Glycerol-Induced Powdery Mildew Resistance in Wheat by Regulating Plant Fatty Acid Metabolism, Plant Hormones Cross-Talk, and Pathogenesis-Related Genes

Our previous study indicated that glycerol application induced resistance to powdery mildew (Bgt) in wheat by regulating two important signal molecules, glycerol-3-phosphate (G3P) and oleic acid (OA18:1). Transcriptome analysis of wheat leaves treated by glycerol and inoculated with Bgt was performed to identify the activated immune response pathways. We identified a set of differentially expressed transcripts (e.g., TaGLI1, TaACT1, and TaSSI2) involved in glycerol and fatty acid metabolism that were upregulated in response to Bgt infection and might contribute to G3P and OA18:1 accumulation. Gene Ontology (GO) enrichment analysis revealed GO terms induced by glycerol, such as response to jasmonic acid (JA), defense response to bacterium, lipid oxidation, and growth. In addition, glycerol application induced genes (e.g., LOX, AOS, and OPRs) involved in the metabolism pathway of linolenic and alpha-linolenic acid, which are precursor molecules of JA biosynthesis. Glycerol induced JA and salicylic acid (SA) levels, while glycerol reduced the auxin (IAA) level in wheat. Glycerol treatment also induced pathogenesis related (PR) genes, including PR-1, PR-3, PR-10, callose synthase, PRMS, RPM1, peroxidase, HSP70, HSP90, etc. These results indicate that glycerol treatment regulates fatty acid metabolism and hormones cross-talk and induces the expression of PR genes that together contribute to Bgt resistance in wheat.


Introduction
Bread wheat (Triticum aestivum L.) is one of the most important food crops worldwide. Blumeria graminis f.sp. tritici (Bgt), is an obligate biotrophic ascomycete fungus that invades the aerial parts of wheat, causing powdery mildew disease. The damage caused by Bgt can result in yield losses from 30% to 40% in years with severe epidemics [1,2]. Deployment of resistance (R) genes and spraying with foliar fungicides are the common methods for diseases control in wheat. However, many R genes lost their effectiveness due to the appearance of new pathogen races, and fungicides are not always economically feasible and environmentally friendly. Therefore, new sustainable effective ways for disease control in wheat are needed.

Bgt Resistance Test in Water-Treated and Glycerol-Treated Wheat Leaves
Four groups of wheat plants were prepared for the identification of the transcriptional responses to glycerol treatment of wheat leaves, with and without Bgt infection: H0 and G0 groups represented water or glycerol-treated plants without Bgt infection; H24 and G24 groups represented water or glycerol-treated plants with Bgt infection ( Figure 1A). The morphology of conidiophores and hyphae was monitored in water and glycerol-treated wheat leaves infected with Bgt (H24 and G24 groups). The obtained results showed that~90% of the spores successfully germinated on water-treated leaves, while only about 10.7% of the spores germinated on the glycerol-treated leaves ( Figure 1B,C). These results suggest that glycerol treatment induced effective resistance against Bgt infection. Therefore, changes in gene expression induced by both glycerol treatment and Bgt infection were explored using whole transcriptome analysis.

Bgt Resistance Test in Water-Treated and Glycerol-Treated Wheat Leaves
Four groups of wheat plants were prepared for the identification of the transcriptional responses to glycerol treatment of wheat leaves, with and without Bgt infection: H0 and G0 groups represented water or glycerol-treated plants without Bgt infection; H24 and G24 groups represented water or glycerol-treated plants with Bgt infection ( Figure 1A). The morphology of conidiophores and hyphae was monitored in water and glycerol-treated wheat leaves infected with Bgt (H24 and G24 groups). The obtained results showed that ~90% of the spores successfully germinated on water-treated leaves, while only about 10.7% of the spores germinated on the glycerol-treated leaves ( Figure 1B,C). These results suggest that glycerol treatment induced effective resistance against Bgt infection. Therefore, changes in gene expression induced by both glycerol treatment and Bgt infection were explored using whole transcriptome analysis.

Transcriptome Sequencing, Assembly, and Functional Annotation
Transcriptome sequencing revealed a total of 681,705,618 clean reads (85.21 G clean data) from twelve cDNA libraries, constituting 41,214,380-77,987,478 clean reads (5.15-9.74 G clean data) for each pool, with a quality score of Q30 ≥89.12%. Following assembly, 81.99%~86.69% of the reads were mapped on the wheat genome reference sequence IWGSC RefSeq v1.0 (Table S1). In total, we annotated 106,755 distinct assembled unigenes after blast searching in six databases (Table S2, Figure  1D). (C) Germination rates of powdery mildew on the water-treated and glycerol-treated wheat leaves at 24 hpi. Each value is the mean ± standard error (SE) of three independent biological repetitions. Asterisk indicates a significant difference from the germination rates of treated leaves at p ≤ 0.05 by Student's t-test. (D) Differentially expressed genes (DEGs) overlapped in different groups comparison. H0 and G0 groups presented water or glycerol-treated plants without Bgt infection; H24 and G24 group presented water or glycerol-treated plants with Bgt infection.

Transcriptome Sequencing, Assembly, and Functional Annotation
Transcriptome sequencing revealed a total of 681,705,618 clean reads (85.21 G clean data) from twelve cDNA libraries, constituting 41,214,380-77,987,478 clean reads (5.15-9.74 G clean data) for each pool, with a quality score of Q30 ≥89.12%. Following assembly, 81.99%~86.69% of the reads were mapped on the wheat genome reference sequence IWGSC RefSeq v1.0 (Table S1). In total, we annotated 106,755 distinct assembled unigenes after blast searching in six databases (Table S2, Figure 1D).

Identification of Differentially Expressed Genes (DEGs)
In the water-treated group (H24 vs. H0), 5616 DEGs were identified of which 3489 were upregulated and 2127 were downregulated in response to Bgt. In the glycerol-treated group (G24 vs. G0), only 2494 DEGs were identified, of which 1723 were upregulated and 771 were downregulated in response to Bgt. A total of 414 DEGs (376 up, 38 down) were identified in the G0 vs. H0 groups, and 463 DEGs (308 up; 155 down) were identified in the H24 vs. G24 groups (Table 1). A total of 1317 upregulated and 467 downregulated DEGs were overlapped in H24 vs. H0 and G24 vs. G0 groups, respectively. More than half of the DEGs in G0 vs. H0 groups overlapped with G24 vs. G0 or H24 vs. H0 groups ( Figure 1D). DEGs identified in different biological replicates were clustered together in a heat map of expression levels, indicating good reproducibility between replicates ( Figure S1). The annotated DEGs of different groups (H24 vs. H0, G24 vs. G0, G0 vs. H0, G24 vs. H24) are attached in Table S3. To better understand the response to Bgt infection in both water-and glycerol-treated plants, we annotated DEGs into Gene Ontology (GO) terms and KEGG pathways (Tables S4 and S5). DEGs were mainly classified into cell killing, immune system process, response to stimulus, signaling, and metabolic process GOs (e.g., biological process terms, Table S4). Those upregulated DEGs both in H24 vs. H0 and G24 vs. G0 groups showed significant enrichment in plant-pathogen interaction, sphingolipid metabolism, and phenylalanine metabolism pathways (Figure 2A,B); while the linoleic acid metabolism pathway was only markedly enriched in upregulated DEGs of H24 vs. H0 groups ( Figure 2A). However, those downregulated DEGs both in H24 vs. H0 and G24 vs. G0 groups showed significant enrichment in photosynthesis-antenna proteins, steroid biosynthesis, fatty acid elongation, starch, and sucrose metabolism, and sesquiterpenoid and triterpenoid biosynthesis showed significant enrichment ( Figure 2C,D). KEGG-enriched analysis showed that 36 DEGs in H24 vs. H0 groups and 22 DEGs in G24 vs. G0 groups belong to plant hormone signal transduction pathways. Those DEGs are involved in the signaling pathways of auxin (AUX/IAA), cytokinin (AHP and A-ARR), abscisic acid (PP2C and SnPK2), ethylene (ETR, EIN2, EIN3, and ERF1/2), and salicylic acid (NPR1, PR-1) ( Figures S2  and S3). Overall, 66 DEGs in H24 vs. H0 groups and 40 DEGs in G24 vs. G0 groups were identified in the plant-fungus interaction pathway. Among them, PPRs (CEBiP), FLS2, MKK1/2, WRKY33, pti1, NHO1, and PR1 were induced by Bgt in the H24 vs. H0 groups, and CaMCML, MKK1/2, WRKY33, NHO1, and PR1 were induced by Bgt in the G24 vs. H0 groups, which might be associated with the PAMP-triggered immunity. Moreover, RPS2 and HSP90 were induced by Bgt in H24 vs. H0 groups, and RPM1, RPS,2 and HSP90 genes were induced in response to Bgt in G24 vs. G0 groups, which may be involved in effector-triggered immunity ( Figures S4 and S5).

Changes in Expression Patterns of Glycerolipid Metabolism and Fatty Acid Biosynthesis Pathways in Response to Bgt Infection
We analyzed the DEGs of glycerolipid metabolism and fatty acid biosynthesis pathways, which responded to Bgt infection in H24 vs. H0. We found that 13 genes from 265 genes which were involved in glycerolipid metabolism pathways were regulated by Bgt infection (Figure 3, Tables S5  and S6). Among them, three TaGLI1 homologous genes (2.7.1.30), identified as glycerol kinase from A, B, and D genomes of wheat, were upregulated by Bgt infection. In addition, 1.1.1.21 identified as an aldehyde reductase, which transfers D-glycer-aldehyde to glycerol, was also upregulated by Bgt infection. While, 2.3.1.15 (named as TaACT1) identified as glycerol-3-phosphate acyltransferase was downregulated by Bgt infection. Moreover, 11 DEGs from 265 genes which were involved in fatty acid biosynthesis pathways were regulated by Bgt infection (Figure 3, Tables S5 and S6). We found that three TaSSI2 genes (1.14.192) from A, B, and D wheat genome, and some upstream genes identified as FabI, FabZ, FabF, and FabD were induced by Bgt infection (Figure 4, Table S6). Moreover, those genes were also regulated by Bgt in G24 vs. G0 groups and showed the same expression pattern The closer the Q-value is to zero, the more significant the enrichment. The horizontal axis indicates the rich factor, which means that the ratio of the DEGs number and the number of genes are annotated in this pathway. The greater the rich factor, the greater degree of enrichment. The size of each circle indicates the number of DEGs in that pathway. The larger the circle, the greater number of DEGs.

Changes in Expression Patterns of Glycerolipid Metabolism and Fatty Acid Biosynthesis Pathways in Response to Bgt Infection
We analyzed the DEGs of glycerolipid metabolism and fatty acid biosynthesis pathways, which responded to Bgt infection in H24 vs. H0. We found that 13 genes from 265 genes which were involved in glycerolipid metabolism pathways were regulated by Bgt infection (Figure 3, Tables S5 and S6). Among them, three TaGLI1 homologous genes (2.7.1.30), identified as glycerol kinase from A, B, and D genomes of wheat, were upregulated by Bgt infection. In addition, 1.1.1.21 identified as an aldehyde reductase, which transfers D-glycer-aldehyde to glycerol, was also upregulated by Bgt infection. While, 2.3.1.15 (named as TaACT1) identified as glycerol-3-phosphate acyltransferase was downregulated by Bgt infection. Moreover, 11 DEGs from 265 genes which were involved in fatty acid biosynthesis pathways were regulated by Bgt infection (Figure 3, Tables S5 and S6). We found that three TaSSI2 genes (1.14.192) from A, B, and D wheat genome, and some upstream genes identified as FabI, FabZ, FabF, and FabD were induced by Bgt infection (Figure 4, Table S6). Moreover, those genes were also regulated by Bgt in G24 vs. G0 groups and showed the same expression pattern with H24 vs. H0 groups (Table S6). The regulation of those genes in response to Bgt infection might contribute to the accumulation of G3P and several fatty acids. Moreover, it was demonstrated in our previous study that the levels of G3P, OA18:1, linoleic acid (LA18:2), and α-linolenic acid (ALA18:3) are induced by Bgt infection in wheat leaves [23]. with H24 vs. H0 groups (Table S6). The regulation of those genes in response to Bgt infection might contribute to the accumulation of G3P and several fatty acids. Moreover, it was demonstrated in our previous study that the levels of G3P, OA18:1, linoleic acid (LA18:2), and α-linolenic acid (ALA18:3) are induced by Bgt infection in wheat leaves [23].

Expression Patterns in Response to Glycerol Application (G0 vs. H0)
A total of 414 DEGs (376 up, 38 down) were identified in G0 vs. H0 groups in response to glycerol. Among them, 77 upregulated DEGs and seven downregulated DEGs were overlapped with those identified in H24 vs. H0 and G24 vs. G0 groups ( Figure 1D, Tables S7 and S8). Those 414 DEGs were annotated in some GO terms such as cell killing, immune system process, and response to stimulus (Table S4). GO enrichment analysis revealed some important GO terms, including response to JA, response to SA, defense response to bacterium, wounding, lipid oxidation, growth, lateral root formation, oxylipin metabolic process, linoleate 9S-lipoxygenase activity, allene oxide synthase activity, linoleate 13S-lipoxygenase activity, and hydroperoxide dehydratase activity ( Figure 5A). KEGG pathway analysis revealed that 21.1% (19/90) of genes in the linoleic acid metabolism, 2.3% (3/131) genes in cutin, suberine, and wax biosynthesis, and 1.9% (5/265) genes in glycerolipid metabolism could be regulated by glycerol (Table S5). KEGGenrichment analysis showed that the most highly enriched pathways were linoleic acid metabolism and alpha-linolenic acid metabolism ( Figure 5B).
Glycerol-induced plant hormone signal transduction pathway was indicated by the upregulation of AUX/IAA, PP2C, and JAZ, annotated in the IAA, ABA, and JA pathways ( Figure S8).

Expression Patterns in Response to Glycerol Application (G0 vs. H0)
A total of 414 DEGs (376 up, 38 down) were identified in G0 vs. H0 groups in response to glycerol. Among them, 77 upregulated DEGs and seven downregulated DEGs were overlapped with those identified in H24 vs. H0 and G24 vs. G0 groups ( Figure 1D, Tables S7 and S8). Those 414 DEGs were annotated in some GO terms such as cell killing, immune system process, and response to stimulus (Table S4). GO enrichment analysis revealed some important GO terms, including response to JA, response to SA, defense response to bacterium, wounding, lipid oxidation, growth, lateral root formation, oxylipin metabolic process, linoleate 9S-lipoxygenase activity, allene oxide synthase activity, linoleate 13S-lipoxygenase activity, and hydroperoxide dehydratase activity ( Figure 5A). KEGG pathway analysis revealed that 21.1% (19/90) of genes in the linoleic acid metabolism, 2.3% (3/131) genes in cutin, suberine, and wax biosynthesis, and 1.9% (5/265) genes in glycerolipid metabolism could be regulated by glycerol (Table S5). KEGGenrichment analysis showed that the most highly enriched pathways were linoleic acid metabolism and alpha-linolenic acid metabolism ( Figure 5B).
In the plant hormone signal transduction pathway, SnPK2, and JAZ (two DEGs), were upregulated in the G24 pool compared with H24 pool ( Figure S11). PR-1 was downregulated in the G24 pool. In the plant-pathogen interaction pathway, HSP90 (two DEGs) and FLS2 were upregulated in the G24 pool ( Figure S12). Moreover, those genes encoding peroxygenase, chitinase 8, peroxidase, beta-glucosidase, UDP-glycosyltransferase, glucan endo-1,3-beta-glucosidase, and xylanase inhibitor were upregulated in the G24 pool (Tables S10 and S3). We found that most of AOC, LOX, and PR genes were also induced in the G0 pool compared with H0 pool (Table S10). It suggested that those genes were significantly changed by glycerol application (G24 vs. H24; G0 vs. H0) and might take part in the resistance pathways in the G24 group.

Plant Hormones Cross-Talk in Response to Glycerol Application
AUX/IAA, PP2C, and JAZ were induced by glycerol, which was annotated in the IAA, ABA, and JA pathways, indicating activation of plant hormone signal transduction pathways ( Figure S8). Moreover, the enriched GO terms in response to jasmonic acid and salicylic acid were identified in the G0 vs. H0 groups ( Figure 5A). These results suggest that glycerol was involved in the regulation of SA, JA, IAA, and ABA pathways. Quantification of plant hormones validated glycerol induction of SA and JA levels, and reduced IAA level, while ABA level showed no change (Figure 7).

Discussion
Glycerol-induced resistance was reported in different plants to different pathogens as a typical non-host resistance, with upregulation of several PR genes and ROS accumulation [15,[20][21][22]. ETI is the main mechanism for host resistance, with a race-specific resistance involving cell death. However, some overlapping defense signaling pathways or genes have been identified between non-host and host resistance, such as PR genes, oxidative burst-associated genes, and cell defense genes [27]. In this study, to better understand the mechanism, the whole transcriptome analysis of wheat plants infected by Bgt after glycerol or water treatment was used to reveal extensive alteration in gene expression. Our results showed that glycerol induced some defense-related pathways or genes which associated with ROS accumulation, cell death, and PR genes, which might contribute to resistance to Bgt in wheat ( Figure 8). PR11), peroxidase (PR9), and oxalate oxidase (PR15 and PR16), and those PR genes can be used as potential candidate genes for improvement of the pathogen resistance of wheat and barley [44]. For example, PR genes encoding hydrolytic enzymes chitinases and β-1,3-glucanases, are very important in plants for invading pathogens, and overexpression of some PRs genes improves the resistance to different pathogens in different plants [45,46]. These results suggest that up-expression of those PRs by glycerol application might contribute to wheat resistance, and those PRs could be used as candidate genes for improving wheat resistance. Figure 8. The glycerol-induced resistance pathway. Upon Bgt infection, the TaGLI1 and TaSSI2 were induced, and TaACT1 was reduced, which might contribute to G3P and OA18:1 accumulation. G3P accumulation might be an active defense process; OA18:1 accumulation might be not good for plant resistance. Exogenous application of glycerol increased the G3P levels and decreased the level of OA18:1. Glycerol induced some genes involved in JA synthesis (e.g., LOX, AOS, and OPRs), then induced the JA level. Glycerol reduced IAA and induced SA levels with cross-talk of plant hormones. Figure 8. The glycerol-induced resistance pathway. Upon Bgt infection, the TaGLI1 and TaSSI2 were induced, and TaACT1 was reduced, which might contribute to G3P and OA18:1 accumulation. G3P accumulation might be an active defense process; OA18:1 accumulation might be not good for plant resistance. Exogenous application of glycerol increased the G3P levels and decreased the level of OA18:1. Glycerol induced some genes involved in JA synthesis (e.g., LOX, AOS, and OPRs), then induced the JA level. Glycerol reduced IAA and induced SA levels with cross-talk of plant hormones. Glycerol induced ROS accumulation, PR and Hsp90 genes expression, and cell death, then contributed resistance to Bgt in wheat.
We identified genes involved in the accumulation of G3P and OA18:1, including the upregulation of three TaGLI1 and three TaSSI2 homologs and downregulation of TaACT1, in response to Bgt infection (Figures 3 and 4; Table S5). Previous reports show that the expression of SSI2 and GLI1 genes are induced by powdery mildew both in wheat and barley [23]. G3P and OA18:1 are important signal molecules involved in plant resistance. Mutants defective in G3P synthesis are compromised in systemic acquired resistance (SAR), while exogenous G3P induces SAR [9,13]. Overexpressing GLI1 increases G3P levels and enhances resistance against bacterial disease in Arabidopsis [11]. Hence, the regulation of genes in the glycerol metabolism pathway, TaGLI1 and TaACT1, which was followed by Bgt infection, might contribute to G3P accumulation and induction of disease resistance ( Figure 3). However, numerous studies showed that downregulation of OA18:1 improves plant resistance [14,15,28]. For example, the ssi2 mutants or RNAi-mediated knockdown of SSI2 show enhanced resistance to different pathogens, with a lower OA18:1 level. Our previous results showed that OA18:1, LA18:2, and ALA18:3 levels were induced by Bgt infection [23]. It was demonstrated that unsaturated fatty acids (OA18:1, LA18:2, and ALA18:3) can inhibit innate immunity responses related to callose deposition in wheat [29]. Plant fatty acids can also be transferred to the pathogenic fungus and are required for colonization by pathogens as a fungal organic carbon source [19]. In this situation, the role of upregulated TaSSI2, FabI, FabZ, FabF, and FabD, following Bgt infection, might not contribute to wheat resistance ( Figure 4). In addition, lipids encompassing fatty acids, fatty acid-based polymers, and fatty acid derivatives are part of the membrane structures of cells and tissues. Upon microbe infestation, the membrane is modified and membrane lipid defense-signaling molecules, including free fatty acids, oxylipins, JA, and the potent second messenger phosphatidic acid are released [30]. However, the role of membrane lipid is still unclear for disease resistance in plants, and the impact on membrane lipid after glycerol application also needs further research.
JA is derived from the fatty acid ALA18:3, and the first step is catalyzed by lipoxygenases (LOXs), as the formation of FA hydroperoxides, then formation of the catalyzed reaction by the AOS, AOC, and OPR3 [31,32]. In our analysis of glycerol application, linolenic acid and alpha-linolenic acid (ALA18:3) metabolism were the two most enriched KEGG pathways ( Figure 5B). The upregulations of LOX2S, LOX1_5, AOS, and OPRs by glycerol application from those two pathways, might contribute to JA accumulation. Further support for these results is provided in the current study by quantification of plant hormones that showed JA accumulation induced by glycerol treatment. However, JA is involved in the defense against necrotrophic pathogens, preventing plant cell death, and inducing defense responses to restrict further pathogen infection [33]. Treatment with JA is shown to protect plants against herbivore attack and reduce the severity of infection by necrotrophic fungi [34,35]; while the application of MeJA does not induce resistance to powdery mildew (biotrophic fungi) in wheat [36]. Furthermore, we found the JASMONATE ZIM DOMAIN (JAZ) in JA signal transduction pathway was induced by glycerol. JAZ is a key target of microbes in their effort to suppress JA-dependent signaling [37]. Previous results suggest that tomato JAZ proteins regulate the progression of cell death during host and non-host interactions with different bacterial strains [38]. So, the upregulation of JAZ by glycerol might contribute to resistance to Bgt in wheat.
Furthermore, we found that glycerol also reduced IAA and induced SA levels. IAA induces the expression of expansins, proteins that loosen the cell wall, which is key for plant growth but may also make the plant vulnerable to biotic intruders. Exogenous application of IAA aggravates some pathogenic diseases progression [39,40]. Auxin-resistant Arabidopsis by mutation of genes functioning in auxin signaling shows enhanced disease resistance [41]. Previous reports showed that Aux/IAAs repressor proteins confer an improved disease resistance against pathogens, regulate the transcript levels of PRs, ROS accumulation, and callose development in the plant defense response, and act as transcriptional repressors of ARF genes to regulate downstream auxin-regulated genes [41][42][43]. Thus, the low level of IAA and upregulated Aux/IAAs in response to glycerol may improve plant wheat resistance to Bgt. SA is recognized as a very important SAR signal molecule, hence, the upregulation of SA by glycerol may also contribute to plant resistance to Bgt. SA causes global repression of auxin-related genes, including the TIR1 receptor gene, resulting in stabilization of the Aux/IAA repressor proteins and inhibition of auxin responses; the inhibitory effect on auxin signaling may be a part of the SA-mediated disease resistance mechanism [41]. Moreover, the amount of ABA was not changed after glycerol application, and other plant hormones-related pathways were not significantly changed with transcriptome analysis. Altogether, our analysis indicated that glycerol application altered plant hormone biosynthesis, signaling, and levels involved in Bgt resistance mechanisms.
Some pathogenesis related genes were induced by glycerol application such as encoding, PR10, PR1, chitinase-1, chitinase-8, peroxidase, glucan endo-1,3-beta-glucosidase, beta-glucosidase, UDP-glycosyltransferase, pathogenesis-related protein PRMS, disease resistance protein RPM1, heat shock 70 kDa protein, and heat shock 90 kDa protein. Previous reports showed that the downstream genes of SAR induce the expression of some PR protein genes, such as PR1, PR2, chitinase (PR3, PR8, and PR11), peroxidase (PR9), and oxalate oxidase (PR15 and PR16), and those PR genes can be used as potential candidate genes for improvement of the pathogen resistance of wheat and barley [44]. For example, PR genes encoding hydrolytic enzymes chitinases and β-1,3-glucanases, are very important in plants for invading pathogens, and overexpression of some PRs genes improves the resistance to different pathogens in different plants [45,46]. These results suggest that up-expression of those PRs by glycerol application might contribute to wheat resistance, and those PRs could be used as candidate genes for improving wheat resistance.

Plant and Fungal Materials
The Bgt susceptible bread wheat cultivar Xuezao was used in this experiment. Powdery mildew isolate E09 was provided by Xiayu Duan, Institute of Plant Protection, Chinese Academy of Agricultural Sciences.

Plants Growth and Glycerol Treatments
Plants potted and maintained in a growth chamber with humidity of 75%, 26/20 • C day/night temperature regime, 12:12 h light/dark cycle, and light intensity of 40 µmol m −2 s −1 . Wheat leaves at seedling stage (about one-week-old) were sprayed with 3% glycerol (containing 0.02% Silwett L-77) when the first leaves were fully expanded. Control plants were sprayed with water, containing 0.02% Silwett L-77. Glycerol and water treatments were preformed two times (once a day) until liquid was dropping from the leaves, to make sure sufficient treatments. Each treatment included three pots of plants, and each pot (12 × 12 cm) included 16 plants.

Pathogen Maintenance and Inoculation
Isolate E09 was maintained on the susceptible wheat line Xuezao through weekly transfer to new plants in the growth chamber with the same controls as described in Section 4.2. The fresh conidia were collected on the dark paper and papered before inoculation. After one day of glycerol or water treatments, wheat plants were put into a vaccination tower; conidia were slowly blown into the vaccination tower by using a blower with an estimated density of 100-150 conidia/mm 2 .

Coomassie Blue Staining
For microscopic observations of fungal development, Bgt-infected leaf segments were collected at 24 h post powdery mildew infection (hpi) for Coomassie blue staining as described previously [23]. Wheat leaf segments (5 cm in length) were stored in 50% glycerol and examined under an Olympus BX-43 microscope (Olympus Corporation, Tokyo, Japan). Three plants in each treatment and three leaves per plant were used for Comassie staining and the germination/penetration rates were calculated. The germination/penetration rates of conidiospores (number of germinated spores or penetrating spores relative to the total number of spores) were calculated and presented as the means of three independent replicated experiments. Statistical significance was determined by paired Student's t test.
4.5. RNA Extraction, cDNA Library Construction, RNA-Seq, and Data Analysis Four groups of wheat plants were prepared for RNA-seq: H0 and G0 groups represent water or glycerol-treated plants without Bgt infection; H24 and G24 groups represent water or glycerol-treated plants infected with Bgt one day after glycerol or water treatments. Each group included three biological repetitions. Ten leaves were pooled together at 24 hpi as one biological repetition for RNA extraction. The leaf samples of H0 and G0 groups were also collected at the same time point as controls of H24 and G24 groups at 24 hpi. The total RNA was extracted using RNA pure Plant Kit (Tiangen Biotech, Beijing, China). Sequencing libraries were generated using NEB Next UltraTM RNA Library Prep Kit for Illumina (NEB, Beijing, China) following manufacturer's recommendations. The parried-end reads were generated on Illumina Hiseq 2500 platform.
Sequencing data were analyzed by using BMKCloud. Trimmomatic was used to remove the adaptors. TopHat2 tools software were used to map and assemble the reads to reference genome IWGSC RefSeq v1.0, [47]. Only reads with a perfect match or one mismatch were further analyzed and annotated based on the reference genome. Gene function was annotated based on the following databases: Nr (NCBI non-redundant protein sequences); Nt (NCBI non-redundant nucleotide sequences); Pfam (protein family); KOG/COG (clusters of orthologous groups of proteins); Swiss-Prot (a manually annotated and reviewed protein sequence database); KO (KEGG ortholog database); GO (Gene Ontology). Differential expression analysis of four groups was performed using the DESeq R package (1.10.1). Genes with p-value ≤ 0.05 were assigned as differentially expressed, with |log2 fold change| ≥1.8. GO enrichment analysis of the differentially expressed genes (DEGs) was implemented by the GOseq R packages based on the Wallenius non-central hyper-geometric distribution, which can adjust for gene length bias in DEGs [48]. KOBAS software was used to test the statistical enrichment of differential expression genes in KEGG pathways [49].

Measurements of Endogenous Plant Hormones
The leaf tissues of both the water treatment (control) and glycerol treatment were collected at the seeding stages one day after glycerol and water treatments. Plant hormones (ABA, IAA, JA, and SA) were extracted and measured from leaf tissue (approximately 50 mg of fresh weight) by simplified icELISA method as described previously [50], slightly modified by using a different antibody which was developed in the Wang Baomin's lab, China Agricultural University.

Concluding Remarks
Bgt infection might contribute to G3P and OA18:1 accumulation in wheat. Glycerol application could improve G3P levels and reduced OA18:1 levels. Furthermore, glycerol altered plant hormones signaling pathways, which were confirmed by increase in JA and SA and reduction of IAA levels. Moreover, glycerol induced the expression of many PR proteins. These glycerol-induced changes probably contributed to wheat resistance to Bgt (Figure 8).
Author Contributions: C.X., Y.L. and Q.S. conceived the project. Y.L. and L.Q. performed most of the experiments. X.L. provided help for using software and analysis of RNA-seq data. Q.Z. and X.Z. provided data curation. Y.L. wrote the manuscript. T.F., T.K. and C.X. improved and revised this manuscript. All authors have read and agreed to the published version of the manuscript.