3.1 Comparison of Morphological Phenotypes from S. miltiorrhiza Purple and Green Tissues
There are significant differences in morphological phenotype between green and purple stems in S. miltiorrhiza. Purple plants have better agronomic phenotype, high plant, thick stems, and purple stem skin has more villi. In addition, the color of buds and leaves on purple plants is also different from that of old leaves (Fig. 1A). The edges of young buds and leaves are purple, and the edge of another part of the old leaves is also purple. For the comparison of the leaves with purple edges and completely green leaves, it was found that the leaves with purple edges is yellowish green, the sharp serrated edges are obvious, and some of their stem is purple. In addition, there were more fluff on the back (Fig. 1B). According to the microscopic observation, the pigment of purple stem is mainly accumulated in several layers of cells below the epidermis (Fig. 1C). Therefore, some stem skin tissues are observed by hand section, and the edge tissues of green and purple leaves are compared. It was found that purple pigment is mainly in the epidermal cells at the edge of leaves (Fig. 1D). Therefore, we established a dynamic accumulation model of stems and leaves biosynthesis pathway of S. miltiorrhiza using transcriptome and metabolome, and explored the biological transport pathway leading to the accumulation and transportation of purple pigment (Fig. 1).
3.2 Identification of Differentially Expressed Genes from S. miltiorrhiza Purple and Green Tissues
A total of 76.95 GB data were measured using DNBSEQ sequencing platform. The clean data of each sample reached 6.3 GB, the Q20 value of the obtained sequence was greater than 96%, the percentage of Q30 base was ≥ 90%, and the GC content was more than 43% (Table S1). A total of 40077 SSR loci were obtained. 19216 unigene sequences included 93 repeat units, of which di-nucleotide repeat was the dominant repeat type (55.27%), followed by mono-nucleotide repeat and tri-nucleotide repeat, with the proportion of 15.5% and 23.1% respectively. A/T and AG/TC were the dominant units of mono-nucleotide and di-nucleotide repeat respectively (Table 1). 114431 unigenes were successfully annotated by comparison with the reference database, 92326 (80.68%) genes were annotated in NR database, of which 60704 (65.75%) genes were consistent with the characteristics of Salvia splendens, and 3566 genes were unique to S. miltiorrhiza. These results showed that the quality of sequencing data was highly, it can meet the needs of transcriptome analysis.
Table 1
Transcriptome SSR repeat type and number
Repeats motifs
|
Duplicate Times
|
Total
|
Ratio (%)
|
4
|
5
|
6
|
7
|
8
|
9
|
10
|
> 10
|
Mono-nucleotide
|
-
|
-
|
-
|
-
|
-
|
-
|
1445
|
6198
|
6,198
|
15.5
|
Di-nucleotide
|
-
|
-
|
6,791
|
4,410
|
3,540
|
2,261
|
1,468
|
3679
|
22,149
|
55.27
|
Tri-nucleotide
|
-
|
4,980
|
2,171
|
1013
|
607
|
141
|
93
|
214
|
9219
|
23.1
|
Quad-nucleotide
|
-
|
205
|
53
|
8
|
5
|
12
|
3
|
6
|
292
|
0.73
|
Penta-nucleotide
|
558
|
172
|
30
|
8
|
7
|
2
|
6
|
1
|
784
|
1.96
|
Hexa-nucleotide
|
1088
|
196
|
95
|
39
|
10
|
3
|
2
|
2
|
1435
|
3.58
|
Total
|
1676
|
5553
|
9140
|
5478
|
4169
|
2419
|
1572
|
10100
|
40077
|
-
|
Ratio(%)
|
4.18
|
13.9
|
22.8
|
13.7
|
10.4
|
6.04
|
3.92
|
25.2
|
-
|
-
|
According to qvalue ≤0.05 and log2 (Fold-change) ≥ 1. 1299 DEGs were identified in S1-S2, including 666 up-regulated genes and 633 down-regulated genes. 1091 DEGs were identified between L1-L2, of which 393 were up-regulated and 698 were down regulated (Fig. S1). GO and KEGG pathway enrichment analysis were performed on DEGs. In GO analysis, 481, 572, and 805 DEGs were assigned to the biological process, molecular function, and cellular component terms, and further classified into 35 functional subcategories in the S1-S2 group (Fig. 2A). In the L1-L2 group, 402, 469, and 676 DEGs were involved in three major categories: biological processes, cellular components, and molecular functions, and further classified into 38 functional subcategories (Fig. 2B).
To understand their biological functions and gene interactions, DEGs were annotated by the KEGG database. In the S1-S2 group (Fig. 3A), 473 out of 1299 DEGs were located in 122 KEGG pathways including cellular processes, environmental information processing, genetic information processing, metabolism, and organismal systems. In the L1-L2 group (Fig. 3B), 438 of 1091 DEGs were located in 113 KEGG pathways. The main enriched metabolic process was phenylpropanoid biosynthesis in S1-S2 group, and there were twelve DEGs involved in flavonoid biosynthesis. In further analysis of the DEGs, we also detected six GST genes, nine ABC transporter genes, 22 MATE genes, and three SNARE genes, which might play important roles in transporting anthocyanin to plant vacuoles (Jiang et al.,2020). The main enriched metabolic process was phenylpropanoid biosynthesis and carbon metabolism in L1-L2 group. Further excavations revealed eight GST genes and seven ABC transporters.
The main components that affect the color of petals, stems, leaves, and fruits are anthocyanins and flavonoids secondary metabolites (Deng et al.,2020; Jiang et al.,2020). In the process of enrichment analysis, In the process of enrichment analysis, we detected PAL, C4H/CYP73A, 4CL, DFR, ANS, FLS, CHI, F3H, and 3AT was significant difference in flavonoid biosynthesis pathway. HCA analysis was performed on DEGs with pheatmap package (Fig. 4, Table S2).
3.3 Identification of Transcription Factor that Regulates Flavonoids Metabolism
The content of flavonoids in plants is strictly regulated by complex regulatory networks. Transcription factors (TFs) include members of MYB, bHLH, MADS, WRKY, bZiP, and WD40 families were play an important role in regulating plant flavonoid biosynthesis (Luo et al.,2019). In this study, the differential expression of MYB, bHLH, bZiP, and WRKY was detected. It is worth mentioning that the MYB has the largest number and difference among all TF genes detected. The differential expression level of these gene was most obvious in stem tissue (Table S3, Fig. 5). MYC2 and WRKY34 were expressed higher levels in green stem compared with purple stem. These findings uncover that these transcription factors were closely related to the color phenotypic differences in different tissues.
3.4 Identification of Differentially Accumulated Metabolites
After data preprocessing, 10636 metabolites were identified in positive ion mode, and 5231 metabolites were identified in negative ion mode in leaves. 6264 metabolites were identified in positive ion mode, 3503 metabolites were identified in negative ion mode in stems (Table 2). Flavonoids, lipids, and sugars without nitrogen are generally more applied in the positive ion mode, and the negative ion mode is one order of magnitude smaller than the positive ion mode, and we get more metabolites in the positive ion mode. Therefore, we choose to analyze the metabolites obtained in the positive ion mode.
Table 2
Statistical table of compound detection
|
mode
|
Number of metabolites
|
Number of metabolites with identification information
|
stem
|
pos
|
6264
|
816
|
neg
|
3503
|
466
|
leaf
|
pos
|
10636
|
944
|
neg
|
5231
|
524
|
Note: pos: The positive ion mode means that the adduct ions are (H+, NH4+, Na+, K+) when the compound is ionized at the ion source. Neg: The negative ion mode means that the adduct ions are (-H, +Cl) plasma when the compound is ionized at the ion source. |
The metabolites identified in S. miltiorrhiza leaves were classified in BGI library. There are 44 metabolites belonging to flavonoids, 50 metabolites belonging to terpenoids, 37 metabolites belonging to phenylpropanoids, 14 metabolites belonging to phenolic acids, 15 metabolites belonging to phenols, 13 metabolites belonging to miscellaneous and 3 metabolites belonging to steroids, 11 metabolites belong to alkaloids (alkaloids and glycosides), and 2 metabolites belong to quinones (quinones and glycosides). Similarly, the metabolites identified in the stem tissue of S. miltiorrhiza were classified in BGI library. It was found that there were 45 metabolites belonging to flavonoids and glycosides, 42 metabolites belonging to terpenes and glycosides, 35 metabolites belonging to phenylpropanoids and glycosides, 18 metabolites belonging to phenolic acids and glycosides, 14 metabolites belonging to phenols and glycosides, 12 metabolites belonging to other classes, and 1 metabolite belonging to steroids and glycosides, 11 metabolites belong to alkaloids and glycosides, and 2 metabolites belong to quinones and glycosides (Fig. 6). It can be seen that flavonoids and glycosides, terpenes and glycosides, phenylpropanoids and glycosides are the main metabolites identified in all samples of above ground tissue of S. miltiorrhiza.
The identified metabolites were classified and annotated in the HMDB and KEGG (Wishart et al.,2018). In the leaves sample, all metabolites were divided into four categories: compounds with biological roles (31.42%), phytochemical compounds (41.09%), others (11.18%), and lipids (16.31%). Flavonoids and terpenoids account for a large proportion of phytochemical compounds, accounting for 37.5% and 36.03%, respectively. In the stems sample, terpenoids (32.8%), flavonoids (32.22%), and phenylpropanoids (15.56%) were still the top three compounds in phytochemistry (Fig. 7).
The DAMs between green and purple cultivar were determined according to the fold-change ≥ 1.2 or ≤ 0.83 and Pvalue < 0.05. In green stems and purple stems, a total of 23 DAMs were screened, of which 7 were up-regulated and 16 were down regulated (Table S4 and Fig. 8A). In green leaves and leaves with purple edges, a total of 34 DAMs were screened, of which 17 were up-regulated and 17 were down regulated (Table S5 and Fig. 8B). Then, HCA was performed on the accumulation pattern of metabolites among different samples (Caesar et al.,2018). The HCA of detected metabolites mainly reveals two clusters, one is the metabolites before the color change (S1, L1), and the other is the metabolites after the color change (S2, L2) (Fig. S2).
In order to better understand the relationship between various DAMs, the metabolic pathways of all DAMs were enriched and analyzed based on KEGG database. The results showed that 48 detected metabolites were enriched in 46 metabolic pathways, and 36 detected metabolites were enriched in the biosynthesis of secondary metabolites. The pathway with significant enrichment of Pvalue< 0.05 has been selected. In the stems sample, the results showed that there were seven DAMs annotated in the metabolic pathway (Guanosine, Aflatoxin G2, Estriol, Arachidonic acid, 5'-S-Methyl-5'-thioadenosine, 4-coumaric acid, and Sinapic acid), and a total of four DAMs in the biosynthesis pathway of secondary metabolites were annotated (Coumarin, 4-Coumaric acid, Sinapic acid, and Phlorizin) (Table 3). In the leaves sample, bubble plots of 14 pathways were drawn for the significant enrichment of DAMs with Pvalue < 0.05, the results showed that the proportion of DAMs annotated to pentose phosphate pathway was the largest. The number of DAMs annotated to ABC transporters was the largest, accounting for 17.65% (D- (+) -glucose, biotin, D- (+) -xylose), D- (+) –glucose, and D- (+) -xylose has involved in pentose phosphate pathway and antibiotic biosynthesis. There is a differential metabolite (D- (-) -quinic acid) involved in the metabolic pathway of phenylalanine, tyrosine and tryptophan biosynthesis. D- (+) -glucose was involved in almost all other metabolic pathways (Fig. 9).
Table 3
Metabolic pathway enrichment with Pvalue < 0.05 in S1-S2
Pathway
|
Count
|
Pvalue
|
DAMs Names
|
Metabolic pathways
|
7
|
0.0018
|
Guanosine, Aflatoxin G2, Estriol, Arachidonic acid, 5'-S-Methyl-5'-thioadenosine, 4-Coumaric acid, Sinapic acid
|
Biosynthesis of secondary metabolites
|
4
|
0.0113
|
Coumarin, 4-Coumaric acid, Sinapic acid, Phlorizin
|
Cysteine and methionine metabolism
|
1
|
0.0463
|
5'-S-Methyl-5'-thioadenosine
|
Arachidonic acid metabolism
|
1
|
0.0549
|
Arachidonic acid
|
Purine metabolism
|
1
|
0.069
|
Guanosine
|
ABC transporters
|
1
|
0.098
|
Guanosine
|
Degradation of aromatic compounds
|
1
|
0.1804
|
4-Coumaric acid
|
Biosynthesis of antibiotics
|
1
|
0.5483
|
Aflatoxin G2
|
Note: S1(green stem), S2 (purple stem) |
3.5 The Metabolome and Transcriptome are Coregulated the Color Phenotypic in Different S. miltiorrhiza Tissues Phenotypes
To further correlate gene expression patterns with metabolite accumulation, the co-expression analysis was applied to metabolome and transcriptome data from S. miltiorrhiza purple and green stems and leaves. This work showed that the expression levels of PAL, 3AT, 4CL, ANS, HCT, MYB43, MYB101, MYB108, MYB94, WRKY24, and WRKY48 were higher in purple samples, however, DFR, CYP73A were higher in green samples (Table S2, and S3). It is speculated that these key genes and transcription factors were involved in the regulation of flavonoid and phenolic acid synthesis, thereby affecting the color of S. miltiorrhiza stem and leaf. And the reason for anthocyanin accumulation in purple characters is closely related to the high expression of these synthase involved in anthocyanin biology.
In the main pathway of phenylpropanoid and flavone biosynthesis, the strong activation and expression of all structural genes eventually lead to the production of a large number of anthocyanins. In the flavonoid biosynthesis pathway, under the catalysis of MYB, bHLH, MADS, WRKY, and bZiP families, the yields of flavones and anthocyanins are also increasing on a large scale. Anthocyanins are a group of water-soluble pigments, which help to form various colors of plants. Flavonoids, especially anthocyanin biosynthesis pathway, lead to the accumulation of purple pigment (Fig. 10).
The transcriptomic results showed that there were great differences in the structural genes controlling flavonoids in the stems and leaves of S.miltiorrhiza. What’s more, metabolomic results showed that the content of rhusflavanone, herbacetin, ipriflavone (leaves) as well as isosakuranin (stems) were higher in purple phenotype than green phenotype. Procyanidin B2, loureirin A, astilbin, and safflomin A were higher in green stem phenotype. Thus, the content of flavonoids also varied greatly in stems and leaves. In S. miltiorrhiza leaves, the expression level of HCT and the content of flavonoids (rhusflavanone, herbacetin, and ipriflavone) in purple phenotype were significantly higher than green phenotype. This may relate to the higher expression of HCT can inhibit the synthesis of lignin, leading to the redirection of the metabolic flux into flavonoids through chalcone synthase activity.
3.6 Distribution Dynamics of Flavonoids by LSCM
To investigate positioning within the organization of flavonoids in the purple and green characters from S.miltiorrhiza, the frozen sections were prepared and observed. At the excitation wavelength of 488 nm, the samples showed chloroplast autofluorescence before staining. This interference is eliminated by subtracting the background signal. After staining with NA solution, the epidermis, glandular hair, and thick horn tissue cells of stems in S. miltiorrhiza were shown strong fluorescence. In addition, A small amount of fluorescence also appeared in phloem. All of these suggests that flavonoids synthesized and accumulated in epidermis, villi, collenchyma, phloem, and fascicle S. miltiorrhiza (Fig. 11A-D). Flavonoids in S. miltiorrhiza leaves are mainly distributed in external tissues such as glandular hair and epidermis (Fig. 11E-H), which may be related to preventing microbial invasion and UV radiation damage, and also accumulate in palisade tissue and spongy tissue (Ferreyra et al.,2021). Because methanol is volatile and has strong solubility in flavonoids, as the solvent of color developing agent, flavonoids will dissolve and diffuse in the dye process, and a small amount of fluorescence will appear in some parts. The content of flavonoids in green phenotype was significantly higher than purple phenotype, which may be because the lower expression level of ANS results in a decrease in the anthocyanin synthesis and accumulation of green samples but an increase in flavonoid synthesis and accumulation.