Abstract

Investigations of host plant-pest interactions have rarely used plant material in which the phenotype could be linked directly to a known genotype. However, in recent years, Arabidopsis thaliana has been intensively studied by molecular biologists and, as a result, there are a large number of genetically characterised A. thaliana mutants available to the scientific community. Our preliminary laboratory investigations have shown that A. thaliana can be used as a host plant by Plutella xylostella (L.) (Lepidoptera: Yponomeutidae) and have demonstrated that larval period, pupal period and pupal weight measurements of P. xylostella developing on A. thaliana are comparable to those on Brassica rapa. These factors, combined with its relatively small size and easy growth, make A. thaliana a suitable model host plant for investigation of both genotypic and phenotypic factors which influence the biology of P. xylostella, the aim of our research.

Our initial investigations have used the A. thaliana lines Col-0 and Col-5. These lines are genetically very similar except that Col-5 is homozygous for a gl1–1 mutant allele that is not present in Col-0. This mutant allele results in an absence of trichomes on the leaves and stem of the plant, which are normally present in Col-0. In our investigations, it was found that female larvae took significantly longer to develop and produced lighter pupae on Col-0 compared with Col-5. However, this effect was not seen in male insects. In both males and females, there was no significant effect on pupal period. It was also found that adults that had developed from larvae on Col-0 produced significantly fewer eggs than those that had developed on Col-5.

Keywords

Plutella xylostella, Arabidopsis thaliana, trichome, development, IPM

Introduction

Successful integrated pest management (IPM) is, almost without exception, based on a large number of investigations of the biology of the pest and its interactions with its environment. One of the most important relationships for study is undoubtedly that between the pest and its host plant, but studies have often been limited by the inherent complexity of these interactions. Model systems using plant material that is genetically well characterised can allow investigation of chosen aspects of plant-pest relationships by limitation of the number of variables during experimentation.

A. thaliana, a member of the Brassicaceae, is one species for which there is a wide range of genetically characterised mutants available. Its relatively small genome (25,498 genes) with almost no repetitive DNA, combined with a compact size and easy growth have resulted in its intensive study by molecular biologists and sequencing of the entire genome (The Arabidopsis Initiative 2000). A. thaliana is not directly of agricultural significance itself, but may be suitable as a model system, as many aspects of the plant-pest relationship are conserved in a broad range of species (Mitchell-Olds 1999).

A. thaliana is known to be susceptible to attack by a significant number of commercially important pests of the Brassicaceae including Plutella xylostella. This pest is a global economic problem due to its large geographical distribution, voracious eating habits and the wide choice of commercial plants which are suitable hosts, including those of agricultural and ornamental importance. The aim of our research is to study the effects of A. thaliana phenotype and genotype on P. xylostella. The first objective was to assess the suitability of A. thaliana as a host plant for studying Brassicaceae-pest interactions, of which there has been some question, due to its comparatively small size and ephemeral nature. Therefore, initial investigations compared P. xylostella growth and development on A. thaliana with Brassica rapa var. pekinensis, a typical host plant species which is also commonly used as a laboratory host for P. xylostella rearing.

Our second objective was to examine the effect of host plant trichomes, or hairs, on the development of P. xylostella, using A. thaliana. A presence or abundance of trichomes has often been linked to increased resistance to phytophagous insects. Non-glandular trichomes, such as those found in many members of the Brassicaceae (Metcalfe and Chalk 1950), can affect pest activity in a variety of ways; from restriction of movement and prevention of settling for feeding (Palaniswamy & Bodnaryk 1994) to provision of support during oviposition (AVRDC 1987).

Morphology of trichomes varies considerably (Metcalfe and Chalk 1950). On A. thaliana leaves, trichomes usually consist of three branches radiating from a stem, however, on stems and sepals, unbranched spikes are more often found. Amongst brassicas, when present, the trichomes also usually have a relatively simple form (Gomez-Campo 1980).

Many different genes are known to affect trichome growth and development in A. thaliana (Marks 1997); one of the most significant of these is GLABROUS1 (GL1). This gene encodes a myb transcription factor whose function is believed to be restricted to controlling trichome development (Koornneef et al. 1982, Oppenheimer et al. 1991). The gl1–1 mutant allele contains a deletion that removes the entire coding region of GL1 and flanking promoter elements (Oppenheimer et al. 1991); phenotypically this results in lack of trichomes in areas other than the leaf margins.

Our studies to investigate the effect of trichomes on P. xylostella development have used the A. thaliana lines Col-0 and Col-5. These lines are genetically very similar except that Col-5 is homozygous for the gl1–1 mutant allele and therefore lacks trichomes that are present in Col-0 plants.

Materials and Methods

Plutella xylostella were kept as a continuous culture on Brassica rapa var. pekinensis at approximately 25°C, 16:8 light:dark cycle. The culture has been the University of Reading since 1994, when insects were provided from a twenty-six year old culture at IACR-Rothamsted. The first instar larvae used experimentally were from eggs oviposited on the inside of plastic containers.

Brassica rapa var. pekinensis were grown in glasshouse facilities, using John Innes No. 2 compost, at 22 ±5 °C with 16:8h light:dark cycle. Plants used experimentally were approximately 4 weeks old and at the 4–6 true leaf stage.

Arabidopsis thaliana plants were grown at 19°C, 16:8 h light:dark cycle in controlled environment conditions using a 6 parts John Innes No.2 compost: 6 parts John Innes Multipurpose: 1 part Perlite soil mix. Col-0 and Col-5 seed material was provided by Dr E. Holub at HRI-Wellesbourne, UK. Plants used experimentally had a rosette diameter of 5-8 cm and were 6-8 weeks old.

Plutella xylostella larvae were reared individually, from first instar, on whole plants (type varies according to treatment) these were arranged randomly and kept at 23-25°C 16:8 h light:dark cycle. The apparatus consisted of individual plants in 9 cm pots, enclosed on the top with inverted plastic tubs (approx. 9.5 cm diameter x 4.6 cm tall) from which the central bases had been removed and replaced with fine netting.

Observations made at 12 h intervals recorded length of larval and pupal period. Approximately 36 h after the onset of pupation, when larval segmentation was no longer visible, pupal weight was also measured.

Following emergence, adults were sexed (Kwapong 1997). Males and females from the same treatment, which had emerged within 12 h of each other, were paired in clear plastic containers and the number of eggs laid on the sides counted after 48 h.

Following preliminary investigations to assess the normality of the data and therefore the assumptions held, a two-way analysis of variance procedure (Genstat 5 1998) was used to compare pupal weights statistically. Larval duration from the start of the experiment and pupal duration from time point first observed as a pupa, were analysed using a Mann-Whitney U test (Genstat 5 1998). For these data, which were measured at 12 h intervals, midpoint values of these categories are recorded, e.g. 6.75 represents 6.5–7. Data concerning male and female insects were analysed separately. Data from individuals that did not reach adulthood were not included in the analyses.

Analysis of eggs laid by insect pairs was executed by two-sample unpaired t-tests, including assessment of variance between data sets (Genstat 5 1998).

Development of P. xylostella on Brassica rapa var. pekinensis was compared with that on A. thaliana line Col-0. Forty of each type of plant were used and, using the procedure outlined above, larval period, pupal period, pupal weight and number of eggs laid were recorded.

The development of P. xylostella on Col-0 and Col-5 A. thaliana was compared using the procedure outlined above. Twenty-one plants of each treatment were used and larval period, pupal period, pupal weight and number of eggs laid recorded.

Results

Of the forty larvae set up on either B. rapa or A. thaliana plants, thirty-seven and thirty-four, respectively, survived to adulthood.

Pupal weight and larval and pupal periods were not significantly different between treatments (P>0.05) for either male or female insects (Table 1). There was no significant interaction between insect sex and treatment for pupal weight.

Treatment had no significant (P>0.05) effect on the number of eggs laid by insect pairs over the 48 h period (variance between the populations was not significantly unequal, P>0.05). Insect pairs which had developed on B. rapa and A. thaliana laid on average 87.7 ±8.1 standard error (SE) and 84.5 ±5.3 SE eggs respectively.

Of the twenty-one larvae set up on either Col-0 orCol-5 plants, seventeen and eighteen, respectively, survived to adulthood. When the duration of the larval period for individual sexes was compared there was a significant treatment effect (P<0.01) on females, but not males (Table 2). However, the duration of the pupal period did not differ significantly, in either male or female insects, between the treatments (Table 2).

Table 1. Summary statistics for development of Plutella xylostella reared on Arabidopsis thaliana versus Brassica rapa var. pekinensis

	Number of replicates	Larval period (days)a	Pupal period (days)a	Pupal weight (mg)b
Males
B. rapa	20	6.75 (6.25, 7.25)	4.75 (4.75, 5.25)	4.709 ± 0.092
A. thaliana	14	6.75 (6.75, 7.25)	4.75 (4.75, 4.75)	4.871 ± 0.102
Females
A. thaliana	20	7.5 (6.75, 7.75)	4.25 (4.25, 4.75)	5.626 ± 0.151
B. rapa	17	7.25 (6.75, 7.75)	4.25 (4.25, 4.75)	5.861 ± 0.186

*For insects of the same sex, significant difference in values between treatments at P= 0.05 level

**For insects of the same sex, significant difference in values between treatments at P= 0.01 level

^a Median (lower quartile, upper quartile)

^b Mean ± standard error

Pupae of females which had developed on Col-0 were significantly lighter (P<0.05), on average by 0.599 mg (0.298, 0.900 ± 2 standard error of difference (s.e.d.)), than those on Col-5. However, no significant difference was found between the two groups in males (Table 2) and no significant interaction was observed between insect sex and treatment.

Table 2. Summary statistics for development of Plutella xylostella reared on Arabidopsis thaliana: Col-0 vs. Col-5

	Number of replicates	Larval period (days)a	Pupal period (days)a	Pupal weight (mg)b
Males
Col-0 (+trichomes)	8	7.5 (7.0, 7.75)	5.25 (5.25, 5.25)	5.229 ± 0.089
Col-5 (-trichomes)	10	7.25 (6.75, 7.25)	5.25 (5.25, 5.75)	5.424 ± 0.104
Females
Col-0 (+trichomes)	9	7.75 (7.75, 7.75)**	4.75 (4.25, 4.75)	6.687 ± 0.113**
Col-5 (-trichomes)	8	7.25 (7.25, 7.5)	4.75 (4.5, 4.75)	7.286 ± 0.097

*For insects of the same sex, significant difference in values between treatments at P= 0.05 level

**For insects of the same sex, significant difference in values between treatments at P= 0.01 level

^a Median (lower quartile, upper quartile)

^b Mean ± standard error

The treatment was found to have a highly significant (P<0.01) effect on the number of eggs laid by insect pairs over the 48 h period (variance between the populations was not significantly unequal, P>0.05). Pairs, which had developed on Col-0, produced on average 23.8 eggs less (10.11, 37.56 ± 2 s.e.d.) during the period than those that had developed on Col-5. The insect pairs developing on Col-0 and Col-5 laying 112.5 ±4.4 SE and 136.3 ±4.4 SE eggs, respectively.

Discussion

These studies have demonstrated that several key developmental measures of P. xylostella raised on A. thaliana are comparable to those on B. rapa, a common host crop. This confirms a certain degree of suitability of A. thaliana as a model host plant for P. xylostella, in our future research. Under these conditions, larval duration is less than ten days, a time period where there is minimal physiological change in the plant. In different conditions, or should an alternate pest species be used however, the ephemeral nature of A. thaliana may become a significant issue.

The preliminary results to compare P. xylostella development on Col-0 and Col-5 A. thaliana suggest that the presence of trichomes has a negative effect on the development of the female larvae. If this is the case, the trichomes may be acting as a physical barrier decreasing the quantity of food that can be consumed in a given time period. This would result in an increase in the time needed to reach adequate nutrition for pupation and may lead to achievement of a lower overall nutritional status, reflected in the lower pupal weight and longer larval period observed.

Although studies have indicated no phenotype other than loss of trichomes associated with the gl1–1 mutant allele (Oppenheimer et al. 1991), it is possible that the results are actually due to a pleiotropic effect of the GL1 gene, which is affected by the gl1–1 mutant allele. A difference in the genetic background of Col-0 and Col-5 could also have had a significant effect. Current research is therefore using previously unavailable lines that do not differ genetically apart from the gl1–1 mutant allele, and hopes to address these queries.

These initial results do not indicate that either the gl1–1 mutant allele or the presence of trichomes affect the development time of male insects. However, further studies in both male and female insects will be necessary to confirm the results obtained.

The results show that the A. thaliana lines used affect the egg laying capacity of the insects within the 48 h period subsequent to pairing. Although oviposition can vary over time between treatments (Hillyer & Thorsteinson 1971), our studies (unpublished) have shown that, under these conditions, P. xylostella oviposit more than 99.99% of eggs in this period. These results may therefore reflect an overall difference in reproductive fitness between the test groups, but further investigation to compare egg viability and an increased sample size would be needed to validate this.

Differences in food quality and availability can result in a decrease of fecundity. Therefore, a lower fecundity in insects reared on Col-0 suggests that these plants are either of lower nutritional value than Col-5, possibly as a result of an undefined genetic difference between the two, or that the availability of the Col-0 as a food source was less than Col-5. Although in both treatments plant material was provided in excess of larval needs, if the trichomes on Col-0 acted as a physical barrier to food acquisition by the larvae, the actual amount of food available per unit time would be decreased. An interesting follow-up study might therefore investigate the quantity of leaf consumed and assimilated over time per treatment.

When experimental measures of P. xylostella raised on Col-0 in the two studies are compared, it is apparent that the results differ substantially. The cause is likely-to be the high sensitivity of P. xylostella development to temperature fluctuations. However, this emphasises the importance of randomization of treatments and standardisation if comparisons are to be made between experiments.

In summary, our investigations have shown a degree of suitability of A. thaliana as a model host plant for P. xylostella and preliminary investigations suggest that the presence of trichomes could have a significant negative effect on their development. However, further investigations are necessary to elucidate and confirm the exact targets of this treatment effect and link these indisputably with the gl1–1 mutant allele.

References