The triterpene profiles in the leaves of four different lettuce cultivars (romaine lettuce, iceberg lettuce, red leaf lettuce, and green leaf lettuce) were determined by gas chromatography–mass spectrometry (GC/MS). Among the 12 triterpenes, six compounds were detected as triterpene esters (α-amyrin acetate, β-amyrin acetate, lupeol acetate, ψ-taraxasterol acetate, taraxasterol acetate, and germanicol acetate) as major constituents in the leaves of all four cultivars (Figures 1A,B). The other four compounds were found to be free triterpenes (β-amyrin, α-amyrin, lupeol, and taraxasterol). All six triterpene esters were esterified by acetic acid, as identified by a comparison with standard triterpene acetate compounds (Figure 1). The peak heights and areas of the triterpene acetates were generally higher than those of the free triterpenes (Figures 1A,B), which indicated that triterpene acetates are more highly accumulated than free triterpenes (Figure 1C). The occurrence of various triterpene acetates in lettuce may be due to the presence of triterpene acetyltransferases that convert free triterpenes into triterpene acetates (Figure 2).

We hypothesized that triterpene acetyltransferases in lettuce may be evolutionarily related to sterol O-acyltransferase in plants because sterol and triterpenes are derived from a common biosynthetic precursor (2,3-oxidosqualene). Interestingly, many acyl-CoA:sterol acyltransferase-like sequences in lettuce were registered in the NCBI GenBank database. Initially, the 10 sequences with similarity to acyl-CoA:sterol acyltransferases (AtASAT1 and SlASAT1) were retrieved from the lettuce (crisphead-type lettuce cultivar, Salinas) NCBI database (BioProject: PRJNA432228). A phylogenetic analysis of the deduced amino acid sequences revealed that three sequences (XM_023872407.1, XM_023887114.1, and XM_023885793.1) were grouped into acyl-CoA:sterol acyltransferase (AtASAT1 and SlASAT1) (Figure 3), but the other eight sequences were positioned in subgroups independent of the sequences of AtASAT1 and SlASAT1 (Figure 3).

Methyl jasmonate acts as an effective elicitor of secondary metabolite production in plants (Cheong and Choi, 2003). MeJA treatment can trigger the biosynthesis of a variety of defense chemicals including triterpenes and saponins (Lee et al., 2004; Misra et al., 2014; Singh and Dwivedi, 2018). Seedlings of lettuce were treated with 100 μM MeJA. GC/MS analysis revealed that all triterpene acetates (α-amyrin acetate, β-amyrin acetate, lupeol acetate, ψ-taraxasterol acetate, taraxasterol acetate, and germanicol acetate) were enhanced by treatment with 100 μM MeJA for 3 days (Figure 4A). A qantitative polymerase chain reaction (qPCR) analysis of 10 isolated lettuce sequences in the control and MeJA-treated lettuce seedlings revealed that one sequence (XM023879520.1) was most highly expressed among the other eight genes and strongly responded to MeJA treatment (Figure 4B). Thus, we selected XM023879520.1 as the best candidate sequence to investigate the functional properties of its enzyme.

Using primer pairs for the XM_023879520.1 sequence, a new open reading frame (ORF) sequence was obtained by sequencing the polymerase chain reaction (PCR) products from cDNAs of a green leaf-type lettuce cultivar (L. sativa var. crispa cv. Cheongchima). The ORF region of the new sequence showed a 2-bp difference from the original XM_023879520.1 sequence from Salinas lettuce, which is registered as the LsTAT1 in GenBank (MZ268019). The full-length cDNA clone of LsTAT1 is 1,062 bp, encodes a protein with 353 amino acids, and has a predicted molecular mass of 40.85 kDa. The deduced amino acid sequence of LsTAT1 was found to exhibit 43% identity with that of AtASAT1 (Supplementary Figure 1).

To examine the enzyme activity of LsTAT1 in triterpene acetate production, microsomal fractions from a yeast strain (INVSc1) expressing the LsTAT1 gene were incubated with various triterpenes using acetyl-CoA as an acyl donor. GC/MS analysis of a control consisting of an assay with the microsomal fractions of yeast cells with an empty vector showed two peaks that were derived from intrinsic sterol compounds of yeast microsomal protein extracts (Figure 5A). GC/MS analysis of extracts from triterpene and enzyme mixtures clearly revealed that the LsTAT1 enzyme is able to convert taraxasterol, ψ-taraxasterol, lupeol, taraxerol, β-amyrin, and α-amyrin into taraxasterol acetate, ψ-taraxasterol acetate, lupeol acetate, taraxerol acetate, β-amyrin acetate, and α-amyrin acetate, respectively (Figures 5B–F). The production of lupeol acetate, taraxerol acetate, β-amyrin acetate, and α-amyrin acetate with tiny peaks was confirmed with the single ion mode (SIM) (Figures 5C–F) and by comparing the mass fraction of the compounds with those of authentic standards (Supplementary Figure 3). All the produced triterpene acetates in the enzymatic reactions were matched with standard triterpene acetates (Figure 5G), all of which have a single acetyl group at the C-3 portion of triterpenes, and this finding indicated that the LsTAT1 enzyme can catalyze the acetylation of the C−3 hydroxyl group in the triterpene backbone. The enzyme exhibited higher activity for taraxasterol- and ψ-taraxasterol acetate formation than other triterpene acetates, and more than 40% of taraxasterol- and ψ-taraxasterol triterpenes were converted into taraxasterol acetate and ψ-taraxasterol acetate, respectively (Figure 6). However, the conversion rate of other triterpenes (lupeol, taraxerol, β-amyrin, and α-amyrin) into their acetate forms did not exceed 10% (Figure 6). To examine the acylation activity of LsTAT1 using fatty acyl-CoA, the enzyme was incubated with palmitoyl-CoA (16:0) and stearoyl-CoA (18:0) as acyl donors and lupeol as an acyl accepter. However, LsTAT1 did not show triterpene acylation activity using fatty acyl-CoA (Supplementary Figure 4).

We previously identified an LsOSC1 enzyme (GenBank accession number, MN107542) in lettuce as a multifunctional oxidosqualene cyclase that produces multiple triterpenes (mainly taraxasterol and ψ-taraxasterol together with minor amounts of β-amyrin, α-amyrin, and dammarenediol-II) (Choi et al., 2020). The LsOSC1 and LsTAT1 genes were cloned into the pESC-URA yeast expression vector with double gene expression cassettes. The total ion chromatogram (TIC) obtained by GC/MS revealed that the expression of LsOSC1 alone in yeast produced four new triterpene peaks at retention times between 32 and 38 min, which were the same retention times as those of the β-amyrin, α-amyrin, ψ-taraxasterol, and taraxasterol standards (Figure 7B). The co-expression of LsOSC1 and LsTAT1 in yeast revealed the production of four additional novel triterpene acetate peaks at retention times between 32 and 38 min, which had the same retention times as ψ-taraxasterol acetate, taraxasterol acetate, β-amyrin acetate, and α-amyrin acetate (Figure 7C). In the control yeast with only an empty vector, no triterpene peak with the exception of yeast products such as ergosterols was detected (Figure 7A). Electron Ionization-mass spectrometry (EI-MS) showed m/z 207 [M-H]⁺ and m/z 249 [M-H]⁺ ion fragments characteristic of taraxasterol and taraxasterol acetate, respectively (Figures 7E–G; Choi et al., 2020). The mass spectrum of taraxasterol acetate (Figure 7G) produced in yeast expressing LsOSC1 and LsTAT1 matched that of standard taraxasterol acetate (Figure 7H). To investigate the possible activity of the LsTAT1 enzyme on the production of ergosterol acetate, the TIC obtained in the SIM for the molecular ion (m/z 438.7) of ergosterol acetate was analyzed together with the TIC of the standard compound (ergosterol acetate). However, ergosterol acetate was not detected in yeast expressing LsOSC1 and LsTAT1 (Figure 7D).

Transgenic tobacco plants co-overexpressing LsOSC1 and LsTAT1 under the control of the CaMV35 promoter were constructed. The transgenic tobacco lines were confirmed by genomic PCR of the BAR (phosphinothricin N-acetyltransferase), LsOSC1, and LsTAT1 genes (Supplementary Figure 5). No signal was detected with wild-type tobacco (Supplementary Figure 5).

The triterpene content in the leaf extracts of transgenic tobacco was analyzed by GC/MS. Transgenic tobacco overexpressing both LsOSC1 and LsTAT1 produced four free triterpenes (α-amyrin, β-amyrin, ψ-taraxasterol, and taraxasterol) and four triterpene esters (α-amyrin acetate, β-amyrin acetate, ψ-taraxasterol acetate, and taraxasterol acetate) (Figure 8C). Transgenic tobacco overexpressing LsOSC1 alone produced four triterpene peaks at retention times between 32 and 38 min (Figure 8B), which were the same retention times as those of the β-amyrin, α-amyrin, ψ-taraxasterol, and taraxasterol standards (Figure 8D). With wild-type tobacco, only phytosterols (campesterol, stigmasterol, and β-sitosterol) were detected (Figure 8A). To review the possible esterification of endogenous free phytosterols (campesterol, stigmasterol, and β-sitosterol) into sterol acetates by the LsTAT1 enzyme, we analyzed the presence of peaks of campesterol acetate, stigmasterol acetate, and β-sitosterol acetate in the GC chromatograms through comparisons with chromatograms of authentic standards of sterol esters. However, no peaks for campesterol acetate, stigmasterol acetate, and β-sitosterol acetate were detected (Figure 8C).

Two functionally characterized acyl-CoA:sterol O-acyltransferases, Arabidopsis AtASAT1 (At3g51970, THAA3) and tomato SlASAT (MG865283.1), and several acyl-CoA-sterol O-acyltransferase-like genes in lettuce (Figure 3) were selected from the NCBI nucleotide sequence database of lettuce registered by Reyes-Chin-Wo et al. (2017). Initially, 10 putative acyl-CoA sterol O-acyltransferase-like genes in lettuce (crisphead-type lettuce cultivar, Salinas, CA, United States) were retrieved from the database. We also isolated one mRNA sequence (XM_0238479520.1) from the lettuce sequence database. Using primer pairs for the XM_023879520.1 sequence, the ORF sequence was obtained by sequencing the PCR products from cDNAs of a green leaf-type lettuce cultivar (L. sativa var. crispa cv. Cheongchima). The ORF region of the new LsTST1 sequence showed a 2-bp difference from that of the original XM_023879520.1 sequence from Salinas lettuce, which is registered as the LsTAT1 gene in GenBank (MZ268019).

Multiple sequence alignments were performed using the CLUSTAL W program (Thompson et al., 1997). The subcellular localization of the selected genes was further predicted using DeepLoc-1.0 (Almagro Armenteros et al., 2017). The predictions of the transmembrane helices and protein topology of the enzyme was analyzed with Protter (Omasits et al., 2014).

The deduced amino acid sequences of 10 ASAT-like genes of lettuce and related genes in other plants obtained from NCBI GenBank were subjected to a phylogenetic analysis.

The analysis was conducted using the neighbor-joining method with MEGA 6.0 software (Tamura et al., 2013). A bootstrap analysis with 1,000 replicates was performed to assess the strength of the nodes in the tree (Felsenstein, 1985).

Two-week-old lettuce seedlings were transferred to liquid medium with 0 and 100 μM MeJA for 3 days. Total mRNA from the lettuce seedlings was isolated and then converted to cDNA using an ImProm-II Reverse Transcription System (Promega, Madison, WI, United States). qPCR was performed using a real-time PCR system with a SYBR Green PCR kit (Qiagen, Germany). The primers for 10 isolated sequences and β-actin are shown in Supplementary Table 1. The expression of target genes was calculated using the ^–ΔΔ C_T method (Livak and Schmittgen, 2001). The lettuce β-actin gene was used for normalization. The data are presented as the means ± SEs from at least three independent experiments.

Total RNA from the leaves of lettuce (L. sativa var. crispa cv. Cheongchima) was extracted using an RNeasy plant mini kit (Qiagen, Germany) following the manufacturer’s instructions. cDNA was obtained from the extracted mRNA using an ImProm-2 Reverse Transcription System (Promega Co., Madison, WI, United States). To construct an expression plasmid vector for yeast, the ORF from LsTAT1 amplified from cDNA was cloned into the pYES2.1/V5-His-TOPO vector (Invitrogen, Thermo Fisher Scientific, Waltham, MA, United States) and transformed into Escherichia coli. The primer pairs used to isolate the LsTAT1 cDNAs are shown in Supplementary Table 1. The ORF was then ligated to the GAL1 promoter in the sense orientation. After confirmation of the nucleotide sequence of the inserted DNA, LsTAT1 and an empty vector were expressed in the Saccharomyces cerevisiae strain INVSc1 (Invitrogen). INVSc1 yeast cells were transformed using a modified lithium acetate procedure, as described previously (Gietz et al., 1992). Transformed cells were selected by SC-T (SC minimal medium lacking tryptophan) and subcultured on YPG medium (Kribii et al., 1997).

Microsomal proteins were extracted from yeast after galactose induction for 1 day. The yeast culture was centrifuged (2,000 × g, 10 min), and the pellet was resuspended in 50 ml of TEK (100 mM KCl in 50 mM Tris–HCl with 1 mM EDTA) and centrifuged at 6,100 × g and 4°C for 5 min. The yeast pellets were ground by agitation with glass beads in 5 mM EDTA, 5 mM Tris–HCl, and 0.6 M sorbitol, pH 7.5. The resultant homogenate was spun at 10,000 × g for 15 min, and the supernatant was transferred to a new tube and centrifuged by ultracentrifugation at 100,000 × g and 4°C for 60 min. The pellet was resuspended in 0.1 M potassium phosphate buffer (pH 7.4), 20% glycerol, and 20 mM β-mercaptoethanol for enzyme assay. The protein concentration was measured using the Bradford protein assay.

The enzymatic activity of LsTAT1 was tested in a total volume of 400 μl of 0.1 M potassium phosphate buffer (pH 7.4), 1 mM dithiothreitol (DTT), 5 mM MgCl₂, 200 μM substrates (α- and β-amyrin, taraxasterol, ψ-taraxasterol, taraxerol, and lupeol), 100 μM acyl donors (acetyl-CoA, palmitoyl-CoA, or stearoyl-CoA), and 200 μg of total microsomal proteins. A reaction without exogenous acetyl-CoA was employed as a control. The reaction mixture was incubated for 1 h at 30°C, and the reaction was extracted twice with the same volume of chloroform. The activity of LsTAT1 was determined by the formation of triterpene acetates from triterpenes per milligram of protein per minute. The reaction mixture was mixed with 100% chloroform by vortexing. After centrifugation, the chloroform layer was transferred to a new tube and filtered using a SepPak C-18 cartridge (Waters, Milford, MA, United States). The in vitro activity of the LsTAT1 enzyme on the conversion of triterpenes to triterpene acetates was monitored by GC/MS analysis.

The ORF of LsOSC1 (GenBank accession number, MN107542.1) was amplified using PCR primers, as shown in Supplementary Table 1. The PCR products were cloned into a pYES2.1/V5-HIS-TOPO expression vector (Invitrogen, United States) under the control of the GAL1 promoter. E. coli were transformed with the recombinant vectors. Gene insertion into the vectors was confirmed by nucleotide sequencing of plasmids isolated from each colony. The isolated plasmid was inserted into yeast strain INVSc1 (MATa his3Δ1 leu2 trp1-289 ura3-52/MATα his3Δ1 leu2 trp1-289 ura3-52), which was obtained from Invitrogen (Carlsbad, CA, United States). The transformed yeasts were grown in 50 ml of synthetic complete medium without uracil (SC-U) containing 2% glucose. After 2 days, cells were collected and resuspended in SC-U medium with 2% galactose and then inducted at 30°C for 16 h. Thereafter, the cells were collected, refluxed with 80% MeOH and sonicated for 30 min at 40°C. After centrifugation, the supernatant was transferred to chloroform and vortexed. The chloroform layer was subsequently filtered using a SepPak C-18 cartridge (Waters Co., Milford, MA, United States) and analyzed by GC/MS.

The coding region of LsTAT1 (GenBank accession No. GU183405) was amplified using the primer pairs mentioned above, subcloned into the pGEM-T Easy vector and sequenced. The ORF fragments were excised through NotI and inserted into the NotI site of the pYES3/CT vector by ligation to construct an expression vector (promoter: GAL1, selection marker: TRP1) (Invitrogen, Thermo Fisher Scientific, Waltham, MA, United States). The vector was introduced into yeast. Gene insertion into the vectors was confirmed by nucleotide sequencing of plasmids isolated from each colony. For cotransformation of the LsTAT1 and LsOSC1 genes in yeast, plasmids containing LsTAT1 were transformed into INVSc1 yeast expressing LsOSC1, and the cells were cultured on medium without uracil and tryptophan. The culture conditions and methods for galactose induction and preparation of the triterpene fraction were the same as those described above. To analyse the production of triterpenes in transgenic yeasts, transgenic yeast cells harvested after centrifugation were mixed with 80% methanol and sonicated for 10 min. After centrifugation, the supernatant was mixed with 100% chloroform by vortexing. The chloroform layer was transferred to a new tube and filtered using a SepPak C-18 cartridge (Waters).

The ORF regions of LsTAT1 (GenBank accession number, MZ268019) and LsOSC1 (GenBank accession number, MN107542.1) were amplified using PCR primers, as shown in Supplementary Table 1. Both genes were cloned into the pCR8/GW/TOPO (Invitrogen, Carlsbad, CA, United States) vector and transferred into the destination vector pPZIP-Bar under the control of the double 35S CaMV promoter. Eventually, the overexpression construct harboring both the LsTAT1 and LsOSC1 genes was introduced into Agrobacterium tumefaciens GV3101 competent cells by the heat shock method. The construction of transgenic tobacco with the LsTAT1 and LsOSC1 genes was performed by A. tumefaciens-mediated transformation.

Seeds of four cultivars (romaine lettuce, iceberg lettuce, red leaf lettuce, and green leaf lettuce) of lettuce were purchased from Asia Seed Co. (Seoul, South Korea). At the two- to three-true-leaf stages, germinated plants were transplanted into plastic pots containing peat moss and perlite (3:1 v/v). The plants were grown in a greenhouse with a 16-h photoperiod and day and night temperatures of 24 and 18°C, respectively. When the plants had grown to the 10- to 12-leaf stage, the third leaves from the top were sampled for triterpene analysis.

To analyse the triterpene profiles in leaves of four lettuce cultivars, MeJA treated lettuce seedlings, and leaves of transgenic tobacco with the LsTAT1 and LsOSC1 genes, the leaves were air dried at 50°C in a drying oven. The milled powders (200 mg) from each of the samples were soaked in 100% methanol (1 ml) and sonicated for 30 min at a constant frequency of 20 kHz and a temperature of 40°C. The supernatant was collected by centrifugation and subsequently filtered using a SepPak C-18 cartridge (Waters Co., Milford, MA, United States).

For the GC/MS analysis prepared from the extracts mentioned above, an aliquot (5 μl) with a split injection (5:1) was obtained for analysis using a gas chromatograph (Agilent 7890A) equipped with an inert MSD system (Agilent 5975C) with a triple-axis detector and an HP-5 MS capillary column (30 m × 0.25 mm, film thickness of 0.25 mm). The inlet temperature was set to 250°C, and the column temperature was programmed to start at 150°C for 5 min, increase to 300°C at a rate of 5°C/min, and remain at 300°C for 20 min. The GC/MS conditions were set as follows: ionizing energy, 70 eV; ion source temperature, 230°C; and MS quad temperature, 150°C. The chromatogram peaks in the GC analysis were identified by comparison with a library database and authentic triterpene and triterpene acetate standards. Triterpenes and triterpene acetates were identified by comparing their retention times with those of authentic standards and mass fragmentation spectra.

The in vitro activity of the LsTAT1 enzyme on the conversion of lupeol to lupeol palmitate by the reaction of palmitoyl-CoA (16:0) as an acyl donor and lupeol as an acyl accepter was monitored by an ACQUITY ultra-performance liquid chromatograph (UPLC) (Waters Co., Milford, MA, United States) equipped with a Photodiode Array (PDA) detector.

A binary gradient method was used to vary the mobile phase composition over time. Mobile phase solvents A and B consisted of water and acetonitrile, respectively. Linear gradient elution at a flow rate of 0.5 ml/min was performed for analysis: initial 15% B from 0 to 0.5 min, linear gradient 15–45% B from 0.5 to 8 min, linear gradient 45–100% B from 8 to 10 min, and a final return to 15% B, which was maintained until 2 min. The total run time was 18 min, and the sample injection volume was 2 μl.