Genome-wide identi cation and characterization of caffeoyl-coenzyme A O-methyltransferase genes related to Fusarium head blight response in wheat

Guang Yang State Key Laboratory of Crop Stress Biology in Arid Areas, College of Agronomy and Yangling Branch of China Wheat Improvement Center, Northwest A&F University, Yangling 712100, Shaanxi, China Wenqiu Pan State Key Laboratory of Crop Stress Biology in Arid Areas, College of Agronomy and Yangling Branch of China Wheat Improvement Center, Northwest A&F University, Yangling 712100, Shaanxi, China Ruoyu Zhang State Key Laboratory of Crop Stress Biology in Arid Areas, College of Agronomy and Yangling Branch of China Wheat Improvement Center, Northwest A&F University, Yangling 712100, Shaanxi, China Yan Pan State Key Laboratory of Crop Stress Biology in Arid Areas, College of Agronomy and Yangling Branch of China Wheat Improvement Center, Northwest A&F University, Yangling 712100, Shaanxi, China Qifan Guo State Key Laboratory of Crop Stress Biology in Arid Areas, College of Agronomy and Yangling Branch of China Wheat Improvement Center, Northwest A&F University, Yangling 712100, Shaanxi, China Weining Song State Key Laboratory of Crop Stress Biology in Arid Areas, College of Agronomy and Yangling Branch of China Wheat Improvement Center, Northwest A&F University, Yangling 712100, Shaanxi, China Weijun Zheng State Key Laboratory of Crop Stress Biology in Arid Areas, College of Agronomy and Yangling Branch of China Wheat Improvement Center, Northwest A&F University, Yangling 712100, Shaanxi, China Xiaojun Nie (  small@nwsuaf.edu.cn ) State Key Laboratory of Crop Stress Biology in Arid Areas, College of Agronomy and Yangling Branch of China Wheat Improvement Center, Northwest A&F University, Yangling 712100, Shaanxi, China


Abstract Background
Lignin is one of the main components of cell wall, which directly associates with the development and defense mechanisms in plants, especially in response to Fusarium head blight (FHB) tolerance. Caffeoyl-coenzyme A Omethyltransferase (CCoAOMT) is the main regulator determining the e ciency of lignin synthesis and composition. Although it has been widely characterized in many plants, the importance of CCoAOMT family in wheat is not well understood up to now.

Results
Here, a total 21 CCoAOMT genes were identi ed in wheat (TaCCoAOMT) through a in silico genome search method and they were classi ed into four groups based on phylogenetic analysis with the members in the same group sharing similar gene structures and conserved motif compositions. Furthermore, the expression patterns and co-expression network which these TaCCoAOMT involved in were comprehensively investigated using 48 RNA-seq samples from Fusarium graminearum-infected and control samples of 4 wheat genotypes. Combined with qRT-PCR validation of 11 Fg-responsive TaCCoAOMT genes, the potential candidates involving in FHB response and their regulation modules were preliminarily revealed. Additionally, we also investigated the genetic diversity and main haplotypes of these CCoAOMT genes in bread wheat and its relative populations based on resequencing data.

Conclusion
This study systematically identi ed and characterized the CCoAOMT gene family in wheat, which not only provided the targets for further functional analysis, but also contribute to the mechanism of lignin biosynthesis and its role in FHB tolerance in wheat and beyond.

Background
Wheat is considered as one of the most important staple crops all over the world, which accounts for approximately 30% of the global cultivated area, and provides 20% of the world's food consumption [1,2]. Wheat is also an important source of human protein and mineral elements intake [3,4]. Continuous increased and stable production of wheat holds the promise for ensuring global food security under the challenge of population booming and climate change as well as limited resource input in future [5]. Fusarium head blight (FHB), that is also called scab and caused mainly by Fusarium graminearum (Fg), is one of the most destructive diseases of wheat, resulting in huge loss of wheat yield and also imposing great health threats on both human beings and livestock due to the DON toxin [6,7]. More seriously, FHB has gradually become the major hazard and limitation of wheat production in recent years because of the climate change and the expansion of conservation agriculture [8]. Thus, revealing the mechanism of FHB resistance and then breeding for FHB-tolerant wheat varieties are crucial to cope with these problems. Cell wall was mainly composed of polysaccharides, phenolic compounds and proteins. In plant cells, cell wall always acts mechanical and regulatory functions [9].
Extensive studies had reported that the components of cell wall endow plants with the ability to resist the invasion of pathogens [10,11]. For example, Giancaspro et al reported the putative markers of FHB resistance through analyzing the expression pattern of pectin methylesterase WheatPME-1 and β, 1-3 Glucanase (Glu-1), which involved in cell wall metabolism as well as regulated the non-speci c lipid transfer protein (nsLTP-1) [10]. Lionetti et al showed that PMEIs could dynamically regulate the PME activity and pectin methylesteri cation during Botrytis infection, and AtPMEI10, AtPMEI11 and AtPMEI12 were veri ed to play the signi cant roles in cell wall metabolism to enhance plant immunity [12].
Lignin is the second most abundant component of the plant cell wall. It has been demonstrated that the dehydrogenative polymerization of three hydroxycinnamyl alcohols, p-coumaryl, coniferyl, and sinapyl alcohols contributed to lignin biosynthesis, of which these three hydroxycinnamyl alcohols give rise to the p-hydroxyphenyl (H), guaiacyl (G), and syringyl (S) units of the lignin polymer, respectively [13]. And O-Methyltransferases (OMTs) play an important role in regulating these secondary metabolic processes invovling in lignin biosynthesis. OMTs are generally classi ed into two types, including caffeic acid O-methyltransferase (COMT) and caffeoyl-coenzyme A O-methyltransferase (CCoAOMT), of which COMT controls the pathway of S unit and CCoAOMT affects the pathway of S and G unit [14]. It is found that COMT could catalyze the O-methylation at the 5 position of the aromatic ring and CCoAOMT function to form the 3 position of the aromatic ring [15,16]. The hydroxylation and methylation steps are crucial to determine the lignin composition and also the S/G ratio is a major determinant of lignin quality [14].
In light of its signi cance, CCoAOMT gene family has been systematically investigated and analyzed in many plants, such as Arabidopsis and rice [17], citrus [18], switchgrass [19], dove tree [20], tea plant [21] and sorghum [22]. In wheat, Nguyen et al analyzed the expression patterns and potential functions of some genes involving in lignin biosynthesis including several CCoAOMTs, and indicated that lodging resistance, tolerance against biotic and abiotic stresses and feedstock quality of wheat biomass were closely associated with its lignin content [23]. Ma and Luo veri ed that TaCCoAOMT1 was an important gene for regulating lignin biosynthesis, which is critical for stem development [24]. Soni et al silenced the function of TaNAC032 via VIGS method and found that the total lignin content in the rachis of TaNAC032 silenced wheat decreased signi cantly and the trangenic plant showed susceptible to Fg infection [25], which demonstrated the role of lignin in FHB tolerance.
Although some CCoAOMT gene has been functionally characterized in wheat, the genomic organization, evolutionary relationship and regulatory modules of wheat CCoAOMT gene family are not well understood up to now, especially those assocaited with FHB tolerance. In this study, we performed a in silico genome-wide search method to identify and characterize CCoAOMT family in wheat using the updated reference genome information. Furthermore, the phylogenetic relationship, conserved motif and cis-elements of them were systematically analyzed. Furthermore, the expression patterns and co-expression network of TaCCoAOMT genes under Fg treatment were studied and FHB-responsive TaCCoAOMTs as well as the regulation modules were obtained. Finally, the genetic diversity and genetic divergence of CCoAOMT genes in different Triticum species were also investigated based on resequencing data to reveal the evolutionary effect of this family during wheat formation.

Results
Genome-wide identi cation of CCoAOMT genes in wheat Using the genome-wide search method described in Methods section, a total of 21 CCoAOMT genes were detected in the wheat genome. Since there is no standard nomenclature, these identi ed CCoAOMT genes were named as TaCCoAOMT1 to TaCCoAOMT21 based on their chromosome location ( Table 1). It is found that these TaCCoAOMT genes were mainly located in chromosome group 7 (66.67%) while not found on chromosome group 2, 3 and 6. The gene size of TaCCoAOMT genes ranged from 447 (TaCCoAOMT17) to 4567 (TaCCoAOMT12) bp in length. The average lengths of CDS and amino acid sequences were 761 bp and 253 aa, respectively. Isoelectric point (pI) of them ranged from 4.89 (TaCCoAOMT15) to 11.11 (TaCCoAOMT17) and and molecular weight (Mw) ranged from 16284.2 (TaCCoAOMT17) and 34158.48 (TaCCoAOMT21), respectively. A search for orthologs of TaCCoAOMT genes revealed that 20 (95.24%) TaCCoAOMTs had the orthologs with Arabidopsis and rice, expect for TaCCoAOMT6. Subcellular localization prediction showed that most of them were located in the cytoplasmic and chloroplast, and only one gene was located in the mitochondrial and nuclear, respectively. In these 21 TaCCoAOMT genes, we also found four homoeologous gene groups with each contained A, B and D homoeologous copies, and all of them were localized on chromosome group 7, resulting in chromosome group 7 having the most abundant TaCCoAOMT genes.
Phylogenetic relationship, exon-intron structure and conserved motifs analysis Phylogenetic tree were constructed using the full-length protein sequences of the CCoAOMT genes in wheat, Arabidopsis and rice ( Figure 1). Result showed they were classi ed into 4 groups based on phylogenetic relationship, with 4, 7, 3 and 7 TaCCoAOMT genes belonging to class I to IV, respectively. TaCCoAOMT and OsCCoAOMT genes were distributed in all of groups. However, AtCCoAOMT genes were mainly clustered in class III and just two genes (AtCCoAOMT5 and AtCCoAOMT6) in class I.
Then, the exon-intron and motif structures of TaCCoAOMTs were further analyzed ( Figure 2). Exon number of TaCCoAOMTs ranged from 1 to 10, of which two genes contained only one exon, and 85.71% genes had 5 exon or less ( Figure 2B). Furthermore, the conserved motif was identi ed and 10 high con dence motifs were predicted ( Figure 2C and Figure S1). Compared to other classes, Motif 6 and 9 was the speci c type in class I and III, respectively. Although few differences of motif were found between Class II and III, the exon number of TaCCoAOMT genes in class II was more than that of class III. Motif 2 was identi ed in 20 (95.24%) TaCCoAOMTs, with the exception of TaCCoAOMT2. Motif 3 was also identi ed in 20 TaCCoAOMTs with the exception of TaCCoAOMT17. In addition, with the exception of TaCCoAOMT5 and TaCCoAOMT17, Motif 1 was identi ed in 19 TaCCoAOMT genes, and Motif 4 was found in 19 TaCCoAOMT genes excepted for TaCCoAOMT2 and TaCCoAOMT17. This result showed that Motif 1, 2, 3, 4 was signi cantly abundant in TaCCoAOMTs, and all of them was found to be related to O-methyltransferase based on PFAM analysis, which further supported the prediction.Meanwhile, the members in the same groups based on phylogenetic relationship shared the similar exon-intron structure and motif compositions.

Cis-element analysis of TaCCoAOMTs
A total 44 types of cis-elements were identi ed in the 1.5-kb genomic sequences upstream from the transcription start sites (TSS) of TaCCoAOMT genes, with the functions primarily associating with stress response (Table S1 and Figure S2). CAAT-box was identi ed in all of TaCCoAOMTs (21), followed by CGTCA-motif (19) and TGACG-motif (19). Together with CGTCA motif and TGACG motif, ABRE related to the abscisic acid responsiveness was identi ed in 18 TaCCoAOMT genes. In addition, CCAAT-box (MYBHv1 binding site), MBS (MYB binding site involved in droughtinducibility), TCA-element (cis-acting element involved in salicylic acid responsiveness), TC-rich repeats (cis-acting element involved in defense and stress responsiveness), MRE (MYB binding site involved in light responsiveness), SARE (cis-acting element involved in salicylic acid responsiveness) and WUN-motif (wound-responsive element) also were identi ed in 13, 7, 7, 2, 1, 1 and 1 TaCCoAOMT genes, respectively.
In order to get some clues about the biological function of TaCCoAOMTs, we performed GO (gene ontology) enrichment analysis of them as the all wheat protein as background. Results showed that they was signi cantly enriched into 10, 10 and 1 terms in biological process, molecular function and cell component classes, respectively ( Figure S3 and Table S2) The spatio-temporal expression patterns of TaCCoAOMTs were investigated using 58 RNA-seq samples of different tissues as well as under low temperature, salt, heat and drought stresses. A total of 20 TaCCoAOMTs were found to express in different tissues and some gene exhibited tissuespeci c expression. Totally, 3 (TaCCoAOMT13, 14, 18) genes were highly expressed in all of the ve differential tissues ( Figure 3A). However, TaCCoAOMT1, 2, 3 were only expressed in leaf and TaCCoAOMT5 was only expressed in root. Most of the TaCCoAOMTs showed high expression in grain and leaf compared to root, spike and stem.
At low temperature treatments ( Figure 3B), we found that the expression level of eight genes was higher at 23 ℃ than that of 4 ℃, indicating these genes may be more responsive to low temperature stress. After salt stress treatments ( Figure 3C), 13 out of 21 TaCCoAOMT genes were downregulated in 6 h, 8 and 11 genes were down-regulated in 12 h and 24 h respectively, indicating their negative effects related to salt stress. Under drought stress ( Figure 3D), 8 and 5 TaCCoAOMT genes were up-regulated in two time points respectively, four genes (TaCCoAOMT7, 8, 10, 11) were up-regulated at both time points. However, 9 genes were down-regulated at both time points, indicating the negative effects of TaCCoAOMTs exerted when in response to drought stress.

Co-expression network and regulation module of FHB-responsive TaCCoAOMTs
To better understand the function and regulation network of the identi ed FHB-responsive TaCCoAOMTs, we further constructed the WGCNA coexpression network based on these RNA-seq data. Through constructing weighted correlation network, 34 co-expression modules were obtained ( Figure 6). Then, we linked the co-expression modules with the available phenotypic data of the Fg inoculation, including Percentage (Fusarium oxysporum inoculum), DON (Deoxynivalenol), GAPDH (Glyceraldehyde-3-phosphate dehydrogenase content) and inoculation time [27]. The module-trait association results showed that saddlebrown, blue, greenyellow, lightcyan and green module were highly correlated with Fg inoculation ( Figure 6A). Meanwhile, tan, darkolivegreen, blue and greenyellow were highly correlated with Fg percentage, DON and GAPDH. These important modules also were correlated with the time of Fg inoculation. Interestingly, 2, 5, 1, 1 and 2 FHB-responsive TaCCoAOMTs were found in blue, green, lightyellow, turquoise and yellow module, respectively. Furthermore, there modules also showed the strong correlation and associated with Fg inoculation ( Figure 6B). A total of 93 genes were associated by TaCCoAOMTs in the modules to show similar expression patterns. GO enrichment analysis of the 93 genes found they mainly enriched into the terms related to defense response, such as GO:0052544 (defense response by callose deposition in cell wall), GO:0010294 (abscisic acid glucosyltransferase activity) and GO:0000165 (MAPK cascade). It is obvious that the CCOAOMT genes located at the hub nodes of these modules. Among them, TaCCoAOMT19, which associated with FHB tolerant, could interact with TaCCoAOMT13 to regulate other 19 wheat genes to form a regulation module (Figure 7). This module might played the potentially important role in FHB tolerance, providing the useful molecular module for FHB resistance improvement.
It is showed that TaCCoAOMT genes were mainly inhibited by miRNAs through transcript cleavage (86.27%). Combined with miRNA-TaCCoAOMT relationship and co-regulation modules TaCCoAOMT involved in, we further obtaiend the miRNA-mediated networks associated with FHB response and resistance that were mainly regulated by CCoAOMT genes in wheat (Figure 7), which provided the useful clues to regulate the expression of TaCCoAOMTs to control lignin biosynthesis and then enhance FHB tolerance through post-transcriptional approach.

Genetic diversity and haplotype analysis of CCoAOMT family in wheat and its relatives
Based on the resequnecing data of Triticum species [28], the genetic variations of CCoAOMT genes in wheat and its diploid and tetraploid relatives were investigated, including the nucleotide diversity (π), population divergence (Fst) and Tajimas'D values. The average value of π in Triticum urartu, Aegilops tasuschii, wild emmer, domesticated emmer, durum wheat and bread wheat was 0.000565, 0.00535, 0.00297, 0.00321, 0.00320 and 0.00233, respectively ( Figure 8A), together with the average value of Tajimas'D of them was -0.834, 1.094, 0.145,0.370, 0.292 and -0.137, respectively ( Figure 8B). Due to too few SNPs identi ed from the re-sequencing data, Triticum urartu showed the abnormally lower value of nucleotide diversity. It is found that Aegilops tasuschii displayed the highest nucleotide diversity in CCoAOMT genes, while bread wheat had the most lower nucleotide diversity except for T. urartu with the value decreased by 2 times, suggesting that signi cant genetic bottleneck occurred in CCoAOMT gene family during wheat evolution. Then, the gene ow and genetic divergence between bread wheat and its relatives were also detected.. In A subgenome, the Fst value between bread wheat and T. urartu was 0.556, ranked the largest, followed by that of wild emmer and T. urartu with the value of 0.480 as well as bread wheat and wild emmer with the value of 0.248, indicating the high divergence of bread wheat with T. urartu,compared to wild emmer at A subgneome level from the perspective of CCoAOMT gene family ( Figure 8C). In B subgenome, the divergence between bread wheat and wild emmer was larger than that of durum and wild emmer, and bread wheat showed closer to durum wheat compared to wild emmer wheat ( Figure 8D). In D subgenome, the Fst value of between bread wheat and Aegilops tasuschii was 0.604 ( Figure 8E). On the whole, the genetic divergence at D subgenome was highest, followed by A subgenome and B.
The samples of resequencing data contained 30 Aegilops tauschii samples, 29 Triticum urartu samples, 28 wild emmer (Triticum turgidum L. ssp. dicoccoides) samples, 29 domesticated emmer (Triticum turgidum L. ssp. dicoccon) samples, 13 durum (Triticum turgidum L. ssp. durum) samples and 163 bread wheat (Triticum aestivum L. ssp. aestivum) samples. Finally, we identi ed the haplotype organization and frequency of each TaCCoAOMT genes in these populations based on the resequencing data (Table S4). A total 13 TaCCoAOMT genes were found to have the genetic variations among these populations, of which 4, 2, 7 were located on A, B, D subgenome, respectively. Then, the main haplotype and its frequencies were investigated ( Figure S4, S5 and S6). It is obviously that the percentage of main haplotype in cultivated wheat was signi cantly larger than that of wild species at all subgenome levels, indicating the arti cal selection exerted on these CCoAOMT genes to result in the decline of genetic diverity and genetic bottleneck during wheat domestication and improvement processes. .

Discussion
Lignin was the main component of cell wall and involved in response to the abiotic and biotic stresses [15]. The characteristic of CCoAOMT genes has been analyzed in Arabidopsis, sorghum, and other plants. However, its signi cance in wheat is not determined, especially its role in Fusarium head blight resistance is not well studied. In this study, we identi ed 21 CCoAOMT in wheat at the whole genome level. Based on the phylogenetic relationship, we divided these TaCCoAOMTs into four groups, and the TaCCoAOMTs belonging to the same group showed similar gene structure and motif organization. In sorghum, the CCoAOMT proteins were classi ed in clade 1a, clade 1b, clade 1c and clade 2, of which clade 1a was the orthologous genes with AtCCoAOMT1 and OsCCoAOMT1 and was considered as the true CCoAOMT gene, while the other classes were considered as CCoAOMT-like genes [22]. In our results, AtCCoAOMT1 and OsCCoAOMT1 also were cluster into same group (class III), which was consistent with that of sorghum. Meanwhile, CCoAOMT genes in rice and wheat can be found in each group, but AtCCoAOMTs just were found in class I and III, indicating that there was some divergence between the monocot and dicot CCoAOMT family. Additionally, class I and III has the speci c motif 6 and 9, respectively. Although no speci c motif found in class II and IV, their gene structure displayed the difference with the obvious difference of exon number, suggesting more type of splice variants or binding-sites may be existing in these two classes.
Based on RNA-seq samples, the expression level of these wheat CCoAOMT genes in different tissue, abiotic and biotic stresses were comprehensively analyzed. A total of 19 TaCCoAOMT genes were found to express in leaves, followed by 17 in roots and 16 in grains. Among them, TaCCoAOMT13, 14 and 18 expressed in ve tissues, while TaCCoAOMT1, 2, 3 was only expressed in leaf and TaCCoAOMT5 only in root. These tissue-constitutive and speci c TaCCoAOMT genes provided the useful targets for further functional study. Most of TaCCoAOMT genes were highly expressed in grains, indicating the lignin biosynthesis in grains were more active than other tissues. Simultaneously, 10 TaCCoAOMT showed signi cantly up-regulated expression after Fg treatment compared to the control, of which 3 (TaCCoAOMT10, 14 and 19) genes were shared in three resistant varieties and one susceptible variety, suggesting they played the important role in response to FHB. Interestingly, TaCCoAOMT10, 14 and 19 were the A, B and D homoeologues copies of the same homoeologues group respectively, of which the expression level of TaCCoAOMT19 was the highest, indicating the dominant role of them in response to FHB and asymmetry expression between homoeologues copies. It is well known that cis-regulatory elements can regulate gene expression level by binding to corresponding transcription factors, and then might determine the speci c expression patterns in different tissue and stresses [29]. CGTCA-motif and TGACG-motif, which is related to MeJAresponsiveness, were identi ed in the promoter regions of all of the up-regulated TaCCoAOMTs. Previous study has proved that MeJA can not only help the wheat to signi cantly delay the necrosis process of susceptible varieties, but also increase the activities of enzymes related to pathogen defense [30]. Therefore, these up-regulated TaCCoAOMTs containing CGTCA-motif and TGACG-motif might play the crucial role in regulating FHB tolerance through MeJA-mediated approach. We further re-constructed the co-expression by WGCNA method based on 48 RNA-seq samples of four wheat genotypes [27]. Results showed that there was a high correlation between seven modules and Fg infection. Meanwhile, TaCCoAOMT genes were found in ve co-expression modules and acted as the hub factors, of which two (blue, green) modules were positively correlated with Fg infection and three (lightyellow, turquoise, yellow) modules were negative. These molecular modules not only provided the vital insight into the genetic basis and regulation network of TaCCoAOMT involved in, but also contributed to reveal the their roles in FHB resistance. Furthermore, we predicted the miRNAs which could target on the TaCCoAOMT genes. And eight microRNAs were found to target on the ve TaCCoAOMT genes in the Fg infection related modules, including tae-miR167a and tae-miR1119. Previous studies had been reported that miR167a could meditate auxin signaling to respond to biotic stresses in tomato [31], and miR1119 was proven to regulate the expression of actin under stress conditions and activate plant defense signaling pathway in barley [32]. We postulated that these two miRNAs also play the regulatory role in the response to Fg infection through regulating on CCoAOMT genes in wheat. The miRNA-CCoAOMT mediated networks that associated with FHB response and resistance provided the insight to better understand the molecular mechanism underlying FHB resistance in wheat.
Finally, we used the resequencing data to investigate the genetic variations and divergence of CCoAOMT family in wheat and its relative populations. Results found that wild species showed high genetic diversity and rich haplotype composition in this family compared to cultivated species, suggesting selection effect was exerted on this family and obvious genetic bottleneck has occurred at it during wheat domestication and improvement processes [33]. Wild populations possessed the speci c haplotypes on the CCoAOMT genes which have been lost in cultivated populations, which hold the promising for enriching the genetic diversity and also improving the traits that CCoAOMT genes controlling in cultivated wheat, such as FHB resistance.

Conclusion
It is the rst study to identify the CCoAOMT family in wheat at the whole genome level. The genomic organization, phylogenetic relationship, exonintron structure, conserved motif composition and cis-elements as well as expression pro les of this family were systematically investigated and characterized. Especially, the expression patterns and co-expression network of these TaCCoAOMTs involved in Fg infection were comprehensively identi ed and the FHB-responsive candidates and their regulation modules were obtained, which provided the insight on the roles of TaCCoAOMTs playing in the FHB response and tolerance. Additionally, we also detected the genetic diversity and haplotype frequencies of these CCoAOMT genes in wheat and its relative populations based on resequencing data. This study not only shed light on the potential function of CCoAOMT family in regulating wheat lignan biosynthesis and FHB tolerance, but also provided the vital clues for the evolution of this family in wheat and beyond.

Identi cation of CCoAOMT genes in wheat
For the identi cation the CCoAOMT family in wheat, the protein sequences of the wheat genome were retrieved from the Ensembl plant database to use as the local protein database (http://plants.ensembl.org/index.html). CCoAOMT gene of Arabidopsis and rice [17] were used to perform BLASTP search against the local protein database with the threshold of E-value < 1e-5 [34]. The domain (PF01596) was downloaded from PFAM database (https://pfam.xfam.org/) and used as the query to search against local protein database using the HMMER 3.0 with the threshold of Evalue < 1e -5 . The results of HMMER and BLASTP were integrated together, and the redundant were manually removed. Then, the putative wheat CCoAOMT genes were submitted to SMART (http://smart.embl-heidelberg.de/) and PFAM database(PF01596) (http://pfam.xfam.org/search) to predic the conserved protein domain, and then those having complete CCoAOMT domain were remained as candidates. The candidate TaCCoAOMT proteins were submitted to ExPASy (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi) database to compute theoretical isoelectric point (pI) and molecular weight (Mw). The cello tool (http://cello.life.nctu.edu.tw/) was used to predict the subcellular localization.
Phylogenetic, gene structure, conservative motif and cis-element analysis of TdPUB Multiple sequence alignment was performed using ClustalX v2.0 [35]. Neighbor-joining method imbedded in MEGA-X program was used to construct the phylogenetic tree and bootstrap was set to 1000 [36]. Additionally, conserved motifs of TaCCoAOMT proteins were identi ed using MEME v5.2.0 with the default parameters. The gene and motif structures were displayed based on GTF annotation les using TBtools [37] The upstream 1500bp region of each TaCCoAOMT genes were extracted and submitted to the PlantCARE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) to predict cis-elements.
FHB related co-expression networks construction and miRNA analysis Co-expression network was constructed using the WGCNA tools based on the 48 RNA-seq samples of four genotype wheat under F. graminearum treatment. To obtain the correct module number and clarify gene interaction, we set the restricted minimum gene number to 30 for each module and used a threshold of 0.25 to merge the similar modules. Genes that have higher weight in important modules were chosen to constructed coexpression network. The traits data publically available , including Fg treatment, Fg time, Fg percent, Fg GAPDH (Glyceraldehyde-3-phosphate dehydrogenase) and DON(Deoxynivalenol) were used for trait-module correlation analysis [27]. miRNA binding sites were predicted using psRNATarget (http://plantgrn.noble.org/psRNATarget/analysis) with default parameters and all of the wheat miRNAs were used. The regulatory network of TaCCoAOMT gene and miRNA were visualized using Cytoscape v3.8.0 [41].
Haplotype and population genetics analysis of TdPUB VCF le of wheat resequencing were downloaded from Genome Variation Map (https://bigd.big.ac.cn/gvm) (accession no. GVM000082) [28], which contained the genome variations of a total of 163 bread wheat accessions, 13 durum wheat accessions, 29 domesticated emmer wheat accessions, 28 wild emmer wheat accessions, 29 T. urartu accessions and 30 Aegilops tauschii accessions. SNPs in the coding region of TaCCoAOMT genes were extracted based on the chromosome location using TBtool tool. Furthermore, the haplotype organization and frequency were investigated by an inhouse Python script.

Validation of the expression of TaCCoAOMTs through qRT-PCR analysis
For experiment veri cation, a FHB resistant RIL line (R75) and a FHB susceptible line (S98) from the wheat RIL population developed by single-seed descent from a cross between the susceptible US wheat variety Wheaton and the Chinese resistant wheat landrace HYZ, which were provided generously by Prof. Tao Li, Yangzhou University, China, are used. The plant material culture and Fg inoculation were performed following the previous described method by Li et al [42]. And the inoculated spikelet samples together with counterpart CK samples were collected from 4 to 5 spikes at 2 days post inoculation (dpi) and three biological replications were adopted. RNA Easy Fast Plant Tissue Kit (Tiangen, Beijing, China) was used to extract total RNA of all sample and RT Master Mix Perfect Real-Time kit (Takara, Dalian, China) was used to synthesized cDNAs according to the manufacturer's instruction. qRT-PCR reaction were performed on the QuantStudioTM 7 Flex System (Thermo Fisher Scienti c, USA) using SYBR® Green Premix Pro Taq HS qPCR Kit (Accurate Biology, Hunan, China) with the thermal cycling condition was 95℃ for 30 s followed by 40 cycles of 95℃ for 3s, 60℃ for 30 s. Three technological replications were applied. The expression levels of these 11 randomly selected TaCCoAOMTs were calculated using the 2−ΔΔCT method with TaActin2 as the internal reference gene. The primers used in this study were listed in additional le (Table S5)       Supplementary Tables   Table S1. Characteristics of cis-acting regulatory elements in the promoter region of CCoAOMT genes in wheat. Table S2. GO enrichment analysis of the identi ed TaCCoAOMTs. Table S3. Identi cation of miRNA bind sites in TaCCoAOMT genes.