Ectopic Expression of JcCPL1, 2, and 4 Affects Epidermal Cell Differentiation, Anthocyanin Biosynthesis and Leaf Senescence in Arabidopsis thaliana

The CAPRICE (CPC)-like (CPL) genes belong to a single-repeat R3 MYB family, whose roles in physic nut (Jatropha curcas L.), an important energy plant, remain unclear. In this study, we identified a total of six CPL genes (JcCPL1–6) in physic nut. The JcCPL3, 4, and 6 proteins were localized mainly in the nucleus, while proteins JcCPL1, 2, and 5 were localized in both the nucleus and the cytoplasm. Ectopic overexpression of JcCPL1, 2, and 4 in Arabidopsis thaliana resulted in an increase in root hair number and decrease in trichome number. Consistent with the phenotype of reduced anthocyanin in shoots, the expression levels of anthocyanin biosynthesis genes were down-regulated in the shoots of these three transgenic A. thaliana lines. Moreover, we observed that OeJcCPL1, 2, 4 plants attained earlier leaf senescence, especially at the late developmental stage. Consistent with this, the expression levels of several senescence-associated and photosynthesis-related genes were, respectively, up-regulated and down-regulated in leaves. Taken together, our results indicate functional divergence of the six CPL proteins in physic nut. These findings also provide insight into the underlying roles of CPL transcription factors in leaf senescence.


Introduction
MYB transcription factors constitute one of the largest families of plant transcription factors. A protein encoded by an MYB gene usually includes one to four incompletely repeated MYB domains, which are named R1, R2 and R3 according to their sequence similarity with the c-MYB protein [1,2]. A MYB domain generally contains 52 amino acid residues which form three a-helices, of which the second and third make up a structure that interacts with the major groove of DNA [1,2]. According to the number of MYB domains, the MYB transcription factor superfamily can be divided into four families: R2R3-MYB, 3R-MYB, 4R-MYB and MYB-related [3]. MYB-related proteins usually contain a single MYB domain, but sometimes there are two MYB domains, and they have been classified into five subfamilies: CCA1-like, CAPRICE (CPC)-like (CPL), TBP-like, I-box-binding-like and R-R-type [4].

Identification and Phylogenetic Analysis of CPL Transcription Factors in Physic Nut
To identify CAPRICE (CPC)-like (CPL) genes in physic nut, BLAST searches were carried out against the public physic nut genome and proteome database [48,50], using the amino acid sequences of Arabidopsis thaliana CPC proteins as queries. As a result, a total of six loci putatively encoding CPLs were identified and named JcCPL1 to JcCPL6; they are located on chromosome 3 (JcCPL1 and JcCPL2), 5 (JcCPL3), 8 (JcCPL4), 7 (JcCPL5), and 9 (JcCPL6). All JcCPL genes consist of three exons and two introns, and the proteins they encode contain 74 to 87 amino acid residues. After analyzing the overall exon-intron arrangements, we observed that JcCPL genes share the same structure as model Ia and Ic R3 type MYB domains [49], which have an intron within the MYB domain coding sequences ( Figure 1A).

Identification and Phylogenetic Analysis of CPL Transcription Factors in Physic Nut
To identify CAPRICE (CPC)-like (CPL) genes in physic nut, BLAST searches were carried out against the public physic nut genome and proteome database [48,50], using the amino acid sequences of Arabidopsis thaliana CPC proteins as queries. As a result, a total of six loci putatively encoding CPLs were identified and named JcCPL1 to JcCPL6; they are located on chromosome 3 (JcCPL1 and JcCPL2), 5 (JcCPL3), 8 (JcCPL4), 7 (JcCPL5), and 9 (JcCPL6). All JcCPL genes consist of three exons and two introns, and the proteins they encode contain 74 to 87 amino acid residues. After analyzing the overall exon-intron arrangements, we observed that JcCPL genes share the same structure as model Ia and Ic R3 type MYB domains [49], which have an intron within the MYB domain coding sequences ( Figure 1A).
To examine the evolutionary relationships of physic nut CPL genes, a phylogenetic analysis was carried out using CPL proteins from physic nut and A. thaliana. The results indicate that CPL proteins were divided into two groups: JcCPL3, JcCPL4 and JcCPL6 were more closely related to the TRY group (group I) of A. thaliana, while JcCPL1, JcCPL2 and JcCPL5 were more closely related to the CPC group (group II) of A. thaliana ( Figure  1B).  and CPL proteins, indicating the exon-intron structure models. Intron positions, relative to the amino acid residues, which are marked in red, are indicated by arrows. Arrows pointing to amino acids indicate that splicing occurred within the amino acid coding sequences. The sequences of two conserved amino acids motifs are marked with black lines. The primary and secondary structures of a typical R3 MYB according to Dubos et al. (2010) [3] are indicated. H, helix; T, turn; F, Phenylalanine; M, Methionine, W, Tryptophan; X, any amino acid. (B) Phylogenetic analysis of the CPC-like subfamily of A. thaliana and physic nut using amino acid sequences. Bootstrap values from 100 replications are indicated at the branch points. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances that were used to infer the phylogenetic tree.
To examine the evolutionary relationships of physic nut CPL genes, a phylogenetic analysis was carried out using CPL proteins from physic nut and A. thaliana. The results indicate that CPL proteins were divided into two groups: JcCPL3, JcCPL4 and JcCPL6 were more closely related to the TRY group (group I) of A. thaliana, while JcCPL1, JcCPL2 and JcCPL5 were more closely related to the CPC group (group II) of A. thaliana ( Figure 1B).
The JcCPL proteins have a conserved R3 MYB domain, but variable N-terminal and C-terminal regions. According to a previous study, the conserved amino acid motif (Motif 1, M1) of [D/E]LX 2 [R/K]X 3 LX 6 LX 3 R (X represents any amino acid), located on helix 1 and helix 2 of the R3 MYB domain, is crucial for interaction between R3 MYB and bHLH proteins [14]. Another motif, WXM (Motif 2, M2), which is located on helix 3 of the MYB domain, is related to cell-to-cell movement of the CPC protein in A. thaliana [31]. Our results show that the two amino acid motifs in group I JcCPL proteins (JcCPL3, JcCPL4 and JcCPL6) were conserved. However, in group II JcCPL proteins (JcCPL1, JcCPL2 and JcCPL5), the first amino acid D/E in the M1 motif was replaced by S, A or T, and the amino acid M in the M2 motif was replaced by S ( Figure 1A).

Expression of JcCPL Genes in Different Organs
After detecting ESTs representing six JcCPL genes by analysing the expressed sequence tag (EST) databases of physic nut [48], we measured the expression levels of six JcCPL transcripts by qRT-PCR using leaves, roots, and young bark of stems from 8-week-old plants, and flowers and developing seeds from 5-year-old plants. The results show that the JcCPL1 gene was highly expressed in flowers and leaves ( Figure 2A). JcCPL2 was expressed highly in young bark and in the middle and late developmental stages of seeds ( Figure 2B). The JcCPL3 gene was expressed at high levels in roots, flowers and in the early and middle developmental stages of seeds ( Figure 2C). JcCPL4 was expressed highly in flowers and in the early developmental stage of seeds ( Figure 2D). JcCPL5 was expressed highly in roots, leaves, flowers and in the early developmental stage of seeds ( Figure 2E), while JcCPL6 was expressed mainly in seeds in the middle and late developmental periods ( Figure 2F). The different expression patterns of JcCPLs in different organs of physic nut implied that they may possess functional diversity.

Subcellular Localization of JcCPL Proteins
To determine the subcellular localizations of JcCPL proteins, we made constructs each consisting of the coding domain sequence of one of the JcCPL genes fused with a YFP gene under the control of the cauliflower mosaic virus (CaMV) 35S promoter in a transient expression vector. Each recombinant was transformed into A. thaliana protoplasts. For JcCPL1, JcCPL2 and JcCPL5 proteins, the fluorescent signals were detected in both cell nuclei and cytoplasm, whilst signals for JcCPL3, JcCPL4 and JcCPL6 were mainly in cell nuclei ( Figure 3).

Overexpression of JcCPL1, 2, and 4 Influences the Formation of Trichomes and Root Hairs in A. thaliana
To investigate the biological roles of JcCPL genes, each of these genes was overexpressed in A. thaliana and three homozygous transgenic A. thaliana lines for each gene were obtained. Semi-quantitative RT-PCR showed that the expression levels of the six JcCPL genes increased to different degrees in the transgenic lines ( Figure 4A

Subcellular Localization of JcCPL Proteins
To determine the subcellular localizations of JcCPL proteins, we made constructs each consisting of the coding domain sequence of one of the JcCPL genes fused with a YFP gene under the control of the cauliflower mosaic virus (CaMV) 35S promoter in a transient expression vector. Each recombinant was transformed into A. thaliana protoplasts. For JcCPL1, JcCPL2 and JcCPL5 proteins, the fluorescent signals were detected in both cell nuclei and cytoplasm, whilst signals for JcCPL3, JcCPL4 and JcCPL6 were mainly in cell nuclei ( Figure 3). A. thaliana R3 MYB genes are involved in trichome formation and root hair formation [51]. In this study, we found that overexpression of JcCPL1, JcCPL2, or JcCPL4 in A. thaliana (OeJcCPL1, 2, 4) resulted in a major reduction in trichome formation on rosette leaves and inflorescence stems ( Figure 4B,C,E,F,H,I). Line 3 (OE3) of OeJcCPL2 formed some trichomes on its leaves and inflorescence stems, probably due to the low expression level in this line. In addition, the number of root hairs increased significantly in OeJcCPL1, 2, and 4 plants compared with WT plants ( Figure 5). Unexpectedly, no significant changes were observed in the root hairs of overexpression JcCPL3, JcCPL5, or JcCPL6 plants (OeJcCPL3, 5, 6) (Supplementary Figure S1C,F,I). Similarly, no abnormal trichome formation was observed on leaves and inflorescence stems in OeJcCPL3, 5 and 6 (data not shown). Accordingly, these three genes were not studied in subsequent experiments.

Overexpression of JcCPL1, 2, and 4 Reduces Anthocyanin Content in A. thaliana
The CPL genes also play roles in regulating anthocyanin biosynthesis in plants [40,52] and CPC is a negative regulator of anthocyanin biosynthesis [11]. The anthocyanin content of A. thaliana shoots was very low under normal growth condition, so the content was measured in plants grown on an anthocyanin-inducing medium (AIM). In this study, the sterilized seeds were sown in 1/2 MS medium for 4 days, and then transferred to AIM containing 5% sucrose for 7 days. Subsequently, a lower level of anthocyanin accumulation in shoots could be easily observed in OeJcCPL1, 2, and 4 shoots ( Figure 6A-C). In most shoots of OeJcCPL1, 2, and 4 plants, the anthocyanin contents were 50% lower than in WT ( Figure 6D-F).

Overexpression of JcCPL1, 2, and 4 Influences the Formation of Trichomes and Root Hairs in A. thaliana
To investigate the biological roles of JcCPL genes, each of these genes was overexpressed in A. thaliana and three homozygous transgenic A. thaliana lines for each gene were obtained. Semi-quantitative RT-PCR showed that the expression levels of the six JcCPL genes increased to different degrees in the transgenic lines ( Figure 4A A. thaliana R3 MYB genes are involved in trichome formation and root hair formation [51]. In this study, we found that overexpression of JcCPL1, JcCPL2, or JcCPL4 in A. thaliana (OeJcCPL1, 2, 4) resulted in a major reduction in trichome formation on rosette leaves and inflorescence stems ( Figure 4B,C,E,F,H,I). Line 3 (OE3) of OeJcCPL2 formed some trichomes on its leaves and inflorescence stems, probably due to the low expression level in The anthocyanin biosynthesis pathway can be divided into early biosynthesis genes (EBGs) (including CHS, CHI and F3H) and late biosynthesis genes (LBGs) (including DFR, ANS, UFGT) [15,35,36]. The FLS1 and BAN genes are involved in the biosynthesis of flavonoids and proanthocyanidins, respectively [15,38]. To determine which steps were down-regulated in the anthocyanin biosynthesis pathway in JcCPL1, JcCPL2 and JcCPL4 overexpressing plants, the expression levels of several relevant genes were tested by qRT-PCR analysis. The results show that the expression levels of CHI, CHS, F3H, FLS1 and BAN gene in OeJcCPL1, OeJcCPL2 and OeJcCPL4 shoots were not significantly changed when compared with WT shoots (Figure 7). However, the expression levels of DFR, ANS and UF3GT were significantly lower in these shoots than in WT (Figure 7). These results indicated down-regulation of the LBGs of anthocyanin synthesis pathway in the OeJcCPL1, OeJcCPL2 and OeJcCPL4 shoots. this line. In addition, the number of root hairs increased significantly in OeJcCPL1, 2, an 4 plants compared with WT plants ( Figure 5). Unexpectedly, no significant changes wer observed in the root hairs of overexpression JcCPL3, JcCPL5, or JcCPL6 plants (OeJcCPL3 5, 6) (Supplementary Figure S1C,F,I). Similarly, no abnormal trichome formation was ob served on leaves and inflorescence stems in OeJcCPL3, 5 and 6 (data not shown). Accord ingly, these three genes were not studied in subsequent experiments.

Overexpression of JcCPL1, 2, and 4 Reduces Anthocyanin Content in A. thaliana
The CPL genes also play roles in regulating anthocyanin biosynthesis in plants [40,5 and CPC is a negative regulator of anthocyanin biosynthesis [11]. The anthocyanin conte of A. thaliana shoots was very low under normal growth condition, so the content w measured in plants grown on an anthocyanin-inducing medium (AIM). In this study, th sterilized seeds were sown in 1/2 MS medium for 4 days, and then transferred to AIM containing 5% sucrose for 7 days. Subsequently, a lower level of anthocyanin accumul tion in shoots could be easily observed in OeJcCPL1, 2, and 4 shoots ( Figure 6A,B,C). most shoots of OeJcCPL1, 2, and 4 plants, the anthocyanin contents were 50% lower tha in WT ( Figure 6D,E,F).

Overexpression of JcCPL1, 2, and 4 Reduces Anthocyanin Content in A. thaliana
The CPL genes also play roles in regulating anthocyanin biosynthesis in plants [40,52 and CPC is a negative regulator of anthocyanin biosynthesis [11]. The anthocyanin conten of A. thaliana shoots was very low under normal growth condition, so the content wa measured in plants grown on an anthocyanin-inducing medium (AIM). In this study, th sterilized seeds were sown in 1/2 MS medium for 4 days, and then transferred to AIM containing 5% sucrose for 7 days. Subsequently, a lower level of anthocyanin accumula tion in shoots could be easily observed in OeJcCPL1, 2, and 4 shoots ( Figure 6A,B,C). I most shoots of OeJcCPL1, 2, and 4 plants, the anthocyanin contents were 50% lower tha in WT ( Figure 6D,E,F).  overexpressing plants, the expression levels of several relevant genes were tested by qRT-PCR analysis. The results show that the expression levels of CHI, CHS, F3H, FLS1 and BAN gene in OeJcCPL1, OeJcCPL2 and OeJcCPL4 shoots were not significantly changed when compared with WT shoots (Figure 7). However, the expression levels of DFR, ANS and UF3GT were significantly lower in these shoots than in WT (Figure 7). These results indicated down-regulation of the LBGs of anthocyanin synthesis pathway in the OeJcCPL1, OeJcCPL2 and OeJcCPL4 shoots.

Overexpression of JcCPL1, 2, and 4 Promotes Leaf Senescence in A. thaliana
Leaf chlorophyll content is an important indicator of plant senescence [53]. When monitoring leaf development, we found that the leaves of OeJcCPL1, 2, and 4 transgenic plants turned yellow earlier than those of WT plants. Most rosette leaves of WT appeared green apart from a few lower leaves under routine cultivation conditions. In contrast, most leaves of OeJcCPL1, 2, and 4 transgenic plants appeared yellow with the exception of some upper leaves under the same cultivation conditions ( Figure 8A). To further investigate the roles of JcCPL1, 2, and 4 in plant senescence, the chlorophyll contents of leaves were measured at different developmental stages. At the 3-week stage (before flowering), only two lines of OeJcCPL1 had significantly lower chlorophyll content compared to WT ( Figure 8B).
At the 4-week stage (after flowering), the chlorophyll content of both OeJcCPL1 and OeJc-CPL2 plants was significantly decreased compared to WT ( Figure 8B). At the 8-week stage (late developmental stage), the chlorophyll contents of OeJcCPL1, 2, and 4 plants were all significantly decreased compared to WT. These results indicate that JcCPL1, 2 and JcCPL4 may promote leaf senescence at different stages of plant growth and development.
plants turned yellow earlier than those of WT plants. Most rosette leaves of WT appeared green apart from a few lower leaves under routine cultivation conditions. In contrast, most leaves of OeJcCPL1, 2, and 4 transgenic plants appeared yellow with the exception of some upper leaves under the same cultivation conditions ( Figure 8A). To further investigate the roles of JcCPL1, 2, and 4 in plant senescence, the chlorophyll contents of leaves were measured at different developmental stages. At the 3-week stage (before flowering), only two lines of OeJcCPL1 had significantly lower chlorophyll content compared to WT ( Figure  8B). At the 4-week stage (after flowering), the chlorophyll content of both OeJcCPL1 and OeJcCPL2 plants was significantly decreased compared to WT ( Figure 8B). At the 8-week stage (late developmental stage), the chlorophyll contents of OeJcCPL1, 2, and 4 plants were all significantly decreased compared to WT. These results indicate that JcCPL1, 2 and JcCPL4 may promote leaf senescence at different stages of plant growth and development.  To explore the relationship between chlorophyll content and senescence or photosynthesis-related genes, we measured the expression levels of senescence-or photosynthesis-related genes using 8-week-old leaves of OeJcCPL1, 2, and 4 transgenic plants. Consistent with their phenotype, the expression levels of several senescencerelated (SAG12, SAG113, NAP and SIRK) were increased ( Figure 8C), while the expression levels of photosynthesis-related genes (RBCL and RBCS2B) were decreased, in OeJcCPL1, 2, and 4 leaves compared to those in WT leaves ( Figure 8C). These results implied that genes JcCPL1, 2, and 4 play roles in the positive regulation of leaf senescence at different stages of plant development.

Discussion
In this study, six CAPRICE (CPC)-like (CPL) genes, JcCPL1-JcCPL6, were identified in the physic nut genome. Analysis of the structure of these revealed that the genes encoding the single repeat R3 MYB region of the six physic nut CPLs have the same exon-intron arrangement and the proteins they encode have high sequence similarity at the amino acid level in the MYB domain region. Our unrooted phylogenetic tree indicated that physic nut CPL proteins fell into two groups: JcCPL3, 4, and 6 (group I), and JcCPL1, 2, and 5 (group II) (Figure 1). This result is consistent with previous phylogenetic analysis which indicated that the seven CPL genes of Arabidopsis thaliana were divided into two groups, AtTCL1, AtTCL2, AtETC2 and AtTRY in group I, and AtCPC, AtETC1 and AtETC3 in group II [54]. The JcCPL proteins exhibited many divergences at the amino acid level, with variable amino acid sequences at the N-and C-terminals as well as in the conserved motifs within the MYB domain. In the M1 and M2 motifs, several amino acid residues were not conserved in group II JcCPL proteins (Figure 1). These results indicate that divergence of motifs within the JcCPL proteins had occurred before the differentiation of A. thaliana and physic nut, especially between group I and group II proteins.
Previous studies indicated that CPL proteins in plants were localized in the nucleus as well as the cytoplasm or plasma membrane [8,[55][56][57], or mainly in the nucleus [54]. In this study, we found that group I JcCPL proteins were localized mainly in the nucleus while group II JcCPL proteins were present in both the nucleus and cytoplasm under the same conditions. These results imply that amino acid divergence between the two group JcCPL proteins determined their different subcellular locations. As well as protein subcellular localization, spatiotemporal expression patterns are assumed to reflect the roles of genes [33]; in the present study, the differential patterns of expression indicate that 6 JcCPL genes may have different roles in the growth and development of physic nut.
However, none of the phenotypes mentioned above were observed in JcCPL3-, 5-, or 6overexpressing plants. Similar results have been observed in plants expressing TRY and ETC2, which did not show any obvious induction of root hairs [60]. Studies have indicated that the TRY and ETC2 members of the CPC family are characterized by rapid proteolysis, which is related to the extended C-terminal domain of these proteins, compared with that of other members [60]. Substitution of amino acids in the C-terminal regions of TRY and ETC2 conferred on them the ability to induce root hair formation [61]. These results indicate that the precise structure of each CPL protein is important in determining its properties. We therefore speculate that this might be one of the reasons why plants overexpressing the JcCPL3, 5, and 6 genes have no visible phenotype. However, further experiments are still needed to explore the specific mechanisms by which the products of JcCPL3, 5, and 6 act.
In addition, we found that the chlorophyll content of both OeJcCPL1 and OeJcCPL2 was significantly decreased compared to WT at four weeks. Compared to WT, OeJcCPL4 showed a significantly decreased chlorophyll content at eight weeks. The expression levels of several senescence-related genes at eight weeks were also changed. These results indicate that these three JcCPL genes function as positive regulators in leaf senescence, especially at the late stage of the plant growth and development. Although there is no literature about CPL genes having such roles, several MYB transcription factors related to the MYB-bHLH-WD40 (MBW) complex have been reported to be involved in the regulation of senescence in plants. GLABRA1 (GL1), a R2R3-MYB transcription factor in A. thaliana, is a component of the MBW complex [9,[62][63][64]. Besides being involving in trichome formation, GL1 has a positive role in dark-induced leaf senescence [65]. Several R2R3-MYB genes have also been reported to play roles in anthocyanin accumulation as well as leaf senescence. OsPL is an R2R3-MYB transcription factor, and the mutant pl exhibited a higher anthocyanin content and a lower chlorophyll content accompanied by down-regulation of photosynthesisrelated genes in flag leaves [66]. In A. thaliana, MYB75/PAP1 plays a significant role in plant development mediated by its phosphorylation status, including a delayed onset of senescence when plants overexpress MYB75 T131E [67]. In Liquidambar formosana, LfMYB113 is a homolog of A. thaliana MYB75. Transgenic tobacco containing 35S::LfMYB113 shows significant anthocyanin accumulation in leaves, and accelerated leaf senescence [68]. HAT1 interacts with MYB75 and thereby interferes with the MBW protein complex. HAT1 constrains anthocyanin accumulation by inhibiting the activities of this protein complex through blocking the formation of the protein complex and recruiting the TPL corepressor to epigenetically modulate the late anthocyanin biosynthetic genes [69]. Whether the regulation of leaf senescence by the genes JcCPL1, 2, and 4 is related to the MBW complex remains to be elucidated.

Plant Materials and Growth Conditions
The physic nut cultivar GZQX0401 was used in this study. The plants were cultivated in a greenhouse and experimental field at South China Botanical Garden (23 • 18 N, 113 • 36 E) under natural sunlight [48,70,71].
Arabidopsis thaliana (Columbia, Col-0) used for functional analysis of JcCPL genes was maintained in a growth cabinet (16 h light/8 h dark, 22 ± 2 • C, 80 µmol m −2 s −1 ). For the analysis of anthocyanin content and related genes, plants were grown vertically from sterilized seeds on 1/2 MS medium (1% sucrose and 1% agar, pH 5.8) for 4 days, then transferred to anthocyanin induction medium (AIM) agar plates (1/2 MS, 5% sucrose and 1% agar, pH 5.8) for 7 days. For observation of growth and developmental stage, plants were grown in pots containing a 1:1 mixture of vermiculite and peat moss. Whole shoots were sampled for qRT-PCR and anthocyanin estimation. For senescence-related gene analysis, the fifth and sixth leaves from the apices of 8-week-old plants were sampled and stored at −80 • C.

Sequence Database Searches, Alignment and Phylogenetic Analysis
The amino acid sequences encoded by the CAPRICE (CPC)-like (CPL) genes of A. thaliana were used to search for CPLs in the genome and predicted proteome of physic nut by tblastn and blastp, and 6 JcCPLs were identified in physic nut. Multiple sequence alignment was performed using Clustal W 1.83 [72] and DNAMAN 6.0 (Lynnon Biosoft Inc., Vandreuil, QC, Canada). The phylogenetic tree was constructed using MEGA 5.0 with neighbor-joining pairwise deletion and bootstrap test 100 times [73].

RNA Isolation and Expression Analysis
Roots, leaves and stems were sampled from 8-week-old physic nut plants. The root samples included all root tips about 10 mm in length, and leaf samples were the fourth fully expanded leaf from the apex, while the stem samples were young bark. Flowers and seeds at points 14,19,25,29,35,41 and 45 days after pollination (DAP) were harvested from 5-year-old physic nut plants. All samples were immediately put into liquid nitrogen and stored at −80 • C. The CTAB method was used for isolating total RNA from physic nut organs according to a previous study [70]. The isolation of total RNA from A. thaliana shoots and leaves was carried out using Trizol reagent (Invitrogen, Burlington, ON, Canada) following the manufacturer's instructions.
About 2 µg total RNA (for reverse transcription PCR, RT-PCR) and 5 µg total RNA (for quantitative real-time PCR, qRT-PCR) were used for synthesizing first-strand cDNAs according to the manufacturer's instructions (Promega, Madison, WI, USA). A JcActin (GenBank accession number HM044307.1) gene was used as internal control for physic nut, while AtACT2 (At3g18780) was used as internal control for A. thaliana. Three independent biological replicates, each with three technical repeats, were employed for semi-quantitative RT-PCR and qRT-PCR. For semi-quantitative RT-PCR, PCR products were separated on 1.5% agarose gels and stained with ethidium bromide. The gels were then photographed using a Gel Imaging System (Shanghai Bio-Tech, Shanghai, China), and an LCS480 system (Roche, Rotkreuz, Switzerland) was used for qRT-PCR [71]. The 2 − CT method was used for calculating gene expression levels [74]. The primers used in this work are listed in Supplementary Table S1.
For transformation into A. thaliana, fragments containing the coding domain sequences (CDS) of the 6 JcCPLs were cloned from the pMD18-T vector using the appropriate restriction enzyme sites (JcCPL1, 2, 3: Sac I/Pst I; JcCPL4, 6: Kpn I/Pst I; JcCPL5: Sac I/Sal I). Then they were ligated into the vector pCAMBIA1301, digested with the same restriction enzyme sites, under the control of the CaMV 35S promoter. The recombinant constructs were transferred into A. thaliana plants by the floral-dipping method [75]. Homozygous lines with single T-DNA insertions were selected for subsequent analysis. The primers used in this work are listed in Supplementary Table S1.

Subcellular Localization
For subcellular localization analysis, complete coding sequences without the stop codon were amplified by RT-PCR using the primers listed in Supplementary Table S1. The six modified JcCPL cDNAs were cloned into the pSAT6-YFP-N1 vector under the control of the CaMV 35S promoter at the EcoR I and BamH I restriction sites.
The naked pSAT6-YFP-N1 vector (YFP) and each of the JcCPL-YFP recombinants (JcCPL-YFP) were transformed into A. thaliana protoplasts using the PEG-Ca 2+ method [76]. The plasmid pSAT6-RFP-C1 containing the nuclear localization signal of endonuclease VirD2 under the control of the 35S promoter was co-transformed. For the study, 200 µL protoplast suspension were mixed with 20 µL plasmids (10 µL YFP/JcCPL-YFP + 10 µL NLS-RFP), and 220 µL PEG-Ca 2+ solution (40% PEG 4000, 0.2 M mannitol, 0.1 M CaCl 2 ,) and incubated at 24 • C for 5 min, then 880 µL solution W5 was added. The mixed suspension was centrifugated at 100× g for 2 min, and resuspended in 1 mL W5 solution for 12-24 h at 22 ± 2 • C in the dark. Then the YFP and RFP fluorescence of the protoplasts were observed with a LSM800 laser scanning confocal microscopy system (Carl Zeiss, Oberkochen, Germany). Images were captured at 488 nm and 543 nm laser excitation, and using 493 to 582 nm and 602 to 656 nm long-pass emission filters, for YFP and RFP, respectively.

Measurements of Root Hairs
Seedlings were cultivated on 1/2 MS medium (containing 1% sucrose and 1% agar, pH 5.8) vertically after surface sterilization. After 4 days, seedlings were photographed using a stereo microscope (LEICA, Wetzlar, Germany). Numbers of root hairs within 5 mm in length from the root tips were counted. Thirty seedlings were used for each line.

Measurement of Anthocyanin and Chlorophyll Content
For measurement of anthocyanin content, weighed shoots (ca. 15 mg) were incubated in 1.5 mL hydrochloric acid-methanol mix solution (v/v = 1:99) at 4 • C for 2 days. The absorbance (A) of the centrifuged extract was determined at 530 nm and 657 nm (A 530 and A 657 , respectively). The total anthocyanin content was expressed as A 530 − 0.25A 657 g −1 fresh weight [71,77].
The chlorophyll content was determined using 95% ethanol according to Ding et al. (2018) [78] with minor modification. The contents of chlorophyll a (Chl a), chlorophyll b (Chl b), and total chlorophyll were estimated from leaves (the 3th, 4th, 5th and 6th leaves from the bottom) of transgenic and WT plants at the 3-week stage and 4-week stage, and from leaves (the fifth and sixth leaves from the apex) of transgenic and WT plants at the 8-week stage [79].

Statistical Analysis
At least three biological repeats and technical repeats were used for all experiments, and data were tested for means using SAS software package version 9 with Duncan tests [80].