Plastid genome sequencing, identification of nuclear SNP markers, and quality assessment of medicinal rhizomatous herb Polygonatum odoratum (Asparagaceae) cultivars

Abstract Polygonatum odoratum (Mill.) Druce (Asparagaceae, Asparagales) is a widely cultivated medicinal herb in China. However, this useful herb is understudied despite being known as a medicinal resource with top grade medical and edible properties since long. In this study, P. odoratum and four cultivars were investigated. The variations in morphological characteristics and vegetative phases of each cultivar were observed. For genetic aspect, the plastid genome of P. odoratum varies in length from 154,569 bp to 155,491 bp, containing a large single‐copy region of 83,486–84,459 bp, a small single‐copy region of 18,292–18,471 bp, and two inverted repeats of 26,302–26,370 bp. A total of 131 genes were predicted, including 85 protein‐coding, 38 tRNA, and eight rRNA genes. Genome comparisons revealed a slight variation in the sequence across the five accessions, but two highly variable regions (trnC‐petN and rpl32‐trnL) were detected when comparing the four different cultivars. For the RAD‐seq markers, a total of 33.64 Gb of clean data, with an average value of 1.08 Gb per sample, were analyzed for the presence of single nucleotide polymorphisms (SNPs). Well‐resolved phylogenies of the P. odoratum cultivars are constructed; the nonmonophyletic relationship in the plastome‐based phylogenetic trees, yet monophyletic form in the RAD‐based linkage map suggested possibility of hybrid cultivar for P. odoratum “Dazhu” (GDDZ), which was further supported by morphological observations. Quality assessment based on the standards of the Chinese Pharmacopoeia on Polygonati Odorati Rhizoma (POR) on the four cultivars used in this study recorded that PORs from P. odoratum “Zhongzhu” (GDZZ) met the minimum criteria for the acceptance as raw material for medicinal drug production. This study has provided insights on the morphological variations, genetic background, and medicinal qualities of P. odoratum cultivars that could be explored for future genetic improvement as well as breeding programs of P. odoratum for POR production.


| INTRODUC TI ON
Over the last decade, there has been an emerging demand for Chinese medicinal herbs due to its popularity across the globe. In order to meet this surge of demand, herbs are now mass-cultivated in open fields and chambers instead of relying on wild resources (Zhao-Seiler, 2013). The World Health Organization (WHO) has always been supportive toward the utilization of native medicine (World Health Organization, 2001), with implementing standards in medicinal plant cultivation practice by prioritizing production quality control for medical purposes. This effort resulted to the implementation of good agriculture practice (GAP) certification system for Chinese medicinal farmers, which provides a guide for farmers to grow these herbs as well as a timely revised Chinese Pharmacopeia that archives desired quality and criteria of raw materials that are meant for medical purposes (Zhang et al., 2010).
Although the practice of propagating the herb might appear to be simple, yet the inconsistency in the chemical and biological nature expressed in its extracts or plant parts to the desired quality appear to be arguable. Obtaining information on the morphological, genetic background, and phytochemical analysis of these medicinal plant species is a requisite step to deter such inadequacies to aid improvement of their growth performance and yield quality extracts.
Polygonatum odoratum (Mill.) Druce, commonly called the Solomon's seal, is a perennial herb from the family of Asparagaceae.
It is mainly natural distributed in the East Asia region, including China, Japan, Korea, and Mongolia, and has also been documented in the northern Europe continent and eastern region of Russia. The P. odoratum produces fibrous rhizome, which is an essential ingredient in traditional medicine. The dried P. odoratum's rhizome, also known as Polygonati Odorati Rhizoma (POR), or Yuzhu (jade bamboo in Chinese), is classified as a medicinal resource with top grade medical and edible properties in ancient medicinal plant records (Liu et al., 2008). It is commonly used as an immune conditioning agent that relieves dryness by quenching thirst and promoting secretion of fluid, promotes blood lipids and glucose reduction, and improves myocardial ischemia. The phytochemical compound-rich POR includes alkaloids, flavonoids, glycosides, phyto-hormones, and saponins (Khan et al., 2012;Zhao et al., 2018). It has been recognized by traditional Chinese medicine (TCM) practitioners and is currently archived in the Chinese Pharmacopoeia (Chinese Pharmacopoeia Committee, 2015). The differences in types of polysaccharides in P. odoratum when compared to other Polygonatum spp. suggested that POR cannot be replaced by other Polygonatum spp or closely related rhizomatous herb species due to variation in pharmacological properties (Guo & Tian, 2001). Its public demand as a functional food and TCM for treating diabetes as well as its potency in healing heart diseases and leukemia has led to promoting the cultivation of P. odoratum on a large scale in China.
Planting and harvesting of Polygonatum odoratum are performed annually during the autumn season. As P. odoratum can be propagated sexually (via seeds) and asexually (via rhizome cuttings), the supply for planting material is often sufficient. With the aid of crop breeding techniques, new P. odoratum cultivars are introduced by medicinal plant breeders through selection breeding, with the aim to yield high-quality PORs. Cultivars often use rhizome cuttings to propagate where identical clones can be replicated through asexual means. The morphological and genetic characteristics of these clones should correspond to the main body, in which the desired quality is thought to be conserved. There are two ways to distinguish types of P. odoratum that the cultivars exhibit in markets. One is from its morphological characteristics, such as rhizome shape, size, and taste, and the other is by its place of origin. According to the former technique, there have been claims that there are more than 10 different cultivars recorded in literatures that used as POR resources, mainly from the provinces of Jilin, Guangdong, and Hunan (Bi et al., 2016;Ma et al., 2014;Yang et al., 2009, personal observation). The latter categorizes P. odoratum into six different types of cultivars, namely Xiangyuzhu (Hunan origin), Haimenyuzhu (Jiangsu origin), Xiyuzhu (Guangdong origin), Dongyuzhu (Zhejiang origin), Guanyuzhu (Northeast China origin), and Jiangbeiyuzhu (Jiangsu and Anhui origins), which all exhibit different phenotypic characteristics.
Initially, the major cultivated POR-producing provinces were confined to Jilin, Guangdong, and Hunan though the high demand in POR has led to the expansion in P. odoratum planting sites in neighboring provinces, such as Jiangsu, Zhejiang, Henan, and Liaoning.
While there are slight morphological differences between P. odoratum cultivars (Chen & Zhou, 2010;Yang, 2004), researches have shown that the harvest and quality of POR may vary across planting regions and types of cultivars (Bi et al., 2016;Bu, 2012;Yang et al., 2009 (Ng & Tan, 2015). The misinterpretation during scoring step and lower reproducibility when comparing with codominant markers may result in incorrect genetic evaluation of the targeted specimens (Agrawal & Shrivastava, 2014). Thus, a DNA sequence-based method is critically needed for the characterization of P. odoratum cultivars.
Owing to the rapid development in DNA sequencing techniques, the next-generation sequencing (NGS) technology warrants an opportunity to generate genome-scale information to provide better resolution in genetic diversity and genome evolution analyses at a much lower cost. The plastid genome (plastome) is reported to be highly conserved in its gene contents and displays a maternal inheritance among angiosperms, thus being useful in phylogenetic studies, population genetics, phylogeography, and species characterization.
Mutation hotspots identified in the plastome due to highly variable regions when comparing across several plastomes of closely related species could be developed for phylogeny and species identification at intraspecies or lower taxa levels. On the other hand, the assembly of the complete nuclear genome is a tedious work; thus, researchers have introduced a substitute method via restriction site-associated DNA (RAD) sequencing technique to provide genome-scale information of the nuclear genome. The RAD markers are short DNA fragments adjacent to the selected restricted enzyme recognition site orders (Baird et al., 2008). The technique is versatile to nonmodel and understudied plant species, in which it could provide thousands of genome-scale single nucleotide polymorphism (SNP) sites that are beneficial for construction of phylogenetic trees and marker-assisted selection (Cui et al., 2018). The genome sequences encompass essential information such as plant origin, evolution, development, physiology, inheritable traits, epigenomic regulation, etc., which are basis in molecular breeding of high-yielding medicinal cultivars and molecular farming of transgenic medicinal strains (Hao & Xiao, 2015).
In this study, we conducted assessments to identify the morphological and vegetative phase differences between the four major Polygonatum odoratum cultivars collected from POR farms in Guangdong, Hunan, and Jilin provinces. We also sequenced and characterized the plastomes of each cultivar to construct the phylogenetic tree to identify potential cultivar-specific barcodes to tell apart the cultivars. A RAD-based linkage map of these four P.
odoratum cultivars was constructed to reveal its phylogenetic relationships at the nuclear genome level. Based on the requirements of the Chinese Pharmacopoeia, quality assessment was also carried out on the POR samples derived from the four P. odoratum cultivars to determine their suitability as raw material for medicine production. Findings from this study will be beneficial to explore molecular breeding of superior cultivars of the P. odoratum in the future.   Figure 1). Three additional wild accessions from Guangdong province (GDWA) were collected as outgroups. The samples were dug out together with its soil and transported immediately to the greenhouse in Sun Yat-sen University for morphology observations.

| Plant materials
Fresh leaf samples were collected from the plants, kept in aluminum ziplock bags, and stored in −20°C prior to the DNA extraction.

| Morphological and vegetative observations
Six accession from each cultivar were used for morphological analy-

| SNP calling
Prior to phylogenetic tree analyses, quality filtering and loci assembly were performed using Stacks v1.40 (Catchen et al., 2013) under default parameters. SNP loci that were found present in more than 12 accessions and with allele frequency greater than 0.05 were included for subsequent analysis. To reduce the inclusion of undesirable sequences, RAD tags with depth above 1,000 were excluded.
SNP identification was carried out in the alignment results using populations, in which alleles with a minimum occurrence of 0.6 times were regarded as true polymorphisms.

| Pharmacopeia standard analyses
Assessments were based on the standard provided by the Chinese

| Polysaccharide quantification assay
A total of 1 g powder sample were diluted with 100 ml of water in a beaker. The mixture was reflux-heated for 1 hr before filtered through absorbent cotton. The process was replicated to obtain two amounts of filtrates. The filtrates were then combined and transferred into a 100-ml measuring flask where water was added until it reached a final volume of 100 ml, which was afterward shaken prior isolating 2-ml into 10 ml of ethanol solution. Thence, the mixture was stirred and centrifuged. Upon extraction, the precipitant was dissolved in water and transferred into a 50-ml measuring flask where the solution was topped-up with water until it reached the final volume of 50 ml. Subsequently, 2 ml of the solution was pipetted and added into 1 ml of 4% phenol solution prior to quantification using light spectrophotometer. Using glucose as the reference, the polysaccharide content in the POR sample should be more than 6%.

| Morphological characteristics and vegetative stages of Polygonatum odoratum cultivars
The four Polygonatum odoratum cultivars shared similar morphological features, such as having alternate leaves, terete rhizomes, and inflorescences with 1-4 flowers. However, there were minor differences that differentiated the cultivars, whereby GDDZ had longer stem size that was recorded at a height of between 0.82 and 1.02 m and at the same time it produced a greater average number of leaves, that is, 20 (

| Plastome variations
The GC contents for plastome of the five Polygonatum odoratum ac- LSC region (n = 34), and IRB (n = 21). A total of 106 parsimony informative sites were detected, in which 79 were detected in the LSC region and 29 were located in the SSC region. Both the IR regions were detected in every three sites. The smallest pairwise distance was detected between GDWA and GDZZ (0.000), while the largest pairwise distance was detected between GDDZ and JLGY (0.0022) (Table 3). Plastome comparison via mVISTA displayed high similarity in nucleotide variability among the five P. odoratum accession. When compared to the reference genome, the plastome of GDDZ exhibited a distinct gap in the intergenic spacer region trnE-trnT of the LSC region that was further identified as a unique insertion of a 11-bp repeat motif (Figure 3). Based on the sliding window analysis, the average nucleotide diversity (Pi) was 0.00129.
With the Pi value cutoff point set at 0.04, two highly variable regions were detected at the aligned sequence region at 28,552 bp to 28,701 bp and 114,586 bp to 114,787 bp, which were the intergenic spacer regions, trnC-petN and rpl32-trnL (Figure 4).

| RAD tag generation and SNP genotyping
This study used 31 accessions and generated 37.51 Gb of raw data that produced 250,068,346 paired-end reads (

| Phylogenetic inferences
Polygonatum odoratum was not recovered as monophyletic in our analyses of the complete plastomes, and two groups were recovered with strong branch supports in both the ML and BI trees ( Figure 5).

| Qualification of POR
For fresh POR samples, JLGZ was recorded with the highest percentage in water content, followed by GDDZ, while both the GDZZ and HNXY recorded the lowest water content percentage (Table 5) However, irregularities were recorded for JLGZ POR samples for the percentage of total ash content as well as GDDZ and HNXY POR samples for percentage of polysaccharide content. The requirements for POR were not more than 3% for the percentage of total ash content, while not less than 6% for the percentage of polysaccharide content.

| Morphological diversity in Polygonatum odoratum cultivars
We observed that the Polygonatum odoratum cultivars had similar morphological features but can still be differentiated when carefully compared. We found out that the distinct morphological characteristics that could differentiate these cultivars were their stem sizes and leaf size. Although the size of the rhizome differs between cultivars, we anticipated that the size of the rhizome could be influenced F I G U R E 3 Plastome comparison of four Polygonatum odoratum cultivars using mVISTA under Shuffle-LAGAN mode. Figure legend describes the direction and types of gene regions using color codes. Probability threshold was set at 50%, and the plastome of P. odoratum from the wild (GDWA) was selected as the reference genome F I G U R E 4 Identification of highly variable sites in the plastome sequences of four different Polygonatum odoratum cultivars by the plant spacing (Tiwari et al., 2019); thus, rhizome size is not a useful morphological trait to differentiate the cultivars. Generally, the number of segments on the rhizome is due to the age of the plant (Ohara et al., 2007), while the planting periods are often three years for plant materials from rhizome cuttings and five years for plant materials from seeds in order to obtain an optimum production yield.

| Potential hybrids in Polygonatum odoratum cultivars
The structure of the plastome for Polygonatum odoratum was well conserved across the five accessions used in this study. under a strong bootstrap support, which separated this cultivar from the rest. This suggested that the evolution of nuclear genome HNXY could be independent and there might be an absence of gene flow between HNXY and other accessions.
On the contrary, the sistership between GDDZ and Polygonatum cyrtonema in the plastome-based phylogenetic trees suggested that the maternal genetic material for GDDZ had a closer relationship with P. cyrtonema than P. odoratum. While there was resemblance between GDZZ and P. cyrtonema in regards to large stem size, the number of inflorescences and rhizome shapes were not.
GDDZ was 2-4-flowered and came with terete-shaped rhizomes whereas P. cyrtonema was 2-7-flowered and had moniliform rhizomes. We note that the basic chromosome numbers for P. cyrtonema and P. odoratum are different, in which P. cyrtonema is x = 9, or 11 and P. odoratum is x = 8, 9, 10, or 11 (Chen & Zhou, 2005;Zhao et al., 2014). Generally, hybridization involving progenitors with two different basic chromosome numbers will lead to sterile hybrids (Shokrpour, 2019). The hypothesis that GDDZ could potentially be a hybrid of P. cyrtonema and P. odoratum was further inclined when we were informed by the farmers that GDDZ was propagated through rhizome cuttings. However, the small sampling size in this study may give a false indication regarding the presence of hybrid, based solely on the results of molecular experiments. Therefore, additional experiments to validate these claims on GDDZ are necessary to provide useful information for further genetic studies. All six GDDZ were categorized within the same clade along with the GDWA accession and a fraction of GDZZ.
Based on the molecular placement of GDDZ in both the plastomebased and SNP-based, we speculate that GDDZ could be a hybrid cultivar between P. cyrtonema and the wild P. odoratum accession from Guangdong. Even though GDZZ could strike off as a potential genetic material for GDDZ, we disregarded the probability because in the province of Guangdong, GDDZ are often grown adjacent to GDZZ within the same locale. Despite the rare and late pollination occurrence of GDZZ with the early blooming of GDDZ in the same field, there might be occurrence of gene flow. Since GDZZ and JLGY were from different regions, thence we were not able to postulate possibilities of gene flow between these two cultivars although the two cultivars were clustered under the same clade in the SNP-based ML tree but were well resolved in the BI tree. The strong sistership between JLGY and P. humile revealed from the plastome-based and the SNP-based phylogenetic trees suggested that they could be closely related to each other, while both taxa shared identical morphological characteristics-P. humile is distinguishable from P. odoratum by having hispidulous leaves abaxially and P. humile is typically 1-flowered compared with P.
An attempt to construct the ITS-based phylogenetic trees for Polygonatum odoratum accessions in this study resulted in both the ML and BI tree displaying weak backbone structure (Figure 7).
Although a handful of SNPs were present in the sequence alignment  Figure 8), the ITS sequence was not powerful enough to resolve the phylogenetic relationships within P. odoratum. Consequently, similar events were reported for other genera in the family Asparagaceae, that is, Asparagus (Fukuda et al., 2012), and Ophiopogon , while the combined data of the plastid intergenic spacer region, petA-psbJ, and ITS sequences are only capable to recognize at best the three distinct sections within Polygonatum (Floden & Schilling, 2018).
In this study, the identity of the SNP loci by SNP calling was not reported as several attempts were reverted with intergenic spacer regions and also largely due to a reference genome on Polygonatum was unavailable at the moment. By means of the reference genome from closely related species, Asparagus officinalis, failed to provide sufficient information in this study, while RAD tags for A. officinalis containing SNPs that complemented to Polygonatum odoratum were found to be very limited (data not shown). Therefore, the proposal to obtain a draft genome for P.
odoratum could certainly aid in the gene identification of desired traits. The estimated genome size for P. odoratum was recorded to be 9,613.73 Mbp (Siljak-Yakovlev et al., 2010), which will undeniably require an enormous account of resource and budget to sequence the full nuclear genome. Hence, it was suggested that the desirable genome coverage when performing analysis with RAD-seq data should be based on its sequencing rigor in order to reveal sufficient informative sites (Feng et al., 2018); the average raw sequence data for P. odoratum samples yielded 9.262 Mbp (Table 4). Despite the low genome coverage rate, the SNP-based phylogenetic trees were able to provide better species resolution when compared to the ITS-based phylogenetic trees which reported higher bootstrap supports along the backbone of the SNP-based phylogenetic tree.

| POR production for TCM resource
Polygonatum odoratum is a well-known traditional medicinal plant recognized for its potential genes in regulating both the polysaccharide and isoflavonoid biosyntheses that produce polysaccharides and isoflavonoids for pharmaceutical purpose (Zhang et al., 2020), as they are as natural chemicals in nutrition for general well-being Miadoková, 2009). In China, the quality standard criteria of raw materials for herbal products are defined in the Chinese Pharmacopoeia, which includes prior testing for characterization and standardization of raw materials before being produced as a medicinal drug. In our study, only the GDZZ POR samples met the minimum criteria of the quality standards set by the Chinese Pharmacopoeia. It is a common phenomenon that distinct spatiotemporal variations in phytochemical profiles in medicinal plants exist (Dhami & Mishra, 2015).
Nonetheless, it is important to note that the Pharmacopoeia's standards are designated for quality control of POR solely for medicinal use or herbal drug productions. Although POR contains a small amount of toxicity (Liu et al., 2008), it is still considered to be safe and can be consumed as a supplement (Chau & Wu, 2006).
The Chinese community has a strong belief in the effectiveness of herbal diet to prevent sickness and diseases (Chen & Yang, 2008;Luo et al., 2019). POR are often purchased and consumed as an ingredient in the preparation of herbal soup as well as in tea and wine making

| CON CLUS ION
In this study, we reported on the morphological variations and differences in vegetative phases of four major Polygonatum odoratum cultivars in China. From the aspect of genetics, we sequenced novel complete plastome sequences of these P. odoratum cultivars and investigated their phylogenetic relationships at genome-scale level for both plastid and nuclear inheritance. As for the medicinal aspect, we assessed the medicinal qualities of the PORs derived from these cultivars based on the criteria described in the Chinese Pharmacopoeia. Outcomes of this study revealed that these cultivars can be differentiated morphologically; the polyphyletic placement of P. odoratum based on the maternally inherited complete plastome sequence, but monophyletic through the biparentally inherited nuclear SNP-based RAD linkage map suggested the possible presence of hybrids among the cultivars; P. odoratum cultivars with PORs that met the criteria described in the Chinese Pharmacopoeia can be considered for selection of sustainable POR production in TCM application. As a valuable medicinal herb that is popular in the Chinese society and nearby regions, these findings would be helpful for future molecular breeding and selection of superior cultivars for POR production in China.

ACK N OWLED G M ENTS
The authors thank Ms.

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to declare.