Distinct Functions of Bombyx mori Peptidoglycan Recognition Protein 2 in Immune Responses to Bacteria and Viruses

Peptidoglycan recognition protein (PGRP) is an important pattern recognition receptor in innate immunity that is vital for bacterial recognition and defense in insects. Few studies report the role of PGRP in viral infection. Here we cloned two forms of PGRP from the model lepidopteran Bombyx mori: BmPGRP2-1 is a transmembrane protein, whereas BmPGRP2-2 is an intracellular protein. BmPGRP2-1 bound to diaminopimelic acid (DAP)-type peptidoglycan (PGN) to activate the canonical immune deficiency (Imd) pathway. BmPGRP2-2 knockdown reduced B. mori nucleopolyhedrovirus (BmNPV) multiplication and mortality in cell lines and in silkworm larvae, while its overexpression increased viral replication. Transcriptome and quantitative PCR (qPCR) results confirmed that BmPGRP2 negatively regulated phosphatase and tensin homolog (PTEN). BmPGRP2-2 expression was induced by BmNPV, and the protein suppressed PTEN-phosphoinositide 3-kinase (PI3K)/Akt signaling to inhibit cell apoptosis, suggesting that BmNPV modulates BmPGRP2-2-PTEN-PI3K/Akt signaling to evade host antiviral defense. These results demonstrate that the two forms of BmPGRP2 have different functions in host responses to bacteria and viruses.


INTRODUCTION
Innate immunity is a self-defense mechanism against infectious non-self entities and is present in all metazoans (1,2). The innate immune system of insects consists of humoral defenses that include the production of soluble effector molecules and cellular response like phagocytosis and encapsulation that are mediated by hemocytes (3). The innate immune response is mediated by germline-encoded pattern-recognition receptors (PRRs) that recognize pathogen-associated molecular patterns (PAMPs) that are present in pathogens but absent in the host (2,4,5). PAMPs include b-glucan, lipopolysaccharides (LPS) of gram-negative (G-) bacteria, and peptidoglycans (PGNs) of both gram-positive (G+) and G-bacteria, as well as bacterial and viral DNA and RNA and related molecules (2,4,5). LPS activates the expression of antimicrobial peptides (AMPs) in Manduca sexta (6) and malpighian tubules of Drosophila (7). PGN-an essential cell wall component in most bacteria-stimulates various immune reactions in insects (2,5,8). Lysine (Lys)-type PGNs are found in many G+ bacteria, whereas diaminopimelic acid (DAP)type PGNs are found in G-and some G+ bacteria (5,9,10). Multiple PRRs have been identified in invertebrates, such as lectin, hemolin, G-binding proteins (GNBPs), Toll-like receptors (TLRs), and PGN recognition proteins (PGRPs); the latter two have been shown to recognize PGN (2,5,10).
PGRP was first purified from the hemolymph of silkworm Bombyx mori (11). Since then, nearly 100 PGRP family members have been identified from insects to mammals. These proteins are highly conserved and have a PGRP domain that is similar to the bacteriophage T7 lysozyme, an N-acetylmuramyl-alanine amidase (9,12,13). Insect PGRPs are categorized as short (PGRP-S), which are small extracellular proteins with signal peptides, or long (PGRP-L), which have long transcripts and are intracellular, extracellular, or transmembrane proteins (12,14). Some PGRP-Ls have multiple splice forms; for example, the 13 Drosophila PGRP genes encode 19 proteins and the seven genes in mosquito (Anopheles gambiae) encode nine proteins (5,12,15,16).
PGRPs have multiple immune-related functions. In Drosophila, PGRP-SB1 and -SB2, -SC1A/1B/2, and -LB have zinc-dependent amidase activity involving the removal of peptides from glycan chains to inhibit or scavenge the biological activity of PGN. Some of these PGRPs modulate the host immune response by eliminating PGN (8,(17)(18)(19)(20). PGRP-SA, -SD, -LA, -LC, -LD, -LE, and -LF lack zinc-binding residues required for amidase activity while retaining the capacity to bind and recognize PGN (8,18). Other PGRPs lacking amidase activity mediate PGN-dependent activation of the prophenoloxidase (proPO) cascade (21,22) in which ProPO is activated to PO, leading to melanization that is toxic to microorganisms (11,23). Transmembrane PGRP-LC and -LE recognize DAP-type PGNs and activate the immune deficiency (Imd) pathway. The former binds PGN and interacts with Imd protein via the extracellular PGRP domain and an intracellular domain, respectively (15,24,25). PGRP-LB has amidase activity to cleave DAP-type PGNs, limiting availability of ligand for PGRP-LC and thus inhibiting the Imd pathway (26). PGRP-SA is an extracellular protein that recognizes Lys-type PGNs and activates Toll signaling. The serine protease cascade is triggered when PGRP-SA binds to PGN, which cleaves the inactive pro Spatzle (Spz) protein to an active form, which in turn binds to and activates Toll receptor (8,(27)(28)(29)(30)(31).
The model lepidopteran B. mori is an important insect because of its production of silk (42)(43)(44); as such, infection by pathogenic bacteria, fungi, and viruses can cause serious economic losses. B. mori nucleopolyhedrovirus (BmNPV) is the major threat to silkworms (45,46). A genome analysis revealed 12 PGRP genes in B. mori (47), some of which the function has been studied. Five BmPGRP-S showed strong amidase activity toward DAP-PGN (48). BmPGRP-S4 (49) and BmPGRP-S5 (50) bound PGNs to increase proPO activation. BmPGRP-S3 could be induced by B. mori cytoplasmic polyhedrosis virus (BmCPV) (51). Recently, our study showed that BmPGRP-S2 was up-regulated upon BmCPV infection (52), overexpression of which can activate the Imd pathway and induce increased AMPs to enhance the antiviral capacity of transgenic silkworm against BmCPV (53). In this study, we cloned two forms of BmPGRP2: BmPGRP2-1 bound to DAP-type PGN to activate Imd signaling, whereas BmPGRP2-2 was induced by BmNPV and negatively regulated the phosphatase and tensin homolog (PTEN)-phosphoinositide 3-kinase (PI3K)/Akt pathway to inhibit cell apoptosis. Our results demonstrate that the two forms of BmPGRP2 have distinct functions in the host response to pathogenic bacteria and viruses. The present study confirms that PGRP was induced by viruses to escape host antiviral immunity.

Protein Expression and Bond Test
The extracellular PGRP domain of BmPGRP2-1 was cloned with the primer BmPGRP2-1pro, which was used for prokaryotic expression in E. coli with the pSKB2-MsyB vector (pSKB2 vector was added with an MsyB tag in our lab). After induction of expression with 0.1 m mol/L isopropyl β-D-thiogalactoside (IPTG, #A100487, Sangon Biotech, China) at 16 • C for 16 h, the protein in the supernatant was purified and used for western blotting with a BmPGRP2-1 antibody. Purified BmPGRP2-1 was tested for binding to PGN-EB, PGN-BS, and LPS-EB by enzymelinked immunosorbent assay.
Transcriptome Analysis and Screening of Genes Downstream of BmPGRP2-2 PGRP2-2I and 932 were orally infected with BmNPV at 2.5 × 10 7 OB/larva on day 3 of the 5th instar. RNA was extracted from the midgut and fat body at 3, 6, 12, and 24 hpi, and used for qPCR analysis of BmPGRP2-2 expression. The RNA libraries of the 16 samples were constructed. Raw sequencing data were generated using an Illumina HiSeq 2000 system, which has been deposited in the NCBI, the BioProject ID is PRJNA521671. The polyA tails, noncoding RNAs and low-quality reads were removed from the raw reads to generate clean reads. The qualities of raw and clean reads were analyzed using FastQC (v0.11.1). Clean reads were mapped to the silkworm genome. The gene ontology (GO) analysis and KEGG annotation of the differentially expressed genes (DEGs) were executed using WEGO online and iPathCons, respectively (52,60). PTEN was cloned using the BmPTEN ORF primer, and its expression after BmPGRP2-2 RNAi and overexpression was evaluated. BmE cells were transfected with 1180 and 1180-A4P-PTEN-SV40; RNA was extracted and analyzed for BmPGRP2-2 expression. The 1180-A4P-GFP-PTEN-SV40 vector was constructed and transfected into BmE cells for localization analysis.

Analysis of Akt Phosphorylation and Cell Apoptosis
BmE cells were infected with BmNPV-GFP, and total protein was extracted for analysis of phosphorylated (p-)Akt expression at 0, 18, and 24 hpi. BmE cells were transfected with 1180 and the BmPGRP2-2 overexpression vector. BmN4-SID1 cells were incubated with dsBmPGRP2-2 and dsRed, and treated cells were infected with BmNPV-GFP (56). GAPDH, p-Akt, and total Akt levels were assessed by western blotting. The antibody of GAPDH (#CB100127, California Bioscience, USA) was used as a control. The PVDF membrane (#03010040001, Roche, Switzerland) was blocked with 5% BSA for 12 h at 4 • C after transfer film, followed incubation with antibody of p-Akt (Ser505, #4054, Cell Signaling Technology, USA) or total Akt (#9272, Cell Signaling Technology, USA) for 12 h at 4 • C. And then, the PVDF membrane was washed three times for 10 min each time with TBST, incubated with secondary antibody for 30 min at 37 • C, washed 3 × 10 min with TBST, and treated with SuperSignal TM West Femto (#34095, Thermo Scientific, USA). The images were analyzed following the manufacturer's instructions. Cell apoptosis was detected at 24 hpi by flow cytometry. The cells were collected from each treatment, washed with cold PBS and incubated with Annexin V-fluorescein isothiocyanate (V-FITC) and propidium iodide (#K201-100, Biovision, USA) at 25 • C for 20 min in the dark on ice. These samples were then analyzed by a fluorescence-activated cell sorter and Cell Quest software (BD, USA) following the manufacturer's instructions. Each test was repeated thrice. BmE cells were treated with the 10 µM PI3K inhibitor LY294002 (#HY-10108, MedChemExpress, USA), with dimethyl sulfoxide (DMSO, #D2650, Sigma, USA) treatment serving as a control (61). The expression of p-Akt was analyzed at 24 hpi and fluorescence was observed at 72 hpi.
A sequence analysis suggested that BmPGRP2-1 and-2 contained a PGRP domain, while a phylogenetic analysis revealed that BmPGRP2-1 and-2 clustered into distinct classes ( Figure S1). The RT-PCR results showed that BmPGRP2-1 was more highly expressed in the larva than that in the egg ( Figure 1B). The qPCR analysis showed that BmPGRP2-1 and -2 levels were lowest in the midgut, and that the expression of BmPGRP2-2 was higher than that of BmPGRP2-1 in all tissues of male and female larvae except for the midgut (Figure 1C). The predicted results from WoLF PSORT program on website showed that BmPGRP2-1 and-2 have no nucleotide localization sequences (NLS). A subcellular localization analysis indicated that BmPGRP2-1 was a transmembrane protein whereas BmPGRP2-2 was an intracellular protein ( Figure 1D). The differences in phylogenetic position (Figure S1), temporal and spatial expression patterns (Figures 1B,C), and subcellular localization ( Figure 1D) between the BmPGRP2-1 and-2 suggest that the two isoforms play distinct roles in host immune response to pathogens.

BmPGRP2-1 Binds to a Dap-Type PGN and Activates Imd Signaling
Multiple sequence alignment revealed that the zinc-binding cysteine residue of T7 lysozyme was replaced by a serine residue in BmPGRP2-1; the other seven PGRP proteins showed no amidase activity (Figure S2), suggesting that BmPGRP2-1 also lacks this activity. We tested whether BmPGRP2-1 can be induced by PGN. BmPGRP2-1 and glv2 but not BmPGRP2-2 levels were induced in silkworm larvae by the G-bacteria E. coli and S. marcescens (Figure S3). A qPCR analysis showed that BmPGRP2-1 expression was induced by DAP-type PGN-EB and PGN-BS but not LPS-EB or Lys-type PGN-SA (Figure 2A); imd and att2 levels were also increased in BmE cells treated with PGN-EB and -BS ( Figure 2B). However, Toll pathway activation in cells and individuals was unaltered by the treatment (data not shown). In addition, BmPGRP2-1 overexpression in BmE cells increased the levels of imd and att2 (Figure 2C), whereas silence of BmPGRP2-1 had the opposite effect ( Figure 2D) and spz level was unaltered. Thus, a  DsRNA targeting BmPGRP2-1 (dsPGRP2-1) and control dsRed were added to BmN4-SID1 cells. BmPGRP2-1, imd, spz, and att2 expressions were evaluated by qPCR. TIF-4A was used as the internal control of qPCR analysis, and each detection was performed thrice. (E) Interaction between BmPGRP2-1 and PGN-EB or -BS as determined by enzyme-linked immunosorbent assay. PGN-EB and -BS were set to each other's competitors; BmPGRP2-1 was pretreated with competitor (or left untreated) and the binding of BmPGRP2-1 with PGN-EB or -BS was evaluated based on alkaline phosphatase activity. LPS-EB was used as a control. Each assay was performed in triplicate. Student's t-tests were used to analyze the statistical data. ns: not significant. Bars represent standard deviation. **P < 0.01.

DAP-type PGN induces BmPGRP2-1 to activate Imd signaling and AMP.
Some studies have reported that PGRP binds to PGN through the extracellular PGRP domain to activate Imd signaling (15,24,25). To determine whether BmPGRP2-1 can bind to a DAPtype PGN, we purified a prokaryotic protein with a PGRP domain ( Figure S4A). Western blotting detected the target band in the purified protein ( Figure S4B). Alkaline phosphatase activity was higher in BmPGRP2-1+PGN-EB and BmPGRP2-1+PGN-BS than in BmPGRP2-1+LPS-EB and BmPGRP2-1. In a competition assay using PGN-EB and PGN-BS, pretreatment of BmPGRP2-1 with a competitor decreased alkaline phosphatase activity (Figure 2E), demonstrating that BmPGRP2-1 directly binds to DAP-type PGN.

BmPGRP2-1 Deficiency Decreases Resistance to Bacterial Infection in B. mori
To further clarify the function of BmPGRP2-1 in immune defense of silkworm against bacteria, we injected the transgenic RNAi vector pb-PGRP2-1I ( Figure S5A) into embryos to generate a transgenic strain of PGRP2-1I ( Figure S5B). The results of inverse PCR showed that a single copy of the transgene was detected in an intergenic region of the genome (Figure S5C). A qPCR analysis showed that BmPGRP2-1 and imd levels were lower in PGRP2-1I as compared to DZ larvae (Figure 3Aa). However, the viability of PGRP2-1I and DZ strains was unaffected by oral infection with E. coli and S. marcescens. The expression of BmPGRP2-1, imd, and att2 was also reduced in PGRP2-1I as compared to DZ pupae (Figure 3Ab), and (c) Analysis of viral DNA content by qPCR 48 h after BmNPV infection. Each DNA sample was extracted from 10 treated larvae. GAPDH (GenBank: AB262581.1) was used as the internal control, and each assay was performed thrice. Student's t-tests were used to analyze the statistical data. Bars represent standard deviation. *P < 0.05; **P < 0.01. the survival rate after S. marcescens injection was lower in PGRP2-1I than in DZ (Figure 3Ac). These results suggest that downregulation of BmPGRP2-1 inhibits Imd signaling and reduces the resistance of silkworm to G-bacteria.

BmPGRP2-2 Is Induced by BmNPV to Promote Viral Replication
BmPGRP2-2 but not BmPGRP2-1 expression was induced in BmE cells by BmNPV (Figure S6). To determine the role of BmPGRP2-2 in BmNPV infection, BmE cells were transfected with a BmPGRP2-2 overexpression vector containing a flag epitope. RT-PCR and western blot analyses confirmed that BmPGRP2-2 was upregulated in the cells (Figure 4Aa). Correspondingly, the fluorescence of the virus was enhanced (Figure 4Ab) and viral load was increased in cells overexpressing BmPGRP2-2 relative to controls (Figure 4Ac). These results suggest that BmPGRP2-2 overexpression promotes BmNPV multiplication. Furthermore, when BmPGRP2-2 dsRNA was added to BmN4-SID1 cells, BmPGRP2-2 expression was suppressed (Figure 4Ba), and viral fluorescence (Figure 4Bb) and viral DNA level (Figure 4Bc) were reduced, implying that BmPGRP2-2 knockdown inhibits BmNPV replication. Together, these data demonstrated that BmNPV induces BmPGRP2-2 to promote viral replication.

Antiviral Effects of BmPGRP2-2 Are Not Mediated via Imd and Toll Pathways
To clarify the role of BmPGRP2-2 in the immune response of silkworm larvae to BmNPV, we used the transgenic RNAi vector pb-PGRP2-2I ( Figure S7A) and generated the transgenic 932 strains PGRP2-2I and PGRP2-2Ia (Figure S7B) containing the transgene insertion in an intergenic region and in the intron of a predicted gene, respectively (Figures S7C). PGRP2-2I was selected for subsequent analysis. BmPGRP2-2 expression was decreased in PGRP2-2I as compared to the control 932 strain (Figure 3Ba). Moreover, mortality after BmNPV infection (Figure 3Bb) and accumulated BmNPV DNA content were significantly lower in PGRP2-2I than in 932 (Figure 3Bc). These results indicate that silencing BmPGRP2-2 enhances the antiviral capacity of silkworm.
Since Imd and Toll signaling pathways are involved in the antiviral immune response (37)(38)(39)(40), we investigated whether BmPGRP2-2 activates these pathways. BmPGRP2-2 was induced by BmNPV in larvae, with lower expression in PGRP2-2I than in 932 ( Figure S8A). However, there was no difference in the levels of imd, Relish, MyD88, Pelle, glv3, and glv4 of PGRP2-2I and 932 upon BmNPV infection (Figure S8B). These results suggest that BmPGRP2-2 mediates its antiviral effects through a pathway other than canonical Imd and Toll signaling.

BmPGRP2-2 Negatively Regulates BmPTEN
To identify the signaling pathway downstream of BmPGRP2-2, RNA was extracted from the midgut and fat body of PGRP2-2I and 932 at 3, 6, 12, and 24 hpi (Figure S9A). We confirmed that BmPGRP2-2 expression in both tissues was significantly lower in PGRP2-2I than in 932 by qPCR ( Figure S9B). And then, the 16 RNA samples were used for  File 1). The FPKM values of silkworm immune-related genes were shown in Table S2 after information analysis. The transcriptome analysis revealed that BmPTEN expression was higher in the midgut and fat body of PGRP2-2I than in those of 932 ( Figure S10); this was confirmed by qPCR ( Figure 5A). We also found that BmPGRP2-2 knockdown and overexpression increased and decreased the expression level of BmPTEN, respectively (Figure 5B), indicating a negative regulatory relationship. We also examined whether BmPGRP2-2 is regulated by BmPTEN; however, BmPTEN overexpression did not affect BmPGRP2-2 expression (Figure 5C). BmPTEN was an intracellular protein ( Figure 5D). These results indicate that BmPTEN is downstream of and negatively regulated by BmPGRP2-2.

BmPGRP2-2 Regulates p-Akt and Cell Apoptosis
Autographa californica multiple nucleopolyhedrovirus (AcMNPV) infection increases Akt phosphorylation in Sf9 cells (61). Akt is a downstream effector of PI3K-dependent cell survival, whereas PTEN is an inhibitor of PI3K/Akt signaling (62,63). We therefore investigated whether p-Akt is upregulated in BmE cells after BmNPV infection. BmNPV induced Akt phosphorylation, with the maximum level detected at 24 hpi ( Figure S11A). We also examined found that BmPGRP2-2 overexpression and knockdown increased and decreased Akt phosphorylation, respectively, at 24 hpi ( Figure 6A). Moreover, pharmacological inhibition of p-Akt by treatment with the PI3K inhibitor LY294002 suppressed BmNPV replication ( Figure 6B).
Previous studied have shown that Akt plays an important role in promoting cell survival by suppressing apoptosis (63,64). We examined whether BmPGRP2-2 regulates apoptosis via modulation of Akt activation. BmNPV infection of BmE cells resulted in the formation of apoptotic bodies (Figure S11B). A flow cytometry analysis showed that BmPGRP2-2 overexpression and knockdown reduced and increased the apoptotic fraction, respectively ( Figure 6C). These results indicate that BmPGRP2-2 induced by BmNPV inhibits apoptosis via Akt activation.

DISCUSSION
PGRP is well-known to play an important role in immune defense of insects against bacteria. In this study, we demonstrated that BmPGRP2-1 binds to DAP-type PGN to activate Imd signaling and inhibit bacteria, whereas BmNPVinduced BmPGRP2-2 suppresses host cell apoptosis to enable viral replication.
Multiple splice forms of PGRP-L have been identified in Drosophila and A. gambiae, although it is unclear how these are generated (5,12,15,16). Our results indicate that the two forms of BmPGRP2 are transcribed from different TIS. BmPGRP2-1 was generated from the distal TIS and contained a long 5 ′ UTR, whereas BmPGRP2-2 with a short 5 ′ UTR was transcribed from the proximal TIS. BmNPV is the primary pathogen to silkworm (46); we speculate that BmPGRP2-1 is the evolutionarily more ancient form, and that BmPGRP2-2 arose through interactions between the silkworm and BmNPV.
The two forms of BmPGRP2 showed distinct expression patterns, subcellular localization, and roles in the immune response to different pathogens. BmPGRP2-1 is a transmembrane protein that binds to DAP-type PGN via the extracellular PGRP domain, thereby activating the Imd pathway and AMP, similar to PGRP-LC in Drosophila. We therefore speculate that BmPGRP2-1 interacts with Imd to activate downstream signaling molecules through an PGRP-LC-like intracellular domain (15,24,25). Previous studies have demonstrated that DAP-type PGNs activates Imd signaling, whereas Lys-type PGNs activate the Toll pathway (8,65). We found here that BmPGRP2-1 was induced by G-bacteria and that silencing BmPGRP2-1 expression suppressed Imd signaling and AMP and reduced the resistance of transgenic silkworm to G-bacteria, confirming that BmPGRP2-1 plays a key role in the immune defense against G-bacteria and possibly against some G+ bacteria with DAP-type PGNs, although the latter requires confirmation.
BmPGRP2-2 was induced by BmNPV to promote viral replication; on the other hand, BmPGRP2-2 knockdown inhibited BmNPV. Some studies have reported that the Imd and Toll pathways are involved in antiviral immune response; inhibition of Imd signaling increased CrPV and alphavirus replication in Drosophila (37,38) whereas Toll pathway activation suppressed dengue virus replication in A. aegypti (40). The expression of PGRP-LA and PGRP-SC1A was up-regulated in Drosophila after infection of Nora virus (66). Sigma virus infection induced the expression of PGRP-SB1, PGRP-SD, and some AMP genes without altering Toll and Imd signaling in Drosophila (67). Imd pathway activation inhibited BmCPV in silkworm (53). The results of the present study demonstrate that antiviral signaling of BmPGRP2-2 against BmNPV is independent of the two pathways. The results of our previous study (53) and present study revealed the different responses of Imd pathway to different viruses in silkworm. JAK/STAT pathway could be activated by BmNPV but not BmCPV (34). This was unexpected, given that BmPGRP2-2 was found to negatively regulate PTEN and activate Akt to inhibit apoptosis. PGRP-LB has been shown to function as an negative regulator of Imd pathways in Drosophila (26). It was previously reported that PI3K-Akt signaling is required for efficient Baculovirus replication (61). The Hepatitis B virus HBx protein suppressed PTEN expression and activated Akt phosphorylation (68), while the hepatitis C virus NS5A protein-induced suppression of PTEN abrogated its inhibitory effect on PI3K/Akt signaling, triggering Akt activation to promote cell survival (69). This study demonstrated that virus-induced PRRs in the host negatively regulate the PTEN-PI3K/Akt pathway. Thus, the mechanism of BmNPV infection is similar to those of some human viruses, suggesting that B. mori and BmNPV can serve as a model of human-virus interaction, for instance to identify new host genes that are involved in viral infection and to screen antiviral drugs targeting the PTEN-PI3K/Akt pathway.
We found that BmPGRP2-2 negatively regulates PTEN-PI3K/Akt signaling to inhibit cell apoptosis, in contrast to the classic Imd and Toll pathways that are activated by PGRP. Apoptosis plays a key role in host defense against viral infection; however, viruses use various strategies to support cell survival, which is in turn beneficial for viral propagation (69). BmPGRP2-1 was induced by G-bacteria to activate Imd signaling and inhibit bacterial proliferation, whereas BmPGRP2-2 was induced by BmNPV to promote cell survival and consequently, viral replication. Thus, PGRP not only mediates host immune defense against bacteria but is also used by viruses to evade host antiviral defense systems (Figure 7). Some open questions are whether BmPGRP2-2 and other PRRs are used by other viruses; which viral components are recognized by BmPGRP2-2; the mechanism underlying negative regulation of PTEN by BmPGRP2-2; and whether there are other host genes or immune components involved in BmPGRP2-PTEN-PI3K/Akt signaling. FIGURE 7 | Model of BmPGRP2 functional differentiation in host immune response to pathogens. The two isoforms of BmPGRP2 were generated from different transcription initiation sites (TIS). BmPGRP2-1 recognizes and binds to DAP-type PGN through the extracellular PGRP domain and activates the Imd pathway and AMP to inhibit bacteria; BmPGRP2-2 is induced by BmNPV to suppress PTEN, thereby relieving its inhibition of PI3K/Akt signaling and triggering an increase in Akt phosphorylation and activation to inhibit cell apoptosis. The increased cell survival is beneficial for viral replication. (Part of the evidence for protein levels is lacking). Some drugs targeting the PTEN-PI3K/Akt pathway can inhibit viral infection. Blocking virus-induced PI3K/Akt signaling inhibited the replication of influenza virus (70). Emodin suppressed p-Akt to induce hepatocellular carcinoma cell apoptosis (71), whereas geridonin combined with paclitaxel induced apoptosis and inhibited the proliferation of gastric cancer cells via upregulation of PTEN and suppression of Akt phosphorylation (72). The PI3K inhibitor LY294002 was shown to block Akt activation and thereby inhibit AcMNPV (61) and BmNPV (this study). These drugs may be used in sericulture to control BmNPV infection if the production process and purity can be optimized. We have generated transgenic antiviral silkworms that block BmNPV infection via overexpression of Bmlipase-1 (55); by suppressing BmNPV mRNA by viral gene RNAi (45); and by inhibiting BmNPV protein synthesis via hycu-ep32 overexpression (73). In the present study, we confirmed that modulating host immune defense can inhibit BmNPV; overexpressing antiviral genes and dsRNA targeting viral genes can further enhance host resistance (74). Hence, combining four antiviral strategies (Bmlipase-1 and hycu-ep32 overexpression and silencing of BmPGRP2-2 and various viral genes) can potentially yield a transgenic silkworm with high resistance to BmNPV.
In conclusion, we cloned two forms of BmPGRP2 generated from different TIS and characterized their distinct functions in the host immune response to pathogens (Figure 7). Our findings indicate that BmPGRP2-2 does not function in canonical immune signaling pathways and PGRP2 is not only involved in host immune defense against invading pathogens, but is also used by viruses to evade host antiviral mechanisms.