Transcriptomic Analysis of Three Differentially Senescing Maize (Zea mays L.) Inbred Lines upon Heat Stress

Maize, one of the world’s major food crops, is facing the challenge of rising temperature. Leaf senescence is the most significant phenotypic change of maize under heat stress at the seedling stage, but the underlying molecular mechanism is still unknown. Here, we screened for three inbred lines (PH4CV, B73, and SH19B) that showed differentially senescing phenotypes under heat stress. Among them, PH4CV showed no obviously senescing phenotype under heat stress, while SH19B demonstrated a severely senescing phenotype, with B73 being between the two extremes. Subsequently, transcriptome sequencing showed that differentially expressed genes (DEGs) were generally enriched in response to heat stress, reactive oxygen species (ROS), and photosynthesis in the three inbred lines under heat treatment. Notably, ATP synthesis and oxidative phosphorylation pathway genes were only significantly enriched in SH19B. Then, the expression differences of oxidative phosphorylation pathways, antioxidant enzymes, and senescence-related genes in response to heat stress were analyzed in the three inbred lines. In addition, we demonstrated that silencing ZmbHLH51 by virus-induced gene silencing (VIGS) inhibits the heat-stress-induced senescence of maize leaves. This study helps to further elucidate the molecular mechanisms of heat-stress-induced leaf senescence at the seedling stage of maize.


Introduction
Heat stress, which is exacerbated by global warming, is considered to be the main cause of global crop losses [1]. To ensure crop production by efficiently improving crop varieties, it is necessary to elucidate the molecular mechanisms of crop response to heat stress [2]. Heat stress affects various physiological processes such as photosynthesis, respiration, and glucose metabolism [3]. Heat stress promotes the accumulation of reactive oxygen species (ROS), which damages the photosynthetic apparatus, particularly PSII, and eventually leads to photoinhibition [4]. In addition, heat stress can damage a plant's cell membrane system by destroying the protein structure of cell membrane, and malondialdehyde (MDA) produced by membrane lipid peroxidation further aggravates the membrane damage [5]. Several studies have shown that plants produce misfolded proteins under heat stress, which triggers the production of ROS and, ultimately, leads to the senescence or death of plants [6][7][8][9].
Leaf senescence is an important feature of plant responses to abiotic stresses. In recent years, the increasing frequency of extreme heat waves has severely affected plant growth and development. However, the molecular mechanisms underlying leaf senescence induced by heat stress remain largely unrevealed [10]. Studies on wheat have demonstrated that high temperature accelerates leaf senescence, which in turn reduces grain filling time and leads to yield reduction [11]. Maize (Zea mays L.) is one of the most planted crops worldwide, with about one-third of the world's population relying on maize as their main

Varied Senescence Phenotypes of Different Maize Inbred Lines in Response to Heat Stress
Seven maize inbred lines were retrieved from local germplasm collections for the initial analysis of their responsiveness to heat stress. It was demonstrated that the leaves of their seedlings senesced differentially upon high-temperature treatment (42 • C day/35 • C night), and three of them with the distinct phenotypes of senescence-the least-senesced line PH4CV (paternal parent of XY335), the most-senesced line SH19B (CIMMYT inbred line), and the intermediately senesced line B73-were chosen for further analysis ( Figure S1). The three-leaf stage seedlings of the three lines were treated with the day/night temperature regime for three days to precisely evaluate their phenotypic response to heat stress. Compared with the untreated controls, SH19B showed a significantly senesced phenotype in both the first lobe (V1) and the second lobe (V2), B73 displayed a significantly senesced phenotype only in V1, while PH4CV exhibited no visible senesced phenotype at all ( Figure 1A). Consistently, the chlorophyll contents in both the V1 and V2 of SH19B decreased significantly, whereas a significant decrease in chlorophyll content only occurred in the V1 of B73 and no significant changes in chlorophyll content were detected in the leaves of PH4CV ( Figure 1B). Accordingly, the changing trend in the ion leakage rate in the respective organs of both SH19B and B73 was just opposite; as expected, the ion leakage rate in the leaves of PH4CV remained largely similar after heat treatment (HT) ( Figure 1C). These observations provoked us to explore the underlying transcriptional changes.

Transcriptome Analysis of the Molecular Basis Underlying the Differentially Senescing Phynotypes in Response to Heat Stress
To explore the molecular basis of differential senescing phenotypes under heat stress, a transcriptome analysis was performed. A total of 12 cDNA libraries were constructed, which consisted of 12 leaf samples of the three inbred lines under HT as well as CK, with two biological replicates for each line. The libraries were then sequenced using Illumina Novaseq 6000 platform. A total of about 85.08 GB of clean data was finally obtained, with an average Q30 value of about 93.7% (Table S1). Using B73 genome (GCF_902167145.1) as a reference, 601 million reads were mapped. There were 84,127,307 (HT_SH19B), 87,847,433 (CK_SH19B), 80,954,154 (HT_B73), 86,897,700 (CK_B73), 87,499,572 (HT_PH4CV), and 90,918,674 (CK_PH4CV) clean paired-end reads used for future analyses (Table S2).

Transcriptome Analysis of the Molecular Basis Underlying the Differentially Senescing Phynotypes in Response to Heat Stress
To explore the molecular basis of differential senescing phenotypes under heat stress, a transcriptome analysis was performed. A total of 12 cDNA libraries were constructed, which consisted of 12 leaf samples of the three inbred lines under HT as well as CK, with two biological replicates for each line. The libraries were then sequenced using Illumina Novaseq 6000 platform. A total of about 85.08 GB of clean data was finally obtained, with (B) chlorophyll contents (SPAD) of the respective organs; (C) ion leakage rates (%) of the above organs. V1 and V2 represent the first and second lobes from the bottom, respectively. Error bars represent ± SD (n = 3 biological replicates); "N.S" indicates no significant difference, "*" represents p < 0.05, "**" represents p < 0.01.
To investigate the global differences in the transcriptomes of PH4CV, B73, and SH19B leaves in response to heat stress, spearman correlation coefficient (SCC) analysis and principal component analysis (PCA) were performed on the average fragments per kilobase million (FPKM) values of all expressed genes in 12 leaf samples. After analysis, the SCC among the biological replicates was around 0.94, indicating a high quality of the replications (Figure 2A). PCA showed that the two biological replicates of each sample were closely clustered, which supported the correlation of the transcriptome sequencing results. In addition, there were differences among the inbred lines and between the HT and CK groups ( Figure 2B).

Common Molecular Responses of the Three Inbred Lines to Heat Stress
In order to fully understand the molecular effects of heat stress on the three inbred lines, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were comprehensively performed on the DEGs of CK_PH4CV vs. We compared the expressed genes of each inbred line (CK_PH4CV vs. HT_PH4CV, CK_B73 vs. HT_B73 and CK_SH19B vs. HT_SH19B) pairwise to determine the genes that specifically respond to heat stress. The results showed 3012 upregulated and 2980 downregulated genes in SH19B, 3196, and 2826 in B73 and 3761 and 3766 in PH4CV, respectively ( Figure 2C, Supplementary File S1). Importantly, it was shown in a Venn diagram that after HT, there were 2020 genes expressed commonly in the three inbred lines, while 2899, 2103, and 1709 genes specifically responded to heat stress in PH4CV, B73, and SH19B, respectively ( Figure 2D).

Common Molecular Responses of the Three Inbred Lines to Heat Stress
In order to fully understand the molecular effects of heat stress on the three inbred lines, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were comprehensively performed on the DEGs of CK_PH4CV vs. HT_PH4CV, CK_B73 vs. HT_B73, and CK_SH19B vs. HT_SH19B. It was found that "photosynthesis" (GO:0009768, GO:0009765, map00196, and map00195), "carbon metabolic pathway" (map00500, map00010, map00620, and map00710), "heat stress response" (GO:0009408 and GO:000) 9266), "ROS response" (GO:0000302 and GO:0042542), "osmotic stress response" ( Other relevant studies revealed that heat stress can cause a decrease in the photosynthesis rate, resulting in a decrease in carbohydrate supply and an increase in carbohydrate demand [17]. We showed the effects of heat stress on photosynthesis-related genes (including those involved in photosystem I, photosystem II, light complement complex, and photosynthetic electron transport) in the three inbred lines through heat maps, and found that photosynthesis-related genes were downregulated in all the three inbred lines ( Figure 3A). In addition, we found that the genes related to carbon metabolism were both upregulated and downregulated in the three inbred lines, to varying degrees ( Figure 3B), indicating that heat stress had significant effects on photosynthesis and carbon metabolism processes.
Recent studies have found that there is a common response mechanism between heat stress and drought, low temperature, and other stress conditions [5,15,18], indicating a close relationship between heat stress and other types of environmental stress. In addition, ROS is a common response source node under various environmental stress conditions [19]. Studies have shown that ROS can induce the expression of heat-response proteins (HSPs) [20,21]. In this study, we found that HSPs and heat-response transcription factors (HSFs) were significantly enriched in the GO terms of " heat stress response " (GO:0009408 and GO:0009266), " ROS response " (GO:0000302 and GO:0042542), "osmotic stress response " (GO:0006970), and " salt stress response " (GO:0009651) ( Figure S2, Supplementary File S2), and most of them were upregulated in the three inbred lines ( Figure 3C). These results demonstrate a common effect of heat stress on the three inbred lines at the transcriptional level and also indicate a general effect of heat stress on maize at the seedling stage.

Distinct Molecular Responses of the Three Inbred Lines to Heat Stress
To investigate the differences in the responses of the three inbred lines to heat stress, based on the GO and KEGG analysis results, we found that genes involved in "ATP synthesis" (GO:0009145, GO:0009206, and GO:0009205) and the "oxidative phosphorylation pathway" (map00190) were significantly enriched only in CK_SH19B vs. HT_SH19B ( Figures S2 and S3, Supplementary Files S2 and S3). In eukaryotes, oxidative phosphorylation is a biological process in which ATP is produced by combining ADP and inorganic phosphate via the proton gradient produced by the mitochondrial inner membrane. The proton gradient is produced by the catalysis of a variety of enzymes, including: NADHubiquinone oxidoreductase, ATP synthase, cytochrome c oxidase, and NADH dehydrogenase [22]. However, oxidative phosphorylation also produces ROS along with ATP, further exacerbating ROS accumulation under heat stress [23]. Several lines of evidence have shown that an increase in the ROS level under heat stress is one of the important factors leading to leaf senescence [24][25][26]. This suggests that heat stress may accelerate the senescence process of SH19B old leaves by producing excessive ROS through the oxidative phosphorylation pathway. Accordingly, we analyzed the differential expression of related genes encoding NADH dehydrogenase, ATP synthase, and cytochrome c oxidase in response to heat stress in the three inbred lines, and found that the expression of these genes was significantly upregulated in SH19B, and some of them were upregulated in B73 as well, while there were no significant changes detected in PH4CV ( Figure 4A), indicating that there is a positive correlation between their expression and the senescence degree among the three inbred lines.
that photosynthesis-related genes were downregulated in all the three inbred lines ( Figure  3A). In addition, we found that the genes related to carbon metabolism were both upregulated and downregulated in the three inbred lines, to varying degrees ( Figure 3B), indicating that heat stress had significant effects on photosynthesis and carbon metabolism processes.
Recent studies have found that there is a common response mechanism between heat stress and drought, low temperature, and other stress conditions [5,15,18], indicating a close relationship between heat stress and other types of environmental stress. In addition, ROS is a common response source node under various environmental stress conditions [19]. Studies have shown that ROS can induce the expression of heat-response proteins (HSPs) [20,21]. In this study, we found that HSPs and heat-response transcription factors (HSFs) were significantly enriched in the GO terms of " heat stress response " (GO:0009408 and GO:0009266), " ROS response " (GO:0000302 and GO:0042542), "osmotic stress response " (GO:0006970), and " salt stress response " (GO:0009651) ( Figure S2, Supplementary File S2), and most of them were upregulated in the three inbred lines ( Figure 3C). These results demonstrate a common effect of heat stress on the three inbred lines at the transcriptional level and also indicate a general effect of heat stress on maize at the seedling stage. Antioxidant enzyme systems can remove ROS and protect plant cells from oxidative damage [14,27]. In this study, a variety of enzymes of an antioxidant defense system were differentially expressed under heat stress, including glutathione S-transferase (GST), ascorbate peroxidase (APX), peroxidase (POX), and glutaredoxin (GRX). We compared the differential expression of these antioxidant oxidase genes among the three inbred lines and found that the five GSTs were all upregulated in PH4CV, while LOC542633_ZmGST23 and LOC541837_ZmGST24 were downregulated in B73 and SH19B. Two APXs (LOC100193740_ ZmAPX6 and LOC100194161_ZmAPX7) were upregulated in PH4CV. All four GRXs were upregulated in PH4CV, LOC103630730_ZmGRXS9, LOC100275991_ZmGRXS4, and LOC100284992_ZmGRX were upregulated in B73, and LOC100284992_ZmGRX was downregulated in SH19B. The three POXs were upregulated in PH4CV and downregulated in B73 and SH19B ( Figure 4B). These data suggest that the upregulated expression of antioxi-dant oxidase genes in PH4CV and B73 may be involved in inhibiting ROS-mediated leaf senescence under heat stress.

Responsive Expressions of Senescence-Related Genes in the Three Inbred Lines to Heat Stress
Leaf senescence-related genes include phytohormone metabolism and signal-transductionrelated genes, transcription factors (TFs), Chl-catabolic genes (CCGs), and senescence-associated genes (SAGs) [28][29][30]. Phytohormones are closely related to the regulation of leaf senescence. Among them, cytokinins, auxins, and gibberellins (GAs) are involved in the negative regulation of leaf senescence, whereas ethylene (ET), abscisic acid (ABA), salicylic acid (SA), jasmonic acid (JA), and brassinosteroids (BRs) were involved in the positive regulation of leaf senescence [28][29][30]. Figure 5A shows the differential expression of phytohormone metabolism and signaling pathway genes in response to heat stress in the three inbred lines. We detected six cytokinin-O-glucosyltransferases (CGTs) genes and one carboxy-lyase (CL) gene involved in cytokinin metabolism, among which, LOC100283506_ZmCGT1_1, LOC100281699_ZmCGT2_1, and LOC100282510_ZmCL were upregulated in PH4CV or B73 but downregulated in SH19B. LOC100191964_ZmCGT1_2, and LOC100282611_ZmCGT2_2 were upregulated in PH4CV but downregulated in B73 and SH19B. LOC103636476_ZmCGT2_3, and LOC100285624_ZmCGT1_3 were downregulated in PH4CV but upregulated in B73 and SH19B. Abscisic acid 8'-hydroxylase 1 (LOC100383693_ZmCYP707A1), gibberellin oxidase 2-beta-dioxygenase (LOC100280480_ ZmGA2ox), and methyl indole-3-acetate methyltransferase (LOC100285262_ZmMIAM) genes were upregulated in PH4CV but downregulated in B73 and SH19B. Moreover, the indole-3-acetic acid-amido synthetase (LOC103648032_ZmGH3.13) involved in auxin metabolism was downregulated in PH4CV but upregulated in B73 and SH19B ( Figure 5A). Genes related to phytohormone signal transduction showed differential responses to heat stress in the three inbred lines ( Figure 5A). Among them, LOC103648545_ZmVAN3B, and LOC100274569_ZmIAA27 involved in auxin signaling and LOC100502464_ZmANT, LOC100285281_ZmPHOS34, LOC103646542_ ZmERF113, and LOC100283060_ZmERF1 involved in ethylene signaling were upregulated in B73 or SH19B but downregulated in PH4CV; LOC542015_ZmSERK2 and LOC103630619_ZmBSL2 involved in BR signaling and LOC103651227_ZmLYK3 involved in ABA signaling were upregulated in PH4CV or B73 but downregulated in SH19B. In addition, LOC100281447_ZmCIGR2 involved in GA signaling and LOC100283895_ZmRING-zf and LOC100283443_ZmDET2 involved in BR signaling were upregulated in PH4CV but downregulated in B73 or SH19B. These results indicate that phytohormone metabolism and signaling impose significant effects on leaf senescence induced by heat stress.
The chlorophyll degradation process has been largely elucidated. It consists of reaction steps catalyzed mainly by three enzymes, NYE/SGR (non-yellowing/stay-green), PPH (pheophytinase), and PAO (pheophorbide a oxygenase) [28,30]. In this study, a total of 3 NYEs, 4 PPHs, and 5 PAOs were detected, and their expression trends in response to heat stress were consistent in the three inbred lines ( Figure 5A), indicating that these CCGs may not be responsible for the inconsistent senescence phenotypes of the three inbred lines under heat stress. Further analysis showed that SAGs expressions were different among the three inbred lines. LOC118474289_ZmSAP was upregulated in SH19B but downregulated in PH4CV and B73. LOC103635052_ZmSAG21 was upregulated in B73 and SH19B but downregulated in PH4CV; LOC100192995_ZmSAG5, LOC541647_ZmSAG12, LOC100272794_ZmRFP, and LOC100384378_ZmPSAT were upregulated in PH4CV and B73 but downregulated in SH19B. In addition, LOC103634911_ZmZC3H2, which encodes the CCCH zinc finger protein, was upregulated in PH4CV but downregulated in B73 and SH19B ( Figure 5C). These are the possible candidate genes involved in heat-stress-induced leaf senescence.

Silencing ZmbHLH51 by VIGS Inhibits Heat-Stress-Induced Leaf Senescence
To temporarily validate the role of the TFs identified in this study on heat-stressinduced leaf senescence, a tobacco rattle virus (TRV)-induced gene silencing (VIGS) system was used. We first used maize PDS (phytoene desaturase) to validate the silencing effect of maize VIGS system [36,37]. It was found that infection with TRV::ZmPDS silenced the ZmPDS-impaired chloroplast development, resulting in a photo-bleached phenotype ( Figure S4), suggesting that the VIGS system used in this study was feasible for silencing maize genes.
In order to determine the inhibitory effect of silencing TFs on leaf senescence, four TFs' genes, LOC100191786_ZmbZIP19, LOC107275227_ZmNAC41, LOC100282509_ZmbHLH104, and LOC100285705_ZmbHLH51, which were upregulated in SH19B and downregulated in PH4CV and B73, were initially tested via VIGS experiments. Among the four tested genes, we observed senescence-inhibited phenotypes on the leaves of TRV::ZmbHLH51 plants in the background of SH19B, consistently within three biological replicates ( Figure 6B). Compared with TRV::00, the expression of ZmbHLH51 in TRV::ZmbHLH51 was greatly reduced ( Figure 6C), indicating that ZmbHLH51 was largely silenced. Leaves of the plants with ZmbHLH51 silenced contained higher chlorophyll contents and displayed lower ion leakage rates under heat stress ( Figure 6D,E). These results indicate that the downregulation of ZmbHLH51 expression could inhibit heat-stress-induced senescence in maize leaves. the three inbred lines (PH4CV, B73, and SH19B) after heat treatment; (B) differential expressions of TFs (MYBs, WRKYs, bZIPs, bHLHs, and NACs); (C) differential expressions of Chl-catabolic genes (CCGs) and senescence-associated genes (SAGs). Each block represents the ratio of the expression of a gene under heat treatment (HT) to untreated (CK), Log10 (HT/CK). Red indicates upregulation, and blue indicates downregulation. The scale indicates the intensity of the colors.

Silencing ZmbHLH51 by VIGS Inhibits Heat-Stress-Induced Leaf Senescence
To temporarily validate the role of the TFs identified in this study on heat-stressinduced leaf senescence, a tobacco rattle virus (TRV)-induced gene silencing (VIGS) system was used. We first used maize PDS (phytoene desaturase) to validate the silencing effect of maize VIGS system [36,37]. It was found that infection with TRV::ZmPDS silenced the ZmPDS-impaired chloroplast development, resulting in a photo-bleached phenotype ( Figure S4), suggesting that the VIGS system used in this study was feasible for silencing maize genes.
In order to determine the inhibitory effect of silencing TFs on leaf senescence, four TFs' genes, LOC100191786_ZmbZIP19, LOC107275227_ZmNAC41, LOC100282509_ZmbHLH104, and LOC100285705_ZmbHLH51, which were upregulated in SH19B and downregulated in PH4CV and B73, were initially tested via VIGS experiments. Among the four tested genes, we observed senescence-inhibited phenotypes on the leaves of TRV::ZmbHLH51 plants in the background of SH19B, consistently within three biological replicates ( Figure 6B). Compared with TRV::00, the expression of ZmbHLH51 in TRV::ZmbHLH51 was greatly reduced ( Figure 6C), indicating that ZmbHLH51 was largely silenced. Leaves of the plants with ZmbHLH51 silenced contained higher chlorophyll contents and displayed lower ion leakage rates under heat stress ( Figure 6D,E). These results indicate that the downregulation of ZmbHLH51 expression could inhibit heat-stress-induced senescence in maize leaves.

Discussion
Heat stress can cause leaf wilting or premature senescence of seedling-stage maize plants, decrease a plant's chlorophyll content, and increase its ion leakage rate [7,18]. In this study, we observed the differentially senescing phenotypes of three maize inbred lines, PH4CV, B73, and SH19B, after heat stress, as evidenced via measurement of the ion leakage rate and chlorophyll content of their leaves. SH19B showed the most obvious leaf senescence symptom, while PH4CV displayed no detectable leaf senescence phenotypes ( Figure 1). As far as the DEGs are concerned, there were 2899, 2103, and 1709 independent response genes in PH4CV, B73, and SH19B, respectively, with 2020 co-response genes among them ( Figure 2D). Vasseur et al. (2011) pointed out that under heat-stress conditions, the plant metabolic rate increases, resulting in an increase in carbohydrate demand, a decrease in photosynthesis rate, and a decrease in carbohydrate supply [17]. Through GO and KEGG enrichment analysis, we found that the photosynthetic and glucose metabolic pathways in PH4CV, B73, and SH19B were significantly enriched ( Figures S2 and S3), suggesting that heat stress could indeed affect both photosynthesis and carbon balance. Other studies have demonstrated that heat stress and drought stress [5,18] or low temperature stress [15] share a common response mechanism. In this study, we found that genes related to heat-stress response, salt-,stress response and osmotic-stress response were significantly enriched in the three inbred lines (Figure S2), indicating a close relationship between heat-stress response and salt-stress and osmotic-stress responses.
Our enrichment analysis further showed that ATP synthesis and oxidative phosphorylation pathways were significantly enriched in SH19B, which provided clues to explore the molecular basis of the differences in response to heat stress among the three inbred lines ( Figure S3). Similar to our results, related transcriptome analyses have shown that oxidative phosphorylation plays an important role in response to environmental stress [38][39][40]. Mitochondria produce ATP through oxidative phosphorylation and also produce ROS, which causes damage to plant cells [41]. Studies in Arabidopsis have found that SHOT1 (inhibitor 1 of hot1-4) is a key component of the oxidative phosphorylation complex, and mutant SHOT1 reduces ROS production, which in turn reduces oxidative damage under heat stress [23]. These studies suggest that the premature senescence of SH19B under heat stress may be caused by excessive accumulation of ROS. ROS is a ubiquitous signal capable of inducing leaf senescence through protein oxidation, enzyme inactivation, altered gene expression, and biofilm breakdown [24,42]. In this study, two GO terms (GO:0000302, GO:0042542) were enriched and related to ROS responses ( Figure S2).
The detoxification mechanism of plants is crucial for crop response to heat stress. In this study, a total of three POXs, five GSTs, four GRXs, and two APXs were identified, which were involved in heat stress response and were differentially expressed in the three inbred lines ( Figure 4B). POXs and APXs are known to break down H 2 O 2 in cells, thereby protecting them from damage caused by heat stress [43,44]. GSTs can combine glutathione with various heterogeneous compounds to form soluble glutathione S-conjugates to achieve plant detoxification [45]. In addition, GRXs were involved in regulating ROS levels under both normal growth and stress conditions [14]. Recent studies have shown that overexpression of Arabidopsis GRXS17 in maize alleviates heat damage and significantly increases yield compared to non-transgenic maize lines [14]. In this study, most of the antioxidant enzyme genes were upregulated in PH4CV but downregulated in SH19B ( Figure 4B), indicating that PH4CV maintained a higher level of antioxidant capacity, better protected the integrity of the cell membrane, and inhibited leaf senescence induced by heat stress.
The increase of ROS level under heat stress is the main cause of SAGs expression and leaf senescence [8,24,46]. In this study, we found that multiple SAGs responded to heat stress and were differentially expressed in three inbred lines ( Figure 5). Several studies have shown that phytohormones play an important role in heat-stress-induced leaf senescence [47][48][49][50]. In this study, we found that the genes related to the metabolism of cytokinins, ABA, GA, SA, and auxin respond to heat stress and are differentially expressed in the three inbred lines ( Figure 5A). More importantly, the genes related to ABA, GA, SA, ET, auxin, and BR signaling in response to heat stress were also differentially expressed in the three inbred lines ( Figure 5A), indicating that these phytohormone metabolism and signal-transduction-related genes may be involved in heat-stress-induced maize leaf senescence. In addition, TFs are widely involved in complex cascades of signaling induced by hormonal signaling and environmental stresses [51]. In this study, we found that numerous genes of the NACs, bHLHs, bZIPs, WRKYs, and MYBs TF families responded to heat stress and were differentially expressed in the three inbred lines ( Figure 5B). Notably, quite a few SAGs were differentially expressed in the three inbred lines ( Figure 5C), indicating that long-term heat stress may regulate leaf senescence in maize through direct induction of SAGs expression. In Arabidopsis, bHLH family TFs PIF4 and PIF5 were found to effectively promote heat-stress-induced leaf senescence through hormonal signaling [10]. In this study, silencing ZmbHLH51, which was upregulated in SH19B in response to heat stress, by VIGS inhibited leaf senescence (Figure 6), suggesting that ZmbHLH51 is involved in promoting heat-stress-induced leaf senescence. These data provide a foundation for further analysis of the cascade signals of leaf senescence induced by heat stress in maize.

cDNA Library Preparation and Illumina Sequencing
RNA was extracted from 12 leaf samples of three maize inbred lines. RNA quality was monitored by agarose-gel electrophoresis. RNA concentration was checked by Nan-odrop2000 (NanoDrop Technologies, Wilmington, CA, USA). RNA integrity was further checked by Agilent 2100 bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Transcriptome sequencing was performed on Illumina Novaseq 6000 sequencing platform (Majorbio, Shanghai, China).

Transcriptome Sequencing Data Analysis
Fastp was used for quality detection, including deletion of joint sequences, low quality read segments, high N-rate (N indicates uncertain base information) sequences, and excessively short sequences. Then, sequence alignment analysis was performed by  [52]. DEGseq2 was used to analyze the identify of DEGs with the criterion of log2 (fold change) > 1 and adjusted p-value < 0.05 [53]. Venn was used to analyze the number of overlapping genes between samples. Goatools (https://github.com/tanghaibao/GOatools, accessed on 23 March 2022) and KOBAS (http://kobas.cbi.pku.edu.cn/home.do, accessed on 23 March 2022) were used for GO and KEGG enrichment analysis. By Fisher's exact test and BH (FDR) multiple test, the corrected p-value < 0.05 was considered as a significant enrichment item.

qRT-PCR
The relative expression of maize genes was detected by qRT-PCR. In brief, cDNA synthesis was performed using a RT Kit with gDNA Clean for qPCR II (Accurate Biology, Changsha, China). The transcript levels of each gene were measured using a 20 µL volume of SYBR (Accurate Biology, Changsha, China) on a CFX96 real-time PCR assay system (Bio-Rad, Hercules, CA, USA). qRT-PCR was performed under the following conditions: initial denaturation at 95 • C for 10 min, 95 • C for 20 s, 60 • C for 30 s, and 72 • C for 20 s for 40 cycles. We used 2 −∆∆Ct method to calculate the relative expression of RT-qPCR. Primers used for RT-qPCR are summarized in Table S3. ZmActin was used as an internal control [54].

VIGS
VIGS experiments were performed as described previously [36,37]. A specific silencing sequence of target gene was screened by the SGN-VIGS tool (https://vigs.solgenomics. net/, accessed on 23 March 2022) and inserted into pTRV2. A. tumefaciens strain GV3101 carrying pTRV1and pTRV2-derived vectors (TRV2::PDS, TRV2::ZmbHLH51) were cultured in LB medium to (OD600) 1.0. Then, pTRV1 and pTRV2 agroinfiltration liquids were mixed in equal volume, adding acetosyringone (20 mg L −1 ), cysteine (400 mg L −1 ), and Tween 20 (5 mL L −1 ). Three-day-old maize seedlings were completely soaked in mixed agroinfiltration liquid, then infiltrated for 10 min with vacuum assistance. The mixture was placed in a flask and co-cultured at 28 • C for 10 h. After co-culture, the seeds were cleaned with sterile water and planted in a light incubator. We verified the silencing effect of the maize VIGS system by ZmPDS. Primers used in VIGS are summarized in Table S4.

Statistical Analysis
In this study, the GraphPad Prism 6.0 software was used to process data and graphs. The SPSS software (version 20; SPSS, Inc., Chicago, IL, USA) was used for analysis of variance. Gene expression heat maps were created in MS Excel using a three-color scale in conditional format.