Host Plant Adaptation in the Two Strains of Fall Armyworm

Louisiana State UniversityLSU Digital Commons

LSU Historical Dissertations and Theses Graduate School

1994

Host Plant Adaptation in the Two Strains of FallArmyworm (Lepidoptera: Noctuidae).Klaas H. VeenstraLouisiana State University and Agricultural & Mechanical College

Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_disstheses

This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion inLSU Historical Dissertations and Theses by an authorized administrator of LSU Digital Commons. For more information, please [email protected].

Recommended CitationVeenstra, Klaas H., "Host Plant Adaptation in the Two Strains of Fall Armyworm (Lepidoptera: Noctuidae)." (1994). LSU HistoricalDissertations and Theses. 5764.https://digitalcommons.lsu.edu/gradschool_disstheses/5764

https://digitalcommons.lsu.edu?utm_source=digitalcommons.lsu.edu%2Fgradschool_disstheses%2F5764&utm_medium=PDF&utm_campaign=PDFCoverPages

https://digitalcommons.lsu.edu/gradschool_disstheses?utm_source=digitalcommons.lsu.edu%2Fgradschool_disstheses%2F5764&utm_medium=PDF&utm_campaign=PDFCoverPages

https://digitalcommons.lsu.edu/gradschool?utm_source=digitalcommons.lsu.edu%2Fgradschool_disstheses%2F5764&utm_medium=PDF&utm_campaign=PDFCoverPages

https://digitalcommons.lsu.edu/gradschool_disstheses?utm_source=digitalcommons.lsu.edu%2Fgradschool_disstheses%2F5764&utm_medium=PDF&utm_campaign=PDFCoverPages

https://digitalcommons.lsu.edu/gradschool_disstheses/5764?utm_source=digitalcommons.lsu.edu%2Fgradschool_disstheses%2F5764&utm_medium=PDF&utm_campaign=PDFCoverPages

mailto:[email protected]

INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

University Microfilms International A Bell & Howell Information C om p any

3 0 0 North Z eeb R oad. Ann Arbor. Ml 4 8 1 0 6 -1 3 4 6 USA 3 1 3 /7 6 1 -4 7 0 0 8 0 0 /5 2 1 -0 6 0 0

Order N um ber 9502151

Host plant adaptation in the two strains of fall armyworm (Lepidoptera: Noctuidae)

Veenstra, Klaas H., Ph.D.

The Louisiana State University and Agricultural and Mechanical Col., 1994

UMI300 N. ZeebRd.Ann Arbor, MI 48106

HOST PLANT ADAPTATION IN THE TWO STRAINS OF FALL ARMYWORM (LEPIDOPTERA: NOCTUIDAE)

A DissertationSubmitted to the Graduate Faculty of the

Louisiana State University and Agricultural and Mechanical College

in partial fulfillment of the requirements for the degree of

Doctor of Philosophyin

The Department of Entomology

byKlaas H. Veenstra

Doctoraal degree (combined B.S. and M.S.) Rijksuniversiteit at Leiden,

The Netherlands, 1987 May 1994

ACKNOWLEDGMENTS

I express my sincere gratitude to Dr. Dorothy P. Pashley, who served as my major advisor, and to Dr. JamesA. Ottea, who provided guidance in execution of the enzyme assays. I would also like to thank the other members of my committee, Drs. D. W. Foltz, J. Geagan, and A. M. Hammond, for their contributions to my education at L.S.U.

I am also much indebted to Julie A. Martin for strain identifications and fall armyworm collecting. Finally I would like to thank the Department of Entomology, through its present head, Dr. F. Guillot, and its former head, Dr. E. A. Heinrichs, for financial assistance.

ii

TABLE OF CONTENTSACKNOWLEDGMENTS..................................... iiABSTRACT............................................ ivCHAPTER

1 INTRODUCTION/LITERATURE REVIEW............... 1References to Chapter 1.................. 5

2 HOST PLANT CONSUMPTION AND UTILIZATIONIN TWO STRAINS OF FALL ARMYWORM ONCORN..................................... 10Introduction............................. 10Materials and Methods..................... 13Results.................................. 17Discussion............................... 25References to Chapter 2.................. 29

3 INDUCTION OF MIXED FUNCTION OXIDASE ANDGENERAL ESTERASE ACTIVITIES IN TWOSTRAINS OF FALL ARMYWORM.................. 32Introduction............................. 32Materials and Methods..................... 3 5Results...».............................. 43Discussion............................... 53References to Chapter 3................... 60

4 THE RELATIONSHIP BETWEEN HOST PLANTPERFORMANCE AND MIXED-FUNCTION OXIDASE ACTIVITY IN THE TWO STRAINS OF FALL ARMYWORM: RESULTS OF SELECTIONEXPERIMENTS.............................. 65Introduction............................. 65Materials and Methods..................... 69Results.................................. 76Discussion............................... 90References to Chapter 4............... 96

5 SUMMARY AND CONCLUSIONS..................... 101References to Chapter 5.................. 105

VITA................................................ 107

iii

ABSTRACT

The fall armyworm (Spodoptera fruoinerda J.E. Smith) Lepidoptera: Noctuidae) consists of two genetically differentiated strains. The corn strain feeds on corn and the rice strain feeds primarily on rice and forage grasses. My research focused on host plant adaptations within each strain as assessed by larval performance and enzymatic responses to feeding upon different hosts. Three sets of experiments were conducted. The first experiment examined consumption and utilization in both strains reared on corn. In the second experiment, mixed-function oxidase (MFO) and general esterase enzyme activities were measured from larvae of both strains that were reared on corn, bermudagrass, or artificial diet. In the third experiment, larvae of both strains were selected for increased larval weight, and associated changes in MFO activities were measured after subsequent feeding upon several hosts.

In the first experiment I determined that, when reared on corn, the corn strain had higher pupal weights and accumulated more biomass during the last instar than the rice strain. This resulted primarily from higher consumption in the corn strain. In addition, the corn

iv

strain was more efficient in converting digested food into biomass than the rice strain. Thus, there are differences in both feeding behavior and nutritional physiology between the strains feeding on corn.

The results from the second experiment indicated that the corn strain had a higher MFO activity than the rice strain when reared on corn or bermudagrass. This result was predicted since there are more reported allelochemicals in corn than in bermudagrass. Within the corn strain, MFO activity was higher when reared on corn than on bermudagrass. Thus, corn plants affected expression of MFO activity only in corn strain larvae. In the rice strain, MFO activity was similar on both hosts. Esterase activity did not differ between strains or host plants.

In the third experiment, lines selected on corn expressed higher MFO activity than either a line selected on bermudagrass, or a control. Differences in larval performance were minor. These results suggest there is genetic variation in MFO activity that could play a role in host plant adaptation.

v

CHAPTER 1 INTRODUCTION/LITERATURE REVIEW

The fall armyworm (Soodoptera fruaiperda;Lepidoptera: Noctuidae) has been recorded feeding on many plant species; however, it strongly prefers members of the grass family (Luginbill, 1928). It is considered a pest on corn, forage grasses, rice and sorghum (Luginbill, 1928; Sparks, 1979), and occasionally feeds on other crops of economic importance, like soybeans and cotton (Pitre et al., 1983). During outbreaks it can cause severe damage on its main hosts (Luginbill, 1928; Sparks, 1979). It is believed that the fall armyworm is not ablt to survive the temperate zone winters, and thus, migrates into temperate regions each year from overwintering areas in the tropics and subtropics (Luginbill, 1928; Johnson, 1987).

The fall armyworm consists of two genetically differentiated strains (Pashley 1986, 1988b), one feeding primarily on corn (the corn strain) and another feeding primarily on forage grasses and rice (the rice strain).The fact that no consistent morphological differences have been found (R.L. Brown, D.P. Pashley, pers. comm.) suggests that the strains are very closely related and probably of recent origin. Caterpillars of both strains can easily be reared on each other's host plants in the laboratory (Pashley, 1988a), and a small amount of overlap

1

in host plant usage occurs in the field (Pashley, 1986; D. P. Pashley; unpub. data). This feature makes the fall armyworm an ideal model for studying evolutionary aspects of host plant adaptation.

Theories on the evolution of host plant specialization differ most significantly from each other with respect to emphasis they place on physiological adaptations versus behavior. Some argue that behavior plays the central role in host plant shifts (Jermy, 1984; Bernays & Graham,1988). This view is supported by the fact that many plant allelochemicals are feeding repellents and are not intrinsically toxic (Usher & Feeny, 1983; Bernays & Chapman, 1987; Bernays & Cornelius, 1992). Another reason for an emphasis on behavior is that physiological adaptations towards a specific host plant can only evolve after that specific plant is accepted as a host (Futuyma, 1987).

From this point of view, populations are likely to evolve a preference for host plants that are either locally abundant (Smiley, 1978; Fox and Morrow, 1981), or on which they experience the lowest rates of predation and parasitism (Bernays & Graham, 1988). Experimental evidence in favor of this view comes from studies that detected genetic variation within populations in behavioral traits associated with host use (host plant preference,

3consumption rates), but not in physiological traits (Wasserman & Futuyma, 1981; Futuyma et al, 1984; Karowe, 1990 (but see Gould et al., 1982). Similar patterns have been observed between populations (Thomas et. al, 1987; Fry, 1988; Scriber et al, 1991).

Others view host plant adaptation mostly or solely in terms of physiological adaptations to host plants (Ehrlich & Raven, 1964; Ahmad et al, 1986; Caprio & Tabashnik,1992). They argue that behavior has evolved as a consequence of physiological adaptations. Some regard behavior to be fundamentally the same as physiology (Lockwood et al, 1984; Slansky, 1990). From this perspective, populations are likely to show preferences towards plants on which they are physiologically best adapted (Via, 1990; Thompson, 1988). This phenomenon is especially likely to occur when specialists perform better than generalists, because adaptation towards a specific host results in reduced fitness on other hosts (Via, 1990; Rausher, 1992).

The nature and extent of physiological host plant adaptations in the two fall armyworm strains are examined in Chapters 2 and 3. Differences between the strains in consumption rates (= behavior) and conversion of ingested and digested food (= physiology) are examined in Chapter 2. Corn was chosen as the host plant for this

experiment because strains differ most from each other in their performance on corn (Pashley, 1988a; Whitford et al., 1988).

During the last two decades much emphasis has been placed on enzymatic detoxication as the key to host plant adaptation (Krieger et al., 1971; Dowd et al., 1983;Ahmad, 1986; Lindroth ,1991). Forage grasses generally are thought to contain few active allelochemicals (Bernays & Barbehenn, 1987), whereas several antibiotic components have been identified from corn (Woodward et al. 1979; Elliger et al. 1980). Thus, the corn strain might be expected to have more active detoxication enzymes than the rice strain.

Results of a test of this hypothesis are presented in Chapter 3. Mixed function oxidase and general esterase enzyme activities were measured in larvae fed various diets. Both enzyme systems have been linked to enzymatic detoxification of allelochemicals in herbivorous insects (Dowd et al, 1983; Ahmad et al., 1986; Lindroth, 1991), although most evidence for the involvement of mixed function oxidase in host plant adaptation is circumstantial (Gould, 1984).

Chapter 4 deals with variation within strains. If, as argued above, behavioral traits are more important in host plant adaptation than physiological traits, genetic

variation for performance should be caused by genetic variation in behavioral traits (like consumption rates). In that case, selection for increased performance would not necessarily lead to physiological adaptations. This concept was tested by performing two artificial selection experiments, in which larvae were selected for increased larval weight on different host plants. After selection, growth rates and mixed-function oxidase activity of selected lines were compared. If mixed-function oxidase activity is crucial to host plant adaptation, selected lines should differ in mixed-function oxidase activity (given sufficient amounts of genetic variation in enzyme activity).

References to Chapter 1Ahmad, S., L. B. Brattsten, C. A. Mullin & S. J. Yu. 1986

Enzymes involved in the metabolism of plant allelochemicals. In: L. B. Brattsten & S. Ahmad [eds.] Molecular aspects of insect-plant associations. Plenum New York.

Bernays, E. A. & R. Barbehenn. 1987. Nutritional ecology of grass foliage-chewing insects. In: F. Slansky & J.G Rodriguez [eds.], Nutritional ecology of insects, mites, spiders and related invertebrates. Wiley-Interscience, New York.

Bernays, E. A. & R. F. Chapman. 1987. The evolution of deterrent responses in plant feeding insects. In: R. F Chapman, E. A. Bernays & J. G. Stoffolano Jr. [eds.], Perspectives in chemoreception and behavior. Springer-verlag, New York.

6Bernays, E. A. & M. Cornelius. 1992. Relationship between

deterrence and toxicity of plant secondary compounds for the alfalfa weevil Hvpera brunneipennis.Entomol. Exp. Appl. 64: 289-292.

Bernays, E. A. & M. Graham. 1988. On the evolution of host specificity in phytophagous arthropods.Ecology 69: 886-892.

Caprio, M. A. & B. E. Tabashnik. 1992. Evolution ofresistance to plant defensive chemicals in insects. In: B. D. Roitberg & M. B. Isman [eds.], Insect chemical ecology. An evolutionary approach. Chapman and Hall,New York.

Dowd, P. F., C. M. Smith & T. C. Sparks. 1983. Detoxification of plant toxins by insects.Insect Biochem. 13: 453-468.

Ehrlich, P. R. & P. H. Raven. 1964. Butterflies andplants: a study in coevolution. Evolution 18: 586-608.

Elliger, C. A., B. G. Chan, A. C. Waiss, Jr., R. E. Lundin & W. F. Haddon. 1980. C-glycosylflavones from Zea mays that inhibit insect development.Phytochemistry 19: 293-297.

Fox, L. R. & P. A. Morrow. 1981. Specialization: species property or local phenomenon. Science 211: 887-893.

Fry, J. D. 1988. Variation among populations of thetwospotted spider mite, Tetranvchus urtica Koch (Acari: Tetranychidae), in measures of fitness and host acceptance behavior on tomato.Environ. Entomol. 17: 287-292.

Futuyma, D. J. 1987. The role of behavior in host-associated divergence in herbivorous insects. In: M. Huettel [ed.], Evolutionary genetics of invertebrate behavior. Plenum, New York.

Futuyma, D. J., R. P. Cort & I. van Noordwijk. 1984. Adaptation to host plants in the fall cankerworm (Alsophila pometaria) and its bearing on the evolution of host affiliation in phytophagous insects.Am. Nat. 123: 287-296.

Gould, F. 1984. Mixed function oxidases and herbivore polyphagy: the devil's advocate position.Ecol. Entomol. 9: 29-34.

7Gould, F., C. R. Carroll & D. J. Futuyma. 1982. Cross

resistance to pesticides and plant defenses: a study of the two-spotted spider mite.Entomol. Exp. Appl. 31: 175-180.

Jermy, T. 1984. Evolution of insect-host plant relationships. Am. Nat. 124: 609-630.

Johnson, S. J. 1987. Migration and the life historystrategy of the fall armyworm, Spodootera frucriperda. in the western hemisphere.Insect. Sci. Applic. 8: 543-549.

Karowe, D. N. 1990. Predicting host range evolution: colonization of Coronilla varia by Colias philodice (Lepidoptera: Pieridae). Evolution 44: 1637-1647.

Krieger, R. I., P. P. Feeny & C. F. Wilkinson. 1971.Detoxification enzymes in the guts of caterpillars: an evolutionary answer to plant defenses?Science 172: 597-581.

Lindroth, R. L. 1991. Differential toxicity of plant allelochemicals to insects: roles of enzymatic detoxication systems. In: E. A. Bernays [ed.], Insect- plant interactions, volume 111. CRC Press, Boca Raton.

Lockwood, J. A., T. C. Sparks & R. N. Story. 1984. Evolution of insect resistance to insecticides: a reevaluation of the roles of physiology and behavior. Bull. Entomol. Soc. Am. 30: 41-51.

Luginbill, P. 1928. The fall armyworm. (U.S.D.A. Tech. Bull. 34). U.S. Depart. Agriculture, Washington D.C.

Pashley, D. P. 1986. Host associated geneticdifferentiation in the fall armyworm (Lepidoptera: Noctuidae): a sibling species complex?Ann. Entomol. Soc. Am. 79: 898-904.

Pashley, D. P. 1988a. Quantitative genetics, development, and physiological adaptation in host strains of fall armyworm. Evolution 42: 93-102.

Pashley, D. P. 1988b. Current status of fall armyworm host strains. Fla. Entomol. 71: 227-234.

8Pitre, H. N., J. E. Mulroony & D. B. Hogg. 1983. Fall

armyworm (Lepidoptera: Noctuidae) oviposition: crop preference and egg distribution on plants.J. Econ. Entomol. 76: 463-466.

Rausher, M. D. 1992. Natural selection and the evolution of plant-insect interactions. In: B. D. Roitberg & M.B. Isman [eds.], Insect chemical ecology. An evolutionary approach. Chapman and Hall, New York.

Scriber, J. M., J. Potter & K. Johnson. 1991. Lack ofphysiological improvement in performance of Callosamia promethea larvae on local host plant favorites. Oecologia 86: 232-235.

Slansky, F. Jr. 1990. Insect nutritional ecology as a basis for studying host plant resistance.Fla. Entomol. 73: 359-378.

Smiley, J. 1978. Plant chemistry and the evolution of host specificity: new evidence from Heliconius and Passiflora. Science 201: 745-747.

Sparks, A. N. 1979. A review of the biology of the fall armyworm. Fla. Entomol. 62: 82-87.

Thomas, C. D., D. Ng, M. C. Singer, J. L. B. Mallet, C. Parmesan & H. L. Billington. 1987. Incorporation of a european weed into the diet of a north american herbivore. Evolution 41: 892-901.

Thompson, J. N. 1988. Evolutionary ecology of the relationship between oviposition preference and performance of offspring in phytophagous insects. Entomol. Exp. Appl. 47: 3-14.

Usher, B. F. & P. Feeny. 1983. Atypical secondarycompounds in the family cruciferae: tests for toxicity to Pieris rapae. and adapted crucifer-feeding insect. Entomol. Exp. Appl. 34: 257-262.

Via, S. 1990. Ecological genetics and host adaptation in herbivorous insects: the experimental study of evolution in natural and agricultural systems.Annu. Rev. Entomol. 35: 421-446.

Wasserman, S.S. and D. J. Futuyma. 1981. Evolution of host plant utilization in laboratory populations of the southern cowpea weevil, Callosobruchus maculatus Fabricius (Coleptera: Bruchidae). Evolution 35: 605-617

9Whitford, F., S. S. Quisenberry, T. J. Riley & J. W. Lee.

1988. Oviposition preference, mating compatibility, and development of two fall armyworm strains.Fla. Entomol. 71: 234-243.

Woodward, M. D., L. J. Corcuera, H. K. Schnoes, J. P. Helgeson & C. D. Upper. 1979. Identification of 1,4-benzoxazin-3-ones in maize extracts by gas-liquid chromatography and mass spectrometry.Plant Physiol. 63: 9-13.

CHAPTER 2HOST PLANT CONSUMPTION AND UTILIZATION IN TWO STRAINS OF

FALL ARMYWORM ON CORN

IntroductionSince the insect-plant coevolution model was first

proposed by Ehrlich & Raven (1964), much emphasis has been placed on the role of host plant allelochemicals in insect-plant interactions. Many view the ability of insects to overcome toxic effects of host plant allelochemicals as the most important aspect of host plant adaptation (Ahmad et al, 1986; Lindroth 1991). Others argue, however, that factors such as natural enemies are more important in determining patterns of host use by phytophagous insects (Bernays & Graham, 1988). Support for the latter comes from the fact that many allelochemicals in plants have little or no intrinsic toxicity, but affect growth negatively by acting as feeding repellents (Usher & Feeny, 1983; Bernays & Graham, 1988; Bernays & Cornelius, 1992).

Most studies in Lepidoptera do not examine directly the consequences of feeding behavior on nutritional physiology (but see Berenbaum, 1986). This has been justified by arguing that behavior is not fundamentally different from physiology (Slansky, 1990; Lockwood et al., 1984).

10

However, most allelochemicals that reduce growth, reduce consumption as well (Bernays & Chapman, 1987) . Thus, it seems relevant to distinguish between actual toxicity and effects resulting from reduced consumption (Berenbaum 1986; Bernays & Cornelius, 1992). In this study, I examined differences in consumption (i.e. behavior) and differences in physiology independent of consumption between larvae of the two strains of fall armyworm, Spodoptera frucriperda J. E. Smith, feeding on corn.

The fall armyworm corn strain feeds primarily on corn, and the rice strain primarily on forage grasses and rice (Pashley, 1986, 1988a). Corn was chosen as a host for this study, because strains differ most from each other in their performance on corn (Pashley, 1988b; Whitford et al., 1988).

Forage grasses are thought to contain few active allelochemicals (Bernays & Barbehenn, 1987), but several antibiotic compounds have been identified from corn (Woodward et al. 1979; Elliger et al. 1980). As expected, the corn strain has a higher mixed function oxidase activity in the midgut than the rice strain (Chapter 3). Results of selection experiments suggest that this difference may be an adaptation to feeding on corn (Chapter 4). Thus, allelochemicals might be important in

12fall armyworm host plant adaptation because the corn strain is better able to metabolize ingested allelochemicals from corn.

However, the importance of detoxication enzymes does not rule out the importance of other physiological or behavioral processes in contributing to differences in host plant performance. For example, plant allelochemicals can exert antibiotic effects before absorption into the midgut by reducing the digestibility of the ingested food (Slansky, 1992). Houseman et. al. (1992) concluded that the allelochemical DIMBOA, which has been isolated from corn, exerts its antibiotic effects in this way on the European corn borer. In addition, allelochemicals may act as feeding repellents or stimulants. Either corn contains feeding repellents more effective toward the rice strain, or the rice strain is less stimulated to feed on corn than the corn strain.

In this chapter strains were compared with respect to consumption rates, the ability to digest corn, and efficiencies of converting ingested and digested corn into biomass. To accomplish this, consumption, frass production and biomass accumulation were measured during the last instar in larvae fed corn leaves.

13Materials and Methods

Experimental FlanFall armyworm larvae were collected on the L.S.U.

research farm at Ben Hur (East Baton Rouge Parish, LA) in August, 1992. Caterpillars were collected from a corn field and a nearby pasture (ca. 1/4 mile apart), in which bermudagrass was the dominant grass. After the field- collected individuals had laid eggs in the lab, the strain identity of each individual was determined by establishing their esterase phenotype (according to Pashley, 1986). Insects were reared for two more generations in the lab on artificial diet prior to the experiments. Inbreeding was avoided by recording ancestry of mating pairs. Corn plants (Pioneer, hybrid # 3165) were grown in a greenhouse. All insect rearing took place in an environmental chamber at 25 "C and a 14:10 (L:D) photoperiod. Forty larvae of each strain were reared individually in containers (120 ml) in which the bottom was covered with 2% agar to avoid desiccation. Corn leaves were replaced on alternate days. Larvae were checked every 12 hours after day 8. Larvae that had started to shed their head capsules were weighed and transferred to containers with pre-measured weights of corn leaves.

Consumption was measured as the amount of corn added minus the amount remaining (on dry weight basis). The

14(approximate) amount digested was determined as consumption minus the amount of frass produced (dry weight). The duration of the last instar was determined as the number of days spent in the larval stage after the onset of the last molt. The prepupal stage, when larvae cease feeding but have not yet pupated, was considered to be part of the last instar. Approximately 12 hours after pupation, pupae were sexed, frozen, dried and weighed. Accumulated weight was estimated as the dry weight of the pupa minus the estimated dry weight of larvae during the last molt. To estimate the amount of dry weight in molting larvae, 8 rice strain larvae and 13 corn strain larvae (all randomly selected) were sacrificed. One estimate of dry weight was used for both strains since they did not differ.

Measurements of consumption and utilization were taken for 2 3 rice strain larvae and 21 corn strain larvae. These measurements were the basis for calculating the following parameters (modified after Waldbauer, 1968):

Relative consumption (RC) = consumption relative to the duration of the last instar

Approximate digestibility (AD) = the amount of food digested relative to consumption

15Efficiency of converting digested food (ECD) = the

efficiency of converting digested food into biomass

Efficiency of converting ingested food (ECI) = theefficiency of converting ingested food into biomass

Data AnalysisStrains were compared for relative consumption by an

ANOVA with strain, sex, and instar duration (and interactions) as sources. The potential effect of larval weight on consumption was initially examined as the relationship between weight of the larvae at the beginning of the last instar and consumption (see Farrar et al.(1989) for why the weight at the beginning of the instar was used, rather than an average weight during the instar). However, no effects of weight at the beginning of the last instar on consumption were detected. Therefore, larval weight was omitted from the analysis of consumption. To assess whether strains differ in their ability to digest corn (AD), the amount digested was regressed on consumption. By conducting this type of covariance analysis (ANCOVA: Neter et al., 1989), a possible correlation between the percentage digested and consumption (= covariate) is avoided (Packard & Boardman, 1987). If, for example, consumption increases, then food

16passes through the gut faster, potentially decreasing the proportion of ingested food that is digested (Slansky & Scriber, 1985). For similar reasons, ECD and ECI were analyzed using an ANCOVA as well. For strain comparisons in ECD and ECI, the numerator of Waldbauer's original index (accumulated biomass) is regressed on its denominator (= covariate: amount digested in ECD, consumption in ECI). In the actual ANCOVA, regression lines of the different treatments (strain and sex) were compared. Significant treatment by covariate interactions indicated differences in slopes of regression lines, whereas significant main effects indicate differences in intercepts.

All models were tested in PROC GLM of SAS (SAS Institute, 1985). Interaction terms with P values larger than 0.3 were removed from the models. The linearity of regression lines was tested by adding quadratic terms of the covariates to the model. Because significant quadratic terms were never obtained, they are not included in the final models. Normality of residuals was tested in PROC UNIVARIATE of SAS. Equality of variance in residuals between strains was tested with a two-tailed F-test (in PROC TTEST of SAS). Because inequality of variance was detected, R2-values were calculated as well for each

17ANCOVA model, using either corn or rice strain only. All means are reported with their standard errors.

ResultsOverall, corn strain larvae performed better on corn

than rice strain larvae (Table 2.1). Larval mortality in both strains was low (corn strain = 10 %, rice strain = 15 %). During the last instar, the corn strain consumed more on average (380 ± 6.0 mg) than the rice strain (337 ± 6.3 mg; F = 24.9 df = 1,41; P < 0.001). Males and females of each strain did not differ significantly in total consumption (F = 1.57; df = 1,41; P < 0.21).

To determine whether duration of the last instar influenced differences in consumption between strains (RC), the above data were reanalyzed by taking last instar duration into account (Fig. 2.1). Larvae of the corn strain consumed more than the rice strain, irrespective of instar duration. Only observations with a last instar duration of 5.0 and 5.5 days were included in the ANOVA to obtain a more balanced design. This ANOVA contained a highly significant strain effect (F = 21.9; df = 1,33; P < 0.001) indicating that corn strain larvae consumed more on a daily basis. There were no strain differences in the duration of the last instar (F = 0.58; df = 1,40;

Table 2.1. Mean developmental variables of the two fall armyworm strains on corn.

Variable: Rice strain Corn strain F df PWeight at day 8 (mg) 38 . 3 (± 3.9) 54 . 4 (± 3.7) 8.93 1,72 0. 004Weight at last molt 146 (± 5.4) 156 (± 5.1) 2.10 1,45 0.15Larval development time3 16.0 (± 0.2) 15. 6 (± 0.2) 1. 69 1,41 0.20Pupal weight6 51.3 (± 1.2) 57.4 (± 1.2) 12.8 1,41 0.0009

a Sex effect: P < 0.92 (F = 0.01). b Sex effect: P < 0.17 (F = 1.96).

4 5 0 i

4 0 0

3 5 0

3 0 0

2 5 0

CONSUMPTION (MG)

2003 .5

A

-t-**

41AA

'"i

AAzNtAA-,.A-f-/\L_\A

A

4 .5 5 5 .5 6L A S T INSTAR D U R A T IO N (DAYS)

/ \

6.5

Fig. 2.1. Consumption in the two fall armyworm strains as a function of the duration of the last instar. Rice strain: a ; corn strain: *.

vo

20P < 0.45). Thus, the corn strain consumed more than the rice strain, in both total amount and on a daily basis.

To determine whether the two strains differed in their ability to digest corn (AD), the amount digested was regressed on consumption (Fig. 2.2). Strains did not differ significantly (Table 2.2), but the corn strain was more variable in the amount it digested than the rice strain (Table 2.3). The average total amounts digested were: corn strain = 120.7 ± 3.4 mg, rice strain = 129.8 ±3.5 mg.

Corn strain larvae accumulated significantly more biomass during the last instar (40.2 ± 0.97 mg, compared to 35.0 ± 0.93 mg for the rice strain; F = 15.0; df =1,41; P < 0.0004). Sex effects were not significant (F = 2.21, df = 1,41, P < 0.15).

Strains were compared in their efficiency of converting digested food into biomass (ECD) by regressing the accumulated biomass on the amount of digested food (Fig. 2.3). The analysis of covariance resulted in significant strain effects and strain-by-digestion interaction (Table 2.2). Because the strain-by-digestion interaction is significant, the difference in efficiency between the strains depended upon the amount digested. At high amounts of digested food, the difference in

170

1 5 0

1 30

1 10

9 0

DIGESTION (MG)

7 0

* *

i i i i i i r r—

2 6 0 2 8 0 3 0 0 3 2 0 3 4 0 3 6 0 3 8 0 4 0 0 4 2 0C O N S U M P T I O N (MG)

4 4 0

Fig. 2.2. Amount of corn digested in the two fall armyworm strains as a function ofconsumption. Rice strain: a , solid line; corn strain: *, broken line.

Table 2.2. Results of the analysis of covariance (ANCOVA) on the data plotted in Fig. 2.2, 2.3 & 2.4.

Analysis:

Effect:

> o 0) ECDb ECIa

F P F P F PStrain 1. 78 0.19 6.67 0. 013 0.09 0. 77Sex 0. 03 0.86 1. 25d 0. 27d 11.2 0. 002CovariatecCovariate*strainCovariate*sex

27 . 0 0.0001 21.7 4.72 2 .17d

0.0001 0.036 0.15d

43 . 8 0. 0001

a df for all effects: 1,40 b df for all effects: 1,38 c Either consumption or digestiond If sex*digestion removed from model: F= 6.51, P < 0.015

tot o

Table 2.3. Comparisons of variability in response between the two fall armyworm strains. Tests of equality of variance in residuals from the ANCOVA models in Table 2.2, and R2 and F-values of each model, when only one strain was used.

R2-values (F-value model)Analysis F (V / V - ) a' corn7 rice7 P Full model Rice CornAD 2 . 7 7 0 . 0 2 0 . 4 9 ( 1 1 . 3 ) 0 . 63 ( 1 6 . 3 ) 0 . 2 8 ( 3 . 4 4 )ECD 1 . 5 4 0 . 1 6 0 . 5 6 ( 1 2 . 4 ) 0 . 5 5 ( 1 2 . 2 ) 0 . 2 1 ( 2 . 4 2 )ECI 1 . 09 0 . 8 5 0 . 66 ( 2 6 . 3 ) 0 . 5 8 ( 1 4 . 1 ) 0 . 5 3 ( 1 0 . 1 )

a df = 20,22

t ow

ACCUMULATED WEIGHT (MG)

4 5 -

3 5 -

3 0 -

2 5 T8 5 9 5 105 115 125 135 145 155

A M O U N T D I G E S T E D (MG)

Fig. 2.3. Efficiency of converting digested food into biomass in the two fallarmyworm strains. Biomass accumulation as a function of amount digested.Rice strain: a , solid line; corn strain: *, broken line.

25efficiency is smaller than at low amounts digested (see Fig. 2.3). At the average amount digested by the corn strain (which is larger than the average for the rice strain), the corn strain is significantly more efficient (t = 2.58; df = 39; P < 0.014, two-tailed). Females were more efficient than males in converting digested food into biomass (Table 2.2).

Strains were also compared in their efficiency of converting ingested food into biomass (ECI) by regressing the accumulated biomass on the amount consumed (Fig. 2.4). Strains did not differ significantly in the conversion of ingested food (Table 2.2), but females were more efficient than males in converting ingested food into biomass.

DiscussionOn average, the corn strain accumulated 15% more

biomass than the rice strain and consumed 13% more corn leaves than the rice strain. Similar efficiencies in the conversion of ingested food indicate the importance of the reduced consumption in the rice strain and explained in a large part, its poorer performance on corn. However, the difference in efficiency of converting digested food into biomass indicates that the corn strain is better able to utilize corn as a host due to physiological differences as well.

50

4 5

4 0

3 5

3 0

ACCUMULATED WEIGHT (MG)

2 5

A A

2 7 0 2 9 0 310 3 3 0 3 5 0 3 7 0 3 9 0 410C O N S U M P T I O N (MG)

4 3 0

Fig. 2.4. Efficiency of converting ingested food into biomass in the two fallarmyworm strains. Biomass accumulation as a function of consumption.Rice strain: &, solid line; corn strain: *, broken line.

The corn strain digests, on average, 7.5% more than the rice strain, but this is because it consumes more as well. When differences in consumption between strains are taken into account, strains digest similar amounts of ingested food. This indicates that interference of host plant allelochemicals in conversion of ingested food into digested food does not seem to be important in causing strain differences on corn. Differences between the strains in utilization of corn are strictly due to conversion efficiencies of food after digestion.

It is not clear why the corn strain is more variable than the rice strain in digesting consumed corn. However, this finding suggests physiological differences between the strains. Conversion of ingested food into biomass can be viewed as a two step process involving digestion and the conversion of digested food into biomass (Slansky & Scriber, 1985). Consequently, the difference in variability of digestion should result in a difference in variance in the conversion of ingested food. However, this was not the case (Table 2.3). The apparent reason for this is that when in the corn strain the amount of food digested is lower than expected from consumption, the efficiency of converting the digested food is increased to compensate for this. As a result, the slope in the conversion of digested food is less steep for the corn

28strain than for the rice strain (Fig. 2.3; hence the significant digestion by strain interaction).

Differences between the sexes in efficiencies of conversion that are reported here for fall armyworm, have been reported for other Lepidoptera as well (Slansky & Scriber, 1985). Sexes are known to differ with respect to protein and lipid content of pupae (Lederhouse et al., 1982; Slansky & Scriber, 1985). Thus, the sex differences likely reflect qualitative differences between the sexes in accumulated biomass.

In conclusion, there are differences in feeding behavior and nutritional physiology between the strains when feeding on corn. Behavioral differences exist with respect to consumption and could be mediated through differential responses to feeding deterrents or stimulants. The physiological differences are caused by different efficiencies of converting digested food into biomass. The difference in conversion of digested food is consistent with expectations based on knowledge of the mixed-function oxidase activity in the midgut. Increased detoxication of allelochemicals in the midgut could negate the impact of corn allelochemicals on the conversion of digested food into biomass. In summary, both physiological and behavioral differences are important components influencing the difference between the strains in their performance on corn.

29References to Chapter 2

Ahmad, S., L. B. Brattsten, C. A. Mullin & S. J. Yu. 1986. Enzymes involved in the metabolism of plant allelochemicals. In: L. B. Brattsten & S. Ahmad [eds.], Molecular aspects of insect-plant associations. Plenum, New York.

Berenbaum, M. 1986. Postingestive effects ofphytochemicals on insects: on Paracelsus and plant products. In: J. R. Miller & T. A. Miller [eds.], Insect-plant interactions, Springer-verlag, New York.

Bernays, E. A. & R. Barbehenn. 1987. Nutritional ecology of grass foliage-chewing insects. In: F. Slansky &J. G. Rodriguez [eds.] Nutritional ecology of insects, mites, spiders and related invertebrates. Wiley-Interscience, New York.

Bernays, E. a. & R. F. Chapman. 1987. Evolution ofdeterrent responses by phytophagous insects. In: R. F. Chapman, E. A. Bernays & J. G. Stoffolano [eds.] Perspectives in chemoreception and behavior. Springer-Verlag. New York.

Bernays, E. A. & M. Cornelius. 1992. Relationship between deterrence and toxicity of plant secondary compounds for the alfalfa weevil Hypera brunneipennis.Entomol. Exp. Appl. 64: 289-292.

Bernays, E. & M. Graham. 1988. On the evolution of host specificity in phytophagous arthropods.Ecology 69: 886-892.



Farrar, R. R., Jr., J. D. Barbour & G. G. Kennedy. 1989. Quantifying food consumption and growth in insects.Ann. Entomol. Soc. Am. 82: 593-598.

30Houseman, J. G. , F. Campos, N. M. R. Thie, B. J. R.

philogene, J. Atkinson, P. Morand & J. T. Arnason.1992. Effects of maize-derived compounds DIMBOA and MBOA on growth and digestive processes of European corn borer (Lepidoptera: Pyralidae)J. Econ. Entomol. 85: 669-674.

Lederhouse, R. C., R. C. Finke, M. D. Scriber. 1982.contributions of larval growth and pupal duration to protandry in the black swallowtail butterfly, Papilio polvxenes. Oecologia 53: 296-300.


Lockwood, J. A., T. C. Sparks & R. N. Story. 1984. Evolution of insect resistance to insecticides: a reevaluation of the roles of physiology and behavior. Bull. Entomol. Soc. Am. 30: 41-51.

Neter, J., W. Wasserman & M. H. Kutner. 1989. Applied linear regression models, Irwin, Homewood IL.

Packard, G. C. & T. J. Boardman. 1987. The misuse of ratios to scale physiological data that vary allometrically with body size. In: M. E. Feder & A. F. Bennet [eds.], New directions in ecological physiology, Cambridge Univ. Press, Cambridge, U.K.

Pashley, D. P. 1986. Host-associated geneticdifferentiation in fall armyworm (Lepidoptera: Noctuidae): a sibling species complex?Ann. Entomol. Soc. Am. 78: 756-762.

Pashley, D. P. 1988a. Current status of fall armyworm host strains. Fla. Entomol. 71: 227-234.

Pashley, D. P. 1988b. Quantitative genetics, development, and physiological adaptation in host strains of fall armyworm. Evolution 42: 93-102.

SAS Institute. 1985. SAS user's guide: Statistics, 5th ed. SAS Institute, Cary, NC.

Slansky, F. Jr. 1990. Insect nutritional ecology as a basis for studying host plant resistance.Fla. Entomol. 73: 359-378.

31Slansky, F. Jr. 1992. Allelochemical-nutrient interactions

in herbivore nutritional ecology, In: G. A. Rosenthal & M. R. Berenbaum [eds.], Herbivores: their interaction with secondary plant metabolites, 2E. Volume 11: Evolutionary and ecological processes, Academic Press, San Diego.

Slansky, F., Jr. & J. M. Scriber. 1985. Food consumption and utilization, In: G. Kerkut & L. I. Gilbert [eds.], Comprehensive insect Physiology, biochemistry and pharmacology, Volume 4, Permagon Press, New York.

Usher, B. F. & P. Feeny. 1983. Atypical secondarycompounds in the family cruciferae: tests for toxicity to Pieris rapae. and adapted crucifer-feeding insects. Entomol. Exp. Appl. 34: 257-262.

Waldbauer, G. P. 1968. The consumption and utilization offood by insects. Adv. Insect Physiol. 5: 229-573.

Whitford, F., S. S. Quisenberry, T. J. Riley & J. W. Lee. 1988. Oviposition preference, mating compatibility, and development of two fall armyworm strains.Fla. Entomol. 71: 234-243.

Woodward, M. D., L. J. Corcuera, H. K. Schnoes, J. P.Helgeson & C. D. Upper. 1979. Identification of1,4-benzoxazin-3-ones in maize extracts by gas-liquid chromatography and mass spectrometry.Plant Physiol. 63: 9-13.

CHAPTER 3INDUCTION OF MIXED-FUNCTION OXIDASE AND GENERAL ESTERASE

ACTIVITIES IN TWO STRAINS OF FALL ARMYWORM

IntroductionMany plants contain chemicals that are thought to have

evolved as defensive mechanisms against herbivores (Berenbaum & Zangerl, 1992; but see Bernays & Graham(1988) for a different view). Insect herbivores, on the other hand, possess a number of enzyme systems that can metabolize potentially toxic compounds. It is assumed by many that the ability of an insect to successfully utilize a specific plant as a host often depends upon its detoxication enzymes (Ahmad et al., 1986; Lindroth, 1991).A problem faced by many herbivorous insects is variable

allelochemical content among and within host species.There are two strategies to deal with this variability.One strategy is to produce amounts of detoxication enzymes high enough at all times to deal with the maximum amount of allelochemicals potentially encountered. It has been argued, though, that detoxication enzymes are energetically costly (Shoonhoven & Meerman, 1978; Brattsten, 1979, but see Neal (1978) and Appel & Martin (1992)), and insects that are well adapted to a specific host plant are less adapted to other hosts because

32

valuable resources are allocated to the production of enzymes that offer minimal or no benefits (Dykhuizen,1978). A second strategy to deal with allelochemical variability is to produce detoxication enzymes in quantities that are proportional to the amount of host plant allelochemicals encountered. The fact that detoxication enzyme activity can, in many cases, be induced when allelochemicals are added to the diet isoften taken as evidence that this is a common strategy inherbivorous insects (Gould, 1984). Thus, enzyme induction can be seen as an adaptive form of phenotypic plasticity that allows herbivorous insects to feed upon plants that vary widely in allelochemical content.

However, if an insect is constantly exposed to inducingallelochemicals over evolutionary time, a canalization of the response could evolve (Weir, 1992). The reason canalization is likely to evolve is because genotypes that are less variable near the optimum phenotype are favored by stabilizing selection relative to more variable genotypes (Schmalhausen, 1949; Weir, 1992). If the response becomes canalized, enzyme activity is no longer a function of the amount of allelochemicals encountered.

The two strains of the fall armyworm (Soodoptera frugioerda; Lepidoptera: Noctuidae) provide an interesting model system with which to study enzymatic adaptation to

34host plants. The fall armyworm has been recorded feeding on plants from more than 80 plant species in 23 families, but it strongly prefers members of the grass family (Pashley, 1986). The two fall armyworm strains, however, show marked differences in host use. The corn strain feeds primarily on corn, and the rice strain feeds primarily on forage grasses and rice (Pashley, 1986 & 1988a). Strains tend to perform better on their preferred host plants (i.e. have higher larval and pupal weights, and shorter larval development time). The largest difference in performance between strains occurs on corn (Pashley,1988b; Whitford et al. 1988? this Chapter).

Forage grasses are believed to contain few active allelochemicals (Bernays & Barbehenn, 1987), whereas several antibiotic compounds have been identified from corn (Woodward et al. 1979; Elliger et al. 1980). We would expect that, if these chemicals are effective against herbivorous insects, either the corn strain would have higher overall activities of detoxication enzymes than the rice strain, or that the corn strain would have a higher enzyme activity on corn than on forage grasses due to enzyme induction.

To test these hypotheses and to examine differences in induction of enzyme activity between strains, mixed- function oxidase (MFO) and general esterase enzyme

activities were measured. The importance of these enzyme systems in insect performance on host plants has been established in other herbivorous insects (Ahmad et al., 1986; Lindroth, 1991). In this study, enzyme activities were measured towards model substrates using midgut homogenates from last instar larvae reared on artificial diet, bermudagrass, and corn. In addition, esterases were examined in more detail by slab gel electrophoresis. Results from these studies suggest that MFO, in contrast to general esterases, play an important role in host plant adaptation.

Materials and MethodsChemicals

g-Nitrophenol (PNP; spectrophotometric grade), NADPH, a-naphthol acetate, a-naphthol (99+ %), and Fast Blue B salt (90% dye content) were obtained from Sigma Chemical Co. (St. Louis, MO). p-Nitroanisole (PNA) was obtained from Aldrich Chemical Co. (Milwaukee, WI). Dichloromethane (spectrophotometric grade) was obtained from EM Industries Inc. (Gibbstown, NY). All other chemicals (highest possible purity) were purchased from commercial suppliers.

36Insects

Fall armyworm larvae were collected on the Louisiana Agriculture Experimental Station research farm at Ben Hur (East Baton Rouge Parish, LA) in August, 1992.Caterpillars were collected from a corn field and a nearby pasture (approximately 1/4 mile apart) in which bermudagrass was the dominant grass. After the field- collected individuals had laid eggs in the lab, the strain identity of each individual was determined by establishing their esterase phenotype (according to Pashley, 1986). The offspring from the field-collected insects were reared for another generation on artificial diet prior to experiments. Inbreeding in the lab was avoided by keeping track of the ancestry of the mating pairs.

General esterase activity was determined in both strains and hybrids of the two strains, whereas MFO activity was determined in both strains. The same insect homogenates used for esterase assays were also used for gel electrophoresis. Insects used for the MFO assays had different parents than the insects used for esterase assays (i.e. MFO activity was determined in a separate experiment).

Enzyme assays were performed with midgut tissues from last instar larvae reared on either artificial diet, corn, or bermudagrass. All rearing took place in an

environmental chamber at 25°C and on a 14:10 (L:D) photoperiod. Larvae were examined for molting every 12 hours after day 8 of development. Molting larvae were weighed and separated from the remaining larvae. The modified pinto bean diet used was the same as that described by Quisenberry & Whitford (1988). Larvae reared on the diet were kept in 3 0 ml containers with approximately 1 cm of diet. All plants were grown in a greenhouse. Bermudagrass was grown from seeds. Corn plants (Pioneer, hybrid # 3165) were grown and used until the first signs of tasseling. Larvae reared on plant material were kept individually in 120 ml containers in which the bottom had been covered with a 2% agar solution to avoid desiccation. Every other day, the plant material was replaced.

For MFO assays, 2 4 larvae of each treatment combination (diet and strain) were reared originally to be used for actual assays. All these larvae were weighed on day 8 of larval development, and half of the larvae (randomly chosen) were kept for use in the MFO assays.

Enzyme AssaysThe activity of MFO was estimated by measuring the

enzymatic O-demethylation of PNA to form PNP by the method of Hansen & Hodgson (1971) as modified by Neal & Berenbaum

38(1989) . Actively-feeding last instar larvae (316 - 555 mg) that had at least doubled their weight since the last molt were used. Larvae were weighed, chilled on ice, and then dissected under a stereo microscope. Trachea and Malpighian tubules were separated from the midgut as much as possible, the midgut was cut open longitudinally, and the contents were removed. The midgut was rinsed twice with water, cut out of the insect, and homogenized in 0.7 ml of ice-cold buffer (HEPES, pH 7.8, 0.1 M) in an allglass homogenizer. The homogenate was centrifuged at 12,000g at 0°C for 12 minutes. From the supernatant, 0.445 ml was used for one MFO assay. The amount of protein used in the reaction mixture ranged from 0.12 to 0.41 mg. Protein concentrations were estimated by the method of Bradford (1976) with bovine serum albumin as standard.From a NADPH generating system, 50 /nl was added to the reaction mixture, so that the reaction mixture contained the following reagents (final concentrations in a final volume of 0.5 ml): HEPES buffer (pH 7.8, 0.1 M), MgCl2 (7.5 mM), NADP (5 mM), glucose 6-phosphate (25 mM) and 0.4 units of glucose 6-phosphate dehydrogenase. The reaction mixture was pre-incubated in a water bath (30°C) for 2 minutes and the reaction was initiated by addition of 5 /Lil of PNA (in ethanol, final concentration 1 mM). After 10 minutes, the reaction was stopped by adding 0.125 ml of 1

N HC1. Pilot studies established that, under these conditions, product formation was linear with respect to tissue equivalents and incubation time used (data not shown). Product was extracted from the reaction mixture by adding 0.62 5 ml dichloromethane, followed by vigorous mixing for one minute and centrifugation at 12,000g for 5 minutes. An aliquot (0.5 ml) was taken from the organic layer, added to 0.5 ml of 0.5 N NaOH, then mixed and centrifuged as described above. Afterwards, 0.3 5 ml was taken from the aqueous layer (which contains the ionized form of PNP) and transferred to a well of a microplate (flat bottomed, 96 well; Costar, Cambridge, MA).Absorbance was read at 4 05 nm in a Thermo Max ™ microplate reader (Molecular Devices Corporation, Palo Alto, CA). Absorbance was compared with a blank containing 0.5 NaOH (no impurities were detected in the reaction mixture, which affect absorbance, and PNA is very stable at room temperature). Absorbance was converted to pmoles produced using an experimentally derived extinction coefficient for PNP.

General esterase activity was measured using the assay of Gomori (1953) as modified by van Asperen (1962). Midgut homogenates were prepared as described above (but homogenized in HEPES buffer: pH 7.6, 0.1 M) and the supernatants were stored at -70°C until used. HEPES buffer

(pH 7.6, 0.1 M) was used instead of phosphate buffer because phosphate buffers interact with SDS to produce turbid solutions. Just prior to assays, supernatants were thawed and diluted 50 X in buffer. From the diluted supernatants 3 0 jUl (containing 0.04 6 - 0.39 tg protein) were used as enzyme source for one assay. The reactions were performed in a water bath at 30°C in a total volume of 1 ml (= 30 Ml diluted homogenate + 960 Ml HEPES buffer (pH 7.6, 0.1 M) + 10 Ml substrate solution). Reactions were started by the addition of substrate solution (a-naphthyl acetate in ethanol; final concentration, 1.35 mM). After 10 minutes, the reaction was stopped by adding 2 ml of a solution containing Fast Blue B (0.2 mg/ml) and SDS (6 mg/ml). Exactly 20 minutes later, absorbance was read at 600 nm in 1 ml cuvets. Blanks were prepared by replacing the diluted supernatant with buffer. Conversion from absorbance to nmols of a-naphthol produced was based on an experimentally derived extinction coefficient for a-naphthol. Esterase assays were done twice for each larva (using same diluted homogenate). Only the means of the two replications were used in statistical analyses. Pilot studies established that product formation was linear for at least 3 0 minutes (data not shown). Product formation was also well within the linear range with respect to the amount of tissue equivalents. A broad pH optimum in

41esterase activity was observed at pH 7.6. The substrate concentration used equals approximately 9 x Kffl (or 90% of V ), as derived from a Lineweaver-Burk plot. Higher concentrations approach the limit of solubility of the substrate.

Vertical polyacrylamide slab gels were prepared as described by Davis (1964) with the following modifications. Polymerization was initiated by using ammonium persulfate only (added from a 10% solution; final concentration 0.069 %), and 7.5% acrylamide was used in the separating gel. Pilot studies (data not shown) showed that the best resolution occurred at this concentration. Model SE600 (Hoefer Scientific Instruments; San Francisco, CA) was used as a gel apparatus. A cooling unit kept the temperature in the lower buffer compartment at approximately 6°C. In each well, a 20 fil sample was applied that contained 75% undiluted homogenate (prepared as described above) and 25% glycerol. Gels (16 X 16 X 0.15 cm) were run at 3 0 mA until the tracking dye (bromophenol blue) reached the end of the gel. After electrophoresis, gels were put in boric acid (250 ml, 0.2 M) and stored in a refrigerator for 3 0 minutes to lower the pH of the gel. Afterwards, the gels were rinsed twice with water and transferred to phosphate buffer (250 ml, pH 6.5) containing 4 0 mg a-naphthol acetate (added from an

42ethanolic solution, just before use; substrate is very unstable at room temperature)• From this point, all incubations occurred at room temperature. After soaking the gels in phosphate buffer for 30 minutes, 100 ml of Fast Blue B solution were added (75 mg / 100 ml) that was filtered before use. After 15 minutes, the gels were removed and fixed in a solution of 10% acetic acid, 40% methanol in water.

Data AnalysisData from enzyme assays were analyzed with an analysis

of covariance (ANCOVA) (Neter et al., 1989). In this approach, activity differences between treatments (strain & diet) were compared at either the same larval weight or at the same larval weight and protein content, by using either weight or weight and protein as covariates. ANCOVA was used instead of ratios (i.e. activity / mg protein) because it is a more efficient method of taking the effects of weight and protein on enzyme activity into account (Packard & Boardman, 1980; Laurie-Ahlberg et al, 1980; Maroni et al., 1982). Least square means were calculated for activity levels at the average value of the covariates.

All statistical models were tested in PROC GLM of SAS (SAS Institute, 1985). Normality of residuals were tested

43with PROC UNIVARIATE of SAS. If data were transformed for ANCOVA, inverse transformations were applied to obtain means used in figures. Equality of variance in residuals between groups was tested in PROC TTEST. Separation of means was accomplished by a two-tailed t-test at P < 0.05.

Adjustments for unplanned multiple comparisons were made according to Bonferroni (SAS Institute, 1985).Planned comparisons for which Bonferroni's correction can be omitted include the following contrasts (a) on corn: corn strain versus rice strain (b) on bermudagrass: corn strain versus rice strain (c) within corn strain: bermudagrass versus corn reared, and (d) within rice strain: bermudagrass versus corn reared. The number of insects (determinations) in each treatment combination ranged from 9 to 14 for the MFO assays, and from 7 to 14 for the esterase assays.

ResultsLarval weights on day 8 of development were highest for

each strain when reared on its preferred host plant (Fig.3.1), although, mean weights were only significantly different between strains when reared upon artificial diet. However, the significant diet-by-strain interaction (F = 4.05; df = 2,12 5; P < 0.020) indicates that the difference between strains depends upon the diet on which

1 RICE STRAIN •M / •j CORN

120

100

8 0

6 0

4 0

20

0

LARVAL WEIGHT (MG)

ART. DIET B. GRASS

HOST PLANT

STRAIN

CORN

Fig. 3.1. Mean larval weights on day 8 of larval development for the two fall armyworm strains on diets used in enzyme assays (134 observations total). Strains are significantly different from each other on artificial diet.

larvae were reared. Thus, strains differed not just in overall average larval weight. In artificial diet reared larvae, MFO activity was high, relative to plant reared insects, but also significantly more variable (test of equal variance in residuals: F = 2 . 9 6 ; df = 1 9 , 4 7 ; P < 0 . 0 0 2 5 , after a square root transformation). Therefore, observations on artificial diet reared insects were separated from those on other diets for analysis. The strains were not significantly different with respect to MFO activity when reared on artificial diet (F = 3 . 7 8 ; df = 1 , 1 7 ; P < 0 . 0 6 9 ; corrected for weight & protein). Least square means (± SE) of MFO activity (in pmole / minute/ midgut) for larvae reared on artificial diet were 44 ± 22

for the rice strain, and 1 0 5 ± 2 0 for the corn strain.When larvae were reared on corn, corn strain larvae had

higher MFO activity than rice strain larvae (Fig. 3 . 2 ) . In addition, when larvae were reared on bermudagrass, corn strain larvae had higher MFO activity than rice strain larvae (no protein correction: P < 0 . 0 1 4 ; corrected for protein: P < 0 . 0 2 1 ). Within strains, effects of larval host on MFO activity did occur in the corn strain (Fig.3 . 2 ) , but not in the rice strain (no protein correction: P < 0 . 6 2 corrected for protein: P < 0 . 9 8 ; planned comparisons therefore Bonferroni's corrections not applied). The significant diet-by-strain effect in the

1 RICE STRAIN

I: C O R R E C T E D F O R W E I G H T

A N D P R O T E I N

M6 0

O

^ 40CT

V 20

TY

B. GRASS CORN

C O R N S T R A I N

I X : C O R R E C T E D F O R W E I G H T ON LY

a

B. GRASSHOST PLANT

CORN

Fig. 3.2. Mean MFO activity (in pmole/ minute/ midgut) of both strains reared on bermudagrass or corn. Means are corrected for either weight and protein content (I), or for weight only (XI)* For ANCOVA see table 3.1. Means with no letter in common are significantly different (P < 0.05; Bonferroni's correction).

ANCOVA, not corrected for protein differences (Table 3.1(b)), indicated that the effect of host plants on MFO activity differs (significantly) between the two strains. Thus, the effect of diet on MFO activity is significantly different between strains. If activity was corrected for protein, both strain and diet main effects were significant (Table 3.1(a)). Because differences in protein content between treatment combinations in the MFO assays were small (Fig. 3.3), and not significant (Table 3.1(c)), the difference between the two ways of expressing MFO activity is marginal. The manner in which activity is expressed causes a change in the significance of the diet- by-strain interaction to one of non-significance, but the overall pattern is the same (Fig. 3.2: 1 versus 11).

There were no strain or diet differences in general esterase activity between larvae reared on bermudagrass and corn (Fig. 3.4 & 3.5). No two treatment combination means were significantly different from each other on these two hosts combined, even when Bonferroni's corrections were omitted (in both Fig. 3.4 & 3.5). Therefore, strain and diet effects measured for general esterase activity (Table 3.2) resulted from artificial diet reared insects. There was a significant (overall) strain effect when esterase activity was corrected for protein (Table 3.2 (a)). In addition, there was a trend

48Table 3.1. ANCOVA MFO activity (data from Fig. 3.2 & 3.3).

Effect Mean Square* F df Pi) Corrected for weight and protein (R2-model==0.51)Diet 1436 5.38 1,41 0. 025Weight 485 1.82 1,41 0.18Protein 863 3.24 1,41 0.079Strain 6993 26.2 1,41 0.0001Strain*diet 806 3.02 1,41 0.0896Error 267 41>) Corrected for weight only (R2-model=0.47)Diet 1146 4.08 1,42 0.050Weight 981 3.49 1,42 0. 069Strain 8112 28.87 1,42 0.0001Strain*diet 1387 4 .94 1,42 0.032Error 281 42:) ANCOVA protein content midguts (R2-model=0. 15)**Diet 4.75 0.98 1,42 0.33Weight 16.76 3.44 1,42 0. 071Strain 6.31 1.30 1,42 0.26Strain*diet 14 . 56 2.99 1,42 0.091Error 4 . 86 42

* x 0.001 in protein contentProtein content was corrected for weight.

49

■ RICE STRAIN M CORN STRAIN

MG PROTEIN PER GUT0.4 r

B. GRASS CORNHOST PLANT

Fig. 3.3. Average protein content corrected for weight, in treatment combinations MFO assays (Fig. 3.2). For ANCOVA see Table 3.1 (c).

C l ] RICE STRAIN M CORN STRAIN HYBRIDS

ESTERASE ACTIVITY1.5

0.5

0

A i~ 5

' J A B, «\> 'J

: 0 B BC v ' V r \[1

- I ' * ' ' W w

a '.. :..........>

pillB B

Ifr

ART. DIET

B B B

B. GRASSHOST PLANT

CORN

Fig. 3.4. Mean general esterase activity (in jumole / minute / midgut) corrected for both weight and protein content in both strains and hybrids. For ANCOVA, see Table 3.2 (a). Means with no letter in common are significantly different (P < 0.05; Bonferroni's correction). uio

RICE STRAIN CORN STRAIN HYBRIDS

ESTERASE ACTIVITY

. ' P H

liiliPIp

«>>} , Will*ART. DIET

B B

Pi

B. GRASS

HOST PLANT

B

B

CORN

Fig. 3.5. Mean general esterase activity (in nmole / minute / midgut) corrected for weight only in both strains and hybrids. For ANCOVA, see Table 3.2 (b). Means with no letter in common are significantly different (P < 0.05; Bonferroni's correction). uiH

9999931

52Table 3.2. ANCOVA general esterase activity (on data Figs. 3.4, 3.5 & 3.7).

Effect Mean Square* F df Pi) Corrected for weight and protein (R2-model=0 .78)1+Diet 211.4 12.19 2,69 0.0001Weight 15.3 0.88 1, 69 0.35Protein 885. 5 51.09 1, 69 0.0001Strain 84.1 4.85 2,69 0. 011Strain*diet 42.8 2.47 4,69 0.053Error 17.3 69i) Corrected for weight only (R2-model=0.61)ttDiet 767 . 2 25. 80 2,70 0.0001Weight 552.7 18.59 1,70 0.0001Strain 43.5 1.46 2,70 0.24Strain*diet 22.2 0.75 4,70 0. 56Error 29.7 70

(c) ANCOVA protein content midguts (R2-model=0. 61)^Diet 6.32 11.83 2,70 0.0001Weight 14.81 27.69 1,70 0.0001Strain 4 .86 9.08 2,70 0.0003Strain*diet 0.70 1.30 4,70 0.28Error 0.53 70

T x 0.C01tt On Square-root transformed data^ On squared transformed data; protein content was

corrected for weight.

53towards a diet-by-strain interaction. The latter is caused by the rice strain's high activity on artificial diet (Fig. 3.4). However, there were also highly significant differences between strains and diets in the protein content of homogenates (Table 3.2(c); Fig. 3.6). No strain or diet-by-strain effects were apparent however, when activity was corrected for weight only (Table 3.2 (b),Fig. 3.5).

Higher protein content in homogenates from insects reared on artificial diet was also apparent in the MFO assays. There was, on average, 54% more protein in homogenates from insects reared on artificial diet than in plant reared insects when results from diet and plant reared insects were combined in the analysis (Diet main effect: F = 49? df = 2,61; P < 0.0001).

Banding patterns measured in electrophoretic gels stained for esterase activity were highly variable among individuals from both strains. Consequently, no diagnostic bands were found that could be used to separated strains or diets. One example of a gel is depicted in Fig. 3.7.

DiscussionDifferences in MFO, but not in esterase, activities

were observed between fall armyworm strains reared on preferred and non-preferred host plants. The corn strain

RICE STRAIN CORN STRAIN HYBRIDS

MG PROTEIN PER GUT

0.3

0.1

0ART. DIET B. GRASS CORN

HOST PLANT

Fig. 3.6. Average protein content, corrected for weight, in treatment combinations general esterase assays (Fig. 3.4). For ANCOVA, see table 3.2 (c).

55Strain:

CS RS HY CS RS HY CS RS HYDiet:

B B B C C C A A A

U K

11

Fig. 3.7. Results of electrophoretic separation of esterases, in corn strain (CS), rice strain (RS), and hybrids (HY) larvae, reared on either bermudagrass (B), corn (C), artificial diet (A). Only one out of 8 gels is shown.

56expressed higher MFO activity when reared on corn, than when reared on bermudagrass. In contrast, no differences in MFO activity were apparent in rice strain larvae reared on these hosts. A possible explanation for this difference is that allelochemicals present in corn induce MFO activity in the corn strain, but not in the rice strain.

Esterase activities were similar in both strains when reared on both host plants, but were significantly higher in both strains when reared on artificial diet. Thus, esterase activity can be influenced (and possibly induced) by the diet upon which larvae are reared. However, from the perspective of differences in host plant adaptation between the fall armyworm strains, only the diet effects in MFO activity seem to be relevant to host use patterns in nature.

The present results indicate that MFO activity differs between strains. Similar activity levels would have indicated no role of MFO in host plant adaptation. If MFO activity can be induced in the corn strain but not in the rice strain, an advantage of high MFO activity in feeding on corn plants is indicated. However, these data can not be directly related to host plant allelochemical metabolism, because the metabolism of allelochemical substrates was not directly measured. Multiple forms of MFO are known to exist that differ in substrate

57specificity (Soderlund & Bloomquist, 1990). Thus, the corn strain could have different isozymes of MFO, which metabolize PNA better than the isozymes of the rice strain.

Overall higher MFO activity in the corn strain compared to the rice strain on both host plants could indicate that some canalization had occurred in the induction response. However, canalization was not of such a magnitude that it has led to the loss of inducibility. This could be interpreted as evidence that metabolic costs due to the over production of detoxication enzymes are ecologically important.

A pattern similar to the one here in MFO activity was observed in a comparison of general esterase activities between two tiger swallowtail subspecies (Lindroth, 1989). The two subspecies differ in their ability to use plants of Salicaceae as hosts, which is associated with a difference in esterase activity. In the subspecies that is able to utilize Salicaceae, higher general esterase activity is measured, and activity can be induced by allelochemicals from Salicaceae, whereas esterase activity in the other subspecies cannot be induced.

An unexpected finding in the present study was that the highest activity in both enzyme systems occurred on artificial diet. Although it is generally assumed that

58artificial diet contains no inducers for detoxication enzymes, Ahmad & Forgash (1978) also found higher MFO activity on an artificial diet than on a plant material diet (oak) and speculated that insecticide residues in components of the artificial diet acted as MFO inducers.In the results presented here, contamination with insecticide residues in the pinto beans or wheat germ, or contamination with any other xenobiotic compound in components of the diet, could explain the poor performance and higher enzyme activities of insects reared on artificial diet. Generally, fall armyworms grow as well or better on artificial diet than on plant material (Pencoe & Martin, 1981; Pantoja et al, 1987; Veenstra, unpub. data).

The results of esterase assays also show the importance of considering differences in protein content between treatments when activity is corrected for protein. Enzyme activity corrected for protein content differs in two ways from activity corrected for weight only. First, statistical power can be increased when activities are corrected for protein differences. Second, when activity is corrected for protein the levels of mean activity are compared at the same protein concentration. Because detoxication enzymes make up only a fraction of the proteins detected by the protein assay, differences in protein between diets and strains are most likely due to

59other proteins. The higher protein content on artificial diet, for example, is likely to be the result of the high protein content of artificial diet.

Protein was, however, a highly significant effect in the model for esterase activity, despite the fact that it is correlated with weight. This means that there is a strong (and positive) correlation within each treatment combination between the amount of protein in the homogenate and esterase activity. A likely explanation for this is that the protein content of the homogenate is an indicator of how much detoxication enzyme is actually liberated from the tissue during homogenization.

In my results of general esterase activity, strains were only significantly different if activity was corrected for protein. Because the difference in protein content between the strains was significant, I concluded that strains do not differ significantly in general esterase activity.

The relevance of considering differences between treatments in protein content of homogenates was also shown by Lindroth et al. (1990 & 1991) who examined the effect of dietary protein on esterase activity. In both cases, insects reared on a diet low in protein exhibited significantly higher esterase activity per mg protein. But given the results here of a higher protein concentration

60in homogenates from larvae fed a high protein diet, the higher esterase activity on a low protein diet could be caused by differences in protein concentrations of homogenates. Strain differences in protein content were also detected by Grant et al (1989), between an insecticide resistant and susceptible strain of mosquito larvae.

In summary, the patterns in MFO activity indicate that enzymatic detoxication by these enzymes could play an important role in host plant adaptation. I found no evidence that indicated the importance of general esterases in host plant adaptation, although activity of these enzymes was affected by the diet upon which they were reared.

References to Chapter 3Ahmad, S. & A. J. Forgash. 1978. Gypsy moth Mixed-function

oxidases: gut enzyme levels increased by rearing on a wheat germ diet. Ann. Entomol. Soc. 71: 449-452.


Appel, H. M. & M. M. Martin. 1992. Significant ofmetabolic load in the evolution of host specificity of Manduca sexta. Ecology 73: 216-228.

Asperen, K. van. 1962. A study of housefly esterase by means of a sensitive colometric method.J. Insect Physiol. 8: 401-416.

61Berenbaum, M. & M. R. Zangerl. 1992. Quantification of

chemical coevolution In: R. F. Fritz & E. L. Simms [eds.], Ecology and evolution of plant resistance, University of Chicago Press, Chicago.

Bernays, E. A. & R. Barbehenn. 1987. Nutritional ecology of grass foliage-chewing insects. In: F. Slansky Jr. & J. G. Rodriguez [eds.], Nutritional ecology of insects, mites, spiders, and related invertebrates, John Wiley & Sons, New York.

Bernays, E. A. & M. Graham. 1988. On the evolution of host specificity in phytophagous arthropods.Ecology 69: 886-892.

Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding.Anal. Biochem. 72: 248-254.

Brattsten, L. B. 197 9. Biochemical defense mechanisms in herbivores against plant allelochemicals. In: G. A. Rosenthal & D. H. Janzen [eds.] Herbivores their interaction with secondary plant metabolites. Academic press, New York.

Davis, B. J. 1964. Disc electrophoresis-11. Method and application to human serum proteins.Ann. N.Y. Acad. Sci., 121: 404-427.

Dykhuizen, D. 1978. Selection for tryptophan auxotrophs of Escherichia coli in glucose-limited chemostats as a test of the energy conservation hypothesis of evolution. Evolution 32: 125-150.

Elliger, C. A., B. G. Chan, A. C. Waiss, Jr., R. E. Lundin & W. F. Haddon. 1980. C-glycosyflavones from Zea mays that inhibit insect development.Phytochemistry 19: 293-297.

Gomori, G. 1953. Human esterases.J. Lab. Clin. Med. 42: 445-453.

Grant, D. F., D. M. Bender & B. D. Hammock. 1989. Quantitative kinetic assays for glutathione S-transferase and general esterase in individual mosquitoes using an EIA reader.Insect Biochem. 19: 741-751.

62Gould, F. 1984. Mixed function oxidases and herbivore

polyphagy: the devil's advocate position.Ecol. Entomol. 9: 29-34.

Hansen, L. G. & E. Hodgson. 1971. Biochemical characteristics of insect microsomes: N- and O-demethylation. Biochem. Pharm. 20: 1569-1578.

Laurie-Ahlberg, C. C., G. Maroni, G. C. Bewley, J. C. Lucchesi, & B. S. Weir. 1980. Quantitative genetic variation of enzyme activities in natural populations of Drosophila melanoaaster.Proc. Natl. Acad. Sci. USA 77: 1073-1077.

Lindroth, R. L. 1989. Biochemical detoxication: mechanism of differential tiger swallowtail tolerance to phenolic glycosides. Oecologia 81: 219-224.


Lindroth, R. L., M. A. Barman & A. V. Weisbrod. 1991. Nutrient deficiencies and the gypsy moth, Lymantria dispar: effects on larval performance and detoxication enzyme activities. J. Insect Physiol. 37: 45-52.

Lindroth, R. L., B. D. Anson, & A. V. Weisbrod. 1990.Effects of protein and juglone on gypsy moths: growth performance and detoxification enzyme activity.J. Chem. Ecol. 16: 2533-2547.

Maroni, G., C. C. Laurie-Ahlberg, D. A. Adams & A. N. Wilton. 1982. Genetic variation in the expression of ADH in Drosophila melanoqaster. Genetics 101: 431-446.

Neal, J. J. 1987. Metabolic costs of mixed-function oxidase induction in Heliothis zea.Entomol. Exp. & Appl. 43: 174-179.

Neal, J. J. & M. Berenbaum. 1989. Decreased sensitivity of mixed-function oxidases from Papilio polvxenes to inhibitors in host plants. J. Chem. Ecol. 15: 439-446.

Neter, J., W. Wasserman, & M. H. Kutner. 1989. Applied linear regression models. Irwin, Homewood IL.

63Packard, G. C. & T. J. Boardman. 1987. The misuse of

ratios to scale physiological data that vary allometrically with body size. In: M. E. Feder & A. F. Bennet [eds.], New directions in ecological physiology, Cambridge Univ. Press, Cambridge, U.K.

Pantoja, A. P., C. M. Smith & J. F. Robinson. 1987. Development of fall armyworm Spodootera fruqiperda (Lepidoptera: Noctuidae), strains from Louisiana and Puerto Rico. Eviron. Entomol. 16: 116-119.

Pencoe, N. L. & P. B. Martin. 1981. Development andreproduction of fall armyworm on several wild grasses. Environ. Entomol. 10: 999-1002.


Pashley, D. P. 1988a. Current status of fall armyworm host strains. Fla. Entomol. 71: 227-233.

Pashley, D. P. 1988b. Quantitative genetics, development, and physiological adaptation in host strains of fall armyworm. Evolution 42: 93-102.

Quisenberry, S. S. & F. Whitford. 1988. Evaluation ofbermudagrass resistance to fall armyworm (Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am. 80: 731-733.

SAS Institute. 1985. SAS user's guide: Statistics, 5th ed. SAS institute, Cary, NC.

Schmalhausen, I. I. 1949. Factors of Evolution: the theory of stabilizing selection. Blakiston, Philadelphia.

Schoonhoven, L. M. & J. Meerman. 1978. Metabolic costs of chances in diet and neutralization of allelochemicals. Entomol. Exp. Appl. 24: 489-493.

Soderlund, D. M. & J. R. Bloomquist. 1990. Molecularmechanisms of insecticide resistance. In: R. T. Roush & B. E. Tabashnik [eds.] Pesticide resistance in arthropods. Chapman and Hall, New York.

64Weir, A. E. 1992. Plant variation and the evolution of

phenotypic plasticity in herbivore performance. In: R. S. Fritz & E. L. Simms [eds.], Plant resistance to herbivores and pathogens. Ecology, evolution, and genetics. University of Chicago Press, Chicago.



CHAPTER 4THE RELATIONSHIP BETWEEN HOST PLANT PERFORMANCE AND MIXED-FUNCTION OXIDASE ACTIVITY IN THE TWO STRAINS OF

FALL ARMYWORM: RESULTS OF SELECTION EXPERIMENTS

IntroductionFactors responsible for specialization in phytophagous

arthropods have been a point of interest and controversy for evolutionary biologists and entomologists alike. The idea that plants contain toxic allelochemicals that protect them against herbivorous insects (Ehrlich & Raven, 1964) has long dominated this discussion. Consequently, many view adaptation to host plants in terms of physiological adaptations needed to overcome toxic effects of host allelochemicals (Ahmad et al., 1986; Slansky,1992) .

Among physiological adaptations to host plants in phytophagous insects, enzymatic detoxication is thought to be one of the most important (Ahmad et al., 1986;Lindroth, 1991). Mixed-function oxidases (MFO) are thought to be especially important in this respect (Dowd et al., 1983; Gould, 1984; Ahmad et al, 1986; Lindroth, 1991). The broad substrate specificity of MFO allows insects to metabolize a wide variety of xenobiotic and endogenous compounds. This broad specificity is thought to be

65

66achieved by the coexistence of multiple forms of MFO, each with its own specific substrate specificity (Soderlund & Bloomquist, 1990).

Although the role of MFO in insecticide resistance has been well established (Agosin, 1985; Soderlund & Bloomquist, 1990), evidence for involvement of MFO in host plant adaptation is mostly circumstantial (Gould, 1984). Generally, MFO activity increases with increased exposure to allelochemicals, but there is little concrete evidence that the increase in activity leads to increased detoxication of host plant allelochemicals. Biochemical studies have shown that compounds that induce MFO activity are not necessarily metabolized by MFO (Gould, 1984).

Because most adaptations to host plants result from selection among genotypes within populations, an examination of genetic variation within populations could elucidate the underlying mechanisms. Studies on a variety of herbivorous arthropods have detected genetic variation for host plant performance within populations (reviewed in Via, 199 0), but only a few studies have attempted to partition variation into genetic variation caused by behavioral traits, like consumption rates, and variation caused by physiological traits. Several studies of Lepidoptera conclude that genetic variation in behavioral characters is more important than that in physiological

67traits (Futuyma et al, 1984; Karowe, 1990, but see Lindroth & Weisbrod, 1991).

The fall armyworm fSpodoptera fruqjperda J. E. Smith) is an ideal insect to study in the context of physiological adaptations because two strains exist in nature feeding on different sets of host plants (Pashley, 1986; 1988b). Fall armyworm has been recorded feeding on plants from more than 80 plant species in 23 families, but it clearly prefers members of the grass family (Pashley, 198 6). The fall armyworm strains include the corn strain, which feeds primarily on corn, and the rice strain, which feeds primarily on forage grasses and rice. Each strain appears to be adapted to its preferred host, but differences in performance between strains are small (Pashley, 1988a; Whitford et al., 1988; Chapters 2 & 3). The largest difference occurs when larvae are reared on corn.

The corn strain has a higher efficiency of converting digested corn into biomass than the rice strain, which suggest physiological differences between the strains when feeding on corn (Chapter 2). An important difference between hosts of the two strains is that low levels of active allelochemicals are typical of forage grasses (Bernays & Barbehenn, 1987), whereas several antibiotic compounds have been identified from corn (Woodward et al.

681979; Elliger et al. 1980). As expected, the corn strain has a higher level of MFO activity than the rice strain (Chapter 3). In addition, MFO activity in the corn strain was higher when reared upon corn than when reared upon bermudagrass, whereas in the rice strain, activity was similar on both hosts. These findings indicate that MFO activity may play a role in host plant adaptation.However, strains also differ in total consumption and in consumption per day (Chapter 2). Thus, feeding behavior also may be important in host plant adaptation.

The presence of significant amounts of genetic variation associated with host plant performance has been established previously (Pashley, 1988a). The specific question addressed here is whether genetic variation in MFO activity is an underlying cause of the genetic variation in host plant performance. This was examined by performing artificial selection experiments for increased larval weight on different host plants. After selection was completed, the effectiveness of selection in producing lines that differed in performance (i.e. larval weight and growth rates) was first established by rearing lines on several hosts. Next, MFO activities were measured in selected lines. If MFO activity is important for rapid development in larvae feeding on corn, then a positive correlation is expected between larval weights and MFO

69activity. If MFO activity is not related to rapid development, when larvae are reared on bermudagrass, no such correlation will occur on that host. As a result, the corn selected line would have a higher MFO activity than the bermudagrass selected line. Results from two such selection experiments indicate that lines selected on corn had higher MFO activities than lines selected on bermudagrass, or a control line.

Materials and Methods Rearing and Selection Procedure

Six lines were selected for increased larval weight in two separate experiments. One experiment used corn strain larvae, and the other rice strain larvae. Corn strain larvae were selected on either bermudagrass, corn, soybeans, or cotton. Rice strain larvae were selected on either bermudagrass, corn or artificial diet. In addition, a control line was established using rice strain larvae in which insects were reared, but not selected, on artificial diet.

All insects were reared in an environmental chamber at 25°C and on a 14:10 (L:D) photoperiod. A modified pinto bean diet was used as artificial diet as described in Quisenberry & Whitford (1988). Larvae reared on artificial diet were kept in 30 ml containers, in which the bottom

70was covered with approximately 1 cm artificial diet.Larvae reared on plant material were kept individually in 12 0 ml containers, to which had been added a 2% agar solution. This agar solution prevented desiccation of insects and plants. All plant material was obtained from greenhouse reared plants. Bermudagrass was grown from seeds. Corn plants (Pioneer, hybrid # 3165) were used until the first signs of tasseling. Soybean plants of the variety Davis were used. Cotton plants (Delta Pine Land Co., DPL 41) were used until flowers appeared. Plant material was replaced every other day.

Selection Using Corn Strain LarvaeA total of 30 larvae collected from corn plants in

Puerto Rico in March 1991 were used to establish the base population from which lines were selected. Within family selection (Falconer, 1981) was applied in the following manner to the four lines on bermudagrass, corn, soybeans and cotton. From each family, 2 5 to 35 larvae were reared initially. At day 8-10, all larvae in each family were weighed and the 5 or 6 largest larvae were used for establishing the next generation. After 2 generations, the cotton line died out and was eliminated from further study. After 4 generations of selection, each of the three remaining lines was reared on bermudagrass, corn, and

soybeans (24 per diet-by-line combination). An original objective was to quantify consumption and utilization during the last instar in selected lines. For this objective, 45 larvae of each line were reared on artificial diet until the last instar and then switched to either bermudagrass, corn, soybeans, or to cotton. Practical problems interfered with the execution of this objective, and consequently only observations on the development of these larvae were obtained. Larvae switched from artificial diet to bermudagrass consumed very little, resulting in nearly no growth during the last instar, and were excluded from the analysis. The following measurements were taken on larvae: weight at day 8, larval development time, and pupal weight (dry). Inbreeding coefficients were calculated from pedigrees for each individual used (Falconer, 1981).

Selection Using Rice Strain LarvaeRice strain larvae were collected in Baton Rouge, LA in

August 1992 from a pasture in which bermudagrass was the dominant grass. Forty, field-collected rice strain individuals were used to establish the base population from which lines were established. Individuals were reared for one generation on artificial diet prior to the selection experiment. Selection proceeded as described for

72corn strain larvae except that artificial diet was substituted for soybean. A control line was established on artificial diet to determine performance and MFO activity in the absence of selection. After selection, each line was reared on bermudagrass and corn (17 per diet-by-line combination). In addition, 45 larvae of each line were reared on artificial diet. The same measurements were taken on these larvae as described for the corn strain.

MFO AssaysAfter selection was completed, siblings of larvae used

to determine growth rates were reared for one generation on artificial diet with no selection applied. Offspring from artificial diet-reared individuals were used to estimate MFO activity. In the corn strain, these offspring were reared on bermudagrass, corn, and soybeans prior to the MFO assays. Activity was measured in rice strain offspring reared on bermudagrass, corn, and artificial diet. The number of observations in each diet-by-line combination was 4 or 5 in both experiments. For all measurements, actively feeding last instar larvae (at least 36 hours into the last instar) were weighed, and midguts were dissected and homogenized.

The activity of MFO was estimated by measuring the enzymatic O-demethylation of p-nitroanisole (PNA) to form

73p-nitrophenol (PNP) by the method of Hansen & Hodgson (1971) as modified by Neal & Berenbaum (1989). Details of the procedure can be found in Chapter 3. Protein concentrations were estimated by the method of Bradford (197 6) with bovine serum albumin as standard. For assays with corn strain larvae, three midguts were homogenized together in 1.5 ml phosphate buffer (pH 7.8; 0.1 M), and 1.35 ml from the 12,000g supernatant (containing 0.74 - 1.93 mg protein) were used as enzyme source for one determination. For the assays with rice strain larvae, two midguts were homogenized in 0.7 ml HEPES buffer (pH 7.8; 0.1 M), and 0.445 ml from the 12,000g supernatant were used as an enzyme source, containing 0.4 0 - 0.87 mg (plant-reared), or 0.62 - 1.27 mg (artificial diet-reared) protein. Reactions were started by adding substrate and run for 10 minutes. Product was extracted from the reaction mixture with dichloromethane. An aliquot of the organic layer was mixed with 0.5 N NaOH, then centrifuged. An aliquot of the aqueous layer was transferred to either a 1 ml cuvette (corn strain assays) or a well of a microplate (for rice strain assays), and absorbance of the NaOH solution was measured using a spectrophotometer (Beckman®, Model 3 5) or a microplate reader, respectively.

74Statistical Analysis

Relative growth rate in larvae was expressed as log (pupal dry weight) / larval development time. A log of weight was used because growth is an exponential process. The growth determination in the experiments that used rice strain larvae was set up as a randomized block design (Sokal & Rohlf, 1981; two blocks representing the two consecutive days on which the experiment was started). All other experiments were set up as complete randomized designs (Sokal & Rohlf, 1981).

All statistical models were tested in PROC GLM of SAS (SAS Institute, 1985). Interactions terms with P values greater than 0.3 were removed from the model with two exceptions: diet-by-line (never removed), and covariate/block-by-line (removed if P > 0.05). Normality of residuals was tested in PROC UNIVARIATE of SAS. If transformations were applied for ANOVA, inverse transformations were applied to obtain means for figures (inverse transformations were not used for mean separations). Separation of means was accomplished by a two-tailed t-test at P < 0.05. Because only differences between lines on the same diet were of interest, only the means of each diet-by-line combination were compared with other diet-by-line means from the same diet. Adjustments for (unplanned) multiple comparison were made according to

Bonferroni (SAS Institute, 1985). Corrections were based on the number of comparisons on each diet, not overall number of comparisons. From the following planned comparisons between selected lines in MFO activity, Bonferroni's correction was omitted: bermudagrass line versus corn line (both strains, higher activity expected in corn line), control versus corn line (rice strain, higher activity expected in corn line), and control versus bermudagrass line (corn strain, no difference expected).

The data from the MFO activity assays were analyzed with an analysis of covariance (ANCOVA) (Neter et al., 1989). In this approach, activity differences between treatments (strain and diet) were compared at either the same larval weight or at the same larval weight and protein content, by using either weight or weight and protein as covariates. ANCOVA was used instead of ratios (i.e. activity / mg protein) because ANCOVA results in higher statistical power and avoids the problem of correlations between activity ratios and their denominators (protein and weight; Laurie-Ahlberg et al., 1980; Maroni et al., 1982; Packard & Boardman, 1987). Enzyme activity is expressed either per individual (by using only weight as a covariate) or per individual corrected for protein content (by using both weight and protein as covariates). Enzyme activity corrected for

76protein results in a higher statistical power and corrects for random differences in protein content. If, however, treatments differ in protein content, then activity corrected for protein differences can become biased. This problem can be circumvented by using activity corrected for weight only. Least square means were calculated (± standard error), which are the estimated mean activity levels corrected for differences in weight and protein or differences in weight only, between diets and lines.

ResultsAverage inbreeding coefficients were modest and

differences between lines in inbreeding were small (Table 4.1). If parents have two grandparents in common, inbreeding coefficients in offspring would be 0.063, with one grandparent in common the value would be 0.031 (after Falconer, 1981).

In experiments with corn strain larvae, differences between lines in larval weights and growth rates were small (Fig. 4.1 & 4.2). But significant diet-by-line interactions in both weight on day 8 and relative growth rate were obtained (Table 4.2). Thus, differences among lines in performance depended upon diet, which indicates that the selected lines exhibited a certain degree of specificity in host plant performance.

77Table 4.1. Average inbreeding coefficients of each selected line in both experiments, as calculated from the pedigree.

Selected Lines: Growth rates MFO activity(a) Experiment using corn strain larvaeBermudagrass 0.027 0.065Corn 0.033 0.058Soybeans 0.074 0.110(b) Experiment using rice strain larvaeControl 0.037 0.044Bermudagrass 0.035 0.077Corn 0.040 0.046Artificial diet 0.027 0.042

L I N E S : E 1 3 G R A S S S O Y B E A N

W EIG H T (MG)8 0

6 0

4 0

20

A R T . D I E T B. G R A S S C O R N S O Y B E A N S

H O S T P L A N T

Fig. 4.1. Average weights at day 8 of larvae in selected lines derived from corn strain larvae (ANOVA: Table 4.2 (a)). Line means that have no letter in common were significantly different (P < 0.05; Bonferroni's correction). Means were only compared with other means on the same diet.

L I N E S : G R A S S C O R N S O Y B E A N

G R O W T H R AT E

0 .3

0.2

0.1

0

X : S W I T C H E D F R O M A R T . D I E T

A A BA A A

AB

B

X L : N O S W I T C H I N G H O S T S

a a a a

lit$ll$l

#1>z§.

iii

i f

i pnn

C O R N S O Y B E A N C O T T O N B. G R A S S

H O S T P L A N T

C O R N S O Y B E A N S

Fig. 4.2. Relative growth rate (=log (dry weight pupa) / larval development time) in selected lines, derived from corn strain larvae (ANOVA: Table 4.2 (b)). Line means that have no letter in common were significantly different (P < 0.05; Bonferroni's correction). Means were only compared with other means on the same plant.

VO

80Table 4.2. ANOVA growth rates in the selection experiment, which used corn strain larvae. Data are shown in Fig. 4.1 & 4.2.

Effect F(a) Weight at day 8 (R2-model=0.77)t.DietLineDiet*line

3883.392.47

df

3,314 2, 314 6, 314

(b) Relative growth rate (R2-model=0.69)DietLineDiet*lineSexDiet*sex

72.82.741.98

12.562.94

5.1942.194 10,1941.194 1, 194

0.0001 0. 035 0.024

0.0001 0.069 0.037 0.0005 0.014

+ Log transformed.^ log (pupal dry weight) / larval development time.

ANOVA performed on 4th power of these values.

Significantly-different mean weight or growth rates were obtained for larvae reared on artificial diet, or switched from artificial diet. The bermudagrass line had highest weights on all diets (Fig. 4.1) and the fastest growth rates on most diets (Fig. 4.2). Although rates of development differed among hosts, selected lines generally did not perform better on their respective hosts. In other words, a pattern consistent with host plant adaptation was not detected. The fact that trends in performance were not in the expected directions, despite the significant diet- by-line interactions, indicates that either the actual differences in performance between lines were small, or selection did not change the lines as expected.

In experiments with rice strain larvae, diet dependent differences in weight and growth rates were more pronounced than in the corn strain, although few values were significant (Fig. 4.3 & 4.4). The analysis of variance resulted in significant diet-by-strain interactions in both weight at day 8 and relative growth rate (Table 4.3), which indicates that performance of a selected line on one diet is not indicative of performance on another diet. The artificial diet line and the bermudagrass line had the highest weights on diets they had been selected upon (Fig. 4.3). The artificial diet line had a higher weight than the corn line when both were

L I N E S : C O N T R O L D I E T C O R N G R A S S

1 6 0 n

120

8 0 -

4 0

W E I G H T ( M G )

0

B

Wmmm,

B

A B

w m m

A R T . D I E T B. G R A S S

H O S T P L A N T

C O R N

Fig. 4.3. Average weights at day 8 of larvae in selected lines derived from rice strain larvae (ANOVA: Table 4.3 (a)). Line means that have no letter in common were significantly different (P < 0.05; Bonferroni1s correction). Means were only compared with other means on the same diet. 00N)

^999999

99999999

L I N E S : C O N T R O L D I E T C O R N G R A S S

G R O W T H R A T E

B

0 .3 -

0.2

0.1 -

M m‘K«NSSKW»»

A R T . D I E T B. G R A S S

H O S T P L A N T

A A A A

C O R N

Fig. 4.4. Relative growth rate (=log (weight pupa) / larval development time) in selected lines derived from rice strain larvae (ANOVA: Table 4.3 (b)). Line means that have no letter in common were significantly different (P < 0.05; Bonferroni's correction). Means were only compared with other means on the same plant.

84Table 4.3. ANOVA growth rates in the selection experiment, which used rice strain larvae. Data are shown in Fig 4.3 - 4.4.

Effect F df P(a) Weight at day 8 (R2-model=0.46).Diet 0.61+ 2,2 0.62Block 0.93* 1,2 0.44Diet*block 58.5 2,243 0. 0001Line 3.69 3,243 0. 013Diet*line 5.59 6,243 0.0001

(b) Relative growth rate (R2-model=0. 46)++.Diet 0.43* 2,2 0.70Block 0.63+ 1,2 0. 51Diet*block 46.2 2,240 0.0001Line 1.60 3,240 0.19Diet*line 4 .13 6, 240 0. 0006Sex 4.11 1, 240 0. 044

+ = Satterthwaite's approx. (Sokal & Rohlf, 1981). = log (pupal dry weight) / larval development time. ANOVA performed on 4th power of these values.

85reared on corn, but both had higher weights than the control. Similar results were obtained for growth rates, although differences were smaller (Fig. 4.4). The artificial diet and bermudagrass lines had the highest growth rates on artificial diet and bermudagrass, respectively. Growth rates in all lines were similar on corn. Thus, the pattern of performance in the rice strain was more consistent with that expected if strains exhibit genetic variation that enhances performance on specific diets.

Selected lines derived from the corn strain differed significantly in MFO activity (Table 4.4). The corn selected line had, as expected, significantly higher activity than the grass selected line. The soybean line had an MFO activity intermediate between these two lines. Greater effects of hosts were observed when activity was corrected for protein (Table 4.5a) compared to weight only (Table 4.5b). Within each line, individuals had highest activity when reared on corn, intermediate activities on bermudagrass, and lowest activities on soybeans.

An analysis of covariance indicated significant effects of line and diet (Table 4.5). Diet had also a significant effect on protein in homogenates (Table 4.5). The significant diet effect on enzyme activity (corrected for protein) could, therefore, be due to differences between

86Table 4.4. MFO activity (pinole / minute / midgut) corrected for weight and protein or for weight only in selected lines derived from corn strain larvae. Line means with no letter in common were significantly different from each other (P < 0.05; Bonferroni's correction).

Selection Diet reared MFO activity: MFO activity: Line upon combination line averagel) Activity corrected for protein.Grass B. grass 244 ± 46 aGrass Corn 311 ± 52 a 227 + 26 aGrass Soybean 124 ± 44 bCorn B. grass 370 ± 52 aCorn Corn 462 ± 58 a 329 + 30 bCorn Soybean 154 ± 54 bSoybean B. grass 312 ± 67 aSoybean Corn 333 ± 52 a 306 + 35 abSoybean Soybean 271 ± 58 a>) Activity corrected for weight only.Grass B. grass 212 ± 52 aGrass Corn 271 ± 57 a 222 + 31 aGrass Soybean 182 ± 46 aCorn B. grass 309 ± 55 abCorn Corn 497 ± 66 b 339 + 35 bCorn Soybean 211 ± 59 aSoybean B. grass 221 ± 70 aSoybean Corn 313 ± 59 a 269 + 38 abSoybean Soybean 273 ± 66 a

87Table 4.5. ANCOVA MFO enzyme activity and protein content homogenates in the selection experiments with corn strain larvae. Enzyme activity was corrected for either weight and protein or for weight only. Data are shown in Table 4.4.

Effect Mean Square1' F df Pi) Activity corrected for protein (R2 ■-model=0.62)Diet 46.7 4.75 2,30 0. 016Weight*diet 68.1 2.31 3,30 0.096Protein 103.1 10.5 1,30 0.0029Line 36.9 3.75 2,30 0. 035Line*diet 15.3 1.56 4,30 0.21Error 9.8 30>) Activity corrected for weight only (R2-model=0.49)Diet 26.2 2.04 2,31 0.14Weight*diet 15.7 1. 22 3,31 0.32Line 44.6 3.48 2 ,31 0. 04Line*diet 17.2 1.34 4,31 0.28Error 12.8 31:) ANCOVA protein content midguts (R2 ■-model=0.59)Weight 86.5 14.95 1,33 0.0005Diet 51.2 8.84 2,33 0.0008Line 18.1 3.12 2,33 0.057Line*diet 7.8 1.34 4,33 0.28Error 5.8 33

x 1000 in MFO activity, x 0.001 in protein content.

88the diets in protein content. The fact that the diet-by- line interactions were not significant indicates that differences between lines were independent of hosts. In summary, results of selection experiment with the corn strain indicated that selection for higher larval weight on corn resulted in higher MFO activity compared to selection for increased larval weight on bermudagrass. Within line patterns supported previous results (Chapter 3) that indicate corn diets for corn strain larvae result in higher MFO activity than bermudagrass diets.

In all lines derived from the rice strain, MFO activity was much higher when larvae were reared on artificial diet relative to larvae reared on plants (Table 4.6). Activity in larvae reared on artificial diet was also significantly more variable (test of equal variance in residuals: F = 4.11; df = 21,37; P < 0.0001, after a square root transformation). Therefore, artificial diet reared insects were eliminated from further analysis. Strains were not significantly different in MFO activity when reared on artificial diet (F = 1.46; df = 3,16; P < 0.26; activity corrected for weight and protein).

As in the experiment with the corn strain, rice strain larvae selected on corn had the highest MFO activity (Table 4.6). Overall activity levels in the rice strain were almost an order of magnitude less than in the

Table 4.6. MFO activity (pinole / minute / midgut) corrected for weight and protein or for weight only in selected lines derived from rice strain larvae. Means with no roman letter (a,b) in common were significantly different from each other (P < 0.05; Bonferroni's correction). Means with no greek letter (i|f,fi) in common were significantly different, when the Bonferroni's corrections were omitted.

Selection Diet reared MFO activity:* MFO activity:Line upon combination line average

(a) Activity corrected for protein.Control B. grass 37.8 + 7.1Control Corn 32.5 + 7.S 35.2 + 5.6 a *Diet B. grass 51. 6 + 6.7Diet Corn 46.2 + 6.5 48.9 + 4.7 a i(i nCorn B. grass 58.5 + 6.7Corn Corn 46.2 + 6.5 52 . 3 + 4.8 a nGrass B. grass 33.2 ± 7.3Grass Corn 40.5 ± 7.9 36.9 + 5 . 3 a 4*12Control Art. diet 174 ± 23Diet Art. diet 124 ± 20Corn Art. diet 111 ± 22Grass Art. diet 123 ± 213) Activity corrected for weight only.Control B. grass 32.2 ± 7.8Control Corn 21.7 ± 8.1 27 . 0 + 5.7 a i(iDiet B. grass 47.7 ± 7.5Diet Corn 48.2 ± 7.4 47.9 + 5.4 ab 12Corn B. grass 64.9 ± 7.3Corn Corn 48.5 ± 7.3 56.7 + 5.2 b nGrass B. grass 31.7 ± 8.3Grass Corn 51.3 ± 8.1 41.5 + 5.8 ab i|inControl Art. diet 174 ± 21Diet Art. diet 124 ± 19Corn Art. diet 111 ± 21Grass Art. diet 123 ± 19

* Means of combinations were not significantly different from means on same diet (even without Bonf. corr.).

90corn strain. The bermudagrass and control lines were not significantly different and had the lowest activities. The artificial diet selected line was similar to the corn line in its high MFO activity.

Lines were significantly different in MFO activity if activity was corrected for weight, but not if activity was corrected for protein content (Table 4.7). However, protein content of homogenates was also significantly different between lines (Table 4.6c). In contrast to results with corn strain larvae, within line MFO activity was not different between corn and bermudagrass reared larvae. In summary, selection for increased larval weight with rice strain larvae was more successful than with corn strain larvae in obtaining lines that had increased performance on the diet they were selected upon, but still resulted in significant differences in MFO activity. As in an earlier experiment (Chapter 3), differences in MFO activity between corn and bermudagrass reared larvae were not apparent.

DiscussionFour generations of selection for increased larval

weights produced selection lines that differed significantly in MFO activity, and to some degree, in performance on various hosts. Patterns of variation were

91Table 4.7. ANCOVA MFO enzyme activity and protein content homogenates in the selection experiment with rice strain larvae. Enzyme activity was corrected for either weight and protein or for weight only. Data are shown in Table 4.6.

Effect Mean Square1’ F df P(a) Activity corrected for protein (R2-model=0.58)Diet 132 0. 65 1,27 0.46Weight 282 1.39 1,27 0.25Protein 1945 9.58 1,27 0.0045Line 541 2.67 3 , 27 0. 068Line*diet 125 0. 62 3,27 0.61Error 203 27>) Activity corrected for weight only (R2-model=0.43)Diet 25 0.10 1,28 0.76Weight 56 0.21 1,28 0.65Line 1306 4.93 3,28 0.0071Line*diet 1618 2.03 3,28 0.13Error 265 28:) ANCOVA protein content midguts (R2--model=0.46)Weight 5.96 1.35 1,28 0.26Diet 2.78 0. 63 1,28 0.43Line 20.05 4.53 3,28 0.010Line*diet 9.61 2.17 3,28 0.11Error 4.42 28

t x 0.001 in protein content.

92consistent with expectations for MFO activity but not performance. Thus, the potential for improvements in host plant performance appears to be small. Because the directions of differences in MFO activity matched expectations based on strain differences (Chapter 3), it is not likely that inbreeding depression or genetic drift could have caused differences.

Selection for increased larval weight on a specific host can lead to three different results. Selection can increase average larval weight, irrespective of which host larvae are reared upon. Selection can result in increased average larval weight on one host with a concomitant decrease on other hosts, commonly referred to as tradeoffs. Finally, selection can result in an increase of average larval weight only when the host reared upon is the same as the one used for selection. In that case, performance on other hosts is not compromised relative to a control. The second and third options indicate adaptation with the second one containing the greatest potential for divergence between lines on different hosts.

Growth rate data do not provide evidence that an increased performance on one host plant leads to a decrease in performance on other host plants. However, the experiments were not specifically designed to examine trade-offs. The only instance where the data could suggest

93a trade-off is in the corn strain, in the grass line on artificial diet. The grass selected line had a lower weight on day 8 on artificial diet than the control line. However, the control line was reared on artificial diet and therefore, selected to some extent for increased performance on artificial diet as well.

A possible explanation for the observed patterns in performance is that the improvements in performance may have been host specific, but were too small to detect given the sample sizes. A small difference in actual performance between selected lines is not unexpected because differences in larval weights between the two strains were previously determined to be small (Pashley, 1988a; Chapter 3). For example, Pashley (1988a) found no significant difference in larval weight between strains when reared on rice (based on 81 and 105 larvae of each strain).

Genetic variation in host plant performance, as measured in the laboratory, reflects to some extent variation in adaptation to the laboratory environment (Fry, 1993; Service & Rose, 1984). If selection acted only on this component, selected lines would differ in their overall performance, resulting only in significant line main effects. The significant diet-by-line interactions thus indicate that selected lines differed in more than

94just in how well they had adapted to the laboratory environment.

Differences in MFO activity between corn and bermudagrass selected lines indicate that MFO activity could play an important role in detoxifying host plant allelochemicals. In that case allelochemicals from corn would be more than just feeding repellents or stimulants for the fall armyworm. These results are in line with Rose et al. (1991), who showed that selection in tobacco budworms for resistance to specific allelochemicals in their artificial diet results in higher MFO activity.

These results do not provide evidence that selection had changed enzyme activity expressed only when reared on specific hosts, since the differences between lines did not depend upon which diet they were reared. However, the scale of this experiment was too small, relative to the variability in activity, to determine whether differences in overall MFO activity, irrespective of hosts, are more important than host specific effects. The small size of the population from which selected lines were established is another cause of the limited utility of these experiments in distinguishing between these two effects.By chance, small samples could contain more genetic variation for either host specific effects or overall effects, irrespective of hosts.

Many studies have shown that populations harbor significant amounts of variation in enzyme activity (references in Chan & Burton, 1992), but few have attempted to assess the ecological importance of this variation (Lindroth & Weisbrod, 1991; Chan & Burton,1992). Genetic variation in MFO activity could be caused by variation in allozyme variants of the MFO enzyme or by variation in regulatory genes affecting MFO activity or both (MaCintyre, 1982).

My results also indicate that differences in physiological adaptations within populations of fall armyworms could be of ecological significance. Some have argued that host plant allelochemicals play only a secondary role in the evolution of insect host ranges (Bernays & Graham, 1988). Others stress the importance of larval feeding behavior in host plant performance (Usher & Feeny, 1983; Bernays & Cornelius, 1992). For example, Waldbauer & Fraenkel (1961) showed that feeding on normally rejected host plants, as a result of experimental manipulation, can result in near normal development in tobacco hornworms. Studies at the population level result in different conclusions regarding the importance of behavioral factors versus physiological factors in adaptation. Fry (1988) concluded that variation between populations of spider mites in host plant performance was

96due to behavioral rather than physiological factors, but Gould et al. (1982) suggested that genetic variation in host plant performance in spider mite populations was caused by physiological differences.

In summary, results from the two selection experiments indicated that populations of fall armyworms harbor significant amounts of genetic variation in MFO activity but little for host plant performance. Because differences between lines in performance were small, increased performance on a specific host is not likely to result in decreased performance on other hosts. Results also indicate that MFO activity could be important for larvae feeding on corn. Thus, differences between the strains in MFO activity could well be related to their ability to successfully utilize corn as a host.

References to Chapter 4Agosin, M. 1985. Role of microsomal oxidation in

insecticide degradation. In: G. A. Kerkut & L. I. Gilbert [eds.], Comprehensive insect physiology biochemistry and pharmacology, volume 12. Pergamon, Oxford.


Bernays, E. A. & R. Barbehenn. 1987. Nutritional ecology of grass foliage-chewing insects. In: F. Slansky & J.G. Rodriguez [eds.]: nutritional ecology of insects, mites, spiders and related invertebrates. Wiley- interscience, New York.

97Bernanays, E. A. & M. Cornelius. 1992. Relationship

between deterrence and toxicity of plant secondary compounds for the alfalfa weevil Hvoera brunneioennis. Entomol. Exp. Appl. 64: 289-292.

Bernays, E. & M. Graham. 1988. On the evolution of host specificity in phytophagous arthropods.Ecology 69: 886-892.

Bradford, M. M. 1976. A rapid and sensitive method for the guantitation of microgram guantities of protein utilizing the principle of protein-dye binding.Anal. Biochem. 72: 248-254.

Chan, J. W. Y. & R. S. Burton. 1992. Variation in alcohol dehydrogenase activity and flood tolerance in white clover, Trifolium repens. Evolution 46: 721-734.

Dowd, P. F., C. M. Smith & T. C. Sparks. 1983. Detoxification of plant toxins by insects.Insect Biochem. 13: 453-468.



Falconer, D. S. 1981. Introduction to quantitative genetics. Longman, London.

Fry, J. D. 1988. Variation among populations of thetwospotted spider mite, Tetranvchus urtica Koch (Acari: Tetranychidae), in measures of fitness and host acceptance behavior on tomato.Environ. Entomol. 17: 287-292.

Fry, J. D. 1993. The "general vigor" problem: can antagonistic pleiotropy be detected when genetic covariances are positive? Evolution 47: 327-332.

Futuyma, D. J., R. P. Cort & I. van Noordwijk. 1984. Adaptation to host plants in the fall cankerworm (Alsophila pometaria) and its bearing on the evolution of host affiliation in phytophagous insects.Am. Nat. 123: 287-296.

98Gould, F. 1984. Mixed function oxidases and herbivore

polyphagy: the devil's advocate position.Ecol. Entomol. 9: 29-34.

Gould, F., C. R. Carroll & D. J. Futuyma. 1982. Crossresistance to pesticides and plant defenses: a study of the two-spotted spider mite.Ent. exp. & appl. 31: 175-180.

Hansen, L. G. & E. Hodgson. 1971. Biochemical characteristics of insect microsomes: N- and O-demethylation. Biochem. Pharm. 20: 1569-1578.

Karowe, D. N. 1990. Predicting host range evolution: colonization of Coronilla varia by Colias philodice (Lepidoptera: Pieridae). Evolution 44: 1637-1647.

Laurie-Ahlberg, C. C., G. Maroni, G. C. Bewley, J. C. Lucchesi, & B. S. Weir. 1980. Quantitative genetic variation of enzyme activities in natural populations of Drosophila melanoqaster.Proc. Natl. Acad. Sci. USA 77: 1073-1077.

Lindroth, R. L. 1991. Differential toxicity of plant allelochemicals to insects:roles of enzymatic detoxication systems. In: E. A. Bernays [eds.], Insect- plant interactions, volume 111. CRC Press, Boca Raton.

Lindroth, R. L. & A. V. Weisbrod. 1991. Genetic variation in response of the gypsy moth to aspen phenolic glycosides. Biochem. Sys. Ecol. 19: 97-103.

Maroni, G., C. C. Lauri-Ahlberg, D. A. Adams & A. N.Wilton. 1982. Genetic variation in the expression of ADH in Drosophila melanoqaster. Genetics 101: 431-446.

MaCintyre, R. J. 1982. Regulatory genes and adaptation. Past, present and future. Evol. Biol. 15: 247-285.

Neal, J. J. & M. Berenbaum. 1989. Decreased sensitivity of mixed function oxidases from Papilio polvxenes to inhibitors in host plants. J. Chem. Ecol. 15: 439-446.

Neter, J., W. Wasserman, & M. H. Kutner. 1989. Applied linear regression models. R. D. Irwin, Homewood IL.

99Packard, G. C. & T. J. Boardman. 1987. The misuse of

ratios to scale physiological data that vary allometrically with body size. In: M. E. Feder & A. F. Bennet [eds.], New directions in ecological physiology, Cambridge Univ. Press, Cambridge, U.K.


Pashley, D. P. 1988a. Quantitative genetics, development, and physiological adaptation in host strains of fall armyworm. Evolution 42: 93-102.

Pashley, D. P. 1988b. The current status of fall armyworm host strains. Fla. Entomol. 71: 227-234.

Quisenberry, S. S. & F. Whitford. 1988. Evaluation ofbermudagrass resistance to fall armyworm (Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am. 80: 731-733

Rose, R. L., F. Gould, P. E. Levi & E. Hodgson. 1991. Differences in cytochrome P450 activities in tobacco budworm larvae as influenced by resistance to host plant allelochemicals and induction.Comp. Biochem. Physiol. 99B: 535-540.

SAS Institute. 1985. SAS user's guide: statistics, 5th ed. SAS Institute, Cary NC.

Service, P. M. & M. R. Rose. 1985. Genetic covariation among life-history components: the effect of novel environments. Evolution 39: 943-945.

Slansky, 1992. Allelochemical-nutrient interactions inherbivore nutritional ecology, In: G. A. Rosenthal & M. R. Berenbaum [eds.], Herbivores: their interaction with secondary plant metabolites, 2E. Volume 11:Evolutionary and ecological processes, Academic Press, San Diego.

Soderlund, D. M. & J. R. Bloomquist. 1990. Molecularmechanisms of insecticide resistance. In: R. T. Roush & B. E. Tabashnik [eds.]. Pesticide resistance in arthropods. Chapman and Hall, New York.

Sokal, R. R. & F. J. Rohlf. 1981. Biometry. Freedman, New York.

100Usher, B. F. & P. Feeny. 1983. Atypical secondary

compounds in the family cruciferae: tests for toxicity to Pieris rapae. and related crucifer-feeding insects. Entomol. Exp. Appl. 34: 257-262.

Via, S., 1990. Ecological genetics and host adaptation in herbivorous insects: the experimental study of evolution in natural and agricultural systems.Annu. Rev. Entomol. 35: 421-446.

Waldbauer, G. P. & G. Fraenkel. 1961. Feeding on normally rejected plants by maxillectomized larvae of the tobacco hornworm, Protoparce sexta (Lepidoptera: Spingidae). Ann. Entomol. Soc. Am. 54: 477-485.



CHAPTER 5 SUMMARY AND CONCLUSIONS

Results from three sets of experiments dealing with larval performance in fall armyworm host strains have provided insights into factors influencing host use. The objectives were to establish whether: 1) physiological or behavioral factors could explain differences in performance on corn, 2) whether enzymes associated with detoxication of allelochemicals differ in activity between strains and 3) whether genetic variation in detoxification enzymes within strains affects performance.

A comparison between strains in consumption and utilization in larvae fed corn showed that the lower performance on corn by the rice strain is, in part, caused by the lower consumption rate of the rice strain on corn. In addition, corn strain larvae had a higher efficiency of converting digested food into biomass than the rice strain. This suggests that the corn strain is better adapted physiologically to utilize corn as a host than the rice strain. There is no indication that the corn strain performs better on corn because it is less sensitive to digestion-reducing allelochemicals from corn.

In the corn strain, mixed-function oxidase (MFO) activity was higher when reared on corn than reared on bermudagrass. In the rice strain, activity did not differ

101

102between these hosts. This could be due to the fact that MFO activity can be induced by corn allelochemicals in the corn strain but not in the rice strain. When reared on bermudagrass, the corn strain expressed higher MFO activity than the rice strain. This could indicate that the corn strain has higher MFO activity irrespective of host. Thus, MFO activity could be important for larvae feeding on corn.

When larvae were reared on corn or bermudagrass, no differences were apparent in general esterase activity between either strains or hosts. Thus, these enzymes do not seem to be important in host plant adaptation.However, esterase activity was significantly higher on artificial diet, which suggests that esterases activity can be altered by diet. Results also indicated the existence of substantial amounts of allozyme variation within the strains for general esterase.

Selection experiments confirmed the presence of genetic variation in host plant performance within each strain (Pashley, 1988). As expected, corn selected lines exhibited higher MFO activity than the bermudagrass selected line, or the control line. This suggests that MFO activity plays an important role in host plant adaptation in the fall armyworm.

103An explanation for the difference within the corn

strain in MFO activity, when reared on either corn or bermudagrass, could be a difference in inhibition of MFO activity by host allelochemicals. In that case activity would be more inhibited when reared on bermudagrass than when reared on corn. However, forage grasses, like bermudagrass, are thought to contain generally few active allelochemicals (Bernays & Barbehenn, 1987). In a survey of hydroxamic acids, which includes DIMBOA and MBOA and is one of the most important groups of allelochemicals in the grass family, Zuniga et al. (1983) found no detectable amounts of these compounds in bermudagrass (see also Niemeyer, 1988). It is, therefore, more likely that the host effects on MFO activity in the corn strain are due to induction of MFO activity by corn allelochemicals. Studies on levels of mRNA are needed to confirm this.

The fall armyworm feeds occasionally on other hosts with active allelochemicals, like cotton or soybeans (Pitre et al. 1983). It is, however, not known which strain feeds more on these hosts. It is possible that the higher MFO activity in the corn strain enables the corn strain to metabolize allelochemicals in these hosts better than the rice strain. For example, Krieger et al. (1971) found an increase of MFO activity with the number of plant families in use as hosts in several species of

104Lepidoptera. Gould (1979) showed that selection for increased survival in mites on cucumber leads to increased survival on potato and tobacco as well (although it leads to decreased survival on lima bean). It is possible that adaptation to a triterpenoid in cucumber leads to cross adaptation to the alkaloids and phenolics of potato and tobacco (Gould, 1988). However, MFO coexist in multiple forms and each form can differ in substrate specificity (Soderlund & Bloomquist, 1990). Thus, the MFO in the corn strain might have specific activities towards corn allelochemicals, but metabolize allelochemicals from other host less well. It is also possible that, depending upon which host larvae are reared, different forms of MFO are induced by allelochemicals. Studies on levels of mRNA could help to resolve this question. If induced MFO forms in the corn strain lead to overall increased metabolism of host plant allelochemicals, these forms are likely induced on other host plants as well. But if these forms can only metabolize corn allelochemicals, other MFO forms can be expected to be induced, when larvae are reared on different hosts. Molecular studies could also reveal structural differences between MFO forms between the two strains. If MFO are important in host plant adaptation, as the results suggest, additional molecular differences between the MFO of the different strains could well exist.

105In conclusion, because MFO activity seems to be

important in host adaptation in the two fall armyworm strains, the two strains seem to be a good system to study the molecular aspects of enzymatic adaptation to host plants.

References to Chapter 5Bernays, E. A. & R. Barbehenn. 1987. Nutritional ecology

of grass foliage-chewing insects. In: F. Slansky & J.G. Rodriguez [eds.], Nutritional ecology of insects, mites, spiders and related invertebrates. Wiley-Interscience, New York.

Gould, F. 1979. Rapid host range evolution in a population of the phytophagous mite Tetranychus urticae Koch. Evolution 33: 791-802.

Gould, F. 1988. Genetics of pairwise and multispecies plant-herbivore coevolution. In: K. C. Spencer [ed.], Chemical mediation of coevolution. Academic Press, San Diego.

Krieger, R. I., P. P. Feeny & C. F. Wilkinson. 1971.Detoxification enzymes in the guts of caterpillars: an evolutionary answer to plant defence?Science 172: 579-581.

Niemeyer, H. M. 1988. Hydroxamic acids(4-hydroxy-l,4-benzoxazin-3-ones), defence chemicals in the gramineae. Phytochemistry 27: 3349-3358.

Pashley, D. P. 1988. Quantitative genetics, development, and physiological adaptation in host strains of fall armyworm. Evolution 42: 93-102.

Pitre, H. N., J. E. Mulroony & D. B. Hogg. 1983. Fall armyworm (Lepidoptera: Noctuidae) oviposition: crop preference and egg distribution on plants.J. Econ. Entomol. 76: 463-466.

106Soderlund, D. M. & J. R. Bloomquist. 1990. Molecular

mechanisms of insecticide resistance. In: R. T. Roush & B. E. Tabashnik [eds.] Pesticide resistance in arthropods. Chapman and Hall, New York.

Zuniga, G. E., V. H. Argandona, H. M. Niemeyer & L. J. Corcuera. 1983. Hydroxamic acid content in wild and cultivated gramineae. Phytochemistry 22: 2665-2668.

VITA

Klaas Hendrik Veenstra was born on July 19, 1962 in Groningen, The Netherlands. He obtained his "V.W.O." diploma from Revius High School in Deventer, The Netherlands in 1981. In the same year he started to study biology at the Rijksuniversiteit in Leiden, The Netherlands. He graduated in November 1987 from the Rijksuniversiteit at Leiden and entered Louisiana State University in August 1988 to pursue a doctoral degree in Entomology.

107

DOCTORAL EXAMINATION AND DISSERTATION REPORT

Candidate: Klaas H. Veenstra

Major Field: Entomology

Title of Dissertation: HOST PLANT ADAPTATION IN THE TWO STRAINS OFFALL ARMYWORM (LEPIDOPTERA: NOCTUIDAE)

Approved:

Major PrpfemSor and Ch j.

Dean of the GraduatfeJSchool

EXAMINING COMMITTEE:

7

rj .7//: J ’/ J

P .Jvi

Date of Examination:

12 / 10 / 1993

Documents

Host Plant Adaptation in the Two Strains of Fall Armyworm