Bighead catfish (Clarias macrocephalus Günther, 1864), North African catfish (Clarias gariepinus) and their hybrid catfish (C. macrocephalus × C. gariepinus), 90–120 g in body weight, were obtained from the Department of Aquaculture, Faculty of Fisheries, Kasetsart University, Thailand. The fish were acclimatized in a quarantine tank with aerated freshwater at temperatures between 28 and 31°C for 2 weeks before the start of the experiment. The experimental procedures performed with aquatic animals were carried out in accordance with the Ethical Principles and Guidelines for the Use of Animals National Research Council of Thailand for the care and use of animals for scientific purposes. The protocol was approved by the Animal Ethics Committee, Kasetsart University, Thailand (Ethics ID: ACKU61-FIS-004).

Catfish genomic (g) DNA was isolated from whole blood tissues using a QIAamp DNA Blood and Tissue Mini Kit (QIAamp, CA, USA) according to the manufacturer’s protocols.

The PBLs of catfish are target tissues used to isolate total RNA and for cDNA synthesis. Total RNA was analyzed using NucleoZOL™ reagent (Clontech Laboratories, CA, USA) according to the manufacturer’s instructions. The obtained total RNA was then used as templates for first strand cDNA synthesis using the protocol described for the Thermo Scientific RevertAid Reverse Transcriptase Kit (Thermo Fisher Scientific, MA, USA). The products of the gDNAs and first-strand cDNA synthesis were stored at -80°C for further experiments.

Degenerative primers were first designed to amplify the internal constant domain region of cDNA encoding Ighμ genes based on the highly conserved regions of Ighμ genes from the NCBI nucleotide database of closely related species of catfish, including Ictalurus punctatus (M27230), Hemibagrus macropterus (JF909893), Silurus meridionalis (KJ659069) and Pelteobagrus fulvidraco (JN202623). The primers were Cla_Ighμ_intr_f: 5′-GTYTMCMSYDTGGCARTGCGGCBC-3′ and Cla_Ighμ_intr_r: 5′-GARVYCTCTGGTGGAGSGAGCAMG-3′ with an approximate amplicon size of 950 bp. Reverse transcription PCRs (RT-PCRs) were carried out using Phusion High-Fidelity DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocols. The PCR cycling conditions are one cycle of 95°C for 5 min, 30 cycles of 95°C for 30 sec, 55°C for 30 sec, and 72°C for 90 sec, followed by a final extension at 72°C for 5 min. Then, PCR products encoding Ighμ genes were ligated to the pJET1.2/blunt cloning vector (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocols. Nucleotide sequencing of the recombinant plasmid was performed by the Macrogen sequencing service (Macrogen Inc., Seoul, South Korea) using pJET1.2 forward and reverse sequencing primers.

The 5′- and 3′- internal constant domain sequences of cDNA encoding Ighμ genes were recovered by rapid amplification of cDNA ends (RACE) PCR techniques using 5′- and 3′-SMARTer^® (Clontech, Mountain View, CA, USA), according to the manufacturer’s protocol, with SMARTer^® universal primers and gene-specific primers that were provided from the kits and designed from 5′- or 3′- internal constant domain sequences, respectively (Cla_Ighμ_5RACE_r: 5′-GTGCCGCTCGCATCCTTCCAAACG-3′ and Cla_Ighμ_3RACE_f: 5′-GGCTCAACTTCTCCAGTTAAGTG-3′). The 5′ and 3′ RACE-PCR products were cloned into the pJET1.2/blunt cloning vector (Thermo Fisher Scientific, Waltham, MA, USA) and sequenced using the Macrogen sequencing service (Macrogen, Inc., Seoul, South Korea) as described above.

To obtain and characterize the Ighμ genes of catfish, we used catfish gDNA to amplify the Ighμ genes of catfish using forward and reverse specific primers designed from the highly conserved region of the 5′ end of Cμ₁ and the 3′UTR of full-length Ighμ cDNAs, respectively. The forward primer was Cla_Ighμ_Cμ_{1_}f: 5′-GAACGTCGGTGACCGTAACTTCA-3′ for all Clarias catfish species. The reverse primers were Cla_mac_Ighμ_3UTR_r: 5′-GAACACACAAGCATCAGACAGACTG-3′ for big head catfish and hybrid catfish and Cla_gar_Ighμ_3UTR_r: 5′-CAAGAACACACAAGCATCAGACAG-3′ for North African catfish and hybrid catfish.

PCRs were carried out using Phusion High-Fidelity DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocols. The PCR cycling conditions were one cycle of 95°C for 5 min, 30 cycles of 95°C for 30 sec, 60°C for 30 sec, and 72°C for 5 min, followed by a final extension at 72°C for 10 min. The PCR products of Ighμ genes were cloned and sequenced as described above.

The obtained nucleotide sequences of Ighμ genes were characterized, analyzed, assembled, and aligned using bioinformatic software DNA sequences including GENETYX version 7.0, BLASTN (https://blast.ncbi.nlm.nih.gov/), BLASTX (https://blast.ncbi.nlm.nih.gov/), ExPASy (https://www.expasy.org/), DAS-Transmembrane Prediction server (www.sbc.su.se/~miklos/DAS/), MatGAT program version 2.0 (http://www.bitincka.com/ledion/matgat), and MEGA version 7.0 (http://www.megasoftware.net). Furthermore, the sequences were aligned with those of related Ighμ gene sequences in other vertebrate species using ClustalW. Phylogenetic trees were constructed using Molecular Evolutionary Genetics Analysis (MEGA) software, version 7.0 (Proprietary freeware, Japan) with neighbor-joining (NJ) algorithms with a bootstrap of 1000 replications.

The protein structures of Ighμ molecules of catfish were determined to predict the possibility of constant domain structure using the structural bioinformatics web server SWISS-MODEL workspace integrates programs (https://swissmodel.expasy.org) according to the programs’ procedures (36). The amino acids in the constant domain of each Ighμ molecule were used to generate the target sequence for protein structure prediction.

Metaphase chromosomes of catfish were obtained from the head kidney of each catfish species treated with colchicine according to Karami et al. (2015) (37). Prepared tissues were stored in fixative solution at -20°C until use. The chromosomal spreads were analyzed by staining with 15% Giemsa for 45 min and examined under a light microscope (Olympus, MA, USA). The Ighμ gene probes for fluorescence in situ hybridization (FISH) used full-length Ighμ genes containing the constant domain Cμ₁ to the 3′UTR of the Ighμ genes of each catfish species. DNA probes were performed according to standard procedures of FISH Tag detection kits (Thermo Fisher Scientific, MA, USA). Metaphase chromosome hybridization was carried out following a previously described method (38). Chromosomes and specific probes were performed with DAPI and Alexa Fluor 488 or 594 dyes following the manufacturer’s instructions (Thermo Fisher Scientific, MA, USA). Digital micrographs were recorded and processed using a Nikon/DigitaL Eclipse C2Si confocal microscope (Nikon Instruments Inc., NY, USA).

The 5′ VH region of first-strand cDNAs encoding Ighμ genes was synthesized from total RNA of PBLs of each catfish species using a 5′ RACE-Ready cDNA synthesis kit (Clontech, Mountain View, CA, USA), as previously described. PCR was amplified using Phusion High-Fidelity DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA) with the 5′ universal primer mix (UPM) and gene-specific primers (GSPs) from the kit and nucleotide sequences located in the 5′ region of the Cμ₁ constant domain of Ighμ genes (Cla_Ighμ_Cμ1_r: 5′-GTGCCGCTCGCATCCTTCCAAACG-3′). The PCR cycling conditions were one cycle of 95°C for 5 min, 25 cycles of 95°C for 30 sec, 68°C for 30 sec, and 72°C for 120 sec, followed by a final extension at 72°C for 5 min. The 5′ RACE VH PCR products of Ighμ cDNA were ligated to the pJET1.2/blunt cloning vector (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. A different hundred recombinant clones were randomly selected for sequencing by the Macrogen sequencing service (Macrogen Inc., Seoul, South Korea) as previously described.

The 5′ VH nucleotide sequences of Ighμ genes were trimmed to three separated regions encoding the V_H , D_H and J_H genes using an Ig variable domain sequence analysis tool (NCBI IgBLAST, https://www.ncbi.nlm.nih.gov/igblast/). The nucleotide sequences of each gene region were classified at the family level based on 80% similarity of nucleotide sequences using the matrix global alignment tool (MATGAT 2.0).

CDR nucleotide regions of each variable domain nucleotide sequence were first characterized using IgBLAST (https://www.ncbi.nlm.nih.gov/igblast/). The D_H segments were predicted using the VDJsolver 1.0 server (https://services.healthtech.dtu.dk). The CDRs were analyzed for their characterization, including CDR length distribution and amino acid composition in each position, using the High V-QUEST IMGT tool (39).

The variability plots of variable domain nucleotide sequences were analyzed using the Shannon (1948) (40) and Kabat and Wu (1971) (41) methods. The variability plots were illustrated using the online bioinformatics software protein variability server (PVS, https://imed.med.ucm.es/PVS/pvs-help.html).

The tissue distribution of Ighμ genes of catfish was addressed in both normal fish and fish that underwent Aeromonas hydrophila challenge. Fifty acclimatized catfish were randomly transferred into tanks containing 500 L of water. Catfish were used to determine the expression of the Ighμ gene in normal fish and infectious fish. The infectious catfish were intraperitoneally injected with 0.1 mL of virulent A. hydrophila AQH0018 bacterium (1 × 10⁶ CFU/mL). A virulent A. hydrophila AQH0018 bacterium was grown in TSB medium at 32°C for 18 hr, and cell pellets were harvested by centrifugation at 2,500 rpm for 10 mins, washed and resuspended in 0.85% NaCl prior to injection. The effective doses were preliminarily determined to validate the optimum dose throughout the median lethal dose (LD₅₀) assay (42). Sixteen organs, brain (BR), dendrite (DR), gall bladder (GB), gills (GIL), head kidney (HK), heart (HR), intestine (IN), liver (LI), muscle (MC), ovary (OV), peripheral blood lymphocytes (PBL), skin (SKN), spleen (SPL), stomach (STO), testes (TES) and trunk kidney (TK), were collected from three fish of each catfish before injection (normal fish) and post injection (infectious fish) at 0, 12, 24, 36, 48, 60, 72, 84, 96, 108, 120, 132, 144, 156 and 168 hr.

Total RNA and first-strand cDNA synthesis from sixteen organs of catfish were isolated using NucleoZOL™ reagent (Clontech Laboratories, CA, USA) and a Thermo Scientific RevertAid Reverse Transcriptase kit (Thermo Fisher Scientific, MA, USA) following the manufacturer’s instructions.

Quantification of Ighμ gene expression was performed using quantitative reverse transcriptase PCR (qRT-PCR) assays. qRT-PCR assays were performed with Brilliant III Ultra-Fast SYBR^® Green (Agilent, CA, USA) in Mx3005P QPCR Systems (Agilent, CA, USA). qPCRs of the Ighμ gene and β-actin genes of each fish species were conducted using the specific primers Cla_Ighμ_qpcr_{_}f: 5′-TGGACTGAGCTACGTTTGGAAGGA-3′ and Cla_ Ighμ_ qpcr_r: 5′-CGCCTGACTCACTGAGGAGTACTT-3′ with an amplicon size of 167 bp and Cla_b-actin_qpcr_{_}f: 5′-GTCCGTGACATCAAGGAGAAGCTC-3′ and Cla_b-actin_qpcr_{_}r: 5′-GGACTCCATACCCAGGAAAGATGG-3′ with an amplicon size of 189 bp for the Ighμ and β-actin genes, respectively. In addition, the expression distribution of two novel transcripts including Cm_mIgM2 and Cm_mIgM3 were performed in sixteen organs of healthy bighead catfish. The specific primers are Cm_mIgM2_f: 5′-GCACAGAGTTCACCTGCAATG-3′ and Cm_mIgM2_r: 5′-AAATGTTAAGGCATACAGACC-3′ with an amplicon size of 154 bp, and Cm_mIgM3_f: 5′-CCACGGATACGTCTGGAGAA-3′ and Cm_mIgM3_r: 5′-ACTCGCCTTGACTCTCACTG-3′ with an amplicon size of 163 bp.

qRT-PCR cycling conditions were one cycle of 95°C for 5 min, 40 cycles of 95°C for 30 sec, 60°C for 30 sec, and 72°C for 90 sec, followed by a final extension at 72°C for 10 min. The expression of the β-actin gene was used as the housekeeping gene to standardize the results by eliminating variation in mRNA and cDNA quantity and quality. The relative expression of the Ighμ gene in catfish tissues was calculated using 2 ^-ΔΔCT analysis according to the protocol by Schmittgen and Livak (2001) (43). The brain RT was used as calibrator. All reactions were done in triplicate.

The relative expression levels of Ighμ genes were calculated using β-actin as a reference. All quantitative data are presented as the mean ± standard deviation (SD). Statistical analysis at each time course was performed using SPSS statistics 24.0 software using one-way analysis of variance (ANOVA) and Duncan’s new multiple range test (DMRT). The level of statistical significance between species in different organs is indicated as * (P<0.05), ** (P<0.01) and *** (P<0.001) using Student’s t-test.

Through analysis of the cDNA transcripts encoding constant regions of the Ighμ genes in bighead catfish, North African catfish and their hybrid catfish, one hundred positive clones of each catfish were first cloned and sequenced to obtain the internal nucleotide sequences of the constant regions of the Ighμ genes. A single band of PCR fragment was shown for each catfish species. Sequence analysis indicated no variation in the nucleotide sequences of individual bighead catfish, North African catfish and their hybrid catfish. There were four exons encoding Cμ_s [Cμ_1(partial)-Cμ₂-Cμ₃-Cμ_4(partial)] of 954 and 949 bp for bighead catfish and North African catfish, respectively ( Figure 1A ), compared to the published Ighμ sequence of channel catfish (Ictalurus punctatus). In the hybrid, two different nucleotide sequences were observed; each sequence exhibited a 100% match with the Ighμ sequence of one of the parents. Among a hundred clones from hybrid catfish, the Ighμ nucleotide sequences shared a distribution of 28% and 72% to bighead catfish and North African catfish, respectively ( Figure 1A ).

The 5′ and 3′ of the Ighμ gene constant regions in each catfish were consequently cloned and sequenced using 5′ and 3′ RACE techniques with specific primers designed from the 5′ end of the Cμ₂ exon and 3′ end of Cμ₁ for recovering the 5′ and 3′ nucleotide regions of Ighμ cDNAs, respectively. The obtained 5′ and 3′ nucleotide sequences of Ighμ genes corresponded to the Cμ₁-Cμ₂-Cμ₃ and Cμ₃-Cμ₄-TM-polyA tails, respectively ( Figure 1A ). No variation difference in the nucleotide sequences was observed in the 5′ Ighμ constant region exon in bighead catfish and North African catfish. In addition, the hybrid catfish exhibited two different nucleotide sequences, one from each parent ( Figures 1A, B ). There was no variation in the internal nucleotide sequences of the constant regions of the Ighμ genes in all studied catfish. ( Figure 1C ). However, the sequences of the 3′ constant regions consisted of two major different Ig forms, (mIg and sIg), which are controlled by alternate mRNA processing events in certain species. Up to six nucleotide sequence patterns of 3′ region exons were observed in bighead catfish ( Figure 1D ): five patterns for mIg and one pattern for sIg. Among the five patterns of mIg, three corresponded to Cμ₁-Cμ₂-Cμ₃-TM1-TM2 with differences in the polyadenylation signal region (polyA signal: TM-2A, TM-2B and TM-2C, respectively) of the untranslated 3′ sequence, one corresponded to Cμ₁-Cμ₂-Cμ₃-Cμ₄-TM1-TM2 containing the polyA signal TM-2C, and one corresponded to Cμ₁-(non-Ighμ exons)-Cμ₂-Cμ₃-TM₁-TM₂ containing the polyA signal TM-2A. One sIg was encoded by exon Cμ₁-Cμ₂-Cμ₃-Cμ₄ and the polyA signal SC-1A ( Figures 1D, E and Supplementary Figures 1 – 6 ).

Three nucleotide sequence patterns were observed in North African catfish, and they corresponded to two patterns for mIg and one pattern for sIg ( Figure 1D ). Two patterns of mIg corresponded to Cμ₁-Cμ₂-Cμ₃-TM1-TM2 with differences in the polyadenylation signal region (polyA signal: TM-2A and TM-2B, respectively) of the untranslated 3′ sequence. Another pattern corresponded to sIg that contained four exons coding Cμ₁-Cμ₂-Cμ₃-Cμ₄ and the polyA signal SC-1A ( Figures 1D, F and Supplementary Figures 7 – 9 ).

Two nucleotide sequence patterns of the 3′ region exon corresponding to mIg (Cμ₁-Cμ₂-Cμ₃-TM1-TM2) and sIg (Cμ₁-Cμ₂-Cμ₃-Cμ₄) were observed in the hybrid. The mIg and sIg nucleotide sequence coding from Cμ₁ through polyA tails were absolutely identical to their parent, bighead catfish and African catfish, respectively ( Figures 1D, G and Supplementary Figures 1 , 9 ).

In situ hybridization with specific probes to the Ighμ gene of individual catfish species, bighead catfish and North African catfish revealed that only a single copy of the Ighμ gene was found in bighead catfish ( Figure 1H ) and North African catfish ( Figure 1I ). In addition, a strong hybridization of both the Ighμ genes of bighead catfish and North African catfish was represented in their hybrid catfish ( Figure 1J ). The expression of Ighμ transcripts in the hybrid catfish from the two Ighμ genes received from the parents was confirmed.

The amino acid sequences of Ighμ transcripts among bighead catfish, North African catfish and their hybrid catfish were aligned and compared to published complete IgH gene loci of channel catfish (Ictalurus punctatus) and zebrafish (Danio rerio). The aligned sequences are illustrated in Figure 2 . The results showed several important features based on the overall relative size of the amino acid sequences of general Ighμ transcripts, with 436 and 430 residues for the constant (Cμ₁-Cμ₄) domain and 44 and 41 residues for the transmembrane domains of bighead catfish and North African catfish, respectively. The mIgM and sIgM forms of the hybrid catfish were similar in size to those of the parents, bighead catfish and North African catfish ( Figure 2 ). The sequence similarity of the mIgM and sIgM forms of the hybrid catfish were absolutely conserved and similar to those of the parents. These results exhibited highly important significant differences between their phenotypic characterization and relationship to their parent.

The translated amino acid sequences of the individual Cμ₁, Cμ₂, Cμ₃, Cμ₄ and transmembrane domains of bighead catfish, North African catfish and their hybrid catfish were used to define significant similarity among their relationships and evolutionary lineage. The amino acid sequences of the Cμ₁ and Cμ₂ domains of all catfish were similar in size, with 103 and 103 residues, respectively ( Figure 2A ). However, the Cμ₁ and Cμ₂ domains were slightly larger and smaller than the channel catfish Cμ₁ and Cμ₂ domains (102 and 104 residues, respectively) ( Figure 2A ). Additionally, bighead catfish shared amino acid similarity of Cμ₁ and Cμ₂ to African catfish, the hybrid catfish (mIgM), the hybrid catfish (sIgM), channel catfish and zebrafish with 82.4, 100.0, 82.4, 72.5 and 38.6% similarity for the Cμ₁ domain and 77.2, 100.0, 727.2, 57.8 and 36.8% similarity for the Cμ₂ domain, respectively ( Supplementary Table 1 ). The amino acid similarity of the Cμ₃ domains of bighead catfish shared 69.7, 100.0, 69.7, 45.9 and 32.9% and 74.4, 100.0, 74.4, 54.03 and 43.2% similarity in the Cμ₄ domain to North African catfish, the hybrid catfish (mIgM), the hybrid catfish (sIgM), channel catfish and zebrafish, respectively. In addition, the amino acid similarity values of transmembrane domains of the Ighμ carboxyl-terminal region of bighead catfish were 97.5, 100.0, 84.0 and 76.0% for North African catfish, the hybrid catfish (mIgM), channel catfish and zebrafish, respectively. Overall, based on amino acid conservation analyses of five complete domains of Ighμ of bighead catfish, the Cμ₁, Cμ₂, Cμ₃, Cμ₄ and TM domains shared 75.8, 100.0, 75.8, 57.5 and 38.4% similarity with North African catfish, the hybrid catfish (mIgM), the hybrid catfish (sIgM), channel catfish and zebrafish, respectively ( Supplementary Table 1 ).

The conservation of the Ighμ domain is evident in the catfish Cμ₁, Cμ₂, Cμ₃, and Cμ₄ of Ighμ structures. The cysteines that likely form the intradomain disulfide bridge (S-S) as well as the tryptophan located within the Cμ₁, Cμ₂, Cμ₃, and Cμ₄ regions were absolutely conserved among catfish, channel catfish and zebrafish ( Figures 2A–E ). The cysteine that likely participates in the disulfide linkage of the catfish chain to the light chain was also conserved in the Cμ₁ regions ( Figure 2A ). Additionally, the predicted results of these protein tertiary structures of Ighμ transcripts in bighead catfish, North African catfish and their hybrid catfish are illustrated in Figure 3 . The Ighμ molecules encoded by general coding exons found in teleost and catfish mIg (Cμ₁-Cμ₂-Cμ₃-TM1-TM2) and sIg (Cμ₁-Cμ₂-Cμ₃-Cμ₄) exhibited similar predicted structures in all the studied catfish ( Figures 3A–C ). Additionally, the predicted results of these protein tertiary structures of Ighμ transcripts of unusual splicing patterns generated novel membrane forms of Ighμ transcripts in teleosts [Cμ₁-Cμ₂-Cμ₃-Cμ₄-TM1-TM2] and vertebrates [Cμ₁-(Cδ₂-Cδ₃-Cδ₄-Cδ₅)-Cμ₂-Cμ₃-TM1-TM2] that were larger in size and rather complex compared to other general predicted structures ( Figures 3D, E ).

The evolutionary relationships of Ighμ constant regions among various vertebrate species are illustrated in Figure 4A , and the relationships of teleost Igh constant regions are illustrated in Figure 4B . In this study, the Clarias catfish Ighμ constant domains were cladded into the catfish group of vertebrate lineages and classified into the Ighμ group among the reported Igh in fish.

The two novel and unusual splicing patterns of the Ighμ membrane form were found in only bighead catfish: one was associated with a new splicing pattern among teleosts, and the other among vertebrates. The constant region of the new Ighμ membrane form in teleosts is encoded by four Cμ₄ exons and directly spliced into the first transmembrane exons (Cμ₁-Cμ₂-Cμ₃-Cμ₄-TM1-TM2), as reported in other vertebrates, such as tetrapods, birds, sharks, and mammals. However, the castratory event was observed by the lack of a cryptic donor splicing site at the 3′ region of the Cμ₄ exon, which is important for splicing to the transmembrane domain to generate the membrane form as reported previously. The predicted crystal protein structure of this molecule is illustrated in Figure 3D . The predicted results of the protein tertiary structures are close to the actual crystal structures of reported Ighμ transcripts in bighead catfish ( Figure 3A ), North African catfish ( Figure 3B ) and their hybrid catfish ( Figure 3C ).

The other novel and unusual splicing pattern of the Ighμ transcript was notable for vertebrates. The largest molecule of the Ighμ membrane-bound form was defined together with the new Ighμ splicing pattern among vertebrate animals. The molecular size of the molecule was approximately 92.64 kDa by encoding two major constant IgH loci in teleosts: Cμ and non-Cμ gene loci. The molecule encoded three Cμ_s and non-Cμ constant exons. The first exon of the 5′N-terminal domain was the Cμ₁ exon of the Ighμ gene, which was directly spliced to the non-Cμ exons (487 residues) ( Figure 2B ) and then spliced directly to link the Cμ₂, Cμ₃, TM1 and TM2 exons of the Ighμ gene of bighead catfish. The translated amino acid sequences of the Cμ exons, including Cμ₁, Cμ₂, Cμ₃, TM1 and TM2, were identical (100%) to the previously described Ighμ gene of bighead catfish ( Figures 1B , 2 ). In addition, the encoded non-Cμ constant region exons were studied via BLAST and compared to the published complete IgH gene loci of channel catfish and zebrafish. The translated amino acid sequences of the entire non-Cμ exons were 57.73 and 34.63% identical to the Ighδ gene locus of channel catfish and zebrafish, respectively ( Supplementary Table 1 ). The evolutionary relationship analysis of non-Cμ exons of bighead catfish was confirmed; it was grouped into Ighδ among other teleosts and shared a branch with the Ighδ of the closely related teleost species channel catfish ( Figure 4B ). In addition, the nearest neighbors’ relationships of presumptive Ighδ constant exons were gilthead bream (Sparus aurata, AXL14358), grass carp (Ctenopharyngodon idella, ACV21058) and Wuchang bream (Megalobrama amblycephala, AGR34025). However, the presumptive non-Cμ exon sequences clustered within teleost Ighδ rather than Ighμ. The predicted protein tertiary structure of this molecule was larger in size and rather complex compared to other predicted structures ( Figure 3E ).

To gain insight into the Ighμ locus of Clarias catfish and hybrid catfish, we also defined the complete genomic organization of the gene locus. The corresponding genomic sequences encoding the Ighμ constant domain were amplified, cloned, sequenced and compared to the published complete Igh gene locus in channel catfish. Consistent with this finding, only a single copy of the genomic Ighμ gene locus was found in bighead catfish and North African catfish ( Figures 1E, F, H, I , and Supplementary Figures 10 , 11 ). The complete constant region of the Ighμ locus, which encodes an Ig μ domain, contained approximately 6,793 kb (six exons, five introns and six polyadenylation signals) and 6,660 kb (six exons, five introns and six polyadenylation signals) of bighead catfish and North African catfish, respectively. However, the individual Ighμ loci of each bighead catfish and North African catfish were also found in the hybrid catfish, which had the transmembrane and secreted Ighμ transcripts, as mentioned before. The nucleotide identity between the two Ighμ gene loci in the hybrid catfish was identical to those of the parents, bighead catfish and African catfish ( Figures 1G, J and Supplementary Figures 10 , 11 ).

A total of 100 unique nonredundant clones of each catfish were characterized for VH sequences of the Ighμ gene. The functional regions of VH encoded by the V_H, D_H and J_H segments were identified, and analysis of framework regions (FWs) and complementarity-determining regions (CDRs) was also performed ( Figure 5A ) according to the international immunogenetics (IMGT) information numbering system for Ig structural definition (59).

Based on 80% nucleotide identity, the functional V_H segments were classified into 15, 15 and 17 families for bighead catfish, North African catfish and their hybrid catfish, respectively. Between the most 3′ V_H segment and the 5′ J_H segment, 9, 11 and 13 functional D_H segment families were defined in bighead catfish, North African catfish and their hybrid catfish, respectively. Altogether, the functional J_H segments of catfish VH have been identified. There were 11, 11 and 8 families of bighead catfish, North African catfish and their hybrid catfish, respectively ( Figure 5B ). The comparisons of the V_H gene family among catfish in this study and among reported vertebrates are illustrated in Figures 5B, C , respectively. The NCBI accession numbers of VH nucleotide sequences are MN934442-MN934541, MN934542-MN934641 and MN934642-MN934741 for bighead catfish, North African catfish and their hybrid catfish, respectively. In addition, adding P- and N-nucleotides was predominantly observed at the junction of the V and J segments for the generation of V(D)J junctional diversity in bighead catfish, North African catfish and their hybrid catfish ( Supplementary Figures 12 – 14 ).

To determine the extent of diversification in the VH repertoire of the Ighμ gene, we calculated the length variation and percent deviations of amino acids in each position of the FWs and CDRs and compared them to those of germline VH genes from the international ImMunoGeneTics information system^® (IMGT^®, http://www.imgt.org) (60).

The FWs of bighead catfish, North African catfish and their hybrid catfish were similar in length and diversity of amino acids. The lengths of amino acids of FW1, FW2, FW3 and FW4 in all catfish were approximately 33, 14, 36 and 14 residues, respectively ( Figure 5D ). To examine the degree of sequence variability of the VH repertoire in Clarias catfish, plots of the variability according to the Shannon (1948) (40) and Kabat and Wu (1971) (41) methods were generated. The overall amino acid sequence variability of the VH region was mostly confined to the CDRs and particularly the CDR3s. In addition, the amino acid variability among CDRs (CDR1, CDR2 and CDR3) was highly similar to that of FR1 to FR4 of all studied catfish. Moreover, the comparisons of variability plots of VH sequences among the studied Clarias catfish were closely related in sequence patterns and variability level to each other based on both the Shannon and Kabat and Wu methods ( Figure 5D ). Shannon entropy plots of all catfish revealed increased diversity at amino acid positions 31 to 42 for CDR1, 58 to 65 for CDR2 and 102 to 120 for CDR3 ( Figure 5D ). They were generally related to the traditional definitions using the Kabat numbering convention ( Figure 5D ).

The amino acid length diagrams of CDRs of bighead catfish, North African catfish and their hybrid catfish in Figure 5E indicate that the amino acid length of CDR1 ranged from 3 to 12 residues with an average of 7.6, 7.5 and 6.9 residues, respectively. The most common length was seven for all catfish, with approximately 58 to 71% frequency.

Within nonredundant CDR1 (n = 100) sequences of each Clarias catfish, 6 major amino acids were observed in bighead catfish, which accounted for 79.7% of all residues: Ala (7.7%), Gly (16.5%), Ser (18.9%), Thr (9.1%), Val (4.1%) and Tyr (23.1%). For North African catfish, 8 major amino acids were observed, which accounted for 93.7% of all residues: Cys (11.7%), Gly (8.5%), Asn (8.0%), Pro (4.0%), Ser (30.2%), Thr (14.5%), Val (8.7%), and Tyr (7.9%). For their hybrid catfish, 7 major amino acids were observed, which accounted for 75.2% of all residues: Asp (6.8%), Gly (8.2%), Leu (10.2%), Ser (17.3%), Thr (8.1%), Val (9.7%), and Tyr (14.6%) ( Figure 5F ).

The amino acid sequences of the CDR2 were shorter overall than those of CDR1, which were 4 to 8 residues in bighead catfish, North African catfish and their hybrid catfish with average amino acid lengths of 7.0, 6.9 and 6.8 residues, respectively. Approximately 76 to 79% of nonredundant CDR2s were 7 amino acid residues in length. Shannon entropy plots of amino acid variation of nonredundant CDR1 and CDR2 were mostly similar in diversity to those of the related V_H families within the species and among Clarias species. Within 8 residues in length of CDR2 revealed highly conserved residues at position one of the CDR1s ( Figure 5F ).

Six major amino acids were observed in nonredundant CDR2s (n = 100) sequences of bighead catfish, which accounted for 78.4% of all residues: Ala (5.4%), Asp (5.8%), Gly (25.3%), Ile (12.4%), Ser (11.7%) and Thr (17.5%). For North African catfish, 5 major amino acids were observed, which accounted for 78.2% of all residues: Asp (5.6%), Gly (21.0%), Ile (22.9%), Ser (17.4%) and Thr (11.1%). For their hybrid catfish, 5 major amino acids were also observed, which accounted for 73.2% of all residues: Asp (5.3%), Gly (31.8%), Ile (12.1%), Ser (11.3%) and Thr (12.4%) ( Figure 5F ).

The length distributions of the Clarias catfish CDR3 repertoire varied from 6 to 19 residues with averages of 11.9, 12.2 and 12.4 residues for bighead catfish, North African catfish and their hybrid catfish, respectively. The most common length was 10 to 15 residues for all catfish, with approximately 13 to 26% frequency. Analysis of amino acid frequencies for nonredundant CDR3s across all individual catfish revealed that the highest mutation rates in the VH repertoire were observed within CDR3 for all studied catfish. Highly conserved amino acid residues at positions one and two were observed in the CDR3 of all catfish. Seven major amino acids were observed in bighead catfish, which accounted for 66.3% of all residues: Ala (13.4%), Lys (7.1%), Leu (17.7%), Pro (3.6%), Gln (3.1%), Arg (17.1%) and Trp (4.0%). For North African catfish, 7 major amino acids were identified, which accounted for 80.3% of all residues: Ala (22.6%), Asp (7.5%), Gly (5.8%), His (9.5%), Ser (16.0%) Thr (4.3%) and Yyr (14.4%). For their hybrid catfish, 8 major amino acids were observed, which accounted for 77.4% of all residues: Ala (7.9%), Asp (5.9%), Gly (5.5%), Lys (10.7%), Leu (16.8%), Arg (21.0%), Ser (5.0%) and Thr (4.3%). Notably, three amino acids, Gly, Ser and Thr, were predominantly distributed in all CDR1s, CDR2s and CDR3s of all catfish in this study ( Figure 5F ).

To examine the distribution of Ighμ in Clarias catfish, we examined the potential functional coding segment of Ighμ, the Cμ₂ region, in sixteen tissues of catfish. Ighμ was highly expressed in the head kidney, spleen and peripheral blood lymphocytes (PBLs) in all studied catfish ( Figure 6A ). The relative expression levels reached approximately 60-, 700- and 600-fold the accustomed expression in head kidney, an important fish immune response organ, of bighead catfish, North African catfish and their hybrid catfish, respectively. However, the lowest expression was also observed in various tissues, including the brain, dendrite, gallbladder, gills, heart, liver, muscle, ovary, skin, stomach, testes and trunk kidney, in all studied catfish ( Figures 6A, B ). Highly significant expression levels were observed in North African catfish and the hybrid catfish in several tissues, including the gills, head kidney, heart, intestine, liver, muscle, ovary, PBLs, skin, spleen, stomach, and testes, compared to those in bighead catfish ( Figure 6B ). No significant difference was observed in tissues of North African catfish and the hybrid catfish. However, the relative expression of Ighμ was not significantly different in dendrites, gall bladder or trunk kidney among all studied catfish ( Figure 6B ).

Intraperitoneal injection of virulent Aeromonas hydrophila clearly exhibited the pattern of immune responses of catfish, and its significance was defined among the three catfish. In the spleen, the expression of Ighμ was moderately increased approximately 200-400-fold after 12–168 hours (hr) of pathogen exposure for all catfish. For the head kidney, the expression of Ighμ was strongly increased approximately 200–300-, 850–950-, and 800–950-fold after 12–24 hr. of pathogen exposure for bighead catfish, North African catfish and their hybrid catfish, respectively. In the PBLs, Ighμ expression showed significant upregulation of approximately 200-, 450- and 500-fold after 12–24 hr. of exposure for bighead catfish, North African catfish and their hybrid catfish, respectively. In addition, an approximate upregulation of 200–300-fold was also observed in all catfish after 12–24 hr. of exposure ( Figure 6C ).

The relative expression of two novel transcripts, including Cm_mIgM2 and Cm_mIgM3 in bighead catfish tissues showed highly expression levels compared to brain, approximately 337- and 122-, 90- and 177-, 40- and 20-, 21- and 17-fold, in PBLs, head kidney, spleen, and liver of Cm_mIgM2 and Cm_mIgM3, respectively ( Figures 7A, B ).