Azadirachtin (95% purity, #A7430, Sigma-Aldrich, St. Louis, MO, USA) was dissolved into dimethyl sulfoxide (DMSO, #D103272, Aladdin, Shanghai, China) and mixed with artificial diet. The antibody of Sl-IBM1 was prepared by our laboratory. Cleaved Caspase-3 (Asp 175) antibody was obtained from Cell Signaling Technology (#9661, USA) and mouse polyclonal anti-cytochrome c was purchased from Beyotime Biotechnology (#AC909, Shanghai, China). Other chemicals were domestic products.

The S. litura larvae maintained in our laboratory (Guangzhou, China) were fed with an artificial diet upon hatching and maintained in a stable condition of 25 ± 1°C, 60–70% relative humidity and a 16:8 h light: dark cycle (Yi et al., 2017). The adults were maintained with 10% honey water. The third-instar larvae fed with the artificial diet supplemented with 1 μg/g azadirachtin for 7 d were used for the following experiments and defined as larvae with azadirachtin treatment. The larvae fed with the artificial diet containing 200 μL/kg DMSO were used as the control group.

Azadirachtin-treated and control larvae were dissected and the midgut was washed in cold phosphate buffered saline (PBS) and fixed with 4% paraformaldehyde (#G1101, Servicebio, Wuhan, China) at 4°C for more than 24 h. Then it was embedded in paraffin wax, sectioned at 4 μm slices, mounted on glass slides and stained by hematoxylin and eosin. Slides were visualized on a microscope (Nikon, Japan).

Total RNA was extracted from the midgut samples using TRIzol (#15596026, ThermoFisher Scientific, USA) as described in the manufacturer's instructions. Purity of total RNA was measured by a NanoDrop® spectrophotometer (Thermo Fihser, MA, USA). RNA integrity was determined by an Agilent 2100 (Agilent Technologies, CA, USA). Two libraries were constructed, one for the midgut of azadirachtin-treated larvae and one for the control. A total of 1 μg qualified RNA per sample was used for the library preparation. The sequencing libraries were generated by VAHTS mRNA-seq v2 Library Prep Kit for Illumina® (Catalog NR601, Vazyme, Nanjing, China) following manufacturer's protocol. The clustering of the index-coded samples was performed on a cBot Cluster Generation System (Illumia, USA) according to the manufacturer's instructions. After cluster generation, the library preparations were sequenced on an Illumina Hiseq X Ten platform with 150 bp paired-end module. The de novo assembly of clean reads was performed by Trinity (settings: –min_contig_length150 –CPU 8 –min_kmer_cov 3 –min_glue 3 –bfly_opts ‘-V 5 –edge-thr = 0.1 –stderr’) (Grabherr et al., 2011). The expression levels of unigenes based on the number of reads uniquely mapped were normalized by Reads Per Kilobase of exon model per Million mapped reads (FPKM) method. The thresholds of differentially expressed unigene were False discovery rate (FDR) ≤ 0.001 and |log₂Fold change| ≥1. Differentially expressed unigenes were subjected to Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis.

The total RNA of midgut samples was extracted by TRIzol reagent as above and reverse transcribed into cDNA with the PrimeScript™ RT reagent Kit (#RR047A, TaKaRa, Japan) following the manufacturer's recommendations. qRT-PCR was performed using the iTaq™ Universal SYBR® Green Supermix (#1725271, BIO-RAD, USA) with a CFX Connect™ Real-Time System (BIO-RAD, USA) under the following thermal program: one cycle at 95°C for 3 min, 40 cycles at 95°C for 10 s; 60°C for 10 s; 72°C for 15 s, and one step from 65 to 95°C for the dissociation stage. The primers used for qRT-PCR were listed in Supplement Table 1. Gene expression levels were calculated by 2^−ΔΔCt method (Livak and Schmittgen, 2001) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the reference gene (Lu et al., 2013).

TUNEL assay is one of the most effective methods for apoptosis detection in tissues (Sarkissian et al., 2014). Paraffin embedded midgut tissue section was prepared as described above. One Step TUNEL Apoptosis Assay Kit (#C1086, Beyotime, Shanghai, China) was used according to the manufacturer's protocol. Briefly, sections were dewaxed in xylene twice for 5 min, fixed with 100% ethanol for 5 min, then dehydrated with 2 min washes in 90% ethanol, 70% ethanol, distilled water. After that, the sections were treated with 20 μg/mL protease K (#1074, Servicebio, Wuhan, China) at 37°C for 30 min and washed three times with PBS. Slides were incubated with the TUNEL reagent in the dark at 37°C for 1 h, washed three times with PBS and counterstained with 4′, 6-diamidino-2-phenylindole (DAPI) prior to mounting in X mounting media. Slides were visualized on a fluorescence microscope (Nikon, Japan).

Whole protein extraction from midgut samples were carried out with the total protein extraction kit (#W034, Nanjing Jiancheng, China) according to the protocol. Cytoplasmic and mitochondrial proteins were extracted using the mitochondrial protein extraction kit (#KGP850, KeyGEN BioTECH, China). Samples were separated on a 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE) gel and transferred to a polyvinylidene fluoride (PVDF) (#IPVH00010, Millipore, USA) with the wet/semi-dry transfer system. The membranes were blocked in tris-buffered saline (TBS: 100 mM Tris–HCl, pH 7.5, 0.9% NaCl) with 5% fat-free milk at room temperature for 2 h and then incubated in primary antibodies (diluted 1:2000 in TBS) at 4°C overnight, washed by TBST (TBS containing 0.1% Tween-20) three times for 5 min and incubated in secondary antibody (1:2000 in TBS) at room temperature for 2 h. Membranes were developed in enhanced chemiluminescence (ECL) reagent (#1705062S, Bio-RAD, USA) and bands were visualized using ECL detection system (Bio-RAD, USA). GAPDH was used as the reference protein to normalize the differences in protein loading.

Approximately 3–10 mg of midgut samples dissected from larvae were prepared. The protein concentrations were determined with Bradford method. The caspase-3 activity was detected by the Caspase-3 Activity Assay Kit (#C1115, Beyotime, Shanghai, China) according to the manufacturer's protocol. Enzyme activities were measured by the absorbance of samples at 405 nm with a microplate reader (ThermoFisher Scietific, USA).

Data collected were expressed as the mean ± standard deviation. Statistical significance between control and treatment was determined with SPSS 20.0 using Student's t-test.

To confirm the growth inhibition phenotype conferred by azadirachtin, third instar larvae were exposed to azadirachtin added or control diet. After 7 days, the larvae fed with azadirachtin exhibited a growth inhibition phenotype. Specifically, the size of larvae with azadirachtin treatment was obviously smaller (Figure 1A). Additionally, the larvae fed with azadirachtin weighed significantly less, with a 43.4% decrease (1.93 g ± 0.093 vs. 3.41 g ± 0.234, P < 0.05, n = 20) (Figure 1B).

In order to determine the effect of azadirachtin on structure, the midgut samples of larvae were dissected, measured and stained with hematoxylin-eosin. After azadirachtin exposure, the size of larval midgut was significantly smaller (Figure 2A). In addition to gross structural changes, azadirachtin also conferred histopathological changes. The midgut of control displayed appropriate, tightly arranged cells (Figure 2B). However, it exhibited cells death, abnormal cell structure, and intestinal wall cracking after azadirachtin treatment (Figure 2B). These data demonstrated that azadirachtin disrupted the normal cell structure and induced cell death in the midgut.

In order to uncover the molecular mechanisms underlying the histological changes in the midgut of azadirachtin-treated larvae, transcriptional differences were measured by DGE profiling. As shown in Figure 3, a total of 4,091 unigenes were differentially expressed in midgut of azadirachtin treatment. Of these unigenes, 3,108 were up-regulated and 983 were down-regulated.

The GO enrichment analysis indicated that the differentially expressed unigenes (DEGs) affected by azadirachtin were classified into three categories and 57 functional groups. Among which 1,181 DEGs were assigned into the category of biological process, 915 DEGs to cellular component and 1,177 DEGs to molecular function (Supplement Figure 1). In biological process category, the DEGs assigned in were divided into 24 functional groups and single-organism process was the biggest group with 881 genes assigned, followed by the group of cellular process (881 DEGs), and metabolic process (740 DEGs). As to cellular component category, the biggest group were cell and cell part, which has the same number of 629 unigenes assigned. The DEGs classified into molecular function category were divided into 16 groups and catalytic activity group has the higher percentage of unigenes and 723 unigenes were concentrated.

The DGE analysis revealed 2,154 differentially expressed unigenes were involved in 247 KEGG pathways. Of these pathways, three pathways with the largest number of annotations were metabolic pathways (358 unigenes, 16.62%), pancreatic secretion (181 unigenes, 8.4%), and protein digestion and absorption (148 unigenes, 6.87%) (Supplement Figure 2).

To verify the results of DGE analysis, parts of unigenes with at least a two-fold difference in expression were selected. Because azadirachtin was considered to be a xenobiotic, 17 differentially expressed unigenes involved in xenobiotics biodegradation and metabolism pathway were validated by qRT-PCR (Table 1) and exhibited the same trends as the DGE profiling results (Figure 4A). These results indicated that three major detoxification enzymes (esterase, UDP-glycosyltransferase and cytochrome P450) were activated in midgut after exposure to azadirachtin.

In addition, many differentially expressed unigenes were mapped into the pathways involved in apoptosis, such as the mitogen-activated protein kinase (MAPK), calcium, and p53 signaling pathways. A total of 55 unigenes in these apoptotic pathways were also validated by qRT-PCR (Table 1). All the selected unigenes expression profiles were consistent with the results of DGE analysis (Figures 4B–D). These results suggested that azadirachtin activated MAPK, Calcium and other apoptosis signaling pathways within larval midgut.

TUNEL assay was utilized to directly demonstrate that azadirachtin induced apoptosis and morphological changes in midgut. The TUNEL-positive cells would fluoresce at 488 nm. TUNEL staining was undetectable in midgut of control larvae, while a high level of TUNEL-positive cells appeared in midgut of azadirachtin exposure (Figure 5). These results indicated that azadirachtin induced apoptosis in midgut.

Many unigenes were up-regulated in DGE of azadirachtin-treated midgut. One such differentially expressed unigene IAP-binding motif 1 (Sl-IBM1) (log₂ Ratio = 4.11), an IAP antagonist and Lepidoptera order homolog to Reaper, was thought to play an important role in apoptosis process (Wu et al., 2013). Therefore, the expression of IBM1 in midgut after azadirachtin treatment was verified by qRT-PCR and western blot. Azadirachtin exposure caused an up-regulation of Sl-IBM1 mRNA expression in midgut (Figure 6A). In addition, it could also induce an increase of Sl-IBM1 on protein level (Figure 6B). Taken together, azadirachtin exposure up-regulated the pro-apoptotic factor IBM1.

To determine if azadirachtin exposure and the increased Sl-IBM1 expression induced caspase expression within the midgut, mRNA expression of four caspase genes, including caspase-1, caspase-3, caspase-5, and caspase-6 were compared after azadirachtin exposure by qRT-PCR. All four genes were demonstrated to be up-regulated at least 2.75 folds after azadirachtin treatment (Figure 7A). Simultaneously, the cleaved caspase 3 protein was increased significantly (Figure 7B). Additionally, in accordance with gene and protein expression data, caspase-3-like protease activity was obviously increased (Figure 7C). The up-regulation of caspases expression and caspase-3-like activity confirmed that azadirachtin activated apoptosis in midgut by regulating the caspase-dependent apoptotic pathway.

To reveal whether mitochondria were involved in midgut apoptosis induced by azadirachtin, the protein levels of cytochrome c within cytoplasm and mitochondria of midgut cells were analyzed by western blot. After azadirachtin treatment, cytochrome c in cytoplasmic increased significantly while the mitochondrial reduced (Figure 8). These results indicated that the event of cytochrome c releasing from mitochondria into cytoplasm was occurred in midgut after azadirachtin treatment.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.