Adult M. coruscus were collected from Gouqi island (30°72’N; 122°77’E), Zhoushan, Zhejiang Province, China. Spawning and larvae culture methods were performed as previously described (Li et al., 2020). Adult mussels were washed in seawater (salinity: 30 ppt) and induced for spawning by ice shock. Mussels that started spawning were transferred to 2L glass beakers. Fertilization was carried out by gently mixing eggs and sperms in filtered seawater (FSW; acetate-fibre filter: 1.2 μm pore size). The excess sperms were filtered through a nylon plankton net (mesh size: 20 μm) and kept them undisturbed in FSW at 18°C for two days. The seawater was renewed every other day. Larvae were fed with microalgae Isochrysis zhanjiangensis and maintained at a density of 5 larvae mL⁻¹ at 18°C. Larvae samples were collected at 1 dpf (trochophore larvae), 2 dpf (D-veliger larvae), 15 dpf (umbo larvae) and 41 dpf (pediveliger larvae) and relaxed with 1M MgCl₂ in FSW, then collected into ice-cold 4% PFA/1×PBS (pH 7.4) for 2h. Samples were replaced with fresh 4% PFA/1×PBS and kept at 4°C overnight. Samples were washed in 1×PBS + 0.1% Tween 20 (PBT) and progressively dehydrated with methanol (MeOH)/PBT (100% MeOH to 0% MeOH) and transferred to 100% MeOH and stored at −20°C for whole-mount in situ hybridization (WISH) and whole-mount immunohistochemistry (IHC). The experimental tests and animal handling procedures performed in this study were approved by the Institutional Animal Care and Use Committee (IACUC) of Shanghai Ocean University with the registration number of SHOU-DW-2018-013.

Fragments of 365 bp of McDx (Accession number: MW928628) and 384 bp of McDy (Accession number: MW928627) genes were amplified from larval cDNA using gene-specific primers containing a T7 promoter sequence (5′- TAATACGACTCACTATAGGG-3′). A T7 RNA polymerase promoter was added in the reverse primer (for antisense probe) or forward primer (for sense probe) for amplifying the probe template. Digoxigenin-labeled sense and anti-sense riboprobes were synthesized by in vitro transcription using purified polymerase chain reaction products as templates using MAXIscript™ T7 (Thermo Fisher, USA) and DIG RNA labeling Mix (Roche), and stored at -80°C. Hybridization with sense riboprobes was conducted as a negative control in the assay.

For the WISH, all steps were performed at room temperature unless otherwise indicated. Larvae samples in 100% MeOH were gradually hydrated by continuous washing in MeOH/PBT (100% MeOH to 0% MeOH) for 5 min. Samples were digested with proteinase K (20 μg mL^-1) in PBT and hybridized with 0.5-1 μg mL^-1 DIG-labeled cRNA probe at 65°C overnight. Specimens were stained with 1/5000 anti-digoxigenin antibody conjugated to alkaline phosphatase (Roche) overnight at 4°C. Samples were washed in PBT and then incubated in NBT (nitro blue tetrazolium)/BCIP (5-Bromo-4-chloro-3-indolylphosphate, Roche) for color development in the dark and washed in PBT to stop the reaction. Images were captured with a Zeiss Axio Imager M2 microscope coupled to an axiocam 305 digital camera.

Larvae samples in 100% MeOH were hydrated as described above and washed with 1×PBS. Hydrated samples were decalcified with 5% EDTA in PBS for 30 min at room temperature and then preincubated with blocking solution (0.25% BSA+0.03% NaN₃+1% Triton X-100+1×PBS) at 4°C for 2-4 h to eliminate non-specific binding sites. The specimens were then washed in 1×PBS+1% Triton X-100 (PBTr) and incubated with primary antibodies at a final dilution of 1:1000 in the blocking solution for 2 days at 4°C using a polyclonal antibodies against FMRFamide or 5-HT (Immunostar, Hudson, WI, USA). Fluorescently tagged AlexaFluor 488 goat anti-rabbit IgG (sigma, USA) was used to detect the primary antisera. After the whole mount-staining, pediveliger larvae were washed with PBTr and stored in 100% glycerol for image acquisition using a Leica TCS SP8 confocal laser scanning microscopes.

The siRNA probe for McDx, nucleotides 240-258 across the open reading frame (ORF; Supplementary Figure 1), and siRNA probe against McDy, spanned nucleotides 549-567 across the ORF (Supplementary Figure 2). The siRNA oligonucleotide sequences specific to McDx (5’-ccauaucgguagagaaauu-3’) or McDy (5’-gccguuuaugguuaagcuu-3’) were confirmed by BLAST search of an in-house M. coruscus transcriptome database. The double-stranded siRNA was synthesized by GenePharma Co. Ltd. (Shanghai, China). The non-sense siRNA (5’-uucuccgaacgugucacgu-3’) was not annealed to non-target gene transcripts of M. coruscus transcriptome database (Li et al., 2020).

siRNA transfection in pediveliger larvae of M. coruscus was conducted following the method described previously (Li et al., 2020). In summary, the pooled pediveliger larvae were transferred to 1.5-ml tubes containing 1 ml autoclaved filtered seawater (AFSW) and 0.4 or 0.8 μg target or NC siRNA. An untreated control group without adding siRNA was set up following the same procedure for transfection.

The siRNA/larvae electroporation was carried out in Gene Pulser Xcell (Bio-Rad, USA) with square-wave pulses in a 0.4 cm electroporation cuvette. The electroporation protocol was used as follows: one pulse of 100 V was administered for 5 ms, followed immediately by 10 pulses of 50 V for 20 ms (with a 1 s interval between pulses). After electroporation, the siRNA-larva mixture was left undisturbed at room temperature for 10 minutes and then transferred to10 L polycarbonate tanks (containing 5 L AFSW at 20°C). 1440 electroporated larvae were maintained in 5 L AFSW for 24 h prior to larval metamorphic bioassays. The larvae electroporated with siRNA (Dx, Dy or NC) or without siRNA (control, C) were used for qPCR analysis and were cultured for 48 hours without feeding prior to sampling. The experimental set-up is shown in Table 1.

Larval metamorphosis bioassays were carried out according to the methods described previously (Yang et al., 2014; Li et al., 2020). Epinephrine (EPI) was used as a strong inducer for larval metamorphosis based on previous results in M. coruscus (Yang et al., 2014; Li et al., 2020). 1440 electroporated larvae were used for the metamorphosis bioassays. Nine replicates of each experimental treatment group were carried out. Treated larvae were transferred into Petri dishes (Ø 60 mm × 19 mm height) containing 20 pediveliger larvae and 20 mL test solution (10^–4 M EPI) for 96 h. The EPI test solution was not changed during the exposure. Blank control (BC) was incubated in 20 mL AFSW by twenty larvae without electroporation or exposure to EPI. The maximum metamorphosis rate was determined using pediveliger larvae exposed to 10^–4 M EPI, which was set up as a positive control (PC). All assays were conducted in a dark environment at 18 ± 1°C. The larvae of each treatment group were observed daily under an Olympus stereoscopic microscope, and the dead larvae were removed and counted. Post-larvae lost their velum and exhibited post-larval shell growth (Li et al., 2020).

Total RNA was extracted from five independent larval pools of treatment groups using RNAiso Plus reagent following the manufacturer’s protocol (Takara, Japan) and genomic DNA was removed by a PrimeScript™ RT kit with gDNA eraser (Takara, Japan) and cDNA synthesis was performed using the PrimeScript Reagent Kit (Takara, Japan) with DNase-treated RNA.

Quantitative real-time PCR (qPCR) reaction was performed duplicate and each reaction containing 1 μL template cDNA, 0.3 μL of each forward and reverse primers (10 mM, Table 2), 5 μL of 2 × FastStart Essential DNA Green Master (Roche) and sigma water to give a final reaction volume of 10 μL. All samples were carried out in 96 multi-well plates using a LightCycler 960 (Roche) with five biological replicates and two technical replicates per sample were carried out. Gene expression was determined by the absolute quantification method using gel isolated amplicons as standards as previously described (Li et al., 2021). The qPCR amplification protocol was carried out as follows: the initial denaturation step for 10 min at 95°C, followed by 45 cycles of 10 s at 95°C and 10 s at 60°C. A melting curve analysis confirmed that a single reaction peak was obtained in the qPCR reaction.

Data were analysed for statistical difference by using JMP™ software. p < 0.05 was the cut-off for statistical difference. The percentage of post-larvae (metamorphosis data) was arcsine-transformed and then tested for normality and homogeneity. A Wilcoxon/Kruskal Wallis test was used to assess the gene expression and larval metamorphosis data.

Transcripts localization of McDx and McDy was examined by WISH in four developmental larval stages (trochophore, D-veliger, umbo and pediveliger larvae) (Figure 1). At the trochophore larvae stage, the expression of McDx and McDy was both detected in telotroch and prototroch (Figures 1A, E; Supplementary Figure 3A). The McDx was also expressed in the central region of the shell field of the larva (Figure 1A). In the stage of the D-veliger larvae, the expression of McDx and McDy was mainly expressed in the mantle region (located at the edge of the shell) (Figures 1B, F). At the umbo larvae stage, McDx and McDy transcripts were detected in the visceral tissues region (Figures 1C, G). In the pediveliger larvae stage, the distinct expression patterns of McDx and McDy transcripts were observed, showing the functional differentiation among two deiodinase genes (Figures 1D, H). McDx was extensively expressed in the visceral tissues and medial region originating from the apical to the caudal region of the pediveliger larvae (Figure 1D). The expression of McDy was found in the edge of the velum (the organ responsible for motor and feeding functions) and mantle region correlated with shell development (Figure 1H; Supplementary Figure 3B).

To understand the McDx signal localized in the median body region of pediveliger larvae, IHC with the antibodies against FMRFamide and 5-HT, markers for the ganglia and developing nervous system, were conducted to visualize the neural development (Figure 2). The FMRFamide immunoreactive (FMRFamide-ir) cell appeared at three pairs of ganglia (cerebral, pedal and visceral) and interconnecting connectives and commissures (Figures 2A, B). The 5-HT-like immunoreactive (5-HT-lir) cells were detected in the cerebral ganglia and were seen to innervate the velum and ventral part of the body, including the mantel area and regions of the pedal and visceral ganglia (Figures 2C, D). The expression of McDx in the median body region of pediveliger larvae likely corresponds to the position of 5-HT-lir neurons (Figures 1D, 2C, D).

The expression of McDx and McDy were significantly reduced in electroporated pediveliger larvae exposed to 0.4 μg mL^-1 and 0.8 μg mL^-1McDx siRNA or McDy siRNA (0.4 McDx and 0.8 McDx; 0.4 McDy and 0.8 McDy) compared to the control group (C) and the non-sense siRNA negative controls (0.4 NC and 0.8 NC, p < 0.05, Figures 3A, 4A). In contrast, there was no significant difference between the non-sense siRNA negative controls and the control group (p > 0.05, Figures 3A, 4A).

No metamorphosis occurred in the electroporation larvae that were not exposed to siRNA and treated with EPI (BC, Figures 3B, 4B). 44% of pediveliger larvae underwent metamorphosis induced by EPI after electroporation without siRNA (PC, Figures 3B, 4B). Electroporation of pediveliger larvae using 0.8 μg mL^-1McDx siRNA significantly reduced larval metamorphosis after 96 h of EPI exposure relative to 0.8 μg mL^-1 non-sense siRNA electroporation and the PC group (p < 0.05, Figure 3B). Larval metamorphosis was significantly inhibited in electroporated pediveliger larvae using 0.4 μg mL^-1 and 0.8 μg mL^-1McDy compared to 0.4 μg mL^-1 and 0.8 μg mL^-1 of NC groups, respectively (p < 0.05, Figure 4B).

Metamorphosed post-larvae were firstly found in 48 h after being treated with EPI, and the number of post-larvae increased gradually with increasing exposure time to 96 h (Figures 3C, 4C). The electroporation of 0.4 μg mL^-1 and 0.8 μg mL^-1McDx siRNA significantly reduced the larval metamorphosis rate by 32% (p < 0.05) and 45% (p < 0.001) compared to the PC group, respectively (Figure 3C). The larval metamorphosis rate of exposure to 0.4 μg mL^-1 and 0.8 μg mL^-1McDy siRNA significantly reduced by 38% (p < 0.05) and 49% (p < 0.001) compared to the PC group, respectively (Figure 4C). Larval viability from 48 h onwards after electroporation significantly declined (p < 0.05) in 0.8 μg mL^-1 of NC groups compared to other groups (Figure 4D).