In silico characterization of immunogenic epitopes presented by HLA-Cw*0401

Immunome Research BioMed Central Research Open Access In silico characterization of immunogenic epitopes presented by HLA-Cw*0401 Joo Chuan Tong1, Zong Hong Zhang1, J Thomas August3, Vladimir Brusic4, Tin Wee Tan2 and Shoba Ranganathan*5,2 Address: 1Institute for Infocomm Research, 21 Heng Mui Keng Terrace, 119613, Singapore, 2Department of Biochemistry, Yong Loo Lin School of Medicine, National University of Singapore, 8 Medical Drive, 117597, Singapore, 3Department of Pharmacology and Molecular Sciences, John Hopkins University School of Medicine, Baltimore, MD, USA, 4Cancer Vaccine Center, Dana-Farber Cancer Institute, Boston, MA, USA and 5Department of Chemistry and Biomolecular Sciences & Biotechnology Research Institute, Macquarie University, NSW 2109, Australia Email: Joo Chuan Tong - ; Zong Hong Zhang - ; J Thomas August - ; Vladimir Brusic - ; Tin Wee Tan - ; Shoba Ranganathan* - * Corresponding author Published: 20 August 2007 Immunome Research 2007, 3:7 doi:10.1186/1745-7580-3-7 Received: 18 May 2007 Accepted: 20 August 2007 This article is available from: http://www.immunome-research.com/content/3/1/7 © 2007 Tong et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Background: HLA-C locus products are poorly understood in part due to their low expression at the cell surface. Recent data indicate that these molecules serve as major restriction elements for human immunodeficiency virus type 1 (HIV-1) cytotoxic T lymphocyte (CTL) epitopes. We report here a structure-based technique for the prediction of peptides binding to Cw*0401. The models were rigorously trained, tested and validated using experimentally verified Cw*0401 binding and non-binding peptides obtained from biochemical studies. A new scoring scheme facilitates the identification of immunological hot spots within antigens, based on the sum of predicted binding energies of the top four binders within a window of 30 amino acids. Results: High predictivity is achieved when tested on the training (r2 = 0.88, s = 3.56 kJ/mol, q2 = 0.84, spress = 5.18 kJ/mol) and test (AROC = 0.93) datasets. Characterization of the predicted Cw*0401 binding sequences indicate that amino acids at key anchor positions share common physico-chemical properties which correlate well with existing experimental studies. Conclusion: The analysis of predicted Cw*0401-binding peptides showed that anchor residues may not be restrictive and the Cw*0401 binding pockets may possibly accommodate a wide variety of peptides with common physico-chemical properties. The potential Cw*0401-specific T-cell epitope repertoires for HIV-1 p24gag and gp160gag glycoproteins are well distributed throughout both glycoproteins, with thirteen and nine immunological hot spots for HIV-1 p24gag and gp160gag glycoproteins respectively. These findings provide new insights into HLA-C peptide selectivity, indicating that pre-selection of candidate HLA-C peptides may occur at the TAP level, prior to peptide loading in the endoplasmic reticulum. Background Major histocompatibility complex (MHC) class I mole- cules, HLA-A, -B, and -C, are cell surface glycoproteins consisting of a polymorphic heavy α chain non-covalently Page 1 of 8 (page number not for citation purposes) Immunome Research 2007, 3:7 linked to a light chain, β2-microglobulin (β2m). HLA-A and -B molecules play critical roles in cell mediated immune responses by binding short antigenic peptide fragments and presenting them on the surface of antigenpresenting cells for recognition by the CD8+ cytotoxic T lymphocyte (CTL). Although several HLA-C specificities with CTL epitopes have been reported [1,2], much remains unknown with regards to their role in the immune response against viral antigens in part due to their poor expression at the cell surface [3,4]. Recent research shows that this group of molecules plays a major role in the control of human immunodeficiency virus type 1 (HIV-1) infection [5]. Improved understanding of peptide binding to this group of molecules is important in the study of HIV-1 disease progression, as well as the design of effective HIV peptide vaccines. The HLA-C allele, Cw*0401, is of particular interest in the study of HIV-1 disease progression because it is the restriction element for HIV-1 proteins [5]. Two HIV-1 proteins (p24gag and gp160gag) are currently known to be restricted by Cw*0401 [5]. Cw*0401 is present in approximately 10% of the general population [6]. The allele is expressed intracellularly in amounts comparable with HLA-A and -B molecules, but is poorly expressed at the cell surface [7,8]. Improved understanding of peptide binding to this molecule is important for elucidating its role in HIV-1 disease progression. Computational strategies for prediction of peptide binding to HLA-A and -B molecules are relatively advanced [9], while sequence-based predictive models for HLA-C molecules have encountered limited success due to the lack of experimental training data [10]. Two matrix-based prediction algorithms for Cw*0401 were reported [11,12], but a sequence independent approach is still lacking. To overcome these limitations, we have developed a structurebased predictive technique that integrates the strength of Monte Carlo simulations and homology modeling [1315]. This method utilizes a probe or "base fragment" to sample different regions of the receptor binding site, followed by loop closure and refinement of the entire class I peptide. The technique has been successfully applied to analyze peptides binding to a variety of MHC class II alleles [14,15]. In this work, we now extend our analysis to peptides presented by the class I HLA-C molecule. We investigated the HIV-1 p24gag and gp160gag peptide binding repertoire of Cw*0401 and illustrate that areas with high concentration of T-cell epitopes or "immunological hot spots" are potentially well distributed throughout both HIV-1 p24gag and gp160gag. We also show that Cw*0401 can possibly bind antigenic peptides in amounts comparable to both HLA-A and -B molecules. Characterization of predicted Cw*0401 binding sequences reveal that Cw*0401 may bind a large variety of http://www.immunome-research.com/content/3/1/7 amino acids at anchor positions with common physicochemical properties which correlate well with existing experimental studies [11]. Results and discussion Cw*0401 predictive model High predictivity (r2 = 0.88, s = 3.56 kJ/mol, q2 = 0.84, spress = 5.18 kJ/mol) is achieved when tested on the training dataset of 6 Cw*0401 peptide sequences. The Cw*0401 predictive model outperforms the predictive models done by Rognan et al. [16] on training datasets of 5 A*0204 (r2 = 0.85, spress = 2.40 kJ/mol) and 37 2Kk (r2 = 0.78, spres (...truncated)