Evaluation and comparison of mammalian subcellular localization prediction methods

Background: Determination of the subcellular location of a protein is essential to understanding its biochemical function. This information can provide insight into the function of hypothetical or novel proteins. These data are difficult to obtain experimentally but have become especially important since many whole genome sequencing projects have been finished and many resulting protein sequences are still lacking detailed functional information. In order to address this paucity of data, many computational prediction methods have been developed. However, these methods have varying levels of accuracy and perform differently based on the sequences that are presented to the underlying algorithm. It is therefore useful to compare these methods and monitor their performance. Results: In order to perform a comprehensive survey of prediction methods, we selected only methods that accepted large batches of protein sequences, were publicly available, and were able to predict localization to at least nine of the major subcellular locations (nucleus, cytosol, mitochondrion, extracellular region, plasma membrane, Golgi apparatus, endoplasmic reticulum (ER), peroxisome, and lysosome). The selected methods were CELLO, MultiLoc, Proteome Analyst, pTarget and WoLF PSORT. These methods were evaluated using 3763 mouse proteins from SwissProt that represent the source of the training sets used in development of the individual methods. In addition, an independent evaluation set of 2145 mouse proteins from LOCATE with a bias towards the subcellular localization underrepresented in SwissProt was used. The sensitivity and specificity were calculated for each method and compared to a theoretical value based on what might be observed by random chance. Conclusion: No individual method had a sufficient level of sensitivity across both evaluation sets that would enable reliable application to hypothetical proteins. All methods showed lower performance on the LOCATE dataset and variable performance on individual subcellular localizations was observed. Proteins localized to the secretory pathway were the most difficult to predict, while nuclear and extracellular proteins were predicted with the highest sensitivity.

Prediction of Golgi Type II membrane proteins based on their transmembrane domains

Motivation: A major issue in cell biology today is how distinct intracellular regions of the cell, like the Golgi Apparatus, maintain their unique composition of proteins and lipids. The cell differentially separates Golgi resident proteins from proteins that move through the organelle to other subcellular destinations. We set out to determine if we could distinguish these two types of transmembrane proteins using computational approaches. Results: A new method has been developed to predict Golgi membrane proteins based on their transmembrane domains. To establish the prediction procedure, we took the hydrophobicity values and frequencies of different residues within the transmembrane domains into consideration. A simple linear discriminant function was developed with a small number of parameters derived from a dataset of Type II transmembrane proteins of known localization. This can discriminate between proteins destined for Golgi apparatus or other locations (post-Golgi) with a success rate of 89.3% or 85.2%, respectively on our redundancy-reduced data sets.

Mécanismes cellulaires et moléculaires impliqués dans la régulation du développement des circuits d’interneurones GABAergiques dans le néocortex : rôle de la molécule d’adhésion cellulaire neurale (NCAM)

Les interneurones GABAergiques constituent une population mineure de cellules par rapport aux neurones glutamatergiques dans le néocortex. Cependant ils contrôlent fortement l'excitabilité neuronale, la dynamique des réseaux neuronaux et la plasticité synaptique. L'importance des circuits GABAergiques dans le processus fonctionnel et la plasticité des réseaux corticaux est soulignée par des résultats récents qui montrent que des modifications très précises et fiables des circuits GABAergiques sont associées à divers troubles du développement neurologique et à des défauts dans les fonctions cérébrales. De ce fait, la compréhension des mécanismes cellulaires et moléculaires impliquant le développement des circuits GABAergiques est la première étape vers une meilleure compréhension de la façon dont les anomalies de ces processus peuvent se produire. La molécule d’adhésion cellulaire neurale (NCAM) appartient à la super-famille des immunoglobulines de reconnaissance cellulaire et est impliquée dans des interactions homophiliques et hétérophiliques avec d’autres molécules. Même si plusieurs rôles de NCAM ont été démontrés dans la croissance neuronale, la fasciculation axonale, la formation et la maturation de synapses, de même que dans la plasticité cellulaire de plusieurs systèmes, le rôle de NCAM dans la formation des synapses GABAergiques reste inconnu. Ce projet visait donc à déterminer le rôle précis de NCAM dans le processus de maturation des synapses GABAergiques dans le néocortex, en modulant son expression à différentes étapes du développement. L’approche choisie a été de supprimer NCAM dans des cellules GABAergiques à paniers avant la maturation des synapses (EP12-18), pendant la maturation (EP16-24), ou durant le maintien de celles-ci (EP24-32). Les méthodes utilisées ont été le clonage moléculaire, l’imagerie confocale, la culture de coupes organotypiques et des techniques morphométriques de quantification de l’innervation GABAergique. Nos résultats montrent que l’inactivation de NCAM durant la phase de maturation des synapses périsomatiques (EP16-24) cause une réduction du nombre de synapses GABAergiques périsomatiques et du branchement de ces axones. En revanche, durant la phase de maintien (EP26-32), l’inactivation de NCAM n’a pas affecté ces paramètres des synapses GABAergiques. Or, il existe trois isoformes de NCAM (NCAM120, 140 et 180) qui pourraient jouer des rôles différents dans les divers types cellulaires ou à des stades développementaux différents. Nos données montrent que NCAM120 et 140 sont nécessaires à la maturation des synapses périsomatiques GABAergiques. Cependant, NCAM180, qui est l’isoforme la plus étudiée et caractérisée, ne semble pas être impliquée dans ce processus. De plus, l’inactivation de NCAM n’a pas affecté la densité des épines dendritiques ou leur longueur. Elle est donc spécifique aux synapses périsomatiques GABAeriques. Finalement, nos résultats suggèrent que le domaine conservé C-terminal KENESKA est essentiel à la maturation des synapses périsomatiques GABAergiques. Des expériences futures nous aiderons à mieux comprendre la mécanistique et les différentes voies de signalisation impliquées.

Interaction of leptospira elongation factor tu with plasminogen and complement factor h: a metabolic leptospiral protein with moonlighting activities

The elongation factor Tu (EF-Tu), an abundant bacterial protein involved in protein synthesis, has been shown to display moonlighting activities. Known to perform more than one function at different times or in different places, it is found in several subcellular locations in a single organism, and may serve as a virulence factor in a range of important human pathogens. Here we demonstrate that Leptospira EF-Tu is surface-exposed and performs additional roles as a cell-surface receptor for host plasma proteins. It binds plasminogen in a dose-dependent manner, and lysine residues are critical for this interaction. Bound plasminogen is converted to active plasmin, which, in turn, is able to cleave the natural substrates C3b and fibrinogen. Leptospira EF-Tu also acquires the complement regulator Factor H (FH). FH bound to immobilized EF-Tu displays cofactor activity, mediating C3b degradation by Factor I (FI). In this manner, EF-Tu may contribute to leptospiral tissue invasion and complement inactivation. To our knowledge, this is the first description of a leptospiral protein exhibiting moonlighting activities

The role of functional tumor suppressor 15-lipoxygenase 2 (15-LOX2) in normal human prostate epithelial cells and prostate cancer development

15-Lipoxygenase 2 (15-LOX2) is a recently cloned human lipoxygenase that shows tissue-restricted expression in prostate, lung, skin, and cornea. The protein level and enzymatic activity of 15-LOX2 have been shown to be down-regulated in prostate cancers compared with normal and benign prostate tissues. We report the cloning and functional characterization of 15-LOX2 and its three splice variants (termed 15-LOX2sv-a, 15-LOX2sv-b, and 15-LOX2sv-c) from primary prostate epithelial (NHP) cells. Western blotting with multiple NHP cell strains and prostate cancer (PCa) cell lines reveals that the expression of 15-LOX2 is lost in all PCa cell lines, accompanied by decreased enzymatic activity. 15-LOX2 is expressed at multiple subcellular locations, including cytoplasm, cytoskeleton, cell-cell border, and nucleus. Surprisingly, the three splice variants of 15-LOX2 are mostly excluded from the nucleus. To elucidate the relationship between nuclear localization, enzymatic activity, and tumor suppressive functions, we established PCa cell clones stably expressing 15-LOX2 or 15-LOX2sv-b. The 15-LOX2 clones express 15-LOX2 in the nuclei and possess robust enzymatic activity, whereas 15-LOX2sv-b clones show neither nuclear protein localization nor arachidonic acid-metabolizing activity. Interestingly, both 15-LOX2- and 15-LOX2sv-b-stable clones proliferate much slower in vitro when compared with control clones. When orthotopically implanted in nude mouse prostate, both 15-LOX2 and 15-LOX2sv-b suppress PC3 tumor growth in vivo. Finally, cultured NHP cells lose the expression of putative stem/progenitor cell markers, slow down in proliferation, and enter senescence. Several pieces of evidence implicate 15-LOX2 plays a role in replicative senescence of NHP cells: (1) promoter activity and the mRNA and protein levels of 15-LOX2 and its splice variants are upregulated in serially passaged NHP cells, which precede replicative senescence and occur in a cell-autonomous manner; (2) PCa cells stably expressing 15-LOX2 or 15-LOX2sv-b show a passage-related senescence-like phenotype; (3) enforced expression of 15-LOX2 or 15-LOX2sv-b in young NHP cells induce partial cell-cycle arrest and senescence-like phenotypes. Together, these results suggest that 15-LOX2 suppress prostate tumor development and do not necessarily depend on arachidonic acid-metabolizing activity and nuclear localization. Also, 15-LOX2 may serve as an endogenous prostate senescence gene and its tumor-suppressing functions might be associated with its ability to induce cell senescence. ^

Measurement of cytosolic, mitochondrial, and Golgi pH in single living cells with green fluorescent proteins

Many cellular events depend on a tightly compartmentalized distribution of H+ ions across membrane-bound organelles. However, measurements of organelle pH in living cells have been scarce. Several mutants of the Aequorea victoria green fluorescent protein (GFP) displayed a pH-dependent absorbance and fluorescent emission, with apparent pKa values ranging from 6.15 (mutations F64L/S65T/H231L) and 6.4 (K26R/F64L/S65T/Y66W/N146I/M153T/V163A/N164H/H231L) to a remarkable 7.1 (S65G/S72A/T203Y/H231L). We have targeted these GFPs to the cytosol plus nucleus, the medial/trans-Golgi by fusion with galactosyltransferase, and the mitochondrial matrix by using the targeting signal from subunit IV of cytochrome c oxidase. Cells in culture transfected with these cDNAs displayed the expected subcellular localization by light and electron microscopy and reported local pH that was calibrated in situ with ionophores. We monitored cytosolic and nuclear pH of HeLa cells, and mitochondrial matrix pH in HeLa cells and in rat neonatal cardiomyocytes. The pH of the medial/trans-Golgi was measured at steady-state (calibrated to be 6.58 in HeLa cells) and after various manipulations. These demonstrated that the Golgi membrane in intact cells is relatively permeable to H+, and that Cl− serves as a counter-ion for H+ transport and likely helps to maintain electroneutrality. The amenability to engineer GFPs to specific subcellular locations or tissue targets using gene fusion and transfer techniques should allow us to examine pH at sites previously inaccessible.

Apg7p/Cvt2p Is Required for the Cytoplasm-to-Vacuole Targeting, Macroautophagy, and Peroxisome Degradation Pathways

Proper functioning of organelles necessitates efficient protein targeting to the appropriate subcellular locations. For example, degradation in the fungal vacuole relies on an array of targeting mechanisms for both resident hydrolases and their substrates. The particular processes that are used vary depending on the available nutrients. Under starvation conditions, macroautophagy is the primary method by which bulk cytosol is sequestered into autophagic vesicles (autophagosomes) destined for this organelle. Molecular genetic, morphological, and biochemical evidence indicates that macroautophagy shares much of the same cellular machinery as a biosynthetic pathway for the delivery of the vacuolar hydrolase, aminopeptidase I, via the cytoplasm-to-vacuole targeting (Cvt) pathway. The machinery required in both pathways includes a novel protein modification system involving the conjugation of two autophagy proteins, Apg12p and Apg5p. The conjugation reaction was demonstrated to be dependent on Apg7p, which shares homology with the E1 family of ubiquitin-activating enzymes. In this study, we demonstrate that Apg7p functions at the sequestration step in the formation of Cvt vesicles and autophagosomes. The subcellular localization of Apg7p fused to green fluorescent protein (GFP) indicates that a subpopulation of Apg7pGFP becomes membrane associated in an Apg12p-dependent manner. Subcellular fractionation experiments also indicate that a portion of the Apg7p pool is pelletable under starvation conditions. Finally, we demonstrate that the Pichia pastoris homologue Gsa7p that is required for peroxisome degradation is functionally similar to Apg7p, indicating that this novel conjugation system may represent a general nonclassical targeting mechanism that is conserved across species.

Activation and targeting of extracellular signal-regulated kinases by β-arrestin scaffolds

Using both confocal immunofluorescence microscopy and biochemical approaches, we have examined the role of β-arrestins in the activation and targeting of extracellular signal-regulated kinase 2 (ERK2) following stimulation of angiotensin II type 1a receptors (AT1aR). In HEK-293 cells expressing hemagglutinin-tagged AT1aR, angiotensin stimulation triggered β-arrestin-2 binding to the receptor and internalization of AT1aR-β-arrestin complexes. Using red fluorescent protein-tagged ERK2 to track the subcellular distribution of ERK2, we found that angiotensin treatment caused the redistribution of activated ERK2 into endosomal vesicles that also contained AT1aR-β-arrestin complexes. This targeting of ERK2 reflects the formation of multiprotein complexes containing AT1aR, β-arrestin-2, and the component kinases of the ERK cascade, cRaf-1, MEK1, and ERK2. Myc-tagged cRaf-1, MEK1, and green fluorescent protein-tagged ERK2 coprecipitated with Flag-tagged β-arrestin-2 from transfected COS-7 cells. Coprecipitation of cRaf-1 with β-arrestin-2 was independent of MEK1 and ERK2, whereas the coprecipitation of MEK1 and ERK2 with β-arrestin-2 was significantly enhanced in the presence of overexpressed cRaf-1, suggesting that binding of cRaf-1 to β-arrestin facilitates the assembly of a cRaf-1, MEK1, ERK2 complex. The phosphorylation of ERK2 in β-arrestin complexes was markedly enhanced by coexpression of cRaf-1, and this effect is blocked by expression of a catalytically inactive dominant inhibitory mutant of MEK1. Stimulation with angiotensin increased the binding of both cRaf-1 and ERK2 to β-arrestin-2, and the association of β-arrestin-2, cRaf-1, and ERK2 with AT1aR. These data suggest that β-arrestins function both as scaffolds to enhance cRaf-1 and MEK-dependent activation of ERK2, and as targeting proteins that direct activated ERK to specific subcellular locations.

Nucleus-associated pools of Rna1p, the Saccharomyces cerevisiae Ran/TC4 GTPse activating protein involved in nucleus/cytosol transit.

Rna1p is the GTPase activating enzyme for Ran/TC4, a Ras-like GTPase necessary for nuclear/cytosolic exchange. Although most wild-type Rna1p is located in the cytosol, we found that the vast majority of the mutant Rna1-1p and, under appropriate physiological conditions, a small portion of the wild-type Rna1p cofractionate with yeast nuclei. Subnuclear fractionation studies show that most of the Rna1p is tightly associated with nuclear components, and that a portion of the active protein can be solubilized by treatments that fail to solubilize inactive Rna1-1p. To learn the precise nuclear locations of the Rna1 proteins, we studied their subcellular distributions in HeLa cells. By indirect immuno-fluorescence we show that wild-type Rna1p has three subcellular locations. The majority of the protein is distributed throughout the cytosol, but a portion of the protein is nucleus-associated, located at both the cytosolic surface and within the nucleoplasm. Mutant Rna1-1p is found at the outer nuclear surface and in the cytosol. We propose that a small pool of the wild-type Rna1p is located in the nuclear interior, supporting the model that the same components of the Ran/TC4 GTPase cycle exist on both sides of the nuclear membrane.

Differential use of signal peptides and membrane domains is a common occurrence in the protein output of transcriptional units

Membrane organization describes the orientation of a protein with respect to the membrane and can be determined by the presence, or absence, and organization within the protein sequence of two features: endoplasmic reticulum signal peptides and alpha-helical transmembrane domains. These features allow protein sequences to be classified into one of five membrane organization categories: soluble intracellular proteins, soluble secreted proteins, type I membrane proteins, type II membrane proteins, and multi- spanning membrane proteins. Generation of protein isoforms with variable membrane organizations can change a protein's subcellular localization or association with the membrane. Application of MemO, a membrane organization annotation pipeline, to the FANTOM3 Isoform Protein Sequence mouse protein set revealed that within the 8,032 transcriptional units ( TUs) with multiple protein isoforms, 573 had variation in their use of signal peptides, 1,527 had variation in their use of transmembrane domains, and 615 generated protein isoforms from distinct membrane organization classes. The mechanisms underlying these transcript variations were analyzed. While TUs were identified encoding all pairwise combinations of membrane organization categories, the most common was conversion of membrane proteins to soluble proteins. Observed within our highconfidence set were 156 TUs predicted to generate both extracellular soluble and membrane proteins, and 217 TUs generating both intracellular soluble and membrane proteins. The differential use of endoplasmic reticulum signal peptides and transmembrane domains is a common occurrence within the variable protein output of TUs. The generation of protein isoforms that are targeted to multiple subcellular locations represents a major functional consequence of transcript variation within the mouse transcriptome.

Proteins accessible to immune surveillance show significant T-cell epitope depletion:implications for vaccine design

T cell activation is the final step in a complex pathway through which pathogen-derived peptide fragments can elicit an immune response. For it to occur, peptides must form stable complexes with Major Histocompatibility Complex (MHC) molecules and be presented on the cell surface. Computational predictors of MHC binding are often used within in silico vaccine design pathways. We have previously shown that, paradoxically, most bacterial proteins known experimentally to elicit an immune response in disease models are depleted in peptides predicted to bind to human MHC alleles. The results presented here, derived using software proven through benchmarking to be the most accurate currently available, show that vaccine antigens contain fewer predicted MHC-binding peptides than control bacterial proteins from almost all subcellular locations with the exception of cell wall and some cytoplasmic proteins. This effect is too large to be explained from the undoubted lack of precision of the software or from the amino acid composition of the antigens. Instead, we propose that pathogens have evolved under the influence of the host immune system so that surface proteins are depleted in potential MHC-binding peptides, and suggest that identification of a protein likely to contain a single immuno-dominant epitope is likely to be a productive strategy for vaccine design.

Toward bacterial protein sub-cellular location prediction:single-class discrimminant models for all gram- and gram+ compartments

Based on Bayesian Networks, methods were created that address protein sequence-based bacterial subcellular location prediction. Distinct predictive algorithms for the eight bacterial subcellular locations were created. Several variant methods were explored. These variations included differences in the number of residues considered within the query sequence - which ranged from the N-terminal 10 residues to the whole sequence - and residue representation - which took the form of amino acid composition, percentage amino acid composition, or normalised amino acid composition. The accuracies of the best performing networks were then compared to PSORTB. All individual location methods outperform PSORTB except for the Gram+ cytoplasmic protein predictor, for which accuracies were essentially equal, and for outer membrane protein prediction, where PSORTB outperforms the binary predictor. The method described here is an important new approach to method development for subcellular location prediction. It is also a new, potentially valuable tool for candidate subunit vaccine selection.

Combining algorithms to predict bacterial protein sub-cellular location:parallel versus concurrent implementations

We describe a novel and potentially important tool for candidate subunit vaccine selection through in silico reverse-vaccinology. A set of Bayesian networks able to make individual predictions for specific subcellular locations is implemented in three pipelines with different architectures: a parallel implementation with a confidence level-based decision engine and two serial implementations with a hierarchical decision structure, one initially rooted by prediction between membrane types and another rooted by soluble versus membrane prediction. The parallel pipeline outperformed the serial pipeline, but took twice as long to execute. The soluble-rooted serial pipeline outperformed the membrane-rooted predictor. Assessment using genomic test sets was more equivocal, as many more predictions are made by the parallel pipeline, yet the serial pipeline identifies 22 more of the 74 proteins of known location.

Subcellular localization of gamma-aminobutyric acid type A receptors is determined by receptor beta subunits.

gamma-aminobutyric acid type A (GABAA) receptors are the major sites of fast synaptic inhibition in the brain. They are constructed from four subunit classes with multiple members: alpha (1-6), beta (1-4), gamma (1-4), and delta (1). The contribution of subunit diversity in determining receptor subcellular targeting was examined in polarized Madin-Darby canine kidney (MDCK) cells. Significant detection of cell surface homomeric receptor expression by a combination of both immunological and electrophysiological methodologies was only found for the beta 3 subunit. Expression of alpha/beta binary combinations resulted in a nonpolarized distribution for alpha 1 beta 1 complexes, but specific basolateral targeting of both alpha 1 beta 2 and alpha 1 beta 3 complexes. The polarized distribution of these alpha/beta complexes was unaffected by the presence of the gamma 2S subunit. Interestingly, delivery of receptors containing the beta 3 subunit to the basolateral domain occurs via the apical surface. These results show that beta subunits can selectively target GABAA receptors to distinct cellular locations. Changes in the spatial and temporal expression of beta-subunit isoforms may therefore provide a mechanism for relocating GABAA receptor function between distinct neuronal domains. Given the critical role of these receptors in mediating synaptic inhibition, the contribution of different beta subunits in GABAA receptor function, may have implications in neuronal development and for receptor localization/clustering.

Multi-class subcellular location prediction for bacterial proteins

Two algorithms, based onBayesian Networks (BNs), for bacterial subcellular location prediction, are explored in this paper: one predicts all locations for Gram+ bacteria and the other all locations for Gram- bacteria. Methods were evaluated using different numbers of residues (from the N-terminal 10 residues to the whole sequence) and residue representation (amino acid-composition, percentage amino acid-composition or normalised amino acid-composition). The accuracy of the best resulting BN was compared to PSORTB. The accuracy of this multi-location BN was roughly comparable to PSORTB; the difference in predictions is low, often less than 2%. The BN method thus represents both an important new avenue of methodological development for subcellular location prediction and a potentially value new tool of true utilitarian value for candidate subunit vaccine selection.