Cyp6g2 is the major P450 epoxidase responsible for juvenile hormone biosynthesis in Drosophila melanogaster

In D. melanogaster, JH is known to play vital roles in post embryonic development and reproduction [30, 31]. To examine whether Cyp303a1, Cyp305a1, and Cyp6g2 have any function in the CA, we performed CA-specific knockdown of Cyp305a1, Cyp303a1, and Cyp6g2 with jhamt-Gal4 and examined phenotypes associated with JH deficiency [5]. Remarkably, only the knockdown of Cyp6g2 resulted in partial pupal lethality, accompanied by decreases in adult ovary size and fecundity (Additional file 1: Fig. S1), suggesting that Cyp6g2’s involvement in JH biosynthesis. Therefore, we focused on Cyp6g2 in subsequent studies.

JH plays an important role in stimulating yolk protein synthesis in the fat body to facilitate mature egg production [37]. Consistently, the mRNA levels of yolk protein 1 (yp1), yolk protein 2 (yp2), and yolk protein 3 (yp3) were found to be downregulated in females lacking Cyp6g2 and jhamt function (Fig. 3D). The composite data show that Cyp6g2 is indispensable for female reproduction.

Significant advancements have been made since the original discovery that JHB3 as the predominant JH in D. melanogaster [20]. These include the identification of the JH receptor Met and Gce as well as the JH primary response gene Kr-h1, and the characterization of the molecular mechanism underlying cooperative regulation of JH and 20E on molting and metamorphosis [6–8, 20, 33, 34]. However, due to the universal JH epoxidase gene CYP15 is not present in the genome of D. melanogaster and its congeners, the identity of JHB3 epoxidase remains elusive [23, 38]. Based on the following criteria, we propose that Cyp6g2 is the major epoxidase responsible for JH biosynthesis in D. melanogaster. First, JH is known to be present in the early larval instars, declines substantially during the last (third) larval stage, and then reappears transiently during pupariation in D. melanogaster [35]. Acting like jhamt, Cyp6g2 is expressed selectively in the CA and its temporal expression pattern is closely correlated with JH signaling activity (Fig. 1) [16, 25, 33]. This parallels the expression profile of the first-described JH epoxidase gene DpCYP15A1 (Diploptera punctate CYP15A1) [15]. Furthermore, Cyp6g2, like DpCYP15A1, also localized to the endoplasmic reticulum (Fig. 1D) [15]. Second, Cyp6g2^−/− mutants displayed JH-dependent lethality and female reproductive deficiencies (Figs. 2 and 3), and pupal lethality of Cyp6g2^−/−::jhamt² double mutants is 100%, identical to that of previously reported genetically allatectomized animals (Aug21 > grim). Third, applying exogenous JH and JH analogs could rescue Cyp6g2^−/−::jhamt², allowing them to develop into adults. Furthermore, these treatments partly restored the female reproductive deficiencies observed in Cyp6g2^−/− (Fig. 4). Fourth, the JHB3 and JH III titer decrease significantly in Cyp6g2^−/− larvae (Fig. 5). Fifth, in D. melanogaster Kc cells, overexpression of Cyp6g2 and jhamt converted FA into JHB3, overexpression of Cyp6g2 converted MF into JHB3 and JH III respectively. Overall, our studies presented compelling evidence supporting the role of Cyp6g2 as the major JH epoxidase in JH biosynthesis in D. melanogaster, mainly for FA and minorly for MF. However, further investigations are required to elucidate the specific enzymatic properties of Cyp6g2, including its substrates and products. Unlike CYP15 enzymes, Cyp6g2 may not have such a strong substrate specificity, as its overexpression conferred increased tolerance to various insecticides, including the organophosphate insecticide diazinon, the neonicotinoid insecticide nitenpyram, and dichlorodiphenyltrichloroethane (DTT) [27–29, 39]. Considering that epoxidated JH is dispensable for insect embryonic development, it is interesting and requires further research to understand why the embryonic lethality of the Cyp6g2^−/−::jhamt² double mutants was lower than that of the single mutant (Fig. 2C) [19]. Additionally, the embryonic lethality of the Cyp15c1^−/−; jhamt^−/− double mutant was also lower than that of the single jhamt^−/− mutant in silkworm [40].

Although JHAMT has been proposed as a rate-limiting enzyme and key enzyme in JH biosynthesis [41–43], the order of the last two steps in JH synthesis has been suggested to depend on the substrate specificity and affinity of the JH epoxidase [44]. In Lepidoptera and higher dipteran species, JH epoxidase exhibited higher affinity for FA, leading to epoxidation preceding methylation, whereas in most other insects, JH epoxidase showed higher affinity for MF, resulting in the esterification of FA to form MF occurred before MF epoxidation [38, 44]. The functional roles of JH epoxidase and JHAMT in development and reproduction have been reported in several insect species, including the domestic silkworm B. mori, the red flour beetle T. castaneum, the mosquito A. aegypti, and the fruit fly D. melanogaster (as demonstrated in this study). Each of these insects displayed distinct phenotypes upon knockdown or knockout of JH epoxidase or JHAMT. In B. mori, which usually undergoes five larval instars, both genes are functionally important for proper metamorphosis and indispensible for JH biosynthesis [14, 40]. In T. castaneum, which typically undergoes 6–8 larval molts, RNAi-mediated knockdown of JHAMT in the fourth or fifth instar caused precocious metamorphosis while knockdown of CYP15A1 did not cause precocious pupation [18, 32]. In A. aegypti, which usually undergoes four larval instars, JHAMT also shows more important role than the JH epoxidase [19]. D. melanogaster larvae invariably develop in three instars, and neither genetic allatectomy nor deficiency in JH receptor genes, Met and gce [8, 34, 35], nor simultaneous knockout of jhamt and Cyp6g2, were able to induce precocious metamorphosis in this insect (Fig. 2). Our studies in D. melanogaster show that the developmental and reproductive deficiencies of the Cyp6g2^−/− mutants are more pronounced compared to jhamt² mutants, which is unique among all the insects studied thus far. At 3 h AIW, Kr-h1 expression was normal in jhamt² larvae but was reduced in Cyp6g2^−/− larvae, possibly because of Kr-h1-inducing ability of JHB3 acid was stronger than MF and jhamt² larvae own JHB3 acid. Additionally, our previous results also showed that Kr-h1 expression was normal in the fat body of jhamt² larvae [5]. An alternative explanation is that Cyp6g2 plays essential roles in the larval development and ovary maturation that is not related to its function in JH biosynthesis. In comparison with MF, Cyp6g2 prefers using FA as the substrate, and it is the only enzyme identified to epoxide two sites in the JH backbone, thus producing JHB3 as the major JH in higher dipterans (Fig. 6).

Moreover, we found the occurrence of precocious metamorphosis when jhamt and CYP15A1 were depleted by RNAi in the hemimetabolous species, the American cockroach, P. americana, which usually shows 14 nymphal instars. And the knockdown of jhamt showed a higher proportion of precocious metamorphosis (S. Z. and S. L., unpublished data). Interestingly, in adult P. americana, both jhamt and CYP15A1 RNAi depletion equally inhibited vitellogenesis and ovarian maturation [45]. In contrast, in another hemimetabolous insect model, the desert locust S. gregaria, the knockdown of jhamt resulted in a delay in sexual maturation while the knockdown of CYP15A1 had no effect [46]. These studies preliminarily showed that the relative importance of JH epoxidase and JHAMT in adult reproductive fitness depends on the order of epoxidation and methylation. However, their functional differences in development also partially depend on unique characteristics of each insect, such as the number of larval molts.

In most insect orders, JH III is the only JH homolog produced by the CA. Compared with other insects that exclusively produce JH III, the last two steps of the JH biosynthetic pathway in D. melanogaster are much less clear [5]. The only established fact is that FA is the common precursor for JHB3, JH III, and MF in D. melanogaster [21]. As the exact biochemical pathway for the production of JH III and JHB3 in D. melanogaster has not been fully characterized, there may be two branches within this pathway, one producing the more abundant juvenile hormone JHB3 and the other producing JH III [47]. Our previous studies showed that overexpression and mutation of jhamt led to an increase and decrease in JHB3 biosynthesis, respectively, without affecting the production of JH III and MF. This suggested that JHAMT governs JHB3 biosynthesis in the CA [5]. Consistent with previous results, in this study, we observed that the whole-body JHB3 titer decreased by 60–70%, JH III titer decreased by 50% in jhamt² larvae in comparison with the w¹¹¹⁸ larvae. Another recent study also showed that JH III titers were reduced by 50% and JHB3 titers were reduced by > 90% in larval hemolymph of a newly generated jhamt mutant line [48]. These results demonstrated that the mutation of jhamt does not completely block JH biosynthesis in D. melanogaster. The MF titer was dramatically increased in Cyp6g2^−/−::jhamt² double mutants. This implied that the existence of other minor methyltransferases that could convert FA into MF and JHA into JH in the CA of D. melanogaster. Similarly, the mutation of Cyp6g2 led to sharply reduced JHB3 levels, but some JHs still remained. This is in contrast to mosquito A. aegypti again, where mutation of JH epoxidase CYP15A1 completely eliminated the production of JH III, causing JH III titer in the hemolymph to completely disappear [19]. Furthermore, the residual JHB3 and JH III levels in Cyp6g2^−/−::jhamt² double mutant larvae were not nil and topical application of FA could partially rescue some Cyp6g2^−/−::jhamt² double mutant larvae into eclosed adults, suggesting the presence of other minor JH epoxidases existed in the JH biosynthesis pathway in D. melanogaster. One potential candidate is Cyp6a2, as its orthologous protein Cyp6a1 has been shown to epoxidase MF into JH III [49]. These findings also indicated that JH biosynthesis in mosquitoes is limited to CA, but in D. melanogaster, the last steps of JH biosynthesis extend beyond the CA and need to be further investigated.

The putative promoter sequence (2002-bp length: − 2002 to 0 bp, from the translational start site of Cyp6g2, the sequence was provided in Additional file 1: Supplementary text) of Cyp6g2 was amplified as a Sac II-BamH I fragment, and cloned into the pChsGAL4 plasmid to generate the Cyp6g2-GAL4 construct. The coding sequences of Cyp6g2 with a C-terminal V5 epitope tag or His epitope tag were cloned into the pUAST vector. Then these constructs were used to produce transgenic flies using P-element-mediated germline transformation by the Core Facility of Drosophila Resource and Technology, Center for Excellence in Molecular Cell Science (CEMCS), Chinese Academy of Science (CAS).

w¹¹¹⁸, jhamt², jhamt-GAL4, and JHRR (JH Response Region)-LacZ, UAS-GFP were reported previously [5, 31]. UAS-Cyp6g2-RNAi, UAS-Cyp305a1-RNAi, and UAS-Cyp303a1-RNAi were obtained from the Vienna Drosophila RNAi Center. Other flies used in this paper were generated by recombination. All fly strains in this paper were grown at 25 ℃ on standard cornmeal/molasses/agar medium.

D. melanogaster Kc cells were cultured in Schneider’s Drosophila medium (21,720,024, Thermo Fisher) supplemented with 5% fetal bovine serum (SH30071.03IH30-45, HyClone). The pActin-Gal4, pUAST-Cyp6g2-V5, and pAC-DsRed2-ER plasmids were transiently transfected as previously described [6, 33].

FA (S-0151), MF (S-0153), JH III (S-0155), and JHB3 (S-0157) were purchased from Echelon Bioscience. Methoprene (S185858) was purchased from Shanghai Aladdin Biochemical Technology Co. Ltd. All JHs or JH analogs used for topical application were dissolved in acetone at a concentration of 6 μg/μl. Third instar larvae (L3D1) were topically treated with 0.5 μl of a variety of concentrations (0–6 μg/μl) of JH mimics diluted in acetone, three biological replicates (each replicate involved 50 L3D1 larvae) were performed. Other details were described in our previous published papers [5, 8].

Total RNA was extracted using AG RNAex Pro Reagent (Accurate Biotechnology, Hunan, China) according to the manufacturer’s instructions. Then 2 μg of total RNA was reversely transcribed into cDNA using HiScript III RT SuperMix for qPCR (+ gDNA wiper) (Vazyme Biotechnology, Nanjing, China). Quantitative real-time PCR (qRT-PCR) was performed using Hieff® qPCR SYBR Green Master Mix (Low Rox Plus) (Yeasen Biotechnology, Shanghai, China) and the Applied Biosystem™ QuantStudio™ 6 Flex Real-Time PCR System. The primers used for qRT-PCR are listed in Table S1 (Additional file 1). Three or more biological replicates (each replicate involved 20 or more individuals) were performed as previously described [6, 33].

Groups of 100 larvae were homogenized in 400 μL ice-cold acetonitrile and were processed by the acetonitrile/pentane extraction method described previously [50]. The recovered organic fraction was dried under N₂ and stored at − 20 ℃ until used. Sesquiterpenoid titers were determined using the recently developed LC–MS/MS protocol by the Center of Pharmaceutical Technology, Tsinghua University [51].

Overexpression of Cyp6g2, JHAMT, and DpCYP15A1 in Kc cells was achieved using a GAL4/UAS system by co-transfecting with pActin-Gal4 construct. UAS-EGFP was used as a negative control. Twenty-four hours after transfection of Kc cells in 100-mm dish. The old medium was replaced with 15 ml of fresh medium. FA or MF was (100 μM at final concentration) added to the medium. After incubation another 24 h, cells were collected by centrifuging for 5 min at 1000 × g. Then these cells were used for extracting sesquiterpenoids.

Brain-RG complexes were dissected from larvae of corresponding time points and stained with antibodies according to standard procedures. The primary antibodies used include LacZ mouse monoclonal antibody (40-1a, Developmental Studies Hybridoma Bank, DSHB), DsRed2 mouse monoclonal antibody (sc-101526, Santa Cruz), V5-Tag (D3H8Q) rabbit monoclonal antibody (#13202S, Cell Signaling Technology), JHAMT rabbit polyclonal antibody (reported previously in [9]), and Cyp6g2 rabbit polyclonal antibody. Secondary antibodies used are Alexa Fluor™ 488 goat anti-rabbit IgG (A-11008; Thermo Fisher), Alexa Fluo™ 594 goat anti-mouse IgG (A-11005; Thermo Fisher). DAPI was used for nuclei labeling. The resulting fluorescence signals were examined under a confocal microscope Olympus FV3000 as previously described [9, 31].

Cyp6g2 rabbit polyclonal antibody was generated by ABclonal Biotechnology Co. Ltd (Wuhan, China). The partial coding sequence of Cyp6g2 (125aa-300aa) was cloned into pET28a-SUMO vector. The recombinant protein was purified from inclusion body with NTA-Ni²⁺ beads. Purified protein was quantified and then used to immunize rabbits.

The ovary in the virgin Drosophila on day 7 after eclosion were dissected and photographed with a Nikon SMZ25 stereomicroscope. Then the areas of at least 10 pairs of ovary were calculated. To analyze fly fecundity, 2 tested virgin females were crossed with 3 wild-type males in a new vial of food with a few grains of dry yeast. Flies were transferred every day, and the number of eggs laid was counted every day. For each genotype, six parallel replicates were set. Other details were described in our previous published papers [5, 31].

All of the data were presented as the mean ± standard deviation of three or more independent experiments and were analyzed using Student’s t test and ANOVA. Asterisks indicate a significant difference as calculated using the two-tailed unpaired Student’s t test (*, p < 0.05; **, p < 0.01; ***, p < 0.001). For ANOVA, the bars labeled with different lowercase letters are significantly different (p < 0.05).

Additional file 1. Figure S1. The changes in JH-associated phenotypes upon CA-specific knockdown of Cyp303a1, Cyp305a1 and Cyp6g2. Table S1. Primers used for qRT-PCR in this study. Supplementary text. The promoter sequence of Cyp6g2 used for generating Cyp6g2-Gal4 transgenic flies. Additional file 2. The individual data values for Fig. 1A, C,  Fig. 2C, D, G, Fig. 3D, Fig. 4A, C, C’, Fig. 5A-C and Fig. S1A-C.