Comparative Analysis of Quantitative Trait Loci Controlling Glucosinolates, Myrosinase and Insect Resistance in Arabidopsis thaliana

Corresponding author: Thomas Mitchell-Olds, Max Planck Institute for Chemical Ecology, Winzerlaer Strasse 10, D-07745 Jena, Germany., tmo{at}ice.mpg.de (E-mail)

Communicating editor: J. A. BIRCHLER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Evolutionary interactions among insect herbivores and plant chemical defenses have generated systems where plant compounds have opposing fitness consequences for host plants, depending on attack by various insect herbivores. This interplay complicates understanding of fitness costs and benefits of plant chemical defenses. We are studying the role of the glucosinolate-myrosinase chemical defense system in protecting Arabidopsis thaliana from specialist and generalist insect herbivory. We used two Arabidopsis recombinant inbred populations in which we had previously mapped QTL controlling variation in the glucosinolate-myrosinase system. In this study we mapped QTL controlling resistance to specialist (Plutella xylostella) and generalist (Trichoplusia ni) herbivores. We identified a number of QTL that are specific to one herbivore or the other, as well as a single QTL that controls resistance to both insects. Comparison of QTL for herbivory, glucosinolates, and myrosinase showed that T. ni herbivory is strongly deterred by higher glucosinolate levels, faster breakdown rates, and specific chemical structures. In contrast, P. xylostella herbivory is uncorrelated with variation in the glucosinolate-myrosinase system. This agrees with evolutionary theory stating that specialist insects may overcome host plant chemical defenses, whereas generalists will be sensitive to these same defenses.

PLANT chemical defense systems and their impact on specialist vs. generalist insect herbivores have intrigued scientists for decades. It is clear that some compounds elicit contrasting behavioral responses from various insect herbivores (CHEW and RENWICK 1994 ). Chemical coevolution theory suggests that specialist insects have adapted to withstand and even utilize some plant defensive chemicals, which nevertheless function as feeding deterrents to generalist herbivores (EHRLICH and RAVEN 1964 ; BERENBAUM and ZANGERL 1992 ). Several researchers have shown that secondary plant compounds that deter feeding of generalist herbivores also stimulate feeding and provide oviposition cues for specialist feeders (DA COSTA and JONES 1971 ; FEENY 1976 ; CHEW and RENWICK 1994 ). However, understanding insect-plant interactions is complicated because different compounds within a chemical class can have heterogeneous effects on specialist herbivores (BOWERS and PUTTICK 1988 ; BARTLET et al. 1994 ). Additionally, these compounds can interact synergistically to alter herbivory patterns (BERENBAUM and NEAL 1985 ).

The glucosinolate-myrosinase system is believed to protect plants from herbivore damage (CHEW 1988 ; GIAMOUSTARIS and MITHEN 1995 ). Glucosinolates are amino-acid-derived thioglycosides. Glucosinolates and their hydrolyzing agent, myrosinase, are spatially separated within plant cells (BONES and ROSSITER 1996 ). When the cell is disrupted, myrosinase cleaves the sugar from the glucosinolate, and a series of toxic compounds are released. These toxins include nitriles, isothiocyanates, oxozaladines, and epithioalkanes. The fact that toxins are produced only when tissues are damaged suggests that they function in plant defense.

Effects of the glucosinolate-myrosinase system on specialist and generalist herbivores display heterogeneous results that do not strictly adhere to chemical defense theory. For example, increasing glucosinolate levels in Brassica juncea reduced feeding by a generalist lepidopteran herbivore, Spodoptera eridania, while the specialist Plutella xylostella was unaffected by glucosinolate concentration in B. juncea. Further, increased glucosinolate levels in B. rapa also led to decreased feeding by both the specialist, Pieris rapae, and the generalist, Trichoplusia ni (STOWE 1998 ). Elevated myrosinase levels had the opposite effect and decreased herbivory by the specialist, P. xylostella (LI et al. 2000 ). In contrast, elevated glucosinolate levels inhibited feeding of P. xylostella on B. rapa plants (SIEMENS and MITCHELL-OLDS 1996 ). Many studies have shown that specialist herbivores are attracted by glucosinolates and their breakdown products (HUANG and RENWICK 1994 ; PIVNICK et al. 1994 ; STADLER et al. 1995 ; BARTLET et al. 1997 ; ROJAS 1999 ; GRIFFITHS et al. 2001 ; MOYES and RAYBOULD 2001 ). In contrast, these compounds may play a defensive role by attracting a parasitoid of aphid herbivores (BRADBURNE and MITHEN 2000 ). Simultaneous analysis of the joint effects of glucosinolates and myrosinase together might clarify the defensive role of this secondary metabolic system.

Effects of the glucosinolate-myrosinase system on generalist and specialist herbivores may be clarified by elucidating genetic control of defensive physiology and its effects on herbivory by generalists and specialists. Arabidopsis thaliana ecotypes differ with respect to glucosinolate content and composition, providing a suitable system for quantitative genetics (KLIEBENSTEIN et al. 2001A , KLIEBENSTEIN et al. 2001B ). Recombinant inbred (RI) lines in A. thaliana have been used to map loci controlling the glucosinolate-myrosinase system (MITHEN et al. 1995 ; MITCHELL-OLDS and PEDERSEN 1998 ; KLIEBENSTEIN et al. 2001A ). Quantitative trait locus (QTL) mapping of specialist and generalist herbivory responses on these same RI lines would uncover genes governing insect feeding responses. The relative effect of each locus could be quantified and interactions among the loci elucidated. In addition, the relative importance of the glucosinolate system can be compared for specialist and generalist insects.

To analyze genetic variation underlying defense against specialist and generalist insect herbivores, we measured feeding rates of T. ni (cabbage looper, T. ni, generalist) and P. xylostella (diamondback moth, P. xylostella, specialist) lepidopteran herbivores on Arabidopsis Ler x Col and Ler x Cvi RI lines. T. ni larvae have a wide host range that includes Brassica crops, while P. xylostella larvae feed only on Brassicaceae (SHOREY et al. 1962 ). Quantitative levels of herbivore damage were used to map insect resistance QTL. The QTL for insect feeding were then compared to QTL regulating the glucosinolate-myrosinase system, to test the prediction that glucosinolates and myrosinase have larger effects on generalist insect herbivores than on crucifer specialists.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Plant and insect growth conditions:
All plant lines were obtained from the Arabidopsis Stock Center (Nottingham, UK). Plants were grown in potting soil mix with timed-release fertilizer (Osmocote). The plants were thinned to a density of one plant per cell in a 96-cell flat (507 plants m^-2) and grown under 14-hr day length with cool white and GrowLux fluorescent bulbs in a controlled environment growth room. After planting, flats were cold stratified at 4° for 5 days and then moved to the growth room. After 4 weeks (before the onset of flowering) the plants were analyzed for insect herbivory. P. xylostella eggs were obtained from Anthony Shelton (Department of Entomology, New York State Agricultural Experimental Station, Geneva, NY) and raised on an artificial diet according to published procedures (PEREZ and SHELTON 1997 ). T. ni eggs were obtained from Entopath (Easton, PA) and reared on Southland artificial diet obtained from Entopath.

Measuring insect herbivory:
Plant diameter was measured when the plants were 4 weeks old; then a single first instar P. xylostella or T. ni larva was placed on each rosette for 48 hr. The insects were taken directly from artificial growth medium and placed on the plants without a starvation period. The percentage of the rosette removed by the insect was estimated by eye with the aid of a transparent 1-cm² grid. Larvae were allowed to roam at will during the experiment. However, there was at least one insect on >95% of the plants at the end of each experiment. Additionally, all plants were investigated for the presence of insect larvae, and any plants lacking an insect were noted and removed from analysis. There was no significant variation among families for the proportions of insects remaining for the full 48 hr (our unpublished data).

Experimental design:
Insect herbivory assays were carried out on the parental ecotypes (Col, Ler, and Cvi) to determine whether they differed genetically. The percentage of the rosette eaten by a single insect larva over 48 hr was measured on at least 30 plants from each ecotype. In addition, 95 RI lines from the Col x Ler cross and 160 lines from the Cvi x Ler cross were scored for herbivory damage (LISTER and DEAN 1993 ; ALONSO-BLANCO et al. 1998B ). From the complete set of 300 Ler x Col RI lines, we chose 95 lines that were maximally informative for mapping previously identified myrosinase activity QTL (MITCHELL-OLDS and PEDERSEN 1998 ).

We used a randomized complete blocks design with 10–13 replicates for both populations . For the Ler x Col RI lines, each 96-cell flat contained one plant from each of the 95 lines being tested and a Col plant. For the Ler x Cvi RI population, pairs of 96-cell flats were analyzed as a single replicate to enable the use of all 160 lines. The 160 lines were divided equally among the two flats and a Ler and Cvi plant were planted in both flats. This was repeated independently to analyze both T. ni and P. xylostella herbivory. Mapping data for the Ler x Col and Ler x Cvi RI lines were obtained from the Nottingham Stock Center (http://nasc.nott.ac.uk/).

Aliphatic glucosinolate QTL mapping:
A total of 300 Ler x Col RI lines were grown for 4 weeks in a randomized design three independent times (LISTER and DEAN 1993 ). The aliphatic glucosinolates were then assayed from each plant using the previously described high-throughput methodologies (KLIEBENSTEIN et al. 2001A , KLIEBENSTEIN et al. 2001B ). QTL were mapped using the same marker data sets and techniques as for the insect herbivory QTL.

Statistical methods:
Genetic variation among RI lines was analyzed as randomized complete blocks ANOVA using the model HERBIVORY = CONSTANT + FLAT + LINE + SIZE x SIZE. SIZE is a covariate included to control for developmental differences that may occur among individuals of the same line due to size-related environmental causes. Because lines were not replicated within replicates, it was necessary to assume that LINE x FLAT interaction was absent. QTL location and effects were estimated by utilizing the family mean for each RI line in conjunction with both interval mapping and composite interval mapping in QTL Cartographer (BASTEN et al. 1999 ). The genome-wide 5% significance threshold was estimated by randomly reshuffling the phenotypic data 500 times in QTL Cartographer (BASTEN et al. 1999 ). Epistatic interactions were tested with SYSTAT by ANOVA, utilizing the mean phenotypic value for each line. Only markers that were individually significant were tested for epistasis.

Genetic correlations, r_G, were estimated from the Pearson product-moment correlation coefficient among family means (typically using the least-squares family means, controlling for flat effects in ANOVA). When traits (such as resistance to several insect species) are measured in separate experiments, r_G provides an unbiased estimator of the genetic correlation (FALCONER and MACKAY 1996 ). In addition, we used ANCOVA to test whether the correlation between P. xylostella and T. ni damage in the Ler x Cvi RI lines was controlled by a QTL near erecta:

where ERECTA is a categorical variable indicating genotype at the erecta locus. If a significant regression of P. xylostella resistance onto T. ni resistance is found in model 1, but is not significant in model 2, then the correlation between resistance to these herbivores is attributable to the QTL located near the erecta.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Variation among parental ecotypes for insect herbivory:
Comparison of the mean levels of herbivore damage for T. ni and P. xylostella feeding showed significant differences between Landsberg erecta (Ler), Cape Verdi Islands (Cvi), and Columbia (Col; Fig 1). All three ecotypes had significantly different T. ni herbivory levels (N = 91, R² = 0.31, P < 0.0001). Cvi was the most resistant while Ler was the most susceptible. P. xylostella feeding also varied significantly, with Ler showing more resistance than Col and Cvi, which had nearly identical feeding scores (Fig 1). These results indicate that resistance to insect herbivory varies among Ler, Col, and Cvi.

View larger version (20K): In this window In a new window Download PPT slide   Figure 1. Natural variation in insect resistance. (A) The mean resistance of the Col, Ler, and Cvi ecotypes to T. ni herbivory measured after 50 hr. (B) The mean resistance of the Col, Ler, and Cvi ecotypes to P. xylostella herbivory measured after 48 hr. The resistance score was determined as described (STOTZ et al. 2000 ).

Variation among RI lines for insect herbivory:
Analysis of rosette damage by P. xylostella and T. ni larvae on 96 Ler x Col RI lines and 160 Ler x Cvi RI lines (Table 1) showed significant differences, indicating that these RI populations can be used to identify loci mediating insect defense (LISTER and DEAN 1993 ; ALONSO-BLANCO et al. 1998A ).

QTL regulating resistance to T. ni herbivory:
The mean levels of feeding damage per line were utilized for mapping QTL controlling resistance to T. ni herbivory. In both Ler x Col and Ler x Cvi, three QTL affecting the level of T. ni herbivory were mapped (Fig 2). None of the QTL overlapped between the two crosses, indicating that at least six loci controlling T. ni resistance segregate in these populations (Fig 2). Increased susceptibility to T. ni herbivory was mediated by the Ler alleles at QTL in the Ler x Col cross and at two of the three QTL in the Ler x Cvi cross (Fig 2). The Ler alleles at the QTL near AOP and EC198L caused increased susceptibility. All six QTL alter insect herbivory by 10–20% (Fig 2).

View larger version (23K): In this window In a new window Download PPT slide   Figure 2. QTL controlling resistance to T. ni herbivory. LOD plots of QTL controlling resistance of 4-week-old plants to T. ni herbivory obtained by composite interval mapping. The five chromosomes are marked by their corresponding Roman numeral and the axes are independently scaled to accommodate chromosome length and maximum LOD scores. The horizontal line represents the genome-wide P = 0.05 LOD score as determined by 500 random permutations of the data. Each QTL is labeled with the marker exhibiting the highest significance and the percentage effect of the Ler allele on resistance. Positive values indicate that plants with the Ler allele increase herbivory resistance in comparison to those carrying the Cvi allele and vice versa for a negative effect. (Left) Map for the Ler x Col RI population. (Right) Map for the Ler x Cvi RI population.

T. ni herbivory is negatively correlated with myrosinase activity:
In the Ler x Col RI population there are two QTL controlling myrosinase activity (MITCHELL-OLDS and PEDERSEN 1998 ). The NCC1 myrosinase QTL cosegregates with one of the T. ni QTL (MITCHELL-OLDS and PEDERSEN 1998 ). Comparison of myrosinase activity with T. ni herbivory revealed a modest but significant negative genetic correlation between the two (Fig 3, r_G = -0.40, P < 0.001, N = 96; MITCHELL-OLDS and PEDERSEN 1998 ), indicating that T. ni herbivory decreases with increasing levels of myrosinase activity. Myrosinase QTL have not been mapped in the Ler x Cvi RI population.

View larger version (23K): In this window In a new window Download PPT slide   Figure 3. Glucosinolate hydrolysis negatively impacts T. ni herbivory. A plot of mean T. ni herbivory vs. mean myrosinase activity for 96 Ler x Col RI lines. The oval is the 95th percentile space obtained by least-squares regression.

T. ni herbivory is negatively correlated with glucosinolate concentration:
In the Ler x Cvi RI population, three QTL regulate leaf aliphatic glucosinolate concentration (KLIEBENSTEIN et al. 2001A ). Two of these QTL, AOP and EC198L, are in the same region as T. ni herbivory QTL, suggesting that aliphatic glucosinolates deter T. ni herbivory (Fig 2 and KLIEBENSTEIN et al. 2001A ). Comparison of leaf aliphatic glucosinolate concentrations vs. the rate of T. ni herbivory in the Ler x Cvi RI lines showed a strong negative genetic correlation between concentrations of aliphatic glucosinolates and T. ni herbivory (Fig 4A, r_G = -0.60, P < 0.001, N = 160). ANOVA comparing the three QTL for glucosinolate concentration (AOP, EC198L, and Elong) in Ler x Cvi to T. ni herbivory showed statistical significance, suggesting that the three glucosinolate loci significantly alter T. ni herbivory (Table 2). The relationship between these QTL appears to be the same for regulation of glucosinolates and herbivory responses: AOP interacts epistatically with EC198L and Elong to regulate leaf aliphatic glucosinolate concentration and T. ni herbivory (ANOVA, Table 2; KLIEBENSTEIN et al. 2001A ). Therefore, glucosinolate concentration appears to be a major determinant of resistance to T. ni herbivory in the Ler x Cvi RI population.

View larger version (18K): In this window In a new window Download PPT slide   Figure 4. Glucosinolates negatively impact T. ni herbivory. Comparison of the mean total leaf aliphatic glucosinolate concentration and insect herbivory. Notice the greater concentration of total aliphatic glucosinolates in Ler x Cvi (scaling of the horizontal axis). (A) Plot of T. ni herbivory vs. leaf aliphatic glucosinolates in Ler x Cvi. Diagonal line is the linear model obtained by least-squares regression. (B) Plot of T. ni herbivory vs. leaf aliphatic glucosinolates in Ler x Col.

In the Ler x Col RI population, the correlation between T. ni herbivory and glucosinolate concentration was not statistically significant (Fig 4B, r_G = 0.08, P = 0.42, N = 93), perhaps because the Ler x Col cross has substantially less variation in glucosinolate concentration. Further, no herbivory or aliphatic glucosinolate QTL overlapped in this cross (Fig 5).

View larger version (15K): In this window In a new window Download PPT slide   Figure 5. QTL controlling total aliphatic glucosinolate levels in Ler x Col leaves. LOD plot of QTL controlling aliphatic glucosinolate concentration in 4-week-old plants in the Ler x Col RI population obtained by composite interval mapping. Only chromosome V contained significant QTL. The horizontal line represents the chromosome-wide P = 0.05 LOD score as determined by 500 random permutations of the data. Each QTL is labeled with the marker exhibiting the highest significance and the percentage effect of the Ler allele on glucosinolate accumulation.

QTL regulating resistance to P. xylostella herbivory:
The line means of feeding damage by P. xylostella on Ler x Col and Ler x Cvi RI lines were utilized to identify QTL regulating P. xylostella herbivory. In the Ler x Cvi RI population two QTL were identified. They mapped near erecta on chromosome II and near the amplified fragment length polymorphism marker DF184L on chromosome V (Fig 6; LISTER and DEAN 1993 ; ALONSO-BLANCO et al. 1998B ). For both QTL, the Ler allele imparts increased resistance to P. xylostella herbivory (Fig 6). In the Ler x Col RI lines, we found no significant variation among RI lines for P. xylostella resistance (Table 1) and no significant QTL affecting this trait (not shown).

View larger version (19K): In this window In a new window Download PPT slide   Figure 6. QTL controlling resistance to P. xylostella herbivory. LOD plots of QTL controlling P. xylostella herbivory on 4-week-old plants in the Ler x Cvi RI population obtained by composite interval mapping. Only the two chromosomes with significant QTL are shown. The axes are independently scaled to accommodate chromosome length and maximum LOD scores. The horizontal line represents the genome-wide P = 0.05 LOD score as determined by 500 random permutations of the data. Each QTL is labeled with the marker exhibiting the highest significance and the percentage effect of the Ler allele on resistance. Positive values indicate that plants containing the Ler allele are more resistant to herbivory than are those containing the other allele and vice versa for a negative effect. No significant QTL were identified in the Ler x Col RI population.

Comparison of P. xylostella herbivory and the glucosinolate-myrosinase system:
Comparison of P. xylostella herbivory to leaf aliphatic glucosinolate concentration in the Ler x Cvi population found no significant correlation between these traits (N = 93, r_G = 0.08, P = 0.416). Further, no QTL controlling resistance to P. xylostella cosegregated with QTL regulating any known aspect of the glucosinolate-myrosinase system (Fig 6; MITCHELL-OLDS and PEDERSEN 1998 ; KLIEBENSTEIN et al. 2001A ).

One QTL regulates resistance to both insects:
Resistance to T. ni and P. xylostella herbivory in Ler x Cvi is positively correlated (Fig 7, r_G = 0.23, P = 0.003, N = 160). Comparison of resistance QTL for the two insects indicates that the erecta region influences damage by both herbivores in the Ler x Cvi lines (Fig 2 and Fig 6). ANCOVA showed that correlated genetic patterns of resistance were completely attributable to the QTL located near erecta: (r_G = 0.50, model 2: ERECTA factor: P < 0.001, TNI covariate: P = 0.495, N = 160).

View larger version (30K): In this window In a new window Download PPT slide   Figure 7. Correlation of resistance to P. xylostella and T. ni herbivory. A plot of the mean resistance to P. xylostella and T. ni herbivory in 160 Ler x Cvi RI lines. The oval is the 95th percentile space obtained by least-squares regression.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We found higher levels of genetic variation for resistance to the generalist herbivore, T. ni, than for the component mediating specialist, P. xylostella (Table 1). Furthermore, five QTL regulating T. ni herbivory overlap with QTL known to regulate the glucosinolate-myrosinase system, while P. xylostella feeding did not appear to be influenced by the glucosinolate-myrosinase system (Fig 2 and Fig 6; KLIEBENSTEIN et al. 2001A ). While generalist feeding was influenced by the glucosinolate-myrosinase system, there was variation for the relative impact of glucosinolate concentration and myrosinase activity between the two RI lines.

The AOP and EC198L QTL for glucosinolate levels overlapped with T. ni herbivory QTL in Cvi x Ler (Fig 2). Both loci influence glucosinolate concentration, while AOP also influences glucosinolate type (KLIEBENSTEIN et al. 2001C ). The allelic status at AOP determines the production of either alkenyl or hydroxy aliphatic glucosinolates. The fact that both QTL control T. ni herbivory and glucosinolate concentration suggests that glucosinolate concentration is more important than type in deterring T. ni herbivory (Fig 4). When a third (Elong) locus regulating glucosinolate amount was included in ANOVA, the AOP locus was epistatic to Elong and EC198L for both T. ni feeding and regulation of glucosinolate amount (Table 2; KLIEBENSTEIN et al. 2001A ). Our findings suggest that glucosinolate loci play an important role in deterring feeding by T. ni and other generalist herbivores on Arabidopsis.

We did not find significant insect resistance QTL near AOP and EC198L in the Ler x Col lines (Fig 2), and glucosinolate amount was not significantly correlated with herbivory in this cross (data not shown). The apparent discrepancy between glucosinolate concentration and feeding damage between Ler x Cvi and Ler x Col may be explained by the differences in maximal glucosinolate levels between the two populations. The Ler x Cvi population has maximal aliphatic glucosinolate levels of 20 µmol per gram dry weight (gDWT^-1) while the Ler x Col population reaches only 5 µmol gDWT^-1 (KLIEBENSTEIN et al. 2001A ). Contrasting results between the two crosses suggest that the rate of T. ni herbivory is little affected by glucosinolate levels up to at least 5 µmol gDWT^-1, but that higher concentrations inhibit T. ni herbivory.

While QTL regulating glucosinolate amount were not found to overlap with T. ni resistance QTL in Ler x Col, one resistance QTL overlapped with the NCC1 myrosinase activity QTL (MITCHELL-OLDS and PEDERSEN 1998 ). Further analysis showed that in Ler x Col, increased myrosinase levels have a significant negative genetic correlation with T. ni feeding (Fig 3). This suggests that myrosinase may play a greater role in deterring herbivory when glucosinolate concentrations are relatively low. The results from these crosses indicate that both glucosinolate and myrosinase levels can inhibit T. ni herbivory and should be included in studies that assess the effects of this system on generalist herbivores.

The other two T. ni herbivory QTL in the Ler x Col cross, nga280 and AthChib, have been shown to regulate the type of aliphatic glucosinolate breakdown product produced after tissue damage (LAMBRIX et al. 2001 ). The QTL near nga280 may be identical to the TASTY locus, which was previously identified as a T. ni herbivory QTL in Ler x Col (JANDER et al. 2001 ). In combination, genes near nga280 and AthChib determine the ratio of nitrile to isothiocyanate glucosinolate breakdown products. Comparison of T. ni herbivory and breakdown products showed that isothiocyanates are stronger feeding deterrents than nitriles (LAMBRIX et al. 2001 ). This indicates that glucosinolate production, rate of glucosinolate breakdown, and type of breakdown product all influence T. ni herbivory in Arabidopsis.

High myrosinase levels have previously been shown to be a feeding deterrent for P. xylostella (LI et al. 2000 ). In Ler x Col, myrosinase and herbivory QTL did not overlap, and myrosinase levels did not influence P. xylostella herbivory. This disparity may reflect species-specific or concentration-specific variation in the mode of action of the glucosinolate-myrosinase system. Alternatively, statistical power to detect P. xylostella resistance QTL may differ between these two studies.

QTL mapping indicated one region that regulated resistance to both T. ni and P. xylostella herbivory. This QTL is tightly linked to the erecta locus in the Ler x Cvi RI population but is not found in Ler x Col, where erecta is also segregating. This disparity between RI populations suggests that resistance to these two insect herbivores is not caused by the erecta mutation. It is possible that this region contains a locus that imparts broad-specificity insect resistance or, alternatively, this region may contain two or more loci that independently control resistance to T. ni or P. xylostella herbivory. Fine-scale QTL mapping experiments are required to differentiate between these alternatives.

The experimental determination of the myrosinase levels, glucosinolate levels, and insect herbivory were conducted on independent plants in two different locations over a several year time span. The environments were maintained as similar as possible by utilizing the same soil type, lights, and growth chambers. However, fluctuations in the environment between the experiments could be affecting our results. However, three independent QTL mapping studies of glucosinolate concentration in the Ler x Col RI populations conducted at both sites identified the same glucosinolate concentration QTL (our unpublished results). This suggests that the major effect of the differing environments may be to diminish the correlation between the glucosinolate/myrosinase system and insect herbivory. However, final confirmation of the herbivory QTL as glucosinolate/myrosinase loci remains to be functionally verified.

A. thaliana contains significant natural genetic variation for resistance to insect herbivory. Mapping QTL responsible for resistance in RI lines allows for comparison with published QTL maps and rapid testing of plant-insect interaction models. Combining the genomics tools available in Arabidopsis with the genetic tools described in this article should allow the cloning of uncharacterized insect herbivory QTL. Finally, molecular characterization of these QTL will enhance our understanding of how plants defend themselves from insect herbivory.

	FOOTNOTES

¹ Present address: Department of Vegetable Crops, University of California, Davis, CA 95616.

	ACKNOWLEDGMENTS

We thank S. Dix for expert secretarial assistance. This work was supported by the Max-Planck-Gesellschaft. T.M.-O. was also supported by the U.S. National Science Foundation, grant DEB-9527725.

Manuscript received November 8, 2001; Accepted for publication February 8, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALONSO-BLANCO, C., S. E. EL-ASSAL, G. COUPLAND, and M. KOORNNEEF, 1998a  Analysis of natural allelic variation at flowering time loci in the Landsberg erecta and Cape Verde Islands ecotypes of Arabidopsis thaliana. Genetics 149:749-764[Abstract/Free Full Text].

ALONSO-BLANCO, C., A. J. M. PEETERS, M. KOORNNEEF, C. LISTER, and C. DEAN et al., 1998b  Development of an AFLP based linkage map of Ler, Col and Cvi Arabidopsis thaliana ecotypes and construction of a Ler/Cvi recombinant inbred line population. Plant J. 14:259-271[Medline].

BARTLET, E., D. PARSONS, I. WILLIAMS, and S. CLARK, 1994  The influence of glucosinolates and sugars on feeding by the cabbage stem flea beetle, Psylliodes chrysocephala.. Entomol. Exp. Appl. 73:77-83.

BARTLET, E., M. M. BLIGHT, P. LANE, and I. H. WILLIAMS, 1997  The responses of the cabbage seed weevil Ceutorhynchus assimilis to volatile compounds from oilseed rape in a linear track olfactometer. Entomol. Exp. Appl. 85:257-262.

BASTEN, C. J., B. S. WEIR and Z-B. ZENG, 1999 QTL Cartographer, version 1.13. Department of Statistics, North Carolina State University, Raleigh, NC.

BERENBAUM, M. and J. J. NEAL, 1985  Synergism between myristicin and xanthotoxin, a naturally cooccurring plant toxicant. J. Chem. Ecol. 11:1349-1358.

BERENBAUM, M. and A. ZANGERL, 1992  Genetics of physiological and behavioral resistance to host furanocoumarins in the parsnip webworm. Evolution 46:1373-1384.

BONES, A. M. and J. T. ROSSITER, 1996  The myrosinase-glucosinolate system, its organisation and biochemistry. Physiol. Plant. 97:194-208.

BOWERS, M. D. and G. M. PUTTICK, 1988  Response of generalist and specialist insects to qualitative allelochemical variation. J. Chem. Ecol. 14:319-334.

BRADBURNE, R. P. and R. MITHEN, 2000  Glucosinolate genetics and the attraction of the aphid parasitoid Diaeretiella rapae to Brassica. Proc. R. Soc. Lond. Ser. B Biol. Sci. 267:89-95[Medline].

CHEW, F. S., 1988 Searching for defensive chemistry in the Cruciferae, or, do glucosinolates always control interactions of Cruciferae with their potential herbivores and symbionts? No!, pp. 81–111 in Chemical Mediation of Coevolution, edited by K. A. SPENCER. Academic Press, New York.

CHEW, F. S., and J. A. A. RENWICK, 1994 Host plant choice in Pieris butterflies, pp. 214–238 in Chemical Ecology of Insects II, edited by R. T. C. A. W. J. BELL. Chapman & Hall, New York.

DA COSTA, C. P. and C. M. JONES, 1971  Cucumber beetle resistance and mite susceptibility controlled by the bitter gene in Cucumis sativa L. Science 172:1145-1146[Abstract/Free Full Text].

EHRLICH, P. R. and P. H. RAVEN, 1964  Butterflies and plants: a study in coevolution. Evolution 18:586-608.

FALCONER, D. S., and T. F. C. MACKAY, 1996 Introduction to Quantitative Genetics. Longman, Harlow/Essex, UK.

FEENY, P., 1976 Plant apparency and chemical defense, pp. 1–40 in Biochemical Interaction Between Plants and Insects, edited by I. WALLACE and R. L. MANSELL. Plenum Press, New York.

GIAMOUSTARIS, A. and R. MITHEN, 1995  The effect of modifying the glucosinolate content of leaves of oilseed rape (Brassica napus Ssp oleifera) on its interaction with specialist and generalist pests. Ann. Appl. Biol. 126:347-363.

GRIFFITHS, D. W., N. DEIGHTON, A. N. E. BIRCH, B. PATRIAN, and R. BAUR et al., 2001  Identification of glucosinolates on the leaf surface of plants from the Cruciferae and other closely related species. Phytochemistry 57:693-700[Medline].

HUANG, X. and J. A. A. RENWICK, 1994  Relative activities of glucosinolates as oviposition stimulants for Pieris rapae and P. napi oleracea. J. Chem. Ecol. 20:1025-1037.

JANDER, G., J. CUI, B. NHAN, N. E. PIERCE, and F. M. AUSUBEL, 2001  The TASTY locus on chromosome 1 of Arabidopsis affects feeding of the insect herbivore Trichoplusia ni.. Plant Physiol. 126:890-898[Abstract/Free Full Text].

KLIEBENSTEIN, D. J., J. GERSHENZON, and T. MITCHELL-OLDS, 2001a  Comparative quantitative trait loci mapping of aliphatic, indolic and benzylic glucosinolate production in Arabidopsis thaliana leaves and seeds. Genetics 159:359-370[Abstract/Free Full Text].

KLIEBENSTEIN, D. J., J. KROYMANN, P. BROWN, A. FIGUTH, and D. PEDERSEN et al., 2001b  Genetic control of natural variation in Arabidopsis thaliana glucosinolate accumulation. Plant Physiol. 126:811-825[Abstract/Free Full Text].

KLIEBENSTEIN, D. J., V. M. LAMBRIX, M. REICHELT, J. GERSHENZON, and T. MITCHELL-OLDS, 2001c  Gene duplication and the diversification of secondary metabolism: side chain modification of glucosinolates in Arabidopsis thaliana.. Plant Cell 13:681-693[Abstract/Free Full Text].

LAMBRIX, V., M. REICHELT, T. MITCHELL-OLDS, D. J. KLIEBENSTEIN, and J. GERSHENZON, 2001  The Arabidopsis epithiospecifier protein promotes the hydrolysis of glucosinolates to nitriles and influences Trichoplusia ni herbivory. Plant Cell 13:2793-2807[Abstract/Free Full Text].

LI, Q., S. EIGENBRODE, G. STRINGHAM, and M. THIAGARAJAH, 2000  Feeding and growth of Plutella xylostella and Spodoptera eridania on Brassica juncea with varying glucosinolate concentrations and myrosinase activities. J. Chem. Ecol. 26:2401-2419.

LISTER, C. and C. DEAN, 1993  Recombinant inbred lines for mapping RFLP and phenotypic markers in Arabidopsis thaliana.. Plant J. 4:745-750.

MITCHELL-OLDS, T. and D. PEDERSEN, 1998  The molecular basis of quantitative genetic variation in central and secondary metabolism in Arabidopsis. Genetics 149:739-747[Abstract/Free Full Text].

MITHEN, R., J. CLARKE, C. LISTER, and C. DEAN, 1995  Genetics of aliphatic glucosinolates. III. Side chain structure of aliphatic glucosinolates in Arabidopsis thaliana. Heredity 74:210-215.

MOYES, C. L. and A. F. RAYBOULD, 2001  The role of spatial scale and intraspecific variation in secondary chemistry in host-plant location by Ceutorhynchus assimilis (Coleoptera: Curculionidae). Proc. R. Soc. Lond. Ser. B Biol. Sci. 268:1567-1573[Medline].

PEREZ, C. J. and A. M. SHELTON, 1997  Resistance of Plutella xylostella (Lepidoptera: Plutellidae) to Bacillus thuringiensis Berliner in central America. J. Econ. Entomol. 90:87-93.

PIVNICK, K. A., B. J. JARVIS, and G. P. SLATER, 1994  Identification of olfactory cues used in host-plant finding by diamondback moth, Plutella xylostella (Lepidoptera, Plutellidae). J. Chem. Ecol. 20:1407-1427.

ROJAS, J. C., 1999  Electrophysiological and behavioral responses of the cabbage moth to plant volatiles. J. Chem. Ecol. 25:1867-1883.

SHOREY, H., L. ANDERSON, and H. REYNOLDS, 1962  Effect of chemical and microbial insecticides on several insect pests of lettuce in southern California. J. Econ. Entomol. 55:5.

SIEMENS, D. H. and T. MITCHELL-OLDS, 1996  Glucosinolates and herbivory by specialists (Coleoptera: Chrysomelidae, Lepidoptera: Plutellidae): consequences of concentration and induced resistance. Environ. Entomol. 80:231-237.

STADLER, E., J. A. A. RENWICK, C. D. RADKE, and K. SACHDEVGUPTA, 1995  Tarsal contact chemoreceptor response to glucosinolates and cardenolides mediating oviposition in Pieris rapae.. Physiol. Entomol. 20:175-187.

STOTZ, H. U., B. R. PITTENDRIGH, J. KROYMANN, K. WENIGER, and J. FRITSCHE et al., 2000  Induced plant defense responses against chewing insects. Ethylene signaling reduces resistance of Arabidopsis against Egyptian cotton worm but not diamondback moth. Plant Physiol. 124:1007-1017[Abstract/Free Full Text].

STOWE, K. A., 1998  Realized defense of artificially selected lines of brassica rapa—effects of quantitative genetic variation in foliar glucosinolate concentration. Environ. Entomol. 27:1166-1174.

This article has been cited by other articles:

				H. C. Rowe, B. G. Hansen, B. A. Halkier, and D. J. Kliebenstein Biochemical Networks and Epistasis Shape the Arabidopsis thaliana Metabolome PLANT CELL, May 1, 2008; 20(5): 1199 - 1216. [Abstract] [Full Text] [PDF]

				J. Wu, C. Hettenhausen, M. C. Schuman, and I. T. Baldwin A Comparison of Two Nicotiana attenuata Accessions Reveals Large Differences in Signaling Induced by Oral Secretions of the Specialist Herbivore Manduca sexta Plant Physiology, March 1, 2008; 146(3): 927 - 939. [Abstract] [Full Text] [PDF]

				G. C. Barker, T. R. Larson, I. A. Graham, J. R. Lynn, and G. J. King Novel Insights into Seed Fatty Acid Synthesis and Modification Pathways from Genetic Diversity and Quantitative Trait Loci Analysis of the Brassica C Genome Plant Physiology, August 1, 2007; 144(4): 1827 - 1842. [Abstract] [Full Text] [PDF]

				H. van Leeuwen, D. J. Kliebenstein, M. A.L. West, K. Kim, R. van Poecke, F. Katagiri, R. W. Michelmore, R. W. Doerge, and D. A. St.Clair Natural Variation among Arabidopsis thaliana Accessions for Transcriptome Response to Exogenous Salicylic Acid PLANT CELL, July 1, 2007; 19(7): 2099 - 2110. [Abstract] [Full Text] [PDF]

				A. J. Heidel, M. J. Clauss, J. Kroymann, O. Savolainen, and T. Mitchell-Olds Natural Variation in MAM Within and Between Populations of Arabidopsis lyrata Determines Glucosinolate Phenotype Genetics, July 1, 2006; 173(3): 1629 - 1636. [Abstract] [Full Text] [PDF]

				Z. Zhang, J. A. Ober, and D. J. Kliebenstein The Gene Controlling the Quantitative Trait Locus EPITHIOSPECIFIER MODIFIER1 Alters Glucosinolate Hydrolysis and Insect Resistance in Arabidopsis PLANT CELL, June 1, 2006; 18(6): 1524 - 1536. [Abstract] [Full Text] [PDF]

				D. J. Kliebenstein, M. A. L. West, H. van Leeuwen, K. Kim, R. W. Doerge, R. W. Michelmore, and D. A. St. Clair Genomic Survey of Gene Expression Diversity in Arabidopsis thaliana Genetics, February 1, 2006; 172(2): 1179 - 1189. [Abstract] [Full Text] [PDF]

				H. S. Callahan Using Artificial Selection to Understand Plastic Plant Phenotypes Integr. Comp. Biol., June 1, 2005; 45(3): 475 - 485. [Abstract] [Full Text] [PDF]

				K. K. Shimizu and M. D. Purugganan Evolutionary and Ecological Genomics of Arabidopsis Plant Physiology, June 1, 2005; 138(2): 578 - 584. [Full Text] [PDF]

				P. Reymond, N. Bodenhausen, R. M.P. Van Poecke, V. Krishnamurthy, M. Dicke, and E. E. Farmer A Conserved Transcript Pattern in Response to a Specialist and a Generalist Herbivore PLANT CELL, November 1, 2004; 16(11): 3132 - 3147. [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. Kroymann, S. Donnerhacke, D. Schnabelrauch, and T. Mitchell-Olds Evolutionary dynamics of an Arabidopsis insect resistance quantitative trait locus PNAS, November 25, 2003; 100(suppl_2): 14587 - 14592. [Abstract] [Full Text]

				O. A. Hoekenga, T. J. Vision, J. E. Shaff, A. J. Monforte, G. P. Lee, S. H. Howell, and L. V. Kochian Identification and Characterization of Aluminum Tolerance Loci in Arabidopsis (Landsberg erecta x Columbia) by Quantitative Trait Locus Mapping. A Physiologically Simple But Genetically Complex Trait Plant Physiology, June 1, 2003; 132(2): 936 - 948. [Abstract] [Full Text] [PDF]

				A. Ratzka, H. Vogel, D. J. Kliebenstein, T. Mitchell-Olds, and J. Kroymann Disarming the mustard oil bomb PNAS, August 20, 2002; 99(17): 11223 - 11228. [Abstract] [Full Text] [PDF]

				D. J. Kliebenstein, A. Figuth, and T. Mitchell-Olds Genetic Architecture of Plastic Methyl Jasmonate Responses in Arabidopsis thaliana Genetics, August 1, 2002; 161(4): 1685 - 1696. [Abstract] [Full Text] [PDF]