Transcriptomic Responses of Fall Armyworms (Spodoptera frugiperda) Feeding on a Resistant Maize Inbred Line Xi502 with High Benzoxazinoid Content

Abstract

The fall armyworm (Spodoptera frugiperda) is a devastating invasive insect herbivore. Its success on its preferred host plant, maize (Zea mays), is supported by numerous specialized detoxification mechanisms that suppress the defense responses of maize. In this study, we used a resistant Chinese maize cultivar, Xi502, which showed slower growth and lower yield-related phenotypes compare with maize inbred line B73. Comparative transcriptomic analyses demonstrated that B73-fed fall armyworm larvae have a significantly faster transcriptomic re-configuration toward maturation compared to their siblings fed with Xi502 leaves, whereas a number of putative aromatic breakdown -related DEGs were specifically induced when feeding on Xi502. Targeted metabolomic quantification demonstrated that Xi502 contains significantly higher levels of various benzoxazinoid compounds. Artificial feeding with the structural analog of a benzoxazinoid compound preferentially accumulated in Xi502 demonstrated a significant growth inhibition effect on FAW larvae. These results provide important genetic material and preliminary evidence for further dissection of the FAW-resistance mechanism in maize.

1. Introduction

The fall armyworm (Spodoptera frugiperda, FAW) is a lepidopteran herbivore originating from tropical and subtropical regions of America that specializes in feeding on maize (Zea mays). Since its invasion of the Old World, it has become a major emergent threat to the local maize industry and to food security [1]. In the Americas, the area from which this insect originated, the FAW is successfully controlled by the prevalent use of transgenic maize producing various Bacillus thuringiensis (Bt) protein toxins [2], whereas synthetic pesticides and agronomic practices have been deployed in regions with limited access to transgenic crops [3,4]. In recent years, increasing cases of field-evolved Bt resistance have been reported in FAW populations around the globe, raising the concern that additional methods are required to control FAW, especially in areas it has invaded [5]. Deployment of genetically pest-resistance crop cultivars is an effective and environmental-friendly pest control strategy. In maize, a crop well-known for its genetic diversity, an FAW-resistant cultivar Mp708 was developed in the 1990s [6], and its genetic resistance was attributed to an insecticidal cysteine proteinase, Mir1-CP [7,8,9,10]. In the last thirty years, a number of FAW-resistant maize genotypes have been identified in field tests, but few have been further studied for their resistance mechanisms [11,12,13,14,15].

In recent decades, FAW has also become an interesting model species due to its preference for feeding on maize compared to other species in the same genus. Benzoxazinoid compounds are a class of specialized metabolites found in maize, wheat, and rye [16]. These compounds play critical roles in the defense against diverse pathogens and insect herbivores [17,18]. However, the most common maize benzoxazinoid compound 2,4-dihydroxy-7-methoxy-1,4-benzoxazin-3-one glucoside (DIMBOA-Glc), is neither deterring nor toxic to FAW larvae, though it is effective against other Spodoptera species [19]. Later studies demonstrated that FAW can detoxify DIMBOA and its break-down products through re-glucosylation with specialized UDP-glucosyltransferases [20,21,22]. However, the methylated derivative of DIMBOA, 2-hydroxy-4,7-dimethoxy-1,4-benzoxazin-3-one glucoside (HDMBOA-Glc) demonstrated a significant repelling and growth-inhibiting effect on FAW larvae [19].

In this study, we identified the novel effects of a FAW-resistant maize cultivar, Xi502 on the physiology of the fall armyworm. Interestingly, this inbred line also showed slower growth, shorter stature, and lower yield compared to the FAW-susceptible check B73. Comparative transcriptomic analyses of the FAW larvae feeding on either maize genotype revealed transcriptomic re-configurations consistent with accelerated development in B73-fed larvae, but preferential up-regulation of aromatic compound catalysis-related genes when feeding on Xi502. Targeted measurement of benzoxazinoid compounds demonstrated that Xi502 constitutively accumulated a higher concentration of HDMBOA-Glc, which may explain the higher FAW resistance level in this inbred line. Artificial feeding assays with a synthetic analog of natural benzoxazinoid degradation product demonstrated a significant growth suppression effect on FAW larvae. Together, these results systematically demonstrate the transcriptomic response of FAW larvae toward maize cultivars of contrasting resistance levels, and shed light on the potential resistance mechanism of Xi502.

2. Materials and Methods

2.1. Plant Growth and Insect Materials

Seeds of maize genotypes were originally obtained from Dr. Jianbin Yan at the Huazhong Agricultural University. Fall armyworm larvae were supplied by Dr. Yutao Xiao at the Agricultural Genomics Institute at Shenzhen, which were originally collected from Ruili, Yunnan, China. For all experiments, maize seeds were in Hawita Aussaaterde an Stecklingserde potting substrate (Hawita Professional), Germany (https://www.hortiways.de/duenger-erden/erden-torf/2024/hawita-aussaaterde-70l) mixed with vermiculite (5:1; v/v). Seedlings were grown under 16:8 h day/night cycle and the temperatures were set at 24 °C (day) and 20 °C (night). To observe the contrasting phenotype of plant growth, data for the plant heights of both genotypes were recorded at the mature stage in the field and the number of seeds was counted after harvesting the cobs.

2.2. RNAseq Sampling, Library Preparation, and Data Analyses

For FAW larvae RNAseq, one second-instar larva was caged on the second expanded leaf of each seedling with a perforated 45 × 30 × 30 cm transparent PVC box for 24 h. For each genotype, ten biological replicates were prepared. The best grown of every two larvae were pooled and frozen in liquid nitrogen for total RNA extraction (i.e., 4 biological replicates for RNAseq per group). The other larvae that showed poor growth were discarded.

RNA from frozen insect tissues was extracted with TRIzoL reagent, and the quality of the extract was assessed with an RNA Nano 6000 Assay Kit of a Bioanalyzer 2100 system (Agilent Technologies, Santa Clara, CA, USA). For each sample, 1 µg of total RNA was used to prepare a paired-end sequencing library using a NEBNext^® UltraTM RNA Library Prep Kit for Illumina^® (NEB, Beverly, MA, USA) following the manufacturer’s protocols, and the index codes were ligated. The quality of the prepared library was assessed with a Bioanalyzer 2100 system.

After removing adaptor and low-quality reads, the remaining clean reads were mapped onto the S. frugiperda reference genome with a STAR aligner [23,24]. The filtered reads are available through the NCBI online depository (PRJNA729598) in fastq format. Raw read counts were summarized with HTseq-count, and differentially expressed genes (DEGs) were calculated with DESeq2 (false discovery rate < 0.05; fold change > 2) [25]. Gene ontology enrichment analyses (singular enrichment analysis) were carried out in the R environment using all expressed genes with available GO annotation as the custom background. Only biological-process-related GO terms were tested for enrichment.

2.3. Benzoxazinoid Extraction and Quantification

Approximately 150 mg of frozen leaf powders obtained from ten-days-old seedlings was extracted with aqueous methanol solutions (50:50, v/v; 0.5% formic acid). After a 15–20 min vortex, samples were centrifuged at 13,000× g for 15 min, and 400 μL of supernatants was taken for analysis on an HPLC-MS/MS (LCMS-8040, Shimadzu) with previously purified analytical standards [26].

2.4. Benzoxazinoid-Feeding Experiment

Commercially-purchased 3-methyl-2-benzoxazolinone (Sigma) was dissolved in DMSO to create a 330 mg/mL of stock solution. The stock solution was added to cooled artificial diet at 1:1000 to reach a final concentration of 330 µg/g FW of artificial diet, a concentration that was previously used to test for in vitro toxicity against FAW larvae (Maag et al., 2014). The same procedure was adopted to make DMSO-only control diets. Twenty-four neonates of FAW were reared on control or 3-MBOA-dosed diets for seven days, and the surviving larvae were weighed individually at the end of the feeding period.

2.5. Statistical Analysis

Benzoxazinoid content was normalized by tissue fresh weight and then Bx levels were compared between B73 and Xi502 with Student’s t-test (p < 0.05). The data from the herbivore bioassay were analyzed with Student’s t-test (p < 0.05). Principal component analysis was conducted and plotted using plotPCA in the DESeq2 package of R [27].

3. Results

3.1. Xi502 Plants Grow Slower and Demonstrate Shorter Stature

FAW resistant maize inbred line Xi502 and a susceptible inbred line B73 were grown in a greenhouse and in a field to observe the comparative phenotypes [28]. We noticed that Xi502 seedlings and mature plants were significantly shorter than B73 plants sowed at the same time (Figure 1a,b). Additionally, Xi502 plants produced fewer seeds per cob when only one cob per plant was hand-pollinated and retained for harvest (Figure 1c,d).

3.2. Feeding on B73 and Xi502 Causes Distinctive Transcriptomic Dynamics in FAW Larvae

As a maize-specialized insect herbivore, FAW has evolved various counter-defense mechanisms to facilitate its success on its preferred host plant. To study the potential physiological impact of feeding on Xi502 versus B73, we collected FAW larvae after 24 h of feeding on these two maize cultivars and an artificial diet (as the shared control group). The principal component analysis results of the RNAseq data demonstrated that transcriptomes of B73-fed larvae were more different from the control group than their siblings fed with Xi502 tissues (Figure 2a; Table S1). Consistently, 1069 induced and 593 suppressed DEGs were found between B73-fed FAW larvae and the control group, whereas the larvae feeding on Xi502 had only 367 up-regulated and 118 down-regulated DEGs compared to the same control group (Figure 2b). There was a clear overlap in the two sets of DEGs (B73-fed larvae versus control, and Xi502-fed larvae versus control), and all but two expression changes were in the same direction (Figure 2c). Additionally, more than 1200 genes were only significantly differentially expressed in B73-fed larvae, while those fed with Xi502 tissues had less than 100 of such group-specific DEGs.

Gene ontology term enrichment analysis with the common DEGs showed that a handful of GO categories related to lipid metabolism were significantly enriched among the induce genes, whereas cellular-transport-related genes were preferentially suppressed among these DEGs (Figure 2c; Tables S2 and S4). Genes induced specifically in B73-fed larvae were over-represented in cytoskeleton organization (GO:0030036), supramolecular fiber organization (GO:0097435), chitin metabolic process (GO:0006030), the development of reproductive systems (GO:0061458), and sensory perception systems (GO:0007605). In contrast, genes specifically suppressed in these larvae were enriched in small molecule metabolic processes (GO:0044281), such as oxoacid metabolism (GO:0043436), carboxylic acid metabolism (GO:0019752), and galactarate catabolism (GO:0046394). Since Xi502-fed FAW larvae had only a few group-specific DEGs, these genes were combined for GO term enrichment analysis. Intriguingly, the aromatic compound catabolic process (GO:0019439) emerged as the most significantly enriched GO term among these DEGs, suggesting that maize-induced regulation of this process may be specific for Xi502-fed FAW larvae.

As for specific genes, the recently cloned UDP-glucosyltransferase-encoding genes (UGT33: FAW_001276; UGT40: FAW_014880) that catalyze the re-glucosylation of benzoxazinoids Israni et al., [22] did not show significant changes in expression with the different diets tested after 24 h of feeding (Figure 3a). Rather, we focused on the 12 DEGs between Xi502-fed and artificial-diet-fed larvae that were putatively involved in aromatic compound catalysis. Ten of these DEGs were up-regulated in Xi502-fed larvae, while only two were down-regulated (Figure 3b). We further examined their expression levels in B73-fed larvae, and found only four of the up-regulated genes in Xi502-fed larvae were also induced in B73-fed ones, whereas the two down-regulated genes were consistently suppressed in both Xi502- and B73-fed larvae (Figure 3b). Interestingly, while all of the shared DEGs between larvae feeding on either maize genotype were expressed at relatively low levels (maximal group mean FPKM < 15; Figure 3c), three of the DEGs that were specifically found in Xi502-fed larvae were expressed at much higher levels (Figure 3d). These three genes (FAW_020866; FAW_003521; FAW_000283) encode a putative kynureninase, a putative uricase, and a putative extracellular endonuclease, respectively, which are potentially important for FAW survival on maize.

3.3. Xi502 Constitutively Contains Higher Levels of Benzoxazinoids

Maize seedlings constitutively produce diverse benzoxazinoid compounds in high abundance as an effective biochemical defense mechanism [16]. Different benzoxazinoid compounds have contrasting efficiency in deterring FAW larvae, which are adapted to maize defense [19] Quantification of seven benzoxazinoid glucosides showed that the resistant Xi502 accumulated significantly higher concentrations of four of these compounds, including HDMBOA-Glc, which was reported to be repel and inhibit the growth of FAW larvae (Figure 4; Table S3; p < 0.05; Student’s t-test; [19].

We also measured three benzoxazinoid breakdown products, namely 6,7-dimethoxy-benzoxazolin-2-one (M₂BOA), 6-methoxy-benzoxazolin-2-one (6-MBOA), and benzoxazolin-2-one (BOA), which were thought to be responsible for the inhibitory effect against insect herbivores [16]. Consistent with the pattern of benzoxazinoid glucosides, the two highly abundant breakdown products, M₂BOA and 6-MBOA, also accumulated at higher levels in Xi502 than in B73 (Figure 4; Table S3; p < 0.05; Student’s t-test).

3.4. 3-Methyl-2-benzoxazolinone Suppresses FAW Growth In Vitro

In addition to HDMBOA-Glc, which was reported to repel and suppress the growth of FAW larvae [19], four other benzoxazinoid compounds presented at high concentration in Xi502 leaf tissues. To measure the FAW growth in vitro, we were able to commercially purchase a structural analog of 6-MBOA, namely 3-methyl-2-benzoxazolinone (3-MBOA). Supplementation of 3-MBOA to FAW neonates in an artificial diet significantly suppressed their growth after 7 days’ feeding (Figure 5), suggesting that the higher levels of certain benzoxazinoid compounds in Xi502 could explain the enhanced FAW resistance level in this genotype.

4. Discussion

In the battle against the invasive FAW, the enlisting of locally adapted insect-resistant maize cultivars could play a pivotal role in supplementing chemical pesticide application, bio-control methods, and the potential adoption of insect-resistant transgenic cultivars. Insect bioassay experiments in a growth chamber showed that Xi502 is highly resistant to the FAW [28], and the subsequent characterization of the heights of seedling and mature plants as well as yield suggests that this heightened resistance level may come with a significant growth penalty (Figure 1). Yet, it would be premature to conclude there is a growth defense tradeoff on a genetic level, as these two traits may be determined by a number of unlinked genetic loci, and it would still be possible to segregate the strong defense alleles from the slow growth and low yield alleles. Hence, it would be interesting to re-assess the FAW-resistance and the growth-related phenotypes among the F2 individuals derived from a bi-parental cross between B73 and Xi502, and look for possible independent segregation of these traits.

As a preliminary probe into the potential mechanism of enhanced resistance in Xi502, we compared the transcriptomic responses of FAW larvae feeding on this maize cultivar versus the susceptible check B73. Though the larvae feeding of the two maize cultivars or an artificial diet were visually similar after just 24 h of feeding, our results demonstrated that significant transcriptomic re-programing had already occurred (Figure 2a,b). Preferential expressions of cytoskeleton re-organization and chitin metabolism-related genes in B73-fed larvae were indicative of the faster larval growth and development on this susceptible cultivar (Figure 2c; Table S2). Meanwhile, the enrichment of various small molecule metabolic processes among genes specifically suppressed in these larvae also supports such interpretation (Figure 2c; Table S2). Yet, it remains unclear as to whether these transcriptomic signatures of better development are the results of the successful suppression of the defense response in B73, or the cause of such suppression.

Host-induced insect transcriptomic plasticity was speculated to play an important role in the adaptation of insect herbivores toward different host plants [29]. Among the published studies of insect herbivore transcriptomic responses, the alteration appeared to be less dramatic in phloem-feeders than in chewing herbivores such as the FAW [30,31,32,33]. In our experiments, the majority of DEGs seemed to be more relevant to the insect developmental process (Figure 2). It is possible that a comparative transcriptomic analysis more focused on the digestive tissues may be more informative in revealing potential detoxification- and nutrient-assimilation-related genes.

In contrast to results in the recent report that cloned the maize metabolite-detoxifying genes in the FAW, in which Israni et al. (2020) showed significant induction of these genes when moving the larvae from artificial diet onto maize leaves [22], we did not observe such induction of these genes in our transcriptomics data (Figure 3). One possible explanation for this disparity is that we performed RNAseq on whole larvae instead of gut tissues only; hence, the induction signals may have been diluted to below significance threshold. Instead, we found six putative aromatic compound catalysis-related genes to be specifically induced when the larvae were feeding on Xi502 leaves (Figure 3). These genes might also be important for the survival of FAW larvae on the more resistant maize genotypes, which probably have stronger biochemical defense against insect herbivores. The exact function of these genes could be further validated by generating targeted CRISPR knock-out mutant lines, and comparison of the performance of such mutants versus their wildtype progenitors.

In congruence with its stronger FAW resistance phenotype, Xi502 contains higher levels of benzoxazinoid glucosides, as well as their bioactive degradation products, than B73 (Figure 4). These include the known FAW-toxic and repelling compound HDMBOA-Glc. In addition to HDMBOA-Glc, which has been reported to repel and suppress the growth of FAW larvae [16,19,34]), four other benzoxazinoid compounds presented in high concentrations in Xi502 leaf tissues. Among them, we tested another benzoxazinoid compound, 3-MBOA, which was not found in maize tissues, and found it to be toxic against FAW larvae in vitro as well (Figure 5). Interestingly, previous in vitro feeding experiments with 6-MBOA showed no significant impact on FAW larvae growth, and analysis of the frass of the larvae suggested that this compound could be re-glucosylated at the N3 position (Maag et al., 2014). Therefore, it is possible that the methylation at N3 position of 3-MBOA prevents effective detoxification through re-glucosylation by the FAW UDP-glucosyltransferases, rendering this compound toxic against FAW [22]. This hypothesis could also explain the higher toxicity of the N-methoxylated HDMBOA-Glc than its N-hydroxylated precursor DIMBOA-Glc.

In summary, FAW-resistant maize genotype Xi502 provides the basic material for further genetic elucidation of the genetic mechanism of FAW resistance in maize. Comparative transcriptomic analyses of FAW larvae feeding on either Xi502 or the susceptible B73 tissues not only demonstrated accelerated larval development on B73 and compromised growth on Xi502. our transcriptomic study also pinpointed a number of putative aromatic compound breakdown -related genes. These were specifically induced when larvae were feeding on Xi502; hence, they may serve as potential targets for FAW control. Finally, the contrasting levels of benzoxazinoids in the susceptible B73 and the resistant Xi502 suggested that the higher resistance against FAW in Xi502 may be associated with genetic variation in the biosynthetic and/or regulatory loci of maize benzoxazinoid metabolism.