No endangered or protected species were involved in this study, and all experimental protocols and methods were approved in July 2021 (Authorization No. 20210715002) by the Animal Care and Use Ethics Committee in the Freshwater Fisheries Research Center (Wuxi, China).

Samples of female M. nipponense (n = 30, BW ± SD: 0.72 ± 0.19 g) used in gene cloning and expression analysis were obtained from Dapu scientific experimental base in Freshwater Fisheries Research Center (Wuxi, China). All prawns were adult and healthy, which were kept in recirculating glass aquarium system under the same environment for at least 1 week. Seven tissues (muscles, M; heart, H; ovary, O; cerebral ganglion, Cg; eyestalks, E; hepatopancreas, He; and gills, G) were collected from adult female prawns (n = 5). Ovarian stage division was according to a previous study (17). The ovaries of living prawn can be distinguished of different stages by color: OI (undeveloped stage, transparent), OII (developing stage, yellow), OIII (nearly-ripe stage, light green), OIV (ripe stage, dark green), and OV (spent stage, gray). Both hepatopancreas and ovaries from 5 stages were collected (n = 5).

The embryos and larvae were sampled according to the criteria of previous studies (17–19). The embryonic stages (n = 20) comprised cleavage stage (CS), blastula stage (BS), gastrulation stage (GS), nauplius stage (NS), and zoea stage (ZS). The larval stages (n = 10) were as follows: L1 (the 1st day after hatching), L5 (the 5th day after hatching), L10 (the 10th day after hatching), L15 (the 15th day after hatching), PL1 (the 1st day after metamorphosis), PL5 (the 5th day after metamorphosis), PL10 (the 10th day after metamorphosis), PL15 (the 15th day after metamorphosis), and PL25 (the 25th day after metamorphosis). All tissues were stored immediately in the liquid nitrogen for further study.

Tissues weighing 0.1 g were used for total RNA isolation. RNAiso Plus Reagent (TaKaRa, Maebashi, Japan) was used for RNA extraction. Agarose electrophoresis (1.5% gel) was used to access RNA quality, and a NanoDrop 1000 was employed to measure the concentration of RNA. An RNA PCR Kit was used to synthesize cDNA (TaKaRa, Japan).

The full-length legumain-like cDNA sequence was obtained from the five ovary stages transcriptomes of M. nipponense (NCBI accession: SAMN11603268–SAMN11603282, Bioproject PRJNA541783). In order to verify the open reading frame sequence (ORF) of legumain-like gene, a pair of primers Mn-Lel-F and Mn-Lel-R was designed (listed in Table 1 ). PCR products were detected on 1.5% agarose gel and then sent to the company for sequencing (Sangon Biotech Shanghai Co., Shanghai, China).

The legumain-like gene sequence was analyzed using BLAST programs (http://www.ncbi.nlm.nih.gov/BLAST/), and SignalP4.1 was applied in signal peptide predicting (http://www.cbs.dtu.dk/services/SignalP-3.0/). The legumain-like amino acid sequences were used to generate the phylogenetic trees with MEGA5.0 based on the neighbor-joining (NJ) method, and the bootstrapping replications were 1,000 (20, 21). Sequence alignment was built with DNAMAN 6.0 software (18). The known sequences of other animals were obtained from the NCBI database (http://www.ncbi.nlm.nih.gov/). NetNGlyc-1.0 (https://services.healthtech.dtu.dk/service.php?NetNGlyc-1.0) was employed to predict potential sites of N-linked glycosylation.

Quantitative Real-Time PCR (qPCR) was used to quantify spatiotemporal mRNA expression levels of the legumain-like gene. EIF gene (eukaryotic translation initiation factor 5A) was used as the internal reference for both tissue distribution and development stage expression analysis (listed in Table 1 ) (22). The qPCR system was detailed in a previous study (23). The expression levels were calculated based on the 2^{− ΔΔCT} method (24). SPSS 23.0 software was used to do statistical analyses, and one-way ANOVA and two-tailed t-test were used to analyze statistical differences. All data were confirmed to the homogeneity of variance and normal distribution. All qPCR data were described as mean ± SD, and a significant difference was indicated by p < 0.05.

The ovaries were dissected from mature female prawns and fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS; pH 7.4) at 4°C overnight. In situ hybridization (ISH) study was performed on 4-μm-thick formalin-fixed paraffin-embedded sections using Zytofast PLUS CISH implementation kit (ZytoVision GmBH, Bremerhaven, Germany) (25). The blank control groups were routine H&E staining sections without probe poured, and the negative control groups had antisense probes poured. The mRNA locations of Mn-Lel were analyzed with sense probe poured over slides. The anti-sense and sense probes in this study with digoxin signal were designed by Primer5 software based on the cDNA sequence of Mn-Lel and synthesized by Shanghai Sangon Biotech Company (Shanghai, China) ( Table 1 ). All slides were examined under a light microscope for evaluation.

The RNA interference (RNAi) primers of legumain-like and green fluorescent protein (GFP) gene, which contained the T7 promoter, were designed for using Snap Dragon (http://www.flyrnai.org/cgibin/RNAifind_primers.pl) (listed in Table 1 ). The GFP gene was selected as a control (26). Transcript AidTMT7 High Yield Transcription kit (Fermentas, Inc., USA) was used to synthesize in vitro dsRNA of both legumain-like and GFP.

In order to calculate the change degree of gonadosomatic index (GSI) better and more clearly, the ovary development level of experimental prawns should be relatively consistent at the beginning of the experiment. At the beginning of the RNAi experiment, most females were in the ovary stage IV. Therefore, stage IV prawns were selected to ensure a sufficient number of experimental prawns. One hundred healthy female prawns (0.83 ± 0.11 g) in stage IV were randomly divided into an experimental group and a control group (N = 50). A 17-day RNAi experiment was carried out (water temperature 25°C ± 1°C). Both experimental and control groups were injected with 4 μg/g.b.w of ds-legumain-like and ds-GFP each, and the injection site is pericardial cavity membrane (27). Ds-RNA was injected every 5 days, and five prawns in each group were randomly sampled on the 1st, 9th, and 17th days. By silencing the legumain-like gene, interference efficiency in both hepatopancreas and ovary and the GSI (gonadal weight/body weight × 100%) were calculated. Ten prawns were randomly sampled from both groups on the 1st, 9th, and 17th days, and the body and ovary weight were recorded.

The full-length legumain-like cDNA sequence of M. nipponense was 1,454 bp, named as Mn-Lel (GenBank No. OM317599), including 19 and 145 bp of 5′ and 3′ untranslated region (UTR), respectively. It had a polyadenylation signal (AATAAA) and a poly(A) tail in the 3′-UTR, indicating the integrity of the Mn-Lel gene ( Figure 1 ). The ORF of Mn-Lel was 1,290 bp and encoded 430 amino acids, and the estimated molecular mass was 48.05 kDa; the theoretical pI was 4.44. SignalP software analysis showed that Mn-Lel had a signal peptide with a length of 19 amino acid residues, and the predicted tangent point was between Ala¹⁹ and Asp²⁰. The results of BLASTP showed that the 21st to 278th amino acid residues constituted a C13 superfamily domain of peptidase. There was a conserved TGD motif (Thr¹¹⁰-Gly¹¹¹-Asp¹¹²) and His¹⁴⁰ and Cys¹⁸¹ catalytic dyad in the C13 superfamily domain. Several highly conserved cysteine residues were located in the 212th, 371th, 383th, 403rd, and 420th. We found that there were no potential N-linked glycosylation sites in Mn-Lel based on NetNGlyc-1.0.

Several amino acid sequences of legumain and legumain-like in vertebrates and invertebrates were deduced, and a phylogenetic tree was built to elucidate the phylogenetic relationships based on the NJ method ( Figure 2 ). The results showed that legumain and legumain-like were highly conserved in evolution; all legumain and legumain-like could be divided into two branches. The Mn-Lel and legumain-like protein of Macrobrachium rosenbergii (AJG06865.1) were isolated and clustered into one branch with other crustaceans. The crustacean legumains were then grouped with vertebrates and insects. The legumain from the parasite was isolated to be one branch.

Multiple sequence alignment was performed using DNAMAN 6.0 to compare Mn-Lel gene with legumain and legumain-like of crustaceans and insects in Arthropoda, including M. rosenbergii, Penaeus japonicus, Penaeus monodon, Penaeus vannamei, Haemaphysalis longicornis, and Ixodes ricinus. The homology with these species was 86.06%, 39.60%, 40.71%, 41.15%, 41.81%, 39.96%, and 37.25%, respectively ( Figure 3 ).

Tissue expression of Mn-Lel was monitored by qPCR using cDNA templates originating from the female muscles (M), heart (H), ovary (O), cerebral ganglion (Cg), eyestalks (E), hepatopancreas (He), and gills (G) tissues as displayed in Figure 4 . The Ct value of qPCR revealed that Mn-Lel was hardly expressed in eyestalks, cerebral ganglion, heart, and gills but was weakly expressed in the muscle. High expression levels were detected in the hepatopancreas and ovary. Mn-Lel was highly expressed in both these tissues, but the expression level in the hepatopancreas was nearly 250 times more than that in the ovaries (p < 0.01).

Mn-Lel expression during embryo and larva development was also investigated ( Figure 5 ). According to the Ct value of qPCR, Mn-Lel was hardly expressed in the embryo development stages (from CS stage to ZS stage), and it was also hardly expressed until the 15th day after hatching (L15) (p < 0.05). Mn-Lel expressed significantly from L5 to the 1st day after metamorphosis (PL1) and then dramatically increased PL5 by about 10 times than PL1 (p < 0.05), which reached the top. In later larval development stages (from PL5 stage to PL25 stage), the Mn-Lel showed a significant fluctuation trend (p < 0.05).

Subsequently, Mn-Lel is expressed in the hepatopancreas and ovaries during the ovary maturation cycle. According to the Ct values, Mn-Lel was highly expressed during the ovary maturation cycle that was examined. The Ct value of Mn-Lel gene in hepatopancreas maintained a remarkably high level in all five ovary stages ( Figure 6A ). It was found that Mn-Lel was higher in stages I and IV (p < 0.05), and its expression was slightly lower as detected in the other three stages (p < 0.05). In ovaries ( Figure 6B ), the Ct value of Mn-Lel gene also showed a high expression level. The Mn-Lel expression level doubled obviously from stage I to II (p < 0.05). Then its expression significantly increased from stage II to III approximately five-fold (p < 0.05) and reached the peak in stage III. Subsequently, it decreased dramatically from stage III to IV (p < 0.05).

Periodic changes of ovarian development were clearly shown in the paraffin section of Figure 7 . In stage I, the space of the ovarian cavity is large, the nucleus of the oocyte is round and large, and the protoplasm is less and transparent. In stage II, the ovarian cavity began to shrink significantly, the oocyte was closely arranged, and the nucleus is relatively small. In stage III, the ovarian cavity becomes narrow, and a layer of vacuoles appears around the oocyte. The oocyte begins to deposit yolk. In stage IV, oocyte development was completed, and the cytoplasm was filled with a large number of oocytes. In stage V, the ovary was filled with empty follicles and a few stage I oocytes. ISH was used to detect the Mn-Lel localization in ovary stages ( Figure 7 ). The results indicated that the Mn-Lel signal was detected in all five stages. The Mn-Lel signal could be found in oocytes including the nucleus, follicle cell, yolk granule, cytoplasmic membrane, and follicle membrane, while follicle membrane, cytoplasmic membrane, and nucleus were the main expressed sites. The color intensity of the signal showed that Mn-Lel is mainly expressed in the membrane of the follicle, cytoplasmic membrane, and nucleus. It showed that the signal of Mn-Lel in the ovary was significantly enhanced during stage II (primary vitellogenesis) and stage III (secondary vitellogenesis) and gradually weakened in other stages.

Mn-Lel was knocked down by a 3-week RNAi. The ovaries of both control and experimental groups were in stage IV (ripe stage, dark green) at the beginning, and as the experiment progressed, the ovaries emptied and became full again. The RNAi efficiency of ds-Mn-Lel in hepatopancreas and ovary was tested. In hepatopancreas, the Mn-Lel expression level was significantly downregulated by 92.67% on the 9th day after injection and then by 96.19% on the 17th day after injection (p < 0.01) ( Figure 8A ). In the ovary, the ds-Mn-Lel showed an obviously decreased efficiency by 81.49% on the 9th day and 99.41% on the 17th day after injection (p < 0.01) ( Figure 8B ). The GSI in both groups was also tracked and calculated ( Figure 8C ). In the beginning, the average GSI of control and experiment groups was 12.68% and 12.00%, respectively; however, the experiment progressed, and there were no obvious changes between the two groups (control: 3.91%, ds-Mn-Lel: 4.01%) on the 9th day after injection (p > 0.05). On the 17th day, the average GSI of the control group increased to 10.45%, while it was only 6.94% in the experiment group (p < 0.05).