Expression pattern of carotenoid biosynthesis genes during papaya fruit development
The fruit color of papaya is determined by accumulation of different types of carotenoid compositions and content, and the expression level of the carotenoid biosynthesis genes play important roles in papaya flesh color, fruit firmness, and nutritional profile during papaya fruit development. To understand carotenoid biosynthesis genes’ expression pattern during papaya fruit ripening, RNA-Seq analysis was performed to detect carotenoid-related genes’ expression level during five stages of fruit development (Supplement Fig. S2). CpCYC-B, CpCHY-B, and CpLCY-B were significantly upregulated during fruit ripening; CpZDS, CpPDS, and CpLCY-E were also upregulated from stage 3 (50% ripening stage) to stage 5 (100% ripening stage) (Fig. 1).
Identification and characterization of putative transcription factors involved in CpCYC-B and CpLCY-B regulation
To further analyze the potential transcription factors regulating carotenoid biosynthesis genes involved in papaya fruit ripening, RNA-Seq analysis was implemented using libraries constructed from papaya varieties SunUp (from S1 to S5) and AU9 (GRN, CB, and RP) fruit at different developmental stages. To calculate reads per kilo bases per million reads (RPKM), significant expressed tags in different samples were identified. The differentially expressed transcription factors in SunUp and AU9 samples were identified based on the following criteria: P < 0.01, FDR < 0.01, and fold change >2. Expression profiles of different TFs were classified as bHLH, GRAS, SBP, and bZIP between the first stage and the final stage in SunUp and AU9 cultivars. In total, 23 and 27 different expressed TFs were identified in SunUp and AU9, respectively. Most TFs were downregulated in two cultivars. Of the 23 TFs in Sunup, 17 TFs, including bHLHs, C2H2s, and GRFs, were downregulated, and 5 TFs, including bZIP and B3, were upregulated. Of the 27 TFs in AU9, 21 TFs, including BES1, WRKY, and TCP, were downregulated, and 6 TFs, including bZIP and AP2, were upregulated (Fig. 2). Some of transcription factors were selected for further analysis.
a Expression pattern of 27 DETFs from green (GRN), color break (CB) to ripeness in AU9 cultivars; b Expression pattern of 23 DETFs from stage 1 (S1) to stage 6 (S6) in SunUp cultivars. The cage of red to blue represents quite different expression level. The highly expressed transcription factors were selected for further analysis. Expression value presented as log2FPKM
To test whether these transcription factors can bind promoters of CpCYC-B and CpLCY-B during papaya fruit development, we analyzed the promoter sequences in silico and constructed a series of chimeric genes containing truncated fragments of CpCYC-B and CpLCY-B promoters and GUS reporter gene. Promoters were analyzed for potential TFs binding sites using the PlantCare database. The potential regulatory elements in upstream −1.5 kb promoters of CpCYC-B and CpLCY-B have been summarized in Table 1. Most of them are responsive to hormones and light. CGTCA-motif and ERF, bHLH boxs were the consensus cis-elements of CpCYC-B and CpLCY-B promoters. For example, bHLH binding site (-CANNTG-), was found at −113 and −164, respectively, in CpCYC-B and CpLCY-B promoters. The distribution of different regulatory elements in different region of promoters implied that some special elements played special function at different developmental stages.
Reporter gene GUS expression was analyzed to check the activity of these cis-elements in transgenic Arabidopsis. Higher expression level of GUS gene was detected at −1.5 and −0.5 kb segments of CpCYC-B (Fig. 3a). Deletion from −1.5 to −1.0 kb decreased expression compared with the other two deletion promoter segments. The −1.0 kb promoter segment of CpLCY-B showed a higher expression level than −1.5 and −0.5 kb promoters. Transgenic plants with pMDC162 alone did not show any GUS expression level (data not shown). These results indicated that −0.5 kb fragment of CpCYC-B and CpLCY-B promoter, respectively, can activate or inactivate the regulation of GUS gene expression. To further verify GUS gene expression patterns in transgenic lines, Arabidopsis transformed with CpCYC-B/CpLCY-B-GUS fusion gene were analyzed by histochemical staining (Fig. 3c–h). The GUS activity with −0.5 kb region of CpCYC-B and CpLCY-B promoter was, respectively, stronger and weaker than another two types of transgenic plants with −1.0 and −1.5 promoters. These results were identical to the relative expression results of different length promoters (Fig. 3a, b), implying that −0.5 kb promoters of CpCYC-B and CpLCY-B were corresponding to chromoplast- and chloroplast-specific expression patterns, respectively.
a The qRT-PCR expression levels of GUS promoted by different promotor promoter fragments of CpCYC-B were presented. b The qRT-PCR expression levels of GUS by different promotor promoter fragments of CpLCY-B were presented. The abscissa represents the different promoter lengths; the ordinate represents the qRT-PCR expression level. The expression level was shown as the mean of determinations made with seven to thirteen independent plants. c–h Histochemical staining of GUS activities driven by different promoter fragments of CpCYC-B and CpLCY-B were obtained from representative transgenic Arabidopsis leaves. c Pro-1.5k CpCYC-B; d Pro-1.0k CpCYC-B; e Pro-0.5k CpCYC-B; f Pro-1.5k CpLCY-B; g Pro-1.0k CpLCY-B; h Pro-0.5k CpCYC-B. Bars represent the standard errors of means. The scale in the figure (c–h) represents 1 mm
We further deleted the −0.5 kb promoters of CpCYC-B and CpLCY-B into −0.2, −0.3, and −0.4 kb fragments. We discovered that Pro–CpCYC-B::GUS had a relatively higher expression level with −0.2 kb promoter than other forms of transgenic Arabidopsis. However, Pro–CpLCY-B::GUS showed similar relative expression levels driven by different promoters fragments (Supplement Fig. S3). In addition, we mutated two elements, -GAAAGAA-(311 bp) and -ATTTCAAA-(ERF-responsive element) in −0.5 kb of CpLCY-B and CpCYC-B promoters, respectively, and the results indicated that −0.5 kb of Pro-CpLCY-B::GUS without -GAAAGAA-element exhibited a relatively higher expression level compared with those containing this element, suggesting -GAAAGAA-element might act as a suppressor in CpLCY-B promoter. Without ERF element in −0.5 kb Pro-CpCYC-B::GUS T1 transgenic plants, there was no obvious expression difference with other transgenic promoter fragment in Arabidopsis (Supplement Fig. S3A and B). However, when we treated the transgenic Arabidopsis containing −1.0 kb CpCYC-B and −1.0 kb CpLCY-B promoter with 80 mg/L ethephon, GUS activity appeared to be weaker (Supplement Fig. S3C-H). This result implied that ERF element might act as a negative cis-element.
By in silico analysis and experimental validation, we identified that −0.5 kb promoter of CpCYC-B and CpLCY-B played important regulatory roles and included some fruit-specific elements, e.g., ERF-responsive element, bHLH-box, in addition to the novel negative regulatory element (-GAAAGAA-) in CpLCY-B promoter.
Combining the result of the expression of TF family members during different fruit ripening stages and TF family motif in −0.5 kb promoters of CpCYC-B and CpLCY-B, we obtained four types of transcription factor families as shown in Fig. 4a. According to their distribution in promoters of CpCYC-B and CpLCY-B, all of their binding sites existed in −0.5 kb CpCYC-B promoter, while, bHLH and ERF binding sites existed in −0.5 kb CpLCY-B promoter. The characterizations of four TF families were shown in Table 2. Eight TFs belonging to four types of TF families were selected from database (Fig. 2). The eight differentially expressed transcription factors were named as CpbHLH1 (evm.model.supercontig_131.7), CpbHLH2 (evm.model.supercontig_8.253), CpbZIP1 (evm.model.supercontig_81.160), CpbZIP2 (evm.model.supercontig_9.75), CpNAC3 (evm.model.Supercontig_594.1), CpERF1 (evm.model.supercontig_131.83), CpERF2 (evm.model.supercontig_13.250), and CpbZIP3 (evm.model.supercontig_244.3). The transcription of the eight TFs was validated by RNA-Seq and qRT-PCR. The expression levels of CpbHLH1 and CpERF1/2 were decreased from the initial stage. As shown in Fig. 4b, the expression level of CpNAC3 had been decreased from stage 2; the expression patterns of CpbZIP1/2/3 were random, and for example, the expression level of CpbZIP1 first increased, then decreased and finally increased again. For CpbHLH2, expression level first increased at color break stage, decreased from stage 2, and then increased significantly from stage 4. These results were further demonstrated by RNA-Seq data (Supplement Fig. S4).
a Different TF familie binding sites existed in −0.5k CpCYC-B and CpLCY-B promoters; b qRT-PCR expression patterns of eight transcription factors. The abscissa represents different stages of fruit development and the ordinate represents qRT-PCR expression levels. Value was shown as mean + SE of three replications
CpbHLH-1/2 binds to CpCYC-B and CpLCY-B promoters and regulates their expression
In view of the expression pattern of eight TFs, it is reasonable to assume that they could regulate the expression of CpCYC-B and CpLCY-B. To test whether these TFs can bind to the promoters of CpCYC-B/CpLCY-B, yeast one-hybrid (Y1H) experiment was performed. Yeast cells were co-transformed with the pGADT7-CpbHLH1 + CpCYC-B190bp promoter, pGADT7-CpbHLH1 + CpCYC-B190bp promoter mutant, pGADT7-CpbHLH1 + target repeated element, pGADT7-bHLH1 + repeated target mutation element, positive control and negative control would be cultured on basic SD medium lacking Leu and Trp element (Fig. 5c). However, yeast cells co-transformed with the positive control, pGADT7-bHLH1 + CpCYC-B190bp promoter, pGADT7-bHLH1 + target element repetition, could grow on triple dropout minimal medium. These results were similar to yeast cells co-transformed with pGADT7-bHLH2 + CpCYC-B promoter, pGADT7-bHLH2 + CpLCY-B promoter, pGADT7-bHLH1 + CpCYC-B promoter. These results indicate that CpbHLH1/2 can bind to CANNTG motifs in CpCYC-B and CpLCY-B promoters in yeast cells (Fig. 5). However, other six TFs (including CpbZIP1/2/3, CpNAC3, and CpERF1/2) failed to act with promoters of CpCYC-B/CpLCY-B (Supplement Fig. S5). We also preliminarily proved that CpbHLH1/2 could bind to CANNTG motif in CpZDS, CpLCY-E, and CpCHY-B promoters though Y1H but fail to bind the promoter of CpPDS. Thus, a co-expression pattern existed between CpbHLH1/2 and CpZDS, CpLCY-E, CpCHY-B, CpCYC-B, and CpLCY-B (Supplement Fig. S6). Meanwhile, the there was no interaction between CpbHLH1 and CpbHLH2 by yeast two-hybrids (Supplement Fig. S7).
a Schema charts of CpbHLH1/2 acting with promoters of CpCYC-B in different control. b Schema chart of CpbHLH1/2 acting with promoters of CpLCY-B in different control. c Yeast one-hybrid interactions between CpbHLH1/CpbHLH2 and CpCYC-B/CpLCY-B. CpbHLH1/2 interacted with CpCYC-B/CpLCY-B in Leu-/Trp- and Leu-/Trp-/His-medium with different controls. Interactione were indicated by the ability of yeast cells to grow on a synthetic medium lacking tryptophan, leucine, histidine. Yeast cells transformed with pGADT7-53m + p53HIS were used as a positive control, while those transformed with Phis2 + pGADT7-53m, Phis2-Pro-Mutant element in CYC/LCY + AD-CpbHLH1/2, Phis2- repeat mutant element + AD-CpbHLH1/2 as negative controls. p53m:Pgad53m; p53:p53HIS; Pro-CYC/LCY: selected about 190 bp promoters, including bHLH elements of CpCYC-B/CpLCY-B; Pro-mutant element in CYC/LCY: mutant bHLH element in 190 bp promoters of CpCYC-B/CpLCY-B; Repeated target element: five repetitions of bHLH element sequence; Repeated mutant target element: five repetitions of mutant bHLH element sequence
To investigate the effects of CpbHLH1/2 on CpCYC-B and CpLCY-B expression, dual-luciferase transient expression assays were performed. The LUC/REN ratio was lower than the positive control and negative control (CpbHLH1/Empty) when either the CpCYC-B or CpLCY-B pro-LUC reporter construct was co-transfected with the CaMV35S-CpbHLH1 effectors, implying that CpbHLH1 repressed CpCYC-B and CpLCY-B expression activities (Fig. 6c, d). In contrast, the LUC/REN ration of CpbHLH2–CpCYC-B/CpLCY-B was higher than the negative control (CpbHLH1/Empty) when previous promoters-LUC reporter constructs was co-transfected with the CaMV35S-CpbHLH2 effectors, indicating that CpbHLH2 may involve the activated regulation of CpCYC-B and CpLCY-B (Fig. 6e, f). Besides, LUC/REN ratios of four types of experimental groups were found to be similar to the empty vector controls. We have not performed the transient expression analysis on the other carotenogenesis genes, e.g, CpZDS, CpCHY-B, and CpLCY-E, although bHLH binding sites existed in their promoters (Supplement sequence1). Collectively, these results indicate that CpbHLH1 and CpbHLH2 could individually repress and promote CpCYC-B and CpLCY-B genes during papaya fruit ripening.
a, b Vivid figures were shown about reporter and effector constructs according to the dual-luciferase reporter assay. a Reporter; b Effector; c–f CpbHLH1/2 regulated activities of CpCYC-B and CpLCY-B. c CpbHLH1 repressed CpCYC-B; d CpbHLH1 repressed CpLCY-B; e CpbHLH2 promoted CpCYC-B; f CpbHLH2 promoted CpLCY-B; the activation of CpCYC-B and CpLCY-B by CpbHLH1/2 were presented by LUC/REN. The ratio of LUC/REN of the empty construct plus promoter vector or transfectors vector was presented. Values showed as mean + SE of eight biological replications
CpbHLH1/2 involved in expression regulation of CpCYC-B and CpLCY-B responded to light
To investigate if CpbHLH1/2 could regulate gene expression on carotenoid biosynthesis pathway (including CpCYC-B and CpLCY-B) in response to light, we collected SunUp fruit from S1 to S3 stage (green, color break, and ripe). Under 28 °C, papaya fruits of each stage were divided into two parts, one part was exposed to strong white light for 2 days, another part was kept in dark. After 2 days of treatment, expression levels of CpbHLH1/2, CpCYC-B, and CpLCY-B were examined by qRT-PCR. In Fig. 7, the expression levels of CpCYC-B and CpLCY-B in light were higher than in dark. In contrast, the expression levels of CpbHLH1/2 were much lower in light than in dark. Especially, at stages S1 and S2 when the carotenoids have not been accumulated, the expression levels of CpCYC-B and CpbHLH2 were twice in light and in dark, respectively. At ripe stage S3, there was no obvious expression difference for CpbHLH1/2, CpCYC-B, and CpLCY-B. In conclusion, CpbHLH1/2 may be among the factors regulating CpCYC-B and CpLCY-B in response to light.