Structural similarity enhances interaction propensity of proteins

We study statistical properties of interacting protein-like surfaces and predict two strong, related effects: (i) statistically enhanced self-attraction of proteins; (ii) statistically enhanced attraction of proteins with similar structures. The effects originate in the fact that the probability to find a pattern self-match between two identical, even randomly organized interacting protein surfaces is always higher compared with the probability for a pattern match between two different, promiscuous protein surfaces. This theoretical finding explains statistical prevalence of homodimers in protein-protein interaction networks reported earlier. Further, our findings are confirmed by the analysis of curated database of protein complexes that showed highly statistically significant overrepresentation of dimers formed by structurally similar proteins with highly divergent sequences ("superfamily heterodimers"). We suggest that promiscuous homodimeric interactions pose strong competitive interactions for heterodimers evolved from homodimers. Such evolutionary bottleneck is overcome using the negative design evolutionary pressure applied against promiscuous homodimer formation. This is achieved through the formation of highly specific contacts formed by charged residues as demonstrated both in model and real superfamily heterodimers.     

Structural similarity and functional diversity in diiron-oxo proteins

 Diiron-oxo proteins currently represent one of the most rapidly developing areas of bioinorganic chemistry. All of these proteins contain a four-helix bundle protein fold surrounding a (μ-carboxylato)diiron core, and most, if not all, of the diiron(II) sites appear to react with O₂ as part of their functional processes. Despite these common characteristics, an emerging functional diversity is one of the most striking aspects of this class of proteins. X-ray crystal structures of diiron(II) sites are now available for four of these proteins: hemerythrin (Hr), the hydroxylase protein of methane monooxygenase (MMOH), the R2 protein of Escherichia coli ribonucleotide reductase (RNR-R2), and a plant acyl-carrier protein Δ⁹-desaturase. The structure of the diiron(II) site in Hr, the sole O₂ carrier in the group, is clearly distinct from the other three, whose function is oxygen activation. The Hr diiron site is more histidine rich, and the oxygen-activating diiron sites contain a pair of (D/E)X_30–37EX₂H ligand sequence motifs, which is clearly not found in Hr. The Hr diiron site apparently permits only terminal O₂ coordination to a single iron, whereas the oxygen-activating diiron(II) centers present open or labile coordination sites on both irons of the center, and show a much greater coordinative flexibility upon oxidation to the diiron(III) state. Intermediates at the formal Fe^IIIFe^III and Fe^IVFe^IV oxidation levels for MMOH and formal Fe^IIIFe^IV oxidation level for RNR-R2 have been identified during reactions of the diiron(II) sites with O₂. An [Fe₂(μ-O)₂]^4+, 3+ "diamond core" structure has been proposed for the latter two oxidation levels. The intermediate at the Fe^IIIFe^IV oxidation level in RNR-R2 is kinetically competent to generate a stable, functionally essential tyrosyl radical. The Fe^IVFe^IV oxidation level is presumed to effect hydroxylation of hydrocarbons in MMOH, but the mechanism of this hydroxylation, particularly the involvement of discrete radicals, is currently controversial. The biological function of diiron sites in three members of this class, rubrerythrin, ferritin and bacterioferritin, remains enigmatic. Received: 31 July 1996 / Accepted: 4 October 1996     

Structural similarity between legumin and vicilin storage proteins from legumes.

The primary structures for several members of both the vicilin and legumin families of storage proteins were examined using a computer routine based on amino acid physical characteristics. The comparison algorithm revealed that sequences from the two families could be aligned and share a number of predicted secondary structural features. The COOH-terminal half of the subunits in both families displayed a highly conserved core region that was largely hydrophobic and in which a high proportion of the residues were predicted to be in beta-sheet conformations. The central region of the molecules which contained mixed areas of predicted helical and sheet conformations showed more variability in residue selection than the COOH-terminal regions. The NH2-terminal segments of subunits from the two different families could not be aligned though they characteristically had a high proportion of residues predicted to be in helical conformations. The feature which most clearly distinguished subunits between the two families was an inserted span in the legumin group with a high proportion of acidic amino acids located between the central and COOH-terminal domains. Residues in this insertion were predicted to exist mainly in helical conformation. Since considerable size variation occurs in this area amongst the legumin subunits, alterations in this region may have a minimal detrimental effect on the structure of the proteins.     

Structural basis for LEAFY floral switch function and similarity with helix-turn-helix proteins

The LEAFY (LFY) protein is a key regulator of flower development in angiosperms. Its gradually increased expression governs the sharp floral transition, and LFY subsequently controls the patterning of flower meristems by inducing the expression of floral homeotic genes. Despite a wealth of genetic data, how LFY functions at the molecular level is poorly understood. Here, we report crystal structures for the DNA-binding domain of Arabidopsis thaliana LFY bound to two target promoter elements. LFY adopts a novel seven-helix fold that binds DNA as a cooperative dimer, forming base-specific contacts in both the major and minor grooves. Cooperativity is mediated by two basic residues and plausibly accounts for LFY's effectiveness in triggering sharp developmental transitions. Our structure reveals an unexpected similarity between LFY and helix-turn-helix proteins, including homeodomain proteins known to regulate morphogenesis in higher eukaryotes. The appearance of flowering plants has been linked to the molecular evolution of LFY. Our study provides a unique framework to elucidate the molecular mechanisms underlying floral development and the evolutionary history of flowering plants.     

Structural and functional similarity between fish antifreeze proteins and calcium-dependent lectins.

A cDNA for a type II antifreeze protein was isolated from liver of smelt (Osmerus mordax). The predicted protein sequence is homologous to that from sea raven (Hemitripterus americanus) and both show homology to a family of calcium-dependent lectins. Smelt and sea raven belong to taxonomic orders believed to have diverged prior to Cenozoic glaciation. Thus, type II antifreeze proteins appear to have evolved independently in these fish species from pre-existing calcium-dependent lectins. Sequence alignment of the antifreezes and the lectins suggest that these proteins adopt a similar fold, that the sea raven antifreeze has lost its Ca2+ binding sites, and the smelt antifreeze has retained one site. Experiments show that smelt antifreeze protein activity is responsive to Ca2+ but that of sea raven antifreeze protein is not. These results suggest that the type II fish antifreeze proteins and calcium-dependent lectins share a common ancestry, related folding structures, and functional similarity.     

Structural similarity and functional diversity in proteins containing the legume lectin fold.

Knowledge of structural relationships in proteins is increasingly proving very useful for in silico characterizations and is also being exploited as a prelude to almost every investigation in functional and structural genomics. A thorough understanding of the crucial features of a fold becomes necessary to realize the full potential of such relationships. To illustrate this, structures containing the legume lectin-like fold were chosen for a detailed analysis since they exhibit a total lack of sequence similarity among themselves and also belong to diverse functional families. A comparative analysis of 15 different families containing this fold was therefore carried out, which led to the determination of the minimal structural principles or the determining region of the fold. A critical evaluation of the structural features, such as the curvature of the front sheet, the presence of the hydrophobic cores and the binding site loops, suggests that none of them are crucial for either the formation or the stability of the fold, but are required to generate diversity and specificity to particular carbohydrates. In contrast, the presence of the three sheets in a particular geometry and also their topological connectivities seem to be important. The fold has been shown to tolerate different types of protein-protein associations, most of them exhibiting different types of quaternary associations and some even existing as complexes with other folds. The function of every family in this study is discussed with respect to its fold, leading to the suggestion that this fold can be linked to carbohydrate recognition in general.     

Structure collisions between interacting proteins

Protein-protein interactions take place at defined binding interfaces. One protein may bind two or more proteins at different interfaces at the same time. So far it has been commonly accepted that non-overlapping interfaces allow a given protein to bind other proteins simultaneously while no collisions occur between the binding protein structures. To test this assumption, we performed a comprehensive analysis of structural protein interactions to detect potential collisions. Our results did not indicate cases of biologically relevant collisions in the Protein Data Bank of protein structures. However, we discovered a number of collisions that originate from alternative protein conformations or quaternary structures due to different experimental conditions.     

Inverse sequence similarity of proteins does not imply structural similarity

There is a debate on the folding of proteins with inverted sequences. Theoretical approaches and experiments give contradictory results. Many proteins in the Protein Data Bank (PDB) show conspicuous inverse sequence similarity (ISS) to each other. Here we analyze whether this ISS is related to structural similarity. For the first time, we performed a large scale three-dimensional (3-D) superposition of corresponding Calpha atoms of forwardly and inversely aligned proteins and tested the degree of secondary structure identity between them. Comparing proteins of less than 50% pairwise sequence identity, only 0.5% of the inversely aligned pairs had similar folds (99 out of 19073), whereas about 9% of forwardly aligned proteins in the same score and length range show similar 3-D structures (1731 out of 19248). This observation strongly supports the view that the inversion of sequences in almost all cases leads to a different folding property of the protein. Inverted sequences are thus suitable as protein-like sequences for control purposes without relations to existing proteins.     

Identification of connexin-43 interacting proteins

Connexin-43 (Cx43), the most ubiquitously expressed vertebrate gap junction protein, has been shown to interact directly with Zonula Occludens-1 (ZO-1). Although several potential functions have been proposed for the ZO-1/Cx43 interaction, the role that ZO-1 and other Cx43-interacting partners play in the regulation of Cx43 trafficking, assembly, gating and turnover are not well understood. We believed a thorough analysis and classification of other Cx43-interacting proteins might help us to understand and better test these roles. We approached this question by utilizing Tandem Mass Spectrometry (MS/MS) analysis to identify proteins from normal rat kidney whole cell lysates that could interact with the C-terminal region of Cx43. Comparison against protein sequence databases identified 19 probable protein matches, including kinases, phosphatases, membrane receptors, cell signaling molecules and scaffolding proteins. We have further characterized some of these interacting proteins, including Zonula Occludens-2 (ZO-2), via western blotting and "pull down" experiments. Further in vitro/in vivo analysis of these interacting proteins will help in our understanding of the global role of connexins in regulating development, cell metabolism and growth.     

Proteomic analysis of nucleoporin interacting proteins

The Saccharomyces cerevisiae nuclear pore complex is a supramolecular assembly of 30 nucleoporins that cooperatively facilitate nucleocytoplasmic transport. Thirteen nucleoporins that contain FG peptide repeats (FG Nups) are proposed to function as stepping stones in karyopherin-mediated transport pathways. Here, protein interactions that occur at individual FG Nups were sampled using immobilized nucleoporins and yeast extracts. We find that many proteins bind to FG Nups in highly reproducible patterns. Among 135 proteins identified by mass spectrometry, most were karyopherins and nucleoporins. The PSFG nucleoporin Nup42p and the GLFG nucleoporins Nup49p, Nup57p, Nup100p, and Nup116p exhibited generic interactions with karyopherins; each bound 6--10 different karyopherin betas, including importins as well as exportins. Unexpectedly, the same Nups also captured the hexameric Nup84p complex and Nup2p. In contrast, the FXFG nucleoporins Nup1p, Nup2p, and Nup60p were more selective and captured mostly the Kap95p.Kap60p heterodimer. When the concentration of Gsp1p-GTP was elevated in the extracts to mimic the nucleoplasmic environment, the patterns of interacting proteins changed; exportins exhibited enhanced binding to FG Nups, and importins exhibited reduced binding. The results demonstrate a global role for Gsp1p-GTP on karyopherin-nucleoporin interactions and provide a rudimentary map of the routes that karyopherins take as they cross the nuclear pore complex.     

Visualization of structural similarity in proteins

Two new methods for the visualization of structural similarity in proteins with known three-dimensional structures are presented. They are based on the degree of equivalence of α-carbon pairs in two proteins. The quantitative measure for residue equivalence is the comparison score generated using the sequence and structure alignment method of Taylor and Orengo, which is based on the comparison of interatomic distances (and other properties that can be defined on a residue basis).The first method uses information on corresponding α-carbon positions to display vectors joining these structurally equivalent residues. These vectors can be defined as target constraints, and their minimization “bends” the two proteins toward a common average structure. In the average structure the corresponding residues virtually superpose, while insertions and deletions become clearly visible.The second method uses the comparison scores to perform a weighted least-squares fit of the two structures. It is further used to color code the two structures according to the score value, i.e., their similarity, on a continuous scale from red to blue. Examples of the methods for the comparison of flavodoxin, chemotaxis Y protein and L-arabinose-binding protein are given.     

Identification of FUSE-binding proteins as interacting partners of TIA proteins

TIA-1 and TIAR are closely related RNA-binding proteins involved in several mechanisms of RNA metabolism, including alternative hnRNA splicing and mRNA translation regulation. In particular, TIA-1 represses tumor necrosis factor (TNF) mRNA translation by binding to the AU-rich element (ARE) present in the mRNA 3' untranslated region. Here, we demonstrate that TIA proteins interact with FUSE-binding proteins (FBPs) and that fbp genes are co-expressed with tia genes during Xenopus embryogenesis. FBPs participate in various steps of RNA processing and degradation. In Cos cells, FBPs co-localize with TIA proteins in the nucleus and migrate into TIA-enriched cytoplasmic granules upon oxidative stress. Overexpression of FBP2-KH3 RNA-binding domain fused to EGFP induces the specific sequestration of TIA proteins in cytoplasmic foci, thereby precluding their nuclear accumulation. In cytosolic RAW 264.7 macrophage extracts, FBPs are found associated in EMSA to the TIA-1/TNF-ARE complex. Together, our results indicate that TIA and FBP proteins may thus be relevant biological involved in common events of RNA metabolism occurring both in the nucleus and the cytoplasm.     

Insulin-like growth factor and interacting proteins

Insulin-like growth factor - IGF-I is a small, 70 aminoacid mitogenic peptide, contributing to processes of growing, cancerogenesis, apoptosis, wound healing and many others. It constitutes so called 'somatotropic axis GH-IGF', composed of many other components. This axis is responsible for regulation of metabolic processes, and its proper functioning conditions organism's homeostasis. Presented work describes concise review of publications concerning IGF-I structure, function, expression and proteins affecting its activity, synthesis and circulation.     

Microtubule dynamics and tubulin interacting proteins

Microtubule dynamics are crucial in generation of the mitotic spindle. During the transition from interphase to mitosis, there is an increase in the frequency of microtubule catastrophes. Recent work has identified two proteins, Op 18/stathmin and XKCM1, which can promote microtubule catastrophes in vitro and in cells or extracts. Although both of these proteins share the ability to bind tubulin dimers, their mechanisms of action in destabilizing microtubules are distinct.     

Beta-adrenergic receptors and their interacting proteins

The three subtypes of beta-adrenergic receptor (beta AR) all interact with G proteins as a central aspect of their signaling. The various beta AR subtypes also associate differentially with a variety of other cytoplasmic and transmembrane proteins. These beta AR-interacting proteins play distinct roles in the regulation of receptor signaling and trafficking. The specificity of beta AR associations with various binding partners can help to explain key physiological differences between beta AR subtypes. Moreover, the differential tissue expression patterns of many of the beta AR-interacting proteins may contribute to tissue-specific regulation of beta AR function.     

Analyzing phosphoinositides and their interacting proteins

Phosphorylated derivatives of phosphatidylinositol (PtdIns), known as phosphoinositides (PIs), are essential regulators of nuclear functions, cytoskeletal dynamics, cell signaling and membrane trafficking. These lipids are found on the cytosolic face of intracellular membranes but can also be detected in membrane-free regions of the nucleoplasm. Their downstream effectors include several proteins that contain various PI-specific domains. Because impaired PI metabolism is associated with disorders such as cancer, cardiovascular disease and immune dysfunction, there is currently great interest in studying PIs and their metabolic enzymes. Here we describe strategies and techniques for quantitative and qualitative measurement of PIs, for characterization of specific PI-binding proteins and for determination of PI kinase and phosphatase activities in vitro and in vivo.