Genetic Variation, DIMBOA Accumulation, and Candidate Gene Identification in Maize Multiple Insect-Resistance

Maize seedlings contain high amounts of 2,4-dihydroxy-7-methoxy-1,4-benzoxazin-3-one (DIMBOA), and the effect of DIMBOA is directly associated with multiple insect-resistance against insect pests such as Asian corn borer and corn leaf aphids. Although numerous genetic loci for multiple insect-resistant traits have been identified, little is known about genetic controls regarding DIMBOA content. In this study, the best linear unbiased prediction (BLUP) values of DIMBOA content in two ecological environments across 310 maize inbred lines were calculated; and their phenotypic data and BLUP values were used for marker-trait association analysis. We identified nine SSRs that were significantly associated with DIMBOA content, which explained 4.30–20.04% of the phenotypic variation. Combined with 47 original genetic loci from previous studies, we detected 19 hot loci and approximately 11 hot loci (in Bin 1.04, Bin 2.00–2.01, Bin 2.03–2.04, Bin 4.00–4.03, Bin 5.03, Bin 5.05–5.07, Bin 8.01–8.03, Bin 8.04–8.05, Bin 8.06, Bin 9.01, and Bin 10.04 regions) supported pleiotropy for their association with two or more insect-resistant traits. Within the 19 hot loci, we identified 49 candidate genes, including 12 controlling DIMBOA biosynthesis, 6 involved in sugar metabolism/homeostasis, 2 regulating peroxidases activity, 21 associated with growth and development [(auxin-upregulated RNAs (SAUR) family member and v-myb avian myeloblastosis viral oncogene homolog (MYB)], and 7 involved in several key enzyme activities (lipoxygenase, cysteine protease, restriction endonuclease, and ubiquitin-conjugating enzyme). The synergy and antagonism interactions among these genes formed the complex defense mechanisms induced by multiple insect pests. Moreover, sufficient genetic variation was reported for DIMBOA performance and SSR markers in the 310 tested maize inbred lines, and 3 highly (DIMBOA content was 402.74–528.88 μg g−1 FW) and 15 moderate (DIMBOA content was 312.92–426.56 μg g−1 FW) insect-resistant genotypes were major enriched in the Reid group. These insect-resistant inbred lines can be used as parents in maize breeding programs to develop new varieties.


Introduction
Maize (Zea mays), is an important agro-economical crop that is utilized globally as food, animal feed, and biofuel products with production of more than 1.14 billion tons in 2018 [1]. Despite the large production area for maize, extensive losses are common in China due to a range of biotic stressors, such as multiple insect injury [1]. Over 350 insect pests have been identified for maize. Of these, Asian corn borer (ACB) (Ostrinia furnacalis; Lepidoptera, Pyralidae) is one of the most destructive insect pests in maize. Newly hatched ACB larvae primarily feed on leaves during the whorl stage; subsequently, the 3rd or 4th instars bore into the stalk. This may cause yield losses of 10 to 30% in the outbreaks recorded in China [2]. In addition, the aboveground parts of maize are susceptible to corn leaf aphid (CLA) (Rhopalosiphum maidis; Homoptera, Aphididae), especially in tropical and warmer temperature regions [3,4]. CLA infestation in maize seedlings slows down plant development, reduces plant height, and decreases grain yield [5]. CLA damage can even occur through the maize tassel in which the accumulation of sticky honeydew can prevent the shedding of pollen, with yield losses of up to 90% [4,6]. Several aphid species can also transmit maize dwarf mosaic virus in a non-persistent manner [7], causing yield losses as high as 70% [8].
Chemical insecticides are widely used to reduce yield losses in maize incurred by insect pests, including ACB and CLA. However, residues from these insecticides can pose a significant health hazard to maize consumers and also cause harm to the environment [9,10]. Furthermore, the overuse of pesticides may also result in the development of chemical resistance to insects and the emergence of secondary pests [9,11]. Fortunately, maize is a genetically diverse crop that exhibits a wide variation in its resistance to multiple insects [9,12,13]; the varieties of maize that show high resistance to insect pests can be used to thwart multiple insect attacks via their various defense mechanisms [9,14]. In maize, the chemical defense metabolite 2,4-dihydroxy-7-methoxy-1,4-benzoxazin-3-one (DIMBOA), a cyclic hydroxamic acid derivative (benzoxazinoids, bxs), is known to confer resistance to leaf-feeding by both ACB [2,15] and CLA [4,16] via inhibitory and toxicity mechanisms. Previous studies reported that the DIMBOA could constitute > 1% of the dry weight of maize seedlings [17] and could be more abundant in adult maize plants after the induction of defense mechanisms [18]. The advances in molecular biology and statistical methods over recent years have led to the identification of multiple quantitative trait loci (QTLs), associated markers, and candidate genes for maize resistance to leaf-feeding by different corn borers and aphids [2,4,12,19,20]. However, only one study reported the involvement of QTLs responsible for DIMBOA content [15], and the low resolution of QTL linkage mapping methodology did not provide a convincing conclusion that genetic variation in maize resistance to multiple insect pests was associated with DIMBOA content and bxs alleles.
Association mapping based on linkage disequilibrium (LD) provides a fine-mapping method that enables researchers to identify functional variation in a broader germplasm background [21]. Both QTL linkage mapping and association mapping methods could be complementary strategies for investigating genetic variation in maize related to DIMBOA content. At present, only a few maize genotypes (Mc37, Mp708, Mp704, D06, CML103, A637, and EP39) are known to be ACB-resistant germplasms [20]; little is known with regard to multiple insect-resistance breeding in maize. Therefore, there is an urgent need to screen genetic resources and characterize the mechanisms associated with multiple insect resistance. The testing of inbred maize lines in terms of their DIMBOA levels in response to ACB and CLA feeding may serve as a potent approach to evaluate and develop new varieties with multiple insect-resistance. Gansu Province in China is one of the core bases for maize seed production and breeding, and a large number of elite inbred lines have been cultivated over the last decade. However, the development of breeding programs for multiple insect-resistance has been slow in this region. For this reason, we collected and investigated 310 elite inbred maize lines with broad genetic backgrounds from different ecological environments in Gansu Province, China. Our objectives were to (i) evaluate the accumulation of DIMBOA in the V6 stage of these different genotypes in different ecological environments and to identify elite multiple insect-resistant parents for use in breeding programs; (ii) explore the genetic diversity and population structure of these maize lines by 186 polymorphic simple repeated sequences (SSRs) in the 10 maize chromosomes, and (iii) identify SSRs that were significantly associated with DIMBOA content and the best linear unbiased prediction (BLUP) values of DIMBOA content in two ecological environments via a general linear model (GLM) and a mixed linear model (MLM). We also detected hot genetic loci and candidate genes associated with multiple insect-resistance by combining our research findings with those of previous studies. These findings have laid a foundation for further multiple insect-resistance marker-assisted selection (MAS) breeding in maize.

DIMBOA Content Variation
The maize-feeding insects begin to damage maize leaves at the V5-V6 seedlings stage in the field [15]. At the V6 stage, sufficient variation was observed for DIMBOA content in 310 maize inbred lines in both Zhangye (E1) and Longxi (E2) ecological environments ( Figure 1A). The average DIMBOA content of 310 inbred lines seedlings in the E1 environment ranged from 8.66 (T58) to 528.88 µg g −1 FW (ZY19-Jiu1101) with an overall mean of 94.21 µg g −1 FW; similarly, the average DIMBOA content of these seedlings in the E2 environment varied from 8.84 (T58) to 493.40 µg g −1 FW (RX20-1006) with an overall mean of 93.26 µg g −1 FW ( Figure 1B). Moreover, factorial analysis of variance also proved significant for the tested variables (genotypes and environments) and their interaction with the DIMBOA content in the seedlings of these maize genotypes ( Figure 1C). These data indicated that the DIMBOA formation and accumulation were controlled by a number of factors, including maize's genetic constitutions, external environments, and their interaction. Therefore, after maize-feeding insects, such as ACB and CLA feed on leaves, DIMBOA can be induced rapidly in maize to confer resistance to these insects, along with other adverse responses, such as anorexia and toxic reactions in these insects.
In addition, the DIMBOA content of these maize materials in the two ecological environments followed a typical partial distribution, as the skewness and kurtosis values of the DIMBOA content were > 1.0 ( Figure 1B). We speculated that this phenomenon may be related to the unequal enrichment of DIMBOA content in the selected maize inbred lines. Further analysis showed that the estimated broad-sense heritability (H 2 B ) and genotype × environment interaction heritability (H 2 GE ) of DIMBOA content in all inbred lines across both ecological environments were 74.64 and 37.32%, respectively ( Figure 1D). The data demonstrated that DIMBOA content was less affected by the environment when compared to other insect-resistant traits, especially the leaf feeding rating of corn borer (LFR) and the number of holes of corn borer (HO), and their H 2 B were 37.5 and 46.4% [2], respectively. Meanwhile, the genetic variation coefficient (CVg) of DIMBOA content among 310 maize inbred lines in both E1 and E2 environments were 108.47 and 106.22%, respectively ( Figure 1B). Thereby, it is necessary to detect the genetic loci responsible for DIMBOA content in maize.

Genetic Diversity and Population Structure Analysis
Genetic diversity and population structure analysis are important tools for germplasm characterization and subsequent utilization in multiple insect-resistance improvements. A total of 186 polymorphic SSRs were used to analyze the distribution of 748 alleles in 310 maize inbred lines. The number of alleles varied from two to eight at a locus. The umc1917 (Bin 1.04, ctg14), umc2314 (Bin 6.01, ctg268) and umc2031 (Bin 8.06, ctg361) exhibited the maximum number of alleles (8 alleles) (Table S1). It is well known that the marker attributes, i.e., the polymorphism information content (PIC) value and Shannon-Wiener's index (I) value are routinely used to evaluate the informativeness of the primers. In this study, the PIC value of each SSR ranged between 0.294 (umc2297, Bin 5.03, ctg220) and 0.826 (umc2314, Bin 6.01, ctg268). Out of 186 SSRs, 69 markers were highly polymorphic, with a PIC ≥ 0.600 (Table S1). Similarly, the average I value was 1.359 and ranged between 0.585 (umc2112, Bin 1.04, ctg21; umc2043, Bin 10.05, ctg415) and 2.410 (umc2314, Bin 6.01, ctg268) (Table S1). These data showed that a PIC value ≥ 0.600 was observed in approximately 40% of the markers, suggesting that these SSRs were very informative and useful in the assessment of genetic diversity, population structure, and marker-DIMBOA content association analysis. icant for the tested variables (genotypes and environments) and their interaction with the DIMBOA content in the seedlings of these maize genotypes ( Figure 1C). These data indicated that the DIMBOA formation and accumulation were controlled by a number of factors, including maize's genetic constitutions, external environments, and their interaction. Therefore, after maize-feeding insects, such as ACB and CLA feed on leaves, DIMBOA can be induced rapidly in maize to confer resistance to these insects, along with other adverse responses, such as anorexia and toxic reactions in these insects.  The statistics including mean ± standard deviation (SD), minimum, maximum, confidence interval (CI), skewness, kurtosis, and genetic variation coefficient (CVg) of DIMBOA content and their BLUP values among all inbred lines in E1 and E2 environments. (C) Frequency of the genotypic variance (σ g 2 ), environmental variance (σ e 2 ), genotype × environment interaction variance (σ ge 2 ), and error variance (σ ε 2 ) for DIMBOA content; *** indicated a significant difference with p < 0.01 (ANOVA). (D) The broad-sense heritability (H 2 B ) and genotype × environment interaction heritability (H 2 GE ) of DIMBOA content.
In addition, the population structure of these inbred lines was analyzed by STRUC-TURE software. When various groups (K value) ranging from 1 to 12 were compared, the ∆K reached the maximum value when K = 5, thus the 310 maize inbred lines were divided into five optimal groups (Figure 2A). The inbred lines with a membership probability of ≥0.500 were assigned to the same groups, and if the inbred lines had a membership probability of less than this value, they were assigned to a mixed group (not assigned to any of the five groups) [22,23]. Of the 310 inbred lines, 294 (accounting for 94.84%) were assigned to either one of the five groups, including the Lüda red cob (LRC; 44 lines) group, Tang si ping tou (TSPT; 81 lines) group, Lancaster (Lan; 52 lines) group, P (47 lines) group, (A) Change curve in the log probability data of ΔK value against K (group number) value. (B) Population structure of 310 inbred lines based on 186 SSR markers at K = 5. Each inbred line was represented by a vertical line, which indicated the membership coefficient for that individual, and the five groups were the Lüda red cob (LRC), Tang si ping tou (TSPT), Lancaster (Lan), P, and Reid groups, respectively.

Evaluation of Multiple Insect-Resistance of Inbred Lines and Dissection of Their Attributive Groups
According to the performance of DIMBOA content among all inbred lines in the two ecological environments, these germplasms were divided into five types (type I to type V  Figure 3C). Furthermore, the three high insect-resistant inbred lines and 15 moderate insect-resistant inbred lines belonged to the Reid (accounted for 50.00%), LRC

Evaluation of Multiple Insect-Resistance of Inbred Lines and Dissection of Their Attributive Groups
According to the performance of DIMBOA content among all inbred lines in the two ecological environments, these germplasms were divided into five types (type I to type V  Figure 3C). Furthermore, the three high insect-resistant inbred lines and 15 moderate insect-resistant inbred lines belonged to the Reid (accounted for 50.00%), LRC (accounted for 22.22%), P (accounted for 16.67%), and Lan (accounted for 11.11%) groups, respectively ( Figure 3B). The data suggested that Reid was an important insect-resistant group, and more attention should be given to the improvement of new insect-resistant varieties using Reid genotypes. (accounted for 22.22%), P (accounted for 16.67%), and Lan (accounted for 11.11%) groups, respectively ( Figure 3B). The data suggested that Reid was an important insect-resistant group, and more attention should be given to the improvement of new insect-resistant varieties using Reid genotypes.

Association Analysis of DIMBOA Content and BLUP Values
In this study, we retrieved the genetic distance (centimorgan, cM) of 186 SSRs on an IBM2 2008 Neighbors map frame (https://www.maizegdb.org/data_center/map (accessed on 18 September 2022)) (Table S1) to build a genetic linkage map, which spanned a total length of 6684.4 cM ( Figure 4). Then, the associated SSR loci of DIMBOA content in both ecological environments and the BLUP values were assessed using Tassel 3.0 software , Lancaster (Lan), P, and Reid group, respectively) in E1 and E2, respectively. (C) Evaluation of multiple insect-resistance among 310 inbred lines in two environments were performed using R package (http://www.R-project.org/; accessed on 20 April 2022) with UPGMA (the K value sets to 5), including type I (high insect-resistant lines), type II (moderate insect-resistant lines), type III (insect-resistant lines), type IV (moderate insect-susceptible lines), and type V (insect-susceptible lines), respectively.

Identification of Hot Genetic Loci and Candidate Genes Associated with Multiple Insect-Resistant Traits
Next, we attempted to obtain hot genetic loci for multiple insect-resistant traits of ACB and CLA to lay a foundation for fine mapping and candidate gene prediction, verification, and breeding application. First, we collected 47 original genetic loci for ACB-/CLA-resistant traits including DIMBOA content, tunnel length of corn borer (TL), aphid incidence rate (AIR), average aphid incidence grade (AIG), LFR, HO, tunnel length/number of holes of corn borer (TL/HO), and aphid resistance (AR) from previous studies ( Table 2), and combined our results (Table 1) to construct a physical map ( Figure 5). Further, 19 hot genetic loci (Loci 1-Loci 19) for multiple insect resistance were identified on 8 chromosomes except chromosomes 3 and 7, resulting in the identification of 49 candidate genes, including 12 controlling DIMBOA biosynthesis, 6 involved in sugar metabolism/homeostasis, 2 participating in peroxidases (POD) activity, 9/12 associated with auxin-upregulated RNAs (SAUR) family member/ v-myb avian myeloblastosis viral oncogene homolog (MYB), and 3/2/1/1 responsible for the lipoxygenase/cysteine protease/restriction endonuclease/ubiquitin conjugating-enzyme ( Figure 5; Tables 3 and S3); these candidate genes may be utilized to determine the multiple insect-resistance to ACB and CLA through their complex interaction network.

Expression Levels of Five Candidate Genes Responsible for DIMBOA Biosynthesis
We randomly selected 5 of the 12 candidate genes that control DIMBOA biosynthesis, including GRMZM2G085381 (bx1), GRMZM6G617209 (bx6), GRMZM2G441753 (bx7), GR-MZM2G311036 (bx10), and GRMZM2G336824 (bx11) to examine their relative expression levels in maize seedlings of RX20-1006 (high insect-resistant line) and T58 (insect-susceptible line) in the E1 environment at the V6 stage using real-time PCR (RT-qPCR). The results showed that the relative expression levels of the five genes were significantly correlated with the DIMBOA content in the two maize genotypes ( Figure 6A-C).

Expression Levels of Five Candidate Genes Responsible for DIMBOA Biosynthesis
We randomly selected 5 of the 12 candidate genes that control DIMBOA biosynthesis, including GRMZM2G085381 (bx1), GRMZM6G617209 (bx6), GRMZM2G441753 (bx7), GRMZM2G311036 (bx10), and GRMZM2G336824 (bx11) to examine their relative expression levels in maize seedlings of RX20-1006 (high insect-resistant line) and T58 (insectsusceptible line) in the E1 environment at the V6 stage using real-time PCR (RT-qPCR). The results showed that the relative expression levels of the five genes were significantly correlated with the DIMBOA content in the two maize genotypes (Figure 6A-C).  The correlational relationships between the relative expression level of five candidate genes and DIMBOA content; lines represent significant correlations (p < 0.05).

Defense Strategies of Maize against ACB-/CLA-Feeding
It is well known that the plethora of ACB and CLA that either simultaneously or concurrently attack multiple maize parts, such as newly hatched ACB feeds on whorl leaves and later instars, tunnel into the stalk or the ear to feed on pith tissues or fresh kernels [25], resulting in a reduction of photosynthetic property, disruption of nutrient and water transport, an increase of stalk lodging and bacterial/fungal infections [25,26], and ultimately complicates harvesting practices and reduces grain yield and quality [27,28]. In addition, CLA is a phloem sap-sucking pest [29], and it can absorb nutrients from phloem sap and alter source-sink patterns [30]; its digestive waste products, i.e., honeydew, can also deposit on the leaf surface of maize and promote mold growth [6]. CLA is also a vector for some plant viral diseases that facilitate pathogen entry and cause maize leaves to curl, discolor, and wilt after serious CLA infestations [6].

Defense Strategies of Maize against ACB-/CLA-Feeding
It is well known that the plethora of ACB and CLA that either simultaneously or concurrently attack multiple maize parts, such as newly hatched ACB feeds on whorl leaves and later instars, tunnel into the stalk or the ear to feed on pith tissues or fresh kernels [25], resulting in a reduction of photosynthetic property, disruption of nutrient and water transport, an increase of stalk lodging and bacterial/fungal infections [25,26], and ultimately complicates harvesting practices and reduces grain yield and quality [27,28]. In addition, CLA is a phloem sap-sucking pest [29], and it can absorb nutrients from phloem sap and alter source-sink patterns [30]; its digestive waste products, i.e., honeydew, can also deposit on the leaf surface of maize and promote mold growth [6]. CLA is also a vector for some plant viral diseases that facilitate pathogen entry and cause maize leaves to curl, discolor, and wilt after serious CLA infestations [6].
In response to insect attack, maize has evolved an extensive array of defense strategies to prevent ACB or CLA feeding and colonization. Increasing evidence has verified that maize-derived compounds bxs, e.g., 2,4-dihydroxy-7-1,4-benzoxazin-3-one glucoside (DIMBOA-Glc), DIMBOA, and 6-methoxy-3h-1,3-benzoxazol-ne (MBOA) are multifunctional defense metabolites that can protect maize against insect pests feeding and pathogens [31][32][33]. DIMBOA acts as a feeding deterrent in maize that can decrease in vivo endoproteinase activity in the larval midgut of the European corn borer, thus limiting the availability of amino acids, reducing larval growth [34], influencing some nervous system and detoxification, and inactivating some hydrolysis enzymes of ACB larvae [35]. Infiltration of DIMBOA into maize leaves stimulated callose accumulation and elevated maize CLA resistance [29]. At 60 h after the 1st instar larvae of Sesamia nonagrwides infestation, the leaves of infested maize were injured, with a significant increase in leaf DIMBOA content of 42-96% [36]. Indeed, these studies indicated that bxs metabolites, especially DIMBOA were involved in maize resistance of corn borer and aphid, which can be a good indicator for screening multiple insect-resistant maize genotypes. Results presented here, together with previous findings [29,34,35], showed that both accumulation and catabolism of DIMBOA varied greatly among maize inbred lines, which may contribute to the resistance of borer and aphid in maize ( Figures 1A-C and 3C). Meihls et al. [16] also reported that DIMBOA content was highly correlated with its precursor DIMBOA-Glc abundance, and these endogenous compounds potentially built up maize's resistance to insect pests. Dafoe et al. [37] further found that jasmonic acid and ethylene were produced rapidly in response to corn borer feeding, and their induction differentially regulated bxs in maize stems; even other phytohormones, i.e., salicylic acid and indole-3-acetic acid, generally considered antagonists of jasmonic acid signaling, were also involved in regulating defense responses [38].

Genetic Loci Comparison between DIMBOA Content and Multiple Insect-Resistant Traits, and Their Hot Loci Identification
Understanding the genetic basis of the multiple insect resistance of maize is critical to the control of combinatorial attacks of ACB and CLA in the field. In this study, we observed a key trait for multiple insect resistance, i.e., the DIMBOA content in two ecological environments and their BLUP values, to detect nine significant associated SSR markers using association mapping via GLM and MLM across 310 diverse maize inbred lines from Gansu Province, China; these nine SSR markers were located in Bin 1.04, Bin 1.11, Bin 2.01, Bin 4.00, Bin 4.01, Bin 6.02, Bin 8.04, and Bin 10.04, respectively ( Figure 4; Table 1). Using the same method, Butrón et al. [15] also identified eight linked single nucleotide polymorphisms (SNP) markers related to DIMBOA content across 281 genetically diverse inbred lines located in Bin 1.04, Bin 2.04, Bin 4.01, Bin 4.04, Bin 5.06, Bin 6.01, Bin 6.05, and Bin 8.06, respectively ( Table 2). They also mapped 12 and 7 QTLs associated with DIMBOA/DIMBOA-Glc/DIMBOA-T content in both B73 × CML322 recombinant inbred lines (RILs) and B73 × IL14H RILs populations, respectively; these QTLs were distributed on chromosomes 1, 3, 4, 6, 7, and 8, respectively [15]. These findings demonstrate that there are multiple genetic loci involved in DIMBOA biosynthesis and decomposition on nine chromosomes except for Chromosome 9. Because of their clear differences in genetic effects and phenotypic variance in these loci, there is a great potential to obtain different multiple insect-resistant lines (materials) based on MAS application. In addition, we further combined the genetic loci from previous foliar studies and our current results of DIMBOA content as well as seven other multiple insect-resistant traits (Tables 1 and 2) to gain a better understanding of hot loci in maize resistance to multiple insect pests and to explore avenues for multiple insect-resistance breeding.
Interestingly, we identified 19 hot loci (Loci 1-Loci 19) involved in maize multiple insect resistance in the present study ( Figure 5; Table 3). Of these, Loci 2 in Bin 1.04 (bnlg147-umc1917 interval) is involved in LFR, HO, AR, and DIMBOA content; Loci 4 in Bin 2.00-2.01 (rs624256-umc2363 interval) controls TL and DIMBOA content; Loci 5 in Bin 2.03-2.04 (rs650025-rs65 interval) is responsible for TL and LFR; Loci 8 in Bin 4.00-4.03 (phi072-umc1509 interval) is related to TL, LFR, AR, AIR, AIG, TL/HO, and DIMBOA content; Loci 12 in Bin 5.03 (phi008-umc1935 interval) is associated with AIG and LFR; Loci 13 in Bin 5.05-5.07 (mmc0081-phi128 interval) is associated with LFR, HO, and DIMBOA content; Loci 15 in Bin 8.01-8.03 (umc1139-umc1627 interval) regulates HO and TL; Loci 16 in Bin 8.04-8.05 (umc1858-bnlg1176 interval) is responsible for TL, LFR, and DIMBOA content; Loci 17 in Bin 8.06 (PZA02746-bnlg1031 interval) is related to TL and DIMBOA content; Loci 18 in Bin 9.01 (phi033-umc1958 interval) is involved in TL, HO, and TL/HO; and Loci 19 in Bin 10.04 (umc1336-umc1054 interval) is associated with LFR, HO, and DIMBOA content. Thus, these 11 hot loci have a pleiotropic effect on 2 to 7 multiple insectresistant traits, and Bin 1.04, Bin 2.00-2.01, Bin 4.00-4.03, Bin 5.05-5.07, Bin 8.04-8.05, Bin 8.06, and Bin 10.04 regions play important roles in conferring DIMBOA accumulation and other aspects of maize multiple-insect resistance to ACB and CLA. Consistent with previous findings, LFR was significantly correlated with HO (phenotypic correlation coefficient (r); r = 0.252) and TL/HO (r = 0.229) in 162 F 3 maize population to ACB resistance [2]. Meanwhile, 3 of 19 hot loci, i.e., Loci 2, Loci 8, and Loci 12 are co-involved in multiple insect resistance to both ACE and CLA; thus, we speculate that the resistance to ACB and CLA for leaf feeding damage is partially controlled by the same mechanisms in these three hot loci. For future research, the contribution of Loci 2, Loci 8, and Loci 12, as well as their functional genes must be examined when developing elite maize varieties with multiple insect resistance to ACE and CLA.

Validation of Candidate Genes in Hot Loci
According to the physical interval of the above 19 hot loci controlling 8 insect-resistant traits and the GO annotations of corresponding genes in these hot intervals, a total of 49 candidate genes were identified (Tables 3 and S3); they may play important roles in maize multiple insect-resistance.
The two candidate genes, GRMZM2G150256 (mir2) and GRMZM2G150276 (mir1) were mapped within Loci 14 and encoded a maize insect resistance-cysteine protease (key defensive protein) against chewing insect pests in maize ( Table 3). The synthetic diet aphid feeding trial bioassays with recombinant mir1-cysteine protease demonstrated that mir1-Cysteine protease triggered direct toxicity to CLA [44], and ethylene acted as a central node in regulating mir1 expression to different feeding guilds of insect herbivores [44].
The MYB transcription factor can interact with mRNA/proteins to form a fine regulatory network to activate the expression of downstream defense genes and induce insect-resistance defense response. Interestingly, the previous findings [19] (Table 3). Thus, these MYBs may be involved in the defense response to ACB and CLA, and further studies are needed to explore the downstream target genes of MYB in herbivore-induced resistance.
Plant POD participates in multiple physiological processes, such as auxin metabolism [51], lignin biosynthesis [52,53], and tolerance against osmotic stress [54]. Additionally, reactive oxygen species (ROS), especially H 2 O 2 induces POD activity, which then oxidizes and polymerizes p-coumaryl/coniferyl-/sinapyl-alcohol into lignin monomers on the cell wall [52]; thereby, POD is involved in the loosening and stiffening of the cell wall during plant development. Moreover, López-Castillo et al. [55] reported that ZmPrx35 as the prevailing POD was involved in defense against pathogens and insects. Similarly, we also identified Zm00001d024752 (POD21) and Zm00001d052335 (POD23) within Loci 19 (Bin 10.04) and Loci 11 (Bin 4.08), respectively (Table 3). Therefore, we speculate that host maize can limit food supplies to multiple insect pests and against larval boring via a key physical barrier, such as cell wall rigidity in the pre-ingestion phase.
In summary, according to the above studies, a possible molecular network underlying maize multiple insect-resistance to ACB and CLA was constructed (Figure 7), which could benefit the development of new maize varieties with multiple insect resistance.
In summary, according to the above studies, a possible molecular network underlying maize multiple insect-resistance to ACB and CLA was constructed (Figure 7), which could benefit the development of new maize varieties with multiple insect resistance.

Germplasms Diversity and Multiple Insect-Resistance Evaluation
When assessing the genetic diversity in maize genotypes, SSR remains the preferred choice due to their co-dominant and multi-allelic nature, abundance, and loci specificity [23]. In the current study, 748 alleles with a range of 2 to 8 per locus were identified among 310 maize inbred lines using 186 polymorphic SSRs (Table S1). The finding showed a wide range of diversity among genotypes [23], which will benefit broadening the genetic base in any breeding program. The average PIC value of all SSR markers was 0.543 (range from 0.294 to 0.826) in our result (Table S1). The data demonstrates the presence of many informative allelic variations in this maize population [22,23]. Moreover, considering the membership probability of ≥ 0.500, the population structure further indicated that the 294 inbred lines were divided into 5 optimal groups (Table S2); this result was consistent with the findings of Liu et al. [22]. Further, we comprehensively evaluated the multiple insect resistance of 310 inbred lines in 2 ecological environments, and screened 3 high and 15 moderate insect-resistant germplasms, which were mainly in the Reid group (accounting for 50.00%) ( Figure 3B-C). These findings have laid a foundation for the improvement of

Germplasms Diversity and Multiple Insect-Resistance Evaluation
When assessing the genetic diversity in maize genotypes, SSR remains the preferred choice due to their co-dominant and multi-allelic nature, abundance, and loci specificity [23]. In the current study, 748 alleles with a range of 2 to 8 per locus were identified among 310 maize inbred lines using 186 polymorphic SSRs (Table S1). The finding showed a wide range of diversity among genotypes [23], which will benefit broadening the genetic base in any breeding program. The average PIC value of all SSR markers was 0.543 (range from 0.294 to 0.826) in our result (Table S1). The data demonstrates the presence of many informative allelic variations in this maize population [22,23]. Moreover, considering the membership probability of ≥0.500, the population structure further indicated that the 294 inbred lines were divided into 5 optimal groups (Table S2); this result was consistent with the findings of Liu et al. [22]. Further, we comprehensively evaluated the multiple insect resistance of 310 inbred lines in 2 ecological environments, and screened 3 high and 15 moderate insect-resistant germplasms, which were mainly in the Reid group (accounting for 50.00%) ( Figure 3B-C). These findings have laid a foundation for the improvement of insect-resistant maize varieties in the future.

Plant Materials
In this study, the collected 310 elite maize inbred lines from different ecological environments (Zhangye, 216 lines; Longxi, 64 lines, Jingtai, 21 lines; Pingliang, nine lines) in Gansu Province, China (Table S2). A total of 20 seeds of these germplasms per row were randomly planted in 4.0 m rows, 0.2 m aisles, and 0.5 m between rows on April 16 in Zhangye (E1; 38.83 • N, 106.93 • E, 1536 m altitude) and in Longxi (E2; 34.97 • N, 104.40 • E, 2074 m altitude), Gansu Province, China in 2020. Before sowing, the soil surface was covered with plastic film (0.08 mm thick, 1.2 m wide). At the V6 stage, the 6th leaf of each inbred line in the E1 and E2 environments was collected, frozen in liquid nitrogen, and stored at -70 • C for subsequent DNA extraction and DIMBOA content assay.

DIMBOA Content Assay
The DIMBOA content was determined using high-performance liquid chromatography (HPLC; Shimadzu LCMS8040 system, Beijing, China). Namely, freeze-dried leaves (0.2 g per sample) were homogenized and weighted into screw-capped 10 mL polypropylene centrifuge tubes, and 5 mL of HPLC grade methanol-methanoic acid solution (0.01%, v/v) was added to each tube. The tubes were rotated and placed in the dark for 12 h and then centrifuged at 12,000 rpm (Centrifuge 5425/5425 R; Eppendorf, Hamburg, Germany) for 20 min at 4 • C. Supernatants (600 µL) were slowly passed the corresponding Millex ® needle filter and transferred into auto-sample vials for analysis by HPLC. Standard DIM-BOA (CAS No.: 15893-52-4) was purchased (Sigma-Aldrich, MA, USA) and was used to optimize the mass spectrometric parameters and fragment spectra.

Genetic Diversity Analysis and Marker-DIMBOA Content Association Mapping
Genomic DNA was extracted from the 6th leaf of 310 inbred lines using the cetyltrimethyl ammonium bromide (CTAB) method [56]. Then, a total of 186 polymorphic SSR markers covering the entire maize genome from the MaizeGDB (http://www.maizegdb.org/ (accessed on 10 January 2020)) were used to perform SSR analysis. Fragments were separated using polyacrylamide gel electrophoresis. The polymorphism information content (PIC) value and Shannon-Wiener's index (I) value were determined as follows [23]: where P i was the i allele frequency. Allele identity was used for an unweighted pairgroup method with arithmetic means (UPGMA) cluster analysis. Population structure [57] was analyzed using STRUCTURE v. 2.3.3 software (http://web.stanford.edu/group/ pritchardlab/structure_software/release_versions/v2.3.4/html/structure.html (accessed on 10 August 2022)) for the assessment of groups and genetic relationships among the 310 maize inbred lines. The project was run with the set parameters of the population admixture model, and the allele frequency correlated. The optimum group number was determined by ∆K value [58]. A linkage map was developed according to the genetic distance (centimorgan, cM) of corresponding SSR markers on an IBM2 2008 Neighbors map frame (https://www.maizegdb.org/data_center/map (accessed on 18 September 2022)) using BioMercator v. 4.2 software (http://www.bioinformatics.org/mqtl/wiki/ (accessed on 18 September 2022)). The linkage disequilibrium (LD) values for r 2 [59] and D [57] between SSR loci on chromosomes were calculated using Tassel 3.0 software (https://tassel.bitbucket.io/ (accessed on 10 August 2022)), following a permutation test of 10,000. The K matrix and marker-DIMBOA content association mapping was completed in Tassel 3.0 software using the genotypic data of 186 polymorphic SSRs and phenotypic data on DIMBOA content for a set of 310 inbred lines. The association analysis was conducted by a GLM with Q matrix (individuals' probability of membership in the population) [60] and an MLM with a K + Q matrix [61]. The associated SSRs for DIMBOA content were filtered out based on the phenotypic variance (R 2 ) of the marker at p < 0.01 level and with the lowest false discovery rate (FDR). Using MaizeGDB (http://www.maizegdb.org/ (accessed on 20 September 2022)) and nucleotide and primer blast tools, the physical locations of associated SSR markers for DIMBOA content were determined on chromosomes.

Hot Genetic Loci and Candidate Genes Detection
Using public databases, i.e., MaizeGDB (http://www.maizegdb.org/ (accessed on 23 September 2022)), NCBI (http://www.ncbi.nlm.nih.gov (accessed on 23 September 2022)), and CNKI (http://www.cnki.net (accessed on 23 September 2022)), we collected information on corresponding QTLs and associated markers for multiple insect-resistant traits from our results and previous studies via QTL analysis, GWAS, and association mapping. The hot genetic loci were the overlapping regions combining multiple genetic loci responsible for multiple insect-resistant traits or single genetic loci that explained the large R 2 (10%) in 10 Mb physical intervals. Further, the physical map of all hot genetic loci and candidate genes involving multiple insect-resistant traits in these hot genetic loci regions was developed by BioMercator v. 4.2 software (http://www.bioinformatics.org/mqtl/ wiki/ (accessed on 28 September 2022)) [62]. The functional annotation of corresponding candidate genes was performed using the tool AgBase v2.00 (https://agbase.arizona.edu/ (accessed on 6 October2022)) [63].

RT-qPCR Analysis
Among 310 maize inbred lines, we selected the two inbred lines with the largest difference in multiple insect resistance, i.e., RX20-1006 with the highest DIMBOA content and T58 with the lowest DIMBOA content. Their total RNAs were extracted from the 6th leaf with TRIZOL reagent (Invitrogen, USA), and cDNA was made using a kit (Starscript II First-stand cDNA Synthesis Mix With gDNA Remover, GenStar, Beijing, China), according to the manufacturer's instructions. RT-PCR was conducted using TransStart Tip Green qPCR SuperMix (Tran, Beijing, China). The primers for five candidate genes [12,64] were designed using Primer Premier 5.0 software (http://www.premierbiosoft.com/ (accessed on 9 October 2022)) ( Table S4). Relative gene expression levels were assessed by the 2 −∆∆Ct method, with GRMZM2G126010 as an internal reference gene [63].

Statistical Analysis
The average DIMBOA content in 310 inbred lines from five biological replicates in each ecological environment were analyzed, respectively. A mixed linear model was fitted using the lmer function in lme4 package of R (http://www.R-project.org/ (accessed on 16 December 2021)) to calculate the BLUP of DIMBOA content values [65]. These data were then compared statistically using IBM-SPSS Statistics v. 19.0 (SPSS Inc., Chicago, IL, USA; http://www.ibm.com/products/spss-statistics (accessed on 16 December 2021)). The significance of the total and residual variances of DIMBOA content in 310 inbred lines under both ecological environments was estimated by a GLM for univariate data and by one-way analysis of variance (ANOVA). The broad-sense heritability (H 2 B ) and genotype × environment interaction heritability (H 2 GE ) values for DIMBOA content under both environments were estimated as follows [53,66,67]: H B 2 = σ g 2 /(σ g 2 + σ ge 2 /n + σ ε 2 /nr), H GE 2 = (σ g 2 /n)/(σ g 2 + σ ge 2 /n + σ ε 2 /nr) (4) where σ g 2 was the genotypic variance, σ e 2 was the environmental variance, σ ε 2 was the error variance, σ ge 2 was the variance of genotype × environment interaction, n was the number of ecological environments (n = 2), and r was the number of replications (r = 5). The CVg [63] of DIMBOA content among all inbred lines under each environment was calculated as follows: where X was the average value of DIMBOA content among all inbred lines in each environment, and δ was the standard deviation.

Conclusions
In summary, the DIMBOA confers significant resistance to ACB and CLA. In this study, SSR analysis revealed a wide genetic diversity in the 310 tested maize inbred lines from four type regions of China's Gansu Province, which is the largest maize seed production and breeding area in China. Population structure indicated that 294 inbred lines were successfully assigned to one or another group at a membership probability of ≥0.500. DIMBOA performance evaluation screened out 3 high and 15 moderate insect-resistant inbred lines, which can be used as parents in breeding programs to develop new maize varieties with multiple insect resistance. Using linkage mapping, we detected nine significant SSRs associated with DIMBOA content in both environments. We then combined the 47 original genetic loci for 8 multiple insect-resistant traits from previous studies to detect 19 hot loci. Among them, 11 hot loci were located in Bin 1.04, Bin 2.00-2.01, Bin 2.03-2.04, Bin 4.00-4.03, Bin 5.03, Bin 5.05-5.07, Bin 8.01-8.03, Bin 8.04-8.05, Bin 8.06, Bin 9.01, and Bin 10.04 regions, and they supported pleiotropy for their association with two or more insect-resistant traits. Further, the 49 candidate genes involved in DIMBOA biosynthesis, sugar metabolism/homeostasis, and other multiple insect-resistant defense mechanisms in maize were identified in all 19 hot loci, and their highly interconnected network may form complex maize, multiple insect, pest-induced defense mechanisms.