Fire ant ovary gene expression analyses revealed immune and insulin pathways underlie the reproductive transition from virgin to mated queen

Abstract

Background

Solenopsis invicta queens experience significant behavioral and physiological changes after mating, which are essential for their reproductive success. We investigated differences in ovary gene expression in virgin alate queens, newly mated queens, and mated queens to identify candidate genes associated with their physiological transition to mature egg-laying queens. Virgin queens and mated queens were obtained from field colonies and newly mated queens were collected from the ground immediately after their mating flight. Whole ovaries of virgin alate queens, and germaria and vitellaria from the ovaries of newly mated and mature mated queens were dissected. Pools of each of these five organs/tissues were used for RNAseq and RT-qPCR analyses.

Results

Principal component analyses revealed a distinct transcriptomic profile among alate virgin ovaries, germaria of newly mated, and germaria of mated queens, highlighting the effect of mating driving significant differences in global gene expression. Mating did not have such a differentiating effect among libraries of newly mated and mated queen vitellaria. Differentially expressed genes (DEGs) were identified between whole ovary transcriptome of virgin alate queens and germaria of newly mated and mated queens, as well as vitellaria of newly mated and mated queens. There were 22 gene ontology terms enriched among the DEGs in the germaria analysis, of note were those enriched in development and phosphorylation. In the vitellarium, terms related to nucleobase-containing molecule processes and fatty acid metabolism were enriched. Sixty-one DEGs were shared between germaria and vitellaria libraries, mainly linked to immunity, lipid metabolism, development, and transcriptional regulation. Phenoloxidase was highly expressed in mated queens in both ovarian regions, suggesting a role in immunity and choriogenesis. Vg3, one S. invicta vitellogenin gene, was upregulated in the vitellaria of mated queens, reinforcing its role in vitellogenesis. Transcripts of the prostaglandin E₂ receptor showed ovary region-specific regulation, suggesting a significant role in immunity, oocyte development and potentially in the release of egg-laying behavior. Insulin-related genes were up-regulated in mated queens, reflecting the metabolic demands for egg production.

Conclusion

This study advances our understanding of immunity and mating and other key signaling pathways in fire ant reproduction.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-025-11900-7.

Background

The red imported fire ant, Solenopsis invicta Buren was accidentally introduced to the United States a hundred years ago [1, 2]. These ants cause significant economic losses in agriculture and are a concern for public health [2–4]. Like in other social insects, fire ant queen development and reproduction are regulated by individual- and colony-level factors. Its success as an invasive species is greatly due to the queens’ high reproductive output, the adaptability of the polygyne colonies to diverse environments, the high aggression of workers and the longevity of queens.

Young queens develop within the colony under the influence of mated queen pheromones that suppress corpora allata activity and reduce juvenile hormone (JH) production, preventing ovarian development in other females [5–9]. After their mating flight, these young queens transition from alates to dealates. The dealation is accompanied by major physiological changes, including ovary development, queen weight gain, and a degree of physogastry due to their enlarged ovaries [6, 10]. As part of their transition to egg-laying these mated queens start producing a queen pheromone, a dealation-inhibitory primer that prevents dealation and the potential reproductive activity of female alates within the colony [5, 11]. This pheromone can be detected as early as 3 days after the drop of the mated queen wings [7]. Underweight mated queens tend to produce less of the dealation-inhibitory pheromone [12].

If mature alate virgin queens are removed from exposure to the mated queen pheromone, they shed their wings, and their ovaries develop. These changes are accompanied by global changes in tissue gene expression [5, 6, 10, 13, 14]. For instance, the Striated Muscle Activator of Rho Signaling (STARS) gene, which is associated with wing muscle atrophy, is upregulated in virgin queens after their withdrawal from the presence of the mature queen, coincidental with their dealation [13, 15].

The two main structures of the ovaries are the germarium and the vitellarium. The germarium, composed by somatic and germ cells give origin to pre-follicular cells and cystoblasts, respectively, is where the oocyte development begins by division and differentiation of germ cells [16, 17]. In hymenopteran queens, the germarium is large and contains several germ cell groups in successive stages of development [18]. The oocyte increases in volume and changes its pigmentation during vitellogenesis. The oocyte development is accompanied by changes in the nurse cells which also increase their volume [18]. Vitellarium is where oocytes mature and yolk deposition occurs [18–20]. This region consists of two chambers, the oocytic and nurse chambers, containing oocytes and nurse cells, respectively. The follicular epithelium lines these chambers and executes different functions depending on the phase of oogenesis [18, 21].

Insect vitellogenesis is regulated by two essential hormones: juvenile hormone (JH) and 20-hydroxyecdysone (20E). JH stimulates vitellogenesis in most insects, while 20E plays a key role in certain insect orders [16, 22]. JH has been demonstrated to regulate reproductive processes and ovarian development in S. invicta queens [23–25]. This hormone stimulates the synthesis of vitellogenin (Vg) in the fat body, which is then deposited in the developing oocyte by the vitellogenin receptor-mediated endocytosis [8, 16, 22, 26]. The vitellogenin receptor has been immunolocalized in queen ovaries [27]. A recent publication reported that feeding fire ant queens with JH reduced the expression of the miR-1175-3p microRNA while the expression of vitellogenin 2 and 3 increased [23], confirming the gonadotropic function of JH in fire ants [9, 28].

Reproduction requires significant nutrient and energy resources, being a tightly regulated process. As in other insects, the amino acid/target of rapamycin (AA/TOR) and insulin are key pathways regulating reproduction and the lifespan of social insects [29]. They control the biosynthesis of JH and ecdysone, vitellogenesis, and the triggering of nutritional responses in specific tissues [29–31]. Insulin and insulin-like growth factors play a role in lipid transport and egg formation in insects by regulating metabolic processes [8, 25, 30, 32, 33].

Oogenesis is a complex phenomenon controlled by several factors such as nutrition, mating, season, and hormonal and pheromonal influences that lead to anatomical changes and behavioral specializations for egg production [6, 8, 10, 29, 34, 35]. In social insects, this process also is regulated by colony factors such as the presence of one or multiple queens. Signaling from these various factors needs to be integrated by the reproducing individual to effectively fulfill its reproductive role. For example, the short neuropeptide F receptor was immunolocalized in fire ant queen ovaries [36], and it is speculated that plays an integrative role sensing the nutritional status of queens.

In a previous study, our group analyzed the expression of genes in the brains of virgin and mature mated fire ant queens, identifying changes in the expression of genes related to juvenile hormone, vitellogenin synthesis, and immune response [14]. Regulation of genes involved in similar functions was also identified when comparing gene expression from whole-body virgin alates and mature mated queens [15, 37]. These findings underscore the interconnection between endocrine, structural, and behavioral changes, and immune responses in fire ant queens after mating. So, we also expected a difference in gene expression in the germarium and vitellarium of mated queens vs. newly mated queens, as these latter females are transitioning from being inhibited by the primer pheromone of mated queens to being disinhibited and mated and becoming laying queens.

In this study, we investigated expected changes in gene expression in Solenopsis invicta polygyne queen’s ovaries. There were two objectives (i) to utilize RNAseq to evaluate changes in gene expression between ovary libraries of virgin, newly mated queens (NMQ) and mated queens (MQ); libraries for the latter two were from germaria and vitellaria, and (ii) to verify the relative expression profiles of differentially expressed genes (DEGs) that are candidates to key regulatory pathways in the transition to fully reproductive queens. The verification of differential expression in ovaries was performed by RT-qPCR. We hypothesized that genes upregulated in MQ ovaries in comparison to NMQ would reflect a heightened synthetic metabolism in response to requirements for oocyte development, nutritional needs, and reproductive output; while genes related to immunity would be more expressed in NMQ because their gene products may be required as newly mated queens return to the soil where pathogens may be present and they begin to lay eggs.

Materials and methods

Insects

This research complements a previous transcriptomic analysis from polygyne fire ant queen brains conducted by Calkins et al. [14]. In the current study we report the analyses of the transcriptomes obtained from ovaries of the same queens. Polygyne fire ant colonies of S. invicta were gathered in the field between June and August 2016 in three locations, as follows: at the Texas A&M University Campus in College Station, Texas; the TAMU pecan orchard (30°31′16.07″N; 96°25′30.40″W, 14.5 km from the TAMU campus) in Somerville, TX, and at the “5 Eagle Ranch” (30°37′49.93″N; 96°40′19.48″W, in Burleson County, TX, 32 km from TAMU). The ants were then brought to the laboratory, and colonies were dripped and maintained following the method described by Chen et al. [24]. For dissections, mated and virgin alate queens were obtained from these fire ant colonies and their age was unknown. During the summer of 2016, also newly mated queens (n = 118) were collected from 3.20 pm until 4 pm after mating flights. Just de-alated mated queens returning from the mating flights were collected from the ground of a parking lot at TAMU. These queens were brought to the laboratory and were dissected the next day.

Ovaries dissections and RNA extraction for transcriptomic analyses

Ovary dissections from alate virgin queens, NMQ and MQ were carried out using the procedure previously described [14].

The tissues were dissected on ice-cold PBS buffer prepared with DEPC-treated water, and placed on dry ice in 100 µl of TRIzol Reagent (Thermo Fisher Scientific, Carlsbad, CA). We refer to the number of queens dissected for their ovaries, to imply that for each queen the pair of ovaries were dissected as a unit. During dissections, we created five pools of ovaries from alate virgin queens, as follows: one pool had ovaries of 33 virgin alate queens, and the other four pools had the ovaries of 12 virgin queens each (4 × 12 ovaries). We did not attempt to separate germaria from vitellaria of virgin alate queen ovaries because these ovaries were small, and by eye gazing no clear structural separation could be visualized. Images of virgin alate queen ovaries compared to mated queen ovaries are in Fig. 4 in [25].

Fig. 4.

Gene Ontology (GO) enrichment analyses of DEGs between the newly mated and mated queen germarium libraries. The g: Profiler analysis identified six driver terms numbered in the figure. Suppl. Table S4.2 shows the total 22 enriched GO terms identified. GO aspects: MF—molecular function, BP—biological process, and CC—cellular component

For newly mated queens, the verification of mating was done during dissection by observing a full white spermatheca. Eighteen queens in which mating could not be confirmed as per spermatheca observation (not white or full) were discarded. About 100 queens were confirmed as being mated, and their ovaries were divided into five pools of ~ 20 each. These ovaries were further separated into germaria and vitellaria. The pools enriched in germaria included the top 1/3 of the ovary, while the rest of the ovary was considered vitellarium, with observable enlarged oocytes. Therefore, samples for RNA extraction consisted in five pools of germaria and five pools of vitellaria (each consisting of ~ 20 respective tissues). For mated queen ovaries, one pool had ovaries of 20 queens, three pools had ovaries of 25 queens each, and one pool had the ovaries of 24 queens. Each ovary was separated into germarium and vitellarium similar to the newly mated queens. The pooled germaria and vitellaria samples from mated queens (MQ) were identified by numbers and corresponded to the pooled brains from the same queens that were similarly numbered [14].

The total RNA was extracted from germaria and vitellaria of the same queen groups, and from whole ovaries of virgin alate queens. Each pool was used to synthesize a library enriched in the respective ovarian structures (see Sect."RNA Sequencing and bioinformatic analyses"). RNA extraction was performed using the Trizol/chloroform method, as previously described [14]. In summary, the tissues were homogenized by vortexing twice for 30 s. Then, 400 µl of TRIzol was added and thoroughly mixed. The mixture was left at room temperature for 10 min. After the incubation step, 100 µl of chloroform was added, and the solution was mixed for 15 s and then incubated for 5 min at room temperature. Following separation by centrifugation (13,000 RCF for 30 min at 4 °C), the aqueous phase was transferred to a new tube, and RNA was precipitated overnight at − 20 °C using 500 µl of isopropanol and 10 µg RNase-free UltraPure glycogen (Thermo Fisher Scientific). After discarding the supernatant, the pellet was washed with 500 µl of 75% ethanol by carefully inverting the tube to ensure thorough washing without disturbing the pellet. DNA was removed using TURBO DNA- free™ Kit (Life Technologies, Carlsbad, CA). The RNA was quantified using Infinite^® 200 NanoQuant (Tecan Trading AG, Switzerland). RNA was then stored at − 80 °C.

Most of the obtained RNA was sent for sequencing at the same time as the brain libraries of the Calkins et al. 2018 study (see Sect."RNA Sequencing and bioinformatic analyses"). The remaining fraction of the total RNA was used to synthesize cDNAs (see Sect."Transcriptome verification by RT-qPCR of selected genes ") from each pool of vitellaria and germaria from mated (MQ) and newly mated queens (NMQ), and from virgin alate queen ovary pools, which were kept at −80 ºC and used for RT-qPCR in the current study.

RNA sequencing and bioinformatic analyses

RNA samples were submitted to the AgriLife Genomic and Bioinformatic Center for transcriptome sequencing. After RNA quality and quantity assessment by Bioanalyzer, two newly mated queen germarium samples (samples 3 and 6) were pooled to reach the required amount of RNA for sequencing. The Illumina TruSeq Stranded Total RNA library preparation kit was used to prepare the libraries, which were sequenced using the Illumina HiSeq 2500 System (single-end 125SE). RNA-seq reads were quality-filtered using FASTQC tools, and Q30 (Phred quality score) evaluation was performed. The processed reads were uploaded to the CyVerse Discovery Environment web interface for bioinformatic analyses. The reads were aligned with the S. invicta transcriptome (Ensembl Metazoa GCA_000188075.1.34) using Kallisto (v.0.43.1) [38] with the bootstrap set to 100. Sleuth was run to identify differentially expressed genes [39]. Principal component analysis (PCA) to evaluate the variance between distinct sample groups as well as similarity within sample replicates was performed and visualized using the Sleuth interactive R Shiny package [40]. Gene expression analyses focused on comparisons between NMQ vs. MQ germaria and NMQ vs. MQ vitellaria. Genes that exhibited a q-value of less than 0.05 were considered differentially expressed. To identify DEGs, there was not a priori set up of a specific fold change threshold, but the fold change for each gene are provided in the supplementary tables. GO term enrichment analyses were performed using g: GOSt in the g: profiler [41] selecting S. invicta as organism, and using the default parameters “only annotated genes” in Statistical domain scope, and “g: SCS threshold” in Significance threshold for multiple testing correct. A Venn diagram was made using the online calculator [42] with the LOC numbers of the DEGs from the germaria and the vitellaria analyses to identify LOCs of DEGs common to both analyses. This approach sought to identify genes that may be under similar or opposite regulation in both structures.

Transcriptome verification by RT-qPCR of selected genes

We selected eight genes of interest (GOIs) to validate and verify the differentially expressed genes identified in the vitellaria and germaria comparisons between mated queen (MQ) and newly mated queens (NMQ). These genes included phenoloxidase 1 (LOC105197777), prostaglandin E₂ receptor EP2 subtype (LOC105193552), protein yellow (LOC105192829), vitellogenin 3 (LOC105205783), insulin-like growth factor 2 mRNA-binding protein 1 (LOC105197398), vitellogenin receptor (LOC105200757), insulin-like growth factor-binding protein complex acid labile subunit (LOC105207810), and insulin gene enhancer protein ISL-1 (LOC105199685). For RT-qPCR, cDNAs had been synthesized in 2016 from fractions of the total RNA extracted from each pool utilized for RNaseq. The cDNA samples were generated using 140 ng of total RNA and the SuperScript^® III First-Strand Synthesis System from Thermo Fisher Scientific, following the manufacturer’s instructions. The final volume of the cDNA synthesis was 20 µL.

For each reaction, 5 µl of PowerUp Sybr™ Master Mix from Applied Biosystems, 300 nM final concentration of each primer (Integrated DNA Technologies, Inc., Coralville, Iowa), and 2 µl of cDNA were added. The total reaction volume was adjusted to 10 µl with nuclease-free water. Reactions were prepared and run with two to three technical replicates for each GOI, cDNA sample, and four to five biological replicates. All reactions included a negative control group (no template control, NTC, two wells per primer) to check for possible contamination. The following thermocycling program was used in a ThermoFisher QuantStudio Flex 6 qPCR thermocycler: 10 min at 95 °C, followed by 40 cycles of 95 °C for 15 s and 60 °C for 30 s. All primer sequences utilized in these analyses are in Table S1. The oligo sequences were designed according to qPCR specifications. Primer specificity was monitored by melting curve analysis in QuantStudio software V1.3 or Design & Analysis (V2.8.0) Apps (Thermo Fisher Scientific) and by sequencing PCR products.

The reference genes elf1-beta (LOC105201189) and rpl18 (LOC105198518) have been validated for S. invicta [14, 43] and were used in this study. Data were analyzed using the ΔΔCt method, with both reference genes utilized for analysis [14, 44]. Statistical differences were determined using the unpaired t-test considering a p-value of less than 0.05, using GraphPad Prism software version 10.1. For RT-qPCR analysis of GOIs, four pools of each ovary or ovary structure (germaria, vitellaria) were used, as to have enough cDNA to include the two reference genes in the analyses.

Results and discussion

Sequencing results

A total of 24 libraries were sequenced: five ovary libraries from virgin queens, four germaria libraries from newly mated queens (NMQ), five vitellaria libraries from NMQ, five germaria libraries from mated queens (MQ), and five vitellaria libraries from MQ. The Q30 analyses of all libraries showed excellent base call accuracy. We obtained over 14 million reads per library (Table S2). We performed a read saturation analysis to evaluate effective sequencing depth on transcript discovery. At baseline we found approximately 18,000 expressed transcripts and despite an increase of read depth of up 4-fold, this only resulted in the detection of additional ~ 800 transcripts (4%), which indicated the vast majority of the transcriptome was captured by our analyzed dataset (Suppl. Fig S1A). The divergence among libraries was evaluated using the Jensen-Shannon divergence analysis (Suppl. Fig S1B).

Principal component analysis of mated queen germaria and Virgin queen ovaries

Because of the high reproductive potential of polygyne fire ant queens, we first examined the effect of mating on the transcriptional activity of ovaries by comparing the virgin alate ovaries to the germaria of newly mated queens (NMQ) and mated queens (MQ) (Fig. 1). Exposure to queen pheromone promotes changes in gene expression in social insects [45]. Because in fire ants older or disinhibited virgin queens can be vitellogenic, we anticipated that the vitellarium of virgin queens would have some expressed genes not directly regulated by mating. To avoid this “gene expression background noise unrelated to mating”, we only used for sequencing the ovaries of virgin queens within the colony that were not observed to be vitellogenic (without visible eggs) and had undeveloped ovaries of small size, in which case the separation of germaria and vitellaria was not possible. In sum, the alate virgin queens used for transcriptomics were “pheromone inhibited” queens.

Fig. 1.

PCA analysis of libraries from alate queen whole ovaries (virgin queens, VQ), NMQ germaria, and MQ germaria. Library groupings showed a clear effect of mating on the germaria along the PC2 axis (MQ, green and NMQ, blue vs. VQ, red). Further, along the PC1 axis a clear differentiation between mated and newly mated queens is seen. Of note is that both, the virgin alate queen ovary libraries and the NMQ germaria libraries were similarly dispersed along the PC1 axis, perhaps reflecting variations linked to hormonal levels and nutritional status or other previous intra-colony effects

The PCA analysis showed a clear separation of the three library types: the least variation along both components was observed for the mated queen germaria, while both the newly mated queen germaria and the ovaries of alate queens showed high variation along the PC1 axis, but a clear separation along the PC2 axis. NMQ germaria were closer to the MQ germaria in this latter axis. We speculate that perhaps the distinction shown along the PC2 axis between virgin alate queen (VQ) ovaries and NMQ germaria reflects a temporal snap-shot of the transition that occurs in gene expression after mating. There is a noteworthy and similar variation along the PC1 axis in both the VQ ovaries and the NMQ germaria libraries. Perhaps, the variation seen in NMQ libraries along the PC1 axis is a remnant of the variation observed among virgin ovary samples and might reflect variation in nutritional status or in the level of virgin queen disinhibition from mated queen pheromone before the mating flight, that is, reflects variation in the overall reproductive maturation status, and this variation may disappear after a certain time, once the NMQ germaria gene expression is sufficiently modified by the “mating effect” and their increased age. PCA analyses of gene expression comparing full ovaries of virgin, newly mated and mated queens also showed a similar clear separation of these groups in honey bees (Apis mellifera) [46]. In our study, details of gene expression in virgin ovaries compared to newly mated or mated germaria are reported in Tables S3.1 and S3.2, respectively. The 42 enriched GO terms identified in the analysis of DEGs between virgin queen whole ovaries and NMQ germaria are shown in Table S3.3. The GO enrichment analysis of these DEGs showed the most significantly enriched term under Molecular Function was protein binding, with 360 genes (Table S3.3). Under Biological Processes (BP) there was an enrichment in terms associated with developmental processes, anatomical structure development and multicellular organism development, supporting that mating activated the NMQ germaria towards oogenesis. Kocher et al. [46] when comparing the effect of mating in honey bees and Drosophila also found protein binding and tissue development as enriched GO categories. Further, in our analysis, 18 DEGs associated with the enriched GO term “Signaling receptor binding” were identified (Table S3.4). Transcripts from four Wnt genes, two laminin genes, three semaforins, two from insulin-like peptides, and one from insulin receptor substrate were found. These genes may be candidates involved in integrating the mating signal and the overall queen reproductive capacity and highlights the prominent role of the insulin pathway in queen reproduction.

The 75 enriched GO terms identified in the analysis of DEGs between virgin queen whole ovaries and MQ germaria are shown in Table S3.5. There was no enrichment in a specific KEGG pathway. For each GO aspect of interest were: within molecular function, there was enrichment in GO terms associated with meiotic cell cycle, kinase activity, nucleotide binding, transcription regulator activity; within the biological processes, there was enrichment in signaling, meiotic cell cycle, developmental process, sexual reproduction and reproductive process; under cellular component there was enrichment in organelle related terms, ATPase complex, and transcription regulator complex, among others.

Taken together these two GO term enrichment analyses identified different terms: they show the activation of the germaria in NMQ in response to mating with many signaling ligands such as insulin-related peptides, and regulation of development, while in the MQ comparison the GO terms are indicative of ongoing oogenesis.

PCA of mated queen vitellaria and germaria

Upon fire ant queen collections, dissections of germaria, vitellaria, and brains were performed on more than 100 fire ant queens, achieving a balance among precision in the dissections, preserving RNA of high quality, and timeliness for maintaining the “newly mated queen” category as such. These constraints required that all dissections be completed within 24 h of collection after their respective mating flight. Principal component analyses were conducted to explore the variation in gene expression between the mated queens germarium and vitellarium libraries (Fig. 2). The PCA analyses revealed a good separation in the first component between the germarium and vitellarium samples. For our dissections, even though roughly the top third of the ovary was considered germarium under the dissecting microscope and no examination was performed under a compound microscope, the germaria and vitellaria libraries showed anatomical structure-specific gene expression. These results supported that the dissections of germaria and vitellaria were performed correctly based on their morphology and indicated a clear separation of gene expression of these ovarian regions reflecting their functional specialization (germarium initiates oocyte development by germ cell division, while the vitellarium supports maturation and yolk deposition). Further, these ovarian structures physiological specializations superseded the hierarchy of mating status along the PC1 axis. However, this hierarchy could be seen along the PC2 axis where the MQ germaria could be clearly differentiated from the NMQ germaria libraries, while the vitellaria did not show separation by mating status along the PC2 axis (Fig. 2).

Fig. 2.

PCA analysis to explore the variation in gene expression among germaria and vitellaria of newly mated queen and mated queen. The PCA showed the smallest variation for the mated queen germaria libraries along both PCs (green) but a large variation in gene expression along the PC1 axis among mated queen vitellaria (turquoise). While the newly mated queen germaria (orange) and vitellaria (purple) were well separated along the PC1 axis by tissue type, the separation among libraries for these tissues was mainly seen along the PC2 axis

PCA of germaria and vitellaria according to mating status

As the regional specialization of germaria and vitellaria is well established in insects and supported in the fire ant queens by the analysis in Fig. 2, we focused our analyses comparing the differences in gene expression between these regions based on their mating status.

The PCA for germaria libraries (Fig. 3A) reflects a greater variation in gene expression among the NMQ germaria libraries (Fig. 3A, orange) than the MQ germaria libraries (Fig. 3A, blue), as the former were more dispersed along the PC1 axis. In contrast, the MQ germaria libraries were closely clustered around similar values for PC1. In addition, the MQ germaria libraries were more dispersed along the PC2 axis (Fig. 3A, blue), indicating a different source of variation in gene expression for MQ germaria. The PC2 appears to be a lesser source of variation for NMQ germaria as the libraries are closer in value along the Y-axis (Fig. 3A).

Fig. 3.

PCA analysis to explore variation in gene expression among fire ant queen ovary libraries emphasizing mating status (A) PCA analysis of germarium libraries of NMQ (orange) and MQ (blue), and (B) PCA analysis of vitellarium libraries of NMQ (orange) and MQ (blue). In library names g refers to germaria and v to vitellaria. The number in the library designation refers to the group of queen tissues extracted from the same queens; i.e., MQg1 and MQv1 correspond to the germaria and vitellaria of the same group of dissected mated queens. NMQg3 + 6 indicates that the total RNA of newly mated queen germaria from groups 3 and 6 were pooled for sequencing. For germaria, PC1 explained 63.48% of the variation among libraries while the PC2, 14.44%. Other three components explained variation as follows: PC3, 9.45%, PC4, 5.84% and PC5, 3.7% (not shown). For vitellaria, PC1 explained 76.77% of the variation, while PC2 13.64%. Other three components explained variation as follows: PC3, 6.13%, PC4 1.35% and PC5, 0.72%

In contrast, there was an overlap of NMQ and MQ vitellaria libraries along both PC axes (Fig. 3B). The large variation in MQ vitellaria libraries (blue) was reflected in the broad distribution of these libraries along the PC1 axis; and in stark contrast, the NMQ libraries were clustered along the PC1 axis, they were widely distributed along the PC2 axis (Fig. 3B, orange). The PC1 and PC2, therefore, appear to represent factors that differentially affect the variation in gene expression in NMQ and MQ vitellaria libraries.

The overall PCA revealed that the germaria libraries were clearly separated by mated status and could be grouped without overlap (Fig. 3A). In contrast the vitellaria libraries could not be grouped by mating status without overlapping (Fig. 3B). Therefore, mating appeared to have a strong detectable effect in the overall gene expression in germarium stem cells as oogenesis proceed at “full throttle” [47]. In virgin queens, vitellogenesis is independent of mating because as virgin queens age within the colony, their level of JH increases as well as the transcript for the vitellogenin receptor, and thus, can develop vitellogenic oocytes [24, 27]. Significantly, the numbers that follow the designations of the libraries in the Figs. 3A for germaria and 3B for vitellaria correspond to the same group of queens. Unfortunately, the libraries for NMQ germaria groups 3 and 6 had to be merged (Fig. 3A) because of low RNA yield for sequencing. Analyzing the patterns in both panels, at least for the mated queens the relative spatial distribution in the graphs is similar, i.e. along PC2 in Fig. 3A, MQg2, MQg6 are closer to each other and separated from the other MQ germaria libraries, and in Fig. 3B MQv2 and MQv6 are also clustered along the PC1 axis. Perhaps this illustrates that the same factor may be regulating or affecting gene expression in the vitellaria and germaria of queens. These factors could be dominance, ovary size, nutritional status, and age, and for NMQ, their pre-mating reproductive maturation/nutritional status and age.

Identification of differentially expressed genes in germaria

The comparison of NMQ and MQ germaria libraries identified 724 differentially expressed genes, those with q-value < 0.05 (Table S4.1). GO enrichment analyses of the DEGs with g: Profiler identified 22 enriched terms (Table S4.2). Among them, the six driver GO terms identified are presented in Fig. 4. The molecular function category included protein kinase and transcription regulator activities; under the biological processes, the terms biological regulation, protein phosphorylation, regulation of DNA-templated transcription, and developmental processes were enriched. No driver GO terms were identified under cellular component (Table S4.2).

Among the GO categories identified herein, “phosphorylation” had been identified among the conserved building blocks involved in the parallel evolution of diverse phenotypic traits in several species of ants that were identified by comparative transcriptomics, and listed within a “Queen” Module 2 (characterized by single queens with sterile workers in a non-invasive species) [48].

Of interest, genes identified within the “developmental process” category included twelve genes. Some of those involved in embryogenesis were up-regulated in the NMQ germarium compared to the MQ germarium (Table S4.1). Examples of those genes were ets DNA-binding protein pokkuri variant X3 and Down syndrome cell adhesion molecule-like protein Dscam2 transcript variant X7. Also upregulated in this category were genes involved in Notch signaling such as the putative ligand protein jagged-1; this pathway is involved in oogenesis and maintenance of germ-stem cells [49]. In the MQ, laminin subunit alpha-2 and titin were up-regulated in germaria compared to NMQ. In Drosophila ovaries, the expression of laminin genes was up-regulated by JH in muscle cells contributing to egg shape and migration [50]. This may be similar in fire ant queens because JH is a gonadotropin.

Identification of differentially expressed genes in vitellaria

The NMQ and MQ vitellaria library comparison identified 424 differentially expressed genes, those with q-value < 0.05 (Table S5.1). GO enrichment analyses identified the Biological Process term GO:0055086: nucleobase-containing small molecule metabolic process, which includes enzymes involved in glycolysis and glycogen synthesis. The second enriched term was KEGG:01212: Fatty acid metabolism (Table S5.2), which included three genes: non-specific lipid transfer protein (LOC105201416), long-chain-fatty acid CoA ligase 3 (LOC105196248), and probable enoyl CoA hydratase-mitochondrial (LOC105193004). This result is similar to previous studies on S. invicta and H. saltator, which also found an increase in the transcription of genes associated with fatty acid metabolism in queens [51, 52].

Identification of differentially expressed genes between NMQ and MQ common to both the germaria and vitellaria

Because mating had a strong effect on the overall gene expression of the germaria (Fig. 2), but not in the vitellaria, we sought to identify which genes, if any, were regulated in both anatomical structures. There were 61 DEGs common to both analyses, the germarium and vitellarium tissues (Fig. 5, Table S6). The Venn diagram (Fig. 5A) was made with the LOC numbers from the germarium (Table S4.1) and vitellarium analyses (Table S5.1). Table S6 shows: the 61 LOC numbers; the various transcripts predicted for each gene; the results of the transcriptome analysis as to whether each transcript was upregulated or downregulated in each of the comparisons (i.e. NMQ Germarium vs. MQ Germarium, and NMQ Vitellarium vs. MQ Vitellarium), including information for those transcripts that were not DEG (i.e. their q-value was non-significant).

Fig. 5.

Shared DEGs between MQ and NMQ in germaria and vitellaria. (A) Venn diagram showing the common DEGs between NMQ and MQ germaria and vitellaria made by comparing genes with q value < 0.05 from transcriptome analyses performed between germarium of newly mated queen and mated queen libraries (Table S4.1) and between vitellarium of newly mated queen and mated queen libraries (Table S5.1). (B) Pie-chart showing the gene ontology categorization of the 61 common DEGs between tissues of NMQ and MQ as a biological process, cellular component, and molecular function

While there were 724 and 424 DEGs identified respectively in the germaria and vitellaria analyses between NMQ and MQ, these numbers corresponded to transcripts. Figure 5A shows the number of genes (LOCs). For these comparative analyses, the LOC number was used instead of the transcripts’ ID for two reasons. First, to avoid including similar multiple transcript isoforms of the same gene in the overall count of shared DEGs, and second, to account for genes with transcript variants differentially regulated in each ovary structure. About the latter, for example, genes that had a transcript differentially expressed in an ovary region-specific manner, such as LOC105200853 (Table S6).

Table S6 also includes the GO terms categorization from Ensembl metazoa and NCBI. Among the 61 shared DEGs, 31% were linked to the three main GO categories: cellular process, molecular function, and cellular component; 16% were only linked to biological process and molecular function; 12% were linked to cellular component and molecular function; 13% were only associated to molecular function, while 2% and 3% were characterized as part of biological process or cellular component, respectively, and 23% were uncharacterized (Fig. 5B).

Overall, the 61 shared DEGs represent a low percentage of the DEGs identified either in the germaria or vitellaria (they were 9.2% of those DEGs identified in the germaria, and ~ 15% of those DEGs identified in the vitellaria). This low overlap in the number of genes regulated in both structures is congruent with the functional biological specialization of these ovarian regions and with the fact that mating had a strong effect on gene expression on the germaria but not the vitellaria. The 61 common DEGs included many associated with immunity, lipid metabolism, development, and transcriptional regulation.

Differentially expressed genes between NMQ and MQ

Among the DEGs identified by the transcriptome analyses were candidate genes previously associated with immunity, reproduction, and development in S. invicta queens and/or other social insects [14, 53].

Among the immunity-associated DEGs, phenoloxidase 1 (PO) (LOC105197777) was up-regulated in both tissue regions in the mated queens compared to the newly mated queens (Tables S4, S5). Further, it had been previously also identified as a DEG and upregulated in the brains of S. invicta mated queens vs. virgin alate queens [14]. In insects, PO is involved in various processes such as egg production, pigment synthesis, molting, or sclerotization [54].

The prostaglandin E₂ receptor EP2 subtype (PGE₂rec) (LOC105193552) is a G protein-coupled receptor involved in the G protein signaling pathway (GO:0007186) and is linked to various aspects of ovarian development and cellular immune responses [55–57]. This LOC is associated with three transcripts (Tables S4.1, S5.1). The transcriptome analyses revealed one transcript (XM_039454095.1) up-regulated in the mated queen vitellarium compared to the newly mated queen, but another transcript (XM_026135469.2) was down-regulated in the mated queen germarium (Tables S4, S5). Both transcripts encode a protein of 398 amino acid residues and their differences lie in the 5’ UTR, which is shorter in XM_039454095.1. It must be noted that it is not known whether these transcripts originated from the ovary tissues per se, because some contamination with fat body, muscles, and trachea was unavoidable during dissections. The three transcripts were expressed in vitellaria.

The protein yellow (Pt Ylw) also known as major royal jelly protein (MRJP), plays a role in fecundity, aging, colony organization, and caste-specific physiology in social insects, particularly in bees, and is important for egg chorion rigidity, nutrient availability, pigmentation, and eggshell formation [32, 58–60]. Transcript XM_026140198.2 (LOC105192829) was down-regulated in the germarium and vitellarium of the mated queens (Table S4, S5). In the vitellarium, two other protein yellow genes were differentially expressed (LOC105199231 and LOC105201282) between MQ and NMQ, both were upregulated in MQ (Table S5). Previously, two unique yellow-g like proteins (CB252011.1 and CB252010.1), the first now corresponding to LOC105199156 (XM_026139259.2), and the second to LOC105199178, both currently annotated as major royal jelly protein 1, were identified as having higher expression in mated queens by northern blot [15]. We found these genes LOC105199156 (XM_026139259.2) and LOC105199178 (XM_011166151.3) downregulated in the MQ germarium (Table S4), but they were not differentially expressed in the vitellarium (Table S5). The existence of numerous copies of MRJP genes has been observed in Hymenoptera through comparative studies, suggesting an origin from the duplication of yellow genes. This finding highlights the intricate nature of their functions and regulatory mechanisms [32, 58].

Other genes of interest only identified as DEGs in the vitellaria were Vitellogenin-3 (Vg3) and the vitellogenin receptor (VgR), which play a role in nutrition, growth. The Vg3 gene is preferentially expressed in fire ants queens compared to workers [31] and expression is higher in mated queens compared to alate virgin queens [31, 51]. In agreement with these previous reports, the Vg3 transcript NM_001304585.1 (LOC105205783) was upregulated in the vitellarium of mated queens (Table S5), and while Vg1 and Vg2 were expressed, they were not differentially regulated. The Vg receptor (VgR) protein was previously found to be highly expressed in the periphery of oocytes in virgin queens approximately two weeks after their emergence, the period necessary for alates to achieve “reproductive maturation” and flying age. While the VgR protein is weakly expressed in small developing oocytes in virgins [27] the proper localization of the receptor in the periphery of the oocyte for successful vitellogenesis coincides with their minimal flying age of about two weeks. As expected, the VgR transcript NM_001304596.1 (LOC105200757) was upregulated in the vitellarium of mated queens (Table S5). Neither vitellogenin nor its receptor were differentially expressed in the germarium between MQ and NMQ. This is not surprising because of the predominant role of the vitellarium in the fast and highly productive fire ant queen [27].

We previously reported two insulin receptors in the fire ant, a species in which the workers have no reproductive ability, which allows us to more clearly define the role of the IIS pathways in nutrition and reproduction when comparing these two castes. When comparing queen tissues, it was concluded that SiInR-1 plays roles in queen physiology (including fat body energy metabolism) and reproduction, and the SiInR-2 plays roles in body size growth and development [25]. When comparing whole ovaries of MQ vs. virgin queens (VQ) no differences in transcript expression were found by RT-PCR [25]. Herein, transcripts for both receptors were identified but they were not differentially expressed. The LOC105207962, identified as insulin receptor-like NM_001304600.1, previously known as SiInR-1 and LOC 105195102, identified as insulin-like peptide receptor (XM_026133454.2) previously named SiInR-2 were not differentially regulated between MQ and NMQ, corroborating our previous results [25].

However, it was of interest to identify DEGs in fire ant queen ovaries potentially involved in insulin signaling in the long-lived reproductive caste of ants [26, 52]. The insulin-like growth factor 2 mRNA-binding protein 1 (ILGF2) (XM_011163719.3, LOC105197398) was upregulated in the MQ vitellarium vs. the NMQ vitellarium (Table S5.1). The insulin-like growth factor-binding protein complex acid labile subunit (ILGF_bpcl) (XM_039457458.1, LOC105207810), and the insulin gene enhancer protein ISL-1 (IGEP_ISL) (XM_039451697.1, LOC105199685) were upregulated in MQ germarium vs. NMQ (Table S4.1). Similar to our findings, an increase in expression of genes encoding secretory proteins, including an acid-labile subunit and an IGF-binding protein 7 also were found in H. saltator gamergate, genes which in queens are candidates associated with reducing the common trade-off between increased reproduction and shorter lifespan [52].

Gene expression analyses in newly mated queens germarium versus mated queens germarium

Among the previously described DEGs between NMQ and MQ in the germarium and/or vitellarium (Tables S4.1, S5.1), we selected eight genes of interest (GOI) to be evaluated by RT-qPCR. These genes were chosen because they are potentially involved in processes of oogenesis and reproductive function, including immune response, insulin pathway signaling, and oocyte development. Genes from the royal jelly and vitellogenin protein families were also selected due to their role in nutrition and reproductive differentiation in social insects: PO1, PGE₂ receptor, protein yellow, ILGF_bpcl, ILGF2, IGEP_ISL, vitellogenin 3, and its receptor. For this selection, gene function categories were given priority over the magnitude of DEGs.

Confirmation of the high correspondence of the transcriptome analyses results with the validation results by RT-qPCR utilizing as template cDNAs synthesized from the same total RNA that was sent for sequencing is shown in Suppl. Figure 2. For example, in panel A of this figure, the box plot for mated queen germaria library 4 shows the highest level of expression for PO1, and similarly, this high level of expression was confirmed by the lowest delta Ct value of 5.89 obtained from the corresponding synthesized cDNA.

PO1 (p < 0.0001) and IGEP_isl (p = 0.0405) were upregulated in MQ germarium compared to NMQ germarium (Fig. 6A, E) with a fold-change of 4.7 for PO1 and of 1.6 for IGEP_isl (not shown). On the other hand, PGE₂rec was down-regulated (p = 0.0038) in the MQ germarium (Fig. 6B) and a fold-change of 0.2 (not shown). These results were consistent with the results of the transcriptome analysis (Tables S4.1). Pt Ylw showed a downward trend in the MQ germarium similar to the transcriptome results, however, the RT-qPCR did not allow detection of significant differences (p = 0.0749) (Fig. 6C). ILGF_bpcl (p = 0.6522) showed no significant differences in expression levels (Fig. 6D).

Fig. 6.

Analysis of expression by RT-qPCR of five GOIs differentially expressed in the germaria of NMQ vs. MQ transcriptome. Gene expression was determined by RT-qPCR and normalized to the average of elf1-beta and rpl18. Horizontal lines represent mean ΔCts ± SEM. *Indicates significant differences between NMQ and MQ germarium determined by unpaired t-test, p < 0.05. PO1 (p < 0.0001) and IGEP_isl, (p = 0.0405) were up-regulated in the germarium of MQ compared to NMQ, while PGE₂rec (p = 0.0038) was down-regulated. P values Pt Ylw, 0.0749; ILGF_bpcl, 0.6522

Gene expression analyses in newly mated queens vitellarium versus mated queens vitellarium

When comparing the vitellarium of MQ and NMQ (Fig. 7), we found that PO1 (p = 0.005) (Fig. 7A) and Vg3 (p = 0.0117) (Fig. 7D) were up-regulated in the MQ compared to the NMQ vitellaria, which was consistent with the transcriptome analysis results (Table S5.1). PO1 had a 3.6-fold-change while Vg3 had 7.1 fold-change (not shown). In contrast, the RT-qPCR results showed that PGE₂rec (p = 0.023) was downregulated in the MQ vitellaria, contradicting the transcriptome data, and with a fold-change of 0.4. This gene has three different splicing forms, while the three were expressed in NMQ and MQ vitellaria, only XM_039454095.1 (Table S5.1), was upregulated in MQ vitellaria with respect to NMQ vitellaria. Because the primers used for RT-qPCR could not discriminate the transcript forms the RT-qPCR could not reflect the higher expression in the vitellarium for the transcript XM_039454095.1. Pt Ylw (p = 0.2319), ILGF2 (p = 0.8606) and Vgrec (p = 0.9033) did not show significant differences in expression levels (Fig. 7).

Fig. 7.

Analysis of expression by RT-qPCR of six GOIs differentially expressed in the vitellaria NMQ vs. MQ transcriptome. Gene expression was determined by qPCR and normalized to the average of elf1-beta and rpl18. Horizontal lines represent mean ΔCts ± SEM. *Indicates significant differences between NMQ and MQ vitellarium as determined by unpaired t-test, p < 0.05. PO1 and Vg3 were up-regulated and PGE₂rec was down-regulated in the vitellarium of MQ compared to NMQ. P values PO1, 0.005; Vg3, 0.0117, PGE₂rec, 0.023; Pt Ylw, 0.2319; ILGF2, 0.8606; Vgrec, 0.9033

Gene expression analyses by RT-qPCR in ovaries of alate virgin queens versus mated queens germaria

Furthermore, we examined by RT-qPCR if the DEGs identified in the transcriptome between germaria of NMQ and MQ (Suppl. Table S4.1) were also DEGs between whole ovaries from virgin alate queens vs. the MQ or NMQ germaria (Suppl. Tables S3.1, S3.2 and Suppl. Figs. S3, S4).

In the transcriptome, no differences in PO1 expression were found when comparing the virgin queen whole ovary to NMQ germaria (Tables S3.1), a similar result was obtained by RT-qPCR (p = 0.3552) (Fig. S3A). In the transcriptome comparison of NMQ vs. MQ germaria PO1 was upregulated (Table S4.1) in the MQ, these results coincided with the RT-qPCR analysis (Fig. 6A). As expected, this gene was also up-regulated in MQ germarium transcriptome compared to the transcriptome of whole ovary of virgin queen (Table S3.2) and the RT-qPCR similarly detected up-regulation in MQ germarium for PO1 (p = 0.0056, and a fold-change of 7.6) (Fig. S4A).

We found that the PGE₂rec gene was up-regulated in NMQ germarium compared to alate queen virgin whole ovary (p = 0.0029) (Fig. S3B) with a fold change value of 4.0. In the latter comparison we also observed a trend for upregulation of protein yellow (p = 0.062) (Fig. S3C), but similarly to other genes, ILGF_bpcl (p = 0.12), and IGEP_ISL (p = 0.6134) (Fig. S3, D and E) were not differentially expressed. In the transcriptome analysis (Table S4.1) the expression of PGE₂rec decreased in the MQ germaria compared to the NMQ (Table S4.1). Therefore, the expression of PGE₂ receptor increases upon mating in the germaria of NMQ compared to the whole ovary of virgin queen and decreases in mated queen germaria, suggesting that this regulation is directly linked to reproductive success for germarium activation immediately after the copulation.

In the RT-qPCR analysis comparing the germarium of the MQ with the whole ovary of virgin queen, ILGF_bpcl (p = 0.033, and a fold-change of 2.0) was upregulated in the MQ germaria (Fig. S4D). However, the expression levels of PGE₂rec (p = 0.9918), Pt Ylw (p = 0.7816), and IGEP_ISL (p = 0.1446) did not show significant differences (Fig. S4 B, C, E).

A summary of all results from the RT-qPCR analyses is presented pictographically in Fig. 8 showing gene upregulation below each queen-type icon for clarity, contrasting gene expression in ovaries of alate virgin queens vs. germaria of newly mated and mated queens; germaria of newly mated vs. mated queens, and finally, vitellaria of newly mated vs. mated queens.

Fig. 8.

Summary of relative expression of genes in fire ant queen ovaries. (A) PGE₂rec was up-regulated in the germarium of the NMQ compared to the entire ovary of AVQ. (B) PO1 and ILGF_bpcl were up-regulated in the germaria of MQ compared to ovaries of AVQ. (C) PO1 and IGEP_isl were up-regulated while PGE₂rec was down-regulated in the MQ germaria vs. NMQ germaria. (D) PO1 and Vg3 were up-regulated, while PGE₂rec was down-regulated in the vitellarium of MQ compared to NMQ vitellaria

Ovarian transcriptomes of germaria and vitellaria revealed reproductive specialization and an immune activation

The transcriptomic analyses revealed a clear effect of mating on the germaria of mated queens of unknown age, as these libraries were differentiated from those of newly mated queens in the PCA analyses (Figs. 1, 2 and 3), and results from RT-qPCR validated these conclusions (Fig. 8B, e.g. PO1 and ILGF_bpcl were upregulated in MQ vs. ovary of alate virgin queens).

Mated queens (MQ) had increased transcriptional expression of genes involved in vitellogenesis in the vitellaria, such as Vg3, and for immunity, especially those genes involved in humoral immunity, like phenoloxidase, in both germaria and vitellaria. Mating prepares the queens to face potential pathogens in the soil as they are ready to lay their first eggs potentially alone, however polygyne queens are known to return to their original or other polygyne nest. Thus, based on the present analysis, phenoloxidase and the prostaglandin receptor genes can be added to the list of previously reported immune genes for S. invicta queens that are significantly upregulated after mating, which include the antimicrobial peptide abaecin and hymenoptaecin [14, 15]. Regulation of immunity-related genes also occurs in queens of other ant species upon copulatory damage or in response to pathogens [37, 61, 62].

In comparison to the germaria libraries that were mostly influenced by mating status, the vitellaria libraries of mated and newly mated queens overlapped in variation (Fig. 3B) and were likely more influenced by nutrition, age, or other factors not related to mating per se, but related to social context, and perhaps the differential social status of queens, such as developing towards dominant or subordinate. We previously demonstrated these effects through experiments manipulating the queens’ social context and then analyzing the brain gene expression of the same queens which ovaries were used in the current study [14]. In the current analyses, one notable aspect was that some genes with more than one transcript variant yielded conflicting or ambiguous results if the primers used for RT-qPCR validation were not specifically designed for the specific differentially expressed variant. Large discrepancies in the number of reads per variant may obscure the physiological significance of the less abundant transcript variants, such as for the prostaglandin E₂- receptor (PGE₂rec) gene.

The regulation of the PGE₂ receptor gene in the ovary is noteworthy. Of the three transcripts identified, XM_039454095.1 was not expressed in the germaria of either NMQ or MQ. However, in the germaria of NMQ a second transcript, XM_026135469.2 was significantly upregulated in comparison to the germaria of MQ. The third transcript, XM_026135472.2 had a trend of higher abundance in the NMQ germaria but it was not significantly different, with a q-value of 0.057 (Fig. 8C, Table S4.1). In contrast, in the MQ vitellaria the transcript XM_039454095.1 was significantly overexpressed with respect to vitellaria of NMQ, while the other two transcripts were not differentially expressed. However, the RT-qPCR analyses revealed a higher expression in NMQ (Figs. 7 and 8D). Despite these apparently contradicting results, the PGE₂ receptor is highly expressed in the vitellarium after mating. In other insects the prostaglandin receptor was implicated in promoting egg development, releasing egg-laying behavior, and functioning in choriogenesis [63], as described in B. mori and D. melanogaster [64]. So, we speculate that in NMQ the transcript XM_026135469.2 and likely XM_026135472.2 may be involved in promoting egg production in the germaria while in the vitellaria of MQ it is likely that the three transcripts are involved in egg-laying behavior and choriogenesis. It will be important to determine if prostaglandin E₂ promotes oocyte development in NMQ germaria and contributes to releasing egg-laying behavior and/or choriogenesis in egg-laying mated fire ant queens. If prostaglandins had gonadotropin roles in fire ant queens as in crickets [55, 63, 65], and other insect species [34, 66, 67], this pathway could be a novel target for population control.

An additional value of this study is the identification of loci with multiple transcript variants that appear to be regulated in an ovary structure- or mating status-specific manner. Further investigation of these transcripts’ regulation may help in identifying gene networks controlled by specific miRNAs, lncRNAs, etc. These ongoing investigations will advance our knowledge on the “Central Nervous System-Ovary” axis for fire ant queen reproduction [14].

Increased expression of insulin genes in germaria after mating (ILGF_bpcl in MQ germaria vs. ovary VQ; IGEP_isl in MQ vs. NMQ (Fig. 8B, C) aligns with findings in other ant species [52, 68], reflecting the high metabolic demands of egg production and insulin signaling in reproductive ants [26, 52].

This study advances our understanding of insect reproduction by focusing on an invasive social species and can help identify marker genes for germarium and vitellarium development, mating signaling, and hormonal response. DEGs associated with phosphorylation, transcription regulation and developmental processes were identified primarily in the germarium, supporting the activation of this structure upon mating. Our hypothesis that mated queens will have a heightened synthetic metabolism in ovaries was supported because genes linked to metabolism of both fatty acids and nucleobases were predominantly regulated in the vitellarium. In contrast our hypothesis related to heightened immunity expected in NMQ was not fully supported because canonical immune candidate genes, such as PO1, showed higher expression in mated queens. Our transcriptome analyses found differential responses in germaria and vitellaria, congruent with the initiation of oocyte development and the completion of oocyte maturation, respectively. While transcriptional changes in ovaries [69] occur upon mating in most insects studied, our dissection strategy clearly revealed the significant regional regulation of ovarian gene expression. Therefore, previous studies using whole ovaries of social insects for analyses likely missed detecting significant regulators. Our comparison of whole ovaries of virgin queens vs. newly mated queen germaria also identified key receptor ligands, including several genes in the insulin pathway. The potential role of these signaling genes as markers for queen reproduction warrants future investigation.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

AA/TOR

Amino acid/Target of rapamycin

AVQ or VQ

Alate virgin queen

DEG

Differentially expressed gene

GO

Gene Ontology

ILPs

Insulin-like peptides

IGEP_isl

Insulin gene enhancer protein ISL-1

ILGF_bpcl

Insulin-like growth factor-binding protein

ILGF2

Insulin-like growth factor 2

JH

Juvenile hormone

KEGG

Kyoto Encyclopedia of Genes and Genomes

PO1

Phenoloxidase 1

PGE₂

Prostaglandin E₂

PGE₂rec

Prostaglandin E₂ receptor EP2 subtype

Pt Ylw

Protein yellow

MQ

Mated queen

NMQ

Newly mated queen

MAPK

Mitogen-activated protein kinase

RIFA

Red imported fire ants

Vg

Vitellogenin

Vg3

Vitellogenin 3

Vgrec

Vitellogenin receptor

Author contributions

TNF: Performed experiments, formal analysis, data curation, visualization, writing original draft, and editing final manuscript; M-EC: Performed experiments, writing methodology of original draft, and editing final manuscript. PS: Formal analysis and editing final manuscript. CT: Investigation, formal analysis, resources, data curation, writing and editing manuscript. PVP: Conceptualization, design, resources, supervision (TNF, M-EC), project administration, funding acquisition, investigation, writing and editing final manuscript.

Funding

Research was partially supported by NSF-IOS 1257837 and Urban Endowment Funding for Fire Ant FY24 from the Dept. of Entomology (TAMU), both to P.V.P. and C.T.

Data availability

The datasets generated and/or analyzed during the current study are available in the NCBI-SRA repository, under BioProject: PRJNA422376 (NOTE: Data were submitted, temporary submission IDs are: SUB15169890, SUB15169934, SUB15169942, SUB15169976, SUB15169968, SUB15170084, SUB15170091, SUB15170124, SUB15170104, SUB15170130, SUB15170136, SUB15170141, SUB15170148, SUB15170155, SUB15170165, SUB15170157, SUB15170172, SUB15170161, SUB15170180, SUB15182543, SUB15182605, SUB15182684, SUB15182684, SUB15182707. The supplementary materials Tables 1-6 provide all the analyzed results of the transcriptomes.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.