Evaluation of different biological data and computational classification methods for use in protein interaction prediction

Protein–protein interactions play a key role in many biological systems. High‐throughput methods can directly detect the set of interacting proteins in yeast, but the results are often incomplete and exhibit high false‐positive and false‐negative rates. Recently, many different research groups independently suggested using supervised learning methods to integrate direct and indirect biological data sources for the protein interaction prediction task. However, the data sources, approaches, and implementations varied. Furthermore, the protein interaction prediction task itself can be subdivided into prediction of (1) physical interaction, (2) co‐complex relationship, and (3) pathway co‐membership. To investigate systematically the utility of different data sources and the way the data is encoded as features for predicting each of these types of protein interactions, we assembled a large set of biological features and varied their encoding for use in each of the three prediction tasks. Six different classifiers were used to assess the accuracy in predicting interactions, Random Forest (RF), RF similarity‐based k‐Nearest‐Neighbor, Naïve Bayes, Decision Tree, Logistic Regression, and Support Vector Machine. For all classifiers, the three prediction tasks had different success rates, and co‐complex prediction appears to be an easier task than the other two. Independently of prediction task, however, the RF classifier consistently ranked as one of the top two classifiers for all combinations of feature sets. Therefore, we used this classifier to study the importance of different biological datasets. First, we used the splitting function of the RF tree structure, the Gini index, to estimate feature importance. Second, we determined classification accuracy when only the top‐ranking features were used as an input in the classifier. We find that the importance of different features depends on the specific prediction task and the way they are encoded. Strikingly, gene expression is consistently the most important feature for all three prediction tasks, while the protein interactions identified using the yeast‐2‐hybrid system were not among the top‐ranking features under any condition. Proteins 2006. © 2006 Wiley‐Liss, Inc.     

Synapse development: still looking for the forest, still lost in the trees

Synapse development in the vertebrate central nervous system is a highly orchestrated process occurring not only during early stages of brain development, but also (to a lesser extent) in the mature nervous system. During development, the formation of synapses is intimately linked to the differentiation of neuronal cells, the extension of their axons and dendrites, and the course wiring of the nervous system. Subsequently, the stabilization, elimination, and strengthening of synaptic contacts is coupled to the refinement of axonal and dendritic arbors, to the establishment of functionally meaningful connections, and probably also to the day-to-day acquisition, storage, and retrieval of memories, higher order thought processes, and behavioral patterns.The authors acknowledge the support of the NIH (grant no. HD38760 DA016758) to C.C.G., the Ruth L. Kirchstein National Research Service Award (NRSA) to C.L.W., and the United States Israel Binational Science Foundation (grant no. 2003176) to C.C.G. and N.E.Z.     

Differential proteomic analysis of the Bacillus anthracis secretome: distinct plasmid and chromosome CO2-dependent cross talk mechanisms modulate extracellular proteolytic activities

The secretomes of a virulent Bacillus anthracis strain and of avirulent strains (cured of the virulence plasmids pXO1 and pXO2), cultured in rich and minimal media, were studied by a comparative proteomic approach. More than 400 protein spots, representing the products of 64 genes, were identified, and a unique pattern of protein relative abundance with respect to the presence of the virulence plasmids was revealed. In minimal medium under high CO(2) tension, conditions considered to simulate those encountered in the host, the presence of the plasmids leads to enhanced expression of 12 chromosome-carried genes (10 of which could not be detected in the absence of the plasmids) in addition to expression of 5 pXO1-encoded proteins. Furthermore, under these conditions, the presence of the pXO1 and pXO2 plasmids leads to the repression of 14 chromosomal genes. On the other hand, in minimal aerobic medium not supplemented with CO(2), the virulent and avirulent B. anthracis strains manifest very similar protein signatures, and most strikingly, two proteins (the metalloproteases InhA1 and NprB, orthologs of gene products attributed to the Bacillus cereus group PlcR regulon) represent over 90% of the total secretome. Interestingly, of the 64 identified gene products, at least 31 harbor features characteristic of virulence determinants (such as toxins, proteases, nucleotidases, sulfatases, transporters, and detoxification factors), 22 of which are differentially regulated in a plasmid-dependent manner. The nature and the expression patterns of proteins in the various secretomes suggest that distinct CO(2)-responsive chromosome- and plasmid-encoded regulatory factors modulate the secretion of potential novel virulence factors, most of which are associated with extracellular proteolytic activities.     

Characterization of the HIV N-terminal fusion peptide-containing region in context of key gp41 fusion conformations

Central to our understanding of human immunodeficiency virus-induced fusion is the high resolution structure of fragments of the gp41 fusion protein folded in a low energy core conformation. However, regions fundamental to fusion, like the fusion peptide (FP), have yet to be characterized in the context of the cognate protein regardless of its conformation. Based on conformation-specific monoclonal antibody recognition, we identified the polar region consecutive to the N36 fragment as a stabilizer of trimeric coiled-coil assembly, thereby enhancing inhibitory potency. This tertiary organization is retained in the context of the hydrophobic FP (N70 fragment). Our data indicate that the N70 fragment recapitulates the expected organization of this region in the viral fusion intermediate (N-terminal half of the pre-hairpin intermediate (N-PHI)), which happens to be the prime target for fusion inhibitors. Regarding the low energy conformation, we show for the first time core formation in the context of the FP (N70 core). The alpha-helical and coiled-coil stabilizing polar region confers substantial thermal stability to the core, whereas the hydrophobic FP does not add further stability. For the two key fusion conformations, N-PHI and N70 core, we find that the FP adopts a nonhelical structure and directs higher order assembly (assembly of coiled coils in N-PHI and assembly of bundles in the N70 core). This supra-molecular organization of coiled coils or folded cores is seen only in the context of the FP. This study is the first to characterize the FP region in the context of the folded core and provides a basic understanding of the role of the elusive FP for key gp41 fusion conformations.     

Inhibitors can activate proteases to catalyze the synthesis and hydrolysis of peptides

Competitive inhibitors can activate proteases (papain, trypsin, and cathepsin S) to catalyze the synthesis of peptide bonds and accelerate the hydrolysis of poor substrates (from 1 to 99%). Reaction mixtures contained intermediate molecules that were formed by the coupling of the inhibitor with the poor substrate. This and other findings suggest the following chain of events. Part of the binding energy of formation of the enzyme-inhibitor complex was used to activate the inhibitor, i.e., to form acyl-enzyme species with a high-energy bond (e.g., a thioester bond in the case of papain) required for coupling the inhibitor with the substrate to form the intermediate molecule. The latter was subjected to successive reactions which led to a stepwise degradation of the substrate, as well as to the regeneration of the inhibitor. One mole of the inhibitor could catalyze rapid hydrolysis of at least 53 mol of substrate. The intermediate molecules were the species undergoing rapid hydrolysis. Therefore, 1 mol of inhibitor was involved in the synthesis of 53 mol of intermediate molecules; i.e., the inhibitor functioned as a cofactor that catalyzed the synthesis of peptides. Thus, the binding energy of formation of the enzyme-inhibitor complex can be utilized to catalyze the synthesis of peptide bonds in the absence of an exogenous energy source (e.g., ATP).     

Herbicide-resistance conferred by expression of a catalytic antibody in Arabidopsis thaliana

Engineering herbicide resistance in crops facilitates control of weed species, particularly those that are closely related to the crop, and may be useful in selecting lines that have undergone multiple transformation events. Here we show that herbicide-resistant plants can be engineered by designing an herbicide and expressing a catalytic antibody that destroys the herbicide in planta. First, we developed a carbamate herbicide that can be catalytically destroyed by the aldolase antibody 38C2. This compound has herbicidal activity on all three plant species tested. Second, the light chain and half of the heavy chain (Fab) of the catalytic antibody were targeted to the endoplasmic reticulum in two classes of Arabidopsis thaliana transformants. Third, the two transgenic plants were crossed to produce an herbicide-resistant F1 hybrid. The in vitro catalytic activity of the protein from F1 hybrids corroborates that catalytic antibodies can be constitutively expressed in transgenic plants, and that they can confer a unique trait.     

The P1812A and P25T BRCA1 and the 5164del4 BRCA2 mutations: occurrence in high-risk non-Ashkenazi Jews

Founder mutations in the BRCA1 and BRCA2 genes have been discovered in the Ashkenazic Jewish population, but a founder mutation(s) has not been discovered among non-Ashkenazi Jews (NAJ). Two BRCA1 mutations (P1812A, P25T), and a BRCA2 mutation (5164del4) have been detected in NAJ high-risk families. We studied the prevalence of these three mutations in 270 high-risk NAJ families, including 85 from Iraq/Iran, 67 from North Africa, 27 from Yemen, 50 from the Balkan region, and 41 with mixed ancestry. The three mutations were detected only in individuals related to the original families. We conclude that the P1812A and P25T BRCA1 and 5164del4 BRCA2 mutations are not likely to be founder mutations in NAJ high-risk families. We also assessed the pathogenicity of the BRCA1 P1812A mutation in vitro using reporter gene assays in yeast and mammalian cells. We found that the BRCA1 P1812A variant activity assays yielded a slightly reduced reporter gene activity. Thus, there is some uncertainty as to the pathogenicity of BRCA1 P1812A.     

p53 and p21 regulate error-prone DNA repair to yield a lower mutation load

Regulation of mutation rates is critical for maintaining genome stability and controlling cancer risk. A special challenge to this regulation is the presence of multiple mutagenic DNA polymerases in mammals. These polymerases function in translesion DNA synthesis (TLS), an error-prone DNA repair process that involves DNA synthesis across DNA lesions. We found that in mammalian cells TLS is controlled by the tumor suppressor p53, and by the cell cycle inhibitor p21 via its PCNA-interacting domain, to maintain a low mutagenic load at the price of reduced repair efficiency. This regulation may be mediated by binding of p21 to PCNA and via DNA damage-induced ubiquitination of PCNA, which is stimulated by p53 and p21. Loss of this regulation by inactivation of p53 or p21 causes an out of control lesion-bypass activity, which increases the mutational load and might therefore play a role in pathogenic processes caused by genetic instability.     

PSMA PET/CT to evaluate response to SBRT for prostate cancer bone metastases

BackgroundIn the current study we evaluated ⁶⁸Ga PSMA PET/ CT to measure local control of bone metastasis in oligometastatic prostate cancer patients treated with SBRT.Materials and methodsAfter the institutional review board approval, a retrospective review of medical records of consecutive prostate cancer patients treated between 2014 and 2018 was conducted. Only medical records of patients that were treated with SBRT for bone metastasis and had pre-and post-SBRT ⁶⁸Ga PSMA PET/CT scans were included in our study. Data extracted from the medical files included patient-related (age), disease-related (Gleason score, site of metastasis), and treatment-related factors and outcomes.ResultsDuring the study period, a total of 12 patients (15 lesions) were included, with a median age of 73 years. The median follow-up was 26.5 months (range 13–45 months). Median time of ⁶⁸Ga PSMA PET/ CT follow up was 17.0 months (range 3–39 months). The median pre-treatment PSA was 2 ng/mL (range 0.56–44 ng/mL) vs. post treatment PSA nadir of 0.01 ng/mL (0.01–4.32) with a median time to nadir of 7 months (range, 2–12). Local control was 93% during the follow up period and there was correlation with PS MA avidity on PE T. None patients developed recurrences in the treated bone. None of the patients had grade 3 or more toxicities during follow-up.ConclusionsSBRT is a highly effective and safe method for treatment of prostate cancer bone metastases. More studies are required to determine if SBRT provides greater clinical benefit than standard fractionation for oligometastatic prostate cancer patients. ⁶⁸Ga PSMA PET/CT should be further investigated for delineation and follow-up.