Crabs were collected from coastal area of Zhanjiang, Guangdong Province, China. Individuals (i.e., 2.1–3.4 g) were selected and stocked in 1 m³ concrete ponds. They were cultured at 24–27°C, salinity at 20‰ and fed daily with oyster. After a week, healthy crabs at D0 (i.e., 2.4–2.9 g) were selected for molting experiment, healthy crabs at D0 and D2 stages (Ong, 1966) were selected for dsRNA-injected experiments.

Crabs were chilled on ice before dissection for tissues. For every tissue that was analyzed, samples were collected in three different pools of five dissected animals each and stored at −80°C until further processing. Total RNA preparation was prepared using a column-based TransZol Up Plus RNA kit (TransGen Biotech, China). After elution, the concentration of total RNA was determined using the Thermo Fisher Scientific the NanoDrop 2000 (Waltham, MA, United States) and the quality of total RNA was tested by 1% agarose gel. For synthesis of first strand cDNA, 5 × All-In-One RT cDNA synthesis kit (ABM Inc., Richmond, BC, Canada) was used. The cDNAs were diluted tenfold for later analyses.

The open reading frame of ETH (SRR3086589, SRR3086590), EH (GeneBank accession No: KR078366.1) and CCAP (GeneBank accession No: MN923209.1) were obtained from NCBI Database (Supplementary Figures 1–3). Primers were designed using the Software Primer 5.0 (Supplementary Table 1) and 18s rRNA and β-actin were used as internal control genes (Chung and Lin, 2006; Liu et al., 2020). Each qPCR reaction contained 10 μL of SYBR qPCR Master Mix (Vazyme, Nanjing, China), 2 μL of cDNA, 7 μL of ultrapure water and 0.5 μL of each Forward and Reverse primer (10 μM). PCR was performed with the Real Time PCR machines (Bio-Rad CFX Connect PCR, Bio-Rad, United States). The following thermal cy-cling profile was applied: 95°C for 2 min, followed by 39 cycles of 95°C for 5 s and 55°C for 30 s, and then a melt curve analysis was performed to verify the specificity of the qRT-PCR reactions. The relative mRNA abundance of each gene was calculated by ΔΔCt method, then normalized by that of 18S and β-actin for each sample. The reaction was performed in triplicated for each sample. All primers used in the present study were given in Supplementary Table 1.

Firstly, primers flanked by the T7 promoter sequence were designed and used to synthesize the dsRNA for EH, ETH, CCAP and control gene (GFP). Secondly, the dsRNA template for each gene were amplified (PCR amplification procedure: 95°C for 5 min; followed by 30 cycles of 95°C for 30 s, 57°C for 30 s and 72°C for 30 s; then 72°C for 10 min) with above primers and purified by means of FastPure Gel DNA Extraction Mini Kit (Vazyme, Nanjing, China). After the concentration detection of DNA tem-plates (NanoDrop 2000, Thermo Fisher Scientific, Inc., Waltham, MA, United States). dsRNA was produced with DNA templates under the method of T7 RiboMAX™ Express RNAi Synthesis kit (Promega, Beijing, China). The final dsRNA was diluted to appropriate concentration (1 μg/μL) with water. Synthetic peptides of ETH and CCAP (Oliphant et al., 2018) were synthesized by GenScript (Nanjing, China) and dissolved in water to 1 μg/μL.

RNA interference experiments (n = 50) were done with the injection of dsEH, dsETH, dsCCAP, and dsGFP, respectively. Each crab was injected in dsRNA at a dose of 5 μg/g. For the synthetic peptide compensation experiments, each of dsEH, dsETH and dsCCAP were co-injected with ETH and CCAP synthetic peptides, respectively, and each co-injected combination named a group (n = 50). Both dsRNA and synthetic peptide were injected at a dose of 5 μg/g, which was an over-added level of synthetic peptides in case of inadequate compensation to the negative effects of gene knockdown. Control group was injected with dsGFP following the same injected dose. Boost injections were given every 5 days to ensure a lasting affection of gene mRNA levels. The number of deaths and molts in each group were counted daily, also the stage of any deaths were observed by microscope. Crabs injected with dsGFP were served as control.

Crabs (n = 20) at D0 and D2 stages were collected and separately injected with dsEH and dsETH at a dose of 5 μg/g at each stage. Also, crabs (n = 20) at D2 stage were selected for co-injection of dsEH and dsETH at a dose of 5 μg/g for each dsRNA. After 48 h, the high-expression tissues of nerves were sampled and analyzed. Crabs injected with dsGFP were served as control.

The relative transcriptional levels of ETH, EH, and CCAP were calculated using 2^–ΔΔCt method. All data were expressed as means ± standard deviation (SD) and analyzed by GraphPad Prism8.0.1 (GraphPad Software Inc., San Diego, CA, United States). Statistical analysis was performed using one-way ANOVA and followed by Tukey’s analysis method with significant difference (P < 0.05). The differences on transcriptional level of genes between treatments and control groups were analyzed by Student’s t test (P < 0.05). For the analyses of molting rates, the log-rank (Mantel-Cox) test was performed and followed by Student’s t-test with Welch’s correction (P < 0.05).

Quantitative real-time PCR (qRT-PCR) results indicated that ETH expression level was the highest in the thoracic ganglion, followed by that in the eyestalks and brain of juvenile crabs (Figure 1A). EH transcripts were mainly localized in the neuronal tissues, and the level was the highest in the thoracic ganglia, followed by the brain and the eyestalks (Figure 1B). CCAP transcript level was the highest in the thoracic ganglia (Figure 1C). A progressive decrease of ETH transcript level was observed from post-molt stages (i.e., A and B) toward the early premolt stage (D0). However, a 100-folds increase in ETH transcript level occurred at the late premolt stage (D2) (Figure 2A). The transcript levels of EH and CCAP at A, B, C, and D0 stages have little changes until the sharp increase at D2 and reach the maximum level at the ecdysis stage (E) (Figures 2B,C).

When juvenile crabs at the D0 stage were injected with dsETH, dsEH or dsCCAP (knockdown condition), the percent of crabs that molt normally was seriously affected compared to the dsGFP-injected group. For dsETH and dsCCAP groups, apolysis occurred in those crabs and most crabs could advance to the late premolt stage. Also, the old exoskeleton and the new epidermis for which could be separated from the new cuticle artificially (Figures 3A,B). However, more than half of the crabs failed to initiate the molting process and eventually death occurred at D2 stage in those crabs (Figure 3C). As a comparison, 58% of the crabs injected with dsEH died before D2 stage, which was significantly higher than that in other groups (Figure 3C).

To further verify the function of molt regulation between ETH, EH, and CCAP, synthetic peptides of ETH and CCAP were co-injected with these three dsRNAs separately. Injection of ETH increased the number of successful molts and survival rate in the ETH-knockdown group from 28 to 72%. However, this has no effect on the EH-knockdown and CCAP-knockdown groups (Figure 3). Injection of CCAP increased the number of successful molts and survival rate in the CCAP-knockdown group from 42 to 72%. However, it has no effect to the EH-knockdown group. However, injection of CCAP the number of successful molts and survival rate in ETH-knockdown group from 28 to 50% (Figure 3).

To further elucidate the relationship between EH and ETH at different molting stages, transcriptional levels of genes in dsETH- and dsEH-injected crabs at D0 and D2 stages were analyzed. For crabs at the D0 stage, injection of dsETH caused a significant reduction of ETH transcript level (−30.82%). The transcript levels of EH and CCAP also decreased significantly (Figures 4A–C). Injection of dsEH to juvenile crabs at stage D0 silenced the transcription of EH to a level of 61.92% (Figure 4D). However, the transcripts level of ETH increased sharply (+360%) (Figure 4E), but there was no change in the transcript level of CCAP (Figure 4F).

When crabs at D2 stage were injected with the same amount of dsETH as at D0 stage, the transcripts level of ETH had a significant increase conversely and CCAP transcripts level was decreased significantly compared to control but the transcripts level of EH had no change (Figures 5A–C). When juvenile crabs at D2 stage were injected with the same amount of dsEH as at D0 stage, the transcripts level of EH was decreased to 23.22% compared to that of the controls (Figure 5D). In addition to the injection of dsETH, knock-down of EH could also reduce the transcripts level of ETH significantly (Figure 5E), in which the transcripts level of CCAP decreased significantly too (Figure 5F).

To verify the regulatory relationship between ETH and EH, co-injection of dsETH and dsEH were conducted to crabs at D2 stage. Results show that ETH expression was knock-down as compared to the controls, with a decrease of 68.06% (Figure 6A), and the transcriptional level of EH decreased by 77.19% (Figure 6B). Similarly, the transcriptional level of CCAP was significantly reduced than that of the control group (Figure 6C).

In insects such as Leptinotarsa decemlineata and Bactrocera dorsalis, ETH is mainly produced by Inka epithelial cells of the endotracheal gland in the trachea. The ETH expression level is the highest at the late premolt stage (Shi et al., 2017; Shen et al., 2021). In Drosophila, EH is expressed in the neuronal tissues and trachea, which participate in the regulation of molting behaviors (Scott et al., 2020). Although the crustaceans did not evolve a tracheal system, Both the Insecta and Crustacea shared a large number of homologous neuropeptides that are highly conserved in structure and functions (Veenstra, 2016; Oliphant et al., 2018). In this study, ETH, EH and CCAP are mainly expressed in thoracic ganglion in S. paramamosain (Figures 1A–C), which indicates that crab nerve tissue assumes part of the similar endocrine regulation function as the insect tracheal tissue. In S. paramamosain, the transcriptional level of ETH fluctuated dramatically with the ecdysis cycle, which indicated that ETH might be the major gene initiating molting behavior (Figure 2A). Unlike ETH, the transcripts of EH and CCAP at D2 and E stages increased gradually (Figures 2B,C). At the initial premolt stage of insects such as Manduca sexta, ETH is at a low level and could regulate the initiation of the entire molting behavior (Gammie and Truman, 1999). In the crab, the transcription level of ETH is also the lowest at D0 stage, it’s interesting to know if crab has a similar ETH regulatory mode.

Results show that ETH, EH, and CCAP are essential for the regulation of molting behavior in mud crabs. In insects, mortality occurred when ETH, EH or CCAP was interrupted. For example, EH null mutant Drosophila and EH-interrupted Tribolium, pre-ecdysis behaviors were disrupted and mass mortality occurred before molting (Arakane et al., 2008; Kruger et al., 2015). Similarly in the red flour beetles Tribolium, the dsETH- or dsCCAP-injected animals have completed material accumulation, water re-absorption, but eventually died instead of shedding the old exoskeleton (Arakane et al., 2008). Interestingly, for dsEH-injected crabs, pre-ecdysis behaviors were disrupted and crabs were unable to advance to D2 stage and died (Figures 3C,F,G). Both ETH and CCAP synthetic peptides could rescue the molting of crabs when their genes were silenced, which suggests they play key roles in crab molting. Compared to ETH and CCAP, EH appears to have a specific function on the formation of new epidermis at premolt stage as synthetic peptides for ETH and CCAP can’t offset the gene knock-down effect of EH (Figures 3D–I). Also, our results show that dsETH- or dsCCAP-injected crabs developed to D2 stage smoothly but failed to molt, at which crabs had completed the formation of new epidermis (Figure 3A). The results were similar to that of the red flour beetles. Remarkably, for the dsETH-injected crabs, injection of CCAP rescued half of them from failed molt, whereas same result didn’t see in the dsCCAP-injected crabs under the rescue of ETH synthetic peptide (Figures 3D,E,H,I). Considering that CCAP may be a downstream gene of ETH and can replace part of the role of ETH in ecdysis behaviors.

In Manduca sexta, 20-E level at early premolt stage is high, which can activate the low-level expression of ETH in the body (Zitnan et al., 2007). The release of ETH then activates the synthesis and release of EH in the brain neurons by binding to its receptor ETHR-A (Gammie and Truman, 1999; Kim et al., 2006). Similarly, the ETH and EH transcriptional levels of S. paramamosain at D0 stage were low (Figures 2A,B), the transcriptional levels of EH and CCAP were increased when ETH was knockdown (Figure 4). In Drosophila, CCAP null mutants did not affect the normal development and molting of the worms (Clark et al., 2004). Whereas the knockdown of CCAP caused failure of molt and eventually died in Tribolium (Arakane et al., 2008). Studies found that CCAP was regulated by 20E in Bactrocera dorsalis (Shi et al., 2019b), which means CCAP is involved in molting-related regulation in B. dorsalis. The above reports indicate that CCAP has functional differentiation in some species such as Drosophila in Insecta. In the present study, the strong dependence of the transcript levels of EH and CCAP on ETH indicated that EH and CCAP have positive correlations with ETH in S. paramamosain at D0 stage.

At late premolt stage of Tribolium castaneum (Arakane et al., 2008), ETH acts on the ventral neurons to release EH into hemolymph, and EH then acts on the ETH storage cells to stimulate large quantities release of ETH. The positive feedback regulation further activates a series of molting-related neuronal genes such as CCAP and bursicon, and initiates molting. However, in Drosophila, the release of ETH has no correlation with EH (Clark et al., 2004), which indicates that the positive feedback regulation mode between ETH and EH is not conserved in Insecta. Our study shows that when crabs developed to D2 stage, ETH and EH transcriptional levels rose rapidly (Figures 2A,B). A strong positive of ETH may exist at D2 stage to ensure its continuous and massive expression. The obvious change in dsETH-injected crabs didn’t cause a fluctuation of EH (Figures 5A,B), which suggests that the gene that directly regulates the retaliatory increase of ETH is not EH. The expression of ETH and ecdysis of crabs were significantly inhibited with the knockdown of EH at D2 stage (Figures 5D,E), which suggesting that although not a direct regulatory gene to ETH at D2 stage, EH plays an important role in the massive expression of ETH. To further verify the relationship between ETH and EH at D2 stage, co-injection of dsETH and dsEH to crabs at D2 stage were down (Figures 6A–C). The results indicated that when EH was disturbed, the unknown factor required for ETH’s surge was also inhibited, and EH may indirectly regulate the increase of ETH at this stage by regulating the unknown factor. Also, the transcription levels of ETH, EH, and CCAP are highest at E stage, whereas the higher levels of mRNA for EH and CCAP should not to be coincident with the highest levels of ETH because stage E lasts just a few minutes in crabs, and it would be very difficult to envisage how an increased levels of ETH or EH might influence transcription/translation of CCAP or ETH in such a short time frame. Therefore, the correlation of these three genes might not be a direct regulation but an inter-relationships with each other.