Artemisinin, an endoperoxide sesquiterpene lactone, is effective against drug-resistant malaria. The ‘high efficacy, fast-action and low-toxicity’ artemisinin-based combination therapy (ACT) is recommended by World Health Organization (WHO) as the best choice for acute malaria (Douglas et al., 2010). Researches show that ACT has a cure rate up to 97% against severe falciparum malaria (Dondorp et al., 2005; WHO, 2012). Artemisia annua L. (sweet or annual wormwood, member of Asteraceae family) is the only plant that synthesizes and accumulates artemisinin (Chinese Pharmacopoeia Commission, 2015; Su and Miller, 2015). A. annua is native of Asia, and now it is widely dispersed throughout the world, including America, Australia, Africa, Bulgaria, France, Argentina, Spain and Hungary (Knudsmark Jessing et al., 2014). The plant has been established as crop in agriculture after the statement of WHO, as a valuable component of combinatorial therapy for malaria (Ferreira, 2007). A. annua is at present the sole commercial source for artemisinin production since the chemical synthesis and biosynthesis is complicated and uneconomical (Corsello and Garg, 2015). Similar to many other secondary metabolites, trace amounts of artemisinin at a range of 0.01 to 1.4% of dry leaf weight of A. annua result in a low ratio of production to demand of this effective drug (Liu et al., 2006). Previous researches show that the Chinese germplasm has higher artemisinin contents than those from Europe, North America, Australia and East Africa (Jain et al., 1996; Zhang et al., 2016).

Chinese scientist Youyou Tu was awarded the 2015 Nobel Prize in Physiology or Medicine for her contribution to artemisinin discovery, which is regarded as a significant breakthrough in 20th century tropical medicine and an important health improvement for people of tropical developing countries (Su and Miller, 2015). Tu screened more than 2,000 Chinese herbal remedies to search for drugs with antimalarial activity. An extract from the wormwood plant A. annua proved especially effective while other species of Artemisia did not contain artemisinin. In Tu’s artemisinin isolation trials, artemisinin content of samples from Beijing Province, Shandong Province, Liaoning Province, and Anhui Province of China, were low, whereas samples from the Sichuan Province, Hunan Province, Chongqing Province of China had high artemisinin contents (Tu, 2009; Zhang et al., 2016). The result suggested that the artemisinin content was variable in A. annua from different geographic origins.

It has been long recognized that plants in general exhibit ecotypic population diversification that reflects influence of environment on a variety of morphological, physiological and chemical traits (Oleksyn et al., 1992; Hancock et al., 2011; Reich et al., 2014). The study of diversification and ultimately speciation is central to evolution and relevant for biology conservation (Weissing et al., 2011). Biogeographic variation has been extensively documented. For instance, Asian cultivated rice consists of two important ecotypes, upland and irrigated, that have, respectively, adapted to either dry land or irrigated cultivation. Upland rice contains abundant drought-related adaptation genes (Lyu et al., 2014). Moringa oleifera Lam. ecotypes show a diverse growth performance and leaf mass production (Forster et al., 2015). Our previous research shows that Panax quinquefolium L. producing in China can be divided into two chemoecotypes: ginsenoside Rb1-Re from outside Shanhaiguan, and ginsenoside Rg2-Rd from inside Shanhaiguan (Huang et al., 2013). However, although evidence shows that ecotypic variation occurs is abundant (Adams et al., 2016; Perreault-Payette et al., 2017), comprehensive characterization of ecologically important intraspecific variation across biogeographic gradients is lacking, greatly hindering progress in the breeding and cultivation of quality varieties. As the demand for artemisinin remains high worldwide, exploring the chemical and genetic intraspecific variation especially knowledge about its genetic mechanisms for high-yielding become more and more important. Public interest has grown in recent decades; however, the market is not always regulated. Consequently, it is also important to develop accurate and efficient authentication methods of differentiating A. annua.

In the present paper, we aim to (i) estimate the chemical variation pattern of A. annua from different regions in China using high-performance liquid chromatography coupled with evaporative light scattering detection (HPLC-ELSD), (ii) reveal intraspecific variation of A. annua based on DNA barcoding analysis using four regions (ITS2, psbA-trnH, matK and rbcL), and differentiate A. annua from its adulterants, and (iii) find possible ecological factors that affect the artemisinin yield. The results will facilitate exploring the genetic basis of chemical variations and developing strategies for the utilization and conservation of A. annua.

As the sole plant producing artemisinin, A. annua has gained much attention during the last several decades. High-quality raw materials face an increase in demand. Information regarding A. annua genetic mechanisms for high-yielding and ecological factors benefiting for producing the plant are central to introduction and cultivation. Given that the variation of secondary metabolite concentration in different plant populations may be largely caused by genetic variation and an interplay of environmental conditions, studies on correlation between chemical constituents variation and genetic diversity can provide better understanding in the ecotypic diversity of medicinal plant (Djabou et al., 2011; Ali et al., 2013; Snyder and Worobo, 2014; Zheng et al., 2014; Xiang et al., 2016a,b). In this paper, a chemical and molecular analysis was combined to reveal the ecotypic population diversification of A. annua.

By analyzing the artemisinin content of 102 A. annua samples from 12 main producing areas in China, we demonstrated that artemisinin yield varied in different areas of China. A. annua grown in the South of the Qinling Mountains-Huaihe River Line had a high artemisinin content, whereas the northern ones had an extremely low artemisinin yields. It can be generally concluded that the artemisinin content reduced from the Wuling Mountain region to the periphery, which was consistent with our previous studies (Huang et al., 2010). In addition, according to the previous research, the artemisinin yield of some provinces located South of the Qinling Mountains-Huaihe River Line, such as Jiangsu province and Zhejiang province were between 0 and 1.00% (Wei et al., 2008; Xiang et al., 2012). And the artemisinin yield of some provinces located North of the Qinling Mountains-Huaihe River Line, such as Shanxi province and Henan province were as low as 0.10% (Liu J.-L. et al., 2007). This proves our conclusion to some extent.

In previous studies, different types of molecular markers have been used to ascertain DNA polymorphism of A. annua. Shi applied RAPD and ISSR to analyze the A. annua germplasm resources, and the results showed that A. annua gathered from Hubei Province of China had a high intraspecific divergence (Shi et al., 2008). RAPD analysis also indicated genetic variation amongst A. annua in the Indian population (Sangwan et al., 1999). By using SRAP markers, Chen et al. found that the A. annua germplasm resources in natural populations of China had an extremely abundant genetic diversity, especially the A. annua germplasm in Chongqing Province (Chen et al., 2011). In EST-SSR analysis, a variety of motifs was detected (Wang et al., 2012). In recent years, DNA barcoding has been suggested as a useful molecular marker technique to complement traditional taxonomic methods for biodiversity inventory construction and fast species identification (Hollingsworth, 2011; Stoeckle et al., 2011; Little and Jeanson, 2013; Nithaniyal et al., 2014; Wu et al., 2015). In the present study, SA contained a high intraspecific divergence in the ITS regions, especially samples from Wuling Mountain region. The NA had a low intraspecific divergence, which was consistent with the distribution pattern of artemisinin. However, a significantly low sequence divergence was observed in the three chloroplast regions investigated (psbA-trnH, matK, and rbcL). One of the possible reasons is that chloroplast genes evolve slowly compared with nuclear genes. The conservation of chloroplast DNA sequences can ensure comparability among groups. Consequently, rbcL sequence is generally applied to the classification of higher taxa above the family level, while matK and psbA-trnH sequences are suitable for identifications at the genus and species level (Hebert and Gregory, 2005; Chase et al., 2007; Kress and Erickson, 2007). The ITS sequence has different copy numbers in different taxa, resulting in high existence of variants in the genome (Alvarez and Wendel, 2003; Nieto Feliner and Rossello, 2007). The abundant variations in ITS can be used in the classification of lower-level taxa (genera, species, subspecies) as far as population level. Our research has confirmed it to a certain extent. Although variations among A. annua from different sampling sites are insufficient to divide into species or subspecies, it reflected an internal variability and DNA polymorphism in A. annua and provide information for the optimal utilization and conservation of A. annua germplasm resource. This work also illustrates the current limitations in the applicability of DNA barcoding to taxonomic surveys. Therefore, further studies are required to protocol improvement and development, particularly design of novel universal primers to extend the barcoding to a broader coverage of plant species.

To identify anti-malaria drugs, A. annua was filtrated from more than 2,000 Chinese herbal remedies. The other species from Artemisia have been proved not to contain artemisinin. After a large-scale national survey, Tu found five main commercial adulterants of A. annua: A. apiacea, A. scoparia, A. capillaries, A. japonica, and A. eriopoda (Dondorp et al., 2005). The adulterants share similar morphological characteristics with A. annua, although they show different pharmacological properties and medicinal values (Xiang et al., 2016b). In the present study, the barcoding ITS effectively identified A. annua from other species of Artemisia, which can significantly reduce the adverse effects on the health of consumers caused by herbal product adulteration and contamination in the traditional herbal medicine market.

The plant content of artemisinin is partly determined by biotic factors, such as germplasm, but also by abiotic factors, such as climate (Hancock et al., 2011). In our PCA analysis, the sampling sites belonging to Wuling Mountain region aggregated closely due to their unique geographical factors. As disclosed in the correlation analysis, humidity and sunshine time were the most distinct positive and negative ecological factors for A. annua, respectively. Temperature, especially temperature in August and July also played an important role in the accumulation of artemisinin. Wuling Mountain region lies in a subtropical monsoon climate that provides suitable temperature and abundant precipitation. Because of the mountains, the region is foggy and lacks sunshine. Consequently, the unique environment facilitates the accumulation of artemisinin. As an annual plant, A. annua blossoms in September. The production of artemisinin usually peaks with flowering. Thus, the plant is very responsive to the temperature in August and July. The result is consistent with our previous study (Huang et al., 2010). Qinling Mountains-Huaihe River Line is one of the most important geographical boundaries in China. The significance of the geographic dividing line is reflected in the boundary of (1) North China and South China; (2) humid areas and semi humid areas; (3) subtropical zones and warm temperate zones; and (4) paddy fields and dry land, namely, rice and wheat main production boundary. Ancient Chinese wisdom indicated that for tangerine plants with similar leaves, tangerine plants grown in the South of Huaihe River would produce tangerines, whereas tangerine plants grown in the North of Huaihe River would produce bitter oranges. China is vast in territory with diversified landforms and mountain ranges and, hence, varying in weather patterns. This diversity of natural ecosystems, climatic conditions and edaphic traits produce abundant diversities among botanical and germplasm resources and contribute to different quality levels of medicinal plants. Environmental stress is an external factor that affects the formation of certain geo-authentic herbs (Zhao et al., 2012). The favorable ecological factors indicated the ecological requirements of A. annua and provided useful guidelines for in vitro and ex situ cultivation.

In this study, a three-dimensional method that integrated chemical detection, molecular identification and ecological analyses was developed for evaluating the ecotypic diversity of A. annua for the first time. By this approach, high ecotypic diversity of A. annua was observed. We demonstrated that the artemisinin content reduced from the Wuling Mountain region to the periphery. We also found high level of intraspecific genetic diversity among A. annua grown in the South of the Qinling Mountains-Huaihe River Line. Humidity and sunshine time were the essential ecological factors influencing the artemisinin content. This work documents the climatic adaptation of A. annua and provides useful guidelines for in vitro and ex situ cultivation. Our intensive research on transcriptomics to test the key enzymes in the biosynthesis of artemisinin are ongoing.

BL and LH collected and authenticated all the samples. LH, LL, SZ, BJ, and SC contributed to the study design. LL and SZ conducted the experiments. LL analyzed all the data and wrote the manuscript. All authors have read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.