De Novo Transcriptome Assembly And Gene Discovery Of Cistanche Deserticola Fleshy Stem-Ⅱ

Functional Classification Of All Expressed Transcripts Based On Gene Ontology And KEGG Databases 

Gene Ontology (GO) annotation was obtained from UniProt annotation and identities association file. In total, 20,907 transcripts, accounting for 32.69% of the total expressed sequences, were assigned to 1,745 functional terms. Of the total functional GO terms, assignments to the biological process made up the majority (1,116, 63.95%) followed by cellular component (329, 18.85%) and molecular function (300, 17.20%). The assigned functions of expressed transcripts covered a broad range of GO categories, and the top 10 GO terms with the most annotated transcripts were listed in Table 3. We provide all expressed transcripts distribution in three Gene Ontology categories (Molecular Function, Cellular component, and biological process) in the supplemental file (S3 Dataset). GO terms related to binding functions and transferase activity were predominantly represented in the molecular function category. Regarding the binding functions, cation binding (4,394 transcripts) represented the most abundant, followed by nucleotide/nucleoside binding (3,404 transcripts on average) and protein binding (2,422 transcripts). While in the transferase activity group, the most are those with transferring phosphorus-containing groups (2,256 transcripts, 65.77%). Among the cellular component category, transcripts were more located intracellular (10,581 transcripts on average), while among the biological process category, transcripts were more involved in the biopolymer metabolic process (6,683 transcripts on average), followed by regulation of cellular process (4,841 transcripts), gene expression (4,678 transcripts) and transport (3,512 transcripts).

NATURAL CISTANCHE TUBULOSA FOR PREVENTING ALZHEIMER DISEASE PHGS75% ECH 30% ACT 12%

To mine genes involved in the biosynthesis of lignin and PhG, 21,358 non-redundant potential protein sequences were searched against gene sequences of 13 plant organisms in the KEGG database, and they were assigned to 275 KEGG pathways with at least 5 hits. The top 10 pathways with most aligned sequences are listed in Table 4. Most pathways were involved in primary metabolic processes, such as amino acid or protein metabolism (ko01230, ko04141, and ko04120), carbohydrate metabolism (ko01200 and ko00500), and nucleotide or nucleoside metabolism (ko03018, ko00230, and ko00240). Besides, there are 27 secondary metabolism-related pathways (Fig 2), such as terpenoid backbone biosynthesis, phenylpropanoid biosynthesis, carotenoid biosynthesis, isoquinoline alkaloid biosynthesis, and tropane, piperidine and pyridine alkaloid biosynthesis. These results provide further indication that active metabolic processes were underway in the C. deserticola stem tissue. All expressed transcripts associated with KEGG pathways were listed in the supplemental file (S4 Dataset). Although there are some significantly changed pathways between C. deserticola and other plants, such as rice(S5 Dataset), our main goal in this study is to reveal the whole transcriptome profile of C. deserticola stem and to picture the related pathways of PhGs biosynthesis which could be useful for guiding the cultivation.

Candidate Genes Encoding Enzymes Involved In The Biosynthesis Of Lignin 

Lignin is the second most abundant natural terrestrial polymer in the plant kingdom, composing up to one-third of the material found in plant cell walls. As an important component of cell walls, lignins help water transport, provide mechanical support and structural integrity, and defend against pathogens and herbivores. Those roles of lignin are very valuable in supporting the underground erective growth of C. deserticola in the desert. In this study, we presented the complete picture of lignin biosynthesis pathways in C. deserticola (Fig 3), in which the lignin monomers are biosynthesized from phenylalanine through a series of enzymatic reactions, including hydroxylation, methylation, reduction, and oxidative polymerization process. Lignin biosynthesis-related enzymes were detected for three mainly synthesized forms in vascular tissue (p-hydroxyl-phenyl (H), guaiacyl (G) and syringyl (S) lignin) and 5-hydroxyl-guaiacyl lignin which was only identified in COMT (caffeic acid 3-O-methyltransferase, EC 2.1.1.68) deficient (such as knock-down) plants.

Phenylalanine ammonia-lyase (PAL, EC 4.3.1.24) is the first key enzyme in the lignin biosynthesis pathway (Fig 3) which transforms phenylalanine into cinnamic acid by non-oxidative deamination. A total of 6,297 PAL reads were sequenced and 7 PAL transcripts were assembled in C. deserticola (Table 5). By sequence similarities comparison, we found that 4 of them (comp28550_c1_seq1/2/3/5) had more than 95% similarity with the known mRNA sequence of C. deserticola (gi|289595227|gb|ADD12041.1|), while comp28550_c1_seq4 and comp25940_c0_seq1 had 77% and 82% similarities, respectively. ORF prediction revealed 5 transcripts had potentials of encoding proteins and carried with aromatic amino acid lyase domain (PF00221.14). Among them, only the comp28550_c1_seq4 transcript could encode a complete protein sequence of 718 amino acid residues. It has been reported that PAL was encoded by a small multigene family in most plant species, such as 4 in Arabidopsis thaliana, 5 in Populus trichocarpa, 3 in Scutellaria baicalensis, and 7 Cucumis sativus, etc. Our phylogenetic analysis suggested that there were 4 PAL-encoding genes in C.

deserticola and we named them CdPAL1, CdPAL2, CdPAL3 and CdPAL4, respectively (S2 Fig). 4-coumarate-CoA ligase (4CL, EC 6.2.1.12) and trans-cinnamate 4-monooxygenase (CYP73A, EC 1.14.13.11) are two enzymes responsible for transforming cinnamic acid to dicoumarol-CoA in two reverse orders. They are also in backbones, and their expression FPKM values are 39.57 and 51.93, respectively.

The four types of lignins were biosynthesized by different pathways which were controlled by three key enzymes, cinnamoyl-CoA reductase (CCR, EC 1.2.1.44), shikimate o-hydroxycinnamoyltransferase (HCT, EC 2.3.1.133), and ferulate-5-hydroxylase (F5H, EC 1.14.-.-). CCR was reported as a control point of the lignins pathway [50, 51] which catalyzed X-CoA (X including dicoumarol, caffeoyl, feruloyl, 5-hydroxyl-feruloyl, and sinapoyl) into Y-aldehyde (Y including p-cougar, caffeoyl, coniferyl, 5-hydroxyl-coniferyl, and snap), while HCT catalyzed p-coumaroyl-CoA to p-coumaroyl shikimic acid/p-coumaroyl quinic acid. The two enzymes, just like a switch, regulated the biosynthesis of P-hydroxyl-phenyl lignins or the other three types of lignins. F5H was another branch switch that regulated syringyl lignin and 5-hydroxyl-guaiacyl lignin. Other important enzymes including caffeic acid 3-O-methyltransferase (COMT, EC 2.1.1.68), caffeoyl-CoA O-methyltransferase (CCoAOMT, EC 2.1.1.104), and cinnamyl-alcohol dehydrogenase (CAD, EC 1.1.1.195) were also detected expressed. Detailed expression information is listed in Table 6. These enzyme genes identified in this study will provide a valuable resource for functional genomic studies in this important medicinal plant. 10 genes related to the lignins biosynthesis pathway in Table 6 were selected for RT-qPCR verification to confirm our RNAseq results (Fig 4), and their high correlations (Pearson correlation coefficient: 0.90343) indicated high accuracy and reproducibility of our transcriptome analysis. S1 Dataset lists the primer sequences used in this analysis.

NATURAL CISTANCHE TUBULOSA FOR IMPROVING SEXUAL FUNCTION PHGS75% ECH 30% ACT 12%

Candidate Genes Encoding Enzymes Involved In The Biosynthesis Of PhGs 

Phenylethanoid glycosides (PhGs) are known to be the primary active ingredients in C. deserticola with activities of improving sexual potency, scavenging free radicals, and anti-aging. Three chemical components of PhGs are organic acid, saccharide, and phenylethanol aglycon (Fig 3). The organic acids including caffeic acid, ferulic acid, and coumalic acid are products of the phenylpropanoid biosynthesis pathway. The components of saccharide including glucose and rhamnose are products of carbohydrate metabolism pathways, such as starch and sucrose metabolism, amino sugar and nucleotide sugar metabolism, fructose and mannose metabolism, etc. However, the biosynthesis pathway of phenylethanol part is not clear yet. Here, we proposed two possible phenylethanol biosynthesis pathways based on our sequence data. One is the reported caffeic acid or ferulic acid pathway, also known as the cinnamic acid pathway which is similar to the lignin biosynthesis backbone pathway. Another is based on the phenylalanine metabolism pathway (Fig 3), in which the phenylalanine to phenylethanol was achieved by a known 'Enrlich pathway' that was first found in yeast one century ago and validated in petunia flowers, tomato, and rose. Four enzyme genes encoding aspartate/tyrosine aminotransferase, histidine-phosphate aminotransferase, and primary-amine oxidase which are responsible for the conversion of phenylalanine to phenylethanol were detected expressed in the stem of C. deserticola. The product of phenyl ethanol may be further oxidized by monooxygenase or methylated by methyltransferase into its derivates (phenyl ethanol aglycon) which take part in PhG biosynthesis. In summary, two putative biosynthesis pathways of phenylethanol aglycon were proposed for C. deserticola but still need more study in further.

Discussions 

In recent years, plant genomics has developed rapidly with the application of next-generation sequencing technology, while few researches have been focused on the genomics of desert medicinal plants. It is urgently necessary to perform genomic or transcriptomic research to understand its adaption to drought and salinity environments and the biosynthesis pathway of the major bioactive components. The de novo transcriptome discovery for some medical plants,

such as Panax ginseng, Ginkgo biloba, and Glycyrrhiza uralensis have been first exploited using the Roche 454 platform for its long read length. Because of the effective assembly ability with short reads, especially the advantaged paired-end reads, Illumina-based transcriptome sequencing and assembly have also been extensively used for model and non-model organisms. In the present study, we generated about 8G of 101 bp paired-end reads and produced longer unigene sequences with 725 bp average length. Large-scale stem-specific transcriptome data could provide useful reference data and be used to mine the secondary metabolism of bioactive components of C. deserticola. There is 81.62% of the total raw reads passed stringent quality filters (including adaptor trimming and low-quality reads discarding) before assembly, suggesting the high quality of our sequencing data, and 82.08% of the high-quality reads were useful for assembly. Other reads that failed to be used for assembly may come from sequencing errors, assembly parameters et al. Those unused high-quality reads remained helpful in improving de novo assembly combined with longer reads from another platform (such as Roche 454) in the future.

NATURAL CISTANCHE TUBULOSA FOR IMPROVING SEXUAL FUNCTION PHGS75% ECH 30% ACT 12%

A great number of assembled transcripts (30,098) showed high sequence similarities to known genes in public databases, suggesting that our Illumina-based paired-end data covered a substantial fraction of transcripts of C. deserticola. Transcripts with no BLAST hits may be due to 3' or 5' untranslated regions, non-coding RNA, or new gene sequences of C. deserticola. The expressed transcripts were annotated to a wide range of GO categories and KEGG pathways (Tables 3 and 4), in which many transcripts were assigned to secondary metabolism-related pathways. As we know, phenylpropanoid can function as an inducible antimicrobial compound with great salutary for an underground lifestyle [1], and also act as a signal molecule in plant-microbe interactions besides its medicinal utility [68, 69]. Terpenoid is used for biosynthesis of bioactive components (such as 6- deoxycatalpol) [70]. We found genes involved in the phenylpropanoid and terpenoid backbone biosynthesis pathway were highly abundant in C. deserticola. More importantly, the discovery of well-represented pathways of lignin biosynthesis (Fig 3) indicated the active metabolic process of lignin in the C. deserticola stem. All known enzyme genes involved in the biosynthesis of lignin (Fig 3) were detected expressed, and four key enzymes including PAL, CCR, HCT, and F5H had lower expression abundance (FPKM 26.47, 3.89, 3.4, and 3.83, respectively) compared with other enzyme genes (Table 6). Whether or not the expression change of those three genes could influence lignin production in C. deserticola is worthy of further study. PAL is a key enzyme in lignin biosynthesis and is also involved in the biosynthesis of phenylpropanoid, resveratrol, flavonoid, and coumarin [71–74]. We detected four distinct PAL genes in the C. deserticola genome (S2 Fig) which was coincident with that PAL was encoded by a small multigene family [39, 43, 45–49] and further proved it may play important roles in metabolic carbon flux.

PhG is the primary active ingredient in C. deserticola. Genes involved in the biosynthesis of phenylethanol are important for the quality of C. deserticola. We deduced two different biosynthesis pathways of phenyl ethanol and 17 enzyme genes involved in PhG biosynthesis in the C. deserticola stem. The possible post-caffeic/ferulic acid processes (Fig 3) were also deduced for the first time based on a structural formula of intermediates and catalytic properties of corresponding enzymes, in which the caffeic/ferulic acid would be first oxidized into phenylpyruvate derivate; then, the carboxyl group was deprived by decarboxylases; finally, aldehyde group was converted back into alcohol group by dehydrogenase. This is the first application of Illumina paired-end sequencing technology to investigate the whole transcriptome of C. deserticola and to assemble RNA-seq reads without a reference genome. This study will provide useful resources and gene sequences for functional genomics and proteomics research on C. deserticola in the future.

Conclusions 

In this study, we profiled the transcriptome of C. deserticola stem based on high-throughput sequencing data, identified genes involved in biosynthesis pathways of lignin, and also inferred the potential biosynthesis pathway of PhGs for the first time, which will certainly accelerate the understanding of the ambiguous physiological processes and the great medicinal value in molecular level. Up to now, this is the first attempt to de novo assemble the whole transcriptome of C. deserticola stem and to detect the biosynthesis pathway of medicinal components using Illumina-based sequencing datasets. Our study may promote the development of natural medicines and the selection of cultivars with medicinal traits.

NATURAL CISTANCHE TUBULOSA FOR IMPROVING SEXUAL FUNCTION PHGS75% ECH 30% ACT 12%