Predicting sequences and structures of MHC-binding peptides: a computational combinatorial approach.

Peptides bound to MHC molecules on the surface of cells convey critical information about the cellular milieu to immune system T cells. Predicting which peptides can bind an MHC molecule, and understanding their modes of binding, are important in order to design better diagnostic and therapeutic agents for infectious and autoimmune diseases. Due to the difficulty of obtaining sufficient experimental binding data for each human MHC molecule, computational modeling of MHC peptide-binding properties is necessary. This paper describes a computational combinatorial design approach to the prediction of peptides that bind an MHC molecule of known X-ray crystallographic or NMR-determined structure. The procedure uses chemical fragments as models for amino acid residues and produces a set of sequences for peptides predicted to bind in the MHC peptide-binding groove. The probabilities for specific amino acids occurring at each position of the peptide are calculated based on these sequences, and these probabilities show a good agreement with amino acid distributions derived from a MHC-binding peptide database. The method also enables prediction of the three-dimensional structure of MHC-peptide complexes. Docking, linking, and optimization procedures were performed with the XPLOR program [1].

A novel member of the STOMATIN/EPB72/mec-2 family, stomatin-like 2 (STOML2), is ubiquitously expressed and localizes to HSA chromosome 9p13.1.

A cDNA encoding a novel second member of the Band7/stomatin-like/SPFH domain family in humans designated stomatin-like 2 (STOML2) has been isolated using the technique of cDNA Representational Difference Analysis. The STOML2 cDNA encoded a 356 amino acid residue polypeptide with a predicted molecular weight of 38.5 kDa. The predicted polypeptide sequence of STOML2 could be delineated into three major domains: an N-terminal alpha-helical region; a domain with significant similarity to a 172 amino acid region of the HSA stomatin polypeptide, composed of an alternating alpha-helical and beta-sheet structure and a C-terminal domain that was mostly alpha-helical. The stomatin-like domain was observed in 51 other proteins with potentially diverse functions. Based on its homology to stomatin, STOML2 was predicted to be cytoplasmically located. However, unlike most of the other proteins containing stomatin-like domains, the predicted STOML2 polypeptide does not contain a transmembrane region although the presence of N-myristoylation sites suggest that it has the potential to be membrane-associated. Northern blot analysis of a panel of poly(A)(+) mRNA from normal human adult tissues showed that a single 1.3-kb mRNA transcript encoding STOML2 was ubiquitously expressed, with relatively higher levels in skeletal muscle and heart compared to other tissues. Comparison of the STOML2 cDNA sequence with human genomic DNA indicated that the gene encoding STOML2 was 3,250 bp long and consisted of ten exons interrupted by nine introns. We have mapped STOML2 to HSA chromosome 9p13.1, a region that is rearranged in some cancers and thought to contain the gene responsible for acromesomelic dysplasia.

Identification of ligand-binding site III on the immunoglobulin-like domain of the granulocyte colony-stimulating factor receptor.

The granulocyte colony-stimulating factor receptor (G-CSF-R) forms a tetrameric complex with G-CSF containing two ligand and two receptor molecules. The N-terminal Ig-like domain of the G-CSF-R is required for receptor dimerization, but it is not known whether it binds G-CSF or interacts elsewhere in the complex. Alanine scanning mutagenesis was used to show that residues in the Ig-like domain of the G-CSF-R (Phe(75), Gln(87), and Gln(91)) interact with G-CSF. This binding site for G-CSF overlapped with the binding site of a neutralizing anti-G-CSF-R antibody. A model of the Ig-like domain showed that the binding site is very similar to the viral interleukin-6 binding site (site III) on the Ig-like domain of gp130, a related receptor. To further characterize the G-CSF-R complex, exposed and inaccessible regions of monomeric and dimeric ligand-receptor complexes were mapped with monoclonal antibodies. The results showed that the E helix of G-CSF was inaccessible in the dimeric but exposed in the monomeric complex, suggesting that this region binds to the Ig-like domain of the G-CSF-R. In addition, the N terminus of G-CSF was exposed to antibody binding in both complexes. These data establish that the dimerization interface of the complete receptor complex is different from that in the x-ray structure of a partial complex. A model of the tetrameric G-CSF.G-CSF-R complex was prepared, based on the viral interleukin-6.gp130 complex, which explains these and previously published data.

Design of inhibitors of Ras--Raf interaction using a computational combinatorial algorithm.

Drugs that inhibit important protein-protein interactions are hard to find either by screening or rational design, at least so far. Most drugs on the market that target proteins today are therefore aimed at well-defined binding pockets in proteins. While computer-aided design is widely used to facilitate the drug discovery process for binding pockets, its application to the design of inhibitors that target the protein surface initially seems to be limited because of the increased complexity of the task. Previously, we had started to develop a computational combinatorial design approach based on the well-known 'multiple copy simultaneous search' (MCSS) procedure to tackle this problem. In order to identify sequence patterns of potential inhibitor peptides, a three-step procedure is employed: first, using MCSS, the locations of specific functional groups on the protein surface are identified; second, after constructing the peptide main chain based on the location of favorite locations of N-methylacetamide groups, functional groups corresponding to amino acid side chains are selected and connected to the main chain C(alpha) atoms; finally, the peptides generated in the second step are aligned and probabilities of amino acids at each position are calculated from the alignment scheme. Sequence patterns of potential inhibitors are determined based on the propensities of amino acids at each C(alpha) position. Here we report the optimization of inhibitor peptides using the sequence patterns determined by our method. Several short peptides derived from our prediction inhibit the Ras--Raf association in vitro in ELISA competition assays, radioassays and biosensor-based assays, demonstrating the feasibility of our approach. Consequently, our method provides an important step towards the development of novel anti-Ras agents and the structure-based design of inhibitors of protein--protein interactions.

Characterization of a comparative model of the extracellular domain of the epidermal growth factor receptor.

The Epidermal Growth Factor (EGF) receptor is a tyrosine kinase that mediates the biological effects of ligands such as EGF and transforming growth factor alpha. An understanding of the molecular basis of its action has been hindered by a lack of structural and mutational data on the receptor. We have constructed comparative models of the four extracellular domains of the EGF receptor that are based on the structure of the first three domains of the insulin-like growth factor-1 (IGF-1) receptor. The first and third domains of the EGF receptor, L1 and L2, are right-handed beta helices. The second and fourth domains of the EGF receptor, S1 and S2, consist of the modules held together by disulfide bonds, which, except for the first module of the S1 domain, form rod-like structures. The arrangement of the L1 and S1 domains of the model are similar to that of the first two domains of the IGF-1 receptor, whereas that of the L2 and S2 domains appear to be significantly different. Using the EGF receptor model and limited information from the literature, we have proposed a number of regions that may be involved in the functioning of the receptor. In particular, the faces containing the large beta sheets in the L1 and L2 domains have been suggested to be involved with ligand binding of EGF to its receptor.

A method for computational combinatorial peptide design of inhibitors of Ras protein.

A computational combinatorial approach is proposed for the design of a peptide inhibitor of Ras protein. The procedure involves three steps. First, a 'Multiple Copy Simultaneous Search' identifies the location of specific functional groups on the Ras surface. This search method allowed us to identify an important binding surface consisting of two beta strands (residues 5-8 and 52-56), in addition to the well known Ras effector loop and switch II region. The two beta strands had not previously been reported to be involved in Ras-Raf interaction. Second, after constructing the peptide inhibitor chain based on the location of N-methylacetamide (NMA) minima, functional groups are selected and connected to the main chain Calpha atom. This step generates a number of possible peptides with different sequences on the Ras surface. Third, potential inhibitors are designed based on a sequence alignment of the peptides generated in the second step. This computational approach reproduces the conserved pattern of hydrophobic, hydrophilic and charged amino acids identified from the Ras effectors. The advantages and limitations of this approach are discussed.

Interaction of granulocyte colony-stimulating factor (G-CSF) with its receptor. Evidence that Glu19 of G-CSF interacts with Arg288 of the receptor.

Granulocyte colony-stimulating factor (G-CSF) forms a tetrameric complex with its receptor, comprising two G-CSF and two receptor molecules. The structure of the complex is unknown, and it is unclear whether there are one or two binding sites on G-CSF and the receptor. The immunoglobulin-like domain and the cytokine receptor homologous module of the receptor are involved in G-CSF binding, and Arg288 in the cytokine receptor homologous module is particularly important. To identify residues in G-CSF that interact with Arg288, selected charged residues in G-CSF were mutated to Ala. To clarify whether there are two binding sites, a chimeric receptor was created in which the Ig domain was replaced with that of the related receptor gp130. This chimera bound G-CSF but could not transduce a signal, consistent with failure of dimerization and loss of one binding site. The G-CSF mutants had reduced mitogenic activity on cells expressing wild-type receptor. When tested with the chimeric receptor, all G-CSF mutants except one (E46A) showed reduced binding, suggesting that Glu46 is important for interaction with the Ig domain. On cells expressing R288A receptor, all the G-CSF mutants except E19A showed reduced mitogenic activity, indicating that Glu19 of G-CSF interacts with Arg288 of the receptor.

Protein-protein recognition: an experimental and computational study of the R89K mutation in Raf and its effect on Ras binding.

Binding of the protein Raf to the active form of Ras promotes activation of the MAP kinase signaling pathway, triggering cell growth and differentiation. Raf/Arg89 in the center of the binding interface plays an important role determining Ras-Raf binding affinity. We have investigated experimentally and computationally the Raf-R89K mutation, which abolishes signaling in vivo. The binding to [gamma-35S]GTP-Ras of a fusion protein between the Raf-binding domain (RBD) of Raf and GST was reduced at least 175-fold by the mutation, corresponding to a standard binding free energy decrease of at least 3.0 kcal/mol. To compute this free energy and obtain insights into the microscopic interactions favoring binding, we performed alchemical simulations of the RBD, both complexed to Ras and free in solution, in which residue 89 is gradually mutated from Arg into Lys. The simulations give a standard binding free energy decrease of 2.9+/-1.9 kcal/mol, in agreement with experiment. The use of numerous runs with three different force fields allows insights into the sources of uncertainty in the free energy and its components. The binding decreases partly because of a 7 kcal/mol higher cost to desolvate Lys upon binding, compared to Arg, due to better solvent interactions with the more concentrated Lys charge in the unbound state. This effect is expected to be general, contributing to the lower propensity of Lys to participate in protein-protein interfaces. Large contributions to the free energy change also arise from electrostatic interactions with groups up to 8 A away, namely residues 37-41 in the conserved effector domain of Ras (including 4 kcal/mol from Ser39 which loses a bifurcated hydrogen bond to Arg89), the conserved Lys84 and Lys87 of Raf, and 2-3 specific water molecules. This analysis will provide insights into the large experimental database of Ras-Raf mutations.

Molecular dynamics simulations of the Ras:Raf and Rap:Raf complexes.

The protein Raf is an immediate downstream target of Ras in the MAP kinase signalling pathway. The complex of Ras with the Ras-binding domain (RBD) of Raf has been modelled by homology to the (E30D,K31E)-Rap1A:RBD complex, and both have been subjected to multiple molecular dynamics simulations in solution. While both complexes are stable, several rearrangements occur in the Ras:RBD simulations: the RBD loop 100-109 moves closer to Ras, Arg73 in the RBD moves towards Ras to form a salt bridge with Ras-Asp33, and Loop 4 of the Ras switch II region shifts upwards toward the RBD. The Ras:RBD interactions (including the RBD-Arg73 interaction) are consistent with available NMR and mutagenesis data on the Ras: RBD complex in solution. The Ras switch II region does not interact directly with the RBD, although indirect interactions exist through the effector domain and bridging water molecules. No large-scale RBD motion is seen in the Ras:RBD complex, compared to the Rap:RBD complex, to suggest an allosteric activation of Raf by Ras. This may be because the Raf kinase domain (whose structure is unknown) is not included in the model.

Disulfide bond structure and N-glycosylation sites of the extracellular domain of the human interleukin-6 receptor.

The high affinity interleukin-6 (IL-6) receptor is a hexameric complex consisting of two molecules each of IL-6, IL-6 receptor (IL-6R), and the high affinity converter and signaling molecule, gp130. The extracellular "soluble" part of the IL-6R (sIL-6R) consists of three domains: an amino-terminal Ig-like domain and two fibronectin-type III (FN III) domains. The two FN III domains comprise the cytokine-binding domain defined by a set of 4 conserved cysteine residues and a WSXWS sequence motif. Here, we have determined the disulfide structure of the human sIL-6R by peptide mapping in the absence and presence of reducing agent. Mass spectrometric analysis of these peptides revealed four disulfide bonds and two free cysteines. The disulfides Cys102-Cys113 and Cys146-Cys157 are consistent with known cytokine-binding domain motifs, and Cys28-Cys77 with known Ig superfamily domains. An unusual cysteine connectivity between Cys6-Cys174, which links the Ig-like and NH2-terminal FN III domains causing them to fold back onto each other, has not previously been observed among cytokine receptors. The two free cysteines (Cys192 and Cys258) were detected as cysteinyl-cysteines, although a small proportion of Cys258 was reactive with the alkylating agent 4-vinylpyridine. Of the four potential N-glycosylation sites, carbohydrate moieties were identified on Asn36, Asn74, and Asn202, but not on Asn226.

Conformation of the Ras-binding domain of Raf studied by molecular dynamics and free energy simulations.

Recognition of Ras by its downstream target Raf is mediated by a Ras-recognition region in the Ras-binding domain (RBD) of Raf. Residues 78-89 in this region occupy two different conformations in the ensemble of NMR solution structures of the RBD: a fully alpha-helical one, and one where 87-90 form a type IV beta-turn. Molecular dynamics simulations of the RBD in solution were performed to explore the stability of these and other possible conformations of both the wild-type RBD and the R89K mutant, which does not bind Ras. The simulations sample a fully helical conformation for residues 78-89 similar to the NMR helical structures, a conformation where 85-89 form a 3(10)-helical turn, and a conformation where 87-90 form a type I beta-turn, whose free energies are all within 0.3 kcal/mol of each other. NOE patterns and H(alpha) chemical shifts from the simulations are in reasonable agreement with experiment. The NMR turn structure is calculated to be 3 kcal/mol higher than the three above conformations. In a simulation with the same implicit solvent model used in the NMR structure generation, the turn conformation relaxes into the fully helical conformation, illustrating possible structural artifacts introduced by the implicit solvent model. With the Raf R89K mutant, simulations sample a fully helical and a turn conformation, the turn being 0.9 kcal/mol more stable. Thus, the mutation affects the population of RBD conformations, and this is expected to affect Ras binding. For example, if the fully helical conformation of residues 78-89 is required for binding, its free energy increase in R89K will increase the binding free energy by about 0.6 kcal/mol.

LIF receptor-gp130 interaction investigated by homology modeling: implications for LIF binding.

Leukemia inhibitory factor (LIF), a member of the gp130 family of helical cytokines, is involved in the hemopoietic and neural systems. The LIF signal transducing complex contains two receptor molecules, the LIF receptor (LIFR) and gp130. The extracellular region of the LIFR is unique in that it includes three membrane-proximal fibronectin type III domains and two cytokine binding domains (CBDs) separated by an immunoglobulin-like domain. Although some mutagenesis data on LIF are available, it is not yet known which regions of LIFR or gp130 bind LIF. Nor is it known whether LIFR contacts gp130 in a manner similar to the growth hormone receptor dimer and, if so, through which of its CBDs. To attempt to elucidate these matters and to investigate the receptor complex, models of the CBDs of LIFR and the CBD of gp130 were constructed. Analyses of the electrostatic isopotential surfaces of the CBD models suggest that gp130 and the membrane-proximal CBD of LIFR hetero-dimerize and bind LIF through contacts similar to those seen in the growth hormone receptor dimer. This work further demonstrates the utility of electrostatic analyses of homology models and suggests a strategy for biochemical investigations of the LIF-receptor complex.

Identification of a ligand-binding site on the granulocyte colony-stimulating factor receptor by molecular modeling and mutagenesis.

Granulocyte colony-stimulating factor (G-CSF) initiates its effects on cells of the neutrophil lineage by inducing formation of a homodimeric receptor complex. The structure of the G-CSF receptor has not yet been determined, therefore we used molecular modeling to identify regions of the receptor that were likely to be involved in ligand binding. The G-CSF receptor sequence was aligned with all the available sequences of the gp130 and growth hormone receptor families and a model of the cytokine receptor homologous domain was constructed, based on the growth hormone receptor structure. Alanine substitution mutagenesis was performed on loops and individual residues that were predicted to bind ligand. Mutant receptors were expressed in factor-dependent Ba/F3 cells and assessed for proliferation response and ligand binding. Six residues were identified that significantly reduced receptor function, with Arg288 in the F'-G' loop having the greatest effect. These residues formed a binding face on the receptor model resembling the growth hormone receptor site, which suggests that the model is reasonable. However, electrostatic analysis of the model provided further evidence that the mechanism of receptor dimerization is different from that of the growth hormone receptor.

A model for the activation of the epidermal growth factor receptor kinase involvement of an asymmetric dimer?

The binding of epidermal growth factor (EGF) to its receptor leads to receptor dimerization, which activates the intracellular kinase domain. Homology models of the inactive and active forms of the EGF-receptor kinase domains have been derived, and these models suggest that the active form can be stabilized by the interaction of helix C and the surrounding area in one receptor monomer with one of two possible complementary surfaces on a second receptor monomer. Both hydrophobic interaction sites are strongly conserved within the EGF-receptor family but not in other tyrosine kinases. Two of the three predicted kinase dimers are symmetric; the other is asymmetric and is predicted to contain only one active kinase. One of the symmetric models and the asymmetric model would account for the effects of two mutations in helix C (Y740F and V741G) on kinase activity. They also provide an explanation for previously reported dominant negative mutants of the EGF receptor and have interesting implications for the signaling through homo- and heterodimers of the family members: EGF receptor, erbB2, erbB3, and erbB4.

Modeling conformational changes in cyclosporin A.

NMR and X-ray structures for the immunosuppressant cyclosporin A (CsA) reveal a remarkable difference between the unbound (free) conformation in organic solvents and the conformation bound to cyclophilin. We have performed computer simulations of the molecular dynamics of CsA under a variety of conditions and confirmed the stability of these two conformations at room temperature in water and in vacuum. However, when the free conformation was modeled in vacuum at 600 K, a transition pathway leading to the bound conformation was observed. This involved a change in the cis MeLeu-9 peptide bond to a trans conformation and the movement of the side chains forming the dominant hydrophobic cluster (residues MeBmt-1, MeLeu-4, MeLeu-6, and MeLeu-10) to the opposite side of the plane formed by the backbone atoms in the molecular ring. The final conformation had a backbone RMS deviation from the bound conformation of 0.53 A and was as stable in dynamics simulations as the bound conformation. Our calculations allowed us to make a detailed analysis of a transition pathway between the free and the bound conformations of CsA and to identify two distinct regions of coordinated movement in CsA, both of which underwent transitions independently.

A potent human interleukin-4 antagonist stimulates the proliferation of murine cells expressing the human interleukin-4 binding chain.

A single-amino-acid substitution mutant form of human interleukin-4 (hIL-4), Y124D.hIL-4, has been described previously as an antagonist of the effects of hIL-4 on various human cells. The murine T-cell leukemic cell line CT.h4S, which expresses the human IL-4 receptor, proliferates in response to both hIL-4 and murine IL-4. Although Y124D.hIL-4 antagonizes the proliferative effects of hIL-4 on human phytohaemagglutinin-stimulated peripheral blood mononuclear cells, Y124D.hIL-4 is a potent stimulator for CT.h4S cells. Molecular modelling studies were performed to investigate the stability of different conformations of residue 124 as well as the efficiency of different molecular mechanics force fields in homology modelling. We suggest that the aspartate substitution alters the C-terminal end of the D-helix in such way that the analogue still binds to the human IL-4 receptor alpha-chain and signals through the murine gamma c-chain. In contrast, the Y124D.hIL-4/IL-4 receptor complex cannot signal through the human gamma c-chain.

Conversion of the biological specificity of murine to human leukemia inhibitory factor by replacing 6 amino acid residues.

Mouse and human leukemia inhibitory factor (mLIF and hLIF) have approximately 80% amino acid identity, but mLIF cannot bind to the hLIF receptor (hLIF-R), while hLIF binds to the alpha-chain of the mLIF receptor (mLIF-R) with a much higher affinity than does mLIF. We have previously shown that the same regions confer both these properties of hLIF and map to an area within the predicted third alpha-helix and two of the loops of hLIF (Owczarek, C. M., Layton, M. J., Metcalf, D., Lock, P., Wilson, T. A., Gough, N. M., and Nicola, N. A. (1993) EMBO J. 12, 3487-3495). The present studies, using interspecies chimeras of mLIF and hLIF, have defined 6 residues (Asp57, Ser107, His112, Ser113, Val155, and Lys158) that contribute most of the binding energy involved in the interaction of hLIF and the hLIF-R alpha-chain, and form a surface at one end of the predicted four alpha-helical bundle of the hLIF molecule. Mouse LIF is unable to bind to the hLIF-R alpha-chain or activate the cellular hLIF-R, but when these 6 residues were substituted onto an mLIF framework, they were able to reconstitute both the binding and biological activities specific to hLIF.

Homology modelling and 1H NMR studies of human leukaemia inhibitory factor.

Human leukaemia inhibitory factor (LIF) is a glycoprotein with a diverse range of activities on many cell types. A molecular model of LIF has been constructed based mainly on the structure of the related cytokine granulocyte colony-stimulating factor, and refined using simulated annealing and molecular dynamics in water. The model was stable during molecular dynamics refinement and is consistent with known stereochemical data on proteins. It has been assessed by comparison with 1H NMR data on the ionization behaviour of the six histidine residues in LIF, the imidazolium pKa values of which range from 3.6 to 7.4. These pKa values were assigned to individual histidine residues from NMR studies on a series of His-->Ala mutants. The environments of the histidine residues in the model account very well for their observed ionization behaviour. Furthermore, the model is consistent with mutagenesis studies which have defined a group of amino acid residues involved in receptor binding.

A dimerization motif for transmembrane alpha-helices.

Specific helix-helix interactions inside lipid bilayers guide the folding and assembly of many integral membrane proteins and their complexes. We report here a pattern of 7 amino acids (LIxxGVxxGVxxT) which when introduced into several hydrophobic transmembrane alpha-helices promotes their specific dimerization. Dimerization is driven by interactions that are specific, dominated by the helix-helix interface, and involve no potentially ionizable groups. The motif may provide a useful tool for the functional analysis of such interactions in a variety of systems. Further, since this particular motif is rare, whilst specific helix association is not, many other such motifs may exist, which could permit sorting within complex membranes as well as guiding folding and oligomerization.

Sequence specificity in the dimerization of transmembrane alpha-helices.

While several reports have suggested a role for helix-helix interactions in membrane protein oligomerization, there are few direct biochemical data bearing on this subject. Here, using mutational analysis, we show that dimerization of the transmembrane alpha-helix of glycophorin A in a detergent environment is spontaneous and highly specific. Very subtle changes in the side-chain structure at certain sensitive positions disrupt the helix-helix association. These sensitive positions occur at approximately every 3.9 residues along the helix, consistent with their comprising the interface of a closely fit transmembranous supercoil of alpha-helices. By contrast with other reported cases of interactions between transmembrane helices, the set of interfacial residues in this case contains no highly polar groups. Amino acids with aliphatic side chains define much of the interface, indicating that precise packing interactions between the helices may provide much of the energy for association. These data highlight the potential general importance of specific interactions between the hydrophobic anchors of integral membrane proteins.