Fig 1.
Graphical abstract: Diagram of the workflow of the proposed methodology of the steps used to reach the possible validity of the molecular mimicry hypothesis in IBD disease through human HSPs and bacterial equivalents.
Each part of the pyramid shows the steps of the method from top to bottom. The Bullitt shape next to each step of the pyramid provided practical explanations of the study results, as well as the number of epitopes predicted from each of the HSPs presented by HLAs at that step. The workflow of 7 steps of the proposed method to predict the epitopes is as follows: In step 1, Protein targets: alignment, and homology analysis (Similarity percent), all bacteria exhibited substantial resemblance similarity with HSP 60: 96-100% (E. coli and S. typhi revealed 100% similarity); HSP 70: Almost 99- 100% (except E. coli with 92% similarity); HSP 90: Almost 90- 93% (except L. monocytogenes = 70% and C. difficile = 82%. In step 2, CTL 9-mer epitopes were selected based on HLA, toxicity, and antigenicity considerations. In step 3, the Optimal HTL 15-mer Epitope Mapping (overlapping in terms of CTL) includes HSP 60 = 3234 (presented by MHC-II: related = 1471; not related 1763), HSP 70 = 3900 (presented by MHC-II: related = 1770; not related = 2130), HSP 90 = 3991 (presented by MHC-II: related = 1813; not related = 2178). In step 3, the HSP 60-related overlapping HTL epitope with CTL epitope presented by related (to IBD) MHC II HLA includes DRB3*01:01 = 295, DQA1*05:01-DQB1*02:01 = 294, DRAB1*03:01 = 294, DRB1*04:01 = 294, DRB1*13:01 = 294. In step 3, the HSP 60-related overlapping HTL epitope with CTL epitope presented by not related (to IBD) MHC II HLA includes: DRB1*01:01 = 295, DRB1*15:01 = 295, DRB1*12:01 = 295, DRB1*13:02 = 295, DRB1*11:01 = 295, DRB1*07:01 = 288. The HSP 70 related overlapping HTL epitope with CTL epitope presented by related (to IBD) MHC II HLA includes: DRB3*01:01 = 355, DQA1*05:01-DQB1*02:01 = 353, DRAB1*03:01 = 355, DRB1*04:01 = 354, DRB1*13:01 = 353. The HSP 70 related overlapping HTL epitope with CTL epitope presented by not related (to IBD) MHC II HLA includes DRB1*01:01 = 355, DRB1*15:01 = 355, DRB1*12:01 = 355, DRB1*13:02 = 355, DRB1*11:01 = 355, DRB1*07:01 = 355. The HSP 90 related overlapping HTL epitope with CTL epitope presented by related (to IBD) MHC II HLA includes: DRB3*01:01 = 363, DQA1*05:01-DQB1*02:01 = 363, DRAB1*03:01 = 363, DRB1*04:01 = 361, DRB1*13:01 = 363. The HSP 90 related overlapping HTL epitope with CTL epitope presented by not related (to IBD) MHC II HLA includes DRB1*01:01 = 363, DRB1*15:01 = 363, DRB1*12:01 = 363, DRB1*13:02 = 363, DRB1*11:01 = 363, DRB1*07:01 = 363. In step 4, the Optimal HTL 15-mer Epitope Mapping (overlapping in terms of CTL-Percentile Rank <10) includes HSP 60 = 136 (presented by MHC-II: related = 69; not related = 67), HSP 70 = 150 (presented by MHC-II: related = 90; not related = 60), HSP 90 = 175 (presented by MHC-II: related = 96; not related = 79). In step 5, the Optimal HTL 15-mer Epitope Mapping (overlapping in terms of CTL-Percentile Rank <10-Antigenicity VaxiJen > 0.4) includes: HSP 60 = 62 (presented by MHC-II: related = 33; not related = 36), HSP 70 = 77 (presented by MHC-II: related = 65; not related = 33), HSP 90 = 64 (presented by MHC-II: related = 49; not related = 32). In step 5, as described in the results section, the candidate 15-mer epitopes were chosen according to study criteria (immunological filters). In step 8, 13 epitopes were chosen considering (Table 1) 1- the high dispersion of restricting HLA class II alleles and 2- related percentile rank. In step 9, Comparative Analysis of Epitopic Areas, the human HSPs exhibited a high degree of sequence similarity with all their bacterial counterparts, specifically HSP 60 (positions 269-283) and HSP 70 (positions 361-375, 386-400, and 404-418). Note: Several HLAs can present each epitope, so the sum of epitopes of each HSP is not necessarily obtained from the sum of epitopes in complex with related and unrelated HLAs in opposite parentheses. (Receptor, HLA; Ligand, Epitope). Abbreviations: HSP, heat shock protein; MHC, major histocompatibility complex; HLA, human leukocyte antigen; CTL, cytotoxic T lymphocyte; HTL, helper T lymphocyte; IFN-γ, interferon-gamma.
Table 1.
Listing of 13 characteristics of selected epitopes. These include overlapping CTL, HTL, IFN-γ, antigenicity, and non-toxicity. We ranked peptides as strong, mild, and weak based on their binding manner (percentile). The potent binding peptides are represented by a shade of Lt Trellis style pattern cells, the mild ones by a shade of Lt Vertical style pattern cells, and the weak ones by white cells. The epitope thresholds for classification as strong, mild, and weak are less than 2% for strong, between 2% and 5% for mild, and between 5% and 10% for weak. The amino acids underlined in the “HTL Epitope Sequence” column represent the overlapped core amino acids with the CTL epitope. The machine learning prediction algorithm of the studied server classifies positive numbers in the SVM score as IFN-γ cytokine epitope (seventh column of the table) and negative numbers as non-toxins (eighth column). The abbreviations used in the table are HSP (heat shock protein), HLA (human leukocyte antigen), CTL (cytotoxic T lymphocyte), HTL (helper T lymphocyte), IFN-γ (interferon-gamma), SVM (Support vector machine).
Fig 2.
3D structural alignment visualization of HSPs and their counterparts using Chimera.
Human protein chains are represented in green, while bacterial proteins are represented in blue. The figure illustrates a clear correspondence between the alpha-helix and beta-sheet structures of human HSPs and their bacterial counterparts. The structural alignment presented indicates a noteworthy similarity between each pair: (A) HSP60 (8g7lH) aligned with GroEL as the HSP60 counterpart of M. tuberculosis (3rtkA). (B) HSP60 (8g7lJ) aligned with GroEL as the HSP60 counterpart of M. tuberculosis (3rtkB). (C) HSP70 (7kw7C) aligned with DnaK as the HSP70 counterpart of M. tuberculosis (6w6eI). (D) HSP70 (7kw7D) aligned with DnaK as the HSP70 counterpart of M. tuberculosis (6w6eI). (E) HSP90 (7zubA) aligned with htpG as the HSP90 counterpart of C. jejuni (AF-Q5HVP5-F1-model_v4). The PDB ID is provided, with the final uppercase letter indicating the respective chain written in parentheses. The RMSD results from the MatchAlign tool in Chimera for pairwise structural alignment show structural similarity, especially for HSP70 pairs. Abbreviations: Three-dimensional (3D); HSP (heat shock protein); Mycobacterium tuberculosis (M. tuberculosis); Campylobacter jejuni (C. jejuni).
Table 2.
The level of similarity between sequences of human HSPs at certain positions, as selected epitopes, and their bacterial counterparts. The abbreviations used in the table are HSP (heat shock protein), C. jejuni (Campylobacter jejuni), C. difficile (Clostridium difficile), E. coli (Escherichia coli), H. pylori (Helicobacter pylori), K. oxytoca (Klebsiella oxytoca), L. monocytogenes (Listeria monocytogenes), MAP (Mycobacterium avium paratuberculosis), M. leprae (Mycobacterium leprae), M. tuberculosis (Mycobacterium tuberculosis), S. typhi (Salmonella typhi), S. dysenteriae (Shigella dysenteriae), S. pneumoniae (Streptococcus pneumoniae), Y. enterocolitica (Yersinia enterocolitica).
Fig 3.
The visualized predicted models of the three-dimensional (3D) structures of MHCs in complex with the four final selected epitopes.
Docking depicts the interaction of selected epitopes with HLA in the active site region. This region illustrates the amino acids of the alpha helix of the agretope in the binding groove (detailed in Figs 6−9, based on MD results). Deep salmon and violet colors represent each epitope related to CTL and HTL, respectively. MHC-I and MHC-II receptors are depicted in aquamarine and green (forest color for chain A and lime green for chain B). Options: HLA Class II alleles, (A) HLA-DQ2 (DQ B1*02/ 6U3M- 1.90 Å resolution) (https://www.rcsb.org/structure/6U3M) carrying KPLVIIAEDVDGEAL (HSP 60- from 269 to 283 AA); (B) DRB1*01:01 (7YX9- 1.76 Å resolution) (https://www.rcsb.org/structure/7YX9) carrying KSINPDEAVAYGAAV (HSP 70- from 361 to 375 AA); (C) DRB1*04:01 (4MD4- 1.95 Å resolution) (https://www.rcsb.org/structure/4MD4) carrying ENVQDLLLLDVAPLS (HSP 70- from 386 to 400 AA); (D) DRB1*11:01 (6CPL- 2.45 Å resolution) (https://www.rcsb.org/structure/6CPL) carrying ETAGGVMTALIKRNS (HSP 70- from 404 to 418 AA). HLA Class I allele: (E) HLA-B*44:02 (B44/ 3KPM- 1.60 Å resolution) (https://www.rcsb.org/structure/3KPM) carrying AEDVDGEAL (HSP 60- from 275 to 283 AA); (F) HLA-B*15:01 (B62/ 5TXS- 1.70 Å resolution) (https://www.rcsb.org/structure/5TXS) carrying INPDEAVAY (HSP 70- from 363 to 371 AA); (G) HLA-A*02:01 (6TDS- 1.70 Å resolution) (https://www.rcsb.org/structure/6TDS) carrying LLLDVAPLS (HSP 70- from 392 to 400 AA); (H) HLA-B*14:02 (3BXN- 1.86 Å resolution) (https://www.rcsb.org/structure/3BXN) carrying TAGGVMTAL (HSP 70- from 405 to 413 AA). Abbreviations: HSP, heat shock protein; MHC, major histocompatibility complex; HLA, human leukocyte antigen; CTL, cytotoxic T lymphocyte; HTL, helper T lymphocyte.
Fig 4.
RMSD plot of HLA-epitope complexes backbone in A-D complexes.
(A) The complex of HLA DQ2 (DQB1*02) HSP60 selected peptide; (B) The complex of HLA DRB1*11:01 HSP70−6 selected peptide; (C) The complex of HLA DRB1*01:01 HSP70−4 selected peptide; (D) The complex of HLA DRB1*04:01 HSP70−5 peptide chosen. (Receptor: HLA; Ligand: Epitope). The RMSD of the HLA as a receptor in complex A initially increased to around 0.55 nm during the first 20 nanoseconds of simulation. Afterward, it stabilized at 3 angstroms from around 40 nanoseconds until the end of the simulation. During the simulation, the epitope, acting as a ligand, quickly reached equilibrium with minimal fluctuations and an average RMSD of 0.2 nm. For further analysis of this complex, we selected the last 60 nanoseconds of the simulation. Complex B achieved stability between 20 and 100 nanoseconds, with the HLA exhibiting an average RMSD value of 0.3 nm and the epitope having an RMSD value of 0.25 nm. Complex C demonstrated an RMSD value of 0.4 nm, and the receptor achieved relative stability after 60 ns. Throughout the simulation, the RMSD value of the epitope remained stable with low fluctuations, averaging 0.35 nm. In complex D, the HLA remained steady at a distance of 0.2 nm for 80 ns before undergoing a conformational change and stabilizing at 0.6 nm. Meanwhile, the epitope remained consistently stable at a distance of 0.2 nM throughout the simulation. Abbreviations: RMSD, root mean square deviation; MD, molecular dynamics; HLA, human leukocyte antigen.
Fig 5.
During simulation, the RMSF plots of HLA-Epitope residues in complexes A to D are shown until the system reaches equilibrium.
(A) The complex of HLA DQ2 (DQB1*02) HSP60 selected peptide; (B) The complex of HLA DRB1*11:01 HSP70−6 selected peptide; (C) The complex of HLA DRB1*01:01 HSP70−4 selected peptide; (D) The complex of HLA DRB1*04:01 HSP70−5 peptide chosen. (Receptor: HLA; Ligand: Epitope). For complexes A to D, the average RMSF of the protein was 1.04, 0.12, 0.2, and 0.22 nm, respectively. On the other hand, the peptides displayed an average RMSF of 0.1, 0.13, 0.24, and 0.14 nm for the same complexes. Abbreviations: RMSF, root mean square fluctuation; MD, molecular dynamics; HLA, human leukocyte antigen.
Fig 6.
Throughout the MD simulation, multiple hydrogen bonds were observed between the epitope and HLA in complexes A through D.
(A) The complex of HLA DQ2 (DQB1*02) HSP60 selected peptide; (B) The complex of HLA DRB1*11:01 HSP70−6 selected peptide; (C) The complex of HLA DRB1*01:01 HSP70−4 selected peptide; (D) The complex of HLA DRB1*04:01 HSP70−5 selected peptide). The number of bonds formed varied, with complex A showing an average of 1.21 bonds (up to 6), complex B exhibiting an average of 4.65 bonds (up to 12), and complex C forming an average of 2.29 bonds (up to 8). Complex D displayed an average of 1.21 bonds (up to 6). Abbreviations: MD, molecular dynamics; HLA, human leukocyte antigen.
Fig 7.
The graphical representation has been provided to illustrate the interaction between the agretope of the epitope (C) as the sequence “KPLVIIAEDVDGEAL” with the HLA binding pockets (chains A and B) in the HLA-DQ2 (DQB1*02)-HSP60 complex.
It has been observed that the epitope does not interact with the A chain of the HLA-DQ2 receptor. However, a hydrophobic interaction has been noted in the B chain, centered on the Ala No. 7, Lys No. 1, and Val No. 4 residues from the 15mer sequence of the selected HSP60 epitope.
Fig 8.
This image depicts the interaction between the agretope of the epitope (C) as the sequence “ETAGGVMTALIKRNS” with residues of the receptor binding pocket of the HLA-DRB1*11:01 (chains A and B) in the HLA-DRB1*11:01-HSP70-6 complex.
Notably, an evident hydrophobic interaction exists between specific epitope residues, including Val No. 6, Gly No. 5, Met No. 7, Thr No. 8, and the A chain of the HLA-DRB1*11:01 receptor. Additionally, we can observe a hydrophobic interaction between residues Ala No. 3, Val No. 6, and Ser No. 15 of the 15mer sequence of the sixth selected epitope of HSP70 protein with the B chain of the receptor.
Fig 9.
From a two-dimensional perspective, an examination of the interactions between the agretope of the epitope (C) as the sequence “KSINPDEAVAYGAAV” with the residues of the receptor binding pocket (chains A and B) in the HLA-DRB1*01:01-HSP70-4 complex is possible.
The data indicate that a hydrophobic interaction is formed between the A chain of the HLA receptor and the epitope residue, facilitated by residue Gly-12 of the epitope. Moreover, a hydrophobic interaction appears between the B chain of the receptor and the fourth selected epitope of the HSP70 protein, and this interaction is facilitated through the cooperation of Ala No. 10, Val No. 9, and Asp No. 6 from the 15mer sequence.
Fig 10.
This image depicts the interaction between the agretope of the epitope (C) as the “ENVQDLLLLDVAPLS” sequence, as well as the residues found in the receptor binding pocket (chains A and B) of the HLA-DRB1*04:01-HSP70-5 complex.
The figure reveals that only one residue of the fifth selected 15mer epitope of HSP70 has entered into a hydrophobic interaction with each of the receptor chains. Specifically, Asn No. 2 and Val No. 3 interact with the A and B chains of the receptor, respectively.