Peptidoglycan recognition protein S2 is crucial for activation the Toll pathway against Israeli acute paralysis virus infection in honey bee Apis mellifera

Although honey bee responses to pathogens have been systematically described in the past decades, antiviral signalling pathways mechanisms are not thoroughly characterized. To decipher direct antiviral roles of an immune pathway, we firstly used the infectious clone of Israeli acute paralysis virus (IAPV) to screen 42 immune genes involved in mTOR, MAPK, Toll, Endocytosis, Jak-STAT pathway and homeobox protein, heat shock protein, as well as antimicrobial peptides (AMPs), and found that Toll pathway was a potential predominant immune pathway in Apis mellifera. Consistent with this, only dsRNA-PGRP-S2 treated A. mellifera significantly exhibited impaired activation of Toll pathway, promoting susceptibility to the IAPV infection. Finally, immunofluorescence results confirmed that the Toll pathway was initiated by peptidoglycan recognition protein S2 (PGRP-S2) interacting with Toll protein. Co-immunoprecipitation findings also further preliminarily confirmed PGRP-S2 directly interacting with viral capsid protein IAPV-VP3 to induce the activation of the Toll pathway in A. mellifera. These findings highlight that the Toll pathway is demanded efficient inhibitions of IAPV replication as a specific antiviral pathway in A. mellifera, and PGRP-S2, acting as a pattern recognition receptor, could be a new approach for control of the viral disease. Author summary Honey bee viruses, particularly IAPV, had been implicated in the colony decline with a global distribution resulting in insufficient pollination services. However, little is known about the antiviral mechanism of honey bee. In this study, we found that the Toll pathway was required for A. mellifera against IAPV infection and initiated by PGRP-S2. We also confirmed that dsRNA-PGRP-S2 treated A. mellifera exhibited impaired Toll pathway activation and promoted susceptibility to the IAPV infection. As a result, we employed co-immunoprecipitation technique to identify the interaction between the PGRP-S2 with Toll. Moreover, it was found the PGRP-S2 directly recognized IAPV-VP3 to activate the immune pathway against IAPV infection. Our work provides novel evidence that honey bees own a specific antiviral immune pathway and suggests that targeting PGRP-S2 could be a new approach for controlling the viral disease.

In this study, we used IAPV infectious clone constructed by our laboratory to screen 6 154 the immune genes involved in the significant immune pathways of honey bee on a large 155 scale [21]. Our results showed that Toll pathway was mainly activated for defending 156 against viral infection of honey bees. A short-type peptidoglycan recognition protein 157 (PGRP-S2) was essential for viral induction of the Toll pathway and activation of AMP 158 genes of defensin 1 and hymenoptaecin to combat the IAPV infection. Our results also 159 revealed that PGRP-S2 was an essential pattern-recognition protein required upstream 160 of the Toll pathway, which was a route-specific to resist viral infection in honey bees. 161 Overall, these findings demonstrated a specific and novel mechanism by which the

IAPV Infection Predominantly Induced Expressions of the Toll Pathway
167 associated genes 168 The previous study has shown that honey bees harboured their antiviral mechanisms 169 [22]. Therefore, the IAPV infectious clone was used to avoid the potential interference 170 from other viruses and explore the precise role in response to viral infection [23]. Our 171 results showed that IAPV infectious clone could effectively reduce honey bee survivals 172 and achieve IAPV proliferation (S1 Fig). Then, according to the reported immune genes 173 and pathway against IAPV in Honey bee (Table S1) [17,19,24], we investigated the 174 expression levels of representative genes involved in immune pathways, including 175 mTOR, MAPK, Toll, endocytosis, Jak-STAT pathways and homeobox protein, heat 176 shock protein genes, as well as AMPs of the honey bee. Analysis of immune genes by 177 qRT-PCR showed that 6 out of 14 genes were significantly up-regulated and the rest 178 were down-regulated in all treated groups (Fig 1). Our results showed that IAPV 179 infection elevated ≤ 2-fold the expression of stat of JAK-STAT pathway but induced 180 the down-regulation in gene of mTOR pathway (Fig 1), which were consistent with that 181 of bees naturally-IAPV-infected [17]. Among these 6 obviously up-regulated genes, 182 peptidoglycan recognition protein S2 (PGRP-S2), homeobox protein Hox-C6b, 7 183 defensin 1, lysozyme 1 and hymenoptaecin were up-regulated at least more than two 184 times, even higher. Especially, PGRP-S2, hymenoptaecin and defensin 1 related to Toll 185 pathway were enhanced more than 4-fold after the IAPV treatment. Furthermore, 186 PGRP-S2 was significantly increased over the survey period in all groups treated with 187 IAPV (p < 0.01). In contrast, apisimin and defensin 2 were down-regulated more than 188 8 times. 189 The results obtained above raised the issue of the immune signalling pathway immune pathways, as well as other genes from Jak-STAT, RNAi, melanization 196 pathway, and two heat shock proteins, had no significant difference before and after 197 viral infection (Fig 2E-T) (Fig 3A). These results showed that the Toll signalling pathway might be the 202 IAPV-specific inducible immune response.

203
Generally, the midgut, hemolymph and fat body, as the most important immune 204 organ in insects, were also used as a candidate tissue to test the expression of genes 205 confirmed by the results above. As shown in Fig 3B, (Fig 3A). As expected, the PGRP-S2, Toll, cactus1, hymenoptaecin and defensin1 were 212 significantly altered after IAPV infection, and they exhibited a higher expression level 8 213 in the fat body than that in the midgut (Fig 3C-M). Moreover, späetzle 3, späetzle 4, 214 Myd88, pellino, dorsal and dorsal-2, to a certain extent, were significantly up-regulated 215 after IAPV infection both in fat body and midgut, especially späetzle 3 in fat body ( Fig   216   3C-M). However, all of the rest genes, such as cactus 3, späetzle 5, dorsal-1a and 217 dorsal-1b did not increase (S2 Fig). Then, immunofluorescence staining using an anti-

218
Toll polyclonal antibody showed that Toll was localized in the fat body cell membrane 219 but not midgut (Fig 3N). These findings suggested that the Toll signalling pathway was 220 the the major response to IAPV infection in the fat body of honey bee. a 50 µg/mL (Fig 4A). The PGN exhibited the expected results at the concentration of 235 10 µg/mL when it was mixed with IAPV, and then this concentration was used in the 236 following study (Fig 4B). As expected, we found that the PGN (10 µg/mL) not only 237 help the bee to activate the Toll pathway and elicit expression of PGRP-S2, Toll, 238 Defensing 1 and hymenoptaecin to against IAPV infection (p < 0.01) (Fig 4C-G), but 239 also reduced the titer of IAPV (Fig 4H). Consistently, compared to the IAPV group, 240 the level of expression of these four genes was up-regulated in bees treated with PGN 241 and IAPV all the time, except for days 7, which may be consistent with less IAPV titer

PGRP-S2 is crucial for the Toll pathway activation
245 PGRP-S2, as a member of the PGRP family, showed the most outstanding levels of up-246 regulation, and its expression level was more than 20 times higher than the control at 247 48 h after IAPV infection in the fat body. Next, to investigate whether PGRP-S2 is 248 crucial for Toll pathway mediated antiviral responses to IAPV infection in honey bee, 249 we silenced Toll and each PGRPs (PGRP-S1, 2 and 3), respectively, which were 250 significantly changed after IAPV infection, and then measured their expression levels 251 by qPCR. Our results revealed that silencing of PGRP-S2 (p <0.01) and Toll (p <0.05) 252 significantly reduced the survival of honey bees, and the mass death of bees mainly 253 ranged from days 3 to 5 (Fig 5A). Meanwhile, silencing of PGRP-S2 but not PGRP-S1 254 or PGRP-S3 significantly down-regulated the expression level of Toll (p <0.05) (Fig   255   5B). As expected, qPCR and western blot assay results showed, silencing of Toll and 256 PGRP-S2 showed increased susceptibility to IAPV (p <0.01) (Fig 5C-D) and 257 progressed to inhibit the generation of IAPV structural proteins, VP2 (Fig 5E). Notably, 258 the virus abundance was promoted significantly at the protein level in the absence of 259 PGRP-S2 and Toll in a time-dependent manner from days 3 to 5 (p <0.01) when IAPV 260 copies were >10 7 genomes/5ng RNA, similar to the detection limit for IAPV-VP2 (Fig   261   5E). This result was further confirmed by quantitatively assessing the target band 262 intensity (Fig 5F). Yet, these inhibition effects were no more until day 6 because the 263 timeliness of dsRNA treatment mainly ranged from days 2 to 6 (S4 Fig). 264 In addition, on the basis of the above results, we reasoned that PGRP-S2 might 265 induce AMPs expression via the Toll pathway. Therefore, we proposed whether PGRP-266 S2 and its downstream genes, AMPs, is reduced treatment with dsRNA on PGRP-S2 S2 and dsRNA-Toll groups was higher than that of hymenoptaecin (S5 Fig A-D). 274 What's more, the dsRNA-PGRP-S1 treated with IAPV-infected bees significantly 275 reduced the expression of defensin1 only at days 3 (S5 Fig E), while the dsRNA-PGRP-276 S3 treated with IAPV-infected bees significantly facilitated the expression of defensin1 277 at days 4 and 5 (S5 Fig F). These results suggested an essential function of PGRP-S2  form dimer and lead to activate the Toll pathway (Fig 6F). Subsequently, the co- , PGRP-S2 was also a natural dimer expressed in bees (Fig 6G). In 301 addition, the interaction between PGRP-S2 and Toll was specific since Toll did not 302 bind biotinylated IgG (Fig 6G).

303
Given that the up-regulated expression of spatzle3/4 and PGRP-S2 was required for 304 activation of Toll pathway after IAPV infection in honey bees, immunofluorescence 305 and Co-immunoprecipitation results also confirmed that Toll pathway was initiated by 306 PGRP-S2 interacting with Toll in honey bee, we sought to determine the possible 307 interaction between PGRP-S2 and capsid proteins of IAPV. According to the genome 308 structure of IAPV (Fig 7A), we obtained the soluble PGRP-S2, VP1, VP2 and VP3 309 proteins of IAPV with His-tagged in vitro (Fig 7B-C) and identified their specificity 310 using rabbit anti-VP1, -VP2 and -VP3 polyclonal antibody (Fig 7D). Co-IP assays and 311 western blot analysis using mouse anti-PGRP-S2 polyclonal antibody with total cell 312 lysates from IAPV-infected bee samples showed that PGRP-S2 can interact with VP3 313 whereas others can not (Fig 7E). We further confirmed this interaction using purified 314 PGRP-S2 in vitro, and the signal was more substantial than that of cell lysates (Fig 7F). 315 Furthermore, a specific interactioln occurred between PGRP-S2 and VP3 as indicated 316 by the results that PGRP-S2 did not bind with either VP1 or VP2. These data suggest 317 that PGRP-S2 may directly recognize the IAPV via directly interacting with VP3 to 318 active the Toll pathway against IAPV infection in honey bees. The essential role for the insect Toll pathway in antimicrobial response is well set 333 up; however, whether the Toll pathway of honey bees serves as antiviral responses have 334 been poorly understood. Honey bee viruses can cause multiple immune responses 335 including Toll pathway but no evidence is associated with activation of this pathway.

336
In this study, we identified and confirmed that the Toll pathway was vital for antiviral 337 response in A. mellifera and was in line with that for DXV in Drosophila qPCR results as well as increased the viral abundance (Fig 3). These results indicated 352 that the Toll pathway might be a distinct antiviral response initiated by specific PGRP.

353
The one critical finding in this study were that honey bee Toll pathway was mainly 354 responsible for fighting IAPV infection. In our study, silencing Toll (Toll-1) exhibited 355 increased IAPV infectivity (Fig 4). Although a previous study had shown that Toll-7  The other critical findings found in this study were that PGRP-S2 directly recognized 384 IAPV-VP3, and passed the signal to Toll with the participation of spätzle 3/4 to activate 385 the Toll pathway. Although honey bees have four PGRP genes, three for short and one 386 for long, their differences in expression level between IAPV and control groups 387 prompted us to investigate the role of PGRP-S2 in the honey bee immune response (Fig   388   1, 2 and 3). Silencing PGRP-S2 exhibited increased IAPV infectivity, and honey bee  (Fig 7). Intriguingly, the interaction between PGRP-S2 406 and VP3 suggested that IAPV interacted directly with viral structural protein, unlike  Likewise, the expression of PGRP-S2 in the fat body was significantly higher than 423 midgut (p < 0.01), which was consistent with that of Bombus ignites but differed from 424 that of Bombyx mori [41,42]. Contrary to a previous study using the bacterial infection 425 in Drosophila, PGRP-LE induced the mainly immune response in the gut [36], our 426 findings did not reveal such role for PGRP-S2. In addition, IAPV infection did not 427 significantly enhance the expression of PGRP-LC (Fig 2), which is a kind of PRRs  Table). 488 The bee samples uninfected with the common viruses mentioned above were placed in   England Biolabs, Ipswich, MA) was used to clone three fragments and a T7 promoter 514 sequence (TAATACGACTCACTATAGGG) was linked at the 5′ end of the first 515 fragment. The reactions were carried out following the manufacturer's instructions. All 516 DNA fragments were purified using a QIA Quick gel extraction kit (Qiagen, Germany).

517
Then, the plasmid containing the full-length IAPV was sequenced to verify the 518 sequence stability. As previously described, a standard cloning procedure made IAPV   Table). The cDNAs of all samples were stored at -20℃ until use.

560
Individual bee was cut into two parts in dorsal midline, and half of the bee was used 561 for qPCR and the other half for western blot. cDNA produced by conventional RT-PCR 562 was subjected to real-time quantitative PCR using a LineGene9600 instrument. The were subsequently used to amplify target fragments using PCR from cDNA (S4 Table). 629 The PCR product was used to the transcript in vitro using MEGAscript RNAi Kit Ipswich, MA, USA) using specific primers (S6 Table). PCR products were inserted  Smirnov test and log-transformed if they did not show normality (S5 Table). For fold   immune defensin1 (C) and hymenoptaecin (D) at days 3, 4 and 5 after dsRNA injection.