Vitellogenesis, including vitellogenin (Vg) production in the fat body and Vg uptake by maturing oocytes, is of great importance for the successful reproduction of adult females. The endocrinal and nutritional regulation of vitellogenesis differs distinctly in insects. Here, the complex crosstalk between juvenile hormone (JH) and the two nutrient sensors insulin/IGF signaling (IIS) and target of rapamycin complex1 (TORC1), was investigated to elucidate the molecular mechanisms of vitellogenesis regulation in the American cockroach, Periplaneta americana. Our data showed that a block of JH biosynthesis or JH action arrested vitellogenesis, in part by inhibiting the expression of doublesex (Dsx), a key transcription factor gene involved in the sex determination cascade. Depletion of IIS or TORC1 blocked both JH biosynthesis and vitellogenesis. Importantly, the JH analog methoprene, but not bovine insulin (to restore IIS) and amino acids (to restore TORC1 activity), restored vitellogenesis in the neck-ligated (IIS-, TORC1- and JH-deficient) and rapamycin-treated (TORC1- and JH-deficient) cockroaches. Combining classic physiology with modern molecular techniques, we have demonstrated that IIS and TORC1 promote vitellogenesis, mainly via inducing JH biosynthesis in the American cockroach.
Vitellogenesis, an imperative event of insect reproduction, involves vitellogenin (Vg) production in the fat body, the release of Vg into the hemolymph, and Vg uptake by maturing oocytes. During the process of Vg uptake, extensive intercellular spaces are formed among follicle cells (follicular patency), allowing the internalization of hemolymph Vg into maturing oocytes by receptor-mediated endocytosis (Raikhel, 2005; Roy et al., 2018). However, the connection between Vg production and the formation of follicular patency has not yet been determined.
Juvenile hormone (JH), a structurally unique sesquiterpenoid hormone found only in arthropods, plays prominent roles in the regulation of reproduction, metamorphosis and other biological processes (Jindra et al., 2013; Li et al., 2019a; Roy et al., 2018). JH is primarily synthesized in the insect corpora allata (CA) through 13 discrete enzymatic steps (Bellés et al., 2005). The last two steps of JH biosynthesis are catalyzed by two crucial regulatory enzymes, juvenile hormone acid methyltransferase (Jhamt) and methyl farnesoate epoxidase (Cyp15a1) (Defelipe et al., 2011; Shinoda and Itoyama, 2003). JH signal transduction involves both genomic and non-genomic actions. In the genomic action, JH binds its intracellular receptor, methoprene-tolerant (Met), to induce the expression of a JH primary-response gene Krüppel-homolog 1 (Kr-h1), which plays a central role in the JH genomic action (Jindra et al., 2013; Li et al., 2019a). Meanwhile, JH signaling might be transduced via receptor tyrosine kinase (RTK) or G protein-coupled receptor (GPCR), putative plasma membrane receptors, to act in a non-genomic manner (Davey, 2000; Liu et al., 2015; Ojani et al., 2016; Bai and Palli, 2016). JH is one of the main gonadotrophic hormones that promote vitellogenesis in insects, including Vg production in the fat body via the genomic action, as well as the development of follicular patency and thus Vg uptake by the maturing oocytes via the non-genomic action (Raikhel, 2005; Jing et al., 2018; Roy et al., 2018).
Doublesex (Dsx), a transcription factor in the last step of the sex determination cascade, participates in reproductive regulation in female insects (Verhulst and Van de Zande, 2015). Dsx was reported to induce Vg expression in the fruit fly Drosophila melanogaster (Burtis et al., 1991), the wild silkmoth Antheraea assama (Shukla and Nagaraju, 2010) and the red flour beetle Tribolium castaneum (Shukla and Palli, 2012). Moreover, a causal link between Dsx and JH signaling was revealed in the stag beetle Cyclommatus metallifer (Gotoh et al., 2014). Therefore, Dsx is probably a transcription factor involved in the JH-induced Vg expression.
Insulin/IGF signaling (IIS) and target of rapamycin complex 1 (TORC1) are highly evolutionarily conserved nutrient sensors with vital regulatory roles in developmental events and disease progression. Insulin/IGF binds the membrane insulin receptor (InR) and activates IIS by a series of protein phosphorylation events, leading to the activation of protein kinase B (Akt) (Li et al., 2019b; Nässel and Broeck, 2016; Wu and Brown, 2006). TORC1 receives IIS and amino acid signals, phosphorylates effector proteins such as S6K and 4EBP to stimulate protein synthesis, ribosome biogenesis and cell growth (Li et al., 2019b; Smykal and Raikhel, 2015). It is well known that a block of IIS or TORC1 decreases JH biosynthesis in the CA and Vg expression in the fat body in a number of insect species. However, the interplay between the nutrient sensors and JH signaling in the regulation of vitellogenesis varies dramatically in insects (Abrisqueta et al., 2014; Corona et al., 2007; Guo et al., 2014; Li et al., 2018; Maestro et al., 2009; Parthasarathy and Palli, 2011; Parthasarathy et al., 2010; Pérez-Hedo et al., 2013; Sheng et al., 2011; Song et al., 2014). Nevertheless, more solid evidence is required to demonstrate whether IIS and TORC1 induce vitellogenesis independent of JH or act indirectly through regulating JH biosynthesis.
The American cockroach Periplaneta americana is an invasive urban pest in warm and humid regions, partially due to its high fecundity. Our previous work showed that JH signaling, IIS, TORC1 and Dsx affect ovarian maturation in this insect species (Li et al., 2018). However, how these signals interact to promote vitellogenesis remains unknown. In this study, we have determined that IIS and TORC1 promote vitellogenesis mainly in an indirect manner by activating JH biosynthesis. Combining classic physiology with modern molecular techniques, our study broadens the understanding of endocrinal and nutritional regulation of female reproduction in insects.
Vg production is required for the development of functional follicular patency
Under our rearing conditions, the first reproductive cycle of adult female American cockroaches was ∼8 days, and the primary oocytes start to accumulate yolk proteins (mainly Vg) on day 4 post-adult emergence (PAE) (Li et al., 2018). Protein mass spectroscopy analysis revealed that the ovarian proteins in the 100 kDa major band were mainly the two Vg proteins, Vg1 and Vg2 (Fig. S1A,B). As this band was most abundant and no antibody against Vg was available at present, it was used to quantify the ovarian Vg content in this study. The Vg protein levels were low during the first 3 days PAE, nearly reached the maximum levels on day 5 PAE, and remained stable until oviposition (day 8 PAE) (Fig. 1A,A′). The results of qPCR revealed that the Vg1 transcription in the fat body began to rise on day 3 PAE, peaked at the middle cycle and steadily decreased thereafter (Fig. 1A′), showing a similar but slightly advanced developmental pattern in comparison with Vg protein levels in the ovary.
The developmental profiles of the sizes of the follicle cells and follicular patency in the maturing oocytes were investigated to monitor ovarian maturation. The size of follicle cells remained constant during the first 3 days PAE, increased significantly from day 3 to day 7 PAE, and reduced on day 8 PAE. Follicular patency appeared on day 4 PAE, indicating that the oocyte started to take up Vg. Although the number of follicular patency decreased in a unit, its diameter linearly increased from day 3 to day 6 PAE, and decreased thereafter (Fig. 1B-B″). Therefore, the number index (patency number/follicle cell number in a unit) profile remained constantly high after day 4 PAE, whereas the diameter index (patency diameter/follicle cell nuclei diameter in a unit) profile mirrored Vg production (Fig. 1B-B″). This quantitative technique (i.e. the number index and the diameter index) serves as an effective method to evaluate the developmental maturation of follicle cells and follicular patency throughout the paper.
Because the Vg1 DNA fragment used to design Vg1 dsRNA showed a relatively high degree of similarity to the corresponding Vg2 fragment, RNAi using Vg1 dsRNA significantly reduced the expression of both Vg1 and Vg2 (Fig. S1C,D). Vg1 RNAi reduced not only Vg1 and Vg2 mRNA levels in the fat body (Fig. S1C) and Vg protein accumulation in the ovary (Fig. 1C,C′), but also the number index and diameter index (Fig. 1D-D″). These results demonstrated that Vg production in the fat body is required for the development of follicular patency and ovarian maturation.
JH signaling regulates vitellogenesis, in part by inducing Dsx expression
To reveal the possible relationship of Dsx, Met and Kr-h1, we first analyzed their expression profiles in the fat body during the first reproductive cycle. In general, the developmental profiles of Dsx, Met and Kr-h1 were analogous to that of Vg1, but Vg1 showed much higher expression levels compared with the other three genes. The expression of Met and Kr-h1 decreased more quickly than Vg1 expression from day 5 PAE, suggesting a lasting JH-induced Vg1 expression. Interestingly, Dsx showed a slightly precocious expression profile compared with Met and Kr-h1, implying that in addition to JH signaling, Dsx expression might be induced by other unknown signals (Fig. 2A,B). A 95% decrease in Dsx mRNA levels was detected in the fat bodies of females injected with Dsx dsRNA, and Vg1 mRNA levels were reduced by 71.8% by Dsx RNAi (Fig. 2C). By contrast, Vg1 RNAi had no effects on Dsx expression (Fig. 2C). Taken together, the results suggest the partial dependence of Vg1 expression on Dsx transcription.
Upon dsRNA treatments, Met and Kr-h1 mRNA levels in the fat body were significantly reduced (Fig. 2C). Interestingly, Met RNAi decreased the Kr-h1 mRNA level, and Kr-h1 RNAi also decreased the Met mRNA level (Fig. 2C), suggesting that Kr-h1 is downstream of Met and Kr-h1 might regulate Met expression through a positive feedback loop. A similar finding was previously shown in the brown plant hopper Nilaparvata lugens (Lin et al., 2015). Knockdown of Met and Kr-h1 led to a significant reduction in Dsx and Vg1 mRNA levels in the fat body (Fig. 2C); thus, the expression of Dsx and Vg1 depends on JH signaling.
Importantly, RNAi depletion of Dsx, Met or Kr-h1 significantly reduced the gonadosomatic index [GSI=ovary weight/(total body weight-ovary weight)×100%] (Fig. 2D; Fig. S2A), primary oocyte length (Fig. 2D′; Fig. S2A), Vg accumulation in the ovary (Fig. 2E; Fig. S2B,C), the number index, and the diameter index (Fig. 2F,F′; Fig. S2D). The changes of all the parameters described above in the maturing oocytes were used to monitor vitellogenesis and ovarian maturation. It is necessary to note that, in contrast with the Met- or Kr-h1-depleted cockroaches, in which neither ovarian growth nor follicular patency was observed, a certain amount of Vg still accumulated in the primary oocytes and size-reduced follicular patency still occurred in the Dsx-depleted cockroaches. The results imply that Dsx is not the single transcription factor downstream of JH signaling in the regulation of vitellogenesis. These experimental data revealed that JH signaling regulates vitellogenesis, in part by activating Dsx expression.
JH biosynthesis is required for vitellogenesis
The mRNA abundance of Jhamt and Cyp15a1 in the head (including the CA) during the first reproductive cycle was determined by qPCR. The expression of the two JH biosynthetic enzyme genes gradually increased and peaked on days 4 and 5 PAE, and decreased thereafter (Fig. 3A), showing developmental patterns similar to JH signaling and vitellogenesis. The effects of JH biosynthesis on vitellogenesis were examined by RNAi knockdown of Jhamt or Cyp15a1. RNAi depletion of Jhamt or Cyp15a1 significantly suppressed the mRNA expression of Met, Kr-h1, Dsx and Vg1 in the fat body (Fig. 3B). Moreover, the depletion of either Jhamt or Cyp15a1 blocked vitellogenesis and ovarian maturation (Fig. 3C-E′; Fig. S3). These data show that JH biosynthesis is required for JH signaling, which then promotes vitellogenesis. Next, we investigated the potential roles of the nutrient sensors (IIS and TORC1) in the promotion of vitellogenesis and the interplay between the nutrient sensors and JH.
IIS and TORC1 are required for JH biosynthesis and vitellogenesis
We measured the developmental profiles of IIS and TORC1 in the fat body during the first reproductive cycle. The total Akt protein level in the first reproductive cycle remained quite constant, whereas the level of phosphorylated Akt (P-Akt, reflecting IIS) gradually increased until day 6 PAE and steadily decreased thereafter (Fig. 4A,A′). The effects of IIS on JH biosynthesis and vitellogenesis were determined by reducing IIS with InR RNAi or LY294002 (a PI3K inhibitor) treatment. Both approaches significantly decreased P-Akt levels and thus IIS (Fig. 4A,A′). RNAi knockdown of InR caused a significant reduction in the expression of Jhamt and Cyp15a1 in the head, as well as the expression of Met, Kr-h1, Dsx and Vg1 in the fat body. The effects of InR RNAi were comparable to those of the LY294002 treatment (Fig. 4B), suggesting that IIS is required for JH biosynthesis and thus the JH-induced Dsx and Vg1 expression. Vitellogenesis and ovarian maturation in the control animals progressed normally but were blocked by InR dsRNA or LY294002 treatment (Fig. 4C-F″), indicating that IIS might act through JH biosynthesis to induce vitellogenesis and ovarian maturation.
The total S6K protein level in the first reproductive cycle remained quite constant, whereas the level of phosphorylated S6K (P-S6K, reflecting TORC1) gradually increased until day 5 PAE and reached a plateau (Fig. 5A,A′). Slimfast (Slif) is a membrane transporter of amino acids involved in TORC1 activation (Colombani et al., 2003). We injected cockroaches with Slif dsRNA, TOR dsRNA or rapamycin (a TORC1 inhibitor) to examine the effects of TORC1 on vitellogenesis. Both rapamycin treatment and RNAi knockdown of Slif or TOR decreased P-S6K levels and thus TORC1 activity (Fig. 5A,A′). Compared with the control, the mRNA levels of Jhamt, Cyp15a1, Met, Kr-h1, Dsx and Vg1 were significantly decreased after TORC1 activity was blocked (Fig. 5B). Furthermore, the block of TORC1 activity arrested vitellogenesis and ovarian maturation (Fig. 5C-F′; Fig. S4). These data indicate that, as with IIS, TORC1 activity is a prerequisite for JH biosynthesis and thus JH-induced vitellogenesis and ovarian maturation.
JH promotes vitellogenesis without additional IIS and TORC1 stimulation
We next examined the effects of starvation (nutrient-deficient; IIS- and TORC1-deficient) and neck ligation (both nutrient- and JH-deficient) on ovarian maturation during the first reproductive cycle. Interestingly, the ovaries were still able to develop to a certain extent after starvation, whereas ovarian development was completely arrested after neck ligation (Fig. 6A-A″). These results not only confirmed the importance of nutrients and the nutrient sensors in stimulating vitellogenesis and ovarian maturation, but also suggested that JH might promote vitellogenesis and ovarian maturation without additional nutritional stimulation. Methoprene is a potent JH analog that is widely used to mimic JH action (Jindra et al., 2013; Li et al., 2019a). Based on these data, we chose the neck ligation approach for the following rescue experiments by injections of methoprene, bovine insulin and an amino acid mixture to restore JH signaling, IIS and TORC1 activity, respectively.
Neck ligation decreased P-Akt and P-S6K levels in both the fat body and ovary, whereas methoprene treatment did not rescue either IIS or TORC1 activity (Fig. 6B-B″). Meanwhile, neck ligation caused a significant reduction in Met, Kr-h1, Dsx and Vg1 mRNA levels in the fat body. Importantly, upon neck ligation, methoprene treatment restored the expression of these genes to levels even higher than that in the feeding group (Fig. 6C). Moreover, vitellogenesis and ovarian maturation were fully restored by methoprene treatment (Fig. 6D-F″). To validate the in vivo rescue experiment results, an in vitro culture experiment was carried out. Fat bodies from newly emerged females (Vg expression was not detected) were in vitro incubated for 6 h in Grace's insect medium containing 2 μM methoprene or the corresponding solvent (DMSO) alone. Again, methoprene treatment increased Met, Kr-h1, Dsx and Vg1 mRNA levels by 1.5-, 149-, 1.9-, and tenfold, respectively (Fig. 6G).
In parallel with the neck ligation and methoprene treatment experiment, we performed an additional rapamycin- and methoprene-treatment experiment. Similar to rapamycin-treatment alone (Fig. 5; Fig. S4), rapamycin- and methoprene-treatment significantly decreased P-S6K levels (Fig. 7A,A′). These results showed that methoprene treatment did not restore rapamycin-inhibited TORC1 activity. Methoprene treatment was not able to rescue rapamycin-inhibited Jhamt and Cyp15a1 expression in the head but it rescued rapamycin-inhibited Met, Kr-h1, Dsx and Vg1 expression in the fat body (Fig. 7B). Moreover, the rapamycin-inhibited vitellogenesis and ovarian maturation (Figs 4, 5) were partially restored by methoprene treatment (Fig. 7C-E″). The composite data demonstrated that JH is able to promote vitellogenesis and ovarian maturation even without additional stimulation by IIS and TORC1.
IIS and TORC1 promote vitellogenesis via inducing JH biosynthesis
To this end, we further examined whether IIS and TORC1 promote vitellogenesis depending on or independent of JH biosynthesis using the neck-ligated cockroaches. Upon neck ligation, bovine insulin and the amino acid mixture restored the fat body P-Akt and P-S6K to the levels observed in the feeding group, respectively, in the fat body. However, the rescuing effects were not observed in the ovary for unknown reasons (Fig. 8A-B′). Surprisingly, JH signaling in the fat body, as well as vitellogenesis and ovarian maturation, was not restored at all by treatments with bovine insulin and the amino acid mixture (Fig. 8C,D; Figs S5, S6). We also determined whether IIS and TORC1 affect JH biosynthesis in vitro. CA incubation with LY294002 or rapamycin significantly decreased the expression of Jhamt and Cyp15a1 (Fig. 8E), consistent with the hypothesis that IIS and TORC1 are required for the promotion of JH biosynthesis. Finally, we examined whether methoprene injection was able to activate JH signaling and induce Vg expression in the adult fat body on day 5 PAE. Methoprene injection significantly activated JH signaling and induced Vg1 expression in adult females in vivo. Importantly, methoprene injection also upregulated Kr-h1 and Vg1 mRNA levels ∼twofold and ∼sixfold in adult males, respectively, in which JH and Vg are normally not produced (Fig. 8F). In conclusion, in the American cockroach, the nutrient sensors IIS and TORC1 promote vitellogenesis mainly via activating JH biosynthesis (Fig. 9).
IIS and TORC1 promote vitellogenesis by inducing JH biosynthesis
The interplay between the nutrient sensors (IIS and TORC1) and JH in the regulation of insect vitellogenesis varies depending on their reproductive strategies. For example, in D. melanogaster, IIS regulates yolk protein uptake into oocytes and autonomously promotes vitellogenesis independent of the roles of JH in the ovary (Richard et al., 2005). In the mosquito Aedes aegypti, IIS and TORC1 stimulate the biosynthesis of JH, which controls a crucial previtellogenic preparatory phase in the fat body but not Vg production; later in the vitellogenesis phase, IIS and TORC1 coordinate with 20-hydroxyecdysone to induce Vg expression in the fat body and during ovarian maturation (Hansen et al., 2004; Pérez-Hedo et al., 2013; Roy et al., 2007, 2018). In T. castaneum and the honeybee Apis mellifera, IIS and TORC1 promote JH biosynthesis, and JH activates insulin production, forming a positive regulatory loop for inducing vitellogenesis (Corona et al., 2007; Parthasarathy et al., 2010; Parthasarathy and Palli, 2011; Sheng et al., 2011). Nevertheless, in the American cockroach, no effects of JH on activating IIS and TORC1 were observed (Figs 6, 7); moreover, in our preliminary experiments, JH was not able to activate insulin production either. In the German cockroach Blattella germanica, IIS promotes both JH biosynthesis and Vg production (Abrisqueta et al., 2014; Maestro et al., 2009), and the authors assumed that IIS acts in parallel with JH signaling to regulate Vg production.
Similar to the results shown in the German cockroach (Abrisqueta et al., 2014; Maestro et al., 2009), we have determined that IIS and TORC1 promote JH biosynthesis and induce vitellogenesis in the American cockroach (Figs 4, 5). More importantly, we demonstrated that IIS and TORC1 promote vitellogenesis in an indirect manner mainly by activating JH biosynthesis, and IIS and TORC1 alone do not directly promote Vg production in the fat body (Figs 6–8, Figs S5, S6). Neck ligation is a classic physiology procedure that was widely used for functional studies by insect physiologists in the past. Neck ligation not only prevented the cockroaches from feeding but also depleted the production of both insulin and JH in the brain. Thus, in the neck-ligated (both nutrient- and JH-deficient) cockroaches, both JH biosynthesis and vitellogenesis were completely abolished. Crucially, JH, but not IIS and TORC1, restored vitellogenesis in the neck-ligated cockroaches. Similarly, JH was able to partially restore rapamycin-blocked vitellogenesis. It is necessary to note that in male adults, no JH is produced in the CA and no Vg is expressed in the fat body, but methoprene injection was able to induce Kr-h1 and Vg1 expression in the fat body (Fig. 8F). It has been previously shown that exogenous JH is able to induce Vg production in the male adult fat body of another cockroach species Diploptera punctata (Mundall et al., 1983) and in nymph fat body of B. germanica (Cruz et al., 2003), in agreement with our results. Therefore, JH is able to promote vitellogenesis without additional IIS and TORC1 stimulation. Combining both classic physiology (i.e. neck ligation and microinjection) and modern molecular techniques (i.e. RNAi), we conclude that IIS and TORC1 promotes vitellogenesis mainly via inducing JH biosynthesis in the American cockroach. It is very likely that the model shown in Fig. 9 applies to a number of insect groups, such as cockroaches, locusts and beetles, in which JH is the major gonadotrophic hormone that promotes vitellogenesis.
Although we have demonstrated that IIS and TORC1 do not induce Vg1 expression in the fat body (Figs 4–8), they should be required for vitellogenesis by stimulating protein synthesis, ribosome biogenesis and cell growth in the fat body and ovarian follicle cells (Li et al., 2019b; Nässel and Broeck, 2016; Smykal and Raikhel, 2015; Wu and Brown, 2006). For example, upon rapamycin- and methoprene-treatment, methoprene treatment fully restored the rapamycin-inhibited JH signaling and Vg1 expression in the fat body but only partially restored the rapamycin-inhibited vitellogenesis and ovarian maturation (Fig. 7). We suppose that in the American cockroach, IIS and TORC1 promote endocycle and cell growth in the fat body, and the ovarian follicle cells thus control a crucial previtellogenic preparatory phase (Fig. S7); later in the vitellogenesis phase, JH induces Vg expression in the fat body and follicular patency for Vg uptake in maturing oocytes (Fig. 9). However, these two phases are not totally separated but are partially overlapped. Interestingly, in the locust Locusta migratoria, JH signaling directly stimulates endocycle and DNA replication in the fat body to prepare for Vg production (Guo et al., 2014; Luo et al., 2017; Song et al., 2014; Wu et al., 2016, 2018). These reports well support our observations that JH promotes vitellogenesis without additional IIS and TORC1 stimulation (Figs 6, 7). It is likely that this two-phase theory should apply to the above insects in which JH is the major gonadotrophic hormone and promotes vitellogenesis.
JH signaling and its regulation of vitellogenesis
As discussed above, in this study, we have conclusively determined that JH induces vitellogenesis, even in the absence of additional stimulation by IIS or TORC1 (Figs 2, 3, 6, 7). Similar to N. lugens (Lin et al., 2015), it is likely that the positive feedback regulatory loop between Met and Kr-h1 (Fig. 3B; Fig. 6C,G) might maximize the JH signaling in vitellogenesis regulation. Based on previous studies, Dsx might directly bind to the promoter region of the Vg gene, thus regulating its expression (Burtis et al., 1991; Gotoh et al., 2014; Shukla and Nagaraju, 2010; Shukla and Palli, 2012). In this study, we found that JH induced Dsx expression and Dsx is required for JH signaling to promote vitellogenesis, establishing Dsx as a molecular connection between JH signaling and vitellogenesis (Fig. 2C; Fig. 6C,G). Interestingly, methoprene injection did not induce the sex-specific Dsx expression in the male adult fat body, partially accounting for the higher JH-induced Vg1 expression observed in males compared with females (Fig. 8F). Nevertheless, RNAi depletion of Dsx did not completely abolish JH-induced Vg1 expression in the fat body (Fig. 2; Fig. S2), suggesting that some other transcriptional factors (i.e. GATA) (Park et al., 2006) should be involved in the JH induction of Vg expression and thus vitellogenesis.
Besides the well-documented JH non-genomic action that induces functional follicular patency in insects (Davey, 2000; Roy et al., 2018), our data show that the JH genomic action is also important for promoting the development of functional follicular patency in the maturing oocytes and thus Vg uptake. Interestingly, the JH-induced Vg production affects the JH-induced follicular patency and thus Vg uptake (Fig. 1C-D′), linking the JH genomic and non-genomic actions in vitellogenesis regulation. Moreover, the JH genomic action is also important for promoting endocycle and DNA replication in the ovarian follicle cells to prepare for Vg uptake in L. migratoria (Wu et al., 2016, 2018). We assume that the coordination of JH genomic and non-genomic actions is likely a general mechanism of vitellogenesis regulation in cockroaches, locusts and beetles, in which JH is the major gonadotrophic hormone.
In summary, our composite data have revealed the molecular mechanisms underlying the endocrinal and nutritional regulation of vitellogenesis in the American cockroach. In this insect species, IIS and TORC1 might promote cell growth and endocycle in the fat body and ovarian follicle cells, and thus control a crucial previtellogenic preparatory phase. Later in the vitellogenesis phase, IIS and TORC1 promote JH production in the CA, and then JH genomic and non-genomic signaling together induce Vg expression in the fat body and activate follicular patency for Vg uptake in the maturing oocytes. The two-phase theory should apply to several groups of insects in which JH is the major gonadotrophic hormone that promotes vitellogenesis.
MATERIALS AND METHODS
The cockroaches used in this study have been described previously (Li et al., 2018). The animals were kept at 28-30°C under a 12/12 h light/dark cycle at a relative humidity of 60-70% in plastic cages and were fed commercial rat food and water. To obtain pools of synchronized animals, newly emerged female adults were collected from the colony and placed in separate containers. Operations were performed after anesthetization with CO2. Dissection and tissue sampling were carried out in cockroach saline solution. For all experiments described in this paper, at least three biological replicates were performed with 30 animals in each replicate.
For RNAi, 300-500 bp fragments of the target genes were amplified from the cDNA (fragment used for designing Dsx is only 100 bp). Then, primers attached to the T7 promoter sequence were used for the PCR amplification of dsRNA templates. The dsRNA was synthesized using a T7 RiboMAX Express RNAi kit (Promega) according to the manufacturer's instructions. A 92 bp noncoding sequence from the pSTBlue-1 vector (CK) was used for the template of a control dsRNA (Gomez-Orte and Belles, 2009). The sequences of all primers used for dsRNA synthesis in this study are listed in Table S1. dsRNA was quantified using a NanoDrop spectrophotometer (Thermo Scientific), and 3 μg of dsRNA was injected into the abdomen of each animal. The injection was carried out on day 2, 4 and 6 PAE. The control animals were injected with the same dose of CK dsRNA at the same time. Tissues were collected after treatments for further use (more details in Fig. S8A). The sequences of all primers used for PCR in this study are listed in Table S1.
LY294002 and rapamycin application
Cockroaches were injected with 4 μg of LY294002 (MedChemExpress) or rapamycin (MedChemExpress) on day 2 and day 4 PAE, respectively. To restore JH activity, 100 μg of methoprene (Cayman Chemical) was injected together with 4 μg of rapamycin. Control groups were injected with the corresponding volume of the solvent. Tissues were collected at day 5 PAE for further use (more details in Fig. S8B).
Methoprene, bovine insulin and amino acid mixture application
In the rescue experiment, the females of day 2 PAE were ligated with dental floss at the neck. A total of 100 μg of methoprene (MedChemExpress) in 2 μl of DMSO (JH analog application), 25 μg of bovine insulin (Shanghai Yuanye Bio-Technology) in 2 μl of 1 M acetic acid or 2 µl of standard solution of amino acid mixture (type H) (Wako) was injected into the abdomen. Control animals were injected with the corresponding volume of the solvent. Tissues were collected at day 5 PAE for further use (more details in Fig. S8C). To activate JH signaling, males or females were injected with 100 μg of methoprene at day 2 PAE and day 4 PAE. Fat bodies were collected at day 5 PAE for RNA extraction.
Tissue imaging and confocal microscopy
Cockroaches were dissected in cockroach saline solution (CSS) using an Olympus SZ61 microscope. To observe the ovary developmental pattern, the ovaries of adult females from day 1 to day 8 PAE were dissected. For the RNAi experiment, the ovaries were dissected on day 5 or day 7 PAE, and body weight, ovarian weight, and primary oocyte length were measured at the same time. Images of the ovaries and ovarioles were captured with a Nikon DS-Ri2 camera and a Nikon SMZ25 microscope. Primary oocyte length was measured using NIS-Elements BR 4.50.00 software (Nikon). For cell staining, ovarioles were fixed in 4% paraformaldehyde for 60 min and permeabilized in 0.1% Triton X-100 for an additional 15 min. F-Actin was stained with TRITC-phalloidin (excitation wavelength 545 nm) (Yeasen). Nuclei were stained with Hoechst 33342 (excitation wavelength 350 nm) (Yeasen). The images were captured with an Olympus FluoView FV3000 confocal microscopy and analyzed with FV31S-SW software (Olympus). The GSI in our study was defined as the portion of the ovary mass to the total body mass excluding ovary mass. The calculation was represented as the following formula: GSI=[ovary weight/(total body weight-ovary weight)]×100%. The size and the number of patency or follicle cells in a unit were measured based on the microscopy image of the ovary using NIS-Elements BR 4.50.00 software. Then we used two indices, the number index and diameter index, to evaluate ovarian maturation. The ratio of patency number to follicle cells number represents the number index, and the ratio of patency diameter to follicle cell nuclei diameter represents the diameter index.
Tissue incubation in vitro
Fat bodies adhered to abdominal tergites and epidermises were dissected from freshly emerged adult females (i.e. day 1 PAE). The tissues were then preincubated for 30 min in 1 ml of Grace's insect medium (Thermo Scientific) at 30°C in the dark (Maestro et al., 2009). After preincubation, fat bodies from females on day 1 PAE were incubated for 6 h in Grace's insect medium supplemented with 2 μM methoprene. The CAs were dissected from adult females of day 5 PAE, then were incubated for 6 h in medium supplemented with 50 mM LY294002 or rapamycin. Control groups were incubated with the corresponding volume of the solvent. After the final incubation, tissues were immediately frozen in liquid nitrogen and stored at −80°C until RNA extraction.
Total RNA extraction and qPCR
Twenty-four hours after the injection of dsRNA or reagents, abdominal fat body tissues and the heads of the cockroaches were collected, flash frozen in liquid nitrogen to prevent RNA degradation and stored at −80°C until further processing. Total RNA was extracted from the fat body, head or CA using a Direct-zol RNA MiniPrep (Zymo Research) according to the manufacturer's instructions. cDNA was synthesized with reverse transcriptase M-MLV (TaKaRa) according to the manufacturer's instructions. qPCR was performed using SYBR Select Master Mix (Applied Biosystems) and the Applied Biosystems QuantStudio 6 Flex Real-Time PCR System. Actin was chosen as a reference gene for qPCR analysis, and the expression levels of target genes are relative to actin for standardization. The sequences of all primers used for qPCR in this study are listed in Table S2.
Total proteins were extracted from the fat bodies of adult females. Tissue lysates in RIPA lysis buffer (Beyotime Biotechnology), with 1 mM phenylmethylsulfonyl fluoride and a phosphatase inhibitor cocktail (CWBIO), were then cleared by centrifugation at 12,000 g at 4°C for 30 min. Extracted proteins were quantified using a BCA Protein Assay kit (Yeasen), fractionated with 10% SDS-PAGE, and then transferred to polyvinylidene fluoride membranes (Millipore). Western blot was performed using antibodies against Akt, P-Akt, P-S6K (9272s, 9271 and 9209s, respectively; Cell Signaling Technology) and S6K (9202; a gift from Prof. Xiaolan Fan, Sichuan Agriculture University, Ya'an, China). Tubulin (AT819, Beyotime Biotechnology) was used as a loading control. All antibodies were used at 1:1000 dilution. Bands were imaged with a Tanon 5200 system and analyzed using ImageJ software.
Statistical analyses were performed using Student's t-test or a Mann–Whitney U-test and SPSS 21.0 software. Data are shown as the mean±s.e.m.
We thank Drs Lynn Riddiford and Marek Jindra for advice and comments on improving the manuscript.
Conceptualization: S.L.; Methodology: S. Zhu, F.L., C.R., Y.Q., S.L.; Software: S. Zhu, F.L., Y.Q., S.L.; Validation: S. Zhu, F.L.; Formal analysis: S. Zhu, F.L., Y.Q., S.L.; Investigation: S. Zhu, F.L., H.Z., C.R., Y.S., S.R.P., J.W., Y.Q., S.L.; Resources: S. Zhu, F.L., Y.Q., S.L.; Data curation: S. Zhu, F.L., Y.Q., S.L.; Writing - original draft: S. Zhu, F.L.; Writing - review & editing: S. Zhu, F.L., C.R., Y.S., S. Zhou, G.W., S.R.P., J.W., Y.Q., S.L.; Visualization: S. Zhu, F.L.; Supervision: Y.Q., S.L.; Project administration: S. Zhu, S.L.; Funding acquisition: N.L., C.R., Y.Q., S.L.
This study was supported by the National Natural Science Foundation of China (31930014, 31620103917, 31900355, 31801968, 31702055 and 31970459 to N.L., C.R., J.W., Y.Q. and S.L.); the Department of Science and Technology in Guangdong Province (2019B090905003 and 2019A0102006 to S.L); and the Shenzhen Science and Technology Program (20180411143628272 to S.L. and G.W.).
The authors declare no competing or financial interests.