Inferring protein function by domain context similarities in protein-protein interaction networks

Background  

Genome sequencing projects generate massive amounts of sequence data but there are still many proteins whose functions remain unknown. The availability of large scale protein-protein interaction data sets makes it possible to develop new function prediction methods based on protein-protein interaction (PPI) networks. Although several existing methods combine multiple information resources, there is no study that integrates protein domain information and PPI networks to predict protein functions.     

Functional coherence in domain interaction networks

MOTIVATION: Extracting functional information from protein-protein interactions (PPI) poses significant challenges arising from the noisy, incomplete, generic and static nature of data obtained from high-throughput screening. Typical proteins are composed of multiple domains, often regarded as their primary functional and structural units. Motivated by these considerations, domain-domain interactions (DDI) for network-based analyses have received significant recent attention. This article performs a formal comparative investigation of the relationship between functional coherence and topological proximity in PPI and DDI networks. Our investigation provides the necessary basis for continued and focused investigation of DDIs as abstractions for functional characterization and modularization of networks. RESULTS: We investigate the problem of assessing the functional coherence of two biomolecules (or segments thereof) in a formal framework. We establish essential attributes of admissible measures of functional coherence, and demonstrate that existing, well-accepted measures are ill-suited to comparative analyses involving different entities (i.e. domains versus proteins). We propose a statistically motivated functional similarity measure that takes into account functional specificity as well as the distribution of functional attributes across entity groups to assess functional similarity in a statistically meaningful and biologically interpretable manner. Results on diverse data, including high-throughput and computationally predicted PPIs, as well as structural and computationally inferred DDIs for different organisms show that: (i) the relationship between functional similarity and network proximity is captured in a much more (biologically) intuitive manner by our measure, compared to existing measures and (ii) network proximity and functional similarity are significantly more correlated in DDI networks than in PPI networks, and that structurally determined DDIs provide better functional relevance as compared to computationally inferred DDIs.     

The Cartographers toolbox: building bigger and better human protein interaction networks

One of the greatest challenges of the post-genomic era is theconstruction of a more comprehensive human protein interactionmap. While this process may take many years to complete, thedevelopment of stringent high throughput techniques and theemergence of complementary assays mean that the aim of buildinga detailed binary map of the human interactome is now a veryrealistic goal. In particular, methods which facilitate theanalysis of large numbers of membrane-protein interactions meanthat it will be possible to construct more extensive networks,which in turn provide new insights into the functional connectivitybetween intra- and extra-cellular processes. This is importantas many therapeutic strategies are designed to elicit effectsvia ‘tractable’ cell-surface proteins. Therefore,the construction of maps depicting the complexity of trans-cellularcommunication networks will not only improve our understandingof physiological processes, it will also aid the design of rationaltherapeutic strategies, with fewer potential side effects. Thisreview aims to provide a basic insight into the approaches currentlybeing used to construct binary human protein interaction networks,with particular reference to newer techniques, which have thepotential to extend network coverage and aid the conditionalannotation of interactome-scale protein interaction maps.      

Cytoprophet: a Cytoscape plug-in for protein and domain interaction networks inference

Cytoprophet is a software tool that allows prediction and visualization of protein and domain interaction networks. It is implemented as a plug-in of Cytoscape, an open source software framework for analysis and visualization of molecular networks. Cytoprophet implements three algorithms that predict new potential physical interactions using the domain composition of proteins and experimental assays. The algorithms for protein and domain interaction inference include maximum likelihood estimation (MLE) using expectation maximization (EM); the set cover approach maximum specificity set cover (MSSC) and the sum-product algorithm (SPA). After accepting an input set of proteins with Uniprot ID/Accession numbers and a selected prediction algorithm, Cytoprophet draws a network of potential interactions with probability scores and GO distances as edge attributes. A network of domain interactions between the domains of the initial protein list can also be generated. Cytoprophet was designed to take advantage of the visual capabilities of Cytoscape and be simple to use. An example of inference in a signaling network of myxobacterium Myxococcus xanthus is presented and available at Cytoprophet's website. AVAILABILITY: http://cytoprophet.cse.nd.edu.     

Biosynthesis of the macrolide antibiotic chlorothricin: basic building blocks

The biosynthesis of chlorothricin (I), a macrolide antibiotic isolated from Streptomyces antibioticus Tü 99, has been studied by feeding experiments with 14C- and 3H-labeled precursors. Acetate and propionate, but not methionine and mevalonate, were incorporated into the macrocylic aglycone of the antibiotic. Glucose and the various carbon atoms of tyrosine, except the carboxyl carbon, also contributed label to the aglycone. Glucose also seems to be a specific precursor of the 2-deoxyrhamnose moiety, probably via a process involving a hydrogen shift from C-4 to C-6 of the hexose. The substituted 6-methylsalicylic acid moiety seems to be derived from acetate and one O-methyl group provided by methionine; shikimic acid is not incorporated.     

Kinetic and structural study of the interaction of myelin basic protein with dipalmitoylphosphatidylglycerol layers

The interaction of myelin basic protein (MBP) with dipalmitoylphosphatidylglycerol films has been investigated by means of a microgravimetric gauge sensitive to the changes in load and structural modifications of the layer deposited onto its surface. Fourier transform infrared spectroscopy, circular dichroism, and x-ray diffraction have confirmed protein uptake by the lipid phase along with a global disordering effect onto the lipid alkyl chains and have shown a temporal evolution of the structure of water penetrating the lipid phase together with the protein. These effects are clearly related to the temporal variation of the microgravimetric gauge signal. Finally, measurements carried out on pre-annealed samples point out the role of mesoscopic morphology in determining the pathways through which MBP penetrates the lipid multilayer. The results obtained in our model system could be useful in clarifying the mechanisms of the myelinating and demyelinating processes that take place in the natural membrane.     

Lipid and basic protein interaction in myelin

1. Purified myelin labelled with [(3)H]myo-inositol or [1-(14)C]acetate was incubated with trypsin or acetylated trypsin at 37 degrees C, pH8.0 for 30min. 2. After incubation and centrifugation analysis of the myelin pellet showed marked digestion of basic protein on polyacrylamide-gel electrophoresis. Proteolipid and Wolfgram proteins remained unchanged. 3. A loss of 15% of total protein and loss of all classes of lipids was also found. Most significant lipid losses were phosphoinositides, phosphatidylserine and sulphatide. 4. A low-density material containing more phospholipid than cholesterol and galactolipid was isolated from the supernatant obtained after centrifugation of trypsin-treated myelin. 5. Interaction of sulphatide and myelin basic protein was shown to take place in a biphasic system. Basic protein does not form any complex either with cerebroside or cholesterol in the same solvent system. 6. The release of acidic lipids from myelin suggests that they may be linked to basic protein by ionic forces and the neutral lipids may be by lipid-lipid interactions. 7. The relevance of these studies as a model of brain degeneration is discussed.     

Engineering building blocks for self-assembling protein nanoparticles

Like natural viruses, manmade protein cages for drug delivery are to be ideally formed by repetitive subunits with self-assembling properties, mimicking viral functions and molecular organization. Naturally formed nanostructures (such as viruses, flagella or simpler protein oligomers) can be engineered to acquire specific traits of interest in biomedicine, for instance through the addition of cell targeting agents for desired biodistribution and specific delivery of associated drugs. However, fully artificial constructs would be highly desirable regarding finest tuning and adaptation to precise therapeutic purposes. Although engineering of protein assembling is still in its infancy, arising principles and promising strategies of protein manipulation point out the rational construction of nanoscale protein cages as a feasible concept, reachable through conventional recombinant DNA technologies and microbial protein production.     

Scale-free behavior in protein domain networks

Several technical, social, and biological networks were recently found to demonstrate scale-free and small-world behavior instead of random graph characteristics. In this work, the topology of protein domain networks generated with data from the ProDom, Pfam, and Prosite domain databases was studied. It was found that these networks exhibited small-world and scale-free topologies with a high degree of local clustering accompanied by a few long-distance connections. Moreover, these observations apply not only to the complete databases, but also to the domain distributions in proteomes of different organisms. The extent of connectivity among domains reflects the evolutionary complexity of the organisms considered.     

Mechanical unfoldons as building blocks of maltose-binding protein

Identifying independently folding cores or substructures is important for understanding and assaying the structure, function and assembly of large proteins. Here, we suggest mechanical stability as a criterion to identify building blocks of the 366 amino acid maltose-binding protein (MBP). We find that MBP, when pulled at its termini, unfolds via three (meta-) stable unfolding intermediates. Consequently, the MBP structure consists of four structural blocks (unfoldons) that detach sequentially from the folded structure upon force application. We used cysteine cross-link mutations to characterize the four unfoldons structurally. We showed that many MBP constructs composed of those building blocks indeed form stably folded structures in solution. Mechanical unfoldons may provide a new tool for a systematic search for stable substructures of large proteins.     

Principles of nanostructure design with protein building blocks

Currently there is increasing interest in nanostructures and their design. Nanostructure design involves the ability to predictably manipulate the properties of the self-assembly of autonomous units. Autonomous units have preferred conformational states. The units can be synthetic material science-based or derived from functional biological macromolecules. Autonomous biological building blocks with available structures provide an extremely rich and useful resource for design. For proteins, the structural databases contain large libraries of protein molecules and their building blocks with a range of shapes, surfaces, and chemical properties. The introduction of engineered synthetic residues or short peptides into these can expand the available chemical space and enhance the desired properties. Here we focus on the principles of nanostructure design with protein building blocks.     

RNA structural motifs: building blocks of a modular biomolecule

RNAs are modular biomolecules, composed largely of conserved structural subunits, or motifs. These structural motifs comprise the secondary structure of RNA and are knit together via tertiary interactions into a compact, functional, three-dimensional structure and are to be distinguished from motifs defined by sequence or function. A relatively small number of structural motifs are found repeatedly in RNA hairpin and internal loops, and are observed to be composed of a limited number of common 'structural elements'. In addition to secondary and tertiary structure motifs, there are functional motifs specific for certain biological roles and binding motifs that serve to complex metals or other ligands. Research is continuing into the identification and classification of RNA structural motifs and is being initiated to predict motifs from sequence, to trace their phylogenetic relationships and to use them as building blocks in RNA engineering.     

Antagonism and bistability in protein interaction networks

A protein interaction network (PIN) is a set of proteins that modulate one another's activities by regulated synthesis and degradation, by reversible binding to form complexes, and by catalytic reactions (e.g., phosphorylation and dephosphorylation). Most PINs are so complex that their dynamical characteristics cannot be deduced accurately by intuitive reasoning alone. To predict the properties of such networks, many research groups have turned to mathematical models (differential equations based on standard biochemical rate laws, e.g., mass-action, Michaelis-Menten, Hill). When using Michaelis-Menten rate expressions to model PINs, care must be exercised to avoid making inconsistent assumptions about enzyme-substrate complexes. We show that an appealingly simple model of a PIN that functions as a bistable switch is compromised by neglecting enzyme-substrate intermediates. When the neglected intermediates are put back into the model, bistability of the switch is lost. The theory of chemical reaction networks predicts that bistability can be recovered by adding specific reaction channels to the molecular mechanism. We explore two very different routes to recover bistability. In both cases, we show how to convert the original 'phenomenological' model into a consistent set of mass-action rate laws that retains the desired bistability properties. Once an equivalent model is formulated in terms of elementary chemical reactions, it can be simulated accurately either by deterministic differential equations or by Gillespie's stochastic simulation algorithm.     

Shaping dots and lines: adding modularity into protein interaction networks using structural information

Determining protein interaction networks and generating models to simulate network changes in time and space are crucial for understanding a biological system and for predicting the effect of mutants found in diseases. In this review we discuss the great potential of using structural information together with computational tools towards reaching this goal: the prediction of new protein interactions, the estimation of affinities and kinetic rate constants between protein complexes, and finally the determination of which interactions are compatible with each other and which interactions are exclusive. The latter one will be important to reorganize large scale networks into functional modular networks.     

Comparative interactomics analysis of protein family interaction networks using PSIMAP (protein structural interactome map)

MOTIVATION: Many genomes have been completely sequenced. However, detecting and analyzing their protein-protein interactions by experimental methods such as co-immunoprecipitation, tandem affinity purification and Y2H is not as fast as genome sequencing. Therefore, a computational prediction method based on the known protein structural interactions will be useful to analyze large-scale protein-protein interaction rules within and among complete genomes. RESULTS: We confirmed that all the predicted protein family interactomes (the full set of protein family interactions within a proteome) of 146 species are scale-free networks, and they share a small core network comprising 36 protein families related to indispensable cellular functions. We found two fundamental differences among prokaryotic and eukaryotic interactomes: (1) eukarya had significantly more hub families than archaea and bacteria and (2) certain special hub families determined the topology of the eukaryotic interactomes. Our comparative analysis suggests that a very small number of expansive protein families led to the evolution of interactomes and seemed to have played a key role in species diversification. SUPPLEMENTARY INFORMATION: http://interactomics.org.     

Patterns,structures, and amino acid frequencies in structural building blocks,a protein secondary structure classification scheme

To study local structures in proteins, we previously developed an autoassociative artificial neural network (autoANN) and clustering tool to discover intrinsic features of macromolecular structures. The hidden unit activations computed by the trained autoANN are a convenient low-dimensional encoding of the local protein backbone structure. Clustering these activation vectors results in a unique classification of protein local structural features called Structural Building Blocks (SBBs). Here we describe application of this method to a larger database of proteins, verification of the applicability of this method to structure classification, and subsequent analysis of amino acid frequencies and several commonly occurring patterns of SBBs. The SBB classification method has several interesting properties: 1) it identifies the regular secondary structures, α helix and β strand; 2) it consistently identifies other local structure features (e.g., helix caps and strand caps); 3) strong amino acid preferences are revealed at some positions in some SBBs; and 4) distinct patterns of SBBs occur in the “random coil” regions of proteins. Analysis of these patterns identifies interesting structural motifs in the protein backbone structure, indicating that SBBs can be used as “building blocks” in the analysis of protein structure. This type of pattern analysis should increase our understanding of the relationship between protein sequence and local structure, especially in the prediction of protein structures. © 1997 Wiley-Liss, Inc.     

In silico protein design by combinatorial assembly of protein building blocks

Utilizing concepts of protein building blocks, we propose a de novo computational algorithm that is similar to combinatorial shuffling experiments. Our goal is to engineer new naturally occurring folds with low homology to existing proteins. A selected protein is first partitioned into its building blocks based on their compactness, degree of isolation from the rest of the structure, and hydrophobicity. Next, the protein building blocks are substituted by fragments taken from other proteins with overall low sequence identity, but with a similar hydrophobic/hydrophilic pattern and a high structural similarity. These criteria ensure that the designed protein has a similar fold, low sequence identity, and a good hydrophobic core compared with its native counterpart. Here, we have selected two proteins for engineering, protein G B1 domain and ubiquitin. The two engineered proteins share approximately 20% and approximately 25% amino acid sequence identities with their native counterparts, respectively. The stabilities of the engineered proteins are tested by explicit water molecular dynamics simulations. The algorithm implements a strategy of designing a protein using relatively stable fragments, with a high population time. Here, we have selected the fragments by searching for local minima along the polypeptide chain using the protein building block model. Such an approach provides a new method for engineering new proteins with similar folds and low homology.     

Greedily building protein networks with confidence

MOTIVATION: With genome sequences complete for human and model organisms, it is essential to understand how individual genes and proteins are organized into biological networks. Much of the organization is revealed by proteomics experiments that now generate torrents of data. Extracting relevant complexes and pathways from high-throughput proteomics data sets has posed a challenge, however, and new methods to identify and extract networks are essential. We focus on the problem of building pathways starting from known proteins of interest. RESULTS: We have developed an efficient, greedy algorithm, SEEDY, that extracts biologically relevant biological networks from protein-protein interaction data, building out from selected seed proteins. The algorithm relies on our previous study establishing statistical confidence levels for interactions generated by two-hybrid screens and inferred from mass spectrometric identification of protein complexes. We demonstrate the ability to extract known yeast complexes from high-throughput protein interaction data with a tunable parameter that governs the trade-off between sensitivity and selectivity. DNA damage repair pathways are presented as a detailed example. We highlight the ability to join heterogeneous data sets, in this case protein-protein interactions and genetic interactions, and the appearance of cross-talk between pathways caused by re-use of shared components. SIGNIFICANCE AND COMPARISON: The significance of the SEEDY algorithm is that it is fast, running time O[(E + V) log V] for V proteins and E interactions, a single adjustable parameter controls the size of the pathways that are generated, and an associated P-value indicates the statistical confidence that the pathways are enriched for proteins with a coherent function. Previous approaches have focused on extracting sub-networks by identifying motifs enriched in known biological networks. SEEDY provides the complementary ability to perform a directed search based on proteins of interest. AVAILABILITY: SEEDY software (Perl source), data tables and confidence score models (R source) are freely available from the author.