Morphology and scanning electron microscopy (SEM) observation during C. migao seed germination
During the four germination stages of C. migao seeds, structural changes were observed in the cotyledon cells under a scanning electron microscope (Figure 2). The diameter of the cotyledon cells were 52–72 μm in the cell full of oil bodies. In the dry seed stage, smooth and plump spherical oil bodies with a diameter of 3–12 μm were present in the cotyledon cells. Globular particles, such as oil bodies, were covered with amorphous enzyme (Figure 2A and B). With seed germination, tangible microstructural changes could be observed in the oil bodies and other storage substances in cells under the action of enzymes. After the seeds absorbed water, the oil bodies were completely wrapped by the activated amorphous enzyme, the structure of the smooth oil bodies disappeared (Figure 2C and D), and some precipitates were observed on oil bodies, which resulted in their rough surfaces. From the seed dehiscence to germination stage, the surface of the oil bodies were hollow and honeycomb and the oil bodies in the cell were significantly consumed (Figure 2E-H).
Physiological characteristics of seed germination
Compared with dry seeds, the SSC, starch, and lipid contents of unit weight seed significantly decreased with the rapid entry of water, which decreased by 11.06%, 4.87%, and 32.25%, respectively; on the other hand, the content of soluble protein significantly increased by 6.59%. Compared with dry seeds, the SSC and soluble protein content continuously increased in the seeds in the dehiscence and germination stages, with the highest increase of 35.88% and 28.66%, respectively; the starch and lipid content of storage materials decreased during seed germination; the contents decreased by 28.69% and 43.86% and 32.25% and 34.79%, respectively (Figure 3).
Transcriptome functional annotation and expression profiling
With the aiming to investigate the transcriptional landscape of C. migao seed germination, we performed RNA-Seq to analyze the variations in the transcripts during the four stages of germination (each stage sample contained three replicates). The raw data were filtered from 12 cDNA libraries, and the Q30 of all samples were more than 95% (Table S2). Assembling of the sequencing data led to the identification of a total of 78,832 unigenes; the length of N50 was 1,560 bp. To obtain comprehensive information on the assembled transcriptomes, the nonredundant sequences were annotated based on a similarity search against the Nr, Swissprot, KOG, KEGG and GO databases using a significant threshold E-value of ≤ 10−5 and the BLAST algorithm. In addition, 78,832 unigenes distributed to each of the databases, with 31,313 (39.72% of the total) for Nr, 17,047 (21.62%) for KOG, 19,363 (24.56%) for SwissProt, 12,305 (15.61%) for KEGG, and 9,284 (11.78%) for GO were investigated (Figure 4). Unfortunately, 47,302 unigenes (60.0%) could not be functionally annotated in the current study, which was likely due to the presence of unique genes in the exceedingly special species of C. migao.
To obtain the transcriptional dynamic expression pattern during seed germination, we used the STEM software to classify the differentially expressed genes (DEGs) in the different germination stages (GZ, SX, LK, and MF). Meanwhile, a total of 26 expression profiles were obtained, of which 12 different expression patterns (K1–K12, P < 0.01) were highly significant (Figure 5A). Using hierarchical clustering, we classified the DEGs into seven coexpression classes (C1–C7), each of which contains genes with highly similar expression patterns (Figure 5B). The expression patterns of the 12 expression profiles were consistent with those of the 7 coexpression classes. The gene expression levels of C1 and C2 were high in the GZ stage, including the genes related to glycosome protein composition, oxidative phosphorylation, and RNA transport. In C3 and C4, a small number of genes was upregulated in the XS stage, whereas a large number of genes in C5 were continuously upregulated in the LK and MF stages. In C6 and C7, the expression levels of the genes involved in substance and energy metabolism were upregulated in the LK and MF stages and reached the peak in the MF stage; these results were consistent with the physiological changes during seed germination (Figure 3).
DEGs and Gene Ontology and KEGG enrichment
We conducted differential expression analysis of the transcription levels of unigenes in four samples to identify potential regulatory genes involved in seed germination. The DEGs were compared in two main ways: (1) using GZ as the reference point, i.e., the fixed reference system (FFS), and (2) by selecting each previous adjacent time point in turn as the reference point, i.e., the continuous comparison system (CCS). In the FFS and CCS groups of C. migao seeds, 43,558 and 40,532 DEGs, respectively, were obtained (Figure 6A). The transcript levels of 26,653 genes significantly increased, whereas those of 8,027 genes decreased in the LK stage (LK versus GZ); in contrast, the transcript levels of 34,133 genes markedly increased, whereas those of 5,021 genes decreased in the LK stage (LK versus XS) (Figure 6A). In the two comparison systems (FFS and CCS), the DEGs of the MF versus GZ and LK versus XS stages accounted for 83.6% and 96.6% of the total DEGs of the two groups; among them, we identified 983 and 54 shared DEGs, respectively (Figure 6B and C). The results showed that the tissue difference between the seeds in the GZ and MF stages was the largest, whereas in the whole germination process, most DEGs start from the XS stage; the difference was most significant from the XS to LK stage. Analysis of the transcriptional expression levels of the genes in seed germination showed that extensive gene expression occurred during germination.
In the CCS group of seed germination, consistent or opposite transcript regulation characteristics were observed. In the FFS group, 4,352 MapMan BINs of the upregulated DEGs and 578 MapMan BINs of the downregulated DEGs were obtained; the upregulated genes accounted for 88.28% of all DEGs. Most DEGs fell into 27 BINs, including RNA, transport, lipid metabolism, protein, carbohydrate, amino acid metabolism, cell, and cell wall (Figure 7). The upregulated genes were mainly involved in RNA/protein biosynthesis, lipid metabolism, cell respiration, and carbohydrate and nutrient absorption (Figure 7A); the downregulated genes were mainly involved in RNA biosynthesis, protein biosynthesis, and solute transport (Figure 7B). In the CCS group, 4,296 MapMan BINs of the upregulated DEGs and 481 MapMan BINs of the downregulated DEGs were obtained; meanwhile, the classification results of DEGs in the CCS system were similar to those of the FFS group (Figure 7C and D).
Because the functional classification generated by MapMan is limited to those genes that have putative Arabidopsis homologs that have been functionally characterized, we have also utilized GO and KEGG annotations to provide further evidence of C. migao functional specialization. Because of the similarity between FFS and CCS, the enrichment analysis of GO and KEGG in DEGs was mainly performed in the CCS group. For GO enrichment, the DEGs in the CCS of XS versus GZ, LK versus XS, and LK versus MF stages were significantly enriched into 48, 81, and 137 GO terms, respectively, which belong to three major functional categories: cell component (C.C.), molecular function (M.F.) and biological process (B.P.). In the LK versus XS stages, the functional categories of the upregulated genes in the LK stage were remarkably enriched in 22 B.P. groups, including cellular process (GO:0009987) and biological regulation (GO:0065007); significantly concentrated in 16 C.C. groups, including cell (GO:0005623), cell part (GO:0044464), and organelle (GO:0043226); and mainly enriched in 11 M.F. groups, including catalytic activity (GO:0003824) and transporter activity (GO:0005215) (Figure 8).
For KEGG enrichment, the DEGs in XS versus GZ, LK versus XS, and LK versus MF stages were significantly enriched into 86, 132, and 114 metabolic pathways. The upregulated genes in the XS stage (XS versus GZ) were concentrated in protein processing in the endoplasmic reticulum and in cysteine and methionine metabolism (Figure 9A); the downregulated genes in the XS stage were enriched in the ribosome, fatty acid elongation, and plant pathogen interaction (Figure 9B). The upregulated genes in the LK stage (LK versus XS) were mainly concentrated in glycerol metabolism, biosynthesis of amino acids, pentose phosphate pathway (PPP), and TCA (Figure 9C); the downregulated genes in the LK stage were concentrated in endoplasmic reticulum protein processing, ABC transporter, and plant hormones (Figure 9D). The upregulated genes in the MF stage (MF versus LK) were mainly enriched in glyceride metabolism, starch and sucrose metabolism, and glycolysis/gluconeogenesis (Figure 9E); the downregulated genes in the MF stage were concentrated in fatty acid biosynthesis and secondary metabolite biosynthesis (Figure 9F). These results show that many categories have certain characteristics in different stages of germination. Our data provide a unique global view of the gene expression related to material and energy metabolism in C. migao seed germination.
Triacylglycerol (TAG) metabolism in seed germination
The oil content varied with seed germination, and the highest oil content was found in dry seeds (Figure 3D). The KEGG annotations indicated that 97 DEGs were annotated to glycerolipid metabolism (ko00561), and 160 DEGs were annotated to fatty acid metabolism during seed germination; among them (Figure 10A and B), 63 were annotated to fatty acid degradation (ko00071) (Figure 10C) in XS versus LK.
The DEGs that were annotated to triacylglycerol (oils) decomposition, unigenes encoding triacylglycerol lipase (LIP), acylglycerol lipase (MGLL), glycerol-3-phosphate O-acyltransferase (GPAT), 1-acyl-sn-glycerol-3-phosphate acyltransferase (plsC), phosphatidate phosphatase, phosphatidate phosphatase, alcohol dehydrogenase (ADH), and aldehyde dehydrogenase (ALDH) were more highly expressed in LK and MF seeds (Fig 11A). Moreover, in DEGs that underwent fatty acid degradation, the unigenes encoding long-chain acyl-CoA synthetase (ACSL), ACOX, ACADM, MPF2, ACAA1, and atoB were obviously upregulated in LK and MF seeds compared with those in GZ and XS seeds (Figure 10D). Hence, the expressions of MGLL, ACSL, and ACOX were upregulated ≥10-fold.
Starch and sucrose metabolism in seed germination
During the germination process, the starch content of C. migao showed a downward trend, whereas the change trend of SSC showed the inverse (Figure 3). Of the 135 DEGs that were annotated to starch and sucrose metabolism (ko00500), most genes were differentially expressed in the LK versus XS stages, and these DEGs maintained higher expression levels during the LK and MF stages (Figure 11A).
The expressions levels of unigenes encoding isoamylase (ISA), α-amylase (AMY), maltase-glucoamylase, 4-α-glucosyltransferase, invertase (Inv), hexokinase (HK), fructokinase (FK), glucose-6-phosphate isomerase (GPI), glycogen phosphorylase (PYG), and phosphoglucomutase (PGM) were significantly upregulated during the LK and MF stages, and their upregulated expressions accelerated the decomposition of starch, maltose, and sucrose in the LK and MF stages. There were still some key genes in starch and sucrose metabolism that were upregulated in the LK and MF stages, including unigenes encoding sucrose synthase (SS), sucrose phosphate synthase (SPS), endoglucanase, β-glucosidase, glucanendo-1,3-β-D-glucosidase endoglucanase, glucose-1-phosphate adenylate transferase (glgC), starch synthase, trehalose 6-phosphatesynthase, and trehalose 6-phosphatephosphatase. Among them, the expression levels of ISA, GPI, SS, and FK increased 7- to 16-fold, and the expression levels of these genes were higher in the MF stage than in the LK stage (Figure 11B).
Energy supply during the germination of seeds
The lipids, starches, and sugars in the seeds decomposed gradually via the germination of seeds and then passed through the PPP (ko00030), glycolysis pathway (EMP) (ko00010), pyruvate metabolism (ko00620), TCA cycle (ko00020), and oxidative phosphorylation (ko00190), although they were not completely decomposed. The decomposed products of the macromolecular substances in seeds provide energy for seed germination by changing the expression levels of key genes in these metabolic pathways (Figure 12A-D).
Glucose was decomposed into D-glyceraldehyde-3P and β-D-fructose-6P in EMP. Of the 49 DEGs annotated to PPP, the unigenes encoding glucose-6-phosphate dehydrogenase (G6PD), phosphor gluconolactonase (PGLS), 6-phosphogluconate dehydrogenase (PGD), ribulose phosphate 3-epimerase, GPI, diphosphate dependent phosphor fructokinase (pfp), 6-phosphofructokinase 1 (pfkA), and fructose bisphosphate aldolase (ALDO) were significantly upregulated in the LK and MF stages (Figure 12A). The expressions of unigenes G6PD, PGLS and pfp in the LK stage were 8- to 10-fold higher than those in the XS stage (Figure 13A). The changes in the expression of the above genes accelerated the decomposition of glucose in the seeds in the LK stage. Of the 32 DEGs annotated to EMP, the unigenes encoding glyceraldehyde 3-phosphate dehydrogenase (GADPH), phosphoglycerate kinase (PGK), phosphoglycerate mutase (pgmI), enolase, and pyruvate kinase (PK) were upregulated in the LK stage, and the expressions of GADPH, pgmI, andPK increased by 9- to 14-fold, which was conducive to the decomposition of D-glyceraldehyde-3P (Figure 13A).
Of the 37 DEGs that were annotated to the pyruvate metabolic pathway, the unigenes encoding pyruvate dehydrogenase E1, E2 (aceE, aceF), dihydrolipoamide dehydrogenase (DLD), malate synthase, malate dehydrogenase (MDH), pyruvate decarboxylase (PDC), acetaldehyde dehydrogenase (ALDH), and acetyl-CoA synthetase (ACSS) were upregulated in the LK stage, and the expressions of MDH, maeB and ALDH increased by 8- to 13-fold (Figure 12C). Of the 26 DEGs that were annotated to the TCA cycle, unigenes encoding citrate synthase (CS), ATP citrate (pro-S)-lyase (ACLY), aconitate hydratase (ACO), isocitrate dehydrogenase (IDH), succinyl-CoA synthetase (LSC), succinate dehydrogenase (SDH), and fumarate hydratase (fum) were upregulated in the LK stage (Figure 12D). In addition, aceE/F, CS, and IDH were the rate-limiting enzymes throughout this process. The expression levels of ACLY, ACO, CS, IDH, LSC, and fum in the LK and MF stages increased by 6- to 10-fold, which increased the synthesis and decomposition of metabolites in each link, increased the speed of cyclic reaction, accelerated the adjustment of the ATP and NADH concentrations in the LK stage, and ensured the supply of energy required for C. migao seed germination (Figure 13B).
Of the 92 DEGs that were annotated to oxidative phosphorylation (Figure 14A), the unigenes encoding NADH dehydrogenase (NDUF, NDHA, NDHB, NDHF), succinate deaminase, cytochrome c reductase, cytochrome c oxidase, and transporting ATPase were upregulated in the LK stage, which accelerated the electron, H+ transfer, and oxygen utilization efficiency so that the seeds generated adequate energy for seed germination in the LK and MF stages (Figure 14B).
Antioxidant capacity of C. migao seeds during germination
Seeds could activate their defense mechanism in response to biotic and abiotic stress during germination. The annotation of DEGs indicated that the enzymes (SOD, CAT, and POD) scavenging macromolecular reactive oxygen species (ROS) were mainly synthesized by peroxisome (ko04146) and phenylpropanoid biosynthesis (ko00940). During seed germination, nine unigenes encoding CAT and SOD were upregulated in the LK stage. The unigenes of 111 DEGs that were annotated to phenylpropanoid biosynthesis (Figure 15), encoded phenylalanine ammonia-lyase (PAL), trans-cinnamate 4-monooxygenase (CYP), 4-coumarate-CoA ligase (4CL), cinnamyl-alcohol dehydrogenase, caffeic acid 3-O-methyltransferase (COMT), ferulate-5-hydroxylase (F5H), and POD. It should be noted that the expression of POD increased by 17-fold, and the other genes increased by 6 to 13.8 times during the LK stage; however, the expressions of PAL, HCT, CoMT and POD were still upregulated during the MF stage, and 12 unigenes encoding POD were upregulated and increased by 6.2-fold, which was consistent with the change in the antioxidant enzyme activity during the germination process (Figure 3F).
Metabolic pathway network analysis
According to the interaction between the KEGG enrichment pathways of DEGs in the seeds of C. migao (Table S3), a metabolic pathway network was established, which elucidated the interaction between important pathways and determined the activation state between pathways (Figure 16). Based on the degree of the seeds, the energy-metabolism-related pathways, TCA, glycolysis/gluconeogenesis, and pyruvate metabolism, were determined to be the center of the network (Figure 16). The highest degree in the network was attributed to TAC, and the expressions of 31 unigenes during the TCA cycle were upregulated during the LK stage, which played a key role in providing seed germination energy. The degree of glycolysis/gluconeogenesis was the second highest, and the expressions of 22 unigenes were upregulated during the LK stage. Pyruvate played a connecting role as an important intermediate product in the metabolic network; therefore, the pyruvate metabolic pathway was critical for seed germination, which is involved in lipid biosynthesis, the TCA cycle, and other target metabolic pathways by converting pyruvate to acetyl-CoA and oxaloacetate; meanwhile, 38 unigenes were upregulated during the LK stage (Figure 13A). At the same time, 49 unigenes in the PPP were upregulated during the LK and MF stages (Figure 13B), which promoted the decomposition of glucose into D-glyceraldehyde-3P and β-D-fructose-6P in the glycolysis pathway. The metabolism of storage materials (glycerolipid, sucrose, starch metabolism, etc.); amino acid metabolism (tyrosine metabolism, arginine and proline metabolism, lysine degradation, etc.); and intermediates and energy supply pathways (glycolysis, the TCA cycle, PPP, etc.) were systematically triggered in C. migao seed germination.