P. saucia larvae were collected from Luoyang, Henan Province, China, and were reared in an incubator under 16  h L: 8 h D cycle at 23°C ± 1°C and 60% relative humidity. Larvae were fed an artificial diet, and adults were provided with 10% sugar water (Choi et al., 2009). For transcriptome sequencing, male and female antennae were collected separately from 80 individuals of 2-3-day-old adult P. saucia. For real-time quantitative PCR (RT-qPCR), male and female antennae, mouthparts, and legs were collected separately from 50 to 80 individuals of 2-3-day-old adult P. saucia. All of the tissue samples were kept in a −80°C freezer until used.

Total RNA from male and female antennae was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, United States). The purity and quantity of the RNA were evaluated with an ND-2000 spectrophotometer (Nanodrop, Wilmington, DE, United States). The RNA integrity was further checked with a 2100 bioanalyzer (Agilent, Santa Clara, CA, United States). Genomic DNA was eliminated from the total RNA with DNase I (Takara, Beijing, China). mRNA was then isolated from ≥1 µg (concentration ≥50 ng/μL) of the total RNA with Dynabeads mRNA purification kit (Invitrogen, United States). RNA-seq libraries were constructed following Illumina’s library construction protocol and then sequenced on the Nova seq6000 platform (Illumina, United States) at Origingene, Shanghai, China. Sequence assembly was performed with a de novo method (Trinity v 2.11) as we previously described (Sun et al., 2020). Putative OBP transcripts were retrieved from the obtained unigenes by searching a pooled non-redundant database using BLASTX algorithm-based method (E-value < 1e-5). Open reading frames (ORFs) of the transcripts encoding candidate PsauOBPs were predicted with ORFfinder (https://www.ncbi.nlm.nih.gov/orffinder). To evaluate the expression abundance, TPM values (Transcripts Per Kilobase of exon model per Million mapped reads) of candidate PsauOBP transcripts were calculated with Salmon (v 1.4.0).

A neighbor-joining tree of candidate PsauOBPs and homologs from other lepidopteran species including Bombyx mori, Spodoptera litura, Helicoverpa armigera, and Agrotis ipsilon was constructed with MEGA 11. The evolutionary distance was calculated with the JTT matrix-based method (Jones et al., 1992). Node supports of branches were evaluated with a bootstrap method of 1,000 replicates. The constructed tree was visualized and edited in FigTree (v 1.4.2). Amino acid sequences of OBPs used in the phylogenetic analysis were listed in Supplementary Table S1.

PCR amplification of PsauGOBP2 was carried out with Premix Taq (Takara) under the following procedure: 94°C for 3 min; 34 cycles of 94°C for 20 s, 55°C for 30 s, and 72°C for 1 min. The crude PCR products were then ligated into a pGEM-T easy vector (Promega, Beijing, China) at a molar ratio of 5: 1 (insert: plasmid). The ligation products were used to transform E. coli Top10 cells. Positive colonies were selected by the T7 and SP6 primers. The colonies were then grown in LB liquid medium and custom sequenced at Origingene, Shanghai, China. The signal peptide of PsauGOBP2 is predicted with SignalP (v 5.0); Protein parameters including the molecular weight and the theoretical isoelectric point were predicted with ProtParam (https://web.expasy.org/cgi-bin/protparam/protparam). Gene-specific primers designed against the PsauGOBP2 cDNA were listed in Supplementary Table S2.

RT-qPCR of PsauGOBP2 in different chemosensory tissues was performed using a Roche LightCycler 480 System (F. Hoffmann-La Roche Ltd., Basel, Switzerland) with a mixture (final volume 20 µL) of 10 μL of TB Green Fast qPCR Mix (Takara), 0.8 μL (10 μM) of each primer, 5 ng of sample cDNA, and appropriate volume of sterilized ultrapure H₂O. The RT-qPCR program was set as: 1 cycle of 94°C for 30 s; 40 cycles of 94°C for 5 s and 60°C for 10 s; followed by 1 cycle of 95°C for 5 s, 60°C for 1 min, and 50°C for 30 s. The primers designed with Primer Premier 6.0 were listed in Supplementary Table S2. Expression levels of PsauGOBP2 in different tissues were normalized with the endogenous gene Psauβ-actin (accession number QQ472022), using the 2^−ΔΔCT method (Livak and Schmittgen, 2001). Three biological replications were performed for each tissue sample, and each biological replication was performed with three technical replicates. The data obtained from different samples were analyzed with a one-way analysis of variance (ANOVA) followed by the Tukey multiple comparison test.

For the expression of recombinant PsauGOBP2, pET-30b containing the sequence encoding mature PsauGOBP2 was used to transform BL21 E. coli cells. Protein expression was induced by the addition of IPTG (final concentration 0.4 mM) when the OD₆₀₀ value of cell culture reached about 0.8. Cells cultured for further 2–4 h at 37°C were harvested by centrifugation. PsauGOBP2 was present as inclusion bodies. 10 mL of 8 M urea and 1 mM DTT in Tris buffer (50 mM, pH 7.4) were then used to solubilise the inclusion body pellet (from 1 L of culture). The dissolved protein was dialyzed against 1 L of Tris buffer (50 mM, pH 7.4). The recombinant protein was purified on the anion exchange resins QFF, following standard protocols previously adopted for other moth OBPs (Sun et al., 2012; Sun et al., 2021).

The fluorescence binding assays were conducted on a Hitachi F-2710 with a 1 cm light path quartz cuvette. To measure the affinity of the fluorescent probe 1-NPN (N-phenyl-1-naphthylamine) to PsauGOBP2, a 2 mM solution of the protein in 50 mM Tris-HCl, pH 7.4, was titrated with aliquots of 1 mM 1-NPN in methanol to a final concentration of 20 µM. The probe was excited at 337 nm, and emission spectra were recorded between 380 and 460 nm. The dissociation constant of 1-NPN (K_1-NPN) to PsauGOBP2 was obtained by processing the data with GraphPad Prism 6.0.

Binding affinities of odorants to PsauGOBP2 were measured by competitive binding assays. A panel of 28 compounds (competitors) including the P. saucia female sex pheromone components Z11-16: Ac and Z9-14: Ac and 26 host plant volatiles were used in the assay. The CAS number, source, and purity of these compounds were listed in Supplementary Table S3. A solution of PsauGOBP2 and 1-NPN, both at the concentration of 2 mM, was titrated with 1 mM of each competitor (dissolved in methanol) at a final concentration of 10 µM (sex pheromones) or 20 µM (host plant volatiles). The dissociation constants (K_i) of competitors to PsauGOBP2 were calculated using the equation: K_i = [IC₅₀]/(1 + [1-NPN]/K_1-NPN), where [IC₅₀] is the concentration of the competitor halving (50%) the initial fluorescence value (100%), [1-NPN] is the free concentration of 1-NPN, and K_1-NPN is the dissociation constant of the protein-1-NPN complex.

A 3D structure of PsauGOBP2 was modeled with Alphafold2 (Jumper et al., 2021). The 3D structure was then evaluated by SAVES (v 6.0). Molecular docking evaluations for PsauGOBP2 with the ligands were performed with AutoDock Vina (v.1.1.2). The default parameters were set as described in the Autodock Vina manual. The top-ranked conformation which was estimated by the Vina docking score was then subjected to PyMOL (v 1.9.0) for visual analyses.

Transcriptomic data of P. saucia antennae was obtained using a Nova seq6000 platform. A total of 44.85 million and 35.53 million clean reads were produced in male and female samples, respectively. All clean reads from male and female data were then merged for de novo assembly, which generated a total of 151,541 unigenes with an N₅₀ length of 1158 bp and a mean length of 683.39 bp (Supplementary Table S4). Moreover, 27.24% of the unigenes were longer than 1000 bp based on the size distribution analysis.

By homologous searching with BLASTX, a total of 35 putative OBP transcripts were identified in the transcriptome of P. saucia antennae (Supplementary Table S5). Except for PsauOBP19, PsauOBP24, PsauOBP26, and PsauOBP30, the other 31 PsauOBPs have full length ORFs based on the presence of start codons, stop codons, and the BLASTP alignment results to other homologous OBPs. Phylogenetic analysis of 156 OBPs from different lepidopteran species including P. saucia (this study), B. mori, H. armigera, S. litura, and A. ipsilon revealed several distinct clades, where three putative PsauPBPs (PsauPBP1/2/3) were clustered with PBPs from other species; PsauGOBP1 and PsauGOBP2 were grouped with other GOBP1s and GOBP2s, respectively (Figure 1). Furthermore, we found that GOBPs share one single original lineage with PBPs (Figure 1), consistent the reports for other lepidopteran species (Vogt et al., 2015). Of the 35 PsauOBPs, 23 were classic OBPs characteristic of six conserved cysteines. Six OBPs (PsauOBP2/9/19/24/27/30) with 4 conserved cysteines were clustered into the minus-C OBP clade. PsauOBP11, PsauOBP13, and PsauOBP22 possessed two extra conserved cysteines and accordingly were clustered into the plus-C OBP clade (Figure 1; Supplementary Table S5). By contrast, PsauOBP14, PsauOBP16, and PsauOBP25 did not match any classification based on the significantly longer sequences (with 258, 337, and 244 amino acid residues, respectively) and the presence of 5 (for PsauOBP25) or 7 (for PsauOBP14 and PsauOBP16) cysteines (Supplementary Table S5).

Transcript abundance analysis based on TPM values showed that seven OBPs, including PsauPBP1, PsauPBP2, PsauPBP3, PsauGOBP1, PsauGOBP2, PsauOBP6, and PsauOBP18, had high transcript levels in the antennae with an average TPM value of >500 in male and female samples (Figure 2). Of these, PsauPBP1 (3191.4/55.5 TPM values for male/female, same below), PsauPBP2 (9151.4/281.5), PsauPBP3 (525/51.9), PsauGOBP1 (1348.2/65.8), and PsauOBP18 (1529.2/2.3) showed higher levels in males than in females. While PsauOBP6 and PsauGOBP2 showed higher TPM values in female (2437.9 and 1384.8, respectively) than in male antennae (709.7 and 702.7, respectively). Other PsauOBPs showed relatively low TPM values and their expressions were variable in male and female antennae (Figure 2).

Based on the TPM values, we focused on a highly abundant transcript in the antennae of P. saucia, PsauGOBP2. First, the full length of PsauGOBP2 was amplified from the P. saucia antennae. The ORF of PsauGOBP2 is 489 bp encoding 162 amino acids, and the predicted matured PsauGOBP2 contains 141 amino acids (Figure 3A). The molecular weight of the mature protein is 16.1 kDa with an isoelectric point of 5.06. The amino acid sequence of PsauGOBP2 has the six-cysteine signature that forms the motif C₁-X_25-30-C₂-X₃- C₃-X_36-42-C₄-X_8-14-C₅-X₈-C₆, a typical feature of classic OBPs. Further multiple alignments revealed distinct sequence similarities between PsauGOBP2 and other lepidopteran GOBP2s (Figure 3B). PsauGOBP2 exhibited the highest identity with SlitGOBP2 of S. litura (90.12%), followed by AipsGOBP2 of A. ipsilon (88.89%).

To investigate the expression profile of PsauGOBP2 in P. saucia, we measured its transcript levels in different chemosensory tissues including antennae, mouthparts, and legs of both sexes. RT-qPCR results showed that the expression of PsauGOBP2 was significantly higher in the antennae than in other tissues. Moreover, PsauGOBP2 expression was slightly higher in female antennae than in male antennae; however, the difference was not significant (Figure 4).

To obtain the recombinant protein, PsauGOBP2 encoding mature protein was cloned and ligated into the expression vector pET-30b. As shown in Figure 5A, the recombinant PsauGOBP2 was abundantly expressed in the transformed E. coli BL21 cells when induced with IPTG. After purification with anion exchange resins, an expected size of the target protein was obtained (Figure 5A). The purified protein was then used in the fluorescence binding assays.

First, the binding pocket of recombinant PsauGOBP2 was saturated by the fluorescent probe 1-NPN, resulting in a K_1-NPN value of 3.2 ± 0.19 μM (Figure 5B). Then 28 ligands including P. saucia female sex pheromone components and host plant volatiles were used as competitors to displace the probe from the binding pocket. The results indicated that the sex pheromone components Z11-16: Ac and Z9-14: Ac were the strongest ligands, with the K_i values of 4.2 ± 0.8 μM and 4.9 ± 0.6 μM, respectively (Figure 6A; Supplementary Table S6). Three host plant volatiles, phenylethyl acetate, β-myrcene, and dodecanol also showed binding affinities to PsauGOBP2, with the K_i values of 6.3 ± 0.3 μM, 8.0 ± 0.3 μM, and 13.0 ± 0.4 μM, respectively (Figure 6B; Supplementary Table S6). The other tested compounds showed no affinities to PsauGOBP2 because the IC₅₀ values were higher than 30 μM (Supplementary Table S6).

As the crystal structure of PsauGOBP2 has not yet been resolved, we applied a highly accurate modeling program, Alphafold2, to build a 3D structure of PsauGOBP2 (Figure 7A). The model evaluation demonstrated that sequence identities between the residues with queries were >80%, the pLDDT score (per-residue predicted local-distance difference test) was >90%, and the PAE value (predicted aligned error) was approximately equal to 0 Å (Supplementary Figure S1). Further PROCHECK evaluation demonstrated that 100% of the non-glycine and non-proline residues were trapped in the allowing areas and 95.4% of the amino acid residues were located in the most favored areas (Figure 7B). These results indicated that the predicted model of PsauGOBP2 was reliable and qualified for further analysis.

The structure prediction with Alphafold2 revealed that PsauGOBP2 comprised seven α-helixes (α1–α7) (Figure 7A), which is reminiscent of the structure of other moth GOBPs (Zhou et al., 2009; Zhou, 2010). Binding energy analysis showed that the docking binding energy between PsauGOBP2 and each ligand was ≤ −6 kcal.mol⁻¹ and the distances of all potential interactive residues were <4 Å. Furthermore, the molecular docking analysis found several residues in PsauGOBP2 involved in the binding with more than one ligand. Four aromatic amino acid residues, i.e., Phe-12, Phe-33, Phe-36, and Phe-118, were needed for the binding to Z11-16: Ac, Z9-14: Ac, phenylethyl acetate, β-myrcene, and dodecanol; three nonpolar amino acid residues, Ile-52, Val-114, and Ala-115, for Z11-16: Ac, Z9-14: Ac, β-myrcene, and dodecanol; two polar amino acid residues, Thr-9 for Z11-16: Ac and phenylethyl acetate and Ser-56 for Z11-16: Ac, Z9-14: Ac, and phenylethyl acetate (Figure 8, Table 1). Notably, the hydrogen bond (2.9 Å) and conjugated bond (3.5 Å) mediated the binding of PsauGOBP2 to dodecanol and phenylethyl acetate, respectively (Figure 8).