Validation of quantitative real-time PCR reference genes and spatial expression profiles of detoxication-related genes under pesticide induction in honey bee, Apis mellifera

Abstract

Recently, pesticides have been suggested to be one of the factors responsible for the large-scale decline in honey bee populations, including colony collapse disorder. The identification of the genes that respond to pesticide exposure based on their expression is essential for understanding the xenobiotic detoxification metabolism in honey bees. For the accurate determination of target gene expression by quantitative real-time PCR, the expression stability of reference genes should be validated in honey bees exposed to various pesticides. Therefore, in this study, to select the optimal reference genes, we analyzed the amplification efficiencies of five candidate reference genes (RPS5, RPS18, GAPDH, ARF1, and RAD1a) and their expression stability values using four programs (geNorm, NormFinder, BestKeeper, and RefFinder) across samples of five body parts (head, thorax, gut, fat body, and carcass) from honey bees exposed to seven pesticides (acetamiprid, imidacloprid, flupyradifurone, fenitrothion, carbaryl, amitraz, and bifenthrin). Among these five candidate genes, a combination of RAD1a and RPS18 was suggested for target gene normalization. Subsequently, expression levels of six genes (AChE1, CYP9Q1, CYP9Q2, CYP9Q3, CAT, and SOD1) were normalized with a combination of RAD1a and RPS18 in the different body parts from honey bees exposed to pesticides. Among the six genes in the five body parts, the expression of SOD1 in the head, fat body, and carcass was significantly induced by six pesticides. In addition, among seven pesticides, flupyradifurone statistically induced expression levels of five genes in the fat body.

Fig 1. C_q distributions of the five candidate reference genes.

Box plots of C_q values for the five reference genes were compared from five body parts from honey bees exposed to seven pesticides (A-E), and the integration of all samples (F). The horizontal lines in the box indicate the 25^th, 50^th, and 75^th percentile values. The dotted lines in the box show the mean median. The error bars denote the maximum and minimum values.

Fig 2. The expression stability values of the five candidate reference genes analyzed by four programs.

The C_q values integrated from all five body parts of honey bees exposed to seven pesticides were used for the analysis with NormFinder (A), BestKeeper (B), geNorm (C), and RefFinder (D). The dotted line indicates the cutoff values for the appropriate reference gene selection.

Fig 3. Optimal number of reference genes for target gene normalization determined by geNorm pairwise variation analysis.

Pairwise variation values (V_n/V_n+1) were calculated from the five body parts of honey bees treated with seven pesticides (A-E), and integration of all samples (F). The dotted lines indicate the cutoff value for the suggestion of an optimal number of reference genes.

Fig 4. The comparison of expression levels of AChE1 calculated with different normalization methods.

The expression levels of AChE1 normalized with a single gene of the five candidate reference genes and combinations of multiple reference genes were statistically compared in the five body parts of honey bees exposed to seven pesticides and non-pesticide treatment control. The reference genes were selected according to the stability ranks analyzed by RefFinder: combinations of two (RAD1a+RPS18), three (RAD1a+RPS18+ARF1), and four (RAD1a+RPS18+ARF1+RPS5) (Fig 2D). The expression levels of AChE1 calculated with different normalization methods were statistically compared with one-way ANOVA with Tukey’s multiple comparison test, and different letters indicate significantly different values (p < 0.05). The hatched lines indicate the expression levels of AChE1 normalized with the combination of multiple reference genes of which the number was obtained from the lowest Pairwise variation values (V_n/V_n+1) (Fig 3). Data are presented as mean ± SE.

Fig 5. Expression levels of AChE1, CYP9Q1-3, CAT, and SOD1 in honey bee body parts treated with pesticides.

The expression levels of genes were normalized with a combination of RAD1a and RPS18, and the expression changes of each gene due to pesticide exposure were calculated with the non-pesticide treatment control. The difference of expression between the control and pesticide-treated sample were statistically compared with the independent samples T-test with Tukey’s comparison analysis, and significant differences were measured with *(p < 0.05), **(p < 0.01), and ***(p < 0.001). Red presents upregulation, while blue presents downregulation.