Prediction of protein signal sequences and their cleavage sites by statistical rulers

Functioning as an "address tag" or "zip code" that guides nascent proteins (newly synthesized proteins in the cytosol) to wherever they are needed, signal peptides (also called targeting signals or signal sequences) have become a crucial tool in finding new drugs or reprogramming cells for gene therapy. To effectively and timely use such a tool, however, the first important thing is to develop an automated method for quickly and accurately identifying the signal peptide for a given nascent protein. With the avalanche of new protein sequences generated in the post-genomic era, the challenge has become even more urgent and critical. In this paper, five statistical rulers were derived via performing a mutual information analysis. By combining these statistical rulers, a new prediction algorithm was established and high success prediction rates were observed. The new algorithm may play a complementary role to the existing algorithms in this area. It is anticipated that the mutual information approach introduced here may be very useful for studying many other sequence-coupling problems in molecular biology as well.     

Distinctive properties of signal sequences from bacterial lipoproteins

We have compared a number of attributes (hydrophobicity, amino acid size, charge and secondary structure propensities) of signal sequences from bacterial lipoproteins with the same attributes of signal peptides from other prokaryotic proteins (non-lipoproteins). Lipoprotein leader sequences tend to be shorter, more hydrophobic and bulky, and they have stronger conformational preferences, the most conspicuous being a predicted beta-turn comprising positions 2 or 3 of the mature protein. Another distinctive feature is a maximum in the local energy profile between positions -1 and +2. With one exception (beta-lactamase III), the lipoproteins do not have Pro in their signal peptides, and they tend to have fewer Ser and Thr but more Gly than non-lipoproteins. Lipoproteins also lack a net negative charge in the N-terminal regions of the mature proteins. The signal peptides of the bacteriocin plasmid-coded lysis proteins appear to be unique in that they have all the ascribed features of lipoprotein signals; these characteristics can be used to guide signal peptide mutagenesis experiments and to construct new secretion vehicles.     

Use of synthetic signal sequences to explore the protein export machinery

The information for correct localization of newly synthesized proteins in both prokaryotes and eukaryotes resides in self-contained, often transportable targeting sequences. Of these, signal sequences specify that a protein should be secreted from a cell or incorporated into the cytoplasmic membrane. A central puzzle is presented by the lack of primary structural homology among signal sequences, although they share common features in their sequences. Synthetic signal peptides have enabled a wide range of studies of how these "zipcodes" for protein secretion are decoded and used to target proteins to the protein machinery that facilitates their translocation across and integration into membranes. We review research on how the information in signal sequences enables their passenger proteins to be correctly and efficiently localized. Synthetic signal peptides have made possible binding and crosslinking studies to explore how selectivity is achieved in recognition by the signal sequence-binding receptors, signal recognition particle, or SRP, which functions in all organisms, and SecA, which functions in prokaryotes and some organelles of prokaryotic origins. While progress has been made, the absence of atomic resolution structures for complexes of signal peptides and their receptors has definitely left many questions to be answered in the future.     

Signal-3L: A 3-layer approach for predicting signal peptides

Functioning as an "address tag" that directs nascent proteins to their proper cellular and extracellular locations, signal peptides have become a crucial tool in finding new drugs or reprogramming cells for gene therapy. To effectively and timely use such a tool, however, the first important thing is to develop an automated method for rapidly and accurately identifying the signal peptide for a given nascent protein. With the avalanche of new protein sequences generated in the post-genomic era, the challenge has become even more urgent and critical. In this paper, we have developed a novel method for predicting signal peptide sequences and their cleavage sites in human, plant, animal, eukaryotic, Gram-positive, and Gram-negative protein sequences, respectively. The new predictor is called Signal-3L that consists of three prediction engines working, respectively, for the following three progressively deepening layers: (1) identifying a query protein as secretory or non-secretory by an ensemble classifier formed by fusing many individual OET-KNN (optimized evidence-theoretic K nearest neighbor) classifiers operated in various dimensions of PseAA (pseudo amino acid) composition spaces; (2) selecting a set of candidates for the possible signal peptide cleavage sites of a query secretory protein by a subsite-coupled discrimination algorithm; (3) determining the final cleavage site by fusing the global sequence alignment outcome for each of the aforementioned candidates through a voting system. Signal-3L is featured by high success prediction rates with short computational time, and hence is particularly useful for the analysis of large-scale datasets. Signal-3L is freely available as a web-server at http://chou.med.harvard.edu/bioinf/Signal-3L/ or http://202.120.37.186/bioinf/Signal-3L, where, to further support the demand of the related areas, the signal peptides identified by Signal-3L for all the protein entries in Swiss-Prot databank that do not have signal peptide annotations or are annotated with uncertain terms but are classified by Signal-3L as secretory proteins are provided in a downloadable file. The large-scale file is prepared with Microsoft Excel and named "Tab-Signal-3L.xls", and will be updated once a year to include new protein entries and reflect the continuous development of Signal-3L.     

Genetic screen for signal peptides in Hydra reveals novel secreted proteins and evidence for non-classical protein secretion

We have screened a Hydra cDNA library for sequences encoding N-terminal signal peptides using the yeast invertase secretion vector pSUC [Jacobs et al., 1997. A genetic selection for isolating cDNAs encoding secreted proteins. Gene 198, 289-296]. We isolated and sequenced 907 positive clones; 88% encoded signal peptides; 12% lacked signal peptides. By searching the Hydra EST database we identified full-length sequences for the selected clones. These encoded 37 known proteins with signal peptides and 40 novel Hydra-specific proteins with signal peptides. Localization of two signal peptide-containing sequences, VEGF and ferritin, to the secretory pathway was confirmed with GFP fusion proteins. In addition, we isolated 105 clones which lacked signal peptides but which supported invertase secretion from yeast. Isolation of plasmids from these clones and retransformation in invertase-negative yeast cells confirmed the phenotype. A GFP fusion protein of one such clone encoding the foot morphogen pedibin was localized to the cytoplasm in transfected Hydra cells and did not enter the ER/Golgi secretory pathway. Secretion of pedibin and other proteins lacking signal peptides appears to occur by a non-classical protein secretion route.     

Genetic screen for signal peptides in Hydra reveals novel secreted proteins and evidence for non-classical protein secretion

We have screened a Hydra cDNA library for sequences encoding N-terminal signal peptides using the yeast invertase secretion vector pSUC [Jacobs et al., 1997. A genetic selection for isolating cDNAs encoding secreted proteins. Gene 198, 289–296]. We isolated and sequenced 907 positive clones; 88% encoded signal peptides; 12% lacked signal peptides. By searching the Hydra EST database we identified full-length sequences for the selected clones. These encoded 37 known proteins with signal peptides and 40 novel Hydra-specific proteins with signal peptides. Localization of two signal peptide-containing sequences, VEGF and ferritin, to the secretory pathway was confirmed with GFP fusion proteins. In addition, we isolated 105 clones which lacked signal peptides but which supported invertase secretion from yeast. Isolation of plasmids from these clones and retransformation in invertase-negative yeast cells confirmed the phenotype. A GFP fusion protein of one such clone encoding the foot morphogen pedibin was localized to the cytoplasm in transfected Hydra cells and did not enter the ER/Golgi secretory pathway. Secretion of pedibin and other proteins lacking signal peptides appears to occur by a non-classical protein secretion route.     

Signal-sequence Trap in Mammalian and Yeast Cells: A Comparison

Membrane associated and secreted proteins are translated as precursors containing a signal peptide that allows protein-insertion into the membrane of the endoplasmic reticulum and is co-translationally removed in the lumen. The ability of the signal peptide to direct a polypeptide into the secretory pathway is exploited in methods developed to select cDNAs encoding such proteins. Different strategies are known in which cDNA libraries can be screened for signal peptides by the ability of the latter to rescue the translocation of signal sequence-less proteins. In one method, a cDNA library is tested for interleukin 2 receptor α chain translocation to the membrane in COS cells, in another one for invertase secretion from yeast. In this work, we compared the two systems by testing six mouse signal peptides in COS and yeast cells. All of them were functional in the mammalian system, whereas only three of them in yeast. Two other sequences needed the 5′ cDNA sequence flanking the ATG codon to be removed in order to enable protein translocation. Although the structure of signal sequences and the functioning of the secretory machinery are well conserved from prokaryotes to eukaryotes, it seems evident that not all signal peptides can be interchanged between different proteins and organisms. In particular, signal peptides that are functional in the mammalian system do not necessarily lead to protein translocation in yeast. Received: 9 March 2001     

Prediction of signal peptides in protein sequences by neural networks

We present here a neural network-based method for detection of signal peptides (abbreviation used: SP) in proteins. The method is trained on sequences of known signal peptides extracted from the Swiss-Prot protein database and is able to work separately on prokaryotic and eukaryotic proteins. A query protein is dissected into overlapping short sequence fragments, and then each fragment is analyzed with respect to the probability of it being a signal peptide and containing a cleavage site. While the accuracy of the method is comparable to that of other existing prediction tools, it provides a significantly higher speed and portability. The accuracy of cleavage site prediction reaches 73% on heterogeneous source data that contains both prokaryotic and eukaryotic sequences while the accuracy of discrimination between signal peptides and non-signal peptides is above 93% for any source dataset. As a consequence, the method can be easily applied to genome-wide datasets. The software can be downloaded freely from http://rpsp.bioinfo.pl/RPSP.tar.gz.     

一种表征蛋白质可分泌性的结构融合度特征

选择适宜的信号肽是实现外源蛋白高效分泌表达的一个重要因素。本研究利用生物信息学方法分析信号肽与外源蛋白之间的相容程度,将其定义为结构融合度,并从数学角度分析拼接信号肽与目的蛋白邻近残基之间的相互作用,提出了信号肽拼接区域与目标蛋白之间的数学模型,利用该模型进行结构融合度特征提取,以此来表征外源蛋白质的可分泌性。模拟结果显示结构融合度特征能有效区分枯草芽孢杆菌宿主的可分泌和不可分泌蛋白。研究结果有助于信号肽的选择,对目的蛋白分泌表达的优化具有一定的指导意义。     

Computational differentiation of N-terminal signal peptides and transmembrane helices

Signal peptides and transmembrane helices both contain a stretch of hydrophobic amino acids. This common feature makes it difficult for signal peptide and transmembrane helix predictors to correctly assign identity to stretches of hydrophobic residues near the N-terminal methionine of a protein sequence. The inability to reliably distinguish between N-terminal transmembrane helix and signal peptide is an error with serious consequences for the prediction of protein secretory status or transmembrane topology. In this study, we report a new method for differentiating protein N-terminal signal peptides and transmembrane helices. Based on the sequence features extracted from hydrophobic regions (amino acid frequency, hydrophobicity, and the start position), we set up discriminant functions and examined them on non-redundant datasets with jackknife tests. This method can incorporate other signal peptide prediction methods and achieve higher prediction accuracy. For Gram-negative bacterial proteins, 95.7% of N-terminal signal peptides and transmembrane helices can be correctly predicted (coefficient 0.90). Given a sensitivity of 90%, transmembrane helices can be identified from signal peptides with a precision of 99% (coefficient 0.92). For eukaryotic proteins, 94.2% of N-terminal signal peptides and transmembrane helices can be correctly predicted with coefficient 0.83. Given a sensitivity of 90%, transmembrane helices can be identified from signal peptides with a precision of 87% (coefficient 0.85). The method can be used to complement current transmembrane protein prediction and signal peptide prediction methods to improve their prediction accuracies.     

Profiling the Cross Reactivity of Ubiquitin with the Nedd8 Activating Enzyme by Phage Display

The C-terminal peptides of ubiquitin (UB) and UB-like proteins (UBLs) play a key role in their recognition by the specific activating enzymes (E1s) to launch their transfer through the respective enzymatic cascades thus modifying cellular proteins. UB and Nedd8, a UBL regulating the activity of cullin-RING UB ligases, only differ by one residue at their C-termini; yet each has its specific E1 for the activation reaction. It has been reported recently that UAE can cross react with Nedd8 to enable its passage through the UB transfer cascade for protein neddylation. To elucidate differences in UB recognition by UAE and NAE, we carried out phage selection of a UB library with randomized C-terminal sequences based on the catalytic formation of UB∼NAE thioester conjugates. Our results confirmed the previous finding that residue 72 of UB plays a “gate-keeping” role in E1 selectivity. We also found that diverse sequences flanking residue 72 at the UB C-terminus can be accommodated by NAE for activation. Furthermore heptameric peptides derived from the C-terminal sequences of UB variants selected for NAE activation can function as mimics of Nedd8 to form thioester conjugates with NAE and the downstream E2 enzyme Ubc12 in the Nedd8 transfer cascade. Once the peptides are charged onto the cascade enzymes, the full-length Nedd8 protein is effectively blocked from passing through the cascade for the critical modification of cullin. We have thus identified a new class of inhibitors of protein neddylation based on the profiles of the UB C-terminal sequences recognized by NAE.     

Amino-terminal amino acid sequences of electron transfer proteins from Gram-negative bacteria as indicators of their cellular localization: the sulfate-reducing bacteria

Abstract Data are presented that all known periplasmic redox proteins from the sulfate reducing bacteria included in the genus, Desulfovibrio have aminoterminal (N-terminal) amino-acid sequences commonly found in other Gram-negative bacteria and are indicative of recognition sites for signal peptides. In contrast, none of the cytoplasmic redox proteins exhibited these unique N-terminal amino-acid sequences. It is proposed that the N-terminal amino-acid residues of a given protein can be used as an indicator of its cellular localization within the bacterial cell.     

Prediction of Cell-Penetrating Peptides

Cell-penetrating peptides, CPPs, are used as delivery vectors for pharmacologically interesting substances, such as antisense oligonucleotides, proteins and peptides. We present a general principle for designing cell-penetrating peptides derived from naturally occurring proteins as well as from randomly generated polyamino acid sequences. Thereby, we introduce a novel pharmacological principle for identification of cell-penetrating peptides for which the applications can be numerous, including cellular transduction vectors and mimics of intracellular protein–protein interactions. The methods of identifying a CPP comprises assessing the averaged bulk property values of the defined sequence, and ensuring that they fall within the bulk property value interval obtained from the training set. Despite this simplistic approach, the search criteria proved useful for finding CPP properties in either proteins or random sequences. We have experimentally verified cell-penetrating properties of 10–20-mer peptides derived from naturally occurring proteins as well as from random poly-amino acids. We note that since CPPs can be found in part of the protein sequences that may govern protein interactions, it is possible to produce cell-penetrating protein agonists or antagonists.     

Evaluating signal peptide prediction methods for Gram-positive bacteria

Gram-positive bacteria have been widely investigated for their huge capability to secrete proteins, such as those involved in gene expression, bacterial surface display and bacterial pathogenesis. The N-terminal signal peptide of a secretory protein is responsible for the translocation of polypeptide through the cytoplasmic membrane. Recently, the signal peptide prediction has become a major task in bioinformatics, and many programs with different algorithms were developed to predict signal peptides. In this paper, five prediction programs (SignalP 3.0, PrediSi, Phobius, SOSUIsignal and SIG-Pred) were selected to evaluate their prediction accuracy for signal peptides and cleavage site using 509 unbiased and experimentally verified Gram-positive protein sequences. The results showed that SignalP was the most accurate program in signal peptide (96% accuracy) and cleavage site (83%) prediction. Prediction performance could further be improved by combining multiple methods into consensus prediction, which would increase the accuracy to 98%, and decrease the false positive to zero. When the consensus method was used to predict Bacillus’s extracellular proteins identified by proteomics, more new signal peptides were successfully identified. It could be concluded that the consensus method would be useful to make prediction of signal peptides more reliable.     

Conformational analysis of invariant peptide sequences in bacterial genomes

The functional significance of evolutionarily conserved motifs/patterns of short regions in proteins is well documented. Although a large number of sequences are conserved, only a small fraction of these are invariant across several organisms. Here, we have examined the structural features of the functionally important peptide sequences, which have been found invariant across diverse bacterial genera. Ramachandran angles (phi,psi) have been used to analyze the conformation, folding patterns and geometrical location (buried/exposed) of these invariant peptides in different crystal structures harboring these sequences. The analysis indicates that the peptides preferred a single conformation in different protein structures, with the exception of only a few longer peptides that exhibited some conformational variability. In addition, it is noticed that the variability of conformation occurs mainly due to flipping of peptide units about the virtual C(alpha)...C(alpha) bond. However, for a given invariant peptide, the folding patterns are found to be similar in almost all the cases. Over and above, such peptides are found to be buried in the protein core. Thus, we can safely conclude that these invariant peptides are structurally important for the proteins, since they acquire unique structures across different proteins and can act as structural determinants (SD) of the proteins. The location of these SD peptides on the protein chain indicated that most of them are clustered towards the N-terminal and middle region of the protein with the C-terminal region exhibiting low preference. Another feature that emerges out of this study is that some of these SD peptides can also play the roles of "fold boundaries" or "hinge nucleus" in the protein structure. The study indicates that these SD peptides may act as chain-reversal signatures, guiding the proteins to adopt appropriate folds. In some cases the invariant signature peptides may also act as folding nuclei (FN) of the proteins.     

Intracellullar peptides as putative natural regulators of protein interactions

Extralysosomal proteolysis by multicatalytic complexes such as the 26S proteasome produces large amounts of peptides in the cytosol, mitochondria and nuclei of eukaryotic cells, and there is increasing evidence that the resulting free intracellular peptides can modulate specific protein interactions. The demonstration that free peptides added to the intracellular milieu can regulate cellular functions mediated by protein interactions suggests new putative roles for these molecules in gene regulation, metabolism, cell signaling and protein targeting. Such interactions frequently involve specific consensus amino acid sequences that can be predicted based on similarities in domain composition. We have recently developed a new strategy for identifying novel natural peptides, the sequences of which correspond to fragments of intracellular proteins and contain putative post-translational modification sites. In this review, we examine the evidence that intracellular peptides released by proteasomes may be involved in regulating protein interactions. In particular, the role of endopeptidase 24.15 (thimet oligopeptidase; EC 3.4.24.15) is discussed in detail as this enzyme has been implicated in intracellular peptide metabolism in vivo in concert with the 26S proteasome.     

Peptides as probes for G protein signal transduction

Triggered by agonist binding to cell surface receptors, the heterotrimeric G proteins dissociate into and βγ subunits, each activating distinct second messenger pathways. Peptides from the primary sequences of receptors, G proteins, and effectors have been used to study the molecular interactions between these proteins. Receptor-derived peptides from the second, third and fourth intracellular loops and certain naturally occurring peptides antagonize G protein interactions and can directly activate G protein. These peptides bind to G protein sites that include the N and C terminal regions of the subunit and a yet to be identified region of the β subunit. Peptides have also been useful in characterizing G protein-effector interactions. The identification of the contact sites between proteins involved in G protein signal transduction should aid in the development of non-peptide mimetic therapeutics which could specifically modify G protein-mediated cellular responses.     

A genetic screen to identify sequences that mediate protein oligomerization in Escherichia coli

Many proteins assemble as oligomeric complexes and in several cases a distinct domain mediates the interaction between the subunits. The identification of new oligomerization modules is relevant to comprehend both the architecture and the evolution of protein sequences and also for protein engineering applications. Using the bacteriophage lambda repressor dimerization assay, we searched Escherichia coli genomic libraries for sequences able to mediate protein oligomerization in vivo. We identified short peptides that can substitute very effectively the dimerizing domain of the repressor. Most of these peptides belong to open reading frames that are normally not expressed in the bacterial cell.     

Efficient display of two enzymes on filamentous phage using an improved signal sequence

Directed protein-evolution strategies generally make use of a link between a protein and the encoding DNA. In phage-display technology, this link is provided by fusion of the protein with a coat protein that is incorporated into the phage particle containing the DNA. Optimization of this link can be achieved by adjusting the signal sequence of the fusion. In a previous study, directed evolution of signal sequences for optimal display of the Taq DNA polymerase I Stoffel fragment on phage yielded signal peptides with a 50-fold higher incorporation of fusion proteins in phage particles. In this article, we show that for one of the selected signal sequences, improved display on phage can be generalized to other proteins, such as adenylate cyclases from Escherichia coli and Bordetella pertussis, and that this is highly dependent on short sequences at the C-terminus of the signal peptide. Further, the display of two enzymes on phage has been achieved and may provide a strategy for directing coevolution of the two proteins. These findings should be useful for display of large and cytoplasmic proteins on filamentous phage.     

Proteomics-based consensus prediction of protein retention in a bacterial membrane

The availability of complete bacterial genome sequences allows proteome-wide predictions of exported proteins that are potentially retained in the cytoplasmic membranes of the corresponding organisms. In practice, however, major problems are encountered with the computer-assisted distinction between (Sec-type) signal peptides that direct exported proteins into the growth medium and lipoprotein signal peptides or amino-terminal membrane anchors that cause protein retention in the membrane. In the present studies, which were aimed at improving methods to predict protein retention in the bacterial cytoplasmic membrane, we have compared sets of membrane-attached and extracellular proteins of Bacillus subtilis that were recently identified through proteomics approaches. The results showed that three classes of membrane-attached proteins can be distinguished. Two classes include 43 lipoproteins and 48 proteins with an amino-terminal transmembrane segment, respectively. Remarkably, a third class includes 31 proteins that remain membrane-retained despite the presence of typical Sec-type signal peptides with consensus signal peptidase recognition sites. This unprecedented finding indicates that unknown mechanisms are involved in membrane retention of this class of proteins. A further novelty is a consensus sequence indicative for release of certain lipoproteins from the membrane by proteolytic shaving. Finally, using non-overlapping sets of secreted and membrane-retained proteins, the accuracy of different signal peptide prediction algorithms was assessed. Accuracy for the prediction of protein retention in the membrane was increased to 82% using a majority-vote approach. Our findings provide important leads for future identification of surface proteins from pathogenic bacteria, which are attractive candidate infection markers and potential targets for drugs or vaccines.