P. trituberculatus were purchased from a local seafood market in Ningbo (Zhejiang Province, China). The crabs were anesthetized by chilling on ice before dissection. For the full-length transcriptome analysis, three immune-related tissues (hemocytes, hepatopancreas and gills) were collected from P. trituberculatus with a body weight of 262.5 g. Crabs with a body weight of 87.72 ± 8.96 g were used for tissue distributions and immune response analyses. Crabs were acclimated in rectangular tanks (60 cm × 40 cm × 25 cm) with aerated seawater (25‰), each tank containing 8 individuals, and the seawater was renewed daily. Hemocytes, hepatopancreas, gills, stomach, intestine, heart, muscle and eyestalk were harvested from five crabs used for tissue distributions. For the immune challenge experiment, crabs were injected with 100 μL V. alginolyticus (2 × 10⁸ CFU/mL) resuspended in 0.1 M phosphate buffered saline (PBS) into the arthrodial membrane of the last walking leg. All of the bacterial strains used in this study were provided by Xiamen University (Xiamen, China). Crabs injected with 0.1 M PBS served as controls. At 3 h and 6 h post injection (hpi), hemocytes, hepatopancreas and gills were collected from five randomly selected crabs from each group. All samples were immediately frozen in liquid nitrogen and stored at −80°C until the extraction of total RNA for analysis.

Total RNA was extracted using RNAiso Plus (TaKaRa, Japan) according to the manufacturer’s instructions. RNA degradation was assessed on 1% agarose gels. The NanoPhotometer^® spectrophotometer (IMPLEN, CA, USA), Qubit^®2.0 Fluorometer (Life Technologies, CA, USA) and Agilent Bioanalyzer 2100 system (Agilent Technologies, CA, USA) were used to detect the purity, concentration and integrity of the RNA samples.

Equal amounts of total RNA extracted from the three tissues were mixed to construct the library for PacBio and Illumina sequencing. The Isoform Sequencing (Iso-Seq) library was prepared according to the Iso-Seq protocol using the SMARTer PCR cDNA Synthesis Kit (Takara, Japan) and the BluePippin Size Selection System protocol as described by Pacific Biosciences (Sage science, USA). The qualified libraries were sequenced on a Pacbio Sequel platform.

Illumina sequencing library was constructed by pooling equal amounts of total RNA extracted from the three tissues. Sequencing libraries were generated using NEBNext^® Ultra™ RNA Library Prep Kit for Illumina^® (NEB, USA) following manufacturer’s recommendations and library quality was assessed on the Agilent Bioanalyzer 2100 system. The qualified libraries were sequenced on an Illumina Novaseq platform and 150 bp paired-end reads were generated.

Raw polymerase reads were processed using the SMRTlink 5.0 software to obtain sub-reads. Circular consensus (CCS) reads were generated from the sub-reads via self-correction, and CCS reads contained the 5’ primer, 3’ primer and polyA tail were considered as full-length transcript sequence. And then, full-length non-chimeric (FLNC) reads were subjected to cluster and non-redundant treatment using hierarchical n*log(n) algorithm and Arrow polishing, and polished consensus reads were obtained. Finally, LoRDEC was used to correct these polished consensus reads with short reads yielded by Illumina sequencing, and CD-HIT program was further used to remove redundancy. Through above data processing procedure, non-redundant and accuracy full-length transcripts were produced for following analysis.

For functional annotation, all transcripts were searched against the databases NCBI non-redundant nucleotide sequences (NT), NCBI non-redundant protein sequences (NR), Protein family (Pfam), Clusters of Orthologous Groups of proteins (COG), Swiss-Prot, Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG) and KEGG Ortholog database (KO). The BLAST programs with an E-value threshold of 10⁻¹⁰ were used in the NT database analysis. The Diamond BLASTX programs with an E-value threshold of 10⁻¹⁰ was used in the NR, KOG, Swiss-Prot and KEGG databases analysis. The Hmmscan programs was used in the Pfam database analysis.

Alternative splicing analysis of non-redundant transcripts were processed with the Coding GENome reconstruction Tool (Cogent v3.1, https://github.com/Magdoll/Cogent).

Identities of deduced aa sequences were determined using BioEdit software (v7.1.3). Functional domains were predicted by SMART program (http://smart.embl-heidelberg.de/).

Four CTLs, named PtCTL-1, PtCTL-2, PtCTL-3 and PtCTL-4, were selected for the immune function analysis. Extraction and qualification of total RNA was conducted as above mentioned. Reverse transcription to synthesize first-strand cDNA was carried out using PrimeScript™ RT reagent Kit with gDNA Eraser (TaKaRa, Japan) following the manufacturer’s protocols. The primers ( Table 1 ) used for cloning were designed based on the sequence of the CTLs identified from the full-length transcriptome of P. trituberculatus. The amplification products were ligated into the pMD19-T vector (Takara, Japan) and sequenced (Sangon, China).

Quantitative real-time PCR (qRT-PCR) was carried out on an ABI 7500 real-time PCR detection system (Applied Biosystems, USA) using TB Green Premix Dimer Eraser kit (Takara, Japan) according to the manufacturer’s instructions. The specific primers used for qRT-PCR were designed based on the ORF sequence of the CTLs and listed in Table 1 . The β-actin gene of P. trituberculatus were served as the internal standardization. The melting curve was recorded to confirm that only a single PCR product was amplified. Three technical replicates were conducted. The relative gene expression levels were determined by the 2^-△△Ct method (23).

The cDNA fragments encoding the mature peptides of the four PtCTLs were amplified using primers with the restriction sites ( Table 1 ). The target cDNA fragments were ligated into the expression vector pET-32a and confirmed by sequencing. The recombinant plasmids were transformed into Escherichia coli BL21 (DE3) pLysS Chemically Competent Cells (TransGen Biotech, China) and induced at 16°C for 12 h with isopropyl-β-thiogalactopyranoside (IPTG) at a final concentration of 0.5 mM. The induced bacterial pellets were collected after centrifugation (10,000 ×g) for 5 min at 4°C and then suspended in Tris-HCl buffer (50 mM Tris-HCl, 500 mM NaCl, 20 mM imidazole, pH 7.8) and sonicated on ice for 30 min with burst duration of 3 s. The supernatants were collected after centrifugation (10,000 ×g) for 25 min at 4°C. The purification of recombinant proteins was conducted using the high affinity Ni-NTA resin (GenScript, China) according to the manufacturer’s protocols. The purity of the recombinant proteins was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Coomassie Blue staining, and the recombinant proteins were further confirmed by western blot analysis as previously described (24). The concentrations of recombinant proteins were examined using BCA Protein Assay Kit (Beyotimie, China).

Gram-negative bacteria (V. alginolyticus and V. parahaemolyticus) and Gram-positive bacteria (S. aureus) were tested in the antibacterial and bacterial binding activities assay. V. alginolyticus, V. parahaemolyticus and S. aureus were suspended in TBS (50 mM Tris-HCl, 50 mM NaCl, 10 mM CaCl2, pH 7.5) at a concentration of 6.5 × 10⁷ CFU/mL, 6.8 × 10⁷ CFU/mL and 6.0 × 10⁷ CFU/mL, respectively. 50 μL of recombinant PtCTL proteins (rPtCTLs) (500 μg/mL) were mixed with the same volume of bacteria suspension and incubated at 37°C (V. alginolyticus and S. aureus) or 30°C (V. parahaemolyticus) for 2 h under slight rotation. The mixtures were diluted 10⁴-fold and spread on Luria-Bertani (S. aureus) or 2216E (V. alginolyticus and V. parahaemolyticus) agar plates, and the CFU were calculated after culturing for 8 h at 37°C/30°C. In the binding assay, the rPtCTLs and bacteria suspension were mixed and incubated as above mentioned. The bacterial pellets were harvested by centrifugation at 5,000 ×g for 3 min and washed five times with TBS. Bindings between rPtCTLs and bacteria was confirmed by western blot analysis.

All data were shown as the mean ± standard deviation (SD) and analyzed by the one-way analysis of variance (ANOVA) using the SPSS 17.0 software. 0.01 < p < 0.05 was considered statistically significant; p < 0.01 was considered to be extremely significant.

In total, 1,172,590 polymerase reads were obtained. The average length of polymerase reads was 75,948 bp, and the N50 length of polymerase reads was 141,113 bp ( Supplementary Table 1 ). A total of 64,159,133 sub-reads with the average length of 1,314 bp and N50 length of 1,900 bp were generated from polymerase reads ( Supplementary Table 1 ). Based on self-correction among sub-reads, 1,061,265 CCS reads were obtained, and the average length and N50 length of CCS reads were 2,006 bp and 2,471 bp, respectively ( Supplementary Table 1 ). Subsequently, among the CCS reads, 668,918 full-length non-chimeric (FLNC) reads with average length of 1,760 bp and N50 length of 2,266 bp were identified ( Supplementary Table 1 ). And then, 43,748 polished consensus reads with N50 length of 2,522 bp were produced from the FLNC reads ( Supplementary Table 1 ). Finally, the polished consensus reads were subjected to correction using short reads produced by Illumina sequencing and removing redundancy. In total, 17,433 nonredundant full-length transcripts with average length of 2,271 bp and N50 length of 2,841 bp were obtained ( Supplementary Table 1 ). Raw PacBio and Illumina sequencing reads are available at NCBI GenBank under accession numbers SRR12442560 and SRR12442561.

In total, 13,978 (80.18%) transcripts were annotated in at least one database. A total of 12,972 (74.41%), 6,371 (36.55%), 10,172 (58.35%), 11,296 (64.80%), 10,172 (58.35%), 10,126 (58.09%), 12,343 (70.80%) transcripts were annotated in the NR, NT, Pfam, SwissProt, GO, KOG and KEGG database, respectively ( Supplementary Table 2 ). For main species distribution matched against NR database, most of the matched transcripts were similar with Hyalella azteca (4,612), followed by Zootermopsis nevadensis (649), Portunus trituberculatus (506), Scylla paramamosain (473), Limulus polyphemus (344), Daphnia magna (332), Eriocheir sinensis (276) and Litopenaeus vannamei (209) ( Supplementary File 1 ).

The distribution of GO second-level terms showed that cellular process (4,907), metabolic process (4,506) and single-organism process (3,678) were the most represented terms in the biological process category ( Figure 1 and Supplementary File 2 ). In the molecular function category, binding (6,075) and catalytic activity (4,173) comprised the large proportion ( Figure 1 and Supplementary File 2 ). As to cellular component category, most transcripts were associated with cell (2,344), cell parts (2,344) and organelle (1,739) ( Figure 1 and Supplementary File 2 ). The results of COG annotation showed that the cluster for general function prediction only (1,837), signal transduction mechanisms (1,368) and post-translation modification, protein turnover, chaperones (1,031) represented the top three class ( Supplementary File 3 ). For the KEGG annotation, most transcripts were mapped to signal transduction (1,281) and transport and catabolism (911) related pathways ( Figure 2 and Supplementary File 4 ). Meanwhile, a total of 561 transcripts were mapped to immune system related pathways, such as chemokine signaling pathway, complement and coagulation cascades, Toll-like receptor signaling pathway, NOD-like receptor signaling pathway, RIG-I-like receptor signaling pathway and IMD signaling pathway ( Figure 2 and Supplementary File 4 ).

In total, 3,960 unique transcript models (UniTransModels) were reconstructed by 8,563 non-redundant full-length transcripts. 1,684 UniTransModels that have more than one isoform were used for alternative event analysis. As a result, a total of 315 AS events were predicted based on 244 UniTransModels and generate 483 full-length transcripts. The retained intron (RI) (167) is the most abundant type of AS events, followed by alternative 3’ splice sites (A3) (72), alternative 5’ splice sites (A5) (61), skipping exon (SE) (9), alternative first exons (AF) (4), alternative last exons (AL) (1) and mutually exclusive exons (MX) (1).

In this study, transcript variants of various immune-related genes such as pattern recognition receptors (PRRs), antimicrobial peptides (AMPs), vital molecules in prophenoloxidase (proPO)-activating system, antioxidant enzymes, and heat shock proteins (HSPs) were identified ( Table 2 ). The ORFs and deduced amino acid (aa) sequences of these transcripts are listed in Supplementary File 5 . A total of twenty-five C-type lectins (CTLs) were identified in this study. The functional domain analysis showed that the deduced aa sequences of the CTLs all contain a carbohydrate recognition domain (CRD), moreover, a signal peptide and a low-density lipoprotein receptor class A domain were also observed in some CTLs ( Figure 3A ). Homology comparisons showed that the deduced aa sequences of these CTLs shared 2.2% to 99.6% identity with each other ( Supplementary File 6 ). In total, four Toll-like receptors (TLRs) were identified, and these TLRs showed 19.8% to 94.2% identity with each other ( Supplementary File 6 ). These TLRs exhibited a typical domain architecture of their counterparts, including extracellular leucine-rich repeats (LRRs), a transmembrane domain and an intracellular Toll/interleukin-1 receptor (TIR) domain ( Figure 3B ). Herein, a total of twelve crustins, seven antilipopolysaccharide factors (ALFs) and fourteen hyastatin were identified. Homology comparisons showed that the deduced aa sequences of the crustins, ALFs and hyastatin shared 7.2% to 99%, 4% to 95.7% and 24% to 93.1% identity with each other, respectively ( Supplementary File 6 ). In the current study, twenty-four proPO-activating factors (PPAFs) and twenty-one serine proteinases transcript variants were identified. Homology comparisons showed that the deduced aa sequences of the PPAFs and serine proteinases shared 10.1% to 98.6% and 7.1% to 96% identity with each other, respectively ( Supplementary File 6 ). Moreover, several antioxidant enzymes, including twelve copper/zinc SODs (Cu/ZnSODs), two cytosolic manganese SODs (CytMnSODs), three CATs, three Gpxs and fifteen GSTs were identified. Additionally, we also identified several HSPs, including a HSP10, two HSP40, two HSP60, six HSP70 and two HSP90.

The cDNA sequences of PtCTL-1, PtCTL-2, PtCTL-3, and PtCTL-4 were deposited in GenBank with the accession number OL848444, OL848445, OL848446 and OL848447, respectively. The four PtCTLs encoded a polypeptide of 161 aa (18.08 kDa), 265 aa (29.68 kDa), 175 aa (20.62 kDa) and 164 aa (18.45 kDa), respectively. All of the four PtCTLs contained a single CRD with a characteristic carbohydrate binding motif (a QPN motif in PtCTL1, an EPS motif in PtCTL2, a FPR motif in PtCTL3 and an EPD motif in PtCTL4) and six cysteine residues (Cys²³, Cys⁴⁰, Cys⁵⁷, Cys¹²¹, Cys¹⁵³ and Cys¹⁵⁶ in PtCTL1; Cys¹³⁶, Cys¹⁴⁷, Cys¹⁶⁴, Cys²³¹, Cys²⁴⁷ and Cys²⁵⁵ in PtCTL2; Cys³³, Cys⁴⁴, Cys⁶¹, Cys¹³⁴, Cys¹⁵⁰ and Cys¹⁵⁸ in PtCTL3; Cys²⁰, Cys³⁵, Cys⁵², Cys¹¹⁹, Cys¹³⁵ and Cys¹⁵⁶ in PtCTL4) ( Figure 4 ). Additionally, the PtCTL1, PtCTL2 and PtCTL3 also contained a signal peptide (Met¹-Ala²¹ in PtCTL1, Met¹-Ala¹⁶ in PtCTL2 and Met¹-Gly²² in PtCTL3) in the N-terminus ( Figures 4A–C ).

Tissue distribution analysis showed that the four PtCTLs were expressed in all tested tissues and showed relatively higher expression levels in the hepatopancreas and gills ( Figure 5 ). Significant induction in the transcription of PtCTL-1, PtCTL-2, and PtCTL-4 were observed in the hemocytes, hepatopancreas and gills of P. trituberculatus challenged with V. alginolyticus, and significant expression of PtCTL-3 was detected in hemocytes and gills when V. alginolyticus challenged the P. trituberculatus ( Figure 6 ). The transcriptional level of PtCTL-1 in the hemocytes was up-regulated 4.59-fold (P<0.01) at 3 hpi ( Figure 6A ), and it was up-regulated 4.50-fold (P<0.01) at 6 hpi ( Figure 6B ) in the hepatopancreas. The transcription of PtCTL-1 in the gills was up-regulated 2.07-fold and 2.65-fold (P<0.05) at 3 and 6 hpi ( Figure 6C ), respectively. The transcripts of PtCTL-2 in hemocytes increased 19.81-fold and 5.57-fold (P<0.05) at 3 hpi and 6 hpi ( Figure 6D ), respectively. The transcripts of PtCTL-2 in hepatopancreas increased 5.73-fold (P<0.05) at 3 hpi ( Figure 6E ), and it was up-regulated 1.98-fold (P<0.05) at 6 hpi in gills ( Figure 6F ). The mRNA expression of PtCTL-3 in hemocytes was significantly up-regulated 7.18-fold and 2.83-fold (P<0.01) at 3 hpi and 6 hpi, respectively ( Figure 6G ). The expression level of PtCTL-3 achieved a 1.84-fold (P<0.05) increment in gills ( Figure 6I ), while it was not changed in hepatopancreas ( Figure 6H ). The transcription of PtCTL-4 in hemocytes ( Figure 6J ) and hepatopancreas ( Figure 6K ) increased 3.23-fold and 4.01-fold, respectively. The transcriptional level of PtCTL-4 in gills was up-regulated 2.10-fold (P<0.05) and 2.62-fold (P<0.01) at 3 hpi and 6 hpi, respectively ( Figure 6L ).

Recombinant proteins of the four PtCTLs were successfully produced and purified. Coomassie Blue staining ( Figure 7A ) and western blot ( Figure 7B ) showed that the recombinant proteins exhibited the expected sizes of approximately 23 kDa (His-Trx), 35 kDa (rPtCTL-1), 34 kDa (rPtCTL-2), 35 kDa (rPtCTL-3) and 36 kDa (rPtCTL-4). Bacterial binding activity assays showed that rPtCTL-1 and rPtCTL-3 could bind to V. alginolyticus, V. parahaemolyticus and S. aureus ( Figures 7C–E ). rPtCTL-2 only exhibited binding activity to V. alginolyticus ( Figures 7C–E ), while rPtCTL-4 could not bind to any tested bacteria ( Figures 7C–E ). The antibacterial activity assay showed that the four rPtCTLs displayed various antibacterial spectra against the tested bacteria. The rPtCTL-1 and rPtCTL-3 exhibited antibacterial activity against V. alginolyticus, V. parahaemolyticus and S. aureus ( Figures 7F–H ). The CFU of V. alginolyticus treated with rPtCTL-1, rPtCTL-2 and rPtCTL-3 significantly decreased 2.41-fold, 5.35-fold and 18.88-fold, respectively. Compared with the blank control group, the CFU of V. parahaemolyticus treated with rPtCTL-1 and rPtCTL-3 were significantly decreased by 3.29-fold and 2.72-fold, respectively. The CFU of S. aureus treated with rPtCTL-1 and rPtCTL-3 significantly decreased 3.32-fold and 3.79-fold, respectively, compared with the blank control group. No obvious antimicrobial activity of rPtCTL-4 was observed ( Figures 7F–H ).