Effects of the pesticide dimethoate on a non-target generalist carabid, Pterostichus melas italicus (Dejean, 1828) (Coleoptera: Carabidae)

Abstract

This study draws attention to the toxicity of dimethoate at sub-lethal concentration to beneficial carabid beetles living in olive groves agroecosystem of Calabria, Italy. Short- and long-term effects of dimethoate on a non-target generalist predator in agroecosystems, Pterostichus melas italicus (Dejean, 1828) (Coleoptera, Carabidae) adults, were quantified by toxicity test, total haemocyte counts and morphometric analyses. In laboratory toxicity tests, beetles of both sexes exposed to field concentration of this toxicant showed a reduction of activity and a mortality of 10% after 72 h. Moreover, the impact of dimethoate on the total haemocyte counts (THCs) was recorded as a significantly lower number of circulating haemocytes in treated animals compared to controls at 48 h. Morphometric analyses showed that dimethoate caused long-term sub-lethal effects as a reduction of some morphometric parameters in P. melas italicus populations from two olive groves (treated and natural). In addition, a significant reduction in body size of females from long-term treated olive grove and a sexual dimorphism alteration were observed. As a result, data suggested that dimethoate may cause sub-lethal effects on this non-target carabid species.

Keywords:

• Pesticides
• morphometric measurements
• haemocyte count
• olive groves
• carabid beetles

Introduction

Agricultural management practices – especially the use of pesticides against a target pest – affect non-target species of agroecosystem communities. Carabid beetles are among the most important groups of beneficial arthropods in the agroecosystem food chain where they are predators of many pests (including aphids, lepidopterans, slugs and diptera). Previous studies refer to their economic relevance as natural enemies in agricultural fields (Bilde & Toft 1994; Kromp 1999; Warner et al. 2008; Zaller et al. 2009) and as environmental indicators (Kielty et al. 1996; Lovei & Sunderland 1996; Rainio & Niemelä 2003). Many factors influence their diversity and abundance within agricultural areas (Kromp 1999; Holland & Reynolds 2003). Field and laboratory tests show that pesticides affect ground beetles both directly and indirectly when exposed to treatment by contact and by ingestion (Holland & Luff 2000). Exposure causes lethal and sub-lethal effects both at the organism level (in behaviour or physiology) (Çilgi et al. 1996; Römbke & Heimbach 1996; Mauchline et al. 2004; Thorbek & Bilde 2004; Desneux et al. 2007; Lagisz & Laskowski 2007; Zaller et al. 2009) and at the population and community levels (Langan et al. 2001, 2004; Bel'skaya et al. 2002; Irmler 2003; Miñarro & Dapena 2003; Navntoft et al. 2006; Prasifka et al. 2008). Pesticides such as organophosphates have high acute toxicity and often lack physiological selectivity. Dimethoate acts as a broad-spectrum, contact and systemic insecticide at the nervous system level and it is responsible for the inhibition of cholinesterase. Usually organophosphates are non-persistent in the environment and do not typically bioaccumulate. Therefore only short-term effects have been thoroughly examined (Fischer et al. 1997; Kennedy et al. 2001; Coeurdassier et al. 2002; Mauchline et al. 2004).

Pterostichus melas italicus (Dejean, 1828) is a generalist predator, eurytopic and thermophilous species of clay soils that shows a bimodal period of activity. Adults appear in April and September and breed during autumn. While adults show activity throughout the winter on the soil surface, larvae hibernate. This species is very common in Calabrian (South Italy) olive grove agroecosystems and acts as a predator against insect pests (i.e. Bactrocera oleae). The experimental hypothesis is that generalist predators, such as P. melas italicus, survive pesticide exposure but show sub-lethal effects at the organism level. Such physiological effects might also cause disturbances in longevity, feeding behaviour, foraging activity reduction and immune capacity variation.

The specific objectives of this study are: (i) laboratory quantification of lethal effects (mortality) on P. melas italicus caused by the chemical treatment (dimethoate) used in olive groves against olive fruit fly Bactrocera oleae Gmelin, 1790 (Diptera, Tephritidae); and (ii) determining sub-lethal effects of pesticides on the physiology and development. Toxicity tests, total haemocyte counts and morphometric measurements were conducted in order to evaluate dimethoate short- and long-term effects on P. melas italicus.

Materials and methods

Experimental animals

All beetles were collected by bait traps containing an attractive substance such as vinegar and fruit juice.

For the laboratory toxicity tests and total haemocyte counts (THC), adult P. melas italicus were collected in an uncontaminated area located in the Botanical Garden of the University of Calabria, Italy (39°4′N and 16°2′E) in October 2008. Tests and control specimens were reared at a light regime of L8:D16, 70% r.h. and at a day/night temperature of 16/13°C and fed on homogenised meat ad libitum before and during the test.

For the morphometric measurements, specimens were collected in two olive grove areas of the Cosenza province (Calabria, Italy). The first area (control-A1) has never been treated with chemicals to control the olive pest Bactrocera oleae. The second area (treated-A2) has been sprayed with dimethoate twice a month from August to September for the past 5 years. In both fields two samples were collected during autumn 2005.

Laboratory toxicity tests

Laboratory toxicity tests were carried out in order to establish the relationship between concentration–mortality rate and dimethoate field concentration. Toxicity tests were performed in plastic boxes (10 cm × 20 cm) containing a 2-cm layer of soil. The working solution was produced by dissolving 0.0004 g/ml of technical dimethoate (Sigma) in distilled water. This concentration equates to the field concentration of dimethoate (i.e. 150 ml per ha of active ingredient in 100 l of water, from 265 g/l of commercial formulation). Before use, the soil was dried out at 100°C overnight, cooled to room temperature and wetted with deionised water both for control and test. After rehydration, spreadability was increased by dissolving 200 μl of working solution in 10 ml of deionised water which was sprayed with a pipette onto the soil surface of each box.

Specimens (10 in the controls and 10 per treatment) were placed in each box and the test was replicated 4 times. The boxes were left uncovered during the test in order to simulate the natural evapotranspiration both of soil and animals. Once the test started, mortality and sub-lethal effects of treatment were assessed at 24 h intervals for 3 days. Three status categories of carabids were recorded: living, alive but immobile (knocked out) and dead. Data were reported in percentage.

Total haemocyte counts (THCs)

In total haemocyte counts (THCs), specimens were treated with dimethoate at field concentration. Forty-eight hours after the beginning of the test, individuals were cold anaesthetised and the last two abdominal segments laterally torn; a 26-gauge needle was inserted in the neck membrane and sterile phosphate-buffered saline (PBS, Sigma) slowly injected. When the first drop exited the tear in the abdomen of the two group animals (5 untreated and 5 treated adults), it was collected and haemocytes counted using a Bűrker's chamber. Normality of measurement data of THCs in untreated and treated animals was checked with a Shapiro–Wilk test and an F test was performed to check the homogeneity of the two variances. Non-parametric statistics, i.e. the Wilcoxon–Mann–Whitney rank sum test, were then used, since the F test null hypothesis could not be rejected. The box and whiskers plots were drawn with the boxplot command. Statistical analyses were performed using R version 2.9.2 software (R Development Core Team 2009).

Morphometric measurements

In morphometric analyses, 83 specimens of P. melas italicus were studied (control-A1 group formed by 38 specimens, 20 males and 18 females, while treated-A2 formed by 45 specimens, 17 males and 28 females). Animals were stored in alcohol (70%). Photographs were taken with a stereoscope (Zeiss Stemi SV 11Apo) and acquired by Matrox PC-VCR software (for Windows® 2000). Body length (mm), elytra length (mm), head width (mm) and prothorax width (mm) were measured for each individual. Measurements were taken using Sigma Scan Pro 5 Software (SPSS® Inc.).

For each morphometric variable, measurements were grouped by ‘control-A1 female’, ‘treated-A2 female’, ‘control-A1 male’ and ‘treated-A2 male’ checking normality of measurement data with a Shapiro–Wilk test. Homogeneity of variance across groups was checked with a Levene test. No significant deviation was observed either from normality or from homogeneity of variance in body length measurements, thus a one-way analysis of variance (ANOVA) was conducted, followed by post-hoc Tukey's HSD (Honestly Significant Difference) multiple comparison test. Elytra length, head and prothorax width data were measured using non-parametric statistics, i.e. Kruskal–Wallis rank sum test with post-hoc Wilcoxon rank sum test pairwise comparisons with Bonferroni correction, since the null hypothesis of the Shapiro–Wilk and/or the Levene tests could not be rejected. Associations among paired morphometric measurements were assessed using Pearson's product moment correlation coefficient. The box and whiskers plots were drawn with the boxplot command. Statistical analyses were performed using R version 2.9.2 software (R Development Core Team 2009).

Results

Laboratory toxicity tests

After a 48 h exposure, field application rate of dimethoate caused a 2.5% mortality rate of P. melas italicus adults and in 7.5% of specimens a reduction of normal activity (knocked out) was recorded. After a 72 h exposure, mortality was 10% whereas no mortality was observed in the control.

Total haemocyte counts (THCs)

The circulating haemocyte numbers in control and experimental groups are shown in Figure 1. After 48 h, THCs recorded in dimethoate-treated animals (n = 5, mean±SD = 2.2×10⁶±1.2×10⁶) showed a significantly lower count compared to control animal THCs (n = 5, mean±SD = 8.1×10⁶±5.4×10⁶; W = 25, p-value = 0.007937**).

Figure 1. Box and whiskers plots of THCs in P. melas italicus. ctrl: control; fc: field concentration.

Morphometric measurements

All correlation coefficients among morphometric measurements are shown in Table I. Body lengths were significantly different between groups (ANOVA: F = 10.403, d.f. = 3, p-value = 7.631e-6***) (Figure 2a). Similarly, groups showed highly significant differences in elytra length (Kruskal–Wallis: chi-squared = 30.7685, d.f. = 3, p-value = 9.51e-07***) (Figure 2b), head width (Kruskal–Wallis: chi-squared = 11.4785, d.f. = 3, p-value = 0.0094**) (Figure 2c), and prothorax width measurements (Kruskal–Wallis chi-squared = 11.8933, d.f. = 3, p-value = 0.007758**) (Figure 2d). All adjusted p-values of post-hoc pairwise comparisons are shown in Table II. P. melas italicus showed a highly significant sexual dimorphism in all variables and females always showed larger dimensions than males. Body and elytra length in control females also showed significant differences compared to measurements recorded in dimethoate-treated females. Morphometric data for dimethoate-treated females, instead, did not show significant differences compared to males coming from both areas (p > 0.9), except for elytra length, which shows an atypical trend (Figure 2b). In this case, treated males showed a significantly higher elytra length compared to males coming from the control site (p < 0.01). So, elytra length in dimethoate-treated females showed significant differences compared to control males (p < 0.001) but not to treated ones (p = 1).

Table I. Correlation coefficients among morphometric measurements and their significances using Pearson's product–moment correlation (*** means p < 0.001)

	Body length	Elytra length	Head width
Elytra length	0.60***
Head width	0.66***	0.51***
Prothorax width	0.65***	0.45***	0.84***

Table II. p-Adjusted values of Tukey's HSD (Honestly Significant Difference) multiple comparisons of means (body length) and Bonferroni p-adjusted values of pairwise comparisons using Wilcoxon rank sum test (elytra length, head and prothorax width). Groups are: ‘female from control-A1 area’ (f), ‘female from treated-A2 area’ (fD), ‘male from control-A1 area’ (m) and ‘male from treated-A2 area’ (mD)

Body length	f	fD	m
fD	0.0000***
m	0.0010***	0.9975
mD	0.0008***	0.9249	0.9765
Elytra length	f	fD	m
fD	0.0221*
m	0.0001***	0.0003***
mD	0.0631	1.0000	0.0023 **
Head width	f	fD	m
fD	0.2299
m	0.0041**	0.9975
mD	0.0256*	0.9249	0.9765
Prothorax width	f	fD	m
fD	0.0960
m	0.0067**	1.0000
mD	0.0143*	1.0000	1.0000

Figure 2. Box and whiskers plots of measured traits. (a) body length (mm), f: body length female from control-A1, fD: body length female from treated-A2, m: body length male from control-A1, mD: body length male from treated-A2; (b) elytra length (mm), f: elytra length female from control-A1, fD: elytra length female from treated-A2, m: elytra length male from control-A1, mD: elytra length male from treated-A2; (c) head width (mm), f: head width female from control-A1, fD: head width female from treated-A2, m: head width male from control-A1, mD: head width male from treated-A2; (d) prothorax width (mm), f: prothorax width female from control-A1, fD: prothorax width female from treated-A2, m: prothorax width male from control-A1, mD: prothorax width male from treated-A2.

Discussion

It is well known that agrochemicals have negative effects on natural enemies of insect pests and each species shows a different response related to its phenology and sensitivity (Kromp 1999; Holland & Luff 2000; Mulligan et al. 2006). Lethal and sub-lethal short-term effects on life cycle stages have been assessed by laboratory bioassays for many species as well as for ground beetles (Römbke & Heimbach 1996; Holland et al. 2000; Kennedy et al. 2001; Desneux et al. 2006).

However, many effects of pesticides on beneficial species are sub-lethal long-term effects at the population level (life span, fertility, sex ratio, feeding and reproductive behaviour) and cannot be quantified by laboratory tests (Stark & Banks 2003). The relatively low toxicity of organophosphates such as dimethoate to predatory carabids has been demonstrated in laboratory bioassays (Çilgi et al. 1996), but field studies have shown that insecticide application was mainly responsible for a decline in carabid species in agroecosystems (Kromp 1999; Navntoft et al. 2006; Nash et al. 2008).

This study quantifies the direct impact of this toxicant on individuals both in laboratory bioassay and in the field. In laboratory toxicity short-term tests, exposure to dimethoate causes a reduction of P. melas italicus activity and a mortality rate that reaches 10% after 72 h. Moreover, a significantly lower THC tendency is recorded in treated animals compared to control ones at 48 h, suggesting an immune disturbance induced by dimethoate, even if the dose is sub-lethal. Nevertheless, long time exposure causes more adverse effects on the physiology of non-target species such as Rhynocoris kumarii (Heteroptera, Reduviidae). A THC increase, correlated with the detoxification action of haemocytes, is recorded in this species (George & Ambrose 2004). Dimethoate concentration used in our trials is comparable to field exposure, so it is possible to say that specimens living in treated fields and field margins potentially suffer lethal effects via the direct and indirect exposure route of consuming contaminated preys, as observed in the carabid Bembidion lampros by Unal and Jepson (1991) and Pterostichus madidus, P. melanarius and Nebria brevicollis by Mauchline et al. (2004).

As far as long-term sub-lethal effects are concerned, the morphological variations occurring on the P. melas italicus populations of two olive groves (treated and natural) show that dimethoate causes a reduction in some of the contaminated specimen morphometric parameters. Interestingly, morphometric analyses show a significant reduction in the body size of females coming from olive groves that have been treated with dimethoate for many years. In this case, the overall size reduction in contaminated specimens can be a result of low food supplies due to prey mortality, since low food availability has an impact on larval development and on the endocrine system determining changes in adult size. Moreover, data suggest a sexual dimorphism alteration induced by dimethoate. Indeed, the morphometric comparison of treated females and males shows that females lose their normal body, prothorax and head larger dimensions, assuming masculine body proportions. This suggests that dimethoate is an endocrine disrupting chemical in this species. Reproduction of adult P. melas italicus occurs during autumn and larvae are active in the soil contaminated by dimethoate. They can, therefore, be exposed in their habitat both directly or indirectly to the action of the chemical. Dimethoate acts at the physiological level and can cause perturbations in development as reported in some studies (Coeurdassier et al. 2002; Desneux et al. 2007). The role of insecticides on arthropods' normal development has been described in a number of insect species (for a review see Desneux et al. 2007) and it can be hypothesised that a smaller body size in P. melas italicus females reduces mating success.

Finally, even if dimethoate does not typically bioaccumulate in the organism, long-term effects at the population level such as morphological variations could be due to habitat change. In fact, insects are sensitive to ecosystem changes or disruptions and their distribution is affected by a decreased number of suitable habitats or by habitats disturbed by anthropogenic activity, such as management practices. At the species level, changes in morphological characteristics can also indicate habitat quality (Magagula 2003; Weller & Ganzhorn 2004). Moreover, there are species-specific morphological peculiarities, which reflect the special demands of the ecological niche in ground beetles with similar body shape (Bauer & Kredler 1993; Bauer et al. 1998; Talarico et al. 2007). Some studies suggest that the variation observed in carabid body size along a gradient is related to habitat alteration caused by urbanisation (Magura et al. 2006) and by exposure to heavy metal pollution (Maryanski et al. 2002). Therefore, it cannot be excluded that a disturbance in the morphometric parameters of P. melas italicus is due to a complex of variables (for example: alteration of the origin sites and food availability) caused by management practices.

This study demonstrates the toxicity of dimethoate on a very common beneficial carabid species living in Calabrian olive grove agroecosystems. In terms of beetle mortality, dimethoate has minimal acute effects on P. melas italicus. However, an interesting dimorphism alteration is found, probably induced by dimethoate and further studies are needed to investigate the role played by this organophosphorous insecticide in endocrine-disrupting effects in carabids.

Acknowledgements

This paper is dedicated to the dear memory of Professor Tullia Zetto, who enthusiastically contributed to this study until her last days. Financial support was provided by grants assigned to T. Zetto from the Ministry of Education, University and Research (MIUR) and to A. Giglio from University of Calabria (Italy) (project ‘PRIN 2007 – Contributo valutazione positiva’). The authors thank Professor P. Brandmayr for his stimulating suggestions concerning the topic of carabid beetles and Dr Antonio Mazzei for supplying us with insects used for morphometric analyses.