Evaluation of a competitive enzyme-linked immunosorbent assay for measurements of soluble HLA-G protein.

The human leukocyte antigen (HLA) class Ib molecule, HLA-G, has gained increased attention because of its assumed important role in immune regulation. The HLA-G protein exists in several soluble isoforms. Most important are the actively secreted HLA-G5 full-length isoform generated by alternative splicing retaining intron 4 with a premature stop codon, and the cleavage of full-length membrane-bound HLA-G1 from the cell surface, so-called soluble HLA-G1 (sHLA-G1). A specific and sensitive immunoassay for measurements of soluble HLA-G is mandatory for conceivable routine testing and research projects. We report a novel method, a competitive immunoassay, for measuring HLA-G5/sHLA-G1 in biological fluids. The sHLA-G immunoassay is based upon a competitive enzyme-linked immunosorbent assay (ELISA) principle. It includes a recombinant sHLA-G1 protein in complex with ß2-microglobulin and a peptide as a standard, biotinylated recombinant sHLA-G1 as an indicator, and the MEM-G/9 anti-HLA-G monoclonal antibody (mAb) as the capture antibody. The specificity and sensitivity of the assay were evaluated. Testing with different recombinant HLA class I proteins and different anti-HLA class I mAbs showed that the sHLA-G immunoassay was highly specific. Optimal combinations of competitor sHLA-G1 and capture mAb concentrations were determined. Two versions of the assay were tested. One with a relatively wide dynamic range from 3.1 to 100.0 ng/ml, and another more sensitive version ranging from 1.6 to 12.5 ng/ml. An intra-assay coefficient of variation (CV) of 15.5% at 88 ng/ml and an inter-assay CV of 23.1% at 39 ng/ml were determined. An assay based on the competitive sHLA-G ELISA may be important for measurements of sHLA-G proteins in several conditions: assisted reproduction, organ transplantation, cancer, and certain pregnancy complications, both in research studies and possibly in the future also for clinical routine use.

Identification of peptides from foot-and-mouth disease virus structural proteins bound by class I swine leukocyte antigen (SLA) alleles, SLA-1*0401 and SLA-2*0401.

Characterization of the peptide-binding specificity of swine leukocyte antigen (SLA) class I and II molecules is critical to the understanding of adaptive immune responses of swine toward infectious pathogens. Here, we describe the complete binding motif of the SLA-2*0401 molecule based on a positional scanning combinatorial peptide library approach. By combining this binding motif with data achieved by applying the NetMHCpan peptide prediction algorithm to both SLA-1*0401 and SLA-2*0401, we identified high-affinity binding peptides. A total of 727 different 9mer and 726 different 10mer peptides within the structural proteins of foot-and-mouth disease virus (FMDV), strain A24 were analyzed as candidate T-cell epitopes. Peptides predicted by the NetMHCpan were tested in ELISA for binding to the SLA-1*0401 and SLA-2*0401 major histocompatibility complex class I proteins. Four of the 10 predicted FMDV peptides bound to SLA-2*0401, whereas five of the nine predicted FMDV peptides bound to SLA-1*0401. These methods provide the characterization of T-cell epitopes in response to pathogens in more detail. The development of such approaches to analyze vaccine performance will contribute to a more accelerated improvement of livestock vaccines by virtue of identifying and focusing analysis on bona fide T-cell epitopes.

Identification of a novel UTY-encoded minor histocompatibility antigen.

Minor histocompatibility antigens (mHags) encoded by the Y-chromosome (H-Y-mHags) are known to play a pivotal role in allogeneic haematopoietic cell transplantation (HCT) involving female donors and male recipients. We present a new H-Y-mHag, YYNAFHWAI (UTY(139-147)), encoded by the UTY gene and presented by HLA-A*24:02. Briefly, short peptide stretches encompassing multiple putative H-Y-mHags were designed using a bioinformatics predictor of peptide-HLA binding, NetMHCpan. These peptides were used to screen for peptide-specific HLA-restricted T cell responses in peripheral blood mononuclear cells obtained post-HCT from male recipients of female donor grafts. In one of these recipients, a CD8+ T cell response was observed against a peptide stretch encoded by the UTY gene. Another bioinformatics tool, HLArestrictor, was used to identify the optimal peptide and HLA-restriction element. Using peptide/HLA tetramers, the specificity of the CD8+ T cell response was successfully validated as being HLA-A*24:02-restricted and directed against the male UTY(139-147) peptide. Functional analysis of these T cells demonstrated male UTY(139-147) peptide-specific cytokine secretion (IFNγ, TNFα and MIP-1ß) and cytotoxic degranulation (CD107a). In contrast, no responses were seen when the T cells were stimulated with patient tumour cells alone. CD8+ T cells specific for this new H-Y-mHag were found in three of five HLA-A*24:02-positive male recipients of female donor HCT grafts available for this study.

Strategies to work with HLA data in human populations for histocompatibility, clinical transplantation, epidemiology and population genetics: HLA-NET methodological recommendations.

HLA-NET (a European COST Action) aims at networking researchers working in bone marrow transplantation, epidemiology and population genetics to improve the molecular characterization of the HLA genetic diversity of human populations, with an expected strong impact on both public health and fundamental research. Such improvements involve finding consensual strategies to characterize human populations and samples and report HLA molecular typings and ambiguities; proposing user-friendly access to databases and computer tools and defining minimal requirements related to ethical aspects. The overall outcome is the provision of population genetic characterizations and comparisons in a standard way by all interested laboratories. This article reports the recommendations of four working groups (WG1-4) of the HLA-NET network at the mid-term of its activities. WG1 (Population definitions and sampling strategies for population genetics' analyses) recommends avoiding outdated racial classifications and population names (e.g. 'Caucasian') and using instead geographic and/or cultural (e.g. linguistic) criteria to describe human populations (e.g. 'pan-European'). A standard 'HLA-NET POPULATION DATA QUESTIONNAIRE' has been finalized and is available for the whole HLA community. WG2 (HLA typing standards for population genetics analyses) recommends retaining maximal information when reporting HLA typing results. Rather than using the National Marrow Donor Program coding system, all ambiguities should be provided by listing all allele pairs required to explain each genotype, according to the formats proposed in 'HLA-NET GUIDELINES FOR REPORTING HLA TYPINGS'. The group also suggests taking into account a preliminary list of alleles defined by polymorphisms outside the peptide-binding sites that may affect population genetic statistics because of significant frequencies. WG3 (Bioinformatic strategies for HLA population data storage and analysis) recommends the use of programs capable of dealing with ambiguous data, such as the 'gene[rate]' computer tools to estimate frequencies, test for Hardy-Weinberg equilibrium and selective neutrality on data containing any number and kind of ambiguities. WG4 (Ethical issues) proposes to adopt thorough general principles for any HLA population study to ensure that it conforms to (inter)national legislation or recommendations/guidelines. All HLA-NET guidelines and tools are available through its website http://hla-net.eu.

T cell responses affected by aminopeptidase N (CD13)-mediated trimming of major histocompatibility complex class II-bound peptides.

Endocytosed protein antigens are believed to be fragmented in what appears to be a balance between proteolysis and MHC-mediated epitope protection, and the resulting peptide-MHC complexes are transported to the surface of the antigen-presenting cells (APC) and presented to T cells. The events that lead to antigenic peptide generation and the compartments where antigen processing takes place remains somewhat enigmatic. The importance of intracellular antigen processing has been well established; however, it is unclear whether additional processing occurs at the APC surface. To follow antigen processing, we have identified a pair of T cell hybridomas that recognize a long vs. a short version of the same epitope. We have used prefixed APC and various protease inhibitors to demonstrate that the APC surface has a considerable potential for antigen processing. Specific antibodies further identified the exopeptidase Aminopeptidase N (APN, CD13) as one of the enzymes involved in the observed cell-surface antigen processing. The NH2-terminal end of the longer peptide could, even while bound to major histocompatibility complex (MHC) class II molecules, be digested by APN with dramatic consequences for T cell antigen recognition. This could be demonstrated both in cell-free systems using purified reagents and in cellular systems. Thus, MHC class II and APN may act in concert to generate the final T cell epitopes.

MHC class II-derived peptides can bind to class II molecules, including self molecules, and prevent antigen presentation.

Seven synthetic peptides corresponding to the polymorphic regions of the alpha and beta chains of the I-Ak molecule were examined for their ability to inhibit the presentation of foreign antigens to antigen-specific, I-A-restricted T cell hybridomas. Two of the peptides, representing the sequences found in the first and third polymorphic regions (PMR) of the A alpha k chain (alpha k-1 and alpha k-3) were capable of inhibiting the presentation of three different HEL-derived peptide antigens to their appropriate T cells. In addition, the alpha k-1 peptide inhibited the presentation of the OVA(323-339) immunodominant peptide to the I-Ad-restricted T cell hybridomas specific for it. Prepulsing experiments demonstrated that the PMR peptides were interacting with the APC and not with the T cell hybridomas. These observations were confirmed and extended by the demonstration that the alpha k-1 and alpha k-3 peptides blocked the direct binding of HEL(46-61) to purified I-Ak and that the alpha k-1 peptide blocked the binding of OVA(323-339) to I-Ad. The binding competition experiments suggest that the alpha k-1 peptide binds to the I-Ak molecule from which it was derived with a Kd approximately 10(-5) M, while the alpha k-3 peptide binds slightly less well. These combined data, suggesting that class II-derived peptides can bind to MHC class II molecules, including the autologous molecule from which they are derived, have important implications for the molecular basis of alloreactivity and autoreactivity. Further, they suggest a possible mechanism by which selecting elements, involving only MHC molecules, may be generated in the thymus.

High-affinity human leucocyte antigen class I binding variola-derived peptides induce CD4+ T cell responses more than 30 years post-vaccinia virus vaccination.

Interferon-gamma secreting T lymphocytes against pox virus-derived synthetic 9-mer peptides were tested by enzyme-linked immunospot in peripheral blood of individuals vaccinated with vaccinia virus more than 30 years ago. The peptides were characterized biochemically as high-affinity human leucocyte antigen (HLA) class I binders (K(D)

Modified human beta 2-microglobulin (desLys(58)) displays decreased affinity for the heavy chain of MHC class I and induces nitric oxide production and apoptosis.

Beta2-microglobulin (beta2m) is the light chain of major histocompatibility complex class I (MHC-I) molecules, and is a prerequisite for the binding of peptides to the heavy chain and their presentation to CD8+ T cells. beta2m can be modified in vivo and in vitro by proteolytic cleavage by complement C1 and subsequent carboxypeptidase B-like activity--processes that lead to the generation of desLys(58) beta2m (dbeta2m). This work aims to study the effect of dbeta2m on peptide binding to MHC-I, the influence of dbeta2m on the binding of beta2m to the MHC-I heavy chain and the biological activity of dbeta2m. Both beta2m and dbeta2m are able to support the generation of MHC-I/peptide complexes at 18 degrees C, but complexes formed in the presence of dbeta2m destabilize at 37 degrees C. Moreover, a 250 times higher concentration of dbeta2m than of beta2m is needed to displace MHC-I associated beta2m from the cell surface. In addition, only beta2m is able to restore MHC-I/peptide complex formation on acid-treated cells whereas dbeta2m appears to bind preferentially to denatured MHC-I heavy chains. In cell cultures, exogenously added dbeta2m, but not beta2m, induces apoptotic cell death in monocytic leukaemic cell lines but spares other kinds of leukaemic cells. Additionally, the presence of dbeta2m, and to a lesser extent beta2m, enhances IFN-gamma-induced NO production by monocytic leukaemic cells. In conclusion, these data show that dbeta2m is not able to support the formation of a stable tri-molecular MHC-I complex at physiological temperature and that dbeta2m exerts other biological functions compared to beta2m when bound to cells.

Dendritic cells engineered to express defined allo-HLA peptide complexes induce antigen-specific cytotoxic T cells efficiently killing tumour cells.

Most tumour-associated antigens (TAA) are non-mutated self-antigens. The peripheral T cell repertoire is devoid of high-avidity TAA-specific cytotoxic T lymphocytes (CTL) due to self-tolerance. As tolerance is major histocompatibility complex-restricted, T cells may be immunized against TAA presented by a non-self human leucocyte antigen (HLA) molecule and transferred to cancer patients expressing that HLA molecule. Obtaining allo-restricted CTL of high-avidity and low cross-reactivity has, however, proven difficult. Here, we show that dendritic cells transfected with mRNA encoding HLA-A*0201, efficiently present externally loaded peptides from the antigen, Melan-A/MART-1 to T cells from HLA-A*0201-negative donors. CD8(+) T cells binding HLA-A*0201/MART-1 pentamers were detected already after 12 days of co-culture in 11/11 donors. The majority of cells from pentamer(+) cell lines were CTL and efficiently killed HLA-A*0201(+) melanoma cells, whilst sparing HLA-A*0201(+) B-cells. Allo-restricted CTL specific for peptides from the leukaemia-associated antigens CD33 and CD19 were obtained with comparable efficiency. Collectively, the results show that dendritic cells engineered to express defined allo-HLA peptide complexes are highly efficient in generating CTL specifically reacting with tumour-associated antigens.

Autologous peptides constitutively occupy the antigen binding site on Ia.

Low molecular weight material associated with affinity-purified class II major histocompatibility complex (MHC) molecules of mouse (Ia) had the expected properties of peptides bound to the antigen binding site of Ia. Thus, the low molecular weight material derived from the I-Ad isotype was efficient in inhibiting the binding of 125I-labeled I-Ad-specific peptide to I-Ad, but did not significantly inhibit the binding of an I-Ed-specific peptide to I-Ed; the reciprocal isotype-specific inhibition was demonstrated with low molecular weight material derived from I-Ed. The inhibitory material was predominantly peptide in nature, as shown by its susceptibility to protease digestion. It was heterogeneous as measured by gel filtration (mean molecular weight approximately 3000), and when characterized by high-performance liquid chromatography, it eluted over a wide concentration of solvent. Such self peptide-MHC complexes may have broad significance in the biology of T cell responses, including generation of the T cell repertoire, the specificity of mixed lymphocyte responses, and the immune surveillance of self and nonself antigens in peripheral lymphoid tissues.

The relation between major histocompatibility complex (MHC) restriction and the capacity of Ia to bind immunogenic peptides.

The capacity of purified I-Ad, I-Ed, I-Ak, and I-Ek to bind to protein derived peptides that have been previously reported to be T cell immunogens has been examined. For each of the 12 peptides studied strong binding to the relevant Ia restriction element was observed. All the peptides bound more than one Ia molecule; however, for 11 of 12 peptides, the dominant binding was to the restriction element, whereas in one instance the dominant binding was to a nonrestriction element. When the peptides were used to inhibit the presentation of antigen by prefixed accessory cells to T cells, an excellent correlation was found between the capacity of a peptide to inhibit the binding of an antigen to purified Ia and the capacity of the peptide to inhibit accessory cell presentation of the antigen. Thus, the binding of peptide to purified Ia is immunologically relevant, and Ia seems to be the only saturable molecule on the surface of the accessory cell involved in antigen presentation. Inhibition analysis also indicated that all peptides restricted to a particular Ia molecule competitively inhibited one another, suggesting that each Ia restriction element has a single binding site for antigen. Cross-linking of labeled peptides to Ia followed by electrophoretic analysis and autoradiography suggested that this single binding site is made up of portions of both alpha and beta chains of Ia.

Immunological self, nonself discrimination.

The ability of immunodominant peptides derived from several antigen systems to compete with each other for T cell activation was studied. Only peptides restricted by a given transplantation antigen are mutually competitive. There is a correlation between haplotype restriction, ability to bind to the appropriate transplantation antigen, and ability to inhibit activation of other T cells restricted by the same transplantation antigen. An exception was noted in that a peptide derived from an antigen, bacteriophage lambda cI repressor, binds to the I-Ed molecule in a specific way, yet is not I-Ed-restricted. Comparison of the sequence of the repressor peptide with that of other peptides able to bind to (and be restricted by) I-Ed and a polymorphic region of the I-Ed molecule itself revealed a significant degree of homology. Thus, peptides restricted by a given class II molecule appear to be homologous to a portion of the class II molecule itself. The repressor-derived peptide is identical at several polymorphic residues at this site, and this may account for the failure of I-Ed to act as a restriction element. Comparison of antigenic peptide sequences with transplantation antigen sequences suggests a model that provides a basis for explaining self, nonself discrimination as well as alloreactivity.

Dwell time verification in brachytherapy based on time resolved in vivo dosimetry.

PURPOSE: This paper presents a method to verify dwell times during High Dose Rate (HDR) Brachytherapy (BT) by means of in vivo dosimetry (IVD), and reports on an afterloader's stability in dwell time control. METHODS: In vivo dosimetry was performed during 20 HDR prostate cancer treatments using a point detector based on a radio-luminescence crystal (Al2O3:C) coupled to a fiber-optic cable. The dose rate was recorded at either 10 Hz or 20 Hz during the treatments. The "time of transit" when the source moved between two dwell positions was identified using the difference in count rate between two measurements. The dwell times were then determined by subtracting two adjacent times of transit. The measured dwell times were matched with the planned dwell times and categorised into two groups: Dwell times matching a single dwell position (identified) and dwell times matching the sum of multiple dwell positions (unidentified). Deviations between measured and planned dwell times were calculated for the identified dwell positions. RESULTS: A total of 3518 dwell positions were analysed. The amount of identified dwell positions were 82%, which increased to 89% if the short dwell times (<1 s) were omitted in the analysis. The largest deviation was -0.4 s seen for a single dwell position, and in 97.1% of the cases, the deviations were <0.15 s. CONCLUSION: The dwell times in BT are well controlled by the afterloader. It is shown that IVD facilitates the detection of dwell time offsets that could have a clinical impact.

MHC-I-restricted epitopes conserved among variola and other related orthopoxviruses are recognized by T cells 30 years after vaccination.

It is many years since the general population has been vaccinated against smallpox virus. Here, we report that human leukocyte antigen (HLA) class I restricted T cell epitopes can be recognized more than 30 years after vaccination. Using bioinformatic methods, we predicted 177 potential cytotoxic T lymphocyte epitopes. Eight epitopes were confirmed to stimulate IFN-gamma release by T cells in smallpox-vaccinated subjects. The epitopes were restricted by five supertypes (HLA-A1, -A2, -A24 -A26 and -B44). Significant T cell responses were detected against 8 of 45 peptides with an HLA class I affinity of K(D) less than or equal to 5 nM, whereas no T cell responses were detected against 60 peptides with an HLA affinity of K(D) more than 5 nM. All epitopes were fully conserved in seven variola, vaccinia and cowpox strains. Knowledge of the long-term response to smallpox vaccination may lead to a better understanding of poxvirus immunity and may aid in the development of new improved vaccines and diagnostic tools.

Description and prediction of peptide-MHC binding: the 'human MHC project'.

MHC molecules are crucially involved in controlling the specific immune system. They are highly polymorphic receptors sampling peptides from the cellular environment and presenting these peptides for scrutiny by immune cells. Recent advances in combinatorial peptide chemistry have improved the description and prediction of peptide-MHC binding. It is envisioned that a complete mapping of human immune reactivities will be possible.

Treatment of transplanted CT26 tumour with dendritic cell vaccine in combination with blockade of vascular endothelial growth factor receptor 2 and CTLA-4.

We investigated the anti CT26 tumour effect of dendritic cell based vaccination with the MuLV gp70 envelope protein-derived peptides AH1 and p320-333. Vaccination lead to generation of AH1 specific cytotoxic lymphocytes (CTL) and some decrease in tumour growth of simultaneously inoculated CT26 cells. After combination with an antibody against VEGF receptor 2 (DC101), a significant increase in survival of the tumour cell recipients was observed. Also, monotherapy with an antibody against CTLA-4 (9H10), led to approximately 100% survival of tumour cell recipients. However, effective treatment of mice with already established tumours was only obtained after combination of vaccination, DC101 and 9H10 treatment in which setting 80% of the mice rejected their tumours.

Receptor-ligand interactions measured by an improved spun column chromatography technique. A high efficiency and high throughput size separation method.

Size exclusion chromatography may under the right circumstances be an easy and powerful way to measure in solution the interaction between a receptor an dits ligand. Spun column chromatography is a fast size exclusion technique of increasing popularity, however, little information exists on the method development essential to obtain efficient separation in particular when used for analytical purposes. In this paper we describe a systematic approach to select the optimal parameters for spun column separation including a simple modification of the technique whereby the spun columns are eluted by high-speed gradient centrifugation. This modification is easy to implement and it considerably improves spun column performance. We hypothesize that the high-speed centrifugation step leads to the release of additional buffer which assists in the complete elution of excluded molecules while the gradient centrifugation helps to achieve equilibrium across the gel matrix during the elution. The new method has been used successfully for several different receptor-ligand interactions, and this paper describes a general approach on how to develop new applications of the technique.

Identification of novel and known mutations in the genes for keratin 5 and 14 in Danish patients with epidermolysis bullosa simplex: correlation between genotype and phenotype.

Epidermolysis bullosa simplex (EBS) is a group of autosomal dominant inherited skin diseases caused by mutations in either the keratin 5 (K5) or the keratin 14 (K14) genes and characterized by development of intraepidermal skin blisters. The three major subtypes of EBS are Weber-Cockayne, Koebner, and Dowling-Meara, of which the Dowling-Meara form is the most severe. We have investigated five large Danish families with EBS and two sporadic patients with the Dowling-Meara form of EBS. In the sporadic Dowling-Meara EBS patients, a novel K14 mutation (N123S) and a previously published K5 mutation (N176S) were identified, respectively. A novel K14 mutation (K116N) was found in three seemingly unrelated families, whereas another family harbored a different novel K14 mutation (L143P). The last family harbored a novel K5 mutation (L325P). The identified mutations were not present in more than 100 normal chromosomes. Six polymorphisms were identified in the K14 gene and their frequencies were determined in normal controls. These polymorphisms were used to show that the K14 K116N mutation was located in chromosomes with the same haplotype in all three families, suggesting a common ancestor. We observed a strict genotype-phenotype correlation in the investigated patients as the same mutation always resulted in a similar phenotype in all individuals with the mutation, but our results also show that it is not possible to predict the EBS phenotype merely by the location (i.e., head, rod, or linker domains) of a mutation. The nature of the amino acid substitution must also be taken into account.

Structural requirements for the interaction between class II MHC molecules and peptide antigens.

Previous work from our and other laboratories indicates that T cells recognize a complex between the MHC restriction element and peptide antigen fragments. This paper reviews the structural characteristics of the formation of such a complex. By analyzing in detail the interactions between purified IA(d) and IE(d) molecules and their peptide ligands, we found that some structural characteristics apply to both antigen-MHC interactions. In particular, we found: 1) each MHC molecule is capable of binding many unrelated peptides through the same peptide-binding site; 2) despite this permissiveness of binding, it is possible to define certain structural features of peptides that are associated with the capacity to bind to a particular MHC specificity (IA(d) or IE(d)); 3) IA(d) and IE(d) molecules recognize different and independent structures on the antigen molecule; 4) only about 10% of the single amino acid substitutions tested on two IA(d)- and IE(d)-binding peptides had significant effect on their MHC-binding capacities, while over 80% of these substitutions significantly impaired T cell recognition of the Ia-peptide complex; 5) based on the segregation between residues that are crucial for T cell activation and Ia binding, the easiest model for the antigen-Ia-T-cell-receptor complex pictures the antigen molecule sandwiched in a planar conformation between the MHC and the T cell.

Predicting binding affinities of protein ligands from three-dimensional models: application to peptide binding to class I major histocompatibility proteins.

A simple and fast free energy scoring function (Fresno) has been developed to predict the binding free energy of peptides to class I major histocompatibility (MHC) proteins. It differs from existing scoring functions mainly by the explicit treatment of ligand desolvation and of unfavorable protein-ligand contacts. Thus, it may be particularly useful in predicting binding affinities from three-dimensional models of protein-ligand complexes. The Fresno function was independently calibrated for two different training sets: (a) five HLA-A0201-peptide structures, which had been determined by X-ray crystallography, and (b) three-dimensional models of 37 H-2K(k)-peptide structures, which had been obtained by knowledge-based homology modeling. For both training sets, a good cross-validated fit to experimental binding free energies was obtained with predictive errors of 3-3.5 kJ/mol. As expected, lipophilic interactions were found to contribute the most to HLA-A0201-peptide interactions, whereas H-bonding predominates in H-2K(k) recognition. Both cross-validated models were afterward used to predict the binding affinity of a test set of 26 peptides to HLA-A0204 (an HLA allele closely related to HLA-A0201) and of a series of 16 peptides to H-2K(k). Predictions were more accurate for HLA-A2-binding peptides as the training set had been built from experimentally determined structures. The average error in predicting the binding free energy of the test peptides was 3.1 kJ/mol. For the homology model-derived equation, the average error in predicting the binding free energy of peptides to K(k) was significantly higher (5.4 kJ/mol) but still very acceptable. The present scoring function is thus able to predict with a good accuracy binding free energies from three-dimensional models, at the condition that the backbone coordinates of the MHC-bound peptide have first been determined with an accuracy of about 1-1.5 A. Furthermore, it may be easily recalibrated for any protein-ligand complex.