Background

For a long period, insects have gradually adapted to the complex and ever-changing physiological environment with their sensitive olfactory system recognizing a large number of odor chemicals, which plays a crucial role in host selection, feeding, mating, and reproduction [1,2,3]. Insect’s antenna, covered by multi-type olfactory sensilla like the basiconic, coeloconic, and trichoid, is the central organ in sensing and recognizing external odors [4]. The sensilla are filled with potassium- and protein-rich fluid called sensillum lymph, which bathes the dendrites [5, 6]. Many chemosensation-related proteins secreted in sensillum lymph are involved in the complex olfactory-perception process, such as Odorant-binding proteins (OBPs), odorant receptors (ORs), chemosensory proteins (CSPs), ionotropic receptors (IRs), sensory neuron membrane proteins (SNMPs), and odorant-degrading enzymes (ODEs) [7, 8].

Among all those olfaction-related proteins, OBPs function as the initial step in odorant recognition and transduction [9, 10]. OBPs were a group of small, soluble, and acidic proteins with a highly-conserved structure [11, 12]. Generally, OBPs are classified into five diverse subtypes based on the number and model of conserved cysteines in their amino acid sequence [13], which includes Classical OBPs (those with 6 conserved cysteines), Minus-C OBPs (those with only 4 conserved cysteines), Plus-C OBPs (those with 8 conserved cysteines), dimer OBPs (those with 12 conserved cysteines) and Atypical OBPs (those with 9~10 conserved cysteines) [14, 15]. Upon encountering external chemical signals, such as pheromones, plant volatiles or odors from other species, odor molecules would enter the sensillum lymph through the massive pores on the sensilla, and OBPs in the lymph immediately recognize, bind and shift the newly-formed odor-OBP complexes to the ORs in sensory dendrites, which transform the chemical signals to electrophysiological signals and eventually trigger the corresponding behavior of insects [16,17,18].

OBPs have been intensively studied since the first report in a moth, Antheraea polyphemus [19]. Various OBPs and multiple functions accordingly have been identified. A class of GOBPs binding and transporting common odor molecules in the antennae of female Antheraea pernyi were identified [20] (Breer et al., 1990). Biochemical binding kinetics studies found the dual role of transporting and inactivating odorous substances [21, 22]. A study of Drosophila melanogaster mutants showed that OBPs are involved in the transport of odor molecules to ORs [23]. Besides, ApisOBP3 in Acyrthosiphon pisum [24], GmolGOBP2 in Grapholita molesta [25] and OBP6 in Meteorus pulchricornis [26] all demonstrated that OBPs could specifically recognize and screen specific chemical signals. Recently, The rapid development of techniques like electrophysiology, RNA interference, and gene knockout has directly revealed the necessity of OBPs for proper functioning in the olfactory system [27,41]. In this study, antennal-specific expressed OBPs, such as MsauOBP2, 3, 4, 8, 11, and 17, were highly possible to possess the olfactory function, which was similar to the fig wasp Wiebesia pumilae,, where this creature located its host Ficus pumila mainly through WpumOBP2 binding the decanal emitted by F. pumila [55]. Furthermore, the O.lotOBP6 of Odontothrips loti could strongly bind to p-Menth-8-en-2-one emitted by its host Medicago sativa and was the most crucial OBP in host-seeking [15]. Consequently, it’s reasonable to hypothesize that M. saussurei locate M. sativa through these highly expressed OBPs binding the single or multiple volatiles emitted by M. sativa to complete their feeding and pollination.

The interaction of MsauOBPs and two alfalfa flower volatiles 3-octanone and (Z)-3-hexenyl acetate was simulated by molecular docking method. Results showed most MsauOBPs could successfully bind with two ligands. It has been confirmed that the lower the binding energy, the better the binding effect [32]. In this study, MsauOBP4 showed the minimum value of binding energy when docking with 3-Octanone, while MsauOBP13 presented the lowest binding energy when docking with (Z)-3-hexenyl acetate. This implied MsauOBP4 and MsauOBP13 may play a crucial role in recognizing these two volatiles and may also contribute to the host location during the pollination process. Although more MsauOBPs tended to bind with 3-octanone, the mean binding energy of (Z)-3-hexenyl acetate was generally much lower, indicating that the combination between MsauOBPs and (Z)-3-hexenyl acetate was much more stable. Results also found the amino acid lysine appeared most frequently in docking simulations, which was also a momentous amino acid in other soluble olfactory proteins such as FoccOBP6 of Frankliniella occidentalis [8], OBP3 of Nilaparvata lugens [56] and MsepCSP14 of Mythimna separata [57]. It was found hydrophilic amino acids are more likely to form hydrogen bonds with ligands [58]. For instance, asparagine and serine in Hymenoptera [59], arginine, threonine, and aspartic acid in Lepidoptera [60, 61], glutamine in Hemiptera [62]. This was consistent with our result, in which lysine was also one of the hydrophilic amino acids.

Conclusions

In this study, we identified the OBPs, and conducted the phylogenetic and expression analysis. The interaction between two alfalfa flower volatiles and MsauOBPs was also simulated. Most OBPs were homologous while a certain degree of differences also existed. Six OBPs (MsauOBP2, 3, 4, 8, 11, and 17) mostly enriched in antennae were possibly involved in the olfactory functions. MsauOBP4 might be the key protein in recognizing 3-Octanone, while MsauOBP13 might be the key protein in binding (Z)-3-hexenyl acetate. These two proteins might contribute to the alfalfa-locating during the pollination process. The relevant results may help determine the highly specific and effective attractants for M. saussurei in alfalfa pollination and reveal the molecular mechanism of odor-evoked pollinating behavior between these two species. Further studies of these highly expressed OBPs using multi-methods are quite necessary, such as fluorescence binding assay, RNAi technique, and corresponding behavioral experiments, etc. Because these methods have been frequently used for the functional prediction and verification of insect OBPs. The relevant results may help determine the highly specific and effective attractants for M. saussurei in alfalfa pollination and reveal the molecular mechanism of odor-evoked pollinating behavior between these two species.

Methods

Antenna sample collection

The M. saussurei adults were captured in a blooming alfalfa field in the Yumen area (40◦45´N, 97◦36´E), Gansu province, China, in July 2022. To attract M. saussurei, the artificial foam nest (polystyrene bee board) was placed near the edges of the alfalfa field with the openings of the artificial nests facing the alfalfa field in a southeast direction [63, 64]. The size of artificial nests was maintained as instructed by Pitts-Singer and Bosch [65]. After M. saussurei was nested in these artificial nests, the emergence status and sex information of the adults were recorded every day. We carefully dissected the antennae from female M. saussurei in the laboratory and placed them in 1.5mL centrifugal tubes containing the RNA later buffer solution (Invitrogen, Carlsbad, CA, USA) [18]. The tubes were preserved at -80℃ until RNA extraction.

RNA extraction and transcriptome sequencing

Fifty pairs of antennae from M. saussurei adult females were used for total RNA extraction using TRIzol Reagent (Invitrogen, Waltham, MA, USA) following the manufacturer’s standard protocols (50 pairs antennae formed one sample, three samples (A1, A2, and A3) were set in total). The concentration and quality of RNA were verified using Fragment Analyzer 5200 (Agilent Technologies, Palo Alto, Canada). The cDNA library construction and transcriptome sequencing were performed on the DNBSEQ-500 platform at Wuhan BGI Technology (Wuhan, China) and a detailed flowchart was displayed in Fig. S1.

De novo assembly and functional annotation

To ensure the data reliability, we obtained clean reads from raw reads by filtering and deleting those reads of low quality, containing adapters and over 5% unknown bases. The clean reads were then assembled with Trinity v2.0.6 (https://github.com/trinityrnaseq/trinityrnaseq/wiki) using default parameters [66]. Then the unigenes from the three samples were pooled together to form the “all-unigene” by clustering reads and removing redundancy with the TGI Clustering Tool (TGICL) [67]. The quality of the assembled transcripts (unigenes) was thereafter evaluated using the BUSCO (Benchmarking Universal Single-Copy Orthologs) (https://busco.ezlab.org/), and the integrity of the transcriptome assembly was illustrated by comparison with conserved genes.

The coding sequence (CDS) in unigenes was identified using TransDecoder software by first extracting the longest open reading frame, and then Blast comparison against the Pfam protein homologous sequences in the SwissProt database and Hmmscan search to predict the coding regions. The unigenes were annotated against seven publicly accessed databases, the Kyoto Encyclopedia of Genes and Genomes (KEGG), the Gene Ontology (GO), the Non-redundant Protein Sequence Database (NR), Nucleotide Sequence Database (NT), the Protein Families Database (Pfam), Swiss-prot protein sequence database (Swiss-prot) and clusters of orthologous groups for eukaryotic complete genomes (KOG) with a threshold E-value < 1e-5. The expression level of each unigene was calculated by RSEM software (RNA-Seq by Expectation Maximization) with default parameters and presented as FPKM (fragments per kilobase of transcript per million mapped fragments) values.

Identification of odorant-binding protein genes and phylogenetic analysis

Candidate unigenes encoding putative odorant-binding proteins (OBPs) were selected from the assembly results. They were manually checked by performing a BLASTx search against the NR database with a threshold E-value < 1e-5 [68]. The open reading frame (ORF) of candidate OBP genes was predicted by NCBI ORF Finder (https://www.ncbi.nlm.nih.gov/orffinder). The N-terminal signal peptides were predicted by Signal P4.0 (http://www.cbs.dtu.dk/services/SignalP/).

We applied multiple amino acid sequence alignment with MUSCLE and constructed phylogenetic trees of putative OBP genes using the neighbor-joining (NJ) method with default parameters in MEGA v11.0 software. The reliability of the tree structure and node support was assessed using a bootstrap method with 1000 replicates and the phylogenetic tree was visualized in the Interactive Tree of Life (iTOL) (https://itol.embl.de/). Sequences of OBP genes from other bees were searched and selected from NCBI and used in the phylogenetic tree construction (Table S1). We finally aligned putative OBPs using GenDoc software and determined the type of putative OBPs.

Expression analysis by quantitative real-time PCR

After we identified the OBPs from the antennal transcriptome, we verified their expression levels in different tissues of M. saussurei using the quantitative real-time PCR method (RT-qPCR). Antenna, heads, legs, wings, and abdomen from 20 individuals were respectively collected and pooled together as one sample. Total RNA was extracted with TRIzol reagent (Invitrogen) and the cDNA was synthesized using the PrimeScript RT Reagent Kit with gDNA Eraser (TaKaRa, Shiga, Japan). The total volume of the PCR reaction system was 25μl, which contains 12.5 µl of SYBR Premix Ex TaqTM, 0.5 µl of forward primer, 0.5 µl of reverse primer, 2 μl of sample cDNA and 8.5 μl of double-distilled H2O. This PCR system was performed under the conditions of 95℃ for 30 s; 40 cycles of 95℃ for 5 s and 60℃ for 30 s; 65℃ to 95℃ in increments of 0.5℃ for 5 s. Negative controls with ddH2O were included. Gene-specific primers (Table S2) were designed using the Primer 3.0 plus server in NCBI. Nuclear β-actin was used as the internal reference gene and abdomen samples were used as the control group. Three biological replicates and three technical replicates were applied for each experiment.

The relative expression level of OBP genes was normalized using the comparative 2−∆∆Ct method [69]. One-way ANOVA analysis was applied to compare the expression levels between tissues, followed by Tukey’s post hoc comparison test for the significant differences. The data analysis and plot-making were both conducted using GraphPad Prism 9.0 software.

Homologous modeling and molecular docking

The online platform SWISS-MODEL (https://swissmodel.expasy.org) was used to predict the three-dimensional structure of all MsauOBPs. Models with similarity >30% were selected as reference templates. The PROCHECK program [70] was used to assess the generated MsauOBP models. 3-Octanone and (Z)-3-hexenyl acetate are two main components of alfalfa flower volatiles with relatively high content [71,72,73,74]. Ligand molecules were obtained from the PubChem database (https://pubchem.ncbi.nlm.nih.gov). The Autodock 4.2.6 and AutoDock Tools 1.5.7 with default parameters were used to conduct the molecular docking between MsauOBPs and two ligands. The docking results were visualized by PYMOL software.