17β-Estradiol (E2) may be involved in the mode of crustacean female sex hormone (CFSH) action in the blue crab, Callinectes sapidus

Animals were reared in the blue crab hatchery, Aquaculture Research Center [Institute of Marine and Environmental Technology (IMET), Baltimore, MD]. After reaching prepuberty, the females were monitored for their molt stages, early premolt and late premolt (23). For the mating study, an individual prepuberty female at the E₀ stage (24) was placed into a 30-liter container, holding 10 liters of 20 ppt artificial seawater (ASW), with an adult male to allow for mating immediately after puberty molting. After mating, adult females with carapace widths of 132.4 ± 3.5 mm (n = 10) were kept in five tanks in a closed circulating aquaculture system (91 cm W × 112 cm H × 58 cm D). Each tank held 300 liters of ASW under the following conditions: 16 hours dark, 8 hours light at 18-20°C (22). The water quality was monitored daily by ZooQuatic Lab (Baltimore, MD). Animals were fed about 10% of their body weight with a piece of frozen squid, shrimp, or mussels daily at 8–9 A.M. during the experimental period (Oct. 2019-Feb. 2020).

The experimental and prepuberty animals were sacrificed in the afternoon (2–6 P.M.). A hemolymph sample (200 µl) was withdrawn directly into an insulin syringe containing crab anticoagulant solution (i.e., 450 mM NaCl, 10 mM KCl, 10 mM NaHCO₃, 1 mM EDTA, 10 mM HEPES, pH 7.3) in a 1:1 ratio (v:v). The animals were then placed on ice for 10 min before dissection. The eyestalk ganglia, brain, and ~100 mg of the hepatopancreas, ovary, spermatheca, ovigerous setae, and plumose setae tissues were dissected in ice-cold DEPC-treated crustacean saline (19) under a stereomicroscope (Leica). After two rinses in fresh saline, the tissues were briefly blotted with Kimwipes, collected into a 1.5-ml tube, and placed on dry ice. The wet weights of tissues were measured, and then the tissues were stored at -80°C until further analysis.

The steroid extraction was carried out by following the procedure described in Chao etal. (25). In brief, a single eyestalk ganglion, brain, or 10 mg (wet weight) tissue samples were homogenized in 135 µl ice-cold phosphate buffer saline (PBS). The steroid was extracted by adding 635 µl diethylether to the homogenized sample. The tubes were vigorously vortexed for 1 min, and then centrifuged at 14000 rpm for 10 min at room temperature. The tubes were gently placed in an ethanol/dry ice bath to freeze the aqueous phase, and the lipid phase was transferred to a 4-ml glass tube (Fisher Scientific). The extraction steps were repeated four times with the aqueous phase of each sample. The pooled lipid phase was either dried by air or with a SpeedVac (Jouan). Samples were kept at -20°C for E2 measurement using ELISA or LC-MS/MS.

E2 levels in the tissues were estimated using an ELISA kit by following the manufacturer’s protocol (Enzo Biochem). The standard curve was generated using E2 solutions with known concentrations of 29.3 – 30,000 pg/ml. The standards and experimental samples were assayed in triplicate, and the plate was read at an absorbance of 405 nm using a SpectraMax M5 reader (Molecular Devices). The EC₅₀ values of ELISAs were 534 ± 12 pg/ml (n = 6). Data were presented as follows: mean ± SE pg/tissue (n = 5) for the eyestalk ganglia and brain; mean ± SE pg/10 mg wet weight (n = 5) of the hepatopancreas, ovary, spermathecae, ovigerous setae, and plumose setae tissues; and, mean ± SE pg/10 µl (n = 5) of hemolymph.

The presence of E2 in the brains (adult females) was confirmed by an LC-MS/MS method modified from previous protocols (26–28). First, 100 ng (100 µl × 1 µg/ml) of the E2-d3 surrogate standard (Sigma-Aldrich) was added to the dried lipid phase collected from brain samples (section 2.3) and incubated overnight. After reconstitution in 205 µl PBS, the sample solution was passed through a Sep-Pak^® C18 solid-phase extraction (SPE) cartridge (Waters) to remove interfering substances (26). The SPE cartridges were washed with 5 ml of 100% isopropanol and then conditioned with 5 ml MilliQ (Waters) or Hydro water. After loading the 200-µl sample onto the SPE cartridge by gravity, the column was washed with 5 ml of water. Next, E2 and the E2-d3 surrogate standard were eluted from the SPE cartridge with 3 ml of 40% isopropanol, and the first five drops of eluate were removed as void volume. The extracts were dried at 40°C in a vacuum centrifuge (Juan), reconstituted in 220 µl PBS, and diluted to 300 µl with 50% methanol containing 200 ng/ml of the 17-ethinylestradiol-2,4,16,16-d4 (EE2-d4; CDN Isotopes) internal standard.

E2, E2-d3, and EE2-d4 were measured using an UltiMate 3000 LC coupled to a Thermo TSQ Quantum Access Max MS/MS (Thermo). Briefly, 50 µl of the reconstituted SPE extracts were injected onto a Waters XBridge C18 guard column (2.1×10 mm, 3.0 µm) connected to an analytical column (2.1×150 mm, 2.5 µm) with the same material. The mobile phase was comprised of (A) LC-MS grade water with 0.1% NH₄OH (pH 10.5) and (B) methanol with 0.1% NH₄OH. Isocratic elution was conducted with 30% A at a flow rate of 200 µl/min. The column compartment was maintained at 40°C, and the overall method run time was 8 min. Negative electrospray ionization mode was employed for LC-MS/MS analysis, and two characteristic product ions were used to quantify and confirm each analyte. The surrogate standard recovery was 89% in the brain sample. The detailed analytical parameters are available in Table S1 of the supporting information.

The total RNA (~1-1.5 μg) from the eyestalk ganglia was subjected to 5’ and 3’ rapid amplification of cDNA ends (RACE) cDNA synthesis using the SMART cDNA amplification kit (BD Bioscience) according to the manufacturer’s protocol. Gene-specific primers for 5’ and 3’ RACE (Table 1) were designed using NCBI Primer-BLAST based on transcripts pulled from our transcriptomes. The amplified products were purified using a QIAquick gel extraction kit (Qiagen) and ligated to a pGEM-T easy vector (Promega). The recombinant vector was used for transforming E. coli competent cells. Clones containing the inserts were isolated and cultured at 37°C overnight for subsequent DNA sequencing (ABI).

The open reading frames were identified using ExPasy (https://web.expasy.org↗). The conserved domain was searched via SMART (http://smart.embl-heidelberg.de/↗). ClustalW (www.genome.ad.jp↗) was used to align the amino acid sequences, which were then modified by ESPript 3.0 (https://espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi↗). The phylogenetic tree was constructed using the neighbor-joining method (MEGA X).

Total RNA was extracted from the tissues using QIAzol Reagent (Qiagen) following the manufacturer’s procedures. RNA concentrations were measured using a Nanodrop (Thermo Scientific). Reverse transcription with 1.5 µg total RNA was carried out using a PrimeScrip^RT reverse transcriptase reagent kit with gDNA eraser (TaKaRa).

After being diluted with RNase-free water at 12.5 ng/μl, cDNA was amplified for the arginine kinase (AK) gene to examine the quality of each sample (19). The spatial distribution of the steroidogenesis genes, StAR3, 3βHSD, 17βHSD8a, 17βHSD8b, and ERR, was examined in the tissues obtained from an adult female using an endpoint RT-PCR assay with the gene-specific primers reported in Table 2, together with AK and eIF4A (eukaryotic translation initiation factor 4A) as reference genes. The PCR conditions were as follows: an initial denaturation step at 94°C for 2.5 min, followed by 35 cycles of 94°C for 30 s, 57°C for 30 s, and 72°C for 30 s, with a final extension 72°C for 5 min. PCR products were resolved on 1.5% agarose gel and stained with ethidium bromide for visualization.

Each cDNA sample containing 25 ng total RNA was assayed in duplicate using Fast SYBR Green Master Mix (Applied Biosystems) and gene-specific primers (Table 2) by following the manufacturer’s protocol. Gene-specific standards were produced as described in Chung etal. (30). The coefficients of determination (R²) and the average amplification efficiency of samples are listed in Table 2. The expression levels of each gene were normalized by the AK and eIF4A levels in the same cDNA samples (29). Data were presented as mean ± SE (n = 5) transcripts/μg total RNA.

CFSH-dsRNA preparation and injection were carried out as described (19). The mated adult females were injected twice a week with 10 μg CFSH-dsRNA in 100 μl crustacean saline (n = 5) or saline-only controls (n = 5). These injection regimes were continued 30 times over four months (Oct. 2019-Feb. 2020). Tissues were sampled (section 2.1) for E2 measurement and expression analysis using RT-qPCR assays.

All data were analyzed using SPSS 25.0 for statistical significance. The data were analyzed in two ways. First, one-way ANOVA was employed with multiple samples (e.g., different tissues), and a non-parametric Kruskal-Wallis test followed by Dunn’s post-hoc test was used for multiple group comparisons. Second, the student’s t-test was used when two groups (e.g., control and test, or the tissue at different stages) were compared. Significance at p < 0.05 was accepted and noted by letters or ‘*’ symbols.

Tissues of the prepuberty females at early and late premolt were measured for E2 using a competitive ELISA (Figure 1A). All tissues at both molt stages had E2, ranging from the lowest level in the hemolymph (9.5 ± 1.4 pg/10 μl, n = 5) to the highest level in the brain (341.0 ± 30.5 pg/tissue, n = 5) at early premolt and from 9.6 ± 1.2 pg/10 μl (hemolymph) to 256.4 ± 64.6 pg/10 mg wet weight (brain) at late premolt (n = 5).

E2 levels differed by molt stage in the following tissues: ovigerous and plumose setae, spermathecae, and ovary. Ovigerous setae at late premolt had E2 levels of 33.1 ± 5.0 pg/10 mg wet weight (n = 5), significantly lower (p < 0.01) than early premolt (104.3 ± 20.9 pg/10 mg wet weight, n = 5). The E2 content in plumose setae and ovary at late premolt significantly decreased (p < 0.05) by about twofold to 112.3 ± 2.0 pg/10 mg wet weight (n = 5) and 49.1 ± 2.2 pg/10 mg wet weight (n = 5), respectively, compared to those at early premolt. Significantly (p < 0.05) reduced levels of E2 were also observed in the spermathecae (early premolt: 85.4 ± 30.4 pg/10 mg wet weight; late premolt: 18.7 ± 4.6 pg/10 mg wet weight, n = 5). However, no changes were identified for the E2 content of the eyestalk ganglia, brain, hepatopancreas, and hemolymph at early and late premolt.

According to the ELISA results, the brain contained the highest E2 level. Therefore, this tissue was chosen for further confirmation of E2 via LC-MS/MS. The LC-MS/MS chromatograms for an E2 standard and the brain samples are shown in Figure 1B. E2 was positively detected in the brain tissues; however, the measured amount was below the method quantitation limit (i.e., 150 pg/tissue, n =3).

Figure 2 shows the potential functions of the steroidogenesis genes, StAR3, 3βHSD, 17βHSD8a, 17βHSD8b, and ERR, in C. sapidus involved in the biosynthetic pathway of vertebrate steroid hormones. The full-length StAR3 and 3βHSD sequences were cloned from the eyestalk using PCR with gene-specific primers (Table 1). StAR3 cDNA was 2207-bp in length with a 188-bp 5’UTR, a 1389-nt open reading frame (ORF), and a 630-bp 3’UTR with a poly-A tail (Figure 3A). The ORF encoded a polypeptide of 462-aa, including a 198-aa cholesterol-capturing domain (MENTAL) and a 177-aa StAR-related transfer domain (START). Multiple alignment analysis showed 32-68% (MENTAL) and 32-38% (START) homologous between vertebrates and invertebrates (Figure S1A).

The full-length 3βHSD cDNA (1604 nt) consisted of a 527 bp 5’UTR, a 912 bp ORF, and a 165 bp 3’UTR, including the poly-A tail (Figure 4A). The conserved domain database identified the deduced amino acid regions from M1 to P225 as the 3βHSD/isomerase family (3β_HSD), which illustrated 28-60% homology to vertebrates and invertebrates (Figure S1B).

The cDNA sequences of the 17βHSD8a and 17βHSD8b genes were isolated, encoding 247 and 256 aa, respectively (Figure 5A). Comparison of the conserved domain sequences of 17βHSD8a and 17βHSD8b derived by comparing sequences with vertebrates and invertebrates showed 49-91% and 45-50% similarity, respectively (Figure S1C). Furthermore, 51% sequence identity was found when the amino acid sequence of 17βHSD8a ORF was compared with that of 17βHSD8b.

The ERR cDNA contained a 1482-bp ORF encoding 493 aa, consisting of two conserved domains, namely zf-C4 (zinc finger, C4 type) and hormone receptor (ligand-binding domain of nuclear hormone) (Figure 6A). The amino acid sequences of conserved regions exhibited 79-90% and 36-90% sequence identities to those of vertebrates and invertebrates, respectively (Figure S1D).

The phylogenetic trees of StAR3, 3βHSD, 17βHSD8, and ERR showed that all of the amino acid sequences of C. sapidus were clustered with the invertebrate (crustaceans and insects) clade and separated from the vertebrate (zebrafish) clade (Figures 3–6B). The C. sapidus StAR3 most closely aligned with Homarus americanus and then formed the decapod subclade (100% support) with two shrimp species (Figure 3B). The C. sapidus 3βHSD was first clustered with P. monodon and P. vannamei, grouped with H. americanus, and then merged into a large branch with isopods, clearly distinguishable from insects and a vertebrate (Figure 4B). All of the crustacean 17βHSD8 proteins formed a subclade, in which the C. sapidus 17βHSD8a and 17βHSD8b were positioned close to the 17βHSD8 of two crab species, P. trituberculatus and Chionoecetes opilio, respectively (Figure 5B). The ERR protein from C. sapidus had the closest evolutionary relationship with S. paramamosain and was classified into an invertebrate clade with those isopods and insects (Figure 6B).

The spatial distributions of steroidogenesis genes were examined in the various tissue cDNAs of an adult female C. sapidus using an endpoint RT-PCR assay (Figure 7). All of the tissues showed the presence of StAR3, 3βHSD, 17βHSD8a, and ERR, and the expression of these genes was relatively high in the eyestalk ganglia, brain, and the adult-female-specific tissues but low in the hepatopancreas. In contrast, 17βHSD8b was mainly expressed in the brain and ovary. AK and eIF4A had high expression in all tissues.

The ovary contained relatively high mRNA levels of 17bHSD8. Hence, the ovary was used for in situ hybridization and immunohistochemistry if the transcripts of 17βHSD8a and 17βHSD8b and 17βHSD8 protein were present (ovarian stage 2) (Figure 8). The signals of 17βHSD8a and 17βHSD8b transcripts were localized in the follicle cells, the periphery of the nuclear membrane of primary oocytes (Figures 8C, D), and yolk granules of vitellogenic oocytes (Figures 8E, F); furthermore, 17βHSD8 protein was visualized in the follicle cells and ooplasm of primary oocytes (Figures 8G–I). Sense probes and pre-immune serum had no positive signals in the ovary (Figure S2).

The presence of E2 in the crustacean tissues has been primarily reported using antibody-based assays. Our ELISA results showed that various tissues of the prepuberty and adult blue crab females contain E2, congruent with earlier findings that report E2 in the hemolymph, hepatopancreas, and ovary of H. americanus (31), Pandalus kessleri (32), and Artemia sp. (33). In this study, an LC-MS/MS spectrometry method confirmed that E2 measured in blue crab tissues is the genuine and endogenous E2. The optimized LC-MS/MS analytical protocol exhibited a good recovery (89%), and the measured E2 concentration matched that from the ELISA analysis. Due to the higher method quantitation limit of the LC-MS/MS method for E2, ELISA was more preferred for E2 measurement in C. sapidus.

This study focused on E2 in early and late premolt C. sapidus females at prepuberty, during which time the ovigerous and plumose setae are associated with development of blue crab females (19). These adult-specific tissues are developing at early premolt and are fully developed at late premolt. Our data showed a marked difference in E2 levels in a molt stage-specific manner, with higher levels at early premolt than at late premolt, when these tissues primarily develop the features specific to adults for mating and maternal care systems. Based on previous findings that CFSH is required to develop adult-specific morphological features, it is reasonable to suggest that higher E2 levels in these tissues may be linked to CFSH function.

In both prepuberty and adult females, maximum E2 levels were present in the brain. In vertebrates, E2 affects the structure and function of the nervous system (34) and provides positive feedback on the development and function of the gonadotropic axis (35–37). The function of E2 in the brain of crustaceans is poorly understood. Considering that mating occurs immediately at the onset of puberty-terminal molting, E2 in the brain of the prepuberty females may be linked with sexual behaviors and involved in the reproductive neuroendocrine responses that occur in the brain of crustaceans (38).

In comparison to the vitellogenic ovaries (ovarian stage 2) in adult females, the undifferentiated ovaries in prepuberty females had higher E2 levels. This finding appears to be rather different from those reported in other decapods, in which E2 levels in the ovary significantly increased with ovarian development and oocyte maturation (7, 39, 40). One plausible explanation is that season affects E2 synthesis in female gonads. In the present study, ovaries were sampled from the prepuberty and adult females during summer and winter months, respectively. Thus, warmer temperatures may stimulate reproductive activities and biosynthesis of E2, while these processes are slowed or halted in colder temperatures.

The levels of E2 measured in the tissues were converted and presented as pg/tissue type, with the same 10 mg wet weight wherever possible. In the present study, E2 levels were lowest in the hemolymph of the prepuberty and adult females, but hemolymph accounted for ~15% of body weight (e.g., 70 ± 10 g for prepuberty, 160 ± 20 g for adult, n = 10); therefore, the hemolymph contained the greatest amount of E2 compared to other tissues, suggesting that E2 may act as a hormone in this species.

The presence of CFSH protein was consistent throughout the female life cycle. Considering the vital role of CFSH in developing adult female-specific tissues, the following question is asked if this hormone has any functions in these tissues, possibly involving E2. Indeed, our data showed that the tissues specific to adult females contained E2. The E2 levels measured in adult tissues were compensated with the increment of their size through the puberty-terminal molting, i.e., a maximal molt increment of 50%, compared to those of the late premolt stage. Interestingly, the ovigerous setae of adults contained E2 levels similar to those of the late premolt prepuberty females, but the other adult tissues had lower E2 contents than at late premolt, which might stem from the lower E2 in the adult hemolymph.

To further examine the involvement of putative E2 steroidogenesis genes to confirm the presence of endogenous E2 in female C. sapidus, several cDNAs encoding these orthologues were isolated and characterized. The blue crab cDNA sequences of steroidogenesis genes were grouped and formed a clade with other crustacean sequences. The sequences of invertebrates, including insects and crustaceans, were separated from a vertebrate sequence, zebrafish Danio rerio, as an out group.

StAR was chosen due to its involvement in the translocation of cholesterol from the outer to the inner mitochondrial membrane, a process that is regarded as the first and rate-limiting step in the steroidogenic pathway of vertebrates (41). Therefore, StAR plays a vital role in the regulation of steroidogenesis. In this study, we identified the presence of StAR3 but not StAR, indicating that a StAR-independent steroidogenic pathway may be adopted by C. sapidus. A similar phenomenon was reported in M. rosenbergi (17).

3βHSD and 17βHSD regulate two different steps in vertebrate steroidogenesis. The former is responsible for both the reduction and oxidation of the 3-keto/hydroxyl and the Δ⁵-Δ⁴-isomerization that generate progesterone and androstenedione, while the latter is involved in the synthesis of active sex steroids that catalyze androstenedione to testosterone, estrone to E2, or vice versa (14). To date, 3βHSD and types 2, 3, 4, 6, 8, 11, and 14 of 17βHSDs were identified in M. rosenbergii and P. trituberculatus (16, 17). In the present study, a 3βHSD and two isoforms of 17βHSD8 were identified in C. sapidus.

Due to the high structure and sequence similarity and close phylogenetic relationship to ERs, the orphan nuclear receptors, ERRs, were identified (42). Unlike ERs, ERRs are not activated by E2 but bind to anthropogenic estrogenic ligands, estrogen, and ERR response elements and, thus, play a physiological role in embryonic and gonadal development (43, 44). In crustaceans, ERRs were detected in Daphnia magna (45), D. pulex (46), and Diaphanosoma celebensis (47), and may be involved in estrogen signaling and metabolism pathways.

The most important cytochrome P450 enzymes, including the P450scc/cyp11a (P450 side chain cleavage), P450c17/cyp17a, and P450arom/cyp19a (P450 aromatase), were not identified in this study. The presence of P450c17 has been documented in echinoderms (48), mollusks (49), and some decapods (17), while P450scc and P450arom have not yet been identified outside of vertebrates (14). However, the conversion of cholesterol to pregnenolone and aromatase-like activity were detected in some invertebrates (50–53), indicating the orthologues of P450scc and P450arom that function in cholesterol catabolism and estrogen synthesis may also exist in invertebrates.

A broad tissue distribution of vertebrate steroidogenesis genes was found in adult female C. sapidus using RT-PCR, possibly explaining the widespread presence of E2. The relatively high expression levels of all steroidogenic genes in the brain were consistent with the ELISA results, wherein the brain had the highest E2 levels among tested tissues. Similarly, most of the enzymes involved in steroidogenesis were highly expressed in the central nervous system of mammals (54), scallops (49), and prawn (17). On the other hand, the low E2 contents in the hepatopancreas may be caused by the relatively low transcript levels of StAR3 and 3βHSD, which are related to E2 synthesis, and the high level of 17βHSD8a, which is associated with E2 metabolism. Interestingly, 17βHSD8 presented two isoforms with different expression patterns: 17βHSD8b was mainly expressed in the brain and ovary, implying its prominent role in E2 metabolism in these tissues; and, 17βHSD8a was ubiquitously expressed in all tissues. An extensive tissue distribution pattern of 17βHSD8 was also found in humans (55), mice (56), and scallops (57). Similar to StAR3, ERR was expressed in all of the examined tissues, and the relatively low levels in the hepatopancreas and plumose setae may corroborate ERR involvement in the transcriptional regulation of StAR (58).

The localizations of 17βHSD8 mRNAs and protein were visualized by in situ hybridization and immunohistochemistry to indirectly support the existence of endogenous E2 in C. sapidus. The positive signals of 17βHSD8 transcripts and protein were clearly visible in the follicle cells and oocytes of the ovary, consistent with results from a previous study conducted at both the gene and the protein levels (57). However, positive hybridization signals were also detected in granulosa cells of growing follicles and luteal cells in the mouse ovary (56). We speculate that the discrepancy may be attributed to the differences in ovary structures between mammals and crustaceans. The 17βHSD8 transcripts were obviously visible in the yolk granules of vitellogenic oocytes, implying the physiological role of 17βHSD8 in vitellogenesis and ovarian development. In mollusks, changes in the expression of 17βHSD8 appear to be related to the maturity of female gonads (57). The high transcript abundance of 17βHSD8 and protein levels may be required to maintain the optimal E2 concentrations for oocyte development and ovarian maturation in C. sapidus.

After finding E2 in the adult-specific tissues of adult females, along with the presence of CFSH protein in adult eyestalks (19), the function of CFSH in these tissues was examined with E2 and steroidogenesis genes using CFSH-dsRNA injection. The dsRNA-mediated RNA interference has been used to define the function of a specific protein. In many decapod species, the injection of a gene-specific dsRNA significantly decreased the target transcripts and protein levels (19, 59–62). Therefore, if CFSH function involves E2, CFSH-dsRNA was expected to reduce E2 and steroidogenesis gene transcript levels in these adult-specific tissues compared to controls.

This hypothesis was confirmed by our data, because E2 levels in the adult-specific tissues were lower than in controls, while the E2 contents in other tissues, such as ovary, brain, eyestalk ganglia, hepatopancreas, and hemolymph, exhibited no differences. Among the examined steroidogenesis genes in the control animals, 17βHSD8a and StAR3 transcripts were the highest, followed by ERR, 3βHSD, and then 17βHSD8b. The most consistent effect of CFSH-dsRNA injections was observed with StAR3 transcripts in the adult-specific tissues, where the levels were lower than controls, congruent with the E2 content in these tissues. It is plausible to suggest that CFSH exerts its functions in the adult female-specific tissues via E2 synthesis by regulating the expression of StAR3. E2 synthesis in crustaceans is the same as described in vertebrates, wherein StAR is required for cholesterol movement. This step is also known to critical for ecdysteroidogenesis in the Y-organ (63).

CFSH-dsRNA injection had no significant effects on transcript levels of 3βHSD in adult-female-specific tissues but decreased expression levels of 17βHSD8a, 17βHSD8b, and ERR in a tissue-specific manner, indicating these genes were unlikely to be directly regulated by CFSH. Previous studies demonstrated that CFSH might not be involved in ovarian development (19, 64). However, the significantly reduced transcript level of 17βHSD8b in the ovary of adult females injected with CFSH-dsRNA may affect the metabolic efficiency of E2 leading to changes in ovarian differentiation. The relationship between transcript levels and protein contents are unknown (65, 66); therefore, the long-term effects of CFSH-dsRNA injection on translation of steroidogenesis genes in the adult-female-specific tissues need to be further investigated. Moreover, the broad distribution of testosterone in males is expected based on the presence of E2 and steroidogenic genes in female C. sapidus.

After a long-term CFSH-dsRNA injection, notable abnormalities were not observed in the length and abundance of ovigerous and plumose setae, wet weight and size of spermathecae, and the presence, position, and size of gonopores (). These results indicated that CFSH and/or E2 may not be required for the maintenance of brooding and mating systems in adult females. In addition, 30 injections of CFSH-dsRNA seemed to be lethal because ~50% mortality (3/8) was obtained, and no mortality was observed in saline-administrated animals. The difference in mortality might have occurred as off-target effects of CFSH-dsRNA. 3