Journal of Clinical Toxicology
Open Access

Synergic Interaction between Permethrin and Bt Toxin Discovered Using RNA-seq

Takuma Sakamoto¹, Toshinori Kozaki² and Norichika Ogata^{3* *}Correspondence: Norichika Ogata, Nihon Bio Data Corporation, 1-2-3 Sakado, Takatsu-ku, Kawasaki, Kanagawa, Japan, Email:

Author info »

Introduction

Insecticides are essential in food production and health protection. For instance, 300 pesticides are in use to control about 900,000 different kinds of insects living within a global population of 7,400,000,000. Insecticides interact with each other and resistance of insects to one chemical is sometimes associated with increased susceptibility to a second chemical [1]. This interaction phenomenon is termed Negatively correlated Cross-Resistance (NCR). New information on NCR would be useful in integrated pest management. Up to now, several methods have been reported for screening and development of NCR toxins. However, only 17 toxin pairs have been revealed to cause NCR in insects (DDT and deltamethrin, DDT and phenylthiourea in Drosophila melanogaster, Bt toxins in Plodia interpunctella, Bt toxins Cry1Ac and Cry1F in Helicoverpa zea, Bt toxins in Helicoverpa armigera, Bt toxin and gossypol, Bt toxin and Steinernema riobrave, Bt toxin and Heterorhabditis bacteriophora in Pectinophora gossypiella, Bt and nucleopolyhedrovirus in Plutella xylostella, Pyrethroids and N-alkylamides, AaIT and pyrethroids in Musca domestica, AaIT and pyrethroids in Heliothis virescens, Pyrethroids and diazinon in Haematobia irritans, N-propylcarbamate and N-methylcarbamate, Nephotettix cincticeps, Organo-phosphates and synthetic pyrethroids in Tetranychus urticae) [2]. Given that next generation sequencing technology has recently been developed, a plan to identify resistant genes by omics analyses and seek drugs which target resistant genes via a knowledge base has become feasible [3]. Thus, it would become possible to predict NCR toxins from omics analyses and then apply agents to induce resistance, and then use the next medicine for the offspring of survivors. In this study, permethrin was selected for omics analyses. The permethrin has two optical isomers, cis-and trans-permethrin and both isomers are mixed in agricultural uses permethrin insecticides. Optical isomers have different physiological activity generally. To simplify the experiment, we examined each isomer independently.

Comparative transcriptomics is a powerful tool. Finding the important gene from comparative transcriptomics is not easy. Every biological sample is unique and every comparison between any sequences data from biological samples show many statistically significant differentially expressed genes. Potentially, biological samples from permethrin-resistant strains or nerve cells (target of permethrin) could be used for comparative transcriptome study. We did not use them, since we do not have any method and technique to control the degree of differences between transcriptomes of treatment and control groups.

The moth, Thysanoplusia intermixta is a pest which feeds on Asteraceae (e.g., burdock) and Umbelliferae (e.g., carrot) [4]. The genome of T. intermixta is unknown. Here, we selected a lepidopteran model insect, the silkworm Bombyx mori, for omics analyses. The silkworm genome was revealed in 2008 and is well annotated [5]. To avoid confusion with omics analyses after surgical removal of the guts and gut lumen, the primary cultured fat bodies of the silkworm in MGM-450 insect medium were used [6]. Permethrin concentrations in the media followed a previous study where an appropriate permethrin concentration (0.25 mM) was obtained, resulting in biologically significant genes by referencing transcriptome diversity as the index of the extent of transcriptome changes [7]. To investigate the effects of permethrin on transcriptomes, we sequenced 6 transcriptomes from larval fat-body tissues exposed to permethrin. Freshly isolated tissues were cultured for 80 hours in MGM-450 insect medium, and then cultured for 10 hours in medium supplemented with 0.25 mM cis-permethrin or 0.25 mM trans-permethrin. The resistant gene was identified by differentially expressed gene analyses. Fortunately, the identified resistant gene was the target of the Bacillus thuringiensis (Bt) toxin [8]. The Bt toxin was predicted as a NCR toxin following the permethrin. For validation of the predicted NCR toxin, the probabilities of survival were compared for the natural population of Thysanoplusia intermixta and the offspring of permethrin survivor T. intermixta. NCR to permethrin and Bt toxins, very common insecticides [9] in T. intermixta is reported in this study.

Materials and Methods

Comparative transcriptomics

Establishment of primary culture: The p50 strain of the silkworm, Bombyx mori, was grown on fresh leaves of the mulberry, Morus bombycis. We harvested the leaves at the mulberry plantation in Tokyo University of Agriculture and Technology. This study complies with relevant institutional, national, and international guidelines and legislation. The larvae of the 5^th instar were aseptically dissected 3 days after the 4^th ecdysis and the fat body was isolated. More than 100 chunks of the tissue (approx. 2 mm³) were excised from the fat bodies of 24 larvae. Those tissue particles were incubated in cell culture dishes (ø=35 mm; BD Biosciences, NJ, USA) with MGM-450 supplemented by 10% FBS (Biowest, Nuaillé, France) with no gas change. The tissue was cultured without antibiotics for 80 hours at 25˚C. The infection of the microbes was checked by microscopic inspection. Infection-free tissues were used in the following induction assays.

Chemicals: All the chemicals used in this study were of analytical grade. cis-permethrin (Wako Pure Chemical, Osaka, Japan) and trans-permethrin (Wako Pure Chemical, Osaka, Japan) were dissolved in acetone. This solution was diluted with three times its volume of ethanol just prior to mixing with the medium. The final concentrations of cis-permethrin and transpermethrin were 0.25 mM. No antibiotics were used in the assays to maintain the primary culture.

Induction assay: The original medium was replaced with the cis-permethrin or trans-permethrin-containing medium in the induction assays. The primary culture tissues were incubated with 0.25 mM cis-permethrin or trans-permethrin for 10 hours. The final concentrations of the inducers and the duration of incubation were determined by the previous report [7]. Inductions were terminated by soaking the tissues in 0.75 ml of TRIzol LS (Invitrogen, CA, USA), and the tissues were kept at -80˚C until analyzed.

RNA isolation: Total RNA was extracted using TRIzol LS (Invitrogen) and the RNeasy Lipid Tissue Mini Kit (Qiagen, Hilden, Germany) following the manufacturers’ instructions as previously described [7]. Silkworm fat bodies (30 mg) soaked in 300 μL TRIzol LS were homogenized. We isolated total RNA from the homoginates as previously described [7]. The integrity of rRNA in each sample was checked using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).

Library preparation and sequencing: RNA-seq: Sequencing was performed according to the TruSeq single-end RNA-sequencing protocols from Illumina for Solexa sequencing on a Genome Analyzer IIx with paired-end module (Illumina, San Diego, CA, USA). A total of 1 μg total RNA was used as the starting material for library construction, using the TruSeq RNA Sample Preparation Kit v2. This involved poly-A mRNA isolation, fragmentation, and cDNA synthesis before adapters were ligated to the products and amplified to produce a final cDNA library. Approximately 400 million clusters were generated by the TruSeq SR Cluster Kit v2 on the Illumina cBot Cluster Generation System, and 36-65 base pairs were sequenced using reagents from the TruSeq SBS Kit v5 (all kits from Illumina). Short-read data were deposited in the DNA Data Bank of Japan (DDBJ)’s Short Read Archive under project ID DRA006123.

Data analysis and programs: Sequence read quality was controlled using FastQC program [10]. Short-read sequences were mapped to an annotated silkworm transcript sequence obtained from KAIKObase (http://sgp.dna.affrc.go.jp) using the Bowtie program. A maximum of two mapping errors were allowed for each alignment. Genome-wide transcript profiles were compared between samples. All the statistical analyses were performed using R software version 2.13.0. For identifying differentially expressed genes specifically from RNA-Seq data, we performed the differentally expression analysis based on a model using the negative binomial distribution called DESeq which is implemented as the DESeq package.

Survival analysis

Worms for the survival analysis: Thysanoplusia intermixta larvae were originally collected near the Fuchu campus of Tokyo University of Agriculture and Technology at Fuchu, Tokyo and then successively raised on an artificial diet [11], at 25 °C (L:D 16:8 h). Adult insects were fed a 10% sugar solution absorbed in cotton. A T. intermixta strain was selected from the parental native strain; pentamolting larvae feed: 1.5 cm3 artificial feeds soaked in racemic mixture 1.0 × 10^-4 % permethrin (Permite 5% Cream; Galderma Ltd., Lausanne, Switzerland) aqueous solution. The strain selection performed one generation. Offspring of surviving larvae selected using diet containing permethrin were used in a following survival analysis.

The survival analysis: Larvae were grown on an artificial diet at 25°C (L:D 16:8 h). Eighty T. intermixta larvae were put into independent plastic cages (φ76 × 38 mm, approx. 90 mL), 40 larvae were the native strain and the other 40 were the permethrin-survivor strain. Twenty pentamolting larvae (10 larvae were native strain and 10 larvae were the permethrinsurvivor strain) were exposed to Bt toxins (Toaro suiwa-zai CT; OAT Agrio Co., Ltd., Tokyo, Japan) from Bacillus thuringiensis serotype kurstaki, containing Cry IAa, Cry IAb and Cry I Ac; pentamolting larvae feed: 1.5 cm³ artificial feeds soaked in 1/50000, 1/60000, 1/70000 and1/80000 Bt toxins dissolved in water. A normal concentration of Toaro suiwa-zai CT is 1/1000 and it is 70 ppm.

Data analysis: Statistical analysis was performed using R software version R 3.0.1 [12]. We developed four models using the GLM function. The response variables were dead or alive of insects assumed to follow the binomial distribution. The explanatory variables were the concentrations of Bt toxins as continuous variables and the permethrin selection as categorical variables. First model is full model; the survival rates were explained by the interating two variables, the concentrations of Bt toxins and the permethrin selection. In second model, the survival rates were simply explained by the concentrations of Bt toxins and the permethrin selection. The survival rates were explained by only the concentrations of Bt toxins in third model and the survival rates were explained by only the permethrin selection in forth model. The Akaike Information Criteria (AIC) values for each model were 28.051, 26.058, 30.282 and 29.182, respectively. The second model was selected given that it had the lowest AIC value. Likely food ratio testing was performed between the first and second models.

Results

Comparative transcriptomics

To identify responding genes to cis-permethrin, a comparison was made between the transcriptomes of silkworm fat bodies cultured in in the medium without permethrin and the transcriptomes of silkworm fat bodies cultured in the medium containing 0.25 mM cis-permethrin. It was found that there were 30 statistically significant differentially expressed genes False Discovery Rate (FDR<0.05) in the comparison data (Figure 1 and Table 1) using a statistical method based on a model using the negative binomial distribution, DESeq. In the case of the transcriptomic study using 0.25 mM trans-permethrin, it was found that there were 50 statistically significant differentially expressed genes (FDR<0.05) (Figure 2 and Table 2). From the gene analysis, BGIBMGA013492, BGIBMGA010010, BGIBMGA009516, BGIBMGA009442, BGIBMGA008815, BGIBMGA007738, BGIBMGA006456, BGIBMGA006099, BGIBMGA005875, BGIBMGA004577, BGIBMGA004172 and BGIBMGA002119 were the biologically significant permethrin induced genes since these twelve genes were included in differentially expressed genes in both comparison studies. The differentially expressed gene (BGIBMGA007738) codes ABC transporter C family protein. ABC transporters play a main role in the mode of action of Bt Cry toxic proteins [13-15]. If permethrin has the effect of increasing the expression of the ABC transporter gene in insects in general, then spraying permethrin not only on silkworms but also on agricultural pests may have increased the expression of the ABC transporter genes in agricultural pests and the amount of ABC transporter proteins in agricultural pests. An increase in the amount of ABC transporter protein in agricultural pests may cause them to be more vulnerable to Bt toxins. We hypothesized Bt toxin is a NCR toxin following the permethrin; Permethrin exposed insect increases the expression of the ABC transporter and the transporter is a target of Bt toxin. We then use the agricultural pest, T. intermixta, to test our hypothesis.

Table 1: Differentially expressed Bombyx mori genes exposed to cis-permethrin.

Table 2: Differentially expressed Bombyx mori genes exposed to trans-permethrin.

Figure 1: Differentially expressed gene analysis of cis-permethrin exposed cultured fat body of Bombyx mori. Scatter plot of log2-fold changes versus mean. Genes marked in red were detected as differentially expressed at the 5% False Discovery Rate (FDR) using DESeq. A=log2((cPer/ Ncp*control/Nco)1/2), M=log2(cPer/Ncp)-log2(control/Nco). "cPer" means expression value of genes in each cPer sequence data. "Ncp" means sum of expression value of genes which measured in each cPer sequence data. "Control" means expression value of genes in each control sequence data. "Nco" means sum of expression value of genes which measured in each control sequence data.

Figure 2: Differentially expressed gene analysis of trans-permethrin exposed cultured fat body of Bombyx mori. Scatter plot of log2-fold changes versus mean. Genes marked in red were detected as differentially expressed at the 5% False Discovery Rate (FDR) using DESeq. A = log2((tPer/Ncp*control/ Nco)1/2), M = log2(tPer/Ntp)−log2(control/Nco).

Survival analysis

The probabilities of survival for the native strain of T. intermixta and the offspring of permethrin survivors T. intermixta were compared. A T. intermixta strain was selected from the parental native strain using artificial feeds containing the permethrin. It is usually performed during at least several generations to select permethrin resistant strain. If we do many generations of selection with permethrin, we should be able to get a strain with distinct properties as a permethrin-resistant strain. Here we envisioned a case where permethrin spraying would be followed by Bt toxins, and we wanted to show that the effect of Bt toxins could be enhanced even when only one generation of permethrin selection had occurred. Therefore, screening against permethrin was performed only one generation. Each of the 10 insects of both strains was fed the artificial feeds containing the Bt toxins at 1/50000, 1/60000, 1/70000 and 1/80000 dilutions (total 80 insects). The probabilities of survival of each insect were statistically analyzed using General Linear Models (GLMs). The permethrin selection was statistically significant (p=0.0126, Figure 3), in a likelihood ratio testing between the Figure 1: Differentially expressed gene analysis of cis-permethrin exposed cultured fat body of Bombyx mori. Scatter plot of log2-fold changes versus mean. Genes marked in red were detected as differentially expressed at the 5% False Discovery Rate (FDR) using DESeq. A=log2((cPer/ Ncp*control/Nco)1/2), M=log2(cPer/Ncp)-log2(control/Nco). "cPer" means expression value of genes in each cPer sequence data. "Ncp" means sum of expression value of genes which measured in each cPer sequence data. "Control" means expression value of genes in each control sequence data. "Nco" means sum of expression value of genes which measured in each control sequence data. selection containing model and the selection ignoring model.

Figure 3: Probabilities of survival of the permethrin survivor strain and the native strain of Thysanoplusia intermixta. Scatter plot for rate of formation of dead insects versus Bt toxin concentration. The insects of the permethrin survivor strain are represented by ○ and the insects of the native strain are represented by +. The rates of formation of dead insects were estimated using 10 insects for each category. A total of 80 insects were used. The mortality decreased for the highest Bt concentration concerning the native strain since experimental variation.

Discussion

Negatively correlated cross-resistance to permethrin and Bt in Thysanoplusia intermixta was predicted using comparative transcriptomics of primary cultured fat bodies of a lepidopteran model insect, the silkworm Bombyx mori. Next generation sequencing technologies that reveal many genomes of nonmodel organisms and the annotating of genomes still represents a challenge. Recently, effective genome annotating tools were developed using genomes including the annotations of other organisms (https://github.com/Hikoyu/FATE) [16]. Wellannotated genomes of model organisms are strongly desired. In this context, there are many harmful insects and NCR studies are needed for all harmful insects. However, the costs of insect species-specific screening and developing NCR strategies are prohibitive. Hence, screening and development programs for NCR toxins using model organisms should be initiated. This approach would not fit to know the differentially expressed genes in T. intermixta. Mechanisms underlying transcriptome expression and proteins function are not exactly same between the two insect species. Further research on the differences in permethrin sensitivity between native strain and offspring of permethrin survivors and on the differences in permethrin dependent changes of transcriptomes using samples prepared from native strain of T. intermixta and offspring of permethrin survivors will help us develop a better IPM strategy.

We used two types of permethrin (cis-and trans-permethrin) on the transcriptome analysis. We obtained 30 and 50 genes which changed their expression level responding the cis-and trans-permethrin, respectively. In both comparison studies, twelve genes were detected as differentially expressed genes. These results suggested that permethrin two optical isomers have different physiological activity. These differentially expressed genes included several carboxylesterases, pyrethroids detoxication enzymes (BGIBMGA004229, BGIBMGA004683, BGIBMGA006456, BGIBMGA007670). BGIBMGA004683 and BGIBMGA006456 increased their expression in cispermethrin treatment samples. BGIBMGA007670 increased their expression in trans-permethrin treatment samples. These results indicated that the most effective detoxication enzyme differs in two permethrin optical isomers.

The transcriptomes from primary cultured tissues were sequenced in this study. Tissue cultures were devised as a means for studying the behavior of animal cells in an environment that is free from systemic variations that might arise in vivo both during normal homeostasis and under the stress of an experiment. In transcriptome analyses, p-values from a comparison of one replicate with another one must be below 0.05 in comparative transcriptomics with Type-I error control. Between replicates, no genes are truly differentially expressed, and the distribution of p-values is expected to be uniform in the interval (0,1) [17]. Whole larval transcriptome analyses would be compromised by the gut lumen. Surgical gut removal would result in changes in the volume ratio between tissues in a body. Differences from these variable factors would be obtained causing statistically significant differences between samples. RNA samples were readily obtained from cultured tissues in the present study thus avoiding such variable factors; in this way biologically significant differentially expressed genes were realized. Drug concentration is also an important factor in comparative transcriptomics as has been reported [7,18,19]. Such methods, predicting NCR toxins using comparative transcriptomics, would be useful not only in pest control but also in research fields affected by biological drug resistance, e.g., antibiotics and cancer research.

Conclusion

NCR is a strategy for acting against the development of resistance to insecticides. However, there is only 11 NCR toxin pairs have been revealed in insects. Here, we reported the novel NCR toxin pair, permethrin-Bt toxin. In vitro transcriptome analyses using cultured fat bodies of silkworm Bombyx mori, a lepidopteron model insect, were performed to predict the NCR toxin pair. The predicted toxin pair was verified using in vivo cross-resistance analyses using larvae of Thysanoplusia intermixta, the agricultural pest moth. In in vitro transcriptome analyses, we cared drug concentration to perform the perfect comparison. In genomic scale analyses, statistical significant differences must appears in any comparison. Comparison designs concentrate biologically significant differences in statistical significant differences.

Acknowledgements

We thank for Tamako Hata for providing the permethrin. We thank for Takeshi Yokoyama for providing the worms and the leaves. We thank for Kikuo Iwabuchi for providing the MGM- 450 insect media.

Funding

Nihon Bio Data Corp. supported this study.

Conflict of Interest

The authors declare no conflicts of interest.

References

1. Casida JE. Pesticide interactions: Mechanisms, benefits, and risks. J Agric Food Chem. 2017;65(23):4553-4561.

   [Crossref] [Google Scholar] [PubMed]

2. Pittendrigh BR, Huesing J, Walters Jr KR, Olds BP, Steele LD, Sun L, et al. Negative cross-resistance: History, present status, and emerging opportunities. InInsect Resistance Management. 2014;373-401.

   [Crossref] [Google Scholar]

3. Mortazavi A, Williams BA, McCue K, Schaeffer L, Wold B. Mapping and quantifying mammalian transcriptomes by RNA-Seq. Nat Methods. 2008;5(7):621-628.

   [Crossref] [Google Scholar] [PubMed]

4. Murakami M, Tsuda K, Kusigemati K. Biological studies of the pests feeding on Gynura bicolor (Willd.) DC. (Asteraceae) II. List of feeding species, seasonal abundance and damage-occurrence in Kagoshima prefecture in 1998. Bull Faculty of Agric, Kagoshima University. 2000(50):9-39.

   [Google Scholar]

5. International Silkworm Genome Consortium. The genome of a lepidopteran model insect, the silkworm Bombyx mori. Insect Biochem Mol Biol. 2008;38(12):1036-1045.

   [Crossref] [Google Scholar] [PubMed]

6. Mitsuhashi J, Inoue H. Obtainment of a continuous cell line from the larval fat bodies of the mulberry tiger moth, Spilosoma imparilis (Lepidoptera: Arctiidae). Appl Entomol Zool. 1988;23(4):488-490.

   [Crossref] [Google Scholar]

7. Ogata N, Kozaki T, Yokoyama T, Hata T, Iwabuchi K. Comparison between the amount of environmental change and the amount of transcriptome change. PLoS One. 2015;10(12):e0144822.

   [Crossref] [Google Scholar] [PubMed]

8. Zhang T, Coates BS, Wang Y, Wang Y, Bai S, Wang Z, et al. Down-regulation of aminopeptidase N and ABC transporter subfamily G transcripts in Cry1Ab and Cry1Ac resistant Asian corn borer, Ostrinia furnacalis (Lepidoptera: Crambidae). Int J Biol Sci. 2017;13(7):835-851.

   [Crossref] [Google Scholar] [PubMed]

9. Bruck DJ, Lopez MD, Lewis LC, Prasifka JR, Gunnarson RD. Effects of transgenic Bacillus thuringiensis corn and permethrin on nontarget arthropods. J Agric Urban Entomol. 2006;23(3):111-124.

   [Google Scholar]

10. Andrews S. FastQC: A quality control tool for high throughput sequence data. 2010.
11. Kawasaki K, Ikeuchi M, Hidaka T. Laboratory rearing method for Acanthoplusia agnata (Lepidoptera: Noctuidae) without change of artificial diet. Jpn J Appl Entomol Zool. 1987;31(1):78-80.

    [Crossref] [Google Scholar]

12. Team RC. R: A language and environment for statistical computing. R Foundation Statistical Computing.

    [Google Scholar]

13. Park Y, González-Martínez RM, Navarro-Cerrillo G, Chakroun M, Kim Y, Ziarsolo P, et al. ABCC transporters mediate insect resistance to multiple Bt toxins revealed by bulk segregant analysis. BMC Biol. 2014;12(1):1-5.

    [Crossref] [Google Scholar] [PubMed]

14. Xiao Y, Liu K, Zhang D, Gong L, He F, Soberón M, et al. Resistance to Bacillus thuringiensis mediated by an ABC transporter mutation increases susceptibility to toxins from other bacteria in an invasive insect. PLoS Pathog. 2016;12(2):e1005450.

    [Crossref] [Google Scholar] [PubMed]

15. Tay WT, Mahon RJ, Heckel DG, Walsh TK, Downes S, James WJ, et al. Insect resistance to Bacillus thuringiensis toxin Cry2Ab is conferred by mutations in an ABC transporter subfamily A protein. PLoS Genet. 2015;11(11):e1005534.

    [Crossref] [Google Scholar] [PubMed]

16. Moriya-Ito K, Hayakawa T, Suzuki H, Hagino-Yamagishi K, Nikaido M. Evolution of Vomeronasal Receptor 1 (V1R) genes in the common marmoset (Callithrix jacchus). Gene. 2018;642:343-353.

    [Crossref] [Google Scholar] [PubMed]

17. Anders S, Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010;15:R106.

    [Crossref] [Google Scholar] [PubMed]

18. Seekaki P, Ogata N. Calculating Kolmogorov complexity from the transcriptome data. Int Conf Intell Comput. 2017;529-540.

    [Crossref] [Google Scholar]

19. Fujita Y, Seekaki P, Ogata N, Chiba K. Physiological effects of a novel artificially synthesized antimalarial cyclic peptide: Mahafacyclin B. PLoS One. 2017;12(11):e0188415.

    [Crossref] [Google Scholar] [PubMed]