Dragon’s Blood from Dracaena cambodiana in China: Applied History and Induction Techniques toward Formation Mechanism

Dragon’s blood that is extracted from Dracaena plants has been widely used as traditional medicine in various ancient cultures. The application of dragon’s blood has a cherished history in China, even though the original plants were not discovered for some period. Dracaena cochinchinensis and Dracaena cambodiana were successively discovered in southern China during the 1970s–1980s. In the last half of the century, Chinese scientists have extensively investigated the production of dragon’s blood from these two Dracaena species, whereas these results have not been previously systematically summarized, as in the present paper. Herein, we present the applied history in ancient China and artificially induced technologies for dragon’s blood development based on these two Dracaena species, in particular, using tissue cultures seedlings and tender plants of D. cambodiana. Big data research, including transcriptomic and genomic studies, has suggested that dragon’s blood might be a defense substance that is secreted by Dracaena plants in response to (a)biotic stimuli. This review represents an effort to highlight the progress and achievements from applied history as well as induction techniques that are used for the formation of dragon’s blood that have taken place in China. Such knowledge might aid in the global conservation of wild Dracaena species and contribute to understanding dragon blood formation mechanisms, eventually assisting in the efficient utilization of limited Dracaena plant resources for the sustainable production of dragon’s blood.


Introduction
Dragon's blood, a crimson resin that is exuded from the injured branch or trunk of Dracaena plants from Asparagaceae family [1,2], has been broadly utilized in the histories of many cultures as a traditional medicine for curing fractures, wounds, diarrhea, piles, and stomach and intestinal ulcers [3,4]. Modern pharmacological research has demonstrated that dragon's blood also has anti-inflammatory, antimicrobial, antioxidant, antitumor, and cytotoxic activities [4]. Further studies have indicated that flavonoids, saponins, terpenes, and steroids that are biosynthesized in plant secondary metabolism pathways are some of the main pharmacodynamic compounds found in dragon's blood [5][6][7][8][9][10].
The sources and identification of dragon's blood have always generated great confusion. Being used as a traditional name for many different resins, dragon's blood has been described in the

Dracaena cochinchinensis
In 1972, the team of Prof. Xitao Cai from Kunming Institute of Botany, Chinese Academy of Sciences, found D. cochinchinensis in Yunnan Province of China and determined its dragon's blood to have equal medicinal efficacy to that sourced from other Dracaena species that were in use. The sapling of D. cochinchinensis can increase by around 10-20 cm in nature, or approximately 23-29 cm under cultivation conditions, every year [18]. The ensuing investigation indicated that D. cochinchinensis was distributed in the subtropical and tropical regions between approximately 21.5 • -23.6 • N, such as in southern China, Vietnam, Cambodia, and Laos [19]. A large number of D. cochinchinensis plants were found in southern Yunnan and Guangxi Province of China in the 1980s, being distributed at altitudes that ranged from approximately 400 to 1700 m. Subsequently, the dragon's blood from D. cochinchinensis and its associated medicines were also developed and named as dragon's blood of Guangxi Xue Jie or Long Xue Jie [20].

Dracaena cambodiana
Dracaena cambodiana was discovered in China during the investigation of sustainable traditional medicine resources in the 1980s, and this Dracaena species was distributed in jungle, stone cracks, cliffs, or desert islands in Hainan Province and southeast Asia, for instance, in Vietnam and Cambodia [21]. As a cork tree, the D. cambodiana plant could not be used as firewood or furniture, and a significant level of moisture was stored in its stem and branch, thus showing excellent drought tolerance. These immense adaptations indicated that the population of D. cambodiana predominantly formed in or around tropical forests and the individuals of D.cambodiana tree are often several hundred years old [18].

Conservation of D. cochinchinensis and D. cambodiana in China
Dracaena plants have been exhausted for their economic and medicinal importance in the period of 1980s~2000s, and many of them were considered to be endangered. In 1998, many of the Dracaena species were list in the Red List of IUCN (The International Union for Conservation of Nature) [22][23][24]. Soon afterwards, wild D.cochinchinensis and D.cambodiana species in China had been list as endangered species in 2001 and were then prohibited to felling. The stem of these two Dracaena species thicken slowly and only the 30~50 years old trees can produce a small amount of dragon's blood [25]. There are the two main causes of raw material of dragon's blood in China relying on import commerce. D.cambodiana is also a horticultural plant species and its tender plant is widely cultured in chamber as bonsai or planted in courtyard as an ornamental tree [26]. The scientists have been focusing on producing dragon's blood while using these tender of D.cambodiana since Dracaena plants discovered in China and many inducing experiments have been performed. These artificial inducing technologies might contribute to reduce the demand for wild harvested dragon's blood.

Wounds and Dragon's Blood Formation
The initial research of Prof. Cai revealed that artificial wounds could promote red resin accumulation in the stem of D. cochinchinensis; hence, the wound became a controlled experiment for further studies on dragon's blood formation in Dracaena species that is still used now. This dragon's blood has both similar chemical constituents and clinical curative effect as dragon's blood from Africa, but its constituents are different to dragon's blood from Southeast Asia, although it has same clinical efficacy as Indonesian dragon's blood that is extracted from the fruit of Daemonorops species [27]. Further textual criticism demonstrated that the dragon's blood that was used throughout the ancient history of Traditional Chinese Medicine was extracted from the Dracaena species planted in the regions of Africa or West Asia, and that dragon's blood from Daemonorops species was the succedaneum of African dragon's blood [28,29].

Microorganisms and Dragon's Blood Formation
Previous studies in model plants have indicated that defense components, such as benzoxazine, camalexins, capsidiol, momilactone, piceids, sakuranetin, and resveratrol, can be induced by microorganisms, and that many of them are flavonoids [30][31][32][33][34]. Modern natural product chemistry has demonstrated that dragon's blood is composed of flavonoids, sterol, lignin, stilbene, and saponin, and that flavonoids are the major constituents [35][36][37]. Microorganisms may also be involved in the formation of dragon's blood. The earliest studies on the relationship between microorganisms and dragon's blood focused on D. cochinchinensis. 303 fungi strains, divided into 23 genera, were isolated from xylem containing dragon's blood from D. cochinchinensis plants, and 52% of strains were Fusarium species. Of these Fusarium strains, 38% were identified as F. graminearum Schw. The other Fusarium strains were F.culmorum (W.G.M.) Sacc. (20%), F. tricinctum (Cd.) Snyd. et Hans (18%), F. solani f. sp. pisi (13%), F. sporotrichioides Sherb. (10%), and with an isolation frequency below 5% were Fusarium strains F. oxysporum Schl., F. moniliforme Sheld., and F. lateritium (Nees) emend. Snyder and Hansen. Although Aureobasidium and Cladosporium strains were also isolated, the Fusarium genus was considered as the predominant strain on account of its high isolation frequency, and subsequent studies also revealed that the Fusarium species could induce dragon's blood formation in D. cochinchinensis stems [38]. Furthermore, 172 fungi strains were isolated from the leaf, stem, or roots of D. cambodiana, with the stem showing the highest abundance of Fusarium species [39].
After being infected by F. graminum at 0, 5, 10, and 15 days, 7,4 -dihydroxyflavanone, 7-hydroxy-4 -methoxyflavane, and loureirin A-the main active components of dragon's blood from D. cochinchinensis-continuously accumulated in the stems of D. cochinchinensis-7,4 -dihydroxyflavanone synthesis occurred later than that of 7-hydroxy-4 -methoxyflavane and loureirin A, but its content was the highest. Six months later, the yield of dragon's blood increased to 67%-120% [38]. F. proliferatum and F. oxysporum can also induce dragon's blood formation in the stems of D. cochinchinensis and D. cambodiana. When compared to the wounded only control, the application of two strains of F. proliferatum increased the yield of dragon's blood in the stem of D. cochinchinensis plants by 2.7 and 3.3 times [40], and it effectively elicited dragon's blood formation from the surrounding leaves and in the inoculation spots in D. cambodiana [41]. Their chemical fingerprint showed that the components of artificial dragon's blood from Fusarium species induction were similar to natural dragon's blood [40,41]. Since then, in about 2007, research on dragon's blood in China gradually shifted its focus to D. cambodiana plant, as it is easier to cultivate than D. cochinchinensis plant.

Plant Hormones and Dragon's Blood Formation
Microorganism infection can activate plant hormone synthesis and increase endogenous hormone content, and exogenous hormones might also induce dragon's blood formation in D. cambodiana. The production of dragon's blood increased by 2.57, 1.6, 2.64, and 4.57 times in response to treatment with gibberellin (GA), indole-3-acetic acid (IAA), brassinolide (BR), and kinetin (KT), respectively. Furthermore, the combination of any two of these three plant hormones can also lead to synergistically increased dragon's blood formation. The application of two of the most important hormones in plants, jasmonic acid (JA) and salicylic acid (SA), did not result in any increases in the yield of dragon's blood at any of the tested concentrations. In addition, 2,4-dichlorophenoxyacetic acid (2,4-D) was demonstrated to have a lethal effect on D. cambodiana plants in experiments to induce dragon's blood formation [42]. The application of the common cytokinin 6-benzylaminopurine (6-BA) could also significantly induce dragon's blood production in the branch or stem of three-year old D. cambodiana plants. In experiments examining the interaction of plant growth regulators (PGRs) with tissue-cultured seedlings of D. cambodiana, KT, GA, IAA, 2,4-D, and 1-naphthylacetic acid (NAA) were respectively mixed in Murashige-Skoog (MS) medium. In tissue cultured seedlings, only 6-BA could promote the secretion of loureirin A and loureirin B into MS medium in culture bottles [43], and this effect did not appear to be dose-dependent, but rather illumination intensity-dependent. In the case of tissue-cultured seedlings of D. cambodiana, none of the PGR treatments resulted in the secretion of dragon's blood into medium under dark conditions, including 6-BA, which was able to induce tissue-cultured seedlings of D. cambodiana in order to secrete dragon's blood into medium under illumination intensities of 1000-3000 Lx, and the inducing effect was positively related to illumination intensity [43].

Suppression Subtractive Hybridization
Before the large-scale application of transcriptomics, suppression subtractive hybridization (SSH) was the popular technology in molecular biology for priming genes with different expression levels [46]. In 2005, 431 recombinants were obtained through SSH by comparing the samples with and without the production of dragon's blood in D. cambodiana tissue culture plantlets. Subsequently, 51 fragments showing differential expression were indicated via sequencing and reverse Northern blot. Of these 51 fragments, 12 cDNA fragments were related to metabolism and energy transfer, five cDNA fragments were involved in signaling transduction, five cDNA fragment were related to transcription and translation, one cDNA fragment was related to photosynthesis, and the other 23 cDNAs were not homologous to genes in the Nr database [47]. In this SSH library, the complete open reading frame encoding calcium-dependent protein kinase 2 (CDPK2) was cloned with RACE, and its expression profile was found to be consistent with dragon's blood accumulation in the stem of D. cambodiana via semi-quantitative RT-PCR. Now we know that the CDPKs play vital roles in plant hormone signaling, growth and development, plant (a)biotic stress responses [48], and they even contribute to plant secondary metabolism by activating calcium binding to their calmodulin regulatory domains [49,50].
Frustratingly, the genes that are directly related to flavonoid or terpene synthesis in plants were not detected by SSH [47].

Transcriptome
In 2014, the transcriptome for the accumulation of dragon's blood was generated while using the stems of three-year-old D. cambodiana plants [44]. Three cDNA libraries were constructed, with the stems being injected by the inducer on the D. cambodiana plant at 0, 3, and 6 days, and 266.57 million raw data from Illumina HiSeq 2000 were assembled into 198,204 unigenes while using Trinity, of which 34,873 unigenes were annotated in the public database. Totals of 2724, 1698, and 2155 unigenes were respectively expressed in the treatments for 0, 3, and 6 days, and the DESeq determined 6986, 7106, and 6085 differentially expressed genes (DEGs) corresponding to the respective pairs of 0 and 3 days, 0 and 6 days, and 3 and 6 days. The KEGG (Kyoto Encyclopedia of Genes and Genomes) classification found 76 DEGs to be involved in flavonoid biosynthesis from the phenylpropanoid pathway, including phenylalanine ammonia-lyase (PAL, six DEGs), cinnamate-4-hydroxylase (C4H, one DEG), 4-coumarate CoA ligase (4CL, 18 DEGs), chalcone synthase (CHS, 10 DEGs), chalcone isomerase (CHI, six DEGs), flavanone 3-hydroxylase (F3H, seven DEGs), flavonol synthase (FLS, 10 DEGs), dihydroflavonol 4-reductase (DFR, 16 DEGs), and leucoanthocyanidin reductase (LAR, one DEG), and of these, 34 DEGs are involved in the catalysis of flavanones into flavonols [44]. As another important chemical component, the total concentration of steroidal saponin, was significantly decreased during dragon's blood formation, and we also detected 122 unigenes that were involved in steroidal saponin biosynthesis, 29 of which encoded 24 kinds of enzymes that had complete open reading frames (ORF) and differential expression [51]. The plant chemical diversity of flavonoids and saponin is dependent on their modification through methylation, glycosylation, or hydroxylation. A total of 27 unigenes encoding O-methyltransferase (OMT, 2 DEGs), UDP-glycosyltransferase (UGT, 14 DEGs), or cytochrome P450 (CYP450, 11 DEGs) were significantly upregulated upon treatment with inducer [44].
Flavonoids are synthesized in the plant cytosol [52], whereas the precursors of steroidal saponin can be synthesized in cytosol or plastid [53]. After synthesis, they are all transported into vacuoles for storage, or to other destinations [54]. The analysis of transcript expression demonstrated that 13 DEGs encoding multidrug resistance-associated proteins (ABC, MRP-type) were significantly induced, while only three DEGs encode ATP-binding cassette transporters (ABC, G-type)-the results might suggest that MRP-type ABC transporters play key roles in secondary metabolism transport during the accumulation of dragon's blood. Furthermore, two DEGs encoding H + PPase, two DEGs encoding vacuolar sorting receptor (VSR), four DEGs encoding soluble N-ethylmaleimide-sensitive factor attachment protein receptors (SNARE), 18 DEGs encoding H + ATPase, and 18 DEGs encoding multidrug and toxic compound extrusion protein transporters (MATE) were also found to be significantly upregulated in dragon's blood formation [44]. Previous studies have indicated that glutathione S-transferase (GST) genes contribute to the transportation of flavonoids from the site of cytosolic synthesis to vacuolar accumulation in plant cells. A total of 20 GST genes were identified based on the transcriptome database, and their transcript profiles were strongly correlated with those of genes that were involved in dragon's blood formation and flavonoid synthesis [55].
As essential transcription factors (TFs), MYB, bHLH, and WD40 play essential roles in flavonoid synthesis and transport [56]. Our transcriptomic analysis also detected 86 TFs with differential expression profiles in the process of dragon's blood accumulation, including 41 DEGs of MYB, 33 DEGs of bHLH, and 12 DEGs of WD40, whereas not all of the TFs selected were induced [41]. The ternary complex of MYB-bHLH-WD40 generally regulating the expression of numerous structural genes might cause these asymmetrical expression profiles [57]. Furthermore, a putative bHLH transcription factor was identified, named as DcHLH1. DcHLH1 could activate the transcription of DcF3'H via binding and regulate the promoter activities of DcF3'H, as shown by yeast one-hybrid screening [58]. F3'H (flavonoid 3 -hydroxylase) can catalyze B-ring hydroxylation of flavonoid derivatives at the 3'-position, forming 3 -hydroxylated flavonoids [59,60].
A model for dragon's blood accumulation in D. cambodiana was proposed based on our transcriptome analysis and previous studies in the model plants [61,62]. It assumes that different environmental factors can induce transient fluctuations in cytosolic Ca 2+ levels in the cell of D. cambodiana, interact with Ca 2+ sensors and resulting in calcium signaling, which always involves transcription factors or upstream kinases, such as, for instance, calcineurin B-like proteins (CBLs), Ca 2+ -dependent protein kinases (CDPKs), and calmodulin-like proteins (CMLs). These sensors subsequently activate or restrain transcriptional activators, such as promoters, enhancers, or suppressors, resulting in the activation of the phenylpropane pathway and suppression of the terpene biosynthetic route. Subsequently, the main chemical constituents of dragon's blood are transferred out of intracellular space to respond to various stimuli through three potential transport mechanisms, including vesicle trafficking-mediated transport, GST transport, or membrane transport. Our present and future studies will focus on verifying this hypothesis via multiple approaches involving omics and genome editing techniques.

Organelle Genome Sequencing of Dracaena cambodiana
Plants generally harbor two independent organellar genomes (chloroplast and mitochondria) that provide invaluable resources for a range of functional, evolutionary, and comparative genomic studies [63]. The D. cambodiana chloroplast (cp) genome can be found in GenBank, with accession number MH293451 [64]. The D. cambodiana plant harbors a circular cp genome 156,697 bp in length. It exhibits the typical plant cp genome structure, including two inverted repeat regions (IRs, 26,526 bp) that are separated by a large single-copy region (LSC, 84,988 bp) and a small single-copy region (SSC, 18,657 bp). The base composition of the D. cambodiana cp genome was uneven (30.81% A, 18.46% G, 19.14% C, 31.59% T), with an overall GC content of 37.6%, while those of the IRs, LSC, and SSC regions were 43%, 35.7%, and 31.1%, respectively. The phylogenetic analyses of entire cp genomes indicated that D. cambodiana might be classified into the Asparagaceae family, and this result was supported by the taxonomy systems of National Center for Biotechnology Information (NCBI) and Angiosperm Phylogeny Group (APG) IV [65]. The cp genome sequence of D. cambodiana encodes a total of 113 cp genes, 76 of which are protein-coding genes, while four are rRNA genes, 30 are tRNA genes, and three (infA, matK, and ndhf ) are pseudogenes. Of these genes, atpF, ndhA, ndhB, petB, petD, rpl2, rpl16, rps16, rpoC1, trnA, trnG, trnL, trnK, and trnV contained a single intron while clpP, rps12, and ycf3 have two introns. These intron sequences might contribute to special DNA barcoding development and species identification with an internal transcribed spacer from rDNA, especially ITS2 (Internal Transcribed Spacer 2) [66]. Moreover, the cp genome of D. cochinchinensis was also released [67]. However, until now, none of the mitochondrial genomes of the Dracaena species have been characterized.

Genome SurveySsequencing of Dracaena cambodiana
The genome of D. cambodiana was surveyed with next-generation sequencing on an Illumina HiSeq 2500 platform, and then the data were assembled and roughly annotated, for obtaining a comprehensive understanding of the genetic characterization of Dracaena species and the formation of dragon's blood. The genome size of D. cambodiana was estimated as 1.12 Gb by k-mer analysis, and its assembly was considered to be straightforward with the PacBio platform, due to the 53.45% of repetitive sequences and 0.38% heterozygosis from k-mer distribution. After assembly with SOAPdevnovo [68], the N50s of the contigs and scaffolds were found to be 1.87 and 3.19 kb. The longest scaffold was 348.12 kb. Although this is neither the draft nor final genome of D. cambodiana, it is the first report of the genome-wide characterization of a Dracaena species [69]. The LTR_FINFER [70] indicated that 30.15% (80,584) scaffolds contained a total of 26,246 simple sequence repeats (SSRs) and the mono-, di-, and tri-nucleotides comprised nearly 98% of the SSRs. The SciRoKo software [71] identified 208 types motif and the most abundant motif was hexanucleotide repeats (106 types). The comprehensive analysis of TRE and RepeatMasker [72] indicated that there were 37.11% retroelements and 2.85% DNA transposons in this D. cambodiana genome, with a total of 39.96% TEs (transposable elements).
The annotation with GeneID [73] predicted 53,700 protein-encoding genes in the survey genome of D. cambodiana and 38,162 (71.07%) genes could be mapped in the public database, such as Swiss-Prot, Pfam, GO, KEGG, KOG, COG, TrEMBL, Nt, or Nr [74][75][76][77][78]. Elaeis guineensis, Phoenix dactylifera, and Musa acuminate were the top three hits for species distribution homologues with D. cambodiana when surveying genome annotation in the Nr database, for which the percentages were 27.80% (6,202), 24.76% (5,524), and 11.45% (2,555), respectively. A total of 38,162 predicated genes were clustered for the gene family analysis with Dendrobium officinale, Asparagus officinalis, Populus euphratica, and Arabidopsis thaliana as the model plant. The results showed that 5375 gene families were mutual among five plant species and 1139 gene families were unique to D. cambodiana.
Interestingly, of the 12,510 genes that matched within Gene Ontology [75], 2,620 genes (20%) were classified into response to stimulus, after cellular process, and metabolic process in the biological process. This revelation suggested that the plant defense system might be involved in dragon's blood formation and the hypothesis was then supported by KEGG mapping [74] and the transcriptome [44]. Of these 38,162 annotated genes, 9258 could be mapped to 125 KEGG pathways and 355 (3.83%) genes corresponded to environment information processing, which was much more than the 267 (2.89%) genes that were mapped to the secondary metabolite biosynthesis and the 184 (1.99%) genes that were involved in terpenoid and polyketide metabolism. Furthermore, this hypothesis guided us to reanalyze the transcriptome of dragon's blood formation by screening the expression profiles of genes that were related to plant defense.
Plant cell walls, antioxidants, and immune systems all contribute to plant defense [79][80][81][82]. For the plant's physical defense, cellulose, chitin, pectin, and polysaccharides are the main structural components of plant cell walls [83]. We detected 41 DEGs that were involved in component synthesis of plant cell walls, including 25 DEGs encoding pectin esterase, 11 DEGs encoding chitinase, four DEGs encoding pectate lyase, one DEG encoding cellulase, and four DEGs encoding galactosidase. The physiological and biochemical barriers of plant defense include antioxidants and immune systems [81,82]. As an antioxidant, ROS is not only a defense substance, but also an important signaling factor in plants, and it might be linked to hormone signaling, programmed cell death, and systemic acquired resistance [82]. However, ROS can also injure normal plant cells, so the plant antioxidant system will be primed during a ROS burst to protect healthy cells from redundant ROS [84]. A total of 79 DEGs that were related to the plant antioxidant system were screened in the RNA-seq data, comprising eight DEGs encoding ascorbate peroxidase, 10 DEGs encoding glutathione s-transferase, two DEGs encoding superoxide dismutase, 22 DEGs that were involved in peroxidase synthase, 12 DEGs encoding permease, 21 DEGs related to plant P450 system, and one DEG for synthesis of each antioxidant, such as glutathione, phytoene, or naringenin. As important defense components, JA, SA, proline, and trehalose regulate plant defense via signaling transduction or relieving osmotic stress [85]. We detected seven DEGs encoding lipoxygenase (LOX), two DEGs encoding allene oxide synthase (AOS), one gene encoding allene oxide cyclase (AOC), and six DEGs encoding 12-Oxo-phytodienoic acid-10,11-reductase (OPAR) in the JA pathway, which are all encoding key enzymes in JA biosynthesis [86]. The SA-related DEGs included one PAL (phenylalanine ammonia lyase), three PBS3 (acyl-adenylate/thioester-forming enzyme), one ICS (isochorismate synthase), and six EPS1 (enhanced pseudomonas susceptibility 1) [87]. In addition, three PDG (proline dehydrogenase) and seven TPS (trehalose phosphate synthase) genes might participate in the formation of dragon's blood via osmotic stress.
These initial results that are based on a genomic survey of D. cambodiana are expected to contribute to further progress in the form of genome sequencing and genetic studies of D. cambodiana and may improve our understanding of the mechanisms of dragon's blood formation. However, most importantly, these results demonstrate effective strategies that involve sequencing platforms and assembly software selection for the construction of the final chromosome-scale genome of D. cambodiana.

Conclusions
This paper presents a summary review of the applied history and current progress in understanding the mechanisms of dragon's blood formation based on Dracaena cambodiana in China. Hence, when considering the vulnerable and severely endangered status of the Dracaena tree population that results from the heavy exploitation of its stem for the production of dragon's blood, the research into artificial induction technologies and dragon's blood formation mechanisms based on D. cambodiana in China, as presented here, will provide valuable information to aid in the global conservation of the wild Dracaena species and contribute to understanding the mechanisms of dragon's blood formation, eventually helping us to understand the evolution of flavonoid genes and Dracaena plants.