Comparative transcription analysis of photosensitive and non-photosensitive eggplants to identify genes involved in dark regulated anthocyanin synthesis

Background Light is a key environmental factor in regulation of anthocyanin biosynthesis. Through a large number of bagging screenings, we obtained non-photosensitive eggplants that still have decent amount of anthocyanin synthesized after bagging. In the present study, transcriptome was made to explore the molecular mechanism of dark-regulated anthocyanin synthesis in non-photosensitive eggplant. Results The transcriptome of the pericarp at 0 h, 0.5 h, 4 h, and 8 h after bag removal were sequenced and analyzed. Comparison of the sequencing data with those of photosensitive eggplant for the same time period showed that anthocyanin synthesis genes had different expression trends. Based on the expression trends of the structural genes, it was discovered that 22 transcription factors and 4 light signal transduction elements may be involved in the anthocyanin synthesis in two types of eggplants. Through transcription factor target gene prediction and yeast one-hybrid assay, SmBIM1, SmAP2, SmHD, SmMYB94, SmMYB19, SmTT8, SmYABBY, SmTTG2, and SmMYC2 were identified to be directly or indirectly bound to the promoter of the structural gene SmCHS. These results indicate that the identified 9 genes participated in the anthocyanin synthesis in eggplant peel and formed a network of interactions among themselves. Conclusions Based on the comparative transcription, the identified 22 transcription factors and 4 light signal transduction elements may act as the key factors in dark regulated anthocyanin synthesis in non-photosensitive eggplant. The results provided a step stone for further analysis of the molecular mechanism of dark-regulated anthocyanin synthesis in non-photosensitive eggplant. Electronic supplementary material The online version of this article (10.1186/s12864-019-6023-4) contains supplementary material, which is available to authorized users.


Background
Anthocyanins are a class of secondary metabolites that contribute to the red, blue, and purple colors in a range of flowers and fruits [1]. Anthocyanins play an important role not only in plant physiology, pollinators and visual appeal of seed communicators, but also in human health, such as preventing cardiovascular disease, controlling obesity and reducing diabetes [2,3]. Anthocyanin biosynthesis in plants is affected by many factors, including light, temperature, sugar and hormones, with light being the most important [4][5][6][7].
Light signaling plays a pivotal role in the control of plant metabolism, growth and development. Light signals are perceived by photoreceptors including phytochrome (PHYA-PHYE), cryptochromes (CRYs), phototropins (PHOTs) and UV resistance locus 8 (UVR8, [8][9][10][11][12]). CRYs are blue light receptor proteins of which some functions have been reported, such as the blue light-dependent hypocotyl inhibition, floral initiation in Arabidopsis, and anthocyanin accumulation in eggplant [13][14][15]. Studies showed that BIC (blue-light inhibitor of cryptochromes) binds to CRY2 to suppress the blue light-dependent dimerization, photobody formation, phosphorylation, and degradation, resulting in the reduction of physiological activities of CRY2 and thus the inhibition of photomorphogenesis [16]. Phototropins are blue light receptors that control many responses to optimize photosynthesis efficiency. It was shown that FaPHOT2 played a role in sensing blue light and mediating anthocyanin biosynthesis in strawberry fruits [17]. The phytochrome family of photoreceptors (phyA-phyE) monitor the environment for informational light signals and induce plant growth and developmental responses appropriate to the prevailing conditions. Arabidopsis phytochrome regulates the accumulation of anthocyanins under nitrogen, phosphorus, and cold induction conditions [18,19]. UVR8 was identified originally as a UV-resistance gene shown to contribute to the UV-B-induced flavonoid accumulation and UV-B protection. In Arabidopsis uvr8 mutant, CHS expression and flavonoid content were reduced significantly [19,20].
Eggplant (Solanum melongena L.) is a horticultural crop of high economic value and has more than one thousand years of plantation history in China. Greenhouse cultivation and low light conditions, however, often lead to poor coloration of eggplant and reduce its quality. In this study, we obtained non-photosensitive eggplant through a large number of bagging screenings that still synthesized a significant amount of anthocyanin after bagging. Twelve RNA samples of non-photosensitive eggplant before and after light induction were sequenced with the Illumina Hiseq platform. Through comparative analysis of transcriptome data of both photosensitive and non-photosensitive eggplant samples at the same time points, we aimed to find key genes that took part in darkregulated anthocyanin biosynthesis in non-photosensitive eggplant.

Dark treatment and determination of anthocyanin content
Our previous study showed that the peel of photosensitive eggplant was white when it was bagged, while the anthocyanin content of peel gradually increased after bag removal [15]. Non-photosensitive eggplant, on the other hand, can synthesize anthocyanins under dark conditions. After the non-photosensitive materials were bagged for 24 days, the color of their pericarp was almost the same as that of the un-bagged eggplant (Fig. 1a). The anthocyanin contents of the un-bagged and the bagged eggplants were 78.7 mg/100 g and 65.4 mg/100 g, respectively. Variance analysis of the anthocyanin content showed that darkness also affected the anthocyanin biosynthesis in nonphotosensitive eggplant; however, the eggplant was evenly colored and the fruit attributes were not affected (P value = 0.0012) (Fig. 1b).

De novo assembly and analyses of RNA-Seq data
In our previous study, RNA-Seq was performed for photosensitive eggplant collected at four time points (0, 0.5, 4 and 8 h after bag removal), and 869 genes were found to be involved in the light-induced anthocyanin biosynthesis [51]. In order to compare the similarities and differences of anthocyanin synthesis in different photosensitive eggplants and to understand the dark-regulated mechanism of anthocyanin synthesis in non-photosensitive eggplant, the same four time points after bag removal were chosen to analyze the transcription of non-photosensitive eggplant. RNA-Seq was performed for three biological replicates at each time point. On average, about 4.45 Gb bases were generated from each sample by Illumina Hiseq platform. After mapping sequenced reads to reference genome and reconstruct transcripts, a total of 27,889 novel transcripts were obtained from all samples. Of these, 14,595 are previously unknown splicing event for known genes, and 1034 are novel coding transcripts without any known features. The remaining 12,260 are long noncoding RNAs. After novel transcript detection, the novel coding transcripts were merged with reference transcript to get the complete reference. The clean reads were then mapped to the reference using Bowtie. The gene expression summary is shown in Additional file 1. To test if the samples chosen were reliable, the correlation value between each two samples was calculated based on FPKM. As shown in Fig. 2, the three replicates at four time points showed a good correlation.

Analysis of the differentially expressed genes (DEGs)
Three thousand eight hundred forty-one DEGs were identified by comparing FPKM values between different libraries under the thresholds of false discovery rate ≤ 0.001, absolute Log 2 Ratio value ≥1, and divergence probability ≥0.8 (Additional file 2). In non-photosensitive eggplant, 721 DEGs were found at 0.  (Fig. 3b). The results indicate that there is much difference between photosensitive and nonphotosensitive eggplants in terms of light response patterns and anthocyanin synthesis mechanisms.

qRT-PCR validation of differentially expressed genes
In order to validate the accuracy of RNA-seq data, and SmWD40(Sme2.5_04405.1_g00003.1) related to anthocyanin biosynthesis were selected and subjected to qRT-PCR analysis. The results showed that qRT-PCR highly repeated RNA-seq data and thus verified the accuracy of RNA-seq data (R = 0.78-0.99) (Additional file 3).

GO analysis of the DEGs
GO is a useful program for annotation and functional categorization of genes [52]. The GO assignment was used here to classify the functions of the DEGs in eggplant peel. A total of 3841 DEGs were assigned to different GO ontologies based on their sequence similarity to the genes with previously known functions; specifically, 804, 1493, and 1124 DEGs were assigned to the molecular function, the biological process, and the cellular component categories, respectively (Fig. 4). Same as the analysis result for the DEGs in photosensitive eggplant [51], the largest two subcategories in each of the 'biological process', the 'cellular component', and the 'molecular function' categories were 'metabolic process' and 'cellular process', the 'cell' and 'cell part', and the 'catalytic activity' and 'binding activity', respectively.

Cluster analysis of expression patterns in the DEGs
Three thousand eight hundred forty-one DEGs were subjected to complete-linkage hierarchical clustering  Correlation coefficients for every two samples. Heat map color represents the correlation coefficient; the darker the color, the higher the correlation using a Euclidean distance metric and divided into 27 clusters (Additional file 4). Genes associated with the anthocyanin biosynthesis were enriched in cluster 12,13,14,16,19,20,24,25,26 and 27 (Fig. 5). The expression levels of genes in cluster 12 were down-regulated during 0 h-8 h. The expression levels of genes in clusters 13 and 16 changed slightly at 0.5 h, but were downregulated at 4 h, and then up-regulated at 8 h. The expression levels of genes in clusters 14, 19, 20, 24, 25, 26, 27 also changed very little at 0.5 h but peaked at 4 h, and were down-regulated at 8 h. Through cluster analysis, it was found that most DEGs related to anthocyanin biosynthesis were down-regulated during the period while in photosensitive eggplant, most such DEGs were up-regulated [51]. This result further illustrates that the regulatory modes for non-photosensitive and photosensitive eggplants were different.

The structural anthocyanin biosynthesis genes
From the cluster analysis, it was found that the structural anthocyanin biosynthesis genes had different expression patterns for photosensitive and non-photosensitive eggplants. In order to find the key genes for the synthesis of anthocyanin in non-photosensitive eggplants under dark conditions, structural anthocyanin biosynthesis genes with significant changes after opening the bags in photosensitive and non-photosensitive eggplants were merged. As show in Fig. 6, 26 structural anthocyanin biosynthesis genes were divided into two clusters. The genes in cluster 1 were down-regulated after bag removal in both photosensitive and non-photosensitive eggplants, and their expression levels at 0 h were similar to each other, suggesting that these genes did not participate in the anthocyanin synthesis in non-photosensitive eggplant under dark conditions. The genes in cluster 2, most of which had high expression levels in dark conditions in non-photosensitive eggplant, were up-regulated after bag removal in photosensitive eggplant but were down-regulated or unchanged in nonphotosensitive eggplant (Fig. 6). The expression levels of PAL (Sme2.5_11682.1_g00004.1) and 4CL (Sme2.5_ 00843.1_g00005.1) in non-photosensitive eggplant were g00020.1) and F3'5'H (Sme2.5_04313.1_g00001.1) in non-photosensitive eggplant were more than 200 times that in photosensitive eggplant (Fig. 6). Therefore, it is believed that these genes were involved in the darkregulated anthocyanin synthesis in non-photosensitive eggplant.

Light signal perception and transduction
In order to clarify the difference between the two eggplants in expression patterns of light signal related genes, the DEGs from photosensitive and non-photosensitive eggplants were combined for comparison. As show in Fig. 7, 9 DEGs were divided into two clusters. In cluster 1, PHYB (Sme2.5_00803.1_g00002.1) had the same expression pattern in both photosensitive and non-photosensitive eggplants, but UVR8 (Sme2.5_03981.1_g00003.1) showed different expression trends in photosensitive and non-photosensitive eggplants after 0.5 h exposure to light. In cluster 2, COP, UVR3 and CRY3 were up- Transcription factors related to dark-regulated anthocyanin synthesis Analysis of the structural and light signal sensing and transduction genes related to anthocyanin biosynthesis showed that those likely involved in dark regulation of anthocyanin synthesis in non-photosensitive eggplant were up-regulated in photosensitive eggplant after bag removal (Fig. 6). However, in non-photosensitive  Genes that are up-regulated appear in red, and those that are down-regulated appear in green eggplant, the structural genes were down-regulated at 0.5 h and 8 h. Although up-regulated at 4 h, only a few genes' expressions reached the significant level; the light signal sensing and transduction genes either had the same expression level or had an up-regulated expression trend, but did not reach the significant level at 4 h and 8 h. Since genes with the same expression trend usually have the same function, TFs related to anthocyanin synthesis can be differentiated based on the expression trends of anthocyanin structural genes and light signal sensing and transduction genes. We obtained 338 TFs that changed significantly after removing the bags in photosensitive or non-photosensitive eggplants by merging (Additional file 5); then, we identified 22 TFs that were up-regulated in photosensitive eggplant, but down-regulated or had no significant change in nonphotosensitive eggplant (Fig. 8). In the 22 TF genes, MYB113, TT8, TTG2 had high expression levels at 0 h in non-photosensitive eggplant. These three TF genes were found to participate in eggplant anthocyanin synthesis in our previous study [15,53]; AP2-EREBP, MYC2, MYB77, bHLH87 and WRKY53 also had high expression level at 0 h in non-photosensitive eggplant, implying their importance to anthocyanin synthesis; Other 14 TF genes were expressed similarly at 0 h in photosensitive and non-photosensitive eggplants. Since one of these TF genes, SmMYB86, was considered as a negative regulator of anthocyanin synthesis in our previous studies [51], it is reasonable to believe that other TF genes of the same type may also have participated in dark-regulated anthocyanin synthesis in non-photosensitive eggplant.

Interaction network construction and target gene prediction of the TFs related to dark regulated anthocyanin synthesis
To verify whether the TFs identified in the analysis results were involved in anthocyanin synthesis, we analyzed the 50-1500 bp upstream of the initiation codon of the anthocyanin structural genes and TFs using the Isanger cloud platform (https://www.i-sanger.com/) for regions that the transcription factors may bind. It was found that these TFs could not only regulate structural genes, but also regulate each other (Additional file 6). Possible interactions among these TFs were also predicted through a protein interaction site (http://netbio. sjtu.edu.cn/arappinet/). As shown in Fig. 9, there were interaction relationships between some of the TFs, and the interaction scores were greater than 0.5 (Additional file 7). These findings suggest that these transcription factors may have the same function and are involved in the dark-regulated anthocyanin biosynthesis in non-photosensitive eggplant.

Experimental verification of regulatory relationship between dark regulated anthocyanin synthesis related TFs and target genes
In order to confirm the predicted regulatory relationship between TFs and target genes, the promoters of the non-photosensitive eggplant anthocyanin synthetic  (Fig. 10). Since previous study showed that overexpression of SmCHS and SmANS can promote anthocyanin synthesis [54]. Thus, TFs that bind to their promoters are likely involved in anthocyanin synthesis.

Discussion
Our previous study shows that photosensitive eggplant did not synthesize anthocyanins under dark conditions, and anthocyanin synthesis genes and regulatory factors were not expressed; however, the anthocyanin synthesis genes and regulatory factors were significantly induced at 0.5 h, 4 h and 8 h after exposure to light, and anthocyanins were synthesized within a few days [15]. In order to find the key regulators of synthetic anthocyanins in dark conditions to solve the problem of poor coloration of eggplant peel under low light conditions, transcriptome data of photosensitive and non-photosensitive eggplant peel at 0 h, 0.5 h, 4 h, 8 h after illumination were compared and analyzed in this study.
Structural genes of anthocyanin biosynthesis are simultaneously regulated by the activated TFs. In this study, through comparative transcriptiome we also found 22 TFs that have the same expression pattern as the structural genes (Fig. 8). The 22 TFs can be divided into 8 categories which contained six MYBs, five bHLHs, two WRKYs, four C2C2s, two SPLs, HD, TCP, and AP2. Among these TFs, three MYBs (MYB113, MYB94 and MYB19), three bHLHs (TT8, MYC2 and BIM1), WRKY (TTG2), HD, and AP2 can bind directly to the promoter of CHS (Sme2.5_13923.1_g00001.1), and one C2C2 (YABBY) can bind to the promoter of TT8 (Fig. 10). Although other transcription factors can not bind directly to the promoters of the structural genes, they were predicted to be involved in anthocyanin synthesis by binding to the promoter of TFs which can bind directly to the promoter of anthocyanin structural genes (Additional file 6). Studies also showed that MYB113 (MYB1 or MYB75) [30,33,62,63], TT8 [40,64,65] and TTG2 (WRKY44, [66,67]) were involved in anthocyanin synthesis in many plant species. In addition, other TFs such as SPL [47], TCP (TCP3 and TCP15, [45,46], AP2 [44], WRKY (WRKY41 and WRKY11, [68,69]) and MYC (MYC1, [29]) were also characterized to be involved in anthocyanin synthesis. At present, there is no report on the involvement of homeodomain-like protein (HD) and C2C2 type zinc finger protein in the synthesis of anthocyanins and flavonoids. In this study, MYB113, TT8, TTG2, MYC2 and AP2 have a high expression level in dark condition in non-photosensitive eggplant, and the expression levels of MYB94, MYB19, BIM1 and HD were not much different between the two eggplants in dark condition, but they have different expression patterns in the two materials. Since these TFs bind directly to the promoter of CHS (Sme2.5_13923.1_g00001.1), they were thought to have participated in anthocyanin biosynthesis in eggplant.
While light is one of the key environmental factors for the anthocyanin synthesis in eggplant, this study found that anthocyanins can still proceed without light. Transcriptome comparison analysis showed that UVR8 has different expression pattern among the two types of eggplants (Fig. 7). It has also been shown that UVR8 can promote anthocyanin synthesis in radish sprouts and Arabidopsis, but the synthesis was dependent of light [70,71]. It was reported that AtUVR8 can interact directly with AtBIM1 and AtWRKY36 to participate in hypocotyl elongation [72,73]. We also confirmed in this study that SmBIM1 and SmWRKY53 can directly bind to the promoter of SmCHS (Sme2.5_13923.1_g00001.1). Based on these results, we hypothesize that SmUVR8 can directly interact with SmWRKY53 and SmBIM1 to regulate eggplant anthocyanin synthesis. COP1, an E3 ubiquitin ligase, can form a complex with SPA to regulate anthocyanin synthesis of Arabidopsis; cop1 and spa mutants can produce anthocyanins also in the dark [26]. Jiang et al. reported that SmCOP1 (Sme2.5_00128.1_ g00013.1) can interact with SmHY5 and SmMYB113 to regulate anthocyanin biosynthesis and restore the phenotype of the Arabidopsis cop mutant [15]. In this study, we found two SPAs and COP1 (Sme2.5_00499.1_ g00010.1) which had different expression patterns in the two materials and thus might have participated in the dark-regulated anthocyanin synthesis in nonphotosensitive eggplant (Fig. 7). Cryptochromes are blue-light receptors that regulate plant development and the circadian clock. CRY2 undergoes blue lightdependent homodimerization to become physiologically active. The binding of BICs to CRY2 inhibits the blue light-dependent dimerization, thereby regulating cell photosensitivity [16,74]. In this study, we found that BIC and BIC2 were up-regulated in photosensitive eggplant after bag removal; yet, there was essentially no expression of these genes in non-photosensitive eggplant, leading us to suspect that the non-photosensitive eggplant peel cells might have lost photosensitivity in some cases.

Conclusions
In this study, comparative transcriptome was applied to elucidate the underlying molecular mechanism for darkregulated anthocyanin biosynthesis in non-photosensitive eggplant. The analyses revealed that the structural genes and the TFs involved in anthocyanin synthesis were down-regulated after bag removal, while most of them have high expression levels in dark conditions in nonphotosensitive eggplant. Through transcription factor target gene prediction and yeast one-hybrid verification, we believe that SmBIM1, SmHD, SmMYB94, SmMYB19, SmTT8, SmYABBY, SmTTG2, SmAP2 and SmMYC2 may be involved in dark-regulated anthocyanin synthesis in non-photosensitive eggplant. Our results shaded some lights on the understanding of molecular mechanism of dark-regulated anthocyanin synthesis in non-photosensitive eggplant.

Plant materials
The non-photosensitive eggplant material '145', obtained through a large number of bagging screening of fruits with different eggplant genotypes, was provided by the Shanghai Academy of Agricultural Sciences, Shanghai, China. '145' was grown in the greenhouse of Shanghai Jiao Tong University, Shanghai, China, under natural light conditions. For fruit bagging treatment, sepals were covered with double-layer Kraft paper bags after full bloom. After the bagging of the materials for 2 weeks, the bags were opened to collect materials at 8:00 am. Three materials were selected as biological replicates, with each individual material being sampled from every bagged plant at 0 h, 0.5 h, 4 h, and 8 h after the bags were removed. All peel samples were frozen immediately with liquid nitrogen and kept at − 80°C.

Anthocyanin content determination
Anthocyanin content was measured with the pH differential spectrophotometry method [75]. 0.5 g materials were used to extract anthocyanin with 0.01% HCl in 0.5 ml methanol at 4°C for 12 h. The extracts were centrifuged at 8500 g for 20 min. The supernatant was transferred to a 10 ml tube and the sediments were further processed with 5 ml extraction solution at 4°C for 6 h. The extracts from the sediments were also centrifuged at 8500 g for 20 min. Spectrophotometrical absorbance was measured at 510 nm and at 700 nm in buffers at pH 1.0 and pH 4.5. The anthocyanin content was calculated according to the following formula: TA = A × MW × 5 × 100 × V/ε, where TA stands for total anthocyanin content (mg/100 g), V is the final volume (ml), A equals to [A510 nm (pH 1.0) − A700 nm (pH 1.0)] − [A510 nm (pH 4.5) − A700 nm (pH 4.5)], ε is the molar absorptivity which is 26,900, MW is the molecular weight with a value of 449.2, and A510 and A700 represent the absorbance values at 510 and 700 nm, respectively [76]. Three measurements were performed for each sample replicate.

RNA extraction, cDNA library construction, and RNA-Seq
Total RNAs were extracted from pericarp using the RNAiso plus kit (TaKaRa, Otsu, Shiga, Japan) and following the manufacturer's instructions therein. RNasefree DNase (Takara, Japan) was used to remove the potential genomic DNAs in extracted RNAs. After total RNAs were extracted and treated with DNase I, Oligo (dT) was used to isolate mRNA. The mRNAs were fragmented through mixing with the fragmentation buffer. cDNA was then synthesized using the mRNA fragments as templates. Short fragments were purified and resolved with EB buffer for end reparation and single nucleotide A (adenine) addition. The short fragments were then connected with adapters. Suitable fragments were selected for the PCR amplification. The quality and quantity of RNAs were determined by Agilent 2100 Bioanalyzer and ABI StepOne Plus Real-Time PC System. Finally, the library was sequenced using Illumina HiSeq™ 2000.

Sequence assembly and gene annotation
Low-quality, adaptor-pollute, and high content of unknown base (N) reads were removed by QC before downstream analyses. After reads filtering, the clean reads were mapped to the reference genome by HISAT. Eggplant reference genome database (http://vegmarks. nivot.affrc.go.jp/VegMarks/jsp/index.jsp) was used as the reference. After the novel transcripts were obtained, their coding transcripts were merged with reference transcripts to get the complete reference which was used for gene expression analysis.

Identification and annotation of DEGs
After the setup of the complete reference, the clean reads were mapped to the reference library using Bowtie2; then gene expression level for each sample was calculated by RSEM. DEGs were detected by DEseq2 and fold change ≥ 2.00 and adjusted p-value ≤ 0.05 were taken as conditions. Gene Ontology (GO) classification and functional enrichment were performed by WEGO software (http://wego.genomics.org.cn/cgi -bin/wego/index.pl). KEGG pathway classification and functional enrichment for DEGs were performed by the database-Kyoto Encyclopedia of Genes and Genomes. Cluster analysis of expression patterns were analyzed by MeV (4.9).

Real-time qPCR analysis
Real-time quantitative reverse transcription-PCR was used to estimate the accuracy of RNA-Seq. Total RNA was treated with DNase to remove trace of DNA. 1 μg RNA was used for synthesis of cDNA with the PrimeScript RT Master Mix Perfect Real Time Kit (Takara). Quantitative real-time RT-PCR (qPCR) was performed on FTC-3000 real-time PCR System (Funglyn Biotech, Canada) using the following program: 95°C for 30 s, followed by 40 cycles of 95°C for 20 s, 58°C for 30 s, and 72°C for 30 s, which was consistent with the manufacturer's instructions of SYBR Premix Ex Taq II Kit (Takara). The Actin gene (GU984779.1) was used as an internal reference, and relative expression was calculated using the 2ΔCt method. Each qPCR analysis was performed in triplicate. The relative expression levels of the amplified products were analyzed using the comparative CT method based on CT values. All primers used in this study are listed in Additional file 8.