Calcineurin-Modulated Antimicrobial Peptide Expression Is Required for the Development of Helicoverpa armigera

Helicoverpa armigera is a universal pest around the world that has been extensively used as a model organism for agricultural pests. Calcineurin (CAN) is an important Ca2+-dependent phosphatase that is participated in various biological pathways. Here, we revealed that CAN inhibition significantly arrested H. armigera larval development by reducing larvae weight, prolonging development time and reducing pupate rates. Furthermore, CAN serves as an immune activator and regulates antimicrobial peptide (AMP; cecropin D, attacin, and gloverin) expression by binding with relish transcript factor in H. armigera. This study provides a potential target to control H. armigera by using synergistic agents for pesticides or plant-mediated RNA interference technology.


INTRODUCTION
Calcineurin (CAN) belongs to the protein phosphatase 2B family and acts as a Ca 2+ -dependent phosphatase. CAN is ubiquitously expressed in most mammalian tissues, and exerts different functions in different tissues (Furman and Norris, 2014). Studies have reported that CAN is predominantly expressed in the brain and plays important roles in various neuronal functions and disorders. This is due to CAN's vital functions in Ca 2+ homeostasis, synaptic plasticity, receptor signaling, and transcription regulation (Jackson et al., 2018). Furthermore, CAN is also found to be expressed in lymphocytes and macrophages, in which CAN has been confirmed to be took part in innate immunity (Bueno et al., 2002;Kang et al., 2007). In addition, In human endothelial cells, CAN functions as a mediator of anaphylaxis (Ballesteros-Martinez et al., 2017). These studies provide key evidence that CAN is a vital signaling molecule that performs a variety of functions in different biological pathways (Liu et al., 1991;Jackson et al., 2018).
In insects, studies on the functions of CAN have mostly been performed in the model organism Drosophila melanogaster. In Drosophila, CAN plays a crucial role in myofilament formation and troponin I isoform transition which affects the development of flight muscles (Gajewski et al., 2006). Interestingly, CAN in Drosophila has also been demonstrated to participate in circadian rhythms regulation (Kweon et al., 2018), axonal transport (Shaw et al., 2013), mitochondrial function (Chang and Min, 2005), olfactory associative learning and long-term memory (Chang et al., 2003), courtship behaviors (Ejima et al., 2004;Sakai and Aigaki, 2010), and female reproductive activities (Ejima et al., 2004;Horner et al., 2006;Takeo et al., 2010). Besides, Drosophila CAN has also been identified to play a role in immune response. When Drosophila were infected with gram-negative bacteria, CAN activity increased (Li and Dijkers, 2015). This promoted the product of relish, a key transcription factor of the immune deficiency (IMD) pathway in the innate immunity, which finally induces antimicrobial gene expression (Dijkers and O'Farrell, 2007;Li and Dijkers, 2015).
Besides Drosophila, relatively few studies have focused on CAN role in other insect species. In the Lepidoptera adult, Bombyx mori, CAN is localized in the cytoplasm of pheromoneproducing cells and regulates the production of sex pheromone (Yoshiga et al., 2002). In Helicoverpa armigera, CAN in adults mediates female attraction to males (Zhao et al., 2018). Furthermore, CAN has been confirmed to activate acetyl CoA carboxylase through dephosphorylation, and thus regulate sex pheromone biosynthesis and female mating acceptance in H. armigera (Du et al., 2017;Zhao et al., 2018). While these studies deepen our understanding of CAN function in adults, CAN function in larvae remains elusive and lacking.
Helicoverpa armigera is a universal pest around the world, and many measures have been taken to control this pest, including planting transgenic crops expressing a chemical pesticide, Bacillus thuringiensis (BT) (Downes et al., 2017;Wei et al., 2018). Although these measures work to some extent, they also present many problems, such as insect resistance. In order to find better methods to control H. armigera, it is critical to study the associated genes that involved in certain physiological process, such as genes involved in the immune process. Insect and mammalian immune systems share many similarities (Kaneko et al., 2004;Kaneko and Silverman, 2005). Insects depend solely upon the innate immune responses to survive. In the IMD pathway, the expression of antimicrobial genes is required to respond to gram-negative bacteria, thus making CAN an attractive target to control H. armigera CANs from different species share high amino acid identity, indicating they may share similar functions (Chen and Zhang, 2013;Zhao et al., 2018). Whether the IMD pathway exists in H. armigera larvae has never been addressed.
In this study, CAN from H. armigera (HaCAN) was found to be widely distributed in different developmental stages and different tissues, especially in fourth and fifth instars. Feeding of H. armigera larvae with FK506 significantly affected their development. FK506 treatments also significantly attenuated the expression of antimicrobial peptide (AMP) transcripts (cecropin D, attacin, and gloverin) caused by infection with gram-negative bacteria. For the first time, HaCAN was confirmed to be involved in larval development. HaCAN was also found to interact with relish. The present study reveals a potential target for synergistic agents such as pesticides or plant-mediated RNA interference (RNAi) technology, to control H. armigera.

Tissue Sampling
Larvae on the second day of every instar (from first to fifth) were collected with three replicates. The collected samples were used for subsequent RNA extraction to analyze the developmental expression profile of HaCAN (GenBank: KR185962.1).
In order to investigate distributions of HaCAN in the different tissues (including midgut, malpighian tubes, salivary glands, fat body, hemocyte, central nervous, and epidermis), the tissues were harvested from more than five larvae on the ice. This served as one biological replicate.
Expressions of endogenous AMPs (cecropin D, attacin, and gloverin) in midgut, fat body, and epidermis were analyzed by quantitative real-time PCR after FK506 treatment or Escherichia coli injection. The FK506 inhibitor was dissolved in DMSO. Either 50 µM FK506 or DMSO was fed to third instar larvae. 72 h after FK506 (or DMSO) treatments larvae were then injected with either 3 µL E. coli cells suspended in PBS (4 × 10 5 cell/µL) or 3 µL PBS. The larvae were incubated for either 1 h or 3 h, after which, the midgut, fat body and epidermis was harvested. All the collected samples were quickly frozen in liquid nitrogen and stored at −80 • C for subsequent RNA extraction. Ten larvae were used for one biological replicate.
All the above treatments were carried out with three biological replicates.

Quantitative Real-Time PCR Analysis of HaCAN Expression
Total RNA was extracted from each sample prepared above with TRIzol reagent (Invitrogen, Carlsbad, CA, United States) according to the manufacturer's instructions. All RNA samples were analyzed by UV spectrophotometry (NanoDrop; Thermo Scientific). A260/A280 ratio of all RNA samples were >1.8. All RNA samples were run on 1% agarose gel to ensure the quality of RNA (data not shown). After that, 1 µg of each DNase I-treated RNA sample was reversely transcribed into first strand cDNA using the PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Daliang, China). The above first-strand cDNA was then used as the template for qRT-PCR. The primers used for qRT-PCR analysis are listed in Table 1. Amplification efficiency of each primer pair was examined on an Applied Biosystems 7500 Fast Real-Time PCR system and calculated using the formula, PCR efficiency = 10 −1/slope − 1 (Bustin et al., 2009). All the primer pairs had an amplification efficiency   were performed individually in a 12 µL reaction system containing 6 µL 2 × SYBR R Premix Ex TaqTM II (Eurogentec, Fremont, CA, United States), 5 µM gene-specific forward primer and reverse primer (0.5 µL each), 1 µL template cDNA, and 4 µL nuclease-free water using an Applied Biosystems 7500 Fast Real-Time PCR system. Thermocycler conditions consisted of an initial denaturation for 3 min at 95 • C, followed by 35 cycles of denaturation at 95 • C for 15 s and annealing/extension at 58 • C for 30 s. A melting curve was generated after the termination of PCR cycles to ensure free of junk products. For each gene in each treatment, at least three biological replicates were tested with three technical replicates. The mRNA expression was quantified using the comparative cross threshold method (CT, the PCR cycle number that crosses the signal threshold) (Livak and Schmittgen, 2001). The CT result of the 18S gene was subtracted from the CT result of the target gene to obtain a CT. Normalized fold changes in target gene mRNA expression were expressed as 2 − CT , where CT is equal to CT treated sample -substracted from the CT control (Schmittgen and Livak, 2008).
The expression levels of cecropin D, attacin, and gloverin in H. armigera larvae were analyzed using qRT-PCR method as the above mentioned after treatments with FK506, E. coli, FK506 + E. coli. The primers of cecropin D, attacin, and gloverin are listed in the Table 1.

Effects on H. armigera Development After Feeding With FK506
Tacrolimus (FK506) is a special CAN inhibitor, which inhibits CAN phosphatase activity and affects the insect immunity (Liu et al., 1991(Liu et al., , 1992Dijkers and O'Farrell, 2007). Three concentrations of FK506 (50, 100, and 200 µM) were used to evaluate its effect on the development of H. armigera larvae (DMSO, FK506 solvent, was set as a control). Third instar larvae, 12 h before molting, were transferred onto the artificial diet that was mixed with either FK506 or DMSO. For each treatment, five replicates (30 third instar larvae per replicates × 5 replicates = 150 larvae total per concentration for each treatment) were employed.
In these experiments, the weights of the treated insects were recorded every other day. After 7 days, dead insects were counted to calculate the mortalities. The days, when the treated larvae began to pupate, were record as the larvae's development time. Successfully pupated insects were counted to calculate pupation rates. For the healthy pupae, the numbers of successfully emerged adults were counted to calculate the adult emergence rate.
Several colonies of NMY51 (Dualsystems Biotech) were picked from YPAD plates and incubated in 3 mL YPAD media at 30 • C, while shaking at 250 rpm for 8-12 h. Upon the culture reaching an OD600 of 0.15-0.3 (about 16-20 h), the culture was centrifuged at 700 × g for 5 min, the supernatant was discarded and pellet was resuspended in 100 mL YPAD. The bacterium solution was centrifuged at 700 × g for 5 min and washed again using 30 mL ddH 2 O. The collected bacteria were then resuspended in 600 µL 1.1 × TE/LiAc solution [1.1 × TE/LiAc (10 mL): 1.1 mL 1 M LiAc, 1.1 mL 10 × TE Buffer, 8.8 mL ddH 2 O].

Statistical Analysis
Significant differences in the relative expression levels of HaCAN in different instars and different tissues were compared using Tukey's test with P < 0.05 [analysis of variance (ANOVA) and DPS7.05]. Significant differences in the relative expression levels of cecropin D, attacin, and gloverin in different treatments were compared using Tukey's test with P < 0.05 (ANOVA and DPS7.05).

Developmental Expression Pattern of HaCAN in Larvae
To investigate the developmental expression pattern of HaCAN in different instars (first-, second-, third-, fourth-, and fifth-instar larvae), qRT-PCR was employed and results revealed that HaCAN was ubiquitously expressed in all of the examined stages. The expression level of HaCAN reached its peak at the first-instar larvae, followed by fifth-instar larvae ( Figure 1A).
The spatial distributions of HaCAN were investigated in different tissues (midgut, malpighian tubes, salivary glands, fat body, hemocytes, central nervous system, and epidermis) during different developmental stages (the first and second day of fourth instar larvae, molting day, and the first and second day of fifth-instar larvae). The distribution patterns of HaCAN in different tissues varied (Figures 1B-F), which generally exhibited higher expression levels in the epidermis during the first and second day in fourth instar larvae, the molting day and the first day in fifth instar larvae. Interestingly, HaCAN transcriptional level showed a sudden increase on the second day in fifth instar larvae (Figures 1B-F).

Effect of CAN Inhibitor on H. armigera Development
Four days of FK506 treatment caused significant decreases in body weight compared with controls (DMSO treatment). However, no significant differences were observed between the different FK506 concentrations used (Figure 2A).
Although seven days of FK506 treatments did not lead to significant changes in larvae mortalities, FK506 treatments significantly arrested the development of the larvae (Figures 2B,C). Pupation was delayed by 1-2 days and the pupation rates were significantly decreased (35.00, 32.47, and 17.36% after treatments with 50, 100, and 200 µM FK506, respectively, versus 93.75% after DMSO treatment) (Figures 2C,D). However, FK506 treatments exhibited no effect on the emergence rate of pupae ( Figure 2E).

Protein-Protein Interactions Between HaCAN and Harelish
To test yeast two-hybrid system, pDHB1-LargeT and pDSL-p53 were transformed as a positive control and were able to see growth on the selection plates, which showed that the yeast two-hybrid system works well (Figure 3). The transformation of pDHB1-N-relish and pOst1-NubI grew well on SD-Leucine-Tryptophan (DDO; Clontech) and SD-Leucine-Tryptophan-Histidine-Adenine (QDO; Clontech) plates, respectively. Moreover, the transformation of pBT3-Nrelish & pOst1-NubI turned blue in the β-galactosidase assay on SD/-LT/X, SD/-LTH/X, SD/-LTHA/X, SD/-LTHA/X/3mM 3-AT and SD/-LTHA/X/10 mM 3-AT plates, respectively (Figure 3). These experiments demonstrated that pBT3-N-relish plasmids worked well. The vector pPR3-N expresses the N-terminal half of ubiquitin (NubG) and transformation of pBT3-N-relish and pPR3-N displayed less growth on normal selection plate while there were no clonies on QDO plates. There were also no blue colonies present in the β-galactosidase assay on SD/-LT/X, SD/-LTH/X, SD/-LTHA/X, SD/-LTHA/X/10 mM 3-AT plates (Figures 1, 2). These results showed that the clones of pDHB1-N-relish and pPR3-N on QDO plates were false positives, and that the split-ubiquitin membrane yeast two-hybrid system can be used to identify interactions between HaCAN and Harelish. The transformation of pDHB1-N-relish & pPR3-N-CAN grew well on both DDO plates and QDO plates. Moreover, the colonies resulting from this transformation turned blue in a β-galactosidase assay on SD/-LT/X, SD/-LTH/X, SD/-LTHA/X, SD/-LTHA/X/3 mM 3-AT and SD/-LTHA/X/10 mM 3-AT plates, indicating that HaCAN and Harelish interact directly (Figure 3).

HaCAN Regulates AMPs Expression
To investigate the regulation of AMPs expression by HaCAN, the CAN inhibitor FK506 was used. The results demonstrated that HaCAN inhibition with different FK506 resulted in a significant decrease in the expression levels of several AMPs, including cecropin D, attacin, and gloverin in the midgut, fat body and epidermis of larvae (Figure 4).
In order to further study the HaCAN role on AMP expressions, gram-negative E. coli was injected into larvae fed with 50 µM FK506 for two days prior. A 1 h treatment with E. coli treatment caused significant increases in AMPs' expressions in the fat body and midgut tissues. Interestingly, the application of FK506 significantly counteracted E. coliinduced AMP expression in the fat body and midgut tissues Frontiers in Physiology | www.frontiersin.org ( Figure 5). While 1 h-E. coli treatment can significantly induced cecropin D expression in the epidermis, FK506 treatment did not significantly hinder this expression (Figure 5). Interestingly, 1 h treatments both E. coli and FK506 had no effect on gloverin expression in the epidermis (Figure 5). Unlike cecropin D and gloverin, the expression pattern of attacin in the epidermis was similar to that the expression pattern seen in the midgut and fat body (Figure 5).
Longer E. coli (for 3 h) treatments resulted in significant increases in the expression levels of attacin and gloverin in the midgut, fat body and epidermis. The E. coli induced increase of these transcripts (attacin and gloverin) was inhibited by FK506 (Figure 6). Although cecropin D expression increases with E. coli stimulation, it was not suppressed by FK506 in the midgut, fat body or epidermis (Figure 6).

DISCUSSION
Many important life processes in the cell depends on protein phosphorylation/dephosphorylation, which plays an important role in many signal transduction pathways (Patarca, 1996;Lapied et al., 2009). HaCAN is a critical dephosphorylase and its ubiquitous expression in different instars and tissues indicates that HaCAN participates in many biological pathways (Figure 2). The importance of CAN has been reported in other species including mammals and Drosophila (Dijkers and O'Farrell, 2007;Jackson et al., 2018;Kweon et al., 2018). CAN shares a high amino acid identity cross different species, which indicates that CAN may have similar functions in these different species (Chen and Zhang, 2013;Zhao et al., 2018).
FK506 is a specific inhibitor of CAN that significantly inhibit CAN dephosphatase activity (Aramburu et al., 2000). In the present study, FK506 was employed to investigate the role of CAN on H. armigera development. FK506 H. armigera insect development by reducing the weight of larvae, prolonging development time, and reducing pupate rates (Figure 2). These results indicate that HaCAN plays an important role in the development of H. armigera larvae. Although treatments with different FK506 concentrations for 7 days exhibited no significant differences in mortality, FK506 treatment clearly affected insect survival. The present study also revealed that HaCAN can regulate AMP's expression, however, the decrease in HaCANregulated AMPs expression alone did not explain the large changes in developmental delay and pupate rate reduction. Thus, the functions of HaCAN needed to be further investigated.  In order to survive, insects must defend themselves against various pathogens. Drosophila employ two pathways (Toll and IMD) against pathogens (Kaneko and Silverman, 2005). Gramnegative bacteria infection promotes the activation of the relish transcription factor through the IMD pathway in Drosophila (Dijkers and O'Farrell, 2007). The expressions of the AMPs attacin, cecropin, and gloverin depend on relish expression in insects (Meng et al., 1999;Xu et al., 2012;Yang et al., 2015). Our results demonstrate that E. coli can induce the expression of these AMPs via the IMD pathway. However, E. coli induced AMP expression can be neutralized by FK506, which indicating that CAN acts as an up-stream regulator of AMP expression. AMP expression depends on the relish transcription factor and our studies confirmed that HaCAN interacts with Harelish (Figure 3). These results provide evidence for the first time that HaCAN can activate relish through the IMD pathway to regulate the expression of cecropin D, attacin, and gloverin under gram-negative bacteria infection. Similar results were also found in Drosophila, in which CAN acted on relish during infection (Li and Dijkers, 2015).
Drosophila utilizes two distinct pathways to express AMP genes (Kaneko and Silverman, 2005). Infection with gram-negative bacteria promotes the activation of the relish transcription factor through the IMD pathway in Drosophila (Dijkers and O'Farrell, 2007). The Toll pathway provides defense against gram-positive bacteria or fungal infection. Activation of the Toll pathway leads to expression of Dorsal and Dif transcription factors, which in turn up-regulate expression of AMP transcripts (Lemaitre et al., 1996(Lemaitre et al., , 1997. Our results show that the expression of cecropin D is regulated in part by the IMD pathway, which suggest that HaCAN may participate in the Toll pathway or another immune mechanism. These pathways contribute to the expression of different AMPs by activating different downstream transcription factors (Zhang et al., 2012;Zeng et al., 2019). It may be interesting to find other transcription factors that interact with HaCAN, and how these transcription factors function within immune response in H. armigera. The role of HaCAN in the immune response must also be further investigated in future studies.
Inhibition of HaCAN activity by FK506 significantly affects the development of cotton bollworms, indicating that FK506 may be used as asynergistic agent for pesticides. Moreover, plantmediated RNAi shows great potential in crop protection. This technology relies on plants to stably express double-stranded RNAs that target essential genes in pest insects (Mamta and Rajam, 2017;Zhang et al., 2017). RNAi-mediated silencing of insect genes took part in various physiological processes were found to be detrimental to the growth, development and survival of these pest insects (Xiong et al., 2013;Zhang J. et al., 2015;Chikate et al., 2016). HaCAN inhibition significantly affects the development of insects (approximately 70% mortality rate with 50 µM FK506 in H. armigera) (Figure 1). It has also been reported that HaCAN can regulate sex pheromone production and change mating behavior (Zhao et al., 2018). An interesting next step will be to study RNAi of HaCAN, as a technology to control H. armigera.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the manuscript/supplementary files.