Five transgenic rice lines, each containing an insecticidal toxin gene from Bacillus thuringiensis (Bt) under control of a different promoter, were tested for effects on two non‐target insects: the brown planthopper, Nilaparvata lugens (Stål) (Homoptera: Delphacidae), and its predator Cyrtorhinus lividipennis (Hemiptera: Miridae). Bt toxin was detected by ELISA in the honeydew of N. lugens that fed on rice lines with the CaMV 35S and actin promoters. Nilaparvata lugens produced greater volumes of acidic honeydew (derived from xylem feeding) on all five Bt rice lines than on non‐transgenic control lines. The amount of honeydew derived from phloem feeding did not differ between Bt and control lines. There were no differences between N. lugens reared on Bt and control lines in any of the five fitness parameters measured (survival to the adult stage, male and female weight, and male and female developmental time). There were no differences between C. lividipennis reared on N. lugens nymphs from Bt and control lines, in any of the three fitness parameters examined (survival to the adult stage and male and female developmental time). Our results indicate that N. lugens and its natural enemies will be exposed to Bt toxins from rice lines transformed with some Bt gene constructs, but that this exposure might not affect N. lugens and C. lividipennis fitness.

The receptor binding step in the molecular mode of action of five delta-endotoxins (Cry1Ab, Cry1Ac, Cry1C, Cry2A, and Cry9C) from Bacillus thuringiensis was examined to find toxins with different receptor sites in the midgut of the striped stem borer (SSB) Chilo suppressalis (Walker) and yellow stem borer (YSB) Scirpophaga incertulas (Walker) (Lepidoptera: Pyralidae). Homologous competition assays were used to estimate binding affinities (K(com)) of (125)I-labelled toxins to brush border membrane vesicles (BBMV). The SSB BBMV affinities in decreasing order was: Cry1Ab = Cry1Ac > Cry9C > Cry2A > Cry1C. In YSB, the order of decreasing affinities was: Cry1Ac > Cry1Ab > Cry9C = Cry2A > Cry1C. The number of binding sites (B(max)) estimated by homologous competition binding among the Cry toxins did not affect toxin binding affinity (K(com)) to both insect midgut BBMVs. Results of the heterologous competition binding assays suggest that Cry1Ab and Cry1Ac compete for the same binding sites in SSB and YSB. Other toxins bind with weak (Cry1C, Cry2A) or no affinity (Cry9C) to Cry1Ab and Cry1Ac binding sites in both species. Cry2A had the lowest toxicity to 10-day-old SSB and Cry1Ab and Cry1Ac were the most toxic. Taken together, the results of this study show that Cry1Ab or Cry1Ac could be combined with either Cry1C, Cry2A, or Cry9C for more durable resistance in transgenic rice. Cry1Ab should not be used together with Cry1Ac because a mutation in one receptor site could diminish binding of both toxins.

Insecticidal activity and receptor binding properties of Bacillus thuringiensis toxins to yellow and striped rice stem borers (Sciropophaga incertulas and Chilo suppresalis, respectively) were investigated. Yellow stem borer (YSB) was susceptible to Cry1Aa, Cry1Ac, Cry2A, and Cry1C toxins with similar toxicities. To striped stem borer (SSB), Cry1Ac, Cry2A, and Cry1C were more toxic than Cry1Aa toxin. Binding assays were performed with 125 I-labeled toxins (Cry1Aa, Cry1Ac, Cry2A, and Cry1C) and brush border membrane vesicles (BBMV) prepared from YSB and SSB midguts. Both Cry1Aa and Cry1Ac toxins showed saturable, high-affinity binding to YSB BBMV. Cry2A and Cry1C toxins bound to YSB BBMV with relatively low binding affinity but with high binding site concentration. To SSB, both Cry1Aa and Cry1Ac exhibited high binding affinity, although these toxins are less toxic than Cry1C and Cry2A. Cry1C and Cry2A toxins bound to SSB BBMV with relatively low binding affinity but with high binding site concentration. Heterologous competition binding assays were performed to investigate the binding site cross-reactivity. The results showed that Cry1Aa and Cry1Ac recognize the same binding site, which is different from the Cry2A or Cry1C binding site in YSB and SSB. These data suggest that development of multitoxin systems in transgenic rice with toxin combinations which recognize different binding sites may be useful in implementing deployment strategies that decrease the rate of pest adaptation to B. thuringiensis toxin-expressing rice varieties. Bacillus thuringiensis, a gram-positive, spore-forming bacterium, produces insecticidal crystal proteins called ␦-endotoxins during sporulation. B. thuringiensis ␦-endotoxins have been used as an alternative to chemical pesticides for managing insect pests. These proteins are toxic to a number of insect larvae in the orders Lepidoptera, Diptera, and Coleoptera (15, 50). B. thuringiensis toxin genes are currently being transferred to crop plant genomes to overcome field degradation problems of conventional B. thuringiensis applications and to improve the efficiency of B. thuringiensis toxins (2, 8, 10, 32, 45). However, the potential for insect resistance to B. thuringiensis toxins raises concern about their long-term effectiveness (27, 39, 41). Recently, high levels of insect resistance have been observed in Plutella xylostella from the field selection (1a, 7, 17, 38

Five entomopathogenic Hyphomycetes were tested under field conditions for biological control of brown planthopper, Nilaparvata lugens (Stiil), in rice. Suspensions of conidia of Metarhizium anisopliae (Metsch.) Sorokin, M. f1avoviride Gams & Roszypal, Beauveria bassiana (Bals.) Vuill., and Hirsutella citriformis Speare were applied at a rate of 4-5.10" conidia per ha. In addition, M. anisopliae and Paecilomyces Ii/acinus (Thorn) Samson were applied as preparations of dry mycelium at a rate of 1.5-2 kg/ha. Mortality due to fungus infection ranged from 63 to 98% 3 weeks after application. There were no consistent differences between fungus species. The mycelium preparation sporulated on the plant and was as effective as the conidia suspension in infecting brown planthopper. Hyphomycetous fungi should be evaluated further for control of brown planthopper in rice.

Bacillus thuringiensisBerliner isolates were detected in 57% of 801 samples of rice grain dust, soil, rice field arthropods, and miscellaneous habitats (rice straw compost and mammal faeces) collected at 100 sites in the Philippines. The collection yielded 3950 isolates ofB. thuringiensis(8.7 isolates/positive sample). Grain dust from rice mills was the richest source (63% of the samples were positive, with 10.2 isolates/positive sample), followed by rice field arthropods, soil, and miscellaneous habitats. Polyclonal antibodies to six δ-endotoxin groups (Cry1A, Cry1B, Cry1C, Cry1D, Cry1E, and Cry3A) were used in enzyme-linked immunosorbent assays (ELISA) to characterize the toxins produced by each isolate. Subsamples of isolates representing the diversity of isolate sources and δ-endotoxin profiles were bioassayed against the yellow stem borer,Scirpophaga incertulas(Walker) and striped stem borer,Chilo suppressalis(Walker). Eighteen isolates highly toxic to both species were selected for characterization of δ-endotoxin genes by polymerase chain reaction (PCR) with primers specific to 14 genes or gene subfamilies, and Western blotting with Cry2A antibodies. At least two novel δ-endotoxin genes, related tocry1Bandcry1F, were detected by DNA sequencing of PCR products.