INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Bacillus thuringiensis is a gram-positive, spore-forming bacterium which, during sporulation, produces protein crystals. It is characterized as a widespread insect pathogen, and its insecticidal activity is attributed to the parasporal crystals. A variety of strains have been isolated from different habitats and, to date, more than 100 crystal protein genes have been sequenced (for a review, see reference (10). A classification with respect to amino acid sequence homology of the full-length crystal proteins is generally employed (2). The toxicity of these crystal proteins against certain insects and their high specificity led to the development of bioinsecticides for the control of pest insect species among the orders Lepidoptera, Diptera, and Coleoptera (10).

In general, most lepidopteran-specific B. thuringiensis toxins are known to be synthesized as protein crystals composed of protoxin molecules of ca. 130 to 140 kDa which, upon ingestion by larvae of a susceptible species, are dissolved by alkaline midgut fluid and proteolytically processed to an active toxin of ca. 60 kDa. Subsequently, the active toxin binds to specific receptors on the surface of midgut epithelial cells, followed by the insertion of the hydrophobic region of the toxin molecule into the cell membrane and formation of a transmembrane pore, which eventually results in cell lysis. Disruption of the gut epithelium leads to starvation, septicemia, and ultimately to death of the intoxicated larvae (5, 9).

To date, most in vivo studies have focused on this general mode of action. In the present study, we investigated some other aspects of the toxicity of Cry1 B. thuringiensis toxins both upon injection in life insects (Lymantria dispar [Lepidoptera: Lymantriidae] and Neobellieria bullata [Diptera: Sarcophagidae]) and in vitro on primary cultures of Lymantria brain cells. Cry1C, which was considered to be a lepidopteran-specific toxin for a long time, has recently been shown to also possess mosquitocidal activity (11). We demonstrated that, upon injection, additional targets for the Cry1C toxin reside in the insects body cavity of both the gipsy moth, L. dispar and the gray fleshfly, N. bullata. The evidence that we provide strongly suggests that the nervous system might be one of them.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Insect rearing. Eggs of the European gypsy moth, L. dispar, were obtained from the Biologische Bundesanstalt für Land und Forstwirtschaft (Darmstadt, Germany). L. dispar larvae were reared at room temperature on an artificial medium (100 ml contains 12 g of germ wheat, 12.5 g of milk powder, 0.8 g of Wesson salt mixture [ICN], 1 g of ICN vitamin diet fortification mixture, 0.2 g of benzoic acid, 0.1 g of nipagin, 1.5 g of Agar powder, and 80 ml of distilled water). Neonate larvae were grown in petri dishes (20 per dish).

Toxin preparation and quantification. The activated toxins were kindly provided by Aventis CropScience (Ghent, Belgium). Cry1Aa, Cry1Ab, Cry1Ac, Cry1C, and Cry1E toxins were prepared by proteolytic activation of protoxins with trypsin and subsequently purified by high-pressure liquid chromatography and quantified as described previously (14). Stock solutions were prepared in 20 mM Tris-HCl-150 mM NaCl (pH 8.6) and kept at 70°C. Dilutions were made to appropriate concentrations immediately prior to use in a saline buffer solution (NaCl, 154 mM; KCl, 2.68 mM; CaCl₂, 1.8 mM; NaHCO₃, 0.7 mM; D-glucose, 11.1 mM; pH 7).

Primary culture of Lymantria brain cells. L. dispar larvae were cold anesthetized, surface sterilized with 70% ethanol in water, and pinned down on a dissection cuvette with the dorsal side up. In a laminar flow cabinet and by using sterile dissection techniques, under direct observation through a stereomicroscope, the heads were cut open dorsally. The cerebral ganglia were carefully excised and transferred to a saline buffer solution (described above) containing 25 µl of gentamicin (Gibco)/ml. The brain was then transferred to a Nunc plastic culture dish (25 mm in diameter) in 100 µl of saline buffer solution. The perineural sheath was removed, and the cells were mechanically dissociated with the help of sterile microneedles. Each culture dish was left undisturbed for 15 min to allow the cells in suspension to settle and adhere to the bottom of the dish. After three rinses to remove unattached cells and cellular debris, the dishes were filled with 1 ml of a serum-free culture medium which consisted of equal parts of Eagle basal medium with Hanks' salts (Gibco) and Grace's insect medium (Sigma). All cultures were kept at 26°C in a humidified atmosphere. After isolation of the Lymantria neurons, the cultures were examined under an inverted phase-contrast microscope.

Assessment of neuronal viability. Cell viability in the presence or absence of B. thuringiensis toxins was determined by trypan blue (0.01 mg/ml, 2-min incubation) exclusion. The percentage of cell death in a culture dish was estimated by calculating the number of dead cells (N_d) divided by the total number of cells (N_t) times 100 (N_d*100/N_t).

Injection assay on L. dispar For the insect bioassays, freshly molted fourth- or fifth-instar larvae were placed individually in plastic vials (with perforated covers for aeration) together with 1 g of artificial food. The larvae received lateral injections at halfway the body length with a Hamilton syringe (20 µl). During the injection the needle was kept parallel to the cuticle to prevent any physical damage of the internal tissues. The different toxins to be tested were dissolved in saline solution to appropriate concentrations and administered in volumes of 2 µl of a given dose per 0.2 g (fresh weight) (g fresh wt) of the larvae. As a result, each larva was given the same dose in relation to its body weight. Control animals were injected with corresponding volumes of the saline buffer solution. To evaluate the effect of the injected toxins on inhibition of food intake in Lymantria, the individual larvae were weighed before (W₀) and 48 h after (W₄₈) injection. The growth of the larvae was expressed as the relative increase in fresh body weight as a percentage: [(W₄₈  W₀) × 100]/W₀. Correspondingly, the amount of food in each vial was weighed before (F₀) and 48 h after (F₄₈) injection. Weight loss due to evaporation of water was never >0.7% and therefore was not factored into the calculations. The relative food intake of each animal was calculated as the total fresh weight of food consumed over a 48-h period relative to W₀, i.e., [(F₀  F₄₈) × 100]/W₀. Data were analyzed with the Student's t test comparing the mean relative growth or relative food intake of the experimental group with a control group. As a benchmark measure for the toxic effect, we determined the effective dose that induced a toxic effect in 50% of the injected animals (ED₅₀). Estimation of ED₅₀ values and statistical analysis was done by probit analysis using the EPA Probit Analysis Program, version 1.5 (source: http://www.its.uidaho.edu/etox/resources.htm).

Pupariation assay. The N. bullata pupariation assay as described elsewhere (15) was used to evaluate the toxic effect of Cry1C. Zd'árek et al. (16) proved this assay to be a valuable and informative means for monitoring effects of drugs, venoms, and other neurotoxic compounds. These authors screened 62 neuroactive compounds and compared their pharmacological actions to their morphogenetic effects on pupation, which led to the following generalizations. Agents that paralyze neuromuscular systems at the peripheral level or suppress or modify basic motor patterns centrally cause the retention of larval morphologic characters in the pupae. Compounds that stimulate convulsive contractions of segmental musculature cause retention of larval segmentation on longitudinally contracted pupae.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of Cry1 B. thuringiensis toxins on food intake and relative growth of L. dispar larvae upon injection. A single dose of 5 µg of the activated Cry1C toxin/0.2 g fresh wt resulted in an irreversible inhibition of food intake in all of the injected animals. The animals also failed to defecate after injection. No lethal effect was observed during the first 48 h. After this period the treated animals gradually died, 60% within the first week after injection. Some of the remaining animals survived up to 2 weeks without resuming food intake. Whether the animals died from a direct toxic effect or starvation was not determined. In contrast, injection of 5 µg of Cry1Ac or Cry1E per 0.2 g fresh wt resulted in no effect after 48 h. An injection of 5 µg of Cry1Aa or Cry1Ab per 0.2 g fresh wt resulted in an inhibitory effect on the feeding behavior of L. dispar larvae but to a much smaller extent then Cry1C (Fig. 1).

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Effect of B. thuringiensis toxins on the relative food intake and relative growth of L. dispar fourth- and fifth-instar larvae. Larvae were injected with a dose of 5 µg/0.2 g fresh wt. The relative growth was calculated as the relative percent increase in the fresh weight of the larvae at 48 h postinjection. The relative food intake is the food intake as a percentage relative to the fresh weight of the larvae (n = 30). Values are means ± the standard deviations (**, P 0.01; *, P < 0.05) .

View larger version (9K): [in this window] [in a new window]   FIG. 2.   Dose-response curve of the Cry1C toxin on the relative growth of L. dispar larvae. Values are means ± the standard error of the mean (n  20).

Comparison of the toxic effect of Cry1 toxins upon injection and upon feeding. The toxicity data from force-feeding assays do not reflect the intrahemocoelic activity of the Cry1 toxins. When the animals were force-fed, the toxicity of Cry1C was much less pronounced (Table 1). Cry1Ab and Cry1Aa, which are particularly active on Lymantria in force-feeding assays, display only a moderate effect upon injection. Finally, the larvae showed very low sensitivity toward Cry1Ac and Cry1E both in force-feeding and upon injection.

Cry1C is toxic to brain cells of L. dispar in vitro The potent toxic effect of Cry1C, when injected into the hemolymph of Lymantria, suggests that other targets are present and accessible for this toxin after injection. We examined the effect of Cry1C on the viability of primary cultures of Lymantria brain cells (Table 2).

Cry1C toxicity to N. bullata larvae. The strong toxicity to Lymantria upon injection led us to perform the same experiment in a dipteran species. Injection of 5 µg of Cry1C/larva into the last larval instars of N. bullata was extremely lethal. It resulted in rapid death, within seconds. In addition, an immediate, extensive, local hemocytic melanization occurred. As a negative control, 5 µg of Cry1E/larva was injected; this had no observable effect on the maggots over a period of 24 h.

View larger version (29K): [in this window] [in a new window]   FIG. 3.   Effect on pupariation of Cry1C (0.05 µg/larva) injected into red spiracle larvae of N. bullata compared to controls. A total of 25% of the treated animals were not able to form a puparium and were therefore not taken into account.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The present study for the first time reports that some B. thuringiensis toxins are lethal to primary cultures of neuronal cells of an insect species. Previous reports have shown that Cry1C is toxic to various cell lines derived from Spodoptera frugiperda, Manduca sexta, Plodia interpunctuella (4), Aedes aegypti, and Anopheles gambiae (11). Spodoptera exigua cell lines show intermediate sensitivity, whereas the sensitivity of Mamestra brassicae and Drosophila melanogaster cell lines is low (6). Cells derived from Choristoneura fumiferana (4) and Culex quinquefasciatus (11) were insensitive.

The doses of Cry1C, used in this study on primary neuronal cultures, which provoke a toxic response of 100%, are considerably smaller (20 µg ml¹) than those reported in toxicity studies on cell lines of A. aegypti and A. gambiae (250 µg ml¹) (11) or S. frugiperda (50 µg ml¹) (6).

It has been known for a long time now that B. thuringiensis toxins, after oral administration, are species specific as well as target specific (the apical membrane of the midgut epithelium cells of susceptible insects). To date, most in vivo studies have focused on this general mode of action. A toxic effect of intrahemocoelic injections of crystals of B. thuringiensis was noted in Pieris brassicae as early as 1967 (7), but since then all attention has been paid to the toxicity of B. thuringiensis toxins after oral intake. Little is known about other putative target organs and cells in insects that are possibly accessible upon injection of the toxin. The potent effect of Cry1C upon injection in the hemolymph strongly suggests that a target other than the midgut epithelial cells must be present for this delta-endotoxin in the larvae of L. dispar. Peyronnet et al. (8) showed that neither Cry1Aa nor Cry1C had any depolarizing effect when applied on the basolateral side of the midgut membrane of L. dispar. On the other hand, Butko et al. (1) demonstrated a membrane-perturbing activity of Cry1C in receptor-free phospholipid vesicles. In vitro, and at low pH, Cry1C undergoes a conformational change that leads to membrane interaction and promotes flux of ions across the lipid bilayer (1). It is conceivable that, in vivo, a transition to a similar conformational state could be triggered by means of another, as-yet-unknown, physiological condition.

The intrahemocoelic activity, however, is not a general property of Cry1 toxins since Lymantria was only weakly sensitive to Cry1Aa and Cry1Ab and not sensitive at all toward Cry1Ac and Cry1E. Comparison of the toxicity upon injection of these toxins with data obtained in force-feeding assays (12, 13) shows quite the opposite pharmacologic profile for Cry1C, Cry1Aa, and Cry1Ab. Cry1Ac and Cry1E, on the other hand, are inactive in vivo both when force-fed and when injected. Moreover, a potent toxic effect was observed after injection of Cry1C into the fleshfly, N. bullata. In sublethal doses, the toxin affected pupation and gave rise to deformed pupae. The morphologic deformation (retaining of larval characteristics in the pupae) strongly resembles the one described earlier in flies treated with bee venom (16) and suggests a suppression of the peripheral neuromuscular system.

These data demonstrate that Cry1C inhibits food intake, and concomitantly growth of the insect, when injected into the hemolymph of Lymantria. Evidently, the mode of action after injection is quite different from the one that occurs after oral intake of the toxins. In vitro toxicity to Lymantria neurons suggests that the insect nervous system is a target for the toxin. In vivo experiments in the fleshfly support this view. According to Kwa et al. (6), sensitivity toward Cry1C of cell lines of S. frugiperda and S. exigua are correlated to the binding of the toxin to a 40-kDa protein. Whether the toxin-specific effects that we observed involve an aspecific mechanism or binding to a distinct receptor remains, however, inconclusive and needs further study. Its putative mode of action upon injection in the hemolymph will require more fundamental research. The question arises if the observed specificity of Cry1C toxicity after oral intake coincides with a specificity after injection.

The model proposed for Cry1C is that injected toxins may reach a neural target, as demonstrated by the good correlation between cytotoxicity and toxicity upon injection. On the other hand, Cry1Ac is nontoxic upon injection but quite cytotoxic to the neural cells in vitro. Apparently, the active Cry1Ac toxin does not reach the brain cells in vivo or the toxin is no longer active when it reaches the cells. Perhaps the toxin is not able to cross the blood-brain barrier or, in the hemolymph, it may be degraded, denatured, or aggregated in a nonfunctional form that would prevent binding or pore formation.

In conclusion, the results of this study suggest that B. thuringiensis toxin targets are not only present in the gut. This is an important result of biological relevance at the level of B. thuringiensis mode of action. Second, insects or other organisms that are susceptible to "injection" (for example, insects or parasitoids that may have ingested the toxin might inject it into their hosts or preys) may be affected. Therefore, the toxic effect described here may be relevant to the more general field of biocontrol, including trophic interactions. The benefit of, or the damage caused by, the phenomenon at the ecological level remains to be investigated. Also, the health protection of workers involved in B. thuringiensis spraying, etc., and the nonspecific effects on nontarget species need to be carefully scrutinized.