Because the protein's function is usually related to its subcellular localization, the ability to predict subcellular localization directly from protein sequences will be useful for inferring protein functions. Recent years have seen a surging interest in the development of novel computational tools to predict subcellular localization. At present, these approaches, based on a wide range of algorithms, have achieved varying degrees of success for specific organisms and for certain localization categories. A number of authors have noticed that sequence similarity is useful in predicting subcellular localization. For example, Nair and Rost (Protein Sci 2002;11:2836-2847) have carried out extensive analysis of the relation between sequence similarity and identity in subcellular localization, and have found a close relationship between them above a certain similarity threshold. However, many existing benchmark data sets used for the prediction accuracy assessment contain highly homologous sequences-some data sets comprising sequences up to 80-90% sequence identity. Using these benchmark test data will surely lead to overestimation of the performance of the methods considered. Here, we develop an approach based on a two-level support vector machine (SVM) system: the first level comprises a number of SVM classifiers, each based on a specific type of feature vectors derived from sequences; the second level SVM classifier functions as the jury machine to generate the probability distribution of decisions for possible localizations. We compare our approach with a global sequence alignment approach and other existing approaches for two benchmark data sets-one comprising prokaryotic sequences and the other eukaryotic sequences. Furthermore, we carried out all-against-all sequence alignment for several data sets to investigate the relationship between sequence homology and subcellular localization. Our results, which are consistent with previous studies, indicate that the homology search approach performs well down to 30% sequence identity, although its performance deteriorates considerably for sequences sharing lower sequence identity. A data set of high homology levels will undoubtedly lead to biased assessment of the performances of the predictive approaches-especially those relying on homology search or sequence annotations. Our two-level classification system based on SVM does not rely on homology search; therefore, its performance remains relatively unaffected by sequence homology. When compared with other approaches, our approach performed significantly better. Furthermore, we also develop a practical hybrid method, which combines the two-level SVM classifier and the homology search method, as a general tool for the sequence annotation of subcellular localization.

Gram-negative bacteria have five major subcellular localization sites: the cytoplasm, the periplasm, the inner membrane, the outer membrane, and the extracellular space. The subcellular location of a protein can provide valuable information about its function. With the rapid increase of sequenced genomic data, the need for an automated and accurate tool to predict subcellular localization becomes increasingly important. We present an approach to predict subcellular localization for Gram-negative bacteria. This method uses the support vector machines trained by multiple feature vectors based on n-peptide compositions. For a standard data set comprising 1443 proteins, the overall prediction accuracy reaches 89%, which, to the best of our knowledge, is the highest prediction rate ever reported. Our prediction is 14% higher than that of the recently developed multimodular PSORT-B. Because of its simplicity, this approach can be easily extended to other organisms and should be a useful tool for the high-throughput and large-scale analysis of proteomic and genomic data.Keywords: subcellular localization; support vector machine; Gram-negative bacteria; machine-learning method; proteome; genome; n-peptide compositions The subcellular location of a protein is closely correlated to its biological function (Jensen et al. 2002). With the rapid increase of sequenced genomic data, the need for an automated and accurate tool to predict protein subcellular localization becomes increasingly important. Many efforts have been made to predict protein subcellular localization. There are methods (Nakai and Kanehisa 1992;Nielsen et al. 1997;Emanuelsson et al. 1999Emanuelsson et al. , 2000Nakai 2000) based on the observation that sequences targeted to specific locations rely on the N-terminal sorting or signal sequences. For example, TargetP (Emanuelsson et al. 2000), a useful tool for analysis of signal peptides, predicts protein subcellular localization for eukaryotic sequences. On the other hand, a number of studies (Cedano et al. 1997;Andrade et al. 1998;Reinhardt and Hubbard 1998;Yuan 1999;Chou 2001;Hua and Sun 2001;Chou and Cai 2002) have shown that amino acid compositions are useful in discriminating protein subcellular localization sites. Cedano et al. (1997) developed a predictive system ProtLock based on a correlation analysis of the amino acid compositions and the cellular locations for five protein classes. Reinhardt and Hubbard (1998) developed a neural network approach, NNPSL, based on amino acid compositions for both eukaryotic and prokaryotic sequences. For the same data sets, Hua and Sun (2001) also developed SubLoc based on support vector machine (SVM) techniques. Chou (2001) developed approaches based on the pseudo amino acid compositions that include sequenceorder information.Gram-negative bacteria have five major subcellular localization sites that include the cytoplasm, the inner membrane, the outer membrane, the periplasm, and the extracelReprint request to:

No abstract

Background: The inhibitory leukocyte receptor PD-1 binds two ligands, PD-L1 and PD-L2.Results: Nuclear magnetic resonance analysis and rigorous binding and thermodynamic measurements reveal the structure of, and the mode of ligand recognition by, PD-1.Conclusion: PD-L1 and PD-L2 bind differently to PD-1 and much more weakly than expected.Significance: Potent inhibitory signaling can be initiated by weakly interacting receptors.

SummaryGlycoproteins present special problems for structural genomic analysis because they often require glycosylation in order to fold correctly, whereas their chemical and conformational heterogeneity generally inhibits crystallization. We show that the “glycosylation problem” can be solved by expressing glycoproteins transiently in mammalian cells in the presence of the N-glycosylation processing inhibitors, kifunensine or swainsonine. This allows the correct folding of the glycoproteins, but leaves them sensitive to enzymes, such as endoglycosidase H, that reduce the N-glycans to single residues, enhancing crystallization. Since the scalability of transient mammalian expression is now comparable to that of bacterial systems, this approach should relieve one of the major bottlenecks in structural genomic analysis.

Cytochrome bc(1) is an integral membrane protein complex essential to cellular respiration and photosynthesis. The Q cycle reaction mechanism of bc(1) postulates a separated quinone reduction (Q(i)) and quinol oxidation (Q(o)) site. In a complete catalytic cycle, a quinone molecule at the Q(i) site receives two electrons from the b(H) heme and two protons from the negative side of the membrane; this process is specifically inhibited by antimycin A and NQNO. The structures of bovine mitochondrial bc(1) in the presence or absence of bound substrate ubiquinone and with either the bound antimycin A(1) or NQNO were determined and refined. A ubiquinone with its first two isoprenoid repeats and an antimycin A(1) were identified in the Q(i) pocket of the substrate and inhibitor bound structures, respectively; the NQNO, on the other hand, was identified in both Q(i) and Q(o) pockets in the inhibitor complex. The two inhibitors occupied different portions of the Q(i) pocket and competed with substrate for binding. In the Q(o) pocket, the NQNO behaves similarly to stigmatellin, inducing an iron-sulfur protein conformational arrest. Extensive binding interactions and conformational adjustments of residues lining the Q(i) pocket provide a structural basis for the high affinity binding of antimycin A and for phenotypes of inhibitor resistance. A two-water-mediated ubiquinone protonation mechanism is proposed involving three Q(i) site residues His(201), Lys(227), and Asp(228).

Summary The cytochrome bc1 complex (bc1) is the mid-segment of the cellular respiratory chain of mitochondria and many aerobic prokaryotic organisms; it is also part of the photosynthetic apparatus of non-oxygenic purple bacteria. The bc1 complex catalyzes the reaction of transferring electrons from the low potential substrate ubiquinol to high potential cytochrome c. Concomitantly, bc1 translocates protons across the membrane, contributing to the proton-motive force essential for a variety of cellular activities such as ATP synthesis. Structural investigations of bc1 have been exceedingly successful, yielding atomic resolution structures of bc1 from various organisms and trapped in different reaction intermediates. These structures have confirmed and unified results of decades of experiments and have contributed to our understanding of the mechanism of bc1 functions as well as its inactivation by respiratory inhibitors.

The cytochrome bc 1 complex (bc 1 ) is a major contributor to the proton motive force across the membrane by coupling electron transfer to proton translocation. The crystal structures of wild type and mutant bc 1 complexes from the photosynthetic purple bacterium Rhodobacter sphaeroides (Rsbc 1 ), stabilized with the quinol oxidation (Q P ) site inhibitor stigmatellin alone or in combination with the quinone reduction (Q N ) site inhibitor antimycin, were determined. The high quality electron density permitted assignments of a new metal-binding site to the cytochrome c 1 subunit and a number of lipid and detergent molecules. Structural differences between Rsbc 1 and its mitochondrial counterparts are mostly extra membranous and provide a basis for understanding the function of the predominantly longer sequences in the bacterial subunits. Functional implications for the bc 1 complex are derived from analyses of 10 independent molecules in various crystal forms and from comparisons with mitochondrial complexes.A central component of the cellular respiratory chain is the cytochrome bc 1 complex (cyt bc 1 or bc 1 ) 2 that catalyzes the electron transfer (ET) from quinol to cytochrome c (cyt c) and simultaneously pumps protons across the membrane, contributing to the electrochemical potential that drives ATP synthesis and many other cellular activities (1). In chloroplasts and cyanobacteria a related membrane protein complex, the cytochrome b 6 f (cyt b 6 f), bridges photosystem I and II, enabling oxygenic photosynthesis and conversion of light energy into a proton gradient for ATP generation (2). For non-oxygenic photosynthetic bacteria, such as R. sphaeroides (Rs), which can grow both aerobically and photosynthetically under anaerobic condition, the bc 1 complex is involved in both growth modes; however it is essential only under anaerobic conditions (3).The critical importance of bc 1 has made it a target for numerous antibiotics, fungicides, and anti-parasitic agents. As a result, resistance to these agents has been documented in a wide variety of organisms (4 -8). Disorders that are related to defects in bc 1 complex are manifest clinically as mitochondrial myopathy (9), exercise intolerance (10), and Leber's optical neuropathy (11). Mounting evidence suggests a correlation between aging and the production of reactive oxygen species from defective bc 1 complexes (12, 13). The elucidation of the molecular mechanisms underlying these phenomena requires a combination of experimental approaches and in particular, structural investigations that can provide a molecular framework for further experiments.Significant advances in elucidating architectural features of this complex have been made by crystal structure determinations of mitochondrial bc 1 (14 -17) and b 6 f from a bacterium (18) and an alga (19). In particular, crystal structures of mitochondrial bc 1 in complex with various bc 1 inhibitors provide important mechanistic insights (20 -27), leading to a significant increase in the number of experimental studies...