Genomic and transcriptomic analyses in Drosophila suggest that the ecdysteroid kinase-like (EcKL) gene family encodes the ‘detoxification-by-phosphorylation’ enzymes of insects
Graphical abstract
Introduction
Toxins play central roles in competition and trophic interactions between species—this is especially true for insects, where they often define ecological niches and can limit the food sources species can exploit. A well-known example is the wide diversity of toxins produced by plants, which aim to kill or otherwise dissuade herbivorous insects feeding on their tissues; the capacity of an insect to tolerate these toxins partially defines which plants it can consume (Mithöfer and Boland, 2012) and contributes to plant-insect co-evolution (Edger et al., 2015; Ehrlich and Raven, 1964). Toxin tolerance has many components, one of the most well-studied of which is metabolic detoxification. In animals, detoxification is carried out by collections of enzyme and transporter systems present in a number of organs and tissues throughout the body, which were originally conceptualised as a series of 'phases” that result in the sequential modification (phase I), conjugation (phase II) and excretion (phase III) of toxins (Williams, 1959). After decades of pharmacological and biomedical research, the specific phase I and II reactions and enzymes present in mammals and other vertebrates are well-known, many of which have also been identified in insects (Berenbaum and Johnson, 2015; Chahine and O'Donnell, 2011; Yu, 2008). However, given that the lineages leading to arthropods and vertebrates diverged approximately 600 m.y.a (Reis et al., 2015). and many gene families are not conserved over deep evolutionary time (Danchin et al., 2006; Lespinet et al., 2002), it is likely that those involved in detoxification may differ between these taxa.
One such difference is xenobiotic phosphorylation, which is rare in mammals but common in bacteria and insects (Mitchell, 2015; Ramirez and Tolmasky, 2010; Wilkinson, 1986; Yang, 1976). Phosphorylation has the potential to be a phase II detoxification reaction, as phosphate groups are highly polar and can be conjugated to hydroxyl moieties, and the formation of phosphorylated metabolites of xenobiotic compounds has been observed in at least 18 insect species across seven insect orders (Table S1). Many of these metabolites were discovered by John Smith and colleagues in the mid 1960s to early 1970s (Binning et al., 1967; Darby et al., 1966; Heenan and Smith, 1974; Smith and Turbert, 1964) before the widespread adoption of modern analytical methods like LC-MS, but more recent papers have characterised phosphate conjugates in some detail (Olsen et al., 2014, 2015). An example can be found in the caterpillars of the gypsy moth, Lymantria dispar, which phosphorylate the glycoside moiety of salicinoids found the leaves of a host plant, Populus tremula x tremuloides, a hybrid poplar tree. These phosphate conjugates—formed in the gut and perhaps also the Malpighian tubules—comprise a substantial proportion of excreted salicinoid-like compounds, especially when caterpillars are previously fed poplar leaves, suggesting this detoxification process is induced by poplar secondary metabolites (Boeckler et al., 2016). Likewise, the detoxicative kinase activity identified in Gromphadorhina portentosa, the Madagascar cockroach, is present in the midgut, fat body and Malpighian tubules of the insect and is inducible by in vivo exposure to phenobarbital (Gil et al., 1974; Yang and Wilkinson, 1973), a compound commonly used to induce detoxification gene expression. Phytoecdysteroid detoxification may also involve phosphorylation (Rharrabe et al., 2007), but the phosphorylation of ingested ecdysteroids has been studied in only one species (Modde et al., 1984), even though ecdysteroids can be phosphorylated in vitro with midgut tissue homogenates from a handful of other species (Webb et al., 1996, 1995; Weirich et al., 1986). Overall, many insects appear able to phosphorylate xenobiotic phenols, glycosides and/or steroids directly, and it is possible other xenobiotics may be metabolised in a similarly direct manner, or after hydroxylation, hydrolysis or glycosylation. To date, however, no detoxicative phosphotransferase enzymes have been cloned or otherwise identified at the genetic level. In this paper, we wish to highlight this outstanding question in insect toxicology and raise a hypothesis about the identity of these unknown enzymes.
The identification of detoxification genes and enzymes is an important part of bridging the gap between toxicology, chemical ecology and functional genomics in insects. In the past, the main method of finding detoxification enzymes involved cloning and biochemical characterisation based on a known detoxification reaction. However, detoxification genes tend to have other characteristic properties, including transcriptional induction by xenobiotics (Willoughby et al., 2006; Giraudo et al., 2010) and enriched expression in tissues with known detoxification roles, like the midgut, Malpighian tubules and fat body (Yang et al., 2007). Phenotypes discovered during genetic experiments can also indicate detoxification functions, including susceptibility via gene disruption (Wang et al., 2018) or tolerance via overexpression (Daborn et al., 2007); genome- and transcriptome-wide association studies can also identify candidate genes for detoxification-related phenotypes using naturally occurring variation (Robin et al., 2019). Additionally, genes encoding detoxification enzymes are thought to undergo gene duplication and loss at a faster rate than those encoding enzymes with important housekeeping functions (Kawashima and Satta, 2014; Thomas, 2007). The broad availability of “-omic” data in some insect taxa, particularly the Drosophila genus, raises the possibility that candidate detoxification genes could be identified by integrating evolutionary and transcriptomic data without any prior knowledge of biochemical functions. However, to our knowledge, this has yet to be attempted in a systematic way.
To validate this approach, we examine the cytochrome P450s (henceforth P450s), which are a large multigene family of enzymes that largely function as monooxygenases and catalyse phase I detoxification reactions such as hydroxylation, although a subset also catalyse a wider variety of reactions, including dealkylation, epoxidation and reduction (Bernhardt, 2006). P450s are an established detoxification family in virtually all animals, including insects (Heidel-Fischer and Vogel, 2015; Yu, 2008), and the number of P450 genes per genome in insects varies dramatically, from 38 in the fig wasp Ceratosolen solmsi to 222 in the little fire ant Wasmannia auropunctata (Rane et al., 2019). It has been suggested that diversity in the size of the P450 gene family may be linked to differences in detoxification capacity between taxa (Calla et al., 2017; Rane et al., 2019, 2016), although this has not been rigorously studied.
Multiple attempts have been made to classify P450s according to their biological functions. Thomas (2007) suggests a split between “endogenous-substrate” and “xenobiotic-substrate” enzymes, while Kawashima and Satta (2014) suggest a similar split with “biosynthesis-type” and “detoxification-type.” For the purposes of this paper, we adopt the classification system proposed by Gotoh (2012) of xenobiotic (X-class), secondary (S-class) and endogenous (E-class) functions. E-class enzymes synthesise or degrade compounds that are important for developmental or physiological processes, such as hormones and cuticular hydrocarbons; S-class enzymes synthesise or degrade secondary metabolites, such as defensive compounds or pigments; and X-class enzymes detoxify xenobiotic compounds found in the diet or otherwise derived from the environment. E- and X-class P450s have been studied for many decades. A classic example of E-class enzymes are the Halloween P450s, which synthesise ecdysteroid moulting hormones from dietary sterols (Rewitz et al., 2007; Niwa and Niwa, 2014), but other E-class P450s belong to the biosynthetic pathways of juvenile hormones (Christesen et al., 2017; Helvig et al., 2004a) and cuticular hydrocarbons (Qiu et al., 2012). Many X-class P450s have also been characterised and implicated in the detoxification of both natural and synthetic toxins (Li et al., 2007). S-class P450s are the least understood of the three classes, although some are involved in the biosynthesis of cyanogenic glycosides (Beran et al., 2019) and the degradation of pheromones (Wojtasek and Leal, 1999).
The evolutionary dynamics of P450s have been well-studied in insects and the Drosophila genus specifically (Feyereisen, 2011, 2006; Good et al., 2014). Within Drosophila, 30 ancestral P450 clades are stable—that is, they contain 1:1 orthologs in all studied species—while 30 clades have gene gain and gene loss in the genus, and 17 have only gene loss (Good et al., 2014). The genome of Drosophila melanogaster specifically contains 87 P450 genes, some of which have been studied in great detail (Chung et al., 2009). While evolutionary stability and developmentally essential (E-class) functions are thought to be linked in D. melanogaster (Chung et al., 2009), the link between evolutionary instability and X-class or S-class functions has yet to be rigorously established in this species.
The ecdysteroid kinase-like (EcKL; Interpro entry IPR004119) gene family is taxonomically restricted, being predominantly present in insect and crustacean genomes (Mitchell et al., 2014). EcKL enzymes are predicted to conjugate phosphate to secondary alcohols, using ATP as a phosphodonor (EC 2.7.1.-) and contain the EcKinase domain (Pfam accession PF02958), formerly known as DUF227 (domain of unknown function 227), a member of the Protein Kinase superfamily (CDD accession cl21453; El-Gebali et al., 2018). This superfamily contains protein kinases, as well as kinases with small molecular substrates, such as choline/ethanolamine kinases, aminoglycoside 3′-phosphotransferases and phosphoinositide 3-kinases (Marchler-Bauer et al., 2015). The EcKLs are currently named after a single member, BmEc22K, which encodes an ecdysteroid 22-kinase in the silkworm, Bombyx mori. BmEc22K phosphorylates the C-22 hydroxyl group of ecdysteroids, producing physiologically inactive ecdysteroid 22-phosphate conjugates (Sonobe et al., 2006). These conjugates are stored in the oocyte where they bind to vitellin in the yolk, and are hydrolysed to their active free form by an ecdysteroid-phosphate phosphatase (EPPase) after fertilisation to supply the ecdysteroid titre required for embryonic development (Sonobe and Yamada, 2004; Yamada and Sonobe, 2003; Yamada et al., 2005). This reciprocal conversion process may also occur in other insects and crustaceans, as orthopteran and other lepidopteran species also store ecdysteroid-phosphate conjugates in their eggs (Feldlaufer et al., 1987; Isaac and Rees, 1984; Isaac et al., 1983; Sonobe and Ito, 2009), as may some crustaceans (Subramoniam, 2000; Young et al., 1991). However, no ecdysteroid kinase/EPPase system has been molecularly characterised in any species besides B. mori, although EPPase orthologs exist in many insect genomes (Sonobe and Ito, 2009) and at least one crustacean genome (Asada et al., 2014).
No EcKLs besides BmEc22K have had their substrates identified, and very few other EcKLs have been functionally characterised. Juvenile hormone-inducible protein 26 (JhI-26) is an EcKL found in D. melanogaster that has been implicated in Wolbachia-mediated cytoplasmic incompatibility (Liu et al., 2014). CHKov1 and CHKov2 are also EcKLs in D. melanogaster; a CHKov1 allele containing a transposable element (TE) insertion confers resistance to the vertically-transmitted sigma virus, and a derived allele containing complete and partial duplications of both the CHKov1-TE allele and its neighbouring gene CHKov2 confers an even greater level of resistance (Magwire et al., 2011). The mechanism underlying the resistance conferred by CHKov1-TE is currently unknown, although it is important to note that the TE insertion likely destroys the kinase function of the encoded polypeptides.
We raise the hypothesis that members of the EcKL gene family encode kinases responsible for the detoxicative phosphorylation seen in insects, based on four observations: first, the apparent taxonomic distribution of EcKLs is consistent with the limited taxonomic distribution of detoxicative phosphorylation in animals (Mitchell, 2015); second, ecdysteroid kinase activity has been linked to detoxicative phosphorylation of phytoecdysteroids in some insects (Rharrabe et al., 2007); third, the size of the EcKL family appears to vary considerably between insect taxa (Mitchell et al., 2014), suggesting not all members encode E-class enzymes; and fourth, EcKLs are at least distantly related to the aminoglycoside 3′-phosphotransferases, known detoxicative phosphotransferase enzymes (Marchler-Bauer et al., 2015).
Here we conduct the first evolutionary analysis of the EcKL gene family, focused on the dipteran genus Drosophila, demonstrating there is wide variability in the stability of EcKL orthologs within this taxon. We then show that integrating evolutionary, xenobiotic induction, transcriptional regulation and tissue expression datasets can be used to predict which members of the P450 gene family are involved in xenobiotic detoxification, and apply this method to the EcKLs, suggesting they also contribute to the insect detoxification system. We also show, using a targeted phenome-wide association study (PheWAS) approach in the Drosophila Genetic Reference Panel (DGRP), that EcKL and P450 genomic and transcriptomic variation is associated with toxic stress phenotypes, providing candidate detoxification functions for members of these two gene families. Finally, we perform RNAi knockdown on a subset of EcKLs in D. melanogaster to find developmental lethality phenotypes and identify candidate E-class genes in this gene family.
Section snippets
Gene family annotation
Eleven Drosophila genomes were accessed from NCBI (Coordinators, 2016; Drosophila 12 Genomes Consortium, 2007) —D. simulans (GCA_000259055.1), D. sechellia (GCA_000005215.1), D. erecta (GCA_000005135.1), D. yakuba (GCA_000005975.1), D. ananassae (GCA_000005115.1), D. pseudoobscura (GCA_000149495.1), D. persimilis (GCA_000005195.1), D. willistoni (GCA_000005925.1), D. mojavensis (GCA_000005175.1), D. virilis (GCA_000005245.1) and D. grimshawi (GCA_000005155.1)—while D. melanogaster (Release 6)
Evolution of the EcKL gene family in the Drosophila genus
We explored the evolution of the EcKLs in the Drosophila genus by collating a dataset of gene models in the genomes of 12 species (Table S2). We manually annotated 564 EcKL gene models in the eleven non-D. melanogaster species in the Drosophila genus, to produce a dataset of 618 total EcKL gene models across the 12 species—of these, 605 are likely full gene models (of which 21 were considered likely pseudogenes), while 13 were partial gene models with clear missing sequence due to genome
EcKL evolution in Drosophila
The evolutionary pattern seen in the EcKL gene family in this study matches that seen in well-established detoxification gene families in Drosophila, such as the P450s, glutathione S-transferases (GSTs) and carboxylcholinesterases (CCEs): a mix of stable clades that have 1:1 orthologs in all species, and unstable clades with high rates of gene gain, gene loss or both (Good et al., 2014; Low et al., 2007; Robin et al., 1996, 2009). We have found that a few EcKLs belonging to stable clades in
Summary
In this study, we have shown that it is possible to use the abundant genomic and transcriptomic resources in the Drosophila genus to test functional hypotheses about a poorly understood gene family, the EcKLs. We have also highlighted that there is much more to discover about the functions of the P450 family in insects, particularly with respect to xenobiotic metabolism. We hope that this work can be used as a springboard for further characterisation of both gene families, and that it might
Author contributions
J.L.S. and P.B.: Writing the manuscript. J.L.S., R.S.G-S., P.B. and C.R.: Editing the manuscript. J.L.S. and R.S.G-S.: Laboratory work. J.L.S. and P.B.: Analysis of data. C.R.: Funding and overall project design.
Acknowledgements
We thank Rob Good and Jin Kee for conversations and analyses that helped motivate this research, as well as Dr Lars Jermiin for his supervisory wisdom and financial support, and Melanie Stewart for feedback on the manuscript. We would also like to thank Pontus Leblanc for assisting with the RNAi knockdown crosses. We also thank Prof. Rene Feyereisen for initial communication about detoxification-by-phosphorylation. Finally we thank two anonymous reviewers whose comments substantially improved
References (175)
- et al.
Comparative analysis of the UDP-glycosyltransferase multigene family in insects
Insect Biochem. Mol. Biol.
(2012) - et al.
Basic local alignment search tool
J. Mol. Biol.
(1990) - et al.
Xenobiotic detoxification pathways in honey bees
Current Opinion in Insect Science
(2015) Cytochromes P450 as versatile biocatalysts
J. Biotechnol.
(2006)- et al.
Metabolism of poplar salicinoids by the generalist herbivore Lymantria dispar (Lepidoptera)
Insect Biochem. Mol. Biol.
(2016) - et al.
Drosophila intestinal response to bacterial infection: activation of host Defense and stem cell proliferation
Cell Host Microbe
(2009) - et al.
The absence of glucuronide conjugates from 1-naphthol dosed flies and grass grubs; Detection of 1-naphthylphosphate
Life Sci.
(1966) Phytoecdysteroids: biological aspects
Phytochemistry
(2001)Origin and evolution of the CYP4G subfamily in insects, cytochrome P450 enzymes involved in cuticular hydrocarbon synthesis
Mol. Phylogenet. Evol.
(2020)Arthropod CYPomes illustrate the tempo and mode in P450 evolution
Biochim. Biophys. Acta
(2011)