Title Page Title: The role of DNA methylation in the accumulation of iridoid glycosides in Rehmannia glutinosa


 Background: Epigenetic regulation plays a significant role in the accumulation of plant secondary metabolites. The terpenoids are the most abundant in the secondary metabolites of plants, iridoid glycosides belong to monoterpenoids which is one of the main medicinal components of R.glutinosa. At present, study on iridoid glycosides mainly focuses on its pharmacology, accumulation and distribution, while the mechanism of its biosynthesis and the relationship between DNA methylation and plant terpene biosynthesis are seldom reports. Results: The research showed that the expression of DXS, DXR, 10HGO, G10H, GPPS and accumulation of iridoid glycosides increased at first and then decreased with the maturity of R.glutinosa, and under different concentrations of 5-azaC, the expression of DXS, DXR, 10HGO, G10H, GPPS and the accumulation of total iridoid glycosides were promoted, the promotion effect of low concentration (15μM-50μM) was more significant, the content of genomic DNA 5mC decreased significantly, the DNA methylation status of R.glutinosa genomes was also changed. DNA demethylation promoted gene expression and increased the accumulation of iridoid glycosides, but excessive demethylation inhibited gene expression and decreased the accumulation of iridoid glycosides. Conclusion: The analysis of DNA methylation, gene expression, and accumulation of iridoid glycoside provides insights into accumulation of terpenoids in R.glutinosa and lays a foundation for future studies on the effects of epigenetics on the synthesis of secondary metabolites.

DNA methylation, in combination with histone modi-cations and non-histone proteins, de nes chromatin structure and accessibility. DNA methylation therefore helps to regulate gene expression, transposon silencing, chromosome interactions and trait inheritance [27].DNA methylation affects the accumulation of plant secondary metabolites and the expression of related enzyme genes in the secondary metabolism pathway.
Transgenic tobacco plants overexpressing Arabidopsis thaliana demethylase gene ROSI signi cantly increased the expression of avonoid metabolism and antioxidant pathway genes, thus improving the salt tolerance of transgenic tobacco [28]. In transgenic hybrid poplars overexpressing chestnut demethylase gene, the expression of avonoid biosynthesis was promoted, the accumulation of avonoids in apical meristem and the maturation of scale bud and apical bud were also promoted [29]. After NaCl stress, 12 genes were signi cantly enriched by hypermethylation in the leaves of tetraploid watermelon seedlings, involving the biosynthesis of cork grease, wax and monoterpenes and transcriptome results showed that diploid watermelon roots were regulated by DNA demethylation and 8 genes in linoleic acid metabolism pathway were up-regulated [30]. Methylation inhibitors affect the expression of related genes in the metabolic pathway, which in turn affects the accumulation of secondary metabolites. 5azaC can activate energy metabolism pathway, pentose phosphate pathway, fatty acid metabolism pathway and other biological pathways and these pathways can provide material and energy basis for paclitaxel synthesis, such as NADPH, energy, acetyl-CoA, Glyceraldehyde triphosphate,at the same time, 5azaC stimulates the expression of key enzyme genes in paclitaxel synthesis pathway, and paclitaxel content increases continuously with the decrease of methylation level [31,32]. The genes in differential methylation region between the roots of Salvia miltiorrhiza at two stages were enriched in the diterpene biosynthesis pathway, 5azaC treatment of Salvia miltiorrhiza hairy roots growing for two months could signi cantly increase the biosynthesis of tanshinones, indicating that DNA methylation was involved in the biosynthesis of tanshinones [33,34]. The expression of key enzyme genes (hmgr,sqs,se and ls) in Ganoderma lucidum biosynthesis was increased after 5azaC treatment, and the accumulation of ganoderic acid in Ganoderma lucidum was also increased, even 5azaC treatment decreased the level of genomic DNA methylation, it did not change the methylation level of sqs,ls and se gene promoter region [35]. NO increased the level of DNA methylation in cultured Robinia pseudoacacia cells, and the change trend was dynamically related to iso avone synthesis and 5azaC blocked iso avone synthesis induced by nitric oxide [36]. The contents of polysaccharides and alkaloids in Dendrobium increased signi cantly under 5azaC treatment [37]. To sum up, DNA methylation is necessary to maintain normal plant growth, and plays an indispensable regulatory role in plant gene expression, cell differentiation, phylogeny and accumulation of secondary metabolites.
Although some progress has been made in the accumulation of secondary metabolites by DNA methylation in plants, it is still mainly focused on the relationship between total DNA methylation level and the content of plant secondary metabolites, and lack of molecular mechanism of DNA methylation in the accumulation of secondary metabolites, especially in R.glutinosa, there are few studies on the relationship between DNA methylation and iridoid glycoside accumulation, and the effect of DNA methylation on the expression of key enzyme genes in iridoid glycoside synthesis pathway. In view of the close relationship between plant secondary metabolism and epigenetics, the purpose of this study is to reveal the changes of gene expression and secondary metabolites at the epigenetic level and hope to explore the mechanism of the formation of plant secondary metabolites.

Plant materials
In this study, R. glutinosa cultivar (85 − 5,Jinjiu) was chosen, and its root tubers were kindly provided by Agricultural Research Institute of Wenxian County, Henan, China. Primers were synthesized by YingjieJi Trade Co., Ltd.,(Shanghai, China) and their sequences are listed in Table S1.
In addition, LYK and LCH, agronomists of Agricultural Research Institute of Wenxian County, have formally identi ed the experimental materials Rehmannia glutinosa 85 − 5 and Jinjiu, and these experimental materials have not been deposited in a publicly available herbarium.

Treatment and Cultivation of R. glutinosa
After sterilized for 10 min by 0.1% HgCl 2 , root tubers of R.glutinosa were washed for 3-5 min with sterile water, and were wrapped with gauze, then were respectively treated for 7d (3 times/d) with 0, 15, 30, 50, 100 or 250 µM 5-azaC. Subsequently, root tubers treated with 5-azaC were grown in test eld, College of Life Science, Henan Normal University, Xinxiang City, Henan Province, China. In addition, 100 sterilized roots of R.glutinosa were treated with 5-azaC in each group, and there were three replicates per group. Tuberous root and leaf of R.glutinosa at each growth stage were frozen in liquid nitrogen and stored at -80℃.
Genomic DNA extraction of R. glutinosa Genomic DNA was extracted from root and leaf of R.glutinosa by CTAB method with modi cations [38]. About 4 g tuberous root or leaf was put into the pre-cooled mortar and ground into the powder, then immediately transferred to 50mL centrifuge tube. In the centrifuge tube, 5mL/g CTAB solution containing two-thousandths of β-mercaptoethanol was added, and put in water bath at 65°C for 2 h accompanied by upside down mixing 1 times/20 min,centrifuged for 10 min at 12000rpm. Afterwards, the supernatant was transferred, extracted by phenol:chloroform:isoamyl alcohol (25:24:1) and centrifuged for 10 min at 12000 rpm. The above process was repeated once, and then the supernatant with isopropyl alcohol was placed for more than 1h and separated by 12000rpm centrifugation for 15 min. Precipitated DNA was washed twice with 70% ethanol, and dissolved in double-distilled water after air dried for at least 5 min, then 1% RNase was added to DNA solution; put at 37°C for 30min and extracted with phenol:chloroform:isoamyl alcohol (25:24:1). After further extraction, DNA solution was precipitated by adding 1/3 volume of 3 M NaAc and 2.5 volumes of absolute ethanol, and centrifuged at 12000 rpm for 10 min, then precipitated DNA was washed twice with 70% ethanol and centrifuged at 12000 rpm for 5 min.Subsequently, dried DNA was dissolved in double-distilled water and was stored at -20°C to analyze genomic DNA methylation.
In addition, the integrity of genomic DNA was detected by 0.8% agarose gel electrophoresis, and the yield and purity of genomic DNA were respectively determined by spectrophotometry at 260nm.
Total RNA extraction and rst-strand cDNA preparation Total RNA was extracted from root and leaf samples using an RNA kit (Bio-Connect, Huissen, Netherlands). RNA concentration and purity were estimated spectro-photometrically based on absorbance at 260 and 280 nm.
First-strand cDNA was synthesized from 1 µg total RNA,which was reversely transcribed using M-MLV reversem transcriptase with oligo (dT) 15 as a primer in a 10µL reaction volume (Takara, Japan).
Cloning of DXS,DXR,10HGO,G10H and GPPS The sequences of DXS, DXR, 10HGO, G10H and GPPS genes of R.glutinosa were obtained by electronic cloning. The partial sequences of the genes obtained by transcriptome sequencing were used as probes, and the online database of R.glutinosa SRA (SequenceReadArchive) sequence in NCBI was searched by Blastn tool. The SRA sequence of R.glutinosa which has the same homology with the probe sequence was obtained, which was compared, spliced and extended by DNAMAN7.0, and then the extended overlap group (Contig) was used as the probe, then Blastn retrieval is carried out through the SRA database until there is no new SRA sequence to be spliced and the overlap group can not be extended.
According to the sequence obtained by electronic cloning technology, the corresponding primers were designed and PCR ampli cation was carried out with R.glutinosa cDNA as template. The PCR ampli cation system was as follows: template 2.0µL, Primer-F(10µM) 1.0µL, Primer-R(10µM) 1.0µl, 2xTaqMix 10µL, sterilized ddH 2 O 6.0µL. PCR ampli cation program is: 95℃, 3min; 95℃ 30 s, annealing 30s (DXS 52℃, DXR 50℃, 10HGO 50℃, G10H 54℃, GPPS 55℃), 72℃ 2min, a total of 34 cycles, 72℃ extension 10min. The PCR product was cut and recovered, and then the target fragment was connected with the T vector, and the ligated product was transformed into Escherichia coli. The PCR product was detected by agarose gel electrophoresis, and the bacterial solution of the same size as the target band was sent to the company for sequencing. The bacterial solution with correct sequencing was mixed with 50% glycerol and stored at -20℃.
Quantitative Real-time PCR TIP41 was used as internal reference gene, the primers of TIP41, DXR,DXS,10HGO,G10H and GPPS are qTIP41-F, qTIP41-R, qDXS-F, qDXS-R, qDXR-F, qDXR-R, q10HGO-F, q10HGO-R, qG10H-F, qG10H-R, qGPPS-F, qGPPS-R, respectively.According to the AceQ Qpcr SYBR Green MasterMix kit (Vazyme, Nanjing, China) instructions, the con guration of the reaction system is carried out on the ice, and the whole operation avoids light and avoids the failure of the reagent. The uorescence quantitative PCR reaction was carried out on the LightCycler96 uorescence quantitative PCR instrument. The PCR ampli cation system is as follows: template 2.0µL, Primer- Excel was used for statistics and calculation, and the expression of the target gene in the sample was calculated by 2 −∆∆Ct method. In the later experiment, GraphPad Prism 5.0 software was used, and the data were analyzed by least signi cant difference (LSD) multiple comparison.
Determination of iridoid glycosides in R.glutinosa The fresh roots and leaves of R.glutinosa (cut into thin slices) were put into the oven, dried at 50℃ for 48 hours, and the powder was pressed by a grinder. Accurately weigh 1g of R.glutinosa powder into a 100mL plug conical bottle, add 10 mL of 70% ethanol solution, ultrasonic extraction for 45 minutes, supplement the weight, lter the ltrate (qualitative lter paper and inorganic lter head double ltration). Put the ltrate into the 50mL volumetric ask and shake,accurately measure 2.5mL in the 50mL volumetric ask and x the volume. Accurate weighing of catalpol reference substance 5.76mg was placed in a 25mL volumetric bottle and 70% ethanol constant volume (0.23mg/mL) control solution was added. Catalpol control solution 1.5ml, 2.0ml, 2.5ml, 3.0ml, 3.5ml and 4.0ml were respectively taken and placed in a 10mL volumetric ask, and then 70% ethanol was added for constant volume.Take the control solution, test solution and 70% ethanol (blank control) 1mL into the 10mL plug test tube, add 1mol LHCL 2ml, 90℃ water bath 15min, room temperature 15min cool,add dinitrophenylhydrazine ethanol test solution 0.5ml, 90℃ water bath 25min, room temperature 15min cool, add 1molL NaOH 70% ethanol solution 30ml, place at room temperature for 1 hour. With 70% ethanol as a blank, the absorbance was determined at 463nm. The data were sorted out by Excel and SPSS software, and multiple comparisons were made by Duncan method in the analysis of variance.

HPLC Detection of Genomic Methylation
The content of 5-methyl cytosine (5mC) usually indicates the level of genomic DNA methylation. In this research,level of DNA methylation was detected by HPLC (high performance liquid chromatography) [38].After genomic DNA of R. glutinosa was orderly hydrolyzed with DNase I, nuclease P1 and alkaline phosphatase, was centrifuged at 12000rpm for 5min, then the supernatant was transferred and was ltered with 0.45µM organic micro ltration membrane, subsequently was detected by HPLC. These chromatography conditions performed in this study were as follows: the mobile phase was composed of 50mM KH 2 PO 4 and 8% methanol (92: 8), the ow velocity was 0.5ml/min, the column temperature was 30°C, the sensitivity was 0.1, the injection volume was 20µl, the analytical column was Agilent C18 Zorbax XDB column (4.6×150mm, 5µm particle size), and the detection wavelength was 285nm.
The standard substance of 0.0044g C and 0.0065g 5mC was separately dissolved with 10ml ultra pure water to 40nM/100µL (mother liquor), and the mother liquor was diluted to 0.02nM/100µl, 0.05nM/100µl, 0.1nM/100µl,0.5nM/100µl, 1nM/100µl, then each concentration of standard substance (20µl) was detected. As shown in HPLC chromatogram, the standard substance of C and 5mC both had a good linear relationship in the range of 0.02-1.0nM/100µl, and the linear equation of C or 5mC as follows: y = 2172.7x-18.537 (R² = 0.9993), y = 2276x-7.9806 (R² = 0.9997), in which x or y represented concentration and peak area of standard substance respectively. According to the above linear equation of standard substance C and 5mC, the concentration of C and 5mC could be respectively calculated, then the content of 5mC was obtained according to the following formula:5mC = [5mC / (5mC + C)] ×100. In addition, precision,repeatability and stability of HPLC were tested to guarantee the reliability of experiment data, and the detection of C and 5mC in genomic DNA was repeated three times.

Methylation-Sensitive Ampli ed Polymorphism (MSAP) Ampli cation
The MSAP experiment was performed with modi cations according to the method of Pieter et.al [39]. Under the condition of without changing methylation status, HpaII and MspI have different sensitivity to DNA methylation.
HpaII is not sensitive to full methylation (double-stranded-methylation) and could cleave hemi-methylation (single-stranded-methylation), MspI is sensitive to the internal cytosine (CmCGG sequence) and not to the external cytosine (mCCGG sequence) in full methylation. Thus,HpaII and MspI were selected to produce different cleavage fragments which can reveal status and extent of genomic DNA methylation. Genomic DNA of R.glutinosa was digested with endonuclease EcoRI/MspӀ, EcoRI/HpaII, in turn, then was ligated with EcoRI adapter and MspI-HpaII adapter by T4 DNA ligase at 16℃ for 15 h, subsequently Enzyme-Ligation product was stored at -20℃.

MSAP Data Analysis
In this research, DNA methylation level was quanti ed by MSAP binary data, the presence or absence of one band was respectively scored as "1" and "0", only clear and reproducible bands were scored after silver staining. Compared with the control, DNA methylation patterns of R. glutinosa treated with 5-azaC were classi ed into methylation polymorphism and methylation monomorphism.DNA methylation monomorphism was regarded Type A,DNA methylation polymorphisms included type B (DNA methylation) and type C (DNA demethylation).

Statistical Analysis
In this research, gene expression, iridoid glycosides content,5mC content and methylation level of R. glutinosa were tested for signi cance level, by using analysis of variance and multiple comparisons of Duncan's multiple range. Gene expression, content of iridoid glycosides ,content of 5mC and level of DNA methyltion were calculated and analyzed by Excel and DPS7.5, and all histograms were drawn by Graphpad Primer 5.0.  prediction shows that the α helix ratio is 24.01-57.82%, and the β-folding ratio is 3.55-8.94%. The prediction of subcellular localization showed that 10HGO was located in the cytoplasm and the others were located in the chloroplast( Table 2).   S2). In addition,build the Phylogenetic tree of DXS, 10HGO, G10H and GPPS, R.glutinosa and Scutellaria barbata, Striga asiatica, Handroanthus impetiginosus, Sesamum indicum,were clustered into one branch, respectively, and DXR branches separately (Fig. S3).

Results
Based on STRING genome, gene co-ocurrence on DXS,DXR,10HGO,G10H,GPPS synthase of R. glutinosa was analyzed, found that they were conservative in organisms, occurred in Archaea, Eukaryota and Bacteria, in which they were more conservative in Viridiplantae. Many genes in Viridiplantae appeared to match DXS,DXR,10HGO,G10H,GPPS synthase of R. glutinosa, gene co-ocurrence for DXS,DXR,10HGO,G10H,GPPS of R. glutinosa, were all found in 19 species of plants, such as Aquilegia coerulea,Citrus clementine,Citrus sinensis,etc. 4 synthase genes were respectively matched from 9 species of plants, 3 synthase genes were respectively matched from 34 species of plants, and only 1-2 synthase genes in 10 species of plants had coocurrence, and mainly belong to GPPS.The co-occurrence ratios of DXS, DXR, 10HGO, G10H and GPPS were 87.5%, 94.4%, 38.9%, 30.6% and 90.3%, respectively (Table S2). Further analysis showed that DXS, DXR, 10HGO, G10H and GPPS had high homology in eukaryotes (Table S3, Fig. S4a,b),it has the highest homology with Erythranthe guttata, and also has high homology with Aquilegia coerulea, Citrus clementina, Citrus sinensis and Gossypium raimondii.

DXS
In stage I-IV, the expression of DXS in Jinju roots was signi cantly higher than that in 85 − 5 roots, then was basically the same at stage V, especially in stage IV, Jinjiu expression level was 2-fold that of 85 − 5. In the stage I-IV, the expression of DXS in Jinjiu leaves was signi cantly higher than that in 85 − 5 leaves, especially in stage IV, Jinjiu expression level was 13-fold that of 85 − 5, but in V stage was lower than 85 − 5 leaves. The overall expression in leaves was higher than that in roots (Fig. 1a).

DXR
In stage I-III and V, the expression of DXR in Jinjiu root was higher than that in 85 − 5 root, but in stage IV was signi cantly lower than that in 85 − 5 root. In stage IV-V(mature stage), DXR in Jinjiu leaf was signi cantly higher than that in 85 − 5 leaf. The expression level in the leaves of stage IV-V(mature stage) was higher than that in the roots (Fig. 1b).

10HGO
In stage II-V, the expression of 10HGO in Jinjiu root was lower than that in 85 − 5, especially in stage V, 85 − 5 expression level was 2-fold that of Jinjiu. In stage IV-V (mature stage), the expression of 10HGO in Jinjiu leaf was signi cantly higher than that in 85 − 5 leaf, especially in stage V, Jinjiu expression level was 2-fold that of 85 − 5 (Fig. 1c).

G10H
In stage I-IV, the expression of G10H in Jinjiu root was higher than that in 85 − 5 roots, especially in stage III, Jinjiu expression level was 10-fold that of 85 − 5. The expression change in leaves was more complex, In stage IV, the expression of G10H in Jinjiu leaves was signi cantly higher than that in 85 − 5 leaves, while in V stage, lower than that in 85 − 5 leaves. The expression level in leaves was signi cantly higher than that in roots (Fig. 1d).

GPPS
In stage II-V, the expression of GPPS in Jinjiu root was lower than 85 − 5, especially in stage I V, 85 − 5 expression level was 3-fold that of Jinjiu. The expression of GPPS in Jinjiu leaf was signi cantly lower than that in 85 − 5 leaf except for stage IV (Fig. 1e).
On the whole, gene expression showed variety speci city, tissue speci city and spatio-temporal speci city, in stage IV-V, genes expression is at a higher level in the whole growth cycle, usually reaching a peak in stage IV or V. The expression of DXS, DXR and G10H genes in Jinjiu roots and leaves was more than 85 − 5, while the expression of 10HGO and GPPS genes in Jinjiu roots and leaves was less than 85 − 5, with the maturity of R.glutinosa, DXS and G10H expression showed a trend of increasing at rst and then decreasing as a whole, and decreased signi cantly in stage V, DXR 10HGO and GPPS expression decreased at rst and then increased, and often continued to increase in stage IV-V.

Clustering heat map analysis
In order to study the expression level of genes in 85 − 5 roots, 85 − 5 leaves, Jinjiu roots and Jinjiu leaves in the same period, the expression of DXS in 85 − 5 roots was used as control, the expression of each gene was calculated and cluster heat map was drawn,so that to study genes clustering expression.
In stage I, the expression of DXS and G10H in leaves was higher than roots, Jinjiu leaves was higher than 85 − 5 leaves, Jinjiu roots was higher than 85 − 5 roots, while the expression of 10HGO, DXR and GPPS in roots was higher than that in leaves (Fig. 2a), cluster analysis showed that the expressions of 10HGO, DXR and GPPS genes were grouped into one group, DXS and G10H were clustered into one group. In stage II, the expression of DXS and G10H in leaves was higher than that in roots, and the expression of DXR, 10HGO and GPPS in 85 − 5 was higher than that in Jinjiu, cluster analysis showed that the expression of 10HGO, DXR and GPPS genes were clustered into one group, and the expression of DXS and G10H were clustered together (Fig. 2b), and the clustering was the same as that of stage I. In stage III, the expression of DXS and G10H in leaves was higher than that in roots, Jinjiu leaves was higher than 85 − 5 leaves, Jinjiu roots was higher than 85 − 5 roots, DXR in Jinjiu was higher than that in 85 − 5, and 10HGO and GPPS in 85 − 5 was higher than that in Jinjiu (Fig. 2c),cluster analysis showed that the expression of DXS, DXR and G10H was grouped into one group, and 10HGO and GPPS were grouped into one group. In stage IV, the expression of DXS, DXR, 10HGO, G10H and GPPS in leaves was higher than that in roots, and Jinjiu leaves was higher than 85 − 5 leaves, 85 − 5 roots was higher than Jinjiu roots (Fig. 2d),cluster analysis showed that the expression of 10HGO, DXR and G10H was clustered into one group, DXS and GPPS was clustered into one group. In stage V, the expression of DXS, DXR, 10HGO, G10H and GPPS in leaves was higher than that in roots, and the expression was more complex, cluster analysis showed that DXR and GPPS gene expression were clustered into one group, DXS and G10H gene expression were clustered into one group, and 10HGO was a separate group (Fig. 2e).

Different varieties
In roots, the content of total iridoid glycosides of 85 − 5 and Jinjiu reached the highest in stage IV and decreased in stage V, and in the mature stage of R.glutinosa (IV-V), the content of iridoterpene glycosides in Jinjiu roots was signi cantly higher than that in 85 − 5 roots. In leaves, the content of total iridoid glycosides of 85 − 5 and Jinjiu reached the highest in stage III and decreased continuously in stage IV-V, in mature stage (IV-V stage), Jinjiu leaves was signi cantly lower than that in 85 − 5 leaves. Further comparison showed that the content of total iridoid glycosides in leaves was higher than that in roots in the early growth stage (I-II stage), from the middle growth stage (III stage), the content of roots was higher than that in leaves, and it was also the same in the later growth stage (IV-V stage) (Fig. 3a). It is worth noting that the content of iridoid glycosides in Jinjiu root was signi cantly higher than that in 85 − 5 roots in the mature stage(IV-V stage).

Different producing areas
In roots, the content of total iridoid glycosides in Wenxian was the highest in stage III, and that in Xinxiang was the highest in stage IV, in the mature stage (IV-V stage), the content of total iridoid glycosides in Wenxian root was lower than that in Xinxiang root. In leaves, the content of total iridoid glycosides in Wenxian and Xinxiang reached the highest in stage III, decreased continuously in stage IV-V, and in mature stage (IV-V), Wenxian leaves was slightly lower than that in Xinxiang leaves. Further comparison showed that in the early growth stage (I-II stage), the content of total iridoid glycosides in leaves was higher than that in roots, but from the middle growth stage (III stage), Wenxian root increased signi cantly, then in stage IV, Xinxiang root increased signi cantly, but both decreased in stage V (Fig. 3b).It is worth noting that in stage V, the content of iridoid glycosides in Wenxian root was almost the same as that in Xinxiang root.
Analysis of the correlation between iridoid glycoside accumulation and gene expression R.glutinosa uses roots as medicine, we use Jinjiu root and 85 − 5 root as research materials, SPSS17.0 software to analyze the correlation between iridoid glycoside accumulation and related enzyme genes expression.The results showed that there was a signi cant negative correlation between the content of total iridoid glycosides and the expression of GPPS in 85 − 5 roots, and the correlation coe cient was − 0.906(P 0.05), and there was no signi cant correlation with DXS, DXR, 10HGO and G10H. In Jinjiu roots, the content of total iridoid glycosides was signi cantly negatively correlated with the expression of DXR, and the correlation coe cient was − 0.991(P 0.01), and there was no signi cant correlation with DXS, GPPS, 10HGO and G10H (Table S4).

Level of Genomic Methylation in R. glutinosa
With the growth of R.glutinosa, the content of 5mC in roots and leaves reached the maximum in stage III, and decreased continuously in stage IV-V, and the content of 5mC in leaves was higher than that in roots during the whole growth cycle (Fig. S5). Without changing the methylation status of the restriction site, the methylation state of DNA sequence can be detected by using HpaII and MspI with different sensitivity to CCGG methylation. HpaII is not sensitive to the full methylation (double-stranded methylation) of single or two cytosines and only cleaves hemimethylation, while MspI is sensitive to fully methylated internal cytosine and insensitive to fully methylated external cytosine. Therefore, the MSAP analysis of the restriction fragment can re ect the degree and status of methylation of the enzyme site. In this study, the band patterns measured by MSAP method are   (Fig. 4a). The ratio of methylated bands of roots reached the highest in stage V, reaching 60.00%, and there was little difference in II-V (58.57%, 58.33%, 56.72%, 60.00%), and the ratio of leaves reached the highest in stage V, reaching 60.22%, and there was little difference between stage IV and V (57.53%, 60.22%), both roots and leaves show the trend of decreasing rst and then increasing (Fig. 4b).
Effects of 5-azaC on accumulation of iridoid glycosides in R.glutinosa The content of total iridoid glycosides in 85 − 5 roots accumulated continuously with the increase of growth period, after treatment with 5-azaC, content of iridoid glycosides increased signi cantly, in the same growing period, the content of total iridoid glycosides increased with the increase of 5-azaC concentration, especially in the mature stage (IV-V), high concentration is more signi cant (100µM or 250µM). It is worth noting that in stage III, 250µM 5-azaC signi cantly inhibited the synthesis of iridoid glycosides (Fig. 5a). The content of total iridoid glycosides in 85 − 5 leaves was more complicated. In stage I, II, III, V, different concentrations of 5-azaC had no signi cant effect on the accumulation of total iridoid glycosides, only in stage IV, 5-azaC has obvious promotion effect. It is worth noting that in stage II and V, low concentration of 5-azaC (15µM-100µM) signi cantly inhibited the synthesis of iridoid glycosides (Fig. 5b).
In Jinjiu roots, 5-azaC promoted the accumulation of iridoid glycosides during the whole growth cycle, in the early growth stage (I-II stage) of R.glutinosa, the accumulation of iridoid glycosides at different concentrations of 5-azaC was very signi cant, but in the middle growth stage (III stage) showed inhibitory effect, then in the late growth stage (IV-V stage), it showed a promoting effect but was not signi cant (Fig. 5c). The content of total iridoid glycosides in Jinjiu leaves was more complicated. In general, 5-azaC promoted the accumulation of iridoid glycosides,and in stage III, IV, V, promoting effect of low concentration and high concentration was signi cant, but middle concentration was not signi cant (Fig. 5d).

Effects of 5-5-azac on gene expression in R.glutinosa
After treatment with different concentrations of 5-azaC, we detected the changes of DXS, DXR, 10HGO, G10H and GPPS expression.
In 85 − 5 roots, 5-azaC promoted the expression of DXS, DXR, 10HGO, G10H and GPPS in varying degrees, with the increase of 5-azaC concentration, the promoting effect increased at rst and then decreased, the promoting effect of 15-50µM 5-azaC was better than other concentration, when the concentration of 5-azaC was more than 100µM, it inhibited the expression as a whole. It is worth noting that the expression of G10H in stage IV, 30µM 5-azaC treatment was nearly 5 times higher than that in the control group, and that of GPPS in stage III, 50µM 5-azaC treatment was nearly 6 times higher than that in the control group (Fig. 6a). In Jinjiu roots, different concentrations of 5-azaC signi cantly inhibited the expression of DXS, which decreased at rst and then enhanced with the increase of concentration, and 5-azaC promoted the expression of DXR, 10HGO and GPPS to a certain extent, especially in the stage I-II, both low concentration (15µM, 30µM) and high concentration (100µM, 250µM) signi cantly promoted the expression. 5-azaC promoted the expression of G10H in varying degrees, with the increase of 5-azaC concentration, the promoting effect of G10H increased at rst and then decreased, 30-50µM 5-azaC has great promoting effect, and when the concentration of 5-azaC was more than 100µM, inhibition of expression. It is worth noting that in stage IV, the expression of DXS under different concentrations of 5-azaC was more than 10 times lower than that of the control group, the expression of DXR in stage II, 250µM 5-azaC was nearly 6 times higher than that of the control group, and the expression of G10H in stage II, 50µM 5-azaC treatment was nearly 10 times higher than that of the control group (Fig. 6b).
In 85 − 5 leaves, 5-azaC promoted the expression of DXS, DXR, 10HGO, G10H and GPPS in varying degrees, and the promoting effect increased at rst and then decreased with the increase of 5-azaC concentration, and promoting effect of 15-50µM 5-azaC was better, when the concentration of 5-azaC was more than 100µM, the promoting effect was not as signi cant as that of low concentration. It is worth noting that the expression of GPPS in stage III, 50µM 5-azaC treatment was nearly 6 times higher than that of the control group, while that of G10H in stage II, treated with various concentrations of 5-azaC was more than 4 times lower than that of the control group, and in stage IV, the expression level of G10H treated with 30µM 5-azaC was 5 times lower than that of the control group (Fig. 6c). In Jinjiu leaves, 5-azaC promoted the expression of DXS, DXR, 10HGO, G10H and GPPS in varying degrees, with the increase of 5-azaC concentration, the promoting effect increased at rst and then decreased. The promoting effect of 15-50µM 5-azaC was better, when the concentration of 5-azaC was more than 100µM, the promoting effect was not as signi cant as that of low concentration. It is worth noting that the expression of DXS in stage III and IV was lower than that in the control group under all concentrations of 5-azaC, while the promoting effect of DXR in stage II was very signi cant, especially at 30µMand 100µM, it increased by 14 and 25 times, respectively (Fig. 6d).
Effects of 5-azaC on 5mC Content in Genomic DNA of R. glutinosa We used different concentrations of 5-azaC to treat roots and leaves, and found that with the maturity of R.glutinosa, the content of genomic 5mC in roots and leaves in mature stage was signi cantly higher than that in seedling stage, and the content of genomic DNA 5mC in roots and leaves decreased with the increase of 5-azaC concentration, 5-azaC signi cantly inhibited the formation of genomic methylation in roots and leaves ( Fig. 7a,b). In different growth periods, the content of 5mC in the control group was the highest, and the inhibitory effect gradually enhanced with the increase of 5-azaC concentration in the same growth period, further analysis showed that the content of genomic 5mC in R.glutinosa root and leaf was different at the same period, and leaf was higher than that in root tuber.  (Table S7). The total methylation ratio of the control group was signi cantly lower than that of the treatment group except stage II, and increased at rst and then decreased with the increase of 5-azaC concentration (Fig. 8a). The change of methylation band ratio (MSAP%) is more complex, MSAP% of the control group in stage I is slightly lower than that in the other four stages, and there is no signi cant difference in the other four stages, on the whole, MSAP% of 5-azaC treatment in the same growth period is basically higher than that of the control group, and the maximum value of 15µM-100µM concentration treatment appeared in stage IV, IV, IV and V, respectively, and the maximum value of 250µM concentration treatment appeared in stage II (Fig. 8b).
In leaves, 6 pairs of primers ampli ed about 61-130 clear bands, of which the number of methylated bands was 24-83 bands, accounting for 33.33%-68.03%, and the proportion of fully methylated bands was 13.04%-35.29%, indicating that most of the genomic CCGG sequences in R.glutinosa leaves are in a state of hemimethylation and total methylation (Table S8). The change of total methylation ratio in leaves was more complex, low concentration of 5-azaC in the same growth period played a promoting role, while high concentration of 5-azaC played an inhibitory role, the control group reached the maximum value in stage III, the maximum value of 15µM-100µM concentration treatment appeared in II, III, III and II stage, and the maximum value of 250µM concentration treatment appeared in V stage (Fig. 8c). Under the treatment of 5-azaC, in I-III stage, the methylation band ratio (MSAP%) was promoted at low concentration, inhibited at high concentration, but both inhibited in IV and V stage. The control group reached the maximum value in stage IV, and the maximum value of 15µM-250µM concentration treatment appeared in stage II, III, IV, II and III, respectively (Fig. 8d).

Change of Methylation Status in R. glutinosa treated with 5-azaC
We used roots and leaves of mature R.glutinosa to analyze changes in DNA methylation status. Compared with the control group, the changes of genomic methylation status of R.glutinosa treated with 5-azaC could be divided into 3 categories and 12 bands. We set class A as a monomorphic band, means that the control group has the same methylation sites as the treatment group, indicating that the methylation status of CCGG sites remains unchanged, while B and C are polymorphic bands, means that methylation sites in the control and treatment groups were different, type B is methylation increased band pattern, indicating that the methylation level of genomic DNA has increased, and the band pattern of DNA methylation increase can be summarized as follows: I→III, I→IV, II→III, II→IV. Type C is the demethylation type, which indicates that the methylation level of genomic DNA has decreased, and the band pattern of demethylation change is: IV→II, IV→I, III→II, III→I. Due to the limitation of MSAP technique, the increase or decrease of DNA methylation could not be determined as I→II, III→IV, IV→III, II→I.
Statistical analysis of genomic DNA methylation increases and decreases in the band pattern (Table 3), we found that in roots most of the bands with increased methylation were II→IV, which changed from semimethylation to full methylation, and a few transformed bands were I→III, at the concentration of 30µM-100µM 5-azaC, the methylation band increased signi cantly. The demethylation band type is mainly IV→II, with the increase of 5-azaC concentration, the number of demethylated bands also increased, and reached the maximum value of 12 bands at 50µM, III→I and IV→I also account for a certain proportion of demethylated bands,and as the concentration increased, the number of demethylated bands also increased and all reached the maximum value at 250µM. Further analysis we found that the number of methylation sites decreased more than the number of methylation sites increased, which was consistent with the trend of MSAP% under 5-azaC treatment in stage V.  The methylation increase band pattern and demethylation band pattern in leaves are more than those in roots (Table 4), we found that in leaves most of the bands with increased methylation were II→IV, which changed from semi-methylation to full methylation, and a few transformed bands were I→III, at the concentration of 30µM-50µM 5-azaC, the methylation band increased signi cantly. The demethylation band type is mainly IV→II, with the increase of 5-azaC concentration, the number of demethylated bands also increased, and reached the maximum value of 16 bands at 250µM, III→I, III→II and IV→I also account for a certain proportion of demethylated bands, and as the concentration increased, the number of demethylated bands also increased and all reached the maximum value at 250µM also. Further analysis we found that the number of methylation sites decreased more than the number of methylation sites increased, especially under high concentration treatment, the number of bands decreased signi cantly, which was consistent with the trend of MSAP% under 5-azaC treatment in stage V.  In general, the number of methylated and demethylated bands increased with the increase of 5-azaC concentration in roots and leaves of R.glutinosa, and effects of 5-azaC on genomic methylation status of roots and leaves were roughly the same.

Discussion
Plant secondary metabolites play an important role in many aspects of plant life activities such as affecting plant growth and development, adapting to the environment, resisting diseases and insect pests [40][41][42]. There are many factors affecting the accumulation of secondary metabolites in plants,such as temperature, light, water, nutrients and so on [43][44][45][46]. The metabolic pathway of terpenoids is very complex, involving a variety of enzymes and genes, and some studies have shown that the accumulation of secondary metabolites is regulated by epigenetics. In this study, we used R.glutinosa 85 − 5 and Jinjiu as experimental materials and found that in the early growth stage (I-II stage), the clustering of gene expression is the same (Fig. 2a,b),which reveals the level and clustering trend of genes expression in different tissues in the early stage growth, and in the middle growth stage (III stage), the clustering of gene expression is similar to that in stage I and II, but in the later growth stage (IV-V stage), it is obviously that the clustering of gene expression has changed signi cantly (Fig. 2c,d,e), although the expression of gene clusters varied in different growth stages, the expression of genes in the late growth period was higher than that in the early growth period, meanwhile, we found that the accumulation of iridoid glycosides also presents an upward trend, which was similar to the genes expression, this results indicated that the level of gene expression affected the accumulation of iridoid glycosides, but the correlation analysis showed that there was no signi cant positive correlation between the accumulation of iridoid glycosides and gene expression, we speculate that this is due to the fact that the accumulation of iridoid glycosides is affected by many internal and external factors, and the DXS, DXR, 10HGO, G10H, GPPS in this study are only part of the genes in the metabolic pathway, so further study is needed.
DNA methylation plays an important role in the growth and development of plants and the accumulation of secondary metabolites, we found that the ratio of total methylation and the ratio of methylated bands (MSAP%) in the roots increased gradually with the maturity of R.glutinosa, and leaves are similar to roots, and it was consistent with gene expression and accumulation of iridoid glycosides, total methylation% and MSAP% reached the maximum in stage IV or V, while the gene expression usually reached the maximum in stage III or IV, indicating that proper methylation had little effect on gene expression in the early stage of R.glutinosa growth, while the methylation site continued to increase in the later stage, resulting in the inhibition of gene expression.
We also found that DNA methylation and demethylation affect the accumulation of secondary metabolites and the expression of related genes in the secondary metabolic pathway. In this study, different concentrations of 5-azaC can promote the accumulation of iridoid glycosides to a certain extent, and the promoting effect of low concentration is not as signi cant as that of high concentration, and the promoting effect is usually the most signi cant when the concentration is 100µM and 200µM, and different concentrations of 5-azaC also can promote DXS,DXR,10HGO,G10H,GPPS expression. Interestingly, in the II stage of 85 − 5 roots, when the concentration of 5-azaC was 15µM and 50µM, the gene expression was inhibited, and the accumulation of iridoid glycosides was also inhibited. In the II, III, V stage of 85 − 5 leaves, gene expression and iridoid glycoside accumulation were inhibited by 5-azaC treatment, so it can be inferred that the effects of 5-azaC on the accumulation of iridoid glycosides and gene expression were consistent, and the increase of gene expression leads to an increase in the accumulation of iridoid glycosides, and vice versa.
The use of 5-azaC will lead to changes in the methylation status of genomic DNA, and affects the normal growth and development of plants [47,48] and the accumulation of metabolites. The Research found that 5-azaC inhibited the formation of genomic DNA methylation, and the inhibition became more signi cant with the increase of concentration, further analysis of total methylation ratio and methylation band ratio (MASP%), we found that low concentration of 5-azaC promotes the formation of methylation, and high concentration of 5-azaC inhibits the formation of methylation. Interestingly, the formation of methylation sites and demethylation in roots by 5-azaC treatment were less signi cant than those in leaves, and the change of complete methylation ratio is more signi cant than methylation band ratio, in order to verify these phenomena, we statistically analyzed the dynamic changes of methylation state under 5-azaC treatment, found that no matter the methylation increase band pattern or the methylation decrease band pattern, the number in the roots is less than that in the leaves, and the number of methylation sites increased and decreased the most when the concentration of 5-azaC was 50µM or 100µM in roots, and 250µM in leaves. In roots, the increased methylation band pattern is mainly the transition from semi-methylation to total methylation, the demethylation bands are mainly characterized by the transition from total methylation to semi-methylation, and the transformation from total methylation to non-methylation also accounts for a considerable proportion, and this situation in the leaves is similar to the roots, these corresponds to the above research results. Further analysis showed that with the increase of 5-azaC concentration, the number of demethylation sites increased gradually, when treated with low concentration(15µM-50µM) of 5-azaC, proper demethylation promoted gene expression, and the accumulation of iridoid glycosides was also promoted, while high concentration(100µM-250µM) 5-azaC treatment led to excessive demethylation, which affected the normal gene expression, but the accumulation of iridoid glycosides was still promoted, and in some cases, the promoting effect of high concentration was more signi cant than that of low concentration, we speculate that this is because the synthesis process of iridoid glycosides is complex, and there are many enzymes and genes involved, so there is no signi cant correlation only from the results of this study, and further research is needed, and under different concentrations of 5-azaC treatment, the dynamic changes of iridoid glycoside accumulation, gene expression and methylation status in mature leaves were more complex and obvious than roots, shows that the treatment of 5-azaC is tissue speci c and plays a more signi cant role in leaves. Therefore, we concluded that 5-azaC inhibited the methylation of genomic DNA, promoted the expression of related enzyme genes in the iridoid glycoside synthesis pathway, and promoted the accumulation of iridoid glycosides.

Conclusion
We compared the iridoid glycoside accumulation, gene expression and genomic DNA methylation of R.glutinosa between the control group and the 5-azaC treatment group, and found that with the maturity of R.glutinosa, iridoid glycosides gradually accumulated, gene expression and methylation sites increased,after treatment with different concentrations of 5-azaC, iridoid glycoside accumulation and gene expression were promoted, and DNA methylation was inhibited. Through analyzing the dynamic changes among these, it is expected to reveal the correlation between DNA methylation and iridoid glycoside accumulation and the expression of key enzyme genes in iridoid glycoside metabolic pathway, which is expected to provide reference for revealing the mechanism of plant secondary metabolic pathways from the perspective of epigenetics.

Declarations
Ethics approval and consent to participate We con rm that all materials and methods were in compliance with relevant institutional, national, and international guidelines and legislation.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analysed during this study are included in this published article [and its supplementary information les].
Competing interests Heat map and Cluster analysis of relate enzyme genes in R.glutinosa. a: I growth period,b: II growth period, c: III growth period, d:IV growth period,e: V growth period Total iridoid glycosides in R.glutinosa. a: different varieties;b: different producing areas.statistical signi cance was determined by a two-sided t-test: * P < 0.05, ** P < 0.01.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Supplementarymaterial.pdf Supplementarymaterial.pdf