Comparative transcriptome analysis provides insight into the molecular mechanisms of long-day photoperiod in Moringa oleifera

Abstract

Moringa oleifera, is commonly cultivated as a vegetable in tropical and subtropical regions because of nutritional and medicinal benefits of its fruits, immature pods, leaves, and flowers. Flowering at the right time is one of the important traits for crop yield in M.oleifera. Under normal conditions, photoperiod is one of the key factors in determining when plant flower. However, the molecular mechanism underlying the effects of a long-day photoperiod on Moringa is not clearly understood. In the present study, deep RNA sequencing and sugar metabolome were conducted of Moringa leaves under long-day photoperiod. As a result, differentially expressed genes were significantly associated with starch and sucrose pathway and the circadian rhythm-plant pathway. In starch and sucrose pathway, sucrose, fructose, trehalose, glucose, and maltose exhibited pronounced rhythmicity over 24 h, and TPS (trehalose-6-phosphate synthase) genes constituted key regulatory genes. In an Arabidopsis overexpression line hosting the MoTPS1 or MoTPS2 genes, flowering occurred earlier under a short-day photoperiod. These results will support molecular breeding of Moringa and may help clarify to genetic architecture of long-day photoperiod related traits.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12298-022-01186-4.

Introduction

Moringa (Moringa oleifera Lam.) is an important edible and medicinal vegetable (Pakade et al. 2013; Omotoso et al. 2018). Moringa fruits, immature pods, leaves, and flowers contain numerous dietary phytonutrients and antioxidants that are essential to human health, such as flavonoids, minerals, vitamins, dietary fiber, and protein (Tumer et al. 2015; Liao et al. 2018; Lin et al. 2018a). According to global distribution data from the Global Biodiversity Information Facility, Moringa has been identified in 1136 countries or territories between 1800 and 2019, including Benin, Mexico, Australia, and Brazil (Hassler 2020). Compared with the distribution of solar energy resources worldwide, Moringa tends to be distributed in areas with high annual average solar radiation (except for the Atacama, Sahara, and Namib deserts). Under natural conditions, external or environmental cues, especially light (light or duration of exposure) or photoperiod (the amount of light and darkness in a daily cycle of 24 h) is a major factor in determining flowering time, which is one of the most important traits with respect to crop yield (Song et al. 2013). In previous studies, many plants use information about photoperiod changes to adjust flowering time and seasonal changes to improve reproductive success (Kantolic and Slafer 2005; Kantolic et al. 2013). For example, it was reported that photoperiod mediated yield determination effect can increase soybean (Glycine max) crop yield under natural field conditions (Nico et al. 2019). However, knowledge about the role of photoperiod and its mechanisms in M. oleifera are not clearly understood due to their long generation time and long juvenile period.

In plants, photosynthetic carbon dioxide assimilation and plant growth development are directly or indirectly dependent on light. At night, plants are completely dependent on metabolites they accumulated when light was available. Accordingly, daylength and nightlength are important for plant growth. Starch is the main storage metabolite in many plants; its turnover is strictly regulated by a biological clock, such that it is nearly exhausted at dawn to prevent premature complete consumption that would lead to carbon starvation at night (Stitt and Zeeman 2012). Previous research has revealed that sugar-related signals are integrated into flowering-related genetic pathways in plants. FT (Flowering locus T) and its earliest target gene, SOC1 (suppressor of overexpression of constans 1), induce the transcription of SWEET10. The overexpression of SWEET10 accelerates flowering slightly under a long photoperiod (Andres et al. 2020). Additionally, the INDETERMINATE DOMAIN transcription factor AtIDD8 plays a role in the molecular link between sugar metabolism and photoperiodic flowering, AtIDD8-overexpressing plants (35S:IDD8) have been shown to exhibit early flowering, whereas idd8 mutants exhibited delayed flowering under long-day conditions (Seo et al. 2011). In plants, Trehalose-6-phosphate (T6P) also plays an important role in promoting flowering with respect to carbohydrate levels. For example, knockdown of T6P synthase 1 (TPS1) expression via artificial miRNA (35S:amiR-TPS1) in Arabidopsis resulted in significant reduction of T6P levels and late flowering. Conversely, enhancement of sucrose levels or misexpression of TPS1 in Arabidopsis led to elevated T6P levels and accelerated flowering, suggesting that T6P acts as a specific signal of sucrose status (Wahl et al. 2013; Schluepmann et al. 2003; Yadav et al. 2014).

Carbohydrates are basic building blocks of life and important nutrients, which also participate in many biological activities such as cell recognition, immune responses, metabolic regulation, fertilization processes, morphogenesis, development, and cancer. Sugar analysis is a precise and widely employed means of studying the transformation and physiological functions of sugar in the living body, and monosaccharides are regarded as the basic units of polysaccharides. In this study, we first analyzed the changes of sugar content in Moringa leaves over a 24-h time course. Twelve monosaccharides and disaccharides were identified, among which the levels of five sugars exhibited pronounced rhythmicity. Levels of D-fructose and maltose were significantly different between the 20-h and 16-h timepoints, as well as between the 4-h and 20-h timepoints. Subsequently, we performed transcriptome analysis of Moringa leaves at three different zeitgeber times (ZT = 4 h, ZT = 16 h, and ZT = 20 h). Dozens of differentially expressed genes (DEGs) related to starch and sucrose metabolism were identified, some were further examined using quantitative real-time polymerase chain reaction (qRT-PCR). Finally, overexpression of trehalose-6-phosphate synthase gene MoTPS1 or MoTPS2 in wild-type Arabidopsis led to significantly shorter flowering time under a short-day condition. This study is helpful to better understand the diurnal changes in sugars and their related genes in Moringa leaves at both metabolic and molecular levels. Furthermore, MoTPS1 and MoTPS2 genes positively regulated flowering.

Materials and methods

Plant materials and sampling

PKM-1 Moringa seeds were planted in an incubator under the following conditions: 16-h light/8-h dark cycle, 8000 lx light intensity, and 32 °C culture temperature. Daylength is perceived by the leaf rather than the flower or shoot apical meristem. Thus, in each treatment, all leaves were collected from three independent biological replicates of Moringa seedlings. Additionally, all leaves of 3-week-old seedlings with a height of 25–30 cm were collected at three ZTs (4, 16, and 20 h) for transcriptomic analysis and at six timepoints over a 24-h course for sugar analysis according to the method of Zhu et al. (2021). Sugar determination includes the following twelve types: sucrose, inositol, glucose, D-fructose, maltose, L-fucose, D-galactose, L-rhamnose, D-arabinose, trehalose, D-sorbitol, and xylitol. All samples were immediately frozen in liquid nitrogen and preserved in a − 80 °C freezer.

RNA extraction, library construction, and sequencing

Total RNA was extracted from frozen Moringa leaves using a TRIzol kit (Promega, Beijing, China), and treated using DNase I (TaKaRa, Shiga, Japan). Then, double-stranded cDNAs were synthesized and purified by a QIAquick PCR extraction kit (Qiagen, Hilden, Germany). The ends were repaired, poly (A) tails were added, and the cDNAs were ligated to sequencing adapters (Illumina, San Diego, CA, USA). Finally, the size of the product was selected by agarose gel electrophoresis and amplified by PCR. The cDNA library was sequenced using an Illumina HiSeq™ 2500 platform (Gene Denovo Biotechnology Co., Guangzhou, China).

RNA-seq analysis

First, raw reads were filtered to obtain high-quality clean reads. Next, rRNA-mapped reads were removed using the short-read alignment tool Bowtie2 (Langmead and Salzberg 2012). Thereafter, the reads from each sample were mapped to a reference genome sequence, provided by Tian et al. (2015) using TopHat2 (Kim et al. 2013; Tian et al. 2015). The raw data have been submitted to the National Center for Biotechnology Information (NCBI) and the bioproject ID is PRJNA835993. The gene expression level was expressed by FPKM (fragments per kilobase of transcript sequence per million base pairs) and principal component analysis (PCA) was performed to reveal the relationship of samples by R package in the present study. Finally, the sequences were annotated based on the BLASTX search against public databases, including the National Center for Biotechnology Information database (NCBI), Swiss-Prot (A manually annotated and reviewed protein sequence database), Clusters of Orthologous Groups. Gene Ontology (GO) (fdr < 0.05) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (p < 0.01) enrichment analysis of DEGs were implemented by the clusterProfiler R package. Differentially expressed genes (DEGs) between samples were identified using R package. A false discovery rate < 0.05 and fold change ≥ 2 were considered thresholds for significant differential expression.

qRT-PCR analysis

Differential expressions of starch and sucrose metabolism genes detected in RNA-seq analysis were confirmed using qRT-PCR. The relative expression level of each gene was calculated via the 2^−ΔΔCt method (Livak and Schmittgen 2001). Three biological replicates were analyzed for each sample. The primers used for qRT-PCR are listed in Table S1 and the corresponding reference gene MoTUB were selected according to previous research reports (Lin et al. 2018b).

Overexpression vector construction and Arabidopsis transformation

Gateway primers (Table S2) containing the attB site were designed to amplify MoTPS1 (MG736634) and MoTPS2 (MG736635), which were cloned into pDONR207 vectors (Lin et al. 2018b). Then, these two genes were cloned into separate pEarlyGate101 vectors, which contains the 35S promoter sequence and the enhanced yellow fluorescent protein tag. Each constructed vector was transformed into Agrobacterium tumefaciens strain GV3101 via the heat-stroke method. Genetic transformation of Arabidopsis was performed using the Floraldip method (Zhang et al. 2006). Finally, 0.1% Basta® (Bayer) was used to screen for positive Arabidopsis soil-grown seedlings. The primers used in vector construction are listed in Table S2.

Subcellular localization of MoTPS

Verified clones were transferred into the pEarleyGate 101 destination vector to produce C-terminal yellow fluorescent protein fusion constructs (Early et al. 2006). Each pEarlyGate101-MoTPS construct was transformed into A. tumefaciens strain C58. C58 cells carrying pEarlyGate101-MoTPS were infiltrated into Nicotiana benthamiana, as previously described (Sparkes et al. 2006). The MoTPS1 and MoTPS2 constructs were co-infiltrated with a Glycine max (soybean) a-mannosidase-mCherry (pm-rk CD3-1007) plasma membrane marker (Nelson et al. 2007).

Results

Transcriptome analysis of Moringa at different photoperiod timepoints

Within a day, the transitions from night to day and from day to night are key timepoints. To further investigate the molecular activity in Moringa leaves at different timepoints under a long-day condition (16-h light/8-h dark), nine cDNA libraries (at ZT = 4 h, ZT = 16 h, and ZT = 20 h) were constructed using high-throughput RNA-seq analysis. In total, 226,132,498 high-quality clean reads were obtained (22,114,928–28,770,920 reads per library; Table 1). The Q30 percentage (sequences with sequencing error rate < 0.1%) was > 89% for each library; the average GC content was 46% for all libraries. For each library, 86.50–96.35% of the clean reads mapped to the reference genome. There were 14,990 known genes and 1228 new genes. These results indicate that these high-quality output data could be used for further analysis.

Table 1.

Summary Statistics of RNA-seq Results in Moringa leaves

Item	HQ Clean Reads	Q30 (%)	GC (%)	Mapping ratio	Known gene number	New gene number
4h-1	25,139,928 (95.49%)	3250937430 (90.42%)	1716988845 (47.75%)	21883363 (87.33%)	14795 (76.01%)	1229
4h-2	22,952,388	2941490503 (89.97%)	1592000460	19792569 (86.75%)	14635 (75.19%)	1203
	(95.07%)		(48.69%)
4h-3	22,114,928	2835631942 (90.01%)	1520095801	19174891 (86.95%)	14792 (75.99%)	1196
	(95%)		(48.25%)
16h-1	24,519,482	3495329149 (96.73%)	1679487562	22279196 (91.02%)	15212 (78.15%)	1249
	(98.57%)		(46.48%)
16h-2	25,926,944	3693027236 (96.62%)	1775712211	23562132 (91.13%)	15207 (78.12%)	1259
	(98.47%)		(46.46%)
16h-3	27,610,142	3929242071 (96.49%)	1887464553	25189784 (91.35%)	15192 (78.05%)	1262
	(98.31%)		(46.35%)
20h-1	28,770,920	3694972148 (90.02%)	1951360481	24887531 (86.75%)	15231 (78.25%)	1235
	(95.07%)		(47.54%)
20h-2	26,880,598	3469797478 (90.25%)	1853287831	23285786 (86.87%)	14921 (76.66%)	1206
	(95.55%)		(48.20%)
20h-3	22,217,168	2847655974 (90.00%)	1498776045	19150606 (86.50%)	14928 (76.69%)	1217
	(94.74%)		(47.37%)

Analysis of DEG expression

To identify DEGs between two timepoints under a long-day condition in Moringa, we first performed gene expression profile clustering and calculated the correlation coefficients between biological replicates (higher than 0.97). In addition, principal component analysis also indicated that the distances between pairs of points were small (Figure S1), implying that biological replicates were very similar and there were no outlier samples. Based on pairwise comparisons with |log₂ (fold change)|> 1 and false discovery rate < 0.05 as thresholds, 7153 DEGs were identified between ZT = 16 h and ZT = 20 h and between ZT = 20 h and ZT = 4 h. There were 1432 DEGs common to both comparisons (Fig. 1b). Between ZT = 16 h and ZT = 20 h, there were 4794 (1536 upregulated and 3258 downregulated) DEGs, between ZT = 20 h and ZT = 4 h, there were 3791 (813 upregulated and 2978 downregulated) DEGs (Fig. 1a). Biological events in Moringa may be more complex between the 16-h and 20-h timepoints than between the 20-h and 4-h timepoints, because there was a greater proportion of DEGs with larger differences in expression in the first period (Fig. 1c and d). All 7153 DEGs were assigned to eight different clusters, among which the clusters associated with profiles 0, 1, and 3 were statistically significant (p < 0.05). These findings indicated that some genes were activated, but more genes were inhibited from ZT = 16 h to ZT = 20 h to ZT = 4 h (Fig. 2).

Fig. 1.

Analysis of differentially expressed genes. Statistical histogram of gene differences between groups (a). Venn diagram showing the overlapping DEGs between two comparison (b), Volcano plots of DEG expression pattern in ZT_20h vs ZT_16h (c) and ZT_4h vs ZT_20h (d). Yellow, red, and blue dots indicate down- and up-regulations or no differences in gene expression, respectively

Fig. 2.

Clustered gene expression profiles from moringa leaves. The clusters were defined on the basis of gene temporal expression profile in R using the k-means method. The y-axis represents the standardized FPKM value of genes, and the x-axis represent different samples

Blast2GO was used for GO-term enrichment analysis to elucidate the biological functions of DEGs. Between ZT = 16 h and ZT = 20 h, the DEGs were significantly associated with the following eight biological processes, according to GO analysis: DNA metabolism, nucleobase-containing compound metabolism, oligosaccharide metabolism, cellular aromatic compound metabolism, disaccharide metabolism, nitrogen compound metabolism, heterocyclic metabolism, and organic cycle compound metabolism (Fig. S2a; Table S3). Between ZT = 20 h and ZT = 4 h, the DEGs were significantly associated with the following seven cellular components, according to GO analysis: photosystem, photosynthetic membrane, thylakoid part, intrinsic membrane component, thylakoid, cell periphery, and external encapsulating structure (Fig. S2b; Table S3). These results suggest that biological activities differed between the two time periods. From 16 to 20 h, the DEGs mainly regulated DNA metabolism, oligosaccharide metabolism, and organic ring compound metabolism. Conversely, from 20 to 4 h, the DEGs were mainly expressed in cell components related to photosynthesis, such as the light system, cooperative light-harvesting membranes, and thylakoid system.

To identify the metabolic pathways involved in DEGs, KEGG analysis was further performed. Nine (ZT = 16 h versus ZT = 20 h) and six (ZT = 20 h versus ZT = 4 h) pathways were significantly enriched (p ≤ 0.01; Table S4). Pathways related to “starch and sucrose metabolism” (ko00500) and “circadian rhythm–plant” (ko04712) were significantly enriched in both time periods (Fig. 3 and Table S4). The starch and sucrose metabolism pathway and plant circadian rhythm pathway may regulate the response of Moringa leaves to light/dark transitions within a photoperiod. As we collected the samples at different timepoints during the photoperiod, we expected DEGs to be enriched in the circadian rhythm pathway. Therefore, we focused on analyzing the DEGs associated with the starch and sucrose metabolism pathway, to elucidate the relationship between this pathway and the regulation of plant photoperiodism.

Fig. 3.

Significantly enriched KEGG pathways from DEGs in ZT_20h vs ZT_16h (a) and ZT_4h vs ZT_20h (b). The enriched starch and sucrose metabolism pathway and plant circadian rhythm pathway are labeled by the red frame

Differential expression of starch and sucrose metabolism genes in Moringa leaves

In starch and sucrose metabolism pathway, 60 DEGs were identified between ZT = 16 h and ZT = 20 h and 51 DEGs were identified between ZT = 20 h and ZT = 4 h. In the first time period, the expression levels of 27 DEGs varied more than four-fold; these included five MoTPS genes, two genes for MoGLGB, MoPME, MoGAE, MoPGLR, MoBAM, and one gene for MoPHS, MoHXK, MoTPP, MoAGL, MoUGDH, MoPHSL, MoBGL, MoBGLU, MoPMEU, MoRG, MoSUS, and MoGLGL (Table S5). Among these DEGs, the expression level of MoTPS2 decreased by 24-fold, whereas that of MoTPS4 increased by > 59-fold. The expression levels of MoTPS, MoGAE, MoUXS, MoPGLR, MoSPSA, MoRG, MoINV1, MoGLGB, MoSUS, and MoPME significantly differed (|log₂ (fold change)|> 2) between ZT = 20 h and ZT = 4 h. Among these DEGs, MoTPS4 (Lamu_10006934) was downregulated by > 26-fold.

Diurnal time course of sugar accumulation over 24 h in Moringa leaves

To investigate the constituents and kinetic patterns of sugars in Moringa leaves over an entire day, total monosaccharides and disaccharides in Moringa leaves were first measured at six timepoints within a 24-h period (ZT = 4 h, 8 h, 12, 16 h, 20 h, 24 h). A total sugar concentration of 106.12 mg/g was detected in Moringa seedling leaves (Figure S3). Sucrose was the dominant sugar in Moringa leaves, comprising 44–58% of total sugars, this was followed by inositol (17–31.4 mg/g), glucose (14.4–25.4 mg/g), D-fructose (1.8–6.9 mg/g), and eight other sugars (Figure S3). No lactose was detected in the leaves tested (Figure S3). Distinct rhythms were observed in sucrose, D-fructose, trehalose, and glucose concentrations over a 24-h time course, and highest levels were observed toward the end of the day, while lowest levels were observed toward the end of the night (Fig. 4 and S4). In contrast, maltose levels followed the opposite trend in terms of diurnal regulation (Fig. 4 and S4). Based on pairwise comparisons, D-fructose and maltose concentrations were significantly different between ZT = 16 h and ZT = 20 h, as well as between ZT = 20 h and ZT = 4 h (Fig. S4).

Fig. 4.

Diurnal time course of sugar accumulation and DEG expression over 24 h in Moringa leaves. The change of key sugar content over 24 h is represented by line chart. Values are the mean ± SD (n = 3). Bars indicate SD. The annotation and the expression patterns of key genes can be found in Supplementary Table S8

qRT-PCR confirmation that diurnal changes in expression of DEGs were related to starch and sucrose biosynthesis

To confirm that the DEGs were associated with starch and sucrose biosynthesis, diurnal changes in seven genes (MoTPS2, MoUGP2, Moscrk7, MoSPSA4, MoPYG, MoBAM3, and MoAGL) were analyzed using qRT-PCR (Fig. 4, Table S6 and Figure S5). The expression patterns of these genes by qRT-PCR analysis were similar with those from RNA-seq analysis, pronounced rhythmicity in the expression levels of these seven genes was observed over the 24-h time course. Among these genes, MoTPS2 exhibited the highest expression level. Most genes homologous to MoTPS2 (e.g., MoTPS1, MoTPS3, MoTPS6, MoTPS7, and MoTPS8) also exhibited rhythmicity. The expression levels of MoTPS1 and MoTPS2 increased upon light exposure and decreased in the absence of light, whereas the MoTPS6–MoTPS8 expression levels followed the opposite trend, suggesting that MoTPS1 and MoTPS2 may be involved in flowering metabolism in Moringa.

Subcellular localization of MoTPS1 and MoTPS2 proteins

To verify the subcellular localization of MoTPS1 and MoTPS2 proteins, we analyzed their heterologous expression in N. benthamiana leaves using MoTPS isoforms fused to enhanced yellow fluorescent protein at the C terminus with a cauliflower mosaic virus 35S promoter and a plasma membrane marker (pm-rk CD3-1007 mCherry). MoTPS1 and MoTPS2 produced diffused fluorescent signals throughout the cytoplasm (Fig. 5). Due to the presence of large vacuoles in N. benthamiana cells, the fluorescence signal is most pronounced in regions close to the cell membrane. In addition, confocal laser scanning microscope videos revealed that enhanced yellow fluorescent protein fluorescent signals from MoTPS1 and MoTPS2 moved continuously with the flow of cytoplasm.

Fig. 5.

Subcellular localization of MoTPS1 and MoTPS2 proteins. YFP fused to the MoTPS1 or MoTPS2 proteins are shown in green, Plasma membrane marker is shown red, the bright MoTPS1-YFP or MoTPS2-YFP label was co-localized with marker

Screening the overexpression line

To further elucidate the functions of MoTPS1 and MoTPS2, the open reading frames of MoTPS1 and MoTPS2 were inserted into the pEarleyGATE101 vector, which was then transformed into Arabidopsis. After screening with 0.1% Basta, an herbicide, more than 10 positive seedlings were obtained (Figure S6). RNA was extracted from the leaves of three positive seedlings and further analyzed using PCR and northern blot imprint hybridization. The results confirmed that Col-35S:MoTPS1 and Col-35S:MoTPS2 were transgenic plants. MoTPS1 and MoTPS2 were strongly expressed in each strain, indicating that the Moringa genes were well-integrated into the Arabidopsis genome and that no contaminating mRNA was present from other organisms.

Phenotypic analysis of MoTPS1 and MoTPS2 overexpression T3 generation seedling lines

The MoTPS1 overexpression line exhibited significantly accelerated flowering (14.48 days earlier) and had fewer leaves, compared with wild-type Arabidopsis plants under a short-day condition (Fig. 6 and Table S7). The MoTPS2 overexpression line exhibited similar phenotypes, with early flowering and < 12.42 rosette leaves (Fig. 6 and Table S7). However, these differences (versus wild-type Arabidopsis) were not statistically significant under a long-day condition (Table S8).

Fig. 6.

Northern blot imprint hybridization (a–f), 84-day-old plants of Col-0, Col-35S:MoTPS1_1, Col-35S:MoTPS1_2, Col-35S:MoTPS1_3, Col-35S:MoTPS2_1, Col-35S:MoTPS2_2, Col-35S:MoTPS2_3 (g)

Discussion

M. oleifera, a highly nutritious vegetable with valuable medicinal properties, is widely distributed across tropical and subtropical regions. Differences in the response to photoperiodic changes strongly influence plant metabolism and the geographical distribution of plants (Seaton et al. 2018). However, little is known regarding the molecular mechanism of photoperiodism in Moringa. In the current study, we performed sugar metabolite and transcriptome analysis to identify the sugar composition in the studied species and characterize the genes that are involved in the regulation of photoperiodism.

In plants, sugars play key regulatory roles in carbon metabolism and act as signal molecules in growth, development, anthocyanin biosynthesis, and responses to environmental changes (Teng et al 2005; Sulpice et al. 2014; Van Dingenen et al. 2016; Vimolmangkang et al. 2016). Sugar profiling analysis revealed the presence of 12 simple sugars (no lactose), including sucrose, inositol, glucose, D-fructose, and maltose, as well as seven other sugars, these findings add to the knowledge base regarding sugars in Moringa leaves. Sucrose was the dominant sugar present in Moringa seedling leaves, similar to results reported from a previous study that used air-dried samples of M. oleifera obtained from south Algeria (Ziani et al. 2019). Upadhyay et al. (2015) reported that L-arabinose, D-galactose, L-rhamnose, D-mannose, and D-xylose, rather than sucrose, were the predominant sugars in purified M. oleifera whole-gum exudates. Their findings suggested that environmental (e.g., topography, temperature, and soil variation) and biotic factors (e.g., genetics and endogenous hormones) may influence sugar synthesis. Our observations of the diurnal changes in sucrose, glucose, fructose, and maltose levels provide clues to guide further studies regarding the nutritional and medicinal value of M. oleifera. For example, because of sugar attachment activity (glycosylation), M. oleifera may be pharmacologically more active than Moringa ovalifolia, as flavonoid glycosides (hydrophilic compounds) are significantly more bioavailable compared to flavonoid aglycones. Thus, dynamic changes in sugar levels due to glycosylation affect the activities and types of metabolites (e.g., flavonoids) (Makita et al. 2016; Gonzales et al. 2016). In addition, although the trehalose content in Moringa leaves was < 0.01 mg/g, it exhibited an obvious circadian rhythm, decreasing in the absence of light and increasing in the presence of light. These results suggest that some kind of sugar accumulation in M.oleifera leaves involved the biological clock and regulated by the photoperiodism pathway, which could be helpful for future breeding and nutritional analysis.

In total, 84 DEGs that encode enzymes associated with starch and sucrose metabolism (e.g., TPS, GAUT, UXS1, UGDH, HK, glgA, glgB, glgC, AGL, XYL4, AMY, and MGAM) were identified via KEGG enrichment analysis and gene functional annotation. Among them, 60 DEGs (34 downregulated and 26 upregulated) were identified during the light-to-dark transition and 51 DEGs (nine upregulated and 42 downregulated) were identified between the 20-h and 4-h timepoints. Notably, gene expression analysis revealed that the expression patterns of TPS, UGP2, scrk, SuSy, PYG, BAM, and AGL exhibited pronounced rhythmicity over a 24-h time course under a long-day condition. The expression patterns of TPS, SuSy, and AGL were correlated with diurnal changes in their products, whereas expression patterns of BAM and maltose followed the opposite trend. Importantly, TPS was expressed at high levels, the expression levels of five genes in the MoTPS gene family in Moringa leaves followed an obvious circadian rhythm, although trehalose content remained very low. These results suggest that the TPS gene family plays a role within Moringa leaves. It has been reported that TPS genes play crucial roles in plant growth and flowering, by regulating the degradation of starch in plants in tandem with the circadian clock, as well as in Arabidopsis flowering through the photoperiodism and age pathways (Wahl et al. 2013; Figueroa and Lunn 2016; Oszvald et al. 2018). In our study, flowering was significantly earlier in Arabidopsis plants that overexpressed MoTPS1 or MoTPS2 than in wild-type plants under a short-day condition. Under this condition, the MoTPS1 and MoTPS2 overexpression lines produced 34.75–39.60 and 35.00–37.19 rosette leaves, respectively. However, under a long-day condition, the overexpression lines did not exhibit phenotypes that significantly differed from those of wild-type plants. This could be due to longer exposure to sunlight, which led to greater accumulation of nutrients during the day, and/or functional redundancy between the MoTPS and AtTPS genes in Arabidopsis.

A second group of 29 DEGs was significantly enriched in the plant circadian rhythm pathway. 21 genes (two upregulated and 19 downregulated) were differentially expressed between ZT = 16 h and ZT = 20 h, while 19 genes (10 upregulated and nine downregulated) were differentially expressed between ZT = 20 h and ZT = 4 h. Of these DEGs, two COP1 genes and two SPA1 genes were synchronously expressed–they were upregulated from ZT = 16 h to ZT = 20 h and downregulated from ZT = 20 h to ZT = 4 h. Reportedly, SPA proteins interact with COP1 to form stable complexes in dark conditions, repressing seedling photomorphogenesis in Arabidopsis (Zhu et al. 2008; Perrella et al. 2020). Moreover, PIF3 and LHY, regulators of the plant circadian clock, were significantly upregulated during the light-to-dark transition in Moringa. In dark conditions, high levels of PIF3 accumulate to maintain skotomorphogenesis, whereas COP1 represses photomorphogenesis (Bauer et al. 2004; Dong et al. 2014). Previous research has revealed that the clock components ELF3 and GI, which physically interact with PHYTOCHROME B, are essential for perception of light input in the oscillator. Therefore, they are essential for photoperiod measurement with respect to growth and development responses (Kolmos et al. 2011; Yeom et al. 2014; Anwer et al. 2020). In Moringa leaves, the expression levels of ELF3, PRR7, GI, and TOC1 changed over the course of the dark-to-light transition. There are three transcription-translation oscillator feedback loops in A. thaliana, including a central oscillator (negative feedback loop involving CCA1/LHY and TOC1/PRR1), a morning loop (CCA1/LHY represses PRR7 and PRR9, which then repress CCA1/LHY), and an evening loop (TOC1 represses GI, which later activates TOC1) (Fukushima et al. 2009). When considered in the context of its geographical distribution, these results indicate that Moringa is sensitive to photoperiod length and influenced by latitudinal photoperiodicity.

In summary, we elucidated the sugar constituents and kinetic patterns of monosaccharide and disaccharide accumulation in Moringa. We also identified sugars that were differentially expressed at different ZTs in Moringa leaves. Notably, DEGs involved in the starch and sucrose metabolism pathways were identified and expression levels of some of these genes were validated using qRT-PCR. Analysis of the DEGs showed that flowering was significantly accelerated in Arabidopsis lines overexpressing the MoTPS1 or MoTPS2 genes. Our results contribute to a better understanding of sugar composition and accumulation patterns in Moringa, and also enhance knowledge regarding the molecular mechanism underlying starch and sucrose metabolism in Moringa leaves and its geographical distribution. Finally, our data provide valuable information to guide further studies regarding the roles of TPS genes in the regulation of flowering in woody plants.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

MoGLGB

Moringa oleifera 1,4-alpha-glucan branching enzyme

MoPME

Moringa oleifera pectinesterase

MoGAE

Moringa oleifera UDP-glucuronate 4-epimerase

MoPGLR

Moringa oleifera pectin lyase-like superfamily protein

MoBAM

Moringa oleifera beta-amylase

MoPHS

Moringa oleifera poor homologous synapsis

MoHXK

Moringa oleifera hexokinase

MoTPP

Moringa oleifera trehalose 6-phosphate phosphatase

MoAGL

Moringa oleifera glycogen debranching enzyme

MoTPS

Moringa oleifera trehalose 6-phosphate synthase

MoUXS

Moringa oleifera UDP-glucuronate decarboxylase

MoSPSA

Moringa oleifera sucrose-phosphate synthase

MoINV1

Moringa oleifera cell wall invertase 1

MoSUS

Moringa oleifera sucrose synthase

MoScrk

Moringa oleifera probable fructokinase