Endophyte Isolate and Host Grass Effects on Chaetocnema pulicaria (Coleoptera: Chrysomelidae) Feeding

Transcription

1 PLANT RESISTANCE Endophyte Isolate and Host Grass Effects on Chaetocnema pulicaria (Coleoptera: Chrysomelidae) Feeding OLIVIER J.-P. BALL, 1,2,3 KIMBERLY D. GWINN, 3 CHARLES D. PLESS, 3 AND ALISON J. POPAY 4 J. Econ. Entomol. 104(2): 665Ð672 (2011); DOI: /EC10262 ABSTRACT Endophytic fungi belonging to the genus Neotyphodium, confer resistance to infected host grasses against insect pests. The effect of host species, and endophtye species and strain, on feeding and survival of the corn ßea beetle, Chaetocnema pulicaria Melsheimer (Coleoptera: Chrysomelidae) was investigated. The grassðendophyte associations included natural and artiþcially derived associations producing varying arrays of common endophyte-related alkaloids or alkaloid groups, peramine, lolitrem B, ergovaline, and the lolines. Preference and nonpreference tests showed that C. pulicaria feeding and survival were reduced by infection of tall fescue with the wild-type strain of N. coenophialum, the likely mechanism being antixenosis rather than antibiosis. In the preference tests, endophyte and host species effects were observed. Of the 10 different Neotyphodium strains tested in artiþcially derived tall fescue associations, eight strongly deterred feeding by C. pulicaria, whereas the remaining two strains had little or no effect on feeding. Infection of tall fescue with another fungal symbiont, p-endophyte, had no effect. Perennial ryegrass, Lolium perenne L., infected with six strains of endophyte, was moderately resistant to C. pulicaria compared with endophyte-free grass, but four additional strains were relatively inactive. Six NeotyphodiumÐmeadow fescue, Festuca pratensis Huds., associations, including the wild-type N. uncinatumðmeadow fescue combination, were resistant, whereas three associations were not effective. Loline alkaloids seemed to play a role in antixenosis to C. pulicaria. Effects not attributable to the lolines or any other of the alkaloids examined also were observed. This phenomenon also has been reported in tests with other insects, and indicates the presence of additional insect-active factors. KEY WORDS Neotyphodium, corn ßea beetle, insect resistance, endophyte, alkaloid Infection of some pasture and turf grasses with endophytic fungi belonging to the genus Neotyphodium (Ascomycota: Clavicipitaceae) has been found to confer resistance to the host against many insect pests (Popay and Rowan 1994, Popay and Bonos 2005). Being self-sustaining due to their vertical transmission, Neotyphodium endophytes are viewed as useful insect biocontrol agents. However, in pasture grass production systems, the usefulness of these endophytes is moderated by the fact that certain livestock disorders, including ryegrass staggers and fescue toxicosis, are a function of their presence (Schmidt and Osborn 1993, Fletcher 1993). These effects against insects and livestock are due to the presence of several alkaloids that are produced by the endophyte in association with the host grass (Gallagher et al. 1981, Siegel et al. 1990, Dahlman et al. 1991, Hill et al. 1994). In perennial 1 AgResearch, Grasslands Research Centre, Private Bag 11008, Palmerston North, New Zealand. 2 Current address: NorthTec, Private Bag 9019, Whangarei, New Zealand. 3 Department of Entomology and Plant Pathology, The University of Tennessee, Knoxville, TN Corresponding author: AgResearch, Ruakura Research Centre, Private Bag 3123, Hamilton, New Zealand ( alison.popay@ agresearch.co.nz). ryegrass, Lolium perenne L., and tall fescue, Festuca arundinacea Schreber, peramine, lolitrem B, ergovaline, and the lolines are the most commonly examined alkaloids or alkaloid groups. By basing selection of grassðendophyte associations on the presence or absence of certain alkaloids, resistance to insect pests can be maintained and health problems in livestock eliminated or much reduced. The corn ßea beetle, Chaetocnema pulicaria Melsheimer (Coleoptera: Chrysomelidae), is an economically important pest of maize, Zea mays L., in the United States that damages leaves and vectors the bacterial plant pathogen Pantoea stewartii. This insect also feeds on numerous other cereal and grass species. In late summer and fall, there is signiþcant movement of C. pulicaria from maize into the grass borders where they overwinter (Esker et al. 2002). Alternative host plants include tall fescue that is often infected with the endophyte Neotyphodium coenophialum (Morgan- Jones and Gams) Glenn, Bacon, and Hanlin. Kirfman et al. (1986) found that populations of the corn ßea beetle were often lower in pastures containing tall fescue infected with N. coenophialum than in pastures of predominantly endophyte-free tall fescue, indicating endophyte-mediated suppression. The aim of the current study was to conþrm the existence of an en /11/0665Ð0672$04.00/ Entomological Society of America

2 666 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 2 dophyte effect against C. pulicaria and to investigate the role in this of antibiosis (toxicity), antixenosis (feeding deterrence), alkaloids, endophyte species and strain, and host plant species. The endophyte species and strain determine the alkaloid proþle produced in the plant, whereas host plant genotypeð endophyte interactions determine the quantities produced (Lane et al. 2000). All of the common alkaloids have known insect activity, but there is considerable variation in insect response to different alkaloids (Popay and Bonos 2005). The effect of endophyte on C. pulicaria was investigated in preference and nonpreference feeding tests by using a variety of grassðendophyte associations producing variable alkaloid spectra. Three agronomically important grasses, tall fescue, perennial ryegrass, and meadow fescue, Festuca pratensis Huds., were used. The majority of plants tested were grown from seed from associations created artiþcially by inoculating seedlings (Latch and Christensen 1985) with different Neotyphodium strains. These strains were selected on the basis of their alkaloid proþle, agronomic promise, or scientiþc interest. Test plants also included commonly recognized natural associations of each host species, i.e., N. coenophialum for tall fescue, Neotyphodium lolii (Latch, Christensen, and Samuels) Glenn, Bacon, and Hanlin for perennial ryegrass, and Neotyphodium uncinatum (Gams, Petrini, and Schmidt) Glenn, Bacon, and Hanlin for meadow fescue. The effect of infection with p-endophytes (previously identiþed as Phialophora-like endophyte) (An et al. 1993) also was investigated in tall fescue. Although unrelated to Neotyphodium endophytes, the p-endophytes form symbiotic associations with their grass hosts, including perennial ryegrass, and they are also vertically transmitted via seed. Little is known of their insect resistance properties, but they can infect tall fescue either alone or in conjunction with N. coenophialum. Materials and Methods Collection and Maintenance of C. pulicaria. Adult C. pulicaria were collected during summer evenings from Johnson grass, Sorghum halepense (L.) Pers., plots at the University of Tennessee Plant Science Farm, Knoxville, TN, by suction trap. Adults were maintained on fresh Johnson grass clippings in the laboratory until use. Individual experiments were conducted using different beetles collected in several batches throughout the summer. Plant Material. The grassðendophyte combinations tested, and their respective alkaloid proþles, are shown in Table 1. All plants were maintained outdoors as described in Ball et al. (2006). Tiller material used in all preference and nonpreference tests was obtained from mature plants, the endophyte status of which were conþrmed using immunological and histological techniques (Ball et al. 2006). All tests included endophyte-free (E ) plant material from the same cultivars as used for the endophyte-infected (E ) material. Table 1. Host plant endophyte combinations used in C. pulicaria feeding tests, and their alkaloid profiles Host plant a (cultivar b ) IdentiÞer c Neotyphodium taxon d Original host e Alkaloid proþle f TF (K31) WT N. coenophialum TF P, EV, L TF (K31) AR501 FaTG-3 TF P, L TF (K31) AR502 FaTG-3 TF P, L TF (K31) AR506 FaTG-3 TF P, L TF (K31) AR508 Neotyphodium sp. TF P TF (K31) AR512 N. coenophialum TF P, L TF (K31) AR513 N. coenophialum TF P, L TF (K31) AR514 N. coenophialum TF P, L TF (K31) AR524 N. coenophialum TF P, L TF (K31) AR525 N. coenophialum TF P, L TF (K31) AR542 N. coenophialum TF P, L g TF (K31) P TF TF (K31) WT/P N. coenophialum TF P, EV, L TF (JS) WT N. coenophialum TF P, EV, L TF (JS) P TF TF (JS) WT/P N. coenophialum TF P, EV, L PRG (Nui) WT N. lolii PRG P, EV, LB PRG (Nui) AR17 Neotyphodium sp. PRG P PRG (Nui) AR19 N. lolii PRG P, LB PRG (Nui) AR20 N. lolii PRG P, LB PRG (Nui) AR21 N. lolii PRG P, LB PRG (Nui) AR22 N. lolii PRG P PRG (Nui) AR23 N. lolii PRG P, LB PRG (Nui) AR24 N. lolii PRG P PRG (Exp.) AR37 N. lolii PRG PRG (Exp.) AR501 FaTG-3 TF P, L h MF (Ensign) AR29 N. lolii PRG P, EV, LB MF (Ensign) AR501 FaTG-3 TF P, L MF (Ensign) AR506 FaTG-3 TF P, L MF (Ensign) AR512 N. coenophialum TF P, L MF (Ensign) AR548 N. coenophialum TF P, EV, L MF (Ensign) AR555 FaTG-2 TF P, EV, LB MF (Ensign) AR565 N. coenophialum TF P, EV, L MF (Ensign) AR583 FaTG-2 TF P, EV MF (Predix) WT N. uncinatum MF L a Plant species in which endophyte was tested. TF, tall fescue; PRG, perennial ryegrass; MF, meadow fescue. b K31, Kentucky 31; JS, Johnstone; Nui, Grasslands Nui; Exp., experimental line. c Endophyte strain or species identiþer. WT, naturally infecting wild-type Neotyphodium endophyte for the particular host; P, p-endophyte. d Christensen et al. (1993). FaTG-2, F. arundinacea taxonomic grouping 2; FaTG-3, F. arundinacea taxonomic grouping 3. e Host plant species from which endophyte originated. f P, peramine; EV, ergovaline; L, lolines; LB, lolitrem B. g Usual lolines (N-formyl and N-acetyl) absent, N-acetyl norloline present. h N-Formyl loline present only. Nonpreference Test. One nonpreference feeding test was performed to compare survival of, and feeding by, adult C. pulicaria conþned on ÔKentucky 31Õ (K31) tall fescue plants either naturally infected with wildtype N. coenophialum or uninfected. The basal 10 mm of a whole E or E tiller was placed in water in a 1.5-ml centrifuge tube, the open end of which was sealed with paraþlm. This was placed into a larger clear plastic tube (10.2 by 3.5 cm diameter) into which two adult C. pulicaria were introduced. A moistened disc of Þlter paper was sealed over the top of the larger tube with paraþlm. The two grass treatments were each replicated 16 times. An unfed control group consisting of 13 replicates not supplied with any grass material also was included. The bioassay was conducted at 26

3 April 2011 BALL ET AL.: EFFECT OF ENDOPHYTE ON C. pulicaria FEEDING C under a photoperiod of 16:8 (L:D) h. Survival of adults was assessed daily. After 4 d, feeding on the tillers was estimated as a score from 0 (no feeding) to 3 (maximum feeding) (Hardy et al. 1985). Daily survival data were analyzed by logistic regression and feeding scores using a MannÐWhitney U test. Preference Tests. The extent of feeding on leaf blade material from endophyte-free tall fescue, meadow fescue, and perennial ryegrass plants was compared with feeding on blades of similar age from Neotyphodium-infected plants of the same species and cultivar. The effect on feeding of the p-endophyte in tall fescue, both alone and in conjunction with N. coenophialum, also was investigated. Leaf blades from the youngest fully emerged leaf were cut as close to the ligule as possible and trimmed at the distal end to a length of 3.5 cm. One endophyte-infected leaf blade and one endophyte-free blade were taped to the base of a culture dish (60 mm in diameter). Between three and seven different host plant genotypes within the same cultivar were used in each test. A moistened Þlter paper disc was attached to the lid of the culture dish to maintain humidity, and seven to 10 adult C. pulicaria were introduced. There were 11 replicate dishes per endophyte treatment. After 48 h at 26 2 C in a photoperiod of 16:8 (L:D) h, the grass blades were examined under a dissecting microscope at 7 magniþcation. Extent of feeding on each blade was estimated as a score (as in the nonpreference test), and the number of feeding scars across three equally spaced transects (6 mm apart), and the number of fecal pellets, on each blade were counted. Data were analyzed using a Wilcoxon signed rank test, and pooled SEs of the differences in the feeding scores, feeding scar numbers, and fecal pellet numbers were calculated. Analysis of Herbage. Herbage samples from plants used in the preference tests were analyzed for peramine, ergovaline, lolitrem B, and loline alkaloids. All analytical methods have been described previously (Ball et al. 2006). Results Non Preference Test. Adult C. pulicaria mortality in the endophyte-free (E ) treatment over the4dwas low (Table 2). Survival of beetles conþned on wildtype N. coenophialum-infected (E ) tall fescue tillers declined steadily and by day 3 was signiþcantly lower than on E tillers. Survival of unfed control beetles was signiþcantly lower than survival of beetles con- Þned to E or E treatments on day 3. On day 4, survival in the unfed control was still lower than in the E treatments, but the difference was not signiþcant. The extent of feeding on E tillers was signiþcantly greater than on E tillers after 4 d (Table 2), even when regression analysis was used to account for the increased mortality of beetles on the E treatment. For the E tillers, regression analysis also showed that greater feeding was associated with increased survival of C. pulicaria (R 2 0.5, df 14, P 0.002). Table 2. Percent survival of unfed adult C. pulicaria (control) and percent survival of, and extent of feeding by, C. pulicaria confined to Kentucky 31 tall fescue either without endophyte (E ) or infected with wild-type N. coenophialum (E ) Feeding Treatment Survival (%) score a (SEM) Day 1 Day 2 Day 3 Day 4 Day 4 E tall fescue a b 90.5a 2.8a (0.1) E tall fescue b 37.5b 1.0b (0.2) Control c 19.0b a Score system based on that used by Hardy et al. (1985), where 0 is no feeding and 3 is extensive feeding. b Within each column, numbers with different letters after them are signiþcantly different at P 0.05 (logistic regression analysis for survival data; MannÐWhitney test for feeding scores). Preference Tests. The comparisons described in all preference tests are between endophyte-infected (E ) and equivalent endophyte-free (E ) blades for each association. Feeding scores, feeding scar numbers, and fecal pellet numbers were all lower on blades taken from all tall fescueðneotyphodium associations compared with endophyte-free controls with the exception of strains AR501 and AR502 (Table 3). Infection of K31 with AR502 did not signiþcantly inßuence C. pulicaria feeding or fecal pellet distribution, whereas feeding scar and fecal pellet numbers, but not feeding scores, were lower on AR501-infected K31. No signiþcant differences in the three parameters measured were observed between blades infected with the p-endophyte alone (both in K31 and JS tall fescue) and endophyte-free blades. Feeding scores, feeding scars and fecal pellet numbers were lower on blades of K31 and ÔJohnstoneÕ (JS) tall fescue infected with the wild-type strain of N. coenophialum alone and when co-infected with both the wild-type strain and the p-endophyte (Table 3). Infection of perennial ryegrass with AR17, AR22, AR24, AR37, AR501, and wild-type N. lolii signiþcantly reduced feeding scores, feeding scar numbers, and fecal pellet numbers compared with equivalent E material (Table 4). Infection with AR19, AR21, and AR23 had no effect on feeding or fecal pellet distribution. Feeding scores were lower in AR20-infected perennial ryegrass but not feeding scar or fecal pellet numbers. Feeding and fecal pellet deposition on meadow fescue leaf blades were lower in AR501-, AR506-, AR512-, AR548-, AR565-, and wild-type N. uncinatum-infected material (Table 5). AR583 reduced the feeding score and fecal pellet numbers only moderately but had a highly signiþcant effect on feeding scars. There were no effects recorded with AR29 and AR555. Herbage Analysis. With the exception of AR29 in meadow fescue, the alkaloid proþles of all grassðendophyte associations were true to type (Table 1, cf. Tables 3Ð5). No detectable levels of ergovaline were found in AR29-infected meadow fescue.

4 668 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 2 Table 3. C. pulicaria feeding scores, feeding scar numbers, and fecal pellet numbers, on endophyte-infected and endophyte-free blades of tall fescue, and mean concentrations of the alkaloids measured in each grass endophyte association used Alkaloid concn ( g/g) c Feeding score d Feeding scars e Fecal pellets f Cultivar a IdentiÞer b P EV LB L K31 AR * 33 88*** 122 K31 AR502 No sample K31 AR *** 2.9 7*** 28 22*** 88 K31 AR *** *** 39 46*** 123 K31 AR ( ) 0.7*** 2.5 6*** 27 12*** 74 K31 AR ( ) 0.0*** 2.8 0*** 29 5*** 81 K31 AR *** 2.9 2*** 27 9*** 89 K31 AR ( ) 0.1*** 2.8 1*** 26 9*** 79 K31 AR *** 3.0 2*** 30 11*** 56 K31 AR ( ) 0.7*** 2.7 4*** 28 14*** 53 K31 WT ( ) 0.3*** 3.0 1*** 23 5*** 69 K31 P K31 WT/P *** 2.9 3*** 32 12*** 72 JS WT ( ) 0.0*** 2.6 0*** 24 6*** 69 JS P JS WT/P ( ) 0.2*** 2.6 1*** 23 9*** 55 Pooled SED *, P 0.05; ***, P 0.001, Wilcoxon signed rank test. a K31, Kentucky 31; JS, Johnstone. b AR, endophyte strain or species identiþer; WT, naturally infecting wild-type N. coenophialum endophyte; P, p-endophyte. c Concentrations are means of all plants used in each test. P, peramine; EV, ergovaline; L, lolines; LB, lolitrem B. Values for the loline alkaloids are the sum of the concentrations of N-acetyl- and N-formyl-lolines. The presence of N-acetyl norloline is indicated by a. A standard for this compound was not available, thus concentrations were estimated using the parameters for N-acetyl loline. ( ), 30Ð100 g/g; ( ), 101Ð300 g/g. d Score system based on that used by Hardy et al. (1985), where 0 is no feeding and 3 is extensive feeding. e Mean of the total number of feeding scars across three transects per blade. f Mean of the total number of fecal pellets on each blade. Discussion These experiments have shown that C. pulicaria is adversely affected by a range of endophyte strains producing different alkaloids in three different grass hosts. They conþrm the deduction of Kirfman et al. (1986) that endophyte-mediated resistance was responsible for reduced numbers of C. pulicaria in the tall fescue pastures these authors observed. There is very little other published information on the effect of endophyte on this insect or on other chrysomelids. However, unidentiþed ßea beetles in the Alticinae sub-family to which C. pulicaria belongs also were reported to be less abundant on E than on E turf tall fescue plots (Davidson and Potter 1995). The tall fescue endophyte also seems to affect the developmental stages of another chrysomelid species, Oulema melanopus (L.) but not adult feeding preferences. Feeding damage and oviposition of O. melanopus was Table 4. C. pulicaria feeding scores, feeding scar numbers, and fecal pellet numbers, on endophyte-infected and endophyte-free blades of perennial ryegrass, and mean concentrations of the alkaloids measured in each grass endophyte association used Alkaloid concn ( g/g) c Feeding score d Feeding scars e Fecal pellets f Cultivar a IdentiÞer b P EV LB L GN AR g *** 2.9 5*** 26 10*** 48 GN AR GN AR * GN AR GN AR g 0 0.9*** 2.7 3*** 18 6* 28 GN AR GN AR *** 2.7 7** 18 6** 34 GN WT *** 2.8 9** 21 32* 53 EX AR *** 3.0 3*** 21 6* 26 EX AR h 1.1*** 2.8 7** 20 10*** 45 Pooled SED *, P 0.05; **, P 0.01; ***, P 0.001, Wilcoxon signed rank test. a GN, Grasslands Nui; EX, experimental. b AR, endophyte strain or species identiþer; WT, naturally infecting wild-type N. lolii endophyte. c Concentrations are means of all plants used in each test. P, peramine; EV, ergovaline; L, lolines; LB, lolitrem B. d Score system based on that used by Hardy et al. (1985), where 0 is no feeding and 3 is extensive feeding. e Mean of the total number of feeding scars across three transects per blade. f Mean of the total number of fecal pellets on each blade. g Detection uncertain, trace quantities if present. h Only N-formyl loline present.

5 April 2011 BALL ET AL.: EFFECT OF ENDOPHYTE ON C. pulicaria FEEDING 669 Table 5. C. pulicaria feeding scores, feeding scar numbers, and fecal pellet numbers, on endophyte-infected and endophyte-free blades of meadow fescue, and mean concentrations of the alkaloids measured in each grass endophyte association used Alkaloid concn ( g/g) c Feeding score d Feeding scars e Fecal pellets f Cultivar a IdentiÞer b P EV LB L EN AR EN AR ( ) 0.0*** 3.0 0*** 30 6*** 40 EN AR ( ) 0.1*** 2.7 0*** 20 4** 37 EN AR ( ) 0.0*** 3.0 0*** 33 3*** 50 EN AR ( ) 0.0*** 2.8 0*** 32 7** 40 EN AR EN AR ( ) 0.0*** 2.8 0*** 25 6* 28 EN AR * *** 25 9* 31 PR WT ( ) 0.1*** 3.0 0*** 32 3*** 44 Pooled SED *, P 0.05; **, P 0.01; ***, P 0.001, Wilcoxon signed rank test. a EN, Ensign; PR, Predix. b AR, endophyte strain or species identiþer; WT, naturally infecting N. uncinatum endophyte. c Concentrations are means of all plants used in each test. P, peramine; EV, ergovaline; L, lolines; LB, lolitrem B. Values for the loline alkaloids are the sum of the concentrations of N-acetyl- and N-formyl-lolines. The presence of N-acetyl norloline is indicated by a. A standard for this compound was not available, thus concentrations were estimated using the parameters for N-acetyl loline. ( ), 30Ð100 g/g; ( ), 101Ð300 g/g; ( ), 300 g/g. d Score system based on that used by Hardy et al. (1985), where 0 is no feeding and 3 is extensive feeding. e Mean of the total number of feeding scars across three transects per blade. f Mean of the total number of fecal pellets on each blade. either the same on wild-type E and E tall fescue and perennial ryegrass (Dubbert and Tscharntke 1997, Clement et al. 2007) or higher on the E tall fescue (Clement et al. 2009) in controlled trials. However, development of larval stages to the fourth instar was slightly longer and survival rates lower on E K31 than on E (Clement et al. 2009). Although Dubbert and Tscharntke (1997) did not detect any differences in adult feeding and oviposition between E and E tall fescue, the survival of pupae was lower for those reared on the wild-type E tall fescue. In the nonpreference trial reported here, there was a signiþcantly lower survival rate of C. pulicaria adults conþned on E wild-type tall fescue compared with E tall fescue. The relationship between feeding and survival established for the E tillers, as well as the feeding differences between E and E tillers and the high mortality in the unfed control group, indicates that antixenosis rather than antibiosis, was responsible for this effect. The greater survival of adults in the E treatment on day 3 compared with the unfed control group was most likely accounted for by the small amount of feeding that occurred on these leaves. The antixenic effect was conþrmed in the preference test. Tall fescue leaf blades from plants infected with the wild-type variety of N. coenophialum suffered less feeding damage and were associated with lower fecal pellet numbers (and therefore a reduction in the time spent on them by C. pulicaria), than leaf blades from E plants. The response of C. pulicaria to Neotyphodium endophyte infection in the preference tests was highly strain dependent. In each host, some endophytes conferred potent deterrent properties whereas others had little or no effect. Most of the Neotyphodium strains investigated in tall fescue deterred C. pulicaria, but one strain had no effect and one was only weakly efþcacious. In meadow fescue, two endophyte strains had no effect, one had a mild effect, and the remaining six conferred potent resistance. For perennial ryegrass, three strains provided no protection and a fourth provided only weak protection. The remaining endophytes in perennial ryegrass were associated with reduced feeding and fecal pellet numbers. Host species also seemed to inßuence endophyte efþcacy against C. pulicaria. Taking the number of feeding scars as the most accurate indicator of feeding, the percentage difference in feeding between the endophyte-infected and noninfected control was calculated for each tall fescue (excluding those with the p-endophyte), meadow fescue and perennial ryegrass association. On average, endophytes in meadow fescue and tall fescue caused a 78% (SEM 12) and 77% (SEM 8) reduction respectively in number of feeding scars compared with ryegrass for which there was a 48% (SEM 11%) reduction. A t-test showed a signiþcantly greater efþcacy of both tall fescue (P 0.045) and meadow fescue (P 0.026) compared with ryegrass. Clearly endophyte associations with meadow fescue and tall fescue have a stronger effect in general on C. pulicaria than those in ryegrass. However, this effect is not always apparent for the same endophyte strain in different hosts. For example, AR501, had only a mild effect in tall fescue, was moderately active in perennial ryegrass, and provided the strongest protection in meadow fescue. C. pulicaria response to two other endophytes, AR506 and AR512, similarly was stronger in a meadow fescue than in a tall fescue host. A comparable host effect was observed in testing of AR501 against fall armyworm, Spodoptera frugiperda (J.E. Smith) (Ball et al. 2006), a result that was attributed to the fact that meadow fescue contained higher concentrations (as measured by enzyme-linked immunosorbent assay) of mycelia than ryegrass, which in turn contained more than tall fescue (unpublished data). Christensen et al. (1997) also have observed that both endophyte and alkaloid con-

6 670 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 2 centrations for several strains tend to be higher in meadow fescue than tall fescue. This may in part also explain our results with C. pulicaria. However, the observation that the wild-type N. lolii strain in ryegrass was moderately active, whereas the very similar (if not identical) strain in meadow fescue, AR29, was inactive, shows that host grass effects are not uniform. AR29 was used in meadow fescue as a representative of the wild-type strains that naturally occur in perennial ryegrass. It produces ergovaline in its perennial ryegrass host but was not found to produce ergovaline in meadow fescue here. Ergovaline has been found previously at low levels in AR29-infected meadow fescue (Ball et al. 2006). The response of C. pulicaria to wild-type N. coenophialum in tall fescue did not seem to be affected by simultaneous infection with the p-endophyte. Infection of JS or K31 tall fescue with the p-endophyte alone also had no effect on feeding. The p-endophyte has not been implicated in antibiosis or antixenosis against invertebrates to date (Schmidt 1994, Ball et al. 2006). On the contrary, where there has been an insect effect, the presence of the p-endophyte in tall fescue and the closely related Gliocladium-like endophyte (also now called a p-endophyte) (An et al. 1993) in perennial ryegrass has been associated with attraction rather than deterrence or antibiosis (Gaynor et al. 1983, Latch et al. 1985). The feeding responses of C. pulicaria on the different ryegrass associations, and those of S. frugiperda on the same associations (Ball et al. 2006), generally correspond very closely. Strains AR19, AR20, AR21, and AR23 were relatively inactive against both insects, whereas AR17, AR22, AR37, AR501, and wild-type N. lolii conferred moderate-to-strong resistance. However, even though the tests on C. pulicaria and S. frugiperda were carried out at similar times by using largely the same plants, the responses of these insects to the different endophytes in tall fescue and meadow fescue were not similar, illustrating the diverse nature of endophyteðinsect interactions. What is the cause of the variation in bioactivity of the different NeotyphodiumÐgrass associations? Endophyte-mediated effects against insects in general have been attributed to the presence of alkaloids (Lane et al. 2000, Schardl et al. 2007) although more often than not, the speciþc alkaloids associated with an insect response have not been conclusively identiþed. The use in our study of different associations producing similar or different arrays of the commonly examined alkaloids, as well as the chemical analysis of material from these plants, enables us to make certain conclusions regarding the role of these alkaloids in endophyte-mediated resistance against C. pulicaria. Research on the grassðfungal symbiosis has focused on four alkaloids or alkaloid groups: lolines, ergot alkaloids, peramine, and lolitrem B. All four groups are known to affect insects, but speciþc insects show varying responses to these different alkaloids. Derivatives of loline alkaloids affect a wide range of insects (Yates et al. 1989, Patterson et al. 1991, Riedell et al. 1991, Dahlman et al. 1997, Dougherty et al. 1999, Wilkinson et al. 2000, Popay and Lane 2000, Jensen et al. 2009, Popay et al. 2009). Both peramine (Rowan and Gaynor 1986) and ergovaline (Popay et al. 1990) are strong feeding deterrents to the adult weevil Listronotus bonariensis (Kuschel), whereas ergovaline but not peramine is known to deter the beetle Heteronychus arator F. (Ball et al. 1997). Lolitrem B which has not been extensively tested against insects reduces growth and development of L. bonariensis larvae (Prestidge and Gallagher 1985). L. bonariensis is one insect which is known to be affected by all four alkaloid groups. Interestingly the adult weevil feeds in a very similar way to C. pulicaria, scraping the leaf surfaces to create a window effect. In this study, we have found that variations in presence of the commonly studied alkaloids do not entirely account for the variations in feeding deterrence to C. pulicaria, suggesting that other factors are involved. Concentrations of these alkaloids in the plants are also very important but, for this set of experiments, it is difþcult to relate actual concentrations to feeding. In part, this is because samples were bulked from individual plant replicates of each treatment for chemical analysis which does not take account of individual variability in alkaloid concentrations within a treatment due to plant genotypeðendophyte interactions (Ball et al. 1995a,b; Easton et al. 2002). In addition, the plant material harvested for chemical analysis included pseudostems that tend to have different concentrations of alkaloids than the leaf blades which were offered to C. pulicaria (Lane et al. 2000, Schardl et al. 2007). Nevertheless, with these limitations in mind, there seems to be a link between C. pulicaria feeding and the presence of a particular set of alkaloids, the lolines. This link was most obvious in the meadow fescue associations tested, but also in tall fescue where the majority contained loline alkaloids and were resistant. The role of lolines was further investigated using regression analysis of the percent reduction in number of feeding scars between endophyte-infected and endophyte-free associations of both tall fescue and meadow fescue against log-transformed loline alkaloid concentrations. Loline concentrations 300 g/g were clearly related to increasing negative effects of the endophyte on feeding. Below that concentration there was no clear effect of lolines, and other factors are likely to be involved in the endophyte-mediated resistance. For example, in tall fescue, there was one instance where lolines were present at low concentrations in a partially susceptible association (AR501), and another instance where lolines were absent in a moderately resistant association (AR508). This result with AR501 indicates that in the absence of other factors, a certain threshold level may need to be reached before strong deterrence is provided. Although AR542-infected tall fescue did not contain the lolines most commonly associated with N. coenophialum infection (N-acetyl- and N-formylloline), N-acetyl norloline was present. The moderateto-strong deterrence recorded in this association suggests that this loline also may be active. Unless there are overriding matrix effects, the deterrence mediated by AR508 cannot be attributed to the rather modest

7 April 2011 BALL ET AL.: EFFECT OF ENDOPHYTE ON C. pulicaria FEEDING 671 peramine content. The result with AR508 therefore indicates the presence of another endophyte-related factor not tested for in our analyses, which is active against C. pulicaria. An examination of the effect of the alkaloid concentration of different associations indicates that neither ergovaline nor lolitrem B is required to confer resistance (e.g., AR555 in meadow fescue, which produced both of these alkaloids but showed no effect on C. pulicaria). A regression analysis between peramine concentration and percent reduction in number of feeding scars also indicated that this alkaloid did not affect feeding. In ryegrass associations where there were high concentrations of peramine, resistance seemed to be strong. However, evidence for a peramine-mediated antixenic effect seems to be negated by the AR29 association with meadow fescue, which had no effect on C. pulicaria yet also produced high concentrations of peramine. Evidence for an additional factor affecting C. pulicaria also can be seen in the results with perennial ryegrass. Two associations (AR17, AR37) strongly reduced feeding but lacked the common alkaloids. AR37 produces epoxy-janthitrems (Tapper and Lane 2004) and has been shown to be active in grasses against a range of insect species, including H. arator (Ball et al. 1994), L. bonariensis (Popay and Wyatt 1995), S. frugiperda (Ball et al. 2006), mealybug Balanococcus poae (Maskell) (Pennell et al. 2005), and aphid Aploneura lentisci (Passerini) (Popay and Gerard 2007). The direct effect of epoxy-janthitrems on insects has yet to be determined. AR17 is not thought to produce epoxyjanthitrems but is also active against S. frugiperda (Ball et al. 2006). Clearly in some Neotyphodium-infected grasses, there are factors other than the alkaloids examined here that are involved in insect resistance. Factors involved in insect responses may not necessarily be alkaloids or only alkaloids. Recently, metabolomic analyses have shown plant-related metabolite concentrations can differ due to presence of endophyte in ryegrass and due to different endophyte strains, and that such differences may correlate with the presence of insects (Rasmussen et al. 2008a,b, 2009). In conclusion, this study has shown that knowledge of endophyte-mediated insect resistance mechanisms is incomplete. The insect resistance of novel associations must be based on experimental studies such as this and not on assumptions from known chemistry. Furthermore, this study, together with others (Ball et al. 1994, 2006; Popay and Wyatt 1995; Pennell et al. 2005), establishes that there are unknown insect resistance factors in some grassðendophyte associations requiring further elucidation. Acknowledgments We thank V. Gibson, G. Clark, K. Fraser, B. Tapper, and E. Davies for technical assistance and N. Cox for valuable assistance with statistics. References Cited An, Z.-Q., M. R. Siegel, W. Hollin, H.-F. Tsai, D. Schmidt, and C. L. Schardl Relationships among non-acremonium sp. fungal endophytes in Þve grass species. Appl. Environ. Microbiol. 59: 1540Ð1548. Ball, O.J.-P., M. J. Christensen, and R. A. Prestidge Effect of selected isolates of Acremonium endophytes on adult black beetle (Heteronychus arator) feeding. Proc. N.Z. Plant Prot. Conf. 47: 227Ð231. Ball, O.J.-P., G. A. Lane, and R. A. Prestidge. 1995a. Acremonium lolii, ergovaline and peramine production in endophyte-infected perennial ryegrass. Proc. N.Z. Plant Prot. Conf. 48: 224Ð228. Ball, O.J.-P., R. A. Prestidge, and J. M. Sprosen. 1995b. Interrelationships between Acremonium lolii, peramine, and lolitrem B in perennial ryegrass. Appl. Environ. Microbiol. 61: 1527Ð1533. Ball, O.J.-P., C. O. Miles, and R. A. Prestidge Ergopeptine alkaloids and Neotyphodium lolii-mediated resistance in perennial ryegrass against adult Heteronychus arator (Coleoptera: Scarabaeidae). J. Econ. Entomol. 90: 1382Ð1391. Ball, O.J.-P., T. A. Coudron, B. A. Tapper, E. Davies, D. Trently, L. P. Bush, K. D. Gwinn, and A. J. Popay The importance of host plant species, Neotyphodium endophyte isolate, and alkaloids on feeding by Spodoptera frugiperda (Lepidoptera: Noctuidae). J. Econ. Entomol. 99: 1462Ð1473. Christensen, M. J., A. Leuchtmann, D. D. Rowan, and B. A. Tapper Taxonomy of Acremonium endophytes of tall fescue (Festuca arundinacea), meadow fescue (F. pratensis) and perennial ryegrass (Lolium perenne). Mycol. Res. 97: 1083Ð1092. Christensen, M. J., O.J.-P. Ball, R. J. Bennett, and C. L. Schardl Fungal and host genotype effects on compatibility and vascular colonization by Epichloe festucae. Mycol. Res. 101: 493Ð501. Clement, S. L., L. R. Elberson, T. Miller, M. Kynaston, and T. D. Phillips Cereal leaf beetle: is performance affected by grass endophytes? Grassl. Res. Pract. Ser. 13: 345. Clement, S. L., V. L. Bradley, and L. R. Elberson Cereal leaf beetle colonises grass germplasm nurseries and impacts seed production activities. Forage Grazinglands doi: /FG-2009Ð RS. Dahlman, D. L., H. Eichenseer, and M. R. Siegel Chemical perspectives on endophyte-grass interactions and their implications to insect herbivory, pp. 227Ð252. In P. Barbosa, V. A. Krischik, and C. G. Jones [eds.], Microbial mediation of plant-herbivore interactions. Wiley, New York. Dahlman, D. L., M. R. Siegel, and L. P. Bush Insecticidal activity of N-formylloline, pp. 13Ð5 Ð 13Ð6. In XVIII International Grasslands Conference, Winnepeg, Manitoba, Canada. Davidson, A. W., and D. A. Potter Response of plantfeeding, predatory, and soil-inhabiting invertebrates to Acremonium endophyte and nitrogen fertilisation in tall fescue turf. J. Econ. Entomol. 88: 367Ð379. Dougherty, C. T., F. W. Knapp, and L. P. Bush Mortality of horn ßy larvae (Diptera: Muscidae) in bovine dung supplemented with ergotamine and N-formyl loline. J. Med. Entomol. 36: 73Ð77. Dubbert, M., and T. Tscharntke Do endophytic fungi protect their host grasses from herbivory? Mitt. Dtsch. Entomol. Ges. 11: 525Ð528. Easton, H. S., G.C.M. Latch, B. A. Tapper, and O.J.-P. Ball Ryegrass host genetic control of concentrations of endophyte-derived alkaloids. Crop Sci. 42: 51Ð57. Esker, P. D., J. Obrycki, and F.W.J. Nutter Temporal distribution of Chaetocnema pulicaria (Coleoptera: Chrysomelidae) populations in Iowa. J. Econ. Entomol. 95: 739Ð747.

8 672 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 104, no. 2 Fletcher, L. R Grazing ryegrass/endophyte associations and their effect on animal health and performance, pp. 115Ð120. In D. E. Hume, G.C.M. Latch, and H. S. Easton [eds.], Proceedings of the 2nd International Symposium on Acremonium/Grass Interactions, plenary papers, 4Ð6 February 1993, AgResearch, Palmerston North, New Zealand. Gallagher, R. T., E. P. White, and P. H. Mortimer Ryegrass staggers: isolation of potent neurotoxins lolitrem A and lolitrem B from staggers-producing pastures. N.Z. Vet. J. 29: 189Ð190. Gaynor, D. L., D. D. Rowan, G.C.M. Latch, and S. Pilkington Preliminary results on the biochemical relationship between adult Argentine stem weevil and two endophytes in ryegrass. Proc. N.Z. Weed Pest Control Conf. 36: 220Ð224. Hardy, T. N., K. Clay, and A. M. Hammond Fall armyworm (Lepidoptera: Noctuidae): a laboratory bioassay and larval preference study for the fungal endophyte of perennial ryegrass. J. Econ. Entomol. 78: 571Ð575. Hill, N. S., F. N. Thompson, D. L. Dawe, and J. A. Stuedemann Antibody binding of circulating ergopeptine alkaloids in cattle grazing tall fescue. Am. J. Vet. Res. 55: 419Ð424. Jensen, J. G., A. J. Popay, and B. A. Tapper Argentine stem weevil adults are affected by meadow fescue endophyte and its loline alkaloids. N.Z. Plant Prot. 62: 12Ð18. Kirfman, G. W., R. L. Brandenburg, and G. B. Garner Relationship between insect abundance and endophyte infestation level in tall fescue in Missouri. J. Kans. Entomol. Soc. 59: 552Ð554. Lane, G. A., M. J. Christensen, and C. O. Miles Coevolution of fungal endophytes with grasses: the signiþcance of secondary metabolites, pp. 341Ð388. In C. W. Bacon and J. F. White, Jr. [eds.], Microbial endophytes. Marcel Dekker, New York. Latch, G.C.M., and M. J. Christensen ArtiÞcial infection of grasses with endophytes. Ann. Appl. Biol. 107: 17Ð24. Latch, G.C.M., M. J. Christensen, and D. L. Gaynor Aphid detection of endophyte infection in tall fescue. N.Z. J. Agric. Res. 28: 129Ð132. Patterson, C. G., D. A. Potter, and F. F. Fannin Feeding deterrency of alkaloids from endophyte-infected grasses to Japanese beetle grubs. Entomol. Exp. Appl. 61: 285Ð289. Pennell, C.G.L., A. J. Popay, O.J.-P. Ball, D. E. Hume, and D. B. Baird Occurrence and impact of pasture mealy bug (Balanococcus poae) and root aphid (Aploneura lentisci) on ryegrass (Lolium spp.) with and without infection by Neotyphodium fungal endophytes. N.Z. J. Agric. Res. 48: 329Ð 337. Popay, A. J., and S. A. Bonos Biotic responses in endophytic grasses, pp. 163Ð185. In C. A. Roberts, C. P. West, and D. E. Spiers [eds.], Neotyphodium in cool-season grasses. Blackwell Publishing, Ames, IA. Popay, A. J., and P. J. Gerard Cultivar and endophyte effects on a root aphid, Aploneura lentisci, in perennial ryegrass. N.Z. Plant Prot. 60: 223Ð227. Popay, A. J., and G. A. Lane The effect of crude extracts containing loline alkaloids on two New Zealand insect pests, pp. 471Ð475. In V. H. Paul and P. D. Dapprich [eds.], 4th International Neotyphodium/Grass Interactions Symposium, 27Ð29 September 2000, Soest, Germany. Popay, A. J., and D. D. Rowan Endophytic fungi as mediators of plant-insect interactions, pp. 83Ð103. In E. A. Bernays [ed.], Insect-plant interactions 5. CRC, Boca Raton, FL. Popay, A. J., and R. T. Wyatt Resistance to Argentine stem weevil in perennial ryegrass infected with endophytes producing different alkaloids. Proc. N.Z. Plant Prot. Conf. 48: 229Ð236. Popay, A. J., R. A. Prestidge, D. D. Rowan, and J. J. Dymock The role of Acremonium lolii mycotoxins in insect resistance of perennial ryegrass (Lolium perenne), pp. 44Ð 48. In S. S. Quisenberry and R. E. Joost [eds.], Proceedings of the 1st International Symposium on Acremonium/Grass Interactions, 3 November 1990, Louisiana Agricultural Experiment Station, Baton Rouge, LA. Popay, A. J., B. A. Tapper, and C. Podmore Endophyteinfected meadow fescue and loline alkaloids affect Argentine stem weevil larvae. N.Z. Plant Prot. 62: 19Ð27. Prestidge, R. A., and R. T. Gallagher Lolitrem BÑa stem weevil toxin isolated from Acremonium-infected ryegrass. Proc. N.Z. Weed Pest Control Conf. 38: 38Ð40. Rasmussen, S., A. J. Parsons, K. Fraser, H. Xue, and J. A. Newman. 2008a. Metabolic proþles of Lolium perenne are differentially affected by nitrogen supply, carbohydrate content, and fungal endophyte infection. Plant Physiol. 146: 1440Ð1453. Rasmussen, S., A. J. Parsons, A. J. Popay, H. Xue, and J. A. Newman. 2008b. Plant-endophyte-herbivore interactions: more than just alkaloids. Plant Signal Beh. 3: 974Ð977. Rasmussen, S., A. J. Parsons, and J. A. Newman Metabolomic analysis of the Lolium perenne-neotyphodium lolii symbiosis: more than just alkaloids? Phytochem. Rev 8: 535Ð 550. Riedell, W. E., R. E. Kieckhefer, R. J. Petroski, and R. G. Powell Naturally-occurring and synthetic loline alkaloid derivatives: insect feeding behaviour modiþcation and toxicity. J. Entomol. Sci. 26: 122Ð129. Rowan, D. D., and D. L. Gaynor Isolation of feeding deterrents against Argentine stem weevil from ryegrass infected with the endophyte Acremonium loliae. J. Chem. Ecol. 12: 647Ð658. Schardl, C. L., R. B. Grossman, P. Nagabhyru, J. R. Faulkner, and U. P. Mallik Loline alkaloids: currencies of mutualism. Phytochemistry 68: 980Ð996. Schmidt, D Du nouveau sur les endophytes de la fetuque des pres. Rev. Suisse Agric. 26: 57Ð63. Schmidt, S. P., and T. G. Osborn Effects of endophyteinfected tall fescue on animal performance. Agric. Ecosys. Environ. 44: 233Ð262. Siegel, M. R., G.C.M. Latch, L. P. Bush, F. F. Fannin, D. D. Rowan, B. A. Tapper, C. W. Bacon, and M. C. Johnson Fungal endophyte-infected grasses: alkaloid accumulation and aphid response. J. Chem. Ecol. 16: 3301Ð3315. Tapper, B. A., and G. A. Lane Janthitrems found in a Neotyphodium endophyte of perennial ryegrass, pp In R. Kallenbach, C. J. Rosenkrans, and T. R. Lock [eds.], 5th International Symposium on Neotyphodium/Grass Interactions, 23Ð26 May 2004, Fayetteville, AR. Wilkinson, H., M. R. Siegel, J. D. Blankenship, A. C. Mallory, L. P. Bush, and C. L. Schardl Contribution of fungal loline alkaloids to protection from aphids in a grass-endophyte mutualism. Mol. Plant Microbe Interact. 13: 1027Ð Yates, S. G., J. C. Fenster, and R. J. Bartelt Assay of tall fescue seed extracts, fractions, and alkaloids using the large milkweed bug. J. Agric. Food Chem. 37: 354Ð357. Received 14 July 2010; accepted 22 December 2010.