HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Cipollini, D. Search for Related Content PubMed Articles by Cipollini, D. Agricola Articles by Cipollini, D.

Variation in the expression of chemical defenses in Alliaria petiolata (Brassicaceae) in the field and common garden¹

Department of Biological Sciences, Wright State University, 3640 Colonel Glenn Highway, Dayton, Ohio 45435 USA

Received for publication September 27, 2001. Accepted for publication April 19, 2002.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
I examined glucosinolates, trypsin inhibitors (TI), and peroxidase (POD) activity in garlic mustard (Alliaria petiolata) plants growing naturally in Wright State University's Forest Preserve and in a common garden experiment in plants from the same populations conducted in the greenhouse. In the field, first-year plants expressed each defense, but defense levels varied significantly in plants from different sites in the forest. Patterns in site variation were consistent for glucosinolate and POD, but not for TI. The TI and POD levels were increased by mechanical wounding, but glucosinolate levels were unaffected. In the greenhouse, plants expressed each defense at higher levels than in the field, but defense levels did not vary among plants collected from each site in the field. The POD activity was increased by wounding, but glucosinolate and TI levels where unaffected. Plants from each site varied in height and leaf length when measured shortly after transplantation, but site differences substantially diminished after 4 wk. Site-based variation in defense expression in the field, which disappeared in the greenhouse, was presumably related to differences in environmental quality among the sites. Sites were shown to vary in soil moisture content, soil pH, nutrient levels, and presumably light quantity or quality. Despite an apparent lack of genetic variation in defense across sites in the field, the constitutive expression of these three chemical defenses, increases due to wounding, and phenotypic variation across sites could reduce herbivore success on garlic mustard individuals and slow the rate of herbivore adaptation to garlic mustard populations.

Key Words: Alliaria petiolata • Brassicaceae • glucosinolates • herbivores • induced defenses • invasive species • peroxidase • trypsin inhibitors

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Garlic mustard, Alliaria petiolata (Bieb.) Cavara & Grande (Brassicaceae), is an introduced biennial herb that is expanding rapidly in northeastern and midwestern forests in the United States and in southern Canada (Nuzzo, 2000 ). Garlic mustard flourishes in mesic woodlands with moderate exposure to light, but it can grow in a diversity of other habitats (Byers and Quinn, 1998 ). Its presence in the field is frequently associated with reduced native plant abundance and/or diversity (McCarthy, 1997 ). Garlic mustard exhibits a high degree of genetic homogeneity among and especially within populations in the midwestern United States (Meekins, Ballard, and McCarthy, 2001 ; S. Schmidt, D. Cipollini, and D. Krane, unpublished data), but it exhibits a high degree of phenotypic plasticity to environmental conditions (Byers and Quinn, 1998 ). It is a good competitor for resources (Meekins and McCarthy, 1999 ), may exude allelopathic chemicals (Kelley and Anderson, 1990 ; but see McCarthy and Hanson, 1998 ), and produces numerous seeds (Cavers, Heagy, and Kokron, 1979 ). It is a combination of these characteristics that most land managers attribute to its success.

Life history attributes associated with resistance to herbivory also appear to contribute to the success of this species in habitats where it is introduced. First, garlic mustard incurs relatively little damage from the native or introduced assemblage of herbivores that typically feed upon brassicaceous plants in the field in North America (Nuzzo, 2000 ; D. Cipollini, personal observation). When damage does occur, populations also appear to vary in the amount of herbivore damage that they receive (Haribal and Renwick, 2001 ; D. Cipollini, personal observation). The success of many invasive plant species may be due to the fact that specialist herbivores that consumed such plants in native habitats did not colonize new environments along with their host plant, permitting plant escape from herbivore pressure (Blossey and Nötzold, 1995 ). As many as 69 different insects are known to consume garlic mustard in its native range in Europe, including five specialists (Szentesi, 1991 ), each of which are lacking in North America. Some herbivores, such as larvae of the introduced cabbage white butterfly (Pieris rapae), will consume field-grown garlic mustard in short-term laboratory bioassays (Renwick and Lopez, 1999 ). However, while adults of P. rapae and related North American native butterflies (such as Pieris napi oleraceae) will oviposit on garlic mustard in the field, the larvae apparently cannot complete development (Bowden, 1971 ; Huang, Renwick, and Chew, 1995 ; Haribal and Renwick, 2001 ; Renwick et al., 2001 ; D. Cipollini, personal observation). Larvae of the introduced cabbage looper moth (Trichoplusia ni) will accept garlic mustard foliage in the laboratory, but grow 50% slower than on either canola (Brassica napus) or wild mustard (Brassica kaber) and cannot complete development (D. Cipollini, unpublished data). In addition, the performance of T. ni varies on leaves of garlic mustard from different sites in the field (D. Cipollini, unpublished data).

Escape from herbivore pressure may contribute to the invasiveness of species such as garlic mustard in at least three non-mutually exclusive ways. First, introduced species lacking herbivores simply rarely experience the well-documented growth- and fitness-suppressing effects of herbivory (Rosenthal and Kotanen, 1994 ) and escape any potential population regulation by herbivores. This is at least partly true for garlic mustard since it experiences little herbivore damage. Second, if escape from herbivore pressure does occur, then selection may favor particularly invasive plant genotypes with greatly reduced constitutive or wound-induced expression of costly and unneeded anti-herbivore chemical defenses. In the absence of benefits resulting from the protection of plants from herbivores, allocation of resources to chemical defense expression is thought to be costly to growth and fitness in plants (e.g., Baldwin, 1998 ). Thus, genotypes with reduced expression of chemical defenses can allocate more resources to growth and reproduction (thus contributing to their competitiveness) than genotypes expressing higher levels of defenses (Blossey and Nötzold, 1995 ). The expression of glucosinolates and cuticular compounds that can influence adult oviposition preference and larval feeding behavior has been examined in garlic mustard (Huang, Renwick, and Chew, 1995 ; Haribal and Renwick, 1998 , 2001 ; Renwick and Lopez, 1999 ; Renwick et al., 2001 ), but the expression of other chemical defenses has been little studied in this plant. Third, invasive genotypes may have retained the ability to express constitutive and wound-inducible chemical defenses, but lack of damage by herbivores simply prevents costly defenses from being induced by wounding to higher levels, leaving more resources available for growth and reproduction (Baldwin, 1998 ). The extent to which garlic mustard expresses wound-inducible chemical defenses has never been demonstrated in order to suggest whether this is a possibility. Given the high degree of genetic homogeneity in this species, observed variation in the amount of damage received by populations of garlic mustard, when it does occur, must be due to such factors as environmental effects on the expression of defenses (Cipollini and Bergelson, 2001 ) and spatial and temporal variation in herbivore loads, but these factors have not been explicitly studied in garlic mustard.

In order to fully address herbivore escape/chemical defense hypotheses about the success of invasive species like garlic mustard, it is important to characterize the production of chemical defenses that can potentially affect herbivore growth and/or the amount of herbivore damage received by the plant. I investigated the expression of one secondary chemical (glucosinolates) and two defense proteins (trypsin inhibitor and peroxidase), known to be toxic or antinutritive for herbivores and pathogens in brassicaceous plants (Broadway and Duffey, 1986 ; Bodnaryk, 1992 ; Stowe, 1998b ), in leaves of garlic mustard plants growing naturally in the field and in plants from the same field sites grown in the common garden in the greenhouse. While garlic mustard is known to express glucosinolates (e.g., Huang, Renwick, and Chew, 1995 ), the expression of trypsin inhibitors and peroxidase activity by this plant has never been examined. The examination of plants growing in diverse habitats in the field enabled a comparison of the expression of these defenses under different environmental conditions, while the greenhouse experiment addressed the degree to which differences seen in the field would persist in a common environment. Phenotypic relationships between levels of these defenses were determined in each experiment to examine the degree of independence of the expression of each of these defenses. In addition to investigating constitutive levels of these defenses, this study is the first to examine the extent to which garlic mustard can alter levels of these defenses in response to wounding and to determine whether wound induction might be influenced by environmental conditions. Furthermore, it has been suggested that wounding can increase variance in chemical levels in wounded plants, in addition to or without affecting mean levels, which can make wounded plants more difficult to efficiently consume and assimilate than unwounded plants. Variation in food quality may decrease individual herbivore performance (Stockhoff, 1993 ) and/or the ability of an herbivore population to adjust physiologically to track a plant population in ecological time or to adapt to a plant population through natural selection (Karban, Agrawal, and Mangel, 1997 ; R. Karban, University of California, Davis, personal communication). Increased variance in chemical levels in wounded plants could contribute to the invasiveness of garlic mustard, if it is related to a reduction in the ability of herbivores to consume and adapt to this plant. The effect of wounding on phenotypic variance in defense traits has been little studied in any plant and is examined for the first time in garlic mustard in this study. Specifically, this study addressed the following questions in two experiments, one conducted on plants growing naturally in Wright State University's Forest Preserve and one conducted on plants collected from the field and performed in the greenhouse. First, do leaves of garlic mustard plants growing naturally in Wright State University's Forest Preserve constitutively express glucosinolates, trypsin inhibitors, and peroxidase activity? Second, does chemical defense expression in these plants vary by site in the forest, and do these differences persist in a common environment? Third, does mechanical wounding of the leaves affect levels of chemical defenses in the field and greenhouse? Fourth, do levels of chemical defenses exhibit more phenotypic variance in leaves of wounded plants than in leaves of unwounded plants? Lastly, are levels of these chemical defenses phenotypically related?

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Field experiment
On 19 November 1999, eight sites supporting populations of garlic mustard were chosen in Wright State University's Forest Preserve (Ohio, USA), located at least 100 m from each other. Sites 1 and 2 were located in a northeast-facing section of the forest, dominated by large white oak, red oak, and sugar maple trees. Sites 3, 4, and 5 were located in a flat stream bottom section of this forest dominated by large oak and sycamore trees. Sites 6, 7, and 8 were located in a southwest-facing section of the forest dominated by younger white ash, black locust, sugar maple, and black cherry trees. Each population had at least 50 first-year individuals present at the time of sampling. These particular sites in the forest were intentionally chosen because they appeared to represent a range of environmental variation. Garlic mustard has been present in this forest for at least 20 yr (J. Runkle, Wright State University, personal communication), and the populations found throughout it exhibit extremely synchronized vegetative and flowering stages. We have found that garlic mustard populations throughout southwestern Ohio exhibit little genetic diversity within and among populations based on several random amplified polymorphic DNA (RAPD) markers (S. Schmidt, D. Cipollini, and D. Krane, Wright State University, unpublished data). Likewise, Meekins, Ballard, and McCarthy (2001) have found that genetic diversity within populations of garlic mustard, including some Ohio populations, can be quite low when assessed using inter-simple sequence repeat (ISSR) markers, although they found higher levels of diversity among populations. Using the same two ISSR primers and the protocol used by Meekins, Ballard, and McCarthy (2001) , we repeatedly found little polymorphism among randomly chosen individuals from each of the eight sites chosen in the forest (S. Walker and D. Cipollini, Wright State University, unpublished data; Fig. 1). This evidence suggests that plants from the different sites chosen in this forest do not represent genetically distinct populations of garlic mustard and have likely originated from one primary introduction (McCarthy, 1997 ). Because leaves had already fallen from surrounding canopy trees, light availability was similar among sites during the field experiment, although light availability may have varied among sites when the canopy was closed. Other site characteristics were not quantified at this time. Within each population, six similarly sized plants exhibiting four fully expanded true leaves and an expanding fifth true leaf were identified. Plants were selected for similarity in size within and among sites, but the age of selected plants was unknown. The fourth true leaf on three of these plants (about 6 cm in length) was wounded by crushing a 0.5 cm band of tissue around the periphery of the leaf with needle-nosed pliers. This technique is known to effectively induce glucosinolates, trypsin inhibitors, and other defenses in related brassicaceous plants, in a manner similar to feeding by herbivores such as larvae of the cabbage looper moth (Broadway and Missurelli, 1990 ; Cipollini and Bergelson, 2000 , 2001 ; D. Cipollini, unpublished data). Induction of these defenses by mechanical wounding is thought to simulate induction by natural herbivory to a large degree, but the absence of herbivore saliva in mechanical wounds and the lack of spatially and temporally variable feeding patterns certainly influences the degree to which this is true (Baldwin and Preston, 1999 ). Despite its limitations, mechanical wounding permits the careful control of the extent and timing of leaf damage. Control plants were unperturbed. Three days later, leaf samples were taken from the wounded leaves on wounded plants and from the matching leaves on unwounded plants, placed individually in 1.5-mL microfuge tubes, and placed on dry ice in a cooler for transport back to the laboratory. Leaf samples were stored at –20°C until chemical analysis. One portion of each leaf sample was subsequently used for the extraction and analysis of glucosinolate content, and another portion was used for the extraction and analysis of trypsin inhibitor (TI) activity, soluble peroxidase (POD) activity, and soluble protein content.

View larger version (94K): [in this window] [in a new window]   Fig. 1. DNA profiles of garlic mustard individuals resulting from the amplification of genomic DNA using the inter-simple sequence repeat (ISSR) primers 17898A and 17898B and the protocol used by Meekins, Ballard, and McCarthy (2001) . Each lane represents a randomly chosen individual from each of the eight sites examined in this study. This analysis was repeated four times with identical results

Greenhouse experiment
On 28 March 2001, several cotyledon-stage garlic mustard seedlings were collected randomly from several locations within each of the eight sites in the forest and brought into the greenhouse. Most, if not all, of these seedlings were offspring of plants present at each site during the field experiment, which set seed during the summer of 2000. At the time of collection, seedlings were approximately 7–10 d old. Eighteen seedlings from each site were transplanted individually to 300-mL square pots in ProMix BX (Grace-Sierra; Premier Horiculture, Inc., Red Hill, Pennsylvania, USA) potting soil. Seedlings were grown on benches in the greenhouse under ambient light supplemented with fluorescent lighting and watered daily. Daylengths were controlled at 14 h : 10 h L : D and temperatures averaged 23 ± 5°C during the experiment. Pots were placed randomly on benches and randomly moved to different locations on the benches every 3 d to minimize microenvironmental effects. To assess variation in size of plants from different sites shortly after transplantation, height of plants and length of the first true leaf were measured on 4 April. To assess whether size differences persisted after several weeks of growth in a common environment, length of the fourth true leaf was measured on 26 April. On 30 April, half of the plants from each site were wounded (N = 9) as plants were wounded in the field. At this time, plants were similar in size and at a similar phenological stage as those used in the field experiment, but were likely much younger, owing to their more rapid growth in the greenhouse than in the field. Leaf samples were harvested from all plants for chemical analysis 3 d after wounding. For glucosinolate analysis, a # 8 cork borer was used to collect two leaf discs from the wounded leaves on wounded plants and the corresponding leaf on unwounded plants, one from each side of the midvein, about 2 cm from the tip of the leaf. Leaf disks were placed individually in 1.5-mL microfuge tubes and placed on dry ice for transport back to the laboratory. One disk was used to determine dry mass, and one was used to determine glucosinolate content, both of which were used to express glucosinolate concentration in this experiment. Another leaf sample was harvested from the same leaf and was used for the extraction and analysis of trypsin inhibitors (TI) activity, peroxidase (POD) activity, and soluble protein content.

Chemical analyses
The portion of each leaf sample used to determine total glucosinolate content in plants from the field was first weighed and then extracted and analyzed using the glucose release method as in Siemens and Mitchell-Olds (1998) . Although this procedure assesses the quantity of total glucosinolates (which can be a diverse mixture of types), garlic mustard is known to contain primarily one type, sinigrin (Renwick and Lopez, 1999 ; J. A. A. Renwick, Boyce Thompson Institute, personal communication). Total glucosinolate concentration in leaf samples was expressed in milligrams of glucose (liberated from extracted and column-bound glucosinolates by the enzyme thioglucosidase) per gram leaf fresh mass. In leaf samples collected from the greenhouse, total glucosinolate concentration was analyzed in one disk using the same protocol, except that glucosinolate concentration was expressed in milligrams of glucose per gram leaf dry mass. Leaf dry mass was determined by drying and weighing the corresponding leaf disk that was taken from each leaf when disks for glucosinolate analysis were taken. Soluble proteins were extracted from the other leaf portion as in Cipollini and Bergelson (2000) , with the exception that soluble proteins were extracted in 0.01 mol/L sodium phosphate buffer, pH 6.8. The TI content in soluble protein extracts was determined by examining the radial diffusion of extracts through a trypsin-containing agar as in Cipollini and Bergelson (2000) . The TI content of extracts was expressed as micrograms of TI per milligram extract protein. The POD activity in soluble protein extracts was determined by following the oxidation of guaiacol in the presence of hydrogen peroxide at 470 nm, as in Cipollini (1998) , in an assay adapted for a microplate reader. The POD activity was expressed as a change () in absorbance at 470 nm · min^–1 · mg extract protein^–1. Soluble protein content was determined in leaf extracts using the Bradford assay (Bradford, 1976 ), using the Bio-Rad protein dye reagent and bovine serum albumin as a standard. All chemical assays were performed in duplicate.

Statistical analyses
To examine whether levels of chemical defenses varied in plants from different sites and wounding treatments, chemical data were analyzed in the field and greenhouse experiments with a two-way univariate ANOVA with site and wounding as main effects, including an interaction term. Because sites were deliberately chosen based upon environmental characteristics, site is considered a fixed effect in all statistical models, as was wounding. Because a few samples were destroyed in handling in each experiment, Type III sums of squares were used. All data were log₁₀ transformed prior to statistical analysis to meet the assumptions of ANOVA. Phenotypic relationships among levels of chemical defenses in each plant were examined using linear regression. To test the hypothesis that variances in chemical levels in wounded plants would be higher than in unwounded plants, comparisons of phenotypic variance in wounded versus unwounded plants were made using a variance ratio test (V_w/V_uw) on log₁₀-transformed data (Ott, 1993 ; R. Karban, University of California, Davis, personal communication). To examine whether seedlings collected from different sites and used in the greenhouse experiment differed in size, height and leaf 1 length were analyzed using one-way ANOVA with site as the main effect, while the phenotypic relationship between these two variables was examined using linear regression. The ratio of height to leaf 1 length was also analyzed with one-way ANOVA with site as the main effect, as was leaf 4 length. To examine whether sites varied in soil characteristics in the field, soil pH, soil moisture percentage, and nitrate nitrogen, phosphate, potassium, and sulfate contents were analyzed first by MANOVA with site as the main effect using the Wilks' lambda test statistic, followed by univariate ANOVA on each variable separately. Mean comparisons after all tests of significance were made using Tukey's tests. All statistical analyses were performed on SAS (Release 6.12, SAS Institute, Cary, North Carolina, USA).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Expression of chemical defenses in the field and in the greenhouse
Glucosinolates
In the field, glucosinolates were detected in leaves of all garlic mustard plants from all sites (Fig. 2A). Glucosinolate levels were similar in plants from most sites, but some sites differed significantly (Site: F_7,32 = 2.62, P = 0.029). In particular, plants from Site 5 had significantly higher glucosinolate levels than plants from Site 8, while plants from Sites 2, 3, 4, 6, and 7 had intermediate levels. Wounding did not significantly affect glucosinolate levels (Wounding: F_1,32 = 0.255, P = 0.617), and there was no significant variation in the effect of wounding across sites (Site x Wounding: F_7,32 = 0.719, P = 0.656). Variance in glucosinolate levels in wounded plants was not different from that in unwounded plants (F_23,23 = 1.32, P > 0.05). Glucosinolate levels were weakly positively related to POD activity levels across all samples in the following manner: square root (glucosinolate level) = 342 x 10^–8 (POD level) + 0.342; F_1,45 = 4.012, P = 0.051, R² = 0.10. In the greenhouse, glucosinolates were detected in leaves of all plants collected in the field and grown in the greenhouse (Fig. 3A). However, there were no significant effects of site (F_7,118 = 0.815, P = 0.577), wounding (F_1,118 = 1.391, P = 0.2407), or their interaction (F_7,118 = 0.536, P = 0.8058) on glucosinolate levels in these plants. Variance in glucosinolate levels in wounded plants was not different from that in unwounded plants (F_67,68 = 1.016, P > 0.05), and glucosinolate levels were not significantly phenotypically related to either POD activity (F_1,131 = 0.0165, P = 0.898) or TI levels (F_1,132 = 0.502, P = 0.480) in plants from all sites.

View larger version (24K): [in this window] [in a new window]   Fig. 2. Levels of chemical defenses in unwounded and wounded garlic mustard plants across eight sites in Wright State University's Forest Preserve, Dayton, Ohio, USA. (A) Glucosinolate levels. (B) Peroxidase activity levels. (C) Trypsin inhibitor levels. Bars represent means + 1 SE. Site means (averaged across wounding treatments) marked with the same letter do not differ significantly at = 0.05

View larger version (31K): [in this window] [in a new window]   Fig. 3. Levels of chemical defenses in unwounded and wounded garlic mustard plants, collected from the same eight sites in Wright State University's Forest Preserve and grown in the greenhouse. (A) Glucosinolate levels. (B) Peroxidase activity levels. (C) Trypsin inhibitor levels. Bars represent means + 1 SE

Trypsin inhibitors
The TIs were also detected in leaves of garlic mustard from all sites in the field (Fig. 2C). The TI levels were similar in plants from most sites, but some sites differed significantly (Site: F_7,29 = 5.46, P = 0.0004). In particular, plants from Site 8 had significantly higher TI levels than plants from Sites 2, 3, 4, and 5, while plants from Sites 1, 6, and 7 had intermediate levels. Wounding significantly increased TI levels in plants across all sites (Wounding: F_1,29 = 42.1, P < 0.0001), by a mean of about 80%. There was no significant variation in the effect of wounding across sites (Site x Wounding: F_7,29 = 0.970, P = 0.498). Variance in TI levels in wounded plants was not significantly different from that in unwounded plants (F_22,22 = 0.455, P > 0.05). The TI levels were not significantly related to either glucosinolate levels (F_1,45 = 0.331, P = 0.568) or POD activity levels (F_1,44 = 0.734, P = 0.396) across all samples. In the greenhouse, TIs were detected in leaves of all plants from all sites (Fig. 3C). However, there was no significant effect of site (F_7,117 = 1.838, P = 0.0864), wounding (F_1,117 = 1.769, P = 0.1861) or their interaction (F_7,117 = 0.733, P = 0.644) on TI levels. Variance in TI levels in wounded plants was not significantly different from that in unwounded plants (F_68,68 = 1.017, P > 0.05).

Growth in the greenhouse
When measured 6 d after transplanting, plants from different sites significantly varied in length of the first true leaf (F_7,136 = 14.70, P < 0.0001; Fig. 4A) and height (F_7,136 = 17.66, P < 0.0001; Fig. 4B). However, these two variables were significantly negatively related to one another in the following manner: (leaf 1 length = –0.190 [height] + 1.28; F_1,142 = 58.8, P < 0.0001, R² = 0.29). A negative relationship between these variables is indicative of the expression of shade-avoidance behavior among seedlings in the experiment (Dudley and Schmitt, 1996 ). Thus, the ratio of these two variables (height: leaf 1 length) was analyzed and proved to significantly vary among sites (F_7,136 = 19.54, P < 0.0001; Fig. 4D), indicating that plants from different sites exhibited different degrees of the expression of shade-avoidance behavior and thus were likely exposed to different light conditions in the field. Specifically, height : leaf 1 length was highest in plants from Site 2 and lowest in plants from Sites 1, 5, 7, and 8. Height : leaf 1 length in plants from Sites 4 and 6 was distinctly intermediate between these two groups, while height : leaf 1 length in plants from Site 3 was significantly lower than in plants from Site 2, but was similar to that in plants from all other sites. Height did not change appreciably after the first measurement date (data not shown). When measured 22 d after transplanting (4 d prior to wounding), plants from different sites significantly varied in length of the fourth true leaf (F_7,136 = 3.45, P = 0.002; Fig. 4C). Site differences in length of the fourth true leaf taken at this time were entirely consistent with site differences in length of the first true leaf measured shortly after transplantation, but were much smaller in magnitude.

View larger version (50K): [in this window] [in a new window]   Fig. 4. Growth characteristics of garlic mustard plants collected from eight sites in Wright State University's Forest Preserve and grown in the greenhouse. (A) Length of the first true leaf 6 d after transplantation. (B) Height of plants 6 d after transplantation. (C) Length of the fourth true leaf 22 d after transplantation. (D) Ratio of height: leaf 1 length of plants 6 d after transplantation. Bars represent means + 1 SE. Site means marked with the same letter do not differ significantly at = 0.05

View this table: [in this window] [in a new window]   Table 1. Soil characteristics of each of eight sites in Wright State University Forest Preserve, Dayton, Ohio, USA, where Alliaria petiolata was examined in this study. Numbers indicate mean ± 1 SE of three samples. Shared letters within columns indicate means that are not significantly different ( = 0.05)

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In the field, patterns in the levels of glucosinolates and POD activity were strikingly similar across sites, and indeed, levels of these two defenses were weakly but positively phenotypically correlated across all samples. In the greenhouse, however, levels of these defenses were not phenotypically correlated. Glucosinolates and POD are not closely related structurally, but their production may be regulated in a similar way by certain environmental factors, such that these two defenses will be phenotypically correlated under diverse conditions in the field, but not under uniform conditions in the greenhouse. In contrast, POD and TI activity levels were strongly positively correlated in the greenhouse, but not in the field. This indicates that the production of TI and POD might be physiologically linked in some basic way (e.g., they are both proteins) such that levels of these defenses exhibit significant phenotypic correlations in a common environment. However, their production may be regulated in the field by different environmental factors (e.g., POD requires iron) such that phenotypic correlations are lost in the field. Patterns in the correlation between levels of closely related defense compounds have been examined in some plants (e.g., for different types of furanocoumarins in wild parsnip) (Zangerl and Berenbaum, 1990 ), but have never been examined for these particular defenses. These results illustrate that chemical defense expression can vary among populations of garlic mustard in the field even in the apparent absence of genetic variation, and that, regardless of the mechanism, the patterns of correlations among chemical defenses can change depending upon environmental conditions (both among sites and among field and greenhouse conditions). As suggested by Haribal and Renwick (2001) , variability in chemical phenotype may partly explain observed variation in the amount of herbivore damage received by different garlic mustard populations.

Mechanical wounding markedly increased TI and POD activity levels in garlic mustard in the field, which is consistent with its effects on other brassicaceous plants, as well as on plants from other families (Broadway and Missurelli, 1990 ; Karban and Baldwin, 1997 ; Cipollini and Bergelson, 2000 ; D. Cipollini and M. Sipe, Wright State University, unpublished data). Although wounding induced smaller, but significant, increases in POD activity in the greenhouse as well, it failed to significantly induce increases in TI levels in the greenhouse, which was surprising given that wounding has been shown to induce TI in other brassicaceous plants in the greenhouse (Cipollini and Bergelson, 2000 ). Since the wounding protocol was the same in the greenhouse and in the field, some aspect of the environment (e.g., UV light exposure) may have made plants more responsive to wounding in the field. Constitutive expression of all defenses was also higher in the greenhouse, which may have related to age differences between plants used in the field and greenhouse experiments (Stout et al., 1996 ; Cipollini and Redman, 1999 ) or environmental differences between the field and greenhouse. Regardless of the mechanism causing the differences, high constitutive expression of defense may have restricted or obscured inducibility, which has been noted in some plants (Stout et al., 1996 ). Wounding appeared to induce increases in glucosinolate levels in plants from some sites, and wounding or elicitor treatment is known to induce glucosinolates in other brassicaceous plants (Bodnaryk, 1992 ; Cipollini and Sipe, 2001 ), but this effect was not significant in the field nor in the greenhouse. Wounding also had no significant effect on phenotypic variance in levels of any of these defenses, so it does not appear to add variability to the chemical phenotype of wounded plants. These findings indicate that garlic mustard is capable of coordinately heightening the expression of at least two chemical defenses in response to wounding (and presumably herbivore attack) in the field, although the manner and degree to which it does so likely varies depending upon environmental conditions and the nature of the damage incurred (Baldwin and Preston, 1999 ). Thus, in addition to expressing several defenses constitutively, induced increases in chemical defense levels should further reduce the quality of wounded garlic mustard as food for herbivores that attempt to consume it (Karban and Baldwin, 1997 ). Changes in defensive chemical phenotype depending upon the presence of herbivory are thought to be a form of adaptive phenotypic plasticity, similar to environmentally induced adaptive changes in morphology (Dudley and Schmitt, 1996 ) that maximize plant fitness in environments with and without herbivory (Cipollini, 1998 ; Agrawal, 1999 ; Baldwin, 1999 ). Garlic mustard is known to exhibit extensive morphological plasticity (Byers and Quinn, 1998 ) but is shown to exhibit defensive plasticity for the first time here.

Overall, this study suggests that the escape from herbivory by garlic mustard in North America is likely due to a combination of the lack of specialist herbivores and the expression of constitutive and wound-inducible chemical defenses that can act against other herbivores. These results do not support the hypothesis that introduced garlic mustard genotypes have lost the ability to express at least some constitutive or wound-inducible chemical defenses, although levels of expression and degree of inducibility may differ in plants from native populations. While the role of different chemical defenses in herbivore resistance is not clear in this system, expression of the defenses illustrated in this study, along with other defenses not characterized (see Haribal and Renwick, 1998 , 2001 ; Renwick et al., 2001 ), may largely explain the relative escape of garlic mustard from herbivore attack and lack of success by herbivores that attempt to consume it (Bowden, 1971 ). For example, although glucosinolates are used in host detection by specialist insects and their presence stimulates adult oviposition and larval feeding (Renwick and Lopez, 1999 ), high glucosinolate levels, as were found in garlic mustard, can reduce the performance of many generalist and even some specialist insects (Stowe, 1998b ). In turn, while introduced garlic mustard genotypes constitutively express several presumably costly defenses (Stowe, 1998a ; Cipollini, 2002), some of which are also wound-inducible, lack of damage by herbivores prevents these defenses from being induced to higher levels, such that growth and fitness can be maximized in the absence of herbivory (Baldwin, 1998 ). Thus, garlic mustard populations rarely experience either the growth- and fitness-reducing effects of herbivory (Rosenthal and Kotanen, 1994 ) or the growth- and fitness-reducing effects of increased allocation to inducible chemical defenses (Baldwin, 1998 ; Cipollini, 2002). Variation in levels and in correlation patterns among defenses by environment may partly explain variation in the amount of damage that garlic mustard plants do receive (Haribal and Renwick, 2001 ), especially since herbivores that consume brassicaceous plants are known to be differentially affected by particular chemical defenses (Agrawal, 2000 ). Although wounding did not increase variance in chemical levels in wounded plants in this study, site-based variation in levels of chemical defenses, along with changes in mean levels due to wounding, could constrain the ability of herbivore populations to consume and adapt to garlic mustard populations. Specific effects of the expression of chemical defenses by native and introduced populations of garlic mustard on herbivore success and plant fitness are the subjects of current study.

	FOOTNOTES

 
¹ The author thanks Caleb Slemmons, Jeremiah Busch, Neela Kumar, Christopher Perkins, Sara Walker, and Stephanie Enright for technical support, Kendra Cipollini for editorial comments, and Rick Karban for discussion. Aaron Ellison and one anonymous reviewer provided valuable comments on the manuscript. This research was supported by a grant from the Ohio Board of Regents, Research Challenge New Investigator program.

	LITERATURE CITED

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
Agrawal A. A. 1999 Induced plant defense: evolution of induction and adaptive phenotypic Plasticity. In A. A. Agrawal, S. Tuzun, and E. Bent [eds.], Induced plant defense against pathogens and herbivores: biochemistry, ecology, and agriculture. APS Press, St. Paul, Minnesota, USA

Agrawal A. A. 2000 Specificity of induced resistance in wild radish: causes and consequences for two specialist and two generalist caterpillars. Oikos 89: 493-500[CrossRef][ISI]

Baldwin I. T. 1998 Jasmonate-induced responses are costly but benefit plants under attack in native populations. Proceedings of the National Academy of Sciences, USA 95: 8113-8118[Abstract/Free Full Text]

Baldwin I. T. 1999 Inducible nicotine production in native Nicotiana as an example of adaptive phenotypic plasticity. Journal of Chemical Ecology 25: 3-30

Baldwin I. T. C. A. Preston 1999 The eco-physiological complexity of plant responses to herbivores. Planta 208: 137-145[CrossRef][ISI]

Blossey B. R. Nötzhold 1995 Evolution of increased competitive ability in invasive non-indigenous plants: a hypothesis. Journal of Ecology 83: 887-889[CrossRef]

Bodnaryk R. P. 1992 Effects of wounding on glucosinolates in the cotyledons of oilseed rape and mustard. Phytochemistry 31: 2671-2677[CrossRef][ISI]

Bowden S. R. 1971 American white butterflies (Pieridae) and English food plants. Journal of the Lepidopterists' Society 25: 6-12

Bradford M. M. 1976 A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Analytical Biochemistry 72: 248-254[CrossRef][ISI][Medline]

Broadway R. M. S. S. Duffey 1986 Plant proteinase inhibitors: mechanism of action and effect on the growth and digestive physiology of larval Heliothis zea and Spodoptera exigua. Journal of Insect Physiology 32: 827-833[CrossRef][ISI]

Broadway R. M. E. L. Missurelli 1990 Regulatory mechanisms of tryptic inhibitory activity in cabbage plants. Phytochemistry 29: 3721-3725[CrossRef][ISI]

Byers D. L. J. A. Quinn 1998 Demographic variation in Alliaria petiolata (Brassicaceae) in four contrasting habitats. Journal of the Torrey Botanical Society 125: 138-149[CrossRef][ISI]

Cavers P. B. M. I. Heagy R. F. Kokron 1979 The biology of Canadian weeds. 35. Alliaria petiolata (M. Bieb.) Cavara and Grande. Canadian Journal of Plant Science 59: 217-229[ISI]

Cipollini D. F. 1998 Induced defenses and phenotypic plasticity. Trends in Ecology and Evolution 13: 200.

Cipollini D. F. 1998 The induction of soluble peroxidase activity in bean leaves by wind-induced mechanical perturbation. American Journal of Botany 85: 1586-1591[Abstract/Free Full Text]

Cipollini D. F. 2002 Does compeletion magnify the fitness costs of induced responses in Arabidopsis thaliana? A manipulative approach. Oecologia 131: 514-520[CrossRef][ISI]

Cipollini D. F. J. Bergelson 2000 Environmental and developmental regulation of trypsin inhibitor activity in Brassica napus. Journal of Chemical Ecology 26: 1411-1422[CrossRef][ISI]

Cipollini D. F. J. Bergelson 2001 Plant density and nutrient availability constrain the constitutive and wound-induced expression of trypsin inhibitor activity in Brassica napus. Journal of Chemical Ecology 27: 593-610[CrossRef][ISI][Medline]

Cipollini D. F. A. M. Redman 1999 Age-dependent effects of jasmonic acid treatment and wind exposure on foliar oxidase activity and insect resistance in tomato. Journal of Chemical Ecology 25: 271-281[CrossRef][ISI]

Cipollini D. F. M. Sipe 2001 Jasmonic acid treatment and mammalian herbivory differentially affect chemical defenses and growth of wild mustard (Brassica kaber). Chemoecology 11: 137-143[CrossRef][ISI]

Dudley S. A. J. Schmitt 1996 Testing the adaptive plasticity hypothesis: density-dependent selection on manipulated stem length in Impatiens capensis. American Naturalist 147: 445-465[CrossRef][ISI]

Haribal M. J. A. A. Renwick 1998 Isovitexin 6''-O-ß-D-glucopyranoside: a feeding deterrent to Pieris napi oleracea from Alliaria petiolata. Phytochemistry 47: 1237-1240[CrossRef][ISI]

Haribal M. J. A. A. Renwick 2001 Seasonal and populational variation in flavonoid and alliarinoside from Alliaria petiolata. Journal of Chemical Ecology 27: 1603-1612

Huang X. P. J. A. A. Renwick F. S. Chew 1995 Oviposition stimulants and deterrents control acceptance of Alliaria petiolata by Pieris rapae and P. napi oleracea. Chemoecology 5/6,2: 79-87

Karban R. A. A. Agrawal M. Mangel 1997 The benefits of induced defenses against herbivores. Ecology 78: 1351-1355[CrossRef][ISI]

Karban R. I. T. Baldwin 1997 Induced responses to herbivory. University of Chicago Press, Chicago, Illinois, USA

Kelley T. R. C. Anderson 1990 Examination of the allelopathic properties of garlic mustard (Alliaria petiolata). Transactions of the Illinois State Academy of Science 83: 31-32

McCarthy B. 1997 Response of a forest understory community to experimental removal of an invasive nonindigenous plant (Alliaria petiolata, Brassicaceae). In J. O. Luken and J. W. Thieret [eds.], Assessment and management of plant invasions, 117–130. Springer-Verlag, New York, New York, USA

McCarthy B. C. S. L. Hanson 1998 An assessment of the allelopathic potential of the invasive weed Alliaria petiolata (Brassicaceae). Castanea 63: 68-73

Meekins J. F. H. E. Ballard Jr. B. C. McCarthy 2001 Genetic variation and molecular biogeography of a north american invasive plant species (Alliaria petiolata, Brassicaceae). International Journal of Plant Sciences 162: 161-169[CrossRef]

Meekins J. F. B. C. McCarthy 1999 Competitive ability of Alliaria petiolata (garlic mustard, Brassicaceae), an invasive nonindigenous forest herb. International Journal of Plant Sciences 160: 743-752[CrossRef][ISI]

Nuzzo V. A. 2000 Element stewardship abstract for Alliaria petiolata (Alliaria officinalis), Garlic Mustard. Nature Conservancy, Arlington, Virginia, USA

Ott R. L. 1993 An introduction to statistical methods and data analysis, 4th ed. Duxbury Press, Belmont, California, USA

Renwick J. A. A. K. Lopez 1999 Experience-based food consumption by larvae of Pieris rapae: addiction to glucosinolates?. Entomologia Experimentalis et Applicata 91: 51-58

Renwick J. A. A. W. Zhang M. Haribal A. B. Attygalle K. D. Lopez 2001 Dual chemical barriers protect a plant against different larval stages of an insect. Journal of Chemical Ecology 27: 1575-1583[CrossRef][ISI][Medline]

Rosenthal J. P. P. M. Kotanen 1994 Terrestrial plant tolerance to herbivory. Trends in Ecology and Evolution 9: 145-148[CrossRef]

Siemens D. H. T. Mitchell-Olds 1998 Evolution of pest defenses in Brassica plants: tests of theory. Ecology 79: 632-646[CrossRef][ISI]

Stockhoff B. A. 1993 Diet heterogeneity: implications for growth of a generalist herbivore, the gypsy moth. Ecology 74: 1939-1949[CrossRef][ISI]

Stout M. J. K. V. Workman J. S. Workman S. S. Duffey 1996 Temporal and ontogenetic aspects of protein induction in foliage of the tomato, Lycopersicon esculentum. Biochemical Systematics and Ecology 24: 611-625[CrossRef][ISI]

Stowe K. 1998a Experimental evolution of resistance in Brassica rapa: correlated response of tolerance in lines selected for glucosinolate level. Evolution 52: 703-712[CrossRef][ISI]

Stowe K. 1998b Realized defense of artificially selected lines of Brassica rapa: effects of quantitative genetic variation in glucosinolate concentration. Environmental Entomology 27: 1166-1174[ISI]

Szentesi A. 1991 Controversial components of plant apparency in Alliaria petiolata Cavara & Grande (Cruciferae). Symposium Biologia, Hungarica 39: 237-244

Zangerl A. R. M. R. Berenbaum 1990 Furanocoumarin induction in wild parsnip: genetics and populational variation. Ecology 71: 1933-1940[CrossRef][ISI]

This article has been cited by other articles:

				S. M. Enright and D. Cipollini Infection by powdery mildew Erysiphe cruciferarum (Erysiphaceae) strongly affects growth and fitness of Alliaria petiolata (Brassicaceae) Am. J. Botany, November 1, 2007; 94(11): 1813 - 1820. [Abstract] [Full Text] [PDF]

				O. Bossdorf, S. Schroder, D. Prati, and H. Auge Palatability and tolerance to simulated herbivory in native and introduced populations of Alliaria petiolata (Brassicaceae) Am. J. Botany, June 1, 2004; 91(6): 856 - 862. [Abstract] [Full Text] [PDF]

				J. E. Taylor, P. E. Hatcher, and N. D. Paul Crosstalk between plant responses to pathogens and herbivores: a view from the outside in J. Exp. Bot., January 2, 2004; 55(395): 159 - 168. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (1) Citing Articles via Google Scholar Google Scholar Articles by Cipollini, D. Search for Related Content PubMed Articles by Cipollini, D. Agricola Articles by Cipollini, D.