Mapping protein-protein interactions in solution by NMR spectroscopy.

NMR is very well suited to the study of especially weak protein-protein interactions, as no crystallization is required. The available NMR methods to this end are reviewed and illustrated with applications from the recent biochemical literature: intermolecular NOEs, cross-saturation, chemical shift perturbation, dynamics and exchange perturbation, paramagnetic methods, and dipolar orientation. Most of these methods are now routinely applied for complexes with total molecular mass of 60 kDa and can likely be applied to systems up to 1000 kDa. A substantial fraction of complexes studied show distinct effects of induced fit affecting structural and dynamical properties beyond the contact interface.     

Utilization of Methyl Proton Resonances in Cross-Saturation Measurement for Determining the Interfaces of Large Protein–Protein Complexes

Cross-saturation experiments allow the identification of the contact residues of large protein complexes (MW>50 K) more rigorously than conventional NMR approaches which involve chemical shift perturbations and hydrogen-deuterium exchange experiments [Takahashi et al. (2000) Nat. Struct. Biol., 7, 220–223]. In the amide proton-based cross-saturation experiment, the combined use of high deuteration levels for non-exchangeable protons of the ligand protein and a solvent with a low concentration of ¹H₂O greatly enhanced the selectivity of the intermolecular cross-saturation phenomenon. Unfortunately, experimental limitations caused losses in sensitivity. Furthermore, since main chain amide protons are not generally exposed to solvent, the efficiency of the saturation transfer directed to the main chain amide protons is not very high. Here we propose an alternative cross-saturation experiment which utilizes the methyl protons of the side chains of the ligand protein. Owing to the fast internal rotation along the methyl axis, we theoretically and experimentally demonstrated the enhanced efficiency of this approach. The methyl-utilizing cross-saturation experiment has clear advantages in sensitivity and saturation transfer efficiency over the amide proton-based approach. Electronic supplementary material Electronic supplementary material is available for this article at and accessible for authorised users.     

Direct determination of a membrane-peptide interface using the nuclear magnetic resonance cross-saturation method

Membrane-peptide interactions are involved in many crucial biological and pharmacological activities. To clarify the interaction mode of membrane-peptide complexes, it is important to analyze both the dynamic properties and the contact residues of the membrane-bound peptide. In this study, we investigated the dynamic properties of a peptide bound to a lipid bilayer, using relaxation and amide-water exchange analyses, and directly determined the membrane-peptide interface, using the cross-saturation method. For the models of a lipid bilayer and a peptide, isotropic bicelles and mastoparan were used, respectively. The results indicate that mastoparan had a heterogeneous distribution of motion over various timescales and interacted with the lipid bilayer by using its hydrophobic side; the molecule was located within the lipid bilayer rather than on the surface, as thought previously. This study shows that the cross-saturation method is useful for determining the interface of not only protein-protein but also membrane-peptide complexes.     

The human cytomegalovirus receptor on fibroblasts is a 30-kilodalton membrane protein.

Previous studies have demonstrated that human cytomegalovirus (HCMV) binding to human foreskin fibroblasts (HFF) is mediated by a single type of molecule, likely a glycoprotein, which serves as a specific receptor for the virus. In the present experiments, HCMV was found to bind to an HFF membrane protein with an approximate molecular mass of 30 kilodaltons (kDa); weak binding to 28- and 92-kDa membrane components was also observed. Binding was specific, as it was inhibited by excess unlabeled HCMV. Radiolabeled HCMV also bound selectively to Raji and Daudi lymphoblastoid cell membrane proteins of the same molecular masses. The 30-kDa radiolabeled HFF membrane protein bound to HCMV in solution; this binding was also specific, as it was blocked by an excess of HCMV. These data suggest that a membrane protein with a molecular mass of approximately 30 kDa mediates HCMV binding to several cell types.     

Secondary structure mapping of DnaK-bound protein fragments: chain helicity and local helix unwinding at the binding site

Little is known about polypeptide conformation and folding in the presence of molecular chaperones participating in protein biosynthesis. In vitro studies on chaperone-substrate complexes have been mostly carried out with small peptide ligands. However, the technical challenges associated with either competing aggregation or spectroscopically unfavorable size and exchange rates have typically prevented analysis of larger substrates. Here, we report the high-resolution secondary structure of relatively large N-terminal protein fragments bound to the substrate-binding domain of the cotranslationally active chaperone DnaK. The all-alpha-helical protein apomyoglobin (apoMb), bearing the ubiquitous globin fold, has been chosen as a model substrate. On the basis of NMR secondary chemical shift analysis, we identify, for the first time, weak helical content (similar to that found in the chemically unfolded full-length protein) for the assigned residues of the chaperone-bound chain away from the chaperone binding sites. In contrast, we found that the residues corresponding to the strongest specific binding site for DnaK, examined via a short 13-mer apoMb peptide fragment matching the binding site sequence, display highly reduced helical content in their chaperone-bound form. Given that the free state of the peptide is weakly helical in isolation, we conclude that the substrate residues corresponding to the chaperone binding site undergo helix unwinding upon chaperone binding.     

Hyaluronan recognition mode of CD44 revealed by cross-saturation and chemical shift perturbation experiments

CD44 is the main cell surface receptor for hyaluronic acid (HA) and contains a functional HA-binding domain (HABD) composed of a Link module with N- and C-terminal extensions. The contact residues of human CD44 HABD for HA have been determined by cross-saturation experiments and mapped on the topology of CD44 HABD, which we elucidated by NMR. The contact residues are distributed in both the consensus fold for the Link module superfamily and the additional structural elements consisting of the flanking regions. Interestingly, the contact residues exhibit small changes in chemical shift upon HA binding. In contrast, the residues with large chemical shift changes are localized in the C-terminal extension and the first alpha-helix and are generally inconsistent with the contact residues. These results suggest that, upon ligand binding, the C-terminal extension and the first alpha-helix undergo significant conformational changes, which may account for the broad ligand specificity of CD44 HABD.     

Collagen-binding mode of vWF-A3 domain determined by a transferred cross-saturation experiment

The nature of the supramolecular complex between fibrillar collagen and collagen-binding proteins (CBPs) has hindered detailed X-ray and NMR analyses of the ligand-recognition mechanism at atomic resolution because of the lack of appropriate approaches for studying large heterogeneous supramolecular complexes. Recently, we proposed an NMR method, termed transferred cross-saturation (TCS), that enables the rigorous identification of contact residues in a huge protein complex. Here we used TCS to study the supramolecular complex between the A3 domain of von Willebrand factor and fibrillar collagen, which allowed the successful determination of the ligand-binding site of the A3 domain. The binding site of the A3 domain was located at its hydrophobic 'front' surface and was completely different from that of the I domain from the a2 subunit of integrin (alpha2-I domain), which was reported to be the hydrophilic 'top' surface of alpha2-I, although the A3 domain and the alpha2-I domain share a similar fold and possess the identical function of collagen binding.     

Identification and characterization of an IgG binding protein in the secretion of the rat coagulating gland

A merocrine released protein (named 115k protein) was highly enriched from the secretion of the rat coagulating gland. The protein has a molecular mass of 115 kDa as calculated by SDS-PAGE under reducing conditions. Furthermore, the 115 kDa protein is glycosylated, and carries Man, GlcNAc, Gal, Fuc and sialic acid residues. For identification, N-terminal amino acid and nucleotide sequence analyses were performed. The sequences obtained showed 86 to 100% identity with human and mouse IgGFc binding proteins. The functional capacity of IgG binding of the 115 kDa protein was shown by overlay experiments, indicating its membership in the IgG binding protein family.     

NMR mapping and secondary structure determination of the major acetylcholine receptor alpha-subunit determinant interacting with alpha-bungarotoxin

The alpha-subunit of the nicotinic acetylcholine receptor (alphaAChR) contains a binding site for alpha-bungarotoxin (alpha-BTX), a snake-venom-derived alpha-neurotoxin. Previous studies have established that the segment comprising residues 173-204 of alphaAChR contains the major determinant interacting with the toxin, but the precise boundaries of this determinant have not been clearly defined to date. In this study, we applied NMR dynamic filtering to determine the exact sequence constituting the major alphaAChR determinant interacting with alpha-BTX. Two overlapping synthetic peptides corresponding to segments 179-200 and 182-202 of the alphaAChR were complexed with alpha-BTX. HOHAHA and ROESY spectra of these complexes acquired with long mixing times highlight the residues of the peptide that do not interact with the toxin and retain considerable mobility upon binding to alpha-BTX. These results, together with changes in the chemical shifts of the peptide protons upon complex formation, suggest that residues 184-200 form the contact region. At pH 4, the molecular mass of the complex determined by dynamic light scattering (DLS) was found to be 11.2 kDa, in excellent agreement with the expected molecular mass of a 1:1 complex, while at pH >5 the DLS measurement of 20 kDa molecular mass indicated dimerization of the complex. These results were supported by T(2) measurements. Complete resonance assignment of the 11.2 kDa complex of alpha-BTX bound to the alphaAChR peptide comprising residues 182-202 was obtained at pH 4 using homonuclear 2D NMR spectra measured at 800 MHz. The secondary structures of both alpha-BTX and the bound alphaAChR peptide were determined using 2D (1)H NMR experiments. The peptide folds into a beta-hairpin conformation, in which residues (R)H186-(R)V188 and (R)Y198-(R)D200 form the two beta-strands. Residues (R)Y189-(R)T191 form an intermolecular beta-sheet with residues (B)K38-(B)V40 of the second finger of alpha-BTX. These results accurately pinpoint the alpha-BTX-binding site on the alphaAChR and pave the way to structure determination of this important alphaAChR determinant involved in binding acetylcholine and cholinergic agonists and antagonists.     

Specific interaction of mouse major satellite with MAR-binding protein SAF-A

A DNA-binding activity specific to the major mouse satellite (satMa) has been detected in a nuclear matrix protein extract by electrophoretic mobility shift assays (EMSA) after fractionation by ion exchange chromatography. An antibody raised against the satMa-protein complexes recovered from preparative EMSA recognizes on Western blots one major polypeptide with an apparent molecular mass of 120 kDa. The protein also has a similar affinity for a matrix-associated region (MAR) fragment. We demonstrate that the protein is a murine homologue of SAF-A which has been shown to bind selectively to MARs and is responsible for the satMa-binding activity in the chromatographic fractions. SatMa has significant homology to the mouse minor satellite fragments, but its binding of SAF-A shows much less affinity. No protected regions of significant length were found by footprinting, but multiple T residues scattered within the satMa sequence are protected, indicating that the whole fragment is involved in the binding to SAF-A. Combined immunofluorescence (SAF-A) and FISH (satMa) with in situ nuclear matrix procedures reveal that SAF-A and satMa colocalize. SAF-A appears as bright dots in interphase nuclei, presumably associated with MARs, predominantly surrounding and covering heterochromatic areas. A scheme based on morphological observations and biochemical data of SAF-A double satMa/MAR specificity is discussed.     

A dominant sulfhydryl-containing protein in the outer membrane of Neisseria gonorrhoeae.

By using a method that labels sulfhydryl-containing proteins in situ, we have detected a major outer membrane protein of Neisseria gonorrhoeae at 41 kDa. A protein of this molecular mass has not previously been shown to be a major outer membrane protein in gonococci. In addition, a minor protein rich in cysteinyl residues was detected at 31.5 kDa.     

Analysis of membrane protein complexes by blue native PAGE

Blue native polyacryamide gel electrophoresis is a special case of native electrophoresis for high resolution separation of enzymatically active protein complexes from tissue homogenates and cell fractions. The method is powerful between 10 and 10,000 kDa. Also membrane protein complexes are separated well after solubilization of complexes with mild neutral detergents. The separation principle relies on binding of Coomassie blue G250 which provides negative charges to the surface of the protein. During migration to the anode, protein complexes are separated according to molecular mass and/or size and high resolution is obtained by the decreasing pore size of a polyacrylamide gradient gel. The principles of 2-dimensional blue native sodium dodecyl sulfate polyacrylamide gel electrophoresis are presented here together with a practical step-by-step guide to performing the method in the laboratory.     

Isolation and characterization of a transferrin binding protein from rat plasma

A transferrin binding protein was isolated from normal rat placenta and from iron-deficient rat plasma using a human transferrin affinity column. The yield of the isolated pure protein from iron-deficient rat plasma was about 0.5 micrograms/ml plasma. The major protein had a molecular mass of 85 kDa and contained carbohydrate. Reduction with mercaptoethanol did not change the molecular mass of the plasma transferrin binding protein whereas the native placental transferrin receptor of 180 kDa was reduced to 90 kDa. The transferrin binding protein reacted with both monoclonal and polyclonal antibodies raised against rat transferrin receptor. Immunoblotting of both normal and iron deficient rat plasma showed that the transferrin binding protein had a molecular mass of 85 kDa. In vitro digestion of purified rat placental transferrin receptor and red blood cells with trypsin provided an identical peptide profile, suggesting that the transferrin binding protein in rat plasma is derived from proteolysis of the extracellular portion of the transferrin receptor of the erythroid tissues.     

Induced fit with replica exchange improves protein complex structure prediction

Despite the progress in prediction of protein complexes over the last decade, recent blind protein complex structure prediction challenges revealed limited success rates (less than 20% models with DockQ score > 0.4) on targets that exhibit significant conformational change upon binding. To overcome limitations in capturing backbone motions, we developed a new, aggressive sampling method that incorporates temperature replica exchange Monte Carlo (T-REMC) and conformational sampling techniques within docking protocols in Rosetta. Our method, ReplicaDock 2.0, mimics induced-fit mechanism of protein binding to sample backbone motions across putative interface residues on-the-fly, thereby recapitulating binding-partner induced conformational changes. Furthermore, ReplicaDock 2.0 clocks in at 150-500 CPU hours per target (protein-size dependent); a runtime that is significantly faster than Molecular Dynamics based approaches. For a benchmark set of 88 proteins with moderate to high flexibility (unbound-to-bound iRMSD over 1.2 Å), ReplicaDock 2.0 successfully docks 61% of moderately flexible complexes and 35% of highly flexible complexes. Additionally, we demonstrate that by biasing backbone sampling particularly towards residues comprising flexible loops or hinge domains, highly flexible targets can be predicted to under 2 Å accuracy. This indicates that additional gains are possible when mobile protein segments are known.     

Determining binding sites in protein–nucleic acid complexes by cross-saturation

Cross-saturation experiments have been shown to give accurate information regarding the interacting surfaces in protein-protein and protein-RNA complexes. The rate of magnetization transfer depends on a number of factors including geometry, spin topology, and effective correlation times. To assess the influence of these variables on such experiments, and to determine the range of applicability of the technique, we have simulated the time-course of magnetization transfer across the interface in a variety of protein-nucleic acid complexes (434 Cro, SRY, MetJ and U1A), each having a different binding geometry. The simulations have been carried out primarily to provide information about the experimentally accessible targets for selective saturation, such as the anomeric protons and the imino protons of the nucleic acid. Saturation of either of these groups of signals leads to partial excitation throughout the nucleic acid molecule, and the resulting transfer of saturation to the labelled protein moiety can be readily detected by the reduction in intensity of particular peaks in the HSQC spectrum of the protein. The simulations show that information can be obtained about the residues in contact with the nucleic acid without any need for deuteration. Experimental cross-saturation data have been obtained from the Mbp1-DNA complex and interpreted in conjunction with the simulations to map out the binding surface in detail.     

The Geodia cydonium galectin exhibits prototype and chimera-type characteristics and a unique sequence polymorphism within its carbohydrate recognition domain

The ancestral galectin from the sponge Geodia cydonium (GCG) is classified on a structural basis to the prototype subfamily, whereas its carbohydrate-binding specificity is related to that of the mammalian chimera-type galectin-3. This dual coordination reveals GCG as a potential precursor of the later evolved galectin subfamilies, which is reflected in the primary structure of the protein. This study provides evidence that GCG is the LECT1 gene product, while neither a previously described LECT2 gene nor a functional LECT2 gene product was found in the specimen under investigation. The electrophoretically separated protein isomers with apparent molecular masses of 13, 15, and 16 kDa correspond to variants of the LECT1 protein-exhibiting peptide sequence polymorphisms that concern critical positions of the carbohydrate recognition domain (13 kDa: Leu51, Asn55, His130, Gly137; 15 kDa: Ser51, Asn55, Asn130, Gly137; 16 kDa: Ser51, Tyr55, Asn130, Glu137). Four residues, highly conserved in the galectin family, are substituted. None of the residues claimed to be involved in interactions with GalNAcalpha1-3 moieties at an extended binding subsite of galectin-3 was identified in the corresponding positions of GCG. Apparently, the substitutions do not confer distinct binding characteristics to the GCG variants as evidenced by binding studies with a recombinantly expressed 15-kDa isoform. The natural isoforms as well as the recombinant 15-kDa isoform oligomerize by the formation of non-covalent heteromeric or homomeric complexes. A phosphorylation of the galectin was confirmed neither by mass spectrometry nor by alkaline phosphatase treatment combined with isoelectric focusing.     

Identification of ligand binding sites of proteins using the Gaussian Network Model

The nonlocal nature of the protein-ligand binding problem is investigated via the Gaussian Network Model with which the residues lying along interaction pathways in a protein and the residues at the binding site are predicted. The predictions of the binding site residues are verified by using several benchmark systems where the topology of the unbound protein and the bound protein-ligand complex are known. Predictions are made on the unbound protein. Agreement of results with the bound complexes indicates that the information for binding resides in the unbound protein. Cliques that consist of three or more residues that are far apart along the primary structure but are in contact in the folded structure are shown to be important determinants of the binding problem. Comparison with known structures shows that the predictive capability of the method is significant.     

Structural basis for the carbohydrate specificities of artocarpin: variation in the length of a loop as a strategy for generating ligand specificity

Artocarpin, a tetrameric lectin of molecular mass 65 kDa, is one of the two lectins extracted from the seeds of jackfruit. The structures of the complexes of artocarpin with mannotriose and mannopentose reported here, together with the structures of artocarpin and its complex with Me-alpha-mannose reported earlier, show that the lectin possesses a deep-seated binding site formed by three loops. The binding site can be considered as composed of two subsites; the primary site and the secondary site. Interactions at the primary site composed of two of the loops involve mainly hydrogen bonds, while those at the secondary site comprising the third loop are primarily van der Waals in nature. Mannotriose in its complex with the lectin interacts through all the three mannopyranosyl residues; mannopentose interacts with the protein using at least three of the five mannose residues. The complexes provide a structural explanation for the carbohydrate specificities of artocarpin. A detailed comparison with the sugar complexes of heltuba, the only other mannose-specific jacalin-like lectin with known three-dimensional structure in sugar-bound form, establishes the role of the sugar-binding loop constituting the secondary site, in conferring different specificities at the oligosaccharide level. This loop is four residues longer in artocarpin than in heltuba, providing an instance where variation in loop length is used as a strategy for generating carbohydrate specificity.     

A New Method to Characterize Hydrophobic Organization of Proteins: Application to Rational Protein Engineering of Barnase

Abstract

We present a new algorithm for characterization of protein spatial structure basing on the molecular hydrophobicity potential approach. The method is illustrated by the analysis of three-dimensional structure of barnase and barnase-barstar complex. Current approach enables identification of amino acid residues situated in unfavorable environment (these residues may be “active” for binding), and to map quantitatively hydrophobic, hydrophilic and unfavorable hydrophobic-hydrophilic intra-and inter-molecular contacts involving backbone and side-chain segments of amino acid residues. Calculation of individual contributions of amino acid residues to such contacts permits identification of structurally-important residues. The contact plots obtained with molecular hydrophobicity potential calculations, provide easy rules to choose sites for mutations, which can increase a strength of intra- or inter-molecular hydrophobic interactions. The unfavorable hydrophobic-hydrophilic contact can be mutated to favorable hydrophobic, and already existing weak hydrophobic contact can be strengthen by increasing hydrophobicity of residues in contact. Basing on the analysis of the contact plots, we suggest several mutations of barnase which are supposed to increase intramolecular hydrophobic interactions, and thus might lead to increased stability of the protein. Part of these mutations was studied previously experimentally, and indeed stabilized barnase. The other of predicted mutations were not studied experimentally yet. Several new mutations of barnase and barstar are also proposed to enhance the hydrophobic interactions on their binding interface.     

Mapping protein receptor-ligand interactions via in vivo chemical crosslinking, affinity purification, and differential mass spectrometry

Protein receptor-ligand interactions play important roles in mediating enzyme catalysis, signal transduction, and other protein functions. Immunoaffinity purification followed by mass spectrometry analysis is a common method for identifying protein receptor-ligand complexes. However, it is difficult to distinguish between specific protein binding partners and non-specifically bound proteins that co-purify with the complex. In addition, weakly interacting binding partners may dissociate from the protein receptor-ligand complexes during immunoaffinity purification. The combination of chemical crosslinking, affinity purification, and differential mass spectrometry analysis provides a direct method for capturing stable, weak, and transient protein interactions that occur in vivo and in vitro. This approach enables the identification of functional receptor-ligand binding partners with high confidence. Herein, we describe a differential mass spectrometry approach coupled with in situ chemical crosslinking and immunoaffinity purification for identifying receptor-ligand binding partners. In particular, we identified a functional, counter-ligand structure of the natural killer cell p30-related protein.