Phenotypic variation in the introduced range of an invasive species can be modified by genetic variation, environmental conditions and their interaction, as well as stochastic events like genetic drift. Recent studies found that epigenetic modifications may also contribute to phenotypic variation being independent of genetic changes. Despite gaining profound ecological insights from empirical studies, understanding the relative contributions of these molecular mechanisms behind phenotypic variation has received little attention for invasive plant species in particular.

This review therefore aimed at summarizing and synthesizing information on the genetic and epigenetic basis of phenotypic variation of alien invasive plants in the introduced range and their evolutionary consequences. Transgenerational inheritance of epigenetic modifications was highlighted focusing on its influence on microevolution of the invasive plant species. We presented a comprehensive account of epigenetic regulation of phenotypic variation and its role in plant invasion in the presence of reduced standing genetic variation, inbreeding depression and associated genomic events which have often been observed during introduction and range expansion of an invasive alien species. Finally, taking clues from the studies conducted so far, we proposed a unified framework of future experimental approaches to understand ecological and evolutionary aspects of phenotypic variation. This holistic approach, being aligned to the invasion process in particular (introduction-establishment-spread), was intended to understand the molecular mechanisms of phenotypic variation of an invasive species in its introduced range and to disentangle the effects of standing genetic variation and epigenetic regulation of phenotypic variation.

epigenetics, evolution, genetic diversity, phenotypic variation, plant invasion, transgenerational inheritance

With the increasing number of reports on negative impacts of invasive species on regional biota (Bellard et al. 2017; Early et al. 2016; Iacarella et al. 2015), biological invasion has become a severe problem globally and for obvious reasons, is in the spotlight of recent research trends. Numerous studies have been conducted to identify how a minor component of native communities has successfully established itself in a new and heterogeneous environment and becomes dominant in the invaded communities (Callaway and Maron 2006).

Multiple hypotheses have been put forward to explain successful invasion, e.g. resource fluctuation, enemy release hypothesis, evolution of increased competitive ability (EICA) [reviewed by (Catford et al. 2009; Inderjit et al. 2005)]. Among these, the influence of genetic diversity on invasion success has been long recognized (Baker and Stebbins 1965) and numerous studies have been conducted to explain this relationship [e.g. (Facon et al. 2006; Lavergne and Molofsky 2007; Roman and Darling 2007)]. Standing genetic variation can result in rapid selection of population showing greater fitness leading to successful establishment and range expansion of the introduced species (Barrett 2015; Sakai et al. 2001). On the other hand, introduction in a new location and range expansion from the point of introduction may cause population size reduction (demographic bottleneck) which can reduce genetic variation (Estoup et al. 2016; Uller and Leimu 2011), and subsequently may lead to inbreeding and considerable loss of fitness (Schrieber and Lachmuth 2017). However, even genetically depleted founder populations may establish and spread successfully if:

i) detrimental inbreeding effects are mitigated (Hufbauer et al. 2013; Rosche et al. 2017; Schrieber and Lachmuth 2017) by one or several of preventive mechanisms like multiple introductions [(Dlugosch and Parker 2008a; Uller and Leimu 2011); but see Hagenblad et al. 2015], genetic admixture among introductions (Roman and Darling 2007), preadaptation to the environment found in the introduced range (Hufbauer et al. 2012), and polyploidy (Pérez et al. 2006), or

ii) the plasticity of ecologically relevant traits of a genotype is enhanced in a way it can take advantage of a wider ecological niche (Bossdorf et al. 2008; Muth and Pigliucci 2007; Richards 2006; Spens and Douhovnikoff 2016; Walls 2010). Evolutionary changes in traits related to log-distance dispersal, growth rate, tolerance to environmental heterogeneity, and competitive ability in response to novel environmental conditions have been found to promote invasiveness (Bhattarai et al. 2017; Davidson et al. 2011; van Kleunen et al. 2010).

Epigenetic modifications in gene expression, being independent of any changes in DNA sequence (Nicotra et al. 2010; Richards 2006; 2011; Scoville et al. 2011), have been recognized as key mechanisms behind the expression of inbreeding depression (Biémont 2010; Nebert et al. 2010; Vergeer et al. 2012) and plastic responses of plant traits to environmental cues (Herrera and Bazaga 2013). Epigenetic changes can be induced by environmental stresses, both biotic and abiotic [e.g. (Dowen et al. 2012; Verhoeven et al. 2010)], and invasive plants are frequently exposed to these interactions in their introduced environment (Blackburn et al. 2011; Nunez-Mir et al. 2017; Zefferman et al. 2015). However, while there has been some progress in epigenetic studies in model and non-model organisms, studies with invasive plants have so far been limited to reviewing broad patterns of epigenetic variation (Richards et al. 2017). Moreover, it has been found that adaptive evolution to local conditions, phenotypic plasticity, or sometimes a combination of both, help invasive species to compete in a range of environments (Liao et al. 2016; Montesinos and Callaway 2018). Experimental studies on local adaptation and phenotypic plasticity are often conducted separately (but see Liao et al. 2016), thereby leaving a gap in comprehensive understanding of relative contribution of genetic differentiation and epigenetically regulated phenotypic variation on invasion success.

In this context, a comprehensive appraisal of the role of genetic and epigenetic variation in plant invasion and future prospects for investigation appears to be timely. This review was therefore framed to i) recognize the factors responsible for phenotypic variation; ii) identify the role of epigenetic processes in maintaining fitness of invasive plants; and iii) to propose a unified framework of experimental approaches to understand the relative importance of genetic differentiation and epigenetic regulation of trait fitness.

One of the major objectives of this review has been finding a comprehensive structural guideline of experimental approaches taking clues from the studies already conducted on invasive and non-invasive, model and non-model species. Phenotypic variation in a plant species in its introduced range is one of the most highly-researched topics in invasion biology in which basic ecological research demonstrated the role of phenotypic variation in the invasion success of exotic species. On the other hand, genetic variation, microevolution and epigenetic processes have been found to play significant roles in the phenotypic variation of traits, and therefore, have been recognized as relevant to understand the mechanisms underlying the natural variation in ecologically important traits (e.g. Colautti and Barrett 2013; Liao et al. 2016; Marchini et al. 2019; Oduor et al. 2016). In this context, a bridge between these parallel but complementary experimental approaches may provide a comprehensive understanding of ecological and evolutionary aspects of phenotypic variation of traits and their roles in the invasion process (introduction-establishment-spread continuum). This empirical framework is, therefore, specifically aimed to broaden the scope of research by including the genetic investigation components into the ecological studies on the phenotypic variation of traits in the invasive species. For this purpose, a model system has been conceptualized based on an invasive plant species which has been reported to have i) phenotypic variation across environmental gradients, and ii) reproduction ability through both sexual and vegetative means. We first aligned the proposed framework with three different stages of invasion (introduction, establishment and spread) to identify the possible locations where genetic differentiation and/or epigenetic regulation can act (Figure 1). We proposed future experimental studies (Figure 2) to understand the relative importance of genetic and epigenetic regulation of trait fitness along the course of the invasion process. The methodologies usually adopted for these experiments have been given in Table 2 with their respective strengths and challenges while detailed methodologies for screening epigenetic variation in invasive plants have been provided in Box 1.

Figure 1.

Conceptual framework for differentiating genetic and epigenetic basis for phenotypic variations across three stages of alien plant invasion process (introduction, establishment, spread). While genetic differentiation between introduced populations may cause phenotypic variation which leads to local adaptation and post-invasion rapid evolution through selection of traits and natural selection of optimal phenotype across environmental conditions, epigenetically regulated phenotypic variations are more prevalent in genetically similar populations. Three sites where epigenetic mechanisms may influence invasion success have been marked with triangles: 1) in case of genetic admixture between different genotypes present in a region, 2) biotic and abiotic stress induced epigenetic alterations among the genetically similar populations, and 3) transmission of epigenetic information from the parents (P) to the offspring (O) making the progeny capable of dealing with similar kinds of parental environment.

Figure 2.

Experimental framework for differentiating genetic and epigenetic regulation of phenotypic plasticity. While a field survey of natural populations may identify plastic traits (1), reciprocal transplant experiments comparing performances of local and foreign populations may give insights into local adaptation, phenotypic plasticity and genetic differentiation as well (2.1). Plants grown in common garden experiments may be subjected to analysis with genetic and methylation-sensitive markers (2.2) or they can be exposed to environmental stresses before analysis (2.3) to identify genetic and epigenetic variation regulating trait plasticity. The use of demethylating agents (2.4) can also provide indirect evidence of transgenerational epigenetic inheritance. Samples from the natural population can also be analyzed with these markers followed by proper statistical analysis to disentangle genetic and epigenetic effects on trait plasticity (3). Characterization of reaction norms of the plants (e.g. comparison between native and invasive lineages) grown in common garden in response to environmental gradients (4) may highlight the trade-offs between maintaining a high performance across a range of conditions (robustness or jack of all trades) and maximizing fitness in an environmental condition (opportunism or master of some) or both (robust to environmental conditions and high performance, the general-purpose genotype).

Box 1.Download as CSV 

Methodologies for screening epigenetic variation in invasive plants.

Molecular markers with methylation sensitive restriction enzymes : A standard Amplified Fragment Length Polymorphism (AFLP) process followed by methylation sensitive AFLP (MS-AFLP or MSAP). In MS-AFLP, pairs of methylation sensitive restriction enzymes (isoschizomers) have been used to survey cystine methylation at restriction sites spread across the investigated genomes. In AFLP, MseI and EcoRI have been used to digestion of DNA extracts whereas HpaII and MspI with EcoRI have been used in MS-AFLP. AFLP and MS-AFLP are usually applied in parallel to compare genetic and epigenetic structures of populations using statistical techniques. Unlike HPLC- and ELISA-based assays which determined the proportion of methylated cytosines across the entire genome, the MS-AFLP can distinguish between different genomic locations or contexts (CG, CHG, CHH) of cytosine methylation from the banding patterns: CpG methylated loci (bands present in EcoRI/MspI only); nonmethylated loci (bands present in both profiles); loci hemimethylated at the external C of the restriction site (bands present in EcoRI/HpaII only) and noninformative loci (bands absent in both profiles). This methodology has been successfully applied to screen epigenetic variation in both invasive species [e.g. Alternanthera philoxeroides (Gao et al. 2010); Fallopia sp. (Iacarella et al. 2015; Richards et al. 2012); Phragmites australis (Spens and Douhovnikoff 2016)] and non-invasive species [Taraxacum officinale (Verhoeven et al. 2010); Viola elatior (Catford et al. 2009; Roman and Darling 2007; Schulz et al. 2014)]. Although the commonly used pair of isoschizomers (HpaII/MspI) can identify changes in methylation pattern, they fail to support data concerning genetic variation exclusively. To circumvent this limitation, some authors suggested use of met-AFLP along with AFLP procedure. In met-AFLP, the restriction enzymes Acc65I/MseI and KpnI/MseI have been used. For example, Chwedorzewska and Bednarek (2012) used AFLP and met-AFLP to characterize genetic and epigenetic variation in invasive Poa annua population in Antarctica. In case of non-invasive species, Abratowska et al. (2012) used this procedure to identify genetic distinctiveness of metallicolous and non-metallicolous populations of a metallophyte, Armeria maritima.

Future directions : Among the advanced and more powerful technologies, bisulfite sequencing-based methods are now being used for screening epigenetic variation (e.g. Schield et al. 2016; Spens and Douhovnikoff 2016; van Gurp et al. 2016). In these methods, unmethylated cytosines are converted to uracil, and methylated cytosines are identified by comparing a treated sample to a reference sample. Quantification of small (s) RNAs that influence de novo establishment and maintenance of DNA methylation at many sites may also provide insights into the heritable epigenetic variation in plants (see Bond and Baulcombe 2014).

Table 1.Download as CSV 

Examples of experimental studies investigating the role of epigenetic variation in phenotypic plasticity in both non-native and native, model and non-model species in controlled as well as field-based experiments. The factors which may influence the experimental designing and outcomes are mentioned here: species reproduction (sexual, vegetative, or both), plant material used, environmental gradient responsible for epigenetically controlled plastic changes and genetic as well as methylation sensitive genetic marker-based analysis (AFLP = Amplified Fragment Length Polymorphism; MS-AFLP/MSAP/met-AFLP = methylation sensitive AFLP).

Obs.	Name of the species	Species status	Species reproduction	Experimental design	Plant material	Environmental gradient	Methodology	Reference
1	Fallopia sp. (Japanese knotweed)	Invasive	Vegetative and sexual	Controlled	Rhizome – Leaf	Diverse habitats	AFLP and MS-AFLP	Richards et al. (2012)
2	Poa annua	Non-native	Sexual	Field based	Shoot	Comparison between native & invasive populations	AFLP and met-AFLP	Chwedorzewska and Bednarek (2012)
3	Alternanthera philoxeroides	Non-native	Vegetative	Field based	Leaf	Habitat – Aquatic and terrestrial	AFLP and MSAP	Gao et al. (2010)
				Common garden	Plant
4	Spartina sp. (5 species – 2 parents, 2 hybrids and 1 allopolyploid	Non-native	Sexual	Controlled	Leaf	Allopolyploid speciation	AFLP and MSAP	Salmon et al. (2005)
5	Phragmites australis	Introduced invasive and native non-invasive subspecies	Facultative clonal	Field based	Leaf	Comparison between native & invasive subspecies	AFLP and MS-AFLP	Spens and Douhovnikoff (2016)
6	Ageratina adenophora (Crofton Weed)	Non-native	Sexual and vegetative	Controlled	Leaf	Cold tolerance	ICE1 gene methylation	Xie et al. (2015)
7	Taraxacum officinale	Endemic	Apomictic	Controlled	Seed – Leaf	Nutrient, Salt, Pathogen attack	AFLP and MS-AFLP analysis	Verhoeven et al. (2010)
8	Arabidopsis thaliana	Model species		Controlled	Seed – Leaf		Demethylating agent 5-azacytidine	Bossdorf et al. (2010)
9	Viola cazorlensis	Endemic	Sexual	Field based	Leaf	Adaptive epigenetic variation	AFLP and MSAP	Herrera and Bazaga (2010)
10	Viola elatior	Endemic	Vegetative and sexual	Field based	Leaf	Light availability	AFLP and MSAP	Schulz et al. (2014)
11	Betula ermanii	Endemic	Sexual	Field based	Leaf	Habitat	AFLP and MS-AFLP	Wu et al. (2013)
12	Armeria maritima	Endemic	Obligatory outbreeding	Controlled	Seed – Leaf	Metal concentration	AFLP and met-AFLP	Abratowska et al. (2012)
13	Ilex aquifolium	Endemic	Sexual	Field based	Leaf – heterophylly	Herbivory	MSAP	Herrera and Bazaga (2013)
14	Laguncularia racemose	Mangrove-endemic	Vegetative and sexual	Field based	Leaf	Habitat	MSAP	Lira-Medeiros et al. (2010)
15	Viola cazorlensis	Endemic	Sexual	Field based (long term: 20 years)	Leaf	Herbivory	AFLP and MSAP	Herrera and Bazaga (2011)

Based on a literature search on Web of Science to find experimental studies conducted to establish relationship between epigenetic variation and phenotypic plasticity. We used the search phrase “(“phenotypic plasticity” AND “plant”) AND (“epigenetic”) AND (“transgenerational plasticity”) AND (“epigenetics” OR “methylation”). The search result yielded 30 papers from which the empirical studies have been summarized in this table. We also considered references cited in these papers and the experimental studies have been included in this table. The references of these studies have been cited in the literature section.

Table 2.Download as CSV 

Experimental designs commonly used for investigating the effect of epigenetic variation on phenotypic plasticity and transgenerational pattern of epigenetic changes across generations. Strengths and challenges associated with each of these approaches have been mentioned.

Experimental design		Examples	Strengths	Challenges
Study system	Study procedure
Natural population	1. Sampling from plant materials (leaf, shoot) of identical developmental stages across a disturbance gradient	(Herrera and Bazaga 2013; Schulz et al. 2014)	1. Consider dynamic ecological factors that exist in wild populations (Spens and Douhovnikoff 2016)	1. Cannot identify whether the observed differences reflect heritable variation or repeated introduction (Richards et al. 2017)
	2. Analysis with molecular markers and methylation sensitive restriction enzymes (Box 1)		2. Three-way relationship (environment x phenotypic plasticity x epigenetic changes) can be established	2. Challenging for sexually reproducing organisms in which genetic and epigenetic variation may be closely intertwined (Herrera and Bazaga 2013)
	3. Statistical analysis to identify epigenetic variation that is not explained by genetic variation
Controlled experiments
Common garden – I	1. Sampling of reproductive materials (rhizomes, seeds) from the field population across a disturbance gradient	(Abratowska et al. 2012; Richards et al. 2012)	1. Minimization of epigenetic differences induced among sampling locations	1. Experimental design may be narrow and therefore, may oversimplify the dynamic ecological factors existing in the wild populations (Spens and Douhovnikoff 2016)
	2. Grow materials in a common environment		2. Detection of stable and heritable (through clonal propagation) epigenetic changes (Bossdorf et al. 2008)	2. Not suitable for outcrossing species as genetic identity of the field population is unknown
	3. Sampling from plants grown in the controlled environment		3. By controlling genetics and environment, quantification of epigenetic variation is possible
	4. Analysis with molecular markers and methylation sensitive restriction enzymes (Box 1)
	5. Statistical analysis to identify epigenetic variation that is not explained by genetic variation
Common garden – II	1. Collection of known genotypes (e.g. from seed stocks, seeds from asexual variants of apomictic plants)	(Bossdorf et al. 2010; Verhoeven et al. 2010)	1. Identification of stress induced DNA methylation patterns	Not suitable for sexually reproducing species in case the genetic variation is unknown and seed stock is not available
	2. Exposure to environmental treatments		2. Heritability of traits
	3. Seeds collected from treated plants and grown in control environment
	4. Samples from controlled environment plants
	5. Analysis with molecular markers and methylation sensitive restriction enzymes (Box 1)
Natural population + Common garden	1. Genetic and epigenetic profiling (Box 1) from field sampled plant materials	(Gao et al. 2010)	1. Identification of epigenetic changes at a temporal scale (a plant’s life time)	Challenging for sexually reproducing plant species
	2. Grow material in a common environment		2. Direction of epigenetic alteration (reversible)
	3. Reciprocal transplantation of the plants grown in common environment
	4. Sampling from the transplanted plants
	5. Analysis with molecular markers and methylation sensitive restriction enzymes (Box 1)
	6. Statistical analysis to identify epigenetic variation that is not explained by genetic variation

Field and controlled experiments are being conducted to characterize phenotypic variation of invading populations, often in comparison to their native congeners and to other species native to the invaded habitat (van Kleunen et al. 2010). Reciprocal transplant and/or common garden experiments are suitable for delineating the effects of local adaptation and phenotypic plasticity on successful invasion of an exotic species (Figure 2). In reciprocal transplant experiments, individuals from different populations are transplanted between the original habitats from where the populations were sampled and population x test habitat interactions are quantified in terms of fitness parameters (Kawecki and Ebert 2004). The alternative approach to this involves creating the properties of different habitats in greenhouse or experimental plots, where fitness functions of different populations have been quantified. This experimental set-up is known as common garden (explant) studies. In a common garden experiment, two or more populations of a species growing in their native and non-native environments are transplanted in a common environment so that the genetic basis of the observed differences among field populations can be identified (Molofsky et al. 2017; Parker et al. 2003). While in field experiments (marked ‘1’ in Figure 2), identifying plastic responses to a set of well-defined stress factors is important for comparative studies (Gratani 2014), manipulation of resource conditions in a biologically meaningful manner is required in reciprocal transplant and/or common garden studies (marked ‘2.1’ in Figure 2) to yield important and relevant information (Davidson et al. 2011). In a reciprocal transplant experiment, plants from different invasive populations can be grown in a common environment to compare fitness traits. Higher mean fitness for all the traits of the local population compared to foreign population will indicate local adaptation (Local versus Foreign comparison). Significant difference in trait values among transplant sites (for a population) will indicate plastic responses whereas difference among populations (for a site) will indicate genetic differentiation. However, this approach is often confounded by intrinsic issues of population quality, e.g., inbreeding and transgenerational effects (Blanquart et al. 2013). The average effects of transplantation can be measured by comparing trait values between the local site and all away sites (Sympatric versus Allopatric comparison); however, this approach has been also found to be confounded by strong local advantage.

For example, morphological differentiation was studied between weedy, non-native and non-weedy, native populations of Centaurea solstitialis in a common garden setting and further compared using neutral genetic variation at simple sequence repeat markers (Eriksen et al. 2012). This study found quantitative variation to be more strongly partitioned among regions than genetic variation, which suggests that local adaptation might play a role in successful invasion of the species. In a recent transplantation experiment, local adaptation and phenotypic plasticity were examined in terms of fitness responses for sexual and clonal reproductive measures and vegetative responses of an invasive plant Fallopia japonica across a broad latitudinal range within North America (van Wallendael et al. 2018). This study reported significant effects of the source population (suggesting genetic differentiation) and transplant sites (suggesting phenotypic plasticity) for all vegetative traits, but no evidence of local adaptation was found for sexual or clonal reproductive traits. Contrasting examples are also found in which phenotypic plasticity can be operational being independent of genetic variation. For example, trait plasticity and genetic variation were examined across 16 populations of Crofton weed (Eupatorium adenophorum) in China in a common garden experiment followed by intersimple sequence repeat (ISSR) marker analysis. This study revealed the presence of high phenotypic plasticity of functional traits despite having low genetically based variation (Zhao et al. 2012).

However, phenotypic differentiation in invading populations may also arise from random shifts in allele frequencies during repeated demographic disequilibrium (i.e., genetic drift). Thus it is necessary to account for non-adaptive evolutionary change when investigating adaptive differentiation in invaders (Keller and Taylor 2008). In this context, information from neutral genetic markers can be used to control for, and quantify the effect of, non-adaptive processes with different statistical approaches (Agrawal et al. 2015; Keller et al. 2009; Meimberg et al. 2010; Schrieber et al. 2017). For example, genetic differentiation in phenotypic traits across environmental gradients was tested between native and introduced populations of two perennial plants Silene vulgaris and S. latifolia in a common garden experiment (Keller et al. 2009) using AFLP loci and statistically controlling neutral processes like colonization history, gene flow and genetic drift. The results revealed evidence of adaptive differentiation for some traits while the role of neutral processes governing phenotypic variation was also found for other traits.

To identify genetic and epigenetic regulation of phenotypic variation, the invasive populations of the common greenhouse environment can be subjected to analysis with genetic and methylation-sensitive markers (marked 2.2 in Figure 2) [e.g. Richards et al. (2012)]. The plants can be exposed to environmental stresses in common garden experimental set-up and stress-induced phenotypic variation can be analyzed using both genetic and methylation-sensitive markers (marked 2.3 in Figure 2) (2.3) (e.g. Verhoeven et al. 2010). Samples from natural populations can also be analyzed with these markers followed by proper statistical analysis to disentangle genetic and epigenetic effects on trait variation (3) (e.g. Herrera and Bazaga 2013). In addition, use of demethylating agents like 5-azacytidine and zebularine inhibits the enzyme methyltransferase activity of DNA demethylation and therefore natural epigenetic variation can be identified from responses of different natural populations to the treatment of these demethylating agents (marked 2.4 in Figure 2). Similar approaches have been found successful to identify genetic and epigenetic regulation of phenotypic variation in model and non-model species (Table 2). For example, a set of natural genotypes of the model species Arabidopsis thaliana was treated with demethylating agent 5-azacytidine, and the effect of reduced DNA methylation was identified as the main cause of the observed phenotypic changes of plant traits (Bossdorf et al. 2010). In another greenhouse experiment, individuals of six genotypes of a perennial grass species Festuca rubra were treated with 5-azacytidine and their performances were measured across different environmental conditions (Münzbergová et al. 2019). This study found interactive effect of demethylation with the environment and genotype, thereby suggesting that epigenetic variation can be influenced by both genetic structure and local environment.

Experiments involving multiple generations of the species may detect the heritability of plastic traits across generations (stage 2.2 in Figure 1). The progeny population can be grown from the seeds or the clonal fragments of the parental population in the common garden experiment, and trait variability can thereafter be analyzed using genetic and methylation-sensitive markers. Alternatively, demethylating agents can also be used to have indirect evidence of transgenerational epigenetic inheritance. While the majority of these studies have been conducted across sexual generations [(e.g. (Herman and Sultan 2016)), very few studies identify adaptive transgenerational effects in clonally reproducing plant species, although clonal reproduction is recognized as the main reproductive strategy for most plant species (Rendina González et al. 2018). For example, adaptive transgenerational effects in clonal offspring of Trifolium repens were tested after exposing parental generation to drought and herbivory stress (Rendina González et al. 2017). 5-azacytidine was used to decrease the global methylation level of DNA relative to control plants. The study found an increased number and size of offspring ramets (branches arising from the transplanted stolon) from the parents grown in drought condition and increased growth of offspring ramets from the parents treated with repeated application of jasmonic acid. Application of 5-azacytidine to the parents exposed to the drought condition and application of jasmonic acid reduced the growth of maternal ramets (transplanted main stolon). These findings provide evidence that parental environment can induce transgenerational effects in the offspring and some of these effects can be adaptive.

Focusing on a specific gene methylation variation can also provide two important insights: in case of genetically uniform species, variation in gene or protein expression (measured using microarrays or 2-D electrophoresis) indicate underlying epigenetic variation (Bossdorf et al. 2008) and secondly, the expression of the methylated gene may highlight the mechanism by which methylation differentiation contributes to the successful invasion (Xie et al. 2015). For example, the C-repeat/dehydration-responsive element binding factor (CBF) pathway governs plant responses to adverse low temperature (Chinnusamy et al. 2003). Demethylated upregulation of cold response regulator ICE1 (inducer of CBF expression 1) was found to be the evolutionary mechanism responsible for northward expansion of the invasive Ageratina adenophora (Crofton weed) in China (Xie et al. 2015). Use of Quantitative Trait Loci (QTL)-mapping approaches can be useful to link the natural epigenetic variation with the observed phenotype. QTLs have been recognized as genetic regions (associated with phenotypic traits) which control the magnitude of a specific trait (Cortijo et al. 2014). Epigenetic QTLs, or the loci associated with different methylation marks, have been found to control flowering time and root length in the model plant Arabidopsis, thereby demonstrating that heritability of some traits can be determined by epigenetic variation (Cortijo et al. 2014).

Finally, a higher degree of phenotypic plasticity in an invasive species does not necessarily mean that the species has become invasive due to the plasticity (Palacio‐López et al. 2015). To infer the role of phenotypic plasticity in successful invasion, observations of trait plasticity should be followed by experimental studies to identify that the plastic response is adaptive or linked to fitness (Davidson et al. 2011; van Kleunen and Fischer 2005). Characterization of reaction norms of the study species (e.g. comparison between native and invasive lineages) grown in common garden in response to environmental gradients (marked ‘4’ in Figure 2) may highlight the trade-offs between maintaining a high performance across a range of conditions (robustness or jack of all trades) and maximizing fitness in an environmental condition (opportunism or master of some) or both (robust to environmental conditions and high performance, the general purpose genotype) (e.g. Drown et al. 2011).

This review, being especially focused on plant invasion, has provided a comprehensive account of the molecular mechanisms of trait fitness of invasive plants. The strength of this review lies in the proposed framework that will encapsulate the ecological and evolutionary aspects of phenotypic variation. Future ecological studies should consider looking into the relative contributions of genetic variation and epigenetic modification to the observed phenotypic variation in invasive plant species, and characterizing the three-way relationship between environmental cue, phenotypic plasticity and epigenetic changes. This framework also suggests that these studies should combine trait and molecular data and include comparative analysis of fitness functions between native and introduced ranges of a species (van Kleunen et al. 2018) and explore adaptive differentiation in invaders, while accounting for non-adaptive evolutionary changes. The unified research framework, therefore, may converge the parallel lines of research towards a better understanding of the mechanism of successful invasion.

This study is supported by grants from the National Natural Science Foundation of China (Grant Nos. 41776166 and 31700178), Special Fund for Science and Technology Development of Guangdong Province (2017A030303014), the Science Foundation of the State Key Laboratory of Biocontrol (32-2017-A30), the Science and Technology Program of Guangzhou (Grant No. 201707020035) and the Chang Hungta Science Foundation of Sun Yat-sen University. The authors would like to thank Prof. Mark van Kleunen of the University of Konstanz, Germany, for providing valuable insights towards improving this manuscript’s quality and readability. The authors are grateful to the editor and three reviewers, Dr. Vit Latzel, Dr. Karin Schrieber and Dr. Stefan Michalski, for their constructive criticisms and valuable suggestions which have immensely improved the manuscript’s quality.