The African clawed frog, Xenopus laevis, is a model species for amphibians. Before metamorphosis, tadpoles do not efficiently express the single classical MHC class I (MHC-I) molecule Xela-UAA, but after metamorphosis, adults express this molecule in abundance. To elucidate the Ag-presenting mechanism of Xela-UAA, in this study, the Xela-UAA structure complex (pXela-UAAg) bound with a peptide from a synthetic random peptide library was determined. The amino acid homology between the Xela-UAA and MHC-I sequences of different species is <45%, and these differences are fully reflected in the three-dimensional structure of pXela-UAAg. Because of polymorphisms and interspecific differences in amino acid sequences, pXela-UAAg forms a distinct peptide-binding groove and presents a unique peptide profile. The most important feature of pXela-UAAg is the two–amino acid insertion in the α2-helical region, which forms a protrusion of ∼3.8 Å that is involved in TCR docking. Comparison of peptide–MHC-I complex (pMHC-I) structures showed that only four amino acids in β2-microglobulin that were bound to MHC-I are conserved in almost all jawed vertebrates, and the most unique feature in nonmammalian pMHC-I molecules is that the AB loop bound β2-microglobulin. Additionally, the binding distance between pMHC-I and CD8 molecules in nonmammals is different from that in mammals. These unique features of pXela-UAAg provide enhanced knowledge of T cell immunity and bridge the knowledge gap regarding the coevolutionary progression of the MHC-I complex from aquatic to terrestrial species.

Major histocompatibility complex molecules are the most important and dominant family of immune response genes that defend against pathogenic microorganisms and function in cancer clearance in the adaptive immunology of jawed vertebrates (1). The classical peptide–MHC class I complex (pMHC-I) is presented on APCs that are mainly recognized by CD8+ CTLs to induce and trigger T cell immunity (2). Current knowledge of T cell structural immunology can be summarized as follows: pMHC-I has evolved with a closed-end peptide-binding groove (PBG) formed by the α1 and α2 domains of the MHC class I (MHC-I) H chain (HC). With two α1 and α2 helices on top and a slightly curved β-sheet at the bottom, the PBG creates enough space for peptides with lengths of 9–13 aa (36). The evolutionary advantage of the classical MHC-I molecule is its polymorphism (7), which enables the formation of diverse PBG conformations and the ability to bind to a variety of T cell epitope peptide receptors (4, 8). Therefore, clarifying the structures of pMHC-I complexes from different species of existing jawed vertebrates will provide new knowledge and functional platforms of T cell immunity and reveal the evolutionary progression of MHC molecules.

The African clawed frog (Xenopus laevis) is an allotetraploid frog that originated from interspecific hybridization of diploid ancestors 17 million years ago (9). Xenopus MHC-I (Xela-MHC-I) molecules have exceptional characteristics, from the genome to the cDNA, that make the model unique. First, X. laevis has evolved a distinct metamorphosis process; during this process, its immune system undergoes tremendous changes, including thymus remodeling and changes in MHC gene expression (10, 11). Second, before metamorphosis, tadpoles do not express classical MHC-I molecules, but after metamorphosis, almost all cells express classical MHC-I molecules (12, 13). Third, there is only one classical MHC-I locus with dominant expression, whereas other classical MHC-I genes are silenced in polyploid individuals of X. laevis (1416). The genome of X. laevis was sequenced (9), which further confirmed that X. laevis maintained only a single classical MHC-I (Xela-UAA) gene (17, 18). In addition, comparative studies demonstrated a marked difference in adaptive T cell immunity between Xenopus adults and mammals (19). The mammalian equivalents of the central and peripheral lymphoid organs (the thymus and spleen), which provide essential immune response components (including classical MHC-I molecules, TCRs, signaling-related glycoproteins, and costimulatory and coinhibitory receptors), have also been identified in Xenopus despite its lack of lymph nodes (2022). Moreover, the involvement of MHC-I–mediated CD8+ T lymphocytes in Xenopus against viral infection and tumors has been discovered (23, 24). The characteristics of these species are fascinating; thus, X. laevis has been chosen as a pivotal “connecting” nonmammalian comparative model for immunological studies (9).

X. laevis has been used as an experimental model to assess T cell immunity in the susceptibility of amphibian hosts to an emerging iridovirus. Adult X. laevis frogs can eliminate frog virus 3 (FV3) iridovirus infection, but larvae cannot fully resist FV3 infection, resulting in the death of most infected larvae (25). Both larvae and adults possess CD8+ T lymphocytes and can produce immunological rejection via immune excretion to kill allogeneic tissue grafts and mount immune responses against viruses (12, 26, 27). However, because classical MHC-I molecules are expressed after metamorphosis, a logical fact is revealed: a real T cell immune response is established in adults (2729).

To elucidate the structure and Ag-presenting function of the single classical MHC-I molecule in X. laevis, we determined the structure of an X. laevis classical MHC-I with β2-microglobulin (β2m) and a peptide using x-ray crystallography. To our knowledge, this study is the first elucidation of a pMHC-I structure in amphibians. The species characteristics of pXela-UAAg are described in detail. The profile of Ag peptides presented by pXela-UAAg was confirmed. In addition, a series of important features were found to be conserved in all jawed vertebrates, and the features of pXela-UAAg are different from those of pMHC-I in mammals. The analysis of the pXela-UAAg molecule, such as the anuran-Ia–specific α2 domain motif, revealed a bulge on top of the peptide-binding domain, which is expected to sterically interfere with many potential TCR interactions. These findings provide core knowledge regarding the MHC structure and functionality and further bridge the knowledge gap regarding MHC evolution from aquatic to terrestrial species.

All peptides that potentially bound to Xela-UAAg were predicted by the NetMHCpan 4.0 Server (http://www.cbs.dtu.dk/services/NetMHCpan/) (30). The nonapeptides used in the study (see Supplemental Table I) and one synthetic random peptide library, Ran_9Xsplitted (9X, XXXXXXXXX, where X is a random amino acid other than cysteine), were synthesized and purified by reverse-phase HPLC and mass spectrometry (MS) (SciLight Biotechnology) with >95% purity. The lyophilized peptides were stored at −80°C and dissolved in DMSO at a concentration of 25 mg/ml before use as previously described (31).

The DNA fragment encoding the extracellular domain of X. laevis classical MHC-I, namely, Xela-UAAg (GenBank accession no. AAF03401.1) (14), which includes residues 1–273 of the mature protein and added NdeI and XholI sites, was synthesized with the codon preferences of Escherichia coli by Shanghai Invitrogen Life Technologies and then cloned into the prokaryotic expression vector pET21a(+) (Novagen) with restriction sites NdeI and XholI. The recombinant Xela-UAAg was expressed in E. coli Rosetta (DE3) as inclusion bodies and purified as previously described (3). Finally, Xela-UAAg was dissolved in 6 M guanidinium chloride buffer (6 M Gua-HCl, 50 mM Tris-HCl [pH 8], 10 mM EDTA, 100 mM NaCl, 10% v/v glycerin, and 10 mM DTT) to a protein concentration of 30 mg/ml. In addition, X. laevis β2m (Xela-β2m) (GenBank accession no. AAF37230.2), containing 99 aa of the mature protein, was synthesized with the codon preferences of E. coli by Shanghai Invitrogen Life Technologies, then cloned and expressed in E. coli Rosetta (DE3) as inclusion bodies, and purified as described above for Xela-UAAg. Finally, the Xela-β2m inclusion bodies were dissolved in 6 M guanidinium chloride buffer to a protein concentration of 30 mg/ml.

Dilution renaturation was used to assemble Xela-UAAg (pXela-UAAg) (32), which was refolded by gradually diluting Xela-UAAg/Xela-β2m/nonapeptide at a 1:1:3 M ratio at 277 K for 24 h using a previously described gradual dilution method (33). Then, the soluble portion of the refolded complex was concentrated and purified with a Superdex 200 16/60 Column, followed by Resource Q anion-exchange chromatography (GE Healthcare). The purified proteins were buffer-exchanged three times with 10 mM Tris-HCl and 50 mM NaCl at a pH of 8.0 (3).

pXela-UAAg folded with a library of random peptides was treated with 0.2 N acetic acid, incubated at 90°C for 5 min, and then concentrated with a 3 kDa filter to collect the peptides. The samples were desalted via a method previously developed in the laboratory (31). In brief, 200 μl of methanol was used to revitalize the desalting tips, and the tips were then equilibrated with 200 μl of 0.1% (v/v) trifluoroacetic acid (TFA). Then, the peptides were washed twice with 200 μl of washing buffer (0.1% [v/v] TFA) and eluted with 200 μl of eluting solution (0.1% [v/v] TFA and 75% [v/v] acetonitrile). The eluted peptides were lyophilized and stored at −80°C. The Easy Nano LC1000 system (Thermo Fisher Scientific, San Jose, CA) was used to separate the peptides. The experimental conditions and procedure were as previously described (31). A Q Exactive HF (Thermo Fisher Scientific, Bremen, Germany) in data-dependent acquisition mode was used to obtain the MS data, and the top 20 precursors by intensity from the mass range m/z 300–1800 were sequentially fragmented with higher energy collisional dissociation, normalized collision energy 27. The dynamic exclusion time was 20 s. Automatic gain control for MS1 and MS2 was set to 3e6 and 1e, and the resolution for MS1 and MS2 was set to 120 and 30 K.

Based on the spectrum information, each of the nonapeptides was determined by the software, and the peptide with the highest probability was derived from each spectrum. The identified polypeptides were adjusted by the detection threshold, sequence length = 9 and score ≥90 (34, 35). Based on the results of MS (Supplemental Table II) and the prediction of the online tools of the NetMHCpan 4.0 Server (http://www.cbs.dtu.dk/services/NetMHCpan/), we selected YMMPRHWPI (YMM9, score = 0.75) and YMMPRHWPL (YMM9-P9L, score = 0.77) with the highest scores to form the pXela-UAAg complex.

The peptide YMM9 and mutant peptides (Supplemental Table I) were synthesized and purified as mentioned above and used to determine the binding pocket. Xela-UAAg was refolded with Xela-β2m, and the peptides were separated and purified by gel filtration chromatography as described above. The thermal stabilities of the pXela-UAAg complexes and their binding affinities with these peptides were tested on a circular dichroism instrument (Chirascan; Applied Photophysics) using previously described protocols (36). The midpoint transition temperature (Tm) was determined by denaturation curve data in the Origin 9.1 program (OriginLab) (37). Based on the position-specific scoring matrices and circular dichroism results, the theoretical nonapeptides derived from the target viral protein were scanned and scored (31, 38, 39) according to previously described methods (31).

The purified pXela-UAAg complex was ultimately concentrated to 10 mg/ml and mixed with reservoir buffer at a 1:1 ratio. The crystallization of pXela-UAAg was screened using the sitting-drop vapor diffusion technique at 277 K. The Index, Crystal Screen I/II and Crystal Screen Cryo I/II Kits (Hampton Research, Riverside, CA) were used to screen for optimal crystal growth conditions. The crystal of pXela-UAAg (with peptide YMM9) was obtained with PEG/Ion 2 solution No. 6 (0.2 M sodium malonate [pH 6], and 20% w/v PEG3350) at 277 K after 1 wk at a protein concentration of 8 mg/ml. Before x-ray diffraction, the crystals were soaked in a cryoprotectant containing 20% glycerol for several seconds in reservoir solution and then flash cooled in a stream of gaseous nitrogen at 100 K (40). Diffraction data were collected to a resolution of 2.8 Å at Beamline BL17U (wavelength 0.97892) of the Shanghai Synchrotron Radiation Facility (Shanghai, China). The data were autoindexed and integrated using the CCP4 program and then scaled and merged using the HKL-2000 software package (HKL Research).

The structure of pXela-UAAg was determined by molecular replacement with the MOLREP and PHASER program using BF2*0401 (Protein Data Bank [PDB] code: 4E0R, with the peptide excluded) as a search model (36). The unit cell of pXela-UAAg belongs to the P212121 space group. A comprehensive model was built manually by COOT (8), and the structure was refined with the REFMACS5 program. Refinement rounds were implemented using phenix as described (41). Finally, the PROCHECK program was used to assess the stereochemical quality of the final model. The details of the refinement are shown in Table I.

The structural illustrations and the related figures were generated by PyMOL (https://www.pymol.org). The multiple sequence alignment was performed by Clustal Οlust (https://www.ebi.ac.uk/Tools/msa/clustalo/) and ESPript 3.0 (http://espript.ibcp.fr/ESPript/ESPript/). Accessible surface area (ASA) and buried surface area values were calculated by the online web site PDBePISA (http://www.ebi.ac.uk/msd-srv/prot_int/pistart.html), and the B factor was calculated with CCP4, using the online tool Swiss model (https://swissmodel.expasy.org/interactive#structure) for homology modeling.

The pXela-UAAg structure was deposited in the PDB (https://www.rcsb.org/structure/6A2B).

The crystal structure of Xela-UAAg-β2m (pXela-UAAg) complexed with the peptide YMM9 from the library of random peptides was determined (Table I). The frame structure of pXela-UAAg is formed by four domains termed α1, α2, α3, and β2m (Fig. 1A). In Fig. 1A, a region is highlighted at the highest point of the α2 domain helical region (with residues P147 to N150) and will be addressed in detail below. The PBG of pXela-UAAg is formed by α1 and α2 domains and consists of two antiparallel helices and eight-stranded β sheets; the YMM9 peptide is located in the PBG (Fig. 1B). The α3 domain and Xela-β2m formed the base of pXela-UAAg and underpin the PBG. Compared with the known pMHC-I structures of other representative species (HLA-A*02:07 PDB: 3OXS, BF2*2101 PDB: 4CVX, and pCtid-UAAg PDB: 5Y91), pXela-UAAg is most similar to pCtid-UAAg (PDB: 5Y91) (3), with a root mean square deviation (RMSD) of 0.948, whereas the RMSDs with the compared chicken pBF2*2101 and human pHLA-A*0207 structures are 1.390 and 1.527, respectively (Fig. 1C).

Table I.
Data collection and refinement statistics (molecular replacement)
Data CollectionCrystal Data
Space group P212121 
Cell dimensions  
a, b, c (Å) 59.139, 68.220, 107.809 
 a, b, γ (°) 90, 90, 90 
 Resolution (Å) 50.0–2.8 
Rsym or Rmergea 0.244 (1.129)b 
I/sI 5.1 (0.8)b 
 Completeness (%) 92.5 (91.7)b 
 Redundancy 2.3 (1.9) 
Refinement  
 Resolution (Å) 40.0–2.8 (14.984–2.800) 
 No. reflections 11,143 
Rwork/Rfreec 0.206/0.2792 
 No. atoms 3076 
 Protein 
 Ligand/ion 
 Water 
 B-factors 8.44 
 Protein 43.907 (40.378) 
 Ligand/ion 28.528 
 Water 
RMSD  
 Bond lengths (Å) 0.0154 
 Bond angles (°) 1.9108 
Ramachandran statistics  
 Most favored (%) 94 
 Disallowed (%) 
Data CollectionCrystal Data
Space group P212121 
Cell dimensions  
a, b, c (Å) 59.139, 68.220, 107.809 
 a, b, γ (°) 90, 90, 90 
 Resolution (Å) 50.0–2.8 
Rsym or Rmergea 0.244 (1.129)b 
I/sI 5.1 (0.8)b 
 Completeness (%) 92.5 (91.7)b 
 Redundancy 2.3 (1.9) 
Refinement  
 Resolution (Å) 40.0–2.8 (14.984–2.800) 
 No. reflections 11,143 
Rwork/Rfreec 0.206/0.2792 
 No. atoms 3076 
 Protein 
 Ligand/ion 
 Water 
 B-factors 8.44 
 Protein 43.907 (40.378) 
 Ligand/ion 28.528 
 Water 
RMSD  
 Bond lengths (Å) 0.0154 
 Bond angles (°) 1.9108 
Ramachandran statistics  
 Most favored (%) 94 
 Disallowed (%) 
a

Rmerge = Σhkl Σi |Ii(hkl) – 〈I(hkl)〉 |/Σhkl Σi Ii(hkl), where Ii(hkl) is the observed intensity, and 〈I(hkl)〉 is the average intensity from multiple measurements.

b

Values in parentheses are for the highest-resolution shell.

c

Rhkl || Fobs | – k | Fcalc | |Σhkl | Fobs |, where Rfree is calculated for a randomly chosen 5% of reflections, and Rwork is calculated for the remaining 95% of reflections used for structure refinement.

RMSD, root mean square deviation.

FIGURE 1.

The frame structure of pXela-UAAg. (A) The structure of the pXela-UAAg and the special features of X. laevis in the PBG. Xela-UAAg, Xela-β2m, and the YMM9 peptide are colored according to B factor. The X. laevis Cα atoms of P147–N150 are shown with electron densities in yellow stick representation. The circled four amino acids are unique to frogs. The conformation of pXela-UAAg is similar to that of the known pMHC-I structures, consisting of four domains: α1, α2, α3, and β2m. The PBG consists of eight β sheets and two α helices. (B) The formation of the YMM9 peptide in the PBG of pXela-UAAg. (C) Conserved amino acids involved in the interaction between Xela-UAAg and Xela-β2m of pXela-UAAg. Representative species of pMHC-I molecules were selected for comparison. Among the amino acids in which the HC interacts with the L chain, Y12, H33, D55, and W62 of the β2m amino acids are conserved in humans, chickens, frogs, and grass carp. The root mean square deviation (RMSD) values are labeled. (D) Conserved residues found in the PBGs of solved pMHC-I structures. HLA-A*02:07 (3OXS) and BF2*2101 (4CVX) were used for comparison with pXela-UAAg. In addition to the conserved disulfide bond, 11 residues were found that are conserved in the known pMHC-I structures. These residues are located in the two ends of the PBG, and their positions are labeled by pXela-UAAg (6A2B) numbering.

FIGURE 1.

The frame structure of pXela-UAAg. (A) The structure of the pXela-UAAg and the special features of X. laevis in the PBG. Xela-UAAg, Xela-β2m, and the YMM9 peptide are colored according to B factor. The X. laevis Cα atoms of P147–N150 are shown with electron densities in yellow stick representation. The circled four amino acids are unique to frogs. The conformation of pXela-UAAg is similar to that of the known pMHC-I structures, consisting of four domains: α1, α2, α3, and β2m. The PBG consists of eight β sheets and two α helices. (B) The formation of the YMM9 peptide in the PBG of pXela-UAAg. (C) Conserved amino acids involved in the interaction between Xela-UAAg and Xela-β2m of pXela-UAAg. Representative species of pMHC-I molecules were selected for comparison. Among the amino acids in which the HC interacts with the L chain, Y12, H33, D55, and W62 of the β2m amino acids are conserved in humans, chickens, frogs, and grass carp. The root mean square deviation (RMSD) values are labeled. (D) Conserved residues found in the PBGs of solved pMHC-I structures. HLA-A*02:07 (3OXS) and BF2*2101 (4CVX) were used for comparison with pXela-UAAg. In addition to the conserved disulfide bond, 11 residues were found that are conserved in the known pMHC-I structures. These residues are located in the two ends of the PBG, and their positions are labeled by pXela-UAAg (6A2B) numbering.

Close modal

Although some variation in the relative α3 domain orientation is observed, as reported in previous studies that compared pMHC-I structures (3), the PBG and β2m domains superimpose very well (Fig. 1C). The similarities in β2m are highlighted by the structural conservation of four β2m residues, Y12, H33, D55 and W62, which form hydrogen bonds with the MHC-I HC (Fig. 1C, see Supplemental Fig. 1 for comparison of β2m sequences). Analysis of the interaction between Xela-UAAg and Xela-β2m showed that Y12, T13, Y30, H33, D55, W62, Y65, and Y66 of Xela-β2m formed hydrogen bonds with Xela-UAAg (Supplemental Fig. 1, in the green boxes). Among them, T13, Y30, Y65, and Y66 are different in humans, monkeys, mice, cats, pigs, caws, chickens, frogs, and grass carp; they showed species specificity, whereas the four amino acids Y12, H33, D55, and W62 were found to be conserved in almost all species (Fig. 1C, Supplemental Fig. 1).

As for the PBG domain, in addition to the disulfide bond (C99-C163), there are 11 residues (Y7, G26, Y58, W59, Y115, T140, K143, W144, Y158, L159, and Y170) that are well conserved in the PBGs of known classical pMHC-I structures (Fig. 1D). Many of these residues, namely, Y7, Y58, T140, K143, W144, Y158, and Y170, can interact directly by hydrogen bonds or indirectly by participating in hydrogen bond networks with the peptide ligand, which has been extensively reported for other pMHC-I structures (33, 36, 4244) and is discussed for pXela-UAAg below.

In addition to these commonalities, because the amino acid sequence identities among the Xela-UAAg and other species are <42.5%, there are obvious differences in the amino acid sequences among humans, mammals, chickens, and fish (Fig. 2). The pXela-UAAg structure showed a prominent distinctive feature with two more amino acids than the structures of other species (Figs. 1A, 2, the green box is marked with a pentagram). These additional amino acids result in a protrusion of ∼3.8 Å at the α2 helix (Fig. 1A); therefore, this protrusion caused by E148 and V149 is a feature unique to pXela-UAAg.

FIGURE 2.

Structure-based amino acid sequence alignment of pXela-UAAg and other representative crystallized pMHC-I molecules, with the secondary structure elements indicated. Black arrows above the alignment indicate β-strands; cylinders denote α helices. Green numbers denote residues that form disulfide bonds. The differences between mammals and the nonmammals are shown in green triangles, and the species-specific characteristics of X. laevis are shown in yellow pentagrams. Residue positions contributing to pXela-UAAg pockets are highlighted by pocket-specific colored shading. The total amino acid identities between Xela-UAAg and the listed MHC-I molecules are shown at the end of each sequence.

FIGURE 2.

Structure-based amino acid sequence alignment of pXela-UAAg and other representative crystallized pMHC-I molecules, with the secondary structure elements indicated. Black arrows above the alignment indicate β-strands; cylinders denote α helices. Green numbers denote residues that form disulfide bonds. The differences between mammals and the nonmammals are shown in green triangles, and the species-specific characteristics of X. laevis are shown in yellow pentagrams. Residue positions contributing to pXela-UAAg pockets are highlighted by pocket-specific colored shading. The total amino acid identities between Xela-UAAg and the listed MHC-I molecules are shown at the end of each sequence.

Close modal

Generally, in classical pMHC-I, a peptide is accommodated in a PBG area divided into six “pockets,” the regions of which were initially defined in human HLA-A2 (6). The six pockets (A–F) in the PBG of pXela-UAAg were confirmed (Fig. 3A). The interactions between these pockets and the anchor sites of the bound peptide are listed in Table II.

FIGURE 3.

Structural analysis and comparison of the six pockets and the peptide in the PBG of pXela-UAAg. Structural analysis and comparison of six pockets and peptide orientation in PBG. (A) Structure of pXela-UAAg PBG. (BG) Six pockets (A–F) of pXela-UAAg in surface polarity representation (blue, positively charged; red, negatively charged; white, nonpolar). The pocket residues were determined based on interaction with the peptide ligand as indicated by CCP4 software (Table II) and on our visual inspection of pocket continuity. The residue-accommodating pockets are listed under the relevant anchors in PBG and are shown in stick form. The hydrogen bonds between peptide and pockets are shown as yellow dashed lines. (H and I) Structure of the YMM9 peptide presented by Xela-UAAg. Electron densities and overall conformations of the structurally defined peptides. Xela-UAAg is shown as a cartoon model and colored according to B factor; the YMM9 peptide is shown in stick representation and colored based on the isotropic B-factors. (J) Side chain orientation of YMM9 peptide in pXela-UAAg structure in the side view from the peptide N terminus to the C terminus, as viewed in profile from the peptide N terminus toward the C terminus. An arrow pointing up indicates that an amino acid residue is oriented toward the TCR, down is toward the floor of PBG, left is toward the α1 helix domain, and right is for the α2 helix domain. ASA and buried surface area represent the exposed and buried surface area of each YMM9 peptide residue, respectively.

FIGURE 3.

Structural analysis and comparison of the six pockets and the peptide in the PBG of pXela-UAAg. Structural analysis and comparison of six pockets and peptide orientation in PBG. (A) Structure of pXela-UAAg PBG. (BG) Six pockets (A–F) of pXela-UAAg in surface polarity representation (blue, positively charged; red, negatively charged; white, nonpolar). The pocket residues were determined based on interaction with the peptide ligand as indicated by CCP4 software (Table II) and on our visual inspection of pocket continuity. The residue-accommodating pockets are listed under the relevant anchors in PBG and are shown in stick form. The hydrogen bonds between peptide and pockets are shown as yellow dashed lines. (H and I) Structure of the YMM9 peptide presented by Xela-UAAg. Electron densities and overall conformations of the structurally defined peptides. Xela-UAAg is shown as a cartoon model and colored according to B factor; the YMM9 peptide is shown in stick representation and colored based on the isotropic B-factors. (J) Side chain orientation of YMM9 peptide in pXela-UAAg structure in the side view from the peptide N terminus to the C terminus, as viewed in profile from the peptide N terminus toward the C terminus. An arrow pointing up indicates that an amino acid residue is oriented toward the TCR, down is toward the floor of PBG, left is toward the α1 helix domain, and right is for the α2 helix domain. ASA and buried surface area represent the exposed and buried surface area of each YMM9 peptide residue, respectively.

Close modal
Table II.
The interactions between the peptide and the PBG of pXela-UAAg
Hydrogen Bonds and Salt Bridgevan der Waals Forces
YMM9H Chain
ResidueAtomResidueAtom
P1-Y Y158 OH Y7, Y58, Q62, Y158, L162, G166, R169, Y170 (R169) 
Y170 OH 
Y7 OH 
P2-M    Y7, Y9, I24, Q62, I65, A66, Y97, Y158 
P3-M Y97 OH Y9, I65, S69, Y97, H111, R154, N155, Y158 
P4-P    I65, S69, R154 
P5-R    G68, S69, V72, H73, R154 
P6-H ND1 Y113 OH V72, H111, Y113, Y130, W144, A151, R154, N155 
P7-W    W144, V149, K143, V72, D76, Y113 
P8-P    D76, V72, H75, T79 
P9-I R83 NH2 W144, T140, K143, D76, T79, F120, R83 (R83) 
T140 OG1 
Hydrogen Bonds and Salt Bridgevan der Waals Forces
YMM9H Chain
ResidueAtomResidueAtom
P1-Y Y158 OH Y7, Y58, Q62, Y158, L162, G166, R169, Y170 (R169) 
Y170 OH 
Y7 OH 
P2-M    Y7, Y9, I24, Q62, I65, A66, Y97, Y158 
P3-M Y97 OH Y9, I65, S69, Y97, H111, R154, N155, Y158 
P4-P    I65, S69, R154 
P5-R    G68, S69, V72, H73, R154 
P6-H ND1 Y113 OH V72, H111, Y113, Y130, W144, A151, R154, N155 
P7-W    W144, V149, K143, V72, D76, Y113 
P8-P    D76, V72, H75, T79 
P9-I R83 NH2 W144, T140, K143, D76, T79, F120, R83 (R83) 
T140 OG1 

In the PBG of pXela-UAAg, the A pocket is composed of L5, Y7, F34, E54, Y58, W59, Q62, Y158, L162, C163, G166, R169, and Y170 and binds P1-Y via hydrogen bonds and multiple van der Waals interactions (Table II). The P1-Y main chain is bound by a hydrogen bond network, which is essentially conserved among pMHC-I structures (3, 32, 33) and involves three direct hydrogen bonds with Y7, Y158, and Y170; residue Y58 is expected to form an indirect hydrogen bond connection with P1-Y by means of a water molecule intermediate (Fig. 3A, 3B). Unusual for the pXela-UAAg A pocket, although (for example) also reported for a bovine pMHC-I allele (40) (Fig. 2), is the absence of the commonly well-conserved W166 residue and its replacement with a small glycine (discussed below). The pXela-UAAg B pocket consists of Y7, Y9, I24, V25, G26, F34, R35, Y36, A43, E44, and A66. This B pocket forms a “deep hole” between the helical region of the α1 domain and the β-sheet floor of the PBG, in which the side chain of P2-M is inserted and fixed (Fig. 3C). The depth of the B pocket is allowed by the small A43 residue. The pXela-UAAg C pocket region is formed by residues Y9, H73, V95, Y97, H111, and Y113 (Fig. 3D), most of which are quite bulky, resulting in little space and an upward orientation of the P5-R side chain. The D pocket consists of the residues Y97, H111, R154, N155, Y158, and L159 (Fig. 3E). The D pocket is a genuine pocket and harbors the side chain of the P3 residue (P3-M in this case), as is common among pMHC-I structures (3, 32). The E pocket is formed by residues H111, Y113, Y130, W144, A151, and N155 (Fig. 3F). The Y130 and A151 residues are smaller than usually found at this position (Fig. 2), which allows the bulky P6-H side chain to be inserted downward into the PBG. W144 is well conserved among MHC-I sequences, and in many elucidated pMHC-I structures, its side chain forms a hydrogen bond with the peptide ligand P8 residue main chain (3, 32); however, the orientation of the W144 side chain can differ among pMHC-I structures (3), and in pXela-UAAg, W144 does not participate in such hydrogen bonds. The F pocket is composed of D76, T79, A80, R83, L93, Y113, F120, I121, T140, K143, and W144 (Fig. 3G). The R83, T140, and K143 residues are well conserved among species, although in mammals, the R83 residue has been replaced by a tyrosine (45), and R/Y83, T140, and K143 can be involved in hydrogen bonding of the peptide ligand P9 residue main chain (32, 43). However, among pMHC-I F pockets, some flexibility can be observed; the orientations of the Y/R83, T140, K143, and W144 residues can differ, and they do not always participate in hydrogen bonding of the peptide ligand main chain (3, 36, 43). In the case of pXela-UAAg, only R83 and T140, but not K143, form hydrogen bonds with the P9 main chain. The side chain of the P9-I residue is inserted perpendicularly downward into the F pocket, which is common among pMHC-I structures. Fig. 3H, 3I are different depictions of the peptide orientation within pXela-UAAg. Moreover, an important feature is that the C–F pockets in the PBG of pXela-UAAg are positively charged and tend to bind to negative or neutral amino acids (Fig. 3D–G). The side chain of P1-Y lies flat in the A pocket, and the orientations of the P2, P3, P6, and P9 side chains are inserted into the B, D, E, and F pockets, respectively, whereas the side chains of residues P4, P5, P7, and P8 extend out of the PBG and into the solvent (Fig. 3H, 3I). The exposed ASA and buried surface area of the residue are shown in Fig. 3J.

YMM9 presented by pXela-UAAg showed the featured “M” conformation in PBG (Fig. 3H, 3I). Compared with peptides of other species, at the N terminus, the peptide presented by pXela-UAAg is relatively high and the side chain of P1-Y lies horizontally on the A pocket, whereas other species present the P1 residues upward (Fig. 4A). This is the result of F34 because Xela-UAAg F34 has a large side chain, whereas other species harbor V or M, which have small side chains. The side chain of F34 positions up to Y170, and the rising Y170 pushes up the main chain C atom of P1-Y (Fig. 4A, 4B), making the P1-Y main chain C atom higher than that from the other species, and a salt bridge between R169 and P1-Y flattens the side chain of P1-Y to the left (Fig. 4B). Because most of other species contain W166, which has a large side chain, the G166 side chain in pXela-UAAg is relatively small, which opens up the A pocket (Fig. 4B, 4C) and allows the ASA of the A pocket, as the largest among all the representative species (Table III), to accommodate amino acids with large side chains. At the same time, we also found that in the solved pMHC-I structure, the A pockets of nonmammals (frog, grass carp, chicken, and duck) are open and wide (Fig. 4D, red box), whereas in mammals, the A pockets are connected or relatively narrow (black box in Fig. 4D).

FIGURE 4.

Structural characteristics of defined peptide and A pocket. (A) Conformation of YMM9 and comparison with other peptides. YMM9 exhibits a typical “M” conformation. The remaining peptides are shown in different colors depending on their species and are depicted as ribbons. F34 of Xela-UAAg has a large side chain, whereas this residue position in other species is occupied by V or M; therefore, the side chain of F34 elevates Y170, which Y170 pushes the main chain of P1-Y upward, making the P1-Y C atom higher than in other species. (B) Characteristics of YMM9 P1-Y. A salt bridge between R169 and P1-Y flattens the side chain of P1-Y to the left. (C) Characteristics of the pXela-UAAg A pocket. Because of the presence of G166 (other species are V or W with large side chains), the A pocket of pXela-UAAg is open, and the ASA of the A pocket is the largest among all species (Table III). (D) pMHC-I PBG conformation of several representative species (nonmammalian A pockets are marked with red squares, mammalian A pockets are marked with black squares). In the determined pMHC-I structures, nonmammalian A pockets are open (frog, grass carp, chicken, and duck), whereas mammalian A pockets are either relatively narrow (HLA-A2 PDB: 3H9H, SLA1*0401 PDB: 3QQ3) or connected together (Mamu-B*17 PDB: 3RWC, H2*Kd PDB: 1VGK).

FIGURE 4.

Structural characteristics of defined peptide and A pocket. (A) Conformation of YMM9 and comparison with other peptides. YMM9 exhibits a typical “M” conformation. The remaining peptides are shown in different colors depending on their species and are depicted as ribbons. F34 of Xela-UAAg has a large side chain, whereas this residue position in other species is occupied by V or M; therefore, the side chain of F34 elevates Y170, which Y170 pushes the main chain of P1-Y upward, making the P1-Y C atom higher than in other species. (B) Characteristics of YMM9 P1-Y. A salt bridge between R169 and P1-Y flattens the side chain of P1-Y to the left. (C) Characteristics of the pXela-UAAg A pocket. Because of the presence of G166 (other species are V or W with large side chains), the A pocket of pXela-UAAg is open, and the ASA of the A pocket is the largest among all species (Table III). (D) pMHC-I PBG conformation of several representative species (nonmammalian A pockets are marked with red squares, mammalian A pockets are marked with black squares). In the determined pMHC-I structures, nonmammalian A pockets are open (frog, grass carp, chicken, and duck), whereas mammalian A pockets are either relatively narrow (HLA-A2 PDB: 3H9H, SLA1*0401 PDB: 3QQ3) or connected together (Mamu-B*17 PDB: 3RWC, H2*Kd PDB: 1VGK).

Close modal
Table III.
The ASA and buried surface area of the A pocket in different representative species
Buried Surface Area (Å2)ASA (Å2)
X. laevis (6A2B) 243.59 279.97 
HLA*0201 (3H9H) 201.7 236.42 
HLA-A*0207 (3OXS) 201.56 256.76 
Mamu-A*01 (1ZVS) 184.62 185.29 
Mamu-B*17 (3RWC) 209.42 224.69 
H2*kd (1VGK) 165.7 167.59 
Murine H2*kd (3TID) 155.69 159.1 
SLA*0401 (3QQ3) 183.94 207.92 
BoLAN*01801 (2XFX) 195.38 206.24 
BoLAN*01801 (3PWU) 191.46 216.9 
BF2*0401 (4E0R) 166.64 197.08 
Grass carp (5Y91) 218.19 238.38 
Buried Surface Area (Å2)ASA (Å2)
X. laevis (6A2B) 243.59 279.97 
HLA*0201 (3H9H) 201.7 236.42 
HLA-A*0207 (3OXS) 201.56 256.76 
Mamu-A*01 (1ZVS) 184.62 185.29 
Mamu-B*17 (3RWC) 209.42 224.69 
H2*kd (1VGK) 165.7 167.59 
Murine H2*kd (3TID) 155.69 159.1 
SLA*0401 (3QQ3) 183.94 207.92 
BoLAN*01801 (2XFX) 195.38 206.24 
BoLAN*01801 (3PWU) 191.46 216.9 
BF2*0401 (4E0R) 166.64 197.08 
Grass carp (5Y91) 218.19 238.38 

To confirm the binding peptide profile and the restrictive pockets of pXela-UAAg, Xela-UAAg and Xela-β2m were refolded with Ran_9Xsplitted instead of the specific purified nonapeptide. As shown in Fig. 5A, the complexes of pXela-UAAg molecules were collected, and the peptides that bound to Xela-UAAg for MS identification, as mentioned above, are shown in Supplemental Table II. The result of pXela-UAAg binding a peptide profile (score >90) determined by Ran_9Xsplitted is shown in Fig. 5B. The primary anchor residues of the YMM9 peptide were investigated by alanine mutation and circular dichroism spectroscopy experiments. The results showed that the peaks of the pXela-UAAg complex with P1-A, P2-A, and P9-A peptides were significantly decreased (Fig. 5C). The Tm value of the wild-type peptide was found to be 53.4°C. Among all of the alanine mutant peptides, the Tm values of the P1-A, P2-A, and P9-A mutant peptides were significantly reduced (48, 42.6, and 45.4°C) (Fig. 5D), indicating that the P1, P2, and P9 residues are the primary anchor residues and play key roles in peptide binding. Mutation of the remaining anchor residues (P3–P8) did not induce significant changes with respect to the wild-type peptide. These circular dichroism results were consistent with the mutation experiment described above (Fig. 5C). Therefore, the pockets (A, B, and F) at the two termini of the PBG of pXela-UAAg are anchor sites, and the basic peptide-binding motif is Y/A-M/A/L/Y-XXXXXX-I/L for pXela-UAAg. In addition, according to our reported method, whole-genome proteins of iridovirus FV3 were scanned, and 682 9-merpeptides were obtained from 87 proteins with a high binding ability to pXela-UAAg (Supplemental Table III).

FIGURE 5.

Investigation of peptide profile and motif of pXela-UAAg. (A) Refolding efficiency of pXela-UAAg. Xela-β2m and Xela-UAAg were co-refolded with Ran_9Xsplitted in vitro. The pXela-UAAg curve is shown in blue. The insets show reducing SDS-PAGE gels (15%) of the peaks that are labeled on the curve. Lane M contains molecular mass markers. A molecular sieve was used to capture the peaks, confirming that peaks1, 2, and 3 are Xela-β2m, complex of Xela-UAAg-β2m, and aggregated Xela-UAAg, respectively. (B) Sequence logo showing the amino acid weighting probabilities at every position of the presented peptide in Shannon representation. (C) The refolded products of pXela-UAAg and Xela-β2m in the presence of mutated peptides tested by gel filtration chromatograms. The refolding efficiencies are represented by the relevant concentration ratios and by the heights of the pXela-UAAg peak for each mutant. Xela-UAAg with wild-type peptide YMM9 was the positive control, and Xela-UAAg without a peptide was included as a negative control. The mutated peptides P1A, P2A, and P9A clearly yielded the lowest refolding efficiency. (D) Thermal stabilities of the pXela-UAAg complex. pXela-UAAg bound to the YMM9 peptide or one of nine mutational peptides was tested by circular dichroism spectroscopy. The denaturation curves of the complexes with the different peptides are indicated in different colors. The results showed that the stability of Xela-UAAg with P1A, P2A, and P9A was significantly decreased, with values more than 5°C lower than that of YMM9.

FIGURE 5.

Investigation of peptide profile and motif of pXela-UAAg. (A) Refolding efficiency of pXela-UAAg. Xela-β2m and Xela-UAAg were co-refolded with Ran_9Xsplitted in vitro. The pXela-UAAg curve is shown in blue. The insets show reducing SDS-PAGE gels (15%) of the peaks that are labeled on the curve. Lane M contains molecular mass markers. A molecular sieve was used to capture the peaks, confirming that peaks1, 2, and 3 are Xela-β2m, complex of Xela-UAAg-β2m, and aggregated Xela-UAAg, respectively. (B) Sequence logo showing the amino acid weighting probabilities at every position of the presented peptide in Shannon representation. (C) The refolded products of pXela-UAAg and Xela-β2m in the presence of mutated peptides tested by gel filtration chromatograms. The refolding efficiencies are represented by the relevant concentration ratios and by the heights of the pXela-UAAg peak for each mutant. Xela-UAAg with wild-type peptide YMM9 was the positive control, and Xela-UAAg without a peptide was included as a negative control. The mutated peptides P1A, P2A, and P9A clearly yielded the lowest refolding efficiency. (D) Thermal stabilities of the pXela-UAAg complex. pXela-UAAg bound to the YMM9 peptide or one of nine mutational peptides was tested by circular dichroism spectroscopy. The denaturation curves of the complexes with the different peptides are indicated in different colors. The results showed that the stability of Xela-UAAg with P1A, P2A, and P9A was significantly decreased, with values more than 5°C lower than that of YMM9.

Close modal

The PBG loops of pXela-UAAg shared a unique shape in nonmammals (Fig. 6). First, the AB loop orientation in nonmammals and mammals is different (Fig. 6A). The AB loop of pXela-UAAg was downward and combined with Xela-β2m (Fig. 6A–C). In the solved pMHC-I structures of nonmammals, such as chicken (Fig. 6D) and grass carp (Fig. 6E), the AB loops combined with β2m. However, the AB loops in the β sheet of the α1 domain cannot combine with β2m in mammals (3) (Fig. 6F). The superposition of mammals and nonmammals (Fig. 6G) shows that the AB loop binds β2m via hydrogen bonds in nonmammalian pMHC-I structures, whereas the AB loop of mammalian pMHC-I does not contact β2m owing to the lack of such interaction. Based on existing three-dimensional data, this point is key to rationally distinguishing nonmammals from mammalian pMHC-I structures.

FIGURE 6.

pXela-UAAg shows some features similar to those in nonmammals (chicken, duck, and grass carp) but different from those in mammals. (A) Unique details of PBG in lower vertebrate pMHC-I (pXela-UAAg: green; grass carp: 5Y91, light orange; chicken: 4E0R, hot pink; duck: 5GJX, magenta; and other mammals: [3OXS, 1ZVS, 3RWC, 3TID, 1VGK, 3QQ3, 3PWU, 2XFX], white). The AB loop of X. laevis is shown with electron density. (B) Sequence characteristics of the lower vertebrates. (CF) The gray areas indicate the discrepant regions of HC and β2m interfaces among X. laevis, grass carp, chicken, and mammals. (C) In grass carp (5Y91), the AB loop can bind to β2m. (D) In chicken BF2*0401 (4E0R), the AB loop can bind to β2m. (E) In pXela-UAAg (6A2B), the AB loop of HC can bind to β2m. (F) In mammalian p/MHC-I structures (3OXS), the AB loop cannot bind to β2m. (G) The AB loop of nonmammalian pMHC-I can interact with β2m via hydrogen bonds, and mammals do not exhibit such interactions.

FIGURE 6.

pXela-UAAg shows some features similar to those in nonmammals (chicken, duck, and grass carp) but different from those in mammals. (A) Unique details of PBG in lower vertebrate pMHC-I (pXela-UAAg: green; grass carp: 5Y91, light orange; chicken: 4E0R, hot pink; duck: 5GJX, magenta; and other mammals: [3OXS, 1ZVS, 3RWC, 3TID, 1VGK, 3QQ3, 3PWU, 2XFX], white). The AB loop of X. laevis is shown with electron density. (B) Sequence characteristics of the lower vertebrates. (CF) The gray areas indicate the discrepant regions of HC and β2m interfaces among X. laevis, grass carp, chicken, and mammals. (C) In grass carp (5Y91), the AB loop can bind to β2m. (D) In chicken BF2*0401 (4E0R), the AB loop can bind to β2m. (E) In pXela-UAAg (6A2B), the AB loop of HC can bind to β2m. (F) In mammalian p/MHC-I structures (3OXS), the AB loop cannot bind to β2m. (G) The AB loop of nonmammalian pMHC-I can interact with β2m via hydrogen bonds, and mammals do not exhibit such interactions.

Close modal

In addition, there are two fewer residues after the 40th amino acid of the PBGs in the frog, grass carp, duck, and chicken than in mammalian MHC-I HC (Fig. 2, green box with green triangle). Therefore, the CD loops of nonmammals are shorter than those of mammals (Fig. 6A, 6B). Moreover, there is one more residue after the 114th amino acid of the PBGs in mammalian structures than in nonmammalian structures (3, 36, 43) (frog, duck, chicken, and grass carp), resulting in the longer EF loop in mammals than in nonmammals (Fig. 6A). There is a hydrophobic core in the GH loops of X. laevis, duck, chicken, and grass carp; therefore, the GH loop is expanded and close to the β2ms. Based on the above data, these features are unique for distinguishing nonmammalian from mammalian pMHC-I structures.

The shift in the loop between the C and D strands (namely, the CD loop) in the α3 domain of MHC-I plays a critical role in binding to the CD8 coreceptor (3). The key amino acid in the CD loop shift of the α3 domain in pXela-UAAg is H224, which is conserved in almost all X. laevis (Fig. 2). In many other species, this residue is Q224, whereas in grass carp it is E219. The superposition of pXela-UAAg with the HLA-A*0201 (PDB: 1AKJ), chicken, and grass carp pMHC-I structures indicated that the distances at the CD loop shift between nonmammals and humans are very significant (Fig. 7). We found that the distance between E219 of pCtid-UAAg and Q226 of HLA-A*0201 (PDB: 1AKJ) is ∼12.4 Å, the distance between H224 of pXela-UAAg (PDB: https://www.rcsb.org/structure/6A2B) and Q226 of HLA-A*0201 (PDB: 1AKJ) is ∼10.7 Å, and the distance between Q222 of BF2*0401 (PDB: 4E0R) and Q226 of HLA-A*0201 (PDB: 1AKJ) is ∼8.4 Å. These results indicated the evolution from lower to higher vertebrates and the decreasing distance between pMHC-I and CD8 (Fig. 7). This finding may reveal a unique way to bind CD8 in nonmammals.

FIGURE 7.

The distance between CD loop shifts reveals a binding mode of CD8 in nonmammals. The major shift in the α3 domain and variation in the key residues for binding CD8αα in pXela-UAAg. The CD loop of X. laevis is shown with electron density. The model of X. laevis CD8αα (green part, surface representation) was established using the online tool Swiss model and superposed on the human HLA-A2 and CD8αα complex (white, PDB code: 1AKJ), Ctid-UAAg (cyan, PDB code: 5Y91), and BF2*0401 (hot pink, PDB code: 4E0R) structures. The distance between E219 of pCtid-UAAg and Q226 of HLA-A*0201 (PDB: 1AKJ) is ∼12.4 Å, the distance between superposed CD loops of pXela-UAAg and HLA-A*0201 is ∼10.7 Å, and the distance between Q222 in BF2*0401 (PDB: 4E0R) and Q226 in HLA-A*0201 (PDB: 1AKJ) is ∼8.4 Å. The residues in the CD loop that are critical for interaction with CD8αα are shown in stick representation.

FIGURE 7.

The distance between CD loop shifts reveals a binding mode of CD8 in nonmammals. The major shift in the α3 domain and variation in the key residues for binding CD8αα in pXela-UAAg. The CD loop of X. laevis is shown with electron density. The model of X. laevis CD8αα (green part, surface representation) was established using the online tool Swiss model and superposed on the human HLA-A2 and CD8αα complex (white, PDB code: 1AKJ), Ctid-UAAg (cyan, PDB code: 5Y91), and BF2*0401 (hot pink, PDB code: 4E0R) structures. The distance between E219 of pCtid-UAAg and Q226 of HLA-A*0201 (PDB: 1AKJ) is ∼12.4 Å, the distance between superposed CD loops of pXela-UAAg and HLA-A*0201 is ∼10.7 Å, and the distance between Q222 in BF2*0401 (PDB: 4E0R) and Q226 in HLA-A*0201 (PDB: 1AKJ) is ∼8.4 Å. The residues in the CD loop that are critical for interaction with CD8αα are shown in stick representation.

Close modal

Sequence alignment showed that the 148–149 positions of the Xela-UAAg and other frogs classical MHC-I sequences all contain two amino acid insertions (E148 and V149) (Fig. 2). The insertion motif creates an additional bulge at the highest-positioned structure of the α2-helical region (Fig. 8A–D), with a kink between helices H1 and H2, and is expected to affect TCR binding. The superposition of the known PBGs in pMHC-I complexes revealed that the insertion of the two amino acids does not result in a large change in the PBGs (Fig. 8C) but does change the kink in the α2 helices in Xela-UAAg, which is significantly the highest among those in the solved pMHC-I PBGs (Fig. 8A, 8B). In brief, the C atom of the main chain at the highest point of the PBG in pXela-UAAg is ∼3.8 Å higher than that in chicken, 4.5 Å higher than that in grass carp, and ∼4–7 Å higher than those in other mammalian pMHC-I complexes. The red regions (Fig. 8E) are the key regions for the interaction between pMHC-I and TCR (46). It has been previously reported that pMHC-I (PBD: 5F1I) in dogs has an insertion of L at position 155 (this site is a key site for MHC to contact TCR) (47). The two inserted amino acids are not located in the key regions (red parts) that contact TCR but do change the highest region of pMHC-I, which is also an important region in contacting TCR. In the superposition of the pXela-MHC-I and HLA-A2–TCR complex, the E148-N150 protrusion is inserted into TCR (Fig. 8E, 8F). This difference may affect the PBG of pXela-UAAg when docking TCRs, reflecting a unique TCR recognition mode for pMHC-I in frogs.

FIGURE 8.

The unique bulged α2 helices of pXela-UAAg. (A) Superposition of α1 and α2 domains of pXela-UAAg (6A2B, green) with another 28 representative p/MHC-I (1VGK, 1T1Y, 2XFX, 2CLZ, 3BO8, 3OXS, 3PWU, 3TID, 5Y91, 1ZVS, 3QQ3, 3RWC, 4E0R, 3X11, 4QRQ, 4MJ6, 4MNX, 4N02, 4NT6, 4O2C, 4QOK, 4QRS, 4CVX, 5XMF, 5GJX, 3BUY, 1ZT7, and 4WJ5 white) shown in surface representation, the region of the Xela-UAAg α2 bulge (P147–N150, yellow) circled by red dotted line ellipse. (B) View from the α2 domain on the surface. (C) Superposition of α1 and α2 domains of pXela-UAAg with representative p/MHC-I shown in cartoon; P147–N150 of pXela-UAAg (yellow) circled by red dotted line ellipse. (D) View from the α2 domain in cartoon. The main chain C atom at the highest region of the pXela-UAAg α2 bulge is ∼3.8 Ǻ higher than that of other representative p/MHC-I molecules. (E) Based on the HLA-A2–TCR complex structure (PDB code: 1BD2), special residues of pXela-UAAg that may contact TCR CDR loops (cartoon, cyan) are shown as yellow and circled by red dotted line ellipse. The regions that may contact TCR in the reported study are colored green in pXela-UAAg. YMM9 was shown on the surface according to its charge. (F) Superposition of pXela-UAAg with HLA-A2–TCR. pXela-UAAg (green) and HLA-A2–TCR complex (cyan) are shown on the surface. The α2 bulge (yellow) is inserted into TCR, revealing a new way for pMHC-I to recognize TCR.

FIGURE 8.

The unique bulged α2 helices of pXela-UAAg. (A) Superposition of α1 and α2 domains of pXela-UAAg (6A2B, green) with another 28 representative p/MHC-I (1VGK, 1T1Y, 2XFX, 2CLZ, 3BO8, 3OXS, 3PWU, 3TID, 5Y91, 1ZVS, 3QQ3, 3RWC, 4E0R, 3X11, 4QRQ, 4MJ6, 4MNX, 4N02, 4NT6, 4O2C, 4QOK, 4QRS, 4CVX, 5XMF, 5GJX, 3BUY, 1ZT7, and 4WJ5 white) shown in surface representation, the region of the Xela-UAAg α2 bulge (P147–N150, yellow) circled by red dotted line ellipse. (B) View from the α2 domain on the surface. (C) Superposition of α1 and α2 domains of pXela-UAAg with representative p/MHC-I shown in cartoon; P147–N150 of pXela-UAAg (yellow) circled by red dotted line ellipse. (D) View from the α2 domain in cartoon. The main chain C atom at the highest region of the pXela-UAAg α2 bulge is ∼3.8 Ǻ higher than that of other representative p/MHC-I molecules. (E) Based on the HLA-A2–TCR complex structure (PDB code: 1BD2), special residues of pXela-UAAg that may contact TCR CDR loops (cartoon, cyan) are shown as yellow and circled by red dotted line ellipse. The regions that may contact TCR in the reported study are colored green in pXela-UAAg. YMM9 was shown on the surface according to its charge. (F) Superposition of pXela-UAAg with HLA-A2–TCR. pXela-UAAg (green) and HLA-A2–TCR complex (cyan) are shown on the surface. The α2 bulge (yellow) is inserted into TCR, revealing a new way for pMHC-I to recognize TCR.

Close modal

Amphibians are a large group of primitive, variable-temperature tetrapods that were among the first land animals. Genome-wide sequencing and various studies have demonstrated the sequences and functions of immune-related genes (9, 48). However, there are no studies on T cell immunity at the structural level in amphibians. In this study, X. laevis was used as a representative amphibian species to elucidate for the first time, to our knowledge, the crystal structure of pXela-UAAg and its Ag-presenting peptide profile.

The length of the amino acid sequence of Xela-UAAg is markedly different from that of MHC-I sequences in mammals (49). The comparison of the classical MHC-I amino acid sequences between X. laevis and mammals, chickens, and fish revealed <45% sequence identity (17, 50). Thus, the differences in amino acid sequences and lengths between species are very significant. However, the intraspecific homology of X. laevis classical MHC-I molecules is >80%, and the variation is mainly concentrated in the α1 and α2 domains. Moreover, the lengths of the amino acid sequences in the α1 and α2 domains that constitute the PBGs are also different from those of mammals (3). There are 88 and 93 aa in the α1 and α2 domains of X. laevis classical MHC-I, respectively. The α1 domain in X. laevis has two fewer amino acids than that in mammals, and the α2 domain in X. laevis has two more amino acids than that in mammals (Fig. 2). Importantly, the differences in the lengths and sequences are vividly reflected in the three-dimensional structure of pXela-UAAg.

Notably, the topology of pXela-UAAg is similar to that of mammals, birds, and bony fish, and the function of the PBG and the presenting peptides are similar; in particular, the peptide is a typical M-type peptide that can activate T cells (33). Hashimoto (45) defines seven “critical” residues, Y7, Y59, T143, K146, W147, Y159, and Y171, in the A, E, and F pockets that bind to the ends of the polypeptide. The Y7, Y59, T143, K146, W147, Y159, and Y171 amino acids are conserved in humans, mammals, and birds and likewise conserved in the PBG of pXela-UAAg (Fig. 1D). Importantly, pXela-UAAg also has distinct characteristics in X. laevis as a species. The A pocket is the largest, and it can accommodate amino acids with large side chains; the C–F pockets are positively charged. In particular, two amino acids (EV) are inserted at the 148 and 149 positions; as a result, the region of this structure extends ∼3.8–6.5 Å higher than those of solved pMHC-I structures. Additionally, X. laevis classical MHC-I sequences are known to have an insertion of two amino acids; therefore, we speculate that X. laevis classical MHC-I docks CTLs differently from the MHC-I complexes of other jawed vertebrates.

In addition, because the AB loop and CD loop in the pXela-UAAg α1 domain and the EF loop in the α2 domain have their own features, the AB loop in pXela-UAAg can directly combine with Xela-β2m; the EF loop and CD loop are shorter than those in mammals, causing the CD loop to contain two fewer amino acids and the EF loop to contain one fewer amino acid in nonmammalian structures than in mammalian structures. However, pXela-UAAg is similar to known nonmammalian structures (chickens, ducks, and grass carp) (3, 5, 36, 43). Moreover, four amino acids in Xela-β2m that interact with Xela-UAAg are conserved in mammals and other vertebrates. The key amino acid that interacts with TCR coreceptor CD8 molecules is H224 on the CD loop in Xela-UAAg, and the distance from H224 to human HLA-A2 Q226 is 10.7 Å (51). The distance from chicken to human HLA-A2 Q226 is 8.4 Å (36), and the distance from grass carp to human HLA-A2 Q226 is 12.4 Å (3). Because of the distance gap, it can be inferred that nonmammalian CD loops may interact differently with CD8 molecules than mammalian CD loops, and each genus if not each species has its own unique interaction.

The pXela-UAAg–binding peptide motif was determined by alanine mutation of a nonapeptide in the experiment. The PBG-binding peptide profile was also determined by the random nonapeptide library (Supplemental Table II). We rationally believe that the peptide profile can cover many antigenic epitopes of viruses and other organisms (31). Based on the profile, we scanned the frog iridescent virus 3, and 682 nonapeptides that may have high affinity to pXela-UAAg were predicted (Supplemental Table III). However, because only adult frogs express classical MHC-I molecules, they can present additional FV3 peptides, whereas tadpoles do not express classical MHC-I molecules. Therefore, we speculate that only adult frogs have T cell immunity and are resistant to FV3 (23).

Although pXela-UAAg is only one three-dimensional structure of X. laevis classical MHC-I alleles, almost all frog pMHC-I complexes can be modeled based on this structure because MHC-I homology among frogs is >55% (52). In other words, a frog epitope vaccine can be preliminarily designed by the Xela-UAAg platform, and this platform can lead to the further development of vaccines for the prevention and treatment of viral and oncological diseases (23, 53). In conclusion, pXela-UAAg revealed not only the species characteristics of amphibian MHC-I sequences but also its differences from the structures of birds, mammals, and fish, providing, to our knowledge, new information about T cell immunity in the animal kingdom.

We acknowledge the assistance of the staff of the Shanghai Synchrotron Radiation Facility of China.

This work was supported by the National Natural Science Foundation of China (Grants 319 726 83 and 31572493).

The pXela-UAAg structure presented in this article has been submitted to the Protein Data Bank (https://www.rcsb.org/structure/6A2B) under accession number 6A2B.

The online version of this article contains supplemental material.

Abbreviations used in this article:

ASA

accessible surface area

FV3

frog virus 3

HC

H chain

β2m

β2-microglobulin

MHC-I

MHC class I

MS

mass spectrometry

PBG

peptide-binding groove

PDB

Protein Data Bank

pMHC-I

peptide–MHC class I complex

RMSD

root mean square deviation

TFA

trifluoroacetic acid

Tm

midpoint transition temperature.

1
Flajnik
,
M. F.
2018
.
A cold-blooded view of adaptive immunity.
Nat. Rev. Immunol.
18
:
438
453
.
2
Bjorkman
,
P. J.
,
M. A.
Saper
,
B.
Samraoui
,
W. S.
Bennett
,
J. L.
Strominger
,
D. C.
Wiley
.
1987
.
The foreign antigen binding site and T cell recognition regions of class I histocompatibility antigens.
Nature
329
:
512
518
.
3
Chen
,
Z.
,
N.
Zhang
,
J.
Qi
,
R.
Chen
,
J. M.
Dijkstra
,
X.
Li
,
Z.
Wang
,
J.
Wang
,
Y.
Wu
,
C.
Xia
.
2017
.
The structure of the MHC class I molecule of bony fishes provides insights into the conserved nature of the antigen-presenting system.
J. Immunol.
199
:
3668
3678
.
4
Trolle
,
T.
,
C. P.
McMurtrey
,
J.
Sidney
,
W.
Bardet
,
S. C.
Osborn
,
T.
Kaever
,
A.
Sette
,
W. H.
Hildebrand
,
M.
Nielsen
,
B.
Peters
.
2016
.
The length distribution of class I-restricted T cell epitopes is determined by both peptide supply and MHC allele-specific binding preference.
J. Immunol.
196
:
1480
1487
.
5
Koch
,
M.
,
S.
Camp
,
T.
Collen
,
D.
Avila
,
J.
Salomonsen
,
H. J.
Wallny
,
A.
van Hateren
,
L.
Hunt
,
J. P.
Jacob
,
F.
Johnston
, et al
.
2007
.
Structures of an MHC class I molecule from B21 chickens illustrate promiscuous peptide binding.
Immunity
27
:
885
899
.
6
Saper
,
M. A.
,
P. J.
Bjorkman
,
D. C.
Wiley
.
1991
.
Refined structure of the human histocompatibility antigen HLA-A2 at 2.6 A resolution.
J. Mol. Biol.
219
:
277
319
.
7
Borghans
,
J. A.
,
J. B.
Beltman
,
R. J.
De Boer
.
2004
.
MHC polymorphism under host-pathogen coevolution.
Immunogenetics
55
:
732
739
.
8
Borbulevych
,
O. Y.
,
K. H.
Piepenbrink
,
B. E.
Gloor
,
D. R.
Scott
,
R. F.
Sommese
,
D. K.
Cole
,
A. K.
Sewell
,
B. M.
Baker
.
2009
.
T cell receptor cross-reactivity directed by antigen-dependent tuning of peptide-MHC molecular flexibility.
Immunity
31
:
885
896
.
9
Session
,
A. M.
,
Y.
Uno
,
T.
Kwon
,
J. A.
Chapman
,
A.
Toyoda
,
S.
Takahashi
,
A.
Fukui
,
A.
Hikosaka
,
A.
Suzuki
,
M.
Kondo
, et al
.
2016
.
Genome evolution in the allotetraploid frog Xenopus laevis.
Nature
538
:
336
343
.
10
Edholm
,
E. S.
,
M.
Banach
,
K.
Hyoe Rhoo
,
M. S.
Pavelka
Jr.
,
J.
Robert
.
2018
.
Distinct MHC class I-like interacting invariant T cell lineage at the forefront of mycobacterial immunity uncovered in Xenopus.
Proc. Natl. Acad. Sci. USA
115
:
E4023
E4031
.
11
Flajnik
,
M. F.
,
C.
Canel
,
J.
Kramer
,
M.
Kasahara
.
1991
.
Evolution of the major histocompatibility complex: molecular cloning of major histocompatibility complex class I from the amphibian Xenopus.
Proc. Natl. Acad. Sci. USA
88
:
537
541
.
12
Salter-Cid
,
L.
,
M.
Nonaka
,
M. F.
Flajnik
.
1998
.
Expression of MHC class Ia and class Ib during ontogeny: high expression in epithelia and coregulation of class Ia and lmp7 genes.
J. Immunol.
160
:
2853
2861
.
13
Flajnik
,
M. F.
,
J. F.
Kaufman
,
E.
Hsu
,
M.
Manes
,
R.
Parisot
,
L.
Du Pasquier
.
1986
.
Major histocompatibility complex-encoded class I molecules are absent in immunologically competent Xenopus before metamorphosis.
J. Immunol.
137
:
3891
3899
.
14
Flajnik
,
M. F.
,
Y.
Ohta
,
A. S.
Greenberg
,
L.
Salter-Cid
,
A.
Carrizosa
,
L.
Du Pasquier
,
M.
Kasahara
.
1999
.
Two ancient allelic lineages at the single classical class I locus in the Xenopus MHC.
J. Immunol.
163
:
3826
3833
.
15
Shum
,
B. P.
,
D.
Avila
,
L.
Du Pasquier
,
M.
Kasahara
,
M. F.
Flajnik
.
1993
.
Isolation of a classical MHC class I cDNA from an amphibian. Evidence for only one class I locus in the Xenopus MHC.
J. Immunol.
151
:
5376
5386
.
16
Du Pasquier
,
L.
,
B.
Blomberg
.
1982
.
The expression of antibody diversity in natural and laboratory-made polyploid individuals of the clawed toad Xenopus.
Immunogenetics
15
:
251
260
.
17
Dijkstra
,
J. M.
,
T.
Yamaguchi
,
U.
Grimholt
.
2018
.
Conservation of sequence motifs suggests that the nonclassical MHC class I lineages CD1/PROCR and UT were established before the emergence of tetrapod species.
Immunogenetics
70
:
459
476
.
18
Goyos
,
A.
,
Y.
Ohta
,
S.
Guselnikov
,
J.
Robert
.
2009
.
Novel nonclassical MHC class Ib genes associated with CD8 T cell development and thymic tumors.
Mol. Immunol.
46
:
1775
1786
.
19
Robert
,
J.
,
Y.
Ohta
.
2009
.
Comparative and developmental study of the immune system in Xenopus.
Dev. Dyn.
238
:
1249
1270
.
20
Bernard
,
D.
,
J. D.
Hansen
,
L.
Du Pasquier
,
M. P.
Lefranc
,
A.
Benmansour
,
P.
Boudinot
.
2007
.
Costimulatory receptors in jawed vertebrates: conserved CD28, odd CTLA4 and multiple BTLAs.
Dev. Comp. Immunol.
31
:
255
271
.
21
Ohta
,
Y.
,
W.
Goetz
,
M. Z.
Hossain
,
M.
Nonaka
,
M. F.
Flajnik
.
2006
.
Ancestral organization of the MHC revealed in the amphibian Xenopus.
J. Immunol.
176
:
3674
3685
.
22
Göbel
,
T. W.
,
E. L.
Meier
,
L.
Du Pasquier
.
2000
.
Biochemical analysis of the Xenopus laevis TCR/CD3 complex supports the “stepwise evolution” model.
Eur. J. Immunol.
30
:
2775
2781
.
23
Morales
,
H. D.
,
J.
Robert
.
2007
.
Characterization of primary and memory CD8 T-cell responses against ranavirus (FV3) in Xenopus laevis.
J. Virol.
81
:
2240
2248
.
24
Goyos
,
A.
,
J.
Robert
.
2009
.
Tumorigenesis and anti-tumor immune responses in Xenopus.
Front. Biosci.
14
:
167
176
.
25
Gantress
,
J.
,
G. D.
Maniero
,
N.
Cohen
,
J.
Robert
.
2003
.
Development and characterization of a model system to study amphibian immune responses to iridoviruses.
Virology
311
:
254
262
.
26
Chinchar
,
V. G.
,
A.
Hyatt
,
T.
Miyazaki
,
T.
Williams
.
2009
.
Family Iridoviridae: poor viral relations no longer.
Curr. Top. Microbiol. Immunol.
328
:
123
170
.
27
Nedelkovska
,
H.
,
E. S.
Edholm
,
N.
Haynes
,
J.
Robert
.
2013
.
Effective RNAi-mediated β2-microglobulin loss of function by transgenesis in Xenopus laevis.
Biol. Open
2
:
335
342
.
28
Flajnik
,
M. F.
2002
.
Comparative analyses of immunoglobulin genes: surprises and portents.
Nat. Rev. Immunol.
2
:
688
698
.
29
Flajnik
,
M. F.
,
M.
Kasahara
.
2001
.
Comparative genomics of the MHC: glimpses into the evolution of the adaptive immune system.
Immunity
15
:
351
362
.
30
Jurtz
,
V.
,
S.
Paul
,
M.
Andreatta
,
P.
Marcatili
,
B.
Peters
,
M.
Nielsen
.
2017
.
NetMHCpan-4.0: improved peptide-MHC class I interaction predictions integrating eluted ligand and peptide binding affinity data.
J. Immunol.
199
:
3360
3368
.
31
Qu
,
Z.
,
Z.
Li
,
L.
Ma
,
X.
Wei
,
L.
Zhang
,
R.
Liang
,
G.
Meng
,
N.
Zhang
,
C.
Xia
.
2019
.
Structure and peptidome of the bat MHC class I molecule reveal a novel mechanism leading to high-affinity peptide binding.
J. Immunol.
202
:
3493
3506
.
32
Garboczi
,
D. N.
,
D. T.
Hung
,
D. C.
Wiley
.
1992
.
HLA-A2-peptide complexes: refolding and crystallization of molecules expressed in Escherichia coli and complexed with single antigenic peptides.
Proc. Natl. Acad. Sci. USA
89
:
3429
3433
.
33
Yao
,
S.
,
J.
Liu
,
J.
Qi
,
R.
Chen
,
N.
Zhang
,
Y.
Liu
,
J.
Wang
,
Y.
Wu
,
G. F.
Gao
,
C.
Xia
.
2016
.
Structural illumination of equine MHC class I molecules highlights unconventional epitope presentation manner that is evolved in equine leukocyte antigen alleles.
J. Immunol.
196
:
1943
1954
.
34
Chi
,
H.
,
H.
Chen
,
K.
He
,
L.
Wu
,
B.
Yang
,
R. X.
Sun
,
J.
Liu
,
W. F.
Zeng
,
C. Q.
Song
,
S. M.
He
,
M. Q.
Dong
.
2013
.
pNovo+: de novo peptide sequencing using complementary HCD and ETD tandem mass spectra.
J. Proteome Res.
12
:
615
625
.
35
Muth
,
T.
,
L.
Weilnböck
,
E.
Rapp
,
C. G.
Huber
,
L.
Martens
,
M.
Vaudel
,
H.
Barsnes
.
2014
.
DeNovoGUI: an open source graphical user interface for de novo sequencing of tandem mass spectra.
J. Proteome Res.
13
:
1143
1146
.
36
Zhang
,
J.
,
Y.
Chen
,
J.
Qi
,
F.
Gao
,
Y.
Liu
,
J.
Liu
,
X.
Zhou
,
J.
Kaufman
,
C.
Xia
,
G. F.
Gao
.
2012
.
Narrow groove and restricted anchors of MHC class I molecule BF2*0401 plus peptide transporter restriction can explain disease susceptibility of B4 chickens.
J. Immunol.
189
:
4478
4487
.
37
Tobita
,
T.
,
M.
Oda
,
H.
Morii
,
M.
Kuroda
,
A.
Yoshino
,
T.
Azuma
,
H.
Kozono
.
2003
.
A role for the P1 anchor residue in the thermal stability of MHC class II molecule I-Ab.
Immunol. Lett.
85
:
47
52
.
38
Jones
,
D. T.
1999
.
Protein secondary structure prediction based on position-specific scoring matrices.
J. Mol. Biol.
292
:
195
202
.
39
Stormo
,
G. D.
,
T. D.
Schneider
,
L.
Gold
,
A.
Ehrenfeucht
.
1982
.
Use of the ‘Perceptron’ algorithm to distinguish translational initiation sites in E. coli.
Nucleic Acids Res.
10
:
2997
3011
.
40
Li
,
X.
,
J.
Liu
,
J.
Qi
,
F.
Gao
,
Q.
Li
,
X.
Li
,
N.
Zhang
,
C.
Xia
,
G. F.
Gao
.
2011
.
Two distinct conformations of a rinderpest virus epitope presented by bovine major histocompatibility complex class I N*01801: a host strategy to present featured peptides.
J. Virol.
85
:
6038
6048
.
41
Adams
,
P. D.
,
R. W.
Grosse-Kunstleve
,
L. W.
Hung
,
T. R.
Ioerger
,
A. J.
McCoy
,
N. W.
Moriarty
,
R. J.
Read
,
J. C.
Sacchettini
,
N. K.
Sauter
,
T. C.
Terwilliger
.
2002
.
PHENIX: building new software for automated crystallographic structure determination.
Acta Crystallogr. D Biol. Crystallogr.
58
:
1948
1954
.
42
Liang
,
R.
,
Y.
Sun
,
Y.
Liu
,
J.
Wang
,
Y.
Wu
,
Z.
Li
,
L.
Ma
,
N.
Zhang
,
L.
Zhang
,
X.
Wei
, et al
.
2018
.
Major histocompatibility complex class I (FLA-E*01801) molecular structure in domestic cats demonstrates species-specific characteristics in presenting viral antigen peptides.
J. Virol.
92
: e01631-17.
43
Wu
,
Y.
,
J.
Wang
,
S.
Fan
,
R.
Chen
,
Y.
Liu
,
J.
Zhang
,
H.
Yuan
,
R.
Liang
,
N.
Zhang
,
C.
Xia
.
2017
.
Structural definition of duck major histocompatibility complex class I molecules that might explain efficient cytotoxic T lymphocyte immunity to influenza a virus.
J. Virol.
91
: e02511-16.
44
Fan
,
S.
,
Y.
Wu
,
S.
Wang
,
Z.
Wang
,
B.
Jiang
,
Y.
Liu
,
R.
Liang
,
W.
Zhou
,
N.
Zhang
,
C.
Xia
.
2016
.
Structural and biochemical analyses of swine major histocompatibility complex class I complexes and prediction of the epitope map of important influenza a virus strains.
J. Virol.
90
:
6625
6641
.
45
Hashimoto
,
K.
,
K.
Okamura
,
H.
Yamaguchi
,
M.
Ototake
,
T.
Nakanishi
,
Y.
Kurosawa
.
1999
.
Conservation and diversification of MHC class I and its related molecules in vertebrates.
Immunol. Rev.
167
:
81
100
.
46
Marrack
,
P.
,
J. P.
Scott-Browne
,
S.
Dai
,
L.
Gapin
,
J. W.
Kappler
.
2008
.
Evolutionarily conserved amino acids that control TCR-MHC interaction.
Annu. Rev. Immunol.
26
:
171
203
.
47
Xiao
,
J.
,
W.
Xiang
,
Y.
Chai
,
J.
Haywood
,
J.
Qi
,
L.
Ba
,
P.
Qi
,
M.
Wang
,
J.
Liu
,
G. F.
Gao
.
2016
.
Diversified anchoring features the peptide presentation of DLA-88*50801: first structural insight into domestic dog MHC class I.
J. Immunol.
197
:
2306
2315
.
48
Hellsten
,
U.
,
R. M.
Harland
,
M. J.
Gilchrist
,
D.
Hendrix
,
J.
Jurka
,
V.
Kapitonov
,
I.
Ovcharenko
,
N. H.
Putnam
,
S.
Shu
,
L.
Taher
, et al
.
2010
.
The genome of the Western clawed frog Xenopus tropicalis.
Science
328
:
633
636
.
49
Bos
,
D. H.
,
B.
Waldman
.
2006
.
Polymorphism, natural selection, and structural modeling of class Ia MHC in the African clawed frog (Xenopus laevis).
Immunogenetics
58
:
433
442
.
50
Dijkstra
,
J. M.
,
T.
Yamaguchi
.
2019
.
Ancient features of the MHC class II presentation pathway, and a model for the possible origin of MHC molecules.
Immunogenetics
71
:
233
249
.
51
Gao
,
G. F.
,
J.
Tormo
,
U. C.
Gerth
,
J. R.
Wyer
,
A. J.
McMichael
,
D. I.
Stuart
,
J. I.
Bell
,
E. Y.
Jones
,
B. K.
Jakobsen
.
1997
.
Crystal structure of the complex between human CD8alpha(alpha) and HLA-A2.
Nature
387
:
630
634
.
52
Lillie
,
M.
,
R.
Shine
,
K.
Belov
.
2014
.
Characterisation of major histocompatibility complex class I in the Australian cane toad, Rhinella marina.
PLoS One
9
: e102824.
53
Mirza
,
M. U.
,
S.
Rafique
,
A.
Ali
,
M.
Munir
,
N.
Ikram
,
A.
Manan
,
O. M. H.
Salo-Ahen
,
M.
Idrees
.
2016
.
Towards peptide vaccines against Zika virus: immunoinformatics combined with molecular dynamics simulations to predict antigenic epitopes of Zika viral proteins. [Published erratum appears in 2017 Sci. Rep. 7: 44633.]
Sci. Rep.
6
:
37313
.

The authors have no financial conflicts of interest.

This article is distributed under The American Association of Immunologists, Inc., Reuse Terms and Conditions for Author Choice articles.

Supplementary data