The adult P. lewisi samples were reared in a controlled environment chamber with a temperature of 25°C ± 3°C, relative humidity (RH) of 65% ± 5%, and a light-dark cycle of 14:10 (L:D). The first-hatched P. lewisi larvae were fed with 10% honey water. After the 2^nd larval instar, P. lewisi larvae were provided with L. separata larvae of the corresponding instar. At the same time, S. litura larvae were fed on fresh tobacco leaves in an artificially controlled environment maintained at 25°C ± 3°C, 40% ± 5% RH, and natural light only.

The Yunnan Tobacco 87 plants were cultivated in a controlled environment with a temperature of 25°C ± 3°C, RH of 65% ± 5%, and a light-dark cycle of 14 L:10 D. The seeds were planted in seedling trays (a hole depth of 5 cm, an upper aperture of 4 cm, and a lower aperture of 2 cm) for 30 days. Subsequently, the seedlings were transferred to two-color pots (upper aperture of 14 cm, lower aperture of 12 cm, and height of 13 cm), and then the plants continued to grow for 30 days. Tobacco plants exhibiting robust and healthy growth and devoid of any pests and diseases were selected for the experiments.

Following 6 h of starvation, six 4^th instar larvae of S. litura were affixed to the 3^rd and 4^th leaves of tobacco plants in a top-to-bottom manner. To prevent the escape of S. litura larvae, the plants were enclosed in cages with 120 mesh screens. The S. litura larvae were removed after consuming 20% of the leaves. In exposure experiments, tobacco plants were used 18–30 h after the S. litura commenced feeding. The root and soil portion of the plants were wrapped in tin foil and placed in a hermetically sealed glass container. The airflow from the air pump passed consecutively through two closed containers, one filled with activated carbon and the other withpure water. The purified and humidified air was directed into the glass containers housing tobacco plants and then out through the air outlet, with the flow rate being regulated by an airflow meter. Once the device was connected, the flow rate of the flow meter was adjusted to 400 mL/min, and the test proceeded once the flow rate had reached an equilibrium level. The gas maintained a consistent flow rate throughout the treatmentand was directly directed into the preservation box housing the P. lewisi adults.

A total of three distinct odor sources for conducting the odor exposure experiments. The first group served as a control, devoid of any odorants (CK group). The healthy and S. litura-infested tobacco plants were used as the source of odor in the HT and IT group, respectively. Each odor exposure experiments group comprises 15 mature P. lewisi individuals (male or female), with three biological replicates. After 1 h of odor exposure treatment, the entire head of the P. lewisi adult samples was collected and promptly placed in RNase-free tubes and stored in liquid nitrogen.

Total RNA was extracted from preserved tissues using TRIzol reagent (Invitrogen, Carlsbad, CA, United States) following the manufacturer’s instructions. The integrity and quantity of RNA samples were assessed by using 1% agarose gel electrophoresis and a spectrophotometer (Eppendorf Bio Photometer Plus, Hamburg, Germany), respectively. A complementary DNA (cDNA) was synthesized from 1 μg of total RNA via reverse transcription, using a Prime-Script II 1st Strand cDNA Synthesis Kit (TaKaRa Bio, Otsu, Japan) according to the pamphlet instructions. This kit uses DNase in the initial step to eliminate the effect of DNA on qRT-PCR.

All samples were conveyed to Personalbio Technology Co. Ltd. (Shanghai, China) for Illumina sequencing using a 100 bp paired-end sequencing strategy. The sequencing data was filtered to obtain high-quality, clean reads for subsequent analysis. For unreferenced transcriptome sequencing, the clean reads were transcribed using Trinity software for further analysis. The unigenes were functionally annotated using the NR, GO, KEGG, eggNOG, Swiss-Prot, and Pfam databases. The number of reads for each sample was compared to each gene to calculate the FPKM value. Differentially expressed genes were identified by the DESeq software package, with thresholds of log2 (fold change) ≥ 1, and P-value ≤0.05 set significant differences.

The transcriptome annotation results were used to select the unigenes that included the annotated content of “odorant-binding protein”, “odorant binding protein”, and “OBPs”. The identification of potential open reading frames (ORFs) and their corresponding amino acid sequences was determined using the ORF FINDER (https://www.ncbi.nlm.nih.gov/orffinder/). Hemipteran-based phylogenetic analysis of PlewOBPs was performed using the MEGA-X program. The N-terminal signal peptide sequences of PlewOBPs were predicted using the SignalP-5.0 server (http://www.cbs.dtu.dk/services/SignalP/). The mature amino acid sequences were aligned using the MAFFT multiple sequence alignment (https://mafft.cbrc.jp/alignment/server/). A neighbor-joining tree was constructed using the p-distance model and pairwise deletion of gaps. The bootstrapping process was performed by resampling the amino acid positions of 1,000 replicas, and branches with bootstrap cutoff of <50% were collapsed. The tree was ultimately examined and modified using the Evolview-v2 (https://evolgenius.info//evolview-v2/) and Adobe Illustrator CC 2019 program.

The relative mRNA expression levels of 15 full-length PlewOBPs genes were determined using qRT-PCR under various treatments. A qRT-PCR reaction was performed on a LightCycler^® 96 System using the Hieff^® qPCR SYBR Green Master Mix (No Rox) (Yeasen, Shanghai, China). Each reaction was systematically run in triplicates with three independent biological replicates. Gene-specific primers of PlewOBPs were designed using the Primer-BLAST service (https://www.ncbi.nlm.nih.gov/tools/primer-blast/index.cgi?LINK_LOC=BlastHome). The comparative 2^−ΔΔCT method was used to calculate the relative transcript levels in each sample. The data was assessed using a one-way analysis of variance (ANOVA) and Tukey’s honestly significant difference (HSD) post hoc test.

A grand number of 765,095,950 Illumina paired-end reads were generated from 18 P. lewisi libraries. Each library was sequenced using samples obtained from the heads of 15 adult individuals (female or male) (Figure 1). After the trimming of adaptors and the filtration of low-quality reads, a total of 754,299,698 high-quality reads were obtained with 96.79% Q30 bases. Subsequently, we used the entire set of reads to create a de novo transcriptome assembly using Trinity (see methods). The assembly process resulted in the formation of 156,008 transcripts, which collectively had 276 million base pairs (bp). These transcripts encoded a total of 75,039 unigenes, with an N50 value of 1728 bp (Supplementary Table S1).

The annotation process involved using BLAST and HMMER software to perform local alignments against various databases, including NR, GO, KEGG, Pfam, eggNOG, and Swissprot. This resulted in annotation for 19,762 (26.34%), 10,287 (13.71%), 7,766 (10.35%), 8,898 (11.86%), 15,770 (21.02%), and 10,639 (14.18%) unigenes, respectively (Figure 2A). Out of the total data set, 55,277 unigenes (75.37%) did not receive a BlastX hit, possibly due to misassembly or insufficient representation in the NR database. The assembled transcriptome of P. lewisi showed a significant resemblance to Halyomorpha halys (Hemiptera: Pentatomidae), with 12,221 (61.84%) of the annotated unigenes showing the closest similarity to that particular species (Figure 2B). It was expected as H. halys has the most thoroughly annotated genome among of stink bugs in the pentatomidae family (Sparks et al., 2020; Paula et al., 2016).

Further functional classification of gene ontology (GO) was performed. The GO database consists of three major categories: biological processes (BP), cellular components (CC) and molecular function (MF). In total, 24 biological processes, 20 cellular components and 14 molecular functions were identified (Figure 2C). The categories more directly related to insect olfaction in the list of BP were a response to stimulus and localization. The unigenes exhibited enrichment in cellular processes, biological regulation and metabolic processes in the BP. Most of the unigenes in the CC list were associated with cells, cell components, and organelles. Finally, in the list of MF, these were annotated to the functional classes of binding, catalytic activity, and transcription regulator activity. These predominant GO annotations in BP, CC, and MF were comparable to those observed in the antennal transcriptomes of the H. halys, Adelphocoris lineolatus (Hemiptera: Miridae), and Apolygus lucorum (Hemiptera: Miridae) (Paula et al., 2016; Xiao et al., 2017; Chen et al., 2017; Ji et al., 2013). This similarity is noteworthy, considering that the transcriptomes were obtained from the head of P. lewisi. However, the annotations differed from the predominant BP and MF detected in the antennal transcriptomes of Chinavia ubica (Hemiptera: Pentatomidae), Dichelops melacanthus (Hemiptera: Pentatomidae), or Euschistus heros (Hemiptera: Pentatomidae) (Farias et al., 2015). The GO enrichment variations could be attributed to their distinct dietary preferences.

OBPs are compact, globular, water-soluble proteins with a signal peptide at the N-terminal region and six Cys residues in conserved positions (Zhou et al., 2008). The Cys motif is a highly conserved tertiary protein structure consisting of six α-helices coordinated by three disulfide bridges (Zhou et al., 2004; Xu et al., 2003; Wang et al., 2014). It is commonly employed as a signature for the identification of OBP. Based on the sequence similarity to insect OBPs, a total of 15 PlewOBPs were identified in the transcriptome. The number of PlewOBPs is much lower compared to H. halys (30 HhalOBPs), A. lineolatus (31 AlineOBPs), and A. lucorum (38 AlucOBPs) (Paula et al., 2016; Yuan et al., 2015; Xiao et al., 2018; Gu et al., 2011). A plausible explanation is that their total RNA is more comprehensive than ours due to the extraction of RNA from various all developmental stages or tissues or because they possess whole genome sequences. Picromerus lewisi had a similar number of OBPs to other stink bugs, specifically 19 in Tropidothorax elegans (Hemiptera: Lygaeidae), 18 in Cyrtorhinus lividipennis (Hemiptera: Miridae), and 17 in Rhodnius prolixus (Hemiptera:Reduviidae) (Song et al., 2018; Wang et al., 2018; Wang et al., 2017; Oliveira et al., 2017).

All 15 PlewOBPs have intact ORFs with lengths ranging from 396 to 663 bp. The authenticity of the nucleotide sequences of all PlewOBPs was confirmed by cloning and sequencing. Out of the 15 PlewOBPs, 11 of them have a signal peptide at their N-terminal. Like other hemipterans OBPs, PlewOBPs sequences were categorized into two types based on the presence of the characteristic OBP Cys signature: “classic” OBP and “plus-C” OBP. Based on the hemipteran “classic” OBP Cys motif (C1-X_22-32-C2-X₃-C3-X_36-46-C4-X8-14-C5-X₈-C6), we have classified 12 PlewOBPs (PlewOBP 1–8, 10–12, and 15) sequences as “classic” OBPs (Supplementary Figure S1). The remaining three PlewOBP proteins (PlewOBP 9, 13, and 14) belong to the “plus-C” OBP family (Supplementary Figure S1), and fit to the Cys spacing pattern (C1-X_8-41-C2-X₃-C3-X_39-47-C4-X_17-29-C4a-X₉-C5-X₈-C₆-P-X_9-11-C6a).

A neighbour-joining tree consisting of 119 mature OBPs was constructed from six Hemipteran species to confirm the intraspecific divergence of their OBPs. The Hemipteran OBP protein family generates an expansive tree, with distinct clades for both “classic” and “plus-C” OBP sequences (Figure 3). In the phylogenetic tree, OBPs of the same subfamily exhibit local clustering, while OBPs of the same subfamily are distributed evenly throughout the entire tree, forming separate central clusters. These results provide evidence of significant duplication and specialization of OBPs within Heteroptera. However, this finding diverges from aphids, which have the most orthologous sequences in different species (Xue et al., 2016; Wang et al., 2019; Song et al., 2021). In this work, we found no intraspecific orthologous genes within the same species of stink bugs. Nevertheless, there is a slightly higher rate of orthologues between related species. 10 of 15 PlewOBPs have homologous sequences to the OBPs found in other Hemiptera insects with a high bootstrap value, indicating a high probability that these sequences come from a common ancestor and are preserved for shared functions in plant bug species. PlewOBPs also have paralogs, such as PlewOBP5 and PlewOBP6, which may have undergone horizontal duplication duplicated from the same ancestor through natural selection to acquire an additional function.

The identification of differentially expressed genes (DEGs) was carried out by the DESeq software package, with an absolute log2 (fold change) value of ≥1 and a P-value ≤0.05 were considered as the threshold for significant differences. Within this framework, the transcript abundance of 325 genes was significantly altered when females were exposed to volatiles emitted by healthy tobacco plants for 1 h. 112 genes demonstrated lower abundance, while 213 showed more abundance when compared to blank control (Figure 4A). Similarly, in males, a total of 93 genes showed decreased abundance out of the 331 genes that were differentially detected. Conversely, 238 genes revealed increased abundance in the HT group compared to the control group (Figure 4B). The S. litura-infested tobacco plants exhibited a higher propensity for genes to change in transcript abundance than healthy plants. Exposure of adult females to infested tobacco resulted in the upregulation of 254 genes and the downregulation of 184 ones (Figure 4C). Additionally, 223 genes were upregulated, and 200 were downregulated in adult males (Figure 4D).

We conducted GO analysis to better understand the genes responsible for the variation between sexes and various treatments. Upon assessing the transcripts in male and female bugs, we observed distinct differences in the GO enrichment patterns of differentially expressed genes induced by healthy or infested tobacco plants (Supplementary Figures S2–S5). There was minimal overlap of GO enrichment terms between males and females within the same treatment and across different treatments within the same sex. The Wayne diagram analysis corroborated these findings, revealing that only a few genes were simultaneously up- or downregulated between different treatments or sexes (Figure 5). Among which, the venom serine protease-like gene was the only annotated gene, potentially serving common roles in digestion and detoxification (He et al., 2024). It was unsurprising that genes can respond to environmental factors, leading to alterations in transcript abundance in vivo. The transcript abundance of a large variety of genes in Anopheles gambiae (Diptera: Culicidae) mosquitoes can be influenced by varied light conditions (Rund et al., 2013). Whereas transcriptomic data in Aedes aegypti (Diptera: Culicidae) revealed that prolonged exposure to 1-octen-3-ol modulated ORs and OBPs genes, as well as cytochrome P450 enzymes, insect cuticle proteins, and glucuronosyltransferases families (Mappin et al., 2023). Abiotic environmental factors and biological factors such as courtship, mating, sex, and age can influence antennal chemosensory-related genes (Tallon et al., 2019; Alonso et al., 2019; Andersson et al., 2014; Siju et al., 2010). However, the transcriptome data in the present study revealed none differentially expressed PlewOBPs. One potential explanation is that this change in transcript levels correlates with the concentration and duration of odor exposure. The duration and concentration of our odor exposure experiments were insufficient (Duan et al., 2023). Conversely, minimal fold differences in gene expression may evade detection using transcriptome sequencing, as our screening criteria were based on thresholds of log₂ (fold change) ≥ 1. To gain a more comprehensive understanding of the pattern of changes in the OBP gene family, we performed qRT-PCR experiments with all PlewOBPs.

To investigate the expression profile of PlewOBPs after exposure to different tobacco volatiles, we carried out qRT-PCR experiments. The results showed that exposure to tobacco volatiles significantly altered the transcript abundance of several PlewOBPs (Figure 6). The transcript abundance of three PlewOBPs (PlewOBP2, 4, and 12) was considerably changed in male adults exposed to S. litura-infested tobacco. PlewOBP2 was the only upregulated gene in this group compared to the blank control. Four PlewOBPs (PlewOBP2, 6, 11, and 13) displayed significantly higher transcript abundance levels than the healthy tobacco exposure treatment. PlewOBP4 had reduced transcript abundance in the IT group compared to the HT group. The transcript abundance of 5 PlewOBPs’ (PlewOBP2, 4, 7, 8, and 12) was significantly downregulated in female adults exposed to S. litura-infested tobacco compared to the blank control.

Compared to the HT group, PlewOBP2, 6, 7, and 8 exhibited significantly lower levels of transcript abundance, while PlewOBP3 showed substantially higher levels of transcript abundance in the IT group. These numerously expressed PlewOBPs might be involved in regulating the behavioral activity of P. lewisi adults towards tobacco plants. The DREAM technique is the prevalent method for screening OBPs by examining substantial changes in mRNA expression levels upon exposure to odors and various biotic or abiotic factors (Chen et al., 2021). An study has shown that cucurbit chlorotic yellows virus (CCYV) infested Bemisia tabaci (Hemiptera: Aleyrodidae) exhibited increased orientation towards the host cucumber plant (He et al., 2023). Transcriptome analysis revealed 429 DEGs (407 upregulated and 22 downregulated) between CCYV-carrying and CCYV-free whitefly adults. Odorant-binding protein 5 (OBP5) was upregulated in CCYV-carrying whiteflies compared to CCYV-free ones and was proved to be involved in odor recognition and host localization. This transcriptional response is likely directly associated with the olfactory capabilities of OBPs, which may also aid insects in host location. In our study, differentially expressed OBPs between the IT and HT groups may be involved in the recognition progress of herbivore-induced plant volatiles released by S. litura-infested tobacco. However, this method is always associated with false positives or false negatives (Koerte et al., 2018). Additional tests, such as RNAi, are supposed to validate the results of this experiment in the future.

Transcription sequencing consistently reveals insights into the molecular mechanisms that regulate several physiological states in insects. For instance, Rinker et al. (2013) showed that the transcript abundance of A. gambiae odorant receptors (AgORs) and their excitatory odorant response profiles corresponded with the shift from host-seeking to oviposition behaviors in blood-fed female mosquitoes (Rinker et al., 2013). In contrast, AgOBPs exhibited a more complex pattern of variation, reflecting the multifaceted roles of OBPs, devoid of any apparent regularity. Besides olfactory-related genes, many genes from various signaling pathways also showed significant changes in transcript abundance, such as cytochrome P450 (Mappin et al., 2023). The molecular mechanisms behind the changes in transcript abundance induced by these genes remain unclear. These substances may possess additional physiological relevance for insects, or OBP may exhibit a diminished interaction with other signaling pathways. Nonetheless, our study demonstrates that the alterations in gene transcript abundance resulting from odor exposure are preserved in insects.

In conclusion, both male and female P. lewisi demonstrated changes in their transcription profiles when exposed to healthy or S. litura-infested tobacco plant odor. The mRNA levels in P. lewisi showed a sex-dependent modification after exposure to odor. The mRNA expression profiles differed significantly between male and female adults. Transcription sequencing identified the presence of 15 PlewOBPs, and 8 showed significant changes in transcript abundance when exposed to S. litura-infested tobacco, compared to the blank control or healthy tobacco. These genes provide novel targets for functional characterization, which may, in turn, lead to the development of tools and strategies for insect behavior regulation.