Molecular Characterization of A New IgZ3 Subclass from Common Carp (Cyprinus Carpio) and Comparative Expression Analysis of IgH Transcripts During Ontogeny of Larvae

Background Immunoglobulins distributing among systemic immune tissues and mucosal immune tissues play an important role in teleost to protect them from infections in a pathogen-rich aquatic environment. Teleost IgZ/IgT subclasses with different tissue expression patterns may have varied immune functions. Results In the present study, a novel secreted IgZ heavy chain gene was cloned and characterized in common carp (Cyprinus carpio) which was different from the reported IgZ1 and IgZ2 in tissue-specic expression prole. The obtained IgZ-like subclass was designated as CcIgZ3 which complete open reading frame contained 1650 bp encoding a protein of 549 amino acid residues. The phylogenetic analysis revealed that CcIgZ3 was grouped with carp IgZ2 and was in the same branch as other teleosts IgZ/IgT. Basal expression detection of IgH in healthy adult common carp showed that CcIgZ3 transcripts were widely expressed in systemic immune tissues and Mucosal-associated lymphoid tissues. It was expressed at a higher level in the head kidney, gill, and gonad, followed by spleen, hindgut, oral epithelium, liver, brain, muscle, foregut, and blood, but at a very low level in the skin. The transcript level of CcIgZ3 mRNA in isolated leukocytes from peripheral blood cells was signicantly higher than that of isolated leukocytes from the spleen. Different groups of common carp were infected with Aeromonas hydrophila via intraperitoneal injection and immersion respectively. The qRT-PCR analysis demonstrated that a signicant difference in CcIgZ3 mRNA level between immersion and injection groups existed in all the detected tissues including head kidney, spleen, liver, and hindgut, especially in hindgut CcIgZ3 mRNA level was higher in immersion than in the injection group. Different routes for Aeromonas hydrophila challenge to common carp had comparatively less effect on IgM response. Further study on the relative expression of the IgH gene during the ontogeny of common carp presented that the tissue-specic expression prole of CcIgZ3 was very different from the others. In the early larval development stage of common carp from 1 dpf to 31 dpf,

Fish, like other vertebrates, possess an extensive defense system, which enables the individual to survive and maintain its integrity in a hostile environment. The humoral immune system responds to a variety of pathogens by producing speci c antibodies. Antibodies produced by B lymphocytes were existed in vertebrates from mammals to cartilaginous sh [1]. There are ve types of heavy chains in mammals, including µ, δ, γ, ε, and , which differ in their constant region. Antibodies that contain different heavy chains belong to different isotypes and are named IgM, IgD, IgG, IgE, and IgA respectively. Each isotype has distinct physical and biological properties and effector functions. Fish are the most primitive group of vertebrates that possess an adaptive immune system capable of generating antibodies in response to pathogenic challenges. Except for IgM and IgD, a novel isotype IgZ(ζ) or IgT (τ) was identi ed in many teleost sh species [2], which were also found to have varied subclasses with different gene sequence and different tissue expression patterns, Such as IgZ1 and IgZ2 in zebra sh (Danio rerio) [3], IgT1, IgT2, IgT3, IgT4, and IgT5 in rainbow trout (Oncorhynchus mykiss) [4], IgZ and a chimeric IgZ (IgZ2) in grass carp (Ctenopharyngoden idellus) [5], IgT1, IgT2, IgT3 and IgT4 in stickleback (Gasterosteus aculeatus) [6], ve IgT-A and three IgT-B in Atlantic salmon (Salmo salar) [7]and IgZ1 and IgM-IgZ (IgZ2) in common carp (Cyprinus carpio) [8]. Most of the IgZ/T molecules have four CH domains which were encoded by the Cζ/Cτ genes, but some others have two CH domains or three CH domains, such as two constant domains in common carp (Cyprinus carpio) IgZ2 [9]and torafugu (Takifugu rubripes) IgH [10], three constant domains in stickleback (Gasterosteus aculeatus) IgT [6] and sea bass (Dicentrarchus labrax) [11]. Except for the diversity of the domain number of IgZ/T subclasses among sh species, those IgZ/T subclasses displayed varied expression patterns and functions. The reported IgZ1 of common carp is mainly expressed in blood against blood pathogens, and IgZ2 chimera is preferentially expressed in the mucosal compartment to respond to mucosal infections [8]. The rainbow trout Igτ1 was mainly expressed in both systemic and mucosal lymphoid tissues, while Igτ2 was largely expressed in systemic lymphoid organs.
After poly (I:C) treatment, Igτ1 and Igτ2 genes exhibited different expression pro les, Igτ1 transcripts reached a peak at 7 d in the spleen and 14 d in the gut, However, Igτ2 increased slightly and reached a peak at 7 d in both the spleen and gut [12]. The previous study about this novel IgH indicated that teleost sh IgZ/T was more diverse than previously thought.
Teleost IgZ/T was considered as a primitive immunoglobulin class specialized in mucosal immunity, which is equivalent to immunoglobulin IgA in mammals and plays an important role in the mucosal immune response [2,13]. IgZ/T is a comparatively new teleost immunoglobulin class and its presence, expression, and tissue distribution during the early development stage are not very clear. Most studies have been investigated on the ontogeny of IgM positive cells and IgM-secreting cells. The previous studies indicated that the rst B cells are most probably generated in the head kidney and B cells populate the gut-associated lymphoid tissue (GALT) much later than the spleen or kidney [14]. In channel cat sh, speci c IgM in eggs could be detected, which provides an immune barrier at the surface of the egg, as well as protection for the developing fry. In carp, surface IgM + cells were rst detected in the head kidney at 2 weeks post-fertilization using WCI12 and WCI4, monoclonal antibodies against IgM H chain, and carp injected with T cell-independent antigen (LPS) developed antibody responses and memory from 4 wpf, while they responded against T cell-dependent antigens (HGG) from 8 wpf [15]. By using speci c primers designed according to the IgH constant sequence, qRT-PCR was performed to detect the IgH presence during the early development stage of some teleost sh species. In zebra sh, all the Ig isotypes effectively responded to the LPS challenge from 21 dpf onwards, IgZ1 responded to LPS challenge faster and stronger than IgM and IgD on 28 dpf. IgZ-2 transcripts could be detected at 14 days after fertilization [16]. In common carp whole embryos showed constitutive expression of all three immunoglobulins (IgM, IgZ1, and IgZ2) as early as 4 dpf with IgM being the predominant form. IgZ1 and IgZ2 expression increased rapidly to reach their maximum at 12 dpf whereas IgM expression reached its maximum at 30 dpf [8]. The appearance of immunoglobulin during the early development stage varies considerably in different teleost species due to important differences in developmental status at hatch and the aquatic environment. The presence of Ig molecules in sh embryos and larva suggests their role as an important barrier against pathogens.
Carp is one of the most popular cultured sh in China and the diseases caused by Aeromonas hydrophila can do great harm to carp culture. Aeromonas hydrophila is resistant to antibiotics attributed to the indiscriminate use of antibiotics in aquaculture. Vaccines based on the immune response mediated by antibodies will help to defend against bacterial infection without the use of antibiotics [17]. In the present study, we cloned a new CcIgZ3 gene from common carp and detected its expression pattern in the adults and during embryonic development of common carp with comparison to the other known IgH of this species. The immune response of CcIgZ3 and IgM after the challenge by Aeromonas hydrophila by immersion and injection were also analyzed comparatively. Our results provide additional experimental evidence to understand the IgZ/T ontogeny and immune function in teleost sh.

Molecular cloning and analysis of common carp CcIgZ3
Identi cation of the constant region of CcIgZ3 cDNA in Cyprinus carpio.
The obtained sequence of full-length CcIgZ3 cDNA was 2144 bp with a 5'-UTR of 112 bp and a 3'-UTR of 382 bp with a putative typical polyadenylation signal sequence AATAAA located upstream of the poly (A) tail. The deduced CcIgZ3 amino acid sequence contained 549 aa and spanned V domain, four constant domains and a secreted tail (Fig. 1.). The analysis of the IgT sequence showed the presence of a putative 20 amino acids signal peptide. Concerning the V domain of the CcIgZ3 sequences the IMGT unique numbering has been used. The entire C-domain could be divided into four CH domains.

Multiple amino acid sequence alignment
The alignment of common carp CcIgZ3 with zebra sh IgZ1 and IgZ2, common carp IgZ1 and IgZ2, grass carp IgZ and IgZ2 showed that it was comprised of four Ig-like constant domains (CH1, CH2, CH3, CH4) and a secretory tail. The conserved cysteine residues for disul de bond formation and tryptophan residues for the folding of IgSF domain were present in each CH domain of CcIgZ3 (Fig. 2.). Five N-linked glycosylation sites were predicted to be present in CcIgZ3 (in CH2, CH3 and CH4) (Fig. 1.). The number and distribution of putative N-glycosylation sites including the sequons NXS or NXT for each domain are varied in different teleost species. CH1 domain could possess none to three putative N-glycosylation sites. There was no N-linked glycosylation site in CcIgZ3 CH1 domain, which is different from the grass carp (three N-linked glycosylation sites in CH1) and zebra sh (two N-linked glycosylation sites in CH1). The rst N-linked glycosylation site of CcIgZ3 existed in the CH2 domain and it is at the very beginning of CH2. CcIgZ3 presented two N-linked glycosylation sites in CH3 and CH4 domains respectively, which is similar to the grass carp [5].
The percentage of identity between the amino acid sequences of all the IgZ subclasses of common carp and that of grass carp, zebra sh and rainbow trout were calculated considering every single domain. The results showed that the value of identity obtained ranged between 37.5% and 74.76% for all the three IgZ subclasses of common carp, 11.59% and 70.21% for the three IgZ subclasses of common carp with grass carp IgZ/IgZ2, 24.18% and 64.13% for the three IgZ subclasses of common carp with zebra sh IgZ1/IgZ2, 21% and 40.82% for the three IgZ subclasses of common carp with rainbow trout IgT1 ~ IgT3. Interestingly, the identity between CcIgZ3 CH1 domain and grass carp IgZ CH2 domain (71.26%) was higher than that between CcIgZ3 CH1 domain and grass carp IgZ CH1 domain (11.69%), the other two IgZ of common carp also showed this characteristic (41.67% vs 13.1 in IgZ1, 66.67% vs 17.07% in IgZ2).

Phylogenetic analysis
The phylogenetic analysis of immunoglobulins in vertebrates was performed to reveal the relationship of common carp CcIgZ3 with other immunoglobulins. The amino acid sequences from different species including mammals, birds, amphibians and sh were aligned by Clustal X2 program. Based on the alignment, we constructed an NJ tree supported by 1000 bootstrap replications. It showed that common carp CcIgZ3 are grouped with grass carp IgZ and in the same branch as IgZ/T, further demonstrated that CcIgZ3 is a homology of IgZ (Fig. 3).

Basal expression comparison of CcIgZ3 with other Ig molecules of common carp
To investigate the tissue-dependent expression pattern, we performed qRT-PCR analysis using genespeci c primers for all the three IgZ and IgM of common carp in normal adult common carp tissues including liver, spleen, head kidney, gill, skin, gonad, brain, muscle, blood, foregut, hindgut and oral epithelium. The expression of the CcIgZ3, IgZ1, IgZ2 and IgM were found in all the tested tissues and IgM was the most abundant Ig in all the tissues. The expression of CcIgZ3 was found to be highest in the head kidney, gill, and gonad, followed by spleen, hindgut, oral epithelium, liver, brain, muscle, foregut, and blood, but at a very low level in skin. IgZ1 and IgZ2 demonstrated similar expression patterns with CcIgZ3 in most of the detected tissues except that the lowest transcription level of IgZ1 and IgZ2 were found in blood (Fig. 4). To compare the expression of the four IgH transcripts in leukocytes, we performed qRT-PCR analysis using gene-speci c primers for all the three IgZ subclasses and IgM of common carp in isolated leukocytes from peripheral blood cells and spleen. The result showed that signi cant difference in CcIgZ3 expression existed between PBL and leukocytes from the spleen. In contrast, no signi cant difference in IgM, IgZ1 and IgZ2 expression existed between PBL and leukocytes of the spleen (Fig. 5).
Basal expression comparison of CcIgZ3 with other Ig molecules of common carp during the different developmental stage When testing for constitutive expression of the four immunoglobulin genes during the early development stage of common carp by qRT-PCR, expression was recorded from 1 to 65 days post fertilization (dpf). ( Fig. 6 and Fig. 7). Expression of CcIgZ3 and the other three immunoglobulin genes were detected from 6 days post fertilization (dpf). IgM was expressed highest and was the dominant immunoglobulin isotype during the early development stages tested. (Fig. 6). Furthermore, to investigate the tissue expression pattern of the IgH transcripts, tissues including gill, spleen, hindgut and liver from common carp at 65 days post fertilization (dpf) were used to evaluate the expression level of CcIgZ3, IgM, IgZ1, and IgZ2 respectively. The expression of IgM was still higher than that of the other three subclasses and it was the only isotype that could be detected in all the tested tissues. The gene expression of CcIgZ3, IgZ1 and IgZ2 exhibited varied tissue expression preference. CcIgZ3 was found in spleen, hindgut and gill but not in liver, however, IgZ1 was expressed at a higher level in liver but its expression level in gill was too low to be detected. IgZ2 was primarily detected in gill and spleen but hardly detected in liver and hindgut (Fig. 7).
Tissue location of CcIgZ3 and IgM mRNA-expressing cells at early development stage of common carp As both of the CcIgZ3 and IgM expression were detected in spleen, gill and hindgut from common carp at 65 dpf, we further detected the CcIgZ3 mRNA-expressing cells and IgM mRNA-expressing cells in tissues by in situ hybridization. Positive expression of both IgM and CcIgZ3 were seen in splenocytes ( Fig. 8F and 8H), gill epithelial cells ( Fig. 8J and 8L) and cells in lamina propria of hindgut ( Fig. 8B and 8D). The staining was speci c for both IgM and CcIgZ3 as only the antisense probes (Fig. 8B, 8D, 8F, 8H, 8J, 8L) and not the sense probes (Fig. 8A, 8C, 8E, 8G, 8I, 8K.) hybridized.

Organ-speci c CcIgZ3 and IgM expression after Aeromonas hydrophila challenge by different routes
The qRT-PCR results showed that both CcIgZ3 and IgM mRNA expressions were up-regulated compared to the blank control in all tested tissues in the injection and immersion group from 3 days to 21 days except for CcIgZ3 in liver in injection group (0.46-fold at 3 days) (Fig. 9F). Compared with the control, the peak value of CcIgZ3 expression was 86.78-fold in spleen (Fig. 9D), 31-fold in hindgut (Fig. 9H), 13.61fold in head kidney (Fig. 9B), and 6.39-fold in liver (Fig. 9F) in the immersion group, and 65.13-fold in spleen (Fig. 9D), 10.96-fold in hindgut (Fig. 9H), 26.91-fold in head kidney (Fig. 9B), and 1.21-fold in liver (Fig. 9F) in the injection group respectively. The peak expression of CcIgZ3 in immersion group, in contrast to injection group was comparatively higher in liver (5.28-fold), hindgut (2.83-fold), and spleen (1.33-fold), but lower in head kidney (0.51-fold). However, the peak value of IgM expression was 13.79fold in head kidney (Fig. 9A), 0.57-fold in spleen (Fig. 9C), 6.5 in liver (Fig. 9E) and 4.07 in hindgut (Fig. 9G) in immersion group and 3.09-fold in head kidney (Fig. 9A), 5.27-fold in spleen (Fig. 9C), 5.85-fold in liver (Fig. 9E), and 5.8-fold in hindgut (Fig. 9G) in injection group. The peak expression of IgM in immersion group, in contrast to injection group were comparatively higher in head kidney (4.46-fod), spleen (1.25-fold), and liver (1.11-fold), but lower in hindgut (0.7-fold). The results of IgM displayed different tissue expression pattern and lower fold change from that of CcIgZ3.
The peak time of CcIgZ3 and IgM were most at 3-7 days in spleen, liver, and hindgut in both immersion group and injection group. However, the peak time of IgM in head kidney was much later in immersion group (14 days) than in injection group (1 day). The peak time of CcIgZ3 in head kidney was also later in immersion group (14 days) than in injection group (3 days). In hindgut, the peak time of CcIgZ3 appeared later in immersion group (3 days) than in injection group (1 day). The immersion challenge in these tissues showed a slower immune response mediated by CcIgZ3 and IgM than injection challenge (Fig. 9).
As for the response in hindgut, the up-regulation of CcIgZ3 expression was signi cantly higher in the immersion group than the injection group from 3 days to 21 days post challenge, while the expression of IgM had no signi cant difference from 3 days to 21 days in the hindgut. The value of CcIgZ3 expression in hindgut was 1.73-fold at 1 day, 31-fold at 3 days, 11.53-fold at 7 days, 5.97-fold at 14 days, and 3.76fold at 21 days in immersion group, and 10.96-fold at 1 day, 1.69-fold at 3 days, 1.13-fold at 7 days, 1.06fold at 14 days, and 1.09-fold at 21 days in injection group. The ratio of expression of CcIgZ3 in immersion group in contrast to injection group at different time point were 0.15-fold at 1 day, 18.34-fold at 3 days, 10.2-fold at 7 days, 15.63-fold at 14 days, and 3.45-fold at 21 days (Fig. 9H). The value of IgM expression in hindgut was 0.75-fold at 1 day, 2.93-fold at 3 days, 4.07-fold at 7 days, 2.07-fold at 14 days, and 1.73-fold at 21 days in immersion group, and 2.94-fold at 1 day, 2.13-fold at 3 days, 5.80-fold at 7 days, 3-fold at 14 days, and 1.62-fold at 21 days in injection group. The ratio of expression of IgM in immersion group in contrast to injection group at different time points was 0.26-fold at 1 day, 1.38-fold at 3 days, 0.7-fold at 7 days, 0.69-fold at 14 days, and 1.07-fold at 21 days (Fig. 9G). The results showed that Aeromonas hydrophila challenge triggered a stronger CcIgZ3 immune response than IgM in hindgut, especially by the immersion route.

Discussion
In teleost, apart from IgM and IgD, the new IgZ/T isotype was discovered and more than one subclass of IgZ/T was found in some sh species. Although the alignment of these subclass sequence with each other indicated that every subclass of IgZ/IgT exhibited some common structural features, speci c gene organization and conserved amino acid residues, their expression and distribution pattern in immune organs and tissues varied obviously, which implying the function difference among these IgZ/T molecules [18,19].
So far in common carp, two IgZ subclasses have been reported, one is IgZ1 and the other is chimeric IgZ2, both of them have the membrane form as well as secretory form. Alignment of CcIgZ3 with other IgZ showed that conserved cysteine residues that participate in the formation of intrachain disul de bonds were found in each CH domain of all the three IgZ subclasses from common carp and the conserved cysteine residue for interchain disul de bond existed in all the three IgZ CH1 domain. However, the conserved the tryptophan residues for the folding of IgSF domain were just found in each CH domain of CcIgZ3 and IgZ2 and the position of tryptophan residue in CH4 domain of CcIgZ3 and IgZ2 was different from rainbow trout and zebra sh. The sequence of the secretory tail is conserved among the three subclasses. Alignment of the CH sequences indicated that the CH4 domain is the most highly conserved domain and CH1 domain exhibited a comparatively lower level of sequence identity.
Phylogenetic analysis showed that CcIgZ3 was grouped with IgZ2 together and the sequence identity between the two molecule CH1 and CH4 domain were 69.52% and 74.49%. In addition, the CcIgZ3 CH1 domain was highly conserved with grass carp IgZ CH2 domain, the sequence identity of their CH1 domain was very low (11.69%). These differences made it possible to discriminate the IgZ subclasses in detecting their expression and position.
In the current study, the results of constitutive expression of different subclasses in normal tissues showed that IgZ2 chimera expressed at a slightly higher level in mucosal tissues including gill, skin, hindgut and oral epithelium compared to IgZ1 and CcIgZ3, however, IgZ2 expressed at a lower level in the blood than the other three subclasses, which may indicate that IgZ2 chimera was preferentially expressed at mucosal sites. This result corroborates the previous ndings which found that the IgZ2 was expressed at relatively higher levels in gut and gills [8]. Compared to all the three subclasses of IgZ, IgM was predominantly expressed in all organs including both lymphoid and non-lymphoid organs and tissues. IgZ1 is expressed systemically since its comparatively higher expression level to IgZ2 and CcIgZ3 was observed in blood. As the third member of new immunoglobulin subclasses, CcIgZ3 expressed higher in gonad and liver compared to IgZ1 and IgZ2. To investigate whether CcIgZ3 has the immune function, the challenged experiment was performed by A. hydrophila stimulation. We found that the CcIgZ3 expression could be induced in the hindgut strongly in the challenge by A. hydrophila immersion, although its basal expression was not very high in hindgut. Teleost gut interacted with the foreign pathogens from water directly and was a main site for pathogens entry in teleost sh. Teleost Igs were proved to elicit the mucosal immune response in GALT by speci c B lymphocytes proliferation and secreting antibodies locally within the mucosa-associated lymphoid tissues (MALTs) [20,21]. The posterior segment of the teleost intestine was discovered to contain several immune cell types and showed to play a more important role in immune responses against pathogen invasion than the rst segment and the second segment of teleost intestine [12,21]. The CcIgZ3 immune response in hindgut of common carp indicated that it may play an important role in protecting the host from A. hydrophila infection, which immune response was even stronger than that of IgM.
The ontogeny studies of sh immunoglobulin have been conducted in several species including common carp IgZ1 and IgZ2. In the previous studies, IgM and IgZ were expressed during the same stage of development in fugu [10]and trout [22], but in zebra sh [4], IgZ was the rst detected antibody. In common carp, the IgZ1 and IgZ2 could already be detected as early as 4dpf and tissue-speci c expression rstly existed in the head and trunk kidney with the highest expression [8,9]. To compare the different subclasses, we rstly performed the detection of IgM, CcIgZ3, IgZ1, and IgZ2 expression during the early development stage from 1 to 31 dpf by qRT-PCR. In the current study, IgM and CcIgZ3 were expressed from the same stage of 6 dpf during all the detected time points and their expression increased gradually.
However, IgZ1 and IgZ2 were expressed from 16 dpf and their expression level even increased exceeding CcIgZ3 at 31 dpf with a greater slope of the curve. Secondly, the results of tissue speci c expression at 65 dpf showed that IgM was the only antibody that could be detected in the four tissues and it was abundant comparatively in spleen, hindgut and gill. The CcIgZ3 dominantly existed in the spleen, followed by gill and hindgut, the IgZ1 gene was expressed at relatively higher levels in liver, then in hindgut and spleen, and the IgZ2 gene was expressed in gill and spleen with absence in hindgut and liver. These observations may indicate the functional difference among these subclasses. Furthermore, the production of these three subclasses of IgZ might attribute to the differentiation and maturation of B lymphocytes residing within the systemic lymphoid tissues and mucosal associated lymphoid tissues. The subpopulations of IgZ positive B lymphocytes need further investigation.
Although much information on IgZ/T characterization has been achieved, few data are available on the in situ detection of IgZ-expressing cells in sh especially during the ontogeny [10,23,24]. Based on the analysis of Ig expression during the ontogeny by qRT-PCR, in situ hybridization (ISH) has been employed to show the CcIgZ3-expressing cells and IgM-expressing cells in the tissues of carp at the development stage of 65 dpf. The ISH showed clearly the location of CcIgZ3-expressing cells in the spleen, gill and hindgut which was similar to that of IgM-expressing cells, however, the amounts of transcripts of the two antibodies were different as their positive reactivity varied. These obtained results implied that we might reveal the distribution and location of CcIgZ3-expressing cells by ISH even in the early development stage of common carp, which might help to clarify the production of CcIgZ3and the maturation of CcIgZ3 positive B lymphocytes.

Conclusion
In conclusion, the third member of common carp CcIgZ3 was cloned and characterized and the study about the expression and location of CcIgZ3 compared with IgM, IgZ1 and IgZ2 in larvae ontogeny was performed. Comparative analysis between CcIgZ3 and IgM in system immune tissues and gut associated lymphoid tissues of the common carp challenged by A. hydrophila injection and immersion was completed. We hope it might help to expand the information of adaptive immunity in common carp and teleost sh. Further research is required to understand the role of different IgZ/T subclasses and their corresponding positive B lymphocytes.

Ethics statement
For all experiments on live animals, we con rm that all methods were carried out in accordance with relevant guidelines and regulations. The protocol was approved by the Animal Experimental Ethics Committee of Shandong Normal University (Permit Number: AEECSDNU2017004). All efforts were made to minimize suffering. The study was carried out in compliance with the ARRIVE guidelines.

Experimental animals
Common carp (Cyprinus carpio) weighing between 150 g to 200 g were obtained from the Fresh Water Fishery Research Institute of Shandong Province (China) and maintained in tanks with aerated freshwater at 22-25℃. the sh were acclimated to the aquarium tanks for at least 2 weeks before being used in experiments. Fish were anesthetized by immersion in a solution of Tricaine (MS222, Sigma Aldrich) and killed. Tissue samples of blood, head kidney, spleen, liver, gill, foregut, midgut, hindgut, oral epithelium, skin, gonad, and muscle were collected immediately, frozen in liquid nitrogen, and stored at -80℃ until use [25,26]. All animal experiments were approved by the Committee on the Ethics of Animal Experiments of Shandong University.

Molecular cloning and analysis of common carp CcIgZ3
Total RNA extraction and rst-strand cDNA synthesis Each frozen sample was ground in a mortar with nitrogen and then total RNA was isolated using the TRNzol Universal reagent (Tiangen, China). The quantity and quality of the total RNA were assessed using the NanoDrop Spectrophotometer (Thermo Scienti c, USA) on all samples. The rst-strand cDNAs were synthesized from 2 µg total RNA according to the manufacture's instructions of FastQuant RT Kit (With gDNase) (Tiangen, China) [27,28]. Total RNA from the collected samples was extracted following the procedure above and then the cDNA was stored at -80 ℃ until used in real-time quantitative PCR (RT-qPCR).

Molecular cloning and sequencing of common carp CcIgZ3
The common carp CcIgZ3 cDNA fragment was rstly ampli ed by PCR with primer IgZ3 F1 and IgZ3 R1 designed based on the known sequences of teleost sh immunoglobulin Z (GenBank Accession No.: Danio rerio AY643750, EU732710.1, AY643750, Ctenopharyngodon idella DQ478943, GQ201421, Cyprinus carpio AB004105, AB598367, AB598368, AB598369). Head kidney cDNA of common carp was used as the template. Reactions steps were: 3 min initial denaturation at 94 ℃, 1 min denaturation at 94 ℃, 30 s annealing at 55 ℃, 1 min extension at 68 ℃ (35 cycles), and 5 min nal extension at 68 ℃. Ex Taq HS (TaKaRa) was used in PCR and PCR products were loaded on 1% agarose gel and visualized by staining the gel in 0.1 mg/ml ethidium bromide. The DNA ampli ed in each reaction by PCR was puri ed using a Gel Extraction Kit (Tiangen), inserted into the pMD19-T vector (Takara) and transformed into competent TOP 10 E. coli cells for sequencing.
Subsequently, rapid ampli cations of cDNA ends (RACE) was performed using the 3'-Full RACE Core Set (Takara) and the SMARTer RACE cDNA ampli cation Kit (TaKaRa) to get the full-length CcIgZ3 cDNA sequences by the speci c primers which were designed based on the obtained partial sequence. The 3'-Full RACE Core Set (TaKaRa) was utilized for obtaining 3'-unknown regions with the speci c forward primers and the adaptor primers listed in Table 1. The rst round of PCR was performed using the primer pair of IgZ3 race-3'outer primer/3'RACE outer primer, under the following conditions: one cycle of 94℃ for 3 min, 30 cycles of 94℃ for 30 s, 57℃ for 30 s, and 72℃ for 2 min, followed by a nal extension of 72℃ for 10 min. The resultant product was diluted and re-ampli ed in the second round PCR using the primer pair of IgZ3 race-3'inner primer/3'RACE inner primer under the same reaction conditions. All PCR products were puri ed using a Gel Extraction Kit (Tiangen) and then were cloned into the pMD19-T vector (TaKaRa) for sequencing. For the 5' RACE reaction, the rst strand cDNA was synthesized from 1 µg of total spleen lymphocyte RNA after adding the 5' RACE adapter to RNA, following the manufacturer's instruction. From the 5' RACE Ready cDNA samples, PCR reactions were performed using the Advantage 2 PCR Kit (Clontech), using the manufacturer's speci cations and the speci c primers shown in Supplemental Table 1. Reaction steps were: one cycle of 95℃ for 1 min, 30 cycles of 95℃ for 30 s and 68℃ for 3 min, followed by a nal extension of 68℃ for 3 min. All PCR products were puri ed using a NucleoSpin Gel and PCR Clean-Up Kit (Takara), and then were cloned into the pRACE vector (TaKaRa) for sequencing.

Bioinformatic analysis of common carp CcIgZ3
The full length of CcIgZ3 was con rmed by RT-PCR using sequence-speci c primers IgZ3-1817F2/R2, IgZ3-1817F3/R3. The open reading frames (ORFs) and deduced protein sequences of CcIgZ3 was predicted using an ORF Finder program, and also by blasting genomic stretches against protein databases at NCBI (blastx). [29]. The location of Ig domains was predicted using the InterProScan program, Prosite Database and NCBI Conserved Domains. Post-translational modi cations were predicted through the NetNGlyc 1.0 program. The theoretical isoelectric point and the molecular weight of the amino acid sequence were calculated using the ExPASy Compute pI/Mw program. Multiple sequence alignment was conducted using CLUSTAL X version 2.1 with default parameters and was adjusted manually [11]. Based on the alignment, the phylogenetic tree was generated based on the deduced amino acid sequence using the Neighbour-Joining method with MEGA-X. All the sequences used for the phylogenetic analysis were listed in Table 2.
Gene expression studies of common carp immunoglobulins Basal expression of immunoglobulin isotypes in organs of common carp For the tissue expression, total RNA of the head kidney, spleen, liver, blood, skin, gill, foregut, midgut, hindgut, oral epithelium, gonad and brain from normal common carp were isolated and then reverse transcripted into rst strand cDNA as described above. PCR was conducted with speci c primers as indicated in Table 1 with SuperReal PreMix Plus (SYBR Green, Tiangen, China). The qPCR ampli cation program was 94℃ for 30 s, 55℃ for 30 s, and 72℃ for 30 s for 30 cycles to amplify S11 (as a standard) using primers S11F/S11R, and 94℃ for 30 s, 55℃ for 30 s, and 72℃ for 30 s for 30 cycles to amplify S11 (as a standard) using primers for IgM, IgZ1, IgZ2, and CcIgZ3 (listed in Table 1) in parallel tubes, respectively. The relative expression of the immunoglobulin gene was calculated and normalized against S11.
Basal expression of immunoglobulin isotypes in isolated leukocytes from different tissues of adult common carp Whole Blood was collected from the caudal vein for isolation of PBL (peripheral blood lymphocytes) with a heparinized syringe and centrifuged at 4 ℃, 500 x g for 10 min and then removed the serum. Diluted the cells 6-fold original volume of blood with RPMI-1640 medium at room temperature, then put on ice. Spleen was dissected from the anesthetized sh and placed in a sterile plastic culture dish containing 5 mL RPMI-1640 with 100U/mL penicillin G and 100 mg/mL streptomycin (Sigma, USA) respectively. Single-cell suspensions from the spleen were obtained by teased apart with sterile dissecting scissors at rst, repeatedly aspirated and then passed through a 100 µm nylon mesh with RPMI-1640 medium. The total volume 10 mL of single-cell suspension was layered upon the same volume of Histopaque 1077 (Sigma, USA) in 50 mL centrifuge tubes gradually, centrifuged 500 x g for 40 min at 4 ℃. Leukocytes were collected from the interface layer and washed three times with medium [30,31]. Cell quantity and viability were determined by 0.4% trypan blue (Sigma, USA), and cells were collected for qPCR analysis [32]. The expression of immunoglobulin genes was calculated by the expression and normalized by S11.

Relative expression of immunoglobulin isotypes during the development of common carp
For the expression analysis of immunoglobulins during different development stages, fertilized eggs (n = 100) were obtained and the total RNA was extracted using the FastQuant RT Kit (With gDNase) (Tiangen) from embryos or larvae 1 day, 6 days, 10 days, 16 days and 31 days post fertilization, and then followed by the steps as described above. The Ig expression in tissues from common carp of 65 days post fertilization including gill, spleen, hindgut and liver were detected by qPCR following the same procedure.
In situ hybridization Synthesis of RNA probes Common carp IgM and CcIgZ3 cDNAs were ampli ed with primers (Table 1) and subcloned into the pSPT18 vector (Roche). The primers IgZ3_situ_F4 and IgZ3_situ_R4 were designed to amplify a 244 bp product corresponding to the constant region of the CcIgZ3 CH4 domain. The primers IgM_situ_F2 and IgM_situ_R2 were designed to amplify a 435 bp product corresponding to the constant region of common carp IgM CH4 domain. The cycling protocol was: denaturation at 94℃ for 3 min, followed by 35 cycles of 94 ℃ for 30 s, 58℃ for 30 s and 72 ℃ for 30 s, with the nal extension step of 72 ℃ for 10 min. The PCR products were visualized on 1% agarose gels containing ethidium bromide. The fragments were puri ed using TIANgel Midi Puri cation Kit (TIANGEN) and inserted into the pSPT18 vector (Roche) and transfected into competent DH5α E. coli cells. Plasmid DNA from four clones was puri ed and sequenced respectively. The sequence analysis was performed to con rm the sequence identity and inserted orientation.
For generating RNA probes, the clones were digested with EcoR I or Hind III, and the fragments were puri ed on an agarose gel and used for in vitro transcription reactions with the DIG RNA Labeling Kit (Roche). Transcription was performed with SP6 RNA polymerase and T7 RNA polymerase according to the protocol so that antisense and sense RNA probes could be generated.
in situ hybridization Spleen, gill and hindgut were aseptically extracted from the sh and xed in 4% paraformaldehyde in PBS-H 2 O DEPC for at least 4 h. the tissues were then immersed in a 15% sucrose solution for 8 h and transferred in a 30% sucrose solution overnight. Embedded tissue in OCT was sectioned in 4 µm thickness and mounted onto poly-L-lysine coated slides. The slides were removed from the freezer, xed in paraformaldehyde (4% in PBS, pH 7.4) for 20 min, washed three times with DEPC-treated PBS buffer (pH 7.4) and permeabilized with protein K (5 µg/ml) buffer at 37℃ with gentle rocking. After washing in PBS-glycine buffer, the sections were washed twice with PBS-H 2 O DEPC . Prehybridization was done by incubating the sections with prehybridization buffer (Servicebio) for 60 minutes at 37℃. DIG-labeled antisense RNA probes (1 µg/ml) were applied with hybridization solution on the tissues and incubated at 55℃ overnight in a moistened chamber. To remove the hybridization solution, the sections were washed sequentially by 5x SSC, 1xSSC, 0.5 x SSC and 20% formamide (50 min). the tissue sections were blocked by 5% serum blocking reagent for 30 min at room temperature. Anti-DIG-AP antibody used for the detection was diluted 1:500 in blocking buffer solution containing 5% serum. The sections were washed twice with PBS, subsequently, BCIP/NBT reagent (Roche) were applied according to the protocol. The reaction was visualized and documented using a bright-eld microscope.
Organ-speci c IgM and CcIgZ3 expression after A. hydrophila challenge by different routes Fifty common carp were divided into two groups for immune stimulation and challenged by Aeromonas hydrophila intraperitoneal (i.p.) injection and immersion respectively as previously described [27]. Brie y, The Aeromonas hydrophila used in the study was obtained from the China Center for Type Culture Collection and incubated in LB medium at 28℃ overnight under continuous shaking. In the injection challenge, the A. hydrophila was inactivated in 0.5% formalin at 4 ℃ overnight and then was suspended in sterile 0.1M phosphate-buffered saline (PBS). Each sh was challenged by i.p. injection of 5 × 10 7 CFU per sh. In the immersion challenge, the cultured A. hydrophila was added to the aquarium tank to the concentration of 1 × 10 8 CFU/ml. After being treated for 40 min, the carp were taken out and transferred to a tank containing fresh water. On day 0, 1, 7, 14, 21 and 28 after stimulation, three sh from each group were anesthetized with Tricaine (MS-222) and the tissues samples including spleen and hindgut were taken and frozen in liquid nitrogen for total RNA extraction and then were detected by qPCR following the same procedure. Relative expression of immunoglobulin mRNA was determined using relative quanti cation by the comparative method 2 (−ΔΔCt), the level of target mRNA was normalized with respect to S11, an internal reference gene, and the results were expressed relative to the unchallenged sh as control (denoted by 0 day) [32]. For all experiments on live animals, we con rm that all methods were carried out in accordance with relevant guidelines and regulations. The protocol was approved by the Animal Experimental Ethics Committee of Shandong Normal University (Permit Number: AEECSDNU2017004). All efforts were made to minimize suffering. The study was carried out in compliance with the ARRIVE guidelines.

Consent for publication
Not applicable.

Availability of data and materials
The dataset supporting the conclusions of this article is available in the GenBank (https://www.ncbi.nlm.nih.gov/nuccore/1806102469) and the accession number is MN170744.1.