Hemolymph protease-17b activates proHP6 to stimulate melanization and Toll signaling in Manduca sexta

Abstract

Manduca sexta hemolymph protease-6 (HP6) plays a central role in coordinating antimicrobial responses, such as prophenoloxidase (PPO) activation and Toll signaling. Our previous studies indicated that HP5 and GP6 activate proHP6 in larval hemolymph and extraembryonic tissues, respectively. Here, we report the characterization of HP17b as another HP6 activating enzyme and its regulation by multiple serpins in hemolymph. The precursor of HP17b expressed in baculovirus infected Sf9 cells became spontaneously cleaved at two sites, and these products were purified together in one preparation named HP17b’, a mixture of proHP17b, a 35 kDa intermediate, and HP17b. HP17b’ converted proHP6 to HP6. As reported before, HP6 converted precursors of PPO activating protease-1 (PAP1) and HP8 to their active forms. HP8 activates proSpätzle-1 to turn on Toll signaling. We found HP17b’ directly activated proSPHI and II to form a cofactor for PPO activation by PAP1. Supplementation of larval hemolymph with HP17b’, HP17b, or proHP17b significantly increased PPO activation. Adding Micrococcus luteus to the reactions did not enhance PPO activation in the reactions containing HP17b’, HP17b, or proHP17b. Using HP17b antibodies, we isolated from induced plasma HP17b fragments and associated proteins (e.g., serpin-4). Serpin-1A, 1J, 1J’, 4, 5, or 6 reduced the activation of proHP6 by HP17b’ through formation of covalent complexes with active HP17b. We detected an activity for proHP17b cleavage in hemolymph from bar-stage pharate pupae but failed to purify the protease due to its high instability. Other known HPs did not activate proHP17b in vitro. Together, these results suggest that HP17b is a clip-domain protease activated by an unknown endopeptidase in response to a danger signal and regulated by multiple serpins.

Graphical Abstract

1. Introduction

Extracellular serine protease (SPs) and their noncatalytic homologs (SPHs) form cascade pathways to mediate innate immune responses in insects (Kanost and Jiang, 2015; Veillard et al., 2016). These include proteolytic activation of phenoloxidase (PO), Spätzle, and stress responsive peptide precursors. PO catalyzes the formation of reactive compounds to kill and sequester pathogens or parasites in melanin sheaths (Kanost and Gorman, 2008). Spätzles, stress responsive peptides, and other cytokines bind to their receptors (e.g., Toll) to trigger intracellular signaling pathways that induce the synthesis and release of immunity-related proteins (e.g., antimicrobial peptides) to kill invading organisms (Kanost and Jiang, 2015; Schrag et al., 2017). Many components of the SP pathways are regulated by serine protease inhibitors of the serpin superfamily (Meekins et al., 2017). Specific interactions among the cascade members are mediated by their regulatory domains. Clip-domain SPs and SPHs, also known as CLIPs, account for a large portion of the protease cascade systems in various insects (Cao and Jiang, 2018). Members of CLIPA and CLIPE subfamilies encode SPHs whereas most members of CLIPB, CLIPC, and CLIPD groups encode clip-domain SPs likely to have specific proteolytic activity. The immune SP-SPH systems have been studied in Drosophila melanogaster (Shan et al., 2023), Anopheles gambiae (Zhang et al., 2023), Aedes aegypti (Ji et al., 2022), Bombyx mori (Satoh et al., 1999), Manduca sexta (Shan et al., 2022), Helicoverpa armigera (Wang et al., 2022a), Ostrinia furnacalis (Zhang et al., 2022), Tenebrio molitor (Jiang et al., 2009), and Holotrichia diomphalia (Park et al., 2010).

M. sexta SPs and SPHs represent one of the best characterized S1A protease families in insects (Cao et al., 2015; Miao et al., 2020). Genome annotation and transcriptome analyses have revealed 246 SP(H) genes, eighty of which are expressed in midgut, including 59 SPs and 11 SPHs in the feeding stages. Among the 166 non-digestive SP(H)s, 56 contain other structural modules, such as clip domains in 43 CLIPs. Thirty-six SP(H)s are detected in larval hemolymph, fifteen of which are functionally characterized, including proHP1* (*, catalytically active), HP2, HP5, HP6, HP8, HP14, HP21, PAP1, PAP2, PAP3, SPH1a, SPH1b, SPH2, SPH4, and SPH101. ProHP14, an initiating protease of the system, is autoactivated in the presence of a β-1,3-glucan recognition protein (βGRP1/2) and fungal β-1,3-glucan (Wang and Jiang, 2006 and 2010; Takahashi et al., 2015). Autoactivation of proHP14 also occurs through the interaction of microbe binding protein (MBP), peptidoglycan recognition protein-1 (PGRP1), and peptidoglycans (DAP-type preferred) (Wang et al., 2022b). HP14 activates proHP2 and proHP21. While HP2 specifically activates proPAP2 during wandering stage, HP21 activates proPAP2 and proPAP3 in hemolymph of feeding larvae (He et al., 2018; Wang and Jiang, 2007; Gorman et al., 2007). PAP3 can autoactivate proPAP3 for feed-forward regulation, generate its own cofactor (SPHI-II), and cleave PPOs at Arg⁵¹ to greatly stimulate melanization (Wang et al., 2014). Both HP21 and PAP3 activate proHP5, HP5 activates proHP6, and HP6 activates proHP8 and proPAP1 (An et al., 2009; Wang et al., 2020). HP8 cleaves proSpätzle-1 to trigger the Toll pathway and antimicrobial protein synthesis (An et al., 2010). Like PAP3, PAP1 activates PPOs directly and via a positive feedback loop in the presence of a complex of SPHI (1a, 1b, 4 or 101) and SPHII (2) (Wang and Jiang, 2008; Jin et al., 2022). Guided by transcriptome and proteome analyses, we have discovered a constitutive SP module of HP14b-SP144-GP6 in the extraembryonic tissues of eggs (Shan et al., 2022). Like their paralogs HP14(a), HP21, and HP5 in larvae, these three constitutively expressed SPs may work with the other CLIPs to promote rapid early immune responses. Then, the inducible module of HP14a-HP21-HP5 may phase in to trigger enhanced activation of PPO and proSpätzle. Crosstalk between the two modules is supported by the activation of proHP5 by SP144 to promote the continuous protection of early embryos against invading pathogens. HP6 resides at the junction for melanization and Toll pathways, and its activating enzymes (HP5 and GP6) are responsible for bacteria and fungi responsive activation of the SP-SPH system in larval hemolymph and extraembryonic tissues. In Drosophila, two orthologs of HP6, Hayan and Persephone, play critical roles in systematic wound responses (Nam et al., 2012), sensing proteases from pathogens (Issa et al., 2018) and endogenous proteases activated via pattern recognition receptors (Buchon et al., 2009), and triggering melanization and the Toll pathway (Shan et al., 2023).

Additional activating proteases may exist in M. sexta to produce active HP6 in response to wounding and development. One candidate is proHP1*, a CLIPD member that hydrolyzes Leu-Asp-Leu-His-p-nitroanilide, stimulates PPO activation, and is then down-regulated by serpin-4 and serpin-5 in covalent complexes (Tong et al., 2005; Yang et al., 2016; He et al., 2017). While LDLH⁹² is the recognition sequence for proHP6 activation, the role of proHP1* and proposed conformational transition of proHP1 to proHP1* have not yet been established. Another candidate HP6-activating protease is HP17, a 35 kDa fragment of which somehow associated with the serpin-protease complexes. After a paralog of HP17 (now named HP17a) was uncovered in the M. sexta genome project, it was named HP17b (Cao et al., 2015). HP17b is 79% identical and 91% similar to the short form of HP17a. In this paper, we report the function of HP17b as a proHP6 activating protease, its inhibition by serpins, and some unusual features of this CLIPD and its activating enzyme.

2. Methods and Materials

2.1. Insect rearing, bacterial challenge, hemolymph collection, tissue dissection, and quantitative reverse transcription (qRT) PCR analysis

M. sexta eggs were purchased from Carolina Biological Supply and larvae were reared on an artificial diet (Dunn and Drake, 1983). Each day 2, fifth instar larva was injected with a mixture of killed bacteria and curdlan in 20 μl H₂O (He et al., 2018). Hemolymph was collected from cut prolegs of the larvae 24 h later and centrifuged at 5000×g for 10 min to harvest induced hemocytes (IH). The supernatants were pooled and labeled as induced plasma (IP). Induced fat body (IF) was dissected for total RNA extraction. Similarly, control plasma (CP) and hemocytes (CH) were collected from three day 3, fifth instar naïve larvae prior to dissection of integument, fat body (CF), midgut, Malpighian tubules, muscles, nervous tissue, salivary glands, and trachea. Hemolymph was collected from bar-stage pharate pupae to prepare plasma (BP) for elicitation with M. luteus and proHP17b in the presence of 0.01% 1-phenyl-2-thiourea (PTU) prior to SDS-PAGE and immunoblot analysis. CP, IP, and BP (100 μl each) were mixed with equal volume of 100% saturated ammonium sulfate to remove most serpins in the supernatant and enrich hemolymph SP(H)s in the pellet, which was dissolved in 50 μl of 20 mM Tris-HCl, pH 7.5, 5 mM CaCl₂, 0.001% Tween-20, and 0.01% PTU for elicitation with M. luteus. To measure and compare HP17b mRNA levels in integument and fat body at different stages, the tissues were dissected from insects ranging from 4^th instar larvae to adults for RNA extraction. All the total RNA samples (1 μg each) were incubated with 1× iScript Reverse Transcription Supermix (Bio-Rad) in a 20 μl reaction incubated at 25 °C for 5 min and 42 °C for 60 min to synthesize cDNA. After the reverse transcriptase was denatured at 85 °C for 5 min, qRT PCR was carried out to determine the mRNA levels of HP17b relative to ribosomal protein S3. Equal amounts of cDNA samples were used as templates by incubating with 1× iTaq^™ Universal SYBR Green Supermix (Bio-Rad) and specific primers. The primer pairs were: j937 (5′- AAAAATGAAGGTGTTTGGTTG) and j938 (5′- AGGGTTAGACTTTTGCTTGGC) for HP17b. The thermal cycling conditions and data normalization were done as described before (He et al., 2018).

2.2. Recombinant expression, purification, and characterization of proHP17b and HP17b from insect cells

The region coding for mature proHP17b was amplified from a fat body cDNA sample of early pupae using primers j933 (5’-GAATTCAAAACGGTGAAACATGCCA) and j505 (5’-GGGCTCGAGCGCGTGCCCGAGCAC). After cleavage with SacII and NdeI to remove possible contamination of HP17a cDNA, the 1.1 kb PCR product was separated by agarose gel electrophoresis, recovered from the gel, and cloned into pGEM-T (Promega). After sequence validation, the EcoRI-XhoI fragment was subcloned into the same sites in pMFH6 which encodes the honeybee melittin signal peptide and a 6×His tag at the C-terminus. The resulting plasmid was used to generate a recombinant baculovirus for producing proHP17b as described before (He et al., 2018). The HP17b precursor, along with active HP17b and 35 kDa intermediate, was isolated from 300 ml of the conditioned medium by cation exchange chromatography on a dextran sulfate Sepharose column (40 ml). The three proteins were further purified by nickel affinity chromatography. After concentration and buffer exchange on Amicon Ultra-30 centrifugal filter devices (Millipore), the mixture of proHP17b, 35 kDa intermediate, and HP17b in 20 mM Tris-HCl, pH 7.5, 50 mM NaCl, designated HP17b’, was aliquoted, rapidly frozen in liquid nitrogen, and then stored at −80 °C. In the second batch, the conditioned medium (1000 ml) was loaded onto another column of dextran sulfate Sepharose (80 ml) to further separate proHP17b and HP17b with a linear gradient of 0–1.0 M NaCl. The proHP17b and HP17b pools were separately recovered by nickel affinity chromatography, concentrated, and exchanged to 20 mM Tris-HCl, pH 7.5, 50 mM NaCl. Half of the native proHP17b (0.5 mg) was used as an antigen to generate a rabbit polyclonal antiserum (Cocalico Biologicals Inc.). N- and O-linked glycosylation of proHP17b was examined by treating with PNGase F (N), a mixture of neuraminidase A and O-glycosidase (O), and a mixture of the three glycosidases (N+O) for 2 h at 37 °C. The reaction mixtures and untreated control (C) were separated by SDS-PAGE under reducing condition followed by immunoblot analysis using antibody to the hexahistidine tag. The proHP17b, HP17b, and HP17b’ were analyzed by SDS-PAGE, followed by Coomassie brilliant blue (CBB) staining and immunoblot analysis.

2.3. Preparation of M. sexta proHPs, HPs, proSPHs and serpins for functional tests

Precursors of HP1a, HP2, HP5, HP6, HP8, HP14a, PAP1, Spatzle-1, SPHI (1a, 1b, 4, and 101) and SPHII (2) were isolated from baculovirus-infected Sf9 cell cultures in the previous studies or produced for this study according to the published protocols (Yang et al., 2016, He et al., 2018, Wang et al., 2020, Jin et al., 2022). Active HP2, HP5, and HP21 were freshly prepared using their activating enzymes as reported before (Wang and Jiang, 2007, He et al., 2018, Wang et al., 2020). Since most of the insect serpins are readily expressed as soluble proteins in E. coli, we produced and purified serpin-1A, 1J, 1J’, 3–6, 9, 12, 13 by nickel affinity chromatography for this study (Wang et al., 2020). Other proteins (e.g., PGRP1, MBP, βGRP1, βGRP2, PPOs) were prepared for functional assays that yielded results not shown (Wang et al., 2022b),

2.4. Measurement of the amidase activity of HP17b’ and HP17b using peptidyl-p-nitroanilides

Aliquots of HP17b’ (1 μl, 800 ng) or HP17b (1 μl, 620 ng) were separately incubated at room temperature with 150 μl of 25 μM FPRpNA, GGRpNA, LDLHpNA, LNNRpNA, VGRpNA, GRpNA, IEGRpNA and IEARpNA (custom synthesized by ChinaPeptides) in 0.1 M Tris-HCl, 0.1 M NaCl, 5 mM CaCl₂, pH 7.8. Increases in absorbance at 405 nm were monitored in kinetic mode on a microplate reader. One unit of activity is defined as the amount of enzyme causing ΔA₄₀₅ of 0.001 per minute.

2.5. Role of HP17b in PPO activation in hemolymph

CP or IP, proHP17b, HP17b’ or HP17b, M. luteus, and buffer A (20 mM Tris-HCl, pH 7.5, 20 mM NaCl, 5 mM CaCl₂, 0.001% Tween-20) and control reactions were incubated for 1 h before PO activity measurement (Jiang et al., 2003). Statistical significance of the differences was analyzed by Students’ t-test. To detect exogenous HP17b activation, CP, IP or BP, M. luteus, proHP17b, and buffer A containing 0.01% PTU were incubated for 15 min prior to 10% SDS-PAGE and immunoblot analysis using 6×His antibody. Activation of endogenous proHP17b was detected in 0–50% ammonium sulfate fraction of the control, induced, and bar-stage hemolymph to remove serpins and increase sensitivity. Three samples were directly subjected to SDS-PAGE. Another three were incubated with M. luteus for 15 min at room temperature, treated with SDS sample buffer, separated along with the other samples, and analyzed using diluted HP17b antiserum as primary antibody.

2.6. Sequential activation of HP17b’, HP6, and PAP1

To test HP6 activation, proHP6 was incubated with HP17b’ or proHP17b for 1 h at room temperature. The reactions and controls were subjected to 10% SDS-PAGE and immunoblot analysis using HP6 antibodies. Activation of proPAP1 by HP6 was verified by incubating HP17b’-treated proHP6 with proPAP1 for 1 h. The reaction mixture and controls were separated by 10% SDS-PAGE under reducing and nonreducing conditions. The proPAP1 and its cleavage products were detected using diluted PAP1 antiserum. Amidase activity of PAP1 was measured using IEARpNA as a substrate (Section 2.4). To find out the optimal pH for proHP6 cleavage, aliquots of proHP6, HP17b’, H₂O, and Polybuffer 96 (Sigma-Aldrich) at different pH were incubated for 1 h at room temperature prior to immunoblot analysis.

2.7. Activation of proSPH1b, 101, 2, and proSpätzle-1A by HP17b’

The proSPHs were separately incubated with HP17b’ and buffer A for 1 h at 37 °C and subjected to 12% SDS-PAGE and immunoblot analysis using diluted SPH antisera. Sizes of immunoreactive bands were compared with the previous study (Jin et al., 2022) to confirm the cleavages at the first and second sites. HP17b’, proHP6, proPAP1, proSPH1b, and proSPH2 were incubated for 1 h and then with PPO to test whether HP17b’-cleaved proSPH1b and 2 act as a cofactor for PPO activation. PO activity was measured using dopamine as a substrate. HP17b’, proSpätzle-1A, and buffer A were incubated for 1 h at 37 °C prior to 12% SDS-PAGE and immunoblot analysis to detect direct activation. Spätzle-1 antiserum was kindly provided by Dr. Michael R. Kanost at Kansas State University.

2.8. Isolation and characterization of serpin-protease complexes using HP17b antibodies

As described previously (He et al., 2017), HP17b antibodies were coupled to Protein A-Sepharose for isolating serpins that form covalent complexes with HP17b in induced plasma. Twenty ml of IP was incubated with M. luteus at 0.5 μg/ml for 30 min at 25 °C in the presence of 0.01% PTU and 10 mM diethyldithiocarbamate. After treating the mixture with 1 mM phenylmethylsulfonyl fluoride (PMSF) for 10 min at 25 °C and flocculant removal, the plasma was loaded onto the antibody column (2.0 ml). Following washing with 20 ml of 1 M NaCl and 20 ml of 10 mM sodium phosphate, pH 6.5, bound proteins were eluted with 10 ml 100 mM glycine (pH 2.5) into tubes containing 50 μl of the neutralizing buffer (1.0 M sodium phosphate, pH 8.0). The fractions were subjected to 10% SDS-PAGE, CBB staining, and immunoblot analysis using different antisera as the primary antibodies.

2.9. Serpin inhibition of HP17b’-mediated proHP6 activation by forming SDS-stable complexes

Aliquots of HP17b’ (800 ng in 1 μl) were separately incubated with serpins (0.5–1.0 μg in 1 μl) and buffer A (17 μl) for 20 min at 25 °C. ProHP6 (90 ng in 1 μl) was added to each reaction and further incubated for 1 h at 37 °C. After 12% SDS-PAGE and electrotransfer, the blot was analyzed using 1:2000 diluted HP6 antiserum as the primary antibody. In order to confirm that the block of proHP6 activation was due to HP17b’ inhibition, HP17b’ (800 ng in 1 μl) was incubated with serpins (0.5–1.0 μg in 1 μl) in buffer A (10 μl) for 20 min at 25 °C. After 10% SDS-PAGE and electrotransfer, the blot was analyzed using 6×His antibody to detect the tagged proteins and their complexes.

2. Results and discussion

3.1. Structural features and expression profiles of M. sexta HP17b

The HP17b gene (XP_030025933.1) resides on chromosome-6 (11,237,600–11,247,000) (Fig. S1), next to HP17a gene (XP_030025924.1, 11,205,300–11,225,000) (Cao et al., 2015). Exon 1 encodes a 19-residue signal peptide and Q¹N of the mature proHP17b, exon 2 encodes a major part of the group-1b clip domain, exon 3 encodes V⁵⁶CCP…NAKL¹²⁸ that spans the linker between the clip (C⁶–C⁵⁸) and protease (V¹²²–L³⁶⁵) domains, and exons 4–8 encode the remaining catalytic domain (G¹²⁹–L³⁷⁵). After predicted proteolytic activation between SFKR¹²¹ and VING¹²⁵, the N-terminal light chain is expected to be attached to the C-terminal heavy chain containing the protease domain via an interchain disulfide bond (C¹¹²-C²³⁵). Note that the dibasic site (KR¹²¹) may be subject to cleavage activation by an intracellular convertase that prefers Arg/Lys at P1 and P2/P4. Separation of the HP17b chains is expected to occur in the presence of a reducing agent such as dithiothreitol (DTT). A part of the linker region (P⁶¹–P¹¹⁰) is acidic, rich in Pro and Thr, and probably O-glycosylated. With D³⁰⁹, G³⁴² and G³⁵² forming an S1 pocket and H¹⁶⁵, D²¹⁵ and S³¹⁵ forming a catalytic triad, active HP17b is predicted to be a trypsin-like S1 protease that cleaves after a positively charged residue (e.g., Arg). Sequence alignment of the clip and protease-related domains of 43 M. sexta CLIPs indicated that HP17b belongs to the CLIPD group, which also includes HP17a, HP1a, HP1b, SP52 and SP60 in subgroup D1 and SP131, SP132, and SP140–143 in subgroup D2 (Cao et al., 2015).

HP17b mRNA increased in larval fat body after an immune challenge, suggesting the protease plays a role in antimicrobial defense responses (Fig. 1A). In day 3, fifth instar naïve larvae, the transcripts were more abundant in muscle and nervous tissues than in fat body. The highest mRNA level was detected in trachea, 36.5- and 15.5-fold higher than those in muscle and nervous tissues (Fig. 1B), respectively, indicative of a potential role in immunity of the respiratory system. Since HP17a expression was high in epidermis (data not shown), we examined the developmental series but found low expression of HP17b in integument (Fig. 1C). In comparison, HP17b expression in fat body was about 10-fold higher (Fig. 1D). A peak of HP17b mRNA was found in wandering and early pupal stages. The highest level was detected in fat body from pharate pupae with a pair of metathoracic bars, indicating a developmental control of the HP17b gene expression. This is consistent with the previous RNA-seq data of HP17b in fat body and muscles from naïve insects (Cao et al., 2015).

Fig. 1.

M. sexta HP17b immune inducibility and expression profiles in various tissues at different stages. (A) Immune inducibility. Relative mRNA levels of control and induced hemocytes and fat body (CH, IH, CF, and IF) were calculated based on Ct values of three biological replicates (3 larvae per sample) and plotted as a bar graph (mean ± SD, n = 3). Pairwise comparisons were performed between C and I using Students’ t-test (*, p < 0.05; ns or not significant, p > 0.05). (B) Tissue distribution. RNA samples of integument (I), fat body (F), hemocytes (H), midgut (G), Malpighian tubules (MT), muscles (Mu), nervous tissue (N), salivary glands (SG), and trachea (T) from day 3, 5^th instar naïve larvae were prepared and analyzed by qPCR using the same method. (C, D) Developmental profiles of HP17b transcripts in integument and fat body. The mRNA levels in the two tissues from larvae, pupae, and adults at different stages were determined. “L4”, 4^th instar larvae, day 2; “L5”, 5^th instar (“e” for early or day 1; “m” for middle or day 2; “1” for late or day 3–4); “W”, wandering (“e” for early or day 1; “m” for middle or day 2.5; “1” for late or day 5.5); “B”, bar stage; “P1/9”, pupal day 1 or 9; “A”, adult day 2.

3.2. Recombinant production and characterization of M. sexta proHP17b

To study its function and regulation, we cloned HP17b cDNA, constructed proHP17b/pMFH6 plasmid, generated a baculovirus to infect Sf9 cells, and isolated 0.4 mg of HP17b’ (a mixture of proHP17b, a 35 kDa HP17b intermediate, and HP17b at a ratio of 16.3 : 1 : 2.4, as estimated by densitometry Fig. 2A, left) from 300 ml of the conditioned medium by ion exchange and nickel affinity chromatography. Reducing and nonreducing SDS-PAGE, Coomassie staining, and immunoblot analyses showed that proHP17b migrated mainly as two bands in the range of 50–60 kDa (Fig. 2A), larger than the theoretical M_r of 42,293 Da for GIQ¹N…HA³⁷⁵LEHHHHHH, where the underlined part corresponds to the mature proHP17b (Fig. S1). This preparation also contained 35 and 30 kDa bands that were recognized by the HP17b antibody. Automated Edman degradation of the 30 kDa band showed this fragment starts with V¹²²INGD, matching the predicted N-terminus exposed after cleavage between R¹²¹ and V¹²². In the absence of DTT, HP17b light and heavy chains migrated as a single band (*) with a mobility similar to that of proHP17b (arrow). The 35 kDa band did not shift to a higher position under non-reducing condition, indicating a cleavage in the linker before C¹¹². This intermediate product is similar in size to the cleaved proHP17 in hemolymph, which was associated with serpin-protease complexes in the pulldowns from serpin-4 and serpin-5 antibody columns (Tong et al., 2005).

Fig. 2.

SDS-PAGE separation of HP17b’, proHP17b, HP17b, and glycosidase-treated proHP17b followed by staining or immunoblot analysis. (A) HP17b’. After cationic exchange and nickel affinity chromatography, the mixture of proHP17b, 35 kDa intermediate, and HP17b was resolved by 12% SDS-PAGE under non-reducing (−DTT) and reducing (+DTT) conditions and subjected to Coomassie brilliant blue staining (CBB, 2.4 μg, left) or immunoblot analysis using 6×His antibody (0.24 μg, right). (B) ProHP17b. CBB, 3.0 μg, left; immunoblot, 0.3 μg, right. (C) HP17b. CBB, 1.0 μg, left; immunoblot, 0.2 μg, right. (D) Deglycosylation of proHP17b. The purified proHP17b (1 μg) was treated with buffer (C for control), PNGase F (N), neuraminidase A and O-glycosidase (O), or all three enzymes (N+O) prior to SDS-PAGE and immunoblot analysis. The proHP17b, HP17b catalytic domain, and HP17b (chain unseparated) bands are marked black arrow, red arrowhead, and red asterisk, respectively. In the absence of DTT, proHP17b and HP17 migrated to similar positions. Sizes and positions of the M_r makers are indicated on the right, with the 50 kDa band highlighted red.

In another trial of purification, proHP17b and HP17b were successfully separated by cation exchange chromatography after one liter of the conditioned medium had been loaded onto a new column of dextran sulfate Sepharose. Using nickel affinity chromatography, we purified 1.0 mg of the proenzyme (Fig. 2B) and 186 μg of HP17b (Fig. 2C) from the pools of dextran sulfate column fractions. Half of proHP17b was used as antigen to prepare a rabbit polyclonal antiserum against the native protein. The N-terminus of the 30 kDa band (Fig. 2B) was determined to be V¹²²INGD, identical to that of the 30 kDa band in HP17b’. In the absence of DTT, most of the HP17b heavy chain shifted to 50–60 kDa (*).

Following PNGase F treatment, no mobility change was observed in the banding pattern of proHP17b (Fig. 2D). However, incubation with O-glycosidases did result in decreased size, confirming that proHP17b was O-glycosylated in the linker region. As the spread of 35 kDa bands was less than that of the 50–60 kDa proHP17b (Fig. 2A), we suspected that the proteolysis occurred somewhere between T⁷⁹ and S⁹⁶ to generate the 5 kDa increase in apparent M_r as compared with the protease domain (Fig. S1). In other words, both N- and C-terminal fragments were O-glycosylated. We did not determine N-terminal sequence of the 35 kDa band because PTH-Thr and PTH-glycoThr peaks are difficult to detect and assign due to low signal sensitivity, changed retention time, and heterogeneity in glycosylation.

The HP17b’ preparation had an amidase activity that hydrolyzed some peptidyl-p-nitroanilides tested. The predicted active form, HP17b (98 ng in 800 ng of the protein mixture), efficiently hydrolyzed IEARpNA (5.6 U), IEGRpNA (5.0 U), FPRpNA (3.0 U), and barely hydrolyzed LNNRpNA, LDLHpNA, VGRpNA, GGRpNA, or GRpNA (≤ 0.3 U, Fig. 3A). IEARpNA has been widely used as a substrate to detect insect PAPs but HP17b’ did not cleave PPO (data not shown). This finding is consistent with the poor hydrolysis of LNNRpNA, which is (nearly) identical to M. sexta PPO1 and PPO2 at the activation site of L(S/N)NR⁵¹*F. In contrast, the poor hydrolysis of LDLHpNA did not agree with the cleavage activation of proHP6 at LDLH⁹²*ILGG (see below). To our surprise, after HP17b (620 ng) was incubated with these synthetic substrates, we only detected a low amidase activity of 0.02–0.40 U (Fig. 3B). The 89-fold decrease in specific activity towards IEARpNA indicated that, upon separation from proHP17b and the 35 kDa intermediate, HP17b experienced a major loss of catalytic activity, with its light and heavy chains still attached by the disulfide bond (C¹¹²-C²³⁵) (Fig. 2C). This is reminiscent of the cleaved but inactive HP1a after Factor Xa treatment of proHP1axa (He et al., 2017). In HP17b’, proHP17b or, more likely the 35 kDa intermediate, may have maintained the active conformation of HP17b. A 35 kDa fragment of HP17 but not proHP17 was a component of the high M_r immune complex pulled down by antibodies to serpin-4 or serpin-5 (Tong et al., 2005). As proHP17b is likely activated by an unknown protease cleaving between R¹²¹ and V¹²² during immune response in hemolymph, we hypothesize that additional hemolymph protein(s) may act as an auxiliary factor to localize and protect the amidase activity of HP17b, more or less like PPO activation by PAPs in the presence of SPHI-II complexes (Jin et al., 2022).

Fig. 3.

Amidase activities of HP17b’ (A) and HP17b (B) measured using peptidyl-p-nitroanilide substrates. As described in Section 2.4, aliquots of HP17b’ (1 μl, 800 ng total, containing 98 ng HP17b) and HP17b (620 ng) were incubated with individual substrates (150 μl, 25 μM) and increases in A₄₀₅ were monitored on a microplate reader. The amidase activities (mean ± SD, n = 3) were plotted as a bar graph for comparison.

3.3. M. luteus-independent PPO activation by HP17b

Do proHP17b/HP17b’/HP17b and M. luteus together or separately induce melanization in control plasma (CP) from naïve larvae? CP represents a low, constitutive state of immunity. To detect possible synergism of HP17b addition and bacterial recognition, we adjusted the experimental conditions so that M. luteus alone, a good immune elicitor, did not cause much PPO activation (0.2 and 0.4 U, Fig. 4A). After proHP17b was added, the PO activity increased to 12.2 and 13.3 U at 1 h in the absence and presence of M. luteus, respectively, suggesting the HP17b proenzyme became active during incubation (see below). When HP17b’ was used, the PO activity rose to 29.0 and 28.9 U (Fig. 4A), some of the increase was likely caused by the high amidase activity of HP17b (Fig. 3A). After HP17b only was added, the PO activities were elevated to 13.1 U and 13.4 U, indicating that the cleaved enzyme regained some of its activity by interacting with plasma factors during incubation. There was no synergistic increase in PPO activation when M. luteus was also present in the paired reactions.

Fig. 4.

Role of HP17b during PPO activation in control and induced plasma (CP and IP) after incubation with proHP17b, HP17b’, or HP17b in the absence or presence of M. luteus. (A, B) CP or IP (1:10 diluted, 5.0 μl), proHP17b, HP17b’ or HP17b (1 μg), M. luteus (1 μg), and buffer A (20 mM Tris-HCl, pH 7.5, 20 mM NaCl, 5 mM CaCl₂, 0.001% Tween-20) to a total volume of 20 μl were incubated 60 min at room temperature for CP and on ice for IP. PO activities (mean ± SD, n = 4) in the reaction mixtures were measured and plotted as a bar graph. Pairwise comparisons were performed using Students’ t-test (paired, 2 tails). ns, p > 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001. (C) Plasma (1 μl, CP, IP or BP), M. luteus (1 μg, 1 μl), proHP17b (0.2 μg, 1 μl), and buffer A containing 0.01% PTU (to 12 μl) were incubated at room temperature for 15 min prior to 10% SDS-PAGE and immunoblot analysis using 6×His antibody. (D) 50% ammonium sulfate (AS) suspensions (200 μl) of the control, induced, and bar-stage hemolymph were spun at 12,000×g for 10 min at 4 °C. After supernatant removal, the protein precipitates were dissolved in 50 μl buffer A containing 0.01% PTU and clarified by centrifugation. Three supernatants (20 μl each, 0–50 AS fraction of CP, IP, and BP) were directly treated with 5 μl 6×SDS buffer (t = 0’). Another three samples (20 μl each) were incubated with 2 μg M. luteus (2 μl) at room temperature for 15 min and treated with 5 μl 6×SDS sample buffer. The reaction mixtures (6 μl each), equivalent to 10 μl plasma, were subjected to 12% SDS-PAGE and immunoblot analysis using 1:1000 diluted HP17b antiserum as primary antibody. The proHP17b and its catalytic domain are marked with black arrow and red arrowhead, respectively. Sizes and positions of the M_r makers are indicated on the left, with the 50 kDa band highlighted red.

Induced plasma (IP) represents an enhanced state that experienced a reduction of constitutive immune factors used in the initial defense against the injected bacteria and a large increase in inducible immune factors from transcriptional upregulation. Repeating the above test with proHP17b and IP (0.1 and 0.2 U) showed a low PO activity increase to 2.0 and 2.8 U at 1 h, because the mixtures were incubated on ice (Fig. 4B). Moderate PO activities were detected when the reaction occurred in 45 min at room temperature (data not shown). Much higher PO levels (33.6 and 35.9 U) were detected after HP17b’ was added to IP. Because the PO activities (39.1 and 41.1 U) were even higher after HP17b treatment, IP seemed to have a larger capacity than CP for purified HP17b to regain protease activity. No significant difference was found between each pair of PO activities, indicating that the presence of M. luteus did not increase HP17b-triggered PPO activation beyond an additive effect.

Analogous to proHP1a* and proHP1b*, which enhanced PPO activation in uncut but active form (Yang et al., 2016; He et al., 2017), HP17b is another member of the CLIPD subfamily that stimulated melanization in an unusual way, i.e., independent of M. luteus, E. coli, S. cerevisiae, or their cell wall components (Fig. 4; data not shown). To better understand this process, we examined the fate of proHP17b alone after incubation with CP, IP, and BP and identified the 30 kDa catalytic domain of HP17b using 6×His antibody (Fig. 4C). BP represents a developmentally controlled state of immunity that is active against microbial infection during this critical stage of development. Again, no difference in the band intensity was observed in the reactions with or without M. luteus. In BP, all proHP17b was converted to HP17b by an unknown protease, likely regulated by a stage-specific signal. The endogenous HP17b was detected at a similar level in 0–50% ammonium sulfate fraction of BP at 0’ and 15’ after M. luteus treatment (Fig. 4D). A lower level of 30 kDa HP17b was present in the AS fraction of IP at 15’, somewhat higher than 0’ but not necessarily related to M. luteus due to the difference in incubation time. Here, AS fractionation of plasma removes serpins and other negative regulators of immune responses and enriches hemolymph proteases that promote immunity and allow better detection of HP cleavage.

3.4. Possible mechanisms of proHP17b activation

We first tested if HP17b may be a cascade initiating protease by activating its zymogen via autocleavage between R¹²¹ and V¹²² in the presence of microbial elicitors. Incubation of proHP17b with peptidoglycans, PGRP1, and MBP or with β-1,3-glucan (curdlan, zymosan, laminarin, or S. aureus) and βGRP1/2 did not generate active HP17b (data not shown). Direct treatment with peptidoglycans, lipopolysaccharide, lipoteichoic acid, lipid A, curdlan, zymosan, laminarin, M. luteus, E. coli, S. cerevisiae failed to activate proHP17b (data not shown). These two results are consistent with the lack of impact from M. luteus on HP17b-induced PPO activation (Fig. 4, A, and B).

Incubation of proHP17b with subtilisin from Bacillus subtilis or Pr1A from Metarhizium anisopliae did not yield active HP17 (data not shown). This suggests that HP17b, unlike Persephone or Hayan (Issa et al., 2018; Shan et al., 2023), did not sense danger signals from the pathogens.

We then tested several wound-exposed structures (chitin, collagen IV, phosphatidylserine, and proHP1a, an abundant hemolymph protein) as activators of proHP17b. The results were negative even after we combined all these molecules and incubated them with proHP17b (data not shown). However, we have not yet excluded the possibility that HP17b is a wound responsive enzyme, since endogenous proteases released during tissue or cell ruptures were not tested. In fact, HP17b was activated by cleavage at the correct site, presumably after exposure to proteases from lysed Sf9 cells.

M. sexta hemolymph proteases (HP2, HP5, HP6, HP8, HP14a, HP21, PAP1–3) did not activate proHP17b (data not shown), which reduced the probability that a low level of HP17b stimulates a positive feedback mechanism through one of these HPs.

Finally, we tried to directly purify the activity of HP17b activating enzyme in BP (Fig. 4C). After only two column steps, the activity became too low to follow (data not shown).

3.5. Sequential activation of proHP6 and proPAP1 by HP17b’

Knowing the critical role of HP6 in the serine protease network, we tested if HP17b’ directly acts on proHP6 as a substrate to generate HP6, which then causes PPO activation in plasma (Fig. 4, A and B). Incubation with HP17b’ generated a 29 kDa doublet from the 38 kDa proHP6 (Fig. 5A). In the absence of DTT, the catalytic domain remained attached to the light chain of HP6 so that the HP6 doublet migrated to a position of 35 kDa. ProHP17b did not activate proHP6 (Fig. 5B). As we expected from the previous study (An et al., 2009), HP6 activated by HP17b’ cleaved proPAP1 into the heavy and light chains, which remained associated by the interchain disulfide bond in the absence of DTT (Fig. 5C, left). Lack of proHP6 failed to yield PAP1 directly by HP17b’ (Fig. 5C, right). Active PAP1 more efficiently hydrolyzed IEARpNA than HP17b’ and HP6 did (Fig. 5D). Therefore, HP17, HP6, and PAP1 constitute an SP branch to induce PPO activation. As reported before (Wang and Jiang, 2008), a positive feedback loop involving PAP1, SPH2, and other HPs greatly enhanced the melanization response.

Fig. 5.

Sequential activation of HP17b’, HP6, and PAP1 in vitro. (A, B) Proteolytic activation of proHP6 by HP17b’ but not proHP17b. The HP6 precursor (1 μl, 0.1 μg) was incubated with buffer A, HP17b’ (1 μl, 0.8 μg), or proHP17b (1 μl, 1.0 μg) in a total volume of 12 μl for 1 h at room temperature. The reactions, controls, and their replicates were treated with 1×SDS sample buffer with or without DTT prior to 10% SDS-PAGE and immunoblot analysis using 1:2000 diluted HP6 antiserum. (C) Activation of proPAP1 by HP17b’-treated proHP6. After proHP6 (0.1 μg) was activated by HP17b’, proPAP1 (1 μl, 0.1 μg) and buffer A (11 μl) were added and incubated for 1 h. Half of the mixture and controls were separated by 10% SDS-PAGE under reducing and nonreducing conditions. The proPAP1 and PAP1 were detected using 1:2000 diluted PAP1 antiserum as the primary antibody for immunoblot analysis. (D) Amidase activity (mean ± SD, n = 3) in the reaction of control mixtures were measured using IEARpNA (Section 2.4) and plotted in the bar graph. (E) Optimal pH for proHP6 cleavage activation. Aliquots of proHP6 (1 μl, 0.1 μg), HP17b’ (1 μl, 0.8 μg), H₂O (8 μl), and Polybuffer 96 (Sigma-Aldrich, 2 μl at indicated pHs) were incubated for 1 h at room temperature. The reaction mixtures were analyzed as described above. Sizes and positions of the M_r makers are indicated on the left, with the 50 kDa band highlighted red.

Since proHP6 has a unique cleavage site after His⁹², we examined the pH-dependence of its cleavage by HP17b’ and found the formation of 29 kDa double bands was most efficient in a range of pH 6.5–7.5 (Fig. 5E). At pH 5.0, proHP6 was cleaved at another site upstream of His⁹², generating a 32 kDa doublet not shifting to the position of proHP6 in the absence of DTT (data not shown). The sidechain of His⁹², located on the surface for cleavage, should have a pKa similar to that in free histidine at 6.0. At pH 6.5, 7.0, and 7.5, 32%, 10%, and 3.2% of His⁹² is expected to carry a positive charge and, thus, can be cleaved by trypsin-like HP17b. While a lower pH favors the protonation of His⁹², it may also cause a conformation change to hide His⁹² and expose the new site (possibly Lys⁶⁸ or Arg⁷²) at pH 5.0. The cleavage efficiency may reflect a balance of the opposite effects at a particular pH.

3.6. Search for plasma proteins affected by HP17b

To better understand the functions of HP17b, we tested if other precursors of HPs (i.e., HP1, HP2, HP5, HP21, PAP1–3) can be directly activated by HP17b’. HP17b’ either did not cleave these proenzymes or cleaved some at a position different from the predicted site. For instance, HP17b’ cleaved purified proHP1a and proHP1b, but the cleavage products did not shift to a higher position near the proenzymes under nonreducing condition (Fig. S2), indicating that the cleavage did not occur at AQGR¹³⁴*FV.

As PAP1 generates active SPH2 and PO through a positive feedback mechanism (Wang and Jiang, 2008), we tested whether HP17b’ plays a similar role and found it did not cleave PPO (data not shown). Interestingly, HP17b’ directly activated proSPH1b, proSPH101, and proSPH2 by cleaving at the same sites as PAP3 did to generate a high M_r cofactor for PPO activation (Fig. 6, A–C). The activation of proSPH4 and proSPH1a was much less prominent, and the order of cleavage efficiency and cofactor activity were identical to those reported before using PAP3 (Jin et al., 2022). The SPH1b and SPH2 generated by HP17b’ were active as a cofactor for PPO activation (Fig. 6D) by PAP1 activated also by HP17b’ through HP6 (Fig. 5).

Fig. 6.

Cleavage activation of proSPH1b, 101, and 2 for PPO activation by PAP1. (A–C) Precursors of SPH1b (A), SPH101 (B), and SPH2 (C) (1 μl, 0.2 μg) were separately incubated with HP17b’ (1 μl, 0.8 μg) and buffer A (22 μl) for 1 h at 37 °C. The reactions and controls were treated with 1×SDS sample buffer with or without DTT prior to 12% SDS-PAGE and immunoblot analysis using 1:2000 diluted SPH antisera as indicated. The proSPHs and their cleavage products at the first and second sites are marked with black arrow and red arrowhead, respectively. Sizes and positions of the M_r makers are indicated on the left, with the 50 kDa band highlighted red. (D) To test whether HP17b’-cleaved proSPH1b and 2 act as a cofactor for PPO activation, HP17b’ (1 μl, 0.8 μg), proHP6 (1 μl, 0.1 μg), proPAP1 (2 μl, 0.1 μg), proSPH1b (1 μl, 0.3 μg), proSPH2 (1 μl, 0.2 μg), and buffer A (18 μl) were incubated for 1 h at room temperature, along with the controls. PPO (1 μl, 160 ng) was added to the reaction mixtures and, after incubation on ice for 1 h, PO activity (mean ± SD, n = 3) was measured using dopamine as a substrate (Jiang et al., 2003) and plotted in a bar graph.

As reported by An and colleagues (2009 and 2010), HP6 also activates proHP8 and HP8 activates Spätzle-1A, a ligand of Toll receptor in M. sexta. We confirmed this extracellular pathway in vitro (data not shown). To test if a more direct activation mechanism exists, we incubated HP17b’ with proSpätzle-1A and found more Spätzle-1A at 16 kDa in the presence of DTT (Fig. 7). Under non-reducing condition, Spätzle-1A dimer migrated to 30 kDa position, consistent with the disulfide-linked dimer. In addition to cofactor-regulated melanization, HP17b’ may also more directly participate in the Toll pathway activation by activating the Spatzle-1A precursor.

Fig. 7.

Direct activation of proSpätzle-1A by HP17b’. HP17b’ (1 μl, 0.8 μg), proSpätzle-1A (5 μl, 0.25 μg), and buffer A (14 μl) were incubated for 1 h at 37 °C, along with the controls. The reactions and controls were treated with 1×SDS sample buffer with or without DTT prior to 12% SDS-PAGE and immunoblot analysis using 1:2000 diluted Spätzle-1 antiserum as primary antibody. ProSpätzle-1A and Spätzle-1A are marked with arrow and red arrowhead, respectively. The positions and sizes of M_r standards are indicated on the left, with the 50 kDa band highlighted red.

3.7. Inhibition of HP17b by M. sexta serpins

To test whether HP17b is inhibited by a serpin, we incubated 20 ml of IP with M. luteus for 30 min at 25 °C to activate immune protease cascades, in the presence of PO inhibitors to prevent protein crosslinking, and then loaded the mixture on an HP17b antibody column to isolate HP17b and its associated proteins. While a large amount of 30–35 kDa HP17b fragments were present in the bound fraction (Fig. 8A), both HP17b and serpin-4 antibodies recognized a 60 kDa band, which could be a cleavage product of the typical 70–75 kDa serpin-protease complex. We then tested if some of the ten M. sexta serpins were able to block HP17b’-mediated proHP6 activation. The 29 kDa doublet appeared after incubation with HP17b’ with buffer, serpin-3, 9, 12, and 13 but not with serpin-1A, 1J, 4, 5, and 6 (Fig. 8B). Serpin-1J’, an allelic variant of serpin-1J, incompletely blocked proHP6 activation. Serpin-4 and serpin-5 formed SDS-stable complexes with the catalytic domain of HP6 at about 75 kDa, as reported previously (Tong et al., 2005). To better understand the inhibition of HP17b, we directly incubated HP17b’ with the serpins that blocked proHP6 activation (Fig. 8B) and detected the serpin-protease complexes at 70–75 kDa (Fig. 8C). These data together indicated that endogenous HP17b was scarce, but it formed a covalent complex with serpin-4 at least. Other serpins including 1A, 1J, 1J’, 4, 5, and 6, may contribute to different extents to the regulation of HP17b.

Fig. 8.

Inhibitory regulation of proHP17b by M. sexta serpins. (A) As described in Section 2.5, the elution fraction containing HP17b (14 μl/lane) was treated with the sample buffer and separated by 10% SDS-PAGE. After electrotransfer, the nitrocellulose membrane was cut into strips which were separately incubated with diluted antisera against M. sexta HP17b and serpin-4. The 60 kDa band are marked with blue arrowhead. (B) Blocking of proHP6 activation by serpins. Aliquots of HP17b’ (1 μl, 0.8 μg) were separately incubated with a serpin (1 μl, 0.5–1.0 μg) and buffer A (15 μl) for 20 min at 25 °C. ProHP6 (90 ng in 1 μl) was added to each reaction and further incubated for 1 h at 37 °C. After 12% SDS-PAGE and electrotransfer, the blot was analyzed using 1:2000 diluted HP6 antiserum as the primary antibody. ProHP6, HP6 catalytic domain, and its covalent complex with serpin are marked with arrow, red arrowhead, and asterisk, respectively. (C) Formation of SDS-stable complexes of HP17b and serpins. HP17b’ (1 μl, 0.8 μg) was incubated with serpins (1 μl, 0.5–1.0 μg) in buffer A (10 μl) for 20 min at 25 °C. After 10% SDS-PAGE and electrotransfer, the blot was analyzed using 6×His antibody to detect the tagged proteins and their complexes (*). The positions and sizes of M_r standards are indicated on the left, with the 50 kDa band highlighted red.

3.8. Concluding remarks

We have characterized an additional clip domain protease, HP17b, which participates in immune cascades in M. sexta plasma. Although the mechanism of proHP17b activation remains a mystery, we established its roles in producing HP6 for melanization through PAP1, SPH1b, and SPH2. Alone or along with HP6 and HP8, HP17b also produce Spätzle-1 that induces the Toll pathway to stimulate synthesis of antimicrobial peptides and other immune factors. Serpin-1, 4, 5, and 6 inhibit HP17b and likely regulate its activity in immune responses to prevent overreaction. The lack of response to bacteria, fungi, microbial surface molecules, and known hemolymph proteases points us to several new directions for future research. For instance, HP17b activation may be responsive to other pathogens such as intracellular bacteria or viruses. The HP17b precursor could be activated by a protease not in the S1A family, e.g., lysosomal cysteine proteases from ruptured cells upon wounding. Consistent with the fact that CLIPDs are evolutionarily more ancient than CLIPCs and CLIPBs, proHP17 activation may be responsible for a conserved function in metazoans, such as sterilization and healing of wounds, but our initial tests of wound-exposed structures did not mimic the process faithfully. Finally, a parallel study on proHP17a would greatly enrich our knowledge on the physiological functions and activation mechanism of the two closely related CLIPDs.

Highlights.

• HP17b, a CLIPD1 closely related to HP17a, mediates immune responses.

• HP17b may be activated in response to danger but not bacteria, fungi, or their wall components.

• HP17b triggers melanization and Toll signaling by directly activating HP6, SPHI-II, and Spätzle-1.

• Serpin-1, 4, 5, and 6 inhibit HP17b to down-regulate immunity.

Abbreviations:

CP, IP, and BP

control, induced, and bar-stage plasma (i.e. cell-free hemolymph)

CH, IH, CF and IF

hemocytes, and fat body from naïve or immune challenged larvae

DTT

dithiothreitol

βGRP

β-1,3-glucan recognition protein

HP

hemolymph (serine) protease

MBP

microbe binding protein

PO and PPO

phenoloxidase and its proenzyme

PAP

proPO activating protease

PG and PGRP

peptidoglycan and its recognition protein

SP and SPH

serine protease and non-catalytic serine protease homolog

Data Availability Statement

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