Behavioral, Physiological, and Molecular Mechanisms Underlying the Adaptation of Helicoverpa armigera to the Fruits of a Marginal Host: Walnut (Juglans regia)

Abstract

In northwest China, changes in cultivation patterns and the scarcity of preferred hosts have forced Helicoverpa armigera to feed on the marginal host walnut (Juglans regia). However, the mechanisms allowing this adaptation remain poorly understood. Here, we investigated the behavioral, physiological, and molecular mechanisms underlying the local adaptation of this pest to walnut fruits. The green husk and shell generally contained higher levels of phytochemicals than the kernel. Bioassays revealed that the phytochemical-rich green husk and shell were less preferred, reduced larval fitness and growth, and elevated the activity of detoxification enzymes compared to the nutrient-rich kernel, which were further supported by a larger number of upregulated detoxification genes in insects fed green husks or shells based on transcriptome sequencing. Together, these data suggest that P450 genes (LOC110371778) may be crucial to H. armigera adaptation to the phytochemicals of walnuts. Our findings provide significant insight into the adaptation of H. armigera to walnut, an alternative host of lower quality. Meanwhile, our study provides a theoretical basis for managing resistance to H. armigera larvae in walnut trees and is instrumental in developing comprehensive integrated pest management strategies for this pest in walnut orchards and other agricultural systems.

1. Introduction

The cotton bollworm, Helicoverpa armigera (Lepidoptera: Noctuidae), is a polyphagous agricultural pest that feeds on a wide range of host plants, including cotton, chickpea, tomato, tobacco, corn, sesame, hemp, sunflower, peanut, okra, and soybean [1,2,3,4,5,6], causing serious economic losses to agriculture worldwide [7,8]. Generalist herbivores such as H. armigera must continually shift to new, alternative, and nutritionally challenging host plants as the availability or quality of other hosts declines. In China, H. armigera has been reported to shift onto some marginal hosts when preferred host plants become scarce either due to seasonality or changes in cultivation or land use patterns. The increased cultivation of high-value horticultural crops in areas formerly dominated by cotton production has led to greater exposure of some marginal hosts to H. armigera feeding [7]. The Southern Xinjiang Uygur Autonomous Region, China, has an arid or semi-arid climate that is ideal for cultivating high-quality fruit of various species, including walnut (Juglans regia), jujube (Ziziphus jujube), apple (Malus pumila), grape (Vitis vinifera), pear (Pyrus sinkiangensis), and apricot (Prunus armeniaca). Recent changes in agricultural production in the region and advancements in farming technology have led to an increased level of exposure of these fruits to H. armigera. For example, since 2017, damage to walnuts by first-generation H. armigera has been reported in Aksu, Southern Xinjiang, resulting in significant economic losses for the local walnut industry [9,10]. However, the mechanisms underlying the local adaptation of H. armigera larvae to walnut fruits are unclear.

Polyphagous generalists such as H. armigera feed on a wide range of host plants [11,12,13], and in doing so their larvae encounter a wide range of plant allelochemicals, against which herbivorous insects have developed various mechanisms to overcome plant defenses [14,15,16], including behavioral, physiological, and transcriptional changes that promote better detoxification of defense compounds [17,18,19,20,21,22]. The metabolic detoxification of such defensive compounds is commonly categorized into three phases [23,24]. In Phase I, enzymes such as cytochrome P450 monooxygenases and carboxylesterases (CarE) act upon the toxin molecule, enhancing its hydrophilicity and the reactivity of its functional groups. In the second stage, Phase II enzymes, including UDP-glycosyltransferases (UGTs) and glutathione S-transferases (GSTs), facilitate the conjugation of endogenous molecules with toxins. Also, other transferases such as methyltransferases, acetyltransferases, phosphotransferases, and sulfotransferases contribute to this conjugation process [25,26]. The resulting conjugates are more hydrophilic and less reactive, diminishing their ability to pass through membranes. Finally, in Phase III, ATP-binding cassette (ABC) transporters bolster the transmembrane transport of toxins, thereby aiding their excretion. Many studies have demonstrated that in H. armigera, key detoxification enzymes, including cytochromes P450 (P450s), are involved in the detoxification of phytochemicals, helping this pest to adapt to diverse toxins produced by host plants [24,27]. For example, CYP6AE14 is involved in the adaptation of H. armigera to gossypol, and CYP303A1 plays a role in the metabolism of the plant allelochemical 2-tridecanone [27].

In recent years, with the rapid development of next-generation sequencing technology, an increasing number of studies have employed RNA-seq to explore adaptive changes in insects in response to different host plants or xenobiotic phytochemicals [28,29]. For example, transcriptomic analysis was used to determine the mechanism of adaptation for the maize pest Diabrotica virgifera to different host plants. This study found that the larvae of this rootworm underwent fewer transcriptional changes when feeding on an alternate host plant than when feeding on a marginal or poor host [30]. Transcriptome profiling of the larvae of the weevil Pagiophloeus tsushimanus exposed to the plant-derived volatile compound linalool showed that linalool exposure induced the expression of genes encoding for P450s, cuticular proteins, antioxidases, heat shock proteins, and digestive enzymes [23]. Moreover, a review of studies that used transcriptomic analyses to investigate how genome-wide transcriptional changes contributed to adaptation over time to novel or poor-quality host plants found that during that process [31], many genes involved in detoxification (P450s, CarE, UGTs, GSTs, ABC transporters), digestion, chemosensory perception, and transcription factors were dynamically expressed. For instance, transcriptome profiling of two tobacco-adapted species of the whiteflies Trialeurodes vaporariorum and Bemisia tabaci found that CYP6DP1 and CYP6DP2 (in T. vaporariorum) and CYP6DP3 (in B. tabaci) played important roles in strong induced tolerance to the toxic allelochemical nicotine [32].

In the present study, we investigated for the first time the effects of walnut fruits (green husk, shell, or kernel parts) on the feeding preferences and fitness costs of H. armigera. Then, the detoxification activities of P450 and carboxylesterase enzymes in H. armigera larvae fed on the above three parts of walnut fruits were also assessed. Finally, the molecular mechanisms underlying the adaptation and detoxification of H. armigera larvae to these three walnut fruit parts were examined using RNA-seq and confirmed by qPCR analysis. This study was designed to develop novel strategies for the sustainable management of H. armigera in walnut orchards and improve understanding of the interactions between polyphagous pests and their host plants.

2. Results

2.1. Secondary Metabolite Content in Walnut Fruit

Principal component analysis (PCA) indicated the variation in secondary metabolite content in different walnut fruit tissues. The first two principal components explained 67.1% and 32.5% of the variance, respectively, with a 99.6% cumulative percentage of explained total variances. Tannin, total polyphenol, total alkaloids, and PHBA had high contributions to the PC1 axis, whereas gallic acid and total flavone had high contributions to the PC2 axis. For the three parts of walnut fruit (green husk, shell, and kernel), the contents of tannin, total polyphenols, and chlorogenic acid were highest in the shell, while the contents of total flavones and gallic acid were highest in the green husk. Generally, the green husk and shell contained higher levels of phytochemicals compared to the kernel (Figure 1, Table S1).

2.2. Feeding Preferences of Larvae on Walnut Fruit Parts

The H. armigera larvae demonstrated a strong preference for the kernel over the green husk or shell of the walnut fruit. At 24 h after the onset of feeding, the consumed area of the kernel was 1.06 cm² compared to only 0.10 cm² and 0 cm² for the shell and green husk. Consumption was significantly affected by food (F = 1259.0; df = 2, 6; p < 0.001) and was significantly greater in kernel-fed larvae than in shell- or green husk-fed ones (Figure 2).

2.3. Development of Larvae on Walnut Fruit Parts

Consumed food had a significant influence on development duration (F = 543.0, df = 3, 8, p < 0.001), mortality rate (F = 154.0, df = 3, 8, p < 0.001), and pupal weight (F = 317.4, df = 3, 8, p < 0.001). Larvae fed walnut fruit parts had developmental delays, higher mortality rates, and reduced pupal weights compared to the control group reared on artificial diets (Figure 3). Both the green husk and shell treatments led to longer developmental periods, significantly higher mortality rates, and lighter pupae compared to the kernel treatment. The green husk treatment resulted in the longest developmental period (41.5 days from the second to the sixth instars; the highest mortality rate was 37.8%) and the lightest pupae (0.08 g).

2.4. Larval Fresh and Dry Weights

Healthy fifth instar larvae were chosen to be fed on different walnut fruit parts for 48 h, after which their fresh and dry weights were recorded (Figure 4A). Larvae fed green husks had an average fresh weight of 0.092 g, while those consuming the shell or kernel weighed 0.144 g and 0.219 g, respectively. By comparison, larvae fed artificial feed weighed 0.284 g on average. The fresh weights of larvae significantly depended on eaten food (F = 382.0; df = 3, 116, p < 0.001), and feeding with green husks led to significantly lower weight than all other treatments.

Average dry weight (Figure 4B) was lowest for larvae fed green husks (0.041 g), compared to 0.067 g and 0.089 g for larvae fed shells or kernels, respectively. Larvae fed on the artificial diet had the highest average dry weight (0.152 g). Dry weight was affected by food treatments (F = 125.9; df = 3, 116, p < 0.001) and larvae fed green husks had significantly lower dry weight than those fed other treatments or the control.

The water content of larvae fed green husks was 5.0%, compared to 8.9% for those fed shells, 13.0% fed kernels, and 15.2% fed the artificial diet (Figure 4C). Significant differences among experimental groups were recorded for water content (F = 65.7; df = 3, 116, p < 0.001). The water content of larvae fed with green husks was significantly lower than that of the other treatments or the control.

2.5. Activities of Detoxification Enzymes Depending on Treatments

The activities of three major detoxifying enzymes—cytochrome P450 (P450), carboxylesterase (CarE), and glutathione S-transferase (GST) (Figure 5)—were significantly different between the various treatments of the larvae. After 24 h of feeding, the P450 enzyme activity in larvae was 1.484 nmol/min/mg prot for green husks, 0.805 nmol/min/mg prot for shells, 1.031 nmol/min/mg prot for kernels, and 0.838 nmol/min/mg prot for the artificial diet (Figure 5A). Notably, P450 activity in larvae fed green husks was significantly greater than that in larvae fed other treatments or the control (F = 25.1, df = 3, 8, p < 0.001). After 48 h of feeding, the P450 activities in larvae fed green husks and shells were both significantly higher than in larvae fed kernels. However, there was no significant difference in P450 enzyme activity between larvae fed green husks versus shells (Figure 5B).

The level of carboxylesterase (CarE) activity after feeding on green husks for 24 h was 6.260 µmol/min/mg prot, compared to 4.903 µmol/min/mg prot after feeding on kernels and 4.352 µmol/min/mg prot on shells. The lowest activity was 2.322 µmol/min/mg prot in the larvae fed the artificial diet. The CarE activity in the larvae fed green husks was significantly higher than those fed shells, but was not significantly different from larvae fed on kernels (Figure 5C). After 48 h, CarE activities in larvae fed green husks and shells were both significantly higher than in larvae fed kernels (Figure 5D). After 24 h, the glutathione S-transferase (GST) enzyme activity in H. armigera larvae that fed on different parts of the walnut fruit was significantly higher than in larvae that fed on the control (Figure 5E). After 48 h, the GST activity in larvae that were fed on green husks was significantly lower than in those fed on kernels, but was not significantly different from those that were fed on shells (Figure 5F).

2.6. Transcriptional Responses in Larvae Depending on Treatments

To identify the molecular mechanisms of the local adaptation of H. armigera to alternative host plants and to identify potential key genes and associated biological processes, transcriptomic changes in the larvae among treatments were monitored with RNA-seq analysis at 24 and 48 h of feeding. After 24 h, the numbers of DEGs in larvae fed green husks, shells, or kernels were 986 (623 up- and 363 downregulated), 1177 (760 up- and 417 downregulated), and 712 (380 up- and 332 downregulated), respectively (Figure 6A). Of these, 62 DEGs were common across all treatments (Figure 6B), while 730, 741, and 252 DEGs were specific to the green husk, shell, and kernel treatments, respectively. After 48 h of feeding, 986 (623 up- and 363 downregulated), 1177 (760 up- and 417 downregulated), and 712 (380 up- and 332 downregulated) DEGs were identified in the green husk, shell, and kernel treatments, respectively (Figure 6C). A Venn diagram showed that 36 DEGs were present in all three treatments (Figure 6D), while 826, 188, and 355 DEGs were unique to the green husk, shell, and kernel treatments, respectively.

2.7. Expression Patterns of DEGs in Different Treatments

After 24 h, a total of 2268 genes were differentially expressed in at least one control-versus-treatment comparison, while 1664 genes were differentially expressed at 48 h. A k-means clustering analysis on expression patterns of these DEGs under different treatments showed distinct clusters. At 24 h, three clusters (of 647 total genes) that exhibited the lowest expression levels were associated with the green husk treatment (clusters a, c, and e). Similarly, three clusters (499 genes) that had the lowest expression levels were associated with the shell treatment (clusters b, g, and h), and two clusters (210 genes) displayed the lowest expression level for the kernel treatment (clusters d and f) (Figure 7A). At 48 h, 72 genes in one cluster that showed the lowest expression level were associated with the green husk treatment (cluster c), while 263 and 303 genes that showed the lowest expression were associated with the shell (clusters b and g) and kernel treatments (clusters a and e), respectively. In contrast, 89 genes that exhibited the highest expression were associated with the shell treatment (cluster f), and 50 genes with the highest level of expression were associated with the kernel treatment (cluster d) (Figure 7B).

2.8. GO and KEGG Functional Enrichment Analysis of DEGs

GO functional enrichment and KEGG pathway analyses revealed that DEGs were mainly enriched in the following biological processes: immune response, insect hormone biosynthesis, chitin synthesis and degradation, detoxification, and digestion (Figure 8). At both 24 and 48 h, the treatments with the green husk and shell diets generally induced more DEGs encoding detoxification enzymes (P450, carboxylesterase, glutathione S-transferase, UDP-glucosyltransferase, and ABC transporter) and digestive enzymes (amylase, maltase, lipase, phospholipase, trypsin, chymotrypsin, carboxypeptidase, aminopeptidase, and serine protease) compared to the kernel treatment (Figure 8). Notably, the general induction of P450 in larvae in treatments with green husks, shells, or kernels at both time points highlights its importance and pivotal role in facilitating H. armigera adaptation to green walnut fruits. At 24 h, 30 P450 genes, including 12 belonging to the CYP4 family, 12 belonging to the CYP6 family, and 3 belonging to the CYP9 family, were differentially expressed in at least one control-versus-treatment comparison, while 14 P450 genes showed differential expression at 48 h, with 4 from the CYP4 family, 7 from the CYP6 family, and 1 from the CYP9 family (Figure 9A). In the three different treatments, the expression levels of certain genes, such as the LOC110371778 gene, were elevated at the two time points of 24 h and 48 h. Green husk-exposed larvae exhibited the greatest number of differentially expressed P450s at both 24 h (17 genes) and 48 h (8 genes). Eight P450 genes were selected and confirmed by subsequent qPCR, and their expression levels were highly consistent with those from RNA-seq, indicating that the RNA-seq results are reliable (Figure 9B). Fourteen P450 genes were differentially expressed in RNA-seq (Figure 9C) at 48 h, including two belonging to the CYP3 family, four belonging to the CYP4 family, and six belonging to the CYP6 family. In addition, two genes belonging to the CYP9 family were differentially expressed in at least one control-versus-treatment comparison. Five P450 genes were selected and confirmed by subsequent qPCR, and their expression levels were highly consistent with those from RNA-seq, indicating that the RNA-seq results are reliable (Figure 9D).

3. Discussion

Helicoverpa armigera is a polyphagous pest known to damage several crops [16,33,34,35]. While the literature on H. armigera is large, few reports mention their capacity to feed on walnuts [7,10]. This study was started following field observations of H. armigera infestation in walnut orchards in the Xinjiang region of China. Its goal was to investigate the mechanisms by which H. armigera larvae were adapting to walnut fruit. In walnut production, the kernel, the edible part of the fruit, is primarily consumed for its nutritive value, while the by-products, particularly the green husk and shell, are widely used in traditional medicine and have received increasing interest in modern pharmacology due to their abundant phytochemicals and excellent antioxidant activities [36]. In this study, we first evaluated the phytochemical contents of different parts of fresh walnut fruits (green husk, shell, and kernel). Similarly to previous publications [36], our findings revealed that the green husk and shell contained substantially higher quantities of phytochemical compounds compared to the nutrient-rich kernel (Figure 1, Table S1).

We further provided H. armigera larvae with various parts of fresh walnut fruit (green husk, shell, and kernel) and observed their feeding preferences. Significant differences in preference for parts of the walnut fruit were noted: larvae showed the highest preference for the kernel, followed by the shell, and the green husk was the least acceptable (Figure 2). In earlier work, H. armigera larvae have shown a stronger affinity for cotton leaves compared to those of tobacco, tomatoes, or peppers. Significant differences in the average relative growth rates of larvae on young leaves from these four hosts were observed, and showed a descending order of cotton, tobacco, tomato, and pepper [20]. In comparison to natural substrates such as okra, tomato, eggplant, pepper, and maize, H. armigera exhibit a pronounced preference for ovipositing on artificial diets, with maize being the least preferred option [37,38]. In our present study, the average mortality rate of H. armigera larvae consuming green husks was 37.8%, followed by 26.7% for shells, 7.8% for kernels, and only 1.1% for the larvae fed on the artificial diet. The developmental duration from the second instar to the sixth instar was longest for larvae fed on green husks (41.5 days), followed by those fed on shells (25.0 days) and kernels (16.4 days), whereas only 11.2 days were needed on the artificial feed. The time required for the H. armigera larvae to develop on green husks was significantly longer than on the shells, kernels, or artificial diet. This pattern was also found in pupal weight. H. armigera larvae fed on green husks were significantly lighter than those fed on shells, kernels, or the artificial diet. Shorter developmental times, higher total reproduction, and higher pupal weight on a host plant are indicative of greater host suitability [39]. Our study similarly determined that larval development on walnut kernels was significantly shorter than on shells and green husks, with a resultant heavier pupal weight. Furthermore, larvae feeding on the green husks exhibited significantly lower average fresh and dry weights compared to those feeding on shells or kernels (Figure 4). These findings suggest that H. armigera larvae are better adapted to walnut kernels than green husks or shells.

Changes in gene expression in H. armigera larvae in the process of adapting to walnut fruits were investigated using RNA-seq analysis and subsequent qPCR validation. Functional classification revealed that the DEGs identified in this process, compared to the larvae fed the artificial diet, were primarily involved in processes essential for growth and development, such as immunity, hormone regulation, cuticle formation, detoxification, and digestion. The green husk and shell treatments induced more upregulated genes than the kernel treatment (Figure 6A,C). For example, DEGs responsible for chitin synthesis and degradation were found in the green husk and shell treatments, with almost all of these genes being upregulated; however, these DEGs were absent in the kernel treatment (Figure 7). Many DEGs responsible for cuticle protein were also identified, with almost all showing upregulation across all three treatments at both time points. Previous studies have shown that several genes encoding cuticle proteins were also induced in the midgut and fat body of Spodoptera litura larvae that were exposed to the botanical quinoline alkaloid compound camptothecin [40,41] and in P. tsushimanus larvae exposed to the plant-derived, volatile compound linalool [23].

P450 enzymes involved in the solubilization phase of detoxification play important roles in the adaptation of herbivores to the detrimental secondary metabolites of novel host plants (i.e., phytochemicals) [34,35,42,43]. For example, the P450 gene CYP321A9 from S. frugiperda was found to be induced by toxic plant metabolites [44] and to be involved in the adaptation of the corn strain of S. frugiperda to the defensive chemicals in rice [45]. Similarly, the expression of CYP321B1 was significantly upregulated in the midgut and fat body of Spodoptera litura larvae in response to tannin levels in host plants [46]. In Mythimna separata, RNA-seq analysis showed that seven P450 genes were significantly upregulated and the enzyme activity of CYP450 was induced by different concentrations of chlorogenic acid, a common secondary metabolite in plants [47]. Similarly, in H. armigera, several P450 genes have been found to be induced by plant allelochemicals, indicating their potential role in the detoxification of these compounds. Several examples include (1) CYP6AE14 induction in the midgut by the cotton metabolite gossypol [48], (2) CYP6B6, CYP315A1, CYP18A1, and CYP306A1 induction by 2-tridecanone in wild tomato [35,49], (3) CYP6B2, CYP6B6, and CYP6B7 induction in the midgut and fat body by four allelochemicals in many plants [50], and (4) CYP321A1 induction by flavone in many plants [51]. In our study on walnut, the phytochemical-rich green husk and shell increased P450 enzyme activities at both time points (Figure 5A,B). Also, H. armigera larvae fed on the green husk, shell, or kernel treatments all upregulated multiple DEGs encoding detoxification enzymes and proteases, especially several P450s (Figure 7). Interestingly, a P450 gene (LOC110371778) from the CYP4 family was upregulated in all three treatments at the 24 h and 48 h time points, suggesting it may play an important role in the detoxification of green walnut allelochemicals. This specific function requires further study. Additionally, several UDP-glucosyltransferases and ABC transporters involved in the conjunction and excretion phases of detoxification were also upregulated across all three treatments, with the green husk treatment showing the most significant gene upregulation. Similarly, both at 24 h and 48 h, CarE and GST genes were significantly upregulated in response to feeding on different parts of the walnut fruit, with the green husk treatment showing particularly significant gene upregulation.

In addition to detoxification enzymes, digestive enzymes also play a role in the adaptation of herbivorous insects to novel plants [28,52]. For example, Hyphantria cunea larvae exhibit varying digestive enzyme activities depending on the host plant, which plays an important role in the adaptive compensatory mechanisms of digestive enzyme inhibitors across different plants. Further studies found that the nutrients needed for the growth and development of larvae came from the activity of digestive enzymes [53]. When feeding on host plants with high nutrient content, Heterolocha jinyinhuaphaga larvae enhance their digestive enzyme activity to obtain more nutrients, ensuring the rapid growth of the organism [54]. These results indicate that insects adjust the activity of digestive enzymes according to their own growth and development when feeding on different host plants. In our study, feeding H. armigera larvae with any of the three walnut fruit treatments, at both time points, changed the expression levels of digestive enzymes (Figure 7). The number of DEGs for the green husk treatment (50 genes at 24 h and 63 at 48 h) and the shell treatment (40 at 24 h and 36 at 48 h) encoding digestive enzymes were elevated compared to the kernel treatment (28 at 24 h and 48 at 48 h) at both time points, suggesting that the green husk and shell treatments had a stronger negative influence on digestion, which in turn would affect larval growth and development. The green husk and shell treatments induced greater differential gene expression, supporting the results of the biological tests and feeding preferences. Additionally, kernel preference was more conducive to the growth and development of H. armigera. Many studies have demonstrated that the digestive system of insect herbivores is key to their adaptation to novel host plants during host shifts, and feeding insects new host plants with novel phytochemicals can result in the upregulation of many digestive genes [23,28,42,43,55].

4. Materials and Methods

4.1. Plant and Insect Sources

Walnut fruits of J. regia cv. Wen 185 were collected from the Xinjiang Experiment Station of the Chinese Academy of Agricultural Sciences, Aksu, Xinjiang, China (80.34° E, 41.12° N). The adults of Helicoverpa armigera collected from the Xinjiang Experiment Station were paired for matin; the resulting offspring larvae were subsequently reared on an artificial diet under controlled conditions (25 ± 1 °C, 60 ± 5% RH, and a 16:8 L:D photoperiod) [56].

4.2. Determination of Secondary Metabolite Content

Samples of 2 g were extracted with 60% acetone and measured using the Folin–Ciocalteu method as described previously [57], with some modifications. The absorbance was measured at 725 nm. The total alkaloid assay was performed with berberine as a reference standard, as reported previously [10]. Total polyphenols were measured by the Folin–Ciocalteu method as previously reported [58], with some modifications. Briefly, 1 g samples were extracted by 60% ethanol. The absorbance was measured at 778 nm. The gallic acid solution was used as a standard. Total flavonoid content was determined using aluminum chloride reagent as described previously [59]. The absorbance was measured at 415 nm. The rutin standard solution was used as a reference standard. Monomeric phenolics including gallic acid, chlorogenic acid, and p-hydroxybenzoic acid (PHBA) were measured on an Agilent 1220 Infinity II gradient liquid chromatography system equipped with an Agilent ZORBAX SB-C18 column (Santa Clara, CA, USA). A total of 2.0 g of the sample was extracted by 75% methanol. Solvent A: 0.1% formic acid; solvent B: acetonitrile. Gradient elution was as follows: 0–5 min, 5% to 20%; 5–20 min, 20% to 100%; 20–21 min, 100% to 5%; 21–30 min, 5%. The flow rate was 0.5 mL/min. An injection volume of 10 µL was used. These metabolites were verified by the synthesized standards.

4.3. Larval Feeding Preferences on Walnut Fruit

Walnut fruit parts (green husk, shell, and kernel) were taken as the initial area, then cut into slices with a thickness of 2 mm with a knife. In each Petri dish, a thin slice of walnut was placed alongside a 3rd-instar larva of H. armigera. Image J software (v1.8.0) was used to measure the area consumed by the larvae before and after a 24 h period of 3rd-instar feeding to determine the relative preferences of larvae among parts. The average feeding area of H. armigera from 30 Petri dishes was used as one replicate, and this was repeated three times, for a total of 90 Petri dishes. The experimental conditions were 25 ± 1 °C and 60 ± 5% RH, and a 16:8 L:D photoperiod.

4.4. Determination of Fresh and Dry Weights of Test Larvae

To establish breeding protocols for H. armigera, we selected newly molted 5th-instar and placed each larva individually in sterile vials. Each larva was fed either green husk, shell, or kernel tissue, with an artificial diet serving as the control. After 48 h, we recorded the fresh weights of 30 larvae from each treatment. The larvae were subsequently dried at 80 °C until a constant weight was reached to determine their dry weights. The water content of the larvae was calculated by subtracting the dry weight from the fresh weight [60].

4.5. Growth and Development, Pupal Weight

The growth and development of H. armigera larvae were evaluated by providing them with green husk, shell, or kernel diets. Initially, neonate larvae were supplied with a specialized artificial diet tailored for the nutritional requirements of H. armigera. Upon entering the second instar, the larvae were placed individually in tubes and provided with one of the three walnut fruit parts. There were 30 larvae in each treatment, and the whole experiment was replicated three times, for 90 larvae per treatment. The experimental conditions were 25 ± 1 °C, 60 ± 5% RH, and a 16:8 L:D photoperiod. Fresh walnut material was supplied daily, and both the survival and developmental stages of the larvae were recorded daily. Pupal weights were measured upon larval pupation.

4.6. Detoxification Enzyme Activities

To assess detoxification enzymes’ activities, H. armigera larvae were reared to the 3rd instar on an artificial diet. At the conclusion of this stage, larvae were fed green husks, shells, or kernels as treatments and an artificial diet as the control. We then examined the impact of these treatments 24 and 48 h after the exposure to the diets by measuring the levels of the primary detoxification enzymes. Larvae were collected at random within treatments, and whole-body tissue was collected. Twenty individuals were used per treatment, and each treatment was repeated three times, for a total of sixty larvae. Protein concentrations were measured using the TP Assay Kit (A045-2) sourced from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). The activities of three key detoxification enzymes—carboxylesterase (CarE), Glutathione S—transferase (GST), and cytochrome P450 (P450)—were determined using the Carboxylesterase Assay Kit (A133-1-1), Glutathione S—transferase (GSH-ST) Assay kit(A004-1-1), and Cytochrome P450 Assay Kit (H303-1-2), respectively, adhering to the protocols provided by the Nanjing Jiancheng Bioengineering Institute (Nanjing, China).

4.7. Total RNA Extraction and Transcriptome Sequencing

Larvae of consistent size at the 3rd instar were selected and fed different walnut fruit parts (green husk, shell and kernel). These larvae were then collected 24 or 48 h after the start of the treatment. Control larvae were reared on an artificial diet. Each treatment contained 30 larvae and the whole experiment was replicated three times (so, 90 [3 × 30] larvae were tested for each treatment). Subsequently, total RNA from the sampled larvae was isolated using the TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) and subjected to RNA sequencing analysis. RNA concentration and purity were assessed using a NanoDrop spectrophotometer (Thermo Scientific, Wilmington, NC, USA). For cDNA library preparation, 3 μg of total RNA per sample was used as the input material with the NEBNext Ultra II RNA Library Prep Kit for Illumina according to the manufacturer’s instructions. Sample sequencing was conducted on the Illumina NovaSeq 6000 platform (Personal Biotechnology Co., Ltd., Shanghai, China). Clean reads were aligned to the H. armigera reference genome available at NCBI (https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_023701775.1, accessed on 26 October 2023) using HISAT2 (v2.1.0). Genes with |log2FoldChange| > 1 and p < 0.05 were identified as differentially expressed genes (DEGs). GO enrichment analysis and KEGG pathway enrichment analyses were performed using the topGO (v2.50.0) and ClusterProfiler (v4.6.0) packages, respectively.

4.8. qPCR Validation

Total RNA used for transcriptome sequencing was reverse-transcribed into cDNA using the PrimeScript™ 1st Strand cDNA Synthesis Kit (TaKaRa, Shiga, Japan). The qPCR validation was performed using the AceQ^® qPCR SYBR^® Green Master Mix kit (Vazyme Biotech, Nanjing, China) following the manufacturer’s protocols [61]. The qPCR amplification conditions were 95 °C for 5 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 30 s. The actin gene of H. armigera was used as a candidate reference gene [62]. All primers used for qPCR are listed in Table S3.

4.9. Statistical Analysis

Principal component analysis (PCA) was used to indicate the distribution of secondary metabolites in various tissues of the walnut fruit—specifically, the green husk, shell, and kernel. PCA was performed by using the “vegan” package [63] in R 4.2.1 software [64]. All values shown are means ± SE (standard errors). Significant differences for feeding preference, fresh weight, dry weight, growth, development, pupal weight, and detoxification enzyme activities between larvae fed walnut fruit parts and the artificial diet were analyzed using one-way ANOVA, followed by Tukey’s tests for mean separation. Statistical difference was considered significant at p < 0.05.

5. Conclusions

Crop diversity in the Aksu region of Xinjiang has increased the exposure of alternate, marginal, or poor-quality crops such as walnut to H. amigera. In this study, we provide insights into the interaction between H. armigera and this alternate host. Different parts of the walnut fruit exert distinct effects on the feeding preferences, larval growth and development, survival rates, and enzyme activity levels of larvae, with the phytochemical-rich green husk and shell having greater negative impacts compared to the nutrient-rich kernel. RNA-seq analysis showed that feeding on different walnut parts strongly affected the expression of genes involved in detoxification, digestion, immunity, and cuticle formation. More P450 genes were upregulated in the green husk and shell treatments than in the kernel treatment. These findings improve our understanding of the adaptive mechanisms being used by H. amigera to feed on walnut fruits specifically, and, more generally, also contribute to our understanding of host adaptation in phytophagous insects to alternate, marginal, or poor-quality hosts.

Finally, considerably more work will need to be carried out to determine the molecular functions of the P450 candidate genes and other genes involved in detoxification metabolism and their co-expression relationship.