Allelopathy and Exotic Plant Invasion: From Molecules and Genes to Species Interactions

See allHide authors and affiliations

Science  05 Sep 2003:
Vol. 301, Issue 5638, pp. 1377-1380
DOI: 10.1126/science.1083245

This article has a correction. Please see:


Here we present evidence that Centaurea maculosa (spotted knapweed), an invasive species in the western United States, displaces native plant species by exuding the phytotoxin (–)-catechin from its roots. Our results show inhibition of native species' growth and germination in field soils at natural concentrations of (–)-catechin. In susceptible species such as Arabidopsis thaliana, the allelochemical triggers a wave of reactive oxygen species (ROS) initiated at the root meristem, which leads to a Ca2+ signaling cascade triggering genome-wide changes in gene expression and, ultimately, death of the root system. Our results support a “novel weapons hypothesis” for invasive success.

Invasive plant species threaten the integrity of natural systems throughout the world by displacing native plant communities (1) and establishing monocultures in new habitats (2). The leading theory for the exceptional success of invasive plants is their escape from the natural enemies that hold them in check, freeing them to utilize their full potential for resource competition (3). Allelopathy, the release of phytotoxins by plants, has been proposed as an alternative theory for the success of some invasive plants (4). Centaurea species are among the most economically destructive exotic invaders in North America, and they have long been suspected of using allelopathic mechanisms to rapidly displace native species (4, 57). Here we demonstrate the allelopathic effects of C. maculosa by integrating ecological, physiological, biochemical, cellular, and genomic approaches.

Previously we have reported the identification of the phytotoxic root exudates from C. maculosa (7, 8). The biologically active fraction was found to be racemic catechin: (–)-catechin was phytotoxic, whereas (+)-catechin had antimicrobial properties. Racemic catechin is abundant in soil extracts from C. maculosa–invaded fields, supporting the position that C. maculosa's invasiveness is facilitated by (–)-catechin release (7). It was observed that (±)-catechin concentration in the soil varied with its proximity to the taproot, differences in soil sampling zones, and age of knapweed's invasion (7). We further tested the “novel weapons hypothesis” by measuring the natural concentration of (–)-catechin in rhizospheres in Europe and North America (9). (–)-Catechin levels were more than twice as high in soils supporting invasive C. maculosa in North America than in Europe (Fig. 1A). Therefore, we analyzed the effect of (–)-catechin on European and North American grasses that interact with C. maculosa. The levels of (–)-catechin observed in North American and European soils were added to pots in which three different North American and European grasses were grown separately. When these levels of (–)-catechin were added to natural field soil in pots, the germination and, to a lesser extent, the growth of Festuca idahoensis and Koeleria micrantha, two native North American grasses, were sharply reduced (9) (Fig. 1, B and C). In contrast, both germination and growth data revealed that European grasses were more resistant to (–)-catechin than their North American counterparts (fig. S1). Similar experiments were conducted using A. thaliana, which was also negatively affected by (–)-catechin (fig. S2). These results provide strong evidence that the root exudation of (–)-catechin accounts at least in part for the displacement of native plant communities by C. maculosa.

Fig. 1.

(A) Natural concentrations of (–)-catechin in soils supporting invasive C. maculosa in North America compared with soils supporting native C. maculosa in Europe. Effect of (–)-catechin [200 μg g1 soil dry weight (DW) basis] on the germination (B) and total biomass (C) of two native North American grasses. Bars, 1 SE. Two-way analysis of variance (ANOVA) for biomass: F(treatment) = 16.92, df = 1,39, P < 0.001; F(species x treatment) = 9.58, df = 1,39, P = 0.004. Two-way ANOVA for germination: F(treatment) = 35.47, df = 1,39, P < 0.001; F(species x treatment) =2.22, df = 1,39, P = 0.145.

For detailed biochemical and molecular analysis of the mode of action of (–)-catechin, we selected A. thaliana and C. diffusa as target species. A. thaliana was chosen to facilitate analysis of genomic responses, whereas C. diffusa was selected because it is closely related to C. maculosa yet is susceptible to (–)-catechin (7). The addition of 100 μg ml1 (–)-catechin to the roots of C. diffusa and A. thaliana led to a condensation of the cytoplasm characteristic of cell death (Fig. 2, A to D; fig. S3; Movies S1 and S2). Such alterations in cytoplasmic structure occurred first in the meristematic zone of the root tip and then in the central elongation zone (CEZ) in both species (Fig. 2, A to D). This cytoplasmic condensation was followed by a cascade of cell death proceeding backward up through the stele (Fig. 2, A to D; fig. S3; Movies S1 and S2). To more fully characterize whether this change in cell morphology represented a wave of cell death upon (–)-catechin administration, we followed the pattern of cell death in C. diffusa and A. thaliana roots using the vital stain fluorescein diacetate (FDA) (9). This dye is retained by living cells but leaks from dead cells, rendering them nonfluorescent. Upon (–)-catechin addition, cells from the meristematic and CEZ were affected first and lost viability, shown by loss of FDA fluorescence, 600 s after exposure to the chemical (Fig. 2, E to G; fig. S4). Cells in the mature region of the root were not detectably affected during this time course (Fig. 2F) but died ∼55 min later (Fig. 2, F and G). Cell death progressed as a wave along the root and was characterized by sequential loss of viability of individual cells (Fig. 2, F and G). In addition, root hairs showed a termination of cytoplasmic streaming over a 5-min time course, followed by bursting at the apex, but only in hair cells that were undergoing tip growth (Fig. 2, C, D, and H; fig. S5; Movie S3). Root border cells do not appear responsive to (–)-catechin treatment (Fig. 2A; fig. S3). Consistent with its resistance to its own toxin, (–)-catechin had no detectable effect on the viability of the main root axis or growth of root hairs of C. maculosa (Fig. 2B; Movies S3 and S4) (10).

Fig. 2.

(–)-Catechin (100 μg ml1) elicits a wave of cell death originating in meristematic and CEZ cells of roots of C. diffusa and A. thaliana but not C. maculosa as evidenced by progressive tissue darkening (A to D) and loss of FDA viability staining (E to G). bc, border cells; m, meristem. Cell death proceeds as sequential loss of viability of individual cells (*) moving back at approximately 25 μm min1. (H) Effect of (–)-catechin (100 μg ml1, added at 20 min) on root hairs of A. thaliana. White arrow, growing root hair; black arrow, root hair bursting. The two focal planes at 40 min depict root hair bursting (black arrow). Bars in (A), (B), (E), and (G), 50 μm; in (C) and (D), 150 μm; in (F), 100 μm; and in (H), 20 μm. Images are representative of ≥15 separate roots.

Plants use sophisticated signal transduction cascades to sense and respond to biotic and abiotic stresses (11, 12). Therefore, we monitored the dynamics of various signaling elements, such as reactive oxygen species (ROS), calcium, and pH, that might be associated with the initial sensing and response to (–)-catechin. ROS production was monitored by imaging the ROS-sensitive fluorescent dye, dichlorofluorescein (DCF), which showed that (–)-catechin induced a rapid (within 10 s) increase in ROS in C. diffusa and A. thaliana (Fig. 3, A and B; fig. S6) but not in C. maculosa (Fig. 3C). The ROS change originated in the meristematic region of the root tip and later moved to the CEZ, suggesting that cells in these regions were the first to sense (–)-catechin. ROS production propagated as a wave through the root apex and then back along the main axis of the root (Fig. 3, A and B; fig. S6; Movie S5). The spatial kinetics of ROS induction are similar to the patterns of cell death induced by (–)-catechin (Fig. 2) but occurred 5 to 10 min before detectable loss of cell viability. In contrast to (–)-catechin, (+)-catechin did not induce ROS change in any of the species tested (10). The intracellular concentration of ROS in the root tips and meristematic zone, as measured by assessing oxidation of ferrous ion (Fe2+) to ferric ion (Fe3+) in vitro (9), increased 12-fold within 10 s upon (–)-catechin administration to A. thaliana and C. diffusa but was unaltered in C. maculosa (fig. S7) (10). Thus, both DCF-based imaging in vivo and the in vitro quantitative analysis indicated that (–)-catechin elicits a burst of ROS in susceptible plant roots. We therefore attempted to block intracellular ROS production using ascorbic acid (13) to characterize its role in the phytotoxicity of (–)-catechin. Ascorbic acid (50 mM) in combination with (–)-catechin (100 μg ml1) blocked the ROS increase seen in response to (–)-catechin (9) (Fig. 3, D to G; fig. S6; fig. S7) and the rhizotoxic response (Fig. 3E; fig. S6) in both C. diffusa and A. thaliana. Therefore, our results indicate (–)-catechin elicits production of ROS that appears required for phytotoxicity and that precedes cell death by 5 to 10 min.

Fig. 3.

(–)-Catechin (100 μg ml1) elicits intracellular ROS generation in C. diffusa (A) and A. thaliana (B), originating at the root tips and progressing back to the elongation zone (arrows) as visualized by dichlorofluorescein (DCF) staining. Note rapid elevation of fluorescence (ROS) in the root hairs (rh). Times represent seconds after treatment. C. maculosa (C) roots showed low stable levels of ROS (DCF fluorescence) throughout the (–)-catechin treatment. Ascorbic acid (50 mM) resulted in failure of (–)-catechin to induce ROS production in A. thaliana (D) and C. diffusa (G) (scale bar, 200 μm), and it kept roots viable (E) even at 600 s after catechin treatment. Bracket in (E) indicates the viable CEZ root section. (F) C. diffusa treated with 25 μM H2O2 shows uniform increased DCF fluorescence, indicating the patterns of ROS generation seen in (–)-catechin treated roots are unlikely to be an artifact of dye distribution. Images are representative of ≥15 separate roots.

We next determined whether a cytoplasmic Ca2+ ([Ca2+]cyt)-dependent signaling system (14, 15) was also involved in (–)-catechin action. To image [Ca2+]cyt dynamics, we acid-loaded the roots of C. diffusa and C. maculosa with the fluorescent Ca2+ indicating dye indo-1 (14, 15). For A. thaliana, [Ca2+]cyt measurements were made in roots of plants expressing the cameleon YC2.1 green fluorescent protein (GFP)-based Ca2+ sensor (9). (–)-Catechin treatment resulted in rapid and transient elevations in root tip–localized [Ca2+]cyt levels in A. thaliana and C. diffusa but not C. maculosa seedlings (Fig. 4, A and D; fig. S8). In contrast, (+)-catechin did not induce any detectable change in [Ca2+]cyt in any of the species tested (10). In C. diffusa and A. thaliana, the (–)-catechin–induced increase in [Ca2+]cyt was initiated in the CEZ and meristem 30 s after treatment (Fig. 4) and after the changes in ROS (initiated within 10 s) but before cell death (initiated at 600 s). The mature region of the root showed no Ca2+ change over this time frame but did show Ca2+ increase at 30 to 40 min, paralleling the delayed cell death seen in this region at 60 min after (–)-catechin treatment (Fig. 4, B and E). The transient elevation in [Ca2+]cyt induced by (–)-catechin only in A. thaliana and C. diffusa suggests that Ca2+ signaling may play a role in the phytotoxic action of (–)-catechin in susceptible species. Ascorbic acid (50 mM) blocked these Ca2+ changes, suggesting that the intracellular ROS increase upon (–)-catechin administration is responsible for triggering the Ca2+ transient within the cell (fig. S8). Consistent with this idea, extracellular application of H2O2 produced an increase of [Ca2+]cyt (fig. S8). Thus, our data are consistent with (–)-catechin eliciting ROS-induced Ca2+-dependent events that trigger a program of cell death, initially characterized by loss of ionic homeostasis, such as a failure to maintain cellular pH control (fig. S9).

Fig. 4.

Effect of (–)-catechin (100 μg ml1) on [Ca2+]cyt levels in roots of C. diffusa (A and B), C. maculosa (C), and A. thaliana (D and E). Roots of C. diffusa and C. maculosa were acid-loaded with the fluorescent Ca2+ indicator indo-1, whereas Ca2+ measurements were made in A. thaliana using plants expressing the Cameleon YC2.1 GFP-based indicator. Plants were treated with (–)-catechin, and confocal ratio images were taken at the indicated times (s). C. diffusa and A. thaliana showed a transient elevation in [Ca2+]cyt (arrows) starting approximately 30 s after (–)-catechin treatment in the meristematic region. (A and D) The mature region of the root did not respond, even after almost 1 hour (E). Representative of ≥10 separate roots. Ca2+ levels have been pseudocolor coded according to the inset scale. Scale bar represents 200 μm.

We next analyzed global gene expression in Arabidopsis to define potential transcriptional events associated with (–)-catechin's phytotoxic response (9). After 1 hour of treatment with (–)-catechin (100 μg ml1), 956 genes were induced twofold or greater whereas by 12 hours many of these same genes were repressed, likely reflecting the onset of cell death (fig. S10). A large number of these induced genes were related to oxidative stress and the phenylpropanoid and terpenoid pathways (table S1). We also conducted a global gene expression profile 10 min after (–)-catechin treatment to identify possible transcriptional events involved in (–)-catechin signaling or early response. We found a cluster of 10 genes upregulated 10 min after (–)-catechin treatment (fig. S10; table S1). These genes were associated with a steroid sulfotransferase-like protein, α-cystathionase, calmodulin, a ribosomal protein L9, peroxidase ATP21a, a chlorophyll binding protein, and four uncharacterized genes (fig. S10). These genes may be implicated in plant-specific early signal transduction events linked to oxidative stress. It seems likely that acclimation to oxidative stress generated by ROS signaling after (–)-catechin treatment involves concerted, long-term potentiation of different sets of antioxidant and defense genes.

The case we have presented here for allelopathy in C. maculosa challenges the conventional ecological perspective that a species' invasiveness is mainly due to enhanced resource competition after escape from natural enemies (1) and highlights the role for the biochemical potential of the plant as an important determinant of invasive success.

Supporting Online Material

Materials and Methods

Figs. S1 to S10

Table S1

Movies S1 to S5


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article