Plant cell wall is a highly sophisticated structure composed of diverse polysaccharides, enzymes, and structural proteins, and it helps to determine cell size and shape, growth and development, intercellular communication and interaction with the external environment (Micheli 2001; Pelloux et al. 2007). Two classes of polysaccharides, pectins and hemicelluloses, are major components of the middle lamellae and primary cell walls. Pectins compose 30-35% of the cell wall dry weight in dicots and non-grass monocots, 2-10% of grass primary walls, and up to 5% of wood tissues. Pectins are important for both cellular adhesion and cell wall plasticity and are a highly complex and heterogenous group of polymers that can be divided into five classes of pectic polysaccharides: homogalacturonan (HGA), rhamnogalacturonans I and II, xylogalacturonan and apiogalacturonan (Pelloux et al. 2007; Wolf et al. 2009; Wu et al. 2018).

HGA is a linear homopolymer of 1,4-linked α-D-galacturonic acid (GalA) and the most abundant of the pectic polysaccharides, constituting 65% of all pectins (Wolf et al. 2009). Plant genes encoding HGA biosynthetic enzymes have been identified from Arabidopsis (Arabidopsis thaliana) and their biochemical features and localization to the Golgi have been studied (Amos et al. 2018; Atmodjo et al. 2011). It is generally accepted that HGA is polymerized in the cis-Golgi, methyl-esterified in the medial-Golgi and modified with side chains in the trans-Golgi and deposited in the cell wall in a highly methyl-esterified form (Wolf et al. 2009). HGA secreted into the cell wall can be de-esterified by the removal of methyl groups by cell wall-localized enzymes during cell growth and development, generating free carboxyl groups and releasing methanol and protons. The removal of methyl groups from GalA residues in the HGA chain is catalyzed by the ubiquitous pectin methylesterase (PME, EC 3.1.1.11, CE8 of CAZy), also called pectinesterase, which is produced by higher plants and microorganisms that have cell-wall degradation activity (Pelloux et al. 2007). Mature PMEs have been proposed to have three modes of action on HGA: single-chain, multiple-chain, and multiple attack. In the multiple-chain mechanism (a type of random action mode), demethylesterification releases protons that promote the action of other cell wall enzymes, such as polygalacturonase and pectin/pectate lyases, which could contribute to cell wall loosening. In contrast, when PMEs act linearly on HGA using the single chain or multiple-attack mechanism, demethylesterification leads to the Ca²⁺ cross-linking of the free carboxyl groups on two adjacent HGA to form “egg-box” structures, which could rigidify the cell wall (Wolf and Greiner 2012; Wu et al. 2018).

A primary structure analysis of PME genes revealed that plant PMEs can be divided into two groups, group 1/type II and group 2/type I, depending on the absence or presence of an N-terminal extension (PRO region) preceding the catalytic PME domain (Pfam01095). The N-terminal PRO region varies in length and shares significant sequence similarity with the PME inhibitor domain (PMEI domain, Pfam04043). Group 1/type II PMEs are characterized by the absence of an N-terminal PRO region. The signal peptide commonly found in both groups of PMEs is required for protein targeting to the endoplasmic reticulum. PMEs are secreted into the cell wall as mature PME form via the Golgi networks. Cleavage of the PRO region can occur before the excretion of mature PME into the cell wall or afterward. PME proteins in bacteria and fungi do not have a PRO region (Pelloux et al. 2007; Sénéchal et al. 2014b). Comparisons of amino acid sequences among PME domains identified five characteristic sequence segments (44_GxYxE, 113_QAVL, 135_QDTL, 157_DFIFG, and 223_LGRPW; carrot numbering) and six strictly conserved residues (Gly44, Gly154, ASP157, Gly167, Arg225, and Trp227), which have all been shown to be functionally important (Markovič and Janeček 2004).

In higher plant species, PME isoforms are encoded by a large multigene family that is differentially expressed in different tissues and in response to environmental stresses (Louvet et al. 2006; Sénéchal et al. 2014b; Wang et al. 2021; Wen et al. 2020; Zhang et al. 2019). In recent years, whole-genome sequencing technologies and bioinformatic analyses have led to the genome-wide identification and characterization of the PME gene family in diverse plant species: 66 in Arabidopsis (Louvet et al. 2006), 89 in poplar (Geisler-Lee et al. 2006), 43 in rice (Jeong et al. 2015), 43 in maize (Zhang et al. 2019), 57 in tomato (Wen et al. 2020) and 127 in soybean (Wang et al. 2021). The lower number of PME genes in grasses is closely related to the finding that methyl-esterified HGA is less abundant in grass species of the Poaceae family than in dicotyledonous species (Wu et al. 2018).

PMEs modulate plant growth and development by regulating the methyl-esterification status of HGA in the cell wall. PMEs are thus involved in diverse physiological processes associated with both vegetative and reproductive plant development and responses to biotic/abiotic stresses. The functional roles of several PMEs including QRT1 and VGD1 have been addressed by isolation and the characterization of loss-of-function Arabidopsis mutants (Wolf et al. 2009; Wu et al. 2018). The roles of PME in fruit ripening have been explored using transgenic tomato plants that express antisense gene constructs (Phan et al. 2007; Wen et al. 2013). A recent report showed that a pollen-expressed PME gene confers male function in unilateral cross-incompatibility in maize by controlling the pectin esterification status required for pollen tube growth (Wang et al. 2022; Zhang et al. 2018).

Citrus is a global economic fruit crop with the largest fruit production in the world. Citrus has a relatively small genome with compact gene content (~30,00 genes). The average haploid genome size in citrus species ranges from 360 Mb (mandarin) to 398 Mb (citron) (Gmitter et al. 2012), and several reference genomes are currently available (https://www.citrusgenomedb.org/). Citrus has several fascinating botanical features, such as long juvenility, self-/cross-incompatibility, sterility in pollen/ovule, polyembryony, and high citrus-specific flavonoid content in the fruits (Kim et al. 2016; Woo et al. 2020), which attract scientists to study its growth and development. In this study, we identify PME gene family members with diverse functional roles from the sweet orange genome and characterize their molecular features, including molecular phylogeny and gene structure analyses. We also performed an expression profile analysis of citrus PME genes using open sources of transcriptome data. Our results provide useful information about the biological functions of PME genes during the growth and development of citrus species and could be used to develop molecular breeding tools.

Identification of PME family genes and sequence analysis in a Citrus genome

PME protein sequences were retrieved from the sweet orange (Citrus sinensis) genome v2.0 (HZAU), which consists of 10 scaffolds, with an assembly size of 328 Mb and contains 29,655 annotated genes (https://www.citrusgenomedb.org/analysis/186/; Xu et al. 2013). To identify candidate citrus PME genes, we performed a BLASTP search against the citrus genome using Arabidopsis PMEs (Louvet et al. 2006) as queries (e-value > 1e-5, identity > = 40, match length > = 100). A Hidden Markov model (HMM) analysis was performed against the citrus genome annotation data with the PME domain (Pfam01095) as a query using the HMMER3.0 package (Eddy and Pearson 2011). Transmembrane helices in PME proteins were predicted using the TMHMM-2.0 server (https://services.healthtech.dtu.dk/service.php?TMHMM-2.0; Moller et al. 2001). The signal peptide domain for the secretion of PME proteins was predicted using the SignalP 6.0 server (https://services.healthtech.dtu.dk/service.php?SignalP; Teufel et al. 2022). We predicted N-glycosylation sites, and calculated isoelectric points and molecular weights using the ScanProsite and Compute pI/Mw tools of the Expasy server (https://web.expasy.org/), respectively. The exon-intron structures of citrus PME genes were analyzed using the Gene Structure Display Server 2.0 (http://gsds.gao-lab.org; Hu et al. 2015) based on the corresponding genomic DNA sequences (Xu et al. 2013).

Phylogenetic analysis

Full-length PME protein sequences derived from citrus (this study), Arabidopsis (Louvet et al. 2006), tomato (Wen et al. 2020), and maize (Moran Lauter et al. 2017; Wang et al. 2022) were used for the phylogenetic analysis. Multiple sequence alignment was performed using ClustalW software with the default parameters. Based on the alignments, phylogenetic trees were constructed with the Neighbor Joining method (Kumar et al. 2016) using MEGA7.0 software (https://www.megasoftware.net/) with the Poisson correction model, complete deletion, 1,000 replicates for the bootstrap analysis, and a 50% cut-off value. Adobe Illustrator (https://www.adobe.com/products/illustrator.html) was used to decorate the resulting phylogenetic trees.

Expression profiling of PME genes in Citrus

To analyze the expression profiles of PME genes in various citrus tissues, raw RNA-seq data were retrieved from the Sequence Read Archive (SRA) of the NCBI (https://www.ncbi.nlm.nih.gov/sra/) (Supplementary Table 1). Low quality reads and adaptor sequences were eliminated from the raw transcriptome data by Trimmomatic software (v.0.33; (http://www.usadellab.org/cms/?page=trimmomatic). Viral, rRNA, human, and bacterial sequences were removed by k-mer matching to NCBI RefSeq database using BBDuk software (v.38.87; https://jgi.doe.gov/data-and-tools/bbtools/) with k = 31 and mcf = 0.5 options. The C. sinensis genome v2.0 (HZAU) was used as a reference for mapping the RNA-seq reads and gene annotation. Pre-processed reads were mapped to the reference genome using HISAT2 software (https://ccb.jhu.edu/software/hisat2/index.shtml). The expression of individual genes was counted using the HTSeq-count method (Putri et al. 2022). Expression deviation among samples was normalized using the normalization procedure using DESeq2 software (Love et al. 2014; http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html). Differentially Expressed Genes (DEGs) among samples were selected from the normalized read counts mapped to the reference sequence (unigene). The selection criteria were as follows: |log₂ fold change| ≥ 1 and adjusted p-value <0.05. A Heatmap was drawn using the FPKM (Fragments Per Kilobase of transcript per Million) value for PME genes identified from the citrus genome.

Identification and analysis of PME family genes from a Citrus genome

To identify putative PME family genes in citrus, BLASTP and HMM searches were performed against the genome of sweet orange (C. sinensis). The BLASTP and HMM searches identified 54 and 53 putative PME proteins, respectively, in the genome. After eliminating repetitive sequences, atypically short sequences, and sequences missing motifs highly conserved in the PME domain, 48 non-redundant PME genes were identified on different chromosomes, except for chromosome 8 (Supplementary Table 2). Based on the chromosome orders and chromosomal locations, the 48 PME genes were designated as CsPME1 to CsPME48. Their detailed molecular features were summarized in Supplementary Table 3. PMEs have a conserved PME domain (Pfam01095) and are divided into two groups/ types depending on the presence of a PME inhibitor (PMEI) domain (Pfam04043): group 1/type II (without PMEI) and group 2/type I (with PMEI) (Pelloux et al. 2007; Wu et al. 2018). In the sweet orange genome, 29 PMEs were classified into group 2/type I and 19 PMEs were identified as group 1/type II. The 48 citrus PME genes encode proteins ranging from 292 (CsPME7, Cs2g16310.1) to 632 amino acid residues (CsPME19, Cs4g06630.1). Their predicted molecular weights ranged from 32.0 to 69.6 kDa, with an average of 52.7 kDa, and their isoelectric points are between 5.11 (CsPME21, Cs4g06670.1) and 9.77 (CsPME40, Cs9g14440.1). PMEs are N-glycosylated cell wall enzymes. Glycosylation of secreted proteins affects folding, enzyme activity, trafficking, localization, and ligand interactions (Ruiz-May et al. 2012). Prediction of the N-glycosylation sites revealed that citrus PMEs have from 1 to 15 N-glycosylation sites, except for CsPME7 (Cs2g16310.1), similar to the prediction that most of the PMEs identified in the tomato genome would have N-glycosylated sites (Wen et al. 2020).

Multiple sequence alignment and phylogenetic analysis of Citrus PME genes

The amino acid sequence alignments of PMEs from plants, fungi and bacteria showed the five conserved sequence regions in the PME domain: 44_GxYxE, 113_QAVAL, 135_QDTL, 157_DFIFG, and 223_LGRPW (according to the carrot PME numbering) (Markovič and Janeček 2004). Six strictly conserved residues have been identified in those segments: Gly44, Gly154, Asp157, Gly161, Arg225, and Trp227. To investigate whether the five segments are highly conserved among the PME domains of 48 citrus PMEs, we performed a multiple sequence alignment using the ClustalW tool embedded in MEGA7.0 software (Supplementary Fig. 1). The sequence alignment revealed that amino acid residues consisting of an active site, two Asp residues (D136 in region III and D157 in region IV), two Gln residues (Q113 in region II and Q135 in region III) and an Arg residue (R225 in region V) were highly conserved among the citrus PMEs, although amino acid substitutions were found in the conserved Q113, Q135, and D136 sites. Amino acid substitutions were also found in two of the six conserved residues, Gly44 and Trp227.

A molecular phylogenetic analysis was performed for the 48 full-length CsPMEs by using MEGA 7.0 with the Neighbor Joining method. The resulting phylogenetic tree revealed that the 48 CsPMEs are divided into 2 groups/ types, group 1/type II and group 2/type I, containing 18 and 30 members, respectively (Fig. 1A). The tree indicated that CsPME46 belongs to group 2/type I, although no PMEI domain was found in its protein sequence (Fig. 1A and Supplementary Table 3). Functional roles of several plant PMEs have been extensively studied in Arabidopsis, tomato, and maize by using mutants and transgenic lines (Wang et al. 2022; Wen et al. 2020; Wu et al. 2018). To explore the phylogenetic relationships among the CsPMEs and other plant species’ PMEs and deduce the possible functions of the CsPMEs during growth and development, we constructed a phylogenetic tree using full-length amino acid sequences from the 48 CsPMEs, 66 Arabidopsis PMEs, 7 tomato PMEs, and 2 maize PMEs. The resulting tree also revealed that plant PMEs, including those of citrus, are divided into 2 groups (Fig. 2), as shown in Fig. 1A. Some of the citrus PMEs showed close relationships with several PMEs whose biological functions have already been investigated (Fig. 2).

Fig. 1. (A) Phylogenetic analysis of 48 CsPME protein sequences from the sweet orange genome. Amino acid sequences were aligned using ClustalW and the phylogenetic tree was drawn using MEGA7 with the Neighbor-Joining method. The bootstrap values indicated on the tree were calculated for 1,000 replicates. (B) The exon-intron structure of 48 CsPME genes. Orange boxes indicate exons. Blue and black lines indicate 5’-/3’-untranslated regions and introns, respectively

Fig. 2. Phylogenetic relationships between 48 CsPMEs and PMEs from other plant species. Full-length protein sequences for 66 Arabidopsis PMEs, 7 tomato PMEs, and 2 maize PMEs were used for the phylogenetic analysis. The functional roles of 18 Arabidopsis genes (green letters), 7 tomato genes (blue letters), and 2 maize genes (purple letters) have been studied. CsPMEs with a close phylogenetic relationship to the PMEs with a known function are indicated with red letters

According to evolutionary research using eukaryotic genome sequences, the exon-intron structure of protein-coding genes appears to have evolved concomitantly with eukaryotic cells, and introns are a major driving force in genome evolution and the functional divergence of gene family members (Liu et al. 2021; Rogozin et al. 2012). To better understand the structural diversity of citrus PME genes, we analyzed the exon-intron structures of the 48 CsPME genes (Fig. 1B). The number of introns varied from 1 to 5, with 1 to 2 in group 2/type 1 genes, and 3 to 5 in group 1/type II genes (Fig. 1B). The exon-intron organization analysis also revealed that CsPME genes that belong to the same group/type generally display very similar gene structures, as shown in Fig. 1A.

Expression profiling of Citrus PME genes in different tissues and fruit developmental stages

To further explore the spatial and temporal expression profiles of the 48 identified CsPME genes in different tissues (leaf, root, stem, flower) and two fruit developmental stages (mature green and full-colored fruit), we retrieved raw RNA-seq data from a public database (NCBI). A Heatmap was generated from the log₂ normalized DEG analysis. As shown in Fig. 3, 46 of 48 CsPMEs showed different expression profiles in the tissues and fruit development stages investigated, indicating that CsPMEs might play diverse functional roles during the growth and development of citrus trees. No transcripts of CsPME2 (Cs1g05720.1) or CsPME18 (Cs3g15260.1) were detected in the tissues used in this analysis (Fig. 3). Two genes, CsPME24 (Cs4g06710.1) and CsPME34 (Cs5g33450.1) have splicing variant forms, CsPME25 (Cs4g06710.2) and CsPME35 (Cs5g33450.2), respectively (https://www.citrusgenomedb.org/). In general, we found no close relationships between the PME group/type and the spatial/temporal expression profiles. The same paralogous gene pairs showed similar expression patterns, that is, high expression in the root (CsPME30/CsPME31 and CsPME8/CsPME9) and high expression in both the flower and mature green fruit (CsPME21/CsPME22), indicating that they might have been formed by segmental duplication and retained their function during genome evolution. The Heatmap analysis showed that the expressed CsPMEs can be classified into 8 groups, A to H. Group A (CsPME24/25 and CsPME28) was highly expressed in root and leaf tissue. Group B was specifically expressed in the root and contains 15 genes (CsPME7, CsPME8, CsPME10, CsPME11, CsPME14, CsPME16, CsPME28, CsPME29, CsPME30, CsPME31, CsPME32, CsPME33, CsPME40, CsPME41, and CsPME42). Group C was highly expressed in the root and flower and contains 7 genes (CsPME1, CsPME6, CsPME9, CsPME17, CsPME36, CsPME43, and CsPME47). Group D was specifically expressed in the flower and contains 7 genes (CsPME3, CsPME5, CsPME26, CsPME34/35, CsPME39, CsPME45, and CsPME46). Group E was highly expressed in the flower and mature green fruit and contains 8 genes (CsPME3, CsPME4, CsPME19, CsPME20, CsPME21, CsPME22, CsPME38, and CsPME48). Group F (CsPME12, CsPME23, and CsPME27) was highly expressed in the leaf tissue. Group G (CsPME44) was specifically expressed in the stem. Group H (CsPME15 and CsPME37) was highly expressed in full colored fruit.

Fig. 3. Expression profiling of the CsPME genes in different tissues and fruit developmental stages. Raw RNA-seq data were retrieved from a public source (SRA data from NCBI). The extracted FPKM values were log₂ normalized to display the Heatmap. This expression profiling was carried out in three biological repeats for each tissue or developmental stage. The red and dark purple boxes denote high levels of expression and low levels of expression, respectively. LF, leaf; RT, root; ST, stem; FW, flower; MGF, mature green fruit; FCF, full-colored fruit

As cell wall-localized enzymes, plant PMEs play crucial roles not only to regulate growth and development, but also to respond to various environmental stresses by contributing to changes in the cell-wall composition/ structure through the remodeling of HGA. In plant genomes, PMEs are encoded by a large multigene family and are differentially expressed in distinct tissues and in biotic/abiotic stress responses. Recent advances in complete plant genome sequences have accelerated the identification and detailed characterization of multigene family members, such as PME isoforms. In this report, we identified PME gene family members in the sweet orange genome and performed detailed molecular characterization, including gene expression profiling, using public transcriptome data.

Up to now, genome-wide analyses of higher land plant genomes revealed that PME isoforms are encoded by a large multigene family, ranging from 41 (rice and Brachypodium distachyon) to 127 genes (soybean). Dicot genomes contain more PME genes than grass species (Pelloux et al. 2007; Sénéchal et al. 2014b; Wang et al. 2021). In this study, we identified 48 PME genes from the sweet orange genome (Fig. 1 and Supplementary Table 2), which is fewer genes than the model plant Arabidopsis, which has 66 genes. The relatively small number of PME genes in grass species is known to be related to the fact that HGA is much less abundant and less methyl-esterified in grasses than in dicots (Pelloux et al. 2007; Sénéchal et al. 2014b).

The 3D crystallographic structure of plant and bacterial PMEs and their co-crystallization with HGA have revealed various biophysical properties, such as their folding topology, carbohydrate-binding sites, catalytic sites, and action mechanism of hydrolysis (Pelloux et al. 2007; Sénéchal et al. 2014b). Previous amino acid sequence alignments of PMEs from plants and microorganisms revealed that six amino acid residues in five characteristic sequence segments are strictly conserved, supporting the crystallization data (Markovič and Janeček 2004). The alignments of the PME domains of the 48 CsPMEs reported here indicate that the segments and amino acid residues are also highly conserved. Three amino acid residues, Gly154, Gly161, and Arg225 were completely conserved among the 48 CsPMEs (Supplementary Fig. 1). However, changes in the Gly44, Asp157, and Trp227 residues were found in some CsPMEs, and Gly44 was missing from CsPME7. Future studies are needed to confirm whether the CsPMEs with variations in the conserved residues have biochemical PME activity. The importance of the Gln135Asp136 residues in the active site of a tobacco PME was confirmed in transgenic plants expressing a mutated form (Dorokhov et al. 2006).

Studies during the past two decades have used sense/ antisense transgenic plants and mutants and found that PMEs play crucial roles in plant vegetative and reproductive development, including pollen development, pollen tube growth, cross incompatibility, embryo/seed development, root development, fruit ripening, and stomatal function, and in responses to biotic/abiotic stresses (Micheli 2001; Sénéchal et al. 2014b; Wang et al. 2022; Wu et al. 2018). Here we discuss the possible functional roles of some CsPME genes by comparing them with the known functions of PME genes that have already been studied in Arabidopsis, tomato, and maize. Table 1 summarizes those CsPME genes, their possible functional orthologs, and their possible biological functions in plant species including citrus.

Table 1

List of CsPME genes showing close phylogenetic relationship with the PMEs whose biological functions have been studied in Arabidopsis, tomato, and maize

CsPMEs	Tissues highly expressed in citrus	Possible functional orthologs in plant species	Functional roles in plant species
CsPME1	Flower	ZmPME3	Cross incompatibility
CsPME2	N.D.*	VGD1 (AT2G47040), VGDH1 (AT2G47030)	Reproductive development (pollen, pollen grains, and pollen tube growth)
CsPME3	Root	AtPME3 (AT3G14310)	Root growth and development. Defense responses against biotic stresses
CsPME12	Leaf	AtPME6/HMS (AT1G23200)	Embryo and seed development. Stomatal movement
CsPME19	Mature green fruit	Pmeu1 (Solyc03g123630)	Fruit ripening
CsPME27	Flower	QRT1 (AT5G55590)	Reproductive development (pollen, pollen grains, and pollen tube growth)
CsPME38	Flower	AtPPME1 (AT1G69940), AtPME48 (AT5G07410)	Reproductive development (pollen, pollen grains, and pollen tube growth)
CsPME41	Root	AtPME17 (AT2G45220)	Root growth and development
CsPME45	Flower	ZmGa1p	Cross incompatibility
CsPME47	Root, Flower	Solyc03g083360	Fruit ripening

*N.D. indicates that transcripts were not detected in the tissues investigated in this study.

Plenty of PME genes in many plant species are dominantly expressed in whole flowers or specific floral tissues such as anthers, pollen, pollen grains, pollen tubes or corn silk (Kim et al. 2010; Kim et al. 2015; Louvet et al. 2006; Moran Lauter et al. 2017; Wang et al. 2021; Wen et al. 2020; Zhang et al. 2019), indicating that they have a specific biological function in reproductive development. Arabidopsis QUARTET1 (QRT1, AT5G55590), which encodes a PME isoform, plays a key role in the separation of pollen grains from pollen tetrads after microsporogenesis in the pollen mother cell (Francis et al. 2006). Defects in Arabidopsis VANGUARD1 (VGD1, AT2G47040) and AtPPME1 (AT1G69940) genes retarded pollen tube growth within the style and transmitting tract, resulting in significant male sterility (Jiang et al. 2005; Tian et al. 2006). VGDH1 (AT2G47030), a homolog of VGD1, was specifically expressed in dry pollen grains and functionally complemented the vgd1 mutant phenotype (Jiang et al. 2005). Leroux et al. (2015) reported that AtPME48 (AT5G07410) was highly expressed in male gametophytes and dry and imbibed pollen grains and was involved in pollen germination by influencing the mechanical properties of the intine wall during maturation of the pollen grain. The CsPME27 gene showed a very close phylogenic relationship with QRT1 (Fig. 2). The CsPME2 and CsPME38 genes shared phylogenetic clades with VGD1/VGDH1 and AtPPME1 (or AtPME48), respectively (Fig. 2). The transcripts of CsPME27 and CsPME38 were also abundantly detected in flowers (Fig. 3), whereas CsPME2 transcripts were not detected in the citrus tissues used in this analysis. The CsPME2 gene could be expressed in response to internal or external stimuli or in specific tissues we did not test, such as grain pollens. Therefore, those three CsPME genes could be involved in reproductive development.

CsPME1 and CsPME45 were highly expressed in flowers (Fig. 3) and belonged to the same clades as two maize PME genes, ZmPME3 and ZmGa1p, respectively (Fig. 2). ZmGa1p and ZmPME3 were specifically expressed in pollen and corn silk, respectively, and act as a male determinant and a female determinant, respectively, in the complex regulation of unilateral cross incompatibility (Moran Lauter et al. 2017; Wang et al. 2022; Zhang et al. 2018). Citrus also has complicated reproductive biology, including male/female sterility and self- and cross-(in)compatibility (Pok et al. 2015). Understanding the underlying molecular mechanisms of (in)compatibility in citrus could be used to improve citrus breeding. Further studies on the CsPME1 and CsPME45 genes will shed light on their functional roles in the reproductive development of citrus.

Several studies using Arabidopsis knockout (KO) mutants and overexpressors revealed that the degree of methylesterification (DM) of HGAs controlled by PMEs plays crucial roles in root development. The atpme3 KO mutant, which showed decreased PME activity, had a 20% reduction in root length compared with the wild type, whereas transgenic lines overexpressing AtPME3 (AT3G14310) produced longer roots than the wild type (Hewezi et al. 2008). A study using different atpme3 mutant allele also revealed that reduced PME activity in the mutant correlated with an increase in the DM of HGA (Guénin et al. 2011). The number of adventitious roots was increased in the atpme3 mutant, suggesting that the AtPME3 isoform plays roles in both root elongation and differentiation (Guénin et al. 2011). AtPME17 (AT2G45220) was highly co-expressed with and processed by a subtilisin-like serine protease (AtSBT3.5) to release a mature apoplastic PME. The root length of the atpme17 KO mutant was longer than that of the wild type (Sénéchal et al. 2014a). CsPME3 and CsPME41, which belong to group 2/type I and are expressed in root tissue, showed very close phylogenetic relationships with AtPME3 and AtPME17, respectively (Figs. 1 to 3). These results indicate that CsPME3 and CsPME41 are possible functional orthologs of their counterparts in Arabidopsis and could play roles in root growth and development.

AtPME6/HMS/HIGHLY METHYL ESTERIFIED SEEDS (AT1G23200) mRNA was abundantly found in both the seed coat and the embryo during mucilage secretion in Arabidopsis. The atpme6/hms KO mutant displayed altered embryo morphology and mucilage extrusion due to defects in cell expansion during embryo development, indicating that AtPME6/HMS is required for cell wall loosening in the developing embryo (Levesque-Tremblay et al. 2015). AtPME58 (AT5G49180) was also specifically expressed in mucilage secretory cells in the seed coat. A detailed analysis of atpme58 KO mutant showed that AtPME58 also plays a role in mucilage structure and organization during embryo development (Turbant et al. 2016). Expression of the CsPME12 gene was highly detected in leaf (Fig. 3), but its mRNA level in the developing seeds has not yet been determined. The phylogenetic tree showed that CsPME12 is a possible functional ortholog of AtPME6/HMS (Fig. 2), and thus it might play a role in embryo/seed development by controlling mucilage structure and organization.

The roles of pectin modification during fruit development and maturation have been extensively studied in tomato (Wang et al. 2018). Fruit softening is also an important issue in post-harvest storage in the citrus industry. The cDNAs of several PMEs have been isolated from tomato fruit tissues, and their antisense suppression lines have been characterized (Phan et al. 2007; Simons and Tucker 1999; Tieman et al. 1992; Wen et al. 2013). Antisense suppression of the Pmeu1 (Solyc03g123630) gene in tomato enhanced the rate of softening during ripening (Phan et al. 2007). Wen et al. (2020) identified PME gene family from a genome-wide analysis of tomato and analyzed gene expression profiles during fruit development. Based on their detailed analysis, they suggested that several PME genes might play a role in tomato fruit ripening. In our results, CsPME19 was highly expressed in mature green fruit (Fig. 3) and belonged to the same clade as Solyc03g123630 in tomato (Fig. 2). The expression of Solyc03g083360, which showed a very close phylogenetic relationship with CsPME47, was detected in young tomato fruits and induced by ethylene treatment (Wen et al. 2020). These results suggest that CsPME19 and CsPME47 might be involved in fruit ripening.

Increased biotic and abiotic stresses caused by global climate change threatens the sustainable production of most crops, including citrus. Plant cells have been shown to develop elaborate systems for responding to a variety of challenges, including environmental stresses. The expression of plant PME genes was up- or down-regulated by abiotic stresses such as salt, cold, heat, drought, and wounding, and by biotic stresses such as viruses, bacteria, fungi, nematodes, insects, and parasitic plants (Pelloux et al. 2007; Sénéchal et al. 2014b; Wu et al. 2018). During an infection of plant-parasitic cyst nematodes into host cells, a cellulose binding protein (CBP) secreted from the pathogen directly interacted with cell wall-modifying PMEs to aid in root penetration and migration (Hewezi et al. 2008). Transgenic Arabidopsis overexpressing AtPME3 (AT3G14310) exhibited increased susceptibility to the parasitic nematode Heterodera schachtii, whereas the atpme3 KO mutant displayed an opposite phenotype, indicating that PME cooperates with CBP to aid cyst nematode parasitism (Hewezi et al. 2008). AtPME3 acts as a susceptibility factor and is necessary for successful colonization by necrotrophic pathogens such as Botrytis cinerea and Pectobacterium carotovorum (Raiola et al. 2011). CsPME3, which belongs to group 2/type I and was highly expressed in the root, flower, and mature green fruit, showed a very close phylogenetic relationship with AtPME3 (Figs. 2 and 3), indicating that CsPME3 is a possible functional ortholog of AtPME3 and might play a role in defense responses against biotic stresses.

Guard cells perceive a multitude of endogenous and environmental stimuli, including hormonal cues, light, CO² concentration, pathogen, drought, and heat. The opening/ closure of stomata pores is strictly regulated by an integration of environmental stimuli and endogenous hormonal signals. The highly specialized walls of guard cells enable them to undergo large and reversible deformation during the construction of stomata (Wu et al. 2018). Guard cell walls are rich in unesterified HGA (Amsbury et al. 2016). Arabidopsis PME6 expression was increased in guard cells, and the guard cells of its KO mutant had walls enriched in methyl-esterified HGA. The mutant also showed a decreased dynamic range in response to triggers of stomatal opening/closure such as drought (Amsbury et al. 2016). Huang et al. (2017) also found that AtPME34 (AT3G49220) is responsible for controlling stomatal movement by regulating the flexibility of guard cell walls under heat stress. CsPME12 was highly expressed in leaf tissue and showed a very close relationship to AtPME6. Therefore, CsPME12 not only plays a role in embryo and seed development, but also can be involved in environmental stress responses by regulating stomatal opening/ closure.

Plant PMEs are known to play crucial roles in regulating a wide range of plant growth and developmental processes through their remodeling of pectin polysaccharide, a major cell-wall component. To better understand the functional role of PME genes in the growth and development of citrus trees, we identified 48 genes that encode CsPME isoforms in a reference Citrus genome. Some CsPMEs showed very close phylogenetic relationships with PMEs whose functional roles have been addressed in other plant species. Expression profiling using transcriptome data revealed that most CsPMEs displayed tissue-specific expression patterns, indicating that each CsPME isoform has a specific functional role. Based on our phylogenetic analysis and expression profiling data, we suggest that CsPMEs also play an important role in regulating reproductive/vegetative development and responding environmental challenges in citrus. Our results will accelerate efforts to unveil the biological functions of individual CsPMEs. Identifying genetic variations among CsPMEs in existing citrus genetic resources will enable the development of molecular markers tightly linked to various agronomic traits, including male sterility, stress resistance, and fruit softening.

This research was supported by the Cooperative Research Program for Agriculture Science & Technology Development, RDA (Project No. PJ01514103), and the Korea Institute of Planning and Evaluation for Technology in Food, Agriculture, and Forestry (IPET) (Grant No. 322072031HD040), Republic of Korea. The authors appreciate Grace Kim (Yieun Kim) for her illustration work.