Protein-protein interaction at crystal contacts

Stability of crystalline proteins

By using two model proteins, glucose oxidase and lipase, we demonstrate that dry crystalline formulations are significantly more stable than their amorphous counterparts. The results of Fourier-transform infrared spectroscopy indicate that crystalline proteins better maintain their native conformation in accelerated stability studies. The lower tendency of crystalline proteins to aggregate is confirmed by size-exclusion chromatography. The data suggest that protein crystallization may significantly improve some aspects of protein handling, and change the way biopharmaceuticals are produced, formulated, and delivered.

A cluster exposed: structure of the Rieske ferredoxin from biphenyl dioxygenase and the redox properties of Rieske Fe-S proteins

BACKGROUND

Ring-hydroxylating dioxygenases are multicomponent systems that initiate biodegradation of aromatic compounds. Many dioxygenase systems include Rieske-type ferredoxins with amino acid sequences and redox properties remarkably different from the Rieske proteins of proton-translocating respiratory and photosynthetic complexes. In the latter, the [Fe2S2] clusters lie near the protein surface, operate at potentials above +300 mV at pH 7, and express pH- and ionic strength-dependent redox behavior. The reduction potentials of the dioxygenase ferredoxins are approximately 150 mV and are pH-independent. These distinctions were predicted to arise from differences in the exposure of the cluster and/or interactions of the histidine ligands.

RESULTS

The crystal structure of BphF, the Rieske-type ferredoxin associated with biphenyl dioxygenase, was determined by multiwavelength anomalous diffraction and refined at 1.6 A resolution. The structure of BphF was compared with other Rieske proteins at several levels. BphF has the same two-domain fold as other Rieske proteins, but it lacks all insertions that give the others unique structural features. The BphF Fe-S cluster and its histidine ligands are exposed. However, the cluster has a significantly different environment in that five fewer polar groups interact strongly with the cluster sulfide or the cysteinyl ligands.

CONCLUSIONS

BphF has structural features consistent with a minimal and perhaps archetypical Rieske protein. Variations in redox potentials among Rieske clusters appear to be largely the result of local electrostatic interactions with protein partial charges. Moreover, it appears that the redox-linked ionizations of the Rieske proteins from proton-translocating complexes are also promoted by these electrostatic interactions.

The beta-slip: a novel concept in transthyretin amyloidosis

Transthyretin is a tetrameric plasma protein associated with two forms of amyloid disease. The structure of the highly amyloidogenic transthyretin triple mutant TTRG53S/E54D/L55S determined at 2.3 A resolution reveals a novel conformation: the beta-slip. A three-residue shift in beta strand D places Leu-58 at the position normally occupied by Leu-55 now mutated to serine. The beta-slip is best defined in two of the four monomers, where it makes new protein-protein interactions to an area normally involved in complex formation with retinol-binding protein. This interaction creates unique packing arrangements, where two protein helices combine to form a double helix in agreement with fiber diffraction and electron microscopy data. Based on these findings, a novel model for transthyretin amyloid formation is presented.

Modeling of loops in protein structures

Comparative protein structure prediction is limited mostly by the errors in alignment and loop modeling. We describe here a new automated modeling technique that significantly improves the accuracy of loop predictions in protein structures. The positions of all nonhydrogen atoms of the loop are optimized in a fixed environment with respect to a pseudo energy function. The energy is a sum of many spatial restraints that include the bond length, bond angle, and improper dihedral angle terms from the CHARMM-22 force field, statistical preferences for the main-chain and side-chain dihedral angles, and statistical preferences for nonbonded atomic contacts that depend on the two atom types, their distance through space, and separation in sequence. The energy function is optimized with the method of conjugate gradients combined with molecular dynamics and simulated annealing. Typically, the predicted loop conformation corresponds to the lowest energy conformation among 500 independent optimizations. Predictions were made for 40 loops of known structure at each length from 1 to 14 residues. The accuracy of loop predictions is evaluated as a function of thoroughness of conformational sampling, loop length, and structural properties of native loops. When accuracy is measured by local superposition of the model on the native loop, 100, 90, and 30% of 4-, 8-, and 12-residue loop predictions, respectively, had <2 A RMSD error for the mainchain N, C(alpha), C, and O atoms; the average accuracies were 0.59 +/- 0.05, 1.16 +/- 0.10, and 2.61 +/- 0.16 A, respectively. To simulate real comparative modeling problems, the method was also evaluated by predicting loops of known structure in only approximately correct environments with errors typical of comparative modeling without misalignment. When the RMSD distortion of the main-chain stem atoms is 2.5 A, the average loop prediction error increased by 180, 25, and 3% for 4-, 8-, and 12-residue loops, respectively. The accuracy of the lowest energy prediction for a given loop can be estimated from the structural variability among a number of low energy predictions. The relative value of the present method is gauged by (1) comparing it with one of the most successful previously described methods, and (2) describing its accuracy in recent blind predictions of protein structure. Finally, it is shown that the average accuracy of prediction is limited primarily by the accuracy of the energy function rather than by the extent of conformational sampling.

Structure of the C-terminal domain of Tup1, a corepressor of transcription in yeast

The Tup1-Ssn6 corepressor complex regulates the expression of several sets of genes, including genes that specify mating type in the yeast Saccharomyces cerevisiae. Repression of mating-type genes occurs when Tup1-Ssn6 is brought to the DNA by the Matalpha2 DNA-binding protein and assembled upstream of a- and haploid-specific genes. We have determined the 2.3 A X-ray crystal structure of the C-terminal domain of Tup1 (accesion No. 1ERJ), a 43 kDa fragment that contains seven copies of the WD40 sequence motif and binds to the Matalpha2 protein. Moreover, this portion of the protein can partially substitute for full-length Tup1 in bringing about transcriptional repression. The structure reveals a seven-bladed beta propeller with an N-terminal subdomain that is anchored to the side of the propeller and extends the beta sheet of one of the blades. Point mutations in Tup1 that specifically affect the Tup1-Matalpha2 interaction cluster on one surface of the propeller. We identified regions of Tup1 that are conserved among the fungal Tup1 homologs and may be important in protein-protein interactions with additional components of the Tup1-mediated repression pathways.

Protein packing: dependence on protein size, secondary structure and amino acid composition

We have used the occluded surface algorithm to estimate the packing of both buried and exposed amino acid residues in protein structures. This method works equally well for buried residues and solvent-exposed residues in contrast to the commonly used Voronoi method that works directly only on buried residues. The atomic packing of individual globular proteins may vary significantly from the average packing of a large data set of globular proteins. Here, we demonstrate that these variations in protein packing are due to a complex combination of protein size, secondary structure composition and amino acid composition. Differences in protein packing are conserved in protein families of similar structure despite significant sequence differences. This conclusion indicates that quality assessments of packing in protein structures should include a consideration of various parameters including the packing of known homologous proteins. Also, modeling of protein structures based on homologous templates should take into account the packing of the template protein structure.

Characterization and molecular basis of the oligomeric structure of HIV-1 nef protein

The Nef protein of human immunodeficiency virus type I (HIV-1) is an important determinant for the onset of AIDS disease. The self-association properties of HIV-1 Nef are analyzed by chemical cross-linking, dynamic light scattering, equilibrium analytical ultracentrifugation, and NMR spectroscopy. The experimental data show that the HIV-1 Nef core domain forms stable homo-dimers and trimers in solution, but not higher oligomers. These Nef homomers are not covalently linked by disulfide bridges, and the equilibrium between these forms is dependent on the Nef concentration. We further provide the molecular basis for the Nef core dimers and trimers obtained by analysis of crystallographic models. Oligomerization of biological polypeptides is a common tool used to trigger events in cellular signaling and endocytosis, both of which are targeted by Nef. The quaternary structure of Nef may be of physiological importance and may help to connect its cellular targets or to increase affinity of the viral molecule for its ligands. The herein described models for Nef dimers and trimers will allow further mutational studies to elucidate their role in vivo. These results provide novel insight into the structural and functional relationships of this important viral protein. Moreover, the oligomer interface may represent a novel target for the design of antiviral agents.

The tertiary structure of full-length bovine adrenodoxin suggests functional dimers

The three-dimensional X-ray crystal structure of full-length oxidized bovine adrenodoxin (Adx) has been determined at 2.5 A resolution by molecular replacement using a structure of a truncated form as a starting model. Crystals of Adx belong to a primitive monoclinic space group P2(1) with four Adx molecules in an asymmetric unit. The unit cell dimensions are a = 59.44 A, b = 77.03 A, c = 59.68 A, and beta = 94.83 degrees. The structure has been refined to an R factor of 23.5%. Structures of the four molecules of full-length Adx (127 amino acids) in the asymmetric unit were compared with each other and also with that of the truncated Adx (4-108). The overall topology of full-length Adx remains the same as described earlier for the truncated protein. Differences that do occur are almost wholly confined to alternate side-chain conformations that reflect differing lattice contacts made by two proteins. Extensive interactions found between molecules 1 and 2 in the full-length Adx asymmetric unit may reflect the ability of Adx to form dimers in vivo and are consistent with hydrodynamic measurements which show that in solution there is an equilibrium between monomeric and dimeric forms of Adx. Dimerization of Adx could explain why the truncated form has greater affinity for the P450 redox partner than the full-length form. From these results it can be considered that the mechanism of electron transfer is not necessarily the same in different mitochondrial P450 systems.

Protein interaction maps for complete genomes based on gene fusion events

A large-scale effort to measure, detect and analyse protein-protein interactions using experimental methods is under way. These include biochemistry such as co-immunoprecipitation or crosslinking, molecular biology such as the two-hybrid system or phage display, and genetics such as unlinked noncomplementing mutant detection. Using the two-hybrid system, an international effort to analyse the complete yeast genome is in progress. Evidently, all these approaches are tedious, labour intensive and inaccurate. From a computational perspective, the question is how can we predict that two proteins interact from structure or sequence alone. Here we present a method that identifies gene-fusion events in complete genomes, solely based on sequence comparison. Because there must be selective pressure for certain genes to be fused over the course of evolution, we are able to predict functional associations of proteins. We show that 215 genes or proteins in the complete genomes of Escherichia coli, Haemophilus influenzae and Methanococcus jannaschii are involved in 64 unique fusion events. The approach is general, and can be applied even to genes of unknown function.

A systematic study of low-resolution recognition in protein--protein complexes

A comprehensive nonredundant database of 475 cocrystallized protein-protein complexes was used to study low-resolution recognition, which was reported in earlier docking experiments with a small number of proteins. The docking program GRAMM was used to delete the atom-size structural details and systematically dock the resulting molecular images. The results reveal the existence of the low-resolution recognition in 52% of all complexes in the database and in 76% of the 113 complexes with an interface area >4,000 A(2). Limitations of the docking and analysis tools used in this study suggest that the actual number of complexes with the low-resolution recognition is higher. However, the results already prove the existence of the low-resolution recognition on a broad scale.

Crystal structure of the two-RRM domain of hnRNP A1 (UP1) complexed with single-stranded telomeric DNA

Human hnRNP A1 is a versatile single-stranded nucleic acid-binding protein that functions in various aspects of mRNA maturation and in telomere length regulation. The crystal structure of UP1, the amino-terminal domain of human hnRNP A1 containing two RNA-recognition motifs (RRMs), bound to a 12-nucleotide single-stranded telomeric DNA has been determined at 2.1 A resolution. The structure of the complex reveals the basis for sequence-specific recognition of the single-stranded overhangs of human telomeres by hnRNP A1. It also provides insights into the basis for high-affinity binding of hnRNP A1 to certain RNA sequences, and for nucleic acid binding and functional synergy between the RRMs. In the crystal structure, a UP1 dimer binds to two strands of DNA, and each strand contacts RRM1 of one monomer and RRM2 of the other. The two DNA strands are antiparallel, and regions of the protein flanking each RRM make important contacts with DNA. The extensive protein-protein interface seen in the crystal structure of the protein-DNA complex and the evolutionary conservation of the interface residues suggest the importance of specific protein-protein interactions for the sequence-specific recognition of single-stranded nucleic acids. Models for regular packaging of telomere 3' overhangs and for juxtaposition of alternative 5' splice sites are proposed.

Structure of the PH domain from Bruton's tyrosine kinase in complex with inositol 1,3,4,5-tetrakisphosphate

BACKGROUND

The activity of Bruton's tyrosine kinase (Btk) is important for the maturation of B cells. A variety of point mutations in this enzyme result in a severe human immunodeficiency known as X-linked agammaglobulinemia (XLA). Btk contains a pleckstrin-homology (PH) domain that specifically binds phosphatidylinositol 3,4,5-trisphosphate and, hence, responds to signalling via phosphatidylinositol 3-kinase. Point mutations in the PH domain might abolish membrane binding, preventing signalling via Btk.

RESULTS

We have determined the crystal structures of the wild-type PH domain and a gain-of-function mutant E41K in complex with D-myo-inositol 1,3,4,5-tetra-kisphosphate (Ins (1,3,4,5)P4). The inositol Ins (1,3,4,5)P4 binds to a site that is similar to the inositol 1,4,5-trisphosphate binding site in the PH domain of phospholipase C-delta. A second Ins (1,3,4,5)P4 molecule is associated with the domain of the E41K mutant, suggesting a mechanism for its constitutive interaction with membrane. The affinities of Ins (1,3,4,5)P4 to the wild type (Kd = 40 nM), and several XLA-causing mutants have been measured using isothermal titration calorimetry.

CONCLUSIONS

Our data provide an explanation for the specificity and high affinity of the interaction with phosphatidylinositol 3,4,5-trisphosphate and lead to a classification of the XLA mutations that reside in the Btk PH domain. Mis-sense mutations that do not simply destabilize the PH fold either directly affect the interaction with the phosphates of the lipid head group or change electrostatic properties of the lipid-binding site. One point mutation (Q127H) cannot be explained by these facts, suggesting that the PH domain of Btk carries an additional function such as interaction with a Galpha protein.

Crystal structure of deoxycytidylate hydroxymethylase from bacteriophage T4, a component of the deoxyribonucleoside triphosphate-synthesizing complex

Bacteriophage T4 deoxycytidylate hydroxymethylase (EC 2.1.2.8), a homodimer of 246-residue subunits, catalyzes hydroxymethylation of the cytosine base in deoxycytidylate (dCMP) to produce 5-hydroxymethyl-dCMP. It forms part of a phage DNA protection system and appears to function in vivo as a component of a multienzyme complex called deoxyribonucleoside triphosphate (dNTP) synthetase. We have determined its crystal structure in the presence of the substrate dCMP at 1.6 A resolution. The structure reveals a subunit fold and a dimerization pattern in common with thymidylate synthases, despite low (approximately 20%) sequence identity. Among the residues that form the dCMP binding site, those interacting with the sugar and phosphate are arranged in a configuration similar to the deoxyuridylate binding site of thymidylate synthases. However, the residues interacting directly or indirectly with the cytosine base show a more divergent structure and the presumed folate cofactor binding site is more open. Our structure reveals a water molecule properly positioned near C-6 of cytosine to add to the C-7 methylene intermediate during the last step of hydroxymethylation. On the basis of sequence comparison and crystal packing analysis, a hypothetical model for the interaction between T4 deoxycytidylate hydroxymethylase and T4 thymidylate synthase in the dNTP-synthesizing complex has been built.

Lattice simulations of protein crystal formation

A new algorithm is presented for the lattice simulation of protein crystal growth. The algorithm allows the calculation of the size distribution of microcrystals in the volume and timescale of experiments and within the framework of the previously-published microscopic model [A.M. Kierzek, W.M. Wolf, P. Zielenkiewicz, Biophys. J. 73 (1997) 571-580]. Simulations for the tetragonal lysozyme crystal show that there are two critical sizes in the development of ordered phase. The first one corresponds to the size of the smallest stable complex which, in the case of the tetragonal lysozyme crystal, is the particular tetramer. In a volume of 5 mul the tetramer appears in the millisecond timescale. The second critical radius of approximately 100 monomers is only reached by a few of all the smallest stable complexes formed in the solution. The model predicts that out of 10(7) tetramers which appear in solution, only eight reach the size of 100 monomers within 8 h. After exceeding the second critical radius the microcrystals grow to the size of 10(4) monomers in the minute timescale and are thus assumed to quickly lead to macroscopic crystals. The predicted number of crystals formed during 8 h of nucleation is in qualitative agreement with arrested nucleation experiments.

Oligomeric structure of the human EphB2 receptor SAM domain

The sterile alpha motif (SAM) domain is a protein interaction module that is present in diverse signal-transducing proteins. SAM domains are known to form homo- and hetero-oligomers. The crystal structure of the SAM domain from an Eph receptor tyrosine kinase, EphB2, reveals two large interfaces. In one interface, adjacent monomers exchange amino-terminal peptides that insert into a hydrophobic groove on each neighbor. A second interface is composed of the carboxyl-terminal helix and a nearby loop. A possible oligomer, constructed from a combination of these binding modes, may provide a platform for the formation of larger protein complexes.

A soft, mean-field potential derived from crystal contacts for predicting protein-protein interactions

We derive a series of novel mean-field potentials from statistical analyses of protein-protein contact regions in crystal structures. These potentials are parameterized in terms of the number of contacts made by an atom in an interface region. Such an explicit number dependence avoids the pairwise assumption and is intrinsically softer than distance-based approaches. It appears well suited to protein-protein docking applications, for which detailed interface geometry is generally lacking. In tests including protein complex reconstitution and docking of independently determined protein structures, we show that a hydrophobic potential of this type performs remarkably well, identifying native-like complexes by their favourable potential energies and in several cases demonstrating a recognition energy gap of 4-8 kcal/mol according to the system.

Molecular origins of osmotic second virial coefficients of proteins

The thermodynamic properties of protein solutions are determined by the molecular interactions involving both solvent and solute molecules. A quantitative understanding of the relationship would facilitate more systematic procedures for manipulating the properties in a process environment. In this work the molecular basis for the osmotic second virial coefficient, B22, is studied; osmotic effects are critical in membrane transport, and the value of B22 has also been shown to correlate with protein crystallization behavior. The calculations here account for steric, electrostatic, and short-range interactions, with the structural and functional anisotropy of the protein molecules explicitly accounted for. The orientational dependence of the protein interactions is seen to have a pronounced effect on the calculations; in particular, the relatively few protein-protein configurations in which the apposing surfaces display geometric complementarity contribute disproportionately strongly to B22. The importance of electrostatic interactions is also amplified in these high-complementarity configurations. The significance of molecular recognition in determining B22 can explain the correlation with crystallization behavior, and it suggests that alteration of local molecular geometry can help in manipulating protein solution behavior. The results also have implications for the role of protein interactions in biological self-organization.

Mode of binding of anti-P-glycoprotein antibody MRK-16 to its antigen. A crystallographic and molecular modeling study

Monoclonal antibody MRK-16 recognizes a discontinuous extracellular epitope on the multidrug resistance-associated ATP-binding cassette transporter, P-glycoprotein. The atomic basis for specificity of this antibody is of interest because of its potential as a modulator of P-glycoprotein activity. The crystal structure of Fab MRK-16 is reported to a resolution of 2.8 A. A structure for a portion of the epitope was derived by comparison to regions of solved structures with similar primary sequence. This has permitted a proposal for the mode of binding of the peptide epitope to the antibody, in which the peptide makes specific contacts with complementarity-determining regions H1, H2, and H3 from the heavy chain and L3 from the light chain. These interactions are consistent with epitope mapping studies and with the observation that MRK-16 is specific for human class I P-glycoprotein. This result identifies side chains in MRK-16 that would be amenable to alteration in antibody engineering experiments to derive improved multidrug resistance inhibitors for clinical use during chemotherapy. In particular, Arg-H97 contacts both Glu-746 and Asp-744 of the peptide, Arg-L96 contacts Asp-743, and Thr-H33 interacts with Thr-747. All of these epitope residues were implicated in mediating specificity by epitope mapping studies.

A structural comparison of the colicin immunity proteins Im7 and Im9 gives new insights into the molecular determinants of immunity-protein specificity

We report the first detailed comparison of two immunity proteins which, in conjunction with recent protein engineering data, begins to explain how these structurally similar proteins are able to bind and inhibit the endonuclease domain of colicin E9 (E9 DNase) with affinities that differ by 12 orders of magnitude. In the present work, we have determined the X-ray structure of the Escherichia coli colicin E7 immunity protein Im7 to 2.0 A resolution by molecular replacement, using as a trial model the recently determined NMR solution structure of Im9. Whereas the two proteins adopt similar four-helix structures, subtle structural differences, in particular involving a conserved tyrosine residue critical for E9 DNase binding, and the identity of key residues in the specificity helix, lie at the heart of their markedly different ability to bind the E9 DNase. Two other crystal structures were reported recently for Im7; in one, Im7 was a monomer and was very similar to the structure reported here, whereas in the other it was a dimer to which functional significance was assigned. Since this previous work suggested that Im7 could exist either as a monomer or a dimer, we used analytical ultracentrifugation to investigate this question further. Under a variety of solution conditions, we found that Im7 only ever exists in solution as a monomer, even up to protein concentrations of 15 mg/ml, casting doubt on the functional significance of the crystallographically observed dimer. This work provides a structural framework with which we can understand immunity-protein specificity, and in addition we believe it to be the first successfully refined crystal structure solved by molecular replacement using an NMR trial model with less than 100% sequence identity.

Adaptation of an enzyme to regulatory function: structure of Bacillus subtilis PyrR, a pyr RNA-binding attenuation protein and uracil phosphoribosyltransferase

BACKGROUND

The expression of pyrimidine nucleotide biosynthetic (pyr) genes in Bacillus subtilis is regulated by transcriptional attenuation. The PyrR attenuation protein binds to specific sites in pyr mRNA, allowing the formation of downstream terminator structures. UMP and 5-phosphoribosyl-1-pyrophosphate (PRPP), a nucleotide metabolite, are co-regulators with PyrR. The smallest RNA shown to bind tightly to PyrR is a 28-30 nucleotide stem-loop that contains a purine-rich bulge and a putative-GNRA tetraloop. PyrR is also a uracil phosphoribosyltransferase (UPRTase), although the relationship between enzymatic activity and RNA recognition is unclear, and the UPRTase activity of PyrR is not physiologically significant in B. subtilis. Elucidating the role of PyrR structural motifs in UMP-dependent RNA binding is an important step towards understanding the mechanism of pyr transcriptional attenuation.

RESULTS

The 1.6 A crystal structure of B. subtilis PyrR has been determined by multiwavelength anomalous diffraction, using a Sm co-crystal. As expected, the structure of PyrR is homologous to those proteins of the large type I PRTase structural family; it is most similar to hypoxanthine-guanine-xanthine PRTase (HGXPRTase). The PyrR dimer differs from other PRTase dimers, suggesting it may have evolved specifically for RNA binding. A large, basic, surface at the dimer interface is an obvious RNA-binding site and uracil specificity is probably provided by hydrogen bonds from mainchain and sidechain atoms in the hood subdomain. These models of RNA and UMP binding are consistent with biological data.

CONCLUSIONS

The B. subtilis protein PyrR has adapted the substrate- and product-binding capacities of a PRTase, probably an HGXPRTase, producing a new regulatory function in which the substrate and product are co-regulators of transcription termination. The structure is consistent with the idea that PyrR regulatory function is independent of catalytic activity, which is likely to be extremely low under physiological conditions.

Molecular basis of sulfite oxidase deficiency from the structure of sulfite oxidase

The molybdenum-containing enzyme sulfite oxidase catalyzes the conversion of sulfite to sulfate, the terminal step in the oxidative degradation of cysteine and methionine. Deficiency of this enzyme in humans usually leads to major neurological abnormalities and early death. The crystal structure of chicken liver sulfite oxidase at 1.9 A resolution reveals that each monomer of the dimeric enzyme consists of three domains. At the active site, the Mo is penta-coordinated by three sulfur ligands, one oxo group, and one water/hydroxo. A sulfate molecule adjacent to the Mo identifies the substrate binding pocket. Four variants associated with sulfite oxidase deficiency have been identified: two mutations are near the sulfate binding site, while the other mutations occur within the domain mediating dimerization.

Structure, catalysis and supramolecular assembly of adenylate kinase from maize

The crystal structure of adenylate kinase from maize ligated with an inhibitor has been determined by molecular replacement and refined to 3.5-A resolution. The enzyme keeps the ATP/ADP/AMP equilibrium in the cell. In the C4 plant maize, it has the special task to recycle the AMP produced in large amounts in primary CO2 assimilation. The established structure explains the side reaction with CMP. Moreover, it shows infinite rods that can be readily discerned in the crystal packing. In comparison with homologues, two structural differences that are crucial for this supramolecular assembly are evident. We propose that the rods represent a natural inactive storage form that assembles at night when maize stops CO2 assimilation and thus most of the AMP production in its C4 cycle. The enzyme is particularly abundant in mesophyll chloroplasts, where such an assembly would release appreciable amounts of water that can be used in other processes during the night.

The crystal structure of HIV-1 Nef protein bound to the Fyn kinase SH3 domain suggests a role for this complex in altered T cell receptor signaling

BACKGROUND

Human immunodeficiency virus (HIV) Nef protein accelerates virulent progression of acquired immunodeficiency syndrome (AIDS) by its interaction with specific cellular proteins involved in signal transduction and host cell activation. Nef has been shown to bind specifically to a subset of the Src family of kinases. The structures of free Nef and Nef bound to Src homology region 3 (SH3) domain are important for the elucidation of how the affinity and specificity for the Src kinase family SH3 domains are achieved, and also for the development of potential drugs and vaccines against AIDS.

RESULTS

We have determined the crystal structures of the conserved core of HIV-1 Nef protein alone and in complex with the wild-type SH3 domain of the p59fyn protein tyrosine kinase (Fyn), at 3.0 A resolution. Comparison of the bound and unbound Nef structures revealed that a proline-rich motif (Pro-x-x-Pro), which is implicated in SH3 binding, is partially disordered in the absence of the binding partner; this motif only fully adopts a left-handed polyproline type II helix conformation upon complex formation with the Fyn SH3 domain. In addition, the structures show how an arginine residue (Arg77) of Nef interacts with Asp 100 of the so-called RT loop within the Fyn SH3 domain, and triggers a hydrogen-bond rearrangement which allows the loop to adapt to complement the Nef surface. The Arg96 residue of the Fyn SH3 domain is specifically accommodated in the same hydrophobic pocket of Nef as the isoleucine residue of a previously described Fyn SH3 (Arg96-->lle) mutant that binds to Nef with higher affinity than the wild type.

CONCLUSIONS

The three-dimensional structures support evidence that the Nef-Fyn complex forms in vivo and may have a crucial role in the T cell perturbating action of Nef by altering T cell receptor signaling. The structures of bound and unbound Nef reveal that the multivalency of SH3 binding may be achieved by a ligand induced flexibility in the RT loop. The structures suggest possible targets for the design of inhibitors which specifically block Nef-SH3 interactions.

Protein-protein crystal-packing contacts

Protein-protein contacts in monomeric protein crystal structures have been analyzed and compared to the physiological protein-protein contacts in oligomerization. A number of features differentiate the crystal-packing contacts from the natural contacts occurring in multimeric proteins. The area of the protein surface patches involved in packing contacts is generally smaller and its amino acid composition is indistinguishable from that of the protein surface accessible to the solvent. The fraction of protein surface in crystal contacts is very variable and independent of the number of packing contacts. The thermal motion at the crystal packing interface and that of the protein core, even for large packing interfaces, though the tendency is to be closer to that of the core. These results suggest that protein crystallization depends on random protein-protein interactions, which have little in common with physiological protein-protein recognition processes, and that the possibility of engineering macromolecular crystallization to improve crystal quality could be widened.

Structural motifs at protein-protein interfaces: protein cores versus two-state and three-state model complexes

The general similarity in the forces governing protein folding and protein-protein associations has led us to examine the similarity in the architectural motifs between the interfaces and the monomers. We have carried out extensive, all-against-all structural comparisons between the single-chain protein structural dataset and the interface dataset, derived both from all protein-protein complexes in the structural database and from interfaces generated via an automated crystal symmetry operation. We show that despite the absence of chain connections, the global features of the architectural motifs, present in monomers, recur in the interfaces, a reflection of the limited set of the folding patterns. However, although similarity has been observed, the details of the architectural motifs vary. In particular, the extent of the similarity correlates with the consideration of how the interface has been formed. Interfaces derived from two-state model complexes, where the chains fold cooperatively, display a considerable similarity to architectures in protein cores, as judged by the quality of their geometric superposition. On the other hand, the three-state model interfaces, representing binding of already folded molecules, manifest a larger variability and resemble the monomer architecture only in general outline. The origin of the difference between the monomers and the three-state model interfaces can be understood in terms of the different nature of the folding and the binding that are involved. Whereas in the former all degrees of freedom are available to the backbone to maximize favorable interactions, in rigid body, three-state model binding, only six degrees of freedom are allowed. Hence, residue or atom pair-wise potentials derived from protein-protein associations are expected to be less accurate, substantially increasing the number of computationally acceptable alternate binding modes (Finkelstein et al., 1995).

Simulations of nucleation and early growth stages of protein crystals

Analysis of known protein crystal structures reveals that interaction energies between monomer pairs alone are not sufficient to overcome entropy loss related to fixing monomers in the crystal lattice. Interactions with several neighbors in the crystal are required for stabilization of monomers in the lattice. A microscopic model of nucleation and early growth stages of protein crystals, based on the above observations, is presented. Anisotropy of protein molecules is taken into account by assigning free energies of association (proportional to the buried surface area) to individual monomer-monomer contacts in the lattice. Lattice simulations of the tetragonal lysozyme crystal based on the model correctly reproduce structural features of the movement of dislocation on the (110) crystal face. The dislocation shifts with the speed equal to the one determined experimentally if the geometric probability of correct orientation is set to 10(-5), in agreement with previously published estimates. At this value of orientational probability, the first nuclei, the critical size of which for lysozyme is four monomers, appear in 1 ml of supersaturated solution on a time scale of microseconds. Formation of the ordered phase proceeds through the growth of nuclei (rather then their association) and requires nucleations on the surface at certain stages.

Extent and nature of contacts between protein molecules in crystal lattices and between subunits of protein oligomers

A survey was compiled of several characteristics of the intersubunit contacts in 58 oligomeric proteins, and of the intermolecular contracts in the lattice for 223 protein crystal structures. The total number of atoms in contact and the secondary structure elements involved are similar in the two types of interfaces. Crystal contact patches are frequently smaller than patches involved in oligomer interfaces. Crystal contacts result from more numerous interactions by polar residues, compared with a tendency toward nonpolar amino acids at oligomer interfaces. Arginine is the only amino acid prominent in both types of interfaces. Potentials of mean force for residue-residue contacts at both crystal and oligomer interfaces were derived from comparison of the number of observed residue-residue interactions with the number expected by mass action. They show that hydrophobic interactions at oligomer interfaces favor aromatic amino acids and methionine over aliphatic amino acids; and that crystal contacts form in such a way as to avoid inclusion of hydrophobic interactions. They also suggest that complex salt bridges with certain amino acid compositions might be important in oligomer formation. For a protein that is recalcitrant to crystallization, substitution of lysine residues with arginine or glutamine is a recommended strategy.

Structural basis for ligand-regulated oligomerization of AraC

The crystal structure of the arabinose-binding and dimerization domain of the Escherchia coli gene regulatory protein AraC was determined in the presence and absence of L-arabinose. The 1.5 angstrom structure of the arabinose-bound molecule shows that the protein adopts an unusual fold, binding sugar within a beta barrel and completely burying the arabinose with the amino-terminal arm of the protein. Dimer contacts in the presence of arabinose are mediated by an antiparallel coiled-coil. In the 2.8 angstrom structure of the uncomplexed protein, the amino-terminal arm is disordered, uncovering the sugar-binding pocket and allowing it to serve as an oligomerization interface. The ligand-gated oligomerization as seen in AraC provides the basis of a plausible mechanism for modulating the protein's DNA-looping properties.

Studies of protein-protein interfaces: a statistical analysis of the hydrophobic effect

Data sets of 362 structurally nonredundant protein-protein interfaces and of 57 symmetry-related oligomeric interfaces have been used to explore whether the hydrophobic effect that guides protein folding is also the main driving force for protein-protein associations. The buried nonpolar surface area has been used to measure the hydrophobic effect. Our analysis indicates that, although the hydrophobic effect plays a dominant role in protein-protein binding, it is not as strong as that observed in the interior of protein monomers. Comparison of interiors of the monomers with those of the interfaces reveals that, in general, the hydrophobic amino acids are more frequent in the interior of the monomers than in the interior of the protein-protein interfaces. On the other hand, a higher proportion of charged and polar residues are buried at the interfaces, suggesting that hydrogen bonds and ion pairs contribute more to the stability of protein binding than to that of protein folding. Moreover, comparison of the interior of the interfaces to protein surfaces indicates that the interfaces are poorer in polar/charged than the surfaces and are richer in hydrophobic residues. The interior of the interfaces appears to constitute a compromise between the stabilization contributed by the hydrophobic effect on the one hand and avoiding patches on the protein surfaces that are too hydrophobic on the other. Such patches would be unfavorable for the unassociated monomers in solution. We conclude that, although the types of interactions are similar between protein-protein interfaces and single-chain proteins overall, the contribution of the hydrophobic effect to protein-protein associations is not as strong as to protein folding. This implies that packing patterns and interatom, or interresidue, pairwise potential functions, derived from monomers, are not ideally suited to predicting and assessing ligand associations or design. These would perform adequately only in cases where the hydrophobic effect at the binding site is substantial.

Hydrophobic regions on protein surfaces. Derivation of the solvation energy from their area distribution in crystallographic protein structures

For the first time, a direct approach for the derivation of an atomic solvation parameter from macromolecular structural data alone is presented. The specific free energy of solvation for hydrophobic surface regions of proteins is delineated from the area distribution of hydrophobic surface patches. The resulting value is 18 cal/(mol.A2), with a statistical uncertainty of +/-2 cal/mol.A2) at the 5% significance level. It compares favorably with the parameters for carbon obtained by other authors who use the the crystal geometry of succinic acid or energies of transfer from hydrophobic solvent to water for small organic compounds. Thus, the transferability of atomic solvation parameters for hydrophobic atoms to macromolecules has been directly demonstrated. A careful statistical analysis demonstrates that surface energy parameters derived from thermodynamic data of protein mutation experiments are clearly less confident.

Principles of protein-protein recognition from structure to thermodynamics

Specific recognition is illustrated by X-ray structures of protease-inhibitor, antigen-antibody and other high affinity complexes including five electron transfer complexes. We attempt to give a physical definition to affinity and specificity on the basis of these data. In a protein-protein complex, specific recognition results from the assembly of complementary surfaces into well-packed interfaces that cover about 1500 A2 and contain about ten hydrogen bonds. These interfaces are larger than between molecules in protein crystals, and smaller than between subunits in oligomeric proteins. We relate the size and chemical nature of interfaces in complexes to the thermodynamical parameters that characterize affinity: the heat capacity and free enthalpy (Gibbs energy) of dissociation at equilibrium, the activation free enthalpy for the dissociation reaction. The same structural and thermodynamical parameters are inadequate for representing the specificity of recognition. We propose instead to describe specificity with the help of statistical physics, and we illustrate the application of the random energy model to antigen-antibody recognition by analyzing results of computer simulations by docking.