Effects of loline-producing endophyte-infected meadow fescue ecotypes on New Zealand grass grub (Costelytra zealandica)

Abstract

Loline alkaloids produced by endophyte-infected meadow fescue were high (>1000 µg/g) in roots. The larvae of New Zealand grass grub are major subterranean pasture pests. In laboratory and field studies, grass grub larvae feeding on roots of endophyte-infected meadow fescue containing lolines at concentrations >450 µg/g either lost weight or gained less weight than the corresponding controls feeding on endophyte-free or low-loline meadow fescue. These results demonstrate that loline alkaloids in roots of meadow fescue have the potential to deter grass grub larvae from feeding and could contribute to control of this pest in a sustainable manner.

Keywords:

• endophyte
• grass grub
• loline alkaloids
• meadow fescue
• Costelytra zealandica
• Festuca pratensis

Introduction

Grass grubs (Costelytra zealandica; Coleoptera: Scarabaeidae) are considered to be a major pest in improved New Zealand pastures. Third-instar larvae are most damaging to the roots of pasture plants in autumn and early winter, especially in the early stages of pasture establishment. However, subterranean insect pests are often difficult to control and have historically required the use of insecticides (Popay 1992). More environmentally acceptable control of insect pests using biological methods can be both effective and less expensive in the longer term (East et al. 1986). Chemicals released from endophytes of some temperate pasture-grass species have the potential to deter feeding of insect pests and provide a sustainable method of control (Popay & Latch 1993). Thus, active pest control agents in endophyte-infected plants present a targeted defence mechanism, usually in the form of naturally produced alkaloids such as ergopeptide alkaloids in perennial ryegrass (Lolium perenne) and meadow fescue (Festuca pratensis) (Cagas et al. 1999).

Perennial ryegrass and tall fescue (Festuca pratensis) are major grasses of agricultural ecosystems in temperate regions and are often infected with seed-borne fungal Neotyphodium endophytes (Easton & Fletcher 2007). The fungus benefits from access to plant nutrients and the infected plant benefits from a number of ecological advantages including resistance to insect herbivory (Richmond 2007). Such host–endophyte symbiotic associations produce a variety of alkaloids, including peramine, ergovaline, lolitrem B and lolines, which can affect insects (Siegel & Bush 1994). However, some alkaloids are toxic to livestock. Fescue toxicosis and ryegrass staggers are caused by ergovaline, lolitrem B and related alkaloids (Hoveland 2000). Meadow fescue, although a significant agricultural crop in Europe, is a minor pasture grass in New Zealand. It is often infected with the endophyte Neotyphodium uncinatum, which has the capacity to synthesise loline alkaloids (Blankenship et al. 2001) that are relatively non-toxic to grazing animals (Schardl et al. 2004; Gooneratne et al. unpubl. obs.). Loline alkaloids can act as toxins and/or feeding deterrents of insects (Jensen et al. 2009; Popay et al. 2009) and potentially can be used as insecticides (Siegel & Bush 1994) against pasture insect pests. The main loline alkaloid in pseudostems and leaves of meadow fescue, N-formyl loline (NFL) (Schardl et al. 2004), acts as a contact and an oral toxin (Bush et al. 1997). The concentration of loline alkaloids in roots is usually much lower than in the crown and shoots of the plant, but may be sufficient to protect against some insects (Bush et al. 1997).

A few studies have explored the action of loline alkaloids on insects. Patterson et al. (1991) showed that grubs of Japanese beetle (Popillia japonica; Coleoptera: Scarabaeidae) were deterred from feeding when offered artificial diets containing NFL and N-acetyl loline (NAL) at 100 µg/g dry weight. However, feeding deterrence did not occur when P. japonica grubs were offered washed tall fescue roots containing 93 µg/g NFL (Potter et al. 1992). Third-instar grass grub larvae in New Zealand fed roots of endophyte-infected meadow fescue in laboratory studies and in the field failed to gain bodyweight compared with those offered roots from plants without endophyte (Fletcher et al. 2000; Popay et al. 2003). However, the concentration of loline alkaloids in the roots of the endophyte-infected meadow fescue in the field study was not reported (Fletcher et al. 2000). A similar effect was observed in studies of grass grub larvae fed on loline alkaloids extracted from seed of tall fescue (Popay & Lane 2000).

In this work, laboratory and field studies were conducted to determine the effect of feeding roots containing known amounts of loline alkaloids in 12 endophyte-infected meadow fescue ecotypes to third-instar grass grub larvae. These studies challenge pre-existing knowledge on the relative concentration of loline alkaloids in roots of fescue required to have a toxic effect on grass grubs. The total root loline concentrations in the field experiment have been published previously in a paper focusing on increases in loline concentration damage induced by grass grubs (Patchett et al. 2008).

Method

Laboratory experiment

Third-instar grass grub larvae were collected from a pasture near Darfield, Canterbury, New Zealand, and placed individually into cells of plastic ice-cube trays. Only actively feeding larvae that consumed all of a fresh carrot cube (∼8 mm³) within 24 h were used in the assays. Four adjacent ‘cells’ were designated as a ‘plot’. The weight of grass grub larvae was considered as the mean of the weight of four grubs (i.e. from one ‘plot’). The experiments were set up as a randomised block design with six replicates (‘plots’).

Roots from 12 meadow fescue genotypes collected from a field near Templeton, Canterbury, were weighed (100 mg) and placed in the ‘cells’ with weighed larvae starved for 24 h. Larvae that were not offered roots acted as controls (no-feed control). Trays were covered with damp paper towels and enclosed in a black plastic bag and kept in an incubator at 16 °C.

After 4 days, faeces of grass grub larvae and remaining unused roots were separately weighed. All larvae were re-offered a weighed cube (8 mm³) of fresh carrot and, after 24 h, the carrot portion remaining was weighed. Roots from the same sample offered to the larvae were freeze-dried, ground through a 0.1 mm aperture sieve and stored at−20 °C for analysis of loline alkaloids.

Two separate experiments were conducted with the same protocol. Experiment 1 used six meadow-fescue endophyte lines plus a no-feed control. Experiment 2 used seven meadow-fescue endophyte lines and a no-feed control. Two of the meadow-fescue lines (lines Fp53 and Fp408) were common to both experiments. All the meadow-fescue endophyte lines were unique ecotypes as determined by amplified fragment length polymorphism (AFLP) (NE Cameron, Cropmark Seeds, pers. comm. 2007).

Field experiment

An irrigated-field experiment was conducted at Templeton, Canterbury, in February 2005. The site was previously sown with a fine fescue (Festuca rubra) and cultivated 6 weeks before transplanting. Two-tillered ramets from each of the 12 meadow fescue–endophyte combinations used in the laboratory experiments were planted in February 2005. This experiment was set up as a split-plot randomised block design, with each plot consisting of two plants (ramets) of one grass–endophyte combination, each separately enclosed by a 20 cm PVC plastic cylinder (internal diameter 16 cm). The plastic cylinders were placed over the plants and driven to ground level 1 month before the grass grubs were introduced. One subplot of each pair contained six third-instar grass grub larvae separately inserted 2.5 cm into the soil on 2 April 2005 to achieve a density equivalent to 300 grass grub larvae per m². The other subplot was free of grass grubs. Twelve grass–endophyte combinations were treated in this way, plus a no-grass-grub or blank control. There were eight replications of each genotype and control. Soil moisture (Quickdraw soil moisture probe 2900F, Soil Moisture Equipment Corp., Santa Barbara, CA 93105, USA) and temperature (Digital temperature probe Checktemp1, ScienceLab.com, Inc., Houston, Texas 777396, USA) at 10 cm depth was measured every second day throughout the experiment. Plants were removed on 11 May 2005, 39 days after introducing the grass grub larvae, complete with the plastic cylinder and enclosed soil. The grass grub larvae were recovered, counted and weighed. Plants were separated into shoots, crown and roots, weighed fresh, and samples taken for dry matter (DM) and loline analysis, but only the root data are reported here. Shoot and crown loline alkaloid concentrations of the same lines are reported elsewhere (Patchett et al. 2011). For loline analysis, root samples were pooled from the replicates, freeze-dried, ground and stored at −20 °C until analysis. Loline alkaloid analyses of each pooled sample were undertaken in duplicate. Samples for DM analysis were also pooled as for the loline alkaloid analysis and dried in a forced-draught oven at 60 °C for 48 h.

Loline analysis

The extraction of loline alkaloids from the plant samples was based on the method of Yates et al. (1990) modified by (Patchett et al.2011). Briefly, dried herbage (0.5 g) was shaken vigorously with 10 ml of dichloromethane: methanol: ammonia (75: 25: 0.5) solvent and 6 mg phenylmorphine (PM) per 100 ml of solvent as the internal standard, for 24 h. After centrifugation at 2000g for 15 min, 1 ml of the supernatant was taken up in a 1 ml plastic syringe and passed through a micro-filter (0.45 µm) into a glass gas chromatograph (GC) vial for loline alkaloid analyses within 24 h. Extraction, purification of loline dihydrogen chloride and derivatisation to NFL, NAL, N-methyl loline (NML) and N-acetyl norloline (NANL) was based on the methods of Petroski et al. (1989) and Blankenship et al. (2001) modified by (Patchett et al.2011). Loline alkaloid concentrations were calculated as µg/g DM of root biomass.

Statistical analysis

Results are expressed as mean±standard deviation (SD). In the laboratory experiment, the bodyweight of grass grub larvae, root consumed and carrot consumed, and the weight of faeces produced by larvae were subjected to analysis of variance (ANOVA) and means separated using least significant difference (LSD) (GenStat for Windows 6th edn 2002 version 6.1.0.2002; VSN International, Rothamsted, UK). In the field experiment, the initial and final bodyweights, change in weight during the experiment and the numbers of grass grub larvae recovered were subjected to ANOVA and the means separated using LSD. A paired t-test was used to analyse the effect of the presence/absence of grass grubs on the total loline alkaloid concentration in roots.

Results

Laboratory experiment 1

Root loline concentration

The total loline concentration in the roots of the six meadow-fescue lines varied markedly, ranging from 53 to 1047 µg/g (Table 1). The mean NFL concentration in them was 69.6% of the total loline>NAL (19.9%)>NANL (8.2%)>NML (2.2%).

Table 1  Food consumption, bodyweight change and faeces weight (mean±SD) of third-instar grass grub (Costelytra zealandica) offered loline-containing roots of meadow fescue (Festuca pratensis) for 5 days in two laboratory experiments. Total root loline concentration is the sum of N-formyl loline, N-acetyl loline, N-methyl loline and N-acetyl norloline. Diff refers to difference between initial and final bodyweights. Means with different superscript letters within a column are significantly different (LSD 5%).

		Weight consumed (mg)		Bodyweight (mg)
Treatment (meadow-fescue line ID)	Total root loline (µg/g DM)	Root	Carrot	Initial	Final	Diff (%)	Faeces weight (g)
Experiment 1
No-feed	—	0	22±3^ab	165±5	154±6^bc	−6.7	2±1^d
Fp53	53±1	102±8^a	8±5^c	166±3	172±10^a	3.6	83±16^a
Fp262	85±45	58±16^b	21±11^ab	164±8	162±5^b	−1.2	43±27^b
Fp391	581±8	44±13^b	15±8^abc	165±8	157±4^bc	−4.8	25±12^bc
Fp358	704±47	50±9^b	13±6^bc	162±4	149±6^c	−8.2	21±5^c
Fp345	778±21	41±8^b	23±6^a	165±4	157±4^bc	−4.8	21±6^c
Fp408	1047±65	37±12^b	18±10^ab	163±5	151±6^bc	−7.4	19±6^c
LSD 5%		29.0	9.3	—	7.8	—	10.5
P value		0.001	<0.05	ns	0.001	—	0.001
Experiment 2
No-feed	—	0	26±3	163±5	152±4^b	−6.7	0^b
Fp53	50±1	136±4^a	23±3	162±5	167±8^a	+3.1	75±33^a
Fp246	64±3	120±3^a	25±3	163±6	172±8^a	+5.5	51±16^a
Fp248	458±33	44±6^b	24±5	164±7	155±7^b	−5.5	16±5^b
Fp430	510±30	49±7^b	27±4	162±5	155±5^b	−4.3	13±5^b
Fp390	613±12	43±13^b	25±4	162±5	152±6^b	−6.2	8±4^b
Fp440	1010±64	37±37^b	25±2	161±5	149±4^b	−7.5	13±5^b
Fp408	1047±39	42±11^b	25±3	164±5	153±3^b	−6.7	12±3^b
LSD 5%	—	23.8		—	6.6	—	17.3
P value	—	0.001		—	0.001	—	0.001

Grass grubs

Grass grub larvae increased in weight when fed for 4 days on line Fp53 roots, a line of meadow fescue putatively free of endophyte and with a low total loline concentration (Table 1). Larvae in the no-feed control and those feeding on the roots of all other grass–endophyte combinations lost weight during the experiment. The weights of larvae feeding on roots of all treatments other than the Fp53 lines, after 5 days, were 1.2–8.2% less than the initial weight. Grass grub larvae consumed significantly more (P<0.001) of the nil-endophyte line Fp53 than all the other treatments, and there was no significant difference (P>0.05) in the weight of roots consumed by those feeding on all the other root treatments. The weight of faeces produced by grass grub larvae was related to the amount of roots consumed, with those feeding on line Fp53 producing more than twice the weight of faeces of any other treatment (P<.001). There were no significant differences in the weight of faeces produced by all the other grass grub larvae feeding on roots of other lines. However, larvae fed roots of line Fp262, which had a relatively low loline concentration (85±45 µg/g root DM), lost less bodyweight and produced more faeces than those larvae fed roots from lines with higher loline concentrations.

The grass grub larvae that lost weight feeding on the root diets (except from lines Fp391 and Fp358) consumed significantly more (P<.05) carrot post-exposure to meadow-fescue roots than larvae on line Fp53. Regression analysis failed to demonstrate a significant relationship between larval weight loss and root consumption, and faecal weight and root loline concentration.

Laboratory experiment 2

Root loline concentration

The relative and absolute concentrations of loline derivatives in roots were similar to experiment 1. The mean concentration of NFL, NAL, NANL and NML of the seven lines were 68.8%, 20.0%, 7.7%, and 3.3% of the total loline concentration respectively. In this experiment, although carrot was consumed by larvae in some treatments, moisture absorption confounded the carrot weights and therefore there was no significant difference in the amount of carrot consumed.

Grass grubs

The weight of roots consumed by grass grub larvae fed meadow-fescue lines Fp53 and Fp246 was significantly higher (P<.001) than those fed roots of other lines. Consequently, the larvae feeding on lines Fp53 and Fp246 produced significantly (P<.001) more faeces than those feeding on other lines and the no-feed control.

Weight gain/loss observed in grass grub larvae in experiment 2 (Table 1) was similar to the pattern in experiment 1. All grass grub larvae feeding on lines except Fp53 and Fp246 lost weight. Weight gains by grass grubs feeding on root diets (except lines Fp246 and Fp53, +3.1% and +5.5% respectively) were not significantly different from each other but were significantly higher (P<.001) than by those feeding on the roots of all the other lines.

The data on loline concentration, the amounts of root consumed, faeces produced and weight gain by grass grub larvae feeding on the two lines common to experiment 1 and experiment 2 (lines Fp53 and Fp408) and the no-feed control were similar in both experiments.

Field experiment

Soil moisture

The mean soil moisture tension throughout the experiment was−15.1 kPa (field capacity=−10 kPa) with a mean daily range of −4 to −25 kPa, sufficient for good plant growth. Soil temperature (at 10 cm) ranged between 4.2°C and 13.2°C during the experiment with a mean of 9.1 C. There was a sprinkling of snow on 24 April that covered the experimental area for a day and reduced soil temperatures on that day to 4.2°C.

Root loline concentration

The total loline concentration measured in the pooled root samples varied markedly (Table 2) with NFL>NAL>NANL>NML (Table 3). The presence of grass grub larvae increased the concentration of, but not the percentage of, all loline alkaloid derivatives (Table 3) and the lines could be separated into two widely disparate groups: one group of lines (Fp53, Fp246 and Fp248) with a mean loline concentration of 89 µg/g or less, and the rest, for which the means were 1509 µg/g or greater (Table 2).

Table 2  Total loline concentration (± SD) in meadow fescue roots and weight change and number of live third-instar grass grub (Costelytra zealandica) larvae recovered following exposure to 12 lines of meadow fescue (Festuca pratensis) in the field over 39 days in autumn in Canterbury, New Zealand. Total loline concentration is the sum of N-formyl loline, N-acetyl loline, N-methyl loline, and N-acetyl norloline. Grass grub larvae results are the mean±SD of six replicates. Means with different superscript letters within a column are significantly different (LSD 5%). Fp53, Fp246 and Fp248 were excluded from the t-test comparing total loline concentration because of very low-loline concentrations.

	Total root loline concentration (µg/g DMmatter)		Grass grub larvae
Treatment	With grass grubs	Without grass grubs	Initial weight(g)	Final weight(g)	Weight change (%)	Number of grass grubs (final)^1
No-feed	—	—	138±7.9	164±7.9^e	18.8	3.3±1.6^d
Fp53	89±57	102±65	141±7.9	192±10.6^a	36.2	5.3±0.7^ab
Fp246	72±61	86±80	141±4.8	190±7.9^ab	35.8	4.6±0.8^abc
Fp248	84±49	123±75	145±6.1	196±10.6^a	35.2	5.3±0.7^ab
Fp408	1509±34	1444±58	139±7.9	181±7.9^bc	30.2	5.4±1.1^a
Fp87	1556±4	1334±126	142±7.9	165±2.9^e	16.2	4.1±1.1^bcd
Fp358	1732±135	1251±56	138±7.9	176±7.9^cd	27.5	3.6±1.3^cd
Fp391	1869±317	1368±103	138±7.9	178±7.1^cd	29.0	4.6±0.8^abc
Fp345	1871±93	1474±78	139±4.5	171±13.2^de	23.0	4.0±1.6^cd
Fp262	1910±133	1330±3	144±4.0	172±7.9 ^cde	19.4	5.4±1.0^a
Fp440	1920±19	1725±18	138±7.9	165±7.9^e	19.6	4.3±0.8^bcd
Fp390	1924±85	1362±48	141±7.9	181±3.4^bc	28.4	4.3±1.6^bcd
Fp430	1937±366	1320±20	141±7.9	171±10.6^de	28.4	4.6±0.8^abc
Mean	1803	1401	140	177	—	4.5
Paired t-test	6.14, 8 d.f.	P<.001	—	—	—	—
LSD (5%)			5.4	9.5		1.2
P value			ns	<0.001		<0.01
^1Means +SD of number of grass grubs recovered at the end of field study of 8 replicates in each meadow-fescue line (treatment).

A paired t-test analysis showed the mean total loline concentration in the roots of meadow fescue subplots with grass grub larvae was higher (26%) than those subplots without (P<.001,Table 2). Lines Fp53, Fp246 and Fp248 were excluded from this analysis because of their very low loline concentration.

Table 3  Mean concentration of different loline alkaloids (µg/g DM) in the roots of nine lines of meadow fescue (Festuca pratensis) in the presence or absence of grass grub (Costelytra zealandica) larvae in the field for 39 days in autumn in Canterbury, New Zealand. Total loline concentration is the sum of NFL+NAL+NANL+NML.

	With grass grubs		Without grass grubs
Total loline alkaloid and derivatives in roots	Concentration (µg/g)	Total loline (%)	Concentration (µg/g)	Total loline (%)
N-formyl loline (NFL)	1302	72.2	1023	73
N-acetyl loline (NAL)	331	18.4	239	17.1
N-acetyl norloline (NANL)	148	8.2	117	8.4
N-methyl loline (NML)	22	1.2	22	1.5
Total loline*	1803		1401

Root weight

There were differences in root dry weight between the grass lines with grass grub larvae (1.33±0.3 g SD) and without them (2.22±0.42 g SD) (LSD 5%). More roots were consumed (55–65%) by larvae in grass lines Fp53, Fp246 and Fp248 with low loline concentrations than by those fed other lines (19–43%) that had high loline concentration, except for line Fp87 (54%).

Recovery of grass grub larvae

Of the grass grub larvae introduced to the field experiment initially, 88% were recovered at the end of the study. Of those recovered, 92% were visibly active and 8% were dead. There was a significant difference (P<.01) in the number of live grass grub larvae recovered between treatments. The lowest recovery (mean of 3.3 grass grubs/treatment) was from the no-feed control (Table 2).

Grass grub larvae in all treatments increased in weight (Table 2). The larvae in the no-feed control gained the least weight (18.8% increase) and the ones on the nil-endophyte and low-loline lines (Fp53, Fp246, Fp248) gained the most (+35.2–36.2%). There was a significant difference (P<.001) in grass grub larval weights between treatments, with those feeding on the high-loline lines showing lower weight gains (+16.2–30.2%). However, there was no difference in the final bodyweights between grass grub larvae in the no-feed control and those feeding on some of the high-loline lines (Fp345, Fp262, Fp440, Fp430).

Discussion

The total loline concentrations measured in roots of endophyte-infected meadow-fescue lines used in the laboratory and field experiments were up to 1937 µg/g, compared with previously reported concentrations of 282 µg/g in tall fescue (Bush et al. 1993), trace amounts reported by Justus et al. (1997) and 93 µg/g reported by Potter et al. (1992) in tall-fescue roots. Thus, it appears there is a considerable range in root loline concentrations depending on the specific endophyte–host combinations and environmental factors. Root samples for the laboratory experiments were collected in either early or mid winter, and this may have coincided with the time when alkaloids would have been translocated to the roots and hence exhibited higher root concentrations at this time of the year (Patchett 2007). Root samples collected in 2005 from the field experiment had higher concentration of lolines (up to 1937 µg/g) than the same lines collected for the laboratory assays from an adjacent area in the previous year (up to 1047 µg/g). Such annual variations occur and may be due to a variety of factors, including the age of the plant, time of year, site on which the plants were grown, site management and weather. Loline concentrations may also vary with the sampling method, extraction and analysis.

The feeding deterrence response to loline alkaloids in grass grubs in these experiments was similar to that previously reported (Popay & Lane 2000). The companion endophyte in the pasture lines and the alkaloids produced are likely to be important contributors to lower consumption of roots by grass grubs and the resultant weight loss (experiments 1 and 2) when grass grub larvae were offered a diet consisting of meadow fescue plants containing loline-producing N. uncinatum endophytes.

Both the laboratory and field experiments showed a negative relationship between the high total loline concentrations in roots and both root intake and weight gain by grass grubs. Similar decreases in weight gain in grass grub larvae when fed on roots of grasses containing N. uncinatum have been previously demonstrated (Fletcher et al. 2000; Popay & Lane 2000) in field studies but the presence of loline alkaloids in the roots of grasses was inferred and not measured. To the authors’ knowledge, this study is the first that demonstrates feeding deterrence in grass grub when exposed to roots of meadow fescue of known loline alkaloid concentration.

Total loline alkaloid concentration in meadow fescue roots varies considerably between grass lines during spring and summer, but increases markedly in all endophyte-infected lines during autumn (Patchett 2007). This study was conducted in late autumn (Table 2). A reduction in feeding by grass grubs in both field and laboratory experiments occurred when they were offered roots containing loline alkaloid concentrations>458 µg/g (Table 1 and 2). Patterson et al. (1991) showed that Japanese beetle (Popillia japonica) grubs (Coleoptera: Scarabaeidae) were deterred from feeding when offered artificial diets containing NFL and NAL at 100 µg/g. However, feeding deterrence did not occur when the grass grubs were offered washed tall fescue roots with 93 µg/g NFL (Potter et al. 1992). In the laboratory study reported here, when total loline concentration was high (>458 µg/g), there was no significant difference between treatments in grass grub weight, root consumption or faeces produced by the grass grubs over a wide range of alkaloid concentrations (Table 1). This suggests that there is probably a minimum (threshold) loline concentration that deters grass grub larvae from feeding on roots containing loline alkaloids. The threshold loline alkaloid concentration will vary according to the composition of the individual alkaloids. The value of 458 µg/g reported in this study is higher than the threshold concentration of 250 µg/g suggested by Popay & Lane (2000), but lower than the concentration of loline alkaloid NFL estimated by Popay et al. (2009) for control of Argentine stem weevil larvae. Root NFL concentration observed in the field study varied between 1203 and 1302 µg/g DM (Table 3). In the field, increasing the loline alkaloid concentration above a threshold may have no effect on grass grub bodyweight gain or loss as larvae cease or reduce feeding on loline-containing roots in preference to more palatable options (Sutherland 1972).

Endophyte infections have been shown to improve grass tolerance to leaf-, stem- and phloem-feeding insects (Breen 1994) but the effects of endophytes on root-feeding insects have been variable (Popay & Bonos 2005; Popay et al. 2009). This has been attributed to the lower alkaloid concentrations normally reported for roots compared with above-ground parts of the plant (Siegel et al. 1987). Lending support to this, despite visible differences attributable to root damage between the plants in some plots, a consistent effect of grass grub damage on plant shoot weight between different meadow-fescue lines was not observed (Patchett et al. unpubl. obs.). This may be due to the relatively low plant numbers used in this experiment (eight per subplot treatment) and the short duration of the experiment (39 days) relative to the natural feeding period of third-instar grubs (4–5 months). The high loline alkaloid concentrations (up to 1937 µg/g in line FP430) in the roots during autumn–winter over the 2 years of this study suggest a more important role for loline alkaloids in the protection of plant roots from subterranean insects than previously thought.

A non-direct effect of loline alkaloids on grass grubs may be mediated through grub nutrition. Nutritional stress has been shown to increase the incidence of disease in grass grubs (Popay 1992). In favourable conditions, grass grub larvae increase in weight rapidly during the autumn in preparation for cooler winter conditions. Any factor(s) that reduces weight gain predisposes these organisms to increased disease risk from a variety of pathogens (Faeth & Bultman 2002; Kain & Atkinson 1997) and increased mortality from several bacterial and protozoan pathogens (Robertson et al. 1999). The reduced weight gain in grass grub larvae feeding on roots containing loline alkaloids exposes these grass grubs to increased risk, resulting in reduced growth and population (Fletcher et al. 2000). A reduction in weight gain due to a reduction in feeding of roots with high loline concentrations by grass grubs suggests increased nutritional stress and vulnerability of grass grubs that can result in increased mortality (Table 2) caused by soil-dwelling entomopathogens. To test this, it would be interesting to conduct future experiments that specifically investigate the interaction of loline-producing plants/endophytes, soil-dwelling organisms including grass grubs and soil entomopathogens. Grass grub larvae in a controlled environment in the laboratory experiments were probably less exposed to soil-borne pathogens (although no check was done to verify this) and other risk factors because of the short-term exposure to loline alkaloids and therefore no deaths occurred. Although mortality of grass grub larvae exposed for 6 weeks to diets containing a combination of NFL + NAL up to 2000 µg/g did not increase at higher loline concentrations, the mean weight of the grass grubs exposed to the high root loline concentrations was approximately half of those fed on diets with loline concentrations of 50 µg/g or less (Popay & Lane 2000); this suggests a stronger deterrent effect rather than an acute toxic effect. In terms of managing pest damage, feeding deterrence may be as effective as mortality.

In conclusion, reducing growth, development and survival of grass grub populations with N. uncinatum endophytes that produce loline alkaloids adds to the growing evidence of insecticidal activity of these alkaloids and offers a potential method of pest control in pastures that is effective, inexpensive and has minimal environmental impact and negligible toxicity to pasture-grazing livestock (Schardl et al. 2007; Gooneratne et al. unpubl. obs.). With the new molecular biology techniques available, there is potential for pasture pest control to be further advanced with high-loline alkaloid pastures to include a wider range of hosts for endophyte inoculation and genetic manipulation of both the host and the endophyte.

Acknowledgements

Statistical advice from Dave Saville, Agresearch, Lincoln, and Alison Lister, Lincoln University, and financial support from Cropmark Seeds, Templeton, New Zealand, and the New Zealand Tertiary Education Commission are gratefully acknowledged.