Abstract

   Plants recognize and respond to feeding by herbivorous insects by
   upregulating their local and systemic defenses. While defense induction
   by aboveground herbivores has been well studied, far less is known
   about local and systemic defense responses against attacks by
   belowground herbivores. Here, we investigated and compared the
   responses of the maize transcriptome to belowground and aboveground
   mechanical damage and infestation by two well‐adapted herbivores: the
   soil‐dwelling western corn rootworm Diabrotica virgifera virgifera
   (Coleoptera: Chrysomelidae) and the leaf‐chewing fall armyworm
   Spodoptera frugiperda (Lepidoptera: Noctuidae). In responses to both
   herbivores, maize plants were found to alter local transcription of
   genes involved in phytohormone signaling, primary and secondary
   metabolism. Induction by real herbivore damage was considerably
   stronger and modified the expression of more genes than mechanical
   damage. Feeding by the corn rootworm had a strong impact on the shoot
   transcriptome, including the activation of genes involved in defense
   and development. By contrast, feeding by the fall armyworm induced only
   few transcriptional changes in the roots. In conclusion, feeding by a
   leaf chewer and a root feeder differentially affects the local and
   systemic defense of maize plants. Besides revealing clear differences
   in how maize plants respond to feeding by these specialized herbivores,
   this study reveals several novel genes that may play key roles in
   plant–insect interactions and thus sets the stage for in depth research
   into the mechanism that can be exploited for improved crop protection.

Significance statement

   Extensive transcriptomic analyses revealed a clear distinction between
   the gene expression profiles in maize plants upon shoot and root
   attack, locally as well as distantly from the attacked tissue. This
   provides detailed insights into the specificity of orchestrated plant
   defense responses, and the dataset offers a molecular resource for
   further genetic studies on maize resistance to herbivores and paves the
   way for novel strategies to enhance maize resistance to pests.

   Keywords: belowground and aboveground defense interactions, Diabrotica
   virgifera , phytohormones, Spodoptera frugiperda , transcriptome, Zea
   mays

1. INTRODUCTION

   Plants have evolved constitutive and inducible defense mechanisms to
   protect themselves from the constant attack by root and shoot
   herbivores (Erb, Glauser, et al., [32]2012; Johnson et al., [33]2016;
   Mithöfer & Boland, [34]2012). Inducible defenses start with the
   recognition of herbivore‐associated molecular patterns (HAMPs) and are
   followed by the activation of signaling networks. Previous studies have
   highlighted the roles of Ca^2+ ion flux, mitogen‐activated protein
   kinase (MAPK) cascades, reactive oxygen species (ROS), and phytohormone
   signaling pathways including jasmonic acid (JA), salicylic acid (SA),
   abscisic acid (ABA), and ethylene (ET) on the expression regulation of
   defense‐related genes, which result in the production of defensive
   compounds (Broekgaarden et al., [35]2015; Erb & Reymond, [36]2019;
   Schuman & Baldwin, [37]2016; Wu & Baldwin, [38]2010). The production of
   defensive secondary metabolites or proteins in plants is referred to as
   direct defense (Erb & Reymond, [39]2019). In addition, plants can
   defend indirectly by emitting herbivore‐induced volatiles that attract
   natural enemies of the herbivores (Turlings & Erb, [40]2018) or
   producing resources for “bodyguards” such as extrafloral nectar
   (Heil, [41]2015). Well‐adapted herbivores may produce effectors that
   suppress plant defenses (Mutti et al., [42]2008; Ye et al., [43]2017)
   and even exploit plant defensive metabolites as foraging cues (Humphrey
   et al., [44]2016; Köhler et al., [45]2015; Machado et al., [46]2021;
   Miles et al., [47]2005; Renwick & Lopez, [48]1999) and/or sequester
   them for their own protection (Kos et al., [49]2011; Kumar et
   al., [50]2014; Robert et al., [51]2012; Singer et al., [52]2009;
   Smilanich et al., [53]2009; Sternberg et al., [54]2012).

   Shoot herbivory induces defenses in both leaves and roots. For example,
   larval performance of western corn rootworm Diabrotica virgifera
   virgifera is attenuated by previous leaf herbivory by fall armyworm
   Spodoptera frugiperda caterpillars (Erb, Robert, et al., [55]2011).
   Similarly, leaf attack by diamondback moth caterpillars Plutella
   xylostella strongly reduces the performance of cabbage root fly larvae
   Delia radicum feeding on roots of cabbage plants Brassica oleracea
   (Karssemeijer et al., [56]2020). In maize, aboveground herbivory by
   cotton leafworm Spodoptera littoralis does not induce JA in roots (Erb,
   Flors, et al., [57]2009). By contrast, an increase in jasmonate levels
   has been observed in roots of tobacco plants 2 h after leaves were
   mechanically damaged and oral secretion (OS) from tobacco hornworm
   Manduca sexta was added to the wounds (Machado et al., [58]2018).
   Cabbage plants also increase JA in roots in response to aboveground
   herbivory by caterpillars, but not by aphids (Karssemeijer et
   al., [59]2020). These plant‐mediated interactions can lead to defense
   facilitation but also suppression, like in tallow trees, where
   different aboveground herbivores induce diverse defensive responses,
   including the differential synthesis of metabolites in roots (Huang et
   al., [60]2014; Xiao et al., [61]2019). Aboveground adults of the tallow
   tree specialist beetle Bikasha collaris thus facilitate development of
   conspecific belowground larvae, but heterospecific aboveground
   herbivory may inhibit B. collaris larval development (Huang et
   al., [62]2014). Thus, the induction of root defenses by shoot herbivory
   can be highly plant‐ and herbivore‐specific.

   Compared with the well‐studied inducible defense mechanism aboveground,
   less is known about the belowground defense of plants against root
   herbivores (Erb, Glauser, et al., [63]2012). As in shoots, the
   responses of plant roots to herbivore attack are insect‐specific
   (Rasmann & Turlings, [64]2008) and different from artificial damage (Lu
   et al., [65]2015). JA is the most important phytohormone that mediates
   plant defense against chewing herbivores (Erb, Meldau, et
   al., [66]2012; Howe & Jander, [67]2008; Wu & Baldwin, [68]2010) and is
   involved in the activation of both local and systemic defenses (Bozorov
   et al., [69]2017; Lortzing & Steppuhn, [70]2016). However, the
   regulation of the JA pathway differs significantly between roots and
   shoot (Acosta et al., [71]2013). Belowground and aboveground herbivore
   attack induces the jasmonate production both in roots (Erb, Flors, et
   al., [72]2009; Lu et al., [73]2015) and shoots (Erb, Flors, et
   al., [74]2009; Erb, Meldau, et al., [75]2012; Wu & Baldwin, [76]2010),
   but jasmonates are less inducible in the roots than in the leaves in
   response to herbivory and mechanical wounding (Erb, Flors, et
   al., [77]2009; Hasegawa et al., [78]2011; Tretner et al., [79]2008). In
   contrast to herbivore‐attacked leaves, there is, at least so far, no
   strong evidence for a role of SA, ABA and ethylene in defenses against
   root herbivory (Erb, Flors, et al., [80]2009; Johnson et al., [81]2016;
   Lu et al., [82]2015), and nothing is known about the involvement of
   other phytohormones in root defense mechanisms. A notable recent study
   shows that root herbivory by D. radicum changes the expression of ABA
   and ethylene biosynthesis genes in cabbage roots after 24 h, suggesting
   the potential role of these phytohormones in later stages of the
   defense response (Karssemeijer et al., [83]2020). Root herbivory not
   only induces reconfiguration of primary metabolites in roots (Lu et
   al., [84]2015; Pan et al., [85]2020), but it also activates systemic
   physiological changes aboveground. For instance, plants infested with
   root herbivores reallocated carbon (Robert et al., [86]2014) and
   nitrogen (Tao & Hunter, [87]2013) to the shoots. Belowground herbivory
   by D. v. virgifera induces water stress, resulting in the accumulation
   of ABA in maize shoots, and enhanced resistance against chewing leaf
   herbivores (Erb, Köllner, et al., [88]2011). Over all, there are still
   large gaps in our understanding of the mechanism of
   root‐herbivory‐induced shoot defense.

   In response to herbivore attack, maize plants accumulate defense
   proteins and toxic secondary metabolites. For example, the
   transcription level of defense‐related genes coding for maize
   proteinase inhibitor (MPI), cystatin‐like proteinase inhibitor, and
   serine protease inhibitor is induced by S. littoralis infestation (Ton
   et al., [89]2007). MPI inhibits the activity of digestive enzymes in
   the gut of S. littoralis (Tamayo et al., [90]2000). Benzoxazinoids
   (BXs), a major group of indole‐derived secondary metabolites, have a
   well‐established role in defense against herbivory in maize (Frey et
   al., [91]2009). BX biosynthesis pathway and enzymes that function in
   the BX production are comprehensively documented (Frey et
   al., [92]2009; Tzin et al., [93]2017). In maize leaves, the content of
   BXs and transcript levels of BX biosynthetic genes are highly induced
   locally in response to caterpillar feeding. BXs such as DIMBOA and
   HDMBOA are toxic and repellent to S. littoralis, respectively (Glauser
   et al., [94]2011). Moreover, Spodoptera exigua and S. littoralis
   caterpillars perform considerably better on maize BX‐deficient mutants
   (Maag et al., [95]2016; Tzin et al., [96]2017). The larger amounts of
   BXs in maize crown roots compared with primary roots play a role in
   deterring feeding by generalist herbivores (Robert et al., [97]2012).
   In contrary, well‐adapted herbivores such as D. v. virgifera and S.
   frugiperda have been shown to tolerate high concentrations of
   benzoxazinoids and use them as foraging cues (Köhler et al., [98]2015;
   Robert et al., [99]2012).

   In addition to non‐volatile defense metabolites, maize plants also emit
   blends of volatile organic compounds (VOCs) that can act as repellents
   of the herbivores (Bernklau et al., [100]2016), foraging cues to
   natural enemies of the pests (Dicke & Sabelis, [101]1988; Rasmann et
   al., [102]2005; Tamiru et al., [103]2011; Turlings et al., [104]1990),
   or airborne signals in systemic defense and plant–plant communication
   (Engelberth et al., [105]2004; Erb et al., [106]2015; Ton et
   al., [107]2007). Volatile indole, for instance, has been shown to prime
   defenses in maize plants (Erb et al., [108]2015). As for direct defense
   responses, the molecular mechanisms that are involved in this
   multifunctional volatile signaling remain to be elucidated.

   While considerable information about aboveground and belowground
   defense responses to herbivory is available, few studies so far have
   directly compared transcriptional responses of roots and shoots in
   response to damage and herbivore attack. To fill this knowledge gap, we
   characterized the local and systemic transcriptional changes of maize
   responses to belowground infestation by D. v. virgifera larvae and
   aboveground herbivory by S. frugiperda caterpillars and compared them
   with mechanical damage. The resulting dataset provides extensive
   insights into the specificity and orchestration of root and shoot
   defense responses to herbivore attack.

2. RESULTS

2.1. Overview of transcriptional changes in maize plants in response to
belowground and aboveground insect herbivory

   To investigate the global transcriptomic changes that occur in response
   to aboveground and belowground insect herbivory, maize plants (var.
   Delprim) were either infested for 72 h by root feeding D. v. virgifera
   larvae, leaf feeding S. frugiperda larvae, or damaged mechanically on
   roots or shoots. The expression levels of eight selected genes were
   confirmed by qRT‐PCR to validate the RNA‐seq results. Similar
   expression patterns and high correlation coefficients of qRT‐PCR and
   FPKM data (Figure [109]S1) confirmed the reliability of the RNA‐seq
   data. Detailed information on RNA sequencing and mapping is provided in
   Table [110]S1.

   Of 46,430 predicted genes in the B73 V4 reference genome, a total of
   37,997 detectable corresponding transcripts could be identified across
   all samples (Data [111]S1). Principal component analyses (PCA) revealed
   that in the shoots, the gene expression profiles of control plants were
   clearly separated from S. frugiperda‐infested, leaf wounded and root
   wounded plants, but overlapping with those of D. v. virgifera‐infested
   plants (Figure [112]1a). Principal component analyses (PCA) of the root
   data show that the gene expression profiles in control root samples
   were separated from D. v. virgifera‐infested and root wounded plants,
   but not from shoot wounded and S. frugiperda‐infested plants
   (Figure [113]1b). Thus, it appears that local responses are generally
   more pronounced than systemic responses, and herbivory elicits specific
   regulation patterns relative to mechanical wounding.

FIGURE 1.

   FIGURE 1
   [114]Open in a new tab

   Overview of maize transcriptome responses to belowground and
   aboveground insect herbivory. (a and b) PCA plots of transcripts
   identified by RNA‐seq of maize shoot (a) and roots (b) from seedlings
   after 72 h of belowground infestation by Diabrotica virgifera virgifera
   (DV) or aboveground infestation by Spodoptera frugiperda (SF), or after
   application of mechanical root (MR) or shoot damage (MS). Non‐treated
   seedlings served as controls (C). (c and d) Total number of transcripts
   that were significantly upregulated or downregulated in maize shoot (c)
   and roots (d) after each treatment compared with non‐manipulated
   controls. (e and f) Venn diagrams illustrating the number of
   transcripts upregulated or downregulated in shoot (e) and roots (f) in
   response to belowground and aboveground treatments

   Genes with a false discovery rate (FDR) adjusted P < .05 and an
   absolute value of log[2]‐transformed fold change
   (treatment/control) > 1 were selected as differentially expressed genes
   (DEGs) for further analysis. Shoot samples from D. v.
   virgifera‐infested, root‐artificially damaged, S. frugiperda‐infested
   and leaf‐artificially damaged plants exhibited 405 (388 up and 17
   down), 1069 (596 up and 473 down), 2438 (1518 up and 920 down), and
   1412 (811 up and 601 down) DEGs, respectively (Figure [115]1c, Data
   [116]S2, [117]S3, [118]S4, and [119]S5). Root samples from D. v.
   virgifera‐infested and root‐mechanically wounded plants exhibited 1266
   (970 up and 296 down) and 4362 (2035 up and 2327 down) DEGs,
   respectively, whereas S. frugiperda‐infestation and leaf‐mechanical
   wounding induced only 264 (159 up and 105 down) and 56 (32 up and 24
   down) DEGs in root samples, respectively (Figure [120]1d, Data [121]S2,
   [122]S3, [123]S4 and [124]S5). Compared with control plants, both
   belowground and aboveground insect herbivory induced local
   transcriptional changes, with systemic changes being less pronounced
   (Figure [125]1c,d). Local mechanical damage also elicited local
   responses and weaker systemic responses. Interestingly, leaf herbivore
   attack triggered stronger responses than mechanical shoot damage, while
   the opposite was the case for root herbivore attack and mechanical root
   damage, where the damage treatment led to stronger responses
   (Figure [126]1c,d). The distribution of upregulated and downregulated
   DEGs in maize shoots and roots in response to each treatment was
   calculated and presented in Venn diagrams (Figure [127]1e,f). In maize
   leaves, the expression of a small number of genes was regulated (62 up
   and 6 down) by all treatments. Two sets of genes were specifically
   regulated by aboveground S. frugiperda herbivory (823 up and 502 down)
   and belowground D. v. virgifera infestation (203 up and 2 down)
   (Figure [128]1e). In the roots, a total of 117 genes (69 up and 48
   down) were specifically regulated by S. frugiperda herbivory, and 646
   genes (409 up and 237 down) were specifically regulated by D. v.
   virgifera infestation (Figure [129]1f). Thus, both local and systemic
   responses are highly specific.

2.2. Differential expression of genes in plants attacked by S. frugiperda

   The DEGs of maize transcriptome in response to aboveground and
   belowground herbivory were further subjected to KEGG pathway enrichment
   analysis to identify pathways that are differentially regulated. The
   DEGs in maize shoots that responded to S. frugiperda attack were
   assigned to 42 significant KEGG pathways (adjusted P < .05) (Data
   [130]S6), of which the top 20 enriched pathways are presented in
   Figure [131]S2A (global and overview maps pathways were excluded). The
   biosynthesis of phenylpropanoids, flavonoids, and benzoxazinoids and
   the metabolism of α‐linolenic acid, as well as other metabolic pathways
   associated with plant defense, signal transduction, and primary
   metabolism, showed strong changes in maize shoots after S. frugiperda
   herbivory (Figure [132]S2A and Data [133]S6). When comparing S.
   frugiperda herbivory and artificial shoot damage, the DEGs are mainly
   involved in energy metabolism, such as the biosynthesis of
   carbohydrates, lipids, and amino acids. Several DEGs are also involved
   in the biosynthesis of certain secondary metabolites and the
   transduction of plant hormone signals (Figure [134]S2B and Data
   [135]S6). DEGs in maize roots that responded to S. frugiperda attack
   were assigned to 11 significant pathways, mainly involving the
   biosynthesis of phenylpropanoid and flavonoids, and some primary
   metabolism pathways including the metabolism of amino acids, nitrogen,
   and carbohydrates (Figure [136]S2C and Data [137]S6). Notably, shoot
   and root responses to S. frugiperda attack comprised the biosynthesis
   of phenylpropanoids, flavonoids, and benzoxazinoids as well as
   phenylalanine, tyrosine, and tryptophan (Figure [138]S2A,C and Data
   [139]S6), implying the potential role of these pathways in general
   systemic stress responses to herbivory. Figure [140]S3 and Data [141]S7
   provide detailed information on the 60 most upregulated genes and 60
   most downregulated genes in the shoot and in response to S. frugiperda
   herbivory.

2.3. Differential expression of genes in plants attacked by D. v. virgifera

   The DEGs in maize roots that responded to D. v. virgifera attack were
   assigned to 52 significant KEGG pathways (adjusted P < .05) (Data
   [142]S6), and the top 20 enriched pathways are presented in
   Figure [143]S4A (global and overview maps pathways were excluded). D.
   v. virgifera herbivory strongly induced the pathways involved in the
   metabolism of phenylpropanoid, α‐linolenic acid, and monoterpenoids, as
   well as primary pathways involved in the metabolism of amino acids,
   lipids, and carbohydrates (Figure [144]S4A and Data [145]S6). Most DEGs
   associated with the biosynthesis of jasmonic acid and methyl jasmonate
   in the α‐linolenic acid metabolism pathway were upregulated in response
   to D. v. virgifera infestation. Of the plant hormone signal
   transduction pathways, genes associated with JA signaling transduction
   (JASMONATE ZIM‐domain [JAZ] and MYC2) and genes responsible for disease
   resistance via SA signaling (transcription factor TGA and
   pathogenesis‐related protein 1 gene PR1) were upregulated by D. v.
   virgifera infestation (Data [146]S2 and Data [147]S6). These results
   suggest that both JA and SA signaling are involved in the defense
   responses of maize roots to D. v. virgifera. When comparing D. v.
   virgifera herbivory and artificial root damage, the DEGs are mainly
   those involved in phenylpropanoid biosynthesis, plant hormone signal
   transduction, plant–pathogen interaction, genetic information
   processing, and cellular processes. Several DEGs are also linked to
   primary metabolism pathways such as the metabolism of carbohydrates,
   amino acids, and lipids (Figure [148]S4B and Data [149]S6). DEGs in
   maize shoot, when comparing D. v. virgifera herbivory and the control
   treatment, were assigned to 13 relevant pathways involved in DNA
   replication, linoleic acid metabolism, translation, carotenoid
   biosynthesis, and other metabolisms of energy, carbohydrate,
   nucleotide, and amino acids (Figure [150]S4C and Data [151]S6). All
   DEGs involved in DNA replication were upregulated in shoot tissue in
   response to belowground D. v. virgifera herbivory, whereas DEGs in
   translation and carbon fixation were downregulated (Data [152]S2 and
   Data [153]S6). Figures [154]S5 and Data [155]S8 provide detailed
   information on the 60 most upregulated genes and 60 most downregulated
   genes in the root in response to D. v. virgifera feeding.

2.4. Plant hormone‐related genes induced by belowground and aboveground
insect herbivory

   To determine phytohormone‐related gene expression changes in response
   to belowground and aboveground insect infestation, we compared the
   expression of genes associated with JA, SA, ABA, and ethylene
   biosynthesis in maize shoot and roots for the five plant treatments
   (Data [156]S9). In general, the expression pattern of genes involved in
   JA (Figure [157]2), SA (Figure [158]3), ABA (Figure [159]4), and
   ethylene pathway (Figure [160]5) were highly induced locally in
   response to belowground and aboveground infestation or artificial
   damage, whereas root and shoot damage by insect herbivory and
   mechanical wounding also systemically induced the strong expression of
   ABA‐related genes (Figure [161]4).

FIGURE 2.

   FIGURE 2
   [162]Open in a new tab

   Effects of belowground and aboveground insect herbivory on jasmonic
   acid (JA) pathway gene expression. (a) Schematic diagram of the JA
   biosynthesis pathway. LOX, lipoxygenase; AOS, allene oxide synthase;
   AOC, allene oxide cyclase; OPR, 12‐oxophytodienoate reductase; JAR,
   jasmonate resistant; JA‐Ile, jasmonoyl‐isoleucine. The dashed arrow
   represents multiple enzymatic steps (Tzin et al., [163]2015). (b and c)
   Heat map of JA biosynthesis‐related gene expression in maize shoot (b)
   and roots (c). Samples were collected from maize plants that were kept
   non‐manipulated (C, control) or after 72 h of belowground infestation
   by Diabrotica virgifera virgifera (DV), mechanical damage on root (MR),
   72 h of aboveground infestation by Spodoptera frugiperda (SF), or
   mechanical damage on shoot (MS). Color coding represents the range of
   log[2](fold change relative to control).

FIGURE 3.

   FIGURE 3
   [164]Open in a new tab

   Effects of belowground and aboveground insect herbivory on salicylic
   acid (SA) pathway gene expression. (a) Schematic diagram of the SA
   biosynthesis pathway. ICS, isochorismate synthase; PAL, phenylalanine
   ammonia lyase; EDS, enhanced disease susceptibility; isochorismate is
   transported by the multidrug and toxin extrusion (MATE) transporter EDS
   to the cytosol. PBS3, avrPphB susceptible 3; EPS1, enhanced pseudomonas
   susceptibility 1. (b and c) Heat map of SA biosynthesis‐related gene
   expression in maize shoot (b) and roots (c). Samples were collected
   from maize plants that were kept non‐manipulated (C, control) or after
   72 h of belowground infestation by Diabrotica virgifera virgifera (DV),
   mechanical damage on root (MR), 72 h of aboveground infestation by
   Spodoptera frugiperda (SF), or mechanical damage on shoot (MS). Color
   coding represents the range of log[2](fold change relative to control).

FIGURE 4.

   FIGURE 4
   [165]Open in a new tab

   Effects of belowground and aboveground insect herbivory on abscisic
   acid (ABA) pathway gene expression. (a) Schematic diagram of the ABA
   biosynthesis pathway. ZEP, zeaxanthin epoxidase; NCED,
   9‐cis‐epoxycarotenoid dioxygenase; SDR, short‐chain
   dehydrogenase/reductase; AO, aldehydeoxidase (Leng et al., [166]2014).
   (b and c) Heat map of ABA biosynthesis‐related gene expression in maize
   shoot (b) and roots (c). VP14, viviparous14, 9‐cis‐epoxycarotenoid
   dioxygenase 1. Samples were collected from maize plants that were kept
   non‐manipulated (C, control) or after 72 h of belowground infestation
   by Diabrotica virgifera virgifera (DV), mechanical damage on root (MR),
   72 h of aboveground infestation by Spodoptera frugiperda (SF), or
   mechanical damage on shoot (MS). Color coding represents the range of
   log[2](fold change relative to control).

FIGURE 5.

   FIGURE 5
   [167]Open in a new tab

   Effects of belowground and aboveground insect herbivory on ethylene
   pathway gene expression. (a) Schematic diagram of the ethylene
   signaling pathway. ER, endoplasmic reticulum; ACS,
   1‐aminocyclopropane‐1‐carboxylate synthase; ACO,
   1‐aminocyclopropane‐1‐carboxylate oxidase; ETR, ethylene receptor;
   EIN2, ethylene insensitive 2 (Tamaoki, [168]2008). (b and c) Heat map
   of ethylene signaling pathway‐related gene expression in maize shoot
   (b) and roots (c). Samples were collected from maize plants that were
   kept non‐manipulated (C, control) or after 72 h of belowground
   infestation by Diabrotica virgifera virgifera (DV), mechanical damage
   on root (MR), 72 h of aboveground infestation by Spodoptera frugiperda
   (SF), or mechanical damage on shoot (MS). Color coding represents the
   range of log[2](fold change relative to control).

   Infestation of maize shoots by S. frugiperda induced the expression of
   JA‐related genes in shoot tissue to a greater extent than artificial
   leaf damage, especially in the first and second steps of JA
   biosynthesis. Among all six 13‐lipoxygenase genes (LOX7, LOX8, LOX9,
   LOX10, LOX11, and LOX13) that enable the production of
   12‐oxo‐phytodienoic acid (12‐OPDA) and its downstream JA synthesis
   (Figure [169]2a), only LOX10 and LOX11 were highly induced by S.
   frugiperda feeding (Figure [170]2b). In contrast, six 9‐LOX genes
   (LOX1, LOX2, LOX3, LOX4, LOX5, and LOX6) that serve in the production
   of 10‐oxo‐11‐phytodienoic acid (10‐OPDA, positional isomer of 12‐OPDA)
   and 10‐oxo‐11‐phytoenoic acid (10‐OPEA) were all highly induced after
   S. frugiperda infestation. Overall, the expression of 9‐lipoxygenases
   was induced to higher levels than 13‐lipoxygenases in shoot tissue in
   response to S. frugiperda feeding. In addition, all the transcripts of
   allene oxide synthase (AOS), allene oxide cyclase (AOC),
   oxo‐phytodienoate reductase (OPR), and jasmonate resistant (JAR) were
   upregulated upon S. frugiperda herbivory (Figure [171]2b). Belowground
   infestation by D. v. virgifera induced the expression of one 13‐LOX
   gene (LOX10) and six 9‐LOX genes (LOX1, LOX2, LOX3, LOX4, LOX5, and
   LOX6) and repressed the expression of LOX7 and LOX12 in maize roots
   (Figure [172]2c). Most of the transcripts of AOS, AOC, JAR, and,
   especially, OPR were upregulated in roots after D. v. virgifera
   infestation, while aboveground infestation by S. frugiperda barely
   modified the expression of JA‐related genes in roots (Figure [173]2c).

   The biosynthesis of SA in plants is regulated by the isochorismate
   synthase (ICS) and phenylalanine ammonia‐lyase (PAL) pathways
   (Figure [174]3a). Between the two distinct pathways, only the
   expression of genes involved in the PAL pathway was clearly upregulated
   in shoots after S. frugiperda feeding (PAL4, PAL5, PAL6, PAL7, and
   PAL8; Figure [175]3b) or in roots after D. v. virgifera infestation
   (PAL4, PAL7, and PAL8; Figure [176]3c).

   Several genes involved in ABA biosynthesis (ZEP, zeaxanthin epoxidase;
   NCED, 9‐cis‐epoxycarotenoid dioxygenase; SDR, short‐chain
   dehydrogenase/reductase; AO, aldehydeoxidase; Figure [177]4a) were
   upregulated in shoots after S. frugiperda herbivory, and the expression
   of ZEPc3, ZEP1, NCED, and AO was higher in S. frugiperda‐infested
   shoots compared with artificially damaged shoots. Moreover, the
   expression of SDR in shoots was also induced by belowground herbivore
   or artificial damage (Figure [178]4b). In roots, D. v. virgifera
   infestation highly induced the transcription of ZEPc2, ZEP1, and NCED,
   while mechanical damage in roots induced the transcription of ZEPc1,
   ZEP1, and SDR (Figure [179]4c).

   S. frugiperda herbivory but not artificial damage induced genes
   involved in ethylene biosynthesis in maize shoot, but repressed the
   expression of ethylene insensitive 2 (EIN2), the central transducer of
   ethylene signal (Figure [180]5a,b). The expression of two
   1‐aminocyclopropane‐1‐carboxylate oxidase (ACO) genes involved in
   ethylene synthesis was highly upregulated in roots after D. v.
   virgifera infestation, whereas the transcription of several ethylene
   biosynthesis genes was highly induced in response to artificial damage
   in roots (Figure [181]5c).

2.5. Benzoxazinoid biosynthesis‐related genes induced by belowground and
aboveground insect herbivory

   We compared the expression of several genes associated with
   benzoxazinoid biosynthesis (Figure [182]6a). Compared with artificial
   leaf damage, S. frugiperda feeding highly induced all genes involved in
   BX biosynthesis except BX1‐igl1 (indole glycerol phosphate lyase) in
   shoot tissue. This was particularly the case for BX1‐igl2, which is
   potentially involved in indole production and several BX genes that are
   required for the synthesis of HDMBOA‐Glc (BX10, BX11, BX12, and BX14),
   TRIMBOA‐Glc (BX13) and HDM[2]BOA‐Glc (BX14) (Figure [183]6b). Moreover,
   belowground infestation by D. v. virgifera and artificial root damage
   significantly upregulated the expression of BX10, BX13, and BX14 in
   shoot tissues (Figure [184]6b). Root herbivory by D. v. virgifera
   induced a similar expression pattern of BX genes in maize roots
   compared with that in leaf tissue after shoot herbivory. Furthermore,
   aboveground herbivory by S. frugiperda slightly upregulated several BX
   genes responsible for DIBOA (BX1, BX2, BX3, BX4, and BX5), HDMBOA‐Glc
   (BX12), and TRIMBOA‐Glc (BX13) synthesis in maize roots
   (Figure [185]6c).

FIGURE 6.

   FIGURE 6
   [186]Open in a new tab

   Effects of belowground and aboveground insect herbivory on
   benzoxazinoid biosynthesis pathway gene expression. (a) Schematic
   diagram of the benzoxazinoid biosynthesis pathway. ER, endoplasmic
   reticulum; IGPS, indole‐3‐glycerolphosphate synthase gene; HBOA,
   2‐hydroxy‐1,4‐benzoxazin‐3‐one; DIBOA,
   2,4‐dihydroxy‐1,4‐benzoxazin‐3‐one; DIBOA‐Glc,
   2,4‐dihydroxy‐1,4‐benzoxazin‐3‐one β‐D‐glucopyranose; TRIBOA‐Glc,
   2‐hydroxy‐1,4‐benzoxazin‐3‐one β‐D‐glucopyranose; DIMBOA‐Glc,
   2,4‐dihydroxy‐7‐methoxy‐1,4‐benzoxazin‐3‐one β‐D‐glucopyranose; DIMBOA,
   2,4‐dihydroxy‐7‐methoxy‐1,4‐benzoxazin‐3‐one; HDMBOA‐Glc,
   2‐hydroxy‐4,7‐dimethoxy‐1,4‐benzoxazin‐3‐one β‐D‐glucopyranose;
   TRIMBOA‐Glc, 2‐2,4,7‐trihydroxy‐8‐methoxy‐1,4‐benzoxazin‐3‐one
   β‐D‐glucopyranose; DIM2BOA‐Glc,
   4‐dihydroxy‐7,8‐dimethoxy‐1,4‐benzoxazin‐3‐one β‐D‐glucopyranose;
   HIDM2BOA‐Glc, 2–2‐hydroxy‐4,7,8‐trimethoxy‐1,4‐benzoxazin‐3‐one
   β‐D‐glucopyranose (modified from Tzin et al., [187]2017). (b and c)
   Heat map of benzoxazinoid biosynthesis‐related gene expression in maize
   shoot (b) and roots (c). Samples were collected from maize plants that
   were kept non‐manipulated (C, control) or after 72 h of belowground
   infestation by Diabrotica virgifera virgifera (DV), mechanical damage
   on root (MR), 72 h of aboveground infestation by Spodoptera frugiperda
   (SF), or mechanical damage on shoot (MS). Color coding represents the
   range of log[2](fold change relative to control).

2.6. Volatile terpene biosynthesis‐related genes induced by belowground and
aboveground insect herbivory

   Lastly, we analyzed the expression of genes coding for terpene
   synthases (TPS) (Figure [188]7), which are enzymes that control the
   synthesis of herbivory‐induced volatile terpenes that may function as
   indirect defenses in plants (Block et al., [189]2019). In maize shoot,
   all TPS genes except TPS6, TPS9, TPS11, and TPS21 were highly induced
   by S. frugiperda feeding. S. frugiperda herbivory also upregulated two
   cytochrome P450 monooxygenases, CYP92C5 and CYP92C6, which respectively
   catalyze transformation of (E)‐nerolidol and (E,E)‐geranyllinalool to
   (3E)‐4,8‐dimethyl‐1,3,7‐nonatriene (DMNT) and
   (E,E)‐4,8,12‐trimethyltrideca‐1,3,7,11‐tetraene (TMTT) (Richter et
   al., [190]2016) (Figure [191]7b). In addition, artificial root damage
   and root herbivory by D. v. virgifera induced several volatile terpene
   biosynthesis‐related genes in shoot tissue but to a much lesser extent
   than aboveground herbivory and damage (Figure [192]7b). In maize roots,
   infestation by D. v. virgifera more strongly induced volatile‐related
   genes (especially TPS2, TPS3, TPS4, TPS5, TPS23, TPS26, and CYP92C5)
   than artificial root damage. Aboveground herbivory by S. frugiperda
   also slightly upregulated the expression of TPS1, TPS9, TPS11, and
   TPS26 in roots (Figure [193]7c).

FIGURE 7.

   FIGURE 7
   [194]Open in a new tab

   Effects of belowground and aboveground insect herbivory on volatile
   terpene biosynthesis gene expression. (a) Enzymes involved in the
   production of volatile terpenes in maize. GPP, geranyl diphosphate;
   GGPP, geranylgeranyl diphosphate; FPP, farnesyl diphosphate; TPS,
   terpene synthase; CYP92C5 and CYP92C6, cytochrome P450 monooxygenases;
   TMTT, (E,E)‐4,8,12‐trimethyltrideca‐1,3,7,11‐tetraene; DMNT,
   (E)‐3,8‐dimethyl‐1,4,7‐nonatriene (Block et al., [195]2019). (b) and
   (c) Heat map of volatile terpene biosynthesis gene expression in maize
   shoot (b) and roots (c). Samples were collected from maize plants that
   were kept non‐manipulated (C, control) or after 72 h of belowground
   infestation by Diabrotica virgifera virgifera (DV), mechanical damage
   on root (MR), 72 h of aboveground infestation by Spodoptera frugiperda
   (SF), or mechanical damage on shoot (MS). Color coding represents the
   range of log[2](fold change relative to control).

3. DISCUSSION

   By analyzing changes in the maize transcriptome, we revealed defense
   responses of plants to two well‐adapted insect herbivores, D. v.
   virgifera and S. frugiperda. Artificial root and leaf damage were used
   for comparison to determine the specific transcriptomic responses of
   maize plants to these specialized insects. The results reveal that
   belowground infestation by D. v. virgifera larvae and aboveground
   feeding by S. frugiperda caterpillar trigger local and systemic
   transcriptome changes that differ in various ways from responses to
   artificial damage. D. v. virgifera and S. frugiperda caused more
   upregulated DEGs than downregulated DEGs in the specific tissue they
   fed on, root and shoot, respectively (Figure [196]1c,d). This is
   similar to transcriptome responses reported for herbivory by the beet
   armyworm Spodoptera exigua (Tzin et al., [197]2017) and Asian corn
   borer Ostrinia furnacalis (Guo et al., [198]2019) and implies that the
   maize plants respond not just to the mechanical damage caused by these
   insects, but also to possible elicitors and effectors that are
   introduced into the plants while they are feeding. Thus far, several
   potent elicitors from OS of Spodoptera caterpillars such as volicitin
   and inceptin have been identified (Alborn et al., [199]1997; Schmelz et
   al., [200]2006; Turlings et al., [201]2000) and the mechanisms
   underlying elicitor‐mediated defense responses have been extensively
   studied (Erb, Meldau, et al., [202]2012). Much less is known about
   effectors and their role in modulating plant defenses. Moreover, the
   identity and mode of action of root herbivore elicitor/effector remain
   unclear (Johnson et al., [203]2016).

   Although root herbivore‐induced leaf resistance has been extensively
   studied for maize (Erb, Flors, et al., [204]2009; Erb, Gordon‐Weeks, et
   al., [205]2009; Erb, Köllner, et al., [206]2011), little is known about
   the molecular mechanism underlying root‐to‐shoot signaling. We found
   that belowground wounding by the root herbivore and artificial root
   damage both also markedly changed the shoot transcriptome
   (Figure [207]1c), offering a dataset to help understand how root
   herbivory systemically affects defenses in maize plants. Even less is
   known about the impact of aboveground infestation on root defense, but
   we know that if S. frugiperda attacks maize before D. virgifera, the
   root herbivore's performance is negatively affected (Erb, Robert, et
   al., [208]2011). Similarly, leaf attack by diamondback moth
   caterpillars Plutella xylostella strongly reduces the performance of
   cabbage root fly larvae Delia radicum on cabbage plants Brassica
   oleracea (Karssemeijer et al., [209]2020). Our transcriptome data
   suggest that aboveground wounding (insect or artificial) causes only
   minor changes in maize roots (Figure [210]1d). This may be due to a
   transient transcriptomic change in the roots that only occurs early
   during leaf‐herbivory and therefore could not be detected in our 3‐day
   experiment. It is also possible that a minor transcription change in
   the roots upon leaf herbivory is enough to trigger an effective root
   defense.

   Feeding by S. frugiperda caterpillars was found to cause significant
   changes in the regulation of primary and secondary metabolism pathways
   in maize shoots. Transcriptomic changes in the biosynthesis of
   phenylpropanoid, flavonoid, benzoxazinoid, and metabolisms related to
   production of phytohormones and volatiles (Figure [211]S2A) indicate
   their role in defense against S. frugiperda herbivory. The much
   stronger responses caused by S. frugiperda herbivory than by artificial
   leaf damage in the photosynthesis pathway (Figure [212]S2B) might be
   explained by a compensatory growth response to consumption of leaf
   tissue by S. frugiperda. It appears that plants can differentiate
   between herbivory and mere mechanical damage and regulate their
   photosynthesis in accordance with growth‐defense trade‐offs (Visakorpi
   et al., [213]2018). Leaf damage by S. frugiperda also induced
   phenylpropanoid, flavonoid and benzoxazinoid biosynthesis in root
   tissues (Figure [214]S2C), confirming that aboveground herbivory may
   affect root‐herbivore performance, as previously shown in insect
   performance assays (Erb, Robert, et al., [215]2011).

   It is known that maize plants, in order to cope with S. frugiperda
   attack, activate the expression of genes involved in direct defense
   such as genes encoding protease and proteinase inhibitors (Pechan et
   al., [216]2002; Ton et al., [217]2007) and indirect defense such as
   genes related to volatile emissions (Köllner et al., [218]2008; Schnee
   et al., [219]2006) (Figure [220]S3A). Interestingly, our results also
   show that S. frugiperda feeding suppresses the expression of several
   candidate stress response‐related genes, such as MYB20 (Zm00001d002545)
   that is involved in secondary cell wall formation (Geng et
   al., [221]2020), MYB111 (Zm00001d026017) that is involved in regulating
   flavonoid biosynthesis (Li et al., [222]2019; Stracke et
   al., [223]2010), and indole‐2‐monooxygenase‐like (Zm00001d035178) that
   is putatively involved in DIBOA‐glucoside biosynthesis. We show that
   genes associated with primary metabolism, like
   monooxygenase/oxidoreductase (Zm00001d021444) that is involved in auxin
   biosynthesis and transcription factor LUX (LUX ARRHYTHMO)
   (Zm00001d041960) necessary for circadian rhythms (Gil &
   Park, [224]2019), are also repressed by S. frugiperda caterpillar
   feeding (Figure [225]S3B). Evidently, maize plants strongly alter
   primary and secondary metabolism in response to S. frugiperda
   herbivory, but S. frugiperda caterpillars may, as a counteradaptation,
   also suppress maize defense (De Lange et al., [226]2020).

   Apart from changes in the regulation of primary metabolism, belowground
   herbivory by D. virgifera larvae was also found to modify secondary
   metabolism pathways such as the biosynthesis of phenylpropanoids and
   monoterpenoids in maize roots (Figure [227]S4A). The difference in
   pathway enrichment between root herbivory and artificial root damage
   (Figure [228]S4B) suggests that maize plants distinguish between root
   herbivore and artificial wounding and reprogram their transcriptome
   accordingly. Importantly, maize also adjusts its DNA replication
   aboveground in response to root attack by D. virgifera
   (Figure [229]S4C), which probably affects the growth and development of
   the shoot (Castellano et al., [230]2004). Interestingly, a putative
   methyl salicylate biosynthesis‐related gene (benzenoid carboxyl
   methyltransferase omt7, Zm00001d052828) is not expressed in leaves of
   maize after leaf herbivory (Köllner et al., [231]2010) but can be
   induced in roots by drought stress (Zheng et al., [232]2020). We found
   that this gene is also induced by D. virgifera feeding
   (Figure [233]S5A), suggesting that D. virgifera attack and drought
   stress both induce root‐specific methyl salicylate. Another interesting
   gene is the one coding for anthranilic acid methyltransferase1 (aamt1,
   Zm00001d044762) responsible for the production of methyl anthranilate
   (Köllner et al., [234]2010), a repellent for D. virgifera (Bernklau et
   al., [235]2016). It was induced by both types of root damage
   (Figure [236]S5A). In contrast, the expression of several genes
   involved in the regulation of plant defense and resistance in shoots,
   for example, two putative LRR protein genes (Bianchet et
   al., [237]2019; Ye et al., [238]2020) and a cysteine proteinase
   inhibitor gene (Ton et al., [239]2007), were downregulated in response
   to D. virgifera feeding (Figure [240]S5B and Data [241]S8). The
   potential role of these genes in belowground plant‐insect interactions
   still needs to be elucidated. Surprisingly, the transcription levels of
   several photosynthetic genes were also repressed in herbivore
   infested‐roots, a nonphotosynthetic organ, but their expression levels
   are much lower than that in leaves (Figure [242]S5B, Data [243]S1, and
   Data [244]S8). Previous research showed that the suppression of
   photosynthetic gene expression is required for sustained root growth in
   Arabidopsis under phosphate deficiency (Kang et al., [245]2014).
   Possibly, herbivore infested‐roots suffer from phosphate deficiency
   caused by root damage; it is also possible that biotic stress in
   general reduces the expression of photosynthetic genes to promote root
   growth to compensate for root consumption by larvae. In summary, it
   appears that maize plants not only switch on their defenses in response
   to D. virgifera infestation but also adjust growth and development in
   both shoot and roots, preparing for tissue regeneration.

   The phytohormone network that comprises JA, SA, ABA, and ET signaling
   is highly important in regulating plant direct and indirect defenses
   against insects (Erb, Meldau, et al., [246]2012; Johnson et
   al., [247]2016; Wu & Baldwin, [248]2010). The essential role of JA
   signaling in the activation of local and systemic defense against
   chewing insect attack is well studied (Lortzing & Steppuhn, [249]2016;
   Lu et al., [250]2015). The start of JA biosynthesis is catalyzed by
   13‐LOX from α‐linolenic acid before being converted to 12‐OPDA by AOS
   and AOC (Lu et al., [251]2015; Figure [252]2a). A similar metabolic
   branch is catalyzed by 9‐LOX from linolenic and linoleic acid to
   produce 10‐OPDA and 10‐OPEA, respectively (Tzin et al., [253]2017).
   Both 10‐OPDA and 10‐OPEA display phytotoxicity, and local production of
   10‐OPEA and associated death acids (DAs) in maize induced by fungal
   southern leaf blight (Cochliobolus heterostrophus) act as a phytoalexin
   by suppressing the growth of fungi and herbivores (Christensen et
   al., [254]2015). A total of six potential 13‐lipoxygenase coding genes
   and seven candidate 9‐lipoxygenase coding genes have been predicted for
   the sequenced B73 maize genome (Woldemariam et al., [255]2018). Among
   these genes, LOX10 has been confirmed to mediate the production of
   green leaf volatiles, jasmonates, and herbivore‐induced plant volatiles
   in maize plants (Christensen et al., [256]2013). In our study, two
   13‐LOX genes (LOX10 and LOX11) and all 9‐LOX genes (especially LOX1,
   LOX2, LOX3, and LOX5) except for LOX12 were highly induced in the shoot
   upon S. frugiperda attack. In general, the expression of 9‐LOX genes
   was more strongly induced than 13‐LOX genes (Figure [257]2b), which is
   largely consistent with the reported expression patterns of LOX genes
   in maize leaves fed upon by the Asian corn borer Ostrinia furnacalis
   (Guo et al., [258]2019) and the beet armyworm Spodoptera exigua (Tzin
   et al., [259]2017), suggesting that the initiation of JA signaling in
   maize is similar in response to different chewing herbivores.
   Considering the strong expression of 9‐LOX genes in maize leaves
   infested by lepidopteran herbivores as well as the local phytoalexin
   activity of DAs produced through 9‐LOX catalyzation, the activity of
   9‐LOX might be involved in the direct defense of maize against
   caterpillar attack. LOX10 appears to be only slightly upregulated by S.
   exigua feeding (Tzin et al., [260]2017), whereas it is relatively
   strongly induced by S. frugiperda (Figure [261]2b) and O. furnacalis
   (Guo et al., [262]2019). This may reflect a difference between
   herbivore species, but may also be due to the use of different numbers
   of caterpillars or different maize lines. All the other genes involved
   in subsequent steps of JA biosynthesis in maize shoot were found to be
   upregulated by S. frugiperda feeding, especially AOS2, OPR1, and OPR2
   (Figure [263]2b), possibly reflecting the respective importance of
   these genes in the defense response to caterpillar attack. Another
   important defense gene, JAR1, mediates the production of
   jasmonoyl‐isoleucine conjugate (JA‐Ile), the active form of JA (Koo &
   Howe, [264]2012). The expression of JAR1a rather than JAR1b is highly
   induced by caterpillar attack on maize leaves (Guo et al., [265]2019;
   Tzin et al., [266]2017), and a similar increase in JAR transcription
   level was observed in our study (Figure [267]2b), further suggesting
   the importance of maize JAR1a in the biosynthesis of JA‐Ile.

   In accordance with the assumed role of JA signaling being involved in
   the local defense of plant roots against belowground herbivores (Lu et
   al., [268]2015), we found that a group of JA‐related genes is induced
   by D. v. virgifera feeding on maize roots (Figure [269]2c). However, in
   comparison with the leaf response to aboveground herbivore feeding,
   maize roots increased their expression levels of JA‐related genes to a
   lesser extent in response to belowground feeding (Figure [270]2 and
   Data [271]S9). Similarly, JA levels in maize roots were found to only
   increase about two fold upon D. v. virgifera attack (Erb, Flors, et
   al., [272]2009), which is considerably less compared with JA increases
   in leaves in response to caterpillar feeding (Schmelz et
   al., [273]2003). This is perhaps due to the different sensitivity of JA
   signaling in roots and shoot to herbivores. In a previous study,
   short‐term JA signaling was differently induced (within 24 h) by
   belowground herbivore attack and artificial root damage, but neither
   the content of JA nor the expression levels of LOX and JAR1 showed
   pronounced differences in roots after 24 h of herbivory or mechanical
   damage (Lu et al., [274]2015). Whether the maize roots can specifically
   recognize herbivores as is known for shoots (Chuang et al., [275]2014;
   Qi et al., [276]2016; Schmelz et al., [277]2009) still needs to be
   explored. We also found that belowground herbivory slightly induced
   several JA‐related genes in the shoot, whereas aboveground herbivory
   hardly changed JA signaling in roots (Figure [278]2 and Data [279]S9),
   suggesting that aboveground JA signaling is mainly responsible for
   local defense, whereas root JA signaling might be involved in
   root‐to‐shoot communication. This appears to also be the case in
   Arabidopsis thaliana, where early systemic JA responses in the shoot
   have been found to be even higher compared with the local responses in
   roots to artificial wounding (Hasegawa et al., [280]2011).

   SA is another important phytohormone for plant immunity that functions
   in basal defense and systemic acquired resistance (SAR) (Huang et
   al., [281]2020). The biosynthesis of SA in plants follows two
   independent pathways, ICS and PAL (Dempsey et al., [282]2011; Huang et
   al., [283]2020). We found that upon aboveground and belowground
   herbivore attack, a number of genes involved in PAL but not ICS pathway
   are induced in maize (Figure [284]3). A similar SA‐related gene
   expression pattern has been reported for maize leaves after O.
   furnacalis infestation (Guo et al., [285]2019). OS application of
   Mythimna separata to maize leaf wound sites also strongly elicits SA
   accumulation (Qi et al., [286]2016). However, aboveground herbivory by
   Spodoptera littoralis and O. furnacalis, or belowground herbivory by D.
   v. virgifera, do not increase SA concentration in maize leaf and roots,
   respectively (Erb, Flors, et al., [287]2009; Guo et al., [288]2019).
   Similarly, belowground attack by cucumber beetle Diabrotica balteata
   and rice water weevil Lissorhoptrus oryzophilus do not increase the SA
   content in rice roots (Lu et al., [289]2015). PAL, which enables the
   production of cinnamic acid and its downstream phenolic products
   caffeic acid and ferulic acid, is involved in the phenylpropanoid
   metabolism pathway. In maize, the levels of caffeic acid and ferulic
   acid have been reported to increase after 6 h infestation by S. exigua
   and to decrease after 24 h, which might be because these
   phenylpropanoids serve as substrate/precursors for the biosynthesis of
   other defensive compounds (Tzin et al., [290]2017). Instead of
   activating SA signaling, maize plants might mobilize phenylpropanoid
   metabolism by increasing the expression of PAL genes to accelerate
   downstream defensive metabolite accumulation, thereby protecting
   themselves against shoot and root attacks. Hydroxycinnamic acid amides
   form a diverse group of specialized phenylpropanoid metabolites in many
   plants. The abundance of several hydroxycinnamic acid amide derivatives
   such as coumaroyltyramine, coumaroyltryptamine, and feruloyltyramine is
   highly increased in maize leaves after S. littoralis attack (Marti et
   al., 2013). The importance of these metabolites in plant defense still
   needs to be examined.

   The regulator function of ABA and ET in plant defense and resistance is
   well documented (Broekgaarden et al., [291]2015; Erb &
   Reymond, [292]2019; Olds et al., [293]2018; Vos et al., [294]2013). For
   instance, ABA‐deficient Arabidopsis mutant plants are more susceptible
   to S. littoralis (Bodenhausen & Reymond, [295]2007). Here, maize plants
   increased the expression of a series of ABA‐related genes in shoot and
   roots in response to herbivory by S. frugiperda and D. v. virgifera,
   respectively (Figure [296]4). This is consistent with previous studies
   of ABA induction in maize plants upon O. furnacalis (Guo et
   al., [297]2019) and D. v. virgifera (Erb, Flors, et al., [298]2009)
   attack. However, S. littoralis infestation does not increase the ABA
   level in maize shoot (Erb, Flors, et al., [299]2009), and in rice
   roots, the biosynthesis of ABA is not induced by belowground D.
   balteata and L. oryzophilus attack (Lu et al., [300]2015). Considering
   the crosstalk between ABA and JA signaling and the role of ABA in
   drought stress response, it is expected that the ABA pathway is
   involved in systemic defenses against herbivores (Erb, Flors, et
   al., [301]2009; Erb, Köllner, et al., [302]2011; Wang et
   al., [303]2018). A previous study showed that exogenous application of
   ABA on maize root boosts aboveground defense (Erb, Gordon‐Weeks, et
   al., [304]2009). We found a few ABA biosynthesis‐related genes to be
   induced in both shoot and root in response to belowground and
   aboveground herbivory, respectively (Figure [305]4). Notably, even
   though artificial leaf damage and aboveground S. frugiperda herbivory
   increase the transcription level of ZEP1 and ZEP2, respectively, in
   maize roots, the expression of NCED was found to be repressed
   (Figure [306]4c), which might lead to the homeostasis of ABA levels in
   roots. Taken together, the results imply that ABA signaling is probably
   not only involved in maize local defenses against S. frugiperda and D.
   v. virgifera herbivory, but also partly responsible for systemic
   defenses against herbivores.

   The effect of ethylene (ET) on plant defense is variable. In maize, it
   positively regulates resistance to S. frugiperda in Mp708, an
   insect‐resistant maize inbred line, but not in Tx610, a susceptible
   maize line (Harfouche et al., [307]2006). The transcription of a rice
   ET biosynthesis‐related gene 1‐aminocyclopropane‐1‐carboxylic acid
   synthase (OsACS2) can be induced by wounding and herbivory, and
   silencing of OsACS2 has been shown to suppress ET production and reduce
   resistance to a chewing herbivore, the striped stem borer Chilo
   suppressalis (Lu et al., [308]2014). Partially consistent with this
   result, simulated caterpillar herbivory (artificial damage plus the
   application of oral secretion from M. separata), in comparison with
   mechanical wounding only, highly increases the concentration of ET in
   maize leaf tissue (Qi et al., [309]2016). Similarly, in our study, the
   transcription of four ET biosynthesis‐related genes in maize shoots was
   induced by S. frugiperda feeding but not mechanical wounding
   (Figure [310]5b). Compared with wild type plants, Arabidopsis ET
   insensitive mutant ein2‐1 is more resistant to generalist S.
   littoralis, but not to specialist diamondback moth Plutella xylostella.
   In addition, exogenous application of ET by treating the plant with
   ethephon (2‐chloroethanephosphonic acid) leads to enhanced resistance
   to S. littoralis (Stotz et al., [311]2000). Furthermore, in Arabidopsis
   thaliana, a double mutant of ET‐stabilized transcription factor ET
   insensitive3 and ET insensitive3‐like 1 (ein3 eil1) shows enhanced
   defense against S. exigua, and this is probably due to the JA and ET
   signaling antagonism in regulating plant wounding response and defense
   against insect attack (Song et al., [312]2014). Interestingly, the
   expression of EIN2, the central component of the ET signaling pathway,
   was repressed in maize shoots in response to S. frugiperda attack
   (Figure [313]5b). Taken together, our data suggest that the
   biosynthesis of ET in maize shoot is activated in response to S.
   frugiperda attack, while downstream the ET signaling pathway might be
   suppressed by JA‐ET antagonism in order to protect maize plants against
   S. frugiperda. In contrast to aboveground herbivory, both root wounding
   by D. v. virgifera feeding and artificial root damage increased the
   expression of several genes involved in ET signaling (Figure [314]5c).
   This is different in rice, where the concentration of ET is not
   increased in response to belowground herbivory by D. balteata (Lu et
   al., [315]2015). In summary, ET appears essential for modulating plant
   defenses against herbivores, but these defenses are plant species‐,
   genotype‐, tissue‐, and herbivore‐specific.

   In addition to these typical plant defense hormones, we also targeted
   benzoxazinoids. These defense metabolites occur in many monocots,
   including maize, and are effective in providing resistance against
   insect herbivores (Tzin et al., [316]2017). However, well‐adapted
   herbivores such as D. v. virgifera and S. frugiperda have been shown to
   tolerate high concentrations of benzoxazinoids and even use
   benzoxazinoids as foraging cues (Köhler et al., [317]2015; Robert et
   al., [318]2012). In maize shoot, aboveground herbivory by S. frugiperda
   caused a significantly higher expression of BX genes compared with
   artificial leaf damage, whereas belowground herbivory and artificial
   root damage resulted in a similar increase of BX genes expression
   pattern in maize roots (Figure [319]6). This was consistent with the
   JA‐related gene expression pattern in maize shoot and roots upon
   herbivory and mechanical damage (Figure [320]2). JA induces the
   production of benzoxazinoids in maize (Tzin et al., [321]2017), and
   this might explain the similarity between the expression pattern of JA‐
   and benzoxazinoid biosynthesis‐related genes in maize roots upon
   herbivory and mechanical damage. Furthermore, compared with the minor
   impact that aboveground S. frugiperda herbivory and artificial shoot
   damage had on root gene expression, belowground D. v. virgifera feeding
   and artificial root damage had a much stronger effect on the expression
   in the shoots of a series of downstream benzoxazinoid
   biosynthesis‐related genes (Figure [322]6). This implies that root
   herbivory and artificial root damage can induce shoot defense and
   resistance against leaf herbivores, and root‐to‐shoot JA signaling
   might be involved in mediating this systemic defense in maize plants.

   Plants have also evolved the ability to attract predators and
   parasitoids with herbivore‐induced plant volatiles (HIPVs) (Dicke &
   Baldwin, [323]2010; Turlings & Erb, [324]2018). Volatile terpenoids
   such as (E)‐β‐caryophyllene (Rasmann et al., [325]2005; Xiao et
   al., [326]2012), DMNT, and TMTT (Tamiru et al., [327]2011) play a
   critical role in this indirect defense. Herbivore‐induced terpene
   production is regulated by the expression of genes of the TPS family
   (Block et al., [328]2019). TPS2 and two cytochrome P450 enzyme coding
   genes, CYP92C5 and CYP92C6, are responsible for the production of DMNT
   and TMTT in maize (Richter et al., [329]2016). In this study, we
   confirm that S. frugiperda and D. v. virgifera attack increases the
   expression of a number of TPS genes in shoot and roots and more so than
   artificial damage (Figure [330]7). These TPS genes are involved in the
   biosynthesis of the major volatile terpenes emitted by
   herbivore‐infested maize plants such as nerolidol (TPS1 and TPS2),
   (E)‐β‐caryophyllene (TPS8, TPS10, and TPS23), (E)‐α‐bergamotene (TPS4,
   TPS5, and TPS10), (E)‐β‐farnesene (TPS1, TPS4, TPS5, and TPS10), and
   DMNT (TPS2 and CYP92C5) (De Lange et al., [331]2020). D. v. virgifera
   herbivory and artificial root damage also slightly but significantly
   induced the expression of a few TPS genes in maize shoots, and S.
   frugiperda attack had the same effect on maize roots (Figure [332]7).
   Hence, our results confirm that maize plants increase their volatile
   terpenoid biosynthesis in response to aboveground and belowground
   herbivory.

   In this study, we evaluated the transcriptomic changes in maize plants
   upon aboveground and belowground attack by the specialized herbivores
   S. frugiperda and D. v. virgifera and compare these changes to those
   triggered by artificially damage. The comprehensive assessment of local
   and systemic transcriptomic changes of herbivore‐infested plants
   provides new insight into the molecular mechanism underlying induced
   resistance in maize against leaf‐ and root‐herbivores, as well as into
   the plant's growth‐defense balance. In addition, the presented data can
   serve as a basis for further exploration of novel crop protection
   strategies that modify and exploit herbivore induced defenses.

4. METHODS

4.1. Plants and herbivores

   Maize seedlings (Zea mays var. Delprim) were grown individually in
   plastic pots (height 10 cm; diameter 4 cm) using a mixture of
   commercial potting soil (Einheitserde Classic, Gebrüder Patzer GmbH &
   Co. KG, Germany) and sand (Sable Capito 1–4 mm, Landi, Dotzigen,
   Switzerland) in equal proportion (1:1; v/v) under controlled conditions
   (28 ± 2°C; 60% relative humidity; 16‐/10‐h light/dark photoperiod) in
   the greenhouse. Two insect species were used for the experiments. The
   leaf herbivore Spodoptera frugiperda (JE Smith) (Lepidoptera:
   Noctuidae) and the root herbivore Diabrotica virgifera virgifera
   LeConte (Coleoptera: Chrysomelidae) were obtained from laboratory
   colonies at the University of Neuchâtel. The larvae of S. frugiperda
   were reared on artificial diet as described by Turlings et
   al. ([333]2004). The larvae of D. v. virgifera were maintained on
   freshly germinated maize roots as described by Erb, Robert, et
   al. ([334]2011).

4.2. Mechanical damage and herbivory treatments

   Twenty‐day‐old maize plants were used for the experiments. We randomly
   assigned 12 plants to each of the following five treatments: roots
   infested by (1) D. v. virgifera or (2) mechanically damaged; shoots
   infested by (3) S. frugiperda or (4) mechanically damaged and (5)
   uninfested controls (hereafter identified as treatments DV, MR, SF, MS,
   and C, respectively). For DV treatment, five second‐instar larvae of D.
   v. virgifera were released onto the soil surface around the stem of
   maize plant to infest the roots. After 72 h infestation, the whole
   roots were harvested. The larvae were removed from the roots
   immediately during root tissue harvest. For MR treatment, the roots
   were mechanically damaged by stabbing with a metal corkborer (diameter,
   7 mm) at a depth of approximately 5 cm into the soil three times daily
   for 3 days based on the methods from Rasmann et al. ([335]2005). For SF
   treatment, three newly molted third‐instar larvae of S. frugiperda were
   caged on a maize leaf using a small clip cage and allowed to feed for
   72 h. The cage was moved to an intact leaf area three times per day.
   For MS treatment, we punched an area of approximately 2 × 10 mm^2 with
   forceps on both sides of the central vein of the third and fourth leaf.
   This was repeated three times daily for 3 days and created a wounded
   leaf area of approximately 2 × 6 cm^2 every day. The whole shoots and
   roots were harvested and flash‐frozen in liquid nitrogen at 72 h after
   treatment.

4.3. Library preparation and transcriptome sequencing

   Tissue from three individual maize seedlings was combined into one
   experimental replicate, and four replicates were prepared for each
   treatment. A total amount of 1 μg RNA per sample was used for library
   construction. Sequencing libraries were generated using NEBNext® Ultra™
   RNA Library Prep Kit for Illumina® (NEB, USA) following manufacturer's
   recommendations and index codes were added to attribute sequences to
   each sample. The PCR products were purified (AMPure XP system), and
   library quality was assessed on an Agilent 2100 (Agilent Technologies,
   Palo Alto, CA, USA). The clustering of the index‐coded samples was
   performed on a cBot Cluster Generation System using PE Cluster Kit
   cBot‐HS (Illumina) according to the manufacturer's instructions. After
   cluster generation, the library preparations were sequenced on an
   Illumina HiSeq 4000 platform and paired‐end reads (2 × 150 bp) were
   generated.

4.4. RNA‐seq data analysis

   Paired‐end clean reads were mapped to the maize reference genome (B73
   RefGen_v4) (Jiao et al., [336]2017) using HISAT2 v2.0.5 program (Kim et
   al., [337]2015) with default parameters. The expression levels of genes
   were analyzed by using HTSeq v0.6.1 software (Anders et al., [338]2015)
   with union mode and were calculated as fragments per kilobase of
   transcript per million fragments mapped (FPKM). Differentially
   expressed genes (DEGs) between different experimental treatments were
   filtered by using DESeq2 R package v1.20.0 (Love et al., [339]2014)
   with false discovery rate (FDR) < .05 (Benjamini & Hochberg, [340]1995)
   and an absolute value of log[2]‐transformed fold change
   (treatment/control) > 1. Pathway enrichment of KEGG (Kyoto Encyclopedia
   of Genes and Genomes) was analyzed by using KOBAS v3.0 (Xie et
   al., [341]2011) (adjusted P < .05 were considered significantly
   enriched). Plant responses in root and shoot samples elicited by the
   leaf‐ and root‐feeding herbivores were compared with those obtained by
   artificial shoot and root damage, and samples from seedlings that were
   kept non‐manipulated served as control. We refer to local plant
   responses for tissue that was directly infested with root or shoot
   herbivores, and systemic plant responses for roots or shoots that were
   not infested but were sampled from a plant damaged in the opposite
   tissue.

CONFLICT OF INTEREST

   The authors declare no conflict of interest.

AUTHOR CONTRIBUTIONS

   W.Y. and T.C.J.T. designed the research. W.Y. performed the
   experiments. W.Y., M.E., T.C.J.T., C.B.S., and T.D. advised on the
   experimental design and wrote and revised the manuscript.

Supporting information

   Data S1. Genes detected in all samples. Gene expression levels were
   shown by FPKMs. Maize shoot (S) and roots (R) were harvested from
   seedlings that were kept non‐manipulated (C, control) or treated with
   belowground infestation by D. v. virgifera (DV), mechanical wounding on
   root (MR), aboveground infestation by S. frugiperda (SF), or mechanical
   wounding on shoot (MS); NA: no annotation.
   [342]Click here for additional data file.^ (21.8MB, xlsx)

   Data S2 All differentially expressed genes (DEGs) in maize shoot and
   roots induced by D. v. virgifera infestation with a cut‐off of two‐fold
   change relative to the control. FC: fold change; NA: no annotation.
   [343]Click here for additional data file.^ (138KB, xlsx)

   Data S3 All DEGs in maize shoot and roots induced by mechanical root
   damage with a cut‐off of two‐fold change relative to the control. FC:
   fold change; NA: no annotation.
   [344]Click here for additional data file.^ (412.1KB, xlsx)

   Data S4 All differentially expressed genes (DEGs) in maize shoot and
   roots induced by S.frugiperda infestation with a cut‐off of two‐fold
   change relative to the control. FC: fold change; NA: no annotation.
   [345]Click here for additional data file.^ (214.6KB, xlsx)

   Data S5 All DEGs in maize shoot and roots induced by mechanical shoot
   damage with a cut‐off of two‐fold change relative to the control. FC:
   fold change; NA: no annotation.
   [346]Click here for additional data file.^ (123.6KB, xlsx)

   Data S6 KEGG pathway enrichment analysis of DEGs in the transcriptome
   of maize induced by different treatments.
   [347]Click here for additional data file.^ (71.4KB, xlsx)

   Data S7 The top 60 DEGs in maize shoot induced by S. frugiperda
   infestation. FC: fold change; NA: no annotation.
   [348]Click here for additional data file.^ (22.5KB, xlsx)

   Data S8 The top 60 DEGs in maize roots induced by D. v. virgifera
   infestation. FC: fold change; NA: no annotation.
   [349]Click here for additional data file.^ (22.2KB, xlsx)

   Data S9 The gene expression pattern of phytohormones, benzoxazinoids
   and terpene volatiles.
   [350]Click here for additional data file.^ (28.1KB, xlsx)

   Figure S1 Expression levels and correlations for maize genes from
   qRT‐PCR (left) and RNA‐seq gene expression data (right). The transcript
   levels (mean + SE, n = 4) of eight genes in shoots of maize seedlings
   are shown after 72 h of belowground infestation by Diabrotica virgifera
   virgifera (DV), aboveground infestation by Spodoptera frugiperda (SF),
   application of root (MR) or shoot mechanical damage (MS). Non‐treated
   seedlings served as controls (C). The following genes were measured:
   BX14 (benzoxazinone synthesis14, Zm00001d004921), ZRP4‐like
   (O‐methyltransferase, Zm00001d038703), PR5 (pathogenesis‐related
   protein5, Zm00001d031158), PR10 (pathogenesis related protein10,
   Zm00001d028816), LOX3 (lipoxygenase, Zm00001d033623), PPO (polyphenol
   oxidase, Zm00001d000001), BBTI13 (Bowman‐Birk type trypsin inhibitor,
   Zm00001d048660), CLH (chlorophyllase1, Zm00001d019758). For qRT‐PCR
   data, fold‐change of gene expression level was calculated using the
   2^−ΔΔCT method. The results (threshold cycle values) of the qRT‐PCR
   assays were normalized to the expression level of ZmCUL (cullin,
   Zm00001d024855). For RNA‐seq data, gene expression levels were
   calculated as FPKM (fragments per kilobase of transcript per million
   fragments mapped).

   Figure S2 KEGG pathway enrichment analysis of differentially expressed
   genes (DEGs) in maize induced by S. frugiperda herbivory. (A) The top
   20 enriched KEGG pathways in maize shoot between S. frugiperda
   herbivory (SF) and non‐manipulated control (C). (B) The top 20 enriched
   KEGG pathways in maize shoot between S. frugiperda herbivory and
   artificial shoot damage (MS). (C) The enriched KEGG pathways in roots
   of maize seedlings between S. frugiperda herbivory and control.
   Enrichment scores are shown as ‐log10(adjusted P value). Number of DEGs
   involved in each pathway are shown above the bar.

   Figure S3 Heatmap of the relative expression levels (fold change after
   log2 transformation) of the 60 most up‐ (A) and down‐regulated (B)
   genes in maize shoot induced by Spodoptera frugiperda feeding. Samples
   were collected from maize plants that were kept non‐manipulated (C,
   control) or after 72 h of belowground (DV) or aboveground infestation
   (SF), or mechanical damage on roots (MR) or shoot (MS). Color coding
   represents the range of log2(fold change relative to control). Color
   bar after gene ID and pie chart under heatmap showing the potential
   gene function. (A) The top 60 most up‐regulated genes in shoots induced
   by S. frugiperda feeding, included 25 defense response‐related genes
   (e.g. proteinase inhibitor, β‐glucosidase, O‐methyltransferase, and
   genes involved in ethylene‐, benzoxazinone‐, flavonoid‐synthesis) and
   11 volatile biosynthesis‐related genes (e.g. terpene synthase,
   germacrene A synthase, dimethylnonatriene synthase, linalool synthase).
   Several genes, such as Bowman‐Birk type trypsin inhibitor
   (Zm00001d048660), O‐methyltransferase ZRP4 (Zm00001d038703), ethylene
   biosynthesis‐related gene 1‐aminocyclopropane‐1‐carboxylate oxidase 3
   (Zm00001d024852), and dirigent protein (Zm00001d004220) were induced by
   both shoot and root damage, implying their potential role in the
   systemic defense response to below‐ and aboveground herbivory. A total
   of 6 highly up‐regulated genes had no annotation. (B) The 60 most
   down‐regulated genes in shoots infested by S. frugiperda , included 15
   genes involved in transcription regulation, 12 genes involved in
   defense response, and 5 genes involved in primary metabolism that were
   highly suppressed by S. frugiperda feeding. They comprised a group of
   MYB‐related transcription factors (Zm00001d026017, Zm00001d002545, and
   Zm00001d037998) and genes involved in the biosynthesis of
   DIBOA‐glucoside (cytochrome P450 71A26, Zm00001d035178), auxin
   (monooxygenase, Zm00001d021444) and starch (α‐amylase 3 chloroplastic,
   Zm00001d043662). A total of 19 highly down‐regulated genes had no
   annotation.

   Figure S4 KEGG pathway enrichment analysis of differentially expressed
   genes (DEGs) in maize induced by D. v. virgifera herbivory. (A) The top
   20 enriched KEGG pathways in maize roots between D. v. virgifera
   herbivory (DV) and non‐manipulated control (C). (B) The top 20 enriched
   KEGG pathways in maize roots between D. v. virgifera herbivory and
   artificial root damage (MR). (C) The enriched KEGG pathways in shoot of
   maize seedling between D. v. virgifera herbivory and control.
   Enrichment scores are shown as ‐log10(adjusted P value). Number of DEGs
   involved in each pathway are shown above the bar.

   Figure S5 Heatmap of the relative expression levels (fold change after
   log2 transformation) of the 60 most up‐ (A) and down‐regulated (B)
   genes in maize roots induced by Diabrotica virgifera virfigera
   infestation relative to control. Color bar after gene ID and pie chart
   under heatmap showing the potential gene function. (A) The top 60
   up‐regulated genes in roots induced by D. v. virgifera infestation
   included 21 defense‐related genes involved in the biosynthesis of
   dhurrin (CYP79A33, Zm00001d013753), suberin (O‐methyltransferase ZRP4,
   Zm00001d038703), DIBOA‐glucoside (tryptophan synthase, Zm00001d034461
   and Zm00001d034453), isoflavonoid (Zm00001d016151), ethylene
   (1‐aminocyclopropane‐1‐carboxylate oxidase3, Zm00001d024852), auxin
   (tyrosine decarboxylase 1, Zm00001d024665), benzoxazinone
   (benzoxazinone synthesis14, BX14, Zm00001d004921), proteinase
   inhibitors (Zm00001d048656, 95 Zm00001d008548, Zm00001d048660) and so
   on, and 13 volatile emission‐related genes (six terpene synthase genes,
   one β‐caryophyllene synthase gene, two salicylate methyltransferase
   genes, one linalool synthase gene, CYP92C5, one NAD(P)‐binding
   Rossmann‐fold superfamily protein related to (Z)‐3‐hexen‐1‐yl acetate
   production, and one HXXXD‐type acyl‐transferase related to volatile
   benzenoid biosynthesis). Of these direct and indirect defense‐related
   genes in roots, polyphenol oxidase1 (Zm00001d000001), 17.4 kDa class I
   heat shock protein (Zm00001d028561), NAD(P)‐binding Rossmann‐fold
   superfamily protein (Zm00001d006796) and salicylate methyltransferase
   (Zm00001d052828) were induced by both shoot and root herbivory.
   Notably, a number of growth and development‐related genes such as GDSL
   esterase/lipase (Zm00001d023984), oil body‐associated protein
   (Zm00001d051459) and several genes coding for embryo protein
   (Zm00001d025434, Zm00001d037985, Zm00001d043709, Zm00001d002360) and
   seed maturation protein (Zm00001d026037, Zm00001d024414) were also
   induced in roots by above‐ and belowground herbivory. (B) From the 60
   most down‐regulated genes in roots in response to D. v. virgifera
   infestation, many of them were only significantly suppressed by D. v.
   virgifera herbivory, whereas the expression of two potential
   defense‐related genes (Leucine‐rich repeat [LRR] family protein,
   Zm00001d026617; cysteine proteinase inhibitor, Zm00001d014219), one
   pumilio homolog 3 (Zm00001d018925) and three genes without annotation
   (Zm00001d046137, Zm00001d007955, Zm00001d045022) were significantly
   suppressed by both below‐ and aboveground herbivory. A group of genes
   involved in photosynthesis (such as chlorophyll a‐b binding protein,
   Zm00001d009589; photosystem I subunit O, Zm00001d003767) and metabolism
   of prophyrin and chlorophyll (protochlorophyllide reductase1,
   Zm00001d001820) showed extremely low expression levels in roots
   compared to their expression in leaves. The expression of these genes
   in maize roots was down‐regulated in response to D. v. virgifera
   infestation.

   Table S2 Primers used for qRT‐PCR.
   [351]Click here for additional data file.^ (5.1MB, pdf)

   Table S1 Summary of RNA sequencing and mapping using the maize genome
   as the reference. The numerical values 1, 2, 3, 4 indicate the
   different biological replicates.
   [352]Click here for additional data file.^ (26.4KB, xlsx)

ACKNOWLEDGMENTS