The retinoid X receptor from mud crab: new insights into its roles in ovarian development and related signaling pathway

In arthropods, retinoid X receptor (RXR) is a highly conserved nuclear hormone receptor. By forming a heterodimeric complex with the ecdysone receptor (EcR), RXR is known to be vital importance for various physiological processes. However, in comparison to EcR, the RXR signaling pathway and its roles in crustacean reproduction are poorly understood. In the present study, the RXR mRNA was detected in the ovarian follicular cells of mud crab Scylla paramamosain (SpRXR) and during ovarian maturation, its expression level was found to increase significantly. In vitro experiment showed that both SpRXR and vitellogenin (SpVg) mRNA in the ovarian explants were significantly induced by 20-hydroxyecdysone (20E) but not methyl farnesoate (MF). However, differing from the in vitro experiment, injection of MF in in vivo experiment significantly stimulated the expressions of SpRXR and SpVg in female crabs at early vitellogenic stage, but the ecdysone and insect juvenile hormone (JH) signaling pathway genes were not induced. The results together suggest that both MF and SpRXR play significant roles in regulating the expression of SpVg and ovarian development of S. paramamosain through their own specific signaling pathway rather than sharing with the ecdysone or the insect JH.

Scientific RepoRts | 6:23654 | DOI: 10.1038/srep23654 et al. found that in the sand fiddler crab Uca pugilator, MF had a high affinity with RXR, and was synergized with ecdysteriod to stimulate the expression of related genes mediated by the RXR-EcR heterodimer complex 2 . Furthermore, the exogenous MF has been proved to enhance the RXR mRNA level in the green crab Carcinus maenas in both in vitro and in vivo experiments 10 . However, the results of a recent study showed that MF failed to transactivate RXR of the water flea D. magna, which was transfected into human hepatocellular carcinoma cells 25 . Thus, the ligands and the regulation mechanism of RXR in crustaceans remain contentious.
In crustaceans, MF has been reported as a naturally occurring terpenoid synthesized by the mandibular organ (MO), whose position has been identified only in the decapod crustaceans 29 . As the immediate precursor of JH, MF participates in various physiological processes of crustaceans 9,30 . It has been shown that in the giant freshwater prawn Macrobrachium rosenbergii, after larvae being fed Artemia enriched with exogenous MF, the adult morphology development in late stage larvae were retarded, showing similar effects to a well-known feature of JH on preventing metamorphosis in insects 28,31 . MF is also known as a major crustacean hormone regulating reproductive maturation in crustaceans 29 . For instance, it has been reported that following injection of exogenous MF, the ovarian index of C. maenas was significantly enhanced when compared to those control crabs injecting crab saline 10 . Similarly, ovarian maturation of the freshwater crab Oziotelphusa senex senex was also reportedly stimulated by the injection of MF 32 . Other functions of MF include stimulating testicular growth 29 , regulating individual morphogenesis and molting 33,34 , involving in sex determination 35,36 and participating in osmoregulation 29 . Therefore, coupled with its signaling molecules, MF plays important roles in many physiological processes and is critical to the normal function of crustaceans. However, to date its receptor remains poorly understood, and the molecular basis of the signaling pathway as well as underlying functional mechanisms are largely unknown.
The mud crab Scylla paramamosain is a large portunid crab species important to both fisheries and aquaculture along the coasts of southern China as well as in many other Indo-Pacific countries 37 . In a previous study, the nuclear receptor EcR and ecdysone was found to promote ovarian development via regulating vitellogenin (SpVg) expression in the species 38 . However, the functional mechanisms of RXR as another important nuclear receptor and MF on ovarian development of the species have not been clearly studied, especially regarding their signaling pathway. Hence in this study, we cloned three splice variants of RXR from S. paramamosain (SpRXR), and quantified the expression and location of SpRXR during ovarian development via quantitative real-time PCR (qRT-PCR) and in situ hybridization. Using multiple molecular techniques in combination with both in vitro and in vivo experiments, the functions of SpRXR and its related signaling pathway were also investigated.

Results
SpRXR sequence identification and phylogenetic analysis. Three cDNA isoforms of SpRXR were derived from the alternative deletion sequence, which were designated as SpRXR1, SpRXR2 and SpRXR3. SpRXR1 consisted of a 364 bp 3′-untranslated region (UTR) with a poly A tail and 192 bp 5′-UTR while 5 aa and 42 aa insert sequences were found in the T-box domain of SpRXR2 and the ligand binding domain (LBD) of SpRXR1, respectively (Fig. 1A,B). Multiple alignments of the SpRXR isoforms with other RXR/USPs from both vertebrates and invertebrates showed that all the three SpRXR isoforms had a high homology to RXRs of other crustaceans and possessed the typical domain organization of nuclear hormone receptors (Fig. 1B). Interestingly, a conserved acidic amino acid residue (lysine, K) in the ligand dependent activation function 2 region (AF-2) of SpRXR was the same as other decapod RXRs but substituted by glutamic acid (E) in vertebrate RXRs and insect USPs. Moreover, another common feature of crustacean RXRs was that a small insert was always found in the hinge region while a large insert in LBD (Fig. 1B).
To determine the relationships of SpRXR with other homologous sequences, the phylogenetic tree of the RXRs/USPs was constructed using NJ method (Fig. 1C). The unrooted radial NJ tree of RXRs/USPs showed that the RXRs from crustaceans were clustered in a separate clade from those RXRs from vertebrates and USPs from insects. Figure 1C also showed that the crustacean RXRs were closer to the clade of the vertebrate RXRs than that of the insect USPs.

Expression pattern of SpRXR in different tissues and during ovarian development. qRT-PCR
analysis demonstrated that the SpRXR (primers used for expression analysis were designed in the conserved region of three isoforms) was extensively expressed in various tissues of the female mud crab, including muscle, eyestalk, heart, thoracic ganglion, hemocyte, stomach, gill, hepatopancreas, brain and ovary ( Fig. 2A). However, it was also clear that the expression level of SpRXR gene in the ovary was many times higher than in other tissues (P < 0.05).
To better understand the correlation of SpRXR with ovarian development, the relative abundance of SpRXR transcripts at different ovarian vitellogenic stages were also determined by qRT-PCR. The SpRXR mRNA level was found to show an increasing trend with ovarian development from the pre-vitellogenic to the late vitellogenic stage, with a significant increase detected at the late vitellongenic stage (Fig. 2B).
Localizing SpRXR in ovary. In situ hybridization of SpRXR mRNA illustrated that in ovaries of the pre-vitellogenic stage, positive signals were localized in the follicular cells distributed along the periphery of the ovarian lobules rather than inside oocytes (Fig. 3A). With the progress of ovarian development, the positive signals were detected in the follicular cells surrounding the oocytes at both the early vitellogenic and the late vitellogenic stage (Fig. 3C,E). No positive signal was ever detected in ovaries of the negative controls in which the sense SpRXR riboprobe was used instead (Fig. 3B,D,F).

Effects of MF and 20E on genes expression by in vivo experiment. The MF in vivo experiment
showed that 24 h after MF injection, the expression of SpRXR in the female crabs at the early vitellogenic stage was significantly induced while the SpVg did not show significant difference from that of the control. However at 48 h post injection, both SpRXR and SpVg transcripts increased significantly in response to the MF administration (P < 0.05) (Fig. 5).
Differing from the results of SpRXR and SpVg, no significant changes in the expression levels of all identified ecdysone signaling pathway related genes (SpEcR, SpE75, SpE74, SpBr-C and SpHR3) were detected at both 24 h and 48 h in the female crabs receiving MF injection (P > 0.05) (Fig. 6A,B). Similarly, of the two JH response genes identified, at 24 h post injection, the expression of SpKr-h1 was lower while SpMet was higher as compared to the control but the differences were both not statistically significant (P > 0.05). At 48 h post-injection, the expression of both SpKr-h1 and SpMet of the treated crabs were almost identical to that of the control crabs (Fig. 6A,B).
The result of 20E injection showed that at 24 h post injection, 20E significantly stimulated the expression of SpVg. For SpRXR, although the expression was higher in the 20E injected crabs as compared to the control, statistical analysis showed that the difference was not significant (Fig. 7).

Discussion
As one of the most conserved and widespread steroid/retinoid nuclear receptors, RXR plays important roles in various physiological processes in both invertebrates and vertebrates [4][5][6][7][8][9][10][11][12][13] . While past studies have been mostly focused on the function of RXR in vertebrates and USP in higher holometabolous insects 22 , there are increasing studies in recent years on the involvement of RXR in major physiological processes in crustaceans, such as reproduction, development and limb regeneration 4,8,10,12,13 . However, the functional mechanisms of RXR in crustaceans are largely remained unknown.
The present study isolated three isoforms of SpRXR from the mud crab S. paramamosain. The deduced amino acid (aa) sequences showed high similarity to RXRs reported from other crustaceans. The three isoforms of SpRXR were produced as a result of an insertion in either the T-box or LBD (Fig. 1B) and no isoform with both insertions was detected, which is different from what was reported in the blue crab Callinectes sapidus 4 . The failure to detect a SpRXR isoform with both insertions in the present study might be explained by either its low expression level or in fact no such an isoform exists in S. paramamosain. The alternative inserting in the conserved T-box of the isoforms, which is related to mediating interactions of hormone response elements and formation of homodimers or heterodimers, may result in different RXR functions 39 . Similar type of RXR isoforms with a short insertion in the T-box and/or with a long insertion in the LBD have been reported in other crustaceans despite the insertion may vary in size and/or position 4,9,10,[39][40][41] . Unfortunately, different from other nuclear receptors, such as EcR, whose isoforms are studied more comprehensively 42 , up to date the specific functions of different RXR isoforms remain unclear, which warrants further investigation.
In additional to the alternative insertion, multiple sequence alignment demonstrated that the conserved K in the AF-2 region, where is responsible for regulating the ligand-dependent transcriptional activity of the nuclear receptor superfamily 43 , of decapods RXRs is generally substituted by E in vertebrate RXRs and insect USPs (Fig. 1B). It showed that significant divergence existed between the crustacean RXRs and insect USPs/ vertebrate RXRs, suggesting that crustacean RXRs might have special transactivation properties that are differed from those of vertebrates and insects 12,44 . Phylogenetic analysis indicated that with the exception of the water flea D. magna, crustacean RXRs were generally clustered in a separate clade, which was closer to vertebrate RXRs than insect USPs (Fig. 1C). In fact, in the shrimp Metapenaeus ensis, it was reported that the LBD of MeRXR was more homologus to the vertebrate RXRs than the insect USPs 12 . Similar results were also reported in the Kuruma shrimp Marsupenaeus japonicas 8 and the green crab C. meanas 10 .
Expression profiling of SpRXR in different tissues of S. paramamosain revealed that SpRXR was ubiquitously existed in various tissues, however the highest expression level was detected in the ovary, which was many times higher than in all other tissues. It was also found that the SpRXR transcripts were concurrently up regulated in the ovary during ovarian development (Fig. 2B). A similar expression pattern during ovarian development was reported in other crustaceans, including the green crab C. maenas 10 , the fiddler crab U. pugilator 40 , the American lobster Homarus americanus 11 , the swimming crab Portunus trituberculatus 15 and the shrimp M. ensis 12 . These results together indicate that RXR probably play important roles in crustacean ovarian development. Similarly in insects, Wang et al. also found that the expression level of USP-B in the ovary of mosquito Aedesa egypti increased during ovarian maturation 45 .
The control of reproduction in crustaceans is complex and reportedly co-regulated by the ecdysone, crustacean hyperglycemic hormone (CHH) neuropeptides family and MF, which are secreted by Y-organ, X-organ/ sinus gland complex and mandibular organ, respectively 11 . In a previous study, we found that coupled with its receptor (SpEcR), 20E, a biologically active form of ecdysone, played important roles in promotion of ovarian maturation in S. paramamosain 38 . It was also reported that EcR could bind with RXR to form a heterodimer to activate DNA regulatory elements in the centipede Lithobius peregrinus 39 . The injection of RXR dsRNA was also reportedly could knockdown RXR transcript abundance, resulting in significant reduction of ecdysone titers in the German cockroach Blattella germanica 46 and the fiddler crab U. pugilator 19 . In this study, the expression profiles of SpRXR in different tissues, during ovarian development, and especially in the in vitro experiment in which ovarian explants were incubated with exogenous 20E, showed almost the same pattern as SpEcR 38 . In fact, it has also been reported that without EcR, RXR or USP could not be activated by the ecdysone agonist in cultured cells 47,48 ; and the binding affinity of ponasterone A, a potent ecdysone agonist, to EcR was remarkably enhanced by the addition of USP 23 . Moreover, in situ hybridization in this study located the SpRXR mRNA in the follicular cells of S. paramamosain, where SpEcR was also located in a previous study 38 . All these together indicate that SpRXR might be involved in the ecdysone signaling pathway regulating ovarian development in S. paramamosain.
However the results of in vivo experiment showed that although the expression of SpRXR was higher in the 20E injected crabs than those from the control, no significant difference was detected. In crustaceans, both EcR and RXR are vital in the ecdysone signaling pathway, but their expressional profiles may not be always consistent with the 20E concentration in different development stages 8 . In addition, RXR might also function by interacting with MF, indicates that RXR is not regulated only by 20E 4 . Moreover, it has also been reported that EcR can, at least under certain conditions, regulate the transcription of certain target genes without binding to USP or RXR but formed a homodimer/heterodimer with other nuclear receptors 6,7,8,42 . Therefore, the different results with 20E between the in vitro and the in vivo experiment may be attributed to the fact that compared to the in vitro experiment, the physiological process was much more complex for the in vivo experiment in which multiple regulating processes co-existed and interacted, and hence SpRXR might also participate in regulating ovarian development via other ways.
Crustacean ovarian development is characterized by significant vitellogenesis, which involving the synthesis of large quantity of Vg to vitelline as the final product 49 . During the ovarian development, the vitelline was found to gradually accumulate in the ovary of S. paramamosain 50 . The increasing trend of SpRXR expression in the ovary of S. paramamosain during ovarian development and its localization in the follicular cells where SpVg was also detected by in situ hybridization 49 suggested that SpRXR might take part in regulating transcription or accumulation of SpVg during vitellogenesis. In fact, Tiu et al. have reported a correlation between RXR and Vg genes expression profiles in ovary and hepatopancreas, the two Vg synthesis sites, in the American lobster H. americanus during the reproductive cycle 11 . More convincingly, several transcription factor binding sites, including USP and RXR, were found in the 5′ promoter region of the Vg gene 51 . Furthermore, two recent studies showed that in the green crab C. maenas and the salmon louse Lepeophtheirus salmonis, knocking-down the expression of RXR diminished the Vg transcription and yolk granules in the maturing oocytes and diluted the follicular substance, leading to follicular cells filled with vacuoles 6,10 .
Although RXR or USP were initially identified as orphan receptors that act as silent binding partners in heterodimeric combinations 8,22 , recent studies have speculated that crustacean RXRs are also regulated by MF 2,8,10,16,41 . However, in the present study, SpRXR transcription in the ovarian explants was not induced by the exogenous MF, even at a very high concentration (Fig. 4B). A possible explanation for such results could be that   SpRXR distributed in the follicular cells actually did not receive MF, MF was alternatively received by receptor(s) in other tissue, organs or cells, which produced secondary factor (s) that induced SpRXR and SpVg expression in the ovary. In fact, it have also been reported that for in vitro experiments with the shrimp M. ensis and the American lobster H. americanus, MF also did not stimulated the expression of response genes as expected 52,53 . It was speculated by the authors that their results might be related to high hydrophobicity of MF which has a high affiliation with plastic surfaces used for the in vitro incubation 54,55 , and then resulted to considerable lower actual concentration than the designated dosages 53 .
Different from the result of in vitro experiment, the expressions of SpRXR and SpVg of the female crabs at the early vitellogenic stage were significantly induced by MF injection. A recent study by Nagaraju et al. also demonstrated that administration of MF significantly enhanced ovarian index, RXR and Vg transcripts levels in both the ovary and hepatopancreas of C. maeans 10 . In crustaceans, MF is secreted by the MO and has been found in the hemolymph of several species as the key endocrine controller for various physiological processes, including molting, reproduction and morphogenesis 29 . As a gonadotropic hormone, MF can stimulates both ovary and testis development and maturation 29,30 . During ovarian development of H. americanus, a significantly increased MF titer was detected in the hemolymph during secondary vitellogenesis 11 . Injection of MF into the freshwater crab Oziotelphusa senex also significantly increased oocyte and follicle diameters and the gonad index in females 32 . In addition, it was reported that implanting MO, where MF is synthesized, from adult male spider crab Libiniae marginata to the immature female crabs stimulated ovarian growth 56 . Taken together, these results suggest that MF may serve as the upstream signaling of RXR and is involved in the Vg synthesis and vitellogenesis during ovarian development of crustaceans.
To better understand the functional mechanisms of MF on regulating ovarian development in crustaceans, it is necessary to investigate its signaling pathways. However, so far few research was performed to clarify the comprehensive MF signaling pathway system in decapods. Interestingly, the present study showed that the activation of SpRXR and SpVg was not synchronous following MF injection. In the in vivo experiment, SpRXR transcription was significantly induced first at 24 h of MF injection, followed by SpVg at 48 h post MF injection, indicating that the effects of MF inducing the expression of RXR and Vg had a time-lag and between them might involve other hormone response genes. It is known that the RXR can form a complex with EcR and ecdysone to activate the transcription of hormone response genes responsible for vitellogenesis, such as E75, E74, HR3 and Br-C 14 , and that MF can enhance the ecdysone secretion of the in vitro Y-organ in crustaceans 57 . On these bases, we once purposed that MF might share the signaling pathway with ecdysone. However, in the present in vivo experiment, the ecdysone response genes were not induced by the injection of MF. Eichner et al. also reported that in the salmon louse L. salmonis, the mRNA levels of ecdysone response genes except EcR were not regulated downward with the RXR knocking down 6 , which supported our new hypothesis that MF might have its own signaling pathway system rather than shared with ecdysone.
In crustaceans, MF is also known as the precursory and active form of insect JH, which plays vital roles in regulation of reproduction in insects and its signaling pathway is better understood 30,58,59 . For instance, it is known that in insects, Methoprene-tolerant (Met), a basic helix-loop-helix-Per-Arnt-Sim (bHLH-PAS) transcriptional regulator, can form a heterodimer with another bHLH-PAS protein known as steroid receptor coactivator (SRC), in response to JH to activate transcription of the downstream gene, Krüppel-homolog 1 (Kr-h1), to regulate various physiological processes 59,60,61 . Although less is known about MF signal transduction in crustaceans, Met and SRC homologs have been identified in the water flea Daphnia pulex and were found to be responsive to MF, suggesting that signaling pathway system similar to that of JH in insects may exist and conserve in the Daphnia 36 . However, in the present study, no significant changes in expression level of SpMet and SpKr-h1was detected in female S. paramamosain receiving MF injection. In D. pulex, although Met and SRC were sensitive to MF, the expression of Kr-h1 also did not change significantly even under a high titer of MF, suggesting that Kr-h1 is not an MF-responsive gene 36 . It hence indicates that response genes of the MF signaling pathway in decapod crustaceans are likely to be somewhat different from the canonical JH signaling pathway in insects.
In summary, in the present study, we have isolated three SpRXR variants from the ovary of S. paramamosain and the expression levels of SpRXR showed an increasing trend with ovarian development. Similar to SpEcR, SpRXR mRNA was also localized in the follicular cells and significantly higher expression levels of SpRXR and SpVg mRNA in the ovarian explants were induced by exogenous 20E. These results together suggest that SpRXR might be involved in the ecdysone signaling pathway regulating ovarian development in S. paramamosain. Meanwhile, it was found that while injection of MF significantly induced the expression of SpRXR and SpVg, no significant effects on the expression of the ecdysone and insect juvenile hormone (JH) signaling pathway genes was detected, indicating MF and RXR might have its own specific signaling pathway rather than sharing with ecdysone or the insect JH in S. paramamosain. To our knowledge, the present study presents the first attempt to comprehensively investigate the MF signaling pathway system in decapod crustaceans. Since specific components of MF signaling pathway (Fig. 8) were not identified in this study, further research based on biochemical and molecular assays are recommended for the definition of the functions and molecular basis of MF action in S. paramamosain.

Materials and Methods
Tissue sampling and RNA isolation. Our study does not involve any endangered or protected species.
All animals used in this study have been approved by the Animal Ethical Committee of Xiamen University and experiments were carried out in accordance with the approved guidelines of the university.
Healthy female S. paramamosain at different vitellogenic stages (i.e. previtellogenic, early vitellogenic and late vitellogenic stage) were purchased from a local fish market in Xiamen city, China. They were transported to Xiamen University and acclimated for three days in several cement tanks filled with natural seawater (salinity 29 ± 1 ppt; temperature 28 ± 1 °C) before experiments. During the acclimatization, the crabs were fed live clam Ruditapes philippinarum, at a ratio of ~30% of their body weight per day.
Past studies have shown that based on vitellogensis, the ovarian maturation in S. paramamosain can be divided into three distinct stages 38,62 : 1) pre-vitellogenic stage: the ovary appears milky white and contains small oocytes; 2) early vitellogenic stage: the ovary is yellowish to orange in color and oocytes contain yolk granules; and 3) late vitellogenic stage: the ovary appears in orange color and oocytes contain larger yolk granules. Based on above ovarian staging system, the female mud crabs at the early vitellogenic stage were anesthetized on the ice and muscle, eyestalk, heart, thoracic ganglion, hemocyte, stomach, gill, hepatopancreas, brain and ovary tissues were sampled. Additionally, ovarian samples were collected from female crabs with at the pre-vitellogenic and the late vitellogenic stages. All tissue samples were immediately frozen in liquid nitrogen and stored at − 80 °C for later total RNA extraction.
For total RNA extraction, the TRIzol reagent was used according to the manufacturer's instructions (Invitrogen, USA). DNase I was then used to remove residual genomic contaminations. The RNA was subsequently quantified by a ND-2000 NanoDrop UV spectrophotometer (nanoDrop Technologies, Inc., USA) and 2 μg total RNA was reversely transcribed using a reversed first strand cDNA synthesis kit (Fermentas, USA) and stored at − 20 °C. Cloning, sequencing of SpRXR and identification gene fragments potentially involved in RXR signaling pathway. RXR F and RXR R, the degenerate primers designed for multiple alignment of conserved domain, were used to amplify a partial sequence of SpRXR (Table 1). The full sequence of SpRXR was completed using 3′ , 5′ full RACE kit (Takara, Japan) with the specific primers RXR3′ and RXR 5′ (Table 1). PCR products were detected by 1% agarose gel, and the expected DNA fragments were gel-purified and ligated to pMD19-T vectors (Takara, Japan) before being transformed into DH5a competent cells. Clones containing the target fragments were isolated and cultured over night for the subsequent DNA sequencing (Sangon Biotech Co., Ltd, China). Sequence alignment was performed with ClustalW software. The neighbor-joining method (NJ) in MEGA5.0 was used to construct the phylogenetic tree based on protein sequences, and the bootstrapping replication number was 1000.
Based on our transcriptome database for female S. paramamosain (data not shown), six cDNA fragments that might be involved in RXR related signaling pathway were identified and re-sequenced. These include JH-inducible genes of SpKr-h1 and methoprene-tolerant (SpMet), as well as ecdysone induced early genes of broad-complex (SpBrC), hormone receptor 3 (SpHR3), SpE75 and SpE74 10,29,59,63 . The full-length cDNA of SpEcR (GenBank accession number JQ821372.1) was cloned in a previous study by us and reported 38   In situ hybridization. A 325 bp length template of SpRXR was amplified using ovary cDNA by the specific primers of TRXR F and TRXR R (Table 1), which were designed in the common domain of different isoforms. The template was subcloned into the pGEM-T easy vector (Promega, USA) and Digoxigenin-labeled probes were synthesized using a DIG-RNA labeling Kit (Roche, Switzerland). After sampling, the ovarian tissues from the three vitellogenic developmental stages were immediately fixed with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) for one night. The fixed ovarian tissues were then dehydrated through a series of increasing concentrations of ethanol, cleared with xylene and infiltrated with liquid paraffin at 55 °C before finally embedded in paraffin blocks. The blocks were trimmed and sliced to 7 μm on a microtome. In situ hybridization was subsequently performed according to the method described in earlier in our previous study 38 . Photographs were taken using an Olympus multifunction microscope (Olympus BX51, Japan).
In vitro experiment with ovarian explants: effects of MF and 20E on gene expression. Ovarian tissues were dissected from the crabs with ovarian development at the early vitellogenic stage and were rinsed nine times with a crab saline solution containing penicillin G at 300 IU/ml and streptomycin at 300 μg/ml (Sigma-Aldrich Chemical Co., USA). The ovarian tissues were then cut into small pieces of ~30 mg and each pieces was placed in a well of 24-well culture plates filled with 0.3 ml of 2 × L15 medium, which contained either 2 μl 20E (Sigma-Aldrich Chemical Co., USA) or MF (Echelon Biosciences, USA) at a designated experimental concentration. Based on previous studies, the dosages used in this study were 0, 0.05, 0.5 and 5 μM for 20E 38 and 0, 8, 80, 800 ng/ml for MF 10,29,64,65 . Each treatment was triplicated and the culture plates were incubated at 25 °C. Total RNAs from the fragments were extracted 3 h after 20E addition but 6 h after MF addition, which were also based on results reported from previous studies 10, 38 .
In vivo experiment: effects of MF and 20E injection on gene expression. In order to further investigate the role of MF and 20E on ovarian development and the possible MF related signaling transduction, in vivo experiment was conducted to detect the effects of MF and 20E on the expression of SpRXR, SpVg, as well as the ecdysone and JH response genes that might be involved in the MF signaling pathway. For the later, genes fragments related to JH and ecdysone signaling pathway were firstly successfully identified from the transcriptome database of S. paramamosain. For the MF in vivo experiment, fourteen female crabs at the early vitellogenic stage of ovarian development (body weight: 235.5 ± 14.3 g) were divided equally into the control and the treatment group. According to previous reports on measurements of MF in hemolymph of crustaceans 29,64,65 , the crabs assigned to the treatment group received the injection of 20 μl MF at 12 ng/g wet body weight, which translated into about 40 ng/ml in the hemolymph as it has been reported that the total hemolymph volume of a decapod is approximately 30% of its wet weight 26,66 . The crabs assigned to the control received the same volume of carrier. All crabs were then maintained as described previously in "Tissue sampling and RNA isolation". At 24 and 48 h post-injection, three crabs were randomly sampled from each group for the extraction of total RNA from their ovaries.
For the 20E in vivo experiment, six female crabs at the early vitellogenic stage were similarly divided equally into two groups of the control and treatment. Based on our previous study 38 , the crabs assigned to the treatment group received injection of 20 μl 20E at 0.2 μg/g wet body weight while the control crabs received the same volume of carrier. All crabs were then maintained as described previously in "Tissue sampling and RNA isolation" and were sampled 24 h after the injection for the extraction of total RNA from their ovaries. The sampling time at 24 h post-injection was based on a previous study showed that exogenous 20E induced the expressions of SpVg in the early vitellogenic crabs 24 h post-injection 38 .
Statistical analysis. The qRT-PCR data obtained were calculated using 2 −△△Ct before subjected to statistical analysis. Student's t-test and one-way analysis of variance (ANOVA) followed by Duncan's test were performed to determine any statistically significant differences among treatments, which was set at P < 0.05 level. All statistical analysis was performed using the SPSS 13.0 software (SPSS, Chicago, USA).