In higher plants, the regulation of anthocyanin biosynthesis by various transcription factors [9–14, 32, 35]. The GLK transcription factors were originally identified in maize, and were subsequently found in Arabidopsis, maize, rice, tomato, and the moss Physcomitrella patens [26, 27, 36–39]. GLK transcription factors belong to the GARP transcription activator family, and the protein sequences are highly conserved among different species, with Myb-like DNA-binding domain and the C-terminal box [26, 39, 40]. In Arabidopsis, AtGLK genes exist as a pair of homologous genes, AtGLK1 and AtGLK2. The previous studies found that AtGLKs mainly regulate the chloroplast development in higher plants [26–28]. In recent years, more and more studies have shown that AtGLKs play important roles not only in responding to biotic and abiotic stresses, but also in regulating leaf senescence [41–45]. The current study found that AtGLKs have an important function in regulating the accumulation of anthocyanins in Arabidopsis.
Anthocyanins are water-soluble, vacuolar pigments in plants that belong to the family of flavonoid compounds [46]. Since sucrose is a strong inducer of flavonoid biosynthesis and is known to induce anthocyanin accumulation in a variety of plant species [16–18], we analyzed the expression patterns of AtGLK1 in response to exogenous sucrose treatment. Real-time quantitative PCR analyses revealed that the mRNA accumulation of AtGLK1 was significantly promoted by sucrose (Fig. 1a). The increased AtGLK1 transcript level in response to sucrose appeared to originate from its promoter activities since it was observed that exogenous sucrose treatments significantly increased GUS expression in the cotyledons and hypocotyl of AtGLK1::GUS transgenic seedlings (Fig. 1b). Such an expression pattern suggested that AtGLK1 may be involved in sucrose-induced anthocyanin accumulation during the early stages of Arabidopsis development.
Through the phenotypic, physiological, and molecular analyses conducted in this work, strong positive correlations were identified between AtGLK1 expression and anthocyanin accumulation to sucrose treatment. First, the loss-of-function glk1 glk2 double mutant was found to have lower anthocyanin levels than the glk2 single mutant, although loss of AtGLK1 alone had not affected the anthocyanin accumulation (Fig. 2). The absence of an anthocyanin-less phenotype for the glk1 mutant may have been due to functional redundancy or compensation between the AtGLK1 and AtGLK2. Similarly, the AtGLK1 and AtGLK2 have been shown to be functionally redundant in the regulation of chloroplast development [26, 27]. During the early developmental stage of Arabidopsis seedlings, single glk mutants (glk1 and glk2) largely resemble wild-type, only the glk1 glk2 double mutant showed a chloroplast-defective phenotype, suggesting that the each of two AtGLK genes acts redundantly to direct monomorphic chloroplast development [26]. The AtGLK genes were found to exhibit partial redundancy since there was an anthocyanin-less phenotype specific to the glk2 mutant allele, but no phenotype specific to the glk1 allele (Fig. 2). The following two aspects of the experimental data may have reflected the fact that the two genes had different expression levels rather than different functions. On the one hand, overexpression of AtGLK1 significantly enhanced anthocyanin accumulation in the 35S::AtGLK1 transgenic Arabidopsis seedlings, even though the expression of AtGLK2 was dramatically impaired (Fig. 3). On the other hand, real-time quantitative PCR results showed that the mRNA accumulation of AtGLK1 was significantly lower than that of AtGLK2 in the wild-type Arabidpsis seedlings (Fig. S2). Second, when overexpressed in Arabidopsis, the 35S::AtGLK1 transgenic seedlings displayed enhanced anthocyanin accumulation (Fig. 3). We also detected the expression of AtGLK2 in the wild-type and 35S::AtGLK1 transgenic seedlings. It was interesting to find that the expression of AtGLK2 was significantly impaired in the AtGLK1-overexpressing plants when compared with the corresponding wild-type plants (Fig. 3b). There were two possible explanations. The first explanation was that the AtGLK1 has an additional function of regulating AtGLK2 expression. The second explanation is that the decreased transcription of the AtGLK2 in the AtGLK1-overexpressing plants were most likely for the purpose of maintaining a constant total mRNA amount of AtGLKs via expressional reprogramming between the two homologous genes. Third, We found that glk mutants (glk1, glk2 and glk1 glk2) seedlings had accumulated lower transcript levels of DFR, F3'H, LDOX, UF3GT, UGT75C1, and UGT75C2, which are known to be involved in the late step of anthocyanin biosynthesis, while the AtGLK1-overexpressing seedlings showed higher transcript levels than those observed in the wild-type seedlings (Fig. 4). In contrast, the transcript levels of the early biosynthesis genes, such as CHS and CHI, were not observed to be greatly altered in the AtGLK1-overexpressing plants (Fig. 4c). Another potential target of AtGLK1 action could be PAP1, which has been shown to trigger the activation of expression of late anthocyanin biosynthesis genes [18, 47]. PAP1 is an R2R3 MYB-type transcription factor that is capable of mediating ectopic activation of an array of genes involved in anthocyanin biosynthesis in several plant species, including Arabidopsis, tobacco, petunia and rose [47–50]. Indeed, our study found that the transcript level of PAP1 was lower in the glk mutants (glk1, glk2, and glk1 glk2) seedlings, but significantly higher in AtGLK1-overexpressing seedlings, when compared with the corresponding wild-type plants (Fig. 4). It therefore appeared that the AtGLK1 regulates sucrose-induced anthocyanin accumulation mainly through influencing the expression of late anthocyanin biosynthesis genes. Therefore, based on the results mentioned above, our study considered that AtGLK1 is potentially a positive regulator of anthocyanin accumulation in Arabidopsis.
The intracellular signaling from the chloroplast to the nucleus is referred to as plastid retrograde signaling. These signaling processes play essential roles in coordinating the expression of nuclear and plastid-encoded genes [51]. In the present study, it was found that norflurazon and lincomycin (two drugs known to block chloroplast biogenesis via different mechanisms), which induce retrograde signaling [33, 34], were found to enhance the anthocyanin accumulation of sucrose-treated Arabidopsis seedlings (Fig. 5; Fig. S1). These findings suggested that the anthocyanin biosynthesis is positively regulated by plastid retrograde signaling. If the positive signals from dysfunctional chloroplasts are transmitted exclusively via AtGLK1, then these signals should be abrogated in glk1 glk2 double mutants. However, the effects of norflurazon and lincomycin on the sucrose-induced anthocyanin accumulation were observed to be greater in the glk1 glk2 double mutants, but lower in AtGLK1-overexpressing seedlings, when compared with wild-type seedlings (Fig. 5c-d). These observations suggested the possibility that AtGLK1 acts as a negative regulator in plastid retrograde signal-mediated anthocyanin accumulation. Consistent with this speculation, the results of the real-time quantitative PCR analysis showed that the AtGLK1 had been strongly down-regulated by the norflurazon and lincomycin treatments at the transcription level (Fig. 5a). Despite this, further studies will be needed in order to unravel the detailed molecular mechanisms of AtGLK1-mediated plastid retrograde signaling pathways which regulate anthocyanin accumulation.
MYBL2 is a negative regulator of anthocyanin biosynthesis. The analyses of the expression patterns of the mybl2 mutant, or transgenic plants overexpressing MYBL2, have demonstrated that MYBL2 regulates the expression of anthocyanin biosynthesis-related genes [13, 14]. Similar expression patterns were observed in the structural and regulatory genes in the anthocyanin biosynthetic pathways in the AtGLK1-overexpressing plants and the glk1 glk2 double mutant in this study (Fig. 4), which raised the possibility that AtGLK1 regulates anthocyanin biosynthesis by modulating MYBL2 expression. However, the MYBL2 transcript levels showed no obvious changes in either the glk1 glk2 double mutant or AtGLK1-overexpressing plants when compared with the wild-type (data not shown). Therefore, it was hypothesized that AtGLK1 may regulate MYBL2 expression at the post-transcriptional level. To determine whether AtGLK1 and MYBL2 act in the same genetic pathway to regulate anthocyanin accumulation in Arabidopsis, we generated transgenic lines overexpressing MYBL2 in AtGLK1-overexpressing plants. The results indicated that the overexpression of MYBL2 completely complemented the anthocyanin overaccumulation phenotype in the AtGLK1-overexpressing seedlings (Fig. 6b-c), which suggested that MYBL2 is epistatic to AtGLK1 in anthocyanin biosynthesis. Also, consistency was found in the transcript levels of the anthocyanin biosynthetic (DFR, F3'H, LDOX, UF3GT, UGT75C1, and UGT75C2) and regulatory (PAP1 and TT8) genes, which were up-regulated in the AtGLK1-overexpressing seedlings, and all down-regulated when MYBL2 was overexpressed (Fig. 6d).