Peucedanum praeruptorum seeds were sown in pots containing soil and grown in a green house at 25°C, with 16 h daylight. These plants were used for further experiments when they were one or 6 months old.

Six-month-old P. praeruptorum plants were used for total RNA isolation. To obtain desired missing sequences, 5′ and 3′ RACE transcripts were generated using the SMARTer^TM RACE DNA Amplification Kit (Clontech Laboratories, Inc., Mountain View, CA, USA). The gene-specific primers used in this experiment are listed in Supplementary Table S1. The full-length cDNAs were verified by reamplification of the open reading frame (ORF) using forward primer and reverse primer (Supplementary Table S1). The PCR products of expected size were excised from the goldview-stained 1% (w/v) agarose gels, purified using the SanPrep Column DNA Gel Extraction Kit and ligated into the pMD19-T (Takara) vector for sequencing. The coding regions of genes were then deposited into the National Center for Biotechnology Information (NCBI) with the accession numbers of KX254614 (Pp4CL1), KX254613 (Pp4CL7), KX254612 (Pp4CL10).

Amino acid sequences were aligned by ClustalX program. Alignment results were analyzed to construct a phylogenetic tree using the neighbor-joining method (Tamura et al., 2011). Statistical robustness was ensured by the bootstrap test with 1,000 replicates. Default parameters were used for all applications. Accession numbers of the sequences used in this study are listed in Supplementary Table S2.

The coding region of Pp4CL1 was PCR-amplied using the forward and reverse primers 5′-ATGGGAGATTATGTAGCACCCAA-3′ and 5′-CATCCGGTGATCTTCCCAAATAA-3′, respectively and cloned into a pMD19-T (TaKaRa) vector for sequencing. Then the gene was subcloned into the Nde I-EcoR I site of the expression vector pET-28a containing a cleavable N-terminal His₆-tag. After verification by sequence analysis, the recombinant plasmid was transformed into E. coli BL21 (DE3) competent cells. Bacterial strain containing recombinant plasmid was grown in Luria-Bertani medium with 50 mg/L-kanamycin at 37°C. The expression of Pp4CL1 was induced with a final concentration of 0.4 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) when the cultures’ density reached A₆₀₀ = 0.5. After 8 h incubation on a rotary shaker (100 rpm) at 18°C, the E. coli cells were harvested by centrifugation and resuspended in buffer containing 50 mM NaH₂PO₄, 300 mM NaCl, pH 8.0. The cells were sonicated for 20 min under cooling conditions. The lysates were centrifuged for 30 min (12,000 rpm, 4°C), and the recombinant protein was purified by Ni-NTA affinity column using FPLC (ÄKTA, GE Healthcare Bio-Sciences) at 0.8 ml/min flow rate. Columns were washed by 50 mL 30 mM imidazole in a buffer containing 50 mM NaH₂PO₄ and 300 mM NaCl, pH 8.0. Pp4CL1 protein was eluted using a solution containing 50 mM NaH₂PO₄, 300 mM NaCl, and 250 mM imidazole, pH 8.0. Eluted fractions (3 mL each) with the highest protein levels were desalted on centrifugal filter devices (Amicon Ultra-4) into the buffer containing 20 mM NaH₂PO₄, pH 8.0. Desalted enzymes were stored at -80°C until use. The purification efficiency was monitored by SDS-PAGE (Supplementary Figure S1) and the protein concentration was determined by the Bradford method (Bradford, 1976).

To examine the biochemical properties of Pp4CLs, enzyme activity was assayed following previous methods with minor modifications (Knobloch and Hahlbrock, 1975). Briefly, the reaction mixture (200 μl) contained 50 mM NaH₂PO₄, 300 mM NaCl, 5 mM ATP, 5 mM MgC1₂, 3 μg of purified recombinant 4CLs protein and 0.3 mM substrates (p-coumaric acid, caffeic acid, ferulic acid, and trans-cinnamic acid, etc.), pH 8.0. The CoA (0.3 mM) was added to the mixture to initiate the reaction. After being incubated for 15 min at 37°C, the mixture was boiled for 10 min and centrifuged (10 min, 12000 rpm) to remove the protein. Blanks contained 0.3 mM substrate, 5 mM ATP, 5 mM Mg²⁺, 50 mM NaH₂PO₄, 300 mM NaCl and 3 μg of purified protein. UV absorptions were recorded continuously for 15 min at an interval of 60 s, using the blanks as background. Formations of the corresponding CoA esters were determined by monitoring the increase in absorption maximum at wavelengths 310 nm for cinnamoyl-, 333 nm for p-coumaroyl-, 333 nm for o-coumaroyl-, 346 nm for caffeoyl-, 346 nm for isoferuloyl- and 346 nm for feruloyl-CoA esters (Lee et al., 1997). The optimum temperature and pH of Pp4CL1 enzyme were analyzed, using 4-coumaric acid as substrate.

The reaction products were dried under nitrogen gas and then were dissolved in methanol for HPLC analysis. The processed samples (10 μl) were subjected to an Agilent 1290 LC System which was coupled to a multiple-wavelength diode array detector and equipped with a ZORBAX SB-C18 column (4.6 × 250 mm, 5 μm) for analysis. Separation column was maintained at 30°C with a flow rate of 1.0 mL/min. The mobile phase consisted of solvents A (0.1% ammonium acetate in double distilled water) and B (methanol). Gradient elution was done for 20-min and the conditions were as follows: 0–5 min, 5–10% B linear; 5–15 min, 10–60% B linear; 15–20 min, 60–90% B linear. Elution of compounds was monitored at 333 nm (p-coumaroyl-CoA and o-coumaroyl-CoA), 346 nm (caffeoyl-CoA, feruloyl-CoA, and isoferuloyl-CoA) and 310 nm (trans-cinnamoyl-CoA).

All reaction products were identified by LC-MS, using the above conditions. The HPLC system was coupled to Q-TOF MS spectrometer (Agilent Technologies, Santa Clara, CA, USA), equipped with an electrospray source in positive ion mode. The conditions of ESI source were as follows: drying gas (N₂) flow rate, 8.0 mL/min; nebulizer, 241 kPa (35 psig); fragmentor voltage, 150 V; collision energy, 30 eV; octopole radio frequency, 250 V and skimmer voltage, 60 V. In addition, the capillary voltage was set at 4.0 kV and the desolvation gas temperature was 325°C.

For methyl jasmonate (MeJA) elicitation, the roots of plants were treated with MeJA (Sigma–Aldrich, 95% purity) at a final concentration of 0.2 mM in 10% ethanol. The control was treated with the same amount of ethanol. All samples were harvested for analysis at 3, 6, 9, 12, and 24 h after elicitation. Meanwhile, the plants were also treated with PEG, NaCl, CuSO₄, heat and cold for a certain time. The above samples were ground in liquid nitrogen with a mortar and pestle and then stored at -80°C for RNA extraction.

Total RNA were extracted from P. praeruptorum roots, stems, leaves, seedlings in turn, using Trizol reagent following the manufacturer’s instructions (Invitrogen). The quantity and quality of RNA were determined using a Spectramax plus384 enzyme-labeling instrument (Molecular Devices, Sunnyvale, CA, USA) and 1% agarose gels. Total RNA were treated with RNase-free DNase I to remove DNA contamination and were reverse transcribed into cDNA using Vazyme’s HisScript QRT SuperMix. Gene expression was assayed by quantitative real-time PCR (qPCR) using the cDNA equivalent of total RNA. Gene-specific primers were designed for PCR amplification (Supplementary Table S1). qPCR was performed on a Roche Light Cycler 480 instrument using a 96 well plate by Vazyme’s SYBR Green PCR Master Mix. The PCR conditions for the amplification of 4CLs transcripts were as follows: 5 min of denaturation at 95°C, 40 cycles of 95°C for 30 s, 60°C for 30 s, followed by one cycle of 95°C for 15 s, 60°C for 60 s, and 95°C for 15 s. The gene expression values were calculated by 2^-ΔΔCt method (Schmittgen, 2006) and were normalized using GAPDH gene as a reference. At least three technical repetitions were performed.

The crystal structure of Populus tomentosa 4CL1 (Hu et al., 2010) was used to conduct a structure model for Pp4CL1. Based on the homology modeling and sequence alignment, some essential residues were selected for site-directed mutagenesis to investigate the key amino acid residues which made an effect on enzymatic specificity and activity. The Pp4CL1 mutants were generated using overlap PCR method and the primers used in this experiment are shown in Supplementary Table S1. Mutated genes were confirmed by sequencing and then transferred into E. coli BL21 (DE3). Purification and activity assays of mutant recombinant proteins were conducted using the same protocol with wild-type Pp4CL1. The substrates used in enzyme assays were p-coumaric, caffeic, ferulic, and isoferulic acids. The reaction products were analyzed by HPLC.

Construction of GFP fusion with the N terminal region of Pp4CLs and the transient expression in Arabidopsis protoplasts were conducted according to previously described protocol (Zhao et al., 2016). For amplification of Pp4CLs (Pp4CL1 and Pp4CL10), primers were designed and listed in Supplementary Table S1. Plasmid pCAMBIA1302-GFP with the cauliflower mosaic virus (CaMV) 35S promoter was used for the generation of GFP fusion constructs. The PCR fragments of two genes were subcloned into the pCAMBIA1302 vector in frame with the GFP. The ORF of PEX7 (Arabidopsis thaliana) was amplified by PCR using PTS1-Per-mcherry-F: CGCTTCTAGAATGCCGGTGTTCAAAGCTCC (Xba I) and PTS1-Per-mcherry-R: AATCCCCGGGACTGGCTCTAGGATCCATCCC (Sma I). The PCR fragment was cloned into P16ΔS:sXVE:mCherryC vector (Schlücking et al., 2013). Arabidopsis protoplasts were prepared and then transformed with the constructed pCAMBIA1302-4CLs plasmids and pCAMBIA1302 empty vector, respectively, using polyethylene glycol (PEG)-mediated transient gene expression (Wu et al., 2009). Meanwhile, Arabidopsis protoplasts were also co-transformed with pCAMBIA1302-Pp4CL1 plasmid and the constructed plasmid P16ΔS:sXVE:mCherryC-PEX7 (peroxisomal marker). Transient expression of GFP fusion proteins was observed 16 h after transformation by a laser scanning confocal microscope (LSM 780; Carl Zeiss, Jena, Germany), equipped with 20 × 0.8 Plan-Apochromat, 40 × 1.2 W C-Apochromat or 63 × /1.4 Oil Plan-Apochromat in multitrack channel mode.

A total of 17 cDNA fragments encoding putative CoA ligases (annotated as 4CL or 4CL-like) were identified in previous experiment (Zhao et al., 2015). Here, we investigated the expression profiles of these putative 4CL genes and three 4CL genes with relatively high expression levels were selected for further study. The full-length cDNA of genes were identified by 5′ and 3′ rapid amplification of cDNA ends (RACE). The three genes were named as Pp4CL1, Pp4CL7 and Pp4CL10. Pp4CL1 shared high nucleotide identity with 4CL1 from Petroselinum crispum, which, like P. praeruptorum, also belongs to Apiaceae family. The ORF lengths of Pp4CL1, Pp4CL7 and Pp4CL10 were 1629 bp (encoding 543 residues), 1626 bp (encoding 542 residues) and 1566 bp (encoding 522 residues), respectively. The amino acid sequences of the three genes contained highly conserved AMP-binding motif (Box I) and a Box II domain (GEICIRG domain).

A phylogenetic tree was constructed based on a series of 4CLs that have actual or putative activity toward p-coumaric acid or derivatives. Protein accession numbers and plant species used in the phylogenetic tree are shown in Supplementary Table S2. Phylogenetic analysis revealed that Pp4CL1, Pp4CL7, and Pp4CL10 belong to three different branches of the evolutionary tree (Figure 2). Pp4CL1 belonged to one clade in which related genes have traditional 4CL protein function. Pp4CL1 had higher homology with As4CL (Angelica sinensis), Pc4CL (Petroselinum crispum) and Dc4CL1 (Daucus carota), and the corresponding plants are all derived from the Umbelliferae family. Whereas, Pp4CL7 and Pp4CL10 were classified as members of other clades in which related proteins have not been demonstrated to possess 4CL activity. The sequence alignment was performed using related 4CL proteins with high sequence similarity. Pp4CL1 contains a highly conserved AMP-binding motif (Box I) and a Box II domain (GEICIRG domain). Some residues essential for the enzyme activity have been marked (Hu et al., 2010). The Ser-240 in Pt4CL1, highly conserved in 4CLs, was replaced by Ala-243 and key residues Lys-303 and Gly-306 in Pt4CL1 were also substituted by Met-306 and Ala-309 here, respectively (Figure 3).

The transcript levels of three 4CL genes in various tissues of P. praeruptorum were quantified by qPCR using gene-specific primers (Supplementary Table S1). The three 4CL genes showed different expression patterns in different tissues. The results showed that expression levels of Pp4CL1, Pp4CL7, and Pp4CL10 in roots were higher than those of leaves and stems. Meanwhile, Pp4CL1 had the highest expression level among all putative 4CL genes (Figure 4A). In addition, transcript levels of Pp4CL1 increased in mature roots compared with those in seedlings, which correlates to the accumulation of bioactive coumarins in roots. MeJA had strong effects on the expression levels of 4CLs in roots. Upon MeJA treatment, the transcript levels of three genes were significantly up-regulated, with the increasing duration of interaction (Figure 4B). Transcript levels of Pp4CL1 and Pp4CL7 in roots reached the maximum after 24 h whereas the expression level of Pp4CL10 was clearly enhanced at 9 h and then declined (Figure 4B). Subsequently, we also investigated the expression patterns of three genes in response to abiotic stresses such as PEG, NaCl, CuSO₄, heat and cold treatments. Upon the above, the expression level of Pp4CL1 significantly decreased. However, transcript abundance of Pp4CL7 was up-regulated under heat treatment, and the expression level of Pp4CL10 was clearly enhanced under PEG treatment (Figure 4C).

For purposes of determining whether Pp4CL1, Pp4CL7, and Pp4CL10 had the CoA ligase function, the coding regions of these cDNAs were cloned and expressed in E. coli BL21 (DE3). The activity of purified recombinant proteins toward several substrates, including p-coumaric, o-coumaric, ferulic, caffeic, cinnamic, isoferulic, and sinapic acids were tested (Figure 5). The results revealed that Pp4CL1 had distinct CoA ligation activity toward the above substrates except sinapic acid. Furthermore, Pp4CL1 exhibited higher activities toward p-coumaric and ferulic acids than other substrates. Pp4CL7 and Pp4CL10 had no activity toward hydroxycinnamic acid compounds (data not shown), indicating that they were not CoA ligases involved in the biosynthesis of coumarins, which was in agreement with the analysis of the phylogenetic tree. Therefore, Pp4CL7 and Pp4CL10 were classified as 4CL-like genes, which were annotated as being closely related to true 4CLs but of unknown biochemical function. To further confirm whether the two genes could function as fatty acid CoA ligases, a series of fatty acids were tested as substrates using recombinant Pp4CL7 and Pp4CL10 proteins. The results showed that neither of them had activity toward the fatty acids (data not shown).

We then focused on the specific properties of the Pp4CL1 enzyme. Pp4CL1 displayed extensive catalytic activities for hydroxycinnamate derivatives by converting them into the corresponding CoA esters. All CoA ligation reactions were monitored by HPLC and the identities of CoA thioester products were established by time-of-flight liquid chromatography-mass spectrometry (LC/TOF-MS) in positive ion mode. The six corresponding product peaks were detected in accordance with analysis results of HPLC profiles (Figure 5) and the reaction products were identified by high resolution mass spectrometry (HRMS). Dominant ions of p-coumaroyl-CoA were singly and doubly charged pseudo molecular ion [M + H]⁺ at m/z 914.1597 and [M + 2H]²⁺ at m/z 457.5840, respectively (Figure 5A). Dominant ions of feruloyl-CoA were m/z 944.1707 and m/z 472.5909 (Figure 5B). Dominant ions of all products identified by HRMS are also listed in Supplementary Table S3. In addition, the optimum pH of Pp4CL1 was approximately 6.5 (Supplementary Figure S2) and its optimum temperature was 37°C (Supplementary Figure S3).

It was predicted that Pp4CL1 and Pp4CL10 are localized in the cytosol using WoLF PSORT program. To further investigate the subcellular localization of the Pp4CL1 and Pp4CL10, their ORFs were cloned into pCAMBIA1302 so that the proteins were fused with green fluorescence protein (GFP) gene under the control of the 35S promoter from CaMV. The designed constructs were transferred into protoplasts of Arabidopsis by PEG-mediated transformation (Wu et al., 2009). The GFP:4CL fusion proteins were expressed in Arabidopsis protoplast. Arabidopsis protoplast containing the empty vector was used as a control. The green fluorescence was present throughout the cytoplasm in Arabidopsis protoplasts containing Pp4CL1-GFP. There was no overlap between GFP and the red autofluorescence of chlorophyll (Figure 6), showing a typical fluorescence pattern of cytosol localization. A similar pattern was observed for the Pp4CL10-GFP fusion protein and the GFP control (Figure 6). The results verified that the Pp4CL1 and Pp4CL10 proteins were localized to the cytosol of plant cells. Furthermore, Arabidopsis protoplasts were co-transformed with Pp4CL1 and PEX7 (Arabidopsis). PEX7 is known to localize in peroxisomes (Singh et al., 2009). There was no overlap between the GFP and the red fluorescent of PEX7, providing evidence that the Pp4CL1 protein is not localized in peroxisomes (Supplementary Figures S4A–D).

We used Populus tomentosa 4CL (PDB ID: 3NI2) as a template to build a meaningful three-dimensional model of Pp4CL1-APP complex (Figure 7A). Key amino acids in the hydroxycinnamate binding pocket were Tyr239, Ala243, Gly308, Ala309, Gly332, Gly334, Thr336, Lys437, Lys441, Gln446 (Figures 7B,C). Some amino acids, such as Lys437 and Gln446, form hydrogen bonds with the substrate (Figure 7D). Based on homology modeling and multiple sequence alignments, some mutants, such as Y239A, Y239F, Y239W, A243S, M306K, M306A, G308A, A309G, G334A, K441A, Q446A, and K526A, were generated. We tested the enzymatic activities of mutants toward four substrates (p-coumaric, ferulic, caffeic, and isoferulic acids). The results of HPLC analysis are shown in Figure 8 and the data are listed in Supplementary Table S4. Mutations of Y239W and G334A almost abolished Pp4CL1 activity, probably because Trp and Ala amino acids decrease the affinity of Pp4CL1 for its substrates. According to the sequence alignment results, the three essential residues of Pp4CL1, namely Ala-243, Met-306 and Ala-309, were different from other 4CLs. The current research focused on mutations of M306A, M306K, A309G and A243S. The results showed that the M306A mutation completely abolished Pp4CL1 activities toward p-coumaric and ferulic acids and diminished its activities toward caffeic and isoferulic acids (Figure 8). However, the mutation of M306K only decreased Pp4CL1 activities. As part of the binding pocket, the amino acid Ala-309 was mutated into Gly and the mutation increased Pp4CL1 activity toward p-coumaric acid, but diminished its activities toward the other three substrates. Ser-240 in Pt4CL was highly conserved in 4CL genes. However, the corresponding amino acid in Pp4CL1 was Ala. The mutation of A243S completely abolished Pp4CL1 activity toward ferulic acid. In addition to the above mutations, there were also other essential mutations in this study and most mutations reduced activities of Pp4CL1 toward substrates to different degrees (Figure 8).