Molecular Identification of Two DNA Methyltransferase Genes and Their Functional Characterization in the Anti-Bacterial Immunity of Antheraea pernyi

Under different physiological conditions, such as microbial infection, epigenetic mechanisms regulate genes at the transcription level in living organisms. DNA methylation is a type of epigenetic mechanism in which DNA methyltransferases modify the expression of target genes. Here, we identified a full-length sequence of DNMT-1 and DNMT-2 from the Chinese oak silkworm, A. pernyi, which was highly similar to the homologous sequences of Bombyx mori. ApDNMT-1 and ApDNMT-2 have unique domain architectures of insect DNMTs, highlighting their conserved functions in A. pernyi. ApDNMT-1 and ApDNMT-2 were found to be widely expressed in various tissues, with the highest levels of expression in hemocytes, the ovary, testis, and fat bodies. To understand the biological role of these genes in microbial resistance, we challenged the fifth instar larvae of A. pernyi by administrating Gram-positive and Gram-negative bacteria and fungi. The results revealed that transcript levels of ApDNMT-1 and ApDNMT-2 were increased compared to the control group. The inhibition of these genes by a DNMTs inhibitor [5-azacytidine (5-AZA)] significantly reduced bacterial replication and larvae mortality. In addition, 5-AZA treatment modified the expression patterns of antimicrobial peptides (AMPs) in the A. pernyi larvae. Our results suggest that ApDNMT-1 and ApDNMT-2 seem to have a crucial role in innate immunity, mediating antimicrobial peptide responses against bacterial infection in A. pernyi.


INTRODUCTION
The Chinese oak silkworm, A. pernyi, is broadly cultured for silk production in Asian countries, such as China, Korea, and India (1). This wild silk moth is also used as a model insect to investigate immunological response, development, metabolism regulation, and breeding technique. A. pernyi, like other insects, has strong and efficient innate immunity, including cellular and humoral components. Encapsulation, nodule formation, and phagocytosis are examples of cellular immune responses (2,3). The humoral immune response includes the production of AMPs following a microbial infection (2,4). In A. pernyi and other insects, complex sets of genes involved in the coordination of the innate immune system are related to the immune pathways such as Toll, Imd, and NF-kB pathways (5). After pathogen detection, the activated cell surface receptors induce the downstream signaling cascade inside cells, stimulating the targeted immune response. Toll and IMD pathways produce AMPs in response to bacterial and fungal infection (5,6).
DNA methylation is one of the critical epigenetic modifications that regulate gene transcription in eukaryotes and control various biological processes. DNA methyltransferases are responsible for DNA methylation (7,8). Compared to mammals, studies on DNMTs and DNA methylation in insects remain relatively poor. However, some progress has been made in insects in recent years due to the development of innovative DNA methylation research methods. DNA methylation has been found in the genomes of insects such as silkworm, Nasonia Vitripennis, Tribolium castaneum, and Drosophila melanogaster, according to whole-genome bisulfite sequencing studies (9)(10)(11)(12). However, the maintenance and regulatory mechanisms of DNA methylation in insects may vary significantly from those in other organisms (13)(14)(15). Insects have been shown to have different combinations of DNMT gene sets. The Apis mellifera genome contains DNMT-1, DNMT-2, and DNMT-3 genes (16). However, in silkworms and Tribolium castaneum, only two DNMT genes (DNMT-1 and DNMT-2) have been identified (16). In the living organisms lacking DNMT-3, including silkworms and beetles, DNMT-1 appears to be a de novo and maintenance DNMT. DNA methylation regulates various physiological processes in insects, such as immunity, memory, aging evolution, and others (17)(18)(19)(20)(21).
Insect immunological responses have not been extensively explored in terms of DNA methylation's biological roles. It has been shown that Aedes aegypti genome wide-patterns of DNA methylation are disrupted by the infection by Wolbachia (22). Cytoplasmic polyhedrosis virus infection caused the differential expression of genes and affected the DNA methylation status in the fat bodies and midguts of silkworms (23). Subsequently, Wu et al. (15) observed the differential expression of DNMT genes in the Galleria mellonella larvae infected with three different strains of Metarhizium anisopliae, suggesting that DNA methylation in insects is likely involved in the immune response against microbial pathogens (15). In the present study, we identified the DNA methylation tool kit in the Chinese oak silkworm A. pernyi. Then we analyzed their spatial and temporal expression patterns in different tissues of A. pernyi larvae. Next, we treated A. pernyi larvae with 5-AZA, a DNA methylation inhibitor, and analyzed the effects of DNA methylation inhibition on the insect's immune responses.

Culture of A. pernyi Larvae
The model insect, A. pernyi larvae, was maintained at room temperature and fed on fresh oak leaves as described in previous reports (2,4). Healthy fifth instar larvae were used in all of the following experiments.

Gene Identification and Bioinformatics Analysis
To identify ApDNMT-1 and ApDNMT-2 sequences, we considered the genomic library of A. pernyi constructed in our laboratory. The retrieved sequence was further analyzed by the NCBI blastn tool for confirmation. Then the fragments of these genes were amplified by polymerase chain reaction using specific primers ( Table 1). The ORF sequences of ApDNMT-1 and ApDNMT-2 were mapped into their respective sequences. The domain architecture was determined by using the blastp program. Multiple sequence alignments were performed with representative DNMT-1 and DNMT-2 sequences from other insect species retrieved from the GenBank database using Clustal software. The three-dimensional protein structure was analyzed by the SWISS-MODEL (http://Swiss-model.expasy. org). Phylogenetic analysis was carried out using Mega 6 software with 1000 bootstraps using the neighbor-joining method based on amino acid sequences.

Tissue Distribution Pattern Analysis of DNMT Genes
To analyze tissue distribution patterns of DNMT genes, hemocytes, fat body, midgut, silk gland, testis, ovary, Malpighian tubules, and integument tissues were collected from the healthy A. pernyi larvae. Total RNA from each sample was extracted using TRIzol reagent (Invitrogen) according to the suppliers' instructions, and 2 mg of total RNA was used to prepare cDNA. Oligonucleotide primers specific for the ApDNMT-1 and ApDNMT-2 sequences and the 18S rRNA sequence for internal control were designed using the Primer 3.0 online primer designing tool (http://bioinfo.ut.ee/primer3-0.4.0/ primer3/) ( Table 1). We performed PCR and analyzed amplicons on the agarose gel before carrying out qRT-PCR to ensure the integrity and stability of 18S rRNA in the cDNA of all samples used in this study. The quantitative real-time polymerase chain reaction assay was performed with specific primers and synthesized cDNAs, and each reaction mixture contained 20 mL with 10 mL of SsoFast EvaGreen SuperMix (Bio-Rad, Hercules, California, USA), 1 mL of forward primer (200 mM), 1 mL of reverse primer (200 mM), 1 mL of diluted cDNA, and 7 mL of sterile water. The thermal cycling conditions were 95°C for 30 s, followed by 40 cycles of 95°C for 5 s and 60°C for 34 s. Amplification was monitored on the iCycleriQ TM RT-PCR Detection System (Bio-Rad, Hercules, California, USA). The specificity of the SYBR-Green PCR signal was further confirmed by melting curve analysis. The experiments were repeated three times as independent biological replicates. The mRNA expression was quantified using the 2 -DDC Tmethod (25).

Microbial Injection and Expression Analysis
For obtaining induction profiles of DNMT genes, fifth instar larvae of A. pernyi on the third day were divided into four groups with 54 larvae in each group. The A. pernyi larvae were injected with 5 mL each of E. coli (DH5a, 1 × 10 6 cells), B. cereus (1 × 10 6 cells), or B. bassiana (1 × 10 6 spores). The larvae injected with PBS were used as a control group. These microbial organisms were injected into the larval abdominal area using microliter syringes (Gaoge, Beijing, China), and injection sites were sealed with vaseline immediately after injection. The larvae were dissected, and immune tissues, including the fat bodies and hemolymph, were sampled at different time intervals (1.5, 3, 6, 12, 24, and 48 h) after injection. Three larvae were collected as one sample, and the biological sampling protocol was repeated three times. The transcript expression patterns of ApDNMT-1 and ApDNMT-2 genes were recorded using a qRT-PCR assay as described in the above section.

5-Aza-2-Deoxycytidine (5-AZA) Administration
To determine the effect of 5-aza-2-deoxycytidine on the DNMT genes and, subsequently, on the immune responses against bacterial pathogens. 5-AZA can prevent de novo and maintenance of DNA methylation by binding covalently to DNMT genes (26). We purchased 5-AZA and diluted it in distilled water with 40 mM/1 mL. The fifth instar larvae of A. pernyi were injected with 1 mL 5-AZA into the abdominal area of the larvae, and the second injection was performed at 24 h intervals. Then we analyzed, the expression patterns of both the ApDNMT-1 and ApDNMT-2 genes using qRT-PCR. Bacterial pathogens (B. cereus or E. coli) were injected 24 h after the second 5-AZA injection. The larvae were used as control groups without any treatment, and the larvae were injected with distilled water.

Bacterial Load Estimation and Survival Analysis
To estimate the bacterial load, we first analyzed the bacterial clearance ability of larvae injected with 5-AZA and bacterial pathogens. After 6 h of bacterial injection, A. pernyi larvae legs were punctured to collect hemolymph from the 5-AZA treated group and control group (distilled water treated and untreated) and then immediately diluted. Fifty mL of each dilution was plated onto agar plates, incubated overnight at 37°C, and the colony-forming units were counted. Furthermore, total genomic DNA was extracted from the hemolymph samples of treated (bacterial-injected and 5-AZA treated) and control A. pernyi larvae. The concentration of the extracted DNA was determined by a spectrophotometer, and 500 ng of total genomic DNA was used for each qRT-PCR reaction, and the analysis was performed as described in section 2.3. Specific primers for bacterial 16s rRNA were used in qRT-PCR, and the A. pernyi 18s rRNA gene was used as an internal control for normalization (24,27). To better understand the insect-pathogen interaction, the mortality of A. pernyi was also analyzed using Graphpad Prism 6.0. Kaplan-Meier survival analysis.

Analysis of AMPs and Their Associated Pathways
A. pernyi larvae were treated with 5-AZA-bacteria (B. cereus or E. coli) or only bacteria or untreated larvae to determine the response of AMPs and their associated pathways. The larvae were dissected, and the fat bodies were used to isolate total RNA, which was then reverse transcribed to prepare cDNA. A quantitative RT-PCR assay was carried out (described in section 2.3) using AMPs (gloverin-like, cecropin, attacin, and lebocin) and their associated pathway genes by specific primers ( Table 1).

Statistical Analysis
In this study, all experiments were executed in triplicate and the obtained data represented the means ± S.E. The one-way analysis of variance (ANOVA) and Tukey's multiple range tests were used to evaluate the difference between groups.

Bioinformatic Analysis of DNMTs of A. pernyi
The cDNA sequences of ApDNMT-1 and ApDNMT-2 were identified by using a genomic library of the Chinese oak silkworm A. pernyi, which was constructed in our laboratory. The full-length sequence of ApDNMT-1 comprised a total length of 5508 base pairs. The open reading frame (ORF) is 4344 bp and contains a 209 bp 5′ untranslated region (5′ UTR) and a 955 bp 3′ UTR (Supplementary Figure S1). The deduced ApDNMT-1 protein is comprisedof1447aminoacidswithanapproximatemolecularweight of 163.182 kDa and isoelectric point of 6.53, respectively. The NCBI Conserved Domain Database was used to determine conserved domains of ApDNMT-1, which showed that this gene has four conserved domains: a Zinc finger domain, a Bromo adjacent homology domain, a DNA-cytosine methyltransferase domain, and a cysteine-rich ADD domain (Supplementary Figure S1;  Figures 1, 3). Further, it was found that ApDNMT-1 has a comparable structure and domain architecture similar to that of silkworm in its tertiary structure ( Figure 3A). Multiple sequence alignment analysis revealed that ApDNMT-1 has the highest similarity to B. mori (77.96%), followed by Athalia rosae (52.12%), Ceratina calcarata (49.74%), and so on. The full-length sequence of ApDNMT-2 contains an 1833 bp nucleotide sequence. The ORF of ApDNMT-2 was 888bp in length, with a 5′ UTR of 205 bp and a 3′ UTRof740bp.ThededucedApDNMT-2proteincontains295amino acid residues. The predicted molecular weight and isoelectric point are 33.89 kDa and 6.01, respectively. In a domain analysis of ApDNMT-2, weobserved that this protein consisted of only one domain (the typical DNMT-2 cyt_C5_DNA methylase superfamily domain of DNMT-2 proteins),anditstertiarystructurewashighlycomparabletothatofthe B. mori DNMT-2 protein (Supplementary Figure S2 and Figures 2,  3B). Additionally, multiple sequence alignment analysis revealed that ApDNMT-2 shares the maximum similarity with B. mori (66.78%), followed by Tribolium castaneum (48.41%), Drosophila melanogaster (43.73%), and other species. A phylogenetic analysis of amino acid sequencesfromdiverseinvertebrateandvertebratespecieswascarried out in order to better understand the evolutionary relationships of ApDNMT-1andApDNMT-2.Thephylogenetictreewasconstructed using the neighbor-joining method, which was performed using the Mega 6 software. The results revealed that ApDNMT-1 and ApDNMT-2 proteins are associated with the DNMT-1 and DNMT-2 proteins of animals, respectively ( Figure 4).

Spatial Expression Patterns of ApDNMT-1 and ApDNMT-2
To understand the biological role of ApDNMT-1 and ApDNMAT-2, we measured their transcription levels in different tissues (hemocyte, fat body, midgut, Malpighian tubules, integument, head, testis, ovary, and silk gland) of A. pernyi by performing qRT-PCR analysis. Using a tissue-specific mRNA expression analysis, it was found that ApDNMT-1 was transcribed in all of the tested tissues, with mRNA expression levels being notably high in the ovary, hemocytes, and fat bodies ( Figure 5A). The ApDNMT-2 gene was observed in all of the tissues examined, although its mRNA expression levels varied in each tissue. ApDNMT-2 expression levels were shown to be higher in the ovary, hemocytes, and fat bodies, similar to ApDNMT-1 expression levels ( Figure 5B). These findings show that these proteins have a broader range of biological roles in insects.

A. pernyi DNMTs Induction Profile Analysis
We examined the transcription levels of the ApDNMT-1 and ApDNMT-2 genes in hemocytes and fat bodies from A. pernyi at 1.5, 3, 6, 12, 24, and 48 h after injection with B. bassiana, E. coli, and B. cereu to investigate the influence of fungal and bacterial infections on ApDNMT-1 and ApDNMT-2 genes. The qRT-PCR analysis revealed that both fungal and bacterial pathogens induced transcription patterns of the ApDNMT-1 and ApDNMT-2 genes in hemocytes and fat bodies that were similar to one another. During the early hours after infection, the expression levels of ApDNMT-1 and ApDNMT-2 genes were slightly altered. ApDNMT-1 gene expression was shown to be notably upregulated in the hemocytes and fat bodies at critical timepoints of infection (e.g., 6, 12, 24 h) ( Figure 6). After the  fungal and bacterial infections, the transcription patterns of ApDNMT-2 were found to be approximately similar to those of ApDNMT-1 (Figure 7). Following these timepoints, however, the transcription levels of these genes start to drop.

Inhibition of DNMTs by 5-AZA Affected the Bacterial Replication and Mortality of A. pernyi
Because of the remarkable variation in ApDNMT-1 and ApDNMT-2 gene transcription that was observed in A. pernyi larvae after fungal and bacterial infections, we next planned to investigate the importance of ApDNMT-1 and ApDNMT-2 involvement in innate immune responses. For this reason, we inhibited the DNMTs in the fifth instar larvae by injecting 5-AZA into the larvae, followed by B. cereus and E. coli injections. Before bacterial injection, we confirmed the suppression of ApDNMT-1 and ApDNMT-2 following the 5-AZA administration ( Figure 8A). Total genomic DNA was extracted from the hemolymph samples of treated (bacterial-injected and 5-AZA treated) and control larvae. To perform qRT-PCR, we used 500 ng total genomic DNA for each qRT-PCR reaction, and the DNA concentrations were measured using a spectrophotometer. In qRT-PCR, specific primers for the 16s rRNA of the bacteria were used, and the insect 18s rRNA gene was used as an internal control for normalization. For each experiment, three biological replicates were replicated three times. Our results showed that 5-AZA treatment of larvae reduced replication of the pathogenic bacteria when compared to control larvae (Figures 8A, B). We also analyzed the mortality rate of bacteria-infected larvae and found that injection of 5-AZA attenuated larval mortality ( Figure 8C).

Antimicrobial Activity of the 5-AZA Treated Plasma
In order to determine whether the presence of 5-AZA, a DNA methylation inhibitor, modulates the antimicrobial activity, we carried out a bacterial clearance assay using plasma collected from larvae that had been treated with 5-AZA and bacteria. The results revealed that plasma from A. pernyi larvae treated with 5-AZA had significantly greater antibacterial activity against gram-positive and gram-negative bacteria than in control groups. For example, the survival rates of bacteria (B. cereus and E. coli) in the 5-AZA treated larvae were much lower than in the distilled water treated larvae and the untreated control larvae ( Figure 9).

Effects of DNMTs Inhibition on AMPs and Their Associated Pathways
To evaluate the effects of 5-AZA on AMP gene expression, we quantified changes in the transcript levels of various AMPs in larvae that had been treated with 5-AZA and then challenged with bacteria. We injected A. pernyi larvae in the control group with distilled water or with bacteria only. We observed that among the four AMPs analyzed, the expression of attacin, cecropin with Gram-positive and gloverin-like, attacin, and lebocin with gram-positive bacteria was strongly increased in 5-AZA-injected larvae after challenge with B. cereus and E. coli as compared to the control group ( Figure 10). To understand further how 5-AZA and bacterial treatments affect these AMPassociated pathways, we analyzed the Toll and IMD pathwayassociated genes (Toll-like receptor, dorsal, and relish), and the results revealed that the expression of these genes was altered in the 5-AZA and bacterial treated larvae compared to control groups ( Figure 11).

DISCUSSIONS
In the present study, we identified two members of the DNA methyltransferase family, including DNMT-1 and DNMT-2, from the Chinese oak silkworm A. pernyi. These genes share high similarities with the DNMT-1 and DNMT-2 of other lepidopterans, particularly with B. mori. Previous studies have also reported the presence of only DNMT-1 and DNMT-2 genes in lepidopteran species (21,28). These studies support the hypothesis of the loss of the DNMT-3 gene from lepidopteran species (29). Conserved domain analysis showed that ApDNMT-1 and ApDNMT-2 proteins have the typical domain composition of DNMT families. ApDNMT-1 has four conserved domains, while ApDNMT-2 has only 1 domain, suggesting that these proteins have different functions. The phylogenetic analysis of DNMTs also exhibited high similarities of ApDNMT-1 and ApDNMT-2 proteins to the homologous proteins in other lepidopterans, particularly B. mori. In the phylogenetic tree, we observed that DNMT-1 and DNMT-2 clades clustered separately, further strengthening the idea that these proteins have different biological roles in insects. Furthermore, we analyzed the tissue distribution of these genes in A. pernyi larvae. Our tissue mRNA expression analysis showed that ApDNMT-1 and ApDNMT-2 expression was detectable in all A. pernyi larval tissues examined, but there were notable mRNA expression levels in the hemocytes, testis, and fat bodies. We presume that these variations in ApDNMT-1 and ApDNMT-2 transcription levels between tissues might have to do with different functions in insects. We suggest this possibility since many previous reports on insects have shown that DNA methylation regulates various physiological processes by modulating a wide range of gene expressions (21,30). Kausar et al. (21) reviewed that in the insect's DNA methylation levels of embryos, larvae, and adults, genes associated with innate immunity, development, growth, etc., respond to the changing levels of DNA methylation. In a previous study, the authors suggested constitutive expression of DNMTs in the insect tissues through their life cycle. This observation might be elucidated by the fact that insects undergo various physical and physiological changes during the life cycle, requiring a precise mechanism for the surveillance of these processes (21,30). It is likely that we are seeing a similar factor affecting the expression of DNMTs. Because of the reasons above, it is not surprising that DNMTs are expressed constitutively in various tissues of A. pernyi larvae. Moreover, mRNA expression analysis revealed that both ApDNMT-1 and ApDNMT-2 gene transcription levels were increased in the Chinese oak silkworm, A. pernyi, after microbial injections, particularly at critical timepoints of infection. The upregulation of DNMT genes transcription seems to be a general response of insects against microbial pathogens. Baradaran and his coworkers (30) reported the induction of DNMT-1 and DNMT-2 genes following bacterial infection. Besides, it has also been shown that E. coli infection induced DNMT-1 and DNMT-2 transcription levels and caused differential DNA methylation status of the genome in Manduca sexta (31) and Galleria mellonella (32). In addition, in Aedes aegypti and Drosophila melanogaster, increased DNMT-2 gene expression and hypermethylation have been observed following the Wolbachia infection (33)(34)(35). Although DNMT-3, a de novo DNA methyltransferase, is absent in lepidopteran insects, including A. pernyi, recent research has shown that in insects, particularly lepidopteran insects, DNMT-1 is responsible for DNA methylation rather than the DNMT-3 gene (28). In addition, this gene has also been reported to be involved in the maintenance of DNA methylation and also in de novo DNA methylation (36). In contrast, the presence of DNMT-2 in dipteran insects (e.g., A. aegypti and D. melanogaster) is responsible for changes in genome methylation (35,37). The molecular mechanism by which pathogens modify DNA methylation is not clear in insects; however, this mechanism has been well explored in vertebrates (21). A previous study suggested that modulation of DNA methylation is caused by natural killer cell activity via IFN (interferon)-gamma (38). The superfamily of tripartite motif-containing proteins induces innate immunity and pathogen defense against a variety of pathogen infections in response to IFNs (39). Lim et al. found that the HBx-mediated reduction of TRIM22 may evade the antiviral effect of IFNs in a host, implying that HBx is one of the factors controlled by the pathogens to regulate DNA methylation. Subsequently, reducing IFN regulatory factor-1 binding affinity thereby decreases the IFN-stimulated induction of TRIM22h, which impedes IFN regulatory factor-1, which subsequently seems to influence DNA methyltransferase tool kit and DNA methylation (40). This process may help pathogens evade host immunity. Given our observations and these studies, it seems that the upregulation of ApDNMT-1 and ApDNMT-2 genes in A. pernyi after microbial infections can affect the DNA methylation status in this insect species, thereby modulating gene expression and affecting different physiological responses, including immunity. Based on the above observations, we predict DNMT genes can affect the immune responses of the insect. Data are limited on the immunological functions of DNA methylation in insects, and we were interested in seeing whether DNA methylation might be implemented in the modulation of innate immune response in A. pernyi. For this purpose, we administrated 5-AZA to inhibit the activity of DNMT genes in A. pernyi. This DNMT inhibitor is commonly used to suppress DNA methylation in the genomes of insects (40)(41)(42)(43). Our results revealed that bacterial replication and mortality rates were lower following 5-AZA treatment compared to the control group. Additionally, colony-forming units were also lower in the 5-AZA treated larvae, indicating that the DNMTs are essential for the optimal growth of bacteria in insects. To the best of our knowledge, there is only one report on the necessity of DNMTs for bacterial replication in insects (30). However, in humans, the tick-born bacterial pathogen Anaplasma phagocytophilum greatly triggers hypermethylation in the neutrophil genome after 24 h postinfection, followed by differential expression of genes. In addition, the number of bacteria (A. phagocytophilum) growing in the presence of 5-AZA was approximately 25% less than in the control group (44).
Finally, we analyzed whether the mRNA expression of A. pernyi AMPs could be stimulated after an E. coli challenge and whether DNA methylation inhibition by 5-AZA had any effect on the transcription levels of these antimicrobial peptides. We observed that inhibition of DNMT inhibitor regulated AMPs following the bacterial infection. These results show that DNA methylation has a potential effect on the expression patterns of these peptides under bacterial infection, leading us to suggest that in A. pernyi, DNA methylation influences transcription of AMPs in response to Gram-negative E. coli. In humoral immunity, AMPs have a crucial role in suppressing microbial infection in insects (45,46). Thus, the variation in the transcription levels of AMPs can shape the insect-bacterial pathogen interactions. Considering the increase of some AMPs transcription and the reduced colony-forming units and bacterial replication in A. pernyi following 5-AZA treatment, it also suggests that changes in the production of AMPs are one of the key molecular mechanisms that attenuate the replication of bacteria. In human cells, it has been shown that 5-AZA has an immunomodulatory function (47). In addition to insects, it has also been demonstrated in human that DNA methylation can be changed in a short period of time in response to microbial infection and that these changes in DNA methylation can influence immune cell responsiveness (48), suggesting that microbes, particularly bacteria, use DNA methylation strategies to evade host immune responses (49)(50)(51). A recent study analyzed the global transcriptional patterns of gene expression after M. tuberculosis, and found that the majority of differentially expressed genes were hypermethylated, which could play a biological role in the activation of innate and adaptive immune cells (51). Helicobacter (H.) pylori are also one of the most studied pathogens in terms of DNA methylation in humans.
H. pylori causes changes in DNA methylation, and this abnormal DNA methylation in H. pylori-infected gastric mucosae has been linked to a higher risk of gastric cancer (52). DNA hypermethylation caused by H. pylori infection can be partially reversible once the bacterium is eliminated or 5-AZA is administered, resulting in a lower risk of gastric cancers triggered by H. pylori infection (53,54). Besides, A noncatalytic biological role has also been described for the DNMT-1 gene since it can control expression levels without any variation in DNA methylation patterns, probably via interactions with other enzymatic proteins associated with epigenetic gene modulation (55,56). Overall, whether the changes in the AMPs transcription patterns are due to the methylation status of these genes, or some other molecular mechanism, is still unexplored; however, the molecular mechanism could also be via variations in the gene transcriptional machinery triggered by the lack of DNMTs' activity. Furthermore, the upregulation of AMPs found in this study was mainly controlled by the Toll and IMD signaling pathways in insects, implying that DNA methylation may influence the activity of these pathways. Our analysis also confirmed that the 5-AZA and bacterial treatments influenced the expression of Toll and IMD pathway-associated genes. Our findings are consistent with previous studies; for example, evidence indicates that despite changes in AMPs transcription caused by DNA methylation (57)(58)(59), the Toll and IMD pathway-associated genes are modulated by epigenetic mechanisms, particularly by DNA methylation (60,61). Thus, we conclude that DNA methylation influences the expression patterns of AMPs directly or indirectly, which helps a host organism to attenuate microbial infections.
In summary, our study is the first to identify and clone DNA methyltransferases (DNMT-1 and DNMT-2 genes) from the Chinese oak silkworm A. pernyi. Phylogenetic and structural analysis revealed a high similarity of these proteins with their respective homologous enzymes in other insects. ApDNMT-1 and ApDNMT-2 were observed in all tissues examined, and their transcription patterns changed in response to microbial infection. As a DNMTs inhibitor, 5-AZA treatment reduced the replication rate, colony-forming units, and mortality of larvae compared to the control group. Additionally, after 5-AZA treatment, transcription patterns of AMPs and their associated pathway-related genes were altered in response to bacterial infection. Altogether, our data uncovered a novel FIGURE 11 | Treatment with 5-AZA modulated the expression of Toll and IMP pathway-associated genes in A. pernyi following bacterial infections. RT-qPCR analysis of relish, Toll-like receptor, and dorsal confirmed that 5-AZA administration altered the expression levels of these pathways-associated genes in response to B cereus and E. coli. 18S was used as the reference gene in RT-qPCR. Reactions from three biological replicates were repeated three times. Bars show mean ± S.E. (n = 3), and asterisks indicate significant differences (*p < 0.05).
immune regulatory function of DNMTs that shapes the insectpathogen interaction in the Chinese oak silkworm A. pernyi.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.

AUTHOR CONTRIBUTIONS
SK, MA, and IG: study design, data acquisition, results analysis, and prepared manuscript. RL, EZ, and QL: analysis of results and modification of the manuscript. HC and ML critical review, modification of manuscript and fund acquisition. All authors contributed to the article and approved the submitted version.