The Interactorium: Visualising proteins,complexes and interaction networks in a virtual 3‐D cell

Here, we describe the Interactorium, a tool in which a Virtual Cell is used as the context for the seamless visualisation of the yeast protein interaction network, protein complexes and protein 3‐D structures. The tool has been designed to display very complex networks of up to 40 000 proteins or 6000 multiprotein complexes and has a series of toolboxes and menus to allow real‐time data manipulation and control the manner in which data are displayed. It incorporates new algorithms that reduce the complexity of the visualisation by the generation of putative new complexes from existing data and by the reduction of edges through the use of protein “twins” when they occur in multiple locations. Since the Interactorium permits multi‐level viewing of the molecular biology of the cell, it is a considerable advance over existing approaches. We illustrate its use for Saccharomyces cerevisiae but note that it will also be useful for the analysis of data from simpler prokaryotes and higher eukaryotes, including humans. The Interactorium is available for download at http://www.interactorium.net .     

Identification and characterization of novel ERC‐55 interacting proteins: Evidence for the existence of several ERC‐55 splicing variants; including the cytosolic ERC‐55‐C

ERC‐55, encoded from RCN2, is localized in the ER and belongs to the CREC protein family. ERC‐55 is involved in various diseases and abnormal cell behavior, however, the function is not well defined and it has controversially been reported to interact with a cytosolic protein, the vitamin D receptor. We have used a number of proteomic techniques to further our functional understanding of ERC‐55. By affinity purification, we observed interaction with a large variety of proteins, including those secreted and localized outside of the secretory pathway, in the cytosol and also in various organelles. We confirm the existence of several ERC‐55 splicing variants including ERC‐55‐C localized in the cytosol in association with the cytoskeleton. Localization was verified by immunoelectron microscopy and sub‐cellular fractionation. Interaction of lactoferrin, S100P, calcyclin (S100A6), peroxiredoxin‐6, kininogen and lysozyme with ERC‐55 was further studied in vitro by SPR experiments. Interaction of S100P requires [Ca²⁺] of ～10^?7 M or greater, while calcyclin interaction requires [Ca²⁺] of >10^?5 M. Interaction with peroxiredoxin‐6 is independent of Ca²⁺. Co‐localization of lactoferrin, S100P and calcyclin with ERC‐55 in the perinuclear area was analyzed by fluorescence confocal microscopy. The functional variety of the interacting proteins indicates a broad spectrum of ERC‐55 activities such as immunity, redox homeostasis, cell cycle regulation and coagulation.     

Cytoskeletal scaffolding proteins interact with Lynch‐Syndrome associated mismatch repair protein MLH1

The involvement of MLH1 in several mismatch repair‐independent cellular processes has been reported. In an attempt to gain further insight into the protein's cellular functions, we screened for novel interacting partners of MLH1 utilizing a bacterial two‐hybrid system. Numerous unknown interacting proteins were identified, suggesting novel biological roles of MLH1. The network of MLH1 and its partner proteins involves a multitude of cellular processes. Integration of our data with the “General Repository for Interaction Datasets” highlighted that MLH1 exhibits relationships to three interacting pairs of proteins involved in cytoskeletal and filament organization: Thymosin β 4 and Actin γ, Cathepsin B and Annexin A2 as well as Spectrin α and Desmin. Coimmunoprecipitation and colocalization experiments validated the interaction of MLH1 with these proteins. Differential mRNA levels of many of the identified proteins, detected by microarray analysis comparing MLH1‐deficient and ‐proficient cell lines, support the assumed interplay of MLH1 and the identified candidate proteins. By siRNA knock down of MLH1, we demonstrated the functional impact of MLH1–Actin interaction on filament organization and propose that dysregulation of MLH1 plays an essential role in cytoskeleton dynamics. Our data suggest novel roles of MLH1 in cellular organization and colorectal cancerogenesis.     

Proteomic,functional and motif‐based analysis of C‐terminal Src kinase‐interacting proteins

C‐terminal Src kinase (Csk) that functions as an essential negative regulator of Src family tyrosine kinases (SFKs) interacts with tyrosine‐phosphorylated molecules through its Src homology 2 (SH2) domain, allowing it targeting to the sites of SFKs and concomitantly enhancing its kinase activity. Identification of additional Csk‐interacting proteins is expected to reveal potential signaling targets and previously undescribed functions of Csk. In this study, using a direct proteomic approach, we identified 151 novel potential Csk‐binding partners, which are associated with a wide range of biological functions. Bioinformatics analysis showed that the majority of identified proteins contain one or several Csk‐SH2 domain‐binding motifs, indicating a potentially direct interaction with Csk. The interactions of Csk with four proteins (partitioning defective 3 (Par3), DDR1, SYK and protein kinase C iota) were confirmed using biochemical approaches and phosphotyrosine 1127 of Par3 C‐terminus was proved to directly bind to Csk‐SH2 domain, which was consistent with predictions from in silico analysis. Finally, immunofluorescence experiments revealed co‐localization of Csk with Par3 in tight junction (TJ) in a tyrosine phosphorylation‐dependent manner and overexpression of Csk, but not its SH2‐domain mutant lacking binding to phosphotyrosine, promoted the TJ assembly in Madin‐Darby canine kidney cells, implying the involvement of Csk‐SH2 domain in regulating cellular TJs. In conclusion, the newly identified potential interacting partners of Csk provided new insights into its functional diversity in regulation of numerous cellular events, in addition to controlling the SFK activity.     

LuMPIS – A modified luminescence‐based mammalian interactome mapping pull‐down assay for the investigation of protein–protein interactions encoded by GC‐low ORFs

The GC content is highly variable among the genomes of different organisms. It has been shown that recombinant gene expression in mammalian cells is much more efficient when GC‐rich coding sequences of a certain protein are used. In order to study protein–protein interactions in Varicella zoster virus, a GC‐low herpesvirus, we have developed a novel luminescence‐based maltose‐binding protein pull‐down interaction screening system (LuMPIS) that is able to overcome the impaired protein expression levels of GC‐low ORFs in mammalian expression systems.     

A domain level interaction network of amyloid precursor protein and Aβ of Alzheimer's disease

The primary constituent of the amyloid plaque, β‐amyloid (Aβ), is thought to be the causal “toxic moiety” of Alzheimer's disease. However, despite much work focused on both Aβ and its parent protein, amyloid precursor protein (APP), the functional roles of APP and its cleavage products remain to be fully elucidated. Protein–protein interaction networks can provide insight into protein function, however, high‐throughput data often report false positives and are in frequent disagreement with low‐throughput experiments. Moreover, the complexity of the CNS is likely to be under represented in such databases. Therefore, we curated the published work characterizing both APP and Aβ to create a protein interaction network of APP and its proteolytic cleavage products, with annotation, where possible, to the level of APP binding domain and isoform. This is the first time that an interactome has been refined to domain level, essential for the interpretation of APP due to the presence of multiple isoforms and processed fragments. Gene ontology and network analysis were used to identify potentially novel functional relationships among interacting proteins.     

Global microRNA level regulation of EGFR‐driven cell‐cycle protein network in breast cancer

The EGFR‐driven cell‐cycle pathway has been extensively studied due to its pivotal role in breast cancer proliferation and pathogenesis. Although several studies reported regulation of individual pathway components by microRNAs (miRNAs), little is known about how miRNAs coordinate the EGFR protein network on a global miRNA (miRNome) level. Here, we combined a large‐scale miRNA screening approach with a high‐throughput proteomic readout and network‐based data analysis to identify which miRNAs are involved, and to uncover potential regulatory patterns. Our results indicated that the regulation of proteins by miRNAs is dominated by the nucleotide matching mechanism between seed sequences of the miRNAs and 3′‐UTR of target genes. Furthermore, the novel network‐analysis methodology we developed implied the existence of consistent intrinsic regulatory patterns where miRNAs simultaneously co‐regulate several proteins acting in the same functional module. Finally, our approach led us to identify and validate three miRNAs (miR‐124, miR‐147 and miR‐193a‐3p) as novel tumor suppressors that co‐target EGFR‐driven cell‐cycle network proteins and inhibit cell‐cycle progression and proliferation in breast cancer.     

Mapping the in vitro interactome of cardiac sodium (Na+)‐calcium (Ca2+) exchanger 1 (NCX1)

The sodium (Na⁺)‐calcium (Ca²⁺) exchanger 1 (NCX1) is an antiporter membrane protein encoded by the SLC8A1 gene. In the heart, it maintains cytosolic Ca²⁺ homeostasis, serving as the primary mechanism for Ca²⁺ extrusion during relaxation. Dysregulation of NCX1 is observed in end‐stage human heart failure. In this study, we used affinity purification coupled with MS in rat left ventricle lysates to identify novel NCX1 interacting proteins in the heart. Two screens were conducted using: (1) anti‐NCX1 against endogenous NCX1 and (2) anti‐His (where His is histidine) with His‐trigger factor‐NCX1_cyt recombinant protein as bait. The respective methods identified 112 and 350 protein partners, of which several were known NCX1 partners from the literature, and 29 occurred in both screens. Ten novel protein partners (DYRK1A, PPP2R2A, SNTB1, DMD, RABGGTA, DNAJB4, BAG3, PDE3A, POPDC2, STK39) were validated for binding to NCX1, and two partners (DYRK1A, SNTB1) increased NCX1 activity when expressed in HEK293 cells. A cardiac NCX1 protein–protein interaction map was constructed. The map was highly connected, containing distinct clusters of proteins with different biological functions, where “cell communication” and “signal transduction” formed the largest clusters. The NCX1 interactome was also significantly enriched with proteins/genes involved in “cardiovascular disease” which can be explored as novel drug targets in future research.     

A computational approach to inferring cellular protein‐binding affinities from quantitative fluorescence resonance energy transfer imaging

Fluorescence resonance energy transfer (FRET) microscopy can measure the spatial distribution of protein interactions inside live cells. Such experiments give rise to complex data sets with many images of single cells, motivating data reduction and abstraction. In particular, determination of the value of the equilibrium dissociation constant (K_d) will provide a quantitative measure of protein–protein interactions, which is essential to reconstructing cellular signaling networks. Here, we investigate the feasibility of using quantitative FRET imaging of live cells to estimate the local value of K_d for two interacting labeled molecules. An algorithm is developed to infer the values of K_d using the intensity of individual voxels of 3‐D FRET microscopy images. The performance of our algorithm is investigated using synthetic test data, both in the absence and in the presence of endogenous (unlabeled) proteins. The influence of optical blurring caused by the microscope (confocal or wide field) and detection noise on the accuracy of K_d inference is studied. We show that deconvolution of images followed by analysis of intensity data at local level can improve the estimate of K_d. Finally, the performance of this algorithm using cellular data on the interaction between yellow fluorescent protein‐Rac and cyan fluorescent protein‐PBD in mammalian cells is shown.     

Combining protein complementation assays with resonance energy transfer to detect multipartner protein complexes in living cells

A variety of fluorescent proteins with different spectral properties have been created by mutating green fluorescent protein. When these proteins are split in two, neither fragment is fluorescent per se, nor can a fluorescent protein be reconstituted by co-expressing the complementary N- and C-terminal fragments. However, when these fragments are genetically fused to proteins that associate with each other in cellulo, the N- and C-terminal fragments of the fluorescent protein are brought together and can reconstitute a fluorescent protein. A similar protein complementation assay (PCA) can be performed with two complementary fragments of various luciferase isoforms. This makes these assays useful tools for detecting the association of two proteins in living cells. Bioluminescence resonance energy transfer (BRET) or fluorescence resonance energy transfer (FRET) occurs when energy from, respectively, a luminescent or fluorescent donor protein is non-radiatively transferred to a fluorescent acceptor protein. This transfer of energy can only occur if the proteins are within 100 Å of each other. Thus, BRET and FRET are also useful tools for detecting the association of two proteins in living cells. By combining different protein fragment complementation assays (PCA) with BRET or FRET it is possible to demonstrate that three or more proteins are simultaneous parts of the same protein complex in living cells. As an example of the utility of this approach, we show that as many as four different proteins are simultaneously associated as part of a G protein-coupled receptor signalling complex.     

A non‐canonical rhodopsin‐mediated insulin receptor signaling pathway in retinal photoreceptor neurons

We previously reported a ligand‐independent and rhodopsin‐dependent insulin receptor (IR) neuroprotective signaling pathway in both rod and cone photoreceptor cells, which is activated through protein–protein interaction. Our previous studies were performed with either retina or isolated rod or cone outer segment preparations and the expression of IR signaling proteins were examined. The isolation of outer segments with large portions of the attached inner segments is a technical challenge. Optiprep? density gradient medium has been used to isolate the cells and subcellular organelles, Optiprep? is a non‐ionic iodixanol‐based medium with a density of 1.320 g/mL. We employed this method to examine the expression of IR and its signaling proteins, and activation of one of the downstream effectors of the IR in isolated photoreceptor cells. Identification of the signaling complexes will be helpful for therapeutic targeting in disease conditions.     

Proteome changes in human bronchoalveolar cells following styrene exposure indicate involvement of oxidative stress in the molecular‐response mechanism

Styrene is a volatile organic compound that is widely used as an intermediate in many industrial settings. There are known adverse health effects at environmentally significant concentrations, but little is known about the molecular effect of exposure to styrene at sub‐acute toxic concentrations. We exposed human lung epithelial cells, at a wide range of concentrations (1 mg/m³–10 g/m³), to styrene and analyzed the effects on the proteome level by 2‐DE, where 1380 proteins spots were detected and 266 were identified unambiguously by MS. A set of 16 protein spots were found to be significantly altered due to exposure to styrene at environmentally significant concentrations of 1–10 mg/m³ (0.2–2.3 ppm). Among these, superoxide dismutase as well as biliverdin reductase A could be correlated with the molecular pathway of oxidative stress, while eukaryotic translation initiation factor 5A‐1, ezrin, lamin B2 and voltage‐dependent anion channel 2 have been reported to be involved in apoptosis. Treatment with styrene also caused the formation of styrene oxide–protein adducts, specifically for thioredoxin reductase 1. These results underline the relevance of oxidative stress as a primary molecular response mechanism of lung epithelial cells to styrene exposure at indoor‐relevant concentrations.     

Penetration into membrane of amino‐terminal region of SecA when associated with SecYEG in active complexes

The general secretory (Sec) system of Escherichia coli translocates both periplasmic and outer membrane proteins through the cytoplasmic membrane. The pathway through the membrane is provided by a highly conserved translocon, which in E. coli comprises two heterotrimeric integral membrane complexes, SecY, SecE, and SecG (SecYEG), and SecD, SecF, and YajC (SecDF/YajC). SecA is an associated ATPase that is essential to the function of the Sec system. SecA plays two roles, it targets precursors to the translocon with the help of SecB and it provides energy via hydrolysis of ATP. SecA exists both free in the cytoplasm and integrally membrane associated. Here we describe details of association of the amino‐terminal region of SecA with membrane. We use site‐directed spin labelling and electron paramagnetic resonance spectroscopy to show that when SecA is co‐assembled into lipids with SecYEG to yield highly active translocons, the N‐terminal region of SecA penetrates the membrane and lies at the interface between the polar and the hydrophobic regions, parallel to the plane of the membrane at a depth of approximately 5 Å. When SecA is bound to SecYEG, preassembled into proteoliposomes, or nonspecifically bound to lipids in the absence of SecYEG, the N‐terminal region penetrates more deeply (8 Å). Implications of partitioning of the SecA N‐terminal region into lipids on the complex between SecB carrying a precursor and SecA are discussed.     

Application of quantitative immunoprecipitation combined with knockdown and cross‐linking to Chlamydomonas reveals the presence of vesicle‐inducing protein in plastids 1 in a common complex with chloroplast HSP90C

Knowledge of the interaction partners of a protein of interest may provide important information on its function. Common to currently available tools for the identification of protein–protein interactions, however, is their high rates of false positives. Only recently an assay was reported that allowed for the unequivocal identification of protein–protein interactions in mammalian cells in a single experiment. This assay, termed quantitative immunoprecipitation combined with knockdown (QUICK), combines RNAi, stable isotope labeling with amino acids in cell culture, immunoprecipitation, and quantitative MS. We are using the unicellular green alga Chlamydomonas reinhardtii to understand the roles of chaperones in chloroplast biogenesis. The goal of this work was to apply QUICK to Chlamydomonas for the identification of novel interaction partners of vesicle‐inducing protein in plastids 1 (VIPP1), a protein required for the biosynthesis/maintenance of thylakoid membranes and known substrate of chloroplast HSP70B. We report here a robust QUICK protocol for Chlamydomonas that has been improved (i) by introducing a cross‐linking step (‐X) to improve protein complex stability and (ii) by including a control for the correction of unequal immunoprecipitation and/or labeling efficiencies. Using QUICK and cross‐linking we could verify that HSP70B and CGE1 form a complex with VIPP1 and could also demonstrate that chloroplast HSP90C is part of this complex. Moreover, we could show that the chaperones interact with VIPP1 also in membrane fractions.     

A method for reverse interactome analysis: High‐resolution mapping of interdomain interaction network in Dam1 complex and its specific disorganization based on the interaction domain expression

An experimental methodology that facilitates functional analysis of numerous protein–protein interactions, which have been found in genome‐wide interactome researches, has long been awaited. We propose herein an antagonistic inhibition‐based approach. The antagonizing polypeptide is generated in the course of interaction domain mapping based on yeast 2‐hybrid (Y2H) screening coupled with in vitro convergence of the Y2H‐selected fragments, which is performed in a formatted procedure. Using the coupled methodology, we first performed a high‐resolution mapping of an interdomain interaction network within budding yeast's Dam1 complex. Dam1 complex is a kinetochore protein complex composed of 10 essential subunits including Spc34p and Spc19p. The high‐resolution mapping revealed the overall network structure within the complex for the first time: Dam1 components form into two separated subnetworks on N‐terminal scaffolding domains of Spc34p and Spc19p, and the coiled‐coil interaction in their C‐terminal domains connects the subnetworks. Secondly, we show that the domain fragments converged in the high‐resolution mapping acted as potent inhibitors for the endogenous interactions when episomally overexpressed. The in vivo Dam1 interaction targeting with the fragments conferred a similar phenotype on the host cells; a critical and irreversible damage, which was accompanied with disturbed budding and chromosome mis‐segregation as a result of disorganized spindle. These phenotypes were strongly related to the cellular function of the Dam1 complex. The results and approach we demonstrated herein not only shed light on the Dam1 molecular architecture but also pave the road to reverse‐interactome analysis and discoveries of novel drugs that target disease‐related protein–protein interactions. © 2010 American Institute of Chemical Engineers Biotechnol. Prog., 2010     

Biotin tagging coupled with amino acid‐coded mass tagging for efficient and precise screening of interaction proteome in mammalian cells

In mammalian cells, when tandem affinity purification approach is employed, the existence of untagged endogenous target protein and repetitive washing steps together result in overall low yield of purified/stable complexes and the loss of weakly and transiently interacting partners of biological significance. To avoid the trade‐offs involving in methodological sensitivity, precision, and throughput, here we introduce an integrated method, biotin tagging coupled with amino acid‐coded mass tagging, for highly sensitive and accurate screening of mammalian protein–protein interactions. Without the need of establishing a stable cell line, using a short peptide tag which could be specifically biotinylated in vivo, the biotin‐tagged target/bait protein was then isolated along with its associates efficiently by streptavidin magnetic microbeads in a single step. In a pulled‐down complex amino acid‐coded mass tagging serves as “in‐spectra” quantitative markers to distinguish those bait‐specific interactors from non‐specific background proteins under stringent criteria. Applying this biotin tagging coupled with amino acid‐coded mass tagging approach, we first biotin‐tagged in vivo a multi‐functional protein family member, 14‐3‐3ε, which was expressed at close to endogenous level. Starting with approximately 20 millions of 293T cells which were significantly less than what needed for a tandem affinity purification run, 266 specific interactors of 14‐3‐3ε were identified in high confidence.