Phylogenetic analysis of all available SNADs sequences in common carp (Cyprinus carpio) (GenBank ID: ADV68699; XP_042609327; XP_042609326; XP_018918723; XP_042610696; XP_042590618; XP_042627763; XP_018965652; XP_018976522; XP_042606770; XP042627685; XP_042596168; XP_042600158; XP_042596170; XP_042599546), zebrafish (Danio rerio) (GenBank ID: NP_001373174; NP_001373175; XP_021322222; NP_001373168) rainbow trout (Oncorhynchus mykiss) (GenBank ID: XP_036825965; XP_021439285; XP_036805675) and Atlantic salmon (Salmo salar) (XP_0140067042) were analyzed for their phylogenetic relationships to other SNAD sequences found in GenBank using the tools available at www.phylogeny.fr (8). Sequence alignment was conducted with the full mode of MUSCLE algorithm (9). Phylogenetic analyses, based on the maximum likelihood method, were performed by SH-like Approximate Likelihood-Ratio Test (aLRT) using the PhyML software (10). Subsequently, the phylogenetic tree was visualized and rendered using TreeDyn (11).

The genomic structure and function of SNAD1 genes was investigated by BLAST analysis of all SNAD1 LOCs. The following databases were checked: common carp reference genome ASM1834038v1 (GCF_018340385), zebrafish reference genome GRCz11 (GCF_000002035), rainbow trout reference genome USDA_OmykA_1.1 (GCF_013265735) and Atlantic salmon reference genome ICSASG_v2 (GCA_000233375). We also searched the expressed sequence tags (EST) and transcriptome shotgun assembly (TSA) sequence databases to check if the gene mRNA had been previously detected. In addition, we screened our in-house common carp blood and seminal plasma samples for the presence of SNAD1 proteins.

Multiple sequence alignment was conducted using the Clustal Omega program available at www.ebi.ac.uk/services (12) and visualize (including 2D structure) using ESPript 3.0 program (13) available at https://espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi. 3D structure models were predicted ab initio for selected SNAD1 members using the RoseTTAFold method (14) on the Robetta server (https://robetta.bakerlab.org/submit.php). Modelled structures were visualized and analysed using UCSF Chimera software (15). To identify signal peptides we used ProtterServer (16) available at www.wlab.ethz.ch/protter/start/. To generate superposition of 3D structures, we used Pairwise Structure Alignment (17, 18) tool available at www.rcsb.org/docs/tools/pairwise-structure-alignment.

The European Bioinformatics Institute Gene Expression Atlas (http://www.ebi.ac.uk/gxa/) was queried for baseline gene expression data for the si:dkey-96g2.1 gene (encoding a putative SNAD1 homolog) in different developmental stages of zebrafish (19). Baseline gene expression of RNA samples extracted from whole zebrafish embryos at 18 different developmental stages from 1 cell to 5 days post-fertilisation was based on data derived from White et al. (20). Transcripts per million (TPM) were calculated from the raw counts by Integrated RNA-Seq Analysis Pipeline (iRAP). Subsequently, they were averaged for each set of technical replicates, and then quantile normalized within each set of biological replicates using Linear Models for Microarray Data (limma). Finally, they were averaged for all biological replicates (20).

We compared the reference of normal gene expression patterns of si:dkey-96g2.1 (ENSDARG00000097725), a putative SNAD1 homolog, among different zebrafish tissues, including: liver, mesonephors, head kidney, spleen, bone element, intestine, granulocyte, tail, swim bladder, larva, zone of skin, muscle tissue, structure with developmental contribution from neural crest and head using the Bgee database (https://www.bgee.org/) (21), which is based exclusively on curated healthy wild-type expression data (e.g., no gene knock-out, no treatment, no disease) and is annotated to the Uberon ontology of anatomy.

Utilizing PubMed and Google web browser we searched previously published articles in the field of immunology, using the following keywords: “fish immunology”, ,,fish infections”, ,,fish immunization”, “fish pathogens”, “fish viruses”, “acclimation to cold in fish” and “stress response in fish” in terms of presence of “uncharacterized proteins” and transcripts, which showed significant changes. Using the NCBI protein and genomic databases and The Zebrafish Information Network (ZFIN), providing genetic and genomic data for the zebrafish (Danio rerio), along with the Basic Local Alignment Search Tool (BLAST) we checked all of found “uncharacterized proteins” sequences and/or “uncharacterized proteins” accession numbers, as well as, to date “uncharacterized” genomic sequences of SNAD1 (gene si:dkey-96g2.1; ENSDARG00000097725) for the presence of SNAD1. Results from NCBI protein database about the similarity to SNAD1 were based on information from Conserved Domain Databases (CDD), a protein annotation resource that consists of a collection of well-annotated multiple sequence alignment models for ancient domains and full-length proteins.

Phylogenetic analysis using all available SNADs in common carp, zebrafish, rainbow trout, and Atlantic salmon, and earlier described SNAD1, 2, 3, and 4 proteins revealed a distinct separation between different SNADs. Furthermore, the SNAD1s show several distinct branches. Within the SNAD1, we found additional separation into four distinct clades withing the branch formed by warm-adopted fish ( Figure 3 ). Interestingly, cold water-adapted salmonids from additional SNAD1 cluster which was different from SNAD1 formed by those of warm water-adapted cyprinids, cichlids, and serrasalmids ( Figure 3 ). Salmonids SNADs are also present in the SNAD3 cluster.

Given the large number of SNAD1 sequences found in common carp, we performed an in silico analysis of these sequences. The SNAD1 sequences were BLAST searched against the common carp genome (ASM1834038v1 in the NCBI database). The number of exons, the genomic location (chromosome number and nucleotide region) and experimental validated expression presence and regulation type of 13 genes encoding potential SNAD1 in carp after environmental and pathogenic challenges were reported (see below 3.72 and 3.73) ( Table 1 ). Sequences were also checked against EST, TSA and proteome databases, which may indicate that genes are expressed or present as protein in their host. Of the thirteen SNAD1 genes, eleven were located on a defined chromosome in the carp genome. One of the genes (LOC109045318) mapped to an undefined chromosome is expressed in the liver (EST CA967515) and present in the seminal plasma proteome and is therefore most likely functional. The mRNA of four genes LOC109051800, LOC109107769, LOC122138962, LOC109104160, LOC109070810 [also known as Cap31, ref (7)] were detected in the transcriptome shotgun analysis sequence database, LOC122138962, LOC122140448 and LOC109070810 were also present in the EST libraries. In zebrafish, three SNAD1 genes have been reported and BLAST analysis against the GRCz11 reference genome showed that they are all located on chromosome 15 and have three exons. Searches of the TSA and EST libraries showed that zebrafish genes are widely expressed in various tissues ( Table 1 ). The same is true for the expression of the single SNAD1 genes in rainbow trout and Atlantic salmon, which have three and four exons respectively ( Table 1 ).

To check a structural similarity between predicted SNAD1 members, we performed primary amino acid sequence alignment of all SNAD1s identified in our phylogenetic analysis. To the comparison, we added also a well-characterized AAD representative (APOBEC3A) for which 3D structure with good resolution is available (PDB: 5SWW). The comparison of the primary structures revealed that almost all of the analyzed sequences (except XP_042609327.1 and XP_014353512.1) contained highly conserved zinc-coordinating motif, characteristic for zinc-dependent deaminases (1) in which two conserved Cys residues, one His residue and a water molecule coordinate a zinc ion, and glutamate (within HxE motif) acts as a proton donor ( Supplementary Figure 1 ). The context of the two conserved Cys residues was marginally different for SNAD1s and AADs (CVxxC versus Cx_2-4C for SNAD1s and APOBECs, respectively). However, this did not influence the common core fold of the catalytic center as evidenced by the location of conserved residues in the primary structure (shown in red in Supplementary Figure 1 ; similarity score 0.7) and the similarity of 2D structures (presented above the text in Supplementary Figure 1 ) for this catalytic region. To check the similarity between 3D structures of SNAD1s and 3D structure of the AAD representative, we performed ab initio modeling using the RoseTTAFold method. For the modeling, we selected 12 sequences, each corresponding to a divergent branch of SNAD1, marked by different colour shading in Figure 3 . Representative examples of the generated 3D structures are presented in Figure 4 and the whole set of predictions is included in Supplementary Figure 3 . We also performed superposition of the predicted 3D structures of SNAD1s from warm-water-adapted (XP_021322222.1 Danio rerio) and cold-water-adapted (XP_014067042.1 Salmo salar) fish ( Supplementary Figure 2A ), as well superposition of 3D structure of SNAD1 representative (XP_021322222.1 Danio rerio) and AAD representative (APOBEC3A, PDB: 5SWW) ( Supplementary Figure 2B ). In SNAD1 3D structures we identified the common core fold consisting of one β sheet surrounded by α helices. The topologic similarity between warm-water-adapted and cold-water-adapted SNAD1s (expressed as template modeling score) was high reaching 0.65. As expected, the template modeling score was slightly lower between AAD and SNAD1 reaching 0.4. The template modeling score measure ranges between 0 and 1, where 1 indicates a perfect match and 0 is no match between structures. Is accepted that scores < 0.2 indicate unrelated proteins while > 0.5 indicate the same protein fold. Moreover, we showed that the zinc-coordinating motif in the predicted catalytic centre is located at a position analogous to that in classic deaminases ( Figure 4 ; Supplementary Figure 3 ), and the catalytic centre is surrounded by structures equivalent to L1, L3, L5, and L7 in classic AADs ( Figure 4 ; Supplementary Figure 3 ). For 26 out of 38 analyzed SNAD1 proteins, we identified N-terminal signal peptides suggesting their secretory nature. The signal peptides differed significantly in both sequence and length (from 17 to 33 amino acids, on average 24). Interestingly, for six analyzed SNAD1 proteins (XP_016091766.1; XP_042627763.1; XP_016422879.1; XP_016105859.1; XP_042609326.1; XP_042609327.1) the N-terminal region formed a membrane-anchored structure. For almost all of the analyzed SNAD1 proteins (except XP_042609327.1 and NP_001373175.1), we identified a C-terminal cluster of three Cys residues, which according to the published data (4) seem to be characteristic of SNADs, and likely form disulfide bonds stabilizing two most C-terminal part of the enzyme backbone.

Among all of the analysed SNAD1 sequences, the only zebrafish gene available in the databases assigned as an “uncharacterized protein” similar to SNAD1 was si:dkey-96g2.1 (ENSDARG00000097725). Therefore, using an Expression Atlas (http://www.ebi.ac.uk/gxa) tool (19), we found that SNAD1 transcript was first transcribed at the end of embryogenesis (protruding mouth stage). In this stage, a low gene expression level (3 transcripts per million, TPM) was observed. In subsequent developmental stages, gene expression levels on larval days 4 and 5 were intermediate (10 TPM) and low (5 TPM), respectively ( Figure 5 ).

Using the Bgee database (21), we found the highest gene expression in the liver, mesonephros (posterior kidney), pronephros (head kidney) and spleen ( Figure 6 ).

We found SNAD1 sequences that were described as “uncharacterized proteins” and/or transcripts in 19 studies (key results, with emphasis on changes in SNAD1, are summarized in Table 2 ; Figure 7 ). The results of our search indicate the involvement of SNAD1 in various immunological processes, including the immune response due to bacterial infection or immunization, and its involvement in interactions with the intestinal microbiota. Moreover, we identified a relationship between SNAD1 and temperature acclimation, as well as its association with environmental pollution. Additionally, we demonstrated that SNAD1 levels are influenced by sex, exhibiting higher abundance in females, and that it plays a role in proper pancreas development (40, 41, 45).

We made a phylogenetic analysis of all available SNADs in common carp, zebrafish and rainbow trout in comparison to SNAD1, SNAD2, SNAD3 and SNAD4 members previously analysed by Kirshnan et al. (4) and SNAD1 and SNAD2 from fish previously examined by Bakke et al. (46). The sequences were analysed for their phylogenetic relationships to other SNAD sequences found in GenBank. Our analyses revealed a distinct separation between SNADs of cold water-adapted salmonids from those of warm water-adapted cyprinids, cichlids and serrasalmids. Within these groups, we identified additional division into two to three separate clades ( Figure 3 ). These shows that the SNADs are far more heterogenic family of proteins in fish. It should be noted, that the observed separation between cold and warm water-adapted fish species might be associated with the presence of overexpressed levels of SNAD1 as a response to cold exposure in fish (as discussed below). Additionally, due to unique cold adaptations of arctic fish, we checked them for the presence of SNAD1 motifs. Using the Motif Finer tool, we found homology of uncharacterized protein LOC117450837 (NCBI Reference Sequence: XP_033944723.1) from South Georgia icefish (Pseudochaenichthys georgianus) with SNAD1 (unpublished data). Further studies are needed to explore the potential role of SNAD1 in exceptional mechanisms of cold adaptation in arctic fish.

Classic AADs share similar core structural features with other zinc-dependent deaminases ( Figure 4A ). Their catalytic centre is characterized by the canonical structural motif HxEx25-30PCx2-4C, in which two conserved Cys residues, one His residue and a water molecule coordinate a zinc ion (47). The motif is located within one or two deaminase domains, whose typical fold comprises five β strands that form the enzyme backbone, which is surrounded by six α helices. The mechanism of enzymatic deamination is conserved within the whole superfamily of zinc-dependent deaminases and consists of nucleophilic attack at position C4 of the cytidine ring (or at position A6 of the adenine ring) by the activated water molecule, coordinated by the zinc ion and the conserved glutamate (which acts as a proton donor) (1). In support of this mechanism, mutation of glutamate or any of the zinc-coordinating residues results in a loss of enzymatic activity (48). In the structure of classic AADs, the conserved catalytic centre is surrounded by loops designated L1, L3, L5, and L7. Variability in loop length, amino acid composition, plasticity, and dynamics is believed to be critical for substrate sequence specificity (2, 49). In 3D structure models generated ab initio for selected SNAD1 members we identified the common core fold consisting of one β sheet surrounded by α helices (see Figure 4 ; Supplementary Figure 3 ) The zinc-coordinating motif in the predicted catalytic centres located at a position analogous to that in classic deaminases (see Figure 4 ; Supplementary Figures 1 , 3 ). Moreover, the catalytic centre was surrounded by structures analogous to L1, L3, L5, and L7, which correspond to those found in classical AADs.

Based on the common structural features of AADs and SNADs, Krishnan et al. proposed that SNADs play a role as cytosine deaminases of single-stranded nucleic acids (4). Our phylogenetic analysis supports this hypothesis, as the vast majority of sequences included in our analysis contained the expected highly conserved zinc-coordinating motif ( Supplementary Figure 1 ), supporting the classification of these proteins into the zinc-dependent deaminase superfamily.

We identified common structural features of SNAD1 members and investigate their potential structural variations in various fish species. As expected, in the vast majority of the analysed proteins (26 out of 38), we identified N-terminal signal peptides, suggesting that these proteins might be secreted. Among proteins, the signal peptides differ significantly in both sequence and length (from 17 to 33 amino acids, on average 24). The high diversity of this region may indicate different secretory pathways for different SNAD1 proteins. Interestingly, in the case of six SNAD1 members (XP_016091766.1; XP_042627763.1; XP_016422879.1; XP_016105859.1; XP_042609326.1; XP_042609327.1), in our predictions, the N-terminal region formed a membrane-anchored structure.

For almost all of the analysed SNAD1 proteins (except XP_042609327.1 and NP_001373175.1) we identified the presence of a C-terminal cluster of three Cys residues, which are characteristic of SNADs (4). They likely form disulfide bonds stabilizing the C-terminal enzyme backbone. All the analysed proteins are also characterized by the presence of loops L1, L3, L5, and L7 surrounding the catalytic centre which corresponds to analogous found in classical AAD. Considering that these loops are likely responsible for substrate recognition (2, 4, 49) and both their length and sequence vary widely among closely related SNADs [see Figure 4 ; Supplementary Figure 3 , and ref ( 4 )], one could hypothesize that SNAD1 proteins recognize different targets or that a given type of target is structurally variable (as would be expected in the case of RNA substrates). The high flexibility of active-site loops may suggest additional degrees of freedom of a substrate in the catalytic pocket (50). Taking into account also secreted nature of SNAD1s we hypothesized RNA as their most likely substrate. Moreover, it is also possible that the variable length of these loops indicates different oligomerization statuses of SNAD1 members. The latter is rationalized by the proven involvement of loop 3 in the dimerization of tRNA adenosine deaminase (TadA) enzyme (from which AAD family members are believed to originate) (51).

In summary, the conservation of structures among SNAD1 members from different fish species supports the classification of these proteins into a common subfamily. The conservation is additionally supported by the relatively high template modeling score between 3D structures of SNAD1 members from cold- and warm-water-adapted fish (see Supplementary Figure 3 ). All analysed SNAD1 proteins exhibit structural features of zinc-dependent deaminases, including zinc-coordinating catalytic centres and similar backbones. The high variability of SNAD1 structures in the region of signal peptides and loops surrounding the catalytic centre suggests possible specialization of these enzymes, especially in terms of different secretory pathways and recognized nucleic acids.

Based on RNA-seq data, SNAD1 is first transcribed in zebrafish after hatching, which suggests its physiological importance during that period and later ( Figure 5 ). This indicates that SNAD1 is not important during embryo development period. Moreover, we found the highest SNAD1 gene expression in the liver, mesonephros (posterior kidney), pronephros (head kidney) and spleen ( Figure 6 ). Notably, all of these organs play a pivotal role in the fish immune system (52). The head kidney is the dominant haematopoietic organ, described as an analogue of bone marrow in higher vertebrates, and functions as the primary haematopoietic tissue in fishes, a description also applied to the spleen and mesonephros (52, 53). In teleost fish, the kidneys contain B and T lymphocytes, dendritic cells, neutrophils, macrophages, granulocytes, thrombocytes and melano-macrophage centres (MMCs). Moreover, mature naive IgM+ cells and IgD+ cells produced by lymphoid progenitors are able to migrate to the mesonephros and spleen. The spleen and head kidney are the main sources of thrombocytes involved in immune responses (e.g., phagocytosis and intracellular killing of pathogens) and express various components of the major histocompatibility complex (MHC) (52).

Due to the presence of AID, B cells and CD4+ T cells, the MMCs present in the kidney and liver are considered as a primitive germinal centres (GCs) able to trap and store antigens (52, 53). The results of constitutive mRNA expression analyses are consistent with our hypothesis that SNAD1 could be a crucial component of the fish immune defence system. This suggestion is strongly supported by the changes in SNAD1 levels during various immunological-related processes (see below).

Despite the potentially important roles of SNADs in fish biology, SNADs have thus far escaped the attention of the majority of scientists. To date, in many publications, raw results regarding SNAD1 expression changes have been available but were not included in the general discussion due to their “uncharacterized” status. However, recently, we identified homology between these sequences and SNAD1, which allowed us to interpret obtained results and speculate about SNAD1 possible functions and its involvement in various biological processes.

We identified SNAD1 as a protein whose expression is highly modulated during immune responses. Possible roles of this protein have been suggested by several studies related to innate immunity, enhancement of innate immunity by cold conditions, stress responses and adaptive immunity in fish ( Table 2 ; Figure 7 ).

The discovery of multiple SNAD1 variants suggests a complex and potentially varied role in fish biology. The exact functions and interactions of these different SNAD1 variants are still unknown. Therefore, our study serves as an initial step in this exploration, highlighting the need for further experiments to fully understand the diversity and functional implications of multiple SNAD1 variants in fish (e.g. carp). Such investigations will be crucial in unraveling the comprehensive role of SNAD1 in fish physiology and its adaptative responses to environmental changes.

Our study revealed that several SNAD1 gene variants responded differently to environmental stressors and pathogenic challenges, underscoring the SNAD1 genes' complex regulatory mechanisms and their potential role in the carp’s adaptative responses. This dual sensitivity to infections and temperature changes highlights the importance of SNAD1 in the carp’s response to environmental changes and pathogenic challenges, possibly contributing to thermotolerance and immune defense mechanisms.

This study’s findings underscore the SNAD1 gene family’s complexity and its potential contributions to the common carp’s adaptative strategies. At present, we can only speculate that the differential expression of SNAD1 variants in response to infection and temperature stressors reflects a sophisticated network of gene regulation, providing insights into the genetic mechanisms underlying stress resilience and pathogen resistance in aquatic organisms - highly characteristic of common carp, one of the most evolutionarily successful fish species, ubiquitous in multiple environments (61). Further research into the specific functions and regulatory pathways of these SNAD1 variants will be crucial in understanding their roles in the common carp’s biology and potential applications in aquaculture and conservation efforts.