Comparative analysis of the PAL gene family in nine citruses provides new insights into the stress resistance mechanism of Citrus species

Abstract

Background

The phenylalanine ammonia-lyase (PAL) gene, a well-studied plant defense gene, is crucial for growth, development, and stress resistance. The PAL gene family has been studied in many plants. Citrus is among the most vital cash crops worldwide. However, the PAL gene family has not been comprehensively studied in most Citrus species, and the biological functions and specific underlying mechanisms are unclear.

Results

We identified 41 PAL genes from nine Citrus species and revealed different patterns of evolution among the PAL genes in different Citrus species. Gene duplication was found to be a vital mechanism for the expansion of the PAL gene family in citrus. In addition, there was a strong correlation between the ability of PAL genes to respond to stress and their evolutionary duration in citrus. PAL genes with shorter evolutionary times were involved in more multiple stress responses, and these PAL genes with broad-spectrum resistance were all single-copy genes. By further integrating the lignin and flavonoid synthesis pathways in citrus, we observed that PAL genes contribute to the synthesis of lignin and flavonoids, which enhance the physical defense and ROS scavenging ability of citrus plants, thereby helping them withstand stress.

Conclusions

This study provides a comprehensive framework of the PAL gene family in citrus, and we propose a hypothetical model for the stress resistance mechanism in citrus. This study provides a foundation for further investigations into the biological functions of PAL genes in the growth, development, and response to various stresses in citrus.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-024-10938-3.

Introduction

Citrus is a common name for species of the genus Citrus in the family of Rutaceae (Rue family). Citrus contains many bioactive compounds, such as carotenoids, fibers, flavonoids, and phenolic compounds, which are highly sought by both the market and consumers [1]. Citrus, among the most widely cultivated fruit crops [2], is grown in more than 140 countries worldwide. According to the Food and Agriculture Organization of the United Nations (FAO, https://www.fao.org/faostat/zh/#data/QCL), the global citrus cultivation area in 2021 was 10,222.42 million hectares, with a production of 161.8 million tons, ranking first among all fruits [3]. At present, cultivated citrus mainly includes Citrus sinensis (Commonly known as sweet orange), Citrus reticulata (Mandarin orange), Citrus limon (Lemon), Citrus grandis (Pummelo), Citrus medica (Citron), Citrus clementina (Clementine), Citrus ichangensis (Ichang papeda), and Citrus australis (Australian lime), and so on.

After a long-term period of continuous domestication and selection, most citrus plants have shown good adaptability to climate, terrain, and various cultivation factors. However, in recent years, with the expansion of citrus cultivation areas and the accelerated deterioration of the ecological environment, various biotic and abiotic stresses, such as drought, low temperature, soil salinization, excessive soil metal ion content, canker disease, citrus huanglongbing (HLB), and other fungal or viral diseases, have caused significant losses to the global citrus industry [4]. Breeding for resistance in citrus has become a top priority in current citrus research. However, owing to the inherent biological characteristics of citrus, such as polyembryony and apomixis, traditional breeding has shown significant limitations in the breeding of new varieties with improved resistance. Genetic engineering, as a new technology for genetic breeding, can overcome the limitations of conventional breeding. With the continuous advancement of modern biotechnology, especially in recent years, genetic engineering technologies such as genetic transformation and gene editing have rapidly developed. Hence, genetic engineering has gained the attention of researchers as a means of improving the resistance of citrus.

The PAL gene, one of the most intensively studied plant defense genes in higher plants, is closely related to plant growth, development, and stress tolerance [5]. With the development of molecular biology techniques, the PAL gene family has been studied in detail in many species, such as Populus trichocarpa [6] (which includes 5 PAL genes.), Salix viminalis [7] (5), Coffea arabica [8] (3), Cephalotaxus hainanensis [9] (4), Arabidopsis thaliana [10] (4), Solanum lycopersicum [11] (14), Sorghum bicolor [12] (8), etc. The plant PAL gene family is highly conserved [3, 13], with the number of PAL genes ranging from three to 20 in most plants. Numerous studies have shown that woody plants generally have fewer PAL genes than herbaceous plants. The PAL enzyme (phenylalanine ammonia lyase, EC 4.3.1.24), encoded by the PAL gene, is one of the most intensively studied enzymes in plant secondary metabolism and is the most studied enzyme in the phenylalanine metabolic pathway [14, 15]. The PAL enzyme participates in the synthesis of secondary metabolites such as lignin, anthocyanins, flavonoids, and phenols; directly regulates plant physiological activities [9] and plays a significant role in plant stress responses [13]. When plants are under stress, substances such as anthocyanins, flavonoids, and phenols can help plants resist stress by clearing the reactive oxygen species (ROS) produced during plant metabolism [16]. Lignin can increase the thickness of plant cell walls to improve plant resistance to external pressures [17].

The functional analysis and utilization of various genes have greatly accelerated with the continuous exploration of an increasing number of plant genetic resources. The plant PAL genes have been identified, and their functions have been verified. In rice, the overexpression of the OsPAL6 and OsPAL8 genes can significantly increase the contents of lignin and SA, leading to increased resistance of rice to brown planthoppers [18]. Potato PAL genes may be involved in the defense against high temperature and drought stress [19]. The ScAPD1-like gene in Syntrichia caninervis enhances resistance against Verticillium dahliae by regulating the expression of PAL genes [20]. Deng et al. [21] reported that the expression level of the PAL gene in the peel of sweet oranges significantly increased after inoculation with Penicillium digitatum, Penicillium italicum, or Geotrichum citriaurantii. Wei et al. [3] demonstrated that overexpression of the PAL gene in sweet oranges could increase the tolerance of sweet oranges to green mold. Notably, research on PAL genes in Citrus species has been reported only in sweet oranges, whereas for other Citrus species, it has hardly been described.

In this study, the PAL gene families of nine Citrus species, namely, C. sinensis, C. reticulata, C. limon, C. grandis cv. ‘Wanbaiyou’, C. grandis cv. ‘Cupi Majiayou’, C. medica, C. clementina, C. ichangensis, and C. australis were comprehensively analyzed, including phylogenetic relationships, gene structure, cis-promoter elements, and gene evolutionary pressures. The response of PAL genes in different citrus plants to biotic or abiotic stresses was explored. Furthermore, to investigate how PAL genes enhance plant resistance, vital genes involved in lignin and flavonoid synthesis metabolism were identified, and their expression profiles were analyzed. This study provides a theoretical basis for further understanding the mechanisms by which citrus resists various biotic and abiotic stresses and serves as a reference for the genetic breeding of citrus.

Materials and methods

Data and materials used in this study

Genomic data: The genomic data of nine Citrus species, namely, C. sinensis, C. reticulata, C. limon, and C. grandis cv. ‘Wanbaiyou’, C. grandis cv. ‘Cupi Majiayou’, C. medica, C. clementina, C. ichangensis, and C. australis were obtained from two Citrus genome databases: the Citrus Pangenome to Breeding Database [22] (CPBD, http://citrus.hzau.edu.cn/) and the Citrus Genome Database (CGD, https://www.citrusgenomedb.org/) (see Supplementary Table S1 for details).

Conserved domain of the PAL gene family: The sequence of the conserved domain Lyase_aromatic [11] (ID: PF00221) of the PAL gene family was downloaded from the Pfam website (https://www.ebi.ac.uk/interpro/entry/pfam/#table).

A. thaliana PAL genes: The PAL gene family of A. thaliana was directly retrieved and downloaded from the TAIR database [23] (https://www.Arabidopsis.org/).

‘Bingtang’ sweet orange (C. Sinensis) was grafted on C. aurantium by Chu Orange Fruit Co., Ltd., located in Xinping, Yunnan Province, China. All the trees were vigorously grown in a greenhouse for one year. The fruits were collected in October 2021 when they matured commercially, and the fruits uniform in color and size without visual defects were used.

Identification of the PAL gene family in nine Citrus species

The PAL gene family identification can be divided into three main steps. In the first step, the BLASTP (Protein-protein Basic Local Alignment Search Tool) program of BLAST + software was used to align the protein sequences of the PAL gene family in A. thaliana as a reference with the protein sequences of nine Citrus species. The protein sequences of citrus that presented high homology (E-value < 1 × 10⁻⁵) [24] with the PAL gene family protein sequences of A. thaliana were filtered out. These sequences were subsequently used for identification. In the second step, the Hmmsearch program of HMMER software (version 3.3.2, http://www.hmmer.org/download.html) was used to search for protein sequences containing the conserved Lyase_aromatic domain (ID: PF00221) in the protein sequences of the nine citrus strains obtained in the first step. The protein sequences with an E value lower than 1 × 10⁻⁵ were screened from the results [25]. Thus, we obtained a preliminary PAL gene family information database for these nine Citrus species. In the third step, to further confirm the accuracy of the identification results, we also performed a conserved domain check on the protein sequences of the PAL gene family in the nine Citrus species via using the online tools CD-search (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi) and SMART (https://smart.embl.de/smart/batch.pl). Sequences that contained incomplete or none of the lyase-aromatic domains were removed. The remaining sequences are the final PAL gene-family identification results for the nine Citrus species. The proteins were renamed and used for subsequent analyses. For species at the chromosomal level of genome assembly, renaming is performed in ascending or descending order based on the distribution of the genes on the chromosome, and random renaming is performed at the scaffold level of genome assembly.

Physicochemical characterization and multiple sequence alignment of PAL proteins in nine Citrus species

The physicochemical properties of the proteins encoded by the PAL genes were predicted via the online tool ProtParam (https://web.ExPASy.org/protparam/) [26]. The online software Plant-mPLoc (http://www.csbio.sjtu.edu.cn/bioinf/plant-multi/) was used to indicate the subcellular localization of the proteins encoded by the PAL genes. MAFFT software (version 7.5, https://mafft.cbrc.jp/alignment/software/) was used for multiple sequence alignment of the amino acid sequences of the PAL proteins from the nine Citrus species. The alignment results and conserved regions were visualized via Geneious (version 9.1.4) and Jalview (version 2.11.2.7).

Construction and analysis of an interspecific phylogenetic tree for the PAL gene family

The PAL protein sequences obtained from the nine Citrus species were combined with the protein sequences encoded by the PAL genes from A. thaliana. Sequence alignment and maximum likelihood (ML) phylogenetic tree construction were performed following a method similar to that described by Xue et al. [27]. First, MAFFT software (version 7.5) was used to align the PAL proteins from the nine citrus and A. thaliana plants. The alignment products were subsequently trimmed via Trimal (version 1.4.1, http://trimal.cgenomics.org/downloads) to remove poorly conserved sites. The most suitable model for phylogenetic construction was selected via the ModelFinder tool of IQ-TREE software, a maximum likelihood tree-building software [28]. Finally, the phylogenetic tree was constructed via the tree visualization tool tvBOT [29] (https://www.chiplot.online/tvbot.html).

Gene structure and promoter cis-acting element analysis of the PAL gene family in nine Citrus species

On the basis of the identified PAL gene IDs, detailed annotation information for each gene was extracted from the nine citrus annotation files (gff3 files). The gene structure can be visualized based on its length and the positions of introns and exons via TBtools software [30]. The gene upstream 2000 bp nucleotide sequences were extracted from the genome using the gene IDs obtained from the identification. The obtained sequences were subsequently imported into the online promoter cis-acting regulatory element analysis software PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) for analysis of the promoter cis-acting elements [31].

Conserved motif analysis of PAL proteins in nine Citrus species

To further analyze the conserved domain composition of the PAL proteins in citrus, the online motif prediction tool MEME (https://meme-suite.org/meme/tools/meme) was used to predict the conserved motifs in the PAL proteins. The parameters for the conserved motif count were set to ten, with a motif width length range of six to 50 amino acids, and all the other parameters were set to default values [32]. The obtained conserved motifs were visualized via TBtools software.

Introduction to the evolutionary history of the PAL gene family in nine Citrus species

To investigate the intraspecific evolution of the PAL genes in nine Citrus species, we utilized the identified PAL gene sequences from these nine Citrus species, and we constructed intraspecific phylogenetic trees for the PAL genes via methods similar to those described in Sect. 3. Gene duplication is a significant mechanism for gene family expansion and plays a crucial role in the evolution of gene families. To understand the gene duplication events in the PAL gene family of Citrus, we employed the MCScanX [33] tool to analyze the duplicated genes within the PAL gene family of each of the nine Citrus species. Furthermore, to investigate the evolutionary pressures on the PAL genes within each species, we used two methods, including NG [34] and YN [35], to calculate the Ka/Ks values for duplicated PAL gene pairs in the nine Citrus species. First, we used the ParaAT (version 2.0, https://ngdc.cncb.ac.cn/tools/paraat) tool to align the duplicated PAL gene pairs in the nine Citrus species. We then employed the KaKs_Calculator (version 3.0, https://ngdc.cncb.ac.cn/biocode/tools/BT000001) to calculate the Ka/Ks values for each duplicated gene pair. Numerous studies have shown that Ka/Ks > 1 indicates positive selection pressure and rapid gene evolution, Ka/Ks = 1 represents neutral selection, and Ka/Ks < 1 indicates purifying selection pressure [36].

Analysis of the contraction and expansion of the PAL gene family in nine Citrus species

Gene family expansion and contraction are common phenomena that occur during genome evolution and play crucial roles in the adaptability and survival of organisms [37]. To further investigate the expansion and contraction of the PAL gene family, this study employed a method similar to that of Li et al. [38] to analyze the contraction and expansion of the PAL gene family. A phylogenetic tree of the nine Citrus species in the PAL gene family was constructed via a method similar to that described in section three. A species phylogenetic tree of the nine citruses was built based on the homologous single-copy genes shared by these species. First, OrthoFinder2 (version 2.5.4, https://github.com/davidemms/OrthoFinder) was used to identify the homologous single-copy genes shared by the nine citruses. Then, MAFFT was used to perform multiple sequence alignment for each homologous single-copy gene shared by the nine citruses. Seqkit (version 2.1.0, https://bioinf.shenwei.me/seqkit/download/) was used to classify and concatenate the alignment results of the homologous single-copy genes based on species, and a species phylogenetic tree was constructed. Finally, Notung (version 2.9.1.5, http://goby.compbio.cs.cmu.edu/Notung) was used to analyze the contraction and expansion of the PAL gene family in the nine Citrus species based on the species phylogenetic tree and the PAL gene family phylogenetic tree generated via parsimony-based optimization criteria. Additionally, this study estimated the divergence times of the nine Citrus species on the basis of the homologous single-copy genes shared by each species via the MCMCtree tool of PAMLX software (version 1.3.1, http://abacus.gene.ucl.ac.uk/software/paml.html).

Expression analysis of citrus PAL genes under biotic and abiotic stresses

To understand the expression patterns of PAL genes in citrus plants under various biotic and abiotic stresses, we collected and analyzed 17 transcriptome data points from different citrus plants under stress (transcriptome data sources are detailed in Table 1). During the transcriptome collection process, we did not find transcriptome data for C. grandis cv. ‘Cupi Majiayou’, C. clementina, C. ichangensis, or C. australis. The 17 transcriptome datasets used in this study were all from five Citrus species: C. sinensis, C. reticulata, C. limon, and C. grandis cv. ‘Wanbaiyou’, and C. medica. Among them, 15 transcriptome datasets can be downloaded from the NCBI-SRA database based on accession numbers (detailed in Table 1), and two transcriptome datasets related to the infestation of P. digitatum in sweet orange fruits and the infestation of Lecanicillium psalliotae in sweet orange leaves were obtained from experiments conducted by our research group.

Table 1.

Sources of transcriptome data

Species	Accession	Stress (Plant tissue)	CK/T		Time	Replicate	Source	References
C. sinensis	PRJNA428873	Aluminum (Leaf)	0 mM	CK	18 weeks	2	NCBI- SRA	[39]
			1.0 mM	T	18 weeks	2
	PRJNA339838	Aluminum (Root)	0 mM	CK	18 weeks	2	NCBI- SRA	[40]
			1.0 mM	T	18 weeks	2
	PRJNA792482	Drought (Leaf)	/	CK	0 day	3	NCBI- SRA	[41]
			/	T	20 days	3
	PRJNA547836	Mg deficient (Leaf)	1 mM	CK	16 weeks	2	NCBI- SRA	[42]
			0 mM	T	16 weeks	2
	PRJNA855348	P. digitatum (Fruit)	Water	CK	7 days	3	Research group experiments	/
			P. digitatum	T	7 days	3
	PRJNA981143	L. psalliotae (Leaf)	Water	CK	14 days	3	Research group experiments	/
			L. psalliotae	T	14 days	3
	PRJNA384780	Citrus tristeza virus (Leaf)	Healthy	CK	6 months	3	NCBI- SRA	[43]
			Citrus tristeza virus	T	6 months	3
C. reticulata	PRJNA623065	Low temperature (Leaf)	0 ℃	CK	0 day	3	NCBI- SRA	/
			0 ℃	T	5 days	3
	PRJNA734968	D. citri (Leaf)	Healthy	CK	14 days	5	NCBI- SRA	/
			D. citri	T	14 days	5
C. limon	PRJNA355134	Drought (Leaf)	/	CK	0 day	3	NCBI- SRA	[44]
			/	T	7 days	3
	PRJDB9486	Low temperature (Leaf)	25 ℃	CK	42 days	3	NCBI- SRA	[45]
			5 ℃	T	42 days	3
	PRJNA348468	Candidatus Liberibacter asiaticus (Leaf)	Healthy	CK	14 weeks	3	NCBI- SRA	[46]
			C. Liberibacter asiaticus	T	14 weeks	3
	PRJNA838230	Citrus yellow vein clearing virus (Leaf)	Healthy	CK	16 h	3	NCBI- SRA	[47]
			Citrus yellow vein clearing virus	T	16 h	3
C. grandis cv. ‘Wanbaiyou’	PRJNA428873	Aluminum (Leaf)	0 mM	CK	18 weeks	2	NCBI- SRA	[39]
			1.0 mM	T	18 weeks	2
	PRJNA339838	Aluminum (Root)	0 mM	CK	18 weeks	2	NCBI- SRA	[40]
			1.0 mM	T	18 weeks	2
	PRJNA702620	Cu-toxicity-3 (Leaf)	0.5 µM	CK	18 weeks	3	NCBI- SRA	[48]
			400 µM	T	18 weeks	3
C. medica	PRJNA688894	Plenodomus tracheiphilus (Leaf)	Water	CK	15 days	3	NCBI- SRA	[2]
			P. tracheiphilus	T	15 days	3

mM mol.l⁻¹, µM µmol.l⁻¹, CK control, T treatment

The transcriptome data of P. digitatum-infested sweet orange fruits were obtained as follows: Step one: P. digitatum was isolated and purified from diseased spots of sweet orange collected from Xinping, Yunnan Province, China, on potato dextrose agar (PDA) medium via the streak plate method. The purified P. digitatum was subsequently inoculated onto PDA media via the spread plate method and incubated in a constant-temperature incubator at 25 °C for three days. A spore suspension of P. digitatum was obtained by rinsing the Petri dish twice with sterilized water, which was subsequently collected in a centrifuge tube and shaken well. In step two, sweet orange fruits of uniform size with no mechanical damage or disease were disinfected with a 4% sodium hypochlorite solution. A hole (4 mm length × 4 mm width × 3 mm depth) was punched at four sites around the equator of each fruit, and 30 µL of spore suspension was injected into it as the treatment group. Treatments with an equal amount of sterilized water were used as the control. Both the treatment and control groups were treated with six sweet orange fruits. In step three, the inoculated fruits were allowed to rest on an ultraclean bench at 25 °C for seven days, and 3 cm of pericarp around the holes was collected and stored in liquid nitrogen, with three biological replicates set and then sent to Bioyi Biotechnology Company for transcriptome sequencing to obtain the transcriptome data of P. digitatum-infested sweet orange fruits. The RNA-Seq data were submitted to the NCBI-SRA (accession no. PRJNA855348).

The transcriptome data of L. psalliotae infested sweet orange leaves were obtained as follows: in step one, via the streak plate method, L. psalliotae was isolated and purified from diseased lesions of sweet orange collected from Xinping, Yunnan Province, China. The purified L. psalliotae was subsequently inoculated onto potato media via the plate coating method and incubated in a constant-temperature incubator at 25 °C for three days. After that, the Petri dishes were rinsed twice with sterilized distilled water, and the rinse solution was collected in a conical flask and shaken well to obtain a suspension of L. psalliotae spores. In step two, sweet orange plants with similar growth, no mechanical damage, and no disease were selected. The young leaves (the top five leaves of each plant) were cleaned and disinfected with a 4% sodium hypochlorite solution. Then, with sterilized toothpicks, wounds were made by puncturing the sweet orange leaves (six wounds per leaf). A suspension of L. psalliotae spores was sprayed onto the surface of the young leaves, which were then sealed in bags as the treatment group. Sterilized distilled water was served as the control treatment. Each treatment had three replicates. In step three, the treated plants were placed in a greenhouse for 14 days. Afterward, leaf samples around each wound (1 cm in diameter) were collected, preserved in liquid nitrogen, and sent to Baiyihui Biological Technology Co. Ltd. for transcriptome sequencing. Three biological replicates were set up for the treatment and control groups, resulting in transcriptome data for L. psalliotae invasion in sweet orange leaves. The RNA-Seq data have been submitted to the NCBI-GEO database (accession no. PRJNA981143).

To avoid the presence of many adapters and low-quality data in the transcriptome data, we used Fastp software (version 0.23.2, https://github.com/OpenGene/fastp) to remove adapters and filter out low-quality data, and FastaQC (version 0.12.0, https://www.bioinformatics.babraham.ac.uk/projects/fastqc) was used to check the quality of the data, ultimately ensuring that all the transcriptome data had a Q value > 30. Reads from transcripts that passed quality control were mapped to the reference genome via HISAT2 [49] (version 2.2.1). The FeatureCounts toolkit of Rsubread software [50] (version 2.12.0) was used to count reads mapped to those in the reference genome and calculate fragments per kilobase of exon model per million mapped fragments (FPKM) for each gene. The expression of each gene was quantified. To make the gene expression estimated from different experiments comparable, we chose the most common gene expression normalization method to calculate the TPM value of each gene, which can integrate the effects of sequencing depth and gene length on reading counts and normalize the expression of each gene.

In addition, we used DESeq2 software (version 1.4.0.2) to analyze PAL genes that were differentially expressed under different stresses. In this study, we defined genes with a differential expression fold change ≥ 2.00 and a P value ≤ 0.05 as significantly differentially expressed genes.

Expression analysis of genes upstream and downstream of the PAL gene in the lignin and flavonoid anabolic pathways in citrus

To further explore the expression patterns of genes involved in lignin and flavonoid synthesis and metabolism in citrus, this study followed a similar approach to that of Li et al. [38] to identify vital genes involved in lignin and flavonoid synthesis in the genomes of five Citrus species (C. sinensis, C. reticulata, C. limon, C. grandis cv. ‘Wanbaiyou’, C. medica) that underwent transcriptome analysis, as described in Sect. 8. The expression patterns of these genes under different stress conditions were explored via transcriptome data. First, we used a method similar to identifying the PAL gene family to identify 17 vital genes involved in lignin and flavonoid synthesis in C. sinensis, excluding PAL. These genes included PAT (bifunctional aspartate aminotransferase and glutamate/aspartate-prephenate aminotransferase), PDT (arogenate/prephenate dehydratase), C4H (trans-cinnamate 4-monooxygenase), 4CL (4-coumarate-CoA ligase), CCR (cinnamoyl-CoA reductase), CAD (cinnamyl-alcohol dehydrogenase), POD (peroxidase), HCT (shikimate O-hydroxycinnamoyl transferase), C3’H (5-O-(4-coumaroyl)-D-quinate 3’-monooxygenase), CCoAOMT (caffeoyl-CoAO-methyltransferase), F5H (ferulate-5-hydroxylase), COMT (caffeic acid 3-O-methyltransferase/acetylserotonin O-methyltransferase), ANS (anthocyanidin synthase), DFR (dihydroflavonol-4-reductase), CitF3H (naringenin 3-dioxygenase), CHI (chalcone isomerase), and CHS (chalcone synthase 1). Then, using the protein sequences of the 17 genes in sweet orange as references, the homologous protein sequences of the 17 genes were identified from the other four species via BLAST (E value < 1 × 10⁻⁵⁰, identity > 75%) as candidate protein sequences for each vital gene of each species. To further confirm the identification accuracy, the protein sequences of the genes involved in lignin and flavonoid synthesis in the five Citrus species were imported into the online tool CD-search for conservative domain verification. Sequences that had incomplete or lacked conserved domains were removed. On the basis of the identification results, we extracted the expression of each vital gene from the transcriptome data and constructed a heatmap of expression patterns via Pheatmap based on log₂^(TPM+1) values. In addition, differential gene identification was performed via the differential gene analysis method described in Sect. 8.

Results

Identification and physicochemical characterization of the PAL gene family in nine Citrus species

In this study, we integrated both Lyase_aromatic (ID: PF00221) conserved structural domain searches and BLAST searches of the genome using the PAL gene family protein sequences in A. thaliana as a reference. As a result, 41 PAL genes (Table 2; detailed sequences are provided in Supplementary Table S2) were identified from the genomes of nine Citrus species. These genes were renamed PAL1 to PAL5 based on their species classification. There was minimal variation in the number of PAL genes among the nine Citrus species, with four or five genes. For species with a genome assembly at the chromosome level (C. sinensis, C. grandis cv. ‘Wanbaiyou’, C. grandis cv. ‘Cupi Majiayou’, C. australis), the PAL genes were located on chromosomes 6, 7, and 8. Specifically, two PAL genes were on chromosome 6, one on chromosome 7, and the remaining on chromosome 8. The results of the physicochemical analysis (see Fig. 1, details in Supplementary Table S3) revealed that the protein sequence lengths encoded by the PAL genes in the nine Citrus species ranged from 714 aa to 1252 aa. The molecular weights ranged from 77.63 kDa to 139.49 kDa, with all the PAL genes except RLI_PAL4 from C. australis encoding proteins with molecular weights less than 90 kDa. The isoelectric points ranged from 5.86 (WBP_PAL4) to 6.76 (LEM_PAL3). The instability coefficients ranged from 28.95 (SWO_PAL2) to 36.76 (RLI_PAL4). The hydrophilicity index ranged from − 0.073 to − 0.333. Overall, the proteins encoded by the PAL family from the nine Citrus species were stable hydrophilic proteins. Subcellular localization prediction revealed that all the PAL genes from the nine Citrus species were located in the cytoplasm.

Table 2.

Identification of PAL family genes in nine Citrus species

Species	Num.	Gene
C. sinensis	4	ID	Cs_ont_6g020600	Cs_ont_6g020620	Cs_ont_7g006400	Cs_ont_8g005310
		Name	SWO_PAL1	SWO_PAL2	SWO_PAL3	SWO_PAL4
		Chr.	6	6	7	8
C. reticulata	4	ID	MSYJ019570	MSYJ088570	MSYJ088590	MSYJ170700
		Name	MAN_PAL1	MAN_PAL2	MAN_PAL3	MAN_PAL4
		Chr.	/	/	/	/
C. limon	4	ID	CL6G057756012_alt	CL6G057757012_alt	CL7G061014012_alt	CL8G065743012_alt
		Name	LEM_PAL1	LEM_PAL2	LEM_PAL3	LEM_PAL4
		Chr.	/	/	/	/
C. grandis cv. ‘Wanbaiyou’	5	ID	Cg6g001770	Cg6g001740	Cg7g006780	Cg8g019990	Cg8g020000
		Name	WBP_PAL1	WBP_PAL2	WBP_PAL3	WBP_PAL4	WBP_PAL5
		Chr.	6	6	7	8	8
C. grandis cv. ‘Cupi Majiayou’	5	ID	CMJ_chr6_004410	CMJ_chr6_004400	CMJ_chr7_020290	CMJ_chr8_005100	CMJ_chr8_005120
		Name	CMP_PAL1	CMP_PAL2	CMP_PAL3	CMP_PAL4	CMP_PAL5
		Chr.	6	6	7	8	8
C. medica	4	ID	Cm082420	Cm082440	Cm083800	Cm160580
		Name	CIT_PAL1	CIT_PAL2	CIT_PAL3	CIT_PAL4
		Chr.	/	/	/	/
C. clementina	5	ID	Ciclev10030821m	Ciclev10011134m	Ciclev10011175m	Ciclev10027912m	Ciclev10027913m
		Name	CLE_PAL1	CLE_PAL2	CLE_PAL3	CLE_PAL4	CLE_PAL5
		Chr.	/	/	/	/	/
C. ichangensis	5	ID	Ci041590	Ci117210	Ci117230	Ci133930	Ci133960
		Name	ICP_PAL1	ICP_PAL2	ICP_PAL3	ICP_PAL4	ICP_PAL5
		Chr.	/	/	/	/	/
C. australis	5	ID	g2243	g2245	g8228	g13929	g13930
		Name	RLI_PAL1	RLI_PAL2	RLI_PAL3	RLI_PAL4	RLI_PAL5
		Chr.	8	8	7	6	6

Num Number of PAL genes in each species, ID Original ID of PAL genes in the genome, Name names of PAL genes after renaming in this study, Chr. chromosome ID where the gene is located, “/“ indicates that the chromosome location cannot be determined

Fig. 1.

Physicochemical properties of the PAL genes in nine Citrus species. a-e Boxplots showing the protein sequence length, relative molecular weight, isoelectric point, instability coefficient, and hydrophilicity index encoded by the PAL genes in nine Citrus species, with different colored boxes representing different species

Multiple sequence comparisons of PAL proteins

Multiple sequence alignments of protein sequences encoded by PAL genes in nine Citrus species (see Fig. 2) revealed that the PAL gene family in citrus is highly conserved, with more than 80% similarity in more than 75% of the aligned positions. All the PAL genes encoded protein sequences in the nine Citrus species containing two vital functional regions, the specific substrate selection switch region “QKELIRFLNAGIFG” and the active site region “GTITASGDLV(L)PLSYIAG”, with an interval of approximately 55 amino acids between these two regions. In the region of specific substrate selection switches, although there was a mutation from K to M (CIT_PAL4) at site 2 and from A to S at site 10 (MAN_PAL4, RLI_PAL3, SWO_PAL3, CLE_PAL1, ICP_PAL1, LEM_PAL3, CIT_PAL3, WBP_PAL3, and CMP_PAL3) in some genes, all the citrus PAL genes had the same specific substrate selection switch, which was “FL”. Within the active site region, all the protein sequences encoded by the PAL genes contained a highly conserved active center, which is a highly conserved 4-methylidene-imidazole-5-one group (MIO) consisting of three amino acids: “ASG (Ala-Ser-Gly)”. However, some mutations were observed at some positions within this region. For example, G can mutate to A at site 1, T can mutate to S at site 2, V can mutate to I at site 10, and L can mutate to F at site 11. Furthermore, our study revealed a correlation between mutations at sites 1, 2, and 11 in the active site region, such that if one site is mutated, the other two sites will also mutate (see Fig. 2(B)). Interestingly, in species with genome assemblies at the chromosomal level, the genes with mutations at sites 1, 2, and 11 within the active site region were all located on chromosome 8.

Fig. 2.

Multiple sequence alignment of protein sequences encoded by PAL genes in nine Citrus species. A Multiple sequence alignment of protein sequences encoded by PAL genes in nine Citrus species, with black and gray shading representing identical and similar amino acid residues, respectively. B Multiple sequence alignments of the active site region and specific substrate selection switch region of the PAL genes in nine Citrus species with yellow, blue, black, and red rectangles representing the specific substrate selection region, specific substrate selection switch, active site region, and active center, respectively

Analysis of an interspecific phylogenetic tree for the PAL gene family

To understand the evolutionary relationship between PAL genes in different plants, we used the maximum likelihood tree-building software IQ-TREE to construct a phylogenetic tree based on amino acid sequences encoded by 41 PAL genes from citrus and 4 PAL genes from A. thaliana. The results are shown in Fig. 3.

Fig. 3.

Interspecies phylogenetic tree of the PAL gene family in nine Citrus species and A. thaliana. This evolutionary tree was constructed based on amino acid sequences encoded by 41 PAL genes from sweet orange and four PAL genes from A. thaliana via the JTT + G4 model as the optimal tree-building model and constructed via the maximum likelihood method with IQ-TREE software. The solid dots of different colors on each branch node represent bootstrap values from 1000 repetitions, where black represents bootstrap values greater than 90% and gray represents bootstrap values between 90% and 70%. No mark indicates bootstrap values less than 70%. The different colored blocks in the inner circle represent the ten various species. However, the diverse colored blocks and numbers one to 11 in the outer circle represent the 11 groups of this phylogenetic tree

The phylogenetic tree revealed that the 45 PAL genes from ten diverse plants could be divided into 11 evolutionary branches. Group 11 contained the greatest number of PAL genes (14). There was only one PAL gene each in Groups 1, 2, 3, 4, and 6. Groups 5, 7, and 9 had 2 PAL genes, and Groups 8 and 10 had 9 and 11 PAL genes, respectively. The PAL genes from A. thaliana were present only in Group 9 and Group 10. The results suggested significant differences in the evolutionary patterns of PAL genes between different species. Furthermore, at the genomic assembly level, which is at the chromosome level for the four species, PAL genes from the same chromosome tend to have closer relationships and cluster together in the same group. For example, PAL genes from chromosome 8 are clustered in Group 11, and genes from chromosome 7 are clustered in Group 10.

Conserved motif analysis of PAL proteins

To understand the function of the proteins encoded by the PAL gene family in citrus, we analyzed the motifs of the proteins encoded by 41 PAL genes via the online software MEME. The results (see Fig. 4) indicate that all the proteins encoded by the PAL genes, except for ICP_PAL4 (lacking motif 5 and motif 10) and ICP_PAL5 (lacking motif 7), contained ten motifs and exhibited a high degree of concordance with both motif arrangement and position. These findings suggest that the PAL gene family is highly conserved and that all the proteins encoded by the PAL genes may have similar or identical functions. However, while two PAL genes from the same species present highly similar motifs, their phylogenetic relationships are not necessarily closer than those of two genes from different species. These findings suggest that PAL genes from the same species may have originated or evolved differently.

Fig. 4.

Phylogenetic and motif analysis of the PAL gene family in nine Citrus species. a Phylogenetic tree of the protein sequences encoded by the 41 PAL genes in nine Citrus species. The tree was constructed via the maximum likelihood method with the JTT + G4 model in IQ-TREE software, and the values at each branch node represent the bootstrap values from 1000 replicates. b The phylogenetic grouping of the protein sequences encoded by the 41 PAL genes in the nine Citrus species, with different colored blocks and numbers 1 to 10 representing the ten groups of the phylogenetic tree. c The composition of the conserved motifs in the protein encoded by the PAL gene family, with different colors representing different motifs. d The logos of the ten conserved motifs. The detailed sequence information for each motif can be found in Supplementary Table S4

Gene structure and promoter cis-acting element analysis of the PAL gene family

The results (as shown in Fig. 5(c)) revealed that most (40) of the PAL genes in the citrus had a length of less than 6 kb, with only RLI_PAL4 being approximately 13 kb. The number of introns in PAL genes ranged from one to 16, with more than four-fifths of the PAL genes (35) having only one intron and only RLI_PAL4 having 16 introns. The number of introns in most citrus PAL genes shows relatively little difference, but their intron lengths exhibit significant variations. Interestingly, we found that PAL genes from the same evolutionary branch were similar, regardless of the number or length of introns. However, PAL genes from different branches, even if they had the same number of introns, presented significant differences in intron length, indicating possible differences in the origin or evolutionary patterns of the PAL genes clustered into different groups.

Fig. 5.

Analysis of PAL gene structure and cis-acting elements in nine Citrus species. a Phylogenetic tree of the protein sequences encoded by 41 PAL genes in nine Citrus species. The tree was constructed via the maximum likelihood method with the JTT + G4 model in IQ-TREE software. The numerical values at each branch node represent the bootstrap values based on 1000 replicates. b Phylogenetic grouping of the protein sequences encoded by 41 PAL genes in nine Citrus species, represented by different colored blocks and numbered 1 to 10 according to the phylogenetic tree. c Gene structure of the PAL gene family in nine Citrus species. The black lines represent introns, the yellow boxes represent coding sequences (CDSs), and the green regions represent untranslated regions (UTRs). The length of each gene can be estimated via the scale at the bottom. d Analysis of cis-acting elements in the promoter regions of the PAL gene family in nine Citrus species. The different colors represent different cis-acting elements. e Heatmap of cis-acting elements in the promoter regions of genes in the PAL gene family in nine Citrus species. A heatmap was generated via the Pheatmap package based on the number of each cis-acting element in the upstream region of each PAL gene. The different colored blocks at the bottom represent various cis-acting elements

To further investigate the regulatory and potential response mechanisms of the PAL genes, the upstream 2000 bp sequences of 41 PAL genes in citrus were analyzed via the online website PlantCARE to predict cis-acting elements. The results (as shown in Fig. 5(d)) revealed that the number of cis-acting elements in the gene upstream 2000 bp sequences ranged from ten to 35. Nine hundred and seven cis-acting elements were found in the upstream regions of the 41 PAL genes in citrus. These elements can be classified into five main categories. The elements were proportional, from largest to smallest, as follows: plant hormone response elements (38.15%), light response elements (29.55%), environmental stress response elements (14.11%), plant-specific expression regulatory elements (9.15%), and MYB binding sites (9.04%). Thirty-eight PAL genes have multiple ABA responsiveness and light response elements in their upstream regions. In summary, 90% of the predicted cis-acting elements are related to plant stress resistance in addition to light response elements. These findings suggest that the PAL genes in the citrus may play crucial roles in resistance to various stresses. In addition, we found that PAL genes from the same group presented similar cis-acting elements in their upstream regions, whereas PAL genes from different branches presented significant differences. Some cis-acting elements are present only in specific groups. For example, the most common cis-acting elements in the upstream regions of PAL genes clustered in group 6 were MYB binding sites involved in flavonoid synthesis. However, the most common cis-acting element upstream of the PAL genes clustered in group 7 was the light response. All the PAL genes in group 8 contained two zein metabolism regulatory elements, and the seed-specific expression regulatory elements were unique to the genes clustered in group 6. This finding suggested that PAL genes from the same group may have similar functions, whereas PAL genes from different groups may have some differences.

Intraspecific evolution analysis of the PAL gene family in citrus

To understand the intraspecific evolutionary relationships of the PAL genes in citrus, we used the maximum likelihood method and IQ-TREE software to construct intraspecific phylogenetic trees for different PAL genes encoded by Citrus species based on their amino acid sequences. The results (as shown in Fig. 6) revealed that the PAL genes of the nine Citrus species could be divided into three evolutionary branches, with only one gene each in Group 1 and Group 2. Notably, from the perspective of chromosomal assembly at the chromosome level, two genes located on the same chromosome tend to have closer genetic relationships and similar evolutionary times. We further analyzed duplicate gene pairs in the PAL gene family of each Citrus species. The results revealed (as shown in Fig. 6) that 18 pairs of PAL genes were identified, ranging from zero to three pairs in the PAL gene family of each Citrus species. Three pairs of PAL genes were found in C. sinensis, C. reticulata, and C. grandis cv. ‘Wanbaiyou’, C. grandis cv. ‘Cupi Majiayou’, and C. australis, while C. limon had two pairs, and C. clementina had one pair. However, no PAL gene pairs were found in C. medica or C. ichangensis. Interestingly, from the perspective of chromosomal assembly at the chromosome level, we found that two genes from the same chromosome were not duplicate gene pairs.

Fig. 6.

Intraspecific phylogenetic relationships and duplicate gene analysis of the PAL gene family in nine Citrus species. a C. sinensis. b C. reticulata. c C. limon. d C. grandis cv. ‘Wanbaiyou’. e C. grandis cv. ‘Cupi Majiayou’. f C. medica. g C. clementina. h C. ichangensis. i C. australis. A phylogenetic tree was constructed on the basis of the protein sequences encoded by all the PAL genes of each Citrus species via the maximum likelihood method with an appropriate model selected via IQ-TREE software. The numbers at the branch nodes represent the bootstrap values obtained from 1000 replicates. The different colored blocks in the phylogenetic trees represent the three groups of the respective trees. The red lines on the right side of the phylogenetic trees connect gene pairs of duplicate genes

To investigate the types of natural selection and evolutionary pressure on PAL genes within citrus, we utilized the NG and YN methods to calculate the selection pressure Ka/Ks for 18 pairs of duplicate PAL genes from different Citrus species. The results (shown in Fig. 7; see Supplementary Table S5 for details) indicated that the Ka values for all 18 pairs of duplicate PAL genes were lower than 0.15, whereas the Ks values were all greater than 1.5, with Ks far exceeding Ka. Overall, the Ka/Ks ratios for the 18 pairs of duplicate PAL genes in citrus were less than one. Therefore, we infer that the PAL gene family in the citrus has undergone purifying selection and that its functional properties have not undergone significant changes during the evolutionary process.

Fig. 7.

Ka/Ks analysis of duplicate gene pairs in the PAL gene family of citrus. A NG method. B YN method. Ka/Ks ratios were calculated via the NG and YN methods for 18 pairs of duplicate PAL genes from different Citrus species and visualized via the R package ggplot2

Analysis of the contraction and expansion of the PAL gene family

To understand the contraction and expansion of the PAL gene family in citrus, a phylogenetic tree was constructed based on homologous genes, and the evolution of the PAL gene family was analyzed via Notung software. The results (shown in Fig. 8.) reported that PAL gene loss occurred more frequently than gene gain in nine Citrus species. Among them, C. limon, C. medica, and C. grandis presented significantly greater PAL gene losses than did the other species. Overall, the PAL gene family exhibited sustained contraction during evolution in the nine Citrus species. Notably, no significant correlation was detected between PAL gene loss and the number of PAL genes. Additionally, we found that species with closer phylogenetic relationships tended to have similar patterns of contraction or expansion in the PAL gene family. For example, C. ichangensis and C. australis, which are closely related, both presented a loss of four PAL genes.

Fig. 8.

Contraction and expansion of the PAL gene family in nine Citrus species. A phylogenetic tree of the nine Citrus species was constructed based on the orthologous single-copy gene identification via OrthoFinder2. The orthologous genes were aligned via MAFFT software and then concatenated and classified by species via Seqkit software. The red and blue numbers represent gene gains and losses in the PAL gene family, respectively. The orange numbers at each node of the phylogenetic tree represent divergence time points estimated via the MCMCtree tool of PAMLX software based on the orthologous genes shared by the nine Citrus species. The green numbers represent the confidence intervals of the divergence times with a confidence level of 95%. The time scale near the phylogenetic tree represents the time scale in millions of years. The different colored blocks at the bottom represent different geological time scales

Expression analysis of citrus PAL genes under biotic and abiotic stresses

To understand whether PAL genes play a role and to determine their expression patterns in citrus plants under biotic and abiotic stresses, we analyzed transcriptome data from five Citrus species under different stresses (transcriptome data sources are detailed in Table 1). We used the R package Pheatmap to generate a heatmap (see Fig. 9) based on log₂^(TPM+1) expression values of each PAL gene in different Citrus species under various stresses. The results revealed that under diverse stresses, all the PAL genes from the five Citrus species presented varying degrees of expression, with 80% of the PAL genes exhibiting high expression. Different Citrus species exhibit distinct expression patterns of PAL genes under the same and various stresses and even exhibit contrasting regulatory modes. For example, in response to drought stress, C. sinensis and C. limon exhibit different regulatory modes, with all PAL genes in C. sinensis showing positive regulation and all PAL genes in C. limon showing negative regulation. We detected significant differences in PAL gene expression patterns in the same citrus under different stresses and even within diverse tissues of the same citrus under the same stress.

Fig. 9.

Heatmaps depicting the expression patterns of PAL genes in different Citrus species in response to various biotic and abiotic stresses. a Heatmap showing the expression patterns of PAL genes in different tissues of C. sinensis under aluminum stress, drought stress, Mg deficiency, L. psalliotae infestation, Citrus tristeza virus infestation, and P. digitatum infestation. b Heatmap showing the expression patterns of PAL genes in C. reticulata leaves under low temperature and after D. citri invasion. c Heatmap showing the expression patterns of PAL genes in different tissues of C. grandis cv. ‘Wanbaiyou’ under Cu toxicity-3 and aluminum metal ion stress. d Heatmap showing the expression patterns of PAL genes in C. limon leaves under drought stress, low temperature, C. Liberibacter asiaticus invasion, and Citrus yellow vein clearing virus invasion. e Heatmap showing the expression patterns of PAL genes in C. medica leaves under P. tracheiphilus invasion. The heatmaps were constructed via the R package Pheatmap based on log₂^(TPM+1) values of each citrus PAL gene under different stresses

We performed differential expression analysis of citrus PAL genes under different stresses via the DESeq2 package to investigate which PAL genes play a prime role in each citrus. The results revealed significant differences in the differential expression of PAL genes in diverse Citrus species under various stresses. Even within the same citrus, there are substantial differences in the number of differentially expressed PAL genes under different stress treatments.

Under aluminum stress in C. sinensis, we detected differential expression of a PAL gene (SWO_PAL2), which was significantly upregulated in both the roots and leaves. Under drought stress, all four PAL genes were significantly upregulated in C. sinensis leaves, with SWO_PAL2 exhibiting the most significant upregulation, an increase of 6.4-fold. Under magnesium deficiency stress in C. sinensis leaves, three PAL genes were differentially expressed with two (SWO_PAL1 and SWO_PAL4) downregulated and one (SWO_PAL2) upregulated. In C. sinensis leaves infected with L. psalliotae, two PAL genes (SWO_PAL1 and SWO_PAL2) were significantly upregulated, with similar fold changes of 2.5. In sweet orange fruits infected with P. digitatum, three PAL genes (SWO_PAL1, SWO_PAL2, and SWO_PAL4) were significantly upregulated, with SWO_PAL2 exhibiting markedly greater than 70-fold greater upregulation than the other two PAL genes. No differential expression of any PAL gene was observed only in C. sinensis leaves invaded by the Citrus tristeza virus. Notably, except for C. sinensis leaves invaded by Citrus tristeza virus, the SWO_PAL2 gene was differentially upregulated to varying degrees under the other stresses. In most cases, this gene was most strongly upregulated. These findings suggest that the SWO_PAL2 gene may be a vital broad-spectrum resistance gene in C. sinensis and may play a more significant role in defense against various biotic or abiotic stresses than other PAL genes.

Under low-temperature stress, all PAL genes, except for MAN_PAL4, were significantly upregulated in C. reticulata leaves. MAN_PAL3 was the most significantly upregulated gene, increasing by 128-fold. In C. reticulata leaves infected by D. citri, only one differential gene (MAN_PAL3) was found and upregulated 4.6-fold. Notably, among the four C. reticulata PAL genes, only the MAN_PAL3 gene was significantly upregulated under biotic and abiotic stress. This finding is highly similar to the results we reported for C. sinensis. Therefore, we speculate that the MAN_PAL3 gene may be a vital broad-spectrum resistance gene in C. reticulata.

Under Cu toxicity-3 stress, three PAL genes were significantly differentially expressed in C. grandis cv. ‘Wanbaiyou’ leaves, with one (WBP_PAL1) significantly upregulated and two (WBP_PAL4, WBP_PAL5) significantly downregulated. No significant differential expression of PAL genes was detected in C. grandis cv. ‘Wanbaiyou’ leaves or roots under Al stress.

Under drought stress, all the PAL genes exhibited varying degrees of significant downregulation in C. limon leaves, with LEM_PAL2 and LEM_PAL3 showing more pronounced downregulation. Under low-temperature stress, three PAL genes (LEM_PAL1, LEM_PAL2, and LEM_PAL3) were significantly upregulated to varying degrees in C. limon fruits, with LEM_PAL1 exhibiting the most significant upregulation with a 128-fold increase. In C. limon leaves invaded by C. Liberibacter asiaticus, only the LEM_PAL3 gene was significantly upregulated. In C. limon leaves invaded by the Citrus yellow vein-clearing virus, two PAL genes (LEM_PAL1 and LEM_PAL2) were significantly upregulated.

In C. medica leaves invaded by P. tracheiphilus, all PAL genes, except for CIT_PAL4, were upregulated, with CIT_PAL1 showing the most significant upregulation, increasing 14-fold.

In conclusion, PAL genes are expressed in different Citrus species under biotic and abiotic stresses, and several PAL genes exhibit significant differential expression. Therefore, it is speculated that PAL genes play crucial roles in the resistance of citrus plants to various biotic and abiotic stresses.

Expression analysis of genes upstream and downstream of the PAL genes in the lignin and flavonoid anabolic pathways in Citrus species

To explore the involvement of PAL genes in the regulation of lignin and flavonoid biosynthesis pathways in Citrus species, we identified vital genes involved in lignin and flavonoid synthesis in the genomes of C. sinensis, C. reticulata, C. limon, and C. grandis cv. ‘Wanbaiyou’, and C. medica, based on extensive previous studies. We constructed lignin and flavonoid biosynthesis pathways in these five Citrus species by integrating previous research and the KEGG database. The results (as shown in Fig. 10 and Supplementary Figures S1-S4) indicate that the synthesis of lignin in Citrus species involves three main pathways, whereas the synthesis of flavonoids involves one pathway. The conversion of prephenate to p-coumaroyl-CoA is a common step in the lignin and flavonoid biosynthesis pathways. Prephenate is ultimately converted to p-coumaroyl-CoA by the sequential catalysis of five enzymes: PAT, PDT, PAL, C4H, and 4CL. P-coumaroyl-Co is further catalyzed by various enzymes, CCR, CAD, POD, HCT, C3’H, CCoAOMT, F5H, and COMT, which are ultimately converted to three important precursors of lignin, guaiacyl lignin, syringyl lignin, and p-hydroxyphenyl lignin. On the other hand, p-coumaroyl-CoA can be transformed into naringenin (flavone) by the catalysis of CHS and CHI and further converted into pelargonidin (flavonoid) by the catalysis of CitF3H, DFR, and ANS.

Fig. 10.

Lignin and flavonoid biosynthesis pathways in C. sinensis. The heatmap shows the expression patterns of 18 vital genes involved in lignin and flavonoid biosynthesis in C. sinensis under different treatments. The values in the heatmap were calculated as log₂^(TPM+1) values based on the expression levels of each gene. The asterisk in the heatmap indicates significant differential gene expression compared with the control (CK) group. The solid black arrows represent the conversion of two substances through the catalysis of specific enzymes. However, the dashed black arrows represent the conversion of specific substances that require the catalysis of multiple enzymes. Since some genes involved in lignin synthesis are involved in all three pathways, different shapes were used to label these genes. The rectangle represents the HCT gene, the hexagon represents the POD (peroxidase) gene, the ellipse represents the CCR gene, and the pentagram represents the CAD gene

Furthermore, we explored the expression patterns of vital genes involved in lignin and flavonoid biosynthesis under different stress conditions in five Citrus species by integrating transcriptomic data. The results (as shown in Fig. 10 and Supplementary Figures S1-S4) revealed that, from the lignin synthesis pathway, the five Citrus species presented a high degree of consistency in both gene expression levels and the number of highly expressed genes. Among the first five crucial enzyme-encoding gene families (PAT, PDT, PAL, C4H, and 4CL) involved in the lignin synthesis pathway in the five Citrus species, all genes were highly expressed, except for individual genes encoding the 4CL enzyme, which presented low or no expression. For the remaining eight key gene families, more than 50% of the genes were highly expressed. In addition, we found that some genes encoding the 13 crucial enzymes in Citrus species exhibited some spatial specificity. For example, multiple POD and CAD genes, which are highly expressed in either roots or leaves, exhibit spatial specificity in C. sinensis (Fig. 10) and C. grandis cv. ‘Wanbaiyou’ (Supplementary Figure S3). However, the genes in the flavonoid biosynthesis pathway show some differences due to species specificity. For example, the DFR gene is expressed only in the leaves of C. limon (Supplementary Figure S2), the roots of sweet orange, and C. grandis cv. ‘Wanbaiyou’ (Supplementary Figure S3), but not in other Citrus species. Therefore, we speculate that the lignin biosynthesis process may be similar among the five Citrus species, whereas the flavonoid biosynthesis process may vary.

We also conducted differential expression analysis of all vital genes in the lignin biosynthesis pathway in the five Citrus species. Many genes involved in lignin and flavonoid biosynthesis were upregulated after stress treatment, and the expression of several genes significantly differed. For example, in the transcriptome of C. sinensis peels under P. digitatum infestation (Fig. 10) and the transcriptome of C. reticulata leaves under low-temperature stress (Supplementary Figure S1), 80% of the genes encoding the 13 vital enzymes involved in lignin synthesis were upregulated to varying degrees. No differential expression was detected for genes encoding the PAT, HCT, or F5H enzymes. However, the remaining vital genes were more or less differentially expressed. According to the transcriptome of C. sinensis leaves (Fig. 10) under drought stress and the transcriptome of C. reticulata under D. citri infestation (Supplementary Figure S1), more than 60% of the genes encoding the ten vital enzymes involved in flavonoid synthesis were upregulated. These findings indicate that lignin and flavonoid synthesis regulation may be a response mechanism for Citrus species to resist various stresses.

In addition, this study revealed significant differences in the expression of vital genes involved in lignin and flavonoid biosynthesis under different stress conditions in the same Citrus species, and some genes even presented opposite trends. This finding suggested that similar to the PAL gene, the vital genes involved in lignin and flavonoid biosynthesis in Citrus species respond specifically to stress.

Discussion

With the rapid expansion of citrus cultivation and the accelerated degradation of the ecological environment, various biotic and abiotic stresses, such as drought, low temperature, soil salinization, excessive soil metal ions, ulcer disease, Citrus Huanglongbing, and viral diseases, have become vital factors restricting the high-quality development of the citrus industry [4]. The Breeding of citrus varieties with high resistance has become a significant priority in citrus research. With the increasing sequencing of citrus plant genomes and the rapid development of genetic engineering technologies such as genetic transformation and gene editing, favorable conditions have been provided for breeding highly resistant plant varieties. The PAL gene family is among the most vital defense gene families in almost all plants and plays a crucial role in plant resistance to various external stresses. However, thus far, researches on PAL genes in citrus have been limited to a small number of cases in C. sinensis [3], and few studies have been conducted on these genes in other Citrus species. Although many studies have shown that there are many naturally highly resistant species in the Citrus genus, such as C. australis and C. medica, the PAL genes of these highly resistant Citrus species have not been comprehensively studied. This scarcity of related research has led to little knowledge about the evolution of PAL genes in citrus and the lignin and flavonoid biosynthesis pathways in which they are involved. In the present study, PAL genes were identified from the genomes of nine Citrus species, namely, C. sinensis, C. reticulata, C. limon, and C. grandis cv. ‘Wanbaiyou’, C. grandis cv. ‘Cupi Majiayou’, C. medica, C. clementina, C. ichangensis, and C. australis, and the characteristics of the PAL genes in these nine Citrus species were analyzed in detail, laying a foundation for the genetic improvement of stress resistance in citrus.

Many studies have shown that the PAL gene family in most plants is highly conserved and small, with a generally small number of members [9, 19], mostly fewer than 20. Forty-one PAL genes were identified from the genomes of nine Citrus species, each containing four or five PAL genes in this study and multiple sequence alignment analysis revealed that more than 75% of the amino acid positions in the protein sequences encoded by the 41 citrus PAL genes presented greater than 80% similarity. These findings are consistent with previous research on species such as P. trichocarpa [6] (including 5 PAL genes), S. viminalis [7] (5), C. arabica [8] (3), C. hainanensis [9] (4), and A. thaliana [10] (4). In addition, previous studies on the structure of the PAL enzyme encoded by PAL genes have revealed a specific substrate selectivity switch region and an active center in almost all dicotyledonous plant PAL enzymes [15]. The substrate selectivity switch region usually contains a conserved motif composed of 14 amino acids, “QKELIRFLNAGIFG”, with “FL” as the specific substrate selectivity switch for PAL [51]. Mutation or deletion of this substrate selectivity switch directly leads to PAL inactivation or functional transformation [52]. Research has shown that when the phenylalanine (F) in the specific substrate selectivity switch region is mutated to histidine (H), PAL loses its function and transforms into a TAL [53]. The active center usually consists of a conserved motif composed of 17 amino acids, “GTITASGDLV(L)PLSYIAG”, which contains a highly conserved 4-methylidene-imidazole-5-one group (MIO) with three amino acids, “ASG (Ala-Ser-Gly)” [54, 55]. The specific substrate selectivity switch region and the active center are indispensable for any PAL enzyme with normal function. Therefore, the presence of a region and center is also considered vital for identifying a gene as a PAL gene [11]. In this study, all the citrus PAL genes contained specific substrate selection switch regions, “QKELIRFLNAGIFG”, and active site regions, “GTITASGDLV(L)PLSYIAG”. In conclusion, as in the majority of plants, the PAL gene family in citrus is highly conserved.

Gene duplication is an important molecular mechanism that drives gene evolution and gene family expansion [56]. Studies have shown that nearly 1/4 of the MYB genes in Liriodendron chinense are generated by segmental duplication [57], and the NBS-LRR gene families in Fabaceae and Solanaceae have expanded primarily through tandem duplication [58]. We identified duplicate gene pairs in the PAL gene family of the nine Citrus species, and the results revealed that PAL gene pairs were present in seven Citrus species, excluding C. medica and C. ichangensis. Among them, three PAL duplicate gene pairs were found in five Citrus species, and 18 PAL gene pairs were found among Citrus species. The results revealed that gene duplication is a vital means of expanding the PAL gene family in citrus. To further determine the type and strength of natural selection acting on the PAL gene family, we calculated the Ka/Ks ratio for the 18 duplicate gene pairs, and we found that all of them had a Ka/Ks ratio less than 1, suggesting that the PAL gene family in citrus has undergone purifying selection. The functional characteristics of the PAL gene family have not undergone significant changes during evolution.

Promoter cis-acting elements are crucial for the transcriptional regulation of plant genes [25, 59, 60]. In our study, numerous cis-acting elements related to resistance were identified in the upstream 2000 bp sequences of 41 PAL genes from citrus, including ABA response elements, MeJA response elements, MYB binding sites, and so on. Previous studies have shown that exogenous ABA and MeJA treatment of plants under stress, such as rice, Fragaria chiloensis, and peach, significantly upregulates the expression of some PAL genes, thereby helping plants resist stress [19, 61]. Therefore, on the basis of the analysis of the cis-acting elements in the PAL gene promoter, the PAL gene family in citrus plays a crucial role in citrus resistance to various stresses.

When plants perceive various biotic and abiotic stresses from the external environment, they regulate the expression of their genes to control the synthesis of resistance substances to maintain normal intracellular activities [62]. It is a vital mechanism for plants to survive under unfavorable environmental conditions. Previous studies have shown that when rice [18], potato [19], and S. caninervis [20] are subjected to stress, they protect themselves from various stresses by regulating the expression of the PAL genes. We analyzed transcriptome data from five Citrus species under various stresses. The results revealed that all the PAL genes in the five Citrus species were expressed to varying degrees under different stress conditions, with 80% being highly expressed. The expression of several PAL genes was significantly upregulated in Citrus species under stress. These findings are in line with those of previous studies in other species. These results confirm the crucial role of PAL genes in the process through which Citrus species resist various stresses, which has also been demonstrated in studies on the resistance of sweet orange fruits to biotic stresses [3, 21]. Notably, we also detected significant differences in the expression levels of different PAL genes in the same Citrus species under the same and various stresses. Previous studies have shown that differences in gene expression levels may be related to factors such as gene copy number, evolutionary time, and selection pressure [63]. By combining the evolutionary relationships of PAL genes in citrus and the expression patterns of PAL genes under different stresses, we found for the first time a strong correlation between the ability of PAL genes to respond to stress and evolutionary time. PAL genes with shorter evolutionary times present broader resistance and are involved in response to more stresses, showing more significant upregulation after stress. We found that these PAL genes associated with broad-spectrum resistance did not undergo gene duplication during the evolution of the PAL gene family, indicating that these PAL genes associated with broad-spectrum resistance are single-copy genes. We speculate that the reasons for this phenomenon may include the following two aspects: on the one hand, citrus PAL genes have to be changed under selective pressure, but it is difficult to preserve the mutated PAL genes because of the conservation of the PAL gene family. On the other hand, the PAL gene, which is associated with broad-spectrum resistance to stress, may be young compared with other PAL genes in the same species and may have no chance to replicate and expand in citrus.

In summary, the PAL genes in Citrus species play crucial roles in resisting various biotic and abiotic stresses. However, the specific substances synthesized and metabolized by PAL genes to help Citrus species resist stress have not been identified. Previous studies have provided some insight. In A. thaliana, PAL gene mutants accumulate relatively little lignin and flavonoids, which results in a decrease in plant tolerance to stress [10]. In tobacco (Nicotiana tabacum) [64] and Salvia miltiorrhiza [65], the downregulation of PAL genes led to a significant change in lignin deposition and diminished plant growth. In addition, secondary metabolism resulting in lignin and carotenoid biosynthesis changes is an important method of the C. sinensis nonhost response, and the PAL genes play crucial roles in this process [66]. Wei et al. [3] demonstrated that overexpression of the PAL gene in sweet oranges could increase resistance to green mold through the increase in flavonoids. Multiple lines of evidence suggest that PAL genes are essential crucial genes for the synthesis of lignin and flavonoids in most plants [18, 19] and that PAL genes can help plants resist various stresses by participating in lignin and flavonoid synthesis. As lignin is a crucial component of plant cell walls, an increase in lignin content leads to thickening of the cell walls, thereby increasing plant stress tolerance [67]. Flavonoids are the most vital antioxidants in plants [68] and have strong reactive oxygen species (ROS) scavenging abilities [69]. When plants are subjected to various stresses, flavonoids can help them quickly eliminate excess ROS generated by stress, thus effectively protecting plants from damage caused by different biotic and abiotic stresses [70]. In this study, we constructed metabolic pathways for lignin and flavonoid synthesis in five Citrus species for the first time. We found that the expression levels of 17 crucial genes, including PAT, PDT, PAL, C4H, 4CL, CCR, CAD, POD, HCT, C3’H, CCoAOMT, F5H, COMT, CHS, CHI, and CitF3H, in the lignin and flavonoid synthesis pathways were upregulated after Citrus species were subjected to stress, with some genes being significantly upregulated. Therefore, we speculate that the PAL genes in the Citrus species help them resist various stresses by participating in lignin and flavonoid synthesis. However, this is suggested by the widely divergent lignin monomer compositions of different plant species [71]. In gymnosperms (such as Pinus spp.), lignin is composed of guaiacyl units and p-hydroxyphenyl units. In angiosperms (such as A. thaliana, Populus spp., and Zea mays), lignin is composed of guaiacyl units, syringyl units, and p-hydroxyphenyl units [72]. We found that all vital genes in the three monolignol and flavonoid synthesis pathways were expressed to varying degrees in our study. These results indicate that, like in most angiosperms, the synthesis of monolignols in Citrus species may involve three pathways, whereas the synthesis of flavonoids may involve only one pathway. In this process, the conversion of prephenate to p-Coumaroyl-CoA occurs via the lignin and flavonoid synthesis pathways. Prephenate is sequentially catalyzed by PAT, PDT, PAL, C4H, and 4CL for conversion into p-coumaroyl-CoA. P-coumaroyl-CoA can be further catalyzed by CCR, CAD, POD, HCT, C3’H, CCoAOMT, F5H, and COMT to produce three crucial precursors of lignin, namely, guaiacyl lignin, syringyl lignin, and p-hydroxyphenyl lignin, respectively. P-coumaroyl-CoA can also be converted into naringenin (flavone) by CHS and CHI and further transformed into pelargonidin (flavonoid) by CitF3H, DFR, and ANS.

In summary, multiple lines of evidence suggest that Citrus species can resist various stresses by regulating the expression of PAL genes and controlling the synthesis of lignin and flavonoids. First of all, promoter cis-acting element analysis suggested that the transcription of citrus PAL genes can be activated by many adversity signals, such as ABA, JA, and MYB transcription factors. Second, transcriptome analysis indicated that the PAL gene family plays a crucial role in resistance to various stresses in Citrus species. Third, pathways for lignin and flavonoid synthesis in Citrus species revealed that the PAL genes in Citrus species play indispensable roles in these pathways. Finally, previous studies have shown that lignin [18, 20] and flavonoids [16] can enhance the physical defense capabilities and reactive oxygen species (ROS) scavenging abilities of plants. Therefore, our study proposes a hypothetical mechanism for stress resistance in Citrus species as follows: When Citrus species are subjected to stress, they activate PAL gene expression through ABA, JA, and MYB transcription factors, thereby regulating the synthesis of lignin and flavonoids and increasing their physical defense capabilities and ROS scavenging abilities to resist various stresses (as shown in Fig. 11.).

Fig. 11.

Hypothetical mechanism of enhanced stress resistance in Citrus species through PAL gene activation. This mechanism suggests that when Citrus species are subjected to stress, PAL gene expression is activated through the ABA, MeJA, and MYB signaling pathways. In turn, it regulates the synthesis of lignin and flavonoids and enhances the physical defense capabilities and reactive oxygen species (ROS) scavenging abilities of Citrus species to help them resist various stresses. Red arrows indicate an increase in the corresponding substance levels, whereas green arrows indicate a decrease in the corresponding substance levels

Conclusion

1. This study investigated the PAL gene family in nine Citrus species and identified 41 PAL genes. Each citrus has four or five PAL genes, indicating that the PAL gene family in citrus is highly conserved and small in size overall.

2. Gene duplication is a vital mechanism for PAL gene family expansion in citrus. During the evolution of the PAL gene family, purifying selection occurred, indicating that the function of this gene family has not undergone significant changes.

3. We discovered that the PAL gene family plays a crucial role in the stress response of citrus. Furthermore, we detected a strong correlation between the ability of PAL genes to respond to stress and their evolutionary duration. PAL genes with shorter evolutionary times present broader resistance to stresses. Additionally, we found that the PAL genes associated with broad-spectrum resistance were all single-copy genes.

4. In this study, we identified vital genes involved in lignin and flavonoid synthesis in the genomes of five Citrus species: C. sinensis, C. reticulata, C. limon, and C. grandis cv. ‘Wanbaiyou’, and C. medica. We also constructed detailed lignin and flavonoid synthesis pathways in these five Citrus species.

5. Finally, we propose a hypothetical mechanism by which Citrus species resist stress. When Citrus species are stressed, PAL gene expression is activated through the ABA, MeJA, and MYB signaling pathways. It, in turn, regulates the synthesis of lignin and flavonoids while also enhancing the physical defense capabilities and reactive oxygen species (ROS) scavenging abilities of citrus plants to help them resist various stresses.

This study provides a comprehensive framework for the PAL gene family in citrus and lays the foundation for further research on the biological functions of PAL genes in the growth, development, and response of citrus plants to various stresses. Meanwhile, this study also provides a solid theoretical basis for selecting citrus varieties with high resistance by choosing the high-expression PAL individuals.

Authors’ contributions

TY, RX: Conceptualization, Data curation, Formal analysis, Methodology and Writing - original draft. LZ, XY and MZ: Formal analysis and Visualization. XL, YZ and KW: Data curation and Validation. HC and KZ: Software. LX and HZ: Funding acquisition, Project administration, Supervision and Writing – review & editing. All the authors have read and agreed to the published version of the manuscript.

Funding

This study is supported by the Rural Revitalization Science and Technology Project-Rural Revitalization Industry Key Technology Integration demonstration Project (202304BP090005), the Yunnan Academician (expert) Workstation Project (202305AF150020), Science and Technology Project for Universities in Yunnan Province Serving Key Industries and the National Natural Science Foundation of China (Grant No. 31760450). The funders had no role in the design of the study, the collection, analysis, or interpretation of the data, or the writing of the manuscript.

Data availability

The data presented in this study are available in the article, Supplementary Materials and online repositories. The RNA-Seq data are publicly available at the National Center for Biotechnology Information, and accession numbers of RNA-Seq data are included in the Table 1 of this article.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.