T-cell receptors (TR) recognize peptide antigens and other pathogen-derived molecules presented by antigen-presenting cells (APCs), a process essential for initiating adaptive cellular immunity. These receptors are expressed on the surface of T lymphocytes and enable specific antigen recognition through their variable extracellular domains. TR-mediated recognition is coupled to signaling via the CD3 complex (conformed by CD3γ, CD3δ, CD3ϵ chains), which transmits activation signals that drive T-cell activation, effector and memory differentiation, and clonal expansion (1, 2).

The general organization of TR genes has remained remarkably conserved throughout 400 million years of gnathostome evolution. Unlike the high variability of immunoglobulin loci, TR loci exhibit structural stability across vertebrates (3). A conventional TR is a disulfide-linked heterodimer composed of α and β chains, or γ and δ chains. Each of these four types of TR chains comprises two immunoglobulin superfamily domains: a membrane-proximal constant domain (C) and the antigen-binding variable domain (V). The V domains of TRβ and TRδ are assembled via somatic recombination of variable, diversity (D), and joining (J) genes, whereas the V domains of TRα and TRγ are assembled only by V and J genes (4). This recombination is mediated by recombination signal sequences (RSS) flanking each gene, consisting of a conserved heptamer and nonamer motif separated by either a 12- or 23-base pair spacer. According to the 12/23 rule, recombination typically occurs between one RSS with a 12-bp spacer and another with a 23-bp spacer, ensuring proper assembly of V(D)J junctions (5).

The TR α, β, γ, and δ chains are found across all jawed vertebrates, exhibiting significant conservation in both sequence and genomic arrangement. A distinctive feature is that the T cell receptor alpha (TRA) locus is embedded within the T cell receptor delta (TRD) locus (TRA-TRD locus), an organization conserved in all jawed vertebrates studied, including fish, amphibians, reptiles, birds, and mammals (6–11). However, the availability of non-model vertebrate genome sequences provides valuable insights into the distant origins of rearranging gene systems and their links to both adaptive and innate recognition processes (12). This approach has led to the identification of additional TR chains, such as the New Antigen Receptor (NAR-TCR) in sharks (13–15) and the T cell receptor μ (TRμ) chain in marsupials and monotremes, which originated from TRδ gene duplication during early mammalian evolution (16). Furthermore, in Xenopus tropicalis, the TRD locus contains canonical variable δ (Vδ) genes and VH-like genes termed VHδ, which are VH domains related to the variable domain of the immunoglobulin heavy chain, adapted as V-domains for TRδ chains (6). VHδ is also found in fish, birds, and monotremes. Overall, this highlights the remarkable evolutionary plasticity of TR evolution, likely due to selective pressure imposed by pathogen recognition (14, 17, 18).

Amphibians are well-suited models for comparative immune system analysis (1, 19), due to their key evolutionary relations as the first tetrapods, bridging aquatic vertebrates (e.g., fishes) and terrestrial vertebrates (20, 21). Their immune system comprises all major components of adaptive immunity, including T and B lymphocytes, immunoglobulins, and Major histocompatibility complex (MHC) molecules, enabling direct comparisons across both ancestral and derived vertebrate lineages. In addition, exhibit unique immunological features, such as unconventional TR gene arrangements or limited receptor diversity (6), offering insights into the evolutionary plasticity of the immune system. The study of TR loci in amphibians like Ambystoma mexicanum is particularly relevant because the immune system is increasingly recognized as a critical player in tissue repair and regeneration. Characterizing the genomic architecture and diversity of these loci not only informs our understanding of adaptive immunity in urodele amphibians but also provides a framework for investigating how immune components modulate regenerative processes. Such knowledge could facilitate the development of species-specific immunological tools, enhancing both biomedical research and conservation strategies.

A. mexicanum, a neotenic urodele amphibian endemic to the Mexico City valley, is an endangered species (22) and has one of the largest genome (32 Gb) among vertebrates sequenced to date (23). Previously, we characterized the immunoglobulin heavy (IGH) and lambda (IGL) loci in the Ambystoma mexicanum, finding that it shares the same general syntenic architecture with X. tropicalis, but lacks the kappa locus (IGK) and other antibody features described in X. tropicalis (24). Pre-genomic studies in A. mexicanum revealed the presence of T lymphocytes found in the spleen and thymus, as well as the presence of T cells expressing α, β, and δ chains (25–27). Of note, the junctional diversity of the TRδ chain is minimal (27, 28) and so far, no description of TRγ chains has been provided. Two genome assemblies are currently available for A. mexicanum: The AmbMex60DD genome assembly, based on a highly inbred laboratory strain (d/d), a two-year-old leucistic male (23, 29) has 27,157 unmapped scaffolds and revealed several positional and orientation inconsistencies in the IGH locus, likely reflecting assembly errors (24). Recently, a new assembly, UKY_AMEXF1_1, generated from an F1 hybrid between A. mexicanum and A. tigrinum, both of wild origin has been publicly released. This assembly, which is currently the reference genome in GenBank, presents an improved chromosomal organization with 21 chromosomes and only 220 unmapped scaffolds (BioProject: PRJNA1165261), suggesting a more accurate annotation of complex loci.

To further investigate the germ-line structure of T cell receptors in A. mexicanum, we present here a genomic characterization and annotation of TR loci in the axolotl compared with X. tropicalis. This is one of the few amphibian species whose adaptive immune system has been extensively characterized at the genomic level (6, 30). In our previous analysis of the IGH and IGL loci in A. mexicanum, we reported the absence of certain components, such as the IGK locus and the pseudogenization of the IgF isotype (24). This feature had also been described in X. tropicalis. Building on these findings, one of the main objectives of the present study is to determine whether, as observed in the case of immunoglobulins, TR loci in A. mexicanum also exhibit missing or divergent components compared with other tetrapods.

The UKY_AMEXF1_1 genome assembly of A. mexicanum enabled a comprehensive characterization of its TR loci, supported by an increased N50 = 1.5 compared with the AmbMex60DD genome version N50 = 1.2. The overall structure of the TRA-TRD locus is conserved, but the TRD locus exhibits strikingly low combinatorial diversity, and the TRG locus is absent, suggesting that bona fide TRγδ T cells may be lacking in axolotls. The TRB locus displays a conserved structure with tandem duplications of the TRBD-TRBJ-TRBC translocon, including a particular TRBC gene with two Cβ domains. As in X. tropicalis, the PTCRA gene is absent.

These findings provide important insights into the genomic organization and evolutionary constraints of TR loci in urodele amphibians. The limited diversity of TRD genes and absence of the TRG locus highlight unique features of the axolotl adaptive immune system, which may have implications for understanding T-cell function in regeneration and immune response. Overall, this work establishes a foundation for comparative immunogenomic studies across amphibians and other vertebrates.

Here we confirm such findings. We found similarities in the overall structure of the TRA-TRD locus, as well as a strikingly low combinatorial diversity of the TRD locus and the absence of the TRG locus, implying the absence of bona fide TRγδ T cells in axolotl. As for the TRB locus, we describe a conserved structure, with tandem duplications of the TRBD-TRBJ-TRBC translocon and a particular TRBC gene composed of two Cβ domains. As in X. tropicalis, the axolotl genome also lacks the PTCRA gene.

A. mexicanum’s exceptionally large genome (32 Gb) posed significant challenges for its assembly. The AmbMex60DD (white strain, d/d) version was released in 2021, based on 30× genomic coverage of 28 chromosomes and 27,157 unmapped scaffolds. This assembly presented several positional and orientation inconsistencies in the IGH locus, likely reflecting assembly errors (24). The current genome assembly (UKY_AMEXF1_1) was released in 2024 and has an increased genomic coverage (48×), 21 chromosomes, and only 220 unmapped scaffolds, suggesting that a more accurate complex loci annotation can be achieved. However, as with AmbMex60DD, no publicly available data on local coverage is currently provided for UKY_AMEXF1_1. Therefore, it is not possible to assess coverage-based metrics for individual TR genes.

In agreement with our previous analyses using the AmbMex60DD assembly, the overall expression patterns of TR genes did not substantially change in the present study ( Supplementary File 4 ; Supplementary Table 3 ). Remarkably, we identified two genes, TRAV_060 and TRDV_001, that exhibit a structurally complete configuration, including the SP, canonical donor and acceptor splice sites, conserved methionine’s at positions 23 and 104, and canonical RSS. However, neither of these genes showed detectable expression in the analyzed tissues. Conversely, two additional genes, TRAV_025 and TRBV_012, also retained an intact genomic organization and displayed transcriptional evidence of the V gene, yet lacked detectable expression of the corresponding SP. According to our classification criteria, these cases were therefore categorized as pseudogenes, since evidence of both V gene and SP expression was required to consider a gene as functional.

Despite the mentioned limitations, the remarkably stable of the TRA-TRD locus across species, maintaining a consistent genomic architecture for over 340 million years of evolutionary history, highlights strong evolutionary constraints on its organization (6). The fact that TRA and TRD remain genetically linked in all examined species (14, 44–46) reinforces the functional importance of their physical association. This overall structure, conserved in X. tropicalis, alligators, birds, and mammals. In all of these species, the TRA-TRD locus is flanked by the METTL3, SALL2, DAD1, and ABHD4 genes (6, 9, 14, 16, 45, 47, 48) suggesting that this syntenic arrangement may be critical for maintaining locus integrity. In contrast, the distinct configuration in A. mexicanum; located on chromosome 13p, near the centromere and flanked by KLHL33, likely reflects a lineage-specific chromosomal rearrangement that preserved the internal gene order, indicating that positional changes do not necessarily disrupt locus function ( Figure 2 ) ( Supplementary File 1 ; Supplementary Figure S1 ).

In A. mexicanum, the residue equivalent to mammalian TRα Arg251 is replaced by Lys, a substitution that is unlikely to affect its functional contact with CD3δ, given the chemical similarity of both residues (34, 35). The connecting peptide motif in the Cα transmembrane region is conserved, maintaining its role in signal transduction from the αβ heterodimer to the CD3/ζ complex (36).

Interestingly, as in X. tropicalis, A. mexicanum lacks the AB-loop, a structural feature essential for CD3ϵδ contact and T cell activation in mammals (34, 49). The absence of the AB-loop in amphibians may reflect a distinct co-evolutionary trajectory of TR and CD3 complexes compared to mammals (50).

We confirm the restricted diversity of germline genes previously reported for the TRD locus. As described by André, et al. (28) the cause of such restricted diversity is determined by a single functional TRDV and two TRDJ genes, but significantly, we confirm the absence of TRDD genes in the germline. As in other tetrapod TRD loci, the V and J genes are flanked by 23-bp and 12-bp spaced RSS, respectively. Hence, direct TRDV-TRDJ junctions do not violate the 12/23 rule (51). TRDV-TRDJ junctions have been described in a subset of human acute lymphoblastic leukemias (52). To our knowledge, A. mexicanum is the first vertebrate capable of non-pathological direct TRDV-TRDJ recombination.

A notable feature of the TRA-TRD locus in the axolotl is that the only TRDV gene is of the conventional Vδ type and not of the VHδ type, in striking contrast to X. tropicalis, which contains 14 VHδ and 2 conventional Vδ genes (6). The VHδ type is widely distributed among non-placental mammals and other vertebrates (9, 10, 15, 16, 48, 53, 54), and it remains to be determined if this type of element was lost in all or some caudates or if it was never present.

Compared with the TRα and β chains, the presence of TRγ and TRδ chains shows more heterogeneity across different taxa (32). In scaled reptiles, there is an absence of the TRD and TRG loci (55). Although the syntenic blocks flanking the TRG locus in Xenopus were identified in A. mexicanum, no genomic and transcriptomic evidence of the presence of the TRG locus was found. These results indicate that axolotl lacks true γδ T cells. It remains to be determined the functional role of the single TRδ chain, and if it pairs at all with itself or with another chain, but due to its invariant structural nature, it may function essentially as a Pattern Recognition Receptor (PRR), similarly to BTNL/Btnl family of innate γδ TR (56).

In this study, we updated and expanded the genomic annotation of the TRB locus located on chr3p previously described using cDNA libraries (57–59). The TRB locus in tetrapods is generally organized into TRBD-TRBJ-TRBC units, resembling the organization of the lambda chain locus (60). For instance, sheep possess three TRBD-TRBJ-TRBC tandem units, while rabbits, mice, and humans have two (42, 61, 62). In comparison, X. tropicalis has a simpler configuration with a single unit. Remarkably, A. mexicanum displays a more complex arrangement, consisting of five tandem TRBD-TRBJ-TRBC units.

In conventional TRβ chains, the Cβ domain interacts with CD3 through the FG-loop, contributing to signal transduction upon MHC-peptide recognition. In mammals, mutations in the FG-loop alter the CD8+ and CD4+ T cell proportions, and it is associated with a poor antigen response (63). Our study in axolotl is in agreement with a previous report by Kim, et al., reporting the absence of the FG-loop in non-mammal vertebrates ( Supplementary File 1 ; Supplementary Figure S8 ) (40, 41). We identified a novel TRBC gene (TRBC_003) that features two Cβ domains, along with transmembrane and cytoplasmic domains, each encoded by a separate exon. The TRBC_003 gene is actively transcribed in the spleen and associated with putatively functional Dβ and Jβ genes. Moreover, in the distal membrane exon (TRBC_003_1), there is a conserved Pro232, which is a relevant position of the FG-loop in mammals with a single N-glycosylation site. In contrast, in the proximal membrane exon (TRBC_003_2), there is an Ala in the 228 position and no N-glycosylation sites. This structural evidence suggests that the proximal exon may have arisen from a duplication of the distal exon. Whether a TRβ product of the TRBC_003 gene pairs with a TRα chain, how it interacts with CD3, and how MHC-peptide interaction takes place remain as open questions.

Genome sequencing of F1 crosses between A. mexicanum and A. tigrinum have revealed genomic regions of high polymorphism (64, 65), however the TR loci are outside these regions. We consider that spurious contributions of allelic variation to gene count are likely minimal. Moreover, genome information derives from a single individual, and our findings may thus fail to capture the extent of intra-species variability, particularly copy number variation in Adaptive Immune Receptor Repertoire (AIRR) loci, which are well documented in mice, macaques, and humans (66–68), that may explain discrepancies between the UKY_AMEXF1_1 and the AmbMex60DD assemblies.

We observed that many TRAV genes in A. mexicanum possess notably long V-introns, some exceeding 1,000 bp; and are considerably larger than IGH and IGL V-introns in A. mexicanum (24). In the Gallus gallus genome, the average TRB V-intron length is 400 bp (69). The biological implication of longer TR V-introns is uncertain. Long introns have been reported to impose evolutionary costs by increasing the energetic demand due to the greater nucleotide investment and extended transcription time required. Additionally, they may compromise the fidelity of mature mRNA and expand the sequence space available for allelic variation and aberrant splicing (70). Intron length has also been suggested to play a functional role in evolutionary dynamics (71). Comeron and collaborators (72) proposed that extensively long introns may enhance the efficiency of natural selection by alleviating Hill-Robertson (HR) interference, a phenomenon where selection acting on linked loci reduces selective efficacy. In this context, long V-introns may act as spacer regions that decouple selective pressures acting on neighboring functional elements.

Moreover, HR interference might prevent their elimination via purifying selection if these introns are linked to active V genes. Consequently, such introns could facilitate the emergence of alternative splicing events or the production of non-functional transcripts, potentially affecting the expression and functionality of the antigen receptor repertoire (70, 73, 74). Collectively, these findings reinforce the idea that introns do not merely represent structural and energetic burdens but may also play key roles in gene generation, conservation, and diversification; particularly in immune-related loci such as those encoding T cell receptors.

It is noteworthy that the functional-to-pseudogene ratio in the A. mexicanum TRB is 3.6 and 3.06 in the TRA locus, compared with 0.9 and 2.7 in the IGH locus (AmbMex60DD) ( 24) and UKY_AMEXF1_1 ¹ assemblies, respectively. Within the framework of the “birth-and-death” model of T and B cell receptor gene evolution (75), higher ratios may indicate a slower accumulation of pseudogenes, potentially reflecting more recent functional gene birth events, stronger purifying selection, or lower rates of gene inactivation in T cell receptors compared with B cell receptors. While such differences could be stochastic, they may also reflect distinct selective pressures acting on these repertoires. The comparison between genome versions further shows that improvements in assembly quality can refine gene counts and alter calculated ratios, underscoring the importance of high-quality chromosome-level genome assemblies and thorough manual curation of AIRR loci for robust evolutionary inferences (76).

Its absence in amphibians implies that early αβ T cell development proceeds via alternative mechanisms, possibly involving different surrogate chains or signaling pathways. This reinforces the hypothesis that PTCRA originated as an amniote-specific innovation, rather than an ancestral gene lost independently in teleosts and amphibians. Identifying how amphibians compensate for the absence of PTCRA could provide insights into the evolution and diversification of T cell developmental programs in vertebrates.

The review and characterization of immune components in A. mexicanum are essential for understanding the cellular processes in which cells interact dynamically and persistently. These components have been shown to play a key role in modulating such interactions in other vertebrates with comparable regenerative capacity, influencing both the persistence of immune responses and the regulation of mechanisms underlying the regeneration of complex structures such as limbs, tail, heart, retina, and spinal cord. However, further studies are needed to clarify the specific role of T cells in this process in A. mexicanum (77–80).

In conclusion, the A. mexicanum TRB locus exhibits greater diversity than the TRA and TRD loci. Notably, no evidence of bona fide γδ T cells was found. This study leaves open questions regarding the composition of T-cell subpopulations and the pairing of TR chains, particularly the δ chain in the absence of the γ chain in axolotl. The presence of two constant domains in TRB_003 warrants further investigation to clarify their role in antigen recognition, functionality, and pairing with the α chain. Such insights are crucial to understanding the impact on this endangered species, which already presents a marked deficit of heterozygosity, reflecting substantial inbreeding and increasing vulnerability to infectious diseases (22).

This work provides valuable insights for comparative evolutionary analyses in tetrapods and advances our knowledge of immune response in caudate amphibians. Moreover, it may aid in solving specific questions regarding the role of acquired immunity in the regulation of the immune response implicated in tissue regeneration (81).