of orthologous proteases will aid the extension of functional information from model species such as M. sexta to other insects of practical importance. The protease network we identified in M. sexta eggs can facilitate studies of innate immunity in eggs of other groups of insects.

Statement of Ethics

An ethics statement was not required for this study type; no human or animal subjects or materials were used.

Conflict of Interest Statement

The authors have no conflict of interest to declare.

Funding Sources

This study was supported by NIH grants GM58634, AI139998, and GM141859. The paper was approved for publication by the Director of Oklahoma Agricultural Experimental Station and supported in part under project OKL03257.

Author Contributions

Tisheng Shan, Haobo Jiang, and Michael R. Kanost designed research and wrote the paper; Tisheng Shan, Yang Wang, and Neal T. Dittmer performed research; Tisheng Shan and Haobo Jiang analyzed data. The authors declare no competing interest.

Data Availability Statement

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.

Supplementary Material

Supplementary data

Supplementary data

Supplementary data

Supplementary data

Supplementary data

Supplementary data

Supplementary data

Supplementary data

Supplementary data

Supplementary data

Supplementary data

Supplementary data

Funding Statement

This study was supported by NIH grants GM58634, AI139998, and GM141859. The paper was approved for publication by the Director of Oklahoma Agricultural Experimental Station and supported in part under project OKL03257.