ABSTRACTWe have designed a high-throughput system for the identification of novel crystal protein genes (cry) fromBacillus thuringiensisstrains. The system was developed with two goals: (i) to acquire the mixed plasmid-enriched genomic sequence ofB. thuringiensisusing next-generation sequencing biotechnology, and (ii) to identifycrygenes with a computational pipeline (using BtToxin_scanner). In our pipeline method, we employed three different kinds of well-developed prediction methods, BLAST, hidden Markov model (HMM), and support vector machine (SVM), to predict the presence of Cry toxin genes. The pipeline proved to be fast (average speed, 1.02 Mb/min for proteins and open reading frames [ORFs] and 1.80 Mb/min for nucleotide sequences), sensitive (it detected 40% more protein toxin genes than a keyword extraction method using genomic sequences downloaded from GenBank), and highly specific. Twenty-one strains from our laboratory's collection were selected based on their plasmid pattern and/or crystal morphology. The plasmid-enriched genomic DNA was extracted from these strains and mixed for Illumina sequencing. The sequencing data werede novoassembled, and a total of 113 candidatecrysequences were identified using the computational pipeline. Twenty-seven candidate sequences were selected on the basis of their low level of sequence identity to knowncrygenes, and eight full-length genes were obtained with PCR. Finally, three newcry-type genes (primary ranks) and fivecryholotypes, which were designatedcry8Ac1,cry7Ha1,cry21Ca1,cry32Fa1, andcry21Da1by theB. thuringiensisToxin Nomenclature Committee, were identified. The system described here is both efficient and cost-effective and can greatly accelerate the discovery of novelcrygenes.
Water influx via aquaporin directly determines necrotic cell death induced by the Bacillus thuringiensis
Cry toxin
