Fine-mapping nicotine resistance loci in Drosophila using a multiparent advanced generation inter-cross population

Abstract

Animals in nature are frequently challenged by toxic compounds, from those that occur naturally in plants as a defense against herbivory, to pesticides used to protect crops. On exposure to such xenobiotic substances, animals mount a transcriptional response, generating detoxification enzymes and transporters that metabolize and remove the toxin. Genetic variation in this response can lead to variation in the susceptibility of different genotypes to the toxic effects of a given xenobiotic. Here we use Drosophila melanogaster to dissect the genetic basis of larval resistance to nicotine, a common plant defense chemical and widely used addictive drug in humans. We identified quantitative trait loci (QTL) for the trait using the DSPR (Drosophila Synthetic Population Resource), a panel of multiparental advanced intercross lines. Mapped QTL collectively explain 68.4% of the broad-sense heritability for nicotine resistance. The two largest-effect loci-contributing 50.3 and 8.5% to the genetic variation-map to short regions encompassing members of classic detoxification gene families. The largest QTL resides over a cluster of ten UDP-glucuronosyltransferase (UGT) genes, while the next largest QTL harbors a pair of cytochrome P450 genes. Using RNAseq we measured gene expression in a pair of DSPR founders predicted to harbor different alleles at both QTL and showed that Ugt86Dd, Cyp28d1, and Cyp28d2 had significantly higher expression in the founder carrying the allele conferring greater resistance. These genes are very strong candidates to harbor causative, regulatory polymorphisms that explain a large fraction of the genetic variation in larval nicotine resistance in the DSPR.

Keywords: DSPR; Drosophila melanogaster; MPP; Multiparent Advanced Generation Inter-Cross (MAGIC); Multiparental populations; QTL mapping; RIL; RNAseq; complex traits; nicotine; quantitative genetics.

Figure 1

Means (filled circles) and 1 SDs (vertical lines) for nicotine resistance across 810 pA RILs (463 pA1 RILs and 347 pA2 RILs). Mean nicotine resistance is the fraction of first instar larvae that emerge as adults on nicotine-supplemented media, averaged over replicate tests.

Figure 2

Genome scan for nicotine resistance QTL. Solid curves are the scans for the pA (blue) and pB (red) data, and the horizontal dotted line represents the genome-wide 5% permutation threshold (LOD = 6.5 for both populations). Genetic distances along the chromosomes are indicated along the x-axis. The centromeres are at positions 54 and 47 on chromosomes 2 and 3, respectively. The positions of the four QTL we describe in the text are indicated on the plot for ease of reference. (Note that while we do identify an above-threshold peak in pB on chromosome 2 at genetic position 65 (LOD = 8.9), we are reluctant to consider this a confirmed QTL given the limited number of lines and replicate assays employed in the B population. See Discussion.)

Figure 3

Founder haplotype means and 1 SDs at all mapped QTL. The number of RILs for which we confidently assign a founder genotype (probability > 0.95) is listed at the bottom of each bar, and only founder means associated with at least five observations are presented. We note that at the genomic region harboring Q4, several founder alleles have been almost completely purged from our panels of RILs.

Figure 4

Expression of three candidate genes in first instar larvae from lines A3 and A4 on both control (C) and nicotine-supplemented (N) food. Bars show the expression level—represented by the FPKM score—for the gene, and the vertical lines are the 95% confidence interval on these values (all taken directly from Cuffdiff output files). Statistical tests comparing expression measures across lines and treatments are presented in Table 2.