Introduction

Urban cockroaches, such as the American cockroach (Periplaneta americana) and the German cockroach (Blattella germanica), have successfully adapted to living in human environments, spreading their distribution worldwide1. They inhabit sewers and other unsanitary areas and are considered pests due to their ability to spread disease and their potential to cause allergic reactions2,3. One of the most prominent characteristics of cockroaches is their use of pheromone communication to aggregate and attract mating partners. Therefore, researchers have been investigating the pheromone reception mechanisms in cockroaches to elucidate the neuronal basis of insect pheromone communication and to develop strategies for pest management4.

Periplaneta cockroaches use periplanone analogs as female sex pheromones to attract males4. The American cockroach (P. americana) used the sex pheromone composed of minor component periplanone-A (PA) and major component periplanone-B (PB)5,6. In the cockroach, PA and PB are mainly detected by two different sensory neurons, PA-responsive sensory neurons (PA-SNs) and PB-responsive sensory neurons (PB-SNs), in the adult male antennae-specific single walled B (sw-B) sensilla, respectively7,8,9,10. We previously identified two odorant receptor (OR) genes, PameOR1 and PameOR2, exclusively expressed in two adjacent olfactory sensory neurons (OSNs) within a single sw-B sensilla11. PameOR1 and PameOR2 genes were first identified as OR genes preferentially expressed in the antennae of adult male P. americana in the transcriptome analysis conducted by Chen et al.12. The phylogenetic analysis conducted of the OR gene in P. americana conducted indicated that PameOR1 and PameOR2 belong to a small independent clade of the other OR gene subfamilies of P. americana, and the clade exhibits linage-specific expansions in Blattodea insects12,13. In our previous study11, we conducted a series of RNAi-based functional analyses of the male-biased OR genes in P. americana. Using the single sensillum recording (SSR) combined with RNAi targeting OR genes, we revealed that PameOR1 acts as a PA receptor in PA-SNs, whereas PameOR2 is tuned to PA and PB and acts as a PA/PB receptor in PB-SNs.

Independent of Tateishi et al.11, Li et al.14 recently reported the identification of two periplanone receptor genes, OR53 and OR100, as male-biased ORx genes in P. americana. Functional analysis suggested that OR100 serves as a PA receptor, while OR53 acts as a PA/PB receptor, potentially corresponding to our PameOR1 and PameOR2, respectively. However, the expression patterns and ligand sensitivities of the periplanone receptor genes reported by Li et al.14 differ from those found in our previous study11.

In our previous study11, we re-constructed transcriptome sequence assembly of the P. americana antennae using previously published RNA-seq raw read data12, and obtained cDNA sequences of OR genes11. This analysis found that the old RNA-seq dataset needed more read length and depth to reconstruct the OR genes adequately. Therefore, in parallel with re-analyzing the old RNA-seq dataset, we conducted a short-read RNA-seq analysis to complement the old RNA-seq dataset. As a result, we identified an additional periplanone receptor-like gene, PameOR1-like, which was only found in our new transcriptome dataset. Then, we conducted a series of expression analyses of PameOR1-like in the cockroach antennae to update our knowledge on the periplanone responsive mechanism in the American cockroach.

Results

Identification of PameOR1-like gene in the antennae of P. americana

We first conducted de novo transcriptome sequence assembly using our new short-read RNA-seq dataset (BioProject ID: PRJDB18621) and conducted a translated BLAST search (tblastn) on the de novo transcriptome sequence assembly data with previously identified periplanone receptor family proteins (PameOR1 and PameOR2)11. As a result, we confirmed the presence of assembled transcripts corresponding to these genes. To our surprise, we found a set of assembled transcripts encoding a protein resembling the PameOR1 protein. Based on the high-level sequence homology to PameOR1, we considered this gene as a new member of the periplanone receptor family gene.

To confirm the presence of the PameOR1-like gene in the adult cockroach antennae, we conducted a series of RT-PCR-based molecular cloning experiments and obtained a 1282-bp cDNA fragment containing a 1203-bp full-length open reading frame. Both PameOR1 and PameOR1-like encode 400 amino acid proteins with seven transmembrane regions (Fig. 1A). The protein-coding sequence and the deduced amino acid sequence of PameOR1-like showed high-level similarity to those of previously identified PameOR1 (77.5% and 85.6% identity, respectively). Those of PameOR1-like and PameOR2 showed 53.5% and 41.1% identity, respectively (see Fig. 1B,C). The high-level sequence similarity between PameOR1 and PameOR1-like proteins supports the idea that PameOR1-like functions as a receptor for periplanone analogs (possibly a receptor for PA). The high-level sequence similarity of the nucleotide sequence of PameOR1 and PameOR1-like suggests that it is difficult to individually knockdown PameOR1 and PameOR1-like genes by systemic RNAi, which requires the injection of dsRNA of several hundred base pairs or more.

Fig. 1
Fig. 1
Full size image

Identification of Periplaneta americana OR1-like (PameOR1-like). (A) Amino acid sequence alignment of PameOR1-like with previously identified PameOR1 and PameOR211. The letters are shaded by alignment strength. The positions of seven transmembrane helixes are illustrated. GenBank IDs: PameOR1, BET24496.1; PameOR2, BET24497.1. (B) Dot plot representation of nucleotide sequence identity between the protein-coding region of PameOR1 and that of PameOR1-like. The red shaded area indicates the region corresponding to the PameOR1 dsRNA and the riboprobes targeting PameOR1 that were used for in situ hybridization in our previous study11. Note that this region shows high-level sequence identity (87.4%). (C) Dot plot representation of protein sequence identity between PameOR1 and PameOR1-like.

Chen et al.12 previously reported that the periplanone receptor family genes (PameOR1 and PameOR2) belong to a small independent clade of the other ORx genes in P. americana. In this study, we conducted the phylogenetic analysis of OR genes with our new transcriptome dataset. We included unique 69 OR genes identified in our new transcriptome dataset, three periplanone receptor family genes in P. americana (PameOR1, PameOR1-like, and PameOR2), previously identified odorant receptor co-receptor gene in P. americana (PameORco)15, and OR genes in two other Blattodea species (Zootermopsis nevadensis and B. germanica) in our analysis (Fig. 2). Consistent with the previous report, the previously identified periplanone receptor family genes, PameOR1 and PameOR2, along with the newly identified PameOR1-like form a small clade separate from other P. americana ORx genes.

Fig. 2
Fig. 2
Full size image

Phylogenetic analysis of Blattodea OR genes. The deduced amino acid sequences of the 73 OR genes expressed in the antennae of male and female adult P. americana, 78 OR genes in Z. nevadensis31, and 135 OR genes in B. germanica32 were subjected to phylogenetic analysis.ORco was set as the outgroup. The deduced amino acid sequences of 69 OR genes annotated in P. americana are provided in the “Supplementary data”. Bootstrap values are represented by circles at the nodes, with the diameter proportional to the value and color-coded using the HSB spectrum. The scale bar corresponds to 0.8 substitutions per site. The phylogenetic tree was visualized using the FigTree (version 1.4.4; https://github.com/rambaut/figtree/releases).

PameOR1 and PameOR1-like are expressed in the PA-SN within the sw-B sensilla

In our previous study, we revealed that the PameOR1 is expressed in a single OSN within the sw-B sensilla, cohabiting with a single PameOR2-expressing OSN11. However, due to the extremely high sequence homology of the protein-coding regions of PameOR1 and PameOR1-like, in situ hybridization performed in our previous study with a riboprobe designed within the protein-coding region of PameOR1 (PameOR1 ORF probe) likely labeled both PameOR1 and PameOR1-like without discriminating between two OR genes (see Fig. 1B)11. To address this, we designed gene-specific riboprobes for PameOR1 and PameOR1-like targeting their 3’ UTRs, where sequence homology is low enough to distinguish between the target genes (see Fig. 3A,B). With single-color fluorescent in situ hybridization (FISH) targeting each gene, we confirmed that the expression of both PameOR1 and PameOR1-like are detected in a single OSN out of the four OSNs within the sw-B sensilla (Fig. 3C, D and Supplementary Fig. 1).

Fig. 3
Fig. 3
Full size image

FISH to discriminatory stain PameOR1 and PameOR1-like. (A) Nucleotide sequence alignment of the 3’ region of the protein-coding sequence and the 3’ UTR region of PameOR1 and PameOR1-like, used as templates for riboprobe synthesis. The letters are shaded according to alignment strength. The protein-coding sequence is outlined with a red line. (B) Dot plot representation of nucleotide sequence identity between the riboprobe region ofPameOR1 and PameOR1-like. (C) Single-color FISH for the discriminatory detection of PameOR1 and PameOR1-like mRNAs in the male adult antennae. As a positive control, we used the riboprobe from our previou study11, which simultaneously detects both PameOR1 and PameOR1-like mRNAs. OSNs located beneath the cuticle were stained (arrowheads). Scale bars = 10 μm. (D) Magnified views of a single sw-B sensilla. In individual sw-B sensilla, a single OSN expressing either PameOR1 or PameOR1-like was detected, respectively (arrowheads). Scale bars = 5 μm.

Next, to examine whether the PameOR1 and PameOR1-like genes are exclusively expressed or co-expressed in a single PA-SN within individual sw-B sensilla, we conducted dual-color FISH targeting either PameOR1 or PameOR1-like with PameOR2 (Fig. 4). The expression of both PameOR1 and PameOR1-like is detected in a single OSN that is always paired with an OSN that expresses PameOR2 (Fig. 4A-C). Although dual-color FISH of PameOR1 and PamrOR1-like was not successful due to their low expression levels, from above results, we concluded that the PA-SNs in which PameOR1 was detected in our previous study co-express both PameOR1 and PameOR1-like (Fig. 4D).

Fig. 4
Fig. 4
Full size image

Dual-color FISH to detect (A) both PameOR1 and PameOR1-like mRNAs, (B) PameOR1 mRNA alone, and (C) PameOR1-like mRNA alone with PameOR2 mRNA in the male adult antennae. An OSN expressing PameOR1 and/or PameOR1-like (PA-SN; magenta) is always paired with an OSN expressing PameOR2 (PB-SN; green). The PA-SN and PB-SN pairs are indicated by arrowheads. Scale bars = 20 μm. (D) Schematic diagram of the sw-B sensilla in the adult male antennae. All four OSNs within the sw-B sensilla express the PameORco protein8. Sex pheromones are received by two OSNs: the PB-SN, which possesses the PameOR2/PameORco receptor complex, and the PA-SN, which co-expresses the PameOR1/PameORco and PameOR1-like/PameORco receptor complexes. Odor-specific ORx protein expressed in the other two OSNs (OSNX and OSNY) are unknown.

Comparison of the mRNA expression levels of PameOR1 and PameOR1-like in the adult antennae

Due to the high-level sequence similarity between PameOR1 and PameOR1-like, the quantification of their expression levels by RNA-seq is not accurate. To compare the expression levels of PameOR1 and PameOR1-like, we conducted an RT-qPCR assay for absolute quantification of the mRNA levels using the standard curve method. We first prepared a plasmid containing the full-length coding sequences of PameOR1 and that of PameOR1-like as a template to generate standard curves for both genes. Then, we quantified and compared the mRNA expression levels of PameOR1 and PameOR1-like expressed in the adult male and female antennae (Fig. 5). The expression level of PameOR1-like is 63-fold higher in the antennae of male adult cockroaches than that of female adults. In the antennae of both sexes, the expression levels of PameOR1 are significantly higher than that of PameOR1-like. The male adult antennae express the PameOR1 at a 2.4-fold higher level of PameOR1-like, and the female adult antennae express the PameOR1 at a 2.8-fold higher level of PameOR1-like.

Fig. 5
Fig. 5
Full size image

Comparison of the mRNA expression levels of PameOR1 and PameOR1-like in the adult cockroach antennae. (A) The expression levels of PameOR1 and PameOR1-like, normalized to PameEf1α, are plotted. Gene expression levels were measured in the antennae of 11-day-old adult males and females. (B) The expression levels of PameOR1 and PameOR1-like in female antennae. The same data as in (A) but with an adjusted vertical axis. Dotted lines in the violin plots represent the 25th to 75th percentile ranges and central values. Asterisks indicate statistical significance between the expression levels of PameOR1 and PameOR1-like (paired t-test; *, p<0.05; **, p<0.01; ***, p<0.001).

Discussion

Urban cockroaches, such as the American cockroach (P. americana) and the German cockroach (B. germanica), use pheromones to regulate aggregation and sexual behaviors4. Research on pheromone reception mechanisms in these insects is a significant topic in insect neuroethology and is also crucial for pest control7,8,9,11,15. Recently, receptor genes for two periplanone analogs (PA and PB), which are released by sexually mature female P. americana, have been identified5,6,11,14. Before our study, Chen et al.12 conducted a transcriptome analysis of the chemoreceptor genes expressed in the antennae and mouthparts and identified 96 odorant receptor genes in P. americana. Since RNAi-mediated knockdown of odorant receptor co-receptor gene in P. americana (PameORco) abolished periplanone reception capability8, we conducted RNAi-based functional analysis for OR genes with male-biased expression in the antennae of adult P. americana reported by Chen et al.12. Eventually, we identified two ORx genes, PameOR1 and PameOR2, as PA receptor and PA/PB receptor genes, respectively11. In this study, we identified another periplanone receptor-like gene, PameOR1-like, selectively expressed in the sex pheromone responsive antennal sensilla of the adult male antennae.

In parallel with Chen et al.12 and Tateishi et al.11, another research group conducted genomic DNA sequencing of P. americana to annotate chemosensory genes16, and functional analyses of two OR genes (OR53 and OR100) in sex pheromone reception14. Unfortunately, the authors did not publish the nucleotide/deduced amino acid sequences of OR53 and OR100. Therefore, it was not easy to map out a correspondence of OR genes identified in our studies (this study and Tateishi et al.11) and those in Li et al.14. Based on PCR primer sequences provided by Li et al.14, we consider that OR53 and OR100 correspond to PameOR1-like and PameOR2, respectively. With the publication of sequence data by Li et al.14, our understanding of the olfactory system of P. americana will be further enhanced, including the unification of gene nomenclature and information on genetic polymorphisms between populations.

In our previous study11, we reported that PameOR1 and PameOR2 are each exclusively expressed in a single OSN out of the four OSNs within the sw-B sensilla. However, at that time, we were unable to distinguish PameOR1 from PameOR1-like because the riboprobe was designed to target a sequence nearly identical in both genes (see Fig. 1B). In the present study, we designed a set of riboprobes to differentiate between PameOR1 and PameOR1-like and successfully detected their expression in the adult male antennae. Consistent with our previous findings, PameOR1 and PameOR1-like were each detected in only one of the four OSNs within the sw-B sensilla (Fig. 3). Furthermore, co-staining with PameOR2 revealed that OSNs expressing either PameOR1, PameOR1-like, or both are always present in pairs with OSNs expressing PameOR2 (Fig. 4A), and that OSNs expressing PameOR2 are always present in pairs with OSNs expressing PameOR1 (Fig. 4B) and in pairs with those expressing PameOK1-like (Fig. 4C). From these results, we concluded that the PA-SNs co-express both PameOR1 and PameOR1-like.

Due to the high nucleotide sequence similarity in the protein-coding regions of PameOR1 and PameOR1-like, our previous study11 likely failed to discriminate between these two genes in dsRNA-mediated systemic RNAi experiments, just as it did in the in situ hybridization experiments (see Fig. 1B). In this study, we confirmed the expression of PameOR1 and PameOR1-like in a single PA-SN in the sw-B sensilla. Therefore, the reduction of PA-reception capability by RNAi of PameOR111 can be interpreted as a combined knockdown effect on two periplanone receptor family genes expressed in the PA-SNs. Besides, RT-qPCR analysis revealed that the mRNA expression level of PameOR1-like is only less than half of that of PameOR1 in the antennae of adult cockroaches. From these data, we conclude that PameOR1 and PameOR1-like function as major and minor PA receptors, respectively, in the antennae of adult male P. americana.

In insects and mammals, it is generally accepted idea that a single OSN expresses only one specific OR gene out of hundreds of receptor genes in the genome17,18. Recently, researchers have found some exceptions to this ‘singular odorant receptor expression’ theory in several holometabolous insect species, including fruit flies, mosquitoes, and ants19,20,21,22. In this study, we revealed that PA-SNs co-express two ORx genes, PameOR1 and PameOR1-like, providing another exception to the ‘singular odorant receptor expression’ theory and representing the first such case identified in hemimetabolous insects. The high-level nucleotide and amino acid sequence homology between PameOR1 and PameOR1-like supports the idea that these genes were produced by a recent gene duplication of the ancestral OR1-like gene. This duplication may have resulted in daughter genes with functionally similar regulatory regions, which could explain why PameOR1 and PameOR1-like share their expression in the PA-SNs. In the present P. americana, these two OR1-like genes maintain a high degree of homology, and therefore, we consider that they play a redundant function in sex pheromone reception (i.e., PA reception). Whether these two genes underwent functional differentiation and became expressed in distinct OSNs or whether gene loss occurred in the related species remains an intriguing topic in the study of the molecular evolution of olfactory receptors.

Conclusion

The male antennae of P. americana express an additional periplanone receptor family gene, PameOR1-like, in addition to previously identified PameOR1 and PameOR2. PameOR1-like is co-expressed with PameOR1 in the PA-SNs within the male-specific sw-B sensilla. PameOR1-like and PameOR1 showed strikingly high-level sequence similarity at nucleotide and protein levels, suggesting these genes were produced by recent gene duplication and play a redundant function in sex pheromone reception in P. americana. Further investigations on the interspecies comparison and molecular evolution studies on the periplanone receptor family genes will shed light on the molecular basis of species-specific pheromone reception mechanisms and the universal principles and diversity of sex pheromone systems in cockroaches.

Materials and methods

Insects

The cockroaches P. americana were obtained from laboratory colonies maintained at 27 ± 1 °C under a 12:12 light-dark cycle at Fukuoka University. Cockroaches were collected from colonies immediately after molting and individually reared.

Molecular cloning of PameOR1-like

RNA-seq and de novo transcriptome sequence assembly: RNA was extracted from the antennae of adult male and female P. americana using the TRIzol reagent (Life Technologies), then purified using the RNeasy Mini Kit (Qiagen, Tokyo, Japan). Genomic DNA contamination was removed by on-column DNase I treatment (RNase-Free DNase Set; Qiagen). Both male and female RNA samples were individually submitted to the Rhelixa Next Generation Sequencing service (Rhelixa, Tokyo, Japan) for high throughput sequencing. Libraries were generated using the NEBNext Poly(A) mRNA Magnetic Isolation Module and the NEBNext Ultra II Directional RNA Library Prep Kit (New England Biolabs, MA, USA). Sequencing was performed in the NovaSeq 6000 System (Illumina). 123FASTQ (ver. 1.3)23 was used to remove the Illumina adapter sequences. SortMeRNA (ver. 4.3.4)24 was used to filter ribosomal RNA fragments. Then, de novosequence assembly was conducted using the Trinity assembler (v2.14.0)25. Raw reads from female and male samples were combined for analysis. The RNA-seq raw sequencing data were registered in the BioProject database (BioProject ID: PRJDB18621).

RT-PCR-based cloning: Since the nucleotide sequences of the 5’ and 3’ untranslated regions (UTRs) of the PameOR1-like assembled transcripts were ambiguous, we conducted 5’ and 3’ rapid amplification of cDNA ends (RACE) to obtain the 5’ and 3’ regions of the gene. Then, the full-length open reading frame (ORF) of the genes was amplified using primers designed at the 5’ and 3’ UTRs of the gene. The 5’ and 3’ RACE were performed using the FirstChoice RLM-RACE kit (Ambion, Austin, TX, USA), according to Watanabe et al.26. All PCRs were performed using the Q5 High-Fidelity DNA polymerase (New England Biolabs). Amplified cDNA fragments were cloned into the pGEM-T Easy vector (Promega, WI, USA). The nucleotide sequence of the obtained cDNA was registered in GenBank (GenBank ID: LC833870). Primers used for 5’ and 3’ RACEs and full-length cDNA amplification are listed in Table 1.

Table 1 List of primers used in this study.
Full size table

Sequence comparison: The deduced amino acid sequences of PameOR1, PameOR1-like, and PameOR2were aligned using the MAFFT algorithm27 and refined by manual inspection on the Geneious Prime program (ver. 2024.0.4; https://www.geneious.com). The alignment was visualized using the pyBoxshade program (v1.2; available from https://github.com/mdbaron42/pyBoxshade/). The nucleotide and deduced amino acid sequence dot plots were visualized using the Geneious Prime program.

Phylogenetic analysis

Genome-guided transcriptome assembly to reconstruct the transcriptome of P. americana OR genes: In addition to the de novo sequence assembly described above, we conducted genome-guided transcriptome assembly. After the removal of ribosomal RNA fragments, raw sequence read data were mapped to two P. americanagenome assemblies (ASM293952v116 and CNA001928128) using the HISAT2 aligner (v 2.2.1)29. Genome-guided transcriptome assembly was conducted using the Trinity assembler (v2.14.0)25. The EvidentialGene tr2aacds.pl pipeline (http://arthropods.eugenes.org/EvidentialGene/) was used to reduce redundancy in the transcriptome assemblies. Functional annotation of the assembled transcriptomes was performed using EnTAP30, and transcripts annotated as 7tm Odorant receptors (7tm_6; PF02949) were extracted. Then, we compared datasets derived from the two genome-guided assemblies and the de novo assembly to exclude overlapping sequences.

Phylogenetic analysis: The deduced amino acid sequences of 73 OR genes of P. americana, 78 OR genes of Z. nevadensis (predicted ORgenes retrieved from the GenBank)31, and 135 OR genes of B. germanica32 were aligned using the MAFFT algorithm on the Geneious Prime program. Poorly aligned regions were removed using the trimAl (version 2.0)33 with the gappyout mode. Then, the bootstrapped maximum likelihood tree (1,000 bootstrap replicates) was generated using the raxmlGUI program (version 2.0)34 and visualized using the FigTree (version 1.4.4; https://github.com/rambaut/figtree/releases). The deduced amino acid sequences of the P. americana OR genes are provided as supplementary data.

Fluorescent in situ hybridization (FISH)

Probe design: Due to the high-level similarity in the nucleotide sequence of the protein-coding regions of PameOR1 and PameOR1-like, we designed a set of gene-specific riboprobes for in situ hybridization in the 3’ UTRs of each gene. For each gene, an aproxmatery   450-bp cDNA fragment corresponding to the coding sequence of the C-terminus of the protein and the 3’ UTR was amplified and inserted into the pGEM-T Easy vector. The nucleotide sequences of the 3’ UTR of the genes were registered in GenBank (GenBank IDs: LC833871 and LC833872, respectively). We also used a riboprobe targeting both PameOR1 and PameOR1-like, as well as that targeting PameOR2from our previous study11.

FISH: Riboprobes were prepared by in vitro transcription using PCR products as a template (Table 1). Digoxigenin- or fluorescein-labeled riboprobes were synthesized using RNA labeling kits (Roche, Basel, Switzerland). Gene expression was visualized using the “enhanced peroxidase (POD)-tyramide signal amplification (TSA) reaction protocol,” as previously described35,36. Briefly, 10 µm fresh frozen sections were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer and hybridized with the riboprobes. Fluorescein- and Cy3-fluorescent signals were developed through a series of POD and TSA reactions with an optimized buffer solution, followed by 4’,6-diamidino-2-phenylindole (DAPI) staining. Fluorescence images were captured using the confocal laser scanning microscopes (DM2500 [Leica Microsystems, Tokyo, Japan] or the LSM980 [Zeiss, Tokyo, Japan]).

RT-qPCR analysis

The reverse transcription for RT-qPCR analysis was conducted following the methods of our previous report11. The quantification of target genes was carried out using the KAPA SYBR Fast qPCR Kit (Kapa Biosystems, MA, USA) and the PCRmax Eco 48 Real-Time qPCR System (PCRmax, Staffordshire, UK). Gene expression levels were measured by the standard curve method using the EcoStudy Software (ver. 5.0; PCRmax). Ten-fold serial dilutions of standard plasmids containing cDNA fragment(s) of target gene(s) were used to plot standard curves. The standard plasmid for PameOR1 and PameOR1-like contained the full-length coding sequence of each gene, which were tandemly inserted into the pGEM-T Easy vector. The standard plasmid for PameEf1α contained the partial coding sequence of the gene (GenBank ID: LC657820.1), which was inserted into the pGEM-T Easy vector. The PCR reaction was performed at 50ºC for 2 min and 95°C for 5 min, followed by 40 cycles of 95°C for 10 sec and 60°C for 30 sec each. The nucleotide sequences of the primers used for RT-qPCR are listed in Table 1. Statistical analysis was conducted using the GraphPad Prism version 10.0 for Mac (GraphPad Software, MA, USA).