Bacillus thuringiensis (Bt) contains a variety of insecticidal proteins and other insecticidal activity factors, including insecticidal crystal protein (including Cry protein and Cyt protein), vegetative insecticidal protein, chitinase, thuringiensin, anticancer protein, zwittermicin A, phospholipase C, AiiA protein and InhA protein (Yu et al., 1997; Shao et al., 2001; Espinasse et al., 2002). Bt was found to have specific activity in Lepidoptera, Coleoplera, Diplera, Orthoplera, Homoptera, Hymenoptera and other insects, nematode, protozoa and cancer cell (Sun et al., 1996; Balaraman, 2005). As the most valuable biological pesticide, Bt has become a useful supplement to chemical pesticides, and is widely used in commercial agricultural production, forest protection and vector insect control. The most striking is the insecticidal gene from Bt, which is the most important genetic resources of genetically modified crops, has been successfully transferred into many important crops to make them have insect resistant ability (Estruch et al., 1997; Schnepf et al., 1998; Kumar et al., 2008).

It is very important to explore more new insecticidal gene resources in Bt research. In recent years, identification and cloning of novel genes with specific activity, such as parasitic nematodes and cancer cells, has attracted much attention. In addition, identification and cloning of novel insecticidal genes is an effective way to overcome the resistance of insects. However, it is more and more powerless to find new insecticidal gene only through the traditional PCR and molecular hybridization and other methods when a large number of insecticidal genes were identified and cloned. In this paper, the second generation sequencing technology Illumina was used to carry out high throughput sequencing for the large plasmid DNA sequence of isolated specific Bt strains, and the novel insecticidal genes were rapidly and efficiently identified combining biological information analysis.

1 Results and Analysis

1.1Bt strain identification

The bacilli which after the high temperature of 70℃ water bath treatment screening and carbolfuchsin staining, were observed under 1000 times microscope, and the bacilli which produced parasporal crystal protein were identified as Bt strains. Then the cry gene and large plasmid DNA of Bt isolates were analyzed by PCR-RFLP and pulsed field agarose gel electrophoresis (results were not listed), and proteins of Bt strain in spore stage were analyzed by SDS-PAGE electrophoresis (Figure 1). Refer to SDS-PAGE characteristic bands of standard Bt strains (Bt subsp. Israelensis: 140 kD, 130 kD, 75 kD and 27 kD; Bt subsp. kurstaki HD1: 130 kD, 70 kD; Bt subsp. sandiego:140 kD) selected out 15 new strains of Bt to carry out subsequent Illumina sequencing.

Figure 1 SDS-PAGE profiles of proteins from Bt isolates during the sporulation growth phase

1.2 Extraction of DNA of large Bt plasmid

DNA of large plasmids of 15 Bt strains were extracted by alkaline lysis method. And 0.8% agarose gel electrophoresis detection showed that these DNA had clear bands (Figure 2). To further clarify the quality of Bt large plasmid extracted DNA, using ultra micro UV spectrophotometer Scan Drop200 (Analytikjena, Germany) measured plasmid DNA samples solubility were more than 200 ng/μL concentration, OD_260/280 value were 1.8~2.0, met the requirements of Illumina sequencing of DNA samples. In order to save the cost, the DNA samples of 15 strains Bt plasmids were mixed into a DNA sample to sequence by 1: 1.

Figure 2 Bt large plasmids DNA detection by agarose gel electrophoresis

1.3 Assembly and analysis of Bt plasmid DNA sequences

Illumina sequencing technology was used to carry out paired-end high-throughput sequencing for Bt large plasmid DNA, thus obtained nearly 500 M data. Deal with the relatively low quality reads. In general, the small fragment data will remove the 10%~20% data. Because the duplication was relatively high, and some pollution of small fragments due to the problem of establishing database, large fragment data removed more data, and some high heterozygosity rate or poor quality sequencing reads would also do other treatment. The obtained sequences were deeply assembled to form 134 scaffolds or contigs by using the software ABySS (Table 1).

Table 1 Statistics of the Bt large plasmids DNA assembly

1.4 Insecticidal gene identification

By using GLIMMER and GeneMark, we predicted ORF of predicted genomes and the possible coding genes. Finally, 1781 genes were selected which the GC content were 35.7%, the ORF size ranged from 114 bp to 5322 bp, and with an average length of 690 bp. And using KEGG automatic annotation server (KAAS) to functional annotate the previously predicted genes (results were not listed). At the same time, analyze target toxin genes by further development and construction of the local blast. Using the 609 amino acid sequences and the nucleotide sequences of the known Bt toxin protein to construct the blast sequence alignment database, and through the blastp and blastn comparison procedure, the genes which could encode the toxin protein were screened out. 25 novel Bt insecticidal protein genes were identified, including 5 vip genes and 3 novel cyt genes (Table 2).

Table 2 Insecticide proteins detection by Illumina sequencing

2 Discussions

Hainan is the only tropical Province in China, which has the largest and best preserved pristine tropical rainforest area. The original tropical rain forest is extremely rich in biological resources and well protected, and contains valuable microbial resources to be excavated. It is very promising to discover new strains of Bt and novel highly efficient insecticidal genes (Zhang et al., 2009a; 2011). In this study, we systematically collected soil samples from the Hainan Five Fingers Group and the Diaoluoshan tropical rain forest area, and screened the strains of Bt. On the basis of above, the second generation sequencing technology Illumina was used to carry out high-throughput sequencing of the isolated Bt strain genomes. More than 20 novel insecticidal protein genes were obtained which the similarity with existing insecticidal proteins ranging from 25% to 49%.

Using Illumina sequencing technology to identify and clone novel Bt insecticidal protein genes can effectively overcome several disadvantages of the molecular hybridization, bioassay, protein N terminal sequencing and gene chip techniques, including time-consuming and labor-intensive, complex technology and expensive cost. Illumina sequencing is bacically not restricted by the homology level of gene sequence, and whether silence or expression, and can obtain the insecticidal protein gene with low homology (Quail et al., 2008; Qin et al., 2011). The most characteristic of Illumina sequencing is high data throughput. And with the combination of bioinformatics analysis, it can quickly identify the insecticidal genes. In this study, a large number of novel insecticidal protein genes were obtained, which would further enrich the insecticidal gene species in China, and provided new genes and new ideas for biological control of agricultural pests and disease vectors.

3 Materials and Methods

3.1 Isolation and identification of Bt strains

From May to July 2012, a total of 506 soil samples were respectively collected from Diaoluoshan and Five Fingers Group, Hainan Province. The Bt strains were isolated by the sodium acetate medium and 70℃ high temperature treatment (Travers et al., 1987; Santana et al., 2008; Zhang et al., 2009b; Wu et al., 2013). In order to determine whether the Bt strain was a new strain, SDS-PAGE electrophoresis, PCR-RFLP method and protein spectrum were adopted to further identify the spore protein and cry gene of the strain (Kuo and Chak, 1996; Song et al., 1998; Zhang, 2011).

3.2 Bt large plasmid DNA extraction

Extraction of Bt large plasmid DNA with reference to Wu et al. (2013) method.

3.3 Illumina sequencing of Bt large plasmid DNA

Bt strains large plasmid DNA were detected, and then constructed library: firstly the Bt large plasmid DNA were broken into small fragments of DNA less than 800 bp by ultrasonic method bioruptor or covaris; then Klenow DNA polymerase or T4 PNK or T4 DNA Polymerase made the small fragment DNA to be blunt-ended, and then in the 3'end of blunt-end added "A", so DNA fragment could efficiently inserted into vectors which 3' end with "T". The agarose gel electrophoresis method selected the target connection products that need to recover and purify. The inserted DNA fragment clusters were amplified by using the universal primers on the vector, and the Illumina HiSeq 2000 platform was used for sequencing.

3.4 Illumina sequencing sequence assembly

After sequencing the DNA library, we respectively intercepted 1~90 bp of reads, and then removed the reads which the sum of base numbers achieved a certain percentage (Set to 4 bp, default was 10%); excluded the reads which base quality value was continuous less than or equal to 20 on a certain degree (set to 20, default was 40%);

removed small fragments contamination, duplication pollution and adapter pollution (set to 15 bp, the default adapter sequence and reads sequence had 15 overlap bp).

Using SOAPdenovo short sequence assembly software (http:// soap.genomics.org.cn/soapdenovo.html; version: 1.05) to assemble the processed reads data. After several adjustments, the main parameter K was set to 83, the optimal assembly results were obtained. The reads were compared to Contigs. According to the relationship of Paired-End and overlap of reads, assembly results were partially assembled and optimized.

3.5 Insecticidal gene identification

By using GLIMMER 3.02 and GeneMark, we predicted ORF of predicted genomes and the possible coding genes (Besemer and Borodovsky, 1999; Delcher et al., 1999). Constructing the blast sequence alignment database, and through the blastp and blastn comparison procedure, the genes which could encode the toxin protein were screened out. In order to improve the accuracy and reliability of the results, the threshold was set to three: low: e-value < 0.0001; middle: e-value < 0.0001, alignment length/max (query length, subject length) > 0.3; high: e-value > 0.0001; alignment length/max (query length, subject length) > 0.7.

Authors’ contributions

Zhang Wenfei was responsible for experimental design, sequencing data analysis and paper writing; Wu Hongping, Xu Zhixia and Jin Yinghong were responsible for sequencing sample preparation; Wu Zhongqi, Huang Nannan, Qian Jiang Chao, Jin Wenjie, Jia Luyu, Wang Ning, Li Xinfeng, Zhao Cong, Liu Xiaojing, Liu Jie, Liu Mingyue and Liu Yifeng were responsible for experiments of soil sampling, strains isolation and cry gene PCR identification; Wang Ruiping was responsible author in this paper.

Acknowledgements

This study was supported by the Hainan Province High School Scientific Research Project (Hjkj2013-19) and the Hainan Normal University Students' Innovation and Entrepreneurship Training Program Funded Project, at the same time thanks for the help of Forestry Bureau of Hainan Province and Diaoluoshan National Forest Park in microbial soil sample collection.