Genetic effects of Red Lettuce Leaf genes on red coloration in leaf lettuce under artificial lighting conditions

Abstract Some cultivars of lettuce accumulate anthocyanins, which act as functional food ingredients. Leaf lettuce has been known to be erratic in exhibiting red color when grown under artificial light, and there is a need for cultivars that more stably exhibit red color in artificial light cultivation. In this study, we aimed to dissect the genetic architecture for red coloring in various leaf lettuce cultivars grown under artificial light. We investigated the genotype of Red Lettuce Leaf (RLL) genes in 133 leaf lettuce strains, some of which were obtained from publicly available resequencing data. By studying the allelic combination of RLL genes, we further analyzed the contribution of these genes to producing red coloring in leaf lettuce. From the quantification of phenolic compounds and corresponding transcriptome data, we revealed that gene expression level‐dependent regulation of RLL1 (bHLH) and RLL2 (MYB) is the underlying mechanism conferring high anthocyanin accumulation in red leaf lettuce under artificial light cultivation. Our data suggest that different combinations of RLL genotypes cause quantitative differences in anthocyanin accumulation among cultivars, and some genotype combinations are more effective at producing red coloration even under artificial lighting.

such as apples, strawberries, turnips, cabbage, and lettuce. Especially lettuce is a major crop grown in artificial light-type plant factories. Artificial light-type plant factories can grow vegetables regardless of abnormal weather conditions. In addition, the growing environment can be controlled for stable and high anthocyanin synthesis. This makes it possible to ensure food safety and a stable supply of functional vegetables.
Anthocyanins are secondary metabolites synthesized via the phenylpropanoid biosynthesis pathway using phenylalanine as precursor (Liu et al., 2018;Petroni & Tonelli, 2011;Saito et al., 2013). (GST) and multidrug resistance-associated protein (MRP) genes have been identified as transporters that transport anthocyanins into vacuoles (Goodman et al., 2004;Sun et al., 2012). The regulatory genes for the anthocyanin biosynthesis have also been identified. Transcriptional activation of biosynthetic genes is triggered by the MBW complex, comprising R2R3-type MYB transcription factor, bHLH transcription factor, and WD40 factor. In most cases, the expression of MYB and bHLH regulatory genes are specific for pigmented tissues, while that of WD40 genes, which are involved in stabilizing the MBW complex, is similar in both anthocyanin-pigmented and non-pigmented tissues (Koes et al., 2005;Ramsay & Glover, 2005).
Four genes involved in leaf coloration of lettuce have been cloned, namely Red Lettuce Leaf (RLL) 1 to 4 (Su et al., 2020). These four genes are genetically polymorphic, and have strong, weak, or no effects on anthocyanin accumulation, depending on the genotype ( Figure S1; Type A and B, respectively). Among these genes, RLL1, encoding a bHLH transcription factor, and RLL2, encoding a MYB transcription factor, have been identified as quantitative trait loci (QTLs) responsible for leaf coloration in a red lettuce cultivar (Su et al., 2020). This implies that the regulatory network of transcription factors and other regulators for anthocyanin biosynthesis is conserved in lettuce. RLL3 is a homolog of MYBL2 in Arabidopsis. RLL4 is a homolog of Arabidopsis RUP1, encoding a negative regulator of UV-B signaling. A genomewide association study (GWAS) using large-scale RNA sequencing (RNA-seq) data sampled from the major cultivars and related species identified six candidate loci, including RLL1, RLL2 and ANS, as expression QTLs (eQTLs) responsible for natural variations in anthocyanin content in lettuce leaves during domestication (Zhang et al., 2017).
Thus, so far, these studies have revealed that natural genetic variations are important for leaf coloration in lettuce under field conditions and discussed how the natural variations controlling leaf color were selected during domestication. However, the identity of those variations that determine anthocyanin accumulation in red lettuce cultivars remains to be elucidated.
Light is not only an essential energy source for photosynthesis but also plays a critical role as an environmental signal in plant development and physiology, such as leaf expansion, stem elongation, and metabolism (Folta, 2004;Kitazaki et al., 2018;McNellis & Deng, 1995). Anthocyanin production in plants can be increased by changing the quality and quantity of light in growth environments (Zoratti et al., 2014). In lettuce, genes whose expression level varies depending on light quality have been reported from omics analysis, and ANS is among them (Kitazaki et al., 2018). The ANS gene is the target of RLL1 and RLL2, while RLL3 and RLL4 are involved in the regulation of the expression of ANS (Su et al., 2020). However, little is known about the differences among lettuce cultivars in the genetic effects on the efficiency of anthocyanin production under artificial light. The advantage of artificial light-type plant factories is that light quality can be controlled according to the purpose. The knowledge of these five genes is expected to be used for environmental control in plant factories. We focused on these five genes as important factors causing quantitative differences in anthocyanin accumulation.
To elucidate the genetic effects of RLL genes on red coloration under artificial lighting, we evaluated leaf lettuce cultivars for the core set of RLL genotypes. While determining the RLL genotype, we found a nonsense mutation in the ANS gene in silico by means of TASUKE (https://tasuke.dna.affrc.go.jp), a comparative genome analysis tool (Kumagai et al., 2013) that we applied to publicly available lettuce next-generation sequencing (NGS) data and new NGS data produced in this study. From the results of RLL genotyping, we selected nine representative cultivars carrying different combinations of RLL genotypes for further quantification of phenolic compounds and transcriptome analysis. Quantitative analysis of phenolic compounds showed a genotype-dependent accumulation of anthocyanin under a given artificial environment. Integrated analysis of anthocyanin levels and RNA-seq data sampled from nine cultivars with different genotype combinations revealed that, in addition to anthocyanin biosynthetic genes, transcript levels of a group of genes, including RLL1 and RLL2, were highly correlated with anthocyanin levels. This result indicated that the genetic network regulating MBW activity plays a critical role in anthocyanin accumulation in lettuce under artificial lighting. Our results supply a genetic basis for conferring red coloration in lettuce and not only offer a strategy for producing new lettuce cultivars containing high levels of phenolic compounds, such as anthocyanin, but also provide valuable information for lettuce-breeding programs using genome-sequencing data.

| RLL genotyping
To genotype the RLL locus, a polymerase chain reaction (PCR)-based assay was performed using genomic DNA as a template with the KAPA2G Fast ReadyMix PCR kit (Kapa Biosystems) in accordance with the manufacturer's instructions. For RLL1, we used a primer that identifies a 5-bp deletion; for RLL3 and RLL4, we used dCAPs primers with MboI and HincII restriction enzyme sites, respectively, that identify a 1-bp substitution (Su et al., 2020). For RLL2, we designed a primer to identify the 15-bp deletion, and for ANS, we used a dCAPs primer designed to create a MboI site at the 1-bp substitution GAG→TAG (stop), which we found by aligning the genomesequencing data of 88 cultivars. The PCR products and their digests with restriction enzymes were separated by agarose gel electrophoresis and fragment sizes were checked except for RLL1. The genotype of RLL1 was determined by whether or not the target sequence was amplified by PCR.

| Genome sequencing using NGS and analysis with TASUKE
Young leaf tissues were powdered with liquid N 2 and total genomic

| Analysis of phenolic compounds
The relative content of phenolic compounds in each extract was determined by a method described by Arapitsas et al. (2008) with minor modifications. Powdered lettuce samples (0.05-0.1 g) were extracted with 250-500 μl (proportional to the frozen weight) of methanol-water-formic acid solution (40:59:1, v/v/v) for 1 h after 3min sonication, followed by two extractions with 250 μl 80% methanol. The combined extract was centrifuged at 13,000 g for 5 min. The supernatant was filtered through a 0.45μm centrifugal filter (Merck-Millipore) and the filtrate made up to 1 ml. Ten μl of the resultant extract was injected into the HPLC system that comprised an Agilent Technologies 1100 Series HPLC (Agilent) equipped with a YMC-Triart C18 column (150 mm × 2.0 mm I.D., S-3 μm, 12 nm; YMC Co. Ltd.). The flavonoids and anthocyanins in the extract were separated with a mobile phase (0.2 ml min −1 ) consisting of (A) 5% (v/v) formic acid aqueous solution and (B) acetonitrile, according to a multistep solvent gradient program (Table S1)

| Quantitative real-time PCR
Frozen samples were powdered with liquid N 2 and RNA was extracted with the RNeasy Plant Mini kit (Qiagen) in accordance with the manufacturer's instructions. RNA (500 ng) was subjected to cDNA synthesis using the ReverTra Ace cDNA synthesis kit with genome remover (TOYOBO). qRT-PCR was performed with the qPCR MasterMix SYBR reagent (TOYOBO) using primers listed in Table S2 and detected by ViiA7 (Thermo Fisher, Applied Biosystems). The cycle program for the PCR was according to the manufacturer's instructions.

| Plasmid construction and agroinfiltration
Full-length cDNAs of RLL1, RLL2 and LsTTG1 were amplified by PrimeSTAR GXL DNA polymerase according to the manufacturer's instructions (TAKARA). Amplified DNA fragments were verified by sequencing. Each full-length cDNA was then amplified by individual primers with the linker sequence (Table S2) and cloned into the binary vector, pMLH7133 (Mochizuki et al., 2003), digested with BamHI and SacI by means of the infusion-HD system (TAKARA).
Expression of each gene was regulated by the 7-fold replicated enhancer of the cauliflower mosaic virus 35S promoter (E7) with the tobacco mosaic virus omega sequence insertion and the first intron of a gene for phaseolin.
The plasmids harboring each cDNA were transformed into Agrobacterium strain GV3101 by electroporation. Resulting recombinant Agrobacteria were separately cultured overnight in liquid media containing 50 mg l −1 kanamycin. Cultured cells were resuspended in agroinfiltration buffer (10 mM MES-KOH, pH 5.7, 10 mM MgCl 2 ) and adjusted to high and low concentration based on absorbance (OD 595 = 0.9 and 0.4, respectively). For suppressing the interfering activity of endogenous RNA, p19 (Lakatos et al., 2004) was co-infiltrated at OD 595 = 0.1. After incubation for 3 h at room temperature, Agrobacterium solutions were infiltrated into leaves of cv. 'Fancy Green' by needleless 1-ml syringe. Leaf samples were collected at 24 h after infiltration.

| RNA sequencing
Total RNA was extracted from frozen and ground leaves using the

| Data processing
Raw reads were filtered to remove adapters and low quality reads with Trimmomatic v0.39 (Bolger et al., 2014) and the quality of the data was checked with FastQC_v0.11.9 (Andrews, 2010). The clean reads were mapped to the L. sativa cv. 'Salinas' genome (Reyes-Chin-Wo et al., 2017) using HISAT2 v2.2.1 (Kim et al., 2015). Gene expression level was calculated based on the transcripts per million (TPM) method in StringTie v2.1.5 (Pertea et al., 2015). In RLL2, there is a tandem duplication of the R2R3-type MYB gene in a red leaf lettuce cultivar (RLL2A, RLL2B; Figure S1b, Type A) (Su et al., 2020). Only one gene (RLL2A) was actively expressed in red leaf cultivars ( Figure S2). To evaluate RLL2 genotypes, we used a structure variant (15-bp in/del) in the last exon of both red and green cultivars of the RLL2 genes ( Figure S1). As shown in Figure 1, the red leaf lettuce cultivars contained the functional type (type A) of RLL1, RLL2, and ANS alleles, indicating that these alleles are important for leaf lettuce to turn red under field conditions. Additional variations producing a specific amino acid substitution in RLL3 and RLL4 genes in red leaf lettuce cultivars were likely to influence quantitative differences in their red coloration among these cultivars, because functional versions of these genes are not necessary for red coloration in leaves. In contrast, green leaf lettuce cultivars harboring functional forms of only ANS or RLL1 and RLL2 were frequently observed. Our results indicate that the FNPs of RLL genes identified in the previous studies (Su et al., 2020) can generally be used for the selection of leaf coloration and suggest that combinations of these alleles might unintentionally have occurred during breeding performed in natural field conditions.

| Confirmation of the RLL genotype of lettuce from around the world using NGS data
To check the distribution of the known RLL genotype in other lettuce accessions, we used genome data collected from diverse lettuce accessions obtained around the world. The sequence variations detected in NGS data were visualized by a comparative genome analysis tool, TASUKE, that enables the user to effectively visualize variations in nucleotide substitutions and in/dels in collections of individually annotated genes (Kumagai et al., 2013). In addition to using NGS data for 88 accessions that are publicly available, we implemented TASUKE with newly added NGS data for nine cultivars used in this study and nine accessions available from the NARO Genebank (https:// www.gene.affrc.go.jp/index_en.php). Although the reference L. sativa cv. 'Salinas' genome had a nonsense mutation in the ANS gene (Lsat_v5_gn_9_97280; Figure 2a), lettuce accessions with red coloration had no mutation in the exon, producing an intact polypeptide of ANS with high similarity to the Arabidopsis ANS protein (Figure 2b).
In silico genotyping for four RLL genes further supported the notion that three functional genotypes, for RLL1, RLL2, and ANS, constituted a core set for red coloration ( Figure S3). However, by expanding the investigation of the RLL genotype into lettuce accessions from around the world, we found an exception to the association between the RLL genotype and leaf coloration in the cultivar, 'Flashy Trout Back'.
Because 'Flashy Trout Back' has been described as having green leaves with red speckles, anthocyanin biosynthesis is potentially active and might be regulated spatially in this cultivar. In addition, we found a putative FNP of GST (Lsat_1_v5_gn_3_87760) that has been identified as a candidate for one e-QTL (Zhang et al., 2017, Figure S1). The highly conserved amino acid, Pro 43 , in the N-terminal region was converted into Ser 43 . The N-terminal region of GST is required for glutathione binding (Dixon et al., 2002).

| Profiles of phenolic compounds in leaf lettuce cultivars with different combination of the RLL genotype grown under artificial light
Green leaf lettuce has been grown commercially in controlled environments to maximize yield in artificial light-type plant factories.
Additional light treatment, such as blue or UV light, can enhance the red color of red leaf lettuce (Ebisawa et al., 2009;Goto et al., 2016;Shoji et al., 2010), implying that stable production of red leaf lettuce with high anthocyanin contents might be difficult in closed-type plant factories. Our genotyping results suggested that differences in the RLL functionality might cause quantitative variation in anthocyanin accumulation among leaf lettuce cultivars. Therefore, we quantified phenolic compounds, including anthocyanin, in different leaf lettuce cultivars grown under fluorescent light. We selected five red leaf lettuce cultivars and four green leaf lettuce cultivars according to the combination of RLL genotype (group I to VII, shown in Figure 1). In our growth conditions, the degree of red coloring in leaves of the selected five red leaf cultivars seemed to be con-  Figure S4). Other flavonoids, namely naringenin, kaempferol, and leucocyanidin and their glycosides, were not detected, possibly because the content was below detectable levels. A similar tendency to that found for anthocyanin accumulation in the five red cultivars and four green cultivars was generally observed for contents of F I G U R E 1 RLL genotype in 36 leaf lettuce cultivars sold in Japan. Genotyping was performed using PCR with primers designed for discriminating the functional nucleotide substitution in individual genes. Gray and white indicate genotypes with strong (A) and weak or no (B) effects on anthocyanin accumulation, respectively. Detailed information about the nucleotide polymorphisms is shown in Figure S1. The genotypes were grouped from I to VII, according to the combination of genotypes, and one or two cultivars from each group were selected for subsequent experiments.

| Expression of RLL genes in leaf lettuce cultivars
The results of quantification of phenolic compounds in the nine cultivars indicated that, in a given environmental condition, cumu-

| Expression profiles of genes associated with anthocyanin accumulation in leaves
As described above, our results from qRT-PCR showed that tran-   (TPM ≥1) in all the red leaf lettuce cultivars. Among those genes, we chose genes whose expression was significantly more than twice as high in five red leaf lettuce cultivars as in four green leaf lettuce cultivars: 187 genes were extracted by the Wald test of DESeq2 as statistically significant (p < 0.05, Figure 6a,b). Because RLL1 and RLL2 expression was highly correlated with anthocyanin levels (Figure 4), we performed co-expression analysis between the expression of those 187 genes and cyanidin-3-O-(6′′-O-malonyl)-glucoside levels using all biological replicate samples and identified 47 genes with a high Pearson's correlation coefficient (r ≥ 0.75) as anthocyanin biosynthesis highly associated genes (ABHAG; Figure S5). The selected 47 genes were then allocated into five clusters using the k-means method. As we expected, the 47 genes included RLL1 and ANS.
Genes in clusters 1, 2, and 3 had a broad range of phenotype from high to low anthocyanin content and the detected absolute values of gene expression were higher than those of clusters 4 and 5. Clusters homolog, whose activities are associated with biosynthesis of phenolic compounds (Borevitz et al., 2000;Ma et al., 2021;Mahmood et al., 2016;Wang et al., 2016). These genes were thought to act as enhancers for biosynthesis and/or accumulation of phenolic compounds in lettuce. Clusters 4 and 5 also contained some abiotic stress response genes, such as peroxidase, UVR3, and LRR family protein (homolog of cold-regulated gene AT3G19320; Vergnolle et al., 2005). In addition, genes encoding TT8, Ser/Thr kinase, and cryptochrome, which have been previously identified as genes coexpressed with flavonoids in Arabidopsis (Yonekura-Sakakibara et al., 2008), were among the genes in clusters 4 and 5.
We mapped the expression profiles of phenylpropanoid biosynthetic genes onto the biosynthetic pathway ( Figure 6c). Our results confirmed that the expression profiles of the series of anthocyanin biosynthetic genes were well coordinated with the accumulation of anthocyanin across different cultivars.
Step of flavonol biosynthesis was also co-regulated with anthocyanin biosynthesis. In contrast, genes involved in the chlorogenic acid and chicoric acid biosynthetic pathways appeared to be expressed independently of anthocyanin biosynthesis. Our results indicate that the ABHAG might be used as biomarkers to precisely monitor a status of production of not only anthocyanin but also other phenolic compounds in a certain cultivar or in a particular artificial environment.

| DISCUSS ION
With the increasing threat of unpredictable or extreme weather conditions, closed-type plant factories with artificial light are commercially attractive as they can stably cultivate high quality vegetables with desirable nutritional value and functional components. Our evaluation of levels of phenolic compounds in leaf lettuce cultivars grown in a controlled environment with fluorescent light showed that accumulation of anthocyanins was dependent on the combination of known RLL genotypes that control red coloration. In addition, transcriptome analysis indicated that the accumulation of anthocyanin coincided with changing transcriptional patterns related to anthocyanin biosynthesis and transport, and abiotic stress responses. By combining the known genetic framework for anthocyanin biosynthesis in plants with targeted metabolomic and transcriptome data, we revealed that a dosagedependent regulation of RLL1 (bHLH) and RLL2 (MYB), which encode core components of the MBW complex, is the underlying mechanism for enhancing accumulation of anthocyanins in leaf lettuce grown under artificial light.

| Potential relationship between anthocyanin accumulation and light intensity or quality
The red color observed in artificial light conditions is much lighter than the red color seen in the open field. The light red color is likely caused by artificial light in growth chamber. The reason for this is probably that light quality and light intensity are different between open field and artificial light condition. As for light intensity, homolog genes encoding ELIP1 (AT3G22840) and ubiquinone methyltransferase (AT2G41040) in ABHAG were reported to respond to high-intensity light in Arabidopsis (Kleine et al., 2007).
They showed that the induction of ELIP1/2 expression is mediated via CRY1 in a blue light intensity-dependent manner. In addition, the ubiquinone methyltransferase misregulated in response to HL in the cry1 mutant. Under our experimental conditions, lettuce CRY1 homologs (Lsat_1_v5_gn_1_3881 and Lsat_1_v5_ gn_9_41700) were expressed but not included in ABHAG because the Log 2 fold change of the red variety compared to the green variety was less than 1.
Regarding light quality, exposure to additional blue light and UV light can enhance accumulation of anthocyanin in a red leaf cultivar (Ebisawa et al., 2009;Goto et al., 2016;Ma et al., 2021;Shoji et al., 2010). Our study revealed that the potential for anthocyanin biosynthesis in individual cultivars might be determined by the presence of particular genotypes. Comparison between genotype and contents of phenolic compounds indicated that a functional difference in the RLL4 gene was important under our experimental conditions (Figures 1 and 3). Presence of the reduced function-type RLL4 allele was associated with remarkable accumulation of phenolic compounds in the cultivars RL and RO. Because

| Enhancement of accumulation of anthocyanin and flavonoid compounds in leaf lettuce
Our results confirmed that the evolutionally conserved MBW complex acts as a central regulator for anthocyanin accumulation in leaf lettuce ( Figure 1) and indicated that the regulation of the MBW complex activity is a major determinant for transcriptional levels of ABHAG, including anthocyanin biosynthesis and transporter genes, hence determining final anthocyanin levels (Figures 3-5). Our results showed that the combination of RLL3 and RLL4 alleles affects the accumulation of anthocyanin.
Functionality of RLL4 might indirectly affect the transcription levels of RLL1 and RLL2. As described above, RLL4 encodes a RUP1 homolog in Arabidopsis (Gruber et al., 2010). RUP1 negatively regulates the UV-dependent expression of HY5, responsible for anthocyanin biosynthesis through transcriptional activation of the downstream transcription factor, MYB75/AtPAP1, a homolog of RLL2 in lettuce (Shin et al., 2013

| Additional transcriptional regulators associated with anthocyanin biosynthesis
Among ABHAG selected in this study, three transcriptional factors in addition to RLL1 and RLL2 were identified ( Figure S5). The first, Lsat_v5_gn_2_124061, encoded a homolog of MYB75/AtPAP1 in Arabidopsis that can activate genes in both early and late steps of anthocyanin biosynthesis through the formation of the MBW complex (Gonzalez et al., 2008;Shi & Xie, 2010 gene, which has the highest identity to CUC3 and is functionally redundant with its homologs CUC1 and CUC2, controls activation of anthocyanin biosynthesis in blood-fleshed peach (Zhou et al., 2015).
Several studies have implicated NAC genes in the regulation of anthocyanin biosynthesis (Ma et al., 2021). Therefore, it is possible that MBW-dependent expression of these transcription factors is required for stable production of phenolic compounds in certain genetic backgrounds of red leaf lettuce.

| Phenolic compounds and their biosynthesis in leaf lettuce
The profiles of phenolic compounds in green and red oak-leaf lettuce have been extensively characterized by ultra-high-performance liquid chromatography coupled online to DAD, electrospray ionization, and QToF-MS systems (Viacava et al., 2017). Consistent with those profiles, cyanidin 3-(6-malonylglucoside), cyanidin 3-(3-malonylglucoside), and cyanidin 3-glucoside were detected only in red leaf lettuce cultivars (Figure 3b-d). Apigenin-glucoside, identified only in the green oak-leaf cultivar (Viacava et al., 2017), was not detected as a major peak in any green leaf cultivars used in this study.
Accumulation of high anthocyanin content can coordinately increase phenolic compounds, including flavonoids. Our quantitative analysis revealed that chlorogenic acid and chicoric acid were Genes selected as anthocyanin biosynthesis highly associated genes are indicated with asterisks (also indicated in Figure S5). Branched pathways for biosynthesis of chlorogenic acid (blue arrows), chicoric acid (green arrows), and quercetin glucosides (orange arrows) are indicated. major phenolic compounds in leaf lettuce cultivars (Figure 3). Both phenolic compounds have long been of interest for human health because of their antioxidant effects and role in suppressing fat accumulation (Lee & Scagel, 2013;Tungmunnithum et al., 2018).

(c)
Among nine cultivars with different RLL genotype combinations, the cultivars RL and RO, both harboring genotype combinations conducive to anthocyanin accumulation, contained not only higher anthocyanin but also higher phenolic compounds than other genotype combinations. Consistent with our data for lettuce, in potato, chlorogenic acid was the most abundant phenolic compound and was present in high levels in high-anthocyanin cultivars (Navarre et al., 2011;Valiñas et al., 2017). Our transcriptome analysis further suggested that full activation of the entire anthocyanin biosynthesis pathway increases chlorogenic acid and chicoric acid contents as by-products (Figure 6c). Therefore, basically, two compounds are constantly synthesized regardless of the RLL genotype (Figure 3e-f). But, in cultivars in which anthocyanin biosynthesis is highly active, such as RL and RO in this study, de novo p-coumaric acid and 4-coumaronyl-CoA production would be expected to be increased, so that the precursor for chlorogenic acid and chicoric acid biosynthesis is more highly available.
Our analyses identified quercetin glucosides as the main type of flavonoid in leaf lettuce. The levels of the major quer- has been identified as a flavonol-specific regulator of phenylpropanoid biosynthesis in Arabidopsis and FLS is a direct target of MYB12 in Arabidopsis (Mehrtens et al., 2005). Coordinated transcriptional regulation of genes involved in branch pathways associated with the anthocyanin biosynthetic pathway would confer a balanced high production of anthocyanin and other phenolic compounds, including flavonoids, in red leaf lettuce cultivars. If the specific MYB transcription factors were solely expressed (Blanco et al., 2018;Liu et al., 2018;Pandey et al., 2015), phenolic compounds might be synthesized independently of anthocyanin biosynthesis. We suggest that, by metabolic engineering for phenolic compounds, which are early by-products of the phenylpropanoid biosynthetic pathway, it might be possible to give green leaf lettuce nutritionally valuable components other than anthocyanin.
Although we have confirmed that evolutionally conserved components play a critical role in anthocyanin accumulation in lettuce, it is also necessary to optimize the quality and quantity of light to maximize the efficiency of anthocyanin synthesis in closed-type plant factories equipped with artificial light. It might be possible to find a light environment that can promote efficient accumulation of anthocyanins by using ABHAG as an indicator. Artificial environments are easily controllable and highly reproducible. In addition, the development of light-emitting diode light sources is progressing steadily. Searching for environments optimized for anthocyanin accumulation in various red lettuce cultivars might help identify novel genetic loci controlling the production of phenolic compounds, including anthocyanin, and will support data-driven strategies for breeding new cultivars appropriate to cultivation in plant factories.