Isolation of biosynthesis related transcripts of 2,3,5,4'-tetrahydroxy stilbene-2-O-β-D-glucoside from Fallopia multiflora by suppression subtractive hybridization

Fallopia multiflora is a traditional Chinese medicinal herb, which has been widely used for thousands of years. Research showed that F. multiflora has potent antioxidative and cytoprotective properties [1], enhanced purgative effects, promoted diuresis and choleretic effects [2], and exerted a neuroprotective effect against glutamate-induced neurotoxicity [3]. 2,3,5,4'-tetrahydroxy stilbene-2-O-ß-Dglucoside (THSG) is an active component in F. multiflora, which possesses anti-hyperlipidemic [4], anti-oxidative, anti-inflammatory, endothelial-protective activities [5]. Furthermore, THSG can protect osteoblastic MC3T3-E1 cells via inhibiting the release of bone-resorbing mediators and oxidative damage of the cells [6], and suppress atherosclerosis by altering the expression of key proteins that may be novel molecular targets responsible for atherogenesis [7]. Sun et al. who reported that THSG might provide a potentially new strategy for preventing and treating neurodegenerative disorders such as Parkinson’s disease, also showed that THSG may protect neurons against MPP+induced cell death through improving mitochondrial function, decreasing oxidative stress and inhibiting apoptosis [8]. The pharmacological research revealed that THSG has a potential impact on human health. As a stilbene, THSG and resveratrol (3,5,4'-trihydroxy-trans-stilbene) belong to phenylpropanoids characterized by a 1,2-diphenylethylene backbone. Plant stilbenes are derived from phenylalanine biosynthesis via the general phenylpropanoid pathway. One p-coumaroylCoA or cinnamoyl-CoA derived from the phenylpropanoid pathway is ligated with three malonyl-CoA under the catalysis of stilbene synthase (STS) and stilbene is the product [9]. STS is a polyketide synthase (PKS) belonging to type III PKSs. Other PKSs III include resveratrol synthase (RS), chalcone synthases (CHS), bibenzyl synthase (BBS), stilbene carboxylate synthase (STCS), 4-coumaroyltriacetate lactone synthase (CTAS) and more [10]. Every PKSs III catalyzes the synthesis of one unique stilbene. Up to now, no enzyme was found to catalyze the biosynthesis of THSG. Sheng et al. isolated a stilbene synthase gene FmPKS from the rhizomes of F. multiflora. The gene expression pattern in the plant correlated with the THSG content in different tissues, but Abstract


Introduction
Fallopia multiflora is a traditional Chinese medicinal herb, which has been widely used for thousands of years.Research showed that F. multiflora has potent antioxidative and cytoprotective properties [1], enhanced purgative effects, promoted diuresis and choleretic effects [2], and exerted a neuroprotective effect against glutamate-induced neurotoxicity [3].2,3,5,4'-tetrahydroxy stilbene-2-O-ß-Dglucoside (THSG) is an active component in F. multiflora, which possesses anti-hyperlipidemic [4], anti-oxidative, anti-inflammatory, endothelial-protective activities [5].Furthermore, THSG can protect osteoblastic MC3T3-E1 cells via inhibiting the release of bone-resorbing mediators and oxidative damage of the cells [6], and suppress atherosclerosis by altering the expression of key proteins that may be novel molecular targets responsible for atherogenesis [7].Sun et al. who reported that THSG might provide a potentially new strategy for preventing and treating neurodegenerative disorders such as Parkinson's disease, also showed that THSG may protect neurons against MPP+-induced cell death through improving mitochondrial function, decreasing oxidative stress and inhibiting apoptosis [8].The pharmacological research revealed that THSG has a potential impact on human health.As a stilbene, THSG and resveratrol (3,5,4'-trihydroxy-trans-stilbene) belong to phenylpropanoids characterized by a 1,2-diphenylethylene backbone.Plant stilbenes are derived from phenylalanine biosynthesis via the general phenylpropanoid pathway.One p-coumaroyl-CoA or cinnamoyl-CoA derived from the phenylpropanoid pathway is ligated with three malonyl-CoA under the catalysis of stilbene synthase (STS) and stilbene is the product [9].STS is a polyketide synthase (PKS) belonging to type III PKSs.Other PKSs III include resveratrol synthase (RS), chalcone synthases (CHS), bibenzyl synthase (BBS), stilbene carboxylate synthase (STCS), 4-coumaroyltriacetate lactone synthase (CTAS) and more [10].Every PKSs III catalyzes the synthesis of one unique stilbene.Up to now, no enzyme was found to catalyze the biosynthesis of THSG.Sheng et al. isolated a stilbene synthase gene FmPKS from the rhizomes of F. multiflora.The gene expression pattern in the plant correlated with the THSG content in different tissues, but THSG was still not detectable in transgenic Arabidopsis thaliana in which FmPKS was inserted and expressed [11].Shao et al. showed precursor feeding of methyl jasmonate and salicylic acid in suspension cultures of F. multiflora that could increase THSG production [12].It is still not clear if THSG is synthesized via phenylalanine pathway or some other way.Suppression subtractive hybridization (SSH) is helpful in identifying differentially expressed genes.Extensive studies showed that SSH is a powerful tool in the analysis of stress resistance [13], pathological mechanism [14] and developmental physiology [15].Moreover, Wang et al. found that light could be effective for activation of the biosynthesis of phenylpropanoids by establishing SSH cDNA libraries of tea calli [16].In this study, we attempted to identify the genes that are related to THSG biosynthesis by SSH.Full-length sequences were obtained using 3'5'RACE, and then the gene functions were annotated by blastn to nucleotide databases nt, and blastx to protein databases nr, Swiss-Prot, Kyoto Encyclopedia of Genes and Genomes (KEGG) and clusters of orthologous groups (COG).Meanwhile, comparative analysis with the transcriptome and digital expression profile revealed the expression differences for each gene in various tissues of F. multiflora.This may indicate that these genes are associated with THSG synthesis.

Plant material collection
F. multiflora plants were gathered from 13 cities (counties) in March 2010, which includes Guangxi Province (Nanning City, Guilin City, Jingxi County, Tianlin County), Guangdong Province (Guangzhou City, Shenzhen City, Dongguan City, Deqing City, Zhaoqing City, Gaozhou City, Zhanjiang City), Jiangxi Province (Xinyu City), Chongqing City.As some samples were gathered in the field, the age was unclear.All plant materials were maintained in the medicinal plant garden of the Department of Pharmacy, Guangzhou Liuhuaqiao Hospital, Guangzhou, China.

HPLC analysis of THSG
To quantify the THSG in F. multiflora roots, fresh plant materials were frozen in liquid nitrogen and ground to fine powder in a mortar.After vacuum freeze-drying, 0.2 g of each sample was taken up in 25 ml 50% (v/v) methanol and then refluxed at room temperature for 16 h.After being filtered through a 0.22 µm film, 10 μl of filtrate was analyzed by HPLC using a Dikma Diamonsil C18 column (250 × 4.6 mm, tablets path 5 µm).Chromatographic separation was performed using a solvent system of H 2 O and CH 3 CN with the ratio of 3:1 (v/v) over 10 min.The flow rate was 1 ml/min, with detection at 320 nm.Each data point represents the average of three independent experiments.

Total RNA extraction and mRNA purification
Total RNA was extracted from leaves and roots using the Plant Total RNA Isolation kit (Bioteke, China) and treated with DNase I (TaKaRa, Dalian, China).Ethidium bromide (EtBr) staining, agarose gel electrophoresis and spectrophotometric (NanoDrop 2000, USA) analysis were performed to examine the quality and concentration of total RNA.mRNA was purified using Oligo tes™-dT 30 <SUPER> mRNA Purification Kit (from Total RNA) according to the manufacturer's instructions (TaKaRa, Dalian, China).

Suppression subtractive hybridization (SSH)
Driver and tester cDNA was synthesized from two samples at equal amounts of purified mRNA, using SMARTer PCR cDNA Synthesis K it (Clontech, USA) according to the manufacturer's instructions.After repeated twice hybridization, target genes were amplified using nested PCR by Ad-vantage™ cDNA PCR Kit (Clontech, USA).To raise the PCR efficiency, an adapter was added to the nested PCR primers (Tab.1; homo nested primer).The following target genes were amplified using Fermentas Dream Taq (Fermentas, USA) with homo PCR primer (Tab.1; homo PCR primer).
The PCR products were inserted into the pMD19-T vector (TaKaRa, Dalian, China), and the ligated products were transformed into Escherichia coli DH5α competent cells by heat shock.Then plated onto LB medium containing 100 µg/ ml ampicillin, 24 µg/ml IPTG and 20 µg/ml X-gal, and incubated overnight at 37°C.Recombinant white colonies were randomly selected for colony PCR with universal primers (Tab.1; universal primers).Colonies containing cloned fragments were sent for sequencing.Besides control sample (Poly A + RNA from human skeletal muscle) provided with the kit, which was used as a control driver cDNA, the experiments were performed with 3 different testers: I -cDNA from Guilin's root sample as tester and Chongqing's root sample as driver; II -Chongqing's root sample as tester and Guilin's root sample as driver; III -Guilin's leaf sample as tester and Guilin's root sample as driver.

Cloning genes with full length by RACE
The 3'5'RACE of the candidate genes was performed in purified mRNA using PCR-select™ cDNA subtraction kit (Clontech, USA) according to the manufacturer's instruction.To improve the specificity, nested primers were designed for Nested-PCR besides gene specific primers.Fragments of 4 genes were selected to do 3'5'RACE from SSH library after BLAST in NCBI (Tab.1).The Nested-PCR reaction was performed by Advantage cDNA Polymerase Mix (Clontech, USA) with the following thermal cycling parameters: 94°C for 4 min, followed by 35 cycles of 94°C for 30 s, 68°C for 30 s and 72°C for 1.5 min, and the final extension was performed at 72°C for 10 min.The PCR product was cloned into the pMD19-T vector (TaKaRa, Dalian, China) and sequenced.

Sequence analysis
Gene fragments obtained from SSH and RACE were compared with the sequences in nucleotide collection database and expressed sequence tag (EST) database at NCBI using the BLASTN algorithm.Further annotation was carried out using nr, Swiss-Prot, gene ontology database (GO), COG and KEGG.Clustalx1.83and Contig Express software were used for multiple sequence alignment, sequence identity and linkage.Moreover the obtained sequences were compared with the transcriptome and digital gene expression tag profiling (DGE) database of four different tissue samples (Cr: roots of Chongqing F. multiflora; Dr: roots of Deqing F. multiflora; Dl: leaves of Deqing F. multiflora; Ds: stems of Deqing F. multiflora) of F. multiflora (unpublished data) using local BLAST.

Quantification of THSG
To select the specific samples with robust difference in THSG concentration, HPLC was used to analyze the THSG contents.Results indicated that there was a significant difference (P-value < 0.001) in THSG content in F. multiflora root from different origins (Fig. 1, Tab. 2).Root samples from Guilin City and Zhaoqin City have the highest accumulation of THSG, with a content of 6.37% and 5.25% respectively.Chongqing samples have the lowest THSG level, with a content of 0.001%.Significant differences were found even within Guilin samples.For example, Guilin samples, the lowest is 1.46%, while the highest is 6.37%.According to this data, we choose Guilin and Chongqing samples as the following materials.

SSH fragments comparison analysis
SSH fragments were amplified and cloned into pMD19-T vector.A total of 136 colonies were picked up, and the insertions were confirmed by PCR using universal primers.75 clones among the 136 colonies were positive (55.1%) and sequenced, with insertion length ranging from 0.2 to 0.5 kb (data not shown).Removing redundant sequences, 12 fragments were obtained, and they were compared with the sequences in nucleotide and EST database at NCBI using the BLASTN algorithm.Results are shown in Tab. 3.There are 4 sequences with no significant similarity in the databases.The 4 sequences were all obtained from the roots and named GE, GW, CF, CA for subsequent research ("G" represent the origin of the sample, Guilin city; "C" represent Chongqing city.).All sequences were submitted to the NCBI GenBank (accession Nos.JZ469200 to JZ469209) for public domain use, except YCTS4 and GTS2 (ribosomal RNA).
Then the 12 sequences were compared with the transcriptome and DEG database of F. multiflora (unpublished data) using local BLAST.Significantly similar genes identified in the transcriptome database were initially aligned by blastx to protein databases nr, Swiss-Prot, KEGG and COG (evalue <0.00001), and then aligned by blastn to nucleotide database nt (e-value <0.00001).Proteins with the highest sequence similarity were retrieved with the given Unigenes and their functional annotations (Tab.4).

3'5'RACE and full-length sequences comparison analysis
The full-length sequences of GE, GW, CF and CA were cloned by 3'RACE and 5'RACE.More than one full-length sequence was obtained for all genes but one, CF.All sequences have been submitted to the NCBI GenBank (accession Nos.KF054163 to KF054169) for public domain.Meanwhile the sequences were compared with the transcriptome and DEG database using local BLAST.Significantly similar Unigenes in transcriptome were aligned by blastx to protein databases, and aligned by blastn to nucleotide databases nt to annotate gene function (Tab.5).

Pathway annotation
Different genes generally interact with each other to sustain their biological functions.Pathway-based analysis Tab.   helps to further understand genes' biological functions.According to KEGG, the pathways of isolated genes are diverse in function.These genes were mainly involved in metabolic pathways; ubiquitin mediated proteolysis, pathogenic E. coli infection, amyotrophic lateral sclerosis (ALS), tryptophan metabolism, photosynthesis, gap junction, RIG-I-like receptor signaling pathway, mRNA surveillance pathway, peroxisome, endocytosis, ribosome, biosynthesis of secondary metabolites, phagosome, protein processing in endoplasmic reticulum, methane metabolism, glyoxylate and dicarboxylate metabolism, microbial metabolism in diverse environments, and RNA transport (Tab.6).

Comparison of genes expression and THSG content in different tissues
Genes expression of SSH fragments in different plant tissues according to DGE database was shown in Fig. 2. Expression of GE, GW, CF and YTS1 in different plant tissues has no obvious difference.Gene expression of full-length sequences in different plant tissues has significant differences (Fig. 3), except GEfl and CFfl.Comparison between the expression of isolated genes and THSG content in different tissues revealed that the expression of GWfl1 (Unigene20297_TrCD), GWfl3 (Unigene27674) and CWfl1 (CL4034.Contig3_TrCD) is in parallel with THSG content in different tissues.Expression of CA, YCTS4 and CWfl1 (Unigene22710_TrCD) is contrary to THSG content in different tissues.Expression of CA, CTS1, GWfl2 (CL3625.Contig1_TrCD), GWfl3 (CL3504.Contig1_TrCD) , YCTS4 and CWfl1 (Unigene22710_TrCD) in Cr are higher than in all D sample tissues.

Discussion
As of any secondary metabolite in plant, the production of THSG in F. multiflora is influenced by multiple factors, such as temperature, climate, altitude, precipitation etc. Gene screening between samples of distinct product content by SSH could display the mRNA information associated with the production of THSG.DGE profile of different plant tissues contains genes expression information in these tissues.Comparison with DGE profile showed the high level of activity of targeted genes in each tissue.The genes whose expression was significantly different were selected.Similar pattern of difference in gene abundance and THSG content in different tissues confirm that the gene might be associated with THSG biosynthesis.
In our study, the expression of GWfl1 (Unigene20297_ TrCD), GWfl3 (Unigene27674_TrCD) and CWfl1 (CL4034.Contig3_TrCD) is in parallel with THSG content in different tissues.Blast in NCBI database showed they might code 40S ribosomal protein, ubiquitin-conjugating enzyme and RNA-binding protein of the Puf family.The predicted protein function indicated there should be active protein degradation and expression behavior in root sample.Expression of CA, YCTS4 and CWfl1 (Unigene22710_TrCD) is contrary to THSG content in different tissues.And CTS1 is higher expressed in Cr than Dr, Ds and Dl.They might be inhibitors of THSG biosynthesis.As roots always contain the highest amount of THSG, we consider the root is most likely the site of THSG synthesis.The fragments, which were obtained from Guilin root tester should be focused on.GW (CL5835.Contig1_TrCD/CL5835.Contig2_TrCD) and GTS1 (CL8017.Contig1_TrCD) might be involved in providing energy for secondary metabolism via GTP metabolism and photosynthesis.GWfl2 (CL3625.Contig1_TrCD) is annotated as Katanin p80 WD40 repeat-containing subunit B1, GWfl3 (CL3504.Contig1_TrCD) as Ubiquitin-protein ligase.As these fragments were identified in SSH when Guilin root tester was applied, it means the intermediate process of THSG biosynthesis should be fast and efficient, and related protein is likely to be subjected to rapid degradation.GTS2 (CL10591.Contig1_TrCD) has high homology with Cytochrome P450   in Medicago truncatula.Cytochrome P450 was found be involved as multifunctional oxidases in the biosynthesis of many secondary metabolites like triterpenoid [17], glycyrrhizin [18], hemolytic saponins [19] and many more.Does it take part in THSG biosynthesis?If THSG is derived from phenylalanine pathway with other stilbenes, then the CTS1 gene obtained from Chongqing root tester may be related with the inhibition of THSG biosynthesis.CTS1 gene is annotated as catalase and participates in "biosynthesis of secondary metabolites (ko01110)"; "glyoxylate and dicarboxylate metabolism (ko00630)" and "tryptophan metabolism (ko00380)".It may regulate the metabolism of tryptophan to influence phenylalanine synthesis.Whether or not, THSG biosynthesis must be complicated and correlate with multiple pathways.More than just a phenylalanine pathway also must have common effect of ATP energy metabolism active and other related reactions to strengthen or weaken the process.Our research isolated several candidate genes that might participate in the THSG biosynthesis.Further research will focus on the characterization of candidate genes and gene screening to transcriptome in more diverse tissues.
represents the origin of the sample, Guilin city; "C" represents Chongqing city.

Fig. 2
Fig. 2 Comparison of SSH fragments expression and THSG content in different tissues.Cr -roots of Chongqing F. multiflora; Dl -leaves of Deqing F. multiflora; Dr -roots of Deqing F. multiflora; Ds -stems of Deqing F. multiflora; TPM -transcripts per million clean tags (all data was calculated as the base of log10).

Fig. 3
Fig. 3 Comparison of full-length sequences expression and THSG content in different tissues .Cr -roots of Chongqing F. multiflora; Dl -leaves of Deqing F. multiflora; Dr -roots of Deqing F. multiflora; Ds -stems of Deqing F. multiflora; TPM -transcripts per million clean tags (all data was calculated as the base of log10).
1 SSH primers and RACE primers of 4 selected gene fragments.
Parameters of high similarity SSH sequence alignments, identified by BLASTN in nucleotide and EST NCBI databases.Parameters and annotation of high similarity full-length sequences alignments, identified by local BLAST in F. multiflora transcriptome databases.
4he character before "TS" represent the tester; "YCTS" means the fragment was found from both "Y" and "C" testers.C -roots of Chongqing F. multiflora; G -roots of Guilin F. multiflora; Y -leaves of Guilin F. multiflora.XM_003517340Tab.4Parametersand annotation of high similarity SSH sequence alignments, identified by local BLAST in F. multiflora transcriptome databases.