Protein interaction methods--toward an endgame

Prediction of protein--protein interaction sites in heterocomplexes with neural networks

In this paper we address the problem of extracting features relevant for predicting protein--protein interaction sites from the three-dimensional structures of protein complexes. Our approach is based on information about evolutionary conservation and surface disposition. We implement a neural network based system, which uses a cross validation procedure and allows the correct detection of 73% of the residues involved in protein interactions in a selected database comprising 226 heterodimers. Our analysis confirms that the chemico-physical properties of interacting surfaces are difficult to distinguish from those of the whole protein surface. However neural networks trained with a reduced representation of the interacting patch and sequence profile are sufficient to generalize over the different features of the contact patches and to predict whether a residue in the protein surface is or is not in contact. By using a blind test, we report the prediction of the surface interacting sites of three structural components of the Dnak molecular chaperone system, and find close agreement with previously published experimental results. We propose that the predictor can significantly complement results from structural and functional proteomics.

Selection of v-abl tyrosine kinase substrate sequences from randomized peptide and cellular proteomic libraries using mRNA display

Methodologies for rapidly identifying cellular protein interactions resulting in posttranslational modification of one of the partners are lacking. Here, we select for substrates of the v-abl tyrosine kinase from two protein display libraries in which the protein is covalently linked to its encoding mRNA. Successive selection cycles from a randomized peptide library identified a consensus sequence closely matching that previously reported for the v-abl tyrosine kinase. Selections from a proteomic library derived from cellular mRNA identified several novel targets of v-abl, including a new member of a class of SH2 domain-containing adaptor proteins. Upon modification, several of the substrates obtained in these selections were found to be effective inhibitors of v-abl kinase activity in vitro. These experiments establish a novel method for identifying the substrates of tyrosine kinases from synthetic and cellular protein libraries.

The impact of microbial genomics on antimicrobial drug development

There is an urgent need to develop novel classes of antibiotics to counter the threat of the spread of multiply resistant bacterial pathogens. The availability of the complete genome sequence of many pathogenic microbes provides information on every potential drug target and is an invaluable resource in the search for novel compounds. Here, we review the approaches being taken to exploit the genome databases through a combination of bioinformatics, transcriptional analysis, and a further understanding of the molecular basis of the disease process. The emphasis is changing from compound screening to target hunting, as the latter offers flexible ways to design and optimize the next generation of broad-spectrum antibiotics.

Fruits of human genome project and private venture, and their impact on life science

A small knowledge base was created by organizing the Human Genome Project (HGP) and its related issues in "Science" magazines between 1996 and 2000. This base revealed the stunning achievement of HGP and a private venture and its impact on today's biology and life science. In the mid-1990, they encouraged the development of advanced high throughput automated DNA sequencers and the technologies that can analyse all genes at once in a systematic fashion. Using these technologies, they completed the genome sequence of human and various other organisms. These fruits opened the door to comparative genomics, functional genomics, the interdisprinary field between computer and biology, and proteomics. They have caused a shift in biological investigation from studying single genes or proteins to studying all genes or proteins at once, and causing revolutional changes in traditional biology, drug discovery and therapy. They have expanded the range of potential drug targets and have facilitated a shift in drug discovery programs toward rational target-based strategies. They have spawned pharmacogenomics that could give rise to a new generation of highly effective drugs that treat causes, not just symptoms. They should also cause a migration from the traditional medications that are safe and effective for every members of the population to personalized medicine and personalized therapy.

Genetic synthetic lethality screen at the single gene level in cultured human cells

Recently, we demonstrated the feasibility of a chemical synthetic lethality screen in cultured human cells. We now demonstrate the principles for a genetic synthetic lethality screen. The technology employs both an immortalized human cell line deficient in the gene of interest, which is complemented by an episomal survival plasmid expressing the wild-type cDNA for the gene of interest, and the use of a novel GFP-based double-label fluorescence system. Dominant negative genetic suppressor elements (GSEs) are selected from an episomal library expressing short truncated sense and antisense cDNAs for a gene likely to be synthetic lethal with the gene of interest. Expression of these GSEs prevents spontaneous loss of the GFP-marked episomal survival plasmid, thus allowing FACS enrichment for cells retaining the survival plasmid (and the GSEs). The dominant negative nature of the GSEs was validated by the decreased resident enzymatic activity present in cells harboring the GSEs. Also, cells mutated in the gene of interest exhibit reduced survival upon GSE expression. The identification of synthetic lethal genes described here can shed light on functional genetic interactions between genes involved in normal cell metabolism and in disease.

A novel yeast system for in vivo selection of recognition sequences: defining an optimal c-Myb-responsive element

Yeast (Saccharomyces cerevisiae) has proved to be a highly valuable tool in a range of screening methods. We present in this work the design and use of a novel yeast effector-reporter system for selection of sequences recognised by DNA-binding proteins in vivo. A dual HIS3-lacZ reporter under the control of a single randomised response element facilitates both positive growth selection of binding sequences and subsequent quantification of the strength of the selected sequence. A galactose-inducible effector allows discrimination between reporter activation caused by the protein under study and activation due to endogenous factors. The system mimics the physiological gene dosage relationship between transcription factor and target genes in vivo by using a low copy effector plasmid and a high copy reporter plasmid, favouring sequence selectivity. The utility of the novel yeast screening system was demonstrated by using it to refine the definition of an optimal recognition element for the c-Myb transcription factor (MRE). We present screening data supporting an extended MRE consensus closely mimicking known strong response elements and where a sequence of 11 nt influences activity. Novel features include a more strict sequence requirement in the second half-site of the MRE where a T-rich sequence is preferred in vivo.

Thermal decomposition of a gaseous multiprotein complex studied by blackbody infrared radiative dissociation. Investigating the origin of the asymmetric dissociation behavior

The blackbody infrared radiative dissociation technique was used to study the thermal decomposition of the gaseous B5 pentamer of the Shiga-like toxin I and its complexes with the Pk trisaccharide and a decavalent Pk-based oligosaccharide ligand (STARFISH, S). Dissociation of the protonated pentamer, (B5 + nH)n+ triple bond B5n+ where n = 11-14, proceeds almost exclusively by the loss of a single subunit (B) with a disproportionately large fraction (30-50%) of the parent ion charge. The degree of charge enrichment of the leaving subunit increases with increasing parent ion charge state. For n = 12-14, a distribution of product ion charge states is observed. The yields of the complementary pairs of product ions are sensitive to the reaction temperature, with higher temperatures favoring greater charge enrichment of the leaving subunit for +13 and +14, and the opposite effect for +12. These results indicate that some of the protons are rapidly exchanged between subunits in the gas phase. Dissociation of B5(14+) x S proceeds exclusively by the loss of one subunit, although the ligand increases the stability of the complex and also reduces the degree of charge enrichment in the ejected monomer. For B5(12+)(Pk)1-3, the loss of neutral Pk competes with loss of a subunit at low temperatures. Linear Arrhenius plots were obtained from the temperature-dependent dissociation rate constants measured for the loss of B from B5n+ and B514+ x S. The magnitude of the Arrhenius parameters is highly dependent on the charge state of the pentamer: Ea = 35 kcal/mol and A = 1,019 s(-1) (+14), 46 kcal/mol and 1,023 S(-1) (+13), 50 kcal/mol and 1026 s(-1) (+12), and 80 kcal/mol and 10(39) (+11). The Ea and A for B5(14+) x S are 59 kcal/mol and 10(30) s(-1), respectively. The reaction pathways leading to greater charge enrichment of the subunit lost from the B5(14+) and B5(13+) ions correspond to higher energy processes, however, these pathways are kinetically preferred at higher temperatures due to their large A factors. A simple electrostatic model, whereby charge enrichment leads to Coulombic repulsion-induced denaturation of the subunits and disruption of the intersubunit interactions, provides an explanation for the magnitude of the Arrhenius parameters and the origin of the asymmetric dissociation behavior of the complexes.

The use of FRET imaging microscopy to detect protein-protein interactions and protein conformational changes in vivo

Intermolecular and intramolecular FRET between two spectrally overlapping green fluorescent protein variants fused to two different host proteins or at two different sites within the same protein offers a unique opportunity to monitor real-time protein-protein interactions or protein conformational changes. By using fluorescence digital imaging microscopy, one can visualize the location of green fluorescent proteins within a living cell and follow the time course of the changes in FRET corresponding to cellular events at a millisecond time resolution. The observation of such dynamic molecular events in vivo provides vital insight into the action of biological molecules.

Similarity of phylogenetic trees as indicator of protein-protein interaction

Deciphering the network of protein interactions that underlines cellular operations has become one of the main tasks of proteomics and computational biology. Recently, a set of bioinformatics approaches has emerged for the prediction of possible interactions by combining sequence and genomic information. Even though the initial results are very promising, the current methods are still far from perfect. We propose here a new way of discovering possible protein-protein interactions based on the comparison of the evolutionary distances between the sequences of the associated protein families, an idea based on previous observations of correspondence between the phylogenetic trees of associated proteins in systems such as ligands and receptors. Here, we extend the approach to different test sets, including the statistical evaluation of their capacity to predict protein interactions. To demonstrate the possibilities of the system to perform large-scale predictions of interactions, we present the application to a collection of more than 67 000 pairs of E.coli proteins, of which 2742 are predicted to correspond to interacting proteins.

Single-molecule observation of protein-protein interactions in the chaperonin system

We have analyzed the dynamics of the chaperonin (GroEL)-cochaperonin (GroES) interaction at the single-molecule level. In the presence of ATP and non-native protein, binding of GroES to the immobilized GroEL occurred at a rate that is consistent with bulk kinetics measurements. However, the release of GroES from GroEL occurred after a lag period ( approximately 3 s) that was not recognized in earlier bulk-phase studies. This observation suggests a new kinetic intermediate in the GroEL-GroES reaction pathway.

Correlated sequence-signatures as markers of protein-protein interaction

As protein-protein interaction is intrinsic to most cellular processes, the ability to predict which proteins in the cell interact can aid significantly in identifying the function of newly discovered proteins, and in understanding the molecular networks they participate in. Here we demonstrate that characteristic pairs of sequence-signatures can be learned from a database of experimentally determined interacting proteins, where one protein contains the one sequence-signature and its interacting partner contains the other sequence-signature. The sequence-signatures that recur in concert in various pairs of interacting proteins are termed correlated sequence-signatures, and it is proposed that they can be used for predicting putative pairs of interacting partners in the cell. We demonstrate the potential of this approach on a comprehensive database of experimentally determined pairs of interacting proteins in the yeast Saccharomyces cerevisiae. The proteins in this database have been characterized by their sequence-signatures, as defined by the InterPro classification. A statistical analysis performed on all possible combinations of sequence-signature pairs has identified those pairs that are over-represented in the database of yeast interacting proteins. It is demonstrated how the use of the correlated sequence-signatures as identifiers of interacting proteins can reduce significantly the search space, and enable directed experimental interaction screens.

Rapid functional analysis of protein-protein interactions by fluorescent C-terminal labeling and single-molecule imaging

Detection of protein-protein interactions is a fundamental step to understanding gene function. Here we report a sensitive and rapid method for assaying protein-protein interactions at the single-molecule level. Protein molecules were synthesized in a cell-free translation system in the presence of Cy5-puro, a fluorescent puromycin, using mRNA without a stop codon. The interaction of proteins thus prepared was visualized using a single-molecule imaging technique. As a demonstration of this method, a motor protein, kinesin, was labeled with Cy5-puro at an efficiency of about 90%, and the processive movement of kinesin along microtubules was observed by using total internal reflection microscopy. It took only 2 h from the synthesis of proteins to the functional analysis. This method is applicable to the functional analysis of various kinds of proteins.

In vitro display technologies: novel developments and applications

In vitro display techniques are powerful tools to select polypeptide binders against various target molecules. Novel applications include maturation of protein affinity and stability, selection for enzymatic activity, and the display of cDNA and random polypeptide libraries. Taken together, these display techniques have great potential for biotechnological, medical and proteomic applications.

Characterization of the interaction between phospholipid and protein by capillary electrophoresis with laser-induced fluorescence detection

We report an electrophoretic mobility shift-based method to study the interactions between phospholipids and proteins by capillary electrophoresis with laser-induced fluorescence detection. A fluorogenic dye, 3-(2-furoyl)quinoline-2-carboxaldehyde (FQ), was used to label phosphatidylserine (PS). Then the FQ labeled PS (FQ-PS) was used as the fluorescent probe for monitoring the association between PS and bovine serum albumin (BSA). Two conjugates were observed to form between each PS species and BSA, indicating that two interactions exist between these PS species and BSA. We can also detect the competitive association with BSA between labeled PS and unlabeled PS. This method only needs a minute volume of sample. It is highly sensitive and can be used to detect the interaction between phospholipids and nanomolar concentrations of proteins, such as BSA.

Recent advances in FRET: distance determination in protein-DNA complexes

Fluorescence resonance energy transfer (FRET) provides information on the distance between a donor and an acceptor dye in the range 10 to 100 A. Knowledge of the exact positions of some dyes with respect to nucleic acids now enables us to translate these data into precise structural information using molecular modeling. Advances in the preparation of dye-labeled nucleic acid molecules and in new techniques, such as the measurement of FRET in polyacrylamide gels or in vivo, will lead to an increasingly important role of FRET in structural and molecular biology.

Towards a protein interaction map of potyviruses: protein interaction matrixes of two potyviruses based on the yeast two-hybrid system

A map for the interactions of the major proteins from Potato virus A (PVA) and Pea seed-borne mosaic virus (PSbMV) (members of the genus POTYVIRUS:, family POTYVIRIDAE:) was generated using the yeast two-hybrid system (YTHS). Interactions were readily detected with five PVA protein combinations (HC-HC, HC-CI, VPg-VPg, NIa-NIb and CP-CP) and weak but reproducible interactions were detected for seven additional combinations (P1-CI, P3-NIb, NIaPro-NIb, VPg-NIa, VPg-NIaPro, NIaPro-NIa and NIa-NIa). In PSbMV, readily detectable interactions were found in five protein combinations (HC-HC, VPg-VPg, VPg-NIa, NIa-NIa and NIa-NIb) and weaker but reproducible interactions were detected for three additional combinations (P3-NIa, NIa-NIaPro and CP-CP). The self-interactions of HC, VPg, NIa and CP and the interactions of VPg-NIa, NIa-NIaPro and NIa-NIb were, therefore, common for the two potyviruses. The multiple protein interactions revealed in this study shed light on the co-ordinated functions of potyviral proteins involved in virus movement and replication.

Identification and kinetic analysis of the interaction between Nck-2 and DOCK180

Nck-2 is a newly identified adapter protein comprising three N-terminal SH3 domains and one C-terminal SH2 domain. We have identified in a yeast two-hybrid screen DOCK180, a signaling protein implicated in the regulation of membrane ruffling and migration, as a binding protein for Nck-2. Surface plasmon resonance analyses reveal that the second and the third SH3 domains interact with the C-terminal region of DOCK180. The interactions mediated by the individual SH3 domains, however, are much weaker than that of the full length Nck-2. Furthermore, a point mutation that inactivates the second or the third SH3 domain dramatically reduced the interaction of Nck-2 with DOCK180, suggesting that both SH3 domains contribute to the DOCK180 binding. A major Nck-2 binding site, which is recognized primarily by the third SH3 domain, has been mapped to residues 1819-1836 of DOCK180. Two additional, albeit much weaker, Nck-2 SH3 binding sites are located to DOCK180 residues 1793-1810 and 1835-1852 respectively. Consistent with the mutational studies, kinetic analyses by surface plasmon resonance suggest that two binding events with equilibrium dissociation constants of 4.15+/-1.9x10(-7) M and 3.24+/-1.9x10(-9) M mediate the binding of GST-Nck-2 to GST fusion protein containing the C-terminal region of DOCK180. These studies identify a novel interaction between Nck-2 and DOCK180. Furthermore, they provide a detailed analysis of a protein complex formation mediated by multiple SH3 domains revealing that tandem SH3 domains significantly enhance the weak interactions mediated by each individual SH3 domain.

Inhibition of p53-dependent transcription by BOX-I phospho-peptide mimetics that bind to p300

The N-terminal BOX-I domain of p53 containing a docking site for the negative regulator MDM2 and the positive effector p300, harbours two recently identified phosphorylation sites at Thr18 or Ser20O whose affect on p300 is undefined. Biochemical assays demonstrate that although MDM2 binding is inhibited by these phosphorylations, p300 binding is strikingly stabilized by Thr18 or Ser20 phosphorylation. Introducing EGFP-BOX-I domain peptides with an aspartate substitution at Thr18 or Ser20 induced a significant inhibition of endogenous p53-dependent transcription in cycling cells, in irradiated cells, as well as in cells transiently co-transfected with p300 and p53. In contrast an EGFP-wild-type BOX-I domain peptide stimulated p53 activity via inhibition of MDM2 protein binding. These results suggest that phosphorylation of p53 at Thr18 or Ser20 can activate p53 by stabilizing the p300-p53 complex and also identify a class of small molecular weight ligands capable of selective discrimination between MDM2- and p300-dependent activities.

A novel approach for the identification of protein-protein interaction with integral membrane proteins

Protein-protein interaction plays a major role in all biological processes. The currently available genetic methods such as the two-hybrid system and the protein recruitment system are relatively limited in their ability to identify interactions with integral membrane proteins. Here we describe the development of a reverse Ras recruitment system (reverse RRS), in which the bait used encodes a membrane protein. The bait is expressed in its natural environment, the membrane, whereas the protein partner (the prey) is fused to a cytoplasmic Ras mutant. Protein-protein interaction between the proteins encoded by the prey and the bait results in Ras membrane translocation and activation of a viability pathway in yeast. We devised the expression of the bait and prey proteins under the control of dual distinct inducible promoters, thus enabling a rapid selection of transformants in which growth is attributed solely to specific protein-protein interaction. The reverse RRS approach greatly extends the usefulness of the protein recruitment systems and the use of integral membrane proteins as baits. The system serves as an attractive approach to explore novel protein-protein interactions with high specificity and selectivity, where other methods fail.

Stability engineering of antibody single-chain Fv fragments

The application of single-chain Fv fragments (scFv) in medicine and biotechnology places great demands on their stability. Only recently has attention been given to the production of highly stable scFvs, and in a number of examples it was found that such fragments indeed perform better during practical applications. The structural parameters influencing scFv stability are now beginning to be elucidated. This review summarizes progress in rational and evolutionary engineering methods, the structural implications of these results, as well as some examples where stability engineering has been successfully applied.

BIND--The Biomolecular Interaction Network Database

The Biomolecular Interaction Network Database (BIND; http://binddb. org) is a database designed to store full descriptions of interactions, molecular complexes and pathways. Development of the BIND 2.0 data model has led to the incorporation of virtually all components of molecular mechanisms including interactions between any two molecules composed of proteins, nucleic acids and small molecules. Chemical reactions, photochemical activation and conformational changes can also be described. Everything from small molecule biochemistry to signal transduction is abstracted in such a way that graph theory methods may be applied for data mining. The database can be used to study networks of interactions, to map pathways across taxonomic branches and to generate information for kinetic simulations. BIND anticipates the coming large influx of interaction information from high-throughput proteomics efforts including detailed information about post-translational modifications from mass spectrometry. Version 2.0 of the BIND data model is discussed as well as implementation, content and the open nature of the BIND project. The BIND data specification is available as ASN.1 and XML DTD.

Protein recruitment systems for the analysis of protein-protein interactions

Following the completion of genome projects in a number of organisms, it is becoming evident that a relatively large proportion of the genes identified encode for proteins that have no sequence homology with known proteins. One possible approach towards understanding protein function is to identify the proteins with which a particular protein associates. Although very powerful, the most commonly used genetic method, the two-hybrid system, is limited in its ability to detect all possible protein-protein interactions. The development of novel approaches, such as the protein recruitment systems, provides attractive alternatives towards identification of protein-protein interactions where other methods have failed to function.

Structural evidence for evolution of the beta/alpha barrel scaffold by gene duplication and fusion

The atomic structures of two proteins in the histidine biosynthesis pathway consist of beta/alpha barrels with a twofold repeat pattern. It is likely that these proteins evolved by twofold gene duplication and gene fusion from a common half-barrel ancestor. These ancestral domains are not visible as independent domains in the extant proteins but can be inferred from a combination of sequence and structural analysis. The detection of subdomain structures may be useful in efforts to search genome sequences for functionally and structurally related proteins.

Domain repertoires as a tool to derive protein recognition rules

Several approaches, some of which are described in this issue, have been proposed to assemble a complete protein interaction map. These are often based on high throughput methods that explore the ability of each gene product to bind any other element of the proteome of the organism. Here we propose that a large number of interactions can be inferred by revealing the rules underlying recognition specificity of a small number (a few hundreds) of families of protein recognition modules. This can be achieved through the construction and characterization of domain repertoires. A domain repertoire is assembled in a combinatorial fashion by allowing each amino acid position in the binding site of a given protein recognition domain to vary to include all the residues allowed at that position in the domain family. The repertoire is then searched by phage display techniques with any target of interest and from the primary structure of the binding site of the selected domains one derives rules that are used to infer the formation of complexes between natural proteins in the cell.

Proteomic analysis of NMDA receptor-adhesion protein signaling complexes

N-methyl-d-aspartate receptors (NMDAR) mediate long-lasting changes in synapse strength via downstream signaling pathways. We report proteomic characterization with mass spectrometry and immunoblotting of NMDAR multiprotein complexes (NRC) isolated from mouse brain. The NRC comprised 77 proteins organized into receptor, adaptor, signaling, cytoskeletal and novel proteins, of which 30 are implicated from binding studies and another 19 participate in NMDAR signaling. NMDAR and metabotropic glutamate receptor subtypes were linked to cadherins and L1 cell-adhesion molecules in complexes lacking AMPA receptors. These neurotransmitter-adhesion receptor complexes were bound to kinases, phosphatases, GTPase-activating proteins and Ras with effectors including MAPK pathway components. Several proteins were encoded by activity-dependent genes. Genetic or pharmacological interference with 15 NRC proteins impairs learning and with 22 proteins alters synaptic plasticity in rodents. Mutations in three human genes (NF1, Rsk-2, L1) are associated with learning impairments, indicating the NRC also participates in human cognition.

Encroaching genomics: adapting large-scale science to small academic laboratories

The process of conducting biological research is undergoing a profound metamorphosis due to the technological innovations and torrent of information resulting from the execution of multiple species genome projects. The further tasks of mapping polymorphisms and characterizing genome-wide protein-protein interaction (the characterization of the proteome) will continue to garner resources, talent, and public attention. Although some elements of these whole genome size projects can only be addressed by large research groups, consortia, or industry, the impact of these projects has already begun to transform the process of research in many small laboratories. Although the impact of this transformation is generally positive, laboratories engaged in types of research destined to be dominated by the efforts of a genomic consortium may be negatively impacted if they cannot rapidly adjust strategies in the face of new large-scale competition. The focus of this report is to outline a series of strategies that have been productively utilized by a number of small academic laboratories that have attempted to integrate such genomic resources into research plans with the goal of developing novel physiological insights.

Chemical ligands, genomics and drug discovery

The sequencing of the human genome and numerous pathogen genomes has resulted in an explosion of potential drug targets. These targets represent both an unprecedented opportunity and a technological challenge for the pharmaceutical industry. A new strategy is required to initiate small-molecule drug discovery with sets of incompletely characterized, disease-associated proteins. One such strategy is the early application of combinatorial chemistry and other technologies to the discovery of bioactive small-molecule ligands that act on candidate drug targets. Therapeutically active ligands serve to concurrently validate a target and provide lead structures for downstream drug development, thereby accelerating the drug discovery process.

Genomic medicine and the future of health care

Genomic technologies and computational advances are leading to an information revolution in biology and medicine. Simulations of molecular processes in cells and predictions of drug effects in humans will advance pharmaceutical research and speed up clinical trials. Computational prognostics and diagnostics that combine with genotyping and molecular profiling may soon cause fundamental changes in the practice of health care.

Toward a protein-protein interaction map of the budding yeast: A comprehensive system to examine two-hybrid interactions in all possible combinations between the yeast proteins

Protein-protein interactions play pivotal roles in various aspects of the structural and functional organization of the cell, and their complete description is indispensable to thorough understanding of the cell. As an approach toward this goal, here we report a comprehensive system to examine two-hybrid interactions in all of the possible combinations between proteins of Saccharomyces cerevisiae. We cloned all of the yeast ORFs individually as a DNA-binding domain fusion ("bait") in a MATa strain and as an activation domain fusion ("prey") in a MATalpha strain, and subsequently divided them into pools, each containing 96 clones. These bait and prey clone pools were systematically mated with each other, and the transformants were subjected to strict selection for the activation of three reporter genes followed by sequence tagging. Our initial examination of approximately 4 x 10(6) different combinations, constituting approximately 10% of the total to be tested, has revealed 183 independent two-hybrid interactions, more than half of which are entirely novel. Notably, the obtained binary data allow us to extract more complex interaction networks, including the one that may explain a currently unsolved mechanism for the connection between distinct steps of vesicular transport. The approach described here thus will provide many leads for integration of various cellular functions and serve as a major driving force in the completion of the protein-protein interaction map.

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.