HLA-B*27:05-Specific Cytotoxic T Lymphocyte Epitopes in Indian HIV Type 1C

Abstract

HLA-B*27:05 is one of the widely reported alleles associated with resistance to HIV, while HLA-A24, HLA-B7, HLA-B*07:02, HLA-B*35:01, HLA-B*53:01, and HLA-B40 are reported to be associated with susceptibility to HIV. Using a bioinformatics approach we attempted to predict potential HLA-B*27:05-specific HIV-1C epitopes that do not bind to susceptibility-associated HLA alleles based on our hypothesis that such epitopes have a greater probability of eliciting a protective immune response in the host. A consensus sequence was built for all proteins of Indian clade C virus. Epitopes specific to HLA-B*27:05 were predicted from the consensus sequence using two different bioinformatics methods to enhance the accuracy of the prediction. Epitopes that were also predicted to bind to any of the susceptibility-associated HLA alleles were excluded from the list. The short-listed epitopes were modeled using MODPROPEP to refine the prediction. Fourteen peptides were identified as epitopes by both sequence-based methods and were found to interact strongly with HLA-B*27:05 by molecular modeling studies. Five of the 14 epitopes were previously reported as immunogenic by other researchers, while the remaining nine are novel. The 14 epitopes have been repeatedly identified by three different methods indicating their potential as useful candidates for an effective HIV vaccine.

Introduction

I t is becoming more and more evident that an ideal HIV vaccine should elicit both humoral and cellular immune responses in the host. A strong cytotoxic T lymphocyte (CTL) response is often associated with better virus control and slower disease progression.^1,2 Mapping of epitopes in vitro is a time consuming and expensive procedure, particularly in the case of pathogens that are ethnically diverse such as HIV. Developments in bioinformatics have aided the process of prioritizing potential epitopes. Computational tools have been widely used to predict MHC-binding peptides and T cell epitopes in infectious diseases, cancer, autoimmunity, and allergy.³ Many such predictions have been validated by in vitro research.^4
–6

HLA alleles influence the response of the host to the pathogen. HLA-A02⁷ and HLA-A11,⁸ specifically HLA-A*11:01⁹ and HLA-B*27,^7,10

–13 particularly HLA-B*27:05¹⁴ and HLA-B*57¹⁵ have been reported to be associated with resistance to HIV or slow progression to AIDS. Likewise, HLA-A24,⁷ HLA-B7,¹⁴ several alleles of HLA-B35-(HLA-B*35:01,⁷ HLA-B*35:02, and HLA-B*35:03),¹² and HLA-B40,⁸ specifically HLA-B40:06⁹ and HLA-B*53,¹² have been reported to be associated with susceptibility to HIV infection or rapid progression to AIDS. Such differential host genetic associations with infection and disease prompted us to hypothesize that epitopes that could bind specifically to resistance-associated HLA alleles and not susceptibility-associated ones could provoke a robust protective immune response in the host. Such epitopes could form critical components of an effective vaccine against HIV.

HIV-1C is the most common clade responsible for the burden of AIDS globally.¹⁶ Strains of HIV-1C circulating in India are known to be biologically distinct from those in other parts of the world.¹⁷ We therefore chose Indian HIV-1C as a model for the present analysis. Gag and Nef proteins of HIV have been widely reported to stimulate cell-mediated immune responses,^5,6,18 but attempts to identify potential epitopes from the rest of the proteome of Indian HIV-1C isolates are limited. The present study was designed to screen the entire expressed genome of Indian HIV-1C for potential epitopes that bind to HLA-B*27:05, since it is one of the most widely reported HLA alleles associated with resistance to HIV, having experimentally solved crystal structures that could be used for modeling studies.

Materials and Methods

Building of consensus sequence for HIV-1C proteins

A total of 1,769 sequences of all expressed proteins (Gag, Nef, gp120, gp41, p31-integrase, p51RT, protease, rev, tat, vif, vpr, and vpu) of Indian HIV-1C were downloaded from the Los Alamos HIV sequence database¹⁹ (as of June 1, 2010). Multiple sequence alignment was performed using Clustal W2²⁰ with default parameters. The Consensus Maker from the Los Alamos online resource for HIV²¹ was used to construct a consensus sequence for each protein.

Prediction of HLA-B*27:05-specific epitopes

Possible epitopes specific to HLA-B*27:05 were predicted from the consensus sequence of each protein of Indian HIV-1C using the Propred1²² and IEDB consensus method.²³ The threshold value was set at 3 in the ProPred1 and 5% in the IEDB consensus method. Predicted epitopes that also bind to HIV susceptibility-associated HLA alleles were excluded from further analysis. Those peptides identified as epitopes by both methods were selected for molecular modeling studies.

Molecular modeling of epitopes with HLA-B*27:05

Short-listed epitopes were modeled with HLA-B*27:05 (PDB ID: 1OGT by Hülsmeyer et al. ²⁴) using MODPROPEP.²⁵ MODPROPEP is a web server that helps to model protein-peptide complexes, specifically with MHC proteins and kinases. Residue-based statistical energy function, Miyazawa and Jernigan pair-potential matrix incorporated in the MODPROPEP was chosen for modeling the peptide in the MHC binding pocket. Among the modeled peptides, nonamers that had a better energy score than that found complexed with the HLA-B*27:05 molecule in the crystal structure (1OGT) were considered to be better binders to HLA, and hence probably better immunogens.

Results and Discussion

The diversity of HIV has made it difficult to come up with an effective vaccine for AIDS. Given the fact that strains of HIV-1C from India are distinct from those in other parts of the world, we undertook a comprehensive analysis for immunodominant epitopes in HIV-1C circulating in India. As it has become increasingly evident that region/subtype-specific vaccines are required for HIV, the study aimed to identify suitable candidates for an effective vaccine against Indian HIV-1C. Other investigators have also attempted to identify immunogenic CTL epitopes from Gag and Nef proteins of HIV-1C.^5,6,18 The epitopes identified by them were longer than nine amino acids. While it is thought that class I MHC are known to bind to peptides 8–12 amino acids in length,²⁶ most of the minimal epitopes reported are 9-mers.²⁷ We therefore attempted to predict minimal epitopes from all proteins of Indian HIV-1 subtype C, and have identified highly potential candidates using in silico methods.

Mapping of epitopes in pathogens in vitro is a time-consuming and expensive procedure, and bioinformatics tools can be effectively employed to hand pick a set of highly potential ones that can be further tested using in vitro and in vivo models. Propred1 is one of the widely used programs for prediction of HLA class I-specific epitopes, having a sensitivity of 66–78% and specificity of 80–81%.²² This tool was initially employed to predict epitopes that specifically bind to HLA-B27:05. As usage of more than one predictive method simultaneously is known to increase the accuracy of the results,²⁸ we used an additional bioinformatics tool (the IEDB consensus method)²³ to identify epitopes. This tool is a consensus algorithm for the prediction of CTL epitopes based on four different methods, and it is reported to be superior to any of the individual methods on which it is based.²⁹ Epitopes that were predicted by both methods were further subjected to molecular modeling.

HLA-B*27:05-specific epitopes

HLA-B*27 is reported to be associated with resistance to HIV or slow progression of AIDS in HIV-infected individuals by several investigators,^7,10

–14 and cytotoxic T lymphocyte responses, activated by HLA-B27, are implicated in playing a protective role against HIV infection.³⁰ The frequency of the HLA-B27 allele in the Indian population varies widely between 1% and 29% across different regions.³¹ Among the HLA-B*27 alleles, HLA-B*27:05 is one of the most characterized and its three-dimensional structure is available²⁴ for molecular modeling studies. Association of HLA-B*27:05 with resistance to HIV implies that epitopes that bind specifically and strongly to the HLA-B*27:05 molecule could provoke a robust protective response in the host.

Two hundred and eighty-eight nonamers were predicted as potential binders to HLA-B*27:05. Of these, 92 nonamers were also predicted to bind to HLA molecules that are known to be associated with susceptibility to HIV infection or rapid progression to AIDS. These nonamers were excluded and the remaining 196 nonamers were identified as epitopes specific to HLA-B27:05 by the first method, viz. ProPred1. Furthermore, 96 of these 196 epitopes were also found to be HLA-B*27:05-specific in the second round of analysis using the IEDB consensus method.²³ These were taken up for further analysis. They were modeled onto the peptide binding site of HLA-B*27:05 using the three-dimensional structure available in the Protein Data Bank (PDB). The binding energy of the nonamer bound to the HLA allele present in the PDB was taken as the threshold for short-listing the predicted nonamers as better binders to HLA-B*27:05, and therefore highly probable immunogens. Fourteen nonamers scored better than the control nonamer. These data are summarized in Table 1. Envelope proteins were found to contain a maximum number of epitopes (five from gp41 and two from gp120), followed by gag and protease (two epitopes each), and one each from RT, vpu, and vpr. The details of the short-listed epitopes are given in Table 2.

Table 1.

Short List of Epitopes in Indian HIV-1C That Specifically Bind to HLA-B ^*27:05

HIV proteins	No. of sequences used for establishing consensus sequence	No. of predicted epitopes that bind to HLAB^27:05 using ProPred1*	No. of predicted epitopes that also bind to susceptibility-associated HLA alleles	No. of epitopes that bind to HLA-B27:05 but not to any of the susceptibility-associated HLAs	Epitopes also identified by IEDB method	No. of epitopes that have stronger binding affinity to HLA-B^27:05*
Gag	51	54	14	40	14	2
Nef	113	18	9	9	3	0
Gp120	79	40	15	25	11	2
Gp41	100	45	17	28	21	5
P31-integrase	20	23	6	17	5	0
P51 RT	65	35	13	22	7	1
Protease	775	8	1	7	2	2
Rev (exon1)	117	3	1	2	0	0
Rev (exon2)	123	15	4	11	10	0
Tat (exon 1)	116	8	2	6	3	0
Tat (exon 2)	125	0	0	0	0	0
Vif	22	19	5	14	9	0
Vpr	24	13	4	9	5	1
Vpu	39	7	1	6	6	1
Total	1769	288	92	196	96	14

Table 2.

List of Potential Epitopes Identified in This Study

S. No.	Potential epitopes and starting position in consensus sequence	Protein	MJ energy score (kcal/mol) ^a	Reported in Los Alamos database as T-helper/CD4 ⁺ epitope	Reported in Los Alamos database as CTL/CD8 ⁺ epitope	HLA alleles (other than HLA-B^27:05 and susceptibility-associated ones) predicted to bind to the epitope*
1	271:VRMYSPVSI ^b	Gag	−41.8	WIILGLNKIVRMYSPVSI in p24 (position 133–150); reported in Subtype C	VRMYSPVSI in p24 (position 142–150); reported in Subtype B, C, F	HLA-B14, HLA-B^2702, HLA-B^3901; HLA B^3901, HLA B^5101, HLA B^7301, HLA A^3201, HLA A^*2301
2	437:GRPGNFLQS^b	Gag	−40.68	GKIWPSHKGRPGNFLQSR in p2p7p1p6 (position 72–89); reported in Subtype B, C	—	HLA-B^*2702
3	1:MRVRGILRN^c	gp120	−39.22	—	—	HLA-B^*2702
4	408:CRIKQIINM^b	gp120	−43.57	TIPCRIKQIINMWQ in gp160 (position 414–428); and CRIKQIINMWQEVGR in gp160 (position 418–432); reported in Subtype C	—	HLA-B14, HLA-B^2702, HLA-B^3901; HLA A^*2603
5	51:QQHLLQLTV ^b	gp41	−40.25	QQHLLQLTVWGIKQL in gp160 (position 562–576); Subtype: not mentioned in HIV Molecular Immunology Database	—	HLA-A^0201, HLA-B^2702, HLA-B^5201, HLA-B62; HLA B^4801, HLA B^3901, HLA A^0206
6	184:LRIIFAVLS^b	gp41	−42.22	—	—	HLA-B^*2702
7	235:IRLVNGFLA ^c	gp41	−39.9	—	—	HLA-B14, HLA-B^2702; HLA B^7301
8	260:LRDFILVTA ^c	gp41	−40.51	—	—	HLA-B^2702; HLA B^7301
9	268:ARVVELLGR ^c	gp41	−39.17	—	—	HLA-B^*2702
10	276:VRQLCKLLR ^b	p51(RT)	−39.09	—	—	HLA-B^*2702
11	7:QRPLVSIRV ^c	Protease	−39.55	—	—	HLA-B14, HLA-B^*2702
12	56:VRQYDQIPI ^c	Protease	−39.27	FIKVRQYDQIPIEICGKK in protease (position 53–70); reported in Subtype C	VRQYDQIPIEI in protease (position 56–66); reported in Subtype B	HLA-B^2702; HLA B^3901
13	72:FRIGCQHSR ^b	Vpr	−39.55	—	—	HLA-B^2702, HLA-B^5401; HLA B^3901, HLA A^3301, HLA A^*6801
14	74:LRLLDVNDL ^b	Vpu	−40.59	—	—	HLA-B14, HLA-B^2702, HLA-B^3901, HLA-Cw^*0301

The MJ energy score of the bound HLA-B^*27:05 peptide was −39.01 kcal/mol. Preferred amino acids in anchor sites (P2 and P9) are underlined. Epitopes with preferred amino acids in P2 and P9 and also reported previously to elicit CD4 and/or CD8 responses are given in bold. Epitopes with preferred amino acids in both P2 and P9 are given in italic.

Epitoes conserved identically in more than 50% of the total sequences (58.46% to 91.67%).

Epitoes conserved in more than 50% of the total sequences (56.90% to 83.23%) with one amino acid difference.

We further examined if any of the short-listed epitopes bind to HLA molecules other than those associated with susceptibility to HIV. We found that each of the short-listed HLA-B*27:05-specific epitopes bound to one or more (up to a maximum of eight) HLA alleles (Table 2). The findings that the 14 short-listed epitopes bind strongly to HLA-B*27:05 but not to any of the known susceptibility-associated alleles, and that they interact promiscuously with other HLA types, prompt us to speculate that these epitopes could be highly significant contributions for the design of a CTL-targeted vaccine against the Indian HIV-1C strains.

Conservation of epitopes

HIV is known to evolve by undergoing escape mutations in CD8⁺ T cell epitopes.³² CTL escape mutants transmit from one host to another,³³ but are likely to revert to the wild type in individuals with HLA mismatch.³⁴ Emergence of escape mutations is determined by various factors, viral as well as human, and is thus a slow process.³³ Experimentally defined HIV CTL epitopes are actually concentrated in relatively conserved and potentially important functional regions of the genome^35,36 and it is believed that the emergence of escape mutations may exact a cost on viral fitness.^37

–40 Conservation of the short-listed epitopes was examined by comparing the sequences of the predicted epitopes with the corresponding sequence in an infectious HIV-1C clone from India.⁴¹ The amino acid sequences of 12 of the 14 epitopes were found to be conserved in the infectious clone. One epitope from gp120 (1:MRVRGILRN) varied from the primary infectious clone by one amino acid at the sixth position, and another epitope from gp41 (260:LRDFILVTA) varied by two amino acids at positions 6 and 8. We also examined the extent of conservation of the short-listed epitopes among Indian HIV-1C protein sequences employed to develop the consensus sequence using the “epitope conservancy analysis” tool available from http://tools.immuneepitope.org/tools/conservancy/iedb_input.⁴² Eight of the 14 short-listed epitopes were found to be completely conserved in more than 50% (58–91%) of the total sequences. Allowing for one amino acid difference, the remaining six epitopes also were found to be conserved in 64–83% of the total sequences. In the present analysis we have used HIV sequences obtained over a period of 15 years; thus, the high level of sequence conservation in the short-listed epitopes augments their suitability as components of a useful HIV vaccine.

Screening of host mimics

It is reported that epitopes that mimic human peptides are tolerated by our immune system while peptides that do not resemble human peptides induce an immunogenic response in the human host.⁴³ Further, Maksyutov et al. ⁴⁴ suggested that epitopes shared by HIV and the human host are likely to be implicated in the immunopathogenesis of AIDS through induction of cross-reacting effectors of the immune system or would cause autoimmune reactions. An inverse relationship has been reported between HIV epitope's similarity with human proteins and their immunogenicity.⁴⁵ Thus, the human immune system preferentially responds to antigenic sequences that are different from human peptides, and hence removal of host-mimics from among the list of epitopes could be crucial while designing an efficacious vaccine. The short-listed epitopes were compared with the human proteome (Homo sapiens, taxid: 9606) using BLASTP (with parameters automatically adjusted for short sequence search)⁴⁶ to look for the presence of identical peptides. None of the 14 short-listed epitopes were found to be maximally conserved in the human peptides. We extended our analysis to evaluate whether peptides sharing eight or less numbers of residues exist in humans. However, none of epitopes was observed to have similar host peptides sharing eight residues. Five of these epitopes differed by two amino acids, and two epitopes differed by three amino acids. Epitopes of HIV differing by a minimum of two residues from human peptides have been previously demonstrated to stimulate immune response against HIV.⁴⁵ The uniqueness of these epitopes implies that they are likely to be recognized by the human immune system as foreign, resulting in the induction of an immunogenic response in man.

Epitopes already reported by others

The set of short-listed epitopes was cross-checked with the HIV Molecular Immunology Database⁴⁷ to see if any of them were already reported by others. We found two (Gag271:VRMYSPVSI and protease 56:VRQYDQIPI) of the 14 short-listed epitopes to be previously identified and reported as CTL epitopes. Though epitopes specific for T-helper cells are generally 9–22 amino acids in length, only nine residues actually fit into the binding groove of the MHC class II molecule.^27,48 Thus, we examined whether any of the 14 epitopes were present in the list of T-helper cell-specific epitopes listed in the Los Alamos Database. Interestingly, five of the 14 CTL epitopes were found to be part of T-helper cell epitopes in the HIV Molecular Immunology Database.⁴⁷ Four of the 5 CD4⁺ epitopes have been reported from HIV-1 subtype C. Furthermore, two of these five epitopes (gag271:VRMYSPVSI and protease 56:VRQYDQIPI) were reported to activate CD8⁺ as well as CD4⁺ cells, thus are polyfunctional. The length, position, sequence, and source proteins for each of these epitopes are listed in Table 2. The fact that five of the 14 short-listed epitopes are already reported to be immunogenic suggests that these could be significant components for a HIV vaccine. Furthermore, our study has identified nine novel epitopes that could be tested in vitro for their immunogenic potential.

Interaction profile of HLA-B*27:05 and potential epitopes

By a comprehensive analysis of all known natural ligands of HLA-B27 and their amino acid composition, Lopez de Castro et al. ⁴⁹ found that a large number of epitopes contained arginine at position two (P2), and a basic, aliphatic, or aromatic amino acid in position 9 (P9). In general, amino acids in position 2 and the extreme C-terminus of the epitope are generally reported to be critical for binding to most class I MHCs and these positions are referred to as anchor positions,⁵⁰ while amino acids in the other positions function as auxiliary anchor residues.⁴⁸ We analyzed the binding interactions of 14 short-listed epitopes with the HLA-B*27:05 molecule. Interestingly, we observed that 13 of the 14 epitopes had amino acid arginine at position 2. This residue was found to interact with the maximum number (five) of residues in the HLA molecule (Fig. 1b and c). To investigate the binding efficiency of the epitopes in the binding groove of HLA-B*27:05, a representative solid ribbon model and electrostatic surface model were employed (Fig. 1a and d). Our observations support the fact that arginine at position two is critical for the interaction of the epitope with HLA-B*27:05, as previously reported by other investigators.^49,51,52 Stewart Jones et al. ⁵² demonstrated that mutation in arginine at P2 eliminates binding of the epitope with HLA-B27, resulting in immune escape and progression to AIDS. We also observed that 10 of the 14 short-listed epitopes had either a basic, aliphatic, or aromatic amino acid in the C-terminus (P9). The amino acid at P9 interacts with two amino acids in the HLA molecule, namely LYS146 and ASP 77 (Fig. 1b and c). Collectively, eight epitopes were observed to have arginine in P2 and either a basic, aliphatic, or aromatic amino acid in P9, highlighting their significance as immunogenic CTL epitopes in HIV.

FIG. 1.

A representative epitope (Gag 271:VRMYSPVSI) modeled onto HLA-B*27:05 using MODPROPEP. (a) Peptide in the binding groove of HLA-B*27:05. (b, c) Interactions between residues in HLA-B*27:05 and epitope (interacting residues of HLA are represented in cyan with a single letter amino acid code and number; hydrogen bonds are denoted by dotted blue lines); preferred residues at P2 and P9 in the epitope are highlighted in yellow and the remaining residues in the epitope are displayed in pink. (d) Figure representing the electrostatic surface of HLA and bound epitope; the epitope is displayed in stick form with its solid surface shown in partly transparent yellow to display the effective fitting of the epitope in the binding grove of the HLA-B*27:05 molecule, represented by an opaque electrostatic solid surface. Color images available online at www.liebertpub.com/aid

Further analysis revealed that two epitopes, one from Gag and one from protease (Gag271:VRMYSPVSI and Prot56:VRQYDQIPI), that were identified as potentially immunogenic epitopes not only by us but by other investigators as well,⁴⁷ also contain the preferred anchor residues at P2 and P9 that hold the binding specificity for HLA-B27. They are also reported to be polyfunctional in nature, and activate both CD4 and CD8 T cells. Furthermore, these two epitopes are present in at least the two major subtypes of HIV-1 (subtypes B and C).⁴⁷ Thus, the above mentioned epitopes may be considered highly significant candidates that warrant further investigation.

Genetic diversity in HIV has been the greatest stumbling block to the development of a successful vaccine. It has become evident that the strategy for effective vaccine development should aim at targeting individual subtypes or subtypes that are genetically closely related. Since host genetics plays a vital role in resistance or susceptibility to diseases, it would be significant to consider resistance-associated factors while designing a protective vaccine for HIV. In this scenario, the set of short-listed epitopes targeting HLA-B*27:05 would prove to be useful components for an effective vaccine for HIV-1 subtype C strains in India. However, for the vaccine to be useful for all individuals in a given population, other epitopes capable of inducing a protective immune response in those with other HLA alleles also need to be included. The strategy employed in this study can be used to identify epitopes specific to other HLA alleles associated with resistance to the pathogen. For the benefit of individuals who have susceptibility-associated HLA alleles, epitopes specific to the other HLA alleles present in such individuals should be identified.

Footnotes

Acknowledgments

We acknowledge the Indian Council of Medical Research, New Delhi, India for providing the Bioinformatics facility to our institute.

Author Disclosure Statement

No competing financial interests exist.

References

Altfeld

, Kalife

, Qi

et al. HLA alleles associated with delayed progression to AIDS contribute strongly to the initial CD8(+) T cell response against HIV-1. PLoS Med, 2006; 3:e403.

Ogg

, Kostense

, Klein

et al. Longitudinal phenotypic analysis of human immunodeficiency virus type 1-specific cytotoxic T lymphocytes: Correlation with disease progression. J Virol, 1999; 73:9153–9160.

De Groot

. Immunomics: Discovering new targets for vaccines and therapeutics. Drug Discov Today, 2006; 11:203–209.

De Groot

, Jesdale

, Martin

et al. Mapping cross-clade HIV-1 vaccine epitopes using a bioinformatics approach. Vaccine, 2003; 21:4486–4504.

Mendiratta

, Vajpayee

, Malhotra

et al. Characterization of Gag and Nef-specific ELISpot-based CTL responses in HIV-1 infected Indian individuals. Med Microbiol Immunol, 2009; 198:47–56.

Thakar

, Bhonge

, Lakhashe

et al. Cytolytic T lymphocytes (CTLs) from HIV-1 subtype C-infected Indian patients recognize CTL epitopes from a conserved immunodominant region of HIV-1 Gag and Nef. J Infect Dis, 2005; 192:749–759.

Singh

, Kaur

, Sharma

, Mehra

. Immunogenetic basis of HIV-1 infection, transmission and disease progression. Vaccine, 2008; 26:2966–2980.

Selvaraj

, Swaminathan

, Alagarasu

, Raghavan

, Narendran

, Narayanan

. Association of human leukocyte antigen-A11 with resistance and B40 and DR2 with susceptibility to HIV-1 infection in south India. J Acquir Immune Defic Syndr, 2006; 43:497–499.

Raghavan

, Selvaraj

, Swaminathan

, Narendran

. Association of HLA-A*1101 with resistance and B*4006 with susceptibility to HIV and HIV-TB: An in silico analysis of promiscuous T cell epitopes. AIDS Res Hum Retroviruses, 2009; 25:1023–1028.

10.

McNeil

, Yap

, Gore

et al. Association of HLA types A1-B8-DR3 and B27 with rapid and slow progression of HIV disease. QJM, 1996; 89:177–185.

11.

Hill

. The genomics and genetics of human infectious disease susceptibility. Annu Rev Genomics Hum Genet, 2001; 2:373–400.

12.

Kaslow

, Dorak

, Tang

. Influence of host genetic variation on susceptibility to HIV type 1 infection. J Infect Dis, 2005; 191:S68–77.

13.

Neumann-Haefelin

. Protective role of HLA-B27 in HIV and hepatitis C virus infection. Dtsch Med Wochenschr, 2011; 136:320–324.

14.

International HIV Controllers Study. Pereyra

, Jia

et al. The major genetic determinants of HIV-1 control affect HLA class I peptide presentation. Science, 2010; 330:1551–1557.

15.

Migueles

, Sabbaghian

, Shupert

et al. HLA B*5701 is highly associated with restriction of virus replication in a subgroup of HIV-infected long term nonprogressors. Proc Natl Acad Sci USA, 2000; 97:2709–2714.

16.

Hemelaar

, Gouws

, Ghys

, Osmanov

. WHO-UNAIDS Network for HIV Isolation and Characterisation: Global trends in molecular epidemiology of HIV-1 during 2000–2007. AIDS, 2011; 25:679–689.

17.

Shankarappa

, Chatterjee

, Learn

et al. Human immunodeficiency virus type 1 env sequences from Calcutta in eastern India: Identification of features that distinguish subtype C sequences in India from other subtype C sequences. J Virol, 2001; 75:10479–10487.

18.

Kaushik

, Vajpayee

, Wig

, Seth

. Characterization of HIV-1 Gag-specific T cell responses in chronically infected Indian population. Clin Exp Immunol, 2005; 142,2:388–397.

19.

HIV Sequence Database

2010 June 1

http://www.hiv.lanl.gov/content/sequence/HIV/mainpage.html.

20.

ClustalW2–Multiple Sequence Alignment

2010 July 21

http://www.ebi.ac.uk/Tools/clustalw2/index.html.

21.

Consensus Maker

2011 January 8

http://www.hiv.lanl.gov/content/sequence/CONSENSUS/consensus.html.

22.

Singh

, Raghava

. ProPred1: Prediction of promiscuous MHC Class-I binding sites. Bioinformatics, 2003; 19:1009–1014.

23.

IEDB consensus epitope prediction method for MHC-I

2012 March 8

http://tools.immuneepitope.org/analyze/html/mhc_binding.html.

24.

Hülsmeyer

, Fiorillo

, Bettosini

et al. Dual, HLA-B27 subtype-dependent conformation of a self-peptide. J Exp Med, 2004; 199:271–281.

25.

Kumar

, Mohanty

. MODPROPEP: A program for knowledge-based modeling of protein-peptide complexes. Nucleic Acids Res, 2007; 35:W549–555.

26.

Tong

, Tan

, Ranganathan

. Modeling the structure of bound peptide ligands to major histocompatibility complex. Protein Sci, 2004; 13:2523–2532.

27.

Lafuente

, Reche

. Prediction of MHC-peptide binding: A systematic and comprehensive overview. Curr Pharm Des, 2009; 15:3209–3220.

28.

Trost

, Bickis

, Kusalik

. Strength in numbers: Achieving greater accuracy in MHC-I binding prediction by combining the results from multiple prediction tools. Immunome Res, 2007; 3:5.

29.

Moutaftsi

, Peters

, Pasquetto

et al. A consensus epitope prediction approach identifies the breadth of murine T(CD8+)-cell responses to vaccinia virus. Nat Biotechnol, 2006; 24:817–819.

30.

den Uyl

, van der Horst-Bruinsma

, van Agtmael

. Progression of HIV to AIDS: A protective role for HLA-B27? AIDS Rev, 2004; 6:89–96.

31.

Chhaya

. HLA-B27 polymorphism in Mumbai, Western India. Tissue Antigens, 2005; 66,1:48–50.

32.

Kawashima

, Pfafferott

, Frater

et al. Adaptation of HIV-1 to human leukocyte antigen class I. Nature, 2009; 458:641–645.

33.

Fryer

, Frater

, Duda

et al. Modelling the evolution and spread of HIV immune escape mutants. PLoS Pathog, 2010; 6:e1001196.

34.

Leslie

, Pfafferott

, Chetty

et al. HIV evolution: CTL escape mutation and reversion after transmission. Nat Med, 2004; 10:282–289.

35.

Yusim

, Kesmir

, Gaschen

et al. Clustering patterns of cytotoxic T-lymphocyte epitopes in human immunodeficiency virus type 1 (HIV-1) proteins reveal imprints of immune evasion on HIV-1 global variation. J Virol, 2002; 76:8757–8768.

36.

Rodriguez

, Addo

, Rathod

et al. CD8+ T lymphocyte responses target functionally important regions of protease and integrase in HIV-1 infected subjects. J Transl Med, 2004; 2,1:15.

37.

Friedrich

, Dodds

, Yant

et al. Reversion of CTL escape-variant immunodeficiency viruses in vivo. Nat Med, 2004; 10:275–281.

38.

Fernandez

, Stratov

, De Rose

et al. Rapid viral escape at an immunodominant simian-human immunodeficiency virus cytotoxic T-lymphocyte epitope exacts a dramatic fitness cost. J Virol, 2005; 79:5721–5731.

39.

Crawford

, Prado

, Leslie

et al. Compensatory mutation partially restores fitness and delays reversion of escape mutation within the immunodominant HLA-B*5703-restricted Gag epitope in chronic human immunodeficiency virus type 1 infection. J Virol, 2007; 81:8346–8351.

40.

Troyer

, McNevin

, Liu

et al. Variable fitness impact of HIV-1 escape mutations to cytotoxic T lymphocyte (CTL) response. PLoS Pathog, 2009; 5:e1000365.

41.

Mochizuki

, Otsuka

, Matsuo

et al. An infectious DNA clone of HIV type 1 subtype C. AIDS Res Hum Retroviruses, 1999; 15,14:1321–1324.

42.

Bui

, Sidney

, Li

et al. Development of an epitope conservancy analysis tool to facilitate the design of epitope-based diagnostics and vaccines. BMC Bioinform, 2007; 8:361.

43.

Ristori

, Salvetti

, Pesole

et al. Compositional bias and mimicry toward the nonself proteome in immunodominant T cell epitopes of self and nonself antigens. FASEB J, 2000; 14:431–438.

44.

Maksyutov

, Bachinskii

, Bazhan

, Ryzhikov

, Maksyutov

. Exclusion of HIV epitopes shared with human proteins is prerequisite for designing safer AIDS vaccines. J Clin Virol, 2004; 31,Suppl 1:S26–38.

45.

Rolland

, Nickle

, Deng

et al. Recognition of HIV-1 peptides by host CTL Is related to HIV-1 similarity to human proteins. PLoS ONE, 2007; 2:e823.

46.

Altschul

, Madden

, Schäffer

et al. Gapped BLAST and PSI-BLAST: A new generation of protein database search programs. Nucleic Acids Res, 1997; 25:3389–3402.

47.

HIV Molecular Immunology Database

2011 July 6

http://www.hiv.lanl.gov/content/immunology.

48.

Lundegaard

, Lund

, Kesmir

, Brunak

, Nielsen

. Modeling the adaptive immune system: Predictions and simulations. Bioinformatics, 2007; 23:3265–3275.

49.

Lopez de Castro

, Alvarez

, Marcilla

et al. HLA-B27: A registry of constitutive peptide ligands. Tissue Antigens, 2004; 63:424–445.

50.

Rammensee

, Bachmann

, Emmerich

, Bachor

, Stevanović

. SYFPEITHI: Database for MHC ligands and peptide motifs. Immunogenetics, 1999; 50:213–219.

51.

Jardetzky

, Lane

, Robinson

, Madden

, Wiley

. Identification of self peptides bound to purified HLA-B27. Nature, 1991; 353:326–329.

52.

Stewart-Jones

, di Gleria

, Kollnberger

, McMichael

, Jones

, Bowness

. Crystal structures and KIR3DL1 recognition of three immunodominant viral peptides complexed to HLA-B*2705. Eur J Immunol, 2005; 35:341–351.