The wheat genome sequences IWGSC RefSeq v2.1 were downloaded from the International Wheat Genome Sequencing Association website (https://wheat-urgi.versailles.inra.fr/Seq-Repository/Assemblies) (Alaux et al., 2018). Firstly, the PAL domain sequence (Pfam accessions: PF00221) was downloaded from the Pfam database (http://pfam.xfam.org/), and were used as queries to search the PAL proteins in wheat genome by HMMER3.0 with cutoff values ≤ 1e-5 (Eddy, 2008; Finn et al., 2016). Secondly, the protein sequences of four Arabidopsis PALs (AtPALs), 18 maize PALs (ZmPALs), and nine rice PALs (OsPALs) were retrieved from the Arabidopsis Information Resource (TAIR10) database (http://www.arabidopsis.org/index.jsp), the Maize Genetics and Genomics Database (MaizeGDB) (https://www.maizegdb.org/), and the Rice Genome Annotation Project (RGAP) database (http://rice.plantbiology.msu.edu/), respectively (Huang et al., 2010; Zeng et al., 2018; Deng et al., 2019). Subsequently, the above mentioned 31 PAL protein sequences were used as queries to search wheat genome data through BLASTp, and the expected cutoff value was set as e ≤ 1e-5 to ensure the reliability of the protein sequence. After merging two rounds of search results and deleting redundant sequences, unique ones were further validated through SMART (http://smart.emblheidelberg.de/) and InterProScan (v71.0, http://www.ebi.ac.uk/InterProScan) to determine whether sequences contain PF00221 domain.

The protein characteristics of the TaPALs, including protein length, isoelectric point (pI), molecular weight (MW), instability index, atomic composition, grand average hydropathicity (GRAVY), and amino acid composition, were analyzed using ExPASy Server10 (SIB Bioinformatics Resource Portal, https://prosite.expasy.org/PS50011). The online tools TMHMM (http://www.cbs.dtu.dk/services/TMHMM/) and SignalP4.1 (http://www.cbs.dtu.dk/services/SignalP/) were used to predict the transmembrane domains and signal peptides in TaPALs (Nielsen, 2017). Subcellular localization prediction of TaPALs were performed using Plant-mPLoc online tool (http://www.csbio.sjtu.edu.cn/cgi-bin/PlantmPLoc.cgi) (Chou and Shen, 2010).

After collecting the protein sequences of four AtPALs, 18 ZmPALs, nine OsPALs, and TaPALs, DNAMAN (version 6.0) was used to align all amino acid sequences. The phylogenetic tree was constructed using the Maximum Likelihood (ML) method (1000 bootstrap trials) based on an LG model in MEGA7.0 software (Kumar et al., 2016). A midpoint rooted base tree was modified using the Interactive Tree of Life (IToL, version 5.5.1, http://itol.embl.de). The TaPALs were classified basing on the phylogenetic relationships.

The TaPAL gene structure was exhibited using the online software GSDS V2.0 (http://gsds.cbi.pku.edu.cn/index.php) according to the wheat genome annotation information (Hu et al., 2014). Conserved motifs of TaPALs were predicted using the online program MEME V 5.1.1 (Multiple Expectation Maximization for Motif Elicitation, http://meme-suite.org/tools/meme) (Bailey et al., 2015). The parameters were as follows: each sequence could comprise any number of non-overlapping occurrences of each motif, the number of different motifs was 20, and motif length ranged from 6 to 50 amino acids (Fang et al., 2019). The conserved domains of these predicted motifs were analyzed using InterPro (http://www.ebi.ac.uk/interpro) and SMART (http://coot.embl-heidelberg.de/SMART). The PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) was used to predict cis-acting elements in the regions 1,500 bp upstream of the start codons of TaPALs (Lescot et al., 2002). TBtools software was used for combing and visualizing the above results (Chen et al., 2020).

The gene annotation of TaPALs was extracted from wheat genome GFF3 file, and the chromosomal location was drawn with MapInspect software (Jiang et al., 2019). Gene duplications were divided into tandem and segmental replications. Tandem repeated events were identified using the following assessment criteria: (1) aligned sequence length > 80% regions of each sequence; (2) identity > 80%; (3) threshold ≤ 10⁻¹⁰; (4) only one duplication can be admitted when genes are linked closely; and (5) intergenic distance is less than 25 kb. If genes met the three conditions (1), (2) and (3) and were located on different chromosomes, they were judged as segmental duplications (He et al., 2020). The Ka (non-synonymous)/Ks (synonymous) ratio were calculated after identification of a duplicated gene, and the selection pressure and selection mode were analyzed. The CDS sequences of the duplicated gene pairs were compared using MEGA 7.0 software. After removal of gaps the alignment results were imported into DnaSP v5.10 software for Ks value and Ka/Ks ratio analyses (Librado and Rozas, 2009) . For estimation of the timing of duplication events the formula T = Ks/2λ × 10^-6 Mya was used to calculate divergence time (T) in millions of years (Mya), where λ = 6.5 × 10^-9 represented the rate of replacement of each locus per herb plant year (Quraishi et al., 2011; Peng et al., 2012). Duplicated gene pairs between species were identified and used to carry out an inter-species synthetic analysis using R package “circlize” (Krzywinski et al., 2009; Song et al., 2019).

The genome sequence data and the annotation information of Triticum urartu (v1.43), Triticum dicoccoides, and Aegilops Tauschii (v4.0.43) were downloaded from Ensembl Plants database (http://plant.ensembl.org/index.html) (Bai et al., 2019). In order to better understand the origin of TaPAL genes, The PAL family genes in T. urartu, T. dicoccoides, and A. Tauschii were identified by the Blastp analysis. The cutoff values (e-value < 10^-10, identity > 80%) were used to ensure the reliability of homologous (Li et al., 2022) . The original definition of an ortholog was two genes from two different species that were derived from a single gene of the last common ancestor of that species. Paralogs were defined as genes derived from a single gene that was repeated within the genome (Zhang et al., 2019). The synteny relationship were displayed with R package “circlize” and TBtools software (Zhu et al., 2015; Chen et al., 2020).

Original data sets, including different development stages and tissues from variable treatments, were downloaded from the NCBI Short Read Archive (SRA) database and mapped to the wheat reference genome using Hisat2 (Li et al., 2022). The SRA numbers of those data were listed in Supplementary file 1 . Cufflinks was used to calculate the expression levels of TaPALs (normalized by Transcripts Per Kilobase of exon model per Million mapped reads, TPM). The R package “pheatmap” was used to produce the heatmap.

The hexaploid common wheat cultivars Jingshuang 16 and Yangmai 158, were moderately sensitive to Puccinia striiformis f. sp. tritici and F. graminearum, respectively. Full-size seeds were surface disinfected with 1% hydrogen peroxide, and thoroughly rinsed with distilled water, then germinated in 25°C saturated water on three-layer filter paper (Zhang et al., 2019). The growth conditions were 25/20°C temperature and 16 h/8 h (day/night) photoperiod. The wild type strain of F. graminearum (PH-1) was cultivated on PDA plate and cultured at 25°C for 3 days. And then, the actively growing mycelial agar plugs (5 mm diameter) were transferred to a 150 mL conical flask containing 50 mL of liquid mung bean medium, and was cultured at 25°C with 150rpm for 3 days to produce conidia (Cao et al., 2016). After the coleoptile reached to 0.5 cm, the seedlings of Yangmai 158 were immersed in spore suspension (5 × 10⁵ spores/mL) for 1 minute (Yang et al., 2010). Then the seedlings were wrapped in a wet filter paper and grown at 25°C with 70% relative humidity. The uninoculated wheat seedlings were used as control. The wheat stems were sampled at 6, 12, 24, and 48 h after inoculation.

The P. striiformis was remained on the susceptible cultivar Jingshuang 16. At two-leaf stage, the seedings were sprayed with fresh sporangium powder of CYR23 (4 mg/mL) and kept in dark place for 24 hours with moisture. The seedlings were cultured in 10-18°C light for 16 h/8 h (day/night) photoperiod. Wheat leaves were sampled at 6, 12, 24, and 48 h after inoculation. The uninoculated wheat was used as a control. The samples were immediately frozen liquid nitrogen and stored at -80°C. Each treatment included three technical replications and three biotic replicates in each replication.

After analyzing the transcriptome data, five TaPAL genes (TaPAL10, TaPAL14, TaPAL32, TaPAL42, TaPAL49) with high expression levels were screened out. The expression levels of these five genes under F. graminearum and P. striiformis infection were analyzed by qRT-PCR. Total RNAs were reverse transcribed into cDNAs with a 5 × All-In-One RT Master Mix (Perfect Real Time) kit (ABM, Canada). Gene-specific primers were designed using Primer Premier 5.0 and listed in Supplementary file 1 . For qRT-PCR assays, the cDNA was diluted into 100 ng/μL with ddH₂O. The 20 μL reaction volume contained 10 μL of 2 × SYBR Green Mix, 1 μL of each primer (10 μM), 2 μL of template, and 6 μL of ddH₂O. All quantitative real-time PCR (qRT-PCR) amplifications were performed on the iCycler iQ instrument (Biorad, Hercules, CA, USA). The following cycling parameters were used: initial denaturation at 95°C for 2min; 50 amplification cycles consisting of denaturation at 95°C for 15s, annealing and extension at 57°C for 45 s. Single-fragment amplification was verified by dissociation curve analysis. Three biotic replicates were performed for each sample, and each replicate contained three technical replicates. ADP-ribosylation factor Ta2291 was used as internal reference gene (Paolacci et al., 2009). Relative expression levels were calculated using the 2^-ΔΔCt method (Livak and Schmittgen, 2001) . The heatmap were plotted using Origin software (Moberly et al., 2018).

To verify the prediction results of TaPAL proteins localization, the CDS region of TaPAL32 and TaPAL42 were amplified using cDNA as template. The vector pART27:GFP was digested with XhoI (Biomarker, China) and then purified by Cycle-pure Kit (Omega, UK). The full-length gene fragments were inserted into the purified pART27:GFP vector using ClonExpress II one-step cloning kit (Vazyme, China). Then the recombinant vector pART27:TaPAL32-GFP and pART27:TaPAL42-GFP were transformed into Agrobacterium tumefaciens GV3101. The leaves from 6 to 8 week-old Nicotiana benthamiana were injected with A. tumefaciens. After 48 h, the GFP signals were detected and were observed under fluorescence microscope (Nikon DS-Ri2, Japan).

The spring wheat cultivar Fielder was used in this experiment. Approximately 120bp of gene fragments of TaPAL32 and TaPAL42 were amplified from wheat cDNA using the primers listed in Supplementary file 1 . Then the fragments were inserted into the BSMV:γ vector. The virus was inoculated on the second leaf of the two-leaf wheat seedlings by mechanical friction method and then placed in an artificial incubator for 25°C cultured. BSMV : TaPDS (TaPDS, wheat octahydrolysin desaturase) was used as a positive control, and 1× Fes buffer (0.1M glycine, 0.06M K2HPO4, 1% w/v tetrasodium pyrophosphate, 1% w/v bentonite, and 1% w/v celite, pH 8.5) treated plants were used as negative control (Mock) (Zhu et al., 2018). At 10 dpi, the leaves with mosaic symptoms were inoculated with fresh uredospores of P. striiformis (CYR23). Inoculated leaves were collected at 0, 24, and 48 h after inoculation to determine the efficiency of silencing. The experiment was repeated at least three times. Relative expression levels were determined using the 2^–ΔΔCt method (Livak and Schmittgen, 2001).

Firstly, 571 wheat candidate PAL proteins were obtained by Blastp analysis. Then, 65 candidate PAL proteins were screened in the wheat genome through SMART and InterProScan. Finally, the results were combined to remove redundant sequences, and 54 PAL members were identified in wheat genome. These TaPAL proteins ranged from 265-714 amino acids in length, with predicted molecular weights ranging from 29.80 kDa to 77.34 kDa ( Supplementary file 2 ). The isoelectric points of 51 TaPAL proteins were less than 7. Only the pI values of TaPAL14, TaPAL26, and TaPAL43 had were greater than 7. The grand average of the hydrophilicity (GRAVY) of all TaPALs were less than zero, suggesting that they were almost all hydrophilic proteins. The proteins stability prediction results showed that all TaPALs were unstable proteins. Subcellular localization predicted results showed that all TaPALs were located in the cytoplasm.

To understand the evolutionary relationships of these 54 TaPAL proteins, a maximum likelihood phylogenetic tree was built based on all PAL proteins in wheat, maize, rice, Arabidopsis ( Supplementary file 3 ). the phylogenetic tree showed that all PAL proteins were divied into four Groups (Group I, II, III, and IV) ( Figure 1 ), which was consistent with the previous classification (Lepelley et al., 2012; Yan et al., 2019). Overall, Group I contained the largest number of TaPALs (26) and Group III had the lowest number of TaPALs (four).

In order to understand gene structure of TaPALs, we analyzed wheat GFF3 format annotations. The results showed that these genes contain no more than two introns, and 12 TaPAL genes were intronless. More interestingly, most of these genes were clustered into Group II and III ( Figure 2 ). In Group I, TaPAL1 and TaPAL11 contained two introns, while the remaining members had only one intron. Compared with genes in the other three groups, the exon-intron structures of genes in Group IV were more complicated. TaPAL39 had no intron, three TaPAL genes (TaPAL37, TaPAL43, and TaPAL49) had two introns, and the remaining members had only one intron.

In order to study the conserved regions of TaPAL proteins, the motifs of 54 TaPAL proteins were analyzed by MEME online software. Twenty conservative motifs were identified ( Figure 2 and Supplementary file 4 ). Eight of these motifs (motif1-2, 4-6, and 10-12) were found to be associated with characteristic functional domains of the typical plant PAL protein. Motif 1 was mostly present in most of TaPAL proteins. Analysis of conserved motifs showed that TaPAL protein domains were distributed in almost all of the 54 members.

Due to the high sequence homology of PAL in plants, the protein sequences of 54 TaPAL were aligned. As shown in Figure 2 , most TaPAL proteins had conserved enzyme active sites in the PAL-HAL and MIO domains, except for the nine core domains of TaPAL16, TaPAL22, TaPAL28, TaPAL37 to 41, and TaPAL43. It was worth noting that TaPAL16 and TaPAL28 lacked the core domains of PAL-HAL and MIO. In the MIO region, two conserved residues, Ala and Gly, were found to be conserved in all TaPAL proteins containing this domain. Although the PAL-HAL domain was conserved among 54 TaPALs, Gly residues had been replaced by Cys and His in TaPAL22 and TaPAL43, respectively. These analyses indicated that most TaPAL proteins may have enzymatic activity, but their activities were different.

Chromosome mapping showed that 54 TaPAL genes were unevenly distributed on 16 chromosomes of common wheat ( Figure 3 and Supplementary file 5 ). TaPALs were mainly distributed on chromosomes 2A, 2B and 2D, which contained eight, eight and seven genes, respectively. It was the least distributed on chromosomes 3A, 4A and 5A, and there was only one gene on each chromosome. In order to adapt to different environmental conditions, tandem replication and fragment replication were necessary for the evolution of gene families, and duplicate events of TaPALs need to be detected. As shown in Figure 4 and Supplementary file 6 , 29 pairs of tandem duplication genes and 115 pairs of segment duplication genes were identified in 54 TaPALs, and most interestingly, the tandem repeat genes were only found in Group I and II.

In order to further infer homology of wheat PAL family, a total of 91 PALs were identified from T. aestivum (45 TaPALs), T. urartu (9 TuPALs), T. dicoccoides (26 TdPALs) and Ae. tauschii (11 AePALs) by a computer-based method ( Figure 4 ). After statistics, there were 393 pairs of orthologous gene pairs, 78 orthologous gene pairs were identified from T. aestivum and Ae. tauschii, 33 orthologous gene pairs were identified from T. aestivum and T. dicoccoides, and 66 orthologous gene pairs were identified from T. aestivum and T. urartu. In addition, T. aestivum, T. urartu, T. dicoccoides and Ae. tauschii had 146, 7, 25 and 6 pairs of paralogous genes respectively.

We further analyzed the Ka (nonsynonymous)/Ks (synonymous) ratio of replicated PAL genes pairs in wheat to understand the evolutionary constraints on TaPALs ( Figure 4 and Supplementary file 7 ). In general, Ka/Ks rates can be used to evaluate the selectivity of coding sequences; the Ka/Ks ratio >1 indicates accelerated evolution with positive selection, a ratio =  1 indicates neutral selection, and a ratio < 1 indicates negative or purifying selection (Zhang et al., 2006; Peterson and Masel, 2009). The Ka/Ks ratios of 144 replicated gene pairs were calculated. The Ka/Ks ratios for all these TaPALs duplicated gene pairs were less than 1, which indicated that all repeated genes of the TaPAL family have undergone strong purifying selections which may help maintain their functional stability.

The cis-regulatory element was specific motif that can be combined with appropriate transcription factors to further regulate the transcription of genes in plants (Wittkopp and Kalay, 2011). To identify putative cis-elements in the TaPAL gene promoter region, we analyzed the 1500 bp upstream promoter region of TaPAL genes. We found that a total of 2292 cis-acting elements were predicted, of which 1422 regulatory-elements were related to growth and development, 485 regulatory elements were related to biotic and abiotic stress, and 385 active-elements were related to phytohormone ( Figure 5 and Supplementary file 8 ). As shown in Figure 5 , growth-related cis-regulatory elements include zein metabolism regulation (O2-site), seed-specific regulation (RY-element), etc. Furthermore, cis-regulatory elements related to stress included light responsive elements (ACE, GT1-motif, AE-box, ATCT-motif, and Box 4), enhancer-like element involved in anoxic specific inducibility (GC-motif), low-temperature responsiveness (LTR). The remaining cis-elements were related to phytohormone response, included cis-acting element involved in the abscisic acid responsiveness (ABRE); gibberellin-responsive element (GARE-motif and P-box). We found that CAAT box and TATA box were the most common cis-acting elements in TaPALs. Meanwhile, many cis-acting elements were related to plant abscisic acid (ABRE), which was distributed in the promoter region of most of TaPALs. Cis-element analysis showed that TaPAL family members may be play an important role in transcriptional regulation and abscisic acid metabolism pathway.

RNA-seq is a powerful tool to explore gene transcription patterns using high-throughput sequencing methods (Wang et al., 2010). Transcriptome data of wheat inoculated with F. graminearum, P. striiformis, and powdery mildew were used for analysis. In the expression profile ( Figure 6 ), 15 TaPALs were not expressed and five TaPALs (TaPAL47, TaPAL48, TaPAL49, TaPAL50, TaPAL53) were generally highly expressed in various periods. The expression levels of 31 TaPALs were up-regulated and the expression level of TaPAL52 was down-regulated after inoculation with F. graminearum. The expression levels of 11 TaPALs were up-regulated and 14 TaPALs were down-regulated after inoculation with P. striiformis. Moreover, the expression levels of 32 TaPALs were higher in powdery mildew treated leaves than that in the control group, while the expression levels of TaPAL18, TaPAL25 and TaPAL27 were lower than that in the control group. These results indicated that TaPALs had different expression patterns under different disease treatments. Under the three stress conditions, the expression levels of ten TaPALs (TaPAL3, TaPAL10, TaPAL13, TaPAL14, TaPAL17, TaPAL30, TaPAL31, TaPAL32, TaPAL42, TaPAL48) were highly induced. It was inferred that these genes may be involved in the regulation of plant disease resistance.

In order to further explore the potential role of five genes (TaPAL10, TaPAL14, TaPAL32, TaPAL42, TaPAL49) under biotic stresses (F. graminearum and P. striiformis), we analyzed their expression levels by qRT-PCR. As shown in Figure 7 , the expression levels of five genes were continuously up-regulated from 12 h to 48 h under F. graminearum treatment, among which only the expression levels of TaPAL32 and TaPAL42 were significantly up-regulated from 0 h to 6 h. In wheat leaves treated with P. striiformis, the expression levels of TaPAL10, TaPAL14, TaPAL32, and TaPAL42 showed a general downward trend. However, the expression level of TaPAL49 at 48 h was much higher than that at 24 h, and the expression level was gradually up-regulated during infection.

The predicted results from plant-MPLOC online tool showed that TaPAL proteins were located in the cytoplasm. To further confirm the predicted results, TaPAL32-GFP and TaPAL42-GFP fusion protein were transiently expressed in N. benthamiana leaves. As shown in Figure 8 , both of TaPAL32 and TaPAL42 were expressed in the whole cell.

We further detected the function of TaPAL32 and TaPAL42 in the wheat cultivar Fielder by barley stripe mosaic virus (BSMV)-mediated virus-induced gene silencing (VIGS) assay. The silenced vectors of two genes (BSMV : TaPAL32, BSMV : TaPAL42) were constructed. About 10 days after BSMV treatment, BSMV : TaPDS leaves showed symptoms of photobleaching, indicating that the virus-induced gene silencing system was working. Then, fresh urediospores of P. striiformis isolates CYR23 was inoculated on the fourth leaves of silenced plants. The expression levels of TaPAL32 and TaPAL42 at 0, 24, and 48 h after inoculation were analyzed by qRT-PCR. Compared with the control (BSMV:γ), the gene expression levels of TAPAL32 and TaPAL42 were significantly down-regulated, and the disease severity was photographed at 14 days after inoculation. The results showed that there were many urediospores on the leaves of silent plants compared with the control group ( Figure 9 ). It indicated that silencing of these two genes weakened the wheat defense responses against P. striiformis.