Genomic insights into neonicotinoid sensitivity in the solitary bee Osmia bicornis

Abstract

The impact of pesticides on the health of bee pollinators is determined in part by the capacity of bee detoxification systems to convert these compounds to less toxic forms. For example, recent work has shown that cytochrome P450s of the CYP9Q subfamily are critically important in defining the sensitivity of honey bees and bumblebees to pesticides, including neonicotinoid insecticides. However, it is currently unclear if solitary bees have functional equivalents of these enzymes with potentially serious implications in relation to their capacity to metabolise certain insecticides. To address this question, we sequenced the genome of the red mason bee, Osmia bicornis, the most abundant and economically important solitary bee species in Central Europe. We show that O. bicornis lacks the CYP9Q subfamily of P450s but, despite this, exhibits low acute toxicity to the N-cyanoamidine neonicotinoid thiacloprid. Functional studies revealed that variation in the sensitivity of O. bicornis to N-cyanoamidine and N-nitroguanidine neonicotinoids does not reside in differences in their affinity for the nicotinic acetylcholine receptor or speed of cuticular penetration. Rather, a P450 within the CYP9BU subfamily, with recent shared ancestry to the Apidae CYP9Q subfamily, metabolises thiacloprid in vitro and confers tolerance in vivo. Our data reveal conserved detoxification pathways in model solitary and eusocial bees despite key differences in the evolution of specific pesticide-metabolising enzymes in the two species groups. The discovery that P450 enzymes of solitary bees can act as metabolic defence systems against certain pesticides can be leveraged to avoid negative pesticide impacts on these important pollinators.

Fig 1. Comparison of the CYPome of O. bicornis with other bee species.

(A) Ortholog analysis of O. bicornis with five other bee species. 1:1:1 indicates common orthologs with the same number of copies in different species, N:N:N indicates common orthologs with different copy numbers in different species, UP indicates species specific paralogs, UC indicates all genes which were not assigned to a gene family, CS indicates clade specific genes. Pie charts show the percentage of genes in the CYPome of each bee species in the CYP2, 3, 4 and mitochondrial clade. (B) Rooted maximum likelihood consensus phylogenetic tree of the CYPome of the same species shown in panel A. Genes are coloured according to their adscription to different P450 clades. (C) Maximum likelihood phylogenetic tree of the CYP9 family of P450s in the same species, P450s belonging to the CYP9Q subfamily are highlighted using filled diamonds.

Fig 2. Toxicodynamics and pharmacokinetics of neonicotinoid sensitivity in O. bicornis.

(A) LD₅₀ values for imidacloprid and thiacloprid in insecticide bioassays for O. bicornis, for comparison data is also shown for A. mellifera and B. terrestris. Sensitivity thresholds are depicted according to EPA toxicity ratings [45]. Data for A. mellifera is taken from [13,14], data for B. terrestris is taken from [3]. Error bars display 95% CLs (n = 4). (B) Specific binding of thiacloprid and imidacloprid to O. bicornis nAChRs. Error bars display standard deviation (n = 3). (C) Penetration of radiolabelled thiacloprid and imidacloprid through the cuticle of O. bicornis. The percentage of the initial ¹⁴C-imidacloprid and ¹⁴C-thiacloprid dose recovered by external cuticular rinsing over 24 hours is shown by dashed lines. The percentage of ¹⁴C-imidacloprid and ¹⁴C-thiacloprid recovered from combusted bees (i.e. internalized compound) is shown by solid lines. Error bars display standard deviation (n = 3). (D) Sensitivity of O. bicornis to imidacloprid and thiacloprid before and after pre-treatment with the insecticide synergist PBO (piperonyl butoxide). Error bars display 95% CLs (n = 3).

Fig 3. Identification of neonicotinoid metabolising P450s in O. bicornis.

(A) Metabolism of thiacloprid and imidacloprid by recombinantly expressed CYP9BU1 and CYP9BU2. Production of 5-hydroxy thiacloprid and 5-hydroxy imidacloprid is displayed per pmol of P450 (*P<0.05, ****P<0.0001; paired t test). Error bars display standard deviation (n = 3). (B) Sensitivity of transgenic flies expressing CYP9BU1 and CYP9BU2 to thiacloprid and imidacloprid in insecticide bioassays. Data is expressed as resistance ratio (RR) compared to a control line (flies of the same genetic background but without the transgene). Significant changes in sensitivity between control and transgenic lines are indicated by an asterisk and are based on non-overlapping 95% fiducial limits of LC50 values (n = 5). See also S6 Table.

Fig 4. Expression of O. bicornis P450s after exposure to neonicotinoids and in different tissues.

(A) Expression heat map of O. bicornis P450s after exposure of female bees to imidacloprid (IMI), thiacloprid (THI) or insecticide dilutent (Ctrl). Expression in each of the four replicates per treatment is derived from scaled FPKM values for each P450 transcript. A maximum likelihood tree of Osmia bicornis P450s is shown to the left of the heatmap. (B) Relative expression (fold change) of O. bicornis thiacloprid metabolising CYP9 genes in different tissues of female bees measured by quantitative PCR. Significant differences (p<0.01) in expression between tissues is denoted using an asterisk above bars as determined by one-way ANOVA with post hoc Tukey HSD.