Mapping the cellular response to small molecules using chemogenomic fitness signatures

Predicting chemotherapeutic drug combinations through gene network profiling

Contemporary chemotherapeutic treatments incorporate the use of several agents in combination. However, selecting the most appropriate drugs for such therapy is not necessarily an easy or straightforward task. Here, we describe a targeted approach that can facilitate the reliable selection of chemotherapeutic drug combinations through the interrogation of drug-resistance gene networks. Our method employed single-cell eukaryote fission yeast (Schizosaccharomyces pombe) as a model of proliferating cells to delineate a drug resistance gene network using a synthetic lethality workflow. Using the results of a previous unbiased screen, we assessed the genetic overlap of doxorubicin with six other drugs harboring varied mechanisms of action. Using this fission yeast model, drug-specific ontological sub-classifications were identified through the computation of relative hypersensitivities. We found that human gastric adenocarcinoma cells can be sensitized to doxorubicin by concomitant treatment with cisplatin, an intra-DNA strand crosslinking agent, and suberoylanilide hydroxamic acid, a histone deacetylase inhibitor. Our findings point to the utility of fission yeast as a model and the differential targeting of a conserved gene interaction network when screening for successful chemotherapeutic drug combinations for human cells.

A robust platform for chemical genomics in bacterial systems

A robust and sensitive platform was developed for chemical-genomics in bacteria. Kinetic acquisitions of colony growth enable calculation of growth rates alongside conventional endpoint volume measurements, generating a wealth of chemical-genetic interactions. This kinetic platform is highly amenable to prokaryotic or eukaryotic strain collections.

While genetic perturbation has been the conventional route to probing bacterial systems, small molecules are showing great promise as probes for cellular complexity. Indeed, systematic investigations of chemical-genetic interactions can provide new insights into cell networks and are often starting points for understanding the mechanism of action of novel chemical probes. We have developed a robust and sensitive platform for chemical-genomic investigations in bacteria. The approach monitors colony volume kinetically using transmissive scanning measurements, enabling acquisition of growth rates and conventional endpoint measurements. We found that chemical-genomic profiles were highly sensitive to concentration, necessitating careful selection of compound concentrations. Roughly 20,000,000 data points were collected for 15 different antibiotics. While 1052 chemical-genetic interactions were identified using the conventional endpoint biomass approach, adding interactions in growth rate resulted in 1564 interactions, a 50–200% increase depending on the drug, with many genes uncharacterized or poorly annotated. The chemical-genetic interaction maps generated from these data reveal common genes likely involved in multidrug resistance. Additionally, the maps identified deletion backgrounds exhibiting class-specific potentiation, revealing conceivable targets for combination approaches to drug discovery. This open platform is highly amenable to kinetic screening of any arrayable strain collection, be it prokaryotic or eukaryotic.

A Signaling Lipid Associated with Alzheimer's Disease Promotes Mitochondrial Dysfunction

Fundamental changes in the composition and distribution of lipids within the brain are believed to contribute to the cognitive decline associated with Alzheimer’s disease (AD). The mechanisms by which these changes in lipid composition affect cellular function and ultimately cognition are not well understood. Although “candidate gene” approaches can provide insight into the effects of dysregulated lipid metabolism they require a preexisting understanding of the molecular targets of individual lipid species. In this report we combine unbiased gene expression profiling with a genome-wide chemogenomic screen to identify the mitochondria as an important downstream target of PC(O-16:0/2:0), a neurotoxic lipid species elevated in AD. Further examination revealed that PC(O-16:0/2:0) similarly promotes a global increase in ceramide accumulation in human neurons which was associated with mitochondrial-derived reactive oxygen species (ROS) and toxicity. These findings suggest that PC(O-16:0/2:0)-dependent mitochondrial dysfunction may be an underlying contributing factor to the ROS production associated with AD.

Chromosome-wide histone deacetylation by sirtuins prevents hyperactivation of DNA damage-induced signaling upon replicative stress

The Saccharomyces cerevisiae genome encodes five sirtuins (Sir2 and Hst1-4), which constitute a conserved family of NAD-dependent histone deacetylases. Cells lacking any individual sirtuin display mild growth and gene silencing defects. However, hst3Δ hst4Δ double mutants are exquisitely sensitive to genotoxins, and hst3Δ hst4Δ sir2Δmutants are inviable. Our published data also indicate that pharmacological inhibition of sirtuins prevents growth of several fungal pathogens, although the biological basis is unclear. Here, we present genome-wide fitness assays conducted with nicotinamide (NAM), a pan-sirtuin inhibitor. Our data indicate that NAM treatment causes yeast to solicit specific DNA damage response pathways for survival, and that NAM-induced growth defects are mainly attributable to inhibition of Hst3 and Hst4 and consequent elevation of histone H3 lysine 56 acetylation (H3K56ac). Our results further reveal that in the presence of constitutive H3K56ac, the Slx4 scaffolding protein and PP4 phosphatase complex play essential roles in preventing hyperactivation of the DNA damage-response kinase Rad53 in response to spontaneous DNA damage caused by reactive oxygen species. Overall, our data support the concept that chromosome-wide histone deacetylation by sirtuins is critical to mitigate growth defects caused by endogenous genotoxins.

Select microtubule inhibitors increase lysosome acidity and promote lysosomal disruption in acute myeloid leukemia (AML) cells

To identify new biological vulnerabilities in acute myeloid leukemia, we screened a library of natural products for compounds cytotoxic to TEX leukemia cells. This screen identified the novel small molecule Deoxysappanone B 7,4' dimethyl ether (Deox B 7,4), which possessed nanomolar anti-leukemic activity. To determine the anti-leukemic mechanism of action of Deox B 7,4, we conducted a genome-wide screen in Saccharomyces cerevisiae and identified enrichment of genes related to mitotic cell cycle as well as vacuolar acidification, therefore pointing to microtubules and vacuolar (V)-ATPase as potential drug targets. Further investigations into the mechanisms of action of Deox B 7,4 and a related analogue revealed that these compounds were reversible microtubule inhibitors that bound near the colchicine site. In addition, Deox B 7,4 and its analogue increased lysosomal V-ATPase activity and lysosome acidity. The effects on microtubules and lysosomes were functionally important for the anti-leukemic effects of these drugs. The lysosomal effects were characteristic of select microtubule inhibitors as only the Deox compounds and nocodazole, but not colchicine, vinca alkaloids or paclitaxel, altered lysosome acidity and induced lysosomal disruption. Thus, our data highlight a new mechanism of action of select microtubule inhibitors on lysosomal function.

Prediction of Synergism from Chemical-Genetic Interactions by Machine Learning

The structure of genetic interaction networks predicts that, analogous to synthetic lethal interactions between non-essential genes, combinations of compounds with latent activities may exhibit potent synergism. To test this hypothesis, we generated a chemical-genetic matrix of 195 diverse yeast deletion strains treated with 4,915 compounds. This approach uncovered 1,221 genotype-specific inhibitors, which we termed cryptagens. Synergism between 8,128 structurally disparate cryptagen pairs was assessed experimentally and used to benchmark predictive algorithms. A model based on the chemical-genetic matrix and the genetic interaction network failed to accurately predict synergism. However, a combined random forest and Naive Bayesian learner that associated chemical structural features with genotype-specific growth inhibition had strong predictive power. This approach identified previously unknown compound combinations that exhibited species-selective toxicity toward human fungal pathogens. This work demonstrates that machine learning methods trained on unbiased chemical-genetic interaction data may be widely applicable for the discovery of synergistic combinations in different species.

The Toxicity of a Novel Antifungal Compound Is Modulated by Endoplasmic Reticulum-Associated Protein Degradation Components

In a search for new antifungal compounds, we screened a library of 4,454 chemicals for toxicity against the human fungal pathogen Aspergillus fumigatus. We identified sr7575, a molecule that inhibits growth of the evolutionary distant fungi A. fumigatus, Cryptococcus neoformans, Candida albicans, and Saccharomyces cerevisiae but lacks acute toxicity for mammalian cells. To gain insight into the mode of inhibition, sr7575 was screened against 4,885 S. cerevisiae mutants from the systematic collection of haploid deletion strains and 977 barcoded haploid DAmP (decreased abundance by mRNA perturbation) strains in which the function of essential genes was perturbed by the introduction of a drug resistance cassette downstream of the coding sequence region. Comparisons with previously published chemogenomic screens revealed that the set of mutants conferring sensitivity to sr7575 was strikingly narrow, affecting components of the endoplasmic reticulum-associated protein degradation (ERAD) stress response and the ER membrane protein complex (EMC). ERAD-deficient mutants were hypersensitive to sr7575 in both S. cerevisiae and A. fumigatus, indicating a conserved mechanism of growth inhibition between yeast and filamentous fungi. Although the unfolded protein response (UPR) is linked to ERAD regulation, sr7575 did not trigger the UPR in A. fumigatus and UPR mutants showed no enhanced sensitivity to the compound. The data from this chemogenomic analysis demonstrate that sr7575 exerts its antifungal activity by disrupting ER protein quality control in a manner that requires ERAD intervention but bypasses the need for the canonical UPR. ER protein quality control is thus a specific vulnerability of fungal organisms that might be exploited for antifungal drug development.

The alternate AP-1 adaptor subunit Apm2 interacts with the Mil1 regulatory protein and confers differential cargo sorting

Adaptor complexes are important for cargo sorting in clathrin-coated vesicles. The µ adaptor subunits Apm1 and Apm2 create functionally distinct versions of the yeast AP-1 complex. A novel regulatory protein is identified that selectively binds Apm2-containing complexes and contributes to their membrane recruitment.

Heterotetrameric adaptor protein complexes are important mediators of cargo protein sorting in clathrin-coated vesicles. The cell type–specific expression of alternate μ chains creates distinct forms of AP-1 with altered cargo sorting, but how these subunits confer differential function is unclear. Whereas some studies suggest the μ subunits specify localization to different cellular compartments, others find that the two forms of AP-1 are present in the same vesicle but recognize different cargo. Yeast have two forms of AP-1, which differ only in the μ chain. Here we show that the variant μ chain Apm2 confers distinct cargo-sorting functions. Loss of Apm2, but not of Apm1, increases cell surface levels of the v-SNARE Snc1. However, Apm2 is unable to replace Apm1 in sorting Chs3, which requires a dileucine motif recognized by the γ/σ subunits common to both complexes. Apm2 and Apm1 colocalize at Golgi/early endosomes, suggesting that they do not associate with distinct compartments. We identified a novel, conserved regulatory protein that is required for Apm2-dependent sorting events. Mil1 is a predicted lipase that binds Apm2 but not Apm1 and contributes to its membrane recruitment. Interactions with specific regulatory factors may provide a general mechanism to diversify the functional repertoire of clathrin adaptor complexes.

Systematic identification and correction of annotation errors in the genetic interaction map of Saccharomyces cerevisiae

The yeast mutant collections are a fundamental tool in deciphering genomic organization and function. Over the last decade, they have been used for the systematic exploration of ∼6 000 000 double gene mutants, identifying and cataloging genetic interactions among them. Here we studied the extent to which these data are prone to neighboring gene effects (NGEs), a phenomenon by which the deletion of a gene affects the expression of adjacent genes along the genome. Analyzing ∼90,000 negative genetic interactions observed to date, we found that more than 10% of them are incorrectly annotated due to NGEs. We developed a novel algorithm, GINGER, to identify and correct erroneous interaction annotations. We validated the algorithm using a comparative analysis of interactions from Schizosaccharomyces pombe. We further showed that our predictions are significantly more concordant with diverse biological data compared to their mis-annotated counterparts. Our work uncovered about 9500 new genetic interactions in yeast.

Functional genomics to uncover drug mechanism of action

The upswing in US Food and Drug Administration and European Medicines Agency drug approvals in 2014 may have marked an end to the dry spell that has troubled the pharmaceutical industry over the past decade. Regardless, the attrition rate of drugs in late clinical phases remains high, and a lack of target validation has been highlighted as an explanation. This has led to a resurgence in appreciation of phenotypic drug screens, as these may be more likely to yield compounds with relevant modes of action. However, cell-based screening approaches do not directly reveal cellular targets, and hence target deconvolution and a detailed understanding of drug action are needed for efficient lead optimization and biomarker development. Here, recently developed functional genomics technologies that address this need are reviewed. The approaches pioneered in model organisms, particularly in yeast, and more recently adapted to mammalian systems are discussed. Finally, areas of particular interest and directions for future tool development are highlighted.

An Antifungal Combination Matrix Identifies a Rich Pool of Adjuvant Molecules that Enhance Drug Activity against Diverse Fungal Pathogens

There is an urgent need to identify new treatments for fungal infections. By combining sub-lethal concentrations of the known antifungals fluconazole, caspofungin, amphotericin B, terbinafine, benomyl, and cyprodinil with ∼3,600 compounds in diverse fungal species, we generated a deep reservoir of chemical-chemical interactions termed the Antifungal Combinations Matrix (ACM). Follow-up susceptibility testing against a fluconazole-resistant isolate of C. albicans unveiled ACM combinations capable of potentiating fluconazole in this clinical strain. We used chemical genetics to elucidate the mode of action of the antimycobacterial drug clofazimine, a compound with unreported antifungal activity that synergized with several antifungals. Clofazimine induces a cell membrane stress for which the Pkc1 signaling pathway is required for tolerance. Additional tests against additional fungal pathogens, including Aspergillus fumigatus, highlighted that clofazimine exhibits efficacy as a combination agent against multiple fungi. Thus, the ACM is a rich reservoir of chemical combinations with therapeutic potential against diverse fungal pathogens.

Identification of a Mitochondrial DNA Polymerase Affecting Cardiotoxicity of Sunitinib Using a Genome-Wide Screening on S. pombe Deletion Library

Drug toxicity is a key issue for drug R&D, a fundamental challenge of which is to screen for the targets genome-wide. The anticancer tyrosine kinase inhibitor sunitinib is known to induce cardiotoxicity. Here, to understand the molecular insights of cardiotoxicity by sunitinib at the genome level, we used a genome-wide drug target screening technology (GPScreen) that measures drug-induced haploinsufficiency (DIH) in the fission yeast Schizosaccharomyces pombe genome-wide deletion library and found a mitochondrial DNA polymerase (POG1). In the results, sunitinib induced more severe cytotoxicity and mitochondrial damage in POG1-deleted heterozygous mutants compared to wild type (WT) of S. pombe. Furthermore, knockdown of the human ortholog POLG of S. pombe POG1 in human cells significantly increased the cytotoxicity of sunitinib. Notably, sunitinib dramatically decreased the levels of POLG mRNAs and proteins, of which downregulation was already known to induce mitochondrial damage of cardiomyocytes, causing cardiotoxicity. These results indicate that POLG might play a crucial role in mitochondrial damage as a gene of which expressional pathway is targeted by sunitinib for cardiotoxicity, and that genome-wide drug target screening with GPScreen can be applied to drug toxicity target discovery to understand the molecular insights regarding drug toxicity.

Biological data analysis as an information theory problem: multivariable dependence measures and the shadows algorithm

Information theory is valuable in multiple-variable analysis for being model-free and nonparametric, and for the modest sensitivity to undersampling. We previously introduced a general approach to finding multiple dependencies that provides accurate measures of levels of dependency for subsets of variables in a data set, which is significantly nonzero only if the subset of variables is collectively dependent. This is useful, however, only if we can avoid a combinatorial explosion of calculations for increasing numbers of variables.

 The proposed dependence measure for a subset of variables,τ, differential interaction information, Δ(τ), has the property that for subsets ofτ some of the factors of Δ(τ) are significantly nonzero, when the full dependence includes more variables. We use this property to suppress the combinatorial explosion by following the “shadows” of multivariable dependency on smaller subsets. Rather than calculating the marginal entropies of all subsets at each degree level, we need to consider only calculations for subsets of variables with appropriate “shadows.” The number of calculations for n variables at a degree level of d grows therefore, at a much smaller rate than the binomial coefficient (n, d), but depends on the parameters of the “shadows” calculation. This approach, avoiding a combinatorial explosion, enables the use of our multivariable measures on very large data sets. We demonstrate this method on simulated data sets, and characterize the effects of noise and sample numbers. In addition, we analyze a data set of a few thousand mutant yeast strains interacting with a few thousand chemical compounds.

Beyond Agar: Gel Substrates with Improved Optical Clarity and Drug Efficiency and Reduced Autofluorescence for Microbial Growth Experiments

Agar, a seaweed extract, has been the standard support matrix for microbial experiments for over a century. Recent developments in high-throughput genetic screens have created a need to reevaluate the suitability of agar for use as colony support, as modern robotic printing systems now routinely spot thousands of colonies within the area of a single microtiter plate. Identifying optimal biophysical, biochemical, and biological properties of the gel support matrix in these extreme experimental conditions is instrumental to achieving the best possible reproducibility and sensitivity. Here we systematically evaluate a range of gelling agents by using the yeast Saccharomyces cerevisiae as a model microbe. We find that carrageenan and Phytagel have superior optical clarity and reduced autofluorescence, crucial for high-resolution imaging and fluorescent reporter screens. Nutrient choice and use of refined Noble agar or pure agarose reduce the effective dose of numerous selective drugs by >50%, potentially enabling large cost savings in genetic screens. Using thousands of mutant yeast strains to compare colony growth between substrates, we found no evidence of significant growth or nutrient biases between gel substrates, indicating that researchers could freely pick and choose the optimal gel for their respective application and experimental condition.

Identifying compound efficacy targets in phenotypic drug discovery

The identification of the efficacy target(s) for hits from phenotypic compound screens remains a key step to progress compounds into drug development. In addition to efficacy targets, the characterization of epistatic proteins influencing compound activity often facilitates the elucidation of the underlying mechanism of action; and, further, early determination of off-targets that cause potentially unwanted secondary phenotypes helps in assessing potential liabilities. This short review discusses the most important technologies currently available for characterizing the direct and indirect target space of bioactive compounds following phenotypic screening. We present a comprehensive strategy employing complementary approaches to balance individual technology strengths and weaknesses.

Sec14-like phosphatidylinositol-transfer proteins and diversification of phosphoinositide signalling outcomes

The physiological functions of phosphatidylinositol (PtdIns)-transfer proteins (PITPs)/phosphatidylcholine (PtdCho)-transfer proteins are poorly characterized, even though these proteins are conserved throughout the eukaryotic kingdom. Much of the progress in elucidating PITP functions has come from exploitation of genetically tractable model organisms, but the mechanisms for how PITPs execute their biological activities remain unclear. Structural and molecular dynamics approaches are filling in the details for how these proteins actually work as molecules. In the present paper, we discuss our recent work with Sec14-like PITPs and describe how PITPs integrate diverse territories of the lipid metabolome with phosphoinositide signalling.

Phosphatidylinositol transfer proteins and instructive regulation of lipid kinase biology

Phosphatidylinositol is a metabolic precursor of phosphoinositides and soluble inositol phosphates. Both sets of molecules represent versatile intracellular chemical signals in eukaryotes. While much effort has been invested in understanding the enzymes that produce and consume these molecules, central aspects for how phosphoinositide production is controlled and functionally partitioned remain unresolved and largely unappreciated. It is in this regard that phosphatidylinositol (PtdIns) transfer proteins (PITPs) are emerging as central regulators of the functional channeling of phosphoinositide pools produced on demand for specific signaling purposes. The physiological significance of these proteins is amply demonstrated by the consequences that accompany deficits in individual PITPs. Although the biological problem is fascinating, and of direct relevance to disease, PITPs remain largely uncharacterized. Herein, we discuss our perspectives regarding what is known about how PITPs work as molecules, and highlight progress in our understanding of how PITPs are integrated into cellular physiology. This article is part of a Special Issue entitled Phosphoinositides.

Experimental and computational analysis of a large protein network that controls fat storage reveals the design principles of a signaling network

An approach combining genetic, proteomic, computational, and physiological analysis was used to define a protein network that regulates fat storage in budding yeast (Saccharomyces cerevisiae). A computational analysis of this network shows that it is not scale-free, and is best approximated by the Watts-Strogatz model, which generates “small-world” networks with high clustering and short path lengths. The network is also modular, containing energy level sensing proteins that connect to four output processes: autophagy, fatty acid synthesis, mRNA processing, and MAP kinase signaling. The importance of each protein to network function is dependent on its Katz centrality score, which is related both to the protein’s position within a module and to the module’s relationship to the network as a whole. The network is also divisible into subnetworks that span modular boundaries and regulate different aspects of fat metabolism. We used a combination of genetics and pharmacology to simultaneously block output from multiple network nodes. The phenotypic results of this blockage define patterns of communication among distant network nodes, and these patterns are consistent with the Watts-Strogatz model.

We discovered a large protein network that regulates fat storage in budding yeast. This network contains 94 proteins, almost all of which bind to other proteins in the network. To understand the functions of large protein collections such as these, it will be necessary to move away from one-by-one analysis of individual proteins and create computational models of entire networks. This will allow classification of networks into categories and permit researchers to identify key network proteins on theoretical grounds. We show here that the fat regulation network fits a Watts-Strogatz small-world model. This model was devised to explain the clustering phenomena often observed in real networks, but has not been previously applied to signaling networks within cells. The short path length and high clustering coefficients characteristic of the Watts-Strogatz topology allow for rapid communication between distant nodes and for division of the network into modules that perform different functions. The fat regulation network has modules, and it is divisible into subnetworks that span modular boundaries and regulate different aspects of fat metabolism. We experimentally examined communication between nodes within the network using a combination of genetics and pharmacology, and showed that the communication patterns are consistent with the Watts-Strogatz topology.

Rapid quantification of mutant fitness in diverse bacteria by sequencing randomly bar-coded transposons

Transposon mutagenesis with next-generation sequencing (TnSeq) is a powerful approach to annotate gene function in bacteria, but existing protocols for TnSeq require laborious preparation of every sample before sequencing. Thus, the existing protocols are not amenable to the throughput necessary to identify phenotypes and functions for the majority of genes in diverse bacteria. Here, we present a method, random bar code transposon-site sequencing (RB-TnSeq), which increases the throughput of mutant fitness profiling by incorporating random DNA bar codes into Tn5 and mariner transposons and by using bar code sequencing (BarSeq) to assay mutant fitness. RB-TnSeq can be used with any transposon, and TnSeq is performed once per organism instead of once per sample. Each BarSeq assay requires only a simple PCR, and 48 to 96 samples can be sequenced on one lane of an Illumina HiSeq system. We demonstrate the reproducibility and biological significance of RB-TnSeq with Escherichia coli, Phaeobacter inhibens, Pseudomonas stutzeri, Shewanella amazonensis, and Shewanella oneidensis. To demonstrate the increased throughput of RB-TnSeq, we performed 387 successful genome-wide mutant fitness assays representing 130 different bacterium-carbon source combinations and identified 5,196 genes with significant phenotypes across the five bacteria. In P. inhibens, we used our mutant fitness data to identify genes important for the utilization of diverse carbon substrates, including a putative d-mannose isomerase that is required for mannitol catabolism. RB-TnSeq will enable the cost-effective functional annotation of diverse bacteria using mutant fitness profiling.

A large challenge in microbiology is the functional assessment of the millions of uncharacterized genes identified by genome sequencing. Transposon mutagenesis coupled to next-generation sequencing (TnSeq) is a powerful approach to assign phenotypes and functions to genes. However, the current strategies for TnSeq are too laborious to be applied to hundreds of experimental conditions across multiple bacteria. Here, we describe an approach, random bar code transposon-site sequencing (RB-TnSeq), which greatly simplifies the measurement of gene fitness by using bar code sequencing (BarSeq) to monitor the abundance of mutants. We performed 387 genome-wide fitness assays across five bacteria and identified phenotypes for over 5,000 genes. RB-TnSeq can be applied to diverse bacteria and is a powerful tool to annotate uncharacterized genes using phenotype data.

The neuronal ceroid lipofuscinoses: Opportunities from model systems

The neuronal ceroid lipofuscinoses are a group of severe and progressive neurodegenerative disorders, generally with childhood onset. Despite the fact that these diseases remain fatal, significant breakthroughs have been made in our understanding of the genetics that underpin these conditions. This understanding has allowed the development of a broad range of models to study disease processes, and to develop new therapeutic approaches. Such models have contributed significantly to our knowledge of these conditions. In this review we will focus on the advantages of each individual model, describe some of the contributions the models have made to our understanding of the broader disease biology and highlight new techniques and approaches relevant to the study and potential treatment of the neuronal ceroid lipofuscinoses. This article is part of a Special Issue entitled: "Current Research on the Neuronal Ceroid Lipofuscinoses (Batten Disease)".

Protein homology reveals new targets for bioactive small molecules

MOTIVATION

The functional impact of small molecules is increasingly being assessed in different eukaryotic species through large-scale phenotypic screening initiatives. Identifying the targets of these molecules is crucial to mechanistically understand their function and uncover new therapeutically relevant modes of action. However, despite extensive work carried out in model organisms and human, it is still unclear to what extent one can use information obtained in one species to make predictions in other species.

RESULTS

Here, for the first time, we explore and validate at a large scale the use of protein homology relationships to predict the targets of small molecules across different species. Our results show that exploiting target homology can significantly improve the predictions, especially for molecules experimentally tested in other species. Interestingly, when considering separately orthology and paralogy relationships, we observe that mapping small molecule interactions among orthologs improves prediction accuracy, while including paralogs does not improve and even sometimes worsens the prediction accuracy. Overall, our results provide a novel approach to integrate chemical screening results across multiple species and highlight the promises and remaining challenges of using protein homology for small molecule target identification.

AVAILABILITY AND IMPLEMENTATION

Homology-based predictions can be tested on our website http://www.swisstargetprediction.ch.

CONTACT

david.gfeller@unil.ch or vincent.zoete@isb-sib.ch.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Large-scale chemical similarity networks for target profiling of compounds identified in cell-based chemical screens

Target identification is one of the most critical steps following cell-based phenotypic chemical screens aimed at identifying compounds with potential uses in cell biology and for developing novel disease therapies. Current in silico target identification methods, including chemical similarity database searches, are limited to single or sequential ligand analysis that have limited capabilities for accurate deconvolution of a large number of compounds with diverse chemical structures. Here, we present CSNAP (Chemical Similarity Network Analysis Pulldown), a new computational target identification method that utilizes chemical similarity networks for large-scale chemotype (consensus chemical pattern) recognition and drug target profiling. Our benchmark study showed that CSNAP can achieve an overall higher accuracy (>80%) of target prediction with respect to representative chemotypes in large (>200) compound sets, in comparison to the SEA approach (60–70%). Additionally, CSNAP is capable of integrating with biological knowledge-based databases (Uniprot, GO) and high-throughput biology platforms (proteomic, genetic, etc) for system-wise drug target validation. To demonstrate the utility of the CSNAP approach, we combined CSNAP's target prediction with experimental ligand evaluation to identify the major mitotic targets of hit compounds from a cell-based chemical screen and we highlight novel compounds targeting microtubules, an important cancer therapeutic target. The CSNAP method is freely available and can be accessed from the CSNAP web server (http://services.mbi.ucla.edu/CSNAP/).

Determining the targets of compounds identified in cell-based high-throughput chemical screens is a critical step for downstream drug development and understanding of compound mechanism of action. However, current computational target prediction approaches like chemical similarity database searches are limited to single or sequential ligand analyses, which limits their ability to accurately deconvolve a large number of compounds that often have chemically diverse structures. Here, we have developed a new computational drug target prediction method, called CSNAP that is based on chemical similarity networks. By clustering diverse chemical structures into distinct sub-networks corresponding to chemotypes, we show that CSNAP improves target prediction accuracy and consistency over a board range of drug classes. We further coupled CSNAP to a mitotic database and successfully determined the major mitotic drug targets of a diverse compound set identified in a cell-based chemical screen. We demonstrate that CSNAP can easily integrate with diverse knowledge-based databases for on/off target prediction and post-target validation, thus broadening its applicability for identifying the targets of bioactive compounds from a wide range of chemical screens.

Genes required for survival in microgravity revealed by genome-wide yeast deletion collections cultured during spaceflight

Spaceflight is a unique environment with profound effects on biological systems including tissue redistribution and musculoskeletal stresses. However, the more subtle biological effects of spaceflight on cells and organisms are difficult to measure in a systematic, unbiased manner. Here we test the utility of the molecularly barcoded yeast deletion collection to provide a quantitative assessment of the effects of microgravity on a model organism. We developed robust hardware to screen, in parallel, the complete collection of ~4800 homozygous and ~5900 heterozygous (including ~1100 single-copy deletions of essential genes) yeast deletion strains, each carrying unique DNA that acts as strain identifiers. We compared strain fitness for the homozygous and heterozygous yeast deletion collections grown in spaceflight and ground, as well as plus and minus hyperosmolar sodium chloride, providing a second additive stressor. The genome-wide sensitivity profiles obtained from these treatments were then queried for their similarity to a compendium of drugs whose effects on the yeast collection have been previously reported. We found that the effects of spaceflight have high concordance with the effects of DNA-damaging agents and changes in redox state, suggesting mechanisms by which spaceflight may negatively affect cell fitness.

Chemical Genomics-Based Antifungal Drug Discovery: Targeting Glycosylphosphatidylinositol (GPI) Precursor Biosynthesis

Steadily increasing antifungal drug resistance and persistent high rates of fungal-associated mortality highlight the dire need for the development of novel antifungals. Characterization of inhibitors of one enzyme in the GPI anchor pathway, Gwt1, has generated interest in the exploration of targets in this pathway for further study. Utilizing a chemical genomics-based screening platform referred to as the Candida albicans fitness test (CaFT), we have identified novel inhibitors of Gwt1 and a second enzyme in the glycosylphosphatidylinositol (GPI) cell wall anchor pathway, Mcd4. We further validate these targets using the model fungal organism Saccharomyces cerevisiae and demonstrate the utility of using the facile toolbox that has been compiled in this species to further explore target specific biology. Using these compounds as probes, we demonstrate that inhibition of Mcd4 as well as Gwt1 blocks the growth of a broad spectrum of fungal pathogens and exposes key elicitors of pathogen recognition. Interestingly, a strong chemical synergy is also observed by combining Gwt1 and Mcd4 inhibitors, mirroring the demonstrated synthetic lethality of combining conditional mutants of GWT1 and MCD4. We further demonstrate that the Mcd4 inhibitor M720 is efficacious in a murine infection model of systemic candidiasis. Our results establish Mcd4 as a promising antifungal target and confirm the GPI cell wall anchor synthesis pathway as a promising antifungal target area by demonstrating that effects of inhibiting it are more general than previously recognized.

Mitochondrial targets for pharmacological intervention in human disease

Over the past several years, mitochondrial dysfunction has been linked to an increasing number of human illnesses, making mitochondrial proteins (MPs) an ever more appealing target for therapeutic intervention. With 20% of the mitochondrial proteome (312 of an estimated 1500 MPs) having known interactions with small molecules, MPs appear to be highly targetable. Yet, despite these targeted proteins functioning in a range of biological processes (including induction of apoptosis, calcium homeostasis, and metabolism), very few of the compounds targeting MPs find clinical use. Recent work has greatly expanded the number of proteins known to localize to the mitochondria and has generated a considerable increase in MP 3D structures available in public databases, allowing experimental screening and in silico prediction of mitochondrial drug targets on an unprecedented scale. Here, we summarize the current literature on clinically active drugs that target MPs, with a focus on how existing drug targets are distributed across biochemical pathways and organelle substructures. Also, we examine current strategies for mitochondrial drug discovery, focusing on genetic, proteomic, and chemogenomic assays, and relevant model systems. As cell models and screening techniques improve, MPs appear poised to emerge as relevant targets for a wide range of complex human diseases, an eventuality that can be expedited through systematic analysis of MP function.

Systems cell biology

Systems cell biology melds high-throughput experimentation with quantitative analysis and modeling to understand many critical processes that contribute to cellular organization and dynamics. Recently, there have been several advances in technology and in the application of modeling approaches that enable the exploration of the dynamic properties of cells. Merging technology and computation offers an opportunity to objectively address unsolved cellular mechanisms, and has revealed emergent properties and helped to gain a more comprehensive and fundamental understanding of cell biology.

Off-target effects of the septin drug forchlorfenuron on nonplant eukaryotes

The septins are a family of GTP-binding proteins that form cytoskeletal filaments. Septins are highly conserved and evolutionarily ancient but are absent from land plants. The synthetic plant cytokinin forchlorfenuron (FCF) was shown previously to inhibit budding yeast cell division and induce ectopic septin structures (M. Iwase, S. Okada, T. Oguchi, and A. Toh-e, Genes Genet. Syst. 79:199-206, 2004, http://dx.doi.org/10.1266/ggs.79.199). Subsequent studies in a wide range of eukaryotes have concluded that FCF exclusively inhibits septin function, yet the mechanism of FCF action in nonplant cells remains poorly understood. Here, we report that the cellular effects of FCF are far more complex than previously described. The reported growth arrest of budding yeast cells treated with 1 mM FCF partly reflects sensitization caused by a bud4 mutation present in the W303 strain background. In wild-type (BUD4(+)) budding yeast, growth was inhibited at FCF concentrations that had no detectable effect on septin structure or function. Moreover, FCF severely inhibited the proliferation of fission yeast cells, in which septin function is nonessential. FCF induced fragmentation of budding yeast mitochondrial reticula and the loss of mitochondrial membrane potential. Mitochondria also fragmented in cultured mammalian cells treated with concentrations of FCF that previously were assumed to target septins only. Finally, FCF potently inhibited ciliation and motility and induced mitochondrial disorganization in Tetrahymena thermophila without apparent alterations in septin structure. None of these effects was consistent with the inhibition of septin function. Our findings point to nonseptin targets as major concerns when using FCF.