Article | Open Access | Published:

Ecology and genetics affect relative invasion success of two Echium species in southern Australia

Scientific Reports volume 7, Article number: 42792 (2017) | Download Citation

  • An Erratum to this article was published on 16 March 2017

This article has been updated


Echium plantagineum and E. vulgare are congeneric exotics first introduced to Australia in the early 1800 s. There, E. plantagineum is now highly invasive, whereas E. vulgare has a limited distribution. Studies were conducted to evaluate distribution, ecology, genetics and secondary chemistry to shed light on factors associated with their respective invasive success. When sampled across geographically diverse locales, E. plantagineum was widespread and exhibited a small genome size (1 C = 0.34 pg), an annual life cycle, and greater genetic diversity as assessed by DNA sequence analysis. It was found frequently in areas with temperature extremes and low rainfall. In contrast, E. vulgare exhibited a larger genome size (1 C = 0.43 pg), a perennial lifecycle, less chloroplast genetic diversity, and occurred in areas with lower temperatures and higher rainfall. Twelve chloroplast haplotypes of E. plantagineum were evident and incidence aligned well with reported historical introduction events. In contrast, E. vulgare exhibited two haplotypes and was found only sporadically at higher elevations. Echium plantagineum possessed significantly higher levels of numerous pyrrolizidine alkaloids involved in plant defence. We conclude that elevated genetic diversity, tolerance to environmental stress and capacity for producing defensive secondary metabolites have contributed to the successful invasion of E. plantagineum in Australia.


Introduced species are of global concern in terms of their inherent economic and environmental costs, with annual losses of USD $1.4 trillion associated with biological invaders around the world1. Australia has endured the ravages of numerous noxious invaders [e.g. prickly pear cactus (Opuntia stricta (Haw.) Haw), cane toad (Rhinella marina L.), carp (Cyprinus carpio L.) and European rabbit (Oryctolagus cuniculus L.)], many of which were intentionally introduced from overseas. In terms of invasive Australian plants, agricultural costs of weed management alone are reported to exceed $4 billion annually and all of the most noxious weeds are non-indigenous2.

Successful plant invaders often rapidly adapt to novel ecosystems3. This can be achieved through rapid occupation of an empty niche4, “evolution of increased competitive ability”5, increasing colonizing ability6, production of large amounts of viable and long-lasting seeds7, a lack of enemies in the non-native range8, synthesis of allelochemicals that promote invasion (‘novel weapons’)9 and modification of local above- and below-ground environments7,9. One particularly successful plant invader, Echium plantagineum L., is self-incompatible in its native range but purportedly became self-compatible after introduction to Australia10, providing further evidence for the hypothesis of increased colonizing ability. Although uniparental reproduction may result in inbreeding depression11, annual self-compatible invaders may operate more independently from conspecifics and ancestral pollinators10.

Evaluation of evolutionary genetics of invaders is critical to develop a better understanding of the mechanism(s) associated with invasion success. With regards to successful plant invasion, sufficient levels of genetic diversity are typically required for species persistence and evolution in a dynamic environment12,13. High levels of genetic diversity may contribute to adaptive potential and resistance to environmental stress, including management practices. However, a plant invader commonly establishes initially with limited genetic variation, as most invasions are derived from small founder populations14. Invader populations may subsequently increase in genetic diversity over time via introduction of new genotypes, rapid evolution and/or cross-species hybridization7,15. Dlugosch and Parker16 highlighted the importance of multiple introductions and adaptive evolution for species invasion. For example, the house sparrow and European starling became problematic only after multiple introductions into North America13.

Numerous studies support a negative relationship between plant monoploid genome size and invasiveness17,18,19,20,21. According to “the large genome constraint hypothesis”, smaller genomes are associated with shorter life cycles, smaller seed, greater specific leaf area and higher photosynthetic rates17,18. Cytogenetic analysis of 156 weedy and 2685 non-weedy species indicated that weedy species tend to have smaller genome size (3.79 pg) compared to non-weedy species (12.14 pg)19. Very small genomes (1 C < 1.40 pg) are very common in the most invasive plant species21.

A direct comparison of the genetics and invasion ecology of both successful and less successful plant invaders introduced at similar timeframes to the same or similar location(s) could result in significantly enhanced understanding of the mechanisms that drive invasion success. Therefore, the congenerics E. vulgare L., commonly called Viper’s bugloss, and E. plantagineum, known regionally as Paterson’s curse or Salvation Jane, were chosen as model species in this study because of their similar introduction history, morphology, reproduction and dispersal22. Both species originated in the Mediterranean and have since naturalized in Africa, America, Asia, Europe and Oceania23,24. Echium vulgare is now commonly encountered in Europe and Canada24,25 but in Australia is restricted to the south-eastern states of South Australia (SA), New South Wales (NSW), Victoria (VIC) and Tasmania (TAS)25. In contrast, E. plantagineum is an economically important weed in Australia26 and has invaded 33 million hectares across southern and western Australia, with an estimated annual economic impact of $250 million27. Both Echium species are drought tolerant, can produce up to 10,000 seeds per plant and rely on mammalian activity for dispersal23,24. Unfortunately in Australia, ‘Paterson’s curse’ has sometimes been used as a common name for either E. vulgare or E. plantagineum28 so the extent of distribution following establishment in the 1800 s is potentially unclear29.

Depending on seasonal growing conditions, Echium plantagineum can exist either as an annual or biennial. It was reportedly introduced to Australia in the mid-1800 s as an ornamental plant23, but quite possibly was repeatedly introduced with the direct importation of merino sheep from northern Spain30. Echium plantagineum is a native of the Iberian Peninsula and today can be found sporadically throughout the Mediterranean region. In contrast, E. vulgare is reported to be a biennial or short-lived perennial, and is widespread across temperate regions of Europe. It is thought to have been introduced to Australia around 182029.

Echium vulgare and E. plantagineum produce two interesting groups of secondary metabolites important in plant defence: pyrrolizidine alkaloids synthesized in above-ground plant tissues and organs, and naphthoquinones produced in living roots and root hairs31,32,33,34. Pyrrolizidine alkaloids play critical roles in plant defence against grazing herbivores and are present in high concentrations in both E. vulgare and E. plantagineum, thus contributing to livestock toxicity across southern Australia due to their direct consumption23,32. The roots of E. plantagineum and E. vulgare also produce high concentrations of naphthoquinones34, red-coloured compounds referred to as shikonins that are also produced by roots of other members of the Boraginaceae35. Shikonins exhibit potent antimicrobial, antifungal, and phytotoxic properties and are frequently used as biomedicinals in Eastern medicine35. In Australia, exposure to stressful conditions is associated with enhanced production of shikonins and pyrrolizidine alkaloids in E. plantagineum, with increased concentrations observed in plants collected from warmer, drier locations32,33. Other Boraginaceae including Lithosperum L. and Arnebia Forssk. also produce shikonins35. Our recent studies suggest that both families of metabolites contribute to plant defence and may serve as important ‘novel weapons’ in the invasion process36.

Past studies of E. plantagineum and E. vulgare in Australia have focused mainly on pollination ecology and floral nectar production related to quality of commercially produced honey37. However, specific information on comparative morphology38, phenology39,40,41,42, genetics, and biology23,24 is limited. Both species have been sparingly included in broader phylogenetic studies of Echium spp.; thus limited information is available regarding their contemporary spatial distributions11,43,44,45. The most recent study of genetic diversity of E. plantagineum in Australia used isozyme markers to study diversity and suggested a similar level of genetic diversity between Australian and native Iberian populations46. As polymorphisms detected by isozyme markers vary among tissues, growth stages and environments47, and methods of specimen preservation often impact isozyme analyses, further studies are warranted.

To shed additional light on the mechanisms of invasion success of these two congeneric species in Australia, a series of field surveys was performed across southern Australia in locations where both species are now naturalized27. Specimen records from Australian herbaria were evaluated to gain an understanding of the historical introduction of each species to Australia. Geographically distinct populations of both species were surveyed for local climatic conditions and coexisting plant diversity. The hypothesis of evolution of increased competitive ability was tested by measuring qualitative and quantitative differences in secondary metabolite production. We also hypothesized that the invasive E. plantagineum has smaller monoploid genome size and higher level of genetic diversity compared to naturalized E. vulgare.


Geographic distribution in Australia

Results obtained from three seasons of field surveys conducted in southern Australia are in general agreement with historical herbarium records obtained for both Echium species in Australia’s Virtual Herbarium (AVH)48. Echium plantagineum was found to be widely distributed across southern Australia (Supplementary Fig. S1a). However, E. vulgare was found only sporadically, and was narrowly restricted to the South Eastern Highlands (SEH) biogeographic region (Supplementary Figs S1b and S2). We noted 1376 and 174 AVH records of E. plantagineum and E. vulgare, respectively, in Australia. Echium plantagineum is widely distributed from eastern Queensland (QLD) to Western Australia (WA), being recorded in Brigalow Belt North, QLD and also across nearly all of the biogeographic regions (around 40) in NSW, VIC, TAS and SA to Carnarvon, WA. In contrast, E. vulgare was restricted to 17 biogeographic regions, with most records coming from one biogeographic region, SEH, which accounts for 59.2% of total records in Australia (Fig. 1). This species was reported sporadically in only four biogeographic regions since 2000: New England Tablelands (NET), SEH, Ben Lomond (BEL), and Tasmanian South East (TSE) (Table 1, Fig. 1 and Supplementary Fig. S2). Historical records of E. vulgare also indicate past occurrences in TAS, southeastern NSW and VIC, where summer rainfall is more common, elevation typically exceeds 400 m and recorded winter temperatures are below 3 °C. In contrast, some records were noted from SA, western NSW and VIC, where summer rainfall is limited, elevation is lower than 300 m and winter temperatures are generally warmer (Table 1).

Figure 1: Distribution of Echium vulgare in Australia.
Figure 1

Red dots indicate the location of herbarium specimen records of E. vulgare. Dots enclosed by solid lines indicate records obtained from the same biogeographic region. See Table 1 for biogeographic regional codes. Image provided by Australia’s Virtual Herbarium48.

Table 1: Climatic conditions experienced (1955–2014) in Australian biogeographic regions supporting Echium vulgare (Supplementary Fig. S2).

Impact of Echium invasion on plant biodiversity in Australia

From 2011 onwards, it proved particularly difficult to find established sites of infestation for E. vulgare across southern Australia. Four sites infested with E. vulgare were noted and analysed; the density of E. vulgare ranged from 2–67 plants m−2, and averaged 27.0 ± 14.3 plants m−2 (mean ± SEM). In contrast, sites infested with E. plantagineum were easily detected and numerous; the density of E. plantagineum in the most heavily infested quadrat was 275 plants m−2, and averaged 80.9 ± 19.3 plants m−2 for the 17 sampled locations. Plant biodiversity decreased when E. plantagineum was present in quadrats but not when E. vulgare was present. The number of all other species per quadrat decreased from 6.6 ± 0.7 to 4.6 ± 0.5 when E. plantagineum was present (P < 0.01), whereas the corresponding values for E. vulgare were 5.3 ± 1.0 vs 4.0 ± 0.7 (difference not significant at P = 0.05). The density of other plants was more heavily impacted by the presence of E. plantagineum. These values declined from 1271.2 ± 219.8 m−2 in quadrats where E. plantagineum was absent to 689.6 ± 130.2 m−2 when E. plantagineum was present (P < 0.01); corresponding values for E. vulgare were 1018.8 ± 240.7 m−2 and 1043.8 ± 82.5 m−2, respectively (P > 0.5). These measures of biodiversity are not directly comparable because of the higher densities of E. plantagineum observed, but when restricting the analysis to quadrats where E. plantagineum spanned a similar range of densities to E. vulgare (13–76 plants m−2, n = 11), a significant decrease in number of other plants was still observed with increased density of E. plantagineum (1112.3 ± 260 uninfested with E. plantagineum vs 552.1 ± 176.9 infested, P < 0.05).

Pyrrolizidine alkaloid content

Metabolic profiling (using ultra high pressure liquid column chromatography coupled to time of flight mass spectrometry, or UPLC MS QToF) of foliage from geographically diverse field- and glasshouse-grown plant populations of both species resulted in detection of 17 pyrrolizidine alkaloids in E. plantagineum leaf extracts and up to 16 pyrrolizidine alkaloids in E. vulgare shoot extracts (Table S1). This corresponds with recent studies noting up to 17 pyrrolizidine alkaloids in Echium spp. shoot extracts32. Of note is the finding that pyrrolizidine alkaloids occurred in E. plantagineum at levels up to three times those observed in E. vulgare, a result confirmed both in controlled glasshouse conditions and in field sampling when species ranges overlapped near Bathurst NSW (Fig. 2). Three pyrrolizidine alkaloids were consistently less abundant in E. vulgare in all environments: 7-O-acetyllycopsamine-N-oxide B, 3′-O-acetylechiumine-N-oxide and 7-O-acetyllycopsamine.

Figure 2: The relative abundance of pyrrolizidine alkaloids and their N-oxides extracted from E. plantagineum (Ep) and E. vulgare (Ev) foliar tissue, averaged over three biological replications for each treatment.
Figure 2

Data was normalized by log2 transformation. Both species were grown (a) under uniform glasshouse condition or (b) at the same field sites near Bathurst. Pyrrolizidine alkaloids were significantly more abundant in Ep as tested by one-way ANOVA (P < 0.05). Ep: Echium plantagineum, Ev: E. vulgare; Ep-A: E. plantagineum collected from Adelong; Ep-S: E. plantagineum collected from Silverton; Ev-A: E. vulgare collected from Adaminaby; Ev-C: E. vulgare collected from Cooma. Please refer to Table S1 for the name of the compounds.

Genome size and genetic diversity

Monoploid genome size (presented as 1 C value) of E. vulgare ranged from 0.41 to 0.45 pg (mean: 0.43 ± 0.003 pg), while the 1 C value of E. plantagineum ranged from 0.30 to 0.39 pg (mean: 0.34 ± 0.002 pg) (Table 2 and Fig. 3). Results obtained are consistent with the previously reported ploidy level of both species in Europe (2n = 32 for E. vulgare and 2n = 16 for E. plantagineum)49. Neither species showed a change in DNA content with variation in ploidy, nor was there any apparent difference in genome size in geographically distinct locations/populations for each species.

Table 2: Genome size of Australian E. vulgare (Ev) and E. plantagineum (Ep) as estimated by flow cytometry using genome size of radish (Raphanus sativus 1 C = 0.55 pg) for standard comparison68.
Figure 3
Figure 3

Flow cytometry histograms of E. plantagineum (a) and E. vulgare (b) using radish (Raphanus sativus 1 C = 0.55 pg) as an internal reference.

PCR and sequencing analysis were 100% successful for all samples at targeted gene regions; 154 sequences were generated for each gene region under scrutiny. Alignments were truncated to 636, 280, 469 and 399 bp for ITS, trnH-psbA spacer, trnL intron and trnL-trnF spacer, respectively. Four alleles were detected in the nuclear ITS region and two haplotypes were found in the concatenated chloroplast regions of E. vulgare; the corresponding values detected for E. plantagineum included two alleles and 12 haplotypes (Table 3).

Table 3: Genetic diversity of Australian E. vulgare and E. plantagineum, as estimated by allele and haplotype (hap) numbers, nucleotide (π) and haplotype (h) diversity.

Echium vulgare showed a similar level of nucleotide (π = 0.0015)50 and haplotype (h = 0.5444) genetic diversity50 in the nuclear region (ITS) to that of E. plantagineum (π = 0.0008, h = 0.4990) (Table 3). However, considerably lower genetic diversity was detected in the chloroplast regions of E. vulgare (π = 0.0014, h = 0.3800) compared to E. plantagineum (π = 0.0021, h = 0.7661).

Evidence of regional chloroplast population structure in E. plantagineum was noted. The distribution of E. plantagineum chloroplast haplotypes (n = 12) showed strong indication of geographic sorting between western NSW and southeastern Australia (Fig. 4), as indicated by shifts in frequency of haplotype 5 (Supplementary Fig. S3), observed as prevalent in eastern NSW and VIC (54.4%), but less so in western NSW (15.4%). Haplotypes 10–13 were not observed in eastern NSW and VIC, but represented 42.3% of the haplotypes found in western NSW. In addition, haplotypes 6 and 8, present at low frequencies in eastern NSW and VIC (2.2 and 5.6%, respectively), were not found in western NSW. A population pairwise Fst test51 showed a significant (Fst = 0.13, P < 0.001) difference between western NSW and eastern NSW and VIC, which strongly suggests the presence of genetic structure. This population structure was not supported at the nuclear ITS gene, where structure was evaluated using an Fst test (Fst = −0.02, P = 0.85) and 95% parsimony network analysis further indicating that the two nuclear ITS alleles in E. plantagineum were generally present at similar frequencies across sampled regions (Supplementary Fig. S4a). The E. plantagineum chloroplast network analysis suggested no apparent phylogenetic basis for haplotype sorting among regions (Supplementary Fig. S4b). Interestingly, one rare haplotype, 14, was unique to WA.

Figure 4: Distribution of haplotypes of E. plantagineum in southeastern Australia.
Figure 4

The dashed line separates southern Australia into eastern NSW and VIC, and western NSW. (specific haplotypes Hap 6 and 8 are found in eastern NSW and VIC, and Hap 10, 11, 12 and 13 in western NSW, respectively. This map is a derivative of “State and Territory ASGC Ed 2011 Digital Boundaries in ESRI Shapefile Format” sourced from the Australian Bureau of Statistics, used under CC BY 2.5 ( and modified using ArcGIS 10.3.1 software by Esri ( and Adobe Illustrator CS5.


Echium plantagineum was first recorded in Australia in MacArthur Garden, located in Camden, NSW (near Sydney, NSW) and introduction from England as an ornamental is postulated23. It is uncertain, however, whether this introduction event resulted in later escape and naturalisation. In Australia, at least three naturalisation events of E. plantagineum have been documented, one near Albury (NSW), one in Gladstone (near Port Pirie, SA) and one in WA, all in the 1880 s52. Considering the similar timing of these events and the great distance between these Australian locations, it is likely that multiple introductions of E. plantagineum occurred29,52. Distribution of the 12 observed chloroplast haplotypes in Australia noted from our analyses is well-aligned with these reported naturalisation events. Regional specific haplotypes were detected in eastern NSW and VIC (haplotypes 6 and 8), western NSW (haplotypes 10–13) and WA (haplotypes 14) (Figs 4, S3 and S4). Although 90 individuals were sampled, samples from eastern NSW and VIC represented only 7 of the 12 detected haplotypes of Australian E. plantagineum, with two specific haplotypes (haplotypes 6 and 8) occurring near Albury, NSW. In contrast, the western part of NSW, located between the SA and NSW introduction events, contained nearly all of the E. plantagineum haplotypes (9 out of 12, except haplotypes 6, 8 and 14) detected in this survey. It is possible that additional sampling in SA might result in the recovery of additional or specific haplotypes (such as 10–13). The Fst test revealed a significant population structure in chloroplast DNA (P < 0.001) but not in nuclear DNA (P = 0.85). Lack of population structure at ITS may be caused by the paucity of available polymorphism at ITS of Australian E. plantagineum and/or higher migration rates of nuclear DNA in contrast to chloroplast DNA. Plastid DNA is maternally inherited in angiosperms53, which means the cpDNA of E. plantagineum and E. vulgare can move only by seed distribution, while the gene flow of the nuclear region can be attributed to both seed and pollen dispersal54.

Echium plantagineum is apparently less prone to genetic bottlenecks because of its greater adaptability across a variety of habitats. Multiple introductions of E. plantagineum to Australia, evidenced by the population structure in south-eastern Australia, may also have contributed to its high genetic diversity. High genetic diversity is associated with invasion success for many plant species7,55,56,57,58. Careful management of each species in local regions may be critical in future years to avoid seed dispersal across Australia and limit out-crossing that may result in further enhancement of genetic diversity among distinct regional genotypes within each species. In addition, considering that E. vulgare is a weed of importance in Europe25 and Canada24, it will also be critical to avoid new introductions of E. vulgare into Australia that might increase the number of genotypes post-introduction.

The invasive species E. plantagineum possesses a distinctly smaller genome size than the non-invasive E. vulgare (Table 2), which supports the large genome constraint hypothesis17. A small monoploid genome size (1 C < 1.40 pg) is often found at high frequency in invasive species21 and is normally also associated with reduced generation time and seed mass and increased relative growth rate and seed numbers21,59. However, studies on Phragmites australis (Cav.) Trin. ex Steud. suggested that smaller genome size can also potentially reduce plant fitness and defence60. There was no significant difference in genome size among 93 invasive and naturalized species in the Czech Republic61, which suggested that small monoploid genomes may be critical for the initial settlement of alien species but less important after establishment21. Small genome is also correlated with an annual life cycle in some plant genera, including Veronica L.62 and Sorghum Moench63. It is not clear in Australia whether genome size is related to the persistent spread of weedy features and/or phenology/life cycle in the genus Echium. When compared with previously reported data on 13 other Echium species20, monoploid genome size of annual Echium species (1 C DNA content range: 0.30–0.32 pg) is considerably smaller than the perennial Echium species (1 C DNA content range: 0.41–0.43 pg). However, as data from only two annual Echium species (E. bonnetii Coincy and E. plantagineum) has been published, it is speculative to generalise that reduced genome size is associated with a shorter life cycle in the genus as a whole. Polyploidy, often reported as occurring in invasive weeds and suspected of enabling certain species to gain plasticity associated with specific habitat and resource requirements resulting in adaptation to broader environmental parameters64, has apparently not been a factor contributing to variable success of E. plantagineum and E. vulgare in colonising Australia (Table 2).

Echium plantagineum in Australia exhibits a considerably shorter life cycle and produces greater leaf area than does E. vulgare42, and also produces larger seeds (3.6–3.9 mg per seed compared to 2.5 mg per seed for E. vulgare)23,24. A shorter life cycle may facilitate the broader adaptation of E. plantagineum to diverse and variable climatic conditions and thereby facilitate escape from environmental stress. Both species are capable of producing similar numbers of seeds per plant, but a shorter life cycle has potentially enabled E. plantagineum to produce more seed over time, as both species are monocarpic15. In addition, E. plantagineum tended to suppress the number of other species growing in close proximity, as suggested by the density of other plants in quadrats where these species were sampled compared with nearby quadrats where Echium spp. were absent. Echium plantagineum also appeared to achieve greater overall densities than E. vulgare, but this result is not definitive because of the small number of observed sites infested with E. vulgare.

Recent records of E. vulgare were found in only four biogeographic regions (NET, SEH, BEL and TSE), where cold winters and reliable summer rainfall (or high humidity) were common. Echium vulgare has also been reported in six biogeographic regions of SA, western NSW and VIC with warmer winter temperatures and limited summer rainfall (Table 1). A comparison of the recent decade (2005–2014) with the previous 50 years (1955–2004) of climate data (Table S2) shows a clear trend toward increased winter temperatures and more frequent summer rainfall events. Increased summer rainfall is likely to promote germination of both species from the existing seed bank, but probably more so for E. plantagineum since its existing seedbanks are likely more plentiful as discussed aboved24. In addition, without exposure to cooler winter temperatures for vernalisation, E. vulgare may become increasingly less abundant in Australia.

A high rate of germination (>40%) is typically achieved at warmer soil temperatures ranging between 20–30 °C in late spring and summer for E. vulgare, or between 10–30 °C in early spring and summer for E. plantagineum42. Germination of E. plantagineum normally occurs after spring and summer rainfall events in Australia23 whereas optimal germination conditions for E. vulgare in the field are associated with higher soil temperature and moisture availability to support maximal emergence; the seedlings of this species therefore emerge weeks to months later than those of E. plantagineum in the same biogeographic region24,42. Echium plantagineum is also highly resistant to water deficit. Most (57%) E. plantagineum seedlings survived after 2–4 weeks under severe moisture stress in Albury (southeastern NSW)23, and we have also observed extreme tolerance of this species to moisture deficit after withholding water for up to 3 weeks in controlled environment experimentation (unpublished data). We do not know of comparable tests for E. vulgare, but in experimentation performed in Canada, only 18% of seedlings survived their first year of establishment and only 5% of all established seedlings reached reproductive maturity, with many seedlings experiencing mortality due to drought following emergence24,42. In inland Australia, rainfall typically occurs more frequently in winter months, when soil temperatures are generally not high enough to support the emergence of E. vulgare. Both summer and early autumn rainfall events in southern and western Australia may induce germination, but are normally followed by severe periods of drought, which could potentially result in high mortality of E. vulgare seedlings. Echium vulgare also has a vernalisation requirement and requires low temperatures “throughout the winter” to induce flowering in potted plants, while warm summers were necessary for vegetative growth42. Without intermittent exposure to cooler winter conditions, E. vulgare has been observed to remain as a vegetative rosette for 10 years in a continuously warm environment24. These factors would undoubtedly result in lower reproductive success of E. vulgare in much of inland Australia.

The higher abundance of pyrrolizidine alkaloids in the foliage of E. plantagineum may limit feeding by animals, both vertebrate and invertebrate, on this species vis-à-vis E. vulgare. Specialist insects are able to successfully feed on Echium spp. and other plant species containing pyrrolizidine alkaloids, but most generalist insects lack the ability to sequester or detoxify these compounds65. The presence of pyrrolizidine alkaloids is readily detected by native or unadapted insect herbivores66, causing these insects to look elsewhere for feed after sampling foliage. Livestock are known to feed on Echium spp. when other species are scarce, but grazers are also able to detect the presence of pyrrolizidine alkaloids and would thus typically avoid feeding on plants containing them. The greater abundance of alkaloids in E. plantagineum is likely to have a stronger protective effect than the reduced levels found in E. vulgare. In addition to foliar alkaloids, naphthoquinones (shikonins) present in the roots of E. plantagineum and E. vulgare are active against a range of biotic threats including microbiota and neighbouring plants, and their variable production may also contribute to the differential invasion success of these two species31,32,33. Although glasshouse grown E. vulgare plants show higher abundance of shikonins than does E. plantagineum, drought conditions experienced in the field may stimulate increased production of shikonins by E. plantagineum to a greater extent than E. vulgare31,32,33, suggesting that the former may be better defended against herbivores and more competitive under stressful conditions.

Herbarium records were essential in this study for documentation of the historical dynamics of dispersal of the weedy invaders E. plantagineum and E. vulgare across Australia42. However, misidentification of Echium species was and continues to be very common in Australia29, and field surveys are clearly required to verify the current infestation rate of each species. Two multi-year surveys performed over 2011–2015 confirmed the previous records of E. plantagineum invasion across southeastern Australia. However, in contrast to past reports, E. vulgare was found only sporadically in the SEH biogeographic region in eastern NSW during this period. Echium vulgare was generally observed near the edges of roadsides in the southern highlands at higher elevations, but at very low densities. In contrast, E. plantagineum was found broadly distributed along roadsides, railroad tracks, in stockyards and grazing lands, but was normally at very high densities, including monocultural stands, and in larger populations.

In summary, greater success of E. plantagineum in contrast to that of E. vulgare in colonising the Australian continent since introduction in the 1800 s corresponds with variation in a number of attributes between the two species: E. plantagineum has 1) a better match between its phenology and the Mediterranean type climate encountered across much of Australia, 2) greater drought tolerance, 3) greater genetic diversity and smaller genome size, and 4) greater abundance of defensive and potentially offensive secondary compounds. The invasion history of this genus in Australia thus provides support for several (non-mutually exclusive) hypotheses previously proposed to explain the ability of plant species to invade new territories.

Materials and Methods

Current and historical survey of E. vulgare and E. plantagineum distribution

The distribution of E. plantagineum and E. vulgare was initially reviewed by examination of herbarium records available from the AVH48. The identity of specimens falling outside of the expected distribution of either taxon was re-examined, and identifications corrected where necessary. A large field survey for presence of E. plantagineum and E. vulgare was conducted in the spring of 2011, 2012 and 2013 across southeastern Australia covering 76 locations aligned with three longitudinal transects32. Additional survey points were included in the Riverina region (Fig. S1) to survey additional geographically distinct populations of each species. As per the Interim Biogeographic Regionalisation for Australia (IBRA) survey, Australia is currently divided into 89 biogeographic regions (Fig. S2) according to climate, geology, landform, species and native vegetation ( For regional climatic analyses, average temperature and annual rainfall for each biogeographic region of collection were obtained from the Spatial Data Analysis Network at Charles Sturt University, Wagga Wagga, NSW.

Ecology survey - impact of infestation on Echium spp. growth and local plant biodiversity

An ecological field survey was conducted at 17 and four sites of E. plantagineum and E. vulgare, respectively (Table S3), in the summer of 2013 and 2014 to investigate the impact of the establishment of these two invaders on local plant biodiversity. For both species, data were collected from two 1 m × 1 m quadrats at each location. The number of Echium sp. individuals, number of other plant species, and total number of other plants present in quadrats were recorded. Means were compared with the Wilcoxon signed-rank test (Statistix ver. 9.067) because ecological data did not meet the requirements for ANOVA.

Comparison of genome size

Echium vulgare leaf tissues were collected from the four known geographically distinct locations of E. vulgare infestation in the SEH biogeographic region, while E. plantagineum seedlings from 11 locations were obtained after seed germination (Table 2). Fresh leaf tissue from numerous individuals (9–15 per population depending on successful germination and establishment) were collected and analysed within 48 hours, depending on availability. A total of 60 E. vulgare and 140 E. plantagineum samples from geographically distinct locations were analysed. Samples were prepared for flow cytometry analysis according to Loureiro, et al.68 using WPB nuclei isolation buffer. Raphanus sativus L. (red globe radish) was used as internal reference for assessment of genome size in E. vulgare and E. plantagineum20. Three samples from each location were examined individually by comparison to the radish genome using a Gallios Flow Cytometer (Beckman Coulter, USA), and any additional samples from each site were pooled for analysis. Three independent repetitions were performed for each sample on separate days, with at least 5,000 nuclei being analysed each time. Genome size of samples from each location was compared using an unbalanced ANOVA (location as factor) with GenStat 17th edition69.

Comparison of genetic diversity

A total of 25 E. vulgare and 129 E. plantagineum plan samples were used for genetic diversity analysis (Table S4), which included preserved herbarium voucher specimens provided by Brendan J. Lepschi of the Australian National Herbarium (four E. vulgare and 20 E. plantagineum). Samples of E. plantagineum and E. vulgare originated from widely distributed locations across the known endemic range of each species in southeastern Australia (Fig. S1). In addition, Wagga Wagga (NSW) experienced several large outbreaks of E. plantagineum which were also monitored and included in sampling. This region was therefore considered a hotspot of diversity following preliminary evaluation and a total of 42 samples were collected from Wagga Wagga NSW for additional haplotype analysis. Several samples from WA, TAS, and the Northern Territory were also sequenced.

Genomic DNA isolation, PCR, sequencing and alignment procedures were performed as described previously70. Samples were sequenced for one nuclear gene (ITS) and three chloroplast DNA regions (trnH-psbA spacer, trnL intron and trnL-trnF spacer). A 25 bp portion was discarded from the trnL intron sequence alignment due to low sequencing quality caused by a homopolymeric region (polyA and polyT) present in the sequence. Indels were coded as single mutations as described previously71 (Appendix 1). In addition, a 4 bp inversion region in trnH-psbA spacer72 was also coded as a single mutation73. The alignments of three linked chloroplast DNA regions were concatenated using FABOX74. Heterozygotes of nuclear ITS sequences were phased into two separate sequences via PHASE 2.175, using 1000 iterations, 10 thinning intervals and 1000 burn-in iterations. The algorithm was run five times using the “-x option” to obtain accurate estimates75. All sequences reported in this study have been deposited in the GenBank database under the GenBank accession numbers KX012007-KX012622.

Nucleotide (h) and haplotype (π) genetic diversity estimates50 were calculated within species using ARLEQUIN ver. 3.576, and 95% statistical parsimony network analyses was performed to investigate the nuclear and chloroplast DNA genealogical relationships in E. plantagineum using TCS ver. 1.2177. Network analysis was not performed for E. vulgare due to the limited number of haplotypes detected in our sampling.

Metabolic profiling, UPLC MS QToF and data analysis of leaf extracts in E. plantagineum and E. vulgare

Both species were evaluated under uniform glasshouse conditions and near identical field conditions in neighbouring collection sites to assess the production of pyrrolizidine alkaloids. Seeds of E. plantagineum were collected from Adelong (N: −35.296, E: 148.057) and Silverton (N: −31.883, E: 141.216), NSW, while seeds of E. vulgare were collected from Adaminaby (N: −35.995, E: 148.791) and Cooma (N: −36.140, E: 149.200), NSW. Plants were cultivated in a glasshouse as described previously32 using a randomised block design experiment (three blocks) and harvested sequentially by blocks when E. plantagineum plants were flowering and E. vulgare were at the rosette stage. Echium vulgare did not flower likely due to its perennial growth habit and lack of vernalisation32. Field plants of both species were collected from a roadside population near Bathurst, NSW (N: −33.463, E: 149.476) at flowering stage. To our knowledge, this is the first reported site where both species were co-located at the same site. Leaves were combined from three or four plants in the field and glasshouse experiment, respectively, to obtain a composite sample of 4.0 g of foliar tissue for extraction. Foliar tissue extraction, solid phase extraction, UPLC MS QToF analysis and statistical analysis were performed as previously described32,78.

Additional Information

How to cite this article: Zhu, X. et al. Ecology and genetics affect relative invasion success of two Echium species in Southern Australia. Sci. Rep. 7, 42792; doi: 10.1038/srep42792 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Change history

  • 16 March 2017

    A correction has been published and is appended to both the HTML and PDF versions of this paper. The error has been fixed in the paper.


  1. 1.

    Biological invasions: economic and environmental costs of alien plant, animal, and microbe species (CRC Press, 2002).

  2. 2.

    Natural Resource Management Ministerial Council. Australian Weeds Strategy – A national strategy for weed management in Australia (Commonwealth of Australia, 2007).

  3. 3.

    , , , & Adaptive evolution in invasive species. Trends Plant Sci. 13, 288–294, doi: 10.1016/j.tplants.2008.03.004 (2008).

  4. 4.

    The ecology of invasions by animals and plants. (Methuen, 1958).

  5. 5.

    & Evolution of increased competitive ability in invasive nonindigenous plants: a hypothesis. J. Ecol. 83, 887–889, doi: 10.2307/2261425 (1995).

  6. 6.

    Self-compatibility and establishment after long-distance dispersal. Evolution 9, 347–349, doi: 10.2307/2405656 (1955).

  7. 7.

    , , & Invasive Plant Ecology (CRC Press, 2013).

  8. 8.

    , , , & The invertebrate fauna on broom, Cytisus scoparius, in two native and two exotic habitats. Acta Oecol. 21, 213–222, doi: (2000).

  9. 9.

    & Invasive plants versus their new and old neighbors: A mechanism for exotic invasion. Science 290, 521–523, doi: 10.1126/science.290.5491.521 (2000).

  10. 10.

    et al. Self-compatibility and plant invasiveness: Comparing species in native and invasive ranges. Perspectives in Plant Ecology Evolution and Systematics 14, 3–12, doi: 10.1016/j.ppees.2011.08.003 (2012).

  11. 11.

    , & Island colonization and evolution of the insular woody habit in Echium L. (Boraginaceae). Proceedings of the National Academy of Sciences of the United States of America 93, 11740–11745, doi: 10.2307/40506 (1996).

  12. 12.

    , & Genetic variability is correlated with population size and reproduction in American wild-rice (Zizania palustris var. palustris, Poaceae) populations. Am. J. Bot. 92, 990–997, doi: 10.3732/ajb.92.6.990 (2005).

  13. 13.

    et al. The population biology of invasive species. Annu. Rev. Ecol. Syst. 32, 305–332, doi: 10.1146/annurev.ecolsys.32.081501.114037 (2001).

  14. 14.

    , , & The inbreeding paradox in invasive species. Interciencia 31, 544–546 (2006).

  15. 15.

    Evolutionary genetics of invasive species. Trends Ecol. Evol. 17, 386–391, doi: 10.1016/s0169-5347(02)02554-5 (2002).

  16. 16.

    & Founding events in species invasions: genetic variation, adaptive evolution, and the role of multiple introductions. Mol. Ecol. 17, 431–449, doi: 10.1111/j.1365-294X.2007.03538.x (2008).

  17. 17.

    , & The large genome constraint hypothesis: evolution, ecology and phenotype. Annals of Botany 95, 177–190, doi: 10.1093/aob/mci011 (2005).

  18. 18.

    , , , & Naturalized plants have smaller genomes than their non-invading relatives: a flow cytometric analysis of the Czech alien flora. Preslia 82, 81–96 (2010).

  19. 19.

    , & DNA Amounts in two samples of angiosperm weeds. Ann. Bot. 82, 121–134 (1998).

  20. 20.

    , & Genome size variation in Macaronesian angiosperms: forty percent of the Canarian endemic flora completed. Plant Syst. Evol. 252, 215–238, doi: 10.1007/s00606-004-0280-6 (2005).

  21. 21.

    , , & The hidden side of plant invasions: the role of genome size. New Phytol. 205, 994–1007, doi: 10.1111/nph.13107 (2015).

  22. 22.

    , , , & Identification and localization of isohexenylnaphthazarins in mature roots and seedling of Paterson’s curse (Echium plantagineum). In Natural Products Chemistry Group Annual One-day Symposium (eds , , & ) 19 (Rayal Australian Chemical Institute; 2014).

  23. 23.

    The biology of Australian weeds. 8. Echium plantagineum L. Journal of the Australian Institute of Agricultural Science 48, 3–16 (1982).

  24. 24.

    , , & The biology of Canadian weeds. 116. Echium vulgare L. Canadian Journal of Plant Science 82, 235–248, doi: 10.4141/P01-058 (2002).

  25. 25.

    & Noxious weeds of Australia 609–612 (CSIRO publishing, Melbourne, 2001).

  26. 26.

    & The Paterson’s Curse Management Handbook 40 (Department of Natural Resources and Environment, 1999).

  27. 27.

    Anon. (2009). Paterson’s curse. Weeds of Southern Tasmania. NRM South and the Southern Tasmanian Councils Authority. Retrieved 25th January 2016, from .

  28. 28.

    Noxious weed of Victoria 32–36 (Inkata Press Proprietary Limited, 1973).

  29. 29.

    The herbaceous species of Echium (Boraginaceae) naturalized in Australia. Muelleria 3, 215–244 (1977).

  30. 30.

    et al. Introduction of Paterson’s curse (Echium plantagineum) to Australia - unravelling the story by DNA sequence analysis. In 20th Australasian Weeds Conference (eds & ) 157–161 (Weeds Society of Western Australia; 2016).

  31. 31.

    et al. Metabolic profiling and identification of shikonins in root periderm of two invasive Echium spp. weeds in Australia. Molecules Under review (2017).

  32. 32.

    et al. Metabolomic profiling of pyrrolizidine alkaloids in foliar of two Echium spp. invaders in Australia – a case of novel weapons? International Journal of Molecular Sciences 16, 26721–26737, doi: 10.3390/ijms161125979 (2015).

  33. 33.

    , & Metabolic profiling in Echium plantagineum: presence of bioactive pyrrolizidine alkaloids and napthoquinones from accessions across southeastern Australia. Phytochem. Rev. 12, 831–837, doi: 10.1007/s11101-013-9306-4 (2013).

  34. 34.

    et al. Identification and localization of bioactive naphthoquinones in the roots and rhizosphere of Paterson’s curse (Echium plantagineum), a noxious invader. J. Exp. Bot. 67, 3777–3788, doi: 10.1093/jxb/erw182 (2016).

  35. 35.

    , , , & The chemistry and biology of alkannin, shikonin, and related naphthazarin natural products. Angewandte Chemie International Edition 38, 270–301, doi: 10.1002/(SICI)1521-3773(19990201)38:3<270::AID-ANIE270>3.0.CO, 2–0 (1999).

  36. 36.

    & Novel weapons: Invasive success and the evolution of increased competitive ability. Front. Ecol. Environ. 2, 436–443, doi: 10.2307/3868432 (2004).

  37. 37.

    , , , & Chemical cues involved in the attraction of the oligolectic bee Hoplitis adunca to its host plant Echium vulgare. Biochem. Syst. Ecol. 39, 498–508, doi: 10.1016/j.bse.2011.07.008 (2011).

  38. 38.

    & Phenotypic plasticity among Echium plantagineum populations in different habitats of Western Cape, South Africa. S. Afr. J. Bot. 74, 746–749 (2008).

  39. 39.

    Factors affecting seed germination of Echium plantagineum L. and Trifolium subterraneum L. Weed Res. 16, 337–344 (1976).

  40. 40.

    Factors affecting seedling establishment and survival of Echium plantagineum L., Trifolium subterraneum L. and Lolium rigidum Gaud. Weed Res. 16, 267–272 (1976).

  41. 41.

    & Reproduction of Echium vulgare L. (Boraginaceae) at contaminated sites. Acta Biologica Cracoviensia Series Botanica 45, 69–75 (2003).

  42. 42.

    , & Characteristics distinguishing invasive weeds within Echium (Bugloss). Weed Res. 26, 351–364, doi: 10.1111/j.1365-3180.1986.tb00718.x (1986).

  43. 43.

    , , & The origins and evolution of the genus Myosotis L. (Boraginaceae). Mol Phylogenet Evol 24, 180–193 (2002).

  44. 44.

    & Phylogenetic relationships of the monotypic genera Halacsya and Paramoltkia and the origins of serpentine adaptation in circum-mediterranean Lithospermeae (Boraginaceae): insights from ITS and matK DNA sequences. Taxon 58, 700–714 (2009).

  45. 45.

    , , & Origin of Mediterranean insular endemics in the Boraginales: integrative evidence from molecular dating and ancestral area reconstruction. J. Biogeogr. 36, 1282–1296, doi: 10.1111/j.1365-2699.2009.02082.x (2009).

  46. 46.

    & Population genetics of Echium plantagineum L.-target weed for biological control. Australian Journal of Biological Sciences 39, 369–378 (1986).

  47. 47.

    , & Allozyme electrophoresis: a handbook for animal systematics and population studies 420 (Academic Press, 1986).

  48. 48.

    AVH. Australia’s Virtual Herbarium (2015).

  49. 49.

    Cytogenetic studies on the Boraginaceae. Brittonia 7, 233–266, doi: 10.2307/2804694 (1951).

  50. 50.

    Molecular Evolutionary Genetics (Columbia University Press, 1987).

  51. 51.

    & Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad. Sci. USA 76, 5269–5273 (1979).

  52. 52.

    The naturalization of Echium plantagineum L. in Australia. Australian Weeds 1, 29–31 (1982).

  53. 53.

    & Rapid screening method to detect potential biparental inheritance of plastid DNA and results for over 200 angiosperm species. Am. J. Bot. 75, 1443–1458, doi: 10.2307/2444695 (1988).

  54. 54.

    The use of chloroplast DNA polymorphism in studies of gene flow in plants. Trends Ecol. Evol. 10, 198–202, doi: (1995).

  55. 55.

    et al. Development of SSR markers for genetic analysis of silverleaf nightshade (Solanum elaeagnifolium) and related species. Plant Molecular Biology Reporter 31, 248–254, doi: 10.1007/s11105-012-0473-z (2013).

  56. 56.

    et al. Evaluation of simple sequence repeat (SSR) markers from Solanum crop species for Solanum elaeagnifolium. Weed Res. 52, 217–223, doi: 10.1111/j.1365-3180.2012.00908.x (2012).

  57. 57.

    et al. Genetic variation and structure of Solanum elaeagnifolium in Australia analysed by amplified fragment length polymorphism markers. Weed Res. 53, 344–354, doi: 10.1111/wre.12029 (2013).

  58. 58.

    et al. Genetic variation in Solanum elaeagnifolium in Australia using SSR markers. Plant Protection Quarterly 28, 88–91 (2013).

  59. 59.

    Invasive alien species: a new synthesis (Island Press, 2005).

  60. 60.

    et al. Do ploidy level and nuclear genome size and latitude of origin modify the expression of Phragmites australis traits and interactions with herbivores? Biol. Invasions 18, 2531–2549, doi: 10.1007/s10530-016-1200-8 (2016).

  61. 61.

    , , , & Naturalized plants have smaller genomes than their non-invading relatives: a flow cytometric analysis of the Czech alien flora. Preslia 82, 81–96 (2010).

  62. 62.

    & Genome size variation and evolution in Veronica. Ann. Bot. 94, 897–911, doi: 10.1093/aob/mch219 (2004).

  63. 63.

    et al. Genome evolution in the genus Sorghum (Poaceae). Ann. Bot. 95, 219–227, doi: 10.1093/aob/mci015 (2005).

  64. 64.

    & Genomic clues to the evolutionary success of polyploid plants. Curr. Biol. 18, R435–444, doi: 10.1016/j.cub.2008.03.043 (2008).

  65. 65.

    , , & Uptake and metabolism of pyrrolizidine alkaloids in Longitarsus flea beetles (Coleoptera: Chrysomelidae) adapted and non-adapted to alkaloid-containing host plants. J. Comp. Physiol., B 173, 483–491, doi: 10.1007/s00360-003-0356-6 (2003).

  66. 66.

    , , , & The “Raison D'être” of pyrrolizidine alkaloids in Cynoglossum officinale: Deterrent effects against generalist herbivores. J. Chem. Ecol. 21, 507–523 (1995).

  67. 67.

    Statistix 9: Analystical software. Talahassee (2009).

  68. 68.

    , , & Two new nuclear isolation buffers for plant DNA flow cytometry: A test with 37 species. Ann. Bot. 100, 875–888, doi: 10.1093/annbot/mcm152 (2007).

  69. 69.

    , , , & An Introduction to GenStat for Windows (14th Edition) (VSN International, 2011).

  70. 70.

    et al. Selection of DNA barcoding regions for identification and genetic analysis of two Echium invaders in Australia: E. plantagineum and E. vulgare. In 19th Australasian Weeds Conference (ed. ) 396–400 (Tasmanian Weed Society; 2014).

  71. 71.

    , , & Evolution of annual species of the genus Medicago: a molecular phylogenetic approach. Mol. Phylogen. Evol. 9, 552–559, doi: 10.1006/mpev.1998.0493 (1998).

  72. 72.

    , & Intraspecific inversions pose a challenge for the trnH-psbA plant DNA barcode. Plos One 5, doi: 10.1371/journal.pone.0011533 (2010).

  73. 73.

    , , & From the Neotropics to the Namib: evidence for rapid ecological divergence following extreme long-distance dispersal. Bot. J. Linn. Soc. 179, 477–486, doi: 10.1111/boj.12334 (2015).

  74. 74.

    FaBox: an online toolbox for fasta sequences. Mol. Ecol. Notes 7, 965–968, doi: 10.1111/j.1471-8286.2007.01821.x (2007).

  75. 75.

    , & A new statistical method for haplotype reconstruction from population data. Am. J. Hum. Genet. 68, 978–989, doi: 10.1086/319501 (2001).

  76. 76.

    & Arlequin ver. 3.0: An integrated software package for population genetics data analysis Evolutionary Bioinformatics Online 1, 47–50 (2005).

  77. 77.

    , & TCS: a computer program to estimate gene genealogies. Mol. Ecol. 9, 1657–1659 (2000).

  78. 78.

    Department of the Environment. Interim Biogeographic Regionalisation for Australia (Subregions) v. 7 (IBRA) [ESRI shapefile]. Available from (2012).

Download references


The authors acknowledge the Australian Research Council (Discovery Project: DP130104346) for funding this project, the Australian National Herbarium for providing information on virtual samples and voucher specimens and the Spatial Data Analysis Network (SPAN) of Charles Sturt University in Wagga Wagga NSW for assistance in creation of sampling maps and provision of climate data. Chris Brodie (State Herbarium of South Australia), John R. Hosking (Herbarium, University of New England), Neville G. Walsh (National Herbarium of Victoria) and Peter G. Wilson (National Herbarium of NSW) are thanked for confirming the identity of herbarium specimens. Dr. Nigel A.R. Urwin is acknowledged for his initial assistance in performing flow cytometry studies. RMC thanks the Montana Institute on Ecosystems and National Science Foundation Experimental Program to Stimulate Competitive Research Track-1 EPS-1101342 (INSTEP 3) for support.

Author information


  1. Graham Centre for Agricultural Innovation (Charles Sturt University and NSW Department of Primary Industries), Charles Sturt University, Wagga Wagga, 2678, Australia

    • Xiaocheng Zhu
    • , Paul A. Weston
    • , Dominik Skoneczny
    • , David Gopurenko
    • , Lucie Meyer
    • , Geoff M. Gurr
    •  & Leslie A. Weston
  2. NSW Department of Primary Industries, Wagga Wagga Agricultural Institute, Wagga Wagga, 2650, Australia

    • David Gopurenko
  3. Australian National Herbarium, Centre for Australian National Biodiversity Research, Canberra, 2601, Australia

    • Brendan J. Lepschi
  4. Division of Biological Sciences, University of Montana, Missoula, 59812, USA

    • Ragan M. Callaway
  5. Institute of Applied Ecology, Fujian Agriculture & Forestry University, Fuzhou 350002, China

    • Geoff M. Gurr


  1. Search for Xiaocheng Zhu in:

  2. Search for Paul A. Weston in:

  3. Search for Dominik Skoneczny in:

  4. Search for David Gopurenko in:

  5. Search for Lucie Meyer in:

  6. Search for Brendan J. Lepschi in:

  7. Search for Ragan M. Callaway in:

  8. Search for Geoff M. Gurr in:

  9. Search for Leslie A. Weston in:


L.A.W., G.M.G. and R.A.C. obtained funding for this project from ARC and X.Z., L.A.W., R.M.C. and G.M.G. conceived and designed the experiments. B.J.L. and X.Z. obtained field and herbarium voucher specimens. X.Z., D.S., and L.M. performed the experiments. X.Z., D.S., D.G., and P.A.W. contributed to data analysis. X.Z., D.S., D.G., B.J.L., P.A.W., R.M.C., G.M.G. and L.A.W. prepared the manuscript.

Competing interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to Xiaocheng Zhu.

Supplementary information

About this article

Publication history





Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing