Proof-of-concept engineering of Escherichia coli expressing a bee-derived cytochrome P450 monooxygenase for thiamethoxam detoxification

Abstract

Bees are unintentionally exposed to pesticides applied to control other insect pests. Although bees possess endogenous detoxification mechanisms, their efficiency may not always be sufficient. To this end, the use of probiotics that can assist in pesticide detoxification may offer an additional layer of protection. In this study, we engineered Escherichia coli BL21(DE3) to heterologously express the bee-derived cytochrome P450 monooxygenase (CYP450), CYP9Q1, from Apis mellifera as a proof-of-concept microbial platform for thiamethoxam biotransformation and future microbiome-assisted pesticide detoxification strategies. We first assessed the toxicity of thiamethoxam, a systemic insecticide, on E. coli, which showed obvious impaired growth at relatively high concentrations (5 and 10 g/L). Subsequently, we computationally modeled the 3D structure and modified the CYP450 to improve its solubility when expressed in E. coli. Molecular docking of the CYP450 and thiamethoxam was also performed to confirm their potential interaction. pRSFDuet-1 plasmid was used to carry the modified CYP450 gene for expression in E. coli and the induction condition was optimized, with 0.5 mM isopropyl 𝛽-D-1-thiogalactopyranoside (IPTG) yielding the most favorable expression level. A whole-cell detoxification assay subsequently showed reduced concentrations of thiamethoxam after 46 h of biocatalytic activity, as monitored by High-Performance Liquid Chromatography (HPLC). Liquid Chromatography-Mass Spectrometry (LC-MS) analysis further revealed the presence of desnitro thiamethoxam corresponding to m/z 247.04105 confirming enzymatic transformation of thiamethoxam. To our knowledge, this study provides the first demonstration of bee CYP9Q1 heterologous expression in E. coli for whole-cell thiamethoxam biotransformation, highlighting its potential as a platform for developing probiotic or gut microbiome-based detoxification strategies to support bee pesticide tolerance.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s13036-026-00686-1.

Highlights

A cytochrome P450 monooxygenase (CYP450) from Apis mellifera (CYP9Q1) was computationally modeled and modified for its expression in Escherichia coli BL21(DE3)

The modified CYP450 was successfully expressed in E. coli, and the optimal induction condition was identified

The expression of CYP450 in E. coli resulted in the reduction of thiamethoxam concentrations in the whole-cell detoxification assay

LC-MS analysis showed that desnitro thiamethoxam was observed after the detoxification assay

Supplementary Information

The online version contains supplementary material available at 10.1186/s13036-026-00686-1.

Introduction

The use of pesticides in agriculture directly affects bees and contributes to the ongoing decline in bee populations. Multiple classes of pesticides are widely applied in agricultural fields, and many of them pose significant risks to bee health. Among these, neonicotinoids are widely applied to control piercing–sucking insects such as aphids, whiteflies, leafhoppers and other insects including flies and moths [1], but their use leads to contamination of pollens and nectars in treated plants [53]. Moreover, the pesticides can disrupt the bee gut microbiome, which plays essential roles in bee health [32]. Bees have several mechanisms to defend themselves from pesticides including the use of detoxifying enzymes. Cytochrome P450 monooxygenases, hereafter CYP450s, are a group of versatile enzymes that are responsible for the oxidation of a wide range of organic compounds. Several CYP450s have also been implicated in pesticide detoxification, as evidenced by elevated gene expression levels following pesticide exposure [27]. Therefore, enhancing the native CYP450-mediated detoxification capacity of bees through microbiome-assisted approaches may represent a promising strategy to improve pesticide tolerance, thereby enhancing bee survival. The use of probiotics in bees has attracted increasing attention as a promising approach to support bee health by stabilizing the gut microbiome [37], improving resistance to pathogens [6], and lower the concentrations of toxic or potentially toxic elements [2]. Overall, this study aims to evaluate the potential of engineered bacteria expressing a bee-derived CYP450 enzyme to detoxify pesticide compounds, with the goal of developing a microbiome-assisted probiotic strategy to enhance pesticide tolerance in bees.

Thiamethoxam is a widely used neonicotinoid pesticide in agriculture [52], particularly on crops that are the sources of pollens for bees [22], for example, oilseed rape [19], and strawberry [26]. It is a second-generation neonicotinoid pesticide that interferes with postsynaptic nicotinic acetylcholine receptors [34] resulting in neurological symptoms such as hyperexcitation and paralysis and, certainly at high doses, death of insects [30]. Moreover, though thiamethoxam has recently been shown to help promote plant growth [36], this pesticide has been reported to negatively influence the microbial community in soil [48], the health of mice [12] and may pose developmental toxicity risks in highly exposed human [46]. It has been reported that thiamethoxam is rapidly metabolized into clothianidin in bees even at a relatively low dose (0.25 ng/bee/day) [10]. However, residue levels of thiamethoxam have been detected at 44.9 ng/g in pollen [40] and approximately 5.1–214 ng/g in vegetable crops [21] with reported half-lives of 4.9–5.5 days in the field and 10.3–15.8 days during storage [39]. Moreover, a thiamethoxam-resistant strain of cotton aphid showed cross-resistance to several insecticides, including clothianidin, which was attributed to the overexpression of detoxification-related enzymes such as CYP450s [47]. This suggests that the resistance may be a result from a versatile mechanism rather than pesticide-specific detoxification mechanisms. To provide further in-depth detail, CYP450s can catalyze several types of reactions (Fig. 1), including carbon hydroxylation (addition of an –OH group to a carbon atom) [14], epoxidation (formation of an epoxide ring on a double bond) [9], sulfoxidation (oxidation of a sulfur atom to a sulfoxide) [44], demethylation (removal of a methyl group) [38] and denitration (stepwise reduction or removal of the nitro group) [4].

Fig. 1.

Proposed detoxification reactions of thiamethoxam via CYP450s (1) Hydroxylation (2) Epoxidation (3) Sulfoxidation (4) Demethylation and (5) Denitration [4, 13]

Here, the expression of different CYP450s in response to thiamethoxam exposure are summarized to facilitate the selection of candidate enzymes (Supplementary Table S1). While several CYP450s showed elevated expression following thiamethoxam exposure, direct biochemical evidence confirming their ability to metabolize thiamethoxam has been unavailable, with one recent study representing the first such demonstration. Specifically, the study showed that CYP6EM1, a CYP450 from Bemisia tabaci (Hemiptera: Aleyrodidae), could transform thiamethoxam, as evidenced by reduced thiamethoxam levels in an in vitro assay [20]. In this work, we aimed to examine the catalytic activity of CYP9Q1, a CYP450 from a honeybee, A. mellifera, toward thiamethoxam using heterologous expression in E. coli as a proof of concept to develop probiotic or microbiome-based detoxification approaches for bees.

Several studies have reported heterologous expression of CYP450s in E. coli; however, the heterologous expression may be challenging due to several factors including its membrane association and solubility. Previously, various strategies have been proposed to achieve the functional expression of CYP450s in heterologous hosts. Given their nature as membrane-anchored proteins [23], the strategies include N-terminal modification and co-expression of auxiliary proteins [31]. To provide more detail, truncation of the N-terminal membrane-anchoring domain upstream of the hydrophilic region has been shown to enhance the activity of plant-derived CYP450s compared with their native forms. Among various modifications investigated, N-terminal truncation that removed the membrane-associated amino acids in front of the hydrophobic region resulted in the highest levels of CYP450 expression and activity [31].

Recently, the use of engineered symbionts re-colonized in bees has attracted attention as an alternative strategy to boost bee immune systems and prevent pathogen infection [18, 49]. Subsequently, the concept of using engineered gut symbionts for other purposes, including metabolic modification to promote bee digestion and pesticide detoxification, was proposed [32]. Although engineering bee gut symbionts has been challenging, a well-established toolkit for genetic manipulation has been reported [24]. Yet, only a few successful cases have been demonstrated. In the context of this study, limitations arise primarily from the difficulty of genetically engineering gut symbionts and the additional challenge of expressing CYP450 enzymes in prokaryotic systems. Therefore, this work adopts a proof-of-concept approach using the well-characterized host, E. coli, to validate the functional activity of a bee-derived CYP450 enzyme prior to implementation in native gut symbionts. Taken together, CYP450s are recognized as enzymes capable of metabolizing thiamethoxam, although slight modifications are often required to achieve successful heterologous expression. While the long-term goal is to develop probiotics or engineer the bee gut microbiome for enhanced pesticide detoxification, E. coli was used in this study as a proof-of-concept host to characterize the enzyme. Accordingly, the toxicity of thiamethoxam was first assessed in the E. coli BL21(DE3) strain. The structure of a bee CYP450 was computationally modeled and the membrane-anchored amino acids were truncated, after which the gene was expressed in E. coli BL21(DE3) and the expression level was optimized. Subsequently, whole-cell detoxification activity of the modified enzyme expressed in E. coli BL21(DE3) was evaluated to verify its functional activity.

Materials and methods

Toxicity of thiamethoxam on E. coli BL21(DE3)

Thiamethoxam was purchased from TCI (Japan). A stock solution (20 g/L) was prepared in distilled water and sterilized by filtration through a 0.22 μm Nylon Syringe Filter (LabFil, China). The toxicity of thiamethoxam toward E. coli was assessed using E. coli BL21(DE3) harboring the plasmid of interest, which was precultured in LB medium before inoculation into fresh LB medium at an initial OD660 of 0.1. Thiamethoxam was added at the start of cultivation at final concentrations ranging from 0 to 10 g/L using a two-fold serial dilution. Cultures were incubated at 37 °C for 48 h in a shaking incubator at 150 rpm. Optical density was measured every 24 h to observe cell growth.

Protein structural modeling and molecular docking analysis of the bee CYP450 and thiamethoxam

The amino acid sequence of the bee CYP9Q1 (NCBI accession no. is XM_006562301.3), hereafter bee CYP450, was retrieved from NCBI database (Supplementary Table S2). The three-dimensional (3D) structure was predicted using AlphaFold2 implemented on the Google Colab platform, ColabFold (Supplementary Table S2) [29]. The default parameters were used apart from the num_relax function that was set to 1 to retrieve the one relaxed structure. The modeled structure was then subjected to the structure validation using Ramachandran plot via Procheck online (Supplementary Table S2). The validated structure was then used for molecular docking analysis using GOLD Protein-ligand docking software [43]. The chemical structure of thiamethoxam was downloaded from PubChem databank (Supplementary Table S2). Asp307 (position before truncation) was used as a docking site corresponding to Asp309 of the reported CYP9Q3 [17, 45]. The docking poses were then visualized using Discovery Studio Visualizer [5].

Heterologous expression and characterization of the bee CYP450

Enzyme modification

As mentioned, a bee CYP450 (CYP9Q1) was the target CYP450 in this work. Firstly, CYP9Q1 was subjected to transmembrane domain prediction using the transmembrane prediction tool namely DeepTMHMM (Supplementary Table S2) [16]. The hydrophobic region in the amino acid sequence was removed according to [31].

Heterologous expression of the bee CYP450 in E. coli BL21(DE3)

Since E. coli BL21(DE3) was used as an expression host to primarily characterize CYP450 expression, the gene sequence was codon-optimized for expression in E. coli and synthesized by Bionics (Korea). pRSFDuet-1 with Kan^R was used as a vector carrying the modified CYP450 gene, hereafter pRSFDuet-1-CYP450. To confirm the expression of CYP450, Western blot analysis was performed. The overnight E. coli BL21(DE3) culture harboring pRSFDuet-1-CYP450 was reinoculated to the LB medium before incubating for 2 h at 37 °C, 180 rpm. Different isopropyl 𝛽-D-1-thiogalactopyranoside (IPTG) concentrations (0, 0.1, 0.5, and 1 mM) were used to induce the protein expression. The cultures were incubated for an additional 4 h before centrifugation to harvest the cells, followed by sonication (3 s on, 1 s off, 10 min, 25% amplitude) for protein extraction. A total of 50 µg of protein for each sample was loaded to the well and proteins were separated on a 12% SDS-PAGE gel and transferred onto nitrocellulose membranes. Western blot analysis was performed using a mouse anti-His antibody (Thermo Fisher Scientific, United States) as the CYP450 gene was designed to include a C-terminal His-tag, followed by detection with an HRP-conjugated goat anti-mouse secondary antibody (Jackson ImmunoResearch, United States). ImageQuant™ LAS 500 Chemiluminescent Imaging System (GE Healthcare, United States) was used to visualize the protein. ImageJ was used to analyze the band intensity [33].

Whole-cell detoxification of thiamethoxam

To assess the ability of the CYP450 to metabolize thiamethoxam, E. coli BL21(DE3) harboring pRSFDuet-1-CYP450 was cultivated in M9 minimal medium and induced with 0.5 mM IPTG immediately after inoculation. After the incubation at 30 °C for 2 h to allow protein expression, different concentrations of thiamethoxam (0, 0.07812, 0.15625, 0.3125 and 0.625 g/L) were fed to the cultures and the incubation was continued for a total incubation time of 48 h.

High Performance Liquid Chromatography (HPLC) was used to quantify the amount of thiamethoxam in supernatant. Samples were prepared by which the cultures were centrifuged at 9,500 xg for 5 min before the supernatants were obtained and filtered through a 0.45 μm nylon syringe filter. Ten µL of each sample was used for analysis. An Agilent 1260 Infinity II HPLC was used for analysis with a C-18 column (Agilent ZORBAX SB-C18; 4.6 × 150 mm 5-micron) as a stationary phase. The mobile phase consisted of 70% deionized water and 30% acetonitrile with a flow rate of 1.0 mL/min. The diode array detector (DAD) was set at a wavelength of 252 nm. For Liquid Chromatography-Mass Spectrometry (LC-MS), chromatographic separation was performed using an Agilent InfinityLab Poroshell 120 EC-C-18 column (2.1 × 100 mm, 2.7-micron) maintained at 50 °C. A 10 µL aliquot of each sample was injected onto the column. The mobile phase consisted of (A) water containing 0.1% (v/v) formic acid and (B) acetonitrile containing 0.1% (v/v) formic acid. The flow rate was set to 0.4 mL/min. The gradient elution program was as follows: 100% A at 0.0–0.5 min; a linear change to 45% A and 55% B at 10.5 min; further decreased to 25% A and 75% B at 12.5 min; then to 0% A and 100% B at 14.0 min, which was maintained until 17.0 min. The mobile phase was returned to initial conditions (100% A) at 17.5 min and held until 20.0 min for column re-equilibration. The supernatant samples were prepared following the same procedure as described for HPLC analysis. For intracellular metabolites, the cells were extracted using 200 µL of 70% (v/v) methanol containing 50 ng/mL sulfadimethoxine as an internal standard. Samples were supplemented with three metal beads and homogenized using a tissue lyser (SWE-FP, ServiceBio, China) at 70 Hz for 3 min, repeated twice to ensure efficient cell disruption. The lysates were subsequently centrifuged at 18,580 xg for 10 min to remove cellular debris. The resulting supernatant was transferred to LC-MS vials for metabolite analysis.

Statistical analysis

Mean ± SD from three biological replicates was used to represent the results in this work. To determine differences within each experiment, a two-sided Student’s t-test was used. All statistical tests were performed on R Studio. Significant differences are indicated by asterisks (*p-value ≤ 0.05, **p-value ≤ 0.01, ***p-value ≤ 0.005, ****p-value ≤ 0.001).

Results

Toxicity of thiamethoxam on E. coli BL21(DE3)

The toxicity of thiamethoxam toward E. coli BL21(DE3) was evaluated across a concentration range of 0–10 g/L. After 24 h of incubation, the growth of E. coli BL21(DE3) reduced significantly in all thiamethoxam concentrations (0.07812–10 g/L). However, the reduction was drastic at 5 and 10 g/L, where a marked decrease in growth, greater than 50%, was observed at 10 g/L (Fig. 2A). Similarly, at 48 h, thiamethoxam at 5 and 10 g/L clearly inhibits the growth of E. coli BL21(DE3) (Fig. 2B).

Fig. 2.

Toxicity of thiamethoxam on E. coli BL21(DE3). LB was used to cultivate E. coli BL21(DE3) with different concentrations of thiamethoxam for (A) 24 h and (B) 48 h. Data are mean of values from three biological replicates. Asterisks indicate statistical significance compared with the 0 g/L thiamethoxam (*p-value ≤ 0.05, **p-value ≤ 0.01, ***p-value ≤ 0.005, ****p-value ≤ 0.001)

Protein structural modeling and molecular docking analysis of the bee CYP450 and thiamethoxam

The amino acid sequence of a CYP450 from a honeybee, namely CYP9Q1, was subjected to transmembrane protein prediction via DeepTMHMM and it showed that the first 22 amino acids starting from methionine (M) to serine (S) were predicted to be signal, outside and membrane amino acids (Supplementary Figure S1). These amino acids were then truncated leaving only methionine as the first starting amino acid (Fig. 3A). Since there is no reported crystal structure of CYP9Q1, the sequence was then modeled for its 3D structure using Alphafold2 via Colabfold with the resultant pLDDT score of 92.9. The Ramachandran plot showed that 92.4% of residues were in most favored regions. The predicted 3D structure was subsequently used for molecular docking analysis, incorporating heme as the catalytic cofactor and thiamethoxam as the ligand (Fig. 3B). The binding pocket was inferred based on CYP9Q3, in which Asp309 has been reported to interact with imidacloprid and thiacloprid. In our analysis, the sequences of CYP9Q1 and CYP9Q3 were aligned using Asp309 of CYP9Q3 as the reference residue, corresponding to Asp307 in CYP9Q1. Based on this alignment, the predicted interacting residues in CYP9Q1 were Phe87, Val99, Phe101, and Ser280 (Fig. 3C).

Fig. 3.

Modification of the bee CYP450 and molecular docking analysis with thiamethoxam. (A) Prediction and truncation site of a CYP450 from bee (CYP9Q1) and (B) Docking pose of thiamethoxam within the active site of CYP9Q1. The protein is shown in light blue, the ligand in magenta, the heme prosthetic group in dark blue, and interacting amino acid residues in yellow. (C) The 2D interaction between CYP9Q1 and thiamethoxam, using the aligned amino acid Asp309 from CYP9Q3 as a reference

Heterologous expression and characterization of the bee CYP450

The gene encoding a truncated bee CYP450 was synthesized in a pRSFDuet-1 plasmid (Fig. 4A) and expressed in E. coli BL21(DE3). The expression of CYP450s has been a challenge in E. coli, IPTG concentrations were then first optimized. Here, the results showed that the expression of CYP450 significantly affected the cell growth, which could be observed with only 4 h after induction (Fig. 4B). This suggests that expression of the bee CYP450 imposed a metabolic burden on E. coli cells, although only a slight reduction in growth was observed. Protein expression was confirmed by Western blot analysis, revealing a band at the expected molecular weight of 48.8 kDa. Expression levels were monitored among E. coli BL21(DE3) strains with and without the CYP450-containing plasmid under different IPTG induction concentrations (0, 0.1, 0.5, and 1 mM). The total protein content was normalized to 50 µg per lane. Western blot analysis revealed different levels of protein expression, with the highest expression observed at 0.5 mM IPTG induction (Fig. 4C, lane 7, Fig. 4D and Supplementary Figure S2). Notably, no leaky expression was detected in the absence of IPTG.

Fig. 4.

Heterologous expression of the truncated bee CYP450 in E. coli BL21(DE3) and tuning the expression via IPTG concentration optimization. (A) The truncated bee CYP450 gene was inserted into a pRSFDuet-1 plasmid and transformed into E. coli BL21(DE3). (B) The growth was observed after 4 h of induction. (C) Western blot analysis of the truncated bee CYP450, where lanes 1–4 contain lysates from control cells, and lanes 5–8 contain lysates from cells harboring the pRSFDuet-1-CYP450 plasmid. Both strains were induced with IPTG at concentrations of 0, 0.1, 0.5, and 1 mM, respectively, and (D) the Western blot band intensity was observed. Data are mean of values from three biological replicates. Asterisks indicate statistical significance compared with 0 mM IPTG in the same experiment (*p-value ≤ 0.05, **p-value ≤ 0.01, ***p-value ≤ 0.005, ****p-value ≤ 0.001)

The activity of the CYP450 was evaluated using a whole-cell detoxification assay. After 2 h of inoculation and IPTG induction, E. coli harboring pRSFDuet-1-CYP450 was supplemented with various concentrations of thiamethoxam (0, 0.07812, 0.15625, 0.3125 and 0.625 g/L). Assays were performed both with and without induction using 0.5 mM IPTG. Thiamethoxam levels in supernatants were quantified using HPLC after 48 h of the total incubation time. The results showed a significant reduction in thiamethoxam across all tested concentrations when comparing the uninduced strain (Fig. 5A) with the strain induced with 0.5 mM IPTG (Fig. 5B). LC-MS analysis was performed to investigate differences in chemical profiles associated with CYP450 expression. The results confirmed a reduction in thiamethoxam levels between the time of thiamethoxam addition and the end of incubation (46 h after thiamethoxam addition) (Supplementary Figure S3). Moreover, LC-MS analysis of the culture supernatant revealed a more pronounced metabolic shift over time (Fig. 5C; Table 1). At the time of thiamethoxam addition, the m/z 261.07559 peak was nearly absent. However, after the incubation, the m/z 261.07559 ion increased dramatically. The substantial accumulation of the m/z 261.07559 ion in the supernatant at 48 h suggests active enzymatic biotransformation.

Fig. 5.

E. coli whole-cell detoxification of thiamethoxam. (A) HPLC analysis of thiamethoxam from a feeding assay without and (B) with 0.5 mM IPTG induction. (C) LC-MS analysis of selected mass with the highest abundance change (m/z 261.07559) between (1) the time of thiamethoxam addition and (2) the end of incubation (46 h after thiamethoxam addition) Data are mean of values from three biological replicates. Asterisks indicate statistical significance (*p-value ≤ 0.05)

Table 1.

Selected masses from LC-MS analysis of the culture supernatants

m/z	Mode	RT (min)	Response	Putative ID	Treatment	Chemical structure
292.02530	+	4.707	1,932,895.80	Thiamethoxam	At the time of thiamethoxam addition
292.02530	+	4.694	1,713,389.38	Thiamethoxam	At the end of the incubation
261.07559	+	3.525	627.31	low score: 2,3,4’-Trihydroxy-4-methoxybenzophenone	At the time of thiamethoxam addition
261.07559	+	3.535	6,661,137.38	low score: 2,3,4’-Trihydroxy-4-methoxybenzophenone	At the end of the incubation

For intracellular metabolite analysis, thiamethoxam and its putative transformation product were detected in both induced and uninduced samples at the end of the incubation. As shown in Table 2, intracellular thiamethoxam (m/z 292.03305) was detected suggesting its ability to be taken up by the cells. Another peak at m/z 247.04105, tentatively identified as NOA407475 (desnitro thiamethoxam), was also detected intracellularly. Notably, this metabolite exhibited an approximately 2.3-fold higher response in induced cells (285,267.60) compared to uninduced cells (122,216.54), which may be a result from pRSFDuet-1 leaky expression as this experiment allowed the incubation up to a total of 48 h, suggesting enhanced transformation activity following IPTG induction; however, it also indicated a leaky expression from the pRSFDuet-1 plasmid or activity of other intrinsic enzymes in E. coli cells.

Table 2.

Selected masses from LC-MS analysis of the intracellular metabolites at the end of the incubation

m/z	Mode	RT (min)	Response	Putative ID	Treatment	Chemical structure
292.03305	+	4.707	6,072,807.44	Thiamethoxam	Induced with 0.5 mM IPTG
292.03305	+	4.694	4,249,584.74	Thiamethoxam	Uninduced
247.04105	+	3.525	285,267.60	NOA407475 (or Desnitro thiamethoxam)	Induced with 0.5 mM IPTG
247.04076	+	3.535	122,216.54	NOA407475 (or Desnitro thiamethoxam)	Uninduced

Discussion

Thiamethoxam is a widely used pesticide for controlling insect pests. The concentrations of thiamethoxam that adversely affect bees are typically between 1.96 and 2.90 ng/bee/day, and the oral LD₅₀ of thiamethoxam has been reported as 9.02 ng/bee. A single oral dose of 1.34 ng/bee caused hyperexcitation and significantly increased flight duration (+ 78%) and distance (+ 72%) in foraging bees. Conversely, in the same study, chronic exposure resulted in decreased flight duration (-54%), distance (-56%), and average velocity (-7%) [41]. This highlights the urgent need to mitigate the toxicity of this pesticide, and we propose the use of probiotic or gut microbiome–based detoxification strategies. Here, E. coli BL21(DE3) was employed as a proof-of-concept host for enzyme engineering and characterization, providing a tractable platform for evaluating detoxification potential. The toxicity of thiamethoxam was first evaluated in this strain and showed a marked growth defect only at a relatively high concentration of 5 g/L indicating a high tolerance to thiamethoxam and supporting its suitability as a host for whole-cell detoxification. Modifications of CYP450s were previously implemented and systematically compared to facilitate heterologous expression. A study has demonstrated that N-terminal truncation of a CYP450 from Arabidopsis thaliana resulted in higher expression (2- to 170-fold higher) in E. coli [31]. Truncation of N-terminal has been identified to be the most effective versatile strategy ([31]; [35, 51]). In this work, the amino acid sequence of CYP450 was obtained from a honeybee, A. mellifera (CYP9Q1), and this is the first report of heterologous expression of this CYP450 in E. coli. Therefore, the expression was tuned using different induction levels. The results indicated that protein expression was highest at an IPTG induction level of 0.5 mM, whereas expression at 0.1 mM was the lowest. In contrast, expression at 1 mM appeared inconsistent, likely due to increased cellular stress. This effect is commonly attributed to the heterologous expression of membrane-bound proteins, which often exhibit toxicity in E. coli, leading to reduced cell growth and unstable protein expression at higher induction levels [28]. However, in this study, the transmembrane region was truncated, yet growth defects at high induction levels were still observed. Therefore, this growth defect may instead be attributed to protein aggregation or misfolding associated with high-level expression.

The bee CYP450 was classified in the CYP9Q subfamily [17] and was also previously shown to catalyze several substrates via oxidation. In one work, nicotine was detoxified via 2′C-oxidation in honey bee larvae [11] and, in another study, CYP9Q6 from Bombus terrestris has been shown to efficiently metabolize the neonicotinoids, thiacloprid and acetamiprid, and its expression is directly associated with increased insecticide tolerance in vivo [42]. A previous study demonstrated molecular docking of CYP9Q3 with clothianidin, which revealed the active-site pocket and its corresponding amino-acid residues [45]. Notably, Phe121 identified in that study corresponds to Phe101 observed in this work. In this work, referencing the binding pocket from the aforementioned work, molecular docking was performed referencing the similar binding pocket with the previous report, and it showed that the binding of thiamethoxam and a heme molecule fit in the active site pocket, indicating a favorable orientation for catalytic interaction.

Moreover, in the whole-cell detoxification assay, the bee CYP450 demonstrated the ability to metabolize thiamethoxam, as evidenced by the reduced amount of thiamethoxam detected in the culture media. Interestingly, a previous work has demonstrated that E. coli was natively capable of bioremediating thiamethoxam at concentrations up to 70 mg/L [50]. However, in the absence of IPTG induction, thiamethoxam levels remained essentially unchanged before and after incubation in this study, indicating that the observed reduction in thiamethoxam was attributable to CYP450 expression. It should also be noted that the reported depletion of 70 mg/L in the previous study occurred over a prolonged incubation period of more than 14 days. LC-MS in positive and negative ion modes produced different sets of putative extracellular metabolites in supernatants. In the positive mode, the top ion that increased from time zero to 48 h after inoculation was m/z 261.07559. Preliminary library matching suggested a possible identity as 2,3,4′-trihydroxy-4-methoxybenzophenone; however, with a low confidence score. Therefore, further examination of this most intense ion, m/z 261.07559, was performed and indicated that this mass is consistent with a metabolite of thiamethoxam, formed when the nitro group (–NO₂) on the nitroguanidine moiety is replaced by an amino group (–NH₂) with the chemical formula C₈H₁₂ClN₅OS. The theoretical monoisotopic mass of this metabolite is 261.0451 Da [7], which agrees closely with the observed m/z 261.07559 in positive mode. Moreover, when intracellular metabolites were analyzed comparing between induced and uninduced cells, the LC-MS library indicated desnitro thiamethoxam with the m/z of 247.04105 with the chemical formula C₈H₁₁ClN₄OS was detected. Although desnitro thiamethoxam is consistently reported as a major transformation product of thiamethoxam [15], the biochemical pathway responsible for its formation remains unclear. No direct evidence has demonstrated which enzyme class, such as CYP450s, nitroreductases, or other reductive systems, catalyzes the conversion of the nitroguanidine moiety to the corresponding desnitro (amine) in denitration reaction derivative as shown in Fig. 1. Therefore, this study provides the first experimental evidence suggesting that the formation of this metabolite may be associated with the CYP450 enzyme activity from A. mellifera.

The intended application of this work is to utilize a bee-derived CYP450 to decrease the effect of pesticides and increase the bee survival rate. This may be achieved either by engineering a probiotic E. coli strain as a proof-of-concept system or by adapting the approach for direct implementation within the native bee gut microbiome. Notably, the core honeybee gut symbionts include Bifidobacterium, Lactobacillus (Firm-4 and Firm-5 group), Gilliamella apicola, and Snodgrassella alvi. Among these, S. alvi has emerged as one of the most genetically tractable members of the bee gut microbiome, owing to the availability of genetic engineering toolkits reported in previous studies [24, 32]. Several works have demonstrated the use of engineered S. alvi for diverse applications, including the activation of bee immune responses against parasites [25] and the development of S. alvi-based biosensors in the bee gut [8]. Taken together with our findings, this supports the potential of engineered S. alvi harboring a CYP450 as a strategy to facilitate pesticide detoxification in the bee gut, which may contribute to improved bee survival following pesticide exposure.

Conclusion

In this work, a bee cytochrome P450 monooxygenase (CYP450) or CYP9Q1 was modified and expressed in E. coli BL21(DE3) to investigate its activity in thiamethoxam detoxification. The enzyme was engineered by truncating its N-terminal transmembrane region to enable soluble expression. However, the results indicated that high-level expression of the modified enzyme (1 mM IPTG) impaired host cell growth and resulted in inconsistent protein expression, whereas induction with 0.5 mM IPTG was optimal. HPLC analysis showed a reduction in thiamethoxam concentrations following incubation with the CYP450, while LC-MS analysis revealed several metabolites with altered abundance across conditions. Further analysis of the intracellular fraction suggested that the bee CYP450 metabolized thiamethoxam, producing desnitro thiamethoxam.

Although this study demonstrates the feasibility of using heterologously expressed bee CYP9Q1 for thiamethoxam biotransformation in E. coli, expression of eukaryotic cytochrome P450 enzymes in prokaryotic hosts remains inherently challenging and imposes a burden on the host, highlighting the need for further optimization of expression strategies. Partial enzyme engineering was already implemented in this study through N-terminal truncation to improve soluble expression; however, inconsistent expression was observed.

Future work should focus on improving detoxification efficiency through additional expression optimization or enzyme engineering, transferring the detoxification system into probiotic-compatible or bee-associated symbiotic bacteria such as S. alvi, and evaluating their colonization performance, including potential effects on gut microbiota composition and thiamethoxam tolerance.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Pachara Sattayawat: Conceptualization, Visualization, Data curation, Investigation, Methodology, Validation, Supervision, Resources, Funding acquisition, Writing – original draft, Writing – review & editing. Kasimaporn Promubon, Visualization, Investigation, Data curation, Writing – review & editing. Kritapas Sripinta: Investigation, Writing – review & editing. Warisa Chunguaongsuk: Investigation. Chonlaphon Juntama: Investigation, Writing – review & editing. Nuttapol Noirungsee: Writing – review & editing, Aussara Panya: Writing – review & editing, Methodology, Resources. Terd Disayathanoowat: Writing – review & editing, Resources, Funding acquisition.

Funding

This work was supported by the Murata Science and Education Foundation 2025 and partially supported by Chiang Mai University.

Data availability

All data that support the findings of this study are included within this paper/supplementary files.

Declarations

Competing interests

The authors declare no competing interests.