J Chem Ecol (2009) 35:495–504 DOI 10.1007/s10886-009-9629-1
Density-Dependent Phytotoxicity of Impatiens pallida Plants Exposed to Extracts of Alliaria petiolata E. Kathryn Barto & Don Cipollini
Received: 23 February 2009 / Revised: 29 March 2009 / Accepted: 2 April 2009 / Published online: 21 April 2009 # Springer Science + Business Media, LLC 2009
Abstract Invasive plants are by definition excellent competitors, either indirectly through competition for resources or directly through allelopathic inhibition of neighboring plants. Although both forms of competition are commonly studied, attempts to explore the interactions between direct and indirect competition are rare. We monitored the effects of several doses of extracts of Alliaria petiolata, a Eurasian invader in North America, on the growth of Impatiens pallida, a North American native, at several planting densities. The density-dependent phytotoxicity model predicts that as plant density increases, individual plant size will decrease, unless a toxin is present in the soil. In this case, individual plant size is predicted to increase as plant density increases, as plants share a limited toxin dose. We tested this model using fractions of an A. petiolata extract enriched in flavonoids or glucosinolates, as well as a combined fraction. The flavonoid-enriched fraction and the combined fraction suppressed I. pallida growth but only when applied at a dose eight times higher than that expected in the field. When treated with a dose equivalent to estimated field exposure levels, I. pallida growth was not distinguishable from that of control plants that received no
E. K. Barto : D. Cipollini Department of Biological Sciences, Wright State University, 3640 Colonel Glenn Hwy, Dayton, OH 45435, USA Present address: E. K. Barto (*) Institut für Biologie, Plant Ecology, Freie Universität Berlin, Altensteinstr. 6, 14195 Berlin, Germany e-mail:
[email protected]
extract, showing that indirect competition for resources was more important for determining the growth of I. pallida than direct allelopathic inhibition by A. petiolata. This is an important reminder that, even though many plants have the demonstrated potential to exert strong allelopathic effects, those effects may not always be apparent when other forms of competition are considered as well. Keywords Allelopathy . Invasion . Glucosinolates . Flavonoid glycosides
Introduction In natural communities, plants compete in complex ways as they vie for the same territory and resources. These interactions can be classified as direct or indirect competition and are difficult to distinguish experimentally since they operate in concert. Indirect competition is widely assumed to be the default natural condition, where two plants require access to a limited resource and the “winner” is more efficient at acquiring that resource (Connell 1990). Direct competition occurs through interference, with allelopathy being one of the most common mechanisms. Allelopathic plants release compounds into the environment that negatively impact surrounding plants, thus giving the allelopath a competitive advantage (Rice 1974). Invasive plant species are by definition excellent competitors, whether by direct competition, indirect competition, or both. The distinction is important for several reasons. Simply removing the invader may control the spread of an invasion by indirect competition. However, in a site where direct competition is occurring, removal of the invader may not be enough. Allelochemicals released by the invader could remain in the soil and continue to inhibit
496
to the density-dependent phytotoxicity model, if resource competition is the dominating factor in an interaction, then individual plant mass will decrease as density increases. However, if allelopathy is the dominant factor, then individual plant mass will decrease more slowly, or even increase, as density increases, until a density is reached where resource competition among target plants becomes the dominant factor (Fig. 1). This model assumes that all plants present must share the available dose of an allelochemical, and as plant density increases the dose per plant decreases. At very high densities, each plant receives a dose so low that allelopathic inhibition disappears, and resource competition among target plants thus becomes dominant (Weidenhamer et al. 1989). This pattern has been observed for plants exposed to purified compounds (Hoffman and Lavy 1978; Andersen 1981; Weidenhamer et al. 1989), ground tissue from suspected allelopaths (Tseng et al. 2003), and soil conditioned by an allelopathic plant (Weidenhamer et al. 1989). Allelopathic effects of A. petiolata may be masked by resource competition among target plants in some cases, and a better understanding of the dose–response relationship of secondary metabolites of A. petiolata would help clarify this issue. Synergism among allelochemicals also may be contributing to the inconsistency of results reported in the literature since fractions of extracts and even purified compounds have been used (Vaughn and Berhow 1999; Roberts and Anderson 2001). The objective of this experiment was to use a density-dependent phytotoxicity approach to determine the allelopathic potential of A. petiolata for a range of extract doses. We used fractionated extracts, enriched in either glucosinolates or flavonoid
Log of yield per plant
growth of other plants, even after removal of the allelopathic invader. For this reason, considerable attention has been paid to exploring the competitive mechanisms of invasive plants (Levine et al. 2004; Theoharides and Dukes 2007), including Alliaria petiolata (M. Bieb.) Cavara and Grande, Brassicaceae (garlic mustard; Meekins and McCarthy 1999; Bauer et al. 2005). A. petiolata, an herbaceous biennial, was introduced into North America from Europe in the 1860s and has since become invasive in much of Canada and the northeast and Midwestern USA (Nuzzo 2002). The success of A. petiolata in North America has been attributed to its high propagule pressure (Cavers et al. 1979), escape from herbivores (Blossey et al. 2001), superior competitive ability (Meekins and McCarthy 1999), and allelopathic inhibition of surrounding plants (Vaughn and Berhow 1999; Roberts and Anderson 2001; Prati and Bossdorf 2004; Stinson et al. 2006; Wolfe et al. 2008). Given the broad range of bioactive secondary metabolites produced by garlic mustard (Haribal and Renwick 1998, 2001; Vaughn and Berhow 1999; Haribal et al. 2001; Renwick et al. 2001; Cipollini and Gruner 2007), it is not surprising that a large part of the effort spent studying the invasive success of this species has focused on allelopathic effects. Crude extracts and purified compounds from A. petiolata inhibited germination and growth of other plants and their symbiotic arbuscular mycorrhizal fungi (Vaughn and Berhow 1999; Roberts and Anderson 2001; Stinson et al. 2006; Callaway et al. 2008). Seeds of Geum spp. sown in soil conditioned by A. petiolata germinated better if the soil also contained activated carbon, presumably indicating the presence of inhibitory organic compounds in the soil (Prati and Bossdorf 2004). Cipollini et al. (2008b) also found positive effects of adding activated carbon on growth of Impatiens capensis growing with A. petiolata in the field. In contrast, A. petiolata extracts have failed to inhibit germination or growth of several plant species in some studies (McCarthy and Hanson 1998; Cipollini et al. 2008a). While A. petiolata clearly has the potential to act allelopathically, whether allelopathic effects are important in more realistic environments where resource competition is also occurring is largely unknown. Activated carbon has been advocated for use as a soil amendment as a simple way to distinguish between allelopathy and resource competition because it should adsorb organic compounds with little effect on inorganic nutrients (Inderjit and Callaway 2003). However, it can alter nutrient levels in unpredictable ways (Lau et al. 2008), and other methods to verify the allelopathic potential of plant species are therefore necessary. Allelopathic effects are highly dependent on densities of target species and can be masked by resource competition at high densities (Weidenhamer et al. 1989; Weidenhamer 1996). According
J Chem Ecol (2009) 35:494–504
Log of density
Fig. 1 Predictions of the density-dependent phytotoxicity model. The bold line represents expectations when resource competition is dominant, and the dashed line represents expectations when allelopathy is dominant. Dotted lines represent expected patterns at intermediate chemical doses
J Chem Ecol (2009) 35:494–504
497
glycosides, to explore the contribution of different classes of compounds to observed allelopathic effects.
Methods and Materials We chose I. pallida (Nutt.) Balsaminaceae (pale jewelweed) as the target plant for this study because it grows in the same habitats invaded by A. petiolata and has been susceptible to allelopathic effects of A. petiolata in prior studies (Barto, unpublished data). I. pallida seeds were collected from a population in Yellow Springs, OH, USA (39° 47.0’ N, 83° 52.5’ W), and nonsterilized seeds were stratified immediately in sterile water at 3°C to stimulate germination. Then, seedlings were transferred to 15-cm (1 L) pots containing sieved field soil mixed 1:1 with sand.
a
Log mean shoot mass per plant
-0.2 0
0
Log plant density 0.2 0.4 0.6
-0.2
0.8
0X 0.25X 0.5X 1X 2X 4X 8X
-0.4 -0.6
1 Slope -0.7089 -0.7001 -0.7967 -0.7943 -0.8503 -0.6441 -0.5416
r2 0.8519 0.9201 0.9355 0.8119 0.9210 0.9070 0.9180
-0.8 -1 -1.2 -1.4
b -0.2 0.3 0.2 Log mean root mass per plant
Fig. 2 Response of Impatiens pallida plants exposed to a flavonoid-enriched fraction of an Alliaria petiolata extract (N=2–3). a Relationship of log mean shoot mass per plant and log plant density. b Relationship of log mean root mass per plant and log plant density
The experiment was fully factorial with four densities (one, two, four, and eight I. pallida plants), three A. petiolata fractions (glucosinolate-enriched fraction, flavonoid-enriched fraction, and a combined fraction), and six concentrations (0.25X, 0.5X, 1X, 2X, 4X, and 8X) with X equaling 3.3-mg tissue equivalents per gram soil. This dose was chosen to represent expected exposure levels in the field (Callaway et al. 2008). A set of controls that received only water was planted at all four densities, and at least three replicate pots were planted for each treatment combination. Extracts were prepared and fractionated, as in Callaway et al. (2008). We boiled A. petiolata tissues in ethanol, then filtered and dried the extract before defatting with hexane to remove chlorophylls. We wanted to focus on the glucosinolates and flavonoids, so the hexane fraction was discarded, even though it may have contained lipophilic
0.1 0 -0.1 -0.2 -0.3 -0.4 -0.5 -0.6 -0.7
0
Log plant density 0.2 0.4 0.6
0.8 0X 0.25X 0.5X 1X 2X 4X 8X
1 Slope -0.6228 -0.4589 -0.7336 -0.6133 -0.6392 -0.6977 -0.5279
r2 0.7281 0.8767 0.7881 0.7736 0.8162 0.8914 0.7984
498
J Chem Ecol (2009) 35:494–504
allelochemicals. We dissolved the dried extract in water, then partitioned with n-butanol to separate the flavonoid glycosides and alliarinoside from the glucosinolates (Callaway et al. 2008). Both extracts were dried and redissolved in water before use. Pots were dosed weekly and watered as needed. Plants were harvested after 6 weeks and air-dried at 25°C for 1 week, and shoot and root dry masses were measured. Root-to-shoot ratios were calculated, and the log of root and shoot dry mass per plant and log of root-to-shoot ratios were regressed against the log of plant density per pot using PROC GLM with density and dose as factors, followed by Duncan’s multiple-range test with α=0.05 (Weidenhamer et al. 1989). SAS version 9.1 (SAS Institute Inc., Cary, NC, USA) was used for all analyses.
Results Across all other factors, mean shoot dry masses were higher in pots dosed with a glucosinolate-enriched fraction than those that received either a combined extract or a flavonoidenriched fraction (F2, 161 =3.55, P=0.031). Mean root dry masses showed a similar pattern, and the effect was almost significant at the P=0.05 level (F2, 161 =2.88, P=0.059). Mean root and shoot mass per plant declined with increasing plant density at every dose. Among pots dosed with a flavonoid-enriched fraction, plant density was the most significant factor that impacted shoot dry mass (t=−19.07, P
