Integrated genomic, transcriptomic, and microsomal metabolic profiling of cytochrome P450s in squirrels Authors: Miku GOTO, Kanami WATANABE, Yared Beyene YOHANNES, Nanako HASEGAWA, Ayuko MORITA, Nelly BANDA, Keita KAKISAKA, Masaki ITO, Yoshinori IKENAKA, Shouta MM NAKAYAMA, Mayumi ISHIZUKA Published: 2026 ABSTRACT AbstractSquirrels (family Sciuridae; hereafter referred to as sciurids) are frequently treated with veterinary medical products, such as fipronil, a topically applied insecticide used for ectoparasite control; however, fundamental information on their drug-metabolizing capacity remains limited. In this study, we examined whether hepatic cytochrome P450 (CYP) profiles and microsomal fipronil metabolism differ among sciurids, Eurasian red squirrel (Sciurus vulgaris orientis), Siberian flying squirrel (Pteromys volans orii), Pallas’s squirrel (Callosciurus erythraeus), and laboratory rodents (Mus musculus and Rattus norvegicus). We integrated sequence, expression, and functional analyses with a primary focus on CYP2C and CYP3A, key enzymes involved in xenobiotic metabolism. No clear evidence of sciurid-specific loss or major divergence was detected in CYP2C/3A, and microsomal fipronil metabolic efficiency did not differ significantly among the examined species. RNA sequencing (RNA-Seq) revealed distinct hepatic CYP expression profiles beyond CYP2C/3A; notably, CYP2B represented <1% of total major CYP expression in the Eurasian red squirrel, a proportion lower than that observed in laboratory rodents. These findings provide baseline information on hepatic CYP profiles and microsomal fipronil metabolism in sciurids and may support future comparative studies of wildlife drug metabolism.Keywords: cytochrome P450, mRNA expression, squirrel, veterinary drug, xenobiotic metabolism INTRODUCTION Sciurids, members of the family Sciuridae and commonly known as squirrels, are widely distributed mammals and include many regionally endemic species. Sciurids can affect ecosystems and public and animal health as reservoir hosts of zoonoses such as plague and tularemia, which are transmitted by ectoparasites [19, 21, 35]. Therefore, ectoparasite control in sciurids remains important for disease prevention and animal management. Fipronil (Fip), a phenylpyrazole insecticide used for ectoparasite control, has been used in wildlife and sciurid studies [9, 21, 31]. Drug-metabolizing capacity in sciurids remains poorly characterized, and dosing regimens are sometimes extrapolated from other rodent species [23, 32]. Extrapolating across species poses a risk of unintended toxicity due to differences in metabolic traits. A mass mortality event has been reported in Japanese squirrel (Sciurus lis), an endemic species in Japan, following the use of multiple veterinary drugs and insecticides, including Fip [36]. Risk assessments for Fip have primarily been conducted using laboratory rodents, such as mice and rats [1]. Reports of poisoning in sciurids suggest that, even within Rodentia, determinants of drug response and toxicokinetics may differ among species, leaving uncertainty when extrapolating findings from laboratory and companion animals to wildlife. Thus, there is a clear need to expand evidence on drug metabolism and toxicokinetics in wildlife. Fip is primarily oxidized to fipronil sulfone by hepatic cytochrome P450 (CYP) enzymes in humans and rats; CYP3A plays a major role in humans, with a minor contribution from CYP2C [22, 29, 34]. CYPs constitute a superfamily of Phase I drug-metabolizing enzymes that catalyze oxidative biotransformation of a wide range of xenobiotics; in mammals, CYP1-CYP3 families expressed predominantly in the liver play major roles in xenobiotic metabolism, including pharmaceuticals and environmental chemicals [8, 44]. Importantly, CYP expression levels and catalytic activities are associated with chemical sensitivity, and metabolic efficiency for the same substrate can vary markedly across animal species [33, 37]. Therefore, investigating hepatic drug-metabolizing enzymes and their relationships to chemical sensitivity in sciurids is essential to prevent recurrence of adverse drug events and to improve the safety of wildlife medicine. In this study, we focused on hepatic CYP enzymes as candidate determinants of species-specific differences in chemical sensitivity in sciurids using in silico phylogenetic analysis, RNA sequencing (RNA-Seq), and an in vitro Fip metabolic assay. MATERIALS AND METHODS Chemicals Fipronil (5-amino-1-(2,6-dichloro-α,α,α-trifluoro-p-tolyl)-4-trifluoromethylsulfinylpyrazole-3‐carbonitrile, CAS No. 120068-37-3), fipronil sulfone (5-Amino-1-[2,6-dichloro-4-(trifluoromethyl) phenyl]-4-(trifluoromethylsulfonyl) pyrazole-3‐carbonitrile, CAS No. 120068-36-2), methanol, acetic acid, acetonitrile, formic acid were purchased from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). glucose-6-phosphate (G6P; cat. No. 45197000) and nicotinamide adenine dinucleotide phosphate (NADPH; cat. No. 44332000), glucose-6-Phosphate dehydrogenase (G6PDH; cat. No. 46854003) were purchased from Oriental Yeast Co., Ltd. (Tokyo, Japan). Magnesium chloride (MgCl2; cat. No. 814733) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Animals Livers from Eurasian red squirrels (ES, Sciurus vulgaris orientis, n=3) and Siberian flying squirrels (FS, Pteromys volans orii, n=4) were obtained from captive individuals that died of natural causes at Sapporo Maruyama Zoo (Sapporo, Japan). The ES samples consisted of two males and one female, and the FS samples consisted of two males and two females. Individual information is provided in Table 1, and diet information provided by Sapporo Maruyama Zoo is provided in Table 1 footnote. Individuals within each species were provided with the same dietary items, although diet composition partly differed between ES and FS. Livers from Pallas’s squirrels (PS, Callosciurus erythraeus, n=3), captured during an invasive species control research program (KAKENHI JP18H02247), were provided by Dr. Masatoshi Yasuda (Kyushu Research Center, Forestry and Forest Products Research Institute, Forest Research and Management Organization, Kumamoto, Japan). Table 1.Sample information for Eurasian red squirrels and Siberian flying squirrelsSampleSexEstimated ageTentative diagnosisTreatmentCollection yearES1Male≥8 yearsHepatitis, pneumoniaDeworming2024ES2Male1 yearCardiac abnormality-2019ES3Female3 monthsUnknownIsoflurane anesthesia2025FS1Female5 yearsGastric distension-2019FS2Female≥4 yearsTumor-2025FS3Male5 monthsPoor recovery from isoflurane anesthesia-2022FS4Male3 yearsFood stagnation-2020ES1–ES3, individual Eurasian red squirrels; FS1–FS4, individual Siberian flying squirrels. Sample information was provided by Sapporo Maruyama Zoo. ES diet: carrot, sweet potato, apple, orange, grape, Komatsuna, grain-based pigeon feed consisting of mixed dried grains, sunflower seeds, dried sardines, and walnuts. FS diet: carrot, apple, orange, grape, Komatsuna, and grain-based pigeon feed consisting of mixed dried grains. Sample information for Eurasian red squirrels and Siberian flying squirrels ES1–ES3, individual Eurasian red squirrels; FS1–FS4, individual Siberian flying squirrels. Sample information was provided by Sapporo Maruyama Zoo. ES diet: carrot, sweet potato, apple, orange, grape, Komatsuna, grain-based pigeon feed consisting of mixed dried grains, sunflower seeds, dried sardines, and walnuts. FS diet: carrot, apple, orange, grape, Komatsuna, and grain-based pigeon feed consisting of mixed dried grains. Male SD rats (250–300 g) and ICR mice (35–40 g) (7 weeks old, n=3/species; Sankyo Lab Service, Sapporo, Japan) were acclimated for at least one week before experiments. They were housed under controlled conditions (22 ± 1°C, 60 ± 5% humidity, 12 hr reversed light/dark cycle) with ad libitum access to food (CE-2; CLEA, Tokyo, Japan) and water. The livers were collected after euthanasia by cervical dislocation under deep anesthesia with isoflurane. All livers were stored at −80°C until use. All animal care and experimental procedures were performed in accordance with the Guidelines of the Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC) and were approved by the Animal Care and Use Committee of the Graduate School of Veterinary Medicine, Hokkaido University (approval numbers: 21-0070, 25-0040). RNA sequencing analysis of hepatic CYP expression RNA extraction, library preparation, and sequencing: Total RNA was extracted from ES (n=3) and FS (n=4) liver tissue (approximately 20 mg; individual information provided in Table 1) using the FastGene RNA Basic Kit with DNase digestion (Nippon Genetics Co., Ltd., Tokyo, Japan). Following NanoDrop and Bioanalyzer QC (Novogene, Tokyo, Japan), libraries were prepared using the TruSeq Stranded mRNA Kit (Illumina, San Diego, CA, USA) and sequenced on a NovaSeq X Plus platform (150-bp paired-end, 6 Gb). All RNA samples used for library preparation had RNA integrity number (RIN) values ranging from 4.3 to 8.8 (Supplementary Table 1) and met the quality control criteria for RNA-Seq library preparation. The mapping-based workflow for ES and the de novo assembly-based workflow for FS are described in the following subsections. For the mapping approach, the gray squirrel (Sciurus carolinensis) genome was used as reference. Because ES belongs to the same genus as the reference species and achieved a mapping rate of ≥70% (Supplementary Table 1), transcript abundance was estimated as transcripts per million (TPM) using the mapping-based pipeline. To avoid assembly-related artifacts commonly seen in highly similar CYP paralogs, a reference-based workflow was adopted. In contrast, FS showed a mapping rate of ≤40%, lacking a closely related reference. Therefore, to minimize reference bias in reference-dependent expression quantification, FS was analyzed using a de novo assembly-based workflow. Expression was summarized by CYP subfamily, except for CYP1A (quantified at isoform levels: CYP1A1/2) due to pronounced sex-associated variation. Read processing and mapping-based analysis for ES: Raw reads were quality-checked (FastQC v0.11.9) and trimmed (fastp v0.23.2; Phred <20 and adapter removal). ES reads were aligned to the gray squirrel (Sciurus carolinensis) reference genome (RefSeq GCF_902686445.1) using STAR (v2.7.10a) with default parameters. Transcript abundance was quantified using StringTie (v2.2.1) guided by the reference annotation, and expression values were summarized as TPM at the CYP subfamily level. For candidate genes showing sample-specific detection, read alignments were visually inspected in Integrative Genomics Viewer (IGV), and splice-junction support was checked using STAR junction output (SJ.out.tab). Since the target species is not annotated in NCBI, sample-specific CYP3A coding sequences reconstructed from RNA-Seq data were incorporated into the phylogenetic analysis, variants were called from STAR-aligned reads with bcftools mpileup/call (v1.13) using the same reference assembly, and consensus sequences were generated with bcftools consensus. Based on the reference General Feature Format version 3 (GFF3) annotation, coding sequence exons for target CYP3A genes were extracted and concatenated using samtools faidx (v1.6). For phylogenetic analysis, the ES among 3 compared individuals with the highest RIN and best mapping metrics (ES2) was used as a representative sequence. De novo transcriptome analysis for FS: Due to the low mapping rates of FS reads against the reference genome, a de novo transcriptome was assembled from trimmed reads using Trinity (v2.8.5). Assembly statistics were obtained with TrinityStats.pl, and assembly completeness was assessed using Benchmarking Universal Single-Copy Orthologs (BUSCO, v5.4.2) with the mammalia_odb10 dataset. For expression quantification, trimmed reads from each FS sample were quasi-mapped to the Trinity-assembled de novo transcriptome using Salmon (v0.14.1), and transcript abundance was calculated as TPM. Open Reading Frames (ORFs) were predicted using TransDecoder (v5.5.0) and translated peptides were annotated by BLASTP (BLAST+ v2.15.0; E-value <1.0 × 10−12, max_target_seqs=1) against custom proteomes (gray squirrel; yellow-bellied marmot, Marmota flaviventris; human, Homo sapiens; Norway rat, Rattus norvegicus). Candidate CYP2C- and CYP3A-like transcripts were further identified by TBLASTN (E-value <1.0 × 10−5) using CYP2C/CYP3A proteins from multiple reference species as queries and screened by BLASTP against a curated CYP2C/CYP3A reference set from gray squirrel using thresholds of E-value <1.0 × 10−20, query coverage ≥50%, and ORF length ≥250 aa. Similarly, for Trinity genes with multiple isoforms, a single representative isoform was selected based on the highest BLAST bitscore (ties resolved by longer ORF length). Retained coding sequences were used for downstream phylogenetic analyses in MEGA. Detailed metrics for RNA integrity, sequencing, and mapping (or assembly) quality are provided in Supplementary Data (Supplementary Tables 1–2). CYP2C and CYP3A sequence collection and phylogenetic analysis CYP2C and CYP3A sequences from ES and FS (RNA-Seq) and additional 18 mammalian species were retrieved from NCBI: the arctic ground squirrel, gray squirrel, thirteen-lined ground squirrel, yellow-bellied marmot, naked mole-rat, domestic guinea pig, American beaver, Norway rat, European wood mouse, house mouse, Chinese hamster, domestic rabbit, European hare, raccoon dog, brown bear, domestic cattle, domestic cat, and human by comparing annotated gene data for each assembly (all scientific name, assembly and annotation details are provided in Supplementary Table 3). To provide phylogenetic context beyond Rodentia, we included representative non-rodent mammals: lagomorphs (domestic rabbit and European hare) as a close relative within Glires, and additional comparator species with distinct ecological and dietary traits, including omnivores (raccoon dog and brown bear), a carnivore (cat), and a herbivore (cattle) [10]. CYP2C and CYP3A homologs were identified via NCBI BLASTN searches using rat CYP2C24 and CYP3A9 as queries. For each species, BLASTN (optimized for highly similar sequences) was performed against the NCBI nucleotide (nr/nt) database using default parameters. The deduced amino acid sequences were aligned using MEGA X (https://www.megasoftware.net/, accessed on 3 November 2025) with MUSCLE. Pseudogenes, sequences shorter than 60% of the average length, and those with premature stop codons were excluded. Human CYP1A1 and CYP1A2 served as outgroups; alignment gaps introduced by them were removed. The resulting alignment was subjected to model selection using the Bayesian Information Criterion (BIC), and maximum-likelihood trees were inferred under the Jones-Taylor-Thornton (JTT) + G model with 100 bootstrap replicates. Because gene symbols derived from LOC identifiers were inconsistent among species, sequence names were reassigned based on their positions within the inferred clades. Microsomal fipronil metabolism assay Fip biotransformation assay using liver microsomes was performed with minor modifications to the method of Suzuki et al. [2, 33]. Microsomal fractions were prepared from male SD rats, ICR mice, FS, and PS (n=3/species). ES and certain FS individuals were excluded due to insufficient liver tissue post-RNA extraction (<1.5 g). CYP concentration was determined by CO-difference spectroscopy [25]. A mixture of Fip substrate (final concentrations: 6.25–2,400 μM in 0.1 M potassium phosphate buffer (KPB)), 0.1 M KPB, 3 mM MgCl2, 5 mM G6P, and 1 mg/mL hepatic microsomes from each individual of each species was pre-incubated for 5 min in a thermo shaker. The reaction was started by adding the mixture with 2 IU/mL G6PDH and 0.5 mM β-NADPH, followed by incubating for 30 min. The temperatures for pre-incubation and incubation were maintained at 37°C. 1% formic acid in acetonitrile was added to stop the reaction. After 15 min on ice, mixtures were centrifuged (15,000 × g, 10 min). Subsequently, 25 µL of 200 ppb internal standard mixture containing two compounds for liquid chromatography-tandem mass spectrometry (LC/MS/MS) analysis (Fipronil-13C4, Fipronil sulfone-13C4-15N2) was added and vortexed. The mixture was filtered through a NANOSEP centrifugal filter (pore size: 0.2 µm, Pall corporation, Tokyo, Japan) by centrifugation at 15,000 × g for 10 min. The concentration of fipronil sulfone was measured using LC/MS/MS, and the results were analyzed, and maximum velocity (Vmax) and Michaelis-Menten constant (Km) values were calculated using GraphPad Prism 8 (GraphPad Software, La Jolla, CA, USA). Analysis of fipronil sulfone by LC/MS/MS Quantification was performed using an Agilent 1290 Infinity II/Ultivo LC/MS/MS system equipped with a Poroshell 120 EC-C18 column (2.7 µm, 2.1 × 150 mm; Agilent Technologies, Santa Clara, CA, USA). Mobile phases A and B were 10 mM ammonium formate in water and methanol, respectively. Gradient and MS conditions are detailed in Supplementary Tables 4a–d. Analytes were quantified using internal standards (fipronil-13C4 and fipronil sulfone-13C4-15N2). The calibration curves were linear over the range of 2.5 to 1,600 ng/mL (r2 >0.98). The lower limit of quantification (LLOQ) was established at 2.5 ng/mL. All sample concentrations fell within the calibration range. CYP3A substrate recognition site alignment and amino acid comparison CYP3A amino acid sequences were aligned using MUSCLE in MEGA X. Retained coding sequences were manually verified for intact substrate recognition site (SRS) regions (SRS-1–SRS-6) [7], conserved motifs (ExxR and the heme-binding invariant cysteine) [40], and the absence of assembly artifacts (Supplementary Fig. 1). SRS segments were extracted using human CYP3A4 (NP_059488.2) numbering as a reference [42]. For interspecies comparisons, the highest-expressed CYP3A-like transcripts with full-length ORFs (e.g., FS CYP3A9-like) were analyzed alongside the gray squirrel (Sciurus carolinensis; NP_047388808.1) and representative rodent functional counterparts of human CYP3A4 (rat CYP3A9, NP_671739.2; mouse CYP3A11, NP_031844.1) [17, 42]. Statistical analysis Kinetic parameters, including Vmax and Km, were estimated by nonlinear regression using GraphPad Prism 8 (GraphPad Software). The Games–Howell post hoc test was used to compare CYP concentrations and kinetic parameters in the microsomal Fip metabolism assay. Statistical analyses were performed using JMP Student Edition version 18 (SAS Institute Inc., Cary, NC, USA). Data are expressed as mean ± standard deviation. Differences were considered statistically significant at P<0.05. RESULTS Phylogenetic profiles of sciurid CYP2C and CYP3A genes Phylogenetic relationships of CYP2C and CYP3A genes are shown in Figs. 1–2. CYP2C sequences from sciurids, rodents, and comparator species were divided into three major groups (Fig. 1). Group 1 included human CYP2C, brown bear and raccoon dog CYP2C41, and CYP2C sequences from sciurids and rodents. Similarly Group 2, contained CYP2C21 from these same species, whereas Group 3 consisted of CYP2C23 orthologs across multiple taxa. Within Groups 1 and 2, rodent CYP2C sequences further clustered into four clades (three in Group 1 and one in Group 2). Each clade contained all species examined from both Myomorpha and Sciuridae. In contrast, within Group 3, although Myomorpha sequences were present in Group 3, no Sciuromorpha sequences were detected in this group. Phylogenetic tree of vertebrate CYP2C. The tree was constructed using the maximum-likelihood method under the JTT + G model. The shaded backgrounds indicate the three major groups identified in this study: Group 1, pink; Group 2, yellow; and Group 3, light blue. Colored triangles represent collapsed clades corresponding to major rodent lineages, with colors denoting suborders as follows: Myomorpha, blue; Sciuromorpha, green; Ctenohystrica, red; and Castorimorpha, orange. For readability, “CYP” was omitted from some individual CYP2C isoform labels in the tree. Plural labels indicate collapsed clades mainly composed of the indicated CYP2C isoform-like sequences together with closely related CYP2C-like sequences. Numbers at nodes indicate bootstrap support values (%) based on 100 replicates; only values ≥50% are shown. Human CYP1A1 and CYP1A2 were used as the outgroup. The scale bar indicates substitutions per site. Phylogenetic tree of vertebrate CYP3A. The tree was constructed using the maximum-likelihood method under the JTT + G model. Sequences within the shaded light-green area constitute the CYP3A9 clade, which includes rodent CYP3A9/CYP3A9-like sequences and related mammalian CYP3A sequences. Colored triangles represent collapsed clades corresponding to major rodent lineages, with colors denoting suborders as follows: Myomorpha, blue; Sciuromorpha, green; Ctenohystrica, red; and Castorimorpha, orange. For readability, “CYP” was omitted from some individual CYP3A isoform labels in the tree. Plural labels indicate collapsed clades mainly composed of the indicated CYP3A isoform-like sequences together with closely related CYP3A-like sequences. Numbers at nodes indicate bootstrap support values (%) based on 100 replicates; only values ≥50% are shown. Human CYP1A1 and CYP1A2 were used as the outgroup. The scale bar indicates substitutions per site. Phylogenetic analysis of CYP3A revealed that rodent CYP3A sequences, including those from sciurids, were divided into two major groups (Fig. 2). The first group consisted of CYP3A9 sequences from rodents and closely related species, and also contained non-rodent mammalian CYP3A genes (hereafter referred to as the CYP3A9 clade). In contrast, CYP3A sequences outside the CYP3A9 clade tended to cluster by rodent suborder, including sciurids. Hepatic CYP expression patterns in ES and FS Expression quantification was conducted using a reference-based mapping approach for ES and a de novo approach for FS. RNA-Seq yielded an average of 26 million 150-bp paired-end reads per sample across 7 liver specimens from two species (Supplementary Table 1). For the FS samples, the de novo assembly yielded a transcriptome with sufficient completeness (BUSCO completeness: 68.3%) for downstream analysis (Supplementary Table 2). Using the resulting TPM values, we calculated the relative contribution of each CYP subfamily (CYP1A, CYP1B, CYP2A–2E, and CYP3A) within individuals (Fig. 3, Supplementary Figs. 2, 3). Transcripts of CYP2C and CYP3A were reliably detected in all liver samples from both sciurid species. Conversely, CYP1B transcripts were generally at low levels across both species. Composition of hepatic xenobiotic-metabolizing CYP families (CYP1A, CYP1B, CYP2A–2E, and CYP3A) quantified as transcript abundance in individual Eurasian red squirrels (ES1-3) and Siberian flying squirrels (FS1-4) based on RNA-Seq expression profiles. Values represent the relative contribution of each CYP family/subfamily to the total transcripts per million of the analyzed hepatic CYPs within each individual. CYP1A expression, particularly CYP1A2, varied among individuals and appeared relatively high in ES1, ES2, FS3, and FS4 (Supplementary Fig. 3). These individuals were male; however, the present sample size was insufficient to evaluate sex-related differences. At the species level, ES and FS showed different hepatic CYP compositional patterns in this dataset. In ES, CYP2C and CYP3A constituted the major hepatic fractions, whereas CYP2E accounted for only 6–15% and CYP2B contributed less than 1%. In FS, CYP2E was the most prevalent subfamily across FS1, FS2, and FS4, with CYP2C and CYP3A also contributing substantially to the total CYP pool. Inter-individual variation was observed within both species. In ES, CYP1B transcripts were detected exclusively in ES1, accounting for approximately 6% of the analyzed CYP subfamilies. ES2 exhibited a higher contribution of CYP1A and a relatively lower contribution of CYP3A compared with other ES individuals. In FS, FS3 showed an atypical profile characterized by reduced CYP2E dominance and relatively increased contributions of CYP1A and CYP2C/3A. Microsomal fipronil sulfone formation across species Microsomal CYP concentrations are summarized in Table 2. FS showed a significantly lower CYP concentration than SD rats and ICR mice and tended to be lower than PS (P=0.08). The CO-difference spectra of FS microsomes showed lower 450-nm peak heights than those of the other species, resulting in lower calculated CYP concentrations, but no obvious abnormal spectral profiles were observed (data not shown). Comparison of kinetic parameters (Vmax, Km, and Vmax/Km) among species (n=3) revealed no statistically significant differences for any metric (Fig. 4 and Table 3). However, FS exhibited a trend toward a lower Km (19.25 ± 25.52 μM) than mice (210.9 ± 60.82 μM; P=0.058). In contrast to this trend, no corresponding difference was observed in Vmax/Km. Table 2.Cytochrome (CYP) concentration in liver microsomes across species (n=3)AnimalSD ratICR mouseFSPSCYP concentration1.00 ± 0.10a1.06 ± 0.17a0.17 ± 0.11b2.02 ± 0.64abValues are represented as mean ± SD in nmol CYP/mg protein. Different superscript letters indicate significant differences among species, as determined by the Games-Howell post hoc test (P<0.05). SD rat, Sprague-Dawley rat; ICR mouse, Crl:CD1 mouse; FS, Siberian flying squirrel; PS, Pallas’s squirrel. Cytochrome (CYP) concentration in liver microsomes across species (n=3) Values are represented as mean ± SD in nmol CYP/mg protein. Different superscript letters indicate significant differences among species, as determined by the Games-Howell post hoc test (P<0.05). SD rat, Sprague-Dawley rat; ICR mouse, Crl:CD1 mouse; FS, Siberian flying squirrel; PS, Pallas’s squirrel. Michaelis-Menten kinetics of fipronil metabolism in liver microsomes from SD rat, ICR mouse, Siberian flying squirrel (FS), and Pallas’s squirrel (PS). Velocity represents the formation rate of fipronil sulfone normalized to microsomal CYP content (per nmol CYP). Data are mean ± SD (n=3). Curves were fitted by nonlinear regression to the Michaelis-Menten equation. Table 3.Kinetic parameters for fipronil sulfone formation in hepatic microsomes across species (n=3)IndicesSD ratICR mouseFSPSVmax206.2 ± 119.51262.2 ± 99.65165.3 ± 137.94335 ± 200.13Km104.3 ± 91.05210.9 ± 60.8219.25 ± 25.52289.9 ± 288.00Vmax/ Km1.98 ± 1.111.24 ± 0.448.59 ± 10.121.16 ± 0.79Values are expressed as mean ± SD. Vmax, Km, and Vmax/Km are expressed as nM/min/nmol cytochrome P450 (CYP), µM, and 10−3/min/nmol CYP, respectively. Vmax, Km, and intrinsic clearance (Vmax/Km) were estimated by nonlinear regression fitted to Michaelis–Menten kinetics using GraphPad Prism 8. No significant differences were detected among species for any kinetic parameter, as determined by the Games–Howell post hoc test (P>0.05). SD rat, Sprague–Dawley rat; ICR mouse, Crl:CD1 mouse; FS, Siberian flying squirrel; PS, Pallas’s squirrel. Kinetic parameters for fipronil sulfone formation in hepatic microsomes across species (n=3) Values are expressed as mean ± SD. Vmax, Km, and Vmax/Km are expressed as nM/min/nmol cytochrome P450 (CYP), µM, and 10−3/min/nmol CYP, respectively. Vmax, Km, and intrinsic clearance (Vmax/Km) were estimated by nonlinear regression fitted to Michaelis–Menten kinetics using GraphPad Prism 8. No significant differences were detected among species for any kinetic parameter, as determined by the Games–Howell post hoc test (P>0.05). SD rat, Sprague–Dawley rat; ICR mouse, Crl:CD1 mouse; FS, Siberian flying squirrel; PS, Pallas’s squirrel. For FS, microsomes were prepared from the same individuals used for RNA-Seq (FS1, FS3, and FS4). Considerable inter-individual variability was observed across kinetic parameters, particularly in Vmax/Km. Specifically, FS3 exhibited a substantially higher Vmax/Km (21.02 nM/min/nmol CYP) than FS1 and FS4 (3.11 and 3.90 nM/min/nmol CYP, respectively). Consistent with the RNA-Seq data, FS3 showed the highest relative contribution and absolute expression of CYP2C and CYP3A (Fig. 3; Supplementary Fig. 2). Amino acid variation in CYP3A substrate recognition sites Comparisons of substrate recognition sites using human CYP3A4 numbering revealed several largely conservative substitutions in SRS-1 and SRS-6, particularly in FS SRS-1 (Tables 4, 5). Table 4.Comparison of CYP3A substrate recognition site 1 (SRS-1) amino acid sequences in sciurids and laboratory rodentsSRS-1, 101-123101102103104105106107108109110111Human CYP3A4VFTNRRPFGPVES2 CYP3A9-likeFFTNRRSFGPVGray squirrel CYP3A9-likeLFTNRRSFGPVFS CYP3A9-likeAFTNRRDMGPLRat CYP3A9TFTNRRNFGPVMouse CYP3A11VFTNRRDFGPVSRS-1, 101-123112113114115116117118119120121122123Human CYP3A4GFMKSAISIAEDES2 CYP3A9-likeGFMKSAISLSEDGray squirrel CYP3A9-likeGFMKSAISLSEDFS CYP3A9-likeGFMRKALAFTEDRat CYP3A9GILKKAISISEDMouse CYP3A11GIMSKAISISKDSRS positions were defined according to human CYP3A4 numbering. The table shows SRS-1 positions 101-123. Sequences include Siberian flying squirrel (FS) CYP3A9-like, Eurasian red squirrel (ES2) CYP3A9-like, gray squirrel CYP3A9-like, rat CYP3A9, mouse CYP3A11, and human CYP3A4. The gray squirrel CYP3A9-like sequence corresponds to LOC124969866. Highlighted residues indicate positions where sciurid residues differ from both rat and mouse residues. Comparison of CYP3A substrate recognition site 1 (SRS-1) amino acid sequences in sciurids and laboratory rodents SRS positions were defined according to human CYP3A4 numbering. The table shows SRS-1 positions 101-123. Sequences include Siberian flying squirrel (FS) CYP3A9-like, Eurasian red squirrel (ES2) CYP3A9-like, gray squirrel CYP3A9-like, rat CYP3A9, mouse CYP3A11, and human CYP3A4. The gray squirrel CYP3A9-like sequence corresponds to LOC124969866. Highlighted residues indicate positions where sciurid residues differ from both rat and mouse residues. Table 5.Comparison of CYP3A substrate recognition site 6 (SRS-6) amino acid sequences in sciurids and laboratory rodentsSRS-6, 476-485476477478479480481482483484485Human CYP3A4KLSLGGLLQPES2 CYP3A9-likeKFGKQGLLQPGray squirrel CYP3A9-likeKFNKQGLLQPFS CYP3A9-likeKFSKQGFLQPRat CYP3A9KLSKQGLLQPMouse CYP3A11KLSRQGLLQPSRS positions were defined according to human CYP3A4 numbering. The table shows SRS-6 positions 476-485. Sequences include Siberian flying squirrel (FS) CYP3A9-like, Eurasian red squirrel (ES2) CYP3A9-like, gray squirrel CYP3A9-like, rat CYP3A9, mouse CYP3A11, and human CYP3A4. The gray squirrel CYP3A9-like sequence corresponds to LOC124969866. Highlighted residues indicate positions where sciurid residues differ from both rat and mouse residues. Comparison of CYP3A substrate recognition site 6 (SRS-6) amino acid sequences in sciurids and laboratory rodents SRS positions were defined according to human CYP3A4 numbering. The table shows SRS-6 positions 476-485. Sequences include Siberian flying squirrel (FS) CYP3A9-like, Eurasian red squirrel (ES2) CYP3A9-like, gray squirrel CYP3A9-like, rat CYP3A9, mouse CYP3A11, and human CYP3A4. The gray squirrel CYP3A9-like sequence corresponds to LOC124969866. Highlighted residues indicate positions where sciurid residues differ from both rat and mouse residues. Conversely, SRS-2 to SRS-5 were highly conserved across sciurids (FS, ES, and the gray squirrel) and laboratory rodents (Supplementary Tables 5a–d). Within SRS-1, the residue corresponding to position 119 was substituted from serine (S) to alanine (A) in FS (S119A; Table 4). In addition, within SRS-6, ES2 exhibited a specific substitution at position 478, resulting in a serine (S) to glycine (G) change (S478G; Table 5). DISCUSSION Evolutionary conservation of CYP2C and CYP3A in sciurids The CYP2C phylogeny observed in this study was broadly consistent with previous phylogenetic studies reporting a three-group separation of mammalian CYP2C genes [15, 38]. Within Groups 1 and 2, sciurid and myomorph CYP2C sequences were included in shared rodent clades, suggesting high sequence similarity among the examined species. Because phylogenetic clades often correlate with shared catalytic properties [15], these results suggest that functional divergence of CYP2C enzymes between sciurids and other rodents may be limited at the sequence level. A notable finding was that no sciurid CYP2C23 ortholog was identified in the present dataset. CYP2C23 is known to preferentially metabolize endogenous substrates, such as arachidonic acid, rather than xenobiotics [12]. This observation raises the possibility that sciurids differ from other rodents in enzymatic pathways related to endogenous lipid metabolism, including epoxygenase activity. The presence of sciurid sequences within the broadly conserved CYP3A9 clade, together with non-rodent mammalian CYP3A genes, is consistent with previous comparative CYP studies [38] and suggests that this core CYP3A lineage has been maintained across mammals. This conservation is also consistent with the proposed functional relationship between rat CYP3A9 and human CYP3A4 [4, 42], and provides a possible sequence-level context for the lack of significant interspecies differences in microsomal Fip metabolic efficiency observed in this study. CYP3A sequences outside the CYP3A9 clade conversely clustered by rodent suborder, suggesting lineage-specific duplication and diversification within Rodentia. However, CYP3A enzymes are characterized by a large and flexible substrate-binding cavity and marked substrate promiscuity [41]. Suborder- or species-specific duplication and sequence divergence do not necessarily imply major functional loss or pronounced phenotypic differences in metabolic capacity. Taken together, the phylogenetic analyses suggest that core CYP2C and CYP3A lineages are largely conserved across rodents, including sciurids. Despite potential lineage-specific diversification in certain isoforms (e.g., CYP2C23 and non-CYP3A9), sequence-level differences in catalytic function are likely modest. The present data therefore does not indicate marked sequence-level divergence in major CYP2C/3A lineages and provides an evolutionary framework for interpreting the microsomal Fip metabolism results. Interpretation of hepatic CYP expression profiles in sciurids Because the present analysis was based on a limited number of non-laboratory sciurid samples, the hepatic CYP expression profiles are best interpreted as exploratory baseline data rather than definitive evidence of species- or sex-dependent differences. Consistent with our phylogenetic findings, CYP2C and CYP3A, key enzymes in Fip metabolism, were detected in the livers of both sciurid species. This finding suggests that species-specific loss or markedly reduced expression of these subfamilies is unlikely to be a primary determinant of interspecies differences in Fip metabolism among the animals examined in this study. This interpretation is also consistent with the absence of significant interspecies differences in microsomal Fip metabolic efficiency observed in the present study. CYP1A expression tended to be higher in male samples than female samples in the present dataset, although this pattern cannot be clearly attributed to sex-related differences because of the limited sample size. This pattern should be carefully considered given the known sex-related and inducible variation of hepatic CYP1A in laboratory rodents [6, 11, 18]. Further studies using a larger number of individuals with well-documented information on sex, diet, exposure history, and health status will be needed to evaluate whether this pattern is reproducible. At the species level, the major hepatic CYP composition in FS generally aligned with that of laboratory rodents such as mice and rats [26, 37], dominated by CYP2E. In contrast, relatively lower CYP2B is implicated in the metabolism of plant-derived compounds [14, 20] and several xenobiotics such as propofol [3, 24], the low contribution of CYP 2B in ES may be noteworthy. Given that ES consumes plant-derived components [39], this observation may provide a useful clue for future studies on species-specific xenobiotic metabolism in sciurids. However, it is important to note that mRNA transcript levels do not always strongly correlate with active protein levels, as CYP enzymes (such as CYP2E) are strongly regulated post-translationally [28]. The low mRNA expression of CYP2E and CYP2B in ES does not necessarily translate to reduced protein abundance or enzymatic activity, highlighting the need for further substrate-specific functional assays. Marked inter-individual variation was also observed. For instance, the detection of CYP1B exclusively in ES1 may be linked to its history of chronic hepatitis, as CYP1B1 is enriched in non-parenchymal cells during hepatic inflammation and fibrogenesis [5, 16, 27]. Although this aligns with the individual’s medical history, the lack of confirmed splice-junction support necessitates cautious interpretation and future qPCR validation. Sample FS3 exhibited an atypical profile, featuring reduced CYP2E dominance and relatively increased contributions of CYP1A and CYP2C/3A. This distinct pattern, along with other observed variations, highlights the inherent challenges associated with using wild or non-laboratory samples. Multiple uncontrolled factors, including prior drug exposure, nutritional status, and tissue preservation quality, likely contribute to this variability. Overall, CYP2C and CYP3A, which are involved in Fip metabolism, were detected in both species. These transcriptomic data did not suggest species-specific loss or very low expression of these subfamilies. In contrast, other CYP subfamilies, including CYP1A, CYP1B, CYP2B, and CYP2E, showed inter-individual and interspecies variation. Although these findings should be interpreted cautiously, they provide useful baseline information and identify candidate CYP subfamilies for future studies on xenobiotic metabolism and chemical sensitivity in sciurids. Functional implications of microsomal Fip metabolism Our functional assays suggested that the overall hepatic oxidative metabolic capacity for Fip in sciurids was broadly comparable to that of laboratory rodents under the present in vitro conditions. First, regarding the microsomal preparations, the lower CYP concentrations observed in FS compared to laboratory rodents and PS may be partly attributable to differences in sample collection conditions. Unlike the laboratory and euthanized wild animals, the FS samples were derived from natural deaths; therefore, post-mortem protein degradation or other sample-related effects may have affected microsomal preparation quality and CYP content. Thus, the lower CYP content observed in FS is difficult to attribute solely to an inherent species difference. Despite differences in absolute CYP concentrations, the kinetic analysis did not support marked interspecies differences in overall Fip metabolic efficiency under the present experimental conditions. Notably, we observed considerable inter-individual variability in metabolic capacity among the FS samples. FS3, which exhibited a substantially higher Vmax/Km, also showed the highest relative contribution and absolute expression of CYP2C and CYP3A in our RNA-Seq analysis. Although definitive identification of the CYP isoforms responsible for Fip metabolism in sciurids requires further investigation, including inhibition assays, these functional results align well with our transcriptomic and phylogenetic data, suggesting that variation in CYP2C/3A expression may be associated with inter-individual differences in metabolic capacity. The overall conservation of CYP3A SRS regions, despite several substitutions in SRS-1 and SRS-6, is consistent with the absence of significant interspecies differences in microsomal Fip metabolic efficiency. Specifically, the SRS-1 S119A substitution is known to affect kinetic parameters like Vmax [13, 30]. Since the substrate (steroids versus Fip) and the experimental system (recombinant enzymes versus liver microsomes) differ from those in previous research, the kinetic trend, which FS showed cannot be definitively attributed to the S119A substitution. Thus, the FS-specific S119A mutation might have contributed to a higher substrate affinity (lower Km), but any impact on overall catalytic efficiency under these conditions appears limited. The individual-specific S478G substitution observed in ES2 within SRS-6 may have structural implications. Because glycine lacks a side chain and can increase local conformational flexibility, this substitution within SRS-6 may influence the plasticity of the substrate-binding pocket and/or substrate orientation [43]. Future studies employing structural modeling, docking analyses, and recombinant enzymes are required to fully elucidate the functional significance of these substitutions. A previous study has reported that mice and rats, species commonly used in risk assessment, convert Fip to fipronil sulfone more efficiently than dogs and cats, in which Fip is widely used clinically [33]. In the present study, Vmax/Km values in FS and PS microsomes were comparable to those in mice and rats. This finding suggests that, from the perspective of hepatic microsomal oxidative metabolism, the present data do not support markedly reduced Fip oxidative metabolic capacity in the sciurids examined. Although microsomal assays do not reflect all in vivo toxicokinetic processes, these results provide a useful functional basis for future evaluation of species-specific Fip sensitivity in sciurids. Conclusion This study provides the first integrated evaluation using genomic, transcriptomic, and functional analyses of hepatic CYP-dependent metabolism as a determinant of chemical sensitivity in sciurids. No clear evidence of sciurid-specific loss or major divergence was detected in CYP2C/3A, and microsomal Fip metabolic efficiency did not differ significantly among the examined species. RNA-Seq analysis revealed distinct hepatic CYP expression profiles in ES and FS, including differences in CYP2E and CYP2B contributions. These findings provide baseline information on CYP sequence, expression, and microsomal Fip metabolism in sciurids and may support future comparative studies of wildlife drug metabolism. Data availability The raw liver RNA-Seq data for seven individuals of Eurasian red squirrel and Siberian flying squirrel are available in the NCBI Sequence Read Archive (SRA) under NCBI BioProject (https://www.ncbi.nlm.nih.gov/bioproject/) with accession number PRJNA1390853. Conflicts of Interest Authors have disclosed any potential conflict of interest, financial or personal, or declared that there are none to disclose, in relation to the publishing of the research. Supplementary Material Acknowledgments We express our sincere gratitude to Dr. Masatoshi Yasuda (Kyushu Research Center, Forestry and Forest Products Research Institute, Forest Research and Management Organization, Kumamoto, Japan) for his assistance with sample collection. This collection was supported by KAKENHI Grant Number JP18H02247. We also thank the members of the Laboratory of Toxicology, Department of Biomedical Sciences, Graduate School of Veterinary Medicine, Hokkaido University, for their academic support. This work was supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan awarded to M. Ishizuka (Nos. 21H04919, 22KK0163 and 25K21816), S.M.M. Nakayama (No. 20K20633, 23H03545, 23K28235 and 25K22863), and Y. Ikenaka (No. 18H0413208, 22K18425, 23H00512). JSPS Core to Core Program (AA Science Platforms) to M. Ishizuka (No. JPJSCCB20250007) and S.M.M. Nakayama (No. JPJSCCB20230006), JSPS Bilateral Open Partnership Joint Research Projects to M. Ishizuka (No. JPJSBP120259902) and S.M.M. Nakayama (No. JPJSBP120259904) and JICA Technical Cooperation Project (SMMN). REFERENCES Associated Data Supplementary Materials Data Availability Statement The raw liver RNA-Seq data for seven individuals of Eurasian red squirrel and Siberian flying squirrel are available in the NCBI Sequence Read Archive (SRA) under NCBI BioProject (https://www.ncbi.nlm.nih.gov/bioproject/) with accession number PRJNA1390853.