Aminopeptidase MNP-1 triggers intestine protease production by activating daf-16 nuclear location to degrade pore-forming toxins in Caenorhabditis elegans

Pore-forming toxins (PFTs) are effective tools for pathogens infection. By disrupting epithelial barriers and killing immune cells, PFTs promotes the colonization and reproduction of pathogenic microorganisms in their host. In turn, the host triggers defense responses, such as endocytosis, exocytosis, or autophagy. Bacillus thuringiensis (Bt) bacteria produce PFT, known as crystal proteins (Cry) which damage the intestinal cells of insects or nematodes, eventually killing them. In insects, aminopeptidase N (APN) has been shown to act as an important receptor for Cry toxins. Here, using the nematode Caenorhabditis elegans as model, an extensive screening of APN gene family was performed to analyze the potential role of these proteins in the mode of action of Cry5Ba against the nematode. We found that one APN, MNP-1, participate in the toxin defense response, since the mnp-1(ok2434) mutant showed a Cry5Ba hypersensitive phenotype. Gene expression analysis in mnp-1(ok2434) mutant revealed the involvement of two protease genes, F19C6.4 and R03G8.6, that participate in Cry5Ba degradation. Finally, analysis of the transduction pathway involved in F19C6.4 and R03G8.6 expression revealed that upon Cry5Ba exposure, the worms up regulated both protease genes through the activation of the FOXO transcription factor DAF-16, which was translocated into the nucleus. The nuclear location of DAF-16 was found to be dependent on mnp-1 under Cry5Ba treatment. Our work provides evidence of new host responses against PFTs produced by an enteric pathogenic bacterium, resulting in activation of host intestinal proteases that degrade the PFT in the intestine.


Introduction
The pore-forming toxins (PFTs) play important roles in bacterial pathogenesis, including many important human pathogens, such as Staphylococcus aureus, Streptococcus pyogenes, Vibrio cholera, Clostridium perfringens, Enterococcus faecalis, among others [1][2][3][4][5][6].The different PFTs bind to membrane receptors such as membrane proteins, lipids, or cholesterol triggering conformational changes in these toxins leading to oligomer formation, and membrane insertion, resulting in their pore formation activity on the cell membrane. It has been documented that even at sublethal concentrations, the action of PFTs would modify the physiological state of cells by activating different cellular responses. Host cells could remove PFTs on plasma membrane by endocytosis and exocytosis, and may induce degradation of PFTs by activating autophagy. In addition, enhance resistance toward PFTs is also carried out by MAPKs pathways or by the inflammasome complex, or by triggering cell death programs such as apoptosis, necrosis or pyroptosis [7][8][9][10][11]. For example, some PFTs could be removed by membrane repair mechanisms such as endocytosis and exocytosis that are the most accepted membrane repair mechanisms [12]. In the case of cholesterol-dependent cytolysins (CDCs), pore formation induces an intracellular [Ca 2+ ] increase in the intoxicated cells, promoting internalization of the PFT in caveolaes, in a Ca 2+ dependent endocytosis process, that finally leads to PFTs degradation in lysosomes [13,14]. Likewise, exocytosis of α-toxin, CDCs or Cry5Ba toxins was shown to be dependent on the rise of intracellular [Ca 2+ ] induced by the PFTs, which in turn enhances shedding of microvesicles on the membrane that contains the PFTs pores [15][16][17]. In addition, different PFT activate signal transduction cascades like MAPKs and autophagy to remove PFTs from the membrane, and resist cell damage [18]. For example, the decrease in intracellular potassium induced by some PFTs, could activate the p38/MAPK pathway, which up-regulate multiple downstream genes involved in the activation of unfolded protein response (UPR) and innate immunity against PFTs intoxication [18,19]. The MAPK cascade is also known to regulate the phosphorylation level of the nuclear receptor fushi tarazu factor 1 (FTZ-F1), and the phosphorylation status of FTZ-F1 directly impacts the expression of the Cry1Ac receptor in Plutella xylostella, ultimately determining the tolerance or sensitivity of P. xylostella to Cry1Ac [20,21]. Moreover, autophagy has been documented as another efficient way to protect host cells against different PFTs, such as V. cholerae cytolysin (VCC), α-Haemolysin (HlyA) and α-toxin [22][23][24]. In the case of Caenorhabditis elegans, survival of this nematode to Cry5Ba PFT depend on autophagy activation that participates in degradation of the toxin pores [25].
Bacillus thuringiensis (Bt) is a spore-forming bacterium that upon sporulation, produces parasporal crystal proteins (Cry) that have specific toxicity to different insects, nematodes, mites, or protozoa species [26]. These Cry toxins disrupt the host intestine after the target organism ingest the parasporal crystal, leading to the host death. The Bt spores germinate and reproduce in the host cadaver [27,28]. Bt insecticidal proteins that accumulated in the crystal inclusions are PFTs which are the major Bt virulence factors. Based on their structural features, they can be divided into several families, such as the Cry three-domain toxins, the α-PFTs toxins now named App and β-PFTs such as aegerolysins now named as Gpp, etc [29]. The threedomain Cry proteins have been proposed to generate pores on the insect larvae or nematodes intestinal cell membrane. Cry proteins bind to receptors (such as aminopeptidase N, alkaline phosphatase or cadherin among others) and undergo conformation changes forming oligomers, that ultimately insert into the membrane to generate pores on the plasma membrane of midgut cells [30].
C. elegans is a powerful model system to study host responses to PFTs. Cry5Ba is a nematocidal Cry toxin, that after binding to a glycolipid receptor, disrupt the nematode intestine cells leading to worm death [31][32][33]. It has been shown that worms resist Cry5Ba intoxication by inducing multiple conserved innate immune responses, including the activation of p38 and c-Jun/MAPKs, the endoplasmic reticulum UPR and hypoxia response pathways, as well as the DAF-2 insulin/insulin-like growth factor-1 signaling pathway [34][35][36][37][38]. Furthermore, it was reported that at the intracellular level, worms could initiate endocytosis, exocytosis, and autophagy as defense strategies from Cry5Ba toxin action. However, how worms reduce toxicity when exposed to lethal concentration of Cry5Ba inside the intestine lumen has not been clearly elucidated.
In this study, we found that C. elegans modulate the transcription factor DAF-16 nuclear location to promote protease expression in the gut lumen, which in turn degrade Cry5Ba toxin. Our work explains how animals clear and inactivate PFTs in the intestine lumen, which represents a novel and additional strategy activated in the hosts to defend against pathogenic bacteria.

MNP-1 is required for protecting C. elegans against Cry5Ba toxin
The identification of C. elegans receptors involved in Cry5Ba mode of action revealed a role of apical cell glycolipids and cadherin CDH-8 [31,39]. In the case of insects, cadherin and aminopeptidase N (APNs), among other proteins, act as Cry toxin receptors [40,41]. Therefore, we analyzed whether APNs may also function as receptors for Cry5Ba in C. elegans. The APN protein family are enzymes that cleave neutral amino acids from the N terminus of polypeptides, which serve a variety of functions in a wide range of species from nematodes to mammals [42]. Two reported APN receptors of Cry proteins reported in insects contain two conserved domains, the catalytic center domain M1_APN_2 and ERAP1-C domain [43]. Thus, we identified all APNs that have the M1_APN_1 like domain in C. elegans genome. We found eight APNs containing both the catalytic center domain M1_APN_1 like and ERAP1-C domain, three APNs that only have the M1_APN_1 like domains, and other three APNs that have the Peptidase_M1_N domains (S1 Table, S1A-S1C Fig). To analyze the potential role of these APN´s as Cry5Ba receptors, we analyzed the toxicity of Cry5Ba toxin to APN mutants, since receptor mutations have been shown to be linked to resistance to different Cry toxins in several organisms [44,45]. A total of eight APNs mutants were obtained from Caenorhabditis Genetics Center (CGC) and evaluated by comparing their susceptibility to Cry5Ba with that of the wild-type N2 worms (S2 Table). The expression of other six APNs was down regulated by RNA interference (RNAi) assays to compare the susceptibility of these silenced animals after RNAi to Cry5Ba toxin in relation to the control nematodes. In these RNAi assays the animals were feed with HT115 E. coli cells that were transformed to express the different dsRNA, and the dsRNA expression was induced with IPTG as reported [46]. Control animals were treated with non-induced transformed E.coli cells. None of the apn mutants nor apn-silenced worms by RNAi that were analyzed showed a significant resistance phenotype to Cry5Ba, compared with the control animals, suggesting that the 14 APNs analyzed do not participate as receptors for Cry5Ba toxin in C. elegans (S2 Table).
However, we found out that one of the APN mutants, the mnp-1(ok2434), was 8.78-fold more sensitive to Cry5Ba (medium lethal concentration [LC 50 ] value of 1.929 ± 0.28 μg/mL) than N2 animals (LC 50 value of 16.926 ±0.8 μg/mL) (Fig 1A, and Table 1), suggesting that MNP-1 may be involved in protecting C. elegans against Cry5Ba toxin. To confirm the defense role of MNP-1 against Cry5Ba in C. elegans, we first analyzed mnp-1 expression in the worms. We generated a transgenic mnp-1p::gfp worms that express GFP protein fused to MNP-1 protein and confirmed that it is expressed in all stages from the egg to adult (S2 Fig). Secondly, the susceptibility to Cry5Ba was analyzed by using a growth inhibitory (GI) assays where the mnp-1(ok2434) mutant worms showed increased susceptibility to Cry5Ba than N2 control animals ( Fig 1B). Moreover, lifespan measurements were also performed in N2 and mnp-1(ok2434) mutant worms upon exposure to Cry5Ba and the control bacterial strain OP50 (Fig 1C,  Table 1). These assays showed that mnp-1(ok2434) mutant worms showed a significantly 2.6 fold lower lifespan than N2 control worms under Cry5Ba treatment. It is important to notice that the lifespan of N2 and mnp-1(ok2434) showed no difference under OP50 treatment. (Fig 1C, Table 1). Images of the mutant worms confirmed that in the presence of the Cry5Ba, the mnp-1(ok2434) mutant worms were more severely intoxicated compared to wild-type N2 worms, as evidenced by their smaller body-size ( Fig 1D). These results confirm that mnp-1 (ok2434) mutant is more sensitive to Cry5Ba than wild-type N2 worms.
To confirm the importance of MNP-1 in a defense response to Cry5Ba toxicity. The expression of MNP-1 was reduced in the wild type N2 worms by using RNAi (S3 Fig). Fig 1E shows that MNP-1-silenced individuals were more sensitive to Cry5Ba toxin, supporting that the Cry5Ba sensitive phenotype is caused by reduction of MNP-1 expression ( Fig 1E, Table 1). Finally, we confirmed the role of MNP-1 by complementing the mnp-1(ok2434) mutant with the wild type mnp-1 gene. The mnp-1(ok2434) mutant worms were transformed with the mnp-1 gene under regulation of its native promotor, showing that the rescued worms had similar sensitivity to Cry5Ba toxin when compared to the wild type N2 worms (Fig 1D and 1F), indicating that the mnp-1 gene was sufficient to restore wild type susceptibility to Cry5Ba toxin in the mnp-1(ok2434) mutant. Taken together, these data demonstrate that MNP-1 is required for C. elegans defense against Cry5Ba toxin.

MNP-1 is specifically involved in defense to Cry5Ba and Cry21Aa but not to other nematicidal toxins such as App6Aa PFT
There are several classes of nematicidal toxins produced by Bt such as Cry5Ba, Cry21Aa and App6Aa (before named Cry6Aa). The Cry5Ba and Cry21Aa belong to the Cry three domain family, while App6Aa toxin belongs to the ClyA-type α-pore-forming toxin family [29,47,48]. To test whether MNP-1 is required for protecting C. elegans against other nematicidal PFTs, the susceptibility of mnp-1(ok2434) mutant to App6Aa and Cry21Aa was compared with the wild-type N2 animals. As in the case of Cry5Ba, both the mortality (Fig 2A) and the GI assays ( Fig 2B) showed that mnp-1(ok2434) mutated animals were more sensitive to Cry21Aa than wild type N2 animals (Fig 2A and 2B, Table 1). However, in the case of App6Aa, the mnp-1 (ok2434) mutant and wild type N2 animals showed similar sensitivity when exposed to App6Aa toxin (Fig 2C and 2D, Table 1).

MNP-1 involved in C. elegans intestinal defensing PFTs
To analyze whether the hypersensitivity of mnp-1 mutant animals to Cry PFTs is not due to general sensitivity of these animals to different stresses, we tested the sensitivity of mnp-1 (ok2434) mutant to two other toxic chemical compounds, the heavy metal CuSO 4 or the oxidative stress agent H 2 O 2 [34]. The dose-dependent mortality assays showed that mnp-1(ok2434) mutant animals have a similar sensitivity as wild type N2 worms when exposed to either To determine whether MNP-1 response involves the up regulation of mnp-1 gene, we determined mnp-1 gene expression by quantitative RT-PCR (qPCR) in N2 worms exposed to the nematicidal Cry PFTs, CuSO 4 or H 2 O 2 as described in Materials and Methods. The expression of mnp-1 gene in N2 worms exposed to E. coli food OP50 at the same condition was used as control. Fig 2E shows that expression of mnp-1 gene was significant induced by exposure to Cry5Ba or Cry21Aa toxins, but showed no significant differences when worms were exposed to App6Aa, CuSO 4 or H 2 O 2 compared to worms treated with the OP50 food control (Fig 2E). These data indicate that mnp-1 gene was up regulated as a specific response to the exposure of C. elegans to the three domain nematicidal Cry5Ba and Cry21Aa PFTs.

MNP-1 defense against Cry5Ba proceed through induction of proteases expression
The mnp-1 gene encodes a membrane-associated APN that in C. elegans is required for embryonic muscle cell migration and neuronal cell migration [49,50]. To analyze if MNP-1 response against PFTs involves the regulation of certain innate immunity pathways, we compared the genomic responses using microarrays of N2 and mnp-1(ok2434) mutant after treatment with Cry5Ba, with the aim to identify genes whose differential-regulation depends on MNP-1. The wild type N2 and mnp-1(ok2434) mutant animals were fed with E. coli cells expressing or nonexpressing Cry5Ba toxin. Then their RNA was isolated, processed, and hybridized to Affymetrix arrays as described in Materials and Methods. The microarray analysis showed that 1575 genes from N2 have significant differential expression after Cry5Ba treatment (FC�2.0, p �0.05) compared to the nematodes that were feed with E. coli cells without Cry protein expression. In the case of mnp-1(ok2434) mutant, we found that 755 genes showed significant differential expression after Cry5Ba treatment (FC�2.0, p �0.05) compared with the control E. coli treatment (Fig 3A, detained information is shown in S1 Data). Gene ontology (GO) terms The p-value was determined by a Two-way ANOVA or Unpaired t-test (Welch's correction for unequal variances), ****p < 0.00001, ***p < 0.001, **p < 0.01, *p < 0.05 show significant differences and ns indicate no significant difference. https://doi.org/10.1371/journal.ppat.1011507.g002

PLOS PATHOGENS
MNP-1 involved in C. elegans intestinal defensing PFTs To analyze if these differentially expressed genes are dependent on mnp-1, we compared the changes in expression of up and down regulated genes following Cry5Ba treatment between N2 and mnp-1(ok2434). First, using the list of up and down regulated genes upon Cry5Ba exposure, we performed the separate linear regression analysis. The results revealed that both the induction and repression of genes in response to Cry5Ba were largely intact in mnp-1(ok2434) animals (r 2 = 0.6672 and 0.4952 respectively, p<0.0001 each) ( Fig 3B). Then, analyzes of the microarray data, showed that 747 up regulated and 255 down regulated genes in Cry5Ba-treated N2 were affected by the mnp-1 mutation ( Fig 3A). Analysis of these data through WormCat, a C. elegans bioinformatics platform that facilitates the further refinement of functional categories within enriched groups, illustrated that signaling, proteolysis proteasome, general proteolysis and stress responses are the top categories within the UP class genes [51]. In the case of DOWN class genes, it was found that genes involved in metabolism, proteolysis and stress responses are the top three categories (Fig 3C and S3 Table). Since mnp-1 mutant showed to be more sensitive to Cry5Ba toxicity, we proposed that it may be due to differences in genes involved in immune regulation. For this reason, we screened the 116 genes involved in proteolysis proteasome, general proteolysis and stress responses. Separate linear regression analysis showed that 106 genes were intact in mnp-1(ok2434) worms (r 2 = 0.5053, p<0.0001) (Fig 3D and S4 Table). In addition, we compared the average differential expression of these genes using a paired t-test. The average difference in induction (X 4 = X mutant -X wildtype ) was calculated. This approach accounts for both the magnitude and direction of attenuation of the response to Cry5Ba, with negative values of X 4 indicating attenuated induction. In mnp-1(ok2434) mutant, the average differences in induction of selected genes was significantly reduced compared with that of wild type (induction: X 4 = -0.856, p<0.0001). This analysis is consistent with the results of regression analysis ( Fig 3D). Also, we compared the transcription levels of ten genes from the 116 genes involved in proteolysis proteasome, general proteolysis and stress responses, in N2 and mnp-1(ok2434) worms, to verify that mnp-1 is involved in the induction of these genes upon Cry5Ba exposure. The transcription levels of the ten selected genes were lower in mnp-1 (ok2434) than in N2 (Fig 3E), which is consistent with the microarray data. All these results support that mnp-1 is required for the induction of many genes involved in proteolysis proteasome, general proteolysis and stress responses during Cry5Ba intoxication.
To confirm that some genes regulated by mnp-1 play a role in defending against Cry5Ba, we directed our attention to genes related in protein degradation. Among the genes found within the UP class, there are 27 genes involved in protein degradation (S4 Table). These data suggest that mnp-1 may regulate the expression of certain proteases and thereby mediate the defense against Cry5Ba action. To analyze this hypothesis, crude extracts samples of N2 or mnp-1 mutant worms were prepared as described in Materials and Methods and these samples were incubated with purified Cry5Ba protein to analyze their capacity for degradation of the Cry5Ba toxin. When Cry5Ba protein was incubated with 170 μg/ml crude extracts from N2 worms, 83.996% of this protein was degraded, while treatment of Cry5Ba with 34 μg/ml or 6.8 μg/ml crude extracts from N2 worms, resulted in 65.314% and 37.910% degradation of Cry5Ba protein, respectively. Fig 3F and 3G show the SDS-PAGE and statistical analysis of Cry5Ba band degradation after incubation with crude extracts at the same concentrations three independent replicates, error bars denote the SD in (E, G and I). The p-value was determined by One-way ANOVA or Unpaired t-test (Welch's correction for unequal variances), ****p < 0.00001, ***p < 0.001, **p < 0.01, *p < 0.05 show significant differences and ns indicate no significant difference. https://doi.org/10.1371/journal.ppat.1011507.g003

PLOS PATHOGENS
MNP-1 involved in C. elegans intestinal defensing PFTs from different worms. In the case of crude extracts from mnp-1 mutant worms, treatment of Cry5Ba with 170, 34 and 6.8 μg/ml extracts, resulted in 55.822%, 29.638% and 14.205% degradation of Cry5Ba protein, respectively (Fig 3F and 3G). In agreement, crude extracts from mnp-1 rescued worms showed recovered Cry5Ba degradation capability when compared with mnp-1(ok2434) mutant worms (Fig 3F and 3G), indicating the mnp-1 gene was sufficient to restore the Cry5Ba degradation capability in the mnp-1(ok2434) mutant background. Finally, the crude extracts of N2 and mnp-1 mutant worms showed no significant degradation of App6Aa protein (Fig 3H and 3I), supporting that MNP-1 is specifically involved in the defense against 3-domain Cry toxins.

Two proteases, F19C6.4 and R03G8.6, are responsible for MNP-1 mediated Cry5Ba degradation
To identify the proteases that are involved in Cry5Ba degradation, we screened the 19 proteases or protease inhibitor genes by gene silencing (RNAi) and analysis of Cry5Ba susceptibility of the silenced animals. S5 Table shows that silencing down two of these genes, F19C6.4 and R03G8.6, resulted in a hypersensitive phenotype to Cry5Ba toxin intoxication similar to mnp-1(ok2434) mutant, suggesting that these proteases are involved in the defense response against Cry5Ba. The dose-dependent mortality assays confirmed that F19C6.4 or R03G8.6 silenced animals were significantly more sensitive to Cry5Ba compared with N2 worms (Fig 4A). The LC 50 values of Cry5Ba against F19C6.4(ok2392) and R03G8.6(ok3143) mutant worms were 4.822±0.716 μg/ml and 7.520±0.751 μg/ml, respectively, while the LC 50 value of Cry5Ba against N2 control was 16.926±0.8 μg/ml (Table 1), indicating 3.5-and 2.25-fold times higher susceptibility in the mutants than in control worms. Moreover, the rescued animals in F19C6.4(ok2392) and R03G8.6 (ok3143) mutant worms with the corresponding protease genes F19C6.4 and R03G8.6 expressed by their native promoters showed that the sensitivity to Cry5Ba of these rescued worms was comparable to that of wild type N2 worms, confirming that F19C6.4 and R03G8.6 are involved in Cry5Ba intoxication (Fig 4A). The qPCR analysis showed that the expression of F19C6.4 and R03G8.6 genes was induced after Cry5Ba toxin treatment, and that this expression depend on MNP-1 since mnp-1(ok2434) mutant did not induce the expression of these genes, showing nosignificant difference in their expression in the presence or absence of Cry5Ba intoxication ( Fig 4B). Furthermore, to determine if Cry5Ba induces the expression of F19C6.4 and R03G8.6 proteins we constructed transgenic worms expressing F19C6.4p::gfp and R03G8.6p::gfp fused proteins. Fig 4C show that F19C6.4 was expressed in the intestine, cephalic sheath cells, and coelomocytes, whereas R03G8.6 protein was primarily found in the nematode gonads ( Fig 4C). The treatment of these transgenic nematodes with Cry5Ba toxin for 4 h showed a clear induction in the expression of F19C6.4 (Fig 4C and 4D). These data indicated that at least F19C6.4 is expressed in the gut and is involved in MNP-1 mediated protection of C. elegans against Cry5Ba.
To further analyze if these two MNP-1 controlled protease proteins are responsible for Cry5Ba degradation, we prepared crude extracts from F19C6.4(ok2392) and R03G8.6 (ok3143) mutants, as well as their corresponding rescued worms, and analyzed their capacity to degrade Cry5Ba toxin. The Cry5Ba-degradation assays showed that the crude extracts from F19C6.4 (ok2392) or R03G8.6(ok3143) mutant worms had significantly lower degradation of Cry5Ba protein than the crude extracts from the wild type N2, in contrast to the rescued worms that showed a similar Cry5Ba-degradation capacity as the wild type N2 (Fig 4E and 4F). In addition, to verify if F19C6.4 and R03G8.6 proteins could degrade Cry5Ba in vitro, we cloned and expressed these two proteins in E.coli strain BL21(DE3) and these protease proteins were purified as indicated in Material and Methods. Both purified proteins were then incubated with purified Cry5Ba, Cry21Aa and App6Aa at 37˚C. The results showed that both F19C6.4 and

PLOS PATHOGENS
MNP-1 involved in C. elegans intestinal defensing PFTs GST-R03G8.6 degraded Cry5Ba and Cry21Aa but not App6Aa (Figs 4G-4J and S6A-S6H). Furthermore, it was previously shown that rhodamine-labeled Cry5Ba could bind to the receptors, and after that it was internalized in the gut cells where it colocalized with granules in wild-type animals [29]. We hypothesized that if F19C6.4 and R03G8.6 proteases participate in the Cry5Ba and Cry21Aa degradation in vivo, this degradation process may attenuate the rate of toxins entering into the intestinal cells. Here, we fed rhodamine-labeled Cry5Ba or -labled Cry21Aa to L4-stage of F19C6.4(ok2392), R03G8.6(ok3143) mutants, wild-type, and to F19C6.4(ok2392) and R03G8.6(ok3143) rescued worms for 40 min. Image analysis showed that rhodamine labeled-Cry5Ba and -Cry21Aa were rapidly internalized into gut cells of F19C6.4 (ok2392) or R03G8.6(ok3143) mutant worms in contrast to that of wild-type and F19C6.4 and R03G8.6 rescued worms (Fig 4K-4N). Overall, these data confirmed that Cry5Ba treatment induce transcription of F19C6.4 and R03G8.6 proteases that is dependent in mnp-1 and also that these proteases participate in Cry5Ba toxin degradation in C. elegans, although, other proteins may also participate in this defense response.

MNP-1-mediated F19C6.4 and R03G8.6 expression by activating FOXO transcription factor DAF-16
Both F19C6.4 and R03G8.6 belong to metalloproteinases, and their mutant nematodes exhibit reduced ability to degrade Cry5Ba (Fig 4E-4F). It has been shown that F19C6.4 is a transcriptional target of the TGF-Dauer pathway in adults, while the transcription level of R03G8.6 is significantly lower in daf-16(mgDf50); daf-2(e1370) compared to daf-2(e1370), suggesting that F19C6.4 and R03G8.6 may be influenced by the DAF-2/DAF-16 signaling pathway [52,53]. The C. elegans DAF-2/DAF-16 signaling pathway has also been reported to participate in the defense of the nematode against external pathogens, with DAF-16 acting as a transcriptional regulator of many stress responses and antimicrobial genes [32]. We hypothesized that the transcription factor DAF-16 may be playing a role in MNP-1 mediated PFTs protection. To analyze this, we quantified the mRNA levels of mnp-1, F19C6.4 and R03G8.6 in wild type N2 worms and in mnp-1(ok2434) and daf-16(mu86) mutant worms fed with E. coli cells expressing or non-expressing Cry5Ba. Upregulation of F19C6.4 and R03G8.6 by Cry5Ba exposure was abolished in mnp-1(ok2434) and in daf-16(mu86) mutants (Fig 5A and 5B), indicating that F19C6.4 and R03G8.6 protease are regulated by both mnp-1 and DAF-16. In addition, under normal conditions, R03G8.6 expression was downregulated in daf-16(mu86) mutant, confirming that basal expression of R03G8.6 is dependent on daf-16 ( Fig 5B). These data support that the DAF-16 activity is required for the induction of these proteases. We also tested whether mutations in DAF-16 protein affected Cry5Ba degradation. The crude extracts from N2, mnp-1(ok2434) and daf-16(mu86) mutant worms were prepared to analyze the degradation of Cry5Ba. Fig 5C and   5D show that the crude extracts from mutant worms were significant less efficient to degrade Cry5Ba than the extracts from wild type N2 control worms, supporting that DAF-16 is involved in Cry5Ba degradation through the induction of F19C6.4 and R03G8.6 proteases.
The daf-16 gene codifies for a transcription factor that activates gene transcription when translocated into the nucleus. As described above, DAF-16 activity is required for the induction of the proteases. Therefore, we analyzed the effect of Cry5Ba exposure on the cellular localization of the DAF-16 using transgenic worms N2-TJ356(daf-16::gfp) that express a functional DAF-16::GFP fusion protein [54,55]. As expected, we found that exposure to Cry5Ba caused significant increase of DAF-16 inside the nucleus compared with the control without toxin treatment (Fig 5E and 5F). To determine the possibility that mnp-1 has a role in DAF-16 translocation into the nucleus after Cry5Ba treatment, we inactivated mnp-1 by RNAi in N2-TJ356(daf-16::gfp) transgenic worms to analyze DAF-16 translocation into the nucleus when treated with Cry5Ba. We found that silencing mnp-1 expression by RNAi significantly reduces the DAF-16::GFP localization into the nucleus after Cry5Ba treatment, compared to DAF-16:: GFP localization in the transgenic worms N2-TJ356(daf-16::gfp) after Cry5Ba treatment ( Fig  5E). Fig 5F shows the quantitative analysis of DAF-16::GFP localization in the different strains categorized as nuclear or cytosolic (named as intermediate and cytosolic) confirming that DAF-16::GFP localization into the nucleus increased after Cry5Ba treatment. In addition, we conducted western blot analyzes with cytosolic or nuclear samples purified from the worms to further confirm the localization of DAF-16 protein after Cry5Ba treatment. The results shows that Cry5Ba-treatment significantly induce DAF-16 protein localization into the nucleus sample, and this was reduced in mnp-1(ok2434) mutant worms (Fig 5G and 5H). Furthermore, to verify that DAF-16 nuclear localization activates F19C6.4 protease expression, we constructed the transgenic animals that express F19C6.4p::gfp in N2 and in daf-16(mu86) worms to compare the expression level of F19C6.4 in these worms under Cry5Ba treatment. The results showed that after 4 h treatment with Cry5Ba, the expression level of F19C6.4 is lower in daf-16 (mu86) than in N2 (Fig 5I and 5J). Overall, these data support that upon Cry5Ba exposure, mnp-1 could induce DAF-16 nuclear localization to activate F19C6.4 protease expression.

PLOS PATHOGENS
MNP-1 involved in C. elegans intestinal defensing PFTs proteins and participating in the pore-formation process [40,41]. The typical lepidopteran APNs proteins possesses six identified motifs and two conserved domains known as M1_APN_1-like and ERAP1_C [56]. In C. elegans, mnp-1 gene encodes a 781 amino acid APN protein that contains M1 peptidase domain and is required for embryonic muscle cell migration and neuronal cell migration [49,50]. Conserved domain and motif analyze of MNP-1, showed that this protein lacks the conserved motifs of Plutella xylostella APNs. Moreover, MNP-1 is presumed to be catalytically inactive because it lacks three of the four essential zinc-binding amino acids in its HENNH + E motif (S1 Fig) [50,56]. Based on the differences in conserved domains and motifs between MNP-1 and insect APNs, which serve as Cry protein receptors, it can be inferred that MNP-1 differs from insect APNs. Here, we showed that MNP-1 functions as a regulator of intestinal innate immunity effectors to protect C. elegans against PFTs. This is the first report that shows a that MNP-1 is associated with the host defense against PFT, supporting a new function for MNP-1 in C. elegans.
Cry5Ba exhibits high toxicity to parasitic and free-living C. elegans nematodes [36]. This toxin binds to the membrane receptor glycolipids and cadherin, and disrupt intestinal cells through forming pores on the cell membrane [31,39]. Meanwhile, worms initiate different pathways to resist Cry5Ba intoxication. Membrane damage activates p38/MAPK that either induces TTM-2-dependent PFTs defense or IRE-1 dependent endoplasmic reticulum UPR to resist Cry5Ba toxin action. In addition, Rab-5 dependent endocytosis and Rab-11 dependent exocytosis machinery assist worms to eliminate the membrane-bound Cry5Ba. For the Cry5Ba that enters the cell, autophagy could be activated by transcription factor HLH-30/TFEB, the Cry5Ba encapsulated into autophagosomes finally fuses with lysosomes for further degradation [25,35,38,57]. Here, we show that MNP-1 is involved in Cry5Ba defense in C. elegans. The Cry5Ba treatment induces the up regulation of mnp-1, which is located in the upstream of daf-16 pathway, and participates in inducing DAF-16 nuclear localization where it can activate the expression of F19C6.4 and R03G8.6 proteases to degrade Cry5Ba in the intestine (Fig 6). MNP-1 dependent Cry5Ba degradation in the intestine represents a novel strategy for C. elegans to resist PFTs. Recent studies have confirmed that treatment with Cry1Ac induces a rise in insect hormones levels within P. xylostella, subsequently triggering the activation of the MAPK cascade. The activated MAPK cascade regulates the phosphorylation level of the nuclear receptor fushi tarazu factor 1(FTZ-F1), resulting in the upregulation of non-receptor gene expression and downregulation of the receptor gene expression. Decreased expression of the receptor gene confers the resistance to Cry1Ac in P. xylostella [20,21]. These findings indicate that the organisms employ a shared defense mechanism to modulate gut gene expression upon detection of pore-forming toxins.
It is worth noting that the WormBase information indicates that the F19C6.4 and R03G8.6 genes are expressed in the nematode intestine, but the transgenic nematodes that we generated showed that F19C6.4 was expressed in nematode intestines, cephalic sheath cells, and coelomocytes, while R03G8.6 was primarily expressed in L4 nematode gonad (Fig 4C) [58,59]. We report here that Cry5Ba treatment can significantly increase the transcription level of F19C6.4 and R03G8.6 genes, Cry5Ba can also induce an increase in fluorescence in F19C6.4p::gfp but not in R03G8.6p::gfp. This could be due to the low expression of R03G8.6 in C. elegans' intestine, as Cry5Ba has not been shown to increase its expression in the intestine. We also show that the degree of Cry5Ba degradation in F19C6.4 and R03G8.6 mutants is higher than that of the mnp-1(ok2434) mutant worms (Fig 4E and 4F), suggesting that these two proteases are functionally redundant or that there may be additional proteases involved in the degradation of Cry5Ba. In addition, the Cry5Ba induced up-regulation of F19C6.4 and R03G8.6 genes was significantly inhibited in both daf-16 and mnp-1(ok2434) mutants (Fig 5A and 5B). However, in daf-16 mutant, Cry5Ba treatment significantly increased the transcription of both genes, indicating that these two genes were also regulated by other pathways besides DAF-16. Furthermore, it remains to be determined the mechanism by which Cry5Ba or Cry21Aa trigger the MPN-1 response. Pore formation is unlikely to be involved, since App6Aa, a different PFT, did not trigger this response. An interesting possibility is that Cry5Ba or Cry21Aa trigger MPN-1 response by a direct interaction of these toxins with MPN-1 since this is a membrane protein. This remains to be analyzed.
Interestingly, insect midgut proteases such as chymotrypsin and trypsin activate the threedomain Cry protoxin at the N and C terminal regions to generate a 55-65 kDa activated toxin. Proteolytic processing of Cry protoxin in the gut lumen, is an important step, involved not only in the activation of Cry toxin fragment, but could also determine the specificity against insects [41]. Recently, the analysis of the midgut transcriptome response of the rice leaf-folder, Cnaphalocrocis medinalis (Guenée) after Cry1C toxin ingestion revealed that a large number of serine proteases, APNs, and carboxypeptidases were upregulated, implying that these proteases may be involved in the response of C. medinalis to Cry1C toxin action [60]. Similarly, analysis of P. xylostella midgut transcriptome response after intoxication with Cry1Ac toxin, revealed that two APN encoding genes are significantly up-regulated upon Cry1Ac treatment [61].Our work revealed a novel defense response for C. elegans against certain PFT in which MNP-1 activates DAF-16 nuclear distribution to modulate intestinal protease expression that degrade Cry5Ba protein. In addition, we also found that the protein extracts obtained from plant parasitic nematodes Meloidogyne incognita and Ditylenchus destructor could also degrade Cry5Ba (S7 Fig). Since the transcription factor DAF-16 and complex intestinal digestive enzymes are found from C. elegans to human, we propose that this novel host mechanism to eliminate PFTs via intestinal protease degradation could be widely used as a defense strategy for different animals including mammals to defend themselves against enteric bacterial pathogens. Thus, the knowledge gained through this study may help to develop novel strategies in the future to treat different diseases induced by pathogenic microorganisms.

Bacterial strains, plasmids, and culture conditions
The bacterial strains and plasmids used in this study are listed in S6 Table. All Escherichia coli and Bt strains were grown on Luria-Bertani (LB) medium supplemented with the appropriate antibiotics at 37 or 28˚C, for E. coli or Bt, respectively.

RNA extraction and qRT-PCR assays
Total RNA was extracted from C. elegans using TRIzol reagent (Invitrogen, Carlsbad, California, USA). The cDNA was reverse transcribed with random primers using Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. The expression analyses of in C. elegans genes were performed using qPCR with the primers listed in S7 Table. The qPCR assays were conducted with Life Technologies ViiA 7 Real-Time PCR system (Life Technologies, California, USA) using the Power SYBR Green PCR Master Mix (Life Technologies, CA, USA) according to the manufacturers' instructions. The experiments were conducted in triplicate. Primer efficiency correction was done with 2 -44 Ct relative quantitation analyses using tba-1 gene as reference and normalization.

Microarray analysis
L1 stage synchronized N2 and mnp-1 (ok2434) worms were transferred to NGM plates spread with OP50 and incubated at 20˚C for about 44 h. The L4 stage worms were washed with M9 buffer for two or three times, and transferred to NGM plates spread with E. coli bacteria BL21 (pET28a) or BL21 (pET28a-cry5B) and incubated for additional 4 h at 20˚C. The worms were washed again with M9 buffer for two or three times, collected into 1.5 ml tubes, and stored at -70˚C after addition of 150 μl Trizol solution. Each treatment was performed three times. Microarray production, hybridization, and scanning assay were performed at Shanghai OE Biotech CO. Ltd. with Affymetrix C. elegans gene 1.0 ST. Default background subtraction and normalization settings were used. Spots were filtered based on foreground to background ratio; values less than 1.2 were flagged. Log (base 2) values were exported. Gene expression on BL21 (pET28a) or BL21 (pET28a-cry5B) were compared by t-test. Differences of 2 folds with p-values of 0.05 were considered significant. Sequencing reads were aligned to WormBase release WS235. Lists of upregulated genes used for comparisons were exported and further sanitized to remove dead genes and update Wormbase.ID to WormBase release WS270.

Gene Ontology analyses
Genes that showed significant differential regulation were categorized as either upregulated (UP) or downregulated (DOWN). To analyze these groups for enriched gene classes based on Gene Ontology (GO) terms, publicly available online resources centered around C. elegans such as https://biit.cs.ut.ee/gprofiler/gost and http://wormcat.com/ were utilized. The Representation Factor was calculated at http://nemates.org/MA/progs/overlap_stats.html.

RNA Interference (RNAi)
The RNAi assays were performed as previously described [44]. To induce dsRNA expression, E.coli strain HT115 transformed with the RNAi plasmids pL4440-mnp-1, and spread onto NG plates supplemented with 50 μg/ml ampicillin and 0.1mM isopropyl β-D-thiogalactopyranoside (IPTG)), and incubated at 25˚C overnight. E.coli HT115 with pL4440-mnp-1 spread on the NG plates that did not contain the IPTG were used as control. The L1 stage worms were cultured on the dsRNA induction or non-induction plates until L4 stage. Then the L4 worms were washed out and transformed to the 96-well plates containing E.coli HT115 RNAi strain and purified Cry5Ba. After a further 5 days of incubation at 25˚C, the survival rate of each well was recorded. Concentration of each toxin was analyzed in a triplicate set-up for each assay, and three independent assays were performed. The LC 50 values were determined by PROBIT statistical analysis [65].

Nematode mortality assays
The Cry5Ba plate bioassays and LC 50 assays were conducted as previously described [66]. Plate assays were used to determine Cry5Ba's toxicity to the worms. Plates containing BL21 E. coli cells expressing Cry5Ba or an empty vector were prepared. We add 20 synchronized L4 worms per plate and incubated them at 20˚C for 72 h. In addition, Cry5Ba liquid toxicity assay were used to quantify the LC 50 values of Cry5Ba against N2 and mutants. In brief, N2 and various mutant worms were exposed to purified Cry5B (the concentration range was 0.0-56 μg/ml) in S medium in 96-well plates with 20-30 worms per well to quantitatively score worms-survival after 5 days incubation at 25˚C. Concentration of each toxin was analyzed in a triplicate set-up for each assay, and three independent assays were performed. The LC 50 values were determined by PROBIT statistical analysis [65].

Nematode growth inhibition assays
L1 stage worms were fed for 48 h on control plates with OP50 food and different concentrations of Cry5Ba toxin. The relative health of each worm was evaluated qualitatively by comparing body size of N2 worms feed with OP50 food control.

Life-span assays of C. elegans
A total of 20 L4 stage synchronized N2 and mnp-1 (ok2434) worms were transferred to fresh NGM plates spread with 300 μl (OD 600 = 0.3) E. coli strain OP50 used as food and containing 0.1 mg/ml FUDR to prevent eggs from hatching. The alive/dead worms were scored every 12 h. The worms that were alive were transferred to new NGM plates until all the worms were dead. Three independent biological repeats were performed.

Fluorescence analysis
The ImageJ program (NIH) was used to analyze the GFP or rhodamine fluorescence intensity. Worms were selected by software and the fluorescence intensity (total pixel strength/area) of each nematode was determined at the same conditions. Nematodes expressing F19C6.4p::gfp (L4 stage) and R03G8.6p::gfp (L3 stage) were collected, these worms were placed either on control plates with E. coli that did not express Cry5B (pET28a vector control) or on plates prepared with E. coli expressing Cry5B (pET28a-Cry5B), with 30 nematodes per plate. These plates were incubated at 25˚C for 4 h, then worms were photographed on 2% agarose pads using fluorescence microscope and phase contrast microscope (Olympus BX51, Olympus, Tokyo, Japan). Three plates were tested per assay, and all experiments were repeated independently three or four times.

DAF-16 nuclear localization assay
N2 worms expressing TJ356 (daf-16::gfp) and its variant where mnp-1 was silenced by RNAi were transferred to ENG plates spread with BMB171 and BMB171 (pBMB0215) bacteria, for 2-3 h incubation at 30˚C. The worms were then washed once with M9 buffer, and transferred to a pad covered with 2% agarose for collecting the fluorescence micrographs using fluorescence microscope and phase contrast microscope (Olympus BX51, Olympus, Tokyo, Japan). Location of DAF-16::GFP in individual worms were classified as: nuclear (dot enrichment) referring to immune response to Cry5Ba, cytosolic (diffusion) referring to none immune response to Cry5Ba, and intermediate (half of dot enrichment). The proportion of worms in each category is a metric for comparing the extent of nuclear localization of DAF-16 between populations.

MNP-1 involved in C. elegans intestinal defensing PFTs
Nuclear and cytosolic distribution of DAF-16 protein L4 stage were prepared as previous described [63]. Nuclear and cytosolic fractions were separated as described by Singh et al [67]. Proteins were analyzed by SDS-PAGE, and DAF-16 proteins were detected by western blot assay. Lamin B1 and GAPDH were used as markers for the nuclear and cytosolic fractions, respectively. Antibodies against GAPDH, DAF-16, and Lamin B1 were obtained from Cell Signaling Technology (CST; USA) and HUABIO (China).

Preparation of crude extracts from C. elegans
Nematode were grown for 4-5 days in 250 ml batch of liquid culture, then they were collected in M9 buffer as previously described [63]. The extraction procedure of worms' crude extract sample was done as reported by Zhang et al [68], with some modifications. Approximately 500 mg worms were washed with Tris-HCl (pH 8.0) buffer for three times to replace M9 buffer. Worms were then grinded with TGrinder (OSE-Y30, TIANGEN BIOTECH (BEIJING) CO., LTD.). The supernatant obtained after centrifugation at 13,800 xg for 20 min was defined as nematode crude extracts. Protein concentration was determined as described below.

MNP-1 involved in C. elegans intestinal defensing PFTs
Quantification of crude extracts from C. elegans 10 mg Coomassie brilliant blue G-250 was dissolved in 5 ml ethyl alcohol and 10 ml orthophosphoric acid (85% (w/v)), and then diluted with ddH 2 O to a final volume of 100 ml. The final liquid solution was defined as G-250 dye. A total of 100 mg bovine serum albumin (BSA) was dissolved and diluted with ddH 2 O to a final volume of 100 ml at 1 mg/ml BSA standard solution. Six different volumes of BSA standard solution (0, 20, 40, 60, 80 and 100 μl) were added to six 20 ml colorimetric tubes, respectively. Then, a final volume of 1 ml, was completed in each tube with ddH 2 O and 3 ml G-250 dye was added and mixed. Absorbance at 595 nm was determined after 2 min incubation to draw the standard curve of BSA. A total of 60 μl nematode extracts was added into another colorimetric tube, and the final volume was completed with ddH 2 O up to 1 ml. The absorbance at 595 nm of these tubes was determined after addition of 3 ml G-250 dye and mixing. The protein concentration was calculated according to the standard curve of BSA.

Data analysis
All experiments were performed a minimum of three times. The data were analyzed using Statistical Package for the Social Sciences (SPSS version 13.0 Chicago, IL, USA) or GraphPad Prism 8. Graphical representations of the lethal concentration assays were generated using nonlinear regression analysis. Lifespan data was analyzed with Kaplan-Meier survival curves. LC 50 values of Cry5Ba, App6Aa, Cry21Aa against worms were determined by PROBIT analysis. The survival of each mutant worms was compared to that of the wild-type in GraphPad Prism 8 with the two-tailed t test. For the data where three or more values were compared, one way ANOVA or two way ANOVA were analyzed by using the Tukey post test. More details were shown in S8 Table. Statistical significance indicated as follows: ns indicates not significant, * indicates p < 0.05, ** indicates p < 0.01, *** indicates p < 0.001, **** indicates p < 0.0001. Data points represent the mean values of three independent replicates, error bars denote the SD. The p-value was determined by Unpaired t-test, ****p < 0.00001, ***p < 0.001, **p < 0.01, *p < 0.05 show significant differences and ns indicate no significant difference. (TIF) S1 Data. Excel spreadsheet containing, in separate sheets, the underlying numerical data for Figure panels 1A, 1B, 1C, 1E, 1F, 2A, 2B, 2C, 2D, 2E, 3E, 3G, 3I, 4A, 4B, 4D, 4E, 4J, 4L,  4N, 5A, 5B, 5D, 5F, 5H and 5J. (XLSX) S1