A cytochrome P450 insecticide detoxification mechanism is not conserved across the Megachilidae family of bees

Abstract Recent work has demonstrated that many bee species have specific cytochrome P450 enzymes (P450s) that can efficiently detoxify certain insecticides. The presence of these P450s, belonging or closely related to the CYP9Q subfamily (CYP9Q‐related), is generally well conserved across the diversity of bees. However, the alfalfa leafcutter bee, Megachile rotundata, lacks CYP9Q‐related P450s and is 170–2500 times more sensitive to certain insecticides than bee pollinators with these P450s. The extent to which these findings apply to other Megachilidae bee species remains uncertain. To address this knowledge gap, we sequenced the transcriptomes of four Megachile species and leveraged the data obtained, in combination with publicly available genomic data, to investigate the evolution and function of P450s in the Megachilidae. Our analyses reveal that several Megachilidae species, belonging to the Lithurgini, Megachilini and Anthidini tribes, including all species of the Megachile genus investigated, lack CYP9Q‐related genes. In place of these genes Megachile species have evolved phylogenetically distinct CYP9 genes, the CYP9DM lineage. Functional expression of these P450s from M. rotundata reveal they lack the capacity to metabolize the neonicotinoid insecticides thiacloprid and imidacloprid. In contrast, species from the Osmiini and Dioxyini tribes of Megachilidae have CYP9Q‐related P450s belonging to the CYP9BU subfamily that are able to detoxify thiacloprid. These findings provide new insight into the evolution of P450s that act as key determinants of insecticide sensitivity in bees and have important applied implications for pesticide risk assessment.

Based on an analysis of 118 structures of resolution of at least 2.0 Angstroms and R-factor no greater than 20%, a good quality model would be expected to have over 90% in the most favoured regions.Based on an analysis of 118 structures of resolution of at least 2.0 Angstroms and R-factor no greater than 20%, a good quality model would be expected to have over 90% in the most favoured regions.Based on an analysis of 118 structures of resolution of at least 2.0 Angstroms and R-factor no greater than 20%, a good quality model would be expected to have over 90% in the most favoured regions.
Comparison of CYPomes from bee species

Figure S5 :
Figure S5: Multiple sequence alignment of A. mellifera CYP9Q3, C. florisomne CYP9BU3 and M. rotundata CYP9DM1 protein sequences.Aligned in Geneious version 10.2.6 (Biomatters) using MUSCLE(Edgar, 2004).The sequences are coloured black to white according to their similarity.Conserved motifs (M) and substrate recognition sites (SRS) are shown in the consensus annotated in yellow and red respectively.Secondary structure of the protein is shown in the consensus and above each protein, with alpha helices, beta sheets and coiled regions shown as dark cyan cylinders, gold cylinders/arrows and grey lines.Secondary structure marked with reference to the AlphaFold2 models predicted in this study.Residues that make up the catalytic pocket were predicted using CAVER Web v1.1(Stourac et  al., 2019)  and are shown in green under each protein sequence.Catalytic pocket residues common to all sequences are annotated blue in the consensus.

Table S1 : Assembly statistics for the transcriptomes of M. centuncularis, M. leachella, M. lapponica and M. willughbiella.
The BUSCO results presented are based on analysis using the Insecta odb10 test gene set.

Table S5 : Statistical analysis of parent compound depletion and hydroxy- metabolite production [TCP-OH and IMI-OH] after incubation of cytochrome P450s with two neonicotinoid insecticides [thiacloprid (TCP) and imidacloprid (IMI)].
Welch's t-tests were used to compare insecticide depletion in the presence and absence of NADPH and the level of TCP-OH versus IMI-OH production in the presence of NADPH.

Table S6 : Statistical analysis of hydroxy-metabolite production [TCP-OH and IMI-OH] after incubation of cytochrome P450s with two neonicotinoid insecticides [thiacloprid (TCP) and imidacloprid (IMI)].
One-way ANOVA and post-hoc pairwise comparisons (Dunnett's multiple comparisons test) of metabolite production for each P450 compared to that of M. rotundata CYP9DM2.

Table S8 : Topology and binding energies for the most relevant access tunnel for TCP and IMI for A. mellifera CYP9Q3, C. florisomne CYP9BU3 and M. rotundata CYP9DM1.
Selection based on energy barrier Emax values, activation energy (Ea) and length (Å).