INTRODUCTION

Many plants are protected from being eaten by their chemical components: feeding deterrents repel while noxious compounds may not prevent ingestion of a plant but later they influence the vitality of the consumer. They inflict damage on one or several subsystems of the organism which can weaken the animal, cause lethal effects or reduce its vitality and fertility. Many studies of essential oil and plant extract compounds from specific plants showed toxic, deterrent activity or repellent effects on a range of insect species (Regnault-Roger, 1997; Prabuseenivasan et al., 2006; Akinwumi et al., 2007; Ikbal et al., 2007; Koul et al., 2008; Soummane et al., 2011; Pandey et al., 2012; El-Gendy and Shaalan, 2012). The typical structure-activity studies in individual insect species provided only limited insights into the mode of action of these compounds. Treatments of flies with essential oils containing monoterpenes or the application of purified constituents resulted in visible symptoms (Ryan and Byrne, 1988), apparently from neurotoxic influences. The overt symptoms were hyperactivity, convulsions and tremors followed by paralysis (‘knock-down’)-similar to those produced by such insecticides as organophosphates and carbamates (Grundy and Still, 1985; Coats et al., 1991; Kostyukovsky et al., 2002).

Many of the deterrent plants belong to the family Apiaceae. They have received much attention at least partly owing to the presence of different kind of terpenes. It was found that the water extracts of fennel (Foeniculum capillaceum Gilib.), garden angelica (Archangelica pfficinalis Hoffm.), common caraway (Carum cavi L.), garden lovage (Levisticum officinale Koch.) and coriander (Coyiandrum sativum L.) significantly affected the feeding effectiveness as well as developmental potential of the Colorado potato beetle (Leptinotarsa decemlineata Say) (Wawrzyniak and Lamparski, 2006).

The seeds of the lovage plant (L. officinale Koch.) are commonly used as a spice. The yield of oil (w/w%) in different stages was in the order: immature fruit (1.5%) > green mature fruit (1.0%) > ripened fruit (0.6%) > flower (0.1%) (Mirjalili et al., 2010) deterrent and insecticidal effects of essential oil of lovage, against Locusts Locusta migratoria (Linneus). The GC-MS analysis of lovage essential oil revealed that the major constituents in oil was limonene followed by beta pinene, pentylbenzene, beta myrcene, ocimene and the hydrocarbons were also detected in high quantity. Limonene was the most abundant headspace component in lovage constituents and revealed a mixture of its enantiomers D and L-limonene. Enantiomers differ from each other only in stereo-configuration but their odors can be perceived as different by insects and mammals (Friedman and Miller, 1971; Leitereg et al., 1971; Heth et al., 1992; Taniguchi et al., 1992; Laska and Teubner, 1999a, b; Laska et al., 1999a, b; Laska and Galizia, 2001; Rubin and Katz, 2001). The higher concentration of lovage oil induced the lowest eaten area percentage (44.22%) of L. migratoria, followed by the D and L-limonene (18.89 and 18.79%), respectively. Antifeedant and protection activity percentage were increased by increasing the concentration. The highest mortality percentage (100%) of L. migratoria nymphs was recorded at 5000 ppm lovage oil while it was 93.1, 86.7 and 82.8% in the case of D-limonene, L-limonene and the mix of D and L-limonene, respectively. Ngoh et al. (1998) stated that Limonene had previously been considered as a potential pesticide for insects since benzene derivatives such as limonene were more toxic and repellent than other terpenes to Periplaneta americana. Ozaki et al. (2003) stated that D-limonene has strong oral toxicity in flies.

Major targets of many toxins, after their uptake in the body, are located in the Central Nervous System (CNS) and the peripheral nervous system (Ngoh et al., 1998) stated that the exact mode of action of the four contact toxicants (eugenol, methyl eugenol, safrole and isosafrole) on P. americana is unknown, the contact toxicity of these chemicals suggest that they can somehow penetrate the integument of the insects, to act presumably on the nervous or respiratory system to produce the killing effect. Ferber et al. (2001) reported that the different toxins can act at various levels of neuronal metabolism: on cell membranes and their ion channels or on synaptic mechanisms of the pathways of neural transmission, such as many pesticides do in insects. The effects can be over-excitation or depression of neural activity that is seen as abnormal motor output of the nervous system.

In larger insects such as locusts, specific physiological preparations or isolated parts of the CNS can be used to study the effects of neuro-active substances if they are super fused by different concentrations of plant extracts or fractions that may be noxious. We used a new preparation from the insect CNS which is very simple to prepare for testing neurotoxic effects: the isolated fifth abdominal ganglion (AG5) of adult specimens of Locusta migratoria. When this ganglion is separated from the neighboring ganglia (AG4 and AG6) by cutting the connectives, it releases a continuing ventilatory motor output in axons of peripheral nerves (Hustert and Mashaly, 2013). The isolated AG5 maintains this autonomous ventilatory motor output for hours if kept in physiological saline with sufficient oxygen supply via the “bath-just…” as shown previously for the more complex locust metathoracic ganglion (Bustami and Hustert, 2000).

In the current study we investigated the acute toxicities and effects of essential oil from lovage (Levisticum officinale Koch.) as insecticide containing monoterpenes. It was tested with increasing concentrations in saline for neural effects on the autonomous ventilatory rhythm of an isolated abdominal ganglion (AG5) which supplies the fifth abdominal segment of Locusta migratoria (Forskal). The inhibition by lovage oil was compared with that of its main components, the enantiomers of limonene (D-and L-limonene).

MATERIALS AND METHODS

Plant material: Seeds of lovage plants (Levisticum officinale Koch.) were harvested in an experimental farm (Department of Horticulture, in the Faculty of Agriculture, Moshtohor, Benha University, Egypt.

Extraction of the essential oil: Lovage essential oil was obtained by hydro-distillation with a Clevenger apparatus (Clevenger, 1928) for 2-3 h. First, 150 g of lovage seeds were powdered coarsely and then hydro-distilled on a sand bath. The collected oil was dried over anhydrous sodium sulfate, filtered, stored and later diluted for bioassays. The identification and quantitative determination of volatile components was performed by GC/MS analysis.

Experimental animals: A colony of gregarious Locusta migratoria (Forskal) was kept in the Institute of Zoology of the University of Göttingen (Germany), reared under crowded conditions and fed on wheat. Lighting had a 12:12 h light-dark cycle and temperature was kept at 32±2°C and 65±5% R.H.

Preparation of the test ganglion and nerve recording: Before isolating the abdominal ganglion (AG5) (Fig. 1) adult males of Locusta migratoria were chilled below 5°C for anesthesia. The head and the rear segments of the abdomen were cut off. The gut was removed and the dorsal part of the insect was cut away with lateral incisions. The ventral part containing the nerve cord was fixed with insect needles in a small petri-dish on Sylgard (Dow Corning) and flooded with saline. The second free abdominal ganglion (AG5) was carefully dissected out by cutting its connectives to the adjacent ganglion and its peripheral nerves at some distance from the ganglion. That severed also the tracheal supply entering at the ventral nerve2 (N2) but the ganglion was kept in a shallow bath of locust saline (pH 6.8, with sucrose (Clements and May, 1974) providing a sufficient oxygen supply.

Fig. 1:	An isolated locust abdominal ganglion provides a simple model for testing neurotoxicity. The regular ventilatory motor output activity is the monitor. Extracts from lovage and its major components D- and L-limonene (enantiomers) are neurotoxic. A short-term suppression and a long-term toxicity can be distinguished

Fig. 2(a-b):

Efferent patterns of the isolated fifth abdominal ganglion. Recordings from the median nerve and from one ventral nerve (nerve 2) of the isolated fifth abdominal ganglion (at room temperature of 23°C). (a) There are four motor units recorded in the median nerve (upper trace): two larger units with similar amplitude (marked with “i”) innervate the inspiratory muscles (M225) and two small units which innervate spiracular muscles in the sixth abdominal segment. In nerve 2 (lower trace) various slow and fast motor units supply dorso-ventral expiratory muscles of the fifth abdominal segment. (b) Simultaneous recording of tonic activity and alternating bursts in the median nerve and nerve 2 during a period of discontinuous ventilation

The typical output of the motoneurons for the ventilatory muscles in this segment alternates between inspiration and expiration (Lewis et al., 1973; Hustert, 1975). That rhythm was recorded routinely (monopolar suction electrodes with tapering tips drawn from thin plastic tubes) from the stumps of the paired ventral nerves 2 (N2; Fig. 2b) and from the unpaired posterior median nerve (MN; Fig. 2a). The main indicator of ventilation was the activity of the two large motoneurons with very similar amplitudes and frequency pattern in the posterior median nerve. These supply the pair of inspiratory muscles M225 (Snodgrass, 1935) of the next (6th abdominal) segment (Miller, 1960). In the same (median) nerve the only other two motoneurones (with very small amplitudes) supply the spiracle closer muscles in the 6th abdominal segment. Their irregular activity only rarely joins in the ventilatory rhythm.

All action potentials were pre-amplified (1000x) and the larger units selected after they had been recorded on magnetic tape or by digital computer recording (Datapac 2K2, Run Technologies). The average frequency of their firing rates (combined from both units) was recorded (1 min standard recording time). That inspiratory activity monitored the neural effects of substances applied at different concentrations on the locust ganglion. The inspiratory motor output of the ganglion displayed occasional changes from bursting to more tonic firing. Therefore, burst intervals were neglected and only the average frequencies over the one minute periods were taken for the analysis.

The saline in the bath could be replaced routinely and substances were added to saline to test their effect on the motor output of the ganglion.

Bath applications: For testing the effects of potentially neurotoxic substances contained in lovage (L. officinale Koch.), the essential oil was used or the pure allochemicals the enantiomers of limonene (D- and L-limonene, obtained from Sigma-Aldrich). These were diluted in locust saline (Clements and May, 1974) with the aid of small amounts of detergent (Tween 20 at 10 μg L^-1 in the stock concentration) and applied on the isolated ganglion in the bath (six concentrations: 0.25, 0.5, 1.25, 2.5, 5 and 10 μg L^-1) for the different test substances: (1) Lovage essential oil, (2) Saline with L-limonene, (3) Saline with D-limonene, (4) Saline with a 50% of D and L-limonene mixture and (5) As a control saline and detergent only at concentrations used for the extracts and substances.

The basic protocol was to record neuronal activity for at least 5 minutes (recording for analysis during the last minute), then to change the bathing fluid with the next and higher concentration of substance in saline. After a recovery time for five minutes, the effects of the new concentration were recorded. After each session of recording from a ganglion, going through all (six) different concentrations for one tested substance, the ganglion was washed in pure saline three times.

Statistical analysis: The inspiratory motoneuron activity (large spikes) was analyzed by counting the total number of spikes occurring within a 60 sec window for every concentration. From the spike intervals (t) the instantaneous frequencies (t^-1) were taken and sorted in 10 Hz intervals from 0 to 300 Hz in order to establish a frequency distribution for each record. For the sake of comparability, a normal distribution was assumed and then the standard deviation of the frequency distribution and the mean frequency were calculated.

Data from concentration-dependent responses (spikes in Hz) of the tested substances were transformed to square roots (Square root values were calculated (N sqrt = N/N max) and subjected to analysis using Wilcoxon Rank-comparison (Wilcoxon, 1945) of lower to next higher within each previous concentration. The values of the total number of spikes (N) the standard deviation (SD) and mean frequency (Mean F) are used to characterize the breathing activity during each record. Also, analysis of variance and “F” tests were used to compare between substances and using a computer software package, “costat” a product of Cohort software Inc., Berkeley. California. USA. Duncan's multiple range test (Duncan, 1955) was applied subsequently to differentiate between means.

The Square root characteristic values of each concentration were averaged between the five ganglia. Means of results for all substances were graphically displayed under “overview”, of the data for each individual ganglion under the respective substance category.

RESULTS

Effects of the lovage essential oil on the ventilatory output of the isolated AG5 of locusts: The application of different concentrations of the lovage essential oil (L. officinale Koch.) in saline changed the activity regularly. It induced strong depression effects (p<0.02) on the ganglionic motor output (Fig. 3).

Fig. 3(a-b):

Depression of neuronal activity in the isolated 5th abdominal ganglion by the potentially neurotoxic lovage oil. Each trace is a sample of spiking activity of the large inspiratory motor units during exposure to different oil concentrations. The sequence is from pure saline (track 1) over saline with just the detergent Tween 20 (track 2) to ascending oil in concentrations (tracks 3 to 8) and finally bathing in pure saline again (tracks 9/10)

The mean frequency of spikes fell to 44 to around 35 Hz (spikes sec^-1) at 0.25 μg of oil concentration (Fig. 4a, e). Higher concentrations resulted not only in significant decreases of the mean frequency of spikes but also in decrease of spike amplitudes (Fig. 3). At the maximum oil concentration (10 μg L^-1) the motoneurones of the isolated ganglion discharged at an average of 6 Hz (p<0.05) or less (Fig. 4a). Increasing the concentrations always resulted in step by step reduction of the ventilatory motor output as measured by the average AP frequency of the inspiratory motoneurons in the median nerve (Fig. 4a, e). In some preparations the motor output also switched from bursting activity to continuous tonic firing. Washing the ganglion with pure saline after the highest concentration had been applied resulted in gradual recovery of the average output from the ganglion (14.1 Hz after 5 min, 19.2 Hz after 30 min and 22.4 Hz after 1 h); (Fig. 4a, e). The spiking activity of the motor units never reached the pre-application level of excitation of 44.0 Hz. Control experiments with concentrations of only the detergent in saline (Tween 20 at 10 μg L^-1), which had been necessary to dilute the oil in saline, showed no significant changes (39.1 Hz in saline+tween20 vs. 44.0 Hz in saline only) of the motor output from the ganglion (Fig. 3).

Effects of the pure enantiomers D- and L-limonene on the ventilatory output of the isolated (AG5) of locusts: Limonene, a monoterpenoid, is the major component of the allochemicals in lovage seed (L. officinale Koch.) essential oil, as identified by GC/MS analysis.

Fig. 4(a-e):

Concentration-dependent decrease of inspiratory spike frequency. Motor output from the isolated fifth abdominal ganglion when treated with different concentrations (each experiment: n = 5) of extracts, (a) Essential oil of lovage, (b) Pure D-limonene, (c) Mixture of 50% D- and L-limonene, (d) Pure L-limonene. For A-D: Results of Wilcoxon-comparison (comparison of lower to next higher within each concentration) as tests of difference between the adjacent results in each series of tests. The letters above the standard deviation indicators are different when they are significantly different between neighboring columns and (e) Comparative graph of the effect of different substances on the inspiratory motor output. Differences between the effects of the substances were tested statistically at each concentration value for the output under saline and the different concentration in saline. concentration in saline. S: Saline only, S+T: Saline with the detergent tween, numerals: Concentration of the respective substances

It is a mix of the limonene enantiomers D- and L-limonene. The effects of pure D-limonene and L-limonene alone and the effect of their mixture (50% of D- and L-limonene) were recorded from the Median Nerve (MN) of the ganglion at different bath concentrations (0.25, 0.5, 1.25, 2.5, 5 and 10 μg L^-1) and after subsequent washing out with pure saline following the highest concentrations (Fig. 4b-e).

When D- and L-limonene were applied separately each produced strong depression effects on the ganglion motor output but D-limonene was more effective than L-limonene (p<0.05). Surprisingly, mixtures of 50% D- and L-limonene with the same final concentrations were less effective than application of the single enantiomers (Fig. 4b-e). All applied chemicals caused at increased concentrations in saline a significant decrease in the ganglion motor output (p<0.02). When after application of the highest limonene concentrations the lowest level of output had been reached, the subsequent washing out with pure saline reversed the effects but spiking activity and the amplitudes of spikes never gained the initial level completely.