Hamiltonella symbionts benefit whitefly fertilization by regulating the maternal protein Tudor–mediated piRNA pathway

Significance

Intracellular symbionts tend to colonize host germ cells induced by maternally derived protein/mRNA. P-element-induced wimpy testis-interacting RNAs (piRNAs) guide insect germline development and reproduction. However, whether intracellular symbionts affect host reproduction by manipulating germline determinants or piRNAs remains unknown. Here, we demonstrated that the facultative symbiont Hamiltonella affects the abundance of a piRNA by altering tud level, thereby regulating the expression of the vacuolar (H+) -ATPase H subunit that maintains the homeostasis of intracellular energy and level of ATP in whitefly ovaries. The changes of ATP level remodel the F-actin pattern in ovaries and eggs, thus manipulating whitefly fertilization. This study suggests that the interaction among multiple maternally transmitted genetic elements: bacterial symbionts, maternal protein, and piRNAs regulates host reproduction.

Abstract

Although it is widely recognized that nutritional symbionts can manipulate host reproduction, the underlying molecular and cellular mechanisms are largely unclear. The facultative symbiont Hamiltonella in bacteriocyte induces female-biased sex ratio of whiteflies. Here, we demonstrate that a maternal gene tudor (tud) and its encoded protein have lower expression levels in ovaries of Hamiltonella-cured whiteflies. Tud family proteins can interlink the various stages of biosynthesis of PIWI-interacting RNA (piRNA), a class of small noncoding RNAs. We find that Hamiltonella affects the abundance of a piRNA through the maternal gene tud, thereby regulating the expression of the vacuolar (H+)-ATPase H subunit (VATPH), which is the switch of activity of the vacuolar (H+)-ATPase that plays a crucial role in maintaining the homeostasis of intracellular energy and supporting mitochondrial respiration. This regulation adjusts the ATP level in ovaries of whiteflies. The ATP level shapes the F-actin pattern in ovaries and eggs of whiteflies, ultimately manipulating whitefly fertilization. Silencing tud inhibited whitefly fertilization by impairing ATP levels and F-actin patterns in ovaries and eggs. This study reveals that symbiont and maternal protein associations can regulate host fertilization by piRNA biosynthesis.

Heritable microbes can regulate insect reproduction in diverse ways (1). For example, some symbionts which have been called reproductive manipulators, cause cytoplasmic incompatibility, male-killing, feminization, or parthenogenesis of their animal hosts (1–3). In contrast, other symbionts influence the reproduction of their hosts by providing essential nutrients that hosts cannot synthesize or by regulating the host hormone level (4–6). However, the molecular and cellular mechanisms underlying reproductive manipulation of animals by nutritional symbionts are largely unclear.

The whitefly Bemisia tabaci MEAM1 and MED are the globally important insect pests in agriculture (7). The facultative symbiont “Candidatus Hamiltonella defensa” (hereafter Hamiltonella) is localized in bacteriocytes, together with the obligate symbiont “Candidatus Portiera aleyrodidarum” (hereafter Portiera), representing a typical feature of whitefly symbiosis, and Hamiltonella has a high prevalence in B. tabaci MEAM1 and MED (5, 8, 9). Hamiltonella induces a female biased sex ratio in B. tabaci MEAM1 by facilitating fertilization and provisioning of five B vitamins (5). Folate derived from Hamiltonella regulates host histone methylation modifications, which subsequently influences ATP levels in the ovaries and ultimately determines the host sex ratio (10). However, the molecular and cellular mechanisms by which Hamiltonella manipulates the sex ratio of B. tabaci are still largely uncharacterized.

During oogenesis, maternally derived protein/mRNA (e.g. Oskar, Vasa, Aubergine, Tudor) localize to the posterior pole for the formation of the germ plasm which possess the ability to induce primordial germ cells (11, 12). Intracellular symbionts prefer to colonize germ cells during vertical transmission (13), so that they may have evolved a mechanism to target maternal protein/mRNA of the host during development (14). Indeed, some Wolbachia strains concentrate at the posterior of host oocytes, promoting Wolbachia incorporation into posterior germ cells during embryogenesis (14, 15). The maternal genes vasa and tudor (tud) impact posterior localization of Wolbachia in eggs of Drosophila (14). Additionally, Wolbachia protein TomO targets nanos mRNA and restores germ stem cells in Drosophila Sex-lethal (Sxl) mutants (16). These studies indicate that Wolbachia can interact with maternal proteins and mRNAs for its own propagation (13). However, it remains unclear whether symbionts can influence insect reproduction by regulating maternal proteins and mRNAs.

In Drosophila and mammal germlines, Tud could interact with P-element-induced wimpy testis (PIWI) proteins such as Aubergine (Aub) or Argonaute 3 (Ago3), thereby controlling the amplification and quality of PIWI-interacting RNAs (piRNAs) (17, 18). piRNAs are a class of small noncoding RNAs ranging from 23 to 30 nucleotides (18, 19). The early study has shown that the function of the PIWI–piRNA complex is to silence transposable elements (TEs) (18). However, subsequent studies have revealed that many piRNA sequences originate from non-TE regions, suggesting that the role of piRNAs extends beyond silencing TEs. Indeed, recent reports have demonstrated that piRNAs regulate protein-coding gene expression through incomplete base-pairing in both germ cells and somatic tissues (20, 21). In Bombyx mori, silencing Masc by Fem-related piRNA is essential for generating female-specific isoforms of Bmdsx in female embryos, thereby impacting silkworm sex determination (22). Similarly, in the Drosophila melanogaster and Locusta migratoria, piRNAs regulate the expression of certain reproduction-related mRNAs, thus regulating germline development and fecundity (19, 23). However, it is unclear whether symbionts can influence insect reproduction by regulating the Tud-mediated piRNA pathway.

The bacteriocyte symbionts in whiteflies have the common feature of posterior localization in oogenesis and embryogenesis, similar to other maternally inherited symbionts (24, 25). Therefore, we investigated whether Hamiltonella can regulate maternal genes and associated piRNAs, thereby impacting whitefly reproduction. We found that Hamiltonella infection regulates maternal protein Tud level in whitefly ovaries. Then, we demonstrate that Hamiltonella enhances whitefly fertilization by influencing the abundance of vacuolar (H+)-ATPase H subunit (VATPH)-related piRNAs and the expression of VATPH through altering tud levels. We further showed that the promotion of ATP synthesis in ovaries and stabilization of F-actin in eggs, both controlled by Hamiltonella ultimately facilitate whitefly fertilization.

Results

Hamiltonella Enhances the Expression of tud in Whitefly Ovaries.

To investigate whether Hamiltonella can influence the expression of maternal genes, Hamiltonella was specifically eliminated by treating B. tabaci adults with an antibiotic cocktail. After F0 whiteflies were treated with antibiotics, the Hamiltonella titer was reduced by 90.6% in Hamiltonella-cured (–HBt) whiteflies compared to Hamiltonella-infected (+HBt) whiteflies (Fig. 1A; P < 0.0001; df = 18), while the titer of Portiera and Rickettsia remained similar (Fig. 1A; P = 0.13 for Portiera and P = 0.88 for Rickettsia; df = 18). Then, mRNA sequencing was performed on the ovaries of +HBt and −HBt female whiteflies with the same genetic background. We found 51 whitefly mRNAs were up-regulated and 55 whitefly mRNAs were down-regulated in the ovaries of −HBt whiteflies compared with +HBt whiteflies (Dataset S1A). Differentially expressed genes between −HBt whiteflies and +HBt whiteflies were visualized using a volcano plot (Fig. 1B). Among these differentially expressed genes, the maternal gene tud exhibited the highest expression level in +HBt whiteflies, and its expression was significantly down-regulated in the ovaries of −HBt whiteflies compared to +HBt whiteflies (Dataset S1A; Fig. 1C; Uncorrected P-value = 0.0086, while FDR-corrected value = 0.9998). We previously demonstrated that Hamiltonella deficiency altered histone H3 lysine 9 trimethylation (H3K9me3) level. Genome-wide chromatin immunoprecipitation-seq analysis of H3K9me3 indicated that Tud was associated with differential H3K9me3 peaks in –HBt whiteflies compared to +HBt whiteflies (dataset S4D in ref. 10; Dataset S1B), indicating that Hamiltonella may impact expression levels of tud in whiteflies by changing the level of H3K9me3. qRT-PCR verified that the expression of tud was significantly decreased in the ovaries of −HBt whiteflies compared to +HBt whiteflies at 4 and 8 d after emergence (Fig. 1D; P = 0.021 and P = 0.042, respectively; df = 4). To examine tud gene expression profiles at various developmental stages of whiteflies or in reproductive tissues, nymphs, and adult females of +HBt whiteflies at various days after emergence were collected and ovaries were dissected. Compared to nymphs, the expression of tud was significantly increased in adult females at 2 and 4 d postemergence, followed by a decline observed between 6 to 12 d after emergence (Fig. 1E; P = 0.0096 for 2 d, P = 0.0379 for 4 d and P = 0.0485 for 12 d; df = 14). Moreover, tud exhibits higher expression levels in the ovaries compared to the whole body of +HBt whiteflies (Fig. 1F; P = 0.00011; df = 4).

Fig. 1.

The expression profiles of tud in whiteflies. (A) Relative abundance of Hamiltonella, Portiera, and Rickettsia in female adults of +HBt and –HBt whiteflies. n = 10. (B) A volcano plot illustrating the differentially expressed mRNAs in ovaries of −HBt and +HBt whiteflies. (C) The FPKM values of tud in –HBt compared to +HBt whitefly ovaries. n = 3. (D) Relative expression of tud in ovaries of +HBt and –HBt whiteflies. n = 3. (E) The expression level of tud in second and third instar nymphs, and adult females of whiteflies 2, 4, 6, 8, 10, and 12 d after emergence. Nymphs were used as reference point when the statistical comparison was made. n = 3. (F) Expression of tud in ovaries compared to the whole body of adult female whiteflies 4 d after emergence. n = 3. (G) Western blot of Tud protein level of +HBt and −HBt whiteflies. We consolidated images of Western blot for +HBt and −HBt whiteflies from the same gel to present a single representative Western blot image. n = 3. (H and I) Localization of Tud (red) and Hamiltonella (green) in ovarioles during vitellogenesis phases III (H) and IV (I) of +HBt and –HBt female adult whiteflies. DNA was stained by DAPI (blue). All images are representative of three biological replicates. The +HBt and −HBt represent Hamiltonella-infected and Hamiltonella-cured whiteflies 7 d after emergence. Data shown are mean ± SEM. The significant differences between treatments are indicated by asterisks (*P < 0.05; **P < 0.01; ***P < 0.001; one-way ANOVA).

To test whether Hamiltonella influenced Tud protein levels, we examined the level of Tud in whiteflies and the subcellular location of Tud in whitefly ovaries after Hamiltonella elimination. A polyclonal antibody against the Tud protein was produced using the purified recombinant protein. The polyclonal antibody against Tud had good specificity, which was verified by western blot (SI Appendix, Fig. S1 A and B) and Label-free quantitative LC/MS proteomics (Dataset S2). The Tud level was decreased in −HBt whiteflies compared with +HBt whiteflies at 4 to 8 d after emergence (Fig. 1G). Immunofluorescence microscopy revealed that Tud is primarily located in the follicle cells and nurse cells of ovarioles at vitellogenesis-phase III (Fig. 1H), and in the posterior region of ovarioles at vitellogenesis-phase IV (Fig. 1I). The localization of Tud in the posterior pole of ovarioles at vitellogenesis-phase IV of whiteflies represents a conserved feature of Tud localization during oogenesis among insects (11, 12). There was no signal of Tud in whitefly ovarioles of negative controls (SI Appendix, Fig. S1C). Hamiltonella was only localized within bacteriocytes but not in the germline or within oocyte nuclei in the ovarioles at vitellogenesis-phase III and IV of whiteflies (Fig. 1 H and I) as also shown previously (5, 25). The posterior localization of the Tud protein is coupled with the posterior localization of bacteriocytes and the associated symbiont Hamiltonella in the ovarioles at the late stage of whitefly oogenesis (24, 25; Fig. 1I). After Hamiltonella was cured, Tud localization noticeably declined in the ovarioles at vitellogenesis-phase III and IV of whiteflies (Fig. 1 H and I).

Hamiltonella Affects the Abundance of VATPH-Related piRNA by Altering tud Level.

A phylogenetic tree analysis showed that the Tud of B. tabaci MEAM1 and other whitefly species clustered within the same clade (SI Appendix, Fig. S2). Then, the domain of Tud was compared among B. tabaci MEAM1, MED, and SSA1. Tud of B. tabaci MEAM1, MED, and SSA1 contains 7, 9, and 10 Tud domains, respectively (Fig. 2A). Tud domain proteins that associate with the symmetrical dimethylarginine modifications of PIWI proteins regulate the amplification and function of piRNA (18). Concurrently, Wolbachia is capable of maintaining the abundance of piRNAs in Aedes aegypti cell lines (26). Thus, we investigated whether Hamiltonella could affect the abundance and function of piRNAs by altering tud levels. Subsequently, we quantified the number of piRNAs present in the ovaries of +HBt and −HBt whiteflies. We found 125,505 piRNAs (Dataset S3A). The 5′ end of piRNA sequences have a strong uracil bias. Therefore, we analyzed the base bias of all piRNAs aligned to the genome. Notably, both sense (Fig. 2B) and antisense (Fig. 2C) sequences exhibit a pronounced preference for uracil as their first base. Differentially expressed B. tabaci piRNAs were analyzed in the ovaries of +HBt and −HBt whiteflies. Among these, 697 whitefly piRNAs were up-regulated, and 220 whitefly piRNAs were down-regulated in −HBt whiteflies compared with +HBt whiteflies (Dataset S3B). Differentially expressed piRNAs between –HBt whiteflies and +HBt whiteflies were visualized using a volcano plot (Fig. 2D).

Fig. 2.

Hamiltonella affects the abundance of VATPH-related piRNA by altering tud level. (A) Conserved domains of Tud in B. tabaci MEAM1, MED, and SSA1. (B and C) The sense (B) and antisense (C) sequence first base bias of piRNAs aligned to the genome. (D) A volcano plot illustrating the differentially expressed piRNAs in ovaries of –HBt and +HBt whiteflies. (E) Cartoon representation of the tripartite complex involving Tud (blue), Aub (green), and piR-t00104691 (red) by AlphaFold3. (F) The hydrogen bonds at the interface of the Tud (blue), Aub (green), and piR-t00104691 (red) complex interaction. (G) Relative expression of piR-t00104691 and VATPH in –HBt compared to +HBt whitefly ovaries. n = 3. (H) Relative expression of piR-t00104691 and VATPH in dsGFP-injected and dstud-injected adult female whitefly ovaries. n = 3. (I) Expression of VATPH in piR-t00104691 mimic-injected whitefly ovaries compared with the negative control at day 3. n = 3. Data are means ± SEM. Significant differences between treatments are indicated by asterisks (*P < 0.05; **P < 0.01; one-way ANOVA).

Maternally derived proteins or piRNAs in the egg could affect sex determination in the developing embryo (22, 27). Our previous work showed that Hamiltonella deficiency results in a male-biased sex ratio by inhibiting whitefly fertilization (5). But, whether Hamiltonella hijacks the maternal protein Tud to regulate piRNA function for whitefly fertilization is unknown. To investigate which piRNA could regulate whitefly fertilization, we analyzed the function of genes targeted by differential expressed piRNAs (Dataset S4). Among them, only vacuolar (H+)-ATPases H subunit (VATPH) is directly related to animal fertilization. VATPH is essential for vacuolar (H+)-ATPases (V-ATPases) activity, which is necessary for fertilization of Caenorhabditis elegans (28, 29). We investigated whether Hamiltonella can adjust the expression of VATPH-related piRNA and VATPH through the maternal gene tud. We found that a piR-t00104691 specifically binds to the 3′UTR of VATPH, as predicted by bioinformatic predictions (Dataset S4). Previous study by Ultraviolet Cross-linking and Immunoprecipitation with High-throughput Sequencing (CLIPSeq) shows that Drosophila Tud interacts with Aubergine (Aub) and piRNAs, forming tripartite complex, thereby to control the amplification of piRNAs in germ plasm (30, 31). The predicted template modeling (pTM) score and the interface predicted template modeling (ipTM) score are critical metrics for evaluating biomolecular interactions using AlphaFold3. A combined pTM + ipTM score exceeding 0.5 indicates high reliability of both overall structure and interface predictions, suggesting a strong likelihood of interaction (32, 33). Our analysis by AlphaFold3 revealed that the Tud–Aub–piRt00104691 complex can also be formed (Fig. 2 E and F; pTM+ipTM = 0.68), indicating that Hamiltonella could affect the abundance of piRNA by altering tud level. qRT-PCR verified the up-regulated expression of piR-t00104691 in the ovaries of −HBt whiteflies compared to +HBt whiteflies (Fig. 2G; P = 0.045; df = 4). We further verified that, after Hamiltonella was cured, VATPH expression significantly decreased in whitefly ovaries (Fig. 2G; P = 0.019; df = 4). Likewise, piR-t00104691 was significantly increased and VATPH was significantly decreased in the ovaries of dstud-injected whiteflies compared to dsGFP-injected whiteflies (Fig. 2H; P = 0.0095 for piR-t00104691; P = 0.012 for VATPH; df = 4). Thus, these data demonstrate that Hamiltonella affects the abundance and function of piR-t00104691 by altering tud levels. To confirm the regulation of VATPH expression by B. tabaci piRNA, female adult whiteflies were injected with piR-t00104691 mimic, and VATPH expression in ovaries was examined by qRT-PCR. The expression of VATPH was significantly reduced in the ovaries of whiteflies injected with piR-t00104691 mimic at 3 d postinjection, compared to the negative control (NC) (Fig. 2I; P = 0.014; df = 4). Collectively, these results demonstrate that piR-t00104691 negatively regulates VATPH expression in whitefly ovaries.

Hamiltonella Regulates Whitefly Fertilization Via Impacting VATPH-Related piRNA and VATPH Expression.

V-ATPase plays a crucial role in maintaining the homeostasis of intracellular energy (namely, ATP) and supporting mitochondrial respiration (34, 35). To further test the role of VATPH in whitefly fertilization, we then investigated whether VATPH can regulate whitefly fertilization by impacting ATP levels in ovaries. Adult females were microinjected with dsVATPH, and the level of ATP in whitefly ovaries was examined. The expression of VATPH in adult females was significantly reduced by 26% and 32%, at 3 and 5 d after RNA interference (RNAi) treatment, respectively (Fig. 3A; P = 0.0014 for 3 d and P = 0.0176 for 5 d; df = 4). We observed that 52.32% and 69.2% of the 1-h eggs (eggs within 1 h of deposition) exhibited malformations at 3 and 5 d following dsVATPH injection, respectively (Fig. 3 B and C). In addition, malformed eggs are unable to develop into nymphs (Fig. 3C). In parallel with the reduced VATPH expression at 3 d postinjection, the ATP level was significantly reduced by 21.28% in the ovaries of whiteflies injected with dsVATPH compared to those injected with dsGFP (Fig. 3D; P = 0.017; df = 8). The fertilization rate of nonmalformed eggs was reduced, on average, by 52.6% in dsVATPH-injected whiteflies compared to dsGFP-injected whiteflies (Fig. 3E; P = 0.0002; df = 6). Furthermore, the ATP level was significantly reduced by 22.25% in the ovaries of whiteflies injected with piRNA mimic compared to those injected with NC (Fig. 3F; P = 0.016; df = 10). The fertilization rate was reduced, on average, by 46.65% in mimic-injected whiteflies compared to NC injected whiteflies (Fig. 3G; P = 0.0001; df = 6). These results demonstrate that VATPH regulates whitefly fertilization by impacting ATP levels in ovaries.

Fig. 3.

Hamiltonella regulates whitefly fertilization by impacting VATPH-related piRNA. (A) Expression of VATPH in adult female whiteflies injected with dsGFP and dsVATPH after 3 and 5 d postmicroinjection of dsRNAs. n = 3. (B) The malformation rate of eggs within 1 h postovulation (n = 3 with 17 to 23 eggs in each replicate). (C) Effect of VATPH silencing on eggs detached from cotton leaves or eggs on cotton leaves within 1 h postovulation as well as the hatching of eggs on cotton leaves at 12 d postovulation. The whitefly samples were imaged using an Extended Depth of Field camera (Keyence VHX-7000). Red arrow denotes the eggshell, and blue arrow denotes the nymphs. (D) The level of ATP in the ovaries of dsGFP-injected and dsVATPH-injected whiteflies. n = 5. (E) The fertilization rate of dsGFP-injected♀×dsGFP-injected♂ and dsVATPH-injected♀×dsGFP-injected♂ whiteflies (n = 4 with 18 to 24 eggs for dsGFP-injected female whiteflies and n = 4 with 11 to 14 eggs for dsVATPH-injected female whiteflies in each replicate). (F) The level of ATP in piR-t00104691 mimic-injected whitefly ovaries compared with the negative control at day 3. n = 6. (G) Fertilization rate of negative control (NC)-injected♀×NC-injected♂ and piRNA mimic-injected♀×NC-injected♂. (n = 4 with 17 to 23 eggs for NC-injected female whiteflies and n = 4 with 16 to 21 eggs for piRNA mimic-injected female whiteflies in each replicate). Data are means ± SEM. Significant differences between treatments are indicated by asterisks (*P < 0.05; **P < 0.01; ***P < 0.001; one-way ANOVA).

The ATP Levels Influence F-Actin Patterns in Whitefly Ovaries.

F-actin polymerization is an ATP-dependent process characterized by the incorporation of ATP-bound actin monomers at the barbed end of F-actin (36, 37). Thus, we investigated whether the repression of ATP levels modulates the F-actin patterns in the 1-h eggs of whiteflies. The distribution of the F-actin patterns in these eggs was assessed after adult females were microinjected with dsVATPH and piRNA mimic. Notably, a decrease in F-actin signal was observed in the 1-h eggs at 3 d postmicroinjection with dsVATPH and piRNA mimic (Fig. 4 A and B). Subsequently, to examine whether ATP supplementation can recover F-actin patterns, dsVATPH-injected and piRNA mimic-injected whiteflies were fed an artificial diet supplemented with ATP (5 μM) for 3 d. Quantification of fluorescence intensity reveals that F-actin patterns in the 1-h eggs of whiteflies injected with dsVATPH (Fig. 4 A and C) and piRNA mimic (Fig. 4 B and D) were restored following ATP supplementation. Collectively, these results indicate that elevated ATP levels may enhance the stability of F-actin in whitefly ovaries.

Fig. 4.

Repressing ATP levels influences F-actin patterns in eggs. (A) Localization of F-actin (green) in 1-h eggs (eggs within 1 h of deposition) of dsGFP-injected and dsVATPH-injected adult female whiteflies feeding on artificial diet supplemented with or without ATP for 3 d. (B) Localization of F-actin (green) in 1-h eggs of NC-injected and piRNA mimic-injected adult female whiteflies feeding on artificial diet supplemented with or without ATP for 3 d. DNA was stained by DAPI (blue). All images are representative of the three replicates. (C) Fluorescence intensity of F-actin in the 1-h eggs of dsGFP-injected and dsVATPH-injected adult female whiteflies feeding on artificial diet supplemented with or without ATP. (D) Fluorescence intensity of F-actin in the 1-h eggs of NC-injected and piRNA mimic-injected adult female whiteflies feeding on artificial diet supplemented with or without ATP.

Disrupting the F-Actin Cytoskeleton Inhibits Whitefly Fertilization.

The cortex is the site of attachment and fusion of sperm and egg (38). Regulation of cortical F-actin is involved in nearly all major aspects of fertilization (39, 40). Our previous study demonstrated that Hamiltonella deficiency impairs egg fertilization (5). Thus, we investigated the role of Hamiltonella in modulating the distribution of F-actin cytoskeleton in 1-h eggs of whiteflies. F-actin signal declined in whitefly eggs within 1 h of deposition after Hamiltonella elimination (Fig. 5A). Then, we ask whether F-actin disruption can impact fertilization in whiteflies. To test that, adult females were microinjected with F-actin inhibitor Cytochalasin B and the fertilization rate was observed. Likewise, F-actin signal declined in the 1-h eggs of whiteflies at 1 d after Cytochalasin B treatment (Fig. 5B). Few sperm pronuclei were observed in the eggs of whiteflies at 1 d after treatment with cytoskeletal inhibitor (Fig. 5 C and D). The fertilization rate was reduced, on average, by 51.43% in Cytochalasin B-treated adult females compared to control water-injected adult females (Fig. 5 C and D; P = 0.0079; df = 4).

Fig. 5.

Disrupting F-actin cytoskeleton inhibits whitefly fertilization. (A) Localization of F-actin (green) in 1-h eggs of +HBt and –HBt female adult whiteflies. (B) Localization of F-actin (green) in 1-h eggs of distilled water-injected (CK) and Cytochalasin B-injected adult female whiteflies. (C) Egg pronuclei and sperm pronuclei in the egg within 1 h after deposition in CK (distilled water-injected) and Cytochalasin B-injected adult female whiteflies. The white arrow denotes the sperm pronuclei, red arrow denotes the egg pronuclei, and green arrow denotes the bacteriocyte nucleus. (D) Fertilization rate of distilled water-injected♀×distilled water-injected♂ and Cytochalasin B-injected♀×distilled water-injected♂. n = 3. DNA was stained by DAPI (blue). All of the images are representative of three replicates. Data are means ± SEM. Significant differences between treatments are indicated by asterisks (**P < 0.01; one-way ANOVA).

tud Regulates Whitefly Fertilization by Impacting the ATP Level in Ovaries and F-Actin Cytoskeleton in Eggs.

To further validate whether tud impacts whitefly fertilization by influencing ATP levels in ovaries and F-actin cytoskeleton in eggs of whiteflies, adult females were microinjected with dstud. The expression of tud in adult females was significantly reduced at 1 and 3 d after RNAi treatment (Fig. 6A; P = 0.009 for 1 d and P = 0.0052 for 3 d; df = 4). Tud localization declined in ovarioles at vitellogenesis-phase IV 1 d after silencing the whitefly tud gene (Fig. 6B). The level of ATP was significantly reduced in the ovaries of dstud-injected whiteflies compared to dsGFP-injected whiteflies (Fig. 6C; P = 0.0047; df = 8). At 1 d after tud silencing, F-actin signal was declined in eggs within 1 h of deposition (Fig. 6D). Furthermore, the sperm pronuclei was frequently observed in eggs of dsGFP-injected whiteflies (Fig. 6 E and F). In contrast, few sperm pronuclei were observed in eggs of whiteflies at 1 d after silencing the whitefly tud gene (Fig. 6 E and F). The fertilization rate was reduced, on average, by 48.7% in dstud-injected adult females compared to dsGFP-injected adult females (Fig. 6E; P < 0.0001; df = 4).

Fig. 6.

tud regulates whitefly fertilization by impacting ATP level in ovaries and F-actin cytoskeleton in eggs. (A) Expression of tud in dsGFP-injected and dstud-injected adult female whiteflies 1 and 3 d after whiteflies were microinjected with dsRNAs. n = 3. (B) Localization of Tud (red) in ovarioles at phase IV of dsGFP-injected and dstud-injected female adult whiteflies 1 d after microinjection with dsRNAs. (C) The level of ATP in ovaries of dsGFP-injected and dstud-injected adult female whiteflies. n = 5. (D) Localization of F-actin (green) in 1-h eggs of dsGFP-injected and dstud-injected adult female whiteflies. (E) Fertilization rate of dsGFP-injected♀×dsGFP-injected♂ and dstud-injected♀×dsGFP-injected♂. n = 3. (F) Egg pronuclei and sperm pronuclei in the egg within 1 h of deposition in dsGFP-injected and dstud-injected adult female whiteflies. (G) Schematic overview of how the symbiont Hamiltonella determines whitefly sex ratio by regulating Tud expression. Expression of the maternal gene tudor (tud) and its encoded protein Tud is down regulated in whitefly ovaries after Hamiltonella elimination. tud is highly expressed in whitefly ovaries. In addition, tud silencing impairs ATP levels by impacting VATPH-related piRNA and VATPH expression in whitefly ovaries. Repressing ATP levels leads to decreased F-actin signal in whitefly ovaries, and disrupting F-actin inhibits whitefly fertilization. tud regulates whitefly fertilization by impacting ATP level in ovaries and F-actin cytoskeleton in eggs. Thus, the maternally inherited symbiont regulates whitefly sex ratio by hijacking maternal protein Tud. This figure was drawn by Figdraw. The white arrow denotes the sperm pronuclei, red arrow denotes the egg pronuclei, and green arrow denotes the bacteriocyte nucleus. DNA was stained by DAPI (blue). All images are representative of three replicates. Data are means ± SEM. Significant differences between treatments are indicated by asterisks (**P < 0.01; ***P < 0.001; one-way ANOVA).

Discussion

The majority of intracellular symbionts colonize germ cells during vertical transmission (13). The maternal protein/mRNA (germline determinants) may also affect sex determination in the developing zygote (27). piRNAs are highly expressed within animal germ cells and play significant roles in gene expression regulation, germ cell maturation, and sex determination (22, 41). However, whether intracellular symbionts affect host reproduction by manipulating germline determinants or piRNAs remains unknown. In this study, we found that Hamiltonella elimination reduced expression of the maternal protein Tud in whitefly ovaries. Furthermore, Hamiltonella inhibited the abundance of a piRNA and induced VATPH expression through maternal gene tud, thereby promoting ATP synthesis in ovaries and the stability of F-actin in eggs, ultimately facilitating whitefly fertilization. In addition, repressing expression of tud inhibited whitefly fertilization by impacting the ATP level in ovaries and F-actin cytoskeleton in eggs (Fig. 6G). We present the function of piRNA in a bacterial symbiont–host interaction. This study reveals the interplay between bacteriocyte symbionts and germline proteins can impact host reproduction.

Intracellular symbionts could target maternal proteins/mRNAs of the host during development (14). The TomO protein of Wolbachia directly targets nanos mRNA in Drosophila Sxl mutants (16). Beyond Wolbachia, whether other symbionts can regulate maternal proteins/mRNAs has been unclear. Our previous work indicated that Tud was associated with differential H3K9me3 peaks in –HBt whiteflies compared to +HBt whiteflies (10). Here, we found that tud and its encoded protein have lower expression levels in whitefly ovaries after Hamiltonella elimination. Therefore, Hamiltonella may impact expression levels of tud in whiteflies by changing the level of H3K9me3. Interestingly, Tud family proteins can serve as versatile effectors of histone methylation (42, 43). For example, UHRF1 Tud functions as a “reader” of H3K9me3 (43). It is speculated that histone methylation modification caused by Hamiltonella may also depend on the regulation of the expression of whitefly tud. Thus, Hamiltonella could lead to the coregulation of histone methylation and tud that can interact with each other in whiteflies. The mechanisms of how other intracellular symbionts manipulate maternal proteins or mRNAs requires future investigation.

Although the expression of the maternal protein Tud in whitefly ovaries was reduced after Hamiltonella elimination by antibiotic treatment, the experimental setup is not suited to fully exclude a direct effect of antibiotic treatment on Tud expression, as opposed to an impact mediated by Hamiltonella. This represents a common challenge in research on symbiotic interactions, where stable separation of symbionts is often unattainable due to host lethality or the inability to maintain symbiont-free hosts as well as limitations in symbiont rearing and/or rescue experiments following antibiotic treatments.

Hamiltonella impacts whitefly sex ratio by regulating the ATP level in ovaries (10). However, the underlying cellular mechanism is unclear. The germ plasm is associated with cortical cytoskeleton in the egg (44). Fertilization occurs at the site of the cortex in the egg when the sperm attaches to and fuses with the egg (38). Remodeling of the cortical F-actin distribution is responsible for nearly all major aspects of fertilization (39, 40). ATP regulates the stability of F-actin (36). Both silencing tud and Hamiltonella elimination reduced F-actin signal in ovarioles and eggs of whiteflies. Microinjection with F-actin inhibitor Cytochalasin B also inhibited the fertilization rate in whiteflies. This suggested Hamiltonella regulates ATP level, thereby impacting the stability of F-actin in ovarioles and eggs of whiteflies, which finally influences whitefly fertilization and sex ratio.

Maternally transmitted genetic elements include bacterial symbionts and mitochondria in cytoplasm and maternal protein and piRNAs in the nucleus. Conflict between cytoplasmic and nuclear genes over sex determination and sex ratios is obvious and widespread (27). Many cytoplasmic sex-ratio distorters are microorganisms that are transmitted through the egg cytoplasm (27). Among them, Hamiltonella impacts ovary mitochondrial function for ATP synthesis, thereby regulating whitefly sex ratio (10). There are also complex interactions between bacterial symbionts and germ plasm within the germline (45). We demonstrate that the elimination of Hamiltonella decreases the expression of the maternal gene tud, thereby inhibiting fertilization. Thus, we reveal the symbiont–maternal protein associations can regulate host sex ratio. The linkage between these maternally transmitted genetic elements can lead to cytonuclear incompatibilities, affecting the persistence of the distorter (46). Our previous work (10) and this study reveal that the interaction among multiple maternally transmitted genetic elements—bacterial symbionts, mitochondria, maternal protein, and piRNAs—impacts host reproduction by jointly influencing ATP level. It would be useful to examine the relationship among these maternally transmitted genetic elements in other insect symbioses and determine how their interactions impact insect reproduction. These data may facilitate our understanding of the mechanisms by which intracellular symbionts regulate host reproduction and complex coevolutionary interactions between cytoplasmic sex-ratio distorter and nuclear genes.

Materials and Methods

Insect Rearing and Plants.

The whitefly B. tabaci MEAM1 colony (mtCO1 GenBank Accession No. GQ332577) was maintained on cotton plants (Gossypium hirsutum, cv. Shiyuan 321) as previously described (5, 10). Cotton plants were cultivated to the 6 to 7 true-leaf stage for use in experiments.

Elimination of Hamiltonella by Antibiotic Treatment.

To specifically eliminate Hamiltonella, hundreds of adult whiteflies of B. tabaci (F0, 0 to 7 d after emergence) were released into each feeding chamber and fed on 25% sucrose solution (w/v) supplemented with the antibiotics ampicillin, gentamycin, and cefotaxime (BBI Life Sciences, Shanghai, China), following previously described protocols (5, 9, 10, 47; SI Appendix, Methods). The F1 B. tabaci with reduced Hamiltonella titers (−HBt), which were obtained by antibiotic treatment, and control F1 B. tabaci (+HBt), which were obtained by feeding sucrose solution not supplemented with antibiotics, were identified.

Sequencing Analyses of mRNAs and Small RNAs.

Ovaries were dissected from 400 to 600 +HBt and −HBt whiteflies for each of three biological replicates at 4 to 10 d after emergence. Total RNAs were extracted. Sequencing analyses of mRNAs and small RNAs were conducted according to the protocols described in refs. 6 and 48, with details in SI Appendix, Methods.

The edgeR package was used to identify differentially expressed genes across samples with the absolute value of log₂ ratio ≥ 1 and a P-value < 0.05. These genes were then used to generate volcano plots.

The putative piRNA with the length of 24 to 33 nt was further matched against piRBase (49) to identify existing piRNAs. In addition, the remaining putative piRNAs were identified by its Ping-Pong structure. The edgeR package was utilized for the analysis of differentially expressed piRNAs, characterized by log₂ ratio ≥ 2 and a P-value < 0.05. These piRNAs were then used to generate volcano plots. Notably, the 5′ end of the piRNA sequence exhibits a strong uracil bias. Consequently, we analyzed the nucleotide bias of all piRNAs aligned to the whitefly genome and performed statistical analysis on the first base bias of piRNA tag sequences across varying lengths, as well as on the base bias at each tag position.

Prediction of Genes Targeted by piRNAs.

piRNAs engage in complementary base pairing with their target genes. Consequently, a gene must have perfect or near-perfect complementary base pairing to be recognized as a piRNA target (50). Based on this premise, the piRNA target sites and target genes for each sample were predicted using BLAST following the previously described protocol (51, 52). Specifically, the blastn software was employed to map piRNA sequences against the whitefly genome (53). Only perfectly complementary matches (allowing up to three mismatches) with an E-value threshold of less than 1E⁻⁵ were retained. Subsequently, each target sequence that could be targeted by piRNAs was scored.

qPCR and qRT-PCR.

Total DNA was extracted, and symbiont DNA was quantified by qPCR as previously described (10, 24; SI Appendix, Methods). Total RNA was extracted, and cDNAs were synthesized. The expression of mRNAs was quantified by qRT-PCR as previously described (10; SI Appendix, Methods). The expression of piRNAs was quantified by qRT-PCR following a previously described protocol (48; SI Appendix, Methods). Two technical replicates were performed for each biological replicate. All of the primers used in this study are shown in SI Appendix. Relative symbiont density and gene expression were calculated using the 2^−ΔCt method (54).

Recombinant Enzyme Generation and Antibody Preparation.

The cDNA sequences were chemically synthesized with optimization for Escherichia coli expression following the previously described protocol (55). The cDNA was cloned in expression vector PET28b (Cloning strategy: NdeI/EcoRI). The recombinant enzyme for whitefly Tud was generated as described previously (56). Custom-made polyclonal antibodies against Tud (predicted size, 314 AAs, 34.99 kDa, Leu803-Tyr1095) protein were produced by Atagenix Laboratories (Wuhan, China) following a previously described method (56).

Label-Free LC/MS Analysis.

To assess the specificity of the Tud polyclonal antibody in whiteflies, proteins were extracted from 50 adult female whiteflies and separated by 12% SDS-PAGE. Proteins around 230.13 kDa verified by Western blot analysis were hydrolyzed into peptide segments, and analyzed by UPLC–MS/MS with a nanoElute UHPLC system (Bruker, Germany) coupled to a timsTOF Pro2 mass spectrometer (Bruker, Germany) at Novogene Co., Ltd. (Beijing, China), following a previously described protocol (57). All resulting spectra were searched against the B. tabaci MEAM1 proteomics database using search engines like Proteome Discoverer (Thermo, HFX and 480) or MaxQuant (Bruker, Tims).

Western Blot Analysis.

Proteins in +HBt and −HBt adult female whiteflies at 4 to 8 d after emergence were extracted, separated by 12% SDS-PAGE, and then blotted with primary antibodies specific against Tud and β-actin following standard procedures using a previously described protocol (56; SI Appendix, Methods). The densitometry for protein levels was analyzed using ImageJ software.

Fluorescence In Situ Hybridization (FISH) Analysis.

To localize Hamiltonella in the ovarioles and eggs of female adult whiteflies, FISH was conducted following a previously described protocol (5, 10). Samples were stained with DAPI (1 µg/mL in PBS, Sigma) for 30 min at room temperature. Three biological replicates were conducted. Images were collected and analyzed on a FV3000 confocal microscope (Olympus, Tokyo, Japan).

Immunofluorescence Microscopy.

Ovarioles from +HBt and −HBt adult female whiteflies or dsRNAs-injected adult female whiteflies were collected, fixed, permeabilized, and incubated with Alexa-Fluor 594 labeled anti-Tud antibodies following a previously described protocol (56). Three biological replicates were conducted. The samples were incubated with no antibody against Tud as the negative control. Images were captured and analyzed using a FV3000 confocal microscope (Olympus, Tokyo, Japan).

Phylogenetic Tree and Protein Domain Analysis.

To determine the homologous genes in other whitefly species, the sequences of tud (Accession No. OQ301635) in B. tabaci MEAM1 were subjected to TBLASTX against the genome of B. tabaci MED, SSA-ECA, and the greenhouse whitefly Trialeurodes vaporariorum (GenBank Accession No.: GCA_003994315.1, GCA_004919745.1, and GCA_011764245.1). The top TBLASTX hits were obtained. To construct the molecular phylogenetic tree for whitefly Tud, a Bayesian inference (BI) analysis was conducted as described previously (56; SI Appendix, Methods).

To analyze the domain of Tud in B. tabaci MEAM1. Batch CD-Search tool (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi) was applied. Then, Tbtools (https://github.com/CJ-Chen/TBtools/releases) were used to compare the domains of Tud in B. tabaci MEAM1, MED, and SSA1 (58).

Analysis of tud Expression in Whiteflies.

The +HBt and −HBt whiteflies at 4 and 8 d after emergence were collected as described above and the ovaries were dissected. Additionally, the 2nd to 3rd instar nymphs of the +HBt whiteflies and adult females of the +HBt whiteflies at 2, 4, 6, 8, 10, and 12 d after emergence were collected. Total RNA was extracted from ovaries and whole body of whiteflies with TRIzol reagent (Sigma-Aldrich, St. Louis, MO). qRT-PCR was conducted as described above. To compare the expression of tud between +HBt and −HBt whiteflies at 4 d and 8 d after emergence, ovaries from 50 whiteflies were used for each of the three biological replicates. To compare the expression of tud between the ovaries and the whole body of +HBt whiteflies at 4 d after emergence, 50 whitefly ovaries and 20 female adult whiteflies were used for each of the three biological replicates. To compare the expression of tud among the whole bodies of the 2nd to 3rd instar nymphs of +HBt whiteflies and adult females of +HBt whiteflies at 2, 4, 6, 8, 10, and 12 d after emergence, 20 female adult whiteflies were used for each of the three biological replicates.

dsRNA Preparation.

dsRNAs specific to whitefly tud (dstud), whitefly VATPH (dsVATPH), and GFP (dsGFP) with the size of 513 bp, 123 bp, and 420 bp, respectively, were synthesized using a T7 RiboMAX Express RNAi System kit (Promega, USA), following the manufacturer’s instructions as described previously (10). The dstud, dsVATPH, and dsGFP correspond to nucleotide regions of 2,470 to 2,982 bp, 90 to 212 bp, and 76 to 495 bp (starting from 5′ end) for the respective targets tud, VATPH, and GFP with the Accession Nos. of OQ301635, LOC109040496, and MN623123, respectively.

Effects of Hamiltonella on piR-t00104691 and VATPH Expression.

To investigate whether Hamiltonella influences piR-t00104691 and VATPH expression, the +HBt and −HBt whiteflies were collected as described above and the ovaries were dissected. RNA was extracted from 30 female adult whiteflies for each of three biological replicates and piR-t00104691 and VATPH expression were examined by qRT-PCR as described above.

Effects of Silencing tud on Tud Localization, piR-t00104691 Expression, and VATPH Expression.

To investigate whether silencing tud influences Tud localization, piR-t00104691 expression, and VATPH expression in whitefly ovaries, approximately 440 female adult whiteflies at 1 d after emergence were injected with 1.5 μg/μL dstud in injection buffer using an Eppendorf microinjection system (Hamburg, Germany) (6, 10, 56). After injection, the whiteflies were transferred onto cotton leaf disks and kept on 1.5% agar plates in an incubator at 26 ± 2 °C, with a 14:10 h (L:D) photoperiod and 60 to 80% RH. Control whiteflies were injected with dsGFP. The survival rate of injected whiteflies was approximately 70% for dsGFP and dstud 1 d after injection. RNA was extracted from ten female adult whiteflies for each of three biological replicates at days 1 and 3 after microinjection and tud expression was examined by qRT-PCR as described above. To examine whether silencing whitefly tud affects Tud localization in ovarioles and eggs, whiteflies were collected 1 d after microinjection, and the whitefly ovarioles were dissected out, fixed, permeabilized, and incubated with antibodies against the whitefly Tud protein as described above. Three biological replicates were conducted. The samples were incubated with no antibodies against Tud as the negative control. Images were analyzed using an FV3000 confocal microscope (Olympus, Japan). To investigate the impact of tud silencing on the expression of piR-t00104691 and VATPH, RNA was extracted from 30 female adult whiteflies ovaries for each of three biological replicates 1 d postmicroinjection. The expression levels of piR-t00104691 and VATPH were subsequently analyzed by qRT-PCR as previously described.

Effects of piRNA Mimic Injection on VATPH Expression in Whiteflies.

The piRNA mimic consists of single-stranded oligonucleotides featuring a phosphate group at the 5′ end and a 2′-O-methyl modification at the 3′ end (59). To assess whether overexpression of piR-t00104691 affects VATPH expression in whiteflies, approximately 100 female adult whiteflies, 1 d postemergence, were injected with a 5 µM mimic using an Eppendorf microinjection system (Hamburg, Germany), following a previously established protocol (48). Whiteflies injected with a 5 µM negative control mimic served as the negative control for the mimic injection experiments. The piR-t00104691 mimic and its corresponding negative control were designed and synthesized by GenePharma (Shanghai, China), with their sequences provided in SI Appendix. To investigate the effects of piRNA mimic injection on VATPH expression in ovaries, whiteflies were collected 3 d postmicroinjection, and RNA was extracted from ovaries of 30 female adult whiteflies for each of three biological replicates. The expression levels of VATPH were subsequently analyzed by qRT-PCR as previously described.

Prediction of the Tud–Aub–piRt00104691 Complex Using AlphaFold3.

The prediction of the Tud–Aub–piRt00104691 complex was conducted using AlphaFold3 (https://alphafoldserver.com/) with default parameters. The sequences used for the prediction of complexes by AlphaFold3 are shown in SI Appendix. The predicted result was visualized and edited by using PyMol (3.1.0).

ATP Measurement by UPLC–MS/MS.

To assess the effect of VATPH inhibition on ATP levels in whitefly ovaries, 50 ovaries were dissected from female adult whiteflies injected with NC, piRNA mimic, dsGFP, and dsVATPH 3 d postmicroinjection for each of six or five biological replicates. To measure effects of tud silencing on levels of ATP in whitefly ovaries, 50 ovaries were dissected from dsGFP-injected and dstud-injected female adult whiteflies at 1 d after whiteflies were microinjected with dsRNA for each of five biological replicates. ATP was subsequently extracted and quantified using UPLC–MS/MS as previously described (10).

Cytogenetics.

To investigate the effects of tud silencing, VATPH inhibition, or Cytochalasin B treatment on fertilization, whiteflies were released onto different cotton plants in the separate whitefly-proof, ventilated cage after gene silencing, piRNA mimic microinjection, or Cytochalasin B treatment. The egg fertilization rate of whiteflies was determined using a previously described cytogenetic approach (5, 10, 60; SI Appendix, Methods).

Cytoskeleton Observation.

Ovaries and 1-h eggs were collected from adult female whiteflies of the +HBt and –HBt lines within 7 d after emergence. Additionally, dsGFP-injected and dstud-injected adult females were sampled 1 d after whiteflies were microinjected with dsRNAs, while NC-injected and piRNA mimic-injected females were collected 3 d postmicroinjection. Furthermore, dsGFP-injected and dsVATPH-injected adult females were also collected 3 d postmicroinjection with dsRNAs. Similarly, distilled water-treated and Cytochalasin B-treated adult females were collected 1 d after microinjection with Cytochalasin B. The samples were fixed in 4% paraformaldehyde (PFA) overnight at 4 °C and then permeabilized with 0.1% Triton X-100 in PBS at room temperature for 2 h. After washing with PBS, the samples were incubated with Phalloidin (phalloidin-Alexa Fluor-488, dilution 1:200, Thermo Scientific) and DAPI (1 µg/mL in PBS, Sigma) at room temperature for 1 h following a previously described protocol (25). Three biological replicates were conducted. Images were collected and analyzed on a FV3000 confocal microscope (Olympus, Tokyo, Japan).

To investigate whether ATP supplementation restores the F-actin pattern, approximately 360 female adult whiteflies within 2 h after emergence were microinjected with dsVATPH or piRNA mimic. These whiteflies were fed 30% (w/v) sucrose solution supplemented with or without ATP at a final concentration of 5 µM for 3 d based on previous studies (37). The controls were dsGFP-injected whiteflies fed with 30% (w/v) sucrose solution. Then, 1-h eggs were for examination of F-actin patterns by immunofluorescence microscopy following the protocol described above.

Statistics.

For symbiont titer, gene expression level, piRNA expression level, ATP titer and fertilization rate, statistical differences were evaluated using a one-way ANOVA at a significance threshold of 0.05 level. Data in percentages were transformed by arcsine square root before analysis. All data analyses were conducted using the STATISTICA v12 software (StatSoft, Inc., Tulsa, OK).

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLS)

Dataset S02 (XLSX)

Dataset S03 (XLSX)

Dataset S04 (XLSX)

Data, Materials, and Software Availability

The raw reads for whitefly mRNAs and piRNA are available at the National Center for Biotechnology Information Short Read Archive, with accession number PRJNA827505 (61) and PRJNA827447 (62) respectively. All study data are included in the article and/or supporting information.