Arabidopsis CAPRICE (MYB) and GLABRA3 (bHLH) Control Tomato (Solanum lycopersicum) Anthocyanin Biosynthesis

In Arabidopsis thaliana the MYB transcription factor CAPRICE (CPC) and the bHLH transcription factor GLABRA3 (GL3) are central regulators of root-hair differentiation and trichome initiation. By transforming the orthologous tomato genes SlTRY (CPC) and SlGL3 (GL3) into Arabidopsis, we demonstrated that these genes influence epidermal cell differentiation in Arabidopsis, suggesting that tomato and Arabidopsis partially use similar transcription factors for epidermal cell differentiation. CPC and GL3 are also known to be involved in anthocyanin biosynthesis. After transformation into tomato, 35S::CPC inhibited anthocyanin accumulation, whereas GL3::GL3 enhanced anthocyanin accumulation. Real-time reverse transcription PCR analyses showed that the expression of anthocyanin biosynthetic genes including Phe-ammonia lyase (PAL), the flavonoid pathway genes chalcone synthase (CHS), dihydroflavonol reductase (DFR), and anthocyanidin synthase (ANS) were repressed in 35S::CPC tomato. In contrast, the expression levels of PAL, CHS, DFR, and ANS were significantly higher in GL3::GL3 tomato compared with control plants. These results suggest that CPC and GL3 also influence anthocyanin pigment synthesis in tomato.


Introduction
Anthocyanins are important chemical compound of polyphenolic pigments derived from the phenylpropanoid biosynthetic pathway. Anthocyanins belong to the group of flavonoids, of which they are noticeable in the wide range of chemical structures [1]. Anthocyanins provide appealing color to leaves, flowers, fruits and seeds in plants. In addition to this obvious feature, they have other essential functions. Anthocyanin synthesis was induced by the stressful occasions, such as low temperature or strong irradiation of the sunlight, against which they protect the plant as scavengers for radical species or a light-screen [2]. Anthocyanins are produced through several enzymatic step [3]. The enzymes which are involved in anthocyanin synthesis are fully analyzed by both biochemical and genetic approaches.
In the previous study, we identified Arabidopsis CPC and GL3 homologous genes from tomato and named them Solanum lycopersicum TRYPTICHON (SlTRY) and Solanum lycopersicum GLABRA3 (SlGL3), respectively [33]. The SlTRY-encoded protein was most closely related to TRY among the CPC-like MYBs [33]. Transformants expressing the tomato TRY homologous gene (SlTRY) in Arabidopsis had a greater number of roothairs and no trichomes, a phenotype similar to that seen in overexpressors of CPC-like MYB genes. On the other hand, transformants expressing the tomato GL3 homologous gene (SlGL3) in Arabidopsis had no obvious GL3-like phenotypes related to non-hair and trichome cell differentiation [33]. We concluded that tomato and Arabidopsis use similar transcription factors for root-hair and trichome cell differentiation and that the SlTRY-like R3 MYB may be a key common regulator of plant root-hair and trichome development [33]. In prior work, we also analyzed the anthocyanin content of SlTRY and SlGL3 transgenic Arabidopsis [34]. We showed that anthocyanin accumulation was repressed in the CPC::SlTRY and GL3::SlGL3 transgenic Arabidopsis plants, suggesting that the tomato genes of SlTRY and SlGL3 are involved in anthocyanin biosynthesis [34].
In this study, we have expressed the Arabidopsis CPC and GL3 genes in tomato to show the effect of these genes on tomato anthocyanin biosynthesis, indicating that GL3 is a positive regulator for anthocyanin biosynthesis, but CPC is a negative regulator.

Plant materials and growth conditions
Tomato, Solanum lycopersicum L. cv. Micro-Tom, was used. Seeds were surface-sterilized with 10% commercial bleach including a detergent (Kitchen Haiter, Kao, Tokyo, Japan), for 20 min and then rinsed with sterilized water three times for 5 min each and sown on 1.5% agar plates containing 0.56 MS medium [35]. Seeded plates were held at 4uC for 2 d and then incubated at 25uC under constant white light (50-100 mmol m 22 s 21 ) for 7 days to produce seedlings for RNA extraction. Some 7-day-old seedlings were transplanted into soil and grown in a photoperiod of 16 h light (50-100 mmol m 22 s 21 ) at 25uC for an additional week to produce mature plant tissues for anthocyanin extraction.

Transgenic plants
Gene constructs of 35S::CPC [13] and GL3::GL3 [36] were introduced into tomato (Micro-Tom) according to a highly efficient transformation protocol for Micro-Tom [37]. Agrobacterium tumefaciens C58C1 was grown for 24 h at 28uC. Cotyledon explants were sectioned, dipped in the bacterial suspension to allow adsorption, and transferred to callus induction medium containing 100 mg L 21 kanamycin, 1.5 mg L 21 zeatin and 375 mg L 21 Augmentin (GlaxoSmithKline, Uxbridge, UK) [37]. Transgenic shoots were selected and rooted on a medium containing 50 mg/L kanamycin.
Homozygous transgenic lines were selected based on kanamycin resistance. We obtained ten and four T2 transgenic tomato lines and selected eight and three homozygous lines of 35S::CPC and GL3::GL3, respectively. The presence of 35S::CPC and GL3::GL3 in the transgenic plants was confirmed by PCR using CPC or GL3 forward and reverse primers (Table 1) ( Figure S1). Only those plants with the expected PCR products (CPC and GL3) were used in the analyses.

Real-time reverse transcription PCR analysis
The sequences of all primers used in this study are listed in Table 1. Total RNA from tomato tissues was extracted with MagDEA RNA 100 (GC) (PSS, Chiba, Japan) using a Magtration System 12 GC (PSS, Chiba, Japan). To remove contaminating genomic DNA, RNA samples were treated with RNase-free DNase I (Ambion, Austin, TX, USA) according to the Magtraction System protocol. Plant tissue (100 mg) was homogenized using a TissueLyser II (Qiagen, Valencia, CA, USA) with 100 mL of RLT buffer (Qiagen, Valencia, CA, USA). Sample supernatants were applied to the instrument, and RNA was eluted with 50 mL of sterile distilled water.
First-strand cDNA was synthesized from 1 mg total RNA in a 20 mL reaction mixture using the Prime Script RT Master Mix (Perfect Real Time) (Takara, Tokyo, Japan). Real-time PCR was performed using a Chromo4 Real-Time IQ5 PCR Detection System (Bio-Rad, Hercules, CA, USA) with SYBR Premix Ex Taq II (Takara, Tokyo, Japan). PCR amplification employed a 30 s denaturing step at 95uC, followed by 5 s at 95uC and 30 s at 60uC with 40 cycles for CPC, GL3, PAL, CHS, DFR, ANS and LeActin. Real-time PCR was used to analyze the mRNA expression level of each transcript encoding CPC, GL3, PAL, CHS, DFR and ANS in transgenic tomato. The relative expression of each transcript was calculated by the DDCT method [38]. The expression levels of CPC, GL3, PAL, CHS, DFR and ANS were estimated after being normalized to the endogenous control gene LeActin (TC116322) [39]. The primers were: CPC-F and CPC-R for CPC; GL3-F and GL3-R for GL3; PAL-F and PAL-R for PAL; CHS-F and CHS-R for CHS; DFR-F and DFR-R for DFR; ANS-F and ANS-R for ANS; LeActin-F and LeActin-R for LeActin [39][40][41].

Extraction and analysis of anthocyanins
Anthocyanin levels were measured according to previously reported protocols [42,43]. Control and transgenic plants were grown together in a growth chamber as described above. Anthocyanins were extracted from cotyledons of 7-day-old seedlings, leaves and stems of three-week-old plants, and fresh weights were determined. Total plant pigments were extracted overnight in 0.3 mL acidic methanol (1% (v/v) HCl). After the addition of 0.2 mL water and an equal volume of chloroform, anthocyanins were separated from the chlorophylls by partitioning into the aqueous methanol phase, and the absorption was measured at 530-657 nm in a spectrometer (GENios, TECAN). Anthocyanin levels were then normalized to the total fresh weight of tissue used in each sample.

Light microscopy
To observe anthocyanin pigment localization in hypocotyls of the control, 35S::CPC and GL3::GL3 transgenic plants, we prepared hand-cut sections from 3-week-old plants and observed them by light microscopy using a Zeiss (Jena, Germany) Axio Imager. Z1 microscope.

Results
Anthocyanin pigmentation of the 35S::CPC and GL3::GL3 transgenic plants To establish whether Arabidopsis CPC and GL3 transcription factors function in tomato, we introduced these genes into one of tomato cultivars (Solanum lycopersicum L. cv. Micro-Tom). Previously, we showed that 35S::CPC transgenic Arabidopsis plants have an unusually large number of root-hairs and no leaf trichomes [13]. Thus, we chose to introduce the 35S::CPC construct into tomato in this experiment. In contrast, the root-hair number of 35S::GL3 transgenic Arabidopsis plants is not significantly different from the wild-type [25], suggesting that the 35S promoter is not suitable for GL3 gene overexpression. The expression of GL3 should be precisely controlled by the GL3 promoter [31]. Therefore, we decided to use the GL3::GL3 construct (a genome fragment of GL3 driven by the GL3 promoter) for transformation of tomato in this study [36].
The 35S::CPC and GL3::GL3 transgenic tomato plants were phenotypically similar to the control plants ( Figure 1; Figure S1). We did not detect any remarkable differences between 35S::CPC or GL3::GL3 transgenic tomato plants and the control tomato plant in root-hair and trichome phenotypes (Figure 1; Figure S1). On the other hand, we observed qualitatively less and more reddish-purple coloration in the stems and leaves of 35S::CPC and GL3::GL3 plants, respectively (Figure 2A, 2D and 2G). The first true leaves of two-week-old 35S::CPC transgenic plants had clearly lower amounts of anthocyanin pigmentation on the adaxial and abaxial sides of the leaves compared with that of the control plants ( Figure 2B   reddish-purple anthocyanin in the leaf veins compared with the control plants ( Figure 2H, 2I).
To determine the tissue distribution of anthocyanin pigments in the 35S::CPC and GL3::GL3 transgenic tomato plants, we examined hand-cut sections prepared from stem samples with a light microscope as shown in Figure 2A, 2D and 2G. In hypocotyls of two-week-old control tomato seedlings, anthocyanin pigments were observed in a few cells, as was previously reported in tomato hypocotyls ( Figure 3A) [44]. Anthocyanins did not accumulate in the hypocotyls of young 35S::CPC tomato seedlings ( Figure 3B). In the hypocotyls of GL3::GL3 seedlings, anthocyanin pigments were present in two to three layers of an epidermal cell and subepidermal cells ( Figure 3C). These results suggest that CPC expression did not induce any remarkable changes in root-hair and trichome formation but reduced anthocyanin accumulation in transgenic tomato. GL3 also did not affect the epidermal phenotype but induced anthocyanin accumulation in transgenic tomato.

Analysis of anthocyanin levels in the cotyledons, leaves and stems of transgenic plants
We examined the effects of CPC and GL3 on anthocyanin accumulation in the different tissues. Expression levels of the introduced CPC gene were checked by PCR, and we selected three lines (35S::CPC#10, 35S::CPC #18 and 35S::CPC #21) among eight transgenic lines for analysis ( Figure S2A). Expression levels of the introduced GL3 gene were also checked by PCR.
Among three GL3::GL3 transgenic lines, only one line, GL3::GL3#12, showed stable expression of GL3. Therefore, we used the GL3::GL3#12 line for further analyses ( Figure S2B). To compare the levels of anthocyanin accumulation in 35S::CPC and GL3::GL3 with those in control tomato, the anthocyanin content in extracts of two-week-old seedlings was determined ( Figure 4). Compared with the control tomato cotyledons, all three lines of 35S::CPC transgenic tomato cotyledons had significantly reduced levels of anthocyanin ( Figure 4A). On the other hand, cotyledons of GL3::GL3 accumulated higher levels of anthocyanin compared with that of the control plants ( Figure 4A). Consistent with the observations shown in Figure 2, very low levels of anthocyanin accumulation were observed in leaves of all three 35S::CPC lines ( Figure 4B). Compared with control tomato leaves, significantly larger amounts of anthocyanin were measured in GL3::GL3 leaves ( Figure 4B). Consistent with the observations shown in Figure 2 and 3, anthocyanin accumulation was also significantly reduced in the stems of all three 35S::CPC lines and increased in GL3::GL3 stems compared with those in the control plants ( Figure 4C). We confirmed that introduction of the CPC gene under the control of the 35S promoter significantly inhibited anthocyanin accumulation in cotyledons, leaves and stems of tomato as observed in Arabidopsis [12]. Introduction of the GL3 gene under the control of the GL3 promoter significantly increased anthocyanin accumulation also in mature leaves and stems of tomato as observed in Arabidopsis [45]. Effect of CPC and GL3 on the expression of anthocyanin pathway genes To characterize more fully the involvement of the introduced CPC and GL3 transcription factors on the regulation of anthocyanin biosynthesis in tomato, we examined the expression levels of genes that encode anthocyanin biosynthetic enzymes. The effects of CPC and GL3 on the expression of anthocyanin biosynthesis genes were examined by real-time RT-PCR, as described in the Materials and Methods section. First and second true-leaf samples of representative 35S::CPC, GL3::GL3 and control plants, harvested from two-week-old seedlings, were homogenized, and total RNA was isolated from each tissue sample. Anthocyanins are synthesized through the flavonoid biosynthetic pathway [46]. Therefore, expression levels of tomato genes for Phe-ammonia lyase (PAL), the flavonoid pathway genes chalcone synthase (CHS), dihydroflavonol reductase (DFR), and anthocyanidin synthase (ANS) were determined and expressed relative to the LeActin gene, a tomato gene that encodes an actin protein [39]. Consistent with the reduced anthocyanin accumulation in 35S::CPC transgenic tomato ( Figure 4B), PAL, CHS, DFR and ANS expression levels were significantly lower in  35S::CPC transgenic tomato compared with the control plants ( Figure 5). In contrast, consistent with the large amount of anthocyanin accumulation in GL3::GL3 transgene tomato ( Figure 4B), PAL, CHS, DFR and ANS expression levels were significantly higher in GL3::GL3 transgenic tomato compared with control plants ( Figure 5). These results suggest that Arabidopsis CPC and GL3 can regulate gene expression of the anthocyanin biosynthetic pathway in tomato and affect the anthocyanin accumulation.

Discussion
In this study, we introduced the Arabidopsis CPC and GL3 genes into tomato under the control of the 35S promoter and the GL3 promoter, respectively. Overexpression of CPC is known to induce root-hair cell differentiation and inhibits trichome forma-tion in Arabidopsis [13]. Overexpression of GL3 is known to reduce root-hair cell differentiation and induce trichome formation in Arabidopsis [25,31]. However, overexpression of CPC and GL3 in tomato did not result in visible differences in the root-hair and trichome phenotypes (Figure 1; Figure S1). The reasons for the differences in CPC and GL3 function between tomato and Arabidopsis may arise from fundamental differences in the way epidermal organs develop in the two plants. Root epidermal development in vascular plants is classified into three types [47]. Tomato root epidermal development belongs to type 1, in which root-hairs can be produced from any root epidermal cell [48]. Conversely, Arabidopsis root epidermal development belongs to type 3 in which root-hair cell files and non-hair cell files are organized in the root epidermis [47]. Regulation of root-hair cell and non-hair cell fate determination by the TTG1-GL3/EGL3- WER complex and CPC might be specific for Arabidopsis but not for tomato.
Trichome phenotypes are also different between Arabidopsis and tomato. Arabidopsis trichomes are normally large single cells with three branches [49], whereas tomato trichomes are chemically and morphologically divergent [50][51][52]. Tomato trichomes are classified into seven types, including glandular (types I, IV, VI and VII), and non-glandular (types II, III and V) trichomes [51,53]. The participation of many regulatory genes might be necessary to form tomato trichomes. Hence, it is likely difficult to change tomato trichome phenotypes by CPC or GL3 overexpression only. Tomato might need other transcriptional factors to change the morphology of the epidermal cell.
In a previous study, we isolated SlTRY and SlGL3 from tomato as orthologous genes of the Arabidopsis CPC and GL3, respectively [33]. The full length SlTRY protein shares 50% amino acid identity with CPC [33]. Phylogenic analysis suggested that SlTRY and CPC originated from a single common ancestor [33]. SlTRY was shown to function quite similarly to the Arabidopsis CPC, including in the formation of ectopic roothairs, in the induction of a no-trichome phenotype and in its action as a repressor of anthocyanin accumulation in Arabidopsis [34]. In summary, SlTRY functions in a similar way as CPC for the epidermal cell differentiation and the anthocyanin accumulation in Arabidopsis. On the other hand, there was no obvious effect on trichome or non-hair cell differentiation in the Arabidopsis GL3::SlGL3 transformants [33]. Rather, anthocyanin accumulation was reduced in the GL3::SlGL3 transgenic Arabidopsis compared with the wild-type [34]. In contrast, GL3 functions as a positive regulator for the anthocyanin accumulation in Arabidopsis [7]. The difference of the sequence between GL3 and SlGL3 might contribute to the opposite functions although they share 45% amino acid identity at the entire region [33]. Taken together, the functions of SlGL3 are completely different from those of GL3.
In this study, we demonstrated that Arabidopsis CPC and GL3 genes regulate anthocyanin biosynthesis in tomato. We made 35S::CPC transgenic tomatoes that accumulated significantly less anthocyanin in comparison with the control plants ( Figure 4). In contrast, anthocyanin accumulation in GL3::GL3 transgenic tomato was greater than the control plants ( Figure 4). CPC and GL3 are known to regulate anthocyanin biosynthesis in Arabidopsis [12,54]. Our study suggests that the regulatory system for anthocyanin biosynthesis by CPC and GL3 is maintained in both Arabidopsis and tomato.
Genes encoding enzymes of the anthocyanin biosynthetic pathway are divided into two groups: early biosynthetic genes including PAL and CHS, and late biosynthetic genes including DFR and ANS. The two groups have independent activation mechanisms in dicotyledonous species [55,56]. Whereas PAL and CHS are involved in the synthesis of precursors and flavonoids, DFR and ANS are more specific for the synthesis of anthocyanins. Analysis of the biosynthetic pathway genes in tomato showed that genes of both groups were regulated by CPC and GL3. Expression levels of PAL, CHS DFR and ANS were significantly lower in 35S::CPC transgenic tomato compared with the control plants ( Figure 5). In contrast, expression levels of PAL, CHS DFR and ANS were significantly higher in GL3::GL3 transgenic tomato compared with the control plants ( Figure 5). GL3 and CPC were strong up-and down-regulators of the entire anthocyanin biosynthesis pathway in tomato, respectively ( Figure 5), which reflect the results form Arabidopsis [7,12]. These results suggest the presence of a TTG1-TT8/GL3-PAP1/2 like protein complex that specifically regulates anthocyanin biosynthesis in tomato [45,[57][58][59].
Many studies contributed to the elucidation of the anthocyanin biosynthetic pathway using Arabidopsis [10,[60][61][62][63][64]. As a result, the molecular genetics of the regulatory system for anthocyanin biosynthesis has greatly progressed [1,46,[65][66][67]. In Arabidopsis, the regulatory protein complex, which includes WD40, bHLH and MYB transcription factors, regulates anthocyanin biosynthesis [10,58,68,69]. WD40 is encoded by TTG1, bHLHs are encoded by TT8, GL3 and EGL3, and MYBs are encoded by PAP1, PAP2, MYB113 and MYB114 [65]. In addition to the WD40-bHLH-MYB complex, CPC, a single repeat R3-MYB, is a negative regulator of anthocyanin biosynthesis in Arabidopsis [12]. MYBL2, another R3-MYB gene, functions as a negative regulator of anthocyanin biosynthesis in Arabidopsis seedlings [70,71]. Our study suggests the existence of a WD40-bHLH-MYB complex that regulates anthocyanin biosynthesis in tomato. CPC may disrupt this putative WD40-bHLH-MYB protein complex, thus inhibiting the activity of downstream anthocyanin biosynthetic genes in tomato. In Arabidopsis, there are a total of seven CPC family R3type MYB genes, including CPC, TRY, ETC1 ETC2, ETC3/ CPL3, TCL1 and TCL2/CPL4 [14][15][16][17][18][19][20][21][22]. In contrast, only SlTRY was identified as a putative tomato ortholog of CPC so far [33]. Although the total number of tomato CPC orthologous gene(s) is still unknown, fewer genes are expected than are present in the Arabidopsis genome. The small number of R3-type MYB gene(s) in tomato might reflect their specific functions in anthocyanin biosynthesis. Because SlGL3 did not induce anthocyanin accumulation in Arabidopsis [34], SlGL3 probably does not participate in the putative WD40-bHLH-MYB protein complex in tomato as is the case in Arabidopsis. A model for regulating anthocyanin biosynthesis in tomato by WD40-bHLH-MYB will be forthcoming with further analyses.