Postintroduction evolution contributes to the successful invasion of Chromolaena odorata

Abstract The evolution of increased competitive ability (EICA) hypothesis states that, when introduced in a novel habitat, invasive species may reallocate resources from costly quantitative defense mechanisms against enemies to dispersal and reproduction; meanwhile, the refinement of EICA suggests that concentrations of toxins used for qualitative defense against generalist herbivores may increase. Previous studies considered that only few genotypes were introduced to the new range, whereas most studies to test the EICA (or the refinement of EICA) hypotheses did not consider founder effects. In this study, genetic and phenotypic data of Chromolaena odorata populations sampled across native and introduced ranges were combined to investigate the role of postintroduction evolution in the successful invasion of C. odorata. Compared with native populations, the introduced populations exhibited lower levels of genetic diversity. Moreover, different founder effects events were interpreted as the main cause of the genetic structure observed in introduced ranges. Three Florida, two Trinidad, and two Puerto Rico populations may have been the sources of the invasive C. odorata in Asia. When in free of competition conditions, C. odorata plants from introduced ranges perform better than those from native ranges at high nutrient supply but not at low nutrient level. The differences in performance due to competition were significantly greater for C. odorata plants from the native range than those from the introduced range at both nutrient levels. Moreover, the differences in performance by competition were significantly greater for putative source populations than for invasive populations. Quantities of three types of secondary compounds in leaves of invasive C. odorata populations were significantly higher than those in putative source populations. These results provide more accurate evidence that the competitive ability of the introduced C. odorata is increased with postintroduction evolution.


| INTRODUC TI ON
The evolution of increased competitive ability (EICA) hypothesis suggests that invasive plants may reallocate resources from defense mechanisms into growth as a response to release from an enemy in their new range (Blossey & Notzold, 1995). Abela-Hofbauerová and Münzbergová (2011) found that invasive Cirsium arvense in North America are larger in most size parameters than the native populations of the same species in Europe. However, most studies that evaluated this hypothesis have not differentiated specialist from generalist enemies, and, in introduced ranges, invasive plants may encounter generalist herbivores rather than total enemy release (Cano, Escarre, Vrieling, & Sans, 2009;Muller-Scharer, Schaffner, & Steinger, 2004). Muller-Scharer et al. (2004) refined the EICA hypothesis and proposed that, in introduced ranges, exotic plant species may adjust the allocation of resources from high-cost quantitative defenses (i.e., resisting specialist herbivores) to growth and low-cost qualitative defenses (i.e., resisting generalist herbivores). Nylund, Pereyra, Wood, Johannesson, and Pavia (2012) found that introduced Fucus vesiculosus increased the dosage of phlorotannins as a defense mechanism against generalist herbivores. The leaf total terpene contents in exotic plant species in Hawaii were 135% higher than that in native species, which facilitate them to resist the generalist herbivores from the introduced range, where specialist herbivores were scarce (Penuelas et al., 2010). Several studies have indicated that plant genotypes from introduced ranges had a more effective types of defenses than genotypes from native ranges (Blair & Wolfe, 2004;Joshi & Vrieling, 2005;Oduor, Kleunen, & Stift, 2017;Puritty, Mayfield, Azcarate, & Cleland, 2018;Turner, Hufbauer, & Rieseberg, 2014). Furthermore, Lin et al. (2015), based on features such as low root-shoot ratio, thin leaves, low leaf cell wall protein contents, and low leaf mass area, proposed that invasive Jacobaea vulgaris had poorer structural defense mechanisms than native genotypes.
Many plant secondary metabolites may act as defense mechanisms against herbivores and have allelopathic effects; if evolutionary mechanisms generate an increase in qualitative defenses against generalist herbivores, the allelopathic effect on indigenous plants may also be strengthened, eventually leading to the increase of the competitive ability of invasive species. Leaf extracts from the invasive Chromolaena odorata in China exerted stronger inhibitory effects on the germination of indigenous plants than the native populations of the same species from Mexico (Qin et al., 2013). Introduced C. odorata had higher resistance to three generalist herbivore species and higher tolerance to simulated herbivory (by shoot removal) than plants from native populations (Liao, Zheng, Lei, & Feng, 2014). Zheng et al. (2015) found that the concentration of odoratin (Eupatorium), a unique compound found in C. odorata with both allelopathic and defensive activities, in the introduced Chromolaena odorata was 2.4 times higher than that from the native range. The introduced population of Taraxacum officinale in the Chilean Andes produced more phenols and anthocyanins as a defensive response to herbivory than the native population in the French Alps (Gonzalez-Teuber, Quiroz, Concha-Bloomfield, & Cavieres, 2017).
Most common garden experiments that tested the EICA hypothesis did not include founder effects among the considered factors, which may have led to misleading conclusions, as source populations are only a fraction of the genotypes among native populations (Dlugosch & Parker, 2008). If the invasive populations are introduced from only one or a few native populations with stronger competitiveness, evidence supporting the EICA hypothesis would be found. However, source populations are weakly competitive, evidence contrary to the EICA hypothesis would be found. Both aforementioned situations could result from founder effects rather than postintroduction evolution. To exclude confounding founder effects, the difference between plants from invasive populations and the ones from their source populations should be compared (Williams & Fishman, 2014). For example, Sakata, Yamasaki, Isagi, and Ohgushi (2014) proposed that the features higher resistance, sexual reproduction, and asexual rhizome reproduction present in introduced populations of Solidago altissima resulted from a long history of pressure by Corythucha marmorata rather than from stochastic events such as genetic drift and founder effects.
Chromolaena odorata is a plant species native to North, Central, and South America, but is a noxious invasive perennial herb or subshrub throughout much of Asia, Oceania, and Africa. It was first introduced into India as an ornamental plant in the middle of the 19th century and has now become one of the most invasive species in southern China (Xie, Li, Gregg, & Dianmo, 2001). There are more than 200 arthropod enemies attacking C. odorata in its native range, and a quarter are specialists; however, some generalist herbivores are documented for Chromolaena odorata in invasive ranges, where specialists are absent (Zhang & Feng, 2007).
Although C. odorata has been introduced into Asia for nearly 100 years, the route of the spread of C. odorata throughout Asia is not well known. A study applying three DNA fragments and six pairs of microsatellite markers (SSRs) revealed that C. odorata in Asia originated from Trinidad and Tobago and adjacent areas in the West Indies (Yu, He, Zhao, & Li, 2014). However, Paterson and Zachariades (2013) indicated that the samples from Asia showed an affinity with samples from Trinidad, Florida, and Venezuela. Therefore, the sources of C. odorata in Asia could not be confirmed based on the existing researches (i.e., Paterson & Zachariades, 2013;Yu et al., 2014). Yu et al. (2014) indicated that the genotypes in Asia (introduced range) have strong competitive ability, which may facilitate the successful invasion of C. odorata.
Therefore, it is reasonable to test the degree to which adaptation contributed to the higher competitive ability of C. odorata plants by comparing the invasive populations with the putative source populations.
During the invasion process, C. odorata strengthens its defense mechanisms against generalist herbivores (Liao et al., 2014) and enhances allelopathic effects (Qin et al., 2013). However, Liao et al. The hypotheses in this study are 1) the genetic diversity of C. odorata in native ranges is higher than that in introduced ranges; and 2) selective pressures in the introduced range cause an increase in  the plant's growth rate and secondary compound production, which in turn increase its competitive ability. Evolution is predicted to increase the production of the secondary compounds active in defense and allelopathy, and to enhance growth traits.

| Plant materials
Chromolaena odorata weeds were collected in the species' native regions in North America and the Caribbean and in the invasive ranges in Asia (Table S1). From each place (defined as a population), 10 plant seeds at least 20-m intervals between any two plants were randomly selected and collected. A total of 10 and 12 geographical populations were collected from the invasive and native range, respectively. The seeds of various groups (populations) were seeded in the nursery bed.

| Genetic analyses
Total genomic DNA was extracted from leaf tissues of C. odorata following the modified cetyltrimethylammonium bromide (CTAB) method described in Yu and Li (2011). In this study, 11 pairs of SSR primers were used to investigate the genetic diversity of 218 individuals from 10 introduced populations and 12 native populations (Table 1). The 11 PCR primers used are described in Table S1. The  (c, d, g, h). Panels a, c, e, and g represent plants grown at high nutrient level; panels b, d, f, and h represent plants grown at low nutrient level. Panels a, b, c, and d represent comparisons between invasive (n = 10) and native (n = 12) regions; panels e, f, g, and h represent comparisons between invasive (n = 10) and putative source (n = 7) regions. Striped columns represent Florida species. Narrow bars indicate mean + SE for each population (n = 10); central thick bars indicate mean + SE for each region (n = 10 for invasive; n = 12 for native). Significant differences between ranges according to one-way nested ANOVAs: * = p < .05; ** = p < .01; *** = p < .001

2013), which is important because intraspecific competition elimi-
nates the potential confounding effects of using a heterospecific as a "phytometer." Pots contained a mixture of 60% forest topsoil and 40% river sand. Topsoil was used as a natural supply of macro-and micronutrients, while river sand provided adequate drainage and facilitated the harvesting of fine roots (Liao, Zhang, Barclay, & Feng, 2013). All seedlings were initially grown in shade with 50% irradiance for 4 weeks to facilitate initial survival; after this period, they were grown in full sun.
Two types of nutrient treatments were set: low nutrient with one-time fertilizer (in August 2013; fertilizer with 0.1 g N + 0.1 g P + 0.1 g K/Kg) and high nutrient with three-time fertilizer (in June, July, and August 2013; fertilizer with 0.1 g N + 0.1 g P + 0.1 g K/Kg).
The high and low nutrient treatments were harvested in September and December 2013, respectively. The entire plants (including roots) were oven-dried at 60°C for 72 hr and weighed.
To evaluate relative competition intensity, the competitive response of each population was measured as the percentage change in performance (i.e., biomass) when grown with competition, and the formula was described by Weigelt and Jolliffe (2003): (P comp -P single )/P single × 100,where P single is plant performance when grown without competition and P comp is plant performance when grown with competition. The competitive effect of each population was measured as the percentage change in the performance of its competitor. In this study, P single was the average of all replicates per population per treatment and P comp was the value of the individual replicate.

| Secondary metabolite extraction and isolation
To detect the increase in production of qualitative defense factors, three types of secondary compounds were extracted and verified: high in defense capacity and allelopathy (4`,5,6,7-tetramethoxyflavone and Acutellerin-4`,6,7-trimethy ether); high in defense capacity but low in allelopathy (Isosakuranetin and 3,5-dihydroxy-7,4`-dimethoxyflavone); and low in defense capacity but high in allelopathy  (c, d, g, h). Panels a, c, e, and g represent plants grown at high nutrient level, and panels b, d, f, and h represent plants grown at low nutrient level. Panels a, b, c, and d represent comparisons between invasive (n = 10) and native (n = 12) regions; panels e, f, g, and h represent comparisons between invasive (n = 10) and putative source (n = 7 for putative source) regions. Striped columns represent Florida species. Narrow bars indicate mean + SE for each population (n = 10); central thick bars indicate mean + SE for each region (n = 10 for invasive; n = 12 for native). Significant differences between ranges according to one-way nested ANOVAs: * = p < .05 | 1259 LI et aL.
( Table S1) grown in the common garden at the Xishuangbanna Tropical Botanical Garden, Chinese Academy of Sciences. All leaves were individually dried under room temperature and ground; then, 500 mg of powder was extracted using 50 ml of methanol for 24 hr. Above six chemicals were measured following the assay method described in Zheng et al. (2015), using ACQUITY ultra-performance liquid chroma- The flow rate of the eluent was 0.6 ml/min, the injection volume of the extract was 5 μL, and the column oven was set at 25°C. Conditions for mass spectrometric detection were as follows: Electrospray ionization (ESI) was performed in positive ion mode at 1.8 kV, ion source temperature was 350°C, solvent temperature was 550°C, sheath gas flow rate was 800 L/h, and auxiliary gas flow rate was 150 L/h. All data for 6 chemicals were collected using multiple reaction monitoring (

| Genetic diversity and structure
Genetic diversity in the invasive populations was significantly lower than in the native ones. Number of alleles, expected heterozygosity, and Shannon's information index (I) in native populations were, respectively, 1.8 (F = 13.15, p < .01), 1.1 (F = 7.66, p < .05), and 1.9 (F = 12.65, p < .01) times higher than those in invasive populations (Table 2).

| Common garden pot experiment
When grown without competition, total biomass and height (Figures   2a and 3a) of plants from the invasive range were larger than those of plants from the native range at high nutrient supply, but not at low nutrient level (Figures 2b and 3b). The relative competition intensity in total biomass and height of plants from the native range were significantly lower than those from the nonnative range (Figures 2c,d   and 3c,d).
Regarding the comparisons conducted between C. odorata populations from invasive ranges and their putative source populations, when grown without competition, no significant differences in total biomass and plant height were observed between the two ranges at both nutrient levels (Figures 2e,f and 3e,f). Competition-driven decreases in total biomass and plant height were significantly greater for C. odorata plants from the putative source populations than for those from the invasive ranges (Figures 2g,h and 3g,h).

| D ISCUSS I ON
The level of genetic diversity of C. odorata plants throughout Asia is significantly lower than that in native populations; similar results were reported by Ye, Mu, Cao, and Ge (2004) and Yu et al. (2014).
In introduced ranges, populations of an invader often originate from only few individuals from the native range, and the invasion into a new territory is associated with frequent founder effects, which potentially lead to a decrease in population-level genetic diversity (Sakai et al., 2001;Tsutsui, Suarez, Holway, & Case, 2000;Ye et al., 2004). Williams and Fishman (2014) proposed that the phenotypic divergence between introduced and native-range populations of Cynoglossum officinale was mainly caused by founder effects. It was believed that small founding sizes reduced genetic variation and fitness but did not prevent adaptation if the founders originated from genetically diverse populations (Szucs, Melbourne, Tuff, Weiss-Lehman, & Hufbauer, 2017).
In this study, we found that total biomass and height of plants from the invasive range were larger than that from the native range at high nutrient supply, but not at low nutrient level. A similar trend was proposed for the invasive plant Poa annua, which exerted a competitive effect on the native plant Deschampsia, but only at high N availability (Cavieres, Sanhueza, Torres-Mellado, & Casanova-Katny, 2018). Liu, Zhang, van Kleunen (2018) and Witkowski (1991) found that the increase in biomass in response to nutrient addition for invasive species is higher than for noninvasive species. At low nutrient levels, soil nutrient is a limiting factor for plant growth and may offset the competitive advantage of invasive species. Competition-driven decreases in total biomass and plant height were significantly greater for C. odorata plants from

F I G U R E 4
Comparison between secondary metabolite productions of Chromolaena odorata plants from invasive and putative source populations. Panels a, b, c, d, e, and f represent comparisons between introduced (n = 10) and putative source (n = 7) ranges. Striped columns represent Florida species. Narrow bars indicate mean + SE for each population (n = 10); central thick bars indicate mean + SE for each region (n = 10 for invasive; n = 7 for putative source). Significant differences between ranges according to one-way nested ANOVAs: *** = p < .001 the putative source populations than for those from the invasive ranges, indicating that evolution actually occurred during the invasion process of C. odorata.
Invasive populations of C. odorata are not completely released from enemies. There are more than 200 herbivores in native ranges of C. odorata, and 25% of them are specialists (Zhang & Feng, 2007), whereas in the species' invasive range in China, few generalists and no specialists on C. odorata have been found (Xu, Xiang, Chen, & Peng, 2011). Evolution occurred in C. odorata plants by increasing biomass, while it also increased the secondary chemical production in response to generalists in introduced ranges. Previous studies also reported the production of higher amounts of odoratin (Eupatorium; considered as qualitative defensive compounds) in introduced C. odorata than in native C. odorata (Zheng et al., 2015). Moreover, in the plant species Triadica sebifera, trading off chemical defenses production occurred as a response to a coevolution with novel natural enemies in introduced ranges; this contributed to its successful invasion by enhancing competitive ability (Wang et al., 2012). Consistent with our result, Ridenour, Vivanco, Feng, Horiuchi, and Callaway (2008) proposed that "evolution occurs at increasing competitive ability and defense traits of Centaurea maculosa in introduced range North American." The secondary chemical productions quantified in this study were all flavonoids. Flavonoids are beneficial for the plant itself as physiologically active compounds, stress protecting agents, attractants, or feeding deterrents and, in general, they play a significant role in plant resistance. Acutellerin-4`,6,7-trimethy ether and 4`,5,6,7-tetramethoxyflavone have defense capacity and allelopathic effect, and the concentration of these compounds in introduced C. odorata was significantly higher than that in putative source populations. These results suggested that evolution indeed occurred in increasing production of qualitative defensive compounds in the process of invasion of C. odorata. Zheng et al. (2015) considered odoratin (Eupatorium) as an important qualitative defensive compound that contributes to the successful invasion of C. odorata. Invasive species could enhance their competitive ability through generating many types of compounds which may only have defending herbivores ability or allelopathic effect. In our study, the other two types of extracted compounds-with high defending capacity but low allelopathy (Isosakuranetin and 3,5-dihydroxy-7,4`-dimethoxyflavone) and low defending capacity but high allelopathy (dihydrokaempferol-3-methoxy ether and Kaempferide-4`-methoxy ether)-also contributed to the plant's strong competition ability.
Although some studies attribute the invasion success of exotic species to founder effects or biased introduction (Williams & Fishman, 2014;Yu et al., 2014) odorata in Asia (Qin et al., 2013), but postintroduction evolution is also essential for the species' establishment and expanding in introduced ranges.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data generated during this study are available from Weitao Li on reasonable request.