"Science, Faculty of"@en . "Botany, Department of"@en . "DSpace"@en . "UBCV"@en . "Sackett, Tara Elizabeth"@en . "2009-07-13T22:12:49Z"@en . "2000"@en . "Master of Science - MSc"@en . "University of British Columbia"@en . "Two furoquinoline alkaloids, skimmianine and dictamnine, were purified from the\r\ndried leaves of Skimmia japonica (Rutaceae) and dried root of Dictamnus albus\r\n(Rutaceae), respectively. The furoquinolines were isolated through acid fractionation\r\nand silica gel column chromatography, and their identity was confirmed through HPLC\r\nand mass spectroscopic analysis.\r\nSkimmianine and dictamnine, as well as the furanocoumarins 8-methoxypsoralen\r\n(8-MOP) and 5-methoxypsoralen (5-MOP) were quantified in leaf tissue of S. japonica\r\nand Ruta graveolens (Rutaceae). Of these compounds, S. japonica contained\r\nskimmianine and 5-MOP, and quantities differed depending on the sex of the plant from\r\nwhich the leaves were taken. The compounds were not present on the surface of the\r\nleaves. R. graveolens was found to contain skimmianine, 5-MOP and 8-MOP, both on\r\nthe surface and interior of the leaves. Neither plant contained detectable levels of\r\ndictamnine.\r\nGrowth assays using Spodoptera litura (Lepidoptera: Noctuidae) and\r\nTrichoplusia ni (Lepidoptera: Noctuidae) demonstrated that both furoquinolines are\r\npotent antifeedants at concentrations below those found naturally. Skimmianine is twice\r\nas deterrent to S. litura as dictamnine on a weight to weight basis. 8-MOP was also\r\ndeterrent to S. litura, but not as potent as either furoquinoline. Skimmianine was\r\ndeterrent to T. ni, but 8-MOP had no effect on its growth. Simulated daylight conditions,\r\nwith appropriate UV wavelengths, were found to have no effect on the antifeedant\r\nproperties of either furoquinoline or 8-MOP in the S. litura trials. In addition, none of the\r\ncompounds were physiologically toxic to larvae.\r\nIt was found that a non-ratio-based method was more accurate than a ratiobased\r\nmethod for the analysis and interpretation of experimental data.\r\nIn conclusion, the furoquinolines skimmianine and dictamnine, and the\r\nfuranocoumarin 8-MOP, are deterrent to S. litura larvae. Skimmianine is deterrent to T.\r\nni. The intensity of experimental light conditions may be essential for inducing toxicity in\r\npotentially phototoxic compounds."@en . "https://circle.library.ubc.ca/rest/handle/2429/10752?expand=metadata"@en . "3280753 bytes"@en . "application/pdf"@en . "Furoquinolines of the Rutaceae and their role in Plant-Lepidopteran Interactions by Tara Elizabeth Sackett BSc , Mount Allison University, 1996 A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in The Faculty of Graduate Studies (Department of Botany) We accept this thesis as conforming to the required standard The University of British Columbia September 2000 \u00C2\u00A9 Tara Elizabeth Sackett In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of JBo'W-v The University of British Columbia Vancouver, Canada DE-6 (2/88) A B S T R A C T Two furoquinoline alkaloids, skimmianine and dictamnine, were purified from the dried leaves of Skimmia japonica (Rutaceae) and dried root of Dictamnus albus (Rutaceae), respectively. The furoquinolines were isolated through acid fractionation and silica gel column chromatography, and their identity was confirmed through H P L C and mass spectroscopic analysis. Skimmianine and dictamnine, as well as the furanocoumarins 8-methoxypsoralen (8-MOP) and 5-methoxypsoralen (5-MOP) were quantified in leaf tissue of S. japonica and Ruta graveolens (Rutaceae). Of these compounds, S. japonica contained skimmianine and 5-MOP, and quantities differed depending on the sex of the plant from which the leaves were taken. The compounds were not present on the surface of the leaves. R. graveolens was found to contain skimmianine, 5-MOP and 8-MOP, both on the surface and interior of the leaves. Neither plant contained detectable levels of dictamnine. Growth assays using Spodoptera litura (Lepidoptera: Noctuidae) and Trichoplusia ni (Lepidoptera: Noctuidae) demonstrated that both furoquinolines are potent antifeedants at concentrations below those found naturally. Skimmianine is twice as deterrent to S. litura as dictamnine on a weight to weight basis. 8-MOP was also deterrent to S. litura, but not as potent as either furoquinoline. Skimmianine was deterrent to T. ni, but 8-MOP had no effect on its growth. Simulated daylight conditions, with appropriate UV wavelengths, were found to have no effect on the antifeedant properties of either furoquinoline or 8-MOP in the S. litura trials. In addition, none of the compounds were physiologically toxic to larvae. It was found that a non-ratio-based method was more accurate than a ratio-based method for the analysis and interpretation of experimental data. In conclusion, the furoquinolines skimmianine and dictamnine, and the furanocoumarin 8-MOP, are deterrent to S. litura larvae. Skimmianine is deterrent to T. ni. The intensity of experimental light conditions may be essential for inducing toxicity in potentially phototoxic compounds. 11 T A B L E OF CONTENTS Abstract ii List of Tables iv List of Figures v List of Appendices vi Acknowledgements vii Chapter 1: General Introduction 1 Chapter 2: Isolation of Furoquinolines and Quantification of Furanocoumarins and Furoquinolines in Rutaceous Plants Introduction 12 Materials and Methods A. Purification of Skimmianine and Dictamnine 13 B. Quantification of Furanocoumarins and Furoquinolines 16 Results and Discussion A. Purification of Skimmianine and Dictamnine 17 B. Quantification of Furanocoumarins and Furoquinolines 19 Chapter 3: Effects of Furoquinolines on the Growth and Feeding of Spodoptera litura (Lepidoptera: Noctuidae) Introduction 26 Materials and Methods 29 Results and Discussion A. Analysis of Final weights and Consumption of Larvae on Various Diets 33 B. A comparison of methods: ratio-based versus non-ratio-based 46 Chapter 4: Test for Synergistic Effects Between Furanocoumarins and Furoquinolines on Trichoplusia ni (Lepidoptera: Noctuidae) Introduction 52 Materials and Methods 53 Results and Discussion 55 Summary and Conclusions 59 References 63 Appendices 68 in LIST OF T A B L E S Table 2.1: Proportion of skimmianine and dictamnine isolated from dry plant material. 18 Table 2.2: Comparison of compound concentrations (Mean \u00C2\u00B1 S.E.) of Skimmia japonica leaf samples, differing in age and sex. 21 Table 2.3: Comparison of compound concentrations in male and female Skimmia japonica plants. 21 Table 2.4: Concentrations (Mean \u00C2\u00B1 S.E.) of various compounds in Ruta graveolens. 24 Table 3.1: Mean final body weights and consumption (mg dry weight) \u00C2\u00B1 95% confidence intervals of S. litura larvae fed skimmianine diets in light. 34 Table 3.2: Mean final body weights and consumption (mg dry weight) \u00C2\u00B1 95% confidence intervals of S. litura larvae fed skimmianine diets in dark. 35 Table 3.3: Mean final body weights and consumption (mg dry weight) \u00C2\u00B1 95% confidence intervals of S. litura larvae fed dictamnine diets in light. 36 Table 3.4: Mean final body weights and consumption (mg dry weight) \u00C2\u00B1 95% confidence intervals of S. litura larvae fed dictamnine diets in dark. 37 Table 3.5: Mean final body weights and consumption (mg dry weight) \u00C2\u00B1 95% confidence intervals of S. litura larvae fed 8-MOP diets in light. 38 Table 3.6: Mean final body weights and consumption (mg dry weight) \u00C2\u00B1 95% confidence intervals of S. litura larvae fed 8-MOP diets in dark. 39 Table 3.7: R G R i and RCRi of Spodoptera litura on dictamnine diets in light. 49 IV LIST OF FIGURES Figure 1.1: Structure of furoquinolines and furanocoumarins. 3 Figure 1.2: Comparative biosynthesis of furanocoumarins and furoquinolines. 4 Figure 1.3: Detoxification products from the metabolism of 8-MOP in Papilio polyxenes. 7 Figure 1.4: The structure of myristicin. 10 Figure 2.1: H P L C elution profiles of C H 2 C I 2 crude extracts of S. japonica leaves from A female B male plants. 20 Figure 2.2: H P L C elution profile of Ruta graveolens C H 2 C I 2 crude extract. 23 Figure 3.1: Weight gain versus initial weight of S. litura larvae on treatment diets containing different concentrations of dictamnine. 47 Figure 3.2: Consumption versus initial weight of S. litura on treatment diets containing different concentrations of dictamnine. 48 Figure 4.1: Mean final weights (\u00C2\u00B1 S.E.) of Trichoplusia ni larvae after feeding for 10 days on control or treatment diets. 55 Figure A1.1: MS Spectra obtained for purified compounds from Skimmia japonica and Dictamnus albus 68 Figure A2.1: Spectra of some furoquinolines and furanocoumarins 69 Figure A3.1: Chart of initial weight versus weight gain of S. litura larvae on diets containing dictamnine. 70 Figure A3.2: The distribution of the final weights of the control treatment before and after adjustment due to initial weight covariance. 71 V LIST OF APPENDICES Appendix I: MS spectra of skimmianine and dictamnine Appendix II: Spectra of some furoquinolines and furanocoumarins, with labelled peak maxima Appendix III: Procedure for A N C O V A analysis VI ACKNOWLEDGEMENTS I would like to acknowledge my supervisor, Dr. Neil Towers, for his encouragement, support and constantly creative ideas. I would also like to thank Dr. Murray Isman, whose guidance and laboratory facilities enabled me to do the insect experiments, and my other committee members, past and present, Dr. lain Taylor and Dr. Gary Bradfield. Two years of my research was funded by an N S E R C P G S A scholarship, and I gratefully acknowledge them for their financial support. Many thanks to Nancy Brard and the other members of the Isman lab, and to Dr. Bertrand Clarke for statistical advice. Essential to my time at U B C were the members of the Towers lab, for their academic advice and companionship. Thanks especially to Zyta Abramowski for all her help, Dr. Malladi for alkaloid isolating advice, Kevin Usher for teaching me to use the HPLC, Fiona Cochrane for her editing, and Rene Orozco for his lessons in practical chemistry. V l l C H A P T E R 1 G e n e r a l I n t r o d u c t i o n Plant chemistry is an integral part of understanding the ecological interactions of plants and insect herbivores. Studying the chemistry of a family of plants, along with the feeding preferences and growth of insect herbivores, can reveal ecological and evolutionary patterns. Studies have identified phytochemicals that are unpalatable and/or toxic for some insect species, suggesting a defensive role for these chemicals (Jacobsen, 1989). Also important to recognize are the responses to plant toxins by different species of herbivores, indicating patterns of specialization and adaptation (Berenbaum, 1983; Stadler, 1992). One of the most chemically diverse plant families is the Rutaceae, or Citrus family. This project focuses on a group of alkaloids unique to the Rutaceae, the furoquinolines, and ascertains the effect of these compounds on generalist lepidopteran larvae. The Rutaceae and their Lepidopteran Herbivores Plants in the Rutaceae family are found in tropical and warm temperate climates throughout the world. The range includes South America, the United States, Central America, Africa, the Mediterranean, Middle East, Australia and New Zealand. There are approximately 155 genera and 1600 species (Chase et al., 1999). In North America, Rutaceous species include various Citrus spp., Zanthoxylum americanum (northern prickly ash), Z. fagara (lime prickly ash), Ptelea trifoliata (hoptree), Amyris elemifera (torchwood) and Choisya dumosa (Mexican orange). Most polyphagous, or generalist, lepidopteran herbivores do not include Rutaceous plants in their diet, although there are exceptions (Tietz, 1972). More l typically, lepidopterans that feed on Rutaceous plants are specialists on this plant family. Species in the genus Papilio have a majority of Rutaceous host plants, although plants from other families, particularly Apiaceae and Lauraceae, can also be host plants (Miller, 1987). Chemistry of the Rutaceae The chemistry of the Rutaceae is diverse and includes a wide variety of compounds. Well known from the Rutaceae are the terpenoid essential oils and volatiles, but also typical of this family are coumarins, flavonoids, tannins, amides, and alkaloids (Hegnauer, 1983). Chemical distributions within the Rutaceae have been particularly important for the study of the systematics of the family (Da Silva et al., 1988). Recent molecular systematic work has found that the chemotaxonomic groupings are more reliable than fruit-type for family and sub-family classifications to the Rutaceae (Chase etal., 1999). From the wide variety of chemicals in the Rutaceae, a few insecticidal and insect antifeedant compounds have already been found. Citrus oils, containing mostly terpenoid compounds, are toxic and deterrent to two species of weevil (Coleoptera) (Jacobsen, 1989; Don-Pedro, 1996). Other terpenoids from this family, such as limonoids, were demonstrated to be deterrent and/or toxic to insects, including various lepidopterans (Jacobsen, 1989). Furanocoumarins, chemicals also found in Apiaceous plants, are widespread in the Rutaceae, and are deterrent and toxic to several species of generalist lepidopteran larvae (Berenbaum and Neal, 1985; Klocke etal., 1989; Berenbaum etal., 1991; Nawrot etal., 1991; Brewer etal., 1995; Berdegue etal, 1997). Several types of alkaloids and amide compounds from Rutaceous plants have also been shown to affect insect growth or feeding. These include zanthophylline, a 2 pyranoquinoline alkaloid, benz(C)-phenanthridine, acridone and indoloquinazoline alkaloids, and isobutylamides (Jacobsen, 1989; Escoubas et al., 1994). Furanocoumarins and Furoquinolines: Structure and Synthesis Furoquinoline alkaloids are the most widespread type of alkaloid in the Rutaceae and also particularly numerous in structural variety (Mester, 1983). They are structurally similar to furanocoumarins; both are planar aromatic compounds with three rings, including a five-carbon furan ring (Figure 1.1). Furanocoumarins Furoquinolines OCH 3 5-methoxypsoralen (5-MOP) skimmianine Figure 1.1: Structure of some furoquinolines and furanocoumarins Although furanocoumarins and furoquinolines share structural similarities, their biosynthetic pathways diverge early on in the synthesis of both types of compounds. Both compounds share the precursor chorismic acid from the shikimic acid pathway 3 (Figure 1.2). Furoquinolines are synthesized from anthranilic acid, which is formed from chorismate and L-glutamine. Anthranilic acid is acetylated and cyclized to form a quinoline, which is subsequently prenylated to form the basic furoquinoline, dictamnine (Cordell, 1981). Furanocoumarins, however, are synthesized from p-coumaryl CoA, an intermediate of the phenylpropanoid pathway. It is synthesized from phenylalanine, which is formed from chorismate. p-Coumaryl CoA is cyclized to form umbelliferone, which is prenylated to form the basic furanocoumarin, psoralen (Stanjek etal., 1999). COOH umbelliferone 2,4-hydroxyquinoline furanocoumarins furoquinolines Figure 1.2: Comparative biosynthesis of furanocoumarins and furoquinolines 4 Light-dependent and independent toxicity of furanocoumarins and furoquinolines: Furanocoumarins and furoquinolines share a mechanism of toxicity, the light-dependent binding to the nucleic acids of DNA. For furanocoumarins, the double bonds of the furan ring and the pyrone ring are both capable of forming intercalations, and upon exposure to UV light, covalent adducts with the pyrimidine bases of DNA. Linear furanocoumarins can thereby form diadducts, cross-linking the strands of DNA with minimal distortion of the double helix. This is thought to block DNA polymerase activity (Murray, 1982). Photobinding of furanocoumarins to enzymes (Veronese et al., 1982), and to unsaturated fatty acids and lecithens (Zare_bska et al., 1998) has also been demonstrated. In additon, activated furanocoumarins can react with oxygen to produce physiologically harmful superoxide and singlet oxygen radicals (Lee and Berenbaum, 1992). Furoquinolines have been shown to form monoadducts with DNA. They also react through the double bond of the furan ring, analogous to the reaction with the double bond in the furan ring of furanocoumarins (Pfyffer et al., 1982a). It is unlikely that the lateral aromatic ring of furoquinolines is photoreactive. 2,3-dihydrodictamnine, which is identical to dictamnine except it has no double bond in the furan ring, is not photoactive (Pfyffer etal., 1982a). The photobinding of dictamnine to DNA is to a degree four times less than the photobinding of the furanocoumarin 8-methoxypsoralen (8-MOP) (Pfyffer et al., 1982a). Dictamnine has also been shown to bind to proteins, but at a capacity twenty times less than photobinding to DNA (Pfyffer and Towers, 1982). The toxicity of furoquinolines has been demonstrated with a variety of microorganisms and cell types. Dictamnine was found to be phototoxic to the bacteria Escherichia coli and Saccharomyces cerevisiae. Skimmianine, however, was not phototoxic (Towers et al., 1981). The photobinding of dictamnine has been 5 demonstrated in vitro with calf thymus DNA (Pfyffer et al., 1982a) and in vitro and in vivo with fungal DNA (Pfyffer and Towers, 1982). The binding to fungal DNA was four times less in vivo than in vitro, possibly due to repair enzymes. Dictamnine, y-fagarine and skimmianine cause frameshift mutations in Salmonella typhimurium without UV light, due to light-independent intercalation with DNA (Mizuta and Kanamori, 1985; Kanamori etal., 1986). Although dictamnine is less phototoxic than 8-MOP, dictamnine was more mutagenic than 8-MOP to E. coli (Ashwood-Smith etal., 1982). In more recent studies, other non-UV dependent actions of furanocoumarins and furoquinolines have been examined. Furanocoumarins have been shown to inhibit insect cytochrome P450 enzymes (Neal and Wu, 1994), and furoquinolines have demonstrated inhibition of cytochrome P450 activity in rats (Goloubkova et al., 1998). The effects of furoquinolines on insects are to this point undocumented. However, there has been extensive research on the effects of furanocoumarins on lepidopteran larvae. Furanocoumarins: Toxicity to Lepidoptera: The effect of dietary furanocoumarins on polyphagous lepidopteran larvae that do not normally encounter furanocoumarins in their host plants has been studied for a variety of species. Furanocoumarins have been shown to deter feeding and to be toxic to Helicoverpa virescens (Klocke etal., 1989), H. zea (Berenbaum and Neal, 1985; Berenbaum et al., 1991), Peridroma saucia , Mamestra configurata (Nawrot et al., 1991), Spodoptera exigua (Diawara etal., 1993; Brewer etal., 1995; Berdegue etal., 1997) and Trichoplusia ni (Zangerl, 1990). The effects of furanocoumarins have also been examined with Papilio polyxenes, which feeds on Rutaceous and Apiaceous plants. 8-MOP has actually been found to 6 stimulate feeding in P. polyxenes, without detrimental effects on growth (Berenbaum, 1981). Papilio polyxenes, and other larvae that feed on plants containing furanocoumarins, appear to have a heightened capacity to detoxify these potentially harmful chemicals. Ivie et al. (1983) found that P. polyxenes metabolized the furanocoumarin 8-MOP to yield either of two metabolites (Figure 1.3). The initial oxidation of the reactive double bond of the furan ring is catalyzed by cytochrome P450 monooxidases located in the midgut and possibly other body tissues of the larvae (Bull et al., 1986). P. polyxenes has multiple isozymes of cytochrome P450, which can be induced to varying degrees by dietary furanocoumarins (Cohen et al., 1989). Larvae of Spodoptera frugiperda, a generalist feeder, were found to detoxify furanocoumarins via the same mechanism as P. polyxenes, but much less efficiently (Bull et al., 1986). Figure 1.3: Detoxification productions from the metabolism of 8-MOP in Papilio polyxenes (Ivie etal., 1983). 7 The capability of furanocoumarins to produce reactive oxygen compounds is also a toxicity risk. Insect larvae do have various antioxidant enzymes, including superoxide dismutase, catalase and peroxidases, that combat these reactants (Brattsten, 1982). However, the efficiency of enzyme activities will vary depending on the insect species. It has been found the P. polyxenes has higher antioxidant enzyme activities in the midgut and fat body tissues than Papilio glaucus, a species that is more polyphagous, and includes very few Rutaceous or Apiaceous plants in its diet (Lee and Berenbaum, 1992). Although furanocoumarins have been shown to be toxic in the diet of non-specialist larvae unaccustomed to the chemicals, the mechanisms of furanocoumarin toxicity have not yet been clarified. Phototoxicity may not have a role in some cases, as furanocoumarins have been shown to also have light-independent toxic effects on lepidopteran larvae (Berenbaum and Neal, 1985). The similarity in structure of furoquinolines to furanocoumarins suggests that the furoquinolines could have similar toxic effects on lepidopteran larvae. Synergy What is the advantage provided to a plant from the production of a multiplicity of compounds with potentially very similar toxic mechanisms? Specifically, why do Rutaceous plants produce both furanocoumarins and furoquinolines? Are there advantages conferred by this diversity of compounds that outweighs any metabolic cost of the production and sequestration of these compounds? There are many reasons why a diversity of compounds may be of benefit to a plant. Secondary metabolites serve a variety of purposes, such as defense against microorganisms, and responses to environmental stresses such as UV radiation, drought, and soil conditions (Harborne, 1982). In addition, having a high number and variety of secondary metabolites provides 8 a pool of chemicals from which a suitable chemical could be found should a new evolutionary selection force arise (Jones and Firn, 1991). Finally, there is also the possibility of synergistic toxicity of compounds with each other, that is, the net toxicity of a combination of chemicals is higher than the sum amount of their individual toxicities. The structural similarity of furoquinolines to the well-studied furanocoumarins could point toward possible synergistic interactions between these two types of chemicals as dietary toxins. The toxicity of the linear furanocoumarin 8-MOP has been found to be synergized by the presence of angular furanocoumarins (Berenbaum and Zangerl, 1996), and other linear furanocoumarins (Brewer etal., 1995). When P. polyxenes larvae were fed diets containing linear and angular furanocoumarins, compared to an equal concentration of linear furanocoumarins alone, there was a greater detrimental effect on growth rate with the mixture of furanocoumarins (Berenbaum and Feeny, 1981). A mixture of the linear furanocoumarins 8-MOP, 5-MOP and psoralen was more detrimental to Spodoptera exigua larvae than the individual furanocoumarins (Brewer et al., 1995). The toxicity of 8-MOP has also been synergized by the presence of a methylenedioxyphenyl compound, myristicin (Figure 1.4) (Berenbaum and Neal, 1985). Myristicin, which is widespread in the Apiaceae family, has no toxic or antifeedant effects on H. zea larvae when in the diet alone. Myristicin did synergize the toxicity of 8-MOP, however. A given concentration of 8-MOP achieved a larger mortality rate of larvae when combined with myristicin than when alone (Berenbaum and Neal, 1985). 9 OCH 3 Figure 1.4: The structure of myristicin There are several potential explanations for the observed synergy in these experiments, which revolve around the structural similarities of the compounds involved. Different furanocoumarins are enzymatically metabolized with different efficiencies depending on the larval species. For example, the metabolism by P. polyxenes of angular furanocoumarins is less efficient than that of linear furanocoumarins (Ivie et al., 1986). Therefore, competition for the active site of the cytochrome P450 would cause a total decrease in metabolic rate, as angular furanocoumarins would occupy the active site for a longer time. Another possible mechanism to explain synergy is that a compound may bind irreversibly to the active site of the cytochrome P450, and render the enzyme inactive and unable to further metabolize other compounds present (Berenbaum and Zangerl, 1996). Furoquinolines may have toxic effects to lepidopteran larvae themselves, or they may act as structural analogues with furanocoumarins to cause a synergistic increase in net toxicity. Project objectives There were three main objectives to this research project. The first objective was to isolate and purify two different furoquinoline compounds from Rutaceous plants. Dictamnine and skimmianine, two of the most common furoquinolines, were selected for 10 isolation. Since these compounds were used in insect feeding trials, a relatively large quantity (approximately 100 mg) of each furoquinoline was required. The second objective was to characterize the concentrations of the furoquinolines in Rutaceous plants. The amounts of the furanocoumarins 8-MOP and 5-MOP were also quantified in these Rutaceous plants. This allowed for appropriate dietary concentrations to be known during the design of insect growth trials. The final objective was to determine the effect of dietary furoquinolines on the growth and feeding of a generalist lepidopteran, Spodoptera litura (Noctuidae). A discussion of growth trial analytical methodology is included, as the common method of analysis, ratio-based analysis, was found to be unsuitable. Feeding experiments were also designed to test for synergy between furanocoumarins and furoquinolines in the diet of larvae. Unfortunately, during the course of the project, the S. litura colony was destroyed by a virus. The synergy experiment was therefore performed with the generalist lepidopteran Trichoplusia ni (Noctuidae). l l C H A P T E R 2 I s o l a t i o n o f F u r o q u i n o l i n e s a n d Q u a n t i f i c a t i o n o f F u r o q u i n o l i n e s a n d F u r a n o c o u m a r i n s in R u t a c e o u s P l a n t s INTRODUCTION Plants in the family Rutaceae have an abundance of secondary metabolites. Although some Rutaceous compounds have been demonstrated to have antifeedant or toxic properties against insects (Jacobsen, 1989), the vast majority of chemicals remain unexplored as to potential physiological or ecological roles. This project seeks to demonstrate whether or not furoquinoline alkaloids have properties as insect antifeedants or toxins. As necessary prerequisites to insect feeding experiments, the pertinent furoquinolines required isolation and purification in a sufficient quantity for insect trials. Hence purification methods were developed to isolate the common furoquinolines, dictamnine and skimmianine. Isolation of these furoquinolines allowed for quantification of the compounds in two Rutaceous plants, Skimmia japonica and Ruta graveolens. Two common furanocoumarins, 8-methoxypsoralen (8-MOP) and 5-methoxypsoraIen (5-MOP) were also quantified in these plants. Since S. japonica is a dioecious species, the concentrations of the compounds were measured separately for male and female leaves. The S. japonica plants were also several years old, so comparisons of chemical concentrations were also performed for leaves of different ages. 1 2 M A T E R I A L S A N D M E T H O D S (A) Purification of skimmianine and dictamnine Plant Material: Skimmianine Leaves from both male and female Skimmia japonica plants were collected from various sites on the campus of the University of British Columbia, Vancouver, in August 1999. Plant material was combined, dried in an oven at 30\u00C2\u00B0 C, and ground using a laboratory mill with a 2mm mesh. Plant Material: Dictamnine The root of Dictamnus albus, available dry as a Chinese medicine (Names: Dictamni radix, Bai Xian Pi (Chinese)), was purchased in Chinatown, Vancouver. The roots were ground in a laboratory mill with a 2 mm mesh. Standards: Pure standards of dictamnine and skimmianine were kindly provided by Dr. Ian D. Spenser, Chemistry department, McMaster University. Solvent Extractions: Skimmianine and Dictamnine The dried, ground plant materials were soaked (separately) in dichloromethane (CH 2CI 2) overnight in a volume of 3 mL solvent per gram dry weight plant material. The extracts were filtered through cheesecloth and Whatman qualitative (1) filter paper. Washing and filtering was repeated a total of three times. The concentrations of skimmianine or dictamnine in the plant materials were determined using these C H 2 C I 2 extracts as further purification resulted in some loss of the compounds. 13 Isolation of Alkaloids: Skimmianine and Dictamnine Alkaloids were isolated from the C H 2 C I 2 extracts through the formation of alkaloid salts (Houghton and Raman, 1998). In this procedure, each C H 2 C I 2 extract was acidified with 1 M HCI in a separatory funnel, to protonate the nitrogen atom of the alkaloids, which causes them to become soluble in the aqueous fraction. This acid extraction was repeated four times, and each aqueous fraction was retained. The pH of the aqueous fraction was raised to 9-10 using concentrated N H 4 O H , thereby de-protonating the alkaloids, which were subsequently extracted into CH 2 CI 2 . Column chromatography: skimmianine The S. japonica alkaloid fraction was further purified using two sequential silica gel columns (70-230 mesh). The extract was applied to the column dry mixed with silica powder. The typical silica gel to extract ratio was 100:1 in dry weight. The first column used an isocratic solvent system of Hexane: C H 2 C I 2 : Methanol at a ratio of 20:12:1. Fractions containing skimmianine were pooled, then concentrated and dried before being applied to a second column. The second column used an isocratic solvent system with Hexane: Ethyl acetate at a ratio of 12:7. By shining a long wave UV lamp on the column the progress of skimmianine through the column was roughly monitored; skimmianine fluoresces blue when exposed to UV light. Pure skimmianine was eluted from the second column. Column Chromatography: dictamnine The D. albus alkaloid fraction was further purified using one silica gel column (70-230 mesh). As in the skimmianine purification, the extract was applied to the column dry mixed with silica powder, and the typical silica gel to extract ratio was 100:1 by dry 14 weight. The column used an isocratic solvent system of Hexane: Ethyl acetate, at a ratio of 4:1. Pure dictamnine was eluted from this column. Confirmation of Purity and Identity of Isolated Skimmianine and Dictamnine Comparison of high performance liquid chromatography (HPLC) elution profiles and spectra (Appendix I) with those of standards confirmed the purity and identity of the isolated compounds as skimmianine and dictamnine. Mass spectroscopy (MS) of the purified compounds and comparison of the MS spectra (Appendix I) with published literature spectra also confirmed their identity (Clugston and MacLean, 1965). Thin Layer Chromatography (TLC) of skimmianine and dictamnine: The presence of skimmianine or dictamnine in an extract or fraction was determined using thin layer chromatography (TLC). Aluminium-backed silica gel 60 F 2 5 4 TLC plates were used. The TLC solvent for separating skimmianine was Hexane: CH 2 CI 2 : Methanol in the ratio of 5:6:1. For dictamnine, the solvent system used was Hexane: CH 2 CI 2 : Methanol in the ratio of 10:10:1. The running distances of compounds were compared by running standards on the same TLC, and detected by the autofluorescence of skimmianine (blue) or dictamnine (yellow) with exposure to long wave UV light. High Performance Liquid Chromatography (HPLC) Analysis: To assess the purity of skimmianine or dictamnine in a fraction, or to quantify the compounds in an extract, H P L C analysis was used. A Waters H P L C with a photodiode array detector was used, allowing the spectra of eluting compounds to be monitored from 200 to 600 nm. A Waters 8 x 100 mm reverse phase 4 u. C 1 8 column in a radial compression module was used. A solvent system of water (H 2 0): acetonitrile (MeCN): 15 methanol (MeOH), either in the ratios of 60:23:18 or 59:21:20, at a flow rate of 2 mL per minute, was used to separate the compounds. The solvent system used depended upon the distribution of the chemical peaks eluting from the extract. These procedures allowed for separation and detection of the furanocoumarins 8-MOP and 5-MOP, as well as the furoquinolines, in the extracts. B. Quantification of furoquinolines and furanocoumarins in Rutaceous plants Plant Material: Ruta graveolens Ruta graveolens L. plants were reared in a growth chamber at 25\u00C2\u00B0C with a 16:8 hour lightdark schedule for three months. The top seven leaflets of nine individual plants were collected for compound analysis. Plant Material: Skimmia japonica One year-old and two year-old leaves from three male and two female Skimmia japonica T. plants were collected in March 2000 from plants growing on a plot on the UBC campus. Standards: The purified furoquinolines from part A were used as standards, and the furanocoumarins 8-MOP and 5-MOP were purchased from Sigma. Whole tissue extraction and quantification: C H 2 C I 2 extractions of the dried, ground plant samples were made and the amount of skimmianine, dictamnine, 8-MOP and 5-MOP in each plant sample was 16 determined using the H P L C procedures outlined in section A. Mean values of compound concentrations within a sampling group were found and compared using analysis of variance (ANOVA) with multiple comparison Tukey tests. Surface extraction: To determine whether or not the compounds were present on the surface of the fresh plant material, fresh leaves from Ruta graveolens and Skimmia japonica were surface extracted by dipping in CH 2 CI 2 fo r 10 seconds. The remaining tissue was ground in liquid nitrogen, and extracted with CH 2 CI 2 overnight. The surface extraction and tissue extraction were analysed by H P L C to qualitatively determine if compounds were present both on the surface and the inside of the leaves. R E S U L T S A N D DISCUSSION A. Purification of skimmianine and dictamnine Several bulk isolations of skimmianine and dictamnine were performed during the development of suitable purification methods. The purified skimmianine or dictamnine was pooled from each isolation. The purity and identity of the compounds were first verified by comparison of HPLC elution profiles and spectra with standards. Mass spectroscopic analysis also confirmed these results; the MS spectra (Appendix 1) were compared with previously published spectra (Clugston and MacLean, 1965) and were identified as skimmianine and dictamnine. Table 2.1 presents the percent of pure compounds isolated from the dry plant materials using the final described isolation procedures. 17 Table 2.1: Proportion of skimmianine and dictamnine isolated from dry plant material. Plant Source compound isolated and purified original cone, of compound. (ug/g DW*) u.g/g DW of compound purified % of total compound isolated Skimmia japonica (leaves) Dictamnus albus (root) skimmianine dictamnine 4240 270 530 110 13 40 *DW = dry weight The skimmianine isolation method recovered 13% of the original skimmianine present in the initial Skimmia japonica CH 2 CI 2 extract. The widespread occurrence of Skimmia japonica as a garden plant, as well as the high concentration of skimmianine in the leaves, 4240 u.g/g dry weight, make this plant an excellent source for bulk isolation of skimmianine. The purification method for dictamnine from the dried roots of Dictamnus albus resulted in a 40% yield of the original amount of dictamnine present in the initial dichloromethane extract of the plant material (Table 2.1). The use of the dried root of Dictamnus albus, available as a Chinese medicine, as a source of dictamnine has been reported previously (Mizuta and Kanamori, 1985). However, the concentration of dictamnine reported to be in the root differs dramatically from the amount detected in the root used in this experiment. Kanamori et al. (1986) report that the concentration of dictamnine found in dried root samples ranges from 3.2 to 9.0 mg per gram dry weight. This is an amount that is 12-33 times higher than the concentration of dictamnine in the plant material used in this study, where a pooled sample of 1 kg dried root had a concentration of 270 u.g per gram dry weight. Due to the potential variability in sources 18 of the medicinal root, and perhaps even the subspecies of Dictamnus albus that are used, this finding is not unreasonable. The purification method used by Mizuta and Kanamori (1985) yielded 2-6.25% of the initial amount of dictamnine estimated to be in their root source. The procedure developed in this project had a more successful yield of 40% the original dictamnine concentration, possibly due to the fewer purification steps required. B. Quantification of furoquinolines and furanocoumarins in Rutaceous plants Skimmia japonica: Differences in compound concentrations due to leafage and sex: Only skimmianine and 5-MOP were detected in leaf samples of S. japonica. Neither dictamnine nor 8-MOP were detected in any of the samples, although they have both been reported to occur in S. japonica (Mester, 1983; Escoubas et al., 1992). Eluting near to the skimmianine and 5-MOP there were two compounds with spectra very similar to spectra of known standard furanocoumarins (Spectra in Appendix 2). The unknowns were called FCIand FC2, referring to their order of elution (Figure 2.1). At least four other furanocoumarins, aside from 5-MOP and 8-MOP, have been reported in Skimmia japonica: isoimperatorin, oxypeucedanin, prangol, and oxypeucedanin methanolate (Gray, 1983). It is possible that one or more of these peaks may include some of these compounds. Relative quantification of these unknowns was possible by calculating the relative peak areas, per gram fresh weight, of these unknown compounds. The elution profiles were different for male and female leaf samples. 19 E c i n O J 03 < 0.25-0.20-0.15J 0.10-J 0.05J 0.00-Solvent: H 2 0 : M e C N : M e O H p 3 : 2 3 : 1 8 ) skimmianine \ 5 - M O P FC1 J F C 2 I i i i r 0.00 10.00 20.00 30.00 Minutes i\u00E2\u0080\u0094i\u00E2\u0080\u0094I\u00E2\u0080\u0094r-^ 40.00 LD CM CO .a < 0.25-0.20J 0.15-0.10-0.05-0.00 0. B Solvent: H 2 0 : M e C N : M e O H unknown 5 - M O P ? skimmianine 60 F C 2 FC1 + I \u00E2\u0080\u0094 i \u00E2\u0080\u0094 i \u00E2\u0080\u0094 r -l O i O O 20I00 Hinutes 63:23:18) III T~^~I\u00E2\u0080\u0094r\u00E2\u0080\u0094i\u00E2\u0080\u0094i\u00E2\u0080\u0094i\u00E2\u0080\u0094r 30.00 40.00 Figure 2.1: HPLC elution profiles of CH 2 CI 2 crude extracts of S. japonica leaves from A female and B male plants. 20 The concentrations of skimmianine and 5-MOP, and the relative concentrations of FC1 and FC2, in categories of leaf age and sex, are presented in Table 2.2. Table 2.2: Comparison of compound concentrations (Mean \u00C2\u00B1 S.E.) of Skimmia japonica leaf samples, differing in age and sex. ng/g F W A relative concentrations leaf sample n skimmianine 5-MOP FC1 FC2 male-young 3 670 + 100 a* 0 1.2 \u00C2\u00B1 0.13 a 2.5 \u00C2\u00B1 0.32 a male-old 3 5 1 0 \u00C2\u00B1 8 0 a 0 1.0 + 0.11 a 1 . 9 \u00C2\u00B1 0 . 1 5 a b female-young 2 1 5 0 0 \u00C2\u00B1 3 0 b 600 \u00C2\u00B1 70 b 1.1 + 0 .09 a 1.1 \u00C2\u00B1 0 . 0 2 b c female-old 2 2050 \u00C2\u00B1 280 b 480 \u00C2\u00B1 80 b 1.3 \u00C2\u00B1 0.03 a 1.0 \u00C2\u00B1 0.08 c A F W = f resh weight * Da ta were ana l yzed by A N O V A with T u k e y test. M e a n s with different superscr ip ts are signi f icant ly different f rom other m e a n s within the s a m e c o l u m n , (p < .05). Data analysis indicated that there were no significant differences between compound concentrations within each sex due to the age of the leaves. However, there were significant differences in compound concentrations depending upon the sex of the plant. Therefore, data from samples of the same plant, young and old leaves, were pooled. The pooled mean values for compound concentrations were then compared for male versus female plants (Table 2.3). Table 2.3: Comparison of compound concentrations in male and female Skimmia japonica plants. ug/g FW relative concentrations leaf sample n skimmianine 5-MOP FC1 FC2 male 3 600 \u00C2\u00B1 70 a* 0 a 0.9 \u00C2\u00B1 0.08 3 2.1 \u00C2\u00B1 0.2 3 female 2 1800 \u00C2\u00B1 90 b 540 +60 b 1.0 +0.05 a 1.0 \u00C2\u00B1 0.04 b A F W = f resh weight * Da ta were ana l yzed by A N O V A with T u k e y test. M e a n s with different superscr ip ts are signi f icant ly different f rom other m e a n s within the s a m e c o l u m n , (p < .05). 21 Female plants had a skimmianine concentration that was three times higher than male plants: 1800 ug/g FW (fresh weight) versus 600 u,g/g FW. In addition, female plants were found to contain 5-MOP at a concentration of 540 ug/g FW, while the male plants contained no 5-MOP. The relative concentrations of FC1 were not significantly different between male and female plants, however male plants had twice as much FC2 than female plants. The different quantities and composition of the furanocoumarins and furoquinolines in male and female plants is an intriguing observation. No conclusions can be confidently drawn from these observations, however, due to the source of the plant material. The leaves were taken randomly from shrubs on the same plot, but the variety(s) of the Skimmia japonica species is unknown, as is whether or not the male and female plants are of the same variety. Therefore the observation of differences in chemical composition of male and female plants indicates that further study of this phenomenon would be of interest, but for this project, no further conclusions can be drawn. Surface extraction: Surface extraction of male and female Skimmia japonica leaves indicated that no furanocoumarins or furoquinolines were present on the epidermis. S. japonica has very thick, waxy leaves, and perhaps the extrusion of these compounds is not necessary from an ecological point of view, either for UV protection or protection from herbivores. Ruta graveolens Whole tissue compound quantification: Compounds identified from Ruta graveolens plants were skimmianine, 8-MOP and 5-MOP (Figure 2.2). 22 0.40J 0.35J Solvent: H 20:MeCN:MeOH (59:21:20) 0.30-E o 0.25-m CN g 0.201] 8-MOP skimmianine 20.00 Minutes 40.00 Figure 2.2: H P L C elution profile of Ruta graveolens crude extract. In addition to 8-MOP and 5-MOP, R. graveolens has also been reported to contain the furanocoumarins psoralen, isoimperatorin, pangelin, isopimpinellin, and chalepensin (Gray, 1983). Dictamnine has been reported to occur in R. graveolens, but was not detected in these extracts. In addition, three other furoquinolines have been reported to occur in R. graveolens, these being pteleine, y-fagarine, and kokusaginine (Mester, 1983). It is possible that they were at concentrations below the limit of detection by the HPLC, or the H P L C method used did not separate them into individual peaks. It is also possible that increased concentrations of furoquinolines and furanocoumarins would be induced by environmental stress or herbivory. It is known that production of furanocoumarins is inducible in Apiaceous plants by herbivorous damage or airborne methyl jasmonate (Zangerl, 1990; Miksch and Boland, 1996). 23 For the known furanocoumarins, the concentration of 8-MOP was found to be 900 ug/g FW, and the concentration of 5-MOP 300 ug/g FW (Table 2.4). The concentration of skimmianine in Ruta graveolens was found to be 500 ug/g FW. The amounts of furanocoumarins detected were comparable to a previous quantification of these furanocoumarins in Ruta graveolens leaf tissue. Zobel and Brown (1989) examined Ruta graveolens growing in the field, and found that 8-MOP concentration varied between 620 and 1600 ug/g FW, and the that concentration of 5-MOP varied between 300 and 490 ug/g FW. The concentration of skimmianine found in R. graveolens was 500 ug/g FW, close to the concentration found in male S. japonica leaves, 600 ug/g FW. Table 2.4: Concentrations (mean \u00C2\u00B1 S.E.) of various compounds in Ruta graveolens. ug/g F W A Plant n 8-MOP 5-MOP skimmianine dictamnine Ruta graveolens 9 950 \u00C2\u00B1100 300 \u00C2\u00B1 30 500 \u00C2\u00B1 40 0 A FW = fresh weight From the surface extraction of fresh Ruta graveolens leaves, it was found that all of the compounds detected from the whole leaf extraction occurred both on the epidermis and inside the leaf tissue. Furanocoumarins in Ruta graveolens tissue have previously been found to be on the epidermis and in the glands of stems and leaves and in the cortex parenchyma of stems and the mesophyll of leaves (Zobel and Brown, 1989). The leaves of R. graveolens are exceptionally thin and delicate, especially compared to other Rutaceous plants such as Skimmia or Citrus, which have thicker, waxy leaves. The presence of furanocoumarins and furoquinolines on the surface of the leaves could act both as protection against UV and as deterrents to herbivores. 24 Conclusions: Pure skimmianine was isolated from the aerial parts of Skimmia japonica, and pure dictamnine was isolated from the root of Dictamnus albus. Furanocoumarins and furoquinolines were detected in the Rutaceous plants Skimmia japonica and Ruta graveolens. Neither type of compound was detected on the leaf surface. The furanocoumarins present in Skimmia japonica were found to differ depending upon the sex of the plant, with 5-MOP being present in females, but not in males. Female plants also contained a greater concentration of skimmianine in their leaves than did male plants. No difference in compound concentration was found between leaves of different ages. However, caution should be used in drawing conclusions about these sex differences because the varieties of the various plants used are unknown. Ruta graveolens was found to contain skimmianine, 5-MOP and 8-MOP, and the furanocoumarins were quantified at concentrations similar to those reported previously in the literature. All three known compounds were present on both the surface and the inside of the leaves. For both S. japonica and R. graveolens, dictamnine was not detected although both plants have been reported to contain the alkaloid (Mester, 1983). There is the possibility that stress or herbivorous damage could induce the production of dictamnine, and perhaps other furanocoumarins or furoquinolines. 25 C H A P T E R 3 E f f e c t s o f F u r o q u i n o l i n e s o n t h e G r o w t h a n d F e e d i n g o f Spodoptera litura ( L e p i d o p t e r a : N o c t u i d a e ) INTRODUCTION There is interest in the potential role of phytochemicals as insect antifeedants and/or insect toxins. Most basically, the question is what are the possible functions of a chemical for a plant, that is, what role do chemicals play for the plant in the context of its physiology and ecology? Many plant chemicals have been found to have a role as toxins or deterrents to herbivorous insects. A single plant may have a variety of chemicals that are deterrent or insecticidal. The complexity and apparent redundancy of phytochemicals suggests resistance to different herbivores, potential synergy between compounds, and may reflect a pattern of coevolution between plants and insect herbivores. In this chapter the effect of the furoquinolines dictamnine and skimmianine on the growth and feeding of the larvae of Spodoptera litura Fab. (Lepidoptera: Noctuidae) is examined. This polyphagous species is a common crop pest in Asia, with host plants from at least 48 families. The only Rutaceous genus reported to be among the host plants for S. litura are Citrus spp. (Pogue, M. 2000), a genus from which only one furoquinoline has been reported (Da Silva et al., 1988). Spodoptera litura is used in these nutritional assays as a typical generalist lepidopteran species, not accustomed to high levels of furoquinolines in its diet. If indeed furoquinolines function as anti-herbivory compounds or toxins, then the compounds should affect this species of Lepidoptera. The effect of UV light is also examined in these experiments. Dictamnine has been shown to have phototoxic effects in vivo and in vitro to bacteria and fungi, although 26 skimmianine was not found to be phototoxic (Towers et al., 1981; Pfyffer etal., 1982a; Pfyffer and Towers, 1982). Rutaceous plants also contain furanocoumarins, well known antifeedants and toxins to many types of insects, including noctuid larvae in the genus Spodoptera (Yajima and Munakata, 1979; Diawara etal., 1993; Brewer etal., 1995; Berdegue etal., 1997). The effects of the furoquinolines on the growth and food consumption of the larvae were compared with the effects of the widespread furanocoumarin, 8-methoxypsoralen (8-MOP). The nutritional analysis experiments were designed to allow for a particular type of data analysis, known as ratio-based analysis. This is a common analytical method found in the literature used to describe the effect of diets on the growth and consumption of lepidopteran larvae (Waldbauer, 1968; Farrar et al., 1989). One advantage of the ratio-based analysis of nutritional assay data is that it theoretically allows separation of larval growth reduction due to antifeedant properties of the treatment diets, versus growth reduction due to physiological toxicity of the treatments. However, there are comments in the literature that raise concerns about potential inaccuracies of ratio-based analysis (Packard and Boardman, 1988; Raubenheimer and Simpson, 1992). In a ratio-based analysis, the indices RGRi and RCRi, which describe relative growth rate and relative consumption rate, are divided by initial body weight. This aims to compensate for any differences in growth or consumption due to the initial size of the insect. The other growth indices, AD, ECland ECD, describe approximate digestibility (AD) of the food, and the efficiency of conversion of ingested or digested food (ECI, ECD) into a change of body mass. These indices are standardized by dividing by the amount of food ingested. In order for these ratios to be mathematically valid, the denominator must vary isometrically (be linear and pass through the origin) with the numerator (Raubenheimer 2 7 and Simpson, 1992). For example, the nutritional index RGRi is found by dividing the weight gain of the larva by the initial body weight, divided by unit time. The assumption is made that weight gain of a larva in a given treatment will differ depending on the initial weight of the insect, and that this relationship between weight gain and initial insect weight is linear and has a y-intercept of zero. If this isometric relationship is not upheld, then the ratio values calculated will vary not only due to any treatment effect on the numerator value, but also because of the inconsistent relationship between numerator and denominator (Raubenheimer and Simpson, 1992). For ECI and E C D , there is the problem that the change in body weight may vary not only due to amount of food ingested, but will also depend on the initial body weight of the insect. This would lead to non-isometry in a plot of weight gain versus amount of food ingested. Another contribution to non-isometry could be due to a certain amount of ingested food going not towards increasing body weight, but towards physiological maintenance (Raubenheimer and Simpson, 1992). The data was examined in light of these criticisms, and the ratio variables were checked for isometry. The ratio-based system was found to be unsuitable. Consequently, the results were re-analyzed by direct analysis of the experimental data. Therefore, the primary methods of analysis used were A N C O V A and A N O V A analysis of directly measured variables. For the dictamnine treatment, the ratio analysis results are presented after the non-ratio analysis results for the purposes of comparison and discussion of analytical methodology. 28 MATERIALS AND METHODS Trials to study growth of and consumption by Spodoptera litura were conducted using the following compounds: the furoquinolines skimmianine and dictamnine, and the furanocoumarin 8-methoxypsoralen (8-MOP). Trials were done under both light and dark conditions. For the light conditions \"Vita-lite\" brand bulbs were used, which mimic the natural daylight spectrum, including the UV wavelengths, as much as is available in an artificial light. This was in order to include the approximate amount of UV radiation that the insects would be exposed to out of doors, and to determine if this affected the toxicity of the test compounds. Growth trial procedure: An artificial diet1 was used for all trials. Treatment diets were prepared by adding the experimental compound, dissolved in 1 mL methanol, to the dry diet base before preparation, and allowing the solvent to evaporate. Methanol was also added to the control diet. Control and treatment diet cubes were weighed to the milligram (initial wet weight was approximately 1 gram) and placed in individual plastic cups. The experimental larvae, newly moulted 5 t h instar Spodoptera litura, were weighed and one each was placed on a diet cube. The plastic cups were then capped, placed in a growth chamber and incubated at 25\u00C2\u00B0C in the light for a 16:8 hour day/night cycle, or in the case of the dark trials, covered from the light. After 3 days, the larvae were killed by freezing. Larvae, frass and remaining diet were separated, dried, and weighed. Drying was done overnight in an oven at 60\u00C2\u00B0C. The conversion ratios of wet to dry weight were determined initially for 20 portions of diet, and 20 sets of 3 larvae. 1 Diet: No. 9795 BioServ.Inc. Frenchtown, N.J., with added ground alfalfa to improve palatability to larvae. Finished diet preparation consists of 17.5% diet base, 3.5% agar, and 80% water. 29 Treatments: Multiple trials were carried out for each compound in both light and dark conditions. The concentration of furoquinolines used was less than the concentration of skimmianine found in Rutaceous plants tested, which ranged from 500 to 1800 ug/g fresh weight (FW). Skimmianine was used in treatments in concentrations of 10 to 80 ug /g FW, and dictamnine in concentrations of 10 to 160 ug / g FW. The low concentrations were used both because of a limit on the amount of chemicals available, and also because significant reductions in growth rate were seen using these concentrations. 8-MOP, which is available commercially, was used in treatments at concentrations ranging from 20 to 640 ug /g FW. Non-ratio-based Data analysis: For the non-ratio-based analysis, the measured variables of final weight and amount of food consumed for the larvae in each treatment were examined directly. First of all, for each treatment, both dependant variables, final body weight and amount of food consumed, were plotted against initial mass. This was to determine if the response to the treatment covaried with the initial weight of the insect. If there was a linear relationship between dependant variable and initial weight, and if the slopes of the regression lines for each treatment within a trial were found not to be significantly different, then this allowed for an analysis of covariance procedure (ANCOVA) (Raubenheimer and Simpson, 1992). Using the average slope, each of the final y-values of the treatment samples was adjusted to the equivalent y value at the mean x value (initial weight) of all larvae in the trial. This essentially adjusts the results of the dependant variable when they vary due to initial weight (Packard and Boardman, 1988). 30 The mean of the adjusted dependant variable for each treatment was then found, and the mean treatment values within each trial were compared by A N O V A with the multiple comparison Tukey test. Simple A N O V A analysis of the mean final weights or amount consumed of the larvae in each treatment was also done for some trials. This was because there were certain cases where A N C O V A was not appropriate, for one of two reasons. In the first case, if regression of the dependant variable (final weight or amount consumed) versus initial weight was not linear for a number of treatments within a trial, this indicated that initial weight was not a variable in the prediction of final weight or consumption. In this case, A N C O V A with initial weight as the covariate was not appropriate, and simple A N O V A of the variable was suitable. In the second case, even if the slopes of all regression lines were linear, if the slopes of the regressions of the treatments within a trial were found to be significantly different from each other, then A N C O V A was also not appropriate. In this second case, A N O V A was also not appropriate, so these treatment data were not included in the overall analysis. Ratio-based data analysis: For the ratio-based analysis, the data obtained was used to calculate the following indices of growth (Waldbauer, 1968): Relative growth rate, arithmetic weight mean (RGRi): Weight gain relative to initial body weight, per day: (AB/BI)/T Relative consumption rate, initial weight (RCRi): Consumption of food relative to initial body weight, per day: (l/BI)/T Approximate Digestibility (AD): percentage of ingested food that is digested: [(I-F)/I]x100. 31 Efficiency of conversion of ingested food (ECI): Change in body weight in proportion to amount of food consumed: (AB/I) x 100. Efficiency of conversion of digested food (ECD): Change in body weight in proportion to amount of food digested: [AB/(I-F)] x 100. I = weight of food ingested F= weight of frass AB = change in body weight Bl = initial body weight T = time (days) Analysis of the dictamnine light trial by the ratio-based method is presented in the results and discussion section for the purposes of comparison to non-ratio-based methods and discussion of analytical methodology. Comparison of multiple trials: Since the trials were not all done at the same time, if results from different trials were to be compared with each other it was necessary to compensate for possible differences in results due to circumstances such as temperature fluctuations, or the exact duration time of the assay. For both ratio-based and non-ratio-based analyses, to compare means from different trials the treatment means were expressed as a value that was a percentage of the mean of the control treatment value. Treatments could then be compared by how much their means decreased from that of the control, set at 100%. However, when comparing mean values as percentages of control between trials, the error term around the control value must also be taken into consideration when deciding whether or not treatment means from different trials are significantly different or 32 not. Essentially, the variability around the control (100%) value must be added to that of the treatment mean variability when means between trials are being compared. Statistical computations were done using the statistical programs S Y S T A T 8.0 and Microsoft Excel '95. A N O V A analysis included multiple comparisons using the Tukey test. R E S U L T S A N D DISCUSSION (A) Analysis of Final weights and Consumption of Larvae on Various Diets For the majority of the trials, A N C O V A analysis was found to be appropriate. Appendix 3 outlines the A N C O V A procedure in more detail, using the dictamnine light trial as an example. For some of the treatments within the trials, it was found that there was no linear relationship between initial weight and the dependant variable being tested. This indicated that initial weight was not a covariate affecting final weight or consumption, or that initial body weight had no effect on the final weight or consumption of the insect over the course of the trial. In these cases, analysis of the treatment final weight or consumption means using direct A N O V A was appropriate. For the skimmianine dark trial, the treatment regressions of initial weight versus final weight were significant, however it was found that the slopes for treatments over 40 ug/g FW were significantly different from each other and from the treatments at lower concentrations. Neither A N C O V A nor A N O V A could be used for analysis, so these treatments have been left out of the discussion. The results for mean final weight and mean amount consumed for the treatments within each trial, analyzed by A N C O V A or ANOVA, are presented in Tables 3.1 to 3.6. 33 Table 3.1: (a) Mean final body weights (mg dry weight) \u00C2\u00B1 95% confidence intervals (b) Mean consumption (mg dry weight) \u00C2\u00B1 95% confidence intervals of Spodoptera litura larvae fed skimmianine diets in light. (a) Final body weight Diet (ug/g FW) A N C O V A M ^ 9 5 % Mean \u00C2\u00B1 ^ j As % of control * A N O V A 95% Mean \u00C2\u00B1 ^ j As % of control * control 42.3 \u00C2\u00B1 1.6 100% \u00C2\u00B1 3.8% a 40.6 \u00C2\u00B1 3.5 100% \u00C2\u00B1 8.7% a 10 36.6 \u00C2\u00B1 2.7 86% \u00C2\u00B1 6.3% b 20 27.9 \u00C2\u00B10 .7 66% + 1.6% c 40 n.l. 21.9 \u00C2\u00B1 1.7 54% \u00C2\u00B1 4.2% b control 80 52.1 \u00C2\u00B12 .9 14.4 + 0.9 100% \u00C2\u00B15 .5% 28% \u00C2\u00B1 1.7% a b (b) Consumpt ion Diet (ug/g FW) A N C O V A Mean \u00C2\u00B1 ^ | As % of control * A N O V A Mean \u00C2\u00B1 ^ j * As % of control control 100.3 \u00C2\u00B15 .8 100% \u00C2\u00B1 5.8% a 95.3 \u00C2\u00B17 .9 100% \u00C2\u00B18 .3% a 10 85.0 \u00C2\u00B14 .2 85% \u00C2\u00B1 4.2% b 20 76.9 \u00C2\u00B12 .5 77% \u00C2\u00B1 2.5% b 40 n.l. 51.1 \u00C2\u00B14 .5 54% \u00C2\u00B1 4.7% b control 80 130.9 \u00C2\u00B19 .8 29.1 \u00C2\u00B1 3.8 100% \u00C2\u00B17 .5% 22% \u00C2\u00B1 2.9% a b * different lowercase letters indicate significant differences between means within each trial (not between trials). Control treatment for each trial always given the letter \"a\". Analysis used ANOVA or ANCOVA with multiple comparison Tukey tests, n.l. = not linear. Indicates that regression of initial weight versus final weight or consumption did not indicate a significant linear relationship. Therefore ANOVA, not using initial weight as a covariate, was valid. 34 Table 3.2: (a) Mean final body weights (mg dry weight) \u00C2\u00B1 95% confidence intervals (b) Mean consumption (mg dry weight) \u00C2\u00B1 95% confidence intervals of Spodoptera litura larvae fed skimmianine diets in dark. (a) Final body weight Diet (ug/g FW) A N C O V A *A , 95% Mean \u00C2\u00B1 c ( As % of control control 52.2 \u00C2\u00B1 1.5 100% \u00C2\u00B1 2.9% a 10 44.9 \u00C2\u00B1 2.2 86% \u00C2\u00B1 4.3% b 20 30.9 \u00C2\u00B1 1.2 59% \u00C2\u00B1 2.3% c 40 d.s. control (2) 80 (2) d.s. d.s. (b) Consumpt ion Diet A N C O V A (ug/g FW) Mean \u00C2\u00B1 J \u00E2\u0084\u00A2 As % of control * control 124.6 \u00C2\u00B14 .0 100% \u00C2\u00B1 3 . 2 % a 10 106.8 + 5.6 86% \u00C2\u00B14 .5% b 20 d.s. 40 d.s. control 106.7 \u00C2\u00B110.1 100% \u00C2\u00B19 .5% c 80 40.3 \u00C2\u00B12 .3 38% \u00C2\u00B1 2 . 1 % d * different l o w e r c a s e letters indicate signif icant d i f ferences be tween m e a n s within e a c h trial (not be tween trials). Cont ro l t reatment for e a c h trial a lways g iven the letter \" a \" . A n a l y s i s used A N C O V A with mult iple compar i son T u k e y tests. d.s.=different s l opes . Indicates A N C O V A and A N O V A a n a l y s e s were inappropr iate b e c a u s e the regress ion of initial weight ve rsus final weight or consumpt ion resul ted in the t reatments hav ing signi f icant ly different s l opes f rom one another and the contro l . 35 Table 3.3: (a) Mean final body weights (mg dry weight) \u00C2\u00B1 95% confidence intervals (b) Mean consumption (mg dry weight) \u00C2\u00B1 95% confidence intervals of Spodoptera litura larvae fed dictamnine diets in light. (a) Final body weight Diet (ug/g FW) A N C O V A Mean \u00C2\u00B1 ^ j As % of control A N O V A M ^ 9 5 % Mean \u00C2\u00B1 ^ ( As % of control control 39.5 \u00C2\u00B12 .3 100% \u00C2\u00B15 .9% a 10 36.1 \u00C2\u00B1 2.1 91% \u00C2\u00B15 .4% a 20 32.4 \u00C2\u00B1 1.7 82% \u00C2\u00B1 4.3% b control 40 80 120 160 63.8 \u00C2\u00B12 .5 45.4 \u00C2\u00B1 2.6 39.0 \u00C2\u00B13 .1 n.l. 19.6 \u00C2\u00B10 .8 100% \u00C2\u00B14 .0% 71% \u00C2\u00B1 4 . 1 % 61% \u00C2\u00B14 .8% 31% \u00C2\u00B1 1.3% a b c d 64.8 \u00C2\u00B1 3.6 26.7 + 2.6 100% \u00C2\u00B1 5 . 5 % a 41% \u00C2\u00B1 4 . 0 % b (b) Consumpt ion Diet A N C O V A A N O V A (ug/g FW) KM ^ 9 5 % Mean \u00C2\u00B1 ^ j As % of control * M ^ 9 5 % Mean \u00C2\u00B1 ^ j As % of control * control 102.9 \u00C2\u00B1 5.0 100% \u00C2\u00B14 .8% ab 10 96.94 \u00C2\u00B1 4.2 94% \u00C2\u00B1 4 . 1 % be 20 87.13 \u00C2\u00B14 .0 85% \u00C2\u00B1 3.8% c control 166.9 \u00C2\u00B19 .2 100% \u00C2\u00B15 .5% a 168.2 \u00C2\u00B1 10.4 100% \u00C2\u00B1 6 . 2 % a 40 117.6 \u00C2\u00B17 .5 70% \u00C2\u00B1 4.5% b 80 n.l. 100.4 \u00C2\u00B19 .83 60% \u00C2\u00B1 5.8% b 120 n.l. 69.66 \u00C2\u00B1 5.80 41% \u00C2\u00B1 3 . 4 % c 160 \u00E2\u0080\u00A2 n.l. 56.75 \u00C2\u00B14.78 34% \u00C2\u00B1 2.8% c * different l owe rcase letters indicate signif icant d i f ferences be tween m e a n s within e a c h trial (not be tween trials). Cont ro l t reatment for e a c h trial a lways g iven the letter \" a \" . A n a l y s i s used A N O V A or A N C O V A with mult iple compar i son T u k e y tests, n.l. = not l inear. Indicates that regress ion of initial weight ve rsus final weight or consumpt ion did not indicate a signi f icant l inear relat ionship. There fo re A N O V A , not us ing initial weight as a covar ia te , w a s val id . 36 Table 3.4: (a) Mean final body weights (mg dry weight) \u00C2\u00B1 95% confidence intervals (b) Mean consumption (mg dry weight) \u00C2\u00B1 95% confidence intervals of Spodoptera litura larvae fed dictamnine diets in dark. (a) Final body weight Diet (ug/g FW) A N C O V A Mean \u00C2\u00B1 ^ ( As % of control A N O V A Mean + c ( As % of control control 10 20 42.4 \u00C2\u00B1 4.4 31.3 \u00C2\u00B1 1.6 35.6 \u00C2\u00B12 .0 100% \u00C2\u00B1 10.3% 74% \u00C2\u00B1 3.8% 84% \u00C2\u00B1 4.6% a b b 42.5 \u00C2\u00B1 5.2 34.8 \u00C2\u00B13 .8 35.1 \u00C2\u00B1 3.6 100% \u00C2\u00B112 .3% a 82% \u00C2\u00B1 9.0% b 83% \u00C2\u00B1 8.4% b control 40 80 120 160 66.7 \u00C2\u00B12 .0 48.3 + 3.7 35.2 \u00C2\u00B12 .9 n.l. n.l. 100% \u00C2\u00B1 3.1% 72% + 5.5% 53% \u00C2\u00B1 4.3% a b c 66.7 \u00C2\u00B13 .3 23.6 \u00C2\u00B12 .4 22.3 \u00C2\u00B12 .1 100% \u00C2\u00B1 4 . 9 % a 35% \u00C2\u00B1 3.5% b 33% \u00C2\u00B1 3 . 1 % b (b) Consumpt ion Diet (ug/g FW) A N C O V A 95% Mean \u00C2\u00B1 ^ | As % of control A N O V A ^ 95% Mean \u00C2\u00B1 ^ | As % of control * control 10 20 104.4 \u00C2\u00B15 .4 90.0 \u00C2\u00B12 .9 83.5 \u00C2\u00B12 .2 100% \u00C2\u00B1 5.2% 86% \u00C2\u00B1 2.8% 80% \u00C2\u00B1 2 . 1 % control 40 80 120 160 n.l n.l n.l n.l n.l 175.5 \u00C2\u00B1 13.4 121.9 \u00C2\u00B1 12.6 90.0 \u00C2\u00B1 12.9 56.6 \u00C2\u00B15 .4 49.1 \u00C2\u00B16 .4 100% \u00C2\u00B1 7 . 6 % a 69% \u00C2\u00B1 7.2% b 51% \u00C2\u00B17 .4% c 32% \u00C2\u00B1 3 . 1 % d 28% \u00C2\u00B1 3.6% d * different lowercase letters indicate significant differences between means within each trial (not between trials). Control treatment for each trial always given the letter \"a\". Analysis used ANOVA or ANCOVA with multiple comparison Tukey tests, n.l. = not linear. Indicates that regression of initial weight versus final weight or consumption did not indicate a significant linear relationship. Therefore ANOVA, not using initial weight as a covariate, was valid. 37 Table 3.5: (a) Mean final body weight (mg dry weight) \u00C2\u00B1 95% confidence interval (b) Mean consumption (mg dry weight) \u00C2\u00B1 95% confidence interval of Spodoptera litura larvae fed 8-MOP diets in light. (a) Final body weight Diet (ug/g FW) A N C O V A ^ 95% Mean \u00C2\u00B1 c j As % of control * A N O V A \u00E2\u0080\u00A2 . , 95% Mean \u00C2\u00B1 c ( As % of control * control 56.1 \u00C2\u00B13 .1 100% \u00C2\u00B15 .5% a 59.1 \u00C2\u00B1 3.5 100% \u00C2\u00B15 .9% a 20 52.3 \u00C2\u00B1 2.5 93% \u00C2\u00B1 4.5% a 40 53.9 \u00C2\u00B13 .6 96% \u00C2\u00B1 6.5% a 80 n.l. 48.6 \u00C2\u00B13 .5 82% \u00C2\u00B1 5.9% b control 160 320 480 640 n.l. n.l. n.l. n.l. n.l. 75.1 \u00C2\u00B14 .0 46.3 \u00C2\u00B14 .5 29.5 \u00C2\u00B1 3.4 24.4 \u00C2\u00B12 .1 21.9 \u00C2\u00B1 2.4 100% \u00C2\u00B15 .4% 62% \u00C2\u00B1 5.9% 39% \u00C2\u00B1 4.5% 33% \u00C2\u00B1 2.8% 29% \u00C2\u00B1 3.3% a b c cd d (b) Consumpt ion Diet (ug/g FW) A N C O V A 95% Mean \u00C2\u00B1 c ( As % of control * A N O V A 95% Mean \u00C2\u00B1 ^ ( As % of control * control 148.6 \u00C2\u00B16 .6 100% \u00C2\u00B14 .5% a 156.6 \u00C2\u00B1 8.6 101% \u00C2\u00B15 .5% a 20 130.1 \u00C2\u00B1 8.4 88% \u00C2\u00B1 5.7% b 40 n.l. 129.9 \u00C2\u00B1 10.1 84% \u00C2\u00B1 6.4% b 80 n.l. 119.7 \u00C2\u00B17 .0 77% \u00C2\u00B1 4.4% b control 160 320 480 640 n.l. n.l. n.l. n.l. n.l. 195.6 \u00C2\u00B17 .5 112.7 \u00C2\u00B1 10.9 72.2 \u00C2\u00B19 .0 54.2 \u00C2\u00B17 .5 62.0 \u00C2\u00B15 .7 100% \u00C2\u00B13 .8% 58% \u00C2\u00B1 5.6% 37% \u00C2\u00B1 4.6% 28% \u00C2\u00B1 3.8% 32% \u00C2\u00B1 2.9% a b c d cd * different lowercase letters indicate significant differences between means within each trial (not between trials). Control treatment for each trial always given the letter \"a\" . Analysis used A N O V A or A N C O V A with multiple comparison Tukey tests, n.l. = not linear. Indicates that regression of initial weight versus final weight or consumption did not indicate a significant linear relationship. Therefore A N O V A , not using initial weight as a covariate, was valid. 38 Table 3.6: (a) Mean final body weight (mg dry weight) \u00C2\u00B1 95% confidence interval (b) Mean consumption (mg dry weight) \u00C2\u00B1 95% confidence interval of Spodoptera litura larvae fed 8-MOP diets in dark. (a) Final body weight Diet A N C O V A A N O V A (ug/g FW) Mean \u00C2\u00B1 95% C.I. As % of control * Mean \u00C2\u00B1 95% C.I. As % of control * control 50.1 \u00C2\u00B1 3.0 100% \u00C2\u00B15 .9% ab 20 46.2 \u00C2\u00B12 .2 92% \u00C2\u00B1 4.4% b 40 51.8 \u00C2\u00B1 3.5 103% \u00C2\u00B1 7.0% a control n.l. 71.4 \u00C2\u00B14 .3 100% \u00C2\u00B1 6 . 1 % a 160 n.l. 35.4 \u00C2\u00B12 .6 50% \u00C2\u00B1 3.7% b 320 n.l. 30.8 \u00C2\u00B12 .0 43% \u00C2\u00B1 2.9% b 480 n.l. 25.1 \u00C2\u00B1 2.6 35% \u00C2\u00B1 3.7% c 640 n.l. 25.4 \u00C2\u00B1 2.2 36% \u00C2\u00B1 3 . 1 % c (b) Consumption Diet A N C O V A A N O V A (ug/g FW) , 95% Mean \u00C2\u00B1 ^ | As % of control * 95% Mean \u00C2\u00B1 ^ j As % of control * control 124.1 \u00C2\u00B16 .6 100% \u00C2\u00B15 .3% a 20 104.6 \u00C2\u00B14 .5 84% \u00C2\u00B1 3.6% b 40 127.9 \u00C2\u00B1 10.0 103% \u00C2\u00B1 8.0% a control n.l. 191.9 \u00C2\u00B1 12.1 100% \u00C2\u00B1 6 . 3 % a 160 n.l. 85.0 \u00C2\u00B18 .6 44% \u00C2\u00B1 4.5% b 320 n.l. 77.0 \u00C2\u00B1 8.7 40% \u00C2\u00B1 4.5% b 480 n.l. 60.0 \u00C2\u00B16 .6 31% \u00C2\u00B1 3.4% c 640 n.l. 59.2 \u00C2\u00B14 .3 31% \u00C2\u00B1 2 . 2 % c * different l o w e r c a s e letters indicate signif icant d i f ferences be tween m e a n s within e a c h trial (not be tween trials). Cont ro l t reatment for e a c h trial a lways g iven the letter \" a \" . A n a l y s i s used A N O V A or A N C O V A with mult iple compar i son T u k e y tests, n.l. = not l inear. Indicates that regress ion of initial weight ve rsus f inal weight or consumpt ion did not indicate a signi f icant l inear re lat ionship. There fo re A N O V A , not us ing initial weight a s a covar ia te , w a s va l id . 39 Furoquinolines: Light effects Both of the furoquinolines, skimmianine and dictamnine, significantly reduced both consumption and final weight of Spodoptera litura larvae. For both furoquinoline trials, the presence of light did not significantly affect the percent decrease in growth or consumption of the treatment larvae as compared to those in the control (Tables 3.1-3.4). Furoquinolines have been reported to have UV light-induced toxic properties to organisms in vivo and in vitro (Pfyffer and Towers, 1982; Pfyffer et al., 1982a, 1982b). There are several possible reasons why UV light did not alter the effect of the furoquinolines in these experiments. First of all, it could be that the insect gut and body tissues are adequately shielded from the light by pigments in the skin of the larvae. If little light energy permeates the integument of the insect, then the danger of phototoxicity is reduced. If light energy does permeate into body tissues, one of the harmful effects of phototoxic compounds is not due to the excited phototoxins themselves, but to reactive oxygen species such as singlet oxygen and superoxide radicals that are produced when these compounds react with oxygen (Berenbaum, 1987). Insects have antioxidant enzymes, for example superoxide dismutase, that react with potentially harmful radicals. Therefore, a certain amount of phototoxicity can be mitigated by larval enzymes. In addition, compounds in the diet of insects can quench radicals, absorbing the energy without producing another harmful radical. These include carotenoids, flavonoids and ascorbic and uric acid (Berenbaum, 1987; Timmerman et al., 1999). Finally, there is also the possibility that the incident UV flux from the light source was not strong enough to elicit the potentially photoactive effects of the chemicals. However, the light conditions were designed to mimic those found outdoors, and thus were intended to reproduce any photoactivity that could be expected in daylight. 40 Due to the lack of significant differences between dark and light treatments, subsequent comparisons between the dictamnine and skimmianine trials refer to those under simulated daylight conditions. Furoquinolines: skimmianine versus dictamnine Skimmianine consistently caused a greater reduction in consumption and growth than dictamnine at similar dietary concentrations. At the lowest experimental concentration used, 10 ug/g FW, skimmianine caused a 15% decrease in consumption and final weight from the values of the control larvae (Table 3.1). The lowest concentration of dictamnine that affected larval consumption and growth was 20 ug/g FW, where there was a 15-20% reduction in consumption and final larval weight (Table 3.1). The highest experimental concentration of skimmianine used, 80 ug/g FW, reduced final weight and consumption by about 75% from the control larvae. In comparison, at this concentration dictamnine reduced larval consumption and final weight by about 40% from that of the control larvae. The highest dictamnine concentration used, 160 ug/g FW, reduced larval growth and consumption by 70% from the control larvae (Table 3.3). These results indicate that for the same decrease in larval consumption and growth, skimmianine is effective at concentrations (in ug/g) approximately one half the amount of dictamnine that is required. However, if the molarity of the compounds is considered, skimmianine is more than twice as effective as dictamnine for a given concentration; the molecular weight of skimmianine is 256 g/mole, that of dictamnine is 199 g/mole. This is the first report of the effects of any furoquinoline on insect feeding or growth. Skimmia japonica has previously been screened for antifeedant compounds against Spodoptera exigua, but only the three furanocoumarins 5-MOP, 8-MOP and 41 oxypeucedanin were reported to be responsible for the antifeedant properties of the plant (Escoubas et al., 1992). Although there has been much study of the toxic mechanisms of furoquinolines (Pfyffer etal., 1982a, 1982b; Ashwood-Smith etal., 1982; Mizuta and Kanamori, 1985; Goloubkova, 1998), the results of the present experiments point to antifeedant action of the furoquinolines being solely responsible for the decrease in growth of the larvae. For both skimmianine and dictamnine, the percent reduction in final larval weight and consumption are not significantly different from each other for all dietary concentrations (Tables 3.1, 3.3). It appears that a given percent reduction in final larval weight is therefore due to an equivalent percentage decrease in consumption. It is possible that at the concentrations used in this experiment, the furoquinolines were solely antifeedant, but at higher concentrations they could also be toxic. The lack of light effects too could indicate that under different experimental conditions, for example, a stronger source of UV radiation, furoquinolines could be toxic. Furanocoumarin: 8-MOP: Light effects Treatment diets containing the furanocoumarin 8-MOP significantly reduced both consumption and final weight of Spodoptera litura larvae as compared to larvae on the control treatment. For 8-MOP, as for the furoquinolines, the presence of light did not significantly affect the percent decrease in growth or consumption of the treatment larvae as compared to those in the control (Tables 3.5, 3.6). The lack of light effect in these trials is probably as previously discussed with respect to the furoquinoline trials. The effect of UV light on the action of 8-MOP in the diet of lepidopteran larvae has been previously examined. However, the literature is somewhat contradictory in terms of the effect of UV light on the action of 8-MOP and other furanocoumarins. Berenbaum and Neal (1985) demonstrated that UV light was not required for 8-MOP to be toxic to 42 neonate Helicoverpa zea. A later study, however, also using neonate H. zea, found that a diet containing 8-MOP had a low effect (less than 20% mortality) on neonate larval survivorship when no UV light was present. The presence of UV light with the same diet increased the mortality of larvae on this diet to about 50% (Berenbaum et al., 1991). The relative effect of UV light on the toxicity of 8-MOP is still to be elucidated. As in the discussion of the furoquinolines, subsequent trial comparisons are using the data from the 8-MOP trial under simulated daylight conditions. Comparison of 8-MOP to furoquinolines The reduction in larval consumption and growth by the furanocoumarin, 8-methoxypsoralen, was significantly less than the reduction by similar concentrations of both skimmianine and dictamnine (Tables 3.1, 3.3, 3.5). 8-MOP started to reduce larval consumption at the concentration of 20 ug/g FW, and this 20% decrease in consumption from that of the control larvae remained constant for diets up to the 80 ug/g FW concentration (Table 3.5). The final larval weight of the insects did not change from the control group until the 80 ug/g FW concentration, when the final weight decreased significantly by 20% from the control larvae. This decrease is much less dramatic than the decrease seen with the furoquinolines. At this concentration, skimmianine had reduced final weight and consumption by about 75% from the control larvae, and dictamnine had reduced larval consumption and final weight by about 40% from that of the control larvae. When larvae were fed diets ranging in concentration of 8-MOP from 160 ug/g FW to 640 ug/g FW, there was a steady decrease in consumption and final weight by approximately 40% to 70% from the means of the control group. At 640 ug /g FW, 8-MOP had still not decreased larval consumption or growth to the same degree as skimmianine had at a concentration 8 times lower. If the molarity of the diets is 43 considered, then the deterrence of skimmianine is greater; the molecular weight of 8-MOP is 216 g/mole, that of skimmianine is 256 g/mole. In addition to the lower potency of 8-MOP as compared to the furoquinolines at a given concentration, the rate of decrease in consumption of growth of larvae was much lower per unit increase in compound dietary concentration for 8-MOP as compared to the furoquinolines. However, similar to the furoquinolines, the decrease in final weight of the larvae appeared to be due to a decrease in consumption. For every treatment, the amount of decrease in consumption as compared to control larvae was not significantly different from the decrease in final weight as compared to control larvae. This indicates that in these experiments, 8-MOP was acting as an antifeedant, not as a physiological toxin. Other studies have also found 8-MOP to have only antifeedant effects without toxicity. Klocke et al. (1989) found that under non-UV light conditions 8-MOP deterred feeding but was not physiologically toxic to neonate Helicoverpa virescens larvae. In their study, the antifeedant properties of 8-MOP reduced the final larval weight by half (as compared to control larvae) at a concentration of 110-115 ug/g FW. Berdegue et al. (1997) also report only antifeedant properties of 8-MOP. The highest 8-MOP concentration tested in their experiment, however, was 18 ug/g FW. Other studies report both antifeedant and toxic effects of dietary 8-MOP to larvae. L C 5 0 2 values of 8-MOP in the literature are variable. For example, the L C 5 0 value for neonate Helicoverpa zea in non-UV light conditions was 96 ug/g FW (Berenbaum and Neal, 1985), while Diawara et al. (1993) found an L C 5 0 value of 245.9 ug/g FW (+/- 59.2) for neonate Spodoptera exigua in simulated daylight conditions. These studies found 8-MOP to be much more effective at lower concentrations than was 2 The concentration of 8-MOP required to reduce larval weight by 50%. 44 found in this study. For example, at 80 ug/g FW of 8-MOP in these experiments, larval final weights had just started to decrease. However, neonate larvae were used in the cited experiments, which have a body weight that is far below the body weight of the fifth instar larvae used in this study. In addition, there have been reports of much higher 8-MOP concentrations required to cause harmful effects. Yajima and Munakata (1979) found that neither 8-MOP nor 5-MOP fed to S. litura larvae caused any mortality up to concentrations of 500 ug/g. The lack of toxicity of 8-MOP in this study, provided in dietary concentrations up to 640 ug/g FW, is still unexpected, considering the reports in the literature. This, combined with the lack of light effect, suggests that the experimental conditions, probably insufficient UV wavelengths reaching larval body tissues, is probably why toxicity was not observed. The question does arise, however, that in an ecological context, would natural daylight provide enough UV radiation to cause furanocoumarin toxicity? Summary: The results of these experiments indicate that both furoquinolines, skimmianine and dictamnine, are antifeedant to Spodoptera litura larvae. 8-MOP was also found only to be antifeedant, but not toxic. The presence of UV light did not affect the antifeedant properties of the furoquinolines or 8-MOP. Most studies in the literature report 8-MOP to be both toxic and antifeedant, and this has been observed under both non-UV (Berenbaum and Neal, 1985) and UV light conditions (Berenbaum etal., 1991; Diawara et al., 1993; Brewer et al., 1995). However, there are some exceptions; Klocke et al., (1989) report only antifeedant action of 8-MOP, under non-UV light conditions, and Berdegue et al. (1997) report that 8-MOP alone is only antifeedant at a concentration of 18 ug/g FW under simulated daylight conditions. 8-MOP toxicity was not elicited in 45 these experiments perhaps because of robust insect physiology that could cope with the amount of toxicity, photo-dependant or otherwise, encountered from the concentrations of chemicals given in the experimental diets. The possibility that the furoquinolines, shown to be potent antifeedants, could also be toxic can therefore not be ruled out. Since the structure of the furoquinolines is similar to that of the furanocoumarins, it could be that the same conditions needed to elicit toxicity from the furanocoumarins, are also required for furoquinolines to be toxic. Further experiments, perhaps using a higher intensity of UV light and higher concentrations of furoquinolines, should be done before ruling out the possibility that these compounds are toxic. What has been clearly demonstrated from the results of this experiment is that the two furoquinolines, dictamnine and skimmianine, are potent antifeedant compounds for Spodoptera litura fifth instar larvae, and at concentrations far below those found in Rutaceous plants. B. A Comparison of Methods: Ratio-based versus non-ratio-based The growth experiments were designed for the use of ratio-based data analysis. As outlined in the procedure section, this involves expressing growth rate and consumption as a ratio to initial weight. For this procedure to be accurate, it requires that the numerator (weight gain or amount consumed) vary isometrically with the denominator (initial weight). To determine whether or not this occurred for these data, the data were plotted and regressed. The results for the dictamnine trial in the light are presented as typical of the trends seen. Weight gain versus initial weight for each treatment is shown in Figure 3.1. 46 Figure 3 .1 : We igh t gain v e r s u s initial weight of S . litura la rvae on t reatment diets conta in ing different concent ra t ions of d ic tamnine. R e g r e s s i o n l ines a l so s h o w n . From Figure 3.1 it is clear that weight gain versus initial weight for the treatments in this trial are not isometrically related. None of the treatment regressions have a y-intercept close to zero, and the 120 ug/g FW treatment does not exhibit a linear relationship between initial and final weights. When a regression line has a y-intercept different from zero, the y value divided by its x value will yield a different ratio at different points along the regression line. The further away the intercept from zero, the larger the variation between ratios of different initial weights. In the case of this data, for treatments with a y-intercept above zero, as the initial weight increases, the ratio value of weight gain divided by initial weight decreases along the regression line. For treatments with a negative y-intercept, the ratio decreases for higher initial body weights. Therefore, finding the average value of treatment ratios that divide weight gain by initial weight distorts the final indices from being representitive. 47 A chart of consumption versus initial weight shows a similar lack of isometry (Figure 3.2). Figure 3.2: C o n s u m p t i o n ve rsus initial weight of S . litura on treatment diets conta in ing different concent ra t ions of d ic tamnine. R e g r e s s i o n l ines are a l so s h o w n . Figure 3.2 illustrates that consumption and initial weight do not have an isometric relationship for the linear treatments, and the treatments from 80 to 160 ug/g FW were not significantly linear according to regression analysis. In these cases as well, the use.' of the average ratios of each treatment, division of amount consumed by initial weight would yield inaccurate results. Since the data were so dramatically non-isometric, the ratios for R G R i and RCRi were calculated in order to compare the results to those obtained by the A N C O V A / A N O V A analysis. 48 Relative consumption and growth ratios based on initial body weights were calculated for the dictamnine treatments and compared with the results from the A N C O V A and A N O V A analyses. Mean values for relative growth rate based on initial body size (RGRi) and relative consumption rate based on initial body size (RCRi) are presented in Table 3.7. Tabulated are the mean nutritional index values, and also the values as percent of the mean of the control treatment of each trial. Table 3.7: RGRi and RCRi of Spodoptera litura on dictamnine diets in light diet RGRi as % of control RCRi as % of control ug/g FW Mean 95%CI Mean 95%CI * Mean 95%CI Mean 95%CI * control 0.79 + 0.12 100% \u00C2\u00B1 14.5% a 3.04 \u00C2\u00B1 0.25 100% \u00C2\u00B18 .4% a 10 0.71 \u00C2\u00B10.06 90% \u00C2\u00B1 7.2% a 2.69 \u00C2\u00B10.16 89% \u00C2\u00B1 5.3% ab 20 0.61 \u00C2\u00B10.07 77% \u00C2\u00B1 8.7% b 2.49 \u00C2\u00B10.15 82% \u00C2\u00B1 5 . 1 % b control 1.51 \u00C2\u00B10.10 100% \u00C2\u00B16 .7% a 4.79 \u00C2\u00B1 0.35 100% \u00C2\u00B17 .4% a 40 0.95 \u00C2\u00B10.07 63% \u00C2\u00B1 4.8% b 3.19 \u00C2\u00B10.24 67% \u00C2\u00B1 5.0% b 80 0.76 \u00C2\u00B10.10 51% \u00C2\u00B16 .5% b 2.91 \u00C2\u00B10.28 61% \u00C2\u00B15 .7% b 120 0.45 + 0.08 30% \u00C2\u00B1 5.2% c 2.07 \u00C2\u00B10.18 43% \u00C2\u00B1 3.7% c 160 0.26 \u00C2\u00B1 0.05 17% \u00C2\u00B13 .5% d 1.69 \u00C2\u00B10.14 35% \u00C2\u00B1 2.9% d * different l owe rcase letters indicate signif icant d i f ferences be tween m e a n s within e a c h trial (not be tween trials). Cont ro l t reatment for e a c h trial a lways g iven the letter \" a \" . A n a l y s i s used A N O V A or A N C O V A with mult iple compar i son T u k e y tests. The RGRi and RCRi values for the larvae in the dictamnine light trial were calculated. The percent decrease in RGRi of larvae as dictamnine concentration increased tended to be greater than those calculated by A N C O V A or A N O V A analysis of change in final weights (Tables 3.3, 3.7). This difference increased as dictamnine concentrations increased. From A N O V A and A N C O V A analysis, the treatments with concentrations from 80 to 160 ug/g FW, the final weight of the larvae decreased by 49 about 40-45% up to 70% from that of the control treatment larvae. The mean RGRi of the larvae on these treatments decreased by about 50-55% up to 80% from the control larvae mean. The RGRi analysis tended to show a larger percent decrease on the higher dictamnine diets than the A N C O V A or A N O V A analysis. For example, the RGRi index for the 160 ug/g FW treatment was 17% that of the control index value. A N C O V A analysis of the same final weight data found that larvae in this treatment had a mean final body weight 30% that of control larvae. In the dictamnine trial, the most obvious difference between the A N C O V A / A N O V A analysis and the ratio-based analysis is in the comparison between change in growth rates and changes in consumption due to the dietary treatments. In the A N C O V A / A N O V A analyses, the decreases, as a percent of control, in final weight and consumption for each treatment are not significantly different. However, from the ratio-based analysis, the RGRi is consistently lower than the R C R i , even if this difference is not significant for every treatment. For example, from non-ratio-based analysis, at the highest dictamnine concentration of 160 ug/g FW, both the final weight and consumption of the larvae have decreased by about 70% from that of larvae on the control diet (Table 3.3). However, the RGRi of the larvae on this treatment has decreased by more than 80% from that of the control larvae, while the consumption rate has only decreased by 65% from that of the control larvae (Table 3.7). From these ratio analysis results, one could propose that the decrease in growth of the larvae was due to more than just antifeedant properties of the compound. This is because the growth of the larvae decreased by more than did the consumption, possibly due to toxicity of the diet. This conclusion conflicts with the conclusions drawn from the non-ratio-based A N C O V A and A N O V A analyses. From those analyses, it appeared that antifeedant properties of the 50 diets, as reflected by decreases in consumption, were solely responsible for the decreases in final weight of the larvae. The A N C O V A / A N O V A analysis is the statistically more accurate method for the analysis of this particular data. The other trials, with skimmianine and 8-MOP, also exhibited lack of isometry when growth index numerator and denominators were plotted. Caution should be used when choosing ratio-based analytical methods for analysis of data. The other major problem encountered in the analysis of these growth trial experiments was the difficulty encountered in the comparison between trials, due to the necessity of combining the variability from the control trial and the relevant treatment when looking at error. As a result of this, the probability of finding significant differences between trials decreased. The solution that bypasses this difficulty is to conduct all trials at the same time, which may not be feasible if there are a large number of trials. Otherwise the solution is to increase the sample size of all treatments, especially that of the control, thereby decreasing treatment variance. Conclusions: Close scrutiny of the methods of analysis showed that the use of ratios for this data was invalid. The direct analysis of the variables final weight and consumption by A N C O V A and A N O V A techniques was more accurate. The use of ratio-based analysis was misleading in that it suggested toxic as well as antifeedant activity for the furoquinolines and furanocoumarin. 5 1 C H A P T E R 4 T e s t f o r S y n e r g y b e t w e e n F u r a n o c o u m a r i n s a n d F u r o q u i n o l i n e s i n t h e d i e t o f Trichoplusia ni ( L e p i d o p t e r a : N o c t u i d a e ) INTRODUCTION The co-occurrence of furanocoumarins and furoquinolines in Rutaceous plants raises the possibility that these compounds may act in synergy as toxins or antifeedant compounds. Synergy occurs when the net toxicity of a combination of compounds is greater than the sum of the toxicities of the individual compounds in the mixture. It is thought that it is the structure of a compound that determines whether it can potentially act as a synergist. Structural analogues may compete with toxins for binding sites on enzymes, thereby decreasing overall metabolic efficiency. In addition, irreversible binding of structural analogues to the enzyme can occur, rendering the enzyme non-functional (Berenbaum and Zangerl, 1996). Furoquinolines, as three-ringed planar molecules with a furan ring, are candidates as potential synergists to 8-MOP. Studies thus far have observed the synergy of the toxicity of one compound by another, but not the synergy of antifeedant properties (Berenbaum and Neal, 1985; Berenbaum etal., 1991). In the experiments presented in Chapter 3, both furoquinolines and 8-MOP were antifeedant to Spodoptera litura, but not physiologically toxic at the dietary concentrations tested. Testing for synergy requires that at least one of the compounds is toxic. Both compounds need not be toxic for analogue synergy to occur. Myristicin, neither antifeedant nor toxic to Helicoverpa zea, synergized the toxic effect of dietary 8-MOP (Berenbaum and Neal, 1985). As mentioned in Chapter 3, the lack of toxicity of 8-MOP to S. litura was unexpected, due to the plethora of literature reporting toxic effects 52 of the compound to generalist Lepidopterous larvae (Berenbaum and Neal, 1985; Berenbaum etal, 1991; Brewer etal., 1995). The concentration of 8-MOP may have been too low to cause detectable toxicity to the 5 t h instar larvae used. Therefore, in these synergy experiments neonate larvae were used rather than the larger 5 t h instar. The smaller larvae would be more vulnerable to potentially toxic effects of either 8-MOP or a furoquinoline. In this study, experiments with Spodoptera litura showed no toxicity even when the simulated daylight conditions were used. This preliminary experiment was performed under light conditions with no UV wavelengths. Prior to the synergy experiments, the Spodoptera litura colony, used for previous experiments in this project, was lost due to a virus. The cabbage looper, Trichoplusia ni, was used as a substitute generalist feeder. T. ni is a polyphagous species that feeds on a wide variety of host plants (Metcalf and Metcalf, 1993). This species was used to run a preliminary experiment to test for the synergistic reduction of larval growth by the combination of the furanocoumarin 8-MOP and furoquinoline skimmianine in the diet. M A T E R I A L S A N D M E T H O D S An initial assay looking for potential synergy of chemical mixtures was used to screen for results before more complex assays are designed. This simple procedure looks at the final weights of neonate larvae fed different diets for ten days. Larval consumption is not measured. Neonate growth assay Control or treatment diet cubes of equal size (about 1 gram) were placed in individual 29.5 mL (1 oz) plastic cups. The control treatment had a sample size of 40, and each experimental treatment had a sample size of 20. For 15 of the diet pieces of 53 each treatment (or 30 for the control), one neonate T. ni larva was placed in the cup. For the remaining five of the diet pieces (10 for control), 2 neonate larvae were placed in the cup. The cups were capped and placed in 16:8 hour light:dark condition using a cool white fluorescent bulbs, at 25\u00C2\u00B0C. After seven days the cups were checked. If any larvae were missing (due to small size and easy damage of neonates during set-up), then they were replaced with a duplicate larva. Extra duplicate larvae were removed. After 10 days, larval weights were recorded. Treatments: The following treatments were prepared in artificial diet, as outlined in Chapter 3: (a) three skimmianine diets, at the concentrations of 50, 100 and 150 ug/g FW; (b) three 8-MOP diets, at the concentrations of 100, 300 and 600 ug/g FW; and (c) three diets that combined 8-MOP and skimmianine in the following ratios (8-MOP:skimmianine): 100:50, 100:100, 300:50, and 300:100 ug/g FW. Data Analysis: The mean final weights of the larvae on each set of control or treatment diets were calculated. The means were compared using A N O V A with the multiple comparison Tukey test. Calculations were performed using the Systat 8.0 statistical program. 54 R E S U L T S A N D D I S C U S S I O N The mean final weights of the T. ni larvae, after 10 days of feeding on control or treatment diets, are presented in Figure 4.1. Vertical bars represent standard error values. 250 , , 5 6 7 Treatments jTreatments (ug/g FW) \" T i C o n t r o l \" \" skim skim skim 8-MOP 8-MOP 8-MOP s+8 ;2 u J5 :6 \7 \8 !9 50 100 150 100 300 600 100:50 s+8 110 s+8 111 s+8 100:100 300:50 300:100 Figure 4.1: M e a n final weights (+/- S E ) of Trichoplusia ni la rvae after feed ing for 10 d a y s on control or t reatment diets. The means of each treatment were compared using A N O V A with a multiple comparison Tukey test. All concentrations of skimmianine significantly reduced larval weight compared to those fed the control diet. The mean final weight of the larvae on the 50 ug/g FW diet was 53% lower than the mean final weight of the larvae on the control diet. The mean final weights of the larvae fed the 100 and 150 ug/g FW skimmianine diets were 87% and 94% lower than control larval weights, respectively. However, the difference between the final weights of the larvae on these two treatments 55 were not significant different. Since consumption was not recorded, this reduction in growth cannot be separated due to antifeedant versus toxic properties of the diet. However, the 150 ug/g FW diet did not cause a significantly greater decrease in growth than the 100 ug/g FW diet. This would suggest that skimmianine was only deterrent. The increase in concentration between these two diets would not make a difference to feeding if the larvae were at the lower limit of consumption to allow for survival. It appears skimmianine is a potent antifeedant, but not a strong enough negative stimulus to cause larval starvation. 8-MOP did not have a detrimental effect on the growth of the larvae. The larvae on treatments containing 100 and 300 ug/g FW 8-MOP did not differ in final weight from the control larvae, and the larvae on the 600 ug/g FW diet were significantly larger than the control larvae. When 8-MOP was combined with skimmianine and 8-MOP, no synergy occurred, and reduction in growth was due to the skimmianine. The diets which had 50 ug/g FW skimmianine and either 100 or 300 ug/g FW 8-MOP were not significantly different from each other or the treatment containing 50 ug/g FW skimmianine alone. Similarly, the treatments containing 100 ug/g FW skimmianine and either 100 or 300 ug/g FW 8-MOP were not significantly different from each other or the treatment containing 100 ug/g FW skimmianine alone. Since skimmianine appears to only have been antifeedant to the 7. ni larvae, it is not unexpected that no analogue synergy was observed. 8-MOP was neither antifeedant nor toxic to Trichoplusia ni up to the concentration of 600 ug/g FW. T. ni does include some Apiacous plants in its diet, including Pastinaca sativa L. (Zangerl, 1990), Petroselinium cripum and Apium gravolens (Metcalf and Metcalf, 1993). But it has been found previously that T. ni is negatively affected by furanocoumarins. Zangerl (1990) found that increased levels of 56 furanocoumarins in the leaves of Pastinaca sativa (wild parsnip) decreased the growth rate of T. ni larvae. However, the larvae were responding to the mixture of furanocoumarins found in P. sativa foliage: imperatorin, 5-MOP, 8-MOP, isopimpinellin and sphondin (Zangerl, 1990). In my experiment diets contained only 8-MOP and it is possible that 8-MOP is not one of the furanocoumarins that affects the growth rate of this species, either through a decrease in consumption or through toxic effects. T. ni has also been reported to feed on Citrus spp. (Lingren etal., 1993), which contain furanocoumarins but only one reported furoquinoline (Da Silva et al., 1988). Depending on the strain of Trichoplusia ni, this species could have the ability to deal with a certain amount of dietary furanocoumarins without difficulty. Another explanation for the lack of toxicity to T. ni by 8-MOP could be the experimental light conditions. The experiment was done under a cool white light source, which contains no UV wavelengths. Although S. litura showed no light dependant toxicity, and while dark toxicity of 8-MOP has been reported (Berenbaum and Neal, 1985), the majority of studies demonstrating 8-MOP toxicity have used simulated daylight conditions. In retrospect, this experiment should have been run under these conditions. Conclusions: The weight of T. ni larvae is significantly reduced as compared to control larvae after feeding on diets containing 50-150 ug/g FW skimmianine. It appears that solely antifeedant, not toxic, properties of the skimmianine containing diet are responsible for the reduction in growth. 8-MOP, in concentrations up to 600 ug/g FW, had no negative effect on T. ni growth. No synergy was observed for the combination of skimmianine and 8-MOP in the diet. 57 Experiments testing for synergy of furanocoumarins and furoquinolines combined in the diet are only informative if at least one of the compounds is toxic to the experimental larvae under the experimental conditions used. The furanocoumarin 8-MOP is toxic to generalist larvae under certain experimental conditions, and it would be prudent to ensure these conditions are met before further synergy experiments are undertaken. 58 SUMMARY AND CONCLUSIONS In this research project, furoquinoline alkaloids were studied from phytochemical and ecological perspectives. Overall, it provided procedural guidelines and data for the purification and quantification of two furoquinolines from Rutaceous plants, and presented the results of several experiments studying the putative role of furoquinolines in plant-lepidopteran ecological interactions. First, a simple and fast method for the isolation of two furoquinolines was designed. Dictamnine was isolated from the dried root of Dictamnus albus, which is available as a Chinese medicine. A quantity of 110 ug/g dry weight dictamnine was purified, this is a 40% yield of the total dictamnine concentration. This is an exceptionally efficient yield, probably due to the fewer steps taken during purification. Skimmianine was purified from the common shrub, Skimmia japonica; 530 ug/g dry weight was isolated, a 13% yield of the total skimmianine concentration. The bulk isolations were done on a mixture of male and female S. japonica leaves. Subsequently it was found that femaleS. japonica leaves have three times the concentration of skimmianine as male leaves, at least during the early spring (March). For future . isolations, it is recommended that the skimmianine concentration in leaf samples be measured before bulk leaf material is picked for extraction. The quantification of skimmianine and dictamnine, and the furancoumarins 5-MOP and 8-MOP in S. japonica and Ruta graveolens leaves yielded some intriguing results. Leaves sampled from R. graveolens and S. japonica did not contain dictamnine, although both have been reported to contain this compound (Mester, 1983). It is possible that dictamnine was present but in quantities below the detection level of the HPLC, or that the plants were not producing dictamnine. Another possibility is that there are fluxes in chemical composition of plants, over the course of the year, or depending 59 upon environmental conditions. Leaf samples from S. japonica were found to have different chemical compositions depending on the sex of the plant they were taken from. This too would be an avenue to investigate. The question is raised: does the chemical composition of male and female plants change over the course of the year, depending upon the reproductive phase of the plants? Feeding experiments were performed with larvae of the generalist lepidopteran, Spodoptera litura. It was found that both skimmianine and dictamnine were antifeedant compounds. Skimmianine was twice as potent as dictamnine; the same extent of reduction in larval growth and feeding was achieved with a skimmianine concentration one-half that of dictamnine. With all dietary concentrations of skimmianine or dictamnine, the larval consumption and growth decreased to the same degree. This indicates it is likely the furoquinolines were only antifeedant, not metabolically toxic. In addition, it was found that the presence of UV wavelengths from simulated daylight conditions did not affect the decrease in growth or consumption on either furoquinoline diet. The lack of effect of UV light, and the lack of metabolic toxicity of the furoquinolines is unexpected. However, the results from the trials with 8-MOP seem to indicate that the experimental conditions were such that phototoxicity was not elicited. Feeding experiments with 8-MOP found that the furanocoumarin was also only antifeedant and not metabolically toxic to the Spodoptera litura larvae. 8-MOP was not as potent a deterrent as either furoquinoline; a concentration of up to 8 times higher than that of skimmianine was required to produce the same percent decrease in growth and consumption. The presence of UV wavelengths also did not affect the deterrence of 8-MOP. However, previous studies have shown 8-MOP to be metabolically toxic to generalist lepidopterans, and that this toxicity is increased when UV light is present (Berenbaum et al., 1991). The possibility is that the experimental conditions were such that this phototoxicity was not elicited. This could result from the concentrations of 60 chemicals being too low, or because the incident UV energy reaching ingested chemicals was insufficient to produce harmful toxic results. Because of the non-observation of toxicity, photo-dependant or independent, with the 8-MOP trial, this suggests that further experiments should be done with the furoquinolines to determine if under certain conditions they too are phototoxic to lepidopterans. In addendum to the S. litura growth trial results, the AN OVA/AN C O V A method used for data analysis was compared to a ratio-based data analysis method. The ratio-based analysis was found to be unsuitable for the data, and when applied led to misleading conclusions about the growth trials. The comparison of the two methods and their derived results illustrated the importance of finding an accurate data analysis method. The final experiment, to test synergy of furoquinolines and furanocoumarins, used larvae of the generalist lepidopteran Trichoplusia ni. This was because during the course of the research time, the S. litura colony was lost to a virus. Diets of skimmianine alone, 8-MOP alone, and four diets combining the two chemicals were fed to T. ni neonate larvae. Skimmianine significantly reduced the final weight of T. ni at concentrations from 50 to 150 ug/g FW, but the pattern of final weight decrease suggests that only antifeedant, not toxic, properties of skimmianine caused the growth weight decrease. T. ni was not detrimentally affected by concentrations of 8-MOP up to 600 ug/g fresh weight. Indeed, the larvae on the treatment with the 600 ug/g FW diet grew to a larger final weight than larvae on the control treatment. The larvae on diets that combined skimmianine and 8-MOP exhibited a reduction in final weight that was the same as that resulting from a diet containing the same concentration of skimmianine alone. Synergy was not observed in this experiment, which was consistent with previous reports of synergy occuring only for the toxic properties of chemicals, not the antifeedant 61 properties. Thus if skimmianine was only antifeedant to T. ni. and 8 - M O P had no effect, then analogue synergy would not occur. It is necessary to ensure experimental conditions are such that at least one compound is toxic to experimental larvae in order to progress to synergy experiments. In conclusion, for the study of the furoquinolines in their role as protection against herbivores, only antifeedant properties have been demonstrated. However, the potential effect of UV light and potential toxic characteristics of the alkaloids should not yet be ruled out, and future research could focus on this area. 62 R E F E R E N C E S Ashwood-Smith, M.J., Towers, G.H.N., Abramowski, Z., Poulton, G.A. and Liu, M. 1982. \"Photobiological studies with dictamnine, a furoquinoline alkaloid\". Mutatation Research, 102. 401-412. Berdegue, M., White, K.K. and Trumble, J.T. 1997. \"Feeding deterrence of Spodoptera exigua (Lepidoptera:Noctuidae) larvae by low concentrations of linear furanocoumarins\". Physiological and Chemical Ecology, 26(4). 912-919. Berenbaum, M. 1981. \"Effects of linear furanocoumarins on an adapted specialist insect (Papilio polyxenes)\". Ecological Entomology, 6. 345-351. Berenbaum, M. 1983. \"Coumarins and caterpillars: a case for coevolution\". Evolution, 37(1). 163-179. Berenbaum, M.R. 1987. \"Charge of the light brigade: phototoxicity as a defense against insects\". In: Light Activated Pesticides, (Eds. Heitz, J.R. and Downum, K.R.). Washington, DC: American Chemical Society. 206-216. Berenbaum, M. and Feeny, P. 1981. \"Toxicity of angular furanocoumarins to swallowtail butterlies: escalation in a coevolutionary arms race?\". Science, 212. 927-929. Berenbaum, M. and Neal, J . J . 1985. \"Synergism between myristicin and xanthotoxin, a naturally cooccuring plant toxicant\". Journal of Chemical Ecology, 11(10). 1349-1358. Berenbaum, M.R., Nitao, J.K. and Zangerl, A.R. 1991. \"Adaptive significance of furanocoumarin diversity in Pastinaca sativa (Apiaceae)\". Journal of Chemical Ecology, 17(1). 207-215. Berenbaum, M.R. and Zangerl, A.R. 1996. \"Phytochemical diversity: adapatation or random variation?\". In: Phytochemical Diversity and Redundancy in Ecological Interactions. New York: Plenum Press. 1-24. Brattsten, L.B. 1982. \"Metabolic defenses against allelochemicals\". In: Herbivores: Their interactions with Secondary Plant Metabolites, Vol. 2, 2 n d ed. (Eds. Rosenthal, G.A. and Berenbaum, M.R). San Diego: Academic Press. 176-242. Brewer, M.J., Meade, T. and Trumble, J.T. 1995. \"Development of insecticide-resistant and -susceptible Spodoptera exigua (Lepidoptera: Noctuidae) exposed to furanocoumarins found in celery\". Physiological and Chemical Ecology, 24(2). 392-401. Bull, D.L., Ivie, G.W., Beier, R.C. and Pryor, N.W. 1986. \"In vitro metabolism of a linear furanocoumarin (8-MOP, xanthotoxin) by mixed-function oxidases of larvae of black swallowtail butterfly and fall armyworm\". Journal of Chemical Ecology, 12(4). 885-892. 63 Chase, M.W., Morton, C M . and Kallunki, J.A. 1999. \"Phylogenetic relationships of Rutaceae: a cladistic analysis of the subfamilies using evidence from rod . and afpB sequence variation\". American Journal of Botany, 86(8). 1191-1199. Clugston, D.M. and MacLean, D.B. 1965. \"Mass spectra of some furoquinoline alkaloids\". Canadian Journal of Chemistry. 43. 2516-2521. Cohen, M.B., Berenbaum, M.R. and Schuler, M.A. 1989. \"Induction of cytochrome P450-mediated detoxification of xanthotoxin in the black swallowtail\". Journal of Chemical Ecology. 15(9). 2347-2355. Cordell, G.A. 1981. Introduction to Alkaloids: A Biogenetic Approach. New York: John Wiley & Sons. Da Silva, M.F.D.G.F., Gottlieb, O.R. and Ehrendorfer, F. 1988. \"Chemosystematics of the Rutaceae: suggestions for a more natural taxonomy and evolutionary interpretation of the family\". Plant Systematics and Evolution, 161. 97-134. Diawara, M.M., Trumble, J.T., White, K.K., Carson, W . G . and Martinez, L.A. 1993. \"Toxicity of linear furanocoumarins to Spodoptera exigua: evidence for antagonistic interactions\". Journal of Chemical Ecology, 19(11). 2473-2484. Don-Pedro, K.N. 1996. \"Investigation of single and joint fumigant insecticidal action of citruspeel oil components\". Pesticide Science, 46. 79-84. Escoubas, P., Fukushi, Y., Lajide, L. and Mizutani, J . 1992. \"A new method for fast isolation of insect antifeedant compounds from complex mixtures\". Journal of Chemical Ecology. 18(10). 1819-1832. Escoubas, P., Lajide, L. and Mizutani, J . 1994. \"Insecticidal and antifeedant activities of plant compounds: potential leads for novel pesticides\". In: Natural and Engineered Pest Management Agents. (Eds. Hedin, P.A., Menn, J . J . and Hollingworth, R.M.). Washington, DC: American Chemical Society. 162-171. Farrar Jr., R.R., Barbour, J.D. and Kennedy, G .G . 1989. \"Quantifying food consumption and growth in insects\". Annals of the Entomological Society of America, 82(5). 593-598. Goloubkova, T.D., Heckler, E., Rates, S.M.K., Henriques, J .A .P. and Henriques, A T . 1998. \"Inhibition of cytochrome P450-dependent monooxygenases by an alkaloid fraction from Helietta apiculata markedly potentiate the hypnotic action of pentobarbital\". Journal of Ethnopharmacology, 60. 141-148. Gray, A. l . 1983. \"Structural diversity and distribution of coumarins and chromones in the Rutales\". In: Chemistry and Chemical Taxonomy of the Rutales: Annual Proceedings of the Phytochemical Society of Europe, Vol. 22. (Eds. Waterman, P.G. and Grundon, M.F.). London: Academic Press. 97-146. Harborne, J .B. 1982. Introduction to Ecological Biochemistry. 2 n d ed. London: Academic Press. 64 Hegnauer, R. 1983. \"Chemical characters and the classification of the Rutales\". In: Chemistry and Chemical Taxonomy of the Rutales: Annual Proceedings of the . Phvtochemical Society of Europe, Vol. 22. (Eds. Waterman, P.G. and Grundon, M.F.). London: Academic Press. 401-440. Houghton, P.J. and Raman, A. 1998. Laboratory Handbook for the Fractionation of Natural Extracts. London: Chapman & Hall. Ivie, G.W., Bull, D.L., Beier, R.C., Pryor, N.W. and Oertli, E.H. 1983. \"Metabolic detoxification: mechanism of insect resistance to plant psoralens\". Science, 221. 374-376. Ivie, G.W., Bull, D.L., Beier, R.C. and Pryor, N.W. 1986. \"Comparative metabolism of [3H]psoralen and [3H]isopsoralen by black swallowtail {Papilio polyxenes Fabr.) caterpillars\". Journal of Chemical Ecology, 12(4). 871-892. Jacobsen, M. 1989. \"Botanical pesticides: past, present and future\". In: Insecticides of Plant Origin, (Eds. Arnason, J.T., Philogene, B.J.R. and Morand, P.). Washington, DC: American Chemical Society. 1-10. Jones, C .G. and Firn, R.D. 1991. \"On the evolution of plant secondary chemical diversity\". Philosophical Transactions of the Royal Society of London. Series B, 333. 273-280. Kanamori, H., Sakamoto, I. and Mizuta, M. 1986. \"Further study on mutagenic furoquinoline alkaloids of Dictamni radicis cortex: isolation of skimmianine and high-performance liquid chromatic analysis\". Chemical and Pharmaceutical Bulletin, Tokyo, 34. 1826-1829. Klocke, J.A., Balandrin, M.F., Barnby, M.A. and Yamasaki, R.B. 1989. \"Limonoids, phenolics, and furanocoumarins as insect antifeedants, repellents, and growth inhibitory compounds\". In: Insecticides of Plant Origin, (Eds. Arnason, J.T., Philogene, B.J.R. and Morand, P). Washington, DC: American Chemical Society. 136-149. Lee, K. and Berenbaum, M.R. 1992. \"Ecological aspects of antioxidant exzymes and glutathione-S-transferases in three Papilio species\". Biochemical Systematics and Ecology, 20(3). 197-207. Lingren, P.D., Bryant, V .M. , Raulston, J.R., Pendleton, M., Westbrook, J . and Jones, G.D. 1993. \"Adult feeding host range and migratory activities of corn earworm, cabbage looper, and celery looper (Lepidoptera: Noctuidae) moths as evidenced by attached pollen\". Journal of Economic Entomololgy, 86(5). 1429-1439. Mester, I. 1983. \"Structural diversity and distribution of alkaloids in the Rutales\". In: Chemistry and Chemical Taxonomy of the Rutales: Annual Proceedings of the Phvtochemical Society of Europe, Vol. 22. ( Eds. Waterman, P.G. and Grundon, M.F.). London: Academic Press. 31-96. Metcalf, R.L. and Metcalf, R.A. 1993. Destructive and Useful Insects: Their Habits and Control, 5 t h ed. New York: McGraw-Hill. 65 Miksch, M. and Boland, W. 1996. \"Airborne methyl jasmonate stimulates the biosynthesis of furanocoumarins in the leaves of celery plants {Apium graveolens)\". Experientia, 52. 739-743. Miller, J .S . 1987. \"Host-plant relationships in the Papilionidae (Lepidoptera): Paralell cladogenesis or colonization?\". CJadistics, 3(2). 105-120. Mizuta, M. and Kanamori, H. 1985. \"Mutagenic activities of dictamnine and gamma-fagarine from Dictamni radicis cortex (Rutaceae)\". Mutatation Research, 144. 221-225. Murray, R.D.H., Mendez, J . and Brown, S.A. 1982. The Natural Coumarins: Occurence, Chemistry and Biochemistry. Toronto: John Wiley & Sons. Nawrot, J . , Koul, O., Isman, M.B. and Harmatha, J . 1991. \"Naturally occurring antifeedants: effects on two polyphagous lepidopterans\". Journal of Applied Entomology, 112. 194-201. Neal, J . J . and Wu, D. 1994. \"Inhibition of insect cytochromes p450 by furanocoumarins\". Pesticide Biochemistry and Physiology, 50(1). 43-50. Packard, G. and Boardman, T. 1988. \"The misuse of ratios, indices, and percentages in ecophysiological research\". Physiological Zoology, 61(1). 1-9. Pfyffer, G.E. , Panfil, I. and Towers, G.H.N. 1982a. \"Monofunctional covalent photobinding of dictamnine, a furoquinoline alkaloid, to DNA as target in vitro\". Photochemistry and Photobiology, 35. 63-68. Pfyffer, G.E. , Pfyffer, B.U. and Towers, G.H.N. 1982. \"Monoaddition of dictamnine to synthetic double-stranded polydeoxyribonucleotides in UVA and the effect of photomodified DNA on template activity\". Photochemistry and Photobiology, 35. 93-7. Pfyffer, G.E. and Towers, G.H.N. 1982. \"Photochemical interaction of dictamnine, a furoquinoline alkaloid, with fungal DNA in vitro and in vivo\". Canadian Journal of Microbiology, 28. 468-473. Pogue, M. 2000. Spodoptera (Lepidoptera) Larval Plant Host Database [online]. Washington, DC: Department of Entomology, National Museum of Natural History. Available: http://160.111.87.10.591/entomology/spodoptera/searchc.html. Raubenheimer, D. and Simpson, S .J . 1992. \"Analysis of covariance: an alternative to nutritional indices\". Entomologia Experimentalis Et Applicata, 62. 221-231. Stadler, E. \"Behavioural responses of insects to plant secondary compounds\" lp: Herbivores: Their Interactions with Secondary Plant Metabolites, Vol. 2. (Eds. Rosenthal, G.H. and Berenbaum, M.R.). San Diego: Academic Press. 45-88. 66 Stanjek, V., Piel, J . and Boland, W. 1999. \"Biosynthesis of furanocoumarins: mevalonate-independent prenylation of umbelliferone in Apium graveolens (Apiaceae)\". Phytochemistry, 50. 1141-1145. Tietz, H.M. 1972. An Index to the Described Life Histories, Early Stages and Hosts of the Macrolepidoptera of the Continental United States and Canada, Vol. 1-2. Florida: Allyn Museum of Entomology. Timmermann, S.E. , Zangerl, A.R. and Berenbaum, M.R. 1999. \"Ascorbic and uric acid responses to xanthotoxin ingestion in a generalist and a specialist caterpillar\". Archives of Insect Biochemistry and Physiology, 42. 26-36. Towers, G.H.N., Graham, E.A., Spenser, I.D. and Abramowski, Z. 1981. \"Phototoxic furanoquinolines of the Rutaceae\". Planta Medica, 41. 136-142. Veronese, F.M., Schiavon, O., Bevilacqua, R., Bordin, F. and Rodighiero, G. 1982. \"Photoinactivation of enzymes by linear and angular furocoumarins\". Photochemistry and Photobiology, 36. 25-30. Waldbauer, G.P. 1968. \"The consumption and utilization of food by insects\". Advances in Insect Physiology, 5. 229-288. Yajima, T. and Munakata, K. 1979. Phloroglucinol-type furanocoumarins, a group of potent naturally-occuring insect antifeedants\". Agricultural and Biological Chemistry. 43. 1701-1706. Zangerl, A.R. 1990. \"Furanocoumarin induction in wild parsnip: evidence for an induced defense against herbivores\". Ecology, 71(5). 1926-1932. Zar, J .H. 1984. Biostatistical Analysis, 2 n d ed. New Jersey: Prentice-Hall. Zare_bska, Z., Waszkowska, E., Caffieri, S. and Dall'Acqua, F. 1998. \"Photoreactions of psoralens with lecithins\". Journal of Photochemistry and Photobiology B, 45. 122-130. Zobel, A .M. and Brown, S.A. 1989. \"Histological location of furanocoumarins in Ruta graveolens shoots\". Canadian Journal of Botany, 67. 915-921. 67 A P P E N D I X 1 A 0 2 0 0 - 2 M S 5 0 Et l\"SCANT3RAPH : T i a g n I i i ' 9=Noi \u00C2\u00AB : M r / z . ' ' skimmianine =206. B a s e M/z=2-4-4.2. 1 0 0 % | n t . - 4 8465K*. 130 ,l.i\u00E2\u0080\u009E... .JJ., , Ill Il 180 t P i l e N a m e ; C r e a t i o n D a l e / T i m e C : \ M A S P E C \ D a l a \ L 4 9 8 7 2 . m s 2 fi/2/OO at 1*:4-4:57 AO^^OO-I \u00C2\u00B0 \" s y s t e m [II32/DD4Q] dictamnine , lnt.\u00C2\u00BB1821S7. Figure A1.1: M S S p e c t r a obta ined for purif ied c o m p o u n d s from Skimmia japonica and Dictamnus albus. Identified as A Sk immian ine B D ic tamnine , by match ing spec t ra f rom C lugs ton and M a c L e a n (1965) 68 APPENDIX 2 ; 250.3 nm skimmianine ) 331.1 nm :2 4 0 9 n n n dictamnine 312.0 nm ;03'.OO ' ' ISG.DO' - 3SO.C0 ' ' i s l o c ' V o i c e ' \u00E2\u0080\u00A2 Vso'.o'o' \u00E2\u0080\u00A2 s'ed.so' ' HS'.O'O ' ' ICfl.w ' ' iSO.OO * K5'.0'0 ' ' i ss ' .oo ' ' ica'.ec-' ' V5Cf.CS ' \u00E2\u0080\u00A2 s'oo'.oo' ' s s l o o 217.3 nm M 250.3 nm 8 -MOP 302.5 nm iOtf.CD ' M c b - ' ISO'.O'O ' ' ISO'.DO ' ' 400.C5 ' ' fStf.OC ' ' 553.00 ' * SSO'.CS ' 222.0 nm 5 -MOP 269.2 nm \u00E2\u0080\u00A2'V : 312.0 nm joo'.oo \u00E2\u0080\u00A2 iscf.o'o' \" j'ecUc ' ' iso ' .os ' ' 4'otf.oij' ' Vso'.ob' ' soo'.cb' 550*.oo 217.3 nm ;i 245.6 nm 302.5 nm FC1 222.0 nm 250.3 nm 312.0 nm F C 2 200.00 ' ' 250.CO ' 353.00' 3SC.00 *C0'.G0 ' ' 450.00 \" 503.00 550'.0 ;o'.co' ' isa.oo ios.oo ' J'SO'.OO ' ' Vco'.oc' ' t ' s c o o ' ' soo'.oo' \" w . s o Figure A 2 . 1 : Spec t ra of s o m e furoquinol ines and fu ranocoumar ins 69 APPENDIX 3 ANCOVA Analysis - Procedure (adapted from Packard and Boardman, 1988; Zar, 1984) Data from dictamnine light trial 1. Plot dependent variable (final weight) versus potential covariate (initial weight). 80 70 60 CD I S O \u00C2\u00A3 40 1 30 c 20 10 cont ro l 6 8 10 initial weight (dry) mg 12 14 Treatments (ug/g FW) O control \u00E2\u0080\u00A2 40 X 8 0 0 1 2 0 (no line) A 1 6 0 Figure A 3 . 1 : Char t of initial ve rsus final weight of S.litura la rvae on diets conta in ing sk immian ine . 2. Determine whether or not the treatment lines have significantly different slopes (Zar, 1984). For this trial, it was found that the slopes of the lines of the dictamnine treatments were not significantly different, with the exception of the 120 ug/g treatment, for which there wasn't a linear relationship between initial and final weight. 3. If H 0 is not rejected, or the slopes are not significantly different, then find the average slope of the regression lines. 70 4. Use the average slope to adjust each final weight value along the common regression line to a single x value, which is the value of the mean initial weight of all treatment samples included in the A N C O V A process. By adjusting the final weights along the average regression line, the variability among the samples decreases. For example, Figure A3 .2 shows the spread of the final weights before and after covariate adjustment. Final weight Adjusted final wt. Figure A3.2: T h e distr ibution of the final weights of the control t reatment before and after ad justment by the covar ia te of initial weight. 5. Find the mean values of the adjusted final weights of each treatment, and compare treatment mean values using A N O V A with multiple comparison Tukey tests. 71 "@en . "Thesis/Dissertation"@en . "2000-11"@en . "10.14288/1.0089597"@en . "eng"@en . "Botany"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "Furoquinolines of the Rutaceae and their role in plant-lepidopteran interactions"@en . "Text"@en . "http://hdl.handle.net/2429/10752"@en .